Isotope Systematics of Gypsum
and its Hydration Water
Nicholas Philip Evans
The Godwin Laboratory for Palaeoclimate Research,
Department of Earth Sciences
University of Cambridge
This dissertation is submitted for the degree of
Doctor of Philosophy
Clare College September 2018

To my parents
&
to my brother
***
It may be that the gulfs will wash us down:
It may be we shall touch the Happy Isles,
And see the great Achilles, whom we knew.
Tho’ much is taken, much abides; and tho’
We are not now that strength which in old days
Moved earth and heaven, that which we are, we are;
One equal temper of heroic hearts,
Made weak by time and fate, but strong in will
To strive, to seek, to find, and not to yield.
Alfred, Lord Tennyson

Declaration
I hereby declare that except where specific reference is made to the work of others, the
contents of this dissertation are original and have not been submitted in whole or in part
for consideration for any other degree or qualification in this University, or any other. This
dissertation is the result of my own work and includes nothing which is the outcome of
work done in collaboration, except where specifically indicated in the text. This disserta-
tion contains fewer than 275 numbered pages of which not more than 225 pages are text,
illustrations and bibliography.
Nicholas Philip Evans
September 2018

Acknowledgements
First and foremost, I would like to thank my supervisor, David Hodell. He has allowed
me the freedom to pursue any research ideas that I wanted to investigate. His drive to step
on untouched ground and to explore new ideas has been inspiring, and has taught me a lot
about how to approach both scientific problems and life.
Secondly, I want to thank Fernando Gázquez and Thomas Bauksa. Experiencing their
passion for science has been particularly motivating, and our collaboration in the lab has
provided hours of fun. They have both been important mentors to me, and their guidance
has been been invaluable.
I also thank Rachel Flecker and Edward Tipper for evaluating this thesis.
Finally, want to thank my family, to whom I dedicate this book. Without their support,
this work would be a castle in the sky. I will be eternally grateful for their love.

Publication List
Chapter 2
• Gázquez, F., Evans, N. P., Hodell, D. A., 2017. Precise and accurate isotope fractiona-
tion factors (α17O, α18O and αD,) for water and CaSO4 ·2H2O (gypsum). Geochim-
ica et Cosmochimica Acta 198, 259–270.
• Gázquez, F., Mather, I., Rolfe, J., Evans, N. P., Herwartz, D., Staubwasser, M., Hodell,
D. A., 2015. Simultaneous analysis of 17O/16O, 18O/16O and 2H/1H of gypsum
hydration water by cavity ring-down laser spectroscopy. Rapid Communications in
Mass Spectrometry 29 (21), 1997–2006.
Chapter 3
• Evans, N. P., Bauska, T. K., Gázquez-Sánchez, F., Brenner, M., Curtis, J. H., Hodell,
D. A., 2018. Quantification of drought during the collapse of the classic Maya civi-
lization. Science 361 (6401), 498–501.
Chapter 5
• Evans, N. P., Turchyn, A. V., Gázquez, F., Bontognali, T. R., Chapman, H. J., Hodell,
D. A., 2015. Coupled measurements of δ18O and δD of hydration water and salinity
of fluid inclusions in gypsum from the Messinian Yesares Member, Sorbas Basin (SE
Spain). Earth and Planetary Science Letters 430, 499–510.
Collaborations
Outside of the remit of this thesis, contributions have been made to a number of different
projects. These include:
1. geochemical analysis of carbonate veins in the oceanic crust (Turchyn, A.V.);
2. reconstruction of the palaeohydrologic changes in lakes of the Bolivian Altiplano
(Bauska, T. K.; Gázquez-Sánchez, F; Baker, P. A.);
x3. investigating the potential on-going subsurface dolomite formation at DSDP Leg 42A
site 374 (McKezsie, J. A.), and;
4. using stable isotopes of gypsum hydration water to ascertain the role of water con-
densation in the formation of subaerial gypsum speleothems (Gázquez-Sánchez, F).
Abstract
Triple oxygen and hydrogen isotope analysis of the structurally-bound water in gypsum can
provide a direct measure of past hydrologic variability. This thesis presents the development
of the water extraction and isotopic measurement procedures, the calculation of the gypsum-
water isotope fractionation factors, and the application of the method to constrain the palaeo-
hydrologic conditions in two temporally and geographically disparate sites.
Measurement of the isotopic composition of gypsum hydration water is used to examine
the hydrological changes that occurred during the Terminal Classic Drought of the Maya
lowlands (~800-1000 CE), coincident with the period when the Classic Maya Civilization
of Mesoamerica collapsed. The data provide a complete and direct archive of hydrological
conditions that have previously been limited to ice core records. Mean annual rainfall is
shown to have decreased by between 41% and 54%, with intervals of up to 70%, compared
to present-day conditions. This study has also shown for the first time that relative humidity
was 2%-7% lower during the Terminal Classic Drought compared to today.
The methodology is also applied to the massive gypsum deposits in the marginal and
deep basins of the Mediterranean to interpret the chemical evolution of parent water bodies
during the Messinian Salinity Crisis (5.97-5.3 Ma). By combining the measurement of
gypsum hydration water with other traditional (e.g. strontium) and novel (e.g. calcium
and barium) isotope tracers, the hydrological changes during the deposition of Primary
Lower Gypsum units of the Sorbas Basin in southeastern Spain, the Upper Gypsum units
of Sicily, and deep basin deposits have been constrained. The results indicate that all
deposits experienced a significant freshwater contribution to the mother fluids from which
they formed. It is proposed that obliquity-controlled sea level and eccentricity-modulated
precession, superimposed on longer-term tectonic restriction of the Mediterranean-Atlantic
exchange, together controlled the varying depositional environments during the formation
of the Messinian Salt Giant.
This thesis demonstrates that the analysis of gypsum hydration water is a powerful tool
for palaeoclimate reconstruction. The methodology can be applied to gypsum (and other
hydrated minerals) in a wide range of settings across geological space and time, providing
a rich source of information about the environmental conditions under which they formed.
Nicholas P. Evans
Isotope Systematics of Gypsum and its Hydration Water

Contents
Dedication iii
Declaration v
Acknowledgements vii
Publication List viii
Abstract xi
Contents xiii
List of Figures xix
List of Tables xxiii
List of Abbreviations, Nomenclature and Units xxx
1 Introduction 1
1.1 Stable water isotopologues . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Statement of the problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Limitations of traditional proxy archives . . . . . . . . . . . . . . 3
1.2.2 Analytical limitations . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Novel approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Objectives and thesis outline . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Calcium Sulfate Systematics 13
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Calcium sulfate formation . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.1 Calcium sulfate solubility and thermodynamic stability . . . . . . . 14
xiv Contents
2.2.2 Natural ion sources for calcium sulfate formation . . . . . . . . . . 16
2.3 Hydration water extraction and analysis procedure . . . . . . . . . . . . . . 22
2.3.1 The vacuum assembly . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.2 Extraction of GHW . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3.3 Isotopic measurement of GHW by CRDS . . . . . . . . . . . . . . 25
2.3.4 Internal and external precision of the CRDS/WASP system . . . . . 26
2.3.5 Discussion of the WASP/CRDS setup . . . . . . . . . . . . . . . . 31
2.3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O . . . . . . . . . . 33
2.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.4.2.1 Hydration of anhydrous CaSO4 . . . . . . . . . . . . . . 35
2.4.2.2 Mixing CaCl2 and Na2SO4 solutions . . . . . . . . . . . 36
2.4.2.3 Gypsum precipitation in brines . . . . . . . . . . . . . . 37
2.4.2.4 Isotopic analyses and calculation of fractionation factors . 37
2.4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4.3.1 Hydration of anhydrite at low salinities . . . . . . . . . . 38
2.4.3.2 Gypsum precipitation from mixing of CaCl2 and Na2SO4
solutions . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4.3.3 Gypsum precipitation from brines . . . . . . . . . . . . . 42
2.4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4.4.1 Effect of temperature and precipitation rate on the isotope
fractionation factors in gypsum hydration water . . . . . 42
2.4.4.2 Effect of salinity on the isotope fractionation factors in
gypsum hydration water . . . . . . . . . . . . . . . . . . 45
2.4.4.3 Triple oxygen isotope fractionation in gypsum hydration
water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5 Gypsum-water isotopic exchange . . . . . . . . . . . . . . . . . . . . . . 48
2.5.1 Gypsum dehydration . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5.2 Thermal dehydration of gypsum monitored by in situ Raman spec-
troscopy and TGA/DSC . . . . . . . . . . . . . . . . . . . . . . . 50
2.5.2.1 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.5.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.5.2.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.5.2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . 60
Contents xv
2.5.3 Gypsum recrystallisation and isotope exchange . . . . . . . . . . . 61
3 Quantification of drought during the collapse of the classic Maya civilisation 65
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.2.1 Sediment cores and sampling . . . . . . . . . . . . . . . . . . . . . 68
3.2.2 Gypsum Hydration Water (GHW) . . . . . . . . . . . . . . . . . . 68
3.2.3 Precipitation, meteoric and lake water samples . . . . . . . . . . . 71
3.2.4 Biogenic carbonate measurements . . . . . . . . . . . . . . . . . . 71
3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.4 Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.4.1 Transient model details . . . . . . . . . . . . . . . . . . . . . . . . 73
3.4.2 Transient model isotope and ionic mass balance . . . . . . . . . . . 75
3.4.3 Transient model parameterisation . . . . . . . . . . . . . . . . . . 80
3.4.4 Age models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.5.1 Monte Carlo modelling scenarios . . . . . . . . . . . . . . . . . . 92
3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4 The Messinian Salinity Crisis and methods of study 95
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.2 Stratigraphy of the MSC . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.2.1 Mediterranean onshore successions . . . . . . . . . . . . . . . . . 98
4.2.2 Mediterranean offshore successions . . . . . . . . . . . . . . . . . 105
4.3 Major / trace elements and isotopic tracers . . . . . . . . . . . . . . . . . . 107
4.3.1 Gypsum Hydration Water . . . . . . . . . . . . . . . . . . . . . . 108
4.3.2 Fluid Inclusions Analysis . . . . . . . . . . . . . . . . . . . . . . . 109
4.3.3 Sulfur isotopes and oxygen isotopes of sulfate . . . . . . . . . . . . 111
4.3.4 Strontium Isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.3.5 Calcium Isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.3.6 Barium Isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.3.7 Trace element chemistry . . . . . . . . . . . . . . . . . . . . . . . 121
4.3.8 Carbonate sample preparation . . . . . . . . . . . . . . . . . . . . 121
4.3.9 X-ray diffraction (XRD) of carbonates . . . . . . . . . . . . . . . . 122
4.3.10 Total inorganic carbon, δ18Ocarb and δ13C of bulk carbonates . . . . 122
4.3.11 Scanning electron microscopy (SEM) . . . . . . . . . . . . . . . . 123
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
xvi Contents
5 Coupled measurements of δ18O and δD of hydration water and salinity of fluid
inclusions in gypsum from the Messinian Yesares Member, Sorbas Basin (SE
Spain) 125
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.2 Geological setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.3.1 Field locations and sample collection methods . . . . . . . . . . . 128
5.3.2 Geochemical methodology . . . . . . . . . . . . . . . . . . . . . . 128
5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.4.1 δ 34S and δ18O of sulphate . . . . . . . . . . . . . . . . . . . . . . 129
5.4.2 Sr isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.4.3 XRD, wt% calcite and wt% dolomite of bulk carbonates . . . . . . 129
5.4.4 δ18O and δ13C of bulk carbonates . . . . . . . . . . . . . . . . . . 130
5.4.5 Gypsum hydration waters . . . . . . . . . . . . . . . . . . . . . . 131
5.4.6 Microthermometric analysis . . . . . . . . . . . . . . . . . . . . . 132
5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.5.1 Strontium, sulfur and oxygen isotopes of gypsum . . . . . . . . . . 135
5.5.2 Gypsum hydration water and microthermometry . . . . . . . . . . 137
5.5.3 Decoupling of the source of ions and water during gypsum formation 139
5.5.4 Deposition of the lower marl unit . . . . . . . . . . . . . . . . . . 139
5.5.5 Formation of the supercone-marl unit . . . . . . . . . . . . . . . . 140
5.5.6 Climate control of gypsum–marl deposition . . . . . . . . . . . . . 142
5.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6 Geochemical evidence for meteoric water, orbital and eustatic control of basin
hydrology during the first stage of the Messinian Salinity Crisis 147
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.2 Geological setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.3 Sample collection methods . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.4.1 Gypsum Hydration Water (GHW) . . . . . . . . . . . . . . . . . . 155
6.4.2 Sr isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.4.3 Ca isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
6.4.4 δ 34S and δ18O of sulfate . . . . . . . . . . . . . . . . . . . . . . . 160
6.4.5 Microthermometric analysis . . . . . . . . . . . . . . . . . . . . . 161
6.4.6 XRD, wt% calcite and wt% dolomite of bulk carbonates . . . . . . 161
6.4.7 δ18Ocarb and δ13C of bulk carbonates . . . . . . . . . . . . . . . . 161
Contents xvii
6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
6.5.1 Brine composition during PLG deposition . . . . . . . . . . . . . . 165
6.5.2 Orbital, tectonic and glacio-eustatic control on basin hydrology . . 167
6.5.2.1 Precessional scale orbital and tectonic control on basin
hydrology . . . . . . . . . . . . . . . . . . . . . . . . . 167
6.5.2.2 Glacio-eustatic control on basin hydrology . . . . . . . . 171
6.5.3 A mechanism for PLG formation: an orbitally forced dissolution flux 174
6.5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
7 Upper Evaporites of the Messinian Salinity Crisis 177
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
7.2 Geological setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.1 Onshore deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.2 Offshore deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.2.1 DSDP Site 372 . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.2.2 ODP Site 654 . . . . . . . . . . . . . . . . . . . . . . . 181
7.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
7.3.1 Gypsum Hydration Water (GHW) . . . . . . . . . . . . . . . . . . 181
7.3.2 Microthermometric analysis . . . . . . . . . . . . . . . . . . . . . 183
7.3.3 Sr isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
7.3.4 Ca isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.3.5 Ba isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
7.3.6 δ 34S and δ18O of sulfate . . . . . . . . . . . . . . . . . . . . . . . 187
7.3.7 δ18Ocarb and δ13C of bulk carbonates . . . . . . . . . . . . . . . . 189
7.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
7.4.1 Isotopic comparison between the onshore Upper Gypsum and Primary
Lower Gypsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
7.4.2 Semi-closed system behaviour during Upper Gypsum formation . . 192
7.4.3 Depositional environment of DSDP/ODP sites 654 and 372 . . . . 196
7.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
References 201
A Tables for Section 2.3 225
B Tables for Section 2.4 231
C Tables for Chapter 3 239
xviii Contents
D Tables for Chapters 5, 6 and 7 251
List of Figures
1.1a Evolution of d-excess vs. δ 18O and 17O-excess vs. δ 18O during evaporation 4
1.1b . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Structural projections of calcium sulfate phases . . . . . . . . . . . . . . . 15
2.2 Solubility curves of gypsum, bassanite, and anhydrite . . . . . . . . . . . . 17
2.3 Calcium sulfate stability . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 The evaporation pathway of seawater in a marine saltwork . . . . . . . . . 20
2.5 Schematic of the Water Analyser Sample Preparation device . . . . . . . . 23
2.6 Typical pressure profile recorded during the extraction of GHW . . . . . . 24
2.7 δ18O, δD, 17O-excess, and d-excess values from repeat analysis of a gypsum
standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.8 Cross-plot of δ18O analysed by CRDS/MCM vs IRMS . . . . . . . . . . . 30
2.9 Cross-plot of 17O-excess in synthetic and natural gypsum analysed by CRDS
and IRMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.10 Gypsum-water isotope fractionation factors at different temperatures . . . . 39
2.11 Gypsum-water isotope fractionation factors at different temperatures and at
different initial concentrations of Ca2+ and SO42− . . . . . . . . . . . . . . 41
2.12 Gypsum-water isotope fractionation factors at different salinities . . . . . . 46
2.13 Raman spectra of gypsum, bassanite, and γ-anhydrite . . . . . . . . . . . . 53
2.14 TGA and DSC profiles of gypsum at different heating rates . . . . . . . . . 54
2.15 TGA and DSC profiles compared with Raman spectra . . . . . . . . . . . . 56
2.16 Isothermal analysis of gypsum dehydration over the temperature range 60°C
to 120°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.17 Gypsum exchange experiments . . . . . . . . . . . . . . . . . . . . . . . . 62
3.1 Map of the Maya Lowlands displaying the locations of proxy climate archives 67
3.2 Water isotopes during drought periods compared to modern water isotopes
of Lake Chichancanab . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
xx List of Figures
3.3 Summary of GHW data obtained from Lake Chichancanab . . . . . . . . . 70
3.4 Measured δ18O and δD of rain, surface water, and ground water . . . . . . 72
3.5 Comparison of measured local meteoric water, modern lake water, and paleo-
lake water data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.6 Transient model summary . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.7 Diagrammatic representation of the two-box model . . . . . . . . . . . . . 77
3.8 Transient model climate forcing . . . . . . . . . . . . . . . . . . . . . . . 79
3.9 Transient model precipitation and relative humidity scenarios . . . . . . . . 81
3.10 The amount effect relationship . . . . . . . . . . . . . . . . . . . . . . . . 85
3.11 Bayesian age model analysis . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.12 Selection of palaeoclimate records from the Maya Lowlands . . . . . . . . 89
3.13 Simulated hydrologic changes of Lake Chichancanab . . . . . . . . . . . . 91
4.1 Modern topographical map of the Mediterranean region . . . . . . . . . . . 96
4.2 Main features of the MSC and the Mediterranean’s exchange history . . . . 97
4.3 Late Miocene chronostratigraphy of the Mediterranean . . . . . . . . . . . 99
4.4 Chronostratigraphy and astronomical tuning of the Upper and Lower Gypsum101
4.5 Diagram of Primary Lower Gypsum Cycle 10 . . . . . . . . . . . . . . . . 102
4.6 Diagram of Upper Gypsum Cycle 4 . . . . . . . . . . . . . . . . . . . . . 104
4.7 Cross plot of δ18O versus salinity . . . . . . . . . . . . . . . . . . . . . . . 110
4.8 Schematic of calcium and strontium cycling in the modern surface environ-
ment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.9 Compilation of calcium isotope data from differing sources . . . . . . . . . 117
5.1 Map of the Messinian evaporites in the Mediterranean . . . . . . . . . . . . 127
5.2 δ34S versus δ18O of sulfate of the Yesares Gypsum . . . . . . . . . . . . . 129
5.3 Sr isotope curve during the Messinian in the Mediterranean Sea and Global
Ocean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.4 Carbonate marl data from the base of PLG Cycle 10 . . . . . . . . . . . . . 131
5.5 Cross plot of carbonates marls δ18O and δ13C . . . . . . . . . . . . . . . . 132
5.6 Predicted Messinian brine mother water values in δ18O vs δD and δ18O vs
d-excess space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.7 Schematic diagram of Cycle 10 containing δ18O, wt% NaCl and d-excess . 134
5.8 Coupled measurements of δ18O and δD of hydration water and salinity of
fluid inclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.9 SEM analyses of the carbonate minerals present in the Yesares Marls . . . . 141
5.10 Locations of the supercone samples . . . . . . . . . . . . . . . . . . . . . 143
List of Figures xxi
6.1 Map and cyclostratigraphy of the PLG in the Eastern Mediterranean . . . . 149
6.2 Cross-plot of palaeo-water data deduced from GHW with measured modern
meteoric water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.3 Mean intra-cycle δ18O, 17O-excess, d-excess and salinity data plotted against
orbitally tuned cycle age . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.4 Isotope data plotted relative to sample high within each cycle of the Yesares
section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.5 High-resolution intra-cycle isotope data from PLG cycles 2, 7, 10 and 12 . . 157
6.6 Mean intra-cycle 87Sr/86Sr, δ44/40Ca, δ18Ocarb and marl dolomite-calcite
ratio data plotted against orbitally tuned cycle age . . . . . . . . . . . . . . 163
6.7 Mean intra-cycle δ 34S and δ18OSO4 data plotted against orbitally tuned
cycle age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6.8 Histogram of Tmice measured in PLG fluid inclusions . . . . . . . . . . . . 166
6.9 Cross-plot of δ18O vs δ13C of carbonates from the marl intervals within the
Yesares Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
6.10 Mean intra-cycle δ44/40Ca data of the gypsum and interbedded carbonates
plotted against orbitally tuned cycle age . . . . . . . . . . . . . . . . . . . 170
6.11 Comparison of 87Sr/86Sr data from the Sorbas and Vena del Gesso Basins . 172
6.12 Evolution of salinity and gypsum saturation at different [Ca2+] and [SO42−]
concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
7.1 Stratigraphic section of the Upper Gypsum . . . . . . . . . . . . . . . . . . 179
7.2 Core logs of ODP Leg 107 site 654 and DSDP Leg 42A site 372 . . . . . . 182
7.3 Comparison of paleo-water data from the PLG, Upper Gypsum and deep
basin sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.4 Comparison of mean intra-cycle δ18O, 17O-excess, d-excess and salinity
data from the Upper and Lower Gypsum units . . . . . . . . . . . . . . . . 187
7.5 High-resolution intra-cycle isotope data from Upper Gypsum cycle 4 and
ODP Leg 107 site 654 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
7.6 Comparison of mean intra-cycle δ44/40Ca, δ138/134Ba and 87Sr/86Sr data
from the Upper and Lower Gypsum units . . . . . . . . . . . . . . . . . . 191
7.7 Comparison of δ 34S vs. δ18O of sulfate of the Primary Lower Gypsum and
Upper Gypsum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
7.8 Comparison of mean intra-cycle δ 34S and δ18OSO4 data from the Upper and
Lower Gypsum units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
7.9 Cross-plot of barium isotopes, strontium, calcium isotopes measured in the
same sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

List of Tables
2.1 Isotope fractionation factors between GHW and its mother solution obtained
by the hydration of anhydrite . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.2 Isotope fractionation factors between GHW and its mother solution obtained
by mixing of CaCl2 and Na2SO4 solutions . . . . . . . . . . . . . . . . . . 40
2.3 Isotope fractionation factors between GHW and its mother solution obtained
by hydration of anhydrite with different concentrations of NaCl . . . . . . . 47
2.4 Reference table of experimental conditions used for TGA, DSC and RS
analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.1 Transient model parameters . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.2 Steady-state parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.1 Barium isotope purification procedure . . . . . . . . . . . . . . . . . . . . 121
A.1a Isotopic analysis of hydration water of a gypsum standard (NEWGYP) ex-
tracted using the WASP and analyzed by CRDS and IRMS . . . . . . . . . 226
A.2 Memory effect tests and experiments of vacuum leakage in the WASP lines
during GHW extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
A.3 Isotopic analysis of hydration water of a gypsum standard (NEWGYP) and
natural samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
A.4 Isotopic composition of water samples distilled using the WASP and anal-
ysed by CRDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
B.1a Fractionation factor between gypsum hydration water and the mother solu-
tion in experiments of hydration of anhydrite at different temperatures . . . 232
B.2a Fractionation factor between gypsum hydration water and the mother solu-
tion in mixing experiments at different ion concentrations . . . . . . . . . . 234
B.3 Fractionation factor between gypsum hydration water and the mother solu-
tion at different concentrations of NaCl . . . . . . . . . . . . . . . . . . . 236
xxiv List of Tables
B.4 XRD analysis of solid recovered in fractionation factor experiments . . . . 237
B.5 Analysis of hydration water of gypsum precipitated from saline brines . . . 237
C.1 Isotopic composition of samples recovered from core CH1 7-III-04, Lake
Chichancanab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
C.2 Paleo-lake temperature calculations . . . . . . . . . . . . . . . . . . . . . 240
C.3a Output table of Bayesian age-depth analysis . . . . . . . . . . . . . . . . . 241
C.4a Probability of Drought derived from Bayesian analysis . . . . . . . . . . . 243
C.5a Stable isotope ratios of lake waters from the Yucatán Peninsula . . . . . . . 245
C.6a Stable isotope ratios of river and freshwaters from the Yucatán Peninsula . . 248
C.7 Stable isotope ratios of rainwater from the Yucatán Peninsula . . . . . . . . 249
C.8 Monte Carlo modelling solutions. . . . . . . . . . . . . . . . . . . . . . . 250
D.1a Isotopic composition of measured GHW of PLG units . . . . . . . . . . . . 252
D.2a Isotopic composition of measured GHW of UG units . . . . . . . . . . . . 255
D.2b Isotopic composition of measured GHW of deep basin units . . . . . . . . 256
D.3 Isotopic composition of measured GHW of modern samples . . . . . . . . 258
D.4a Strontium, calcium and barium isotopic results from PLG units . . . . . . . 259
D.5a Strontium, calcium and barium isotopic results from UG units . . . . . . . 262
D.5b Strontium, calcium and barium isotopic results from deep basin units . . . . 263
D.6a Fluid inclusion salinity estimates, Sr/Ca and sulfur isotope results obtained
from PLG units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
D.7a Fluid inclusion salinity estimates, Sr/Ca and sulfur isotope results obtained
from UG units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
D.7b Sulfur isotope results obtained from deep basin units . . . . . . . . . . . . 268
D.8a Isotopes of PLG carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . 270
D.9a Oxygen and hydrogen isotope data of Mediterranean freshwaters . . . . . . 272
List of Abbreviations, Nomenclature and
Units
List of Abbreviations
AMOC Atlantic meridional overturning circulation
BSE bulk silicate Earth
CE common (or current) era
CIESM Mediterranean Science Commission (Commission Internationale pour l’Exploration
Scientifique de la Méditerranée)
CP chemically pure
CRDS cavity ring-down laser spectroscopy
CRM confocal Raman microscopy
DSC differential scanning calorimetry
DSDP Deep Sea Drilling Project
E/I evaporation/inflow
E/P evaporation/precipitation
EDX energy-dispersive X-ray spectroscopy
EPS extracellular polymeric substance
E east
FIM fluid inclusion microthermometry
xxvi List of Abbreviations, Nomenclature and Units
GC gas chromatography
GHW gypsum hydration water
GRA gamma-ray attenuation
HR-ICP-MS high resolution inductively-coupled plasma mass spectrometry
IAEA International Atomic Energy Agency
ICP-OES inductively-coupled plasma optical emission spectroscopy
ICSD Inorganic Crystal Structure Database
IRMS isotope ratio mass spectrometry
IR infra-red
LN2 liquid nitrogen
LE Lower Evaporites
LLNL Lawrence Livermore National Laboratory
LMWL local meteroric water line
MC-ICP-MS multiple-collector inductively-coupled plasma mass spectrometry
MCM micro-combustion module
MES Messinian Erosion Surface
MOW Mediterranean outflow water
MSA methyl-sulfonic acid
MSC Messinian Salinity Crisis
MWL meteroric water line
NARR North American regional reanalysis
ND neutron diffraction
NMR nuclear magnetic resonance
N north
List of Abbreviations, Nomenclature and Units xxvii
ODP Ocean Drilling Program
OD outer diameter
P/E precipitation/evaporation
PLG primary lower gypsum
RHn normalised relative humidity
RLG resedimented lower gypsum
RS Raman spectroscopy
SC supercone
SEM scanning electron microscopy
SE south-east
SRM standard reference material
S south
TC/EA high temperature conversion elemental analyzer
TCP Terminal Classic Period
TGA thermogravimetric analysis
TIMS thermal ionisation mass spectrometry
Tmice final melting temperature of ice
T temperature
UA Upper Abad
UE Upper Evaporites
UG Upper Gypsum
VCDT Vienna-Canyon Diablo Troilite
VPDB Vienna Pee Dee Belemnite
VSMOW Vienna Standard Mean Ocean Water
xxviii List of Abbreviations, Nomenclature and Units
WASP water analyser sample preparation
W west
XRD X-ray diffraction
Nomenclature
α fractionation factor
αeq equilibrium fractionation
αkin kinetic fractionation
∆ gross change in a variable
δ incremental change in a variable
ε transformed fraction as a function of total mass loss
λ1/2 half life
Ω electrical resistance
θ the triple-isotope exponent; ln
(
α 17O
)
/ ln
(
α 18O
)
mbs f metres below sea floor
R2 correlation coefficient
D Deuterium (2H)
d-excess Deuterium-excess
SIgyp saturation index of gypsum
F flux
V atmospheric water vapour
w wind-induced turbulence
Xe fraction of water loss by evaporation with respect to the inflow of a system
(0 = no evaporation, 1 = all water is lost to evaporation)
List of Abbreviations, Nomenclature and Units xxix
Prefixes
M mega (106)
k kilo (103)
c centi (10−2)
m milli (10−3)
µ micro (10−6)
n nano (10−9)
p pico (10−12)
List of Units
Å angstrom (0.1 nanometer)
A ampere
◦C temperature in degrees centigrade
a annum
g gram
h hour
J joule
l litre
M Molar
mol mole
m metre
mbar unit of pressure; standard atmospheric pressure at sea level is defined as
1013.25 mbar
min minute
N normality
xxx List of Abbreviations, Nomenclature and Units
pH logarithm to the base 10 of the reciprocal of the hydrogen ion concentration
ppm parts per million
ppt parts per thousand
rpm revolutions per minute
S Siemens
s second
V volt
wt% weight percent
yr year
ϕ porosity
% per cent
h per mille, parts per thousand


Chapter 1
Introduction
1.1 Stable water isotopologues
Measurements of stable water isotopologues (e.g. H216O, H217O, H218O, and HD16O)
have been utilised since the 1950s with the aim of improving our understanding of the
hydrological cycle and its links to climate (Urey et al., 1951; Epstein and Mayeda, 1953;
Friedman, 1953; Emiliani, 1955; Craig, 1961; Dansgaard, 1964). Water isotopologues differ
by their mass and molecular symmetry, resulting in slightly different behaviours during
phase transitions. Fractionation processes that lead to the enrichment or depletion of the
heavy isotopes (i.e. 18O and D) during phase changes can be quantified if isotope ratios
are expressed relative to a standard. The isotopic composition is traditionally given as a
deviation from the ocean reservoir, and the delta (δ ) notation is used to quantify stable
isotopes as relative ratios expressed in parts per thousand (h):
δ =
(
RSample
RStandard
−1
)
×1000
RSample and RStandard are the molar ratio of the heavy to light isotopes in the sample and
in the Vienna Standard Mean Ocean Water (VSMOW) standard, respectively (Craig, 1961;
Dansgaard, 1964).
At first order, variations in oxygen (δ17O,δ18O) and hydrogen (δD) isotope ratios occur
due to the difference in saturation vapour pressures between the light and heavy isotopes,
resulting in an equilibrium fractionation. The larger diffusivity of light isotopes with respect
to heavy isotopes during evaporation can also results in a kinetic (non-equilibrium) frac-
tionation (Gat, 1996). There is an added value in combining δ18O and δD measurements
because kinetic effects have a larger influence on δ18O than on δD (Merlivat, 1978; Merlivat
and Nief, 1967; Risi et al., 2010). The effects of non-equilibrium fractionation are quantified
2 Introduction
as the Deuterium excess (d-excess) (Dansgaard, 1964):
d -excess= δD−8·δ18O
where 8 is the reference slope of the meteoric water line (MWL) worldwide. Non-equilibrium
fractionation during evaporation causes the fractionation ratio of the two molecules to devi-
ate from the ~1:8 ratio observed under equilibrium conditions. The extent of this deviation
is influenced by factors including the relative humidity gradient above the water surface,
air and water temperature, and evaporative cooling and turbulence on the water surface
(Merlivat and Jouzel, 1979; Cappa, 2003; Gázquez et al., 2018) (Fig. 1.1 A-E).
The measurements of triple oxygen isotopes (18O, 17O, 16O) allows derivation of the
parameter 17O-excess (Barkan and Luz, 2007; Luz and Barkan, 2010; Steig et al., 2014),
expressed in per meg (10−3h):
17O-excess=
(
ln
[
δ17O+1
]−0.528ln[δ18O+1])
The value of 0.528 has been proposed to describe the δ17O and δ18O global MWL
relationship (Luz and Barkan, 2010). The trajectory of δ18O and 17O-excess in evaporated
water is relatively insensitive to temperature and salinities below 100 g l-1 (Luz and Barkan,
2010; Passey et al., 2014); however, it is significantly affected by other parameters such as
atmospheric relative humidity (Fig. 1.1 F-I) (Gázquez et al., 2018). As δ17O, δ18O and δD
show different sensitivities to equilibrium and kinetic fractionation processes, the simultan-
eous measurement of triple oxygen and hydrogen isotopes (and utilisation of the derived
d-excess and 17O-excess parameters) in climatic archives can provide detailed information
on climate conditions of the past.
1.2 Statement of the problem
Geochemical proxies derived from climatic archives are used to reconstruct past climatic
variations. To this end, examination of the mineral chemistry of substrates that derive a
common atom with the water from which they form should be directly linked to the changes
in water balance of the regional environment during formation. An ideal proxy should
vary quantitatively and with respect to variations in the climatic parameter of interest.
Derived palaeoclimatic interpretations from traditional climate proxy archives are often
under-constrained, however, due to a combination of (i) assumptions that must be made
to isolate the parameter of interest, and (ii) analytical constraints. Below, the limitations of
the traditional methodologies used to reconstruct palaeoclimate variations are described, and
1.2 Statement of the problem 3
a new method that permits the simultaneous measurements of triple oxygen and hydrogen
isotopes in hydrated minerals is proposed.
1.2.1 Limitations of traditional proxy archives
The majority of studies that attempt to quantify past climate changes use either oxygen-
bearing proxies (such as carbonate) or hydrogen-bearing proxies (such as plant waxes and
alkenones). Few studies to date have combined the two isotope systems, because the ma-
terials used for analysis preclude simultaneous measurement of the multiple isotopologues
of water. Indeed, the relative influence of climatic variables such as temperature, relative
humidity and wind turbulence on a water body are impossible to deconvolve from the
individual analysis of δ18O or δD (Fig. 1.1). Even when δ18O and δD are combined,
the relationship between the isotope pairs is complicated if two or more of the climate
variables are unknown, preventing a unique, quantitative interpretation of the environmental
parameters influencing the water body. Additionally, in evaporative systems such as marine
sabkhas, the isotope ‘salt effect’ affects the activity ratios between free water and water in
ionic hydration shells, thus producing different hydrogen and oxygen isotope fractionations
(Feder and Taube, 1952; Taube, 1954; Sofer and Gat, 1972; Driesner et al., 2000; Herwartz
et al., 2017).
The isotopic content of the proxies themselves may also be controlled by:
i. Temperature-dependent fractionation factors (Urey, 1947; Epstein et al., 1953; Ep-
stein and Mayeda, 1953; Friedman and O’Neil, 1977; Chappell and Shackleton, 1986);
ii. Mineralogy (e.g. Tarutani et al., 1969);
iii. The pH of the solution (Deuser et al., 1968; Zeebe and Wolf-Gladrow, 2001);
iv. Any kinetic effects manifested during mineral precipitation (McConnaughey, 1989)
and, in the case of carbonate minerals, additional biochemical ‘vital’ effects (Gat,
1996).
Because both palaeoclimatic archives and the water from which they form are sensitive
to several climatic and non-climatic processes and conditions, some of which may be un-
known, isotopic datasets can often be underdetermined.
Combined analysis of δ17O, δ18O and δD is a powerful method to estimate hydrologic
changes quantitatively because hydrogen and triple oxygen isotopes each undergo slightly
different fractionation during evaporation, leading to changes in the derived d-excess and
17O-excess parameters (Surma et al., 2015; Herwartz et al., 2017; Gázquez et al., 2017b,
4 Introduction
50
0
-50
-100
-10 -5 0 5 10 15
A
d-
ex
ce
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
B
d-
ex
ce
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
C
d-
ex
ce
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
D
d-
ex
ce
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
E
d-
ex
ce
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
F
17
O
-e
xc
es
s 
(p
er
 m
eg
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
G
17
O
-e
xc
es
s 
(p
er
 m
eg
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
H
17
O
-e
xc
es
s 
(p
er
 m
eg
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
I
17
O
-e
xc
es
s 
(p
er
 m
eg
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
J
17
O
-e
xc
es
s 
(p
er
 m
eg
)
δ18OVSMOW (‰)
Wetter
Drier
40%50%
60%
70%
RHn = 80%
Sensitivity to RHnFreshwater
T = 25°C T = 15°C
Warmer Colder
w = 0.5
w = 0.6
w = 0.7
w = 0.8
E/I = 1
E/I = 0.1
E/I = 0.5
Terminal Lake
Throughflow Lake
90%
80%
70%
Freshwater
Turbulent
Less Turbulent
Freshwater
Freshwater
Freshwater
Freshwater
Freshwater
Freshwater
Freshwater
Freshwater
90% 80%
70%
w = 0.5
w = 0.6
w = 0.7
w = 0.8
T = 25°C
T = 15°C
Warmer
Colder
RHn = 80%
70%
60%
50%
40%
Wetter
Drier
Sensitivity to E/I
(Evaporation/Inflow)
Sensitivity to Temperature
Sensitivity to Turbulence
(wind speed)
Sensitivity to Atmospheric 
Vapor Composition
(equilibrium with 
freshwater input)
Sensitivity to RHn
Sensitivity to E/I
(Evaporation/Inflow)
Sensitivity to Temperature
Sensitivity to Turbulence
(wind speed)
Sensitivity to Atmospheric 
Vapor Composition
(equilibrium with 
freshwater input)
E/I = 1
E/I = 0.1
E/I = 0.5 Terminal Lake
Throughflow Lake
Fig. 1.1a. Sensitivity of d-excess vs. δ18O (A-E) and 17O-excess vs. δ18O (F-J) to different environmental
parameters during the evaporation of a water body in partial equilibrium with atmospheric vapour
(after Gázquez et al. 2018). The isotopic composition of the freshwater endmember is δ18O =
−8h; δD = −54h; d-excess = 10h; 17O-excess = 30 per meg. When all parameters remain
constant, the isotopic compositions of water pools with different ratios of Evaporation/Inflow
(E/I) are also represented in figures A and F. The isotopic composition of a terminal lake has
then been modelled under different conditions of normalised relative humidity (RH ) (B and G),
temperature (C and H), boundary layer turbulence (created by wind) (D and I), and the degree of
equilibrium between the atmospheric vapour and the freshwater input (E and J).
1.2 Statement of the problem 5
50
0
-50
-100
-10 -5 0 5 10 15
D
d
-e
xc
e
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
E
d
-e
xc
e
ss
 (
‰
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
I
1
7
O
-e
xc
e
ss
 (
p
e
r 
m
e
g
)
δ18OVSMOW (‰)
50
0
-50
-100
-10 -5 0 5 10 15
J
1
7
O
-e
xc
e
ss
 (
p
e
r 
m
e
g
)
δ18OVSMOW (‰)
w = 0.5
w = 0.6
w = 0.7
w = 0.8
90%
80%
70%
Turbulent
Less Turbulent
Freshwater
Freshwater
Freshwater
Freshwater
90% 80%
70%
w = 0.5
w = 0.6
w = 0.7
w = 0.8
Sensitivity to Turbulence
(wind speed)
Sensitivity to Atmospheric 
Vapor Composition
(equilibrium with 
freshwater input)
Sensitivity to Turbulence
(wind speed)
Sensitivity to Atmospheric 
Vapor Composition
(equilibrium with 
freshwater input)
Fig. 1.1b. Continuation of Fig. 1.1a. See left.
2018). The differing sensitivities of δ17O and δ18O to equilibrium and kinetic fractionation
processes means that the 17O-excess parameter provides a greater constraint on hydrological
conditions than does the analysis of δ18O and/or δD alone (Barkan and Luz, 2007) (Fig. 1.1
F-J). The main advantages of the 17O-excess parameter over d-excess are:
i. The ratio of the triple-isotope exponent, θ (ln 17α/ ln 18α), is almost identical within a
30°C temperature range, making 17O-excess largely temperature-independent (Barkan
and Luz, 2005; Cao and Liu, 2011; Surma et al., 2015); and
ii. 17O-excess is insensitive to elevated salt concentrations; while both cations and anions
affect the 2H/1H salt effect (Driesner et al., 2000), making d-excess a function of
the salt composition, the activity ratios of 18O/16O and 17O/16O are both controlled
by the cations and thus 17O-excess should remain unaffected (Sofer and Gat, 1972;
Herwartz et al., 2017).
The significant benefits of the triple oxygen isotope system have resulted in its recent
application to reconstruct evaporative processes, such as the effect of evapotranspiration
photosynthesis on atmospheric O2 (Landais et al., 2006; Li et al., 2017b), glacial-interglacial
hydrologic conditions at the ocean surface (e.g. Landais et al., 2008), and the parent waters
6 Introduction
of carbonates (e.g. Benedix et al., 2003; Clayton and Mayeda, 1984; Farquhar, 1998; Passey
et al., 2014), as well as being employed as a tracer of biosphere productivity (e.g. Luz et al.,
1999) and to examine water-rock interactions (Pack and Herwartz, 2014). To date, however,
analysis of the full suite of water isotopologues has been limited to polar ice cores (e.g. Risi
et al., 2010; Landais et al., 2012; Winkler et al., 2013) and modern environments such as
lakes (e.g. Surma et al., 2015, 2018) where the material analysed is the water substance
itself. Because the predicted trends of both d-excess and 17O-excess in evaporating waters
display different responses to climate variables, they can be evaluated individually using
iterative computer models to deconvolve individual hydrological parameters (e.g. Gázquez
et al., 2018). If analysed together, triple oxygen and hydrogen isotopes and their derived
parameters provide a powerful constraint on the hydrological conditions during proxy form-
ation.
1.2.2 Analytical limitations
Despite the benefits of measuring all water isotopologues, analytical challenges have pre-
vented the widespread application of 17O-excess to date. Firstly, 17O comprises less than
0.04% of all naturally occurring stable oxygen, resulting in small signal variability. Measur-
ing 17O therefore requires instrumentation capable of extremely high precision (better than
0.01h, or 10 per meg). Secondly, traditional isotope ratio mass spectrometry (IRMS) does
not permit the direct measurement of 17O-excess from water or via conversion to CO2 (often
used for measuring δ18O in water) due to isobaric interferences; the 17O-containing CO2 iso-
topologue, 12C17O16O (~760 ppm), is masked by the more abundant 13C16O2 isotopologue
(~11,000 ppm) of the same nominal mass (Passey et al., 2014). Fluorination chemistry (e.g.
using cobalt fluorine) has therefore been used to convert H2O to O2 for measurement via
IRMS:
2H2O+4CoF3 −−→ O2+4CoF2+4HF
with the resultant O2 measured for long time periods to produce the required precision.
The conversion of H2O to O2 also precludes the simultaneous measurement of oxygen and
hydrogen isotopic ratios, thereby requiring multiple instrumental calibrations and increasing
potential for isotopic fractionation during sample preparation. These analytical difficulties,
coupled with the numerous safety risks associated with the fluorination method, have resul-
ted in only a small number of laboratories measuring δ17O in water.
Recent developments in cavity ring-down laser spectroscopy (CRDS) have now made
it possible to measure water isotopes directly as H2O vapour at a precision comparable
to or better than IRMS. CRDS is a time-based measurement system that uses a laser to
1.3 Novel approach 7
quantify spectral absorption lines unique to H216O, H217O, H218O, and HD16O in an optical
cavity. Unlike IRMS, CRDS allows simultaneous measurement of oxygen and hydrogen
isotopic ratios without the need to convert the water into another gas, thereby minimising the
opportunity for isotopic fractionation (Gázquez et al., 2015). Measuring all isotopologues
on a single phase (H2O vapour) also simplifies calibration, resulting in greater precision and
accuracy than is achievable by IRMS, especially for hydrogen isotopes.
1.3 Novel approach
Many minerals found in nature and in the geological record incorporate molecular water
(H2O) or hydroxyl (OH) into their crystal structures during formation, theoretically provid-
ing a vector to simultaneously analyse triple oxygen and hydrogen isotopes through the
geological record. Examples of hydrated (or hydoxylated) minerals include iron oxides
(goethite), phyllosilicates, and evaporitic minerals. While phyllosilicates and goethite are
both predominantly the product of chemical weathering reactions, evaporites are water-
soluble mineral sediments (Goldschmidt, 1937) that form from the concentration and crys-
tallisation of ions from an aqueous solution (Hardie and Eugster, 1980). The isotopic
composition of the crystallisation water of these hydrated minerals has the potential to
provide information on the isotopic composition of waters from which they formed (Horita,
1989; Hodell et al., 2012).
The oxygen and hydrogen isotopes of the original fluid from which a hydrated mineral
precipitated can be determined if:
i. isotopic equilibrium is achieved between the mother and hydration water;
ii. the fractionation factors (α) during mineral formation are known (and temperature-
independent or the temperature dependence is known);
α =
(
18O/16O
)
crystal -water(
18O/16O
)
mother -water
or
(
17O/16O
)
crystal -water(
17O/16O
)
mother -water
or
(D/H)crystal -water
(D/H)mother -water
iii. and no isotope exchange has occurred between environmental and hydration water
after mineral deposition (Hodell et al., 2012).
If all conditions are met, the measurement of oxygen and hydrogen isotopes and their de-
rived parameters (d-excess and 17O-excess) could provide quantitative hydrological inform-
ation, including changes in atmospheric relative humidity, wind velocity and water source,
which have been difficult to achieve and calibrate in traditional archives (e.g. Gázquez
8 Introduction
et al., 2018). Additionally, hydrated minerals have a greater occurrence, both geographically
and throughout time, compared to all other archives that yield triple oxygen and hydrogen
isotope data (e.g. Warren, 2016).
Theoretically, gypsum (CaSO4 ·2H2O) is a favourable mineral for the study of hydration
water systematics both because of its widespread occurrence (e.g. Warren, 2016), and
because the structurally bonded water in gypsum is thought to form in equilibrium with the
ambient gypsum mother water during gypsum formation (e.g. Sofer, 1978; Hodell et al.,
2012; Herwartz et al., 2017). The corresponding isotope fractionation factors for D/1H
and 18O/16O have been traditionally viewed as temperature- (Gonfiantini and Fontes, 1963;
Hodell et al., 2012) and salinity- (Sofer and Gat, 1972; Horita, 1989) independent, but the
17O/16O fractionation factor has not been determined.
Despite the abundance of gypsum in the sedimentary record, relatively few studies have
utilised gypsum hydration water (GHW) for palaeoenvironmental reconstruction. Concerns
that the isotopic composition of GHW can be modified by post-depositional recrystallisation
(i.e. dissolution and reprecipitation) or isotopic exchange with environmental water have
limited the use of GHW as a palaeoenvironmental proxy to date (e.g. Pierre and Fontes,
1978; Sofer, 1978), although other studies have indicated gypsum displays relatively low
rates of exchange (e.g. Charola et al., 2007; Hodell et al., 2012). The post-depositional
isotopic exchange of GHW can be examined via exchange and dehydration experiments
(see Section 2.5) and by comparison to other climatic proxy datasets (e.g. Hodell et al.,
2012; Grauel et al., 2016; Li et al., 2017a). Additionally, positive correlation between proxy
and GHW indicates that, at least in some cases, the primary isotopic composition of GHW
is preserved in time (Hodell et al., 2012; Grauel et al., 2016; Gázquez et al., 2017a).
As well as the potential derivation of all isotopes of the water from which gypsum
precipitated, it is also possible to use GHW to derive palaeo-water temperatures in en-
vironments where gypsum and carbonate co-precipitate. The δ18O of calcite (δ18Ocal)
or aragonite (δ18Oarag) is dependent upon both the temperature (T) and the δ18O of the
water (δ18Owater) from which the carbonate was precipitated (Urey et al., 1951; Epstein
et al., 1953). Tandem measurements of both GHW and carbonate δ18O that were deposited
concurrently permit the deconvolution of the δ18O carbonate signal into its temperature and
δ18O-water components via carbonate palaeo-temperature equations (Hodell et al., 2012).
For example, calculation of the temperature at which the calcitic shells form can be conduc-
ted using the equation of Anderson and Arthur (1983):
T (◦C)= 16−4.14 · (δ18Ocal (PDB)−δ18Owater (SMOW ))+0.13 · (δ18Ocal (PDB)−δ18Owater (SMOW ))2
and calculation of the temperature at which the aragonitic shells form can be conducted
1.4 Objectives and thesis outline 9
using the equation of Grossman and Ku (1986):
T (°C) = 21.8−4.59 · (δ18Oarag(PDB)−δ18Owater (SMOW ))
This method provides a powerful constraint for oxygen isotope palaeothermometry (Hodell
et al., 2012; Grauel et al., 2016).
1.4 Objectives and thesis outline
The advent of novel analytical methods such as CRDS coupled with the ability to extract
and simultaneously analyse all isotopes of water from a geologically abundant mineral
such as gypsum provides an exciting new direction for palaeoclimatological research. The
opportunity to pioneer novel methods for palaeoclimatic reconstruction led Professor D.
A. Hodell to bring together a multidisciplinary team of scientists in geochemistry, mineral
physics, and ab initio molecular dynamics to study hydration water in minerals.
The primary objective of this thesis is to understand how the chemistry of the fluids
from which gypsum formed is reflected in the geochemistry and the mineralogy of deposits
in geologic space and time. This research specifically provides quantitative and semi-
quantitative information to establish criteria for the reconstruction of the palaeoenviron-
mental conditions during the deposition of gypsum in marine and non-marine environments.
To achieve this, this thesis examines the triple oxygen and hydrogen isotopic composition
of GHW, and the isotopic and elemental variations of both traditional and novel tracers
(including calcium and barium isotopes) to deduce the origin of ancient gypsum deposits.
The work presented in this thesis would not have been possible without the assistance of the
multiple members of the ‘Water Isotopes in Hydrated Minerals’ team. To provide clarity,
the work carried out by Nicholas P. Evans is highlighted at the beginning of each chapter or
section.
This thesis is structured as follows:
Chapter 2 outlines the mechanisms of gypsum formation and water incorporation upon
precipitation. The chapter includes a description of the methods that have been developed
over the course of this research to extract and analyse GHW, as well as the calculation of the
fractionation factors between gypsum and free water. The conditions under which gypsum’s
structurally-bound water can undergo exchange with secondary waters are subsequently
discussed, including primary research into the thermal dehydration mechanics of gypsum
at elevated temperatures. This work provides a basis on which to study GHW as a proxy
of the actual isotopic composition of palaeo-waters, and to provide quantitative information
10 Introduction
about the relative effects of climatic factors, such as humidity, precipitation, and temperature
change at the time of gypsum formation in evaporative environments.
Chapter 3 describes the application of the novel geochemical method to reconstruct the
triple oxygen and hydrogen isotopes of palaeo-lake water, specifically to gypsum that was
deposited during peak drought conditions that occurred between 600 and 1100 AD on the
Yucatán Peninsula, Mexico. This period included the Terminal Classic Period when the
Classic Maya Civilisation of Mesoamerica collapsed (Hodell et al., 1995; Kennett et al.,
2012). Our data provide a complete and direct archive of hydrological conditions that have
previously been limited to ice core records.
The measurement of all isotopes of water permits the application and use of numerical
models, previously limited to modern lakes, to constrain palaeo-lake hydrology, thus provid-
ing robust, quantitative estimates of past changes in rainfall and relative humidity. In con-
trast to a previous estimate that inferred a ‘modest drought’ based on a less robust method
using a speleothem proxy (Medina-Elizalde and Rohling, 2012), mean annual rainfall was
found to have decreased between 41% and 54%, with intervals of up to 70%, compared to
today. It has also been shown for the first time that relative humidity was between 2-7%
lower than present-day conditions during the Terminal Classic Drought.
Chapter 4 outlines the current understanding, discoveries and controversies of the period
known as the ‘Messinian Salinity Crisis’ (MSC; 5.97-5.33 Ma). Interpretation of the de-
positional environments of the Mediterranean’s evaporite facies is critical to producing a
reliable interpretation of the MSC. The contribution of continental and marine source waters
to Mediterranean brines during precipitation of the ‘Primary Lower Gypsum’ and ‘Upper
Gypsum’ units are poorly constrained, however, and the factors that trigger the appearance
of different evaporite facies are largely unknown. The isotope tracers used to interpret the
hydrological history during gypsum deposition are outlined, and the methodologies used in
the subsequent chapters are described.
Chapter 5 describes a single-cycle study of Primary Lower Gypsum unit in the Sorbas
Basin, South-eastern Spain. The objective was to reconstruct the changing environment
of deposition and its relation to astronomically-forced climate change. The δ18O and δD
of GHW and salinity of fluid inclusions were measured in the same samples to test if
they record the composition of the mother fluid from which gypsum was precipitated.
Water isotopes are highly correlated with fluid inclusion salinity suggesting the hydration
water has not exchanged after formation. The relatively low water isotope values and
fluid inclusion salinities indicate a significant influence of meteoric water, whereas δ 34S,
δ18OSO4 and
87Sr/86Sr support a dominantly marine origin for the gypsum deposits. The
discrepancy between water and elemental isotope signatures can be reconciled if meteoric
1.4 Objectives and thesis outline 11
water dissolved previously deposited marine sulfates supplying calcium and sulfate ions
to the basin which maintained gypsum saturation. This recycling process accounts for the
marine δ34S, δ18OSO4 and
87Sr/86Sr signatures, whereas the low δ18O and δD values of
gypsum hydration water and fluid inclusion salinities reflect the influence of freshwater.
Chapter 6 builds on the methods used in Chapter 5. Here, a high-resolution, multi-
isotope approach is utilised to reconstruct the paleo-hydrological dynamics of the Sorbas
Basin and the wider Mediterranean region during the first stage of the Messinian Salinity
Crisis. Measurements of GHW from 12 cycles of the Primary Lower Gypsum are combined
with salinity inferred from gypsum fluid inclusions, 87Sr/86Sr, δ44/40Ca, δ34S, and δ18OSO4
in the same samples, as well as isotopic analysis δ44/40Ca, 87Sr/86Sr, δ18Ocarb, δ13C of
the interbedded carbonate marls. Results indicate that obliquity-controlled sea level and
eccentricity-modulated precession, superimposed on longer-term tectonic restriction of the
Mediterranean-Atlantic exchange, together played a significant role in modulating gypsum-
marl deposition in the Sorbas Basin. Strontium isotopes in gypsum display a gradual, but
non-unidirectional, divergence from the seawater curve up-section, with slight reversals in
the general trend of decreasing 87Sr/86Sr at ~5.84 Ma and 5.72 Ma. Maximum divergence
of 87Sr/86Sr from the seawater curve, as well as anomalous authigenic carbonate miner-
alogy (dolomite) and high δ18Ocarb is coincident of glacial stage TG20 and an eccentricity
minimum at ~5.75 Ma. Calcium isotopes track the variability in 87Sr/86Sr, overprinted by
Rayleigh fractionation effects. The non-unidirectional changes in strontium and calcium
isotopes in combination with the patterns of the expression (i.e. thickness and facies) of the
cycles in the Sorbas section indicate that sea level variability influenced basin hydrology.
Chapter 7 investigates the origin of gypsum deposition of both onshore and deep basin
deposits in the Mediterranean basin during the last stage of the MSC (5.55-5.33 Ma). Stron-
tium, calcium and barium isotope data obtained from the onshore Upper Gypsum have
significantly lower values than data obtained from the Primary Lower Gypsum deposits,
whereas oxygen and hydrogen isotopes of GHW display higher values. These data demon-
strate that the onshore Upper Gypsum deposits formed from the intense evaporation of
a dominantly freshwater fluid, during a period where the connection to the Atlantic was
periodic and highly restricted. The deep-basin ‘Upper Evaporites’ from Deep Sea Drilling
Project (DSDP) / Ocean Drilling Program (ODP) sites 372 and 654 are characterised by
similar strontium, calcium and sulfate isotope values to those of the onshore Upper Evapor-
ites. The deep basin water isotope data deviates from the majority of the onshore deposits,
however, indicating that variable and heterogeneous hydrologic conditions characterised the
deep basins during evaporite formation.

