Monte Carlo-based uncertainties of surface UV estimates from models and from spectroradiometers

Although some of the adverse effects of the ultraviolet (UV) radiation may be strictly proportional to cumulative UV doses, others may relate to the frequency of extreme UV events. Therefore, an improved understanding of the UV global climate, including variability and trends, has become of great interest. Variability and trend analyses require quality-ensured surface UV series. The quality of surface UV data depends on their uncertainty. Building upon our prior efforts, we have used a Monte Carlo-based method to compute, under different conditions, the uncertainties affecting UV data rendered by models (1D radiative transfer models) and by spectroradiometers (double monochromator-based and CCD array-based). We found that the uncertainty of spectral UV measurements is driven by the signal-to-noise ratio of the detector, while the uncertainty of spectral UV calculations strongly depends on the uncertainty of the ozone input. The presented uncertainty figures allow comparison of the performance of modern UV gathering techniques (models and instruments), and provide a frame to assess the significance of differences when intercomparing.