Standard skies for maritime climates

Abstract

A set of 15 standard skies proposed by Kittler, Darula and Perez was compared with a large sample of measured sky luminance distributions from Singapore, Japan and the United Kingdom. It was found that (a) the standard set gives a good overall framework for categorising actual skies; (b) subsets of about four luminance distributions were adequate to describe the skies that occurred at each site; (c) inhomogeneity of actual skies in maritime climates leads to uncertainty in interior daylight predictions (typically the room illuminance computed on the best-fitting standard sky might have a variance such that 80% of calculated values lie between 0.8 and 1.3 of the real value); and (d) this error range is significantly less than when calculations are based on an overcast sky alone.