Abstract
Evapotranspiration (ET) is a key process of the water cycle in general and in ecohydrology in particular. Measuring ET in forest eco-hydrosystems allows us to gain a better understanding of the response of forests to drought events, and to better anticipate the effects of climate change. Punctual (e.g. lysimeters or sapflow measurements) or integrative measurement methods (e.g. eddy covariance tower) can be used to estimate ET at the forest stand scale but these methods are not without limitations (e.g., resolution issues, representativeness, not adapted to mountainous areas).Superconducting gravimeters can be used to study ET. These gravimeters can be deployed in both flat and mountainous environments. In this work, we studied the hydrological residuals (i.e., hydrologically induced gravity variations) of 5 superconducting gravimeters located in different contexts. We interpreted the daily decreases in the stacked hydrological residual as the loss of water mass due to evapotranspiration. These results were compared with those of the SimpKcET water balance model.The results underline that the detectability of the ET signal depends strongly on the configuration of the gravimetric station, the topography and the type of ecosystem. We show that gravimeters located on summit area and in a forested context can detect the seasonality of ET. Conversely, gravimeters located in flat or underground areas and with a significant masking effect are unable to detect ET.Gravimetry therefore has a strong complementarity with conventional methods used to study ET and could contribute to a better understanding of water fluxes in forested ecosystems.