Abstract Daily values of solar global ultraviolet (UV) B and UVA irradiation as well as erythemal irradiation have been parameterized to be estimated from pyranometer measurements of daily global and diffuse irradiation as well as from atmospheric column ozone. Data recorded at the Meteorological Observatory Potsdam (52°N, 107 m asl) in Germany over the time period 1997–2000 have been used to derive sets of regression coefficients. The validation of the method against independent data sets of measured UV irradiation shows that the parameterization provides a gain of information for UVB, UVA and erythemal irradiation referring to their averages. A comparison between parameterized daily UV irradiation and independent values of UV irradiation measured at a mountain station in southern Germany (Meteorological Observatory Hohenpeissenberg at 48°N, 977 m asl) indicates that the parameterization also holds even under completely different climatic conditions. On a long-term average (1953–2000), parameterized annual UV irradiation values are 15% and 21% higher for UVA and UVB, respectively, at Hohenpeissenberg than they are at Potsdam. Daily global and diffuse irradiation measured at 28 weather stations of the Deutscher Wetterdienst German Radiation Network and grid values of column ozone from the EPTOMS satellite experiment served as inputs to calculate the estimates of the spatial distribution of daily and annual values of UV irradiation across Germany. Using daily values of global and diffuse irradiation recorded at Potsdam since 1937 as well as atmospheric column ozone measured since 1964 at the same site, estimates of daily and annual UV irradiation have been derived for this site over the period from 1937 through 2000, which include the effects of changes in cloudiness, in aerosols and, at least for the period of ozone measurements from 1964 to 2000, in atmospheric ozone. It is shown that the extremely low ozone values observed mainly after the eruption of Mt. Pinatubo in 1991 have substantially enhanced UVB irradiation in the first half of the 1990s. According to the measurements and calculations, the nonlinear long-term changes observed between 1968 and 2000 amount to +4%, …, +5% for annual global irradiation and UVA irradiation mainly because of changing cloudiness and +14%, …, +15% for UVB and erythemal irradiation because of both changing cloudiness and decreasing column ozone. At the mountain site, Hohenpeissenberg, measured global irradiation and parameterized UVA irradiation decreased during the same time period by −3%, …, −4%, probably because of the enhanced occurrence and increasing optical thickness of clouds, whereas UVB and erythemal irradiation derived by the parameterization have increased by +3%, …, +4% because of the combined effect of clouds and decreasing ozone. The parameterizations described here should be applicable to other regions with similar atmospheric and geographic conditions, whereas for regions with significantly different climatic conditions, such as high mountainous areas and arctic or tropical regions, the representativeness of the regression coefficients would have to be approved. It is emphasized here that parameterizations, as the one described in this article, cannot replace measurements of solar UV radiation, but they can use existing measurements of solar global and diffuse radiation as well as data on atmospheric ozone to provide estimates of UV irradiation in regions and over time periods for which UV measurements are not available.