Statistical comparisons of model output with satellite observations: A severe weather case

Geostationary Operational Environmental Satellite-R (GOES-R) and National Polar-Orbiting Operational Environmental Satellite System (NPOESS) Preparatory Project (NPP) risk reduction activities involve the creation of synthetic imagery and using them to develop new products in advance of satellite launch. It is important to analyze the performance of our models when reproducing actual weather events in order to assess the usefulness of any of the products developed before satellite launch. Also such comparisons can be used to quantify model error statistics for data assimilation purposes. We will present results from such a statistical comparison performed on the simulation of an outbreak of thunderstorms in Oklahoma and Kansas on May 8, 2003. CSU Regional Atmospheric Modeling System (RAMS) was used to model the thunderstorms while a forward observational operator was used to compute the observable satellite radiances. Observed radiances from the GOES imager and MODIS are used for the comparison.