Accueil > Enviro'Fos / Gardanne > Plate-forme Enviro'Fos / Gardanne > Coupling geophysical and isotopic approaches to better simulate saltwater intrusion into coastal aquifers : a case study in the Crau aquifer
La recherche a porté sur les contraintes reliées à la modélisation des écoulements des eaux souterraines en aquifère côtier.
Dans ces aquifères, la présence concomitante d'eaux douces et d'eaux salées modifient les patrons d'écoulement classiquement représentées par l'équation de Darcy. L'influence de la salinité de l'eau sur ses propriétés (densité) et sur la perméabilité du milieu complexifie la représentation mathématique des équations d'écoulement et la résolution des systèmes d'équations. Par ailleurs, l'influence des hétérogénéités du milieu rend plus difficiles et hasardeux le paramétrage des modèles et leur calibration. Pour ce faire, nous avons proposé d'utiliser les résultats d'investigations géophysiques et isotopiques des eaux. Une meilleure description du milieu souterrain et l'imagerie de l'interface eau douce/eau salée peut en effet aider à un meilleur paramétrage et à une validation accrue de la simulation des écoulements des nappes côtières. Les mesures géochimiques telles celles des isotopes du radium ont aidé à la validation des flux et concentrations simulés. Notre approche a été appliquée sur la partie aval de la nappe de la Crau. La comparaison des résultats simulés par le modèle aux données de terrain résultant des suivis réalisés ont montré l'intérêt et l'applicabilité et la validation de notre approche méthodologique.
The use of groundwater models is prevalent in the field of environmental science to investigate a wide variety of hydrogeological conditions. More recently, groundwater models are being applied to predict the transport of contaminants or saltwater intrusion for risk evaluation. The applicability or usefulness of a model depends on how closely the mathematical equations approximate the physical system that is being modeled. In order to evaluate the applicability or usefulness of a model, it is necessary to have a thorough understanding of the physical system and the assumptions embedded in the derivation of the mathematical equations.
To simulate the variable-density flow system in heterogeneity hydrogeological conditions is complicated and sensitive due to the non-linear coupling between the flow and transport equations. Thus, observed data of salinity from wells and boreholes are normally not enough to validate a saltwater intrusion model. The aim of this study was to propose an approach to control and reduce the effects of uncertainties and heterogeneities of variable- density groundwater modeling into coastal aquifers. Combination of different methods from geophysics and isotope can help to better parameterize and validate the saltwater intrusion modeling.
First, radon has been used to assess groundwater velocities in the study aquifer, to highlight pattern of groundwater discharges and to constraint water-mass balance. A monitoring of 222Rn in boreholes and wells have been done in the Crau aquifer. It is likely that the radon activity in groundwater is controlled by salinity and grain size of sediments in the aquifer where the wells are located. The radon activity measurements in boreholes have been used to estimate the groundwater (Darcy) flow velocity applying the approach of Schubert and Hamada. In our study, the decay of Radon was measured at different depths in boreholes to achieve information on hydraulic conductivity and heterogeneity of aquifer. The groundwater velocities evaluated with radon activity vary from few cm/day to more than 1 m/day. Based on salinity and radon activity observed in boreholes, the origin of saltwater in the aquifer seems to be related to ancient seawater and not current seawater intrusion, this is in agreement with previous studies from Vella et al. 2005 and De Montety. 2008. Our approach is useful to areas where it is difficult to operate pumping tests (lack of groundwater wells; due to the limited diameter of wells it may not be possible to install measuring devices) or tracing tests (in areas where the tracers can change the natural hydro-geochemical conditions).
Radon activity in groundwater and surface water has been used to estimate the exchange between groundwater and surface water. Using continuous measurements of 222Rn and the balance box model of Burnett et al., the water mass-balance of groundwater and surface water was calculated. Radon surveys in surface waters suggest no significant discharges of groundwater to surface water (55 l/m2/day). This method is useful for large areas of surface water and/or too deep to install a seepage meter.
Geophysical methods are very useful techniques to inform on the hydrogeological characteristics and calibrate models. Inversed 2D and 3D geophysical models provide high resolution datasets of subsurface structure at a low cost and in a short time. However, this technique still faces difficulty while different resistivity models may produce the same apparent geophysical effect. Therefore, to reduce the errors and uncertainties in geoelectrical models, it is necessary to compare these models with geological and hydrogeological data.
In the study area, the saltwater front is located around the marsh area where it is distributed in a dense network of ponds, canals and the bulls grazing fields of the bull husbandry industry. Because space for geophysical measurements is very constrained, electromagnetic methods appear useful and need little measurement space. An electromagnetic method using a EM34 device have been applied for mapping the saltwater intrusion with a total investigation length of more than 30km. This method is very simple and rapidly operated at a low cost and on a restrained surface. In order to reduce the errors and uncertainties and to validate results of EM method, a combination of difference geophysics techniques is strongly recommended. Electrical Resistivity Tomography (ERT) was the second choice and developed within this study in three profiles. The advantage of the ERT method is the quality of the electrical resistivity data obtained with relatively high spatial resolution. Both the EM and ERT can be coupled to obtain a continuous coverage of the underground in 2D and 3D spaces. The EM34 data have been interpreted using the software EM4Soil and ERT resistivity interpretation have been done with RES2DINV.
