Studying the Atmosphere Over Africa
using Astronomical Data: I - Extinction Measurements

Abstract. The article highlights the potential uses of astronomical extinction measurements in the study of the transport and concentration of aerosols, which in turn have an effect on the global radiation balance, as well as cloud formation. Some examples of cases where astronomical extinction measurements have facilitated atmospheric research in Africa are presented, these being (a) the properties of Saharan dust and its transportation to the Canary Islands; (b) the Pinatubo ash-cloud and its evolution; (c) the brown haze in Cape Town; and (d) the passage of pyrogenic aerosol clouds over Sutherland.

Sommaire. L'article met l'accent sur les utilisations potentielles des mesures de l'extinction astronomique dans l'étude du transport et de la concentration des aérosols qui, à leur tour, ont un effet sur le bilan global des radiations comme sur la formation des nuages. Quelques exemples de cas où les mesures d'extinction astronomique ont facilité la recherche atmosphérique en Afrique, sont présentés: (a) les propriétés de la poussière du Sahara et son transport vers les Iles des Canaries, (b) le nuage de cendres du Pinatubo et son évolution, (c) la brume foncée à Cape Town, et (d) le passage des nuages d'aérosols pyrogéniques au-dessus du Sutherland en Afrique du Sud.

1. Atmospheric transmission: Atmospheric extinction measurements, which are regularly made by astronomers engaged in all-sky photometry, can be used to determine the concentration, transportation and typical particle size of atmospheric aerosols.

1. Diffuse radiation: Night-sky brightness depends on, amongst other things, aerosol absorption and reflection properties, altitude, azimuth and lunar position. The theory for deriving night sky intensity has been developed and applied by several authors^[3,7]. By fitting these models to the night sky measurements (often recorded during routine astronomical observations) it is possible to determine parameters such as aerosol concentration and reflectivity.
2. Airglow: Sky comparison spectra recorded during spectroscopic observing programmes contain usually unutilised information about atmospheric molecular absorption and fluorescence lines.
3. Atmospheric micro-turbulence: Astronomical "seeing" is the apparent size of an intrinsically point-like image of a star after passage of the wavefront through the atmosphere. "Seeing" measurements are frequently recorded in astronomical work, either explicitly during site testing operations, or as a by-product of imaging observations, or even when estimates are recorded in observing logbooks. "Seeing" is indicative of the degree of atmospheric instability and can be compared with meteorological and topographic data to investigate micro-turbulence.
4. Cloud formation studies: Despite the astronomers' understandable preoccupation with clouds, almost no attempts have been made to use astronomical instruments and facilities for the study of the reflectivity, transmissivity, polarisation and growth of clouds. It is conceivable that such studies could be carried out with astronomical data recorded during partly cloudy conditions. It is an area of study that awaits development.

Apart from their use as tracers in atmospheric circulation studies, aerosols have more recently been recognised as important contributors to weather phenomena and climate change. This is partly due to their role as nuclei on which water droplets can grow, and also partly because of their effect on the global radiation balance.

Three processes determine the concentration and particle size distribution of an aerosol ensemble:
 

• The injection of aerosol into the atmosphere from ground level through a variety of mechanisms described below;
• The growth of particles through the coalescing of smaller particles;
• The deposition of airborne particles on the ground through precipitation.

(a) Volcanic ash: Propelled skywards in the course of volcanic eruptions, these sulphur-rich aerosols are occasionally lifted as high as the stratosphere, where they have typical lifetimes of several years, much longer than their tropospheric counterparts. Recent such events include the eruptions of Agung (1963), El Chichon (1982) and Pinatubo (1991). Characteristically, the aerosols get dissipated throughout the stratosphere within a few months. The particles then coalesce until they become too large to be supported and fall to the ground.
(b) Pyrogenic aerosols: These are in essence the smoke from forest and savannah fires. High concentrations of these aerosols are usually recorded over sub-Saharan Africa during and just after the dry season.
(c) Windborn sand and dust: Such aerosols are usually generated in arid regions and tend to be rich in silicates. Significant generation of dust also occurs in wetter areas following the ploughing season or even through traffic on dirt roads.
(d) Maritime aerosols: These result from the uplifting of sea spray through wind. These particles characteristically have high abundances of sodium chloride. Though prevalent over the oceans, these aerosols can be transported far inland.
(e) Biogenic emissions: Biogenic processes are more commonly responsible for trace gas generation, which may contribute to the formation of aerosols. They also produce airborne microscopic organisms such as pollen.
(f) Industrial and other anthropogenic emissions: Aerosols originating in this fashion include the emissions from coal burning power stations, dust generated by opencast mining operations and domestic wood and coal burning.

where k is the standard astronomical extinction coefficient, defined as

The Rayleigh extinction is almost constant at any particular location and altitude, while ozone only affects specific parts of the spectrum. Outside these spectral regions any variations in the extinction are thus due to changes in the aerosol concentration or characteristics.

