Opening Opportunities for High-Resolution Isotope Analysis Quanti ﬁ cation of δ 15 N NO3 and δ 18 O NO3 in Di ﬀ usive Equilibrium in Thin − Film Passive Samplers

: The fate of nitrate transported across ground-water-surface water interfaces has been intensively studied in recent decades. The interfaces between aquifers and rivers or lakes have been identi ﬁ ed as biogeochemical hotspots with steep redox gradients. However, a detailed understanding of the spatial heterogeneity and potential temporal variability of these hotspots, and the consequences for nitrogen processing, is still hindered by a paucity of adequate measurement techniques. A novel methodology is presented here, using Di ﬀ usive Equilibrium in Thin- ﬁ lm (DET) gels as high-spatial-resolution passive-samplers of δ 15 N NO3 and δ 18 O NO3 to investigate nitrogen cycling. Fractionation of δ 15 N NO3 and δ 18 O NO3 during di ﬀ usion of nitrate through the DET gel was determined using varying equilibrium times and nitrate concentrations. This demonstrated that nitrate isotopes of δ 15 N NO3 and δ 18 O NO3 do not fractionate when sampled with a DET gel. δ 15 N NO3 values from the DET gels ranged between 2.3 ± 0.2 and 2.7 ± 0.3 ‰ for a NO 3 − stock solution value of 2.7 ± 0.4 ‰ , and δ 18 O NO3 values ranged between 18.3 ± 1.0 and


* S Supporting Information
ABSTRACT: The fate of nitrate transported across groundwater-surface water interfaces has been intensively studied in recent decades. The interfaces between aquifers and rivers or lakes have been identified as biogeochemical hotspots with steep redox gradients. However, a detailed understanding of the spatial heterogeneity and potential temporal variability of these hotspots, and the consequences for nitrogen processing, is still hindered by a paucity of adequate measurement techniques. A novel methodology is presented here, using Diffusive Equilibrium in Thin-film (DET) gels as high-spatialresolution passive-samplers of δ 15 N NO3 and δ 18 O NO3 to investigate nitrogen cycling. Fractionation of δ 15 N NO3 and δ 18 O NO3 during diffusion of nitrate through the DET gel was determined using varying equilibrium times and nitrate concentrations. This demonstrated that nitrate isotopes of δ 15 N NO3 and δ 18 O NO3 do not fractionate when sampled with a DET gel. δ 15 N NO3 values from the DET gels ranged between 2.3 ± 0.2 and 2.7 ± 0.3‰ for a NO 3 − stock solution value of 2.7 ± 0.4‰, and δ 18 O NO3 values ranged between 18.3 ± 1.0 and 21.5 ± 0.8‰ for a NO 3 − stock solution of 19.7 ± 0.9‰. Nitrate recovery and isotope values were independent of equilibrium time and nitrate concentration. Additionally, an in situ study showed that nitrate concentration and isotopes provide unique, high-resolution data that enable improved understanding of nitrogen cycling in freshwater sediments.
