Academia.eduAcademia.edu

Figure 7 – uploaded by Gábor Újvári

See full PDFdownloadDownload figure
Fig. 6. Measured and modeled ‘wet’ threshold friction velocity as a function of gravimetric soil moisture for a) loamy fine sand and sandy loam, and b) clay loam and clay soils. Measurements for different soils are from Selah and Fryrear (1995). The empirical model, U-pw = 0.305 + 0.022(0¢/0¢1.5) + 0.506(42/0¢1.5)°, of Selah and Fryrear (1995) is compared with theoretical models of Fécan et al. (1999) (Eqs. (3.13)-(3.14)) and Cornelis et al. (2004a) (Eqs. (3.15)-(3.16)). For all model calculations u., = 0.31 ms~! has been used, as published by Selah and Fryrear (1995) for the oven-dried soils without abrasion. The CGH (2004a) model is only given for the loamy fine sand soil, with parameters of D, = 130 um and surface tension of water at 15 °C, yu = 0.0735 N m_!. The rest of the parameters (Aco1, Aco2, Aco3) are as defined in the text.

Figure 6 Measured and modeled ‘wet’ threshold friction velocity as a function of gravimetric soil moisture for a) loamy fine sand and sandy loam, and b) clay loam and clay soils. Measurements for different soils are from Selah and Fryrear (1995). The empirical model, U-pw = 0.305 + 0.022(0¢/0¢1.5) + 0.506(42/0¢1.5)°, of Selah and Fryrear (1995) is compared with theoretical models of Fécan et al. (1999) (Eqs. (3.13)-(3.14)) and Cornelis et al. (2004a) (Eqs. (3.15)-(3.16)). For all model calculations u., = 0.31 ms~! has been used, as published by Selah and Fryrear (1995) for the oven-dried soils without abrasion. The CGH (2004a) model is only given for the loamy fine sand soil, with parameters of D, = 130 um and surface tension of water at 15 °C, yu = 0.0735 N m_!. The rest of the parameters (Aco1, Aco2, Aco3) are as defined in the text.

Related Figures (26)

Contents Loess deposits are recorders of aeolian activity during past glaciations. Since the size distribution of loess deposits depends on distance to the dust source, and environmental conditions at the source, during transport, and at de- position, loess particle size distributions and derived statistical measures are widely used proxies in Quaternary paleoenvironmental studies. However, the interpretation of these proxies often only considers dust transport processes. To move beyond such overly simplistic proxy interpretations, and toward proxy interpretations that consider the range of environmental processes that determine loess particle size distribution variations we pro- vide a comprehensive review on the physics of dust particle mobilization and deposition. Furthermore, using high-resolution bulk loess and quartz grain size datasets from a last glacial/interglacial sequence, we show that, because grain size distributions are affected by multiple, often stochastic processes, changes in these distri- butions over time allow multiple interpretations for the driving processes. Consequently, simplistic interpreta- tions of proxy variations in terms of only one factor (e.g. wind speed) are likely to be inaccurate. Nonetheless using loess proxies to understand temporal changes in the dust cycle and environmental parameters requires (i) a careful site selection, to minimize the effects of topography and source distance, and (ii) the joint use of bulk and quartz grain size proxies, together with high resolution mass accumulation rate calculations if possible. © 2016 Elsevier B.V. All rights reserved.
Contents Loess deposits are recorders of aeolian activity during past glaciations. Since the size distribution of loess deposits depends on distance to the dust source, and environmental conditions at the source, during transport, and at de- position, loess particle size distributions and derived statistical measures are widely used proxies in Quaternary paleoenvironmental studies. However, the interpretation of these proxies often only considers dust transport processes. To move beyond such overly simplistic proxy interpretations, and toward proxy interpretations that consider the range of environmental processes that determine loess particle size distribution variations we pro- vide a comprehensive review on the physics of dust particle mobilization and deposition. Furthermore, using high-resolution bulk loess and quartz grain size datasets from a last glacial/interglacial sequence, we show that, because grain size distributions are affected by multiple, often stochastic processes, changes in these distri- butions over time allow multiple interpretations for the driving processes. Consequently, simplistic interpreta- tions of proxy variations in terms of only one factor (e.g. wind speed) are likely to be inaccurate. Nonetheless using loess proxies to understand temporal changes in the dust cycle and environmental parameters requires (i) a careful site selection, to minimize the effects of topography and source distance, and (ii) the joint use of bulk and quartz grain size proxies, together with high resolution mass accumulation rate calculations if possible. © 2016 Elsevier B.V. All rights reserved.
