Isotopic characterization of nitrate sources and transformations in Lake Winnipeg and its contributing rivers, Manitoba, Canada

Journal of Great Lakes Research 38 (2012) 135–146

Contents lists available at SciVerse ScienceDirect

Journal of Great Lakes Research journal homepage: www.elsevier.com/locate/jglr

Isotopic characterization of nitrate sources and transformations in Lake Winnipeg and its contributing rivers, Manitoba, Canada Bernhard Mayer a,⁎, Leonard I. Wassenaar b, 1 a b

Department of Geoscience, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada T2N 1N4 Environment Canada, 11 Innovation Blvd., Saskatoon, Canada S7N 3H5

a r t i c l e

i n f o

Article history: Received 21 June 2011 Accepted 19 January 2012 Available online 8 March 2012 Communicated by Marley J. Waiser Keywords: Nitrate Stable isotopes Nitrogen isotope ratios Oxygen isotope ratios Nitrification Lake Winnipeg

a b s t r a c t Lake Winnipeg (Manitoba, Canada) is in a eutrophic state from a century of increased riverine loadings from agricultural and urban nitrogen (N) and phosphorus (P) sources. This study investigated seasonal patterns of the isotopic composition of nitrate (NO3−) in Lake Winnipeg and its contributing rivers to gain insight into current N nutrient sources and in-lake N dynamics. Elevated NO3− concentrations in Lake Winnipeg tributaries between 0.36 and 2.44 mg/L NO3−–N were associated with high δ 15N values between + 5.0 and + 13.9‰, while δ 18ONO3 values were b+ 15.0‰. The three major riverine inputs had distinctive mean δ 15NNO3 values of + 8.1‰ for the Red River, -0.6‰ for the Winnipeg River, and + 5.0‰ for the Saskatchewan River. The isotopic composition of NO3− in Lake Winnipeg was partly controlled by the isotopic composition of the riverine nitrate for instance via the predominant nitrate input to the South basin from the Red River. Nitrate assimilation and late season mineralization of phytoplankton and N2 fixing cyanobacteria were identified as important additional processes affecting the isotopic composition of lake NO3− resulting in low δ 15NNO3 values, especially in the North basin. In the South basin, elevated δ 15NNO3 values in spring that changed to lower values by summer indicated a dynamic N cycle within the lake. Agreement between δ 15N values of lake NO3−, PON and fish suggests that dissolved nitrate partially affects the flow of nitrogen in the aquatic food webs of Lake Winnipeg. © 2012 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Introduction Over the past 60 years human activity and agricultural industrialization with intensive fertilizer application have greatly affected the nitrogen (N) cycle of terrestrial and aquatic ecosystems (Vitousek et al., 1997). As a consequence of enhanced anthropogenic nitrogen loading, many key aquifers and water supplies, rivers, and estuaries around the world are negatively affected by increased dissolved inorganic nitrogen concentrations (Glibert et al., 2006; Goolsby et al., 2000; Mayer et al., 2002; Rupert, 2008). Consequences of excess dissolved inorganic N entering receiving waters include human health concerns, water quality degradation, eutrophication, red tides, and in some cases hypoxia of coastal oceans and inland lakes (Diaz and Rosenberg, 2008). Lake Winnipeg, a dual-basin great lake in Western Canada (Fig. 1) is currently in a eutrophic state due to decadal-scale increases in

⁎ Corresponding author. Tel.: + 1 403 220 5389. E-mail addresses: [email protected] (B. Mayer), [email protected] (L.I. Wassenaar). 1 Current address: Isotope Hydrology Laboratory, Department of Nuclear Science and Applications, Wagramer Strasse 5, P.O. Box 100, A-1400 Vienna, Austria. Tel.: + 43 1 2600 21766.

riverine loadings of agricultural and urban derived nitrogen (N) and phosphorus (P), primarily from the Red River and also from the Winnipeg River (Stewart et al., 2000; Yates et al., 2012). Extensive cyanobacterial blooms in Lake Winnipeg, particularly over the past decade (McCullough, 2009; McCullough et al., 2012), are generally attributed to this increased anthropogenic nutrient loading (Manitoba, 2008). Lake Winnipeg receives discharge and nutrients principally from three major riverine inputs; the Red and Winnipeg Rivers in the South basin, and the Saskatchewan River on the west side of the North basin (Fig. 1). Other potential nutrient sources are from smaller agriculturally affected rivers to the west (e.g. Dauphin and Icelandic Rivers), and the boreal Canadian Shield to the east although these are characterized by comparatively lower discharge and a lack of nutrient monitoring data. The movement and mixing of waters entering Lake Winnipeg follow counter-clockwise mixing gyres – one gyre in the south and two in north basin (Zhao et al., 2012). Mean water residence time is 4.4 years for the entire lake, 1.4 years for the South basin, and 3.6 years for the North basin (Zhang and Yerubandi, 2012; Zhao et al., 2012). Lake Winnipeg drains into Hudson Bay through the Nelson River outflow in the North basin (Fig. 1). Long-term increases in nutrient loadings to the lake stemmed from the Red River and Winnipeg River, albeit with considerable temporal variability. A 58% increase in the total dissolved nitrogen (TDN) load from the Red River was reported for the period 1978–1999

0380-1330/$ – see front matter © 2012 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jglr.2012.02.004

136

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Fig. 1. Map of Lake Winnipeg, Manitoba, Canada, its bathymetry, riverine inputs and outflow, and names of official long-term water quality sampling stations.

(Jones and Armstrong, 2001; McCullough et al., 2012). TDN concentrations in the Red River from 1992 to 2007 were with on average 2.5 mg/L, markedly higher than those of the Winnipeg and Saskatchewan Rivers with 0.52 and 0.49 mg/L, respectively, with no discernible temporal trend. The TDN loading to Lake Winnipeg from all sources averages 90,701 t per year from 1994 to 2007, with 34% coming from the Red River, 25% from the Winnipeg River, and 10% from the Saskatchewan River. The remaining smaller and/or unmeasured rivers plus atmospheric N deposition are estimated to contribute ~ 30% of the TDN loading to Lake Winnipeg (Environment Canada, 2011). The annual and seasonal riverine nutrient loading to Lake Winnipeg is strongly dependent on the amount of discharge and nutrient levels in each of the contributing watersheds (McCullough, 2001; McCullough et al., 2012; Schindler et al., 2012). In the Red River, for example, the month of April contributes twice the N loading of all other months, owing to higher discharge from snowmelt and spring runoff in the Red River watershed. Other N sources to Lake Winnipeg are less well characterized, but could include nitrate delivered via smaller rivers, diffusive groundwater inputs, and decay and subsequent mineralization of N2 fixing cyanobacteria (Environment Canada, 2011; Kling et al., 2011). Monitoring of TDN within Lake Winnipeg by Manitoba Water Stewardship (MWS) revealed distinctive differences between the North and the South basins, averaging 0.65 mg TDN/L and 0.87 mg TDN/L, respectively, from 1999–2007 (Environment Canada, 2011). Similarly, average surface water nitrate (NO3−–N) concentrations in the South basin are correspondingly higher than those in the North basin (0.16 mg/L vs. 0.11 mg/L NO3−–N). Not surprisingly, highest concentrations of TDN are observed in the South basin near the mouth of the Red River. Seasonally, the TDN concentrations in Lake

