Venomous animal incidents represent an important public health issue and may be influenced by meteorological conditions as well as land use and land cover characteristics. However, it remains unclear which predictive approaches perform best in estimating these events and how increasing temperatures may alter their spatial distribution. This study compared different predictive models to estimate monthly incidence rates of scorpion, snake, and spider envenomation across the municipalities of São Paulo State, Brazil, from 2010 to 2024, and evaluated the potential impact of hypothetical warming scenarios (+ 1 °C, + 2 °C, and + 4 °C). Multiple linear regression, Generalized Additive Models (GAM), Extreme Gradient Boosting (XGBoost), and Random Forest models were compared using meteorological and land use/land cover variables. Random Forest achieved the best predictive performance for all three animal groups. Land use and land cover variables, particularly the proportion of urban area, were the most important predictors of the current spatial distribution of incidents. In contrast, temperature was the main factor associated with the projected changes under the simulated warming scenarios. Increasing temperatures resulted in higher predicted incidence rates for all animal groups, with the greatest increases observed for scorpion envenomation, revealing substantial spatial heterogeneity among municipalities. These findings indicate that the factors associated with the current distribution of venomous animal incidents are not necessarily the same as those driving their response to rising temperatures, highlighting the importance of integrating meteorological and land use and land cover variables to improve future projections and support public health surveillance and planning.
da Silva Freitas, L., da Silva Júnior, F.M.R. Comparison of predictive models for estimating scorpion, snake, and spider envenomation under warming scenarios in São Paulo State, Brazil. Int J Biometeorol 70, 294 (2026). https://doi.org/10.1007/s00484-026-03353-8
