The NASA Global Energy and Water-cycle Experiment (GEWEX) SRB project has now updated and improved a global data set of surface radiative fluxes on a 1-degree grid for a 24-year period (July 1983 to June 2007). Cloud and surface properties for input to the algorithms were derived from International Satellite Cloud Climatology Project pixel level (DX) data. However, the temperature and humidity profiles were derived from the NASA Global Modeling and Assimilation Office (GMAO) Goddard Earth Observing System version 4.0.3 reanalysis products. Errors in the temperature and humidity profiles, particularly in the boundary layer and surface, produce errors in the resulting radiative fluxes. Zhang et al. (2006) assessed GEOS1, ECMWF 15-year assimilation, the NCEP reanalysis, the TIROS Operational Sounder (TOVS) and the ISCCP modified TOVS meteorological distributions for sensitivities of the radiative fluxes to these inputs. Studies like Zhang et al. have shown that the largest uncertainties in the computation of the longwave radiative fluxes remain the diurnal mean and variation of the boundary temperatures, humidity and Earth skin temperature. Future versions of the GEWEX SRB will require improved estimates of these quantities and one candidate is the use of the MERRA data as input.

This study builds upon that Zhang et al. work by contrasting the GEOS-4, ERA-40, and early results of the latest GMAO assimilation product called Modern Era Reanalysis for Research and Applications (MERRA). Comparisons are made for those quantities most important for the computation of surface radiative fluxes including total column precipitable water (TPW), skin-temperatures, and near-surface temperatures and humidity. The monthly averages and a limited analysis of monthly averaged diurnal cycles will be contrasted at a few validation sites. Preliminary comparisons between GEOS-4 and test MERRA skin temperatures show that on a monthly average, MERRA is cooler over Antarctica during January 2004, but slightly warmer during July. MERRA has 20-30% greater TPW in July 2004 than GEOS-4. While these differences in meteorology produce only slight changes on surface shortwave fluxes (less than 0.5 W/m2 global average difference using MERRA vs. GEOS-4), surface longwave flux differences can be much greater both globally (1.5 W/m2) and regionally (greater than 5.5 W/m2 for polar regions). This analysis will be completed for a select number mid-seasonal months for years containing official MERRA results (currently 1989-1990, and 1998 have been completed) to provide a preliminary estimate of these sensitivities with the updated assimilation products.