The role of land cover in bioclimatic models depends on spatial resolution

Aim We explored the importance of climate and land cover in bird species distribution models on multiple spatial scales. In particular, we tested whether the integration of land cover data improves the performance of pure bioclimatic models. Location Finland, northern Europe. Methods The data of the bird atlas survey carried out in 1986–89 using a 10 × 10 km uniform grid system in Finland were employed in the analyses. Land cover and climatic variables were compiled using the same grid system. The dependent and explanatory variables were resampled to 20‐km, 40‐km and 80‐km resolutions. Generalized additive models (GAM) were constructed for each of the 88 land bird species studied in order to estimate the probability of occurrence as a function of (1) climate and (2) climate and land cover variables. Model accuracy was measured by a cross‐validation approach using the area under the curve (AUC) of a receiver operating characteristic (ROC) plot. Results In general, the accuracies of the 88 bird–climate models were good at all studied resolutions. However, the inclusion of land cover increased the performance of 79 and 78 of the 88 bioclimatic models at 10‐km and 20‐km resolutions, respectively. There was no significant improvement at the 40‐km resolution. In contrast to the finer resolutions, the inclusion of land cover variables decreased the modelling accuracy at 80km resolution. Main conclusions Our results suggest that the determinants of bird species distributions are hierarchically structured: climatic variables are large‐scale determinants, followed by land cover at finer resolutions. The majority of the land bird species in Finland are rather clearly correlated with climate, and bioclimate envelope models can provide useful tools for identifying the relationships between these species and the environment at resolutions ranging from 10 km to 80 km. However, the notable contribution of land cover to the accuracy of bioclimatic models at 10–20‐km resolutions indicates that the integration of climate and land cover information can improve our understanding and model predictions of biogeographical patterns under global change.     

Do community‐level models describe community variation effectively?

Aim The aim of community‐level modelling is to improve the performance of species distributional models by taking patterns of co‐occurrence among species into account. Here, we test this expectation by examining how well three community‐level modelling strategies (‘assemble first, predict later’, ‘predict first, assemble later’, and ‘assemble and predict together’) spatially project the observed composition of species assemblages. Location Europe. Methods Variation in the composition of European tree assemblages and its spatial and environmental correlates were examined with cluster analysis and constrained analysis of principal coordinates. Results were used to benchmark spatial projections from three community‐based strategies: (1) assemble first, predict later (cluster analysis first, then generalized linear models, GLMs); (2) predict first, assemble later (GLMs first, then cluster analysis); and (3) assemble and predict together (constrained quadratic ordination). Results None of the community‐level modelling strategies was able to accurately model the observed distribution of tree assemblages in Europe. Uncertainty was particularly high in southern Europe, where modelled assemblages were markedly different from observed ones. Assembling first and predicting later led to distribution models with the simultaneous occurrence of several types of assemblages in southern Europe that do not co‐occur, and the remaining strategies yielded models with the presence of non‐analogue assemblages that presently do not exist and that are much more strongly correlated with environmental gradients than with the real assemblages. Main conclusions Community‐level models were unable to characterize the distribution of European tree assemblages effectively. Models accounting for co‐occurrence patterns along environmental gradients did not outperform methods that assume individual responses of species to climate. Unrealistic assemblages were generated because of the models’ inability to capture fundamental processes causing patterns of covariation among species. The usefulness of these forms of community‐based models thus remains uncertain and further research is required to demonstrate their utility.     

Biotic interactions influence the projected distribution of a specialist mammal under climate change

Aim

To measure the effects of including biotic interactions on climate‐based species distribution models (SDMs) used to predict distribution shifts under climate change. We evaluated the performance of distribution models for an endangered marsupial, the northern bettong (Bettongia tropica), comparing models that used only climate variables with models that also took into account biotic interactions.

Location

North‐east Queensland, Australia.

Methods

We developed separate climate‐based distribution models for the northern bettong, its two main resources and a competitor species. We then constructed models for the northern bettong by including climate suitability estimates for the resources and competitor as additional predictor variables to make climate + resource and climate + resource + competition models. We projected these models onto seven future climate scenarios and compared predictions of northern bettong distribution made by these differently structured models, using a ‘global’ metric, the I similarity statistic, to measure overlap in distribution and a ‘local’ metric to identify where predictions differed significantly.

