Northeastern Naturalist 120 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 22001155 NORTHEASTERN NATURALIST 2V2(o1l). :2122,0 N–1o4. 31 Observations of Greenhouse Gases and Nitrate Concentrations in a Maine River and Fringing Wetland Susan R. Bresney1,2,*, Serena Moseman-Valtierra3, and Noah P. Snyder1 Abstract - In the Sheepscot River, ME, we measured percent saturations of dissolved methane (CH4) and carbon dioxide (CO2) and concentrations of nitrate (NO3 -) four times in the main stem and once in the West Branch. River water was super-saturated with CH4 at all sites throughout the study, and measurements were generally higher at lower-gradient sites (1000–5000% saturation) than higher-gradient sites (generally <1000%). Percent saturations of CO2 in the main stem varied in both time and space and were under-saturated at some sites. CO2 percent saturations and NO3 - concentrations in the more-developed West Branch were significantly higher than the main stem, likely because of the position of mainstem sites downstream of Sheepscot Pond where primary production and degassing could occur. We also measured CH4 and CO2 fluxes from wetland soil adjacent to the main stem, which averaged 710 (± 59) μmolCH4/m2/h and -51 (± 6.4) mmolCO2/m2/h. Our findings suggest that rivers and fringing wetlands in the formerly glaciated northeastern US contribute to the production of greenhouse gasses, and that dissolved methane shows spatial variations with channel morphology. Introduction Methane (CH4) and carbon dioxide (CO2) are greenhouse gases (GHGs) that have increased by 48% and 35%, respectively, in the atmosphere since preindustrial times (Forster et al. 2007). Although CO2 is the most abundant GHG in the atmosphere, CH4 has a greater global warming potential (per molecule) than CO2 by 25-fold in a 100-year period (IPCC 2007). About one third of all CH4 emissions are estimated to be from natural ecosystems (Denman et al. 2007, Townsend-Small and Czimczik 2010), and natural freshwater wetlands are estimated to contribute the most CH4 to the atmosphere among all natural and anthropogenic sources (Gedney et al. 2004). River systems are major conduits of C and N from watershed landscapes to the ocean and atmosphere (Abril et al. 2014, Beaulieu et al. 2010). Riverine wetlands can play an important role in the retention, alteration, and eventual discharge of C and N to the atmosphere and coasts (Johnston 1993, LaFleur 2009, Wetzel 1990). For example, recent estimates suggest that wetlands in the Amazon river basin rapidly export half of their gross primary production as dissolved CO2 and organic carbon in the river water (Abril et al. 2014). However, the role that GHGs play in riverine carbon budgets of northeastern North America is poorly quantified. Typical 1Earth and Environmental Sciences Department, Boston College, 140 Commonwealth Avenue, Chestnut Hill, MA 02467. 2Current address - Department of Civil and Environmental Engineering, Tufts University, Medford, MA 02155. 3Biological Sciences Department, University of Rhode Island, 45 Upper College Road, Kingston, RI 02881. *Corrresponding author - sbresney@gmail.com. Manuscript Editor: Peter Raymond Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 121 conditions of freshwater riverine wetlands, such as high organic content and low oxygen availability of soils, are conducive to the microbial production of some GHGs, particularly CH4. In anoxic wetland soils, microbial communities produce CH4 during the terminal stage of organic decomposition (Wang et al. 1996). This process of methanogenesis dominates C metabolism in anoxic conditions in freshwater (Capone and Kiene 1988). CO2 concentrations in stream water are governed mainly by groundwater inputs, reflecting soil respiration which produces CO2, and by in-stream primary productivity, reflecting photosynthesis which reduces CO2 (Jones and Mulholland 1998, Salisbury et al. 2008), as well as decomposition of organic matter (Humborg et al. 2010) and weathering reactions. Like CH4, production of CO2 is regulated by the availability of dissolved organic carbon (DOC), and is also influenced by available N, O2, and temperature (Neal et al. 1998). While microbial and in-stream metabolic processes may drive greenhouse-gas production and fluxes in soils, the morphology of rivers, especially in northeastern North America, may promote the absorption, retention, and release of GHGs to the atmosphere. The rivers in this region flow through a landscape conditioned by glaciation and retreat during the late Pleistocene (e.g., Snyder et al. 2009, 2013). The result was longitudinal profiles of river channels that alternate among steep (slope >5×10-3), gravel-bedded reaches, low-gradient reaches (slope <10-3) often with fringing wetlands, and lakes. Just as one study found a significant correlation between channel gradient and partial pressure of GHGs (Lauerwald et al. 2013), these geomorphic characteristics could be promoting the cycling of GHGs, through absorption and retention in the low-gradient wetlands and release by off-gassing in the high-gradient reaches. Differences in land use and the presence or absence of wetlands may influence riverine GHG concentrations, but these systems are complex and require extensive studying to understand how they contribute to overall C and N budgets on a larger scale. Systems of channels and fringing freshwater wetlands may be producing and retaining GHGs that ultimately are emitted to the atmosphere. In this study, we measured spatial patterns of percent saturations of dissolved GHGs and concentrations of NO3 - upstream, within, and downstream from a fringing wetland along the main stem of the Sheepscot River in midcoast Maine (Fig. 1) during 4 sampling periods between June 2010 and July 2011. For comparison, we measured these parameters in a segment of the West Branch with contrasting land-use and morphology (Fig. 1) in June 2011. We also studied net vertical fluxes of GHGs from wetland soils in a fringing wetland of the main stem. This study has 2 objectives: (1) to quantify the dissolved GHG and NO3 - concentrations and basic water-quality parameters (temperature, DO, pH, and conductivity) in surface waters of a 4.5-km section of the main stem of the Sheepscot River that passes through a wetland and includes large variations in channel gradient; and (2) to compare GHG concentrations between the main stem and West Branch of the river, which have contrasting watershed land use. We based our study on 2 hypotheses: (1) dissolved GHG percent saturations in the river surface-water reflect channel morphology, being highest in the low-gradient reach, adjacent to the Northeastern Naturalist 122 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 wetland and lowest at the high-gradient sites; and (2) GHG concentrations in river water are higher in the West Branch than in the main stem, reflecting different landuse patterns in the watershed. Methods Study area The 590-km2 Sheepscot River watershed is located in midcoast Maine and drains into Sheepscot Bay and then the Gulf of Maine (Fig. 1). The watershed is mainly forested or used for agriculture (Fig. 1; McLean et al. 2007). The Sheepscot River main stem is 74 km long and flows northeast to southwest along the Norumbega Fault Zone (Osberg et al. 1985). Due to bedrock and glacial geology, the Sheepscot River channel morphology varies from gravel and bedrock beds in the high-gradient reaches to wetlands and lakes where the stream crosses flatlands (Fig. 2 SVCA 2005; Snyder et al. 2009, 2013). Figure. 1. Map of the northern Sheepscot River watershed showing sampling stations and land cover (inset table) for the 2 watershed areas relevant to this study. Land-cover key only includes categories with >1% coverage in either subwatershed (land-cover dataset from MaineGIS 2006). Inset map shows location of the study area within Maine. Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 123 The Sheepscot River has been the site of several water-quality monitoring efforts (Arter 2004, McLean et al. 2007, Whiting 2006) because it, like other rivers that discharge into the Gulf of Maine, contains important spawning and rearing habitat for the endangered native Salmo salar (Atlantic Salmon; SVCA 2005). Studies have found high nutrient levels in sections of this river (Arter 2004, McLean et al. 2007, SVCA 2005, Whiting 2006). One study declares that the Sheepscot River is “highly enriched” in nitrogen and phosphorus compared to similar rivers (Whiting 2006), and another considers Sheepscot Bay to be one of the most eutrophic coastal areas in the US (Bricker et al. 1999). Sampling sites Our 6 main sampling sites (MS1–MS6) are located along a 4.5-km segment of the main stem Sheepscot River between Sheepscot Pond and Long Pond (Fig. 1), where the river is a third-order stream. This segment is just downstream from Sheepscot Figure 2. Longitudinal profiles and graphs showing the spatial pattern of percent saturations of GHGs in river water across the 6 main stem (A–C) and 4 West Branch (D–F) sample sites. Error bars are calculated as the standard error for the mean of samples collected at river right, left, and thalweg at each site or sites (except in October 2010 when only right and thalweg positions were sampled), but statistical analyses were conducted using the 3 samples as replicates. Sites are classified as high gradient or low grad ient based on slope (Table 2). Northeastern Naturalist 124 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 Pond and the Palermo Rearing Station, a facility run by the Maine Department of Inland Fisheries and Wildlife that rears Salvelinus fontinalis (Mitchill) (Brook Trout) and Salmo trutta L. (Brown Trout). MS1 is located in a gravel-bedded, high-gradient upstream reach, MS2–MS5 are located in a 2.65-km reach that flows through a ~0.6- km2 vegetated wetland, and MS6 is in a gravel-bedded, high-gradient downstream reach (Fig. 2a). Our vertical flux measurements and soil collections (detailed below) were focused in wetland sites near MS3 (Fig. 1) where dominant vegetation cover was live and dead Eleocharis obtusa Willd. (Blunt Spikerush). Our study also included 4 sampling sites along a 9-km segment of the West Branch Sheepscot River (WB1–WB4; Fig. 1), which is also a third-order stream. This segment is characterized by a relatively low-gradient upstream reach, a gravel-bedded higher-gradient reach at WB2, and a low-gradient reach downstream (Fig. 2d). WB1–WB4 are each located just downstream of the confluences of 4 tributaries with the West Branch (Hewett Brook, Dearborn Brook, Griffin Brook, Wingwood Brook, respectively). NO3 - concentrations above 80 μM have been reported in these tributaries (Whiting 2006). In contrast to the main stem, the West Branch study segment does not contain a large riverine wetland or a nearby lake. The source of the West Branch is Branch Pond, which is >16 km upstream from WB1. This upstream segment is mostly low gradient (slope <10-3) and includes several fringing wetlands. Land cover upstream of both our main stem and West Branch study areas is predominately forested (Fig. 1). The area upstream of our West Branch study area (drained through WB4) is covered by significantly more developed and agricultural land (6% and 13%, respectively) and significantly less forested land (73%) than the area upstream of our main stem study area, drained through MS6 (2% developed, 5% agricultural, 79% forested) (t6 = -2.99, P = 0.02). Surface-water sampling and data collection To characterize the dissolved GHG concentrations of the main stem surface water, we collected three 40-ml water samples from each site (MS1–MS6), in July 2010, June 2011, and July 2011. All samples were collected from 3 points (about 1 m from each bank and in the thalweg) separated by an average of 7 m across the channel at each site. In October 2010, only 2 water samples from each site were collected (at points 1 m from the river right bank and the thalweg) from sites MS2– MS5, one sample was collected from the thalweg at site MS6, and no samples were taken at MS1 due to limitations on time. All sampling occurred during periods of relatively low discharge (Fig. 3). Also, we collected three 20-ml water samples from each site for analyses of NO3 - + NO2 - concentrations (which will hereafter be referred to just as NO3 - for brevity) during each summer sampling period (July 2010, June 2011, and July 2011). To contrast surface waters of the main stem with those of the West Branch, we collected 3 water samples for GHG and NO3 - analyses from each site in the West Branch (WB1–WB4) in June 2011. We collected all of the water samples from 15–20 cm below the surface or if water depth was less than 15 cm, from half way between the surface and riverbed. Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 125 Water samples for GHG analysis were collected based on protocols outlined by the Environmental Protection Agency (EPA 2001). For preservation, 0.6 ml of 50% w/v ZnCl2 was added to the 40-ml samples. We collected samples for NO3 - analysis by filling a plastic 60-ml syringe and filtering water through a Millipore Sterivex Eastar co-polyester filter with a 0.22-μm polyethersulfone membrane into 20-ml high-density polyethylene containers. These samples were kept on ice while in the field and frozen within 8 hours of collection until analysis. The samples were later analyzed for concentration of NO3 - by the Woods Hole Oceanographic Institution (WHOI) Nutrient Analytical Facility (July 2010 samples) with a Lachat Instruments QuickChem 8000 four-channel continuous-flow injection system, and the University of New Hampshire (UNH) Water Quality Analysis Lab (June and July 2011 samples) with a SmartChem® 200 discrete wet chemistry analyzer. Along with each water-sample collection, we measured conductivity (μS/cm), temperature (°C), and dissolved oxygen (DO) content (in mg/L and percent saturation) using a YSI Professional Plus Multiparameter meter. The YSI probe was placed at the sampling site at the same depth explained above immediately after GHG and NO3 - samples were collected. Figure 3. Graph showing discharge in the Sheepscot River between 1 May 2010 and 1 August 2011. Dots indicate dates that samples and/or data were collected; the 4 sampling periods are marked with arrows. Discharge values between 4 January 2011 and 10 March 2011 are estimated values while the river was frozen. Discharge data was measured by the USGS at gauging station 01038000 in North Whitefield, ME (Fig. 1). Northeastern Naturalist 126 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 Surface-water GHG analysis To analyze dissolved GHG concentrations in the water samples, we equilibrated 35 ml of each water sample with 25 ml of N2 headspace gas, and shook samples by hand for 1 minute (based on EPA 2001). In the lab, we recorded water temperature by placing a thermometer into the water sample immediately after each equilibration. Salinity of the water was measured with a refractometer. We determined concentrations of CH4 and CO2 in the headspaces of each sample via gas chromatography (Shimadzu GC-2014 equipped with a split/splitless injection port) and a hydrogen flame ionization detector. Column temperatures and flow rates followed manufacturer specifications (Shimadzu). The detection limits are 0.1 ppm for CH4 and 10 ppm for CO2. We calculated the solubility and percent saturation of CH4 and CO2 in these riverine surface-water samples using formulas derived from Walter et al. (2005). We calculated the equilibrium constant of each gas, K0, based on Weiss and Price (1980) and used the gas constant, R, based on Eby (2004). GHG fluxes from wetland sediments In the wetland, 5 plots with similar plant composition were selected for chamber experiments in June 2011. These plots were positioned within approximately 600 m of each other, at and upstream from MS3 (Fig. 1). We installed metal collars (28 cm diameter, 6 cm deep) in the wetland soil at each of these sites 2 weeks prior to carrying out the experiments. To determine CH4 and CO2 fluxes, we placed transparent polycarbonate chambers (28 cm diameter, 34 cm height, 22,940 cm3 volume) equipped with small fans and coiled stainless steel tubes (Moseman-Valtierra et al. 2011) on top of the previously installed steel collars sealed with water-filled channels in foam rings. We retained intact live plants inside of these chambers and folded any