【摘要】 蒸散发(ET)涵盖了土壤、植物和大气间的水分交换，是水循环和能量平衡中的关键环节。冠层阻力(r_c)和气孔阻力(r_s~c)是影响ET准确性的重要阻力参数。本研究选用两种阻力模型Katerji-Perrier(KP)、Todorovic(TD)对P-M模型和S-W模型在中国西北地区夏玉米田ET模拟的适用性进行了评价。基于2015—2018年田间试验数据利用最小二乘法对阻力模型关键参数进行了校准和验证。结果表明，KP模型在P-M模型和S-W模型中表现优于TD模型，SW-KP模型ET模拟精度最高，决定系数R~2均值为0.816,均方根误差RMSE均值为0.79 mm/d;PM-KP模型在夏玉米苗期对ET低估仅为0.86%,SW-KP模型在拔节期、吐丝期和乳熟期表现最佳，R~2为0.718±0.071,RMSE为(0.85±0.11)mm/d;模型敏感性分析显示，KP模型r_c、r_s~c对净辐射(R_n)和饱和水汽压差(VPD)变化最为敏感，在TD模型中对空气温度(T_a)最为敏感，T_a变化±20%时r_c、r_s~c变化超过10%;PM-KP模型中ET对R_n最敏感，降低20%时ET减少19.69%,PM-TD模型中对T_a的敏感性最高，T_a变化±20%时ET变化约30%。在S-W耦合模型中，ET对环境变量的敏感性由大到小依次为R_n、T_a、土壤含水率(SWC)、VPD,R_n变化±20%时ET变化超过20%。通过最小二乘法校正的KP模型精度更高，PM-KP模型适用于苗期ET模拟，SW-KP模型在夏玉米拔节期、吐丝期和乳熟期模拟的ET精度更高。本研究可为玉米不同生育阶段ET模拟提供更优模型组合形式，为半干旱地区夏玉米田耗水量精量计算和灌溉制度制定提供理论依据。更多还原

【Abstract】 Evapotranspiration(ET), which covers the exchange of water between soil, plants and the atmosphere, is a key link in the water cycle and energy balance. Canopy resistance(r_c) and stomatal resistance(r_s~c) are important resistance parameters that affect the accuracy of ET.Two resistance models, Katerji-Perrier(KP) and Todorovic(TD) were selected to evaluate the applicability of P-M model and S-W model for ET simulation in summer maize fields in Northwest China. Based on the field experiment data from 2015 to 2018, the key parameters of the resistance model were calibrated and verified by least square method. The results showed that KP model performed better than TD model in P-M model and S-W model. The SW-KP model had the highest ET simulation accuracy, with mean value of coefficient of determination R~2 of 0.816 and mean value of root mean square error(RMSE) of 0.79 mm/d. The underestimation of ET by PM-KP model in summer maize seedling stage was only 0.86%, while the performance of SW-KP model was the best at the stage of ascending, spinning and milking, with R~2 ranging of 0.718±0.071 and RMSE ranging of(0.85±0.11) mm/d. Model sensitivity analysis showed that r_c and r_s~c in KP model were most sensitive to changes in net radiation(R_n) and saturated vapor pressure difference(VPD). In the TD model, the air temperature(T_a) was the most sensitive. When T_a changed ±20%, r_c and r_s~c changed more than 10%. In PM-KP model, ET was the most sensitive to R_n, and the sensitivity to T_a was 19.69% when T_a was reduced by 20%. In PM-TD model, ET was the most sensitive to T_a, and ET changed by about 30% when T_a changed by ±20%. In the S-W coupling model, the sensitivity of ET to environmental variables, ranked from high to low was R_n, T_a, soil moisture content(SWC),VPD,and the change of ET exceeded 20% when R_n changed ±20%. The KP model corrected by the least square method had higher accuracy, the PM-KP model was suitable for ET simulation at seedling stage, and the SW-KP model had higher accuracy for ET simulation at summer maize’s articulation stage, spinneret stage and milk ripening stage. The research result can provide a better model combination form for ET simulation at different growth stages of maize, and provide a theoretical basis for accurate calculation of water consumption and establishment of irrigation system in summer maize fields in semi-arid areas.更多还原