Fig. 1. Typical loess/soil bulk and quartz particle size distributions from the studied profile at Dunaszekcs6 with the most widely used bulk loess grain size proxies. Samples shown are Dsz. GS-018 (loess, bulk/quartz: thin/bold black line), Dsz-GS-104 (loess, bulk/quartz: thin/bold red line) and Dsz-GS-290 (paleosol, bulk/quartz: thin/bold blue line). Abbreviations: M.sd. - medium sand, C.sd. — coarse sand, C.s. — coarse silt, Md grain size — median grain size, GSI — grain size index, PM — pipette method (based on Stokes sedimentation), LPS — laser particlé sizer (based on forward scattering of monochromatic coherent light). Size limits of clay, silt and sand fractions determined by laser particle sizer are different from those given by the pipette method (e.g. clay = <2 um for the PM, while it is <4.6/5.5 tum for LPS; Konert and Vandenberghe, 1997). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) The studied loess—paleosol section is located at Dunaszekcs6 (Fig. 2), Southern Hungary, on the right bank of the Danube river (46°05’25’N, 18°45'45’E, and 135 ma.s.l.) and exposes last glacial-interglacial sedi- ments with a thickness of 17 m. A detailed sedimentological description of the profile can be found in Ujvari et al. (2014). In 2008, an enormous bank failure exposed the uppermost 15-20 m part of the ca. 70 m thick Although the physical background of loess particle transport and de- position mechanisms was reviewed and discussed in detail in the late nineteen eighties by Pye (1987), Tsoar and Pye (1987) and Pye and Tsoar (1987), this knowledge does not always seem to be reflected in loess grain size proxy interpretations, which tend to be overly simplistic (Wang and Lai, 2014) and only consider transport mechanisms while essentially disregarding mobilization and deposition processes. As a
Fig. 1. Typical loess/soil bulk and quartz particle size distributions from the studied profile at Dunaszekcs6 with the most widely used bulk loess grain size proxies. Samples shown are Dsz. GS-018 (loess, bulk/quartz: thin/bold black line), Dsz-GS-104 (loess, bulk/quartz: thin/bold red line) and Dsz-GS-290 (paleosol, bulk/quartz: thin/bold blue line). Abbreviations: M.sd. - medium sand, C.sd. — coarse sand, C.s. — coarse silt, Md grain size — median grain size, GSI — grain size index, PM — pipette method (based on Stokes sedimentation), LPS — laser particlé sizer (based on forward scattering of monochromatic coherent light). Size limits of clay, silt and sand fractions determined by laser particle sizer are different from those given by the pipette method (e.g. clay = <2 um for the PM, while it is <4.6/5.5 tum for LPS; Konert and Vandenberghe, 1997). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) The studied loess—paleosol section is located at Dunaszekcs6 (Fig. 2), Southern Hungary, on the right bank of the Danube river (46°05’25’N, 18°45'45’E, and 135 ma.s.l.) and exposes last glacial-interglacial sedi- ments with a thickness of 17 m. A detailed sedimentological description of the profile can be found in Ujvari et al. (2014). In 2008, an enormous bank failure exposed the uppermost 15-20 m part of the ca. 70 m thick Although the physical background of loess particle transport and de- position mechanisms was reviewed and discussed in detail in the late nineteen eighties by Pye (1987), Tsoar and Pye (1987) and Pye and Tsoar (1987), this knowledge does not always seem to be reflected in loess grain size proxy interpretations, which tend to be overly simplistic (Wang and Lai, 2014) and only consider transport mechanisms while essentially disregarding mobilization and deposition processes. As a
Fig. 2. Location of the Dunaszekcs6 loess sequence in the Carpathian Basin.
Fig. 2. Location of the Dunaszekcs6 loess sequence in the Carpathian Basin.
