Attribution analysis of actual and potential evapotranspiration changes based on the complementary relationship theory in the Huai River basin of eastern China

Accurate terrestrial evapotranspiration (ET) estimation and understanding the causes of ET variation are essential for water resource management and irrigation planning. In this research, the complementary relationship (CR) between actual evapotranspiration (ETa) and potential evapotranspiration (ETp) was evaluated. An Advection-Aridity (AA) model was calibrated and validated using eddy covariance measurements. The spatiotemporal variations of ETa, ETp and associated meteorological variables were examined through Mann-Kendall (MK) testing and Theil-Sen's estimator, using daily weather data from 137 meteorological stations in Huai River basin (HRB) during 1961-2014. Moreover, the influences of meteorological factors on ETa and ETp were quantified by the differential equation method. The results indicate that CR theory is applicable in the HRB and the calibrated AA model can simulate the ETa well. ETa exhibited a significant increasing trend before 1990 and then decreased significantly. However, ETp decreased significantly before 1990 and then declined slightly. During 1961-1990, except for significant increasing relative humidity (RH), other meteorological variables exhibited decreasing trends. The aerodynamic component dominated ETa and ETp trends in general. The wind speed at 2-m height (u(2)) dominated ETa trends except for summer and growing season and ETp trends except for summer, when the dominant factor is net radiation (R-n). During 1991-2014, mean temperature (T-a) and RH showed distinct increasing and decreasing trends, respectively, whereas significant decline in u(2) palpably slowed. The absolute value of the radiative component was larger than that of the aerodynamic one. The dominant factor of ETa trends shifted from u(2) to RH in spring and to R-n in autumn, winter and annual timescales. Moreover, the dominant factor of ETp trends changed from u(2) to RH in spring, winter and annual timescales and to R-n in growing season and autumn. R-n always played a pivotal role in both ETa and ETp trends in summer.