Evaluation of the electrical properties of dust storms by multi-parameter observations and theoretical calculations

Earth and Planetary Science Letters 461 (2017) 141–150

Contents lists available at ScienceDirect

Earth and Planetary Science Letters www.elsevier.com/locate/epsl

Evaluation of the electrical properties of dust storms by multi-parameter observations and theoretical calculations Huan Zhang a,b , Tian-Li Bo b , Xiaojing Zheng a,∗ a b

School of Mechano-Electronic Engineering, Xidian University, Xi’an, 710071, China Key Laboratory of Mechanics on Western Disaster and Environment, Lanzhou University, Lanzhou, 730000, China

a r t i c l e

i n f o

Article history: Received 13 August 2016 Received in revised form 21 December 2016 Accepted 1 January 2017 Available online xxxx Editor: T.A. Mather Keywords: dust storms electrical properties electric field dust concentration early warning

a b s t r a c t Dusty phenomena, such as wind-blown sand, dust devils, and dust storms, play key roles in Earth’s climate and geological processes. Dust electrification considerably affects the lifting and transport of dust particles. However, the electrical properties of dust storms remain poorly understood. Here, we conducted multi-parameter measurements and theoretical calculations to investigate the electrical properties of dust storms and their application to dust storm prediction. The results show that the vertical electric field (E-field) decreases first, then increases, and finally decreases with the height above the ground, reversing its direction at two heights, ∼8–12 and ∼24 m. This suggests that the charge polarity of dust particles changes from negative to positive and back to negative again as the height increases. By carefully analyzing the E-field and dust concentration data, we further found that there is a significant positive linear relationship between the measured E-field intensity and dust concentration at the given ambient conditions. In addition, measurements and calculations demonstrate that a substantial enhancement in the vertical E-field can be observed several hours before the arrival of the external-source dust storms, indicating that the E-field can be used to provide an early warning of external-source dust storms. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Dust storms frequently occur in arid and semi-arid areas and can cause serious damage to the economy, environment and human health (Shao and Dong, 2006; Akhlaq et al., 2012). Therefore, achieving a clear understanding of the physical processes involved in dust storms and high accuracy early warning of dust storms can considerably minimize these hazards. Preliminary dust storm Efield measurements have shown upward vertical E-fields with values of 5–10 kV/m (Rudge, 1913). Subsequent measurements have detected both downward and upward vertical E-fields, with values of ∼50–200 kV/m at the ground level (e.g., Demon et al., 1953; Stow, 1969; Williams et al., 2009). Recently, comprehensive threedimensional E-field measurements recorded both the upward vertical E-field and downwind streamwise E-field (Bo and Zheng, 2013). Moreover, the magnitude of the streamwise E-field was larger than that of the vertical E-field, reaching up to ∼60 kV/m at a height of several meters above the ground. Measurements and theoretical models have suggested that the intense E-field had pronounced effects on the trajectories of saltating particles (Zheng et al., 2003; Kok and Renno, 2008), the evolution of wind-blown

*

Corresponding author. E-mail address: [email protected] (X. Zheng).

http://dx.doi.org/10.1016/j.epsl.2017.01.001 0012-821X/© 2017 Elsevier B.V. All rights reserved.

sand (Zheng et al., 2006), and the lifting of dust particles from the ground (Kok and Renno, 2006, 2008; Esposito et al., 2016). To date, the vertical E-field was measured at only one height (Rudge, 1913; Kamra, 1972; Williams et al., 2009; Esposito et al., 2016) or at four heights ranging from 0.4 to 4 m (Bo and Zheng, 2013). The measured vertical E-field was upward (Rudge, 1913; Stow, 1969; Bo and Zheng, 2013) or downward (Demon et al., 1953; Esposito et al., 2016) at a near-ground height of ∼2 m. The reason for this difference in the measured E-field direction is still unclear. In addition, the measured vertical E-field decreased rapidly with height in the saltation layer (Schmidt et al., 1998), but initially increased and then decreased with height in the region of 0.4–4 m (Bo and Zheng, 2013). The entire profile of the vertical E-field in dust storms has not been completely investigated. Existing technologies that are used for dust storm detection and early warning can be classified into two categories, namely, traditional ground-based observations and spaceborne observations. The former method is generally based on monitoring the wind speed, temperature, and dust concentration, etc. However, it is difficult to forecast upcoming dust storms because the large increases in these quantities can only be measured when the dust storms arrive. For the latter method, the position, propagation direction, and average propagation speed of dust storms are determined by analyzing the continuous monitoring of satellite images and thus can

142

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

Fig. 1. (a) An overview and (b) accurate configuration of the instrument array at Minqin, Qingtu Lake, China. All instruments were mounted on an observed tower (32 m in height). The CSAT3B and DustTrak II Aerosol Monitor were mounted at heights of 0.9, 1.71, 2.5, 3.49, 5, 7.15, 8.5, 10.24, 14.65, 20.96, and 30 m above the ground. To measure the E-field in the near-ground region (<0.9 m), the VREFMs were mounted at heights of 0.1, 0.2, 0.3, 0.4, 5, 7.15, 8.5, 10.24, 14.65, 20.96, and 30 m. The SPC-91 and dust particle collector were mounted at 0.1 and 5 m, respectively.

