Inferring transpiration from evapotranspiration: A transpiration indicator using the Priestley-Taylor coefficient of wet environment

Abstract

Abstract Transpiration (T) from plant tissues is an important variable in ecological and agricultural studies. T is usually estimated using the Priestley-Taylor (PT) equation by an analog to the wet environment evaporation (ETw), because vegetation can use water from deep soils. However, the relation between the PT coefficients of T and ETw (αc and αe) was still unclear. In this paper, we tested whether αe can be used as a reasonable indicator of αc at intra-annual and annual scales. We obtained αe by inverting a complementary relationship model with measured evapotranspiration (ET) at the flux sites and obtained αc after ET was partitioned into evaporation (E) and T. We found that αc tends to converge to αe at annual scales, and a linear equation can adequately fit the αc = f(αe) functions at annual scales. With the αc = f(αe) functions determined, satisfactory accuracy (R2 = 0.89) was achieved in estimating T from ET following the “ET → αe → αc → T” calculating process at annual scales. In comparison, the αc ~ αe relationship at 8-day scales are modulated by leaf area index and vapor pressure deficit. However, similar αc = f(αe) functions are identified in the region with the same climate type at 8-day scales, indicating that a single αc = f(αe) function is applicable in the same region. The accuracy in T predictions at 8 day scales deteriorated compared to those at annual scales. However, R2 were still larger than 0.58 at most of the sites. Because great progress in ET estimation has been achieved with the help of thermal remote sensing technique in the last two decades, our results here can provide a new method to estimate T from remotely sensed ET.