CLDAS气温数据在中国区域的适用性评估

Applicability assessment of CLDAS temperature data in China

以中国48708个地面气象自动站逐小时气温数据为基础,采用平均偏差(Bias)、相关系数、均方根误差(RMSE)、平均绝对误差(MAE)等评估指标,对比分析2017年10月、2018年1月、4月、7月CLDAS-V2.0气温(分辨率为0.0625°),探讨中国8个分区春、夏、秋、冬4个季节CLDAS与站点气温的相关性及偏差分布特征。结果表明:1)CLDAS气温较好地反映了中国气温的年际变化,非独立性检验、独立性检验与站点气温的平均相关系数分别为0.995、0.991,东北地区相关性最高,西南地区相关性最低。2)CLDAS与站点气温的Bias为-0.011℃,非独立性检验的RMSE、MAE分别为1.275、1.645℃,独立性检验的RMSE、MAE分别为0.867、1.089℃,总体上CLDAS气温误差小,可信度较高。3)春、秋季的偏差小于夏、冬季;东北、华北、江淮、华南地区的偏差小于西北、西南地区;84.6%站的冷偏差或暖偏差在1℃内,冷暖偏差空间分布均匀。4)CLDAS的最高气温存在冷偏差,最低气温存在暖偏差,夏季最高气温的最大误差为-0.59℃。5)CLDAS平均偏差的日变化为-0.23~0.07℃,白天呈冷偏差,夜间呈暖偏差,夏季平均偏差的日变化较显著,偏差的日较差为0.26℃;全国8个分区夏季平均偏差日变化最大为1.06℃,秋、冬、春季变幅相似,西南地区平均偏差日变化最大而江淮地区最小。

Based on the hourly temperature data derived from 48708 surface automatic weather stations in China,the CLDAS-V2.0 temperature data in October 2017,January 2018,April 2018 and July 2018(resolution:0.0625°)were analyzed by using the evaluation indexes such as mean deviation(Bias),correlation coefficient,root mean square error(RMSE)and mean absolute error(MAE).This paper studied the correlation and deviation distribution characteristics between CLDAS and station temperature in spring,summer,autumn and winter in eight zones of China.Results show that:(1)CLDAS temperature reflects the interannual variation of temperature in China,and the average correlation coefficients of non-independence test,independence test and station temperature are 0.995 and 0.991 respectively.The correlation coefficient is the highest in Northeast China and the lowest in Southwest China.(2)The Bias of CLDAS and station temperature is-0.011℃,RMSE and MAE of non-independence test are 1.275℃and 1.645℃,and RMSE and MAE of independence test are 0.867℃and 1.089℃,respectively.In general,CLDAS has small temperature error and high reliability.(3)The deviation in spring and autumn is less than that in summer and winter.The deviation in Northeast,North China,Jianghuai and South China is smaller than that in Northwest and Southwest China.In 84.6%stations,the cold deviation or warm deviation is within 1℃,and the spatial distribution of cold and warm deviation is uniform.(4)The maximum temperature of CLDAS has cold deviation,the minimum temperature has warm deviation,and the maximum deviation of maximum temperature in summer is-0.59℃.(5)The diurnal variation of mean deviation of CLDAS is-0.23—0.07℃,colder in the daytime and warmer in the night.The diurnal variation of mean deviation of CLDAS is significant in summer,and the diurnal range of deviation is 0.26℃.The maximum diurnal variation of summer mean deviation in the eight zones of China is 1.06℃,and the variation ranges in autumn,winter and spring are similar.The diurnal variation of mean deviation is the largest in Southwest China and the smallest in Jianghuai.