issue an early warning to the downwind regions. Nevertheless, this method is also limited by the low spatial and temporal resolution of satellite imaging devices. For example, GOES and GERB provide data with a temporal resolution of up to ∼15 min, but the spatial resolution is as low as ∼10 km. The poor spatial resolution is insufficient to determine the accurate location of the leading edge. By contrast, Quickbird has a high spatial resolution (∼0.65 m) but low temporal resolution (revisit time is 1–3.5 days). Thus, it is ineffective at providing real-time monitoring of dust storms. As discussed above, the existing traditional methods have some limitations. Unlike wind speed and the dust concentration, the Efield is something real that extends throughout a volume of space, and therefore the E-field can be detected at a distance between the source and measurement point. Although the existing measurements of dust storms cannot directly confirm long-distance electrical effects (e.g., Bo and Zheng, 2013), measurements in other natural granular systems, such as dust devils (e.g., Freier, 1960; Crozier, 1964) and volcanic plumes (Rakov and Uman, 2003; Mather and Harrison, 2006; Aizawa et al., 2016), are well-documented. If the exterior E-field is large enough in dust storms, detection of dust storms can be achieved by monitoring strong increases in the atmospheric E-field. However, whether this method is feasible remains unknown. In this study, we present a series of dust storm E-field measurements up to a height of 30 m above the ground in Minqin, China. Based on the measured results, we also propose a simple Efield model. The objectives of this study are as follows: (1) reveal the vertical E-field profile; (2) obtain the relationship between the dust concentration and vertical E-field; (3) determine the chargeto-mass ratios of airborne dust particles; and (4) discuss whether the prediction method based on monitoring the E-field is feasible for a dust storm early warning system. 2. Field measurements Measurements were taken in the flat-bottomed dry lakebed of the Qingtu Lake (longitude: 103◦ 40 03 ; latitude: 39◦ 12 27 ), approximately 90 km northeast of Minqin, Gansu, China. This field site is located between the Badain Jaran Desert and Tengger Desert, which are source areas of sand and dust particles. Measurements

were performed continuously during the spring season from March to June 2015. Measurements of the wind velocity, ambient temperature, vertical E-field, mass concentration of PM10 (dust particles with diameter of <10 μm), number density of saltating particles, and particle size distribution (PSD) of airborne dust particles were taken simultaneously. The accurate configuration of the instrument array is shown in Fig. 1b. The wind velocity and ambient temperature were measured by sonic anemometer (model CSAT3B, Campbell Scientific), and the PM10 dust concentrations were measured by DustTrak II Aerosol Monitor (Model 8530EP, TSI Incorporated). The vibrating-reed electric field mill (VREFM) developed by Lanzhou University was used to measure the E-field. The number density of saltating sand particles in the range of 36–490 μm (with 32 bin steps) was obtained using sand particle counter (SPC-91, Niigata Electric Co., Ltd.). Measurements were performed at a sampling rate of 1 Hz (expect CSAT3B with 50 Hz). In addition, the PSD of total airborne dust particles was determined by measuring the PSD of sample dust particles collected in a dust particle collector mounted at a height of 5 m above the ground. It is worth noting that the DustTrak II Aerosol Monitor measures dust concentrations based on the principle of light scattering, and thus the measurements are highly dependent on the particle size and material properties. The 8530EP Monitor is factory calibrated to the reparable fraction of the standard ISO 12103-1 A1 ultrafine test dust, which has a nominal 0–10 micron size. In the present study, we use the factory defaults calibration factor (i.e., User Cal is 1.0) to measure the PM10 dust concentration in the dust events because the aerosols are very similar to ISO 12103-1 A1 test dust. Therefore, the measurement error could be considered to be negligible. Additionally, to obtain accurate measurements for other specific aerosols, such as PM1 and PM2.5 , an additional calibration experiment is required and implemented following Morawska et al. (2003) and Zhou et al. (2016). The working principle of VREFM is based on detecting the charge induced on the sensor electrode (Zheng, 2013). As the electrode oscillates, it charges and discharges periodically, and the electric current is linearly proportional to the electric field intensity. The VREFM is calibrated using a large (1 m square plates) parallel-plate electric field calibrator. The results of the laboratory

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

and field calibrations demonstrate that the VREFM can accurately measure electric fields ranging from −200 to 200 kV/m with a resolution of <0.1 kV/m. More detailed descriptions of the VREFM and dust particle collector are shown in the Text S1 of the supplementary materials. 3. E-field theory 3.1. Interior E-field For the purpose of further discussing the electrical properties of dust storms, a simple E-field theoretical model is also proposed in the present study. We refer to the E-field inside the dust storms and far from the edges as the interior E-field and the E-field outside the dust storms as the exterior E-field. Under natural conditions, sand/dust transport generally occurs in turbulent flows, thus causing high spatial and temporal variability (Durán et al., 2011; Kok et al., 2012). This variability is one of the factors leading to the generation of the streamwise E-field (Bo and Zheng, 2013; Zhang et al., 2013, 2014). However, sand/dust transport averaged over a time scale of ∼10 min is approximately constant in the steady stage of dust storms (see Fig. S5 in the Supplementary Materials). In this time-averaged steady transport, the concentration of charged dust particles is constant with time and horizontal distance. In this section, we are mainly concerned with long time-averaged transport in which the horizontal variability (e.g., dust concentration and space charge density) is not included and therefore only consider the mean vertical E-field. The E-field in dust storms is produced by charged sand/dust particles (grounded sand particles and airborne dust particles) and atmospheric ions. Since the streamwise and spanwise spatial dimensions of dust storms are much greater than its vertical extent, the infinite horizontal plane is a good approximation of the E-field far from the edges. Hence, the vertical interior E-field E z in the time-averaged steady state can be expressed as (see Text S2 in the Supplementary Materials for details)

E z ( z) = −0.21 −

1

ε0

h

 

ρ z dz



(1)

dz

=

ρ ( z) ε0

ρ (z )dxdy dz σs dxdy  r1 + r2 3 4πε0 r1 4πε0 r23

(3)

where r1 = (−x , − y  , z − z ) and r2 = (−x , − y  , z) are the vectors in the direction from the source point p 1 and p 2 toward the observation site p, respectively; r1 = |r1 |, r2 = |r2 |. During the dust storms, dust/sand particles exchange charge by undergoing multiple collisions or contacts. Under the time-averaged uniform and steady conditions, the charge is approximately conserved locally in an air column. Hence, the surface charge density can be expressed as (Zheng et al., 2004; Zhang et al., 2014)

h

 

