Since July 1971, total ozone measurements have been performed at Uccle (50¿48'N, 4¿21'E) with the Dobson spectrophotometer 40. All the individual measurements were reevaluated after a careful homogenization, using the detailed calibration history of the instrument. Special care was taken to assure the consistency between direct Sun (DS) and zenith sky (ZS) measurements. The monthly mean values of percentage differences between the quasi-simultaneous direct Sun and zenith sky measurements show a weak modulation, probably induced by long-term changes in the mean vertical ozone distribution. The overall standard deviation of the percentage differences between the two types of observations amounts to 1.72%. From a detailed comparison of our revised data over the period 1971-1986 with the results of an approximate revision on a month-by-month basis (used in the Report of the International Ozone Trends Panel), it is concluded that the latter method of revision is not reliable.

The trend of the reduced thickness of the SO2 amount in the lower troposphere, as measured with the Brewer instrument 16 located at the same site as the Dobson spectrophotometer, is in agreement with the trend of the SO2 density near the ground which has been measured in the urban area of Brussels since 1968. From these two SO2 data sets the variation of the reduced SO2 amount since 1968 and the corresponding correction of the Dobson instrument data of Uccle were calculated on a month-by-month basis. Without the SO2 correction the revised Dobson data for the period from August 1971 through July 1991 show a statistically significant downward trend (¿ standard deviation) that amounts to -1.38¿0.50% per decade, while after the SO2 correction a nonsignificant trend of 0.31¿0.51% per decade is found. This means that the decrease of SO2 over the last two decades at Uccle has induced a fictitious Dobson total ozone trend of -1.69% per decade. At the subperiod for which reprocessed total ozone mapping spectrometer (TOMS) data are available (November 1978 to May 1990) our corrected Dobson data show a statistically significant downward trend (¿ standard deviation) of -2.93¿1.17% per decade which is consistent with the results from TOMS; this proves that analysis of total ozone data over time periods as short as one solar cycle does not allow to make conclusive statements about long-term trends. The spectacular decrease of SO2 densities is not a regional phenomenon but has been generally ascertained in European and North American urban areas; therefore it should not be neglected in any study about global ozone trends derived from Dobson data. ¿ American Geophysical Union 1992