T he transport and transformation of nitrate across groundwater-surface water interfaces has been intensively studied over the past few decades, resulting in the identification of hotspots of increased biogeochemical turnover in these areas. 1−4 However, our understanding of the spatial patterns and temporal dynamics of nitrogen processing at the sediment interfaces between aquifers and rivers or lakes is still hampered by a critical lack of adequate monitoring methodologies. 5−9 In particular, there is a vital need for in situ data providing a more detailed insight into gradients of nutrient cycling at small spatial scales. 5 Isotopic data is particularly crucial as it is able to provide additional source and process information that concentration data alone cannot. 10,11 Such information is crucial for improving mechanistic process understanding of ecosystem functioning across spatial and temporal scales and to support integrated river and groundwater management and restoration so that freshwater systems are managed effectively. 12−15 A promising technological advancement has been the emergence of Diffusive Equilibrium in Thin-film (DET) gel samplers, to passively collect chemical constituents in water, soil, and sediment ( Figure 1). Besides a wide range of contaminants, DET gels have been applied to analyze vertical profiles of nitrate concentrations at high spatial resolutions of 1 cm, providing significant advantages over traditional sampling methods, such as multilevel piezometers. 14,16−19 Recently, this spatial resolution has been further improved to millimeter scale using colorimetry and hyperspectral imagery to obtain simultaneous nitrate/nitrite profiles. 20 The application of DET gels at groundwater-surface water interfaces supports the identification of discrete zones of concentrations of nitrate, nitrite, and ammonium, including the characterization of differing redox zones and hotspots of biogeochemical reactivity. 14 DET gels have been used recently to investigate coupled nitrification-denitrification and dissimilatory nitrate reduction to ammonium; however, no evidence was provided demonstrating there was no fractionation on diffusion of nitrate through the DET gel, and only δ 15 N NO3 was considered. 21,22 Here we present a new method, which combines the advantages of high-resolution sampling by DET technology with the analysis of nitrate isotope ratios to quantify nitrogen cycling at groundwater-surface water interfaces. Recognizing the limitations of inferring biogeochemical cycling and nutrient dynamics from concentration data alone, we propose the use of DET gels as a high resolution, in situ sampler, of nitrate isotopes in addition to concentration data. The measurement of δ 15 N NO3 and δ 18 O NO3 provides useful information on the processes controlling nitrate concentrations in hotspots of biogeochemical turnover in areas such as aquifer-lake or aquifer-river interfaces. 10 Additionally, the sources of nitrate measured may be identified, as nitrate from differing sources often has distinct isotopic values of δ 15 N NO3 and δ 18 O NO3 , enabling identification of nitrate sources affecting freshwater systems. 10 This combination of high-resolution sampling and process inference from tracer analysis provides significant potential for increasing our understanding of hotspots of biogeochemical turnover and differing redox zones in the hyporheic zone, therefore, allowing more effective management of freshwater systems.

■ EXPERIMENTAL SECTION
Laboratory experiments were performed to determine the potential for fractionation of δ 15 N NO3 and δ 18 O NO3 during the diffusion of nitrate through DET gels. Thereby, during laboratory experiments, the influences of two key controls were investigated for their impacts on fractionation; (1) the concentration of nitrate in the initial solution and (2) the time allowed for diffusive equilibrium of the nitrate from the initial solution into the DET gel. An initial proof of concept study was conducted using an isotope technique requiring 1 mg NO 3 − -N resulting in the requirement of high nitrate concentrations (up to 7.0 g NO 3 − L −1 ). Subsequently, a more environmentally relevant experiment was performed to verify the results, using an isotope technique requiring 0.7 μg NO 3 − -N, and therefore, much lower nitrate concentrations between 20.1 and 100.5 mg NO 3 − L −1 could be used. A field trial was then conducted to demonstrate the additional insight gained through high spatial resolution δ 15 N NO3 and δ 18 O NO3 data; in addition to nitrate concentration data, this was also conducted using the isotope technique requiring 0. − L −1 were used, with each concentration being equilibrated for three different time periods: 24, 48, and 168 h. A river sample spiked with nitrate (7.0 ± 0.0 g NO 3 − L −1 ) was also used to test for matrix effects; the river sample was collected from the River Tern, U.K. and filtered (0.2 μm). All experiments were performed in triplicate, and the concentration of the stock solution was compared to that of the solution in the beaker after the gel was removed. A concentration 97.7% that of the stock solution was expected due to the gel volume being 2.3% of the solution volume. The solution concentrations were found to be 95.0 ± 3.0% that of the stock (n = 12).
Back-Equilibration from DET Gels. At the end of the equilibration period DET gels were removed from solution and

Analytical Chemistry
Article weighed in a preweighed centrifuge tube to determine the weight of each gel. The solution volume of each gel was calculated from the weight multiplied by the assumed water content of the saturated gel (95%). Twenty-five ml of ultrapure water was added to each gel, and the gels were shaken on a reciprocating shaker for 24 h, after which the gels were removed and the back-equilibrated samples were frozen for chemical analysis. Nitrate concentrations were determined using ion chromatography (Dionex ICS1100); standards were used as quality controls and gave an accuracy of 0.4 mg L −1 , precision of ±0.4 mg L −1 , and a limit of detection of 0.5 mg L −1 .