ρ z dz

σs = −

(4)

z0

To obtain the exterior E-field, Eq. (4) is substituted into Eq. (3), and then the equation is integrated over (−∞, −a] × (−∞, +∞) × z [ z0 , h], which gives the vertical exterior E-field E ext as z E ext ( z , a)

= −0.21 −

+

1 2πε0

h

1 2ε0

h

 

ρ z dz

z

  

ρ z

  arctan

a z

 − arctan

z0

a z − z



dz (5)

where a is the horizontal streamwise distance between the observation site p (0, 0, z) and leading edge of the dust storm. The first term on the right-hand side of Eq. (5) also represents the fairweather atmospheric E-field. In addition to the vertical exterior E-field, a streamwise exterior E-field also exists due to the asymmetry of the dust particles in the streamwise direction. Similarly, x the streamwise exterior E-field E ext can be expressed as x E ext ( z , a) =

1 4πε0

h

 



ρ z ln

z0

a2 + z 2 a2 + ( z − z )2



dz

(6)

z

where ε0 = 8.85 × 10−12 is the permittivity of the vacuum; h is the maximum height that the dust particles reach; and ρ ( z) is the space charge density. The first term on the right-hand side of Eq. (1) explains the downward fair-weather atmospheric E-field produced by the atmospheric ions, and the second explains the E-field produced by charged dust particles. Equation (1) indicates that E z ( z) is a function of the total electric charge above the height z. From Eq. (1), the derivative of E z with respect to z is

dE z ( z)

dE =

143

(2)

This means that the monotonicity of E z depends only on the sign of ρ ( z). That is, if ρ ( z) is positive (negative) at every height of interval (z, z +  z), then E z is increasing (decreasing) over that interval. 3.2. Exterior E-field To find the exterior E-field, consider a space charge element,

ρ (z )dx dy  dz , located at point p 1 (x , y  , z ), and a ground surface charge element, σs dx dy  , located at point p 2 (x , y  , 0), where ρ (z ) and σs are the space and surface charge density, respectively. From Coulomb’s Law, the E-field produced by the space and surface charge elements at the observation site p (0, 0, z) is

4. Results 4.1. Measurement results During the field campaign, we have successfully recorded 24 dust events, including several long-term dust storms and a large number of short-term blowing dust events. An overview of these dust events is given in Table 1. Even though the properties of the dust events differed substantially from event to event, they could be separated into three different categories based on the wind direction. From a total of 24 dust events, 6 and 14 events had a stable wind direction of 245◦ and 45◦ , respectively (see Fig. 2c). In another 4 events, the wind directions varied widely with time (e.g., between 3.7◦ and 290.9◦ in the No. 4 event). Additionally, during the field measurement campaign, the observed E-field at 20.96 m was generally downward, with occasional and short reversals of the E-field direction in the unstable wind direction events, as shown in Fig. 2b. Examples of the complete time series of measured wind speed, PM10 concentration, saltating particle number density, E-field intensity, ambient temperature, and wind direction of the long-term dust storms originating on 29 March are shown in Figs. 3a–3f, respectively. The wind speed gradually increased starting at 09:30 on 29 March and reached the fluid threshold (the lowest wind speed at which ground sand particles can be directly lifted by wind

144

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

Table 1 Properties of the dust events. The grey-shaded region indicates the long-term external-source dust storm, which occurred from 29 March to 30 March. T denotes the ambient temperature. Values labeled “Max wind” refer to the maximum streamwise wind speed at a height of 5 m. Values labeled “Max PM10 ” refer to the maximum hourly mean PM10 dust concentration at a height of 1.71 m. Values labeled “Max E-field” refer to the maximum magnitude of hourly mean E-field at a height of 20.96 m. In several events, the PM10 dust concentration could not be measured (N/M). No. of dust events

Start date

Duration (h)

T (◦ C)

Wind direction (deg.)

Max wind (m/s)

Max PM10 (mg/m3 )

Max E-filed (kv/m)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

28 29 17 19 20 22 26 28 29 06 07 08 09 10 11 12 13 14 15 16 17 22 27 30

6.9 22.9 1.0 3.6 2.9 3.2 5.3 9.7 11.3 6.7 8.4 11.1 9.1 14.1 8.2 10.7 11.3 8.9 12.4 8.8 9.6 3.5 10.9 9.1

19.1–26.7 7.7–24.7 12.5–13.8 20.2–23.1 9.8–14.5 22.7–26.9 26.5–30.3 19.2–32.1 21.9–34.8 18.6–24.5 12.9–27.1 12.1–18.5 16.4–25.3 11.6–31.2 12.5–22.6 16.5–33.8 16.3–31.5 23.5–28.3 15.1–28.1 24.5–33.2 20.9–28.9 17.1–28.9 14.9–30.8 21.1–32.7

273.7–319.8 274.7–302.5 12.9–38.1 3.7–290.9 147.9–340.6 197.3–248.3 43.5–62.6 233.8–262.3 156.4–211.4 41.1–65.8 43.7–80.4 26.4–56.5 22.6–119.2 51.7–282.6 34.1–77.1 43.5–73.2 22.6–45.9 29.2–59.9 281.7–336.9 93.3–116.6 16.2–46.8 77.3–93.6 12.2–38.5 32.3–66.8

6.7 11.6 9.9 10.8 9.6 7.7 8.2 8.1 4.9 8.4 10.8 12.5 11.8 11.4 3.8 5.6 7.1 10.4 7.8 5.1 15.4 5.1 10.9 9.1

0.21 1.66 0.45 1.71 0.41 0.52 0.24 0.29 0.15 N/M N/M 1.95 N/M 1.83 N/M 0.12 0.23 0.74 0.19 0.99 4.7 0.24 0.68 0.50

8.13 46.6 71.8 35 6.6 21.1 14.5 21.1 16.1 36.7 43.6 46.1 40.2 23.5 19.5 8.6 47.6 13.6 3.1 18.9 50.1 39.3 18.1 46.2

