Advances in microclimate ecology arising from remote sensing Florian Zellweger1, Pieter De Frenne2, Jonathan Lenoir3, Duccio Rocchini4,5,6, David Coomes1 1 Forest Ecology and Conservation Group, Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge CB23EA, UK 2Forest & Nature Lab, Ghent University, Geraardsbergsesteenweg 267, BE-9090 Gontrode, Belgium 3 UR “Ecologie et dynamique des systems anthropisés” (EDYSAN, UMR 7058 CNRS-UPJV), Université de Picardie Jules Verne, 1 Rue des Louvels, 80037 Amiens Cedex 1, France 4 University of Trento, Center Agriculture Food Environment, Via E. Mach 1, 38010 S. Michele all'Adige (TN), Italy 5 University of Trento, Centre for Integrative Biology, Via Sommarive, 14, 38123 Povo (TN), Italy 6Fondazione Edmund Mach, Research and Innovation Centre, Department of Biodiversity and Molecular Ecology, Via E. Mach 1, 38010 S. Michele all'Adige (TN), Italy Correspondence: fz255@cam.ac.uk (Zellweger, F.); dac18@cam.ac.uk (Coomes, D.) Keywords: Biodiversity, Climate Change Ecology, Light Detection and Ranging LiDAR, Thermal imaging, Topography, Vegetation Cover Abstract Microclimates at the land-air interface affect the physiological functioning of organisms which, in turn, influences the structure, composition and functioning of ecosystems. We review how remote sensing technologies that deliver detailed data about the structure and thermal composition of environments are improving the assessment of microclimate over space and time. Mapping landscape-level heterogeneity of microclimate advances our ability to study how organisms respond to climate variation, which has important implications for understanding climate-change impacts on biodiversity and ecosystems. Interpolating in-situ microclimate measurements and downscaling macroclimate provide an organism-centred perspective for studying climate-species interactions and species distribution dynamics. We envisage that mapping of microclimate will soon become commonplace, enabling more reliable predictions of species and ecosystem responses to global change. Importance of microclimate maps Local modification of the climate (See Glossary) by topography and vegetation produces microclimates at the land-air interface which can differ greatly from the climatic means [1,2]. Surface temperatures between north- and south-facing mountainsides, for example, can vary by 20 °C, equivalent to a latitudinal gradient of about 2000 km [3]. Additionally, forest canopies can buffer the diurnal amplitude of air temperature in the forest understorey by 7 °C [4]. Such differences in temperature within landscapes matter to organisms, affecting processes such as respiration, heat and energy exchange which, in turn, set thermodynamic constraints on species behaviour, growth, reproduction and survival [5–7]. Innumerable papers over the past century have quantified microclimates and their influences on ecological processes at all levels of organization, from physiological processes of single organisms to ecosystem-level productivity and nutrient cycling [4–6,8–10]. Microclimate is also relevant to evolution because phenotypic and genotypic adaptations are driven by environmental conditions actually experienced by the organisms [11]. Moreover, microclimate mapping and monitoring have been recognised as key to effective natural resource management, with forestry, agroforestry and agriculture being prominent examples [6,12]. Microclimate ecology is attracting renewed attention due to its fundamental importance in understanding how organisms respond to climate change [2]. Species distributions are typically modelled using macroclimate data obtained from national networks of weather stations [13,14]. These standard meteorological data are measured in open areas at 1.5 - 2 m height above short grass, and capture synoptic conditions that are unrepresentative of a range of microclimates that most organisms experience [15,16]. These inaccuracies and biases can have serious implications when predicting organismal responses to climate change. For example, recent studies suggest that many plant and animal communities are accumulating a climatic debt because they are migrating more slowly than needed to keep up with macroclimate warming [17–22]. However, temperature buffering near the ground – due to local radiation regimes, soil characteristics and topography – means that organisms may not have to migrate, or adapt, as quickly as previously thought to keep pace with the shifting macroclimate [23,24] (Box 1). Thus, extinction risk from climate change for plants and insects is considerably reduced by the occurrence of microrefugia within landscapes with highly heterogeneous microclimate [25]. Yet, the modulating effects of microclimate variability on climate change impacts have only recently started to be quantified [3,21,25–29]. A key impediment to progress in incorporating microclimate into models of climate change impacts on organisms has been our limited ability to map and monitor microclimatic variation over large spatial scales and over time. Networks of microclimate sensor provide point-based measurements and weather stations provide macroclimate data, but we have lacked approaches to effectively interpolate and downscale this information. Remote sensing is now offering opportunities to lift this technical barrier, by producing detailed and spatially continuous data-layers that can be used as explanatory variables to understand and model the horizontal and vertical variation in microclimatic conditions over large spatial and temporal scales. Here, we review how these emerging technologies are advancing microclimate modelling and mapping, and highlight some of the opportunities they provide for ecology, conservation and climate change research. Box 1 Shifts in species distributions in response to global warming Microclimate – the local modulation of macroclimate by vegetation canopies and topographic position – affects species re-distribution under climate change (Figure I). Maps of microclimate predicted from remote sensing data can improve habitat suitability maps and predictions of how species will respond to climate change. Eureka: remote sensing advances for modelling and mapping microclimate Remote sensing technologies are increasingly capable of mapping the structural complexity and thermal composition at the ground-atmosphere boundary at scales relevant to studying organismal responses to environmental variation [27]. We discuss the contributions that laser scanning, photogrammetry, hyperspectral imaging and thermal imaging are making. Airborne Light Detection and Ranging (LiDAR) (aka airborne laser scanning) is particularly valuable for modelling and mapping microclimate because it provides spatially continuous, sub-metre-scale information on two key modifiers of climate at the ground-atmosphere interface: ground topography and vegetation structure [30]. To construct maps, microclimate measurements taken on the ground using sensor networks are related to LiDAR structural information, such as topographic position and light incidence at very high resolutions (Boxes 2-4), using statistical modelling approaches, and the function generated by this approach is then used to predict microclimate across the entire LiDAR-mapped landscape (Figure 1) [13,31–35]. Effective interpolation requires that the sensor networks sample contrasting sites within the study area. The sensor data must also be summarised in ecologically meaningful ways, guided by clear research questions [36]. For example, the frequency of extremely cold or hot temperatures, calculated over timescales relevant for the growth and survival