Université de Bordeaux 1, CNRS UMR 5805 Environnements Paléoenvironnements Océaniques (EPOC), avenue des facultés, 33405 Talence, France
bUniversité de Bordeaux 1, GHYMAC EA 4134, Talence, France
cUniversité de Bordeaux 3, Institut EGID, Pessac, France
a r t i c l e i n f o a b s t r a c t
Article history:
Received 17 November 2009 Received in revised form 20 April 2010 Accepted 22 April 2010
Available online 15 May 2010
Groundwaters impacted by mature landfill leachate are generally enriched in ammonium. In order to assess the dynamics of ammonium exchanges between leachates and the water system inside a sandy permeable catchment we measured ammonium, nitrate and chloride concentrations in the stream and in sediment pore waters of the streambed of a landfill impacted aquifer. Geophysical investigation methods complemented the biogeochemical survey. The studied zone is a 23 km² catchment located in a coastal lagoon area sensitive to eutrophication risk. Ammonium concentrations in the river were up to 800 µmol l−1during low water period in summer. Three surveys of the river chemistry showed a regular increase in ammonium, nitrate and chloride concentrations along a 1 km section of the watercourse, downstream the landfill, implying that the leachate plume exfiltrates along this section.
Sediment cores collected within this section showed all an increase in ammonium concentrations with depth in pore waters as a consequence of the landfill leachate dispersion, as attested by a simultaneous increase in chloride concentrations. Nitrate enrichment in the river water was due to nitrification of ammonium at the interface between groundwater and streamwater. The apparent nitrification rate obtained was within values reported for turbid estuaries, although the river contained very little suspended particulate matter. Actually, pore water chemistry suggests that nitrification occurred for the most part in subsurface permeable sediments, rather than in stream water. The overall topographic, hydrological, geochemical, and geoelectrical data set permit to estimate the extension of the chloride and ammonium plume. The estimation of the apparent ammonium plume velocity is 23 m year−1whereas the chloride plume velocity should be 50 m year−1. The river is the outlet of the impacted groundwaters. Considering that the input of ammonium from the landfill is balanced by the present day output via the river, the residence time of ammonium in the aquifer is between 7 and 18 years.
© 2010 Elsevier B.V. All rights reserved.
Keywords:
Landfill leachate Contaminated plume Ammonium Nitrification Hyporheic zone porewater Arcachon Bay
Electrical Resistivity Tomography Self Potential
1. Introduction
The amount of domestic and industrial waste has increased during the last 50 years. In France the annual domestic wastes production is about twice more in 2004 than in 1960 and reaches
353 kg hab−1year−1(ADEME, 2007). Landfilling is a common and cheap technique of waste management, which permits to treat 54% of domestic wastes in the USA (Hanson and Caponi, 2008) and 39% in France (ADEME, 2007). Consequently, environ- mental laws have been created in order to reduce pollution from landfills. Environmental laws of the European Union focused on the water quality via the Water Framework Directive (2000/
60EC). Despite these improvements, several landfill plants continue to pollute groundwaters because of the several-decade lifetime of waste (Chu et al., 1994; Ehrig, 1989). Recent landfills
⁎Corresponding author.
E-mail address:p.anschutz@epoc.u-bordeaux1.fr(P. Anschutz).
1Present address: BRGM (French Geological Survey), Orléans, France.
0169-7722/$–see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jconhyd.2010.04.006
contribute to pollution when leaks of compartment storage occur.
The frequent identified defects are stamping of geotextiles, breaking joints (Darilek and Parra, 1989; Touze-Foltz et al., 2001), or diffusion through geomembranes (Foose et al., 2002).
These defects favor input of leachate in subsoil and unconfined aquifers after percolation. Because sandy catchments have a high hydraulic conductivity coefficient, leachates easily flow in the hydraulic gradient. Moreover, sands enable rapid exchanges between aquifer and rivers assimilated to a Darcy flow (Sophocleous, 2002). Thus, leachates may be subjected to variable redox conditions between aquifer and rivers.
Leachates from older landfills are enriched in ammonium because of the hydrolysis and fermentation of biodegradable organic nitrogen (Carley and Mavinic, 1991; Kulikowska and Klimiuk, 2008). Ammonium is the stable form of nitrogen in reduced waters. Ammonium is oxidized in oxic environment such as rivers or upper portions of free aquifers. This oxidation called nitrification is a two-step reaction, with nitrite as an intermediate form, completed by autotrophic bacteria. Bacterial nitrification kinetics is regulated by dissolved oxygen and inorganic carbon concentrations, temperature and pH. Denitri- fication occurs in anoxic environments. Denitrification is a dissimilatory reduction of nitrate to dinitrogen carried out by heterotrophic bacteria. The process is controlled by organic carbon concentration. Dinitrogen is inert but the other forms of nitrogen have an environmental impact. First, ammonium contributes to leachate toxicity (Pivato and Gaspari, 2006;
Ward et al., 2002), as nitrite does (Dave and Nilsson, 2005).
Second, ammonium and nitrate are highly bioavailable for primary producers and high concentrations may be responsible for eutrophication. Confined ecosystems like coastal lagoons or lakes are highly sensitive to eutrophication (e.g.Adriano et al., 2005; Billen et al., 2007; Jickells, 2005). Eutrophication is most often the result of an elevated supply of nutrients to surface waters, particularly nitrogen and phosphorus, that enhance the production biomass of algae and phytoplankton (Prepas and Charette, 2003). It is therefore important to characterize the impact of landfills on river nutrientfluxes.
Here we present for the first time the monitoring of ammonium in a stream, the Ponteil River, and in pore waters of the streambed sediments of a landfill impacted superficial aquifer. The outlet of the Ponteil River reaches the Arcachon Bay, which is the largest intertidal lagoon of the French Atlantic coast covering 156 km2. This lagoon is highly sensitive to nutrient supply and eutrophication risk (De Wit et al., 2001). The study focused on nitrogen compounds, which are adapted tracers of leachates originating from mature landfills (Carley and Mavinic, 1991; Kulikowska and Klimiuk, 2008). The aims were to constrain the dynamics of the exchanges between leachates and the water system inside a permeable catchment and to assess the fate of leachate-originating ammonium in natural environment. In order to achieve these aims, we have applied biogeochemical and geophysical investigation methods.
2. Studied site
2.1. Catchment
The studied zone is a small catchment located near the Arcachon bay in the South-western part of France (Fig. 1). The catchment of 23 km2(Laplana et al., 1992) is drained by the
Ponteil stream. The largest portion of the catchment is covered by pine forest and is moderately urbanized down- stream. The landfill plant represents 0.4 km2, located at about 5 km upstream of the river mouth (Fig. 1). The catchment is a typical lowland as the average slope is 0.25% (De Wit et al., 2005). The soil is composed of quaternary wind-driven sands that are poorly cemented and highly permeable. The superficial cover is enriched in organic matter.
