1991 1992 1993 1994 1995 1996 Year

FIGURE 13.6 Fits to monthly mean tropospheric concentrations of some CFC alternates in the Northern (-) and Southern

(---) Hemispheres (adapted from Montzka et al., 1996a).

1992 1993 1994 1995 1996 Year

FIGURE 13.7 Effective total tropospheric concentration of chlorine from halocarbons from 1992 to 1996 in the Northern (-)

and Southern (---) Hemispheres (adapted from Montzka et al,


air parcels (e.g., see Patra et al., 1997; and Harnisch et al., 1998). For example, CF4 and C2Fh are generated during aluminum production, and SF6 during its use in electrical insulation and switching and as an inert gas for handling reactive molten metals such as aluminum. The tropospheric concentrations of all of these compounds have been increasing rather dramatically over the past several decades (Harnisch et al., 1996a; Maiss and Levin, 1994; Maiss et al., 1996; Maiss and Bren-ninkmeijer, 1998). For example, the growth rate for SFft in early 1996 was measured to be 6.9 ± 0.2% (0.24 ± 0.01 ppt) per year on top of a global average mixing ratio of 3.4 ppt (Geller et al., 1997). Similarly, Maiss and Brenninkmeijer (1998) showed that SF6 increased from 0.24 ppt in 1970 to 3.8 ppt in 1996; furthermore, the observed increase in its atmospheric concentrations was best fit by a function with a squared dependence on time.

For CF4, there is some evidence (Harnisch et al., 1996b) that the observed increase has occurred on top of a "background" due to natural emissions whose sources are as yet not well understood. Both CF4 and SFft have been found in natural fluorites (CaF2) and hence degassing of these compounds from the earth's surface may be at least one natural source of these compounds, particularly CF4 (Harnisch et al., 1998).

Although not fully fluorinated, HFC-23 (CHF3) also has a long lifetime (~250 years; WMO, 1995) because its reaction with OH is slow, k2™ = 2.8 x 10"16 cm3 molecule"1 s"1 (DeMore et al., 1997). A major source is the production of HCFC-22 (CHC1F2), where HFC-23 is a byproduct. As the use of HCFC-22 and its atmospheric levels have increased (Fig. 13.6), the levels of HFC-23 would be expected to increase as well, and indeed, this is the case. Oram et al. (1998) reported the first atmospheric measurements of this compound at Cape Grim, Australia. HFC-23 increased from ~ 2 ppt in 1978 to 11 ppt in 1995, with a growth rate of 5% per year in 1995.

In short, the trends in the tropospheric concentrations of CFCs, halons, and their substitutes follow trends in their emissions. The effects of the controls imposed by the Montreal Protocol and its subsequent amendments are evident in the trends and have been used to show that the associated impact on ozone destruction is expected to begin about the turn of the century. The following section briefly describes the observed trends in stratospheric ozone.

2. Trends in Stratospheric 03

As discussed in Chapter 12, trends in stratospheric ozone in the Antarctic spring during formation of the "ozone hole" are clear. However, as treated in detail in that chapter, there are unique meteorological and chemical circumstances that are responsible for the dramatic loss of ozone in the Antarctic spring. It is perhaps of even greater interest as to whether there is evidence for a trend in midlatitude stratospheric ozone, since this is where the bulk of the population resides.

Detecting and quantifying ozone trends in midlati-tudes from anthropogenic perturbations is complex due to the effects of natural variations in stratospheric ozone and to the interactions between various effects (e.g., Krzyscin, f994; Brasseur et al., 1995; Callis et al., 1997; Zerefos et al., f997; Hood, 1997; Callis et al.,

1997). Thus, long-term trends in ozone must be extracted from variability due to the solar cycle, which has an 11-year period associated with it, as well as the quasi-biennial cycle (QBO), which is an oscillation of zonal winds in the stratosphere around the equator and which has a 26-30 month cycle (e.g., see Kane et al.,

1998). For example, Bjarnason et al. (1993) examined column 03 measurements made at Reykjavik from 1957 to 1990 using a Dobson spectrometer and applied a stratospheric model that included variations due to seasons, the solar cycle, the QBO, and a linear trend. The combination of the data and model showed a variation of 3.5 + 0.8% in column 03 over a solar cycle and 2.1 + 0.6% over a QBO, on top of a linear trend of decreasing 03.

