COUPLED NITROGEN AND OXYGEN ISOTOPE STUDY OF NITRATE AT A RURAL UNLINED LANDFILL IN WESTERN VICTORIA, AUSTRALIA

Understanding of site-scale physical and chemical heterogeneities will inform remediation strategies for unlined rural landfills worldwide. The aim of this study was to characterize potential sources of nitrogen contamination to groundwater at an unlined landfill in rural western Victoria, Australia. Results revealed simultaneously high concentrations of both nitrate and ammonia within several wells, indicative of heterogeneous redox potentials within the unconfined aquifer. Combined isotopic analyses of δ 15 N and δ 18 O in nitrate identified a leachate-derived source and active denitrification across two sites hydraulically downgradient from the landfill cell. Groundwater at an up-gradient site, as well as nearby surface water samples from a riparian creek, reflected primarily the isotopic signature of agriculturally-derived nitrate with overprinting from atmospheric exchanges of nitrogen and oxygen. Nitrate and ammonia in the creek were interpreted to result from a mixture of leachate and agricultural sources. Results illustrate how redox gradients across a landfill leachate plume impact biogeochemical nitrogen cycling.


INTRODUCTION
Effective remediation of groundwater contamination by leachate from unlined (and possibly decommissioned) landfill sites requires site-scale hydrogeological characterization (Tiller, 1992;McLay et al., 2001;Allen, 2001). Groundwater flow direction and chemistry can influence the mobility and fate of leachate contaminants (Poeter and Gaylord, 1990;Assmuth and Strandberg, 1994). Variations in hydraulic conductivity in a sedimentary aquifer, e.g., due to changes in lithology (Genereux and Guardiario, 2010), can induce redox gradients that control the distribution of microbial consortia and thus the potential for contaminant biodegradation (Bjerg and Christensen, 1992;Jankowski and Beck, 2000). Understanding of site-scale physical and chemical heterogeneities is therefore important for effectively harnessing microbial activity for bioremediation (Cozzarelli et al., 2011), particularly at unlined rural landfills, a common historical type of landfill worldwide (e.g., USGS, 2003). Of particular interest is how nitrogen-bearing contaminants (i.e., nitrate or ammonia) can be transported or transformed via a number of physical and biogeochemical pathways. Various approaches exist for understanding the distribution of potential electron donors for denitrification in leachate plumes (Cozzarelli et al., 2000).
This study investigated the origin and distribution of nitrate and ammonia in groundwater receiving leachate from a rural unlined landfill site in western-central Victoria, Australia. The lack of published data on contaminated rural unlined Australian landfills (WCS, 2010) was a primary motivator for this study and the field site for this study provided a typical example of an unlined rural Australian landfill with irregular historical waste collection records. The site is located in a region of Victoria where agricultural nitrate contamination of groundwater was historically of concern and is adjacent to a protected riparian zone. Our study focused on understanding the distribution and type of nitrogen Science Publications AJES species within the unconfined aquifer receiving leachate from the landfill cell. Based on previous studies at other sites (Cozzarelli et al., 2011) and the lithological heterogeneity of the unconfined aquifer reported in core logs, we hypothesized that nitrogen redox transformations may vary considerably throughout the leachate plume. Although many studies have been conducted of nitrate contamination of groundwater by agricultural or waste water sources, as expressed in coupled nitrogen and oxygen isotope ratios (in nitrate), a paucity of data exists for rural unlined landfills. Here we applied field-based spectrophotometry of redox-sensitive chemical species and coupled nitrogen and oxygen isotope ratio analyses of nitrate in groundwater, to differentiate potential nitrate sources and delineate the extent of denitrification associated with a rural unlined Australian landfill.