Winnipeg increase towards the summer and fall, respectively, in part as a result of biological N2 fixation by cyanobacteria in the summer months (Environment Canada, 2011). Identifying the sources of dissolved NO3−–N and elucidating its biogeochemical cycling in riverine and lotic systems is challenging, but is crucial to gaining a better understanding of the sources and role of N in nutrient cycling in Lake Winnipeg. Stable isotope techniques are a promising tool for assessing sources of dissolved NO3−– N in surface water bodies, especially if the dual-isotope (δ 15NNO3 and δ 18ONO3) approach is used to characterize key sources and assess nutrient transformation processes (Aravena and Mayer, 2010; Kendall, 1998; Kendall et al., 2007; Mayer, 2005). Nitrate from animal manure and human sewage is characteristically typified by high δ 15NNO3 values ranging from +7‰ to +20‰ (Aravena et al., 1993; Aravena and Robertson, 1998; Kreitler, 1979; Wassenaar, 1995). Animal and human waste derived N generally has much higher δ 15NNO3 values than synthetic agricultural fertilizers (~0‰), nitrate generated via nitrification processes in soils, or from atmospheric nitrate deposition (−11 to +8‰)(Kendall, 1998; Kendall et al., 2007). Unfortunately, nitrate derived from ammonium fertilizers, soils, and atmospheric deposition cannot be clearly differentiated on the basis of δ15N alone because of their overlapping isotopic values (Fig. 2). Nitrate from atmospheric deposition, however, has very positive δ 18ONO3 values ranging from >+ 50 to + 94‰ (Durka et al., 1994; Kendall, 1998; Kendall et al., 2007; Voerkelius, 1990). Nitratecontaining synthetic fertilizers have δ 18ONO3 values around +22 ± 3‰ (Amberger, 1987; Voerkelius, 1990; Wassenaar, 1995). Nitrate derived from nitrification, for instance, in soils or resulting from mineralization of aquatic phytoplankton and N2 fixing cyanobacteria followed by ammonification and nitrification has δ 18ONO3 values between − 15 and + 15‰ (Durka et al., 1994; Hollocher, 1984; Mayer et al., 2001; Rock and Mayer, 2004) that are dependent on the δ 18O of water. Nitrates from manure and sewage have also δ 18ONO3 values between −15 and +15‰ (Aravena et al., 1993; Wassenaar, 1995). Therefore, the combined use of δ 15 N and δ 18O values of NO3− can provide a diagnostic tool for discerning among four major nitrate sources: (1) atmospheric nitrate deposition, (2) nitrate based fertilizers, (3) nitrate derived from nitrification in soils and aquatic systems (including decay of in-lake organic matter), and (4) nitrate from manure and sewage (Fig. 2).

Fig. 2. Typical range of δ15N and δ18O values of nitrate derived from natural and anthropogenic N sources (adapted from Kendall et al., 2007).

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Isotope source tracking of low levels of dissolved NO3−–N in riverine and lake systems can be complicated by the fact that numerous biogeochemical transformations in soils and in the aquatic nitrogen cycle are associated with significant and variable isotope effects. During ammonium volatilization (e.g. from sewage treatment plants, soils), the conversion of NH4+ to NH3, 14N is preferentially converted to NH3 leaving the remaining NH4+ enriched in 15N (Letolle, 1980; Wellman et al., 1968). Nitrification, the conversion of NH4+ to NO3−, can also proceed with significant nitrogen isotope fractionation, accumulating 14N preferentially in the produced NO3− provided that the NH4+ substrate is not limiting (Mariotti et al., 1981). During nitrification, three new oxygen atoms are introduced into the newly formed NO3− molecule. Two of these oxygen atoms are typically derived from ambient H2O and one from atmospheric O2 (Hollocher, 1984), resulting in δ 18ONO3 values between b−5‰ to + 15‰ depending on the δ 18O values of water and O2 gas (Durka et al., 1994; Mayer et al., 2001; Wassenaar and Hendry, 2007) and potential oxygen exchange (Snider et al., 2010). Another process that can cause significant alteration of the isotopic composition of NO3− in aquatic systems is microbial denitrification under anaerobic conditions, whereby 14N and 16O are preferentially metabolized by microorganisms, causing isotopic enrichment of 15N and 18O in the remaining NO3− as its concentration decreases (Mariotti et al., 1981, 1982). The increase in δ 15N values due to anaerobic microbial denitrification is about twice that of δ 18O (e.g. Böttcher et al., 1990). Hence, the residual unconsumed nitrate progressively assumes elevated δ 15N and δ 18O values along a 2:1 line in a δ 15NNO3 versus δ 18ONO3 plot under closed system conditions (Fig. 2). In aquatic systems, assimilation of NO3− by phytoplankton has also been found to fractionate nitrogen and oxygen isotope ratios (Granger et al., 2004, 2010). Here nitrate reduction to nitrite enriches the cell-internal nitrate in 15N and 18O, followed by nitrate re-flux from the cells back to the aquatic system. The observed increase in δ 15N and δ 18O values of NO3− was found to be similar regardless of phytoplankton species (Granger et al., 2004), yielding a 1:1 slope on a δ 15NNO3 versus δ 18ONO3 plot (Fig. 2). By contrast, cyanobacterial N2 fixation of atmospheric nitrogen is not accompanied by significant nitrogen isotope fractionation effects, resulting in cyanobacterial δ 15N values of around 0 ± 2‰ (Gu et al., 2006; MacGregor et al., 2001). Subsequent mineralization and nitrification of cyanobacteria yields nitrate with δ 15N values equal to or less than 0‰, dependent on the extent of N isotope fractionation during this process (Hadas et al., 2009; Patoine et al., 2006) with δ 18O values of the newly formed nitrate of b+15‰. The objective of this study was to use δ 15N and δ 18O values to investigate temporal and spatial patterns of the isotopic composition of NO3− in Lake Winnipeg and its contributing rivers in order to gain insights into N nutrient sources and in-lake N dynamics. Our objectives were attained by 1) spatio-temporal chemical and isotopic (δ 15N and δ 18O) analyses of NO3− in Lake Winnipeg from surface and bottom waters at selected stations in the North and South basins in 2007, 2) chemical and isotopic analyses of NO3− in several key contributing rivers from 2008 to 2009, and 3) chemical and isotopic analyses of NO3− from several sewage treatment plant (STP) effluents that discharge into the rivers or lake from 2008 to 2009. Due to logistical considerations we were unable to conduct our in-lake, watershed and STP sampling within the same year. However, our intention was to assess whether nitrate isotopic analyses had the potential to reveal insights into current nitrate provenance and in-lake N processing over a relatively short timeframe of a few years.

137

open water stations (Fig. 3) were visited to obtain representative spatial coverage of the lake, with seasonal subsets (winter, spring, summer) selected for isotopic assays. A total of 110 coupled in-lake NO3− concentration and stable isotopic assays were obtained. The research vessel PV Namao of the Lake Winnipeg Research Consortium (www.lakewinnipegresearch.org) was used to access open-water stations on lake-wide scientific cruises. In 2007, the scientific cruises on Lake Winnipeg were done over a period of 3 weeks in spring following ice-off (May–June), summer (July–August) and fall (September– October). To obtain depth specific samples, a 2-stage 12 V submersible pump (2 L/min flow) with clear Tygon™ tubing (5/8” hose) was run to the deck of the Namao. Targeted sampling depths were 1 m below surface (e.g. within photic zone) and ~ 0.2 to 0.5 m above the lake bottom. For the wintertime samples, a small subset of lake stations was visited by helicopter. Holes were drilled through the ice and a Niskin sampler was manually lowered and triggered to obtain 1 m and bottom water 2 L samples. All lake water samples for NO3− isotope analyses were filtered (120 ml) through a 0.45 μm cellulose acetate filter (Millipore) and immediately frozen in order to prevent biodegradation. A second filtered aliquot (120 ml) was stored at 4 °C for the determination of NO3− concentration at the National Laboratory for Environmental Testing (NLET, Saskatoon) at Environment Canada within 1–2 weeks of collection. Watershed riverine samples were collected at 8 different sites from 2008 to 2009 to characterize the NO3− concentrations and isotopic composition of nitrate inputs to Lake Winnipeg (Fig. 3). All watershed stations used were established federal or provincial monitoring stations with sampling occurring approximately monthly at the major riverine inflows to Lake Winnipeg. A total of 52 coupled isotope and concentration riverine data points were obtained for samples with NO3−–N concentrations >0.1 mg/L. On the Red River, water samples were taken at Selkirk, MB (n = 17) (downstream of Winnipeg and the closest station to the mouth of the Red River) and at the Floodway (Red River upstream of the City of Winnipeg, n = 3). The

Materials and methods Water samples for NO3− concentrations and stable nitrogen and oxygen isotope ratio assays were collected from Lake Winnipeg and its tributaries between 2007 and 2009. In Lake Winnipeg up to 46

Fig. 3. Sampling locations for stations in Lake Winnipeg in 2007–2008 (circles), and sampling locations for riverine (triangles) and STP (squares) sites from 2008 to 2009.

