Monitoring of the terrestrial solar ultraviolet irradiance by using a radiometer is often considered as expensive
and laborious or the data collected as insufficient in spatial coverage and in some cases in its temporal
resolution, too. Therefore, alternative methods, all relying on modelling in one way or the other, have been
developed. They differ in which input they receive, either standard meteorological information, space-based
radiance measurements or ground-based irradiances from broadband or multiband UV radiometer or from
pyranometer. A comparison of performance is presented between three methods during a 15-month period.
The ground reference instrument is the Brewer Mk-III #107 spectroradiometer of the Observatory of
Jokioinen, Finland. Compared to the reference, the space-based method overestimates the UV irradiance at
noon by 14.6% and the pyranometer-based by 0.9% with root-mean-square differences of 35.5% and 10.4%,
respectively. Daily erythemal doses agree by 3.8% for the space-based and 0.4% for the pyranometer-based
method with a scatter of 16.5% and 4.6%, respectively. Spectral irradiances generated by the pyranometerbased
model agree within 0.4% on average with a standard deviation of 17%. A rough estimate on the cost of
each approach suggests that none of them is clearly superior to the others and the actual nature of the data
needed may be used in decision making concerning monitoring strategies.
The objectives of the COST action 726 are to establish long-term changes of UV-radiation in the past, which can only be derived by modelling with good and available proxy data. To find the best available models and input data, 16 models have been tested by modelling daily doses for two years of data measured at four stations distributed over Europe. The modelled data have been compared with the measured data, using different statistical methods. Models that use Cloud Modification Factors for the UV spectral range, derived from co-located measured global irradiance, give the best results.
Solar spectral UV-monitoring data for 8 European sites with 5-10 years of data, and covering a latitudinal range from 41 degrees North to 67 North have been re-evaluated and resubmitted to the European UV-database (EUVDB) in Finland as part of the EU-project SCOUT-O3. All resubmitted spectra (420000) were quality checked, flagged, and corrected with respect to wavelength scale errors and spectral distortions using the SHICrivm software package. Additional data products provided by the software are standardized spectra, spectral atmospheric transmissions, and biologically weighted UV-irradiances for a wide variety of biological action spectra. The resubmitted spectra were shown to have improved based on the EUVDB quality flagging criteria. Spectral and effective irradiances were integrated and summed in a standardized way to obtain daily, monthly, and seasonal UV-doses. The measured summer sums varies from 478 kJ/m<sup>2</sup> for Thessaloniki to 228 kJ/m<sup>2</sup> for Sodankyla. Clouds reduced the exposure during summer time by 30% on average, in Bilthoven this was 35%, while in Thessaloniki only 17% was reduced. Using co-located ozone and pyranometer measurements results of a generic UV-modelling approach, derived in a specific low albedo and low surface elevation environment, are systematically compared to the UV-doses obtained for all sites. Generally, a good agreement is found, measured and modelled total UV-doses agree within a few percent with a standard deviation of 15 typically. Deviations with respect to the application in a high surface albedo and high altitude environment have been identified and handles to improve the modelling have been assigned.