WITSEG sampler: a segmented sand sampler for wind tunnel test

Geomorphology 59 (2004) 119 – 129 www.elsevier.com/locate/geomorph

WITSEG sampler: a segmented sand sampler for wind tunnel test Zhibao Dong *, Hongyi Sun, Aiguo Zhao Laboratory of Blown Sand Physics and Desert Environments, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, No. 260, West Donggang Road, Lanzhou, Gansu Province 730000, People’s Republic of China Accepted 16 July 2003

Abstract The flux profile of blowing sand is the reflection of a large number of sand particles moving in different trajectories. To describe the function of the flux profile requires measuring the flux of blown sand at different heights. A segmented sand sampler for wind tunnel study (WITSEG sampler) has been designed and evaluated in a wind tunnel. The sampler is 0.6 m high and sectioned into 60 openings that can measure the flux profile of blowing sand at 10-mm intervals. Special attention was paid in designing the sampler to its practicability and efficiency. For the convenience of use, the leading edge with inlets, sand chambers and side cover are removable. To maximize the sampling efficiency, the leading edge is wedge-shaped so that the width of the inlets is 5 mm, while the width of the sand chamber is 15 mm. Each sand chamber has two screened vents that are connected to a vertical vent to minimize the air pressure inside the sand chamber and maximize its efficiency. Wind tunnel evaluation shows that the sampler is good for measuring aeolian sand flux. Both the measured flux profile and total flux show a reasonably good agreement with the widely accepted flux profile function and sand transport equation. It has an overall efficiency of 0.91. The WITSEG sampler that can measure the detailed variation of blown sand flux with height proves to be a good tool for studying the flux profile of blowing sand. D 2003 Elsevier B.V. All rights reserved. Keywords: Aeolian sand transport; Sampling efficiency; Flux profile

1. Introduction Aeolian transport is an important geomorphological process both in the geological eras and at present (Liu, 1985; Pye and Tsoar, 1990; Lancaster, 1995; Livingstone and Warren, 1996). Past studies demonstrate that aeolian processes are also active on the planet of Mars and possibly Venus and Titan (Greeley and Iversen, 1985). One of the central concerns in aeolian research is the sediment transport rate, which is important for the development of aeolian landforms and the management * Corresponding author. E-mail address: [email protected] (Z. Dong). 0169-555X/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2003.09.010

of aeolian hazards (Lancaster, 1995; Livingstone and Warren, 1996). However, estimating the rate of sediment transport by wind with accuracy proves to be a difficult task. Difficulties in understanding the particle movement mechanism, along with complications such as moisture, vegetation and surface crusting limit the development of theoretical relations linking sediment discharge to wind or wind friction velocity. The relations based on wind tunnel tests and field observations are largely dependent on the reliability of the direct measurement of sediment flux. Though it is necessary for the confirmation and calibration of theoretically derived equations (Nickling and McKenna Neuman, 1997), some researchers (Livingstone and Warren,

120

Z. Dong et al. / Geomorphology 59 (2004) 119–129

1996) think that the direct physical measurement of sediment transport by wind is a less accurate method compared with the estimation from wind data. Seeking reliable sand traps or samplers for direct sediment transport measurement has been of continuing significance for decades and many types of sand samplers have been made for different purposes (e.g., Leatherman, 1978; Stout, 1990; Arens and van der Lee, 1995; Nickling and McKenna Neuman, 1997; Goosens and Offer, 2000; Goosens et al., 2000). Aeolian sand samplers fall into two broad groups (Nickling and McKenna Neuman, 1997): those with horizontal sampling orifices (e.g., Pollet et al., 1998) and those with vertical sampling orifices (Bagnold, 1941; Nickling and McKenna Neuman, 1997). Samplers can be classified as either passive or active according to the way in which the air inside the sampler is exhausted. Passive samplers are more popular because they are easy to use and relatively inexpensive (Nickling and McKenna Neuman, 1997). Aeolian sand samplers can be stationary or rotating: the stationary samplers that are usually used in a wind tunnel are always oriented to a single direction, while the rotating samplers that are needed in field measurements are able to change their direction in response to the wind direction. Samplers can be designed as integrating samplers that collect aeolian sediment flux within a relatively greater layer or single-point samplers (e.g., Wilson and Cook, 1980; White, 1982; Fryrear, 1986; Janssen and Tetzlaff, 1991; Spaan and van den Abeele, 1991) that collect sediment flux passing a small area (point). The integrated samplers can be either single-slot samplers (e.g., Nickling and McKenna Neuman, 1997) or segmented (e.g., Bagnold, 1941) samplers. The segmented samplers can collect the sediment flux at different positions respectively and are useful in studying flux profiles. The flux profile of blowing sand is the reflection of a large number of sand particles moving in different trajectories. To describe the function of the flux profile requires measuring the sand flux at different heights and creates the need for a segmented sand sampler. In wind tunnel studies, the sampler need not be rotating. Here, we report on the design and efficiency of a newly designed segmented sampler for wind tunnel tests (WITSEG sampler). The main purpose is to show if the WITSEG sampler is accurate and efficient in sampling aeolian sand move-