Results

Inclusion of food resource biotic interactions improved model performance. Over moderate climate changes, up to 3.0 °C of warming, the climate‐only model for the northern bettong gave similar predictions of distribution to the more complex models including interactions, with differences only at the margins of predicted distributions. For climate changes beyond 3.0 °C, model predictions diverged significantly. The interactive model predicted less contraction of distribution than the simpler climate‐only model.

Main conclusions

Distribution models that account for interactions with other species, in particular direct resources, improve model predictions in the present‐day climate. For larger climate changes, shifts in distribution of interacting species cause predictions of interactive models to diverge from climate‐only models. Incorporating interactions with other species in SDMs may be needed for long‐term prediction of changes in distribution of species under climate change, particularly for specialized species strongly dependent on a small number of biotic interactions.     

Species distribution models for Peruvian plantcutter improve with consideration of biotic interactions

Biotic interactions have been controversial in distributional ecology, mainly in regards to whether they have effects over broad extents, with the negative view known as the Eltonian noise hypothesis (ENH). In this study, we evaluated the ENH for Phytotoma raimondii, a restricted‐range Peruvian endemic bird species: we developed models based on 1) only abiotic conditions, 2) only host plant distributions, and 3) both abiotic conditions and host plant distributions; models were evaluated with partial receiver operating characteristic test and Akaike information criteria metrics. We rejected the ENH for this case: biotic interactions improved the model. The frequency with which exceptions to the ENH are detected has important implications for distributional ecology and methods for estimating distributions of species.     

Fish introductions reveal the temperature dependence of species interactions

A major area of current research is to understand how climate change will impact species interactions and ultimately biodiversity. A variety of environmental conditions are rapidly changing owing to climate warming, and these conditions often affect both the strength and outcome of species interactions. We used fish distributions and replicated fish introductions to investigate environmental conditions influencing the coexistence of two fishes in Swedish lakes: brown trout (Salmo trutta) and pike (Esox lucius). A logistic regression model of brown trout and pike coexistence showed that these species coexist in large lakes (more than 4.5 km²), but not in small, warm lakes (annual air temperature more than 0.9–1.5°C). We then explored how climate change will alter coexistence by substituting climate scenarios for 2091–2100 into our model. The model predicts that brown trout will be extirpated from approximately half of the lakes where they presently coexist with pike and from nearly all 9100 lakes where pike are predicted to invade. Context dependency was critical for understanding pike–brown trout interactions, and, given the widespread occurrence of context-dependent species interactions, this aspect will probably be critical for accurately predicting climate impacts on biodiversity.     

Species distribution modelling for the people: unclassified landsat TM imagery predicts bird occurrence at fine resolutions

Aim

Assessing the influence of land cover in species distribution modelling is limited by the availability of fine‐resolution land‐cover data appropriate for most species responses. Remote‐sensing technology offers great potential for predicting species distributions at large scales, but the cost and required expertise are prohibitive for many applications. We test the usefulness of freely available raw remote‐sensing reflectance data in predicting species distributions of 40 commonly occurring bird species in western Oregon.

Location

Central Coast Range, Cascade and Klamath Mountains Oregon, USA.

Methods

Information on bird observations was collected from 4598 fixed‐radius point counts. Reflectance data were obtained using 30‐m resolution Landsat imagery summarized at scales of 150, 500, 1000 and 2000 m. We used boosted regression tree (BRT) models to analyse relationships between distributions of birds and reflectance values and evaluated prediction performance of the models using area under the receiver operating characteristic curve (AUC) values.

Results

Prediction success of models using all reflectance values was high (mean AUC = 0.79 ± 0.10 SD). Further, model performance using individual reflectance bands exceeded those that used only Normalized Difference Vegetation Index (NDVI). The relative influence of band 4 predictors was highest, indicating the importance of variables associated with vegetation biomass and photosynthetic activity. Across spatial scales, the average influence of predictors at the 2000 m scale was greatest.

Main Conclusions

We demonstrate that unclassified remote‐sensing imagery can be used to produce species distribution models with high prediction success. Our study is the first to identify general patterns in the usefulness of spectral reflectances for species distribution modelling of multiple species. We conclude that raw Landsat Thematic Mapper data will be particularly useful in species distribution models when high‐resolution predictions are required, including habitat change detection studies, identification of fine‐scale biodiversity hotspots and reserve design.

