Hydraulic and geochemical impact of occasional saltwater intrusions through a submarine spring in a karst and thermal aquifer (Balaruc peninsula near Montpellier, France)

14 Submarine springs are a common discharge feature of the karst aquifers along the 15 Mediterranean coast. In some instances, occasional and localized saltwater intrusions can occur 16 through the submarine spring and negatively impact the quality of the groundwater resource. 17 The hydraulic and geochemical behavior of a submarine spring discharging into the Thau 18 lagoon just offshore of the Balaruc peninsula near Montpellier, France has been characterized 19 to determine the impacts of such phenomena to better understand the dynamics of a regional 20 karst aquifer and improve its groundwater management. This work is based on both historical 21 and new hydrogeological and geochemical data, illustrating 6 occasional saltwater intrusion 22 events (from 1967 to 2014) in the Thau lagoon area (southern France). 23 Hydraulic perturbation of the aquifer is propagated instantly within the Balaruc-les-Bains 24 peninsula and reaches a distance of about 5 km upgradient within 9 days. Comparison of 25 hydraulic heads during seawater intrusion events in 2010 and 2014 indicates an aggravation of 26 the phenomenon with an increase in hydraulic head variations. 27 In contrast, isotopic tracers ( 87 Sr/ 86 Sr, D/H, 18 O/ 16 O,) and Rare Earth Elements (REE) 28 demonstrate that the geochemical impact of these inversac events is only observed at the local by acid titration with 0.1N HCl. Major ion (Cl - , NO 3- , SO 42- , Ca 2+ , Mg 2+ , Na + , and 215 K + ) were analyzed by ion chromatography (ICS 1000 Dionex ® ). Precision error was < ± 5%. 216 After acidification with 1% HNO 3- , trace elements (Li, B, Sr, REE and U) were analyzed by 217 inductively-coupled mass spectrometry using a Thermo Scientific ® iCAP Q at the platform of at the The use of an in-line Argon Gas Dilution system permits the direct injection of highly mineralized samples without prior dilution. Precision error was typically <5%. Fresh water reference material SLRS-6, and seawater reference materials CASS-6 and for trace metals analyzed every 20 samples to monitor analytical accuracy. Mean results are and recovery

these springs can support economic development or be exploited for drinking water supply.

42
In certain instances, especially during periods of low discharge, flow can reverse and instead 43 intrude into the upgradient karst aquifer, causing an adverse impact on the quality of the 44 groundwater resource and the economic activities that depend on it (Arfib and Gilli, 2010; 45 Drogue and Bidaux, 1986). This phenomenon, termed "inversac" in French (Gèse, 1987; 46 Pinault et al., 2004), was previously described by Albéric (2004) and Joigneaux et al. (2011) in 47 a related context of river backflooding into karst springs. 48 The factors responsible for initiating and terminating inversac events are not particularly well 49 understood and the recovery of carbonate aquifers after a seawater intrusion is poorly described 50 (Han et al., 2015). Yet, it is necessary to understand the mechanisms controlling this 51 phenomenon, as well as the long-term impact of occasional and localized saltwater intrusions, 52 to appropriately manage the groundwater resources of such karst systems.

53
In this context, a hydraulic and geochemical study was undertaken to characterize the The Thau hydrosystem in the Balaruc-les-Bains area is an example of a karst and thermal 62 aquifer affected by occasional saltwater intrusions through a submarine spring. This complex 63 hydrologic system has been described by Aquilina et al. (1997Aquilina et al. ( , 2002Aquilina et al. ( , 2003, Doerfliger et al.  Table 1. Summary of the observed occasional saltwater intrusions (inversac events) in the karst 96 and thermal aquifer of the Thau basin.

97
A multi-tracer approach was followed to better define the origin and contribution of the thermal,   Séranne, 1995;Séranne, 1999;Thaler, 1962). Then, the sea level rise of the Early Miocene 121 (Burdigalian) deposited transgressive sediments across the region (Fig. 1b). The overall 122 sedimentary record associated with these Upper Cretaceous to Middle Miocene geological 123 events, includes a wide range of detrital and carbonate facies deposited in marine, lacustrine 124 and fluvial environments (Combes, 1990; Marchand, 2019) (Fig. 1a).

125
The sea level fall related to the Messinian Salinity Crisis caused major river incision across 126 southern France which, as a consequence, deepened the karst system (Clauzon, 1982;Hsu, 127 1973; Ryan, 1976). The Messinian paleovalley in the coastal area was then flooded during the           within the range of certified uncertainties. Precision error for all analyses was typically <5%.

224
The REE data are represented in profiles, after normalization to a reference geological material, 225 the North American Shale Composite (NASC) for natural waters, which corresponds to an 226 average sample of North American shales (Taylor and McLennan, 1985).      This hydrogeological setting illustrated in Figure 2 was confirmed by the F13 exploratory well 276 (Table 2), that displayed contrasting electrical conductivity and temperature over depth.

277
Under normal hydrologic conditions (Fig. 2a), groundwater from the karst aquifer discharges 288 Figure 1a.   Table 3 Characteristics of the main wells and springs in the study area      hydrologic conditions during these periods (Fig.5 a). However, the hydraulic head difference, 374 with respect to the reference measured before the inversac (Fig. 5b)   Results of the mixing calculations are included in the Supplement (Table S1). This calculation 455 gives an indication of the magnitude of the karst water and saltwater component in of each 456 water sample, but does not provide additional insight into the origin of groundwater salinity.

522
The contribution of each end-member (saltwater, thermal water and karst water) to groundwater Results of the mixing calculations are in the Supplement (Table S2).

529
The Sr-isotopic composition of the Cauvy spring is quite variable during the reference interval,

563
The thermal well F8 does not seem to be impacted by a direct intrusion of lagoon water during 564 the inversac event but instead influenced by seawater already within the hydrosystem. There is a good alignment of the data points in Fig. 9, which illustrates the process of water/rock 597 interaction and indicates that the B in the waters across the Balaruc peninsula is derived from 598 clay sediments. That 11B/10B ratios are well above the seawater values, suggests that water 599 from the Thau lagoon interacts with clay sediments within the karst system during the  The impact on water quality is local whereas the hydraulic impact is perceived at the peninsula. The geochemical impact is, therefore, restricted to within a 1-kilometer radius from 680 the Vise spring (Fig. 10).

681
This slow geochemical recovery is best explained by the characteristic internal heterogeneity 682 of the karst aquifer, which is typically described in terms of a triple porosity model comprising:  from the Cauvy spring, followed by its subsequent abandonment because the quality of this 713 major drinking water supply was strongly impacted by the inversac events.

714
The Issanka spring is another major source of drinking water in the area for a population of The results of this study support groundwater management across the Balaruc peninsula. The