Operational Assimilation of Data Retrieved by GNSS Observations into GRAPES_Meso 3DVar System

Radio occultation (RO) observations using the Global Navigation Satellite System (GNSS) provides valuable data to support operational numerical weather prediction, and it is proved that the assimilation of RO data has the potential to significantly improve the accuracy of global and regional meteorological analysis and weather prediction. RO observations have many advantages such as high precision, global coverage and high vertical resolution compared with other observations. RO observations provide phase (Level 1a), bending angle (Level 1b), refractivity (Level 2a), retrieved pressure, temperature and humidity profiles (Level 2b), all of which can be assimilated into numerical model. Bending angle and refractivity are proved better choices for assimilation as observation operators are less complicated, and have no disadvantages associated with the modeling of ionospheric effects in the assimilation model. However, assimilation of water vapor, temperature or pressure derived from RO observations have advantages that data forms are same as model variables, and it is proved that assimilation of retrieved data also improve the accuracy of global analyses and forecasts significantly.In the operational GRAPES (Global/Regional Assimilation and Prediction System) regional (GRAPES_Meso) 3-dimensional variational (3DVar) system, the analysis is performed in model vertical level, but it can only assimilate the upper-level wind, pressure and humidity from radiosonde report (TEMP), upper-level wind and humidity by AIREP, cloud drift wind (SATOB), the pressure from surface station report over land (SYNOP) and over sea, integrated column precipitable water vapor (IPW) retrieved by ground-based GPS observations and radar wind profiles. In order to assimilate RO retrieved atmosphere data and address their impacts on analyses and forecasts, observation operators of retrieved humidity and pressur eassimilation are developed first, and then quality control and vertical thinning scheme are formed and discussed. Results of one month experiments show that the assimilation of GNSS/RO retrieved pressure and humidity data have considerable positive impacts on analyses of geopotential height, minor positive impacts on humidity analyses, geopotential height, humidity forecasts, rainfall forecasts, and major positive impacts on rainfall forecast for Qinghai-Tibet region.