This paper promotes an understanding of the mineralogical, chemical, and
physical interrelationships of re-suspended mineral dusts collected as grab
samples from global dust sources. Surface soils were collected from arid
regions, including the southwestern USA, Mali, Chad, Morocco, Canary Islands,
Cabo Verde, Djibouti, Afghanistan, Iraq, Kuwait, Qatar, UAE, Serbia, China,
Namibia, Botswana, Australia, and Chile. The
Globally, airborne mineral dust has the largest mass emission rate, average
column mass burden, and average optical depth of all aerosol types (Satheesh
and Moorthy, 2005). More than half (
In the past only particles less than
The mineralogy of dust from different geological domains differs from each other. How these relate to variations in optical properties influencing radiative effects, visibility impairment, and their potential health effects is not clearly understood. An interdisciplinary approach to characterizing mineral dust from different geological source regions provides important information for modeling of radiative effects and climate change (e.g., Satheesh and Moorthy, 2005), remote sensing (e.g., Ginoux et al., 2012), visibility impairment (e.g., Watson, 2002), as well as impact on solar energy production (e.g., Sayyah et al., 2014), ecosystems (e.g., Field et al., 2010), and human health (Middleton et al., 2008; Norboo et al., 1991; Wiggs et al., 2003; Fubini and Fenoglio, 2007; Ghio et al., 2014). The prime objective of this publication is to provide speciated information on mineral dust aerosol optical properties, mineralogy, chemistry, as well as particle morphology and size distribution, all required for assessments of regional and global impacts of mineral dust.
Related previous publications provided geographically limited information on mineral dust optical properties (e.g., Moosmüller et al., 2012; Müller et al., 2009; Petzold et al., 2009; Schladitz et al., 2009; Von Hoyningen-Huene et al., 2009), chemical composition, mineralogy, and particle size distribution (e.g., Al-Dousari and Al-Awadhi, 2012; Avila et al., 1997; Kahlaf et al., 1985; Kandler et al., 2007, 2009; Molinaroli, 1996). This paper expands on these previous contributions by presenting properties of a global set of re-suspended dust samples applying similar sample collection, preparation, and analysis methods.
Surface soil samples had previously been collected at more than 100 sites
worldwide, of which 65 were selected for re-suspension, monitoring, and
analysis (Fig. 1, Supplement S1). Included are samples from the
Middle East (Iraq (6), Kuwait (5), Qatar (1), United Arab Emirates (UAE) (1), Afghanistan (3),
Djibouti (1)), USA (Arizona (4), California (2), Nevada (3), Colorado (1),
Utah (2)), China (5), Spain (Canary Islands (8)), Morocco (1), Mali (3), Cabo Verde (1), Chad (3), Serbia (3), Australia (3), Botswana (4), Namibia
(3), Chile (1) (Fig. 1), as well as one sample of hematite (Fe
Surface soil sampling regions shown as green stars on a global map of dust sources as identified from Total Ozone Mapping Spectrometer (TOMS) data (Prospero et al., 2002). The size of the marker is approximately proportional to the number of samples collected in the region.
The grab soil samples collected in the field were generally sieved to
The
Mineral grain mounts for optical microscopy of the
XRD is the single most important nondestructive technique for characterization of minerals such as quartz, feldspars, calcite, dolomite, clay minerals, and iron oxides in fine dust. A Bruker D8 system with sample stacker and spinner was used to collect the XRD spectra, while Bruker Topas® software employing the relative intensity ratio (RIR) procedure was applied for semi-quantitative analyses of the XRD spectra of the dust samples (Rietveld, 1969; Chung, 1974; Esteve et al., 1997; Sturges et al., 1989; Caquineau et al., 1997) (Supplement S2.2).
Laser particle size analysis was performed on each of the soil
samples. The system measures the size-class fractions of a soil or sediment
sample in an aqueous suspension, based on the principle that light scatters
at angles inversely proportional to, and with intensity directly proportional
to, particle size (Gee and Or, 2002). The grab samples were dry sieved to
A dust entrainment facility (Fig. 2) was configured with an array of
instrumentation, including (i) three-wavelength photoacoustic instrument,
(ii) beta attenuation gauge, (iii) aerodynamic particle size analyzer,
(iv) desktop computer, and (v) PM filter samplers
(
Integrated dust entrainment facility: three-wavelength photoacoustic instrument (DMT PASS-3) on rack to the left, re-suspension chamber on trolley in the center, with a particle size analyzer and beta-attenuation gauge installed below the chamber, and an array of vacuum pumps on rack to the right.
