Many mangrove communities form bands parallel to the shoreline with each community dominated by a single species. However, the key determinants of mangrove species distribution across the intertidal zone are not well understood. We aimed to quantify the relationship between species' dominance and the hydroperiod (defined as the duration of inundation in a year), soil salinity and the salinity of inundating water for three dominant species, Sonneratia alba, Rhizophora stylosa and Ceriops tagal.

An extensive (20,000 ha), largely intact mangrove forest in northern Australia, of some note as mangrove forests are threatened globally.

We related species dominance to the explanatory variables by applying two statistical modelling approaches: generalized linear models (GLMs), where a set of competing models were evaluated; and boosted regression tree models (BRTs), an approach that automatically captures interactions and nonlinear relationships between variables.

Both GLM and BRT models achieved strong predictive performance for all species based on cross-validation, with receiver operating characteristics above 0.85 for all species, and 88% of deviance explained for S. alba, 42% for R. stylosa and 35% for C. tagal. All models indicated that the hydroperiod was the key variable influencing distribution, followed by soil salinity. The salinity of inundating water was the least informative variable in the models. Ecological space, determined by gradients in hydroperiod and soil salinity, was partitioned between the three species with little overlap.

As anticipated changes in sea level will alter the hydroperiod, our findings are critical for global forecasting of future distributions of mangrove communities, and for the design of mitigation and adaptation measures.

To investigate whether latitudinal diversity gradients differ among biogeographic realms for members of a closely related clade and to examine whether differences can be explained by environmental differences such as the temperature gradient.

Indo-Pacific, eastern Pacific and western Atlantic temperate to tropical coastal intertidal.

We digitized the ranges of fiddler crabs (Decapoda, Ocypodidae, genus Uca) and calculated standing diversity as a function of latitude in the Indo-west-Pacific, eastern-Pacific Americas and western Atlantic regions. We examined correlations between diversity and summer sea surface temperature, water column primary productivity, and also investigated the contribution of spatial autocorrelation.

There was a latitudinal diversity gradient with a peak in the tropics or subtropics, but richness as a function of latitude differed by region. The western Atlantic had a broad zone of equal diversity with a peak that corresponds to the Gulf of Mexico–Caribbean Basin. The eastern Pacific had a distinct peak of diversity at about 10° N latitude corresponding to Panama. The Indo-west-Pacific had a broad relatively flat upper level of diversity, reaching a peak at about 20° S latitude corresponding to the north coast of Australia. In both the eastern Pacific and western Atlantic, Northern Hemisphere diversity was greater than Southern Hemisphere. Species richness of the three regions was positively and strongly correlated with air and sea surface temperature at the start of summer. In contrast, diversities were weakly and inconsistently correlated with productivity.

This paper shows that a physical factor is more important in explaining latitude distributions than regional cladal structure or presumed dispersal patterns. While observed diversity–latitude functions are region specific, the feature seen across regions to most strongly explaining the pattern is temperature.

Plant root traits regulate belowground C inputs, soil nutrient and water uptake, and play critical roles in determining sustainable plant production and consequences for ecosystem C storage. However, the effects of elevated CO₂ on root morphology and function have not been well quantified. We reveal general patterns of root trait responses to elevated CO₂ from field manipulative experiments.

North America, Europe, Oceania, Asia.

The meta-analysis approach was used to examine the effects of CO₂ elevation on 17 variables associated with root morphology, biomass size and distribution, C and N concentrations and pools, turnover and fungal colonization from 110 published studies.

Elevated CO₂ increased root length (+26.0%) and diameter (+8.4%). Elevated CO₂ also stimulated total root (+28.8%), fine root (+27.7%) and coarse root biomass (+25.3%), demonstrating strong responses of root morphology and biomass. Elevated CO₂ increased the root:shoot ratio (+8.5%) and decreased the proportion of roots in the topsoil (–8.4%), suggesting that plants expand rooting systems. In addition, elevated CO₂ decreased N concentration (–7.1%), but did not affect C concentration, and thus increased the C:N ratio (+7.8%). Root C (+29.3%) increased disproportionately relative to root N pools (+9.4%) under elevated CO₂. Functional traits were also strongly affected by elevated CO₂, which increased respiration (+58.9%), rhizodeposition (+37.9%) and fungal colonization (+3.3%).

These results suggest that elevated CO₂ promoted root morphological development, root system expansion and C input to soils, implying that the sensitive responses of root morphology and function to elevated CO₂ would increase long-term belowground C sequestration.

This paper aims to project areas of olive cultivation into future scenarios. Accordingly, we first asked the question whether global circulation models (GCMs) are able to reproduce past climatic conditions and we used historical ranges of olive cultivation as a palaeoclimate proxy.

The Mediterranean basin.

We used an ecological model, calibrated and validated for modern times, to test the reliability of a general circulation model (NCAR-CSM GCM) in reproducing past ranges of olive tree cultivation inferred from the literature, archaeo-botanical investigations and fossil pollen analyses.

The re-constructions of olive growing areas, obtained for the Medieval Climate Anomaly (MCA, 1200-1300 AD) and the Little Ice Age (LIA, 1600-1700 AD) by coupling the outputs of NCAR-CSM to the ecological model, were in agreement to those observed. Simulations of olive growing areas for future time-windows showed that a northwards expansion of the species is expected to occur by 2100.

These results demonstrate that the NCAR-CSM can provide an accurate reconstruction of past climate with results sensitive to climate forcing factors and thus, it is more likely to give reliable projections for the future. Additionally, the warming and drying conditions expected in the coming decades may determine changes across the Mediterranean basin that is unprecedented.

We hypothesized that mechanisms underlying beta diversity in rivers would differ between gradients where (1) natural stressors result in progressive species turnover with high specificity and (2) anthropogenic stressors result in the loss of specialist taxa thus giving rise to nestedness.

Great Britain, the Iberian Peninsula and the Himalayan Mountains.

We analysed five datasets describing benthic macroinvertebrates sampled along natural (elevation, salinity) and anthropogenic (acidity, metals, land use) stress gradients. Predictions were tested by fitting models relating species richness and beta-diversity components (total, turnover and nestedness dissimilarities) to putative stress intensity (i.e. the degree to which a particular environmental constraint filters species occurrence).

Stress intensity accounted for most of the variability in species richness (r² = 0.64–0.93), which declined with increasing stress. Dissimilarity in community composition between locations increased with the difference in stress intensity for all datasets. For natural stressors, beta-diversity patterns mainly reflected species turnover, whilst for anthropogenic stressors beta diversity mainly reflected nesting of subsets of species as stress intensity increased.

