Abstract

The Simplified Surface Energy Balance (SSEB) model uses satellite imagery to estimate actual evapotranspiration (ET_a) at 1-kilometer resolution. SSEB ET_a is useful for estimating irrigation water use; however, resolution limitations restrict its use to regional scale applications. The U.S. Geological Survey investigated the downscaling potential of SSEB ET_a from 1 kilometer to 250 meters by correlating ET_a with the Normalized Difference Vegetation Index (NDVI) from the Moderate Resolution Imaging Spectroradiometer instrument (MODIS). Correlations were studied in three arid to semiarid irrigated landscapes of the Western United States (Escalante Valley near Enterprise, Utah; Palo Verde Valley near Blythe, California; and part of the Columbia Plateau near Quincy, Washington) during several periods from 2002 to 2008. Irrigation season ET_a-NDVI correlations were lower than expected, ranging from R² of 0.20 to 0.61 because of an eastward 2–3 kilometer shift in ET_a data. The shift is due to a similar shift identified in the land-surface temperature (LST) data from the MODIS Terra satellite, which is used in the SSEB model. Further study is needed to delineate the Terra LST shift, its effect on SSEB ET_a, and the relation between ET_a and NDVI.

First posted September 18, 2012