A dust plume was generated by puffing the dust into the re-suspension chamber
with compressed air, from which particles were sampled onto filters, together
with continuous optical, particulate mass concentration, and particle size
measurements (Moosmüller et al., 2012). Dust measurements were focused
on the size fraction associated with long-range transport, typically less
than 9
The most important optical properties (Bergstrom et al., 2007; Chýlek and
Wong, 1995; Hassan et al., 2015) for calculating aerosol radiative impacts
are scattering (
The optical coefficients have dimensions of inverse distance (Mm
Teflon® membrane and quartz fiber filters were conditioned and weighed prior to and after sampling for subsequent mass and chemical composition determinations of the dust samples.
In the course of this study, about 130 Teflon® filters were analyzed by energy dispersive X-ray fluorescence spectrometry (EDXRF) (Watson et al., 1999; US EPA, 1999) for chemical elements, including magnesium (Mg), aluminum (Al), silicon (Si), potassium (K), calcium (Ca), titanium (Ti), and iron (Fe). After completion and validation of the EDXRF results, the same 130 Teflon® filters were dissolved in nitric/hydrochloric acid, and the solutions analyzed by inductively coupled plasma mass spectrometry (ICP-MS) for 12 trace metals: antimony (Sb), arsenic (As), beryllium (Be), cadmium (Cd), chromium (Cr), lead (Pb), manganese (Mn), nickel (Ni), vanadium (V), zinc (Zn), mercury (Hg), and strontium (Sr).
Water extractions were performed on one half of each of the 130 quartz fiber
filters. Aliquots of the extractions were analyzed by ion chromatography (IC)
(Chow and Watson, 1999) for water soluble anions: sulfate (SO
Chemical abundances for PM
SEM-based individual particle analysis was performed on 65 PM
CCSEM was performed using a combination of backscattered electron and
secondary electron (SE) imagery along with energy dispersive spectroscopy
(EDS). A large number (often more than 1000) of individual particles were
analyzed for chemical composition, particle size, and shape by this method.
The measurements on the particles are grouped into “bins” by chemical
composition and particle size. The CCSEM results were routinely grouped into
29 chemical categories or “bins”. Following on preliminary statistical
evaluation of the chemical data and a priori information of the mineralogical
composition (from XRD, optical microscopy, and particle morphology) the
number of chemical “bins” were reduced to 15. Mineral names were assigned
to each of the chemical categories, e.g., Si-rich category as quartz, Ca-rich
category as calcite, Ca, Mg category as dolomite, Ca, S category as gypsum,
Fe-rich category as iron oxides, Ca, Al, Mg category as clay, Ca, Al, Si
category as Ca feldspar, Na, Al, Si category as Na feldspar, K, Al, Si
category as K feldspar. Summary plots of the inferred mineral
compositions were compiled. Individual particle morphology, including aspect
rations, was measured by CCSEM. Although labor intensive, manual SEM
analysis allows for collection of superior-quality, high-resolution SE
images. This technique further allows for the study of particle shape,
surface coatings, and chemical composition. Approximately five secondary electron images (SEIs) with
EDS for each of the 65 PM
All results are compiled in the Supplement to this paper, with the sampling
sites listed in Supplement S1, mineralogy by optical microscopy in
Supplement S2.1, mineralogy by XRD in Supplement S2.2, and particle size
fractions (volume %) distributions of the
Descriptions of samples from each geographic region are summarized below.
The dried lakebed contains minerals washed in from the adjoining Sierra and
Inyo mountains by the Owens River as well as small amounts of evaporite
deposits. Since about 1926 the lake remained dry as a result of overuse of
its water for agriculture and human consumption by the city of Los Angeles
(Wilkerson et al., 2007). The coarser sand and silt fractions contain
plagioclase, calcite, quartz, biotite mica, and hornblende, with halite,
kaolinite, and trona in the finer fractions. These minerals are abundant in
local aerosols generated during dust events, thereby creating a major health
issue along the Owens Valley and poor visibility at the nearby China Lake
Naval Air Station (Reheis, 1997; Wilkerson et al., 2007). This resulted in
ongoing mitigation steps to contain alkaline dust being blown off the playa.
