Developing and applying a soil erosion model in a data-poor context to an island in the rural Philippines

Soil erosion in many parts of the developing world poses a threat to rural livelihoods, to the sustainbility of the agricultural sector, and to the environment. Most erosion prediction models are mechanistic and unsuited to quantify the severity of soil erosion in a data-limited developing world context. The model developed in this paper for Negros Island, in the central Philippines, is based on the Revised Universal Soil Loss Equation, but contains important innovations such as the movement of eroded soil over the landscape, simulating deposition on lower slopes and in waterways. It also includes a term describing farmer strategies to reduce soil erosion, which are typically ignored in erosion prediction models. A two-sample t-test found that model-predicted sediment loading values were not significantly different from field-measured sediment loading values when corrected for watershed size (P = 0.857). The model predicts an annual loss of 2.7 million cubic meters of sediment to waterways such that by 2050 more than 416,000 ha of agricultural land will be rendered unproductive due to erosion. Farmer behavior conserves soil, but on the steepest slopes soil conservation practices are not adequate to prevent erosion. Of two proposed strategies to control soil erosion in the rural Philippines, the model suggests that a complete switch to tree crops would conserve more soil than universal terrace adoption. However, even under these conservation scenarios, erosion threatens the areal extent of upland agriculture on Negros, and hence the sustainability of the island’s food supply.