降雨和蒸散对夏玉米灌溉需水量模型估算的影响

Impact of rainfall and evapotranspiration on irrigation requirement of summer maize

为揭示降雨和蒸散年际波动对作物灌溉需水量的影响机理，以华北平原南部地区夏玉米为研究对象，基于土壤水分概率密度函数建立灌溉需水量计算模型，分析了降雨、蒸散和灌溉需水量的年际波动特征，结合蒙特卡洛方法探讨了降雨和蒸散对灌溉需水量年际波动的贡献。结果表明：研究区夏玉米降雨参数年际波动显著，平均降雨量波动于4.958～25.003 mm，变异系数为0.326；降雨频次波动于0.143～0.457 d-1，变异系数为0.170；平均降雨量和降雨频次均可采用Logistic分布描述：平均降雨量服从Logistic（11.273,2.022），降雨频次服从Logistic（0.318,0.029）。潜在蒸散量年际波动相对平稳，可采用对数正态分布描述：潜在蒸散量服从LogNormal（1.370,0.076）。灌溉需水量多年平均值为133.1 mm，波动于8.1～381.8 mm，变异系数为0.673，年际波动显著高于降雨和蒸散。平均降雨量对灌溉需水量年际波动的贡献率最大，其次为降雨频次，潜在蒸散量最小；在三者的共同作用下，灌溉需水量年际波动呈现出更大的不确定性。因此，在估算灌溉需水量时有必要对这种不确定性进行定量评估，针对不同降雨和蒸散条件制定合理的灌溉策略，为农业水资源管理和决策提供更可靠的科学依据。

Irrigation requirement is important for agricultural water management and irrigation decision making, and it is strongly influenced by interannual fluctuations of rainfall and evapotranspiration. The summer maize is one of the main crops in the China North plain, and can be greatly affected by annual variation of rainfall and potential evapotranspiration during its growing season. For that reason, supplementary irrigation estimated based on the rainfall and potential evapotranspiration is necessary. In this study, the summer maize was selected as a research object in the southern area of the China North Plain, and the analytic relation between irrigation requirement and rainfall parameters associated with potential evapotranspiration was described on the basis of probability density function of soil moisture in order to investigate the influence of rainfall and potential evapotranspiration on irrigation requirement. The annual variations and statistical characteristics of rainfall, potential evapotranspiration and irrigation requirement were analyzed firstly, and then the Monte Carlo method was employed to investigate the long-term influences of rainfall parameters and potential evapotranspiration on irrigation requirement. The long series of rainfall parameters and potential evapotranspiration were generated by a pseudo-random number generator according to their probability density functions. The standard deviation and variation coefficient of irrigation requirement were adopted to assess the contribution of rainfall parameters and potential evapotranspiration to irrigation requirement fluctuations. The results showed that the annual variability of rainfall parameters was remarkable. Rainfall mean depth fluctuated between 4.958 and 25.003 mm with a variation coefficient of 0.326, and rainfall frequency fluctuated between 0.143 and 0.457 d-1 with a variation coefficient of 0.170. Rainfall mean depthconformed to a Logistic distribution with the expected value of 11.273 and the standard deviation of 2.022, and rainfall frequency conformed to the same distribution but with the expected value of 0.318 and the standard deviation of 0.029. Both rainfall mean depth and rainfall frequency presented an unmarked downward trend by Mann-Kendall tend test. While the annual fluctuation of potential evapotranspiration, which subjected to a LogNormal distribution with the expected value of 1.370 and the standard deviation 0.076, was smooth and steady, and there was a significant downward trend for it. Irrigation requirement fluctuated between 8.1 and 381.8 mm around the average value （133.1 mm） with a variation coefficient of 0.673, and the variation was significantly larger than rainfall parameters and potential evapotranspiration. Irrigation requirement conformed to a Gamma distribution with the shape parameter of 2.173 and the scale parameter of 61.271. There was an unremarkable upward trend for irrigation requirement. When taking no account of the annual fluctuations of rainfall parameters and potential evapotranspiration, the calculated average value of irrigation requirement was 114.3 mm. It indicated that it would underestimate the annual average value of the irrigation requirement without consideration the annual fluctuations of rainfall parameters and potential evapotranspiration. When only rainfall mean depth was considered as random variable, the standard deviation and variation coefficient of irrigation requirement were 63.6 mm and 0.485. When only rainfall frequency was considered as random variable, the standard deviation and variation coefficient were 41.9 mm and 0.354. When only potential evapotranspiration was considered as random variable, the standard deviation and variation coefficient were 28.4 mm and 0.245. While all of them were considered as random variables, the standard deviation and variation coefficient were 88.5 mm and 0.651. Rainfall mean depth contributed to the largest share of irrigation requirement fluctuations, followed by rainfall frequency and potential evapotranspiration. Annual fluctuations of irrigation requirement presented a larger uncertainty under the mutual effects of rainfall parameters andpotential evapotranspiration. Consequently, it is necessary to assess the annual uncertainty of irrigation requirement, and to develop a reasonable strategy according to different rainfall and evapotranspiration conditions in order to reduce the negative impacts of climate fluctuations and guarantee food safety in China.