Sources and fate of nitrate in groundwater at agricultural operations overlying glacial sediments

Leaching of nitrate (NO3) from animal waste or fertilizers at agricultural operations can result in NO3 10 contamination of groundwater, lakes, and streams. Understanding the sources and fate of nitrate in groundwater systems in glacial sediments, which underlie many agricultural operations, is critical for managing impacts of human food production on the environment. Elevated NO3 concentrations in groundwater can be naturally attenuated through mixing or denitrification. Here we use isotopic enrichment of the stable isotope values of NO3 to quantify the amount of denitrification in groundwater at two confined feeding operations overlying glacial 15 sediments in Alberta, Canada. Uncertainty in δNNO3 and δONO3 values of the NO3 source and denitrification enrichment factors are accounted for using a Monte Carlo approach. When denitrification could be quantified, we used these values to constrain a mixing model based on NO3 and Cl concentrations. Using this novel approach we were able to reconstruct the initial NO3-N concentration and NO3-N/Cl ratio at the point of entry to the groundwater system. Manure filtrate had total-nitrogen (TN) of up to 1820 mg L, which was predominantly 20 organic-N and NH3. Groundwater had up to 85 mg L TN, which was predominantly NO3. The addition of NO3 to the local groundwater system from temporary manure piles and pens equalled or exceeded NO3 additions from earthen manure storages at these sites. On-farm management of manure waste should therefore increasingly focus on limiting manure piles in direct contact with the soil, and encourage storage in lined lagoons. Nitrate attenuation at both sites is attributed to a spatially variable combination of mixing and denitrification, but is dominated by 25 denitrification. Where identified, denitrification reduced agriculturally-derived NO3 concentrations by at least half and, in some wells, completely. Infiltration to groundwater systems in glacial sediments where NO3 can be naturally attenuated is likely preferable to off-farm export via runoff or drainage networks, especially if local groundwater is not used for potable water supply.

Concentrations of NO3will also decrease along groundwater flow paths due to attenuation via dilution by hydrodynamic dispersion (referred to hereafter as mixing). Because of these natural attenuation mechanisms, 5 infiltration to groundwater may be preferable to off-site drainage and runoff of nitrate-rich waters. Many agricultural operations are undertaken on fertile soils associated with glacial sediments (Spalding and Exner, 1993;. Understanding the sources and fate of agriculturally derived nitrate in groundwater systems in glacial sediments is therefore critical for managing impacts of human food production on the environment. 10 Identification of the sources and fate of NO3at agricultural operations can be challenging because of spatial and temporal variations in sources (e.g. earthen manure storage, temporary manure piles, or fertilizer) and heterogeneity in hydrogeologic systems (Spalding and Exner, 1993;Rodvang et al., 2004;Showers et al., 2008;Kohn et al., 2016). These spatial and temporal variations can result in complex subsurface solute distributions that are difficult to interpret using classical transect studies or numerical groundwater models (Green et al., 2010;15 Baily et al., 2011).
Groundwater containing significant agriculturally derived NO3also typically has elevated chloride (Cl -) concentrations (Saffigna and Keeney, 1977;Rodvang et al., 2004;Menció et al., 2016). Decreasing NO3-N/Cl -(or NO3 -/Cl -) ratios have been used to define denitrification based on the assumption that NO3is reactive while Clis non-reactive (conservative), such that denitrification results in a decrease in the NO3-N/Clratio (Kimble et 20 al., 1972;Weil et al., 1990;Liu et al., 2006;McCallum et al., 2008). However, NO3-N/Clratios can also change in response to mixing of groundwater with different NO3-N/Clratios or when groundwater sampling traverses hydraulically disconnected formations (Bourke et al., 2015b). If NO3-N/Clratios vary among potential sources and the NO3-N/Clratio at the point of entry to the groundwater system can be reconstructed, this information could be used to show that anthropogenic NO3at different locations within an aquifer is derived from the same 25 or different sources.
The stable isotopes of NO3 -(δ 15 NNO3 and δ 18 ONO3) provide an alternative approach to characterize the source and fate of NO3in groundwater systems. In agricultural areas, multiple sources of NO3are common and could include precipitation, soil NO3 -, inorganic fertilizer, manure, and septic waste (Komor and Anderson, 1993;Liu et al., 2006;Pastén-Zapata et al., 2014;Xu et al., 2015). While source identification is theoretically 30 possible using δ 15 NNO3 and δ 18 ONO3 (particularly with a dual-isotope approach), in practice this can be difficult due to geologic heterogeneity, overlapping source values, and the complexity of biologically mediated reactions (Aravena et al., 1993;Wassenaar, 1995;Mengis et al., 2001;Choi et al., 2003;Vavilin and proportion of substrate that has undergone a given reaction, if enrichment factors and source values are known; as in the case of evpoarative loss of water, for example . To date, there have been very few attempts to quantify denitrification using dual-isotope enrichment, largely due to uncertainty in source values and enrichment factors , Xue et al., 2009).
The only published calculations of the fraction of NO3remaining after denitrification the that we are aware of 5 assumed a constant enrichment factor and the same isotopic source values across the field site (Otero et al., 2009).
However, the enrichment factor will vary across a field site in response to reaction rates (Kendall and Aravena 2000), and isotopic values of even the same type of source (e.g. manure) can vary substantially (Xue et al., 2009).
If the varation in source values and enrichment factors can be characterized from measured data then these uncertainties can be accounted for using a Monte Carlo approach (Joerin et al., 2002;Ji et 10 al., 2017), thereby extending the application of the dual-isotope technique to allow for a robust quantitative assessment of denitrification in agricultural settings.
A synthesized analysis of stable isotopes of NO3with additional ionic tracers can further improve the assessment of NO3attenuation mechanisms and sources of NO3in agricultural settings (Showers et al., 2008;Vitòria et al., 2008;Xue et al., 2009;Xu et al., 2015;Ji et al., 2017). We hypothesise that if the amount of denitrification can 15 be quantified based on δ 15 NNO3 and δ 18 ONO3, then this estimate of the fraction of NO3-N removed through denitrification can be used to constrain a mixing model based on NO3-N and Clconcentrations. This novel approach allows for the ratio of NO3-N/Clat the point of entry to the groundwater system to be reconstructed from measured NO3and Clconcentrations (see Section 2.3). Where the NO3-N/Clratio varies between sources, this ratio can then be used to assess the source of the NO3in groundwater (e.g. temporary manure piles or feeding 20 pens). These data can also then be used to estimate the initial concentrations of NO3and Clat the point of entry to the groundwater system and quantify attenuation by mixing.
In this study, we present the application of this approach at two confined feeding operations (CFOs) in Alberta, Canada, with differing lithologies and durations of operation (Fig. 1). Concentrations of Cland nitrogen species (N-species) and the stable isotopes of NO3were measured in groundwater samples collected from monitoring 25 wells and continuous soil cores, as well as manure filtrate at both sites. These data were interpreted to (1) assess the extent of agriculturally derived NO3in groundwater, (2) identify sources and initial concentrations of NO3at the point of entry to the groundwater system, and (3) assess mixing and denitrification as attenuation mechanisms at these sites.

Experimental sites
This study was conducted using data from two of the five sites investigated by Alberta Agriculture and Forestry during an assessment of the impacts of livestock manure on groundwater quality (Lorenz et al., 2014). To the best of our knowledge (including discussions with farm operators) fertilizers have not been applied at either of these sites. As such, manure waste from livestock is assumed to be the sole source of agricultural nitrogen (N) and 35 elevated NO3concentrations in groundwater at these sites.
The first study site (CFO1) is located 25 km northeast of Lethbridge, Alberta (Fig. 1). Agricultural operations at this site were initiated with the construction of a dairy in 1928, which has the capacity for 150 dairy cattle. A feedlot for beef cattle was added in 1960s along with an earthen manure storage (EMS) facility for storing liquid dairy manure (approx. 4 m deep) and a catch-basin that receives surface water runoff. This feedlot was expanded in the 1980s to the 2000 head capacity it was at the time of this study. There is also a dugout (or slough, a shallow wetland) on site that receives local runoff and an irrigation drainage canal at the southern boundary of the property.
The second study site (CFO4) is located approximately 30 km north of Red Deer, Alberta and 300 km north of 5 CFO1. This dairy and associated EMS (approx. 6 m deep) were constructed in 1995 and the facility had 350 head of dairy cattle at the time of the study. Runoff will drain either to the small dugout in the north-west of the site, or the natural drainage features (ephemeral ponds or a creek approx. 1.5 km east).

Groundwater monitoring wells 10
Groundwater samples were collected from water table wells and piezometers (hereafter both are referred to as wells) installed at both sites (Table 1). At CFO1, groundwater samples were collected from six individual water table wells (DMW1, DMW2, DMW3, DMW4, DMW5, DMW6) and eight sets of nested wells with one well screened at the water table and one well screened 20 m below ground (BG) (DP10-2 and DP10-1, DMW10 and DP11-10b, . Wells DP10-2 and DP10-1 were located directly adjacent to the EMS on the hydraulically downgradient side. At CFO4, groundwater samples were collected from eight water table wells (BC1, BC2, BC3, BC4, BC5, BMW1, BMW3, BMW7) and four sets of nested wells, with wells screened across the water table and at 15 m BG. Two of these nests were located adjacent to the EMS (BMW2 and BP10-15e, BMW4 and BP10-15w) and two were hydraulically downgradient of the EMS (BMW5 and BP5-20 15,. Groundwater samples were collected for ion analysis (Cland N-species) quarterly between April 2010 and August 2015. All water samples were collected using a bailer after purging (1-3 casing volumes) and stored at ≤ 4 °C prior to analysis. Samples for δ 15 NNO3 and δ 18 ONO3 were collected from wells at CFO1 on 1 January 2013 and 1 May 2013. Samples for δ 15 NNO3 and δ 18 ONO3 at CFO4 were collected on 27 October 2014. Wells were purged 25 prior to sample collection (1-3 casing volumes), and samples filtered into high-density polyethylene (HDPE) bottles in the field and frozen until analysis.
Hydraulic heads in monitoring wells were determined using manual measurements (approximately monthly, [2010][2011][2012][2013][2014][2015]. Hydraulic head response tests were conducted on the majority of the wells at the sites to determine hydraulic conductivity (K) of the formation media surrounding the intake zone. These tests were either a slug test 30 (water level decline after water addition), or bail test (water level recovery after water removal) depending on the location of the water table within the well at the time of testing. K was determined from hydraulic the head responses using the method of Hvorslev (1951).  to depths of up to 15 m below surface and distances of up to 100 m from the EMS between wells DMW3 and DP11-14.
Continuous core samples were retrieved using a hollow stem auger (1.5-m core lengths) with 0.3-m sub-samples collected at approximately 1-m intervals ensuring that visually consistent lithology could be sampled. Core samples for Clwere stored in Ziploc TM bags and kept cool until analysis. Core samples for N-species analysis 5 were stored in Ziploc bags filled with an atmosphere of argon (99.9% Ar) to minimize oxidation and kept cool until analysis. Subsamples of each core (250-300 g) were placed under 50 MPa pressure in a Carver Series NE mechanical press with a 0.5-μm filter placed at the base of the squeezing chamber, which was placed within an Ar atmosphere to minimize oxidation. A syringe was attached to the base of the apparatus and 15 mL of filtered pore water were collected for analyses within 3.5 to 6.0 h (Hendry et al., 2013). 10

Liquid manure storages
Samples of liquid manure slurry were collected directly from the EMS at both sites and the catch basin (containing local runoff from the feedlot) at CFO1 using a pipe and plunger apparatus to sample from approximately 0.5 m below the surface. The slurry collected was subsequently filtered (0.45 μm) to separate the liquid and solid components. The water filtered from samples collected from the EMS or catch basin is hereafter referred to as 15 manure filtrate.

Laboratory analysis
Groundwater samples from wells were analysed by Alberta Agriculture and Forestry (Lethbridge, Alberta).
Pore-water samples squeezed from continuous core were analysed at the University of Saskatchewan (Saskatoon, Canada) for Cl -, NO3-N, and NO2-N using a Dionex IC25 ion chromatograph (IC) coupled to a Dionex As50 30 autosampler (EPA Method 300.1, accuracy and precision of 5.0%) (Hautman and Munch, 1997). Ammonia as N (NH3-N) was measured by Exova Laboratories using the automated phenate method (APHA Standard 4500-NH3 G, detection limit of 0.025 mg L -1 , accuracy of 2% of the measured concentration, and a precision of 5% of the measured concentration). δ 15 NNO3 and δ 18 ONO3 in groundwater samples (from wells and pore water from continuous core) and manure filtrate 35 were measured at the University of Calgary (Calgary, Alberta) using the denitrifier method (Sigman et al., 2001) with an accuracy and precision of 0.3‰ for δ 15 NNO3 and 0.3‰ for δ 18 ONO3. Groundwater samples collected for NO3isotope analysis in January 2013 were also analyzed for NO3-N by the University of Calgary (denitrifier technique, Delta+XL).

Quantification of denitrification based on δ 15 NNO3 and δ 18 ONO3
Nitrate in groundwater that has undergone denitrification is commonly reported as being identified by enrichment 5 of δ 15 NNO3 and δ 18 ONO3 with a slope of about 0.5 on a cross-plot . However, published studies of denitrification in groundwater report slopes of up to 0.77 (Mengis et al., 1999;Singleton et al., 2007). The relationship between isotopic enrichment of 15 NNO3 and 18 ONO3 and the fraction of NO3-N remaining during denitrification can be described by a Rayleigh equation: where R0 is the initial isotope ratio (relative to the standard) of the NO3 -(δ 18 ONO3 or δ 15 NNO3), R is the isotopic ratio when fraction fd of NO3remains, and β is the kinetic fractionation factor (> 1) Otero et al., 2009;Xue et al., 2009). Kinetic fraction effects are commonly also expressed as the enrichment factor, ε = 1 1000( −1) . In the case of a constant enrichment factor, fd can be calculated from measured δ 15 NNO3 (or δ 18 ONO3), if the initial δ 15 NNO3 (δ 15 N0) is known; 15 The fraction of NO3-N removed from groundwater through denitrification is then given by (1-fd). The concentration of NO3-N that would have been measured if mixing was the only attenuation mechanism (NO3-Nmix) can also be calculated by dividing the measured concentration by fd.
A sub-set of 20 samples with isotopic values of NO3indicative of denitrification were identified, and for each of 20 these samples fd (mean and standard deviation) was calculated from Eq. (2) using a Monte Carlo approach with 500 realizations.). The distribution of ε values was defined based on measured data. If the initial δ 15 NNO3 is known, ε for δ 15 NNO3 (ε15N) can be determined from the slope of the linear regression line on a plot of ln(fd) vs. δ 15 NNO3 . If the initial δ 15 NNO3 and fd are not known, as is the case here, ε15N can be determined from the slope of the regression line on a plot of ln(NO3-N) vs. δ 15 NNO3, which will be the same as on a plot of ln(fd) 25 vs. δ 15 NNO3. In-situ variations in temperature and reaction rates may affect the enrichment factor (Kendall and Aravena, 2000) and this was accounted for by allowing for variation in ε15N within the Monte Carlo analysis. The enrichment factor for δ 18 ONO3 (ε18O) was calculated by multiplying the δ 15 NNO3 by a linear coefficient of proportionality determined for each CFO from the slope of the denitrification trend on an isotope cross-plot (see Section 3.2). 30 For each realization, initial isotopic values (δ 15 N0 and δ 18 O0) were determined by Solver such that the difference between fd calculated from δ 15 NNO3 and δ 18 ONO3 was minimized (<1% difference). The ranges of δ 15 N0 and δ 18 O0 were limited based on measured data and literature values (see 3.2). This approach neglects the effect of mixing of groundwater with differing isotopic values, and is valid if the concentration of NO3in the source is much greater than background concentrations such that the isotopic composition of NO3is dominated by the 35 agriculturally derived end-member.

Quantification of mixing and initial concentrations of Cland NO3-N
A binary mixing model that also accounts for decreasing NO3-N concentrations in response to denitrification was used to quantify NO3attenuation by mixing and estimate the initial concentrations of Cland NO3-N. The measured concentration of Clwas assumed to be a function of two end-member mixing, described by where Cl is the measured concentration of Clin the groundwater sample, Cli is the concentration of Clat the initial point of entry of the agriculturally derived NO3to the groundwater system, Clb is the concentration of Clin the background ambient groundwater, and fm is the fraction of water in the sample from the source of agriculturally derived Cl -(and NO3 -) remaining in the mixture.
The concentration of NO3-N was also assumed to be a function of two end-member mixing but with an additional 10 coefficient, fd (the fraction of NO3-N remaining after denitrification), applied to account for denitrification. The measured NO3-N concentration was thus described by where NO3-N is the concentration of NO3-N measured in the groundwater sample, NO3-Ni is the concentration of NO3-N in the source of agriculturally derived NO3at the initial point of entry to the groundwater system, and 15 NO3-Nb is the concentration of NO3-N in the background ambient groundwater. This mixing calculation was only conducted on samples for which NO3dominated total-N (NH3-N <10% of NO3-N) so that nitrification of NH3 could be neglected.
If Cli is much greater than Clb and NO3-Ni is much greater than NO3-Nb, then fm is insensitive to background concentrations and these terms can be neglected (see 4.2 for further discussion of this assumption). In this case, 20 Eqs. (3) and (4) reduce to Solving Eq. (6) for fm and substituting into Eq. (5) yields Thus, for each groundwater sample, the ratio of NO3-N/Clat the initial point of entry of the agriculturally derived NO3to the groundwater system ( 3i i ) can be simply calculated using measured concentrations, and fd estimated from NO3isotope data. This provides a relatively simple method to identify agriculturally derived NO3from different sources (e.g., EMS vs. manure piles) if they have different NO3-N/Clratios. Estimated Cli and NO3-Ni are reported as the mid-range value with uncertainty described by the minimum and maximum values. 30 These initial concentrations are at the water table for top-down inputs, or at the saturated point of contact between the EMS and the aquifer for leakage from the EMS. This analysis assumes that a sampled water parcel consists of water with agriculturally derived NO3that entered the aquifer from one source at one point in time and space and has since mixed with natural ambient groundwater. Any NO3produced during nitrification after the anthropogenic source water enters the aquifer is implicitly included in NO3-Ni. The error in 3i NO3-Ni and Cli were defined by measured concentrations (e.g., if i = , fm=1). Maximum values of NO3-Ni and Cli were defined based on measured concentrations of NO3-N and Clin groundwater and manure filtrate (NO3-N ≤ 150 mg L -1 and Cl -≤ 1300 mg L -1 ; see Section 3.2). These maximum values of NO3-Ni and Cli correspond to the minimum fm. The value of fd was assumed to be the mean fd estimated from NO3isotopes using Eq. (2), and 3i i was required to be within one standard deviation of the estimate from Eq. (7). 5 The resulting estimates of fm are reported as the mid-range, with uncertainty described by the minimum and maximum values. Larger values of fm indicate less mixing (a shorter path for advection-dispersion) and suggest a source close to the well. Smaller values of fm indicate extensive mixing (a longer path for advection-dispersion) and suggest a source further away from the well. The relative contributions of mixing and denitrification to NO3attenuation at each site were evaluated by comparing fm and fd for each sample. This analysis was conducted using 10 isotope values from the samples collected on 1 May 2013 at CFO1, which were combined with the Cland NO3-N data from 6 June 2013. At CFO4, results from stable isotopes collected on 27 October 2014 were combined with Cland NO3-N data collected on 7 October 2014.

CFO1
The geology at CFO1 consists of clay and clay-till interspersed with sand layers of varying thickness to the maximum depth of investigation (20 m BG, bedrock not encountered). Hydraulic conductivities (K) calculated from slug tests on wells ranged from 1.2×10 -7 to 4.2×10 -5 m-s -1 (n=10) for sand, 1.1×10 -8 to 2.8×10 -8 m s -1 (n=2) for clay-till, and 1.6×10 -9 to 3.0×10 -7 m s -1 (n=8) for clay. Depth to the water table throughout the study site ranged 20 from 0.5 m at DMW14 to 3.8 m at DMW11. Seasonal water table variations were about 0.5 m with no obvious change in the annual average during the 6-year measurement period. Water table elevation was highest at DMW10 and DMW1 on the west side of the site and lowest at DMW11 on the northeast side of the site (see Supplementary Material). Measured heads indicate groundwater flow from the vicinity of the EMS to the northeast and southeast.
Mean horizontal hydraulic gradients at the water table ranged from 4.4×10 -3 to 1.4×10 -2 m m -1 . Vertical gradients 25 were predominantly downward in the upper 20 m of the profile (mean gradients ranging from 1.8×10 -3 to 0.18 m m -1 ), with the exception of DMW11 where the vertical gradient was upward (mean gradient -2.8×10 -2 m m -1 ).
Using the geometric mean K for the sand (5.0 x 10 -6 m s -1 ) and a lateral head gradient of 1.4×10 -2 m m -1 yields a specific discharge (Darcy flux, q) of 2.2 m y -1 . Assuming an effective porosity of 0.3 (Rodvang et al., 1998), the average linear velocity ( ̅ ) is 7.4 m y -1 . This suggests that, in the absence of attenuation by mixing or 30 denitrification, agriculturally derived NO3could have been transported through the groundwater system by advection about 400 m from the EMS since 1960 and 630 m since 1930.

CFO4
The geology at CFO4 consists of about 5 m of clay (with minor till) underlain by sandstone, to the maximum depth investigated (20 m BG). Hydraulic conductivities measured using slug tests on wells were 1.0×10 -8 to the sandstone) and 1.0×10 -5 to 2.9×10 -5 m s -1 (n=4) for the sandstone. The depth to water table ranged from 1.0 to 3.4 m, increasing from west to east across the study site. Seasonal water table variations were on the order of 1.5 m with water table declines on the order of 0.3 m y -1 . The horizontal hydraulic gradient was consistently from west to east, with a mean gradient at the water table of 3.9×10 -3 m m -1 between BC2 and BMW2 and 4.3×10 -3 m m -1 between BMW2 and BMW7. Vertical hydraulic gradients were 4.2×10 -2 to 4.6×10 -2 m m -1 downward. Using 5 the geometric mean K for the site (2.9×10 -5 m s -1 ) and a lateral head gradient of 4.3×10 -3 m m -1 yields a q of 0.4 m y -1 . Assuming an effective porosity of 0.3 yields a ̅ of 1.3 m y -1 . These values suggest that, in the absence of attenuation by mixing or denitrification, anthropogenic NO3could have been transported through the groundwater systems about 10 m by advection between 1995 and the time of sampling.

Values and evolution of stable isotopes of nitrate 10
The range of isotopic values of NO3in groundwater was similar at both sites ( Fig. 2). At CFO1, δ 18 ONO3 ranged from -5.9 to 20.1‰ and δ 15 NNO3 from -5.2 to 61.0‰. At CFO4, δ 18 ONO3 ranged from -1.9 to 31.6‰ and δ 15 NNO3 from -1.3 to 70.5‰. The isotopic values of δ 18 ONO3 in groundwater are commonly assumed to be derived from a mix of a 1/3 atmospheric-derived oxygen (+23.5‰) and 2/3 water-derived oxygen (Xue et al., 2009). Given the average δ 18 OH2O for both sites (-16‰, see Supplementary Material), a 1/3 atmospheric 2/3 groundwater mix would 15 result in a δ 18 ONO3 of -3.7‰. Manure filtrate from the EMS at CFO1 had δ 15 NNO3 ranging from 0.4 to 5.0‰ and δ 18 ONO3 ranging from 7.1 to 19.0‰. A curve showing the co-evolution of δ 18 ONO3 (mixing of atmospheric δ 18 O with groundwater-derived δ 18 O) and δ 15 NNO3 (Rayleigh distillation, β = 1.005) during nitrification is shown in Fig.   2. Isotopic values in DMW3, where direct leakage from the EMS was evident, are consistent with partial nitrification following this trend of isotopic evolution (δ 18 ONO3 of -1.2‰ and δ 15 NNO3 of 7.8‰). 20 At both sites, co-enrichment of δ 18 ONO3 and δ 15 NNO3 characteristic of denitrification was evident in some samples (slopes of 0.42 and 0.72 in Fig. 2a). At CFO1, this includes samples from DP10-2, DMW5, DMW11, DMW12, DP11-12b, and DMW13 (and associated core) and some pore water from cores DC15-22 and DC15-23. These samples had NO3-N concentrations of 0.6 to 23.7 mg L -1 , δ 18 ONO3 ranging from 4.8 to 20.6‰, and δ 15 NNO3 ranging from 22.9 to 61.3‰. At CFO4, samples exhibiting evidence of denitrification were from BMW2, BMW5, BMW6, 25 BMW7, and BC4. These samples had NO3-N concentrations ranging from 0.4 to 35.1 mg L -1 , δ 18 ONO3 ranging from 1.6 to 22.1‰, and δ 15 NNO3 ranging from 20.9 to 70.1‰. Although the isotopic values of DMW5 suggest enrichment by denitrification, the data plot away from the rest of the CFO1 data and close to the denitrification trend at CFO4 (Fig. 2), suggesting these samples were affected by some other process (possibly mixing or nitrification); therefore, the fraction of NO3-N remaining in this well was not calculated. Also, well DMW3, which 30 clearly receives leakage from the EMS, did not contain substantial NO3-N and so fd was not calculated.
In the Monte Carlo analysis the potential range of original isotopic values of the NO3source prior to denitrification (δ 15 N0 and δ 18 O0) varied from 5 to 27‰ for δ 15 NNO3 and from -2 to 7‰ for δ 18 ONO3 based on isotopic values measured during this study (Fig. 2a). These values are consistent with literature values for manure-sourced NO3 -, which report δ 15 NNO3 ranging from 5 to 25‰ and δ 18 ONO3 ranging from -5 to 5‰ (Wassenaar, 1995;Wassenaar et 35 al., 2006;Singleton et al., 2007;McCallum et al., 2008;Baily et al., 2011). ε15N was defined by a normal distribution with a mean of -10‰ and standard deviation of 2.5‰ (Fig. 2b). At CFO1, the coefficient of proportionality between the enrichment factor of δ 15 NNO3 and δ 18 ONO3 was described by a normal distribution with mean of 0.72 and standard deviation of 0.05. At CFO4, the coefficient of proportionality was also described by a normal distribution with a mean of 0.42 and standard deviation of 0.035 (see Fig. 2a). These enrichment factors are consistent with values from denitrification studies that report ε15N ranging from -4.0 to -30.0‰ and ε18O ranging from -1.9 to -8.9‰ (Vogel et al., 1981;Mariotti et al., 1988;Spalding and Parrott, 1994;Mengis et al., 1999;Pauwels et al., 2000;Otero et al., 2009).

Distribution and sources of agricultural nitrate in groundwater 5
At both sites TN concentrations in filtrate from the EMS and catch-basin were generally an order of magnitude larger than concentrations in groundwater ( Table 2). The one exception is well DMW3 at CFO1 which intercepted direct leakage from the EMS (see 3.3.1 for further discussion of this well). The dominant form of N differed between manure filtrate and groundwater. In the EMS filtrate, N was predominately organic-N (TON up to 71%) or NH3-N (up to 90%), with NOx-N <0.1% of TN. In the catch-basin at CFO1 TON was >99% of TN. In 10 groundwater TN concentrations ranged from <0.25 to 84.6 mg L -1 , and this N was predominantly NO3 -(again, with the exception of DMW3).

CFO1
Agriculturally derived NO3was generally restricted to the upper 20 m (or less) at CFO1 (NO3-N ≤ 0.2 mg L -1 and Cl -≤ 57 mg L -1 in seven wells screened at 20 m). The one exception was DP11-12b, which had up to 4.1 mg L -1 15 of NO3-N. The southeast portion of the site also does not appear to have been significantly contaminated by agriculturally derived NO3 -, with NO3-N concentrations < 1 mg L -1 in five water table wells (DMW4, DMW6, DMW14, DMW15, DMW16). In DMW6, Cland TN concentrations were elevated (see Supplementary Material) but NO3-N concentrations were < 2 mg L -1 . Collectively, these data suggest the catch basin is not a significant source of NO3to the groundwater at this site. 20 Leakage of manure slurry from the EMS at CFO1 is clearly indicated by the data from DMW3, which feature the highest concentrations of TN in groundwater (up to 548 mg L -1 ) and elevated Cl -, HCO3 -, and DOC in concentrations similar to EMS manure filtrate (see Supplementary Material). Nevertheless, NO3-N concentrations in this well were consistently low (1.1 ± 2.7 mg L -1 , n=22). The potential for nitrification in the vicinity of this well is indicated by NO2-N production (2.7 ± 8.3 mg L -1 , n=22). However, the data demonstrate that only a small 25 proportion of the NH3-N in DMW3 (373.4 ± 79.4 mg L -1 , n=22) could have been converted to NO3within the subsurface (NO3-N in groundwater ≤ 66 mg L -1 ). Further work is required to assess the importance of cation exchange as an attenuation mechanism for direct leakage from the EMS at this site.
Given the evidence of partial nitrification in DMW3 (and low NO3-N concentrations), the NO3-N/Clratio of contamination from the EMS was assumed to be best represented by DP10-2, which is located directly downgradient of the EMS. Data for this well indicate values of NO3-N/Clpredominantly ranging from 0.1 to 0.3 with NO3-Ni/Cli estimated at 0.3 ± 0.13 (Fig. 4). 35 The maximum NO3-N concentration in groundwater at CFO1 (66.4 mg L -1 ) was measured in core sample DC15-23 (clay at 2 m bgl, 7 m hydraulically downgradient of DMW3). Pore water extracted from the unsaturated zone (sand) at the top of this core profile contained 865 mg L -1 of NO3-N and had a NO3-N/Clratio of 1.04, consistent with the ratio of 0.95 in the core sample. Given this consistency, and that NO3-N concentrations in the well immediately up-gradient were low (DMW3), the NO3-N in this core sample was most likely introduced into the groundwater system by vertical infiltration or diffusion from above. In contrast, elevated NO3-N (up to 21.1 mg L -1 ) within the sand between 6 and 12 m depth in this core had NO3-N/Clratios consistent with an EMS source (0.07 to 0.31). Stable isotope values in pore water from this sand layer do not indicate substantial 5 denitrification (δ 18 O ≤ 5.9‰, δ 15 N ≤ 16.7‰), suggesting these ratios will be similar to the initial ratios at the point of entry to the groundwater system.
In DMW13 (33 m downgradient from DP10-2) the ratio of NO3-Ni/Cli was 0.75 ± 0.29, similar to the NO3-N/Clratio in DC15-23 at 2 m (0.95), which is interpreted as reflecting a top-down source. The NO3in DMW13 is therefore unlikely to be sourced solely from leakage from the EMS, and could be sourced from the adjacent dairy 10 pens or a temporary manure pile that was observed adjacent to this well during core collection in 2015 (or a combination of EMS and top-down sources).
In DMW12 the NO3-Ni/Cli ratio was not inconsistent with an EMS source, but the hydraulic gradient between DMW2 and DMW12 is negligible, indicating a lack of driving force for advective transport from the EMS towards DMW12. This is also the case for well DMW1, which is up-gradient of the EMS but had elevated NO3-N 15 concentrations (6.5 ± 3.6, n=18). The source of nitrate in these wells is therefore unlikely to be related to leakage from the EMS, but alternative sources (i.e., nearby temporary manure piles) are not known.
Well DMW11, 470 m from the EMS, had consistently low NO3-N/Clratios (< 0.05) similar to DP10-2, but estimates of Cli were three-fold higher than Cli for DP10-2 ( Fig. 4b). NO3-Ni and Cli estimated for DMW11 were consistent with measured values in that well, indicating a local top-down source. Well DMW11 is located 20 hydraulically downgradient of feedlot pens and adjacent to a solid manure storage area, in a local topographic low. Elevated NO3-N in this well is therefore interpreted to be from surface runoff and top-down infiltration, rather than lateral advection from the EMS.

CFO4
At CFO4, measured data indicate that effects from agricultural operations on NO3concentrations in groundwater 25 are restricted to the upper 15 m of the subsurface. NO3-N concentrations in wells screened at 15 m depth were < 0.5 mg L -1 , with the exception of one sample from BP10-15w (May 2012) with 4.3 mg L -1 of NO3-N. Water table wells in the west and north of the study site (BC1, BC2, and BC3) also indicate negligible impacts of agricultural operations, with Cl -< 10 mg L -1 and NO3-N < 0.1 mg L -1 .
Given that the estimated subsurface travel distance during operations at this site is 10 m, agriculturally derived NO3in other wells not immediately adjacent to the EMS is unlikely to be related to leakage from the EMS. Wells BMW5 and BMW7 are 60 and 140 m hydraulically downgradient from the EMS, respectively. NO3-Ni/Cli ratios in these wells were not inconsistent with BMW2 (i.e., the range of values overlap), but given the distance from the EMS the source of NO3-N in these wells is most likely the adjacent dairy pens. Concentrations of NO3-N > 10 mg L -1 were also measured in BC4, which is located 95 m hydraulically upgradient of the EMS. The ratio of NO3-Ni/Cli at BC4 was the highest at CFO4 (0.6) and did not overlap with BMW2. The NO3in this well is 5 interpreted to have been sourced from an adjacent manure pile, which was observed during the study.

Mechanisms of attenuation of agriculturally derived NO3 -
Attenuation of agriculturally derived NO3in groundwater is dominated by denitrification at both CFO1 and CFO4, with estimates of fm consistently higher than estimates of fd (Table 3, Fig. 7, Table S10). Calculated fd values indicate that where denitrification was identified, at least half of the NO3-N present at the initial point of 10 entry to the groundwater system has been removed by this attenuation mechanism. Comparison of NO3-Nmix (the concentration of NO3-N that would be measured if mixing was the only attenuation mechanism) with measured concentrations (which reflect attenuation by both mixing and denitrification) suggests that the sample from 20 m depth (DP11-12b) is the only sample that would be below the drinking water guideline if mixing was the only attenuation mechanism (Fig. 8) . 15 At both sites, the stable isotope values of NO3indicate that denitrification proceeds within metres of the source.
At CFO1, calculated fd in well DP10-2 (2 m from the EMS) is 0.52 ± 0.22; at CFO4, fd in well BMW2 (3 m from the EMS) is 0.13 ± 0.06. Denitrification also substantially attenuated NO3-N concentrations in wells where the source is not the EMS but instead is adjacent solid manure piles (e.g., DMW11 at CFO1, BC4 at CFO4). In BMW6 at CFO4, denitrification completely attenuated the agriculturally derived NO3 -. This well had negligible NO3-N 20 (0.4 ± 0.2 mg L -1 , n=8) and the lowest fd of 0.01. Measured DOC in this well was consistent with other wells at both sites (6.9 ± 1.7 mg L -1 , n=3), suggesting DOC depletion does not limit denitrification at these CFO operations.

Implications for on-farm waste management
Agriculturally derived NO3at these two sites with varying lithology was generally restricted to depths < 20 m, 25 consistent with previous studies at CFOs (Robertson et al., 1996;Rodvang and Simpkins, 2001;Rodvang et al., 2004;Kohn et al., 2016). Attenuation of agriculturally derived NO3in groundwater was a spatially varying combination of mixing and denitrification, with denitrification playing a greater role than mixing at both sites. In the samples for which fd could be determined, denitrification reduced NO3concentrations by at least half and, in some cases, back to background concentrations. Given that the range of source isotopic composition was allowed 30 to vary to its maximum justifiable extent, these quantitative estimates of denitrification based on stable isotopes of NO3are likely to be conservative. Redox conditions within the groundwater system were not able to be determined in this study due to the sampling method used to collect groundwater from wells screened across low-K formations (well bailed dry then sample collected after water level recovery). However, denitrification appears to proceed within metres of the NO3source, suggesting relatively short sub-surface residence times are required 35 and that redox conditions close to the water table are conducive to denitrification reactions .
The substantial role of denitrification within the saturated glacial sediments at these study sites indicates the potential for significant attenuation of agriculturally derived NO3by denitrification in similar groundwater systems across the North American interior and Europe . Denitrification in the unsaturated zone is limited by low water contents and oxic conditions, resulting in substantial stores of NO3in vadose zones (Turkeltaub et al., 2016;Ascott et al., 2017). NO3in water that is removed rapidly from site is 5 also unlikely to be substantially attenuated by denitrification due to oxic conditions and rapid transit times . Therefore, water management focussed on reducing the effects of NO3contamination in similar hydrogeological settings to this study should aim to maximize infiltration into the saturated zone where NO3concentrations can be naturally attenuated, provided that local groundwater isn't used for potable water supply. 10 At both sites there is evidence of elevated NO3due to leakage from the EMS, but the impact appears to be limited to within metres of the EMS. This suggests that saturation within the clay lining of the EMS has limited the development of extensive secondary porosity that would allow rapid water percolation (Baram et al., 2012).
Infiltration of NO3rich water that has passed through temporary solid manure piles and dairy pens has resulted in groundwater NO3-N concentrations as high as those associated with leakage from the EMS (e.g., DMW11, 15 BC4). At CFO4, this is in spite of the presence of clay at surface, reflecting secondary porosity in the upper part of the profile that has led to hydraulic conductivities comparable to sand. This is consistent with the findings of Showers et al. (2008), who investigated sources of NO3at an urbanized dairy farm in North Carolina, USA.

Construction of EMS facilities in Alberta has been regulated under the Agriculture Operation Practices Act since
2002, which requires them to be lined with clay to minimise leakage (Lorenz et al., 2014). On-farm waste 20 management should increasingly focus on minimising temporary manure piles that are in direct contact with the soil to reduce NO3contamination associated with dairy farms and feedlots.

Critique of this approach and applicability at other sites
At both sites, leakage from the EMS had NO3-Ni/Cli of between 0.1 and 0.4, but this alone was not diagnostic of the source. The sources of manure-derived NO3 -(manure piles vs. EMS) are distinguishable based on NO3-Ni/Cli 25 ratios, provided there is also an understanding of the history of each site, local hydrogeology, and potential sources. Calculated fd and fm generally decreased with increasing subsurface residence time and distance from source, providing additional evidence for source attribution. For example, at CFO4, well BMW2, which is adjacent to the EMS, had the highest fm (0.92), indicating the least attenuation of NO3 by mixing and consistent with the EMS being the source of NO3to this well. 30 Calculation of NO3-Ni/Cli assumed that background concentrations could be neglected in the mixing model. At these study sites, background concentrations are likely to be < 20 mg L -1 for Cland < 1 mg L -1 for NO3-N. which depends on measured Cland NO3-N. The largest uncertainties in NO3-Ni and Cli correspond to the lowest measured concentrations (i.e., furthest from the upper limit), with less uncertainty at higher measured concentrations as they approach the maximum values. Temporal variability in NO3-Ni/Cli for each source could 5 not be determined based on the snapshot isotope sampling conducted, but this could be investigated by measuring NO3isotopes in conjunction with NO3-N and Clat multiple times.
Although applicable at these sites, this approach may not be valid at other sites if additional sources of NO3 in groundwater (e.g. fertilizer or nitrification) are significant, or if NO3 concentrations in groundwater are naturally elevated (Hendry et al., 1984). The combination of the approach outlined here with measurement of groundwater 10 age indicators would allow for better constraints on groundwater flow velocities and determination of denitrification rates Katz et al., 2004;McMahon et al., 2004;.

Comparison with isotopic values of NO3in previous studies
Nitrate isotope values in groundwater at the two CFOs studied were generally consistent with previous studies reporting denitrification of manure-derived NO3at dairy farms (Wassenaar, 1995;Wassenaar et al., 2006;15 Singleton et al., 2007;McCallum et al., 2008;Baily et al., 2011). However, the isotopic values of NO3in the manure filtrate from the EMS at CFO1, were not consistent with values for manure-sourced NO3reported in other groundwater studies (Wassenaar, 1995;Wassenaar et al., 2006;Singleton et al., 2007;McCallum et al., 2008a;Baily et al., 2011). This is likely to be because nitrification within the EMS was negligible (NO3-N <0.7 mg L -1 ), such that the isotopic values of NO3-N in the manure filtrate reflect volatilization of NH3 and partial nitrification 20 within the EMS. δ 18 ONO3 values may also have been affected by evaporative enrichment of the δ 18 OH2O being incorporated into NO3 - (Showers et al., 2008).
A number of groundwater samples collected during this study had relatively enriched δ 18 ONO3 (> 15 ‰) with depleted δ 15 NNO3 (< 15‰). Some of these isotopic values are within the range previously reported for NO3derived from inorganic fertilizer (δ 15 NNO3 from -3 to 3‰ and δ 18 ONO3 from -5 to 25‰), with the δ 18 ONO3 depending on 25 whether the NO3is from NH4 + or NO3in the fertilizer (Mengis et al., 2001;Wassenaar et al., 2006;Xue et al., 2009). To the best of our knowledge, however, no inorganic fertilizers have been applied at these study sites.
Another potential source is NO3derived from soil organic N, but this should have δ 15 NNO3 values of 0 to 10‰ and δ 18 ONO3 values of -10 to 15‰ Mayer et al., 2001;Mengis et al., 2001;Xue et al., 2009;Baily et al., 2011). Incomplete nitrification of NH4 + can result in δ 15 NNO3 lower than the manure source (Choi et 30 al., 2003), but as there was no measurable NH3-N in these samples this is also unlikely. These isotope values may reflect the influence of NO3from precipitation, which usually has values ranging from -5 to 5‰ for δ 15 NNO3 and 40 to 60‰ for δ 18 ONO3, and has been reported to dominate NO3isotope values of groundwater under forested landscapes . Alternatively, they may be affected by microbial immobilization and subsequent mineralization and nitrification, which can mask the source δ 18 ONO3 in aquifers with long residence times (Mengis

Conclusions
A mixing model constrained by quantitative estimates of denitrification from isotopes substantially improved our understanding of nitrate contamination at these sites. This novel approach has the potential to be widely applied as a tool for monitoring and assessment of groundwater in complex agricultural settings. NO3-N concentrations in excess of the drinking water guideline were measured at both sites, with sources including manure piles, pens 5 and the EMS. Even though these sites are dominated by clay-rich glacial sediments, the input of NO3to groundwater from temporary manure piles and pens resulted in comparable (or greater) NO3-N concentrations than leakage from the EMS. This is attributed to the development of secondary porosity within unsaturated clays.
Nitrate attenuation at both sites is dominated by denitrification, which is evident even in wells directly adjacent to the NO3source. In the wells for which denitrification was identified, concentrations of agriculturally-derived 10 NO3had been reduced by at least half and, in some wells, completely. In the absence of denitrification all but one of these wells would have had NO3-N concentrations above the drinking water guideline.
These results indicate that infiltration to groundwater systems in glacial sediments where NO3can be naturally attenuated is likely to be preferable to off-farm export via runoff or drainage networks, provided that local groundwater isn't a potable water source. On-farm management of manure waste at similar operations should 15 increasingly focus on limiting manure piles that are in direct contact with the soil to limit NO3contamination of groundwater.

Figure 6 (a) Estimated NO3-Ni/Cli ratios (mean and st. dev.) in water table wells with evidence of denitrification at CFO4, plotted with distance from earthen manure storage (EMS), where dashed lines are upper and lower bounds of BMW2 (EMS source) and values are maximum measured NO3-N (mg L -1 ). (b) Estimated concentrations of NO3-Ni and
Cli at CFO1 (mid-range, error bars are max. and min. values).

Figure 7 Relative contributions to NO3attenuation by mixing and denitrification, as indicated by estimated fm and fd at (a) CFO1 and (b) CFO4, for groundwater samples with denitrification indicated by stable isotope values of NO3 -.
5

Figure 8 Measured concentrations of NO3-N (blue circles -attenuation by mixing and denitrification) and NO3-Nmix (red triangles -attenuation by mixing only) vs mid-range estimate of NO3-Ni at a) CFO1 and b) CFO4. Dashed lines
10 are drinking water guideline (10 mg L -1 of NO3-N).

