Habitat-based spatial prediction of human-elephant conflict risk in Minas, Riau, Indonesia

FADHIL NAUVALDY SYAHRAN (Tropical Biodiversity Conservation Study Program, Department of Forest Resources Conservation and Ecotourism, Faculty of Forestry and Environment, Institut Pertanian Bogor)
DEDE AULIA RAHMAN (Tropical Biodiversity Conservation Study Program, Department of Forest Resources Conservation and Ecotourism, Faculty of Forestry and Environment, Institut Pertanian Bogor)
YANTO SANTOSA (Tropical Biodiversity Conservation Study Program, Department of Forest Resources Conservation and Ecotourism, Faculty of Forestry and Environment, Institut Pertanian Bogor)

Biodiversitas Journal of Biological Diversity
Vol. 26 No. 9 (2025)

Abstract. Syahran FN, Rahman DA, Santosa Y. 2025. Habitat-based spatial prediction of human-elephant conflict risk in Minas, Riau, Indonesia. Biodiversitas 26: 4465-4478. Human-Elephant Conflict (HEC) presents a pressing threat to the endangered Sumatran elephant (Elephas maximus sumatranus) in Riau, Indonesia. This study utilized Maximum Entropy (MaxEnt) modeling to predict HEC risk zones in Minas, focusing on a small subpopulation of 11 GPS-collared elephants. Conflict occurrence data from 2020-2022 were compiled from validated reports by the Riau Natural Resources Conservation Agency (BBKSDA) and Global Positioning System (GPS) movement records. Fourteen environmental variables were initially considered, and multicollinearity analysis reduced these to eleven key predictors. MaxEnt was run using 25 cross-validation replicates, achieving excellent performance (AUC: 0.958). Proximity to open land (40.9% contribution), industrial forest plantations (22.8%), settlements (9.2%), roads (8.8%), and oil palm plantations (8.3%) emerged as the strongest predictors. Conflict probability peaked within 0-2 km of anthropogenic features, confirming the critical role of disturbed habitats. High- to very-high-risk zones (>70% probability) covered 19.94 km² (1.76% of the landscape) and were identified as priority areas for intervention, while a moderate-risk zone of 221.32 km² (19.48%) requires targeted monitoring. Low and minimal-risk zones dominated, spanning over 1,000 km² (>80%). These spatial predictions offer actionable insights for conservation planning by directly linking risk zones to mitigation needs. Strategies such as establishing buffer crops, strengthening community-based patrols, and developing early warning systems should be prioritized in high-risk areas. At the same time, habitat connectivity and ecological restoration are crucial for mitigating long-term conflict pressure. Overall, this study advances HEC modeling in Sumatra by combining empirical conflict records with predictive mapping, offering a practical framework to guide evidence-based management and promote human-elephant coexistence in fragmented tropical landscapes.