Laboratory Experiment at Environmentally Relevant Concentrations. The proof of concept experiment outlined above was repeated to investigate isotope fractionation at environmentally relevant nitrate concentrations, using an equilibration time of 24 h. Solutions of 20.1 ± 0.0, 50.8 ± 0.2, and 100.5 ± 0.3 mg NO 3 − L −1 were used, as well as a filtered (0.2 μm) river sample (Wood Brook, Mill Haft, U.K.) with a concentration of 23.1 ± 0.0 mg NO 3 − L −1 . During backequilibration 20 mL of ultrapure water was added to each gel, and nitrate concentrations were determined on a Continuous Flow Analyzer (Skalar Sans++). Standards were used as quality controls and gave an accuracy of 0.4 mg NO 3 − L −1 , precision of ±0.1 NO 3 − L −1 , and a limit of detection of 0.9 mg NO 3 − L −1 . The concentration of the stock solution was compared to that of the solution in the beaker after the gel was removed. A concentration 97.7% that of the stock solution was expected due to the gel volume being 2.3% of the solution volume, and solution concentrations were found to be 96.8 ± 1.1% that of the stock (n = 4).
In Situ Field Trial. Field Trial Study Site. For a field trial, proving the concept of DET isotope analysis in sediments, gel probes were deployed at the Urban River Laboratory (URL)  (Figure 1). The deployment period exceeded estimated exposure times required to ensure concentration equilibrium by diffusion in order to account for resettling of any potential sediment disturbances during the probe deployment. Gel 1 was deployed closest to the inflow: 3.86 m from the beginning of the flume. Gel 2 was deployed 1.15 m downstream of gel 1, and gel 3 was deployed 3.19 m downstream of gel 2 and 3.80 m from the end of the flume ( Figure 1).
All gels were extracted from the sediment within 10 min and immediately sliced at 2.5 cm intervals within 40 min. The DET gels were sliced (ultrapure water-rinsed blade) on an acidwashed (10% HCl) chopping board. Once sliced, the gels were placed into 50 mL centrifuge tubes and stored at 4°C.
Nitrate Elution from DET Gel Probes. The DET gels were back-equilibrated, on ice, with 6 mL of ultrapure water on a reciprocating shaker for 24 h. Once equilibrated the gels were removed and weighed, the eluate was then filtered (0.2 μm), and the eluate was frozen for later analysis. Samples were analyzed for nitrate concentration on a continuous flow analyzer (Skalar San++); standards were used as quality controls and gave an accuracy of 0.1 mg NO 3 − L −1 , precision of ±0.1 NO 3 − L −1 , and a limit of detection of 0.9 mg NO 3 − L −1 . Isotope Analysis. For the laboratory proof of concept experiments the nitrate was extracted from the samples using anion and cation exchange columns and converted to silver nitrate using the method in Chang 23 and Heaton 24 or a modified version of this as subsequently described. For river samples with the presence of interfering anions, the method by Chang 23 and Heaton 24 was used; for the pure NO 3 − solutions, the samples were not passed through anion and cation exchange columns. Instead the nitrate was converted to silver nitrate, and the above method was used from the point of adding the first batch of Ag 2 O. The silver nitrate was analyzed by mass spectrometry as in Heaton. 24 The international isotope reference materials used for δ 15 N were IAEA-N-1 and IAEA-N-2, with δ 15 N values vs air of +0.4 and +20.4‰, respectively, with a measurement precision of ±0.3 and ±0.5‰, respectively. The international isotope reference materials used for δ 18 O were IAEA-NO 3 , USGS-34, and USGS-35, with δ 18 O vs SMOW of +26.0, −28.0, and +56.4‰, respectively, with a measurement precision of ±1.2, ±1.7, and ±1.9‰, respectively. Analysis was performed at the NERC Isotope Geoscience Laboratory, British Geological Survey.
For the laboratory experiment at environmentally relevant concentrations and the field study data, the denitrifier method was used as this requires a lower mass of nitrate for analysis (0.7 μg NO 3 − -N). 25,26 This method utilizes denitrifying bacteria to convert sample nitrate to N 2 O, with a long-term measurement precision of ±0.3 and ±0.4‰ and an accuracy of 0.0 and 0.0‰ for δ 15 N NO3 and δ 18 O NO3 , respectively, and a measurement limit of 2 μM NO 3 − . The international isotope reference materials used were IAEA-NO 3 , USGS-34, and USGS-35, with δ 15 N of +4.7, −1.4, and +3.4‰, respectively, with a measurement precision of ±0.3, ±0.6, and ±0.5‰, respectively, and δ 18 O of +25.7, −28.0, and +57.4‰, respectively, with a measurement precision of ±0.7, ±0.6, and ±0.6‰, respectively. Analysis was performed by the Analytical Facilities, University of East Anglia.
