/AnMtgsAbsts2009.55532 Mapping Evapotranspiration and Moisture Stress in An Advective Environment Using Multi-Scale Thermal Remote Sensing Data.

Tuesday, November 3, 2009: 10:45 AM
Convention Center, Room 326, Third Floor

Martha Anderson1, William Kustas1, Christopher M.U. Neale2, John Prueger3, Derek Williamson4, Steven Evett5, Jose Chavez6 and Prasanna Gowda7, (1)Hydrology and Remote Sensing Laboratory, USDA-ARS, Beltsville, MD
(2)Utah State Univ., Logan, UT
(3)National Soil Tilth Laboratory, USDA-ARS, Ames, IA
(4)Civil, Construction, and Environmental Engineering, Univ. of Alabama, Tuscaloosa, AL
(5)Conservation and Production Research Laboratory, USDA-ARS, Bushland, TX
(6)Irrigation Engineering Department, Colorado State Univ., Fort Collins, CO
(7)USDA-ARS, Bushland, TX
Robust remote-sensing methods for mapping evapotranspiration (ET) and vegetation stress are needed to develop sustainable strategies for regional water management.  Such methodologies are particularly critical in areas of rapidly depleting groundwater reserves, such as in the Ogallala Aquifer system in the U.S. High Plains where extractions for irrigation and other uses significantly exceed recharge rates. Thermal infrared (TIR) remote sensing imagery, available at multiple spatial and temporal resolutions, provides valuable diagnostic information about the surface energy balance and sub-surface moisture status at field to regional scales.  However, application of TIR-based energy balance algorithms over irrigated landscapes in arid and semi-arid climates can be challenging, as local advection can significantly enhance ET, often in excess of the available energy.   In this study, a multi-scale two-source TIR model of surface energy balance is evaluated over the BEAREX08 experimental field site in Bushland, TX during various stages of crop development.  Regional fluxes were modeled with the Atmosphere-Land Exchange Inverse (ALEXI) model using 10-km resolution TIR imagery from the Geostationary Operational Environmental Satellites (GOES), while an associated flux disaggregation algorithm (DisALEXI) used Landsat TIR data to generate flux distributions at 120-m resolution.  Flux estimates from ALEXI/DisALEXI were compared with eddy covariance and weighing lysimeter observations collected during BEAREX08, as well as with spatially averaged regional flux estimates acquired by flux aircraft.   Model performance was also evaluated using 120-m Landsat TIR imagery sharpened to 30-m resolution using Landsat visible/near infrared vegetation index data – scales better resolving the small experimental fields at the BEAREX08 site.   These model-derived flux maps provide a spatial basis for intercomparing aircraft, tower and lysimeter data collected during BEAREX08.