Fig. 3. Scanning electron microscopy images of isolated quartz particles. a) Angular quartz grain from sample Dsz-GS-150, b) surface with v-shaped impact features of the same quart particle from sample Dsz-GS-150, c) Quartz grain with conchoidal fractures and v-shaped percussion cracks from sample Dsz-GS-293, d) Blow-up of the quartz surface with v-shape percussion cracks and breakage with sharp edges (sample Dsz-GS-293). To reach robust and solid interpretations of loess PSD variations in the context of paleoenvironmental changes, the nature and characteris- tics of major factors exerting control on grain size distributions must be clearly understood. Such influential factors include the main flow char- acteristics of the atmospheric boundary layer, the mobilization and transport modes of particles, and properties of soils and aeolian sur- faces. These issues will be reviewed in this section below. Note, howev- er, that in this paper we avoided discussing the production mechanisms of silt-sized mineral particles, but the interested reader is referred to The absolute frequency-dependent susceptibility, Yep = XLe — Xue, allows the determination of the concentration of magnetic particles over a small grain size window across the superparamagnetic (SP)/sta- ble single domain (SSD) boundary (Liu et al., 2012). By changing the ob- servation time (i.e. frequency) a fraction of SSD grains turn
Fig. 3. Scanning electron microscopy images of isolated quartz particles. a) Angular quartz grain from sample Dsz-GS-150, b) surface with v-shaped impact features of the same quart particle from sample Dsz-GS-150, c) Quartz grain with conchoidal fractures and v-shaped percussion cracks from sample Dsz-GS-293, d) Blow-up of the quartz surface with v-shape percussion cracks and breakage with sharp edges (sample Dsz-GS-293). To reach robust and solid interpretations of loess PSD variations in the context of paleoenvironmental changes, the nature and characteris- tics of major factors exerting control on grain size distributions must be clearly understood. Such influential factors include the main flow char- acteristics of the atmospheric boundary layer, the mobilization and transport modes of particles, and properties of soils and aeolian sur- faces. These issues will be reviewed in this section below. Note, howev- er, that in this paper we avoided discussing the production mechanisms of silt-sized mineral particles, but the interested reader is referred to The absolute frequency-dependent susceptibility, Yep = XLe — Xue, allows the determination of the concentration of magnetic particles over a small grain size window across the superparamagnetic (SP)/sta- ble single domain (SSD) boundary (Liu et al., 2012). By changing the ob- servation time (i.e. frequency) a fraction of SSD grains turn
Fig. 4. Relationship between yp (= Xie — Yur) and low field susceptibility (7p). xp is the background susceptibility. works by Smalley and Vita-Finzi (1968), Whalley et al. (1982), Wright and Smith (1993), Wright (1995, 2001), Smalley (1995), Assallay et al. (1998), Wright et al. (1998), Smith et al. (2002), and Muhs (2013). It must also be noted here that some of these mechanisms, such as frost weathering or granitoid weathering, produce quartz grains with differ- ent size characteristics. While frost weathering is capable of producing more silt-sized quartz grains, granitoid weathering profiles are often considered to be deficient in silt-sized material and rich in sand- and clay-sized particles (Wright, 2007). Subsequently, these grains are re- leased to sedimentary systems by glacial, fluvial and/or aeolian erosion and become the starting material of loess formation. Although the effect of these production mechanisms on loess PSDs is likely to be diminished by sorting during aeolian transport and depositional processes, they may still exert control on PSDs through sediment availability, as will be discussed below.
Fig. 4. Relationship between yp (= Xie — Yur) and low field susceptibility (7p). xp is the background susceptibility. works by Smalley and Vita-Finzi (1968), Whalley et al. (1982), Wright and Smith (1993), Wright (1995, 2001), Smalley (1995), Assallay et al. (1998), Wright et al. (1998), Smith et al. (2002), and Muhs (2013). It must also be noted here that some of these mechanisms, such as frost weathering or granitoid weathering, produce quartz grains with differ- ent size characteristics. While frost weathering is capable of producing more silt-sized quartz grains, granitoid weathering profiles are often considered to be deficient in silt-sized material and rich in sand- and clay-sized particles (Wright, 2007). Subsequently, these grains are re- leased to sedimentary systems by glacial, fluvial and/or aeolian erosion and become the starting material of loess formation. Although the effect of these production mechanisms on loess PSDs is likely to be diminished by sorting during aeolian transport and depositional processes, they may still exert control on PSDs through sediment availability, as will be discussed below.
