Partitioning of evapotranspiration into soil evaporation and plant transpiration using isotopes of water in controlled conditions

Rainfall recycling by evapotranspiration from continental surfaces is certainly the most unknown component of the global water cycle. This is due to the large variability of rainfall as well as the heterogeneity of these continental surfaces, both in time and space. Traditional measuring methods such as sap flow, micro lysimeter, water and energy balance estimation (Bowen ratio, eddy correlation) have been used since the 70s for a monitoring of real evapotranspiration fluxes over crops and others plant covers. A complementary method consists in using isotopic biogeochemistry. When making specific hypothesis, it is possible to identify and quantify the different sources of the atmospheric water vapour (vegetation and soil at different scales). Analysis of the heavy stable isotopic ratios of water in both liquid and vapour phases: 18O and 2H can allow determining the history of the water in the soil since the last rainfall event (infiltration, re-evaporation) or the root extraction depths. Field campaigns measurements (plants and soils), interpreted using the Keeling Plot method allowed some progress in the partition between evaporation and transpiration understanding. But the experimental design is not sufficient to mechanistically describe the water processes involved. The study of all the interactions is difficult due to the large number of controlling variables describing climate, vegetation and soil characteristics. A monolith experiment (including soil and growing plant) was carried out in a reactor called RUBIC (Reactor Used for Continental Isotopic Biogeochemistry, Bariac et al., Geochim. Cosmochim. Acta., 1991). Controlled conditions allowed a monitoring and regulation of climatic parameters (net radiation, air temperature, vapour pressure deficit, CO2 partial pressure, and wind speed). It was also necessary to fix soil (structure, texture, and water content) and vegetation (specie and seeding density) parameters. The collected data allow us to improve our understanding of the partition of evapotranspiration into soil evaporation and plant transpiration and to assess the hypothesis (often made in isotopic biochemistry) of a stationary state reached in the two reservoirs (soil and plant). These data also allow the evaluation of the hypothesis included in a transfer module of heavy stable isotopes of water within the bare soil and the plant (Braud et al., Journ. of Hydrol., 2005). The latter is coupled to a SVAT model (Soil-Plant-Atmosphere Transfer) called SiSPAT (Simple Soil Plant Atmosphere Transfer model, Braud et al., Journ. of Hydrol., 1995) and was extended to take into account isotopes transfer within the vegetation (root extraction and transpiration). The experimental design of RUBIC as well as the first modelling results will be presented.