Information about diffuse pollutant sources in river catchments and the processes responsible for their generation is often obtained using geochemical tracing, in‐stream monitoring, or process based modeling. Yet these methods are rarely combined. In this study, we combine these three methods to produce a catchment sediment budget for the Lake Burragorang catchment in Australia. The physically based process model SedNet was modified using local catchment data to constrain predicted hillslope erosion rates in forested areas, the rate and texture of sediment supply from gully erosion, and the predicted floodplain deposition rates. The model's capacity to match the geochemical tracer (both spatial source and erosion process) data and gauging station load data improved markedly. Hillslope erosion, primarily in the steeper, mainly forested areas near the reservoir, is identified by the model as the main source of sediment delivered to Lake Burragorang. The contribution from gully and river bank erosion is comparatively small, though gully erosion can be locally dominant. This result is consistent with the tracer data. Utilizing multiple independent data sets allows for a more rigorous evaluation of model performance and, it is argued, increases the likelihood that the model is correctly representing the main components of the catchment's sediment budget. The identification of diffuse sediment sources in other catchments can clearly benefit from combining a range of observational data with a structured model evaluation process.