Our results support the hypothesis that natural and anthropogenic stressors generate contrasting patterns in beta diversity that arise through different mechanisms.

Predicted distributions of invasive species are often not congruent between their realized native and introduced ranges, but the reasons for this are rarely investigated empirically. We tested for niche shift in an invasive species using a simple framework combining environmental niche models (ENMs) and niche-limiting thermal tolerance traits.

Australia and South Africa.

The red-legged earth mite, Halotydeus destructor, native to South Africa, is a major agricultural pest in Australia and has expanded its range to areas not predictable from its native range in the last 40 years. Revisiting recently constructed ENMs for H. destructor, we select populations in both native and invasive ranges that appear to occupy different niches. We characterize thermal tolerance traits and test for acclimation patterns of cold tolerance of these H. destructor populations to test for niche shifts.

Australian populations had an increased upper thermal threshold for movement and were able to recover from cold stress more rapidly than South African populations. Australian populations also differed in trait means from the likely source population in South Africa. Acclimation patterns were conserved across ranges for most populations, with 10 °C acclimation lowering the onset of and recovery from cold tolerance and 15 °C raising them when compared with field-acclimated populations.

These results support the prediction, based on ENMs, that H. destructor may have undergone a niche shift by adapting to environmental conditions in Australia. The increase in thermal resistance has implications for how this invasive species will respond to future climate change.

Spatial patterns in biodiversity along environmental gradients are a central theme in ecology. However, the ways in which local assembly processes control changes in species turnover (β-diversity) along broader gradients have been less well documented. In this study, we aimed to elucidate factors and processes governing the altitudinal gradients in the β-diversity of woody plants and ground-dwelling oribatid mites.

Shiretoko National Park in Hokkaido, Japan.

The diversity of plants and oribatids was investigated in seven plots (each containing 10 subplots) at different altitudes, and the β-diversity of the two organism groups was calculated for each altitude. The dependence of β-diversity on the size of the species pool (γ-diversity) is an issue of long-standing importance. We therefore used null modelling, which randomly shuffles individuals among subplots while preserving the γ-diversity, the relative abundance of each species per plot and the number of individuals per subplot. This approach enabled us to estimate how much the observed β-diversity deviates from the expected β-diversity under stochastic assembly processes. Environmental data were collected to evaluate the possible effects of habitat condition/heterogeneity on community processes.

In plants, deterministic processes dominated in the low-productivity, high-altitude stands because of the finer-scale niche partitioning seen among small individuals within less-stratified stands. In the structurally developed, low-altitude stands, the community structure was more strongly affected by stochasticity, probably resulting from one-sided competition such that the canopy trees intercept the majority of light, a primary resource for plants, and therefore the small understorey individuals had limited access to light. Among the oribatids, the altitudinal gradient of β-diversity was less evident than among the studied plants. However, this nonlinearity does not support the notion that local assembly processes contribute little to the spatial pattern of β-diversity. Indeed, local-scale environmental heterogeneity favoured a more deterministic assembly of oribatids at a given altitude.

The biogeographical patterns of β-diversity are not independent of community processes and, in reality, are shaped by local stochastic/deterministic factors that change within a landscape.

Understanding the stability of realized niches is crucial for predicting the responses of species to climate change. One approach is to evaluate the niche differences of populations of the same species that occupy regions that are geographically disconnected. Here, we assess niche conservatism along thermal gradients for 26 plant species with a disjunct distribution between the Alps and the Arctic.

European Alps and Norwegian Finnmark.

We collected a comprehensive dataset of 26 arctic-alpine plant occurrences in two regions. We assessed niche conservatism through a multispecies comparison and analysed species rankings at cold and warm thermal limits along two distinct gradients corresponding to (1) air temperatures at 2 m above ground level and (2) elevation distances to the tree line (TLD) for the two regions. We assessed whether observed relationships were close to those predicted under thermal limit conservatism.

We found a weak similarity in species ranking at the warm thermal limits. The range of warm thermal limits for the 26 species was much larger in the Alps than in Finnmark. We found a stronger similarity in species ranking and correspondence at the cold thermal limit along the gradients of 2-m temperature and TLD. Yet along the 2-m temperature gradient the cold thermal limits of species in the Alps were lower on average than those in Finnmark.

We found low conservatism of the warm thermal limits but a stronger conservatism of the cold thermal limits. We suggest that biotic interactions at the warm thermal limit are likely to modulate species responses more strongly than at the cold limit. The differing biotic context between the two regions is probably responsible for the observed differences in realized niches.

Xylem structures are closely related to a tree's hydraulic efficiency and mechanical stability, both of which affect the life history and ecological strategy of a species. Although mechanical strength and hydraulic capacity can be shaped by the environment, no such associations between hydraulic efficiency and climatic variables have been reported across a wide range of tree species.

Yunnan, south-west China.

We compiled a data set for vessel density, vessel diameter (D), potential hydraulic conductivity (K_p), wood density (WD), modulus of rupture (MOR) and modulus of elasticity (MOE) from 316 angiosperm tree species. Our objective was to examine the correlations among xylem traits and climatic variables. We hypothesized that both hydraulic efficiency and mechanical strength would vary along climatic gradients, but that a trade-off would occur between them.

All xylem traits varied significantly across species, but the magnitudes of variation were greater for vessel traits than for mechanical properties. Values for K_p and D increased with mean annual temperature (MAT) in both evergreen and deciduous trees, but they were significantly correlated with aridity index (AI) in evergreen species only. Both WD and MOR were significantly correlated with MAT only in evergreen trees. MOR decreased with increasing AI in the evergreens, but not in deciduous trees. These findings indicated that xylem development in evergreens is more sensitive to environmental changes than in deciduous trees. However, stem hydraulic traits are independent of mechanical properties.

Consistent with our hypothesis, both hydraulic efficiency and mechanical strength of angiosperm trees are influenced by the environment, with temperature having a more important effect on hydraulic efficiency than precipitation. However, no trade-off exists between efficiency and strength. This absence of a link is explained because angiosperms have xylem tissue that specifically functions in either mechanical strength or water transport.

To examining whether niche breadth, niche position and a compound measure of the two variables help to explain the range sizes of forest vascular plants.

Deciduous forests in two regions of Germany, the Weser–Elbe region and Bavarian Alps.