Fig. 5. Measurements and models of threshold friction velocities required to initiate particle motion on dry sand surfaces. Models were run for quartz spheres (Pp.q = 2650 kg m~3). Some of the measurements of the fluid threshold were done on materials other than sand and dust (Fletcher, 1976; Iversen et al., 1976; Iversen and White, 1982), and for these the effect of different densities were taken into account by calculating the equivalent particle diameter (Dp-eqy = DpPp/Pp-a, Chepil, 1951). The fluid threshold was calculated with air parameters of pg = 1.225 kg m~? and vy = 1.47 x 10-° m? s~! (at 15 °C) in the Iversen and White (1982) model (for the equations the reader is referred to the original paper or Kok et al., 2012), while y = 3 x 10-4N m“! was used in the Shao and Lu (2000) model (Eq. 3.9). The Bagnold (1941) model is given by Eq. (3.7) in the text. Fi = BeDp + WnAc:
Fig. 5. Measurements and models of threshold friction velocities required to initiate particle motion on dry sand surfaces. Models were run for quartz spheres (Pp.q = 2650 kg m~3). Some of the measurements of the fluid threshold were done on materials other than sand and dust (Fletcher, 1976; Iversen et al., 1976; Iversen and White, 1982), and for these the effect of different densities were taken into account by calculating the equivalent particle diameter (Dp-eqy = DpPp/Pp-a, Chepil, 1951). The fluid threshold was calculated with air parameters of pg = 1.225 kg m~? and vy = 1.47 x 10-° m? s~! (at 15 °C) in the Iversen and White (1982) model (for the equations the reader is referred to the original paper or Kok et al., 2012), while y = 3 x 10-4N m“! was used in the Shao and Lu (2000) model (Eq. 3.9). The Bagnold (1941) model is given by Eq. (3.7) in the text. Fi = BeDp + WnAc:
with model parameters of Aco1, Aco2 and Aco; being 0.013 N m~!, 1.7 x 10-4N m7!, and 3 x 107 '4N~! m7! (Cornelis et al., 2004a,b). Wma is the matric potential at oven dryness (=— 10? MPa), 0, is gravi- metric water content, 6, is the gravimetric water content at —1.5 MPa, and y,; is the surface tension of water (Cornelis et al., 2003, 2004a,b). As shown in Fig. 6 which represents measured data and three models including that of Cornelis et al. (2004a), soil moisture significantly affects the threshold friction velocity of different soil sur- faces. Compared to their uniform oven-dried value of 0.31 ms~! the threshold friction velocities required to initiate particle motion for soils from loamy fine sand to clays are higher by ~20-220% for gravi- metric water contents ranging from 1.2 to 16.4% (Selah and Fryrear, 1995; Fig. 6a). While capillary forces caused by liquid-bridge bonding dominate at high soil moisture conditions, these forces are negligible at low humid- ity conditions (i, < —10 MPa; Tuller and Or, 2005), and thus adhesion forces are substantial for soils with either a high clay content and/or a low soil moisture content (Cornelis and Gabriels, 2003; Cornelis et al., 2004a; Ravi et al., 2004, 2006). In the model developed by Cornelis et al. (2004a), both the capillary and adhesion forces are included, and Usp is given by
with model parameters of Aco1, Aco2 and Aco; being 0.013 N m~!, 1.7 x 10-4N m7!, and 3 x 107 '4N~! m7! (Cornelis et al., 2004a,b). Wma is the matric potential at oven dryness (=— 10? MPa), 0, is gravi- metric water content, 6, is the gravimetric water content at —1.5 MPa, and y,; is the surface tension of water (Cornelis et al., 2003, 2004a,b). As shown in Fig. 6 which represents measured data and three models including that of Cornelis et al. (2004a), soil moisture significantly affects the threshold friction velocity of different soil sur- faces. Compared to their uniform oven-dried value of 0.31 ms~! the threshold friction velocities required to initiate particle motion for soils from loamy fine sand to clays are higher by ~20-220% for gravi- metric water contents ranging from 1.2 to 16.4% (Selah and Fryrear, 1995; Fig. 6a). While capillary forces caused by liquid-bridge bonding dominate at high soil moisture conditions, these forces are negligible at low humid- ity conditions (i, < —10 MPa; Tuller and Or, 2005), and thus adhesion forces are substantial for soils with either a high clay content and/or a low soil moisture content (Cornelis and Gabriels, 2003; Cornelis et al., 2004a; Ravi et al., 2004, 2006). In the model developed by Cornelis et al. (2004a), both the capillary and adhesion forces are included, and Usp is given by
Fig. 7. Particle terminal velocity as a function of grain size. Settling tube experimental data originate from Cui et al. (1983) and Malcolm and Raupach (1991). Newton's Impact Law (Eq. (3.24)), the Stokes Law (Eq. (3.26)), and the Ferguson and Church (2004) and Farrell and Sherman (2015) models are also shown. For calculating w; using the Impact and Stokes Laws quartz density of 2650 kg m~? and air density at 15 °C of 1.2256 kg m7? are used. Cy has been derived for Eq. (3.24) by solving Eqs. (3.27)-(3.28) numerically. For calculations using Eq. (3.26) a kinematic viscosity of air at 15 °C of 1.455 x 10-> m? s~! has been used and C, is computed following Rader (1990). Fall velocity has been calculated using the Ferguson and Church (2004) expression (their Eq. (4)) with parameters for natural sand of C; = 18 and C2 = 1. Terminal velocity calculations based on the Farrell and Sherman (2015) model (their Eq. (18)) are only given for the very fine to medium sand fractions, since this expression is valid only for the sand fraction. w; can be calculated numerically for the whole range of particle sizes in loess. Such a solution together with other expressions for w, and mea- sured values of natural sands are shown in Fig. 7. As can be seen from this figure and demonstrated by Eqs. (3.24) and (3.26) w, is proportion- al to Dj for small particles (Rep: < 1) and D,’? for larger (Shao, 2008). In fact, the fall velocity is influenced by both the aerodynamic properties of particles (surface area, shape, density) and the fluid properties (density and viscosity; Malcolm and Raupach, 1991; Chen and Fryrear, 2001; Farrell and Sherman, 2015). For two particles having the same size and shape, but different density the heavier (more dense) will settle
