Sensitivity analysis of a pulse nutrient addition technique for estimating nutrient uptake in large streams

The constant nutrient addition technique has been used extensively to measure nutrient uptake in streams. However, this technique is impractical for large streams, and the pulse nutrient addition (PNA) has been suggested as an alternative. We developed a computer model to simulate Monod kinetics nutrient uptake in large rivers and used this model to evaluate the sensitivity of the PNA technique. We parameterize our model using the average hydrogeomorphological estimates from a pulse release study of ammonium in the Snake River, WY, and used this study to demonstrate how data from a field experiment can be effectively analyzed using a simulation model. To evaluate the sensitivity of the PNA technique, we manipulated the hydrogeomorphology and uptake kinetics of our stream model, simulated a pulse ammonium addition, and measured the downstream response in our model as if it were a field experiment, while ammonium areal uptake at ambient concentration was kept unchanged in the model. Ammonium uptake estimates by the PNA technique were different from the uptake in our model and these differences were nonrandom. The difference was greatest when velocity was high and there was little solute spread, either in the water column or from exchange with transient storage. The difference was also high when the half saturation coefficient for uptake was low. Our estimates of ammonium uptake under the assumption of Monod kinetics were higher than those under the assumption of first-order kinetics based on direct calculation from the experimental data. *Corresponding author: E-mail: [email protected] Acknowledgments We thank Jennifer L. Tank for providing data and information for this study and Robert L. Runkel and Robert A. Payn for constructive suggestions. We also thank Dr. Luc Claessens for guidance during this study. DOI 10.4319/lom.2012.10.718 Limnol. Oceanogr.: Methods 10, 2012, 718–727 © 2012, by the American Society of Limnology and Oceanography, Inc. LIMNOLOGY and OCEANOGRAPHY: METHODS nique (PNA technique) was suggested as an alternative for uptake measurement in large streams. This technique was first used by Meals et al. (1999) and more recently used to estimate nutrient uptake parameters by Tank et al. (2008) and Powers et al. (2009). Using the PNA technique, a slug of nutrient and biological conservative tracer is released, and their concentrations are monitored at several downstream stations. To measure nutrient uptake, Tank et al. (2008) (TPNA technique) calculated the total mass of nutrient and conservative tracer at each station by time integration. The logarithmic change in the ratio of nutrient mass to conservative tracer mass over downstream distance represented the longitudinal uptake rate. In the constant nutrient addition technique, the addition can be kept very small to minimize saturation effects. However, the PNA technique requires a relatively large but shortterm nutrient elevation and sometimes leads to large enrichment relative to ambient conditions. Although the limitations of the PNA technique may be understood conceptually, the quantitative impacts of different hydrogeomorphology and different uptake saturation levels on the PNA technique have not been documented. In this study, our objectives were (1) to develop a computer model to simulate nutrient uptake in a large river and quantitatively evaluate the sensitivity of the TPNA technique to different hydrogeomorphology and uptake saturation levels; and (2) to demonstrate how data from a field PNA experiment can be effectively analyzed to estimate nutrient uptake using a simulation model. We parameterize our model using the average hydrogeomorphological estimates of the Snake River, WY (Tank et al. 2008). Materials and procedures Model development We began with the basic model of advection and longitudinal dispersion, including inflow, transient storage, and nutrient uptake in the water column and transient storage (e.g., Bencala and Walters 1983; Runkel 1998), but we did not include lateral inflow as the data of Tank et al. (2008) did not indicate a downstream increase in discharge: