Printer-friendly Version Interactive Discussion Tracing Atmospheric Nitrate in Groundwater Using Triple Oxygen Isotopes: Evaluation Based on Bottled Drinking Water Bgd Printer-friendly Version Interactive Discussion

Biogeosciences Discussions This discussion paper is/has been under review for the journal Biogeosciences (BG). Please refer to the corresponding final paper in BG if available. Abstract The stable isotopic compositions of nitrate dissolved in 49 types of bottled drinking water collected worldwide were determined, to trace the fate of atmospheric nitrate (NO − 3 atm) that had been deposited into subaerial ecosystems, using the 17 O anomalies (∆ 17 O) of nitrate as tracers. The use of bottled water enables collection of groundwater 5 recharged at natural, background watersheds. The nitrate in groundwater had small ∆ 17 O values ranging from −0.2 ‰ to +4.5 ‰ (n = 49). The average ∆ 17 O value and average mixing ratio of atmospheric nitrate to total nitrate in the groundwater samples were estimated to be 0.8 ‰ and 3.1 %, respectively. These findings indicated that the majority of atmospheric nitrate had undergone biological processing before being ex-10 ported from the surface ecosystem to the groundwater. Moreover, the concentrations of atmospheric nitrate were estimated to range from less than 0.1 µmol l −1 to 8.5 µmol l −1 , with higher NO − 3 atm concentrations being obtained for those recharged in rocky, arid or elevated areas with little vegetation and lower NO − 3 atm concentrations being obtained for those recharged in forested areas with high levels of vegetation. Additionally, many 15 of the NO − 3 atm-depleted samples were characterized by elevated δ 15 N values of more than +10 ‰. Uptake by plants and/or microbes in forested soils subsequent to deposi-tion and the progress of denitrification within groundwater likely plays a significant role in the removal of NO − 3 atm .


Introduction
Anthropogenic activities have increased emissions of fixed nitrogen from land to the atmosphere.Indeed, the amount of these emissions has almost doubled globally, with much greater increases being observed in some regions, and this fixed-nitrogen flux is expected to double again by 2030 (Galloway et al., 2008).Excess fixed-nitrogen input is linked to various environmental problems, including forest decline (e.g., Fenn et al., 1998), degradation of groundwater quality (e.g., Murdoch and Stoddard, 1992;Williams et al., 1996), eutrophication of the hydrosphere (e.g., Paerl, 1997;Duce et al., 2008), and shifts in biodiversity (e.g., Tilman et al., 1996).
Previous detailed studies of forested catchments have offered considerable insight into the link between atmospheric nitrate deposition and nitrate discharge to streams (Grennfelt and Hultberg, 1986;Williams et al., 1996;Tietema et al., 1998;Durka et al., 1994).Although water chemistry has been routinely measured through many programs on regional to international scales, our understanding of the mechanisms that regulate the discharge of atmospheric deposition from surface ecosystems is still limited.This is because the fate of atmospheric nitrate deposited onto the subaerial ecosystem is determined through the complicated interplay of several processes that include (1) dilution through nitrification, (2) uptake by plants or microbes, and (3) decomposition through denitrification.
In a previous study, we determined the 17 O anomalies of nitrate in groundwater recharged on the island of Rishiri, Japan to trace the fate of atmospheric nitrate (NO − 3 atm ) that had been deposited onto and passed through the island F. Nakagawa et al.: Tracing atmospheric nitrate in groundwater ecosystem.In that study, we estimated the average mixing ratio of atmospheric nitrate to total nitrate in the groundwater to be 7 % (Tsunogai et al., 2010).Based on this mixing ratio, we estimated that direct drainage accounts for 10.5 ± 5.2 % of atmospheric nitrate that has been deposited onto the island, and that the residual portion has undergone biological processing before being exported from the island ecosystem at the groundwater recharge zones.
Anomalies in 17 O can be quantified using the 17 O signature, which is defined by the following equation (Miller, 2002;Kaiser et al., 2007): where β is 0.5247 (Miller, 2002;Kaiser et al., 2007),  (Michalski et al., 2004;Tsunogai et al., 2010Tsunogai et al., , 2011;;Dejwakh et al., 2012).Measuring the 17 O values of nitrate for water eluted from various watersheds can increase our understanding of fixed-nitrogen processing and fixed-nitrogen retention efficiencies for surface natural ecosystems that are subjected to atmospheric fixed-nitrogen deposition.
In this study, we determined the 17 O values of nitrate in various types of bottled drinking water collected from different markets by applying those representing groundwater recharged at natural watersheds worldwide (Tsunogai andWakita, 1995, 1996).Using bottled drinking water for the samples reduced the time, labor, and costs relative to those that would be required for direct groundwater sampling.Moreover, the use of bottled water enables collection of groundwater recharged at natural background watersheds where there have been few artificial alternations and little contamination.The use of bottled water also prevents samples of groundwater systems recharged in declining or damaged forested watersheds as well as anoxic nitrate-exhausted groundwater.
Using the 17 O tracer, we quantified the fraction of NO − 3 atm within the total nitrate output in groundwater to gain insight into the processes controlling the fate and transport of NO − 3 atm deposited onto natural subaerial ecosystems.The quantified NO − 3 atm output will be useful in future studies as a basic background data set for evaluation of the amounts of NO − 3 atm eluted into groundwater from subaerial ecosystems in general, including the same watersheds under elevated NO − 3 atm input in the future.In addition to clarifying the fate of atmospheric nitrate, this study was also conducted to gain insight into the processes controlling the total nitrate concentrations in groundwater.Because nitrate enrichment of drinking water causes infantile methemoglobinemia, major factors controlling nitrate concentrations must be clarified for each groundwater system.The determination of stable isotopic compositions of nitrate in groundwater will provide insight into the processes controlling the nitrate concentrations in natural groundwater, especially for the contribution of atmospheric nitrate into groundwater.

