How good perform complementary evapotranspiration models? – A comparison of modeled versus measured evapotranspiration derived from eddy covariance estimates of latent heat flux

Actual evapotranspiration (AET) is a major component in the water balance of hydrological systems and accurate assessments are needed. Two complementary relationships have been used in this paper to estimate AET for a grassland experimental site, the antecedent precipitation index (API) model and the advection aridity (AA) model. Results of both models were subsequently compared with daily AET obtained by the eddy covariance method, assumed as observed data. The comparison of observed and modeled timeseries comprises a two year period, from July 2013 to July 2015. The accuracy of the complementary relationships has been evaluated with quantitative, statistical and fitting analysis. Both empirical models predict an overestimation of observed data and models errors exibith a typical seasonal pattern. Larger errors are in particular detected during the winter and autumn seasons. The results of the analysis also suggest that the API method is more suitable for practical applications because of its higher accuracy compared to the AA model.