Physics interpretation of ISO/CIE sky types

Abstract Sky luminance distribution, itself immensely variable solar radiation product, is an indispensable prerequisite for any daylight availability assessment. Currently, prediction-based modeling is a key ingredient of daylight qualification worldwide, while ISO/CIE general sky standard is an essential tool accommodated for simulation purposes. Due to its long term perspective and a number of benefits (e.g. simple mathematics and easy numerical implementation), the model is applied routinely and independent of meteorological conditions prevailing on a regional scale. ISO/CIE classifies sky states into fifteen subtypes on statistical base, however, with no relation to e.g. local sources of pollution. In result, the statistics found for two different territories may accidentally cohere for which there is really no theoretical reason. Here we show that the physics interpretation of sky states is needed for a proper use of the ISO/CIE model in different territories through linking the luminance distributions with atmospheric properties, specifically τ and g . The latter shapes the angular distribution of scattered light, while τ is interpreted in term of atmospheric optical thickness. The new approach we present here not only profits from its theoretical foundation, but both τ and g can be easily derived from the momentary local atmosphere state. This way the characterization of sky states and a use of ISO/CIE skies can be more selective and more systematic.