138

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Winnipeg River was sampled at Pine Falls, MB (n = 12). The Saskatchewan River was sampled at Grand Rapids, MB, from outflow of the hydroelectric dam that discharges reservoir water directly into Lake Winnipeg (n = 9). Several other Manitoba rivers upstream in the contributing watersheds were sampled less frequently, including the Assiniboine River near Headingly, MB (n = 14), the Souris River near Treesbank, MB (n = 9), and the Dauphin (n = 4) and Icelandic rivers (n = 2) at their mouths (Fig. 3). Riverine samples were collected midstream from bridges using clean 5 L buckets that were rinsed three times with sample. River water samples for NO3− concentration and isotope analyses were filtered, preserved and stored as described previously for lake water samples. Sewage treatment plant (STP) effluents in the watershed were sampled several times in 2008–2009 in order to chemically and isotopically characterize NO3− in the effluent discharge. Sewage treatment plants included the City of Winnipeg (North, South and West STPs) and two smaller STPs adjacent to the lake at Gimli, MB, and on Hecla Island (Fig. 3). Two one-liter bottles of fresh effluent were provided by STP plant staff. STP effluent samples were filtered, preserved and stored as described above. For all lake, river and STP samples, a 2 ml aliquot was taken for δ18OH2O determination of the water. Oxygen isotope ratios of water were determined using laser spectroscopy on a Liquid Water Isotope Analyzer (DLT-100, Los Gatos Research Inc.) on filtered 2 ml water samples using the method described by Lis et al. (2008) at the Isotope Hydrology and Ecology Laboratory of Environment Canada in Saskatoon, Canada. Nitrate concentrations were determined on 0.45 μm field-filtered samples by automated colorimetric EPA standard method SM 4500 (ASTM, 1992) NO3-F with a detection limit of 0.010 mg/L as NO3−–N. Field-frozen NO3− isotope samples were analyzed at the Isotope Science Laboratory at the University of Calgary (Calgary, Alberta, Canada). Nitrogen and oxygen isotope ratios of NO3− were determined using the denitrifier method (Casciotti et al., 2002; Sigman et al., 2001). Samples were thawed and depending on the NO3− concentration, between b0.1 and 10 mL of sample was used in order to obtain targeted N2O yields for isotope ratio mass spectrometry (IRMS). Denitrifying bacteria were injected into each aqueous sample and dissolved NO3− was completely converted to nitrous oxide (N2O) gas. The 15N/ 14N and 18O/ 16O ratios were measured on N2O gas using a ThermoFinnigan MAT Delta Plus XL IRMS. The δ 15N–NO3− values are reported in the internationally accepted delta notation in per mil (‰) relative to the AIR reference, with an analytical precision of ±0.3‰. The δ 18O–NO3− values are referenced to VSMOW (Vienna Standard Mean Ocean Water) with an uncertainty of ±0.5‰. Several international reference materials were used to ensure precise and accurate results: USGS 34 (δ 15 N = − 1.8 ± 0.2‰; δ 18O = −27.9 ± 0.6‰), USGS 35 (δ 15 N = +2.7 ± 0.2‰; δ 18O = +57.5 ± 0.6‰) and IAEA NO3 (δ 15 N = +4.7 ± 0.2‰; δ 18O = +25.6 ± 0.4‰). Results and discussion Nitrate in contributing rivers Over the course of this study, riverine dissolved NO3−–N concentrations were determined for the main and some lesser tributaries of Lake Winnipeg, with mean values ± standard deviations reported in Table 1. Of the major rivers entering the lake, the Winnipeg River at Pine Falls had the lowest average non-flow weighted dissolved NO3−–N concentration of 0.12 ± 0.21 mg/L (n = 11). The Red River had a mean nitrate-N concentration of 0.94 ± 0.82 mg/L (n = 13) at Selkirk, and 0.83 ± 0.38 mg/L NO3−–N (n = 3) at the Floodway upstream of the City of Winnipeg and the confluence with the Assiniboine River (n = 3). The Saskatchewan River was characterized by a mean concentration of 0.75 ± 1.45 mg/L NO3−–N (n = 9). The Souris and Assiniboine tributaries of the Red River west of Winnipeg had

Table 1 Mean NO3−–N concentrations and δ15N and δ18O values of nitrate (± standard deviation, number of samples) in rivers and tributaries flowing into Lake Winnipeg sampled between May 2008 and October 2009. Station

Location

Saskatchewan River Red River

Grand Rapids Selkirk

Icelandic River

At mouth

Average Average NO3− N δ15Nnitrate [‰] [mg/L]

0.75 (9) 0.94 (13) Red River Floodway 0.83 (3) Winnipeg River Pine Falls 0.12 (11) Assiniboine Headingly 0.80 River (14) Souris River Treesbank 0.36 (9) Dauphin River At mouth 0.48 2.44

+ 5.0 ± 3.4 (5) + 8.1 ± 1.4 (12) + 8.9 ± 1.1 (3) − 0.6 ± 3.8 (10) + 10.9 ± 6.1 (10) + 13.9 ± 4.8 (6) + 2.3 ± 10.6 (4) + 10.4 ± 2.4 (2)

Average δ18Onitrate [‰]

Average δ18Owater [‰]

+ 5.6 ± 10.1 (5) + 3.2 ± 5.3 (12) + 5.4 ± 5.7 (3) + 2.9 ± 9.1 (10) + 5.5 ± 11.2 (10) + 6.0 ± 6.7 (6) + 14.7 ± 13.9 (4) + 0.3 ± 0.2 (2)

− 14.4 ± 0.7 (4) − 10.4 ± 2.2 (13) − 12.2 ± 5.0 (2) − 8.6 ± 0.4 (9) − 13.1 ± 1.2 (11) − 7.8 ± 2.4 (5) − 8.6 ± 0.2 (2) − 14.1 (1)

mean concentrations of 0.36 ± 0.46 mg/L NO3−–N (n = 9) and 0.80 ± 0.77 mg/L NO3−–N (n = 14), respectively. The Dauphin River was characterized by a mean NO3−–N concentration of 0.48 ± 0.82 mg/L (n = 4). The highest mean NO3−–N concentration was observed for the Icelandic River at 2.44 mg/L (n = 2). All of the reported nitrate concentrations were within the long-term ranges of NO3− values reported for the above-mentioned rivers from 1994 to 2007 (Environment Canada, 2011). Average monthly nitrate riverine loadings (data from this study) were determined for the Saskatchewan River, the Red River at Selkirk, and the Winnipeg River (Table 2). The Red River at Selkirk was characterized by the highest average monthly nitrate + nitrite-N loading of 887 t per month, constituting 41.5% of the total nitrogen load, with the remainder being predominantly particulate organic nitrogen (PON). The Winnipeg River at Pine Falls and Point du Bois had an intermediate nitrate-N load with 382 and 206 t per month respectively (17.1 and 14.4% of total N flux). The lowest average monthly nitrate-N loading was observed in the Saskatchewan River at 108 t per month, which constituted only 10.2% of the total N flux contributed by this river. Since riverine fluxes of nitrogen compounds were calculated as concentrations multiplied by discharge, there was significant variability in monthly nitrate fluxes owing to dependence on river flow. These loading estimates, however, were consistent with previous long-term loading estimates (Bourne et al., 2002) despite the shorter observation timeframe. The highest monthly nitrate fluxes from rivers generally occur in April during high river flow, as noted in previous studies (McCullough, 2001). Over the course of this study the Red River contributed the highest annual loading of NO3− to Lake Winnipeg, as noted in other summary reports (Environment Canada, 2012; McCullough et al., 2012). Average δ 15N and δ 18O values of riverine NO3− are summarized in Table 1 with a frequency distribution of δ 15NNO3 values for main rivers plotted in a range-histogram in Fig. 4. The Winnipeg River at Pine Falls had the lowest mean NO3−–N concentration accompanied Table 2 Mean monthly total N and nitrate/nitrite-N fluxes in the Saskatchewan River, Red River and Winnipeg River from January 2007 to December 2008. River

Location

Total N flux [t/month]

Nitrate + nitrite-N [t/month]

N

Saskatchewan River Red River Winnipeg River Winnipeg River

Grand Rapids Selkirk Pine Falls Point du Bois

1062 2135 2228 1434

108 887 382 206

24 24 24 24

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Fig. 4. Frequency histogram and range of δ15N values of NO3− entering Lake Winnipeg from the Red, Winnipeg and Saskachewan Rivers in 2008–2009.

