Users of the Version 1.1 WCRP/SRB SW data set should be aware of the following issues in their use of the data.
1. SURFACE ALBEDO - The SW GEWEX (Pinker method) and SW QUALITY-CHECK (Staylor method) algorithms that produce surface albedo values differ by significant amounts. Surface albedo has a second-order effect on downward surface irradiance and a first-order effect on net surface irradiance. Albedo measurements applicable to the 280 km cell size SRB results have been obtained from short-term experiments in Greenland, the Amazon rainforest, and the U.S. Department of Energy (DOE) Southern Great Plains (SGP) Atmospheric Radiation Measurement (ARM) site. These very limited results indicate that surface albedos in the QUALITY-CHECK method are within 10 percent of measured values, whereas values of the GEWEX method differ by as much as 40 percent.
2. DIRECT-DIFFUSE RATIO - Experimental direct-diffuse ratio data has been obtained from the DOE National Renewable Energy Resources Laboratory (NREL) for several sites in the U.S. Comparisons with SRB results indicate that the GEWEX algorithm may be in error by as much as a factor of 5, particularly under clearer-sky conditions. The QUALITY-CHECK algorithm does not calculate direct-diffuse ratio.
3. AEROSOL ERRORS IN CENTRAL AFRICA AND SOUTH AMERICA - Discussions at previous WCRP/GEWEX/WGRF meetings have suggested that many SW ground site measurements from central Africa were not to be trusted. Neither SDAC nor the WGRF have been concerned by large differences between the satellite estimates and surface measurements in this region during previous reviews of the SRB product. For the Version 1.1 SRB product, 46-month average bias values range between 20 and 65 W/m**2, depending on location. Subsequent investigation indicates that clear-sky biases appear to correlated with the length of time into the biomass burning season. The same result was obtained for a site in the Amazon rain forest. SDAC is currently investigating the possibility of developing an improved aerosol climatology, which includes biomass burning for these regions based on clear-sky downward SW daily data from the World Radiation Data Center in combination short term aerosol experiments that have been conducted in these regions.
4. INCONSISTENCIES IN THE WESTERN PACIFIC REGION - SDAC has used a SUBSET of Global Energy Budget Archive (GEBA) sites which were limited to moderate terrain consistent with usual Global Climate Model (GCM) assumptions for its previous error analyses. ALL SITES from GEBA have been included in a new error analysis as per recommendations of the GEWEX Workshop on SRB Algorithms and Data, Williamsburg, VA, November 1993. The new analysis shows significant biases in the western Pacific region. Sites in Japan, the Marcus Island, the Ryukyu Islands, Malaysia, Thailand, Hong Kong, the Bonin Islands, South Korea, and North Korea have 46-month average bias values between 20 and 52 W/m**2 depending on location. Another issue is the Central Warm Pool region compared to the TOGA COARE area. The ratio of ISCCP daytime cloud parameters (and SRB SW surface cloud forcing values) for the two regions is approximately 1.0 for each parameter when averaged over 46 months. Earth Radiation Budget Experiment (ERBE) data indicates that the ratio of top-of-atmosphere cloud forcing between the two regions is 1.07 for the same time period. The Central Warm Pool region contains thousands of small islands whereas the TOGA COARE region is mostly open ocean with fewer islands. SDAC is concerned as to whether these inconsistencies in the western Pacific are acceptable to the WCRP for Version 2 GEWEX/SRB processing.