Environ. Sci. Technol. 2005, 39, 539-548
Tracking the Sources of Nitrate in Groundwater Using Coupled Nitrogen and Boron Isotopes: A Synthesis D A V I D W I D O R Y , * ,† EMMANUELLE PETELET-GIRAUD,† P H I L I P P E N EÄ G R E L , † A N D BERNARD LADOUCHE‡ BRGM, 3 Avenue Claude Guillemin, BP 6009, 45060 Orle´ans, Cedex 2, France, and BRGM, 1039 Rue de Pinville, 34000 Montpellier, France
Nitrate (NO3) is one of the world’s major pollutants of drinking water resources. Although recent European Directives have reduced input from intensive agriculture, NO3 levels in groundwater are approaching the drinking water limit of 50 mg L-1 almost everywhere. Determining the sources of groundwater contamination is an important first step toward improving its quality by emission control. It is with this aim that we review here the benefit of using a coupled isotopic approach (δ15N and δ11B), in addition to conventional hydrogeological analyses, to trace the origin of NO3 in water. The studied watersheds include both fractured bedrock and alluvial (subsurface and deep) hydrogeological contexts. The joint use of nitrogen and boron isotope systematics in each context deciphers the origin of NO3 in the groundwater and allows a semiquantification of the contributions of the respective pollution sources (mineral fertilizers, wastewater, and animal manure).
impossible to discriminate between multiple NO3 sources based solely on their N isotopic composition whenever there is heterogenic or autogenic denitrification; therefore, the need for co-migrating discriminators of NO3 sources arises. Due to its ubiquitous nature, boron (B) is a minor groundwater constituent (13). The wide range of B isotope ratios observed in nature seems to indicate that significant contrasts between B sources in groundwater are possible (14). Previous studies of B isotopes as tracers of human impact on water resources have focused on the identification of wastewater and sewage dominated by synthetic B products (15) and on the impact of fly ash leachate (16). Bassett et al. (17) and Komor (18) were the first to use B isotopes as comigrating tracers of NO3, but only Komor (18) and Widory et al. (19) report the B isotope signatures of input from agriculture (e.g., hog manure, cattle feedlot runoff, synthetic fertilizers) and combine N and B isotopes in order to distinguish between different NO3 sources in groundwater and surface water. The coupled use of geochemical and isotopic tracers (N, B) provides a sensitive method for tracing sources of NO3 in contaminated groundwater. N isotopes, being an intrinsic tracer of the NO3 molecule, reflect both the sources and the fate (e.g., denitrification) of NO3 in groundwater. B isotopes, because they are not affected by denitrification (20), bear the signature of the solute sources but may nevertheless fractionate through processes such as adsorption on clay minerals (15). Studied were carried out in three different hydrogeological contexts in Francesfractured bedrock (Arguenon, Brittany), deep alluvial groundwater (Pia, Pyrenees), and subsurface alluvial groundwater (Ile du Chambon, Allier)sin order to determine whether, regardless of the hydrogeological context, the combined use of nitrogen and boron would allow both the identification of pollution sources and the semiquantification of their respective contributions to the observed NO3 levels in groundwater.