139

oxygen exchange and associated isotope fractionation (Snider et al., 2010) may also affect the oxygen isotope ratios of nitrate formed by nitrification. In the Dauphin River, an elevated δ 18ONO3 value of ~+26‰ was observed in connection with an anomalously low NO3− concentration event of b0.13 mg/L NO3−–N, suggesting some possible contribution from atmospheric NO3− deposition. In contrast, the δ 18O value of NO3− in the Dauphin River was very low (− 3.0‰) at a very high NO3− concentration of 1.71 mg/L NO3−–N, suggesting negligible contribution of atmospheric NO3− at elevated riverine nitrate levels. In general, low mean concentrations of riverine NO3−–N below ~0.2 mg/L were associated with the lowest δ 15NNO3 values, as exemplified by the Winnipeg River at Pine Falls (Table 1). In contrast, elevated NO3− concentrations in all Lake Winnipeg tributaries between 0.36 and 2.44 mg/L NO3−–N were associated with the highest δ 15N NO3 values between + 5.0 and +13.9‰, while the corresponding δ 18ONO3 values were b+15.0‰. Moreover, each of the main riverine inputs into Lake Winnipeg had distinctive mean δ 15N values; +8.1‰ for the Red River at Selkirk, −0.6‰ for the Winnipeg River at Pine Falls, and +5.0‰ for the Saskatchewan River at Grand Rapids (Table 1, Fig. 4). This provided a suitable isotopic label for differentiating nitrate from the three major riverine inputs to Lake Winnipeg, and enabled us to test whether the fate of these isotopically distinctive riverine nitrate inputs can be traced in the dynamic N cycle of Lake Winnipeg. Nitrate in waste water treatment effluent

by the lowest mean δ 15N value of −0.6 ± 3.8‰ (n = 10). The mean δ 15N value of riverine NO3− in the Saskatchewan River was +5.0 ± 3.4‰ (n = 5). The mean δ 15N values of NO3− in the Red, Assiniboine, Souris and Icelandic Rivers were higher, with δ 15N values that ranged between +8.1 ± 1.4‰ and +13.9 ± 4.8‰. The Dauphin River had extremely variable δ 15N values with b−5.0‰ at low NO3− concentrations (b0.15 mg/L NO3−–N) and a high δ 15N value of + 14.7‰ at its peak concentration of 1.71 mg/L NO3−–N. This resulted in a nonrepresentative mean δ 15N value of +2.3‰ with a very high standard deviation of ±10.6‰ (n = 4; Table 1). Rivers that drained agricultural landscapes (e.g. Red, Assiniboine, and Icelandic Rivers) had the highest nitrate-N concentrations and δ 15N values. With the exception of the Dauphin River, average δ 18O values of NO3− in all other tributaries ranged between + 0.3 ± 0.2 and +6.0 ± 6.7‰. This is a typical range of δ 18O values for NO3− derived from microbial nitrification in soils, manure, or sewage treatment systems (Kendall et al., 2007). The δ 18O values of NO3− produced by nitrification can be calculated following the equation:     18 18 18 − δ O−NO3 ½‰ ¼ 2=3 δ O−H2 O þ εH20 þ 1=3 δ O−O2 þ ε02 where εH20 and ε02 are the isotopic enrichment factors during the addition of oxygen from H2O and from the atmospheric O2, respectively. Assuming no oxygen isotope fractionation during the addition of oxygen (Amberger and Schmidt, 1987; Böttcher et al., 1990; Durka et al., 1994; Mayer et al., 2001), the equation simplifies to: 18

−

18

18

δ O−NO3 ½‰ ¼ 2=3  δ O−H2 O þ 1=3  δ O−O2 Atmospheric O2 has a δ 18OO2 value of +23.5‰ (Kroopnick and Craig, 1972), also similar to dissolved O2 in Lake Winnipeg which averages +23.3‰ (Wassenaar, 2012). The average δ 18O-H2O values of river water (Table 1) varied from − 14.4‰ (Saskatchewan River) to −7.8‰ (Souris River). Thus, the δ 18O values for NO3− produced by nitrification of N in manure and urea or ammonium fertilizers in the watershed were expected to be in the range of −2 to +4‰. This range was generally in good agreement with the observed range of δ 18O values of riverine NO3− (Table 1), especially since

The average NO3−–N concentrations and amount-weighted mean δ 15N and δ 18O values of effluents from five STPs are summarized in Table 3 and Fig. 5. The amount-weighted mean average δ 15NNO3 value of nitrate in STP effluents from the City of Winnipeg combined was +4.5‰. Nitrate concentrations in the effluents of the STPs from the City of Winnipeg varied from b1.0 mg/L to occasionally over 10 mg/L NO3−–N. The City of Winnipeg STP with the highest average NO3−–N concentration of 7.8 mg/L in its effluent (STP North) had the lowest concentration-weighted average δ 15NNO3 value of +2.6 ± 5.7‰, whereas the STP with the lowest average NO3−–N concentration in its effluent of 1.3 mg/L in Winnipeg (STP West) was characterized by the highest average δ 15N value of +13.9 ± 6.1‰. Effluents from the Winnipeg South STP ranged widely between these extremes, with an average NO3−–N concentration of 4.0 mg/L and an average δ 15N value of +5.4 ± 2.2‰ (Table 3). Increased biological nitrogen reduction (BNR) efficiencies in city waste water treatment plants are typically achieved via repeated nitrification and denitrification cycles, generally resulting in increasing δ 15NNO3 values as the NO3− concentrations in the effluents decrease (Voss et al., 2005, 2006). Only one of the three STPs in Winnipeg has implemented full biological nutrient removal. Effluents from the STPs in the towns of Gimli and Hecla, non-BNR operations north of Winnipeg, were characterized by elevated amount-weighted average δ 15N values of +14.2 ± 6.7 and + 15.1 ± 5.5‰, respectively, while average NO3−–N concentrations were 3.3 and 2.0 mg/L (Table 3). Amount-weighted mean δ 18O values of NO3− in all measured STP effluents ranged from − 4.6‰ (Hecla) to + 9.6‰ (Winnipeg South). This wide range of oxygen Table 3 Mean NO3−–N concentrations and amount-weighted average δ15N and δ18O values for nitrate in the effluent of STPs around Lake Winnipeg. Location

Average NO3− N [mg/L]

Average δ15N [‰]

Average δ18O [‰]

N

Winnipeg South Winnipeg West Winnipeg North Gimli Hecla

4.0 1.3 7.8 3.3 2.0

+ 5.4 + 13.9 + 2.6 + 14.2 + 15.1

+ 9.6 + 3.4 + 8.0 + 1.1 − 4.6

6 6 6 5 6

140

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Fig. 5. Crossplot of δ15N and δ18O values of riverine nitrate and nitrate from STPs (grey circles) in the Lake Winnipeg watershed compared to the isotopic compositions of nitrate from various natural and anthropgenic N sources.

isotope ratios is typical for NO3− generated via microbial nitrification and potentially affected by some treatment process denitrification (Kendall et al., 2007). Sources of nitrate in contributing rivers Although not the major focus of this study, it was found that the isotopic composition of riverine nitrate (Fig. 4) entering Lake Winnipeg was consistent with isotopic ranges typical of nitrate derived from animal manure and/or municipal waste waters (e.g. Red and Assiniboine Rivers), from ammonium based fertilizers or soil mineralization (e.g. Winnipeg River), or from a combination of these sources (e.g. Saskatchewan River) (Fig. 5). The observed ranges of nitrate concentrations and N and O isotopic compositions of the urban STPs, albeit variable, were also consistent with the elevated δ 15N values of NO3− observed in several rivers and tributaries discharging into Lake Winnipeg. It must, however, be noted that manure-derived NO3− from agricultural sources in the watersheds constitutes another major NO3− source that cannot be conclusively distinguished from STP derived nitrate due to similar N and O isotope ratios (Fig. 5). Thus, nitrate source apportionment between manure and STPs for the Red River remains equivocal based on isotopic characterizations alone, and hence must rely on conventional monitoring nitrate mass balance data. For example, in a study of nutrient source quantification in the Red River basin (Yates et al., 2012), the largest potential source of N leaching to streams was believed to be associated with synthetic fertilizer and manure applications, while urban derived sewage and lagoon inputs played a minor role. Other watershed N mass balance studies have shown that most of the nitrate in the Red River comes from agricultural sources in the United States and Canada, with the City of Winnipeg contributing only about 5–11% to the total nitrate load from its STPs (Bourne et al., 2002). Bunting et al. (2011) concluded that human wastes are a relatively small part of the riverine nutrient input to Lake Winnipeg, while decadal-scale increases in livestock and the associated applications of fertilizers and manure constitute the main sources of nutrients draining into Lake Winnipeg. The isotopic data presented here are consistent with mass balance studies that suggest nitrate at the mouth of the Red River watershed was indicative of predominantly upstream manure and waste N sources owing to the similarity in nitrate concentrations and isotopic compositions with sites upstream of the City of Winnipeg (Floodway) and those draining agricultural landscapes (e.g. Assiniboine River).