Sample description
We collected as many bottled drinking water samples as possible from supermarkets in Sapporo and Nagoya, Japan.Because of increasing numbers of people becoming healthconscious in Japan, it is possible to obtain a wide variety of bottled drinking water from supermarkets.In addition to samples collected from Japanese supermarkets, we asked acquaintances abroad to send bottled drinking water to Japan.
We collected a total of 48 brands of bottled drinking water (Nos. 2 to 49 in Table 1) that had been sealed in either polyethylene terephthalate bottles or glass bottles and distributed on the market as drinking water.About 70 % of the samples collected were from abroad (Fig. 1).The longest storage period between bottling and analysis was less than a year.In addition to the water being collected in this study, a brand of bottled drinking water collected and analyzed in Tsunogai et al. (2010) (Rishiri water) has been included in our discussions (No. 1 in Table 1).
The methods of sterilization are presented in the table as well.Because the EU directive for bottled mineral water prohibits sterilization, most of the samples from Europe were non-sterile.Conversely, most samples from Japan were sterilized by either (1) microfiltration (0.2 µm), and/or (2) heating to 130 • C for at least 2 s; therefore, both concentrations and stable isotopic compositions may have been altered from the original to some extent.We previously determined both the concentrations and stable isotopic compositions of nitrate in four sets of the same brand of bottled water (Rishiri water) sold at supermarkets together with those sampled directly at the source well (Meisui factory) (Tsunogai et al., 2010).The differences between the average concentrations, δ 15 N, δ 18 O and 17 O in the bottled water and those in water from the source well were less than 2 %, 0.2 ‰, 0.2 ‰, and 0.2 ‰, respectively (Tsunogai et al., 2010).Because the Rishiri water had been sterilized by a combination of filtration and heating  (Table 1), we concluded that alternations in both concentrations and stable isotopic compositions of nitrate in response to these treatments were small enough to enable bottled water to serve as samples of its source groundwater.
In addition to filtration and heating, some of the samples were sterilized by UV radiation for several seconds (Table 1).UV sterilization could reduce nitrate and increase both the δ 15 N and δ 18 O of residual nitrate; however, the extent of any reduction during the short reaction period would be small because the decomposition rates of nitrate estimated for the wastewater treatments were small, especially under neutral conditions (Kosaka et al., 2002).Moreover, the triple oxygen isotopic compositions remained stable during the reduction.Therefore, we used these UV sterilized samples to represent groundwater from each site.However, caution should be taken before applying the values obtained from such samples to detailed discussions, especially those based on the values of δ 15 N and δ 18 O.
Many of the bottled water samples produced in North America and Asia were sterilized using ozone, which is known to cause large 17 O anomalies (Mauersberger et al., 1999).To exclude those with altered triple oxygen isotopic compositions of nitrate, bottled water sterilized using ozone was excluded.As a result, no samples from North America were evaluated (Fig. 1).