Materials and Methods Introduction Nitrate (NO3), naturally present at moderate concentrations in groundwater (around 10 mg L-1), is often greatly enriched by anthropogenic activities that involve nitrogenous compounds (e.g., the spreading of mineral fertilizers) and byproducts of organic compounds from agriculture, septic systems, and animal manure (1-5). Despite increasing efforts at the national and European (EC Directive 91/976/EEC) levels to reduce NO3 input from intensive agriculture, NO3 is still one of the world’s major groundwater and surface water contaminants. The isotopic composition of dissolved nitrogen (N) has been used extensively to better constrain the sources and fate of N in groundwater (6). However, the possibility of quantifying both the origin of and secondary processes affecting N concentrations by means of a single tracer appears to be more limited. Nitrogen cannot be considered conservative because it is biologically modified by nitrification and denitrification, both during infiltration and within the aquifer, causing isotopic fractionation, which modifies the nitrogen isotope signatures of the dissolved N species (3, 7-12). It is * Corresponding author: e-mail:
[email protected]; fax: +33 (0)2 38 64 37 11. † BRGM, Orle ´ ans. ‡ BRGM, Montpellier. 10.1021/es0493897 CCC: $30.25 Published on Web 12/04/2004
2005 American Chemical Society
Groundwater samples were collected after pumping a volume corresponding to three times that of the well or borehole. They were then filtered through a 0.45-µm membrane and stored at 4 °C for anion determination or acidified to pH 2 with ultrapure HNO3 for trace-element determination. For isotope measurements, the groundwater samples were collected in either 10-L jerry cans (for nitrogen isotopes determination) or 1-L bottles (for boron isotopes). All containers were first rinsed several times with groundwater and then filled leaving no air space in the neck. To chemically and isotopically characterize anthropogenic sources, chemical fertilizers, animal manure (types are regiondependent), and raw sewage effluents were collected and analyzed for each field site. The chemical fertilizers were generally obtained directly from local farmers, and raw sewage effluents were sampled according to the groundwatersampling procedure. Animal manure samples were collected by leaching 100 g of dried solid samples from manure pits with 1 L of Milli-Q water. Chloride and NO3 concentrations were measured by ion chromatography. Boron concentrations were determined by ICP-MS. NO3 was reduced to NH3, and the mass spectrometry determination of nitrogen isotope compositions was done on N2 liberated by the reaction of NH4Cl with LiOBr on a Finnigan MAT Delta S mass spectrometer. For boron isotope compositions determination, B was analyzed from the Cs2BO2+ ion (21) on a Finnigan MAT 261 mass spectrometer VOL. 39, NO. 2, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
539
(22). The 11B/10B ratio obtained for the NBS951 boric acid standard after oxygen correction was 4.0467 ( 0.0022 (2σ, n ) 147). The isotope ratios are reported as per mil deviation (δ15N or δ11B) from the 15N/14N or 11B/10B ratios relative to air and NBS951 standards, according to the following equations:
[ ] ( ) ( ) 15
15
δ N)
N N sample - 1 × 103 15 N 14 N AIR
14
11
δ B)
and
[ ] ( ) ( ) 11
B B 11 B 10 B 10
sample
- 1 × 103 (1)
NBS951
vNH3gas
Analytical precision on δ15N and δ11B is (0.2‰ and (0.5‰, respectively. Nitrogen and Boron Isotopes in Anthropogenic Groundwater Systems. When using the N isotopic signature as a tracer of the NO3 source, there might be interference between dilution of the polluted groundwater and natural denitrification (both affecting the δ15N of the dissolved nitrate) (3, 11, 19, 23). A simple binary-mixing model can describe the pollution process (i.e., dilution of the polluting end-member by a baseline end-member that represents the unpolluted groundwater). Each end-member is characterized by its N concentration (N) and corresponding δ15N, as described by the following mixing equation system:
[
Nr × f + Np × (1 - f) ) N Nr × f × δ15Nr + Np × (1 - f) × δ15Np ) N × δ15N
(2)
where r and p are the reference and pollution end-members, respectively, and f is the proportion (0 e f e 1) of the baseline end-member in the mixing. Pauwels et al. (10) showed that both autotrophic and heterotrophic denitrification can occur in the same catchment. If an aquifer rock contains pyrite (FeS2), the presence of Fe2+ and a general tendency toward the presence of SO42in conjunction with a decrease in NO3- may indicate NO3reduction coupled with pyrite oxidation in groundwater (autotrophic denitrification; 24):
5FeS2 + 14NO3- + 4H+ f 7N2 + 10SO42- + 5Fe2+ + 2H2O (3) Heterotrophic denitrification (oxidation of organic matter catalyzed by heterotrophic bacteria) contributes to the production of CO2 without increasing SO4 as follows (note that CH2O is a simplified formula for organic matter; 24):
4 4 2 7 CH2O + NO3- + H+ f N2 + CO2 + H2O 5 5 5 5
(4)
