The Assimilation of ATOVS Radiances in the JMA GIobal Analysis System

[摘要] A scheme for the assimilation of radiance data from the Advanced TIROS Operational Vertical Sounder (ATOVS) into the global three-dimensional variational (3DVar) analysis system at the Japan Meteorological Agency (JMA) is described. It makes better use of ATOVS observations than the previous ATOVS retrieval assimilation scheme. Several procedures have been developed: advanced thinning, identification of cloud/rain-affected radiances, removal of radiance data which are erroneous or not well simulated with a current fast radiative transfer model and/or numerical weather prediction (NWP) model, selection of adequate channels and assignment of observation errors for different observation conditions, such as clear/cloud/rain and surface type, and correction of radiance biases in which the effects of NWP model biases are minimized.It was found from parallel experiments that the radiance assimilation shows significant improvement over the retrieval assimilation in many respects. Analyzed temperature and water vapor are improved when verified against radiosonde and the Special Sensor Microwave/Imager (SSM/I) observations. Impacts on forecasts are positive on the globe, especially for short-range forecasts. The reduction in the root mean square errors by the radiance assimilation reaches up to 0.5 K for the analyzed temperature against RAOBs, and about 5-10% for the 500 hPa geopotential height at day 1 to day 4 forecast. As a result of these findings, the ATOVS radiance assimilation was operationally implemented in the JMA global analysis system on 28 May 2003.

[学科分类] 大气科学