Analysis
The concentration of nitrate together with chloride and other major anions was determined by ion chromatography.The stable isotopic compositions were then determined by converting the nitrate in each sample to N 2 O using the chemical method originally developed to determine the 15 N/ 14 N and 18 O/ 16 O isotope ratios of seawater and freshwater nitrate (McIlvin and Altabet, 2005), with modifications (Tsunogai et al., 2008(Tsunogai et al., , 2010(Tsunogai et al., , 2011)).
The stable isotopic compositions were determined for N 2 O using our Continuous-Flow Isotope Ratio Mass-Spectrometry (CF-IRMS) system (Tsunogai et al., 2008;Hirota et al., 2010), which consists of an original helium purge and trap line, a gas chromatograph (Agilent 6890) and a Finnigan MAT 252 (Thermo Fisher Scientific, Waltham, MA, USA) with a modified Combustion III interface (Tsunogai et al., 2000(Tsunogai et al., , 2002(Tsunogai et al., , 2005) ) and a specially designed multicollector system (Komatsu et al., 2008).An aliquot of N 2 O was introduced, purified, and then carried continuously into the mass spectrometer via an open split interface to monitor isotopologues of N 2 O + at m/z ratios of 44, 45 and 46 to determine δ 45 and δ 46 .Each analysis was calibrated using a machine-working reference gas (99.999 % N 2 O) that was introduced into the mass spectrometer via an open split interface according to a definite schedule to correct for sub-daily temporal variations in the mass spectrometry.In addition, a working-standard gas mixture containing a known concentration of N 2 O (ca. 1000 ppm N 2 O in air) that was injected from a sampling loop was analyzed in the same way as the samples at least once a day to correct for daily temporal variations in the mass spectrometry.
After the analyses based on N 2 O + monitoring, another aliquot of N 2 O was introduced to determine the 17 O for N 2 O (Komatsu et al., 2008).Using the same procedures as those used in the N 2 O + monitoring mode, purified N 2 O was introduced into our original gold tube unit (Komatsu et al., 2008), which was held at 780 • C for the thermal decomposition of N 2 O to N 2 and O 2 .The produced O 2 purified from N 2 was subjected to CF-IRMS to determine the δ 33 and δ 34 by simultaneous monitoring of O + 2 isotopologues at m/z ratios of 32, 33 and 34.Each analysis was calibrated with a machine-working reference gas (99.999 % O 2 gas in a cylinder) that was introduced into the mass spectrometer via an open split interface according to a definite schedule to correct for sub-daily temporal variations in the mass spectrometry.In addition, a working-standard gas mixture containing N 2 O of known concentration (ca.1000 ppm N 2 O in air) that was injected from a sampling loop was analyzed in the same way as the samples at least once a day to correct for daily temporal variations in the mass spectrometry.
The obtained values of δ 15 N, δ 18 O and 17 O for N 2 O derived from the nitrate in each sample were compared with those derived from our local laboratory nitrate standards that had been calibrated using the internationally distributed isotope reference materials (USGS-34 and USGS-35) (Böhlke et al., 2003;Kaiser et al., 2007) to calibrate the δ values of the sample nitrate to an international scale, as well as to correct for both the isotope fractionation during the chemical conversion to N 2 O and the progress of oxygen isotope exchange between the nitrate-derived reaction intermediate and water (ca.20 %).All δ values are expressed relative to air (for nitrogen) and VSMOW (for oxygen) in this paper.
In this study, we adopted the internal standard method for accurate calibrations to determine the δ 15 N, δ 18 O or 17 O values of N 2 O derived from our laboratory nitrate standards.Specifically, we added each of the nitrate standard solutions (containing ca. 10 mmol L −1 nitrate with known δ 15 N, δ 18 O or 17 O values) to additional aliquots of the samples until the nitrate concentration was 3 to 5 times larger than the original, and then converted it to N 2 O and determined the values of δ 15 N, δ 18 O or 17 O in a similar manner as was used for each pure sample.After correcting the contribution of N 2 O from the nitrate in each sample, we could obtain the stable isotopic compositions for N 2 O derived from our laboratory nitrate standards.The δ 15 N, δ 18 O and 17 O values in the samples were then simply calibrated using calibration curves generated from the N 2 O derived from the nitrate standards.
Most samples had nitrate concentrations of more than 1 µmol L −1 , which corresponds to nitrate quantities of more than 30 nmol in a 30 mL sample and is sufficient to determine δ 15 N, δ 18 O and 17 O values with high precision.Indeed, we even attained similar high precision for nitrate-depleted samples showing concentrations of less than 1 µmol L −1 through repeated measurements using another aliquot of water.Thus, all isotopic data presented in this study have an error better than ±0.3 ‰ for δ 15 N, ±0.5 ‰ for δ 18 O and ±0.2 ‰ for 17 O.
Because we used the more precise power law (Eq.1) to calculate 17 O, the estimated 17 O values were somewhat different from those estimated based on traditional linear approximation (Michalski et al., 2002).While the differences were insignificant for groundwater samples newly evaluated in this study (less than 0.05 ‰), the differences would be 0.9 ± 0.2 ‰ for the 17 O values of NO − 3 atm (Tsunogai et al., 2010).When using the linearly approximated 17 O values of NO − 3 atm available in the literature, we recalculated the 17 O values based on the power law.
Nitrite (NO − 2 ) in the samples also interferes with the final N 2 O produced (McIlvin and Altabet, 2005) when the chemical conversion method is used to determine the stable isotopic compositions of nitrate (NO − 3 ).Therefore, it was necessary to correct for the contribution to accurately determine stable isotopic compositions of the sample nitrate.However, all samples analyzed in this study were originally for drinking water, and were characterized by oxic conditions and little biological activity.As a result, they contained NO − 2 at concentrations less than the detection limit, which corre- sponded to NO − 2 /NO − 3 ratios less than 10 %.Thus the results were used without any corrections.
The δD and δ 18 O values of H 2 O in the samples were also analyzed using Cavity Ring-Down Spectroscopy (Picarro L2120-I with an A0211 vaporizer and autosampler) within an error of ±0.1 ‰.Both VSMOW and VSLAP were used to calibrate the values to the international scale.

