Simple modeling of the global variation in annual forest evapotranspiration

Annual forest evapotranspiration (ET) is highly variable among various sites. Zhang's model (Zhang et al., 2001) has been widely used for predicting the spatial variation in ET. The forest component of the model employs limiting theory and assumes constant annual potential evaporation (E ₀=1410mm) by regressing data recorded at 56 forest sites. However, most of the data used in determining E ₀ were recorded for limited regions (Australia, African countries, and the United States). We summarized 829 forest ET data items obtained for sites around the world from earlier publications. Using the dataset, we showed that Zhang's model overestimates forest ET in temperate and boreal regions with low annual mean temperature (T) owing to the E ₀ value. We revised the E ₀ term of Zhang's model so as to consider the dependency of E ₀ on T using the dataset. The revised model did not overestimate forest ET in temperate and boreal regions. Consequently, we recommend revising the E ₀ term of Zhang's model when predicting forest ET in these regions.