Isotopic Apportionment of Atmospheric and Sewage Nitrogen Sources

Aug 11, 2007 - to Long Island Sound. In contrast to previous work, we found that sewage effluent NO3. - was not consistently enriched in 15N. However,...
0 downloads 0 Views 227KB Size
Environ. Sci. Technol. 2007, 41, 6363-6369

Isotopic Apportionment of Atmospheric and Sewage Nitrogen Sources in Two Connecticut Rivers S H I M O N C . A N I S F E L D , * ,† REBECCA T. BARNES,† MARK A. ALTABET,‡ AND TAIXING WU‡ School of Forestry and Environmental Studies, Yale University, New Haven, Connecticut 06511 and School for Marine Science and Technology, University of Massachusetts, New Bedford, Massachusetts 02744

We used the dual isotope approach to identify sources of nitrate (NO3-) to two mixed land-use watersheds draining to Long Island Sound. In contrast to previous work, we found that sewage effluent NO3- was not consistently enriched in 15N. However, these effluents followed a characteristic denitrification line in δ15N-δ18O space, which could be used as a source signature. We used this signature, together with those of atmospheric deposition and microbial nitrification, to calculate ranges of possible contributions from each of these sources. These estimates are unaffected by any denitrification that may have taken place in soils or streams. Our estimates for atmospheric nitrogen only include unprocessed atmospheric deposition, i.e., NO3that is not taken up in watershed soils before being delivered to rivers. Using this method, the contribution of atmospheric NO3- could be assessed with good precision and was found to be very low at all our sampling sites during baseflow. During a moderate storm event, atmospheric deposition contributed up to ∼50% of stream NO3-, depending on the site, with the sites that experienced more stormflow showing a greater contribution of atmospheric NO3-. Our estimates of sewage contribution generally had too large a range to be useful.

Introduction Excess nitrogen loading to coastal ecosystems is a widespread phenomenon that can lead to a variety of harmful effects, including hypoxia, harmful algal blooms, and loss of submerged aquatic vegetation (1). As a result, there has been great interest in developing methods for quantifying the contributions of different anthropogenic and natural sources of nitrogen to rivers and estuaries. One such method involves the use of natural variations in stable isotopic ratios to identify sources of riverine nitrate (NO3-). Methodological advances (2-5) now permit the use of both N and O isotopes, which allows greater separation of source signatures and identification of a larger number of sources (6). In particular, the high 18O content of atmospherically derived NO3- can facilitate estimation of the contribution of atmospheric deposition to N budgets (7-12), an issue that has been of great interest (13). * Corresponding author phone: (203)432-5748; fax: (203)432-3929; e-mail: [email protected]. † Yale University. ‡ University of Massachusetts. 10.1021/es070469v CCC: $37.00 Published on Web 08/11/2007

 2007 American Chemical Society

Isotopic source apportionment is complicated by the fact that the isotopic composition of a given NO3- sample reflects not just mixing of sources, but also any isotope fractionation associated with denitrification (6). However, denitrification can frequently be assessed from dual N and O studies due to the characteristic upward shift in both the δ15N and the δ18O of the remaining NO3- (14-17). Dual isotope source apportionment has been used successfully in forested watersheds to distinguish between atmospheric nitrate and nitrate derived from microbial nitrification (7-11), and in agricultural watersheds to distinguish between agricultural, atmospheric, and groundwater sources (12). Similarly, this technique has been used in urban and mixed land-use settings to identify the contributions of NO3- from human waste, fertilizer, manure, and atmospheric deposition (18-23), although this has been somewhat more challenging because of the poorer isotope signature separation between sources. Many of these dual isotope studies have attempted to identify NO3- contributions from human or animal waste by taking advantage of its relatively high 15N signature (6, 24), which is a result of food web enrichment as well as loss of isotopically light nitrogen during waste treatment (e.g., through ammonia volatilization). In addition, several researchers have traced this enriched 15N signal through aquatic foodwebs, and proposed that a high δ15N in an aquatic ecosystem can be used as a marker of wastewater inputs (25-27). However, several sewage-tracing studies have not actually measured the isotopic signature of sewage released to their systems, but instead have relied on literature values, which are based on a relatively small number of measurements, most of which have been for manure and septic systems, not sewage treatment plants. Long Island Sound (LIS) has experienced recurrent summertime hypoxia that has been attributed to excess nitrogen inputs. As a result, a large-scale effort to control nitrogen sources to LIS is underway, within the framework of the total maximum daily load (TMDL) program (28). Current estimates (based on nonisotopic methods) are that almost 75% of the in-basin N loading to LIS is derived from point sources, namely sewage treatment plants (STPs) (28). Other important sources of N to LIS include septic tanks, fertilizer runoff, animal waste, and release from soils. In addition, there has been great interest in recent years in atmospheric deposition of N, both to LIS directly and to its watershed, and questions remain about the importance of this source. We describe here our work in using the dual isotope approach to identify NO3- sources during both baseflow and stormflow in two urbanized rivers draining to LIS. We focus primarily on the contributions of sewage and atmospheric deposition.

