Abstract:
Evapotranspiration (ET) is a key element in the land water and energy cycles and as such plays an important role in the climate system. Accurate estimates of ET are thus critical for climate and hydrological studies. Several global multi-year ET datasets are available from remote sensing measurements, diagnostic approaches or reanalysis data. A major constraint of the various ET datasets is the difficulty of their validation, since a direct comparison to in-situ measurements is only possible on a point scale. An intercomparison of differently derived ET datasets can partly stand in for a validation with in-situ measurements. In the framework of the LandFlux-EVAL project (www.iac.ethz.ch/url/research/LandFlux-EVAL), we compare observation based ET datasets with land-surface model output, reanalyses and IPCC AR4 simulations. The intercomparison aims at providing insight in the spatial differences and uncertainties of ET data products. We therefore present global uncertainty estimations (e.g. triple collocation errors), analyses of the relationship between the datasets as well as a comparison of the datasets in selected river basins. The presentation will further focus on the validation of ET in IPCC AR4 climate models.