■ RESULTS AND DISCUSSION Laboratory Experiments. Nitrate Concentrations from DET Gels. Nitrate concentrations were recovered from the DET gels with ranges between 3.6 ± 0.1 and 3.7 ± 0.1 for a 3.3 ± 0.0 g NO 3 − L −1 stock solution, between 5.2 ± 0.3 and 5.3 ± 0.1 for a 4.8 ± 0.0 g NO 3 − L −1 stock solution, between 6.7 ± 0.1 and 6.9 ± 0.0 for a 7.0 ± 0.1 g NO 3 − L −1 stock solution, and 6.3 ± 0.0 and 6.8 ± 0.1 for a 7.0 ± 0.0 g NO 3 − L −1 spiked river sample (Table S- were all between 90.4 and 112.1%, and we believe the differences to come from varying dilution factors between the lowest and highest concentrations used, as solutions of 7.0 g NO 3 − L −1 were diluted by the same factor, which was twice that of the solutions of concentration 3.3 and 4.8 g NO 3 − L −1 , which were also diluted by the same factor, to allow for machine analysis. The nitrate recoveries observed were, therefore, independent of initial solution concentration. This was supported by the independence of recovery and initial solution concentration when performing the environmentally relevant concentration experiments (Table S- The independence from equilibration time evidenced that as long as equilibrium is reached, the time the gel is left in solution should not affect nitrate recovery, and that the DET gels were expected to equilibrate by 24 h. When utilizing the DET gels in situ the deployment time should be longer than the equilibrium time required for the gel thickness used. This is because the natural conditions of the system need to be re-established after gel deployment. Field deployment times for DET gels of 72 h have been recommended previously. 14 The high nitrate concentrations used in these experiments (up to 7.0 g NO 3 − L −1 ) resulted from practical limitations of the isotope analysis method used in the proof of concept study, which required a minimum of 1 mg of NO 3 − -N. As evidenced by our results, these high concentrations did not prevent the reaching of equilibrium by diffusion into the gel. This proves that DET gels can also be applied in high nitrate conditions (e.g., artificial wetlands, wastewater treatment plant outputs), as the recovery of nitrate was not dependent on the solution concentration. It is acknowledged that the large nitrate concentrations used here, due to method limitations, are much higher than those found in most natural environments; therefore, the experiment was repeated with environmentally relevant concentrations as discussed above.
Nitrate Isotope Ratios in DET Gels. δ 15 N NO3 . The δ 15 N NO3 for the nitrate solutions were the same as that of the stock solution, within error, with a range between 2.3 ± 0.2‰ and 2.7 ± 0.1‰, compared with 2.7 ± 0.4‰ of the stock ( Figure  2). The nitrate-spiked river water had a different δ 15 N NO3 value, 2.3 ± 0.5‰, than that of the nitrate solutions; this was expected as distinct sources of nitrate have different δ 15 N NO3 and δ 18 O NO3 values. The δ 15 N NO3 for the river solutions equilibrated for 24 and 48 h were found to be the same as the stock, within error, with δ 15 N NO3 of 1.8 ± 0.7‰ and 1.7 ± 0.1‰, respectively. δ 15 N NO3 for 168-h equilibrium time was lower than the stock with a value of 1.5 ± 0.1‰; however, this was not a statistically significant difference (paired t test, p-value = 0.16).
δ 18 O NO3 . The δ 18 O NO3 of the nitrate solutions were predominantly found to be the same as that of the stock solution, within error, with a range between 18.3 ± 1.0‰ and 20.8 ± 1.0‰, compared with 19.7 ± 0.9‰ of the stock (Figure  This is believed to be due to the unusually small measurement error in the stock solution (0.2‰), as the lowest value the 24-h sample could be is 20.8‰ and the highest value the stock could be is 20.4‰, which are similar values and would not otherwise be interpreted as having fractionated. This is evidenced by the uncertainty found in the IAEA-NO 3 standard of 0.85 and 1.46‰ (from multiple analyses), showing that two results within this range cannot be distinguished using this technique.