Figure 9 Effect of neglecting background concentrations (Clb or NO3-Nb) in the mixing model on calculated fm over the range of values in this study.
contamination of groundwater, lakes, and streams. Understanding the sources and fate of nitrate in groundwater systems in glacial sediments, which underlie many agricultural operations, is critical for managing impacts of human food production on the environment. Elevated NO3concentrations in groundwater can be naturally attenuated through mixing or denitrification. Here we use isotopic enrichment of the stable isotope values of NO3to quantify the amount of denitrification in groundwater at two confined feeding operations overlying glacial 15 sediments in Alberta, Canada. Uncertainty in δ 15 NNO3 and δ 18 ONO3 values of the NO3source source and denitrification enrichment factors are accounted for using a Monte Carlo approach. When denitrification could be quantified, we used these values to constrain a mixing model based on NO3and Clconcentrations. Using this novel approach we were able to reconstruct the initial NO3-N concentration and NO3-N/Clratio at the point of entry to the groundwater system. Manure filtrate had total-nitrogen (TN) of up to 1820 mg L -1 , which was 20 predominantly organic-N and NH3. Groundwater had up to 85 mg L -1 TN, which was predominantly NO3 -. The addition of NO3to the local groundwater system from temporary manure piles and pens equalled or exceeded NO3additions due to leaching from earthen manure storages at these sites. On-farm management of manure waste As such, on-farm management of manure waste sto limit NO3contamination of groundwater should therefore increasingly focus on limiting manure piles in direct contact with the soil, and encourage storage in lined lagoons. 25 Nitrate attenuation at both sites is is attributed to a spatially variable combination of mixing and denitrification, but is dominated by denitrification. On-siteWhere identified, denitrification reduced agriculturally-derived NO3concentrations by at least half and, in some wells, completely. Therefore, iInfiltration to groundwater systems in glacial sediments where NO3can be naturally attenuated is likely preferable to off-farm export via runoff or drainage networks, especially if local groundwater is not used for potable water supply. 30

Introduction
The contamination of soil and groundwater with nitrate from agricultural operations is a global water quality issue that has been extensively documented (Power and Schepers, 1989;Spalding and Exner, 1993;Rodvang and Simpkins, 2001;Galloway et al., 2008;Arauzo, 2017;Ascott et al., 2017). Leaching of nitrate (NO3 -) from animal waste or fertilizers can result in groundwater NO3concentrations that exceed drinking water discharge of high-NO3groundwater, runoff, or drainage can contaminate streams and lakes, resulting in eutrophication and ecosystem decline Kaushal et al., 2011). In saturated groundwater systems with low oxygen concentrations, elevated NO3can be naturally attenuated by microbial denitrification (Wassenaar, 1995;Robertson et al., 1996;Smith et al., 1996;Tesoriero et al., 2000;Singleton et al., 2007).
Concentrations of NO3will also decrease along groundwater flow paths due to attenuation via dilution by 5 hydrodynamic dispersion (referred to hereafter as mixing). Because of these natural attenuation mechanisms, infiltration to groundwater may be preferable to off-site drainage and runoff of nitrate-rich waters. Many agricultural operations are undertaken on fertile soils associated with glacial sediments (Spalding and Exner, 1993;. Understanding the sources and fate of agriculturally derived nitrate in groundwater systems in glacial sediments is therefore critical for managing impacts of human food production on 10 the environment. Identification of the sources and fate of NO3at agricultural operations can be challenging because of spatial and temporal variations in sources (e.g. earthen manure storage, temporary manure piles, or fertilizer) and the complexity heterogeneity inof hydrogeologic systems (Spalding and Exner, 1993;Rodvang et al., 2004;Showers et al., 2008;Kohn et al., 2016). These spatial and temporal variations can result in complex subsurface solute 15 distributions that are difficult to interpret using classical transect studies or numerical groundwater models (Green et al., 2010;Baily et al., 2011).
Groundwater containing significant agriculturally derived NO3also typically has elevated chloride (Cl -) concentrations (Saffigna and Keeney, 1977;Rodvang et al., 2004;Menció et al., 2016). Decreasing NO3-N/Cl -(or NO3 -/Cl -) ratios have been used to define denitrification based on the assumption that NO3is reactive while 20 Clis non-reactive (conservative), such that denitrification results in a decrease in the NO3-N/Clratio (Kimble et al., 1972;Weil et al., 1990;Liu et al., 2006;McCallum et al., 2008). However, NO3-N/Clratios can also change in response to mixing of groundwater with different NO3-N/Clratios or when groundwater sampling traverses hydraulically disconnected formations (Bourke et al., 2015b). If NO3-N/Clratios vary among potential sources and the NO3-N/Clratio at the point of entry to the groundwater system can be reconstructed, this information 25 could be used to show that anthropogenic NO3at different locations within an aquifer is derived from the same or different sources.
The stable isotopes of NO3 -(δ 15 NNO3 and δ 18 ONO3) provide an alternative approach to characterize the source and fate of NO3in groundwater systems. In agricultural areas, multiple sources of NO3are common and could include precipitation, soil NO3 -, inorganic fertilizer, manure, and septic waste (Komor and Anderson, 1993;Liu et al., 30 2006;Pastén-Zapata et al., 2014;Xu et al., 2015). While source identification is theoretically possible using δ 15 NNO3 and δ 18 ONO3 (particularly with a dual-isotope approach), in practice this can be difficult due to geologic heterogeneity, overlapping source values, and the complexity of biologically mediated reactions (Aravena et al., 1993;Wassenaar, 1995;Mengis et al., 2001;Choi et al., 2003;Vavilin and Rytov, 2015;Xu et al., 2015). 35 NO3attenuation by denitrification in groundwater systems can be identified based on the characteristic enrichment of δ 15 NNO3 and δ 18 ONO3. Numerous studies have made qualitative assessments that identified denitrification in groundwater using the stable isotope approach Wassenaar, 1995;Singleton et al., 2007;Baily et al., 2011;Xu et al., 2015). Recently published papers have also used stable isotopic values of NO3and water as the basis for mixing models in agricultural settings (Ji et al., 2017 ;Lentz and Lehersch, 2019). Isotopic fractionation effects can also allow for quantitative assessment of the proportion of substrate that has undergone a given reaction, if enrichment factors and source values are known; as in the case of evpoarative loss of water, for example . To date, there have been very few attempts to quantify denitrification using dual-isotope enrichment, largely due to uncertainty in source values and enrichment factors , Xue et al., 2009. 5 The only published calculations of the fraction of NO3remaining after denitrification the that we are aware of assumed a constant enrichment factor and the same isotopic source values across the field site (Otero et al., 2009).
However, the enrichment factor will vary across a field site in response to reaction rates (Kendall and Aravena 2000), and isotopic values of even the same type of source (e.g. manure) can vary substantially (Xue et al., 2009).
If the varation in source values and enrichment factors can be characterized from measured data then these 10 uncertainties can be accounted for using a Monte Carlo approach (Joerin et al., 2002;Ji et al., 2017), thereby extending the application of the dual-isotope technique to allow for a robust quantitative assessment of denitrification in agricultural settings.
A synthesized analysis of stable isotopes of NO3with additional ionic tracers can further improve the assessment of NO3attenuation mechanisms and sources of NO3in agricultural settings (Showers et al., 2008;Vitòria et al., 15 2008;Xue et al., 2009;Xu et al., 2015;Ji et al., 2017). We hypothesise that if the amount of denitrification can be quantified based on δ 15 NNO3 and δ 18 ONO3, then this estimate of the fraction of NO3-N removed through denitrification can be used to constrain a mixing model based on NO3-N and Clconcentrations. This novel approach allows for the ratio of NO3-N/Clat the point of entry to the groundwater system to be reconstructed from measured NO3and Clconcentrations (see Section 2.3). Where the NO3-N/Clratio varies between sources, 20 this ratio can then be used to assess the source of the NO3in groundwater (e.g. temporary manure piles or feeding pens). These data can also then be used to estimate the initial concentrations of NO3and Clat the point of entry to the groundwater system and quantify attenuation by mixing.
In this study, we present the application of this approach at two confined feeding operations (CFOs) in Alberta, Canada, with differing lithologies and durations of operation (Fig. 1). Concentrations of Cland nitrogen species 25 (N-species) and the stable isotopes of NO3were measured in groundwater samples collected from monitoring wells and continuous soil cores, as well as manure filtrate at both sites. These data were interpreted to (1) assess the extent of agriculturally derived NO3in groundwater, (2) identify sources and initial concentrations of NO3at the point of entry to the groundwater system, and (3) assess the dominantmixing and denitrification as attenuation mechanisms controlling subsurface NO3distributions at these sites. 30

Experimental sites
This study was conducted using data from two of the five sites investigated by Alberta Agriculture and Forestry during an assessment of the impacts of livestock manure on groundwater quality (Lorenz et al., 2014). To the best of our knowledge (including discussions with farm operators) fertilizers have not been applied at either of these 35 sites. As such, manure waste from livestock is assumed to be the sole source of agricultural nitrogen (N) and elevated NO3concentrations in groundwater at these sites.
The first study site (CFO1) is located 25 km northeast of Lethbridge, Alberta (Fig. 1). Agricultural operations at this site were initiated with the construction of a dairy in 1928, which has the capacity for , with the capacity for the150 dairy cattle since the 1960s. A feedlot for beef cattle was added in 1960s along with an earthen manure storage (EMS) facility for storing liquid dairy manure (approx. 4 m deep) and a catch-basin that receives surface water runoff. This feedlot was expanded in the 1980s to the 2000 head capacity it was at the time of this study. 5 There is also a dugout (or slough, a shallow wetland) on site that receives local runoff and an irrigation drainage canal at the southern boundary of the property.
The second study site (CFO4) is located approximately 30 km north of Red Deer, Alberta and 300 km north of CFO1. This dairy and associated EMS (approx. 6 m deep) were constructed in 1995 and the facility had 350 head of dairy cattle at the time of the study. Runoff will drain either to the small dugout in the north-west of the site, or 10 the natural drainage features (ephemeral ponds or a creek approx. 1.5 km east).

Groundwater monitoring wells
Groundwater samples were collected from water table wells and piezometers (hereafter both are referred to as wells) installed at both sites (Table 1) Groundwater samples were collected for ion analysis (Cland N-species) quarterly between April 2010 and 25 August 2015. All water samples were collected using a bailer after purging (1-3 casing volumes) and stored at ≤ 4 °C prior to analysis. Samples for δ 15 NNO3 and δ 18 ONO3 were collected from wells at CFO1 on 1 January 2013 and 1 May 2013. Samples for δ 15 NNO3 and δ 18 ONO3 at CFO4 were collected on 27 October 2014. Wells were purged prior to sample collection (1-3 casing volumes), and samples filtered into high-density polyethylene (HDPE) bottles in the field and frozen until analysis. 30 Hydraulic heads in monitoring wells were determined using manual measurements (approximately monthly, [2010][2011][2012][2013][2014][2015]. Rising Hydraulic head head response tests (slug or bail tests) were conducted on the majority of the wells at the sites to determine hydraulic conductivity (K) of the formation media surrounding the intake zone on the majority of the wells at the sites. These tests were either a slug test (water level decline after water addition), or bail test (water level recovery after water removal) depending on the location of the water table within the well

Continuous core
Continuous core was collected at CFO1 immediately adjacent to well DP11-13b on 1 May 2013 (Fig. 1).
Continuous core samples were retrieved using a hollow stem auger (1.5-m core lengths) with 0.3-m sub-samples collected at approximately 1-m intervals ensuring that visually consistent lithology could be sampled. Core samples for Clwere stored in Ziploc TM bags and kept cool until analysis. Core samples for N-species analysis 10 were stored in Ziploc bags filled with an atmosphere of argon (99.9% Ar) to minimize oxidation and kept cool until analysis. Subsamples of each core (250-300 g) were placed under 50 MPa pressure in a Carver Series NE mechanical press with a 0.5-μm filter placed at the base of the squeezing chamber, which was placed within an Ar atmosphere to minimize oxidation. A syringe was attached to the base of the apparatus and 15 mL of filtered pore water were collected for analyses within 3.5 to 6.0 h (Hendry et al., 2013). 15

Liquid manure storages
Samples of liquid manure slurry were collected directly from the EMS at both sites and the catch basin (containing local runoff from the feedlot) at CFO1 using a pipe and plunger apparatus to sample from approximately 0.5 m below the surface. The slurry collected was subsequently filtered (0.45 μm) to separate the liquid and solid components. The water filtered from samples collected from the EMS or catch basin is hereafter referred to as 20 manure filtrate.

