Evapotranspiration Modeling over Irrigated Crops in South Central Arizona.

Models for estimating evapotranspiration (ET) of irrigated crops using thermal infrared remote sensing have been developed over the past several decades and are now considered mature and well-validated. Approaches such as the Two Source Energy Balance (TSEB) and METRIC have been implemented and evaluated for multiple environments, crops and irrigation methods. Results from these experiments show reasonable results with daily ET estimates commonly better than 1.5 mm/day. However there are still few studies using measured water deliveries within an irrigation district for entire growing seasons. Such measurements complement micrometeorological data sets and can help improve estimates of irrigation efficiencies and can assist ET model comparisons. Using Landsat 5 & 7 data collected in 2008 over the Central Arizona Irrigation District (CAIDD) we implemented three ET estimation approaches, two thermally driven models, TSEB and METRIC, and one NDVI/crop coefficient model. Time series ET estimates over cotton, wheat and alfalfa were made at weekly to bi-weekly time steps. Volumetric water delivery data provided at the farm-level were converted to apparent depths by repartitioning according to remotely sensed vegetation cover. Generally all three models appeared to closely agree with each other and to overestimate seasonal ET relative to water delivery data. This outcome suggests that simpler, non-thermal, vegetation index-based models may be adequate to meet needs for typical on-farm irrigation practice.