of organisms, are more meaningful for biogeographic applications than average conditions [2,36]. Aerial photography provides an alternative approach to assessing topography and forest structure, using photogrammetry and structure-from-motion (SfM) techniques to construct 3D surfaces (Figure 2) [37]. These inexpensive and easy-to-use methods are increasingly applied, but are less accurate than LiDAR at deriving terrain elevation beneath tree canopies, or for measuring vertical vegetation structure, because photos only record reflectance off the upper surface [38,39]. One-off mapping of large areas using LiDAR and aerial photography is normally conducted from manned aircrafts, while unmanned aerial vehicles (UAV, e.g. drones) equipped with miniaturised cameras and LiDAR sensors are becoming available to map smaller areas at even higher spatiotemporal resolutions. Using UAVs and SfM techniques, Milling et al. [40] found that summer maximum temperatures may vary up to four degrees Celsius over just a few metres within sagebrush-steppe landscapes – habitats that were previously considered relatively homogeneous. A key advantage of UAVs is that deployment is very flexible, enabling the collection of time-series of aerial imagery over a period of interest at relatively low costs. SfM techniques applied to image time-series offer novel opportunities for monitoring microclimate in ecosystems in which phenology creates strong temporal variation in microclimate [41,42]. Terrestrial laser scanning (TLS) provides immensely detailed datasets of vegetation structure that can be used to model microclimate. Complementary to airborne laser scanning, which maps 3D vegetation structure from above, TLS maps vegetation in extraordinary detail from below, thus providing information on the understorey structure. Kong et al. [43] found that TLS-based reconstructions of canopy volumes coupled with microclimate measurements revealed cooling effects in the understorey that varied among tree species, suggesting that TLS can pick up subtle effects of different leaf sizes on understorey microclimate. Moreover, Ehbrecht et al. [44] found that TLS-derived measurements of canopy openness were positively related to diurnal temperature ranges in managed temperate forests in Germany. TLS measurements are restricted to a few hectares and are of limited use, compared to airborne laser scanning, for modelling microclimate over large areas. Yet, the forest understorey-structure information TLS provides at the plot level has been shown to improve landscape-level vertical vegetation structure mapping based on full-waveform airborne LiDAR [45]. Complementing maps of 3D vegetation structure, maps of leaf functional traits and species obtained by hyperspectral remote sensing [46,47] are likely to improve the statistical fit of microclimate models. We expect this improvement because the quality and quantity of solar radiation transmitted by canopies vary according to leaf traits and tree species, leading to species-specific microclimatic conditions in the understorey [48]. However, we are unaware of studies using hyperspectral remote sensing to map microclimate (see Outstanding Questions). Box 2: Measuring how plant canopies affect solar radiation fluxes Solar radiation flux has strong effects on the energy budget and performance of organisms living beneath vegetation canopies. Radiation regimes along the vertical canopy profile of forests can be estimated from a Light Detection and Ranging (LiDAR) point cloud by creating a 3D map of foliage presence/absence in voxels (i.e. 3D pixels) and then apply ray tracing algorithms to evaluate whether beams entering the canopy in different locations and angles are likely to be intercepted [31,49]. Alternatively, LiDAR data can be used to generate synthetic hemispherical images from which fluxes of non-directional diffuse sky radiation and direct solar radiation, or light extinction following the Beer-Lambert law [50], can be calculated for any time in the day or year (Figure I). These approaches are computationally intensive but better represent light conditions experienced by forest organisms than simple approaches based on canopy cover [51]. Vegetation structure thus drives the interception of solar radiation, which means that the importance of vegetation structure for microclimate will vary between day and night and different weather conditions, with the temperature offsets highest on bright sunny days. Advances in physically based radiative transfer modelling now make it possible to estimate the 3D radiative budget in forests and open lands at an ever-increasing detail, e.g. by accounting for foliar-specific filtering of different wavelengths [52]. Box 3: Temperature buffering and offset Solar radiation reaching the land-atmosphere interface is mostly reflected, or absorbed and re-emitted as thermal radiation, or drives evapotranspiration. Vegetation canopies lift energy-exchange surfaces off the ground, and in doing so modulate radiant fluxes, air temperature and humidity at ground level. The capacity of plant canopies to sustain a different temperature below canopy compared to free-air conditions (i.e. a temperature offset) is thus closely related to canopy structure and composition. Under canopy, diurnal changes in temperatures are less extreme than above canopy, and this temperature buffering is modulated by canopy height and cover, both of which can now be precisely mapped [13,32,53]. Sensor networks sampling environmental gradients (cf. Figure 1) are increasingly combined with remote sensing data to map microclimate. The current scientific literature often makes crude assumptions about the shading and temperature buffering effect of vegetation when modelling microclimate, and usually neglects systematic changes in the temperature offset over time, i.e. the offset trend (Figure I) [24,28]. The degree to which temperatures below the canopy are offset compared to free-air conditions will not be constant over time and depend on successional processes driving dynamics in canopy structure and composition. Long time series of below-canopy temperature records thus need to be related to forest dynamics to better understand the drivers of long-term microclimatic dynamics [24]. Such data are scarcely available [54] and global long-term networks such as FLUXNET may prove very valuable in this respect. Forest microclimates are also affected by landscape features such as distances to forest edges, urban areas and large water bodies. Many of these landscape features can be retrieved from remote sensing data [4,29,32,35,55] and integrated into predictive models used to map microclimate. Another key influence on spatiotemporal dynamics in microclimate is topographic position, because it determines the influences of cold air drainage and pooling on a site [56,57]. Topographic position and cold air drainage can be estimated from high-resolution digital terrain models (DTMs), further increasing our ability to map and model microclimate across broad spatial and temporal scales. Box 4: Water and wind Plant canopies not only buffer temperature, but also precipitation, relative humidity and vapour pressure deficit (VPD), which is exponentially related to air temperature. VPD drives transpiration in plants and growth and survival can be impeded when VPD is high (responses vary greatly among species and depend on water supply and leaf temperature). In a degraded tropical forest landscape, models of understorey VPD generated by interpolating sensor-network data with LiDAR imagery (see Figure 1) suggest that tropical tree regeneration will be severely affected by global