気象庁における数値天気予報へのAMSR-Eデータ利用

抄録

Observation data by Advanced Microwave Scanning Radiometer for EOS (AMSR-E) on board the U.S. earth observation satellite Aqua are used for Numerical Weather Prediction (NWP) models of Japan Meteorological Agency, namely MesoScale Model (MSM) and Global Spectral Model (GSM). The AMSR-E data have been assimilated in the Meso-Analysis (MA) since November 2004 and in the Global-Analysis (GA) since May 2006. This paper overviews some technical issues related the assimilation of AMSR-E data for MA and GA, respectively. For MA, 1) an efficient retrieval algorithm for precipitable water and rain rate, 2) a bias correction between the observations by similar instruments and between the observation and the model, 3) a four dimensional variational (4D-Var) scheme, 4) data thinning and smoothing scheme and 5) appropriate setting of observation error are key issues for successful assimilation of the AMSR-E data in operation. A pre-operational Observation System Experiment (OSE) shows the AMSR-E precipitable-water and rain-rate data improve the forecast in a heavy rain event. On the other hand, for GA, 1) a 4D-Var scheme, 2) a direct assimilation of AMSR-E radiance based on the effective radiative transfer scheme named RTTOV-7, and 3) a Variational Bias Correction (Var-BC) are key issues. A pre-operational OSE shows the improvement of typhoon-track forecast and precipitation distribution over the Indian Ocean. The accuracy of typhoon-track forecast has been improving recently by the improvement of the NWP models and the associated analysis system as well as the introduction of new observation data including microwave imager data. The observation network of the microwave radiometers by the Global Precipitation Measurement (GPM) mission will be constructed in the near future. Another promising data are provided by ground-based Global Positioning System (GPS) precipitable-water data for MA and GPS occultation temperature and water vapor data for GA.