P5 An examination of background error modeling for the WRF-Var system

Wang, Hongli, Xiang-Yu Huang, Juanzhen Sun, Dongmei Xu, National Center for Atmospheric Research, Shuiyong Fan, and Jiqin Zhong, Institute of Urban Meteorology, CMA, China

The Rapid Update and Cycling Data assimilation and Forecasting System at Beijing Meteorological Bureau (BJ-RUC), which includes the WRF model and the WRF 3D-Var system, has been run in operation since June 2008. In this paper, the features of background errors are investigated to further improve the performance of the BJ-RUC system. The background error statistics are extracted from operational forecasts in Summer 2012 using WRF Var gen_be utility. It is found that 1) u- and v-wind variances are underestimated and their auto-correlations are overestimated; 2) the background error variances for wind and temperature in June and July are larger than that in August, 3) the error variances for wind and temperature in the low atmosphere at 12Z are larger than that at 00Z. The possible reasons that lead to the above characteristics of the background errors are discussed. A new background error formulation, which uses u-, v-wind, temperature, surface pressure and pseudo relative humidity as control variables, is proposed and developed in WRF-Var for mesoscale data assimilation.