the vertical distribution of ozone. They typically use chemical analytical techniques such as oxidation of I~ to I2; as discussed in Chapter 11, although such instruments respond to 03, they also respond positively to some extent to other compounds such as NOz and negatively to others such as S02.
The SBUV instrument is based on satellite measurements of backscattered light at 12 wavelengths between 255 and 340 nm; from comparisons of wavelengths where ozone absorbs to those where Rayleigh light scattering occurs, total column ozone and vertical profiles with low resolution (5-15 km, depending on the altitude) can be obtained. One such instrument operated from 1978 to f990 and another from 1989 to 1994, with two more operating in 1996. Some of the difficulties in extracting the vertical 03 distribution using SBUV data are described by Ziemke and Chandra (1998).
TOMS is also a satellite-based method based on a similar approach to that of SBUV, in which the earth's albedo is measured at several wavelengths around 300 nm. Unlike SBUV, vertical distributions of 03 are not derived using TOMS, but it provides better horizontal resolution (Ziemke et al., 1998). TOMS has also been used to measure tropospheric aerosols using the wavelength dependence of UV reflectivity at wavelengths that are not absorbed by 03 (e.g., Hsu et al., 1996; Herman et al., 1997; Torres et al., 1998).
SAGE instruments have been operating since 1978, although with a gap between 1981 and 1984. Figure 13.8 illustrates the geometry associated with these measurements (Rusch et al., 1998). These instruments are solar occultation instruments that measure the light intensity at seven wavelengths at satellite sunrise and sunset, giving about a dozen such measurements in one 24-h day. Three of the wavelengths are used to derive 03 (0.6 nm), N02 (0.448 ftm), and water vapor (0.94 fim), whereas the other four (1.02, 0.525, 0.453, and
0.385 /jum) are used to extract aerosol particle concentrations. Light transmission data are generated by comparison to the unattenuated solar flux and, along with ancillary measurements of other parameters, can be used to obtain extinction profiles for 03 and for other species, including aerosol particles, water vapor, and N02 (Rusch et al., 1998).
However, extracting ozone concentrations and their vertical profiles from SAGE data (and indeed other long-path, remote techniques) is not straightforward. Typically, the contribution from Rayleigh scattering is first removed and then the contribution of each gas to the total extinction is derived as a function of the tangent altitude (z, in Fig. 13.8) using the wavelength chosen to be characteristic of light absorption by gas (the so-called species inversion). Finally, the contribution of each gas as a function of altitude (z;) is extracted (the "geometric" or "altitude" inversion) (e.g., see Cunnold et al., 1989). Rusch et al. (1998) have shown that the order of applying the inversions can make a difference; for example, carrying out the altitude inversion first and then the species inversion improves agreements for 03 below 22 km with concentrations derived using ozonesondes.
An important factor in deriving 03 concentrations is the presence of aerosol particles, which also scatter light at 0.6 yu,m. Thus, correction for their contribution to extinction at this wavelength must by applied to derive the ozone concentrations. This requires some assumptions regarding aerosol particle properties such as the size distribution, which is not known. It is also commonly assumed that the optical properties of particles do not change with altitude. Such problems introduce uncertainties into the calculation of the particle contribution (e.g., Steele and Turco, 1997a, 1997b; Thomason et al., 1997; Fussen, 1998) and hence into the ozone concentrations extracted from such data.
For example, Fig. 13.9a shows the trends in column 03 as a function of latitude in the stratosphere (at altitudes above pressures corresponding to 82.5 mbar) for the period from November 1984 to May 1991 derived using SAGE II measurements (Cunnold et al., 1996). Also shown are the trends derived from TOMS data made for the same locations; the latter include changes in tropospheric ozone as well. Figure 13.9b shows similar data for a smaller period of time, from January 1988 to May 1991, during which time stratospheric aerosol concentrations were quite stable and small. In the 1984-1988 period included in Fig. 13.9a, the aerosol particle concentrations were larger and were decreasing with time. The agreement between the TOMS and SAGE II data is much better for the period with small particle concentrations, suggesting that they were responsible for much of the discrepancy seen in
Was this article helpful?