All-Sky Evapotranspiration and its Diurnal Asymmetry Using Physically Based Modelling and Geostationary Land Surface Temperature Data

Evapotranspiration (ET) is an Essential Climate Variable (ECV) that plays a key role in the energy-water cycle, as it can influence precipitation and temperature dynamics, and it also has a direct impact on irrigation water demand in agricultural areas. However, its accurate estimation is still under debate, with a major unresolved challenge: cloudy-sky conditions. This issue arises because most evaporation satellite-based models rely on instantaneous land surface temperature (LST) as an input to solve the energy balance. As a result, capturing ET dynamics under all-sky conditions remains challenging.

Moreover, these models rely solely on daily data and fail to capture the full dynamics of ET throughout the day. This partial representation of daytime ET dynamics is closely related to the asymmetric relationship between ET and radiation, which is strongly linked to LST. This asymmetry becomes even more complex in arid environments, where environmental factors such as vapour pressure deficit and air temperature modulate vegetation behaviour, as well as in irrigated areas, where water inputs can sometimes be uncertain. To interpret and represent ET dynamics under both clear-sky and cloudy-sky conditions, it is necessary to use models capable of simulating ET without relying on data availability affected by cloud presence.

This work presents preliminary results of the hybrid version of the FEST-EWB model, which is able to compute energy fluxes under all-sky conditions, merged with LST data from the Meteosat Second Generation. Evapotranspiration estimates over the entire MSG disk will be analysed and validated against eddy covariance data. The hybrid approach combines the FEST-EWB model—which continuously simulates soil moisture and ET over time and space, resolving LST and closing the energy–water balance equations (Corbari et al., 2011), thus providing a physically based framework capable of filling gaps when satellite LST is not available due to cloud cover—and the residual version of the FEST model, which uses the available LST under clear-sky conditions. Preliminary results from the FEST-Hybrid approach highlight the strong potential of the model due to its adaptability across different spatiotemporal scales and all-sky conditions. The integration of satellite LST data when available allowed to properly represent ET dynamic also in irrigated areas.