Inter-seasonal and cross-treatment variability in single-crop coefficients for rice evapotranspiration estimation and their validation under drying-wetting cycle conditions

A two-year experiment was conducted to investigate the inter-seasonal and cross-treatment variability in measured rice evapotranspiration (ETcMea), measured single-crop coefficients (KcMea), and treatment-specific calibrated coefficients (KcCal), under different drying-wetting cycles in a subtropical monsoon climate in East China. For each drying-wetting treatment, ETcMea was determined based on data collected in lysimeters, and KcMea was calculated from ETcMea, reference evapotranspiration, and soil moisture deficit coefficient. Following the single-crop coefficient method, KcCal was determined by matching KcMea. In 2012 and 2013, ETcMea varied from 459.5 to 486.7 mm, and 544.5-605.1 mm, respectively. Its inter-seasonal variability was larger than the cross-treatment variability. Stage-wise average KcMea were 1.07-1.17, 1.30-1.51, 1.49-1.54, and 1.17-1.29 in 2012, and 1.06-1.12, 1.31-1.49, 1.43-1.57, and 1.26-1.27 in 2013 during the initial, crop development, mid-season, and late season stages, respectively. Treatment-specific KcCal were calibrated as 1.09-1.20, 1.51-1.60, and 0.74-0.78 in 2012, and 1.05-1.14, 1.47-1.64, and 0.96-1.01 in 2013 for the initial, mid-season, and end-season stages, respectively. The inter-seasonal and cross-treatment variability in KcMea and KcCal was low. Each treatment-specific KcCal set performed similarly when rice ETc was calculated under different drying-wetting treatments. Cross validation indicated that large daily uncertainty in ETc estimation occurred when daily ETcMea was high, and uncertainty in seasonal ETc calculated using different treatment-specific KcCal sets ranged from 45.7 to 60.1 mm (approximate to one irrigation). Calibrating KcCal using more data (season-specific KcCal or mixed treatment KcCal) would improve the accuracy of KcCal in ETc estimation.