Tracking the multiple sources of sediment within an agricultural catchment in Northeast China

Abstract

Understanding the multiple sources of sediment is essential for effective soil erosion mitigation and catchment-scale management. However, in the Mollisol region of Northeast China, the complex interplay of human activities, land use changes and developing gullies present significant challenges for accurately identifying sediment sources at the catchment scale. This difficulty hinders a comprehensive understanding of erosion processes in this region. This study employed geochemical fingerprinting techniques to estimate sediment contributions within an agricultural catchment in Northeast China, utilizing three source classification schemes. The goodness-of-fit and virtual mixtures by using a suite of evaluation metrics were employed to evaluate the performance and uncertainty of the mixing model, thereby confirming the applicability of this technique to the study area. Compared to the two-source approach, the multi-source classification schemes indicated lower variability in geochemical elements and reduced model uncertainty. All three source classification schemes consistently identified cultivated land and gully banks as the primary sediment sources. In the four-source classification scheme, the average sediment contributions from cultivated land, gully banks, ephemeral gullies and unsealed roads were 48.1 % ± 26.3 %, 27.8 % ± 18 %, 19.9 % ± 21.6 %, and 4.2 % ± 3.9 %, respectively. Moreover, spatial heterogeneity in sediment contributions across sub-catchments suggests that variations in key erosion drivers, such as land use intensity and gully density, may influence the primary sediment sources within the catchment. These findings provide actionable insights for targeted catchment management and the preservation of Mollisols in Northeast China, advancing strategies to mitigate soil degradation in agriculturally sensitive regions.