广东省GPS/PWV资料的质量控制及其对前汛期降水预报影响的初步研究

A PRELIMINARY STUDY ON THE QUALITY CONTROL METHOD FOR GUANGDONG GPS/PWV DATA AND ITS EFFECTS ON PRECIPITATION FORECASTS IN THE ANNUALLY FIRST RAINING SEASON OF GUANGDONG

对2009年8月—2012年8月广东省地基GPS反演的可降水量(GPS/PWV)资料进行分析,建立了该资料的质量控制方案。该方案能合理剔除取值不合理的"离群"资料及与背景场偏差较大的资料。选取2010年5月6—7日及2012年4月27—28日两次前汛期降水过程进行数值试验,经过质量控制,只允许与背景场偏差较小的GPS/PWV资料进入GRAPES三维变分同化系统。并对初始湿度场进行合理调整,使数值预报模式的24 h累积降水量的量级和落区的预报得到明显改进,从而提高数值预报模式的降水预报能力。

First analyzed GPS precipitable water vapor（GPS/PWV） available from a ground-based GPS observation network in Guangdong from 1 August 2009 to 27 August 2012 and then developed a method of quality control before GPS/PWV data is assimilated into the GRAPES 3DVAR system. This method can reject the outliers effectively. After establishing the criterion for quality control, we did three numerical experiments to investigate the impact on the precipitation forecast with the GPS/PWV data quality-controlled or not before they are assimilated into the GRAPES 3DVAR system. In the numerical experiments, two precipitation cases（on 6 to 7 May, 2010 and 27 to 28 April, 2012 respectively） that occurred in the annually first raining season of Guangdong were selected. The results indicated that after quality control, only the GPS/PWV data that deviates little from the NCEP/PWV data can be assimilated into the GRAPES 3DVAR system, has reasonable adjustment of the initial water vapor above Guangdong, and eventually improves the intensity and location of 24-hour precipitation forecast significantly.