四维数据同化在一次梅雨锋暴雨过程中的应用研究

Research on four-dimensional data assimilation applied in a Meiyu front rainstorm case

通过2008年梅雨季内的一次暴雨过程来着重研究WRF模式中四维测站同化技术的应用情况以及不同的同化策略对模拟结果的影响,使用的同化资料包括常规高空和地面探测资料。结果表明:高空探测资料的同化能够较好地修正中α尺度系统的移动发展速度,从而有效地改进强降雨带的落区位置;湿度观测资料的同化反而降低了强降雨带的降水强度,这可能因为同化湿度观测量可能减小近饱和区域内的水汽混合比,从而导致这些区域内的潜热加热减小,对流强度变弱,降水减少;地面观测资料的同化能够对模式行星边界层进行修正,这种同化影响通常在模式最低层上表现得较为明显,向上逐渐减小,而边界层模拟质量对降水至关重要,因此地面观测资料的加入使得模拟降水更加接近于实况。区域嵌套模拟试验表明对粗网格区域进行四维分析同化处理可以有效抑制中α尺度模式误差的增长,对细网格区域进行四维测站同化处理则可以进一步抑制中β尺度模式误差增长。

This paper mainly applied four-dimensional observation nudging assimilation in WRF model to a rainstorm case in the Meiyu season of 2008 and studied the impact of different assimilation strategies on the simulation results.The assimilation data include 12 h rawinsonde observations and 3 h surface observations.The results show that the assimilation of rawinsonde observations can correct the displacement and development of meso-α system,therefore,effectively improve the location of the heavy rain band.Oppositely,the assimilation of moisture observations underestimates the intensity of the heavy rain band,because reducing water vapor mixing ratio in the near moist saturation regions results in decreasing latent heat and weakening the intensity of convection and precipitation in these regions.The assimilation of the surface observations can improve planetary boundary layer in the model.Usually this kind of effect is evident in the model surface layer and tends to decreases with increasing height upward.The simulation quality of the PBL is very important to model precipitation,so that the inclusion of surface observations makes model precipitation more real.Regional grid-nested simulation tests indicate that the data assimilation on coarse grid can effectively restrain the error growth of meso-α system and the one on fine grid can further reduce the error growth of meso-β system.