Developing surrogate pixels for comparing SEBAL ET with lysimeter ET measurements

SEBAL (Surface Energy Balance Algorithm for Land) is an image-processing model comprised of twenty-five submodels for calculating evapotranspiration (ET) as a residual of the surface energy balance. SEBAL was developed in the Netherlands by Bastiaanssen and has been applied in many developing countries. SEBAL has now been applied with Landsat Thematic Mapper (TM) images in southern Idaho to predict monthly and seasonal ET for water management. The validation of SEBAL on the Snake River Plain of Idaho has centered on the use of two precision weighing lysimeter systems for ET measurement that were in place at Kimberly, Idaho. Because of the relatively small dimensions of the lysimeter fields at Kimberly it was not possible to find a band 6 pixel of the Landsat 5 TM completely inside a lysimeter field. This caused "blurring" or contamination of the band 6 pixel from surface temperatures of surrounding fields that were different from those for the lysimeter field. A procedure was developed to map surrogate pixels that had the same spectral characteristics as the lysimeter fields and use these as surrogates to compare SEBAL ET with lysimeter ET. Results indicate a wide distribution of ET computed by SEBAL for groups of fields that appear spectrally similar to the lysimeter fields. Much of the cause of the variation is probably due to differences in the thermal properties (surface temperatures) of individual fields. This illustrates the utility of applying the energy balance process to predict ET that incorporates surface temperature along with other aerodynamic characteristics of the surface.