River systems are dynamic, highly connected water transfer networks that integrate a wide range of physical and biological processes. We used a river network nitrogen (N) removal model with daily temporal resolution to evaluate how elevated N inputs, saturation of the denitrification and total nitrate removal processes, and hydrologic conditions interact to determine the amount, timing and distribution of N removal in the fifth‐order river network of a suburban 400 km2 basin. Denitrification parameters were based on results from whole reach 15NO3 tracer additions. The model predicted that between 15 and 33% of dissolved inorganic nitrogen (DIN) inputs were denitrified annually by the river system. Removal approached 100% during low flow periods, even with the relatively low and saturating uptake velocities typical of surface water denitrification. Annual removal percentages were moderate because most N inputs occurred during high flow periods when hydraulic conditions and temperatures are less favorable for removal by channel processes. Nevertheless, the percentage of annual removal occurring during above average flow periods was similar to that during low flow periods. Predicted river network removal proportions are most sensitive to loading rates, spatial heterogeneity of inputs, and the form of the removal process equation during typical base flow conditions. However, comparison with observations indicates that removal by the river network is higher than predicted by the model at moderately high flows, suggesting additional removal processes are important at these times. Further increases in N input to the network will lead to disproportionate increases in N exports due to the limits imposed by process saturation.