Chapter 2
Calcium Sulfate Systematics
2.1 Introduction
The isotopic composition of the crystallisation water of gypsum can potentially be utilised
as a palaeoclimatic proxy to trace geological and hydrogeological processes (e.g. Fontes and
Gonfiantini, 1967; Matsubaya and Sakai, 1973; Sofer, 1978). To examine the potential for
gypsum and its structural hydration water to accurately record the oxygen and hydrogen
isotopic signature of the fluid from which it formed, it is important to understand the
mechanisms of gypsum formation and how water is incorporated into the mineral during
crystallisation. Knowledge of the associated gypsum-water fractionation factors (α) is
particularly crucial, as small deviations in assumed α 18Ogypsum -water, α 17Ogypsum -water,
and αDgypsum -water result in variability that impacts the derived d-excess and 17O-excess
parameters, thus affecting the quantitative calculation of environmental conditions during
formation. To interpret the water isotope signatures, the complete extraction and accurate
measurement of the gypsum hydration water (GHW) and the isotopic ratios is fundamental
in providing reliable information about the conditions under which the gypsum formed and
any subsequent interaction with other fluids after deposition.
In this chapter, the mechanisms of calcium sulfate formation are outlined to provide a
framework to understand the conditions under which gypsum forms (Section 2.2). Sub-
sequently, the development of a semi-automated procedure to extract GHW and measure
the evolved water by cavity ring-down laser spectroscopy (CRDS) is discussed (Section
2.3). The methodology presents several analytical advantages over earlier methods, in-
cluding better long-term precision, higher sample throughput, reduced sample size and less
memory effect between consecutive samples. Using this methodology, precise and accurate
isotope fractionation factors for gypsum and water (α 18Ogypsum -water, α 17Ogypsum -water and
αDgypsum -water) are calculated (Section 2.4). Finally, the dehydration of gypsum and sub-
14 Calcium Sulfate Systematics
sequent ion exchange mechanisms are examined to ascertain the ability of gypsum to record
the isotopic signature of the fluid from which it formed and the range of environmental
conditions over which it is stable (Section 2.5).
2.2 Calcium sulfate formation
2.2.1 Calcium sulfate solubility and thermodynamic stability
Calcium sulfate can exist in various structural forms and the CaSO4-H2O system has been
classically characterised into the three main groups:
i. calcium sulfate dihydrate, gypsum (CaSO4 ·2H2O) with 20.9 wt% water;
ii. calcium sulfate hemihydrate, bassanite (CaSO4 ·0.5H2O) with ~6.2 wt% water, and;
iii. anhydrite (CaSO4) (Fig. 2.1).
Other intermediate and complimentary crystalline phases of calcium sulfate have also been
reported. These include differing forms of hemihydrate (α and β CaSO4 ·0.5H2O and
CaSO4 ·nH2O where 0 < n < 1.0, n ̸= 0.5), distinguished on the basis of their specific surface
area, crystal sizes, habit and surface topography (Freyer and Voigt, 2003; Ballirano and
Melis, 2009; Hildyard et al., 2011), whereas anhydrite can be subcategorised into soluble
(AIII−CaSO4 or γ-CaSO4) and insoluble phases (AI−CaSO4 or α-CaSO4 and AII−CaSO4
or β -CaSO4) (Christensen et al., 2008; Lou et al., 2011; Prieto-Taboada et al., 2014). Sol-
uble anhydrite (γ-CaSO4) derives its name from its metastable nature, and hydrates to
bassanite under normal atmospheric conditions (e.g. Harrison, 2012). When exposed to
temperatures above ~300°C, γ-anhydrite transforms into the insoluble β -anhydrite and no
longer rehydrates (Prieto-Taboada et al., 2014).
The direct precipitation of the different calcium sulfate phases is thought to be dependent
on the temperature, salinity and the supersaturation of calcium and sulfate in the mother
solution (e.g. Ostroff, 1964; Hardie, 1967; D’Ans, 1968; Billo, 1987; Freyer and Voigt,
2003; Ossorio et al., 2014) (Fig. 2.2). At a given temperature, salinity and supersaturation,
the solid phase with the lowest solubility should represent the stable phase (Freyer and
Voigt, 2003). Early solubility measurements and thermodynamic calculations indicated that
gypsum should precipitate from pure calcium sulfate aqueous solutions below temperatures
of ~50°C (D’Ans, 1968; Hardie, 1967), whereas anhydrite should directly precipitate above
this temperature. Determination of the stability region of each of the calcium sulfate phases
has proved elusive, however, due to inconsistencies between solubility measurements and
2.2 Calcium sulfate formation 15
Fig. 2.1. Structural projections (along the c-axis) of the three main calcium sulfate phases: anhydrite (AII
form) (A), bassanite (B), and gypsum (C). Tetrahedra are sulfate ions, large spheres are Ca2+ ions,
and small spheres are water molecules. Structural projection (along the a-axis) of anhydrite (AII
form) (D), bassanite (E), and gypsum (F) (from Stawski et al., 2016).
thermodynamic predictions of stability, and data obtained in laboratory precipitation studies
(e.g. Billo, 1987; Freyer and Voigt, 2003; Ossorio et al., 2014). Experimental data indicates
that γ-anhydrite does not directly precipitate within its predicted stability field, and that
gypsum forms at temperatures above its theoretical equilibrium (Ossorio et al., 2014). The
intrinsic slow growth dynamics of γ-anhydrite mean that even when stable microscopic
anhydrite nuclei are formed in an aqueous solution, a longer time period is required for
volumetrically significant amounts of anhydrite to form (Ossorio et al., 2014).
Direct gypsum formation from aqueous solutions was originally assumed to occur via a
single-step process following the classical nucleation and growth mechanism (e.g. Packter,
1974). Recent work using synchrotron-based in situ X-ray techniques has shown that
gypsum precipitation occurs via a predictable four-stage solution-based nucleation and growth
16 Calcium Sulfate Systematics
pathway (Stawski et al., 2016). During this process, amorphous calcium sulfate and/or
nano-crystalline bassanite forms as a precursor to gypsum (Saha et al., 2012; Van Driessche
et al., 2012; Wang et al., 2012). If the water volume available during formation is limiting,
the reaction will be directly halted at the intermediate point and bassanite will emerge as
the bulk crystalline phase (Stawski et al., 2016). Low water activity also increases the
residence time of the bassanite precursor: in highly saline solutions (e.g. water activity
~0.8), bassanite is the primary precipitate for up to 2 months at 80°C before turning into
γ-anhydrite (Ossorio et al., 2014; Rapin et al., 2016). At higher water activities and lower
temperatures, bassanite is unstable and will fully hydrate to gypsum (Badens et al., 1998)
(Fig. 2.3). Due to the instability bassanite and and slow growth kinetics of γ-anhydrite
formation, gypsum is the major primary precipitate in natural environments on Earth.
In summary, the major and volumetrically most important gypsum deposits worldwide
are formed from direct deposition from a solution saturated with respect to gypsum at
temperatures below ~60°C, while smaller volumes of gypsum may form via the hydration
of γ-anhydrite.
2.2.2 Natural ion sources for calcium sulfate formation
The formation of calcium sulfate in natural environments is dependent on both the pro-
portion of specific dissolved ions in a water body, and the mechanism that results in the
separation of dissolved calcium sulfate from the water solute (i.e. promoting precipita-
tion). The evaporation of waters in (semi-) enclosed environments is the most common
mechanism that results in a rise in concentration of dissolved ions in a water solution.
‘Freshwater’, containing >1 g l-1 total dissolved solids, can be evaporatively concentrated
so that the waters become brackish, saline, brine, and finally a bittern (Ba¸bel and Schreiber,
2014). While these terms are not defined unequivocally, brackish commonly refers to the
transitional zone between marine and freshwater, saline waters have salinities similar to or
greater than seawater (35 g l-1 total dissolved solids), brines are waters that contain sodium
chloride and are significantly more saline than seawater, and bitterns are usually defined as
brines that are stripped of their sodium chloride content (Ba¸bel and Schreiber, 2014).
The ion source for all evaporite deposits originate from soluble chemical compounds
(salts) dissolved in natural waters on Earth. There are many natural types of salt-containing
waters, including lakes, rivers, and groundwater; the ocean, however, is the volumetrically
largest source (~47.6×1018 kg of salts; Hay et al., 2006). Of the many elements dissolved
in seawater, Na+ (30.5), Cl− (55.1), SO42− (7.7), Mg2+ (3.7), Ca2+ (1.2), K+ (1.1), and
HCO3− (0.4) (proportion of total salts by weight, in %, is given in brackets; Hay et al., 2006)
create the stable and volumetrically significant salts as the product of seawater evaporation
2.2 Calcium sulfate formation 17
100
80
60
40
20
0
C
aS
O
4 
(m
M
)
0 20 40 60 80 100 120
0 20 40 60 80 100 120
0 20 40 60 80 100 120
0 20 40 60 80 100 120
20C
aS
O
4 
(m
M
)
0
40
60
80
100
20C
aS
O
4 
(m
M
)
0
40
60
80
100
20C
aS
O
4 
(m
M
)
0
40
60
80
100
Temperature (°C)
Freyer and Voigt (2003)
PHREEQC Database
LLNL Database
1M NaCl
3M NaCl
4.5M NaCl
BassaniteAnhydrite
Gypsum
Fig. 2.2. Solubility curves of gypsum, bassanite and anhydrite as a function of temperature (0°C to 120°C)
and salinity (0 M, 1 M, 3 M and 4.5 M NaCl) calculated from experimental and modelled data (after
Ossorio et al., 2014). Solubility data (crosses) were obtained from Freyer and Voigt (2003) and the
references therein. Calculated solubility curves of the calcium sulfate phases were obtained from
the PHREEQC database (solid lines) and the Lawrence Livermore National Laboratory (LLNL)
database within operated within PHREEQC (dashed lines) (Parkhurst and Appelo, 2013). Black
vertical lines mark the gypsum-anhydrite transition temperature obtained with the PHREEQC
(solid lines) and LLNL (dashed lines) databases.
18 Calcium Sulfate Systematics
Anhydrite
Bassanite
Gyp
sum
Gypsum
150
100
50
0
-50
-100
0.4 0.5 0.6 0.7 0.8 0.9 1.0
Te
m
pe
ra
tu
re
 (
°C
)
Activity of Liquid Water
3 days
3 da
ys
Fig. 2.3. Calcium sulfate stability as a function of temperature and water activity in aqueous solution (after
Rapin et al., 2016). The solid black line represents the gypsum-anhydrite transition (Freyer and
Voigt, 2003). Kinetic boundaries (i.e. boundaries delimiting the phases found in solution after
a specific timing) from Ossorio et al. (2014) are reported (red dashed lines) and represent the
precursor phases found in solution after 3 days, with subsequent evolution path leading to the
formation of anhydrite (red arrow). The red box denotes the range covered by the experiments of
Ossorio et al. (2014). Extrapolated trends to lower water activities are shown in dotted red lines,
and are only indicative.
2.2 Calcium sulfate formation 19
(see Ba¸bel and Schreiber, 2014).
The precipitation of evaporitic minerals induced by the evaporative concentration of
seawater results in the modification of the ionic composition of the mother fluid due to the
combined constraints of equilibrium and mass balance (Eugster, 1980; Ba¸bel and Schreiber,
2014). During equilibrium crystallisation, both the ion activity product of the solution
must remain constant (at constant pressure and temperature), and the appropriate molar
proportions of ions must be removed from the solution (Hardie and Eugster 1971; Eugster
and Hardie, 1978). Thus, during the evaporation of modern seawater, the concentrations of
major ions systematically change in a predictable way and result in three stages of mineral
precipitation (Usiglio, 1849);
calciumcarbonate−−→ calciumsul f ate−−→ sodiumchloride
This sequence is also well known from geochemical studies in both solar saltworks (Fig.
2.4) and laboratory seawater evaporation experiments (e.g. Schreiber and Kingsman, 1975;
Warren, 1982; Ortí et al., 1984; McCaffrey et al., 1987).
Calcium carbonate: Evaporitic alkaline earth carbonates are the first to precipitate
at elevated salinities (>35h) induced by evaporation of seawater due to their relatively
low solubility in water. Calcium carbonate precipitation eliminates HCO3− ions from the
solution according to the reaction:
Ca2++2HCO3− −−→ CaCO3+H2O+CO2
The greater the disparity between the initial concentration of Ca2+ and HCO3− ion in
solution, the more rapid the enrichment-depletion process will be (Eugster, 1980; Holland
et al., 1996; Ba¸bel and Schreiber, 2014). Thus, fluids that are initially rich in carbonate will
experience a relative enrichment of carbonate and depletion of calcium and vice versa. As
modern seawater contains a relatively high Ca2+/HCO3− ratio, the evaporated seawater
brine will experience a relative depletion of carbonate and enrichment of calcium after
calcite precipitation (Eugster, 1980; Ba¸bel and Schreiber, 2014).
During the precipitation of calcium carbonate the Mg/Ca ratio in the solution also
increases, first producing low-Mg calcite or aragonite (CaCO3) (Sandberg, 1983), and even-
tually high-Mg calcite (Ca,Mg)CO3 and protodolomite (CaMg(CO3)2) (Eugster, 1980).
In natural environments, several other nonevaporite-driving mechanisms commonly induce
CaCO3 deposition within a basin (e.g. Decima et al. 1988). Evaporation-driven precipita-
tion of carbonate occurs up to the salinity characterising the precipitation of gypsum (i.e.
~140h).
20 Calcium Sulfate Systematics
Fig. 2.4. Schematic of a marine saltwork, displaying the evaporation pathway of seawater with increasing
salinity and water density, modified after Ortí (2011) and Ba¸bel and Schreiber (2014). Remaining
sulfate-rich brine flows back to the ocean or, in some saltworks, back to the concentration pans to
promote more precipitation of gypsum, which crystallises due to a mixture of brines (Ba¸bel and
Schreiber, 2014).
Calcium Sulfate: Modern seawater contains relatively high Ca2+/HCO3− and SO42−/HCO3−
ratios. As such, Ca2+ is not fully consumed during the prior precipitation of CaCO3, result-
ing in the supersaturation of the brine for gypsum during evaporation. Gypsum saturation
occurs at salinities of ~140h when ~80% of the original seawater has evaporated (Ortí,
2011; Rossi et al., 2015). The resultant deposits of gypsum precipitation can be divided into
two major facies groups:
i. Fine-grained gypsum facies are characterised by fine-grained laminae in thin beds, of-
ten found in deep basins and in association with fine- to coarse-grained slope clastics
(e.g. gypsum turbidites and gypsum breccias). On platforms, gypsum cumulates
often display more distinct lamination, and the clastic facies are mainly composed
of disorganised gypsum blocks (Ortí, 2011). This facies also includes microcrystal-
line laminated gypsum microbialites (stromatolites), an organo-sedimentary deposit
2.2 Calcium sulfate formation 21
formed by the interaction of a microbial community (Ba¸bel, 2004) that often show
vertically stratified coloured zones (white, green, brown and black) of a few mm to
~1 cm thick.
ii. Selenite facies are autochthonous, coarse-grained (>2 mm), massive deposits that dis-
play euhedral crystal structures. Selenite can occur as tabular, reticular, and columnar
crystals or develop twins or intergrowths along the (001) plane. Crystals commonly
display upward competitive growth and exhibit zonation due to the presence of matrix
inclusions (Ortí, 2011). The term selenite also includes other ‘crystalline varieties’ of
gypsum such as satin spar, desert rose, and gypsum flower. Individual selenite crystals
can reach up to several tens of centimetres or even metres in length (Warren, 2016).
Because Ca2+ < SO42− in seawater, gypsum precipitation ceases when Ca2+ ions are ex-
hausted.
Sodium chloride: Halite (NaCl) begins to crystallise when seawater is evaporated to
~10% of the original volume. Minor amounts of gypsum form during the early stages of
halite crystallisation due to the overlap of the crystallisation fields (Ba¸bel and Schreiber,
2014). Halite is a cubic structure and displays three perfect cleavage directions at right
angles. Millimetre-scale zoning, often arranged into chevron-like pattern, is often seen in
halite crystals growing in bottom crusts. The rapid rate of growth results in trapped fluid
inclusions within the crystalline structure (Lowenstein et al., 2001; Rigaudier et al., 2011).
Unlike seawater brines, the evaporative concentration of typical continental waters does
not draw on the near iso-geochemical reservoir of the open ocean (Warren, 2016). Contin-
ental waters have a more variable suite of dissolved ions derived from the weathering of local
material, resulting in a larger and less predictable range of minerals that precipitate during
evaporation. For example, flow through of limestone aquifers produces waters rich in Ca2+
to HCO3−, dolomite dissolution generates Mg2+, and erosion of igneous and metamorphic
matrixes often yields fluids containing Ca2+, Na+ and HCO3− (e.g. Hardie, 1991; Warren,
2016).
In typical freshwater environments, the anionic content is often dominated by bicar-
bonate (Eugster and Hardie, 1978). Therefore as the Ca2+ present in the fluid is removed
during the evaporative CaCO3 phase of precipitation, further evaporation cannot result in
the precipitation of Ca-bearing evaporites such as gypsum (Hardie and Eugster, 1971).
To increase the Ca2+ to HCO3− ratio of freshwaters to facilitate gypsum precipitation,
fluids need to be enriched in calcium and/or sulfate, either via mixing with seawater, or
from surrounding lithologies that are leached in the drainage basin (e.g. marine-derived
evaporites) around the evaporative system (Rossi et al., 2015). Equally, in marginal marine
22 Calcium Sulfate Systematics
environments, the input of differing continental waters may result in the alteration of the
ion ratios and salinity thresholds to reach mineral saturation. Whereas high salinities are
required to saturate evaporated seawater (a Na−Cl−SO4−Mg brine) with gypsum (~140h;
Warren, 2016), gypsum saturation can be reached at low salinities (~5h) when additional
calcium and sulfate are added from non-marine sources (Rossi et al., 2015).
In summary, the geochemistry of calcium sulfate is linked to the process of evaporation
of a water body leading to salinity increase and concentration of calcium and sulfate ions.
As water incorporation into calcium sulfate occurs via a predictable pathway (e.g. Stawski
et al., 2016), the triple oxygen and hydrogen isotope composition of the GHW should
provide reliable and reproducible information regarding the isotope composition of the
fluid at the time of mineral precipitation. While the volumetrically most abundant gypsum
precipitates forms from these major reservoirs, a small percentage of gypsum also forms
from hydrothermal fluids (e.g. Gázquez et al., 2013), volcanic vapours (e.g. Zimbelman
et al., 2005), and during brine freezing (e.g. Marion, 1997; Stark et al., 2003). It is therefore
important to extract and analyse GHW precisely, and investigate gypsum-water fraction-
ation factors as a function of the large range of environmental conditions (i.e. variable
temperatures and salinities) under which gypsum may form.
2.3 Simultaneous analysis of 17O/16O, 18O/16O and 2H/1H
of gypsum hydration water by cavity ring-down laser
spectroscopy
Adapted from the publication: Fernando Gázquez, Ian Mather, James Rolfe, Nicholas
P. Evans, Daniel Herwartz, Michael Staubwasser and David A. Hodell (2015). Rapid
Commun. Mass Spectrom, 29, 1997–2006.
The modification of the extraction procedure introduced by Hodell et al. (2012) is de-
scribed. F.G., I.M., J.R., and D.A.H. designed and built the vacuum assembly. F.G. and
N.P.E. conducted the extraction of GHW from a synthetic gypsum standard and a variety
of natural gypsum minerals, as well as the subsequent triple oxygen and hydrogen isotope
measurements using a new generation CRDS analyser (Picarro L2140-i). To validate the
17O-excess values obtained by CRDS, D.H. and M.S. measured the δ17O and δ18O of
GHW by fluorination followed by isotope ratio mass spectrometry (IRMS) of O2. The
procedure presents several analytical advantages over earlier methods, including better long-
term precision, higher sample throughput and reduced sample size.
2.3 Hydration water extraction and analysis procedure 23
Fig. 2.5. Schematic of the WASP (Water Analyser Sample Preparation) device developed for the extraction
of gypsum hydration water. P: Two-stage rotation pump; G: Pirani gauges; V: Swagelok® valves;
S: 12 mm Pyrex tubes for samples; LN2: Cryogenic traps and 6 mm OD tubes for freezing GHW.
2.3.1 The vacuum assembly
A semi-automated procedure is developed for extracting hydration water from gypsum
by slowly heating the sample to 400°C in vacuo and cryogenically trapping the evolved
water. The Water Analyser Sample Preparation (WASP) system comprises an adapted,
programmable gas chromatography (GC) oven and six individual extraction lines attached
to a common backing line, coupled to a two-stage rotary pump (Edwards® E2M2, Crawley,
UK) (Fig. 2.5). A liquid nitrogen (LN2) trap is fitted before the vacuum pump to improve
pumping efficiency and avoid backstreaming of pump oil.
Each line is ~1 m in length and composed of ¼-inch outer diameter (OD) stainless
steel tubing with an internal diameter of ~4 mm (Fig. 2.5). The six lines can be isolated
individually by means of Swagelok valves (Swagelok SS-DSS4, London, UK). Samples
are loaded into disposable 12 mm Pyrex tubes located in the GC oven (ThermoFinnigan
TRACE GC; Thermo Scientific, Bremen, Germany). The programmable oven provides
complete control over heating steps and rapid cooling following the analysis; this permits
several consecutive runs of the apparatus per day. The hydration water is recovered outside
the oven in 6 mm break-seal tubes by cryogenic trapping in LN2 (Fig. 2.5). The Pyrex and
break-seal tubes are connected to the stainless steel tubing via ¼-inch Swagelock Ultra-torr
unions fitted with VitonTM O-rings, which are replaced after every five or six runs of the
apparatus. All the stainless steel lines outside the GC oven are wrapped in heating tape and
held at 70°C to prevent condensation of water in the extraction lines.
24 Calcium Sulfate Systematics
Fig. 2.6. Typical pressure profile recorded during the extraction of GHW. The temperature ramp used along
the dehydration process is given. The final pressure differs depending on whether the line is leaking
or functioning properly. Vacuum leaks can affect the isotopic composition of the extracted GHW.
∆ = raise LN2 trap; # = heat cold spots; + = pump non-condensable gases.
A Pirani gauge (Edwards® APG100 Active Pirani vacuum gauge) is fitted to each of
the six lines for monitoring pressure via an Edwards TIC 6 head instrument controller. The
pressure in each line is continuously monitored during the extraction using a computer inter-
face to the WASP system. The change in pressure as a function of time allows identification
of any abnormalities (e.g. leaks) during the GHW extraction procedure (Fig. 2.6).
2.3.2 Extraction of GHW
Prior to GHW extraction via the WASP system, gypsum samples (~200 mg) are ground to
a homogenous powder in an agate mortar. Powders are loaded into 12 mm OD Pyrex tubes,
topped with 3 cm of quartz wool (to prevent loss of gypsum powder during dehydration)
and dried at 45°C for 24 h. To remove adsorbed water, samples are then either inserted into
the vacuum lines in the WASP and placed under vacuum to a pressure of ~5× 10−3 mbar
for at least 3 h at room temperature, or pumped offline for 3 h in a desiccator connected
to a two-stage rotary pump (Edwards) and a LN2 trap fitted in-line between the pump and
the desiccator. This low vacuum pumping (10-2 mbar to 10-3 mbar) is effective at removing
adsorbed water with no detectable loss of hydration water (Playà et al., 2005). No difference
has been found in the isotopic results between the two pumping methods. This allows up
2.3 Hydration water extraction and analysis procedure 25
to 18 gypsum samples (15 unknowns and 3 standards) to be extracted each day, thereby
considerably increasing the sample throughput compared with earlier methods (e.g. Hodell
et al., 2012). An analytical-grade powdered gypsum standard (C/2360/50; Fisher Scientific,
Hemel Hempstead, UK), denoted as the NEWGYP standard hereafter, is utilised in all
experimental runs to evaluate the external precision of the method.
After the lines have been completely evacuated (pressure ~5×10−3 mbar), each branch
is isolated from the vacuum and the glass 6 mm break-seal tubes at the end of each of the
lines are immersed in LN2 for water trapping. GHW is exhumed by slowly heating the
sample to 400°C (Fig. 2.6). The temperature ramp is 5°C min-1, between 25°C and 70°C,
and 2°C min-1 between 70°C and 120°C. In the temperature range of 70°C and 120°C,
gypsum dehydration occurs under low-to-medium vacuum conditions (McConnell et al.,
1987). The slower heating during the second step is required to avoid rapid dehydration
of the sample that can lead to gypsum powder blowing through the line. At 110°C, the
cryogenic traps are raised (~3 cm) to increase the surface area for freezing the water. The
ramp is increased to 5°C min-1 between 120°C and 200°C. At 200°C, the cryogenic traps
are again raised (~5 cm). After this step, the temperature ramp is increased to 10°C min-1
between 200°C and 400°C. Once the oven reaches 400°C, this temperature is held constant
for 40 min to ensure complete recovery of all water in the lines by the cryogenic traps.
Finally, non-condensable gases are pumped away for 30 s before the 6 mm break-seal tubes
are flame sealed. The GHW extraction process takes approximately 2 h per run. The
pressure in the lines typically reaches 1.2 mbar at 130°C (after 30 min) and then gradually
decreases to ~10-2 mbar at the end of the heating stage (Fig. 2.6).
After water extraction, the 6 mm OD tubes containing 40-60 μl of hydration water are
stored until ready for analysis. Prior to analysis, water is frozen into the base of the break-
seal tube by immersion in LN2. Subsequently, the tube is scored with a diamond cutter,
broken to a fixed height of 25 mm, and then quickly inserted into the 2 ml septum-capped
vials used by an A0325 autosampler (Picarro).
2.3.3 Isotopic measurement of GHW by CRDS
Triple oxygen and hydrogen isotopes of water are measured simultaneously by CRDS using
a L2140-i Picarro water isotope analyser (Steig et al., 2014), interfaced with an A0211
high-precision vaporiser (both from Picarro, Santa Clara, CA, USA). A Picarro micro-
combustion module (MCM®), comprising of an 8 cm-long cylindrical cartridge filled with
metallic wool coated with a proprietary catalyst, is placed in-line between the vaporiser and
the water isotope analyser. The MCM has been shown to remove combustible compounds
from water samples (Brand et al., 2009; Saad et al., 2013). For experiments that utilised the
26 Calcium Sulfate Systematics
MCM (denoted as ‘MCM On’ hereafter), the cartridge was held at a constant temperature of
200°C. For experiments that did not utilise the accessory (denoted as ’MCM Off’ hereafter),
the WARM mode (~80°C) configuration was used. This configuration avoids water vapour
condensation in the MCM transfer line, thereby reducing the memory effect between injec-
tions; however, this MCM mode does not remove organics from the vapour stream. Both
the MCM On and MCM Off analyses use dry air (containing 21% O2) as the carrier gas.
Each sample is injected ten times into the vaporiser, which is heated to 110°C.
2.3.4 Internal and external precision of the CRDS/WASP system
Internal precision of the CRDS is calculated from the repeat analysis of an in-house water
standard (SPIT). Memory effects from previous water samples are avoided by rejecting
the first three analyses of the ten injections into the vaporiser. The values for the final
seven injections are averaged with a typical in-sample precision (±1σ ) of ±0.02h for δ17O
values, ±0.04h for δ18O values, and ±0.19h for δD values (n = 23). The 17O-excess and
d-excess were calculated for each injection using the corrected δ17O, δ18O and δD values.
The 17O-excess and d-excess from the seven repeated analyses were then averaged. Typical
in-sample 17O-excess and d-excess precisions (±1σ ) in water standards (SPIT, n = 23) were
13 per meg and 0.2h, respectively. This agrees with the guaranteed precision given by the
manufacturer (15 per meg for 17O-excess). No 17O-excess or d-excess values were rejected.
Results of CRDS analysis are normalised against the VSMOW standard by analysing
internal standards before and after each set of 12 to 15 samples. To this end, four internal
water standards (JRW, BOTTY, SPIT and ENR-15) were calibrated against VSMOW, using
δ17O and δ18O values of 0.0h, respectively, SLAP using δ17O of -29.6986h and δ18O
values of -55.5h (Lin et al., 2010; Schoenemann et al., 2013), and the GISP standard water
was analysed as an unknown. GISP yielded an average 17O-excess value of 25 ± 12 per meg.
This value is in good agreement with the results reported by Schoenemann et al. (2013) (22
± 11 per meg). The in-run drift of the instrument was corrected (when necessary) by the
analysis of an internal standard every 3 or 4 samples.
The long-term precision of the method, including GHW extraction using the WASP
system and sample preparation, was ±0.07h for δ17O values, ±0.13h for δ18O values, and
±0.49h for δD values (n = 37, 7 injections each, ±1σ ) for analyses of NEWGYP (δ17O =
0.18h, δ18O = 0.31h and δD = -51.02h). No significant differences (within analytical
errors) in the value of the in-house water standard, NEWGYP, were found and no systematic
drift has been observed in the δ17O, δ18O and δD measurements of the standard over time
(Fig. 2.7). In principle, any long-term drift can be monitored and corrected as the gypsum
standard is extracted with each five unknown samples. No such drift was observed. If a drift
2.3 Hydration water extraction and analysis procedure 27
is observed, a correction could be applied to improve the long-term reproducibility of the
method.
The WASP system was tested for memory effects by extracting the hydration water
of isotopically depleted and enriched gypsum in two consecutive runs of the apparatus.
These gypsum standards (DPL-gyp and ENR-gyp) were produced by hydrating anhydrous
CaSO4 with isotopically depleted (-6.1h, -8.6h, and -131.4h for δ17O, δ18O and δD
values, respectively) and enriched (11.2h, 21.7h and 57.8h for δ17O, δ18O and δD,
respectively) waters, following the procedure of Conley and Bundy (1958). Prior to the
first extraction, the lines were evacuated to a pressure of ~5×10−3 mbar for 4 days before
the first experiments to remove potential adsorbed water remaining from previous samples.
The system was evacuated for 30 min between the two consecutive runs. The samples
extracted in the first run were found not to influence the δ17O, δ18O and δD values of
the samples extracted in the second run of the apparatus. The values for the isotopically
enriched gypsum (δ17O = 13.05h, δ18O = 25.10h and δD = 34.83h, n = 3, 7 injections
each) showed analytical errors (±1σ ) of ±0.08h, ±0.16h and ±0.24h for the δ17O, δ18O
and δD values, respectively, and the errors for the isotopically depleted gypsum (δ17O =
-6.81h, δ18O = -12.86h, and δD = -150.39h, n = 4, 7 injections each) were ±0.06h,
±0.12h and ±0.53h, respectively (Appendix A; Table A.2). These results show that there
is no measurable memory effect in the WASP system with the pumping protocol.
The presence of organic matter and other contaminants in gypsum, including gases
derived from microbial activity such as H2S and CH4 trapped in the mineral lattice, may
produce spectral interferences in the CRDS measurements when GHW is analysed, as has
also been reported for the analysis of water extracted from the xylem of plants (Brand
et al., 2009; Saad et al., 2013). Such contamination is likely to be greater in gypsum
deposited in environments associated with organic matter degradation, such as lakes and
marine settings. To evaluate gypsum formed in a wide variety of depositional environments,
the natural sample set included gypsum formed by evaporation in a lake (PI 6A-13H-2 8 cm
and PI 6C-7H-2 26 cm; Hodell et al. 2012), gypsum derived from seawater evaporation in a
coastal setting (Salina 1; Evans et al., 2015), gypsum speleothems generated by drip water
evaporation in a gypsum cave (SBL-2.3 and SBL-8; Gázquez and Calaforra, 2014), and
two samples of natural gypsum formed in hydrothermal environments (CRI-01 and BG-10;
García-Guinea et al., 2002; Gázquez et al., 2013).
For natural gypsum samples, CRDS using the MCM produced values ranging from -
1.00h to 5.61h for δ17O, from -1.93h to 10.74h for δ18O and from -53.36h to 13.21h
for δD (Appendix A; Table A.3). The in-sample reproducibility (±1σ ) of the CRDS ana-
lyser for the analysis of natural gypsum samples was typically ±0.03h for δ17O values,
28 Calcium Sulfate Systematics
Fig. 2.7. The δ18O (A), δD (B), 17O-excess (C) and d-excess (D) values of GHW from the repeated analysis
of a gypsum standard (NEWGYP), extracted using the WASP and analysed by CRDS (n = 37) and
IRMS (n = 8). δ17O displayed a similar trend to δ18O. Data are displayed following the order
in which the samples were extracted using the WASP. Error bars refer to the internal error (±1σ
obtained from the repeated analysis of the same hydration water (7 injections for CRDS and 3-4
for IRMS). The long-term means (dashed line) and external errors are shown for ±1σ (deep-grey
shade) and ±2σ (light-grey shade).
2.3 Hydration water extraction and analysis procedure 29
±0.04h for δ18O values, and ±0.24h for δD values, found by taking the average of seven
consecutive injections for each GHW sample. These results are similar to those observed for
the analysis of water standards. Regarding the long-term reproducibility of natural samples,
the replicate GHW extraction and subsequent analysis in different runs of sample Salina
1 (n = 3, 7 injections each) gave an error (±1σ ) of ±0.02h for δ17O values, ±0.03h for
δ18O values, and ±0.16h for δD values (Appendix A; Table A.3). This demonstrates the
long-term precision obtainable for the hydration water of natural gypsum samples using the
method.
To validate the values obtained by CRDS, D. Herwartz and M. Staubwasser measured
the GHW δ17O and δ18O of the NEWGYP standard and the natural gypsum samples by
fluorination followed by IRMS of O2 at the University of Cologne (Cologne, Germany).
~800 mg of sample was loaded into the WASP, and the subsequent water samples (2.8 μl)
were injected via an autosampler into a CoF3 reactor held at 375°C which converts H2O into
O2 and HF, following the method developed by Barkan and Luz (2005). Helium carries the
products through a double cold trap immersed in LN2 to freeze out the HF. The O2 is then
trapped using 5 Å molecular sieve held at LN2 temperature. Helium is pumped away after
disconnecting the oxygen trap from the He stream and sample O2 is transferred to one tube in
a 12-fold sample manifold immersed in LN2. For this study a manifold typically contained
four reference water extractions, leaving eight sample tubes for O2-extractions from two or
three unknown water samples. To minimize memory effects at least three sample injections
were discarded when switching between waters of different isotopic values. The CoF3 line
is LabView (National Instruments, Austin, TX, USA) controlled and runs fully automated
overnight. The sample O2 was then analysed on a MAT253 IRMS (Thermo Scientific)
in dual-inlet mode at m/z 32, 33 and 34. Each measurement comprises 66 samples to
reference comparisons divided into 3 blocks of 22 cycles each. Each block begins with
a peak centre and adjustment of bellow pressure, such that the ion beam intensity at m/z
32 is 8.5 V. The idle time was set to 17 s and the integration time to 12 s. The baseline was
measured for 240 s prior to the first sample. The δ17O sample and δ18O sample values were
normalised to VSMOW via internal lab standards that were measured with the samples. As
for the CRDS data, SLAP of δ17O = -29.6986h and δ18O = -55.5h was used for SMOW-
SLAP scaling (Schoenemann et al., 2013). Analysis of SMOW (n = 15) and SLAP (n = 15)
during the course of this study gave -28.699 ± 0.011h (1σ ) and -53.622 ± 0.024h (1σ )
for the δ17O measured-SLAP and δ18O measured-SLAP values, respectively. This slightly
contracted scale relative to the IAEA-scale appears to be typical for MAT253 instruments
(Schoenemann et al., 2013).
The IRMS analysis of NEWGYP (n = 8, 3 or 4 analyses each) yielded mean values and
30 Calcium Sulfate Systematics
Fig. 2.8. Cross-plot of δ18O values in GHW of natural samples analysed by CRDS/MCM compared with
measurements of the same samples by IRMS. The long-term ±2σ error of the method (0.26h for
the δ18O values from the CRDS measurements and 0.28h for the δ18O values from the IRMS
analyses) is given for all samples. The in-sample ±2σ errors (7 injections for CRDS and 3-4 for
IRMS) are smaller than the data symbols.
reproducibility (±1σ ) of 0.28 ± 0.07h and 0.51 ± 0.14h for the δ17O and δ18O values,
respectively, which are within the 2σ error of the CRDS measurements. For most natural
samples, the δ17O, δ18O and δD values of GHW did not differ significantly whether or
not the MCM was used (within the 2σ error) from the results obtained by IRMS (Fig.
2.8). However, the 17O-excess of both the NEWGYP standard and natural samples differed
considerably between CRDS and IRMS when the MCM was not used. The 17O-excess
values measured by CRDS were considerably higher than those obtained by IRMS when
the MCM was turned off. By contrast, the 17O-excess results from CRDS were similar to
those achieved by IRMS when the MCM was used (Appendix A; Table A.3), resulting in a
close 1:1 relationship (Fig. 2.9). The difference in 17O-excess values with MCM On and
Off reflects the degree of spectral interference caused by contamination.
Using the MCM, the long-term precision (±1σ ) of the method for 17O-excess determin-
ation by CRDS of NEWGYP was ±8 per meg (n = 37) (Appendix A; Table A.1). In addition,
no systematic drift in the 17O-excess of the standard has been observed with time (Fig. 2.7).
Remarkably, this reproducibility is better than the typical in-sample precision obtained from
seven consecutive injections (±13 per meg) observed in both water standards and GHW.
Similar precision (typically ±15 per meg) is also obtained by randomly choosing the 17O-
excess of injections from NEWGYP samples, extracted on different days and analysed
in different CRDS analyser runs. This is a consequence of the minimised importance of
drift, memory effect and potential isotopic fractionation occurring during the 17O-excess
2.3 Hydration water extraction and analysis procedure 31
Fig. 2.9. Cross-plot of 17O-excess in synthetic and natural gypsum analysed by CRDS and IRMS. The
analyses by CRDS that used the MCM yielded results similar to those obtained by IRMS. Error
bars refer to the consecutive analyses (7 injections for CRDS and 3-4 analyses for IRMS) of the
same hydration water.
measurements, compared with the measurement of the δ17O and δ18O values. The repeated
extraction and measurement (n = 3, 7 injections each) of one natural gypsum sample (Salina
1) by CRDS using the MCM produced an analytical error (±1σ ) of ±7 per meg. This
indicates that the long-term 17O-excess precision of the method using CRDS is similar for
synthetic and natural gypsum when the MCM is powered on. The long-term precision is also
similar to that obtained for the analysis of water standards, as observed from the repeated
measurement of an internal water standard (SPIT, ±8 per meg, n = 23) in different runs over
the period of this study.
2.3.5 Discussion of the WASP/CRDS setup
A distinct advantage of CRDS over IRMS is the possibility of simultaneously determin-
ing D/1H, along with 17O/16O and 18O/16O, on the same hydration water sample. This
enables the calculation of both d-excess and 17O-excess in GHW. The long-term d-excess
of the NEWGYP standard (n = 37) was -53.4 ± 1.1h (Fig. 2.7). This reproducibility is
comparable with the long-term precision observed from the repeated measurement of the
SPIT water standard (±0.9h, n = 23). No systematic drift in the d-excess of the NEWGYP
standard has been observed with time (Fig. 2.7). As for the natural gypsum samples, the
d-excess in GHW analysed by CRDS using the MCM ranged from -39.7h to -71.9h. As
expected, the values are positively correlated with those of the 17O-excess across the dataset.
Use of the MCM removes impurities that affect the 17O-excess determination. The
differences in 17O-excess in natural samples in the MCM On/Off experiments are greater
in the case of gypsum samples generated in organic- and microbe-rich environments (e.g.
lake sediments). This, as expected, points to higher concentration of contaminants in these
32 Calcium Sulfate Systematics
types of material and suggests that the cause of the spectral interferences could be H2S
or organic compounds released by natural gypsum samples during dehydration. Further
evidence for the presence of impurities in GHW is given by a strong smell of H2S in water
after extraction, especially in water from lacustrine gypsum samples.
Results reveal that the use of the MCM accessory is crucial for accurate determination
of the 17O-excess in GHW by CRDS. In earlier work, this device was found to be necessary
for isotopic analysis of water samples with high concentrations of organic compounds, such
as water extracted from the xylem of plants that usually contains alcohols that spectro-
scopically interfere with the CRDS analysis (Brand et al., 2009; Saad et al., 2013). To
date, the efficiency of the MCM has been demonstrated for the removal of alcohols and
other organic compounds from water (Saad et al., 2013). However, catalytic oxidation
of other substances also occurs, such as H2S, volatile organic compounds and long-chain
compounds, and this is important for the analysis of GHW samples. Although this accessory
removes most impurities, we suspect that organic-rich gypsum samples (e.g. those from
lake sediments) may require additional pre-treatment to remove organic compounds (H2O2,
sodium hypochlorite, etc.). However, any pre-treatment procedure will require testing to
ensure that it does not alter the δ18O and δD values of the hydration water.
The WASP system also offers distinct advantages over previous methodologies. One
critical advance is the ability of the WASP extraction line to monitor internal pressure during
GHW extraction, thus detecting vacuum leaks in individual lines (Fig. 2.6). When operating
correctly, the pressure in the lines displayed similar profiles, first increasing up to 1 mbar
at around 120°C, and then decreasing to below 5×10−2 mbar at the end of the extraction
process. When O-rings are compromised, profiles showed slightly greater pressures (up to
1.2 mbar) at 120°C (Fig. 2.6). The pressure decreased by as much as 0.2 mbar when the LN2
trap was raised for a second time, but subsequently increased to around 1.4 mbar by the end
of the run. In the line run as a blank with no sample, the pressure increased constantly to
reach a final value of 0.7 mbar, and no drop in pressure occurred during the run (Fig. 2.6).
The isotopic composition of the NEWGYP standard extracted in the leaky line showed
no difference within error (0.03h for the δ17O value, 0.03h for the δ18O value and -
51.24h for the δD value) from that obtained from the properly functioning line (0.07h for
the δ17O value, 0.09h for the δ18O value, and -51.39h for the δD value). However, slight
differences were observed in the results of the ENR-gyp samples. The enriched gypsum
extracted in the leaking line showed slightly lower isotope values (12.78h for the δ17O
value, 24.58h for the δ18O value, and 33.56h for the δD value) than those from the
properly functioning line in the same run (12.96h for the δ17O value, 24.93h for the
δ18O value, and 34.63h for the δD value). This suggests that vacuum leaks can alter the
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 33
isotopic ratio values of samples extracted using the WASP, especially for hydration waters
with an isotopic composition significantly different from that of the ambient water vapour in
the laboratory atmosphere. Therefore, it is vital to monitor the pressure in the WASP during
GHW extraction to ensure that no leaks occur during extraction.
2.3.6 Conclusions
The method presented provides precise and simultaneous isotopic measurements of δ17O,
δ18O and δD values. The derived d-excess and 17O-excess, can be obtained from GHW
using the new extraction apparatus and subsequent measurement by CRDS.
2.4 Precise and accurate isotope fractionation factors (α 17O,
α 18O and αD) for water and CaSO4 ·2H2O (gypsum)
Adapted from the publication: Fernando Gázquez, Nicholas P. Evans and David A. Hodell
(2017). Geochimica et Cosmochimica Acta, 198, 259–270.
Isotope fractionation factors (α 17O, α 18O and αD) between GHW and free water of the
mother solution were determined in the temperature range from 3°C to 55°C at different
salinities and precipitation rates. F.G. and D.A.H. designed the experimental setup and
F.G. conducted the experiments. F.G. and N.P.E. extracted and analysed GHW and N.P.E.
performed all other analytical procedures.
2.4.1 Introduction
The isotopic fractionation factor (α) between the mother water and GHW is defined as:
αgypsum -water =
αgypsum+1000
αwater+1000
where αgypsum and αwater denote the isotopic ratio (i.e. 17O/16O, 18O/16O and 2H/1H) of
the hydration water and mother water, respectively, relative to VSMOW.
Calculating the α 18Ogypsum -water requires measurement of the relative difference in the
isotopic composition of GHW and the free water of the mother solution. Multiple methods
of gypsum precipitation have been employed, including evaporation of solutions saturated
in CaSO4 (e.g. Gonfiantini and Fontes, 1963), the hydration of γ -CaSO4 (e.g. Hodell et
34 Calcium Sulfate Systematics
al., 2012), and mixing of CaCl2 and Na2SO4 solutions (e.g. Gonfiantini and Fontes, 1963;
Fontes and Gonfiantini, 1967; Tan et al., 2014).
The first stable isotopic measurement of GHW was made by Baertschi (1953), who
reported an α 18Ogypsum -water of 1.0035. Seminal papers by Gonfiantini and Fontes (1963)
and Fontes and Gonfiantini (1967) investigated gypsum-water fractionation factors over the
temperature range of 17°C to 57°C. Gonfiantini and Fontes (1963) argued that their rehyd-
ration experiments gave a precise α 18Ogypsum -water = 1.0040, although two ‘steady state’
evaporation experiments produced an α 18Ogypsum -water = 1.0031. Overall, α 18Ogypsum -water
and αDgypsum -water were reported as 1.0037 (±0.0005) and 0.985, respectively. Within
analytical uncertainty, both α 18Ogypsum -water and αDgypsum -water were found not to be sens-
itive to temperature variability across the range investigated (Gonfiantini and Fontes, 1963;
Fontes and Gonfiantini, 1967). Fractionation factors produced by Matsubaya (1971) and
Sofer (1975) apparently confirmed α 18Ogypsum -water = ~1.0040 and αDgypsum -water = 0.980
although, because the datasets are not accessible, the precision and accuracy of these ex-
periments remain unclear. More recently, Hodell et al. (2012) found α 18Ogypsum -water =
1.0039±0.0004 and αDgypsum -water = 0.981 ± 0.002, which did not differ significantly from
previous values. As the majority of the reported gypsum-water fractionation factors are
within analytical error, most investigations that utilise αgypsum -water use rounded values of
α 18Ogypsum -water = 1.004 and αDgypsum -water = 0.98 (e.g. Sofer, 1978; Bath et al., 1987;
Khademi et al., 1997; Kasprzyk and Jasinska, 1998; Farpoor et al., 2004; Buck and Van
Hoesen, 2005; Hodell et al., 2012; Chen et al., 2016).
The opposite signs of α 18Ogypsum -water and αDgypsum -water mean that the oxygen iso-
topic composition GHW is enriched by ~4h compared to the isotopic composition of
mother water, whereas 2H is depleted by ~20h (e.g. Gonfiantini and Fontes, 1963; Fontes
and Gonfiantini, 1967). The enrichment of the oxygen isotope can be explained if the GHW
is related to the water of the hydration sphere of Ca2+, which is likely to be enriched in 18O
because of the difference in vibration frequencies of the two isotopic forms of water along
the bonds formed between water and cation (Taube, 1954; Gonfiantini and Fontes, 1963;
Longinelli, 1979; Oi and Morimoto, 2013). In contrast, the depletion of GHW in 2H by
~20hmay be ascribed to the fact that the hydration sphere of SO42− in solution is depleted
in 2H with respect to the free water (Oi and Morimoto, 2013).
Although the fractionation factors from different studies are in general agreement, the
reason for the slight scatter in reported αgypsum -water values and temperature/salinity effects
on the gypsum-water isotopic fractionation remains poorly understood. Firstly, the slight
scatter in reported αgypsum -water values may imply that full equilibrium between GHW and
the mother fluid are not always reached in experiments (possibly due to sluggish exchange
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 35
between the solution and GHW; Herwartz et al., 2017), or that αgypsum -water differs due
to the different methods of gypsum precipitation employed (which have traditionally been
thought of as equivalent). Alternatively, the drying process prior to GHW extraction may
differ between studies, and thus adsorbed water on the surface of the crystals may not
have been not effectively removed. Secondly, Hodell et al. (2012) found a slight positive
temperature dependence (0.00012 per °C) for αDgypsum -water between 12°C and 37°C,
although the full extent of this temperature dependence was not fully investigated. Any
temperature dependence will impact investigations of hydrothermal gypsum that use the
accepted α 18Ogypsum -water and αDgypsum -water of 1.004 and 0.98 (e.g. Matsuyaba and Sakai,
1973), as αgypsum -water may differ at the higher temperatures of hydrothermal gypsum pre-
cipitation (e.g. 55°C; Garofalo et al., 2010; Gázquez et al., 2012, 2013). Finally, the derived
uncertainty of αgypsum -water measurements is also unsatisfactory for many geological and
palaeoclimate applications. Standard errors of ±0.0004 and ±0.002 for α 18Ogypsum -water
and αDgypsum -water reported by Hodell et al. (2012) result in a derived mother fluid error of
~0.4h and ~2h for δ18O and δD, respectively.
Here the fractionation factors for gypsum are calculated by measuring the relative dif-
ference in the isotopic composition of the hydration water and the free water of the mother
solution. Gypsum precipitation experiments were conducted at a range of temperatures (3°C
to 55°C) and salinities (0 to 300 g l-1). The effect of precipitation rate on isotope fractiona-
tion factors was evaluated by changing the initial concentration of CaCl2 and Na2SO4.
2.4.2 Methods
2.4.2.1 Hydration of anhydrous CaSO4
100 ml of 0.5 M Na2SO4 solution was added to 1.5 g of analytical grade powdered anhydrite
(Acros, UK) to induce hydration to gypsum (Conley and Bundy, 1958). Sodium sulfate
acts a catalyst during the hydration reaction of anhydrite, and acceleration of the reaction
takes place through the medium of transient surface complexes that are unstable in dilute
solution and finally evolve to gypsum (Conley and Bundy, 1958). Importantly, the hydration
sphere of Na+ does not isotopically affect the activity of water in dilute solutions; thus, the
addition of Na2SO4 does not interfere with isotopic fractionation during gypsum precipita-
tion (Taube, 1954; Gonfiantini and Fontes, 1963). The conversion of anhydrite to gypsum
occurs in two steps via the intermediate hemihydrate of calcium sulfate (i.e. bassanite) (Van
Driessche et al., 2012; Wang et al., 2012).
Prior to the experiment, anhydrite was heated to 400°C for 3 h to completely remove
all hydration and adsorbed water. Subsequent analysis by ATR-FTIR (Attenuated Total
36 Calcium Sulfate Systematics
Reflectance, Bruker Platinum accessory, coupled to a Fourier Transformation Infrared Spec-
trometer instrument, Bruker, Tensor II; Department of Earth Sciences, University of Cam-
bridge, UK) detected only anhydrite, with no traces of gypsum or bassanite. X-ray diffrac-
tion (XRD) analysis of the same material showed insignificant amounts of bassanite (0.5%)
and gypsum (0.4%).
Experiments were performed in duplicate or triplicate in a conical flask with a greased
ground glass stopper and joint to prevent evaporation. Before mixing, the anhydrite and the
Na2SO4 solutions were temperature equilibrated in an oven at 20°C, 25°C, 30°C, 35°C and
40°C for at least 2 h. Anhydrite was added to the solution and then constantly stirred at
700 rpm for 24 h using a wireless magnetic stirrer. After mixing, the flasks containing the
solution and the anhydrite were placed again in the oven and temperature was maintained
to a precision of ±0.1 °C (1σ ).
Experiments at the lower temperatures of 8°C and 3°C were conducted in a refrigerator
and cold room, respectively, in which temperatures were monitored for the duration of the
experiments and varied by less than 0.5°C.
At temperatures greater than 45°C, the conversion of anhydrite to gypsum was incom-
plete (e.g. 42 wt% of gypsum at 45°C or 0 wt% at 60°C) because of the greater stability
of anhydrite relative to gypsum at temperatures above ~42°C (Ostroff, 1964; D’Ans, 1968).
Thus, we only consider the experiments in which the hydration of anhydrous CaSO4 resulted
in over 98 wt% conversion to gypsum (i.e. experiments performed at temperature below
40°C). A water sample (200 µl) from each experiment was stored for subsequent isotopic
analysis. We found that during temperature equilibration in the experiments at 3°C, Na2SO4
precipitation occurred because of a rapid decrease in solubility at temperatures below 10°C.
For this reason, experiments at 3°C used a 0.05 M Na2SO4 solution.
After 24 h, the solutions were vacuum filtered using Millipore nitrocellulose filters (0.45
µm ϕ pore). Samples were then dried at 45°C for 48 h. Thermogravimetric analysis
(Netzsch STA 449 F1 Jupiter) showed that this drying method removed all adsorbed water,
but did not result in the loss of hydration water. The mineralogy of the dry precipitates was
analysed by XRD.
2.4.2.2 Mixing CaCl2 and Na2SO4 solutions
Gypsum saturation was achieved by mixing solutions of CaCl2 ·6H2O and Na2SO4 of
varying concentrations to control the rate of precipitation. Experiments with three different
initial Ca2+ and SO42− concentrations (0.5 M, 0.33 M and 0.125 M) were conducted by
diluting mother solutions of 0.5 M CaCl2 ·6H2O (analytical grade, Sigma–Aldrich) and
0.5 M Na2SO4 (analytical grade, Fisher Scientific).
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 37
Experiments were conducted at 5°C, 20°C, 25°C, 45°C and 55°C using a water bath
(±0.1°C) and solutions were temperature equilibrated for 2 h prior to mixing. A sample of
each solution was collected after mixing. In order to promote slower gypsum precipitation,
no stirring or shaking of these gypsum precipitation experiments occurred (referred to as
free-drift experiments hereafter). Experiments at 25°C and 55°C were repeated by setting
the shaking mode of the bath at 110 rpm to promote solution homogenisation and fast
gypsum precipitation (referred to as shaking experiments hereafter). Both sets of experi-
ments lasted 10 days. Subsequently, solutions were filtered and the precipitate was dried and
analysed by XRD using the same method employed in the anhydrite hydration experiments
(Section 2.4.2.1). A sample of each solution was stored after filtering. The saturation index
of gypsum (SIgyp) in the initial solution was calculated using PHREEQC (3.1.7) (Parkhurst
and Appelo, 2013).
2.4.2.3 Gypsum precipitation in brines
Experiments examining gypsum precipitation from brines of varying salinity used the same
methodology described in Section 2.4.2.1, but varying amounts of NaCl were added to the
initial 0.5 M Na2SO4 solutions. The NaCl concentrations in the final solution were 30 g l-1,
80 g l-1, 150 g l-1, 200 g l-1 and 300 g l-1. All the experiments were conducted in duplicate at
20°C and used the same procedure for filtering, drying, and mineralogical analysis described
in Sections 2.4.2.1 and 2.4.2.2.
2.4.2.4 Isotopic analyses and calculation of fractionation factors
The GHW was extracted and water isotopes were analysed using the same methods de-
scribed in Section 2.3. The mother waters collected from each experiment and the corres-
ponding hydration water extracted from the gypsum were measured consecutively by CRDS
under the same instrument conditions. This direct comparison minimises the effect of drift
and provides precise and accurate estimates of the fractionation factor that is calculated
as the isotopic difference between the sequential samples of the mother water and GHW.
By analysing the waters before and after gypsum precipitation, we found that the isotopic
composition of the solution did not change over the course of the experiments within error
(Appendix B; Table B.1); therefore, the values of the initial solutions were used for the
calculation of the isotope fractionation factors.
38 Calcium Sulfate Systematics
Temperature (˚C) α 17O* 1σ α 18O* 1σ θ 1σ n* α 18O# 1σ αD# 1σ n#
3 1.00197 0.00005 1.00372 0.0001 0.53 0.0011 4 1.0037 0.0001 0.9788 0.0003 4
8 1.00201 0.00009 1.0038 0.00016 0.5293 0.0014 4 1.0037 0.0002 0.9801 0.0006 6
20 n.a - n.a - - - - 1.0034 0.0001 0.9814 0.0011 4
25 1.00181 - 1.00341 - 0.5298 - 1 1.0035 0.0001 0.981 0.0006 5
30 1.00174 0.000 1.00329 0.00001 0.5289 0.0014 1 1.0033 0.0001 0.9803 0.0009 2
35 n.a - n.a - - - - 1.0034 0.000 0.9812 0.0003 4
40 1.00169 - 1.00318 - 0.5314 - 1 1.0034 0.0002 0.9821 0.0009 5
Table 2.1. Isotope fractionation factors (α 17Ogypsum -water, α 18Ogypsum -water and αDgypsum -water) between
GHW and its mother solution obtained experimentally by the hydration of anhydrite. See
Appendix B for complete report. (*Analyses conducted with Picarro L-2140i analyser;
#averaged results of samples analysed by Picarro L-1102i and L-2140i analysers).
2.4.3 Results
2.4.3.1 Hydration of anhydrite at low salinities
The α 18Ogypsum -water values of the anhydrite hydration experiments varied from 1.0033 ±
0.0001 to 1.0037 ± 0.0001 in the temperature range from 3°C to 40°C, with the lowest values
at 30°C and the highest values at 3-8°C. There was no statistically significant trend over this
temperature range given the analytical uncertainty of the measurements (R2 = 0.56; p-value
>0.05). In contrast, αDgypsum -water increased with temperature from 0.9788 ± 0.0003 at
3°C to 0.9821 ± 0.0009 at 40°C and showed significant dependence with temperature (R2 =
0.89; p-value <0.05) (Table 2.1, Fig. 2.10, Appendix B; Table B.1).
The α 17Ogypsum -water was determined in the experiments of hydration of CaSO4 and
varied in the same manner as α 18Ogypsum -water reflecting mass-dependent fractionation in
the triple oxygen isotope system (Cao and Liu, 2011), with minimum value of 1.0017
± 0.0001 in the temperature range from 25°C to 40°C and maximum of 1.0020 at 8°C
(Table 2.1). The relationship between α 17Ogypsum -water and α 18Ogypsum -water is given by
the parameter θ (Mook, 2000):
α 17Ogypsum -water = α 18Ogypsum -waterθ
where θ = ln
(
α 17Ogypsum -water
)
/ ln
(
α 18Ogypsum -water
)
.
θ was found to be 0.5297 ± 0.0012 (n = 11) and displayed no correlation with tem-
perature (Table 2.1). All gypsum samples yielded a weight loss of hydration water of
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 39
Fig. 2.10. Isotope fractionation factors (A: α 18Ogypsum -water; B: αDgypsum -water) between GHW and its
mother solution at different temperatures obtained experimentally by hydration of anhydrite and
mixing of CaCl2 and Na2SO4 solutions. Note that mixing experiment results are averaged values
of gypsum precipitation at different initial SIgyp of the solution (see Table 2.2 and Appendix B).
Error bars denote 1σ . Grey shades represent the 95% confidence limits.
40 Calcium Sulfate Systematics
Temperature (°C) 0.25 Ca2+/0.25 SO42− (mol l
-1
) 0.166 Ca2+ /0.166 SO42− (mol l
-1
) 0.065 Ca2+ /0.065 SO42− (mol l
-1
)
α 18O 1σ αD 1σ α 18O 1σ αD 1σ α 18O 1σ αD 1σ
5 1.0035 0.0001 0.978 0.001 1.0035 0.0002 0.977 0.003 1.0035 0.0003 0.976 0.000
20 1.0038 0.0001 0.983 0.001 1.0036 - 0.98 - 1.0037 0.000 0.979 0.000
25 1.0036 0.0001 0.982 0.000 1.0035 0.0001 0.981 0.000 1.0035 0.000 0.98 0.001
45 1.0032 0.0001 0.984 0.000 1.0029 0.000 0.983 0.000 1.0032 - 0.982 -
55 1.0034 0.0001 0.984 0.001 1.0033 0.0001 0.984 0.001 1.0031 0.000 0.983 0.001
Table 2.2. Isotope fractionation factors (α 17Ogypsum -water, α 18Ogypsum -water and αDgypsum -water) between
GHW and its mother solution obtained experimentally by mixing of CaCl2 and Na2SO4 solutions
at different initial concentrations of Ca2+ and SO42− and different temperatures. Initial Ca2+ and
SO42− concentrations are given. SIgyp ranged from 0.66 to 1.50 (Appendix B). Analyses were
conducted with a Picarro L-1102i analyser.
over 20%, similar to that of the gypsum standard (20.5 ± 0.3%). This suggests that there
was complete hydration of anhydrite to gypsum and complete mineral dehydration in the
extraction procedure for GHW. Insignificant amounts of bassanite and anhydrite (less than
1 wt%) were detected in all samples (Appendix B; Table B.4).
2.4.3.2 Gypsum precipitation from mixing of CaCl2 and Na2SO4 solutions
Experiments of gypsum precipitation from mixing of CaCl2 and Na2SO4 solutions produced
a mean α 18Ogypsum -water value of 1.0034 ± 0.0003 in the temperature range from 5°C to
55°C. No measurable differences were found between the free-drift gypsum precipitation
and the shaking experiments. No significant relationship was observed with temperature
considering the analytical uncertainty of the measurements (R2 = 0.73; p-value >0.05). The
α 18Ogypsum -water value was unaffected by the different initial SIgyp of the solution within
the analytical error (Figs 2.10 and 2.11 and Table 2.2).
Unlike the oxygen isotope fractionation factor, αDgypsum -water was affected by changes
in both temperature and initial SIgyp. The average αDgypsum -water was 0.9815 ± 0.0025
between 5°C and 55°C. No measurable differences were found between the free-drift gypsum
precipitation and the shaking experiments. The experiments conducted at 55°C produced
the highest values of αDgypsum -water (0.984 ± 0.001) compared with gypsum precipitation
at lower temperatures (e.g. 0.977 ± 0.001 at 5°C). Hydrogen isotope fractionation shows a
clear increasing trend with temperature by 0.0001 units per °C (R2 = 0.94; p-value <0.05;
taking the average of the experiments at different initial SIgyp). The αDgypsum -water was
affected by the SIgyp of the initial solution, increasing by 0.0033 for each increase of 1 unit
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 41
Fig. 2.11. Isotope fractionation factors (A: α 18Ogypsum -water; B: αDgypsum -water) between GHW and its
mother solution obtained experimentally by mixing of CaCl2 and Na2SO4 solutions at different
initial concentrations of Ca2+ and SO42− and temperatures. Error bars denote 1σ .
42 Calcium Sulfate Systematics
in the SIgyp. This linear trend is similar (±0.0012, 1σ ) at different temperatures. Less than
1 wt% of bassanite and anhydrite has been detected in all samples (Appendix B; Table B.4).
2.4.3.3 Gypsum precipitation from brines
The addition of NaCl to the solution does not affect α 18Ogypsum -water (1.0033 ± 0.0001)
below 150 g l-1 of NaCl, and results were similar to gypsum precipitation at 20°C when no
NaCl was added (1.0034 ± 0.0001) (Table 2.3 and Fig. 2.12). In contrast, α 18Ogypsum -water
increased to 1.0038 ± 0.0002 at a salinity of 200 g l-1 and to 1.0047 ± 0.0003 at 300 g l-1.
The αDgypsum -water increased linearly with salinity from 0 to 300 g l-1 NaCl by 0.00003 units
per gram of NaCl in solution and showed the greatest value of 0.9893 ± 0.0002 at 300 g l-1
(Table 2.3). The water content of the solids was 20.7 ± 0.3% in experiments performed
at salinities between 30 and 150 g l-1 of NaCl, whereas the water yield was slightly lower
(19.7 ± 0.1%) in the experiments at 200 g l-1 and 300 g l-1. Mineralogical analyses by XRD
detected only gypsum in the experiments at NaCl concentrations below 150 g l-1 and small
amounts of bassanite and unconverted anhydrite of up to 2.5% and 4.6% respectively at
300 g l-1(Appendix B; Table B.3).
2.4.4 Discussion
2.4.4.1 Effect of temperature and precipitation rate on the isotope fractionation factors
in gypsum hydration water
Most equilibrium fractionation factors between the solution and the solid phase approach
unity (i.e. no fractionation between the solution and the mineral, α = 1) with increasing
temperature (Friedman and O’Neil, 1977). In the experiments, αDgypsum -water increases
with temperature in the gypsum formed using two methods (conversion of anhydrite to
gypsum and mixing of CaCl2 and Na2SO4 solutions). These methods produce similar
results for α 18Ogypsum -water and αDgypsum -water in the range of temperature studied within
analytical uncertainties (Fig. 2.10). In the case of α 18Ogypsum -water, no significantly stat-
istical relationship with temperature was observed when examining separately the results
from each method of gypsum precipitation (p-value >0.05 in both cases). However, when
combining the results from both methods, a slight dependence of α 18Ogypsum -water with
temperature becomes significant (R2 = 0.70; p-value <0.05; Fig. 2.10) because of the
increased number of observations. The equation for α 18Ogypsum -water (expressed as 1000
ln(α 18Ogypsum -water)), as a function of temperature (°C) is:
1000ln
(
α 18Ogypsum -water
)
=−0.009 (±0.004)T +3.68 (±0.13)
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 43
whereas the temperature dependence of αDgypsum -water is given by:
1000ln(αDgypsum -water) =−0.104 (±0.053)T −22.05 (±1.99)
This very small dependence of α 18Ogypsum -water on temperature is not relevant for many
geological and palaeoclimate applications; however, the greater sensitivity of αDgypsum -water
to temperature has implications for the calculation of δD of the mother solution, especially
when gypsum forms in hydrothermal environments from hydration of anhydrite (e.g. Mat-
suyaba and Sakai, 1973; Bath et al., 1987). For example, using the revised fractionation
factor value at 55°C (1.0033 ± 0.0002), the inferred value of δ18O in the mother water
increases by 0.7h compared to using the fractionation factor of 1.004 (Gonfiantini and
Fontes, 1963; Sofer, 1978). δD value decreases by ~4h using the revised fractionation
factor (0.984) instead of the accepted value (0.980; Sofer, 1978). Hence, the temperature
of gypsum formation should be considered when choosing which fractionation factors to
apply.
At lower temperatures (i.e. from 20°C to 40°C), the two methods of gypsum precipit-
ation produce slightly lower values for α 18Ogypsum -water, but similar αDgypsum -water values
(within analytical error) compared to previously proposed values (Gonfiantini and Fontes,
1963; Fontes and Gonfiantini, 1967; Sofer, 1978; Hodell et al., 2012). For α 18Ogypsum -water,
we obtained a value of 1.0034 ± 0.0001 from 20°C to 40°C, which is within the error
reported previously (i.e. 1.0037 ± 0.0005 in Gonfiantini and Fontes, 1963; 1.0039 ± 0.0004
in Hodell et al., 2012), but mostly closer to the lower range of values and more precise. For
αDgypsum -water, obtained values from 20°C to 40°C (0.9812 ± 0.0007) are within error of
previous measurements (0.981 ± 0.002 between 12 and 37°C; Hodell et al., 2012), but are
also more precise.
By using the proposed α 18Ogypsum -water value of 1.0034 ± 0.0001, any calculation of
mother water in the range from 20°C to 40°C produces a δ18O of water that is ~0.6h
higher than if the fractionation factor of 1.004 is used instead (Appendix B; Table B.5).
This results in significant differences for quantitative isotopic studies using GHW, such as
the tandem carbonate-GHW palaeothermometer (Hodell et al., 2012). Values of δ18O of
mother water that are 0.6h greater will lead to water temperatures that are approximately
2°C higher than those calculated using a fractionation factor of 1.004.
CaCl2/Na2SO4 mixing experiments at different initial SIgyp were performed to calculate
the effect of precipitation rate on the isotope fractionation factors. Results show that, at
different initial SIgyp, within the studied range of saturations (SIgyp = 0.66-1.50) there are no
measurable differences in α 18Ogypsum -water (Table 2.2), independent of gypsum formation
temperature. Unlike α 18Ogypsum -water, the hydrogen fractionation factor, however, increases
44 Calcium Sulfate Systematics
at a constant rate with SIgyp at all temperatures by 0.003 ± 0.001 per unit of SIgyp.
The studied range of supersaturation is ~5 to ~20 times greater than gypsum saturation
under equilibrium conditions (e.g. 0.014 M of dissolved CaSO4 at 25°C). Equilibrium or
near-equilibrium conditions are expected during the precipitation of most natural gypsum
deposits. However, the experiments at relatively elevated saturations are useful to ascertain
the role of kinetic effects on isotopes fractionation during gypsum precipitation.
The fractionation factors between GHW and free water obtained experimentally are the
net result of equilibrium and kinetic effects for oxygen and hydrogen isotopes. The relative
importance of the two is governed by the rate of gypsum precipitation. Pure equilibrium-
controlled fractionation may occur during slow gypsum precipitation, whereas kinetic iso-
topic fractionation is more likely at higher precipitation rates. This is demonstrated by
the increase in αDgypsum -water with increasing saturation (SIgyp). This suggests that the
equilibrium fractionation (i.e. SIgyp = 0) for hydrogen isotopes may be lower than the
values obtained from CaCl2/Na2SO4 mixing experiments at different temperatures, in which
αDgypsum -water is partially controlled by kinetic effects.
The fact that α 18Ogypsum -water does not show measurable trends with SIgyp may indicate
that kinetic effects are minimal for oxygen isotopes. The α 18Ogypsum -water is controlled by
isotopic fractionation between the free solution and the hydration sphere of Ca2+ in solution
(Taube, 1954; Gonfiantini and Fontes, 1963; Oi et al., 2013). The results suggest that
different calcium concentrations in the solution and SIgyp do not affect the isotopic values
of the hydration sphere of Ca2+, within the range of experimental conditions investigated.
No measurable differences in α 18Ogypsum -water and αDgypsum -water were observed between
the free-drift and the shaking experiments performed at the same SIgyp and temperature.
This is because the initial saturations used in the experiments are relatively far from the
gypsum precipitation equilibrium. At these saturation levels, there is little difference in the
rate of gypsum nucleation and precipitation between both types of experiments, as gypsum
crystallisation occurs immediately after mixing the initial CaCl2 and Na2SO4 solutions.
The relative difference in αDgypsum -water between the experiments conducted at lower
initial SIgyp (0.66-0.70) and those at higher SIgyp (1.39-1.44) is 0.0024 ± 0.0010. This
suggests that holding all other parameters constant (isotopic composition of the solution,
temperature, etc.), faster precipitation of gypsum causes enrichment of hydrogen isotopes in
GHW by ~2.4h with respect to the experiments with the lowest initial SIgyp (and therefore
the slowest rates of precipitation), within the studied range of SIgyp. Unlike the insignific-
ant effect of different initial SIgyp and calcium concentrations on the hydration sphere of
Ca2+, the isotopic values of the hydration sphere of SO42− are affected by different ionic
concentrations and gypsum precipitation rates, within the range of experimental conditions
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 45
investigated. Although more experiments are needed to explore αDgypsum -water closer to
the saturation point of gypsum, we assume that the relationship between αDgypsum -water
and SIgyp follows the same linear relationship at lower saturations. The αDgypsum -water
values are extrapolated to for SIgyp = 0 and to estimate the expression of pure equilibrium
fractionation, with application to cases when gypsum precipitates under near-equilibrium
conditions.
The results indicate that at SIgyp = 0, the value of αDgypsum -water should be 0.0046 ±
0.0019 lower than in the experiments with faster gypsum precipitation (i.e. SIgyp = 1.39–
1.50). Therefore, at the slowest gypsum growth rates, δD in GHW is depleted by ~4.6h
when compared with the gypsum formed at the faster precipitation rates in the experiments
at any given temperature.
This finding has potentially important implications for accurate calculations of δD of
the fluid from GHW (particularly for determining d-excess values), especially in gypsum
crystals formed at low saturation state. This is the case for the megacrystals of the caves in
the Naica mine (Chihuahua, Mexico) (García-Ruiz et al., 2007; Gázquez et al., 2012, 2013),
where gypsum speleothems grew from a solution with SIgyp close to 0 and temperature
around 47-55°C.
2.4.4.2 Effect of salinity on the isotope fractionation factors in gypsum hydration
water
The initial salinity (NaCl concentration) of the solution also controls the fractionation factor
between water and gypsum. Results suggest that a α 18Ogypsum -water is not affected by NaCl
concentrations below 150 g l-1. Above 150 g l-1, the fractionation of oxygen isotopes in-
creases gradually from 1.0034 below 150 g l-1 to 1.0047 ± 0.0003 at 300 g l-1. For hydrogen,
the αDgypsum -water increases linearly with salinity from 0.9806 to 0.9893 between 0 g l-1
and 300 g l-1 (Table 2.3 and Fig. 2.12). Therefore, gypsum that precipitates from a 300 g l-1
NaCl solution has a δD that is ~10h lower than gypsum precipitated from freshwater with
the same isotopic composition (when all other variables are held constant). Given that
gypsum formed from evaporated seawater starts to precipitate when the solution reaches
a salinity of ~130 g l-1, the effect of salinity on δ18O hydration water is small for most
gypsum precipitates. However, the hydrogen isotopes fractionate by 3h less at a salinity
of 130 g l-1 compared to freshwater. This implies that when the conventional fractionation
factors are applied to gypsum formed from evaporated seawater (i.e. ~130 g l-1), the values
of d-excess are 10h more positive than when the revised fractionation factors are used
instead (assuming no salinity effect on δ18O).
The effect of NaCl concentration on the fractionation factor between water and gypsum
46 Calcium Sulfate Systematics
Gyp Halite
Fig. 2.12. Isotope fractionation factors (A: α 18Ogypsum -water; B: αDgypsum -water) between GHW and its
mother solution at different salinities obtained experimentally by hydration of anhydrite in
solutions with different amounts of NaCl. Error bars denote 1σ . The vertical dashed lines indicate
the approximate salinities at which gypsum and halite begin to precipitate during the evaporation
of seawater.
2.4 Isotope fractionation factors for water and CaSO4 ·2H2O 47
NaCl concentration (g l-1) α 18O 1σ αD 1σ n
0 1.0034 0.0001 0.9806 0.001 4
30 1.0032 0.0001 0.9807 0.0003 2
80 1.0034 0.0003 0.9824 0.0004 2
150 1.0033 0.0001 0.9846 0.0001 3
200 1.0038 0.0002 0.9886 0.0004 2
300 1.0047 0.0003 0.9893 0.0002 2
Table 2.3. Isotope fractionation factors (α 18Ogypsum -water and αDgypsum -water) between GHW and its
mother solution obtained experimentally by hydration of anhydrite from solutions with different
concentrations of NaCl at 20°C. See Appendix B for complete data report. Measurements were
made using a Picarro L-1102i analyzer.
can be attributed mainly to decreases in the activity of water as salinity increases, which is
related to the effect of Cl− on the hydration spheres of Ca2+ and SO42− (Di Tommaso et al.,
2014). As a consequence, the activity and isotopic ratios of water in brines are not the same
as for freshwater (Sofer and Gat, 1975). Another explanation of the effect of NaCl on the
fractionation factor is that the precipitation of intermediate hydrated calcium sulfate phases
(e.g. bassanite) could affect the fractionation factors between water and gypsum at high
salinities. Indeed, the stability of bassanite increases with increasing NaCl concentration
(Ostroff, 1964; Ossorio et al., 2014). Considering these results, the salinity of the solution
from which gypsum precipitated should be considered for calculations of the original δ18O,
δD, and derived d-excess from gypsum precipitated from brines (i.e. evaporative marine
gypsum).
2.4.4.3 Triple oxygen isotope fractionation in gypsum hydration water
The parameter θ , which describes the relationship between α 17O and α 18O (α 17Ogypsum -water
= α 18Ogypsum -waterθ ), has been determined for water-GHW. We observed θ = 0.5297 ±
0.0012 (1σ ) in the experiments of hydration of anhydrite and θ is independent of temperat-
ure. The observed θ value is close to the greatest theoretical values of this parameter in any
mass-dependent fractionation process of triple oxygen isotope, which ranges from 0.52 to
0.5305 (Matsuhisa et al., 1978; Cao and Liu, 2011; Bao et al., 2016). This θ value agrees
with that given by Barkan and Luz (2005) in vapour-liquid water equilibrium (θ = 0.529 ±
0.001), as well as other equilibrium mass-dependent reactions for the triple oxygen isotope
system (Miller, 2002; Cao and Liu, 2011; Bao et al., 2016). Although no dependence of
θ with temperature has been detected in the experiments, this parameter increases with
48 Calcium Sulfate Systematics
temperature in most water–mineral systems during oxygen isotope fractionation. However,
in the temperature range from 0°C to 50°C no measurable trends (within analytical un-
certainties of the current methods) are expected for most geochemical systems, including
CO2-water, quartz-water and calcite-water (Cao and Liu, 2011), agreeing with observations
in the GHW-water system.
2.4.5 Conclusions
The αDgypsum -water is 0.9812 ± 0.0007 at 20°C. Between 3°C and 55°C and salinities
<150 g l-1 of NaCl, αDgypsum -water was found to increase by 0.0001 units per °C. The
α 18Ogypsum -water displays a very small decrease with temperature of 0.00001 per °C, which
is not significant over much of the temperature range considered for palaeoclimate applica-
tions. Between 3°C and 55°C, α 18Ogypsum -water averages 1.0035 ± 0.0002. Precipitation of
gypsum in NaCl concentrations below 150 g l-1 did not significantly affect α 18Ogypsum -water,
but αDgypsum -water increases linearly with NaCl concentrations even at relatively low salin-
ities. It is necessary, therefore, to apply a salt correction for gypsum formed from brines.
Fractionation factors for 17O and 18O were found to follow the theoretical mass-dependent
fractionation on Earth (θ = 0.529 ± 0.001). Specific examples of the importance of using
the revised fractionation factors when calculating the isotopic composition of the fluids are
given in Gázquez et al. (2017).
2.5 Gypsum-water isotopic exchange
Once accurate and precise gypsum-water fractionation factors are known, it is necessary
to understand the conditions under which gypsum’s structurally-bound water can undergo
exchange with secondary waters. GHW can only be used for accurate palaeoenvironmental
reconstruction if gypsum deposits are both primary and have not undergone (i) dehydration
or (ii) recrystallisation and isotope exchange, thus retaining the original water isotopic
composition (Hodell et al., 2012). Knowledge of the mechanisms of gypsum dehydration,
recrystallisation and exchange is needed in ascertain the extent to which GHW is retained
within the crystalline structure, and thus for the subsequent interpretation of hydration water
signatures.
2.5.1 Gypsum dehydration
Gypsum may dehydrate due to elevated temperatures during sediment burial and exhuma-
tion cycles or hydrothermal processes (e.g. Warren, 2016). Numerous studies have ex-
2.5 Gypsum-water isotopic exchange 49
amined the thermally induced dehydration of gypsum, focusing on the dehydration mech-
anisms from gypsum through to α-anhydrite, the associated empirical activation energies,
and the occurrence and duration of the intermediate phases during the dehydration (see
Ballirano and Melis, 2009). There is a currently a consensus that the thermal dehydration of
gypsum occurs in a two-step process (e.g. McAdie, 1964; Abriel et al., 1990; Christensen
et al., 2008; Paulik et al., 1992; Deutsch et al., 1994; Chang et al., 1999; Robertson and
Bish, 2013; Prieto-Taboada et al., 2014), the final dehydrated product being reached via the
formation of the intermediate hemihydrate, bassanite:
CaSO4 ·2H2O−−→ CaSO4 ·0.5H2O+1.5H2O−−→ γ -CaSO4+0.5H2O
During the analysis of large, single crystals, however, it is reported that γ-anhydrite can
be directly produced without the appearance of a hemihydrate intermediate (e.g. Badens
et al., 1998; Sarma et al., 1998; Prasad et al., 2005). Reported activation energies of the de-
hydration reaction also range from 60 kJ mol-1to 500 kJ mol-1 between studies (e.g. Strydom
et al., 1995; Ballirano and Melis, 2009), and inconsistencies in experimental analysis have
resulted in a wide range of reported initial dehydration temperatures (~57°C to ~100°C;
Prasad et al., 2001; Prieto-Taboada et al., 2014).
Discrepancies between previous studies have been attributed to multiple factors that
include the rate at which gypsum is heated, atmospheric relative humidity and/or water
vapour pressure, crystalline size and sample purity (e.g. Khalil, 1982; Paulik et al., 1992;
Badens et al., 1998; Harrison, 2012; Robertson and Bish, 2013). Results may also be
affected by the nature of the samples analysed; the majority of previous studies examine
powdered gypsum samples, while some utilise polycrystalline samples but with very small
mean grain sizes (e.g. between 45 µm and 123 µm; Fusseis et al., 2012). The samples
previously analysed are often not applicable to gypsum seen in natural environments, such
as the mm to cm scale crystals examined in palaeoclimatological studies of Hodell et al.
(2012), Gázquez et al. (2013), and Grauel et al. (2016) (among others).
Here, in situ confocal Raman spectroscopy (RS) and tandem thermogravimetric (TGA)
and differential scanning calorimetry (DSC) measurements were made on different crystal-
line grain sizes of gypsum to both determine the stability of the intermediate bassanite phase
during the thermal dehydration of gypsum, and to understand the timing of the dehydration
and rehydration reactions.
50 Calcium Sulfate Systematics
2.5.2 Thermal dehydration of gypsum monitored by in situ Raman
spectroscopy and TGA/DSC
A systematic study was conducted to determine (1) the timing of the dehydration of gypsum
as a function of temperature and grain size of the starting material, and (2) to investigate
the stability of the intermediate bassanite phase during the dehydration reaction. in situ
confocal RS and TGA and DSC measurements were made on four different crystalline grain
sizes of gypsum. To minimise intra-method variability, TGA/DSC and RS experiments were
conducted at the same internal conditions – this included the same sample housing, gas flow
(and thus water vapour pressure) and heating rates. Experiments were conducted at both
isothermal temperatures and at continuous incremental temperature ramps (e.g. 1°C min-1,
5°C min-1 etc.). Experiments were designed by N.P.E. and carried out by N.P.E. and Gernot
Nehrke at the Alfred Wegener Institute, Bremerhaven.
2.5.2.1 Methods
An analytical-grade synthetic gypsum powder, NEWGYP (Fisher Scientific, C/2360/50;
Gazquez et al., 2015), was used for experimental analysis. Differing grain size fractions
were obtained by sieving the powder so that the sample grain sizes were within the range of
two mesh diameters; 38 to 64 µm, 64 to 125 µm, and 125 to 300 µm. Gypsum with larger
grain sizes (>500 µm) was obtained from fragments of a natural gypsum sample obtained
from the Messinian deposits of the Sorbas Basin (Evans et al., 2015). The individual crystals
were examined under a confocal microscope prior to analysis. Only crystals that contained
a negligible volume of fluid inclusions were used. Hereafter, the differing grain sizes are
referenced respect to their mean grain size (rounded to the nearest 50 µm), and are therefore
referred to as ‘50 µm gypsum’, ‘100 µm gypsum’ ‘200 µm gypsum’, and ‘500 µm gypsum’,
respectively.
Isothermal experiments were conducted at 60°C, 70°C, 80°C, 90°C, 100°C and 120°C,
where a ramp rate of 1°C min-1 was used to reach the desired temperature. Non-isothermal
experimental conditions implemented during individual runs using TGA, DSC and RS are
outlined in Table 2.4. To maintain constant relative humidity between analyses, all experi-
ments were performed under a dry N2 atmosphere.
The dehydration of gypsum was investigated by means of a Mettler Toledo TGA/DSC
instrument equipped with an auto sampler. Data was collected and processed using the
Mettler Toldeo Star software version 11. The gas flow was set to 40 ml min-1 N2 (20 ml
sample gas, and 20 ml protective gas).
RS measurements were made using a WITec alpha 300 R (WITec GmbH, Germany)
2.5 Gypsum-water isotopic exchange 51
Experiment ID Temperature Ramp (°C min-1) Tmax (°C) Crucible lid on/off
1 1 200 Off
2 5 200 Off
3 10 200 Off
4 20 200 Off
5 1 200 On
6 5 200 On
7 10 200 On
8 20 200 On
Table 2.4. Reference table of experimental conditions used for TGA, DSC and RS analysis.
CRM housed at the BioGeosciences section at the Alfred Wegener Institute, Bremerhaven.
A 488 nm diode laser was used for all Raman measurements. Spectra were acquired using
an UHTS 300 spectrometer equipped with an EMCCD camera. Raman spectra have been
obtained using a grid with 600 lines and for high spectral resolution with 2400 lines. A
Linkam THMSG600 temperature controlled stage connected to an LNP95 LN2 cooling
system was placed under the microscope and a Zeiss objective (LD Plan-Neofluar 20x / NA
0.4 corr) was used for all CRM measurements. The chamber of the temperature-controlled
stage was sealed with a thin cover glass and the N2 flow inside the chamber adjusted to
the same flow used in the TGA runs (40 ml min-1). The gas flow was determined using an
Agilent Technologies, ADM1000 universal gas flowmeter connected to the outflow of the
temperature-controlled stage. A wash bottle could also be fitted to the inlet N2 flow to allow
for controlled changes in relative humidity of the chamber.
Prior to TGA/DSC and CRM analysis, 10 mg (±0.5 mg) of gypsum was placed in 30 ml
alumina crucibles (Mettler Toledo material no.: 51140843). The same 30 µl alumina cru-
cibles used in the TGA were housed inside the chamber of the temperature-controlled stage
of the CRM.
The addition of a lid to the crucible that houses the gypsum in all experiments is known
to increase the partial pressure of water exerted on the sample (e.g. Lou et al., 2011). This
is a result of the restriction to the removal of water from the boundary layer around the
gypsum as it is expelled from the structure upon dehydration, thus causing an increase in
pressure in the local environment. An alumina lid (Mettler Toledo material no.: 51140477)
that has a hole of diameter ~1 mm was utilised in ‘lid on’ experiments, whereas no lid was
applied in ‘lid off’ experiments (Table 2.4).
52 Calcium Sulfate Systematics
When the lids of crucibles were on in CRM experiments, the laser was focused onto the
gypsum through the hole in the lid. As the crucible that houses the gypsum is heated from
its base in the temperature-controlled stage and the Raman laser is focused onto the top of
the sample, there is minor lag between the reference temperature and the real temperature
of the sample when measured. Modelling of this transient condition demonstrates that this
lag is on the order of 3°C min-1 at a heating rate of 20°C min-1, becoming negligible at a
heating rate of 1°C min-1. The reference temperature is reported without correction.
Fig. 2.13A shows the Raman spectra obtained from analysis gypsum, bassanite and sol-
uble γ-anhydrite. The 300 to 4000 cm-1 spectra can be split into sulfate (300–1300 cm-1) and
water (3400–3700 cm-1) peak regions. The flexion vibrations have been further subdivided
into ν1, ν2, ν3, and ν4 (SO4) and an H2O region (Fig. 2.13C) (e.g Prieto-Taboada et al.,
2014).
Black body radiation causes growth of the baseline at elevated temperatures, especially
to the higher Raman shift frequencies. This results in partial masking of the spectra in the
H2O region. This effect was increased during slower heating profiles, as well as when the
lid of the crucible was added. Therefore, the Raman band of the ν1 (SO4) region was se-
lected for analysis. The majority of other studies employ deconvolution analysis to separate
the relative contribution of the individual peaks in the 1008 cm-1 to 1025 cm-1 range (e.g.
Prieto-Taboada et al., 2014). Here, a high resolving grid was employed between 900 cm-1
and 1200 cm-1 so that the individual peaks representing gypsum, bassanite, and soluble γ-
anhydrite at 1008 cm-1, 1015 cm-1, and 1025 cm-1, respectively, can be fully separated (Fig.
2.13B).
2.5.2.2 Results
Transformation experiments monitored by TGA/DSC – Non-isothermal analysis
TGA determines the mass loss of a sample over the course of thermal dehydration, and is
represented as a curve displaying percentage weight loss over time (e.g. Fig. 2.14). DSC
is a measurement of the difference in heat flow between the sample and a reference, thus
providing a measure of the heat absorbed or released during phase transitions. The resultant
output of DSC experiments is a curve of heat flux over time, where exothermic reactions are
shown by a divergence of the curve below the baseline. Experiments are compared when the
lid of the crucible that houses the gypsum is omitted (Fig. 2.14A) and applied (Fig. 2.14B).
Fig. 2.14A shows the TGA and DSC curves of the dehydration of 100 µm gypsum at
heating ramps of 1 min-1, 5 min-1, 10 min-1 and 20°C min-1 when the lid of the crucible
that houses the sample is omitted (experiments 1-4; Table 2.4). A comparison of the TGA
2.5 Gypsum-water isotopic exchange 53
600 1200 1800 2400 3000 3600
900 1000 1100 1200
Raman Shift (cm-1)
Raman Shift (cm-1)
R
am
an
 I
nt
en
si
ty
 
(a
rb
itr
ar
y 
un
its
)
A
C
1025 - CaSO4
1015 - CaSO4•0.5H2O
1008 - CaSO4•2H2O
1008
1015 1025
3553
3401 3491
B
Compound
Gypsum
Bassanite
γ-Anhydrite 
β-Anhydrite 
α-Anhydrite 
415
428
420
417
417 
495
489
490
499
497
609
609 
620
627
630
628
628 
673
669
673
675
676 
1008
1015
1025
1017
1017 
1111
1108 
1134
1128
1128
1127 
1167
1160
1158 
3401 3491
2553
v2(SO4) v4(SO4) v1(SO4) v3(SO4) (H2O)
Fig. 2.13. A: Raman spectra (0 cm-1 to 4200 cm-1) identified during the analysis of gypsum, bassanite, and
soluble γ-anhydrite. B: Raman spectrum of the v1(SO4) stretching region (900 cm-1 to 1200 cm-1)
produced using a high resolving grid. C: Suggested Raman Spectra in the CaSO4–H2O System
(Prieto-Taboada et al., 2014).
54 Calcium Sulfate Systematics
2202102001901801701601501401301201101009080
2202102001901801701601501401301201101009080
0
5
10
15
20
40
20
0
60
40
20
0
60
T
G
A
 - %
 w
t lo
ss
D
S
C
 -
 m
W
0
5
10
15
20
20 K min-1
10 K min-1
5 K min-1
1 K min-1
20 K min-1
10 K min-1
5 K min-1
1 K min-1
B
A
Temperature (°C)
Temperature (°C)
D
S
C
 -
 m
W
T
G
A
 - %
 w
t lo
ss
Fig. 2.14. Thermogravimetric analysis (TGA; dashed lines) and differential scanning calorimetry (DSC;
solid lines) profiles of 100 µm gypsum at heating rates of 1°C min-1, 5°C min-1, 10°C min-1 and
20°C min-1: A: Lid of the crucible is not applied. B: Lid of the crucible is applied. Black arrows
denote points on the TGA curve where the rate of mass loss changes.
2.5 Gypsum-water isotopic exchange 55
curves obtained for 100 µm gypsum at all heating rates shows that the onset of dehydration,
represented by the initial reduction in mass, occurred at lower temperatures when lower
heating rates are used and at higher temperatures when heating rates are higher. The rate of
mass loss appeared to gradually increase from the onset until the termination of dehydration.
Over all experiments, the total mass loss was found to be 20.9%, indicating all water had
been removed from the crystal structure. All DSC curves display a single peak.
Fig. 2.14B shows the TGA and DSC results of the dehydration of 100 µm gypsum
when the crucible lid is applied, but all other conditions remain constant (i.e. experiments
5-8; Table 2.4). ‘Lid on’ experiments display three observable differences to ‘lid off’
experiments: Firstly, DSC curves displayed two distinct peaks. Secondly, a reduction in the
gradient of the TGA curve was observed, indicating that the rate of mass loss was retarded
at ~75% of the total mass loss (black arrows; Fig. 2.14B). Thirdly, the termination of the
dehydration reaction in experiments 5 to 8 occurred at higher temperatures compared to
experiments 1 to 4 (Fig. 2.14A).
To compare the effect of grain size on the dehydration profiles, Fig. 2.15 displays 50 µm,
100 µm and 500 µm gypsum at a heating ramp of 10°C min-1 with both crucible lid off and
on (experiments 3 and 7, respectively). The DSC curves of 50 µm gypsum show two distinct
peaks, both when the crucible lid is on and off, and the TGA of both profiles show a step
at ~75% of the total mass loss. 500 µm gypsum shows a single DSC peak and continuous
weight loss in both experiments, however. Results from these analyses show that the onset
of dehydration occurs at lower temperatures for smaller grain sizes. Analysis of 50 µm,
100 µm and 500 µm gypsum at heating rates of 1°C min-1, 5°C min-1 and 20°C min-1 all
produce similar profiles to those observed at a heating rate of 10°C min-1 (i.e. Fig. 2.15),
but the onset of dehydration occurs at lower temperatures when heating rates are lower and
vice versa.
Transformation experiments monitored by RS – Non-isothermal analysis
To compare the effect of grain size on the dehydration profiles with TGA/DSC analysis,
all experiments (i.e. Table 2.4; experiments 1 to 8) were conducted on 50 µm, 100 µm
and 500 µm gypsum. Fig. 2.15 displays the temperatures over which the ν1 (SO4) modes
of 1008 cm-1, 1015 cm-1 and 1028 cm-1 were observed during the dehydration of 50 µm,
100 µm and 500 µm gypsum at ramp rate of 10°C min-1 with crucible lid off (experiment 3;
dashed line) and on (experiment 7; solid line). During the dehydration of 50 µm and 100 µm
gypsum, the ν1 (SO4) mode shifted from 1008 cm-1 to 1015 cm-1 and 1028 cm-1 over dif-
ferent timescales. Lid-on experiments showed a systematic increase in the time period
over which the 1015 cm-1 mode was observed. As observed in TGA/DSC experiments,
56 Calcium Sulfate Systematics
100 110 120 130 140 150 160 170 180 190 200 210
D
S
C
 -
 m
W
0
TG
A
 - %
 w
t loss
5
10
15
20
lid on
lid off
A
40
20
0
lid off
lid on
1008 cm-1 - Gypsum
1015cm-1 - Bassanite
1025cm-1 - Anhydrite
50µm  
100 110 120 130 140 150 160 170 180 190 200 210
D
S
C
 -
 m
W
0
TG
A
 - %
 w
t loss
5
10
15
20
lid on
lid off
B
40
20
0
lid off
lid on
100µm  
100 110 120 130 140 150 160 170 180 190 200 210
D
S
C
 -
 m
W
0
TG
A
 - %
 w
t loss
5
10
15
20
lid on
lid off
C
40
20
0
lid off
lid on
500µm  
Temperature (°C)
Fig. 2.15. Thermogravimetric (TGA) and differential scanning calorimetry (DSC) profiles of (A) 50 µm,
(B) 100 µm, and (C) 500 µm gypsum at a heating rates of 10 ˚C min-1. Horizontal bars indicate
presence of raman spectra at 1008 cm-1(green), 1015 cm-1 (blue), and 1025 cm-1(red), representing
gypsum, bassanite, and soluble γ-anhydrite, respectively.
2.5 Gypsum-water isotopic exchange 57
analysis of gypsum at heating rates of 1°C min-1, 5°C min-1 and 20°C min-1 all produce
similar profiles to those observed at a heating rate of 10°C min-1, where the transition to
each phase occurred at lower temperatures for lower heating rates and visa versa. Notably,
during the dehydration of 500 µm gypsum at all ramp rates and during both lid-on and lid-
off experiments, the ν1 (SO4) mode shifts directly from 1008 cm-1 to 1028 cm-1, omitting
the intermediate ν1 (SO4) mode at 1015 cm-1.
Transformation experiments monitored by TGA/DSC – Isothermal analysis
Experimental data obtained from isothermal analysis of 50 µm, 100 µm, 200 µm and 500 µm
gypsum at 60°C, 70°C, 80°C, 90°C, 100°C and 120°C displayed continuous weight loss and
one DSC peak for all grain sizes in both lid-on and lid-off experiments. The transformed
fractions (ε) were calculated as a function of the total mass loss:
ε =
Wi−Wt
Wi−Wf
where Wi = initial gypsum sample weight; Wt = sample weight after heating for time t; and
Wf = sample weight corresponding to complete water loss (Khalil, 1982). Representative
ε% values of the investigated gypsum materials at different isothermal temperatures are
shown in Fig. 2.16 as a function of time. The termination of dehydration reaction (i.e. ε = 1)
occurred after a shorted time period for both higher temperatures and in lid-off experiments.
Dehydration of 50 µm, 100 µm, 200 µm and 500 µm gypsum at 60°C and 70°C (both lid-on
and lid-off experiments) was incomplete after 3 days (i.e. ε did not reach 1).
Transformation experiments monitored by RS – Isothermal analysis
1008 cm-1, 1015 cm-1 and 1028 cm-1 modes were visible simultaneously during the dehyd-
ration of 50 µm and 100 µm gypsum, while only the 1008 cm-1 and 1028 cm-1 modes were
present during the dehydration of 500 µm gypsum. Dehydration of 50 µm and 100 µm
gypsum at 60°C only displayed the 1008 cm-1 modes after 3 days, while dehydration reac-
tions at 70°C displayed both 1008 cm-1and 1015 cm-1 modes after 3 days.
2.5.2.3 Discussion
Timing of gypsum dehydration:
An accurate understanding of the isotopic modification of GHW due to dehydration
and rehydration induced by exposure to elevated temperatures is needed for the accurate
58 Calcium Sulfate Systematics
interpretation of GHW signatures. Isothermal TGA/DSC experiments show that the onset of
the dehydration of gypsum (represented by the initial mass loss of a sample) occurs earlier
both at higher temperature and with decreasing grain size (Fig. 2.16). While the partial
dehydration of finer grain sizes (e.g. 50 µm and 100 µm gypsum) occurs at temperatures of
60°C, larger grain sizes (i.e. 500 µm gypsum) do not show weight loss over the time period
studied. Therefore, assuming a normal geothermal gradient, massive, crystalline gypsum
should be the stable phase at depths up to ~2 km.
The heterogeneity of dehydration temperature with grain size has not been discussed
in detail previously. It has been hypothesised that the differing phases produced during the
dehydration process require different reaction energies, accounting for both the separation of
water molecules from the crystal structure as well as evaporation of the released dehydration
water (e.g. van der Heijden et al., 2011). Changes to parameters such as the vapour pressure
and temperature will therefore influence the dehydration kinetics, changing the efficiency of
dehydration (e.g. McAdie, 1964; Badens et al., 1998; Ballirano and Melis, 2009; Lou et al.,
2011; Fusseis et al., 2012). During gypsum dehydration, the local environment is associated
with a net volume increase as water exhumed has a greater volume than the pore spaces that
develop. If water expulsion is restricted (e.g. due to crystalline size), the fluid pressure will
increase, slowing the dehydration reaction (Bedford et al., 2017). The increased stability
of larger grain sizes of gypsum may therefore be linked to the increased stabilisation of the
inner domain of the crystalline structure caused by the increased pressures during heating.
The timing and rate of dehydration of differing crystalline grain sizes of gypsum is therefore
likely controlled by a feedback of pressure release and pore formation on the grain scale.
Stability of the intermediate bassanite phase:
The presence of two peaks within a DSC profile during the dehydration of powdered
gypsum has previously been interpreted as the two-step dehydration process of gypsum
through the bassanite intermediate (CaSO4 ·2H2O −−→ CaSO4 ·0.5H2O+ 1.5H2O −−→
CaSO4 + 0.5H2O), while the 3:1 mass loss in the TGA curve is consistent with the stoi-
chiometry of this dehydration reaction (Paulik et al., 1992; Lou et al., 2011) (Fig. 2.15). A
single DSC peak and no change in the rate of mass loss is interpreted to represent a single
step transition (i.e. CaSO4 ·2H2O−−→ CaSO4+2H2O), as observed in isothermal experi-
ments below 100°C and non-isothermal experiments when the crucible lid was omitted (e.g.
van der Heijden et al., 2011; Lou et al., 2011). During the dehydration of 100 µm gypsum
at a heating ramp of 10°C min-1 (lid off), the CRM detected multiple phases of calcium
sulfate coexisting simultaneously at the same temperature (i.e. at 147°C; Fig. 2.15B).
2.5 Gypsum-water isotopic exchange 59
0 5 10 15 20 25 70
0 5 10 15 20 25 70
0 5 10 15 20 25 70
0
0.2
0.4
0.6
0.8
1.0
0 5 10 15 20 25 70
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
Time (hours)
Tr
an
sf
or
m
ed
 F
ra
ct
io
n 
(ε)
 
Tr
an
sf
or
m
ed
 F
ra
ct
io
n 
(ε)
 
Tr
an
sf
or
m
ed
 F
ra
ct
io
n 
(ε)
 
Tr
an
sf
or
m
ed
 F
ra
ct
io
n 
(ε)
 A
B
C
D
60°C
70°C
80°C
90°C
100°C
120°C
Fig. 2.16. Isothermal analysis of (A) 500 µm, (B) 200 µm, (C) 100 µm and (D) 50 µm gypsum over the
temperature range 60˚C to 120°C with the crucible lid applied. Transformed fractions (ε), a
function of the total mass, loss versus time curves for the dehydration (gypsum to γ-anhydrite).
Note broken x-axis between 25 h and 70 h.
60 Calcium Sulfate Systematics
Direct comparison with the DSC curve shows this temperature coincides with the crest of
the single DSC peak. The existence of intermediate phases can therefore be concealed in
the TGA/DSC analysis (Putnis et al., 1990). It can be concluded that the stability of the
bassanite intermediate is dependent on the rate of moisture release in the local environment,
which may be due to increased temperatures (increasing the rate of moisture expulsion),
and/or restricted removal of the expelled water due to the application of a lid.
During the dehydration of 500 µm gypsum at all temperatures (Fig. 2.15C), the ν1
(SO4) mode shifts directly from 1008 cm-1 to 1028 cm-1, while the 1015 cm-1 mode is
suppressed beyond detectible limits. This ν1 (SO4) mode shift (in both lid on and lid off
experiments) occurs at only ~5% total weight loss. As the Raman laser only samples the
surface of the crystal, the transition in the internal crystal structure is still in progress. The
timing of the Raman shift and the lack of 1015 cm-1 mode is consistent with the observation
of a sharp dehydration front that propagates inward from the margin of polycrystalline
samples in time-series synchrotron X-ray micro-tomography experiments (Fusseis et al.,
2012). Gypsum breakdown across the dehydration front of larger crystals is therefore very
efficient and most likely controlled by the orientation of the gypsum lattice with respect to
the advancing front (Fusseis et al., 2012). Dehydration of smaller grain sizes (< 200 µm),
however, produces a prolonged bassanite phase due to both the reduced stability of smal-
ler crystalline lattice and the increased efficiency of fluid drainage, lowering the internal
crystalline water pressure. The presence of the bassanite intermediate is therefore likely
controlled by a feedback of pressure release on the grain scale.
2.5.2.4 Conclusions
The initial temperature of dehydration increases with increasing grain size. Isothermal
temperature experiments (up to 3 days in duration) conducted on crystalline grain sizes
larger than 500 µm did not result in weight loss at temperatures below 65°C. In contrast,
powdered gypsum with crystalline grain sizes less than 200 µm experienced partial weight
loss between 60°C and 70°C, but did not fully dehydrate. Above 70°C, all gypsum fully
dehydrated over differing timescales, where increasing temperature resulted in an increase
in the rate of dehydration.
RS demonstrated that gypsum with crystalline grain sizes less than 200 µm dehydrated
to γ-CaSO4 via an observable bassanite intermediate during all isothermal temperatures
and ramped temperatures experiments. The observed duration of bassanite was increased
at both smaller grain sizes and increased partial pressure. At grain sizes larger than 500
µm, bassanite was undetectable by RS and TGA/DSC during the dehydration process.
Overall, results suggest that stable isotopes of GHW can be used to reconstruct the isotopic
2.5 Gypsum-water isotopic exchange 61
composition of palaeo-waters in environments where large (>500 µm) crystals are present,
burial/exhumation of the sample has not exceeded ~2 km depth and recrystallisation has not
occurred.
2.5.3 Gypsum recrystallisation and isotope exchange
Post-depositional isotope exchange of GHW may occur if gypsum is in contact with water
that is isotopically different from its hydration water. The isotopic exchange reactions for
oxygen and hydrogen isotopes between gypsum hydration water and pore water are:
CaSO4 ·2H216O+H218O−−→ CaSO4 ·2H218O+H216O
CaSO4 ·2H216O+HD18O−−→ CaSO4 ·2DH18O+H216O
The kinetics governing the rate of exchange via these reactions are not well understood,
but are likely mediated by factors such as temperature, pressure, humidity, crystal size,
and disequilibrium time (Hodell et al., 2012). Equally, the rate of oxygen versus hydrogen
isotope exchange is likely to differ due to the differing rates of diffusion of the H218O and
DH16O isotopologues into and out of the crystal structure. Thus, if isotope exchange were
to occur, the ∆δD/Δδ18O relationship would deviate from its original value (Hodell et al.,
2012).
Long-term isotope exchange experiments were set up in the Godwin Laboratory, Cam-
bridge, to investigate the rate of GHW exchange (Fig. 2.17). Gypsum from the Cabezo de
Maria, southeastern (SE) Spain, of 1 mm-2 mm grain size was placed in CaSO4-saturated
waters at 45°C with significantly differing isotopic composition (enriched: δ18O = 21.6h,
δD = 58h; depleted: δ18O = -19.5h, δD = -151h) to that of the original GHW (δ18O =
-0.89h, δD = -55.08h). After two years, no exchange was found to occur. Results suggest
that gypsum of mm-sized grain size will retain the isotopic composition of the water from
which they form in most natural settings.
The isotopic composition of GHW can also be altered by dissolution in under-saturated
waters and secondary precipitation of gypsum in sediment pore waters, however (Hodell et
al., 2012; Pierre, 2018). Gypsum may undergo diagenetic transformations due to gypsum’s
relatively high solubility in water (2.0 g l-1 to 2.5 g l-1) and halokinetic properties, making
deposits somewhat mobile, both physically and chemically, in sedimentary environments
(Orti, 2011; Ba¸bel and Schreiber, 2014). If dissolution and secondary precipitation were to
occur, the hydration water of the recrystallised gypsum will have an isotopic composition
in equilibrium with sediment pore waters (Hodell et al., 2012). The isotopic signature of
62 Calcium Sulfate Systematics
δ1
8 O
 (
‰
) δD
 (‰
)
17O
-excess (per m
eg)
δ1
7 O
 (
‰
)
δ1
8 O
 (
‰
) δD
 (‰
)
δ1
7 O
 (
‰
)
17O
-excess (per m
eg)
Time (weeks)
Time (weeks) Time (weeks)
Time (weeks)
A
C D
B
-56.0
-55.5
-55.0
-54.5
-54.0
-53.5
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0 20 40 60 80 100 120
0
5
10
15
20
25
30
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0 20 40 60 80 100 120
-56.0
-55.5
-55.0
-54.5
-54.0
-53.5
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0 20 40 60 80 100 120
0
5
10
15
20
25
30
35
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0 20 40 60 80 100 120
Fig. 2.17. Gypsum exchange experiments. 1 mm to 2 mm gypsum crystals (from the Cabezo de Maria,
SE Spain) are placed in CaSO4-saturated waters at 45°C. Initial GHW composition is -0.89h, -
55.08h and 22 per meg for δ18O, δD and 17O-excess, respectively. Enriched water (A and B) has
the isotopic composition δ18O = 21.6h, δD = 58h; depleted water (C and D) has the isotopic
composition δ18O = -19.5h, δD = -151h. No isotopic exchange between the gypsum hydration
water and the solution water is seen after >2 years.
GHW after recrystallisation is easily recognisable, and multiple samples will display no
evaporative slope (Sofer, 1978). Equally, extensive recrystallisation results in recognisable
changes to crystalline morphology, where primary features, i.e. clear, euhedral crystals
with swallow-tail twinning, are replaced by gypsum entombing sediments and secondary
diagenetic features (e.g. Hodell et al., 2012; Warren, 2016).
GHW is also believed to be free of ‘vital effects’ that influence standard δ18O meas-
urements of carbonates (e.g. Herwartz et al., 2017). In carbonates, vital effects generally
result in lighter isotope compositions than expected for equilibrium fractionation (McCon-
naughey, 1989) due to the rate limiting step inducing the kinetic effect is conversion of
CO2 −−→ HCO−3 and slow re-equilibration with water, especially at elevated pH (Herwartz
et al., 2017). Thus a similar ‘vital effect’ is not expected for GHW (Herwartz et al., 2017).
Additionally, biological effects have not been observed to impact GHW in environments
where gypsum precipitation was likely mediated by microorganisms, such as the moder-
ately saline (<100 g l-1) ponds of Atacama Desert, Chile, where gypsum displayed variable
2.5 Gypsum-water isotopic exchange 63
coloration (black, white, brown and greenish) and a friable structures (Thompson and Ferris,
1990; Vogel et al., 2009; Vogel et al., 2010; Farías et al., 2014; Herwartz et al., 2017).
Overall, the triple oxygen and hydrogen isotopic composition of GHW may serve as a
useful palaeoenvironmental proxy, given that the isotopic composition of GHW is stable
over a wide range of natural conditions. Analysis of GHW permits the actual isotopic
composition of palaeo-waters to be determined with little to no effect of temperature, and
thus GHW may be used as a tool for quantitative palaeohydrological reconstructions.

Chapter 3
Quantification of drought during the
collapse of the classic Maya civilisation
Adapted from the manuscript: Nicholas P. Evans, Thomas K. Bauska, Fernando Gázquez,
Mark Brenner, Jason H. Curtis, and David A. Hodell. Science 361, 498-501 (2018).
The demise of Lowland Classic Maya civilisation during the Terminal Classic Period
(TCP, ~800-1000 CE) is a well-cited example of how past climate may have impacted
ancient societies. Attempts to estimate the magnitude of hydrologic change, however, have
met with equivocal success because of the qualitative and indirect nature of available climate
proxy data. The past isotopic composition (δ18O, δD, 17O-excess and d-excess) of water in
Lake Chichancanab, Mexico, was reconstructed by analysing the structurally-bound water
in sedimentary gypsum, which was deposited under peak drought conditions. The triple
oxygen and hydrogen isotope data provide a direct measure of past changes in lake hydro-
logy. We modelled the data and conclude that annual precipitation decreased between 41%
and 54% (with intervals of up to 70% rainfall reduction during peak drought conditions) and
that relative humidity declined by between 2% and 7% compared to present-day conditions.
Our modelling results provide robust, quantitative estimates of the magnitude of hydrologic
change during the TCP that can be used to assess how drought affected agriculture in the
northern Maya Lowlands.
D.A.H., N.P.E., and F.G. developed the analytical method and designed the study; M.B.,
J.H.C., and D.A.H. collected the original sediment cores from Lake Chichancanab; N.P.E.
sampled the cores; N.P.E. and F.G. performed all isotopic analyses; T.K.B. designed the
66 Quantification of drought during the collapse of the classic Maya civilisation
transient model and performed drought simulations; and N.P.E. wrote the paper with contri-
butions from all other authors.
3.1 Introduction
More than two decades ago, a sediment core from Lake Chichancanab (Yucatán Peninsula,
Mexico; Fig. 3.1) provided the first physical evidence of a temporal correlation between
drought and the sociopolitical transformation of the Classic Maya civilisation during the
TCP (Hodell et al., 1995). Presence of gypsum horizons and a concomitant increase in the
oxygen isotope ratio (18O/16O) in shells of ostracods and gastropods suggested the TCP
was among the driest periods of the Holocene in northern Yucatán. Paleoclimate records
produced subsequently provided additional evidence for drought during the TCP (Curtis
et al., 1996; Hodell et al., 2001; Haug, 2003; Medina-Elizalde et al., 2010; Kennett et al.,
2012; Wahl et al., 2013; Douglas et al., 2015), but the magnitude of hydro-climate change
and its influence on Maya agricultural and sociopolitical systems remains controversial
(Douglas et al., 2016a). The qualitative nature of most climate proxy archives, combined
with dating uncertainties, has prevented detailed assessment of the relationship between past
climate and cultural changes (Yaeger and Hodell, 2009; Douglas et al., 2016a,b).
Recent attempts to quantify estimates of past changes in rainfall amount and assess the
impact on ancient Maya agriculture have utilised isotopes of either oxygen (δ18O) (Medina-
Elizalde et al., 2010; Medina-Elizalde and Rohling, 2012) or hydrogen (δD) (Douglas
et al., 2014, 2015, 2016a). No study to date has combined the two isotope systems be-
cause materials used for analysis (i.e. carbonates and leaf waxes) preclude simultaneous
measurement of the multiple isotopologues of water. Combined analysis of δ18O, δ17O
and δD is a powerful method to estimate past hydrologic changes quantitatively because
hydrogen and triple oxygen isotopes each undergo slightly different fractionation during
evaporation, leading to changes in the derived d-excess and 17O-excess parameters (see
Chapter 1). In an effectively closed hydrological basin such as Lake Chichancanab, the
primary controls on the isotopic fractionation of lake water during evaporation include: the
fractional loss of precipitation to evaporation (P/E), normalised relative humidity (RHn),
temperature, and changes in the precipitation source (Hodell et al., 1995, 2001, 2005a;
Gázquez et al., 2018). The d-excess is largely dependent on RHn and temperature, whereas
17O-excess is controlled mainly by RHn (Surma et al., 2015; Herwartz et al., 2017; Gázquez
et al., 2018). Because the predicted trends of d-excess and 17O-excess in evaporating waters
display different responses to climate variables, they can be evaluated individually using an
iterative model (see Section 3.4).
3.1 Introduction 67
1000 200
km
21°N
20°N
19°N
18°N
17°N
16°N
84°W85°W86°W87°W88°W89°W90°W91°W92°W93°W94°W95°W96°W97°W
A A
C
B
B
C
D E
F
A
B
C
A
B
C
Caves: Lakes:
Punta Laguna 
Aguada X’caamal
Lake Chichancanab 
Lake Salpetén 
Lake Puerto Arturo 
Laguna Yaloch
Tzabnah Cave
Macal Chasm 
Yok Balum Cave
D
E
F
Fig. 3.1. Map of the Maya Lowlands displaying the locations of proxy climate archives (north to south); the
Chaac speleothem of Tzabnah Cave (Medina-Elizalde et al., 2010); Punta Laguna (Curtis et al.,
1996); Aguada X’caamal (Hodell et al., 2005b); Lake Chichancanab (this study) (Hodell et al.,
2005a, 2001, 1995; Douglas et al., 2014); Lake Puerto Arturo (Wahl et al., 2014); Laguna Yaloch
(Wahl et al., 2013); Macal Chasm (Webster et al., 2007); Lake Salpetén (Douglas et al., 2015); the
Yok I speleothem of Yok Balum Cave (Kennett et al., 2012).
We took advantage of the benefits of using all isotopologues of water and their derived
parameters (d-excess and 17O-excess) by measuring triple oxygen and hydrogen isotopes in
the hydration water of gypsum (CaSO4 ·2H2O) in sediment cores from Lake Chichancanab
(Hodell et al., 2005). Today, the lake water is near saturation for gypsum and during past
periods of drier climate, when the lake volume shrank, gypsum precipitated from the lake
water and was preserved as distinct layers within the accumulating sediments (Hodell et al.,
1995, 2001, 2005a). When gypsum forms, water molecules are incorporated directly into
its crystalline structure and this GHW records the isotopic composition of the parent fluid,
with known isotopic fractionations (Sofer, 1978; Hodell et al., 2012; Herwartz et al., 2017;
Gázquez et al., 2017a,b, 2018). Unlike oxygen isotope fractionation during formation of
carbonate minerals (Lachniet, 2009; Kluge and Affek, 2012), fractionation during gypsum
crystallisation is practically independent of temperature (Chapter 2), biological or kinetic
68 Quantification of drought during the collapse of the classic Maya civilisation
(non-equilibrium) effects (Herwartz et al., 2017). Additionally, isotopes of GHW that are
measured in the sedimented gypsum inherently record the driest periods, offering a distinct
advantage over other traditional climate archives such as speleothems or mollusk shells,
which may fail to register peak drought conditions because of growth hiatuses. Absolute
differences in the δ18O, δD, 17O-excess and d-excess, between modern and paleo-lake
water, provide an estimate of differences between the lake hydrologic budget during the
TCP and today (Fig. 3.2). Results were evaluated using a numerical isotope mass balance
model that must satisfy all isotope variables, and thus provide a more robust constraint on
past hydrology than does modelling δ18O or δD alone.
3.2 Materials and methods
3.2.1 Sediment cores and sampling
Lake Chichancanab is located at ~19°51’21”N 88°45’49”W, Yucatán Peninsula, SE Mexico
(Fig. 3.1). In March 2004, core CH1 7-III-04 was collected from Lake Chichancanab
with a piston corer in 14.7 m of water (Hodell et al., 2005). Shortly after collection, cores
were split, wrapped in plastic film and stored at 4°C. Core sections were measured for
bulk density by gamma-ray attenuation (GRA) by Hodell et al. (2005), and contain high-
density gypsum bands interbedded with low-density organic layers (Fig. 3.3). Gypsum was
sampled at 0.5 cm intervals – individual crystals were picked from the >350 µm size fraction
and ground to a powder. Powdered samples were dried in an oven at 45°C for 15 h to 20 h
and placed under vacuum (~10-3 mbar) for ~3 h to remove adsorbed water prior to hydration
water extraction (Playà et al., 2005).
3.2.2 Gypsum Hydration Water (GHW)
GHW was extracted from each sample (150-200 mg) by heating to 400°C and trapping the
evolved water, in vacuo, using a bespoke offline extraction system described in Chapter 2.
1σ was ±0.07h for δ17O, ±0.12h for δ18O, ±0.63h for δD, ±0.8h for d-excess and ±6
per meg (±0.006h) for 17O-excess.
We suggest the gypsum in Lake Chichancanb preserves the isotopic composition of the
lake water at the time of deposition and has not undergone post-depositional diagenesis or
exchange with modern lake water. After applying fractionation factors, the isotopic values
of the palaeo-lake waters are considerably enriched compared to the modern lake (Fig. 3.2;
Fig. 3.3). If the GHW had exchanged with sediment pore water, a relatively homogeneous
3.2 Materials and methods 69
S
ed
im
en
t d
en
si
ty
 
(g
/c
m
3 )
 CLASSIC TC POSTCLASSIC CONTACTPRECLASSIC
MODERN LAKE WATER RANGE
POSTCLASSIC CLASSIC TERMINAL CLASSIC
Age (C.E.)
Age (C.E.)
Drier
MODERN LAKE WATER RANGE
MODERN LAKE WATER RANGE
MODERN LAKE WATER RANGE
A
ge
 U
nc
er
ta
in
ty
(Y
ea
rs
)
0 400 800 1200 1400 2000
1.0
1.5
-12
-16
-20
10
15
20
1.5
2.0
2.5
3.0
4.0
5.0
600 700 800 900 1000 1100
600 700 800 900 1000 1100
+200
-200
d-
ex
ce
ss
 
(‰
 V
S
M
O
W
)
δ17
O
(‰
 V
S
M
O
W
)
δD
(‰
 V
S
M
O
W
)
δ18
O
(‰
 V
S
M
O
W
)
Fig. 3.2. Water isotopes during drought periods compared to modern water isotopes of Lake Chichancanab.
(Lower) Sediment density record of core CH1 7-III-04 from 0 CE to 2000 CE (shown relative to
Maya chronology) (Hodell et al., 2005). Periods of gypsum precipitation are indicated by density
values >1.1 g cm-3. Age uncertainty (95% confidence intervals) are derived from Bayesian age-
depth analysis and normalised to the best-fit age model (Fig. 3.11). (Upper) δ18O, δ17O, δD and
d-excess (δD–8 ·δ18O) of palaeo-lake water data (yellow circles) from 550 to 1150 CE shown after
correction of measured GHW for known fractionation factors at 26°C. Horizontal blue band defines
the mean (±1σ ) isotopic composition recorded in the modern lake. Positive δ18O, δ17O and δD
values and negative d-excess values reflect periods of drought. Note d-excess axis is reversed.
70 Quantification of drought during the collapse of the classic Maya civilisation
1.
8
1.
6
1.
4
1.
2
1.
0
17
0
16
0
15
0
14
0
13
0
12
0
S
ed
im
en
t 
D
en
si
ty
 
(g
/
cm
3
) 
Core Depth (cm)
3.
1 
±
 0
.4
1.
6 
±
 0
.2
δ18
O
(‰
) 
12
.7
 ±
 1
.9
-2
8 
±
 7
-1
1.
8 
±
 1
.1
δ17
O
(‰
) 
δD (‰
) 
1
7
O
-e
xc
es
s
(p
er
 m
eg
) 
d
-e
xc
es
s
(‰
) 
A
g
e
lo
w
er
 b
ou
n
d
 
(±
 e
rr
or
)
4.
1 
±
 0
.3
4.
3 
±
 0
.3
4.
1 
±
 0
.4
2.
1 
±
 0
.2
2.
2 
±
 0
.2
2.
1 
±
 0
.2
16
.1
 ±
 1
.2
16
.9
 ±
 0
.9
16
.0
 ±
 1
.6
-3
6 
±
 5
-3
8 
±
 6
-2
9 
±
 1
2
-1
6.
5 
±
 1
.7
-1
7.
6 
±
 1
.7
-1
6.
5 
±
 1
.2
19
94
 ±
 0
 
93
9 
±
 1
00
72
0 
±
 1
00
 6
02
 ±
 7
1
61
1 
±
 7
2
84
9 
±
 1
06
10
92
 ±
 9
0
20
10
 ±
 0
A
g
e
u
p
p
er
 b
ou
n
d
 
(±
 e
rr
or
)
Fig. 3.3. Image of split core CH1 7-III-04 (Hodell et al., 2005). Sediments are composed of interbedded
gypsum- and organic-rich strata containing abundant shell material. Solid black line represents the
GRA bulk density record measured on core CH1 7-III-04 (Hodell et al., 2005). Periods of gypsum
precipitation are indicated by positive density excursions >1.2 g cm-3. The δ18O, δ17O, δD and
d-excess (h, VSMOW) and 17O-excess (per meg, VSMOW) (±1σ ) of the modern lake water and
measured GHW during each of the three periods of gypsum deposition (after correction for known
fractionation factors at 26°C) are displayed.
3.3 Results 71
isotopic profile, with values similar to the current lake water, would be expected. Further-
more, the burial depth of the gypsum is shallow (< 2 m) and the sediments are porous, thus
isotopic gradients in pore water would be strongly attenuated by diffusion and advection
with overlying lake water (Gázquez et al., 2018).
3.2.3 Precipitation, meteoric and lake water samples
The triple oxygen and hydrogen isotopic measurement of lake water (n = 156), river and
ground water (n = 92), and rainwater (n = 31) samples from stations across the Yucatán Pen-
insula are reported from a collection period of 1994 to 2010 (Fig. 3.4; Appendix C; Tables
C.5, C.6 and C.7). Measurements were made using an L2140-i Picarro CRDS water isotope
analyser. The majority of rainwater samples were collected at 20°00’59”N 89°01’13”W,
~30 km west of Lake Chichancanab (Hodell et al., 2012). All water samples were collected
and stored in Qorpak bottles with Polyseal cone-lined caps to prevent evaporation.
3.2.4 Biogenic carbonate measurements
Shells and shell fragments of the gastropod Pyrgophorus coronatus were picked from 0.5 cm
intervals of core CHI 7-III-04. Shells were cracked and sonicated in methanol to remove
contaminant debris. Subsequently, samples were treated with 10% H2O2 for 30 min to re-
move organic matter, dried and ground to a fine powder. Stable oxygen isotopes of carbonate
were measured using a ThermoScientific GasBench II, equipped with a CTC autosampler
coupled to a MAT253 mass spectrometer (Spötl and Vennemann, 2003). Samples were
flushed with CP grade helium then acidified with 104% H3PO4 and reacted at 70°C for 1
hour. Repeat analysis of the Carrara Marble standard yielded a 1σ analytical precision of
±0.1h for δ18O. Results are reported relative to the Vienna Pee Dee Belemnite (VPDB).
Small sample fragments were run on a Kiel III carbonate preparation device interfaced with
a Finnigan MAT 252 mass spectrometer. Analytical precision was estimated at ±0.08h for
δ18O.
3.3 Results
The modern regional climate around Lake Chichancanab is characterised by a mean annual
precipitation of ~1200 mm, a mean annual surface water temperature of ~26°C and a net
annual water deficit of 300–400 mm yr-1 (Hodell et al., 2005). Large changes in precipitation
and RHn occur between the dry (November to May) and rainy seasons (June to October)
(Wassenaar et al., 2009; Medina-Elizalde and Rohling, 2012). Measured δ18O and δD of
72 Quantification of drought during the collapse of the classic Maya civilisation
Local Meteoric Water Line
y = 7.70x + 9.0
R² = 0.95
Yucatán Peninsula Lake Water
Evaporative Line
y = 6.03x - 5.1
R² = 0.94
-60.0 
-50.0 
-40.0 
-30.0 
-20.0 
-10.0 
0.0
10.0
20.0
30.0
40.0
-10.0 -8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 10.0
δD
(‰
VS
M
OW
)
δ18O
(‰ VSMOW)
Fig. 3.4. δ18O and δD of rain, surface water, and ground water (blue squares). Solid blue line represents
the local meteoric water line estimated by least-squares linear regression through rain, surface and
groundwater data. Red squares indicate measurements made on lakes with varying hydrologic
budgets and evaporative losses across the Yucatán Peninsula. Solid red line represents the
evaporative line (with slope = 5.8) estimated by least-squares linear regression using lake data.
Note that this slope is comparable to the slope of modern data from Lake Chichancanab alone
(slope = 5.1), and is significantly different from that of the slope of measured GHW data (slope =
3.1) (Fig. 3.5).
3.4 Modelling 73
precipitation and groundwater samples from the Yucatán Peninsula, collected from 1994 to
2010, define a local meteoric water line (LMWL) with slope 7.7 (Fig. 3.4). Evaporation
enriches the lake in the heavier isotopes of oxygen and hydrogen in water (2.6h < δ18O <
3.8h and 10.1h < δD < 17.2h), evolving along an evaporative line defined by δD = 5.1
· δ18O – 3.1. This evaporation line intersects the LMWL at δ18O = -4.7(±1.2)h and δD
= -27.5(±10.7)h, which is within error of the mean oxygen and hydrogen isotope values
recorded in local rivers and groundwater from the International Atomic Energy Agency’s
(IAEA) regional Global Network of Isotopes in Precipitation stations (δ18O = -4.1h; δD =
-24.3h) (Wassenaar et al., 2009) and this study (δ18O = -4.0h; δD = -23.5h).
The gypsum deposited during the droughts of the Terminal Classic and early Postclassic
Periods (~720 to 1100 CE) was used to calculate the δ18O, δ17O and δD of the palaeo-lake
water, which ranged from 3.6h to 4.9h for δ18O, 1.9h to 2.5h for δ17O, and 13.7h
to 18.8h for δD (Fig. 3.5). Mean values of the palaeo-lake waters (δ18O = 4.2h; δ17O
= 2.2h; δD = 16.4h) during drought episodes are significantly greater than modern lake
values (δ18O = 3.1h; δ17O = 1.6 h; δD = 12.7h). Age uncertainty associated with the
lake record and periods of gypsum precipitation was calculated using Bayesian age-depth
analysis of radiocarbon ages obtained from the sediment cores (see Section 3.4.4). High
drought probabilities are found specifically during the onset (~750 and ~850 CE) and the
end (~950 to ~1050 CE) of the TCP (P > 0.85 and P > 0.95, respectively). Multiple proxy
climate records across the Maya Lowlands also provide evidence of drought synchronicity,
with only slight temporal variations across the region (Douglas et al., 2016a).
3.4 Modelling
3.4.1 Transient model details
To estimate quantitatively the magnitude of drought during the TCP, we employed a transi-
ent model that explicitly simulates the evolution of the isotopic and chemical composition
of the lake water, including the gypsum flux to the lake sediments (Fig. 3.6).
The transient box model presented here is a lake basin with a surface and deep box (Fig.
3.7A). The volume (4.57×107 m3), surface area (1.02×107 m2), and surface area to depth
relationship conform to data presented in Hodell et al. (2005) (Fig. 3.7B). This geometry
equates to a mean depth of 4.5 m in the model with the modern deepest area of the basin
equal to ~15 m.
The basin is fed by a precipitation flux determined by North American Regional Reana-
lysis (NARR) data (~1 m water equivalent/m2/a; Fig. 3.8) with a constant catchment size
74 Quantification of drought during the collapse of the classic Maya civilisation
LM
WL
δD = 7.7 · δ18O + 8.7
R2 = 0.96
d-excess = -0.3 · δ18O + 8.7
R2 = 0.04
Modern Lake Water
Modern Fresh Water
Paleo-lake Water
LMWL
8
12
16
2 3 4 5
-20
-10
0
2 3 4 5
-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0
-50
-40
-30
-20
-10
0
10
20
30
-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0
-50
-40
-30
-20
-10
0
10
20
30
-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0
-60
-40
-20
0
20
40
20
δD = 5.1 · δ18O - 3.1
R2 = 0.95
δD = 3.1 · δ18O + 3.4
R2 = 0.81
d-excess = -4.9 · δ18O + 3.4
R2 = 0.91
d-excess = -2.9 · δ18O - 3.0
R2 = 0.83
P/EMW
Wind
Atm. 
vapor
T
RH
P/E
MW
Wind
Atm. 
vapor
T
RH
LMWL
δD
 
(‰
 V
S
M
O
W
)
17
O
-e
xc
es
s 
(p
er
 m
eg
)
d-
ex
ce
ss
 
(‰
 V
S
M
O
W
)
δ18O 
δ18O 
δD
 
A
C
B
δ18O (‰ VSMOW)
δ18O (‰ VSMOW)
δ18O (‰ VSMOW)
3.4 Modelling 75
Fig. 3.5. Left: (A) Comparison of measured local meteoric water, modern lake water, and palaeo–lake
water data. Paleo–lake water data (yellow circles) are shown after correction of measured GHW
for known fractionation factors at 26°C. (A) δ18O versus δD space. (B) δ18O vs d-excess space.
(C) δ18O versus 17O-excess space. Local meteoric water measurements (grey circles) define the
local meteoric water line (LMWL). Relative to modern lake waters (blue circles), paleo–lake water
displays greater δ18O and δD values along an evaporative trend. The grey ellipses define the
relative influence of variables that can affect the isotopic composition of water in δ18O vs d-excess
and δ18O vs 17O-excess; the effects of precipitation/evaporation (P/E), normalised relative humidity
(RHn), temperature (T), changes to source composition (MW), the degree of equilibrium between
atmospheric vapour and fresh water (Atm. vapour), and turbulence created by wind are indicated.
The size of each arrow is derived from the tolerance given for each input parameter in Table 3.2.
through time. The basin hydrology is effectively closed (Perry et al., 2002). Groundwater
flux to and from the lake (equivalent to ~10% of precipitation) delivers Ca2+, SO42−, Na2+,
K+, Mg2+ and Cl− and maintains the lake near the modern salt balance, with gypsum near
saturation. Water is lost via evaporation, which is held constant at a rate of 1.07 m water
equivalent/m2/a with the absolute flux evolving through time, depending on the surface area
of the lake.
The saturation state of gypsum was calculated offline with the PHREEQC model for a
wide range of solutions, starting at modern conditions. The lake model incorporates these
solutions in a look-up table to calculate the gypsum saturation at every time step. If the
lake exceeds saturation, Ca2+ and SO42− are removed by gypsum precipitation at a rate
that maintains the lake below saturation, with a response time of 1 year. Combining the
mass of gypsum precipitated and surface area of the lake at a given time-step allows us to
calculate the gypsum accumulation in the lake sediment. This is then used to calculate a
synthetic core log of density, assuming the density of accumulating gypsum is 2.31 g cm-3
and that the accumulation of other sediments is constant at 0.941 mm yr-1, with a density
of 1.06 g cm-3 based on the mean sediment accumulation that brackets the drought periods
(Fig. 3.9).
3.4.2 Transient model isotope and ionic mass balance
The isotope mass balance model employs a Craig-Gordon evaporation scheme as formulated
in Criss (1999). The overall isotopic fractionation during evaporation (αevp) depends on the
equilibrium fractionation (αeq), kinetic fractionation (αkin), relative humidity (h) and the
isotopic ratio of water vapour (Rv) and the evaporating surface of the basin (Rb), whereby:
αevp = αeqαkin
(
1−h
1−αeqhRvRb
)
(3.1)
76 Quantification of drought during the collapse of the classic Maya civilisation
NARR Relative 
Humidity
NARR 
Precipitation
Transient Model Solutions (n = 300)
Select time periods where any gypsum 
precipitates (density > 1.1 g/cm3) 
Means of all model 
variables
PDFs of all model 
variables
Select only those scenarios 
where the means of the 
modeled isotopes are within 1σ 
of the mean isotope data and 
density > 1.2 g/cm3 
Time-series
PDFs of all the model 
parameters throughout the 
droughts
Means and ranges of the 
combined drought periods
1.  Reanalysis data input
2.  Create synthetic time-
series of climate forcing 
with intensity of droughts 
modulated with density 
data
3.  Selection of gypsum 
layers
4.  Aggregating model output 
in average drought 
conditions
5.  Selection of scenarios 
consistent with data
6.  Output
	