We compiled range size data for vascular plant species (30 in the Weser–Elbe region, 35 in the Bavarian Alps) on regional, national and continental scales by determining the area of occupancy (number of occupied grid cells) in both regions, Germany and Eurasia. Estimates of realized niche breadth and niche position (ecological optimum) for soil pH and light were based on measurements in 46 sites for all species. Frequency distributions of pH values on regional and national scales served to calculate the ‘available niche breadth’, i.e. niche breadth values corrected for the different availabilities of pH values in the region and country.

Regional range size in the Weser–Elbe region increased with increasing niche breadth for both soil pH and light and with decreasing pH niche position. pH niche breadth was positively correlated to national range size in the Weser–Elbe region and to Eurasian range size in the Bavarian Alps. In the latter region, all other relationships with range size were (partly marginally) non-significant. Available niche breadth was generally more closely related to the regional and national distribution of species in both regions than either niche breadth or position alone.

Niche breadth and position performed well as predictors of range size only for soil pH and in the Weser–Elbe region, which shows more homogeneous environmental conditions than the Bavarian Alps. If the frequencies of different ranges of pH values can be quantified, the calculation of available niche breadth for soil pH appears to be a promising approach for assessing the possible effects of niche variables on the range sizes of species.

Ecological niche modelling (ENM) and species distribution modelling (SDM) have been used extensively to study biogeographic and macroecological patterns of terrestrial fauna and flora. Few studies to date have applied ENM and SDM to marine ecosystems, and those that have treated the marine environment as a two-dimensional space owing to limitations of the implementations of current ENM/SDM tools. For many marine organisms, ENM/SDM should be performed in three-dimensional space, taking into account latitude, longitude and depth. We present a case study demonstrating a strategy for three-dimensional ENM/SDM.

Open ocean; global.

We decompose the three-dimensional structure of marine environmental and species occurrence data into a series of two-dimensional spaces using an easy-to-implement transformation, after which existing ENM/SDM tools can be used to analyse the data. We demonstrate our approach by modelling the potential distribution of a deep-sea-dwelling jellyfish with two commonly used algorithms. Potential effects of missing data and spatial sampling biases were assessed using resampling approaches.

We demonstrate that it is feasible to derive predictive models of three-dimensional distributions of marine species using existing software tools developed with two-dimensional terrestrial situations in mind. The strategy presented here allowed us to model the distribution of a jellyfish species that inhabits the deep sea. We assessed the effects of missing occurrence data and spatial bias of occurrence data, and found that interpolation among occurrence data-points and extrapolation into unsampled conditions present distinct challenges that may require different modelling algorithms and interpretations.

Our modified ENM/SDM approach is straightforward, and can be used to model situations that have heretofore been beyond the reach of ENM/SDM applications. In particular, geographic distributions and ecological niches of organisms inhabiting three-dimensional habitats such as water columns in marine and freshwater environments can be modelled using the framework presented here.

To provide the first analysis of predictors of both establishment and spread, both within and across taxa, for all vertebrate taxa within a region. We used Florida, USA, as our study system because it has a well-documented history of introduction and invasion, and is a hotspot for biological invasions.

Florida, USA.

We analysed non-indigenous species (NIS) data from peninsular Florida – which included both successful and unsuccessful introductions from all vertebrate classes – to determine the best predictors of both establishment and spread for fish (65 species), herpetofauna (63 species), birds (71 species) and mammals (25 species). We used 10 variables proposed to be associated with the establishment and spread of NIS: body mass, geographic origin, reproductive rate, diet generalism, native-range size, latitude of native range, number of NIS present at date of introduction, presence of NIS congeners, morphological proximity to other NIS (in terms of body mass) and propagule pressure. A multimodel selection process was used with an information-theoretic approach to determine the best fit models for predicting establishment and spread of NIS. We selected a priori plausible predictive models for establishment and spread.

Large native-range size and small body mass best predicted establishment of non-indigenous herpetofauna. The presence of NIS congeners had the largest positive effect on the establishment of non-indigenous fish. For mammals, the number of NIS present at the time of introduction best explained establishment. No single model best explained bird establishment. For all taxa but birds, the number of NIS present at time of introduction was included in at least one of the best-supported models for explaining spread.

Our analyses suggest that predictors of establishment and spread differ across vertebrate taxa at the scale studied. Most predictive variables can be interpreted as measures of competitive interactions among species.

Evolutionary radiations into novel areas or niches require innovative adaptations. However, rapid subsequent changes in these novel conditions might demand rapid re-adaptations to secure population persistence and prevent extinction. We propose that reptilian viviparity (live birth) is consistent with such a scenario. Using the Liolaemus lizard radiation, we investigate the hypotheses that historical invasions of cold climates have been permitted by transitions to viviparity, and that this parity mode is irreversible. Then, we investigate whether these combined factors restrict viviparous lizards to cold climates, and hence, whether viviparous species are particularly threatened by climate change.

South America.

We employ phylogenetic analyses to investigate evolutionary transitions in reproductive modes and their consequences for environmental restrictions in viviparous lizards. We then employ climatic projections to predict the impact of climate change on the future persistence of these organisms.

The oviparity-to-viviparity transition is consistently associated with colonization of cold climates, and appears to be irreversible. Since viviparity seems less viable (compared with oviparity) in warm climates, species that evolve viviparity in cold climates are likely to remain adaptively constrained to such environments. Therefore, upward–poleward advances of climate warming will cause severe shifts and contractions of viviparous species ranges, threatening major extinctions over the next half century.

Viviparity has been largely responsible for the successful radiation of Liolaemus into cold climates, but since this adaptation is predominantly viable in these environments and is unlikely to re-evolve into oviparity, viviparity may prove to be an evolutionary dead-end for lizards facing rapid climate change.

Temperature influences most components of animal ecology and life history – but what kind of temperature? Physiologists usually examine the influence of body temperatures, while biogeographers and macroecologists tend to focus on environmental temperatures. We aim to examine the relationship between these two measures, to determine the factors that affect lizard body temperatures and to test the effect of both temperature measures on lizard life history.

World-wide.

We used a large (861 species) global dataset of lizard body temperatures, and the mean annual temperatures across their geographic ranges to examine the relationships between body and mean annual temperatures. We then examined factors influencing body temperatures, and tested for the influence of both on ecological and life-history traits while accounting for the influence of shared ancestry.

