Surface radiation budget parameters are being produced operationally in real time at NOAA/NESDIS from GOES observations at hourly time scale, at 0.5-degree spatial resolution, and delivered to the GCIP/GAPP community and to the public, in real time. The current product is being used extensively in a wide range of applications, such as validation of numerical weather prediction models, land data assimilation activities, hydrological modeling, net primary productivity assessment, snow melt estimation, solar energy applications, and air quality modeling. Extensive product evaluations against ground observations have shown that the product is of high quality at both daily and hourly time scales, during average snow-free conditions. It was learned that the product needs to be improved under situations frequented with sub-pixel clouds, clouds over snow, and in respect to calibration of satellite sensors. To meet such needs, Clouds for AVHRR (CLAVR) methods were adapted for the GOES satellites, including contrast signatures, spectral signatures, and spatial signatures, and new snow detection techniques were incorporated. New surface type classifications have been implemented in the radiative transfer schemes and new and improved aerosol climatologies have been developed and incorporated. In this presentation we report on progress made towards improving the GCIP/GAPP surface radiation product.