Fig. 3.6. Transient model summary of forcing (blue boxes and arrows), scenario selection by model/data
comparison (green boxes and arrows), and model output (red boxes and arrows).
3.4 Modelling 77
Groundwater Flux
Precipitation Water Vapour
Precipitation Flux Evaporation Flux
Mixing
Gypsum Precipitation
Lake Level
Basin 1
Basin 2
A
B
Fig. 3.7. (A) Diagrammatic representation of the two-box model used in transient model scenarios. (B) The
bathymetry of Lake Chichancanab used in the transient model (Hodell et al., 2005), and generated
normalised surface area and volume vs normalised depth.
78 Quantification of drought during the collapse of the classic Maya civilisation
αeq for H218O (α 18Oeq) and DHO (αDeq) are calculated using an assumed lake surface
temperature of 25°C and the equations of Horita and Wesolowski (1994). The equilibrium
fractionation for H217O is then a function α 18Oeq, where:
α 17Oeq = α 18O
θeq
eq
and the equilibrium fractionation parameter between 18O and 17O (θeq) is 0.529 (Barkan
and Luz, 2005).
Kinetic fractionation factors for all three minor isotopes are calculated as a function
of wind-induced turbulence (w), lake surface temperature (T ) and the kinetic fractionation
parameter between 18O and 17O (θkin):
α 18Okin = 1.0283w
αDkin = (1.25−0.02T )(α 18Okin−1)+1
α 17Okin = α 18O
θkin
kin
We assume Rv depends on the degree to which the atmospheric water vapour (Veq) is in
equilibrium (Gibson et al., 2016) with isotopic ratio of precipitation (Rp) where:
Rv = Rp
(
αeq(p−v)Veq
)
The general form of the mass-balance equation for the mass of a given species (mX) of
either isotope or ion in basin box 1 (b1) is:
dmXb1
dt
= FpRp−FevpRevp+Fgb1Rg−Fb1gRb1+Fb2b1Rb2−Fb1b2Rb1
Similarly the mass balance for basin box 2 (b2) is:
dmXb2
dt
= Fb1b2Rb1−Fb2b1Rb2
In this set of equations F represents the various fluxes of water in terms of mass of the
major water isotope (1H216O) and R represents the ratio of minor isotope or ionic species
relative to 1H216O. As examples, the ratio of deuterium in box 1 is:
RDHO,b1 =
mDHOb1
m1H216Ob1
3.4 Modelling 79
16
12
8
4c
ou
nt
 (n
)
-40 -35 -30 -25 -20
17O-excess (VSMOW)
6
4
2
0
count (n)
20
15
10
5
co
un
t (
n)
-16 -14 -12 -10
D-excess (VSMOW)
8
6
4
2
0
count (n)
20
15
10
5c
ou
nt
 (n
)
16141210
δD (VSMOW)
6
4
2
0
count (n)
20
15
10
5
co
un
t (
n)
4.03.53.02.5
δ
18O  (VSMOW)
6
4
2
0
count (n)
0.4
0.3
0.2
0.1
0.0
Pr
ec
ipi
ta
tio
n 
(m
/m
on
th
)
2008200420001996
Model Year (C.E.)
0.90
0.85
0.80
0.75
0.70
0.65
0.60
RH
 (%
)
4.0
3.5
3.0
2.5
δ1
8 O
  (
VS
M
OW
)
16
14
12
10
δD
 (V
SM
OW
)
-16
-14
-12
-10
D-
ex
ce
ss
 (V
SM
OW
)
-36
-34
-32
-30
-28
-26
17
O-
ex
ce
ss
 (V
SM
OW
)
Model Climate Forcing (NARR Reanalysis)
Model Water Isotope Solutions 
Model Solutions
Modern Data
Fig. 3.8. Transient model climate forcing. The transient model is forced with North American Regional
Reanalysis (NARR) data for local precipitation and relative humidity (RHn) across the modern
sampling period from 1994 to 2010 (black lines). Modelled water isotope solutions (red lines)
are then compared to measured modern data (blue lines) as histograms (right panel). The model
successfully reproduced the mean of modern isotope data, with insignificant gypsum precipitation.
The time interval from 1994 to 2010 is subsequently used as the baseline for comparison to palaeo-
simulations.
80 Quantification of drought during the collapse of the classic Maya civilisation
The concentration of Ca in box 1 is:
R[Ca],b1 =
Cab1
m1H216Ob1
When the overall water mass balance is calculated (i.e. the variation in volume):
R1H216Ob1 = 1
Ratios for external isotopes such as the precipitation and groundwater fluxes are refer-
enced relative to VSMOW such that:
RDHO,p =
(
δD
103
+1
)
RDHO,VSMOW
The isotopic ratios of the evaporation flux (Revp) are determined from Equation 3.1 and
the isotopic ratios of the basin surface (Rb1)
Revp =
Rb1
αevp
3.4.3 Transient model parameterisation
When modelling the isotopic composition of the palaeo-lake, the (i) lake surface temperature
(T ), (ii) wind-induced turbulence (w), (iii) isotopic composition of the atmospheric vapour,
(iv) salinity effect on isotope fractionation, (v) the isotopic composition of the freshwater
input, and (vi) any variability caused by the amount effect must be known or assumed (Table
3.1).
i. To constrain water temperature changes at Lake Chichancanab during the TCP, tan-
dem measurements of both GHW and carbonate δ18O that were deposited concur-
rently permit the deconvolution of the δ18O carbonate signal into its temperature and
δ18O-water components via the carbonate palaeo-temperature equation (Hodell et al.,
2012). To calculate the temperature at which the aragonitic shells of Pyrgophorus
coronatus formed, we used the equation of Grossman and Ku (1986) that is based on
analysis of foraminifera (Hoeglundina elegans) and gastropods:
T °C = 21.8 – 4.59
(
δ18Oarag (PDB)−δ18Owater (SMOW )
)
The δ18O of gastropod aragonite and GHW is used to estimate δ18Oarag and δ18Owater,
3.4 Modelling 81
1.6
1.4
1.2
1.0
De
ns
ity
		(g
/c
m
3 )
150010005000
Age	C.E	(years)
Mean	=	1.09
SD	=	0.11
350
300
250
200
150
100
50
0
Co
un
t	(n
)
1.81.61.41.21.0
Density	(g/cm3)
Signficant	Gypsum	PrecipitaGon
-60
-50
-40
-30
-20
-10
0
%
	Re
du
cG
on
	in
	Pr
ec
ip
ita
Go
n	
1.401.351.301.251.201.151.10
Modeled	Mean	Density	of	Gypsum	Layers	(g/cm3)
Minimal	Constraint	from	Gypsum	Precipitated
-10
-8
-6
-4
-2
0
RH
	 De
cr
ea
se
	( %
)
5.04.54.03.5
Modeled	δ18O	(VSMOW)
1.301.251.201.15
Density	(g/cm3)
Mean	=	4.16
SD	=	0.33
Range	=	3-7%
-10
-8
-6
-4
-2
0
RH
	De
cr
ea
se
	 ( %
)
-22 -20 -18 -16 -14
Modeled	DEXCESS
Range	=	3-7%
Mean	=	-17
SD	=	1.7
-10
-8
-6
-4
-2
0
RH
	De
cr
ea
se
	 ( %
)
-44 -42 -40 -38 -36 -34 -32
Modeled	17OEXCESS
Mean	=	-36
SD	=6
weak	constraint
A
B
C
D
E
F
Fig. 3.9. Transient model precipitation and relative humidity scenarios. (A) GRA bulk density record of
core CH1 7-III-04 (5). From 500 CE to 2000 CE, mean (±1σ ) sediment density = 1.09±0.11 g cm-3
(red horizontal band). (B) Histogram of counts from GRA bulk density record from 500 CE to
2000 CE (C) Example of the relationship between modelled mean core density (g cm-3) and the
modelled reduction in precipitation (%). Varying the minimum cut-off point from which modelled
data are compared to GHW data changes the baseline precipitation reduction estimate. Example
of modelled δ18O (D), d-excess (E), and 17O-excess (F) data plotted as a function of modelled
densities. Mean GHW (±1σ ) data are then compared to modelled runs to derive %RHn decrease.
82 Quantification of drought during the collapse of the classic Maya civilisation
Category Parameter Units Value
L
ak
e
E
va
po
ra
tio
n Lake Temperature °C 25
w - 0.5
Veq - 0.6
Vapour Temperature °C 25
θeq - 0.529
θkin - 0.518
Pr
ec
ip
. δ18O
(h, VSMOW) -4.5d-excess 10
17O-excess 2.5
G
ro
un
dw
at
er
δ18O
(h, VSMOW) -4.5d-excess 10
17O-excess 2.5
Ca
g kg-1
0.607
SO4 2.455
Na 0.2
Mg 0.2
K 0.011
Cl 0.234
Fl
ux
es
Fp m water equivalent/m2/a
Reanalysis data = ~1
Fe 1.07
Fgb
kg (~m water equivalent/m2/a)
1.02e9 (~0.1)
Fbg 1.02e9 (~0.1)
Fb1b2 1.02e10 (~1.0)
Fb2b1 1.02e10 (~1.0)
Table 3.1. Transient model parameters. See text for details of parameters.
3.4 Modelling 83
respectively. Gypsum and gastropod samples were only used if they were in direct
contact with each other, or the shell fragments were found embedded within gypsum.
The derived temperature from Pyrgophorus coronatus and gypsum from the same
bed averaged 25.9 ± 1.7°C (Appendix C; Table C.2). This temperature is indistin-
guishable from the mean annual temperature of the lake today (Hodell et al., 2012).
Equally, because 17O-excess is minimally affected by temperature changes, moderate
variations in mean temperature result in insignificant effects on the trajectories of the
evaporated waters. For example, 5°C of temperature change has a small effect on the
model results for 17O-excess (up to ~±2 per meg in a terminal lake), whereas d-excess
changes by as much as ~3h in a terminal lake, when all other parameters remain
constant (Gázquez et al., 2018). Thus, the model sensitivity to temperature changes
is low.
ii. It is known that the proportion of α 18Okin may be suppressed by turbulent flow
induced by wind (Uemura et al., 2010), and therefore wind could alter isotope mass
balance, especially for d-excess and 17O-excess. The exponent ‘w’ is set between
0.5 (pure turbulence) and 1 (no wind). Measured wind speeds in the region of Lake
Chichancanab are ~3 m s-1 and relatively constant over the year (Soler-Bientz et al.,
2010), resulting in a well-mixed lake-surface layer. When turbulence is not con-
sidered, the model yields d-excess and 17O-excess values that are systematically too
low compared to the analytical data for some modern periods. We also tested this
variable by analysing modern and palaeo-lake water data using previously published
Monte Carlo models (Gázquez et al., 2018) (Section 3.5.1). These tests provide the
best fit to modelled data when w is kept constant at ~0.5 when modelling both the
modern and the TCP GHW data.
iii. The isotopic composition of modern atmospheric vapour is not well constrained in
the Yucatán Peninsula, nor are there estimates available for how this variable has
changed in the past. Assuming equilibrium with local meteoric water, the isotopic
composition of atmospheric water vapour can be approximated (Gázquez et al., 2018).
Gibson et al. (2016), however, suggested that the isotopic composition of atmospheric
vapour is often only in partial equilibrium with that of local freshwater. We assumed a
degree of equilibrium of 60%, a reasonable estimate for most tropical and intertropical
regions (Gibson et al., 2016). Note that Gázquez et al. (2018) show that the triple
oxygen and hydrogen system is relatively sensitive to the isotopic composition of the
vapour, especially for coastal lakes affected by advection of marine vapour masses.
In coastal lakes, the isotopic composition of the modern atmospheric vapour is in
84 Quantification of drought during the collapse of the classic Maya civilisation
equilibrium with seawater rather than freshwater. This is probably not the case at
Lake Chichancanab, which is located ~140 km inland.
iv. High concentrations of NaCl and other salts within a lake can cause the water iso-
topic activity ratios to diverge from the corresponding concentration ratios, as a con-
sequence of isotopic fractionation between free water and water in ionic hydration
shells (Sofer and Gat, 1975; Herwartz et al., 2017). This ‘salt effect’ is different for
hydrogen and oxygen isotope fractionation, resulting in complications when inter-
preting the relationship between δ18O and δD at salt concentrations >100,000 mg l-1
(Sofer and Gat, 1975; Herwartz et al., 2017). The total dissolved salt concentra-
tion in Lake Chichancanab today is ~4000 mg l-1 (Hodell et al., 2005). Substantial
oxygen and hydrogen isotope fractionation effects caused by high salinity would not
be expected at these low concentrations (Sofer and Gat, 1975; Criss, 1999). During
periods of lake drawdown, the transient model displays elevated salt concentrations,
but concentrations that approach 100,000 mg l-1 are never experienced in model runs.
v. The mean (±1σ ) modern isotopic composition of regional groundwater is -4.0 ±1.7h
for δ18O, -2.10 ±0.9h for δ17O and -23.5 ±12.8h for δD. We use δ18O = -4.5 and
δD = -26 (so that d-excess = 10h) and set 17O-excess = 2.5 per meg to model present-
day conditions. These variables are held constant throughout the transient model runs,
although they can be systematically varied to reflect isotope effects such as the amount
effect (see below and Section 3.5). We tested the effect of non-systematic variability
on the isotopic composition of meteoric water (using bounds of ±0.5h for δ18O, ±5h
for d-excess, and ±9.5 per meg for 17O-excess), using the steady state Monte Carlo
model of Gázquez et al. (2018) to quantify derived uncertainty (Section 3.5.1).
vi. The amount effect (i.e. correlation between depletion of heavy isotopes in rain-
fall with greater amount of rain) is thought to play a role in the Yucatán Peninsula
and thus the isotopic composition of meteoric water may vary over the timescales
modelled (Medina-Elizalde et al. 2010; Medina-Elizalde and Rohling, 2012). The
amount effect and its perturbation in the δ18O of precipitation (P) over a seasonal
cycle was calculated from isotopic composition of precipitation collected at Hobonil,
20°00’59”N 89°01’13”W (Hodell et al., 2012), and rainfall amounts from the prox-
imal metrological station of Dziuche, 19°54’00”N 88°48’40”W, from 2006 to 2009.
The linear regression (δ18O = −0.0176(P)− 0.1204; R² = 0.92) is very similar to
that found by Medina-Elizalde et al. (2010) from the IAEA station in Veracruz,
México (δ18O = −0.0118(P)− 0.64; R² = 0.80). We compiled data from Veracruz
between 1969 and 1985 (omitting years with poor data coverage) and the records
3.4 Modelling 85
y = -0.0163x + 11.0
R² = 0.16
y = -0.0261x + 12.8
R² = 0.33
-20 
-10 
0
10
20
30
0 100 200 300 400
d-
ex
ce
ss
 (‰
 V
SM
OW
) 
Precipitation (mm/month)
y = -0.0121x - 0.41
R² = 0.73
y = -0.0117x - 0.94
R² = 0.73
0 100 200 300 400
-8.0 
-6.0 
-4.0 
-2.0 
0.0
2.0
δ1
8 O
(‰
 V
SM
OW
)
Precipitation (mm/month)
A B
Fig. 3.10. The amount effect relationship between monthly precipitation amount and monthly mean rainfall
(A) δ18O and (B) d-excess during the dry (November to May) and rainy seasons (June to October)
from the IAEA station in Veracruz, México (years 1969 to 1985) and Hobonil (20°00’59"N
89°01’13"W; 2006 to 2009). Both raw (blue squares) and amount-weighted (red squares) show
that rainfall δ18O is negatively correlated to the amount of precipitation on seasonal time scales,
whereas there is no significant amount effect displayed by d-excess.
from Chichancanab between 2006 and 2009 (Hodell et al. 2012). The compiled linear
regression for δ18O (δ18O=−0.0121(P)−0.41) displays a significant correlation (R²
= 0.73), whereas there was no significant amount effect displayed by d-excess or 17O-
excess (Fig. 3.10).
3.4.4 Age models
Age uncertainty associated with the lake record and periods of gypsum precipitation was cal-
culated using Bayesian age-depth analysis of radiocarbon ages obtained from the sediment
cores (Hodell et al., 2005). We estimated calendar ages with 95% confidence intervals using
the Bayesian Age–Depth Modelling software ‘BACON’ in R (Fig. 3.11; Appendix C; Table
C.3) (Blaauw and Christen, 2011). Derived age-depth error (95% confidence intervals) was
normalised to the best-fit line to produce age-depth errors for the Chichancanab density
record (Fig. 3.2). To quantify age uncertainty in relation to the timing of dry intervals,
we inverted the density record of core CH1-III-04 and identified data outside one standard
deviation of the mean sediment density from core CH1-III-04. This would be equivalent
to selecting sediment densities >1.2 g cm-3, equivalent to the periods from which GHW
was extracted. The BACON function ‘Events’ was then used to quantify the probability of
arid conditions at any calendar age. 50-year window widths were used, with the windows
moving at 10-year steps from the core’s bottom ages to its top (Blaauw and Christen, 2011).
86 Quantification of drought during the collapse of the classic Maya civilisation
A B 
D C 
Lake Chichancanab 
 (Hodell et al. 2005) 
Lake Chichancanab 
(Douglas et al. 2014) 
Laguna Yaloch 
(Wahl et al. 2013) 
Punta Laguna 
(Curtis et al. 1996) 
Fig. 3.11. Bayesian age models produced for (A) core CHI 7-III-04 of Lake Chichancanab (Hodell et al.,
2005), (B) Plant waxes of Lake Chichancanab (Douglas et al., 2014), (C) Punta Laguna (Curtis
et al., 1996) and (D) Laguna Yaloch (Wahl et al., 2013) by the program BACON. Dark shading
indicates age uncertainty at each depth and the red lines indicate the best-fit age model.
The probabilities of drought occurring specifically during the onset (~750 to ~850 CE) and
the end (~950 to ~1050 CE) of the TCP are high (Probability > 0.85 and Probability >
0.95, respectively). Multiple proxy climate records across the Maya Lowlands also provide
evidence of drought synchronicity, with only slight temporal variations across the region
(Fig. 3.12) (e.g. Douglas et al., 2016a).
We also used Bayesian age modelling techniques to synthesise the regional proxy re-
cords of Curtis et al. (1996), Wahl et al. (2013) and Douglas et al. (2014) (Fig. 3.12). To
quantify age uncertainty in relation to the timing of dry intervals, we inverted the records
and identified the lowest 10th percentile of the raw data for Wahl et al. (2013) and Douglas
et al. (2014), and the 5-point smooth of the δ18O Cytheridella ilosvayi record from Curtis
et al. (1996).
All proxy records dated using radiometric techniques yielded a significant age uncer-
3.5 Discussion 87
tainty. Overall, although individual records show subtle variations in drought timing, the age
uncertainty results in the chance of drought occurring any time between 500 and 1300 CE
(Bhattacharya et al., 2017). Gypsum deposition at Lake Chichancanab displays a significant
chance of occurring from ~750 to ~850 CE (Probability of drought at 800 CE = 0.85),
coinciding with the onset of the collapse of Terminal Classic Maya Civilisation (Fig. 3.12).
3.5 Discussion
To estimate quantitatively the magnitude of drought during the TCP, we employed a transi-
ent model that explicitly simulates the evolution of the isotopic and chemical composition of
the lake water, including the gypsum flux to the lake sediments. The modelled gypsum flux
can be compared to observed variations in the gypsum content of the sediments, as expressed
by variations in sediment bulk density (Hodell et al., 2005). Changes in lake surface-area-to-
volume ratio were obtained from the lake bathymetry (Fig. 3.7). The model was run at sub-
monthly resolution in a series of millennial-duration experiments, forced with NARR data
for local precipitation and RHn. The model was first tested using the climate forcing across
the modern sampling period from 1994 to 2010 (Fig. 3.8). It successfully reproduced the
mean of modern isotope data, with insignificant gypsum precipitation. This time interval,
which was fortuitously one of the driest of recent decades, was then used as the baseline for
comparison to palaeo-simulations.
To provide scenarios that are directly comparable to the GHW data, long transient
simulations were performed in which rainfall and RHn were reduced by variable amounts
to simulate a series of multi-decadal-scale droughts. The use of a model allows us to
compare, directly and quantitatively, climate conditions affecting the modern lake with those
of plausible drought conditions. Precipitation and/or RHn forcings are directly modulated by
the density record of core CH1-III-04 (Hodell et al., 2005); the maximum variability during
a run was set to the maximum density point over time, and precipitation and/or RHn were
reduced linearly across all months. Absolute reductions in P/E and RHn were referenced
relative to mean NARR data over the period from 1994 to 2010. In other words, a 50%
reduction in precipitation and a 5% reduction in RHn would equate to 50% less rainfall in
each month and a 5% reduction in RHn each month (e.g. 75% to 70%) during a modelled
time period compared to the 1994 to 2010 baseline.
First, only the intervals over which the model produced gypsum deposition (modelled
sediment density >1.1 g cm-3) were selected. The periods of modelled gypsum accumulation
were then aggregated into drought conditions for a given scenario via two pathways: (i)
all model variables were averaged across all of the droughts, and (ii) probably density
88 Quantification of drought during the collapse of the classic Maya civilisation
-6.0
-5.0
-4.0
-3.0
-2.0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
-8.0
-7.0
-6.0
-5.0
-4.0
-3.0
-2.0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
0
0.2
0.4
0.6
0.8
1
0
10
20
30
40
50
60
0
0.2
0.4
0.6
0.8
1
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
Speleothem 
δ18O 
(‰, VPDB)
Drought
Probability 
Yok I 
(Kennett et al. 2012)
Drought
Probability 
Drought
Probability 
Drought
Probability 
Speleothem 
δ18O 
(‰, VPDB)
CaCO3 
(wt%)
Cytheridella 
ilosvayi
δ18O 
(‰, VPDB)
Plant wax 
δDwax-corr
(‰, VSMOW)
Sediment 
density 
(g/cm3)
Laguna Yaloch 
(Wahl et al. 2013)
Lake Chichancanab 
(Hodell et al. 2005)
Lake Chichancanab 
(Douglas et al. 2012)
Punta Laguna 
(Curtis et al. 1996)
Chaac 
(Medina-Elizalde et al. 
2010)
Age (C.E.)
Age (C.E.)
0
0.2
0.4
0.6
0.8
1
-4.0
-3.0
-2.0
-1.0
0.0
1.0
2.0
A
B
C
D
E
F
0
0.2
0.4
0.6
0.8
1
-40
-30
-20
-10
0
10
20
3.5 Discussion 89
Fig. 3.12. Left: Selection of palaeoclimate records from the Maya Lowlands, arranged from north to south:
(A) Chaac speleothem (Tzabnah Cave) δ18O (Medina-Elizalde et al., 2010); (B) Punta Laguna
δ18O (Curtis et al., 1996); (C) Lake Chichancanab δDwax-corr (Douglas et al., 2014); (D) Lake
Chichancanab sediment density (Hodell et al., 2005) (E) Laguna Yaloch weight percent calcium
carbonate (CaCO3) (Wahl et al., 2013); (F) Yok I speleothem (Yok Balum Cave) δ18O (Kennet et
al., 2012). Age uncertainty analysis for proxy sites displays drought probabilities in given time
intervals for the proxy data. All age models were calculated using Bayesian age analyses. The
vertical grey bar indicates the TCP between 800 and 1000 CE. δDwax-corr values indicate δDwax
values corrected for the influence of vegetation change (Douglas et al., 2014).
functions were constructed incorporating the variability within and between each decadal-
length drought (Fig. 3.6). Data consistent scenarios were then selected by excluding
those model runs that fell outside the 1σ range of the isotope data and where, on average,
the model failed to produce significant gypsum accumulation (cut-off of average density
<1.2 g cm-3 based on 1σ range; Fig. 3.9A).
A major constraint on the modelled P/E variability is the sediment density range from
which the modelled data is selected. As displayed in Fig. 3.9, if the modelled sediment
density threshold is changed from 1.2 g cm-3, the lower bound of P/E and RHn estimates will
vary systematically. Importantly, increasingly conservative estimates for periods of gypsum
precipitation (i.e. modelled sediment density >1.2 g cm-3) produce increasingly more severe
reductions in baseline P/E and RHn.
Two possible scenarios were tested subsequently: (i) a reduction in precipitation with
accompanying shifts in the isotopic composition of rainwater (i.e. the amount effect) and,
(ii) a reduction in precipitation with accompanying decreases in RHn (Fig. 3.13).
In the first scenario, precipitation δ18O was reduced according to the amount effect
relationship (i.e. δ18Oprecipitation/ΔPrecipitationvolume = –0.0121 mm-1; Fig. 3.10) with
associated changes in δD and δ17O that track the Global MWL (i.e. no changes in d-
excess or 17O-excess). No scenarios with these assumptions are able to reproduce the
relationship between δ18O, d-excess and 17O-excess observed in the data. If the constraints
provided by d-excess and 17O-excess are removed and only δ18O and gypsum precipitation
are employed, our model permits reductions in precipitation that average 50% over all
drought intervals (Fig. 3.13, blue lines). This estimate is in broad agreement with previous
work that relied on carbonate δ18O-derived precipitation estimates (using the local amount
effect), which predicted reductions of up to 40% (Medina-Elizalde et al., 2010; Medina-
Elizalde and Rohling, 2012). Our greater estimate of 50% is in part a consequence of
the peak drought δ18O values recorded by gypsum, as well as the integration of simulated
gypsum formation and true lake bathymetry in the model. Crucially, however, the added
information from the d-excess and 17O-excess data suggests that multi-decadal shifts in
90 Quantification of drought during the collapse of the classic Maya civilisation
x10-3
P
re
ci
pi
ta
tio
n
(m
m
/y
ea
r)
1000
500
0
80
75
70
65
60
600 800 1000
A
R
el
at
iv
e
H
um
id
ity
 (%
)
δ18
O
(‰
 V
S
M
O
W
)
6
5
4
3
-10
-15
-20
-25
-30
d-
ex
ce
ss
(‰
 V
S
M
O
W
)
15
14
13
12
11
La
ke
 D
ep
th
(m
)
2.0
1.8
1.6
1.4
1.2
1.0
D
en
si
ty
 (g
/c
m
3 )
M
od
el
ed
 G
yp
su
m
P
re
ci
pi
ta
tio
n
600 800 1000
Age (C.E.) δ18O  (‰ VSMOW)
δD
 (‰
 V
S
M
O
W
)
20
18
16
14
12
10
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Modern Lake Water 
Paleo-lake Water 
Modeled Gypsum Precipitation 
No gypsum precipitation
Model Scenario 1
Model Scenario 2
-10
-12
-14
-16
-18
-20
-22
-20
-30
-40
-50
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
d-
ex
ce
ss
 (‰
 V
S
M
O
W
)
17
O
-e
xc
es
s
 (p
er
 m
eg
)
Model Scenario 2
Model Scenario 1
Model Scenario 2
Model Scenario 1
B
C
D
Probability Density 
Function (PDF)
Model Scenario 1
Model Scenario 2 
3.5 Discussion 91
Fig. 3.13. Left: (A) Transient model of the lake system from 550 to 1200 CE GHW data (yellow
circles) and core density are plotted against sampling ages derived from Bayesian age-depth
analysis. Multi-decadal-scale droughts were simulated by forcing (1) a reduction in precipitation
with accompanied shifts in the isotopic composition of rainwater (i.e. the amount effect,
δ18Oprecipitation/∆Precipitationvolume = –0.0121h/mm; Scenario 1, blue line) and (2) reductions
in precipitation with accompanied decreases in RHn (Scenario 2, red field). Probably density
functions incorporate the variability within and between each decadal-length drought (GHW data
= yellow line; scenario 1 = blue line; scenario 2 = red line). Scenario 1 fails to match
the d-excess data derived from GHW. Scenario 2 successfully reproduces all and d-excess data.
When all model variables were averaged across all droughts, the mean precipitation and reduction
(closed red boxes adjacent to PDFs) is 47% (with a 1σ level of 41-54%) and 4% (1σ level 2-7%),
respectively. The 1±σ range determined from probability density functions (open red boxes
adjacent to PDFs) shows the variability of precipitation and throughout the droughts. Scenarios
1 and 2 are also plotted as (B) δ18O vs δD, (C) δ18O vs d-excess and (D) δ18O vs 17O-excess.
Open circles indicate points in the model at which gypsum is precipitating; dots indicate modelled
data points when gypsum is not precipitating. Error bars (±1σ ) are shown or are smaller than the
symbols.
the δ18O of precipitation (caused by the amount effect) were not the dominant factor that
affected the isotope budget of Lake Chichancanab during the TCP.
In the second scenario, precipitation was reduced without changes in the δ18O of pre-
cipitation, but instead with concurrent changes in RHn. In this case excellent agreement
between the modelled evolution of all isotopic data is observed, with increases in δ18O
accompanied by decreases in d-excess and 17O-excess (Fig. 3.13, red lines). This analysis
yielded plausible scenarios of precipitation reduction that average 47% across all droughts
(with a 1σ level of 41% to 54%) accompanied by RHn reductions of 4% (1σ level 2% to
7%). This result provides a robust, quantitative estimate of the mean annual hydrological
conditions of the combined drought periods during the TCP at Lake Chichancanab.
It should also be noted that a combination of RHn reduction and the amount effect could
potentially reproduce the observed data during the TCP. Although a combination of RHn
reduction and the amount effect cannot be definitively ruled out, given the great number of
possible outcomes when modelling three variable parameters, our experiments show that
the amount effect does not dominate the isotopic budget of the lake. Thus, any influence of
the amount effect (in combination with RHn reductions) at Lake Chichancanab during the
TCP will have little effect on modelled outcomes because of the dominance of RHn in the
17O-excess and d-excess signal.
Although the time evolution of our model is not a direct reconstruction of climate condi-
tions, the model permits heterogeneity within and between each decade-long drought. The
±1σ range determined from the probability density functions indicates that the precipitation
reduction could vary from 20% to 70% throughout the modelled droughts (Fig. 3.13). This
variability represents the transition into and out of drought phases and demonstrates that
92 Quantification of drought during the collapse of the classic Maya civilisation
the severity of the droughts could be intense (up to a 70% reduction in precipitation), while
maintaining the isotope balance and without desiccating the lake. Although variability in the
seasonal delivery of rainfall (or lack thereof) is difficult to constrain because the residence
time of the lake water is greater than an annual cycle, our results provide quantitative
estimates for the total annual reduction in the water available to the ancient Maya for
agricultural and domestic use. Note that recorded Colonial-period accounts of later droughts
(e.g. 1535-1560 and 1765-1773), during which high mortality, famines, and population
displacement were reported (Hoggarth et al., 2017), are not manifest as intervals of gypsum
precipitation in Lake Chichancanab. The lack of gypsum formation is likely a result of
shorter duration and/or lower severity of these droughts, providing further evidence that the
TCP was an unusually dry period for the Holocene on the Yucatán Peninsula.
3.5.1 Monte Carlo modelling scenarios
In addition to transient model experiments described above, a previously described steady
state model of the lake was used in a series of Monte Carlo experiments to determine the
range of climatological conditions that simultaneously satisfy all stable isotope results of
GHW, in combination with statistical estimates of uncertainty (Gázquez et al., 2018). In
these scenarios, the parameter ‘Xe’ represents the hydrologic balance of the lake. A scenario
in which all water is lost by outflow and no evaporation occurs is represented as Xe =
0, whereas a scenario in which all water is lost to evaporation (i.e. a terminal basin) is
represented as Xe = 1. The Xe of Lake Chichancanab is not believed to have changed over
the last ~1500 years and near-terminal conditions are expected (Perry et al. 2002). Equally,
Gázquez et al. (2018) show that when Xe ranges from 0.75 to 1, changes in this parameter
barely affect the RHn values derived from this Monte Carlo model. Here, conservative
estimates of Xe were chosen between 0.8 and 0.9 to cover all likely variability (Table 3.2).
Modelled GHW data suggest there was a reduction in RHn of between 2% and 9%
during the TCP, compared to the period when modern lake water was sampled (1994-2010),
in good agreement with our transient experiments (2% to 7%) (Appendix C; Table C.8).
Estimated errors for RHn are smaller than 4% (1σ ) when all variables listed in Table 3.2
are considered. The slightly greater estimates for reductions in RHn relative to the transient
model arise because the steady-state assumption in the Monte Carlo model precludes an ac-
curate simulation of isotopic balance during lake level fluctuations. Because the model does
not account for the additional isotopic enrichment in the lake due to decreased P/E during a
lake-level drawdown, it is slightly biased towards higher estimates of RHn reduction.
3.6 Conclusions 93
Category Parameter Units Values [Range]
L
ak
e
E
va
po
ra
tio
n Lake Temperature °C [22 to 28]
w - [0.5 to 0.6]
Veq - [0.6 to 0.7]
Xe - [0.8 to 0.9]
Vapour Temperature °C 25
θeq - 0.529
θkin - 0.518
Pr
ec
ip
. δ18O
(h, VSMOW) [-5 to -4]d-excess [5 to 15]
17O-excess [-7 to 12]
Table 3.2. Steady-state parameters for Monte Carlo simulations. See text for details of parameters.
3.6 Conclusions
Using triple oxygen and hydrogen isotope data to independently deconvolve climate vari-
ables precipitation, RHn and the amount effect, the changing hydrological conditions at Lake
Chichancanab have been constrained. This approach provides a significant advance over
previous attempts to estimate the magnitude of rainfall reduction during the TCP droughts
(e.g. Medina-Elizalde and Rohling, 2012). Furthermore, these quantitative estimates of
past rainfall and RHn can serve as input variables in crop models, to better understand how
drought affected agriculture (e.g. maize production) in the northern Maya Lowlands during
the TCP (Yaeger and Hodell, 2009).

Chapter 4
The Messinian Salinity Crisis and
methods of study
4.1 Introduction
‘Saline giants’, extensive accumulations of evaporitic salts of up to several kilometres in
thickness (Hsü et al., 1973), have formed throughout the Phanerozoic, but have no mod-
ern equivalent. The largest Quaternary evaporite basins (e.g. Salar de Uyuni, 9654 km2,
~120 m thick; Salar de Atacama, 3064 km2, >900 m thick) are largely insignificant in scale
when compared to ancient evaporative systems such as the Cretaceous Aptian salt basins of
the southern Atlantic (741,000 km2, >900 m thick) and the Triassic Hith Formation of the
Middle East (1,107,000 km2, ~100 m thick) (Evans, 2006; Warren, 2010). Saline giants
are commonly the focus of applied research by the petroleum industry because of the
sealing capacity of salt rock and the association hydrocarbon fluids. In the absence of
modern analogues, however, there are major difficulties in establishing a comprehensive
understanding of the hydrological, sedimentological, and geochemical processes that may
operate in such extensive evaporative basins (Ba¸bel, 2007; Ba¸bel and Schreiber, 2014).
The geologically most recent of the known saline giants was deposited in the Mediter-
ranean region during the late Miocene. Plate tectonic convergence of Africa and Eurasia
formed the various Mediterranean basins, and subduction dynamics in the Alborán region
resulted in a reduction of the hydrological exchange between the Mediterranean and the At-
lantic (Flecker et al., 2015). A combination of this tectonic forcing (e.g. Garcia-Castellanos
and Villaseñor, 2011), eustatic factors (e.g. Manzi et al., 2013), and astronomically-forced
climate change (e.g. Krijgsman et al., 1999; Hilgen et al., 2007; Lugli et al., 2010) is
thought to have driven fluctuation in Mediterranean salinity, resulting in the precipitation
96 The Messinian Salinity Crisis and methods of study
Fig. 4.1. Modern topographical map of the Mediterranean region. Messinian evaporite deposits, the major
rivers/drainage basins are indicated. White arrows show Mediterranean outflow water (MOW).
Numbers refer to Messinian and modern marginal and intermediate basins. From Simon (2017).
of over one million cubic kilometres of evaporite deposits during the period known as the
Messinian Salinity Crisis (MSC; 5.97-5.33 Ma) (Hsü et al., 1973; Krijgsman et al., 1999;
Rouchy and Caruso, 2006; Roveri et al., 2014a).
The concept of a late Miocene ‘salinity crisis’ (Selli, 1960) was first formulated from
onshore studies, while the deep basin extent of the Messinian salt was only recognized
after seismic investigations carried out during the 1960s and 1970s were confirmed by the
discovery of evaporites during Deep Sea Drilling Project (DSDP) Leg 13 (Hsü et al., 1977,
1973). The deposits of the Mediterranean have been subsequently recognised as unique
amongst the saline giants due to their relatively young age and relative lack of significant
tectonic modification (Warren, 2010; Roveri et al., 2014b). It has also been suggested that
the MSC impacted global climate systems and ocean chemistry by contributing warm, saline
water to northern latitudes (white arrows, Fig. 4.1), influencing the Atlantic Meridional
Overturning Circulation (AMOC) (Hernandez-Molina et al., 2014), in addition to seques-
4.1 Introduction 97
Fig. 4.2. Illustration of the main features of the MSC and Mediterranean’s exchange history including
lithology, Mediterranean salinity, a qualitative representation of gateway size and the probable
drivers (tectonics, erosion, sea level) of changing dimensions and the Mediterranean’s freshwater
budget (E = evaporation; P = precipitation over the Mediterranean; R = river discharge into the
Mediterranean Sea). From Simon (2017), modified from Flecker et al. (2015).
tering 5% (Ryan, 2009) to 10% (Garcia-Castellanos and Villaseñor, 2011) of the global salt
content into the Mediterranean. The age and scale of this single, giant evaporite system
has resulted in its repeated use as a cornerstone for the investigation of evaporite deposition
(Meilijson et al., 2018).
As with the other saline giants that developed episodically over geological time, the
causes of and mechanisms by which large volumes of gypsum, anhydrite and halite accu-
mulated on the Mediterranean seafloor over a relatively brief period of ~640 ka are contro-
versial. This is in part because of the absence of modern evaporitic systems comparable
in terms of both size and mineralogy (Evans, 2006; Warren, 2010). Moreover, the absence
of a complete cored succession from the deep basinal Mediterranean prevents a complete
synthesis of onshore and offshore records (Roveri et al., 2014a). Thus, the evolution of the
MSC has been traditionally explained by using observations and data from uplifted marginal
areas (Fig. 4.1). In onshore successions, a stratigraphic consensus model has been proposed
comprising of three sedimentary units characterised by distinct evaporitic associations (see
CIESM, 2008 and Roveri et al., 2014a). In stratigraphic order, these units can be broadly
categorised as: the Primary Lower Gypsum (Lower Evaporites), followed by halite with
98 The Messinian Salinity Crisis and methods of study
clastic gypsum, ending in either brackish deposits (Lago Mare) or gypsum sequences (Upper
Evaporites) (Fig. 4.2). Because the related deposits are mostly hidden from scrutiny in the
deep basin, correlation between onshore and offshore sediments is difficult, hampering the
development of a comprehensive stratigraphic model (Simon and Meijer, 2017).
The subsequent chapters are individually focused on the interpretation of the two time-
distinct periods of gypsum deposition - The Primary Lower Gypsum (Chapters 5 and 6)
and the Upper Gypsum (Chapter 7) - to provide a robust framework to interpret the causes
of and mechanisms by which large volumes of gypsum accumulated on the Mediterranean
sub-basins during the MSC. In order to give clear stratigraphic context for this subdivision,
the currently accepted stratigraphy of the MSC is outlined in this chapter (Section 4.2).
Subsequently, both traditional and novel isotope tracers, and their potential to interpret
the hydrological variability during gypsum deposition, are described (Section 4.3). These
methods provide new insights into the conditions under which gypsum forms, and help
distinguish changes in Mediterranean-Atlantic connectivity from the regional freshwater
budget during the generation and evolution of evaporite formation across the MSC.
4.2 Stratigraphy of the MSC
In 2008, the CIESM (Mediterranean Science Commission) defined a stratigraphic consensus
model consisting of three main stages during the MSC (Fig. 4.3). Although this consensus
model can explain most (currently available) data, knowledge of the MSC sedimentary
evolution is heavily biased by onshore exposures. Within the CIESM (2008) framework,
we outline new offshore and modelling data that potentially calls for a revaluation of the
commonly accepted stratigraphic models for the MSC (e.g. Simon and Meijer, 2017;
Meilijson et al., 2018).
4.2.1 Mediterranean onshore successions
In marginal basins of the circum-Mediterranean, onshore pre-evaporite successions consist
of cyclic alternations of homogeneous marls, sapropels and diatomites. The pre-evaporite
carbonate-marls sediments were deposited in response to astronomically-driven climate
oscillations (e.g. Krijgsman et al., 1999; Sierro et al., 2001, 1999) and have been correlated
and cyclostratigraphically tuned across the Mediterranean basin (e.g. Hilgen et al., 2007).
The lowermost gypsum bed marks the onset of the MSC (Krijgsman et al., 1999; Manzi
et al., 2013), which was originally dated at 5.96 Ma by Krijgsman et al., (1999) using the
same integrated high-resolution stratigraphy. This age was later redefined to 5.971 Ma by
4.2 Stratigraphy of the MSC 99
Base of Pliocene
Top Salt
Messinian 
Erosional 
Surface (MES)
On-shoreOff-shore
Core 
Record
Shallow 
basins
Intermediate 
basins
MSC
stages
Age
[Ma]
Pliocene
ODP-DSDP Outcrop, seismic, core data (CIESM, 2008) 
MSC
stages
Pre-MSC
MSC Onset
3.2
3.1
2
1
Upper
Gypsum
Lago Mare
hiatus
of variable 
amplitue
Primary 
Lower 
Gypsum
3.2
1
NO EVAPORITES
euxinix shale
and dolomites
Lago Mare
Gypsum/
dolomite
Anhydrite/
Halite
?
Pliocene
5.33
5.42
5.55
5.60
5.97
MES
Fig. 4.3. Late Miocene chronostratigraphy of the Mediterranean (modified after Roveri et al., 2014a).
Messinian Salinity Crisis (MSC) Stage 1 is marked by the deposition of the Lower Evaporites
or Primary Lower Gypsum in the shallow basins and euxinic conditions in the deep basins. MSC
Stage 2 is characterised by erosion on the shelves, and halite deposition in the deep basins. MSC
Stage 3 is subdivided in the Upper Evaporites or Upper Gypsum (3.1) and the brackish to fresh
Lago Mare facies (3.2).
Manzi et al. (2013), based on a reinterpretation of the lowermost gypsum bed in the Sorbas
basin, south-eastern Spain.
Stage 1 (~5.97-5.61 Ma) – Primary Lower Gypsum The first period of evaporite depos-
ition, known as the Primary Lower Gypsum (PLG), comprises up to 16 beds of massive
gypsum (1 m to 35 m in thickness), interbedded with thin layers (~0.5 m to 5 m in thickness)
of finely laminated marls or shale (Roveri et al., 2009; Lugli et al., 2010). The deposition
of these evaporate-marl cycles is also reported to be controlled by the variations of Earth’s
orbital parameters, mainly the ~21 kyr precession cycle (Krijgsman et al., 2001, 1999; Lugli
et al., 2010; Manzi et al., 2013). According to this interpretation, evaporite deposition
occurred at the precession maxima (i.e. minimum summer insolation when summer solstice
and aphelion coincided) when evaporation exceeded precipitation during periods of dry
climate, and marl deposition occurred at precession minima (maximum summer insolation
when summer solstice and perihelion coincided) when rainfall increased (Krijgsman et al.,
2001, 1999) (Fig. 4.4). Assuming precession does control the periodicity of gypsum
deposition, the top of Stage 1 Lower Evaporites has been determined to be 5.61 Ma by
cyclostratigraphy (Krijgsman et al., 2001; Hilgen et al., 2007; CIESM, 2008). In situ PLG
deposits are well preserved in the western and central Mediterranean regions such as the
Betic Cordillera and the Apennines (Roveri et al., 2014a). In other locations, the develop-
100 The Messinian Salinity Crisis and methods of study
ment of PLG evaporites can be deduced from their occurrence as isolated blocks (e.g. Za-
kynthos, Aegean Sea; Ba¸bel, 2004; Lugli et al., 2010) or within resedimented deposits (e.g.
Cyprus, Levant margin; Lugli et al., 2013). The lower ~5 cycles of all PLG sections consist
of massive, vertically-elongated, twinned selenite crystals described as ‘arrow-headed’ or
‘swallow-tailed’ (Ba¸bel, 2004; Lugli et al., 2010) (Fig. 4.4). The crystals have a preferred
vertical orientation and produce a ‘palisade’-like structure (Dronkert, 1985). The crystalline
size of gypsum is generally large (up to 50 cm in length) and homogenous, although a slight
reduction in crystal size is seen up-section in some cycles. The massive aspect of these
selenite beds is thought to indicate that the brine concentration did not drop to carbonate
saturation (e.g. Lugli et al., 2010). The lower contact between the overlying gypsum and
underlying carbonate marls is often undulating, and contains structures interpreted to be
‘nucleation cones’ (Dronkert, 1985). These load structures consist of conical clusters of
crystals that occur at the base of the gypsum beds, representing the initial nucleation points
of gypsum that sank into the relatively soft marls (Lugli et al., 2010). The occurrence of the
branching ‘supercones’ (Dronkert, 1985) from the sixth cycle upwards (Lugli et al., 2010) is
of particular importance as it provides a marker bed for correlation to other sections across
the Mediterranean (Roveri et al., 2014a). Indeed, PLG sections across the Mediterranean
are sufficiently similar in terms of vertical facies distribution and cycle stacking pattern
that they have been used for basin-wide correlation of individual cycles (Lugli et al., 2010;
Roveri et al., 2014a). These basin-wide stratigraphic similarities, combined with measured
strontium isotope values which have been reported to be ‘within error’ of, or close to,
open ocean values (see section 4.3.5), suggest that the PLG precipitated from a relatively
homogeneous Atlantic-fed water body (Lugli et al., 2007, 2010; Roveri et al., 2014a; García-
Veigas et al., 2018).
From the sixth cycle upwards, there is a significant change in facies across all PLG
sections (Fig. 4.4). While the base of the gypsum beds in each individual cycle generally
consist of the largest crystals (up to ~30 cm in length), the crystal size becomes progressively
smaller (up to a few centimetres) up section (Fig. 4.5). Massive palisade gypsum often
gives way to a ‘banded selenite’ facies, consisting of finer (~1 cm to 10 cm) crystals that
are separated by thin carbonate laminae (Lugli et al., 2010). According to Ba¸bel (2007),
such features reflect characteristic fluctuations of the pycnocline (i.e. the gypsum-saturation
interface) that repeatedly interrupts selenite growth. Above the banded selenite, the gypsum
grades into giant cauliflower-shaped supercones (SC) in some cycles. Supercone structures
are created from clusters of small gypsum crystals creating curved, horizontal branches up
to ~2 m in length. The branches originate from a central nucleation region and gradually in-
crease in length towards the top of the section, thus forming the inverted cauliflower-shaped
4.2 Stratigraphy of the MSC 101
δ
1
8O
 (‰
, V
P
D
B
)
A
in
 E
l B
e
id
a
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
S
u
m
m
e
r 
In
s
o
la
tio
n
 6
5
°N
O
b
liq
u
ity
 
E
T
P
C
a
rb
o
n
a
te
 M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
C
a
rb
o
n
a
te
s
H
a
lite
0
m
5
0
m
1
0
0
m
1
5
0
m
MSC Stages
321
Age [Ma]
5
.4
5
.3
3
5
.5
5
.6
5
.7
5
.8
5
.9
6
.0
Pre-MSC
H
a
lite
B
re
ccia
te
d
Lim
e
sto
n
e
C
h
a
o
tic
G
y
p
su
m
Resedimented 
Lower Gypsum
U
p
p
e
r 
G
y
p
s
u
m
P
rim
a
ry
 
L
o
w
e
r 
G
y
p
s
u
m
D
e
e
p
 B
a
s
in
S
ite
 3
7
2
S
ite
 6
5
4
1
5
0
m
2
0
0
m
2
4
5
m
3
0
1
m
3
1
2
m
M
e
s
s
in
ia
n
 
E
ro
s
io
n
a
l 
S
u
rfa
c
e
O
N
S
E
T
o
f M
S
C
A
rid
W
e
t
-6
-3
0
3
6
-0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
0
0.5
1
1.5
2
2.5
3
3.5
380
440
500
560
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
Fig. 4.4. Messinian MSC chronostratigraphy and astronomical tuning of the Mediterranean Messinian
Lower Gypsum (Yesares Member, Sorbas basin, SE Spain), halite and Upper Gypsum (Eraclea
Minoa, Sicily). Schematic representation of the vertical facies arrangement of an ideal PLG and
UG section orbitally tuned, with gypsum beds calibrated with the aridity peak of the insolation
curve. The revised benthic δ18O record of site 982 from Drury et al. (2018) is also compared
to the benthic δ18O record from Ain El Beida (van der Laan et al., 2005). The Primary Lower
Gypsum unit was deposited during stage 1 of MSC. During stage 3 of the MSC, the Upper Gypsum
was deposited in Sicily and Eastern Mediterranean, while the Northern Apennines and the Betics
are characterised by deposition of terrigenous units. Proposed correlation between the onshore
sections and the offshore deposits recovered from DSDP/ODP cores are also shown (see Roveri
et al., 2014a).
102 The Messinian Salinity Crisis and methods of study
Marl
Supercone
Banded 
Selenite
Palisade 
Gypsum
Marl
A
1m
Nucleation Cone
B
Fig. 4.5. Cycle 10 of the PLG (Yesares Member, Río de Aguas section, SE Spain), displaying the differing
crystal morphology and structures (A) relative to a schematic diagram (B) (after Dronkert, 1985).
Carbonate marls occur at the onset and termination of the cycle. Nucleation cones represent the
onset of gypsum precipitation. Palisade gypsum directly above the nucleation cones contains large,
vertically orientated selenite crystals. Selenite palisade is overlain by banded selenite that are, in
turn, overlain by asymmetric conical structure (supercones) juxtaposed to carbonate marls.
cones (Dronkert, 1977; Lugli et al., 2010). Laminated marls can occur in pods between
the branches of the supercones and carbonate minerals are composed almost entirely of
dolomite (Evans et al., 2015; Chapter 5). Some of the carbonate laminae can be traced
into the gypsum crystals where they thin and disappear, indicating the gypsum and marl are
syndepositional.
Stage 2 (~5.61-5.55 Ma) – resedimented gypsum and halite While the halite phase of
Stage 2 only developed in the deep basins, the upper boundary of the PLG in some marginal
basins is incised by a major erosive event known as the Messinian Erosion Surface (MES).
The marginal erosion surface was caused by a sea level drop, which resulted in reworking
and transport of sediments to the Mediterranean basin centre (see Roveri et al., 2014a and
references therein). Glacial stages TG14 and TG12 occurred during this interval and may
have caused further restriction of the Mediterranean-Atlantic exchange (e.g. Hilgen et al.,
4.2 Stratigraphy of the MSC 103
2007). The total sea level fall during this period is uncertain. The drop in sea level is thought
to have promoted slope instability and gravity failure around the basin margins, resulting in
the Resedimented Lower Gypsum (RLG) unit (Roveri et al., 2008).
Stage 3 (~5.55-5.33 Ma) – Upper Evaporites and Lago Mare Stage 3 deposits developed
in the deep basins, which have been subsequently uplifted into marginal settings. Sub-basins
of the southern and eastern Mediterranean (i.e. Sicily and Cyprus; Rouchy and Caruso,
2006; Manzi et al., 2013, 2009) contain both selenite and cumulate gypsum facies (the
‘Upper Evaporites’ or ‘Upper Gypsum’ (UG); Manzi et al., 2009), while in the northern
and western areas, evaporite-free ‘Lago Mare’ (e.g. Orszag-Sperber, 2006) units are com-
prised of grey-to-white marls interbedded with deltaic or fluvial conglomerates (Fortuin
and Krijgsman, 2003). These sediments contain fauna that live in brackish water (Bassetti
et al., 2004; Orszag-Sperber, 2006). The development of potential shallow-water environ-
ments during this period is also implied from strontium isotope values (measured on both
evaporites and fossils) that diverge away from the global oceanic water value, indicating
a substantial dilution of surface waters, locally punctuated by episodic, precession-driven,
evaporitic events (e.g. Flecker et al., 2002; Flecker and Ellam, 2006; Roveri et al., 2014b).
The Upper Evaporites of Stage 3 consist of repeated gypsum-carbonate marl beds that,
similar to PLG evaporitic cycles, represent a response to precession-driven climate oscil-
lations (see Manzi et al., 2009 and references therein). Exposures on Sicily and Cyprus
show alternations of gypsum beds (3 m to 8 m thick) and marl horizons (5 m to 20 m thick)
containing typical Lago-Mare flora and fauna (e.g. Rouchy et al., 2001; Gliozzi et al., 2007;
Londeix et al., 2007). The most complete stage 3 succession occurs at Eraclea Minoa,
Sicily, where 10 sedimentary cycles (7 with gypsum) have been recognised and correlated
regionally between the top of the stage 2 deposits and the base of the Pliocene (Van Der
Laan et al., 2006; Hilgen et al., 2007; Manzi et al., 2009) (Fig. 4.4).
The UG shows different facies and cycles thickness with respect to stage 1 PLG deposits
(Manzi et al., 2009) (Fig. 4.6). Much of the Eraclea Minoa section, for example, consists
of marl and thinly-bedded sandstone packages that are tens of metres thick, showing a
slight coarsening and thickening up-section. Before the onset of gypsum beds, the marl
and sandstone deposits are overlain sharply by decimetre-thick tabular graded sandstone
beds (unit S1). The first gypsum consists of decimetre-thick, fine-grained gypsum (<5 mm)
and grey marl graded beds (unit G1a) that sit conformably on the underlying marl and
sandstone layers. This unit is characterised by traction and fallout sedimentary structures
indicating deposition from high-density to low-density gravity flows (Manzi et al., 2009).
Fine-grained laminar gypsum beds (unit G1b) overlie unit G1a, each bed up to a few tens
104 The Messinian Salinity Crisis and methods of study
Sandstone
Fine-grained gypsum-marl 
Laminar gypsum beds 
Selenite Crusts
Massive Selenite
Domed Selenite
Fine-grained gypsum-marl 
G2a
G2b
G2c
G1a
G1b
G1a
S1
Fig. 4.6. Cycle 4 of the UG (Eraclea Minoa section, Sicily), displaying the main recognised facies and facies
associations after Manzi et al. (2009). See text description for description.
of centimetres thick. Above, the laminar gypsum transitions into centimetre-thick, selenite
crusts (unit G2a), likely formed as bottom growth primary evaporites. Unit G2a is followed
by decimetre-thick massive selenite beds (each bed up to ~1 m thick), which are separated
by thin marl horizons (up to ~30 cm thick) (unit G2b). At the top of the gypsum sections,
domed structures (unit G2c) are present. These selenite domes have been compared to the
dome-like features found in modern marine salinas and marine-fed lakes (Warren, 1982).
Locally, reworked selenite beds up to a few centimetres in thickness can be found at the tops
of the domes (Manzi et al., 2009). While these features have been traditionally interpreted
as erosional surfaces (e.g. Rouchy and Caruso, 2006), Schreiber (1997) interprets these
truncations as the point at which gypsum growth stops at the saturation water interface (i.e.
the pycnocline).
4.2 Stratigraphy of the MSC 105
4.2.2 Mediterranean offshore successions
In the deep basin, seismically-imaged evaporite bodies have been subdivided into three
units characterised by different seismic facies. From oldest to youngest, they are labelled as
the Lower Unit, Mobile Unit, and Upper Unit. These seismic units have been historically
correlated with the PLG, Halite and UG of the onshore record, respectively (Roveri et al.,
2014a). The offshore and onshore Messinian units are physically disconnected, however,
and establishing reliable correlations is difficult.
Stage 1 (~5.97-5.61 Ma) Little is known about the deposition of the offshore sediment-
ary units coeval to the onshore PLG, but evidence from the Apennines and geochemical
relationships suggest that organic-rich shales and dolomites accumulated in the deep basins
(De Lange and Krijgsman, 2010; Flecker et al., 2015). De Lange and Krijgsman (2010)
argued that the possible lack of gypsum in these deep environments is because sulfate acts
as an oxidant for organic matter mineralisation in deep water. In hypersaline settings, water
masses become anoxic due to high-density stratification and lack of ventilation. Sustained
sulfate consumption results in a diminished deep water sulfate concentrations and thus
undersaturation with respect to gypsum (De Lange et al., 1990; Torfstein and Turchyn,
2017). The balance between organic matter remineralisation-related sulfate consumption
and the supply of sulfate from shallow-water gypsum formation may therefore determine
if the deep-water sulfate concentration becomes too low to permit gypsum formation and
preservation (De Lange and Krijgsman, 2010).
Stage 2 (~5.61-5.55 Ma) The primary evaporitic deposits during this phase include the
thick and extensive halite and K–Mg salts (mainly represented by kainite, MgSO4KCl ·3H2O,
but also; carnallite, KMgCl3 ·6H2O; bischofite, MgCl2 ·6H2O; and sylvite, KCl) (see Yoshimura
et al., 2016 and references therein). The presence of K-Mg salt minerals reflects extreme hy-
persaline environments during this stage. This extreme stage is often linked to a desiccated
Mediterranean (e.g. Hsü et al., 1973).
In the present stratigraphic scenario, the deposition of halite and K-Mg salts occurs in a
short period of ~50 ka during the MSC acme (CIESM, 2008; Roveri et al., 2014a). Recent
analysis from cuttings and well-log data from industry drilling in the eastern Mediterranean
Levant Basin, however, has shown that halite may have started to precipitate in the deep
basins around the onset of the PLG deposition at 5.97 Ma (Meilijson et al., 2018). A box
model developed by Simon and Meijer (2017) has recently suggested the possibility of
synchronous deposition of deep-basin halite and shallower deposition of gypsum during
Atlantic connectivity. Indeed, while the MES at the top of the PLG can be traced off-
106 The Messinian Salinity Crisis and methods of study
shore in seismic lines, it splits into several surfaces, not all of which are erosional (Lofi
et al., 2011; Flecker et al., 2015). The cores from the Levant Basin contradict a ‘chaotic’
sedimentation and unconformity scenario for the base of the MSC in the deep basins because
of the consistent recovery of the drilled record, lack of any indication of truncation in the
seismic section, and no sedimentary evidence indicating substantial reworked or transported
deposits in this interval (Meilijson et al., 2018). A coeval initiation of basinal halite and
marginal PLG precipitation would therefore call for a revaluation of the commonly accepted
stratigraphic model for the MSC (CIESM, 2008). Additionally, it would necessitate the
revaluation of global-scale climatic and oceanographic models of the MSC, taking into
account the possibility of an increase in the duration of halite precipitation and the older
age for the initiation of halite deposition (Meilijson et al., 2018).
Stage 3 (~5.55-5.33 Ma) ODP and DSDP holes have recovered material that correlates to
the offshore seismic unit labelled as the ‘Upper Unit’, and indicate that Lago Mare sediments
overlie deep-sea evaporites (Hsü et al., 1973; Roveri et al., 2014b). The correlative relation-
ship between these offshore deposits and the marginal sediments remains unclear, however
(Flecker et al., 2015). To avoid a possible hydrocarbon blowout in the Mediterranean,
only the upper surfaces of the deep evaporites were sampled. With drilling penetration
in evaporites at 80 m or less, 95% of the Messinian salts are unsampled beneath the deep
Mediterranean basin and its continental margins. Deep offshore Messinian successions that
have been cored, such as site 654 in the Tyrrhenian Sea, display a cyclical stacking pattern
in the Upper Unit that could be tentatively correlated with the onshore succession of the UG
(Roveri et al., 2014a). Preliminary strontium isotope measurements for this section imply
that the correlation of the ‘Upper Unit’ with the onshore UG succession may be not correct,
however (Roveri et al., 2014b). In a scenario proposed by Roveri et al. (2014b), the ‘Upper
Unit’ would largely correlate to stage 2 (including onshore RLG and halite) rather than the
previously assumed stage 3. Only the uppermost part of the Upper Unit, often not detectable
from seismic data because of its limited thickness (<50 m), may be considered the offshore
counterpart of stage 3 onshore deposits.
The ‘Zanclean flooding’ at the end of the MSC saw the return of the Mediterranean
basin to fully and stable marine conditions. This substantial palaeoceanographic change
is typically recorded by a sharp lithological and palaeontological boundary, suggesting a
geologically instantaneous event (Roveri et al., 2014a). This event has been used to define
the Zanclean/Pliocene stage GSSP in the Eraclea Minoa section (Sicily; Van Couvering
et al., 2000).
Across stages 1 to 3 of the MSC, questions remain regarding both the contribution of
4.3 Major / trace elements and isotopic tracers 107
continental and marine source waters to Mediterranean brines during precipitation of the
PLG and UG (Flecker and Ellam, 2006; Lugli et al., 2010; Natalicchio et al., 2014), and
the origin of the hydrological change that triggers the appearance of different carbonate,
chlorite and sulfate facies (e.g. the branching selenite facies from the 6th cycle upwards of
the PLG units; Lugli et al., 2010). Indeed, numerical models that simulate the formation
of Messinian evaporites (e.g. Garcia-Castellanos and Villaseñor, 2011; Topper and Meijer,
2013) are currently limited by the lack of reliable geochemical constraints for both the
onshore and offshore evaporitic sequences, and a lack of knowledge about their spatial
dimensions and heterogeneity. An integrated approach that combines external factors that
force Mediterranean hydrology, e.g. tectonics, climate, eustatic sea level change, with
an accurate understanding basin level hydrological variability is critical for developing a
unified theory that explains the formation of the Mediterranean Salt Giant. In order to
address these open questions, we present an overview of current and novel geochemical
methods that may, in conjunction, provide an integrated understanding of the Upper and
Lower Gypsum of the MSC.
4.3 Differentiation of ancient gypsum deposits – major /
trace elements and isotopic tracers
In order to precipitate gypsum in natural environments such as the late Miocene Mediter-
ranean basin, supersaturation with respect to gypsum must be achieved by the concentration
of calcium and sulfate ions within a basin. Supersaturation can occur if evaporation in a
basin exceeds the inflow, thus creating a negative water balance, or by the inflow being
concentrated in calcium and/or sulfate ions (or a combination of both). For marine-sourced
gypsum to form, an evaporative basin is usually restricted from the open ocean by a to-
pographic barrier inducing a negative water balance. Such a barrier can either emerge or
subside (e.g. due to sea level rise/fall or tectonic activity) and maintain a connection to the
open ocean, or the barrier may sporadically restrict open water connections between the
basin area and the ocean (e.g. a salt lagoon) (Ba¸bel and Schreiber, 2014). By definition,
therefore, the majority of ‘marine’ evaporative basins are fundamentally ‘marginal marine’
basins, surrounded by land and thus always under some influence of non-marine sources of
water. Thus, a marginal marine basin can contain fluids that range from strictly marine, to a
mixture of marine and non-marine (hybrid), to brackish (during intervals of wetter climate)
(Ba¸bel and Schreiber, 2014). Methods to interpret and distinguish the geological and hydro-
geochemical character of (i) marine/marginal marine or (ii) continental/non-marine gypsum
108 The Messinian Salinity Crisis and methods of study
deposits are therefore fundamental to the understanding evolution of ancient gypsum for
palaeo-climatic and palaeo-environmental reconstructions.
Environmental tracers provide an important tool for understanding the source, flow and
mixing dynamics of water in both the terrestrial and oceanic realm. Tracers, such as ion
ratios, fluid inclusion analysis, and the isotopic compositions of minor/trace elements and
incorporated water molecules, can be utilised individually or together to reconstruct the
hydrogeochemical processes operating within a system. The degree of sensitivity of an
isotopic geochemical tracer is dependent on both the isotopic and concentration contrast
between oceanic and freshwater inputs. The isotopic compositions of such binary mixtures
are therefore related to their end-member compositions in simple ways that can be expressed
by relatively straightforward algebraic expressions (i.e. mixing lines).
Here I describe the various tracers and the analytical methods employed to measure the
trace element concentrations, isotopic ratios, and fluid inclusion- based salinity estimates
that will be utilised in Chapters 5, 6 and 7. This section describes the preparation of the
gypsum and carbonate samples and the methods developed over the course of this research
to produce data that can be utilised to interpret the evolution of the environment of gypsum
precipitation during the deposition of the Upper and Lower Gypsum units of the MSC.
4.3.1 Gypsum Hydration Water
The measurement of triple oxygen and hydrogen isotopes of GHW provides independent
information that may be used to disentangle the competing effects of fractionation during
evaporation, transport, and precipitation (as discussed in Chapter 1 and in Gázquez et al.,
2018). GHW also provides an excellent marker to record the derivation of H2O in brines;
with SMOW taken as a reference standard (and its oxygen and hydrogen isotopes values
equal to zero), fluids dominated by meteoric waters generally display depleted isotopic
values (Craig, 1961; Dansgaard, 1964). Evaporated seawater brines mixed with meteoric
or other waters may have distinct δ18O, δ17O and δD values due to mixture of H2O with
different isotopic characteristics.
Although the GHW of many freshwater gypsum deposits have lower δ18O, δ17O and δD
values than their marine equivalents, this is not always the case. Freshwater gypsum from
highly evaporative freshwater environments may lead to brines that are enriched in 18O and
2H even above that of marine gypsum. For example, gypsum from the marine salt works of
the Cabo de Gata, SE Spain, display a maximum δ18O value of 8.8h, whereas the GHW of
freshwater gypsum from Lake Estanya, Southern Pyrenees, reach values of 10.6h. While a
distinction between marine and non-marine gypsum is no longer possible based on isotopic
analysis of the GHW alone, triple oxygen isotopes can be used to examine basin recharge
4.3 Major / trace elements and isotopic tracers 109
(e.g. Surma et al., 2018). This provides important information regarding the hydrological
evolution of palaeo-basins.
Methodology: The extraction and analysis of GHW was conducted using the methods
described in Chapter 2.
Sampled gypsum crystals were initially divided into two parts: whole crystals were
preserved for fluid inclusion analysis, while crystals for GHW, trace element and isotope
analysis were ground to a powder and dried for 24 h at 45°C to remove adsorbed water.
Approximately 200 mg of dried powder was used for GHW extraction, while the excess
powder was homogenised and stored trace element and isotope analysis.
4.3.2 Fluid Inclusions Analysis
During the precipitation of gypsum, the crystal growth surface inevitably contains imper-
fections. Typically, continued growth of the crystal will engulf such imperfections, thus
producing a vacuole containing the fluid present at the moment of sealing. The trapped
fluids that are enclosed during the growth of the crystal are termed primary, and contain
a sample of the fluid responsible for the precipitation of the diagenetic phase (Goldstein
and Reynolds, 1994). Primary fluid inclusions in gypsum are identified by a relationship to
growth surfaces, where elongate inclusions in gypsum occur along the (001) plane and more
equidimensional inclusions occur normal to (103) plane (Goldstein and Reynolds, 1994). A
careful distinction must be made between primary and secondary fluid inclusions in gypsum,
where the latter display prismatic to elongated shape up to 100 µm in length. Secondary
inclusions are arranged along the (010) cleavage plane, with the major axis parallel to the
(100) plane (Natalicchio et al., 2014), lack a zonal arrangement, are present on different
planes than primary fluid inclusions, and are unrelated to growth bands of host selenite
crystals (Attia et al., 1995).
Chemical analysis of fluid inclusions in gypsum and other saline minerals is a powerful
tool for the estimation of brine composition (e.g. Knauth and Beeunas, 1986; Lowenstein
et al., 2001; Rigaudier et al., 2011). A second technique, fluid inclusion microthermometry
(FIM), can be used to deduce the salinity of the parent fluid. Fluids that contain dissolved
salts depress the freezing temperature of water and the melting temperature of ice; thus, the
melting temperature of frozen fluid inclusions is related to the salinity of the parent brine. To
date, only a few studies have used this technique on fluid inclusions in modern and ancient
gypsum to determine brine salinities (e.g. Attia et al., 1995; Natalicchio et al., 2014). This
reticence likely stems from concerns about the lack of knowledge regarding secondary fluid
inclusion formation and how interaction with secondary fluids may impact the chemistry
110 The Messinian Salinity Crisis and methods of study
Sea
water
Marine Bri
ne Phase
Flushing Phase
LAGOON
Evaporating Marine
Brine
Brackish Waters
Meteoric Water
Ultrafiltrate Continental Mixing 
Line
Fresh water dissolving evaporites
Continental Brine
Lake
-4
Salinity (gl-1, TDS) 
-2
0
2
4
6
8
-6
50 100 150 200 250 300 350
δ
1
8
O
 (
‰
)
Fig. 4.7. Schematic of changes in salinity and isotopic composition of formation waters in δ18O-salinity
space, illustrating evaporation, continental mixing and freshwater dissolution of evaporites. After
Gat (1996).
of primary inclusions. Equally, FIM is limited to mm-sized selenite crystals where growth
zonation is clearly defined, and thus cannot be performed on microcrystalline gypsum.
A combination of water isotopes deduced from GHW and salinity deduced from gypsum
fluid inclusions provide information on the composition of formation waters and fluid beha-
viour during evaporation. Together, changes in the salinity and water isotopic composition
allow derivation of the processes of evaporation, continental water mixing, and freshwater
dissolution/reprecipitation cycles (Fig. 4.7). Additionally, coupling of gypsum water iso-
topes and salinity of fluid inclusions in the same samples can be used to test if they record
the composition of the mother fluid from which gypsum was precipitated (Chapter 5).
Methodology: The method for microthermometric analysis closely followed that described
by Attia et al. (1995). Thin (<1 mm) samples of gypsum were obtained by cleaving the
mineral along 010 planes using a razor blade. The fragments were placed in a Linkam
THMSG600 heating-freezing stage attached to a Zeiss Axio Scope.A1 microscope equipped
with a 100× objective.
Primary inclusions in modern and Messinan gypsum crystals are all single phase, con-
4.3 Major / trace elements and isotopic tracers 111
taining only liquid. By freezing the inclusions to −70°C and then heating to 30°C at a
rate of 20°C min-1, vapour bubbles were formed due to the stretching of the inclusion
during ice formation. The first melting temperature of fluid inclusions and the final melting
temperature of ice (Tmice) were observed by cooling to −70°C and subsequently heating
at a rate of 5°C per minute to −50°C, 2°C per minute to −15°C and at 0.3°C per minute
to 0°C. Ice crystals were identified by their low birefringence, low relief and a rounded
shape near Tmice (Attia et al., 1995). At Tmice, ice crystals decreased in size until no
longer visible. Tmice was recorded to within 0.1°C and reproduced at least twice for each
inclusion with an error of ±0.2°C. Salinities of the Messinian and modern fluid inclusions
were calculated from the final melting temperatures of ice and expressed as both wt% NaCl
equivalent (Bodnar, 1993) and parts per thousand (Goldstein and Reynolds, 1994).
4.3.3 Sulfur isotopes and oxygen isotopes of sulfate
Sulfur isotopes have previously been used to study the origin of gypsum and other sulfur
containing minerals in sedimentary systems (Fontes et al., 1973; Longinelli, 1979; Pierre
and Rouchy, 1990; Lu et al., 2001). Modern seawater has a δ 34S of ~21h and a δ18O
of ~8h (Turchyn and Schrag, 2004). Gypsum precipitates with a negligible sulfur isotope
fractionation, but with as much as a 3h offset for oxygen isotopes (Lloyd, 1968). The
oxygen isotope fractionation factor remains poorly constrained, however, and is often con-
sidered negligible (e.g. Torfstein and Turchyn, 2017). The oxygen isotope composition of
the precipitating sulfate depends on the variable contribution of oxygen atoms from water
and from atmospheric oxygen (Calmels et al., 2007). Abiotic oxygen isotope exchange
between aqueous sulfate and water does not occur over geologically relevant timescales
(Lloyd, 1968; Rennie and Turchyn, 2014). If the sulfate from seawater trapped in a basin
was quantitatively precipitated, then, by mass conservation, the δ 34S and δ 18OSO4 of the
bulk precipitated gypsum should reflect that of seawater.
The multiple valence states of sulfur results in its frequent use in biological processes,
and the sulfur isotopic composition is often utilised as a primary tracer of the redox dy-
namics in lacustrine environments both in the modern environment and over geological
time (Claypool et al., 1980; Canfield et al., 2004 and references therein). The microbially-
mediated reduction of sulfate to sulfide, termed ‘microbial sulfate reduction’, involves a
sulfur isotopic fractionation where the light isotope (32S) is preferentially distilled into the
product sulfide relative to the heavier isotope (34S) (Torfstein and Turchyn, 2017). Microbial
sulfate reduction coupled to organic carbon oxidation is described through the simplified
equation:
2CH2O+SO42− −−→ 2HCO3−+H2S
112 The Messinian Salinity Crisis and methods of study
The rate at which the 34S/32S ratio in the residual sulfate increases is mainly controlled
by factors such as the rate of microbial sulfate reduction, temperature, sulfate concentration,
and organic carbon supply (e.g. Torfstein and Turchyn, 2017). Microbial sulfate reduc-
tion also fractionates the sulfate-bound oxygen isotopes, whereby the residual sulfate is
preferentially enriched in the heavy isotope (18O). The relative change in the sulfur and
oxygen isotope composition of sulfate over the onset of microbial sulfate reduction has
been shown to reflect the degree of recycling of sulfur within cells (over the amount which is
unidirectionally reduced). The δ 18OSO4 therefore reflects the latest oxidation pathway of the
reduced sulfur, plus any equilibration with water along the oxidation or reduction pathway
(Torfstein and Turchyn, 2017). Thus, the oxygen isotope ratio of sulfate has become a
powerful tool for exploring sedimentary and aqueous sulfur redox dynamics.
The δ 34S and δ 18OSO4 isotopes also provide a possible means of distinguishing whether
gypsum was deposited in a marine, non-marine, or hybrid brine. The δ 34S and δ 18OSO4 of
marine dissolved sulfate in reservoirs are globally uniform (20h ±0.5h for δ 34S; 9.5h
±0.5h for δ 18OSO4) without significant change from the Tertiary to present (Thode and
Monster, 1965; Lloyd, 1968; Turchyn and Schrag, 2004). Thus, the sulfur isotopic compos-
ition of aqueous sulfate (and therefore the gypsum that precipitates) in a marine system in-
fluenced by continental waters reflects the balance between runoff (generally with depleted
values compared to seawater) and the amount of water column microbial sulfate reduction,
which is controlled by the stability and length of water column stratification leading to deep-
basin anoxia (Torfstein and Turchyn, 2017). The oxygen isotopic composition of sulfate
is controlled by the same processes, but is particularly sensitive to the incorporation of
oxygen atoms during sulfide oxidation, and the degree of intracellular equilibration between
intermediate valence state sulfur species and water (Torfstein and Turchyn, 2017). Sulfate
concentrations in most terrestrial waters are low, with most rivers containing three orders of
magnitude less sulfate than seawater (Lu and Meyers, 2003). Because of the much higher
sulfate concentration in marine brines than in freshwater, the δ 34S and δ 18OSO4 of marine
gypsum are relatively insensitive to non-marine contributions.
Methodology: Powdered gypsum was initially dissolved in deionised water, and a super-
saturated barium chloride solution (~30 g l-1) was subsequently added to the solution to
precipitate the aqueous sulfate from gypsum dissolution as barite (BaSO4). Prior to analysis
the barite was cleaned in 6 N HCl and rinsed in deionised water to purify it from any co-
precipitating minerals. The barite was weighed into tin and silver capsules and analysed
for sulfur and oxygen isotopes, respectively. For oxygen isotopes of sulfate, barite was
pyrolysed at 1450°C in a High Temperature Conversion Elemental analyser (TC/EA), and
4.3 Major / trace elements and isotopic tracers 113
the resulting carbon monoxide (CO) was measured by continuous flow IRMS with a Ther-
moFisher Delta V Plus. For sulfate isotope analysis, the barite was combusted at 1030°C
in a Flash Element analyser (EA), and the resulting sulfur dioxide (SO2) was measured
by continuous flow IRMS (Delta V Plus). Samples were run in triplicate for δ 18OSO4 and
the average and standard deviation of these replicate analyses are presented. Samples were
bracketed by NBS127, IAEA-SO5 and IAEA-SO6, international barite standards, which
were corrected to δ 34S of 20.3h, 0.5h and -34.1h, respectively (1σ = 0.2h). δ 18OSO4
was bracketed by NBS127, IAEA-SO5 and IAEA-SO6 which were corrected to 8.6h,
12.1h and -11.4h, respectively, and reported uncertainties reflect the average and standard
deviation of triplicate analyses of each δ 18OSO4 sample (1σ < 0.5h).
4.3.4 Strontium Isotopes
Strontium isotopes provide a natural fingerprint of rock-water interactions and have been
widely utilised in studies of weathering and hydrology (e.g. Hodell et al., 1989; Muller et al.,
1990; Müller and Mueller, 1991; Palmer and Edmond, 1992; McArthur et al., 2001; Flecker
and Ellam, 2006). Many biogenic carbonate, phosphate and evaporitic minerals (including
gypsum) incorporate strontium isotopes during precipitation without fractionation at low
temperatures or modification due to biotic processes. Thus, unless post-depositional diagen-
esis has altered the primary strontium isotopic signature, measurement of these materials can
provide a record of the strontium isotopic composition of the fluids from which they formed.
The abundance of 87Sr is expressed relative to the abundance of stable 86Sr (i.e. 87Sr/86Sr).
87Sr/86Sr is commonly measured by multiple-collector inductively-coupled plasma mass
spectrometry (MC-ICP-MS) or thermal ionisation mass spectrometry (TIMS). TIMS in-
strumental mass fractionation effects are overcome by linear normalisation of measured
isotopic ratios to constant values of stable isotope pairs; by convention, 87Sr/86Sr ratios are
normalised to 86Sr/88Sr = 0.1194 (Elderfield, 1986).
Heterogeneity in 87Sr/86Sr ratios of dissolved strontium across geologically diverse en-
vironments provides a useful tool quantifying end-member mixing processes over time. At
the global scale, temporal variation in the 87Sr/86Sr ratio of seawater occurs due to the
variable input of strontium derived from two main sources: (i) the continental crust and
(ii) the upper mantle. Continental crust supplies dissolved strontium to the oceans domin-
antly through the chemical weathering of old silicate lithologies, and thus provide waters
with high 87Sr/86Sr ratios (>0.720), but with low strontium concentrations. Upper mantle
sources supply strontium principally through hydrothermal activity at mid-oceanic ridges,
submarine weathering of basalts and chemical weathering of continental basalts (Elderfield,
1986). Weathering of such atypical continental crust result in low 87Sr/86Sr ratios (0.706-
114 The Messinian Salinity Crisis and methods of study
seawater       
Fig. 4.8. Schematic of calcium and strontium cycling in the modern surface environment, with sources
labelled in bold and sinks italicised. Isotopic data from Fantle and Tipper (2014) and Palmer and
Edmond (1992). For calcium and strontium, major sources include riverine calcium sourced from
continental weathering and hydrothermal fluids at mid-ocean ridge systems. CaCO3 formation and
burial is the major sink. Modified after Husson et al. (2015).
0.709) but high strontium concentrations (up to 1000 ppm) (Palmer and Edmond, 1992).
The residence time of strontium is significantly longer than the mixing time of the ocean
(~10,000 years), and thus the seawater 87Sr/86Sr ratio is the homogenous at any one point in
time, resulting in a single temporal strontium isotope record (Elderfield, 1986).
At the local scale, the 87Sr/86Sr ratio of water in a semi-enclosed basin connected to the
global ocean is controlled by a simple mixing between seawater and other strontium-bearing
water sources feeding the basin (e.g. river run-off and groundwater) (Flecker and Ellam,
2006) (Fig. 4.8). Importantly, average global river waters have strontium concentrations
~40 times lower than that of seawater (Palmer and Edmond, 1992).
In the Mediterranean today, the 87Sr/86Sr values of the Rhone (~0.70872) and the Nile
(~0.70600) are lower than the global seawater average, and make up >80% of the total Medi-
terranean riverine input (McArthur et al., 2001; Flecker et al., 2002; Topper et al., 2014).
During the MSC, river input was likely largely made up by North African rivers that are
dry today (Simon et al., 2017). On the marginal basin scale, the strontium characteristics of
the river input are also dependent on locally outcropping rocks in small catchments (Topper
4.3 Major / trace elements and isotopic tracers 115
et al., 2014). While previous studies conducted on marginal Mediterranean basins, such
as the Sorbas Basin (SE Spain), have inferred a higher 87Sr/86Sr (~0.70910) riverine value
than that of coeval ocean water values (~0.70900) on the basis of 87Sr/86Sr measurements of
ostracods from the lacustrine Zorreras Member (located two units above gypsum deposited
during the MSC; Modestou et al., 2017) and secondary gypsum veins (Lu and Meyers,
2003), only ~0.1% of the river discharge into the Mediterranean flows directly into marginal
basins (Topper et al., 2014). The Mediterranean freshwater budget was therefore dominated
by rivers with low 87Sr/86Sr.
Methodology: Contamination and its potential effect on strontium (and calcium) isotope
signatures was tested in our samples by leaching gypsum in three steps under clean laborat-
ory conditions. Lu et al. (1997) demonstrated that a leaching process involving treating the
sample with alcohol, deionised water, and HNO3 washes significantly reduced certain ele-
mental contaminations. Here, samples were washed in methanol to mechanically separate
clays, followed by two deionised water washes to remove ions liberated from fluid inclu-
sions during the grinding process. This was followed with three washes in 10% ammonium
hydroxide to remove adsorbed ions from any remaining clays and three washes with 0.001 M
HCl to dissolve any carbonates present. Finally, the sample was exposed to three washes of
deionised water to remove any traces of the leaching reagents. Approximately 30 samples
that were treated via this process were compared to gypsum samples directly dissolved
in deionised water. The concentrations of contaminant elements (such as aluminium and
sodium) were very low (below detection) in both sample sets, and isotopic measurements
made on both leached and non-leached samples were all within analytical error.
An aliquot of sample solution containing approximately 300 ng of strontium was dried
down and dissolved in 200 µl 3 M HNO3. Strontium was separated using Eichrom Sr Spec
resin with 100 µm to 150 µm mesh particle size in clean lab conditions. 87Sr/86Sr measure-
ments were made by TIMS (Thermo-Scientific Triton Plus MC-TIMS). The separated Sr
was dried down, refluxed in 2 N HNO3 and subsequently dried down once again. Samples
were then loaded onto degassed single Re filaments together with 1 µl of tantalum phosphate
activator. A current of ~0.8 A was continuously applied during the loading. The evaporation
filament is heated manually until a stable signal between 4 V and 6 V of 88Sr on a 1011Ω
resistor is reached. Data acquisition was comprised of 10 blocks of 20 measurements with
a ~8 s integration time in static mode. Results were normalised to 88Sr/86Sr 0.1194 with an
exponential fractionation correction. Runs were bracketed with the NBS 987 standard. The
11 analyses of NBS 987 during the course of this study gave a mean value of 0.710253 (2σ
= 0.000007). Blanks were <350 pg and negligible for the Sr concentration of these samples.
116 The Messinian Salinity Crisis and methods of study
A small number of samples (n = 6) were analysed with a VG Sector 54 TIMS. In this
case, separated strontium was dried down, refluxed in HCl to ensure crystallisation in a
chloride form, and subsequently dried down once again. The resultant crystals were loaded
and analysed on a VG Sector TIMS to determine the 87Sr/86Sr ratio. Runs were bracketed
with an internal NBS 987 standard. NBS 987 gave 0.710262 (2σ = 0.000016) on 45 separate
measurements over the last two years. Procedural blanks for strontium were negligible.
4.3.5 Calcium Isotopes
Calcium is one of the major dissolved components of seawater (~400 ppm) and the most
abundant cation in rivers (~20 ppm, but may constitute up to ~100 ppm in carbonate catch-
ments; Fantle and Tipper, 2014). The high abundance of calcium is related to the stability of
its nucleus, forming six naturally occurring stable or quasi-stable isotopes, 40Ca (96.94%),
42Ca (0.65%), 43Ca (0.14%), 44Ca (2.09%), 46Ca (0.01%) and 48Ca (0.19%), where 48Ca is
treated as stable because of its long half-life of 6×1018 years (Fantle and Tipper, 2014). The
isotopic composition of calcium in minerals may vary over long timescales as 40Ca can be
produced via natural β decay of 40K (λ1/2 = 1.27×109) (Marshall and DePaolo, 1982). For
most samples, enrichments in 40Ca due to radiogenic decay of 40K will be minimal owing to
its long half-life and usually limited to less than the range of current analytical uncertainty
(DePaolo, 2004).
The stable isotopic composition of Ca is expressed in the δ -notation notation relative
to a normalisation standard using parts per thousand (h) units. There are four reference
standards that have been used to date: CaF2, ‘bulk silicate Earth’ (BSE), seawater, and SRM-
915a (Fantle and Tipper, 2014). Here, δ 44/40Ca values are referenced against BSE, which
gives modern seawater δ 44/40Ca as 0.92h. The 44Ca/40Ca ratio is measured by TIMS,
while the 44Ca/42Ca ratio (expressed as δ44/42Ca) is measured by MC-ICP-MS because
of large interferences on 40Ca from 40Ar+. Assuming 44Ca/40Ca and the 44Ca/42Ca ratios
follow mass-dependent behaviour, a simple correction can be made to compile values into
a single database on a uniform scale (Fantle and Tipper, 2014). Analytical fractionation of
calcium isotopes during measurement via TIMS is corrected for via the addition of enriched
42Ca-48Ca double spike.
Calcium fractionates isotopically during mineral precipitation, where calcite, aragonite
and barite incorporate isotopically light calcium into their structure (e.g. Gussone et al.,
2005, 2003). The degree of this fractionation varies with mineralogy (calcite vs. aragonite;
Gussone et al., 2005), temperature (Gussone et al., 2005), and precipitation rate. Calcium
isotope fractionation during mineral precipitation is believed to be kinetically controlled,
and related to aqueous ion-mineral surface interactions (Fantle and DePaolo, 2007; DePaolo,
4.3 Major / trace elements and isotopic tracers 117
1.0
Plants
River waters
Carbonates
Ocean water
Hydrothermal fluids
Silicates
2.00.0-1.0-2.0-3.0
δ
44/40Ca (‰)
Fig. 4.9. Compilation of calcium isotope data from seawater, hydrothermal fluids, silicates, carbonates,
rivers and plants (after Fantle and Tipper, 2014). Values are referenced against BSE.
2011). The calcium isotopic fractionation associated with gypsum precipitation is close to
equilibrium and unity for some natural systems over long time scales (Fantle and DePaolo,
2007). The fractionation factor close to chemical equilibrium over short experimental time
scales is in the range of ~0.9995 (Harouaka et al., 2014), a value that is in agreement with
predictions based on surface reaction kinetic model of calcite precipitation (e.g. DePaolo,
2011). Thus, during the evaporation of a fluid such as seawater, calcium isotopes should
follow a Rayleigh fractionation path, where the light isotope is preferentially incorporated
into precipitated calcium minerals and the heavy isotope is relegated to the concentrating
brine (e.g. Hensley, 2006).
The sizes of the main surficial calcium reservoirs and their calcium isotope composition
are given in Fig. 4.9, displaying a total range of δ 44/40Ca values of ~4h (Fantle and
Tipper, 2014). Calcium in the modern ocean is isotopically homogenous (0.92 ±0.20h
2σ ), owing to its long residence time of 1 Ma (Fantle and Tipper, 2014). Calcium inputs
to the ocean system are mainly from riverine and hydrothermal sources. Because of mass-
dependent fractionation during terrestrial calcium cycling, interpreting the proportion of
riverine calcium derived from carbonate, silicate, and evaporite lithologies is more complex
than interpretations of radiogenic isotopes (e.g. 87Sr/86Sr), where all process-related mass-
dependent fractionation is removed by normalisation during data reduction (Tipper et al.,
2010, 2008; Fantle and Tipper, 2014) (Fig. 4.8). While there are no coherent spatial trends
in riverine δ 44/40Ca, this can be beneficial to elucidate marine-freshwater mixing profiles,
as riverine inputs to a local system will always drive calcium isotopes towards lower values.
118 The Messinian Salinity Crisis and methods of study
The only quantitatively significant calcium output to the ocean system is CaCO3 precip-
itation and burial (Fantle and Tipper, 2014; Fig. 4.8). CaCO3 minerals are fractionated from
seawater, where 40Ca is preferentially taken up into the solid, leaving seawater enriched
in 44Ca. At steady state, the average δ 44/40Ca composition of the CaCO3 output flux
must equal the δ 44/40Ca value of the input flux (~-0.1h), which is thought to be stable
over geologic timescales (Tipper et al., 2010). Unlike other trace isotopic systems (e.g.
87Sr/86Sr), calcium species are major constituents of carbonates and gypsum, and thus are
not subject to the same concerns regarding diagenesis and isotopic resetting (Husson et al.,
2015).
Methodology: Calcium isotope (44Ca/40Ca) analysis was conducted using a Thermo Sci-
entific Triton Plus MC-TIMS, closely following the method of Bradbury and Turchyn (2018).
Sample aliquots containing 6 µm of calcium were combined with a 42Ca-48Ca double–
spike at a ratio of 10:1 (sample-to-spike) in acid-cleaned Teflon vials. The 48:42 ratio of
the double–spike is 1:1, similar to the optimum ratio of 3:2 for a 42Ca-48Ca double-spike
(Rudge et al., 2009). Dissolved gypsum and carbonate samples were dried down, refluxed
in ultra-pure nitric acid to ensure crystallisation in a nitric form, and then combined with
the double spike. The samples were then dried and re-dissolved in 0.5% nitric acid and
calcium was separated using a Dionex ICS-5000+ IC system coupled with a Dionex AS-AP
fraction collector. The samples were run through a high-capacity carboxylate-functionalised
column (Dionex CS-16) using 30 mM methyl-sulfonic acid (MSA) at a flow rate of 1 ml
min-1. The conductivity of the samples is continuously measured during the separation, and
a minimum peak slope of 0.003 µS/s determines the sample collection during a set time
window (Bradbury and Turchyn, 2018). The procedural blank on the Dionex as determined
independently by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) is
96 ng of calcium, when 7 ml of eluent was collected from the Dionex. During the collection
of 4.4 µg of calcium using the Dionex, this represents ~2% of the collected calcium.
The collected calcium was subsequently dried down before being re-eluted in 1 µl of
2 M nitric acid and 4 µg of calcium was loaded on an outgassed 0.7 mm Re filament with
0.5 µl of 10% trace metal purity phosphoric acid as an activator. A current of ~0.8 A
was continuously applied during the loading. The samples are run using a double filament
method, with the ionisation filament being heated to 1400˚C (as measured on a pyrometer),
and the evaporation filament is heated manually until a stable signal between 5 V and 10 V
of 40Ca on a 1011Ω resistor is reached. The method consists of a dynamic collection routine
(40-44 and 42-48) with an integration time of 8.389 s. 10 blocks of 20 cycles are collected,
with peak centring and baseline being run at the start of the run, and lens focusing occurring
4.3 Major / trace elements and isotopic tracers 119
every 2nd block. Total run-time not including heating is 1.5 h. The average external 2σ
standard deviation over 9 months on the standard NIST915B was 0.10h (mean = -0.28, n
= 82).
4.3.6 Barium Isotopes
Naturally occurring barium is a mix of six stable isotopes, 132Ba (0.10%), 134Ba (2.42%),
135Ba (6.59%), 136Ba (7.85%), 137Ba (11.23%), 138Ba (71.70%) and one long-lived ra-
dioactive primordial isotope, 130Ba (0.11%), was recently identified as being unstable by
geochemical means. Barium has been used as a water-mass tracer of riverine input to the
coastal ocean by directly measuring seawater concentrations (Guay and Falkner, 1997) or
Ba/Ca ratios in corals (McCulloch et al., 2003) and foraminifera (Lea and Boyle, 1989),
but barium isotopes have never been applied to constrain endmember-mixing systems. Ba-
isotope data of samples are reported in the δ -notation (δ 138/134Ba) using parts per thousand
(h) units relative to the Ba standard NIST3104a.
Previous studies have reported Ba isotopic compositions variously as δ 137/134Ba norm-
alised to the Fluka Ba(NO3)2 standard (e.g. Cao et al., 2016) or δ 138/134Ba as normalised
to NIST3104a (e.g. Horner et al., 2015). Horner et al. (2015) and Pretet et al. (2015) have
both demonstrated that the differences between the Fluka and NIST Ba standards in their Ba
isotopic compositions are within the analytical uncertainty. Therefore, one can be simply
converted to the other directly by multiplying by ~1.33 (Hsieh and Henderson, 2017).
Compiled global upper-ocean barium data show relatively constant concentrations (35-
45 nM) in near-surface waters, with the exception of areas near river inputs or strong up-
welling (Hsieh and Henderson, 2017). The relatively uniform barium concentrations in the
upper ocean indicate that barium removal is slow relative to supply and mixing, implying
that near-surface Ba isotope values are controlled by basin-scale balances rather than by re-
gional or short-term processes (Hsieh and Henderson, 2017). Measured seawater δ 138/134Ba
values vary from 0.24h to 0.65h (Hsieh and Henderson, 2017; Bridgestock et al., 2018).
Riverine input of barium to the surface ocean is the major external source, with potential
to influence surface-water compositions near the input source, and to control the global
barium isotope composition of the oceans on long timescales (Hsieh and Henderson, 2017).
Cao et al. (2016) measured barium in eight rivers, recording low δ 138/134Ba values (relative
to typical seawater) between -0.1h and 0.4h. Importantly, barium concentrations in rivers
range from 100 nM to 500 nM (Cao et al. 2016), which is higher than seawater (30 nM to
170 nM). The combination of distinct marine and non-marine isotopic signatures and higher
barium concentrations in riverine waters suggest that barium isotopes could potentially
provide a powerful tracer for riverine inputs in marginal oceanic settings.
120 The Messinian Salinity Crisis and methods of study
Methodology: Barium isotope (138Ba/134Ba) analysis was conducted using a Thermo Sci-
entific Triton Plus MC-TIMS at the University of Oxford by Luke Bridgestock. Between
0.1 g and 0.4 g of powdered gypsum was leached in deionised water (~4 ml water, per
1 g of gypsum) over a period of 24 h with regular treatments in an ultrasonic bath. The
supernatant was separated from the residual solid material and evaporated to dryness, before
re-dissolving in 10 ml to 15 ml of 7.5 M HNO3. Samples contained between 25 ng and
1283 ng of Ba. Samples were equilibrated with appropriate quantities of a Ba double spike
solution (Hsieh and Henderson, 2017) prior to purification by a 2 stage cation exchange
chromatography procedure (adapted from Bridgestock et al., 2018), utilising 2 ml of AG50-
X8 resin (200 µm to 400 µm mesh size) (Table 4.1).
Spiked sample solutions were evaporated to dryness and re-dissolved in 5 ml to 15 ml of
3 M HCl, depending on the quantity of solid residue. To maximise the separation efficiency
of Ba from Ca and Sr, sample solutions were split into two aliquots (2.5 ml to 7.5 ml) that
were processed through the cation exchange chromatography procedure separately (Table
4.1; stages 1A and 1B). The volume of 3 M HCl used to elute the matrix was adjusted based
on the volume sample solution loaded onto the column to give a total volume of 18 ml.
The Ba elutions from stages 1A and 1B were then recombined and processed through the
chromatography procedure again (Table 4.1; stage 2). Organics leached from the cation
exchange resin were oxidised by the sequential addition and evaporation of 60 µl 15 M
HNO3 and 9.8 M H2O2. The procedural blank was assessed to be <0.48 ng of Ba (n = 2),
representing <1.95% of the sample Ba processed though the procedure.
Ba isotope measurements were made using TIMS featuring a double Re filament as-
sembly. The purified Ba was loaded onto the evaporation filament in 2 M HCl without an
activator gel. Ba ionisation was achieved by applying currents of 2000 mA to 2500 mA
to the ionisation filament and 50 mA to 300 mA to the evaporation filament. Ion beam
intensities of between 4 V to 12 V were achieved for 138Ba. Ba isotope ratios were measured
following the protocol described by Hsieh and Henderson (2017).
Isotopic analyses of the Ba standard (NIST3104a) were conducted every measurement
session, alongside analyses of the samples. The long-term reproducibility of δ 138/134Ba
values determined for NIST3104a was between ±0.02h and ±0.03h (2σ ) over a period of
3 months. This level of agreement (±0.03h) is taken to represent the analytical uncertainty
of the sample analyses (Bridgestock et al., 2018). A higher level of uncertainty (±0.05h) is
assigned for the samples ‘BdI C15 5 cm’ and ‘EM4 Dome 30 cm’ due to the lower ion beam
intensities (~4 V for 138Ba) achieved during TIMS analyses. A single measurement of the
barite standard NBS-127 yielded δ 138/134Ba = -0.30 ± 0.03h, in agreement with the values
published by Horner et al. (2017) (δ 138/134Ba = -0.27 ± 0.02h; 2σ , n = 4).
4.3 Major / trace elements and isotopic tracers 121
Procedure Stages 1A and 1B Stage 2
1. Clean resin 20 ml H2O & 20 ml 6 M HCl 20 ml H2O & 20 ml 6 M HCl
2. Condition resin 20 ml 3 M HCl 20 ml 3 M HCl
3. Load sample 2.5-7.5 ml 3 M HCl 1 ml 3 M HCl
4. Elute matrix 11.5-15.5 ml 3 M HCl 18 ml 3 M HCl
5. Elute Ba 10 ml 4 M HNO3 10 ml 4 M HNO3
Table 4.1. Cation exchange chromatography procedure employed for purification of Ba from gypsum
samples, conducted using AG50-X8 resin (200-400 mesh) and Teflon shrink fit columns with
a resin bed volume of 2 ml and aspect ratio (diameter: length) of 0.05.
4.3.7 Trace element chemistry
Major and trace element concentrations of gypsum and carbonate samples were determ-
ined by ICP-OES after dilution in 0.1 M HNO3 to a constant calcium concentration of
between 10 ppm and 50 ppm. Samples were run on a Varian Vista Axial ICP-OES using
the 315.887 nm Ca emission line. In-house calibration standards were prepared to closely
match the concentration matrix of the media solutions.
Trace element chemistry was also determined using a Thermo Element XR, a single
collector sector field high-resolution inductively-coupled plasma mass spectrometry (HR-
ICP-MS), using the process described by Misra et al. (2014). An aliquot of each sample
solution was taken and diluted with sufficient 0.1 M HNO3 and 0.5% HF matrix such that
the calcium concentration was 10 ppm and the total volume of the resultant solution was
300 µl.
4.3.8 Carbonate sample preparation
Carbonate samples were prepared for isotope analysis by treating the samples with meth-
anol, ammonia and water washes, each performed in triplicate. Unlike the leaching process
for gypsum, the acid wash was omitted for carbonate samples to prevent the dissolution of
calcite, which could affect the dolomite-calcite ratio of the sample and potentially cause
isotopic fractionation. After leaching, the samples still possessed a small percentage of
contaminants (e.g. clays). To isolate the carbonate phase and prevent the samples from
being excessively contaminated during dissolution, the solid sample was dried, weighed
and placed in a volume of 0.1 M HCl containing 110% of the acid required for complete
dissolution of the carbonate, calculated to avoid elemental variation generated by incom-
plete dissolution. Samples that were not treated via this process displayed significant clay
contamination and isotope values varied by more than 2σ in some cases.
122 The Messinian Salinity Crisis and methods of study
4.3.9 X-ray diffraction (XRD) of carbonates
XRD analysis was conducted on carbonates primarily to estimate the relative proportion of
calcite to dolomite in the sample. Samples were ground to <5 µm particles, slurry mounted
onto glass slides with acetone, and analysed with a D8 Bruker diffractometer equipped with
primary Gobbel mirrors for parallel Cu Ka X-rays and a Vantec position sensitive linear
detector. Collection conditions were: 10° to 70° in 2θ , 0.02 step size, 450 s/step, divergence
slit 0.2 mm. Rietveld refinements were performed with software Topas 4.0 (Cohelo, 2007).
Crystal structures of all phases were retrieved from the inorganic crystal structure database
(ICSD) (Allmann and Hinek, 2007).
Rietveld quantitative analysis is known to be unreliable for minor phases (<5 wt %). The
accuracy is considered to be ±1-2% relative for major phases. As the estimated standard
deviations from the Rietveld calculation are related to the mathematical fit of the model,
they have no bearing on the accuracy or quantification itself (Madsen and Scarlett, 2008).
4.3.10 Total inorganic carbon, δ18Ocarb and δ13C of bulk carbonates
Carbonate samples were digested in 50% H3PO4 for 20 h at 70°C and total inorganic car-
bon was measured using an AutoMateFX autosampler coupled to a UIC (Coulometrics)
5011 CO2 coulometer (Engleman et al., 1985). Analytical precision was estimated by
analysis of reagent-grade CaCO3 (100%) and MgCa(CO3)2 (100%) and yielded a mean and
1σstandard deviation of 99.5 ±0.4% and 94.6 ±0.8% for calcite and dolomite, respectively.
The dolomite is likely underestimated because of incomplete reaction even after 20 h. Wt%
calcite and dolomite of samples were calculated using the dolomite/calcite ratio obtained by
XRD multiplied by the total inorganic carbon measured by coulometric titration. Results
are expressed as wt% CaCO3 and wt% MgCa(CO3)2 assuming typical stoichiometry.
For stable isotope analysis, bulk sediment samples were ground to a fine powder, flushed
with CP grade helium then acidified with 104% H3PO4 and left to react for 1 h at 70°C.
Stable carbon and oxygen isotopes of carbonate were measured using a ThermoScientific
GasBench II, equipped with a CTC autosampler coupled to a MAT253 mass spectrometer
(Spötl and Vennemann, 2003). Analytical precision was estimated at ±0.08h for δ13C
and ±0.1h for δ18Ocarb by repeated analysis of the Carrara Marble standard. Results are
reported relative to the VPDB.
4.4 Summary 123
4.3.11 Scanning electron microscopy (SEM)
SEM analyses of carbonate samples were performed with a Zeiss Supra 50 VP equipped
with an energy dispersive X-ray detector for element analysis (EDX). SEM images and
EDX analysis were obtained with a secondary electron detector, a working distance of
9 mm to 10 mm, and an accelerating voltage of 15 kV. In order to reduce the possibility
of misinterpreting recent microbial contamination with fossilised biological structure, only
the surface of freshly broken samples is investigated. Prior to analysis, the samples were
coated with 10 nm of platinum. SEM analysis was carried out by T.R.R. Bontognali at ETH,
Zurich.
4.4 Summary
Trace elements, isotopes, and salinity measurements deduced from fluid inclusions provide
a useful means to determine the origin of gypsum precipitated during the MSC. Most
information on MSC brine chemistry to date has been inferred from the analysis of gypsum
using traditional tracers such as strontium concentrations, 87Sr/86Sr, δ 34S and δ 18OSO4 .
Taking into account secular variation of oceanic 87Sr/86Sr and δ 34S during the Phanerozoic
(McArthur et al., 2001), 87Sr/86Sr and δ 34S of gypsum that are close to expected oceanic
values supports a marine origin (e.g. Cendón et al., 2004). This criterion is particularly
useful where the isotopic composition of weathered continental material is markedly dif-
ferent from the expected oceanic value (Rossi et al., 2015). This is the case in many
Tertiary circum-Mediterranean basins, for example, where continental sulfate is mainly
recycled from Triassic marine evaporites, thus being markedly different in 87Sr/86Sr and
δ 34S from marine Tertiary sulfate (e.g. Rossi et al., 2015). 87Sr/86Sr, δ 34S and δ 18OSO4
isotopes of gypsum are, however, insensitive to minor non-marine contributions due to the
large concentration differences between marine and continental fluids, and consequently
near-marine 87Sr/86Sr, δ 34S and δ 18OSO4 isotopes do not necessarily indicate pure marine
depositional environments.
Coupling traditional tracers to novel tracers such as δ 44/40Ca, δ 138/134Ba and triple
oxygen and hydrogen isotopes of GHW can provide additional constraints on the relative
continental contribution, redox processes and reservoir effects influencing the deposition of
gypsum during the MSC. Increasing the number of isotopic tracers employed with differing
endmember mixing sensitivities delimits the number of endmember sources that can be
partitioned. By directly comparing trace elements, isotopes, and salinities deduced from
fluid inclusions in the same sample, it is possible to test to see if all tracers agree as to the
origin of the sample, or whether certain systems have undergone post-depositional resetting.
124 The Messinian Salinity Crisis and methods of study
Importantly, this method provides a significant advance over previous attempts to constrain
the hydrological budget of the MSC as continental contributions, redox processes and reser-
voir effects can be individually examined in each sample, and subsequently deconvolved on
both a local and regional scale.
Chapter 5
Coupled measurements of δ18O and δD
of hydration water and salinity of fluid
inclusions in gypsum from the Messinian
Yesares Member, Sorbas Basin (SE
Spain)
Adapted from the publication: Nicholas P. Evans, Alexandra V. Turchyn, Fernando Gázquez,
Tomaso RR Bontognali, Hazel J. Chapman, and David A. Hodell. (2015). Earth and
Planetary Science Letters, 430, 499-510.
One cycle (cycle 10) of gypsum–marl deposition from the Messinian Yesares Member
in Sorbas Basin, Spain was studied with the objective of reconstructing the changing envir-
onment of deposition and its relation to astronomically-forced climate change. The δ18O
and δD of gypsum hydration water (CaSO4 ·2H2O) and salinity of fluid inclusions were
measured in the same samples to test if they record the composition of the mother fluid from
which gypsum was precipitated. Water isotopes are highly correlated with fluid inclusion
salinity suggesting the hydration water has not exchanged after formation. The relatively
low water isotope values and fluid inclusion salinities indicate a significant influence of met-
eoric water, whereas δ 34S, δ 18OSO4 and
87Sr/86Sr support a dominantly marine origin for
the gypsum deposits. The discrepancy between water and elemental isotope signatures can
be reconciled if meteoric water dissolved previously deposited marine sulfates supplying
calcium and sulfate ions to the basin which maintained gypsum saturation. This recycling
126 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
process accounts for the marine δ 34S, δ 18OSO4 and
87Sr/86Sr signatures, whereas the low
δ18O and δD values of gypsum hydration water and fluid inclusion salinities reflect the
influence of freshwater.
The cyclic deposition of gypsum and marl in the Yesares Member has previously been
interpreted to reflect changing climate related to Earth’s precession cycle. The δ18O, δD and
salinity of the parent brine increased from low values at the base of the cycle to a maximum
in the massive gypsum palisade, and decreased again to lower values in the supercones at
the top of the cycle. This pattern, together with changes in mineralogy (calcite–dolomite–
gypsum), is consistent with a precession-driven change in climate with wettest conditions
(summer insolation maxima) associated with the base of the calcium carbonate marls and
driest conditions (summer insolation minima) during formation of the gypsum palisade.
5.1 Introduction
During the Messinian salinity crisis (MSC), the Mediterranean Sea was transformed into
a giant brine pool where more than one million cubic kilometres of salt was deposited in
~640 ka (Krijgsman et al., 1999; Rouchy and Caruso, 2006). Most of the deposits in the deep
basins have remained unsampled thus studies have relied upon onshore Messinian deposits
in marginal basins to infer the stratigraphic history of the MSC (Roveri et al., 2014a) (Fig.
5.1).
In marginal basins of the circum-Mediterranean, the lowermost gypsum bed marking
the onset of the MSC is dated at ~5.97 Ma (Krijgsman et al., 1999; Manzi et al., 2013).
The overlying unit known as the Primary Lower Gypsum (PLG) comprises up to 16 beds of
massive gypsum interbedded with finely laminated marls or shale (Roveri et al., 2009; Lugli
et al., 2010). The deposition of these evaporite-marl cycles is reported to be controlled by
the variations of Earth’s orbital parameters, mainly the ~21 kyr precession cycle (Krijgsman
et al., 2001, 1999; Manzi et al., 2009; Lugli et al., 2010). According to this interpretation,
evaporite deposition occurred at the precession maxima (i.e. minimum summer insolation
when summer solstice and aphelion coincided) when evaporation exceeded precipitation
during periods of dry climate and marl deposition occurred at precession minima (maximum
summer insolation when summer solstice and perihelion coincided) when rainfall increased
(Krijgsman et al., 2001, 1999).
Whereas orbital forcing may explain the observed large-scale lithological cyclicity, the
composition and salinity of the brines from which large volumes of sulfate minerals formed
are still debated. Most information on brine chemistry has been inferred using strontium
(87Sr/86Sr), sulfur (δ 34S), and oxygen isotopes in sulfate (δ 18OSO4) (e.g. Longinelli, 1979;
5.1 Introduction 127
Messinian Reef
Messinian Gypsum
Neogene and 
Quaternary 
Neogene Volcanic 
Rocks
Betic Basement 
Sierra de Gador
Sierra de los Filabres
Sie
rra
 C
ab
re
ra
Sier
ra A
lham
illa
Camp
o de 
Dalia
s
ALMERIA
CABO DE GATA
Si
er
ra
 d
el
 C
ab
o 
de
 G
at
a
Se
rra
ta
 Fa
ult
Ve
ra
 B
as
in
10 km
Messinian Evaporites DSDP-ODP Sites
Gypsum/Halite/Gypsum "Trilogy" 
Undifferentiated Gypsum-Halite
Only Gypsum 
Evaporites Recovered 
No Evaporites Recovered
45°
40°
35°
10° 20°
Orogenic front 
Passive margin 
Transcurrent fault 400 km 
IBERIA
AFRICA
AEGEAN 
SEA
ASIA
SORBAS 
BASIN
ATLANTIC 
OCEAN
SOR
BAS 
BASI
N
Ni
jar
 B
as
in
121 976 977 978
122
123 656
133
974
373
127
377
126
125
973
966
965
967
968129
121
975
372
124975
371
134
654
132
654 652
374 378
375
376
A
B
Fig. 5.1. (A) Location map of the Sorbas Basin and surrounding areas modified after Krijgsman et al. (2001).
Indicated are the distributions of the outcropping marginal Messinian reefs (Cantera Member) and
gypsum deposits (Yesares Member). (B) Map of the Messinian evaporites in the Mediterranean
modified after Rouchy and Caruso (2006) and Roveri et al. (2014a). Also shown is the location of
the DSDP-ODP boreholes in which Messinian evaporites were recovered.
Müller and Mueller, 1991; Lu and Meyers, 2003; Lugli et al., 2010). Here the traditional
isotope tracers are combined with the measurement of GHW and fluid inclusions to recon-
struct the chemical composition of the parent brine solution.
The measurement of oxygen and hydrogen isotopes in GHW is a potentially powerful
tool for studying the nature of the parent water from which gypsum was precipitated (Sofer,
1978; Longinelli, 1979; Hodell et al., 2012). An assumption of the method is that once
formed, gypsum hydration water retains its isotope composition and does not undergo
postdepositional isotopic exchange (Sofer, 1978). To test this assumption, we compare the
isotope composition of the hydration with the salinity of fluid inclusions in gypsum estim-
ated by microthermometry (Goldstein and Reynolds, 1994; Attia et al., 1995; Natalicchio
et al., 2014). We expect the δ18O and δD of hydration water to be well correlated with
the salinity of primary fluid inclusions if it has retained the isotopic signature of the mother
water.
In addition to 87Sr/86Sr, δ 34S, and δ 18OSO4 isotope analysis of gypsum, oxygen and
hydrogen isotopes in gypsum hydration water and salinity of fluid inclusions were measured
in gypsum from the sixth cycle of gypsum deposition from the Río de Aguas section, Sorbas
Basin. We also measured modern gypsum deposits from the Cabo de Gata Salina, Almería,
128 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
Spain, to provide a comparison to Messinian gypsum. We demonstrate that the hydration
water in Messinian gypsum of Sorbas Basin has retained its original isotopic composition
and can be used to infer paleoenvironmental conditions of gypsum deposition. We also show
a well-defined cyclicity in the mineralogy and δ18O and δD of gypsum hydration water
and fluid inclusion salinity over a single gypsum cycle that is consistent with astronomical
forcing of Messinian climate.
5.2 Geological setting
The Neogene Sorbas Basin in SE Spain is an elongated intra-montane depression surroun-
ded by basement highs of the Betic Internal Zone (Fig. 5.1). The deposition of evaporites
begins with gypsum/marl cycles of the basal Yesares Member. These evaporites consist of
gypsum deposits that are interbedded with laminated marls. The Yesares Member consists
of a total of 15 gypsum cycles and is best exposed in the Río de Aguas section in the Sorbas
Basin (Krijgsman et al., 2001; Roveri et al., 2009). The facies description for cycle 10 is
given in Section 4.2.1.
5.3 Methods
5.3.1 Field locations and sample collection methods
We sampled gypsum cycle 10 that is exposed at the base of the Río de Aguas section
(37°05’23.2” N 002°06’54.2” W) in Sorbas Basin, SE Spain (Dronkert, 1985; Krijgsman et
al., 2001). Samples were collected in April 2013 and March 2014. The zero depth reference
was taken as the horizontal base of the palisade gypsum layer; therefore, gypsum crystals
sampled in nucleation cones have a negative depth reference. Marl samples were obtained
from the base of the section and from marls juxtaposed to the supercones.
5.3.2 Geochemical methodology
All geochemical methodology is given in Section 4.3.
5.4 Results 129
20.0
δ34S (‰)
6.0
8.0
10.0
12.0
14.0
16.0
21.0 22.0 23.0 24.0
PLG Cycle 10
Modern Salina Gyspum 
δ18
O
S
O
4
 (
‰
)
Messinian 
marine field Oxygen isotope 
exchange with 
water 
Py
rit
e o
xid
at
ion
 
