Climate change is significantly altering species distributions and abundances globally. However, the impacts on desert ecosystems remain understudied, with limited research beyond polar regions or specific countries. This is concerning because warm deserts possess surprisingly high biodiversity, including many endemic species. Deserts are already experiencing the effects of climate change more severely than many other ecosystems and are projected to face substantial temperature increases. Many desert species exist near their physiological limits, meaning further warming could lead to local extinctions. Predicting climate change's effects on desert species is crucial. While mapping air temperature changes provides a starting point, this is insufficient as species operate within microclimates and exhibit behavioral and physiological responses that can either buffer or exacerbate climate change impacts. Physiological models, incorporating microclimatic data, offer more realistic assessments of species' responses. This study uses a microclimate model, a physiologically explicit biophysical model, climate change projections, and biodiversity maps to assess the impacts of climate change on desert bird communities globally. The research addresses four key questions: 1) How will warm deserts be affected, and do the projected impacts vary between and within major desert realms? 2) How do physiological model results compare to air temperature-only models? 3) Which areas will likely serve as refugia? 4) How much of these refugia are within existing protected areas (PAs)? Birds are the focus due to their diurnality, limited use of thermally buffered microsites (like burrows), and high evaporative water loss rates, making them particularly vulnerable to warming. Two physiological metrics—total evaporative water loss (TEWL) and acute dehydration risk (ADR)—are used to assess climate change impact. These metrics reflect the critical importance of water balance for desert birds, as they represent a trade-off between thermoregulation and dehydration.

The literature review section highlights the scarcity of research on climate change impacts in warm desert ecosystems. Existing studies primarily focus on either polar regions or specific geographic locations within deserts, leaving significant gaps in global understanding. The reviewed literature emphasizes the high biodiversity in these environments, the disproportionate impact of climate change on deserts compared to other biomes, and the precarious physiological state of many desert organisms, with many already living close to their thermal limits. The existing literature supports the need for more sophisticated predictive models that account for microclimates and species-specific physiological responses, instead of relying solely on broad-scale air temperature projections. The authors point to previous research highlighting declines in avian abundance in some desert regions as a result of climate change, further emphasizing the urgency of addressing these concerns.

The study combined a microclimate model and a physiologically explicit biophysical model with climate change projections and biodiversity maps. Global warm deserts were mapped using data from the IUCN Habitat Classification Scheme, focusing on habitat types associated with deserts and arid regions (less than 500 mm annual precipitation). These deserts were then divided into six major realms using an updated map of Wallace's zoogeographic regions. Historical and projected monthly climate data (50 km resolution) from TerraClimate were used. Two future climate scenarios were considered: a 2°C and a 4°C increase in global mean temperature above pre-industrial levels. The climate projections account for geographic patterns in warming and changes in radiation, humidity, and wind speed. The ‘micro-terra’ function of NicheMapR disaggregated monthly climate data to hourly values. Three model bird species representing small, medium, and large body sizes were created. The physiological model calculated hourly water loss (cutaneous and respiratory) using a modified version of the “endoR_devel” function in the R package “NicheMapR.” This function simulates heat and mass exchange and thermoregulatory responses (e.g., ptiloerection, posture changes, increased respiratory rate) to maintain minimal metabolic rates. TEWL and ADR were calculated for each species, and the results for the three species were averaged, weighted by the number of species in each size category for each grid cell. Desert bird species were defined as species with ≥90% of their habitat within warm deserts. Rarity-weighted species richness (RWR) was calculated for each grid cell. Climate change impact was represented by the overlap between current and future values of TEWL and ADR. Refugia were defined as areas with high RWR and low climate change impact. Two thresholding approaches were used: a fixed threshold (75th percentile) and a floating threshold to ensure at least 5% of each desert realm was classified as refugia. Finally, the extent of PA coverage of the refugia was assessed using the World Database on Protected Areas (WDPA). Model validation involved comparing model predictions of body temperature and water loss against empirical data for several well-studied species, and also comparing predicted changes in TEWL and ADR against observed occupancy declines in the Mojave Desert. Sensitivity analyses were conducted to assess the robustness of the results to interspecific variation in species traits.

The analysis revealed significant heterogeneity in predicted climate change impacts on desert birds both between and within major desert realms. The greatest changes in air temperature, TEWL, and ADR were projected for the Saharo-Arabian desert realm. The physiological model produced spatially distinct results from air temperature-only models, indicating that air temperature is not a sufficient proxy for physiological impacts. Most identified climate change refugia were located near coastlines, possibly due to an oceanic buffering effect. However, only a small proportion (generally <20%) of the identified refugia fell within the borders of existing protected areas (PAs). The proportion of area classified as refugia varied significantly between desert realms, both using fixed and floating thresholds for determining refugia. The Australian desert realm had the highest percentage of its area classified as refugia, while the Neotropical desert realm had the lowest. Comparisons of refugia identified using physiological metrics and air temperature showed substantial under-protection (false negatives) and over-protection (false positives) when using air temperature alone. These findings were robust across different scenarios (4°C warming, no access to shade).

The study demonstrates that understanding climate change impacts on desert species requires an integrated approach combining biodiversity patterns and physiological responses. The findings highlight the limitations of using air temperature alone as a proxy for physiological impacts, emphasizing the need for physiologically explicit models incorporating microclimate data. The identified climate change refugia, concentrated near coastlines, are vulnerable to sea-level rise and human disturbances. The low percentage of refugia within existing PAs underscores an urgent need for expanded conservation efforts to protect these areas, especially in the Neotropical desert realm, which shows the least protection. Future conservation strategies should consider the spatial distribution of refugia identified in this study. The study provides a physiologically relevant framework for prioritizing conservation areas in the world's warm deserts. The study also advocates for expanding the range of taxonomic groups investigated using these more sophisticated physiological models to better understand the impacts of climate change on desert biodiversity in its entirety.

This study highlights the heterogeneous impacts of climate change on desert bird communities globally and emphasizes the necessity of incorporating physiological models into climate change impact assessments. The identification of climate change refugia, predominantly located near coastlines but largely unprotected, underscores the urgent need for increased conservation efforts in these areas to mitigate the negative effects of climate change on desert bird populations. Further research using this method could expand on the analysis to other taxonomic groups for a more complete understanding of conservation needs in global warm desert ecosystems.

The study focused on birds, and its findings may not fully represent the climate change impacts on all taxa in warm deserts. The physiological model uses three representative bird species, which may not capture the full range of physiological responses across all desert bird species. The accuracy of the projections is dependent on the reliability of climate change models and the availability of microclimate and species distribution data. Future research could expand the model to incorporate a broader range of species and include additional factors affecting species resilience, such as habitat alteration and human interactions.