Ideally, some of the more useful quantities presented herein could be included as CPFM data products. This would require additional work beyond what has been conducted here to `automate' the data reduction, quality control, retrieval and analysis aspects. The best candidates would be VCDs of ozone, NO2, and BrO below the aircraft as the DOAS retrieval involves a straightforward least-squares fit and conversion to VCDs could be achieved using precalculated air mass factor look-up tables. In general, air mass factors would have to be functions of solar zenith angle, surface albedo, ER-2 altitude, and perhaps air column/latitude. Additionally, ozone VCDs are a function of the ozone profile due to the larger optical depths. Retrievals using limb spectra might prove more problematic as air mass factors would have to be calculated on-line due to their sensitivity to elevation angle.
It may be possible to extract further information by making use of the temperature dependence of the ozone cross-sections in the UV. The DOAS multiple-regression coefficient was fairly sensitive to the choice of temperature and this can be exploited to retrieve ozone information below the plane possibly into the troposphere using an algorithm similar to that currently employed by GOME (Munro, 1998).
There remains many other useful applications of the CPFM measurements not addressed in this study. One possible example is using the nadir radiance and polarization to extract aerosol information in the troposphere. Tropospheric aerosols are much more variable in type, size, and number density and so even limited information would be very important. Concomitant with this is the issue of surface polarization properties, the impact of which will likely be much larger when using the nadir for aerosol retrievals over using the limb. It may be possible to retrieve both simultaneously using multiple wavelengths where surface effects are both important at one and marginal at another. This would be especially useful if applied to the polar sunrise as the relationship between BrO and aerosol would elucidate the role of heterogeneous chemistry in the free troposphere. Another useful study would be developing a better and more physically realistic algorithm for representing the mean radiance from the CPFM-measured fields for use in J-values calculations. This would be achieved by simulating all fields, calculating the mean radiance based on these, and comparing with the complete angularly-integrated mean radiance.
From the tropospheric BrO results related to polar sunrise ozone depletion phenomena, a natural next step is to mount the CPFM (or similar instrument) on a conventional aircraft and make measurements while flying in the troposphere. This would eliminate the uncertainties associated with accounting for the stratospheric column. By flying at different altitudes, much greater vertical resolution could be obtained. In addition, flying in the troposphere will also allow the simultaneous measurement of ozone as roughly 90% of the column would be above the aircraft. The spatial extent of enhanced BrO appears to be extremely large but it is difficult to equate this to the spatial extent and magnitude of ozone depletion. This also makes the previous suggestion concerning the retrieval of tropospheric aerosols using the nadir even more attractive.