     

Species interactions weakly modify climate‐induced tree co‐occurrence patterns

Aims

Species distributions are hypothesized to be underlain by a complex association of processes that span multiple spatial scales including biotic interactions, dispersal limitation, fine‐scale resource gradients and climate. Species disequilibrium with climate may reflect the effects of non‐climatic processes on species distributions, yet distribution models have rarely directly considered non‐climatic processes. Here, we use a Joint Species Distribution Model (JSDM) to investigate the influence of non‐climatic factors on species co‐occurrence patterns and to directly quantify the relative influences of climate and alternative processes that may generate correlated responses in species distributions, such as species interactions, on tree co‐occurrence patterns.

Location

US Rocky Mountains.

Methods

We apply a Bayesian JSDM to simultaneously model the co‐occurrence patterns of ten dominant tree species across the Rocky Mountains, and evaluate climatic and residual correlations from the fitted model to determine the relative contribution of each component to observed co‐occurrence patterns. We also evaluate predictions generated from the fitted model relative to a single‐species modelling approach.

Results

For most species, correlation due to climate covariates exceeded residual correlation, indicating an overriding influence of broad‐scale climate on co‐occurrence patterns. Accounting for covariance among species did not significantly improve predictions relative to a single‐species approach, providing limited evidence for a strong independent influence of species interactions on distribution patterns.

Conclusions

Overall, our findings indicate that climate is an important driver of regional biodiversity patterns and that interactions between dominant tree species contribute little to explain species co‐occurrence patterns among Rocky Mountain trees.     

Dominant climate influences on North American bird distributions

Aim Geographic distributions of species are constrained by several factors acting at different scales, with climate assumed to be a major determinant at broad extents. Recent studies, however, have challenged this statement and indicated that climate may not dominate among the factors governing geographic distributions of species. Here, we argue that these results are misleading due to the lack of consideration of the geographic area that has been accessible to the species. Location North America. Methods We generated null distributions for 75 North American endemic and 19 non‐endemic bird species. For each species, climatic envelopes of observed and null distributions were modelled using neural networks and generalized linear models, and seven climatic predictors. Values of the area under the receiver–operating characteristic curve (AUC) based on models of observed distributions were compared with corresponding AUC values for the null distributions. Results More than 82% of the endemic species showed AUC higher for the observed than for the null distributions, while 63% of the non‐endemic species showed such a pattern. Main conclusions We demonstrate a dominant climatic signal in shaping North American bird distributions. Our results attest to the importance of climate in determining species distributions and support the use of climate‐envelope models for estimating potential distributional areas at the appropriate spatial scales.     

Species distributions models may predict accurately future distributions but poorly how distributions change: A critical perspective on model validation

Aim

Species distribution models (SDMs) are widely used to make predictions on how species distributions may change as a response to climatic change. To assess the reliability of those predictions, they need to be critically validated with respect to what they are used for. While ecologists are typically interested in how and where distributions will change, we argue that SDMs have seldom been evaluated in terms of their capacity to predict such change. Instead, typical retrospective validation methods estimate model's ability to predict to only one static time in future. Here, we apply two validation methods, one that predicts and evaluates a static pattern, while the other measures change and compare their estimates of predictive performance.

Location

Fennoscandia.

Methods

We applied a joint SDM to model the distributions of 120 bird species in four model validation settings. We trained models with a dataset from 1975 to 1999 and predicted species' future occurrence and abundance in two ways: for one static time period (2013–2016, ‘static validation’) and for a change between two time periods (difference between 1996–1999 and 2013–2016, ‘change validation’). We then measured predictive performance using correlation between predicted and observed values. We also related predictive performance to species traits.

Results

Even though static validation method evaluated predictive performance as good, change method indicated very poor performance. Predictive performance was not strongly related to any trait.

Main Conclusions

Static validation method might overestimate predictive performance by not revealing the model's inability to predict change events. If species' distributions remain mostly stable, then even an unfit model can predict the near future well due to temporal autocorrelation. We urge caution when working with forecasts of changes in spatial patterns of species occupancy or abundance, even for SDMs that are based on time series datasets unless they are critically validated for forecasting such change.     