Groundwater nitrate: overview
The concentrations and δ 15 N, δ 18 O, and 17 O values of nitrate in the samples are presented in Table 2 with their Cl concentrations and the δ 2 O and δ 18 O values of water.We could not determine the 17 O values of nitrate for one of the water samples (No. 18) that had a nitrate concentration of less than 0.3 µmol L −1 .
The average and 1σ variation for concentration, δ 15 N, and δ 18 O of nitrate were 84 ± 197 µmol L −1 , +5.8 ± 6.6 ‰, and +1.2 ± 7.6 ‰, respectively (Table 2, Fig. 2), which are typical values for nitrate in natural groundwater free from artificial pollution (e.g., de Walle and Sevenster, 1998;Hiscock et al., 1991;Kendall, 1998;Moore et al., 2006).A very low δ 15 N value of −11 ‰ was found for No. 13.This is a hot spring water with a temperature of 47 • C at the source well.Some unusual nitrate production such as microbial nitrification could be responsible for the 15 N depletion observed at this site.
The average and maximum 17 O values of nitrate were +0.8 ‰ and +4.5 ‰, respectively (Table 2, Fig. 2).While most of the samples showed positive 17 O values, seven (Nos. 15,28,29,34,37,40,and 45) showed negative 17 O values as low as −0.2 ‰.Because the 17 O value of tropospheric O 2 is around −0.2 ‰ (Luz and Barkan, 2000), the contribution of the oxygen atom derived from tropospheric O 2 during the production of remineralized nitrate from ammonium or organic nitrogen could be partly responsible for  the observed negative 17 O values less than 0. However, even if the contribution was significant, the possible 17 O value of produced remineralized nitrate would include 0 ‰ within the error of our analytical precision (±0.2 ‰).Accordingly, 0 ‰ is used for the 17 O value of remineralized nitrate and observed 17 O values less than 0 ‰ are considered to be 0 ‰ for the remainder of this paper.In other words, possible errors in our further discussions derived from the deviations in the 17 O values of remineralized nitrate from 0 ‰ were included within the error of the analyses.
To verify that atmospheric nitrate was responsible for the elevated 17 O values up to 4.5 ‰ in the samples, the δ 18 O values of nitrate in the samples were plotted as a function of 17 O (Fig. 3).Because atmospheric nitrate is also enriched in 18 O (Durka et al., 1994;Kendall, 1998) 3 atm (Durka et al., 1994;Morin et al., 2009;Alexander et al., 2009;Tsunogai et al., 2010).These findings indicate that the δ 18 O values primarily reflect the contribution of NO − 3 atm as well as 17 O for the samples in Group A. Conversely, data plotted outside of Group A (hereafter referred to as Group B) did not present such a linear correlation.While they showed 18 O enrichment up to +11.6 ‰ in δ 18 O, the 17 O values were always close to 0 ‰.Although the 17 O values of nitrate are stable during the postdepositional isotopic fractionation processes, the δ 18 O values of nitrate can vary.The isotopic fractionations during the partial removal through nitrate uptake or denitrification likely play a significant role in the 18 O enrichment because many of the 18 O-enriched samples in Group B showed the 15 N-enriched δ 15 N values of more than +11 ‰ (Fig. 3).
As shown in Fig. 3, some data in Group B did not show 15 N enrichment, despite showing 18 O enrichment of up to +9.2 ‰.Without the 17 O data for these samples, such 18 Oenrichment would be interpreted as the result of the contribution of atmospheric nitrate (Durka et al., 1994).However, since the 17 O values of these samples were close to zero, we must assume alternative processes for them, such as reducing δ 15 N values through the contribution of some highly 15 N-depleted nitrate as well as their partial removal through nitrate uptake or denitrification.The presence of such data implies that interpretation based only on the values of δ 18 O and δ 15 N of nitrate is not always straightforward.Accordingly, 17 O data would be essential to quantify the contribution of atmospheric nitrate in the hydrosphere.