Mar. Mar. Apr. Apr. Apr. Apr. Apr. Apr. Apr. May May May May May May May May May May May May May May May

Fig. 2. The complete time series of the measured (a) dust concentration and (b) E-field at 20.96 m during the field measurement campaign from March to June 2015. The dust events are denoted by the vertical arrows. Two dashed rectangles indicate the reversals of the E-field direction (i.e., upward). (c) The wind direction of the 24 observed dust events are shown; the black triangles denote 4 dust events having an unstable wind direction, and the other symbols denote 20 dust events having a stable wind direction.

shear stress) at approximately 10:24. From 10:24 to 12:14, the sand and dust particles were transported in the near-ground region of <2.5 m (Fig. 3c) because of the very low dust concentration at 2.5 m (Fig. 3b). The dust storm arrived at 12:14 and then remained at an approximately steady state from 12:14 to 13:24, with a maximum PM10 concentration of 0.48 mg/m3 (2.5 m), and gradually weakened from 13:24 to 18:30. After 18:30, the dust storm strengthened slowly with frequent fluctuations and the mean PM10 concentration reached 1.5 mg/m3 (2.5 m) at 03:52 on 30 March. Finally, the dust storm vanished at approximately 10:00 on 30 March. It is clear that we can reasonably classify the dust storm into two stages: the mild stage, which lasted from 12:14 to 18:30 on 29 March, and the severe stage, which lasted from 18:30 on 29 March to ∼10:00 on 30 March. Additionally, the wind direction remained over a narrow range of 274.7◦ to 302.5◦ during this dust

storm (Fig. 3f), suggesting that the dust particles were lifted from the same source area. Figs. 4a and 4b show examples of the measured average vertical E-field profiles for the dust events with stable wind directions of 245◦ and 45◦ , respectively. As seen in Figs. 4a and 4b, different categories of dust events exhibit different vertical profiles in detail, but all of the events follow a similar profile in which the E-field gradually decreases and then increases with height, and two direction reversals of the E-field exist. That is, the E-field reverses first at the lower height for the events with a wind direction of 245◦ (∼8 m) than for the events with a wind direction of 45◦ (∼12 m), but it reverses again at the same height (∼24 m) for both two stable wind direction categories. This complex vertical profile is quite different from the monotonic and upward E-field of wind-blown sand (Schmidt et al., 1998; Kok and Renno, 2008) due to the presence of highly charged air-

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

145

Fig. 3. Examples of the measured time series: (a) streamwise wind speed U 5 at 5 m; (b) PM10 concentration M 10 at 2.5 m; (c) saltating particle number density N s at 0.1 m; (d) E-field E z at 20.96 m; (e) ambient temperature T at 5 m; and (f) hourly mean wind direction θ at 5 m of the long-term dust storms originating on 29 March, where θ denotes the angle between the wind flow and the South. Vertical dashed lines show the two stages: the mild stage (between two dashed lines) and the severe stage (after the second dashed line).

Fig. 4. Examples of the measured vertical E-field profiles for the dust events having mean wind directions of (a) 245◦ and (b) 45◦ .

borne dust particles. To determine the relationship between the E-field and dust concentration, we define the height average of the PM10 dust concentration  M 10  in the region of [z1 , z2 ] as

 M 10  =

1 z2 − z1

z2 M 10 ( z)dz

(7)

z1

because the E-field directly depends on the whole distribution of the charged dust particles. The results of our data analysis show that a significant positive linear relationship exists between the Efield intensity and dust concentration. Note that the magnitude of the E-field increases linearly with the PM10 concentration but the slopes vary from event to event even at the different temperature stages of one event (Fig. 5). As an example, for the mild stage of a dust storm that occurred from 29 March to 30 March, with a temperature from 19.1 to 24.7 ◦ C, the increase rate of the E-field intensity with respect to the average dust concentration at 20.96 m was 123.7 kV m2 /mg, but it decreased to 28.4 kV m2 /mg in the severe stage (7.7–19.1 ◦ C). We also compared the detection time between the E-field and dust cloud to investigate the long-distance electrical effects of dust storms. Based on the dust source regions all dust events can

be divided into two categories: (1) local-source events in which dust particles are lifted from the observation site; (2) externalsource (or long-range transport) events in which dust particles are transported from other source regions. During the field measurement campaign, only one external-source event was observed on 29 March (i.e., the No. 2 event in Table 1), and the others were found to be local-source events. Figs. 6a and 6b show examples of the difference in the detection time between the dust concentration and E-field of local- and external-source events, respectively. For local-source events, once the dust particles had lifted, the E-field increased immediately. Thus, the enhancements in the dust concentration and E-field were recorded nearly simultaneously. For external-source events, however, the E-field changes occurred even when the dust cloud was far from the observation site. Thus, large increases in the E-field occur prior to the arrival of the dust cloud (e.g., No. 2 event). That is, from 10:24 to 12:14 on 29 March, an intense E-field was observed (one order of magnitude larger than the fair-weather atmospheric E-field), but fewer dust particles were transported to the higher altitudes (>2.5 m), indicating that there was a strong enhancement in the E-field of approximately ∼1.83 h before the arrival of the dust storm (Fig. 6b). During this period, there just existed saltation process in the near ground region. The E-field produced by

146

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

Fig. 5. Examples of the quantitative relationship between the 30 minutes time-averaged PM10 mass concentration  M 10  and E-field for the different dust events on (a) 28 March, (d) 10 May, (e) 14 May and (f) 30 May, and at the different temperature stages of (b) 19.1–24.7 ◦ C and (c) 7.7–19.1 ◦ C in the event occurring from 29 March to 30 March. The black squares, green triangles, red circles, and blue diamonds denote the measured data at heights of 10.24, 14.65, 20.96, and 30 m, respectively. The lines show the linear fit (i.e., E z = k M 10 ) to the measured data with correlation coefficient R 2 values of ∼0.74–0.97 and p < 0.0001). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the saltating particles decreased rapidly with height, and became zero at a few centimeters from the ground (Schmidt et al., 1998; Kok and Renno, 2008). Therefore, the observed intense E-field was mainly caused by the charged dust cloud far from the observation site rather than the saltating sand cloud. Here, we provide direct observational evidence that it is possible to detect external-source dust storms (dust storms affecting residential and urban areas are generally external-source) by monitoring the variation of the atmospheric E-field. To some extent, E-field monitoring is able to predict dust storms earlier than traditional ground-based technologies.