Linking macroecology and community ecology: refining predictions of species distributions using biotic interaction networks

Macroecological models for predicting species distributions usually only include abiotic environmental conditions as explanatory variables, despite knowledge from community ecology that all species are linked to other species through biotic interactions. This disconnect is largely due to the different spatial scales considered by the two sub‐disciplines: macroecologists study patterns at large extents and coarse resolutions, while community ecologists focus on small extents and fine resolutions. A general framework for including biotic interactions in macroecological models would help bridge this divide, as it would allow for rigorous testing of the role that biotic interactions play in determining species ranges. Here, we present an approach that combines species distribution models with Bayesian networks, which enables the direct and indirect effects of biotic interactions to be modelled as propagating conditional dependencies among species’ presences. We show that including biotic interactions in distribution models for species from a California grassland community results in better range predictions across the western USA. This new approach will be important for improving estimates of species distributions and their dynamics under environmental change.     

The importance of biotic interactions for modelling species distributions under climate change

Aim There is a debate as to whether biotic interactions exert a dominant role in governing species distributions at macroecological scales. The prevailing idea is that climate is the key limiting factor; thus models that use present‐day climate–species range relationships are expected to provide reasonable means to quantify the impacts of climate change on species distributions. However, there is little empirical evidence that biotic interactions would not constrain species distributions at macroecological scales. We examine this idea, for the first time, and provide tests for two null hypotheses: (H₀ 1) – biotic interactions do not exert a significant role in explaining current distributions of a particular species of butterfly (clouded Apollo, Parnassius mnemosyne) in Europe; and (H₀ 2) – biotic interactions do not exert a significant role in predictions of altered species’ ranges under climate change. Location Europe. Methods Generalized additive modelling (GAM) was used to investigate relationships between species and climate; species and host plants; and species and climate + host plants. Because models are sensitive to the variable selection strategies utilised, four alternative approaches were used: AIC (Akaike's Information Criterion), BIC (Bayesian Information Criterion), BRUTO (Adaptive Backfitting), and CROSS (Cross Selection). Results In spite of the variation in the variables selected with different methods, both hypotheses (H₀ 1 and H₀ 2) were falsified, providing support for the proposition that biotic interactions significantly affect both the explanatory and predictive power of bioclimatic envelope models at macro scales. Main conclusions Our results contradict the widely held view that the effects of biotic interactions on individual species distributions are not discernible at macroecological scales. Results are contingent on the species, type of interaction and methods considered, but they call for more stringent evidence in support of the idea that purely climate‐based modelling would be sufficient to quantify the impacts of climate change on species distributions.     

Modelling the distribution of a threatened habitat: the California sage scrub

Aim Using predictive species distribution and ecological niche modelling our objectives are: (1) to identify important climatic drivers of distribution at regional scales of a locally complex and dynamic system – California sage scrub; (2) to map suitable sage scrub habitat in California; and (3) to distinguish between bioclimatic niches of floristic groups within sage scrub to assess the conservation significance of analysing such species groups. Location Coastal mediterranean‐type shrublands of southern and central California. Methods Using point localities from georeferenced herbarium records, we modelled the potential distribution and bioclimatic envelopes of 14 characteristic sage scrub species and three floristic groups (south‐coastal, coastal–interior disjunct and broadly distributed species) based upon current climate conditions. Maxent was used to map climatically suitable habitat, while principal components analysis followed by canonical discriminant analysis were used to distinguish between floristic groups and visualize species and group distributions in multivariate ecological space. Results Geographical distribution patterns of individual species were mirrored in the habitat suitability maps of floristic groups, notably the disjunct distribution of the coastal–interior species. Overlap in the distributions of floristic groups was evident in both geographical and multivariate niche space; however, discriminant analysis confirmed the separability of floristic groups based on bioclimatic variables. Higher performance of floristic group models compared with sage scrub as a whole suggests that groups have differing climate requirements for habitat suitability at regional scales and that breaking sage scrub into floristic groups improves the discrimination between climatically suitable and unsuitable habitat. Main conclusions The finding that presence‐only data and climatic variables can produce useful information on habitat suitability of California sage scrub species and floristic groups at a regional scale has important implications for ongoing efforts of habitat restoration for sage scrub. In addition, modelling at a group level provides important information about the differences in climatic niches within California sage scrub. Finally, the high performance of our floristic group models highlights the potential a community‐level modelling approach holds for investigating plant distribution patterns.     