warming, because of the close link between temperature and VPD [13]. The effect of remotely sensed canopy structure and composition on below-canopy VPD and moisture availability warrants further research, e.g. to better understand how moisture influences air and topsoil temperatures, and vice-versa [31,58]. Topographic features, such as slope angle, affect the lateral surface and subsurface water flow. Airborne LiDAR-derived maps of topographic wetness and ruggedness are thus suitable to analyse the fine-scale variation of soil moisture and air humidity [59,60]. Detailed ecosystem structure data also delivers input parameters to better account for the effects of wind on microclimate. Canopy surface roughness and vertical canopy structure, for instance, improve wind modelling in heterogeneous forests and offer promising opportunities to make predictions of the near-surface wind fields more accurate [61,62]. Thermal imaging using thermal infrared (TIR) cameras can be applied to map surface temperatures. As opposed to LiDAR technologies TIR cameras directly record longwave infrared radiation (i.e. 7.5-14 µm) emitted by an object or organism, which is linked to surface temperature according to Boltzmann’s law when surfaces have high emissivity [3,63]. The surface and body temperature of an organism is related to its energy budget and thus to the functioning and performance of plants and animals [7,64]. Yet, the surface temperature is not necessarily related to the air temperature an organism experiences. For instance, plants respond to water shortage by closing stomata and reducing transpiration, which causes leaf surface temperatures to rise – irrigated and non-irrigated plants can differ in leaf temperature by several degrees but have similar air temperature in their surrounds, as measured with shielded temperature sensors (Figure 2). TIR images recorded by UAVs have centimetre resolution [63], providing valuable means for the fine-scale monitoring and management of water use and water stress by plants, e.g. in crop production [66,66], or to assess the temperature experienced by insects living on a leaf’s surface. However, TIR images might not necessarily reflect atmospheric or soil microclimatic temperatures experienced by plants, i.e. their thermal niche. Surface temperatures from high-resolution TIR images have been applied for fire and disease detection, phenotyping in plant breeding, wildlife monitoring and microclimate ecology (reviewed in [42,66]). Senior et al. [67], for example, used TIR images to show that selective logging of tropical forests had a very little impact on thermal buffering compared to primary forests, suggesting that selectively-logged tropical forests may play an important role in retaining species with temperature niches that are disappearing under climate change. In aquatic systems, TIR images provide the means for landscape-level mapping of cold water patches (thermal refuges) along rivers – an important habitat element for riverine salmonids in times of climate warming [68]. Such maps provide valuable information to guide conservation efforts. We currently know little about the extent to which canopy surface temperatures measured by TIR images are coupled to the temperatures prevailing in the layers beneath the canopy surface, e.g. in forest understoreys or at the soil surface, although this knowledge would be helpful for using TIR images to model and map microclimatic air temperature. The difference between canopy leaf temperatures and ambient air temperatures can be highly variable and depends on canopy structure and species-specific leaf traits, such as aerodynamic leaf boundary-layer resistance and associated levels of atmospheric coupling [69]. Such analysis will also be subject to effects deriving from the ability of plants to regulate leaf temperature [64]. Research into the relationship between below-canopy temperatures measured by sensor networks (Box 3) and canopy temperature measured by TIR images is needed to further understanding of these linkages. TIR radiation flux is affected by a number of factors besides leaf temperature, including the relative humidity, ambient temperature, wavelength dependency of the emissivity and range of the camera, wind speed and shadows [70]. Accurate surface temperature assessments using TIR imagery can thus be challenging. A key point is the emissivity, which is the ability of the surface of an object to emit thermal radiation [63,71]. The mean emissivity of surfaces from plants, soil and rocks range from 0.903-0.997 and deriving surface temperature data from TIR images is thus complicated by the fact that not all surfaces in the image have similar thermal emissivity [71]. Furthermore, the spatiotemporal resolution of TIR imagery needs to be considered. Representing the climate conditions at a site requires TIR images taken across the full range of weather conditions, at day and night, and across seasons [2]. While this may be feasible for terrestrial and potentially airborne TIR imagery, the high spatiotemporal resolution of such datasets comes at the cost of limited spatial coverage. Satellite TIR imagery provides surface temperature data with global coverage, although at too coarse a resolution to directly quantify microclimate (Figure 2). Yet, satellite TIR images can improve the interpolations of temperature data from weather stations in areas with a low station density [72]. Despite these challenges and limitations, the potential of TIR imagery in fundamental and applied microclimate ecology is substantial and should be explored in more detail. Another approach to microclimate mapping is to downscale macroclimate data obtained from macroclimatic grids [2,24], such as WorldClim 2 [72] and CHELSA [73], which are published at relatively coarse scales (typically 30’’ resolution, equivalent to 1 km2 at the equator). High-resolution remote sensing products, such as digital terrain models (DTMs), canopy height models (CHMs) or detailed ground and canopy albedo measurements, are used to generate indices of microclimatic processes related to solar radiation, cold-air drainage or topographic wetness from the grid data, and these indices are then related statistically to macroclimatic variables using regression [74–76]. Software such as R-packages implementing these approaches using freely available input data are now becoming available [14]. Because these models are based on macroclimate data that are available at a high temporal resolution, such models allow for predictions of how microclimate conditions vary in time, thus tackling a key limitation of temporally limited approaches based on microclimate measurements from sensor networks (cf. Box 3). Mechanistic models may also use predictor variables derived from remote sensing data but are fundamentally different in that they model heat and mass exchange between organisms and their environments, relying on functional relationships derived from the physical processes involved in creating microclimate [77,78]. Perhaps the most advanced mechanistic model is Niche MapperTM [77], which downscales air temperatures based on a set of abiotic variables such as soil characteristics, macroclimatic meteorological variables including cloud cover, air temperatures and wind speeds and shading. The model has been parameterized to predict lizard distributions in open habitats in Australia and the US but does not currently include detailed modulating influences of plant canopies among its input variables [28,77]. Implications and avenues for microclimate ecology Ecologists are starting to appreciate the ways in which microclimate mapping