In-stream Denitrification: Pilot Studies and Site Characteristics on a Headwater Stream by Alison J. Milliman a Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science in Environmental Science University of Rhode Island

Increased loads of watershed nitrogen (N) are a major cause of water quality degradation, especially in coastal waters. According to regional budgets, only 20-30% of watershed N inputs reach coastal waters indicating there are N “sinks” across the landscape. Headwater streams can generate substantial N processing; however, several studies have observed minimal N retention within-streams. We assessed the reach-scale in-stream N removal and denitrification rates in one Rhode Island headwater stream. Using the constant rate injection method, we tracked the fate of isotopically enriched nitrate along a 500 meter stream reach. The objectives of this study were to assess the resolution of this method for use in Rhode Island, to investigate the sensitivity of the method to seasonal variation and to evaluate the extent of the intra-seasonal variation. In-stream NO3-N removal was estimated using three approaches: NO3 disappearance using NO3: Br ratios, denitrification gas flux using the N signature of the dissolved N2 and N2O gases, and mass balance of the NO3-N in the stream water. Removal was examined in summer and fall with three trials per season. During each trial the stream was dosed at a rate of 0.002 L second for 2.9 hours. The dosing solution contained N-enriched potassium nitrate (N-KNO3), a conservative solute tracer potassium bromide (KBr) and a conservative gas tracer sulfurhexafluoride (SF6). Stream water samples were collected at six downstream stations and analyzed for nitrate, isotopic enrichment in the stream water, dissolved gases and concentration of conservative tracers. DOC, pH, and temperature were also assessed. A solute transport model, One Dimensional Transport with Inflow and Storage (OTIS) was used to estimate stream hydrologic parameters such as residence time and degree of entrainment. Uptake length, which is used to determine the average distance traveled by a nitrate molecule before it is denitrified in a stream reach, was calculated directly from denitrification rates and by using the Spatially Referenced Regressions on Watershed Attributes (SPARROW) model. Although flow rate, depth, temperature and DOC varied between the summer and fall trials, stream velocity and ambient NO3-N concentrations were very similar between seasons. Given the assumption that NO3-N transformation rates will follow first order kinetics (Mulholland et al., 2000), the lack of seasonal differences in these characteristics could explain the lack of difference in reach-scale NO3-N transformation between the summer and fall. In one trial of both the summer and fall experiments, the NO3 disappearance approach indicated significant N removal, but this method displayed high variability within the reach and within a season. The N gas flux and the mass balance of NO3-N consistently generated results indicating little or no N removal for all trials in the summer and fall. The observed uptake length was greater than 170 km, which agreed with predictions generated using the SPARROW model. Results from the OTIS model indicate that the stream storage zone area and storage zone exchange were small in the summer and fall. Seasonal variation in stream hydrologic and chemical conditions did not appear to affect the consistent, negligible removal rates found with the N data. Similarly, the intra seasonal variations in-stream characteristics did not have a notable impact on the rates of N removal. The denitrification rates observed in this study are comparable to several studies; but a number of studies found markedly higher amounts of N disappearance. The minimal denitrification rates coupled with the low retention times associated with this headwater stream argue that it is a negligible sink for watershed N. The uncertainties surrounding in-stream N removal warrant further investigation, with consideration given to headwater streams with extended retention times and benthic interactions. ACKNOWLEDGEMENTS I would like to thank my thesis committee members/ advisors: Dr. Art Gold, Dr. D.Q. Kellogg, Dr. Candace Oviatt and Dr. Charles Roman. For help with everything I would like to thank Kelly Addy. I would like to thank Marissa Kelly for all her hard work I the field and lab. In addition, I would like to thank Sophia Narkiewicz, Opeyemi Odewale, Tara Watson, Yuyu Zhou, Holly Meehan, Matt Richardson, Meagan Hopkins, Brian Clarkin, Peg Pelletier, Cathy McNally and NRS 361 2007 students for assistance in the field. Thanks to Linda Green, Dr. Tom Boving, and Dr. Barbara Nowicki for analytical assistance and the use of their equipment.