Body temperatures and mean annual temperatures are uncorrelated. However, accounting for activity time (nocturnal species have low body temperatures), use of space (fossorial and semi-aquatic species are ‘colder’), insularity (mainland species are ‘hotter’) and phylogeny, the two temperatures are positively correlated. High body temperatures are only associated with larger hatchlings and increased rates of biomass production. Annual temperatures are positively correlated with clutch frequency and annual longevity, and negatively correlated with clutch size, age at first reproduction and longevity.

Lizards with low body temperatures do not seem to have ‘slower’ life-history attributes than species with high body temperatures. The longer seasons prevalent in warm regions, and physiological processes that operate while lizards are inactive (but warm enough), make environmental temperatures better predictors of lizard life-history variation than body temperatures. This surprisingly greater effect of environmental temperatures on lizard life histories hints that global warming may have a profound influence on lizard ecology and evolution.

Climate change and loss of apex predators can affect ecosystem structure and function through modified limitation processes. We investigated, on a continental scale, whether mesopredator abundance is limited from the top down by large predators, as predicted by the mesopredator release hypothesis, or by bottom-up factors. The mesopredator in focus is the red fox Vulpes vulpes, a key predator in many ecosystems due to its strong effects on prey abundance.

Europe and northern Asia.

Data on red fox density were compiled from published papers and reports. For each site, we collated presence–absence data on large carnivores (Lynx lynx, Canis lupus, Canis aureus) and remote sensing data for factors potentially related to bottom-up limitation (winter severity, summer temperature, human density, primary productivity, tree cover). The data were analysed through structural equation modelling.

The presence of lynx had a direct negative effect on red foxes, suppressing fox abundance. Also winter severity had a negative effect on red fox abundance, and in Eurasia as a whole this effect was partially mediated through lynx. Within the lynx distribution range, winter severity was the only bottom-up factor significantly affecting red fox abundance. Outside the lynx distribution range, primary productivity, summer temperature and human density had a positive effect on red fox abundance.

Our results show that apex predators can limit mesopredator abundance on a continental scale, thus supporting the mesopredator release hypothesis. Winter severity also affected red fox abundance, partially due to an interaction between lynx and winter conditions. On the continental scale a complex network of processes operates with varying effects depending on mediation processes. Our results imply that apex predators can have an important effect on ecosystem structure by limiting mesopredator abundance, and we suggest that apex predators may dampen increases in mesopredator abundance driven by global warming.

Niche conservatism, or the extent to which niches are conserved across space and time, is of special concern for the study of non-native species as it underlies predictions of invasion risk. Based on the occurrence of 28 non-native birds in Europe, we assess to what extent Grinnellian realized niches are conserved during invasion, formulate hypotheses to explain the variation in observed niche changes and test how well species distribution models can predict non-native bird occurrence in Europe.

Europe.

To quantify niche changes, a recent method that applies kernel smoothers to densities of species occurrence in a gridded environmental space was used. This corrects for differences in the availability of environments between study areas and allows discrimination between ‘niche expansion’ into environments new to the species and ‘niche unfilling’, whereby the species only partially fills its niche in the invaded range. Predictions of non-native bird distribution in Europe were generated using several distribution modelling techniques.

Niche overlap between native and non-native bird populations is low, but niche changes are smaller for species having a higher propagule pressure and that were introduced longer ago. Non-native birds in Europe occupy a subset of the environments they inhabit in their native ranges. Niche expansion into novel environments is rare for most species, allowing species distribution models to accurately predict invasion risk.

Because of the recent nature of most bird introductions, species occupy only part of the suitable environments available in the invaded range. This signals that apart from purely ecological factors, patterns of niche conservatism may also be contingent on population-specific historical factors. These results also suggest that many claims of niche differences may be due to a partial filling of the native niche in the invaded range and thus do not represent true niche changes.

Plant roots are crucial for water and nutrient absorption, but large-scale patterns and underlying mechanisms of root trait variation and evolution are poorly understood. Here we quantify the degree of variation in functional traits for the first-order roots across large geographical scales and examine the potential mechanisms underlying these patterns.

China.

We collected first-order roots (stream-based ordering system) and leaf samples of 65 tree species in six forests from subtropical to temperate zones and determined the key morphological, architectural and chemical traits.

We found wider variation in root morphology in the subtropical zone, where species with very thick and very thin first-order roots coexisted, in contrast with narrower variation among species in the temperate zone. Inconsistent with the predictions of trait economics spectrum, root nitrogen concentration was uncorrelated with root morphology. Furthermore, average root diameter at the plant family level decreased markedly with their divergence time in both piecewise and phylogenetic independent contrast regression analyses.

Higher variation in root morphology in the subtropical zone appears to result from the coexistence of tree species with thick and thin roots, probably because of a more favourable water supply. Patterns of root evolution towards thinner roots were driven by both phylogeny and possible adaptation of newly diverged species to drier habitats from mid to late Cretaceous. Our findings reveal contrasting selection pressures at the root and leaf level in different climatic zones during plant evolution, and suggest that a single vector of variation linking morphology to resource acquisition or life span (‘trait economics’) remains elusive in roots.

Despite increasing evidence for plant growth often being limited by sink (meristem) activity rather than source (photosynthesis) activity, all currently available dynamic global vegetation models (DGVMs) simulate plant growth via source-limited processes. For a given climatic region, this may lead to an overestimation of carbon stock per unit surface area, particularly if a model fails to correctly predict forest cover. Our aim is to improve the Lund–Potsdam–Jena (LPJ) DGVM by replacing the source-limited (SoL) tree growth algorithm by a sink-limited (SiL) one.

Our analysis focuses on the cold tree line at high latitudes and altitudes. We study two altitudinal transects in the Swiss Alps and the northern tree line.

We limit annual net primary productivity of the LPJ DGVM by an algorithm based on the annual sum of growing degree-days (GDD), assuming that maximum plant growth is reached asymptotically with increasing GDD.

Comparing simulation results with observational data, we show that the locations of both the northern and the alpine tree line are estimated more accurately when using a SiL algorithm than when using the commonly employed SoL algorithm. Also, simulated carbon stocks decrease in a more realistic manner towards the tree line when the SiL algorithm is used. This has far-reaching implications for estimating and projecting present and future carbon stocks in temperature-limited ecosystems.

In the range of 60–80° N over Europe and Asia, carbon stored in vegetation is estimated to be c. 50% higher in the LPJ standard version (LPJ-SoL) compared with LPJ-SiL, resulting in a global difference in estimated biomass of 25 Pg (c. 5% of the global terrestrial standing biomass). Similarly, the simulated elevation of the upper tree line in the European Alps differs by c. 400 m between the two model versions, thus implying an additional overestimation of carbon stored in mountain forests around the world.

