2025/07/02 by Carleen Schwarz, Juliana Masseloux, S. Blair Hedges · 1 voice
Environmental Science · Biochemistry, Genetics and Molecular Biology · #Species Distribution and Climate Change #Wildlife Ecology and Conservation #Genetic diversity and population structure
paper · doi:10.1016/j.gecco.2025.e03719
openalex publication_date 2025/07/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/23
Human-elephant conflict (HEC) is a pervasive conservation challenge impacting both human and elephant populations where their habitats overlap. Understanding the spatial dynamics of HEC through Species Distribution Models (SDMs) is crucial for devising effective mitigation strategies like proactive land-use planning. While Maximum Entropy (MaxEnt) models have been widely used, comparative assessments of alternative SDMs remain limited. This study evaluates the performance of five SDMs—Generalized Linear Model (GLM), Generalized Additive Model (GAM), Gradient Boosting Machines (GBM), Random Forest (RF), and MaxEnt—alongside three ensemble approaches (Mean, Median, and Weighted Mean) to predict HEC risk in two regions of Thailand using eight environmental variables. Ensemble models outperformed individual algorithms at predicting HEC risk (Ensembles: AUC = 0.955–0.985; TSS = 0.868–0.950; Kappa = 0.736–0.854), effectively mitigating the limitations of single-model approaches. Where ensembles are impractical, MaxEnt models demonstrated robust performance (AUC = 0.934–0.960; TSS = 0.741–0.910; Kappa = 0.657–0.729), offering a viable alternative to the overly conservative RF and GBM models. Variable importance analysis revealed context-specific drivers of HEC, with water-related factors being more influential in one study site than the other. These insights highlight the need for adaptive strategies tailored to local conditions. The resulting risk maps provide a crucial foundation for targeted HEC mitigation and inform policy development aimed at fostering human-elephant coexistence. Future research should focus on refining models for projections, as this could be a novel and useful way to overcome limitations of scarce historical data on crop-raiding when developing HEC insurance programs.