Estimation of Ultraviolet-B Irradiance under Variable Cloud Conditions

Abstract

Methods to estimate the irradiance of ultraviolet-B (UVB; 280?320 nm) radiation are needed to assess biological effects of changes in atmospheric composition. Measurements of the spatial distribution of sky cloud cover, temporal variability of photon flux density of photosynthetically active radiation (PAR; 400?700 nm), and UVB irradiance (I-UVB) were made on 23 days during the summer of 1993 in a rural area (West Lafayette, Indiana). Prediction equations for the measured UVB irradiance under partly cloudy skies were developed based on the photosynthetically active photon flux density (PPFD), cloud cover fraction, probability of cloud obstruction of the sun, and a semiempirical combination of cloud probability and cloud cover. The I-UVB was linearly related to the PPFD, with the variability in PPFD accounting for 77% of the I-UVB variability. Normalized PPFD (PAR F) and I-UVB (UVB F) values, calculated by dividing the observed value by the expected cloud-free sky PPFD and I-UVB, were also linearly related. Linear models based on either the spatial cloud fraction or the probability of clouds blocking the sun accounted for less than 30% of the UVB F variability. A two-component semiempirical model was developed to predict UVB F based on cloud-cover fraction, probability of sun obscuration by clouds, the predicted cloud-free sky diffuse fraction, and solar zenith angle. This model accounted for 60% of the variability in UVB F. Results indicate the best estimation of I-UVB under the partly cloudy sky conditions is made using PPFD measurements and referencing the measured PPFD to cloud-free sky PPFD at the same sun angle. Alternative approaches, such as the developed two-component model, should be used only if on-site PPFD measurements are unavailable.