Comparison of evapotranspiration measurements between eddy covariance and lysimeters in paddy fields under alternate wetting and drying irrigation

Evapotranspiration from paddy fields under alternate wetting and drying (AWD) irrigation was measured using lysimeters (ETL) and eddy covariance (ETec) during the late 2016 and early 2017 rice seasons. Four energy balance closing methods were explored, and discrepancy of ETL and ETec (with or without adjustment) in different growth stages and wet/dry conditions were examined. The results indicated that both energy balance non-closure and wet/dry conditions of the field affect the discrepancy of ETL and ETec. Energy balance was suggested to be closed each half hour by 'residual-LE closure' method only at daytime, and nighttime flux stayed unchanged, while a careful data control and interpolation should be noted for nighttime flux. In the meantime, wet-to-dry patterns of AWD irrigation can affect the discrepancy between ETec and ETL. In wet periods, ETec with suitable adjustment method was well matched with the ETL. However, ETL was underestimated in dry periods due to lower water content in the lysimeter. Finally, net radiation (Rn), vapor pressure deficit (VPD), air temperature (Ta), soil and water temperature (Ts and Tw), and wind speed (u) were selected to conduct path analysis on ETL and ETec (with adjustment) for both wet and dry conditions. The direct dependencies of each meteorological factor on ETL and ETec were similar under wet conditions; however, under dry conditions, VPD and Rn are the respective main controls for ETL and ETec. Moreover, ETL is more sensitive to Ta and Tw under wet conditions.