ment and provide some helpful experiences in sampler design.

2. Design of the sampler The WITSEG sampler is a vertically integrating, passive type that follows an earlier design by Bagnold (1941). The WITSEG sampler is designed to measure the flux profile of blowing sand in the sand wind tunnel of the Laboratory of the Blown Sand Physics and Desert Environments, Cold and Arid Regions Environmental and Engineering Research Institute, the Chinese Academy of Sciences. The cross-sectional area of the working section of the wind tunnel is 1.2
1.2 m. Shao et al. (1993) and Nickling and McKenna Neuman (1997) considered a desirable property for any sampler was that it be isokinetic, meaning that the flow speed through the intake is the same as the local instantaneous ambient air speed. Isokinetic sampling does not distort the flow at the sampling orifice, resulting in more accurate sampling of the streamwise sediment flux (Shao et al., 1993; Nickling and McKenna Neuman, 1997). However, few passive samplers are accurately isokinetic, and active samplers are not isokinetic unless the actively driven flow through the sampler is matched to independently sensed ambient wind speed (Shao et al., 1993). Some compromises must be made in designing a sand sampler to meet the requirement of isokinetic sampling. Nickling and McKenna Neuman (1997) suggested that a sampler should be compact and streamlined to cause as little interference with the flow as possible. This is most important for a wind tunnel that has very limited size. The sampler should be practical, easy to use and inexpensive. All the above factors were taken into account in designing the WITSEG sampler. The sampler is constructed of 0.5-mm stainless steel. The sampler has four main components (Fig. 1): a removable side cover, a wedge-shaped leading edge, a support and 60 sand chambers. The total height of the sampler is 700 mm and the total width 160 mm. The total thickness of the sampler is designed to be 25 mm so that it has no significant interference with the airflow while maintaining enough sampling volume of the sand chambers. The wedge-shaped leading edge

Z. Dong et al. / Geomorphology 59 (2004) 119–129

121

Fig. 1. Design of the WITSEG sand sampler (1: removable cover, 2: wedge-shaped leading edge, 3: support, 4: inlet, 5: sand chamber, 6: vertical vent, 7: screened vent).

122

Z. Dong et al. / Geomorphology 59 (2004) 119–129

has 60 nozzle orifices connecting to 60 sand chambers. Each orifice is 10 mm high and 5 mm wide. The wedgeshaped leading edge was chosen to reduce the interference of the sampler with the airflow at the inlet and avoid the too short sampling time of the sand chambers by reducing the sampling width of the inlets. The spacers between the orifices are milled very sharp to reduce the particle rebound on them and errors in the measurement of total flux. The leading edge and a side cover are removable so that the sand chambers can be removed and the collected sand weighed. The size of the sand chambers is 140
15
6 mm. Each chamber has a full sand capacity of about 18 g and is inclined 30j with respect to the horizontal. A key design in the leading edge of the sampler is the vent. The function of the vent is to ensure the isokinetic sampling. Jones and Willetts (1979) and Nickling and McKenna Neuman (1997) have suggested the importance of vent for a passive sampler. Jones and Willetts (1979) found that the stagnation of air in front of the sampler inlet produced by closing the vent reduced the transport rate by approximately half compared with those deemed to be correct values. Nickling and McKenna Neuman (1997) measured the airflow characteristics of their wedge-shaped trap. They found that the non-vented sampler was characterized by a zone of very high positive stagnation pressure (low velocity) at the front of the inlet with flow separation to each side of it. A strong gradient in static pressure developed along the sides of the trap toward the rear corners of the wedge. Venting of the trap reduced the positive stagnation at the inlet significantly and eliminated the flow separation to each side of the sampler, and weakened the pressure gradients toward the rear of the trap. To avoid the problem of stagnation and the related errors in measuring the sediment flux, each chamber of the WITSEG sampler is vented. There are vent holes of diameter 2 mm on both sides of the orifice, which are open to a sand chamber and connected to the common vertical vent. The airflow goes out of the vent and sand particles in the sand-laden wind entering an inlet fall into the sand chamber by gravity. To prevent the blown sand particles from going out with the wind, the vent holes are covered with the same fine stainless steel wire mesh (200 mesh, 62.5 Am openings, 60% porosity) as that used in Nickling and McKenna Neuman’s (1997) wedge-shaped sand trap.

