Inferring nitrogen removal in large rivers from high-resolution longitudinal profiling

We present a method for estimating nitrogen (N) removal based on high-resolution longitudinal profiling, which facilitates repeated measurement in larger rivers. The Lagrangian reference frame allows removal to be spatially disaggregated, enabling identification of removal hot spots,'' and potentially passive assessment of reaction kinetics using ambient longitudinal concentration gradients. Applying the method in six spring-fed rivers in North Florida, we tested the hypothesis that removal is controlled by spatial variation in channel hydraulics and vegetation. Removal estimates obtained using this method were statistically robust, spatially consistent across replicate profiles, and comparable to measurements made in these and other systems of similar size using alternative methods. We observed significantly increased removal in reaches with lower specific discharge and in more heavily vegetated reaches. Assessing uptake kinetics directly from longitudinal profiles assumes negligible sampling effects from temporally variable removal, non-instantaneous sampling velocities, and the mixing of water along the flowpath. Results from a reactive transport model suggest these effects are significant, producing complex profile geometries. Relatively small differences between alternative kinetic models substantially limit inferences about reaction order when utilizing the small concentration gradients observed longitudinally within rivers. Across rivers, N removal follows either efficiency-loss (exponent = 0.39) or Michaelis-Menten kinetics, but not zero- or first-order kinetics.