GRAPES全球模式背景误差协方差水平结构特征分析

Horizontal Structural Characteristics Analysis of Global GRAPES Model Background Error Covariance

基于1个月的GRAPES全球模式模拟资料,采用NMC方法统计了GRAPES全球模式的背景误差协方差,从背景误差的谱曲线和水平相关特征尺度2方面对背景误差协方差的水平结构进行分析,并与欧洲中心的计算结果进行比较。结果表明:2种模式的谱曲线变化趋势较为一致,背景误差的动能谱和位势高度功率谱均在1—7的波数范围逐渐增大,在7—14的波数范围取得最大值,之后随着波数的增大,谱值逐渐减小;在波数>20的范围内,总动能谱随波数的衰减主要由旋转动能谱随波数的减小造成。采用谱方法和格点方法进行水平相关特征尺度的计算,结果表明二者计算得到的背景误差的水平相关特征尺度随高度的升高而增大,随波数的增大而减小,但谱方法的计算结果大于格点方法。

Based on GRAPES - Global model forecast data in one month, the statistics of background error covariance of GRAPES - Global model was done by using NMC method. The spectrum curve and horizontal correlation characteristic scale of background error covariance were calculated and analyzed. Compared the results of GRAPES with ECMWF, the rotational kinetic energy spectrum, divergent kinetic energy spectrum, total kinetic energy spectrum and geopotential height power spectrum increased gradually when wave-number ranged from 1 to 7, and they reached maximum when wavenumber ranged from 7 to 14, then decreased slowly with increasing of wavenumber. The calculation values of GRAPES and ECMWF were also very close when wavenumber was less than 60. Meanwhile, the horizontal correlation characteristic scales were calculated separately by using spectral method and grid method, and results of both methods show that the horizontal correlation characteristic scales of background error increased with height and decreased with wavenu-mber, but calculation values of spectral method were larger than those of grid method.