In addition, there is an observed correlation between total column ozone and the El Niño Southern Oscillation (ENSO) in the tropical troposphere, with decreases in total ozone in middle and sometimes polar latitudes following the ENSO by several months; the period associated with the ENSO is ~43 months (Zerefos et al., 1992). While the association between the ENSO and ozone is not well understood, it has been proposed that the warming of the troposphere in the tropics over the Pacific Ocean causes increases in the upper troposphere air temperatures and tropopause height and an upwelling in the lower stratosphere. If sufficiently large, this could have more widespread impact than just in the tropics (e.g., see Zerefos et al., f992; and Kalicharran et al., 1993).

There is also a significant correlation between temperature fluctuations in the lower stratosphere and fluctuations in total ozone. There are two sources of this correlation, radiative and dynamical (McCormack and Hood, 1994). Thus, increased ozone leads to increased absorption of solar radiation and increased heating. In addition, dynamical effects associated with vertical and meridional air motions also give a positive correlation between ozone and stratospheric temperature. For example, Randel and Cobb (1994) analyzed total column 03 and temperatures in the lower stratosphere from 1979 to 1992. Correlations between 03

and temperature associated with the solar cycle, the QBO, and the ENSO were identified. The variations were ~13 DU per K for the solar cycle effect, 14—16 DU per K for the tropical and subtropical QBO, and -6-9 DU per K for the ENSO.

Heterogeneous chemistry leading to ozone destruction can also lead to ozone-temperature correlations since many of the important reactions forming active chlorine are faster at lower temperatures (see Chapter 12). In addition, there is more PSC formation at the poles and hence more ozone destruction in these regions associated with lower temperatures (see Fig. 13.14 and associated discussion below).

Some trend analyses have also observed correlation between total column 03 and the temperature of the free troposphere, which also has dynamical origins. For example, Staehelin et al. (1998b) report that at Arosa, Switzerland, the change in 03 per degree change in tropospheric temperature was —(4.34 ± 0.20) DU per K, attributed to the cold advection of polar air having high total 03 or warm subtropical air containing low levels of 03.

A further complication in such trend analyses is potential systematic errors associated with various measurement techniques used to determine ozone. Several different techniques have been in use for sufficient lengths of time that they can be used to assess longer term trends in ozone; ground-based techniques include Dobson measurements, Umkehr techniques and ozonesondes, and satellite approaches include TOMS (Total Ozone Mapping Spectrometer), SBUV (Solar Ziackscatter LTtraKiolet), and SAGE (Stratospheric Aerosol and Gas Experiment satellite measurement) (Kaye, 1995).

Dobson instruments are based on the measurement of solar radiation transmitted through the atmosphere at pairs of wavelengths around 300 nm, one of which ozone absorbs more strongly. Such ground-based measurements have been made since 1926 at Arosa, Switzerland. Given that this is the longest data record for column 03, there has been considerable effort expended to provide "quality control" on the data as different instruments, detection devices, monitoring wavelengths, and calibration procedures were substituted over the years (Staehelin et al., 1998a). More than 30 such instruments have been operating worldwide since 1957. The Umkehr technique uses a Dobson spectrometer to make measurements at dawn and dusk, i.e., at high solar zenith angles, to obtain the vertical distribution of ozone from the effects on light absorption of changing the total path through the atmosphere.

Ozonesondes are in situ measurements launched from the ground to altitudes of ~30 km to measure

0 0

Post a comment