Physical and Chemical Properties of Australian Mineral Dust Aerosols
Three field campaigns have been undertaken within the Lake Eyre Basin (LEB) - Birdsville in 2006; Muloorina Station in 2007; and Fowlers Gap Station in 2009 (see Figure 2) - to investigate the optical properties, size-resolved mass and chemical properties, and mineralogy of Australian mineral dust aerosol.
In addition, dust aerosol was collected in Sydney when a strong dust storm hit the city during September 2009. During all these campaigns we collected size-resolved aerosol samples by using a 12 stage Micro Orifice Uniform Deposition Impactor (MOUDI) (Figure 3) for subsequent analysis: ion beam analysis, ion chromatography, and mineralogy. In addition, we have three years of ground-based remote sensing data from Birdsville.
Figure 3: 12 stage Micro Orifice Uniform Deposition Impactor (MOUDI).
Particle Size Distribution
Analysis of the mass distributions showed that the fine (PM2.5: 'particulate matter with diameter less than 2.5 μm') and coarse modes each accounted for approximately 50% of the particulate mass in the LEB atmosphere during background (non-dusty) days. Small particles in the atmosphere will have longer lifetimes, travelling long distances, with more chance to undergo chemical and physical processing in the air. Thus Australian mineral dust could be good source to the open and remote ocean, particularly for Fe, an essential ingredient in plankton growth. However, the dust storm days saw a very significant enhancement of the coarse mode, a clear indication of the nature of these dust storms when higher winds are able to more easily mobilise, and subsequently loft such larger particles.
Figure 4 shows size distributions obtained under dust storm conditions, and also on the day after a dust storm, measured at Sydney, Muloorina, Fowlers Gap and Birdsville. (Note the different scales of the two panels.) The dust storm at Muloorina was quite strong, and it is clear that there was still a lot of dust in the air the day after it. By contrast, the Birdsville dust storm was much weaker.
Figure 4: Aerosol size distributions
measured at Sydney, Muloorina, Fowlers Gap and Birdsville:
We used Ion Beam Analysis (IBA) at ANSTO – specifically PIXE (Proton-Induced X-ray Emission) and PIGE (Proton-Induced Gamma ray Emission) - to provide the (relative) concentrations of many elements: this was the first step in our analysis. IBA showed that Si and Fe occurred in all sizes and all samples from all sites, and these elements were well correlated. The Fe/Si mass ratios were higher than the value in the Earth's crust in all samples, which indicates that soils in the Australian arid region are rich in iron - as expected. In turn, the scatter plot slopes of Fe vs. Al were found to 0.77, 0.9, 0.94 and 0.9 for Birdsville, FGS, Muloorina and Sydney, respectively (see Figure 5), which is higher than the values reported in the literature for other parts of the world (typically 0.4 - 0.7).
Figure 5: Scatter plots of Fe vs. Al at Muloorina, Birdsville, Fowlers Gap and Sydney.
Scatter plots are a good way of determining which elements have a common source. Thus we found strong correlations between Si, Al, Fe and Ti, as all are standard components of crustal material. In some cases we found evidence of two populations, which indicates that some of a particular element has a soil source, but some has a different origin. Figure 6 shows two examples of this.
Figure 6: Images of minerals particles analysed by QENSCAN technique.
We used QEMSCAN, a powerful new tool developed by CSIRO, to examine the mineralogy of many thousands of our collected particles. Figure 4 shows some sample images. QEMSCAN analysis showed that, as expected, Fe is found in a number of different minerals, including iron oxides. However, this analysis suggested a somewhat lower Fe content, based on the concentrations of iron-containing minerals, than did the IBA results. The most likely explanation of this is that some of the Fe seen by IBA is substitutional in other minerals ('fixed iron'), rather than contained in standard iron minerals ('free iron'). It is also clear from Figure 6 that Australian dust particles are internally mixed and well aggregated, although some of this aggregation seems likely to have happened on the filter, not in the atmosphere.
We used suppressed Ion Chromatography (IC) at CSIRO to measure the concentrations of soluble ions from selected filter sets. Sodium and chlorine occurred in most size fractions during all events and in all sites. The Na/Si and Cl/Si mass ratios were higher than the values in the Earth's crust during non dust storm days at all sites (especially when the wind blew from, or passed over the LEB). This strongly argues that these ratios are a good signature of salt lifted from dry lakes in the LEB (e.g. Lakes Eyre, Frome, Callabonna and Blanch). The Cl/Na mass ratio was higher than the value in the Earth's crust in all events and in all sites. We estimated that roughly 0.5% by mass of the dust was NaCl, although this fraction decreased somewhat during dust storms.
Sulfate in our samples could be from many sources. The burning of fossil fuels often releases SO2 which is oxidised to H2SO4. This may further react to form (fine mode) aerosols. Sea spray contains coarse mode SO42-. Finally, marine biogenic emissions also release sulphur compounds (DMS; dimethyl sulphide) which are progressively oxidised, first to MSA (methane sulfonic acid), and then to SO42-.
We used standard techniques to subtract the sea salt SO42-. Non-sea salt SO42- was found to be mostly in the fine fraction, in all samples collected in the LEB. A high percentage of nssSO42- was found to be non-biogenic. In addition the mass ratios of nssSO42- to MSA were in the range 31-164, suggesting a large input of pollutant emission to atmospheric nssSO42- over this region. Mt Isa is one possible source.
More than 90% of the NH4+ ion exists in the fine mode for the samples collected in LEB which suggests that the NH4+ in this region is not derived from sea salt or soil but most likely from the gas phase, leading to reaction products in the ammonium sulfate family. The mass ratios of NH4+ to SO42- support this conclusion. Figure 7 shows the size distributions of MSA, SO42- and NH4+ at Birdsville: their similarities are striking.
Figure 7: Size distributions of MSA, sulfate and ammonium ions at Birdsville.
Figure 8 shows a scatter plot between NH4+ and SO42- at Fowlers Gap, with an excellent correlation. The slope of this plot confirms the presence of (NH4+)2SO42- to a very high accuracy.
Figure 8: Scatter plot of ammonium vs. sulfate at Fowlers Gap.
Finally the presence of the ions of three organic acids - formic, acetic and oxalic - in our samples demonstrates that the region is not totally devoid of biogenic inputs.
Radhi, M., Box, M.A., Box, G.P., Mitchell, R.M., Cohen, D.D., Stelcer, E. and Keywood, M.D., 2010. Optical, physical and chemical characteristics of Australian continental aerosols: results from a field experiment. Atmos. Chem. Phys., 10, 5925-5942.
Radhi, M., Box, M.A., Box, G.P., Mitchell, R.M., Cohen, D.D., Stelcer, E. and Keywood, M.D., 2010. Size-resolved mass and chemical properties of dust aerosols from Australia’s Lake Eyre Basin. Atmos. Environ., 44, 3519-3529.
With thanks to
Dr Ross Mitchell, Dr Melita Keywood, CSIRO Marine and Atmospheric
Dr David Cohen, Mr Ed Stelcer, ANSTO
Dr David French, CSIRO Energy Technology