4.2. Theoretical results In this section, the E-field theoretical model is used to determine the relationship between the horizontal distance a and exterior E-field. To do this, we should first calculate the chargeto-mass ratio of airborne dust particles as a key parameter of the E-field model. A large number of studies found negatively charged saltating particles (on average) in the saltation layer, with a charge-to-mass ratio of ∼−60 μC/kg (Schmidt et al., 1998; Zheng et al., 2003; Bo et al., 2014). However, it is difficult to mea-

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

147

Fig. 6. Ten minutes time-averaged series of the dust concentration and magnitude of the E-field for the different measured heights on (a, b) 14 May, 2015 and (c, d) 29 March, 2015. The light grey shading indicates the period of E-field enhancement before the arrival of dust storm.

20  .96

M ( z)dz + c 3

8.26

20.96

30

+∞

c2

M ( z)dz + c 3

23.87

sure the charge-to-mass ratio of airborne dust particles directly because dust particles are difficult to capture in the Faraday cup (Merrison, 2012; Harrison et al., 2016). Here, we determined the charge-to-mass ratio of airborne dust particles by calibrating the measured E-field with the theoretical model. The PSD of total airborne dust particles is essential to calculate the charge-to-mass ratios (see below), and it is difficult to measure the PSD precisely in short-term dust events due to the sparsity of collected dust particles. Consequently, we only evaluated the charge-to-mass ratios of airborne dust particles from 12:14 to 13:24 on 29 March (in a long-term dust storm lasting ∼22.9 h). Combining Eq. (2) with the measured E-field profile, we conclude that the airborne dust particles are negatively charged below 20.96 m, but positively charged from 20.96 to 30 m. In our field observations, the E-field above 30 m was not directly measured. However, according to Eq. (1), the total electric charge above 30 m must be negative because the vertical E-field at 30 m is upward. Therefore, a reasonable assumption is that the dust particles above 30 m are negatively charged, which is confirmed by the following agreement between our model predictions and measurements (see Fig. 7). To determine the charge-to-mass ratio of airborne dust particles, let c 1 , c 2 and c 3 be the charge-to-mass ratios of dust particles corresponding to the regions of 0–20.96 m, 20.96–30 m, and >30 m, respectively. The E-field intensity is zero at 8.26 and 23.87 m, and hence from Eq. (1) we have

+∞

M ( z)dz + c 2

c1

Fig. 7. Comparison of the measured (open squares) and theoretical calculated E-field (solid line) in the period from 12:14 to 13:14 on 29 March, 2015 as a function of height. The light grey shading denotes the uncertainty of c 1 = −1142 ± 256 μC/kg.

30

M ( z)dz = −0.21ε0 (8a) 30

M ( z)dz = −0.21ε0

(8b)

30

Thus, we obtain c 2 = −8.2c 1 and c 3 = 4.7c 1 . In Eq. (8), M ( z) is the mass concentration of total airborne dust particles, which is determined by the measured PM10 mass concentration and PSD of airborne dust particles (see Text S3 in the Supplementary Materials). By calibrating the E-field theoretical model with the measurements, we obtain c 1 of approximately −1142 ± 256 μC/kg in the period from 12:14 to 13:24 on 29 March 2015 (see Fig. 7). This charge-to-mass ratio corresponds to dust electrification values of ∼4.28 × 10−14 C/grain, which is consistent with the laboratory results of 10−16 –10−14 C/grain (Merrison, 2012). Equations (5) and (6) are respectively the quantitative analytical expressions for the vertical and streamwise exterior E-fields of dust storms. Using Eqs. (5) and (6), the variation of the exterior E-fields with streamwise distance a can be obtained. We define the horizontal distance, which is sufficient for measuring the Efield with a magnitude of 1 kV/m as acr . As shown in Fig. 8, with increase of height, the acr of the exterior streamwise E-field decreases first and then increases, with a maximum value of 5.5 km at heights from 1.5 to 1000 m. However, the acr of the vertical exterior E-field is approximately constant with height, with a value of up to ∼28 km. The theoretical calculations also demonstrate that substantial E-field enhancement occurs in advance of the arrival of dust storms. 5. Discussion 5.1. Charge structure in the dust events In the present study, we first found the decrease–increase– decrease vertical E-field profiles in dust storms and blowing dust events. According to Eq. (2), we conclude that the charge polarity of dust particles changes from negative to positive and back to negative again as the height increases. Previous studies explained the E-field in wind-blown sand (e.g., Schmidt et al., 1998; Kok and Renno, 2008) and dust devils (e.g., Freier, 1960; Jackson and Farrell, 2006; Lacks and Sankaran, 2011) by a negative electric dipole model (i.e., monopolar dust clouds), which was expected to create a monotonic and upward E-field. However, we show here

148

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

x Fig. 8. (a) The streamwise exterior E-field E ext as a function of streamwise distance a for different heights above the ground. The solid and dashed lines denote the upwind z E-field at the higher elevation and the downwind E-field at the lower elevation, respectively. (b) The vertical exterior E-field E ext as a function of streamwise distance a for the different heights above the ground. In this case, the charge-to-mass ratios are taken as the calculated values in Fig. 7, whereas the mass concentration of the airborne PM10 dust particles M ( z) is taken as the average concentration during the period of 12:14–13:14 on 29 March, 2015, respectively. The dark oblique line with the slope of z −6/5 indicates that E ext can be approximately expressed as 40.3a−6/5 in the range of distance a greater than ∼3 km.

