三种再分析资料在西沙海域的适用性分析

Applicability Analysis of Three Reanalysis Data in Xisha Sea Area

摘要: 利用2008年南海西沙永兴岛观测站(16˚50'N, 112˚20'E)的地面资料与2020年季风爆发前后的高空资料，与CRA40、ERA5和CFSR三种再分析资料进行了对比分析，讨论了三种再分析资料在西沙海域的适用性。结果表明：对于地面资料三种再分析在西沙站的适用性较好，均能够反映出海气要素和海气通量的变化趋势及南海季风爆发时的变化，CRA40和CFSR的海温与观测值相关性一般且显著低于观测值，ERA5的10 m风速和观测值相关性较差且明显偏低，对于海表温度的相关性则较好，优于其他两种再分析资料。CRA40的感热通量与观测数据相关性一般，其主要原因可能是CRA40的海温偏低，ERA5的相关性略好于CFSR，但其均方根误差和平均偏差偏高，三种再分析资料对潜热通量均表现出了较高的相关性，但三种再分析资料的潜热通量均偏高。总体上CFSR的海气通量在西沙海域的适用性优于其他两种再分析资料。对于高空资料，CRA40在各高度与观测值均有很好的相关性，整体上明显优于其他两种再分析资料，对于风速、相对湿度和比湿CFSR和ERA5适用性较差，在各层的相关性均较低，均方根误差和平均偏差也偏大，CRA40的高空资料在南海的适用性优于CFSR和ERA5。

Abstract: Based on the surface data of Yongxing Island observation station (16˚50'N, 112˚20'E) in Xisha, South China Sea in 2008 and the upper air data before and after the onset of monsoon in 2020, the three reanalysis data of CRA40, ERA5 and CFSR are compared and the applicability of the three reanalysis data in Xisha sea area is discussed. The results show that the three reanalyses of surface data have good applicability in Xisha site, and can reflect the changing trend of air sea elements and air sea flux during the outbreak of the South China Sea monsoon. The correlation between the sea surface temperature of CRA40 and CFSR and the observed value is general and significantly lower than the observed value, and the correlation between the 10m wind speed and the observation data of ERA5 is poor and significantly lower. The correlation of sea surface temperature is better than the other two reanalysis data. The sensible heat flux of CRA40 is generally correlated with the observed data. The main reason may be that the sea surface temperature of CRA40 is lower, and the correlation of ERA5 is slightly better than CFSR, but its root mean square error and average deviation are higher. The three reanalysis data show high correlation to the latent heat flux, but the latent heat flux of the three reanalysis data is much higher. In general, the applicability of CFSR sea air flux in Xisha sea area is better than the other two reanalysis data. For the upper air data, CRA40 has a good correlation with the observation data on each height, which is obviously better than the other two reanalysis data on the whole. For wind speed, relative humidity and specific humidity, CFSR and ERA5 have poor applicability. The correlation in each layer is low. The root mean square error and average deviation are also large. The applicability of CRA40’s upper air data in the South China Sea is better than CFSR and ERA5.