Comparison of evaporation estimation methods for water surface under floating coverage in arid areas

The main purpose of this study is to evaluate evaporation estimation methods for water surfaces under floating coverage in arid areas. The natural water surface evaporation methods (namely, the mass transfer method (MT), Bowen ratio energy balance (BREB) method, Penman method, the Penman-Monteith method (PM), and the Priestley-Taylor method (PT)) are corrected to establish the covered water surface evaporation methods, which are based on the micrometeorology near the covered water surface, energy balance components, and Bowen ratio. The radiation term of the natural water surface evaporation model is corrected by introducing the Bowen ratio of the entire surface, which includes the uncovered water surface and the floating cover surface. The mass transfer coefficient and aerodynamic impedance in the aerodynamic terms of the natural water surface evaporation model were corrected. Taking the evaporation of the water surface covered by spheres (coverage rate of spheres is 76.4%) as an example, based on the rose change characteristics of the above parameters, the effects of these modified evaporation models are comprehensively evaluated from three different time scales (even months, odd months and the complete test period). The evaluation methods include the root mean square error (RMSE), index of agreement (D), mean bias error (MBE), and coefficient of determination (R2). The results show that under three different time scales, the Penman-Monteith model is the most promising and superior in estimating water surface evaporation under floating coverage (for the complete test period, RMSE=0.11 mm·d−1, D=0.99, MBE= 0.07 mm·d−1, and R2 =0.99); the Priestley-Taylor model is slightly less effective in estimating evaporation than the Penman-Monteith model, and the mass transfer model and the Penman model are less effective in estimating evaporation. Furthermore, the analysis of the Penman-Monteith model shows that the evaporation of the radiation term (69.5%) is higher than that of the aerodynamic term (30.5%) during the evaporation of the water surface under floating coverage, while the evaporation of the radiation term is dominant. Moreover, during the complete test period, the weights of the radiation evaporation and aerodynamic evaporation changed significantly with time. The proportion of radiation evaporation varies from 54.8% to 81.7%, while that of aerodynamic evaporation varies from 18.3% to 45.2%.