Dust transport and deposition in the Southern Ocean
Figure 7 : Vertical distribution of the meridional mean annual dust concentration (µg.m-3) over the Southern Ocean between 50°S and 75°S based on modelling results. Source:
After the emission, dust is sent to high altitude and can be transported over long distance to the open ocean. The modelling study of Li et al. (2008) simulated the vertical distribution of the meridional mean annual dust concentration (µg.m-3) in the troposphere over the Southern Ocean (50°S and 75°S) (Figure 7). Maximum values are observed at 60° – 0°W and 160° – 180°E, corresponding to the longitude of South America and Australia, respectively, with an eastward shift of 30° as a result of the continuous westerly winds. The maximum dust concentration corresponding to South America is found from ground level to 600 mb, whereas the maximum dust concentration corresponding to Australia is located at high altitude between 800 mb and 600 mb, which could be explained by the long-range transport of dust emitted from Australia. Dust export from Southern Africa seems to be negligible compared to South America and Australia, according to Figure 7. Compared to the dust originated from Australia, Patagonian dust travels at low altitude over the Southern Ocean (Gasso and Stein, 2007; Johnson et al., 2011).
Dust are deposited into surface seawater by dry deposition (mostly by gravitational settling and turbulent deposition) and by wet deposition (Mahowald et al., 2005; Rosenfeld et al., 2014; Gao et al., 2003; Slinn, 1976; Textor et al., 2006). Along the atmospheric transport, large dust particles are preferentially removed by the gravitational sedimentation. Particles larger than 10 µm generally have a lifetime for a few hours and most dust particles are deposited rapidly near the source area. Although large dust particles (>100 µm) were observed to transport for a long distance in previous studies (Betzer et al., 1988), long-range transported dust generally has a mode diameter around 2.5~3.5 µm (Schulz et al., 1998; Maring et al., 2003). Dust particles can be removed through washout process by precipitation or can act as cloud condensation nuclei (CCN) and through in-cloud scavenging (Rosenfeld et al., 2014; Berthet et al., 2010; Chate et al., 2011; Chate and Pranesha, 2004; Chate et al., 2003).
In the past decades, both field measurements and modeling studies have contributed to the quantification of mineral aerosol deposition in Southern Ocean. Wagener et al. (2008) estimated total dust deposition flux in Southern Ocean based on shipboard measurement of atmospheric aerosol concentration. Dry deposition was calculated as the product of aerosol concentration and deposition velocity. Wet deposition fluxes were derived from aerosol concentration and precipitation data by defining a fixed wet scavenging ratio (SR). Wagener et al. (2008) finally found that dust fluxes are up to two orders of magnitude lower than previous model predictions (Mahwald et al., 2005) and extrapolations of land-based measurements (Duce et al., 1991). The later studies of Heimburger et al. (2012) at Kerguelen Island (49°18’S, 70°07’E) and Chance et al. (2015) in the southeastern Atlantic Ocean sampled simultaneously atmospheric aerosol and rainwater to determine the dry deposition fluxes, wet deposition fluxes and ultimately the total deposition fluxes. The two studies found much higher deposition fluxes than those estimated by Wagener et al. (2008), and show better agreements with dust modelling results (Johnson et al., 2010; Mahowald et al., 2005; Mahowald et al., 2007). Particularly, the study of Heimburger et al. (2012) and Chance et al. (2015) indicated that he scavenging ratio (SR) used by Wagener et al. (2008) to estimate the deposition flux is too low and hence the computation of wet deposition based on atmospheric aerosol concentration seriously underestimated the wet dust deposition flux. Estimating the deposition flux from atmospheric aerosol concentration will introduce serious bias. On the other side, another study of Grand et al. (2015) estimated total dust deposition from the mixed layer concentration of dissolved aluminum (dAl) in the eastern Indian Ocean by assuming a steady state between addition of dAl due to the dissolution of atmospheric deposition and the removal of dAl via particulate scavenging. Grand et al. (2015) finally found similar total deposition flux to the modelling results of Mahowald et al. (2005) in the Southern Ocean.