Biotic predictors with phenological information improve range estimates for migrating monarch butterflies in Mexico

Although long-standing theory suggests that biotic variables are only relevant at local scales for explaining the patterns of species' distributions, recent studies have demonstrated improvements to species distribution models (SDMs) by incorporating predictor variables informed by biotic interactions. However, some key methodological questions remain, such as which kinds of interactions are permitted to include in these models, how to incorporate the effects of multiple interacting species, and how to account for interactions that may have a temporal dependence. We addressed these questions in an effort to model the distribution of the monarch butterfly Danaus plexippus during its fall migration (September–November) through Mexico, a region with new monitoring data and uncertain range limits even for this well-studied insect. We estimated species richness of selected nectar plants (Asclepias spp.) and roosting trees (various highland species) for use as biotic variables in our models. To account for flowering phenology, we additionally estimated nectar plant richness of flowering species per month. We evaluated three types of models: climatic variables only (abiotic), plant richness estimates only (biotic) and combined (abiotic and biotic). We selected models with AICc and additionally determined if they performed better than random on spatially withheld data. We found that the combined models accounting for phenology performed best for all three months, and better than random for discriminatory ability but not omission rate. These combined models also produced the most ecologically realistic spatial patterns, but the modeled response for nectar plant richness matched ecological predictions for November only. These results represent the first model-based monarch distributional estimates for the Mexican migration route and should provide foundations for future conservation work. More generally, the study demonstrates the potential benefits of using SDM-derived richness estimates and phenological information for biotic factors affecting species distributions.     

Habitat shifts of endangered species under altered climate conditions: importance of biotic interactions

Predicting changes in potential habitat for endangered species as a result of global warming requires considering more than future climate conditions; it is also necessary to evaluate biotic associations. Most distribution models predicting species responses to climate change include climate variables and occasionally topographic and edaphic parameters, rarely are biotic interactions included. Here, we incorporate biotic interactions into niche models to predict suitable habitat for species under altered climates. We constructed and evaluated niche models for an endangered butterfly and a threatened bird species, both are habitat specialists restricted to semiarid shrublands of southern California. To incorporate their dependency on shrubs, we first developed climate‐based niche models for shrubland vegetation and individual shrub species. We also developed models for the butterfly's larval host plants. Outputs from these models were included in the environmental variable dataset used to create butterfly and bird niche models. For both animal species, abiotic–biotic models outperformed the climate‐only model, with climate‐only models over‐predicting suitable habitat under current climate conditions. We used the climate‐only and abiotic–biotic models to calculate amounts of suitable habitat under altered climates and to evaluate species' sensitivities to climate change. We varied temperature (+0.6, +1.7, and +2.8 °C) and precipitation (50%, 90%, 100%, 110%, and 150%) relative to current climate averages and within ranges predicted by global climate change models. Suitable habitat for each species was reduced at all levels of temperature increase. Both species were sensitive to precipitation changes, particularly increases. Under altered climates, including biotic variables reduced habitat by 68–100% relative to the climate‐only model. To design reserve systems conserving sensitive species under global warming, it is important to consider biotic interactions, particularly for habitat specialists and species with strong dependencies on other species.     

Climate change and the future distributions of aquatic macrophytes across boreal catchments

