Modeling climate-driven changes in potential zoonotic virus density across Australia

Abstract

Summary Over 70% of emerging infectious diseases originate from zoonotic sources, making spillover risk prediction important for public health preparedness. This study investigates how potential zoonotic virus density (PVD) may change across Australia in 2050 under climate change. PVD captures ecological hazard, while actual spillover risk additionally depends on wildlife-human contact and pathogen transmission dynamics. Using an Extra Trees Regressor trained on 2020 climate outputs from eight global climate models (GCMs) under two emissions scenarios, and using projected 2020 PVD as the response variable, we generated spatial re-projections of climate suitability for PVD under 2050 climate conditions. Re-projections revealed spatial redistribution rather than a uniform change, with higher projected suitability across parts of western and southwestern Australia and lower suitability in some eastern areas relative to the 2020 projections. Relative humidity was the strongest overall climatic predictor, while the SHapley Additive exPlanations (SHAP) analysis highlighted model-specific effects of precipitation, water vapor, and solar radiation. Local Moran’s I showed persistent spatial clustering across coastal and inland areas.