Classification of the IDMP data of Kyoto into Kittler-Perez model of sky luminance distribution

The purpose of this paper is to validate Kittler-Perez model of classification of sky luminance distribution about the IDMP data in Kyoto Japan. Data were classified by the nearest centroid sorting method on the combination of nine indices of insolation condition. The indices are, 1) cloud ratio, 2) sky clearness index, 3) ratio of zenith luminance to diffuse illuminance, 4) standardized diffuse illuminance, 5) brightness, 6) standardized global illuminance, 7) standardized direct illuminance, 8) permeability, and 9) turbidity. Data measured every minute for a year over five degree of solar altitude were used to calculate these indices. Parameters of the types of K-P model in the relationship between solar altitude and ratio of zenith luminance to diffuse illuminance were estimated by non-linear regression. Classified cases were matched to the nearest types of the model by parameter comparison. Almost classified groups could be matched to the types of K-P model except for the type of turbid sky condition and the types of non-homogeneous gradation. This fact shows that the classification of the model is valid and the method of sorting classification works effective to some extent. But it is notable that the types of non-homogeneous gradation were found or could not be separated for the data of Kyoto.