Based on the apparent resistivity/conductivity values obtained from ERT and EM investigations, 3D saltwater distribution was characterized. The low resistivity area is located in the southwest of the study area. Low resistivities were found from 4-5 m.asl close the southwest boundary to more than 20 m.asl near X34, X35 (about 1.7km from the boundary). Observations of water in wells and boreholes also indicate brackish water. In marsh area, low resistivity was found near the surface confirming the presence of a top clay layer and the salinity of surface water caused by evaporation. Outside this area no indication of saltwater or saltwater intrusion has been found.
Based on ERT results, porosity of the aquifer was estimated using the Archie's law for each layers of the aquifer. Data provide detailed information on geometry of the aquifer that can be further interpreted using the lithological descriptions from boreholes.
A 3D finite element model was built with FEFLOW software to simulate groundwater flow and the transition between saltwater and freshwater. All mechanisms which affect saltwater intrusion in the aquifer are considered in this model, including advection, dispersion, adsorption, diffusion and chemical reactions. Our model has combined existing data from different sources to estimate the groundwater flow, the mass balance and the effects of recharge
and pumping rates in difference management scenarios. With knowledge obtained from geophysics and isotope measurements, hydrogeological parameters such as hydraulic conductivity, porosity and lithology have been achieved to reduce the errors and the uncertainties of modeling.
Coupling geophysics and groundwater modeling may help to describe the spatial characteristics of the subsurface. Geometric constraints were used to delineate geological units (elevation and thickness of layers) and distribution of heterogeneous properties. The variation of electrical conductivity is also responding to the variation of the lithology. ERT and EM methods are thus applied since the electrical conductivity is sensitive to both clay content and groundwater salinity. With comparison to geological information provided from boreholes, a detailed geological unit distribution has been done. Based on this coupling approach, the simulated numerical model is validated and flux and mass budgets are results of this study.
Flow modeling was calibrated by fitting the simulated and observed hydraulic head in observation points and validated using isotope results and geophysics. A good match was achieved between observed and simulated groundwater heads and concentrations for both models in the steady state and in the transient state. The simulated results from groundwater modeling have detailed the recharge/discharge between groundwater and surface water in marsh area. Results of the numerical model show a significant intrusion of the saltwater front in the southwestern part of the study area.
The developed model was applied to simulate displacement of saltwater intrusion resulting from different scenarios of recharge and pumping rates. The first scenario considers no pumping while the second scenario considers doubling the current pumping rate. The third and fourth scenarios test a non-recharge and double recharge condition respectively over the study area. Simulated results clearly show the effects of pumping and recharge on water table and on saltwater intrusion. Compared to the current situation, water head would drawdown approximately 0.3-0.5 m if pumping rate double. Conversely, if all pumping wells were to stop then the water head would increase by 0.2-1.0 m. Water heads simulated for non-recharge over this area show a global decrease of about 0.2-0.5 m. The simulated mass transport model also indicated a significant change in saltwater intrusion due to activities such as irrigation and abstraction. Doubling the pumping rate induced a landward displacement of saltwater front up to 50 m in one year in SE area (near Pissarotte pumping station). On the other hand, reducing the pumping rate in this area will help mitigate the seawater intrusion. In case of no pumping wells, the saltwater front displaced downward of approximate 130 m/year.
The coupling geophysics and isotope techniques appears to be a very useful tool to help saltwater intrusion modeling in coastal aquifers. Such methods can be widely applied to other sectors or others scales because of their ease of application. This approach could be applied to
study of the effects of climate change, sea level rise and over pumping leading to saltwater intrusion in coastal aquifers.
Future works may increase the potential and the reliability of the proposed approaches as follows:
- More applications should be carried out to achieve more information on hydrogeological parameters (porosity, hydraulic conductivity and storativity) to better constraint the effect of errors and uncertainties on the models, such as pumping tests or slug tests in boreholes.
- A complicated relationship between groundwater and surface water in this area could be clarified by using seepage meters in the field to quantify fluxes at the surface water.
- Extent the boundaries of the current model to the Mediterranean Sea in the southern area and to Rhone River in the western area to test the possible effect of such extended boundary conditions and simulate the effect of sea level rise due to climate change.
- Others geophysical and isotopic (Radon and Radium) investigations can be carried out in the same profiles during rainy season to consider the impact of recharge to groundwater flow pattern and the saltwater distribution.
- To clarify the hydrogeological information (porosity, lithology) in a vertical distribution, a borehole logging method in this area is recommended to clarify the heterogeneity of the aquifer.
- Scenarios develop for long periods. - Apply this method to other areas.
Vous souhaitez apporter votre contribution à Enviro'Fos/Gardanne ?Un formulaire de contact est à votre disposition afin de nous faire parvenir de nouvelles études ou vos remarques et éventuelles modifications à apporter au contenu présenté.
Aucun jugement, ni dans la qualité, ni dans l'exactitude des résultats des études, n'est porté par le SPPPI PACA et les partenaires du projet Enviro'Fos/Gardanne.