CHAPTER 2 • The Chemical and Physical Properties of Marine Aerosols: An Introduction
71
els have recently addressed this issue (see for example, Marticorena et al. 1997; Miller
and Tegen 1998; Tegen and Miller 1998; Ginoux et al. 20m). Although these do replicate some features of the global dust cycle, there are major discrepancies. Most notable is that models tend to show very large amounts of dust transport in the Southern Hemisphere, whereas satellites show very little dust activity there (Pro spero et al.
2002). For example, most models show very large plumes of dust emerging from Australia, when in fact Australia is a very weak dust source. This is apparent from satellite
images such as AVHRR (Fig. 2.3) and TOMS (Fig. 2.4). The development of better
models will hinge on a better understanding of the factors affecting dust mobilization and a more complete knowledge of the physical environments in the source regions. They will also require more and better data on the distribution of dust over the
oceans.
There are major issues to be resolved with regard to the temporal and spatial variability of dust transport and deposition. Over the open ocean where the mass median
diameter is only a few micrometres at most, the dominant deposition mechanism is
generally by precipitation (i.e. "wet" removal). Removal by "dry" processes (principally
by sedimentation) could be important in coastal regions close to sources (such as along
the western coast of North Africa) where the dust size distribution is skewed towards
large particles. There are very few measurements of dust deposition rates to the ocean;
consequently, they must be estimated. Estimates of wet deposition rates are based on
calculations using scavenging ratios (defined as the concentration of a substance in
rain divided by the corresponding concentration in air). Scavenging ratios are empirically derived from measurements made with collocated precipitation and aerosol samplers. However, because of the dearth of long-term measurements of dust in precipitation and aerosols (Prospero 1996a,b), it has been necessary to extrapolate scavenging ratios to the world ocean. The most recent and most comprehensive estimate of
dust deposition to the oceans is presented in Duce et al. 1991. For various reasons, Duce
et al. 1991, used a scavenging ratio of 200 for the North Atlantic; for the remainder of
the world ocean, they used a ratio of 1 000 (which will yield a deposition flux that is
five times that obtained with the 200 value, all other things being equal). The Duce
et al. estimates are shown in Table 2.4. To demonstrate the sensitivity of the estimated
global fluxes to the scavenging ratio, Table 2.4 also shows wet deposition rates obtained
using a global scavenging ratio of 200. The difference in the global flux is almost a
factor of three. Despite the large differences in the flux estimates, it is clear that the
highest basin rates are in the northern North Atlantic, the northern Indian Ocean, and
the North Pacific.
Ultimately, the estimates of dust deposition to the ocean must be reconciled with
the accumulation rates in the deep-sea sediments. (However, it should be noted that a
major shortcoming of such comparisons is that sediment accumulation rates are in
effect long-term averages, while the atmospheric deposition estimates are based on
present day conditions.) Rea (1994) estimated dust deposition rates based on the analysis of aeolian materials in pelagic sediment cores. Rea's measured accumulation rates
for aeolian materials in Holocene sediments in the central North Pacific Ocean seem
to be in reasonable agreement with the estimates by Duce et al. (1991). Data on the
Atlantic are sparse and concentrated in the eastern equatorial regions; but here too,
the agreement seems acceptable. In contrast, Rea finds that the Duce et al. deposition
rates to the southern oceans are too high by a factor of 5 to 10; however, Rea's data
71
els have recently addressed this issue (see for example, Marticorena et al. 1997; Miller
and Tegen 1998; Tegen and Miller 1998; Ginoux et al. 20m). Although these do replicate some features of the global dust cycle, there are major discrepancies. Most notable is that models tend to show very large amounts of dust transport in the Southern Hemisphere, whereas satellites show very little dust activity there (Pro spero et al.
2002). For example, most models show very large plumes of dust emerging from Australia, when in fact Australia is a very weak dust source. This is apparent from satellite
images such as AVHRR (Fig. 2.3) and TOMS (Fig. 2.4). The development of better
models will hinge on a better understanding of the factors affecting dust mobilization and a more complete knowledge of the physical environments in the source regions. They will also require more and better data on the distribution of dust over the
oceans.
There are major issues to be resolved with regard to the temporal and spatial variability of dust transport and deposition. Over the open ocean where the mass median
diameter is only a few micrometres at most, the dominant deposition mechanism is
generally by precipitation (i.e. "wet" removal). Removal by "dry" processes (principally
by sedimentation) could be important in coastal regions close to sources (such as along
the western coast of North Africa) where the dust size distribution is skewed towards
large particles. There are very few measurements of dust deposition rates to the ocean;
consequently, they must be estimated. Estimates of wet deposition rates are based on
calculations using scavenging ratios (defined as the concentration of a substance in
rain divided by the corresponding concentration in air). Scavenging ratios are empirically derived from measurements made with collocated precipitation and aerosol samplers. However, because of the dearth of long-term measurements of dust in precipitation and aerosols (Prospero 1996a,b), it has been necessary to extrapolate scavenging ratios to the world ocean. The most recent and most comprehensive estimate of
dust deposition to the oceans is presented in Duce et al. 1991. For various reasons, Duce
et al. 1991, used a scavenging ratio of 200 for the North Atlantic; for the remainder of
the world ocean, they used a ratio of 1 000 (which will yield a deposition flux that is
five times that obtained with the 200 value, all other things being equal). The Duce
et al. estimates are shown in Table 2.4. To demonstrate the sensitivity of the estimated
global fluxes to the scavenging ratio, Table 2.4 also shows wet deposition rates obtained
using a global scavenging ratio of 200. The difference in the global flux is almost a
factor of three. Despite the large differences in the flux estimates, it is clear that the
highest basin rates are in the northern North Atlantic, the northern Indian Ocean, and
the North Pacific.
Ultimately, the estimates of dust deposition to the ocean must be reconciled with
the accumulation rates in the deep-sea sediments. (However, it should be noted that a
major shortcoming of such comparisons is that sediment accumulation rates are in
effect long-term averages, while the atmospheric deposition estimates are based on
present day conditions.) Rea (1994) estimated dust deposition rates based on the analysis of aeolian materials in pelagic sediment cores. Rea's measured accumulation rates
for aeolian materials in Holocene sediments in the central North Pacific Ocean seem
to be in reasonable agreement with the estimates by Duce et al. (1991). Data on the
Atlantic are sparse and concentrated in the eastern equatorial regions; but here too,
the agreement seems acceptable. In contrast, Rea finds that the Duce et al. deposition
rates to the southern oceans are too high by a factor of 5 to 10; however, Rea's data