Abstract

The Weather Research and Forecasting (WRF) model is a new generation advanced mesoscale (10km-1km) model for both operational numerical weather prediction (NWP) and atmospheric research and it is the successor of the MM5, Eta, and RUC model system. GPS precipitable water (PW) which was obtained f... Show moreThe Weather Research and Forecasting (WRF) model is a new generation advanced mesoscale (10km-1km) model for both operational numerical weather prediction (NWP) and atmospheric research and it is the successor of the MM5, Eta, and RUC model system. GPS precipitable water (PW) which was obtained from dense ground-based GPS networks in Japan and the U.S. was assimilated into the WRF 3DVAR System. The two networks used are (1) GEONET in Japan: one of the densest operational networks in the world with averaged spacing of about 20km, and (2) U.S. GPS network (FLSNET, SUOMINET, and CORSNET): these networks cover the continental U.S. with station spacing from several ten km in some dense regions to several hundred km. The horizontal resolution in the WRF simulation is 5km and 10km for domains of the Japanese islands and the continental U.S., respectively. We have tuned the optimal scale length of specific humidity used in the recursive filter in the 3DVAR for each network. Several cases of dynamic weather conditions from synoptic to mesoscale phenomena from the summer of 2003 to the summer of 2004 were chosen for the data assimilation experiment. Improved rainfall forecast due to assimilation of GPS PW were observed for several hours into the forecast. The impact for longer forecasts was neutral. Our results suggest that GEONET PW can be used for short term forecasts. Long-term forecasts do not seem to benefit from GPS PW presumably because no observations are available from the oceans surrounding Japan. The results of the model simulations in each case will be presented and discussed. The tropospheric gradient parameters estimated by using ray-traced slant delay in the model grid show similar variations with those estimated by GPS analysis in the dynamic weather condition, suggesting the gridded point value (GPV) of the NWP with the high space and temporal (1 hour) resolution may be helpful information for GPS analysis. The positive feedback system between NWP system and GPS analysis system will be also discussed. Show less