The Winnipeg River, on the other hand, had low δ 15N values suggesting that the main nitrate sources in this boreal watershed (e.g. outflow of Lake of the Woods) were from natural soil N and possibly also synthetic ammonium and urea-based fertilizers (Fig. 5). The Saskatchewan River had intermediate δ 15N values compared to the two other major riverine nitrate inputs to Lake Winnipeg, indicating a mixture of various nitrate sources potentially modified by N transformation processes. The Saskatchewan River is regulated by numerous dams and impoundments between its headwaters in Alberta and Lake Winnipeg. It was beyond the scope of this study to assess how upstream agricultural and urban nitrogen sources, N sequestration and N mineralization affect the concentrations and isotopic compositions of nitrate in the Saskatchewan River basin. However, Rock and Mayer (2006) showed that manure-derived NO3− with δ 15N values between +12 and +15‰ contributed markedly to the riverine NO3− load in the Oldman River in Alberta, while wastewater effluents from the City of Calgary with a δ15NNO3 value around +9‰ were a dominant nitrate source in the Bow River, both tributaries to the South Saskatchewan River (Mayer et al., in press). Dissolved nitrate in Lake Winnipeg Dissolved NO3−–N concentrations in Lake Winnipeg were determined seasonally at selected stations for surface and bottom water, with spatial concentration distribution results summarized in Fig. 6. On an annual depth-integrated basis, mean NO3−–N concentrations were slightly higher and less variable in the South basin (0.19 ± 0.12 mg/L) compared to the North basin (0.14 ± 0.34 mg/L) although this difference was not significant based on a t-test (p = 0.137). Seasonally, across the entire lake during open water season, depthintegrated mean NO3−–N concentrations were generally similar with 0.11 ± 0.01 mg/L (n = 60) in the fall and 0.12 ± 0.15 mg/L (n = 67) and 0.12 ± 0.12 mg/L (n = 89) in spring and summer, while highest mean NO3−–N concentrations were observed in winter under ice (1.0 ± 0.4 mg/L, n = 11). Spatially, the highest proportion of “below detection” nitrate samples were found in surface waters of the north eastern part of the North basin in summer and early fall (Fig. 6) when nitrate uptake by phytoplankton may have consumed the available dissolved nitrate. The highest NO3−–N concentrations (>1–2 mg/L, n = 7) were found along the eastern side of the North basin in wintertime under ice. With a few exceptions, no significant differences were found in the mean NO3−–N concentrations between surface and bottom water (Fig. 6) (t-test; p = 0.312), similar to trends observed for dissolved oxygen (Wassenaar, 2012) due to the well mixed nature of Lake Winnipeg (Zhao et al., 2012). Considered spatially, during the open water season in spring, summer and fall, NO3−–N concentrations were generally lower in the North basin compared to the South basin, typically not exceeding 0.2 mg/L in surface or bottom water samples and often b0.1 mg/L especially in the summer and fall (Fig. 6). Across all open water seasons, there was a distinctive nitrate “plume” in the South basin, which clearly originated from the Red River. The Red River is the major riverine nitrate input to Lake Winnipeg with highest nitrate fluxes in April. Around the mouth of the Red River, NO3−–N concentrations between 0.2 and 0.4 mg/L were often observed, with station W10 (Fig. 1) registering even higher values for bottom water in spring 2007 (0.5–1.0 mg/L). These spatial patterns were consistent with the nitrate plume from the Red River entering the South basin. Subsequent mixing with other nitrate sources (e.g. from the Winnipeg River), uptake by phytoplankton and bacteria (Kling et al., 2011), dilution and dispersal towards the north restricted the elevated nitrate-N concentrations to the south-western portion of the South basin during the open water season. The spatial distribution of δ 15N values of lake water NO3− is summarized in Fig. 7. Overall, across all seasons the North basin had a significantly lower mean δ 15NNO3 value (+ 3.3 ± 3.2‰, n = 105)

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

141

Fig. 6. Interpolated (kriged) map of surface (1 m depth — upper panel) and bottom water (50 cm from lake bottom — lower panel) nitrate–N (mg/L NO3−–N) concentrations in 2007. Black dots are the sampling sites. White areas indicate areas that were not visited (winter) or had “below detection” or insufficient spatial data for interpolation (fall).

compared to the South basin (+6.2 ± 3.0‰, n = 105, t-test; p = b0.001). In wintertime under ice, surface water samples at 6 stations around the lake had δ 15NNO3 values that varied between +2.5 and +9.7‰ with the highest values found in the central parts of the

lake in both basins. Bottom water samples collected in winter had slightly elevated δ 15NNO3 values, ranging between + 6.7 and + 13.1‰. During the spring, elevated δ 15NNO3 values between + 6.0 and + 12.1‰ were observed for surface and bottom water samples in

142

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Fig. 7. Interpolated (kriged) map of the δ15N values of dissolved NO3− in surface (top – 1 m depth) and bottom waters (~ 50 cm from bottom) in Lake Winnipeg in 2007. White areas indicate regions that were not visited (winter) or had insufficient nitrate concentrations to allow for stable isotopic analysis (e.g. summer surface).

the South basin and in the North basin near the inflow of the Saskatchewan River. These high in-lake δ 15NNO3 values were consistent with observed high riverine nitrate inputs with elevated δ 15N values delivered via snow melt and spring runoff entering the South Basin from the Red River and to a lesser degree the North Basin from the Saskatchewan River. In the rest of the North basin δ 15N values of NO3− were generally below + 4.1‰ during spring. In summertime, δ 15N values in both surface and bottom waters of the North basin ranged between − 2.0‰ for samples with the lowest

NO3− concentrations to values up to +6.5‰, although nitrate was below detection levels for a number of North basin stations in the summer. The highest δ 15N values in the North basin were associated with the highest NO3− concentrations of up to 0.4 mg/L NO3−–N near the mouth of the Saskatchewan River and hence appeared to be caused by this riverine nitrate input. The low δ 15N values of nitrate associated with low nitrate concentrations in the North basin revealed a different nitrate source. Fixation of atmospheric N2 results in cyanobacterial δ 15N values of around 0 ± 2‰ (Gu et al., 2006;

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

MacGregor et al., 2001). Subsequent decay of cyanobacteria followed by nitrification yields nitrate with δ 15N values equal or less than 0‰ (Hadas et al., 2009; Patoine et al., 2006). These processes appeared to be responsible for the low δ 15N values of nitrate in the summer months, especially in the North basin. N2 fixing cyanobacteria are a major cause for the increased phytoplankton abundance in Lake Winnipeg in recent years (Schindler et al., 2012). In the South basin, a marked change from elevated δ 15N values of nitrate in spring (mean δ 15N of +8.4‰) to lower nitrogen isotope ratios in summer (mean δ 15N of +4.0‰) was also observed. There may be several reasons for this rapid change in δ 15N values of nitrate, including recycling of N from N2 fixing cyanobacteria. The rapid change in δ 15NNO3 values revealed a short residence time of nitrate in the South basin. This could be due to the short flushing rate of water in the South basin (months) and eventual mixing of the spring nitrate plume with high δ 15N values from the Red River with the Winnipeg River nitrate input having low δ 15N values. Alternately, rapid nitrate assimilation and uptake could also cause short residence times of nitrate in the South basin accompanied by an enrichment of 15N in the remaining nitrate (Granger et al., 2004, 2010), thereby explaining some of the higher δ 15N values observed during spring sampling. The spatial distribution of δ 18ONO3 values is summarized in Fig. 8. Overall, across all seasons, the North basin had a lower mean δ 18ONO3 value (− 0.3 ± 4.7‰, n = 105) compared to the South basin (+3.0 ± 7.1‰, n = 105). In wintertime, surface and bottom water samples under ice at 6 stations had low δ 18ONO3 values that varied between −5.5 and + 2.5‰. During the open water season in summer, similar ranges of δ 18ONO3 with values from −5.2 to +5.8‰ were observed in surface and bottom waters. In the spring, elevated δ 18O values of NO3− were observed ranging between +6.9 and + 21.8‰ for both surface and bottom water samples across the entire South basin, and in the North basin near the inflow of the South Saskatchewan River. In the eastern part of the North basin δ 18ONO3 ranged generally between −3.3 and +4.1‰ during spring. The elevated δ 18ONO3 values of up to +21.8‰ were associated with elevated δ 15N values (Fig. 7) and high nitrate concentrations. This unique isotope composition of nitrate required either that the 2007 spring runoff in the Red River and Saskatchewan River delivered nitrate with unusually high δ 18O values, or alternately, a process enriching both 15N and 18O in the remaining nitrate could be responsible for the observed trends. Fig. 9 compares the isotopic composition of lake water NO3− samples to the average isotopic compositions of NO3− in the three major nitrate contributing rivers. In summer, the lake NO3− average δ 15N values of +3.2 ± 2.3‰ and δ 18O values of −2.8 ± 2.2‰ were slightly lower than those of NO3− in most of the contributing rivers (Table 1, exceptions were the Winnipeg and Dauphin Rivers). This isotopic composition of in-lake nitrate suggests that there is another source of NO3− to Lake Winnipeg having low δ 15N and low δ 18O values, especially later in the season. External nitrate from atmospheric deposition and from fertilizer NO3− can be ruled out because of the high δ 18O values typical of these nitrate sources. The decay of accumulated phytoplankton, however, from spring and early summer algal blooms (see POM data in Fig. 10) and subsequent nitrification seems the most likely pathway for N recycling and lake-internal generation of NO3−. Specifically, the decay and nitrification of N2 fixing cyanobacteria would cause low δ 15N and δ 18O values in the newly formed nitrate, as was observed in some parts of Lake Winnipeg especially in summer and fall. In springtime, lake water NO3− had elevated average δ 15N (+ 6.5 ± 3.4‰) and δ 18O values (+6.7 ± 6.2‰) compared to those observed in summer. The mean isotopic composition of lake water NO3− in spring was similar to that of NO3− from most riverine inputs (Table 1). This suggests that the riverine NO3− fluxes (especially from the Red River), which peak during spring runoff, were the major source of lake nitrate in the South Basin in spring. However, a number of lake water samples in spring in the South basin displayed highly elevated

