Solar ultraviolet radiation risk estimates—a comparison of different action spectra and detector responsivities

Abstract

Studies assessing the dose–response relationship for human skin cancer induction by solar ultraviolet radiation (UVR) apply a range of methods to quantify relevant UVR doses, but information about the comparability of these datasets is scarce. We compared biologically weighted effectivities applying the most relevant UVR action spectra in order to test the ability of certain UVR detectors to mimic these biological effects at different times during the day and year. Our calculations were based on solar spectra measured at Dortmund, Germany (51.5◦ N) and at Townsville, Australia (19.3◦ S), or computed for latitudes 20◦ S and 50◦ N. Convolutions with the CIE action spectra for erythema and non-melanoma skin cancer (NMSC) and with ICNIRP’s weighting function showed comparable solar zenith angle (SZA) dependences with little influence of season or latitude. A different SZA dependence was found with Setlow’s action spectrum for melanoma induction. Calculations for a number of UVR detector responsivities gave widely discrepant absolute irradiances and doses, which were nevertheless related to those calculated with both CIE spectra by correction factors largely independent of the SZA. Commonly used detectors can thus provide quite accurate estimates of NMSC induction by solar UVR, whereas they may be inadequate to mimic melanoma induction.