2.2. River properties
The Ponteil River is a minor source of freshwater to the Arcachon lagoon. It represents less than 1% of fresh water input (Canton et al., 2009). The stream water discharge is between 60 and 800 l s−1and the averageflow is 200 l s−1 estimated by point gauges and comparison with daily gauging tests of similar catchment (Canton et al., 2009). The mouth of the Ponteil River is located in a protected and confined area inside the Arcachon lagoon.
2.3. Aquifer properties
The aquifer is a highly connected multi-layered system (Astié et al., 1971). We have focused on the shallow free aquifer called the“Sable des Landes”covering more than 4000 km2 close to the Arcachon lagoon. The water table is not deeper than 3 m and the total thickness is variable but does not exceed 25 m near the coast. It is formed by a homogeneous quaternaryfine sand formation (Borneuf, 1968). Its substrate is composed of 1 to 4 m thick clay deposits (Legigan, 1979). The “Sable des Landes”hydraulic conductivity is 3.7.10−4ms−1(Rimmelin et al., 1998) and the hydraulic gradient is about 5‰in the Ponteil catchment. The resulting flow velocity is about 50 m an−1 calculated from the Darcy law.
2.4. Landfill
The studied landfill has been active since 1973. Between 1973 and 1997, it received domestic wastes of nearby municipalities. Wastes were buried inside soil without leachate management. Since 1997, geotextiles and geomem- branes have been used to limit infiltrations. Leachates are collected and filtered by inverse osmosis process. On the other hand, the annual waste weight has increased up to 170·106kg year−1. Organic wastes (domestic and industrial) represent 90% of total wastes. Asbestos and crushed cars represent 10%. The landfill has been closed since 2008.
3. Materials and methods
A bi-monthly survey of nitrate and ammonium concen- trations, temperature and pH was performed near the outlet of the Ponteil stream between October 6, 2006 and October 30, 2007. Waters from the river were sampled three times at high spatial frequency between the outlet of the stream and upstream the landfill on 10 April 2006, 29 June 2007, and 29 April 2008 (Fig. 2). Waters were sampled with a 50 ml syringe and filtered through a 0.20 µm cellulose acetate syringe membrane. Samples were kept frozen until analyses of dissolved nutrients. Temperature and pH were recorded in situ using WTW probes.
In addition, we sampled 20 cm long sediment cores in the riverbed at the water/sediment interface in June 2007 and April 2008 (Fig. 2). Stream water temperature and discharge were 15 °C and 133 l s−1in June 2007 and 13 °C and 214 l s−1 in April 2008. Sediment cores were collected manually using pre-sliced acrylic tubes (6 cm diameter and 30 cm length).
The device enabled collecting pore waters from sandy sediments (Deborde et al., 2008). Interstitial waters were immediately extracted from their cores in order to analyze dissolved nitrate, ammonium and chloride. For that, the core was lain down horizontally to prevent pore water migration, the pre-sliced core was sub-samples with a 5 cm resolution from the surface to 20 cm. Sub-samples were preserved in centrifuge vial under inert N2-atmosphere. We used centri- fuge 0.2 µm filter vials. Pore waters were extracted by centrifuge at 4000 rpm for 20 min (Deborde et al., 2008).
Filtered interstitial waters were frozen at−25 °C for nutrient analysis. At each station we sampled and handled three cores in order to assess spatial heterogeneity. Finally, we drilled and sampled 6 wells around the landfill. Dissolved nitrate and nitrite were analyzed byflow injection analysis according to standard colorimetric methods (Anderson, 1979; Hall and Aller, 1992). Dissolved ammonium was analyzed by colori- metric procedures (Mullin and Riley, 1955; Murphy and Riley,
1962; Stookey, 1970; Strickland and Parsons, 1972). The precision of these procedures was ±10%.
Wells located inside the waste disposal were sampled and analyzed by the former manager. The company Antea gave us the results.
Between March and July 2009, we performed geophysical investigations, which could provide large-scale information on the aquifer. Wefirstly realizedfive electrical resistivity tomography (ERT) profiles (see location onFig. 2). ERT is an active geoelectrical prospecting technique used to obtain 2D, 3D and 4D (with time) images of the subsurface electrical resistivity distribution. The electrical resistivity of rocks and soils is highly dependent on their water content, ionic concentrations and clay content. Because the electrical resistivity values can drastically change in the presence of contaminated groundwater due to an increase in ionic concentration, ERT has been widely used to study contami- nated aquifers (Atekwana et al., 2000;Naudet et al., 2004).
This geoelectrical method consists in the injection of current into the ground through a set of electrodes, and the measurement of the resulting electrical potential differences between another set of electrodes at the surface. The mathematical association between electric currents and voltage measurements provides the apparent resistivity Fig. 1.General view of the Ponteil River catchment.
distribution of the ground called pseudo-sections. In order to have a better vertical resolution, we used a Wenner–
Schlumberger configuration (Samouëlian et al., 2005). For 2D profiles acquisition, we used a fast resistivity-meter (Syscal-Pro-Switch, Iris Instruments) with 48 electrodes spaced at 2 m (Fig. 2) leading to 94 m long profiles. The apparent electrical resistivity pseudo-sections were inverted using RES2DINV software (Loke and Barker, 1996) in order to obtain a model of the true electrical resistivity distribution.
We secondly carried out a self potential (SP) survey. This method involves a passive measurement along the profile of a natural electrical potential at the ground surface between two non-polarizing electrodes. This natural electrical field is associated with different charge polarization mechanisms occurring at depth. In the case of contaminated area, the SP sources are due to electrokinetic, electrochemical and redox effects (e.g.Jouniaux et al., 2009).Naudet et al. (2003)showed that in the case of organic contaminated groundwater, the origin of SP signal comprises two main components: an electrokinetic component associated with groundwaterflow (Birch, 1998; Fournier, 1989; Titov et al., 2002) and an electrochemical component associated with redox reactions.
The aim of our SP prospect was to confirm the boundary of redox front in a linear profile running parallel to the landfill and extending 500 m downstream. We used two non-polarizable Pb/PbCl2Petiau electrodes with an electrode spacing of 10 m, a high impedance voltmeter (100 MΩ) and insulated single- conductor wire as described inNaudet et al. (2004).
To quantify water exchanges between the aquifer and the stream during a mean hydrological period (outside a winter flood event), we performed three gauging tests between the upstream and downstream of the river in June 2008 using two
different methods. First, we used a current velocity profile technique. We measured and summed current velocity along many vertical profiles across the stream with an electromag- netic current velocity meter type Flo-Mate. Second, we used the salt dilution technique, which is well established in small headwater streams. It consists of adding a mass of brine upstream and measuring conductivity variations down- stream. This method was less accurate than current velocity measurements, but it permitted to consider water residence time characteristics.