(
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1052 from Reno, Nevada: a fragment of a diatom containing a major component of silicon (Si), with trace amounts of aluminum (Al) pointing to a silica particle (amorphous?) with a minor amount (coating?) of clay, and possibly kaolinite.
The Black Rock playa about 150 km north of Reno is known for its very flat
lakebed, with approximately 1 m altitude variation over an area of
300 km
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1065 (red clay) from Carbondale, California. Composite particle containing major components of aluminum (Al) and silicon (Si), possibly the clay mineral kaolinite, as well as iron (Fe) in hematite or goethite.
Two sedimentary samples were collected along the Peavine mountain foothills in northwestern Reno. The first (S1052) is an off-white diatomaceous lakebed deposit (Smedman, 1969), containing diatoms (Fig. 3), plagioclase, quartz, and lesser amounts of clay (kaolinite, sepiolite, illite, montmorillonite), stilbite, and biotite. The second (S1053) is a yellowish brown soil, the color being ascribed to iron oxides (goethite), in addition containing quartz and plagioclase and smaller amounts of hornblende, biotite, muscovite, and clays (palygorskite, montmorillonite, kaolinite).
Carbondale Red Clay is a commercially available fireclay, considered to be the equivalent of Martian dust (White et al., 1997). The material is composed of very finely ground quartz and kaolinite with a high percentage of hematite (Fig. 4).
The two grab samples were collected from improved artillery firing pads at the Yuma Proving Ground (YPG) (Gillies et al., 2007). The samples contain major amounts of quartz and calcite, together with lesser amounts of dolomite, mica (biotite, muscovite), feldspars (plagioclase, microcline), hornblende, gypsum, and clays. To compact the soil and contain the dust during the gun firing activities, brines containing potassium sulfate and other salts were sprayed onto the firing pads. Trace amounts of copper, bismuth, and lead from the spent ordnance are present in the finer dust fractions (Engelbrecht et al., 2012).
This surface soil sample was collected during a field campaign measuring dust emissions, created by rotary-winged aircraft (Gillies et al., 2010). The sample contains rounded quartz grains together with calcite, feldspars (plagioclase, microcline), and mica (biotite, muscovite) and small amounts of hornblende, clay (kaolinite), dolomite, gypsum, and apatite (Engelbrecht et al., 2012).
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S3017, from Dugway, Utah. Particle with major amount of calcium (Ca) from calcite, with aluminum (Al), silicon (Si), and magnesium (Mg) and potassium (K); there is evidence of a coating of clay and possibly illite.
Fort Carson is located in central Colorado at the foot of the Rocky Mountains and to the south of Colorado Springs. The area is underlain largely by fluvial and alluvial clays, silts, and sands. The surface soil sample was collected from an unpaved road on the military base. It contains major amounts of quartz and plagioclase, with lesser amounts of biotite, potassium feldspars, calcite, hornblende, muscovite, dolomite, gypsum, and clays (Engelbrecht et al., 2012).
The two samples from Utah are similar to each other, both collected along built-up dirt roads on the military base. The road gravel material had been excavated from a local quarry. Both samples contain major amounts of quartz and calcite (Fig. 5) and lesser amounts of dolomite, aragonite, plagioclase, with traces of biotite, muscovite, illite, and orthoclase.
Djibouti lies in the Afar triangle, which is part of the African continental
rift zone, underlain by basaltic lava flows and associated sediments.
Minerals in the screened dust sample including pyroxenes, amphiboles,
olivines, plagioclases, ilmenites, and magnetites (Fe
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S2016 from Camp Leatherneck, Helmand Province, Afghanistan. Well-rounded grain of quartz (containing Si) with coating of finger-shaped clay mineral containing aluminum (Al), magnesium (Mg), iron (Fe), and possibly palygorskite.
The first two sites are in northeastern Afghanistan in the mountainous region close to the Pakistan border, while the third is in the Helmand province. Dust storms originating in the southwestern part of the country move in a counterclockwise direction, funneling wind-blown dust toward Kabul, Bagram, and Khowst. The Bagram site is underlain by loess (wind deposited silt), sand, and clay and the Khowst site by alluvium containing gravel, sand, silt, and clay. These last mentioned two samples from Afghanistan are mineralogically similar, both containing major amounts of quartz, calcite, and clays (kaolinite, illite, palygorskite), with smaller amounts of feldspar, mica (biotite, muscovite), chlorite, and amphibole (Engelbrecht et al., 2009b, 2008). The sample from Helmand Province (S2016) contains largely quartz (Fig. 6) and calcite with small amounts of clays, gypsum, and silicates.
Qatar and UAE are impacted largely by windblown dust transported from the Arabian Peninsula, Iraq, and Kuwait. Samples from both sites contain large amounts of calcite, dolomite, quartz, and feldspar and lesser amounts of chlorite, clay minerals, gypsum, and anhydrite. The Qatar sample also contains small amounts of hornblende and ilmenite, probably transported from weathered volcanics in Saudi Arabia to the west of the sampling site (Engelbrecht et al., 2009b, 2008).
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S2017 from Kuwait. Orthoclase identified from its silicon (Si), aluminum (Al), and potassium (K) content as well as crystal morphology. Finger-shaped particles on the feldspar surface may be palygorskite.
Satellite imagery shows dust storms, blowing across Iraq and Kuwait from salt flats and wadis along the border with Turkey and Syria (Engelbrecht et al., 2009b), being major sources of windblown dust in the region. The Baghdad site, as well as the three sites further north and close to the Tigris River (Taji, Balad, and Tikrit), contains major amounts of quartz, calcite, and feldspar, plus smaller amounts of gypsum, micas, and clays. The Tallil sample, collected further to the south and close to the Tigris–Euphrates river confluence, also contains the evaporite mineral gypsum. This sampling site is underlain by poorly drained river-basin soils, and the gypsum could be wind-entrained from the gypsiferous desert immediately to the southwest of Tallil. Al Asad, closer to the Euphrates River and 200 km to the west of Baghdad, differs from the other five Iraqi sites in that its soil contains largely dolomite, with small amounts of gypsum, and quartz. This region is underlain by soils containing gypsum, limestone, and sandstone (Buringh, 1960). XRD analysis of the sieved dust samples confirms that the limestone in this region is largely dolomite, instead of calcite identified at the other five Iraq sampling sites (Engelbrecht et al., 2009b, 2008).
The samples from Kuwait contain higher fractions of silt than those from adjacent Iraq, with major amounts of quartz and calcite. Samples from all five sites also contain a significant amount of feldspar, both plagioclase and potassium feldspar (Fig. 7) as well as micas (muscovite, biotite) and clay minerals. Previous studies on fallout dust in Kuwait (Al-Awadhi, 2005) identified the clay minerals as palygorskite, together with illite-montmorillonite and a smaller proportion of kaolinite. Al-Awadi (2005) and Kahlaf et al. (1985) suggest that most of the dust in Kuwait to have been transported from the dry Mesopotamian floodplain in southern Iraq and other parts of the Arabian peninsula (Engelbrecht et al., 2009b, 2008).
The Canary Island archipelago lies along the westward pathway of Saharan dust blowing off the west coast of North Africa, the islands being impacted by severe dust events in that region (Alastuey et al., 2005; Bergametti et al., 1989; Criado and Dorta, 2003; Gelado-Caballero et al., 2005; Maring et al., 2000; Menéndez et al., 2009; Pérez-Marrero et al., 2002; Viana et al., 2002; Kandler et al., 2007; Engelbrecht et al., 2014; Coudé-Gaussen et al., 1987; Mizota and Matsuhisa, 1995). It was shown that fractions of transported Saharan dust had been deposited on the islands, contributing to the formation of local soils (Jahn and Stahr, 1996; Menéndez et al., 2014, 2007; Mizota and Matsuhisa, 1995; Muhs et al., 2010; von Suchodoletz et al., 2013, 2009; Zarei, 1989; Williamson et al., 2004; Moreno et al., 2001). It can be safely assumed that samples of soils collected on these islands contain Saharan dust in variable proportions. Two samples from Fuerteventura are from dust traps in lava flows, on the eastern side of the island, approximately 100 km off the coast of Morocco. The sieved dust samples from the Canary Islands generally contain quartz as the major component, with minor amounts of calcite, feldspars, amphibole, pyroxenes, micas, clays, iron oxides, and volcanic glass.
