Open-channel estimation of denitrification

Estimates of denitrification based on standard methods often cannot easily be extrapolated to entire ecosystems because of high spatial heterogeneity in rates of denitrification. Diel changes in the concentration of dinitrogen (N2) in running waters provide the basis for estimating denitrification by an open-channel technique in much the same way that ecosystem metabolism can be estimated from diel changes in the concentration of dissolved oxygen. The open-channel N2 method was field tested at a site on the South Platte River downstream from Denver, CO; concentrations of N2 were measured by membrane inlet mass spectrometry. For a date in November 1998, the rate of denitrification was estimated to be 0.19 mol N2 m –2 d–1 and was similar to rates reported in previous studies based on mass-balance analysis. The open-channel N2 method is the first method to provide direct, whole-system estimates of denitrification in flowing waters and may help to expand our understanding of nitrogen cycling in running waters. Acknowledgments We thank S. Chavez de Baca, S. Kaushal, C. McGrath, and T. Rosenstiel for help with sampling and L. Smith for assistance in the laboratory. We also thank S. Joye, P. Kemp, and an anonymous reviewer for their comments and suggestions. This work was funded by the Regional Integrated Sciences and Assessments (RISA) Program of the National Oceanic and Atmospheric Administration (NOAA; award NA17RJ1229), a fellowship from the Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, and a grant from the Andrew W. Mellon Foundation to G. Likens, Institute of Ecosystem Studies (IES). This paper is a contribution to the Program of the Institute of Ecosystem Studies. Limnol. Oceanogr.: Methods 1, 2003, 74–81 © 2003, by the American Society of Limnology and Oceanography, Inc. LIMNOLOGY and OCEANOGRAPHY: METHODS MIMS allows rapid and precise measurement of N2 and is gaining widespread use in studies of sediment cores (Eyre et al. 2002). Typically, determination of N2 concentration by MIMS has been based on measurements of the N2 : Ar ratio and assumptions about the concentration of Ar. If Ar concentration in water is always at equilibrium with the atmospheric mixing ratio for ambient temperature and pressure (i.e., at saturation), N2 concentration can be computed from the N2 : Ar ratio. In natural aquatic systems, however, the assumption that Ar is at saturation rarely is valid. Even though Ar is biologically inert, the concentration of Ar changes on a diel basis in response to changes in temperature and pressure; diel variations in temperature and pressure result in continuous variation of the saturation concentration. The ambient concentration for a gas dissolved in water is driven toward saturation through exchange with the atmosphere (reaeration), but there is always a time lag between changes in saturation and changes in gas concentration. Because it cannot be assumed that Ar remains at saturation in natural systems, especially where the diel amplitude of temperature is large and the reaeration rate coefficient is low, application of the open-channel method to estimation of denitrification is most feasible through measurements of N2 concentration that do not depend on assumptions about Ar concentrations. Laursen and Seitzinger (2002) estimated concentrations of N2 from direct measurements of Ar concentration and the N2 : Ar ratio. Direct measurements of N2 concentration are possible with MIMS but require greater care than measurements of the N2 : Ar ratio. The purpose of this paper is to describe an open-channel method for the estimation of denitrification in running waters. The technique described here depends on measurement of the concentration of dissolved N2 independently of other gas concentrations by MIMS. Hourly measurements were made over a 24-h period at a single station and estimation of denitrification was based on calculations of N2 flux, after correction for reaeration flux and groundwater flux. As shown here for a site on the South Platte River in Colorado, gas concentrations can be measured independently and precisely with MIMS, but only after careful calibration and correction for drift. Estimates presented here do not include losses of N2O because N2 accounts for nearly all (>99%) of the gaseous losses associated with denitrification in the South Platte River (McMahon and Dennehy 1999). Materials and procedures Site description—Diel changes in dissolved N2 were measured on 14 November 1998–15 November 1998 at a site on the South Platte River. The South Platte drains an area of almost 63,000 km2 in Colorado, Wyoming, and Nebraska. Wastewater from the city of Denver is a major source of fixed nitrogen to the South Platte, as is agriculture. Previous studies of the South Platte have shown that denitrification is an important component of the mass balance for nitrogen (McMahon and Bohlke 1996; Sjodin et al. 1998). This study was conducted at a site near Ft. Lupton, 55 km downstream of Denver (40°14′12′′N, 105°38′3′′W). At low flow, which characterized the period of sampling, the channel averages 40to 50-m wide and is 30to 70-cm deep. Sampling and field measurements—Water samples for analysis of gas concentrations were collected hourly at three locations spaced evenly along a transect across the river (i.e., perpendicular to the channel). N2 samples were collected in 40-mL glass vials with 3.2-mm Teflon/silicone septa (I-Chem part nr. SB46-0040). Vials were filled with a sampler that flushed the vials with approximately 5 volumes of water, thus minimizing exchange of gases with the atmosphere (Kilpatrick et al. 1989). Vials were immersed in cold water (ca. 0°C) immediately after filling and were held there until analysis (within 12 to 24 h); no preservative was added to samples. Shallow groundwater also was collected from piezometers adjacent to the river, as necessary to correct for changes in N2 concentration caused by seepage (McCutchan et al. 2002). Water temperature and barometric pressure were recorded at 10-min intervals; temperature was measured with a Hydrolab Datasonde III, and pressure was measured with a field barometer (Oregon Scientific). Discharge was estimated from measurements of stream depth and cross-sectional velocity profiles, as measured with a Marsh-McBirney flow meter, and the rate of groundwater seepage (m3 s–1 d–1) was estimated from paired measurements of discharge (i.e., at the sampling location and at another point upstream). Samples for analysis of nitrate and ammonium concentrations were collected hourly. Propane was used as a volatile tracer to estimate the reaeration rate (Kilpatrick et al. 1989). The rate coefficient for nitrogen (knitrogen) was derived from the rate coefficient for propane using the indexing method of Gulliver et al. (1990). The reaeration rate was corrected for temperature using the following equation (Thomann and Mueller 1987): knitrogen,T = knitrogen,20°1.024 (T – 20) (1) where knitrogen,20° is the reaeration rate for nitrogen at 20°C. Measurement of gas concentrations—Concentrations of dissolved N2 were measured with a Balzers Prisma quadruple mass spectrometer with a membrane inlet (Kana et al. 1994). Precise and independent measurement of N2 concentration depended on careful preparation of standards for calibration, slight modifications to the inlet line, and correction for machine drift. Analytical precision was further improved by correcting measured concentrations for diffusion during storage. Samples suspected of having high gas concentrations were diluted prior to analysis to avoid the formation of bubbles in the inlet line. For this study, the membrane inlet of the MIMS was partially immersed in a water bath at constant temperature (5.0 ± 0.05°C), and samples were drawn through the inlet line with a peristaltic pump. The water bath was covered with Styrofoam insulation and a submersible aquarium pump was used to irrigate the upper portion of the inlet. Maintenance of the McCutchan et al. Open-channel estimation of denitrification