Assimilation of dual-polarimetric radar observations and its impact on two convective storms

Previous studies showed that the assimilation of Doppler radar data can improve the forecast of convective storms. On the other hand, radar data assimilation still remains a challenging problem due to several issues including the uncertainties in radar observations and the effectiveness of radar information to model variables.

Dual-polarimetric radars typically transmit both horizontally and vertically polarized radio wave pulses. Owing to the enhanced information, studies demonstrated that dual-pol Doppler variables provide more accurate measurement of liquid and solid cloud and precipitation particles than non-polarimetric weather radars. The assimilation of dual-pol radar data may be a potential way to improve the performance of radar data assimilation. However, not much effort has been given into the dual-pol radar data assimilation research field in past studies, partly due to unavailability of dual-pol radar data and the special requirements for processing and assimilating these data. With the upgrade of the current U.S. NEXRAD radar network to include dual-polarimetric capabilities, the time is upon us to begin exploring how the polarimetric data can improve forecast initialization.

Our main goals in this study are to develop the methodology to assimilate polarimetric radar variables; to seek improvement in storm initialization and short-term quantitative precipitation forecast with the assimilation of dual-pol radar variables; and to understand how the dual-pol variables can be used more efficiently in numerical weather prediction. High resolution (~1 km) WRF model and its 3DVAR data assimilation system are used in our work. The dual-polarimetric radar data used in our study is collected by the C-band Advanced Radar for Meteorological and Operational Research (ARMOR) radar (located at Huntsville International Airport (34.6804°N, 86.7743ºW)). We will present our recent work on assimilating the ARMOR radar data for two convective storms, a squall line case on 15 March 2008 and a thunderstorm on 23 June 2008. Details of the methodology of data assimilation and the results for the two storms will be presented in the conference. We will also compare and discuss the influences of different dual-pol variables on model initial condition and on the short-term prediction of precipitation.