Nitrate depletion dynamics and primary production in riverine benthic chambers

Aquatic N cycling arises from multiple overlapping and interacting pathways. In lotic ecosystems, time-varying upstream delivery complicates isolation of internal biogeochemical processing signals. To experimentally evaluate process-specific reaction kinetics in response to changing nitrate (NO3-) concentrations, we evaluated high-frequency time series from 38-wk-long deployments of clear benthic chambers in 4 spring-fed rivers in north Florida. Three of these rivers are heavily enriched in NO3-, and in these rivers we observed rapid NO3- depletion over time as well as diel variation that suggested daytime autotrophic assimilation (U-A). We evaluated the fit of models of varying complexity to these data and the best-fit models typically explained >70% of observed NO3- uptake variability. These models suggest both U-A and net dissimilatory uptake, or denitrification minus nitrification (U-D), follow efficiency-loss kinetics, but that U-D is far more concentration dependent (mean exponent = 0.88 vs 0.47). Notably, U-A often declined with NO3- depletion, but gross primary production did not, suggesting autotrophs adjust to alternative N sources (e.g., ammonium [NH4+]). This conclusion was supported by marked alterations in NO3- depletion in response to NH4+ additions. In the unenriched river we noted minimal net uptake of NO3-, suggesting low rates of denitrification relative to nitrification. Furthermore, despite high gross primary productivity, we observed diel variation suggesting limited NO3- assimilation, presumably in favor of NH4+. These conclusions were supported by a shift in the NO3- signal when NO3- was added such that it matched the diel signal observed in the enriched rivers. In all sites, we observed evidence of diel variation in non-assimilatory pathways such as denitrification and nitrification. As U-A of NO3- declines with concentration, these alternative pathways potentially become the dominant source of diel variation. However, including additional parameters to represent these processes led to poor model identifiability. Thus, additional solute time series (e.g., NH4+ or N-2) may be necessary to adequately resolve the overlapping processes responsible for complex NO3- time series.