Modelling Habitat Suitability for African Elephants (Loxodonta africana) Preferred Forage in a Savannah Ecosystem in Kenya

Abstract

Understanding the environmental determinants of habitat suitability for African elephants’ preferred forage plants is essential for effective conservation, especially in the face of climate change. This study assessed the current and future habitat suitability for elephant forage species within a savannah ecosystem in Kenya. An optimised Maximum Entropy (MaxEnt) modelling approach was employed using 152 spatially independent occurrence data points and eight non-correlated environmental variables. The model used a Linear-Quadratic feature combination and a regularisation multiplier of 1.5. Occurrence data were collected through field surveys, while environmental layers were obtained from publicly available datasets including WorldClim, ASTER GDEM, USGS, HydroSHEDS, and OpenStreetMap. Future climate projections were based on the CanESM2 model under SSP245 and SSP585 scenarios for the periods 2041–2060 and 2081–2100. Model performance was evaluated using the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC), yielding strong results (AUC = 0.883 ± 0.007). Precipitation Seasonality (bio_15), Precipitation of the Driest Quarter (bio_17), elevation, and Land Use Land Cover (LULC) were the most influential predictors. Under current conditions, unsuitable habitats dominate the landscape (73.42%), while low, moderate, and high suitability zones are relatively limited. Future projections indicate continued dominance of unsuitable habitats with minor changes (1.28%–3.78%), slight increases in high and moderate suitability areas, and notable declines in low suitability zones across all climate scenarios. These findings highlight the vulnerability of forage plant habitats to climate change and underscore the need for climate-informed conservation strategies to ensure the long-term availability of key food resources for African elephants in this dryland ecosystem.