UC Davis

Evapotranspiration (ET) is a major hydrologic flux in water resources planning and irrigation management. While recent advances in remote sensing (RS) have enabled availability of high spatial and temporal resolution ET data, a lack of information related to error in the estimations has made it challenging to use this data for on-farm irrigation management decision making. In this study, three commonly used single-source RS based ET models (pySEBAL — a new version of the Surface Energy Balance Algorithm for Land; SEBS — Surface Energy Balance System algorithm; and METRIC — Mapping Evapotranspiration at High Resolution with Internalized Calibration) were used to estimate daily actual evapotranspiration (ETa) for almonds, processing tomatoes, and maize in the Central Valley of California. Model evaluation was conducted by comparing the predicted ETa from RS with in-situ measured ETa using surface renewal. Results indicated that the RS-based ETa estimations for all three models were within acceptable levels of uncertainty and agreed well with surface renewal estimates except for the underestimation by pySEBAL and METRIC during early season growth stages of processing tomatoes. This underestimation was attributed to the lack of accuracy when using single source ET models under lower vegetation cover condition (when ET is dominated by soil evaporation). Better estimates of ETa with pySEBAL and METRIC were detected at full cover, which explains the applicability of these two models to irrigation management during peak crop water demand. SEBS performed the best among the three RS-based models for daily ETa estimation for all crops. This suggests that SEBS-based ETa estimates can be adopted in operational irrigation management programs for farms that have not installed in field ETa sensors such as Tule Sensors (Tule Technologies Inc.). In addition, RS based ET is spatially distributed which can help to identity spatial variability between different irrigation zones.