Résumé
Characterization of plant water use, generally determined from evapotranspiration, together with water stress, derived from remote sensing data in the thermal infrared domain, are needed to better manage water resources. Evapotranspiration and water stress can be simulated by a dual source energy balance model that combines satellite and in situ hydrometeorological information. Available hydrometeorological observations are often insufficient to account for the spatial and temporal variability of the area of interest. To address this issue, we developed a stochastic weather generator that relies on ERA5 reanalyses and provides spatio-temporal scenarios of multiple hydrometeorological variables. The generator is evaluated and compared with two bias correction methods in terms of their ability to reproduce both observed hydrometeorological variables and simulated evapotranspiration and water stress in central Tunisia. Our analyses show that the stochastic generator offers interesting advantages to perform gap-filling and to extend the hydrometeorological time series in the past.