基于探空观测的CMA-RA资料在中国的适用性评估

Applicability Evaluation of CMA-RA Data Based on Radiosonde Observation in China

摘要:

基于中国125个探空站观测资料对比CMA-RA与ERA5、CFSv2、MERRA-2再分析资料，从整体效果、时间序列、空间分布3方面评估4类资料的位势高度、气温、相对湿度、风速产品在中国的适用性。结果表明：（1）再分析位势高度数据总体呈负偏差，对流层中下层CMA-RA表现优于ERA5，其余等压面层ERA5优于CMA-RA，其次为CFSv2，MERRA-2表现相对较差。（2）再分析气温数据总体呈负偏差，CMA-RA综合表现最优，ERA5与CMA-RA接近，MERRA-2、CFSv2次之。（3）再分析相对湿度数据总体呈正偏差，MERRA-2相对湿度整体表现最优，其次为ERA5，CMA-RA与CFSv2相对湿度空间分布类似，在对流层中下层适用性相对较好，在对流层中上层正偏差相对明显，且冬季偏差高于夏季。（4）再分析风速数据总体呈负偏差，CMA-RA风速整体表现最优，尤其是对流层中下层，其次为ERA5资料，MERRA-2资料表现最差，4类资料冬半年风速均方根误差普遍高于夏半年。CMA-RA气温、相对湿度、风速在2019年10月—2020年2月均方根误差偏大，使用前应进行偏差订正。

Abstract:

Using observation data of 125 radiosonde stations in China, comparing CMA-RA with ERA5, CFSv2, and MERRA-2 reanalysis data, the applicability of geopotential height, air temperature, relative humidity and wind speed in China of four types of reanalysis data were compared and evaluated from three aspects: overall effect, time series, and spatial distribution. The results showed that: (1) Reanalysis geopotential height data presents a negative bias generally. CMA-RA performs better than ERA5 in the lower-middle troposphere, while ERA5 performs better than CMA-RA in other isobaric surface, followed by CFSv2, and MERRA-2 performs relatively poorly. (2) Reanalysis air temperature data presents a negative bias generally. CMA-RA air temperature data performs best overall, and ERA5 is similar to CMA-RA, followed by MERRA-2 and CFSv2. (3) Reanalysis relative humidity data presents a positive bias generally, and MERRA-2 relative humidity data performs best overall, followed by ERA5. The spatial distribution of relative humidity data of CMA-RA is similar to that of CFSv2, which performs relatively well in the lower-middle troposphere but has a relatively significant positive bias in the upper-middle troposphere, and the bias in winter is higher than that in summer. (4) Reanalysis wind speed data presents a negative bias generally. CMA-RA wind speed data performs best overall, especially in the lower-middle troposphere, followed by ERA5, MERRA-2 performs worst, and the root mean square error of wind speed in winter half year is generally higher than that in summer half year for the four types of reanalysis data. The root mean square errors of temperature, relative humidity, and wind speed data of CMA-RA are relatively large between October 2019 and February 2020. Therefore, bias correction is needed before using to better improve data applicability.