基于温度的参考作物蒸散量计算方法的适用性评价

参考作物蒸散量是表征气候干湿程度、植被耗水量、生产潜力及水资源供需平衡的重要指标之一。以海口和敦煌两个气候相差较大的站点为例,利用Irmark-Allen、Hargreaves、Jensen-Haise 3种基于温度的ET 0计算方法,计算了 2013 2015 年两个站点的参考作物蒸散量,以FAO98 Penman-Monteith方法计算所得结果为标准,依据相关系数(R)及其显著性(P)、均方根误差(RMSE)和平均偏差(MBE)等量化指标,分别对3种方法计算结果在两个站点月和日序列的适用性进行评价,并对这3种方法进行本地化修正优化和检验。结果表明:本地化前,Irmark-Allen方法在海口的计算与Penman-Monteith的偏差最小且相关性好( R =0.97, P <0.01,RMSE=0.38 mm/d,MBE=-0.01 mm/d),其他两种方法均高估。3种基于温度的ET 0方法在敦煌都有很大的误差,其中Irmark-Allen方法在夏季偏低,在冬季偏高;Hargreaves方法整体偏高;Jensen-Haise方法在冬季不适用,出现无效负值,而在其他时段偏高。本地化后,3种基于温度的ET 0方法在两个地区都得到明显改善,其中Jensen-Haise方法在海口效果最好( R =0.96, P< 0.01,RMSE=0.61 mm/d,MBE=0.003 mm/d),在敦煌效果也是最好的( R =0.96, P <0.01,RMSE=0.69 mm/d, MBE=-0.02 mm/d)。

Applicability Evaluation of Temperature based Methods for Estimating Reference Evapotranspiration

The reference evapotranspiration (ET0) is one of the important indicators to characterize the degree of climate drying,vegetation water consumption,potential productivity and water supply demand balance.In this study,we evaluated the applicability of 3 temperature based ET0 methods including Irmark Allen,Hargreaves,and Jensen Haise in Haikou and Dunhuang stations with different climates compared to FAO98 Penman Monteith method and optimized and checked out in each local site.The daily and monthly series of ET0 were calculated from 2013 to 2015 using FAO98 Penman Monteith and 3 temperature based methods.The quantitative indices included correlation coefficient (R),significant (P) root mean square error (RMSE) and mean variation (MBE).Results showed that before localization,Irmark Allen method performed best with the lowest deviation and highest correlation (R=0.97,P<0.01,RMSE=0.38 mm/d,MBE=-0.01 mm/d) in Haikou,while other two temperature based methods overestimated ET0 compared to FAO98 Penman Monteith method.On the other hand,all 3 temperature based methods were with large biases in Dunhuang,the ET0 calculated by Irmark Allen method was underestimated in summer but overestimated in winter,by Hargreaves method was overestimated for the whole period,by Jensen Haise method was overestimated and invalid (ET0<0) in winter.After localization,the 3 temperature based methods were improved significantly in 2 regions.The optimized Jensen Haise method was best performed in Haikou (R=0.96,P<0.01,RMSE=0.61 mm/d,MBE=0.003 mm/d) and Dunhuang (R=0.96,P<0.01,RMSE=0.69 mm/d,MBE=-0.02 mm/d).