Estimation of organic carbon source composition and riverine outflow using an integrated watershed hydrological–carbon modelling approach

Carbon source apportionment and outflow estimation are the primary scientific considerations for reducing carbon output from watershed ecosystems to ocean. However, carbon loss and transportation mechanisms from soil to river system driven by watershed hydrological cycle, remain unclear. Our study developed a process–based watershed organic carbon model that integrates soil biogeochemical processes, overland loss, riverine metabolism and transportation driven by hydrological processes, and estimates the sources, outflows and their spatial distributions. The proposed model was validated using long-term field observations of runoff and labile particulate, dissolved, and total organic carbon (LOC, DOC and TOC) loads across the Xiangxi Watershed in China. The biases within ±0.25 were for all runoff simulations and for 71.4 % (30/42) of carbon load simulations, and both Nash–Sutcliffe efficiency and correlation coefficient were over 0.60 for runoff simulation and for 83.3 % (35/42) of carbon load simulations. Annual average TOC load flowing into rivers was 11.3 ton^.km^-2.yr^-1, with resistant particulate organic carbon (ROC) as the main form, accounting for 88.7 % of the TOC load. Atmospheric deposition was the primary TOC source with a contribution of 87.9 %, followed by soil loss. Annual average riverine TOC outflow was 3.8 ton^.yr^-1, with LOC and DOC accounting for 57.5 % and 40.0 %, respectively. This indicates that a majority of ROC decomposed into DOC and LOC via riverine metabolism and sedimentation. Our study provides insights into integration mechanisms of watershed hydrological and carbon cycles, and contributes to strategies for controlling water and carbon losses to strengthen terrestrial carbon sequestration.