ObjectiveEstimating evapotranspiration (ET) based on the semi-empirical and semi-theoretical Priestly-Taylor model (PT) mainly relies on accurately determining the appropriate value of the model coefficient α in a specific study area. This study expored the applicability of the model coefficient α. Localization studies have been conducted to more accurately estimate evapotranspiration in arid and semi-arid regions.

MethodThis paper used the eddy covariance technique combined with micro-meteorological data to monitor the water and heat transfer processes of a typical Artemisia ordosica shrubland in growing season in Mu Us Sandy Land of northwestern China to quantify the seasonal variations in α coefficients.

ResultThe results showed that the actual α coefficient had significant seasonal variations, peaking during the leafing period, and remaining relatively plateau during expanded leaf period and the leaf discoloration period. The maximum α coefficient was 0.66 and the minimum value was 0.03, averaging 0.23 in growing season. There was a logarithmic positive correlation between α coefficient and canopy conductance and saturated vapor pressure difference in the growing season. Soil water content (30 cm belowground) and leaf area index were positively correlated with α coefficient. Evapotranspiration estimated from the PT model with α=1.26 and with α=0.50, which was calculated from the wind speed data at 2 m height, was significantly larger than the measured evapotranspiration. The recommended localization value of the improved PT model coefficient α was 0.23, which gave better estimation of ET, linear gradient was 0.72, R² was 0.57.

ConclusionTherefore, the modified PT model can be used to accurately estimate evapotranspiration in arid and semi-arid regions.