An 8 year (July 1983 through June 1991) time series of daily and monthly mean surface solar irradiance has been produced for the globe using data from the International Satellite Cloud Climatology Project (ISCCP) and a revised Bishop and Rossow  algorithm. We present a detailed validation analysis of the ISCCP solar irradiance fields with contemporaneous surface observations at buoys, at remote islands, and from the Global Energy Balance Archives (GEBA) [Ohmura et al., 1991]. The validation is hampered to some degree by the scale difference between the 280 km ISCCP product and the single-point measurements, some of which are affected by orographic clouds and other local meteorological effects. Our analysis suggests criteria for siting of island or coastal monitoring locations to minimize such biases. Particularly, eastward or poleward facing oceanic exposures are to be avoided. In addition, we suggest that deep sea buoys should be investigated for validation of oceanic surface fluxes. At open-ocean, clean-air sites, the ISCCP product is shown to be good to within 10 W m−2 in the monthly mean. The high-frequency (daily) systematics of solar irradiance variability at the open-ocean sites are also well duplicated by the ISCCP product. An identifiable error in the revised solar irradiance product is the neglect of spatially and temporally varying aerosol extinction. This error, when translated into an equivalent aerosol extinction coefficient, can be as large as 0.6 in known polluted and mineral dust-affected regions. We cannot determine additional satellite sensor calibration errors beyond those already corrected in the ISCCP processing. This uniquely long data set has been publicly available since 1994 at the National Center for Atmospheric Research. The data documents significant differences in solar fluxes received by the major oceans as well as significant flux variability on seasonal to interannual timescales.