Spatial and temporal variability of evapotranspiration derived from MODIS and FLUXNET over US from 2001 to 2008

We propose a novel approach to map evapotranspiration. We couple information from MODIS with flux towers to assess the drivers and parameters of evapotranspiration. Specifically, we derive canopy conductance via gross primary productivity at the continental scale instead of climate drivers. Radiation components (incoming/outgoing short- and long-wave radiation, PAR, diffuse PAR) are modeled using an atmospheric radiative transfer model, FluxNet (neural network version of Streamer). All forcing data are prepared from a diverse atmosphere and land products of MODIS in support of Microsoft cloud computing service. Gross primary productivity is modeled using a two-leaf model (sunlit and shaded leaf) and it is used to calculate canopy conductance via Ball-Berry model. Then, we apply Penman-Monteith equation to calculate evapotranspiration. Finally we explore the spatial and temporal variability of evapotranspiration in US over 8 years.