Fig. 7. Particle terminal velocity as a function of grain size. Settling tube experimental data originate from Cui et al. (1983) and Malcolm and Raupach (1991). Newton's Impact Law (Eq. (3.24)), the Stokes Law (Eq. (3.26)), and the Ferguson and Church (2004) and Farrell and Sherman (2015) models are also shown. For calculating w; using the Impact and Stokes Laws quartz density of 2650 kg m~? and air density at 15 °C of 1.2256 kg m7? are used. Cy has been derived for Eq. (3.24) by solving Eqs. (3.27)-(3.28) numerically. For calculations using Eq. (3.26) a kinematic viscosity of air at 15 °C of 1.455 x 10-> m? s~! has been used and C, is computed following Rader (1990). Fall velocity has been calculated using the Ferguson and Church (2004) expression (their Eq. (4)) with parameters for natural sand of C; = 18 and C2 = 1. Terminal velocity calculations based on the Farrell and Sherman (2015) model (their Eq. (18)) are only given for the very fine to medium sand fractions, since this expression is valid only for the sand fraction. w; can be calculated numerically for the whole range of particle sizes in loess. Such a solution together with other expressions for w, and mea- sured values of natural sands are shown in Fig. 7. As can be seen from this figure and demonstrated by Eqs. (3.24) and (3.26) w, is proportion- al to Dj for small particles (Rep: < 1) and D,’? for larger (Shao, 2008). In fact, the fall velocity is influenced by both the aerodynamic properties of particles (surface area, shape, density) and the fluid properties (density and viscosity; Malcolm and Raupach, 1991; Chen and Fryrear, 2001; Farrell and Sherman, 2015). For two particles having the same size and shape, but different density the heavier (more dense) will settle
Fig. 8. Transport modes of quartz grains as a function of friction velocity and particle diameter. The most widely used grain size proxies are also shown in this context. Terminal velocities are from Ferguson and Church (2004) with parameters as defined in Fig. 7. Transport modes with different w, / u. values as limits are from Gillette et al. (1974), Hunt and Nalpanis (1985), Nalpanis (1985), Tsoar and Pye (1987) and Shao (2008). Typical friction velocities in dust storms originate from Tsoar and Pye (1987), Li and Zhang (2011), while typical Mg grain size of loess are from Tsoar and Pye (1987), Derbyshire et al. (1995), Pye (1995), Shi et al. (2003), Ding et al. (2005), Prins et al. (2007), Yang and Ding (2008), Varga et al. (2012), and Vandenberghe (2013). Transport modes of particles are determined by the balance between the terminal velocity and the mean Lagrangian vertical velocity at which air parcels are dispersed upward by turbulence (Shao, 2008). Under neutral atmospheric conditions the latter is approximately Ku. (Raupach and Lu, 2004, Shao, 2008) and particles tend to move disper- sively, i.e. in suspension, if w, « Ku.. Pure suspension, in which particles essentially move with the fluid, occurs when the particle's terminal ve- locity is small compared to the friction velocity (Tsoar and Pye, 1987). Gillette et al. (1974) have found that the upper limit of pure suspension is W; / Us & 0.12 to 0.7, while Tsoar and Pye (1987) further subdivided this for short-term (0.1 < w, / u- < 0.7) and long-term suspension re- gimes (w, / u- < 0.1; Fig. 8). The residence time of long-term suspended dust (<~20 tum) particles may be days or weeks and they can be transported thousands of kilometers from source regions (Tsoar any Pye, 1987, Pye, 1987). Pure saltation, when the particle trajectories are not affected by the vertical turbulent velocity fluctuations, occurs from w, / u» > ~1-2 (Tsoar and Pye, 1987; Shao, 2008; Nickling and McKenna Neuman, 2009), and these larger particles saltate following ballistic trajectories (Ungar and Haff, 1987; Anderson and Haff, 1988). A sharp boundary between saltation and pure suspension does not exist and particles having semi-random trajectories (~70-150 um; Anderson, 1987; Kok et al., 2012) influenced by both inertia and termi- nal velocity are considered to be transported in modified saltation (0.7 < w; / u» < 1; Hunt and Nalpanis, 1985; Nalpanis, 1985). As seen in Fig. 8, the transport modes of particles having different size are strongly dependent on the friction velocity, i.e. the intensity of atmo- spheric turbulence. For instance, a quartz particle with a diameter of 70 jum may be transported in saltation or modified saltation in flows with weak wind shear and turbulence, as well as transported in short- term suspension in airflows with strong friction and turbulence.