in 
riv
er
s 
Sulfate
reduction 
Fig. 5.2. δ34S versus δ18O of sulfate of the Yesares Gypsum. Open box defines Messinian seawater range
(Paytan, 1998; Turchyn and Schrag, 2004).
5.4 Results
5.4.1 δ 34S and δ18O of sulphate
Ten samples were analyzed for sulfur and oxygen isotopes of sulfate (Appendix D; Table
D.6). These data are shown in Fig. 5.2 relative to the expected range of gypsum precipitated
from Miocene seawater. δ 34S varies between 21.9 and 23.3h VCDT and δ 18OSO4 between
11.3 and 14.5h VSMOW. There are negligible differences in the isotope composition
among the nucleation cone, palisade, banded selenite, and supercone samples and δ 34S
and δ18OSO4 display no systematic changes with depth in the section.
5.4.2 Sr isotopes
Eleven samples from cycle 10 were measured for Sr isotopic analysis (Appendix D; Table
D.4). The range of 87Sr/86Sr is small (29 ppm), varying between 0.708941 and 0.708970.
These values fall slightly below the range of strontium isotope seawater expected for the
Messinian (McArthur et al., 2001) (Fig. 5.3).
5.4.3 XRD, wt% calcite and wt% dolomite of bulk carbonates
Thirty one marl samples were analysed by XRD (Appendix D; Table D.8). The results
are semi-quantitative and indicative of the relative abundances of common mineralogical
components only. In addition to clay, the major phases are gypsum, calcite, and dolomite
and each occurs in varying proportions in cycle 10. The marls at the base of the section
130 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
5.3 5.4 5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4
Age (Myr)
8
7
S
r/
8
6
S
r
0
.7
0
8
6
0
.7
0
8
5
0
.7
0
8
7
0
.7
0
8
8
0
.7
0
8
9
0
.7
0
9
0
6.5
Zanclean Messinian
PLG
O
u
tc
ro
p
O
D
P
-D
S
D
P
Sorbas 
Stratigraphy 
Zorreas 
Member 
Yesares 
Member 
Abad 
Member 
Sorbas 
Member 
M
S
C
 O
n
se
t 
 5
.9
7
1
B
a
se
 o
f 
P
lio
ce
n
e
  
5
.3
3
0
Primary Lower Gypsum (PLG) 
Halite 
Gypsum cumulate 
Gypsum cumulate with clastic 
Clastic gypsum 
Ostracods 
This study 
Primary Lower Gypsum (PLG) 
Upper Gypsum 
Halite and K-Mg-salts 
RGL 
"Colombacci" limestone 
Foraminifera 
Paratethyan fossil assemblages
Fish 
GLOBAL 
OCEAN
MEDITERRANEAN 
SEA
Fig. 5.3. Sr isotope curve during the Messinian in the Mediterranean Sea and Global Ocean. The progressive
change in the isotopic composition of Mediterranean waters during the MSC occurred because of
the restricted exchange with the global ocean due to closure of the gateways as well as a contribution
of freshwater with low 87Sr/86Sr from rivers (e.g. Nile and Rhone) or from the Paratethys (Roveri
et al., 2014a). 87Sr/86Sr data from this study are plotted at 5.795 Ma in accordance with the
cyclostratigraphic dating of cycle 10. Sorbas Basin stratigraphy after Krijgsman et al. (2001).
Figure modified after Roveri et al. (2014b).
contain both calcite and dolomite. The ratio of calcite to dolomite in the lower marls
decreases up-section such that dolomite becomes more dominant toward the gypsum–marl
contact (Fig. 5.4). Fine-grained gypsum within the marl is abundant in the upper marl from
8 cm below the marl-gypsum contact. Marls juxtaposed to supercones are purely dolomitic
and contain varying proportions of gypsum.
5.4.4 δ18O and δ13C of bulk carbonates
The oxygen and carbon isotope composition of thirty-seven carbonate samples vary widely
from -3.7h to 3.5h for δ18Ocarb and from -4.9h to 1.5h for δ13C (Appendix D; Table
D.8; Fig. 5.5). The carbon and oxygen isotope composition of the carbonates are within
the range of other Messinian carbonates associated with evaporites in the Mediterranean
5.4 Results 131
-30
-25
-20
-15
-10
-5
0
D
e
p
th
 (
cm
)
0.0 2.0-2.0-4.0 0.0 0.4 0.8 1.2 1.6-0.4
0 10 20 30 40
δ18Ocarb δ13C PDB
wt % Dolomite
B-30
-25
-20
-15
-10
-5
0
D
e
p
th
 (
cm
)
A
0.0 0.2 0.4 0.6
Dolomite:Calcite
0.8 1.0
0 10 20 30 40 50 60
% Gypsum
Fig. 5.4. (A) The ratio of dolomite to calcite of the marls at the base of the section. % gypsum is obtained
from XRD data; (B) δ18Ocarb, δ13C and % dolomite in the marls at the base of the section. 0 cm is
the gypsum–marl contact.
(Longinelli, 1979). Marls at the base of cycle 10 ranged from −3.7h to 1.9h for δ18Ocarb
and between −0.1h and 1.5h for δ13C. In contrast, marls juxtaposed with the super-cones
show relatively high δ18Ocarb (2.8 < δ18Ocarb < 3.5h) and low δ13C values (-4.9 < δ13C <
−3.8h).
5.4.5 Gypsum hydration waters
Twenty-six samples representing massive, banded and branching selenite were analysed for
their GHW (Appendix D; Table D.1). The measured δ17O, δ18O and δD values of GHW
gypsum from cycle 10 range from 1.38h to 3.38h, 2.02h to 6.32h, and -37.71h to -
6.23h, respectively. Eleven modern gypsum samples were also analysed from the Salina
at Cabo de Gata, Almería, Spain. The measured δ17O, δ18O and δD values of GHW range
from 5.50h to 6.27h, 12.05h to 10.25h, and 9.57h to 19.71h, respectively.
The oxygen and hydrogen isotope composition of the parent water in which the gypsum
formed is calculated by correcting for gypsum-water fractionation factors (Chapter 2). Frac-
tionation factors were assumed to be 1.0035 and 0.981 for oxygen and hydrogen isotopes,
respectively. The fractionation factors were selected based on water temperature (~15˚C
to 20°C; Section 5.5.8) and salinity (<150 g l-1; Section 5.5.2) estimates at the time of
formation. Following correction, the δ17O of the gypsum mother water for gypsum from
cycle 10 is estimated to have ranged between -0.47h to 1.52h, the δ18O from -1.47h to
2.81h, and δD from -13.97h to 13.01h. The calculated d-excess values of the mother
132 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
0.0 2.0 4.0-2.0-4.0
-6.0
-5.0
-4.0
-3.0
-2.0
-1.0
0.0
1.0
2.0
δ
18
 Ocarb (‰)
δ
1
3
C
 (
‰
)
Fig. 5.5. δ18O and δ13C of carbonates from the marl intervals within the Yesares Formation. Diamonds are
marls taken from the base of Cycle 10; squares are marls taken from beds juxtaposed to supercones.
water range from -11.2h to -2.2h, whereas the 17O-excess ranges from -37 per meg to
-10 per meg. The d-excess and 17O-excess values are negatively correlated with those of
the δ18O across the dataset. The calculated gypsum mother water δ17O, δ18O and δD of
the modern Salina samples ranged from 3.68h to 4.45h, 6.82 to 8.81h, and 29.12h
to 39.46h, respectively. The calculated d-excess values of the mother water range from
-29.7h to -23.0h, and the 17O-excess ranges from -82 per meg to -51 per meg (Fig. 5.6).
The δ18O and δD of gypsum mother water show a systematic change through cycle 10
(Fig. 5.7). The most negative δ18O and δD values are found at the base of the section in
the nucleation cone, which represents the first gypsum to precipitate above the marl. δ18O
and δD increase up section into the gypsum palisade reaching maximum values of 2.8h
and 13.0h, respectively, at 56 cm. Above this peak, δ18O and δD fall to average values of
0.6h and -1.0h, respectively, in the banded selenite above 130 cm. Isotope values in the
supercones are lower than those below with average values of 0.0 for δ18O and -6.4h for
δD. The top of the supercone has the lowest δ18O and δD values of -0.7h and -11.5h,
respectively. Thus, the entire section displays a systematic change from low isotopic values
at the base (nucleation cone), to maximum values in the middle (palisade gypsum), and a
return to low isotopic values near the top (supercone).
5.4.6 Microthermometric analysis
Eleven samples from the Messinian gypsum of the Sorbas Basin and two samples from
the modern Salina at Cabo de Gata, Almería, Spain, were selected for microthermometric
analysis (Appendix D; Table D.6). Primary inclusions in Messinian gypsum froze at tem-
peratures be-tween -40˚C and -60°C, with simultaneous shrinkage of the vapour bubble.
5.4 Results 133
-80
-60
-40
-20
0
20
40
-15 -10 -5 0 5 10 15
-80
-60
-40
-20
0
20
40
-12 -8 -4 0 4 8 12
A
αD = 0.981 
α18O = 1.00341
LM
W
LModern Salina Gypsum
Rain and Groundwater
Gy
ps
um
 H
yd
ra
tio
n 
W
at
er
Gy
ps
um
 M
ot
he
r 
W
at
er
αD = 0.981 
α18O = 1.0035
LMWL
δ18O (‰ VSMOW)
δD
 (
‰
 V
S
M
O
W
)
B
d
-e
xc
e
ss
 (
‰
 V
S
M
O
W
)
-20
-15
-10
-5
0
5
-4 0 4
y = -2.4x - 6.9
δ18O (‰ VSMOW)
Rio de Aguas Gypsum:
Nucleation Cone
Massive Selenite
Banded Selenite
Supercone
Fig. 5.6. Comparison of measured local meteoric water (grey circles) and predicted Messinian brine mother
water values for the differing facies of the Yesares Gypsum after correction for fractionation factors
(coloured diamonds) (Chapter 2). In (A) δ18O vs δD and (B) δ18O vs d-excess space, local meteoric
water measurements define the local meteoric water line (LMWL). The blue line represents the
evaporative line estimated by least-squares linear regression using mother-water data of Messinian
gypsum (in panel A, δD = 6.3 · δ18O – 5.4; R2 = 0.97). Light blue triangles are modern gypsum
data from the Cabo de Gata Salina. Red square is modern Mediterranean Seawater. The trajectories
for the evaporation of Mediterranean seawater for high (90%; slope ~7) and low (40%; slope ~3)
humidities are shown, defining a range of possible δ18O and δD values (pink field). Gypsum
precipitation takes place at 3 to 7 times concentrated seawater; gypsum saturation (blue field) is
calculated using the Rayleigh Distillation equation, assuming initial δ18O and δD values of ~1h
and 8h, respectively, 75% equilibrium between atmospheric vapour and fresh water, and gypsum
saturation occurring at an evaporation ratio of 3.8× (evaporation ratio = weight of H2O of original
seawater divided by weight of H2O in residual evaporated brine). Errors (1σ ) are less than the
width of the symbols.
134 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
0.0 2.0 4.0-2.0
δ18O (‰) 
5.03.02.0
eq % NaCl
H
e
ig
h
t 
(c
m
) 
4.0 6.0 0.0 -4.0
d-excess (‰) Insolation 
+
A B C D E
T1
T2
T3
T1
R
e
la
ti
v
e
 T
im
e
 
0
100
200
400
600
300
500
0
100
200
400
600
300
500
0
100
200
400
600
300
500
0
100
200
400
600
300
500
0
100
200
400
600
300
500
-8.0 -12.0
Fig. 5.7. (A) Schematic diagram of Cycle 10 scaled to height of the section, showing carbonate marls at the
base and top, a nucleation cone at the base of the gypsum, gypsum palisade, banded selenite and
supercone. Data points in B, C and D are plotted relative to the height from which the samples
were taken. (B) δ18O (VSMOW) of gypsum mother water vs. height in section; 1σ is the width of
the symbol. (C) wt% NaCl eq vs height in section. Wt% NaCl calculated from the final melting
temperatures of ice using the equation of Bodnar (1993). Error bars denote 1σ . (D) d-excess vs.
height in section; note that the d-excess axis is reversed. 1σ is the width of the symbol. All trend
lines are 5th Order Polynomial Regressions. (E) Hypothetical insolation curve, where T1 denotes
the humid period during marl deposition, T2 denotes the period of gypsum palisade formation at the
peak of the cycle, and T3 denotes the time period during banded selenite and supercone growth;
T1+T2+T3 is equal to approximately 21 ka, although the relative duration of T1, T2 and T3 is
unknown.
During heating, first melting was observed between -42˚C and -32°C, but this transition is
easily missed.
The Tmice of 206 primary Messinian inclusions ranged between -0.2°C and -6.0°C.
The average Tmice was -2.71°C. Using the revised equation and table for determining the
freezing point depression of H2O-NaCl solutions of Bodnar (1993), salinities estimates
range from 0.35% to 9.21% (average 4.45 wt% NaCl equivalent). Tmice of modern gypsum
from the Cabo de Gata Salina ranged between -6.5˚C and -10.1°C (average -8.0°C), corres-
ponding to an average salinity of 11.7%.
The Tmice of fluid inclusions show a systematic change during cycle 10. When Tmice
values are converted to wt% NaCl equivalent, the curve co-varies with the δ18O and δD of
gypsum hydration water data (Fig. 5.7). Nucleation cone samples at the base of the section
display the highest melting temperatures (averaging a Tmice of −1.7°C) and lowest salinities.
Moving up into the palisade gypsum, the lowest Tmice and greatest salinity values occur at
56 cm, averaging -3.6°C and 5.9 wt% NaCl equivalent, respectively. Above this maximum,
5.5 Discussion 135
values increase throughout the banded selenite to average Tmice values of -2.5°C at 231 cm.
Supercone Tmice values are collectively very similar to the banded selenite, averaging a
Tmice of -2.6°C.
5.5 Discussion
An outstanding question regarding the depositional environment of Messinian gypsum in
marginal basins of the Mediterranean is whether the gypsum was precipitated directly from
seawater or was influenced by meteoric water (e.g. Longinelli, 1979; Müller and Mueller,
1991; Attia et al., 1995; Playà et al., 2000; Flecker et al., 2002; Lu and Meyers, 2003; Lugli
et al., 2010; Natalicchio et al., 2014). Strontium and sulfur isotope results favour a domin-
antly marine origin whereas hydration water and fluid inclusion data support a significant
contribution by meteoric water. Below, these seemingly contradictory interpretations are
discussed and an explanation is proposed to reconcile the geochemical data.
5.5.1 Strontium, sulfur and oxygen isotopes of gypsum
Our measured δ 34S of gypsum of the Sorbas Basin ranges between 21.9h and 23.3h
VCDT (Fig. 5.2), which agrees with those measured in other Messinian evaporite deposits
(e.g. Longinelli, 1979; Lu et al., 2001; Lugli et al., 2007), and is similar to Miocene seawater
as reconstructed through pelagic marine barite from the equatorial Pacific (~21h). This
suggests that the Mediterranean Sea in the Miocene was connected, and isotopically similar,
to the global ocean. Similarly, the measured δ 18OSO4 of Yesares gypsum (11.3h to 14.5h)
is similar to the δ 18OSO4 of Miocene seawater sulfate, reconstructed from pelagic marine
barite (~12h; Turchyn and Schrag, 2004). The fact that near-seawater values are measured
for sulfur and sulfate-oxygen isotopes in the gypsum of the Yesares Member suggest that
deposition was dominantly in a marine environment (e.g. Müller and Mueller, 1991; Lu and
Meyers, 2003).
It has been suggested that gypsum precipitates with a negligible sulfur isotope fraction-
ation, but as much as a 3h offset for oxygen isotopes (Lloyd, 1968). However, this oxygen
isotope fractionation factor remains poorly constrained, and the isotope composition of a
multiply evaporated basin may evolve over time. If the sulfate from Miocene seawater
trapped in the Mediterranean were quantitatively precipitated, then, by mass conservation,
the δ 34S and δ 18OSO4 of the bulk precipitated gypsum should reflect Miocene seawater.
Subsequent solution-reprecipitation of the gypsum in water with a similar δ 34S or δ 18OSO4
will not alter the δ 34S or δ 18OSO4 of the gypsum significantly. Abiotic oxygen isotope
136 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
exchange between aqueous sulfate and water does not occur over geologically relevant
timescales (Lloyd, 1968; Rennie and Turchyn, 2014).
The sulfur and sulfate-oxygen isotope composition of meteoric water that could have
been supplied to the site of gypsum formation would largely reflect the isotope composition
of sulfur-bearing minerals that are being locally weathered. Measurements of the δ 34S and
δ 18OSO4 in Spanish rivers today are heavily biased towards evaporite values (high δ
34S
and δ 18OSO4) owing to the ongoing weathering of the Messinian evaporites. However,
underneath these evaporites exist sedimentary deposits containing shale beds that plausibly
bear pyrite, and it is possible that oxidative pyrite weathering could have dominated the
terrestrial sulfate–isotope signal during the MSC. Oxidative weathering of pyrite produces
river sulfate that is enriched in the 32S isotope and typically with a δ 18OSO4 similar to the
δ18O of the water (e.g. ~-5h) (Calmels et al., 2007). Thus, it is hypothesised that if a
significant amount of sulfate used to make the Yesares gypsum was derived from terrestrial
sources, then the δ 34S and δ 18OSO4 of the gypsum would be significantly lower than the
measured values.
87Sr/86Sr of a semi-enclosed marginal basin is similarly a function of the balance between
ocean and riverine input, both of which are characterised by different strontium concentra-
tions and 87Sr/86Sr. Because of the long residence time of strontium relative to the mixing
time of the ocean, 87Sr/86Sr is homogeneous in the global ocean. The 87Sr/86Sr of Messinian
seawater is well established with values varying between 0.70900 and 0.70903 during the
MSC (McArthur et al., 2001). There is no fractionation of strontium isotopes during gypsum
precipitation; thus, a direct comparison can be made between the 87Sr/86Sr of gypsum and
the parent solution.
Measured 87Sr/86Sr of the gypsum samples in cycle 10 of the Río de Aguas section vary
between 0.708941 and 0.708970. The values fall slightly below the range of the strontium
isotope curve for seawater (McArthur et al., 2001), suggesting that continental waters with
a lower 87Sr/86Sr than contemporaneous Miocene seawater may have modified the brine
within the Sorbas Basin. The values are within the range given by Roveri et al. (2014b)
for the 87Sr/86Sr isotopic composition of Mediterranean seawater values between 5.97 Ma
and 5.60 Ma (Fig. 5.3), who interpret the near global ocean values as a dominantly marine
formation environment. If the interpretation of Roveri et al. (2014b) is correct, the 87Sr/86Sr
data, coupled with δ 34S and δ 18OSO4 data, imply there was little contribution of meteoric
water to the brine that formed the Yesares Gypsum.
5.5 Discussion 137
5.5.2 Gypsum hydration water and microthermometry
The observed range of δ18O and δD of gypsum mother water (-1.47h to 2.81h; -13.97h
to 13.01h, respectively) is significantly lower than the expected range from gypsum formed
solely from the evaporation of seawater. Assuming a starting point of seawater with modern
δ18O and δD values, Fig. 5.6A shows the possible trajectories of δ18O and δD for evapor-
ative concentration under conditions of high and low humidity. Modern gypsum measured
from the Salina at Cabo de Gata falls in the region of predicted gypsum saturation for
evaporated seawater. If the gypsum precipitated in the Sorbas Basin were solely of marine
origin, we would expect the values to fall along a mixing line within this zone of gypsum
saturation. Instead all δ18O and δD values from cycle 10 fall well below the saturation point
for gypsum on the evaporative line. This implies that the Yesares Gypsum was precipitated
from a hybrid brine, consisting of a mixture of seawater and meteoric water.
An alternative explanation of the low δ18O and δD of hydration water is that the values
do not reflect the original brine but have undergone postdepositional isotopic exchange
of the crystallisation water with meteoric water. The kinetics of the isotopic exchange
of GHW is not well studied, but Sofer (1978) concluded GHW only retains its isotopic
composition under dry conditions (e.g. Negev Desert) and exchanges isotopically under
humid conditions. Sofer suggested the GHW method can only be applied to primary gypsum
that has not undergone dehydration, exchange, or recrystallisation, and thus has retained its
original water isotopic composition. To test if the isotope composition of GHW is reliable
in the Sorbas Basin, GHW isotopic results were directly compared with microthermometric
analysis of primary fluid inclusions from the same samples in cycle 10.
Primary fluid inclusions trap the parent brine and reflect the composition of the solution
from which the gypsum precipitated (Attia et al., 1995). Microthermometric data show that
no Messinian fluid inclusions have salinities high enough to be within the range of gypsum
saturation if evaporated from seawater. Gypsum first precipitates during the evaporation of
seawater at a wt% NaCl equivalent of ~11%, which should correspond to a Tmice of −7˚C to
−8°C (Attia et al., 1995; Natalicchio et al., 2014). In comparison, the fluid inclusions from
the modern Salina samples melted at −8.0°C corresponding to 11.7 wt% NaCl equivalent,
and are above the point of gypsum saturation. In contrast, Messinian fluid inclusions from
cycle 10 have average wt% NaCl equivalent ranging from 2.87% to 5.76% (averaging 4.5
wt% NaCl equivalent), indicating that the fluids from which gypsum precipitated were
moderately saline to brackish (Attia et al., 1995; Natalicchio et al., 2014). Additionally,
the average strontium contents of selenite from cycle 10 samples range from 506 ppm to
989 ppm, whereas the gypsum from the Cabo de Gata Salina also have an average strontium
concentration of 1730 ppm. As the partition coefficient of strontium into gypsum is less
138 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
-8
-6
-4
-2
0
2
4
6
8
10
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Salinity = 0.08(δ18O) - 3.8
R2 = 0.66
δ
1
8
O
 (
‰
 V
S
M
O
W
)
Salinity (ppt)
Fig. 5.8. Coupled measurements of δ18O of hydration water and salinity of fluid inclusions in Messinian
Gypsum. Salinities calculated from the final melting temperatures of ice using the equation of
Goldstein and Reynolds (1994). The dark grey band denotes 95% confidence limits. The y intercept
at zero salinity defines the freshwater δ18O endmember. Error bars denote 1σ . 1σ in δ18O space is
the edge of the symbol.
than unity (Kushnir, 1980), the strontium concentration of gypsum precipitated directly
from seawater after evaporative enrichment should correspond to ~1500 ppm to 4000 ppm
(Rosell et al., 1998; Rossi et al., 2015).
Using the equation of Goldstein and Reynolds (1994), we converted wt% NaCl equival-
ent to salinity in parts per thousand (ppt) seawater. The salinity of primary Messinian fluid
inclusions is correlated with δ18O and δD of gypsum mother water, yielding an R2 of 0.66
and 0.65, respectively (Fig. 5.8). The intercepts of the regression equations and salinity are
-3.8 ±2 for δ18O and -30.2 ±15 for δD. These values define the isotope composition of the
freshwater endmember and are within error of the average isotope values of precipitation
and groundwater data from the local region of Almería today (δ18O = -4.3h and δD =
-22.2h; IAEA, 2005). These results provide strong evidence that both the fluid inclusion
and gypsum hydration water are representative of the mother brines in which the gypsum
formed and postdepositional isotopic exchange has not occurred. These results also imply
that the mother fluid from which the gypsum was precipitated was influenced by meteoric
water.
5.5 Discussion 139
5.5.3 Decoupling of the source of ions and water during gypsum form-
ation
An inconsistency exists between the marine origin of gypsum implied by sulfur and oxygen
isotopes of sulfate and the significant meteoric influence inferred from the δ18O and δD of
gypsum hydration water and the salinities of fluid inclusions. The formation of the PLG is
most commonly attributed to the evaporation from brines produced by continuous inflow of
seawater from the Atlantic Ocean, providing the ion supply necessary for gypsum precipit-
ation, coupled with episodes of reduced outflow under drier climate conditions (Krijgsman
and Meijer, 2008; Topper and Meijer, 2013). A non-marine origin of Messinian gypsum
in marginal basins raises questions concerning the marine signatures sulfur and oxygen
isotopes in sulfate. An alternate mechanism is needed to explain the lower salinities inferred
from fluid inclusion and isotopic results in the Sorbas PLG.
One interpretation is that strontium isotopes, sulfur and oxygen isotopes in sulfate are
relatively insensitive to freshwater inputs and are therefore not recording the dilution of
seawater by non-marine fluids. Even if this is the case, a mechanism is needed to explain
the source of the ion required to reach gypsum saturation. A plausible explanation is sulfate
molecules are decoupled from the source of the water for the gypsum hydration water, which
was a mixture of marine and terrestrial sources. The source of the sulfate and strontium
may come from dissolution of previously deposited marine gypsum that recharged the basin
and maintained gypsum saturation. Evaporation of this fluid would produce a brine with
the observed δ18O and δD values and salinities, while retaining marine δ 34S and δ 18OSO4
values and near-marine 87Sr/86Sr values, because of low concentrations of these elements
in meteoric water. In addition, the dissolution of marine gypsum (with a strontium content
of ~1500 ppm to 4000 ppm) in meteoric water will result in molar Sr/Ca ratios lower than
seawater. Subsequent gypsum precipitation will reflect the low Sr/Ca ratios of the hybrid
fluid, which is indicative of the low strontium concentrations observed in the majority
of the Sorbas PLG sample. This recycling mechanism has been proposed previously by
Natalicchio et al. (2014) to explain salinity inferred from fluid inclusions in gypsum from
the Piedmont Basin (Italy).
5.5.4 Deposition of the lower marl unit
A change to the isotopic composition of the carbonate can be seen at the base of cycle 10,
which is made up of a mixture of calcite and dolomite (Fig. 5.4). There is an increase in
the proportion of dolomite/calcite towards the top of the marl where it transitions into the
gypsum unit above. The proportion of gypsum in the marls increases in the 8 cm below the
140 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
transition into the massive gypsum above the marl. This progression of calcite to dolomite
to gypsum is indicative of fractional crystallisation of minerals by evaporative enrichment
(Eugster, 1980).
The δ18Ocarb and δ13C values of bulk carbonate in the lower marl unit of cycle 10
increase from -3.7h and -0.1h, respectively, at the base to a maximum of 1.9h and 1.5h
at 10 cm below the transition into the massive gypsum before decreasing in the top of the
lower marl unit (Fig. 5.4). The δ18Ocarb and δ13C of bulk carbonate is expected to be higher
in the marls dominated by dolomite because of the different isotope fractionation factor for
oxygen and carbon isotopes between calcite and water, and dolomite and water; the oxygen
and carbon isotope composition of dolomite is ~2.6h and ~2.4h greater than calcite when
formed in equilibrium (Sheppard and Schwarcz, 1970; Vasconcelos et al., 2005). This
indicates that a good correlation exists between the percent dolomite and bulk δ18Ocarb and
δ13C. The relatively positive δ13C in the marls at the base of the section suggest there is not
significant oxidation of organic carbon, which would drive the δ13C lower (Fig. 5.4). SEM
investigations show that both calcite and dolomite present in the lower unit are authigenic
(Fig. 5.9). Indeed, no angular or fractured mineral surfaces indicative of transport and
mechanical erosion were observed. The elongated, almost acicular, morphology of the
calcite crystal as well as the rhombohedral habit of the dolomite is commonly observed
in carbonate minerals that form in the peritidal and supratidal zone of modern evaporitic
environments (Bontognali et al., 2010). We found no clear morphological evidence (i.e.
dissolution/re-crystallisation structures, hybrid mineral shapes) suggesting that the dolomite
formed through a penecontemporaneous replacement of calcite. Therefore, the progressive
enrichment in dolomite versus calcite observed in the studied sequence may have been
caused by an environmental change from fresher peritidal hybrid-waters to more evaporated
intertidal/supratidal conditions.
5.5.5 Formation of the supercone-marl unit
The origin of the supercone structures and associated laminated marls in the Río de Aguas
section have prompted much speculation since their initial description by Dronkert (1985).
We have identified important differences in the mineralogical and isotopic composition of
the marl at the base of the cycle and the marl associated with the supercones near the top.
The carbonate in the marl of the supercone unit is entirely dolomite and display higher
δ18Ocarb and lower δ13C values than those in the lower marl unit (Fig. 5.4).
The supercone selenite and the associated marl are considered to be syndepositional as
individual lamina thin and disappear into gypsum crystals. Some laminae wrap around the
supercone branches and can be traced above and below (Fig. 5.10). The syndepositional
5.5 Discussion 141
A
B
C
2 μm
2 μm
1 μm
Figure X
D
1 μm
C
O Mg Pt
Ca
1.00 2.00 4.003.00
co
un
ts
kV
Ca
Si0.2k
0.4k
C
Mg Ca
Ca
1.00 2.00 4.003.00
co
un
ts
kV
0.2k
0.4k Pt
Si
O
1.00 2.00 4.003.00
co
un
ts
kV
0.2k
0.4k
C Pt
Ca
Ca
O
Fig. 5.9. Scanning electron microscopy (SEM) photomicrographs and energy dispersive X-ray (EDX)
analyses of the carbonate minerals present in the studied samples. (A) Image of a representative
sample of the lower marl unit (i.e., RdA1 -25 cm). Calcite crystals are characterized by an
elongated, almost acicular morphology, while dolomite crystals display a rhombohedral habit.
No morphological evidence indicates that dolomite formed through a diagenetic replacement of
calcite or whether it is a primary precipitate. Both calcite and dolomite are authigenic. (B) Typical
calcite crystal present in the lower unit and corresponding EDX spectrum, which is consistent with
a Ca-carbonate. (C) Typical dolomite crystal present in the lower unit and corresponding EDX
spectrum, which is consistent with dolomite. (D) Image of representative dolomite crystals present
in the upper unit (i.e., SC marl 100) and corresponding EDX spectrum. Rhombohedral crystals
of dolomite seem to form through aggregation of dolomite nanoglobules (white arrow), which
suggests that dolomite in the ‘supercone unit’ is a primary precipitate. Dolomite crystals that form
through aggregation of nanoglobules are typically observed in microbially mediated dolomite. Pt
peaks in all EDX spectra are due to the metal coating used for sample preparation. Si peaks in the
spectra measured from dolomite are interpreted as a contamination from the surrounding authigenic
silicate minerals.
142 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
nature of the dolomitic marl and gypsum permits palaeotemperature to be estimated if the
gypsum and dolomite were formed in isotopic equilibrium with the same water. The δ18O
of the gypsum hydration water provides the δ18O of the water whereas the δ18Ocarb of
the dolomite provides the carbonate value (Hodell et al., 2012). Marl sample SC50-10
(δ18Odolomite = 3.05h) is directly juxtaposed to gypsum sample SC50B (δ18O = 1.08h),
and together yield a temperature of 19.2°C using the dolomite–water palaeotemperature
equation of Vasconcelos et al. (2005). This temperature is close to the average sea surface
temperature from the local region of Almería today (15°C to 25°C). It also agrees with
Messinian sea surface temperatures (17°C to 18°C) obtained from fluid inclusion analyses
in Messinian halite (Speranza et al., 2013). These reasonable temperature estimates provide
support that the dolomite and gypsum of the marl-supercone beds were formed at the same
time from water with similar δ18O values.
The marls associated with the supercones are very finely laminated and lack any evid-
ence of bioturbation indicating anoxic conditions, and are reminiscent of the anoxic dolo-
mite found in the Messinian Tripoli Formation in the deep basin (McKenzie et al., 1979).
The δ13C of the dolomite is low and varies between -3.8h and -4.9h, indicating a source
of dissolved inorganic carbon at least partially derived from the oxidation of organic matter;
thus, the formation of the dolomite may be microbially mediated (Vasconcelos et al., 1995;
Vasconcelos and McKenzie, 1997). This interpretation is also supported by the morphology
of the dolomite crystals, which form through aggregation of spherical/amorphous nanoglob-
ules (Fig. 5.9). In a study conducted with sulfate reducing bacteria capable of mediating
dolomite formation, Bontognali et al. (2008) demonstrate that such nanoglobules represent
the early stage of mineral nucleation within microbially produced extracellular polymeric
substances (EPS). Nanoglobules that merge to form rhombohedral dolomite have also been
observed in microbial mats that occur in the intertidal and supratidal zone of the sabkha
of Abu Dhabi (Bontognali et al., 2010). In this modern evaporitic environment, dolomite
can co-occur with gypsum crystals. As the sediment ages, only the dolomite (and not the
microbial mat) is preserved, similar to observations in the marls juxtaposed to supercones.
5.5.6 Climate control of gypsum–marl deposition
The gypsum-marl cycles of the PLG are purported to represent climatic changes paced by
precession with gypsum deposited during more arid conditions and sapropelic marl during
more humid conditions (Krijgsman et al., 2001, 1999; Krijgsman and Meijer, 2008). The
mineralogical and isotopic trends support this interpretation for cycle 10:
i. The peak wet conditions occurred at the precession minimum (maximum summer
5.5 Discussion 143
Fig. 5.10. Sample locations of Supercone (‘SC’) gypsum and marls juxtaposed to supercones (‘SC Marls’),
cycle 10, Río de Aguas section.
insolation) and is represented by the deposition of the laminated marl at the base
of the studied section. Climate conditions became progressively drier promoting the
formation of dolomite, and finally gypsum towards the top of the lower marl unit (Fig.
5.4).
ii. As climate became progressively drier, detrital input declined and pure gypsum began
to precipitate in nucleation cones. The hydration water of the gypsum in the nucle-
ation cones has relatively low δ18O and δD and the fluid inclusions indicate relatively
low salinity compared to the gypsum above (Fig. 5.7).
iii. Salinity, δ18O and δD progressively increased above the nucleation cones reaching
a maximum in the massive selenite palisade unit, marking peak aridity associated
with the precession maxima (summer insolation minimum). A highly stratified brine,
which in relation to total water depth produces a high and stable pycnocline (the
gypsum saturation interface), creates the condition necessary for massive selenite
growth (Ba¸bel, 2007). The δ18O and δD and salinity of the palisade gypsum support
144 Measurements of δ18O and δD of hydration water and salinity of fluid inclusions
formation at a low saturation state needed for slow crystal growth (Ba¸bel, 2007).
iv. After formation of the gypsum palisade, the trend in aridity reversed indicating a
progressive decrease in salinity related to increased freshwater input as recorded by
decreasing salinity, δ18O and δD values of the banded selenite unit.
v. As precipitation increased with increasing summer insolation, the supercones and
associated dolomitic marls began to be deposited. The salinity of fluid inclusions
and δ18O and δD of gypsum in the supercones are lower than the banded and massive
selenite units below. Detrital input also increased and led to the deposition of the
dolomitic marls juxtaposed to the supercones (Section 5.5.5). Above the supercone
the basal clay unit of Cycle 10 was deposited, marking the return of wetter climate
conditions associated with the precession minimum.
5.6 Conclusions
We made tandem measurements of the isotopic composition of GHW and the salinity of
fluid inclusions to test whether Messinian gypsum retained the original isotopic composition
and salinity of the mother water. After correction of oxygen and hydrogen isotopes of GHW
for known fractionation factors, we show that the δ18O and δD of the mother water is highly
correlated with salinity of fluid inclusions in gypsum deposits of cycle 10 within the Yesares
Member. The intercepts of the regression equations (i.e. at zero salinity) define the isotope
composition of the freshwater endmember. These values are within error of the average
isotope composition of precipitation and groundwater data from the local region of Almería
today. This agreement provides strong evidence that the GHW has retained its isotope
composition and has not undergone postdepositional exchange.
The isotope and salinity values indicate a significant contribution of meteoric water
during formation of Messinian gypsum in the Sorbas Basin. This observation contrasts
with sulfur and oxygen isotopes in sulfate (21.9 < δ 34S < 23.3h; 11.3 < δ 18OSO4 <14.5h)
and strontium isotopes (0.708942 < 87Sr/86Sr < 0.708971) that are similar to those measured
in other Messinian evaporites of the Mediterranean. We suggest the source of the ions for
gypsum formation is recycled from previously deposited units, whereas the source of the
water is a mixture of marine and meteoric sources. The 87Sr/86Sr, δ 34S, and δ 18OSO4 retain
the signature of the parent sulfate deposits because of low concentrations of these elements
in meteoric water, whereas the δ18O and δD and salinity reflect a mixed signal of marine and
freshwater. The recycling of solutes explains the discrepancies between strontium, sulfur
5.6 Conclusions 145
and water isotopes by decoupling the geochemical signatures of the ions and water during
gypsum formation.
Lastly, we demonstrate that the patterns displayed by the δ18O and δD and salinity of
the gypsum, together with the overall changes in mineralogy (calcite–dolomite–gypsum) in
cycle 10, are consistent with a precession-driven change in climate. Additional studies are
needed to determine whether the patterns observed here for cycle 10 apply to other cycles
in the Yesares Member.

Chapter 6
Geochemical evidence for meteoric
water, orbital and eustatic control of
basin hydrology during the first stage of
the Messinian Salinity Crisis
Nicholas P. Evans, Fernando Gázquez, Thomas K. Bauska, Harold J. Bradbury, Alexandra
V. Turchyn, José María Calaforra, and David A. Hodell.
Twelve cycles of gypsum-marl deposition from the Messinian Yesares Member in the
Sorbas Basin, Spain were examined using a high-resolution, multi-isotope approach to
reconstruct the palaeo-hydrological dynamics of the basin and the wider Mediterranean
region during the first stage of the Messinian Salinity Crisis (MSC). The past isotopic
composition (δ18O, δD, 17O-excess and d-excess) of basin waters during evaporite form-
ation were determined by analysing the structurally bound water in sedimentary gypsum
(CaSO4 ·2H2O). These measurements are combined with salinity inferred from gypsum
fluid inclusions, strontium (87Sr/86Sr), calcium (δ 44/40Ca), sulfur (δ 34S), and oxygen iso-
topes of sulfate (δ 18OSO4) in the same samples, as well as isotopic analysis (δ
44/40Ca,
87Sr/86Sr, δ18Ocarb, δ13C) of the interbedded carbonate marls, to determine the changing
environment of deposition and its relation to astronomically-forced climate change.
This study provides data to show that obliquity-controlled sea level and eccentricity-
modulated precession, superimposed on longer-term tectonic restriction of the Mediterranean-
Atlantic exchange, together played a significant role in modulating gypsum-marl deposition
in the Sorbas Basin. Strontium isotopes in gypsum display a gradual, but non-unidirectional
148 Meteoric water, orbital and eustatic control of basin hydrology
divergence from the seawater curve up-section, with slight reversals in the general trend of
decreasing 87Sr/86Sr at ~5.84 Ma and 5.72 Ma. Maximum divergence of 87Sr/86Sr from the
seawater curve, as well as anomalous authigenic carbonate mineralogy (dolomite) and high
δ18Ocarb is coincident of glacial stage TG20 and an eccentricity minimum at ~5.75 Ma.
Calcium isotopes track the variability in 87Sr/86Sr, overprinted by Rayleigh fractionation
effects. The non-unidirectional changes in strontium and calcium isotopes in combination
with the patterns of the expression (i.e. thickness and facies) of the cycles in the Sorbas
section indicate that sea level changes influenced basin hydrology.
6.1 Introduction
During the first stage of the Late Miocene Messinian Salinity Crisis (MSC, 5.97 Ma to
5.60 Ma), large volumes of sulfates and interbedded carbonate marls (the Primary Lower
Gypsum, PLG) were deposited in silled sub-basins of the Mediterranean. The Sorbas Basin
(SE Spain) has become a type-locality for this first stage of cyclic evaporite deposition, and
is important in the study of the oceanographic evolution of the Miocene Mediterranean Sea
owing to its proximity to the Atlantic gateway. Questions remain, however, about the factors
that mediated the precipitation of gypsum and carbonate during each of the 15 gypsum-marl
cycles, as well as the primary ion source during each depositional episode.
Previous studies that utilised traditional isotope tracers, such as strontium (87Sr/86Sr),
sulfur (δ 34S), and oxygen isotopes in sulfate (δ 18OSO4), suggest that the deposition of the
PLG occurred under a permanent open-ocean connection between the Mediterranean and
the Atlantic, driven by astronomically controlled climatic oscillations (e.g. Longinelli,
1979; Müller and Mueller, 1991; Krijgsman et al., 1999; Lu and Meyers, 2003; Lugli
et al., 2010; Manzi et al., 2013; Roveri et al., 2014a). The average freshwater budget in
the Mediterranean remains largely unconstrained, however, because of the insensitivity of
the traditional isotope tracers to continental water input. There is also evidence suggesting
Mediterranean salinity may have been lower than the saturation horizon at which gypsum
precipitation occurs (~130 g l-1) during the evaporation of only seawater (e.g. Rosell et
al., 1998; Natalicchio et al., 2014; Dela Pierre et al., 2014; Rossi et al., 2015; Evans et al.,
2015). Because of difficulties reconstructing both palaeosalinity and the relative influence of
continental water contributions on the PLG gypsum-carbonate cycles (e.g. Lu and Meyers,
2003; Rossi et al., 2015), neither the amplitude nor the phasing of salinity changes in the
Mediterranean basin are clear (Flecker et al., 2015). Additionally, the correlation between
isotopic variations, eustatic sea-level changes and orbital oscillations has remained tentative
owing to the relatively low sampling resolution in the majority of previous studies.
6.1 Introduction 149
Messinian Evaporites DSDP-ODP Sites
Gypsum/Halite/Gypsum "Trilogy"
Undifferentiated Gypsum-Halite
Only Gypsum
Evaporites Recovered
No Evaporites Recovered
45°
40°
35°
20°
Orogenic front
Passive margin
Transcurrent fault400 km
IBERIA
AFRICA
ATLANTIC
OCEAN
121 976 977 978
122
123 656
133
974
373
973
121
975
372
124975
371
134
654
132
654 652
374
CALTANISSETTA
BASIN
SORBAS
BASIN
Sampled Sites
Other Western Mediterranean 
sites containing Primary 
Lower Gypsum 
0 0.05-0.05-0.1 0.1
Eccentricity
Climatic Precession 
50
100
150
0
50
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
7
6
5
4
3
H
e
ig
h
t 
(m
)
Marls
Gypsum Supercones
Banded Gypsum
Massive Gypsum
50
Monte Grotticelle
(Caltanissetta Basin, Sicily)
Yesares Member
(Sorbas Basin, SE Spain)
xyz
Irpinia
Maiella
Vena del 
Gesso
Hoya de 
la Sima
Nijar
6.0
5.9
5.8
5.7
5.6
A
g
e
 (
M
a
)
P
S
M
Q
B
d
I
R
d
A
C
C
LM
Fig. 6.1. (Upper) Map of the Messinian evaporites in the Mediterranean and the sampled basins. Also
shown is the location of the DSDP-ODP boreholes in which Messinian evaporites were recovered,
as well as other PLG sections found in marginal basin settings. Modified after Rouchy and Caruso
(2006) and Roveri et al. (2014a). (Lower) Astronomical tuning of the PLG units in the Sorbas and
Caltanissetta Basins. Also shown are the localities of the sampled sections relative to the Sorbas
Basin log: PS (the Perales section); MQ (the Majadas Viejas quarry); CC (Covadura Cave); BdI
(the Barranco del Infierno section); RdA (the Río de Aguas section); LM (the Los Molinos section).
See text for details.
150 Meteoric water, orbital and eustatic control of basin hydrology
A unified theory that explains the formation of the Mediterranean Salt Giant requires
an accurate understanding of the ion source and underlying mechanism by which the PLG
formed. Here, a high-resolution (sub-precessional scale), multi-isotope study of 12 cycles
of the Yesares Member, Sorbas Basin, is performed to constrain the hydrological variability
during the first stage of the MSC. In addition to 87Sr/86Sr, δ 34S, and δ 18OSO4 isotope
analysis of gypsum, we measured the triple oxygen and hydrogen isotopes in GHW, calcium
isotopes (δ 44/40Ca), and salinity deduced from fluid inclusions. Together, these measure-
ments provide detailed information on the parent water from which gypsum and carbonate
were precipitated and offer high-resolution, independent evidence for the relative influence
of basin stratification, sea-level variability and continental influences in marginal basins. We
demonstrate that the Messinan PLG formed from evaporation of seawater that was strongly
influenced by abundant freshwater input. We also show non-unidirectional changes in intra-
cycle strontium and calcium isotope values are consistent with eustatic sea-level variability,
and suggest that the appearance of the differing selenite facies can be linked to periods of
increased northern hemisphere glaciation.
6.2 Geological setting
The Sorbas Basin, SE Spain: The Sorbas Basin (Fig. 6.1) preserves the transitional
interval between the marine marls of the Upper Abad (UA) Member to the stage 1 evap-
orites of the Yesares Formation (Krijgsman et al., 2001). This area was likely part of the
Betic Corridor that connected the Atlantic and the Mediterranean during the Late Miocene
(Sierro et al., 2001, 1999). It therefore represents a key area for understanding the Western
Mediterranean palaeoceanographic evolution during the onset of the MSC (Reghizzi et al.,
2017).
The karstified Yesares Member occurs within a ~140 m thick sequence consisting of
gypsum-carbonate cycles that have been astronomically tuned to precession (e.g. Krijgsman
et al., 2001, 1999; Roveri et al., 2009; Lugli et al., 2010). The selenitic gypsum units are up
to 30 m thick. The facies description is given in Section 4.2.1.
Dronkert (1985) was the first to describe the sedimentary cycles of the Yesares Member
of the Sorbas Basin, reporting 12 cycles of alternating gypsum and carbonate-marl. In
subsequent studies, however, the reported stacking pattern and number of cycles of the
Yesares Member differs significantly from this original interpretation. Rosell et al. (1998)
described 13 cycles in their stratigraphic log of the Los Molinos section, whereas Krijgsman
et al. (2001) described 14 cycles in the same section. Lugli et al. (2010) and Roveri et al.
(2009) reported 15 and 16 cycles, respectively, although these studies interpret the basal
6.2 Geological setting 151
cycle of the Los Molinos section to be cycle 2. This is because ‘cycle 1’, a carbonate-
sapropel couplet that occurs at the base of this locality, has been correlated with a recently
discovered gypsum cycle at the base of the Perales section (Manzi et al., 2013). As well
as the addition of this cycle at the bottom of the PLG, the variable number of total cycles
described is likely related to the differing interpretation of the upper half of the Los Molinos
section where outcropping gypsum units are obscured.
We reinvestigated the cyclicity and stacking pattern of the Yesares Member in the Sorbas
Basin (Fig. 6.1). We focus our study on gypsum and carbonates from the passages of
Covadura Cave (Bosque branch; 37°07’21”N 002°04’8”W); the Majadas Viejas quarry
(37°04’41”N, 002°05’51”W); the Perales section (37°06’24” N 002°02’33” W); the Río
de Aguas section (37°05’23.2” N 002°06’54.2” W); the Los Molinos section (37°06’19”
N 002°04’44” W); and the Barranco del Infierno section (37°04’58” N 002°07’07” W).
Our field surveys concluded that twelve distinct gypsum-carbonate cycles are present in the
wider basin, whereas an additional three gypsum-carbonate cycles are present in uppermost
part of the Los Molinos section.
The lowermost gypsum bed is located at the Perales section. Manzi et al. (2013) recog-
nized this highly discontinuous gypsum bed within the ‘transitional interval’ at the Abad-
Yesares boundary (Sierro et al., 2001) below the lowermost continuous gypsum layer (i.e.
cycle 2). The subsequent two cycles are exposed at the Majadas Viejas quarry, displaying
massive selenite with little crystal variation. Cycles 3-8 can be observed in the passages of
Covadura Cave. The sixth cycle marks the first appearance of branching selenite supercones,
and the stacking pattern in the Covadura Cave is generally recognizable in the lower half of
the Los Molinos section.
The upper part of the Yesares Member is best exposed in the Rio de Aguas and Barranco
del Infierno sections, where the transition into the lowermost Sorbas Member is also exposed
(Krijgsman et al., 2001). Eight distinct gypsum cycles are present in the Barranco del
Infierno section; the uppermost gypsum cycle is a relatively discontinuous layer consisting
of massive selenite, whereas the lower seven cycles in this section display both massive
selenite and supercone structures. The presence of seven cycles containing supercones
allows us correlate the lowermost cycle of the Barranco del Infierno section to cycle 6 in
the Covadura Cave. Thus, the eight cycles of the Barranco del Infierno section correspond
to cycles 6-13. The three cycles exposed at the base of the Rio de Aguas section also
contain supercone structures and are overlain by a discontinuous gypsum/carbonate unit.
These cycles at Rio de Aguas are recognizable as the lateral equivalent of the upper part of
the Barranco del Infierno section, thus corresponding to evaporite cycles 10-12. Dronkert
(1985) observed a similar correlation between the Rio de Aguas and Barranco del Infierno
152 Meteoric water, orbital and eustatic control of basin hydrology
-80
-60
-40
-20
0
20
40
-15 -10 -5 0 5 10 15
-80
-60
-40
-20
0
20
40
-12 -8 -4 0 4 8 12
-80
-60
-40
-20
0
20
40
-12 -8 -4 0 4 8 12
A
αD = 0.981 
α18O = 1.00341
LM
W
LPLG Cycles 1-5
PLG Cycles 6-15
Modern Salina Gypsum
Rain and Groundwater
Gy
ps
um
 H
yd
ra
tio
n 
W
at
er
Gy
ps
um
 M
ot
he
r 
W
at
er
αD = 0.981 
α18O = 1.0035
LMWL
LMWL
E/I = 0.5
E/I = 1
E/I = 0.5
E/I = 1
Recharge
Recharge
δ18O (‰ VSMOW)
δD
 (
‰
 V
S
M
O
W
)
B
d-
ex
ce
ss
 (
‰
 V
S
M
O
W
)
C
17
O
-e
xc
es
s 
(p
er
 m
eg
)
-20
-15
-10
-5
0
5
-4 0 4
y = -1.6x - 5.8
R2 = 0.78 
y = -1.9x - 3.5
R2 = 0.70 
δ18O (‰ VSMOW)
δ18O (‰ VSMOW)
6.2 Geological setting 153
Fig. 6.2. Left. Comparison of measured local meteoric water (grey circles) and predicted mother water
values after correction for fractionation factors (Gázquez et al., 2017). In (A) δ18O vs δD, (B)
δ18O vs d-excess and (C) δ18O vs 17O-excess space, local meteoric water measurements define the
local meteoric water line (LMWL). Light blue triangles are modern gypsum data from the Cabo de
Gata Salina. Red square is Modern Mediterranean Seawater. The trajectories for the evaporation
of Mediterranean seawater for high (90%; slope ~7) and low (40%; slope ~3) humidities are shown
in A, defining a range of possible δ18O and δD values (pink field). Gypsum precipitation takes
place at 3 to 7 times concentrated seawater; gypsum saturation (blue field) is calculated using the
Rayleigh Distillation equation, assuming initial δ18O and δD values of ~1h and 8h, respectively,
75% equilibrium between atmospheric vapour and fresh water, and gypsum saturation occurring at
an evaporation ratio of 3.8× (evaporation ratio = weight of H2O of original seawater divided by
weight of H2O in residual evaporated brine). Mixing lines in B and C are derived from the model of
Gázquez et al. (2018) at a humidity of 75% and a starting point of modern Mediterranean seawater.
The mixing curve does not match the data, indicating that basin recharge with freshwaters was
dominant. Errors (1σ ) are less than the width of the symbols, and ±6 for 17O-excess.
sections, although he incorrectly correlated the cycles to the Los Molinos section and does
not log the discontinuous gypsum layer above the uppermost supercone cycle.
The uppermost part of the Los Molinos section contains two further gypsum-carbonate
cycles. The carbonate-marls below the uppermost gypsum layer display signs of slumping.
The uppermost cycle contains massive, metre-thick intervals of selenite interbedded with
distinct layers of carbonate marls (up to 30 cm thick), whereas the two gypsum layers below
are composed of discontinuous gypsum beds, often interbedded with limestone. As the
two uppermost cycles of the Los Molinos section are not present in the Rio de Aguas and
Barranco del Infierno sections, the beds are unlikely to be laterally continuous across the
basin.
The Caltanissetta Basin, Sicily: We compare our samples from the Sorbas Basin to the
samples collected from the PLG units of the Caltanissetta Basin (Fig. 6.1). The PLG units
occur as large disarticulated units that are detached from their original substratum, but still
preserve most of their original sedimentary features (Caruso et al., 2015). The chaotic nature
of the blocks has been interpreted either as redeposition of previously intact PLG units from
small northern sub-basins in the central foredeep via gravity flows (Roveri et al., 2006a),
or the result of the combined effects of the salt deformation under regional compressive
events and halokinesis followed by collapse dissolution (Rouchy and Caruso, 2006). The
Monte Banco (37°26’30”N 13°29’00”E), Monte Grotticelle (37°26’10”N 13°29’30”E) and
Santa Elisabetta sections display differing numbers of cycles and limited interbedded car-
bonates owing to the reworking. The stacking pattern is discussed in Section 4.2 and can be
correlated with the PLG of the Sorbas basin (e.g. Lugli et al., 2010).
154 Meteoric water, orbital and eustatic control of basin hydrology
-60
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3Cycle
21 3 4 5 6 7 8 9 10
1
1
1
2
1
5
1
3
1
5
1
4
-40
-20
0
Height (m)
A
g
e
 (M
a
)
0.0
-2.0
δ
1
8O
 
(‰
 V
S
M
O
W
)
O
b
liq
u
ity
 
E
T
P
A
rid
W
e
t
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
d
-e
x
c
e
s
s
(‰
 V
S
M
O
W
)
SMOW
5.6
5.7
5.8
5.96
6.1
-6
-3
0
3
6
-0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
2
2.5
3
3.5
2.0
4.0
-12.0
-4.0
0.0
-8.0
0
50
100
150
1
7O
-e
x
c
e
s
s
 
(p
e
r m
e
g
)
S
a
lin
ity
 (p
p
t)
T
G
3
2
T
G
2
6
T
G
2
2
T
G
2
0
M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
5
0
1
0
0
1
5
00
Gypsum Supercones
Non-evaporated 
seawater
Seawater 
evaporated to 
gypsum saturation
(La
ska
r e
t a
l., 2
0
0
4
)
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
Fig. 6.3. Mean intra-cycle δ18O, 17O-excess and d-excess data (calculated from measurement of GHW) and
salinity data (deduced from fluid inclusions and calculated using the equation of Goldstein and
Reynolds, 1994) are plotted against their designated cycle. Error bars define ±1σ of the data in
each cycle. Each gypsum bed of the Yesares Member is plotted in relation to its proposed orbital
tuning (Laskar et al., 2004; Lugli et al., 2010; Manzi et al., 2013), where gypsum precipitation
correlates with insolation minima. The benthic δ18O (Hodell et al., 2001; modified by Drury et al.,
2018) is also shown, and the major glacial stages are indicated. The horizontal grey band indicates
the appearance of gypsum supercone cycles (cycles 6-12).
6.3 Sample collection methods 155
6.3 Sample collection methods
In the Sorbas Basin, individual in situ gypsum crystals (n = 135) were removed with a
hammer and chisel. Cycles 2, 7, and 10 were subject to high-resolution sampling, where >22
discrete, evenly spaced gypsum samples were analysed in each cycle. Carbonate samples
(n = 129) were taken using a battery-powered drill fitted with a coring drill bit. The corer
was inserted parallel to the bedding plane after removal of the outer 5 cm of marl to avoid
contamination. In the Caltanissetta Basin, gypsum samples (n = 24) were collected from
Monte Grotticelle and Monte Banco using the same techniques.
6.4 Results
6.4.1 Gypsum Hydration Water (GHW)
One hundred and fifty-nine samples representing massive, banded and branching selenite
were analysed (Appendix D; Table D.1). The oxygen and hydrogen isotope composition
of the parent water in which the gypsum formed is calculated by correcting the measured
values for the gypsum-water fractionation factors (Chapter 2). Fractionation factors were
assumed to be 1.0035 and 0.981 for oxygen and hydrogen isotopes, respectively, which
represents a water temperature of ~15°C to 20°C and salinity of <150 g l-1 (Chapter 5). The
sensitivity of the fractionation factors to temperature is low, especially for oxygen isotopes
(Gázquez et al., 2017). Following correction, the δ17O, δ18O and δD of the gypsum mother
water for the Sorbas Basin PLG are shown in Fig. 6.2.
Inter-cycle trends: The range of δ18O and δD does not vary greatly among gypsum
cycles, with little change up-section (Fig. 6.3). Mean (±σ ) δ18O and δD values across
the entire section are 0.99 (±1.5)h and 1.1 (±9.6)h, respectively. The lowest mean δ18O
and δD values occur in cycle 5 (δ18O = -1.1h; δD = -13.3h) and cycle 11 (δ18O = -0.1h;
δD = -6.2h). The highest mean intra-cycle values occur in cycle 8 (δ18O = 2.1h; δD =
9.0h) and cycle 13 (δ18O = 1.9h; δD = 7.1h).
Deuterium excess is also relatively constant across all gypsum cycles with a mean (±σ )
value of -6.8 (±3.0)h (Fig. 6.3). 17O-excess is -44 (±6) per meg at the base of cycle 1, but
rises to a mean value of -19 (±9) per meg over cycles 2 to 11. In cycles 12 to 15, 17O-excess
returns to lower values (-37 (±13) per meg), similar to cycle 1. d-excess and 17O-excess
are also negatively correlated with δ18O and δD; the most positive mean intra-cycle values
occur in cycle 5 (d-excess = -1.1h; 17O-excess = -13 per meg) and cycle 12 (d-excess =
-0.1h; 17O-excess = -16 per meg).
156 Meteoric water, orbital and eustatic control of basin hydrology
H
e
ig
h
t 
(m
)
-5.0 0.0 5.0
δ18O (‰ VSMOW)
0.7089 0.7090
87Sr/86Sr
0.0 0.4 0.8 1.2
44/40Ca (‰ BSE)
0.0 2.0 4.0
Sr/Ca
20.0 22.5 25.0
δ34S (‰)
10.0 12.5 15.0
δ18OSO4 (‰)
1
2
3
4
5
6
7
8
9
10
11
12
13
15
14
δ
Marls
Gypsum Supercones
Banded Gypsum
Massive Gypsum
50
100
150
0
Fig. 6.4. The δ18O deduced from GHW, strontium isotopes, calcium isotopes, bulk Sr/Ca ratio, sulfur and
oxygen isotopes of sulfate of individual samples are shown relative to their sample location in
relation the Yesares section.
6.4 Results 157
-5.0 0.0-10.0
Marls
Gypsum Supercones
Banded Gypsum
Massive Gypsum
0.0 0.5 1.0 0 50 100 150 200
C
y
c
le
 2
δ44/40Ca (‰)
0m
5
10
15
Salinity (ppt) d-excess (‰)
0.7089 0.7090 -2.0 0.0 2.0 4.00.7091
1
0m
1
5
0m
1
4
6
0m
1
2
3
4
5
87Sr/86Sr δ18O (‰)
C
y
c
le
 1
0
2
3
5
2
3
4
6
7
8
9
6
C
y
c
le
 7
C
y
c
le
 1
2
G
lo
b
a
l 
O
ce
a
n
C
a
 S
e
a
w
a
te
r
Fig. 6.5. Schematic diagram of cycles 2, 7, 10 and 12 scaled to height of each of the sections. Calcium
isotope data, strontium isotope data, salinity data (deduced from fluid inclusions and calculated
using the equation of Goldstein and Reynolds, 1994), δ18O and d-excess data deduced from GHW
are plotted relative to the height from which the samples were taken. Dashed lines are a guide for
the eye, highlighting the similarity in δ18O values in cycles 7, 10 and 12.
158 Meteoric water, orbital and eustatic control of basin hydrology
Intra-cycle trends: On the basis of oxygen and hydrogen isotope compositions, the Sorbas
section can be divided into three parts: (i) a lower section (cycles 1 to 4) with homogeneous
isotope trends within cycles, (ii) a middle section beginning with cycle 5, where cycles
display systematic, decreasing upward trends of δ18O and δD values, and (iii) an upper
section beginning with cycle 13, where cycles display systematic, increasing upward trends
of δ18O and δD values (Fig. 6.4). The middle section coincides with the appearance of
banded selenite and branching supercone structures within the cycles.
The δ18O, δD, 17O-excess and d-excess of gypsum mother water within cycles 1 to 4
exhibit low amplitude oscillations, with no clear intra-cycle trends (Fig. 6.4; Fig. 6.5).
These beds consist of massive selenite with little change in facies.
Cycles 7, 10, 11 and 12 display a clear and systematic inter-cycle change in δ18O and δD
values (Fig. 6.4; Fig. 6.5). The δ18O and δD values increase upwards from the base of the
section to maximum values in the gypsum palisade at ~50 cm to 100 cm. Above this peak,
δ18O and δD fall to their lowest values at the base of the supercone structures. Isotope values
in the supercones are more variable, but a small cyclic trend is apparent where the δ18O and
δD increase to more positive values ~150 cm above the base of the supercones, and then
fall to lower values at the top of the cycle. Overall, these cycles display a systematic change
of increasing δ18O and δD from the base of the section that reach maximum values in the
lower-middle (palisade gypsum), and a fall to low isotopic values near the top (supercone).
These trends are also apparent in cycles 5 and 6, despite the comparatively low sampling
resolution.
Cycles 13 and 15 show a general trend towards higher δ18O and δD values towards the
top of each cycle (Fig. 6.4). Cycle 13 is composed of a discontinuous bed of massive selen-
ite, whereas cycle 15 is composed of massive, metre-thick intervals of selenite interbedded
with distinct layers of carbonate marls (up to 30 cm thick) and topped by domed gypsum
layers.
6.4.2 Sr isotopes
Seventy-one samples from Messinian gypsum and 12 samples from the interbeded carbon-
ates from the Sorbas Basin were processed for strontium isotopic analysis (Appendix D;
Table D.4). For gypsum samples, the 87Sr/86Sr ranged between 0.708916 and 0.708979.
These values fall slightly below the strontium isotope range expected for Messinian seawa-
ter, which varies from ~0.708989 at 5.971 Ma to ~0.709014 and 5.6 Ma relative to a NIST
SRM 987 value of 0.710248 (McArthur et al., 2001).
6.4 Results 159
Inter-cycle trends: The lowermost gypsum cycle (cycle 1) has 87Sr/86Sr values that are
indistinguishable from seawater (difference in mean 87Sr/86Sr from seawater = 0.000016),
but the difference increases up-section to a maximum departure from the global seawater
curve in cycles 10 and 11 (difference in mean 87Sr/86Sr from seawater = 0.000068) (Fig.
6.6). Overall, mean 87Sr/86Sr values decrease away from the seawater curve up-section,
but this trend reverses at cycles 7 and 12 with a slight recovery towards the global ocean
seawater curve. This defines a ~100 ka cyclicity within the 87Sr/86Sr profile, and transition
from cycles 6 to 7 and 11 to 12 correlate temporally with periods of eccentricity minima.
Intra-cycle trends:
The 87Sr/86Sr values of gypsum analysed of cycles 1 through 4 display very little intra-cycle
variation; isotope values display a slight variance throughout the gypsum bed, but there is
no significant correlation towards higher or lower values up-section within each layer (Fig.
6.4; Fig. 6.5). Cycles 5 through 15 display a significant and systematic change in 87Sr/86Sr
within each cycle; the lowest Sr isotope ratios are recorded at the base of each gypsum layer,
and values increases up-section toward maximum values the top of each cycle. No sharp
changes in the Sr isotopic composition of the gypsum are observed at the transition between
gypsum facies.
The range of 87Sr/86Sr for the interbedded carbonates (n = 9) varies between 0.708936
and 0.708423, with an average value of 0.708720 (Appendix D; Table D.8). These values
are significantly lower than 87Sr/86Sr of the massive gypsum layers. The 87Sr/86Sr for the
carbonate marls juxtaposed to supercones (n = 3) ranges from 0.709025 to 0.708991, with
an average value of 0.709011, which is within the range of the Sr isotopic composition of
the surrounding gypsum.
6.4.3 Ca isotopes
The results of Ca isotope analysis are reported in Appendix D (Table D.4). For gypsum
samples (n = 75), the range of δ 44/40Ca varies between 0.30h and 1.16h.
Inter-cycle trends: The lowest δ 44/40Ca values are found in gypsum cycles 1 to 5, where
mean (±σ ) values of all 5 cycles are 0.53 (±0.13)h (Fig. 6.6). A step change in Ca isotopes
occurs at cycle 7. Mean δ 44/40Ca of cycles 7 to 11 are 0.79 (±0.10) h. Up-section, Ca
isotope values of cycle 12 also display a stepped increase compared to the lower cycles,
with mean δ 44/40Ca values of 1.01 (±0.10)h. Calcium isotope values of cycles 13 to 15
(0.68 ± 0.10 h) return to similar values as cycles 7 to 11. Overall, δ 44/40Ca of gypsum
160 Meteoric water, orbital and eustatic control of basin hydrology
increases up-section, with step changes defining a ~100 ka cyclicity similar to those seen in
87Sr/86Sr results.
The range of δ 44/40Ca for the interbedded carbonates (n = 19) varies between -0.35h
and 0.46h, and average intra-cycle δ 44/40Ca decreases up-section (Appendix D; Table
D.8). The δ 44/40Ca for the carbonate marls juxtaposed to supercones ranges from 0.18h to
0.21h.
Intra-cycle trends: The lowest Ca isotope values in each cycle are observed at the base
of the bed. In general, all cycles display a trend towards higher δ 44/40Ca toward the top of
the bed, but this correlation is weak. Ca isotopes exhibit low amplitude oscillations in all
cycles. There is no correlation between Ca isotope variability and gypsum facies or grain
size (Fig. 6.4).
6.4.4 δ 34S and δ18O of sulfate
Fifty-two samples were analysed for sulfur isotopes and oxygen isotopes of sulfate (Ap-
pendix D; Table D.6). δ 34S of gypsum from the Sorbas Basin show a narrow range of values
between 21.1h and 23.3h VCDT, and δ 18OSO4 varies between 10.4 and 14.9h VSMOW.
These values fall on or slightly above the δ 34S and δ 18OSO4 isotope range expected for
sulfate minerals precipitated from evaporated Late Miocene seawater (i.e. δ 34S ~22h and
δ 18OSO4 ~12h, relative to NBS-127 values of 21.1h and 8.6h for δ 34S and δ 18OSO4 ,
respectively) (Fig. 6.7).
Inter-cycle trends: Gypsum δ 34S values remain homogenous across the entire section,
with mean (±σ ) values of 22.4 (±0.4)h (Fig. 6.7). The δ 18OSO4 is also relatively unchan-
ging across the section, displaying mean (±σ ) values of 12.9 (±1.0)h. From cycles 1 to 12,
there is a slight trend toward lower δ 18OSO4 values up-section, although cycle 15 reverses
this trend and displays anomalously high values of 13.9 (±0.5)h.
Intra-cycle trends: Sulfate isotope compositions are also relatively homogenous within
individual cycles of the Sorbas Basin (Fig. 6.4). There are negligible differences in the
isotope composition among the nucleation cone, palisade, banded selenite, and supercone
facies, and changes in δ 34S and δ 18OSO4 with depth are not repeatable between cycles.
The δ 34S and δ 18OSO4 of gypsum from the Caltanissetta Basin show similar values to
those found in the Sorbas Basin, where δ 34S and δ 18OSO4 vary between 22.6h and 22.9h
VCDT and 12.7h and 14.6h VSMOW, respectively.
6.4 Results 161
6.4.5 Microthermometric analysis
Thirty-four samples from the Messinian gypsum of the Sorbas Basin, 4 samples from the
Caltanissetta basin, and 2 samples from the modern Salina at Cabo de Gata, Almería, Spain,
were selected for microthermometric analysis (Appendix D; Table D.6).
The final melting temperature (Tmice) of 405 primary Messinian inclusions ranged between
-0.2°C and -6.1°C, with an average Tmice of -3.2°C (Fig. 6.8). Using the revised equation
and table for determining the freezing point depression of H2O–NaCl solutions of Bodnar
(1993), average salinity estimates of individual samples range from 2.9% to 8.5% (aver-
age 5.3 wt% NaCl equivalent). Tmice of modern gypsum from the Cabo de Gata Salina
ranged between -6.5°C and -10.1°C (average -8.0°C), corresponding to an average salinity
of 11.7%.
Within the section, the average wt% NaCl equivalent of gypsum from cycles 1 to 9 is
5.1 (±1.1) % (Fig. 6.3). These values are well below the threshold for gypsum precipitation
from the evaporation of seawater alone. The mean wt% NaCl of cycle 11 is 7.2 (±1.2) %,
representing a significant increase compared to the lower cycles.
6.4.6 XRD, wt% calcite and wt% dolomite of bulk carbonates
Eighty five marl samples were analysed by XRD (Appendix D; Table D.8). The results
are semi-quantitative and indicative of the relative abundances of common mineralogical
components only. In addition to clay, the major phases are gypsum, calcite, and dolomite.
The marls interbedded with the lower 7 cycles contain dominantly calcite. The marls
at the base of cycle 10 contain both calcite and dolomite (Fig. 6.6). The ratio of calcite to
dolomite in the lower marls of cycle 10 decreases up-section such that dolomite becomes
more dominant toward the gypsum–marl contact. Fine-grained gypsum within the marl is
abundant in the upper marl from 8 cm below the marl-gypsum contact. Between cycles 10
and 11, marls are dominantly composed of dolomite with minor amounts of calcite. Marls
juxtaposed to supercones in cycles 10 and 11 are purely dolomitic and contain varying
proportions of gypsum.
6.4.7 δ18Ocarb and δ13C of bulk carbonates
The oxygen and carbon isotope composition of 129 carbonate samples vary from -8.3h to
5.8h for δ18Ocarb and from -4.9h to 2.0h for δ13C (Appendix D; Table D.8; Fig. 6.6 and
Fig. 6.9). The carbon and oxygen isotope composition of the carbonates are within the range
of other Messinian carbonates associated with evaporites in the Mediterranean (Longinelli,
162 Meteoric water, orbital and eustatic control of basin hydrology
0
.7
0
8
9
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3Cycle
21 3 4 5 6 7 8 9 10
1
1
1
2
1
5
1
3
1
5
1
4
-8.0
-6.0
0.0
4.0
D
o
l:C
a
l R
a
tio
1.0
0.8
0.6
0.4
0.2
0.0
M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
5
0
1
0
0
1
5
00
Gypsum Supercones
0
.7
0
8
9
5
0
.7
0
9
0
.7
0
9
0
5
8
7S
r/ 8
6S
r
Height (m)
A
g
e
 (M
a
)
0
.0
0
.4
1
.2
0
.8
δ
4
4
/
4
0C
a
 
(‰
, B
S
E
)
Global Ocean
O
b
liq
u
ity
 
E
T
P
A
rid
W
e
t
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
δ
1
8O
 (‰
, V
P
D
B
)
Ca Seawater
5.6
5.7
5.8
5.96
6.1
-6
-3
0
3
6
-0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
2
2.5
3
3.5
T
G
3
2
T
G
2
4
T
G
2
2
T
G
2
0
TG20
TECTONICSINCREASING 
MEDITERRANAEN-ATLANTIC 
RESTRICTION
In
cre
a
se
d
 
R
u
n
n
o
ff/
D
e
cre
a
se
d
 
S
e
a
w
a
te
r In
p
u
t
D
e
cre
a
se
d
 
R
u
n
n
o
ff/
In
cre
a
se
d
 
S
e
a
w
a
te
r In
p
u
t
(La
ska
r e
t a
l., 2
0
0
4
)
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
6.5 Discussion 163
Fig. 6.6. Left. Mean intra-cycle 87Sr/86Sr and δ44/40Ca data from gypsum beds, and δ18Ocarb and dolomite-
calcite ratio of the interbedded carbonates are plotted against their designated cycle. Error bars
define ±1σ of the data in each cycle. Note that carbonate data plotted between the dashed
lines denote carbonate samples located between gypsum beds, whereas carbonate samples that
are plotted on the dashed lines denote carbonate samples that are located in supercones. Each
gypsum bed of the Yesares Member is plotted in relation to its proposed orbital tuning (Laskar
et al., 2004; Lugli et al., 2010; Manzi et al., 2013), where gypsum precipitation correlates with
insolation minima. The benthic δ18O (Hodell et al., 2001; modified by Drury et al., 2018) is also
shown, and the major glacial stages are indicated. The light grey horizontal grey band indicates
the appearance of gypsum supercone cycles (cycles 6-12). The dark grey horizontal band highlight
the glacial period TG20. The blue arrow highlights the period of increased cooling and global
sea-level fall through TG24, TG22 and TG20. The orange arrow indicates the expected, gradual
influence of the tectonic restriction of the Mediterranean-Atlantic exchange of the isotope systems
shown. The black arrows highlight key trends in 87Sr/86Sr, δ44/40Ca, δ18Ocarb and dolomite-calcite
data. Note that there is significant correlation between shifts in 87Sr/86Sr, δ44/40Ca, δ18Ocarb and
dolomite-calcite data and glacial TG24, TG22 and TG20.
1979). The δ18Ocarb of the marls from cycles 1 to 7 average -6.1 (±1.4)h, whereas the δ13C
averages -0.8 (±1.4)h. At the base of cycle 10, mean (±σ ) δ18Ocarb and δ13C increase to
-1.2 (±1.9)h and 0.6 (±1.9)h, respectively. Marls juxtaposed to supercones in cycles 10
and 11 show the highest δ18Ocarb values, but the lowest δ13C values, averaging 3.1 (±0.2)
h and -4.4 (±0.4)h, respectively. Additionally, the marls between cycle 10 and 11 show
high δ18Ocarb and low δ13C values, averaging 4.1 (±1.4)h and -1.6 (±2.1)h, respectively.
In this cycle, negative excursions in the δ18Ocarb are well correlated with an increase in the
calcite:dolomite ratio. Above cycle 12, the δ18Ocarb and δ13C return to values similar to
those of cycles 1 to 7, averaging -5.8h and -1.1h, respectively.
6.5 Discussion
The causes and mechanisms leading to deposition of the PLG remain controversial. Out-
standing problems include (i) the composition and salinity of brines that formed the gypsum
and carbonate layers, and (ii) the relative influence of orbital forcing, Mediterranean-Atlantic
exchange dynamics, and eustatic sea-level variability during evaporate deposition. The
majority of previous studies have utilised strontium and sulfur isotopes to constrain the water
budget during evaporate formation; however, because these tracers are relative insensitive
to meteoric input, the multiple factors that can influence basin hydrology have been difficult
to deconvolve. Below we discuss our multi-isotope approach to address the outstanding
questions of PLG formation and propose an explanation to reconcile the geochemical data.
164 Meteoric water, orbital and eustatic control of basin hydrology
1
0
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3Cycle
21 3 4 5 6 7 8 9 10
1
1
1
2
1
5
1
3
1
5
1
4
M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
5
0
1
0
0
1
5
00
Gypsum Supercones
Height (m)
A
g
e
 (M
a
)
2
0
δ
3
4S
 
