Environ. Sci. Technol. 2008, 42, 8759–8765
Monitoring Reverse Osmosis Treated Wastewater Recharge into a Coastal Aquifer by Environmental Isotopes (B, Li, O, H) W . K L O P P M A N N , * ,† E . V A N H O U T T E , ‡ G. PICOT,† A. VANDENBOHEDE,| L. LEBBE,| C. GUERROT,† R. MILLOT,† I . G A U S , †,§ A N D T . W I N T G E N S ⊥ ´ BRGM 3 av. C. Guillemin, B.P. 36009, F-45060 Orleans cedex 2, France, IWVA, (Intercommunale Waterleidingsmaatschappij van Veurne Ambacht), Doornpannestraat 1, B-8670 Koksijde, Belgium, Research Unit Groundwater Modelling, Department of Geology and Soil Science, Ghent University, Krijgslaan 281, S8, B-9000 Ghent, Belgium, and RWTH (RheinischWestfälische Technische Hochschule), Aachen University, Institut für Verfahrenstechnik, Turmstrasse 46, D-52064 Aachen, Germany
Received April 23, 2008. Revised manuscript received August 18, 2008. Accepted August 22, 2008.
Artificial recharge (AR) is gaining importance as a management tool in water stressed regions. The need to prove recovery performance requires new monitoring tools for AR systems. A novel combination of environmental isotope tracers (B, Li, O, H stable isotopes) was tested for the monitoring of AR of tertiary treated, desalinated domestic wastewater into a coastal dune aquifer in Flanders, Belgium. No significant isotope fractionation was observed for the treatment process, which includes low pH RO desalination. The wastewater, after infiltration through ponds and before recovery through pumping wells is characterized by low molar Cl/B ratios (3.3 to 5.2), compared to 130 to 1020 in the wider study area, δ11B values close to 0‰, rather homogeneous δ7Li values (10.3 ( 1.7‰), and a 18O and 2H enrichment with respect to ambient groundwater due to evaporation in the infiltration ponds. This confers to the AR component a unique isotopic and geochemical fingerprint. Immediately downstream of the pumping wells and in the deeper part of the aquifer no evidence of AR wastewater could be found, indicating a high recovery efficiency. In the wider area and in the deeper part of the aquifer, isotopes evidence mixing of coastal rain and a fresh paleo-groundwater component with residual seawater as well as interaction with the aquifer material. Combining several isotope tracers provides independent constraints on groundwater flow and mixing proportions as a complement to hydrodynamic modeling and geochemical studies.
Introduction Worldwide increasing pressure on groundwater resources is due to a globally growing demand and, regionally, decreasing * Corresponding author e-mail:
[email protected]. † BRGM. ‡ IWVA. § Now at NAGRA Switzerland | Ghent University. ⊥ RWTH Aachen. 10.1021/es8011222 CCC: $40.75
Published on Web 10/29/2008
2008 American Chemical Society
availability (1, 2). This evolution has led to an intensive search for alternative water management options including the use of unconventional water resources as saline waters or wastewaters. Management aquifer recharge (MAR), commonly termed “artificial recharge” (AR), is gaining importance as a means to actively manage the balance of groundwater resources exploited for human needs, mainly drinking water and agriculture (3). Whereas the beneficial impact of MAR on the quantitative status of groundwater is obvious, there are potential quality impacts that can be both beneficial and adverse: Recharge through a biologically and chemically active unsaturated zone and the aquifer itself will lead to an amelioration of the infiltrating water quality (Soil Aquifer Treatment concept). Freshwater injection can ameliorate groundwater quality in coastal aquifers endangered by saline intrusion or in brackish continental aquifers. But the use of low quality waters like domestic wastewaters for AR bears risks for groundwater and end users. Treatment processes not only can be selectively inefficient with respect to specific pollutants (e.g., pharmaceutical residues or boron) and pathogens but can even introduce new types of contaminants like disinfection byproduct (4). A major challenge of AR is therefore to assess and monitor the penetration of artificially introduced waters into the natural groundwater system and to provide reliable estimations of mixing proportions. Environmental isotopes have proved, over the last decades, a valuable tool for elucidating groundwater flow conditions, hydrogeochemical background conditions, and contamination sources (5, 6). Isotope methods have been successfully applied for the conception and monitoring of MAR systems as tracers of groundwater residence times (3H, 3H/3He, 14C) (7-10), tracers of groundwater dynamics, dispersion and mixing proportions (mainly δ18O, δ2H),and tracers of reactive transport and nutrient cycling (δ13C, δ34S, δ15N) (11-14, 10, 8). The specific case of treated wastewater recharge (7, 8, 10) has been first investigated at the large-scale AR site of Shafdan, Israel, (15) showing the potential of boron isotopes for monitoring this type of injectate. The use of intrinsic isotopic tracers (tracers present in the injectate) has several advantages with respect to artificial tracing (16): No tracers are introduced that may contaminate the system; injection of the intrinsic tracer is permanent so that monitoring integrates over time and over variations of operation conditions; logistic constraints are less; and monitoring can be started at any moment of the lifetime of the AR system. Potential drawbacks are linked to the less well-known input function compared to artificial tracing tests, as variations of concentrations and flows in the past are often not recorded. Geochemical reactions and physical transport may alter the isotope signal through fractionation. Furthermore, the principal condition of the use of environmental isotope tracer is a significant contrast of isotope signatures of the injectate and the receptor system. The isotopic “toolbox” offers nevertheless a choice of tracers adapted to different types of injectate (sewage, stormwater, and desalinated water, treated with various techniques) and to different geochemical milieus of the receptor. Combining more than one tracer provides complementary constraints on groundwater flow, mixing, and biogeochemical processes. Here we test a new combination of environmental isotopes (B, Li, O, H) to assess an Aquifer Storage, Transfer, and Recovery (ASTR) system where, after primary (mechanical) and secondary (biological) and tertiary (disinfection, ultrafiltration, reverse osmosis desalination) treatment urban wastewater is recharged into dune aquifer at the Flemish coast (St. Andre´ site, Figure 1a). For this site, both 3D finite element flow modeling (17) and VOL. 42, NO. 23, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8759