SEM-based EDS spectrum and SEI of a particle from re-suspended
sample S1013 from Cabo Verde Islands. Particle of apatite
(Ca
The Il de Sal Island is part of the Cabo Verde archipelago, approximately 700 km off the west coast of Africa, and considered to be within the Sahel geographic zone. As with the Canary Islands, the Cabo Verde islands are impacted by dust blowing off the west coast of Africa. The soil sample collected on the northern side of Sal Island is assumed to contain some trapped Sahara and Sahel dust. The sample contains plagioclase and pyroxene in the coarse fraction, possibly from local weathered volcanic rocks, together with quartz, illite clay, calcite, carbonate apatite, halite, and kaolinite in the fine fraction. The carbonate apatite (francolite) (Fig. 8) is possibly of local biogenetic origin.
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1025 from Iriki, Morocco. From the relative intensities of the silicon (Si), aluminum (Al), magnesium, potassium (K), and iron (Fe) spectral peaks as well as the morphology shown by the SEI, it is concluded that the particle is a cluster illite (platy) and palygorskite (finger shaped).
Morocco is considered a major source of dust, blowing across the Mediterranean into Europe. Several studies were performed on airborne mineral dust from this region, some as part of the Sahara Mineral Dust Experiment (SAMUM) (Heintzenberg, 2009; Kandler et al., 2009; Schladitz et al., 2009; Müller et al., 2009; Petzold et al., 2009; Von Hoyningen-Huene et al., 2009; Scheuvens et al., 2013; Molinaroli, 1996; Molinaroli et al., 1993; Blanco et al., 2003; Avila et al., 1997; Jahn and Stahr, 1996).
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1009 from Bamako, Mali. Composite grain composed of silicon (Si) and aluminum (Al) components of a clay mineral and possibly kaolinite with lesser amounts of palygorskite (finger shaped) and illite, together with iron (Fe) in hematite or goethite.
The major minerals in this sample are quartz and illite, with smaller amounts of calcite, feldspars (plagioclase, orthoclase), hornblende, dolomite palygorskite (Fig. 9), and kaolinite.
The three grab samples were collected in a field approximately 5 km to the
northwest of Bamako in Mali, all within 200 m of each other. The samples are
from soils immediately to the south of the Sahel, the latter being defined as
that zone adjoining the southern Sahara, which receives on average between
It was shown (Oldfield et al., 2014) that the magnetic properties of dusts and their hematite/goethite contents could be applied to distinguishing dust sources from North Africa impacting the Caribbean region. Moosmüller et al. (2012) showed that the optical properties such as SSA were controlled by the Fe content of airborne dust. Roquin et al. (1989) were able to correlate the mineralogical and chemical information from Dagadamou region in southern Mali with the SPOT satellite imagery.
SEM-based EDS spectra and SEI's of particles from re-suspended
samples S1049 (above) and S1051 (below) from the Bodélé region of
Chad. Two examples of diatom fragments, the particles contain largely
silicon (Si), evidence of amorphous silica or quartz (SiO
The three dry lakebed deposits were collected from the Bodélé depression in the southern Sahara region of North Africa. Bodélé is considered the largest single source of global dust (Chappell et al., 2008; Goudie and Middleton, 2001; Tegen et al., 2006; Todd et al., 2007; Washington and Todd, 2005; Middleton and Goudie, 2001), contributing to the fertilization of the Atlantic Ocean and land as far as the Amazon basin (Ben-Ami et al., 2010; Koren et al., 2006; Chadwick et al., 1999; Formenti et al., 2001; Bristow et al., 2010).
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1014 from northwestern China. Composite particle of calcite as seen from the high calcium (Ca) content together with a clay mineral (coating?) and possibly illite as is evident from the silicon (Si), aluminum (Al), magnesium (Mg), and potassium (K) contents.
The samples contain largely amorphous silica (diatoms) (Fig. 11), quartz, and calcite with small amounts of plagioclase, potassium feldspar, kaolinite, illite, and halite.