Laboratory analysis
For gGroundwater samples from wells were analysed by Alberta Agriculture and Forestry (Lethbridge, Alberta). and manure filtrate, cConcentrations of Clwere determined using potentiometric titration of H2O, with a detection limit of 5.0 mg L -1 and accuracy of 5% (APHA 4500-Cl -D). Concentrations of NH3 as N (NH3-N), NO3as , and NO2as N (NO2-N) in groundwater samples from wells and manure filtrate were measured by air- Pore-water samples squeezed from continuous core were analyzedanalysed at the University of Saskatchewan 35 (Saskatoon,Canada) for Cl -, NO3-N, and NO2-N using a Dionex IC25 ion chromatograph (IC) coupled to a Dionex As50 autosampler (EPA Method 300.1, accuracy and precision of 5.0%) (Hautman and Munch, 1997). Ammonia as N (NH3-N) was measured by Exova Laboratories using the automated phenate method (APHA Standard 4500-NH3 G, detection limit of 0.025 mg L -1 , accuracy of 2% of the measured concentration, and a precision of 5% of the measured concentration). δ 15 NNO3 and δ 18 ONO3 in groundwater samples (from wells and pore water from continuous core) and manure filtrate were measured at the University of Calgary (Calgary, Alberta) using the denitrifier method (Sigman et al., 2001) with an accuracy and precision of 0.3‰ for δ 15 NNO3 and 0.3‰ for δ 18 ONO3. Groundwater samples collected for 5 NO3isotope analysis in January 2013 were also analyzed for NO3-N by the University of Calgary (denitrifier technique, Delta+XL).

Quantification of denitrification based on δ 15 NNO3 and δ 18 ONO3
Nitrate in groundwater that has undergone denitrification is commonly reported as being identified by enrichment 10 of δ 15 NNO3 and δ 18 ONO3 with a slope of about 0.5 on a cross-plot . However, published studies of denitrification in groundwater report slopes of up to 0.77 (Mengis et al., 1999;Singleton et al., 2007). The relationship between isotopic enrichment of δ 15 NNO3 and δ 18 ONO3 and the fraction of NO3-N remaining during denitrification can be described by a Rayleigh equation: where R0 is the initial isotope ratio (relative to the standard) of the NO3 -(δ 18 ONO3 or δ 15 NNO3), R is the isotopic ratio when fraction fd of NO3remains, and β is the kinetic fractionation factor (> 1) Otero et al., 2009;Xue et al., 2009). Kinetic fraction effects are commonly also expressed as the enrichment factor, ε = and tThe fraction of NO3-N removed from groundwater through denitrification is then given by (1-fd). The concentration of NO3-N that would have been measured if mixing was the only attenuation mechanism (NO3-Nmix) can also be calculated by dividing the measured concentration by fd. 25 A sub-set of 20 samples with isotopic values of NO3indicative of denitrification were identified, and for each of these samples fd (mean and standard deviation) was calculated from Eq. (2) using a Monte Carlo approach with 500 realizations. The value of R was given by the measured isotopic ratio for each sample (δ 18 ONO3 or δ 15 NNO3).
R0 was allowed to vary randomly within a range of values determined from measured data and literature values.
The distribution of ε values was defined based on measured data. If the initial δ 15 NNO3 is known, ε for δ 15 NNO3 30 (ε15N) can be determined from the slope of the linear regression line on a plot of ln(fd) vs. δ 15 NNO3 . If the initial δ 15 NNO3 and fd are not known, as is the case here, ε15N can be determined from the slope of the regression line on a plot of ln(NO3-N) vs. δ 15 NNO3, which will be the same as on a plot of ln(fd) vs. δ 15 NNO3. Insitu variations in temperature and reaction rates may affect the enrichment factor (Kendall and Aravena, 2000) and this was accounted for by allowing for variation in ε15N within the Monte Carlo analysis. The enrichment 35 factor for δ 18 ONO3 (ε18O) was calculated by multiplying the δ 15 NNO3 by a linear coefficient of proportionality determined for each CFO from the slope of the denitrification trend on an isotope cross-plot (see Section 3.2).
For each realization, initial isotopic values (δ 15 N0 and δ 18 O0) were determined by Solver such that the difference between fd calculated from δ 15 NNO3 and δ 18 ONO3 was minimized (<1% difference). The ranges of δ 15 N0 and δ 18 O0 were limited based on measured data and literature values (see 3.2). This approach neglects the effect of mixing of groundwater with differing isotopic values, and is valid if the concentration of NO3in the source is much greater than background concentrations such that the isotopic composition of NO3is dominated by the 5 agriculturally derived end-member.

Quantification of mixing and initial concentrations of Cland NO3-N
A binary mixing model that also accounts for decreasing NO3-N concentrations in response to denitrification was used to quantify NO3attenuation by mixing and estimate the initial concentrations of Cland NO3-N. The measured concentration of Clwas assumed to be a function of two end-member mixing, described by 10 where Cl is the measured concentration of Clin the groundwater sample, Cli is the concentration of Clat the initial point of entry of the agriculturally derived NO3to the groundwater system, Clb is the concentration of Clin the background ambient groundwater, and fm is the fraction of water in the sample from the source of agriculturally derived Cl -(and NO3 -) remaining in the mixture. 15 The concentration of NO3-N was also assumed to be a function of two end-member mixing but with an additional coefficient, fd (the fraction of NO3-N remaining after denitrification), applied to account for denitrification. The measured NO3-N concentration was thus described by where NO3-N is the concentration of NO3-N measured in the groundwater sample, NO3-Ni is the concentration of 20 NO3-N in the source of agriculturally derived NO3at the initial point of entry to the groundwater system, and NO3-Nb is the concentration of NO3-N in the background ambient groundwater. This mixing calculation was only conducted on samples for which NO3dominated total-N (NH3-N <10% of NO3-N) so that nitrification of NH3 could be neglected.
If Cli is much greater than Clb and NO3-Ni is much greater than NO3-Nb, then fm is insensitive to background 25 concentrations and these terms can be neglected (see Section 4.2 for further discussion of this assumption). In this case, Eqs. (3) and (4) reduce to Solving Eq. (6) for fm and substituting into Eq. (5) yields 30 Thus, for each groundwater sample, the ratio of NO3-N/Clat the initial point of entry of the agriculturally derived NO3to the groundwater system ( 3i i ) can be simply calculated using measured concentrations, and fd estimated from NO3isotope data. This provides a relatively simple method to identify agriculturally derived NO3from different sources (e.g., EMS vs. manure piles) if they have different NO3-N/Clratios. Estimated Cli and 35 NO3-Ni are reported as the mid-range value with uncertainty described by the minimum and maximum values.
These initial concentrations are at the water table for top-down inputs, or at the saturated point of contact between the EMS and the aquifer for leakage from the EMS. This analysis assumes that a sampled water parcel consists of water with agriculturally derived NO3that entered the aquifer from one source at one point in time and space and has since mixed with natural ambient groundwater. Any NO3produced during nitrification after the anthropogenic source water enters the aquifer is implicitly included in NO3-Ni. The error in 3i i -was assumed to be dominated by error in the estimated fd, with the measurement error in NO3-N and Clconsidered negligible.
Using the geometric mean K for the sand (5.0 x 10 -6 m s -1 ) and a lateral head gradient of 1.4×10 -2 m m -1 yields a specific discharge (Darcy flux, q) of 2.2 m y -1 . Assuming an effective porosity of 0.3 (Rodvang et al., 1998), the denitrification, agriculturally derived NO3could have been transported through the groundwater system by advection about 400 m from the EMS since 1960 and 630 m since 1930.

CFO4
The geology at CFO4 consists of about 5 m of clay (with minor till) underlain by sandstone, to the maximum depth investigated (20 m BG). Hydraulic conductivities measured using slug tests on wells were 1.0×10 -8 to 5 1.0×10 -5 m s -1 (n=12) for the clay and sandstone (many shallow wells were screened across the clay-till and into the sandstone) and 1.0×10 -5 to 2.9×10 -5 m s -1 (n=4) for the sandstone. The depth to water table ranged from 1.0 to 3.4 m, increasing from west to east across the study site. Seasonal water table variations were on the order of 1.5 m with water table declines on the order of 0.3 m y -1 . The horizontal hydraulic gradient was consistently from west to east, with a mean gradient at the water table of 3.9×10 -3 m m -1 between BC2 and BMW2 and 4.3×10 -3 m 10 m -1 between BMW2 and BMW7. Vertical hydraulic gradients were 4.2×10 -2 to 4.6×10 -2 m m -1 downward. Using the geometric mean K for the site (2.9×10 -5 m s -1 ) and a lateral head gradient of 4.3×10 -3 m m -1 yields a q of 0.4 m y -1 . Assuming an effective porosity of 0.3 yields a ̅ of 1.3 m y -1 . These values suggest that, in the absence of attenuation by mixing or denitrification, anthropogenic NO3could have been transported through the groundwater systems about 10 m by advection between 1995 and the time of sampling. 15

Values and evolution of stable isotopes of nitrate
Manure filtrate from the EMS at CFO1 had δ 15 NNO3 ranging from 0.4 to 5.0‰ and δ 18 ONO3 ranging from 7.1 to 19.0‰. A curve showing the co-evolution of δ 18 ONO3 (mixing of atmospheric δ 18 O with groundwater-derived δ 18 O) and δ 15 NNO3 (Rayleigh distillation, β = 1.005) during nitrification is shown in Fig. 2. Isotopic values in DMW3, where direct leakage from the EMS was evident, are consistent with partial nitrification following this 20 trend of isotopic evolution (δ 18 ONO3 of -1.2‰ and δ 15 NNO3 of 7.8‰).
At both sites, co-enrichment of δ 18 ONO3 and δ 15 NNO3 characteristic of denitrification was evident in some samples (slopes of 0.42 and 0.72 in Fig. 2a). At CFO1, this includes samples from DP10-2, DMW5, DMW11, DMW12, DP11-12b, and DMW13 (and associated core) and some pore water from cores DC15-22 and DC15-23. These 35 samples had NO3-N concentrations of 0.6 to 23.7 mg L -1 , δ 18 ONO3 ranging from 4.8 to 20.6‰, and δ 15 NNO3 ranging from 22.9 to 61.3‰. At CFO4, samples exhibiting evidence of denitrification were from BMW2, BMW5, BMW6, BMW7, and BC4. These samples had NO3-N concentrations ranging from 0.4 to 35.1 mg L -1 , δ 18 ONO3 ranging from 1.6 to 22.1‰, and δ 15 NNO3 ranging from 20.9 to 70.1‰. Although the isotopic values of DMW5 suggest enrichment by denitrification, the data plot away from the rest of the CFO1 data and close to the denitrification trend at CFO4 (Fig. 2), suggesting these samples were affected by some other process (possibly mixing or nitrification); therefore, the fraction of NO3-N remaining in this well was not calculated. Also, well DMW3, which clearly receives leakage from the EMS, did not contain substantial NO3-N and so fd was not calculated. 5 In the Monte Carlo analysis tThe potential range of original isotopic values of the NO3source prior to denitrification (δ 15 N0 and δ 18 O0R0) varied from 5 to 27‰ for δ 15 NNO3 and from -2 to 7‰ for δ 18 ONO3 based on isotopic values measured during this study (Fig. 2a). These values are consistent with literature values for manuresourced NO3 -, which report δ 15 NNO3 ranging from 5 to 25‰ and δ 18 ONO3 ranging from -5 to 5‰ (Wassenaar, 1995;Wassenaar et al., 2006;Singleton et al., 2007;McCallum et al., 2008;Baily et al., 2011). 10 The enrichment factor of δε15NNO3 was defined by a normal distribution with a mean of -10‰ and standard deviation of 2.5‰ (Fig. 2b). At CFO1, the coefficient of proportionality between the enrichment factor of δ 15 NNO3 and δ 18 ONO3 was described by a normal distribution with mean of 0.72 and standard deviation of 0.05. At CFO4, the coefficient of proportionality was also described by a normal distribution with a mean of 0.42 and standard deviation of 0.035 (see Fig. 2a). These enrichment factors are consistent with values from denitrification studies 15 that report ε15N ranging from -4.0 to -30.0‰ and ε18O ranging from -1.9 to -8.9‰ (Vogel et al., 1981;Mariotti et al., 1988;Spalding and Parrott, 1994;Mengis et al., 1999;Pauwels et al., 2000;Otero et al., 2009).

