Operational monitoring of daily evapotranspiration by the combination of MODIS NDVI and ground meteorological data: Application and evaluation in Central Italy

Time-varying crop coefficients (Kc) can be obtained from remotely sensed data and combined with daily potential evapotranspiration estimates for the operational prediction of actual evapotranspiration (ETA). This approach, however, presents relevant limitations when applied in mixed, water stressed ecosystems. The current paper addresses these issues by introducing two innovations. First, fractional vegetation cover (FVC) is derived from NDVI and utilized to split evaporating and transpiring surfaces, whose behavior is simulated under fully watered conditions by the use of generalized Kc. Next, the short term effect of water shortage is taken into account by means of downregulating factors which are based on meteorological observations (potential evapotranspiration and rainfall) and act differently for vegetated and not vegetated surfaces. The new method is tested against latent heat of evaporation (LE) measurements taken by the eddy covariance technique in six sites of Central Italy representative of various forest and herbaceous ecosystems. In this experiment the method is driven by 1-km meteorological data obtained from a pan-European archive and by 250 m MODIS NDVI imagery. Satisfactory accuracies are obtained in all experimental situations, which encourages the application of the method for the operational monitoring of ETA on regional scale. (C) 2014 Elsevier Inc. All rights reserved.