基于格点的中国东北中北部2 m温度数值预报检验及偏差订正

Grid Temperature Numerical Forecast Verification and Correction over Central and North Region of Northeast China

基于中国气象局陆面数据同化系统(CLDAS-V2. 0)实时产品数据集中2 m气温数据对ECMWF高分辨率数值模式2 m气温预报产品在中国东北中北部的预报能力进行初步检验并利用递减平均法对系统性偏差进行订正。结果表明,气温平均预报准确率与海拔高度呈显著负相关,山区预报准确率偏低,系统性偏差较大。气温预报偏差还表现为明显的日变化特征,在夜间表现为预报较实况显著系统性偏高,白天系统性偏差不明显。冬季夜间气温,特别是最低气温系统性偏高的特征变得更加明显。递减平均法对系统性偏差的订正效果好,订正后,东北中北部地区冬季夜间气温及最低气温预报能力有大幅度提高。另外,递减平均法对东北中北部山区3、4及9月以外其他月份夜间气温及最低气温,对冬季白天气温及最高气温有显著的订正能力。

Based on 2-meter temperature products from CMA land date assimilate system,forecast ability about 2-meter temperature produced by high resolution numerical model of ECMWF was tested and systematic errors were corrected by using decaying averaging method.The results show that there was negative relation between the average forecast accuracy of temperature and altitude,and the average forecast accuracy was lower and the systematic error was larger in mountain area.The errors of temperature forecast showed a diurnal variation,and systematic error was higher during nighttime,and random error was the main error during daytime.The systematic error of temperature in winter night,especially for the minimum temperature was more significantly higher.The decaying averaging method performed better in systemic error correction.The forecast ability of minimum temperature in winter greatly increased after revision in the central and north region of Northeast China.In addition,the decaying averaging method worked well for 2-meter temperature during nighttime and minimum temperature in the mountain areas over the central and north region of Northeast China in the other months with the exception of March,April and September,and maximum temperature and temperature during daytime in winter.