Hotspots and ecological drivers of human-sloth bear conflicts in Tamil nadu, India

Human-wildlife conflict (HWC) poses a significant challenge to conservation and socioeconomic stability, primarily driven by anthropogenic pressures. Sloth bears (Melursus ursinus), classified as Vulnerable by the IUCN, frequently engage in conflict with humans. Understanding the spatial and temporal patterns of conflict is crucial for developing effective mitigation strategies and informed conservation planning. This study analyzes Human-Sloth Bear Conflict (HSBC) events across 48 forest divisions in Tamil Nadu from 2016 to 2021, examining spatial hotspots, seasonal trends, and key predictors using ensemble modeling. A total of 17 divisions reported HSBC incidents, with a mean (± SD) conflict frequency of 3.69 ± 4.36 conflicts per affected division. Pollachi Division recorded the highest number of conflicts (n = 17), with the majority of incidents concentrated in the Valparai and Manambolly forest ranges. Seasonal trends indicate strong fluctuations in conflict frequency, with (n = 61) HSBC incidents recorded across 12 months, averaging 5.08 conflicts per month with a standard deviation of 2.31. Conflict incidents remain relatively low between January and May (n = 2 to n = 5 incidents per month), reaching their lowest point in April (n = 2). However, conflict levels peak in July (n = 9) and remain consistently high during the monsoon and post-monsoon months (June, July, and October), necessitating targeted management strategies during these high-risk periods. Human injuries dominate HSBC cases, accounting for 83.87% (n = 52), followed by human fatalities, which account for 9.68% (n = 6). The distance to the forest boundary revealed that over 92% of HSBC incidents occurred within 10 km of reserved forests, with nearly half concentrated within just 1 km of the forest edge. The LULC analysis reveals that agricultural and low-density rural landscapes adjacent to forests are the most susceptible to HSBC. The Random Forest (RF) model demonstrated high predictive accuracy (AUC = 0.95, TSS = 0.77, correlation = 0.72), identifying high-risk zones in the Nilgiris, Sathyamangalam, Coimbatore, and Anamalai regions, as well as parts of Tirunelveli and Kanyakumari. Predictor analysis reveals that temperature (Bio1), precipitation (Bio12), road proximity, elevation, and human modification indices (HMI) have a strong influence on conflict likelihood. Conflicts are more frequent in moderate temperature zones, near water sources, and in semi-modified landscapes. These findings provide essential insights for strategic conflict mitigation efforts, emphasising seasonally adapted interventions and landscape-specific conservation strategies.