Estimating actual evapotranspiration in a semiarid region by simple remote sensing techniques

The use of remote sensing techniques to estimate regional actual evapotranspiration has become more and more widespread during the last decade. A large number of models – of varying complexity – exists today, and even some near operational products are available online for download by users (The LandSaf Evapotranspiration product based on Meteosat Second Generation SEVIRI (Ghilain et al. 2011) and the MODIS ET product, Mu et al., 2007). Typically, models incorporate information from different wavelengths of the electromagnetic spectrum together with some ancillary data depending on the complexity of the model. For operational purposes however, it is not always feasible to apply models that highly demand for other input/ancillary data because of their scarcity and their formats not readily compatible for modeling regional evapotranspiration. We therefore have to look for more simple models that meet the requirements. In a first step, it is necessary to test the validity of the performance of the models with more complex physically based models and observations, followed by an investigation of how simplifications affect the performance of the models.

In this study, we compare simple approaches to modeling evapotranspiration, the SEBAL model (Bastiaanssen et al., 2005) and the simplified triangle approach combined with Priestley Taylor (Stisen et al, 2008), for the semiarid West Africa. Both models are based on a surface energy balance approach, where land surface temperature (LST) is a key variable. In order to assess the effect of temporal and spatial resolutions, the models are run with two sets of input data with different spatio-temporal characteristics. Data from the MODIS sensor has a moderate spatial resolution of (500m-1 km) compared to data from the SEVIRI sensor on board the geostationary Meteosat Second Generation satellite with a spatial resolution of 4km. The overall performances of the models are evaluated on the basis of their spatial/temporal performance for different – cloud free – climatic conditions during 2006 and 2008. Model output validation is done for two savanna landscapes in Senegal (the Dahra test site) and Mali (the Agoufou test site) respectively, where flux tower data are available.