Predicting sky types and zenith luminance from the cloud index derived from geostationary satellite data

Abstract

This paper presents methods to predict sky types and hourly zenith luminance from cloud index n derived from the visible channel data of the geostationary satellites. Three sky conditions based on the CIE standard general sky can be classified by cloud index: n > 0.38, 0.13 ≤ n ≤ 0.38 and n < 0.13, corresponding to overcast, partly cloudy and clear sky conditions, more specifically to sky 1, sky 8 and sky 13 of the ‘Hong Kong Representative Sky’. Hourly zenith luminance can be derived as a function of the combination of relative standard overcast sky zenith luminance and relative standard clear sky zenith luminance determined by cloud index. The mean bias error and the root mean square error of hourly zenith luminance are 0.49 and 3.37 kcd/m², respectively.