Hydrochemical assessment of the Rochefort marsh: Role of surface and groundwater in the hydrological functioning of the wetland

Abstract

The interaction between a wetland (the Rochefort agricultural marsh) and its underlying aquifers was studied using both geochemical and isotopic approaches. Water from the various components of the hydrosystem (water from ditches, shallow ‘Bri’ groundwater (silty clay deposits), deep ‘Bri’ groundwater and groundwater in the Kimmeridgian aquifer (confined and unconfined) was sampled and analysed for major and trace elements (Sr, Br), stable water isotopes ( 18O and 2H) and Sr isotopes. The salinity of groundwater in the marsh (Cl up to 15.5 g/l) can be attributed to seawater trapped in the clay sediments during the Flandrian (12,000–6400 BP). The groundwater has since undergone more or less major chemical evolution due to water–rock interactions and mixing with fresh water (rainwater, ditch water and Kimmeridgian groundwater from the edge of the marsh). The chemical characteristics of the groundwater in the confined Kimmeridgian underlying the Bri formation exclude the possibility of a salt wedge. The combined use of δ 18O, δ 2H, Cl, Br and Sr contents and 87Sr/ 86Sr ratios enabled exchanges between surface water in marsh canals and ditches and groundwater in the shallow Bri aquifer (<2 m), and between groundwater in the deep Bri (>7 m) and in the confined Kimmeridgian (>18 m) aquifers, to be defined and quantified. During the summer, the ditches recharge the shallow Bri aquifer by sub-irrigation (gravity flow) by up to 26%. The presence of a very compact layer of clay, 1–2-m deep, prevents the rapid recharge of the deep Bri aquifer by rainfall and direct hydraulic connections between the ditches and the deep Bri aquifer. On the scale of the marsh, the confined Kimmeridgian groundwater recharges the Bri aquifer by upward flow (maximum contribution estimated to be 76±13%), this flow being relatively limited and preferentially on the edge of the marsh. The deep Bri groundwater contributes very locally to recharging the confined Kimmeridgian aquifer (maximum contribution estimated to be 61±36%). This study demonstrates the interest in combining hydrological and geochemical tools in order to identify and quantify the hydraulic interconnections between a wetland and its groundwater components. A prior understanding of how wetlands function is crucial for the sustainable management of such ecosystems.