(‰
)
O
b
liq
u
ity
 
E
T
P
A
rid
W
e
t
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
5.6
5.7
5.8
5.96
6.1
-6
-3
0
3
6
-0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
2
2.5
3
3.521
2
2
2
3
2
4
1
1
1
2
1
3
1
4
1
5
δ
1
8O
S
O
4
(‰
)
T
G
3
2
T
G
2
4
T
G
2
2
T
G
2
0
(La
ska
r e
t a
l., 2
0
0
4
)
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
Fig. 6.7. Mean intra-cycle δ 34S and δ18OSO4 data are plotted against their designated cycle. Error bars define
±1σ of the data in each cycle. Each gypsum bed of the Yesares Member is plotted in relation to its
proposed orbital tuning (Laskar et al., 2004; Lugli et al., 2010; Manzi et al., 2013), where gypsum
precipitation correlates with insolation minima. The benthic δ18O (Hodell et al., 2001; modified by
Drury et al., 2018) is also shown, and the major glacial stages are indicated. The horizontal grey
band indicates the appearance of gypsum supercone cycles (cycles 6-12). The vertical grey band
denotes the range of Late Miocene marine sulfates (Paytan et al., 1998).
6.5 Discussion 165
6.5.1 Brine composition during PLG deposition
The commonly accepted palaeo-hydrologic models for the formation of the PLG suggest
that gypsum precipitated from the evaporation of brines formed by continuous inflow of
seawater from the Atlantic Ocean, providing the necessary ion supply (Krijgsman and
Meijer, 2008; Topper and Meijer, 2013). Water isotopes of GHW (e.g. Evans et al., 2015)
and salinities determined from fluid inclusions (e.g. Natalicchio et al., 2014) suggest that
gypsum precipitation may have occurred at lower salinities than those expected during the
evaporation of only seawater (~130 g l-1). We applied the approach of Evans et al. (2015) to
12 of the 15 Sorbas PLG cycles.
The δ18O, δ17O and δD of the water from which gypsum formed ranges between -
1.71h to 2.27h, -3.23h to 4.38h and -26.76h to 23.48h, respectively. If the Yesares
Gypsum were to have formed directly from the evaporation of seawater, the expected results
would be significantly higher than those measured, probably similar to the evaporated sea-
water that formed the Salinas gypsum (Fig. 6.2). Additionally, microthermometric salinity
data indicate that Messinian fluid inclusions do not have salinities high enough to be within
the range of gypsum saturation if derived from the evaporation of seawater alone. This im-
plies that all cycles of the Yesares Gypsum were precipitated from a hybrid brine, consisting
of a mixture of seawater and meteoric water. Analysis of the evaporation trajectories (i.e.
δ18O vs d-excess and δ18O vs 17O-excess) and isotope mass-balance (Gázquez et al., 2018)
demonstrates that the recorded isotopic composition cannot be formed from the simple
evaporation of water in a closed basin, where evaporating water evolves along a trajectory
towards the isotopic end-point (Surma et al., 2018). Basin recharge (mixing) drives the
water’s isotopic composition along a trajectory below the expected evaporation trend (Fig.
6.2), as has been demonstrated to occur in other modern hydrologic systems (Herwartz et
al., 2017; Surma et al., 2018). Basin recharge by meteroric water is the only mechanism
that can produce the observed water isotope variability for the Sorbas PLG, suggesting that
gypsum formed at a steady state with continuous inflow of freshwater.
The significant freshwater contribution to the mother brines recorded by GHW and
salinity deduced from fluid inclusions is confirmed by strontium isotope data obtained from
the same samples. Our results from the PLG of the Sorbas Basin show that 87Sr/86Sr values
are significantly lower than global seawater values, with an increasing departure from global
ocean 87Sr/86Sr values up-section (Fig. 6.6). Given the uncertainties on both the riverine
strontium concentration and isotopic ratio during the Messinian, early studies concluded that
these near-oceanic 87Sr/86Sr values required the PLG to precipitate from water dominated
by seawater (e.g. Flecker et al., 2002; Lu and Meyers, 2003; Flecker and Ellam, 2006; Lugli
et al., 2007). Indeed, Evans et al. (2015) argued that if a significant amount of strontium and
166 Meteoric water, orbital and eustatic control of basin hydrology
-1.00.0
Tmice (°C)
0
10
20
30
40
50
60
70
N
u
m
b
e
r 
o
f 
In
cl
u
si
o
n
s
-2.0 -3.0 -4.0 -5.0 -6.0 -7.0
0.0 2.0 4.0 6.0 8.0 10.0
Salinity (wt % eq. NaCl)
Freshwater saturated 
for gypsum
Non-evaporated 
seawater
Evaporated 
seawater-freshwater 
mixture
Evaporated seawater to 
gypsum saturation
80
Fig. 6.8. Histogram of Tmice measured in PLG fluid inclusions.
sulfate used to make the PLG were derived from terrestrial sources, the 87Sr/86Sr, δ 34S and
δ 18OSO4 composition of the gypsum would fall toward lower riverine values (Chapter 5).
Simple mixing simulations using the Messinian marine δ 34S (~22h, 2600 ppm sulfate) and
the weighted modern major Mediterranean river inputs (~0h, ~30 ppm sulfate) (Longinelli
and Cortecci, 1970) indicate that a dilution of over 90-95% of the Mediterranean Sea is
required to shift the δ 34S away from the marine value. Sulfate isotopes are therefore highly
insensitive to continental contributions. Additionally, Topper (2013) demonstrated that
when Mediterranean river discharge constitutes around 40% of the water fluxes into a basin,
the basinal 87Sr/86Sr value will deviate measurably from relatively high oceanic values
(~0.70900) toward lower riverine values of the Rhone (~0.70872) and the Nile (~0.70600)
(Flecker and Ellam, 2006). The high precision of our Sr isotope results allow us to con-
firm that gypsum precipitation occurred in a water body strongly influenced by freshwater
input. The increased departure of measured 87Sr/86Sr from global ocean values indicates
a progressive reduction of marine inflow and the increasing influence of continental input
up-section.
The high proportion of freshwater inflow into the Mediterranean combined with surface
evaporation leading to a denser fluid that flowed towards the deeper parts of the basin
6.5 Discussion 167
would have led to significant water column stratification (Simon and Meijer, 2017). Indeed,
homogenous δ 34S and δ 18OSO4 of gypsum of both the Yesares Member (21.1h < δ 34S
< 23.3h; 10.4h < δ 18OSO4 < 14.9h) and Caltanissetta Basin (22.4h < δ 34S < 23.2h;
12.7h < δ 18OSO4 < 14.6h) are higher than expected for Late Miocene seawater (δ 34S ~
21h; δ 18OSO4 ~ 12h), likely driven to higher values by cyclic bacterial sulfate reduction
in a stratified water column (Fig. 6.7) (De Lange and Krijgsman, 2010; García-Veigas et
al., 2018). A scenario of freshwater inflow resulting in water column stratification is further
supported by evidence of a changing carbon source in the Sorbas Basin: low δ13C values
in some interbedded carbonates indicate that freshwater inputs resulted in an increase in
primary production in the surface water supplying the primary carbon source for microbial
sulfate reduction in the water column (Torfstein and Turchyn, 2017) (Fig. 6.9).
6.5.2 Orbital, tectonic and glacio-eustatic control on basin hydrology
Our high resolution, multi-isotope approach allows us to more accurately link isotopic vari-
ations to the climate-driven hydrological changes within a single precessional cycle. Within
the Sorbas Basin PLG, oxygen and hydrogen isotopes of GHW, 87Sr/86Sr and δ 44/40Ca
display distinctive patterns of isotopic variability both (i) within individual gypsum-marl
cycles, related to the arid-humid climate oscillations controlled by precessional forcing, and
(ii) over longer time scales in relation to eustatic sea-level changes superimposed on longer-
term tectonic restriction and the amplitude modulation of precession by eccentricity.
6.5.2.1 Precessional scale orbital and tectonic control on basin hydrology
Gypsum deposition in the PLG is thought to occur at the precession maxima (i.e. minimum
summer insolation when summer solstice and aphelion coincided) when evaporation ex-
ceeded precipitation during periods of dry climate, whereas marl deposition is assumed to
have occurred at precession minima (maximum summer insolation when summer solstice
and perihelion coincided) when rainfall increased (Krijgsman et al., 1999, 2001). Within
a depositional cycle, variations in water isotopes (Evans et al., 2015) and 87Sr/86Sr (e.g.
Flecker and Ellam, 2006; Lugli et al., 2010; Topper et al., 2014; Reghizzi et al., 2018) have
been shown to correlate with the changing freshwater budget forced by orbital precession.
Wide sampling intervals and lack of results from the interbedded carbonate layers from
previous studies have resulted in tenuous links between measured isotopic variations and
climatic forcing.
Firstly, we are able to confirm the inference that carbonate marl deposition occurred
at precession minima when rainfall increased (Krijgsman et al., 1999, 2001; Lugli et al.,
168 Meteoric water, orbital and eustatic control of basin hydrology
-6.0
-5.0
-4.0
-3.0
-2.0
-1.0
0.0
1.0
2.0
3.0
-10.0 -8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 10.0
Cycles 1-8
Cycle 10
Cycle 10 Supercone
Cycle 11
Cycle 11 Supercone
Cycle 14
Cycle 15
Su
pe
rco
ne
s
δ1
3 C
 (
‰
)
δ18Ocarb (‰)
Mixing 
+ Evap
oration?
Re
du
cin
g?
Fig. 6.9. Cross-plot of δ18O vs δ13C of carbonates from the marl intervals within the Yesares Formation.
2007; Roveri et al., 2014a). By isolating the carbonate phase from the surrounding clays
(see Chapter 4), we could subsequently measure the Sr and Ca isotopes to identify the
isotopic composition of the mother fluids from which they formed. The 87Sr/86Sr results
of the carbonate marls demonstrate that these horizons are always lower than those of the
interbedded gypsum layers. These data agree with Sr isotope analysis of mollusc shells that,
with one exception, yield values equal to or lower than that of gypsum measured in the same
section (Reghizzi et al., 2017). Additionally, δ 44/40Ca values of the interbedded carbonates
are generally lower than the δ 44/40Ca of the gypsum (both before and after correction for
fractionation factors), strongly suggesting that the freshwater flux to the basin was greater
during the deposition of the carbonate marls (Fig. 6.10).
During the climatic transition from precession minimum to precession maximum, river
input is expected to decrease while E/P increases, leading to the precipitation of gypsum
horizons. In Chapter 5, it was shown that, in a gypsum single cycle, δ18O and δD of
GHW increased from low values at the base of the cycle to a maximum in the massive
gypsum palisade, and decreased again to lower values in the supercones at the top of the
cycle. This pattern, together with changes in mineralogy in the carbonates at the base of the
cycle (calcite–dolomite–gypsum), is consistent with a precession-driven change in climate.
Additionally, box modelling has shown that 87Sr/86Sr is expected to increase toward global
ocean values during the arid phase as the inflow from the riverine input of Sr decreases
during the precession maximum phase (Topper et al., 2014).
We have identified important differences in the thickness and isotopic composition of
the gypsum in relation to the appearance of differing facies within the PLG. Cycles 1-5
6.5 Discussion 169
are composed of thick homogenous massive selenite deposits; our GHW data show that
δ18O and δD display no trends within the cycles and do not mirror the cyclic trend observed
previously in cycle 10 (Chapter 5; Evans et al., 2015). From the first occurrence of branching
supercones in cycle 6, however, the cyclic trends in δ18O and δD become regular and
highly consistent (Fig. 6.4). Strontium concentrations (e.g. Rosell et al., 1998) and salinity
deduced from fluid inclusions in gypsum (Fig. 6.3) also show relatively uniform values in
the lower beds (i.e. cycles 1 to 4), but more variable values that often increase upwards in
the sixth gypsum cycle and above.
The δ 44/40Ca and 87Sr/86Sr values obtained from the PLG also display consistent vari-
ability within each gypsum layer. The lowest δ 44/40Ca and 87Sr/86Sr values are usually
found at the base of each gypsum bed, although the standard deviation of both δ 44/40Ca
and 87Sr/86Sr within each cycle is low (Fig. 6.4 and Fig. 6.5). Importantly, there is no
significant relationship between δ 44/40Ca and crystal size, and thus the growth kinetics
cannot be considered as the main controlling factor of δ 44/40Ca in the selenitic gypsum
facies. The low intra-cycle isotopic variability in the lower gypsum beds (i.e. cycles 1 to
4) results in a weak correlation between δ 44/40Ca and 87Sr/86Sr within each cycle. This
lack of correlation suggests that the slight increase in δ 44/40Ca from the base to the top of
individual cycles could reflect ongoing Rayleigh fractionation in the brine during gypsum
precipitation, rather than changes to the meteoric and/or marine input. This indicates that a
steady state was achieved after sufficient calcium was removed from the brine such that the
input flux was equivalent to the output precipitation. Thus, crystallisation took place beneath
a stable and protective pycnocline in marginal Mediterranean water bodies, resulting in near-
constant δ 44/40Ca and 87Sr/86Sr values within each cycle.
The upper cycles containing supercones (i.e. cycles 6 to 12) show a stronger correlation
between δ 44/40Ca and 87Sr/86Sr (R2 = 0.53), where low isotope values are usually found at
the base of each gypsum bed that increase to higher values at the top. These data suggest that
gypsum precipitation at the base of each cycle started from a water body strongly influenced
by freshwater input with a progressive weakening of riverine input towards the top of each
cycle. Isotopes of GHW, however, are relatively high in the massive selenite at the base of
the gypsum cycles and fall towards lower values in the supercone structures. This indicates
that the freshwater-dominated fluid that formed the massive gypsum at the base of each
cycle required relatively high levels of evaporation to reach gypsum saturation, resulting in
increased δ17O, δ18O and δD values. The supercone structures likely formed from fluids
that were composed of less meteoric input compared to the base of the cycle, but were
more enriched in calcium and sulfate ions, requiring less evaporative enrichment to reach
gypsum saturation. The necessary ionic input may have resulted from evaporite recycling
170 Meteoric water, orbital and eustatic control of basin hydrology
2
1
3
4
5
6
7
8
9
10
11
12
14
13
C
y
c
le
15
C
a
 S
e
a
w
a
te
r
αgypsum = 0.999
αcarbonate = 0.9989
2
1
3
4
5
6
7
8
9
10
11
12
15
13
14
Marls
Gypsum Supercones
Banded Gypsum
Massive Gypsum
50
100
150
0
H
e
ig
h
t 
(m
)
-0.08-0.0400.040.08
380 440 500 560
6.1
6.0
5.9
5.8
5.7
5.6
Age (Ma)
0.0
δ44/40Ca 
(‰, BSE)
0.5 1.0 1.5 2.0
Measured Carbonate
Measured Gypsum
Corrected Carbonate
Corrected Gypsum
Eccentricity
Climatic Precession 
Summer 
Insolation 65°N
Fig. 6.10. Mean intra-cycle δ44/40Ca data of the gypsum and interbedded carbonates are plotted against
their designated cycle. Error bars define ±1σ of the data in each cycle. Note that carbonate
data plotted between the dashed lines denote carbonate samples located between gypsum beds,
whereas carbonate samples that are plotted on the dashed lines denote carbonate samples that are
located in supercones. Measured values (open diamonds) are corrected for calcium fractionation
factors (Tipper et al., 2010; Harouaka et al., 2014). Note that the relative difference in gypsum
and carbonate δ44/40Ca increases up-section.
6.5 Discussion 171
from the lower cycles (Evans et al., 2015; Rossi et al., 2015), and confirms that gypsum
crystallisation was likely controlled by the oscillating level of a pycnocline (Manzi et al.,
2009; Lugli et al., 2010).
The influence of decreasing Mediterranean-Atlantic exchange driven by tectonic re-
striction can largely explain the intra-cycle isotopic profiles and changes in gypsum bed
thickness observed in the Sorbas section. The lower 4 cycles display relatively homogenous
isotopic profiles and massive gypsum deposits, reflecting deposition under conditions of
hydrological stability and relatively deep (~200 m) waters, whereas the more variable results
from the upper cycles and thinner gypsum beds result from a shallowing of the basin.
Tectonic modification of the Atlantic gateway directly controlled the seawater input to the
Mediterranean (Simon et al., 2017; Krijgsman et al., 2018), resulting in a reduction in
seawater input, and thus the reduction in gypsum bed thickness, up-section. If tectonics
were the only cause of the decrease in Mediterranean-Atlantic exchange and the patterns of
the expression (i.e. thickness and facies) of the gypsum cycles, however, a uni-directional
trend in facies and geochemical profile reflecting uplift and the gradual constriction of the
Mediterranean-Atlantic gateways would be expected. This is not observed: strontium iso-
topes in gypsum, for example, display a gradual, but non-unidirectional divergence from the
seawater curve up-section, with slight reversals in the general trend of decreasing 87Sr/86Sr
at ~5.84 Ma and 5.72 Ma (Fig. 6.6). Step changes in calcium isotopes track the variability
in 87Sr/86Sr. Additionally, the gypsum supercones facies are deposited during periods of
both high low eccentricity, indicating that their formation cannot be unilaterally related to
long-term changes in precession forcing. Thus, factors other than tectonics and precessional
scale orbital dynamics must be involved in the formation of the differing gypsum facies.
6.5.2.2 Glacio-eustatic control on basin hydrology
Global eustatic sea-level changes, as recorded by deep-sea benthic δ18O variations, po-
tentially influenced the connection between the Atlantic and Mediterranean (Hodell et al.,
2001). A comparison of δ 44/40Ca and 87Sr/86Sr in gypsum cycles and δ18O of the carbonate
marls with the benthic δ18O record of ODP Site 982 (Hodell et al., 2001; Drury et al., 2018)
provides information on sea-level control of the Mediterranean-Atlantic water exchange
during the time of PLG deposition (Fig. 6.6). We suggest that glacio-eustatic changes,
superimposed on longer-term tectonic restriction of the Mediterranean-Atlantic exchange,
were together responsible for some of the facies and geochemical changes observed in the
Sorbas Basin sequence.
A eustatic sea-level influence on PLG deposition was suggested by Manzi et al. (2013),
after re-correlation of the MSC-onset to an eccentricity minimum roughly coincident with
172 Meteoric water, orbital and eustatic control of basin hydrology
0.7089 0.70895 0.709 0.70905
87Sr/86Sr
Vena del Gesso Basin
Sorbas Basin
2
1
3
4
5
6
7
8
9
10
11
12
14
13
C
y
c
le
15
2
1
3
4
5
6
7
8
9
10
11
12
15
13
14
Marls
Gypsum Supercones
Banded Gypsum
Massive Gypsum
50
100
150
0
H
e
ig
h
t 
(m
)
-0.08-0.0400.040.08
380 440 500 560
6.1
6.0
5.9
5.8
5.7
5.6
Age (Ma)
Eccentricity
Climatic Precession 
Summer 
Insolation 65°N
Fig. 6.11. Sr isotope data of the PLG from Vena del Gesso Basin, Northern Italy (Reghizzi et al., 2018),
and the Sorbas Basin (this study) are calibrated with the orbital eccentricity and precession curves
(Laskar et al., 2004). Note that the data from the Vena del Gesso Basin displays a progressive
divergence from the oceanic curve up-section, whereas data from the Sorbas Basin show shifts
returning towards the seawater curve in PLG cycles 7 and 12.
6.5 Discussion 173
glacial stage TG32. It is interesting to note that TG32 is weakly expressed in the Site
982 record, potentially indicating that tectonic restriction was the main driving force that
initiated the start of the MSC (Fig. 6.6). However, the appearance of gypsum supercone
facies in cycle 6 and a stepped increase in δ 44/40Ca and 87Sr/86Sr during the transition from
cycle 6 to 7 can be linked to the start of a cooling phase, characterised by glacial conditions
in the northern hemisphere and global sea-level fall recorded by increasing benthic δ18O at
Site 982 (Lugli et al., 2010; Dela Pierre et al., 2011; Manzi et al., 2013). The step change in
our δ 44/40Ca and 87Sr/86Sr data is coincident with glacial stages TG 26 and TG24 and the
transition out of an eccentricity minimum phase, indicative of a reduction in both seawater
and freshwater input due to the modulation of precession by eccentricity. The resultant
change in sea level governed by both eustatic variability and eccentricity modulation of
precession likely resulted in the fluctuation of the pycnocline and the formation of branching
gypsum units.
The δ 44/40Ca and 87Sr/86Sr in both the gypsum and the interbedded carbonates progress-
ively diverge from global ocean values from cycle 7 to 11, during the period of increased
cooling and global sea-level fall through TG24, 22 and 20. The strong correlation between
δ 44/40Ca and 87Sr/86Sr in the upper part of the PLG (i.e. cycles 6 to 12) indicate that both
systems are reposing to changes in water source and an increase in the freshwater fraction
up-section. The departure of the 87Sr/86Sr from the global oceanic curve over these cycles
in the Sorbas PLG is also characterised by a similar evolution of 87Sr/86Sr in PLG deposits
from the Vena del Gesso basin (northern Italy; Reghizzi et al., 2018) (Fig. 6.11). Due to
gaps in the Vena del Gesso cycle analysis, however, the Sr profile does not display the slight
reversals in the general trend of decreasing 87Sr/86Sr up-section.
Glacial stage TG20, the largest glaciation during the deposition of the PLG, is expressed
both in variations in the isotopic composition of gypsum in cycles 9 to 11, as well as a
change in carbonate mineralogy during this period (Fig. 6.6). The 87Sr/86Sr of cycle 9 to
11 displays the greatest divergence away from the global ocean curve. Additionally, the
progressive increase in carbonate δ18O and transition from calcite- to dolomite-dominated
carbonate marls from cycle 9 to 11 indicate an highly restricted oceanic inflow during this
period, providing clear evidence that global sea-level variability influenced the mineralogy
and geochemistry of the gypsum-marl cycles. The terminal supercone cycle and the final
step change in δ 44/40Ca and 87Sr/86Sr coincides with the termination of TG20, indicat-
ing the occurrence slight seawater incursion after this phase (Fig. 6.6). Thus, the non-
unidirectional divergence of δ 44/40Ca and 87Sr/86Sr confirms that glacio-eustatic changes,
superimposed on longer-term tectonic restriction of Mediterranean-Atlantic exchange and
orbital forcing, were responsible for the differing geochemical profiles observed in the
174 Meteoric water, orbital and eustatic control of basin hydrology
-2
-1.5
-1
-0.5
0
0.5
1
1.5
0 20 40 60 80 100 120 140 160
48 62 75 89 104
138
Gypsum precipitation
42.1 mmol Ca/L (at 35 g/L)
21.1 mmol Ca/L   
15.8 mmol Ca/L 
12.6 mmol Ca/L 
10.5 mmol Ca/L (normal seawater) 
6.32 mmol Ca/L  
Salinity (g/l)
G
y
p
su
m
 S
a
tu
ra
ti
o
n
 (
S
I g
y
p
)
S
e
a
w
a
te
r 
Fig. 6.12. Graph showing the salinity at which the gypsum saturation (SIgyp) is reached during the
evaporation seawater containing different calcium [Ca2+] and sulfate [SO42−] concentrations.
Evaporation of normal seawater is represented by curve with 10.5 mmol l-1 Ca at normal seawater
salinity and shows that gypsum will precipitate at salinities of ~104 g l-1. Other curves represent
an addition or removal of calcium relative to the starting seawater concentration, where the stated
ratio of [Ca2+] to [SO42−] (i.e. x mmol l-1 Ca2+: 28.9 mmol l-1 SO42−]) increases/decreases at the
same rate.
Yesares Member.
After cycle 12 and TG20, global sea-level was likely higher, but the continued tectonic
uplift likely reduced the exchange to a point whereby thick gypsum deposits were prevented
from forming in the sequence. As tectonic restriction continued, the gypsum facies disap-
pear form the upper sequence and are replaced by freshwater limestone beds.
6.5.3 A mechanism for PLG formation: an orbitally forced dissolution
flux
An inconsistency exists between the large proportion of freshwater implied from Sr and
GHW isotopes during the formation of the PLG, and the source of calcium and sulfate ions
needed for the formation of such large gypsum deposits. A mechanism other than simple
evaporation of seawater is needed to explain the lower salinities inferred from fluid inclusion
and isotopic results in the PLG.
6.5 Discussion 175
Natalicchio et al. (2014) and Evans et al. (2015) proposed that gypsum formed from a
mix of seawater and Ca2+- and SO42−-enriched freshwaters derived from partial dissolution
and recycling of coeval marginal evaporite deposits. The recycling of carbonate and gypsum
deposits would have occurred in marginal areas of the basin where evaporites were deposited
previously. Alternatively, recycling could also occur via in situ early recrystallisation, which
will not affect the δ 34S or δ 18OSO4 values, but will change the isotopic composition of
the GHW, salinity and strontium concentration (e.g. Longinelli, 1979). To date, a late
Miocene gypsum deposit that formed solely from the evaporation of seawater has not been
found in the Mediterranean, and the consistent 87Sr/86Sr profiles between differing circum-
Mediterranean basins indicate that the original fluids were composed of a large percentage
of freshwater. While recycling processes likely enhanced ionic input at a local scale, basin-
wide recycling without evidence for evaporative marine gypsum (either due to physical
absence or lack of discovery) is harder to justify.
Measurement of both Ca and Sr isotopes in PLG sections provide clues to the origin of
the ion source for gypsum formation. The δ 44/40Ca values generally track the variations in
87Sr/86Sr of the gypsum layers up-section, suggesting that the same mechanism of enhanced
continental runoff is impacting both isotope systems. The low 87Sr/86Sr on the freshwater
input is indicative of the weathering of carbonate rocks, and weathering of carbonate with
low δ 44/40Ca (~-0.1h) would also result in a lowering of δ 44/40Ca during increased fluvial
discharge. The dissolution of carbonate and addition of calcium to the Mediterranean basin
may provide the additional ion source needed to form gypsum at lower than predicted
salinities.
The increased weathering of carbonates inferred from fluctuations in calcium and stron-
tium isotopes provides a potential explanation for the unexpectedly low salinities found
within many of the PLG of many Mediterranean marginal basins. Both the solubility of
gypsum (Ksp) and the product of dissolved calcium [Ca2+] and sulfate [SO42−] determines
gypsum precipitation. If
[
Ca2+
]×[SO42−]> Ksp, then gypsum super-saturation is reached
and will precipitate. Because normal seawater (0.01 mol/l [Ca2+], 0.028 mol/l [SO42−];
KspCaSO4·2H2O =∼ 0.28× 10−3) is under-saturated for gypsum (~2.6× 10−3 ) (Broecker,
1982), evaporation is needed to increase the concentration of dissolved [Ca2+] and [SO42−].
Seawater salinities are required to rise from ~35 ppt to ~130 ppt (Warren, 2016) before
gypsum saturation is reached, requiring the loss of ~70% of the starting volume. By adding
[Ca2+] and/or [SO42−] to the system, the product of
[
Ca2+
]×[SO42−] will rise without the
need for large amounts of evaporation (Fig. 6.12). The addition of calcium ions from the
large scale weathering of the African continent will mean that the gypsum saturation level
can be reached at lower salinities. Additionally, unlike the weathering of ancient evaporites
176 Meteoric water, orbital and eustatic control of basin hydrology
that could potentially alter the δ 34S or δ 18OSO4 of the gypsum, the dissolution of carbonates
would only alter the calcium and strontium isotopic signatures.
6.5.4 Conclusions
Multi-isotope analysis of 12 of the 15 cycles of the Yesares Member indicate that PLG
deposition was influenced by hydrologic dynamics that were governed by precessional-scale
climate oscillations, tectonics and eustatic sea-level change. The homogeneous isotopic
contents recorded in the lower cycles of the sequence suggest hydrological stability and
relatively deep waters, whereas these conditions are progressively modified upwards and
greater isotopic variability is seen in the upper part of the section, suggesting a shallowing
trend driven by tectonic restriction. Critically, reversals of the general trend of decreasing
in Sr isotopes up-section together with step changes in Ca isotopes, as well as high δ18Ocarb
and anomalous mineralogy of the co-occurring marls, indicates eustatic sea level variability
also controlled the seawater input by affecting the rate of Mediterranean-Atlantic exchange.
The gypsum of all cycles was formed in a hybrid brine consisting of a seawater body
strongly influenced by abundant freshwater input that increased up-section. The freshwater
input likely enhanced the dissolved carbonate ion flux and stratification of the Mediterranean
basin, facilitating gypsum precipitation.
Chapter 7
Upper Evaporites of the Messinian
Salinity Crisis
Nicholas P. Evans, Fernando Gázquez, Thomas K. Bauska, Luke Bridgestock, Harold J.
Bradbury, Alexandra V. Turchyn, and David A. Hodell.
This chapter aims to address fundamental questions related to mechanisms responsible
for the salinity fluctuations experienced by the Mediterranean basin during the terminal
stages of the Messinian Salinity Crisis (MSC). Analysis of gypsum hydration water (GHW)
from both onshore and deep basin deposits is conducted to obtain the δ18O, δD, 17O-excess,
and d-excess of the mother fluids from which they formed. The measurements are combined
with salinity deduced from fluid inclusion melting temperatures, strontium, calcium, barium
and sulfate isotopes in order to reconstruct the hydrological changes in the Mediterranean
basin during the late stage of the MSC. The results from the Upper Gypsum (UG) are
compared with those of the Primary Lower Gypsum (PLG) deposits (Chapter 6) to contrast
the different environments of deposition.
Strontium, calcium and barium isotope data obtained from the onshore UG are signi-
ficantly lower than results of the PLG deposits, whereas oxygen and hydrogen isotopes of
measured gypsum hydration water display higher values. These data demonstrate that the
onshore UG deposits formed from intense evaporation of a dominantly freshwater fluid,
with a periodic and highly restricted connection to the Atlantic. The deep-basin ‘Upper
Evaporites’ from DSDP/ODP sites 372 and 654 are characterised by similar strontium,
calcium and sulfate isotope values to those of the onshore Upper Evaporites. The deep basin
water isotope data deviates from the majority of the onshore deposits, however, indicating
that variable and heterogeneous hydrologic conditions characterised the deep basins during
178 Upper Evaporites of the Messinian Salinity Crisis
evaporite formation.
7.1 Introduction
The reduction of the hydrological exchange between the Mediterranean Sea and Atlantic
Ocean during late Miocene ultimately culminated in the deposition of large volumes of
sulfate and chloride minerals in the Mediterranean basin during the event known as the
Messinian Salinity Crisis (MSC; 5.97 Ma to 5.33 Ma) (Roveri et al., 2014a). According to
the consensus model of the MSC (CIESM, 2008), the acme of evaporite precipitation was
associated with the deposition of halite between 5.60 Ma and 5.55 Ma (stage 2) coupled with
the incision of deep canyons at the basin margins, both a result of a significant sea level drop
of greater than 1500 m. The subsequent deposits include the Upper Evaporites (stage 3.1;
5.55 Ma to 5.42 Ma) and brackish-water ‘Lago-mare’ (stage 3.2; 5.42 Ma to 5.33 Ma) units
in marginal to intermediate settings. The Upper Evaporites typically consist of gypsum-marl
alternations that represent a response to precession-driven climate oscillations (e.g. Manzi
et al., 2009), similar to stage 1 PLG evaporitic cycles (Chapter 6). An evaporite layer in
the deep Mediterranean basin has also been identified from seismic data and is thought to
be coeval to the onshore Upper Evaporates (see Roveri et al., 2014a). The topmost part of
these basinal units were sampled during DSDP Leg 13 in 1970, DSDP Leg 42A in 1975,
ODP Leg 107 in 1986, and Legs 160 and 161 in 1995.
Debate is still ongoing regarding the palaeoenvironmental and hydrological conditions
that prevailed during stage 3 of the MSC, and how they differ from the conditions that
occurred during the formation of the older PLG. The occurrence of brackish water fauna
such as ostracods with Black Sea affinity (Gliozzi et al., 2007; Stoica et al., 2016) suggest
that shallow, brackish lake environments existed both at the bottom of a desiccated Mediter-
ranean and around its margins (e.g. Hsü et al., 1973; Orszag-Sperber, 2006). The discovery
of marine fish remains (e.g. Carnevale et al., 2006), however, indicates a continuous or
episodic flux of Atlantic waters into the Mediterranean during deposition of the stage 3
evaporites.
Here, the hydrological history of the last stage of the MSC from both onshore (Caltan-
issetta Basin, Eraclea Minoa section, Sicily) and offshore (DSDP Leg 42A site 372 and
ODP Leg 107 site 654) deposits are determined. To estimate the composition of the fluids
that formed the Upper Evaporites in stage 3, triple oxygen and hydrogen isotopes of GHW,
strontium (87Sr/86Sr) and calcium (δ 44/40Ca), barium (δ 138/134Ba), sulfur (δ 34S), oxygen
isotopes of sulfate (δ 18OSO4) and salinity (deduced from fluid inclusion melting temperat-
ures) were measured. The 87Sr/86Sr, δ 44/40Ca, δ 138/134Ba and GHW isotopes values are
7.1 Introduction 179
Primary Lower
Gypsum Block
0
1
2
3
4
5
6
6'
6"
7
Trubi Fm. Miocene–
Pliocene 
boundary
Sandstone bar
Partially 
covered
50
100
150
200
H
e
ig
h
t 
(m
)
Marls
Domed selenite
Massive selenite
Fine-grained gypsum-marl
Laminar gypsum beds
Sandstone beds
Resedimented Lower Gypsum 
Messinian Evaporites DSDP-ODP Sites
Gypsum/Halite/Gypsum "Trilogy"
Undifferentiated Gypsum-Halite
Only Gypsum
Evaporites Recovered
No Evaporites Recovered
45°
40°
35°
20°
Orogenic front
Passive margin
Transcurrent fault400 km
IBERIA
AFRICA
ATLANTIC
OCEAN
121 976 977 978
122
123 656
133
974
373
973
121
975
372
124975
371
134
654
132
654 652
374
CALTANISSETTA
BASIN
Sampled Sites
Other Western Mediterranean 
sites containing Primary 
Lower Gypsum 
xyz
Irpinia
Maiella
Vena del 
Gesso
Hoya de 
la Sima
Nijar
Fig. 7.1. (Left) Stratigraphic section of the 7 gypsum-marl cycles found at Eraclea Minoa, Caltanissetta
Basin, Sicily (after Manzi et al., 2009). (Right) Map of the Messinian evaporites in the
Mediterranean and the sampled sites including onshore basins, ODP Leg 107 site 654 and DSDP
Leg 42A site 372. Also shown is the location of other DSDP-ODP boreholes in which Messinian
evaporites were recovered, as well as other PLG sections found in marginal basin settings. Modified
after Rouchy and Caruso (2006) and Roveri et al. (2014a).
180 Upper Evaporites of the Messinian Salinity Crisis
significantly different from both the global ocean and PLG deposits. The data indicate
that the proportion of oceanic water reaching the Mediterranean was significantly reduced
during stage 3, and large variations in the hydraulic regime occurred during the gypsum-
marl cycles.
7.2 Geological setting
7.2.1 Onshore deposits
The onshore record of the MSC is widely accepted to have originally formed in ‘marginal’
or ‘peripheral’ basins, i.e. sub-basins formed in the more proximal settings of the Mediter-
ranean basins (Roveri et al., 2014a). Exposures on Sicily, for example, show alternations of
gypsum and marl beds. The most complete stage 3 successions are found at Eraclea Minoa,
Sicily, where 10 sedimentary cycles (7 with gypsum) have been recognized (Van der Laan
et al., 2006; Hilgen et al., 2007; Manzi et al., 2009) overlaying a ‘Resedimented Lower
Gypsum’ (RLG) block (Fig. 7.1).
According to Manzi et al. (2009) the UG facies include: marl and thin-bedded sandstone
packages; graded sandstone beds (unit S1); fine-grained gypsum and grey marl graded
beds (unit G1a); fine-grained laminar gypsum beds (unit G1b); selenite crusts (unit G2a);
decimetre-thick massive selenite beds (which are separated by thin marl horizons up to
~30 cm thick) (unit G2b); and domed selenite structures (unit G2c) (Fig 4.6; Chapter 4).
Sixty-seven gypsum samples and 8 carbonate marl samples were chosen from cycles 3,
5 and 7 for isotopic analysis, and cycle 4 was chosen for high-resolution sampling. We also
examined 4 samples taken from the RLG at the base of the section.
7.2.2 Offshore deposits
7.2.2.1 DSDP Site 372
We selected 14 gypsum samples of laminated gypsum from of core 8 (189 mbsf to 190 mbsf)
of DSDP Site 372 (Fig. 7.2). The site is located at the East Menorca Rise in the Balearic
Sea, western Mediterranean (40°01.86’N, 04°47.79’E). The major lithologies are nanno-
fossil marls and clayey mudstones with extremely low coarse terrigenous clastic content,
together with gypsum and dolomitic marls in the evaporitic sequence (Hsu and Montadert,
1978). Primary cumulative and clastic gypsum and dolomitic marls are found 150 mbsf to
199.5 mbsf. Lofi et al. (2011) report this unit as a pinch-out wedge of a seismic unit that is
7.3 Results 181
ascribed to the ‘Upper Evaporite’, while Roveri et al. (2014a) separate the lower and upper
halves of the unit into stage 2 and stage 3, respectively, based on 87Sr/86Sr stratigraphy.
7.2.2.2 ODP Site 654
Fifty-five samples were taken from cores 27-34 (242-291 mbsf) of ODP Site 654 (Fig. 7.2).
Samples were mainly composed of regular laminar gypsum or wavy laminar gypsum (Fig.
6.3; Lugli et al., 2015). ODP Site 654 is located on the upper Sardinia margin of the
Tyrrhenian Sea (40°34.76’N, 10°41.80’E) (Shipboard Scientific Party, 1987). Rhythmic
alternation of clastic or cumulate gypsum and marls occurs between 242.7 and 313 mbsf
(Borsetti et al., 1990). Similar to other deep basin sites, Roveri et al. (2014b) separate the
upper and lower half of this unit and correlate the respective layers to MSC stages 3 and 2.
7.3 Results
7.3.1 Gypsum Hydration Water (GHW)
One hundred and thirty-six samples representing both onshore and offshore gypsum deposits
were analysed (Appendix D; Table D.2). The oxygen and hydrogen isotope composition
of the parent water in which the gypsum formed is calculated by correcting for gypsum-
water fractionation factors. Fractionation factors were assumed to be 1.0035 and 0.981 for
oxygen and hydrogen isotopes, respectively (Chapter 2). The fractionation factors used were
selected based on estimated water temperature (~15°C to 20°C) and salinity (<150 g l-1) at
the time of gypsum formation, although the corrected values are relatively insensitive to the
temperature and salinity chosen (Gázquez et al., 2017). Corrected results are displayed in
Fig. 7.3.
Onshore Deposits: All cycles display significantly higher δ18O and δD values and lower
d-excess and 17O-excess values compared to the PLG of the Sorbas Basin (Fig. 7.3 and Fig.
7.4). The δ18O and δD values remain relatively constant through the section, averaging 4.5
(±1.5) h and 21.5 (±10.9) h, respectively, while d-excess and 17O-excess average -14.6
(±2.1)h and -44 (±14) per meg, respectively.
The δ18O and δD of gypsum mother water show a systematic change through cycle 4 of
the UG in the Caltanissetta Basin (Fig. 7.5). The lowest δ18O and δD values are found in the
fine-grained laminar gypsum beds at the base of the cycle. The δ18O and δD values steadily
increase to maximum values of 6.49h and 37.24h, respectively, in the gypsum palisade at
~6 m. The δ18O and δD values maintain their high values into the dome structures above,
182 Upper Evaporites of the Messinian Salinity Crisis
Fig. 7.2. Schematic depiction of the facies that were recovered from the cores of ODP Leg 107 site 654 and
DSDP Leg 42A site 372 in the Eastern Mediterranean Sea (after Lugli et al., 2015).
7.3 Results 183
although there is some variability as two samples display low values. Cycle 3 shows a
similar trend, where δ18O and δD increase up-section. The δ18O and δD of Cycle 5 and
7 do not follow the same increasing trend up-section, but display relatively constant values
over the cycle (δ18O = 4.6 (±2.1)h and 4.3 (±0.7)h, respectively).
Offshore Deposits: Mean δ18O and δD from ODP Site 654 averages 1.0 (±0.8) h and
4.9 (±4.2)h, respectively, similar to that of the onshore PLG (1.0 [±0.4]h and 0.9 [±9.2]
h, respectively), although the isotopic range of obtained from the PLG is greater. The δ18O
and δD range of gypsum from site 372 is 1.6 < δ18O < 3.0h and 7.5 < δD < 12.2h (Fig.
7.3 and Fig. 7.4). The δ18O and δD also display relatively little variability down-core in
both sites 654 and 372 (Fig. 7.5).
7.3.2 Microthermometric analysis
Six samples from the UG of the Caltanissetta Basin were selected for microthermometric
analysis (Appendix D; Table D.7). The Tmice of 46 primary inclusions ranged between -
1.2°C and -8.2°C, with an average Tmice of -4.4°C. Using the revised equation and table
for determining the freezing point depression of H2O–NaCl solutions of Bodnar (1993),
average salinity estimates of individual samples range from 5.7% to 9.0% (average 7.1
wt% NaCl equivalent) (Fig. 7.5). Samples from the deep basin sites are composed of
fine-grained laminar gypsum beds, and thus individual crystals could not be selected for
microthermometric analysis.
7.3.3 Sr isotopes
Seventeen samples from the onshore UG, 7 samples from ODP site 654 and 4 samples from
the RLG were processed for strontium isotopic analysis (Appendix D; Table D.5).
Onshore Deposits: For UG samples, the 87Sr/86Sr ranged between 0.708719 and 0.708795
(Fig. 7.6). These data show a strong divergence from the strontium isotope values expected
for Messinian seawater, which varies from ~0.709015 at 5.55 Ma to ~0.709025 at 5.33 Ma
(McArthur et al., 2001). Mean 87Sr/86Sr decreases up-section, with mean values in cycle 2
of 0.708783 to 0.708730 in cycle 7.
Within individual cycles of the onshore UG, there is a slight trend toward more positive
87Sr/86Sr values toward the top of each gypsum layer (Fig. 7.5). Within cycle 4, strontium
concentrations increase from 595 ppm at the base of the cycle to 1333 ppm at the top.
184 Upper Evaporites of the Messinian Salinity Crisis
-80
-60
-40
-20
0
20
40
-15 -10 -5 0 5 10 15
-80
-60
-40
-20
0
20
40
-12 -8 -4 0 4 8 12
-80
-60
-40
-20
0
20
40
-12 -8 -4 0 4 8 12
A
αD = 0.981 
α18O = 1.00341
LM
W
LPLG Cycles 1-15
UG Cycles 3-7
DSDP Site 372
ODP Site 654
Modern Salina Gypsum
Rain and Groundwater
Gy
ps
um
 H
yd
ra
tio
n 
W
at
er
Gy
ps
um
 M
ot
he
r 
W
at
er
αD = 0.981 
α18O = 1.0035
LMWL
LMWL
E/I = 0.5
E/I = 1
E/I = 0.5
E/I = 1
Recharge
δ18O (‰ VSMOW)
δD
 (
‰
 V
S
M
O
W
)
B
d-
ex
ce
ss
 (
‰
 V
S
M
O
W
)
C
17
O
-e
xc
es
s 
(p
er
 m
eg
)
-20
-15
-10
-5
0
5
-1 1 2 3 4 5 6 70
δ18O (‰ VSMOW)
δ18O (‰ VSMOW)
y = -2.7x - 0.6 
y = -4.8x - 1.8 
7.3 Results 185
Fig. 7.3. Left. Comparison of measured local meteoric water (grey circles) and predicted Messinian brine
mother water values for the differing facies of the gypsum from sites 372 (red diamonds) and 654
(green diamonds), Upper Gypsum (yellow diamonds) and PLG (blue circles) after correction for
fractionation factors (Gázquez et al., 2017). In (A) δ18O vs δD, (B) δ18O vs d-excess and (C)
δ18O vs 17O-excess space, local meteoric water measurements define the local meteoric water line
(LMWL). Light blue triangles are modern gypsum data from the Cabo de Gata Salina. Red square
is Modern Mediterranean Seawater. The trajectories for the evaporation of Mediterranean seawater
for high (90%; slope ~7) and low (40%; slope ~3) humidities are shown in A, defining a range
of possible δ18O and δD values (pink field). Gypsum precipitation takes place at 3 to 7 times
concentrated seawater; gypsum saturation (blue field) is calculated using the Rayleigh Distillation
equation, assuming initial δ18O and δD values of ~1h and 8h, respectively, 75% equilibrium
between atmospheric vapour and fresh water, and gypsum saturation occurring at an evaporation
ratio of 3.8× (evaporation ratio = weight of H2O of original seawater divided by weight of H2O in
residual evaporated brine). Mixing lines in B and C are derived from the model of Gázquez et al.
(2018) at a humidity of 75% and a starting point of modern Mediterranean seawater. The mixing
curve does not match the data, indicating that basin recharge with freshwaters was dominant. Errors
(1σ ) are less than the width of the symbols, and ±6 for 17O-excess.
Offshore Deposits: The 87Sr/86Sr variation of gypsum from ODP site 654 is large (0.708730
to 0.708806) and also significantly lower than late Miocene seawater. There is a slight
decreases in 87Sr/86Sr values up-core (Fig. 7.5).
The 87Sr/86Sr of the RLG is very similar to mean PLG values (Fig. 7.6).
7.3.4 Ca isotopes
Six samples from the onshore UG and 11 samples from ODP site 654 were processed for
calcium isotope analysis (Appendix D; Table D.5).
Onshore Deposits: The δ 44/40Ca of the onshore UG is significantly lower than that of the
PLG (Fig. 7.6). There are no trends in Ca isotopes up-section.
The δ 44/40Ca of Cycle 4 of the UG varies between 0.03 and 0.54h. δ 44/40Ca generally
increase from low values at the base to maximum values at the top of the cycle (Fig. 7.5).
Offshore Deposits: Calcium isotope values of gypsum from ODP Site 654 are signific-
antly lower than both the onshore UG and the PLG, ranging between -0.13 and 0.46h (Fig.
7.6). The highest values occurred near the base of the sampled interval, and values decrease
towards to the top of the core (Fig. 7.5).
186 Upper Evaporites of the Messinian Salinity Crisis
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3Cycle
1
5 21 3 4 5 6 76'
6
''
R
L
G
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3
1
5 21 3 4 5 6 76'
6
''
A
g
e
 (M
a
)
O
b
liq
u
ity
 
E
T
P
A
rid
W
e
t
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
S
ite
 6
5
4
S
ite
 3
7
2
0
.0
2
.0
T
G
3
2
4
.0
6
.0
-6
0
-4
0
-2
0
0
0
-5
-1
0
-1
5
-2
0
0
5
0
1
0
0
1
5
0
2
0
0
δ
1
8O
 
(‰
, V
S
M
O
W
)
1
7O
-e
x
c
e
s
s
 
(p
e
r m
e
g
)
S
a
lin
ity
 
(p
p
t)
d
-e
x
c
e
s
s
 
(‰
, V
S
M
O
W
)
SMOW
5
0
1
0
0
1
5
00
Height (m)
M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
 
S
a
n
d
sto
n
e
 b
e
d
s
T
G
2
4
T
G
2
2
T
G
2
0
T
G
1
4
T
G
1
2
5.3
5.4
5.5
5.6
5.7
5.8
5.96
6.1
6
3
0
3
6
0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
2
2.5
3
3.5
-
-
-
(La
ska
r e
t a
l., 2
0
0
4
)
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
7.3 Results 187
Fig. 7.4. Left. Mean intra-cycle δ18O, 17O-excess and d-excess data (calculated from measurement of
GHW) and salinity data (deduced from fluid inclusions and calculated using the equation of
Goldstein and Reynolds, 1994) are plotted against their designated cycle. Error bars define ±1σ of
the data in each cycle. The vertical grey bars indicate the ±1σ range of the data from all PLG and
all UG cycles, respectively. Each gypsum bed is plotted in relation to its proposed orbital tuning
(Laskar et al., 2004; Lugli et al., 2010; Manzi et al., 2013), where gypsum precipitation correlates
with insolation minima. Note that the the position of data from the RLG, ODP Leg 107 site 654
and DSDP Leg 42A site 372 is arbitrary and does not reflect the age of the deposit.
7.3.5 Ba isotopes
Four samples from the onshore UG, 17 samples from the PLG of Sorbas Basin (SE Spain),
and 1 sample from the modern Cabo de Gata Salina, SE Spain, were processed for barium
isotope analysis (Appendix D; Table D.5).
The δ 138/134Ba of the PLG ranged from 0.00h to 0.23h, with a mean value of 0.10h
(Fig. 7.6). δ 138/134Ba display no significant variability between different cycles, although
similarities between mean values of δ 138/134Ba and δ 44/40Ca are apparent up-section. There
is no defined variability in δ 138/134Ba within any of the PLG cycles.
The δ 138/134Ba of the UG ranged from -0.06 to 0.01, with a mean value of -0.01h (Fig.
7.6). Isotope values within cycle 4 show no trends with position in the cycle.
The δ 138/134Ba of the sample from the Cabo de Gata Salina yielded a value of 0.08h.
7.3.6 δ 34S and δ18O of sulfate
Sixty two samples were analysed for sulfur isotopes and oxygen isotopes of sulfate (Ap-
pendix D; Table D.7). These data are shown in Fig. 7.7 relative to the expected range of
gypsum precipitated from Miocene seawater and the PLG reported in Chapter 6.
Onshore Deposits: The δ 34S of gypsum from the onshore deposits of the Caltanissetta
Basin varies between 20.8h and 22.6h VCDT and δ 18OSO4 between 13.2h and 18.8h
VSMOW. The δ 34S and δ 18OSO4 of onshore UG are positively correlated with a slope of
~2.6 (Fig. 7.7).
Onshore UG deposits are characterised by similar δ 34S compositions the PLG (δ 34S
~22.5h), although mean values are slightly lower and mean δ 34S decreases from 22.4h in
cycle 3 to 21.1h in cycle 7 (Fig. 7.8). The δ 18OSO4 values (15.9 ±1.8h) are significantly
higher than PLG (δ 18OSO4 ~13h), and above values expected from the precipitation of
Late Miocene marine evaporites (δ 18OSO4 ~12h). Mean δ 18OSO4 also decreases up-section
from 18.5h in cycle 3 to 16.0h in cycle 7.
188 Upper Evaporites of the Messinian Salinity Crisis
0
.7
0
8
7
0
.7
0
8
8
0
.0
2
.0
4
.0
6
.0
2
1
.0
-0
.5
0
.0
0
.5
0
5
0
1
0
0
1
5
0
2
0
0
0
.7
0
8
9
ODP Leg 107 Site 654Cycle 4
δ44
/
4
0
C
a
 (
‰
)
2
8
0
2
6
0
2
5
0
0
m2468
S
a
li
n
it
y
 (
p
p
t)
8
7
S
r/
8
6
S
r 
δ18
O
 (
‰
)
M
a
rl
s
D
o
m
e
d
 s
e
le
n
it
e
M
a
ss
iv
e
 s
e
le
n
it
e
Fi
n
e
-g
ra
in
e
d
 g
y
p
su
m
-m
a
rl
La
m
in
a
r 
g
y
p
su
m
 b
e
d
s
2
7
0
2
9
0
3
0
0
3
1
0
m
b
sf
2
0
.0
2
7
2
8
2
9
3
0
3
1
3
2
3
3
3
4
3
5
3
6
C
o
re
1
0
.0
1
5
.0
2
0
.0
δ34
S
 (
‰
)
δ18
O
S
O
4
 (
‰
)
2
2
.0
2
3
.0
2
4
.0
Fig. 7.5. Schematic diagram of UG cycle 4 and ODP Leg 107 site 654 scaled to height of each of the sections.
Calcium isotope data, strontium isotope data, salinity data (deduced from fluid inclusions and
calculated using the equation of Goldstein and Reynolds, 1994), δ18O and d-excess data deduced
from GHW are plotted relative to the height from which the samples were taken. Note there is no
salinity data available from ODP Leg 107 site 654 due to the fine-grained nature of the gypsum
present.
7.4 Discussion 189
The δ 34S and δ 18OSO4 show a systematic decrease with increasing height in cycle 4
(Fig. 7.5). The δ 34S and δ 18OSO4 values decrease from ~21.0h and 14.0h, respectively,
at the base, to ~22.0 and 17.5h, respectively, at the top.
Offshore Deposits: The δ 34S of gypsum from the offshore deposits from DSDP site 372
and ODP site 654 varies between 19.5h and 22.9h and δ 18OSO4 between 15.3h and
20.3h (Fig. 7.8). Both δ 34S and δ 18OSO4 show no correlations with depth (Fig. 7.5).
Compared to the mean PLG values, the δ 34S and δ 18OSO4 of the offshore deposits display
lower and higher values, respectively. The δ 34S and δ 18OSO4 of DSDP site 372 and ODP
site 654 sulfates are positively correlated with a slope of ~1.2 (Fig. 7.7).
The δ 34S and δ 18OSO4 RLG are in line with those of the PLG (22.3h and 12.8h,
respectively).
7.3.7 δ18Ocarb and δ13C of bulk carbonates
The oxygen and carbon isotope composition of eight carbonate samples vary from -2.0 to
1.1h for δ18Ocarb and from -1.6 to -0.3h for δ13C (Appendix D; Table D.8). The carbon
and oxygen isotope composition of the carbonates are within the range of other Messinian
carbonates associated with evaporites in the Mediterranean (Longinelli, 1979).
7.4 Discussion
The onshore UG unit of the Caltanissetta Basin (Sicily), and the offshore sulfate deposits re-
covered from DSDP and ODP sites are believed to record the last phase of the MSC, but the
hydrological conditions under which the geographically disparate units formed are poorly
understood. Our GHW and salinity data from the onshore deposits partially overlap with the
range expected from the evaporation of seawater, but 87Sr/86Sr, δ 44/40Ca, δ 138/134Ba data
support a significant contribution by meteoric water at both deep and shallow water sites.
Below we discuss how the composition of PLG deposits differ from both the (i) onshore UG
and (ii) deep basin units, and propose a mechanism that explains the geochemical data.
7.4.1 Isotopic comparison between the onshore Upper Gypsum and
Primary Lower Gypsum
The corrected oxygen and hydrogen isotopic composition of GHW provide a record of the
parent water from which gypsum precipitated (e.g. Evans et al., 2015; Herwartz et al., 2017;
190 Upper Evaporites of the Messinian Salinity Crisis
0
.7
0
8
9
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3Cycle
1
5 21 3 4 5 6 76'
6
''
R
L
G
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3
1
5 21 3 4 5 6 76'
6
''
0
.7
0
8
9
5
0
.7
0
9
0
.7
0
9
0
5
8
7S
r/ 8
6S
r
A
g
e
 (M
a
)
0
.0
0
.4
1
.2
0
.8
δ
4
4
/
4
0C
a
 
(‰
, B
S
E
)
Global Ocean
0
.0
0
.2
0
.4
δ
1
3
8
/
1
3
4B
a
 
(‰
)
O
b
liq
u
ity
 
E
T
P
A
rid
W
e
t
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
0
.7
0
8
8
5
0
.7
0
8
8
0
.7
0
8
7
5
0
.7
0
8
7
Ca Seawater
Ba Seawater
S
ite
 6
5
4
S
ite
 3
7
2
5
0
1
0
0
1
5
00
Height (m)
M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
 
S
a
n
d
sto
n
e
b
e
d
s
T
G
3
2
T
G
2
4 TG
2
2
T
G
2
0
T
G
1
4 T
G
1
2
5.3
5.4
5.5
5.6
5.7
5.8
5.96
6.1
6
3
0
3
6
0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
2
2.5
3
3.5
-
-
-
(La
ska
r e
t a
l., 2
0
0
4
)
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
7.4 Discussion 191
Fig. 7.6. Left. Mean intra-cycle δ44/40Ca, δ138/134Ba and 87Sr/86Sr data plotted against their designated
cycle. Error bars define ±1σ of the data in each cycle. The vertical grey bars indicate the ±1σ
range of the data from all PLG and all UG cycles, respectively. Each gypsum bed is plotted in
relation to its proposed orbital tuning (Laskar et al., 2004; Lugli et al., 2010; Manzi et al., 2013),
where gypsum precipitation correlates with insolation minima. The benthic δ18O (Hodell et al.,
2001; modified by Drury et al., 2018) is also shown, and the major glacial stages are indicated.
Note that the the position of data from the RLG, ODP Leg 107 site 654 and DSDP Leg 42A site
372 is arbitrary and does not reflect the age of the deposit. All isotope systems display a significant
shift toward lower values in the UG compared to the PLG, indicative of an environment dominated
by meteoric input.
Gazquez et al., 2018), whereas the ionic constituents of the mother fluid are preserved in
primary fluid inclusions that trap the parent brine during gypsum precipitation (e.g. Attia
et al., 1995). The δ18O and δD of the UG units display values that are greater than results
obtained from the PLG, and partially overlap with the isotopic range expected for evaporated
seawater (Fig. 7.3). Salinity estimates deduced from the melting temperature of fluid
inclusions from the UG (mean cycle 4 (±σ ) = 102 (±34) ppt) partially overlap with the
range of gypsum saturation expected if evaporated from seawater (> ~135 ppt) (Goldstein
and Reynolds, 1994). Additionally, the δ 34S of the UG is similar to results obtained from the
PLG and the range expected from sulfate minerals precipitated from normal Messinian sea-
water (~21h) (Fig. 7.8). This may indicate that the sulfate molecule for gypsum formation
may have been derived from a marine source, although simple mixing model calculations
demonstrate that sulfur isotopes are highly insensitive to continental inputs, requiring >95%
contribution from freshwater sources to cause a deviation in seawater δ 34S values (Chapter
6). Together, these results indicate that the UG formed from (predominantly) marine fluids
that had undergone evaporation. However, the other isotopes systems suggest a different
origin.
Strontium isotope data argues for a dominantly freshwater-sourced parent fluid. The
onshore UG yields low 87Sr/86Sr values compared to the coeval global ocean. No significant
overlap exists between the UG and the PLG (Fig. 7.6). The 87Sr/86Sr values of the Rhone
(~0.70872) and Nile (~0.70600) are lower than the global seawater average during the
Messinian (~0.709), and make up >80% of the total Mediterranean riverine input today
(McArthur et al., 2001; Flecker et al., 2002; Topper et al., 2014). As the strontium isotope
ratio of water in a marginal oceanic basin is a simple mixture between ocean water and
meteoric sources feeding a basin, the strontium isotope data indicate that inflow from the
global ocean was severely restricted and that continental waters dominated the Mediter-
ranean during the precipitation of the UG.
Measurements of barium and calcium isotopes provide additional information regarding
the relative freshwater flux to marginal settings. The evolution of both δ 138/134Ba and
192 Upper Evaporites of the Messinian Salinity Crisis
δ 44/40Ca within a semi-enclosed marginal basin, similar to 87Sr/86Sr, is a function of the bal-
ance between ocean and riverine input. Riverine sources of barium constitute an important
input to the surface ocean, where freshwater sources have lower isotope values (δ 138/134Ba
= -0.1h to 0.4h; Cao et al., 2016) than typical seawater (0.4 h to 0.6h) (Hsieh and
Henderson, 2017). The δ 44/40Ca of ocean and riverine input are also characterised by
different calcium concentrations and δ 44/40Ca values (rivers ~-0.1h; seawater = 0.92h).
Both δ 138/134Ba and δ 44/40Ca measured in UG display lower values than PLG samples
or Messinian seawater (Fig. 7.6), and display a significant correlation with measured 87Sr/86Sr
in the Upper and Lower Gypsum (Fig. 7.9). The lower 87Sr/86Sr, δ 138/134Ba and δ 44/40Ca
values in the UG are consistent with a dominant contribution of meteoric water to the
fluids that formed Stage 3 deposits. Simple mixing simulations using the Messinian marine
(0.70902, 8×10−3 kg m-3 Sr) and the weighted modern major river inputs of the Rhone and
Nile Rivers (0.70763, 3×10−4 kg m-3 Sr) suggests a dilution of >80% of the Mediterranean
Sea to reach the measured average of the UG (0.708771). Additionally, the influence of
freshwater likely increased toward the top of the UG unit, as mean 87Sr/86Sr of cycle 7 is
less (0.708725) than the cycles below in the section (e.g. cycle 4 = 0.708777).
The upper bound of the continental contribution can be constrained from the application
of mixing models to the δ 34S values obtained from the gypsum deposits. This analysis
yields results that suggest that the freshwater contribution did not exceed 95%. Volumes
of freshwater input above this threshold would result in a lowing of the measured sulfur
isotope results towards riverine values. Together, these data indicate that Mediterranean-
Atlantic exchange persisted, although it was suppressed and episodic.
7.4.2 Semi-closed system behaviour during Upper Gypsum formation
The discrepancy between salinity, δ 34S and GHW signatures, and 87Sr/86Sr, δ 138/134Ba
and δ 44/40Ca measurements made on the UG can be reconciled if significant evaporation
occurred in a semi-closed system with large inputs of meteoric water. The isotopic com-
position of a mother fluid during the gypsum precipitation phase is a function of the mixing
between seawater (δ17O ~0h; δ18O ~0h; δD ~0h) and freshwater (e.g. Sicilian and
Spanish freshwater; δ17O = -2.79h; δ18O = -5.3h; δD = -32.5h) and the evaporative
enrichment that is needed to reach gypsum saturation (dependent on the concentration
calcium and sulfate ions in the fluid). During the evaporation of a brackish fluid, for
example, a greater proportion of water needs to be lost during evaporation to reach gypsum
saturation compared to the evaporation of seawater alone (e.g. Rossi et al., 2015). Using
the method of Gázquez et al. (2018), we modelled the δ18O, d-excess and 17O-excess
obtained from the GHW to estimate the hydrologic balance during gypsum precipitation.
7.4 Discussion 193
20.0
δ34S (‰)
6.0
8.0
10.0
12.0
14.0
16.0
21.0 22.0 23.0 24.0
PLG 
Upper Gypsum
ODP Site 654
DSDP Site 372
Modern Salina Gyspum 
δ18
O
S
O
4
 (
‰
)
Messinian 
marine field 
Oxygen isotope 
exchange with 
water 
Py
rit
e o
xid
at
ion
 
in 
riv
er
s 
Cyclic sulfate
reduction 
18.0
20.0
Fig. 7.7. Comparison of δ 34S vs. δ18O of sulfate of the PLG and UG. Open box defines Messinian seawater
range (Paytan, 1998; Turchyn and Schrag, 2004).
The Monte Carlo experiments are used to determine the range of climatological conditions
that simultaneously satisfy all stable isotope results of GHW, in combination with statistical
estimates of uncertainty. In these scenarios, the parameter ‘Xe’ represents the hydrologic
balance. A scenario in which all water is lost by outflow and no evaporation occurs is
represented as Xe = 0, whereas a scenario in which all water is lost to evaporation (i.e.
a terminal basin) is represented as Xe = 1. Assuming (i) the isotopic composition of
freshwater input and relative humidity (~75%) is similar to today, and (ii) the mother fluid
was dominated by freshwater (80-95%) as evidenced by the 87Sr/86Sr, δ 138/134Ba and
δ 44/40Ca results, we can estimate the hydrologic balance during UG formation.
Model results indicate that the UG was formed in a near-terminal basin (Xe ~ 0.8) with
limited exchange with the wider Mediterranean water body during the evaporative phase.
Open system behaviour (Xe > 0.5) would result in lower oxygen and hydrogen isotope
values than those observed. The measured values also fall below a simple mixing line
between Xe = 0 and Xe = 1 in δ18O vs 17O-excess space (Fig. 7.3). Constant evaporation in
combination with recharge has been shown to drive the isotopic composition of the ambient
brine water to a steady state value below pure evaporation mixing lines (e.g. Surma et al.,
2018). It is therefore likely that freshwater basin recharge occurred during humid phases at
194 Upper Evaporites of the Messinian Salinity Crisis
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3Cycle
1
5 21 3 4 5 6 76'
6
''
R
L
G
21 3 4 5 6 7 8 9 10
1
1
1
2
1
4
1
3
1
5 21 3 4 5 6 76'
6
''
A
g
e
 (M
a
)
O
b
liq
u
ity
 
E
T
P
A
rid
W
e
t
δ
1
8O
 (‰
, V
P
D
B
)
S
ite
 9
8
2
S
ite
 6
5
4
S
ite
 3
7
2
2
0
2
1
2
2
2
3
1
0
1
2
1
4
1
6
1
8
2
0
2
2
δ
1
8O
 S
O
4
(‰
)
δ
3
4S
 
(‰
)
5
0
1
0
0
1
5
00
Height (m)
M
a
rls
G
y
p
su
m
 S
u
p
e
rco
n
e
s
B
a
n
d
e
d
 G
y
p
su
m
M
a
ssiv
e
 G
y
p
su
m
 
S
a
n
d
sto
n
e
 b
e
d
s
T
G
3
2
T
G
2
4
T
G
2
2
T
G
2
0
T
G
1
4 T
G
1
2
5.3
5.4
5.5
5.6
5.7
5.8
5.96
6.1
6
3
0
3
6
0.08
-0.04
0
0.04
0.08
0.38
0.4
0.42
2
2.5
3
3.5
-
-
-
(La
ska
r e
t a
l., 2
0
0
4
)
E
c
c
e
n
tric
ity
C
lim
a
tic
 P
re
c
e
s
s
io
n
 
Fig. 7.8. Mean intra-cycle δ 34S and δ18OSO4 data are plotted against their designated cycle. Error bars define
±1σ of the data in each cycle. The vertical grey bars indicate the ±1σ range of the data from all
PLG and all UG cycles, respectively. Each gypsum bed is plotted in relation to its proposed orbital
tuning (Laskar et al., 2004; Lugli et al., 2010; Manzi et al., 2013), where gypsum precipitation
correlates with insolation minima. The benthic δ18O (Hodell et al., 2001; modified by Drury et al.,
2018) is also shown, and the major glacial stages are indicated. Note that the the position of data
from the RLG, ODP Leg 107 site 654 and DSDP Leg 42A site 372 is arbitrary and does not reflect
the age of the deposit.
7.4 Discussion 195
the sub-precessional scale, overprinting the large evaporative flux.
In semi-restricted settings where mother fluids are undergoing intense evaporative con-
centration, strontium and calcium isotopes of gypsum will be correlated if mixing of seawa-
ter and meteoric water occur, but uncorrelated if Rayleigh distillation predominates (Hens-
ley, 2006). By directly comparing the calcium measurements with 87Sr/86Sr in the same
sample, it is possible to distinguish if the shift toward higher values are related to the influx
of seawater or Rayleigh distillation. In UG cycle 4, there is no measurable variance in
87Sr/86Sr results (Fig. 7.5), indicating a significant shift in the relative flux of meteoric
and/or marine fluids to the basin is therefore unlikely. The δ 44/40Ca, however, ranges from
a low value of 0.14h at the base of the cycle to the greatest value of 0.54h, found near
the top (Fig. 7.5). The progressive increase in δ 44/40Ca values from the top of the cycle
to the bottom is indicative of substantial Rayleigh distillation of calcium in a semi-enclosed
basin. This is in contrast to PLG cycles that displayed limited isotopic variability (Fig. 6.5).
The progressive increase in δ 44/40Ca from the base of cycle 4 to the top is also observed in
the δ18O and δD of the gypsum mother water and salinities deduced from fluid inclusions,
driven by increased aridity associated with periods of precession maxima.
Variable Rayleigh distillation and isotope fractionation effects are also evidenced by
the poor correlation between δ 138/134Ba and δ 44/40Ca results (Fig. 7.9). The δ 44/40Ca
fractionation associated with gypsum formation results in an offset of between 0.5h and
2.2h in a closed system, leading to enrichment in δ 44/40Ca of the residual fluid (Harouaka
et al., 2014; Blättler and Higgins, 2014). This fractionation may also vary due to additional
factors such as temperature, precipitation rate, and crystal morphology (Harouaka et al.,
2014). Similarly, the incorporation of the light isotope during mineral formation is expected
to lead to enrichment of 138Ba in the residual fluid. While stable barium isotope fractionation
has been investigated in experimental materials and various minerals (e.g. von Allmen et al.,
2010; van Zuilen et al., 2016), further study is needed to determine the isotope fractionation
upon incorporation into gypsum. Barite, a well-known carrier for barium removal and
remineralization (e.g. Griffith and Paytan, 2012), is a good candidate for a first order
comparison owing to similar mineral structure (e.g. Böttcher et al., 2018). Laboratory
precipitation of barite has indicated a fractionation of 1.00032 ±0.00005 (von Allmen et al.,
2010), and Böttcher et al. (2018) found barite formed with a similar fractionation during
the transformation of gypsum to BaSO4. Thus, the variable nature of the gypsum-water
fractionation during Ba and Ca incorporation likely results in the poor direct correlation
between the δ 138/134Ba and δ 44/40Ca systems. When combined with 87Sr/86Sr, however,
they provide important information to confirm the freshwater influence in the parent fluids
of the UG.
196 Upper Evaporites of the Messinian Salinity Crisis
Finally, sulfur and oxygen isotopes of sulfate also show distinctive trends within gypsum
beds. The sulfate isotope compositions are high at the base of each bed and decrease up-
cycle (Fig. 7.5), and δ 34S and δ 18OSO4 are systematically lower and higher compared
to the PLG, respectively (Fig. 7.8). The range of δ 34S and δ 18OSO4 values of the UG
are positively correlated (Fig. 7.9) and vary considerably (20.8h to 22.3h for δ 34S and
13.2h to 17.9h for δ 18OSO4) compared to expected normal marine values (δ 34S = ~21h,
δ 18OSO4 = ~12.5h), which is likely the result of sulfate-reducing bacteria increasing the 34S
and 18O in the residual sulfur pool. Density-driven stratification during increased freshwater
influx is known to enhance bacterial sulfate reduction (Torfstein and Turchyn, 2017), which
potentially occurred during wet stages (at precession minima). During the transition from
precession minima to precession maxima, the regional shift from wet to dry conditions may
have caused sea level changes within the Mediterranean Basin, enhancing water column
mixing and massive precipitation of gypsum. Indeed, García-Veigas et al. (2018) interprets
the decreasing upwards δ 34S and δ 18OSO4 trends within each of the UG cycles to reflect
increased diffusion, or mixing, with an upper oxygenated water layer that increasingly
occurs towards the top of each cycle.
Overall, the isotopic values indicate that the UG formed from a dominantly freshwater
fluid that underwent intense evaporation. The evaporation likely occurred in a semi-enclosed
environment, and led to significant Rayleigh distillation of the fluid as well as water column
destabilisation during the precipitation of each of the gypsum cycles. These results differ
from those of the PLG, where isotope results indicate deposition occurred at a steady state
and under a greater influence of marine-sourced fluids.
7.4.3 Depositional environment of DSDP/ODP sites 654 and 372
Previous studies have compared the onshore units to those of the offshore deposits based on
observations of crystal morphology and Sr isotope stratigraphy (e.g. Roveri et al., 2014b).
Our measured 87Sr/86Sr values of gypsum from site 654 range from 0.708730 to 0.708806
and overlap the range of 87Sr/86Sr from the UG of Sicily (0.708719 and 0.708795), but
are significantly lower than the values obtained from the PLG (Fig. 7.6). The trend in
87Sr/86Sr seen down-core from ~0.708750 at the top to ~0.708800 at the base (Fig. 7.5) is
similar to the isotopic variability of the onshore UG from the top of the section to the base
(mean 87Sr/86Sr of cycle 2 =0.708783; mean 87Sr/86Sr of cycle 7 = 0.708725) (Fig. 7.6).
As both successions consist of similarly low strontium isotope values as well as similar
cumulate gypsum laminites, Roveri et al. (2014b) concluded that evaporate deposits from
onshore and offshore sites formed from a unified water body with a single, evolving isotopic
composition.
7.4 Discussion 197
Our δ 44/40Ca, δ18O and δD of gypsum mother water values show significant differences
between the onshore and offshore deposits, questioning the singular evaporitic basin theory
outlined by Roveri et al. (2014a). The range of gypsum mother water values from both site
372 (1.6 < δ18O < 3.0; 7.5 < δD < 12.2h) and site 654 (0.0 < δ18O < 4.5; -0.4 < δD <
25.9h) overlap with those of the PLG and UG, but the total range is significantly smaller
and values are greater than SMOW (Fig. 7.3). The slopes of the δ18O and δD data from the
two sites are also significantly different to the PLG and UG (site 654 = 5.3; site 372 = 2.8;
PLG = 6.3; UG = 7.3), and the 17O-excess and d-excess values fall below the evaporation
trajectory expected from the simple evaporation of water. Similar to the onshore UG, these
results can only be achieved from recharge driving the water’s isotopic composition along
a trajectory below the evaporation trend (e.g. Surma et al., 2018) (Fig. 7.3). The slopes
of the evaporation trajectories intercept with SMOW, strongly suggesting that this recharge
was composed of Atlantic waters.
Calcium isotope values at site 654 are measurably lower than those of the onshore UG
(Fig. 7.6). Additionally, there is a progressive increase in δ 44/40Ca down-core at site
654 (Fig. 7.5). The relatively high δ 44/40Ca values at the base of the sampled interval
are opposite to that expected from progressive enrichment up-section due to the Rayleigh
distillation of calcium. This provides further evidence of basin recharge overprinting the
process of evaporative enrichment during the period after major salt deposition (Blättler and
Higgins, 2014).
The sulfur and oxygen isotope values of sulfate are also measurably different at site
654 compared to the onshore UG. The highest δ 34S values (22.1h to 23.4h) and δ 18OSO4
values (17.7h to 20.3h) are measured in the laminated gypsum of site 654 (Fig. 7.8). The
values are indicative of intense sulfate-reducing bacterial activity during gypsum deposition,
resulting in an increase of both oxygen and sulfur isotope ratios of the remaining aqueous
sulfate. These conclusions are supported by SEM observations which have revealed the
abundance of bacteria trapped within the gypsum crystals (Pierre and Rouchy, 1990; Panieri
et al., 2010).
The isotope data from the onshore UG and sites 654 and 372 can be reconciled if
gypsum precipitated in an evaporitic system of coalescent units with differing degrees of
hydrological restriction. In order to satisfy the isotopic variability at all sites, the Mediter-
ranean Sea level was likely below the Atlantic connection, and strong fluctuations in the
pycnocline due to intense evaporation was likely. Salinity fluctuations were also large over
a precessional cycle (brackish to >150 g l-1). This is confirmed by the presence of ostracod
taxa such as Cyprideis, which tolerate much higher salinities (~2 g l-1to 120 g l-1; e.g. Gitter
et al., 2015, and references therein) and have been recovered in deposits from both deep
198 Upper Evaporites of the Messinian Salinity Crisis
-0.20
-0.10
0.00
0.10
0.20
0.30
0.40
0.50
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
-0.20
-0.10
0.00
0.10
0.20
0.30
0.40
0.50
0.708700 0.708800 0.708900 0.709000
0.708650
0.708700
0.708750
0.708800
0.708850
0.708900
0.708950
0.709000
0.709050
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
A
B
C
δ44/40Ca (‰)
δ13
8/
13
4 B
a
87Sr/86Sr
δ44/40Ca (‰)
87
S
r/8
6 S
r
δ13
8/
13
4 B
a
Upper Gypsum
Primary Lower Gypsum
% Freshwater
100%
90%
50%
0%
0%
50%
100%
90%
0%
100%
90%
50%
Rayleigh 
Fractionation
R
ay
le
ig
h
Fr
ac
tio
na
tio
n
Ra
yle
igh
Fra
ctio
nat
ion
Fig. 7.9. Cross-plots of (A) δ138/134Ba vs 87Sr/86Sr, (B) δ138/134Ba vs δ44/40Ca, and (C) δ44/40Ca vs
87Sr/86Sr data measured in the same sample. Red lines define mixing between seawater (0%) and
freshwater (100%). δ138/134Ba vs 87Sr/86Sr and δ44/40Ca vs 87Sr/86Sr display clear mixing trends
between the PLG and UG samples, and δ138/134Ba and δ44/40Ca show differing degrees of Rayleigh
fractionation. The combined effect of both Rayleigh fractionation and fractionation associated with
gypsum formation results in a poor direct correlation between δ138/134Ba and δ44/40Ca.
7.5 Conclusions 199
and marginal settings (e.g. Stoica et al., 2016, and references therein). Oceanic inflow to
the Mediterranean, while weakened, was persistent, as evidenced from near marine δ 34S
and δ 18OSO4 isotopes and the progressive decrease in
87Sr/86Sr, both up-section in the UG
and site 654. During periods of precession maxima, an increased ratio of evaporation to
inflow resulted in mixing and partial drawdown of the Mediterranean water mass due to
the highly restricted Mediterranean-Atlantic connection, resulting in gypsum precipitation
both in deep and intermediate settings. Within each unit, large variations occurred in the
hydraulic regime as evidenced by the changes in the redox potentials. In these semi-
restricted basins, activity of sulfate reducing bacteria, evidenced by the variations values
of sulfate, was likely conditioned by the amounts of organic material available (Pierre and
Rouchy, 1990).
7.5 Conclusions
We investigated the hydrological history of Mediterranean Sea during the last stage of the
MSC using isotopic compositions of sulfate and carbonate sediments recovered from three
sites (Site 372 in Balearic Sea in the western Mediterranean, Site 654 in the Tyrrhenian
Sea, and the onshore Upper Evaporites from the Caltanissetta basin, Sicily). We show that
strontium, barium and calcium isotopes are indicative of a dominantly freshwater environ-
ment during the deposition of the upper evaporites, comprising stage 3 of the Messinian
salinity crisis. Intense evaporation was needed to reach gypsum saturation, and gypsum
precipitation occurred in heterogeneous, semi-restricted settings. Our results indicate that
both marginal and deep basin deposits are sourced from similar fluids, although each site
had specific depositional environments induced by local physiographic conditions.