technologies could improve their science. Correlative species distribution modelling (SDM) is often criticised for its reliance on coarse climate information [24] and its failure to incorporate physiological knowledge [8]. Using detailed spatiotemporal microclimate data in such models will allow for more organism-centred approaches to determine species range boundaries and their climate change-related dynamics. This especially applies at the temperature-driven leading and trailing edges, where the response of organisms may be particularly susceptible to the availability of suitable microclimate and associated microrefugia [8,24,26,79,80]. Incorporating microclimate layers into SDMs thus holds a large potential, but is still in its infancy. Using simulations and focusing on maximum temperature of the warmest month, Lenoir et al. [24] found that using airborne LiDAR-derived variables to model microclimate decreases the extirpation risk of a virtual plant species under climate change compared to predictions based on downscaled climate data at coarser resolutions (Box 1). Such modelling results are physiologically more meaningful because they derive from the comparison of the species’ temperature niche to realistic temperature dynamics driven by vegetation shading and cold air drainage. Microclimate data will also help to shed new light into microclimatic effects on phenology – potentially quantified by remotely sensed vegetation indices such as the Normalized Difference Vegetation Index (NDVI) – and how these effects affect species distributions and species interactions. For instance, plant species range limits may be driven by temperature extremes during key stages of phenology, such as extreme cold during bud-break of broad-leaved tree species [81]. Such extreme events are not represented in currently available climate data with coarse spatiotemporal resolutions. Using remote sensing data to derive climate data at resolutions similar to those at which organisms perceive and respond to climate conditions is thus a timely task and will pave the way for more reliable predictions of species range dynamics in response to climate change [27]. Microclimate mapping could also refine our understanding of species diversity patterns. Following the environmental heterogeneity hypothesis, microclimate heterogeneity is expected to be positively related to species richness (alpha diversity) [82], but this remains understudied. Similarly, investigating how spatial and/or temporal changes in microclimate contributes to beta diversity through environmental filtering deserves more attention [82,83]. For example, a recent study found that microclimate on cooler, north-facing slopes affected plant community responses to climate change by delaying extinctions of species with low-temperature requirements [84]. Increased short distance microclimatic variation is expected to affect the climatic debt in bird assemblages, e.g. by lowering the risk of population decline due to the ability to avoid harmful climatic variation or by increasing landscape permeability which facilitates the spatial tracking of climate change [85]. Spatiotemporal mapping of microclimate will thus be crucial to understanding how local phenomena give rise to large-scale processes. Estimating connectivity among fragmented habitats to evaluate the functionality of ecological networks, for instance, depends on reliable landscape-level representations of microrefugia, stepping stones, 3D-habitat structure and associated microclimate (e.g. ecological corridors such as hedgerows), as these attributes are critical for species migration and gene flow [79,80,86]. The potential of remote sensing technologies to better understand and model microclimate is already recognised implicitly in the ecological literature. The widespread use of LiDAR to model species occurrence from habitat structure, for instance, relies implicitly on the assumption that LiDAR data can be used to assess microclimate conditions that are, at least partially, responsible for the fitness and distribution of an organism [87–89]. What is missing in such indirect approaches is how the measured environmental features actually drive and interact with the microclimate variables that are physiologically relevant to the species or the biological phenomena of interest, e.g. the minimum and maximum air temperatures relevant to an organism’s temperature tolerance [7]. In forests, our mechanistic understanding of how canopy structure and composition drive and interact with vertical radiation and temperature regimes to determine species habitat preferences and vertical niche partitioning is still incomplete. Indeed, the steep vertical microclimatic gradients within forests are increasingly appreciated for structuring arboreal biodiversity, particularly in the tropics [90–92] and the remote sensing approaches described here play a key role in filling this knowledge gap. Microclimatic changes arising from forest management have been shown to exert strong controls on local plant communities and their response to macroclimate warming [21]. Thus, mapping of microclimate has far-reaching implications for conservation and other fields, such as forestry and agriculture [35]. Successful tree regeneration – planted or natural – strongly depends on microclimate conditions [12,74]. Maps of thermal and light regimes below different canopy conditions or in clear-cuts can help managers optimise planting in accordance to tree species-specific temperature and light adaptations [12]. Similarly, microclimate maps would be helpful for managing agroforestry systems, such as those associated with coffee and cacao, where microclimates affect yield and the susceptibility to climate extremes [93]. In agriculture, precision farming of speciality crops increasingly relies on remote sensing technologies capturing the spatiotemporal variability of the micro-environmental conditions [94,95]. Mapping the thermal heterogeneity across landscapes improve the analysis and management of crop water status [66,94] and how microclimate affects the occurrence and dynamics of pests [63,96]. Current limitations and future directions Field measurements of microclimate recorded with sensor networks are crucial for the development of landscape-scale maps, but sensor and sampling designs vary greatly between studies, making it difficult to synthesise results [2]. The need for standardised sampling approaches, centred around the following principles, is increasingly recognised: (1) field surveys are designed to represent the entire spatial and temporal gradients of the microclimate conditions in the study system; (2) time span between the collection of field and remote sensing datasets is short enough to prevent significant discrepancies; and (3) measurement sites are georeferenced precisely using a differential Global Positioning System, so that the data can be spatially co-registered with the imagery. Simulations show that registration errors as small as 1 m when working with 10-m radius plots can create major uncertainty in forest properties estimated from airborne LiDAR [97]. Thus, also precisely locating species records, particularly of less mobile species, is a prerequisite for sound inference about species-microclimate relationships. The presented airborne remote sensing tools and data (i.e. involving airborne LiDAR and/or SfM) to map the effects of vegetation structure on microclimate near the ground work best in tall habitats, such as forests, wood- and shrublands. In short stature vegetation, such as grassland, heath or crops, the level of structural detail picked up by airborne LiDAR and SfM