Finally, local groundwater flow direction was estimated from a survey of water table elevation in punctual piezo- meters. They permit to calculate the hydraulic gradientΔh! (in‰). The groundwaterflow velocityvis assessed using theh
Darcy law:ν¼K:Δhh whereK is the hydraulic conductivity equal to 3.7.10−4m s−1 according to Legigan (1979). The piezometers inside the landfill were dug, sampled, and analyzed by landfill managers. The piezometers outside the landfill were dug manually with an auger and they were not deep enough to collect water quality samples.
4. Results
4.1. Aquifer
The less saline groundwaters were sampled in wells P1 and P2 located upstream of the landfill (Fig. 3). Concentrations of dissolved chloride and ammonium were about 1000 µmol l−1 and 20 µmol l−1respectively. Water conductivity was about 160 µS cm−1. Values increased downstream up to about 2000 µmol l−1of ammonium, about 2600 µmol l−1of chloride, Fig. 2.Map of the studied area and sampling strategy. The sampling point 30 is the location of the 2-years monitoring of the downstream Ponteil River chemistry.
and 1300 µS cm−1 (Table 1). The overall drill-hole data set showed a significant increase in ammonium concentrations in 2009, even upstream. Chloride concentrations were patchier.
Piezometric head elevations at low and high water level are shown inFig. 3. The landfill influences piezometric lines, probably because of waste compartment management.
Therefore, the groundwaterflow direction is altered. Upper- most groundwater storey is, however, drained by the river Ponteil. The hydraulic gradientΔh!
his about 7% near the river according to piezometric data inFig. 3i.e. a groundwaterflow velocity close to 750 m year−1.
4.2. Stream
Data of the two-year monitoring of ammonium concen- tration downstream the Ponteil river showed variations between 0 and 100 µmol l−1except in summer, where the concentrations increased up to 800 µmol l−1(Fig. 4). Maxima of concentrations were observed during low water period in summer. Minima were associated to high water level. Nitrate concentrations varied between 10 and 35 µmol l−1. They were the lowest during high water periods in winter.
The three surveys of ammonium, nitrate and chloride concentrations along the stream showed the same trends (Fig. 5). The most complete survey of 2008 showed that ammonium concentrations were below 5 µmol l−1upstream the gauging point J1. Then, at the sampling point 6, ammonium increased regularly to reach about 100 µmol l−1 1 km downstream, at the gauging point J2. Beyond this point ammonium concentrations remained constant until the gauging point J3, and then, it decreased slightly by about 10 µmol l−11 km farther. Chloride data were more scattered.
They showed, however, an increase from about 700 µmol l−1 upstream to values above 900 µmol l−1 downstream. The shape of the conductivity profile was the same as that of the ammonium profile as far as the point J3. Beyond this point, the conductivity did not decrease and remained constant.
Nitrate concentrations showed an upstream concentration of 12 µmol l−1in June 2007 and 5 µmol l−1in April 2006 and 2008. At the sampling point 6, nitrate concentrations began to increase regularly with about 5 µmol l−1km−1in 2007 and 2008 and 3 µmol l−1km−1 in 2006. Chemical variations between gauging stations are summarized inTable 2.
The stream waterflow was estimated at the stations J1, J2 and J3 in June 2008 (Table 3). Between J1 and J2, the waterflow increased by 33 l s−1i.e. by 25% according to the currentmeter method, and by 25 l s−1i.e. by 15% according to the salt dilution technique (Table 3). Downstream, the stream water flow decreased little between J2 and J3 by 13 l s−1i.e. 7% according to the currentmeter method, and by 6 l s−1i.e. by 3% according to the salt dilution technique. The mean water residence time in river sections as deduced from propagation of the maximum conductivity anomaly during the salt spiking experiment was 71 min between J1 and J2 and 167 min between J1 and J3, i.e. a mean residence time of 111 min km−1.
4.3. Aquifer/stream interface
Sediments of the riverbed consisted of permeable median sands. Locally we observe accumulations of tree fragments.
Fig. 6 shows the concentration of ammonium, nitrate, and chloride in pore waters extracted from 20 cm long sediment cores collected at the aquifer/stream interface in triplicate.
Cores 07A, 07B, 07C, and 08B collected between J1 and J2 all Fig. 3.Interpolation of the piezometric level data from the wells located inside the landfill in high water level (winter) and low water level (fall).
Table 1
Mean (2007–2009) conductivity, ammonium and chloride concentrations in boreholes inside of the waste disposal.
P12 landfill P11 landfill P10 landfill P9 landfill P8 landfill P2 upgradient P1 upgradient
Conductivity (µS cm−1) 1295 839 644 317 1141 162 180
Ammonium (µmol l−1) 2000 1571 1500 786 1357 19 10
Chloride (µmol l−1) 1963 2657 1640 1086 2314 979 1157
showed an increase in ammonium concentrations with depth.
Concentrations up to 2600 µmol l−1were measured in cores 07B and 07C. These cores showed also an increase in chloride concentrations with depth that reached up to 5000 µmol l−1. In cores sampled upstream J1 (08A) and downstream J2 (08C and 08D), pore water ammonium concentrations remained close to the concentrations measured in stream waters. In these cores, pore water chloride concentrations are higher than in stream water, but they remain below 2000 µmol l−1. Nitrate concen- trations were below detection limit (1 µmol l−1) in pore waters of the core 08A collected upstream J1. Cores collected between J1 and J2 showed nitrate concentrations between 1 and 16 µmol l−1 that decreased generally with depth. Cores collected downstream contained the highest concentrations of nitrate in the upper sediment layer with values above 50 µmol l−1that reach 205 µmol l−1in one core of site 08C and 118 µmol l−1in one core of site 08D. Triplicate cores present some differences in the concentration vs. depth profiles. They show, however, always the same trend, except for site 07B, where one core displays low concentrations of ammonium and chloride while two others present concentrations above 2000 µmol l−1. These two cores were collected under 50 cm of water, in the deepest part of the riverbed cross section at site 07B. The core with low concentrations was collected on a sand bank in the middle of the riverbed, at a depth close to the water surface. This sediment was probably less exposed to ground- water seeping.
4.4. Electrical resistivity tomography and self potential One ERT profile (profile A) has been performed upstream the landfill to determine the electrical resistivity distribution in a non-contaminated area (see location onFig. 2). The others are located at the south-eastern part (profiles B, D and E) and downstream (profile C) from the landfill. All the profiles show high resistivity values in thefirst 3 m below the ground surface, corresponding to the vadose layer, with resistivity values between 1000 and 2000Ωm (Fig. 7). Variability between different profiles in the vadose zone can be explained byfield
observations. For example, the soil was dryer when performing profiles D and E compared to other profiles. This could explain the highest resistivity values measured there. The vegetation cover changed along profile C and we crossed a trenchfilled with heterogeneous materials (around position 35–40 m), which could explain the variation of values close to the surface.