Mixing ratios of atmospheric nitrate within total nitrate in groundwater
The 17 O data of nitrate in each sample can be used to estimate the mixing ratio of NO − 3 atm to total NO − 3 (C total ) in each groundwater sample by applying Eq. ( 2): where C atm /C total is the mixing ratio of NO − 3 atm to NO − 3 total and 17 O avg denotes the average 17 O value of NO − 3 atm (Tsunogai et al., 2010).In a previous 17 O study conducted on Rishiri Island, we used the 17 O avg value obtained through continuous monitoring on the island.However, in this study, the 17 O avg value at each recharge area was not available for most samples.Because the 17 O value of atmospheric nitrate varies primarily as a result of the reaction path from NO to nitrate in the atmosphere (Michalski et al., 2003;Morin et al., 2009;Alexander et al., 2009), it is difficult to apply the values obtained from monitoring at distant stations.Therefore, we used the 17 O avg values of +20 ± 3 ‰, +25 ± 3 ‰, and +30 ± 3 ‰ for samples collected at latitudes of 20 • S to 20 • N, 20 • N to 50 • N, and 50 • N to 70 • N, respectively, based on the global 17 O distribution of atmospheric nitrate estimated by Alexander et al. (2009) as well as those determined in previous investigations of atmospheric nitrate (Michalski et al., 2003;Savarino et al., 2007;Kaiser et al., 2007;Ewing et al., 2007;Kunasek et al., 2008).Because the 17 O value of NO − 3 atm is a function of the reaction path from NO to nitrate in the atmosphere (Michalski et al., 2003;Morin et al., 2009;Alexander et al., 2009)

F.
Fig. 1.Maps showing the distribution of the source wells worldwide (a), and regionally (b) and (c).

Fig. 2 .
Fig. 2. Histograms of the δ 15 N (a), δ 18 O (b), and 17 O (c) of nitrate in the bottled mineral water samples.Those classified to the outside of each histogram were included in the nearby uppermost/lowermost classes.

Fig. 3 .
Fig. 3. Relationship between 17 O and δ 18 O in NO − 3 from bottled water samples.The open circles denote those showing δ 15 N values of more than +11 ‰.
, we can anticipate 18 O enrichment for those showing elevated 17 O values.As shown in Fig. 3, those showing high 17 O values always accompanied high δ 18 O values; thus, we can find a linear correlation for the samples categorized as Group A in the figure.By extrapolating the linear correlation to the region of NO − 3 atm having 17 O = +20 to +30 ‰, we obtain δ 18 O = +80 ± 30 ‰, which corresponds to the values for NO − 1 and R is the 18 O/ 16 O ratio (or the 17 O/ 16 O ratio in the case of δ 17 O or the 15 N/ 14 N ratio in the case of δ 15 N) of the sample and each international standard.By using the 17 O values of nitrate, we can distinguish NO − 3 atm ( 17 O > 0) from the remineralized nitrate (NO − 3 re ) ( 17 O = 0).Additionally, 17 O is stable during usual massdependent isotope fractionation processes; therefore, we can use 17 O as a conserved tracer of NO − 3 atm and trace NO − irrespective of partial removal through denitrification and/or uptake reactions that occur after deposition.Previous studies of the 17 O values of nitrate in freshwater environments demonstrated that 17 O values can serve as robust tracers of the fate of NO − 3 atm deposited onto a surface ecosystem without using artificial tracers

Table 2 .
Concentration and stable isotopic compositions (δ 15 N, δ 18 O, and 17 O) of nitrate dissolved in the bottled water samples, together with the other parameters.