9.1    On the Bias-Correction Strategy of Radiance Data Assimilation with the Limited-Area Model

Liu, Zhiquan, Hui-Chuan Lin, Craig S. Schwartz, and Ying-Hwa Kuo, National Center for Atmospheric Research

Bias correction (BC) is crucial to properly assimilate radiance data and obtain the positive impact. At the operational NWP centers, BC is usually applied to both global and regional data assimilation system through a global set of BC coefficients.  For community research applications using a limited-area model such as WRF, a straight way is to apply BC coefficients obtained over the domain under investigation through the advanced scheme such as Variational BC (VarBC).  In this study, several 4-month long (August ~ November, 2008) experiments were conducted to investigate the sensitivity of radiance data impact on the BC strategy.  Firstly, we found that using a set of pre-trained BC coefficients at the beginning of data assimilation cycles is beneficial when comparing to start cycling from no knowledge of BC coefficients (i.e., cold-start BC).  Moreover, we found that the global statistics of BC coefficients are different from the regional ones over the Northern or Southern high-latitude regions ASR or Antarctic-centered domain, and radiance data assimilation with applying global BC coefficients outperforms those using regional BC coefficients.