JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 100, NO. D2, PAGES 3023-3032, FEBRUARY 20, 1995 Total ozone and synoptic weather relationshipsover southern Africa and surrounding oceans J. Barsby and R.D. Diab Department of Geographical and Environmental Sciences, University of Natal, Durban, SouthAfrica Abstract. Relationships between Total OzoneMappingSpectrometer (TOMS) total ozone and synoptic weather systems for a region bounded by longitudes 10øW-50øE and latitudes 0ø-50øS are presented. Least squares regression analysis revealed a generallypoor but negative relationship between total ozoneand the heights of the 500, 300, and 100 hPa geopotential surfaces at nine selected stations for the period 1987 to 1988. The relationship strengthens with increasing latitudeand suggests a physical link with the midlatitude cyclone which facilitates the injection of strato- spheric ozonerich air into the troposphere. In the tropics the link with middleto upper tropospheri c meteorological parameters is weak. Analysis of daily relationships betweengridded TOMS ozoneand griddedEuropean Centre for Medium-Range Weather Forecasts (ECMWF) geopotential height data over the study area showed considerable variations, indicative of varyingsynoptic weather situations. Casestudies of a cutoff low sequence and a subtropical anticyclone system were selected as examples of synoptic situations which accounted for the highest and lowestcorrela- tions, respectively. The contrasting dynamics operating in these systems are high- lighted. Introduction Dobson et al. [1929] were the first to notice the relation- ship between day-to-day variations in totalozone andsurface weather. They found that maximum positivedeviations of daily values from monthly means (high ozone values) generallyoccurred to the rear of surface low-pressure sys- tems (west of cyclonic centers), while maximum negative deviations (low ozone values) were found to the rear of surface anticyclones. Later, Reed[1950]proposed a model to describe the relationship between totalozone distribution and weather patterns for the northern hemisphere. His suggestion was that at any one time, ozone distributionwould result from the effects of horizontal advection or vertical motion or a combination of both in association with a baroclinic wave. Since total ozone amountsincreasetoward the poles, advect•ion of airfrom the north into the trough would be expected to cause an increase of ozone,while advection from the south would result in a decrease in ozone amount. In addition, subsiding motion associated with the upper level trough would cause an increase in ozone, whereas upward displacements in association with an upperair ridge would lead to decreasing ozone amounts. The combined effect of verticalmotionand horizontal advection produces positive deviations in upper level troughs and negative deviations in upper levelridges. Since in a baroclinic atmosphere theupper air systemis displaced to the rear or west of its surface counterpart, positivedeviations should occur to the rear of the surface cyclone andnegative deviations to the rear of the surface anticyclone. Copyright 1995 by the American Geophysical Union. Papernumber 94JD01987. 0148-0227/95/94JD-01987505.00 The relationship between ozone amounts and day-to-day weather parameters is well documented in the literature [e.g., Meetham, 1937; Fritz and Stevens, 1950; Normand, 1953; Ohring a•wl Muench,1960;Orlanski et al. , 1989; Wakamatsu et al., 1989]. Much attention hasbeen devoted to tropopause foldingwhichoccurs in conjunction with intense upper level frontogenesis [Reed, 1955;Danielsen,1968; Danielsen and Mohnen, 1977] andwhichis accompanied by the intrusion of stratospheric air, rich in potential vorticity and ozone,into the mirIdle and lower troposphere [Shapiro, 1980]. Most studies relating ozone andweather parameters have based the explanation for the link on the midlatitude cyclone model. Nonehaveexplored the relationship in tropical or subtropical regions. South Africa, situated astride the subtropics between latitudes 22øS and 35øS, is ideally located to examine the influences of both midlatitude and tropicalweather disturb- ances on total ozone. Data Version6 TOMS (totalozone mapping spectrometer) data were obtained on CD-ROM from the National Space Science Data Centerat the Goddard Space FlightCenter in Maryland. The dataused here extended over the 2-yearperiod, 1987to 1988. Gridded TOMS data,with a grid spacing of 5o, for an area bounded by longitudes 10øW-50øE and latitudes 0ø-50øS and data for nine individual stations were used in this investigation. The stations are Marion Island (46ø55'S;37ø45'E), Gough Island (40ø20'S;10øW), Cape Town(33ø56'S; 18 ø28'E), Port Elizabeth (33ø58'S;25 ø36'E), Durban (29ø53'S;31øE), Bloemfontein (29ø07'S;26ø14'E), Pretoria (25ø45'S;28ø12'E),Harare(17ø43'S;31 ø05'E), and Nairobi (1ø17'S;36ø50'E). The locations of the stations and the studyarea are shownin Figure 1. 3023