FIGURE 12.34 Vortex-averaged total O, (DU) from TOMS satellite data for October at 75°S compared to model predictions using the assumption of a constant aerosol surface area or the measured surface areas (adapted from Portmann et al., 1996).

year compared to model calculations with the assumption of either a constant aerosol or the observed aerosol surface areas (Portmann et al., 1996). Clearly, the data are better fit by the model that includes variable aerosols due to the volcanic eruptions, which contributes significantly to continued ozone destruction during the spring by maintaining enhanced levels of active chlorine over a longer period of time. A similar conclusion has been reached by Shindell and de Zafra (f997). It should be noted that including the effects of not only the particle surface area but also temperature is important. Thus, Solomon et al. (1998) show that including temperature fluctuations in the model for midlatitudes improves the match of the model predictions and the observations, since lower temperatures enhance the net chlorine activation through the heterogeneous chemistry described in this chapter.

In short, it is predicted that the combination of high aerosol concentrations and low temperatures will lead to the development of particularly strong ozone holes.

Although decreases in total column ozone are anticipated due to this heterogeneous chemistry, there are also indirect effects of volcanic injections on stratospheric ozone and indeed, these may predominate under some conditions. For example, the volcanic aerosol can absorb both long-wavelength radiation emitted by the earth's surface and direct solar radiation, both of which lead to local heating in the stratosphere (although the former is the larger effect; see Chapter f4). This can cause upwelling of the aerosol layer, bringing air with lower ozone concentrations to higher altitudes that normally have higher 03 concentrations (e.g., see Schoeberl et al., 1993a).

On the other hand, a reduction in stratospheric ozone leads to less heating of the stratosphere; Zhao et al. (1996) suggest that this feedback could lead in the mid- and high-latitude lower stratosphere to sufficient net cooling that the breakup of the Arctic polar vortex (see Section C.7) could be delayed. Another indirect effect is the reduction in light intensity below the aerosol layer due to absorption and backscattering of sunlight, accompanied by an increase in actinic flux above the layer. The net effect is an increase in the photolysis rate of 03 above the layer, which is not countered by an increased rate of production via 02 photolysis. These effects are thought to be most important in the first 6 months to a year after an eruption, with the effects of heterogeneous chemistry predominating subsequently after the aerosol has been dispersed globally (e.g., see Kinne et al., 1989; Michelan-geli et al., 1992; Pitari et al., 1993; and Tie et al., 1994).

In addition to these indirect effects of volcanic emissions, there are a variety of nonvolcanic parameters that, of course, can change 03 as well, and these must be taken into account in assessing the role of the volcanic emissions alone. For example, there is a natural solar variability, part of which cycles on a time scale of about f f years and part of which is on a much longer time scale (Lean, 1991; Lean et al., 1995a, 1995b; Labitzke and van Loon, 1996). In addition, stratospheric ozone levels vary with the quasi-biennial oscillation (QBO), which is associated with a periodic variation in the zonal winds at the equator between 20 and

35 km (Garcia and Solomon, 1987; Chipperfield et al., 1994; WMO, 1995). The mean period for the QBO is about 27 months, but it can vary from 23 to 34 months (Zawodny and McCormick, 1991; WMO, 1995). For example, Chandra and Stolarksi (1991) point out that while a decrease in total ozone of 5-6% occurred in the winter following the El Chichon eruption, much of this could be due to the QBO and at most 2-4% could be attributed to the El Chichon emissions.

Long-term trends due to CFCs must also be removed from the data to examine the effects of volcanic emissions. Finally, one must take into account the possible contributions of air that has been processed through the polar vortices and of meteorological influences that are unique to certain locations (e.g., see Ansmann et al., 1996).

Despite these difficulties in quantifying the effects of volcanic emissions on stratospheric ozone and the uncertainties in the relative importance of direct versus indirect effects, there are ample data to support a decrease in stratospheric ozone due to volcanic emissions. Figure 12.35, for example, shows the ozone above Brazzaville, Congo, measured using electrochemical (ECC) sondes in the 16- to 28-km altitude range from 1990 to 1992 (Grant et al., 1992, 1994). Also shown is the expected ozone based on satellite measurements (SAGE II) from October 1984 to June 1991, corrected for normal cyclical variations and long-term trends. This "ozone climatology" can be compared to the measured ozone before the eruption of Mount Pinatubo as well as afterward. It is seen that before the eruption, the two are in reasonably good agreement, but afterward, the measured values lie some 15-33 Dobson units (DU) below the values expected based on the climatology, which is outside two standard deviations associated with the satellite data. The maximum decreases correspond to the loss of about 12% of the average total ozone. How much of this decrease is due to heterogeneous chemistry and how much is due to changes in heating, radiation, etc. are not known.

While the data in Fig. 12.35 are for a tropical site, similar data have been gathered at mid- and high-latitude locations as well (e.g., Gleason et al., 1993; McGee et al., 1994; Rodriguez et al., 1994; Hofmann et al., 1994b). For example, decreases of approximately 10% of the total column ozone over the Observatoire de Haute Provence in southern France were observed in July and August 1992, with decreased ozone observed at altitudes that overlapped those having increased volcanic aerosols (McGee et al., 1994).

A similar relationship was observed in Germany. Figure 12.36, for example, shows the deviation of the monthly mean ozone concentration after corrections for seasonal variations, long-term trends, the QBO and vortex effects, and the associated particle surface area concentration from 1991 to 1994 (Ansmann et al., 1996). The increase in the particle surface area due to Mount Pinatubo is clear; associated with this increase in aerosol particles are negative monthly mean deviations in ozone that persist until fall 1993, when the surface area approaches the preemption values. Similarly, the decrease in the total column ozone from f980-1982 to 1993 observed at Edmonton, Alberta, Canada, and shown at the beginning of this chapter in Fig. 12.1 has been attributed to the effects of the Mount Pinatubo eruption (Kerr et al., f 993).

7. Ozone Depletion in the Arctic

Given the dramatic decrease in stratospheric ozone in the Antarctic during spring, a similar phenomenon might be expected in the Arctic as well. However, it is now clear that while ozone depletion occurs in the

0 0

Post a comment