The ERT profiles are restricted by a basement with high resistivity. This boundary may correspond here to the clay basement of the Sable des Landes aquifer. It is positioned at 11 or 12 m depth downstream, and below 12.5 m depth upstream.
Between the vadose zone and the aquifer basement i.e.
between 3 and 11 m depth, we clearly identify the aquifer zone with low resistivity. Profiles B to E show lower resistivity values (between 9 and 42Ωm) compared to profile A (between 90 and 450Ωm) performed upstream. The SP profile performed along the southern edge of the landfill showed constant values upstream the landfill, followed by a negative anomaly that reached−100 mV at 600 m (Fig. 8). From 600 m to 900 m, the self potential increased up to 0 mV. The maximum at 900 m occurred right at the level of an overhead high tension wire. The SP decreased again at the end of the profile.
5. Discussion
5.1. Water exchange between groundwater and stream water The gauges along the fluvial continuum suggest that the Ponteil River drains groundwater between J1 and J2, because the river discharge increases. The hydraulic gradientΔh!
hclose to the Ponteil River is estimated to about 7% from the water table elevation (Fig. 3). The resulting Darcy velocityvis of about 750 m year−1. The theoretical waterflow change between J1 and J2 due to groundwater discharge can be deduced from the equationΔQ=v.Se, whereΔQis the stream waterflow between J1 and J2,vis the Darcy velocity andSeis the exchange surface area between groundwater and stream water.Seis close to 2000 m2.ΔQshould be equal to 46 l s−1, which is consistent with the measured increase in waterflow between J1 and J2, suggesting that the piezometric map and the gauges are correct.
Fig. 4.Temporal variations of daily streamwater discharge (black line), ammonium concentrations (black square) and nitrate concentrations (grey triangle) concentrations;Xaxis is the time in month between January 2007 and December 2008; the leftYaxis represents ammonium concentrations in µmol l−1in streamwater and the streamwater discharge in l s−1; The right Y axis represents nitrate concentration in µmol l−1.
5.2. Impact of the waste disposal on groundwater chemistry 5.2.1. Aquifer
Waters collected in wells or in sediment cores that have the highest concentrations of ammonium also reveal high
concentrations of chloride. Ammonium accumulation is the result of anaerobic mineralization of organic nitrogen, and chloride enrichment is typical of landfill leachates (Gettinby et al., 1996). The map of piezometric levels (Fig. 3) suggests that the groundwaterflow direction is from the landfill to the
Table 2
Variations of nitrate, ammonium and chloride concentrations in µmol l−1between gauging points in June 2008 (first value) and in April 2007 (second value);Δdis the distance between the gauging stations;Δtis the mean duration of water transit deduced from the salt spiking experiment.
∆d(m) ∆q(min) ∆NO3−(µmol l−1) ∆NH4+(µmol l−1) ∆Cl−(µmol l−1)
J1–J2 700 71 +5 to +8 +60 to +135 + 120 to +95
J2–J3 800 96 + 5 +17 +135
J3-most downstream 750 + 5 −19 −82
Fig. 5.Survey of the River Ponteil in 2006, 2007, and 2008. J1, J2 and J3 show the position of the gauging test stations.
river. The increase in ammonium concentrations with depth in pore waters of the riverbed sediments is not the result of in situ organic matter mineralization, but the consequence of the landfill leachate dispersion, as attested by the simulta- neous increase in chloride concentrations. Wells P8 to P12 (Table 1, landfill downgradient) are enriched in chloride and ammonium relative to waters collected upstream. In the river the enrichment of dissolved ammonium and chloride can be estimated assuming a background concentration of ammoni- um and chloride of 5 µmol l−1and 700 µmol l−1, respectively (Fig. 5). This background has been measured in the most upstream sample of the Ponteil River. In groundwater, the background concentrations of ammonium and chloride are estimated from the average wells P1 and P2 to 14.5 µmol l−1 and 1068 µmol l−1, respectively (Table 1). Concentrations of both compounds reach up to 2000 µmol l−1in wells located in the western part of the landfill (Table 1). Values are twice higher in pore waters collected in riverbed sediments at stations 07B and 07C, which suggests that the groundwater located currently beneath the landfill is not the most contaminated one. Groundwater that now reaches the riverbed at the level of 07B and 07C is more impacted by the landfill leachate, which suggests that the transfer of matter between the waste disposal and groundwater was higher in the past than now.
Conductivity increases with ionic concentration according to the relation:
Sf=∑λiCi
whereSfis thefluid conductivity in S m−1,λiis the molar ionic conductivity in S m2mol−1, withλCl=7.63·10−3and λNH4= 7.35·103S m2mol−1,Ciis the molar concentration of the specie i in mol m3. The resulting contribution of ammonium and chloride concentrations on the total conduc- tivity change between P1 and P12 is 207 µS cm−1, whereas the measured conductivity change is 1115 µS cm−1. Then, ammonium and chloride explain only 20% of the conductivity increase in P12, which points out the complex composition of the leachate.
5.2.2. River
The three sampling campaigns along the Ponteil River show an increase in ammonium concentration from about 100 m upstream the gauging test station J1, which suggests that impacted groundwater reaches the river at this point (Fig. 5).
The only sediment core that has been collected upstream this station (08A) shows the lowest concentrations of pore water ammonium, which confirms that the ammonium enriched plume does not reach this point. Ammonium, nitrate, and chloride concentrations and conductivity rise during about
1 km until a point located about 200 m downstream the station J2. Then, ammonium concentration decreases slightly, whereas conductivity remains steady and nitrate concentrations con- tinue to increase. The enrichment of ammonium and chloride on a 1 km long river section indicates a large-scale dispersion mechanism of leachates rather than a point source. The increase in waterflux between J1 and J2 (Table 3) suggests groundwater seepage in the stream that may explain the diffusive enrich- ment. Such a hydrological behavior is consistent with the regional hydrology since 95% of the stream waterflow is due to upper aquifer drainage (Borneuf, 1968; Saint Pe, 1966).
Considering that the drainage of the aquifer is the only source of water toward the stream, we can estimate an average ammonium and chloride concentration of the groundwater seeping toward the riverbed. The rise in ammonium was between 60 µmol l−1in 2008 and 135 µmol l−1in 2007 along the J1 and J2 interval (Table 2), and the water discharge increased by between 15% and 25% (Table 3). Therefore, the excess of ammonium concentration in groundwater that fed the river was between 240 and 900 µmol l−1. This range is large because of error margins on data. Chloride concentrations increased by 120 and 95 µmol l−1between J1 and J2 (Table 2).
