Long-term validation and variability of the shortwave and longwave radiation data of the GEWEX Surface Radiation Budget (SRB) project

The Surface Radiation Budget (SRB) project of NASA is a principal component the Global Energy and Water cycle Experiment (GEWEX) program of the World Climate Research Programme. The project has recently produced and archived a global 21.5-year (July 1983 - December 2004) climatology of shortwave (SW) and longwave (LW) surface radiation fluxes along with other variables using the upgraded algorithms (Version 2.5). This development is based on the previous 12-year SRB data for the years from July 1983 to October 1995 (Version 2.0 for shortwave and 2.1 for longwave) and contains new key inputs including the Global Earth Observing System (GEOS) Version 4.0.3 meteorological information. The data sets cover the entire globe with a grid-scheme which attempts to divide the Earth's surface into equal-area grids that are 1 degree longitude by 1 degree latitude in tropical regions.

Up to the present, we have systematically compared the SRB data, which are presented as 3-hourly, 3-hourly-monthly, daily and monthly means, with the Baseline Surface Radiation Network (BSRN) data, World Radiation Data Center (WRDC) data and Global Energy Balance Archive (GEBA) data. The original BSRN measurements of shortwave and longwave radiation were made at 1-, 2-, 3- or 5-minute intervals, and we processed the data to generate 3-hourly, 3-hourly-monthly, daily, and monthly means. The WRDC data are in daily means, and monthly means are computed therefrom. The GEBA data provide only monthly means, but a large number of GEBA sites have long enough records to cover the entire 21.5-year duration of the SRB data. Systematic comparisons show that the satellite-based SRB data are in generally very good agreement with the ground-based observations throughout the world.

In this paper, we focus on those SRB grids whose corresponding ground stations have the longest records. The variabilities, correlations, and trends of the satellite-based and ground-based data are analyzed and compared. We introduce the applications of the statistical technique of Weatherhead et al. (JGR, 1998) to better account for the noise and the uncertainties in both the surface and satellite estimates. The results show that, despite the daunting challenge the global mapping of radiation presents, the SRB products are in surprisingly good agreement with the measurements made at the Earth's surface not only in temporal means and trends at the majority of ground sites, but in correlations and frequency distributions as well.