However, although the measurements taking into account the contribution of wet deposition (Chance et al., 2015; Grand et al., 2015; Heimburger et al., 2012) show better agreement with previous modelling studies, discrepancies still exist between these studies. Being the only available long term time series of deposition flux in marine locations far from the dust sources in subantarctic region, the two-year measurements of dust deposition at Kerguelen and Crozet islands display a seasonal pattern with higher deposition fluxes in austral winter and spring (Heimburger et al., 2012; Heimburger et al., 2013). The time scale of sampling and the sporadic nature of dust fluxes can be a source of uncertainty and variability between observations (Grand et al., 2015).
Briefly, according to previous studies, wet deposition is the dominant mechanism to deposit atmospheric particles into the Southern Ocean. The dominant wet deposition means that most of dust particles deposited into the Southern Ocean is incorporated into rainwater, which may have further impact on the bioavailability of elements in dust for the marine ecosystem.
Mineral Dust as Micronutrient Supplier
For the HNLC Southern Ocean, dust deposition is supposed to be an important source of micronutrients. Hence, the key issue for the marine ecosystem is trace elements that dust contains rather than dust particles themselves. Elemental composition and bioavailability of elements are two key issues of dust that determine the amount of elements assimilated by marine ecosystem.
Elemental composition of mineral dust
As the product of wind erosion of soils, mineral dust contains several chemical elements that could be important for the biological processes of marine ecosystem. Measuring the dust elemental compositions is important to estimate the emission inventory of trace elements from dust source areas and evaluate the biological impact of dust input (Baker et al., 2003; Zhang et al., 2015).
Chemical compositions of dust can differ from its parent soil since the dust materials contain only the fine fraction of soil particles. In bulk soils, large particles, especially the sandy fraction, are dominated by quartz and calcium containing materials (e.g. calcite and gypsum), whereas smaller particles contain clay minerals, feldspars, quartz, micas, carbonates and iron oxides (Journet et al., 2014; Schütz and Rahn, 1982). Quartz contains mainly Si; Clay minerals are mainly composed of Si and Al; Fe content is found principally in clays, feldspars and iron oxides; Ca and Mg exist mostly in gypsum, calcite and dolomite (Journet et al., 2014). The variation of chemical compositions with mineral species results in finally size dependence of elemental concentrations in desert soils (Eltayeb et al., 1993; Eltayeb et al., 2001; Castillo et al., 2008; Schütz and Rahn, 1982; Miller et al., 1972). For example, Schütz and Rahn (1982) studied African and American soils and found that elemental concentrations for most elements, except for Si, increase to the highest when the particle size decreases to 20 µm. This increase is greater in higher weathered and more winnowed soils, and is negligible in humus-rich soils. Eltayeb et al. (1993) found a nearly constant concentration of Al, K, Sr and Rb, a positive fractionation for the elements Ca, Ti, Mn, Fe, Y, and a negative fractionation for Si and Zr in aerosol fraction of Namibian soils. Therefore, both major elements and trace elements exhibit preferential partitioning with size fractions. As a result of different fractionation behavior of elements, elemental composition dust may differ from the chemical composition of parent soil.
Previous modelling studies generally take the average iron concentration of Earth’s crust (3.5%) (Taylor and McLennan, 1995) as the iron content in dust (e.g. Duce and Tindale, 1991; Luo et al., 2008). However, spatial heterogeneities of dust elemental composition have been shown in former studies. For example, Formenti et al. (2008) found an average iron content equaling to 8.6 ± 0.2% (mean ± std) for local dust and 7.6 ± 0.6% for transported dust in Banizoumbou, Niger. Different iron content in dust from local source and dust from remote sources reflects regional variability of iron content. In addition, both dust sources showed much higher iron compositions than the values used by dust models. Iron composition of dust fallout measured by Gaiero et al. (2007) at four sites in Patagonian coast, despite the fact that dust deposited closing source areas is different from dust transported for long distance, also showed higher iron concentrations (4.3 ± 0.6%) than values used by models. Spatial variability of dust elemental composition must be taken into account to better evaluate the emission inventory of trace elements associated with dust.