that the dust storms and blowing dust events consist of three distinct charge regions, with a bottom negative charge region, a middle positive charge region and a top negative charge region. It is generally accepted that the primary charging mechanism for dust particles in dusty phenomena is contact electrification, especially asymmetric contact (Lacks and Sankaran, 2011; Hu et al., 2012). In this process, smaller (larger) particles tend to be negatively (positively) charged (Forward et al., 2009; Kok and Lacks, 2009; Lacks and Sankaran, 2011; Yoshimatsu et al., 2016). Therefore, the net charge held by a dust particle depends on the relative frequency of collisions with other sand and dust particles and could be both positive and negative. During dust events, the charge transfer process is thought to involve Earth’s surface charging and volume charging, and surface charging is generally expected to result in monopolar dust clouds, whereas volume charging is generally expected to create multipolar dust clouds (Williams et al., 2009). Our measured results confirm that volume charging is dominant in dust storms. Interestingly, the multipolar charge structure is also common in thunderclouds and volcanic plumes (Mather and Harrison, 2006; Dwyer and Uman, 2014). In addition to dust storms and blowing dust events, we did not observe dust devils (i.e., spikes in the E-field, with wind and dust data lasting several minutes) during our field measurements, although several studies have reported 30–50 encounters per 100 days (Lorenz and Jackson, 2015; Harrison et al., 2016). This difference is presumably caused by environmental factors, such as weather, soil and land-surface conditions. Further measurements are required to understand the E-field and charge structure in dust devils. 5.2. Factors affecting the E-field The results of our study showed that there was a significant positive linear relationship between the E-field intensity and dust concentration at a certain ambient temperature (see Fig. 5), and that the slopes varied widely in different events, implying that the dust particle properties (e.g., grain composition and PSD of dust particles) and ambient temperature might have a strong impact on the dust charging process. This positive correlation was also reported in a recent study with only one measurement height of 2 m. That is, Esposito et al. (2016) found a linear trend between the E-field and dust concentration and that the slope was influenced by the ambient relative humidity, i.e., the E-field intensity decreased with the ambient relative humidity at a specific dust

concentration. In the present study, due to the absence of available ambient relative humidity data, we did not show the effects of ambient relative humidity on the E-field. In fact, the ambient relative humidity is closely related to the ambient temperature. For example, Bo and Zheng (2013) reported that a strong negative correlation existed between the ambient relative humidity and ambient temperature in a dust storm. Therefore, it is difficult to determine each single effect on the E-field in the field measurements, and further research is required. 5.3. Prediction based on E-field precursor phenomena As discussed in section 4, both observation data and a theoretical calculation revealed that there is a strong enhancement in the vertical E-field prior to external-source dust storms, providing a new method for dust storm prediction and early warning. To our knowledge, this is the first direct observational evidence of the long-distance electrical effects in dust storms. The theoretical calculations demonstrate that the vertical exterior E-field is one order of magnitude larger than that of the streamwise exterior E-field, and the reason for this is as follows: the positively charged dust particles on the ground produce upward vertical exterior E-fields and downwind streamwise exterior E-fields, whereas the negatively charged airborne dust particles (on average) also produce upward vertical exterior E-fields but upwind streamwise exterior E-fields. Note that the horizontal variability in dust transport is neglected in our E-field model, and this simplification may underestimate the actual value of the horizontal exterior E-field, because such horizontal variability is one of the factors producing the horizontal E-field (Bo and Zheng, 2013; Zhang et al., 2014). Consequently, the horizontal E-field is possibly more useful than the vertical E-field for dust storm prediction. However, since the actual horizontal E-field is usually much more complex than the vertical E-field, the prediction is better established by the vertical E-field. Furthermore, we suggest that one straightforward method for dust storm prediction might be to set a group of increasing Efield thresholds (i.e., E th,1 < E th,2 < · · · < E th,i < · · · , which are better defined from the local statistical observed data) according to ground-based E-field observations. If the observed E-field intensity continuously exceeds the increasing E-field thresholds, a warning should be submitted to decision makers. This will give decision makers sufficient time to exclude false alarms and take any precautionary actions. An accurate and real-time prediction

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

Fig. 9. A ground-based E-field observation array. The m E-field sensors are installed in a line with a spacing of l, where θ represents the angle between the dust cloud propagation direction and the aligned observation array.

is usually based on well-established empirical or statistical laws, such as the volume charge structure, dust concentration profile, and dust particle composition, etc. Fig. 9 shows a linear arrangement of the ground-based E-field observation array. Suppose that m E-field sensors are mounted at the same height z . Then, according to Eq. (5) we have





z E ext z , a − E 1 = 0





z z , a + l cos θ − E 2 = 0 E ext



.. .





z E ext z , a + (m − 1)l cos θ − E m = 0

height above the ground, with two direction reversals at ∼8–12 and ∼24 m. By combining with a theoretical model, our findings firstly showed that the dust storms and blowing dust events consisted of three distinct charge regions, with a bottom negative charge region, middle positive charge region and top negative charge region. There was a strong linear relationship between the measured E-field and dust concentration at the given ambient conditions, and the slopes varied from event to event even in the different temperature stages of one event. The charge-to-mass ratios of dust particles were determined by calibrating the theoretical model with the measured E-field, and we found that c 1 was approximately equal to −1142 ± 256 μC/kg during the period from 12:14 to 13:14 on 29 March 2015. Additionally, a strong enhancement in the E-field was observed approximately 1.83 h before the arrival of the external-source dust storm. Therefore, it is possible to detect a dust storm several hours earlier based on monitoring the variation of the atmospheric E-field than by using traditional ground-based technologies. The multi-level E-field thresholds can be used to reduce the false alarm rate, and a ground-based E-field observation array is established to provide real-time monitoring. Note that the wellestablished empirical (or statistical) laws in the local region can significantly improve the prediction accuracy. In conclusion, the prediction method based on monitoring the variation of the vertical E-field is feasible for external-source dust storm early warning. Compared with traditional ground-based and spaceborne observations, the method based on E-field monitoring has advantages, such as early detection and high accuracy. Our study clarifies the importance of the electrical effects associated with dust storms and blowing dust events for which such observations are rare and knowledge is sparse. Acknowledgements



z E ext z , a + (i − 1)l cos θ − E i = 0

.. .

