Quality control and flux gap filling strategy for Bowen ratio method: revisiting the Priestley–Taylor evaporation model

Micrometeorological measurements were made over an irrigated rice paddy during the Huaihe River Basin Experiment (HUBEX) in 1999. This study addresses the quality control and gap filling strategy for the heat fluxes with the Bowen ratio energy balance (BREB) method. It also endeavors to benefit future studies through comparing five methodologies to estimate the net water exchange. First, a three-step quality control strategy is constructed. Its first two steps guarantee the correct flux directions and reject suspicious data, respectively. The third step forbids supersaturation by considering the Bowen ratio ranges for different flux combinations (termed as “scenarios”). The quality-controlled latent heat (LE) and sensible heat (H) fluxes fall in three scenarios, namely I (LE > 0, H > 0), II (LE > 0, H < 0), and IV (LE < 0, H < 0). Second, the Priestley–Taylor evaporation model (PTEM) is applied to fill the gaps of LE, while the energy balance relationship is used to gap-fill H, namely H = R _N − G − LE (R _N : net radiation; G: soil heat flux). Central to the success of this strategy is the idea to calibrate the Priestley–Taylor parameter (α) in a scenario-specific manner. On average, α values are calibrated as 1.20 and 1.35 for scenario I and II, respectively. For scenario IV, most α values lie in a narrow range, namely from 0.9 to 1.0. Then, α is calibrated as 0.97 to extend the applicability of the PTEM to condensation (negative LE). The scenario-specific treatment explicitly explains the diurnal variation of α derived without distinguishing the scenarios. Third, five methodologies are compared in the calculation of net water exchange, including PTEM-gap-filled BREB method (M1), energy-balance-based eddy covariance method (M2), and three simplified models to estimate LE, respectively as R _N − H, R _N − G, and R _N . A major finding is that G and H have a similar effect of about 3% in the net water exchange. Thus, if either is neglected, the net evaporation should be slightly lowered to avoid overestimate, by 3% as a rule of thumb from this work.