plants taller than the chamber without breaking them. During the flux chamber measurements, we measured temperature and light levels at sediment surfaces inside and outside of chambers with Onset Hobo pendant data loggers. We did not find a significant difference in temperature (paired t4 = 0.88, P = 0.42) or light intensity (paired t4 = 0.24, P = 0.82) between the interior and exterior of the chambers. We collected a series of seven 60-ml gas samples from each chamber in intervals lasting 2–8-minutes, beginning immediately after the chamber was in place until 22 minutes later. Gas samples were collected from a 2-m Tygon tube connected to each chamber, following methods similar to those of Moseman-Valtierra et al. (2011). The tubing allowed for us to collect samples from at least 2 m away to minimize sediment disturbance. Before the collection, the tubing was flushed by repeatedly filling the syringe and expelling to the atmosphere. Total volume of flushing and sampling was <1% of total chamber volume. We conducted gas-flux measurements at all 5 sites on 6 July 2011. To store the gas samples, we transferred them underwater into pre-evacuated 12-ml Labco Exetainer vials within 6 hours of collection until analysis. Prior testing of the plastic syringes and Exetainers showed no change of gas concentration during a period of Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 127 2–3 weeks, which is the length of time the samples were stored (Moseman-Valtierra et al. 2011). We determined concentrations of CH4 and CO2 in gas samples via gas chromatography as described above. We calculated gas fluxes from the change in gas concentration within the chamber headspace over time in each chamber and chamber height using Fick’s law (Healy et al. 1996, Moseman-Valtierra et al. 2011). Statistical analyses We tested differences in the percent saturation of GHGs in river surface water between dates and among sampling sites using 2 separate two-factor ANOVAs in which factor 1 was date and factor 2 was site. For both of these analyses, 3 measurements at each site (right bank, left bank, and thalweg) were used as replicates because they did not differ from each other on a given date. The first two-factor ANOVA included all of the 6 sites (MS1–MS6) in the main stem of the Sheepscot River for 3 dates (July 2010, June 2011, July 2011), and the second examined all 4 dates (July 2010, October 2010, June 2011, July 2011) but only included the sites that were sampled all 4 times (MS2–MS5). As only 2 samples were collected from each site in October 2010 (at the left bank and in the thalweg of the river), only samples from those 2 positions were used per site. These 2 analyses agreed with each other except for data that were present in 1 model exclusively. The former analysis (all sites) was used primarily to infer spatial patterns, while the latter (all dates) was used primarily for temporal patterns. When significant effects of site or date were found, Tukey’s honestly significant difference tests were used in post hoc analyses to determine which specific means differed from each other. Percent saturation of CH4 saturation was log transformed to meet assumptions of normality. We also tested differences in the concentration of NO3 - and percent saturation of GHGs from the main stem and West Branch study areas with t-tests in which the 3 replicate measurements at each site served as replicates for comparisons between branches. To compare pH, temperature and DO in the main stem between sampling dates and sites, we applied the same two-factor ANOVAs described for dissolved gas concentrations above. For vertical GHG-flux measurements, the significance of fluxes in each chamber was determined via regressions of gas concentrations and time (with r2 ≥ 0.90 for significant fluxes). Relationships between GHG fluxes from wetland soils and environmental factors (temperature and light intensity) were evaluated with linear regressions. The differences in temperature and light intensity between the interior of chambers and the exterior of the chambers (testing artificial chamber effects) were evaluated with paired t-tests. Results Spatial patterns of percent saturations of GHGs in surface water of the Sheepscot River In the main stem of the Sheepscot River, percent saturations of CH4 were generally greater in the low-gradient sites (within the riverine wetland) than in the high-gradient sites (Fig. 2b). When all sites were compared among July 2010, June Northeastern Naturalist 128 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 2011, and July 2011, the percent saturation of CH4 in water samples from the 2 highgradient sites (MS1 and MS6) were significantly lower than at 3 of the low-gradients sites (MS3–MS5) (F5,50 = 17.45, P < 0.0001; Tables 1 and 2, Fig. 2b). This result is generally consistent with hypothesis 1, in terms of CH4, that dissolved GHG concentrations would be greatest in the low-gradient reach and lowest at the high-gradient sites (Fig. 2b). However, the upstream low-gradient site (MS2) was intermediate between the other low-gradient sites (MS3–MS5) and the high-gradient sites (MS1 and MS6). When all 4 sampling dates were included (sites MS2–MS5 only), percent saturations of CH4 were highest at the middle low-gradient site (MS4) and lowest at the first low-gradient site (MS2) (F3,43 = 7.57, P < 0.001). In contrast with CH4, spatial variations in the percent saturation of CO2 did not as clearly relate to low- or high-gradient sites (Tables 1, 2; Fig. 2c). When all sites were compared in July 2010, June 2011, and July 2011, a high-gradient site (MS6) had significantly lower percent saturation of CO2 than 3 low-gradient sites (MS3– MS5) (F5,50 = 7.66, P < 0.001). The high-gradient site MS1 was intermediate, as was the low-gradient site MS2 (Table 2, Fig. 4b). When all 4 dates were included (sites MS2–MS5 only), the first low-gradient site (MS2) had significantly lower percent saturation of CO2 than MS5, while the other low-gradient sites (MS3–MS4) were intermediate (F3,43 = 5.55, P = 0.004). Temporal variability was notable for both GHGs in the main stem. The percent saturation of both dissolved GHGs varied significantly between the 4 sampling periods among sites MS2–MS6 (CO2: F3,43 = 198.78, P < 0.001; CH4: F3,,43 = 9.53, Table 1. (A) Effects of site and time on dissolved GHGs and water-quality properties of the main stem of the Sheepscot River. Two-factor ANOVAs were used to compare upstream high-gradient, lowgradient, and downstream high-gradient sites across sampling dates. Due to incomplete sampling, one of the two-factor ANOVAs included all 4 dates but a subset of sites, while the other included all sites but only 3 of 4 dates. (B) t, degrees of freedom (df), and P-values for t-tests used to compare GHG saturations and NO3 - concentrations between the main stem and West Branch. Cond. = conductivity, Temp. = temperature. A. Two-factor ANOVAs All dates All sites (sites MS 2–MS5 only) (July 2010, June 2011, July 2011 only) Site Time Site x time Site Time Site x time % CO2 F3,43 = 5.5, F3,43 = 198.78, F9,43 = 1.17, F5,50 = 7.66, F2,50 = 576.50, F10,50 = 6.24, P = 0.004 P < 0.001 P = 0.349 P < 0.001 P < 0.001 P < 0.001 % CH4 F3,43 = 7.57, F3,43 = 9.53, F9,43 = 1.44, F5,50 = 17.45, F2, 50= 0.19, F10,50 =1.11, P = 0.0007 P = 0.0002 P = 0.22 P < 0.0001 P = 0.82 P = 0.38 Cond. F3,44= 2.20, F3,44 = 93.32, F9,44 = 0.89, F5,51 = 6.73, F2,51 = 35.52, F10,51 = 0.88, P = 0.11 P < 0.0001 P = 0.54 P = 0.002 P < 0.001 P = 0.55 Temp. F3,44 = 12.46, F3,44 =2431.91, F9,44 = 11.57, F5,51 = 51.78, F2,51 = 993.65, F10,51 = 13.23, P < 0.001 P < 0.001 P < 0.001 P < 0.0001 P < 0.0001 P < 0.0001 B. t-test t df P NO3 - June 2011 -5.29 8 0.01 CO2 June 2011 10.31 8 <0.01 Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 129 Table 2. Data for slope (from a lidar digital elevation model), discharge (Q), NO3 - concentration, percent saturation (% sat) and concentration (ppm) of CH4 and CO2, specific conductivity (SC), water temperature (T), dissolved oxygen (DO), pH, water velocity (V), water depth (h), and channel width (w) for each site during each sampling period. n per site refers to the number of samples or data collected at that site. Reported values