References
Abriel, W., Reisdorf, K., Pannetier, J., 1990. Dehydration reactions of gypsum: A neutron and X-ray
diffraction study. Journal of Solid State Chemistry 85 (1), 23–30.
Allmann, R., Hinek, R., 2007. The introduction of structure types into the Inorganic Crystal Structure
Database ICSD. Acta Crystallographica Section A: Foundations of Crystallography 63, 412–417.
Anderson, T. F., Arthur, M. A., 1983. Stable isotopes of oxygen and hydrogen and carbon and
their application to sedimentologic and paleoenvironmental problems. In: Stable Isotopes in
Sedimentary Geology. SEPM (Society for Sedimentary Geology), Ch. 1, pp. 1–151.
Attia, O. E., Lowenstein, T., Wali, A., 1995. Middle Miocene Gypsum, Gulf of Suez: Marine or
Nonmarine? SEPM Journal of Sedimentary Research Vol. 65A (4), 614–626.
Ba¸bel, M., 2004. Models for evaporite, selenite and gypsum microbialite deposition in ancient saline
basins. Geologica Polonica Acta 54 (2), 313–337.
Ba¸bel, M., 2007. Depositional environments of a salina-type evaporite basin recorded in the
Badenian gypsum facies in the northern Carpathian Foredeep. Geological Society, London,
Special Publications 285 (1), 107–142.
Ba¸bel, M., Schreiber, B., 2014. Geochemistry of Evaporites and Evolution of Seawater. In: Treatise
on Geochemistry. Vol. 9. Elsevier, pp. 483–560.
Badens, E., Llewellyn, P., Fulconis, J. M., Jourdan, C., Veesler, S., Boistelle, R., Rouquerol, F., 1998.
Study of Gypsum Dehydration by Controlled Transformation Rate Thermal Analysis (CRTA).
Journal of Solid State Chemistry 139 (1), 37–44.
Baertschi, P., 1953. Über die relativen Unterschiede im H2 18O-Gehalt natürlicher Wässer. Helvetica
Chimica Acta 36 (6), 1352–1369.
Ballirano, P., Melis, E., 2009. Thermal behaviour and kinetics of dehydration of gypsum in air from
in situ real-time laboratory parallel-beam X-ray powder diffraction. Physics and Chemistry of
Minerals 36 (7), 391–402.
Bao, H., Cao, X., Hayles, J. A., 2016. Triple Oxygen Isotopes: Fundamental Relationships and
Applications. Annual Review of Earth and Planetary Sciences 44 (1), 463–492.
Barkan, E., Luz, B., 2005. High precision measurements of 17O/16O and 18O/16O ratios in H2O.
Rapid Communications in Mass Spectrometry 19 (24), 3737–3742.
202 References
Barkan, E., Luz, B., 2007. Diffusivity fractionations of H216O/H217O and H216O/H218O in air and
their implications for isotope hydrology. Rapid Communications in Mass Spectrometry 21 (18),
2999–3005.
Bassetti, M. A., Manzi, V., Lugli, S., Roveri, M., Longinelli, A., Ricci Lucchi, F., Barbieri, M.,
2004. Paleoenvironmental significance of Messinian post-evaporitic lacustrine carbonates in the
northern Apennines, Italy. Sedimentary Geology 172 (1-2), 1–18.
Bath, A., Darling, W., George, I., Milodowski, A., 1987. and changes during progressive hydration
of a Zechstein anhydrite formation. Geochimica et Cosmochimica Acta 51 (12), 3113–3118.
Bedford, J., Fusseis, F., Leclère, H., Wheeler, J., Faulkner, D., 2017. A 4D view on the evolution of
metamorphic dehydration reactions. Scientific Reports 7 (1), 1–7.
Benedix, G., Leshin, L., Farquhar, J., Jackson, T., Thiemens, M., 2003. Carbonates in
CM2 chondrites: constraints on alteration conditions from oxygen isotopic compositions and
petrographic observations. Geochimica et Cosmochimica Acta 67 (8), 1577–1588.
Bhattacharya, T., Chiang, J. C., Cheng, W., 2017. Ocean-atmosphere dynamics linked to 800-1050
CE drying in mesoamerica. Quaternary Science Reviews 169, 263–277.
Billo, S. M., 1987. Petrology and kinetics of gypsum-anhydrite transitions. Journal of Petroleum
Geology 10 (1), 73–85.
Blaauw, M., Christen, J. A., 2011. Flexible paleoclimate age-depth models using an autoregressive
gamma process. Bayesian Analysis 6 (3), 457–474.
Blättler, C. L., Higgins, J. A., 2014. Calcium isotopes in evaporites record variations in Phanerozoic
seawater SO4 and Ca. Geology 42 (8), 711–714.
Bodnar, R., 1993. Revised equation and table for determining the freezing point depression of H2O-
NaCl solutions. Geochimica et Cosmochimica Acta 57 (3), 683–684.
Bontognali, T., Vasconcelos, C., Warthmann, R., Bernasconi, S., Dupraz, C., Strohmenger, C.,
McKenzie, J., 2010. Dolomite formation within microbial mats in the coastal sabkha of Abu
Dhabi (United Arab Emirates). Sedimentology 57 (3), 824–844.
Bontognali, T. R., Vasconcelos, C., Warthmann, R. J., Dupraz, C., Bernasconi, S. M., McKenzie,
J. A., 2008. Microbes produce nanobacteria-like structures, avoiding cell entombment. Geology
36 (8), 663.
Borsetti, A., Curzi, P., Landuzzi, V., Mutti, M., Ricci Lucchi, F., Sartori, R., Tomadin, L., Zuffa,
G., 1990. Messinian and Pre-Messinian Sediments from ODP Leg 107 Sites 652 and 654 in the
Tyrrhenian Sea: Sedimentologic and Petrographic Study and Possible Comparisons with Italian
Sequences. In: Proceedings of the Ocean Drilling Program, 107 Scientific Results. Ocean Drilling
Program.
Böttcher, M. E., Neubert, N., Escher, P., von Allmen, K., Samankassou, E., Nägler, T. F., 2018. Multi-
isotope (Ba, C, O) partitioning during experimental carbonatization of a hyper-alkaline solution.
References 203
Chemie der Erde 78 (2), 241–247.
Bradbury, H. J., Turchyn, A. V., 2018. Calcium isotope fractionation in sedimentary pore fluids from
ODP Leg 175: Resolving carbonate recrystallization. Geochimica et Cosmochimica Acta 236,
121–139.
Brand, W. A., Geilmann, H., Crosson, E. R., Rella, C. W., 2009. Cavity ring-down spectroscopy
versus high-temperature conversion isotope ratio mass spectrometry; a case study on δ2H and δ18O
of pure water samples and alcohol/water mixtures. Rapid Communications in Mass Spectrometry
23 (12), 1879–1884.
Bridgestock, L., Hsieh, Y. T., Porcelli, D., Homoky, W. B., Bryan, A., Henderson, G. M., 2018.
Controls on the barium isotope compositions of marine sediments. Earth and Planetary Science
Letters 481, 101–110.
Broecker, W. S., 1982. Tracers in the Sea. Eldigio.
Buck, B., Van Hoesen, J., 2005. Assessing the applicability of isotopic analysis of pedogenic gypsum
as a paleoclimate indicator, Southern New Mexico. Journal of Arid Environments 60 (1), 99–114.
Calmels, D., Gaillardet, J., Brenot, A., France-Lanord, C., 2007. Sustained sulfide oxidation by
physical erosion processes in the Mackenzie River basin: Climatic perspectives. Geology 35 (11),
1003–1006.
Canfield, D. E., Sorensen, K. B., Oren, A., 2004. Biogeochemistry of a gypsum-encrusted microbial
ecosystem. Geobiology 2 (3), 133–150.
Cao, X., Liu, Y., 2011. Equilibrium mass-dependent fractionation relationships for triple oxygen
isotopes. Geochimica et Cosmochimica Acta 75 (23), 7435–7445.
Cao, Z., Siebert, C., Hathorne, E. C., Dai, M., Frank, M., 2016. Constraining the oceanic barium
cycle with stable barium isotopes. Earth and Planetary Science Letters 434, 1–9.
Cappa, C. D., 2003. Isotopic fractionation of water during evaporation. Journal of Geophysical
Research 108 (D16), 4525.
Carnevale, G., Caputo, D., Landini, W., 2006. Late Miocene fish otoliths from the Colombacci
Formation (Northern Apennines, Italy): Implications for the Messinian ’Lago-mare’ event.
Geological Journal 41 (5), 537–555.
Caruso, A., Pierre, C., Blanc-Valleron, M. M., Rouchy, J. M., 2015. Carbonate deposition
and diagenesis in evaporitic environments: The evaporative and sulphur-bearing limestones
during the settlement of the Messinian Salinity Crisis in Sicily and Calabria. Palaeogeography,
Palaeoclimatology, Palaeoecology 429, 136–162.
Cendón, D., Peryt, T., Ayora, C., Pueyo, J., Taberner, C., 2004. The importance of
recycling processes in the Middle Miocene Badenian evaporite basin (Carpathian foredeep):
palaeoenvironmental implications. Palaeogeography, Palaeoclimatology, Palaeoecology 212 (1-
2), 141–158.
204 References
Chang, H., Huang, P. J., Hou, S. C., 1999. Application of thermo-Raman spectroscopy to study
dehydration of CaSO4 ·2H2O and CaSO4 ·0.5H2O. Materials Chemistry and Physics 58 (1), 12–
19.
Chappell, J., Shackleton, N. J., 1986. Oxygen isotopes and sea level. Nature 324, 137–140.
Charola, A. E., Pühringer, J., Steiger, M., 2007. Gypsum: a review of its role in the deterioration of
building materials. Environmental Geology 52 (2), 339–352.
Chen, F., Turchyn, A. V., Kampman, N., Hodell, D., Gázquez, F., Maskell, A., Bickle, M., 2016.
Isotopic analysis of sulfur cycling and gypsum vein formation in a natural CO2 reservoir. Chemical
Geology 436, 72–83.
Christensen, A. N., Olesen, M., Cerenius, Y., Jensen, T. R., 2008. Formation and Transformation
of Five Different Phases in the CaSO4-H2O System: Crystal Structure of the Subhydrate β -
CaSO4 ·0.5H2O and Soluble Anhydrite CaSO4. Chemistry of Materials 20 (6), 2124–2132.
CIESM, 2008. The Messinian Salinity Crisis from mega-deposits to microbiology - A consensus
report. In: In: Briand, F. (Ed.). Ciesm Workshop Monographs no. 33. p. 168.
Claypool, G. E., Holser, W. T., Kaplan, I. R., Sakai, H., Zak, I., 1980. The age curves of sulfur and
oxygen isotopes in marine sulfate and their mutual interpretation.
Clayton, R. N., Mayeda, T. K., 1984. The oxygen isotope record in Murchison and other
carbonaceous chondrites. Earth and Planetary Science Letters 67 (2), 151–161.
Cohelo, A., 2007. TOPAS-Academic. Coelho Software, Brisbane, Australia.
Conley, R., Bundy, W., 1958. Mechanism of gypsification. Geochimica et Cosmochimica Acta 15 (1-
2), 57–72.
Craig, H., 1961. Isotopic Variations in Meteoric Waters. Science (New York, N.Y.) 133, 1702–1703.
Criss, R. E., 1999. Principles of Stable Isotope Distribution. Oxford University Press, Oxford.
Curtis, J. H., Hodell, D. A., Brenner, M., 1996. Climate Variability on the Yucatan Peninsula
(Mexico) during the Past 3500 Years, and Implications for Maya Cultural Evolution. Quaternary
Research 46 (01), 37–47.
D’Ans, J., 1968. Der Ubergangspunkt Gips-Anhydrit. Kali Steinsalz 5, 109–111.
Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus.
De Lange, G. J., Boelrijk, N. A. I. M., Catalano, G., Corselli, C., Klinkhammer, G. P., Middelburg,
J. J., Müller, D. W., Ullman, W. J., Van Gaans, P., Woittiez, J. R. W., 1990. Sulphate-related
equilibria in the hypersaline brines of the Tyro and Bannock Basins, eastern Mediterranean.
Marine Chemistry 31 (1-3), 89–112.
De Lange, G. J., Krijgsman, W., 2010. Messinian salinity crisis: A novel unifying shallow
gypsum/deep dolomite formation mechanism. Marine Geology 275 (1-4), 273–277.
Decima, A., McKenzie, J. A., Schreiber, B. C., 1988. The origin of "evaporitive" limestones; an
References 205
example from the Messinian of Sicily (Italy). Journal of Sedimentary Research 58 (2), 256–272.
Dela Pierre, F., Bernardi, E., Cavagna, S., Clari, P., Gennari, R., Irace, A., Lozar, F., Lugli, S.,
Manzi, V., Natalicchio, M., Roveri, M., Violanti, D., 2011. The record of the Messinian salinity
crisis in the Tertiary Piedmont Basin (NW Italy): The Alba section revisited. Palaeogeography,
Palaeoclimatology, Palaeoecology 310 (3-4), 238–255.
Dela Pierre, F., Clari, P., Natalicchio, M., Ferrando, S., Giustetto, R., Lozar, F., Lugli, S.,
Manzi, V., Roveri, M., Violanti, D., 2014. Flocculent layers and bacterial mats in the mudstone
interbeds of the Primary Lower Gypsum unit (Tertiary Piedmont basin, NW Italy): Archives of
palaeoenvironmental changes during the Messinian salinity crisis. Marine Geology 355, 71–87.
DePaolo, D. J., 2004. Calcium Isotopic Variations Produced by Biological, Kinetic, Radiogenic and
Nucleosynthetic Processes. Reviews in Mineralogy and Geochemistry 55 (1), 255–288.
DePaolo, D. J., 2011. Surface kinetic model for isotopic and trace element fractionation during
precipitation of calcite from aqueous solutions. Geochimica et Cosmochimica Acta 75 (4), 1039–
1056.
Deuser, W., Degens, E., Guillard, R., 1968. Carbon isotope relationships between plankton and sea
water. Geochimica et Cosmochimica Acta 32 (6), 657–660.
Deutsch, Y., Nathan, Y., Sarig, S., 1994. Thermogravimetric evaluation of the kinetics of the gypsum-
hemihydrate-soluble anhydrite transitions. Journal of Thermal Analysis 42 (1), 159–174.
Di Tommaso, D., Ruiz-Agudo, E., de Leeuw, N. H., Putnis, A., Putnis, C. V., 2014. Modelling
the effects of salt solutions on the hydration of calcium ions. Phys. Chem. Chem. Phys. 16 (17),
7772–7785.
Douglas, P. M., Demarest, A. A., Brenner, M., Canuto, M. A., 2016a. Impacts of Climate Change
on the Collapse of Lowland Maya Civilization. Annual Review of Earth and Planetary Sciences
44 (1), 613–645.
Douglas, P. M., Pagani, M., Eglinton, T. I., Brenner, M., Hodell, D. A., Curtis, J. H., Ma, K. F.,
Breckenridge, A., 2014. Pre-aged plant waxes in tropical lake sediments and their influence on
the chronology of molecular paleoclimate proxy records. Geochimica et Cosmochimica Acta 141,
346–364.
Douglas, P. M. J., Brenner, M., Curtis, J. H., 2016b. Methods and future directions for
paleoclimatology in the Maya Lowlands. Global and Planetary Change 138, 3–24.
Douglas, P. M. J., Pagani, M., Canuto, M. A., Brenner, M., Hodell, D. A., Eglinton, T. I., Curtis,
J. H., 2015. Drought, agricultural adaptation, and sociopolitical collapse in the Maya Lowlands.
Proceedings of the National Academy of Sciences 112 (18), 5607–5612.
Driesner, T., Ha, T.-K., Seward, T., 2000. Oxygen and hydrogen isotope fractionation by hydration
complexes of Li+, Na+, K+, Mg2+, F– , Cl– , and Br– : a theoretical study. Geochimica et
Cosmochimica Acta 64 (17), 3007–3033.
206 References
Dronkert, H., 1977. The evaporites of the Sorbas basin. Instituto de Investigaciones Geológicas
Diputación Provincial Universidad de Barcelona (32), 55–76.
Dronkert, H., 1985. Evaporite Models and Sedimentology of Messinian and Recent Evaporites. In:
GUA Papers of Geology, Series 1, No. 24. p. 283.
Drury, A. J., Westerhold, T., Hodell, D., Röhl, U., 2018. Reinforcing the North Atlantic backbone:
Revision and extension of the composite splice at ODP Site 982. Climate of the Past 14 (3), 321–
338.
Elderfield, H., 1986. Strontium isotope stratigraphy. Palaeogeography, Palaeoclimatology,
Palaeoecology 57 (1), 71–90.
Emiliani, C., 1955. Pleistocene Temperatures. The Journal of Geology 63 (6), 538–578.
Engleman, E. E., Jackson, L. L., Norton, D. R., 1985. Determination of carbonate carbon in
geological materials by coulometric titration. Chemical Geology 53, 125–128.
Epstein, S., Buchsbaum, R., Lowenstam, H. A., Urey, H. C., 1953. Revised Carbonate-Water Isotopic
Temperature Scale. Geological Society of America Bulletin 64, 1315–1326.
Epstein, S., Mayeda, T., 1953. Variation of 18O content of waters from natural sources. Geochimica
et Cosmochimica Acta 4 (5), 213–224.
Eugster, H. P., 1980. Geochemistry of Evaporitic Lacustrine Deposits. Ann. Rev. Earth Planet. 8,
35–63.
Eugster, H. P., Hardie, L. A., 1978. Saline lakes. In: Lerman, A. (Ed.), Chemistry. Geology, and
Physics of Lakes. Springer, New York, pp. 237–293.
Evans, D. A., 2006. Proterozoic low orbital obliquity and axial-dipolar geomagnetic field from
evaporite palaeolatitudes. Nature 444 (7115), 51–55.
Evans, N. P., Turchyn, A. V., Gázquez, F., Bontognali, T. R., Chapman, H. J., Hodell, D. A., 2015.
Coupled measurements of δ18O and δD of hydration water and salinity of fluid inclusions in
gypsum from the Messinian Yesares Member, Sorbas Basin (SE Spain). Earth and Planetary
Science Letters 430, 499–510.
Fantle, M. S., DePaolo, D. J., 2007. Ca isotopes in carbonate sediment and pore fluid from ODP
Site 807A: The Ca2+(aq)-calcite equilibrium fractionation factor and calcite recrystallization rates in
Pleistocene sediments. Geochimica et Cosmochimica Acta 71 (10), 2524–2546.
Fantle, M. S., Tipper, E. T., 2014. Calcium isotopes in the global biogeochemical Ca cycle:
Implications for development of a Ca isotope proxy. Earth-Science Reviews 129, 148–177.
Farías, M. E., Contreras, M., Rasuk, M. C., Kurth, D., Flores, M. R., Poiré, D. G., Novoa, F.,
Visscher, P. T., 2014. Characterization of bacterial diversity associated with microbial mats,
gypsum evaporites and carbonate microbialites in thalassic wetlands: Tebenquiche and La Brava,
Salar de Atacama, Chile. Extremophiles 18 (2), 311–329.
References 207
Farpoor, M., Khademi, H., Eghbal, M., Krouse, H., 2004. Mode of gypsum deposition in
southeastern Iranian soils as revealed by isotopic composition of crystallization water. Geoderma
121 (3-4), 233–242.
Farquhar, J., 1998. Atmosphere-Surface Interactions on Mars: 17O Measurements of Carbonate from
ALH 84001 . Science 280 (5369), 1580–1582.
Feder, H. M., Taube, H., 1952. Ionic Hydration: An Isotopic Fractionation Technique. The Journal
of Chemical Physics 20 (8), 1335–1336.
Flecker, R., De Villiers, S., Ellam, R. M., 2002. Modelling the effect of evaporation on the salinity-
87Sr/86Sr relationship in modern and ancient marginal-marine systems: The Mediterranean
Messinian Salinity Crisis. Earth and Planetary Science Letters 203, 221–233.
Flecker, R., Ellam, R. M., 2006. Identifying Late Miocene episodes of connection and isolation in
the Mediterranean-Paratethyan realm using Sr isotopes. Sedimentary Geology 188-189, 189–203.
Flecker, R., Krijgsman, W., Capella, W., de Castro Martins, C., Dmitrieva, E., Mayser, J. P.,
Marzocchi, A., Modestu, S., Ochoa, D., Simon, D., Tulbure, M., van den Berg, B., van der Schee,
M., de Lange, G., Ellam, R., Govers, R., Gutjahr, M., Hilgen, F., Kouwenhoven, T., Lofi, J.,
Meijer, P., Sierro, F. J., Bachiri, N., Barhoun, N., Alami, A. C., Chacon, B., Flores, J. A., Gregory,
J., Howard, J., Lunt, D., Ochoa, M., Pancost, R., Vincent, S., Yousfi, M. Z., 2015. Evolution
of the Late Miocene Mediterranean-Atlantic gateways and their impact on regional and global
environmental change. Earth-Science Reviews 150, 365–392.
Fontes, J., Gonfiantini, R., 1967. Fractionnment isotopique de I’hydrogene dans I’eau de
crystallization du gypse. Comptes Rendus de l’Académie des Sciences Paris 265, 4–6.
Fontes, J.-C., Letolle, R., Nesteroff, W., Ryan, W., 1973. Oxygen, Carbon, Sulfur, and Hydrogen
Stable Isotopes in Carbonate and Sulfate Mineral Phases of Neogene Evaporites, Sediments, and
in Interstitial Waters. In: Initial Reports of the Deep Sea Drilling Project, 13. U.S. Government
Printing Office.
Fortuin, A. R., Krijgsman, W., 2003. The Messinian of the Nijar Basin (SE Spain): Sedimentation,
depositional environments and paleogeographic evolution. Sedimentary Geology 160 (1-3), 213–
242.
Freyer, D., Voigt, W., 2003. Crystallization and Phase Stability of CaSO4 and CaSO4- Based Salts.
Monatshefte fur Chemie 134 (5), 693–719.
Friedman, I., 1953. Deuterium content of natural waters and other substances. Geochimica et
Cosmochimica Acta 4 (1-2), 89–103.
Friedman, I., O’Neil, J., 1977. Compilation of stable isotope fractionation factors of geochemical
interest. In: Fleischer, M. (Ed.), Data of Geochemistry. United States Geological Survey
Professional Paper, p. 12.
Fusseis, F., Schrank, C., Liu, J., Karrech, A., Llana-Fúnez, S., Xiao, X., Regenauer-Lieb, K., 2012.
208 References
Pore formation during dehydration of a polycrystalline gypsum sample observed and quantified
in a time-series synchrotron X-ray micro-tomography experiment. Solid Earth 3 (1), 71–86.
Garcia-Castellanos, D., Villaseñor, A., 2011. Messinian salinity crisis regulated by competing
tectonics and erosion at the Gibraltar arc. Nature 480 (7377), 359–363.
García-Guinea, J., Morales, S., Delgado, A., Recio, C., Calaforra, J., 2002. Formation of gigantic
gypsum crystals. Journal of the Geological Society 159 (4), 347–350.
García-Ruiz, J. M., Villasuso, R., Ayora, C., Canals, A., Otálora, F., 2007. Formation of natural
gypsum megacrystals in Naica, Mexico. Geology 35 (4), 327.
García-Veigas, J., Cendón, D. I., Gibert, L., Lowenstein, T. K., Artiaga, D., 2018. Geochemical
indicators in Western Mediterranean Messinian evaporites: Implications for the salinity crisis.
Marine Geology 403 (October 2017), 197–214.
Garofalo, P. S., Fricker, M. B., Günther, D., Forti, P., Mercuri, A.-M., Loreti, M., Capaccioni, B.,
2010. Climatic control on the growth of gigantic gypsum crystals within hypogenic caves (Naica
mine, Mexico)? Earth and Planetary Science Letters 289 (3-4), 560–569.
Gat, J. R., 1996. Oxygen and hydrogen isotopes in the hydrological cycle. Annual Review of Earth
and Planetary Sciences 24 (1), 225–262.
Gázquez, F., Calaforra, J. M., 2014. The Gypsum Karst of Sorbas, Betic Chain. In: F. Gutiérrez,
M. G. (Ed.), Landscapes and Landforms of Spain. Springer, pp. 127–135.
Gázquez, F., Calaforra, J. M., Evans, N. P., Hodell, D. A., 2017a. Using stable isotopes (δ17O, δ18O
and δD) of gypsum hydration water to ascertain the role of water condensation in the formation
of subaerial gypsum speleothems. Chemical Geology 452, 34–46.
Gázquez, F., Calaforra, J.-M., Forti, P., Rull, F., Martínez-Frías, J., 2012. Gypsum-
carbonate speleothems from Cueva de las Espadas (Naica mine, Mexico): mineralogy and
palaeohydrogeological implications. International Journal of Speleology 41 (2), 211–220.
Gázquez, F., Calaforra, J. M., Stoll, H., Sanna, L., Forti, P., Lauritzen, S.-E., Delgado, A., Rull,
F., Martínez-Frías, J., 2013. Isotope and trace element evolution of the naica aquifer (Chihuahua,
Mexico) over the past 60,000 yr revealed by speleothems. Quaternary Research 80 (03), 510–521.
Gázquez, F., Evans, N. P., Hodell, D. A., 2017b. Precise and accurate isotope fractionation factors
(α 17O, α 18O and αD) for water and CaSO4 ·2H2O (gypsum). Geochimica et Cosmochimica Acta
198, 259–270.
Gázquez, F., Mather, I., Rolfe, J., Evans, N. P., Herwartz, D., Staubwasser, M., Hodell, D. A., 2015.
Simultaneous analysis of 17O/16O, 18O/16O and 2H/1H of gypsum hydration water by cavity ring-
down laser spectroscopy. Rapid Communications in Mass Spectrometry 29 (21), 1997–2006.
Gázquez, F., Morellón, M., Bauska, T., Herwartz, D., Surma, J., Moreno, A., Staubwasser, M.,
Valero-Garcés, B., Delgado-Huertas, A., Hodell, D. A., 2018. Triple oxygen and hydrogen
isotopes of gypsum hydration water for quantitative paleo-humidity reconstruction. Earth and
References 209
Planetary Science Letters 481, 177–188.
Gibson, J., Birks, S., Yi, Y., 2016. Stable isotope mass balance of lakes: a contemporary perspective.
Quaternary Science Reviews 131, 316–328.
Gitter, F., Gross, M., Piller, W. E., 2015. Sub-Decadal Resolution in Sediments of Late Miocene
Lake Pannon Reveals Speciation of Cyprideis (Crustacea, Ostracoda). PLOS ONE 10 (4), 1–21.
Gliozzi, E., Ceci, M. E., Grossi, F., Ligios, S., 2007. Paratethyan Ostracod immigrants in Italy during
the Late Miocene. Geobios 40 (3), 325–337.
Goldschmidt, V. M., 1937. The principles of distribution of chemical elements in minerals and rocks.
The seventh Hugo Müller Lecture, delivered before the Chemical Society on March 17th, 1937. J.
Chem. Soc., 655–673.
Goldstein, R. H., Reynolds, T. J., 1994. Systematics of fluid inclusions in diagenetic minerals. In:
SEPM Short Course 31, Society for Sedimentary Geology. p. 199.
Gonfiantini, R., Fontes, J. C., 1963. Oxygen isotopic fractionation in the water of crystallization of
gypsum. Nature 200, 644–646.
Grauel, A.-L., Hodell, D. A., Bernasconi, S. M., 2016. Quantitative estimates of tropical temperature
change in lowland Central America during the last 42 ka. Earth and Planetary Science Letters 438,
37–46.
Griffith, E. M., Paytan, A., 2012. Barite in the ocean - occurrence, geochemistry and
palaeoceanographic applications. Sedimentology 59 (6), 1817–1835.
Grossman, E. L., Ku, T.-L., 1986. Oxygen and carbon isotope fractionation in biogenic aragonite:
Temperature effects. Chemical Geology: Isotope Geoscience section 59, 59–74.
Guay, C., Falkner, K., 1997. Barium as a tracer of Arctic halocline and river waters. Deep Sea
Research Part II: Topical Studies in Oceanography 44 (8), 1543–1569.
Gussone, N., Bohm, F., Eisenhauer, A., Dietzel, M., Heuser, A., Teichert, B. M., Reitner, J.,
Wörheide, G., Dullo, W.-C., 2005. Calcium isotope fractionation in calcite and aragonite.
Geochimica et Cosmochimica Acta 69 (18), 4485–4494.
Gussone, N., Eisenhauer, A., Heuser, A., Dietzel, M., Bock, B., Bohm, F., Spero, H. J., Lea,
D. W., Bijma, J., Nagler, T. F., 2003. Model for kinetic effects on calcium isotope fractionation
in inorganic aragonite and cultured planktonic foraminifera. Geochimica et Cosmochimica Acta
67 (7), 1375–1382.
Hardie, L. A., 1967. The gypsum-anhydrite equilibrium at one atmosphere pressure. The American
Mineralogist 52, 171–200.
Hardie, L. A., 1991. On the Significance of Evaporites. Annual Review of Earth and Planetary
Sciences 19 (1), 131–168.
Hardie, L. A., Eugster, H. P., 1971. The Depositional Environment of Marine Evaporites: a Case for
210 References
Shallow, Clastic Accumulation. Sedimentology 16 (3-4), 187–220.
Hardie, L. A., Eugster, H. P., 1980. Evaporation of Seawater: Calculated Mineral Sequences. Science
208 (4443), 498–500.
Harouaka, K., Eisenhauer, A., Fantle, M. S., 2014. Experimental investigation of Ca isotopic
fractionation during abiotic gypsum precipitation. Geochimica et Cosmochimica Acta 129, 157–
176.
Harrison, T. N., 2012. Experimental VNIR reflectance spectroscopy of gypsum dehydration:
Investigating the gypsum to bassanite transition. American Mineralogist 97 (4), 598–609.
Haug, G. H., 2003. Climate and the Collapse of Maya Civilization. Science 299 (5613), 1731–1735.
Hay, W. W., Migdisov, A., Balukhovsky, A. N., Wold, C. N., Flögel, S., Söding, E., 2006. Evaporites
and the salinity of the ocean during the Phanerozoic: Implications for climate, ocean circulation
and life. Palaeogeography, Palaeoclimatology, Palaeoecology 240 (1-2), 3–46.
Hensley, T. M., 2006. Calcium Isotopic Variation in Marine Evaporites and Carbonates: Applications
to Late Miocene Mediterranean Brine Chemistry and Late Cenozoic Calcium Cycling in the
Oceans. Ph.D. thesis, University of California.
Hernandez-Molina, F. J., Stow, D. A. V., Alvarez-Zarikian, C. A., Acton, G., Bahr, A., Balestra,
B., Ducassou, E., Flood, R., Flores, J.-A., Furota, S., Grunert, P., Hodell, D., Jimenez-Espejo, F.,
Kim, J. K., Krissek, L., Kuroda, J., Li, B., Llave, E., Lofi, J., Lourens, L., Miller, M., Nanayama,
F., Nishida, N., Richter, C., Roque, C., Pereira, H., Goni, M. F. S., Sierro, F. J., Singh, A. D.,
Sloss, C., Takashimizu, Y., Tzanova, A., Voelker, A., Williams, T., Xuan, C., 2014. Onset of
Mediterranean outflow into the North Atlantic. Science 344 (6189), 1244–1250.
Herwartz, D., Surma, J., Voigt, C., Assonov, S., Staubwasser, M., 2017. Triple oxygen isotope
systematics of structurally bonded water in gypsum. Geochimica et Cosmochimica Acta 209,
254–266.
Hildyard, R. C., Llana-fúnez, S., Wheeler, J., Faulkner, D. R., Prior, D. J., 2011. Electron backscatter
diffraction (EBSD) analysis of bassanite transformation textures and crystal structure produced
from experimentally deformed and dehydrated gypsum. Journal of Petrology 52 (5), 839–856.
Hilgen, F., Kuiper, K., Krijgsman, W., Snel, E., Laan, E. V. D., van der Laan, E., 2007. Astronomical
tuning as the basis for high resolution chronostratigraphy: The intricate history of the Messinian
Salinity Crisis. Stratigraphy 4, 231–238.
Hodell, D. A., Brenner, M., Curtis, J. H., 2005a. Terminal Classic drought in the northern Maya
lowlands inferred from multiple sediment cores in Lake Chichancanab (Mexico). Quaternary
Science Reviews 24, 1413–1427.
Hodell, D. A., Brenner, M., Curtis, J. H., Guilderson, T., 2001. Solar Forcing of Drought Frequency
in the Maya Lowlands. Science 292 (5520), 1367–1370.
Hodell, D. A., Brenner, M., Curtis, J. H., Medina-González, R., Ildefonso-Chan Can, E., Albornaz-
References 211
Pat, A., Guilderson, T. P., 2005b. Climate change on the Yucatan Peninsula during the Little Ice
Age. Quaternary Research 63, 109–121.
Hodell, D. A., Curtis, J. H., Brenner, M., 1995. Possible role of climate in the collapse of Classic
Maya civilization. Nature 375 (6530), 391–394.
Hodell, D. A., Mueller, P. A., McKenzie, J. A., Mead, G. A., 1989. Strontium isotope stratigraphy
and geochemistry of the late Neogene ocean. Earth and Planetary Science Letters 92, 165–178.
Hodell, D. A., Turchyn, A. V., Wiseman, C. J., Escobar, J., Curtis, J. H., Brenner, M., Gilli, A.,
Mueller, A. D., Anselmetti, F., Ariztegui, D., Brown, E. T., 2012. Late Glacial temperature and
precipitation changes in the lowland Neotropics by tandem measurement of δ18O in biogenic
carbonate and gypsum hydration water. Geochimica et Cosmochimica Acta 77, 352–368.
Hoggarth, J. A., Restall, M., Wood, J. W., Kennett, D. J., 2017. Drought and Its Demographic Effects
in the Maya Lowlands. Current Anthropology 58 (1), 82–113.
Holland, H. D., Horita, J., Seyfried, W. E., 1996. On the secular variations in the composition of
Phanerozoic marine potash evaporites. Geology 24 (11), 993.
Horita, J., 1989. Stable isotope fractionation factors of water in hydrated saline mineral-brine
systems. Earth and Planetary Science Letters 95 (1-2), 173–179.
Horita, J., Wesolowski, D. J., 1994. Liquid-vapor fractionation of oxygen and hydrogen isotopes of
water from the freezing to the critical temperature. Geochimica et Cosmochimica Acta 58 (16),
3425–3437.
Horner, T. J., Kinsley, C. W., Nielsen, S. G., 2015. Barium-isotopic fractionation in seawater
mediated by barite cycling and oceanic circulation. Earth and Planetary Science Letters 430, 511–
522.
Horner, T. J., Pryer, H. V., Nielsen, S. G., Crockford, P. W., Gauglitz, J. M., Wing, B. A., Ricketts,
R. D., 2017. Pelagic barite precipitation at micromolar ambient sulfate. Nature Communications
8 (1), 1342.
Hsieh, Y. T., Henderson, G. M., 2017. Barium stable isotopes in the global ocean: Tracer of Ba
inputs and utilization. Earth and Planetary Science Letters 473, 269–278.
Hsu, K., Montadert, L., 1978. Initial Reports of the Deep Sea Drilling Project, 42 Pt. 1. Vol. 42 Pt. 1
of Initial Reports of the Deep Sea Drilling Project. U.S. Government Printing Office, Washington.
Hsü, K. J., Montadert, L., Bernoulli, D., Cita, M. B., Erickson, A., Garrison, R. E., Kidd, R. B.,
Mèlierés, F., Müller, C., Wright, R., 1977. History of the Mediterranean salinity crisis. Nature
267 (5610), 399–403.
Hsü, K. J., Ryan, W. B. F., Cita, M. B., 1973. Late Miocene Desiccation of the Mediterranean. Nature
242 (5395), 240–244.
Husson, J. M., Higgins, J. A., Maloof, A. C., Schoene, B., 2015. Ca and Mg isotope constraints on
the origin of Earth’s deepest δ13C excursion. Geochimica et Cosmochimica Acta 160, 243–266.
212 References
Kasprzyk, A., Jasinska, B., 1998. Isotopic composition of the crystaUization water of gypsum in
the Badenian of the northern Carpathian Foredeep: a case study from the cores Przyborow 1 and
Strzegom 143. Geological Quarterly 42 (3), 301–310.
Kennett, D. J., Breitenbach, S. F. M., Aquino, V. V., Asmerom, Y., Awe, J., Baldini, J. U. L.,
Bartlein, P., Culleton, B. J., Ebert, C., Jazwa, C., Macri, M. J., Marwan, N., Polyak, V., Prufer,
K. M., Ridley, H. E., Sodemann, H., Winterhalder, B., Haug, G. H., 2012. Development and
Disintegration of Maya Political Systems in Response to Climate Change. Science 338 (6108),
788–791.
Khademi, H., Mermut, A., Krouse, H., 1997. Isotopic composition of gypsum hydration water in
selected landforms from central Iran. Chemical Geology 138 (3-4), 245–255.
Khalil, A. A. A., 1982. Kinetics of gypsum dehydration. Thermochimica Acta 55 (2), 201–208.
Kluge, T., Affek, H. P., 2012. Quantifying kinetic fractionation in Bunker Cave speleothems using
∆47. Quaternary Science Reviews 49, 82–94.
Knauth, L., Beeunas, M. A., 1986. Isotope geochemistry of fluid inclusions in Permian halite with
implications for the isotopic history of ocean water and the origin of saline formation waters.
Geochimica et Cosmochimica Acta 50 (3), 419–433.
Krijgsman, W., Capella, W., Simon, D., Hilgen, F. J., Kouwenhoven, T. J., Meijer, P. T., Sierro, F. J.,
Tulbure, M. A., van den Berg, B. C., van der Schee, M., Flecker, R., 2018. The Gibraltar Corridor:
Watergate of the Messinian Salinity Crisis. Marine Geology 403 (March), 238–246.
Krijgsman, W., Fortuin, A. R., Hilgen, F. J., Sierro, F. J., 2001. Astrochronology for the Messinian
Sorbas basin (SE Spain) and orbital (precessional) forcing for evaporite cyclicity. Sedimentary
Geology 140 (1-2), 43–60.
Krijgsman, W., Hilgen, F. J., Raffi, I., Sierro, F. J., Wilson, D. S., 1999. Chronology, causes and
progression of the Messinian salinity crisis. Nature 400 (August), 652–655.
Krijgsman, W., Meijer, P. T., 2008. Depositional environments of the Mediterranean "Lower
Evaporites" of the Messinian salinity crisis: Constraints from quantitative analyses. Marine
Geology 253 (3-4), 73–81.
Kushnir, J., 1980. The coprecipitation of strontium, magnesium, sodium, potassium and chloride ions
with gypsum. An experimental study. Geochimica et Cosmochimica Acta 44 (10), 1471–1482.
Lachniet, M. S., 2009. Climatic and environmental controls on speleothem oxygen-isotope values.
Quaternary Science Reviews 28 (5-6), 412–432.
Landais, A., Barkan, E., Luz, B., 2008. Record of δ18O and 17O-excess in ice from Vostok Antarctica
during the last 150,000 years. Geophysical Research Letters 35 (2), L02709.
Landais, A., Barkan, E., Yakir, D., Luz, B., 2006. The triple isotopic composition of oxygen in leaf
water. Geochimica et Cosmochimica Acta 70 (16), 4105–4115.
Landais, A., Steen-Larsen, H., Guillevic, M., Masson-Delmotte, V., Vinther, B., Winkler, R., 2012.
References 213
Triple isotopic composition of oxygen in surface snow and water vapor at NEEM (Greenland).
Geochimica et Cosmochimica Acta 77, 304–316.
Laskar, J., Robutel, P., Joutel, F., Gastineau, M., Correia, A. C. M., Levrard, B., 2004. A long-term
numerical solution for the insolation quantities of the Earth. Astronomy & Astrophysics 428 (1),
261–285.
Lea, D., Boyle, E., 1989. Barium content of benthic foraminifera controlled by bottom-water
composition. Nature 338 (6218), 751–753.
Li, J., Li, M., Fang, X., Zhang, G., Zhang, W., Liu, X., 2017a. Isotopic composition of gypsum
hydration water in deep Core SG-1, western Qaidam basin (NE Tibetan Plateau), implications for
paleoclimatic evolution. Global and Planetary Change 155, 70–77.
Li, S., Levin, N. E., Soderberg, K., Dennis, K. J., Caylor, K. K., 2017b. Triple oxygen isotope
composition of leaf waters in Mpala, central Kenya. Earth and Planetary Science Letters 468,
38–50.
Lin, Y., Clayton, R. N., Gröning, M., 2010. Calibration of δ17O and δ18O of international
measurement standards - VSMOW, VSMOW2, SLAP, and SLAP2. Rapid Communications in
Mass Spectrometry 24 (6), 773–776.
Lloyd, R. M., 1968. Oxygen isotope behavior in the Sulfate-Water System. Journal of Geophysical
Research 73, 6099.
Lofi, J., Sage, F., Deverchere, J., Loncke, L., Maillard, A., Gaullier, V., Thinon, I., Gillet, H.,
Guennoc, P., Gorini, C., 2011. Refining our knowledge of the Messinian salinity crisis records
in the offshore domain through multi-site seismic analysis. Bulletin de la Societe Geologique de
France 182 (2), 163–180.
Londeix, L., Benzakour, M., Suc, J.-P., Turon, J.-L., 2007. Messinian palaeoenvironments and
hydrology in Sicily (Italy): The dinoflagellate cyst record. Geobios 40 (3), 233–250.
Longinelli, A., 1979. Isotope geochemistry of some Messinian evaporates: Paleoenvironmental
implications. Palaeogeography, Palaeoclimatology, Palaeoecology 29, 95–123.
Longinelli, A., Cortecci, G., 1970. Isotopic abundance of oxygen and sulfur in sulfate ions from river
water. Earth and Planetary Science Letters 7 (4), 376–380.
Lou, W., Guan, B., Wu, Z., 2011. Dehydration behavior of FGD gypsum by simultaneous TG and
DSC analysis. Journal of Thermal Analysis and Calorimetry 104 (2), 661–669.
Lowenstein, T. K., Timofeef, M. N., Brennan, S. T., Hardie, L. A., Demicco, R. V., 2001. Oscillations
in phaneroizoic seawater chemisrty: Evidence from fluid inclusions. Science 294, 1086–1088.
Lu, F. H., Meyers, W. J., 2003. Sr, S, and O(SO4) Isotopes and the Depositional Environments of the
Upper Miocene Evaporites, Spain. Journal of Sedimentary Research 73, 444–450.
Lu, F. H., Meyers, W. J., Schoonen, M. A., 1997. Minor and trace element analyses on gypsum: an
experimental study. Chemical Geology 142 (1-2), 1–10.
214 References
Lu, F. H., Meyers, W. J., Schoonen, M. A., 2001. S and O(SO4) isotopes, simultaneous modeling, and
environmental significance of the Nijar messinian gypsum, Spain. Geochimica et Cosmochimica
Acta 65 (18), 3081–3092.
Lugli, S., Bassetti, M. A., Manzi, V., Barbieri, M., Longinelli, A., Roveri, M., 2007. The Messinian
‘Vena del Gesso’ evaporites revisited: characterization of isotopic composition and organic matter.
Geological Society, London, Special Publications 285 (1), 179–190.
Lugli, S., Gennari, R., Gvirtzman, Z., Manzi, V., Roveri, M., Schreiber, B. C., 2013. Evidence of
Clastic Evaporites In the Canyons of the Levant Basin (Israel): Implications For the Messinian
Salinity Crisis. Journal of Sedimentary Research 83 (11), 942–954.
Lugli, S., Manzi, V., Roveri, M., Schreiber, B. C., 2015. The deep record of the Messinian salinity
crisis: Evidence of a non-desiccated Mediterranean Sea. Palaeogeography, Palaeoclimatology,
Palaeoecology 433, 201–218.
Lugli, S., Manzi, V., Roveri, M., Schreiber, C. B., 2010. The Primary Lower Gypsum in the
Mediterranean: A new facies interpretation for the first stage of the Messinian salinity crisis.
Palaeogeography, Palaeoclimatology, Palaeoecology 297 (1), 83–99.
Luz, B., Barkan, E., 2010. Variations of 17O/16O and 18O/16O in meteoric waters. Geochimica et
Cosmochimica Acta 74 (22), 6276–6286.
Luz, B., Barkan, E., Bender, M. L., Thiemens, M. H., Boering, K. A., 1999. Triple-isotope
composition of atmospheric oxygen as a tracer of biosphere productivity. Nature 400 (6744), 547–
550.
Madsen, I., Scarlett, N., 2008. Quantitative Phase Analysis. In: Dinnabier, R. (Ed.), Powder
Diffraction: Theory and Practice. Royal Society of Chemistry, London, pp. 298–331.
Manzi, V., Gennari, R., Hilgen, F., Krijgsman, W., Lugli, S., Roveri, M., Sierro, F. J., 2013. Age
refinement of the Messinian salinity crisis onset in the Mediterranean. Terra Nova 25 (4), 315–
322.
Manzi, V., Lugli, S., Marco, R., Schreiber, B., 2009. A new facies model for the Upper Gypsum of
Sicily (Italy): chronological and palaeoenvironmental constraints for the Messinian salinity crisis
in the Mediterranean. Sedimentology 56 (7), 1937–1960.
Marion, G. M., 1997. A theoretical evaluation of mineral stability in Don Juan Pond, Wright Valley,
Victoria Land. Antarctic Science 9 (1), 92–99.
Marshall, B., DePaolo, D., 1982. Precise age determinations and petrogenetic studies using the KCa
method. Geochimica et Cosmochimica Acta 46 (12), 2537–2545.
Matsubaya, O., 1971. Oxygen isotopic fractionation between the water of crystallization of gypsum
and the mother solution. Abstract Paper of Annual Meeting of Geochem. Soc. Japan, 17A01 (in
Japanese).
Matsubaya, O., Sakai, H., 1973. Oxygen and hydrogen isotopic study on the water of crystallization
References 215
of gypsum from the Kuroko type mineralization. Geochemical Journal 7 (3), 153–165.
Matsuhisa, Y., Goldsmith, J. R., Clayton, R. N., 1978. Mechanisms of hydrothermal crystallization
of quartz at 250°C and 15 kbar. Geochimica et Cosmochimica Acta 42 (2), 173–182.
McAdie, H., 1964. The effect of water vapor upon the dehydration of CaSO4 ·2H2O. Canadian
Journal of Chemistry 42, 792–801.
McArthur, J., Howarth, R., Bailey, T., 2001. Strontium isotope stratigraphy: LOWESS version 3:
best fit to the marine Sr-isotope curve for 0-509 Ma and accompanying look-up table for deriving
numerical age. The Journal of Geology 109 (2), 155–170.
McCaffrey, M. A., Lazar, B., Holland, H. D., 1987. The evaporation path of seawater and the
coprecipitation of Br– and K+ with halite. Journal of sedimentary petrology 57 (5), 928–938.
McConnaughey, T., 1989. 13C and 18O isotopic disequilibrium in biological carbonates: I. Patterns.
Geochimica et Cosmochimica Acta 53 (1), 151–162.
McConnell, J. D. C., Astill, D. M., Hall, P. L., 1987. The pressure dependence of the dehydration of
gypsum to bassanite. Mineralogical Magazine 51 (361), 453–457.
McCulloch, M., Fallon, S., Wyndham, T., Hendy, E., Lough, J., Barnes, D., 2003. Coral record
of increased sediment flux to the inner Great Barrier Reef since European settlement. Nature
421 (6924), 727–730.
McKenzie, J., Jenkyns, H., Bennet, G., 1979. Stable isotope study of the cyclic diatomite - Claystones
from the tripoli formation, Sicily: A prelude to the Messenian salinity crisis. Palaeogeography,
Palaeoclimatology, Palaeoecology 29 (143), 125–141.
Medina-Elizalde, M., Burns, S. J., Lea, D. W., Asmerom, Y., von Gunten, L., Polyak, V., Vuille,
M., Karmalkar, A., 2010. High resolution stalagmite climate record from the Yucatán Peninsula
spanning the Maya terminal classic period. Earth and Planetary Science Letters 298 (1-2), 255–
262.
Medina-Elizalde, M., Rohling, E. J., 2012. Collapse of Classic Maya Civilization Related to Modest
Reduction in Precipitation. Science 335 (6071), 956–959.
Meilijson, A., Steinberg, J., Hilgen, F., Bialik, O. M., Waldmann, N. D., Makovsky, Y., 2018. Deep-
basin evidence resolves a 50-year-old debate and demonstrates synchronous onset of Messinian
evaporite deposition in a non-desiccated Mediterranean. Geology 46 (3), 243–246.
Merlivat, L., 1978. Molecular diffusivities of H2 16O, HD16O, and H2 18O in gases. The Journal of
Chemical Physics 69 (6), 2864.
Merlivat, L., Jouzel, J., 1979. Global climatic interpretation of the deuterium-oxygen 18 relationship
for precipitation. Journal of Geophysical Research 84 (C8), 5029.
Merlivat, L., Nief, G., 1967. Fractionnement isotopique lors des changements d’état solide-vapeur et
liquide-vapeur de l’eau à des températures inférieures à 0°C. Tellus 19 (1), 122–127.
216 References
Miller, M. F., 2002. Isotopic fractionation and the quantification of 17O anomalies in the oxygen
three-isotope system. Geochimica et Cosmochimica Acta 66 (11), 1881–1889.
Misra, S., Owen, R., Kerr, J., Greaves, M., Elderfield, H., 2014. Determination of 11B by HR-ICP-
MS from mass limited samples: Application to natural carbonates and water samples. Geochimica
et Cosmochimica Acta 140, 531–552.
Modestou, S., Simon, D., Gutjahr, M., Marzocchi, A., Kouwenhoven, T. J., Ellam, R. M., Flecker, R.,
2017. Precessional variability of 87Sr/86Sr in the late Miocene Sorbas Basin: An interdisciplinary
study of drivers of interbasin exchange. Paleoceanography 32 (6), 531–552.
Mook, W. G., 2000. Environmental Isotopes in the Hydrological Cycle: Principles and Applications,
Vol. I: Introduction - Theory, Methods, Review. UNESCO, IAEA, Paris.
Muller, D., Mueller, P. a., McKenzie, J. a., 1990. Strontium Isotopic Ratios As Fluid Tracers in
Messinian Evaporites of the Tyrrhenian Sea (Western Mediterranean Sea). Proceedings of the
Ocean Drilling Program, Scientific Results, Vol. 107 107, 603–614.
Müller, D. W., Mueller, P. A., 1991. Origin and age of the Mediterranean Messinian evaporites:
implications from Sr isotopes. Earth and Planetary Science Letters 107 (1), 1–12.
Natalicchio, M., Dela Pierre, F., Lugli, S., Lowenstein, T. K., Feiner, S. J., Ferrando, S., Manzi, V.,
Roveri, M., Clari, P., 2014. Did Late Miocene (Messinian) gypsum precipitate from evaporated
marine brines? Insights from the Piedmont Basin (Italy). Geology 42 (3), 179–182.
Oi, T., Morimoto, H., 2013. Oxygen and Hydrogen Isotopic Preference in Hydration Spheres of
Chloride and Sulfate Ions. Zeitschrift für Physikalische Chemie 227 (6-7), 807–819.
Orszag-Sperber, F., jun 2006. Changing perspectives in the concept of “lago-mare” in mediterranean
late miocene evolution. Sedimentary Geology 188-189, 259–277.
Ortí, F., 2011. Selenite facies in marine evaporites: a review. In: Quaternary carbonate and evaporite
sedimentary facies and their ancient analogues. John Wiley & Sons Ltd., Chichester, West Sussex,
UK, pp. 431–463.
Ortí, F., Pueyo, J. J., Geisler-Cussey, D., Dulau, N., 1984. Evaporitic sedimentation in the coastal
salinas of Santa Pola (Alicante, Spain). Revista d’Investigacions Geologiques 38 (39), 169–220.
Ossorio, M., Van Driessche, A. E., Pérez, P., García-Ruiz, J. M., 2014. The gypsum-anhydrite
paradox revisited. Chemical Geology 386, 16–21.
Ostroff, A., 1964. Conversion of gypsum to anhydrite in aqueous salt solutions. Geochimica et
Cosmochimica Acta 28 (9), 1363–1372.
Pack, A., Herwartz, D., 2014. The triple oxygen isotope composition of the Earth mantle and
understanding ∆17O variations in terrestrial rocks and minerals. Earth and Planetary Science
Letters 390, 138–145.
Packter, A., 1974. The precipitation of calcium sulphate dihydrate from aqueous solution. Journal of
Crystal Growth 21 (2), 191–194.
References 217
Palmer, M., Edmond, J., 1992. Controls over the strontium isotope composition of river water.
Geochimica et Cosmochimica Acta 56 (5), 2099–2111.
Panieri, G., Lugli, S., Manzi, V., Roveri, M., Schreiber, B., Palinska, K. A., 2010. Ribosomal RNA
gene fragments from fossilized cyanobacteria identified in primary gypsum from the late Miocene,
Italy. Geobiology 8 (2), 101–111.
Parkhurst, D. L., Appelo, C. A. J., 2013. Description of input and examples for PHREEQC version
3: a computer program for speciation, batch-reaction, one-dimensional transport, and inverse
geochemical calculations. Tech. rep., US Geological Survey.
Passey, B. H., Hu, H., Ji, H., Montanari, S., Li, S., Henkes, G. A., Levin, N. E., 2014. Triple oxygen
isotopes in biogenic and sedimentary carbonates. Geochimica et Cosmochimica Acta 141, 1–25.
Paulik, F., Paulik, J., Arnold, M., 1992. Thermal decomposition of gypsum. Thermochimica Acta
200 (C), 195–204.
Paytan, A., 1998. Sulfur Isotopic Composition of Cenozoic Seawater Sulfate. Science 282, 1459–
1462.
Perry, E., Velazquez-Oliman, G., Marin, L., 2002. The Hydrogeochemistry of the Karst Aquifer
System of the Northern Yucatan Peninsula, Mexico. International Geology Review 44 (3), 191–
221.
Pierre, C., 2018. The isotopic record of gypsum diagenesis in diluted solutions: Observations in
natural salinas and experiments. Chemical Geology 493, 451–457.
Pierre, C., Fontes, J., 1978. Isotope Composition of Messinian Sediments from the Mediterranean
Sea as Indicators of Paleoenvironments and Diagensis. In: Initial Reports of the Deep Sea Drilling
Project, 42 Pt. 1. U.S. Government Printing Office.
Pierre, C., Rouchy, J. M., 1990. Sedimentary and diagenetic evolution of Messinian evaporites in the
Tyrrhenian Sea (ODP Leg 107, sites 652, 653, and 654); petrographic, mineralogical, and stable
isotope records. Proceedings of the Ocean Drilling Program, Tyrrhenian Sea, covering Leg 107 of
the cruises of the Drilling Vessel JOIDES Resolution, Malaga, Spain to Marseille, France, sites
650-656, 20 December 1985-18 February 1986. 107, 187–210.
Playà, E., Recio, C., Mitchell, J., 2005. Extraction of gypsum hydration water for oxygen isotopic
analysis by the guanidine hydrochloride reaction method. Chemical Geology 217, 89–96.
Prasad, P. S. R., Chaitanya, V. K., Prasad, K. S., Rao, D. N., 2005. Direct formation of the γ-CaSO4
phase in dehydration process of gypsum: In situ FTIR study. American Mineralogist 90 (4), 672–
678.
Prasad, P. S. R., Pradhan, A., Gowd, T. N., 2001. In situ micro-Raman investigation of dehydration
mechanism in natural gypsum. Current Science 80 (9), 1203–1207.
Pretet, C., van Zuilen, K., Nägler, T. F., Reynaud, S., Böttcher, M. E., Samankassou, E.,
2015. Constraints on barium isotope fractionation during aragonite precipitation by corals. The
218 References
Depositional Record 1 (2), 118–129.
Prieto-Taboada, N., Gómez-Laserna, O., Martínez-Arkarazo, I., Olazabal, M. Á., Madariaga, J. M.,
2014. Raman Spectra of the Different Phases in the CaSO4-H2O System. Analytical Chemistry
86 (20), 10131–10137.
Putnis, A., Winkler, B., Fernandez-Diaz, L., 1990. In situ IR spectroscopic and thermogravimetric
study of the dehydration of gypsum. Mineralogical Magazine 54 (374), 123–128.
Rapin, W., Meslin, P. Y., Maurice, S., Vaniman, D., Nachon, M., Mangold, N., Schroder, S.,
Gasnault, O., Forni, O., Wiens, R. C., Martinez, G. M., Cousin, A., Sautter, V., Lasue, J., Rampe,
E. B., Archer, D., 2016. Hydration state of calcium sulfates in Gale crater, Mars: Identification of
bassanite veins. Earth and Planetary Science Letters 452, 197–205.
Reghizzi, M., Gennari, R., Douville, E., Lugli, S., Manzi, V., Montagna, P., Roveri, M., Sierro,
F. J., Taviani, M., 2017. Isotope stratigraphy (87Sr/86Sr, δ18O, δ13C) of the Sorbas basin (Betic
Cordillera, Spain): Paleoceanographic evolution across the onset of the Messinian salinity crisis.
Palaeogeography, Palaeoclimatology, Palaeoecology 469, 60–73.
Reghizzi, M., Lugli, S., Manzi, V., Rossi, F. P., Roveri, M., 2018. Orbitally Forced Hydrological
Balance During the Messinian Salinity Crisis: Insights From Strontium Isotopes (87Sr/86Sr) in
the Vena del Gesso Basin (Northern Apennines, Italy). Paleoceanography and Paleoclimatology
33 (7), 716–731.
Rennie, V. C. F., Turchyn, A. V., 2014. Controls on the abiotic exchange between aqueous sulfate
and water under laboratory conditions. Limnology and Oceanography: Methods 12 (4), 166–173.
Rigaudier, T., Lécuyer, C., Gardien, V., Suc, J. P., Martineau, F., 2011. The record of temperature,
wind velocity and air humidity in the δD and δ18O of water inclusions in synthetic and Messinian
halites. Geochimica et Cosmochimica Acta 75 (16), 4637–4652.
Risi, C., Landais, A., Bony, S., Jouzel, J., Masson-Delmotte, V., Vimeux, F., 2010. Understanding the
17O excess glacial-interglacial variations in Vostok precipitation. Journal of Geophysical Research
Atmospheres 115 (10).
Robertson, K., Bish, D., 2013. Constraints on the distribution of CaSO4 ·nH2O phases on Mars and
implications for their contribution to the hydrological cycle. Icarus 223 (1), 407–417.
Rosell, L., Orti, F., Kasprzyk, A., Playa, E., Peryt, T. M., 1998. Strontium geochemistry of Miocene
primary gypsum; Messinian of southeastern Spain and Sicily and Badenian of Poland. Journal of
Sedimentary Research 68 (1), 63–79.
Rossi, C., Vilas, L., Arias, C., 2015. The Messinian marine to nonmarine gypsums of Jumilla
(Northern Betic Cordillera, SE Spain): Isotopic and Sr concentration constraints on the origin
of parent brines. Sedimentary Geology 328, 96–114.
Rouchy, J., Orszag-Sperber, F., Blanc-Valleron, M.-M., Pierre, C., Rivière, M., Combourieu-Nebout,
N., Panayides, I., 2001. Paleoenvironmental changes at the Messinian–Pliocene boundary in
References 219
the eastern Mediterranean (southern Cyprus basins): significance of the Messinian Lago-Mare.
Sedimentary Geology 145 (1-2), 93–117.
Rouchy, J. M., Caruso, A., 2006. The Messinian salinity crisis in the Mediterranean basin: A
reassessment of the data and an integrated scenario. Sedimentary Geology 188-189, 35–67.
Roveri, M., Bertini, A., Stefano, A. D., Gennari, R., Gliozzi, E., Grossi, F., Iaccarino, S. M., Lugli,
S., Manzi, V., Taviani, M., A, V. G. P. U., 2008. A high-resolution stratigraphic framework for the
latest Messinian events in the Mediterranean area. Stratigraphy 5, 323–342.
Roveri, M., Flecker, R., Krijgsman, W., Lofi, J., Lugli, S., Manzi, V., Sierro, F. J., Bertini, A.,
Camerlenghi, A., De Lange, G., Govers, R., Hilgen, F. J., Hübscher, C., Meijer, P. T., Stoica, M.,
2014a. The Messinian Salinity Crisis: Past and future of a great challenge for marine sciences.
Marine Geology 352, 25–58.
Roveri, M., Gennari, R., Lugli, S., Manzi, V., 2009. The Terminal Carbonate Complex: the record of
sea-level changes during the Messinian salinity crisis. GeoActa 8, 63–77.
Roveri, M., Lugli, S., Manzi, V., Gennari, R., Schreiber, B. C., 2014b. High-resolution strontium
isotope stratigraphy of the Messinian deep Mediterranean basins: Implications for marginal to
central basins correlation. Marine Geology 349, 113–125.
Rudge, J. F., Reynolds, B. C., Bourdon, B., 2009. The double spike toolbox. Chemical Geology
265 (3-4), 420–431.
Ryan, W. B. F., 2009. Decoding the Mediterranean salinity crisis. Sedimentology 56 (1), 95–136.
Saad, N., Hsiao, G., Chapellet-volpini, L., Vu, D., 2013. Two-Pronged Approach to Overcome
Spectroscopically Interfering Organic Compounds with Isotopic Water Analysis. Geophysical
Research Abstracts (15), EGU2013–13296.
Saha, A., Lee, J., Pancera, S. M., Bräeu, M. F., Kempter, A., Tripathi, A., Bose, A., 2012.
New Insights into the Transformation of Calcium Sulfate Hemihydrate to Gypsum Using Time-
Resolved Cryogenic Transmission Electron Microscopy. Langmuir 28 (30), 11182–11187.
Sarma, L. P., Prasad, P. S. R., Ravikumar, N., 1998. Raman spectroscopic study of phase transitions
in natural gypsum. Journal of Raman Spectroscopy 29 (9), 851–856.
Schoenemann, S. W., Schauer, A. J., Steig, E. J., 2013. Measurement of SLAP2 and GISP δ17O
and proposed VSMOW-SLAP normalization for δ17O and 17O excess. Rapid Communications in
Mass Spectrometry 27 (5), 582–590.
Schreiber, C., Kingsman, D. J. J., 1975. New Observations on the Pleistocene Evaporites of
Montallegro, Sicily and a Modern Analog. SEPM Journal of Sedimentary Research Vol. 45.
Selli, R., 1960. Il Messiniano Mayer-Eymar (1867). Proposta di un neostratotipo. Giornale di
Geologia 28, 1–33.
Sheppard, S. M. F., Schwarcz, H. P., 1970. Fractionation of carbon and oxygen isotopes and
magnesium between coexisting metamorphic calcite and dolomite. Contributions to Mineralogy
220 References
and Petrology 26 (3), 161–198.
Sierro, F., Hilgen, F., Krijgsman, W., Flores, J., 2001. The Abad composite (SE Spain): a Messinian
reference section for the Mediterranean and the APTS. Palaeogeography, Palaeoclimatology,
Palaeoecology 168 (1-2), 141–169.
Sierro, F. J., Flores, J. A., Zamarreño, I., Vázquez, A., Utrilla, R., Francés, G., Hilgen, F. J.,
Krijgsman, W., 1999. Messinian pre-evaporite sapropels and precession-induced oscillations in
western Mediterranean climate. Marine Geology 153, 137–146.
Simon, D., 2017. Gateway exchange, climatic forcing and circulation of the Mediterranean Sea
during the late Miocene: A model perspective. No. 129.
Simon, D., Marzocchi, A., Flecker, R., Lunt, D. J., Hilgen, F. J., Meijer, P. T., 2017. Quantifying
the Mediterranean freshwater budget throughout the late Miocene: New implications for sapropel
formation and the Messinian Salinity Crisis. Earth and Planetary Science Letters 472, 25–37.
Simon, D., Meijer, P. T., 2017. Salinity stratification of the Mediterranean Sea during the Messinian
crisis: A first model analysis. Earth and Planetary Science Letters 479, 366–376.
Sofer, Z., 1975. Isotopic Compositions of Hydration Water in Gypsum. Ph.D. thesis, The Feinberg
Graduate School, Rehovot, Israel (in Hebrew).
Sofer, Z., 1978. Isotopic composition of hydration water in gypsum. Geochimica et Cosmochimica
Acta 42, 1141–1149.
Sofer, Z., Gat, J., 1972. Activities and concentrations of oxygen-18 in concentrated aqueous salt
solutions: Analytical and geophysical implications. Earth and Planetary Science Letters 15 (3),
232–238.
Soler-Bientz, R., Watson, S., Infield, D., 2010. Wind characteristics on the Yucatan Peninsula based
on short term data from meteorological stations. Energy Conversion and Management 51 (4),
754–764.
Speranza, G., Cosentino, D., Tecce, F., Faccenna, C., 2013. Paleoclimate reconstruction during the
Messinian evaporative drawdown of the Mediterranean Basin: Insights from microthermometry
on halite fluid inclusions. Geochemistry, Geophysics, Geosystems 14, 5054–5077.
Spötl, C., Vennemann, T. W., 2003. Continuous-flow isotope ratio mass spectrometric analysis of
carbonate minerals. Rapid communications in mass spectrometry : RCM 17, 1004–1006.
Stark, S. C., O’Grady, B. V., Burton, H. R., Carpenter, D., 2003. Frigidly concentrated seawater and
the evolution of Antarctic saline lakes. Australian Journal of Chemistry 56 (2-3), 181–186.
Stawski, T. M., Van Driessche, A. E., Ossorio, M., Diego Rodriguez-Blanco, J., Besselink, R.,
Benning, L. G., 2016. Formation of calcium sulfate through the aggregation of sub-3 nanometre
primary species. Nature Communications 7, 1–9.
Steig, E. J., Gkinis, V., Schauer, A. J., Schoenemann, S. W., Samek, K., Hoffnagle, J., Dennis, K. J.,
Tan, S. M., 2014. Calibrated high-precision 17O-excess measurements using cavity ring-down
References 221
spectroscopy with laser-current-tuned cavity resonance. Atmospheric Measurement Techniques
7 (8), 2421–2435.
Stoica, M., Krijgsman, W., Fortuin, A., Gliozzi, E., 2016. Paratethyan ostracods in the Spanish Lago-
Mare: More evidence for interbasinal exchange at high Mediterranean sea level. Palaeogeography,
Palaeoclimatology, Palaeoecology 441, 854–870.
Strydom, C. A., Hudson-Lamb, D. L., Potgieter, J. H., Dagg, E., 1995. The thermal dehydration of
synthetic gypsum. Thermochimica Acta 269-270 (C), 631–638.
Surma, J., Assonov, S., Bolourchi, M. J., Staubwasser, M., 2015. Triple oxygen isotope signatures
in evaporated water bodies from the Sistan Oasis, Iran. Geophysical Research Letters 42 (20),
8456–8462.
Surma, J., Assonov, S., Herwartz, D., Voigt, C., Staubwasser, M., 2018. The evolution of 17O-excess
in surface water of the arid environment during recharge and evaporation. Scientific Reports 8 (1),
4972.
Tan, H.-B., Huang, J.-Z., Zhang, W.-J., Liu, X.-Q., Zhang, Y.-F., Kong, N., Zhang, Q., 2014.
Fractionation of hydrogen and oxygen isotopes of gypsum hydration water and assessment of
its geochemical indications. Australian Journal of Earth Sciences 61 (6), 793–801.
Tarutani, T., Clayton, R. N., Mayeda, T. K., 1969. The effect of polymorphism and magnesium
substitution on oxygen isotope fractionation between calcium carbonate and water. Geochimica
et Cosmochimica Acta 33 (8), 987–996.
Taube, H., 1954. Use of Oxygen Isotope Effects in Study of Hydration of Ions. The Journal of
Physical Chemistry 58 (7), 523–528.
Thode, H., Monster, J., 1965. Sulfur isotope geochemistry of petroleum, evaporites and ancient seas.
American Association of Petroleum Geologists Memoir 4, 367–377.
Thompson, J. B., Ferris, F. G., 1990. Cyanobacterial precipitation of gypsum, calcite, and magnesite
from natural alkaline lake water. Geology 18 (10), 995.
Tipper, E. T., Gaillardet, J., Galy, A., Louvat, P., Bickle, M. J., Capmas, F., 2010. Calcium isotope
ratios in the world’s largest rivers: A constraint on the maximum imbalance of oceanic calcium
fluxes. Global Biogeochemical Cycles 24 (3).
Tipper, E. T., Galy, A., Bickle, M. J., 2008. Calcium and magnesium isotope systematics in rivers
draining the Himalaya-Tibetan-Plateau region: Lithological or fractionation control? Geochimica
et Cosmochimica Acta 72 (4), 1057–1075.
Topper, R. P. M., Lugli, S., Manzi, V., Roveri, M., Meijer, P. T., 2014. Precessional control of Sr ratios
in marginal basins during the Messinian Salinity Crisis? Geochemistry, Geophysics, Geosystems
15 (5), 1926–1944.
Topper, R. P. M., Meijer, P. T., 2013. A modeling perspective on spatial and temporal variations in
Messinian evaporite deposits. Marine Geology 336, 44–60.
222 References
Torfstein, A., Turchyn, A. V., 2017. Rates and Cycles of Microbial Sulfate Reduction in the Hyper-
Saline Dead Sea over the Last 200 kyrs from Sedimentary δ34S and δ18O(SO4). Frontiers in Earth
Science 5, 1–14.
Turchyn, A. V., Schrag, D. P., 2004. Oxygen isotope constraints on the sulfur cycle over the past 10
million years. Science (New York, N.Y.) 303 (5666), 2004–2007.
Uemura, R., Barkan, E., Abe, O., Luz, B., 2010. Triple isotope composition of oxygen in atmospheric
water vapor. Geophysical Research Letters 37 (4), 1–4.
Urey, H., Lowenstam, H., Epstein, S., McKinney, C., 1951. Measurement of Paleotemperatures and
Temperatures and the Southeastern United States. Bulletin of the geological society of America
62 (4), 399–416.
Urey, H. C., 1947. The thermodynamic properties of isotopic substances. Journal of the Chemical
Society (Resumed), 562–581.
Usiglio, J., 1849. Analyse de l’eau de la Méditerranée sur les côtes de France. Annales de Chimie et
de Physique 3eme ser 27, 92–107.
Van Couvering, J., Castradori, D., Cita, M., Hilgen, F., Rio, D., 2000. The base of the Zanclean Stage
and of the Pliocene Series. Episodes 23, 179–187.
van der Heijden, G. H., Pel, L., Huinink, H. P., Kopinga, K., 2011. Moisture transport and
dehydration in heated gypsum, an NMR study. Chemical Engineering Science 66 (18), 4241–
4250.
van der Laan, E., Gaboardi, S., Hilgen, F. J., Lourens, L. J., 2005. Regional climate and glacial
control on high-resolution oxygen isotope records from Ain el Beida (latest Miocene, northwest
Morocco): A cyclostratigraphic analysis in the depth and time domain. Paleoceanography 20 (1),
1–22.
Van Der Laan, E., Snel, E., De Kaenel, E., Hilgen, F. J., Krijgsman, W., 2006. No major deglaciation
across the Miocene-Pliocene boundary: Integrated stratigraphy and astronomical tuning of the
Loulja sections (Bou Regreg area, NW Morocco). Paleoceanography 21 (3), 1–27.
Van Driessche, A. E. S., Benning, L. G., Rodriguez-Blanco, J. D., Ossorio, M., Bots, P., Garcia-
Ruiz, J. M., 2012. The Role and Implications of Bassanite as a Stable Precursor Phase to Gypsum
Precipitation. Science 336 (6077), 69–72.
van Zuilen, K., Müller, T., Nägler, T. F., Dietzel, M., Küsters, T., 2016. Experimental determination
of barium isotope fractionation during diffusion and adsorption processes at low temperatures.
Geochimica et Cosmochimica Acta 186, 226–241.
Vasconcelos, C., McKenzie, J., A., Bernasconi, S., Grujic, D., Tiens, A., J., 1995. Microbial
mediation as a possible mechanism for natural dolomite formation at low temperatures. Nature
377, 220–222.
Vasconcelos, C., McKenzie, J. A., 1997. Microbial Mediation of Modern Dolomite Precipitation
References 223
and Diagenesis Under Anoxic Conditions (Lagoa Vermelha, Rio de Janeiro, Brazil). Journal of
Sedimentary Research 67, 378–390.
Vasconcelos, C., McKenzie, J. a., Warthmann, R., Bernasconi, S. M., 2005. Calibration of the
δ18O paleothermometer for dolomite precipitated in microbial cultures and natural environments.
Geology 33 (4), 317–320.
Vogel, M. B., Des Marais, D. J., Parenteau, M. N., Jahnke, L. L., Turk, K. A., Kubo, M. D., 2010.
Biological influences on modern sulfates: Textures and composition of gypsum deposits from
Guerrero Negro, Baja California Sur, Mexico. Sedimentary Geology 223 (3-4), 265–280.
Vogel, M. B., Des Marais, D. J., Turk, K. A., Parenteau, M. N., Jahnke, L. L., Kubo, M. D., 2009.
The Role of Biofilms in the Sedimentology of Actively Forming Gypsum Deposits at Guerrero
Negro, Mexico. Astrobiology 9 (9), 875–893.
von Allmen, K., Böttcher, M. E., Samankassou, E., Nägler, T. F., 2010. Barium isotope fractionation
in the global barium cycle: First evidence from barium minerals and precipitation experiments.
Chemical Geology 277 (1-2), 70–77.
Wahl, D., Byrne, R., Anderson, L., 2014. An 8700 year paleoclimate reconstruction from the
southern Maya lowlands. Quaternary Science Reviews 103, 19–25.
Wahl, D., Estrada-Belli, F., Anderson, L., 2013. A 3400 year paleolimnological record of
prehispanic human-environment interactions in the Holmul region of the southern Maya lowlands.
Palaeogeography, Palaeoclimatology, Palaeoecology 379-380, 17–31.
Wang, Y.-W., Kim, Y.-Y., Christenson, H. K., Meldrum, F. C., 2012. A new precipitation pathway
for calcium sulfate dihydrate (gypsum) via amorphous and hemihydrate intermediates. Chem.
Commun. 48 (4), 504–506.
Warren, J. K., 1982. The hydrological setting, occurrence and significance of gypsum in late
Quaternary salt lakes in South Australia. Sedimentology 29 (5), 609–637.
Warren, J. K., 2010. Evaporites through time: Tectonic, climatic and eustatic controls in marine and
nonmarine deposits. Earth-Science Reviews 98 (3-4), 217–268.
Warren, J. K., 2016. Depositional Chemistry and Hydrology. In: Evaporites. Springer International
Publishing, Cham, pp. 85–205.
Wassenaar, L. I., Van Wilgenburg, S. L., Larson, K., Hobson, K. A., 2009. A groundwater isoscape
(δD, δ18O) for Mexico. Journal of Geochemical Exploration 102 (3), 123–136.
Webster, J. W., Brook, G. A., Railsback, L. B., Cheng, H., Edwards, R. L., Alexander, C.,
Reeder, P. P., 2007. Stalagmite evidence from Belize indicating significant droughts at the time
of Preclassic Abandonment, the Maya Hiatus, and the Classic Maya collapse. Palaeogeography,
Palaeoclimatology, Palaeoecology 250 (1-4), 1–17.
Winkler, R., Landais, A., Risi, C., Baroni, M., Ekaykin, A., Jouzel, J., Petit, J. R., Prie, F.,
Minster, B., Falourd, S., 2013. Interannual variation of water isotopologues at Vostok indicates
224 References
a contribution from stratospheric water vapor. Proceedings of the National Academy of Sciences
110 (44), 17674–17679.
Yaeger, J., Hodell, D. A., 2009. The Collapse of Maya Civilization: Assessing the Interaction
of Culture, Climate, and Environment. In: Sandweiss, D. H., Quilter, J. (Eds.), El Niño,
Catastrophism, and Culture Change in Ancient America. Harvard University Press, Dumbarton
Oaks, Washington, D.C., Ch. 10, pp. 187–242.
Yoshimura, T., Kuroda, J., Lugli, S., Tamenori, Y., Ogawa, N. O., Jiménez-Espejo, F. J., Isaji, Y.,
Roveri, M., Manzi, V., Kawahata, H., Ohkouchi, N., 2016. An X-ray spectroscopic perspective
on Messinian evaporite from Sicily: Sedimentary fabrics, element distributions, and chemical
environments of S and Mg. Geochemistry, Geophysics, Geosystems 17 (4), 1383–1400.
Zeebe, R., Wolf-Gladrow, D., 2001. Stable Isotope Fractionation. CO2 in Seawater: Equilibrium,
Kinetics, Isotopes: Equilibrium, Kinetics, Isotopes.
Zimbelman, D., Rye, R., Breit, G., 2005. Origin of secondary sulfate minerals on active andesitic
stratovolcanoes. Chemical Geology 215 (1-4), 37–60.
Appendix A
Tables for Section 2.3
226 Tables for Section 2.3
Table A.1a. Isotopic analysis (δ17O, δ18O and δD values) of hydration water of a gypsum standard
(NEWGYP) extracted using the WASP and analyzed by CRDS and IRMS.
Line/date Method δ17O 1σ δ18O 1σ δD 1σ d-excess 1σ 17O−excess 1σ H2O
(h) (h) (h) (h) (per meg) (per meg) (h)
L3-28/1/15 CRDS 0.12 0.05 0.21 0.07 -51.23 0.14 -52.3 0.4 21 15 21.3
L4-29/1/15 CRDS 0.2 0.04 0.35 0.04 -50.83 0.12 -52.8 0.3 27 15 20.8
L5-30/1/15 CRDS 0.23 0.04 0.42 0.1 -50.91 0.23 -53.3 0.6 19 11 20.9
L4-30/1/15 IRMS 0.36 0.05 0.67 0.09 - - - - 7 3 20.5
L5-30/1/15 IRMS 0.35 0.01 0.65 0.02 - - - - 10 8 20.6
L2-30/1/15 IRMS 0.29 0.03 0.52 0.05 - - - - 11 3 20.9
L1-30/1/15 IRMS 0.34 0.09 0.62 0.17 - - - - 10 4 20.7
L2-30/1/15 IRMS 0.29 0.1 0.53 0.19 - - - - 9 7 20.6
L3-30/1/15 IRMS 0.17 0.11 0.32 0.2 - - - - 0 8 20.8
L4-30/1/15 IRMS 0.17 0.03 0.31 0.06 - - - - 10 4 20.8
L5-30/1/15 IRMS 0.28 0.06 0.51 0.1 - - - - 9 4 20.4
L6-30/1/15 CRDS 0.16 0.04 0.25 0.07 -50.89 0.52 -54.2 0.5 18 7 20.9
L1-03/2/15 CRDS 0.13 0.05 0.16 0.09 -50.34 0.32 -51.6 0.6 33 8 20.8
L2-03/2/15 CRDS 0.14 0.02 0.25 0.03 -51.28 0.17 -53.3 0.2 8 14 21.9
L4-04/2/15 CRDS 0.17 0.02 0.3 0.05 -51.48 0.29 -53.9 0.4 20 16 20.4
L5-05/2/15 CRDS 0.13 0.02 0.21 0.02 -51.81 0.1 -53.5 0.1 23 14 20.3
L1-09/2/15 CRDS 0.16 0.02 0.27 0.04 -51.59 0.27 -53.8 0.1 13 14 20.7
L2-09/2/15 CRDS 0.12 0.02 0.18 0.03 -51.88 0.09 -53.3 0.2 21 13 20.5
L3-10/2/15 CRDS 0.18 0.03 0.29 0.04 -51.63 0.19 -54 0.2 18 16 20.9
L6-18/2/15 CRDS 0.07 0.03 0.11 0.04 -51.8 0.18 -53.4 0.2 20 12 20.8
L1-22/2/15 CRDS 0.2 0.02 0.33 0.03 -51.24 0.2 -54.7 0.2 32 10 20.7
L3-04/2/15 CRDS 0.17 0.03 0.3 0.04 -50.39 0.23 -52.9 0.1 11 13 20.7
L4-02/3/15 CRDS 0.16 0.02 0.29 0.02 -50.8 0.08 -53.1 0.2 12 11 20.1
L1-03/3/15 CRDS 0.19 0.02 0.32 0.03 -51.27 0.19 -53.8 0.2 17 13 20.9
L2-05/3/15 CRDS 0.23 0.02 0.41 0.02 -50.36 0.35 -53.5 0.5 18 7 20.6
L3-06/3/15 CRDS 0.37 0.04 0.66 0.05 -50.8 0.18 -56.1 0.3 17 15 20.2
L4-08/3/15 CRDS 0.18 0.02 0.32 0.03 -51.74 0.13 -54.3 0.2 12 15 21.9
L5-09/3/15 CRDS 0.28 0.03 0.5 0.05 -50.64 0.19 -54.6 0.2 19 10 19.9
L6-11/3/15 CRDS 0.22 0.01 0.4 0.01 -51.31 0.1 -54.5 0.2 3 11 20.7
L1-12/3/15 CRDS 0.28 0.03 0.52 0.03 -51.16 0.06 -55.3 0.2 12 13 20.8
L6-13/3/15 CRDS 0.18 0.02 0.31 0.05 -50.63 0.36 -53.1 0.2 15 15 20.7
L3-22/3/15 CRDS 0.25 0.02 0.45 0.03 -50.5 0.3 -54.5 0.5 6 16 20.4
L4-23/3/15 CRDS 0.21 0.03 0.39 0.03 -51.29 0.18 -53.2 0.4 9 15 20.5
L6-26/3/15 CRDS 0.22 0.03 0.43 0.03 -50.28 0.21 -53.7 0.2 -5 16 20.8
L6-27/3/15 CRDS 0.24 0.01 0.41 0.02 -50.81 0.22 -54.1 0.1 18 10 20.7
L5-30/3/15 CRDS 0.21 0.03 0.37 0.03 -51.1 0.23 -54 0.2 16 12 21.2
L4-01/4/15 CRDS 0.16 0.03 0.29 0.03 -50.81 0.15 -53.1 0.4 0 12 20.9
227
Table A.1b. Table A.1a continued. Isotopic analysis (δ17O, δ18O and δD values) of hydration water of a
gypsum standard (NEWGYP) extracted using the WASP and analyzed by CRDS and IRMS.
Line/date Method δ17O 1σ δ18O 1σ δD 1σ d-excess 1σ 17O−excess 1σ H2O
(h) (h) (h) (h) (per meg) (per meg) (h)
L3-02/4/15 CRDS 0.22 0.02 0.37 0.03 -50.64 0.39 -53.7 0.2 19 15 20.8
L1-20/4/15 CRDS 0.16 0.03 0.29 0.03 -50.01 0.27 -52.4 0.4 5 13 20.5
L2-21/4/15 CRDS 0.04 0.06 0.04 0.12 -50.66 0.65 -50.3 0.4 27 9 20.3
L5-21/4/15 CRDS 0.13 0.04 0.21 0.06 -50.87 0.38 -52.6 0.3 6 13 20.9
L3-22/4/15 CRDS 0.18 0.02 0.32 0.02 -50.59 0.27 -53.1 0.2 13 15 20.9
L4-22/4/15 CRDS 0.07 0.02 0.09 0.02 -51.39 0.18 -52.4 0.2 20 14 20.8
L3-22/4/15 CRDS 0.14 0.03 0.24 0.06 -50.99 0.39 -52.8 0.4 8 13 20.9
L4-22/4/15 CRDS 0.03 0.02 0.03 0.02 -51.24 0.18 -51.7 0.1 17 14 20.7
AVG CRDS 0.18 0.31 -51.02 -53.4 15 20.7
STD ±0.07 ±0.13 ±0.49 ±1.1 ±8 ±0.4
AVG IRMS 0.27 0.49 n/a n/a 8 20.7
STD ±0.07 ±0.14 ±4 ±0.2
Table A.2. Memory effect tests and experiments of vacuum leakage in the WASP lines during GHW
extraction. The term “fresh” refers to a line in which the previous sample analyzed was close in
isotopic value to the next sample and memory effects should be minimized.
Sample/Line Line status δ17O 1σ δ18O 1σ δD 1σ 17O−excess 1σ
(h) (h) (h) (per meg)
ENR-gyp L5 “Fresh” line 12.96 0.04 24.93 0.05 34.73 0.19 -126 9
ENR-gyp L6 “Fresh” leaking line 12.78 0.04 24.58 0.05 33.56 0.39 -120 8
ENR-gyp L1 After depleted sample 13.12 0.03 25.24 0.05 35.1 0.38 -125 14
ENR-gyp L2 After depleted sample 13.08 0.03 25.14 0.05 34.66 0.36 -115 13
NEWGYP- L3 “Fresh” line 0.07 0.02 0.09 0.02 -51.39 0.18 20 14
NEWGYP-L4 “Fresh” leaking line 0.03 0.02 0.03 0.02 -51.24 0.18 17 14
DPL-gyp L1 “Fresh” line -6.79 0.04 -12.83 0.06 -149.66 0.24 1 14
DPL-gyp L2 “Fresh” line -6.89 0.02 -13.04 0.03 -150.35 0.16 10 9
DPL-gyp L5 After Enriched sample -6.75 0.03 -12.77 0.03 -150.75 0.04 11 19
DPL-gyp L6 After Enriched sample -6.78 0.02 -12.79 0.02 -150.8 0.11 -10 21
228 Tables for Section 2.3
Table A.3. Isotopic analysis (δ17O, δ18O and δD values) of hydration water of a gypsum standard
(NEWGYP) and natural samples. *Samples analyzed in replicate (in these cases, the analytical
error refers to the external precision, taking the average of the indicated number of samples).
δ1
7 O
δ1
8 O
δ D
d-
ex
ce
ss 
(‰
)
17
O-
ex
ce
ss
δ1
7 O
δ1
8 O
δ D
d-
ex
ce
ss
17
O-
ex
ce
ss 
δ1
7 O
δ1
8 O
17
O-
ex
ce
ss 
(‰
)
(‰
)
 (‰
)
 (‰
)
 (p
er 
me
g)
(‰
)
(‰
)
 (‰
)
 (‰
)
 (p
er 
me
g)
(‰
)
(‰
)
 (p
er 
me
g)
-1.
00
-1.
93
 
-53
.36
-37
.9
20
-0.
89
-1.
78
-52
.9
-38
.6
57
-1.
01
-1.
96
28
± 0
.02
± 0
.02
± 0
.21
± 0
.2
± 1
3
± 0
.01
± 0
.03
 ± 
0.1
4
± 0
.2
± 1
1
± 0
.02
± 0
.04
± 7
-0.
06
-0.
15
-41
.91
-40
.7
15
-0.
04
-0.
16
-42
.3
-42
.9
54
0.0
9
0.1
2
24
± 0
.02
± 0
.04
± 0
.23
 ± 
0.2
± 7
± 0
.02
 ± 
0.0
4
± 0
.30
 ± 
0.3
± 1
2
± 0
.01
± 0
.01
± 5
0.1
7
0.3
-51
.03
-53
.4
15
0.1
2
0.1
5
-51
.23
-53
42
0.2
8
0.5
1
8
± 0
.07
± 0
.13
± 0
.48
 ± 
1.0
± 8
± 0
.04
 ± 
0.0
5
± 0
.35
± 0
.6
± 1
2
± 0
.07
± 0
.14
± 4
1.5
5
2.9
2
-27
.71
-51
.6
-7
1.8
8
3.4
-27
.47
-54
.7
77
1.8
5
3.5
2
-11
± 0
.01
± 0
.03
± 0
.27
 ± 
0.3
± 1
5
± 0
.02
± 0
.02
± 0
.11
± 0
.1
± 1
3
± 0
.01
± 0
.02
± 1
1.6
8
3.1
3
-27
.72
-53
.2
7
1.7
4
3.1
7
-28
.31
-53
.7
61
1.9
3.6
5
± 0
.02
± 0
.02
± 0
.07
± 0
.1
± 1
5
± 0
.02
± 0
.02
± 0
.07
± 0
.2
± 1
4
± 0
.01
± 0
.03
± 5
4.4
6
8.5
3
3.7
6
-64
.5
-26
5.0
8
9.4
1
6.9
3
-68
.4
12
3
4.8
7
9.3
-32
± 0
.01
± 0
.02
± 0
.14
 ± 
0.2
± 1
2
± 0
.04
± 0
.05
± 0
.20
± 0
.4
± 1
8
± 0
.00
± 0
.01
 ± 
2
4.8
9
9.3
9.7
6
-64
.7
-17
5.3
2
9.7
4
11
.21
-66
.8
19
0
4.9
8
9.5
1
-32
± 0
.03
± 0
.05
± 0
.26
± 0
.3
± 9
± 0
.02
± 0
.02
± 0
.06
± 0
.2
± 7
± 0
.02
± 0
.03
± 2
5.6
1
10
.74
13
.21
-72
.7
-48
5.7
4
10
.74
13
.55
-72
.4
72
5.8
8
11
.26
-51
± 0
.02
± 0
.03
± 0
.16
 ± 
0.2
± 8
± 0
.04
± 0
.04
± 0
.18
 ± 
0.4
± 1
6 
± 0
.03
± 0
.07
± 5
Sa
lin
a 1
 (n
=3
)*
PI
 6A
-13
H-
2  
7-8
 cm
PI
 6C
-7H
-2 
25
-26
 cm
SB
L-
2.3
SB
L-
8 
CR
I-0
1
NE
W
GY
P 
(n
=3
6/8
)*
BG
-10
CR
DS
 (M
CM
 O
N)
CR
DS
 (M
CM
 O
FF
)
IR
M
S
    
    
   S
am
ple
229
Table A.4. Isotopic composition of water samples distilled using the WASP and analysed by CRDS.
δ1
7 O
1σ
δ1
8 O
1σ
δ D
1σ
d-
ex
ce
ss
1σ
17
O-
ex
ce
ss 
1σ
(‰
)
(‰
)
(‰
)
(‰
)
(‰
)
(‰
)
(‰
)
(‰
)
(pe
r m
eg
)
(pe
r m
eg
)
JR
W
-9.
99
0.0
1
-18
.85
0.0
2
-14
6.0
1
0.1
9
5.6
0.2
-4
13
JR
W
-W
AS
P
-10
.02
0.0
3
-18
.9
0.0
5
-14
6.1
9
0.2
5.9
0.2
2
13
BO
TT
Y
-3.
95
0.0
2
-7.
52
0.0
2
-50
.18
0.2
1
10
.6
0.2
20
13
BO
TT
Y-
W
AS
P
-3.
96
0.0
2
-7.
54
0.0
3
-50
.22
0.1
1
10
.7
0.3
24
10
SP
IT
-0.
07
0.0
2
-0.
13
0.0
3
0.1
6
0.1
4
1.5
0.3
-6
11
SP
IT
-W
AS
P
-0.
08
0.0
2
-0.
17
0.0
2
-0.
17
0.2
1.4
0.3
6
13
EN
R
5.6
4
0.0
2
10
.76
0.0
1
40
.23
0.1
-46
.4
0.1
-36
18
EN
R-
W
AS
P
5.5
5
0.0
2
10
.59
0.0
2
39
.58
0.1
2
-45
.6
0.1
-37
12
DE
PO
-03
A
3.0
3
0.0
3
5.8
1
0.0
2
21
.34
0.1
6
-25
.3
0.2
-43
21
DE
PO
-03
B
3.0
3
0.0
2
5.8
0.0
2
21
.33
0.0
8
-25
.3
0.2
-40
17
DE
PO
-06
A
4.1
5
0.0
1
7.9
5
0.0
1
34
.02
0.1
8
-29
.9
0.2
-47
10
DE
PO
-06
B
4.2
0.0
3
8.0
3
0.0
3
34
.4
0.1
1
-30
.1
0.2
-36
16
Sa
mp
le

Appendix B
Tables for Section 2.4
232 Tables for Section 2.4
Table B.1a. Fractionation factor between gypsum hydration water and the mother solution in experiments
of hydration of anhydrite at different temperatures.
Ex
pe
rim
en
t I
D
Te
mp
er
atu
re
 (o C
)
CR
DS
 an
aly
ze
r
H 2
O 
%
mo
th
er
 w
ate
r/G
HW
δ1
7 O
1σ
δ1
8 O
1σ
δ D
1σ
α
17
O
α
18
O
α
D
θ
#3
2 w
ate
r 3
 o C
3
L2
14
0-i
 P
ica
rro
wa
ter
-3.
85
0.0
4
-7.
35
0.0
7
-48
.45
0.4
7
1.0
01
98
1.0
03
73
0.9
79
0.5
30
4
#3
2 g
yp
su
m 
3 o
C
3
L2
14
0-i
 P
ica
rro
19
.7
GH
W
-1.
88
0.0
4
-3.
65
0.0
6
-68
.86
0.4
0
#3
2 w
ate
r 3
 o C
3
L2
14
0-i
 P
ica
rro
wa
ter
-3.
93
0.0
2
-7.
33
0.0
3
-49
.69
0.1
4
1.0
02
04
1.0
03
86
0.9
79
0.5
29
8
#3
2 g
yp
su
m 
3 o
C
3
L2
14
0-i
 P
ica
rro
20
.1
GH
W
-1.
81
0.1
2
-3.
51
0.2
1
-68
.78
0.2
7
#3
3 w
ate
r 3
 o C
3
L2
14
0-i
 P
ica
rro
wa
ter
-3.
84
0.0
4
-7.
32
0.0
8
-47
.92
0.2
9
1.0
01
94
1.0
03
66
0.9
78
0.5
28
7
#3
3 g
yp
su
m 
3 o
C
3
L2
14
0-i
 P
ica
rro
20
.2
GH
W
-1.
91
0.0
2
-3.
68
0.0
2
-68
.58
0.1
2
#3
3 w
ate
r 3
 o C
3
L2
14
0-i
 P
ica
rro
wa
ter
-3.
88
0.0
2
-7.
39
0.0
3
-48
.68
0.0
8
1.0
01
93
1.0
03
64
0.9
78
0.5
31
2
#3
3 g
yp
su
m 
3 o
C
3
L2
14
0-i
 P
ica
rro
20
.5
GH
W
-1.
92
0.0
2
-3.
71
0.0
2
-68
.89
0.0
9
#2
6 w
ate
r 8
 o C
8
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
50
0.0
4
-50
.03
0.3
8
-
1.0
03
75
0.9
80
-
#2
6 g
yp
su
m 
8 o
C
8
L1
10
2-i
 P
ica
rro
20
.1
GH
W
-
-
-3.
78
0.0
5
-68
.58
0.3
1
#2
9 w
ate
r 8
 o C
8
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
49
0.0
4
-49
.11
0.3
2
-
1.0
03
11
0.9
80
-
#2
9 g
yp
su
m 
8 o
C
8
L1
10
2-i
 P
ica
rro
20
.6
GH
W
-
-
-4.
40
0.0
6
-68
.12
0.3
4
#3
0 w
ate
r 8
 o C
8
L2
14
0-i
 P
ica
rro
wa
ter
-3.
82
0.0
2
-7.
28
0.0
4
-48
.34
0.1
0
1.0
01
96
1.0
03
71
0.9
79
0.5
27
2
#3
0 g
yp
su
m 
8 o
C
8
L2
14
0-i
 P
ica
rro
20
.6
GH
W
-1.
87
0.0
3
-3.
59
0.0
5
-67
.85
0.2
8
#3
0 w
ate
r 8
 o C
8
L2
14
0-i
 P
ica
rro
wa
ter
-3.
86
0.0
2
-7.
36
0.0
2
-48
.27
0.1
0
1.0
01
91
1.0
03
61
0.9
79
0.5
29
9
#3
0 g
yp
su
m 
8 o
C
8
L2
14
0-i
 P
ica
rro
20
.4
GH
W
-1.
95
0.0
1
-3.
77
0.0
1
-67
.93
0.1
7
#3
1 w
ate
r 8
 o C
8
L2
14
0-i
 P
ica
rro
wa
ter
-3.
85
0.0
3
-7.
33
0.0
5
-48
.51
0.1
0
1.0
02
06
1.0
03
90
0.9
81
0.5
30
0
#3
1 g
yp
su
m 
8 o
C
8
L2
14
0-i
 P
ica
rro
20
.9
GH
W
-1.
79
0.0
2
-3.
46
0.0
3
-66
.65
0.1
9
#3
1 w
ate
r 8
 o C
8
L2
14
0-i
 P
ica
rro
wa
ter
-3.
90
0.0
2
-7.
43
0.0
2
-48
.36
0.0
6
1.0
02
09
1.0
03
96
0.9
81
0.5
30
0
#3
1 g
yp
su
m 
8 o
C
8
L2
14
0-i
 P
ica
rro
20
.8
GH
W
-1.
81
0.0
2
-3.
50
0.0
3
-67
.02
0.1
4
#6
 w
ate
r 2
0 o
C
20
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
46
0.0
6
-50
.03
0.2
3
-
1.0
03
54
0.9
81
-
#6
 gy
ps
um
 20
 o C
20
L1
10
2-i
 P
ica
rro
20
.4
GH
W
-
-
-3.
95
0.0
3
-68
.13
0.3
0
#6
 w
ate
r 2
0 o
C
20
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
48
0.0
4
-50
.24
0.3
9
-
1.0
03
26
0.9
80
-
#6
 gy
ps
um
 20
 o C
20
L1
10
2-i
 P
ica
rro
20
.6
GH
W
-
-
-4.
23
0.0
3
-68
.64
0.4
1
#7
 w
ate
r 2
0 o
C
20
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
48
0.0
4
-50
.38
0.2
6
-
1.0
03
47
0.9
81
-
#7
 gy
ps
um
 20
 o C
20
L1
10
2-i
 P
ica
rro
20
.2
GH
W
-
-
-4.
04
0.0
3
-68
.82
0.3
5
#7
 w
ate
r 2
0 o
C
20
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
51
0.0
2
-51
.23
0.4
3
-
1.0
03
42
0.9
80
-
#7
 gy
ps
um
 20
 o C
20
L1
10
2-i
 P
ica
rro
20
.3
GH
W
-
-
-4.
08
0.0
5
-68
.91
0.4
2
#2
4 w
ate
r 2
5 o
C
25
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
50
0.0
5
-49
.91
0.2
3
-
1.0
03
32
0.9
80
-
#2
4 g
yp
su
m 
25
 o C
25
L1
10
2-i
 P
ica
rro
20
.0
GH
W
-
-
-4.
21
0.0
2
-68
.72
0.2
9
#2
4 w
ate
r 2
5 o
C
25
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
54
0.0
6
-50
.52
0.4
9
-
1.0
03
54
0.9
81
-
#2
4 g
yp
su
m 
25
 o C
25
L1
10
2-i
 P
ica
rro
20
.1
GH
W
-
-
-3.
99
0.0
7
-68
.19
0.0
9
#2
5 w
ate
r 2
5 o
C
25
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
58
0.0
6
-49
.98
0.4
0
-
1.0
03
44
0.9
81
-
#2
5 g
yp
su
m 
25
 o C
25
L1
10
2-i
 P
ica
rro
20
.2
GH
W
-
-
-4.
17
0.0
4
-67
.74
0.2
9
#2
5 w
ate
r 2
5 o
C
25
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
59
0.0
4
-49
.90
0.4
3
-
1.0
03
70
0.9
82
-
#2
5 g
yp
su
m 
25
 o C
25
L1
10
2-i
 P
ica
rro
20
.8
GH
W
-
-
-3.
92
0.0
5
-67
.17
0.4
1
#2
5 w
ate
r 2
5 o
C
25
L2
14
0-i
 P
ica
rro
wa
ter
-3.
93
0.0
3
-7.
50
0.0
3
-49
.13
0.1
5
1.0
01
81
1.0
03
41
0.9
81
0.5
29
8
#2
5 g
yp
su
m 
25
 o C
25
L2
14
0-i
 P
ica
rro
20
.7
GH
W
-2.
13
0.0
4
-4.
11
0.0
5
-67
.53
0.3
0
#2
3 w
ate
r 3
0 o
C
30
L1
10
2-i
 P
ica
rro
wa
ter
-3.
91
0.0
2
-7.
48
0.0
2
-49
.18
0.1
7
1.0
01
74
1.0
03
30
0.9
81
0.5
27
9
#2
3 g
yp
su
m 
30
 o C
30
L1
10
2-i
 P
ica
rro
20
.7
GH
W
-2.
17
0.0
3
-4.
21
0.0
5
-67
.07
0.1
5
#2
8 w
ate
r 3
0 o
C
30
L2
14
0-i
 P
ica
rro
wa
ter
-3.
90
0.0
2
-7.
49
0.0
3
-49
.29
0.2
2
1.0
01
74
1.0
03
28
0.9
81
0.5
29
9
#2
8 g
yp
su
m 
30
 o C
30
L2
14
0-i
 P
ica
rro
20
.1
GH
W
-2.
17
0.0
3
-4.
23
0.0
4
-67
.05
0.2
6
233
Table B.1b. Table B.1a continued. Fractionation factor between gypsum hydration water and the mother
solution in experiments of hydration of anhydrite at different temperatures.
Ex
pe
rim
en
t I
D
Te
mp
er
atu
re
 (o C
)
CR
DS
 an
aly
ze
r
H 2
O 
%
mo
th
er
 w
ate
r/G
HW
δ1
7 O
1σ
δ1
8 O
1σ
δ D
1σ
α
17
O
α
18
O
α
D
θ
#1
2 w
ate
r 3
5 o
C
35
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
45
0.0
5
-50
.38
0.2
4
-
1.0
03
43
0.9
81
-
#1
2 g
yp
su
m 
35
 o C
35
L1
10
2-i
 P
ica
rro
19
.8
GH
W
-
-
-4.
04
0.0
2
-68
.42
0.0
9
#1
2 w
ate
r 3
5 o
C
35
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
45
0.0
1
-50
.44
0.0
9
-
1.0
03
43
0.9
81
-
#1
2 g
yp
su
m 
35
 o C
35
L1
10
2-i
 P
ica
rro
20
.0
GH
W
-
-
-4.
04
0.0
2
-68
.56
0.0
9
#1
3 w
ate
r 3
5 o
C
35
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
44
0.0
2
-50
.36
0.0
6
-
1.0
03
42
0.9
81
-
#1
3 g
yp
su
m 
35
 o C
35
L1
10
2-i
 P
ica
rro
19
.9
GH
W
-
-
-4.
05
0.0
2
-68
.09
0.1
1
#1
3 w
ate
r 3
5 o
C
35
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
47
0.0
2
-50
.53
0.0
9
-
1.0
03
41
0.9
82
-
#1
3 g
yp
su
m 
35
 o C
35
L1
10
2-i
 P
ica
rro
20
.0
GH
W
-
-
-4.
08
0.0
2
-67
.99
0.1
1
#1
8 w
ate
r 4
0 o
C
40
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
47
0.0
2
-50
.44
0.0
6
-
1.0
03
53
0.9
82
-
#1
8 g
yp
su
m 
40
 o C
40
L1
10
2-i
 P
ica
rro
21
.1
GH
W
-
-
-3.
96
0.0
4
-67
.96
0.0
9
#1
8 w
ate
r 4
0 o
C
40
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
47
0.0
2
-50
.58
0.0
7
-
1.0
03
38
0.9
81
-
#1
8 g
yp
su
m 
40
 o C
40
L1
10
2-i
 P
ica
rro
20
.8
GH
W
-
-
-4.
11
0.0
2
-68
.42
0.0
8
#1
9 w
ate
r 4
0 o
C
40
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
46
0.0
2
-50
.83
0.1
0
-
1.0
03
49
0.9
83
-
#1
9 g
yp
su
m 
40
 o C
40
L1
10
2-i
 P
ica
rro
20
.1
GH
W
-
-
-4.
00
0.0
8
-66
.62
0.3
7
#1
9 w
ate
r 4
0 o
C
40
L1
10
2-i
 P
ica
rro
wa
ter
-
-
-7.
36
0.0
5
-50
.30
0.0
5
-
1.0
03
43
0.9
82
-
#1
9 g
yp
su
m 
40
 o C
40
L1
10
2-i
 P
ica
rro
21
.0
GH
W
-
-
-3.
96
0.0
3
-67
.05
0.1
4
#1
9 w
ate
r 4
0 o
C
40
L2
14
0-i
 P
ica
rro
wa
ter
-3.
91
0.0
2
-7.
48
0.0
2
-49
.05
0.0
6
1.0
01
69
1.0
03
18
0.9
82
0.5
31
4
#1
9 g
yp
su
m 
40
 o C
40
L2
14
0-i
 P
ica
rro
20
.9
GH
W
-2.
23
0.0
4
-4.
32
0.0
5
-65
.85
0.3
5
AV
G 
θ
SD
 θ
234 Tables for Section 2.4
Table B.2a. Fractionation factor between gypsum hydration water and the mother solution in mixing
experiments (CaCl2 and Na2SO4) at different temperatures and initial Ca2+ and SO42−
concentrations. Experiment were performed in an oven with no agitation and in a shaking
bath.
Ex
pe
rim
en
t I
D
In
iti
al 
Ca
2+
 an
d S
O 4
2- 
co
nc
en
t
(m
ol/
L)
Te
mp
er
at
ur
e 
(o C
)
Dr
ift
ing
/Sh
ak
ing
CR
DS
 an
aly
ze
r
mo
th
er
 
wa
ter
/
GH
W
%
 H
2O
δ1
7 O
1σ
δ1
8 O
1σ
δ D
1σ
α
18
O
α
D
1A
 5 
DR
F W
AT
0.2
5
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
35
0.0
3
-50
.25
0.1
8
1.0
03
6
0.9
79
1A
 5 
DR
F G
YP
0.2
5
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.7
-3.
80
0.0
4
-70
.39
0.3
4
1B
 5 
DR
F W
AT
0.2
5
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
36
0.0
4
-50
.24
0.1
7
1.0
03
4
0.9
78
1B
 5 
DR
F G
YP
0.2
5
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
21
-4.
01
0.0
3
-71
.28
0.2
9
2A
 5 
DR
F W
AT
0.1
66
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
34
0.0
3
-49
.97
0.1
9
1.0
03
6
0.9
78
2A
 5 
DR
F G
YP
0.1
66
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.6
-3.
75
0.0
6
-71
.32
0.2
8
2B
 5 
DR
F W
AT
0.1
66
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
31
0.0
4
-50
.07
0.2
5
1.0
03
3
0.9
77
2B
 5 
DR
F G
YP
0.1
66
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.8
-4.
01
0.0
2
-71
.94
0.3
1
3A
 5 
DR
F W
AT
0.0
65
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
32
0.0
3
-49
.91
0.2
1
1.0
03
7
0.9
76
3A
 5 
DR
F G
YP
0.0
65
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.5
-3.
66
0.0
5
-72
.81
0.2
9
3B
 5 
DR
F W
AT
0.0
65
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
24
0.0
2
-49
.94
0.2
8
1.0
03
3
0.9
76
AV
G
1.0
03
5
0.9
77
3B
 5 
DR
F G
YP
0.0
65
5
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
19
.8
-3.
93
0.0
2
-72
.74
0.1
4
1σ
0.0
00
2
0.0
01
1
1A
 20
 D
RF
 W
AT
0.2
5
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
22
0.0
3
-49
.19
0.1
8
1.0
03
9
0.9
84
1A
 20
 D
RF
 G
YP
0.2
5
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.4
-3.
37
0.0
4
-64
.78
0.3
4
1B
 20
 D
RF
 W
AT
0.2
5
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
31
0.0
4
-48
.83
0.1
7
1.0
03
8
0.9
83
1B
 20
 D
RF
 G
YP
0.2
5
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.1
-3.
58
0.0
3
-65
.19
0.2
9
2B
 20
 D
RF
 W
AT
0.1
66
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
26
0.0
4
-49
.36
0.2
5
1.0
03
6
0.9
80
2B
 20
 D
RF
 G
YP
0.1
66
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.5
-3.
70
0.0
2
-67
.94
0.3
1
3A
 20
 D
RF
 W
AT
0.1
25
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
31
0.0
3
-49
.17
0.2
1
1.0
03
6
0.9
79
3A
 20
 D
RF
 G
YP
0.1
25
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.8
-3.
72
0.0
5
-68
.97
0.2
9
3B
 20
 D
RF
 W
AT
0.1
25
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
24
0.0
2
-48
.83
0.2
8
1.0
03
6
0.9
79
AV
G
1.0
03
7
0.9
81
3B
 20
 D
RF
 G
YP
0.1
25
20
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.7
-3.
70
0.0
2
-68
.73
0.1
4
1σ
0.0
00
1
0.0
02
1
1A
 25
 SH
 W
AT
0.2
5
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
40
0.0
5
-50
.52
0.2
3
1.0
03
6
0.9
82
1A
 25
 SH
 G
YP
0.2
5
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.4
-3.
80
0.0
5
-67
.49
0.2
4
1B
 25
 SH
 W
AT
0.2
5
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
50
0.0
4
-50
.76
0.3
6
1.0
03
8
0.9
82
1B
 25
 SH
 G
YP
0.2
5
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.6
-3.
74
0.0
4
-67
.86
0.3
3
2A
 25
 SH
 W
AT
0.1
66
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
51
0.0
4
-50
.95
0.2
2
1.0
03
4
0.9
81
2A
 25
 SH
 G
YP
0.1
66
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.5
-4.
16
0.0
6
-68
.85
0.2
6
2B
 25
 SH
 W
AT
0.1
66
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
51
0.0
6
-50
.50
0.2
8
1.0
03
4
0.9
81
 2B
 25
 SH
 G
YP
0.1
66
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.7
-4.
11
0.0
3
-68
.90
0.4
2
 3A
 25
 SH
 W
AT
0.1
25
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
51
0.0
5
-50
.86
0.2
3
1.0
03
6
0.9
80
 3A
 25
 SH
 G
YP
0.1
25
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.3
-3.
93
0.0
8
-69
.48
0.4
5
 3B
 25
 SH
 W
AT
0.1
25
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
45
0.0
4
-50
.30
0.2
8
1.0
03
2
0.9
78
AV
G
1.0
03
5
0.9
81
 3B
 25
 SH
 G
YP
0.1
25
25
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.8
-4.
25
0.0
7
-70
.88
0.5
6
1σ
0.0
00
2
0.0
02
1A
 25
 D
RF
 W
AT
0.2
5
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
63
0.0
9
-51
.00
0.4
2
1.0
03
6
0.9
83
1A
 25
 D
RF
 G
YP
0.2
5
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.3
-4.
03
0.0
8
-67
.42
0.4
5
1B
 25
 D
RF
 W
AT
0.2
5
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
54
0.0
5
-50
.67
0.1
6
1.0
03
4
0.9
82
1B
 25
 D
RF
 G
YP
0.2
5
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.5
-4.
14
0.0
7
-67
.48
0.5
8
235
Table B.2b. Table B.2a continued. Fractionation factor between gypsum hydration water and the mother
solution in mixing experiments (CaCl2 and Na2SO4) at different temperatures and initial Ca2+
and SO42− concentrations. Experiment were performed in an oven with no agitation and in a
shaking bath.
Ex
pe
rim
en
t I
D
In
iti
al 
Ca
2+
 an
d S
O 4
2- 
co
nc
en
t
(m
ol/
L)
Te
mp
er
at
ur
e 
(o C
)
Dr
ift
ing
/Sh
ak
ing
CR
DS
 an
aly
ze
r
mo
th
er
 
wa
ter
/
GH
W
%
 H
2O
δ1
7 O
1σ
δ1
8 O
1σ
δ D
1σ
α
18
O
α
D
2A
 25
 D
RF
 W
AT
0.1
66
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
62
0.0
4
-50
.67
0.4
1
1.0
03
7
0.9
80
2A
 25
 D
RF
 G
YP
0.1
66
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.4
-3.
98
0.0
6
-69
.37
0.3
3
2B
 25
 D
RF
 W
AT
0.1
66
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
42
0.0
5
-50
.44
0.2
4
1.0
03
5
0.9
81
2B
 25
 D
RF
 G
YP
0.1
66
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.2
-3.
98
0.0
5
-68
.38
0.3
2
3A
 25
 D
RF
 W
AT
0.1
25
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
47
0.0
4
-50
.39
0.2
5
1.0
03
6
0.9
80
3A
 25
 D
RF
 G
YP
0.1
25
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
19
.8
-3.
92
0.0
6
-69
.80
0.3
9
3B
 25
 D
RF
 W
AT
0.1
25
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
45
0.0
2
-50
.61
0.3
0
1.0
03
5
0.9
80
AV
G
1.0
03
5
0.9
81
3B
 25
 D
RF
 G
YP
0.1
25
25
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
19
.8
-3.
96
0.0
4
-69
.50
0.2
0
1σ
0.0
00
1
0.0
01
 1A
 45
 SH
 W
AT
0.2
5
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
27
0.0
2
-49
.81
0.2
1
1.0
03
2
0.9
84
 1A
 45
 SH
 G
YP
0.2
5
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.6
-4.
07
0.0
4
-64
.71
0.3
8
 1B
 45
 SH
 W
AT
0.2
5
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
26
0.0
4
-49
.88
0.3
7
1.0
03
3
0.9
84
 1B
 45
 SH
 G
YP
0.2
5
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.6
-4.
02
0.0
3
-65
.01
0.2
9
 2A
 45
 SH
 W
AT
0.1
66
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
18
0.0
5
-49
.43
0.2
4
1.0
02
9
0.9
83
 2A
 45
 SH
 G
YP
0.1
66
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.2
-4.
30
0.0
9
-65
.83
0.2
1
 2B
 45
 SH
 W
AT
0.1
66
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
20
0.0
4
-49
.72
0.2
0
1.0
02
9
0.9
83
 2B
 45
 SH
 G
YP
0.1
66
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
19
.9
-4.
30
0.0
4
-65
.76
0.3
9
 3A
 45
 SH
 W
AT
0.1
25
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
19
0.0
2
-49
.81
0.3
6
1.0
03
2
0.9
82
AV
G
1.0
03
1
0.9
83
 3A
 45
 SH
 G
YP
0.1
25
45
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.1
-4.
06
0.0
3
-67
.00
0.3
8
1σ
0.0
00
2
0.0
01
0
 1A
 55
 SH
 W
AT
0.2
5
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
53
0.0
2
-50
.72
0.2
2
1.0
03
5
0.9
84
 1A
 55
 SH
 G
YP
0.2
5
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.3
-4.
09
0.0
5
-65
.90
0.4
3
 1B
 55
 SH
 W
AT
0.2
5
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
63
0.0
5
-50
.98
0.2
6
1.0
03
5
0.9
84
 1B
 55
 SH
 G
YP
0.2
5
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.7
-4.
13
0.0
6
-66
.45
0.2
4
 2A
 55
 SH
 W
AT
0.1
66
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
55
0.0
2
-50
.22
0.3
4
1.0
03
2
0.9
83
 2A
 55
 SH
 G
YP
0.1
66
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.5
-4.
34
0.0
3
-66
.17
0.3
1
 2B
 55
 SH
 W
AT
0.1
66
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
54
0.0
4
-50
.64
0.2
1
1.0
03
5
0.9
84
 2B
 55
 SH
 G
YP
0.1
66
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.1
-4.
11
0.0
3
-65
.63
0.4
4
 3A
 55
 SH
 W
AT
0.1
25
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
49
0.0
4
-50
.27
0.3
9
1.0
03
1
0.9
83
 3A
 55
 SH
 G
YP
0.1
25
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.4
-4.
36
0.0
5
-66
.20
0.2
5
 3B
 55
 SH
 W
AT
0.1
25
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
59
0.0
4
-50
.30
0.1
5
1.0
03
2
0.9
83
AV
G
1.0
03
3
0.9
83
 3B
 55
 SH
 G
YP
0.1
25
55
Sh
ak
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.3
-4.
39
0.0
4
-66
.84
0.2
6
1σ
0.0
00
2
0.0
01
1A
 55
 D
RF
 W
AT
0.2
5
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
44
0.0
6
-50
.64
0.2
8
1.0
03
4
0.9
84
1A
 55
 D
RF
 G
YP
0.2
5
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.3
-4.
03
0.0
5
-65
.85
0.3
2
1B
 55
 D
RF
 W
AT
0.2
5
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
49
0.0
8
-49
.65
0.6
2
1.0
03
3
0.9
85
1B
 55
 D
RF
 G
YP
0.2
5
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.2
-4.
21
0.0
5
-63
.67
0.3
7
2A
 55
 D
RF
 W
AT
0.1
66
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
41
0.0
5
-50
.70
0.3
3
1.0
03
2
0.9
84
2A
 55
 D
RF
 G
YP
0.1
66
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.1
-4.
22
0.0
5
-66
.05
0.1
9
2B
 55
 D
RF
 W
AT
0.1
66
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
43
0.0
8
-50
.18
0.4
7
1.0
03
2
0.9
85
2B
 55
 D
RF
 G
YP
0.1
66
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.4
-4.
23
0.0
8
-64
.64
0.4
2
3A
 55
 D
RF
 W
AT
0.1
25
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
44
0.0
3
-49
.85
0.2
0
1.0
02
7
0.9
82
3A
 55
 D
RF
 G
YP
0.1
25
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.6
-4.
79
0.0
4
-67
.26
0.1
8
3B
 55
 D
RF
 W
AT
0.1
25
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
wa
ter
-7.
44
0.0
6
-50
.37
0.4
9
1.0
03
5
0.9
84
AV
G
1.0
03
2
0.9
84
3B
 55
 D
RF
 G
YP
0.1
25
55
Dr
ift
ing
L1
10
2-i
 P
ica
rro
GH
W
20
.6
-3.
95
0.0
3
-65
.16
0.3
4
1σ
0.0
00
3
0.0
01
236 Tables for Section 2.4
Table B.3. Fractionation factor between gypsum hydration water and the mother solution in experiments of
hydration of anhydrite adding different concentrations of NaCl.
Ex
pe
rim
en
t I
D
Na
Cl
 (g
/L
)
α
H 2
O
M
oth
er
 w
ate
r/G
HW
H 2
O 
%
CR
DS
 an
aly
ze
r
δ1
8 O
1σ
δ D
1σ
α
18
O
α
D
α
18
Oa
vg
1σ
α
Da
vg
1σ
0
1
da
ta 
in 
tab
le 
SI
 1
L1
10
2-I
 an
d L
21
40
-I 
Pic
arr
o
1.0
03
4
0.0
00
10
0.9
80
6
0.0
01
0
Na
Cl
1
30
0.9
8
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
32
0.0
2
-49
.01
0.1
4
Na
Cl
1
30
GH
W
20
.8
L1
10
2-I
 Pi
ca
rro
-4.
17
0.0
3
-67
.56
0.1
9
1.0
03
2
0.9
80
5
Na
Cl
2
30
0.9
8
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
34
0.0
4
-49
.38
0.2
4
Na
Cl
2
30
GH
W
21
.0
L1
10
2-I
 Pi
ca
rro
-4.
07
0.0
4
-67
.54
0.3
4
1.0
03
3
0.9
80
9
1.0
03
2
0.0
00
08
0.9
80
7
0.0
00
3
Na
Cl
3
80
0.9
42
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
28
0.0
4
-49
.84
0.3
2
Na
Cl
3
80
GH
W
20
.7
L1
10
2-I
 Pi
ca
rro
-4.
10
0.0
2
-66
.80
0.1
9
1.0
03
2
0.9
82
1
Na
Cl
5
80
0.9
42
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
33
0.0
2
-49
.40
0.1
2
Na
Cl
5
80
GH
W
21
.0
L1
10
2-I
 Pi
ca
rro
-3.
80
0.0
2
-65
.88
0.2
2
1.0
03
6
0.9
82
7
1.0
03
4
0.0
00
25
0.9
82
4
0.0
00
4
Na
Cl
6
15
0
0.8
8
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
12
0.0
5
-48
.83
0.1
5
Na
Cl
6
15
0
GH
W
20
.9
L1
10
2-I
 Pi
ca
rro
-3.
78
0.0
2
-63
.55
0.2
6
1.0
03
4
0.9
84
5
Na
Cl
7
15
0
0.8
8
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
26
0.0
4
-49
.29
0.2
9
Na
Cl
7
15
0
GH
W
20
.2
L1
10
2-I
 Pi
ca
rro
-3.
99
0.0
3
-63
.85
0.3
7
1.0
03
3
0.9
84
7
Na
Cl
8
15
0
0.8
8
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
24
0.0
6
-49
.66
0.1
6
Na
Cl
8
15
0
GH
W
20
.4
L1
10
2-I
 Pi
ca
rro
-4.
09
0.0
6
-64
.26
0.2
8
1.0
03
2
0.9
84
6
1.0
03
3
0.0
00
09
0.9
84
6
0.0
00
1
Na
Cl
9
20
0
0.8
29
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
58
0.0
3
-51
.11
0.2
5
Na
Cl
9
20
0
GH
W
19
.8
L1
10
2-I
 Pi
ca
rro
-3.
64
0.0
7
-61
.66
0.3
1
1.0
04
0
0.9
88
9
Na
Cl
10
20
0
0.8
29
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
54
0.0
6
-51
.24
0.6
7
Na
Cl
10
20
0
GH
W
19
.6
L1
10
2-I
 Pi
ca
rro
-3.
89
0.0
4
-62
.38
0.2
0
1.0
03
7
0.9
88
3
1.0
03
8
0.0
00
21
0.9
88
6
0.0
00
4
Na
Cl
11
30
0
0.6
99
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
21
0.0
4
-50
.15
0.2
8
Na
Cl
11
30
0
GH
W
19
.8
L1
10
2-I
 Pi
ca
rro
-2.
70
0.0
8
-59
.17
0.1
8
1.0
04
5
0.9
90
5
Na
Cl
12
30
0
0.6
99
wa
ter
L1
10
2-I
 Pi
ca
rro
-7.
35
0.1
1
-49
.04
0.4
1
Na
Cl
12
30
0
GH
W
19
.7
L1
10
2-I
 Pi
ca
rro
-2.
46
0.0
6
-60
.35
0.1
6
1.0
04
9
0.9
88
1
1.0
04
7
0.0
00
27
0.9
89
3
0.0
01
7
237
Table B.4. XRD analysis of solid recovered in experiments of mixing Ca2+ and SO42− solutions (mixing
experiments), hydration of anhydrous CaSO4 in low salinity solutions (hydration of anhydrite)
and hydration of anhydrous in NaCl brines.
Mixing Experiments
Experiment ID Initial SIgyp Temperature (oC)   Gypsum (%) Bassanite (%) Anhydrite (%)
EXP #1A-5oC 1.5 5 99 0.6 0.1
EXP #2A-5oC 1.26 5 99.4 0.6 0
EXP #3A-5oC 0.75 5 99 0.7 0.3
EXP #1A-25oC 1.42 25 99 0.8 0.2
EXP #2A-25oC 1.18 25 98.9 0.7 0.3
EXP #3A-25oC 0.68 25 99.4 0.6 0
EXP #1A-55oC 1.39 55 99.1 0.9 0.1
EXP #2A-55oC 1.15 55 99.1 0.7 0.3
EXP #3A-55oC 0.66 55 99.2 0.8 0
Hydration of Anhydrite
Experiment ID Temperature (oC)   Gypsum (%) Bassanite (%) Anhydrite (%)
#32-3oC 3 99 0.3 0.7
#24-25oC 25 99.2 0.6 0.6
#13-35oC 35 98.1 0.5 0.3
Hydration of Anhydrite in brine
Experiment ID NaCl (g/l) Temperature (oC)   Gypsum (%) Bassanite (%) Anhydrite (%)
NaCl 30 g/L Exp2 30 20 99.2 0.2 0.8
NaCl 150 g/L Exp7 150 20 98.6 1.2 0.2
NaCl 200 g/L A Exp9 200 20 97.7 1.5 0.8
NaCl 300 g/L A Exp12 300 20 92.9 2.5 4.6
Table B.5. A. Stable isotopes in hydration water of gypsum precipitated from brines in a natural salt factory
in Cabo de Gata (Almeria, SE Spain) (Evans et al., 2015)1. Mother water is calculated using
published fractionation factors of Gonfiantini and Fontes (1963)2 and Sofer (1978)3 and those
proposed in this study, at 20˚C and salinity of 150 gl-1 of NaCl. B. Isotopic compositions of
brines from the same salt factory.
A. Gypsum
Sample δ18O 1σ δD 1σ δ18O 1σ d-excess 1σ δ18O 1σ d-excess 1σ
#1SALINA 10.56 0.17 12.94 0.35 6.53 33.62 -18.6 0.3 7.23 28.37 -29.5 0.3
#2SALINA 10.25 0.13 9.57 0.45 6.23 30.17 -19.7 0.2 6.93 24.94 -30.5 0.2
#3SALINA 11.37 0.07 15.94 0.30 7.34 36.67 -22.1 0.1 8.05 31.41 -33.0 0.1
#4SALINA 11.60 0.10 18.62 0.24 7.57 39.41 -21.1 0.3 8.27 34.13 -32.0 0.3
#5SALINA 10.86 0.04 16.69 0.31 6.83 37.44 -17.2 0.2 7.53 32.17 -28.1 0.2
B. Brines
Sample δ18O 1σ d-excess 1σ Residue (mg/L)
Brine 1 2.35 0.04 10.61 0.44 51.2
Brine 2 1.99 0.12 6.06 0.44 109.3
Brine 3 1.17 0.07 7.98 0.42 40.6
Brine 4 6.83 0.07 31.59 0.54 88.6
Brine 5 8.45 0.02 36.50 0.29 162.0
Brine 6 8.60 0.09 33.46 0.45 227.7
Brine 7 9.81 0.07 34.17 0.37 306.7
Brine 8 2.46 0.06 13.86 0.47 49.6
Brine 9 3.51 0.06 7.62 0.29 302.6
Brine 10 4.15 0.09 12.81 0.31 265.7
Mother water (1.004, 0.980)(2,3)Gypsum hydration water(1) Mother water (1.0033, 0.985)
Mother water

