Gravity effect of water storage changes in a weathered hard-rock aquifer in West Africa: results from joint absolute gravity, hydrological monitoring and geophysical prospection

International audienceAdvances in groundwater storage monitoring are crucial for water resource management and hydrological processes understanding. The evaluation of water storage changes (WSC) often involve point measurements (observation wells, moisture probes, etc.), which may be inappropriate in heterogeneous media. Over the past few years, there has been an increasing interest in the use of gravimetry for hydrological studies. In the framework of the GHYRAF (Gravity and Hydrology inAfrica) project, 3 yr of repeated absolute gravity measurements using a FG5-type gravimeter have been undertaken at Nalohou, a Sudanian site in northern Benin. Hydrological data are collected within the long-term observing system AMMA-Catch. Once corrected for solid earth tides, ocean loading, air pressure effects, polar motion contribution and non-local hydrology, seasonal gravity variations reach up to 11 μGal, equivalent to a WSC of 260-mm thick infinite layer of water. Absolute temporal gravity data are compared to WSC deduced from neutron probe and water-table variations through a direct modelling approach. First, we use neutronic measurements available for the whole vertical profile where WSC occur (the vadose zone and a shallow unconfined aquifer). The RMSD between observed and modelled gravity variations is 1.61 μGal, which falls within the error bars of the absolute gravity data. Second, to acknowledge for the spatial variability of aquifer properties, we use a 2-D model for specific yield (Sy) derived from resistivity mapping and Magnetic Resonance Soundings (MRS). The latter provides a water content (θMRS) known to be higher than the specific yield. Hence, we scaled the 2-D model of θMRS with a single factor (α). WSC are calculated from water-table monitoring in the aquifer layer and neutronic measurements in the vadose layer. The value of α is obtained with aMonte-Carlo sampling approach, minimizing the RMSD between modelled and observed gravity variations. This leads to α = Sy/θMRS = 0.63 ± 0.15, close to what is found in the literature on the basis of pumping tests experiments, with a RMSD value of 0.94 μGal. This hydrogeophysical experiment is a first step towards the use of time-lapse gravity data as an integrative tool to monitor interannual WSC even in complicated subsurface distribution