多套土壤温湿度资料在青藏高原的适用性

Applicability of Soil Temperature and Moisture in Several Datasets over Qinghai-Xizang Plateau

利用青藏高原中部和东部土壤温度和湿度观测资料,通过计算两套再分析资料(ERA-Interim和CFSR)和六套陆面模式资料(ERA/land、MERRA/land、GLDAS-NOAH、GLDAS-CLM、GLDAS-M OSAIC和GLDAS-VIC)分别与观测资料之间的平均偏差、偏差标准差、相关系数、标准差比等统计参数,结合Brunke排名法,综合评估了再分析资料和陆面模式资料中土壤温湿度数据在青藏高原的适用性。结果表明:对于土壤温度,CFSR与观测值最接近,其次是MERRA/land和GLDAS-CLM,而ERA-Interim和ERA/land与观测值相差较大;除GLDAS-CLM土壤温度比观测值偏高外,其他资料土壤温度在大部分站点比观测值偏低,其中ERA-Interim和ERA/land土壤温度比观测值偏低较多,部分站点平均偏差超过-20℃。对于非冻结期(5 10月)土壤湿度,GLDAS-CLM与观测值最接近,其次是GLDAS-NOAH或ERA-Interim;与观测值相比,CFSR、ERA-Interim和ERA/land的土壤湿度偏湿,平均偏差大部分在0.05~0.20 m3·m-3之间,而GLDAS-NOAH、GLDAS-CLM和GLDAS-M OSAIC的土壤湿度偏干。

In situ soil temperature and moisture observations at 7 stations and one region( Naqu) over the Qinghai-Xizang Plateau are used to validate two reanalysis products( i. e. ERA Interim and CFSR) and six land surface model products( i. e. ERA / land,MERRA / land,GLDAS- NOAH,GLDAS- CLM,GLDAS-MOSAIC and GLDAS- VIC). Four statistical quantities,i. e. mean bias( BIAS),standard deviation of differences( σd),correlation coefficient( R) and ratio of standard deviations( σr/ σobs),are calculated at each site,and the Brunke ranking method is applied to quantify the relative performance of the eight datasets for each variable and statistical quantity. The results showthat for daily soil temperature CFSR has the best overall performance,followed by MERRA / land and GLDAS- CLM,while ERA Interim and ERA / land perform the worst. GLDAS-CLM tends to overestimate daily soil temperatures,while other datasets tend to underestimate soil temperatures at most observation sites,with ERA Interim and ERA / land having large cold bias exceeding- 20℃. For soil moisture during unfreezing period( May to October),GLDAS- CLM shows the best overall performance,followed by GLDAS- NOAH and ERA Interim. CFSR,ERA Interim,and ERA / land have wet biases,with most of the biases between 0. 05 and 0. 20 m3·m- 3,while GLDAS- NOAH,GLDAS- CLM and GLDAS- MOSAIC tend to be drier than observations.