Résumé
The rainfall isotopic composition, which is a prerequisite for any isotope water cycle study, is generally characterized by important spatial and temporal variations. While δ18O or δ2H are driven, at first order, by the Rayleigh distillation mechanism, leading to a progressive decrease as vapor condensation progresses (leading to the so-called amount, orographic, and continental effects), the combination of both tracers (d-exc = δ2H – 8 x δ18O) generally informs about vapor sources and mixing processes. In tropical areas, the rainfall isotope composition is largely controlled by convective processes associated with the monsoon. Nevertheless, the huge observed variations reflect complexatmospheric processes, from the initial water vapor formation above the ocean until the final rainfall drop at the soil surface, making it difficult to decipher simple dominant controls. These processes combine different phase changes (e.g. condensation processes in the clouds, rainfall re-evaporation) and mixing between different vapor masses, including continental vapor produced by evaporation and plant transpiration.Recent studies highlighted the added-value of using an additional tracer, namely the 17O-excess of water (17O-excess = ln (δ17O + 1) – 0.528 × ln (δ18O + 1)) for characterizing evaporative processes in rainfall (e.g. Luz and Barkan, 2010; Landais et al 2010 ; Tian et al., 2018), surface water (Surma et al., 2015; Surma et al., 2018), and plant water (Landais et al., 2006; Li et al., 2017; Alexandre et al., 2018). In contrast to d-exc, and except at very high latitudes, the 17O-excess is not significantly impacted by temperature and is much less sensitive to Rayleigh distillation processes. Its variations aretherefore principally driven by evaporation processes and may provide key information on evaporation rates, relative humidity conditions, and more generally non-equilibrium processes. In rainfall water, small variations in 17O-excess are expected, mainly controlled by evaporation conditions at the oceanic water source (Luz and Barkan, 2010). In convective monsoon systems, raindrop re-evaporation may play an additional control (Landais et al., 2010). However available data are still insufficient to fully characterize and understand this tracer in atmospheric water. n this study, we characterize the stable isotope compositions (δ2H, δ18O and δ17O) of rainfall in the monsoonal sudanian climatic zone of Benin (West Africa) from two sampling stations installed in April and June 2018. The first station is located at the University of Abomey-Calavi in Cotonou (lat. 6°26’ N; long. 2°21’ E, sampling every ten days), close to the shoreline of the Gulf of Guinea. The second station is located 400 km northern, ‘downwind’ from the main oceanic air mass trajectories, in the AMMA-Catch observatory site (http://www.amma-catch.org/) in Djougou (lat. 9°44’ N; long. 1°34’ E, bi-weekly sampling). Both stations have an average annual rainfall close to 1200 mm with different seasonal distribution. At Djougou, 90% of the annual rainfall occurs in seven months (April o October), while the rainfall season in Cotonou extends from March to November, and is split into two periods separated by a short dry period (mid-July to mid-September). Stable isotopes of the water molecule (δ2H, δ18O and δ17O) are measured at CEREGE on a WS-CRDS Picarro L2140-i. In addition to a classical δ18O and δ2H characterization of rainfall times series, we will compare the 17O-excess and d-exc seasonal variations at the two stations and investigate how evaporation at the oceanic source, water transport, potential raindrop re-evaporation and relative humidity at the rainfall sites, imprint the isotope signatures. Beyond the rainfall isotope characterization, these results willprovide a rainfall background for explorating the triple oxygen isotopic fractionation at play in the water cycle at the soil/plant/atmosphere interface, in the framework of an ongoing ANR project (HUMI-17, https://www.cerege.fr/fr/2017-anr-humi-17).