Abstract

Topsoil erosion poses a significant threat to agricultural productivity and environmental sustainability, particularly in regions with intensive land use. This study evaluates topsoil erosion in a 500-hectare agricultural watershed in central Iowa, USA, using Unmanned Aerial Vehicle (UAV) imagery and machine learning techniques. High-resolution multispectral and RGB images were collected using a DJI Phantom 4 Pro UAV, complemented by ground truth data from 120 soil erosion sampling points. Random Forest (RF) and Gradient Boosting Machine (GBM) models were employed to predict erosion rates, incorporating variables such as slope, vegetation cover, and rainfall intensity. The RF model achieved a prediction accuracy of 88% with a root mean square error (RMSE) of 2.1 t/ha/year, outperforming GBM (84%, RMSE 2.5 t/ha/year). UAV-derived digital elevation models (DEMs) and vegetation indices significantly enhanced prediction accuracy. The study demonstrates the potential of UAV-based remote sensing combined with machine learning for high-resolution erosion mapping, providing actionable insights for soil conservation