Fig. 8. Transport modes of quartz grains as a function of friction velocity and particle diameter. The most widely used grain size proxies are also shown in this context. Terminal velocities are from Ferguson and Church (2004) with parameters as defined in Fig. 7. Transport modes with different w, / u. values as limits are from Gillette et al. (1974), Hunt and Nalpanis (1985), Nalpanis (1985), Tsoar and Pye (1987) and Shao (2008). Typical friction velocities in dust storms originate from Tsoar and Pye (1987), Li and Zhang (2011), while typical Mg grain size of loess are from Tsoar and Pye (1987), Derbyshire et al. (1995), Pye (1995), Shi et al. (2003), Ding et al. (2005), Prins et al. (2007), Yang and Ding (2008), Varga et al. (2012), and Vandenberghe (2013). Transport modes of particles are determined by the balance between the terminal velocity and the mean Lagrangian vertical velocity at which air parcels are dispersed upward by turbulence (Shao, 2008). Under neutral atmospheric conditions the latter is approximately Ku. (Raupach and Lu, 2004, Shao, 2008) and particles tend to move disper- sively, i.e. in suspension, if w, « Ku.. Pure suspension, in which particles essentially move with the fluid, occurs when the particle's terminal ve- locity is small compared to the friction velocity (Tsoar and Pye, 1987). Gillette et al. (1974) have found that the upper limit of pure suspension is W; / Us & 0.12 to 0.7, while Tsoar and Pye (1987) further subdivided this for short-term (0.1 < w, / u- < 0.7) and long-term suspension re- gimes (w, / u- < 0.1; Fig. 8). The residence time of long-term suspended dust (<~20 tum) particles may be days or weeks and they can be transported thousands of kilometers from source regions (Tsoar any Pye, 1987, Pye, 1987). Pure saltation, when the particle trajectories are not affected by the vertical turbulent velocity fluctuations, occurs from w, / u» > ~1-2 (Tsoar and Pye, 1987; Shao, 2008; Nickling and McKenna Neuman, 2009), and these larger particles saltate following ballistic trajectories (Ungar and Haff, 1987; Anderson and Haff, 1988). A sharp boundary between saltation and pure suspension does not exist and particles having semi-random trajectories (~70-150 um; Anderson, 1987; Kok et al., 2012) influenced by both inertia and termi- nal velocity are considered to be transported in modified saltation (0.7 < w; / u» < 1; Hunt and Nalpanis, 1985; Nalpanis, 1985). As seen in Fig. 8, the transport modes of particles having different size are strongly dependent on the friction velocity, i.e. the intensity of atmo- spheric turbulence. For instance, a quartz particle with a diameter of 70 jum may be transported in saltation or modified saltation in flows with weak wind shear and turbulence, as well as transported in short- term suspension in airflows with strong friction and turbulence.
Fig. 9. Measured and modeled streamwise saltation flux as a function of friction velocity. Wind tunnel data for the transport rate of 230 and 242 «wm diameter sands are from Iversen and Rasmussen (1999) and Sorensen (2004) (transport-limited situation), while for non-cohesive, clay and salt crusted surfaces (all three undisturbed) are from Macpherson et al. (2008) (supply-limited situations). Saltation mass flux is calculated for sand with a diameter of 250 pm and an air density of 1.2256 kg m~? (at 15 °C) using model equations of 3.30-3.32 and parameters as defined in the text (Section 3.1.1.8). These theoretical models of Bagnold (1941), Kawamura (1951), Duran et al. (2011) and Kok et al. (2012) are proposed for transport-limited situations.
