Long-term comparison of filter method and sensitive analyser in monitoring of sulphur dioxide

Abstract Co-located long-term routine measurements of sulphur dioxide with the impregnated filter method and a pulsed fluorescence analyser at four Finnish background stations were compared. The data consist of 9 years of parallel measurements at the Virolahti station and 8, 7 and 2 years in Uto, Oulanka and Ahtari, respectively, between the years 1993–2001. During the whole study period, the daily SO 2 concentrations measured with the analyser and the filter method showed excellent agreement in Virolahti (number of daily data pairs 2748, r 2 = 0.99 ) and Uto ( n = 2328 , r 2 = 0.97 ), and no systematic relative error between the two methods was found. Also in Oulanka ( n = 2131 , r 2 = 0.99 ) and Ahtari ( n = 643 , r 2 = 0.97 ) the linear model described the data well. However, in Oulanka the analyser produced 16% and in Ahtari 10% higher results than the filter method. The difference would decrease to 12% in Oulanka and to 8% in Ahtari if the concentrations were given at the same reference pressure. At all stations also a small constant error (analyser>filter) of the order of 0.03–0.06 μg S m −3 was discerned. The role of nitrogen oxide, volatile organic compounds and water vapour interferences as well as extreme temperatures were studied but no clear systematic influence by them on the daily differences between the two methods was detected.