is unlikely to capture microclimate variation resulting from fine-scale differences in vegetation structure. High-resolution TIR, however, provides the means to measure surface temperatures in both tall and short stature vegetation, but does not provide structural information required to interpolate microclimate measurements from sensor networks. There is pressing need to gather georeferenced microclimate data from different types of habitat across the globe and a global archive and data portal facilitating data access would significantly promote progress in microclimate ecology. To complement temporal dynamics of microclimate data gained from downscaled macroclimate we need long-term microclimate data series [24]. This will enable an improved understanding of the drivers of microclimate dynamics and how they deviate from the macroclimate, which will have important implications for estimating the velocity, and thus impact, of climate change. Many of the remote sensing approaches described here rely on data whose spatial coverage is growing but does not yet expand over continental and global scales. In the future, remedial satellite LiDAR data experiments, e.g. the Global Ecosystem Dynamics Investigation LiDAR (GEDI) or the ICESat-2 satellite project, may provide new avenues to arrive at analysing and monitoring microclimate variation at larger temporal and spatial scales. Satellite missions employing synthetic aperture radar (SAR) systems, such as the launched TanDEM-X and planned Tandem-L missions, provide worldwide, repeated and spatially detailed data for digital terrain elevation and forest height modelling. Incorporating these data into microclimate models may play a key role for increasing their spatial and temporal cover, and is expected to facilitate tracking microclimatic changes in habitats with dynamic structural attributes, such as forest vegetation structure. Concluding Remarks We have shown that advances in remote sensing technologies are making it possible to map microclimate at fine spatiotemporal resolutions and over large areas for the first time. This offers new opportunities to scale up ecological knowledge about the organism-environment interactions at fine scales, to understand species and ecosystem responses to environmental changes over broad scales. Topographically controlled microclimate gradients have historically been studied in more detail than those controlled by 3D-vegetation structure. LiDAR and photogrammetry provide key structural data to fill this gap, which is critical, given the contribution that vegetation structure makes to biodiversity. However, methodological efforts taking an ecological perspective in approximating microclimate via remote sensing tools are required to make most out of the available data and resources. The technological advances in remote sensing and the methodological advances in microclimate modelling call for coordinated efforts between remote sensing experts, climatologists and ecologists to improve our predictive abilities on the role of microclimate in biodiversity and global change ecology. Outstanding questions Improved, open-access and easy-to-use methods based on remotely sensed canopy structure and composition for modelling and predicting microclimate in forests are needed. Such methods are required to further our understanding of light, temperature and relative humidity regimes, and how they affect species behaviour, performance and distribution. Emphasis should also be given to quantifying the effect of local wind dynamics on microclimate. How will microclimatic conditions change in response to climate warming? This will depend on the extent to which vegetation structure and topography modulate air temperature and how changes in vegetation structure change solar radiation and wind regimes. How does horizontal and vertical microclimate variation affect alpha- and beta-diversity? What is the influence of microclimate buffering on species range dynamics, biodiversity and the climatic debt of species and communities? How are microclimate gradients related to plant functional traits derived from hyperspectral imaging? How could landscape-level mapping of microclimate contribute to our understanding of habitat connectivity and the functionality of ecological networks? How are below canopy and soil microclimate linked to vegetation surface temperatures measured by thermal infrared (TIR) imagery? How important is microclimate for driving phenological responses to climate change, and what are the implications thereof for species interactions and distributions? Glossary Airborne LiDAR: a remote sensing technology used for 3D analysis of earth surface environments. LiDAR is short for Light Detection and Ranging (aka laser scanning). A LiDAR sensor emits about 200,000 laser pulses per second towards the ground and measures the energy waveform returning from backscattering objects. When used to measure vegetation structure, the light pulse is wider than a typical leaf by the time it reaches the upper canopy, meaning that some of its energy passes through the upper canopy to lower layers and even the ground. The sensor converts the continuous waveform of returning energy into ‘discreet returns’ and, by precisely recording return times and its location in the air, creates a 3D point cloud of the position of objects. The point cloud is used to derive high-resolution of topography and canopy height (see DTM and CHM) and detailed information on vertical vegetation structure, spatially continuous across large areas. Some LiDAR sensors record the full-waveform, providing detailed information about the entire vertical forest profile. The added value of full-waveform over discrete LiDAR for microclimate mapping remains to be tested. Alpha diversity: species diversity in sites or habitats at the local scale (e.g. point-based surveys), often expressed as the total number of species (species richness) or abundances weighted indices, such as the Shannon index or the Simpson index. Beta diversity: diversity measure expressing variation (turnover and nestedness) in community composition among habitats gradients, can be calculated based on taxonomic (e.g., species identities), functional (e.g., functional traits) and phylogenetic (e.g., branches) units. Canopy Height Model (CHM): continuous digital surface – usually in the form of a raster dataset – representing the height of the canopy above the underlying terrain. Climate: synthesis of atmospheric conditions characteristic of a particular place in the long-term (usually 30-year averages) expressed by averages of various elements of weather and probabilities distributions of extreme events. Climate debt: biotic responses observed in nature are often slower than expected under the assumption of complete synchrony with climate change; climate debt describes the spatiotemporal lag accumulated by a species or a community compared to the actual shift in climate. Cold air drainage: gravity-induced, downslope flow of relatively cold air near the ground, pooling in local depressions and valley constrictions. A prominent phenomenon in mountain valleys at night and during winter. Digital Terrain Model (DTM): continuous digital surface representing the elevation height of the bare earth. Sometimes also referred to as digital elevation model (DEM). Hyperspectral remote sensing: image analysis based on the spectral reflectance across a wide range of the electromagnetic spectrum; also known as hyperspectral imaging or imaging spectroscopy. Macroclimate: the climate conditions above ground or above the canopy (e.g. > 2 m) at a relatively large scale, e.g. across spatial dimensions of 1 km or more, and temporal dimensions of days to weeks or longer. Microclimate: the climate conditions close to the ground (e.g. < 2 m) or along vertical forest profiles at relatively fine spatiotemporal resolutions, e.g., across spatial dimensions of centimetres to meters, and temporal dimensions of minutes or shorter. Microclimate conditions include temperature, precipitation, humidity, wind and radiation regimes. Microrefugia: spatially-restricted local habitats that sustain a climate that has become, or is becoming, lost due to climate change and that enables species to persist in an otherwise inhospitable region. Remote sensing: acquiring information about an object of phenomena from a distance. Temperature buffering: below plant, especially forest canopies, daily air temperatures may be substantially buffered, increasing less during the day and decreasing less during the night than outside the forest canopies. Terrestrial Laser Scanning (TLS): the process of gathering 3D data using a LiDAR instrument on the ground. 