Distribution and sources of agricultural nitrate in groundwater
At both sites TN concentrations in filtrate from the EMS and catch-basin were generally an order of magnitude 20 larger than concentrations in groundwater ( Table 2). The one exception is well DMW3 at CFO1 which intercepted direct leakage from the EMS (see 3.3.1 for further discussion of this well)., The dominant form of N differed between manure filtrate and groundwater. In the EMS filtrate, N was predominately organic-N (TON up to 71%) or NH3-N (up to 90%), with NOx-N <0.1% of TN. In the catch-basin at CFO1 TON was >99% of TN. In groundwater TN concentrations ranged from <0. 25 with the exception of DMW3).

CFO1
Agriculturally derived NO3was predominantly generally restricted to the upper 20 m (or less) at CFO1 (NO3-N ≤ 0.2 mg L -1 and Cl -≤ 57 mg L -1 in seven wells screened at 20 m). The one exception was DP11-12b, which had up to 4.1 mg L -1 of NO3-N. The southeast portion of the site also does not appear to have been significantly 30 contaminated by agriculturally derived NO3 -, with NO3-N concentrations < 1 mg L -1 in five water table wells (DMW4, DMW6, DMW14, DMW15, DMW16). In DMW6, Cland TN concentrations were elevated (see Supplementary Material) but NO3-N concentrations were < 2 mg L -1 . Collectively, these data suggest the catch basin is not a significant source of NO3to the groundwater at this site.
Leakage of manure slurry from the EMS at CFO1 is clearly indicated by the data from DMW3, which feature the 35 highest concentrations of TN in groundwater (up to 548 mg L -1 ) and elevated Cl -, HCO3 -, and DOC in concentrations similar to EMS manure filtrate (see Supplementary Material). Nevertheless, NO3-N concentrations in this well were consistently low (1.1 ± 2.7 mg L -1 , n=22). The potential for nitrification in the vicinity of this well is indicated by NO2-N production (2.7 ± 8.3 mg L -1 , n=22). However, the data demonstrate that only a small proportion of the NH3-N in DMW3 (373.4 ± 79.4 mg L -1 , n=22) could have been converted to NO3within the subsurface (NO3-N in groundwater ≤ 66 mg L -1 ) (NO3-N/Clratio of 0.95). Further work is required to assess the importance of cation exchange as an attenuation mechanism for direct leakage from the EMS at this site.
Given the evidence of partial nitrification in DMW3 (and low NO3-N concentrations), the NO3-N/Clratio of contamination from the EMS was assumed to be best represented by DP10-2, which is located directly downgradient of the EMS. Data for this well indicate values of NO3-N/Clpredominantly ranging from 0.1 to 0.3 10 with NO3-Ni/Cli estimated at 0.3 ± 0.13 (Fig. 4).
The maximum NO3-N concentration in groundwater at CFO1 (66.4 mg L -1 ) was measured in core sample DC15--23 (clay at 2 m bgl, 7 m hydraulically downgradient of DMW3). The NO3-N in this core sample was most likely introduced into the groundwater system by vertical infiltration or diffusion from above. PPore water extracted 15 from the unsaturated zone (sand) at the top of this core profile contained 865 mg L -1 of NO3-N and had a NO3-N/Clratio of 1.04, consistent with the ratio of 0.95 in the core sample. Given this consistency, and that The NO3-N concentrations in the well immediately up-gradient were low (DMW3), the NO3-N in this core sample was most likely introduced into the groundwater system by vertical infiltration or diffusion from above. In contrast, Contamination by agricultural NO3that exceeds the drinking water guidelines (NO3-N > 10 mg L -1 ) was observed 20 in wells DMW2 and DMW12 also had NO3-N concentrations that were elevated but did not exceed the drinking water guideline (≤ 3.7 mg L -1 ). up to 40 m hydraulically downgradient of the EMS (DMW13, DP10-2) and in well DMW11 situated 470 from the EMS (Fig. 3). DMW1, located upgradient of the EMS, also had concentrations of NO3-N > 10 mg L -1 with an increasing trend, but the source of this NO3is not clear. DMW2 and DMW12 also had NO3-N concentrations that were elevated but did not exceed the drinking water guideline (≤ 3.7 mg L -1 ). 25 Given the evidence of incomplete partial nitrification in DMW3, the NO3-N/Clratio of contamination from the EMS was assumed to be best represented by DP10-2, which is located directly downgradient of the EMS. Data for this well indicate values of NO3-N/Clpredominantly ranging from 0.1 to 0.3 with NO3-Ni/Cli estimated at 0.3 ± 0.13 (Fig. 4). Advective transport from DMW3 is also the likely source ofelevated NO3-N (up to 21.1 mg L -1 ) within the sand between 6 and 12 m depth in this core had in DC15-23. NO3-N/Clratios consistent with an EMS 30 source (0.07 to 0.31)in these samples ranged from 0.07 to 0.31, consistent with DP10-2.. Stable isotope values in pore water from this sand layer do not indicate substantial denitrification (δ 18 O ≤ 5.9‰, δ 15 N ≤ 16.7‰), suggesting these ratios will be similar to the initial ratios at the point of entry to the groundwater system.
In DMW13 (33 m downgradient from DP10-2) tIn contrast, thehe ratio of NO3-Ni/Cli in DMW13 (33 m downgradient from DP10-2) was 0.75 ± 0.29, is more similar to the NO3-N/Clratio in DC15-23 at 2 m (0.95), 35 which is interpreted as reflecting a top-down source. The NO3in DMW13 is therefore unlikely to be sourced solely from leakage from the EMS, and could be sourced from the adjacent dairy pens or a temporary manure pile that was observed adjacent to this well during core collection in 2015 (or a combination of EMS and top-down sources).
The NO3-Ni/Cli ratio Iin DMW12 the NO3-Ni/Cli ratio is was not inconsistent with an EMS source, but the hydraulic gradient between DMW2 and DMW12 is negligible, indicating a lack of driving force for advective transport from the EMS towards DMW12. This is also the case for well DMW1, which is up-gradient of the EMS but had elevated NO3-N concentrations (6.5 ± 3.6, n=18). The source of nitrate in these wells is therefore unlikely to be related to leakage from the EMS, but alternative sources (i.e., nearby temporary manure piles) are not known. 5 Well DMW11, 470 m from the EMS, had had conconsistently low NO3-N/Clratios (< 0.05). The NO3-Ni/Cli ratio indicated by DMW11 was similar to DP10-2, but estimates of Cli were three-fold higher than Cli for DP10-2 (Fig. 4b). , but estimates of Cli indicate Clsourced from inputs with three-fold higher Clconcentrations than the source to DP10-2 (Fig. 4b). NO3-Ni and Cli estimated for DMW11 were consistent with measured values in that well, indicating a local top-down source. Well DMW11 is located hydraulically downgradient of feedlot pens 10 and adjacent to a solid manure storage area. Well DMW11 is also, in a local topographic low. and Elevated NO3-N in this well is therefore interpreted to be likely receiving NO3-N and Clfrom surface runoff and top-down infiltration, rather than lateral advection from the EMS in addition to subsurface groundwater flow. Well DMW11 had high NO3-Ni and Cli consistent with measured values in that well, indicating a local top-down source that is likely the nearby solid manure pile. 15

CFO4
At CFO4, measured data indicate that effects from agricultural operations on NO3concentrations in groundwater are restricted to the upper 15 m of the subsurface. NO3-N concentrations in wells screened at 15 m depth were < 0.5 mg L -1 , with the exception of one sample from BP10-15w (May 2012) with 4.3 mg L -1 of NO3-N. Water 20 table wells in the west and north of the study site (BC1, BC2, and BC3) also indicate negligible impacts of agricultural operations, with Cl -< 10 mg L -1 and NO3-N < 0.1 mg L -1 .
Concentrations of NO3-N > 10 mg L -1 were measured in three water table wells (BMW2, BMW3, BMW4) installed adjacent to the EMS, indicating that they have been impacted by the EMS (Fig. 5). Of these, BMW2 had much higher Clconcentrations (502 ± 97 mg L -1 , n=22 in BMW2 compared to 182 ± 81 mg L -1 in BMW3 and 25 188 ± 74 mg L -1 in BMW4), and therefore lower NO3-N/Clratios (< 0.05). Given the elevated Clconcentrations in this wellBMW2 were consistent with concentrations in the EMS suggesting , direct leakage , while sfrom the EMS was assumed to be the source. Sttable isotopes of NO3and initial concentrations (NO3-Ni ≥ 127 mg L -1 ) indicate substantial denitrification (Table 2, Fig. 6) in BMW2, with estimated NO3-Ni ≥ 127 mg L -1 . The and an NO3-Ni/Cli ratio in BMW2 is consistent with of 0.1 to 0.3 measured NO3-N/Cl - (Fig. 6) in . This ratio is consistent 30 with data from well BMW4, which is immediately adjacent to the EMS (on the upgradient side) and therefore likely reflects leakage from the EMS without denitrification (based onconsistent with stable isotopes of values of NO3 -). NO3-N/Clratios measured in BMW4 were predominantly 0.1 to 0.3, consistent with the reconstructed NO3-Ni/Cli ratio in BMW2.
Given that the estimated subsurface travel distance during operations at this site is 10 m, Aagriculturally derived 35 NO3in other wells not immediately adjacent to the EMS is unlikely to be related to leakage from the EMS. Wells BMW5 and BMW7 are 60 and 140 m hydraulically downgradient from the EMS, respectively. NO3-Ni/Cli ratios in these wells were not inconsistent with BMW2 (i.e., the range of values overlap), but given the distance from the EMS but advective transport is only likely to have transported solutes around 10 m since the EMS was installed (see Section 3.1.2). As such, the source of NO3-N in these wells is most likely the adjacent dairy pens rather than the EMS. Concentrations of NO3-N > 10 mg L -1 were also measured in BC4, which is located 95 m hydraulically upgradient of the EMS. The ratio of NO3-Ni/Cli at BC4 was the highest at CFO4 (0.6) and did not overlap with BMW2. This indicates that tThe NO3in this well is interpreted to have been was sourced from an adjacent manure pile, which was observed during the study. 5