143

δ 15N and δ 18O (Fig. 9) reaching values higher than mean NO3− isotopic compositions that were measured in riverine inputs (Table 1). This increase in δ 15N and δ 18O values appeared to follow slopes between 0.6:1 and 1:1 (Fig. 9), which are typical for assimilation of NO3− under closed system conditions (Granger et al., 2010). An alternate explanation could be that the spring pattern was driven by a short-term flush of Red River snow melt water high in nitrate concentrations (but not captured in our infrequent riverine sampling) having elevated δ 15N and δ 18O values, given that large algal blooms tend not to occur frequently in the South basin in the spring. Indeed, the apparent linear trend of δ 15N and δ 18O values for spring samples shown in Fig. 9 can also be interpreted as a simple mixing line between a plume of 15N and 18O enriched nitrate, since all of the most positive δ 15N and δ 18O values follow a mixing gyre from the Red River through the South basin. The few samples collected in winter had average δ 15N values of + 6.9 ± 3.2‰, which were similar to the nitrogen isotope ratios of riverine NO3− inputs. The average δ 18O value of winter NO3− was − 0.5 ± 2.5‰, slightly lower than that of most riverine NO3− inputs. Enhanced ammonification and nitrification of accumulated algal remains from the previous seasons could be responsible for the low δ 18O values of lake water NO3− in winter. While two samples displayed increased δ 15N and δ 18O values that plotted close to a 2:1 line typical for denitrification (Fig. 9), this process seems unlikely given that Lake Winnipeg is polymictic and highly aerobic throughout the entire year.

Dissolved nitrate affects N flow in aquatic food webs of Lake Winnipeg The utilization of dissolved nitrate by autotrophic algae in lakes is accompanied by an average negative nitrogen isotopic discrimination of about −3‰ (range of −1.0 to −4.7‰) due to the preferential uptake of 14N by phytoplankton (Lehmann et al., 2004; Teranes and Bernasconi, 2000) or by cell-internal reduction to nitrite by nitrate reductase (Granger et al., 2004; Needoba et al., 2004; Shearer et al., 1991) in case that nitrate concentrations are not limiting. Nitrogen isotope effects are lower if nitrate concentrations are rate-limiting. Thus, if dissolved nitrate assimilation were the primary control on phytoplankton δ 15N values in Lake Winnipeg, then the δ 15N values of PON should be ~1 to 3‰ lower than the corresponding δ 15N values of nitrate. Fig. 10 shows a plot of mean δ 15N values of lake nitrate by basin (all seasons), compared to δ 15N values of particulate organic nitrogen (PON; Environment Canada, unpublished data) from corresponding stations. While the diagram confirmed the distinctive nitrogen isotope differences between the North and South basins described previously, it also indicated a mean apparent 15N enrichment factor between dissolved NO3− and PON of −3.3 and −2.1‰, respectively, strikingly similar to that expected for nitrogen isotope fractionation due to nitrate assimilation. A concurrent study of 15N/ 14N ratios in food webs of Lake Winnipeg (Hobson et al., 2012) found that upper trophic level organisms were highly correlated with the spatial patterns of δ 15N of nitrate with an appropriate trophic level isotopic enrichment, as demonstrated for emerald shiners in Fig. 10. This observation suggests that dissolved nitrate affects, at least in part, N flows in the aquatic food webs of Lake Winnipeg. The elevated δ 15N values of nitrate (and PON) in the South basin are indicative of landbased animal and human waste nutrient sources delivered at particularly high rates during spring runoff (e.g. Red River). Differences between the δ 15 N values of nitrate and PON between the South and North basins of Lake Winnipeg suggest that the uptake of landderived nitrate in the South basin appears to be rapid resulting in short mean residence times (b1 year) for dissolved nitrate. Therefore, other nitrate sources are required especially in the North basin. The lower δ 15N values for PON and dissolved nitrate in the North basin (Fig. 10) strongly suggest that nitrification of N2 fixing

144

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Fig. 8. Interpolated (kriged) map of the δ18O values of dissolved NO3− in surface and bottom water of Lake Winnipeg in 2007. White areas indicate regions that were not visited (winter) or had insufficient nitrate concentrations to allow for stable oxygen isotopic analysis (e.g. summer surface).

cyanobacteria constitutes an important additional source of lakeinternal nitrate. The relationships between δ 15N values of lake nitrate and PON in Fig. 10 also suggest that paleolimnological δ 15N analyses of PON in lake bottom cores should be a direct recorder of past nitrate sources to Lake Winnipeg in each basin (Leavitt et al., 2006). Indeed, a recent study (Bunting et al., in review) showed that the δ 15N values of “modern” (~year 2000) total organic N from cores in the South basin (δ 15N = + 7.5 to + 8.5‰) closely corresponded to the riverine δ 15N values of nitrate in the Red River reported in this study (δ15N ~ +8‰; Table 1) and to those of lake nitrate in the South basin (δ 15N = + 8 to 12‰). Bunting et al. (in review) further demonstrated that the

δ 15N values of PON in sedimentary cores from the South basin of Lake Winnipeg were characterized by a systematic increase from values of + 4‰ (e.g. pre-settlement) starting around 1900 to the current more positive δ 15N values of + 8‰ a century later, and attributed this to a systematic increase in N loading from animal and human waste derived nitrogen in the South basin, consistent with our findings. Conclusions In summary, elevated NO3−–N concentrations in rivers entering Lake Winnipeg between 0.36-2.44 mg/L were associated with highest

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

Fig. 9. Crossplot of seasonal nitrate δ15N and δ18O values of Lake Winnipeg in 2007 (circles) compared to mean isotopic compositions of major riverine nitrate inputs (Red, Winnipeg and Sasaktchewan Rivers) and other tributaries in the watershed from 2008 to 09. Trend lines are for those expected from closed-system dentrification (2:1) and phytoplankton nitrate assimilation (1:1).

δ 15N values from + 5.0 to +13.9‰, while δ 18ONO3 values were b+15.0‰. The δ 15N and δ 18O values of nitrate along the Red and Assiniboine Rivers were consistent with animal manure and/or municipal waste water sources. Nitrate in the Winnipeg River, had δ 15N and δ 18O values characteristic of soil mineralization or possibly ammonium or urea-based fertilizers. The rivers had distinctive mean δ 15NNO3 values of + 8.1‰ for the Red River, −0.6‰ for the Winnipeg River, and + 5.0‰ for the Saskatchewan River, thereby providing a useful isotopic label for tracing the fate of the riverine nitrate entering Lake Winnipeg. Within Lake Winnipeg, the isotopic composition of NO3− appeared to be only partly controlled by the isotopic composition of the riverine nitrate with a distinctive plume of nitrate to the South Basin from the Red River. Nitrate assimilation, and decay of N2 fixing cyanobacteria followed by nitrification were identified as potential lakeinternal processes modifying the isotopic composition of dissolved NO3−, leading to low δ 15NNO3 values especially in the North basin. In the South basin, elevated δ 15NNO3 values in spring changed to lower