Determination of Fractionation. The analysis of δ 15 N NO3 and δ 18 O NO3 revealed no significant difference between the values of the stock solution and those of the gel solutions. We, therefore, found no evidence of fractionation during the process of nitrate diffusion into and out of the gel during equilibrium and elution. This also demonstrates that solution concentration and equilibrium time did not affect δ 15 N NO3 and δ 18 O NO3 (Figure 2). There was also no relationship between the δ 15 N NO3 and δ 18 O NO3 of the gel solution and nitrate recovery, with fitted linear models for the KNO 3 for δ 15  4‰. These were found to be significantly different than the stock (paired t test, p-value = 0.022 and 0.000, respectively, for 20.1 and 50.8 mg NO 3 L −1 ); however, this is believed to be due to the small standard deviation of sample replicates (n = 3), of 0.2, 0.0, and 0.0‰, respectively, for 20.1 and 50.8 mg NO 3 L −1 , and the stock. Given that the difference between the highest value the stock ratio could be and the lowest value the solution ratio could be is only 0.2 for both solutions, this would not usually be considered fractionated. This is evidenced by the long-term reproducibility of the isotope technique, which is ±0.4‰, showing that two results within this range cannot be distinguished using this technique. The river sample had a δ 15 N NO3 value of 10.3 ± 0.3‰ for a stock of 10.2 ± 0.3‰ and a δ 18 O NO3 value of 4.7 ± 0.4‰ for a stock of 4.2 ± 0.4‰.
The nitrate-spiked river samples were used to test for any matrix effects, which may affect δ 15 N NO3 and δ 18 O NO3 values when this method is utilized in situ. There was no fractionation of δ 15 N NO3 and δ 18 O NO3 in the river water samples, confirming the applicability of this method in environments where interfering ions are present.
Concentration and Isotope Analysis from in Situ DET Application. Three example profiles of in situ DET sampling are discussed as proof of concept for the proposed combined DET-isotope methodology. The DET gels captured a large range of nitrate concentrations and δ 15 N NO3 and δ 18 O NO3 values in a 15 cm profile (Table S-2, Figure 3), with all nitrate concentrations from the DET gels (18.6 ± 0.1 to 82.5 ± 0.9 mg NO 3 − L −1 ) greater than that of the average inflow concentration of nitrate (13.4 ± 0.7 mg NO 3 − L −1 ). The largest concentration range was observed in gel 2, with a minimum of 20.0 ± 0.1 mg NO 3 − L −1 at 11 cm depth to a maximum of 82.5 ± 0.9 mg NO 3 − L −1 at 1.25 cm depth. The largest range of δ 15 N NO3 and δ 18 O NO3 was shown in gels 1 and 2, respectively. δ 15 N NO3 values in gel 1 ranged from 0.2 ± 0.3‰ at 11 cm depth to 17.9 ± 0.3‰ at 3.75 cm depth, and δ 18 O NO3 values in gel 2 ranged from −9.9 ± 0.4‰ at 1.25 depth to 9.3 ± 0.4‰ at 6.25 cm depth. These results highlight the ability of DET gel-based passive sampling to capture hotspots of biogeochemical activity and spatially small redox zones, which would be missed with more traditional methods i.e. multilevel piezometers, as was found previously with nitrate concentrations. 14 This is particularly shown in profile 1, where there appears to be an area of denitrification at 3.75 cm, indicated by low nitrate concentration combined with high δ 15 N NO3 and δ 18 O NO3 values. Concentrations at all depths in gel 1 vary over a small range between 37.3 ± 0.4 and 47.4 ± 0.4 mg NO 3 − L −1 , except for at a depth of 3.75 cm, where the concentration has decreased to 18.6 ± 0.1 mg NO 3 − L −1 . This is also reflected in the isotopic data where δ 15 N NO3 varies over a small range between 0.2 ± 0.3‰ and 3.3 ± 0.3‰, except at 3.75 cm, where the δ 15 N NO3 value has increased to 17.9 ± 0.3‰, and δ 18 O NO3 varies over a small range between −2.3 ± 0.4‰ and −3.5 ± 0.4‰, except at 3.75 cm where the δ 18 O NO3 value has increased to 11.4 ± 0.4‰.