To investigate the importance of autumn phenology in controlling interannual variability of forest net ecosystem productivity (NEP) and to derive new phenological metrics to explain the interannual variability of NEP.

North America and Europe.

Flux data from nine deciduous broadleaf forests (DBF) and 13 evergreen needleleaf forests (ENF) across North America and Europe (212 site-years) were used to explore the relationships between the yearly anomalies of annual NEP and several carbon flux based phenological indicators, including the onset/end of the growing season, onset/end of the carbon uptake period, the spring lag (time interval between the onset of growing season and carbon uptake period) and the autumn lag (time interval between the end of the carbon uptake period and the growing season). Meteorological variables, including global shortwave radiation, air temperature, soil temperature, soil water content and precipitation, were also used to explain the phenological variations.

We found that interannual variability of NEP can be largely explained by autumn phenology, i.e. the autumn lag. While variation in neither annual gross primary productivity (GPP) nor in annual ecosystem respiration (R_e) alone could explain this variability, the negative relationship between annual NEP and autumn lag was due to a larger R_e/GPP ratio in years with a prolonged autumn lag. For DBF sites, a longer autumn lag coincided with a significant decrease in annual GPP but showed no correlation with annual R_e. However, annual GPP was insensitive to a longer autumn lag in ENF sites but annual R_e increased significantly.

These results demonstrate that autumn phenology plays a more direct role than spring phenology in regulating interannual variability of annual NEP. In particular, the importance of respiration may be potentially underestimated in deriving phenological indicators.

At a global scale, the relationship of leaf longevity (LL) to mean annual temperature (MAT) is positive for deciduous species but negative for evergreen species. The aim of this paper is to understand the mechanisms underlying these contrasting patterns of leaf longevity, from a cost–benefit perspective.

Global.

We tested our hypothesis that contrasting LL–MAT relationships in evergreen and deciduous species result from differing adaptations to variation in the length of the annual favourable period. We defined f as the portion of the year when monthly temperature and water availability were favourable. We examined whether the contrasting LL patterns with MAT can be also seen with f. Next, we calculated the optimal LL that maximizes carbon gain per unit time across a range of f.

The contrasting LL patterns across MAT were also found across f. Our optimization model successfully reproduced the contrasting LL patterns across f for the evergreen (LL longer than 1 year) and deciduous plants. The model shows that longer LL is required to maximize carbon gain for evergreen plants in shorter f, while LL of deciduous plants decreases with decreasing f. Without any a priori trait association, the model reproduced the well-known LL–leaf mass per area (LMA) relationship. The model also reproduced observed shifts in LL–LMA relationships across MAT or f. Evergreen leaves in long f need greater LMA to maintain LL than those in shorter f.

Observed contrasting LL–MAT patterns in deciduous and evergreen species can be reproduced via the simple rule of maximizing carbon gain across different lengths of favourable periods. Our model provides a mechanistic explanation for the empirical global patterns of several key leaf traits and their relationships.

We tested the energy and metabolic theories for explaining diversity patterns of crustacean zooplankton in Canadian lakes, and evaluated the influence of regional and local environments on community structure.

The 1665 studied lakes are distributed across Canada in 47 ecoprovinces.

Our database included the occurrence of 83 pelagic crustacean species. The regional species richness in each ecoprovince was estimated using the average local species richness per lake and the first-order jackknife diversity index. Using a principal component plot and forward selection in a multiple regression we identified the most important predictors of regional species richness estimates. We tested the predictions of the species richness-energy hypothesis using climatic variables at regional scale, and of the metabolic theory using the inverse of air temperature. To evaluate the influence of regional and local environmental drivers, we carried out a redundancy analysis between crustacean species occurrences and regional climate and lake environmental factors on a subset of 458 lakes.

Estimates of pelagic crustacean species richness in Canadian ecoprovinces varied from 3 to 10 species per lake (average local species richness) or 8 to 52 species per ecoprovince (Jackknife diversity index). Our study fully supports the species richness-energy hypothesis and partially the metabolic theory. Mean daily global solar radiation was the most important regional predictor, explaining 51% of the variation in the regional species richness among ecoprovinces. Together, regional climate and local lake environment accounted for 31% of the total variation in community structure. Regional-scale energy variables accounted for 24% of the total explained variation, whereas local-scale lake conditions had less influence (2%).

The richness-energy theory explains diversity patterns of freshwater crustacean zooplankton in Canadian ecoprovinces. Solar radiation is the best predictor explaining regional species richness in ecoprovinces and community structure of pelagic crustaceans in Canadian lakes.

We compared the upper limits of 18 deciduous tree species with respect to elevation in Switzerland and latitude in Europe. We hypothesized that species would exhibit the same relative positions along elevation and latitude, which can be expected if species have reached their thermal cold limit along both gradients.

Europe and Switzerland.

We developed a method to identify a least biased estimate of the elevational and latitudinal cold temperature limits of species and for comparing the relative rank positions with respect to these two limits. We applied an algorithm to calculate the elevation of the potential tree line for each point in the gridded landscape of Europe and Switzerland. For each occurrence of each species, the elevation was extracted from digital elevation models. The vertical distance between the elevation of the potential regional climatic tree line and the uppermost species occurrences was calculated and used for comparisons between elevation and latitude.

We found a strong relationship between the thermal latitudinal and elevational distances of species’ cold limits to the potential tree line, with only marginally significantly different rank positions (P = 0.057) detected along elevational and latitudinal gradients. A first group of nine species showed very similar thermal distances to the potential tree lines along elevation and latitude. Among these species, eight showed a significant decrease in their elevational limit towards high latitudes across mountainous regions of Europe. A second group of seven species occupied a climatic niche closer to the tree line at the edge of their latitudinal range, and only two species did not fill their thermal niche.

Our study provides support for the common concept of a species range–environment equilibrium. Notably, we did not detect a stronger deviation for the filling of thermal niches at latitudinal limits compared with elevational limits, although the former involves a species covering a much greater geographic distance.

The increased incidence of large fires around much of the world in recent decades raises questions about human and non-human drivers of fire and the likelihood of increased fire activity in the future. The purpose of this paper is to outline a conceptual framework for examining where human-set fires and feedbacks are likely to be most pronounced in temperate forests world-wide and to establish and test a methodology for evaluating this framework using palaeoecological records.