3D point clouds produced by TLS are typically much denser than those obtained by airborne LiDAR. Thermal imaging: technique to produce an image based on the heat emitted by an object or an organism. Understorey: a layer of vegetation close to the floor beneath the main canopy of a forest. Vapour pressure deficit (VPD): the difference between saturation vapour pressure and the actual vapour pressure, at a given temperature. Acknowledgements FZ was funded by the Swiss National Science Foundation (grant no. 172198). PDF received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (ERC Starting Grant FORMICA 757833). DR was partially funded by the the H2020 project ECOPOTENTIAL (Grant Agreement no. 641762) and the H2020 TRuStEE - Training on Remote Sensing for Ecosystem modElling project (Grant Agreement no. 721995). DAC was funded by NERC (grant number NE/K016377/1) and a Leverhulme International Fellowship. We thank the Swiss NFI for providing LiDAR data and two anonymous reviewers and Christian Körner for commenting on earlier versions of the manuscript. References 1 Geiger, R. et al. (2003) The climate near the ground, Rowman and Littlefield, Oxford. 2 Bramer, I. et al. (2018) Advances in Monitoring and Modelling Climate at Ecologically Relevant Scales. Adv. Ecol. Res. 58, 101–161 3 Scherrer, D. and Körner, C. (2010) Infra-red thermometry of alpine landscapes challenges climatic warming projections. Glob. Chang. Biol. 16, 2602–2613 4 Chen, J. et al. (1999) Microclimate in forest ecosystem and landscape ecology: Variations in local climate can be used to monitor and compare the effects of different management regimes. Bioscience 49, 288–297 5 Porter, W.P. and Gates, D.M. (1969) Thermodynamic Equilibria of Animals with Environment. Ecol. Monogr. 39, 227–244 6 Jones, H.G. (2014) Plants and microclimate. A quantitative approach to environmental plant physiology. Third Edition., Cambridge: Cambridge University Press. 7 Huey, R.B. et al. (2012) Predicting organismal vulnerability to climate warming: roles of behaviour, physiology and adaptation. Philos Trans R Soc L. B Biol. Sci. 367, 1665–1679 8 Kearney, M. and Porter, W. (2009) Mechanistic niche modelling: combining physiological and spatial data to predict species’ ranges. Ecol. Lett. 12, 334–350 9 Novick, K.A. et al. (2016) Cold air drainage flows subsidize montane valley ecosystem productivity. Glob. Chang. Biol. 22, 4014–4027 10 Uvarov, B.P. (1931) Insects and climate. Trans. R. Entomol. Soc. London 79, 1–232 11 Penuelas, J. et al. (2013) Evidence of current impact of climate change on life: a walk from genes to the biosphere. Glob. Chang. Biol. 19, 2303–2338 12 Aussenac, G. (2000) Interactions between forest stands and microclimate: Ecophysiological aspects and consequences for silviculture. Ann. For. Sci. 57, 287–301 13 Jucker, T. et al. (2018) Canopy structure and topography jointly constrain the microclimate of human-modified tropical landscapes. Glob. Chang. Biol. 24, 5243-5258 14 Maclean, I.M.D. et al. (2018) Microclima: an R package for modelling meso- and microclimate. Methods Ecol. Evol. DOI: 10.1111/2041-210X.13093 15 WMO, World Meterological Organization (2008) Guide to Meteorological Instruments and Methods of Observation. 16 Suggitt, A.J. et al. (2011) Habitat microclimates drive fine-scale variation in extreme temperatures. Oikos 120, 1–8 17 Bertrand, R. et al. (2011) Changes in plant community composition lag behind climate warming in lowland forests. Nature 479, 517–520 18 Barnosky, A.D. et al. (2012) Approaching a state shift in Earth’s biosphere. Nature 486, 52–58 19 Devictor, V. et al. (2012) Differences in the climatic debts of birds and butterflies at a continental scale. Nat. Clim. Chang. 2, 121–124 20 Dullinger, S. et al. (2012) Extinction debt of high-mountain plants under twenty-first-century climate change. Nat. Clim. Chang. 2, 619–622 21 De Frenne, P. et al. (2013) Microclimate moderates plant responses to macroclimate warming. Proc. Natl. Acad. Sci. 110, 18561–18565 22 Alexander, J.M. et al. (2018) Lags in the response of mountain plant communities to climate change. Glob. Chang. Biol. 24, 563–579 23 Scherrer, D. and Körner, C. (2011) Topographically controlled thermal-habitat differentiation buffers alpine plant diversity against climate warming. J. Biogeogr. 38, 406–416 24 Lenoir, J. et al. (2017) Climatic microrefugia under anthropogenic climate change: implications for species redistribution. Ecography. 40, 253–266 25 Suggitt, A.J. et al. (2018) Extinction risk from climate change is reduced by microclimatic buffering. Nat. Clim. Chang. 8, 713–717 26 Lenoir, J. et al. (2013) Local temperatures inferred from plant communities suggest strong spatial buffering of climate warming across Northern Europe. Glob. Chang. Biol. 19, 1470–1481 27 Potter, K.A. et al. (2013) Microclimatic challenges in global change biology. Glob. Chang. Biol. 19, 2932–2939 28 Kearney, M.R. et al. (2014) Microclimate modelling at macro scales: a test of a general microclimate model integrated with gridded continental-scale soil and weather data. Methods Ecol. Evol. 5, 273–286 29 Maclean, I.M. et al. (2017) Fine-scale climate change: modelling spatial variation in biologically meaningful rates of warming. Glob. Chang. Biol. 23, 256–268 30 Lefsky, M.A. et al. (2002) Lidar remote sensing for ecosystem studies. Bioscience 52, 19–30 31 Tymen, B. et al. (2017) Quantifying micro-environmental variation in tropical rainforest understory at landscape scale by combining airborne LiDAR scanning and a sensor network. Ann. For. Sci. 74, 32 32 Frey, S.J.K. et al. (2016) Spatial models reveal the microclimatic buffering capacity of old-growth forests. Sci. Adv. 2:e1501392 33 Pradervand, J.-N. et al. (2014) Very high resolution environmental predictors in species distribution models. Prog. Phys. Geogr. 38, 79–96 34 George, A.D. et al. (2015) Using LiDAR and remote microclimate loggers to downscale near-surface air temperatures for site-level studies. Remote Sens. Lett. 6, 924–932 35 Greiser, C. et al. (2018) Monthly microclimate models in a managed boreal forest landscape. Agric. For. Meteorol. 250–251, 147–158 36 Körner, C. and Hiltbrunner, E. (2018) The 90 ways to describe plant temperature. Perspect. Plant Ecol. Evol. Syst. 30, 16–21 37 Westoby, M.J. et al. (2012) ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 179, 300–314 38 White, J. et al. (2013) The Utility of Image-Based Point Clouds for Forest Inventory: A Comparison with Airborne Laser Scanning. Forests 4, 518–536 39 Fonstad, M.A. et al. (2012) Topographic structure from motion: a new development in photogrammetric measurement. Earth Surf. Process. Landforms 38, 421–430 40 Milling, C.R. et al. (2018) Habitat structure modifies microclimate: An approach for mapping fine-scale thermal refuge. Methods Ecol. Evol. 2018, 1648–1657 41 Anderson, K. and Gaston, K.J. (2013) Lightweight unmanned aerial vehicles will revolutionize spatial ecology. Front. Ecol. Environ. 