Aim Aquatic–terrestrial ecotones are vulnerable to climate change, and degradation of the emergent aquatic macrophyte zone would have severe ecological consequences for freshwater, wetland and terrestrial ecosystems. Our aim was to uncover future changes in boreal emergent aquatic macrophyte zones by modelling the occurrence and percentage cover of emergent aquatic vegetation under different climate scenarios in Finland by the 2050s. Location Finland, northern Europe. Methods Data derived from different GIS sources were used to estimate future emergent aquatic macrophyte distributions in all catchments in Finland (848 in total). We used generalized additive models (GAM) with a full stepwise selection algorithm and Akaike information criterion to explore the main environmental determinates (climate and geomorphology) of emergent aquatic macrophyte distributions, which were derived from the national subclass of CORINE land‐cover classification. The accuracy of the distribution models (GAMs) was cross‐validated, using percentage of explained deviance and the area under the curve derived from the receiver‐operating characteristic plots. Results Our results indicated that emergent aquatic macrophytes will expand their distributions northwards from the current catchments and percentage cover will increase in all of the catchments in all climate scenarios. Growing degree‐days was the primary determinant affecting distributions of emergent aquatic macrophytes. Inclusion of geomorphological variables clearly improved model performance in both model exercises compared with pure climate variables. Main conclusions Emergent aquatic macrophyte distributions will expand due to climate change. Many emergent aquatic plant species have already expanded their distributions during the past decades, and this process will continue in the years 2051–80. Emergent aquatic macrophytes pose an increasing overgrowth risk for sensitive macrophyte species in boreal freshwater ecosystems, which should be acknowledged in management and conservation actions. We conclude that predictions based on GIS data can provide useful ‘first‐filter’ estimates of changes in aquatic–terrestrial ecotones.     

The pace of past climate change vs. potential bird distributions and land use in the United States

Climate change may drastically alter patterns of species distributions and richness, but predicting future species patterns in occurrence is challenging. Significant shifts in distributions have already been observed, and understanding these recent changes can improve our understanding of potential future changes. We assessed how past climate change affected potential breeding distributions for landbird species in the conterminous United States. We quantified the bioclimatic velocity of potential breeding distributions, that is, the pace and direction of change for each species’ suitable climate space over the past 60 years. We found that potential breeding distributions for landbirds have shifted substantially with an average velocity of 1.27 km yr^?1, about double the pace of prior distribution shift estimates across terrestrial systems globally (0.61 km yr^?1). The direction of shifts was not uniform. The majority of species’ distributions shifted west, northwest, and north. Multidirectional shifts suggest that changes in climate conditions beyond mean temperature were influencing distributional changes. Indeed, precipitation variables that were proxies for extreme conditions were important variables across all models. There were winners and losers in terms of the area of distributions; many species experienced contractions along west and east distribution edges, and expansions along northern distribution edges. Changes were also reflected in the potential species richness, with some regions potentially gaining species (Midwest, East) and other areas potentially losing species (Southwest). However, the degree to which changes in potential breeding distributions are manifested in actual species richness depends on landcover. Areas that have become increasingly suitable for breeding birds due to changing climate are often those attractive to humans for agriculture and development. This suggests that many areas might have supported more breeding bird species had the landscape not been altered. Our study illustrates that climate change is not only a future threat, but something birds are already experiencing.     

From species distributions to meta‐communities

The extent that biotic interactions and dispersal influence species ranges and diversity patterns across scales remains an open question. Answering this question requires framing an analysis on the frontier between species distribution modelling (SDM), which ignores biotic interactions and dispersal limitation, and community ecology, which provides specific predictions on community and meta‐community structure and resulting diversity patterns such as species richness and functional diversity. Using both empirical and simulated datasets, we tested whether predicted occurrences from fine‐resolution SDMs provide good estimates of community structure and diversity patterns at resolutions ranging from a resolution typical of studies within reserves (250 m) to that typical of a regional biodiversity study (5 km). For both datasets, we show that the imprint of biotic interactions and dispersal limitation quickly vanishes when spatial resolution is reduced, which demonstrates the value of SDMs for tracking the imprint of community assembly processes across scales.     

Why is the choice of future climate scenarios for species distribution modelling important?

Species distribution models (SDMs) are common tools for assessing the potential impact of climate change on species ranges. Uncertainty in SDM output occurs due to differences among alternate models, species characteristics and scenarios of future climate. While considerable effort is being devoted to identifying and quantifying the first two sources of variation, a greater understanding of climate scenarios and how they affect SDM output is also needed. Climate models are complex tools: variability occurs among alternate simulations, and no single 'best' model exists. The selection of climate scenarios for impacts assessments should not be undertaken arbitrarily - strengths and weakness of different climate models should be considered. In this paper, we provide bioclimatic modellers with an overview of emissions scenarios and climate models, discuss uncertainty surrounding projections of future climate and suggest steps that can be taken to reduce and communicate climate scenario-related uncertainty in assessments of future species responses to climate change.