Tasmania, north-western USA, southern South America and New Zealand.

We outline a conceptual framework for predicting the sensitivity of ecosystems to human impacts on fire regimes and then use a circum-Pacific comparison of existing historical reconstructions of fire, climate, human settlement and vegetation to evaluate this approach.

Previous research investigating important controls on fire activity shows that the sensitivity of temperate ecosystems to human-set fires is modulated by the frequency of natural fire occurrence, fuel moisture and fuel type and availability. Palaeoecological data from four temperate regions suggest that the effects of anthropogenic burning are greatest where fire is naturally rare, vegetation is poorly adapted to fire and fuel biomass is abundant and contiguous. Alternatively, where fire activity is naturally high and vegetation is well adapted to fire, evidence of human influence on fire and vegetation is less obvious.

Palaeofire records suggest that the most dynamic and persistent ecosystem transitions occur where human activities increase landscape flammability through fire–vegetation feedbacks. Rapid forest transitions in biomass-rich ecosystems such as New Zealand and areas of Tasmania and southern South America illustrate how landscapes experiencing few fires can shift past tipping points to become fire-prone landscapes with new alternative stable state communities. Comparisons of palaeoecological data from different regions with similar biophysical gradients but different human settlement histories can provide new opportunities for understanding ecosystem vulnerability to fire–climate–human interactions.

We incorporate diversity-dependent colonization and extinction rates into process-based models of species geographic range dynamics to explore their effects on species richness gradients, and on extent and occupancy of species ranges. In particular, we investigate whether diversity dependence promotes or inhibits the emergence of mid-domain effects (MDEs) in homogeneous environments.

A theoretical one-dimensional domain.

We formulated diversity-independent (DI) and diversity-dependent (DD) models that simulated colonization, local extinction and speciation within a homogenous domain. In the DD model, colonization and extinction probabilities were functions of diversity, whereas in the DI model, they were constants. For a wide range of parameter values, we examined local and regional species richness gradients and species range size frequency distributions (RSFDs).

In contrast to the DI model, for which MDEs only occurred in a very narrow parameter range, the DD model generated a MDE in regional richness (range overlap) that was robust to colonization and extinction parameters, over a broad range of speciation rates. However, neither model could produce gradients in local richness (patch occupancy). The DD model also produced more realistic RSFDs than the DI model. In the latter, all species generally either became highly pandemic or went globally extinct, depending on the balance of colonization and extinction probabilities.

Diversity-dependent colonization and extinction rates can have strong effects on species richness gradients and distributions of range extent and occupancy. Models with such diversity dependence amplify MDEs in regional richness, but largely eliminate MDEs in local richness, relative to DI models. DD models also generate more realistic RSFDs. These findings suggest that diffuse species interactions can strongly influence patterns of range size and overlap, but also that environmental gradients are likely to be necessary to explain many species richness patterns in nature, which exhibit both local and regional diversity gradients.

Intercomparison of mechanistic and empirical models is an important step towards improving projections of potential species distribution and abundance. We aim to compare suitability and productivity estimates for a well-understood crop species to evaluate the strengths and weaknesses of mechanistic versus empirical modelling.

South Africa.

We compared four habitat suitability models for dryland maize based on climate and soil predictors. Two were created using maximum entropy (MAXENT), the first based on national crop distribution points and the second based only on locations with high productivity. The third approach used a generalized additive model (GAM) trained with continuous productivity data derived from the satellite normalized difference vegetation index (NDVI). The fourth model was a mechanistic crop growth model (DSSAT) made spatially explicit. We tested model accuracy by comparing the results with observed productivity derived from MODIS NDVI and with observed suitability based on the current spatial distribution of maize crop fields.

The GAM and DSSAT results were linearly correlated to NDVI-measured yield (R² = 0.75 and 0.37, respectively). MAXENT suitability values were not linearly related to yield (R² = 0.08); however, a MAXENT model based on occurrences of high-productivity maize was linearly related to yield (R² = 0.62). All models produced crop suitability maps of similarly good accuracy (Kappa = 0.73–75).

These findings suggest that empirical models can achieve the same or better accuracy as mechanistic models for predicting both suitability (i.e. species range) and productivity (i.e. species abundance). While MAXENT could not predict productivity across the species range when trained on all occurrences, it could when trained with a high-productivity subset, suggesting that ecological niche models can be adjusted to better correlate with species abundance.

Previous studies comparing conditions of high- versus low-resource environments have pointed at differences in key traits that would allow aliens to perform better than natives under high-resource conditions. We generalize and test the robustness of this idea by exploring how trait differentiation between aliens and natives changes along continuous resource gradients.

Global.

We constructed a database of three leaf traits (specific leaf area, SLA; photosynthetic capacity, A_mass; leaf nitrogen content, N_mass) that are important for carbon capturing strategies in plants. The database includes 2448 native and 961 alien species over 88 locations world-wide. Using rank correlations and mixed-effect linear models, we assessed the relations between plant traits and climatic, edaphic and human disturbance gradients. Then we determined how the differences in traits between natives and aliens changed along the same gradients.

Across all environments, aliens were found to have higher SLA, N_mass and A_mass than natives. These differences were observed both globally and when controlling for co-occurrence. Also, higher average trait values were found in higher resource supply environments. However, trait differences between natives and aliens remained constant along the evaluated environmental and disturbance gradients. When compared in a multidimensional trait space defined by the leaf economics spectrum, co-occurring aliens and natives showed no between-group differences and no relation with any of the evaluated gradients.

We suggest that although increased resource availability is positively related to higher carbon capture strategies (determined via higher plant leaf trait values), these benefits remain the same for aliens and natives. Therefore, we conclude that high-resource environments do not specifically cause aliens to outperform natives with respect to carbon capture, or at least not more than in other environments.

Ecologists have generally agreed that β-diversity is driven at least in part by ecological processes and mechanisms of community assembly and is a key determinant of global patterns of species richness. This idea has been challenged by a recent study based on an individual-based null model approach, which aims to account for the species pool. The goal of the present study is twofold: (1) to analyse data sets from different parts of the world to determine whether there are significant latitude–β-diversity gradients after accounting for the species pool, and (2) to evaluate the validity of the null model.

Global.

A total of 257 forest plots, each being 0.1 ha in size and having 10 0.01-ha subplots, were used. We conducted four sets of analyses. A modified version of Whittaker's β-diversity index was used to quantify β-diversity for each forest plot. A randomization procedure was used to determine expected β-diversity.