Both autotrophic and heterotrophic reactions are accompanied by an isotopic fractionation inducing a 15N enrichment of the residual NO3 (25) that is described by the classic Rayleigh distillation law:
C δ - δ0 ) ln C0
(5)
where δ is the δ15N of the residual NO3, δ0 is the δ15N of the initial NO3 (i.e., before denitrification), C is the NO3 concentration, C0 is the initial NO3 concentration, and is the isotopic enrichment factor. 540
9
The isotopic composition of B, as a NO3 co-migrant, is not affected by denitrification and can therefore be used as a tracer of mixing processes. The pollution process is then described by an equation system similar to the nitrogen systematic (eq 2; N is then replaced by B). Nevertheless, interaction with the aquifer matrix leading to dissolution of B-bearing silicates or adsorption-desorption processes on clays or ferrihydroxides may affect the isotopic composition and concentration of dissolved B (26, 27). Characterization of Anthropogenic Inputs. N and B isotope composition ranges of the principal anthropogenic sources that we measured in France, as well as data from the literature are summarized in Figure 1. Manure. Nitrogen in excreted waste is present mainly as urea, which is hydrolyzed to NH3 and converted to NH4 and finally NO3 in the soil:
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 2, 2005
CO(NH2)2 f NH3 98 NH4+ f NO3-
(6)
The hydrolysis of urea produces a temporary rise in pH, which favors the formation of ammonia (NH3), easily lost to the atmosphere. Both the kinetic fractionation associated with this hydrolysis and the equilibrium fractionation between ammonia and ammonium (NH4+) in solution result in a strong 15N depletion of the NH3 lost from the system, leaving the remaining NH4+ strongly enriched in 15N. Most of this NH4+ is subsequently oxidized to 15N-enriched NO3 (28). Animal manure is thus transformed into NO3 with δ15N values typically in the range of 10-20‰ (29, 30). The range obtained in France (5-35‰) is wider than previously observed (2). This difference may result from local environmental conditions (moisture, temperature, wind speed, etc.), which influence the volatilization of NH3, which in turn determines the degree of 15N enrichment in the final NO3 product (28, 31). The δ11B values in hog manure vary from 19.5 to 42.4‰ (Figure 1), which is higher than those reported by Komor (18). The difference between the American (i.e., Komor’s study) and French values may be explained by the fact that the B isotope composition reflects the animals’ diet and physiology (18). The δ11B of cattle and poultry manure cannot yet be compared to literature values. Mineral Fertilizers. The NO3 and NH4+ in fertilizers are synthesized through industrial fixation of atmospheric N2 by quantitative processes that only slightly fractionate the nitrogen isotope composition (28). Komor (18) reports δ11B values of -2 to 0.7‰ for ammonium nitrate and urea (n ) 3) and of 14.8‰ for phosphate fertilizer (n ) 1). Fertilizers from France have δ15N similar to those reported by previous studies except for one sample from the Allier region, which yields a lower isotopic composition (urea, -12.7‰). Corresponding δ11B range from -8.0 to 7.0‰ (n ) 4). Wastewater. Studies report δ15N values ranging from 10.3 to 23.5‰ (32-34). δ15N from our study yield a slightly lower range from 4.3 to 17.4‰. δ11B values are in good agreement with data from the literature measured on sewage and nonmarine evaporites such as sodium borate, which are generally between 0 and 10‰ (14, 21, 35). Sodium borate is widely used for the production of sodium perborate, a whitening agent found in most detergents.
Different Hydrogeological Contexts Fractured Bedrock: The Arguenon Watershed. Groundwater in Brittany has the highest NO3 concentrations in France. These commonly exceed the 50 mg L-1 limit for drinking water (36). Two small adjacent catchmentssNoe¨ Ronde (33 ha) and Loges (125 ha)swere studied. They are located within the Arguenon watershed approximately 70 km northwest of Rennes (Figure 2A). The basement lithology is granitic gneiss to the north and Brioverian mica schist to the south.
FIGURE 1. δ15N and δ11B characterization of the main NO3 sources. Isotope compositions measured during these studies are compared to values from the literature (data compiled from ref 19 and references included). The Arguenon River flows from an elevation of 150 m down to an elevation of 25 m, with an average discharge of ≈1 m3 s-1. The piezometric level in most of the watershed follows the surface topography, which suggests a fairly low overall aquifer permeability as is common in hard-rock environments (e.g., ref 37). Most of the sample wells are screened in the weathered zone or in both the weathered zone and the fresh basement rock. The local agriculture includes a high level of indoor pig farming, poultry, and cattle breeding and cultivation for which the land is extensively fertilized. Domestic wastewater is treated in a sewage treatment plant (direct discharge to infiltration ponds) downstream from the village of St. Igneuc.