Study Sites Our study sites consisted of the Quinnipiac and Naugatuck Rivers in central Connecticut (Figure 1). For both rivers, we defined our study area as the nontidal portion of the watershed. The nontidal Quinnipiac Riversas defined by the U.S. Geological Survey (USGS) gauging station at Wallingford, CT, no. 01196500shas a basin area of 298 km2 and an average flow of 6.14 m3 sec-1 (1931-2005). Three municipal STP’s (Southington, Cheshire, Meriden) discharge secondarytreated effluent into the river. At the time of our sampling, all three STP’s were carrying out nitrification but were not designed to carry out denitrification (though see below), so that the N in their effluent was largely (>80%) NO3-. The VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6363

TABLE 1. Summary of Sampling Dates and Conditions, and Results of the Mass Balance Calculation for the Contribution of STPs to River NO3-

FIGURE 1. Location of the study sites. The three types of sampling sites are indicated: the outlets of the (non-tidal) watersheds; the STP’s where effluents were collected; and the tributary sites (including the upper portion of the main stem). nontidal Naugatuck Riversas defined by the USGS gauging station at Beacon Falls, CT, no. 01208500shas a basin area of 673 km2 and an average flow of 14.8 m3 sec-1 (1929-2005). Three municipal STP’s (Torrington, Thomaston, Waterbury) discharge tertiary-treated effluent into the river. At the time of our sampling, all three plants were carrying out denitrification, resulting in relatively low effluent N concentrations. For each river, we collected three types of samples: final effluents from each of the STPs (with the exception of Torrington); samples from the outlets of all the important tributaries (including the upper portion of the main stem); and a sample from the main stem at the outlet of the watershed (i.e., at the USGS gauging stations listed above). The STP’s all discharge into the main stem, so none of the tributaries receive any direct sewage inputs, with the exception of the upper Naugatuck, which receives effluent from the Torrington STP. Land use in the tributary watersheds ranged from 21 to 70% developed (including agricultural, urban, and suburban land; Table S1, Supporting Information). To assess the potential for inputs of N from septic systems, we calculated the percent of each watershed that consisted of urban/suburban land that was not sewered; this ranged from 9.9 to 25% for the tributaries that did not receive sewage effluent (Table S1). The Quinnipiac sites were sampled on three dates in the summer of 2005 (two baseflow dates and one storm), whereas the Naugatuck was sampled once during baseflow (Table 1). Hydrographs for the summer of 2005 are shown in Figure S1 (Supporting Information).

Materials and Methods Samples for nutrient and isotopic analysis were filtered through 0.45 µm membrane filters and frozen for later analysis. NH4+ and NO2- were measured using an Astoria2 6364

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007

date

watershed

6/16/05 7/25/05 7/28/05 8/19/05

Quinnipiac Naugatuck Quinnipiac Quinnipiac

flow at flow outlet conditions (m3/sec) baseflow baseflow stormflow baseflow

3.11 3.37 3.99 1.98

STP flow STP NO3as % of flux as % of outlet outlet NO3- flux flow 22% 30% 17% 26%

80% 80% 57% 81%

colorimetric flow analyzer. NO3- was measured using a Dionex ion chromatograph. Total dissolved N (TDN) was measured on a Shimdazu TDN analyzer. Isotopic analysis was carried out using the azide reduction method (5). The average standard deviations for lab replicates (n ) 11) and field replicates (n ) 6) were, respectively, 0.63 and 0.54‰ for δ15N and 0.57 and 0.55‰ for δ18O. Note that isotope values reported for NO3- are actually for the sum of NO3- + NO2-; however, [NO2-] was always 80% of TDN. NH4+ was almost always