Characterizing the Impact of Hyperspectral Infrared Radiances near Clouds on Global Atmospheric Analysis

Hyperspectral infrared measurements from modern instruments, namely the Atmospheric Infrared Sounder (AIRS), the Cross-Track Infrared Sounder (CrIS), and the Infrared Atmospheric Sounding Interferometer (IASI), are highly sensitive to the presence of clouds within a field-of-view. The Goddard Earth Observing System (GEOS) atmospheric data assimilation system, which utilizes the Gridpoint Statistical Interpolation (GSI) algorithm, is employed at the Global Modeling and Assimilation Office (GMAO) at NASA Goddard Space Flight Center. In routine forward processing, it only considers observations that are insensitive to clouds. This screening is performed by retrieving the cloud height of an observation using a minimum residual method and then excluding channels sensitive to clouds by comparing the retrieved cloud height with their respective transmittance profiles.

Certain assumptions within the minimum residual retrieval, however, can be violated in any measured spectrum affected by clouds. Specifically, there is no guarantee that the clouds will have only a single layer or that the cloud emissivity is constant in wavenumber. Additionally, the channels used in the cloud height retrieval may be suboptimal. These issues will all be evaluated under the framework of both traditional observation statistical metrics (e.g. observation departures, observation bias correction) as well as the observation impact calculations, as described in Gelaro et al. (2010). This work is being performed in conjunction with an effort to expand hyperspectral infrared radiance assimilation to further assimilated cloud-affected radiances. This effort will analyze the “gray” area between clear-only infrared radiance assimilation and clear and near-opaque cloudy infrared radiance assimilation.