MATERIALS AND METHODS
The Beaufort Transfer Station (BTS) is a rural landfill in western-central Victoria ( Fig. 1) that lacks a liner and leachate collection system. The landfill cell sits approximately 4-5 meters in topographic elevation above the surrounding flood terrace sedimentary deposits to the northeast and is thought to be approximately 6-7 meters in depth, although the accuracy of historical records has been questioned. The landfill therefore essentially slopes downward towards the northeast across the scale of the entire site.
Investigations were conducted at the BTS in 2009-2010; groundwater samples were collected from three usable boreholes (MB01, MB02 and MB04), which penetrated the unconfined aquifer to shallow depths of 3, 4.5 and 4.5 meters, respectively (Fig. 2). Screens for these bores were 1.5, 3 and 3 meters long, respectively and opened to mainly clay and silt (MB01, MB04) or sands and paleo-stream gravels (MB02). Borehole MB03 was partially collapsed and could not be used. A single borehole penetrating to an unknown depth and with an unknown screen length within the landfill cell was sampled as well, although this well was considered, by visual inspection, to be poorly installed and maintained. This well consisted of a simple PVC pipe casing with no cap, no apparent annular seal or well packing of any kind and no sealing of the borehole at the surface, compared to properly installed wells at MB01, MB02 and MB04.
Historical standing water level data for the wells (Supplementary Fig. 3) show that the general groundwater flow direction is from MB04, located up-gradient hydraulically of the landfill cell, towards MB02 and MB01 either due northward or northeastward, depending on seasonal or annual recharge rates. Samples were collected with a bailer, after purging of two well volumes. Boreholes were not purged dry in order to prevent interference of field spectrophotometric assays by turbidity. Bailers were rinsed with decontaminant solution and ultrapure (Milli-Q) water on site before use.
Groundwater and surface water chemistry measurements (pH, EC, alkalinity and Eh) were obtained in the field ( Fig. 4 and Table 1). Three-point calibration was used on an Orion 5-Star meter with automatic temperature correction for pH and EC; Eh was measured using an Orion 250A meter and redox probe and values were corrected to the standard hydrogen electrode. Total alkalinity (calculated as mg/L of CaCO 3 ) was measured using a HACH digital titrator with 100ml of groundwater or surface water titrated using 0.16 N sulphuric-acid to a pH value of 4.35 using the inflection endpoint method. All groundwater or surface water samples were filtered through a 0.45 µm pore-size filter, with groundwater samples for cations and trace metals acidified with trace metal grade HNO 3 to a pH <2. Sulfate, nitrate, chloride, nitrite and ammonia were measured using a DR2800 HACH Spectrophotometer in the field and in the lab. \Nitrate, ammonia and nitrite analyses were conducted on site using the appropriate colorimetric assays for both groundwater and surface waters. Each assay was analyzed with a reference standard matrix designed to correct for Cl − (>1000mg L −1 ) in the groundwater. Major cations and trace metals were analyzed using ICP-AES (ALS Laboratory Group, Springvale, VIC, Australia) and major anions were analyzed using IC at the Australian National University. Sampling bottles were rinsed with filtered sample water before sample collection. Field blanks were used to quantify any contamination during sample handling and transport.
Sampling and processing of groundwater and surface waters for δ 15 N and δ 18 O followed the protocol of the Colorado Plateau Stable Isotope Laboratory (CPSIL; Flagstaff, AZ, USA) (Sigman et al., 2001). Analyses were performed using the Gas Bench Precon isotope ratio mass spectrometer (GB-Precon IRMS). Samples were collected in triplicate in acid-washed Nalgene bottles that were triple-rinsed with nanopure (Milli-Q) water. Two sets of triplicates were collected for each sampling site; one set of 100 ml each for total nitrate concentration and the other set of 100 ml each for isotopic analysis. Each sample was filtered in the field with 0.22 µm pore-size filters and immediately placed on ice.
Groundwater sampled from within the landfill cell showed the highest concentrations of dissolved iron and nitrate (2.5 and 34 mg L −1 , respectively), but the lowest concentrations of chloride and sulfate (240 and 7 mg L −1 , respectively).

Fig. 4. Plot of nitrogen and oxygen isotope ratios in nitrate sampled from monitoring wells at BTS landfill and nearby Yam Holes
Creek sites. Data are superimposed on the schematic figure from (Clark and Fritz, 1997) Science Publications

AJES
Surface water samples along the Yam Holes Creek showed a substantial variation along its long profile. Background surface water (YH01) had moderate nitrate concentrations (1.9 mg/L) and low concentrations of ammonia (.01 mg/L) and HCO 3 − (93 mg/L). Surface water sites closer to the BTS, however, at YH02 and YH04 exhibited NO 3 − concentrations above freshwater guidelines (10.4 mg/L and 10.9 mg/L respectively). Still, elevated nitrate concentrations were not consistent along the down-gradient portion of the creek as YH03 showed very low concentrations of nitrate (.46 mg/L), ammonia (.04 mg/L) and HCO 3 − (35 mg/L). Figure 5 shows δ 18 O and δ 15 N values of nitrate sampled in groundwater and surface waters, plotted alongside typical δ 18 O and δ 15 N values for various sources of nitrate defined by (Clark and Fritz, 1997 and MB01 and YH01-YH04 with MB04, yielded isotope ratio data that plotted towards increasingly heavier (more positive) isotopic ratios for both nitrogen and oxygen.