The analysis of vertical profiles of nitrate isotope ratios and concentrations indicates differences in concentration patterns at the three locations ( Figure 3). Gel 1 shows slightly higher concentrations at greatest depths (43.4 ± 0.5 to 47.4 ± 0.4 mg NO 3 − L −1 ) compared to the shallowest depth (37.3 ± 0.4 mg NO 3 − L −1 ), with a local minimum of 18.6 ± 0.1 mg NO 3 − L −1 at 3.75 cm. This could be due to zones of nitrification associated with the mineralization of nitrogen from the macrophytes. 14,27−29 The δ 15 N NO3 and δ 18 O NO3 profiles show little variation with depth, only varying between 0.2 ± 0.3‰ and 3.3 ± 0.3‰ and −2.3 ± 0.4‰ and −3.5 ± 0.4‰, respectively, except at 3.75 cm where there is a large increase in δ 15 N NO3 to 17.9 ± 0.3‰ and δ 18 O NO3 to 11.4 ± 0.4‰. Although similar, δ 15 N NO3 values between 6.25 and 11 cm depth do decrease with depth, perhaps pointing to the onset of denitrification at 6.25 cm depth. The combination of low nitrate concentration, with high δ 15 N NO3 and δ 18 O NO3 values, is indicative of denitrification; therefore, there appears to be a localized zone of denitrification at 3.75 cm. 30 In gel 2 an overall decrease in nitrate was observed at greater depths than in gel 1, although there was a concentration of 46.9 ± 0.3 mg NO 3 − L −1 at 14.25 cm, which was intermediate of the concentrations found at 1.25 and 3.75 cm (82.5 ± 0.9 and 72.4 ± 0.7 mg NO 3 − L −1 , respectively) and 6.25 to 11 cm (20.0 ± 0.1 to 22.4 ± 0.2 mg NO 3 − L −1 ). The δ 15 N NO3 values showed little variation between the shallowest and the largest depth; δ 15 N NO3 was slightly higher at the greatest depth than at the

Analytical Chemistry
Article shallowest depth, with values of 5.1 ± 0.3‰ and 2.7 ± 0.3‰, respectively, and the lowest δ 15 N NO3 value of 0.9 ± 0.3‰ was found at 3.75 cm depth. The shallowest depth δ 18 O NO3 values were also lower than at the greatest depth; however, the difference from −9.9 ± 0.4‰ to −0.9 ± 0.4‰ was greater than seen for δ 15 N NO3 . A substantial increase in δ 15 N NO3 values was observed between 6.25 and 11 cm depth, with values between 16.8 ± 0.3 and 16.9 ± 0.3‰. This large increase was also present in the δ 18 O NO3 profile, where peak concentrations ranged between 8.6 ± 0.4‰ and 9.3 ± 0.4‰, between 6.25 and 11 cm. In combination with δ 15 N NO3 and δ 18 O NO3 values, the concentrations observed at these depths indicate denitrification.
Vertical variation of nitrate concentrations in gel 3 seems to be minimal, with a narrow concentration range of 22.2 ± 0.2 to 33.1 ± 0.2 mg NO 3 − L −1 for the whole 15 cm gel. It is worth noting that the maximum concentration found in gel 3 is at 1.25 cm, which is the same as is found in gel 2. The δ 15 N NO3 and δ 18 O NO3 values in gel 3 did not cover as wide a range as in gels 1 and 2; the range in δ 15 N NO3 and δ 18 O NO3 here was just 11.6 ± 0.3 to 16.8 ± 0.3‰ and 4.1 ± 0.4 to 12.4 ± 0.4‰, respectively. δ 15 N NO3 values showed little variation ranging from 11.6 ± 0.3 to 14.2 ± 0.3‰, at all depths except 3.75 and   18 O NO3 is also useful in identifying areas of denitrification, a ratio between approximately 2.1 and 2.5 is considered indicative of denitrification. 10 In gel 3 this ratio is found at depths of 3.75, 6.25, 11.25, and 13.75 cm, where the δ 15 N NO3 :δ 18 O NO3 is 2.3, 2.1, 2.2, and 2.4, respectively, providing further evidence of denitrification. This denitrification throughout the profile is likely related to an overall increase in denitrification toward the downstream end of the flume as the residence time of the porewater and nitrate increased, and due to the cumulative effect of vegetation described previously.