As a result the groundwater that mixed with river water had a concentration between 380 and 800 µmol l−1 above the upstream river water concentration (i.e. 520 µmol l−1). These concentrations are below the highest values found in sediment pore waters collected between J1 and J2. Pore water ammoni- um and chloride concentrations vary from a core to another;
however, this suggests that groundwaters that reach the riverbed are heterogeneous. Waters coming from the left (southern) side of the river are probably less contaminated than waters coming from the landfill side.
The aquifer downgradient the landfill has a lower resistivity than upgradient aquifer as shown by the ERT profiles B to E (Fig. 7). Resistivities are between 9 and 42Ωm downgradient and between 90 and 450Ωm upgradient. This suggests that a high ionic contamination plume occurs downgradient the landfill. The link between the fluid resistivity (measured in wells) and the soil resistivity (measured by ERT method) is given by the Archie law:
ρs=ρfaΦ−m
whereρfandρsarefluid and soil resistivities, respectively;a is the coefficient of saturation equal to 1 in saturated aquifer;
Φis the porosity equal to 40% (Anschutz et al., 2009) andmis the coefficient of cementation equal to 1.3 in non cohesive sands. Values ofρfdeduced from Archie law are between 2.3 and 10.6Ωm downgradient, corresponding to conductivities between 4400 and 940 µS cm−1. Fluid resistivities are between 22.8 and 113.8, i.e. conductivities between 440 and 90 µS cm−1 upgradient. Then, conductivities measured in wells are consistent with conductivities assessed from ERT.
5.3. Flux of dissolved inorganic nitrogen
Concentrations of ammonium in the downstream section of the river were between 50 and 130 µmol l−1in the longitudinal surveys (Fig. 5). These concentrations are in agreement with the concentrations measured close to the outlet of the river in spring during the two years monitoring (Fig. 4). Concentration Table 3
Comparison of the both gauging methods.
J1 J2 J3
Stream waterflow (l s−1) curentmeter method 144 177 164 Stream waterflow (l s−1) salt spiking method 168 193 187 Average stream waterflow (l s−1) 156 185 176
in the river reached up to 800 µmol l−1in August 2007 and 350 µmol l−1in summer 2008. Such values remain still below the concentrations measured in pore waters of the riverbed
(Fig. 6), which suggests that the mechanism of ammonium enrichment is the same in summer as in the other seasons.
These maxima of concentrations are observed during low water Fig. 6.Ammonium, nitrate and chloride profiles in pore waters of the Ponteil riverbed sediments. Cores were collected in triplicate at each station. From upstream to downstream: 08A, 08B, 07A, 07B, 07C, 08C, and 08D (seeFig. 2).
period. Therefore, ammonium concentrations are the highest probably because landfill leachates that seep through the sandy aquifer to the riverbed are less diluted by waters of the upstream portion of the stream. Nitrate concentrations are the same in summer 2007 and in summer 2008, although ammonium enrichment is not the same. It implies that nitrification rate is not governed by ammonium concentrations, but rather by other parameters such as temperature.
Annual fluxes of dissolved inorganic nitrogen species are calculated by integration of the product of concentration by water flow. Annualfluxes of ammonium and nitrate at the outlet of the Ponteil stream in 2007 were 893 103mol year−1(12.5·103kg- N year−1) and 93·103mol year−1 (1.3·103kg-N year−1), re- spectively; the fluxes in 2008 were 364 103mol year−1 (5.1·103kg-N year−1) and 129·103mol year−1 (1.8·103kg- N year−1), respectively (Fig. 4). At the lagoon scale, the Ponteil River corresponds to only 0.9% of freshwater discharge but 1.6% of
dissolved inorganic nitrogenflux to the Arcachon Bay, and 25% of ammoniumflux to the bay in 2007. In 2008, the Ponteil River represented only 0.7% of the total DINflux and 14% of the total ammonium flux. These differences were due to hydrological context: Rainfalls were lower in 2007 than in 2008 with 808 mm and 1036 mm, respectively (data source: Meteofrance at Temple station). Ammonium enrichment in Ponteil River is due to groundwater discharge. During dry years such as 2007, ground- water that exfiltrates is probably enriched in ammonium, because it is less diluted with rainwater. The average rainfall between 1971 and 2008 was 961 mm. In these conditions, 2007 is representative of dry year and 2008 is representative of wet year. Moreover, the main portion of DINflux from the other catchments that fed the Arcachon lagoon is due to fertilizer. The leaching of farming area is higher during wet years (Rimmelin et al., 1998). Then, the proportion of ammonium supplied by the Ponteil River relative to the overallflux of DIN becomes lower.
Fig. 6(continued).
5.4. Nitrification in the river sediment
Nitrate concentrations continuously increase between J1 and J2 (Fig. 5), whereas pore waters are depleted in nitrate relative to river waters along this section (Fig. 6). This suggests that the
groundwater seeping should dilute river nitrate concentration rather than augment it. The common pathway of nitrate enrichment in oxic waters that contain dissolved ammonium is nitrification (Cébron et al., 2003; Garnier et al., 2007, 2002).
Assuming that nitrate enrichment is due to nitrification, we can Fig. 7.2D electrical resistivity tomography (ERT) obtained in Wenner configuration with an electrode spacing of 2 m and an inversion realized with RES2DINV;
Black lines represent shallow aquifer boundary estimated by wells. Profile A has been performed upstream from the landfill and the others downstream (see location onFig. 2).
estimate a nitrification rate between J1 and J3 using the residence time of waters in the river section and the mean waterflux (Table 2). The value obtained is 65 µmol l−1day−1in
June 2008. This value is high, but of the same order of magnitude comparative to 37 µmol l−1day−1reported in the Seine River downstream Paris (Brion and Billen, 2000), where nitrifying bacteria originating from wastewater treatment plant support nitrification (Montuelle et al., 1996; Teissier et al., 2002). The nitrification rate obtained is within values reported for turbid estuaries, which are efficient nitrifying reactors due to suspended particulate matter (Bonnet et al., 1997). A nitrite oxidation maximum rate of 32 µmol l−1day−1(Dai et al., 2008), and a nitrification between 45 and 80 µmol l−1day−1(De Bie et al., 2002) have been reported for the Scheldt estuary. Nitrifica- tion rates higher than those estimated in the Ponteil River have been measured in the maximum turbidity zone of the Gironde estuary with values between 240 and 336 µmol l−1day−1(Abril et al., 2000). The Ponteil River aquatic system contains very little suspended particulate matter in comparison with turbid estuaries. Nevertheless, flowing waters are in contact with permeable sediments. Cores 08C and 08D collected in the riverbed downstream J2 show an accumulation of nitrate between 100 to 200 µmol l−1 in the uppermost sediment suggesting that subsurface sediments are probably the biogeo- chemical reactor where nitrification occurs (Fig. 6). Nitrification is probably enhanced in the interstitial environment because the residence time of water is higher than it is in stream. Therefore, the nitrate in stream water is probably produced in pore water and supplied by groundwater discharge toward the river.
