Abstract
Irradiance data obtained over a long period at Vancouver and Toronto, Canada, and covering a range of slope orientations are used to validate four models that estimate either the direct or diffuse solar irradiances for inclined surfaces. Evaluations are initially performed for daily and hourly time integrals. A simple parametrization of the diffuse sky radiance dramatically improves estimates of the diffuse irradiance. Both of the direct irradiance models have difficulty accommodating the diurnal characteristics of the irradiance, and consequently modelling errors are substantial for slopes not directly facing the equator. For equator‐facing slopes a saving in data requirements and computational effort through the use of daily integrals can be achieved with little additional error. A substantial portion of the differences between the measured and estimated irradiances is non‐systematic in nature and is therefore reduced through temporal averaging.