145

values by summertime, which suggested a short residence time of nitrate, possibly due to rapid assimilation of nitrate. All 15N enriched nitrate stemming from the Red and Assiniboine Rivers appeared to be fully assimilated and sequestered in the South basin, with little, if any transfer to the North basin. The low summertime δ 15N values of nitrate in the North basin of Lake Winnipeg suggest that decay of N2 fixing cyanobacteria constitutes a significant additional nitrate source. This suggests that the whole lake riverine nutrient N loading models currently being used could be overly simplistic in predicting Lake Winnipeg's response to eutrophication due to the very dynamic lake-internal N cycle. An improved understanding of nutrient N impacts may require separate models for the North and South basins. Comparison of δ 15N values of lake nitrate with those of PON and fish in both basins revealed that dissolved nitrate partially affects the 15N distribution and hence the nitrogen flow in the aquatic food webs of Lake Winnipeg. From a landscape-scale nutrient management perspective, seasonal and long term monitoring of N and O isotopic composition of nitrate in streams and Lake Winnipeg may prove to be a useful tool to assess the efficacy of best management practices (BMPs) and longterm changes in agricultural practices (e.g. changes in fertilizer types and application rates, manure management, types of livestock and crops) and urban nutrient reduction strategies (N removal from STPs). For example, a shift from current predominantly animalwaste derived nitrate in most agriculturally contributing rivers towards better managed and optimized targeted inorganic fertilizers like urea in the Red River basin could eventually be accompanied by a decrease in riverine nitrate fluxes concomitant with a decrease in δ 15N values of nitrate. An example of isotopic monitoring of nutrient management is the decadal-scale δ 15N changes in coastal aquatic food webs of the Caribbean that revealed the relative increase in the proportion of sewage derived 15N enriched nutrients from the reduction and better management of synthetic fertilizers (Baker et al., 2010). Numerous questions regarding nitrate and N cycling in Lake Winnipeg and its watershed remain that provide opportunities for future targeted research. Little quantitative information is available concerning the residence time of nitrate and reactivity and availability of N forms in the contributing watersheds and the time scales of in-lake nitrate transformations. The upstream characterization of the nitrate and N isotopic composition of riverine systems (e.g. in biota and aquatic plants), particularly in the Red and Assiniboine River watersheds, may also help to distinguish impacts from STPs from other 15N hotspots where excessive leaching of animal waste nutrients to streams occurs. The relationship between in lake nitrate concentration and algal uptake, PON re-mineralization and algal N sequestration remains qualitative for each basin, and requires further controlled, preferably in-lake, experimentation. We submit that stable isotope analyses can play an important role in further elucidating the very dynamic N cycle in Lake Winnipeg affected by several N sources and rapid N transformations.

Acknowledgements

Fig. 10. Mean annual δ15N values of dissolved nitrate, particulate organic matter (PON), and fish (e.g. emerald shiners) in the North and South basin of Lake Winnipeg. Fish data are from (Hobson et al., 2012), PON δ15N data are unpublished data from Environment Canada.

This research was supported by the Lake Winnipeg Basin Initiative and funding from Environment Canada (LIW) and by support from the Natural Sciences and Engineering Research Council (NSERC) of Canada (BM). We thank Amy Ofukany, Mike Maksymchuk and Daryl Halliwell for assistance with field sampling. Emily Ritson-Bennett assisted with the GIS mapping. Sue Watson and Peter Leavitt provided useful comments on an earlier version of this manuscript. The constructive comments of two reviewers and the associate editor for improving this manuscript are also gratefully acknowledged. We especially thank the Lake Winnipeg Research Consortium (www. lakewinnipegresearch.com) and crew of the PV Namao for logistical and ship support.

146

B. Mayer, L.I. Wassenaar / Journal of Great Lakes Research 38 (2012) 135–146

References 15

18

Amberger, A., 1987. Natürliche N und O Gehalte als Indikatoren für die Herkunft von Nitrat in Boden und Grundwasser. Ph.D. thesis, Technical University Munich, Munich (Germany). Amberger, A., Schmidt, H.-L., 1987. Naturliche Isotopengehalte von Nitrat als Indikatoren fur dessen Herkunft. Geochim. Cosmochim. Acta 51, 2699–2705. Aravena, R., Mayer, B., 2010. Isotopes and processes in the nitrogen and sulfur cycles. In: Aelion, C.M., et al. (Ed.), Environmental Isotopes in Biodegradation and Bioremediation. Lewis Publishers, New York, pp. 203–236. Aravena, R., Robertson, W.D., 1998. Use of multiple isotope tracers to evaluate denitrification in ground water: study of nitrate from a large-flux septic system plume. Ground Water 36, 975–982. Aravena, R., Evans, M.L., Cherry, J.A., 1993. Stable isotopes of oxygen and nitrogen in source identification of nitrate from septic systems. Ground Water 31, 180–186. ASTM, 1992. Standard Methods for the Examination of Wastewater. American Public Health Association. 4–101 pp. Baker, D.M., Webster, K.L., Kim, K., 2010. Caribbean octocorals record changing carbon and nitrogen sources from 1862 to 2005. Glob. Chang. Biol. 16 (10), 2701–2710. Böttcher, J., Strebel, O., Voerkelius, S., Schmidt, H.-L., 1990. Using isotope fractionation of nitrate–nitrogen and nitrate–oxygen for evaluation of microbial denitrification in a sandy aquifer. J. Hydrol. 114, 413–424. Bourne, A., Armstrong, N., Jones, G., 2002. A preliminary estimate of the total nitrogen and total phosphorous loading to streams in Manitoba, Canada. Manitoba Water Stewardship, Winnipeg. 49 p. Bunting, L., Leavitt, P.R., Wissel, B., Laird, K.R., Cumming, B.F., St. Amand, A., Engstrom, D.R., 2011. Sudden ecosystem state change in Lake Winnipeg, Canada, caused by eutrophication arising from crop and livestock production during the 20th century. Manitoba Water Stewardship 72 p. Casciotti, K.L., Sigman, D.M., Hastings, M., Boehlke, J.K., Hilkert, A., 2002. Measurement of the oxygen isotopic composition of nitrate in seawater amd freshwater using the denitrifier method. Anal. Chem. 74, 4905–4912. Diaz, R.J., Rosenberg, R., 2008. Spreading dead zones and consequences for marine ecosystems. Science 321 (5891), 926–929. Durka, W., Schulze, E.D., Gebauer, G., Voerkelius, S., 1994. Effects of forest decline on uptake and leaching of deposited nitrate determined from 15N and 18O measurements. Nature 372 (6508), 765–767. Environment Canada, 2011. State of Lake Winnipeg: 1997–2007. Winnipeg, MB, 209 p. Glibert, P.M., Harrison, J., Heil, C., Seitzinger, S., 2006. Escalating worldwide use of urea — a global change contributing to coastal eutrophication. Biogeochemistry 77 (3), 441–463. Goolsby, D.A., Battaglin, W.A., Aulenbach, B.T., Hooper, R.P., 2000. Nitrogen flux and sources in the Mississippi River Basin. Sci. Total. Environ. 248 (2–3), 75–86. Granger, J., Sigman, D.M., Needoba, J.A., Harrison, J.A., 2004. Coupled nitrogen and oxygen isotope fractionation of nitrate during assimilation by cultures of marine phytoplankton. Limnol. Oceanogr. 49 (5), 1763–1773. Granger, J., Sigman, D.M., Rohde, M.M., Maldonado, M.T., Tortell, P.D., 2010. N and O isotope effects during nitrate assimilation by unicellular prokaryotic and eukaryotic plankton cultures. Geochim. Cosmochim. Acta 74 (3), 1030–1040. Gu, B.H., Chapman, A.D., Schelske, C.L., 2006. Factors controlling seasonal variations in stable isotope composition of particulate organic matter in a soft water eutrophic lake. Limnol. Oceanogr. 51 (6), 2837–2848. Hadas, O., Altabet, M.A., Agnihotri, R., 2009. Seasonally varying nitrogen isotope biogeochemistry of particulate organic matter in Lake Kinneret, Israel. Limnol. Oceanogr. 54 (1), 75–85. Hobson, K.A., Ofukany, A., Soto, D.X., Wassenaar, L.I., 2012. An isotopic baseline (δ13C, δ15N) for fishes of Lake Winnipeg: Implications for investigating impacts of eutrophication and invasive species. J. Great Lakes Res. 38 (Suppl. 3), 58–65. Hollocher, T.C., 1984. Source of the oxygen atoms of nitrate in the oxidation of nitrite by Nitrobacter agilis and evidence against a P-O-N anhydride mechanism in oxidative phosphorylation. Arch. Biochem. Biophys. 233, 721–727. Jones, G., Armstrong, N., 2001. In: Conservation, M. (Ed.), Long-term trends in total nitrogen and total phosphorus concentrations in Manitoba streams. Winnipeg, pp. 154 pp. Kendall, C., 1998. Tracing nitrogen sources and cycling in catchments. In: Kendall, C., McDonnell, J.J. (Eds.), Isotope Tracers in Catchment Hydrology. Elsevier, Amsterdam, pp. 521–576. Kendall, C., Elliott, E.M., Wankel, S.D., 2007. Tracing anthropogenic inputs of nitrogen to ecosystems. In: Michener, R.H., Lajtha, K. (Eds.), Stable Isotopes in Ecology and Environmental Science. Blackwell Publishing, Oxford, pp. 375–449. Kling, H.J., Watson, S.B., McCullough, G.C., Stainton, M.P., 2011. Bloom development and phytolankton succession in Lake Winnipeg: a comparison of histrorical records with recent data. Aquat. Ecosyst. Heal. Manag. 14 (2), 219–224. Kreitler, C.W., 1979. Nitrogen-isotope ratio studies of soils and groundwater nitrate from alluvial fan aquifers in Texas. J. Hydrol. 42, 147–170. Kroopnick, P., Craig, H., 1972. Atmospheric oxygen: Isotopic composition and solubility fractionation. Science 175, 54–55. Leavitt, P.R., Brock, C.S., Ebel, C., Patoine, A., 2006. Landscape-scale effects of urban nitrogen on a chain of freshwater lakes in central North America. Limnol. Oceanogr. 51 (5), 2262–2277. Lehmann, M.F., et al., 2004. Seasonal variation of the delta C-13 and delta N-15 of particulate and dissolved carbon and nitrogen in Lake Lugano: Constraints on biogeochemical cycling in a eutrophic lake. Limnol. Oceanogr. 49 (2), 415–429. Letolle, R., 1980. Nitrogen-15 in the natural environment. In: Fritz, P., Fontes, J.C. (Eds.), Handbook of Environmental Isotope Geochemistry: The Terrestrial Environment. Elsevier, Amsterdam, pp. 407–433.

