Inferior Performance of the Generalized Complementary Relationship for Evaporation in the Mediterranean Climates

The complementary principle was widely used in evaporation (E) estimation. However, it has been found that the complementary principle has biases in estimating evaporation in the Mediterranean climates, which affects its application. In this study, a deep analysis by using the original sigmoid generalized complementary equation and its modified version by introducing soil moisture content were applied on 24 eddy flux sites located in the Mediterranean climates. We found that (a) the performance of the generalized complementary relationship is inferior at most of the sites in the Mediterranean climates, which results in low accuracy on E estimation with the mean absolute error (MAE) of 12.06 ± 5.62 W m⁻²; (b) rescaling the complementary function with surface moisture information can substantially reduce the simulation biases with MAE decreases by 26%; and (c) the modification works well in most of the ecosystem types with the MAE decreases from 24% to 48% except in the forests. The results indicate that when precipitation and heat conditions are not synchronized (i.e., in the Mediterranean climate), the original complementarity principle cannot provide a reasonable explanation for the real world. It is necessary to introduce certain surface information into the complementary framework reasonably to improve the accuracy of evaporation simulation.