Open AccessA method for reconstruction of past UV radiation based on radiative transfer modeling: Applied to four stations in northern Europe
Author(s) -
Lindfors Anders,
Kaurola Jussi,
Arola Antti,
Koskela Tapani,
Lakkala Kaisa,
Josefsson Weine,
Olseth Jan Asle,
Johnsen Bjørn
Publication year - 2007
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/2007jd008454
Subject(s) - radiative transfer , albedo (alchemy) , environmental science , irradiance , atmospheric radiative transfer codes , atmospheric sciences , radiation , altitude (triangle) , water vapor , cloud cover , relative humidity , snow , meteorology , remote sensing , geology , physics , optics , mathematics , cloud computing , geometry , art , performance art , computer science , art history , operating system
A method for reconstruction of past UV radiation has been developed. The idea of the method is to use measurements of global radiation (300–3000 nm) for determining the influence of clouds on UV radiation. In order to transfer the information contained in the global radiation data into a cloud effect in the UV range, a so‐called cloud modification table was developed, which is based on physical relationships determined through radiative transfer calculations. The method was given as input the measured global radiation and total ozone column, the total water vapor column from the ERA‐40 data set, the surface albedo as estimated from snow depth, and the altitude of the location. Using this method, erythemally weighted UV irradiances were reconstructed back to the early 1980s at four stations in northern Europe: Bergen in Norway, Norrköping in Sweden, and Jokioinen and Sodankylä in Finland. The reconstructed daily UV doses are in good agreement with measurements. For the summer season, the systematic error was found to vary between 0% at Bergen and 4% at Jokioinen, and the correlation coefficient was 0.99 at all stations. The summer root‐mean‐square error was 5% at all stations except Jokioinen, where it was 9%. The method performs well also for spring and autumn, whereas for winter conditions of low Sun, a systematical underestimation was found. A large part of this underestimation was found to be due to the plane‐parallel approximation used in the radiative transfer calculations. The time series of reconstructed UV exhibit a clear increase since the early 1980s at both Sodankylä (4.1%/decade; statistically significant) and Norrköping (3.3%/decade; not significant). At Jokioinen, a weak increase was found, while at Bergen there was no considerable overall change. At both Sodankylä and Norrköping, the increase in the reconstructed UV radiation was primarily driven by an increase in the global radiation, that is, by decreased cloudiness. The method is general in the sense that it can be applied also to other stations.