Identifying source and formation altitudes of nitrates in drinking water from Réunion Island, France, using a multi-isotopic approach

Abstract

Nitrate concentrations, water isotopes (δ²H and δ¹⁸O_water) and associated nitrate isotopes (δ¹⁵N_nitrate and δ¹⁸O_nitrate) from 10 drinking water wells, 5 fresh water springs and the discharge from 3 wastewater treatment stations in Réunion Island, located in the Indian Ocean, were analysed. We used a multi isotopic approach to investigate the extent of nitrate contamination, nitrate formation altitude and source of nitrates in Réunion Island's principal aquifer. Water from these study sites contained between 0.1 and 85.3 mg/L nitrate. δ¹⁵N_nitrate values between + 6 and + 14‰ suggested the main sources of contamination were animal and/or human waste, rather than inorganic (synthetic) fertilisers, infiltrating through the subsurface into the saturated zone, due to rainfall leaching of the unsaturated zone at various altitudes of precipitation. Based on δ¹⁵N_nitrate values alone, it was not possible to distinguish between animal and human activities responsible for the contamination of each specific catchment. However, using a multi isotope approach (δ¹⁸O_water and δ¹⁵N_nitrate), it was possible to relate the average altitude of rainfall infiltration (δ¹⁸O_water) associated with the nitrate contamination (δ¹⁸O_nitrate). This relationship between land use, rainfall recharge altitude and isotopic composition (δ¹⁵N_nitrate and δ¹⁸O_water) discriminated between the influences of human waste at lower (below 600 m elevation) or animal derived contamination (at elevations between 600 and 1300 m). By further comparing the theoretical altitude of nitrate formation calculated by the δ¹⁸O_nitrate, it was possible to determine that only 5 out of 15 fresh water wells and springs followed the conservative nitrate formation mechanism of 2/3δ¹⁸O_water + 1/3δ¹⁸O_air, to give nitrate formation altitudes which corresponded to land use activities.

Introduction

We used δ¹⁵N_nitrate, δ¹⁸O_water and δ¹⁸O_nitrate values to identify the key source(s) of nitrate contamination, their formation and their entry altitude in Réunion Island's principal drinking water aquifer. The island offers a unique opportunity to investigate the effects of nitrate transportation from altitudes between 0 and 1300 m elevation (where human activity occurs) due to the specific contaminant sources in each catchment, and the island's simple cone shaped drainage system formed by two volcanoes with altitudes up to 3000 m.

Increasing urbanisation and intensive farming are placing stress on ecosystems that previously assimilated small amounts of nitrate, resulting in higher concentrations of nitrate in drinking water than those recommended. In Réunion Island (an overseas department of France), the maximum permissible concentration of nitrate in drinking water is 50 mg/L NO₃⁻. This permissible nitrate concentration is slightly higher than the recommended World Health Organisation (WHO) drinking water standard of 45 mg/L NO₃⁻.

Nitrate concentrations in drinking water are recorded around Réunion Island by Direction Régionale des Affaires Sanitaires et Sociales de la Réunion (DRASS Réunion, 2004). In the last 30 years, nitrate concentrations in some Réunion Island drinking water wells have risen from trace concentrations pre-1970's (0.1 mg/L NO₃⁻) to modern values well in excess of the acceptable limits (up to 85.3 mg/L NO₃⁻). The contamination can be directly attributed to the increased amount of discrete nitrous compounds such as mineral fertiliser, animal effluent (pork, poultry or dairy), domestic sewage and wastewater leaching into the groundwater, due to increased agriculture and urbanisation.

