Submitted to: American Geophysical Union
Publication Type: Abstract Only
Publication Acceptance Date: March 25, 2008
Publication Date: May 30, 2008
Repository URL: http://handle.nal.usda.gov/10113/36843
Citation: Choi, M., Kustas, W.P., Anderson, M.C., Li, F., Allen, R.G. 2008. An intercomparison of three remote sensing-based evapotranspiration (ET) schemes for a corn and soybean production region (Iowa, U.S.) during SMACEX [abstract]. EOS Transactions, American Geophysical Union, Joint Assembly Supplements. 89(23):H43A-02. Technical Abstract: Reliable estimation of evapotranspiration (ET) over a range of spatial and temporal scales is crucial for agricultural, meteorological, and hydrological applications. However, accurate estimation is hampered by the complexity in land-atmosphere interaction and heterogeneity in land surface states. In this study, three remote-sensing based ET models with Land Surface Temperature (LST) as a key boundary condition and having different algorithm complexity are evaluated using remote sensing imagery and ground-truth data from the 2002 Soil Moisture/ Atmosphere Coupling EXperiment (SMEX02/SMACEX) in Iowa, U. S. The SMACEX was designed to provide a variety of observations of vegetation, soil moisture, and atmospheric conditions under a wide range of temporal and spatial scales. The three models include: 1) the Trapezoid Interpolation Method (TIM), which uses available energy and LST-vegetation index space to adjust the Priestley-Taylor based potential ET algorithm for mapping actual ET; 2) the Mapping EvapoTranspiration with Internalized Calibration (METRIC), that uses energy balance constraints and hydrologic (wet and dry) extremes in the image to define LST-air temperature difference/heat flux relationship; and 3) the Two-Source Energy Balance model (TSEB) that has a more complex treatment of the energy exchange between soil-plant-atmosphere interface. Comparison of model output with flux tower observations and differences in the flux/ET maps generated using Landsat imagery will be described. Through such an inter-comparison, model differences associated with algorithms and model complexity, sensitivity to uncertainty in model inputs, and other factors will be presented and each model’s strengths and limitations will be described. Potential improvements in model algorithms and opportunities for incorporating the strengths of the different approaches in the development of a hybrid remote sensing-based ET model will be discussed.