Satellite-Based Estimates of Evapotranspiration Compared with Estimates from Land Surface Models and Flux Towers: Drivers of Discrepancies and their Spatial and Temporal Patterns
Abstract
Monthly-average estimates of latent heat flux and evapotranspiration have been derived from a combination of satellite-derived microwave emissivities, day-night differences in land surface temperature (from microwave AMSR-E), downward solar and infrared fluxes from ISCCP cloud analysis, and MODIS visible and near-infrared surface reflectances. The estimates are produced by a neural network. These estimates have been compared with data from the NOAH land surface model, as produced for GLDAS-2, and we have analyzed areas with persistent, substantial discrepancies between the satellite and model products. Some of these discrepancies persist year-to-year, while others relate to different responses of the satellite/NN product and the model to anomalous conditions. Data from flux towers have been used to identify model and data deficiencies responsible for these discrepancies. The satellite-derived products are compared also with latent heat flux estimates derived by Jung et al., who used model tree ensembles to upscale Fluxnet tower data to the global scale.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2013
- Bibcode:
- 2013AGUFM.H21P..04L
- Keywords:
-
- 1818 HYDROLOGY Evapotranspiration