A study on the application of FY-2C cloud drift wind in a mesoscale numerical model.

Though the cloud drift wind (CDW) has displayed its good application perspective in numerical weather prediction (NWP), up to the present the CDW data are not actually applied to the daily operation of NWP. The main reason is that the CDW data have abnormal error, and moreover, study on the errors and quality control of CDW data is very lacking. Further research is needed in order to effectively utilize the CDW data in the daily NWP. This paper attempted to explore the systematic error character of FY-2C CDW and its effects on the initial fields and forecast results of NWP model so as to promote the application of CDW data in operational NWP. At first, statistical test and error analysis to the CDW data of FY-2Cwere performed in terms of the radiosonde observations; the quality control of each record of the CDW data was realized individually through the systematic bias correction and the check of thermal wind relation; then, the CDW records after quality control were incorporated into and assimilated in the GRAPESMeso model by the GRAPES 3Dvar scheme; and at last, the effect of the CDW data were examined through the contrast experiments of a torrential rain event in Northwest China in the period from 00:00 UTC 1 to 00:00 UTC 2 July 2005. The analysis results of the CDW data indicate that: the number of CDWs vertically, mainly distributed in the midupper troposphere above 500 hPa, with the maximum number at about 250 hPa. The CDW under 500 hPa was characteristic of wind direction error, large wind speed error and divergent error distribution, and therefore it has a bad usability; while the CDW above 500 hPa was characteristic of smaller wind direction and speed errors and a Gaussian error distribution, and therefore it has a better usability. Seeing that the CDW at higher levels had systemic bias error and the wind was coherently distributed in the vertical direction, so the bias correction and thermal wind relation were employed to control the quality of the FY-2 CDW data. In the quality control, the CDWs with wind direction error or bigger wind speed error were eliminated, the quality of the CDW used in the assimilation is improved. After the assimilation of CDWs, the southwest airflows near the torrential rain area became stronger in the initial wind fields of numerical experiments, which led to the enhancement of vapor transportation and convergence and finally to the evident improvement of the 24 hours forecast of the torrential rain in rain intensity and area. As a whole the proportional distribution of CDW data was inferior to the radiosonde observations data, but the local CDW was much more denser than the radiosonde observations' near the area of torrential rain occurred. The CDW data reflected information of the local torrential rain synoptic system more particularly and accurately, which is the reason why the 24 hrain forecast in CDW dataassimilated experiment is better than the forecast in radiosonde data assimilated experiment. 