The number of individuals per species, which characterizes species abundance distribution, alone explains 56.8–84.2% of the variation in observed β-diversity. Species pool (γ-diversity) explained only an additional 2.6–15.2% of the variation in observed β-diversity. Latitude explains 18.6% of the variation in raw β deviation in Gentry's global data set, and explains 11.0–11.6% of the variation in standardized β deviation in the global and three regional analyses. Latitude explains 33.2–46.2% of the variation in the number of individuals per species.

Species abundance distribution, rather than species pool size, plays a key role in driving latitude–β-diversity gradients for β-diversity in local forest communities. The individual-based null model is not a valid null model for investigating β-diversity gradients driven by mechanisms of local community assembly because the null model incorporates species abundance distributions, which are driven by mechanisms of local community assembly and in turn generate β-diversity gradients.

One of the main gaps in the assessment of biodiversity is the lack of a unified framework for measuring its taxonomic and functional facets and for unveiling the underlying patterns.

Europe, 25 large river basins.

Here, we develop a decomposition of functional β-diversity, i.e. the dissimilarity in functional composition between communities, into a functional turnover and a functional nestedness-resultant component.

We found that functional β-diversity was lower than taxonomic β-diversity. This difference was driven by a lower functional turnover compared with taxonomic turnover while the nestedness-resultant component was similar for taxonomic and functional β-diversity.

Fish faunas with different species tend to share the same functional attributes. The framework presented in this paper will help to analyse biogeographical patterns as well as to measure the impact of human activities on the functional facets of biodiversity.

Determining the relative influence of niche-based and neutral processes in driving the spatial turnover of community composition is a central challenge in community ecology. Spatial patterns of functional turnover, or functional beta diversity, may capture important signals of niche-based assembly processes, but these patterns have not been quantified for communities across broad geographic and environmental gradients. Here, we analyse continental-scale patterns of species and functional beta diversity in relation to space and the environment to assess the relative importance of niche-based and neutral community assembly mechanisms.

Eastern North America.

We use a continental-scale forest plot dataset and functional trait data to quantify spatial patterns of species and functional beta diversity. We use redundancy analysis-based variance partitioning to evaluate the influence of space, soil and climate on beta-diversity metrics. We use a null model approach to test for non-random functional beta diversity given the observed patterns of species turnover across spatial scales.

Species and functional beta diversity increased with increasing geographic distance (i.e. distance decay of community similarity). Results of variance partitioning analysis show that species and functional beta diversity were spatially structured and significantly related to environmental, particularly climatic, variation. Results of null model analysis show that functional beta diversity was lower than expected based on species turnover at fine scales (< 600 km) and higher than expected at broad scales (> 1800 km).

The observed patterns of functional beta diversity support a niche-based model of community assembly, driven by the deterministic filtering of species across environmental gradients based on their functional traits.

Vegetation optical depth (VOD) is an indicator of the water content of both woody and leaf components in terrestrial aboveground vegetation biomass that can be derived from passive microwave remote sensing. VOD is distinct from optical vegetation remote sensing data such as the normalized difference vegetation index in that it is: (a) less prone to saturation in dense canopies; (b) sensitive to both photosynthetic and non-photosynthetic biomass; and (c) less affected by atmospheric conditions. Our primary objective was to analyse a recently developed long-term VOD record and investigate how the vegetation water content of various land-cover types responded to environmental changes and human influences from 1988 to 2008.

Global.

We first conducted Mann–Kendall trend tests on annual average VOD to identify regions with significant changes over the period 1988–2008. To diagnose the underlying cause of the observed changes, patterns for these identified regions were further compared with independent datasets of precipitation, crop production, deforestation and fire occurrence.

(1) Over grassland and shrubland, VOD patterns corresponded strongly to temporal precipitation patterns. (2) Over croplands, annual average VOD showed a general increase that corresponded to reported crop production patterns and was attributed to a combination of precipitation patterns and agricultural improvements. (3) Over humid tropical forest, the spatial pattern of VOD decline agrees well with deforestation patterns; the 2005 Amazon drought corresponded with a temporary VOD decrease. (4) Over boreal forests, regional VOD declines are attributed to a combination of fires and clear cutting.

Passive microwave remote sensing of VOD can be used to monitor global changes in total aboveground vegetation water content and biomass over various land-cover types. This new observational record can help in hydrological, agricultural, ecological and climate change studies, and provides new insights into large-scale vegetation change and its drivers.

To evaluate the climate sensitivity of model-based forest productivity estimates using a continental-scale tree-ring network.

Europe and North Africa (30–70° N, 10° W–40° E).

We compiled close to 1000 annually resolved records of radial tree growth for all major European tree species and quantified changes in growth as a function of historical climatic variation. Sites were grouped using a neural network clustering technique to isolate spatiotemporal and species-specific climate response patterns. The resulting empirical climate sensitivities were compared with the sensitivities of net primary production (NPP) estimates derived from the ORCHIDEE-FM and LPJ-wsl dynamic global vegetation models (DGVMs).

We found coherent biogeographic patterns in climate response that depend upon (1) phylogenetic controls and (2) ambient environmental conditions delineated by latitudinal/elevational location. Temperature controls dominate forest productivity in high-elevation and high-latitude areas whereas moisture sensitive sites are widespread at low elevation in central and southern Europe. DGVM simulations broadly reproduce the empirical patterns, but show less temperature sensitivity in the boreal zone and stronger precipitation sensitivity towards the mid-latitudes.

Large-scale forest productivity is driven by monthly to seasonal climate controls, but our results emphasize species-specific growth patterns under comparable environmental conditions. Furthermore, we demonstrate that carry-over effects from the previous growing season can significantly influence tree growth, particularly in areas with harsh climatic conditions – an element not considered in most current-state DGVMs. Model–data discrepancies suggest that the simulated climate sensitivity of NPP will need refinement before carbon-cycle climate feedbacks can be accurately quantified.

Questions about the relative importance of top-down and bottom-up regulation of populations have long occupied ecologists. Recent work has explored how the strength of consumer- versus resourced-based limitation can shift in strength across both time and space, and indicates the need for larger scales of study across spatial resource gradients. A better understanding of how climate gradients affect population regulation is integral to conservation ecology, particularly in the context of global change.

California.