11, 138–146 42 Maes, W. et al. (2017) Optimizing the Processing of UAV-Based Thermal Imagery. Remote Sens. 9, 476 43 Kong, F. et al. (2016) Retrieval of three-dimensional tree canopy and shade using terrestrial laser scanning (TLS) data to analyze the cooling effect of vegetation. Agric. For. Meteorol. 217, 22–34 44 Ehbrecht, M. et al. (2019) Effects of structural heterogeneity on the diurnal temperature range in temperate forest ecosystems. For. Ecol. Manage. 432, 860–867 45 Hancock, S. et al. (2017) Measurement of fine-spatial-resolution 3D vegetation structure with airborne waveform lidar: Calibration and validation with voxelised terrestrial lidar. Remote Sens. Environ. 188, 37–50 46 Asner, G.P. et al. (2015) Quantifying forest canopy traits: Imaging spectroscopy versus field survey. Remote Sens. Environ. 158, 15–27 47 Schneider, F.D. et al. (2017) Mapping functional diversity from remotely sensed morphological and physiological forest traits. Nat. Commun. 8, 1441 48 Canham, C.D. et al. (1994) Causes and consequences of resource heterogeneity in forests: interspecific variation in light transmission by canopy trees. Can. J. For. Res. 24, 337–349 49 Musselman, K.N. et al. (2013) Estimation of solar direct beam transmittance of conifer canopies from airborne LiDAR. Remote Sens. Environ. 136, 402–415 50 Campbell, G.S. (1986) Extinction coefficients for radiation in plant canopies calculated using an ellipsoidal inclination angle distribution. Agric. For. Meteorol. 36, 317–321 51 Alexander, C. et al. (2013) Airborne laser scanner (LiDAR) proxies for understory light conditions. Remote Sens. Environ. 134, 152–161 52 Gastellu-Etchegorry, J.-P. et al. (2015) Discrete Anisotropic Radiative Transfer (DART 5) for Modeling Airborne and Satellite Spectroradiometer and LIDAR Acquisitions of Natural and Urban Landscapes. Remote Sensing 7, 1667-1701 53 von Arx, G. et al. (2013) Microclimate in forests with varying leaf area index and soil moisture: potential implications for seedling establishment in a changing climate. J. Ecol. 101, 1201–1213 54 De Frenne, P. and Verheyen, K. (2016) Weather stations lack forest data. Science (80). 351, 234 55 Latimer, C.E. and Zuckerberg, B. (2016) Forest fragmentation alters winter microclimates and microrefugia in human-modified landscapes. Ecography. 40, 158–170 56 Pepin, N.C. et al. (2011) The influence of surface versus free-air decoupling on temperature trend patterns in the western United States. J. Geophys. Res. 116, 57 Daly, C. et al. (2010) Local atmospheric decoupling in complex topography alters climate change impacts. Int. J. Climatol. 30, 1857–1864 58 Ashcroft, M.B. and Gollan, J.R. (2013) Moisture, thermal inertia, and the spatial distributions of near-surface soil and air temperatures: Understanding factors that promote microrefugia. Agric. For. Meteorol. 176, 77–89 59 Leempoel, K. et al. (2015) Very high-resolution digital elevation models: are multi-scale derived variables ecologically relevant? Methods Ecol. Evol. 6, 1373–1383 60 Kemppinen, J. et al. (2017) Modelling soil moisture in a high-latitude landscape using LiDAR and soil data. Earth Surf. Process. Landforms DOI: 10.1002/esp.4301 61 Boudreault, L.-É. et al. (2017) How Forest Inhomogeneities Affect the Edge Flow. Boundary-Layer Meteorol. 162, 375–400 62 Schlegel, F. et al. (2012) Large-Eddy Simulation of Inhomogeneous Canopy Flows Using High Resolution Terrestrial Laser Scanning Data. Boundary-Layer Meteorol. 142, 223–243 63 Faye, E. et al. (2016) A toolbox for studying thermal heterogeneity across spatial scales: from unmanned aerial vehicle imagery to landscape metrics. Methods Ecol. Evol. 7, 437–446 64 Michaletz, S.T. et al. (2015) Plant Thermoregulation: Energetics, Trait-Environment Interactions, and Carbon Economics. Trends Ecol. Evol. 30, 714–724 65 Möller, M. et al. (2007) Use of thermal and visible imagery for estimating crop water status of irrigated grapevine. J. Exp. Bot. 58, 827–838 66 Jones, H.G. et al. (2009) Thermal infrared imaging of crop canopies for the remote diagnosis and quantification of plant responses to water stress in the field. Funct. Plant Biol. 36, 978–989 67 Senior, R.A. et al. (2018) Tropical forests are thermally buffered despite intensive selective logging. Glob. Chang. Biol. 24, 1267–1278 68 Dugdale, S.J. et al. (2015) Spatial distribution of thermal refuges analysed in relation to riverscape hydromorphology using airborne thermal infrared imagery. Remote Sens. Environ. 160, 43–55 69 Leuzinger, S. and Körner, C. (2007) Tree species diversity affects canopy leaf temperatures in a mature temperate forest. Agric. For. Meteorol. 146, 29–37 70 Vollmer, M. and Möllmann, K.-P. (2017) Infrared Thermal Imaging: Fundamentals, Research and Applications, John Wiley & Sons. 71 Rubio, E. et al. (1997) Emissivity measurements of several soils and vegetation types in the 8–14, μm Wave band: Analysis of two field methods. Remote Sens. Environ. 59, 490–521 72 Fick, S.E. and Hijmans, R.J. (2017) WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37, 4302-4315 73 Karger, D.N. et al. (2017) Climatologies at high resolution for the earth’s land surface areas. Sci. Data 4, 170122 74 Dingman, J.R. et al. (2013) Cross-scale modeling of surface temperature and tree seedling establishment in mountain landscapes. Ecol. Process. 2, 30 75 McCullough, I.M. et al. (2016) High and dry: high elevations disproportionately exposed to regional climate change in Mediterranean-climate landscapes. Landsc. Ecol. 31, 1063–1075 76 Meineri, E. and Hylander, K. (2017) Fine-grain, large-domain climate models based on climate station and comprehensive topographic information improve microrefugia detection. Ecography. 40, 1003–1013 77 Kearney, M.R. and Porter, W.P. (2017) NicheMapR – an R package for biophysical modelling: the microclimate model. Ecography. 40, 664–674 78 Kearney, M.R. et al. (2014) microclim: Global estimates of hourly microclimate based on long-term monthly climate averages. Sci. Data 1, 140006 79 Hannah, L. et al. (2014) Fine-grain modeling of species’ response to climate change: holdouts, stepping-stones, and microrefugia. Trends Ecolgy Evol. 29, 390–397 80 Dobrowski, S.Z. (2011) A climatic basis for microrefugia: the influence of terrain on climate. Glob. Chang. Biol. 17, 1022–1035 81 Kollas, C. et al. (2013) Spring frost and growing season length co-control the cold range limits of broad-leaved trees. J. Biogeogr. 41, 773–783 82 Opedal, Ø.H. et al. (2015) Linking small-scale topography with microclimate, plant species diversity and intra-specific trait variation in an alpine landscape. Plant Ecol. Divers. 8, 305–315 83 Zellweger, F. et al. (2017) Beta diversity of plants, birds and butterflies is closely associated with climate and habitat structure. Glob. Ecol. Biogeogr. 26, 898–906 84 Maclean, I.M.D. et al. (2015) Microclimates buffer the responses of plant communities to climate change. Glob. Ecol. Biogeogr. 24, 1340–1350 85 Gauzere, P. et al. (2017) Where do they go? The effects of topography and habitat diversity on reducing climatic debt in birds. Glob Chang Biol 23, 2218–2229 86 Milanesi, P. et al. (2017) Three-dimensional habitat structure and landscape genetics: a step forward in estimating functional connectivity. Ecology 98, 393–402 87 Davies, A.B. and Asner, G.P. (2014) Advances in animal ecology from 3D-LiDAR ecosystem mapping. Trends Ecol. Evol. 29, 681–691 88 Simonson, W.D. et al. (2013) Remotely sensed indicators of forest conservation status: Case study from a Natura 2000 site in southern Portugal. Ecol. Indic. 