The four samples are from far northwestern China, two (S1017, S1018) being from a locality referred to as Daemon City, named after the unusual wind eroded sandstone formations. The two samples are composed of major amounts of quartz and calcite, with plagioclase, kaolinite, mica (biotite, muscovite), and montmorillonite. The two samples further north (S1014, S1016) from the Daemon City area contain quartz and biotite, with lesser amounts of plagioclase, calcite (Fig. 12), hornblende, kaolinite, illite, sepiolite, and halite.
This region in China contains the greatest thickness of loess in the world. It is an un-stratified silt deposit formed by wind-blown fluvioglacial deposits related to the geologically recent ice ages (Derbyshire et al., 1998; Kohfeld and Harrison, 2003; Sun et al., 2000; Lewis et al., 1984). This silt sample contains quartz with smaller amounts of calcite, plagioclase, potassium feldspar, biotite, hornblende, clay (kaolinite, illite), and mica (muscovite, biotite).
The samples are from the dry lakebed of Makgadikgadi Pan in central Botswana. The samples contain largely calcite, quartz, the evaporate minerals trona and halite, and smaller amounts of sepiolite, illite, dolomite, plagioclase, orthoclase, muscovite, hornblende, and dolomite. During the dry winter months, the Makgadikgadi Pan is a major source of dust in southern Africa.
Three soil samples were collected along the southern and western edge of Etosha Pan in northern Namibia. They all contain major amounts of the evaporate minerals halite, thenardite, and calcite, with sepiolite and small amounts of dolomite, plagioclase, muscovite, and amphibole (Buch and Rose, 1996). Etosha Pan, especially during the dry winter season, is a major source of dust in Namibia, exacerbated by disturbance of the lakebed being from the passage of large herds of game.
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1057 from Lake Eyre, Australia. Round quartz particle as evident from its silicon (Si) content, with small amounts of aluminum (Al), magnesium (Mg), and iron (Fe) from clay (coating) and possibly illite or montmorillonite.
The sample was collected at the Yungay (Rock Garden) site in the Atacama Desert of Chile, considered to be one of the most arid regions of the world (McKay et al., 2003), also described as Mars-like (Navarro-González et al., 2003). The sample contains quartz as its main constituent, together with clay (illite, montmorillonite), gypsum, anhydrite, plagioclase, calcite, orthoclase, biotite, and amphibole.
SEM-based EDS spectrum and SEI of a particle from re-suspended sample S1062 from Kostolac, Serbia. The silicon (Si) content points to a rounded quartz particle contained in the loess.
The three samples are from riverbed or lakebed deposits in the Lake Eyre and Lake Frome regions of South Australia (Marx et al., 2009; McGowan and Clark, 2008; Moreno et al., 2009), considered to be a major source of dust in the southern hemisphere, being deposited on the snow as far as the New Zealand Alps (McGowan et al., 2005; Marx et al., 2009). Evidence of dust from this region was also found in East Antarctica (Revel-Rolland et al., 2006). The samples are composed largely of quartz grains (Fig. 13) coated with clays, together with small particles of clay minerals (illite, kaolinite, palygorskite), gypsum, calcite, feldspar (plagioclase, orthoclase), and dolomite.
The loess samples were collected along the Danube valley in Serbia, representing loosely bound wind-blown silt deposits, related to the ice periods of the recent geological past (Marković et al., 2009, 2011). The three samples are similar, being composed mainly of rounded particles of quartz (Fig. 14), with muscovite, biotite, clay (illite, kaolinite, montmorillonite), plagioclase, dolomite, and calcite.
Particle size analysis on the
From field and laboratory measurements on dust from the western USA, Engelbrecht et al. (2012) found dust emissions to be in part dependant on the particle size distributions of the soils being disturbed. Surface soils with silt contents greater than approximately 50 % and less than about 10 % clay had the greatest likelihood of being sources of substantial dust emissions. On the USGS soils ternary plot (Fig. 15), soils with the greatest potential to generate dust when disturbed are those that plot in the bottom one-third of the “silt loam” field. Included are two previously characterized dust sources from Yakima Test Center (Engelbrecht et al., 2012), three from the Bodélé Depression (S1049–S1051) (Washington and Todd, 2005; Koren et al., 2006), three loess samples from the Danube River valley (S1060, S1062, S1064) (Marković et al., 2011, 2009), one sample from Kuwait (S2017) (Al-Dousari and Al-Awadhi, 2012; Draxler et al., 2001; Kahlaf et al., 1985), and two loess samples from China (S1017, S1055) (Burbank and Jijun, 1985; Derbyshire et al., 1998). It also includes four silt samples collected on Fuerteventura (S1033, S1034), Lanzarote (S1007), and Gran Canaria (S1027) in the Las Canarias archipelago, assumed to consist largely of dust from the western Sahara (Menéndez et al., 2014), as well as one diatomaceous silt sample from Reno, USA (S1052). This particle size criterion provides an important measure for whether a site or region has the potential to be a significant dust source.