Extinction by aerosols is largely the result of Mie scattering, and its dependence on wavelength may be described by the following relation ^[4]:

The coefficient α ranges from 0 for very large particles to 4 for very small particles.

Saharan dust is occasionally transported as far as the Canary Islands in the northern hemisphere summer months. It manifests itself as an almost fog-like haze at the various astronomical sites on the archipelago, such as the Roque de los Muchachos observatory on La Palma. Through the measurement of the extinction during such events it has been possible to not only monitor the passage and density of the dust clouds, but also to determine the colour dependence of the aerosol opacity (and hence particle size distribution) of Saharan dust. Stickland et al (1987) found that the aerosol opacity at La Palma is independent of wavelength to a good approximation^[9]. This confirmed the theoretical work of several authors^[10], who showed that the refractive properties of typical Saharan dust grains are expected to be colour-neutral. Kidger^[5] and Andrews & Williams^[1] found a small wavelength dependence on the extinction coefficients in the infrared and optical regimes respectively, which is likely to be the result of mixing of the type of grains modelled by Whittet, Bode & Murdin with smaller particles^[10].

The Pinatubo ash-cloud and its evolution

The volcanic eruption of Mount Pinatubo in the Philippines in 1991 injected huge quantities of volcanic ash into the stratosphere. Within a couple of months these volcanic aerosols became distributed around the globe. The development of the volcanic ash clouds over the South African Astronomical Observatory in Sutherland can be traced by plotting the measured extinction coefficients^[6]. The study showed that enhanced aerosol concentrations persisted for several years. It also illustrated the patchy nature of the stratospheric ash clouds.

Figure 1 shows the volcanic ash extinction coefficients calculated by Kilkenny as a function of wavelength for two high-extinction events following the eruption. These were obtained by subtracting the Sutherland "normal" (i.e. pre-Pinatubo clear day) values from the measured extinction coefficients. Note that the value of a i.e. the slope of the graph) is much smaller on 26 September 1992 than on 10 September 1991. This illustrates the change in the particle size distribution in the intervening period - the smaller particles that had dominated the distribution soon after the eruption had coalesced into bigger units a year later.

The brown haze in Cape Town

Where telescopes equipped with photometers exist in urban areas, the extinction measurements may be utilised to study pollutants. The city with the largest available extinction value database in Africa is probably Cape Town, as a result of the extensive standard star work by Cousins at the South African Astronomical Observatory headquarters. Cousins has described extinction coefficient behaviour as a function of meteorological conditions^[2]. He has been able to detect maritime aerosols and the "brown haze", which is caused by domestic fires in the Cape Flats. Future measurements of the extinction at the site will provide the opportunity to monitor the severity of brown haze type pollution as a result of further urbanisation and electrification.

Passage of pyrogenic aerosol clouds over Sutherland

During the winter months an anti-cyclonic air circulation pattern frequently develops over southern Africa.

The late winter months are a period of intense woodland burning in the belt just to the south of the Intertropical Convergence Zone, centred on Zambia and including neighbouring countries.

The pyrogenic aerosols thus placed into circulation are frequently transported southward and form layers of haze over the subcontinent, occasionally moving as far south as Sutherland. On the night of 29-30 September 1997, an aerosol cloud passed over Sutherland observatory, and the extinction was measured regularly throughout the night. Brownish haze was spotted above the horizons at dawn, making it unlikely that the aerosols were locally generated dust.

Figure 2 illustrates the change of the U, B and V-band extinction coefficients during the course of the night. The 23h00 arrival time of the aerosol cloud and its intensification just before dawn can clearly be seen on the graph.

Such events can be interpreted in conjunction with meteorological data to estimate the generation and transport of the aerosols.

1. Andrews, P.J., Williams, I.P. 1989, The Observatory, 109, 15.
2. Cousins, A.W.J. 1985, Mon. Not. Astr. Soc. South Africa, 44, 10.
3. Garstang, R.H. 1991, Publs Astr. Soc. Pacific, 103, 1109.
4. Hayes, D.S. & Latham, D.W. 1975, Astrophys. J., 197, 593.
5. Kidger, M.R. 1988, The Observatory, 108, 226.
6. Kilkenny, D. 1995, The Observatory, 115, 25.
7. Krisciunas, K., Schaefer, B.E. 1991, Publs Astr. Soc. Pacific, 103, 1033.
8. Spencer Jones, J.H. 1980, Mon. Not. Astr. Soc. South Africa, 39, 89.
9. Stickland, D.J., Lloyd, C., Pike, C.D. & Walker, E.N. 1987, The Observatory, 107, 74.
10. Whittet, D.C.B., Bode, M.F. & Murdin, P. 1987, Vistas in Astronomy,30, 135.