24, 636–647 89 Zellweger, F. et al. (2016) Environmental predictors of species richness in forest landscapes: abiotic factors versus vegetation structure. J. Biogeogr. 43, 1080–1090 90 Scheffers, B.R. et al. (2014) Microhabitats reduce animal’s exposure to climate extremes. Glob. Chang. Biol. 20, 495–503 91 Nakamura, A. et al. (2017) Forests and Their Canopies: Achievements and Horizons in Canopy Science. Trends Eco.l Evol. 32, 438–451 92 Ashton, L.A. et al. (2016) Vertical stratification of moths across elevation and latitude. J. Biogeogr. 43, 59–69 93 Lin, B.B. (2007) Agroforestry management as an adaptive strategy against potential microclimate extremes in coffee agriculture. Agric. For. Meteorol. 144, 85–94 94 Lee, W.S. et al. (2010) Sensing technologies for precision specialty crop production. Comput. Electron. Agric. 74, 2–33 95 Mulla, D.J. (2013) Author ’ s personal copy Special Issue : Sensing in Agriculture Review Twenty five years of remote sensing in precision agriculture : Key advances and remaining knowledge gaps 5. Biosyst. Eng. 114, 358–371 96 Faye, E. et al. (2017) Does heterogeneity in crop canopy microclimates matter for pests? Evidence from aerial high-resolution thermography. Agric. Ecosyst. Environ. 246, 124–133 97 Frazer, G.W. et al. (2011) Simulated impact of sample plot size and co-registration error on the accuracy and uncertainty of LiDAR-derived estimates of forest stand biomass. Remote Sens. Environ. 115, 636–649 98 Metz, M. et al. (2014) Surface Temperatures at the Continental Scale: Tracking Changes with Remote Sensing at Unprecedented Detail. Remote Sensing 6, 3822-3840 99 Moeser, D. et al. (2014) Canopy closure, LAI and radiation transfer from airborne LiDAR synthetic images. Agric. For. Meteorol. 197, 158–168 100 Bennie, J. et al. (2008) Slope, aspect and climate: Spatially explicit and implicit models of topographic microclimate in chalk grassland. Ecol. Modell. 216, 47–59 Figure 1. Conceptual overview of the approach used to generate microclimate maps from a sensor network. A: Microclimate data are recorded using a network of sensors measuring air/soil temperature and humidity conditions, e.g., placed in the open (S1) and below tree canopies (S2) as shown by 3D airborne Light Detection and Ranging (LiDAR) data in the top panel. The microclimate data from each sensor (S1, S2, and black dots) are then summarised in ecologically meaningful ways, e.g. to daily minimum (Tmin) and maximum (Tmax) temperatures as shown in the middle left panel, and related to vegetation structure and the topography mapped using remote sensing technologies, e.g., LiDAR, as shown for canopy height and elevation across a landscape in the tropical lowlands [13]. B: Statistical models are then used to predict microclimate across the entire mapped landscape and over time. In this example, maximum canopy height and topographic position were strong predictors of maximum daily air temperatures in the understorey (left), which explained small-scale variation of maximum vapour pressure deficit (VPD) (right), as indicated by the black arrows (taken from Jucker et al. [13]). Figure 2. Thermal infrared (TIR) imaging reveals spatially detailed information about surface temperatures. Images A and B show land surface temperatures (LSTs) for Europe (EuroLST) derived from freely available MODIS satellite images with a pixel size of 250 m [98]. On the other hand, data for images C to E were recorded at sub-metre resolution by an UAV flown at 70 m height above ground during an exceptional drought in June 2017 in a tree diversity experiment in Belgium (www.treedivbelgium.ugent.be). Panel C is conventional red-green-blue (RGB) photography, panel D shows the vegetation height (m) determined by structure-from-motion analysis of overlapping photos and panel E shows the surface temperature derived from the TIR image. We see that surface temperatures of plants on the ground are considerably higher than those of tree surfaces, due to different transpiration rates as a response to water shortage. The data was processed following Maes et al. [42]. Box 1 Figure I. Probability of occurrence maps based on a virtual species approach, for which the realized niche is known, predicted with current-day macroclimate (A) and microclimate data (B), and projected into the future under a 2 oC warming scenario (C and D respectively). The temperature data for images A and C refer to long-term (30-yr averages during the period 1970-2000) maximum temperature of the warmest month and were obtained by downscaling macroclimate at 25-m resolution to incorporate topoclimatic processes. Spatial variation in microclimate (temperature in this case) generated by trees (i.e. canopy cover) and topography (i.e. topographic concavity) were modelled using 50-cm resolution maps (images B and D) derived from 3D airborne Light Detection and Ranging (LiDAR). Note that microclimatic models indicate much larger areas of suitable habitat than macroclimatic models. In particular, many potential microrefugia are identified in image D which could continue to provide suitable habitat under climate warming (adapted from Lenoir et al. [24]). Box 2 Figure I. Using airborne Light Detection and Ranging (LiDAR) to map solar radiation fluxes in a mountainous region. A: Potential clear sky solar radiation predicted to reach the ground on a summer day if vegetation is absent (i.e. based on a digital terrain model generated by LiDAR); B: Forest canopy height measured over the same region; C: Potential clear sky solar radiation calculated to reach the ground having penetrated through the forest canopy, assuming an increase of shading with increasing vegetation cover and height. It can be seen that much of the landscape is deeply shaded by trees and shrubs, making it suitable for shade-tolerant plant species. D: 3D airborne LiDAR-derived elevation data of a forest (black rectangle in B) is used to construct synthetic hemispherical images at 1 m and 25 m height above the forest floor [99]. E: Reconstructed hemispherical images, taken at the red point position in B, show portions of the sky obscured by trees (black) and the terrain (blue), from which diffuse and direct light transmission can be calculated. These images can be calculated for any point in the landscape and at any height in forest canopies providing unprecedented opportunities to estimate the microclimate in the neighbourhood of individual organisms. Note that ground topography (elevation, aspect and slope) have strong influences on solar radiation [100], and high-resolution DTMs from LiDAR surveys provide critical input data for quantifying these effects [13,14]. Box 3 Figure I. A: Weather stations as illustrated on the left provide long-term climate data for synoptic conditions (right panel). B: Microclimate data from sensor networks (cf. Figure 1) are currently available mostly for short time periods only, e.g. months to a few years (right panel). The left image shows a shielded sensor placed on the north side of a tree trunk. C: Maximum air temperatures below canopies (i.e. microclimate) are frequently offset by several degrees compared to free-air conditions (i.e. macroclimate) and the offset trend over time may vary. Long-term data series are required to assess the differences in spatiotemporal dynamics between macro- and microclimate (see text). 1 image2.tiff image3.tiff image4.tiff image5.tiff image1.tiff