are the average of collections from each site. The July 2010 sample set has two dates and discharge values because DO and pH data were collected on the second date. Discharge is the average discharge for that date reported by the USGS (Fig. 3). In October 2010, no samples or data were collected from MS 1 and NO3 -, and V, h, and w data were not collected due to limited time in the field. n per Q NO3- CH4 CH4 CO2 CO2 SC T DO DO V h w Site site Date Slope (m3/s) (uM) % sat (ppm) (% sat) (ppm) (uS/cm) (°C) (%) (mg/L) pH (m/s) (cm) (m) MS1 3 7/15/10, 7.7×10-3 2.0, 3.4 358.4 0.9 188.8 329.9 36.8 23.5 90.5 7.92 7.50 0.19 31 11.5 7/20/10 1.2 MS2 3 7/15/10, 8.8×10-4 2.0, 3.7 1523.1 9.0 159.7 301.8 36.1 24.6 97.2 8.45 7.38 0.10 40 10.6 7/20/10 1.2 MS3 3 7/15/10, 8.5×10-4 2.0, 2.9 3009.9 15.1 183.4 288.4 37.8 25.4 93.3 8.09 7.21 0.08 64 11.2 7/20/10 1.2 MS4 3 7/15/10, 4.6×10-5 2.0, 1.7 3039.4 18.5 193.7 430.8 40.4 26.4 89.6 7.85 7.21 0.00 95 0.0 7/20/10 1.2 MS5 3 7/15/10, 1.9×10-4 2.0, 2.1 3092.6 14.1 209.6 340.9 38.0 25.7 82.3 6.86 7.02 0.08 50 35.0 7/20/10 1.2 MS6 3 7/22/10, 7.1×10-3 1.6, 3.3 1095.9 7.1 108.6 373.5 37.7 23.4 91.0 7.55 7.09 0.35 22 5.5 7/20/10 1.2 MS2 2 10/24/10 8.8×10-4 3.7 395.2 3.6 87.8 250.0 35.8 9.4 94.4 10.73 7.36 MS3 2 10/24/10 8.5×10-4 3.7 542.6 5.3 99.5 239.1 36.1 8.8 83.5 9.71 7.40 MS4 2 10/24/10 4.6×10-5 3.7 852.4 18.5 113.1 430.8 36.3 8.3 88.4 10.19 7.13 MS5 2 10/24/10 1.9×10-4 3.7 664.1 14.1 138.8 340.9 36.2 7.5 81.5 9.75 7.29 MS6 1 10/24/10 7.1×10-3 3.7 475.1 5.3 115.5 315.9 36.2 7.6 86.2 10.27 7.10 Northeastern Naturalist 130 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 Table 2, continued. n per Q NO3- CH4 CH4 CO2 CO2 SC T DO DO V h w Site site Date Slope (m3/s) (uM) % sat (ppm) (% sat) (ppm) (uS/cm) (°C) (%) (mg/L) pH (m/s) (cm) (m) MS3 3 6/16/11 8.5×10-4 3.7 2.2 2149.8 14.7 31.3 59.5 35.4 18.7 92.9 8.68 7.53 0.30 30 3.8 MS1 3 6/15/11 7.7×10-3 4.5 1.6 121.3 1.0 20.6 54.9 35.5 16.9 93.6 9.12 7.57 0.25 24 4.2 MS2 3 6/16/11 8.8×10-4 3.7 1.8 991.3 7.3 26.8 64.9 35.9 17.8 91.9 8.78 7.63 0.19 42 3.8 MS4 3 6/16/11 4.6×10-5 3.7 1.2 5016.3 38.3 28.3 65.5 33.6 19.4 88.0 7.98 7.48 0.02 76 14.8 MS5 3 6/16/11 1.9×10-4 3.7 2.1 2642.6 20.1 30.3 95.0 37.9 18.4 77.8 7.17 7.42 0.16 41 5.4 MS6 3 6/15/11 7.1×10-3 4.5 1.6 1453.6 11.0 35.9 85.5 37.0 17.0 88.4 8.42 7.57 0.60 31 2.6 MS1 3 7/7/11 7.7×10-3 2.3 2.9 111.3 0.8 22.8 53.8 35.8 19.9 89.9 8.20 8.07 0.28 13 4.1 MS2 3 7/7/11 8.8×10-4 2.3 2.5 1370.4 9.7 31.9 70.0 36.7 21.5 93.8 8.26 7.98 0.10 33 3.3 MS3 3 7/7/11 8.5×10-4 2.3 2.0 3078.8 19.4 57.6 113.4 37.2 22.5 90.1 7.79 8.75 0.23 30 3.1 MS4 3 7/7/11 4.6×10-5 2.3 2.0 3668.2 21.1 52.0 101.6 37.7 23.9 86.2 7.27 7.98 0.03 85 0.0 MS5 3 7/7/11 1.9×10-4 2.3 0.6 2147.2 12.4 50.2 86.4 37.9 18.4 77.8 7.17 7.42 0.16 41 5.4 MS6 3 7/7/11 7.1×10-3 2.3 0.8 840.2 5.0 41.5 73.5 37.0 23.7 91.5 7.71 7.98 0.33 11 1.8 WB1 3 6/15/11 5.1×10-4 4.5 4.8 1186.4 11.2 404.2 843.6 90.7 17.6 94.5 9.01 7.84 0.05 48 5.8 WB2 3 6/17/11 6.0×10-3 3.4 3.2 884.1 8.5 409.8 817.6 81.0 20.1 78.8 7.22 7.85 0.08 40 0.0 WB3 3 6/17/11 1.5×10-3 3.4 5.5 2172.6 6.2 526.9 799.7 84.5 18.7 86.0 8.03 7.72 0.46 33 3.5 WB4 3 6/15/11 7.5×10-4 4.5 6.8 1042.6 7.6 298.7 705.9 81.1 16.5 100.0 9.79 7.76 0.81 23 2.3 Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 131 P = 0.0002; Tables 1 and 2, Fig. 2). CH4 percent saturations were significantly higher in the summer sampling periods of both years (July 2010, June and July 2011) than during October 2010. In June and July 2011, CO2 was under-saturated and significantly lower than both dates (July and October) of th e previous year. River water properties did not closely match the spatial and temporal patterns of dissolved GHG percent saturations, but there were some patterns with river gradient. For instance, conductivity was greatest in one of the low-gradient sites (MS5) and lowest at sites MS1 and MS2, while other sites were intermediate (F5,51 = 6.73, P = 0.0002). Surface temperatures of the river water were greatest at the lowgradient sites MS4 and MS5, and lowest at the high-gradient site MS1, while other sites were intermediate (F5,5 1= 51.78, P < 0.0001). Neither the percent of dissolved oxygen nor pH were related in a simple way to river gradient, as both were lowest at MS5 (DO: F5,51 = 6.90, P = 0.01; pH: F17,51 = 17.50, P = 0.004), while all remaining sites did not significantly differ from each other. In temporal comparisons of water properties, conductivity and temperature were lower in October 2010 than on all other dates (F3,44 = 93.32 P < 0.0001 and F3,44 = 2431.91, P < 0.0001, respectively). The temperature differed significantly between all 4 dates, with July 2010 being greatest, followed by July 2011, June 2011, and October 2010. The surface-water pH was highest in July 2011, with June 2011 being intermediate and July 2010 being lowest among all 4 sampling dates (F3,44 = 61.53, P < 0.0001). In contrast, the percent saturation of dissolved oxygen was similar across all dates (F3,44 = 0.46, P = 0.71). Dissolved GHGs in branches of the Sheepscot River with contrasting land use The study area in the West Branch of the Sheepscot River, where the watershed contains more development, roads, and agricultural land cover (Fig. 1), had significantly higher surface-water concentrations of NO3 - and higher percent saturations of CO2 than the main-stem study area during our measurements in June 2011 (Tables 1, 2; Fig. 4a, c). This finding is consistent with hypothesis 2 (that GHG saturations are higher in the West Branch than the main stem) for one of two measured GHGs (CO2). All sampling sites on the West Branch were super-saturated with CH4 and CO2 relative to the atmosphere, while the main stem was under-saturated with CO2 when this comparison was conducted. Although we measured qualitatively higher percent saturations of CH4 in the surface water of the main stem study area than in the West Branch, we did not find a significant difference between the 2 study areas (Figs. 2b, 2e, 4b). Spatially, percent saturations of GHGs in the West Branch did not show any significant pattern (Fig. 2e, f). This result is consistent with our hypothesis 1 that GHGs would reflect changes in geomorphology, because the West Branch study reach is more consistently low gradient than the main stem and does not include an extensive riparian wetland. Fluxes of GHGs from wetland soil During July 2011, significant positive fluxes of CH4 were measured from wetland soils, ranging from 4.2 to 43 mmol m-2/d-1, while significant negative fluxes Northeastern Naturalist 132 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 of CO2, ranging from -5200 to -340 mmol m-2d-1, were found in all plots except C3 (Table 3). These fluxes were not significantly related to either temperature or light. Discussion Patterns of dissolved riverine GHGs relative to geomorphic gradients High percent saturation of CH4 in surface water of the Sheepscot River in the vicinity of the riverine wetland and lower percent saturation of CH4 at the high- Figure 4. Bar graphs showing concentrations of NO3 - (A) and percent saturations of GHGs (B–C) in river water at the main-stem sample sites and the West Branch sample sites. Bars represent the mean of all samples collected in June 2011 from all sites in each study area, and error bars are calculated as the standard error of these samples. Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 133 gradient sites (Fig. 2) suggests that the main stem of the river is likely releasing CH4 to the atmosphere in the study area. This spatial pattern is consistent with gas absorption and retention in the low-gradient reach, and off gassing in the highgradient, gravel-bedded reaches (hypothesis 1; Fig. 2b). One major control on methanogenesis is availability of organic carbon (Jones and Mulholland 1998), which is frequently related to gradient and riverbed substrate grain size. Therefore, the lower concentrations of CH4 in the high-gradient reaches may be due to larger grain size and less available organic material as opposed to the low-gradient reach, which is underlain by organic-rich mud. The transition from high to low gradient occurs in the main stem between sites MS1 and MS2. The slope decreases and the wetland forest begins >0.5 km upstream of MS2. The residence time increases through this transition, as shown by a decrease in stream velocity from MS1 to MS2 (Table 2), and this slowdown becomes more pronounced farther downstream in the low-gradient reach (MS3–MS5). Slower water velocities may allow more CH4 from the riverbed and vegetated banks to build up in stream water, while the highgradient reaches are more turbulent, likely causing more off-gassing. An increase in temperature at the lower-gradient sites (Table 2) may also contribute to an increase in methanogenesis at those locations. There is a qualitative (but not significant) decrease in percent saturation of CH4 between MS4 and MS5, two low-gradient sites, during the summer sampling periods (Fig. 2b). This result is likely related to the increase in slope and velocity between these two sites (Table 2), suggesting an increase in off-gassing. The entire length of the Sheepscot River is characterized by these alternating steep and low-gradient reaches (SVCA 2005; Snyder et al. 2009, 2013), suggesting that this river as a whole and other rivers in the region with similar geomorphology could be sources for CH4 to the atmosphere. Although CO2 showed a trend of increase in the wetland and off-gassing in the high-gradient reaches during the summer sampling periods (Table 1, Fig. 2c), the pattern is not as clear as with CH4, and does not include all sites. CO2 saturations in river water are influenced by abiotic processes such as the weathering of carbonate and silicate rocks, but we cannot speculate on how variations in weathering rates may influence our observations. Biological processes, such as methanogenesis, may better reflect the change in morphology between non-wetland and wetland sites, given that the organic wetland soils fuel microbial activities. CO2 percent saturations, in contrast, may reflect more physical and chemical reactions in the river, Table 3. Mean fluxes of greenhouse gases from five plots during the July 2011 experiments, measured in the field. CH4 CO2 Site mmol/m2/day mmol/m2/day C1 15.0 -520 C2 4.2 -1700 C3 9.0 0 C4 43.0 -340 C5 22.0 -5200 Northeastern Naturalist 134 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 and changes in CO2 may be buffered by carbonate equilibria within the river water. Thus, CO2 percent saturations likely do not respond to the morphology changes in the same way as CH4. In this study, we found elevated levels of NO3 - and CO2 in the West Branch when compared to the main stem (Fig. 4a, c). We attribute the generally higher NO3 - in the West Branch to the more-developed land use in the area that it drains (Figs. 1, 4a; Table 2). NO3 - concentrations in the West Branch are relatively low (Table 4), even if higher than the main stem, and are likely not high enough to cause an increase in CO2 production. More likely, low percent saturations of CO2 in the main stem are reflective of high primary productivity (Salisbury et al. 2008) and possible degassing during the relatively long water-residence time in Sheepscot Pond, the source of water for the main-stem sites (Figs. 1, 2a). While site MS1 is about Table 4. Average concentrations of NO3 - from different rivers, regions, and EPA criteria. Rio Tempesquito Sur, located in South America is considered undisturbed because it has 80% or more natural vegetative cover, population density of less than 5 individuals/km2, and less than 2.5 kg/ha/yr anthropogenic deposition of NO3 -. “Ecoregion” refers to the level III Omernik ecoregions of the continental US, created by the EPA based on similar ecosystems with similar types, qualities, and quantities of environmental resources. This value represents the average concentration of NO3 - in streams and rivers in Maine’s ecoregion. The EPA water-quality standards for the state of Florida are for streams, there is a range due to the different criteria for different watersheds and this is the only state so far that has standards for NO3 -. WWTP is wastewater treatment plant. Area Avg NO3 -(μmol N/L) Source Main stem Sheepscot River 1.75 This study West Branch Sheepscot River 5.07 This study Average Sheepscot River watershedA 20 Whiting 2006 Highest Sheepscot River watershedA 225B Whiting 2006 Rio Tempesquito Sur (undisturbed watershed) 10 Lewis et al. 1999 Rivers in Maine ecoregion 19.1 EPA 2010b EPA water quality standards for Florida 48–134C EPA 2010a Husdson (tidal) 60 Cole and Caraco 2001 Temmesjoki (eutrophic) 36–132 Silvennoinen et al. 2008 Upstream seine (upstream of WWTP and high 179 Garnier et al. 2009 population density) Downstream seine (downstream of WWTP and 286 Garnier et al. 2009 high population density) Boghall drainage ditch (drainage from highly 221–428 Reay et al. 2003 fertilized agricultural land) South Platte (N enriched by WWTP discharge) 425–700 McMahon and Dennehy 1999 AWatershed refers to the river itself as well as tributaries to that river. BEstimated 90th percentile from the SVCA main stem site with the highest NO3 - values (Whiting 2006). CThese numbers represent the concentration of total N, not just NO3 -; to compare, total N is about double NO3 -. Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 135 2 km downstream from the outlet of Sheepscot Pond, WB1 is more than 16 km downstream from Branch Pond. Therefore, in our study areas, the main stem river waters may be more representative of lake water, while higher CO2 saturations in the West Branch are more characteristic of stream water. High productivity should be reflected by higher saturation of dissolved oxygen in the main stem (Table 2), which was also well oxygenated as a result of turbulence. Despite the trend of higher CH4 in surface water of the main stem than in the West Branch (Figs. 2b, 2e, 4b), which may suggest a link between proximity of surface water to a wetland and CH4 percent saturations, we did not see a significant difference in percent saturation of CH4 between these two branches. This finding is likely due to the high variability within the main stem, where values ranged from 1000 to 5000 percent saturation within the low-gradient sites. The lack of significant difference may also be explained by: (1) CH4 addition in several wetlands along the West Branch 2–15 km upstream of WB1, and (2) little opportunity for off-gassing in the relatively low-gradient segment from WB1 to WB4 (Fig. 2). This interpretation suggests that although there may not be a direct correlation between wetland proximity and surface water CH4, there is likely a link between stream morphology and CH4, as discussed above. Dissolved GHG and NO3 - within the Sheepscot River and comparisons to other rivers As we observed in both the main stem (on most dates) and West Branch of the Sheepscot River (Fig. 2b, d; Table 2), river channels are frequently super-saturated with CH4 and CO2 with respect to the atmosphere (Tables 5, 6; Hope et al. 2004, Jones and Mulholland 1998, Neal et al. 1998, Raymond et al. 2013). Although water in the Sheepscot River during this study was generally super-saturated with these gases, the concentrations we observed were relatively low when compared to published data from other rivers (Tables 5, 6). Studies have found contradicting patterns between different rivers pertaining to the fluctuation of percent saturations of CH4 temporally, some having the highest saturations in winter (Silvennoinen et al. 2008), others finding higher saturations in summer (Jones and Mulholland 1998), or one study finding no seasonal patterns at all (Hope et al. 2004). Seasonal effects may be the cause for some variation between our small data set for the summer and fall and other larger data sets focused only in the summer, winter, or year round (Table 5). Additionally, the forested, temperate rivers most closely related to the Sheepscot River (Tyne, Elbe, Ouse, Hudson) are much larger and therefore may have the capacity to produce and retain more CH4 (Table 5). The relatively low percent saturation of CO2 observed in this study, and particularly the under-saturation of CO2 that was observed in June and July 2011 in the main stem, is unusual for freshwater systems. There are many factors that affect CO2 saturations in stream water, and any combination of them