Fig. 9. Measured and modeled streamwise saltation flux as a function of friction velocity. Wind tunnel data for the transport rate of 230 and 242 «wm diameter sands are from Iversen and Rasmussen (1999) and Sorensen (2004) (transport-limited situation), while for non-cohesive, clay and salt crusted surfaces (all three undisturbed) are from Macpherson et al. (2008) (supply-limited situations). Saltation mass flux is calculated for sand with a diameter of 250 pm and an air density of 1.2256 kg m~? (at 15 °C) using model equations of 3.30-3.32 and parameters as defined in the text (Section 3.1.1.8). These theoretical models of Bagnold (1941), Kawamura (1951), Duran et al. (2011) and Kok et al. (2012) are proposed for transport-limited situations.
Fig. 10. Vertical dust flux as a function of shear velocity due to a) direct aerodynamic entrainment and b) saltation bombardment. In panel a), long-term dust flux measurements and the best-fit model are from Loosemore and Hunt (2000) (=LH2000), while the rest of the data are from experiments of Macpherson et al. (2008) (=M2008) carried out on supply-limited non-cohesive, clay crusted and salt crusted surfaces. In panel b), field measurements during different erosion events are from Gomes et al. (2003) (=G2003) and Sow et al. (2009) (=$2009), while the bold lines are Eqs. (3.37)-(3.38), and represents models of Gillette and Passi (1988) (=GP1988) and Shao et al. (1993) (=S1993). The model by Kok et al. (2012) (=K2012) is given as Fy,,=QU»(u.?-u%), which uses u,, = 0.20 m s~! and is normalized to yield 10,000 ug m~* s~! atu, =1ms7!.
Fig. 10. Vertical dust flux as a function of shear velocity due to a) direct aerodynamic entrainment and b) saltation bombardment. In panel a), long-term dust flux measurements and the best-fit model are from Loosemore and Hunt (2000) (=LH2000), while the rest of the data are from experiments of Macpherson et al. (2008) (=M2008) carried out on supply-limited non-cohesive, clay crusted and salt crusted surfaces. In panel b), field measurements during different erosion events are from Gomes et al. (2003) (=G2003) and Sow et al. (2009) (=$2009), while the bold lines are Eqs. (3.37)-(3.38), and represents models of Gillette and Passi (1988) (=GP1988) and Shao et al. (1993) (=S1993). The model by Kok et al. (2012) (=K2012) is given as Fy,,=QU»(u.?-u%), which uses u,, = 0.20 m s~! and is normalized to yield 10,000 ug m~* s~! atu, =1ms7!.
Fig. 12. a) Contributions to total collision efficiency between a raindrop (D,;q = 1 mm) and dust particles and b) total collision efficiency (€,) for different rain droplet sizes. Collection efficiencies from Brownian diffusion (Egy), interception (E;yw) and inertial impaction (Ejmw) are calculated based on the Slinn (1984) model with modifications to include thermophoresis (E7py), diffusiophoresis (Eppy), and electrostatic mechanisms (Egsw) after Davenport and Peters (1978), Andronache (2004) and Andronache et al. (2006). Parameters (15 °C): pp = 1000 kg m-*, po = 1.225 kg m3, by = 1.783 x 10°-°kgm~!s~1, vy = 1.455 x 10-9 m2 s~!, uw, = 1.139 x 10-7 kg m-! 57}, Ca = 1005 J kg~! K~!, kg = 0.02534 J m=! s-! K7!, Ty - Tras = 3 °C, Dw = 2.35 x 107° m? s~!, RH = 80%, des = 2.
Fig. 12. a) Contributions to total collision efficiency between a raindrop (D,;q = 1 mm) and dust particles and b) total collision efficiency (€,) for different rain droplet sizes. Collection efficiencies from Brownian diffusion (Egy), interception (E;yw) and inertial impaction (Ejmw) are calculated based on the Slinn (1984) model with modifications to include thermophoresis (E7py), diffusiophoresis (Eppy), and electrostatic mechanisms (Egsw) after Davenport and Peters (1978), Andronache (2004) and Andronache et al. (2006). Parameters (15 °C): pp = 1000 kg m-*, po = 1.225 kg m3, by = 1.783 x 10°-°kgm~!s~1, vy = 1.455 x 10-9 m2 s~!, uw, = 1.139 x 10-7 kg m-! 57}, Ca = 1005 J kg~! K~!, kg = 0.02534 J m=! s-! K7!, Ty - Tras = 3 °C, Dw = 2.35 x 107° m? s~!, RH = 80%, des = 2.