Because dust chemical compositions are different from bulk soils and vary with the emission regions, investigations into the elemental composition of dust from sources are necessary. Considering that many active dust exist as “hot spots” in small areas, which is quite common in the Southern Hemisphere, rather than in large homogeneous dust emission areas (Gillette, 1999), these kinds of investigations should be done at smaller scale. A database of dust elemental composition will be quite useful to evaluate the emission inventory of trace elements from dust sources and to reduce the uncertainties in dust modelling (Zhang et al., 2015).
Bioavailability of trace elements in dust
After deposited into seawater, only a fraction of trace elements in dust is bioavailable for the marine biota, where “being bioavailable” means being effective in causing a biological effect on the phytoplankton. For elements like Fe, the processes making iron bioavailable are complicated and different forms of iron are bioavailable, but all bioavailable iron are in dissolved phase including colloidal phase or soluble phase (Barbeau et al., 2001; Rich and Morel, 1990; Fitzsimmons and Boyle, 2014). Although not all forms of dissolved iron are bioavailable (Visser et al., 2003), the iron bioavailability is generally evaluated by the common known “fractional solubility” (hereafter “solubility”) that is defined as the percentage ratio of dissolved amount to the total amount.
Factors controlling the solubility of micronutrients in mineral dust: the case of iron
For marine ecosystem, the bioavailability of micronutrients associated with dust depends on multiple factors: 1) the mineralogical composition of source dust, 2) the chemical processing history of dust during atmospheric transport, 3) the deposition process of dust into the ocean, 4) the composition of seawater (Baker and Jickells, 2006; Baker et al., 2006b; Desboeufs et al., 1999; Journet et al., 2008; Paris and Desboeufs, 2013; Gierlus et al., 2012; Losno et al., 1991; Schulz et al., 2012; Shi et al., 2012). The following content will discuss the impact of these factors with a focus on iron.
Mineralogical composition of dust may influence the solubility of elements in dust. As indicated by Journet et al. (2008), solubility of iron-containning minerals can vary by different orders of magnitude, particularly iron contained by clays generally show much higher solubility than iron (hydr-)oxide. Despite the variation of solubility among Ca-containing minerals (Krueger et al., 2004; Chou et al., 1989), some minerals such as calcium carbonate in dust can act as alkalinity buffers and neutralize the acidic conditions in cloud droplets or rainwater during atmospheric transport (Losno et al., 1991; Loye-Pilot et al., 1986) and hence prevent the enhancement of solubility by atmospheric acid processing that is presented below.
Atmospheric processing is suggested to result in greater uncertainty of bioavailability of iron in dust (Shi et al., 2012). Previous studies generally found fractional iron solubility less than 0.5% for non-atmospheric processed dust but ranging from 0.1% to ~90% for transported aerosol (Mahowald et al., 2005; Hand et al., 2004; Chen and Siefert, 2004; Baker and Jickells, 2006; Sedwick et al., 2007; Heimburger et al., 2013a). The difference of solubility suggests an enhancement of dust solubility by atmospheric processing (Shi et al., 2012). During the atmospheric transport, dust particles can incorporate into cloud droplets as cloud condensation nuclei (CCN) or as interstitial particles (Andreae and Rosenfeld, 2008; Gierlus et al., 2012). Cloud water is an effective medium for heterogeneous chemical reactions. Nitrate and sulfate produced by the oxidation process by H2O2, O3, O2 and NO2 acidify the cloud waters (Rosenfeld et al., 2014). Enhanced acidity of cloud water finally enhanced the elemental solubility of dust particles (Spokes and Jickells, 1995; Spokes et al., 1994; Desboeufs et al., 2001; Desboeufs et al., 2005). In addition to the acid processing in cloud, acid processing as wet particles outside cloud can also enhance the iron solubility of dust (Spokes and Jickells, 1995; Spokes et al., 1994; Desboeufs et al., 2001; Desboeufs et al., 2005; Shi et al., 2015). After the evaporation of cloud droplets, pH values of the water content in dust particles could decrease to 2 or even lower (Meskhidze et al., 2003; Zhu et al., 1992). As indicated by the recent study of Shi et al. (2015), the highly acidic wet particles outside cloud resulting from the evaporation of cloud droplets is the main process increasing readily dissolved iron during atmospheric processing. Furthermore, the exposition of dust particles to the solar ration in the presence of acidic solutions can also enhance the solubility of iron due to the photoreduction reaction (Hand et al., 2004; Fu et al., 2010) that converts the relatively insoluble Fe(III) into the more soluble Fe(II) (Kieber et al., 2005). Fu et al. (2010) found that dust in HCl solution showed higher increase of iron solubility under irradiation compared to the dark reaction.
