Introduction

In recent years, stratospheric aerosols have received much attention due to the large role they play in atmospheric chemistry and climate. It was in the wake of the newly-discovered ozone hole (Farman et al., 1985) that the importance of stratospheric sulphates and polar stratospheric clouds were fully realized. Aerosols serve as surfaces and volumes which catalyze reactions which would not occur in the gas-phase. One important example at cold temperatures is,

$$\rm HCl + ClONO_2 \stackrel{\rm aerosol} \longrightarrow \rm Cl_2 + HNO_3$$ (9.1)

$$\rm Cl_2 + h\nu \longrightarrow \rm 2 Cl$$ (9.2)

which, following a photolysis reaction, converts chlorine from two reservoir species (which do not react with ozone) to a form which can catalytically destroy ozone. Stratospheric aerosols also perturb the radiation budget and can impact the global climate. They are efficient scatterers of solar radiation and act to increase the global albedo and hence cool the planet. Specific to the ER-2 flights, it is important to have an accurate representation of the aerosol for the forward radiative transfer modeling necessary to support photochemical modeling or in the retrieval of trace-gases such as NO₂ and BrO.

Instruments which measure aerosols can be roughly grouped into four classes: (1) in-situ, or direct collection (e.g.: Wilson et al., 1983); (2) active remote sensors such as lidar; (e.g.: Thomason and Osbourne, 1992); (3) occultation, or transmission (e.g.: Wang et al., 1989); and (4) scattering (e.g.: Kaufman et al., 1994). Each offers advantages and disadvantages. Lidar, for example, can give excellent vertical resolution but no horizontal resolution (other than what can be inferred from air trajectories). In-situ instruments can be used to measure size distribution and number, surface area and volume densities. The collected sample can also be analysed for composition. The problem with in-situ measurements is that aerosol density may not be homogeneous. Global coverage is obtained by either transmission or scattering instruments mounted on satellite platforms. The Satellite Aerosol and Gas Experiment (SAGE) II, the Stratospheric Aerosol Measurement (SAM) II, and other satellite instruments, have been making global maps of extinction coefficient and surface area density since the late 1970s.

Optical instruments, in general, measure only the total radiance. However, as will be shown this chapter, radiance is largely insensitive to aerosol size. A much better quantity which yields size information is the linear polarization (Hansen and Travis, 1974). Mie and Rayleigh scatterers behave quite differently as polarizers. All Rayleigh phase matrix elements are of roughly the same size and polarization varies slowly with scattering angle between 0 and 1. The Mie P₁₁ element is generally much larger than the others and so polarization rarely exceeds 0.4 (and is usually less than 0.2) for realistic size distributions. The `depolarizing' nature of aerosols is physically the result of interference, which also produces rainbows, halos and glories, features which result in sharply defined regions of high polarization (Hansen and Travis, 1974).

Polarization is not frequently used to retrieve aerosol information. A rare example was that of a polarimeter, mounted on a balloon platform, which measures both radiance and polarization at 850 and 1650 nm (Herman et al., 1986). Data was recorded with the balloon at 20 km and with a solar zenith angle near 90$^{\circ }$. The instrument was rotated in the azimuth through the tangent plane so that scattering angles from 0 to 180$^{\circ }$ are sampled. The retrievable quantities include the effective radius and variance, refractive index, and the extinction coefficient between 16 and 23 km (Brogniez et al., 1997). One very successful application of polarization to the remote sensing of aerosols was in the identification of the thick clouds which cover the surface of Venus. Hansen and Hovenier (1974) compared linear polarization measurements of sunlight reflected by Venus with multiple-scattering model calculations and deduced the size distribution and refractive index of the cloud aerosol. They ascertained that the clouds were composed of sulphuric acid droplets with an effective radius of 1.05 $\mu$m and an effective variance of 0.07. It is of interest to note that some of the measurements they used dated back to 1929.

The purpose of this chapter is two-fold. First, a sensitivity study is presented to gain insight into how various properties of aerosol impact limb radiance and polarization. Based on these results, it will become clear which aerosol properties are potentially retrievable. In addition, a better understanding of the multiple-scattering and polarization phenomena will be gained. The second goal is to devise a method of aerosol retrieval consistent with the knowledge gained from the sensitivity study.