Incorporating food plant distributions as important predictors in the habitat suitability model of sumatran orangutan (Pongo abelii) in Gunung Leuser National Park, Indonesia

Abstract

Gunung Leuser National Park (GLNP) is a vital habitat for the Sumatran orangutan. This study enhances habitat suitability models (HSM) for Sumatran orangutans by incorporating the predictive distributions for 21 key orangutan food plants, which had not been previously explored. Using machine learning algorithms—support vector machine, random forest, boosted regression trees, and maximum entropy—along with an ensemble model, seven important food plants, including Ixora insularum and Calamus manan, were identified as critical predictors of habitat suitability. The ensemble model revealed that 70.29 % (5834.84 km²) of GLNP is suitable for Sumatran orangutans, with the key influencing factors being land cover, distance to settlements, and temperature seasonality or annual variation of temperature. As this study only successfully incorporated a few food plant species, future research should include a wider range of food plant species, especially preferred fruits like figs, to improve Sumatran orangutan HSM predictions. The Sumatran orangutan faces significant habitat challenges in GLNP due to habitat fragmentation, and limited monitoring coverage. To address these issues, conservation strategies should prioritize restoring connectivity between forest blocks, particularly through canopy corridors. Expanding monitoring efforts with advanced technologies, such as thermal drones, can improve data collection in remote areas. Additionally, enhancing habitat quality by restoring and protecting key food plant species will ensure sustainable orangutan populations.