Observations and stochastic modeling of soil moisture control on evapotranspiration in a Californian oak savanna

The study of water exchange between soil, plants, and the atmosphere in response to seasonal or periodic droughts is critical to modeling the hydrologic cycle and biogeochemical processes in water‐controlled ecosystems. An essential step in such studies is to characterize changes in evaporation and transpiration under water stress. The objectives of this study are to investigate how soil moisture controls the evapotranspiration in a Californian oak savanna that experiences seasonal droughts, using multiyear field observations at the daily and stand scale, and to model these controls stochastically. The influence of soil moisture on evapotranspiration at the stand scale is studied using correlations between tower‐based evapotranspiration measurements and representative soil moisture obtained by aggregating point measurements. The observed pattern of this effect is found in agreement with an existing model that features a linear reduction of the evapotranspiration when soil moisture falls below a critical value. The model parameters are inferred using a Bayesian framework, and they are found to vary from year to year because of climate variability. The comparison between various aggregations of soil moisture at the stand scale from point measurements demonstrates that the spatial variability of the soil moisture and the water uptake capacity limited by the root biomass need be taken into account to produce a model that is most resistant to interannual variability. Finally, the parameterized model is used to predict the actual evapotranspiration with uncertainty estimates determined using the joint distribution of the parameters derived from the Bayesian framework. The satisfactory agreement between the predicted and measured evapotranspiration suggests that the calibrated model can be incorporated into water balance studies in the future.