Fingerprinting generates reliable sediment provenance information which supports devising policy or practical strategies for soil conservation and sediment management, but it remains challenging in areas with fragmented landscapes and diverse land use practices. This study evaluated the seasonality of biomarker signatures and their variability among particle size fractions, and apportioned target time-integrated suspended sediment to land use-based sources in an intensive agricultural watershed with mosaic land use patch configurations and crop-specific farming practices. Source materials (i.e., topsoil) from dry croplands, paddy fields and citrus orchards were sampled, and target time-integrated suspended sediment samples were collected at the watershed outlet. The content and compound-specific delta C-13 of long-chain saturated n-alkanes (C-23-C-33) were determined for two particle size fractions (i.e., <25 mu m, 25-63 mu m). The delta C-13 of monomeric n-alkanes displayed insignificant variabilities between particle size fractions and temporal variability across the sampling period. The MixSIAR Bayesian model was employed to quantify sediment source contributions. Due to land disturbance by tillage and crop plantation, our results revealed that paddy fields act as an important temporary secondary sediment source despite such fields conventionally being recognized as sediment sinks. Regardless, dry farmland remains the largest contributor to watershed sediment loss. A range of measures such as soil virginization, returning straw to fields, and pasture cultures in orchards are recommended for precision sediment management at watershed scale.(c) 2025 International Research and Training Center on Erosion and Sedimentation, China Water & Power Press, and China Institute of Water Resources and Hydropower Research. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).