Solar irradiance modelling using satellite retrieved cloudiness : A pilot study

Accurate measurements of incoming solar radiation at the Earth' s surface are made by present ground based networks. The data are of high quality and high sampling frequency but there is a lack of spatial coverage. Areas of great interest may not have any observations at all. Since clouds are the major modulators of solar radiation, the possibility to use satellite retrieved cloudiness as a basis for irradiance calculations has been investigated. Data from the cloud classification model SCANDIA, which utilises NOAA AVHRR data as input, have been analysed and the influence of SCANDIA classed clouds on incoming solar radiation at the ground has been approximately quantified. Hourly mean values of global irradiance, measured at the twelve stations in the solar radiation network of SMHI, were taken as reference data. Derived cloud transmittances are relatively close to those found by others using manually observed and classified clouds. In this study satellite data from only one month (April 1995) have been analysed. To determine mean cloud transmittances of high accuracy far more data (from >1 year) have to be used. In spite of the division into subgroups of cloud classes made by SCANDIA, it is shown that individual cloud transmittances for one and the same cloud class scatter a lot. Therefore it is difficult to accurately calculate solar irradiances, representative for short time intervals (>1 hour), with only cloud transmittances (and solar elevation) as input data. Though, for modelling accumulated solar irradiance during longer time intervals (days or months), satellite derived cloudiness is thought to be very valuable information. It is also shown that for accurate calculations of global irradiance under clear skies, additional information about atmospheric turbidity and water vapour content is urgently needed