Appendix C
Tables for Chapter 3
240 Tables for Chapter 3
Table C.1. Isotopic composition of measured gypsum hydration water (GHW) and calculated lake mother
water (MW) from samples recovered from core CH1 7-III-04, Lake Chichancanab.
Sample name Depth(cm)
Age
(C.E.)
δ17O
 (GHW)
δ18O
 (GHW)
δD 
(GHW)
δ17O 
(MW)
δ18O
 (MW)
δD
 (MW)
17O-excess 
(MW)
d-excess
(MW)
A
CHI 07-III-04  128.0-128.5 cm 128 1092 4.08 7.81 -2.66 2.24 4.33 17.0 -46 -17.7
CHI 07-III-04  129.0-129.5 cm 129 1085 3.83 7.33 -4.58 1.99 3.85 15.0 -38 -15.8
CHI 07-III-04  130.5-131.0 cm 130.5 1073 3.70 7.08 -5.86 1.87 3.60 13.7 -33 -15.1
CHI 07-III-04  131.0-131.5 cm 131 1069 3.91 7.46 -4.57 2.07 3.99 15.0 -31 -16.9
CHI 07-III-04  132.0-132.5 cm 132 1059 4.21 8.04 -1.35 2.37 4.57 18.3 -35 -18.2
CHI 07-III-04  133.0-133.5 cm 133 1050 3.73 7.12 -3.90 1.90 3.65 15.7 -30 -13.5
CHI 07-III-04  133.5-134.0 cm 133.5 1046 3.84 7.34 -4.41 2.01 3.87 15.2 -34 -15.8
CHI 07-III-04  134.0-134.5 cm 134 1041 4.11 7.85 -2.99 2.27 4.37 16.6 -31 -18.3
CHI 07-III-04  135.5-136.0 cm 135.5 1026 3.90 7.45 -3.74 2.06 3.98 15.9 -36 -15.9
CHI 07-III-04  136.0-136.5 cm 136.5 1013 4.01 7.65 -2.63 2.17 4.17 17.0 -31 -16.4
CHI 07-III-04  136.5-137.0 cm 137.5 1000 4.04 7.71 -3.20 2.20 4.24 16.4 -36 -17.5
CHI 07-III-04  137.0-137.5 cm 138 993 3.83 7.34 -4.14 2.00 3.87 15.5 -42 -15.5
CHI 07-III-04  139.5-140.0 cm 139.5 973 3.77 7.21 -4.19 1.93 3.73 15.4 -40 -14.5
CHI 07-III-04  140.0-140.5 cm 140 967 3.89 7.45 -3.27 2.05 3.98 16.3 -47 -15.5
CHI 07-III-04  140.5-141.5 cm 140.5 960 3.83 7.31 -4.04 1.99 3.84 15.6 -32 -15.1
CHI 07-III-04  142.0-142.5 cm 142 939 4.36 8.33 -1.27 2.52 4.85 18.4 -36 -20.4
B
CHI 07-III-04  148.0-148.5 cm 148 849 3.95 7.54 -3.18 2.11 4.07 16.4 -33 -16.1
CHI 07-III-04  149.0-149.5 cm 149 833 4.00 7.65 -3.72 2.16 4.18 15.9 -43 -17.5
CHI 07-III-04  150.0-151.0 cm 150 817 4.21 8.04 -1.92 2.37 4.56 17.7 -36 -18.8
CHI 07-III-04  151.0-151.5 cm 151 801 4.13 7.89 -2.62 2.29 4.42 17.0 -36 -18.3
CHI 07-III-04  151.5-152.0 cm 151.5 793 4.31 8.24 -1.40 2.47 4.77 18.2 -43 -19.9
CHI 07-III-04  153.0-153.5 cm 153 768 3.75 7.17 -3.70 1.91 3.69 15.9 -33 -13.6
CHI 07-III-04  153.5-154.0 cm 153.5 760 4.05 7.74 -2.34 2.22 4.26 17.3 -32 -16.8
CHI 07-III-04  154.0-154.5 cm 154 752 4.28 8.18 -0.88 2.44 4.70 18.8 -35 -18.8
CHI 07-III-04  154.5-155.0 cm 154.5 744 3.98 7.60 -3.69 2.14 4.12 15.9 -33 -17.1
CHI 07-III-04  154.5-155.0 cm 154.5 744 4.11 7.84 -3.01 2.27 4.37 16.6 -32 -18.3
CHI 07-III-04  155.0-155.5 cm 155 736 4.12 7.89 -2.76 2.28 4.42 16.9 -45 -18.5
CHI 07-III-04  155.5-156.0 cm 155.5 728 3.90 7.48 -3.44 2.07 4.01 16.2 -45 -15.9
CHI 07-III-04  156.0-156.5 cm 156 720 4.08 7.81 -3.47 2.24 4.34 16.1 -48 -18.6
C
CHI 07-III-04  166.0-166.5 cm 166 611 3.84 7.32 -4.48 2.00 3.85 15.1 -26 -15.7
CHI 07-III-04  167.0-167.5 cm 167 605 4.15 7.93 -1.91 2.31 4.46 17.7 -43 -17.9
CHI 07-III-04  167.5-168.0 cm 167.5 602 3.86 7.34 -4.59 2.02 3.86 15.0 -19 -15.9
Table C.2. Paleo-lake temperature calculations. The equation of Grossman and Ku (1986) was used to
calculate the temperature at which the aragonitic shells of Pyrgophorus coronatus formed.
Sample δ18O
(Carbonate)
δ18O
 (GHW)
Temperature
(°C)
CHI 07-III-04  133.0-133.5 cm 2.96 3.65 25.0
CHI 07-III-04  133.5-134.0 cm 3.46 3.87 23.7
CHI 07-III-04  139.5-140.0 cm 3.40 4.07 24.9
CHI 07-III-04  140.5-141.0 cm 3.23 3.99 25.3
CHI 07-III-04  153.0-153.5 cm 3.06 4.07 26.4
CHI 07-III-04  154.0-154.5 cm 3.26 4.70 28.4
CHI 07-III-04  166.0-166.5 cm 2.59 3.85 27.6
Mean 25.9
1σ 1.7
241
Table C.3a. Output table of Bayesian age-depth analysis. Mean ages display single ‘best’ model based
on the weighted mean age for each depth. Positive and negative age errors represent 95%
confidence intervals.
Depth 
(cm)
Positive Age Error 
(C.E.)
Negative Age Error
(C.E.)
Mean Age
(C.E.)
Depth 
(cm)
Positive Age Error 
(C.E.)
Negative Age Error
(C.E.)
Mean Age
(C.E.)
0 2025.6 1953.5 2004 60 1752.8 1386.9 1574.3
1 2021.3 1948.7 2000.8 61 1742.5 1382.3 1567
2 2019.6 1938.9 1993.6 62 1735.3 1377.7 1559.7
3 2018.6 1920.1 1986.5 63 1730 1370.6 1552.4
4 2018.5 1897.6 1979.7 64 1722.8 1363.1 1544.9
5 2018.7 1874.6 1972.9 65 1716.8 1353.5 1537.5
6 2014.2 1867.4 1965.4 66 1706.5 1349.3 1530.2
7 2011.8 1859.3 1958 67 1698.3 1344.9 1522.8
8 2010 1846.3 1950.5 68 1692.2 1338.1 1515.2
9 2008.4 1831.4 1943 69 1687.9 1331.8 1507.7
10 2006.6 1814 1935.6 70 1683 1323.9 1500.3
11 2001.4 1805.9 1928.4 71 1671.5 1319.3 1493.2
12 1997.7 1796.8 1921 72 1664.5 1313.4 1486.2
13 1995.4 1787.2 1913.7 73 1658 1305.6 1479.2
14 1993.2 1773.1 1906.5 74 1653.6 1297.6 1472
15 1991.6 1758.2 1899.2 75 1648.5 1290.1 1465
16 1985.5 1751.1 1891.9 76 1638.8 1285.7 1457.7
17 1981.2 1744 1884.4 77 1631.8 1280.5 1450.5
18 1978 1731.8 1877.2 78 1625.4 1273.7 1443.2
19 1975.3 1719 1869.9 79 1620.4 1267.4 1435.9
20 1973.4 1705 1862.8 80 1614 1258.6 1428.6
21 1967.1 1698.7 1855.6 81 1604.9 1253.8 1421.8
22 1961.7 1692.4 1848.4 82 1595.6 1248.3 1415.1
23 1957.2 1684.3 1841.2 83 1588.2 1241.6 1408.2
24 1953.8 1673.3 1833.9 84 1581.9 1234.6 1401.5
25 1950.8 1661.2 1826.7 85 1576.8 1225.3 1394.6
26 1942.6 1654.2 1819.5 86 1567.5 1221.4 1387.7
27 1937.1 1648.5 1812.4 87 1557.7 1216.6 1380.6
28 1932.8 1640.3 1805.3 88 1549.7 1211.5 1373.7
29 1928.9 1630.3 1798.2 89 1543 1204.8 1366.7
30 1925.5 1619.9 1791.1 90 1537.7 1195.6 1359.8
31 1917.6 1612.2 1783.7 91 1529.9 1191.4 1352.9
32 1911.4 1605 1776.4 92 1522 1186.4 1345.8
33 1907.3 1597.2 1769.1 93 1517.1 1180.5 1338.8
34 1902.1 1588.6 1761.7 94 1510.2 1174.9 1331.8
35 1899.1 1576.3 1754.3 95 1505 1167.6 1324.8
36 1889.9 1569.6 1747.2 96 1494.9 1164 1317.7
37 1884.8 1562.5 1740 97 1483.8 1159.3 1310.6
38 1878.6 1555.9 1732.8 98 1476.3 1153.2 1303.4
39 1874.2 1546.3 1725.6 99 1469.1 1146.4 1296.2
40 1870.9 1535.3 1718.4 100 1464.9 1138.6 1288.9
41 1863.1 1530 1711.1 101 1452.1 1135.1 1281.8
42 1856.5 1524.1 1703.8 102 1440.8 1130.8 1274.5
43 1851 1515.5 1696.5 103 1432.4 1124.8 1267.4
44 1846.6 1506.5 1689.3 104 1426.3 1118.1 1260.1
45 1843 1499.4 1682 105 1419.7 1110.5 1252.9
46 1836 1494.1 1674.9 106 1407.6 1107.3 1245.3
47 1829.3 1488 1667.6 107 1397.7 1103.1 1237.8
48 1824.2 1480.7 1660.4 108 1389.7 1098.2 1230.2
49 1820 1471.6 1653.1 109 1383.5 1092.9 1222.7
50 1817.4 1463.2 1646 110 1376.9 1085.2 1215.1
51 1808.4 1457.4 1638.8 111 1365.3 1082.2 1208.3
52 1801.1 1450.9 1631.7 112 1354.9 1078.7 1201.6
53 1794.7 1443.7 1624.5 113 1345.4 1074.8 1194.9
54 1788.8 1433.9 1617.5 114 1337 1070.4 1188.2
55 1783.3 1424 1610.6 115 1329.3 1064.3 1181.5
56 1774.4 1418.9 1603.3 116 1318.1 1061.3 1174.6
57 1767.7 1411.9 1596 117 1308.8 1058.7 1167.8
58 1761.3 1404.7 1588.7 118 1300.9 1055.5 1160.8
59 1756.1 1395.4 1581.3 119 1293.2 1050.4 1154
242 Tables for Chapter 3
Table C.3b. Table C.2a continued. Output table of Bayesian age-depth analysis. Mean ages display single
‘best’ model based on the weighted mean age for each depth. Positive and negative age errors
represent 95% confidence intervals.
Depth 
(cm)
Positive Age Error 
(C.E.)
Negative Age Error
(C.E.)
Mean Age
(C.E.)
Depth 
(cm)
Positive Age Error 
(C.E.)
Negative Age Error
(C.E.)
Mean Age
(C.E.)
120 1288.3 1044 1147 180 602.3 393.1 508.2
121 1274.1 1042.7 1140.1 181 591.9 386.2 499.2
122 1262.6 1040.7 1133.3 182 584 379.1 490.1
123 1252.7 1038 1126.3 183 577.8 372 481.2
124 1244.4 1034.8 1119.2 184 573.4 359.3 472.3
125 1237.9 1029 1112.2 185 569.9 345.3 463.6
126 1224.3 1027.2 1105.6 186 556.8 339.1 453
127 1211.8 1025.1 1098.6 187 547.3 333.1 442.5
128 1203 1021.4 1091.7 188 541.6 321.9 432
129 1195.7 1016.6 1084.8 189 535.7 307.6 421.3
130 1190.9 1010.3 1078.1 190 531.3 291.9 410.7
131 1175.8 1005.6 1068.7 191 510.2 284.1 397.5
132 1165.4 997.3 1059.4 192 497.5 274.8 384.3
133 1157.9 985 1050.2 193 487.3 260.9 371.1
134 1151.4 967.4 1041 194 480.5 243.4 357.9
135 1146.8 948 1032.2 195 474.6 220.4 344.7
136 1127.1 940.2 1019.3 196 445.6 214.7 331.6
137 1110.5 925.7 1006.2 197 418.9 207.4 318.4
138 1099.2 902.2 992.9 198 400.5 198.4 305.3
139 1092 874.6 979.7 199 387.3 187.3 292.3
140 1087 847.5 966.9 200 377.2 170.3 279.3
141 1056.6 838.5 953.1 201 362.3 165.8 269.7
142 1029.7 828.4 939.1 202 351.9 158.6 260
143 1011.1 814.5 925 203 344 148.5 250.1
144 997.8 795.5 911.1 204 338.9 133.1 240.1
145 989.4 770.4 897.2 205 334.5 116.1 230.2
146 965.8 761 881.1 206 325.9 110.1 222
147 954.1 749.2 865 207 318.2 101.8 213.7
148 946 733.4 848.9 208 312.6 92.2 205.4
149 940.7 710.8 832.9 209 308.2 81.8 197.3
150 936.9 681 817 210 304.2 67.6 189.2
151 915.2 676.2 801 211 292.6 61.5 180.9
152 901.2 670.8 784.7 212 284.1 54.1 172.5
153 891.6 663.9 768.4 213 277.8 44.4 164.3
154 882.9 654.6 752.2 214 273.3 33.5 155.9
155 876.9 636.7 736.1 215 269.3 21.1 147.7
156 835.6 635.2 720.2 216 258.6 13.7 138.5
157 800.2 632.2 704.1 217 250.6 4.4 129.4
158 772.6 628.7 687.7 218 243.8 -8.6 120.3
159 753.8 622.7 671.2 219 239 -23.7 111.1
160 741.2 610 654.8 220 233 -42.4 102.2
161 729.3 605.5 647 221 223.8 -52.3 92.1
162 722.7 597.6 639.3 222 216.4 -67 82.2
163 717.3 586.6 631.9 223 210.5 -82.4 72.5
164 712.6 573.6 624.3 224 206 -101.8 62.7
165 706.5 557.8 616.7 225 202.7 -127.1 52.9
166 695.1 552.1 610.7 226 192.2 -136.3 42.9
167 685.3 543.5 604.5 227 185.8 -147 32.8
168 676.1 531.9 598.5 228 179.4 -160.1 22.8
169 668.6 517.4 592.4 229 174.2 -176 12.7
170 664.3 501 586.5 230 169.9 -196.4 2.6
171 652.7 493.8 579.4 231 160.7 -207.1 -7.1
172 644.9 485.1 572.3 232 153.4 -217 -16.7
173 636.9 475.4 565.4 233 145.9 -228.7 -26.6
174 632.7 462.1 558.6 234 141 -243.2 -36.3
175 629.9 446.8 551.8 235 136.8 -264.2 -45.6
176 620.8 439.9 543.2 236 126.2 -271.6 -55.4
177 614.5 432.8 534.4 237 116.9 -281.5 -65.1
178 610 421.9 525.5 238 108.7 -292 -74.8
179 605.7 408.1 516.8 239 103 -306.6 -84.6
243
Table C.4a. The BACON function ‘Events’ was used to quantify the probability of arid conditions at any
calendar age. Fifty-year window widths were used, with the windows moving at 10-year steps
from the core’s bottom ages to its top.
Cal Year (BP) Age (C.E.) Probability of Drought Cal Year (BP) Age (C.E.) Probability of Drought Cal Year (BP) Age (C.E.) Probability of Drought
105.1 1844.9 0 705.1 1244.9 0.0034 1305.1 644.9 0.49422
115.1 1834.9 0 715.1 1234.9 0.0068 1315.1 634.9 0.529752
125.1 1824.9 0 725.1 1224.9 0.012581 1325.1 624.9 0.586535
135.1 1814.9 0 735.1 1214.9 0.024991 1335.1 614.9 0.654029
145.1 1804.9 0 745.1 1204.9 0.040972 1345.1 604.9 0.681231
155.1 1794.9 0 755.1 1194.9 0.064264 1355.1 594.9 0.650119
165.1 1784.9 0 765.1 1184.9 0.090275 1365.1 584.9 0.556273
175.1 1774.9 0 775.1 1174.9 0.121047 1375.1 574.9 0.426726
185.1 1764.9 0 785.1 1164.9 0.153859 1385.1 564.9 0.314519
195.1 1754.9 0 795.1 1154.9 0.189221 1395.1 554.9 0.212853
205.1 1744.9 0 805.1 1144.9 0.223563 1405.1 544.9 0.134988
215.1 1734.9 0 815.1 1134.9 0.252975 1415.1 534.9 0.080755
225.1 1724.9 0 825.1 1124.9 0.289527 1425.1 524.9 0.046073
235.1 1714.9 0 835.1 1114.9 0.33526 1435.1 514.9 0.025842
245.1 1704.9 0 845.1 1104.9 0.394594 1445.1 504.9 0.015471
255.1 1694.9 0 855.1 1094.9 0.465148 1455.1 494.9 0.009011
265.1 1684.9 0 865.1 1084.9 0.551513 1465.1 484.9 0.00561
275.1 1674.9 0 875.1 1074.9 0.67069 1475.1 474.9 0.0034
285.1 1664.9 0 885.1 1064.9 0.794798 1485.1 464.9 0.00289
295.1 1654.9 0 895.1 1054.9 0.890513 1495.1 454.9 0.0017
305.1 1644.9 0 905.1 1044.9 0.949677 1505.1 444.9 0.00119
315.1 1634.9 0 915.1 1034.9 0.976029 1515.1 434.9 0.00068
325.1 1624.9 0 925.1 1024.9 0.985379 1525.1 424.9 0.00085
335.1 1614.9 0 935.1 1014.9 0.985889 1535.1 414.9 0.00051
345.1 1604.9 0 945.1 1004.9 0.984189 1545.1 404.9 0.00051
355.1 1594.9 0 955.1 994.9 0.980279 1555.1 394.9 0.00085
365.1 1584.9 0 965.1 984.9 0.970758 1565.1 384.9 0.00221
375.1 1574.9 0 975.1 974.9 0.958007 1575.1 374.9 0.00476
385.1 1564.9 0 985.1 964.9 0.939646 1585.1 364.9 0.011051
395.1 1554.9 0 995.1 954.9 0.908535 1595.1 354.9 0.019211
405.1 1544.9 0 1005.1 944.9 0.870962 1605.1 344.9 0.033152
415.1 1534.9 0 1015.1 934.9 0.824549 1615.1 334.9 0.048453
425.1 1524.9 0 1025.1 924.9 0.777967 1625.1 324.9 0.067834
435.1 1514.9 0 1035.1 914.9 0.735124 1635.1 314.9 0.097756
445.1 1504.9 0 1045.1 904.9 0.694492 1645.1 304.9 0.138728
455.1 1494.9 0 1055.1 894.9 0.66763 1655.1 294.9 0.198742
465.1 1484.9 0 1065.1 884.9 0.653859 1665.1 284.9 0.266576
475.1 1474.9 0 1075.1 874.9 0.658279 1675.1 274.9 0.319279
485.1 1464.9 0 1085.1 864.9 0.6695 1685.1 264.9 0.356681
495.1 1454.9 0 1095.1 854.9 0.693472 1695.1 254.9 0.374192
505.1 1444.9 0 1105.1 844.9 0.717613 1705.1 244.9 0.374532
515.1 1434.9 0 1115.1 834.9 0.746855 1715.1 234.9 0.359572
525.1 1424.9 0 1125.1 824.9 0.779667 1725.1 224.9 0.345291
535.1 1414.9 0 1135.1 814.9 0.807038 1735.1 214.9 0.33543
545.1 1404.9 0 1145.1 804.9 0.83288 1745.1 204.9 0.32897
555.1 1394.9 0 1155.1 794.9 0.859572 1755.1 194.9 0.316389
565.1 1384.9 0 1165.1 784.9 0.883203 1765.1 184.9 0.287657
575.1 1374.9 0 1175.1 774.9 0.900374 1775.1 174.9 0.246515
585.1 1364.9 0 1185.1 764.9 0.916865 1785.1 164.9 0.212173
595.1 1354.9 0 1195.1 754.9 0.920945 1795.1 154.9 0.17494
605.1 1344.9 0 1205.1 744.9 0.914655 1805.1 144.9 0.135668
615.1 1334.9 0 1215.1 734.9 0.897314 1815.1 134.9 0.109657
625.1 1324.9 0 1225.1 724.9 0.870622 1825.1 124.9 0.088575
635.1 1314.9 0 1235.1 714.9 0.83662 1835.1 114.9 0.070724
645.1 1304.9 0 1245.1 704.9 0.792758 1845.1 104.9 0.054573
655.1 1294.9 0 1255.1 694.9 0.739544 1855.1 94.9 0.046753
665.1 1284.9 0 1265.1 684.9 0.680551 1865.1 84.9 0.042333
675.1 1274.9 0.00017 1275.1 674.9 0.620707 1875.1 74.9 0.040292
685.1 1264.9 0.00017 1285.1 664.9 0.554403 1885.1 64.9 0.040802
695.1 1254.9 0.0017 1295.1 654.9 0.49966 1895.1 54.9 0.041823
244 Tables for Chapter 3
Table C.4b. Table C.4a continued. The BACON function ‘Events’ was used to quantify the probability of
arid conditions at any calendar age. Fifty-year window widths were used, with the windows
moving at 10-year steps from the core’s bottom ages to its top.
Cal Year (BP) Age (C.E.) Probability of Drought Cal BP Age (C.E.) Probability of Drought
1905.1 44.9 0.048113 1905.1 44.9 0.048113
1915.1 34.9 0.058824 1915.1 34.9 0.058824
1925.1 24.9 0.072934 1925.1 24.9 0.072934
1935.1 14.9 0.087385 1935.1 14.9 0.087385
1945.1 4.9 0.102006 1945.1 4.9 0.102006
1955.1 -5.1 0.122407 1955.1 -5.1 0.122407
1965.1 -15.1 0.142809 1965.1 -15.1 0.142809
1975.1 -25.1 0.16627 1975.1 -25.1 0.16627
1985.1 -35.1 0.186331 1985.1 -35.1 0.186331
1995.1 -45.1 0.207242 1995.1 -45.1 0.207242
2005.1 -55.1 0.231554 2005.1 -55.1 0.231554
2015.1 -65.1 0.253655 2015.1 -65.1 0.253655
2025.1 -75.1 0.285957 2025.1 -75.1 0.285957
2035.1 -85.1 0.305338 2035.1 -85.1 0.305338
2045.1 -95.1 0.32812 2045.1 -95.1 0.32812
2055.1 -105.1 0.348011 2055.1 -105.1 0.348011
2065.1 -115.1 0.361102 2065.1 -115.1 0.361102
2075.1 -125.1 0.377083 2075.1 -125.1 0.377083
2085.1 -135.1 0.397144 2085.1 -135.1 0.397144
2095.1 -145.1 0.410405 2095.1 -145.1 0.410405
2105.1 -155.1 0.421285 2105.1 -155.1 0.421285
2115.1 -165.1 0.428936 2115.1 -165.1 0.428936
2125.1 -175.1 0.437946 2125.1 -175.1 0.437946
2135.1 -185.1 0.450697 2135.1 -185.1 0.450697
2145.1 -195.1 0.458348 2145.1 -195.1 0.458348
2155.1 -205.1 0.464298 2155.1 -205.1 0.464298
2165.1 -215.1 0.463618 2165.1 -215.1 0.463618
2175.1 -225.1 0.469228 2175.1 -225.1 0.469228
2185.1 -235.1 0.470928 2185.1 -235.1 0.470928
2195.1 -245.1 0.468038 2195.1 -245.1 0.468038
2205.1 -255.1 0.455457 2205.1 -255.1 0.455457
2215.1 -265.1 0.450867 2215.1 -265.1 0.450867
2225.1 -275.1 0.429106 2225.1 -275.1 0.429106
2235.1 -285.1 0.410745 2235.1 -285.1 0.410745
2245.1 -295.1 0.381843 2245.1 -295.1 0.381843
2255.1 -305.1 0.359062 2255.1 -305.1 0.359062
2265.1 -315.1 0.33322 2265.1 -315.1 0.33322
2275.1 -325.1 0.314349 2275.1 -325.1 0.314349
2285.1 -335.1 0.286297 2285.1 -335.1 0.286297
2295.1 -345.1 0.259436 2295.1 -345.1 0.259436
2305.1 -355.1 0.233594 2305.1 -355.1 0.233594
2315.1 -365.1 0.202992 2315.1 -365.1 0.202992
2325.1 -375.1 0.195682 2325.1 -375.1 0.195682
2335.1 -385.1 0.186161 2335.1 -385.1 0.186161
2345.1 -395.1 0.175281 2345.1 -395.1 0.175281
2355.1 -405.1 0.17103 2355.1 -405.1 0.17103
2365.1 -415.1 0.156239 2365.1 -415.1 0.156239
2375.1 -425.1 0.141789 2375.1 -425.1 0.141789
2385.1 -435.1 0.124277 2385.1 -435.1 0.124277
2395.1 -445.1 0.106426 2395.1 -445.1 0.106426
2405.1 -455.1 0.092146 2405.1 -455.1 0.092146
2415.1 -465.1 0.081605 2415.1 -465.1 0.081605
2425.1 -475.1 0.070044 2425.1 -475.1 0.070044
2435.1 -485.1 0.062394 2435.1 -485.1 0.062394
2445.1 -495.1 0.051853 2445.1 -495.1 0.051853
2455.1 -505.1 0.047603 2455.1 -505.1 0.047603
2465.1 -515.1 0.041313 2465.1 -515.1 0.041313
2475.1 -525.1 0.037062 2475.1 -525.1 0.037062
2485.1 -535.1 0.033152 2485.1 -535.1 0.033152
2495.1 -545.1 0.026692 2495.1 -545.1 0.026692
245
Table C.5a. Stable isotope ratios of lake waters from the Yucatán Peninsula.
Sample - Lakewater δ17O (1σ) δ18O (1σ) δD (1σ) d-excess 17O-excess Latitude (N) Longitude (W)
Lake Chichancanab
Chichancanab 09-XI-04  2 m 1.55 0.04 2.98 0.04 13.34 0.25 -10.5 -20 19° 52.771' 88° 46.026'
Chichancanab 09-XI-04  3 m 1.40 0.01 2.70 0.02 10.39 0.11 -11.2 -25 19° 52.771' 88° 46.026'
Chichancanab 09-XI-04  12 m 1.43 0.02 2.76 0.03 11.18 0.16 -10.9 -29 19° 52.771' 88° 46.026'
Chichancanab 27-II-05 A 1.76 0.03 3.41 0.04 13.08 0.32 -14.2 -41 19° 52.771' 88° 46.026'
Chichancanab 27-II-05 1.90 0.02 3.68 0.03 15.31 0.16 -14.2 -40 19° 52.771' 88° 46.026'
Chichancanab #1 B SURF 1.78 0.03 3.44 0.04 14.67 0.19 -12.8 -32 19° 52.771' 88° 46.026'
Chichancanab DEEP #2 A 1.76 0.02 3.39 0.02 14.59 0.11 -12.5 -28 19° 52.771' 88° 46.026'
Chichancanab DEEP #2B 1.76 0.01 3.40 0.04 14.47 0.14 -12.7 -31 19° 52.771' 88° 46.026'
Chichancanab DEEP #5 B 1.76 0.02 3.38 0.02 14.31 0.15 -12.8 -25 19° 52.771' 88° 46.026'
Chichancanab DEEP 13 #5 1.76 0.02 3.41 0.03 14.30 0.09 -13.0 -36 19° 52.771' 88° 46.026'
Chichancanab 22-V-00 A 1.97 0.03 3.82 0.04 17.21 0.09 -13.4 -42 19° 52.771' 88° 46.026'
Chichancanab 22-V-00 B 1.65 0.03 3.20 0.03 13.37 0.15 -12.7 -36 19° 52.771' 88° 46.026'
Chichancanab 19-VI-96 1.32 0.03 2.56 0.04 10.13 0.17 -10.5 -30 19° 52.771' 88° 46.026'
CH Lake water surface #1A 1.79 0.03 3.45 0.02 14.86 0.12 -12.7 -28 19° 52.771' 88° 46.026'
CH 8-III-04 0m 1.55 0.04 2.98 0.06 11.95 0.29 -11.6 -23 19° 52.771' 88° 46.026'
CH 8-III-04 1m 1.54 0.02 2.95 0.03 12.16 0.19 -11.1 -21 19° 52.771' 88° 46.026'
CH 8-III-04 2m 1.49 0.02 2.86 0.03 11.73 0.10 -10.8 -24 19° 52.771' 88° 46.026'
CH 8-III-04 3m 1.38 0.03 2.67 0.04 10.39 0.16 -11.0 -32 19° 52.771' 88° 46.026'
CH 8-III-04 4m 1.36 0.03 2.62 0.04 10.08 0.17 -10.9 -23 19° 52.771' 88° 46.026'
CH 8-III-04 5m 1.49 0.03 2.87 0.04 11.73 0.17 -11.2 -19 19° 52.771' 88° 46.026'
CH 8-III-04 6m 1.49 0.03 2.86 0.04 11.80 0.23 -11.1 -19 19° 52.771' 88° 46.026'
CH 8-III-04 8m 1.51 0.04 2.91 0.06 12.04 0.35 -11.2 -21 19° 52.771' 88° 46.026'
CH 8-III-04 10m 1.50 0.03 2.89 0.05 11.80 0.18 -11.3 -25 19° 52.771' 88° 46.026'
CH 8-III-04 12m 1.38 0.03 2.67 0.03 10.37 0.23 -11.0 -26 19° 52.771' 88° 46.026'
CH 8-III-04 14m 1.44 0.02 2.77 0.03 11.47 0.21 -10.7 -21 19° 52.771' 88° 46.026'
Peten Itza Lake
PI 13-VIII-02 0 m 1.53 0.03 2.94 0.03 16.05 0.21 -7.5 -23 17° 0.170' 89° 47.956'
PI 13-VIII-02 0 m II 1.54 0.03 2.95 0.04 15.34 0.15 -8.3 -21 17° 0.170' 89° 47.956'
PI 13-VIII-02 10 m 1.54 0.03 2.95 0.03 16.31 0.16 -7.3 -16 17° 0.170' 89° 47.956'
PI 13-VIII-02 10 m II 1.56 0.03 3.00 0.05 15.56 0.10 -8.4 -25 17° 0.170' 89° 47.956'
PI 13-VIII-02 20 m 1.50 0.06 2.85 0.10 15.50 0.52 -7.3 -8 17° 0.170' 89° 47.956'
PI 13-VIII-02 20 m II 1.54 0.03 2.95 0.03 15.47 0.10 -8.2 -18 17° 0.170' 89° 47.956'
PI 13-VIII-02 30 m 1.45 0.03 2.78 0.07 15.51 0.38 -6.7 -14 17° 0.170' 89° 47.956'
PI 13-VIII-02 30 m II 1.42 0.03 2.73 0.03 14.21 0.17 -7.6 -19 17° 0.170' 89° 47.956'
PI 13-VIII-02 40 m 1.43 0.04 2.75 0.04 14.98 0.29 -7.0 -24 17° 0.170' 89° 47.956'
PI 13-VIII-02 40 m II 1.44 0.03 2.79 0.03 14.50 0.09 -7.8 -32 17° 0.170' 89° 47.956'
PI 13-VIII-02 60 m 1.42 0.04 2.74 0.05 15.19 0.27 -6.7 -24 17° 0.170' 89° 47.956'
PI 13-VIII-02 60 m II 1.43 0.02 2.75 0.04 14.45 0.11 -7.6 -19 17° 0.170' 89° 47.956'
PI 13-VIII-02 90 m 1.34 0.03 2.59 0.07 14.28 0.54 -6.4 -29 17° 0.170' 89° 47.956'
PI 13-VIII-02 90 m II 1.38 0.04 2.65 0.04 13.17 0.14 -8.1 -22 17° 0.170' 89° 47.956'
PI 13-VIII-02 120 m 1.41 0.05 2.72 0.08 15.01 0.49 -6.8 -27 17° 0.170' 89° 47.956'
PI 13-VIII-02 120 m II 1.34 0.01 2.58 0.02 12.85 0.10 -7.8 -26 17° 0.170' 89° 47.956'
PI 13-VIII-02 150 m 1.40 0.04 2.68 0.05 14.68 0.31 -6.8 -17 17° 0.170' 89° 47.956'
PI 13-VIII-02 150 m II 1.41 0.03 2.71 0.04 14.01 0.17 -7.7 -22 17° 0.170' 89° 47.956'
PI1 06-VIII-02-0m 1.58 0.03 3.01 0.04 15.74 0.14 -8.3 -10 17° 0.170' 89° 47.956'
PI1 06-VIII-02-5m 1.48 0.02 2.83 0.02 15.04 0.10 -7.6 -15 17° 0.170' 89° 47.956'
PI1 06-VIII-02-10m 1.47 0.03 2.81 0.03 14.90 0.11 -7.6 -14 17° 0.170' 89° 47.956'
PI1 06-VIII-02-15m 1.16 0.02 2.23 0.02 12.62 0.16 -5.2 -11 17° 0.170' 89° 47.956'
PI1 06-VIII-02-20m 1.10 0.02 2.10 0.02 11.79 0.11 -5.0 -14 17° 0.170' 89° 47.956'
PI5 06-VIII-02-0m 1.44 0.02 2.77 0.02 15.00 0.08 -7.2 -23 17° 0.170' 89° 47.956'
PI5 06-VIII-02-5m 1.51 0.01 2.90 0.03 15.46 0.17 -7.7 -17 17° 0.170' 89° 47.956'
PI5 06-VIII-02-10m 1.49 0.02 2.85 0.02 15.62 0.05 -7.2 -18 17° 0.170' 89° 47.956'
PI2-06-VIII-02-0m 1.45 0.02 2.79 0.03 14.95 0.12 -7.4 -20 17° 0.170' 89° 47.956'
PI17-06-VIII-02-0m 1.50 0.03 2.85 0.03 15.22 0.10 -7.6 0 17° 0.170' 89° 47.956'
PI17-06-VIII-02-7m 1.47 0.03 2.81 0.03 14.97 0.08 -7.5 -13 17° 0.170' 89° 47.956'
PI13-06-VIII-02-150m 1.45 0.01 2.74 0.03 15.01 0.14 -6.9 -1 17° 0.170' 89° 47.956'
PI07-06-VIII-02-0m 1.24 0.02 2.39 0.02 11.91 0.08 -7.2 -22 17° 0.170' 89° 47.956'
PI07-06-VIII-02-7m 1.39 0.03 2.66 0.02 14.00 0.18 -7.3 -18 17° 0.170' 89° 47.956'
Lake Peten #193 B 12-VIII-2005 1.53 0.02 2.94 0.04 15.56 0.14 -7.9 -19 17° 0.170' 89° 47.956'
Peten Itza Lake 6-VII-1995 core site surface 1.39 0.03 2.69 0.07 13.86 0.38 -8.1 -31 17° 0.170' 89° 47.956'
Peten Itza Remate PI-22-VIII-99 core site 1.51 0.04 2.95 0.04 14.89 0.24 -8.7 -41 17° 0.170' 89° 47.956'
Peten Itza Central 29-VIII-99 surf 1.57 0.03 3.02 0.02 15.86 0.42 -8.9 -25 17° 0.170' 89° 47.956'
246 Tables for Chapter 3
Table C.5b. Table C.5a continued. Stable isotope ratios of lake waters from the Yucatán Peninsula.
Sample - Lakewater δ17O (1σ) δ18O (1σ) δD (1σ) d-excess 17O-excess Latitude (N) Longitude (W)
Salpeten Lake
SP-13-VI-02 0m ST1 A A 2.22 0.03 4.23 0.04 20.13 0.32 -13.7 -13 16°58.8' 89°40.0'
SP-13-VI-02 2m ST1 A B 2.23 0.02 4.26 0.02 20.62 0.14 -13.5 -22 16°58.8' 89°40.0'
SP 13-07-02 5 m STIA 2.24 0.05 4.31 0.10 20.23 0.41 -14.3 -36 16°58.8' 89°40.0'
SP-13-VI-02 10m ST1 A A 2.23 0.04 4.27 0.08 21.28 0.45 -12.9 -19 16°58.8' 89°40.0'
SP-13-VI-02 15m ST1 A B 2.09 0.04 4.01 0.06 19.99 0.44 -12.1 -19 16°58.8' 89°40.0'
SP-13-VI-02 20m ST1 A B 2.07 0.02 3.97 0.03 19.95 0.15 -11.8 -27 16°58.8' 89°40.0'
SP-13-VI-02 30m ST1 A A 1.87 0.03 3.59 0.04 16.67 0.21 -12.1 -25 16°58.8' 89°40.0'
SP shore 10-VIII- 2005 2.16 0.02 4.18 0.02 18.32 0.40 -14.9 -44 16°58.8' 89°40.0'
Salpeten 20-I-2006 #210 2.45 0.02 4.70 0.03 23.06 0.37 -14.5 -31 16°58.8' 89°40.0'
SP2 19-VIII-99  bottom 16 m 1.93 0.03 3.70 0.03 16.23 0.11 -13.3 -27 16°58.8' 89°40.0'
Xcaamal Lake
XCA 30-VIII-2 0m 1.08 0.03 2.06 0.05 3.36 0.30 -13.1 -8 20° 36.599' 89° 42.907'
XCA 30-VIII-2 2m 0.89 0.04 1.68 0.04 1.87 0.33 -11.6 -3 20° 36.599' 89° 42.907'
XCA 30-VIII-2 4m 0.93 0.02 1.80 0.03 3.20 0.19 -11.2 -17 20° 36.599' 89° 42.907'
XCA 30-VIII-2 6m 0.90 0.04 1.73 0.07 3.19 0.32 -10.7 -13 20° 36.599' 89° 42.907'
XCA 30-VIII-2 8m 0.91 0.02 1.74 0.04 3.52 0.13 -10.4 -8 20° 36.599' 89° 42.907'
XCA 30-VIII-02 10 m 1.01 0.03 1.90 0.02 4.59 0.13 -10.6 7 20° 36.599' 89° 42.907'
XCA 30-VIII-02 1.02 0.02 1.95 0.02 5.08 0.07 -10.5 -9 20° 36.599' 89° 42.907'
XCA 5-III-04 0m 0.87 0.03 1.67 0.05 3.14 0.29 -10.2 -11 20° 36.599' 89° 42.907'
XCA 5-III-04 1m 0.52 0.02 1.03 0.04 -2.53 0.15 -10.8 -19 20° 36.599' 89° 42.907'
XCA 5-III-04 2m 0.38 0.03 0.74 0.04 -5.56 0.25 -11.5 -16 20° 36.599' 89° 42.907'
XCA 5-III-04 3m 0.46 0.04 0.88 0.07 -1.78 0.36 -8.8 -4 20° 36.599' 89° 42.907'
XCA 5-III-04 4m 0.41 0.04 0.78 0.07 -1.81 0.18 -8.0 0 20° 36.599' 89° 42.907'
XCA 5-III-04 5m 0.78 0.03 1.48 0.06 2.20 0.32 -9.6 -3 20° 36.599' 89° 42.907'
XCA 5-III-04 6m 0.19 0.02 0.38 0.01 -5.13 0.07 -8.2 -15 20° 36.599' 89° 42.907'
XCA 5-III-04 7m 0.43 0.01 0.86 0.01 -4.36 0.14 -11.2 -26 20° 36.599' 89° 42.907'
XCA 5-III-04 8m 0.36 0.02 0.70 0.04 -2.47 0.24 -8.0 -7 20° 36.599' 89° 42.907'
XCA 5-III-04 9m 0.21 0.05 0.43 0.06 -5.53 0.30 -9.0 -15 20° 36.599' 89° 42.907'
XCA 5-III-04 10m 0.46 0.03 0.87 0.04 -0.99 0.27 -8.0 -6 20° 36.599' 89° 42.907'
XCA 5-III-04 11m 0.46 0.03 0.88 0.06 -1.25 0.27 -8.3 -9 20° 36.599' 89° 42.907'
CXCA 5III 04 Cladium H2O 0.64 0.03 1.27 0.04 1.11 0.34 -9.1 -29 20° 36.599' 89° 42.907'
Xcaamal A 26-II-05 1.03 0.03 1.99 0.06 4.58 0.51 -11.3 -15 20° 36.599' 89° 42.907'
Xcaamal 08-VIII-01 used to cond 1.28 0.03 2.51 0.06 7.39 0.27 -12.6 -39 20° 36.599' 89° 42.907'
San Jose Lake
SJ 4-III-04 0m -0.41 0.02 -0.75 0.02 -8.06 0.18 -2.1 -10 20° 52.316' 90° 8.252'
SJ 4-III-04 1m -0.45 0.02 -0.86 0.03 -8.94 0.07 -2.0 4 20° 52.316' 90° 8.252'
SJ 4-III-04 2m -0.16 0.02 -0.33 0.03 -4.60 0.13 -1.9 17 20° 52.316' 90° 8.252'
SJ 4-III-04 3m -0.31 0.03 -0.63 0.03 -6.91 0.16 -1.9 19 20° 52.316' 90° 8.252'
SJ 4-III-04 4m -0.37 0.03 -0.72 0.04 -7.08 0.14 -1.3 15 20° 52.316' 90° 8.252'
SJ 4-III-04 5m -0.28 0.02 -0.55 0.02 -5.71 0.22 -1.3 10 20° 52.316' 90° 8.252'
SJ 4-III-04 6m -0.25 0.02 -0.50 0.03 -5.41 0.08 -1.4 10 20° 52.316' 90° 8.252'
SJ 04-II-04 7 m -0.35 0.01 -0.65 0.03 -6.31 0.17 -1.1 -5 20° 52.316' 90° 8.252'
SJ 4-III-04 8m -0.23 0.02 -0.46 0.02 -5.07 0.07 -1.4 8 20° 52.316' 90° 8.252'
SJ 4-III-04 9m -0.18 0.02 -0.37 0.02 -4.73 0.08 -1.8 11 20° 52.316' 90° 8.252'
SJ 4-III-04 10m -0.29 0.03 -0.58 0.03 -6.57 0.14 -2.0 15 20° 52.316' 90° 8.252'
San Jose B surface #5 0.09 0.03 0.17 0.03 -3.87 0.09 -5.3 1 20° 52.316' 90° 8.252'
247
Table C.5c. Table C.5b continued. Stable isotope ratios of lake waters from the Yucatán Peninsula.
Sample - Lakewater δ17O (1σ) δ18O (1σ) δD (1σ) d-excess 17O-excess Latitude (N) Longitude (W)
Punta Laguna
Punta Laguna 3-IV-05 1.32 0.03 2.58 0.04 11.42 0.09 -9.4 -41 20° 38.888' 87° 38.128'
Punta Laguna 26-V-00 mid basin 0.76 0.02 1.45 0.03 5.83 0.17 -5.8 -6 20° 38.888' 87° 38.128'
Punta Laguna 26-V-00 east basin 1.05 0.03 2.00 0.05 8.40 0.17 -7.6 -8 20° 38.888' 87° 38.128'
Punta Laguna 3-VI-95 0.88 0.03 1.69 0.03 5.64 0.28 -8.9 -12 20° 38.888' 87° 38.128'
Punta Laguna 20-I-96 end of the dock 0.20 0.03 0.41 0.04 0.40 0.16 -3.9 -16 20° 38.888' 87° 38.128'
Punta Laguna 2-5-96 0.22 0.02 0.45 0.05 0.05 0.33 -4.5 -18 20° 38.888' 87° 38.128'
Other lakes
Sacnab surf water 13-VIII-03 2.16 0.05 4.11 0.06 22.07 0.12 -10.8 -8 17° 03.0' 89° 22.0'
Sacnab west end 31-VIII-99 1.39 0.03 2.68 0.03 12.04 0.13 -9.3 -22 17° 03.0' 89° 22.0'
Lake Petenxil surface 16-VIII-03 2.12 0.02 4.05 0.02 15.03 0.17 -17.4 -16 17° 0.170' 89° 47.956'
Sayacil # 2 deep 2.28 0.02 4.38 0.03 16.29 0.17 -18.8 -26 20° 41.037' 88° 48.73'
Sayaucil surface A #6 2.70 0.02 5.18 0.04 19.79 0.19 -21.6 -28 20° 41.037' 88° 48.73'
Sayaucil #2 deep 2.35 0.03 4.54 0.04 17.14 0.31 -19.2 -47 20° 41.037' 88° 48.73'
Sayaucil #6 surface 2.65 0.02 5.11 0.05 19.73 0.15 -21.1 -41 20° 41.037' 88° 48.73'
Yalahau #5 1.66 0.03 3.21 0.03 14.30 0.16 -11.3 -29 20° 39.4' 89° 13.1'
Lake Sacpuy Surface 1.46 0.02 2.84 0.03 9.68 0.28 -13.0 -39 16° 59.4'  90° 3.1'
Lake Sacpuy Surface 16-VIII-03 2.43 0.02 4.69 0.02 22.07 0.38 -15.5 -39 16° 59.4'  90° 3.1'
Lake Yaxha 13-8-03 4.2 m 1.72 0.04 3.29 0.04 16.31 0.21 -10.0 -12 17° 3.5'  89°24.5'
Tikal Aguada 13-VIII-01 2.68 0.02 5.14 0.03 19.46 0.15 -21.6 -30 17° 13.741' 89° 36.131'
Lake Quexil 16-VIII-03 surface 2.16 0.02 4.16 0.03 17.95 0.14 -15.3 -32 16° 55.5' 89° 48.1'
Uaxactun Aguada 13-VIII-01 0.57 0.03 1.10 0.04 -1.43 0.13 -10.2 -12 17° 23.643' 89° 38.077'
Santa Ana Vieja Aguada 14-VIII-01 -0.36 0.02 -0.66 0.04 -4.87 0.21 0.4 -8 16° 38.6' 89° 45.0' 
Aguada Zacpeten  30-I-2006 2.64 0.03 5.08 0.03 21.87 0.40 -18.8 -36 16° 59.255' 89° 39.602'
Laguna Milagros 15-V-02 1.23 0.03 2.34 0.04 11.42 0.34 -7.3 -6 18° 30.2' 88° 25.5'
Sayil little basin 4-III-05 1.92 0.02 3.69 0.04 14.29 0.21 -15.2 -21 20° 10.684' 89° 39.128'
Laguna near Xcaamal A 31-III-05 1.39 0.03 2.67 0.04 15.17 0.34 -6.5 -20 20° 36.599' 89° 42.907'
Laguna near Xcaamal B 31-III-05 1.40 0.03 2.71 0.04 15.66 0.16 -13.8 -35 20° 36.599' 89° 42.907'
Cenote 05-III-04 261B 0m 2.42 0.01 4.65 0.02 18.69 0.16 -18.5 -29 20° 35.707' 89° 42.70'
Cenote 261 A Surf 3.01 0.02 5.82 0.05 23.21 0.32 -23.3 -57 20° 35.707' 89° 42.70'
Cenote 261 95B 2-VIII-2005 1.69 0.02 3.26 0.02 9.98 0.17 -16.1 -27 20° 35.707' 89° 42.70'
Cenote 261 B 26-II-05 2.59 0.01 4.98 0.03 20.23 0.23 -19.6 -37 20° 35.707' 89° 42.70'
Cenote 261 26-II-05 2.35 0.04 4.55 0.05 18.19 0.11 -18.3 -48 20° 35.707' 89° 42.70'
Cenote 261 21-VIII-05 1.23 0.02 2.43 0.03 4.78 0.06 -14.6 -47 20° 35.707' 89° 42.70'
Quexil shore 7-III-2005 2.03 0.05 3.93 0.07 17.57 0.49 -13.8 -36 20° 35.707' 89° 42.70'
Cenote X'caamal #96B 2-VIII-2005 0.63 0.02 1.22 0.03 -0.55 0.09 -10.3 -13 20° 35.707' 89° 42.70'
Monifata 7-VIII-97 0.79 0.02 1.52 0.05 2.66 0.25 -9.5 -10 16° 55.343' 89° 50.247'
La Gloria 21-VIII-09 0.87 0.02 1.72 0.05 3.80 0.33 -10.0 -36 16° 56.745' 90° 22.495'
Uxmal Fountain 0.10 0.03 0.20 0.03 -1.45 0.18 -3.0 -3 20° 21.672' 89° 46.091'
Yalahau surface #3 B 1.62 0.05 3.10 0.06 12.74 0.52 -12.1 -20 20° 39.447' 89° 12.979'
Yalahau deep #1 A 1.62 0.04 3.13 0.04 13.22 0.28 -11.8 -37 20° 39.447' 89° 12.979'
Well Macanche 28-VIII-99 1.46 0.02 2.80 0.03 13.31 0.25 -9.1 -23 16° 58.0' 89° 38.5'
Macanche surface 28-VIII-99 1.62 0.04 3.12 0.06 14.28 0.45 -10.7 -30 16° 58.0' 89° 38.5'
Amatitlan 13-III-2000 1m -2.57 0.03 -4.91 0.03 -39.95 0.12 -0.7 28 14° 28.908' 90° 36.103'
Amatitlan 14-III-2000 core site -1.63 0.03 -3.09 0.05 -29.25 0.26 -4.5 4 14° 28.908' 90° 36.103'
Amatitlan 15-III-2000 core site -2.77 0.04 -5.28 0.09 -42.11 0.51 0.1 23 14° 28.908' 90° 36.103'
Amatitlan 14-III-2000 hot spring -2.98 0.05 -5.66 0.08 -48.06 0.65 -3.1 14 14° 28.908' 90° 36.103'
Coba deep #5 1.31 0.03 2.51 0.04 11.41 0.23 -8.6 -14 20° 29.652' 87° 44.308'
Coba deep #2 1.27 0.01 2.43 0.02 11.22 0.17 -8.3 -11 20° 29.652' 87° 44.308'
Coba surface #1A 1.30 0.03 2.47 0.04 10.86 0.16 -8.9 -8 20° 29.652' 87° 44.308'
Coba surface #1 1.26 0.01 2.41 0.02 10.81 0.15 -8.5 -9 20° 29.652' 87° 44.308'
Lake Coba 2-V-96 -0.15 0.02 -0.25 0.04 -3.69 0.17 -2.6 -16 20° 29.652' 87° 44.308'
L. Coba 2-VI-96 0.54 0.02 1.03 0.02 2.75 0.14 -5.5 0 20° 29.652' 87° 44.308'
248 Tables for Chapter 3
Table C.6a. Stable isotope ratios of river and freshwaters from the Yucatán Peninsula.
Sample - Groundwater δ17O (1σ) δ18O (1σ) δD (1σ) d-excess 17O-excess Latitude (N) Longitude (W)
Bladen River, Toledo, 13-I-02 -2.18 0.03 -4.14 0.02 -21.31 0.08 11.8 7 16° 28.290' 88° 38.766'
Rio Blanco, Toledo, BZ, 11-I-02 -2.09 0.03 -3.96 0.05 -21.07 0.15 10.6 5 15° 19.941' 91° 0.104'
Sibun River, Cayo, BZ 14-I-02 -2.19 0.04 -4.15 0.06 -19.47 0.27 13.7 2 17° 6.566' 88° 39.601'
South Stann Creek, BZ, 13-1I-02 -2.07 0.03 -3.93 0.02 -19.20 0.12 12.3 5 16° 43.449' 88° 25.820'
Rio Sibun 21-VIII-01 -1.64 0.02 -3.10 0.03 -16.67 0.12 8.2 -1 17° 6.566' 88° 39.601'
Rio Dolores 21-VIII-01 -2.35 0.02 -4.47 0.04 -21.20 0.34 14.6 11 15° 41.012' 90° 24.513'
San Simon V 21-VIII-01 -2.59 0.03 -4.93 0.05 -26.00 0.39 13.4 14 15° 50.078' 90° 17.245'
Uspantan 20-VIII-01 -4.68 0.03 -8.90 0.03 -58.81 0.25 12.4 32 15° 19.684' 90° 57.355'
Arroyo 21-VIII-01 -1.80 0.03 -3.45 0.04 -18.74 0.31 8.9 20 16° 24.445' 90° 6.651'
Rio Candelaria  21-VIII-01 -2.10 0.04 -4.00 0.05 -20.30 0.23 11.7 16 15° 53.068' 90° 11.245'
Seasonal wetland Peten 21-VIII-01 -1.23 0.02 -2.36 0.02 -9.20 0.13 9.7 20 16° 8.675' 90° 10.833'
Rio Passion 21-VIII-01 -1.83 0.04 -3.50 0.04 -16.41 0.22 11.6 19 16° 31.868' 90° 11.313'
Coban Rio Cahabon 21-VIII-01 -3.56 0.04 -6.76 0.07 -41.11 0.18 13.0 19 15° 27.926' 90° 22.339'
Rio Ixlu  13-VIII-01 -2.00 0.04 -3.81 0.05 -24.09 0.24 6.4 11 16° 58.460' 89° 41.199'
Rio Dulce  14-VIII-01 -1.70 0.04 -3.26 0.07 -17.37 0.21 8.7 25 15° 39.438' 89° 0.014'
Rio Ixbobo 14-VIII-01 -1.83 0.04 -3.48 0.04 -16.94 0.20 10.9 6 16° 8.866' 89° 24.167'
Town of Copan Rio Aguas Calientes 15-VIII-01 -2.92 0.03 -5.50 0.04 -31.20 0.23 12.8 -9 14° 50.434' 89° 9.212'
Rio Copan at site 15-VIII-01 -3.09 0.02 -5.87 0.04 -33.42 0.30 13.6 12 14° 50.011' 89° 8.487'
Rio Machaquila 14-VIII-01 -1.78 0.03 -3.34 0.04 -18.66 0.18 8.1 -8 16° 23.612' 89° 26.615'
Rio Cienega 15-VIII-01 -1.49 0.02 -2.81 0.02 -11.84 0.06 10.6 -1 15° 44.074' 89° 4.626'
Rio San Pedro 14-VIII-01 -1.68 0.02 -3.18 0.02 -13.79 0.26 11.7 -4 15° 56.833' 89° 14.697'
Rio San Juan 14-VIII-01 -2.06 0.01 -3.90 0.03 -19.93 0.34 11.3 7 16° 37.871' 89° 36.251'
Rio Camotan  15-VIII-01 -3.14 0.03 -5.95 0.03 -36.06 0.25 11.6 13 14° 51.300' 89° 19.424'
Rio Matagun 15-VIII-01 -3.90 0.04 -7.38 0.05 -48.48 0.26 10.6 1 14° 58.6'  89° 31.45'
Cenote Cristal 11-VIII-01 -2.23 0.02 -4.27 0.03 -25.04 0.30 9.1 25 20° 35.707' 89° 42.70'
Cenote Aktun Ha -2.14 0.02 -4.08 0.05 -23.60 0.25 9.0 19 20° 35.707' 89° 42.70'
Cenote Ik Kil near Chichen 12-VIII-01 -2.06 0.03 -3.90 0.04 -21.87 0.19 9.4 0 20° 35.707' 89° 42.70'
Celestun Cenote 18-VIII-94 -0.59 0.02 -1.13 0.02 -6.01 0.22 3.0 7 20° 35.707' 89° 42.70'
Rio Grande 15-VIII-01 -3.01 0.02 -5.73 0.04 -36.49 0.39 9.4 18 14° 57.278' 89° 32.231'
Kikil 21-VI-94 -1.30 0.03 -2.49 0.06 -13.56 0.22 6.4 14 21° 11.511' 88° 10.117'
Bacalar surface 12-VI-02 -1.27 0.04 -2.46 0.06 -15.16 0.15 4.5 35 18° 39.100' 88° 24.522'
Cenote Azul Bacalar 12-VI-02 -2.04 0.03 -3.89 0.05 -24.26 0.15 6.9 16 18° 39.100' 88° 24.522'
E of Paxcaman #142 Source 1-XIII-2005 -1.55 0.05 -2.97 0.07 -15.91 0.40 8.0 12 16° 56.7'  89° 46.3'
Xcambo water sample paso B 3-III-05 -2.05 0.04 -3.90 0.07 -22.71 0.46 8.5 6 21° 18.54' 89° 21.11'
Xcambo 2nd pool A 3-III-05 -1.11 0.02 -2.16 0.03 -14.69 0.25 2.6 28 21° 18.54' 89° 21.11'
Xcambo, paso 3-III-05 -2.09 0.02 -3.98 0.04 -23.85 0.36 8.0 7 21° 18.54' 89° 21.11'
Rio Lagartos Cenote -1.49 0.03 -2.83 0.05 -15.21 0.18 7.4 6 21° 35.759' 88° 8.745'
Telchaquillo Cenote 29-III-05 -2.18 0.04 -4.18 0.04 -24.54 0.15 8.9 26 20° 38.58' 89° 27.29'
Mayapan drip 3-4-5 29-III-05 -2.15 0.02 -4.11 0.03 -25.06 0.26 7.8 18 20° 37.751' 89° 27.631'
Cenote Chi-Huan Holca 30-III-05 -2.25 0.02 -4.27 0.05 -26.76 0.07 7.4 8 20° 45.46'  88° 55.52'
Mayapan pool watertable A 2-III-05 -2.32 0.02 -4.42 0.04 -25.88 0.15 9.5 19 20° 28.564' 89° 11.864'
Mayapan dripwater2 29-II-05 -1.81 0.02 -3.47 0.02 -22.35 0.31 5.4 23 20° 37.751' 89° 27.631'
Mayapan dripwater2 29-II-05 -1.81 0.02 -3.47 0.02 -22.35 0.31 5.4 23 20° 37.751' 89° 27.631'
Calcehtok 2  25-II-05 -1.32 0.03 -2.49 0.02 -9.86 0.26 9.9 -1 20° 33.052' 89° 54.733'
Calcehtok 3 pool -1.42 0.02 -2.70 0.02 -14.36 0.12 6.9 10 20° 33.052' 89° 54.733'
Rio Ixtul 7-VIII-2005 #146 -1.94 0.03 -3.69 0.05 -20.74 0.19 8.5 13 20° 55.37 87° 7.39
Rio Ixtul 7-VIII-2005 #146 -1.95 0.03 -3.68 0.05 -20.60 0.12 9.0 -5 20° 55.38 87° 7.40
Calcehtok 4 cascada -1.58 0.02 -3.00 0.03 -14.40 0.11 9.8 9 20° 33.041' 89° 54.740'
Calcehtok 3 pool -1.43 0.02 -2.69 0.02 -14.02 0.13 7.4 -7 20° 33.041' 89° 54.740'
Calcehtok drip 3 -1.87 0.03 -3.54 0.04 -19.64 0.26 8.7 -3 20° 33.041' 89° 54.740'
Rio Madre Vieja 18-VIII-01 -3.21 0.03 -6.08 0.04 -43.25 0.37 5.5 6 16° 41.818' 89° 44.859'
Rio Benque Viejo 16-VIII-01 -3.67 0.04 -6.96 0.05 -51.06 0.18 4.6 7 17° 4.264'  89° 8.374'
Rio Asuchillo 17-VIII-01 -3.06 0.02 -5.82 0.02 -36.58 0.12 10.1 15 15° 19.941' 91° 0.104'
Rio Blanco 20-VIII-01 -4.77 0.02 -9.05 0.03 -62.84 0.07 9.7 10 15° 19.941' 91° 0.104'
Rio Hato 16-VIII-01 -4.10 0.02 -7.79 0.03 -52.61 0.08 9.9 12 15° 19.941' 91° 0.104'
L Catemaco, MX 21-V-03 01F2 -1.06 0.01 -2.03 0.03 -7.50 0.17 8.8 13 18° 25.44' 95° 7.54'
L. Izabal 3-VI-02 -1.51 0.04 -2.90 0.02 -14.98 0.40 8.0 8 15° 38.052' 88° 59.491'
Edzna water bath #105 2-VIII-05 -2.04 0.02 -3.90 0.03 -24.26 0.28 6.9 17 19° 35.51'  90° 13.42'
Uxmal Fountain 0.12 0.04 0.20 0.08 -1.57 0.19 -3.2 15 20° 21.39' 89° 45.58'
Jalacte Creek 08-I-02 Toledo District -1.78 0.03 -3.40 0.06 -17.44 0.38 9.8 18 16° 10.1' 89° 4.5'
Rio Cunen El Molino 20-VIII-01 -4.70 0.03 -8.94 0.05 -61.18 0.37 10.3 24 15°17.48  91° 4.14'
Madre Vieja Trib 18-VIII-01 -2.98 0.02 -5.67 0.03 -42.81 0.36 2.5 15 16° 41.818' 89° 44.859'
CA-2 Km 90 Rio 18-VIII-01 -3.17 0.02 -6.05 0.02 -37.94 0.11 10.5 32 16° 41.818' 89° 44.859'
Rio Chixoy 20-VIII-01 -4.73 0.02 -9.00 0.02 -64.21 0.35 7.8 30 15° 21.283' 90° 39.361'
Xbuya Ha 10-VI-94 13:53 -1.34 0.03 -2.53 0.03 -14.15 0.19 6.0 -4 21° 23.41'  88° 53.42
Xbuya Ha  940614 10:30 3m VOCAL -2.04 0.05 -3.86 0.08 -22.39 0.33 8.4 -4 21° 23.41'  88° 53.43
Noc Ac 10:30  sup 940614 -2.05 0.03 -3.87 0.07 -22.18 0.21 8.7 -6 21° 23.41'  88° 53.44
Dzibilchaltun 28-5-94 sample 1 -2.10 0.05 -3.97 0.06 -22.98 0.21 8.7 2 21° 5.30' 89° 35.48'
Dzitya cenote 18-5-96 -2.08 0.02 -3.90 0.05 -22.97 0.25 8.2 -13 21° 3.09 89° 40.53'53
Rio Lagartos Chiquila cenote 11-VI-1994 -1.49 0.04 -2.87 0.04 -17.19 0.53 5.7 24 21° 35.759' 88° 8.745'
Rio Lagartos Chiquila cenote 30-VI-1994 -1.47 0.02 -2.83 0.02 -16.56 0.23 6.1 24 21° 35.759' 88° 8.745'
Celestun Cenote 18-6-94 -0.46 0.03 -0.90 0.03 -6.08 0.06 1.1 14 20° 51.37'  90° 23.43'
Dzitnup (Xkeken) 14-VI-96 -1.94 0.05 -3.69 0.09 -21.38 0.41 8.2 16 20° 39.4' 88° 14.34'
Cenote Dziuche 25-V-96 -2.00 0.03 -3.83 0.03 -22.21 0.23 8.4 19 19° 53.57  88° 48.28
Peten Itza 31-X-99 -2.26 0.03 -4.30 0.04 -24.76 0.47 9.6 14 17° 0.170' 89° 47.956'
Cave Actuum Caan 1 #194 15-Aug-2005 -1.33 0.04 -2.51 0.04 -17.21 0.24 2.9 -5 17° 7.47' 88° 51.00'
Cave Actuum Caan 1B #194 15-Aug-2005 -1.37 0.02 -2.58 0.02 -18.30 0.16 2.3 -10 17° 7.47' 88° 51.00'
Cave Actuum Caan 2 #195 15-Aug-2005 -1.83 0.02 -3.47 0.02 -19.47 0.05 8.3 4 17° 7.47' 88° 51.00'
Cave Actuum Caan 3 B 15-Aug-05 -1.90 0.02 -3.58 0.03 -18.57 0.15 10.1 -6 17° 7.47' 88° 51.00'
Cave Tiki Tiki drip (1-2-3) 28-02-05 -0.38 0.03 -0.73 0.04 -2.65 0.16 3.2 0 19° 58.04' 88° 59.642'
Cave Tikitiki pool 28-III-14 -2.64 0.04 -5.00 0.06 -32.48 0.21 7.8 -4 19° 58.04' 88° 59.642'
Cave Tikitiki pool B 28-III-14 -2.65 0.04 -5.05 0.05 -32.73 0.51 7.6 19 19° 58.04' 88° 59.642'
Cave Ixinche water pool 1-III-05 -2.07 0.04 -3.95 0.05 -25.41 0.27 6.2 21 20° 9.608' 88° 47.624'
Cave Ixinche 2nd water sample 1-II-05 -2.40 0.02 -4.54 0.03 -26.61 0.09 9.8 0 20° 9.608' 88° 47.624'
Cave Ixinche water pool B 1-III-05 -2.05 0.02 -3.86 0.03 -26.35 0.15 4.6 -7 20° 9.608' 88° 47.624'
Cave Ixinche 2nd water sample B 1-II-05 -2.36 0.05 -4.46 0.09 -26.94 0.31 8.8 3 20° 9.608' 88° 47.624'
Loltun Cave 21-V-94 Site 1 -1.72 0.04 -3.26 0.06 -16.68 0.25 9.3 -3 20° 15.183' 89° 27.397'
Loltun Cave 21-V-94 Site 2 -2.17 0.03 -4.11 0.04 -22.68 0.31 10.2 0 20° 15.183' 89° 27.397'
Loltun Cave 21-V-94 Site 2 B -2.14 0.02 -4.06 0.03 -23.25 0.32 9.1 9 20° 15.183' 89° 27.397'
Loltun Cave 26-V-96 -1.94 0.02 -3.67 0.04 -19.85 0.16 9.4 -1 20° 15.183' 89° 27.397'
Loltun Cave 26-II-05 -0.60 0.03 -1.13 0.03 -1.86 0.12 7.2 2 20° 15.183' 89° 27.397'
Loltun Cave B 26-II-05 -0.63 0.03 -1.20 0.02 -3.29 0.25 6.3 -2 20° 15.183' 89° 27.397'
249
Table C.6b. Table C.6a continued. Stable isotope ratios of river and freshwaters from the Yucatán
Peninsula.
Sample - Groundwater δ17O (1σ) δ18O (1σ) δD (1σ) d-excess 17O-excess Latitude (N) Longitude (W)
Peten Itza 31-X-99 -2.26 0.03 -4.30 0.04 -24.76 0.47 9.6 14 17° 0.170' 89° 47.956'
Cave Actuum Caan 1 #194 15-Aug-2005 -1.33 0.04 -2.51 0.04 -17.21 0.24 2.9 -5 17° 7.47' 88° 51.00'
Cave Actuum Caan 1B #194 15-Aug-2005 -1.37 0.02 -2.58 0.02 -18.30 0.16 2.3 -10 17° 7.47' 88° 51.00'
Cave Actuum Caan 2 #195 15-Aug-2005 -1.83 0.02 -3.47 0.02 -19.47 0.05 8.3 4 17° 7.47' 88° 51.00'
Cave Actuum Caan 3 B 15-Aug-05 -1.90 0.02 -3.58 0.03 -18.57 0.15 10.1 -6 17° 7.47' 88° 51.00'
Cave Tiki Tiki drip (1-2-3) 28-02-05 -0.38 0.03 -0.73 0.04 -2.65 0.16 3.2 0 19° 58.04' 88° 59.642'
Cave Tikitiki pool 28-III-14 -2.64 0.04 -5.00 0.06 -32.48 0.21 7.8 -4 19° 58.04' 88° 59.642'
Cave Tikitiki pool B 28-III-14 -2.65 0.04 -5.05 0.05 -32.73 0.51 7.6 19 19° 58.04' 88° 59.642'
Cave Ixinche water pool 1-III-05 -2.07 0.04 -3.95 0.05 -25.41 0.27 6.2 21 20° 9.608' 88° 47.624'
Cave Ixinche 2nd water sample 1-II-05 -2.40 0.02 -4.54 0.03 -26.61 0.09 9.8 0 20° 9.608' 88° 47.624'
Cave Ixinche water pool B 1-III-05 -2.05 0.02 -3.86 0.03 -26.35 0.15 4.6 -7 20° 9.608' 88° 47.624'
Cave Ixinche 2nd water sample B 1-II-05 -2.36 0.05 -4.46 0.09 -26.94 0.31 8.8 3 20° 9.608' 88° 47.624'
Loltun Cave 21-V-94 Site 1 -1.72 0.04 -3.26 0.06 -16.68 0.25 9.3 -3 20° 15.183' 89° 27.397'
Loltun Cave 21-V-94 Site 2 -2.17 0.03 -4.11 0.04 -22.68 0.31 10.2 0 20° 15.183' 89° 27.397'
Loltun Cave 21-V-94 Site 2 B -2.14 0.02 -4.06 0.03 -23.25 0.32 9.1 9 20° 15.183' 89° 27.397'
Loltun Cave 26-V-96 -1.94 0.02 -3.67 0.04 -19.85 0.16 9.4 -1 20° 15.183' 89° 27.397'
Loltun Cave 26-II-05 -0.60 0.03 -1.13 0.03 -1.86 0.12 7.2 2 20° 15.183' 89° 27.397'
Loltun Cave B 26-II-05 -0.63 0.03 -1.20 0.02 -3.29 0.25 6.3 -2 20° 15.183' 89° 27.397'
Table C.7. Stable isotope ratios of rainwater from the Yucatán Peninsula.
Sample - Rainwater δ17O (1σ) δ18O (1σ) δD (1σ) d-excess 17O-excess Latitude (N) Longitude (W)
RAIN Chichancanab 7-3-94 -1.44 0.04 -2.74 0.04 -14.22 0.54 7.7 14 19° 52.771' 88° 46.026'
RAIN Chichancanab 7-3-94 B -1.20 0.25 -2.29 0.47 -13.38 1.52 7.3 11 19° 52.771' 88° 46.026'
RAIN-9-I-2007 3:30 PM -0.09 0.02 -0.18 0.05 13.73 0.17 15.2 6 20° 0.993' 89° 1.218'
RAIN-30-III 3:00 PM -1.50 0.03 -2.84 0.04 -18.16 0.36 4.6 -1 20° 0.993' 89° 1.218'
RAIN-25-I-2007 7:00 PM 0.26 0.05 0.47 0.07 15.40 0.41 11.6 6 20° 0.993' 89° 1.218'
RAIN 23-II-2007 8:00 PM -0.08 0.04 -0.15 0.06 8.66 0.26 9.9 3 20° 0.993' 89° 1.218'
RAIN 8-III-2007 8:00 -0.96 0.02 -1.84 0.01 6.51 0.10 21.2 14 20° 0.993' 89° 1.218'
RAIN 15-I-2007 9:00 0.11 0.02 0.20 0.02 13.77 0.21 12.1 2 20° 0.993' 89° 1.218'
RAIN 5-VIII-2007 8:00 -2.80 0.02 -5.33 0.04 -30.52 0.29 12.2 16 20° 0.993' 89° 1.218'
RAIN 23-VI-2007 2:40PM -1.29 0.03 -2.45 0.02 -14.16 0.25 5.5 1 20° 0.993' 89° 1.218'
RAIN 11-V-2007 9:30 PM 0.23 0.03 0.45 0.02 12.61 0.26 9.0 -4 20° 0.993' 89° 1.218'
RAIN 08-II-2017 1:00PM -0.02 0.02 -0.05 0.04 11.26 0.37 11.7 3 20° 0.993' 89° 1.218'
RAIN 13-I-2007 2:15AM 0.43 0.02 0.80 0.03 16.43 0.32 10.0 5 20° 0.993' 89° 1.218'
RAIN 11-V-2007 9:30 PM 0.22 0.02 0.36 0.03 11.02 0.78 8.1 8 20° 0.993' 89° 1.218'
RAIN 08-II-2017 1:00PM -0.03 0.02 -0.11 0.04 10.50 0.28 11.4 11 20° 0.993' 89° 1.218'
RAIN 13-I-2007 2:15AM 0.46 0.03 0.83 0.06 16.37 0.28 9.7 10 20° 0.993' 89° 1.218'
RAIN 06-II-2007 6:00PM -0.42 0.05 -0.84 0.08 9.55 0.45 16.3 16 20° 0.993' 89° 1.218'
RAIN 12-I-2007 2:30PM 0.10 0.05 0.13 0.07 14.37 0.36 13.4 13 20° 0.993' 89° 1.218'
RAIN 6-II-2007 6.00 pm -0.38 0.02 -0.78 0.03 10.11 0.12 16.3 29 20° 0.993' 89° 1.218'
RAIN 11-II-2007 5:00 PM -1.09 0.02 -2.10 0.02 0.79 0.09 17.6 22 20° 0.993' 89° 1.218'
RAIN 12-I-2017 2:30PM 0.08 0.02 0.12 0.04 14.23 0.12 13.3 18 20° 0.993' 89° 1.218'
RAIN  29-IV-2007 0.10 0.02 0.18 0.02 15.12 0.29 13.7 4 20° 0.993' 89° 1.218'
RAIN 31-V-2017 6:00 -3.90 0.05 -7.37 0.07 -46.60 0.61 12.4 -2 20° 0.993' 89° 1.218'
RAIN 31-VII-2005 -0.60 0.03 -1.13 0.03 4.94 0.18 13.8 -1 20° 0.993' 89° 1.218'
RAIN Roof drain 31-VII- 2005 -0.60 0.03 -1.16 0.03 4.33 0.31 13.3 16 20° 0.993' 89° 1.218'
RAIN Punta Laguna 2-VI-96 0.23 0.05 0.42 0.06 11.63 0.16 8.2 7 20° 38.888' 87° 38.128'
RAIN Calcehtok  25-II-05 -0.48 0.03 -0.92 0.03 3.04 0.05 9.9 5 20° 33.041' 89° 54.740'
RAIN Peten Itza 21-I-96 -0.74 0.03 -1.41 0.03 1.50 0.16 12.8 4 17° 0.170' 89° 47.956'
RAIN Merida 16-VI-94 -0.63 0.03 -1.16 0.05 -3.03 0.08 6.2 -17 20° 58.654' 89° 37.407'
RAIN X'Caamal 8-VIII-01 0.20 0.02 0.39 0.03 4.97 0.20 1.8 -4 20° 36.599' 89° 42.907'
RAIN Piste 9-VI-96 -3.24 0.02 -6.15 0.03 -39.02 0.50 10.2 10 20° 41.53 88° 35.22
250 Tables for Chapter 3
Table C.8. Monte Carlo modelling solutions. n defines number of successful simulations.
Age 
(years 
C.E.)
δ18O δD d-excess
17O-excess Mean 
δ18O
Mean 
δD
Mean 
17O-ex
Mean 
δ18O
Mean 
δD
Mean 
17O-ex
Mean 
RH Mean T
Mean 
Xe
min 
RH
Max 
RH
min 
d-excess
Max 
d-
excess
Min 
17O-ex
Max 
17O-ex n
Modern 2.98 13.34 -10.47 -20 -4.35 -22.02 7 2.98 13.34 -21 0.80 24.4 0.85 0.76 0.85 -11.44 -9.56 -26 -14 3338
Modern 2.70 10.39 -11.19 -25 -4.45 -23.58 2 2.70 10.38 -25 0.81 24.8 0.85 0.75 0.86 -12.15 -10.30 -31 -19 5379
Modern 2.76 11.18 -10.87 -29 -4.46 -23.14 -1 2.76 11.16 -29 0.80 24.7 0.85 0.75 0.86 -11.84 -9.98 -35 -23 4305
Modern 3.41 13.08 -14.20 -41 -4.53 -22.95 -4 3.42 13.06 -39 0.76 24.8 0.85 0.71 0.81 -15.12 -13.30 -45 -35 2019
Modern 3.42 14.71 -12.68 -31 -4.37 -21.95 1 3.42 14.69 -31 0.78 24.3 0.85 0.73 0.83 -13.59 -11.72 -37 -25 3476
Modern 3.49 15.00 -12.69 -26 -4.35 -21.90 6 3.49 15.00 -26 0.77 24.3 0.85 0.72 0.83 -13.84 -12.00 -32 -20 3373
Modern 3.44 14.46 -13.08 -30 -4.38 -22.14 2 3.44 14.45 -30 0.77 24.5 0.85 0.72 0.83 -13.99 -12.13 -36 -24 3770
Modern 3.43 14.88 -12.56 -35 -4.39 -21.76 -2 3.43 14.86 -35 0.77 24.3 0.85 0.73 0.83 -13.48 -11.65 -41 -29 2828
Modern 3.38 14.50 -12.52 -29 -4.37 -21.99 3 3.38 14.49 -29 0.78 24.4 0.85 0.73 0.83 -13.48 -11.61 -35 -23 3596
Modern 3.41 14.68 -12.57 -27 -4.37 -21.95 5 3.41 14.67 -27 0.78 24.4 0.85 0.73 0.83 -13.53 -11.70 -33 -21 3432
Modern 3.39 14.56 -12.55 -38 -4.43 -21.91 -4 3.40 14.53 -37 0.77 24.4 0.85 0.73 0.82 -13.49 -11.64 -42 -32 2075
Modern 3.40 14.38 -12.83 -25 -4.37 -22.13 6 3.40 14.37 -25 0.78 24.5 0.85 0.73 0.83 -13.75 -11.89 -31 -19 3489
Modern 3.40 14.34 -12.86 -20 -4.31 -22.27 9 3.40 14.33 -22 0.78 24.4 0.85 0.73 0.83 -13.79 -11.93 -26 -15 2018
Modern 3.36 14.27 -12.65 -30 -4.38 -22.09 2 3.36 14.26 -30 0.78 24.5 0.85 0.73 0.83 -13.54 -11.68 -36 -24 3813
Modern 3.68 15.31 -14.18 -40 -4.43 -21.98 -4 3.69 15.29 -39 0.76 24.5 0.85 0.71 0.81 -15.06 -13.23 -45 -34 2368
Modern 2.98 11.98 -11.28 -31 -4.45 -22.97 -1 2.98 11.98 -31 0.79 24.7 0.85 0.74 0.85 -12.79 -10.93 -37 -25 4283
Modern 2.92 11.94 -10.75 -24 -4.41 -22.93 4 2.92 11.93 -24 0.80 24.6 0.85 0.75 0.85 -12.35 -10.49 -30 -18 4700
Modern 2.88 11.63 -10.81 -24 -4.42 -23.10 4 2.88 11.63 -24 0.80 24.6 0.85 0.75 0.85 -12.35 -10.48 -30 -18 4805
Modern 2.69 10.18 -11.36 -32 -4.52 -23.62 -3 2.69 10.17 -31 0.80 24.9 0.85 0.75 0.85 -12.27 -10.41 -38 -26 3635
Modern 2.62 9.59 -11.37 -27 -4.48 -24.06 0 2.62 9.59 -27 0.81 24.8 0.85 0.75 0.87 -12.31 -10.44 -33 -21 5323
Modern 2.88 11.49 -11.52 -23 -4.42 -23.21 5 2.88 11.49 -23 0.80 24.6 0.85 0.75 0.85 -12.46 -10.62 -29 -17 4742
Modern 2.85 11.47 -11.36 -26 -4.43 -23.13 2 2.85 11.46 -26 0.80 24.6 0.85 0.75 0.86 -12.26 -10.39 -32 -20 4745
Modern 2.90 11.61 -11.55 -24 -4.42 -23.17 4 2.90 11.61 -24 0.80 24.6 0.85 0.75 0.85 -12.51 -10.66 -30 -18 4837
Modern 2.91 11.54 -11.74 -33 -4.49 -23.11 -3 2.91 11.53 -32 0.79 24.8 0.85 0.74 0.85 -12.68 -10.81 -39 -27 3656
Modern 2.65 10.02 -11.19 -25 -4.46 -23.82 2 2.65 10.02 -25 0.81 24.8 0.85 0.75 0.86 -12.10 -10.26 -31 -19 5425
Modern 2.79 11.39 -10.93 -33 -4.49 -22.94 -3 2.80 11.37 -32 0.80 24.7 0.85 0.75 0.85 -11.86 -10.01 -39 -27 3062
Modern 3.82 17.21 -13.37 -42 -4.37 -21.01 -5 3.83 17.17 -40 0.76 23.9 0.85 0.71 0.80 -14.26 -12.50 -44 -36 903
Modern 3.13 13.05 -12.96 -36 -4.47 -22.45 -4 3.13 13.03 -35 0.78 24.6 0.85 0.73 0.83 -12.92 -11.07 -41 -30 2617
Modern 3.27 13.69 -12.48 -36 -4.44 -22.29 -3 3.27 13.68 -35 0.78 24.5 0.85 0.73 0.83 -13.40 -11.54 -41 -30 2902
Modern 2.50 9.92 -10.10 -28 -4.48 -23.43 -2 2.50 9.92 -28 0.81 24.8 0.85 0.76 0.87 -10.99 -9.16 -34 -22 4495
Modern 2.62 10.33 -10.87 -31 -4.50 -23.35 -3 2.62 10.32 -30 0.81 24.8 0.85 0.75 0.85 -11.55 -9.70 -37 -25 3618
Modern 3.41 14.30 -12.95 -36 -4.42 -22.17 -3 3.41 14.29 -35 0.77 24.5 0.85 0.72 0.83 -13.91 -12.06 -42 -30 3113
1040.10 4.33 16.97 -17.68 -46 -4.44 -22.19 -4 4.34 16.95 -44 0.72 24.5 0.85 0.67 0.79 -18.58 -16.74 -51 -40 2473
1032.30 3.85 15.01 -15.82 -38 -4.43 -22.66 -1 3.86 15.00 -38 0.75 24.6 0.85 0.69 0.80 -16.76 -14.91 -44 -32 4148
1019.90 3.60 13.70 -15.13 -33 -4.43 -23.12 1 3.61 13.69 -33 0.76 24.6 0.85 0.70 0.82 -16.04 -14.21 -39 -28 4896
1015.30 3.99 15.02 -16.89 -31 -4.40 -23.00 6 3.99 15.01 -31 0.74 24.6 0.85 0.68 0.80 -17.82 -15.96 -37 -25 4586
1006.00 4.57 18.31 -18.22 -35 -4.35 -21.82 6 4.57 18.28 -35 0.72 24.3 0.85 0.66 0.77 -19.14 -17.29 -41 -29 3295
1001.40 4.37 16.48 -18.47 -4.41 -22.73 3 4.37 16.47 -38 0.72 24.5 0.85 0.65 0.79 -19.41 -17.53 -53 -19 11064
996.80 3.65 15.71 -13.50 -30 -4.36 -21.79 4 3.65 15.69 -30 0.77 24.2 0.85 0.71 0.82 -14.41 -12.56 -36 -24 3296
992.20 3.87 15.18 -15.78 -34 -4.40 -22.59 2 3.87 15.17 -34 0.75 24.5 0.85 0.69 0.81 -16.71 -14.85 -40 -28 4563
987.60 4.37 16.63 -18.34 -31 -4.37 -22.60 8 4.37 16.62 -32 0.72 24.5 0.85 0.66 0.78 -19.27 -17.40 -37 -26 3557
973.10 3.98 15.87 -15.94 -36 -4.40 -22.28 1 3.98 15.86 -36 0.75 24.4 0.85 0.69 0.80 -16.86 -15.02 -42 -31 3969
962.40 4.17 17.00 -16.37 -31 -4.35 -21.89 6 4.17 16.99 -32 0.74 24.4 0.85 0.68 0.79 -17.29 -15.44 -37 -25 3437
951.60 4.24 16.41 -17.48 -36 -4.40 -22.50 3 4.24 16.41 -36 0.73 24.5 0.85 0.67 0.79 -18.41 -16.55 -42 -30 4427
946.30 3.87 15.46 -15.49 -42 -4.46 -22.30 -4 3.87 15.44 -41 0.75 24.5 0.85 0.69 0.80 -16.41 -14.58 -47 -36 2401
930.20 3.73 15.41 -14.46 -40 -4.43 -22.05 -3 3.74 15.39 -39 0.76 24.5 0.85 0.70 0.81 -15.39 -13.53 -45 -34 2637
924.90 3.98 16.35 -15.48 -47 -4.49 -21.91 -5 3.99 16.31 -44 0.74 24.3 0.85 0.69 0.78 -16.35 -14.57 -48 -41 694
919.50 3.84 15.56 -15.13 -32 -4.38 -22.23 3 3.84 15.55 -32 0.75 24.4 0.85 0.70 0.81 -16.07 -14.20 -38 -26 4018
903.10 4.85 18.38 -20.42 -36 -4.37 -22.44 7 4.85 18.38 -36 0.70 24.4 0.85 0.63 0.76 -21.36 -19.49 -42 -30 3638
835.30 4.07 16.44 -16.10 -33 -4.37 -22.07 4 4.07 16.43 -33 0.74 24.4 0.85 0.68 0.80 -17.03 -15.17 -39 -28 3741
823.70 4.18 15.88 -17.52 -43 -4.44 -22.64 -3 4.18 15.87 -42 0.73 24.6 0.85 0.67 0.79 -18.45 -16.62 -48 -37 3760
812.10 4.56 17.72 -18.77 -36 -4.38 -22.22 5 4.56 17.71 -36 0.72 24.4 0.85 0.65 0.77 -19.71 -17.85 -42 -30 3969
800.40 4.42 17.01 -18.33 -36 -4.39 -22.46 4 4.42 16.99 -36 0.72 24.5 0.85 0.66 0.78 -19.26 -17.39 -42 -30 4324
794.50 4.77 18.25 -19.89 -43 -4.38 -22.21 0 4.77 18.23 -43 0.70 24.4 0.85 0.64 0.76 -20.82 -18.96 -49 -37 4039
776.90 3.69 15.91 -13.62 -33 -4.35 -21.70 1 3.69 15.90 -33 0.76 24.2 0.85 0.70 0.82 -14.56 -12.70 -39 -27 3277
771.00 4.26 17.30 -16.80 -32 -4.35 -21.91 6 4.26 17.28 -32 0.73 24.3 0.85 0.68 0.79 -17.74 -15.88 -38 -26 3378
765.20 4.70 18.78 -18.81 -35 -4.34 -21.74 6 4.70 18.77 -35 0.71 24.3 0.85 0.64 0.77 -19.74 -17.89 -41 -29 3206
759.30 4.12 15.92 -17.07 -33 -4.40 -22.61 5 4.13 15.91 -33 0.74 24.5 0.85 0.68 0.79 -18.01 -16.15 -39 -27 4217
759.30 4.37 16.61 -18.31 -32 -4.37 -22.61 7 4.37 16.61 -32 0.72 24.5 0.85 0.66 0.78 -19.24 -17.37 -37 -26 3658
753.40 4.42 16.86 -18.46 -45 -4.44 -22.46 -3 4.42 16.85 -44 0.72 24.6 0.85 0.66 0.77 -19.40 -17.56 -51 -39 3330
747.40 4.01 16.17 -15.88 -45 -4.47 -22.08 -5 4.01 16.13 -43 0.74 24.4 0.85 0.69 0.78 -16.81 -14.97 -48 -40 1354
741.50 4.34 16.14 -18.57 -48 -4.51 -22.80 -4 4.34 16.12 -46 0.72 24.7 0.85 0.66 0.77 -19.49 -17.64 -52 -42 2273
634.30 3.85 15.11 -15.67 -26 -4.35 -22.70 8 3.85 15.11 -27 0.75 24.4 0.85 0.69 0.81 -16.58 -14.73 -32 -20 3045
625.30 4.46 17.73 -17.93 -43 -4.39 -21.92 -2 4.46 17.71 -43 0.72 24.4 0.85 0.66 0.78 -18.86 -17.01 -49 -37 3296
620.70 3.86 15.00 -15.91 -19 -4.21 -23.35 10 3.86 15.00 -22 0.77 24.0 0.85 0.72 0.81 -16.74 -15.05 -25 -17 461
Lakewater UncertaintyData Input Assumed Freshwater input Modeled Lakewater Derived Parameters RH Uncertainty
Appendix D
Tables for Chapters 5, 6 and 7
252 Tables for Chapters 5, 6 and 7
Table D.1a. Isotopic composition of measured gypsum hydration water and calculated mother water values
(d-excess and 17O-excess) from samples obtained from PLG units, SE Spain.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
Cycle 1
CdA C1 10 cm 0.1 5969 3.31 0.03 6.34 0.05 -8.40 0.15 -38 -11.9
CdA C1 40 cm 0.4 5969 2.89 0.03 5.56 0.03 -14.19 0.15 -50 -11.5
CdA C1 70 cm 0.7 5969 2.14 0.04 4.13 0.04 -21.90 0.17 -45 -8.0
Cycle 2
Q1 0.4 5951 2.26 0.00 4.31 0.01 -17.78 0.12 -19 -5.2
Q2 0.8 5951 2.16 0.02 4.11 0.02 -19.46 0.01 -10 -5.3
Q3 1.29 5951 2.06 0.01 3.95 0.01 -19.37 0.31 -26 -3.9
Q4 1.64 5951 2.32 0.01 4.44 0.02 -15.60 0.10 -23 -4.0
Q5 2.32 5951 1.82 0.01 3.47 0.03 -23.31 0.30 -19 -4.1
Q6 3.15 5951 1.85 0.04 3.50 0.05 -20.31 0.44 -5 -1.3
Q7 4.41 5951 1.59 0.01 3.03 0.02 -24.94 0.11 -8 -2.3
Q8 5.1 5951 1.50 0.04 2.86 0.04 -26.32 0.19 -17 -2.4
Q9 5.3 5951 3.34 0.02 6.35 0.03 -3.34 0.11 -15 -6.8
Q10 5.84 5951 2.02 0.03 3.82 0.05 -20.19 0.11 -1 -3.8
Q11 6.65 5951 1.41 0.03 2.65 0.06 -27.03 0.42 3 -1.4
Q12 7.59 5951 1.97 0.02 3.72 0.05 -20.83 0.32 -1 -3.6
Q13 8.64 5951 2.30 0.02 4.39 0.02 -16.48 0.11 -20 -4.6
Q14 9.57 5951 0.84 0.03 1.58 0.03 -32.94 0.20 -7 1.1
Q15 10.32 5951 2.09 0.04 3.97 0.06 -18.45 0.40 -6 -3.2
Q16 11.78 5951 3.09 0.02 5.88 0.02 -6.09 0.22 -16 -5.8
Q17 12.63 5951 1.60 0.04 3.07 0.04 -23.86 0.10 -23 -1.6
Q18 13.51 5951 1.77 0.03 3.37 0.03 -22.32 0.05 -12 -2.4
Q19 14.7 5951 3.39 0.02 6.45 0.02 -3.89 0.14 -17 -8.1
Q20 15.24 5951 2.23 0.03 4.23 0.03 -17.18 0.25 -9 -4.0
Q21 15.54 5951 3.31 0.03 6.28 0.04 -4.26 0.13 -9 -7.2
Q22 16.81 5951 2.04 0.02 3.89 0.02 -19.24 0.03 -12 -3.3
Q23 17.87 5951 1.91 0.03 3.64 0.05 -20.63 0.12 -15 -2.8
Cycle 3
G.6.1 18.2 5931 4.46 0.03 -16.39 0.34 -5.0
Cycle 4
G.5.3 0 5910 2.30 0.04 4.38 0.06 -17.88 0.23 -19 -5.9
G.5.2 6 5910 2.99 0.05 5.69 0.05 -10.97 0.30 -17 -9.3
G.5.1 15.5 5910 5.65 0.11 -11.61 0.46 -9.6
Cycle 5
G.4.2 0 5887 1.92 0.02 3.65 0.04 -25.45 0.21 -14 -7.8
G.4.1 10.7 5887 0.58 0.02 1.10 0.04 -38.55 0.38 -6 -0.8