DISCUSSION
The increases in alkalinity and pH, with a decrease in reduction potential (Eh), from MB04 to MB02 or MB01, are all consistent with an input of bicarbonate from microbially-mediated anaerobic degradation of organic matter within the landfill cell (Brad et al., 2008). The presence of both nitrate and ammonia in the leachate bore and MB02 and to some degree in MB04, suggests that high spatial heterogeneity in redox potential of groundwater in the unconfined aquifer was reflected by the chemistry sampled in these wells. In other words, groundwater from wells MB02 and MB04 (setting the leachate bore aside for reasons discussed above) likely represents a mixture of reduced and oxidized nitrogen species from heterogeneous "pockets" of the unconfined aquifer with varying redox potentials that has not yet reacted with atmospheric oxygen within the wells. We infer that groundwater from MB02 and MB04 may reflect the presence of heterogeneous zones of microbial denitrification and ammonia oxidation within the aquifer, presumably separated and mediated by different populations within the aquifer microbial community structure. Also, the presence of paleo-stream gravels with higher hydraulic conductivity in MB02 serves to focus meteoric recharge and locally increase groundwater flow rates, resulting in more oxic conditions, shorter pore water residence times and less chemical potential for complete denitrification.
For detailed discussion of the placement of fields describing the nitrogen and oxygen isotopic compositions of various possible nitrate sources and the changes in these ratios that would result from denitrification, we refer the reader to the original source from which we derived the (modified) Fig. 4 (Clark and Fritz, 1997). Isotopic ratios for wells MB01 (δ 15 N = +26% o , δ 18 O = +9% o ) and MB02 (δ 15 N = +22% o , δ 18 O = +6% o ), located down gradient of the landfill cell, exhibited values associated with manure and/or septic sources of nitrate (Fig. 4). This finding is consistent with the idea that these two wells groundwater contaminated with landfill leachate-derived nitrate and organic compounds. We could not resolve, with the data obtained in this study, a significant difference in the isotopic signatures of landfill leachate-derived and septic-derived nitrate, a result similar to those of (Aravena et al., 1993;Aravena and Robertson, 1998 (Aravena and Robertson, 1998) (Table 1) with the exception that our δ 18 O ratios possibly reflect greater effects from evaporation.
Linear trends in the isotope ratio data reflect the increasing extent of denitrification detected in both groundwater and surface waters. However, waters from Yam Holes Creek contained the N and O isotopic signatures of a different nitrate source from landfill leachate. Therefore, although the YH02 and YH04 sites appear to have elevated nitrate levels more similar to MB02 than to YH01 or YH03, some agriculturally derived nitrate (i.e., from synthetic or natural fertilizer) also contributes to the δ 15 N and δ 18 O values of surface waters. We hypothesize that Science Publications AJES nitrate contamination from the landfill is being effectively channeled along near-surface (<5 meters depth) paleostream channel gravels and has possibly already delivered elevated nitrate levels to portions of Yam Holes Creek. Recent conversion of local blue gum plantations to the north and west of Beaufort to canola farming (http://www.abc.net.au/news/2013-06-05/blue-gumdemise/4735044), with consequent increased fertilizer requirements (Hocking et al., 1997), may also have contributed to the observed isotope ratios of surface waters. Future research should focus on testing the hypothesis, or further substantiating evidence, that landfill leachatederived nitrate has already impacted the local surface riparian ecosystem in and around Yam Holes Creek.

CONCLUSIONS
The findings of this study reveal the extent of nitrogen transformations and biogeochemical cycling across a landfill leachate plume in a lithologically heterogeneous and shallow unconfined aquifer. Cooccurrence of nitrate and ammonia in some wells supports the interpretation that significant heterogeneity in redox potential exists, most likely as a result of the degree of lithological variability (i.e., variations in hydraulic conductivity). These lithological variations are primarily the result of buried paleo-stream channels typical of riverine plains across much of the Australian Murray-Darling Basin. Coupled nitrogen and oxygen isotope ratio analyses allowed the resolution of disparate nitrate sources in groundwater impacted by landfill leachate.