Reduction of solar UV by clouds: A comparison between satellite-derived cloud effects and ground-based radiation measurements

Assessment of the effects of ozone depletion on biologically effective solar UV at ground level has been greatly advanced through the use of remote sensing data. Satellite data on atmospheric properties allow the construction of geographically distributed surface UV radiation maps based on radiative transfer calculations. In this respect, clouds play a dominant but rather complex role. We compared the reduction of daily UV doses due to clouds, as derived from satellite cloud data, with the reduction derived from routine ground-based measurements of global solar radiation (i.e., broadband total solar irradiances with wavelengths between 0.3 and 2.8 mu m). An empirical relationship is used to link the reduction due to clouds of global solar radiation and UV radiation. The abundance of global solar radiation measurements (data from over 125 stations in 30 satellite grid cells) for the European region ensured a sound basis for the data analysis for the period considered (May, June, and July of 1990, 1991, and 1992). Approximately 6500 daily UV-reduction factors, defined as the ratio of daily UV doses calculated with and without clouds, were thus obtained applying both methods. The daily UV-reduction factors (and 10-day averaged UV reduction factors) from the two independent sources correlated well, with r(2) = 0.83 (r(2) = 0.89), and had a standard deviation of 0.06 (0.03). Over 90% of the satellite-derived results agreed within a range of +/-0.14 (+/-0.07) with the ground-based measurement-derived results. We evaluated sources of uncertainty related to spatial and temporal resolution, and optical properties, and estimated their consequences and range. Among these different sources the largest uncertainties are caused by the sampling error, i.e., grid-cell average versus station average, which is on average 0.10 for daily UV-reduction factors. Information on the atmospheric optical properties during the measurements may reduce the stated range of uncertainty from +/-0.14 to +/-0.07. The variation of the measurements from station to station is then the limiting factor. We concluded that the reduction of daily UV based on satellite-derived cloud cover and cloud optical thickness relates well with the UV reduction due to clouds derived from ground-based global solar radiation measurements.