Producing high quality proxy datasets from NWP simulations for the future ABI are important in developing and testing new products and algorithms. This study evaluates high-resolution simulations of top-of-atmosphere (TOA) radiances in cloudy regions using GOES-12 imager data to determine whether these proxy datasets have sufficient realism for testing cloud products and algorithms. The WRF model was used to produce retrospective forecasts on 5 June 2005 using 50 vertical levels and 6-km horizontal grid spacing that spanned nearly the entire region covered by the GOES-12 imager during full-disk mode. These simulations also provide a higher resolution (2 km) CONUS domain and a fine resolution (667 m) small domain located in the central U.S. To compute TOA radiances and reflectances for all of the imager bands, the gas absorption model in the NOAA Community Radiative Transfer Model was used together with the SOI (Successive Order of Interaction) radiative transfer model.

Our validation approach examines imager channel differences and relationships between different channels. How well the simulations reproduce observed cloud types will be determined by looking at 2D histograms of the visible reflectances from band 1 and brightness temperatures from the infrared window channel (band 4). Brightness temperature differences between the water vapor band (channel 3) and band 4 will reveal clues as to the ability of the WRF simulations to produce realistic physical properties for very cold (band 4 brightness temperature < 210 K) clouds. Comparisons of 3.9 microns (band 2) reflectance data will also be done to evaluate model-simulated cloud particle sizes. Results will be presented at the conference.