VARIATIONAL ASSIMILATION OF GPS PRECIPITABLE WATER INTO MM5 MESOSCALE MODEL

The GPS precipitable water (PW) from 13 GPS sites in the Yangtze delta is explor ed to investigate the tune of background errors and the three dimensional variat ional assimilation of MM5 on rainfall event occurred from June 27 to 28,2002. T he results show: Background errors (BE) play a key role in the three dimensional variational assimilation of MM5. The horizontal scalelength of model variables ( u,v,T,pand q) is closely related to the average time of NMC technique. The scalelength of model variables is different for each other, which value is associated with the vertical height of the variable on the MM5 level. The BE cal culated by NMC technique reach the true BE more closely than that provided by MM 5-3DVAR system.GPS PW data can be assimulated into MM5 by using 3DVAR technique. After GPS PW d ata assimulation, the initial humidity field can be reanalyzed while the initial temperature, pressure and wind fields also being modified. 3DVAR of GPS PW is b enefitial to reduce analysis bias of PW in initial fields which can result in re straining PW prediction bias during the earlier period (3-6 h) of model integrat ion so as to improve PW prediction. The PW prediction improvement is related to the GPS receiver location in the area covered by GPS networks. By comparing the results with no GPS IPW ass imilation, we find that GPS PW assimilation can increase the accuracy of 6 h and 24 h accumulated precipitation prediction so as to improve the precipitation pr ediction ability of MM5. On the whole, GPS PW data assimilation will improve the precipitation prediction of MM5. However, there are several difficulties that will impact the GPS assimilation, w hich are how to get the true background errors and GPS PW errors at real time. I t will be further researches to develop a new technique to calculate more reason able background errors and GPS PW errors.