3. Wind tunnel evaluation The main tasks of the evaluation were to examine if the sampler could capture accurate flux profiles or total flux and to determine its sampling efficiency. The collected sand flux at different heights and the total flux were analyzed and compared with the widely accepted flux profile function and sand transport equation to determine its accuracy. The sampling efficiency was determined by comparing the total flux collected by the sampler with that considered to be the ‘‘correct’’ total flux. The evaluation was carried out in the sand wind tunnel of the Laboratory of Blown Sand Physics and Desert Environments, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences. The blow-type noncirculating wind tunnel has a total length of 37 m and is composed of a power section (2.6 m long), expansion section (6.4 m long), stabilization section (1.5 m long, where drag screens and honeycombs are set to reduce large scale eddying), compression section (2.5 m long), working section (21 m long) and diffusion section (3 m long). The maximum cross-sectional area at the expansion section is 2.4
1.2 m. The 21-m long working section has a cross-sectional area of 1.2
1.2 m. Wind speed can be changed continuously from 2 to 30 m s 1. The stability coefficient of air flow is g < 3.0%, the uniformity coefficient of wind speed = du < 2.0%, the average turbulence e = 0.6% and the axial static pressure gradient jdCp/dxj = 0.01 m 1. The depth of boundary layer can reach 0.4– 0.5 m in the working section. 3.1. Experimental procedures The sand used for the evaluation test was typical dune sand from the Shapotou area, southeast of the Tengger Desert of China. Fig. 2 shows the grain size distribution of this sand. The mean diameter was 2.42F (0.18 mm), standard deviation 0.41F (well sorted), skewness 0.05 and kurtosis 1.02. The threshold friction velocity of the sand was 0.16 m s 1 (Ha and Wang, 2001). In the test, the sand was put in a 4.0
0.9
0.025m sand tray, which was set 14 m downwind from the start of the working section of the wind tunnel and tested at different wind speeds. The length of the sand tray chosen ensured a significant development of the saltating sand cloud (Dong et al., 2003). At each test,

Z. Dong et al. / Geomorphology 59 (2004) 119–129

123

Fig. 2. The grain size distribution of the sand used in evaluation test.

both the correct total flux and the flux collected by the sampler were determined. To determine the correct total flux, the sand sample put in the sand tray was weighed before and after a test by an electronic scale underlying it with an accuracy of 1 g. The weight difference before and after test was the total erosion amount, which was converted to the correct total flux per unit time and per unit width perpendicular to the wind. To determine the sand flux collected by the sampler, the sampler was set in the middle of the tunnel floor, about 18 m downwind from the entrance of the working section of the wind tunnel, 20 mm apart from the downwind end of sand tray. The bottom of the lowest opening of the sampler was set flush with the upwind sandy surface and downwind tunnel floor. Sampling times ranged from about 15 to 320 s, depending on the wind velocity. Because it took some time for the wind tunnel to reach a pre-set free-stream wind speed, the sand tray was covered by a piece of geo-textile until the required wind speed was reached so that the sand was not blown away. The geo-textile was uncovered rapidly to let the sandy surface be eroded when the wind attained equilibrium velocity. When the bottom of the sand tray was exposed, the wind tunnel was powered off and the sampler was removed. To determine the amount of sand collected by the sampler, the removable side cover and leading edge of the sampler were taken off, and each sand chamber was removed and emptied to weigh the collected sand

