Evapotranspiration Estimates over Non-Homogeneous Mediterranean Land Cover by a Calibrated "Critical Resistance" Approach

An approach based on the Penman-Monteith equation was used to estimate the actual evapotranspiration from local meteorological data over non-homogeneous land cover in a Mediterranean site in the south-east of Italy, with two six month data sets from two different years of measurements (2006 and 2009). The "critical resistance" formulation was used in different forms to model the surface resistance, together with some modifications to take into account the soil moisture content. One, two, or three model parameters were estimated, one of them related to the atmospheric resistance and the others to the surface resistance, and the calibration was made by either linear regression or nonlinear minimization of a proper cost function, depending on the applicability. Two kinds of cost functions were tested, the first depending on both the latent heat flux and the difference between screen air temperature and surface radiometric temperature, and the second depending on the temperature difference only. In all cases the calculated fluxes give better results with respect to both a flux-gradient approach and a complementarity based method, that require comparable data inputs. However the calibration by the temperature differences only, that requires no turbulent flux measurements, considerably increases the statistical uncertainty of the calibration parameters. The inclusion of the soil moisture did not significantly improve the model results in the considered site.