Abstract:

Quality control (QC) is a key for microwave sounding radiance data assimilation in numerical weather prediction (NWP) models. In this study, a new approach of QC for Advanced Microwave Sounding Unit-A (AMSU-A) radiance was developed by introducing brightness temperature gradient of view point. In this approach, the data were considered to be contaminated by precipitation if their gradients were anomalously large, and would be discarded. Two typhoon cases, Kalmaegi in 2008 and Kompasu in 2010, were simulated by WRF (Weather Research and Forecasting) model to see the impacts of the new QC method on AMSU-A radiance. In the new QC cases, as the contaminated data removed from the three-dimensional variational assimilation (3DVar), the tracks of both typhoons were much closer to the observations due to the increments of initial geopotential height and wind fields. In addition, the typhoons’ intensities were much closer to the observations with more reasonable temperature field for the warm structure of typhoon core development. In the original cases, with much more contaminated AMSU-A radiance in WRF, both typhoons’ tracks and intensities were much worse simulated. An improvement ratio to simulated track error was 12, that is an improvement to 540 km compared to the track without AMSU-A radiance assimilated, in the Kalmaegi case. And the ratio was 13 with 118 km improvement to simulated track in the Kompasu case.

Key words: AMSU-A, gradient information, quality control, numerical prediction of typhoon, variational assimilation