Performance of the two-source energy budget (TSEB) model for the monitoring of evapotranspiration over irrigated annual crops in North Africa

The main objective of this study was to evaluate the performance and the domain of validity of the two source energy balance model (TSEB) for the monitoring of actual evapotranspiration (ETa) as a first step towards its use for irrigation planning. Secondary objectives were to analyze the ability of TSEB model to detect water stress and to evaluate evapotranspiration partition between evaporation (E) and transpiration (T) over irrigated annual crops. Within this context, TSEB was compared to the calibrated FAO-56 dual approach, taken as a reference tool for the monitoring of crop water consumption. TSEB computes ET,, as the residual of a double component energy balance driven by the radiative surface temperature (T-s) used as a proxy of crop hydric conditions; the FAO-56 dual crop coefficient approach uses the Normalized Difference Vegetation Index (NDVI) as a proxy of Basal Crop Coefficient (K-cb) and assesses the hydric status directly by solving a two layer soil water budget. Both approaches were evaluated over four plots of wheat and sugar beet located in the Haouz plain (Marrakech, Morocco) that were instrumented with eddy covariance systems during the 2012 and 2013 growing seasons. Series of ASTER images were acquired during the first agricultural season. Both models offered fair performances compared to ET observations with Root Mean Square Error (RMSE) lower than 1 mm day-1 apart from the FAO-56 dual approach on the sugar beet plot because of uncertain irrigation inputs. This highlights a major weakness of this model when water inputs are uncertain; a very likely case at the plot scale. By contrast, the TSEB model offered smoother performances in all cases. The potentialities of both approaches to predict a water stress index based on the departure from potential evapotranspiration (ETa) was evaluated: although the FAO-56 dual was better suited to detect high water stresses, the TSEB model was able to detect moderate stresses without a need to prescribe water inputs. Finally, the partition of ETa between soil evaporation and plant transpiration was estimated indirectly by confrontation between simulated soil evaporation and surface (0-5 cm) soil moisture acquired spatially with Theta Probe sensors and taken as a proxy of soil evaporation. TSEB evaporation was well correlated to surface soil moisture (r = 0.82) for low Leaf Area Index (LAI) values (<1.5 m(2) m(-2)). In addition, TSEB predicted partition compared well to snapshot measurements based on the stable isotope method, This in-depth comparison of two simple tools to monitor ET, leads us to the conclusion that the TSEB model can reasonably be used to map ETa on large scale and possibly for the decision-making process of irrigation scheduling. (C) 2017 Elsevier B.V. All rights reserved.