Downstream J2, ammonium concentrations are constant, and
Fig. 9.Assessment of both, chloride and ammonium plume; black triangles refer to core station results inFig. 6; black arrow refer to the river continuum survey (stations inFig. 2and concentrations inFig. 5); ERT profiles and wells are represented.
Fig. 8.Self potential profile obtained along the southern edge of the landfill with an electrode spacing of 10 m; distance origin corresponds to the base electrode.
Diamonds correspond to raw data and the line to a mobile average on 7 data.
65 and 110 µmol l−1(Fig. 6) and the ammonium deficit is between 35 to 50 µmol l−1. In core 08C, nitrate concentra- tions are between 150 and 200 µmol l−1 (Fig. 6) and the ammonium loss is between 30 and 50 µmol l−1. Therefore, nitrification at the interface between groundwater and stream water explains ammonium attenuation. Below 5 cm depth in cores 08C and 08D, nitrate concentrations drop significantly. It suggests that nitrogen is lost, probably because of denitrification, which is an efficient pathway of N2 production in hypoxic pore waters (Brun et al., 2002;
Thornton et al., 2005).
Chloride concentrations and conductivities did not de- crease in the downgradient portion of the river survey (Fig. 5) highlighting the non-reactivity of chloride in river compared with N-compounds, as observed in other landfills (Brun et al., 2002; Thornton et al., 2005).
5.5. Redox boundary of the leachate plume
Despite the little bit noisy SP profile, the averaging curve shows a decrease in the signal with respect to distance downgradient (Fig. 8). Without the presence of the landfill, we should have observed an increase in the signal due to the electrokinetic effect, which depends on water circulation in permeable aquifers. The coupling coefficient relating piezo- metric level gradient to SP signal is generally comprised between−10 mV m−1and−1 mV m−1for water conductivity values from 0.01 to 0.1 S m−1respectively, for quartz sands, sandstones, granites and volcanic ashes (Jouniaux et al., 2009).
For example, a piezometric level diminution of−1 m induces a resulting electrokinetic current positive, with values reaching +10 mV in the most permeable aquifers. In the prospected area, from the base electrode station of the SP profile located upstream the landfill, to the end of the profile, the piezometric gradient is around 6 m (Fig. 3). Taken into account a water conductivity of P1 upstream (0.018 S m−1), we should expect to observe an increase in the SP profile up to +60 mV downstream. In fact, we observe a trend toward negative values. SP negative anomalies have been observed in organic contaminant plumes originating from landfills (Arora et al., 2007; Linde and Revil, 2007; Naudet et al., 2004).Naudet et al.
(2004)suggested that contaminant plumes with high organic content can behave as biogeobatteries and that they can be mapped by a negative SP anomaly with a drastic increase at its edges due to the redox front. In this model, a bacterial biofilm at
from the start of the profile is probably due to contaminant coming from the old portion of the landfill without geomembranes. Then, this negative SP anomaly indicates the upgradient limit of the leachate plume. We interpret this high negative SP gradient between 400 and 600 m as the upgradient redox front. At 900 m, where crushed cars storage starts (dark portion of the landfill onFig. 8), the SP signal reaches a maximum before decreasing again. The SP signal decreases until 1300 m, where it starts to increase again. If the profile would have been extended further, and the SP signal still would have increased, we could have located the downgradient redox front. Nevertheless, we can conclude that the SP profile clearly shows the presence of a reactive contaminant plume at least until 1300 m inFig. 8, i.e., 300 m downstream the landfill. The different disposal wastes seem to have a specific SP signature along this profile.
5.6. Contour of the impacted aquifer
The topographic, hydrological, geochemical, and geoelec- trical data set permits to estimate the extension of the leachate plume (Fig. 9) and the volume of contaminated aquifer. The overall profiles of ERT close to the landfill show a similar structure (Fig. 7). In the studied environment, the electrical resistivity is driven by two parameters: water saturation of soil and ionic concentration of fluid. Geology consists of a homogeneous sand formation until the aquifer substrate and cannot modify the electrical resistivity. Vadose zone has a high resistivity because pores are not saturated with water. The ERT profiles are restricted by a basement with high resistivity. This boundary may correspond here to the clay basement of the Sable des Landes aquifer. It is positioned at 11 or 12 m depth downstream, and below 12.5 m depth upstream. Compared to profile A performed upgradient the landfill, the profiles B to E show lower resistivity values in aquifer due to leachate contamination. The thickness of the leachate plume corre- sponds to the thickness of the aquifer. Since all downgradient profiles are in the contaminated area, the plume extends at least 300 m downgradient the landfill.
The southern boundary is defined by the river. The northern boundary is not precisely characterized, but it is deduced from the aquiferflow direction (Fig. 3). SP profile (Fig. 8) permits to locate the upgradient limit of the redox front shown by the black dotted line in Fig. 9. Since negative potentials are measured up to 1300 m after the electrode basement (i.e.
800 m downgradient the landfill) the leachate plume should extends up to 800 m. Data of wells (Table 1), ERT profiles B to E (Fig. 7) and cores 07B, 07C and 08B permit to define thefirst domain of the plume with maximum concentrations of 2000 µmol l−1of ammonium and 4000 µmol l−1of chloride (Fig. 9). It covers a surface area of about 0.653 km2i.e. a volume of groundwater of about 2.87·106m3, based on a porosity of 40%. The second domain is 1000 µmol l−1of ammonium and 2500 µmol l−1 of chloride. The downgradient limit is not directly given by in situ measurements but it is assumed to be between the cores 07C and 08C. The core 08B is included in this domain but not 07A, although they were close to other. So the upgradient limit is close to 08B and 07A. It covers a surface area of about 0.146 km2 i.e. a volume of groundwater of about 0.64·106m3. The third domain of the leachate plume is 50 µmol l−1 of ammonium and 1000 µmol l−1 of chloride (Fig. 9). The upgradient limit corresponds to the beginning of ammonium enrichment in stream water (arrow 5 inFig. 9). The downgradient limit corresponds to the end of the ammonium enrichment (arrow 16 inFig. 9). It is downstream to 08C, for which pore water contains up to 150 µmol l−1of ammonium.
This third domain is the maximum extent of the ammonium plume i.e. the ammonium plume is about 800 m in the landfill downgradient toward the coast. It covers a surface area of 0.385 km2i.e. a volume of groundwater of about 1.69·106m3. The related volume of contaminated aquifer is about 5.2 106m3 i.e. a quantity of 6.5·106mol of ammonium (91·103kg of N).
