Abstract. Results from an intercomparison campaign of ultraviolet spectroradiometers that was organized at Nea Michaniona, Greece July, 1-13 1997, are presented. Nineteen instrument systems from 15 different countries took part and provided spectra of global solar UV irradiance for two consecutive days from sunrise to sunset every half hour. No data exchange was allowed between participants in order to achieve absolutely independent results among the instruments. The data analysis procedure included the determination of wavelength shifts and the application of suitable corrections to the measured spectra, their standardization to common spectral resolution of 1 nm full width at half maximum and the application of cosine corrections. Reference spectra were calculated for each observational time, derived for a set of instruments which were objectively selected and used as comparison norms for the assessment of the relative agreement among the various instruments. With regard to the absolute irradiance measurements, the range of the deviations from the reference for all spectra was within ñ20%. About half of the instruments agreed to within ñ5%, while only three fell outside the ñ 10% agreement limit. As for the accuracy of the wavelength registration of the recorded spectra, for most of the spectroradiometers (14) the calculated wavelength shifts were smaller than 0.2 nm. The overall outcome of the campaign was very encouraging, as it was proven that the agreement among the majority of the instruments was good and comparable to the commonly accepted uncertainties of spectral UV measurements. In addition, many of the instruments provided consistent results relative to at least the previous two intercomparison campaigns, held in 1995 in Ispra, Italy and in 1993 in Garmisch-Partenkirchen, Germany. As a result of this series of intercomparison campaigns, several of the currently operating spectroradiometers operating may be regarded as a core group of instruments, which with the employment of proper operational procedures are capable of providing quality spectral solar UV measurements.

The anamnestic skin phototypes (ASP) I-IV of 22 Caucasian volunteers wee compared with their phototested skin phototypes (PSP) using solar simulating, broadband UV radiation. The Commission Internationale de l'Eclairage (CIE)-weighted (i.e. erythemally effective) minimal erythema doses (MED) for solar simulating radiation varied from 20 mJ/cm2 (PSP type 1) to 57 mJ/cm2 (PSP type 4). In only 11 of 21 volunteers did the ASP (I-IV) and PSP (1-4) classifications coincide, and the MED values of the volunteers within the different ASP groups (I-IV) overlapped considerably. To compare the reactivity to erythematogenic radiation of different wavelengths, narrowband monochromator irradiations were performed at 298 nm, 310 nm and 330 nm. The CIE-weighted MED values at these wavelengths (20-80 mJ/cm2) corresponded well with those obtained in the broadband testing. Our results indicate that, with classification by interrogation, Caucasian skin can reliably be classified into only two subtypes, corresponding to Fitzpatrick phototypes I-III and phototype IV, respectively. A classification into four sensitivity types can be achieved by phototesting, only. We propose that the concept of ASP should be used with caution. The concept of PSP 1-4 should be favored.

No abstract

The first international intercomparison of erythemally weighted (EW) broadband radiometers was arranged in 1995 to improve the accuracy and comparability of the measurements carried out by solar UV monitoring networks. The intercomparison was arranged at the Radiation and Nuclear Safety Authority in Helsinki, Finland, in cooperation with the University of Innsbruck and with support from the World Meteorological Organization. Altogether 20 EW meters of six different types from 16 countries were (1) tested in the laboratory by measuring the spectral and angular responsivities and (2) calibrated in solar radiation against two reference spectroradiometers. Calibration factors (CFs) for the EW meters were determined by using simultaneously measured EW solar UV spectra as a calibration reference. The CFs averaged over solar elevations higher than 35° varied from 0.87 to 1.75, with the estimated uncertainty being ±10%. As a result of this intercomparison, for the first time the calibrations of more than 100 EW radiometers around the world are possible to trace to the same origin. The present experience indicates that the accuracy of temperature‐controlled EW radiometers is not significantly lower than the accuracy of spectroradiometers provided that strict quality assurance/quality control procedures are followed.

The increase in the UV exposure of the Finnish population associated with the combined effects of ozone depletion and snow reflection was studied with the aid of theoretical calculations based on Green's clear sky UV model. A simple formula was utilized to transform horizontal irradiances to vertical irradiances averaged over 360 degrees azimuth angle. The model was verified with spectral and broadband measurements. The difference between the theoretical and measured UV radiation falling to horizontal surfaces was in most cases less than +/- 10%, and the additional error to theoretical vertical irradiances was less than +/- 10%. The calculations show that the annual horizontal doses in Helsinki (60.2 degrees N, 25 degrees E) are about 35% higher than in Saariselkä (68.4 degrees N, 27.5 degrees E), but the difference is only 16% for vertical doses owing to the stronger contribution to vertical (facial) surfaces of the reflection of UV from snow. At Saariselkä, the maximum vertical irradiance at the end of April approaches the midsummer values. The ozone depletions up to 40% in February and March 1992 had no significant effect on the annual doses because the total ozone returned to normal before the UV increased to biologically significant levels.