Impact of a Unified DMSP Meteorological Sensor Retrieval Methodology on Global Numerical Weather Prediction.

摘要

An observing system simulation experiment (OSSE) is conducted using simulated DMSP microwave sensor data, surrogate data for cloud imagery (all based on a natural run), and the application of the unified retrieval (UR) method for DMSP to obtain meteorological variables as proposed by Isaacs (23). Both the DMSP SSM/T-1 temperature sounder and SSM/T-2 moisture sounder data are simulated. The unified retrieval uses physical considerations based on a forecast or other first guess by comparing the sensor data with those simulated using the first guess profiles. Although the preliminary UR did not perform better than the statistical retrieval, enhanced UR are found to be more accurate in terms of integrated water vapor amounts. The retrieved variables, primarily temperature and moisture, are used to construct initial conditions for input to the AFGL global spectral numerical weather prediction model. The resulting analyses and forecasts based on the present statistical retrieval scheme. The impact of the preliminary UR used in this experiment was small. However the enhanced UR should provide improved humidity analyses. Keywords: Satellite meteorology; Water vapor; Clouds; Unified retrieval; Numerical weather prediction. (JHD)