In cores 08A, 08C and 08D, chloride is more concentrated in pore waters than in stream water, suggesting that these stations are impacted by the chloride plume. Conversely ammonium concentration is lower in pore water than in stream water, suggesting that the ammonium plume does not reach these positions. These data suggest that the non-reactive chloride plume is more extended than the reactive ammonium plume. Based on the local groundwater velocity of 50 m year−1 the chloride plume should have extended 1700 m ahead the landfill since the beginning of the landfill exploitation (35 years). The maximum extent of the ammonium plume corresponds to a reactive plume velocity of 23 m year−1i.e.
twice slower than non-reactive plume. This behavior is frequently reported in landfill studies (Lorah et al., 2009) and plume dispersion modelling (Brun et al., 2002; Thornton et al., 2005). Sorption by cation exchange reactions at negatively charged clay mineral surface may control ammonium migra- tion in aquifers (Buss et al., 2004; Lorah et al., 2009). The cationic exchange capacity reported for soils of the studied area is between 5 and 10 meq 100 g−1(Lemoine et al., 1988; Vernier et al., 2003). Values obtained experimentally in sands sampled in the aquifer impacted by the plume are close to 1 meq 100 g−1. These values are low, but high enough to support a partial retention of ammonium relative to chloride.
6. Conclusion
This study presents an original association of geochemical and geophysical methods. The overall methods are comple- mentary and converge to give an estimation of a contami- nated plume extension. The study of pore waters of riverbed sediment was a simple approach to examine directly the link between groundwater and chemical changes in the river water (Baez-Cazull et al., 2007). Spatial and temporal
monitoring allowed us to calculatefluxes of nitrogen and to evaluate the attenuation capacities of the system. Geoelec- trical methods allowed us to detect the contaminant plume and the vertical extension of the aquifer.
The chloride plume size is about 1700 m, but the ammonium plume does not exceed 800 m, which points out the efficient attenuation process in the aquifer. The apparent ammonium plume velocity is about 23 m year−1. Both nitrification and sorption by cation exchange limit the migration of the ammonium flux. Nitrification occurs pre- dominantly at the river–groundwater interface, in the upper sediment layer, where groundwater ammonium met dis- solved oxygen of the river. Nitrification rate in June is close to that observed in estuarine turbidity maximum. The river modifies the groundwater flow and actually it drains the contaminant plume. This drainage is the outlet for the plume.
Exchanges between groundwater and river water experience a seasonal dynamics, which should be better characterized in future works. The annual ammoniumflux of Ponteil River is between 364 and 893·103mol year−1. Hence, the landfill represents about 20% of the N–NH4flux to the bay as nearly all the N–NH4 in the river comes from the landfill. The thickness of the aquifer, the extension of the plume and its mean ammonium concentration indicate that the present plume contains about 6.5·106mol. This stock is small in comparison with the annualflux of N-nitrate that is supplied from fertilizer by other rivers to the Arcachon Bay. Therefore, the landfill studied is not a potential risk for eutrophication in future years. Considering that the input of ammonium from the landfill is balanced by the present day output via the river, the residence time of ammonium in the aquifer is between 7 and 18 years. Then, if the source of contamination ceases, the ammonium plume should not reach the lagoon because at the reactive plume velocity, the groundwaters need about 100 years toflow from the landfill to the Arcachon Bay.
Acknowledgments
The authors thank S. Bujan, A. Riberi, P. Polsenaere, and J.
Deborde for their assistance during sampling campaigns and laboratory work and F. Huneau for providing the current velocity meter. Furthermore, they thank N. Le Yondre, M. Dia and ANTEA for providing information on the landfill. This work was funded by the Region Aquitaine Ph.D. fellowship to M.C. The authors gratefully acknowledge the financial support from the French program PNEC-Littoral Atlantique, and the ANR06 PROTIDAL. The French CNRS/INSU is thanked for providingfinancial support to the Biogeophysics project.
References
Abril, G., et al., 2000. Transient, tidal time-scale, nitrogen transformations in an estuarine turbidity maximum —fluid mud system (the Gironde, South-West France). Estuar. Coast. Shelf Sci. 50 (5), 703–715.
ADEME, 2007. Les déchets en chiffre : données et références, Angers.
Adriano, S., Chiara, F., Sonia, C., Antonio, M., 2005. Recording the occurrence of trophic level changes in the lagoon of Venice over the '90s. Environ.
Int. 31 (7 SPEC. ISS), 993–1001.
Anderson, L., 1979. Simultaneous spectrophotometric determination of nitrite and nitrate byflow injection analysis. Anal. Chem. Acta 110, 123–128.
Anschutz, P., et al., 2009. Tidal sands as biogeochemical reactors. Estuar.
Coast. Shelf Sci. 84 (1), 84–90.
Res. 34 (12), 3213–3221.
Brun, A., Engesgaard, P., Christensen, T.H., Rosbjerg, D., 2002. Modelling of transport and biogeochemical processes in pollution plumes: Vejen landfill, Denmark. J. Hydrol. 256 (3–4), 228–247.
Buss, S.R., Herbert, A.W., Morgan, P., Thornton, S.F., Smith, J.W.N., 2004. A review of ammonium attenuation in soil and groundwater. Q. J. Eng.
Geol. Hydrogeol. 37 (4), 347–359.
Canton M. Dynamique des éléments biogènes dans le continuumfluvio- estuarien de la zone hydrologique d'influence du Bassin d'Arcachon.
Thesis. Université Bordeaux 1, 2009, pp. 169.
Carley, B.N., Mavinic, D.S., 1991. The effects of external carbon loading on nitrification and denitrification of a high ammonia landfill leachate. J.
Water Pollut. Control Fed. 63 (1), 51–59.
Cébron, A., Berthe, T., Garnier, J., 2003. Nitrification and nitrifying bacteria in the lower Seine River and Estuary (France). Appl. Environ. Microbiol. 70 (11), 6726–6737.
Chu, L.M., Cheung, K.C., Wong, M.H., 1994. Variations in the chemical properties of landfill leachate. Environ. Manage. 18 (1), 105–117.
Dai, M., et al., 2008. Nitrification and inorganic nitrogen distribution in a large perturbed river/estuarine system: the Pearl River Estuary, China.
Biogeosciences 5 (5), 1227–1244.
Darilek, G.T., Parra, J.O., 1989. The electrical leak location method for geomembrane liners. J. Hazard. Mater. 21 (2), 177–187.
Dave, G., Nilsson, E., 2005. Increased reproductive toxicity of landfill leachate after degradation was caused by nitrite. Aquat. Toxicol. 73 (1), 11–30.
De Bie, M.J.M., Starink, M., Boschker, H.T.S., Peene, J.J., Laanbroek, H.J., 2002.
