Calibration of an evapotranspiration algorithm in a semiarid sagebrush steppe using a 3-ha lysimeter and Landsat normalized difference vegetation index data

In arid and semiarid environments, evapotranspiration (ET) is the primary discharge component in the water balance, with potential ET exceeding precipitation. For this reason, reliable estimates of ET are needed to construct accurate water budgets in these environments. Remote sensing affords the ability to provide fast, accurate, field-scale ET estimates, but these methods have largely been restricted to deep rooted (phreatophytic) plant communities underlain by shallow groundwater. We used 13 years of data from a 3-ha drainage lysimeter in a semiarid sagebrush steppe and Landsat normalized difference vegetation index (NDVI) data to calibrate a generalized least squares model capable of predicting vadose zone ET in a high elevation upland ecosystem. Annual precipitation was the best predictor of annual ET, as they were nearly balanced every year analysed (mean difference = 3 mm). We incorporated reference crop ET and a linear combination of NDVI and precipitation to capably predict ET on a subannual, lag-determined interval of 48 days, with a mean error of only 9.92% across all observations. To our knowledge, this is the first vegetation index-ET algorithm calibrated in a semiarid upland plant community using field-scale lysimetry. Vadose zone ET is particularly important at waste disposal sites in the Desert Southwest, where accurate and spatially explicit ET estimates are needed for monitoring potential mobilization and transport of contaminants past the root zone into local aquifers and for monitoring and modelling effects of recharge on flow and transport of contaminants in underlying aquifers.