A method to estimate erythemal UV from total solar irradiance measurements based on 9 years of 1-minute data at Lauder, New Zealand

Abstract

The radiative effect of clouds on the incident surface solar radiation highly depends on the spectral band of the solar spectrum that is considered. In particular, cloud effects are known to be different for the broadband erythemal ultraviolet (hereafter, UVE) irradiance and the total global solar irradiance (TR) due to the much more important molecular (Rayleigh) scattering on the UV than on the visible and near infrared wavelengths. In this work we investigate these differences by analyzing 9 years (2000-2008) of 1-minute UVE and TR measured at Lauder (45.04S, 169.68E, 370m asl), New Zealand. Clear sky models for UVE and TR are considered and their performances are tested. Effective cloud transmittance, also known as cloud modification factor (CMF) is calculated as the ratio between the 1-minute measurements and the clear sky estimation from modeling, both for UVE (CMF(UV)) and TR (CMF(TR)). The two CMF are then compared. The analyses are undertaken as a function of solar zenith angle (SZA), and Sun visibility derived from a Total Sky Imager (TSI) device. Differences between CMF(UV) and CMF(TR) are maximum for intermediate values of the latter, being CMF(UV) always greater than CMF(TR) (that is, UVE less affected by clouds). A methodology to calculate CMF(UV) as a function of CMF(TR) and SZA is proposed, and subsequently validated using 2 years of independent data (2009-2010). The agreement between estimated and measured UVE is remarkably good: absolute differences are less than 0.5 UVI units (being the greatest at lower SZA), and relative differences are less than 10% (the greatest at higher SZA), despite of a systematic error that needs to be corrected