Cycle 6
G.3.2 0 5867 3.03 0.03 5.79 0.04 -8.22 0.18 -28 -7.2
G.3.1 9.1 5867 2.59 0.03 4.95 0.04 -14.61 0.20 -24 -7.1
Cycle 7
G.2.1 8.9 2.35 0.02 4.52 0.03 -16.89 0.16 -37 -6.0
BOSQUE 00 8.82 5837 1.21 0.03 2.31 0.03 -33.62 0.29 -18 -5.4
BOSQUE 01 8.4 5837 1.51 0.03 2.88 0.05 -29.54 0.10 -15 -5.8
BOSQUE 02 7.98 5837 2.13 0.03 4.06 0.04 -22.59 0.16 -23 -8.2
BOSQUE 03 7.56 5837 1.84 0.02 3.52 0.02 -25.24 0.10 -19 -6.5
BOSQUE 04 7.14 5837 1.72 0.01 3.30 0.02 -26.45 0.15 -21 -6.0
BOSQUE 05 6.72 5837 1.28 0.03 2.44 0.04 -31.71 0.19 -15 -4.5
BOSQUE 06 6.3 5837 0.85 0.03 1.61 0.02 -36.81 0.14 -5 -3.1
BOSQUE 07 5.88 5837 1.98 0.03 3.80 0.04 -23.90 0.12 -25 -7.4
BOSQUE 08 5.46 5837 2.24 0.02 4.28 0.04 -19.94 0.19 -24 -7.2
BOSQUE 09 5.04 5837 2.60 0.02 4.94 0.03 -16.52 0.06 -15 -8.9
BOSQUE 10 4.62 5837 3.21 0.04 6.12 0.05 -10.07 0.26 -23 -11.8
BOSQUE 11 4.2 5837 2.69 0.02 5.15 0.03 -14.25 0.06 -30 -8.3
BOSQUE 12 3.78 5837 3.03 0.05 5.78 0.10 -11.42 0.61 -19 -10.4
BOSQUE 13 3.36 5837 2.83 0.03 5.41 0.04 -13.60 0.07 -24 -9.7
BOSQUE 14 2.94 5837 3.37 0.03 6.42 0.02 -6.97 0.06 -14 -11.0
BOSQUE 15 2.52 5837 2.99 0.02 5.70 0.01 -11.08 0.08 -18 -9.5
BOSQUE 16 2.1 5837 3.36 0.02 6.42 0.02 -5.46 0.10 -32 -9.4
BOSQUE 17 1.68 5837 3.36 0.03 6.40 0.02 -6.36 0.12 -21 -10.2
BOSQUE 18 1.26 5837 3.64 0.03 6.95 0.05 -2.86 0.26 -27 -11.1
BOSQUE 19 0.84 5837 3.43 0.04 6.55 0.04 -5.18 0.30 -35 -10.3
BOSQUE 20 0.42 5837 3.32 0.03 6.33 0.06 -5.36 0.15 -25 -8.6
BOSQUE 21 0.2 5837 2.87 0.04 5.49 0.04 -10.97 0.20 -26 -7.7
G.2.2 0 5837 4.77 0.04 -13.54 0.37 -4.6
Cycle 8
G.1.1 5.55 5816 2.92 0.03 5.54 0.04 -10.48 0.32 -10 -7.6
G.1.2 0 5816 2.97 0.03 5.68 0.03 -9.84 0.38 -33 -8.0
GYPSUM HYDRATION WATER MOTHER WATER
253
Table D.1b. Table D.1a continued. Isotopic composition of measured gypsum hydration water and
calculated mother water values (d-excess and 17O-excess) from samples obtained from PLG
units, SE Spain.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
Cycle 10
RdA -35 cm nuc cone -0.35 5774 2.02 0.05 -32.71 0.33 -13 -2.2
0cm 5cm 0 5774 4.95 0.06 -15.44 0.22 -7.9
10cm 11cm 12cm 0.1 5774 5.19 0.05 -14.14 0.27 -16 -8.5
50 cm 0.5 5774 4.93 0.03 -14.33 0.20 -6.6
56cm 57cm 0.55 5774 6.32 0.06 -6.23 0.15 -15 -9.5
87cm 91cm 0.87 5774 5.39 0.06 -11.62 0.28 -7.5
102cm 105cm 1.02 5774 4.43 0.04 -18.82 0.29 -7.2
107cm 1.07 5774 5.67 0.04 -12.98 0.26 -11.2
117cm 1.17 5774 4.63 0.04 -19.11 0.24 -9.1
137cm 1.37 5774 3.68 0.04 -24.22 0.49 -6.8
153cm 1.53 5774 5.25 0.05 -13.63 0.20 -29 -8.4
161cm 1.61 5774 3.65 0.04 -23.85 0.27 -6.1
192cm 1.92 5774 4.39 0.02 -17.82 0.28 -37 -5.9
231cm 2.31 5774 3.30 0.04 -23.98 0.28 -25 -3.5
272cm 2.72 5774 5.24 0.04 -12.55 0.15 -32 -7.3
306cm 3.06 5774 3.29 0.01 -24.12 0.30 -30 -3.6
SC50A&B 4.5 5774 4.31 0.03 -19.90 0.30 -7.4
SC100A 5 5774 3.34 0.04 -25.71 0.25 -5.6
SCTOP 6 5774 2.83 0.02 -30.26 0.28 -10 -6.1
Cycle 11
C7-01 4.0 5754 1.42 0.02 2.69 0.05 -29.65 0.28 -3 -4.4
C7-02 4.0 5754 2.15 0.03 4.10 0.06 -21.27 0.26 -23 -7.1
C7-04 4.0 5754 1.21 0.04 2.31 0.05 -29.36 0.17 -19 -1.1
C7-05 3.0 5754 1.43 0.04 2.72 0.04 -29.83 0.19 -7 -4.8
C7-06 1.0 5754 2.24 0.01 4.30 0.01 -19.82 0.38 -29 -7.2
C7-07 0.0 5754 2.21 0.02 4.20 0.03 -20.36 0.14 -12 -7.0
Cycle 12
BdI C15 -5 cm -0.05 5721 2.92 0.04 5.58 0.06 -14.41 0.40 -24 -11.9
BdI C15 10 cm 0.1 5721 3.29 0.02 6.29 0.07 -7.25 0.29 -29 -10.3
BdI C15 46 cm 0.46 5721 3.16 0.05 6.03 0.09 -9.73 0.27 -30 -10.8
BdI C15 130 cm 1.3 5721 3.04 0.18 5.82 0.34 -11.62 0.99 -34 -11.0
BdI C15 200 cm 2 5721 2.38 0.04 4.59 0.06 -17.23 0.30 -44 -6.9
BdI C15 400 cm 4 5721 0.98 0.04 1.89 0.08 -33.86 0.34 -20 -2.3
BdI C15 450 cm 4.5 5721 0.87 0.04 1.72 0.07 -36.28 0.44 -41 -3.4
BdI C15 480 cm 4.8 5721 1.89 0.03 3.63 0.06 -25.41 0.28 -29 -7.6
BdI C15 520 cm 5.2 5721 2.30 0.05 4.46 0.07 -23.99 0.41 -55 -12.7
BdI C15 550 cm 5.5 5721 1.94 0.05 3.75 0.12 -21.54 1.24 -42 -4.6
BdI C15 600 cm 6 5721 1.12 0.03 2.15 0.07 -34.07 0.25 -23 -4.6
Cycle 13
BdI C16 0m 0 5702 2.64 0.02 5.08 0.05 -15.05 0.18 -43 -8.6
BdI C16 10m 0.1 5702 0.97 0.03 1.84 0.06 -34.74 0.44 -7 -2.8
BdI C16 25m 0.25 5702 3.54 0.05 6.81 0.06 -5.37 0.50 -53 -12.5
BdI C16 35m 0.35 5702 1.55 0.02 2.99 0.06 -27.17 0.48 -34 -4.2
BdI C16 55m 0.55 5702 2.91 0.03 5.61 0.05 -10.74 0.24 -57 -8.4
BdI C16 75 cm 0.75 5702 1.83 0.04 3.52 0.05 -23.09 0.46 -35 -4.3
BdI C16 95 cm 0.95 5702 2.79 0.03 5.36 0.06 -12.92 0.18 -43 -8.7
BdI C16 120 cm 1.2 5702 3.60 0.06 6.87 0.17 -1.79 0.51 -33 -9.3
BdI C16 140 cm 1.4 5702 3.82 0.03 7.33 0.02 0.31 0.27 -48 -10.8
BdI C16 160 cm 1.6 5702 3.16 0.04 6.08 0.07 -6.39 0.16 -49 -7.7
BdI C16 240 cm 2.4 5702 4.13 0.02 7.90 0.04 4.04 0.12 -43 -11.6
Cycle 15
RdA C16 -35 cm -0.35 5663 2.37 0.05 4.58 0.08 -18.81 0.38 -49 -8.4
RdA  C16 5 cm 0.05 5663 0.14 0.05 0.26 0.09 -45.27 0.86 -5 -0.9
RdA C16 45 cm 0.45 5663 2.37 0.02 4.55 0.05 -19.15 0.38 -40 -8.6
RdA C16 85 cm 0.85 5663 3.36 0.04 6.45 0.08 -7.27 0.71 -44 -11.6
RdA C16 200 cm 2 5663 3.40 0.05 6.50 0.06 -5.81 0.15 -35 -10.5
RdA C16 270 cm 2.7 5663 2.13 0.04 4.11 0.06 -20.45 0.24 -38 -6.3
RdA C16 320 cm 3.2 5663 3.12 0.03 6.00 0.06 -9.35 0.18 -53 -10.1
Nijar Basin
GY-NIJAR-01 -1.21 0.02 -2.34 0.02 -62.86 0.21 25 1.9
GY-NIJAR-02 1.84 0.02 3.49 0.03 -22.05 0.18 -11 -3.1
GY-NIJAR-03 2.89 0.03 5.51 0.03 -10.80 0.18 -18 -7.7
GY-NIJAR-04 -1.21 0.03 -2.38 0.04 -58.48 0.58 36 6.6
GY-NIJAR-05 -0.53 0.01 -1.05 0.02 -58.82 0.17 15 -4.3
GY-NIJAR-06 -0.76 0.02 -1.47 0.04 -57.94 0.59 8 -0.1
GY-NIJAR-07 2.63 0.03 5.00 0.04 -15.43 0.34 -20 -8.4
GYPSUM HYDRATION WATER MOTHER WATER
254 Tables for Chapters 5, 6 and 7
Table D.1c. Table D.1b continued. Isotopic composition of measured gypsum hydration water and
calculated mother water values (d-excess and 17O-excess) from samples obtained from PLG
units, Sicily.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
 Resedimented Lower Gypsum, Sicily 
EM MS massive selenite A 5579 3.94 0.04 7.55 0.06 0.74 0.27 -38 -12.1
EM MS massive selenite B 5579 3.38 0.04 6.49 0.05 -8.77 0.22 -43 -13.4
EM lower gypsum banded in place 2B 5579 3.04 0.01 5.79 0.01 -10.19 0.14 -20 -9.3
EM lower banded gypsum hard bit location 2A 5579 2.95 0.02 5.65 0.03 -10.80 0.22 -34 -8.8
 Primary Lower Gypsum, Sicily 
Cycle 3
MG-C3 22m 22 5929 7.35 0.09 0.68 0.51 -10.7
MG-C3 28m 28 5929 7.84 0.07 3.37 0.33 -11.8
MG-C3 44m recrys? 44 5929 5.86 0.03 -9.83 0.48 -9.5
Cycle 4
MG-C4 BASE 0 5910 2.46 0.03 4.74 0.05 -19.20 0.22 -48 -10.1
MG-C4-3m 3 5910 7.26 0.08 -1.36 0.38 -12.0
MG-C4 6m 6 5910 3.10 0.03 5.95 0.07 -14.20 0.22 -44 -14.7
MG-C4 8m 8 5910 5.95 0.07 -14.11 0.24 -14.5
MG-C4-15m banded 15 5910 7.16 0.06 -2.61 0.33 -12.5
MG-C4 15m 15 5910 3.70 0.04 7.11 0.04 -2.52 0.14 -58 -12.0
MG-C4-19m 19 5910 6.65 0.05 -5.62 0.17 -11.5
MG-C4 24m (rep) 24 5910 3.45 0.05 6.61 0.10 -3.57 0.33 -38 -9.0
 MG-C4-24m 24 5910 6.93 0.06 -3.35 0.22 -11.4
 MG-C4-28m 28 5910 6.38 0.05 -8.81 0.35 -12.5
Cycle 5
MG-C5 1m selen espagetti 1 5887 6.02 0.08 -10.42 0.19 -11.3
Cycle 6
MG-C6 1m selenite 10cm 1 5867 6.63 0.06 -6.56 0.21 -12.3
Monte Banco Cycles 1 to 6
MBC1 bottom good crystal 7.28 0.05 -5.47 0.18 -16.3
MBC1 middle drilled pieces 6.24 0.03 -8.57 0.13 -11.2
MB-C2 middle drilled pieces 7.90 0.11 4.48 0.38 -11.2
MBC3 Selenite surface 7.10 0.06 -2.32 0.31 -11.7
MBC3 Selenite surface (rep) 3.83 0.01 7.30 0.02 -2.03 0.12 -29 -13.0
MBC4 middle bulk 6.62 0.05 -6.90 0.24 -12.5
MBC4 middle bulk (rep) 3.39 0.02 6.47 0.02 -7.22 0.10 -29 -11.6
MBC6 30 cm drilled pieces 6.26 0.08 -9.80 0.15 -12.6
MBC6 30 cm drilled pieces (rep) 3.45 0.07 6.58 0.13 -7.71 0.76 -27 -13.1
Rocca d'Entella
RdE massive selenite by the road 7.48 0.03 0.36 0.27 -12.0
RdE selenite photo 6.11 0.03 -8.93 0.27 -10.6
RdE banded photo 7.04 0.06 -2.61 0.36 -11.5
RdE banded cones above observatory 6.26 0.06 -9.54 0.35 -12.3
RdE upper part of mountain road 4.33 0.01 -20.20 0.08 -7.9
Santa Ninfa
SN banded gypsum 6.39 0.04 -4.40 0.19 -8.2
SN well preserved selenite 7.21 0.07 0.20 0.28 -10.0
SN well preserved selenite (rep) 3.82 0.04 7.27 0.06 0.27 0.15 -17 -10.4
SN massive selenite 6.39 0.04 -4.58 0.32 -8.4
MGU
MGU marl-gyp top C1 6.27 0.07 -9.13 0.37 -12.0
 MGU selenite bottom C1 7.72 0.08 -0.94 0.44 -15.3
 MGU selenite bottom C1 (rep) 3.85 0.01 7.34 0.02 -1.31 0.06 -27 -12.6
 MGU massive selenite C2 8.56 0.06 5.70 0.25 -15.1
GYPSUM HYDRATION WATER MOTHER WATER
255
Table D.2a. Isotopic composition of measured gypsum hydration water and calculated mother water values
(d-excess and 17O-excess) from samples obtained from UG units, Sicily.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
Eraclea Minoa
Cycle 2
EM2-6 5511
EM2-2 5511
Cycle 3
EM3-12 5482
EM3-7 5482
EM3 0 cm 0 5482 8.33 0.05 4.78 0.38 -14.3
EM3 -30 c -0.3 5482 8.81 0.04 8.74 0.34 -14.1
EM3 -60 cm -0.6 5482 9.12 0.12 9.97 0.33 -15.3
EM3 -150 cm -1.5 5482 9.18 0.05 13.42 0.26 -12.3
EM3 -200 cm -2 5482 9.02 0.04 9.75 0.22 -14.7
EM3 -205 cm -2.05 5482 8.48 0.06 7.22 0.26 -13.0
EM3 -220 cm -2.2 5482 9.34 0.03 12.32 0.25 -14.6
EM3 -230 -2.3 5482 7.09 0.02 -2.61 0.15 -11.9
EM3 -260 -2.6 5482 8.10 0.04 3.31 0.38 -13.9
EM3 -290 cm -2.9 5482 7.58 0.06 1.67 0.29 -11.5
EM3 -320 cm -3.2 5482 8.20 0.07 5.23 0.27 -12.7
EM3 -360 cm -3.6 5482 8.63 0.06 6.02 0.24 -15.4
EM3 -400 -4 5482 6.93 0.07 -2.20 0.21 -10.3
Cycle 4
EM4-20 5459
EM4-13 5459
EM 4 dome top 9 5459 4.86 0.02 9.29 0.04 7.93 0.20 -40 -18.7
EM 4 dome 0-26cm 8.76 5459 2.60 0.04 4.98 0.04 -21.86 0.35 -29 -14.7
EM 4 dome 30 cm 8.7 5459 4.48 0.04 8.55 0.06 3.90 0.17 -29 -16.9
EM 4 dome 39-60 cm 8.4 5459 2.69 0.03 5.13 0.06 -20.61 0.40 -24 -14.7
EM4 dome 60-81 8.29 5459 4.43 0.02 8.51 0.02 5.91 0.10 -55 -14.5
EM 4 dome 81-98 cm 8.02 5459 4.90 0.03 9.36 0.02 9.64 0.36 -43 -17.5
EM4 dome-119-146cm 7.54 5459 4.62 0.03 8.88 0.05 5.03 0.32 -63 -18.4
EM4 dome 300 cm 6 5459 5.20 0.04 9.98 0.06 17.54 0.30 -63 -14.4
EM4 palisade A 5.25 5459 5.13 0.03 9.82 0.05 14.12 0.20 -48 -16.6
EM4 palisade B 5 5459 4.78 0.03 9.16 0.03 9.85 0.41 -54 -15.7
EM4 palisade C 4.75 5459 4.77 0.02 9.11 0.03 8.88 0.13 -36 -16.3
EM4 palisade D 4.5 5459 3.76 0.05 7.21 0.08 -1.45 0.17 -43 -11.7
EM4 palisade E 4.25 5459 4.72 0.02 9.02 0.03 8.37 0.11 -45 -16.1
EM4 palisade F 4 5459 5.11 0.03 9.79 0.04 12.58 0.36 -53 -18.0
EM4 palisade G 3.75 5459 4.46 0.04 8.53 0.06 3.44 0.13 -37 -17.2
EM4 palisade H 3.5 5459 4.05 0.04 7.80 0.05 3.71 0.26 -64 -11.1
EM4 palisade I 3.25 5459 4.85 0.03 9.30 0.06 11.27 0.31 -50 -15.4
EM4 palisade K 3 5459 4.94 0.05 9.43 0.08 12.86 0.35 -40 -14.8
EM4 UGF 4B 1-01-20 cm 2.92 5459 3.99 0.04 7.62 0.07 -2.00 0.24 -30 -15.5
EM laminated gypsum 40 cm below unit K 2.6 5459 4.01 0.05 7.68 0.08 -0.83 0.34 -40 -14.8
EM4 UGF 4-01-130 cm 2.48 5459 3.41 0.02 6.52 0.05 -9.77 0.07 -34 -14.7
EM4 UGF 4-02-200cm 2.2 5459 3.43 0.03 6.57 0.06 -8.90 0.40 -40 -14.2
EM4 3A-02-190 cm 0.94 5459 3.22 0.04 6.14 0.09 -11.71 0.27 -19 -13.6
EM4 Gypsorudite 0 5459 2.76 0.04 5.26 0.05 -18.63 0.18 -20 -13.6
EM4 veins 60 cm 5459 5.72 0.03 -14.70 0.56 -13.3
 EM4 UGF 3a VEINS 120 cm 5459 5.76 0.02 -13.24 0.21 -12.1
Cycle 5
EM5-31 5436
EM5-22 5436
EM5 A 4.5 5436 3.34 0.05 -32.01 0.36 -12.0
EM5 B 4 5436 5.30 0.01 -18.95 0.29 -14.3
 EM5 C 3.5 5436 8.84 0.06 6.39 0.29 -16.7
 EM5 D 3 5436 10.48 0.05 18.26 0.20 -17.7
EM5 E 2.5 5436 8.62 0.07 4.65 0.26 -16.7
EM5 F 2 5436 9.33 0.03 10.41 0.18 -16.5
 EM5 G 1.5 5436 8.72 0.07 6.44 0.32 -15.7
 EM5 H 1 5436 8.60 0.05 3.40 0.16 -17.8
EM5 I 0.5 5436 8.94 0.04 7.67 0.33 -16.2
EM5 BASE 0 5436 8.75 0.05 5.28 0.27 -17.1
Cycle 6
EM6-39 5416
EM6-33 5416
Cycle 7
EM7-44 5343
EM7-43 5343
EM7-42 5343
EM7-41 5343
EM Aren -21 to -17 cm -0.17 5343 4.31 0.02 8.30 0.04 8.05 0.25 -69 -10.7
EM AREN  upper 0-6 cm 0.06 5343 4.18 0.01 8.02 0.01 2.03 0.24 -51 -14.6
EM AREN 76-83 cm 0.83 5343 3.59 0.03 6.89 0.06 -3.71 0.53 -53 -11.5
EM Aren 112-122 cm 1.12 5343 3.70 0.03 7.13 0.04 -1.89 0.20 -60 -11.5
EM Aren 143-148 cm 1.48 5343 4.25 0.01 8.18 0.02 5.93 0.27 -65 -11.9
EM Aren 191-211 cm 2.11 5343 4.50 0.04 8.65 0.09 7.84 0.57 -60 -13.7
GYPSUM HYDRATION WATER MOTHER WATER
256 Tables for Chapters 5, 6 and 7
Table D.2b. Table D.2a continued. Isotopic composition of measured gypsum hydration water and
calculated mother water values (d-excess and 17O-excess) from samples obtained from deep
basin units.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
Leg 42A Site 372
42 372Z8RW 30.5-33.5cm 0.305 5325 6.05 0.02 -9.65 0.46 -10.8
42 372Z8RW 30.5-33.5cm 0.305 5325 3.11 0.03 5.98 0.02 -9.82 0.09 -48 -10.4
42 372Z8R1W 46-48.5cm 0.46 5325 5.74 0.10 -10.96 0.20 -9.7
42 372Z8R1W 46-48.5cm 0.46 5325 3.16 0.06 6.04 0.08 -10.78 0.19 -26 -11.9
42 372Z8RW 91-94cm 0.91 5325 6.07 0.11 -8.95 0.47 -10.3
43 372Z8RW 91-94cm 0.91 5325 3.33 0.04 6.37 0.07 -8.47 0.21 -36 -12.2
42 372 Z8 R1 W 96-99cm 0.96 5325 5.47 0.07 -11.65 0.53 -8.2
42 372 Z8 R1 W 96-99cm 0.96 5325 3.16 0.05 6.04 0.06 -10.48 0.28 -30 -11.6
 42 372 Z8 R2 W 98-100 cm 0.98 5325 6.51 0.04 -7.04 0.28 -11.8
 42 372 Z8 R2 W 98-100 cm 0.98 5325 3.11 0.03 5.92 0.09 -9.17 0.25 -14 -9.2
42 372Z8R1 W 126-128.5 1.26 5325 6.37 0.09 -8.37 0.32 -12.0
42 372Z8R1 W 126-128.5 1.26 5325 3.22 0.03 6.20 0.02 -7.28 0.13 -57 -9.6
42 372 Z8 R2 W 146-149 cm 1.46 5325 6.08 0.04 -8.88 0.33 -10.2
42 372 Z8 R2 W 146-149 cm 1.46 5325 2.61 0.01 5.06 0.02 -11.34 0.16 -70 -4.7
Leg 42A Site 374
      42 374 Z11 R2 58.0-59cm 0.58 5315 3.65 0.03 -19.32 0.29 -1.5
42 374Z12R1W 125-127cm 1.25 5315 3.99 0.05 -19.19 0.20 -4.1
42 374Z12R2W 127-130 cm 1.27 5315 3.98 0.06 -19.48 0.23 -4.3
42 374 Z12 R2 W 112-114 cm 1.12 5315 4.11 0.12 -17.15 0.30 -3.0
 42 374 Z16 R1 W 89-90cm 0.89 5315 2.30 0.03 4.34 0.04 -13.26 0.16 7 -0.9
42 374Z16R1W 130-135cm 1.3 5315 4.30 0.03 -15.42 0.32 -2.7
42 374Z16R1W 136-135cm 1.36 5315 4.62 0.04 -17.24 0.21 -7.1
 42 374Z17R1W 37-39cm 0.37 5315 4.45 0.08 -11.46 0.52 0.1
42 374Z17R1W 58-59cm 0.58 5315 4.22 0.06 -13.08 0.48 0.3
42 374Z17R1W 82-86cm 0.82 5315 4.08 0.03 -14.89 0.21 -0.4
 42 374 Z17 R1 W 110-115 cm 1.1 5315 4.38 0.04 -14.00 0.47 -1.9
 42 374Z17R1W 115-119cm 1.15 5315 4.85 0.07 -15.23 0.59 -6.9
 42 374 Z19 R1 W 19-21 cm 0.19 5315 2.27 0.01 4.32 0.03 -9.94 0.23 -13 2.7
  42 374 Z19 R1 W 25-27cm 0.25 5315 3.94 0.07 -14.01 0.33 1.6
42374Z19R1W 77-80cm 0.77 5315 4.57 0.06 -11.23 0.48 -0.6
42374Z19R1W 83-84cm 0.83 5315 4.82 0.07 -18.11 0.33 -9.6
 42 374 Z19 R1 W 103-104 cm 1.03 5315 4.65 0.05 -16.82 0.44 -6.9
 42 374Z19R1W 124-126cm 1.24 5315 4.31 0.05 -13.31 0.32 -0.7
42 374Z19R1W 136-138cm 1.36 5315 4.44 0.07 -14.04 0.23 -2.5
 42 374 Z20 R1 W 51-55 cm 0.51 5315 2.35 0.05 4.47 0.07 -10.73 0.36 -10 0.7
42 374 Z20 R1W 33-36cm 0.33 5315 4.37 0.06 -14.13 0.41 -1.9
42 374Z20R1W 77-81cm 0.77 5315 4.81 0.08 -15.27 0.33 -6.7
42 374Z20R1W 100-104cm 1 5315 4.79 0.05 -17.97 0.29 -9.3
GYPSUM HYDRATION WATER MOTHER WATER
257
Table D.2c. Table D.2b continued. Isotopic composition of measured gypsum hydration water and
calculated mother water values (d-excess and 17O-excess) from samples obtained from deep
basin units.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
Leg 107 Site 654
107 654 A 27 R1w 6-9 cm 0.06 5310 2.44 0.04 4.67 0.07 -13.33 0.21 -27 -3.6
107 654A 27R1 w 12-15 0.12 5310 2.65 0.04 5.07 0.06 -13.15 0.15 -28 -6.6
 107-654 A 27-R1w 19-23cm 0.19 5310 2.62 0.04 5.01 0.06 -11.86 0.17 -31 -4.8
 107-654 A 27-R1w 20cm 0.2 5310 2.65 0.04 5.09 0.04 -10.74 0.16 -45 -4.3
 107-651 A 27-R1w 26cm 0.26 5310 2.88 0.03 5.53 0.04 -8.95 0.28 -37 -5.9
107 654A 27 R1 w 30-33cm 0.3 5310 2.47 0.04 4.69 0.08 -14.56 0.58 -8 -5.0
107 654 A 27 R1w 34-36 cm 0.34 5310 2.48 0.05 4.76 0.11 -14.50 0.23 -33 -5.4
107 654A 27R1 w 43-45 0.43 5310 2.82 0.03 5.36 0.03 -10.80 0.10 -12 -6.5
 107-654 A 27R CC W 10cm 1 5310 2.57 0.03 4.92 0.03 -12.56 0.21 -29 -4.8
 107-654 A 28-R1w 8cm 0.08 5310 2.81 0.06 5.37 0.13 -10.22 0.41 -29 -6.0
 107-654-A28-R1W-25-28 cm 0.25 5310 2.16 0.04 4.15 0.07 -16.05 0.30 -35 -2.2
 107-654-A28-R1W-33-34 cm 0.33 5310 2.02 0.02 3.86 0.04 -18.65 0.47 -24 -2.5
 107-654-A28-R1W-51-54 cm 0.51 5310 2.14 0.08 4.11 0.14 -16.46 0.47 -30 -2.3
107 654 A 29 R2 21-22 cm 0.21 5310 2.41 0.05 4.60 0.07 -15.47 0.14 -17 -5.2
107 654 30 R1 w 75-76 cm 0.75 5310 2.23 0.04 4.22 0.06 -15.62 0.22 -8 -2.3
107 654 A 30 R1 121-123 cm 1.21 5310 2.30 0.06 4.41 0.07 -15.00 0.23 -30 -3.2
107 654 A 30 R1 135-137 cm 1.35 5310 2.44 0.02 4.65 0.06 -14.73 0.19 -13 -4.8
107 654 A 30 R1 135-137 cm 1.35 5310 1.82 0.05 3.48 0.09 -19.40 0.29 -21 -0.2
107 654 31 R1 w 2-3 cm 0.02 5310 4.18 0.03 8.01 0.05 6.43 0.35 -44 -10.0
107 654A 31R1 w 105-107 cm 1.05 5310 2.50 0.03 4.76 0.03 -16.22 0.29 -14 -7.2
107 654A 31R1 w 89-90 cm 0.89 5310 2.18 0.03 4.17 0.06 -15.35 0.28 -27 -1.6
107 654A 31R1 w 127-129 cm 1.27 5310 2.33 0.04 4.45 0.05 -13.53 0.09 -23 -2.0
107 654A 31R1 w 134-135 cm 1.34 5310 2.11 0.02 4.05 0.03 -15.73 0.04 -32 -1.0
107 654 A 32 R1   3-5 cm 0.03 5310 2.07 0.03 3.95 0.05 -17.16 0.19 -23 -1.7
107 654 32 R1 w 107-110 cm 1.07 5310 2.06 0.02 3.93 0.02 -17.24 0.13 -19 -1.7
107 654 32 R1 w 144-145 cm 1.44 5310 2.35 0.04 4.45 0.08 -14.43 0.48 -4 -2.9
107 654 32 R1 w 27-28.5 cm 0.27 5310 2.21 0.04 4.24 0.04 -15.77 0.12 -32 -2.6
107 654 32 R2 w 9-10 cm 0.09 5310 2.30 0.03 4.34 0.04 -11.23 0.13 0 1.2
107 654 32 R2 w 9-10 cm 0.09 5310 2.16 0.01 4.14 0.03 -16.41 0.11 -27 -2.5
107 654 32 R2 w 41-43 cm 0.41 5310 2.25 0.04 4.30 0.05 -16.33 0.20 -23 -3.7
107 654 3 2R2 w 144-145 cm 1.44 5310 2.30 0.02 4.37 0.02 -15.71 0.12 -13 -3.6
107 634 A 33 R1 w 72-74 cm 0.72 5310 3.53 0.05 6.74 0.05 -2.78 0.33 -33 -9.3
107 634 A 33 R1 w 63-66 cm 0.63 5310 3.16 0.02 6.03 0.03 -7.75 0.10 -24 -8.7
107 654 34 R1 w 30-37 cm 0.3 5310 2.30 0.05 4.38 0.07 -16.49 0.22 -9 -4.4
 107 634 A 34 R1 w 35-37 cm 0.35 5310 1.88 0.02 3.58 0.02 -18.07 0.18 -15 0.3
107 654 34R1 w 51-53 cm 0.51 5310 2.51 0.02 4.79 0.04 -14.09 0.11 -24 -5.3
107 654 A 34R1w 99-106 cm 0.99 5310 2.40 0.02 4.56 0.02 -14.13 0.12 -10 -3.5
107 654 34R1 w 101-102.5 cm 1.01 5310 2.06 0.04 3.93 0.06 -16.26 0.31 -19 -0.6
107 654 A 34 R1 119-122 cm 1.19 5310 2.20 0.05 4.21 0.07 -15.75 0.16 -23 -2.4
107 654 34 R2 w 2-3 cm 0.02 5310 2.60 0.02 4.97 0.02 -12.61 0.45 -23 -5.2
 107 634 A 34 R2 w 18-20 cm 0.18 5310 2.28 0.03 4.38 0.04 -14.21 0.19 -32 -2.2
107 654 A 34R2w 32-33 cm 0.32 5310 2.18 0.04 4.16 0.08 -16.58 0.52 -17 -2.8
107 654 A 34R2w 75-76 cm 0.75 5310 2.22 0.04 4.24 0.06 -15.68 0.36 -20 -2.5
107 634 A 34 R2 w 90-94 cm 0.9 5310 2.24 0.03 4.27 0.05 -14.70 0.33 -13 -1.7
 107 634 A 34 R2 w 144-145 cm 1.44 5310 2.19 0.03 4.19 0.06 -15.52 0.22 -21 -2.0
107 634 A 34 R3 w 2-3 cm 0.02 5310 2.27 0.02 4.34 0.04 -13.47 0.17 -24 -1.0
 107 634 A 34 R3 w 27-28 cm 0.27 5310 2.23 0.03 4.27 0.05 -14.99 0.16 -23 -2.0
107 654 A 34R3w 44-48 cm 0.44 5310 2.20 0.06 4.17 0.10 -16.56 0.26 -10 -2.9
 107 634 A 34 R3 w 62-63 cm 0.62 5310 2.02 0.03 3.86 0.04 -16.19 0.21 -18 0.0
107 654A 34R3 w 90-91 0.9 5310 2.06 0.05 3.95 0.08 -18.31 0.57 -22 -2.8
 107 634 A 34 R3 w 120-123 cm 1.2 5310 2.18 0.03 4.16 0.05 -15.83 0.18 -21 -2.0
107 654 A 34R3w 146-147 cm 1.46 5310 2.29 0.03 4.36 0.05 -15.14 0.19 -14 -3.0
 107-654-A35-R1W-58-59 cm 0.58 5310 1.06 0.03 2.01 0.05 -33.59 0.32 -5 -3.0
GYPSUM HYDRATION WATER MOTHER WATER
258 Tables for Chapters 5, 6 and 7
Table D.3. Isotopic composition of measured gypsum hydration water and calculated mother water values
(d-excess and 17O-excess) from samples obtained from the Cabo de Gata Salinas, SE Spain.
Sample Name Height in cycle (m) Age !17O σ !18O σ !D σ 17O-Excess d-excess
Modern Marine Gypsum
#1SALINA 10.56 0.17 12.94 0.35 -24.4
#2SALINA 10.25 0.13 9.57 0.45 -25.4
#3SALINA 11.37 0.07 15.94 0.30 -27.9
#4SALINA 11.60 0.10 18.62 0.24 -26.9
#5SALINA 10.86 0.04 16.69 0.31 -23.0
Salina 1 10.49 0.02 12.39 0.24 -24.5
Salina 4 2015 12.00 0.02 19.71 0.08 -29.0
Salina gypsum on water wheel 11.81 0.09 18.97 0.47 -28.2
DEPO2 beside wheel 11.61 0.02 17.73 0.09 -27.9
DEPO3 11.25 0.04 15.87 0.12 -27.0
DEPO10 12.05 0.01 19.46 0.12 -29.7
Modern marine gypsum (IRMS)
Salina5-1 5.75 0.07 11.01 0.12 -59
Salina 5-3 5.80 0.36 11.09 0.69 -51
Salina 5-4 6.08 0.39 11.66 0.74 -70
Salina 5-5 5.90 0.12 11.31 0.25 -66
Depot1 6.27 0.10 12.03 0.20 -74
desert roses_2_Depot 1 5.89 0.07 11.30 0.12 -71
Depot_3 6.00 0.13 11.53 0.24 -82
lost pond-6 5.50 0.01 10.56 0.01 -72
Calentador2 6.12 0.09 11.72 0.17 -62
Noria2015 6.05 0.03 11.61 0.04 -74
BRINES CABO DE GATA SALINAS
D5-26/03 1.04 0.06 1.99 0.12 6.06 0.44 -6 -9.9
C-23/07 0.62 0.05 1.17 0.07 7.98 0.42 8 -1.3
D1-23/07 1.24 0.04 2.35 0.04 10.61 0.44 14 -9.2
D3-23/07 3.57 0.03 6.83 0.07 31.59 0.54 -31 -23.1
D4-23/07 4.40 0.01 8.45 0.02 36.50 0.29 -58 -31.1
D5-23/07 4.12 0.95 7.94 1.82 30.43 8.47 -60 -33.1
D6-23/7 5.10 0.05 9.81 0.07 34.17 0.37 -64 -44.3
D1-10/09 1.28 0.04 2.46 0.06 13.86 0.47 -15 -5.8
D5-10/09 1.83 0.04 3.51 0.06 7.62 0.29 -20 -20.4
D6-10/9 2.16 0.05 4.15 0.09 12.81 0.31 -30 -20.4
GYPSUM HYDRATION WATER MOTHER WATER
259
Table D.4a. Results of strontium, calcium and barium isotope analysis on PLG units, SE Spain.
Sample Name Height in cycle (m) Age (ka) 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ δ138/134Ba 
Cycle 1
CdA C1 10 cm 0.1 5969 0.708974 10 0.34 0.03 0.00
CdA C1 40 cm 0.4 5969 0.708975 8 0.40 0.03
CdA C1 70 cm 0.7 5969 0.708971 6 0.45 0.03
Cycle 2
Q1 0.4 5951 0.708974 10 0.39 0.02
Q2 0.8 5951 0.708968 10
Q3 1.29 5951
Q4 1.64 5951 0.708961 10
Q5 2.32 5951
Q6 3.15 5951
Q7 4.41 5951
Q8 5.1 5951
Q9 5.3 5951 0.708972 10 0.44 0.03
Q10 5.84 5951 0.76 0.02
Q11 6.65 5951
Q12 7.59 5951
Q13 8.64 5951 0.708973 6 0.54 0.03
Q14 9.57 5951
Q15 10.32 5951
Q16 11.78 5951 0.708979 6 0.64 0.03 0.02
Q17 12.63 5951
Q18 13.51 5951 0.35 0.03
Q19 14.7 5951
Q20 15.24 5951
Q21 15.54 5951 0.70897 8
Q22 16.81 5951 0.67 0.03
Q23 17.87 5951 0.708953 7 0.38 0.02 0.14
Cycle 3
G.6.1 18.2 5931 0.708965 6 0.86 0.03 0.08
Cycle 4
G.5.3 0 5910 0.708955 9 0.30 0.03 0.06
G.5.2 6 5910 0.80 0.03
G.5.1 15.5 5910 0.708935 9 0.78 0.03 0.18
Cycle 5
G.4.2 0 5887 0.70894 6 0.42 0.03
G.4.1 10.7 5887 0.708957 7 0.71 0.12
Cycle 6
G.3.2 0 5867 0.708933 7 0.36
G.3.1 9.1 5867 0.708958 6 0.56 0.02
Cycle 7
G.2.1
BOSQUE 00 8.82 5837 0.708964 7 0.77 0.02
BOSQUE 01 8.4 5837 0.708959 7 0.85 0.03 0.14
BOSQUE 02 7.98 5837 0.708971 6 0.77 0.02
BOSQUE 03 7.56 5837 0.708957 5 0.82 0.03
BOSQUE 04 7.14 5837 0.708959 6 0.90 0.03
BOSQUE 05 6.72 5837 0.708963 5 0.86 0.02 0.11
BOSQUE 06 6.3 5837 0.708961 6 0.86 0.03
BOSQUE 07 5.88 5837 0.708957 6 0.79 0.02
BOSQUE 08 5.46 5837 0.708956 5 0.81 0.03
BOSQUE 09 5.04 5837 0.70896 6 0.72 0.02
BOSQUE 10 4.62 5837 0.70896 6 0.83 0.03 0.14
BOSQUE 11 4.2 5837 0.708962 6 0.95 0.02
BOSQUE 12 3.78 5837 0.708961 6 0.95 0.02
BOSQUE 13 3.36 5837 0.708955 6 0.91 0.02
BOSQUE 14 2.94 5837 0.70896 6 0.87 0.03
BOSQUE 15 2.52 5837 0.708958 6 0.79 0.02 0.09
BOSQUE 16 2.1 5837 0.708951 7 0.92 0.03
BOSQUE 17 1.68 5837 0.708954 7 0.84 0.02
BOSQUE 18 1.26 5837 0.708955 6 0.81 0.03
BOSQUE 19 0.84 5837 0.708952 6 0.80 0.02
BOSQUE 20 0.42 5837 0.708963 7 0.89 0.03 0.14
BOSQUE 21 0.2 5837 0.708951 6 0.66 0.03
G.2.2 0 5837
Cycle 8
G.1.1 5.55 5816 0.70896 6 0.82 0.03
G.1.2 0 5816 0.708945 6 0.67 0.03
260 Tables for Chapters 5, 6 and 7
Table D.4b. Table D.4a continued. Results of strontium, calcium and barium isotope analysis on PLG units,
SE Spain.
Sample Name Height in cycle (m) Age (ka) 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ δ138/134Ba 
Cycle 10
RdA -35 cm nuc cone -0.35 5774 0.708926 10 0.53 0.02
0cm 5cm 0 5774
10cm 11cm 12cm 0.1 5774 0.75 0.03
50 cm 0.5 5774
56cm 57cm 0.55 5774 0.708938 10 0.70 0.03
87cm 91cm 0.87 5774 0.708934 10 0.71 0.03
102cm 105cm 1.02 5774 0.708938 9 0.71 0.03
107cm 1.07 5774
117cm 1.17 5774 0.708933 10
137cm 1.37 5774
153cm 1.53 5774 0.70894 10 0.70 0.03
161cm 1.61 5774 0.708942 9 0.73 0.03
192cm 1.92 5774 0.85 0.03
231cm 2.31 5774 0.56 0.03
272cm 2.72 5774
306cm 3.06 5774 0.708945 6 0.74 0.02
SC50A&B 4.5 5774 0.70896 5 0.88 0.03
SC100A 5 5774 0.708955 14 0.85 0.03
SCTOP 6 5774 0.708954 10 0.81 0.03 0.05
Cycle 11
C7-01 4.0 5754 0.708947 5 0.67 0.02
C7-02 4.0 5754
C7-04 4.0 5754
C7-05 3.0 5754 0.708942 5
C7-06 1.0 5754
C7-07 0.0 5754 0.708916 6 0.63 0.02
Cycle 12
BdI C15 -5 cm -0.05 5721 0.708962 9 0.91 0.03 0.23
BdI C15 10 cm 0.1 5721
BdI C15 46 cm 0.46 5721 0.708966 8 1.16 0.03
BdI C15 130 cm 1.3 5721
BdI C15 200 cm 2 5721 0.70895 10 1.00 0.02
BdI C15 400 cm 4 5721 0.708971 10 0.87 0.02
BdI C15 450 cm 4.5 5721 0.708966 10 1.10 0.03 0.07
BdI C15 480 cm 4.8 5721 0.708972 11 1.02 0.03
BdI C15 520 cm 5.2 5721 0.708954 9 0.94 0.03
BdI C15 550 cm 5.5 5721
BdI C15 600 cm 6 5721 0.708959 9 1.06 0.03 0.01
Cycle 13
BdI C16 0m 0 5702 0.70896 6 0.58 0.02
BdI C16 10m 0.1 5702
BdI C16 25m 0.25 5702
BdI C16 35m 0.35 5702
BdI C16 55m 0.55 5702
BdI C16 75 cm 0.75 5702
BdI C16 95 cm 0.95 5702 0.708978 7 0.62 0.02
BdI C16 120 cm 1.2 5702
BdI C16 140 cm 1.4 5702
BdI C16 160 cm 1.6 5702
BdI C16 240 cm 2.4 5702 0.708949 6 0.71 0.02
Cycle 15
RdA C16 -35 cm -0.35 5663 0.708944 7 0.48 0.03
RdA  C16 5 cm 0.05 5663
RdA C16 45 cm 0.45 5663 0.78 0.03
RdA C16 85 cm 0.85 5663 0.708944 7 0.66 0.03
RdA C16 200 cm 2 5663 0.708952 10 0.71 0.03
RdA C16 270 cm 2.7 5663 0.70896 9 0.77 0.03
RdA C16 320 cm 3.2 5663 0.708958 10 0.80 0.03
Nijar Basin
GY-NIJAR-01 0.708987 10
GY-NIJAR-02
GY-NIJAR-03
GY-NIJAR-04
GY-NIJAR-05
GY-NIJAR-06
GY-NIJAR-07
261
Table D.4c. Table D.4b continued. Results of strontium, calcium and barium isotope analysis on PLG units,
Sicily.
Sample Name Height in cycle (m) Age (ka) 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ δ138/134Ba 
 Resedimented Lower Gypsum, Sicily 
EM MS massive selenite A 5579 0.708953 10
EM MS massive selenite B 5579 0.708982 10
EM lower gypsum banded in place 2B 5579 0.70895 10
EM lower banded gypsum hard bit location 2A 5579 0.708949 11
 Primary Lower Gypsum, Sicily 
Cycle 3
MG-C3 22m 22 5929
MG-C3 28m 28 5929
MG-C3 44m recrys? 44 5929
Cycle 4
MG-C4 BASE 0 5910
MG-C4-3m 3 5910 0.55 0.03
MG-C4 6m 6 5910
MG-C4 8m 8 5910
MG-C4-15m banded 15 5910
MG-C4 15m 15 5910 0.74 0.04
MG-C4-19m 19 5910
MG-C4 24m (rep) 24 5910 0.85 0.03
 MG-C4-24m 24 5910
 MG-C4-28m 28 5910
Cycle 5
MG-C5 1m selen espagetti 1 5887
Cycle 6
MG-C6 1m selenite 10cm 1 5867
262 Tables for Chapters 5, 6 and 7
Table D.5a. Results of strontium, calcium and barium isotope analysis on UG units, Sicily.
Sample Name Height in cycle (m) Age (ka) 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ δ138/134Ba 
Eraclea Minoa
Cycle 2
EM2-6 5511 0.708785 14
EM2-2 5511 0.708780 14
Cycle 3
EM3-12 5482 0.708778 14
EM3-7 5482 0.708782 14
EM3 0 cm 0 5482
EM3 -30 c -0.3 5482
EM3 -60 cm -0.6 5482
EM3 -150 cm -1.5 5482
EM3 -200 cm -2 5482
EM3 -205 cm -2.05 5482
EM3 -220 cm -2.2 5482
EM3 -230 -2.3 5482
EM3 -260 -2.6 5482
EM3 -290 cm -2.9 5482
EM3 -320 cm -3.2 5482
EM3 -360 cm -3.6 5482
EM3 -400 -4 5482
Cycle 4
EM4-20 5459 0.708774 14
EM4-13 5459 0.708788 14
EM 4 dome top 9 5459 0.708762 11
EM 4 dome 0-26cm 8.76 5459
EM 4 dome 30 cm 8.7 5459 0.708766 10 0.03 0.03 0.01
EM 4 dome 39-60 cm 8.4 5459
EM4 dome 60-81 8.29 5459 0.708777 11 0.54 0.03 0.00
EM 4 dome 81-98 cm 8.02 5459 0.708795 10
EM4 dome-119-146cm 7.54 5459
EM4 dome 300 cm 6 5459 0.708775 10
EM4 palisade A 5.25 5459 0.708779 10
EM4 palisade B 5 5459
EM4 palisade C 4.75 5459 0.708777 8 0.35 0.03 -0.06
EM4 palisade D 4.5 5459 0.708769 10
EM4 palisade E 4.25 5459 0.708788 10 0.43 0.03 -0.01
EM4 palisade F 4 5459 0.708769 10
EM4 palisade G 3.75 5459 0.708782 10
EM4 palisade H 3.5 5459
EM4 palisade I 3.25 5459 0.708779 10
EM4 palisade K 3 5459 0.708779 10
EM4 UGF 4B 1-01-20 cm 2.92 5459 0.708779 10 0.25 0.03 0.09
EM laminated gypsum 40 cm below unit K 2.6 5459
EM4 UGF 4-01-130 cm 2.48 5459
EM4 UGF 4-02-200cm 2.2 5459
EM4 3A-02-190 cm 0.94 5459 0.708775 10 0.14 0.02
EM4 Gypsorudite 0 5459
EM4 veins 60 cm 5459
 EM4 UGF 3a VEINS 120 cm 5459
Cycle 5
EM5-31 5436 0.708756 14
EM5-22 5436 0.708752 14
EM5 A 4.5 5436
EM5 B 4 5436
 EM5 C 3.5 5436
 EM5 D 3 5436
EM5 E 2.5 5436
EM5 F 2 5436
 EM5 G 1.5 5436
 EM5 H 1 5436
EM5 I 0.5 5436
EM5 BASE 0 5436
Cycle 6
EM6-39 5416 0.708729 14
EM6-33 5416 0.708724 14
Cycle 7
EM7-44 5343 0.708737 14
EM7-43 5343
EM7-42 5343
EM7-41 5343 0.708734 14
EM Aren -21 to -17 cm -0.17 5343
EM AREN  upper 0-6 cm 0.06 5343 0.708719 10
EM AREN 76-83 cm 0.83 5343
EM Aren 112-122 cm 1.12 5343
EM Aren 143-148 cm 1.48 5343
EM Aren 191-211 cm 2.11 5343 0.708730 10
263
Table D.5b. Table D.5a continued. Results of strontium, calcium and barium isotope analysis on beep basin
units.
Sample Name Height in cycle (m) Age (ka) 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ δ138/134Ba 
Leg 107 Site 654
107 654 A 27 R1w 6-9 cm 0.06 5310
107 654A 27R1 w 12-15 0.12 5310
 107-654 A 27-R1w 19-23cm 0.19 5310
 107-654 A 27-R1w 20cm 0.2 5310
 107-651 A 27-R1w 26cm 0.26 5310
107 654A 27 R1 w 30-33cm 0.3 5310 0.708772 8 -0.10
107 654 A 27 R1w 34-36 cm 0.34 5310
107 654A 27R1 w 43-45 0.43 5310
 107-654 A 27R CC W 10cm 1 5310
 107-654 A 28-R1w 8cm 0.08 5310 0.70874 7 -0.33
 107-654-A28-R1W-25-28 cm 0.25 5310
 107-654-A28-R1W-33-34 cm 0.33 5310
 107-654-A28-R1W-51-54 cm 0.51 5310
107 654 A 29 R2 21-22 cm 0.21 5310 -0.39
107 654 30 R1 w 75-76 cm 0.75 5310 0.708789 8 -0.21
107 654 A 30 R1 121-123 cm 1.21 5310
107 654 A 30 R1 135-137 cm 1.35 5310
107 654 A 30 R1 135-137 cm 1.35 5310
107 654 31 R1 w 2-3 cm 0.02 5310 -0.10
107 654A 31R1 w 105-107 cm 1.05 5310
107 654A 31R1 w 89-90 cm 0.89 5310
107 654A 31R1 w 127-129 cm 1.27 5310
107 654A 31R1 w 134-135 cm 1.34 5310 0.708784 8 -0.03
107 654 A 32 R1   3-5 cm 0.03 5310
107 654 32 R1 w 107-110 cm 1.07 5310
107 654 32 R1 w 144-145 cm 1.44 5310 -0.17
107 654 32 R1 w 27-28.5 cm 0.27 5310
107 654 32 R2 w 9-10 cm 0.09 5310
107 654 32 R2 w 9-10 cm 0.09 5310
107 654 32 R2 w 41-43 cm 0.41 5310
107 654 3 2R2 w 144-145 cm 1.44 5310
107 634 A 33 R1 w 72-74 cm 0.72 5310 0.708802 8
107 634 A 33 R1 w 63-66 cm 0.63 5310 0.03
107 654 34 R1 w 30-37 cm 0.3 5310
 107 634 A 34 R1 w 35-37 cm 0.35 5310
107 654 34R1 w 51-53 cm 0.51 5310 0.19
107 654 A 34R1w 99-106 cm 0.99 5310
107 654 34R1 w 101-102.5 cm 1.01 5310
107 654 A 34 R1 119-122 cm 1.19 5310
107 654 34 R2 w 2-3 cm 0.02 5310
 107 634 A 34 R2 w 18-20 cm 0.18 5310
107 654 A 34R2w 32-33 cm 0.32 5310 0.708804 9 0.21
107 654 A 34R2w 75-76 cm 0.75 5310
107 634 A 34 R2 w 90-94 cm 0.9 5310
 107 634 A 34 R2 w 144-145 cm 1.44 5310
107 634 A 34 R3 w 2-3 cm 0.02 5310
 107 634 A 34 R3 w 27-28 cm 0.27 5310
107 654 A 34R3w 44-48 cm 0.44 5310
 107 634 A 34 R3 w 62-63 cm 0.62 5310 0.70879 8 0.09
107 654A 34R3 w 90-91 0.9 5310
 107 634 A 34 R3 w 120-123 cm 1.2 5310
107 654 A 34R3w 146-147 cm 1.46 5310
 107-654-A35-R1W-58-59 cm 0.58 5310
Modern Marine Gypsum
#1SALINA 0.92 0.03
264 Tables for Chapters 5, 6 and 7
Table D.6a. Fluid inclusion salinity estimates, Sr/Ca and sulfur isotope results obtained from PLG units,
SE Spain.
Sample Name Height in cycle (m) Age (ka) δ34S δ18OSO4 Sr/Ca Tmice wt % NaCl ppt σ
Cycle 1
CdA C1 10 cm 0.1 5969 1.35 -3.05 5.00 61 20
CdA C1 40 cm 0.4 5969 22.5 12.1 1.33
CdA C1 70 cm 0.7 5969 1.51
Cycle 2
Q1 0.4 5951 22.1 14.9 1.64 -3.60 5.85 73 12
Q2 0.8 5951 22.9 13.9 1.53
Q3 1.29 5951 1.50 -4.24 6.78 92 21
Q4 1.64 5951 1.75
Q5 2.32 5951 1.44 -3.25 5.33 64 10
Q6 3.15 5951 1.44
Q7 4.41 5951 1.56 -3.74 6.07 77 7
Q8 5.1 5951 1.47
Q9 5.3 5951 22.1 13.4 1.75 -4.33 6.91 94 19
Q10 5.84 5951 1.59
Q11 6.65 5951 1.50 -2.36 3.96 45 10
Q12 7.59 5951 22.5 13.1 1.60
Q13 8.64 5951 1.51
Q14 9.57 5951 1.42
Q15 10.32 5951 1.82
Q16 11.78 5951 1.63
Q17 12.63 5951 1.59 -2.86 4.71 57 24
Q18 13.51 5951 1.93
Q19 14.7 5951 1.67 -3.73 6.04 77 17
Q20 15.24 5951 1.74
Q21 15.54 5951 22.2 13.2 1.85
Q22 16.81 5951 22.8 12.0 1.55
Q23 17.87 5951 1.20 -2.07 3.49 39 10
Cycle 3
G.6.1 18.2 5931 22.4 14.0 1.47
Cycle 4
G.5.3 0 5910 22.6 14.0 1.31
G.5.2 6 5910 22.2 11.0 1.89
G.5.1 15.5 5910 21.9 13.1 1.51
Cycle 5
G.4.2 0 5887 22.6 12.8 1.36
G.4.1 10.7 5887 22.2 12.5 2.47
Cycle 6
G.3.2 0 5867 22.6 13.1 1.42
G.3.1 9.1 5867 22.1 13.3 1.74
Cycle 7
G.2.1 22.2 13.0 2.52
BOSQUE 00 8.82 5837 2.29
BOSQUE 01 8.4 5837 21.9 12.5 1.97 -3.94 6.34 84 23
BOSQUE 02 7.98 5837 1.60
BOSQUE 03 7.56 5837 1.91
BOSQUE 04 7.14 5837 1.86
BOSQUE 05 6.72 5837 1.94
BOSQUE 06 6.3 5837 22.6 12.5 2.23
BOSQUE 07 5.88 5837 13.0 1.60
BOSQUE 08 5.46 5837 21.9 11.8 1.52
BOSQUE 09 5.04 5837 14.1 1.36
BOSQUE 10 4.62 5837 1.61 -3.94 6.34 84 21
BOSQUE 11 4.2 5837 1.54
BOSQUE 12 3.78 5837 1.49
BOSQUE 13 3.36 5837 1.52
BOSQUE 14 2.94 5837 1.47
BOSQUE 15 2.52 5837 1.42
BOSQUE 16 2.1 5837 1.68
BOSQUE 17 1.68 5837 1.49
BOSQUE 18 1.26 5837 1.42 -3.69 5.98 76 17
BOSQUE 19 0.84 5837 22.5 14.5 1.47
BOSQUE 20 0.42 5837 1.55
BOSQUE 21 0.2 5837 22.3 12.5 1.31 -2.31 3.85 45 21
G.2.2 0 5837 22.6 14.1
Cycle 8
G.1.1 5.55 5816 22.4 12.3 1.48
G.1.2 0 5816 22.4 12.8 1.72
265
Table D.6b. Table D.6a continued. Fluid inclusion salinity estimates, Sr/Ca and sulfur isotope results
obtained from PLG units, SE Spain.
Sample Name Height in cycle (m) Age (ka) δ34S δ18OSO4 Sr/Ca Tmice wt % NaCl ppt σ
Cycle 10
RdA -35 cm nuc cone -0.35 5774 22.5 13.0 1.11 -1.71 2.87 32 19
0cm 5cm 0 5774 23.3 14.5
10cm 11cm 12cm 0.1 5774 22.7 12.9 1.46 -2.44 4.06 47 20
50 cm 0.5 5774 22.4 13.3 1.73
56cm 57cm 0.55 5774 22.3 12.6 1.72 -3.58 5.76 77 36
87cm 91cm 0.87 5774 22.6 12.4 1.91 -3.31 5.36 69 33
102cm 105cm 1.02 5774 22.5 12.4 1.82 -2.94 4.83 59 22
107cm 1.07 5774 22.7 13.2
117cm 1.17 5774 22.5 11.5
137cm 1.37 5774 22.6 13.2
153cm 1.53 5774 1.72
161cm 1.61 5774 22.5 11.9 2.17 -2.68 4.42 53 21
192cm 1.92 5774 1.58 -2.46 4.08 48 24
231cm 2.31 5774 22.7 13.6 1.23
272cm 2.72 5774 1.57
306cm 3.06 5774 1.70 -2.83 4.58 60 40
SC50A&B 4.5 5774 22.3 11.3 1.89 -2.77 4.54 57 36
SC100A 5 5774 1.93 -2.30 3.81 45 26
SCTOP 6 5774 1.55
Cycle 11
C7-01 4.0 5754 1.54
C7-02 4.0 5754 1.51
C7-04 4.0 5754 1.58
C7-05 3.0 5754 1.36
C7-06 1.0 5754 1.33
C7-07 0.0 5754 1.45
Cycle 12
BdI C15 -5 cm -0.05 5721 1.584 -3.66 5.93 75 17
BdI C15 10 cm 0.1 5721 22.3 10.4 2.092
BdI C15 46 cm 0.46 5721 1.874
BdI C15 130 cm 1.3 5721 22.4 12.8 3.335
BdI C15 200 cm 2 5721 22.1 13.0 2.158 -4.54 7.21 101 16
BdI C15 400 cm 4 5721 2.529
BdI C15 450 cm 4.5 5721 2.875
BdI C15 480 cm 4.8 5721 2.650
BdI C15 520 cm 5.2 5721 2.874 -5.46 8.49 134 19
BdI C15 550 cm 5.5 5721 21.2 13.1 2.144
BdI C15 600 cm 6 5721 2.261
Cycle 13
BdI C16 0m 0 5702 21.1 13.4 1.27
BdI C16 10m 0.1 5702 1.39
BdI C16 25m 0.25 5702 1.43
BdI C16 35m 0.35 5702 1.54
BdI C16 55m 0.55 5702 1.59
BdI C16 75 cm 0.75 5702 1.59
BdI C16 95 cm 0.95 5702 22.7 12.5 1.63
BdI C16 120 cm 1.2 5702 1.64
BdI C16 140 cm 1.4 5702 22.5 12.1 2.28
BdI C16 160 cm 1.6 5702 1.72
BdI C16 240 cm 2.4 5702 22.2 10.9 2.01
Cycle 15
RdA C16 -35 cm -0.35 5663 22.7 14.6 1.19
RdA  C16 5 cm 0.05 5663 1.40
RdA C16 45 cm 0.45 5663 0.68
RdA C16 85 cm 0.85 5663 22.6 14.0 1.27
RdA C16 200 cm 2 5663 22.5 13.3 2.01
RdA C16 270 cm 2.7 5663 22.2 13.5 2.02
RdA C16 320 cm 3.2 5663 22.4 13.9 2.20
Nijar Basin
GY-NIJAR-01 8.7 10.2 0.48
GY-NIJAR-02 21.9 10.6 2.95
GY-NIJAR-03 21.8 10.6
GY-NIJAR-04 21.2 15.1 1.11
GY-NIJAR-05 6.7 9.0 0.45
GY-NIJAR-06 -18.3 7.1
GY-NIJAR-07 20.5 10.8
266 Tables for Chapters 5, 6 and 7
Table D.6c. Table D.6b continued. Fluid inclusion salinity estimates, Sr/Ca and sulfur isotope results
obtained from PLG units, SE Spain.
Sample Name Height in cycle (m) Age (ka) δ34S δ18OSO4 Sr/Ca Tmice wt % NaCl ppt σ
 Resedimented Lower Gypsum, Sicily 
EM MS massive selenite A 5579 22.5 12.9
EM MS massive selenite B 5579 22.7 12.3
EM lower gypsum banded in place 2B 5579 21.6 12.6
EM lower banded gypsum hard bit location 2A 5579 22.4 13.3
 Primary Lower Gypsum, Sicily 
Cycle 3
MG-C3 22m 22 5929 -4.31 6.87 95 29
MG-C3 28m 28 5929 -4.26 6.79 93 28
MG-C3 44m recrys? 44 5929
Cycle 4
MG-C4 BASE 0 5910
MG-C4-3m 3 5910 22.8 13.0 1.47 -3.80 6.13 80 21
MG-C4 6m 6 5910 22.9 13.4
MG-C4 8m 8 5910
MG-C4-15m banded 15 5910 22.7 13.2
MG-C4 15m 15 5910 2.13 -4.63 7.34 103 13
MG-C4-19m 19 5910 22.8 14.0
MG-C4 24m (rep) 24 5910 1.92
 MG-C4-24m 24 5910 22.8 13.0
 MG-C4-28m 28 5910 22.6 12.7
Cycle 5
MG-C5 1m selen espagetti 1 5887 22.9 14.6
Cycle 6
MG-C6 1m selenite 10cm 1 5867
Monte Banco Cycles 1 to 6
MBC1 bottom good crystal 22.4
MBC1 middle drilled pieces 22.6
MB-C2 middle drilled pieces 23.1
MBC3 Selenite surface
MBC3 Selenite surface (rep)
MBC4 middle bulk 22.9
MBC4 middle bulk (rep) 23.2
MBC6 30 cm drilled pieces
MBC6 30 cm drilled pieces (rep)
267
Table D.7a. Fluid inclusion salinity estimates, Sr/Ca and sulfur isotope results obtained from UG units,
Sicily.
Sample Name Height in cycle (m) Age (ka) δ34S δ18OSO4 Sr/Ca Tmice wt % NaCl ppt σ
Eraclea Minoa
Cycle 2
EM2-6 5511
EM2-2 5511
Cycle 3
EM3-12 5482
EM3-7 5482
EM3 0 cm 0 5482
EM3 -30 c -0.3 5482
EM3 -60 cm -0.6 5482
EM3 -150 cm -1.5 5482 22.1 18.2
EM3 -200 cm -2 5482
EM3 -205 cm -2.05 5482
EM3 -220 cm -2.2 5482
EM3 -230 -2.3 5482 22.6 18.8
EM3 -260 -2.6 5482
EM3 -290 cm -2.9 5482
EM3 -320 cm -3.2 5482
EM3 -360 cm -3.6 5482 22.4 18.6
EM3 -400 -4 5482
Cycle 4
EM4-20 5459
EM4-13 5459
EM 4 dome top 9 5459 21.6 14.5 -3.74 6.07 77 3
EM 4 dome 0-26cm 8.76 5459
EM 4 dome 30 cm 8.7 5459 1.49
EM 4 dome 39-60 cm 8.4 5459 20.9 14.9
EM4 dome 60-81 8.29 5459 2.93 -5.23 8.15 126 30
EM 4 dome 81-98 cm 8.02 5459 20.8 13.7
EM4 dome-119-146cm 7.54 5459
EM4 dome 300 cm 6 5459 20.9 13.9 -5.88 8.99 158 63
EM4 palisade A 5.25 5459 21.2 13.2
EM4 palisade B 5 5459 21.9 16.5
EM4 palisade C 4.75 5459 21.3 13.9 1.87 -4.00 6.44 84 11
EM4 palisade D 4.5 5459
EM4 palisade E 4.25 5459 21.3 13.9 2.02 -4.50 7.16 99 16
EM4 palisade F 4 5459
EM4 palisade G 3.75 5459 21.4 13.6
EM4 palisade H 3.5 5459 -3.52 5.74 71 9
EM4 palisade I 3.25 5459
EM4 palisade K 3 5459 21.7 13.7
EM4 UGF 4B 1-01-20 cm 2.92 5459 22.0 16.4 1.48
EM laminated gypsum 40 cm below unit K 2.6 5459 21.4 14.9
EM4 UGF 4-01-130 cm 2.48 5459 22.0 16.5
EM4 UGF 4-02-200cm 2.2 5459 22.1 17.1
EM4 3A-02-190 cm 0.94 5459 21.4 17.9 1.31
EM4 Gypsorudite 0 5459 22.3 17.5
EM4 veins 60 cm 5459 22.2 17.9
 EM4 UGF 3a VEINS 120 cm 5459
Cycle 5
EM5-31 5436
EM5-22 5436
EM5 A 4.5 5436
EM5 B 4 5436
 EM5 C 3.5 5436
 EM5 D 3 5436
EM5 E 2.5 5436
EM5 F 2 5436 22.0 16.3
 EM5 G 1.5 5436
 EM5 H 1 5436 22.1 16.9
EM5 I 0.5 5436
EM5 BASE 0 5436
Cycle 6
EM6-39 5416
EM6-33 5416
Cycle 7
EM7-44 5343
EM7-43 5343
EM7-42 5343
EM7-41 5343
EM Aren -21 to -17 cm -0.17 5343
EM AREN  upper 0-6 cm 0.06 5343 21.1 14.8
EM AREN 76-83 cm 0.83 5343
EM Aren 112-122 cm 1.12 5343
EM Aren 143-148 cm 1.48 5343 21.2
EM Aren 191-211 cm 2.11 5343 21.0 17.1
268 Tables for Chapters 5, 6 and 7
Table D.7b. Table D.7a continued. Sulfur isotope results obtained from deep basin units.
Sample Name Height in cycle (m) Age (ka) δ34S δ18OSO4 Sr/Ca Tmice wt % NaCl ppt σ
Leg 42A Site 372
42 372Z8RW 30.5-33.5cm 0.305 5325
42 372Z8RW 30.5-33.5cm 0.305 5325 21.4 16.3
42 372Z8R1W 46-48.5cm 0.46 5325
42 372Z8R1W 46-48.5cm 0.46 5325 21.0 15.8
42 372Z8RW 91-94cm 0.91 5325
43 372Z8RW 91-94cm 0.91 5325 19.5 16.6
42 372 Z8 R1 W 96-99cm 0.96 5325
42 372 Z8 R1 W 96-99cm 0.96 5325 21.1 15.5
 42 372 Z8 R2 W 98-100 cm 0.98 5325
 42 372 Z8 R2 W 98-100 cm 0.98 5325 21.8 17.9
42 372Z8R1 W 126-128.5 1.26 5325
42 372Z8R1 W 126-128.5 1.26 5325 21.1 15.3
42 372 Z8 R2 W 146-149 cm 1.46 5325
42 372 Z8 R2 W 146-149 cm 1.46 5325 21.9 18.2
Leg 42A Site 374
      42 374 Z11 R2 58.0-59cm 0.58 5315 20.6 12.9
42 374Z12R1W 125-127cm 1.25 5315 11.8
42 374Z12R2W 127-130 cm 1.27 5315 20.2 12.3
42 374 Z12 R2 W 112-114 cm 1.12 5315
 42 374 Z16 R1 W 89-90cm 0.89 5315 22.0 15.7
42 374Z16R1W 130-135cm 1.3 5315 23.4 15.4
42 374Z16R1W 136-135cm 1.36 5315
 42 374Z17R1W 37-39cm 0.37 5315
42 374Z17R1W 58-59cm 0.58 5315 22.2 11.2
42 374Z17R1W 82-86cm 0.82 5315
 42 374 Z17 R1 W 110-115 cm 1.1 5315
 42 374Z17R1W 115-119cm 1.15 5315 20.1 14.6
 42 374 Z19 R1 W 19-21 cm 0.19 5315 22.2 13.0
  42 374 Z19 R1 W 25-27cm 0.25 5315
42374Z19R1W 77-80cm 0.77 5315 22.9 15.4
42374Z19R1W 83-84cm 0.83 5315 21.1 15.4
 42 374 Z19 R1 W 103-104 cm 1.03 5315 21.1 15.9
 42 374Z19R1W 124-126cm 1.24 5315
42 374Z19R1W 136-138cm 1.36 5315
 42 374 Z20 R1 W 51-55 cm 0.51 5315 22.0 17.0
42 374 Z20 R1W 33-36cm 0.33 5315
42 374Z20R1W 77-81cm 0.77 5315
42 374Z20R1W 100-104cm 1 5315 22.3 14.7
269
Table D.7c. Table D.7b continued. Sulfur isotope results obtained from deep basin units.
Sample Name Height in cycle (m) Age (ka) δ34S δ18OSO4 Sr/Ca Tmice wt % NaCl ppt σ
Leg 107 Site 654
107 654 A 27 R1w 6-9 cm 0.06 5310
107 654A 27R1 w 12-15 0.12 5310
 107-654 A 27-R1w 19-23cm 0.19 5310
 107-654 A 27-R1w 20cm 0.2 5310
 107-651 A 27-R1w 26cm 0.26 5310 1.53
107 654A 27 R1 w 30-33cm 0.3 5310 22.4 18.1
107 654 A 27 R1w 34-36 cm 0.34 5310
107 654A 27R1 w 43-45 0.43 5310
 107-654 A 27R CC W 10cm 1 5310
 107-654 A 28-R1w 8cm 0.08 5310 1.42
 107-654-A28-R1W-25-28 cm 0.25 5310 22.4 17.7
 107-654-A28-R1W-33-34 cm 0.33 5310
 107-654-A28-R1W-51-54 cm 0.51 5310 22.1 18.3
107 654 A 29 R2 21-22 cm 0.21 5310 22.9 20.3 1.20
107 654 30 R1 w 75-76 cm 0.75 5310 22.1 18.2 1.07
107 654 A 30 R1 121-123 cm 1.21 5310
107 654 A 30 R1 135-137 cm 1.35 5310 22.1 18.3
107 654 A 30 R1 135-137 cm 1.35 5310
107 654 31 R1 w 2-3 cm 0.02 5310 2.01
107 654A 31R1 w 105-107 cm 1.05 5310
107 654A 31R1 w 89-90 cm 0.89 5310
107 654A 31R1 w 127-129 cm 1.27 5310
107 654A 31R1 w 134-135 cm 1.34 5310 22.1 18 1.28
107 654 A 32 R1   3-5 cm 0.03 5310
107 654 32 R1 w 107-110 cm 1.07 5310
107 654 32 R1 w 144-145 cm 1.44 5310 22.2 18.7 1.60
107 654 32 R1 w 27-28.5 cm 0.27 5310
107 654 32 R2 w 9-10 cm 0.09 5310
107 654 32 R2 w 9-10 cm 0.09 5310
107 654 32 R2 w 41-43 cm 0.41 5310 22.1 18.8
107 654 3 2R2 w 144-145 cm 1.44 5310
107 634 A 33 R1 w 72-74 cm 0.72 5310 21.3 17.8
107 634 A 33 R1 w 63-66 cm 0.63 5310 1.53
107 654 34 R1 w 30-37 cm 0.3 5310 22.9 18.3
 107 634 A 34 R1 w 35-37 cm 0.35 5310
107 654 34R1 w 51-53 cm 0.51 5310 1.34
107 654 A 34R1w 99-106 cm 0.99 5310
107 654 34R1 w 101-102.5 cm 1.01 5310
107 654 A 34 R1 119-122 cm 1.19 5310
107 654 34 R2 w 2-3 cm 0.02 5310
 107 634 A 34 R2 w 18-20 cm 0.18 5310
107 654 A 34R2w 32-33 cm 0.32 5310 1.53
107 654 A 34R2w 75-76 cm 0.75 5310
107 634 A 34 R2 w 90-94 cm 0.9 5310
 107 634 A 34 R2 w 144-145 cm 1.44 5310
107 634 A 34 R3 w 2-3 cm 0.02 5310
 107 634 A 34 R3 w 27-28 cm 0.27 5310
107 654 A 34R3w 44-48 cm 0.44 5310
 107 634 A 34 R3 w 62-63 cm 0.62 5310 3.57
107 654A 34R3 w 90-91 0.9 5310
 107 634 A 34 R3 w 120-123 cm 1.2 5310
107 654 A 34R3w 146-147 cm 1.46 5310
 107-654-A35-R1W-58-59 cm 0.58 5310
270 Tables for Chapters 5, 6 and 7
Table D.8a. Oxygen, carbon, strontium, calcium and XRD data of PLG carbonates
Sample Name Depth from base of gypsum (m) δ18O δ13C 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ % Calcite % Dolomite% Aragonite Cal:Dol
Base of PLG
CdA C1 -10 cm -0.1 -6.8 -3.7 100%
Sorbas Cycle 2-3
Q24 -0.2 -8.2 -2.0 100%
Q25 -0.1 -8.3 -2.3 100%
Sorbas Cycle 3-4
M5.1 -0.5 -4.2 1.6 61 0 29.8 100%
M5.2 -1.1 -4.5 1.5 65 0.2 27.3 100%
M5.3 -1.6 -6.5 0.1 0.39 0.03 91.8 0.1 1.1 100%
Sorbas Cycle 4-5
M4.1b -0.3 -7.9 -3.1 0.25 60.7 1 0 98%
Sorbas Cycle 5-6
M3.1 -0.75 -7.6 -2.0 0.15 55 0.4 5.8 99%
M3.2 -1.5 -7.2 -2.7 57.1 2.6 1 96%
Sorbas Cycle 6-7
M2.1 -0.35 -6.3 -1.4 0.70872 6 0.34 0.02 89 0.3 1.9 100%
M2.2 -0.6 -4.9 0.2 0.70877 6 0.06 0.02
M2.3 -1 -4.0 1.5 0.708802 7 -0.02 0.03 22.1 0.8 50.2 97%
M2.4 -1.2 -3.4 -1.8 16.6 0.2 8.3 99%
M2.5 -1.5 -3.7 1.6 0.708803 9 -0.10 0.03 20.6 0.3 53.6 99%
M2.6 -1.95 -5.9 -1.3 86 0.3 1.6 100%
M2.7 -2.5 -6.3 -1.2 0.708936 9 0.46 0.02 87.6 0.4 0.8 100%
Sorbas Cycle 7-8
HIGUERA MARL 0 0 -4.9 -1.3
HIGUERA MARL 5 -0.05 -5.8 -0.8
HIGUERA MARL 10 -0.1 -6.6 -1.6
HIGUERA MARL 15 -0.15 -7.0 -1.2
M1.1 -0.18 -7.3 -1.1 0.708423 6 0.26 0.02 64.1 0.8 4.1 99%
HIGUERA MARL 20 -0.2 -7.6 -1.1
HIGUERA MARL 25 -0.25 -7.6 -1.1
HIGUERA MARL 30 -0.3 -7.6 -1.0
HIGUERA MARL 35 -0.35 -7.1 -0.8
HIGUERA MARL 40 -0.4 -5.8 0.1
HIGUERA MARL 45 -0.45 -5.3 0.5
M1.2 -0.48 -5.3 -0.3 0.708645 6 0.20 0.02 34 0.1 23.3 100%
HIGUERA MARL 50 -0.5 -4.1 1.7
HIGUERA MARL 55 -0.55 -5.4 0.3
M1.3 -1.2 -6.2 -1.4 0.70855 10 0.26 0.02 29.1 0.4 5.8 99%
Sorbas Cycle 9-10
RdA M5 0 0 -2.5 0.5 0.00 0.02 0 42.6 0.0 0%
RdA M5 -2 -0.02 -1.1 0.2 10.8 20.2 0.0 35%
RdA M5 -4 -0.04 0.2 0.6 1.9 12.4 0.0 13%
RdA M5 -6 -0.06 1.3 0.9 9.5 13.1 0.0 42%
RdA M5 -8 -0.08 1.5 0.8 9.7 15.5 0.0 38%
RdA M5 -10 -0.1 1.9 1.5 9.5 14.0 0.0 40%
RdA M5 -12 -0.12 1.4 1.3 18.7 33.9 0.0 36%
RdA M5 -14 -0.14 -1.2 1.1 19.1 30.9 0.0 38%
RdA M5 -16 -0.16 -1.3 1.1 -0.02 0.02 36.4 25.8 0.0 59%
RdA M5 -18 -0.18 -1.2 0.5 39.9 25.7 0.0 61%
RdA M5 -20 -0.2 -2.1 0.4 20.6 12.2 0.0 63%
RdA M5 -22 -0.22 -3.7 -0.1 35.9 16.1 0.0 69%
RdA M5 -24 -0.24 -3.0 -0.1 40.0 10.1 0.0 80%
RdA M5 -26 -0.26 -3.1 -0.1 44.5 12.8 0.0 78%
RdA M5 -28 -0.28 -3.5 -0.1 0.708832 7 -0.13 0.02 37.8 10.7 0.0 78%
RdA M5 -30 -0.3 -3.0 0.3 40.4 7.4 0.0 84%
RdA M5 BULK 0.0 0.6 44.1 11.8 0.0 79%
Sorbas Cycle 10 SC
RdA C6 SC 130 1.3 3.3 -3.8 0.3 24.4 0.0 1%
RdA C6 SC 120 1.2 3.4 -4.0 0.5 15.7 0.0 3%
RdA C6 SC 110 1.1 3.1 -3.9 0.3 20.4 0.0 2%
RdA C6 SC 100 1 3.3 -4.3 0.1 30.0 0.0 0%
RdA C6 SC 90 0.9 2.9 -4.2 0.2 22.6 0.0 1%
RdA C6 SC 80 0.8 3.2 -4.3 0.2 33.7 0.0 1%
RdA C6 SC 70 0.7 3.0 -4.6 0.0 46.9 0.0 0%
RdA C6 SC 60 0.6 3.0 -4.8 0.0 69.3 0.0 0%
RdA C6 SC 50 0.5 3.0 -4.6 0.708986 12 0.21 0.02 0.0 42.1 0.0 0%
RdA C6 SC 40 0.4 3.1 -4.7 0.1 41.8 0.0 0%
RdA C6 SC 30 0.3 2.9 -4.4 0.1 30.5 0.0 0%
RdA C6 SC 20 0.2 3.2 -4.0 0.0 24.0 0.0 0%
RdA C6 SC 10 0.1 3.0 -3.8 0.2 31.9 0.0 0%
RdA C6 SC 50-80 0.5 3.5 -4.3 0.18 0.02 0.2 52.2 0.0 0%
RdA C6 SC 50-70 0.5 2.9 -4.3
RdA C6 SC 50-60 0.5 3.2 -4.8
RdA C6 SC 50-50 0.5 3.2 -4.8
RdA C6 SC 50-40 0.5 3.0 -4.9
RdA C6 SC 50-30 0.5 2.8 -4.9
RdA C6 SC 50-20 0.5 2.9 -4.9
RdA C6 SC 50-10 0.5 3.1 -4.8 0.709025 6 0.16 0.03 0.0 42.6 0.0 0%
271
Table D.8b. Table D.8a continued. Oxygen, carbon, strontium, calcium and XRD data of PLG and UG
carbonates
Sample Name Depth from base of gypsum (m) δ18O δ13C 87Sr/86Sr 2σ (E-6) δ44/40Ca 2σ % Calcite % Dolomite% Aragonite Cal:Dol
Sorbas Cycle 9-10
RdA M7 0 -1.8 4.5 -3.4 0 73.6 1.7 0%
RdA M7 5 -1.75 4.9 -3.3 0.4 41.6 4.3 1%
RdA M7 10 -1.7 5.0 -3.6 0.4 41.8 5.7 1%
RdA M7 15 -1.65 4.7 -3.8 0.7 54 1.2 1%
RdA M7 20 -1.6 4.2 -4.0 0.3 73.7 1.4 0%
RdA M7 25 -1.55 4.0 -3.9 0.3 36.5 3.6 1%
RdA M7 30 -1.5 4.0 -3.9 0.5 70.6 2.4 1%
RdA M7 35 -1.45 3.4 -3.9 -0.19 0.03 0.3 54.5 5.1 1%
RdA M7 40 -1.4 4.5 -3.7 0.8 39.4 4.2 2%
RdA M7 45 -1.35 3.9 -3.2 1.7 18.6 2.4 8%
RdA M7 50 -1.3 5.1 -3.4 0.3 61.1 2.4 0%
RdA M7 55 -1.25 4.3 -3.3 0.3 53.5 1.7 1%
RdA M7 57.5 -1.225 5.0 -3.3 0.4 51.5 4.1 1%
RdA M7 60 -1.2 4.0 -3.2 0.5 28.5 2.4 2%
RdA M7 65 -1.15 4.5 -3.2 0.3 40.7 3 1%
RdA M7 70 -1.1 4.3 -3.3 0.4 57.9 5.3 1%
RdA M7 75 -1.05 2.0 -3.2 16.8 44.2 3.1 28%
RdA M7 80 -1 1.2 -3.1 18.5 33.4 3.4 36%
RdA M7 85 -0.95 0.1 -2.7 0.16 0.02 22.1 18.6 2.8 54%
RdA M7 90 -0.9 5.0 -2.5 0.3 47.2 5.1 1%
RdA M7 95 -0.85 4.7 -2.1 1.6 39.1 4.2 4%
RdA M7 100 -0.8 4.9 1.2 0.7 66.6 1.5 1%
RdA M7 105 -0.75 0.8 0.9 28.7 34.7 0.5 45%
RdA M7 110 -0.7 0.2 1.1 39.6 56 0.5 41%
RdA M7 115 -0.65 3.8 2.0 7.8 64.8 0.8 11%
RdA M7 120 -0.6 3.9 1.7 0.6 33.5 1.2 2%
RdA M7 125 -0.55 4.5 1.2 0.1 83.8 1.1 0%
RdA M7 130 -0.5 3.4 0.9 0.5 26.3 1.1 2%
RdA M7 135 -0.45 5.0 1.2 0 82.3 1.2 0%
RdA M7 140 -0.4 5.3 1.3 0.708894 6 -0.4 0.02 0.1 81.2 2.1 0%
RdA M7 145 -0.35 5.3 1.2 0 71.9 2.9 0%
RdA M7 150 -0.3 4.6 0.5 0.3 16 1.6 2%
RdA M7 155 -0.25 5.8 0.4 0.1 74.6 2.9 0%
RdA M7 160 -0.2 5.4 -0.1 0.3 59.3 3 1%
RdA M7 165 -0.15 5.1 -0.1 0.2 22.8 1.4 1%
RdA M7 170 -0.1 5.2 -0.8 0.3 44.1 6.2 1%
RdA M7 175 -0.05 5.2 -0.8 0.1 59.1 5.4 0%
RdA M7 180 0 5.4 -0.9 -0.25 0.02 0 71.8 4 0%
Sorbas Cycle 10 SC
SC7-M1 4 4.0 -3.8
SC7-M2 4 3.5 -4.8
SC7-M3 4 4.0 -3.9
SC7-M4 4 5.8 -2.6
Sorbas Cycle 12-13
BdI C16 -25 -0.25 -7.3 -0.6
BdI C16 -20 -0.2 -7.2 -0.6
BdI C16 -15 -0.15 -7.2 -0.9
Sorbas Cycle 13-14
BdI C16 +100 1 -7.3 -1.0
BdI C16 +100 Bulk 1 -5.3 0.8
Sorbas Cycle 14-15
RdA C16 -5 -0.05 -5.3 -3.3
RdA C16 -10 -0.1 -5.5 -2.7
RdA C16 -15 -0.15 -5.2 -2.3
RdA C16 -23 -0.23 -5.7 -1.4
RdA C16 -25 -0.25 -5.3 -1.5
RdA C16 -30 -0.3 -4.8 -1.9
RdA C16 -42 -0.41 -5.3 -2.2
RdA C16 -60 -0.6 -5.2 -1.8
RdA C16 -65 -0.65 -5.4 -1.2
RdA C16 -75 -0.75 -5.6 0.3
RdA C16 -80 -0.8 -5.7 0.0
RdA C16 -90 -0.9 -5.5 0.6
Sorbas within Cycle 15
RdA C16 37 0.37 -5.4 -2.7
Sorbas Cycle 15+
RdA C16 +100 1 -5.5 0.6
Upper Gypsum
EM Cycle 4
EM4 dome 60cm -0.5 -1.3
EM4 dome 125 cm -1.2 -1.3
EM4 dome 148 cm -2.0 -1.6
EM4 pal G-H 0.1 -0.9
EM4 pal I-J 0.0 -0.9
EM4 pal J-K 1.1 -1.6
EM Cycle 7
EM Aren 86-87 cm 0.3 -0.3
EM Aren 148-157 cm -0.3 -0.7
272 Tables for Chapters 5, 6 and 7
Table D.9a. Oxygen and hydrogen isotope data of Mediterranean freshwaters
δ18O δD d-excess Sample Name δ18O δD 17O-ex d-excess
-5.92 -34.21 13.14 rain 6 -1.81 -7.09 23 7.43
-7.83 -55.42 7.23 rain 7 -4.18 -25.10 16 8.31
-8.17 -53.73 11.62 rain 9 -4.93 -29.17 21 10.29
-5.46 -39.20 4.47 rain 18 -4.97 -35.07 15 4.68
-4.50 -24.22 11.81 rain 20 -2.96 -16.10 16 7.62
-5.22 -34.95 6.79 rain 26 -4.40 -25.59 10 9.62
-2.22 -8.09 9.66 MOLINOS 01-03-2012 -5.86 -36.54 12 10.30
-4.40 -27.23 7.94 MOLINOS 18-03-2012 -5.77 -36.27 4 9.92
-7.22 -38.35 19.42 MOLINOS 24-03-2012 -5.06 -34.10 0 6.35
-7.76 -46.16 15.92 MOLINOS 07-04-2012 -5.80 -35.49 12 10.94
-3.49 -9.93 17.96 MOLINOS 10-08-2012 -5.99 -36.75 6 11.15
-5.27 -33.86 8.32 MOLINOS 10-08-2012 -5.40 -34.74 18 8.44
-3.04 -11.05 13.29 APAS 11-05-2011 -4.21 -27.21 11 6.48
-4.97 -31.94 7.81 APAS 10-08-2011 -5.09 -29.84 28 10.86
-4.71 -27.39 10.29 APAS 01-03-2012 -5.33 -30.59 28 12.06
-6.09 -30.48 18.27 APAS 06-06-2012 -5.26 -30.17 18 11.92
-4.61 -12.52 24.35 PERAL 11-05-2011 -4.94 -31.46 19 8.03
-11.12 -73.92 15.03 PERAL 01-03-2012 -5.33 -30.59 28 12.06
-4.51 -13.05 23.01 PERAL 06-06-2012 -5.26 -30.17 18 11.92
-6.40 -39.10 12.12 PERAL 11-05-2011 -4.94 -31.46 19 8.03
-2.50 -17.63 2.35 PERAL 10-08-11 -5.04 -32.70 20 7.64
-1.89 -7.55 7.54 PERAL 01-3-12 -5.81 -34.98 23 11.49
-4.97 -29.89 9.84 PERAL 06-06-12 -5.66 -34.48 26 10.78
-5.03 -35.81 4.40 VINICA 1-3-12 -5.18 -30.09 24 11.36
-4.26 -26.13 7.94 VINICA 7-4-12 -4.86 -28.91 18 9.97
-3.11 -17.08 7.83 RAIN-01 -5.9 -34.2 13.00
-4.29 -25.81 8.48 RAIN-02 -7.8 -55.4 7.00
-5.55 -30.03 14.35 RAIN-03 -8.2 -53.7 11.90
-5.29 -29.10 13.22 RAIN-04 -5.5 -39.2 4.80
-5.38 -29.50 13.52 RAIN-05 -4.5 -24.2 11.80
-5.23 -28.60 13.21 RAIN-06 -5.2 -35 6.60
-5.39 -30.30 12.79 RAIN-07 -2.2 -8.1 9.50
-5.36 -29.30 13.55 RAIN-08 -4.4 -27.2 8.00
-5.36 -29.99 12.89 RAIN-09 -7.2 -38.4 19.20
-5.55 -30.00 14.40 RAIN-10 -7.8 -46.2 16.20
-5.55 -30.68 13.75 RAIN-11 -3.5 -9.9 18.10
-5.39 -28.86 14.26 RAIN-12 -5.3 -33.9 8.50
-5.36 -29.16 13.71 RAIN-13 -3 -11.1 12.90
-6.20 -33.66 15.90 RAIN-14 -5 -31.9 8.10
-6.26 -35.21 14.85 RAIN-15 -4.7 -27.4 10.20
-6.23 -34.98 14.84 RAIN-16 -6.1 -30.5 18.30
-6.30 -35.55 14.87 RAIN-17 -4.6 -12.5 24.30
-6.33 -35.21 15.42 RAIN-18 -11.1 -73.9 14.90
-5.42 -30.59 12.79 RAIN-19 -1.81 -7.09 23 7.39
-6.29 -36.21 14.08 RAIN-20 -4.18 -25.1 16 8.34
-6.23 -36.84 13.01 RAIN-21 -4.93 -29.17 21 10.27
-6.31 -36.10 14.36 RAIN-22 -4.97 -35.07 15 4.69
-6.01 -35.66 12.41 RAIN-23 -2.96 -16.1 16 7.58
-5.91 -33.45 13.83 RAIN-24 -4.4 -25.59 10 9.61
-6.13 -35.05 14.01
-6.27 -36.69 13.46
-5.76 -33.69 12.38
-5.77 -33.05 13.12
-5.66 -31.56 13.73
-5.57 -31.95 12.64
-5.64 -32.21 12.93
-5.72 -31.06 14.67
-4.4 -26 9.20
-5.2 -23 18.60
Rainwaters, Sicily Groundwaters, Sorbas Basin, SE Spain
273
Table D.9b. Table D.9a continued. Oxygen and hydrogen isotope data of Mediterranean freshwaters
δ18O δD d-excess δ18O δD d-excess δ18O δD δ18O δD
-5.29 -32.5 9.82 -10.8 -71 15.40 -7.83 -66.4 -10.02
-6.32 -36.6 13.96 -4.1 -24 8.80 -4.76 -36 -12.49 -90.7
-7.62 -51.2 9.76 -3.8 -27 3.40 -8.45 -71.3 -11.85 -84.6
-8.99 -56.9 15.02 -5 -24 16.00 -3.25 -31.6 -8.76 -59.3
-8.94 -52.5 19.02 -6.1 -37 11.80 -0.89 -10.6 -4.83 -27.3
-7.5 -45.8 14.20 -6.8 -39 15.40 -3.03 -27.3 -4.95 -32.1
-6.05 -33.4 15.00 -7.1 -39 17.80 -3.38 -19.9 -5.16 -34.7
-4.22 -17.9 15.86 -5.2 -29 12.60 -7.07 -51.8 -2.33 -13
-6.94 -40 15.52 -11.7 -77 16.60 -6.26 -31 -5.05 -28.4
-9.47 -55 20.76 -10 -59 21.00 -10.38 -66.4 -4.88 -24.3
-10.4 -67.1 16.10 -11.2 -69 20.60 -10.72 -74.5 -7.26 -45.5
-9.4 -62.4 12.80 -9.5 -55 21.00 -14.03 -104.2 -7.32 -44.4
-3.6 -17.1 11.70 -6.8 -36 18.40 -8.95 -51.4 -11.46 -84
-2.9 -9 14.20 -6.9 -39 16.20 -9.43 -58.7 -10.86 -75.1
-3.3 -17 9.40 -4.05 -28.9 -7 -44
-7 -44 12.00 -0.46 -7.1 -8.37 -53.1
-5.1 -30 10.80 -0.46 -4.5 -7.38 -50.7
-6.9 -39 16.20 -0.52 -7.8 -5.08 -35.3
-7.4 -44 15.20 -3.93 -11.8 -7.61 -54.3
-7.1 -37 19.80 -12.11 -89.4 -3.7 -27.3
-11.1 -66 22.80 -12.76 -84.4 -2.04 -17.7
-11 -68 20.00 -10.51 -70.7 -4.71 -31.2
-9.1 -53 19.80 -7.8 -50.1 -1.6 -17.2
-8.1 -53 11.80 -7.77 -46.9 -9.09 -61
-10.3 -72 10.40 -3.91 -19.4 -9.08 -54.9
-5.1 -37 3.80 -5.32 -26.9 -13.69 -113.1
-5.3 -38 4.40 -8.12 -59.4 -7.98 -50.9
-5.9 -36 11.20 -4.35 -28.8 -7.06 -45.3
-6.6 -41 11.80 -7.04 -40.7 -6.15 -39.7
-7.2 -40 17.60 -16.44 -113.5 -9.48 -61.7
-4.1 -24 8.80 -11.27 -79.4 -4.99 -30.6
-11.4 -78 13.20 -9.63 -65.1 -5.53 -32
-9.6 -63 13.80 -8.92 -64.8 -5.49 -29.7
-10.7 -69 16.60 -6.56 -46.7 -5.32 -31.1
-7.7 -45 16.60 -4.5 -37.8 -5.37 -36.6
-6.4 -31 20.20 -3.3 -26.9 -5.52 -27.6
-5.7 -38 7.60 -3.07 -22.3 -13.4 -88.8
-4 -23 9.00 -2.99 -22.9 -4.93 -32.9
-4.1 -25 7.80 -10.84 -73.3 -4.35 -22.4
-3.7 -25 4.60 -9.8 -64.7 -6.46 -33.5
-5.8 -38 8.40 -10.55 -76.4 -11.75 -78
-7.4 -44 15.20 -10.62 -70.2 -11.58 -71.5
-7.4 -44 15.20 -6.74 -42.1 -9.59 -66.8
-10.6 -64 20.80 -0.01 -1.1 -9.37 -60.1
-9.8 -61 17.40 -5.08 -39.64 -3.63 -28.3
-9.8 -57 21.40 -4.17 -33.5 -3.72 -28.3
-8.2 -54 11.60 -3.95 -23.9 -3.52
-4.2 -25 -2.55 -17.2
-5.47 -32.2 -4.88 -38.6
-6.62 -39.3 -8.14 -51.3
-5.06 -33.42 -9.41 -73.8
-1.95 -21.22 -9.26 -61.4
-7.94 -46.39 -6.25 -37.1
-11.35 -73.79 -4.7 -28.2
-11.3 -72.16 -11.81 -82.5
-15.21 -107.89 -7.08 -44.6
-15.77 -113.9 -6.29 -40.2
-16.88 -120.64 -0.28 -6
-12.62 -99.08 -6.19 -47.8
-4.63 -38.36 -6.19 -38.3
-5.76 -38.04 -7.87 -50.8
-3.6 -42.38 -13.2 -92.5
5.36 13.89 -6.72 -36.92
Rainwaters, Ancona, Central Apennines Freshwater, South Appenines Freshwater, Central Anatolia