The utility of sediment budget analysis is explored in revealing spatio‐temporal changes in the sediment dynamics and morphological responses of a fluvial system subject to significant human impacts during the recent Anthropocene. Sediment budgets require a data‐intensive approach to represent spatially‐differentiated impacts adequately and are subject to numerous estimation uncertainties. Here, field and topographic surveys, historical data, numerical modelling and a representative‐area extrapolation method are integrated to construct a distributed, process‐based sediment budget that addresses historical legacy factors for the highly regulated Lagunitas Creek (213 km2), California, USA, for the period 1983–2008. Independent corroboration methods and error propagation analysis produce an uncertainty assessment unique to a catchment of this size. Current sediment yields of ~20 000 t a‐1 ± 6000 t a‐1 equate to unit rates of ~300 t km‐2 a‐1 ± 90 t km‐2 a‐1 over the effective sediment contributing area of 64 km2. This is comparable with yields associated with early Euro‐American settlement in the catchment, despite loss of sediment supply upstream of the two large dams. It occurs because ~57% of the sediment is now derived from incision‐related channel erosion. Further, the highly efficient routing of channel‐derived sediments in these incised channels suggests an efflux of 84% of contemporary sediment production, contrasting with the efflux of ≈10–30% reported for unregulated agricultural catchments. The results highlight that sediment budgets for regulated rivers must accommodate channel morphological responses to avoid significantly misrepresenting catchment yields, and that volumetric precision in sediment budgets may best be improved by repeat, spatially dense, channel cross‐section surveys. Human activities have impacted every aspect of the sediment dynamics of Lagunitas Creek (production, storage, transfer, rates of movement through storage), confirming that, while distributed sediment budgets are data demanding and subject to numerous error sources, the approach can provide valuable insights into Anthropocene fluvial geomorphology. Copyright © 2017 John Wiley & Sons, Ltd.