Size-class fractions grouped as clay (
Optical properties such as SSA are also dependent on the particle size
distributions (van Staveren et al., 1991; Moosmüller and Arnott, 2009) as well as
particle shape (Alexander et al., 2013) and mineralogy (Moosmüller et
al., 2012). For this project we measured scattering and absorption
coefficients for PM
Dust samples are composed of mineral mixtures in variable proportions, either individual mineral grains, intergrowths, coatings on other particles, or composite mineral clusters. Their chemical compositions can in many instances be related to their mineral constituents, as verified by optical microscopy, XRD, and SEM-based individual particle analysis.
The correlation matrix (Table 1) provides a general assessment of chemical
species associated with specific minerals in the suite of samples. The high
correlation of 0.91 between water soluble sodium (Na
Linear correlation coefficients between pairs of chemical species,
with pairs
Al is highly correlated with Fe, as well as Ti
and Mn, pointing to these elements occurring together or attached
to clay minerals such as montmorillonite and palygorskite. Also, the
correlation of Si with Al and K points to
the clay mineral illite or possibly potassium feldspar. The correlation
(0.42) between Ca and Mg is ascribed to these species
occurring together in the carbonate mineral dolomite (CaMg[CO
The scatter of CaO values shown in Fig. 16 is ascribed to the fact that
calcium occurs not only in calcite (CaCO
There is a reasonably strong linear correlation of 0.79 between Fe and
Al for the complete set of 165 samples (Table 1), with most (122)
samples falling in one oblong-shaped cluster with an average slope of 0.76
(Fig. 17). This subset of samples has an average Fe
Triangular-shaped field for the majority of the 130 dust
re-suspensions, with a maximum for Al
Plot showing the linear relationship between iron (Fe) and
aluminum (Al) for 130 chemically analyzed soil re-suspensions. Three sample
clusters can be identified with most (122) falling in one field (brown oval)
with an average Fe
A set of 28 Saharan dust samples collected from the Bodélé Depression
in Chad (Bristow et al., 2010) were analyzed for a number of chemical species
and have a Fe/Al mass ratio in the range of 0.53–1.71. In contrast, soil
samples collected along the southern Sahel region in Mali (Fig. 17) were
different (Roquin et al., 1990), with Fe
Average concentrations and elemental ratios for (a) PM
Single scattering albedo (SSA) at 405 nm for the PM
The high correlation between Fe and Al is attributed to the
co-existence of oxide minerals (hematite, goethite, magnetite) with various
clay minerals, chlorite, and other silicates, in varying proportions,
depending on factors of provenance, chemical weathering, and pedogenesis.
Comparisons of chemical results of re-suspended soils from this study and
those published on biotites (Deer et al., 1962) and clays (Weaver and
Pollard, 1973) are shown in Table 2. Clay minerals identified by XRD and SEM
include illite, palygorskite, and kaolinite, with montmorillonite in few
samples. The average Al concentrations of the re-suspended soils
(3.58 % for PM
To explain the elevated concentrations of Fe, Ti, Ca, K, and Fe
Sample S1056 from Lake Eyre, South Australia. SEM-generated backscattered electron image (BEI) and chemical maps of a quartz grain with a prominent coating. The coating is composed of an amorphous clay matrix containing colloidal particles of sheet silicates (clay, mica), oxyhydroxides (goethite, hematite), and adsorbed ions, explaining the uniform distribution of magnesium (Mg), silicon (Si), aluminum (Al), iron (Fe), and potassium (K).
