Résumé
A superconducting gravimeter (SG) has been installed since 2010 in Nalohou, northern Benin, within the framework of the GHYRAF (Gravity and Hydrology in Africa) project. This site was chosen to monitor the strong annual monsoon signal with both local and non-local hydrological contributions within the humid Sudanian zone of West-Africa. The area is also part of the long-term observing system AMMA-Catch, and thus under intense hydrological monitoring. We present here the results of the first two years relative gravity monitoring. The signal includes predominantly solid earth tides, ocean loading, polar motion, atmospheric pressure effects, drift and water storage changes (WSC). Retrieving WSC needs thorough corrections of other components, and detailed tide and barometric analysis are thus undertaken. Pressure effects are of major concern in the equatorial band, because they are governed by S1 and S2 thermal pressure waves. These waves dominate both the local Newtonian effect (an increase in local pressure decreases the gravity) and the smaller non-local loading effect (an increase in regional pressure decreases the gravity by a subsidence effect of the elastic earth) because of their coherency at the regional scale. FG5 absolute gravity data are used for calibration and drift estimate of the SG. Residuals clearly show interesting WSC behaviors at two predominant frequencies for which different accuracies are involved: the seasonal and the rainfall event time-scales. The weathered hard rock shallow aquifer of the site is known to produce WSC only in the 0-7 m depth range (7 m being the lowest interannual level of the water table within our record). WSC are coming both from soil moisture and water table variations. These are monitored 1) by weekly neutronic measurements over the whole vertical profile (every 50 cm) in a borehole close to the SG, and 2) by water table level observations for the saturated zone only. The contribution to gravity of the latter is evaluated with a 2D model for specific yield (Sy), inferred from resistivity mapping and magnetic resonance soundings (MRS), and previously calibrated with FG5 gravity data. This work was undertaken to account for known spatial variability in this metamorphic basement context. The main results arising from this study can be summarized in two points. First, gravity effect of seasonal WSC is calculated and compared to SG record. Results show a good fit, confirming a previous study based on FG5 data only. Second, rainfall effects on SG records are cautiously analyzed for each significant event. This includes both the rapid increase in gravity following a precipitation (which is always lower than expected, the difference being attributed to runoff during the time span of the event), and the slower decrease afterwards, related to evapotranspiration rates and lateral underground water redistribution. Results confirm the absolute need for most appropriate corrections in SG data to properly retrieve the local hydrological signal. However, significant integrative information on WSC can be derived at these two frequencies, enlightening once more the potential for gravity monitoring in hydrological studies.