Proxy-based reconstruction of erythemal UV doses over Estonia for 1955–2004

Abstract. A proxy-based reconstruction of the erythemally-weighted UV doses for 1955-2004 has been performed for the Tartu-Tõravere Meteorological Station (58°16' N, 26°28' E, 70 m a.s.l.) site. The pyrheliometer-measured daily sum of direct irradiance on partly cloudy and clear days, and the pyranometer-measured daily sum of global irradiance on overcast days were used as the cloudiness influence related proxies. The TOMS ozone data have been used for detecting the daily deviations from the climatic value (averaged annual cycle). In 1998–2004, the biases between the measured and reconstructed daily doses in 55.5% of the cases were within ±10% and in 83.5% of the cases within ±20%, on average. In the summer half-year these amounts were 62% and 88%, respectively. In most years the results for longer intervals did not differ significantly, if no correction was made for the daily deviations of total ozone from its climatic value. The annual and summer half-yearly erythemal doses (contributing, on average, 89% of the annual value) agreed within ±2%, except for the years after major volcanic eruptions and one extremely fine weather year (2002). Using the daily relative sunshine duration as a proxy without detailed correction for atmospheric turbidity results in biases of 2–4% in the summer half-yearly dose in the years after major volcanic eruptions and a few other years of high atmospheric turbidity. The year-to-year variations of the summer half-yearly erythemal dose in 1955–2004 were found to be within 92–111% relative to their average value. Exclusion of eight extreme years reduces this range for the remaining to 95–105.5%. Due to the quasi-periodic alternation of wet and dry periods, the interval of cloudy summers 1976–1993 regularly manifests summer half-yearly erythemal dose values lower than the 1955–2004 average. Since 1996/1997 midwinters have been darker than on average.