Coupled socio-ecological modelling of land-use dynamics and erosion risks in Bolango Watershed, Indonesia

Abstract

This study examines how land-use change alters soil erosion in the Bolango Watershed, Gorontalo Province, Indonesia, through an integrated framework that couples the revised universal soil loss equation (RUSLE) with agent-based modelling (ABM). Across the 52,494 ha watershed, marked transformations were projected between 2013 and 2033, including a 14.25% decline in forest cover, a 42.37% expansion of dryland agriculture, and a 122.24% increase in settlements. Spatial analyses implemented in geographic information system (GIS) software showed that mean soil loss decreased from 59.08 Mg∙ha−1∙y−1 in 2013 to 49.22 Mg∙ha−1∙y−1 in 2023, before rising slightly to 50.17 Mg∙ha−1∙y−1 in 2033. Rainfall erosivity (R) declined from 921.56 to 760.80 MJ∙mm∙ha−1∙h−1∙y−1, whereas the slope length–steepness factor (LS) ranged from 0.1 to 276.82, underscoring strong terrain control. Scenario simulations showed that contour farming reduced total erosion by 27%, while strict protection of slopes >25% reduced erosion by 51%, demonstrating the effectiveness of targeted conservation measures. The ABM represented interactions among farmers, government, and investors and showed that policy enforcement and environmental awareness strongly shape land-use decisions. Model validation yielded R2 = 0.83 and RMSE = 12.7 Mg∙ha−1∙y−1, indicating robust predictive performance. The coupled ABM–RUSLE framework therefore provides a rigorous basis for sustainable watershed management and spatial planning in tropical environments.