Partitioning of evapotranspiration using a stable isotope technique in an arid and high temperature agricultural production system

Abstract

Agricultural production in the hot and arid low desert systems of southern California relies heavily on irrigation. A better understanding of how much and to what extent irrigated water is transpired by crops relative to being lost through evaporation would improve the management of increasingly limited water resources. In this study, we examined the partitioning of evapotranspiration (ET) over a field of forage sorghum (Sorghum bicolor), which was under evaluation as a potential biofuel feedstock, based on isotope measurements of three irrigation cycles at the vegetative stage. This study employed customized transparent chambers coupled with a laser-based isotope analyzer to continuously measure near-surface variations in the stable isotopic composition of evaporation (E, δE), transpiration (T, δT) and ET (δET) to partition the total water flux. Due to the extreme heat and aridity, δE and δT were very similar, which makes this system highly unusual. Contrary to an expectation that the isotopic signatures of T, E, and ET would become increasingly enriched as soils became drier, our results showed an interesting pattern that δE, δT, and δET increased initially as soil water was depleted following irrigation, but decreased with further soil drying in mid to late irrigation cycle. These changes are likely caused by root water transport from deeper to shallower soil layers. Results indicate that about 46% of the irrigated water delivered to the crop was used as transpiration, with 54% lost as direct evaporation. This implies that 28 − 39% of the total source water was used by the crop, considering the typical 60 − 85% efficiency of flood irrigation. The stable isotope technique provided an effective means of determining surface partitioning of irrigation water in this unusually harsh production environment. The results suggest the potential to further minimize unproductive water losses in these production systems.