Species Distribution Modeling Under Climate Change Scenarios

Species distribution models (SDMs) have become indispensable tools for forecasting how climate change will reshape the geographic ranges of flora and fauna. By relating occurrence records to environmental predictors, these models generate projections of habitat suitability under alternative future climates. Recent advances leverage machine learning, ensemble forecasting and high‐resolution climate data to capture both broad patterns and fine‐scale refugia. Integrating dispersal mechanisms and demographic constraints has improved assessments of species’ capacity to track shifting climates. At the same time, efforts to quantify and communicate model uncertainty—arising from algorithm choice, climate scenarios and data quality—have become central to informing conservation planning, policy decisions and adaptive management worldwide.

Research from Nature Portfolio

Innovative deep‐learning approaches have been applied to large citizen‐science datasets to jointly model thousands of plant species at fine spatial and temporal resolution. By explicitly accounting for seasonal observation biases and projecting neural network outputs into the future, these methods provide nuanced forecasts of distributional shifts, changes in community composition and phenological phases. Separately, global assessments using ensemble frameworks have evaluated the relative influence of modelling algorithms, emissions pathways, circulation models and dispersal assumptions on biodiversity projections for amphibians, birds and mammals out to mid‐ and late century. These studies reveal that algorithm selection and scenario choice dominate overall uncertainty, emphasising the necessity of multi‐model ensembles and transparent reporting of underlying assumptions.

Species Distribution Modeling Under Climate Change Scenarios publication trend

The graph below shows the total number of articles in species distribution modeling under climate change scenarios across all publications each year (not limited to Nature Index journals).

Technical terms

Species distribution model (SDM): A statistical or machine‐learning framework relating species occurrence records to environmental variables to predict geographic suitability.

Ensemble forecasting: The practice of combining predictions from multiple models or scenarios to characterise central tendencies and uncertainty.

Representative concentration pathway (RCP): A greenhouse-gas concentration trajectory used to project future climate conditions under differing emissions scenarios.

Global circulation model (GCM): A numerical simulation of Earth’s climate system, providing large-scale projections of temperature, precipitation and atmospheric dynamics.

Statistical downscaling: Methods to translate coarse-resolution climate model outputs into finer-scale environmental data for use in local or regional SDMs.

References

1. Multispecies deep learning using citizen science data produces more informative plant community models. Nature Communications (2024).
2. Climatologies at high resolution for the earth’s land surface areas. Scientific Data (2017).
3. Uncertainty in ensembles of global biodiversity scenarios. Nature Communications (2019).
4. A standard protocol for reporting species distribution models. Ecography (2020).
5. Improving the Use of Species Distribution Models in Conservation Planning and Management under Climate Change. PLOS ONE (2014).