Nitrification in the Schelde estuary: methodological aspects and factors influencing its activity. FEMS Microbiol. Ecol. 42 (1), 99–107.
De Wit, R., et al., 2001. Robust: the role of buffering capacities in stabilising coastal lagoon ecosystems. Cont. Shelf Res. 21 (18–19), 2021–2041.
De Wit, R., et al., 2005. Relationship between land-use in the agro-forestry system of les Landes, nitrogen loading to and risk of macro-algal blooming in the Bassin d'Arcachon coastal lagoon (SW France). Estuar.
Coast. Shelf Sci. 62 (3 SPEC. ISS), 453–465.
Deborde, J., et al., 2008. Role of tidal pumping on nutrient cycling in a temperate lagoon (Arcachon Bay, France). Mar. Chem. 109 (1–2), 98–114.
Ehrig, H.J., 1989. Leachate Quality. Sanitary Landfilling, pp. 213–229.
Foose, G.J., Benson, C.H., Edil, T.B., 2002. Comparison of solute transport in three composite liners. J. Geotech. Geoenvironmental Eng. 128 (5), 391–403.
Fournier, C., 1989. Spontaneous potentials and resistivity surveys applied to hydrogeology in a volcanic area: case history of the Chaine des Puys (Puy-de-Dome, France). Geophys. Prospect. 37 (6), 647–668.
Garnier, J., et al., 2002. Modelling the transfer and retention of nutrients in the drainage network of the Danube river. Estuar. Coast. Shelf Sci. 54 (3), 285–308.
Garnier, J., Billen, G., Cébron, A., 2007. Modelling nitrogen transformations in the lower Seine river and estuary (France): impact of wastewater release on oxygenation and N20 emissions. Hydrobiologia 588 (1), 245–259.
Gettinby, J.H., Sarsby, R.W., Nedwell, C., 1996. The composition of leachate from landfilled refuse. Proceedings of the Institution of Civil Engineers.
Munic. Eng. 115 (1), 47–59.
Hall, O.J., Aller, R.C., 1992. Rapid, small volume,flow injection analysis for ΣCO2 and NH4+ in marine and freshwaters. Limnol. Oceanogr. 37, 1113–1119.
Hanson, D.L., Caponi, F.R., 2008. US landfill disposal: the big picture. MSW Manage. 18 (3).
Jickells, T., 2005. External inputs as a contributor to eutrophication problems.
J. Sea Res. 54 (1 SPEC. ISS), 58–69.
sediment–alluvial aquifer interface in a landfill leachate plume. J.
Contam. Hydrol. 105 (3–4), 99–117.
Montuelle, B., Volat, B., Torio-Fernandez, M.M., Navarro, E., 1996. Changes in nitrobacter serotypes biodiversity in a river: impact of a wastewater treatment plant discharge. Water Res. 30 (5), 1057–1064.
Mullin, J.B., Riley, J.P., 1955. The spectrophotometric determination of silicate–silicon in natural waters with special reference to sea water.
Anal. Chem. Acta 12, 162–170.
Murphy, J., Riley, J.P., 1962. A modified single solution method for determination of phosphate in natural waters. Anal. Chem. Acta 42, 31–36.
Naudet, V., Revil, A., Bottero, J.-Y., Bégassat, P., 2003. Relationship between self-potential (SP) signals and redox conditions in contaminated groundwater. Geophys. Res. Lett. 30 (21) HLS 2-1 - HLS 2_4.
Naudet, V., Revil, A., Rizzo, E., Bottero, J.-Y., Bégassat, P., 2004. Groundwater redox conditions and conductivity in a contaminant plume from geoelectrical investigations. Hydrol. Earth Syst. Sci. 8 (1), 8–22.
Pivato, A., Gaspari, L., 2006. Acute toxicity test of leachates from traditional and sustainable landfills using luminescent bacteria. Waste Manage. 26 (10), 1148–1155.
Prepas, E.E., Charette, T., 2003. Worldwde eutrophication of water bodies:
causes, concerns, controls. In: Lollar, B.S. (Ed.), Treatise on Geochemistry.
Elsevier, pp. 311–331.
Rimmelin, P., Dumon, J.C., Maneux, E., GonCalves, A., 1998. Study of annual and seasonal dissolved inorganic nitrogen inputs into the Arcachon Lagoon, Atlantic Coast (France). Estuar. Coast. Shelf Sci. 47 (5), 649–659.
Saint Pe, M.A., 1966. Etude Hydrogéologique et Géochimique: la Nappe Phréatique du Nord-est du Bassin d'Arcachon et son Influence sur l'Alimentation des Aquifères Sous-Jacents., Université de Bordeaux. 439 pp.
Samouëlian, A., Cousin, I., Tabbagh, A., Bruand, A., Richard, G., 2005. Electrical resistivity survey in soil science: a review. Soil Tillage Res. 83 (2), 173–193.
Sophocleous, M., 2002. Interactions between groundwater and surface water: the state of the science. Hydrogeol. J. 10 (1), 52–67.
Stookey, L.L., 1970. Ferrozine—a new spectrophotometric reagent for iron.
Anal. Chem. Acta 42, 779–781.
Strickland, J.D.H., Parsons, T.R., 1972. A practical handbook of seawater analysis. Fish. Res. Board Can. Bull. 167, 1–31.
Teissier, S., Garabetian, F., Torre, M., Dalger, D., Labroue, L., 2002. Impact of an urban centre on the nitrogen cycle processes of epilithic biofilms during a summer low-water period. River Res. Appl. 18 (1), 21–30.
Thornton, S.F., Tellam, J.H., Lerner, D.N., 2005. Experimental and modelling approaches for the assessment of chemical impacts of leachate migration from landfills: a case study of a site on the triassic sandstone aquifer in the UK East Midlands. Geotech. Geol. Eng. 23 (6), 811–829.
Titov, K., Ilyin, Y., Konosavski, P., Levitski, A., 2002. Electrokinetic spontane- ous polarization in porous media: petrophysics and numerical model- ling. J. Hydrol. 267 (3–4), 207–216.
Touze-Foltz, N., Rowe, R.K., Navarro, N., 2001. Liquidflow through composite liners due to geomembrane defects: nonuniform hydraulic transmissiv- ity at the liner interface. Geosynthetics Int. 8 (1), 1–26.
Vernier, F., Beuffe, H., Chossat, J.C., 2003. Forêt et ressource en eau : étude de deux bassins versants en sol sableux (Landes de Gascogne). Revue Forestière Française, 6.
Ward, M.L., Bitton, G., Townsend, T., Booth, M., 2002. Determining toxicity of leachates from Florida municipal solid waste landfills using a battery-of- tests approach. Environ. Toxicol. 17 (3), 258–266.
