An All-Weather Observational Operator for Radiance Data Assimilation with Mesoscale Forecast Models

Abstract

Assimilating satellite radiance data under all weather conditions remains an outstanding problem in numerical weather prediction. This study develops an observational operator for use in radiance assimilation under both clear and cloudy conditions specifically for mesoscale models containing explicit microphysics. It is part of a larger research effort to build a 4D variational radiance assimilation system for optimal use of satellite data. The operator is suitable for radiance calculations at visible/infrared wavelengths and is adaptable to the different spectral characteristics of many types of narrowband satellite sensors. The new operator makes use of a gas extinction model and fast, multiple-scattering radiative transfer models, and relies on physical approximations for deriving cloud optical properties. One property, the asymmetry factor, is estimated through a new application of anomalous diffraction theory. A test of the observational operator's ability to estimate cloudy radiances was performed by forecasting a continental stratus system using the Regional Atmospheric Modeling System and computing radiances at all channels of the Geostationary Operational Environmental Satellite-9 imager. The forecasted radiances were found to reproduce very well the frequency distributions of observed cloudy radiances, particularly, the subtle temporal changes in the distributions that occurred between the early and late stages of development of the cloud system. These results are very encouraging and hold promise for future application of this observational operator in a full radiance assimilation system for satellite data across a wide range of wavelengths.