黄土高原丘陵沟壑区多种参考作物需水量模型对比分析

Comparative Study on Calculation Methods for Reference Crop Evapotranspiration in Hill and Gully Area of the Loess Plateau

利用米脂地区2000—2005年逐日气象资料,以FAO56-PM方程计算结果为标准,通过线性回归、均方根误差(RMSE)和平均偏差,在日、月及年时间尺度上分析、评价10种ET0计算模型。结果表明,除Hargreaves和Turc模型外,其他8种模型计算的日ET0值与FAO56-PM的R2均大于0.9。不同模型与FAO56-PM日ET0值的RMSE为0.35～1.78。多数计算模型在冬季(11、12、1月)和春季(2、3、4月)的准确性高,而在5—10月表现较弱。通过年值分析发现Priestley-Taylor、FAO24-BC和Penman这3种模型与FAO56-PM计算值最为接近。而对于FAO24-Rad、Makkink、Hargreave与Turc等参数较少的公式,在应用时利用当地气象要素进行经验系数校正。

Ten kinds of ET0 methods were compared based on the daily meteorological data during the years of 2000 to 2005 from Mizhi stations in hill and gully area of the Loess Plateau.The Penman-Monteith equation standardized by the Food and Agriculture Organization（FAO56-PM） was used to compare with other methods at daily,monthly and yearly scale.The results of linear regression,root mean squared error（RMSE） and mean bias error（MBE） showed that the correlation coefficients between Hargreaves,Turc and FAO56-PM were lower than 0.9,and the other methods were higher than 0.9 at daily scale.The RMSE between FAO56-PM and other methods were 0.35～1.78.Most methods had the high accuracy during winter（November,December and January） and spring（February,March and April）,however,the low accuracy during May to October.The annual ET0 estimated by Priestley-Taylor,FAO24-BC and Penman model were closed to FAO56-PM.The FAO24-Rad,Makkink,Hargreave and Turc equations with fewer parameters should be regulated by the meteorological elements in given local stations.