On the Relation between Atmospheric Ozone and Sunspot Number

Abstract

Based on data from the Dobson network, between 1960 and 1987 there has been a zero-lag correlation of 0.48 between the 112 unsmoothed seasonal values of sunspot number and global total ozone, significant at the 1% level taking into account the considerable serial correlation in these data. The maximum correlation of 0.54 is found when sunspot number lags total ozone by two seasons, the result mainly of a phase difference early in the record. On the basis of only 2½ solar cycles, the global total ozone has increased by 1.4% for an increase in sunspot number of 100. The correlation between sunspot number and total ozone has been significant at the 5% level in north temperate and tropical zones—the zones with the most representative data. In the north temperate zone, the correlation between sunspot number and total ozone has been much higher in the west-wind phase of the 50 mb equatorial QBO than in the east-wind phase, but in the tropics the correlation has been much higher in the east-wind phase. Umkehr measurements between 1966 and 1987 in the north temperate zone indicate that the correlation between sunspot number and ozone amount has been higher (0.35, almost significant at the 5% level) in the low stratosphere where transport processes dominate than in the high stratosphere where photochemical processes dominate. During 1932–60 there was a significant correlation of 0.35 between sunspot number and Arosa total ozone 14 seasons later, very different from the nearly in-phase relation found after 1960. Considered is the possible impact of long-term change in transport processes in the low stratosphere on the total-ozone record at a single station such as Arosa. Between 1966 and 1985 there has been very good agreement between observed global total ozone, and global total ozone calculated from three 2-D stratospheric models that take into account the solar cycle, the time variation in trace gases, and nuclear tests; both observed and calculated variations are closely related to the variation in sunspot number. Between 1960 and 1966, however, the agreement between observation and calculation is poor, the models indicating a pronounced minimum in global total ozone in 1963 due to the nuclear tests of the early 1960s—a minimum not found in this analysis. The observed variation in global total ozone has been compared with the variation predicted by one of the models up to the sunspot maximum in 1990, and the agreement is shown to be good through the northern summer of 1988 if the impact of the QBO on global total ozone is taken into account. On the basis of the present analysis, there has been a 1.0 ± 0.9% decrease in global total ozone between solar cycles 20 and 21, a decrease 70% larger than that indicated by the three stratospheric models.