Generally, enhanced denitrification appears to be correlated with the occurrence of vegetation in the flume, which has been previously observed to particularly affect depths between 5 and 12 cm. 27,28 Possible mechanisms by which vegetation enhances sediment denitrification can be of biotic or abiotic nature, generally leading to high biogeochemical reactivity. 31 These include, uptake by macrophytes, increased surface water downwelling, and enhanced residence times of water in the sediment that are facilitated by vegetation and may, therefore, have led to the increased denitrification seen here. 3,28 Similar zones of vegetation-associated denitrification were found in the River Leith, U.K. 28  Horizontal patterns along the flume indicate a general trend of denitrification with an increased observation of low nitrate concentration samples from gel 3 through to gel 1, combined with an increased frequency of high δ 15 N NO3 and δ 18 O NO3 values. This longitudinal profile is overlain by local effects, where hotspots of biogeochemical reactivity can be seen, thought to be influenced by the vegetation effect described previously. Comparing the gel data to the nitrate concentration of the inflow shows an increase in nitrate from the inflow to the subsurface water of the flume, indicating net nitrification within the flume itself. This is the opposite of the denitrification trend shown by the gel profiles and seems to show high nitrification within the flume, even upstream of the first gel, which increased nitrate concentrations before these decreased again through denitrification.
The in situ results of the DET gel and isotope method presented here have allowed the investigation of detailed processes at a spatial scale, which exceeds that of previous studies. Particularly, hotspots of denitrification were easily identified using both concentration and isotope data. The isotopic data were invaluable in showing that the gel profiles indicated generally high nitrate concentrations, with zones of denitrification leading to low nitrate concentrations. This is in contrast to the concentration data alone, which, along with the inflow concentration, indicates varying degrees of nitrification in the gel profiles. This is increasingly important in the study of nutrient fate at aquifer and river or lake interfaces; research areas that are often limited by a lack of sufficient monitoring methods. Thus, this field trial has successfully demonstrated the value of this new approach for in situ applications.
To assess natural waters in which this DET-isotope technique could be applied, the concentration of nitrate in porewaters needed to provide a solution of 50 nmol NO 3 − required to perform the isotope analysis was calculated. Using a 1.56 mm thick DET gel and slicing at 1 cm would require a porewater nitrate concentration of 10.5 mg L −1 , and slicing at 2.5 cm would reduce this to 4.2 mg NO 3 − L −1 . The nitrate concentrations required limit this technique to sediments in nonpristine environments. 3,32 ■ CONCLUSIONS The laboratory proof of concept demonstrates that δ 15 N NO3 and δ 18 O NO3 do not fractionate when sampled with a DET gel. Nitrate recovery and δ 15 N NO3 and δ 18 O NO3 values were independent of both equilibrium time and nitrate concentration, suggesting the applicability of DET technology for sampling isotope ratios from sediment pore-water at high spatial resolution.
Additionally, the in situ application of DET gel probes in a field trial provides evidence of the potential of this methodology to sample nitrate concentration and isotopic data with DET technology in the field at higher resolution than previously possible.
Based on the promising results of the presented lab and field trials we recommend the application of this combined methodology at aquifer-river and aquifer-lake interfaces in order to enhance mechanistic process understanding of hotspots in nitrogen cycling. Future research may elaborate to what degree the application of the proposed methodologies can be extended also to brackish and marine systems. analysis and data interpretation, Dr. Eugeǹia Martíat CEAB and the Urban River Laboratory for allowing us to use their facilities and data, and Dr. Hao Zhang of DGT Research and Lancaster Environment Centre, University of Lancaster, for providing the DET gels and advice on the gel's capabilities. We would also like to thank the Natural Environment Research Council (NERC) and the Central England NERC Training Alliance for their funding of this project, as well as the NIGL for providing an in-kind grant for isotopic analysis.