H33B-1576
Evapotranspiration Estimates in the Mountain West Utilizing Multi-platform Remote Sensing
Wednesday, 16 December 2015
Poster Hall (Moscone South)
Kyle Randall Knipper, Colorado School of Mines, Golden, CO, United States, Terri S Hogue, Colorado School of Mines, Civil and Environmental Engineering, Golden, CO, United States and Kristie Franz, Iowa State University, Ames, IA, United States
Abstract:
Understanding the linkages between energy and water cycles through evapotranspiration (ET) in the western U.S. is uniquely challenging given the climatic and ecological heterogeneity coupled with the added complexities of population growth, land use change and ongoing disturbance (beetle infestation, wildfire, etc.). A number of models have been developed to estimate ET based primarily on remote-sensing observations [MODIS global ET (MOD16) and Operational Simplified Surface Energy Balance (SSEBop)]. However, the scale of MOD16 and SSEBop (1km/monthly) may be too large to provide information on the spatial and temporal variability of ET that occurs over regions with acute or chronic land cover change and precipitation driven fluxes. The current study aims to improve the spatial and temporal resolution of ET utilizing only satellite-based observations by incorporating a previously developed potential evapotranspiration (PET) methodology with satellite-based soil moisture estimates. Initially, remotely-sensed soil moisture observations (AMSR2) are compared to ground-based estimates (AmeriFlux) and land-surface model (NLDAS) outputs at four sites within the Colorado River Basin. Soil moisture estimates are then downscaled to a final 250m resolution to estimate ET. Final values are compared to ground-based estimates, satellite-based estimates (SSEBop, MOD16), as well as land-surface model estimates (NLDAS). Results indicate poor AMSR2 performance, with slightly improved estimates reported by the downscaling approach (average RMSE of 0.076 m3/m3 and 0.065 m3/m3, respectively). However, the approach provides an improved spatial resolution, capable of deciphering small scale changes in soil moisture following a precipitation event. Subsequently calculated ET estimates display patterns representative of the basin’s elevation and vegetative characteristics, with improved spatial resolution and temporal heterogeneity when compared to MOD16 and SSEBop.