SEM-based EDS spectrum and SEI of a particle from re-suspended
sample S1009 from Mali. The composite particle contains iron (Fe) as the
mineral goethite (FeO(OH)) and/or hematite (Fe
SSA for each of the re-suspended PM
As was previously shown (Moosmüller et al., 2012), there is a direct
relationship between the iron (expressed as Fe
As expected, pure hematite (S1000) has the lowest SSA (0.407) at
405 nm; this also holds true at 532 nm (SSA
The purpose of our study was to provide an understanding of the optical,
physical, mineralogical, and chemical interrelationships of a suite of
re-suspended dust samples. It can be argued that soil samples re-suspended
in the laboratory may be representative of surface wind-generated dust, part
of which (e.g., PM
However, it is recognized that dusts may be modified as they mix with other aerosols and age under changing meteorological conditions and over time, in some cases forming aggregates of larger particles. In our study, SEM-based SEIs of the re-suspended dust samples are evidence of such mineral clusters, as well as of coatings on quartz, feldspar, and other mineral grains, similar to that reported on in ambient aerosols (Cuadros et al., 2015; Díaz-Hernández and Párraga, 2008; Engelbrecht et al., 2009a; Jeong et al., 2014). The suite of surface soil samples investigated here is representative of many recognized dust source regions of the globe. At the time of this study, we did not have samples from all dust regions of North Africa, the Middle East, Asia, Australia, or South America.
A holistic approach (Smuts, 1986) to the analysis of mineral dusts provided us with a comprehensive understanding of their optical, physical, mineralogical, and chemical properties, and useful information that can be applied in remote sensing, modeling, global climate, health-related, and other studies.
The individual analytical procedures each provided important perspectives towards a better
understanding of mineral dusts. Laser-based particle size
analysis of the
Two mineralogically distinct phase assemblages are distinguished. The first
and largest group of samples is described as colloids in amorphous clay
(allophane) (Kars et al., 2015) clusters, surface coatings on particles, and
interstitial soils, characterized here by an average Fe
The SSA measured on the PM
Understanding the dust mineralogy, especially of the clay minerals and iron oxides, is of prime importance to our understanding of the optical properties of airborne mineral dusts. This paper greatly expands on our previous results from a limited sample set, showing a linear relationship between SSA and iron content (Moosmüller et al., 2009), here illustrating that this relationship is also controlled by sample mineralogy and mineralogical interrelationships.
Mineralogical, chemical, physical, and optical results are presented in tables and figures in the Supplement to this paper. Requests for data from this study can be directed to Johann Engelbrecht (johann@dri.edu) at the Desert Research Institute (DRI) in Reno, Nevada.
Johann Engelbrecht was the principal investigator and manager for the research project, responsible for the design of the program, configuring the hardware, experimental activities, data interpretation, compilation of results, and compiling this paper. Hans Moosmüller was the co-principal investigator for the project, formulating the problem, advising on the experimental design, interpreting results, and compiling parts of this paper. Samuel Pincock was responsible for a major part of the experimental work in the re-suspension laboratory, including the collection and validation of measurements. R. K. M. Jayanty from RTI International was responsible for the chemical analysis of filter samples. Gary Casuccio and Traci Lersch from RJ Lee Group were responsible for the SEM-based individual particle analysis.
The following persons or organizations provided grab soil samples for this program: US Army Public Health Command for samples from the Middle East, including Iraq (6), Kuwait (5), Afghanistan (3), Djibouti (1), Qatar (1), and UAE (1); Virginia Garrison for samples from Mali (3) and Cabo Verde (1); Edward Derbyshire for a sample from China (1); Earle Williams for samples from Chad (3); Samuel Marx for samples from Australia (3); Estelle Stegmann for samples from Namibia (3); Johann Engelbrecht for samples from USA (8), Canary Islands (8), Serbia (3), and Botswana (4); Henry Sun for a sample from Chile (1); Ross Edwards for samples from China (4); Michał Skiba for a sample from Morocco (1); Davis Zhu for samples from USA (2); Steven Kohl for a sample from USA (1).
We acknowledge analytical support from the laboratories of RTI International and RJ Lee Group. Greatly appreciated is technical support from the Desert Research Institute (DRI), including from Larry Sheets for help with the design and building of the re-suspension chamber, and Dave Campbell for installing and calibrating the equipment.
This project was funded under the US Army Medical Research and Materiel Command grant W81XWH-11-2-0220. Edited by: D. Knopf Reviewed by: R. Reynolds and one anonymous referee