To characterize the nonpolluted state of the watershed’s groundwater, water from the Bonne Fontaine spring (Figure 2A), located in the forested part of the Arguenon Watershed, away from any human influence, was analyzed. It has a low specific conductance (195 µS cm-1; 19) and a sodium chloride composition inherited from rainwater derived from sea salts. The corresponding δ15N is 6.7 ( 0.2‰ (19), consistent with values from the literature for unpolluted groundwater (28, 38-41). δ11B (38.5 ( 0.4‰; 19) is consistent with the seawater value of 39.5‰ (42) and therefore with a marine origin. NO3 concentrations in groundwater vary between 3.2 and 245 mg L-1 (Table 1), with a mean value of 106 ( 78 mg L-1 (median value of 104 mg L-1), which greatly exceeds the 50 mg L-1 drinking-water level. Chemical correlations of B and VOL. 39, NO. 2, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
541
FIGURE 2. Groundwater study sites in France: (A) Brittany, (B) Pyrenees, (C) Allier. The sedimentary deposits include g3M (Oligocene limestones and marls), C2 (g3M-derived colluvium), A (sandy colluvium), Fz, Fy-z, FuA (alluvial deposits, sometimes marly calcareous and made of sands, gravels and pebbles derived from granite, undifferentiated volcanic and metamorphic rocks). NO3 indicate that data plot within a three end-member model that is delimited by unpolluted groundwater (i.e., recharge), 542
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 2, 2005
wastewater, and nitrate resulting from the spreading of animal manure (19). δ15N in groundwater samples vary between 2.7
TABLE 1. Chemical and Isotopic Characterization of the Water Samples (Brittany) sample
date
rainwater rainwater Bonne Fontaine FJH1 Noe¨ Ronde Noe¨ Ronde Noe¨ Ronde Noe¨ Ronde Noe¨ Ronde Loges Loges Loges Loges Loges FCA1 FCA1 FCA1 FCA1 PCA2 PCA2 PCA2 PCA2 PCA2 PPE2 PPE2 PPE2 PPE2 PPE2 FVE1 FVE1 FVE1 FVE1 PVE1 PVE2 PVE2 PVE2 PVE2 PVE2
10-11 2000 06/99 11/00 06/99 11/99 03/00 06/00 11/00 06/99 11/99 03/00 06/00 11/00 06/99 06/99 03/00 06/00 06/99 11/99 03/00 06/00 11/00 06/99 11/99 03/00 06/00 11/00 11/99 11/99 03/00 06/00 06/99 06/99 11/99 03/00 06/00 11/00
a
Eh (mV)
405 388 409 439 484 396 439 389 439 466 422 464 387 288 320 349 441 440 273 425 458 271 380 346 413 416 426 434 487 381 389 487 346 393 396 434
cond. (dS m-1)
AlK (mequiv L-1)
195 267 462 661 587 403 338 562 582 531 465 289 465 422 426 342 773 744 501 428 444 480 468 498 435 415 729 630 694 513 697 811 736 677 570 519
pH
0.30 1.11 1.41 1.47 0.99 1.23 2.11 1.16 1.05 0.88 1.13 1.16 0.75 0.91 0.89 0.88 0.24 0.71 0.77 1.11 0.72 1.40 1.37 1.63 2.20 1.46 0.31 0.37 0.26 0.37 0.20 0.06 0.16 0.13 0.14 0.21
δ15N ((0.2‰)a
5.7 5.8 7.8 7.7 6.5 7.6 7.6 7.7 7.4 6.2 7.6 7.2 6.1 6.1 5.4 6.1 5.7 5.9 5.4 5.7 5.5 6.4 6.3 6.0 6.3 6.0 6.0 5.8 6.0 5.9 5.9 5.4 5.1 4.8 5.3 4.9
δ11B (‰)
NO3 (mg L-1)
B (µg L-1)
30.0 ( 0.2
1.6 5.5 10.1 3.2 104.6 127.0 131.0 106.5 39.2 91.7 90.2 106.0 78.2 44.9 59.9 33.8 35.7 31.8 205.0 200.0 124.0 82.4 104.0 15.9 25.1 34.0 17.4 16.3 196.1 192.7 225.0 172.2 198.0 245.0 230.0 231.0 213.4 229.0
2.4 7 11.0 10.5 32.0 35.0 23.0 27.0 30.8 68.0 51.0 42.0 55.0 46.0 23.0 17.0 16.0 14.0 40.0 39.0 24.0 30.0 31.4 15.0 22.0 17.0 23.0 18.4
