Global Land Surface Evaporation from Satellite-based Observations

Estimating the magnitude of the different fluxes in the hydrological cycle is essential if we are to predict the impacts of climate change. However, climate change is acting on a dynamic three dimensional globe where change in one region may produce impacts in another. This creates a need to expand the current climate change studies to encompass the entire globe. Evaporation is a key component of the hydrological cycle as it can affect both feedbacks on large scale water processes (e.g.[1]) and the dynamics of the atmosphere due to changes in the Bowen ratio (e.g.[2]). The uncertainty in predictions of future climate must be reduced if we want to effectively manage adaptation to climate change. Consequently, there is a need to create observation-based benchmarks, against which GCM performance can be judged (e.g.[3]). Hydro-meteorological benchmarks will allow selection between GCMs and lead to improvement in model simulations of the hydrological cycle. The first priority is to combine the available remote sensing data products with hydrological models and land surface parameterization schemes to create such an evaporation benchmark. In the last two decades several attempts have been made to build global evaporation products using remote sensing information. In 1997, Choudhury pioneered a potential evaporation product based on data assimilation of satellite observations ([4]). His approach applied the Penman-Monteith equation and derived monthly evaporation maps for the years 1987 and 1988. Years later [5] also used the Penman-Monteith equation to derive global evaporation, but applying more advanced remote sensing products from the MODIS sensor. Several model and satellite-based evaporation products are now available, but the majority of these models lack the key elements of estimating forest rainfall interception loss and the coupling of transpiration with observed soil moisture. Global Land surface Evaporation: the Amsterdam Model (GLEAM), a unique methodology to derive evaporation from remotely sensed observations, has been developed at the VU University of Amsterdam to fill these gaps. GLEAM uses an extensive range of satellite observations as a basis for estimating daily actual evaporation at a global scale and 0.25 degree spatial resolution. Central to the approach is the Priestley and Taylor evaporation formula ([6]). This radiation-based model appears appropriate for the level of available driving data at the spatial scale of the study. The use of a variety of independent remote sense observations is only one of the main advantages of GLEAM over other evaporation models. Coupling of transpiration to soil moisture conditions and detailed, separate rainfall interception loss estimation are two key features that allow the approach to be applied in land-atmosphere feedback studies and tests of GCM performance. The proposed work describes the modeling framework and tests the applicability of the methodology in combination with the available remote sensing driving data. The final global evaporation flux has been validated, applying a selection of stations from the Fluxnet global network of micrometeorological flux measurements. Results corresponding to a wide range of years will be presented, showing the global spatial distribution of the model-estimated actual evaporation and analyzing the factors controling this distribution.