Stable isotopes are increasingly used to trace various anthropogenic inputs of nitrogenous compounds in groundwaters and ecosystems (Kendall et al., 2007). Historically, nitrogen isotopes (δ¹⁵N) of fertiliser has been documented to range between − 2 and + 2‰ (Kendall, 1998). However these values can be increased by isotopic fractionation processes, often making it similar to soil organic nitrogen, which has δ¹⁵N values around + 3 to + 5‰ (Kendall, 1998). Nitrogen derived from animal manure or human sewage is often indistinguishable on the basis of δ¹⁵N values alone and both are characterised by δ¹⁵N values usually higher than + 6‰, depending if volatilization (loss of volatile NH₃ compounds) has occurred (Aravena and Robertson, 1998, Aravena et al., 1993, Wassenaar, 1995). Therefore the nitrogen in wastewater (animal or human effluent) is distinct from nitrogen in atmospheric deposition (− 10 to + 8‰) nitrogen in most synthetic fertiliser (− 2 to + 2‰), and natural soil organic nitrogen (+ 3 to + 5‰).

The chemical reaction of isotopically distinct nitrogen sources with localised rainfall is well known (Kellman and Hillaire-Marcel, 2003, Wassenaar, 1995). Hence isotopic values of nitrogen (δ¹⁵N_nitrate) and oxygen (δ¹⁸O_nitrate) can be used to identify the origins of various groundwater nitrate contaminants within an aquifer. However, little work has been conducted on the reaction between nitrate contaminants and rainfall occurring at various altitudes, mostly due to the isotopic altitude effect where oxygen in water OHO and nitrate ONO₂ change with increasing altitude (Ambach et al., 1968, Barbier, 2005). Multiple point sources, dilution of nitrates by local groundwater, subsequent oxygen exchange between the water and nitrates from various altitudes, soil microbial action and denitrification can also complicate the identification and apportioning of the various nitrate sources.

In Réunion Island, nitrates enter the groundwater over a range of altitudes from 0 m (at sea level) to 1300 m (highest altitude for housing and farming). Nitrates form with rainwater at specific altitudes, then infiltrate groundwater systems through the relatively porous, regionally homogeneous large-scale lava flows that form the island. As oxygen isotopes in water vary according to the altitude of rainfall (Dansgaard, 1975, Rozanski et al., 1993), it would seem therefore possible to also use both the δ¹⁸O_water and δ¹⁸O_nitrate as comparative tracers to determine the infiltration altitude of groundwater nitrates.

The contribution of δ¹⁸O_water to nitrate formation is theoretically derived from 1/3 oxygen from air (δ¹⁸O_air = + 23.5‰; Amberger and Schmidt, 1987, Kroopnick and Craig, 1972), present in the unsaturated and saturated zones which interacts with 2/3 oxygen from water (rainfall, surface water or irrigation water). This suggests that the oxygen isotopic value of groundwater nitrate may be a good tracer of nitrate source, but is based on the assumption that δ¹⁸O_air does not vary from the theoretical value of + 23.5‰. Mayer et al. (2001) note that in ammonium-rich systems where nitrification rates are high, up to two of the three oxygen atoms in newly formed nitrate are derived from water. In ammonium-limited systems with lower nitrification rates, less nitrate oxygen is derived from water, resulting in more positive δ¹⁸O_water values.

Soil-nitrogen is predominantly organically bound as a result of immobilisation of ammonium (NH₄⁺) and nitrate (NO₃⁻) or nitrogen (N₂) fixation. The process by which organic nitrogen compounds are converted to nitrates consists of ammonification (N_org⁻ ➔ NH₃ ➔ NH₄⁺) and nitrification (NH₄⁺ ➔ NO₂⁻ ➔ NO₃⁻). During nitrification of soils, nitrosomas and nitrobacter (the chemolithoauthotrophic bacteria responsible for nitrification) derive two of the oxygen atoms in nitrate from soil water and one from atmospheric oxygen (Böhlke et al., 2003, Kendall, 1998, Mayer et al., 2001, Wassenaar, 1995).

Here we have analysed nitrate concentration, water isotopes (δ²H and δ¹⁸O_water) and their associated nitrate isotopes (δ¹⁵N_nitrate and δ¹⁸O_nitrate) from 10 drinking water wells, and 5 fresh water springs in Réunion Island. We compared them with potential nitrogen contaminant sources, such as organic and inorganic fertilisers and discharged water from three wastewater treatment stations.