149

(9)

where E 1 , E 2 , . . . , E i , . . . , E m are the measured values of the Efield sensors. A nonlinear system (9) of m equations of two unknown variables (i.e., a and θ ) typically has no solution when m > 2 (called inconsistent). In practice, a large number of E-field sensors are required to obtain more accurate real-time horizontal distance a, and hence m is much greater than 2. To find the best approximation, consider applying the Gauss–Newton method to solve a nonlinear least squares problem. See Sauer (2012) for general information on this topic. Strictly speaking, the exterior z vertical E-field E ext is expressed as Eq. (5). However, it can be z simplified as a power law E ext = μa−λ over a range of horizontal distances greater than ∼3 km (see Fig. 8b). Further work is also needed to assess the feasibility of a ground-based E-field observation array and establish a more detailed prediction system. 6. Conclusions Airborne dust particles affect the Earth’s climate by absorbing and scattering radiation, by warming the atmosphere at some altitudes, and by acting as cloud condensation and ice nuclei (Kok and Renno, 2006; Shao, 2008). Additionally, E-fields in dust storms have been found to lift ten times more dust into the atmosphere than wind alone (e.g., Esposito et al., 2016). Therefore, an understanding of dust electrification has significant implications for global climate studies. In this study, we presented multi-parameter measurements and theoretical calculations that were used to investigate the electrical properties of dust storms in Minqin, China. The measured E-field initially decreased and then increased with the

This research was supported by a grant from the National Natural Science Foundation of China (Nos. 11121202, 11232006, 11490553, and 11230226). The authors express their sincere appreciation for this support. Thanks are also extended to the editor, Tamsin Mather, and to the reviewers, Ralph Lorenz and an anonymous reviewer, for their very constructive suggestions and comments which greatly improved the manuscript. Appendix A. Supplementary material Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.epsl.2017.01.001. References Aizawa, K., Cimarelli, C., Alatorre-Ibargüengoitia, M.A., Yokoo, A., Dingwell, D.B., Iguchi, M., 2016. Physical properties of volcanic lightning: constraints from magnetotelluric and video observations at Sakurajima volcano, Japan. Earth Planet. Sci. Lett. 444, 45–55. http://dx.doi.org/10.1016/j.epsl.2016.03.024. Akhlaq, M., Sheltami, T.R., Mouftah, H.T., 2012. A review of techniques and technologies for sand and dust storm detection. Rev. Environ. Sci. Biotechnol. 11, 305–322. http://dx.doi.org/10.1007/s11157-012-9282-y. Bo, T.L., Zheng, X.J., 2013. A field observational study of electrification within a dust storm in Minqin, China. Aeolian Res. 8, 39–47. http://dx.doi.org/10.1016/j.aeolia. 2012.11.001. Bo, T.L., Zhang, H., Zheng, X.J., 2014. Charge-to-mass ratio of saltating particles in wind-blown sand. Sci. Rep. 4, 5590. http://dx.doi.org/10.1038/srep05590. Crozier, W.D., 1964. The Electric field of New Mexico dust devil. J. Geophys. Res. 69, 5427–5429. http://dx.doi.org/10.1029/JZ069i024p05427. Demon, L., DeFelice, P., Gondet, H., Kast, Y., Pontier, L., 1953. Premiers résultats obtenus au cours du printemps. J. Rech. Cent. Natl Rech. Sci. 24, 126–137. Durán, O., Claudin, P., Andreotti, B., 2011. On aeolian transport: grain-scale interactions, dynamical mechanisms and scaling laws. Aeolian Res. 3, 243–270. http:// dx.doi.org/10.1016/j.aeolia.2011.07.006. Dwyer, J.R., Uman, M.A., 2014. The physics of lightning. Phys. Rep. 534, 147–241. http://dx.doi.org/10.1016/j.physrep.2013.09.004.

