基于年内尺度的中国大陆地区Angstrom公式参数校正与优选

Calibrating and Optimizing the Parameters in Angstrom Equation for Calculating Evapotranspiration from China's Mainland

联合国粮农组织(FAO)推荐Angstrom公式作为参考作物需水量计算过程中太阳总辐射(Rs)参数的建议算法,因此,该公式的as、bs参数取值对准确计算Rs至关重要。【目的】探讨Angstrom公式中2个经验系数as和bs在中国陆地范围内的空间分布特征,并给出各农业区的校正值。【方法】基于中国辐射月值数据集和中国地面气候资料月值数据集中共有的121个站点1957—2016年的逐月日平均太阳辐射和日照百分比有效观测记录,基于各站点1957—2010年的多年平均数据,利用最小二乘回归计算并统计得到中国38个农业区的年内统一as、bs系数,并以2011—2016年的Rs观测值为真值,比较了分别以as、bs系数回归值和FAO的建议值计算的Rs相对精度,筛选出了各农业区年内as、bs系数的优化取值。【结果】大部分站点的n/N与Rs/Ra的相关系数在0.8以上;以线性回归分析求得的Angstrom公式系数as在绝大部分区域维持在0.1~0.2之间,且空间分布稳定,而系数bs则波动性变幅较大,大部分区域在0.40~.8之间;以回归系数计算的Rs整体上要优于FAO推荐值计算的Rs,但在具体应用中则存在FAO推荐值计算的Rs优于回归系数计算的Rs的情况。【结论】在中国,基于n/N与Rs/Ra的相关关系,可以直接回归分析求取Angstrom公式的as、bs系数;各农业区回归求取的as、bs系数与FAO 56的建议值存在显著差别,且有各自的表现特征,但二者的相对大小也呈此消彼长的规律;现有数据和方法条件下,as、bs系数的回归值和建议值相结合推广使用是最佳方案。

【Objective】Evapotranspiration is an important process of hydrological cycle and the Food and Agriculture Organization(FAO)of United Nations has recommended the Angstrom Equation to calculate the total solar radiation(Rs)parameters required to calculate the reference crop evapotranspiration,in which the parameters as and bs are crucial for accurate estimation of Rs.We calibrated and optimized the two parameters in this paper for China's Mainland【.Method】The analysis was based on solar radiation data and percentage of sunshine data measured from 1957 to 2010 from 121 weather stations across the country.We calculated the values of as and bs for 38 agricultural production regions using least-square regression,based on daily mean solar radiation and percentage of sunshine.We then used the Rs measured from 2011 to 2016 as“true”date and compared it with those calculated using the estimated as and bs and recommended by FAO,respectively,from which we further calculated the optimal as and bs in each month for the 38 agricultural regions.【Result】The correlation coefficient between n/N and Rs/Ra was>80%for most stations.The coefficient as obtained from the linear regression model was in the range of 0.1~0.2 for most regions,and it was spatially insensitive.In contrast,the coefficient bs oscillated temporally and changed spatially in the range of 0.4~0.8 in most regions.In general,the Rs calculated using the parameters estimated from the regression model was superior to that calculated using the FAO recommended value,although exception existed.【Conclusion】The correlation between n/N and Rs/Ra can be used to estimate the parameters as and bs in the Angstrom formula directly from the regression model,and as and bs estimated from the regression model for each of the 38 agricultural areas differed considerably from that suggested by FAO.However,there are circumstances that the FAO value worked better.Therefore,with the available data and method,combining the regression model and the recommended as and bs by FAO appears to be the best solution to estimate the reference crop evapotranspiration.This paper has implication as it provides further case study on calibrating the parameters in the Angstrom Equation and can thus help improve the accuracy of Rs estimation in China.