Lis, G., Wassenaar, L.I., Hendry, M.J., 2008. High-precision laser spectroscopy D/H and O-18/O-16 measurements of microliter natural water samples. Anal. Chem. 80 (1), 287–293. MacGregor, B.J., et al., 2001. Microbiological, molecular biological and stable isotopic evidence for nitrogen fixation in the open waters of Lake Michigan. Environ. Microbiol. 3 (3), 205–219. Manitoba, 2008. Manitoba Water Stewardship 2007–2008 Annual Report, pp. 84–85. Winnipeg. Mariotti, A., et al., 1981. Experimental determination of nitrogen kinetic isotope fractionation: some principles; illustration for the denitrification and nitrification processes. Plant Soil 62, 413–430. Mariotti, A., Germon, J.C., Leclerc, A., 1982. Nitrogen isotope fractionation associated with the NO2− -> N2O step of denitrification in soils. Can. J. Soil Sci. 62, 227–241. Mayer, B., 2005. Assessing sources and transformations of sulphate and nitrate in the hydrosphere using isotope techniques. In: Aggarwal, P.K., Gat, J.R., Froehlich, F.O. (Eds.), Isotope in the Water Cycle: past, present and future of a developing science. Springer (IAEA), Dordrecht, pp. 67–89. Mayer, B., Bollwerk, S.M., Mansfeldt, T., Hutter, B., Veizer, J., 2001. The oxygen isotope composition of nitrate generated by nitrification in acid forest floors. Geochim. Cosmochim. Acta 65 (16), 2743–2756. Mayer, B., et al., 2002. Sources of nitrate in rivers draining sixteen watersheds in the northeastern U.S.: Isotopic constraints. Biogeochemistry 57 (58), 171–197. Mayer, B., Sebilo, M., Wassenaar, L.I., in press. Isotopic asessment of nitrogen cycling in river basins: potential and limitations for nutrient management purposes, Vienna, pp. McCullough, G., 2001. Organic carbon, nitrogen and phosphorous fluxes in rivers flowing into and out of Lake Winnipeg. Department of Fisheries and Oceans, Winnipeg, MB. 112 p. McCullough, G., 2009. Satellite images of Manitoba's great lakes. McCullough, G.K., Page, S.J., Hesslein, R.H., Stainton, M.P., Kling, H.J., Salki, A.G., Barber, D.G., 2012. Hydrological forcing of a recent trophic surge in Lake Winnipeg. J. Great Lakes Res. 38 (Suppl. 3), 95–105. Needoba, J.A., Sigman, D.M., Harrison, P.J., 2004. The mechanism of isotope fractionation during algal nitrate assimilation as illuminated by the 15N/14N of intercellular nitrate. J. Phycol. 40 (517–522). Patoine, A., Graham, M.D., Leavitt, P.R., 2006. Spatial variation of nitrogen fixation in lakes of the northern Great Plains. Limnol. Oceanogr. 51 (4), 1665–1677. Rock, L., Mayer, B., 2004. Isotopic assessment of sources of surface water nitrate within the Oldman River basin, Southern Alberta, Canada. Water Air Soil Pollut. Focus 4, 545–562. Rock, L., Mayer, B., 2006. Tracing nitrates and sulphates in river basins using isotope techniques. Water Sci. Technol. 53 (10), 209–217. Rupert, M.G., 2008. Decadal-scale changes of nitrate in ground water of the United States, 1988–2004. J. Environ. Qual. 37 (5), S240–S248. Schindler, D.W., Hecky, R.E., McCullough, G., 2012. The rapid eutrophication of Lake Winnipeg: greening under global change. Journal of Great Lake Research. 38 (Suppl. 3), 6–13. Shearer, G., Schneider, J.D., Kohl, H.D., 1991. Separating the efflux and influx components of net nitrate uptake by Synechococcus-R2 under steady-state conditions. J. Gen. Microbiol. 137, 1179–1184. Sigman, D.M., et al., 2001. A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem. 73, 4145–4153. Snider, D.M., Spoelstra, J., Schiff, S.L., Venkiteswaran, J.J., 2010. Stable Oxygen Isotope Ratios of Nitrate Produced from Nitrification: (18)O-Labeled Water Incubations of Agricultural and Temperate Forest Soils. Environ. Sci. Technol. 44 (14), 5358–5364. Stewart, A.R., et al., 2000. Influence of the 1997 Red River flood on contaminant transport and fate in southern Lake Winnipeg, Winnipeg, p. 190. Teranes, J.L., Bernasconi, S.M., 2000. The record of nitrate utilization and productivity limitation provided by delta N-15 values in lake organic matter - A study of sediment trap and core sediments from Baldeggersee, Switzerland. Limnol. Oceanogr. 45 (4), 801–813. Vitousek, P.M., et al., 1997. Human alteration of the global nitrogen cycle: sources and consequences. Ecol. Appl. 7, 737–750. Voerkelius, S., 1990. Isotopendiskriminierungen bei der Nitrifikation und Denitrifikation: Grundlagen und Anwendungen der Herkunfts-Zuordnung von Nitrat und Distickstoffmonoxid. TU Munich, Munich. 119 pp. Voss, M., Emeis, K.C., Hille, S., Neumann, T., Dippner, J.W., 2005. Nitrogen cycle of the Baltic Sea from an isotopic perspective. Global Biogeochem. Cycles 19 (3). Voss, M., et al., 2006. Source identification of nitrate by means of isotopic tracers in the Baltic Sea catchments. Biogeosciences 3 (4), 663–676. Wassenaar, L.I., 1995. Evaluation of the origin and fate of nitrate in the Abbotsford Aquifer using the isotopes of 15N and 18O in NO3−. Appl. Geochem. 10, 391–405. Wassenaar, L.I., 2012. Dissolved oxygen status of Lake Winnipeg: Spatio-temporal and isotopic (δ18O–O2) patterns. J. Great Lakes Res. 38 (Suppl. 3), 123–134. Wassenaar, L.I., Hendry, M.J., 2007. Dynamics and stable isotope composition of gaseous and dissolved oxygen. Ground Water 45 (4), 447–460. Wellman, R.P., Cook, F.D., Krouse, H.R., 1968. Nitrogen-15: Microbiological alteration of abundance. Science 161, 269–270. Yates, A.G., Culp, J.M., Chambers, P.A., 2012. Estimating nutrient production from human activities in subcatchments of the Red River. Manitoba. J. Great Lakes Res. 38 (Suppl. 3), 106–114. Zhang, W., Yerubandi, R.R., 2012. Application of a eutrophication model for assessing water quality in Lake Winnipeg. J. Great Lakes Res. 38 (Suppl. 3), 158–173. Zhao, J., Yerubandi, R.R., Wassenaar, L., 2012. Numerical modeling of hydrodynamics and tracer dispersion during ice-free period in Lake Winnipeg. J. Great Lakes Res. 38 (Suppl. 3), 147–157.