150

H. Zhang et al. / Earth and Planetary Science Letters 461 (2017) 141–150

Esposito, F., Molinaro, R., Popa, C.I., Molfese, C., Cozzolino, F., Marty, L., Taj-Eddine, K., Di Achille, G., Franzese, G., Silvestro, S., Ori, G.G., 2016. The role of the atmospheric electric field in the dust-lifting process. Geophys. Res. Lett. 43, 5501–5508. http://dx.doi.org/10.1002/2016GL068463. Forward, K.M., Lacks, D.J., Sankaran, R.M., 2009. Charge segregation depends on particle size in triboelectrically charged granular materials. Phys. Rev. Lett. 102, 028001. https://doi.org/10.1103/PhysRevLett.102.028001. Freier, G.D., 1960. The electric field of a large dust devil. J. Geophys. Res. 65, 3504. http://dx.doi.org/10.1029/JZ065i010p03504. Harrison, R.G., Barth, E., Esposito, F., Merrison, J., Montmessin, F., Aplin, K.L., Borlina, C., Berthelier, J.J., Déprez, G., Farrell, W.M., Houghton, I.M.P., Renno, N.O., Nicoll, K.A., Tripathi, S.N., Zimmerman, M., 2016. Applications of electrified dust and dust devil electrodynamics to Martian atmospheric electricity. Space Sci. Rev. 203, 299–345. http://dx.doi.org/10.1007/s11214-016-0241-8. Hu, W., Xie, L., Zheng, X., 2012. Contact charging of silica glass particles in a single collision. Appl. Phys. Lett. 101, 114107. http://dx.doi.org/10.1063/1.4752458. Jackson, T.L., Farrell, W.M., 2006. Electrostatic fields in dust devils: an analog to Mars. IEEE Trans. Geosci. Remote Sens. 44, 2942–2949. http://dx.doi.org/10. 1109/TGRS.2006.875785. Kamra, A.K., 1972. Measurements of electrical properties of dust storm. J. Geophys. Res. 77, 5856–5869. http://dx.doi.org/10.1029/JC077i030p05856. Kok, J.F., Lacks, D.J., 2009. Electrification of granular systems of identical insulators. Phys. Rev. E 79, 051304. http://dx.doi.org/10.1103/PhysRevE.79.051304. Kok, J.F., Renno, N.O., 2006. Enhancement of the emission of mineral dust aerosols by electric forces. Geophys. Res. Lett. 33, L19S10. http://dx.doi.org/10. 1029/2006GL026284. Kok, J.F., Renno, N.O., 2008. Electrostatics in wind-blown sand. Phys. Rev. Lett. 100, 014501. http://dx.doi.org/10.1103/PhysRevLett.100.014501. Kok, J.F., Parteli, E.J.R., Michaels, T.I., Karam, D.B., 2012. The physics of windblown sand and dust. Rep. Prog. Phys. 75, 106901. http://dx.doi.org/10.1088/ 0034-4885/75/10/106901. Lacks, D.J., Sankaran, R.M., 2011. Contact electrification of insulating materials. J. Phys. D, Appl. Phys. 44, 453001. http://dx.doi.org/10.1088/0022-3727/44/45/ 453001. Lorenz, R.D., Jackson, B.K., 2015. Dust devils and dustless vortices on a desert playa observed with surface pressure and solar flux logging. GeoResJ 5, 1–11. http:// dx.doi.org/10.1016/j.grj.2014.11.002. Mather, T.A., Harrison, R.G., 2006. Electrification of volcanic plumes. Surv. Geophys. 27, 387–432. http://dx.doi.org/10.1007/s10712-006-9007-2. Merrison, J.P., 2012. Sand transport, erosion and granular electrification. Aeolian Res. 4, 1–16. http://dx.doi.org/10.1016/j.aeolia.2011.12.003. Morawska, L., He, C., Hitchins, J., Mengersen, K., Gilbert, D., 2003. Characteristics of particle number and mass concentrations in residential houses in

Brisbane, Australia. Atmos. Environ. 37, 4195–4203. http://dx.doi.org/10.1016/ S1352-2310(03)00566-1. Rakov, A.V., Uman, M.A., 2003. Lighting-Physics and Effects. Cambridge Univ. Press, New York. Rudge, W.A.D., 1913. Atmospheric electrification during South Africa dust storm. Nature 91, 31–32. http://dx.doi.org/10.1038/091031a0. Sauer, T., 2012. Numerical Analysis, 2nd ed. Pearson Education Inc., Boston. Schmidt, D.S., Schmidt, R.A., Dent, J.D., 1998. Electrostatic force on saltating sand. J. Geophys. Res., Atmos. 103, 8997–9001. http://dx.doi.org/10.1029/98JD00278. Shao, Y.P., Dong, C.H., 2006. A review on East Asian dust storm climate, modelling and monitoring. Glob. Planet. Change 52, 1–22. http://dx.doi.org/10.1016/ j.gloplacha.2006.02.011. Shao, Y.P., 2008. Physics and Modelling of Wind Erosion, 2nd ed. Springer, Heidelberg, Germany. Stow, C.D., 1969. Dust and sand storm electrification. Weather 24, 134–144. http:// dx.doi.org/10.1002/j.1477-8696.1969.tb03165.x. Williams, E., Nathou, N., Hicks, E., Pontikis, C., Russell, B., Miller, M., Bartholomew, M.J., 2009. The electrification of dust-lofting gust fronts (‘haboobs’) in the Sahel. Atmos. Res. 91, 292–298. http://dx.doi.org/10.1016/j.atmosres.2008.05.017. Yoshimatsu, R., Araújo, N.A.M., Shinbrotd, T., Herrmann, H.J., 2016. Field driven charging dynamics of a fluidized granular bed. Soft Matter 12, 6261–6267. http://dx.doi.org/10.1039/c6sm00357e. Zhang, H., Zheng, X.J., Bo, T.L., 2013. Electrification of saltating particles in windblown sand: experiment and theory. J. Geophys. Res., Atmos. 118, 12086–12093. http://dx.doi.org/10.1002/2013JD020239. Zhang, H., Zheng, X.J., Bo, T.L., 2014. Electric fields in unsteady wind-blown sand. Eur. Phys. J. E 37 (13). http://dx.doi.org/10.1140/epje/i2014-14013-6. Zheng, X.J., Huang, N., Zhou, Y.H., 2003. Laboratory measurement of electrification of wind-blown sands and simulation of its effect on sand saltation movement. J. Geophys. Res., Atmos. 108, 4322. http://dx.doi.org/10.1029/2002JD002572. Zheng, X.J., He, L., Zhou, Y.H., 2004. Theoretical model of the electric field produced by charged particles in windblown sand flux. J. Geophys. Res., Atmos. 109, D15208. http://dx.doi.org/10.1029/2004JD004863. Zheng, X.J., Huang, N., Zhou, Y.H., 2006. The effect of electrostatic force on the evolution of sand saltation cloud. Eur. Phys. J. E 19, 129–138. http://dx.doi.org/10. 1140/epje/e2006-00020-9. Zheng, X.J., 2013. Electrification of wind-blown sand: recent advances and key issues. Eur. Phys. J. E 36, 138. http://dx.doi.org/10.1140/epje/i2013-13138-4. Zhou, Z., Liu, Y., Yuan, J., Zuo, J., Chen, G., Xu, L., Rameezdeen, R., 2016. Indoor PM2.5 concentrations in residential buildings during a severely polluted winter: a case study in Tianjin, China. Renew. Sustain. Energy Rev. 64, 372–381. http://dx.doi. org/10.1016/j.rser.2016.06.018.