๐ AI predicts where river soil erosion comes from
Researchers developed a machine learning framework to predict sediment sources for better watershed management. This study analyzed 142 sediment tracing studies across 267 watersheds globally. The model achieved robust accuracy (R2 = 0.40โ0.53) across subsurface, cultivated, non-cultivated, and infrastructure sources. Testing in the US and UK showed different erosion patterns, like subsurface dominance in the Upper Mississippi. This tool offers a data-driven way to map sediment sources for targeted environmental action. ๐