基于探空资料评估再分析资料中的QBO特征

EVALUATING THE QUASI-BIENNIAL OSCILLATION IN REANALYSISDATASETS BASED ON RADIOSONDE DATA

关键词:  QBO  年际变化  气候趋势  再分析资料  探空资料  评估

Key words:  QBO  inter-annual variation  climate trend  reanalysis dataset  radiosonde data  assessment

摘要:

使用柏林自由大学发布的1980—2020年近赤道站点探空月平均纬向风资料，首先评估ERA5赤道地区平流层中低层纬向风，然后结合ERA5、NCEP1、NCEP2、CFSR、JRA55和MERRA2等六套再分析资料和探空资料，分析了热带平流层准两年振荡（Quasi-Biennial Oscillation，QBO）位相转换时间、周期和振幅的年际变化以及QBO周期和振幅的长期变化趋势。结果表明，ERA5的赤道平流层中低层纬向风存在明显的QBO信号，且ERA5的各层纬向风与探空资料的相关系数都大于0.97，除10 hPa外，均方根误差（RMSE）都小于3；再分析资料与探空资料的QBO位相转换时间的差异主要为超前或滞后1个月，总体上，ERA5和MERRA2与探空资料最接近；再分析资料与探空资料 QBO周期以及东、西风位相持续时间都很接近，相关系数高达0.97以上，RMSE几乎都小于1，且几乎都无显著长期变化趋势，但都表现出随时间越来越不稳定的特征，其中ERA5和MERRA2同样与探空资料更接近；探空资料的QBO振幅在10 hPa和50 hPa显著减弱，30 hPa和70 hPa显著增强，但在20 hPa变化趋势不明显，ERA5的QBO振幅变化趋势和年际变化与探空资料最接近，其次是JRA55和MERRA2。

Abstract:

Using the radiosonde observation dataset produced by Free University of Berlin (FUB) from 1980 to 2020, this study first evaluates the zonal wind in the middle and lower equatorial stratosphere of ERA5. Then this study analyzes the inter-annual variation of QBO phase conversion time, period and amplitude and the long-term variation trend of QBO period and amplitude in combination with six datasets of reanalysis including ERA5, NCEP1, NCEP2, CFSR, JRA55 and MERRA2 and radiosonde data. The results show that the zonal wind in the middle and lower equatorial stratosphere of ERA5 shows QBO signal. The correlation coefficient between ERA5 and radiosonde data in each layer is greater than 0.97 and the root mean square error (RMSE) is less than 3 (except for 10 hPa). The difference of QBO phase conversion time between reanalysis datasets and radiosonde data is mainly one month ahead or behind. Generally, ERA5 and MERRA2 are the closest to radiosonde data. The QBO period and the phase duration of the easterly and westerly from the reanalysis data are very close to those from the radiosonde data. The correlation coefficient is more than 0.97, the RMSE is almost less than 1, and there is almost no significant long-term change trend. However, all of them show the characteristics of becoming more and more unstable with time, ERA5 and MERRA2 are also closer to the radiosonde data. The QBO amplitude from radiosonde data decreases significantly at 10 hPa and 50 hPa while increases significantly at 30 hPa and 70 hPa, but the change trend is not obvious at 20 hPa. The long-term variation trend and inter-annual variation of QBO amplitude from ERA5 are the closest to the radiosonde data, followed by JRA55 and MERRA2.