FIGURE 1. a. Map of the wider area of the St. Andre´ site, schematic cross section, and isotope sampling points. b. Situation of sampling points in the overall treatment process including secondary treatment, tertiary treatment, SAT, and a polishing step before drinking water distribution. a long-term water quality monitoring were performed (18) so that the results from the isotope study can be validated through independent evidence. The most promising element for tracing AR with treated wastewater is boron, a highly mobile element that is present in municipal sewage in concentrations often exceeding 1 mg/L due to the use of perborates in washing powders. It is not removed by standard treatment techniques and only partially by RO (19) and constitutes therefore an ideal 8760
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 23, 2008
tracer for the monitoring of wastewater infiltration into aquifers (15). The range of isotope signatures observed in nature is in fact particularly high for boron (>70‰). Furthermore, boron is a pollutant of drinking and irrigation water so that boron monitoring may also be required as part of risk assessment. Lithium isotopes bear a high potential as environmental tracer due to the mobility of the element and the strong fractionation due to high relative mass difference between
FIGURE 2. Boron isotope composition of raw and treated domestic wastewaters from refs15, 30, 32-38. 6 Li and 7Li (20). Applications to environmental problems are yet rare (21, 19, 16). The potentially reactive tracers Li and B have been combined with stable isotope ratios of the water itself (δ18O and δ2H) as an ideal conservative tracer of groundwater flow and mixing, particularly sensitive to surface evaporation before infiltration and therefore adapted to infiltration ponds as the injection system (11).
Materials and Methods Site Description. The case study is located in the dune belt at Koksijde at the Belgium North Sea Coast. An average of 285 m3/h tertiary treated municipal effluent is infiltrated using infiltration ponds to reduce the extraction of natural groundwater for potable water production and hold back saline intrusion (17, 22). The surface area of the infiltration pond is 18 200 m2. The wastewater effluent is from a municipal wastewater treatment works which operates primary treatment and activated sludge secondary treatment. The secondary effluent is then polished with ultrafiltration (UF), disinfection, and reverse osmosis (RO) before infiltration (Figure 1b). The produced RO filtrate is reconditioned by pH adjustment and then transported by a pipeline (2.5 km) to the recharge/ extraction site of St. Andre´ (Figure 1). The infiltration occurs in ponds into the unconfined sandy aquifer. The recharged water is recaptured after a minimum residence time of approximately 30 days in the dune aquifer using 112 pumping wells. The extracted groundwater is further treated with aeration, rapid sand filtration, and UV disinfection for drinking water production. Hydrogeology. The unconfined aquifer in the dune belt is formed by 25 to 35 m thick quaternary sandy deposits, in which, in some areas, thin layers of fine grained sediments occur. They lie upon 110 m thick tertiary clay, itself underlain by a confined sandy to silty aquifer. Details about geology are provided in the Supporting Information. Salinity patterns are complex: Under the dunes, a freshwater lens formed by infiltrating rainwater. Under the shore, a salt water lens is present above a freshwater tongue. In the polder area, south of the dunes, freshwater filled the upper and relict salt water the lower part of the aquifer (23). Overexploitation, causing the inflow of salt water from the aquifer under the beach and polder areas, threatened the quality of the dune aquifer so that the AR project was initiated to re-equilibrate recharge and extraction. Sampling and Analytical Procedures. The investigated area extends on a profile (Figure 1a) stretching from the
upstream southern part of the dune aquifer over the infiltration pond, and the pumping well galleries toward the sea, following roughly the natural groundwater flow gradient. Water was sampled throughout the treatment process (Figure 1b). Samples further include water from the extraction wells, distributed drinking water, one rainwater grab sample, and a seawater sample (Table S1, Supporting Information). Lithium and boron concentrations were measured by ICPMS and major ion concentrations by ion chromatography and ICP-AES. Analytical errors for all dissolved constituents did not exceed 5%. Boron isotopes, expressed as δ11B versus NBS951, were determined by positive thermal ionization mass spectrometry (P-TIMS). Lithium isotopic compositions (δ7Li vs L-SVEC) were measured by Multi-Collector ICP-MS, and oxygen and hydrogen isotopes (δ18O, δ2H values; vs. VSMOW) were determined by IRMS after gas-water equilibration. Details on analytical techniques and uncertainties are provided in the Supporting Information.
Results and Discussion Boron Isotopes. Seawater and Rainwater. Seawater boron concentrations (4400 µg/L) and B isotope composition (39.3‰) measured on the Flemish coast show the values for global seawater (24).The coastal instantaneous rain sample of this study (July 2007) falls close to the seawater δ11B value with 34.9‰; its boron concentration of 4.7 µg/L is in the lower range measured for costal rains (from
