基于温度信息和历史阈值的蒸散发预报及修正

蒸散发（ET_c）是陆面和大气之间物质与能量交换的关键要素，其精准预报对于制定合理的灌溉计划和提高水分利用效率至关重要。本文以华北平原冬小麦和夏玉米为研究对象，在作物充分供水条件下，重点考虑了温度对作物生长的影响，发展了基于温度效应的作物系数（K_c）和参考作物蒸散发（ET_o）估算模型，构建了ET_c预报模型（简称：M_T）；考虑极端天气情况的不确定性影响，通过构建历史阈值修正函数进一步发展了M_T模型，形成了M_T-threshold模型，并以涡度相关法实测值为依据，验证了M_T-threshold模型在1~15 d预见期的适用性。结果表明：基于温度效应的K_c预报值与实测值拟合效果较好，1~15 d预见期内具有较好的适用性；率定后Hargreaves-Samani模型的ET_o估算效果较好，1~15 d预见期的预报准确率均在72%以上，且冬小麦生长季的预报效果优于夏玉米生长季；M_T模型在1~7 d预见期具有较好的适用性（预报准确率>83%）；M_T-threshold模型在1~15 d预见期的预报效果有明显提升，1~7 d预见期的预报精度提高2%以上，预报精度均在86%以上；8~15 d预见期预报准确率提高8%以上，预报精度均在80%以上。因此，基于温度预报信息及作物生长历史阈值能够较好的实现ET_c短中期预报。

Evapotranspiration forecast and modification based on temperature and historical threshold

Evapotranspiration (ET_c) is a key element of material and energy exchange between land surface and atmosphere.Its accurate prediction is very important for making reasonable irrigation plan and improving water use efficiency.In this study, winter wheat and summer maize in North China Plain were studied.Under the condition of adequate water supply for crops, considering the influence of temperature on crop growth, the estimation models of crop coefficient (K_c) and reference crop evapotranspiration (ET_o) based on temperature effect were developed, and the evapotranspiration prediction model was constructed (M_T).Considering the uncertain influence of extreme weather conditions, the M_T model was further developed by constructing the historical threshold correction function, and the M_T-threshold model was formed.Based on the measured value of eddy covariance method, the applicability of M_T-threshold model in 1-15 d prediction period was verified.The results show that the forecast value of K_c based on temperature effect fits well with the measured value, and has good applicability in the prediction period of 1-15 d.After calibration, the ET_o estimation effect of Hargreaves-Samani model is good, the prediction accuracy of 1-15 d prediction period are more than 72%, and the prediction effect of winter wheat growing season is better than that of summer corn growing season.M_T model has good applicability in 1-7 d prediction period (prediction accuracy>83%).The prediction effect of M_T-threshold model has been significantly improved in the prediction period of 1-15 d, the prediction accuracy of 1-7 d prediction period has been improved by more than 2%, and the prediction accuracy are more than 86%, the prediction accuracy in 8-15 d prediction period increases by more than 8%, and the prediction accuracy are more than 80%.Therefore, based on temperature forecast information and crop growth history threshold, the short and medium-term forecast of ET_c can be better realized.