Mechanisms of attenuation of agriculturally derived NO3 -
Attenuation of agriculturally derived NO3in groundwater is dominated by denitrification at both CFO1 and CFO4, with estimates of fm consistently higher than estimates of fd (Table 3, Fig. 7, Table S10). Calculated fd values indicate that where denitrification was identified, suggest that at least half of the NO3-N present at the initial point of entry to the groundwater system has been removed by denitrificationthis attenuation mechanism. 10 Comparison of NO3-Nmix (the concentration of NO3-N that would be measured if mixing was the only attenuation mechanism) with measured concentrations (which reflect attenuation by both mixing and denitrification) suggests that the sample from 20 m depth (DP11-12b) is the only sample that would be below the drinking water guideline if mixing was the only attenuation mechanism (Fig. 8).
At both sites, the stable isotope values of NO3indicate that denitrification proceeds within metres of the source. 15 At CFO1, calculated fd in well DP10-2 (2 m from the EMS) is 0.52 ± 0.22; at CFO4, fd in well BMW2 (3 m from the EMS) is 0.13 ± 0.06. Denitrification also substantially attenuated NO3-N concentrations in wells where the source is not the EMS but instead is adjacent solid manure piles (e.g., DMW11 at CFO1, BC4 at CFO4). In BMW6 at CFO4, denitrification completely attenuated the agriculturally derived NO3 -. This well had negligible NO3-N (0.4 ± 0.2 mg L -1 , n=8) and the lowest fd of 0.01. Measured DOC in this well was consistent with other wells at 20 both sites (6.9 ± 1.7 mg L -1 , n=3), suggesting DOC depletion does not limit denitrification at these CFO operations.
Calculated fd and fm should decrease with increasing subsurface residence time and distance from source. Data from wells support the source identification based on concentrations of NO3-N and Cland NO3-N/Clratios (see Section 3.3). Well DMW11 (470 m from the EMS) had the highest fm at CFO1 (0.83), indicating less mixing and suggesting the anthropogenic source of NO3in this well is relatively close, which is consistent with the adjacent 25 the solid manure pile being the source of NO3to this well. At CFO4, well BMW2, which is adjacent to the EMS, had the highest fm (0.92), indicating the least attenuation of NO3 by mixing and consistent with the EMS being the source of NO3to this well. 4.1 Implications for on-farm waste management Agriculturally derived NO3at these two sites with varying lithology is was generally restricted to depths < 20 m, consistent with previous studies at CFOs (Robertson et al., 1996;Rodvang and Simpkins, 2001;Rodvang et al., 2004;Kohn et al., 2016). Attenuation of agriculturally derived NO3in groundwater is was a spatially varying combination of mixing and denitrification, with denitrification playing a greater role than mixing at both sites. In 35 the samples for which fd could be determined, denitrification reduced NO3concentrations by at least half and, in some cases, back to background concentrations. Given that the range of source isotopic composition was allowed to vary to its maximum justifiable extent, these quantitative estimates of denitrification based on stable isotopes of NO3are likely to be conservative. Redox conditions within the groundwater system were not able to be determined in this study due to the sampling method used to collect groundwater from wells screened across low-K formations (well bailed dry then sample collected after water level recovery). However, Ddenitrification appears to proceed within metres of the NO3source, suggesting relatively short sub-surface residence times are 5 required and that redox conditions at theclose to the water table may beare conducive to denitrification reactions .

Discussion
The substantial role of denitrification within the saturated glacial sediments at these study sites indicates the potential for significant attenuation of agriculturally derived NO3by denitrification in similar groundwater systems across the North American interior and Europe . Denitrification 10 in the unsaturated zone is limited by low water contents and oxic conditions, resulting in substantial stores of NO3in vadose zones (Turkeltaub et al., 2016;Ascott et al., 2017). NO3in water that is removed rapidly from site is also unlikely to be substantially attenuated by denitrification due to oxic conditions and rapid transit times . Therefore, water management focussed on reducing the effects of NO3contamination in similar hydrogeological settings to this study should aim to maximize infiltration into the saturated zone where 15 NO3concentrations can be naturally attenuated, provided that local groundwater isn't used for potable water supply.
At both sites there is evidence of elevated NO3due to leakage from the EMS, but the impact appears to be limited to within metres of the EMS. This suggests that saturation within the clay lining of the EMS has limited the development of extensive secondary porosity that would allow rapid water percolation (Baram et al., 2012). 20 Infiltration of NO3rich water that has passed through temporary solid manure piles and dairy pens has resulted in groundwater NO3-N concentrations as high as those associated with leakage from the EMS (e.g., DMW11, DMW13, BC4). At CFO4, this is in spite of the presence of clay at surface, which is attributable toreflecting secondary porosity in the upper part of the profile that has led to hydraulic conductivities comparable to sand. This is consistent with the findings of Showers et al. (2008), who investigated sources of NO3at an urbanized 25 dairy farm in North Carolina, USA. Construction of EMS facilities in Alberta has been regulated under the Agriculture Operation Practices Act since 2002, which requires them to be lined with clay to minimise leakage (Lorenz et al., 2014). The results of this study suggest that oOn-farm waste management should increasingly focus on minimising temporary manure piles that are in direct contact with the soil to reduce NO3contamination associated with dairy farms and feedlots. 30 The absence of direct leakage from the EMS at CFO4 suggests that saturation within the clay lining of the EMS has limited the development of extensive secondary porosity that would allow rapid water percolation (Baram et al., 2012). Elevated NH3-N concentrations in the water table well at the southeast corner of the EMS at CFO1 (DMW3) do indicate direct leakage from the EMS, but because nitrification within the EMS is minimal, this has not resulted in elevated NO3-N in this well. Two possibilities for the fate of NH3-N in DMW3 are attenuation by 35 cation exchange and oxidation to NO3-N within the groundwater system. Measured NO3-N concentrations in groundwater represent only a small fraction (≤ 10%) of NH3-N within the EMS (or DMW3), suggesting oxidation to NO3within the aquifer may be limited. Further work is required to assess the importance of cation exchange as an attenuation mechanism for direct leakage from the EMS at this site.

Critique of this approach and applicability at other sites
The sources of manure-derived NO3 -(manure piles vs. EMS) are distinguishable based on NO3-Ni/Cli ratios, provided there is also an understanding of the history of each site, local hydrogeology, and potential sources. At both sites, leakage from the EMS had NO3-Ni/Cli of between 0.1 and 0.4, but this alone was not diagnostic of the source. The sources of manure-derived NO3 -(manure piles vs. EMS) are distinguishable based on NO3-Ni/Cli 5 ratios, provided there is also an understanding of the history of each site, local hydrogeology, and potential sources. Calculated fd and fm generally decreased with increasing subsurface residence time and distance from source, providing additional evidence for source attribution. For example, at CFO4, well BMW2, which is adjacent to the EMS, had the highest fm (0.92), indicating the least attenuation of NO3 by mixing and consistent with the EMS being the source of NO3to this well. is primarily related to uncertainty in the initial concentrations (Cli and NO3-Ni), which depends on measured Cland NO3-N. The largest uncertainties in NO3-Ni and Cli correspond to the lowest measured concentrations (i.e., 25 furthest from the upper limit), with less uncertainty at higher measured concentrations as they approach the maximum values. Temporal variability in NO3-Ni/Cli for each source could not be determined based on the snapshot isotope sampling conducted, but this could be investigated by measuring NO3isotopes in conjunction with NO3-N and Clat multiple times.
Although applicable at these sites, this approach may not be valid at other sites if additional sources of NO3 in 30 groundwater (e.g. fertilizer or nitrification) are significant, or if NO3 concentrations in groundwater are naturally elevated (Hendry et al., 1984). The combination of the approach outlined here with measurement of groundwater age indicators would allow for better constraints on groundwater flow velocities and determination of denitrification rates Katz et al., 2004;McMahon et al., 2004;.

Comparison with isotopic values of NO3in previous studies 35
Nitrate isotope values in groundwater at the two CFOs studied are were generally consistent with previous studies reporting denitrification of manure-derived NO3at dairy farms (Wassenaar, 1995;Wassenaar et al., 2006;Singleton et al., 2007;McCallum et al., 2008;Baily et al., 2011). However, Tthe isotopic values of NO3in the manure filtrate from the EMS at CFO1, were generally inot nconsistent with values for manure-sourced NO3 -reported in other groundwater studies (Wassenaar, 1995;Wassenaar et al., 2006;Singleton et al., 2007;McCallum et al., 2008a;Baily et al., 2011). This is likely to be because nitrification within the EMS was negligible (NO3-N <0.7 mg L -1 ), such that the isotopic values of NO3-N in the manure filtrate reflect volatilization of NH3 and partial nitrification within the EMS. δ 18 ONO3 values may also have been affected by evaporative enrichment of the δ 18 OH2O being incorporated into NO3 - (Showers et al., 2008). 5 However, aA number of groundwater samples collected for during the presentthis study had relatively enriched δ 18 ONO3 (> 15 ‰) with depleted δ 15 NNO3 (< 15‰). Some of these isotopic values are within the range previously reported for NO3derived from inorganic fertilizer (δ 15 NNO3 from -3 to 3‰ and δ 18 ONO3 from -5 to 25‰), with the δ 18 ONO3 depending on whether the NO3is from NH4 + or NO3in the fertilizer (Mengis et al., 2001;Wassenaar et al., 2006;Xue et al., 2009). To the best of our knowledge, however, no inorganic fertilizers have been applied at 10 these study sites. Another potential source is NO3derived from soil organic N, but this should have δ 15 NNO3 values of 0 to 10‰ and δ 18 ONO3 values of -10 to 15‰ Mayer et al., 2001;Mengis et al., 2001;Xue et al., 2009;Baily et al., 2011). Incomplete nitrification of NH4 + can result in δ 15 NNO3 lower than the manure source , but as there was no measurable NH3-N in these samples this is also unlikely. These isotope values may reflect the influence of NO3from precipitation, which usually has values ranging from -5 to 15 5‰ for δ 15 NNO3 and 40 to 60‰ for δ 18 ONO3, and has been reported to dominate NO3isotope values of groundwater under forested landscapes . Alternatively, they may be affected by microbial immobilization and subsequent mineralization and nitrification, which can mask the source δ 18 ONO3 in aquifers with long residence times (Mengis et al., 2001;Rivett et al., 2008).
The isotopic values of NO3in the manure filtrate from the EMS at CFO1, were generally inconsistent with values 20 for manure-sourced NO3reported in other groundwater studies (Wassenaar, 1995;Wassenaar et al., 2006;Singleton et al., 2007;McCallum et al., 2008a;Baily et al., 2011). This is likely to be because nitrification within the EMS was negligible (NO3-N <0.7 mg L -1 ), such that the isotopic values of NO3-N in the manure filtrate reflect volatilization of NH3 and partial nitrification within the EMS. δ 18 ONO3 values may also have been affected by evaporative enrichment of the δ 18 OH2O being incorporated into NO3 - (Showers et al., 2008). 25

Conclusions
A mixing model constrained by quantitative estimates of denitrification from isotopes substantially improved our understanding of nitrate contamination at these sites. This novel approach has the potential to be widely applied as a tool for monitoring and assessment of groundwater in complex agricultural settings. NO3-N concentrations in excess of the drinking water guideline were measured at both sites, with sources including manure piles, pens 30 and the EMS. Even though these sites are dominated by clay-rich glacial sediments, the input of NO3to groundwater from temporary manure piles and pens resulted in comparable (or greater) NO3-N concentrations than leakage from the EMS. This is attributed to the development of secondary porosity within unsaturated clays.
On-farm management of manure waste should increasingly focus on limiting manure piles that are in direct contact with the soil to limit NO3contamination of groundwater. Nitrate attenuation at both sites is dominated by 35 denitrification, which is evident even in wells directly adjacent to the NO3source. In the wells for which denitrification was identified, concentrations ofOn-site denitrification agriculturally-derived reduced agriculturally derived NNO3concentrations had been reduced by at least half and, in some wells, completely. In the absence of denitrification all but one of these wells would have had NO3-N concentrations above the drinking water guideline.
These results indicate that infiltration to groundwater systems in glacial sediments where NO3can be naturally attenuated is likely to be preferable to off-farm export via runoff or drainage networks, provided that local groundwater isn't a potable water source. On-farm management of manure waste at similar operations should 5 increasingly focus on limiting manure piles that are in direct contact with the soil to limit NO3contamination of groundwater.  * NOx-N of 50 mg L -1 in DMW3 consisted of 12.6 mg L -1 as NO3-N and 37.4 mg L -1 as NO2-N. ** NOx-N max in groundwater measured in core (NO3-N = 66.4 mg L -1 , NOx-N = 67.8 mg L -1 ) ^R ange across three replicates measured on 25 August 2011