Two long-term solar UV measurement campaigns, in the USA and in Austria, using Robertson-Berger radiometers found opposite trends for measured levels of ultraviolet-B radiation reaching the earth's surface. This could be a consequence of the method used to calibrate the radiometers. Changes or differences in responsivity were compensated for by adjusting the sensitivity of the field radiometers to match the output of a reference radiometer.
This radiometer intercomparison procedure has been evaluated in terms of the normalization wavelength to which the Robertson-Berger effective irradiance refers. There are small differences in spectral responsivities apparent in the radiometers used in the USA campaign, which require the field radiometers to be normalized at different wavelengths to match the response of the reference radiometer. This normalization wavelength is shown to depend on the time-averaged spectrum experienced by the instruments during the intercomparison. As a result there are substantial interradiometer variations in their calculated response to solar radiation when the measured spectral distribution is different from the spectrum used for the radiometer intercomparison procedure.