OR26 - Using ensemble modelling for predicting habitat suitability of sun bears in the Kampar Peninsula, Sumatra
David Lee, University of South Wales, UKMain tent
Dede Aulia Rahman, Yudi Setiawan, Yoan Dinata, Muhammad Iqbal, Dian Andi Syahputra, David Lee
Sumatra, Indonesia, has experienced significant tropical forest degradation and loss over the last 30 years. The province of Riau, which includes the peat swamp forest of Kampar Peninsula, has experienced the highest rate of forest loss in Sumatra during this time. Alongside monoculture plantations, Kampar Peninsula has one of the largest remaining tropical forest peat forests in Sumatra, while including several wildlife reserves and the ecosystem restoration concessions of Restorasi Ekosistem Riau (RER), representing an important landscape for sun bears. Here, we collated sun bear presence records from camera trapping (2015-24) and bear sign surveys (2024) in RER to model habitat suitability across the wider peninsula. We modelled bear presence against a series of biophysical and anthropogenic predictors using four algorithms within a weighted ensemble framework. While all single models showed good predictive performance, the ensemble model produced the most stable and reliable predictions, highlighting the benefit of combining multiple algorithms to reduce model uncertainty. Precipitation, land cover, canopy height, and distances to settlements, roads and water bodies emerged as the most influential predictors, indicating that climatic conditions, forest structure, and human disturbance all contribute to sun bear habitat suitability in this peatland ecosystem. Habitat suitability was highest in secondary peat swamp forest, and mostly within RER, highlighting their value and that of ecosystem restoration concessions as core bear habitat. These findings provide the first landscape-scale prediction of habitat suitability for sun bear in this important landscape and offer a spatial basis for prioritising peat swamp forest protection and restoration, as well as adaptive management of plantation-dominated matrices to support the long-term conservation of this globally threatened species.