Inhomogeneous Background Error Modeling for WRF-Var Using the NMC Method

• DOI: 10.1175/jamc-d-13-0281.1
• Publication Date: 2014-08-22T15:33:34Z
• Authors: Hongli Wang; Xiang-Yu Huang; Juanzhen Sun; Dongmei Xu; Man Zhang; Shuiyong Fan; Jiqin Zhong
• Keywords: Quantitative precipitation forecast; Errors-in-Variables Models

Abstract Background error modeling plays a key role in variational data assimilation system. The National Meteorological Center (NMC) method has been widely used systems to generate forecast ensemble from which the climatological background covariance can be modeled. In this paper, characteristics of via NMC are investigated for system Weather Research and Forecasting (WRF-Var) Model. statistics extracted short-term 3-km-resolution forecasts June, July, August 2012 over limited-area domain. It is found 1) that variances vary month also have feature diurnal variations low-level atmosphere 2) u - υ -wind underestimated their autocorrelation length scales overestimated when default control variable option WRF-Var used. A new approach transform (CVT) proposed model based on method. capable extracting inhomogeneous anisotropic information obtained Single observation experiments show not only merit incorporating geographically dependent information, but able produce multivariate analysis. results assimilaton study real convective case use CVT improves synoptic weather precipitation up 12 h.