We examined the effects of bottom-up (water limitation) and top-down (small mammal herbivory) forces on valley oak (Quercus lobata) sapling recruitment. We used existing literature on valley oak planting experiments and field data on natural populations at 26 sites across the species' distribution to explore how the factors regulating oak populations shift along a regional precipitation gradient. This approach allowed us to study time-integrated ecological processes in a long-lived system that would be impossible to manipulate in a strictly experimental setting.

We found that the relative importance of top-down and bottom-up forces in limiting oak recruitment shifted across the precipitation gradient, with small mammalian herbivory (top-down) more important as long-term mean site precipitation increased, and factors related to moisture stress (bottom-up) increasingly important at lower precipitation levels.

Our findings emphasize the importance of expanding the spatial scale of studies on population and community dynamics in order to better understand patterned variation in the multiple factors regulating them across resource gradients.

It has been suggested that on a global scale, fire activity changes along the productivity/aridity gradient following a humped relationship, i.e. the intermediate fire–productivity hypothesis. This relation should be driven by differing relative roles of the main fire drivers (weather and fuel) along the productivity gradient. However, the full intermediate fire–productivity model across all world ecosystems remains to be validated.

The entire globe, excluding Antarctica.

To test the intermediate fire–productivity hypothesis, we use the world ecoregions as a spatial unit and, for each ecoregion, we compiled remotely sensed fire activity, climate, biomass and productivity information. The regression coefficient between monthly MODIS fire activity and monthly maximum temperature in each ecoregion was considered an indicator of the sensitivity of fire to high temperatures in the ecoregion. We used linear and generalized additive models to test for the linear and humped relationships.

Fire occurs in most ecoregions. Fire activity peaked in tropical grasslands and savannas, and significantly decreased towards the extremes of the productivity gradient. Both the sensitivity of fire to high temperatures and above-ground biomass increased monotonically with productivity. In other words, fire activity in low-productivity ecosystems is not driven by warm periods and is limited by low biomass; in contrast, in high-productivity ecosystems fire is more sensitive to high temperatures, and in these ecosystems, the available biomass for fires is high.

The results support the intermediate fire–productivity model on a global scale and suggest that climatic warming may affect fire activity differently depending on the productivity of the region. Fire regimes in productive regions are vulnerable to warming (drought-driven fire regime changes), while in low-productivity regions fire activity is more vulnerable to fuel changes (fuel-driven fire regime changes).

To estimate the concentrations, stoichiometry and storage of soil microbial biomass carbon (C), nitrogen (N) and phosphorus (P) at biome and global scales.

Global.

We collected 3422 data points to summarize the concentrations and stoichiometry of C, N and P in soils, soil microbial biomass at global and biome levels, and to estimate the global storage of soil microbial biomass C and N.

The results show that concentrations of C, N and P in soils and soil microbial biomass vary substantially across biomes; the fractions of soil elements C, N and P in soil microbial biomass are 1.2, 2.6 and 8.0%, respectively. The best estimates of C:N:P stoichiometry for soil elements and soil microbial biomass are 287:17:1 and 42:6:1, respectively, at global scale, and they vary in a wide range among biomes. The vertical distribution of soil microbial biomass follows the distribution of roots up to 1 m depth.

The global storage of soil microbial biomass C and N were estimated to be 16.7 Pg C and 2.6 Pg N in the 0–30 cm soil profiles, and 23.2 Pg C and 3.7 Pg N in the 0–100 cm soil profiles. We did not estimate P in soil microbial biomass due to insufficient data and insignificant correlation between soil total P and climate variables used for spatial extrapolation. The spatial patterns of soil microbial biomass C and N were consistent with those of soil organic C and total N, i.e. high density in northern high latitude, and low density in low latitudes and the Southern Hemisphere.

Elevational Rapoport's rule, proposed in 1992 by Stevens, predicts that species ranges on mountains become larger in elevational extent with increasing elevation. Here we test this prediction using 160 datasets of range size measured by maximum elevational extents for bats, birds, frogs, non-volant small mammals, reptiles, and salamanders from mountains around the globe.

Mountains distributed globally and spanning 36.5° S to 48.2° N.

We compare three methods: (1) the Stevens method, which uses the average range size of all species within each elevational band (100-m bands); (2) the midpoint method, which uses the average range size of species whose midpoints occur in each elevational band; and (3) a quartile method that examines the distribution of only the smallest ranges (less than one-quarter of the mountain height) to see if their frequency distribution is negatively related to elevation.

Support for the elevational Rapoport's rule was weak across all groups of montane vertebrates. For the Stevens method, the mean r² value was 0.32, and strong support (positive relationship, r² value > 0.50) was detected in 40% of the studies, ranging from 20% for salamanders to 57% for frogs. For the midpoint method, the mean r² value was 0.06, and none of the datasets showed strong support. For the quartile method, the mean r² value was 0.26, and strong support (negative relationship, r² value > 0.40) was detected in 38% of the studies, ranging between 10.5% in salamanders and 58% in reptiles.

Across vertebrates, and within the literature for plants and invertebrates, more empirical studies find a lack of trend than the predicted trend of increasing range size with increasing elevation. Thus, elevational Rapoport's rule is not a consistently predictive pattern for understanding montane patterns in range size.

The distributions of many organisms are spatially autocorrelated, but it is unclear whether including spatial terms in species distribution models (SDMs) improves projections of species distributions under climate change. We provide one of the first comparative evaluations of the ability of a purely spatial SDM, a purely non-spatial SDM and a SDM that combines spatial and environmental information to project species distributions across eight millennia of climate change.

Eastern North America.

To distinguish between the importance of climatic versus spatial explanatory variables we fit three Bayesian SDMs to modern occurrence data for Fagus and Tsuga, two tree genera whose distributions can be reliably inferred from fossil pollen: a spatially varying intercept model, a non-spatial model with climatic variables and a spatially varying intercept plus climate model. Using palaeoclimate data with a high temporal resolution, we hindcasted the SDMs in 1000-year time steps for 8000 years, and compared model projections with palynological data for the same periods.

For both genera, spatial SDMs provided better fits to the calibration data, more accurate predictions of a hold-out validation dataset of modern trees and higher variance in current predictions and hindcasted projections than non-spatial SDMs. Performance of non-spatial and spatial SDMs according to the area under the receiver operating curve varied by genus. For both genera, false negative rates between non-spatial and spatial models were similar, but spatial models had lower false positive rates than non-spatial models.

The inclusion of computationally demanding spatial random effects in SDMs may be warranted when ecological or evolutionary processes prevent taxa from shifting their distributions or when the cost of false positives is high.