Fig. 13. Measured and modeled wet scavenging coefficients as a function of particle size and rain rate (J). The model is based on work by Slinn (1983), Loosmore and Cederwall (2004), Seinfeld and Pandis (2006), Davenport and Peters (1978), Andronache (2004) and Andronache et al. (2006). Parameters as defined in the caption of Fig. 12.
Fig. 13. Measured and modeled wet scavenging coefficients as a function of particle size and rain rate (J). The model is based on work by Slinn (1983), Loosmore and Cederwall (2004), Seinfeld and Pandis (2006), Davenport and Peters (1978), Andronache (2004) and Andronache et al. (2006). Parameters as defined in the caption of Fig. 12.
Fig. 14. Calculated changes in the volume size distribution of the remote continental model aerosol after various rainfall durations (t,) at a rain rate (J) of 10 mm hr.~!. The model aerosol distribution parameters are from Jaenicke (1993): N; = 3200 cm~?, Nz = 2900 cm~?, N3 = 0.3 cm~?, Dp; = 0.02 pum, D2 = 0.116 pm, Dp3 = 1.8 pm, log 0; = 0.161, log 0; = 0.217, log 0; = 0.380.
Fig. 14. Calculated changes in the volume size distribution of the remote continental model aerosol after various rainfall durations (t,) at a rain rate (J) of 10 mm hr.~!. The model aerosol distribution parameters are from Jaenicke (1993): N; = 3200 cm~?, Nz = 2900 cm~?, N3 = 0.3 cm~?, Dp; = 0.02 pum, D2 = 0.116 pm, Dp3 = 1.8 pm, log 0; = 0.161, log 0; = 0.217, log 0; = 0.380.
Fig. 15. Bulk and quartz grain size proxy variations as a function of depth in the Dunaszekcs6 sequence. Infra Red Stimulated Luminescence (IRSL) ages are pIR-IRSLy2s (first number) anc pIR-IRSL299 (second number) ages as published and defined in Ujvari et al. (2014). Ages at depths of 10 and 14.9 mare yet unpublished age data, both are pIR-IRSLy99 ages. Legend: 1. loess 2. recent soil, 3. weathered loess, 4. red-brown, well-developed pedocomplex, 5. IRSL sampling points. Black arrows denote in-phase coarse GS events between bulk and quartz grain size proxies.
Fig. 15. Bulk and quartz grain size proxy variations as a function of depth in the Dunaszekcs6 sequence. Infra Red Stimulated Luminescence (IRSL) ages are pIR-IRSLy2s (first number) anc pIR-IRSL299 (second number) ages as published and defined in Ujvari et al. (2014). Ages at depths of 10 and 14.9 mare yet unpublished age data, both are pIR-IRSLy99 ages. Legend: 1. loess 2. recent soil, 3. weathered loess, 4. red-brown, well-developed pedocomplex, 5. IRSL sampling points. Black arrows denote in-phase coarse GS events between bulk and quartz grain size proxies.
Fig. 16. Internal relationships of median diameter, U-ratio and grain size index in the Dunaszekcs6 sequence.
Fig. 16. Internal relationships of median diameter, U-ratio and grain size index in the Dunaszekcs6 sequence.
Fig. 17. Relationships between bulk loess and quartz grain size proxies in the studied sequence.
Fig. 17. Relationships between bulk loess and quartz grain size proxies in the studied sequence.
Fig. 18. Relationships between coarse and fine fractions in the U-ratio and the grain size index (GSI) in the Dunaszekcs6 sequence.
Fig. 18. Relationships between coarse and fine fractions in the U-ratio and the grain size index (GSI) in the Dunaszekcs6 sequence.
Fig. 19. Minimally and fully dispersed particle size distributions of three loess samples from the studied section.
Fig. 19. Minimally and fully dispersed particle size distributions of three loess samples from the studied section.
Sweeney, M.R., Mason, J.A., 2013. Mechanisms of dust emission from Pleistocene loess de- posits, Nebraska, USA. J. Geophys. Res. 118, 1460-1471.
Sweeney, M.R., Mason, J.A., 2013. Mechanisms of dust emission from Pleistocene loess de- posits, Nebraska, USA. J. Geophys. Res. 118, 1460-1471.
Related topics:
SedimentologySedimentary geology and stratigraphyQuaternary environmentsGrain size distributionEolian dustParticle Size Studies
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved