Modeling hillslope sediment yield using rainfall simulator field experiments and partial least squares regression: Cahaba River watershed, Alabama (USA)

Abstract

Hillslope erosion processes such as rainsplash erosion and slopewash are a major source of fine sediment that causes water-quality problems in many river systems. Hillslope sediment yield is controlled by both climatic and watershed characteristics. Understanding the relative sensitivity of sediment yield to soil and watershed characteristics under different precipitation intensities is important for identifying hotspots of sediment yield and the likely response of sediment yield to changing climate and land use. Here, field experiments on sediment yield were conducted in Alabama’s (United States) Cahaba River watershed under varying precipitation intensities using a rainfall simulator. Using data on soil and watershed characteristics, partial least squares (PLS) regression models were developed of sediment yield under both “more intense” (recurrence interval ≈ 10 years) and “less intense” (recurrence interval ≈ 1 year) simulated rainfall events. The resulting models were used to create spatially explicit estimates of sediment yield under different precipitation intensities for the Cahaba River watershed. The optimized PLS models had a cumulative R 2 of 0.41 and 0.73 for the “more intense” and “less intense” rainfall events, respectively. The higher R2 of the “less intense” model is attributed to higher significance of the heterogeneity of watershed variables under less intense precipitation. Significant variables that were retained in both models included percent sand, percent clay, percent organic matter, and slope. The results support the use of PLS modeling to analyze the results of field experiments using rainfall simulators to examine both climatic and watershed controls on hillslope sediment yield.