Rivers can rapidly process nitrate via assimilatory and dissimilatory pathways. We typically measure these fluxes using experimental additions of nitrate, but the advent of high-frequency sensors allows rates to be estimated by fitting ambient nitrate dynamics to process models. Here, we couple high-frequency nitrate data collected with co-calibrated SUNAs deployed in a spring stream and in the Nyack reach of the Middle Fork Flathead River. We confronted these data with two different two-station models of uptake based on zero- and first-order kinetics, i.e., where nitrate uptake depends on its concentration. One-station models with first-order kinetics can satisfactorily estimate both assimilatory uptake via photoautotrophs and the overall rate of gross nitrate turnover, but they assume long upstream reaches. Two-station models account for reach heterogeneity. Models with zero-order kinetics showed no net change in nitrate flux across the reaches. Models with first-order kinetics showed that although there was no net change in nitrate flux, turnover was high, indicating high gross processing rates along each reach. Sensor data coupled with models allow estimation of nitrate dynamics in rivers that are far too large to easily conduct tracer experiments.