Dry and wet deposition process can affect differently the solubility of mineral aerosol. Baker and Jickells (2006) argued that the specific surface area of mineral aerosol particles is the primary controlling factor of aerosol iron solubility. The dry deposition process removes preferentially larger particles and consequently results in higher solubility of mineral aerosol.
Table of contents :
Chapter 1 Background, Significance and Approaches of Research
1. Dust Emission processes
2. Sources, Transport and Deposition of Mineral Dust to the Southern Ocean
2.1. Distribution and contribution of dust sources in the Southern Ocean
2.1.1. Distribution of dust sources
2.1.2. Contribution of dust sources in the Southern Ocean
2.2. Dust transport and deposition in the Southern Ocean
3. Mineral Dust as Micronutrient Supplier
3.1. Elemental composition of mineral dust
3.2. Bioavailability of trace elements in dust
3.2.1. Factors controlling the solubility of micronutrients in mineral dust: the case of iron
3.2.2. Common methods of elemental solubility estimation
4. Research Topics and Strategies
4.1. Research topics
4.2. Research Strategies
4.2.1. Long-term dust concentration measurements in Patagonia
4.2.2. Spatial heterogeneity of source dust elemental compositions
4.2.3. Some aspects of the solubility of continental dust
Chapter 2 Long-term dust concentration measurements in Patagonia
2. Materials and methods
2.1. Aerosol sampling location and methods
2.2. Elemental analysis
2.3. Chemical compositions of the crustal fraction of the aerosol
2.4. Air mass back trajectories
2.5. Wind simulation and meteorological records
3. Results and discussion
3.1. Chemical composition of the dust fraction
3.2. Atmospheric concentration of sea salt and mineral dust
3.3. Seasonal pattern of the aerosol concentration
3.4. Meteorological dependence of seasonal dynamics of dust concentrations and emission
Chapter 3 Spatial Heterogeneity of source dust compositions
2. Study area
2.1. Patagonia Desert
2.2. Namibia: Namib Desert and Kalahari Desert
3. Materials and methods
3.1. Soil-derived aerosol generation
3.2. Soil sample collection
3.3. Elemental analysis
3.4. Principal component analysis of compositional data
3.5. Accumulation factor and enrichment factor of dust relative to parent soil
4. Results and discussion
4.1. Elemental composition of soil and aerosol
4.1.1. Element concentration of topsoil and soil-derived dust in Patagonia and Namibia
4.1.2. Spatial variation of elemental composition in regional scale
4.1.3. Robust principle component analysis
4.2. Variation of elemental composition from bulk soil to aerosol
Chapter 4 Contribution to Bioavailability Study of Mineral Dust from Patagonia and Namibia
2. Materials and methods
2.1. Mineral aerosol samples
2.2. Dissolution experiments of aerosol sample
2.3. Centrifugation separation of suspension
2.4. Chemical analysis
3. Results and discussion
3.1. Comparison of solubility values between centrifugation and filtration
3.2. Variation of solubility with elements and its dependence on pH
3.3. Dependence of solubility on types of dust sample
Conclusions and prospects