may be affecting the saturations measured in this study. Biological productivity in the source waters (Sheepscot Pond) of the main stem and off-gassing of CO2 within that lake may be the main factor contributing to the low CO2 percent saturations that we observed Northeastern Naturalist 136 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 Table 5. Values of percent saturation of CH4 in surface water of different rivers in the literature. Standard errors are in parentheses. Rivers are roughly listed from low to high based on average percent saturation. River % saturation Measuring period Source West Branch Sheepscot 1110 (22.7) June This study study area Main stem Sheepscot 1710 (19.8) October, June, July This study study area Tyne 75–4129 December Upstill-Goddard et al. 2000 Rio San Pedro (tidal salt 514–5000 Year round Ferron et al. 2007 marsh creek) Elbe 1750–3500 Summer Wernecke et al. 1994 Ouse 3861 (667) December Upstill-Goddard et al. 2000 Amazon 6300 (1050) Spring Bartlett et al. 1990 Temmesjoki (eutrophic, 10,270 (1620) Summer Silvennoinen et al. 2008 freshwater) Hudson (tidal, 4400-42,400 Summer de Angelis and Scranton 1993 freshwater) Temmesjoki (eutrophic, 15,440 (1420) Year round Silvennoinen et al. 2008 freshwater) Table 6. Values of percent saturation and partial pressures of CO2 in surface water of different rivers in the literature. Standard errors are in parentheses. Rivers are listed from low to high based on average percent saturation and partial pressure. Partial pressures were used due to limited percent saturation data in the literature. % pCO2 Measuring River saturation (μatm) period Source Main stem Sheepscot River 88.7 (0.99) 498.4 (6.0) October, This study June, July Third-order streams in Sweden 1950 Year round Humborg et al. 2010 (mainly boreal forested) West Branch Sheepscot River 343.0 (8.9) 1969.3 (20) June This study 1120 rivers in North America 2091 Year round Lauerwald et al. 2013 Rivers in conterminous US 2109 Jones et al. 2003 Third-order streams in the 3000 Year round Butman and Raymond 2011 northern US Rivers in the temperate zone 3200 Aufdenkampe et al. 2011 (25°–50° latitiude) Streams in the temperate zone 3500 Aufdenkampe et al. 2011 (25°–50° latitiude) Rio San Pedro (tidal salt 295–1270 Feb–May, Ferron et al. 2007 marsh creek) July, Sept Temmesjoki (eutrophic, 560 (69) Summer Silvennoinen et al. 2008 freshwater) Temmesjoki (eutrophic, 890 (89) Year round Silvennoinen et al. 2008 freshwater) Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 137 (as discussed above). Additionally, we may have seen low saturations because our samples were collected during relatively low discharge (Fig. 3). Both Butman and Raymond (2011) and Lauerwald (2013) suggest that precipitation may affect CO2 concentrations on a short timescale, specifically by flushing fringing wetland soils and delivering CO2 to the river water. Our sampling periods may have missed times after rain events when concentrations are higher. As previously mentioned, CO2 production is regulated by availability of dissolved organic carbon (DOC), N, temperature, and O2 (Neal et al. 1998) and also may be correlated with pH (Lauerwald et al. 2013). When we tested these factors (pH, concentration of NO3 -, temperature, and DO) with the percent saturation of CO2, we did not find any significant relationships, but we did see some trends. During this study, we measured pH values that were generally higher than those measured in other similarly sized Maine rivers, which may suggest a link to lower CO2. The US EPA (2012) reports an average pH of 6.30 for 19 other wadeable streams in Maine, which is much lower than our average of 7.56 (Table 2). Because, like CO2 concentrations, pH may be controlled by and negatively correlated with precipitation (Lauerwald et al. 2013), it is possible that our data shows higher pH values because we collected samples during relatively low discharge (Fig. 3). Also, both pH and CO2 in streams are heavily influenced by biogeochemical processes as well as gas exchange between the water and atmosphere (Lauerwald et al. 2013). Possibly, processes that are beyond the scope of this study are influencing the pH and CO2 content of river water. Because pH and CO2 are negatively correlated (Lauerwald et al. 2013), these findings of higher pH during our study may suggest why overall CO2 percent saturations were low. Although various studies have detected high NO3 - levels in some parts of the Sheepscot River, and considered the main stem to contain significant nutrient pollution (Arter 2004, Bricker et al. 1999, McClean et al. 2007, Whiting 2006), we found low NO3 - concentrations during this limited study in all sites that we examined (Table 2). This result is especially evident when the average concentration from this study is compared with other “undisturbed” watersheds, or rivers with known NO3 - pollution sources (Table 4). The generally low NO3 - concentrations in both the main stem and the West Branch are consistent with a study of anthropogenic nitrogen sources to 16 rivers in the northeastern US, which concluded that 4 rivers in Maine had the lowest nutrient concentrations (Boyer et al. 2002). Because water chemistry and, specifically, nutrient concentrations are temporally dynamic and vary with dry and wet seasons (Peterson and Benning 2013, Shields et al. 2013), our measurements likely missed nutrient pulses that could have occurred during periods with higher runoff that other studies may have observed, as our measurements were focused in times of relatively low discharge (Fig. 3). Low nutrient availability may lead to low organic content and low respiration, which might have contributed to the low CO2 percent saturations we observed as compared to other rivers (Table 6), but further investigation of this and other rivers is needed to determine a significant link. Additionally, because our data set is limited mainly to the summer during the day time when primary productivity is high, and under-saturation of CO2 is likely Northeastern Naturalist 138 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 the result of high phytoplankton productivity (Salisbury et al. 2008), our sampling may have missed periods throughout the day and year when respiration is greater and primary production is less. However, high productivity should be reflected by super-saturation of dissolved oxygen, which we did not see during this study (Table 2). Water sourcing, precipitation, pH, NO3 -, stream temperature, and primary productivity might all be contributing factors to the low CO2 percent saturations observed in this study, but none alone are sufficient to support our findings. Overall, carbon cycling in river systems is complex, and with our small snapshot of data from only a few days of sampling in the Sheepscot River, it is difficult to determine a single explanation for our low reported CH4 and CO2 percent saturations. The possible sources and sinks for CO 2 and CH4 in the Sheepscot River, as well as in other rivers where these greenhouse gases have been investigated (e.g., Butman and Raymond 2011, Striegl et al. 2012), are still not fully quantified, and therefore a more detailed study of these processes is needed to determine their influence on GHGs in the atmosphere. GHG fluxes from wetland soils Significant positive fluxes of CH4 from wetland soil adjacent to the main stem (Table 3), suggests that the wetland itself is also likely a contributor of CH4 to the atmosphere. Interestingly, CH4 fluxes from wetland soils (average = 12.5 mg CH4 m-2 h-1) during this study exceeded those measured in salt marshes in Massachusetts (average range = 0.05–5 mg CH4 m-2 h-1) by Moseman-Valtierra et al. (2011) by about two orders of magnitude despite identical techniques. Fluxes measured in this study also greatly exceeded fluxes from a brackish coastal marsh in the Chesapeake Bay (average range = 0.6–2.6 mg CH4 m-2 h-1; Bartlett et al. 1987). This striking difference is consistent with higher CH4 production from freshwater wetlands than salt- or brackish-water wetlands that has been reported previously (Kang and Freeman 2002, Wang et al. 1996). However, it is important to note that the exceptionally large fluxes measured in this study are likely also due to our small sample set of one hot day in July when microbial activity was high. In the main-stem wetland soil, CO2 fluxes were negative (indicating consumption) at all plots except C3 (Table 3) where no significant CO2 flux was detected. The CO2 consumption rates that we observed in the wetland along the main stem are consistent with CO2 uptake reported in other natural wetlands, where photosynthesis generally exceeds respiration (Liikanen et al. 2009, Whiting and Chanton 2001), especially during the day or during the growing period (Lafleur 2009). Again, our chamber deployments were limited to a sunny day in July, when plant productivity was likely at its peak, and so the CO2 consumption rates we measured (average = 2920 mg CO2 m-2 h-1) are exceptionally high relative to other studies that observed CO2 production including another freshwater wetland (average range = 21–35; Roobroeck et al. 2010), another study on the US east coast (average = 226 mg CO2 m-2 h-1; Morris and Whiting 1985), and Moseman- Valtierra et al.’s (2011) study in New England salt marshes with similar chamber techniques (average = 380 mg CO2 m-2h-1). Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 139 Conclusions We found dissolved CH4 to be super-saturated in surface waters in a section of the main stem of the Sheepscot River relative to the atmosphere during 4 sampling periods between June 2010 and July 2011 (Fig. 4a). Spatial patterns of percent saturation of dissolved CH4 in river water appear to generally be linked to channel geomorphology, with CH4 absorption and retention in the low-gradient reach and off-gassing in the high-gradient reaches (Fig. 2a, b), suggesting that during this study this channel was likely a source of CH4 to the atmosphere. During June and July 2011, we observed under-saturation of CO2 in the river water (Fig. 2c). This result may be linked to primary productivity and off-gassing in the lake upstream of the study area, as well as low precipitation and higher pH in the river water at that time (Tables 1, 2) and a combination of other water-chemistry factors. Our small sample size, focused during the day in mainly summer months when primary productivity is high, also likely influenced the low levels. Additionally, soils in a fringing wetland showed positive fluxes of CH4, suggesting that the wetland was a source of CH4 during the time of the study (Table 3), and negative fluxes of CO2, suggesting that the wetland was a sink for CO2 on the sunny summer day when we sampled (Table 3). Concentrations of NO3 - and the CO2 in river water were both significantly greater in the West Branch of the Sheepscot River than in the main stem, likely suggesting a relationship of land development in the watershed with nutrient loading, but there is not enough evidence to support a link between NO3 - and CO2 (Fig. 4). Elevated CO2 levels in the West Branch are likely reflective of river-water processes, as opposed to lake-water processes that may have influenced the main stem. During our study period (focused in summer months with low river discharge; Fig. 3), concentrations of NO3 - in the Sheepscot River were generally low. Temporal dynamics and biogeochemical responses of riverine systems to more chronic nitrogen loads, including responses of wetlands within them, should be investigated more thoroughly to better understand impacts of watershed development and water quality on GHG emissions. These data from relatively pristine reaches of the Sheepscot River in Maine may serve as initial references for more disturbed ecosystems in the northeastern US and elsewhere. Acknowledgments This work was funded by National Science Foundation award 0645343. We thank the University of New Hampshire and the Woods Hole Oceanographic Institution for surfacewater NO3 - analyses, the US Geological Survey for discharge data provided and assistance in equilibration procedures and calculations, and the University of Maine Darling Marine Center for accommodations during field work. We thank Dan Hallstrom, Kim Rhodes, Billy Armstrong, Andrew Nesheim, and Stephanie Strouse for assistance during sampling and data collection, and the Rafuse family for river access. This manuscript was greatly improved by comments from 2 anonymous reviewers and Manuscript Editor Peter Raymond. Northeastern Naturalist 140 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 Literature Cited Abril, G., J. Martinez, L.F. Artigas, P. Moreira-Torcq, M.F. Benedetti, L. Vidal, T. Meziane, J. Kim, M.C. Bernardes, N. Savoye, J. Deborde, E.L. Souza, P. Alberic, M.F. Landim de Souza, and F. Roland. 2014. Amazon River carbon dioxide outgassing fuelled by wetlands. Nature 505:395–398. Arter, B.S. 2004. Sheepscot River water quality monitoring strategic plan. Krisweb. Available on line at http://www.krisweb.com/biblio/sheepscot_share_arter_2004_wqmp.pdf. Accessed June 2010 Aufdenkampe, A.K., E. Mayorga, P. Raymond, J.M. Melack, S.C. Doney, S.R. Alin, R.E. Aalto, and K. Yoo. 2011. Riverine coupling of biogeochemical cycles between land, oceans, and atmosphere. Frontiers in Ecology and the Environment 9:53–60. Bartlett, K.B., P.M. Crill, J.A. Bonassi, J.E.Richey, and R.C. Harriss. 1990. Methane flux from the Amazon River floodplain: Emissions during rising water. Journal of Geophysical Research: Atmospheres 95:16773–16788. Beaulieu, J.J., C.P. Arango, S.K. Hamilton, and J.L. Tank. 2010. The production and emission of nitrous oxide from headwater streams in the Midwestern USA. Global Change Biology 14:878–894. Boyer, E.W., C.L. Goodale, N.A. Jaworski, and R.W. Howarth. 2002. Anthropogenic nitrogen sources and relationships to riverine nitrogen export in the northeastern USA. The Nitrogen Cycle at Regional to Global Scales 57/58:137–169. Bricker, S.B., C.G. Clement, D.E. Pirhalla, S.P. Orlando, and D.R.G. Farrow. 1999. National estuarine eutrophication assessment: Effects of nutrient enrichment in the nation’s estuaries. NOAA, National Ocean Service, Special Projects Office and the National Centers for Coastal Ocean Science. Available online at http://ian.umces.edu/neea/pdfs/ eutro_report.pdf. Accessed 10 June 2010. Butman, D., and P.A. Raymond. 2011. Significant efflux of carbon dioxide from streams and rivers in the United States. Nature Geoscience 4:839–842. Capone, D.G., and R.P. Kiene. 1988. Comparison of microbial dynamics in marine and freshwater sediments: Contrast in anaerobic carbon metabolism. Limnology and Oceanography 33:725–749. Cole, J.J., and N.F. Caraco. 2001. Emissions of nitrous oxide from a tidal, freshwater river, the Hudson River, New York. Environmental Science and Technology 35:991–996. de Angelis, M.A., and M.I. Scranton. 1993. Fate of methane in the Hudson River and estuary. Global Biogeochemical Cycles 7:509–523. Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C. Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X. Zhang. 2007. Coupling between changes in the climate system and biogeochemistry. Pp. 500–587, In S. Solomon, D. Qin, M. Manning, Z.Chen, et al. (Eds.) Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, New York, NY.. Eby, G.N. 2004. Principles of Environmental Geochemistry. Brooks/Cole Publishing Company, Cole, CA. Environmental Protection Agency (EPA). 2001. Technical guidance for the natural attenuation indicators: Methane, ethane, and ethane. Available online at http://www.epa.gov/ region1/measure/Natatten.pdf. Accessed 20 June 2010. EPA. 2010a. 40 CFR Part 131. Water quality standards for the State of Florida’s lakes and flowing waters. Federal Register Vol. 75 No. 233. Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 141 EPA. 2010b. National nutrient database Available online at http://water.epa.gov/scitech/ swguidance/standards/criteria/nutrients/database_index.cfm. Accessed 20 June 2010. EPA. 2012. Maine Field Chemistry. Available online at http://www.epa.gov/region1/lab/ reportsdocuments/wadeable/sites/me_state/fieldche.html. Accessed 1 December 2013. Ferron, S., T. Ortega, A. Gomez-Parra, and J.M. Forja. 2007. Seasonal study of dissolved CH4 CO2 and N2O in a shallow tidal system of the bay of Cadiz (SW Spain). Journal of Marine Systems 66:244–257. Forster, P., V. Ramaswamy, P. Artaxo, T. Bernsten, et al. 2007. Changes in atmospheric constituents and in radiative forcing. Pp. 130–234, In S. Solomon, D. Qin, M. Manning, Z.Chen, et al. (Eds.). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, New York, NY. Garnier, J., G. Billen, G. Vilian, A. Martinez, and M. Silvestre. 2009. Nitrous oxide in the Seine River and basin: Observations and budgets. Agriculture, Ecosystems, and Environment 133:223–233. Gedney, N., P.M. Cox, and C. Huntingford. 2004. Climate feedback from wetland methane emissions. Geophysical Research Letters 31. doi:10.1029/2004GL020919. Healy, R.W., R.G. Striegl, T.F. Russell, G.L. Hutchinson, and G.P. Livingston. 1996. Numerical evaluation of static-chamber measurements of soil-atmosphere gas exchange: Identification of physical processes. Soil Science Society of America Journal 60:740–747. Hope, D., S.M. Palmer, M.F. Billett, and J.J. Dawson. 2004. Variations in dissolved CO2 and CH4 in a first-order stream and catchment: An investigation of soil-stream linkages. Hydrological Processes 18:3255–3275. Humborg, C., C.M. Morth, M. Sundbom, H. Borg, T. Blenckner, R. Geisler, and V. Ittekkot. 2010. CO2 supersaturation along the aquatic conduit in Swedish watersheds as constrained by terrestrial respiration, aquatic respiration, and weathering. Global Change Biology 16:1966–1978. Johnston, C.A. 1993. Material fluxes across wetland ecotones in northern landscapes. Ecological Applications 3:424–440. Jones, J.B., and P.J. Mulholland. 1998. Influence of drainage basin topography and elevation on carbon dioxide and methane supersaturation of stream water. Biogeochemistry 40:57–72. Jones, J.B., E.H. Stanley, and P.J. Mulholland. 2003. Long-term decline in carbon dioxide supersaturation in rivers across the contiguous United States. Geophysical Research Letters 30:2-1–2-4. Kang, H., and C. Freeman. 2002. The influence of hydrochemistry on methane emissions from two contrasting northern wetlands. Water Air and Soil Pollution 141:263–272. LaFleur, P.M. 2009. Connecting atmosphere and wetland: Trace gas exchange. Geography Compass 3/2:560–585. Laurwald, R., J. Hartmann, N. Moosdorf, S. Kempe, and P.A. Raymond. 2013. What controls the spatial patterns of the riverine carbonate system? A case study for North America. Chemical Geology 337–338:114–127. Lewis, W.M., J.M. Melack, W.H. McDowell, M. McCleain, and J.E. Richey. 1999. Nitrogen yield from undisturbed watersheds in the Americas. Biogeochemistry 46:149–162. Liikanen, A., H. Silvennoinen, A. Karvo, P. Rantakokko, and P.J. Martikainen. 2009. Methane and nitrous oxide fluxes in two coastal wetlands in the northeastern Gulf of Bothnia, Baltic Sea. Boreal Environment Research 14:351–368. Northeastern Naturalist 142 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 Vol. 22, No. 1 Maine Office of Geographic Information Systems (MaineGIS). 2006. MELCD land-cover map. Available online at http://megis.maine.gov/catalog/. Accessed April 2010. McLean, J., L. Sewall, and L. Pugh. 2007. Sheepscot River watershed management plan. Prepared for Time and Tide Resource Conservation and Development Area, Augusta, ME. Available online at http://www.kcswcd.org/Projects/Sheepscot/WHOLE%20 PLAN%20Jan26.07.pdf. Accessed June 2010. McMahon, P.B., and K.F. Dennehy. 1999. N2O emissions from a nitrogen-enriched river. Environmental Science and Technology 33:21–25. Morris, J.T., and G.J. Whiting. 1985. Gas advection in sediments of a South Carolina salt marsh. Marine Ecology-Progress Series 27:187–194. Moseman-Valtierra, S., R. Gonzalez, K.D. Kroeger, J. Tang, W.C. Chao, J. Crusius, J. Bratton, A. Green, and J Shelton. 2011. Short-term nitrogen additions can shift a coastal wetland from a sink to a source of N 2O. Atmospheric Environment 45:4390–4397. Neal, C., House, W.A., H.P. Jarvie, and A. Eatherhall. 1998. The significance of dissolved carbon dioxide in major lowland rivers entering the North Sea. Science of the Total Environment 210/211:187–203. Osberg, P.H., A.M. Hussey, III, and G.M. Boone. 1985. Bedrock Geologic Map of Maine. Maine Geological Survey, Department of Conservation, Augusta, ME. Peterson, E.W., and C. Benning. 2013. Factors influencing nitrate within a low-gradient agricultural stream. Environmental Earth Sciences 68:1233–1245. Raymond, P.A., J. Hartmann, R. Lauerwald, S. Sobek, C. McDonald, M. Hoover, D. Butman, R. Striegl, E. Mayora, C. Humborg, P. Kortelainen, H. Durr, M. Meybeck, P. Ciais, and P. Guth. 2013. Global carbon dioxide emissions from inland waters. Nature 503:355–359. Reay, D.S., K.A. Smith, and A.C. Edwards. 2003. Nitrous oxide emission from agricultural drainage waters. Global Climate Change Biology 9:195–203. Roobroeck, D., K. Butterbach-Bahl, N. Brueggemann, and P. Boeckx. 2010. Dinitrogen and nitrous oxide exchanges from an undrained monolith fen: Short-term responses following nitrate addition. European Journal of Soil Science 61:662–670. Salisbury, J.E., D. Vandemark, C.W. Hunt, J.W. Campbell, W.R. McGillis, and W.H. Mc- Dowell. 2008. Seasonal observations of surface waters in two Gulf of Maine estuaryplume systems: Relationships between watershed attributes, optical measurements, and surface pCO2. Estuarine, Coastal, and Shelf Science 77:245–252. Saunders, D.L., and J. Kalff. 2001. Nitrogen retention in wetlands, lakes, and rivers. Hydrobiology 443:205–212. Sheepscot Valley Conservation Association (SVCA). 2005. KRIS Sheepscot. Available online at http://www.krisweb.com/krissheepscot/krisdb/html/krisweb/index.htm. Accessed 20 June 2010. Shields, F.D., R.E. Lizotte, and S.S. Knight. 2013. Spatial and temporal water-quality variability in aquatic habitats of a cultivated floodplain. River Research and Applications. 29:313–329. Silvennoinen, H., A. Liikanen, J. Rinalta, and P.J. Martikainen. 2008. Greenhouse gas fluxes from the eutrophic Temmessjoki river and its estuary in the Liminganlahti Bay (the Baltic Sea). Biogeochemistry 90:193–208. Snyder, N.P., M.R. Castele, and J.R. Wright. 2009. Bedload entrainment in low-gradient paraglacial coastal rivers of Maine, USA: Implications for habitat restoration. Geomorphology 103:430–46. Snyder, N.P., A.O. Nesheim, B.C. Wilkins, and D.A. Edmonds. 2013. Predicting grain size in gravel-bedded rivers using digital elevation models: Application to three Maine watersheds. Geological Society of America Bulletin 125:148–163. Northeastern Naturalist Vol. 22, No. 1 S.R. Bresney, S. Moseman-Valtierra, and N.P. Snyder 2015 143 Striegl, R.G., M.M. Dornblaser, C.P. McDonald, J.R. Rover, and E.G. Stets. 2012. Carbon dioxide and methane emissions from the Yukon River system. Global Biogeochemical Cycles. 26:GB0E05 Townsend-Small, A., and C.I. Czimczik. 2010. Carbon sequestration and greenhouse gas emissions in urban turf. Geophysical Research Letters 37: doi:10.1029/2009GL041675. Upstill-Goddard, R.C., J. Barnes, T. Frost, S. Punshon, and N.J.P. Owens. 2000. Methane in the southern North Sea: Low-salinity inputs, estuarine removal, and atmospheric flux. Global Biogeochemical Cycles 14:1205–1217. Walter, S., I. Peeken, K. Lotche, A. Webb, and W. Bange. 2005. Nitrous oxide measurements during EIFEX, the European Iron Fertilization Experiment in the subpolar South Atlantic Ocean. Geophysical Research Letters 32. doi:10.1029/2005GL024619. Wang, Z., D. Zeng, and W.H. Patrick. 1996. Methane emissions from natural wetlands. Environmental Monitoring and Assessments 42:143–161. Weiss, R.F., and B.A. Price. 1980. Nitrous oxide solubility in water and seawater. Marine Chemistry 8:347–359. Wernecke, G., G. Floser, S. Korn, C. Weitkamp, and W. Michaelis. 1994. First measurements of the methane concentration in the North Sea with a new in-situ device. Bulletin of the Geological Society of Denmark 41:5–1 1. Copenhagen, Denmark. Wetzel, R.G. 1990. Land–water interfaces: Metabolic and limnological regulators. Verhandlungen- Internationale Vereinigung Für Theoretische und Angewandte Limnologie 24:6–24. Whiting, M. 2006. Water-quality summary for the Sheepscot River. Gulf of Maine. Available online at www.gulfofmaine.org/kb/uploads/1723/sheepscotsummary1.doc. Accessed July 2010. Whiting, G.J., and J.P. Chanton. 2001. Greenhouse carbon balance of wetlands: Methane emission versus carbon sequestration. Tellus B 53:521–528.