by using a 1/1000-g electronic balance. The total sand flux collected by the sampler was obtained by summing up the collected flux in every sand chamber. The evaluation test was conducted at eight free-stream wind velocities: 8, 10, 12, 14, 16, 18, 20 and 22 m s 1, which was measured by a Pitot static tube mounted at the centerline position of the start of the working section of the wind tunnel. The eight free-stream wind velocities corresponded to wind friction speeds of 0.21, 0.26, 0.31, 0.35, 0.39, 0.44 and 0.53 m s 1. For each wind velocity, three repetitions were made to get the mean values. Fig. 3 illustrates the layout of the test. 3.2. Flux profile and total flux Fig. 4 illustrates the measured flux profiles. The decay curves of the sand flux with height at different wind speeds are very similar and can be described by an exponential function: q ¼ aexpðh=bÞ

ð1Þ

where q is the sand transport rate at height h (in kg m 1 s 1), h is height in m, a and b are regressive coefficients. Table 1 indicates that both a and b are functions of wind speed. Eq. (1) is in the same form as the exponential decay function proposed by Williams (1964) for the 0.16-m near-surface layer based on wind tunnel tests. Many other published results also suggest that

124

Z. Dong et al. / Geomorphology 59 (2004) 119–129

Fig. 3. Layout of the tests.

the blown sand transport rate in saltation decays with height by a natural exponential law (Takeuchi, 1980; Greeley et al., 1983; Gerety and Slingerland, 1983; Anderson and Hallet, 1986; Wu, 1987; Fryrear et al., 1991; Fryrear and Saleh, 1993). Coefficient a represents the sand transport rate on the surface (h = 0), or the transport rate in true creep increases with wind speed. The true creep proportions in the total blown sand transport within the 0.6-m layer decrease with the wind speed, ranging from 12.09% to 24.2%, with an average of 16.65%. Bagnold (1941) suggested that the creep fraction was about 25%, close to the true creep fraction at 0.21 m s 1 wind friction speed in our test. Because the more intense saltation results in more particles moving in higher levels, the creep fraction should decrease as the wind speed increases.

The curves in Fig. 4 are converted to straight lines if we plot transport rate q on a log scale. The slope of the straight lines can be derived by: k ¼ ðlnq2  lnq1 Þ=ðh2  h1 Þ ¼ 1=b

ð2Þ

where k is the slope, q1 and q2 are the transport rates at heights h1 and h2, h1 and h2 are two different heights. Eq. (2) shows that the coefficient b in (Eq. (1)) characterizes the relative decay rate of the sand transport rate with height. Table 1 indicates that the relative decay rate decreases as the wind speed increases, revealing that more particles are moving at higher levels, resulting from more intense saltation at higher wind speed, decreasing both the creep fraction and the relative decay rate of the sediment flux. This is qualitatively in agreement with some

Fig. 4. The measured sediment flux profiles at different wind friction speeds.

Z. Dong et al. / Geomorphology 59 (2004) 119–129 Table 1 Correlation between sand flux q and height h Serial V* Qo (kg a no. (m s 1) m 1 s 1)

b

a/Qo (%)

k =  1/b r2

01 02 03 04 05 06 07 08

0.049 0.061 0.067 0.070 0.072 0.075 0.084 0.092

24.21 19.29 17.20 16.55 15.65 15.03 13.21 12.09

 20.41  16.39  14.93  14.29  13.89  13.33  11.90  10.87

0.21 0.26 0.31 0.35 0.39 0.44 0.49 0.53

0.037 0.086 0.178 0.325 0.550 0.721 0.952 1.261

0.0090 0.0166 0.0306 0.0538 0.0861 0.1084 0.1258 0.1525

0.92 0.89 0.92 0.92 0.95 0.96 0.92 0.87

V* is the friction wind speed, Qo is the total flux, fitted function: q = a exp(  h/b).

previously reported results (Znamenskii, 1960; Wu and Lin, 1965; Chen et al., 1996) that stated the blown sand was transported higher as wind speed increased. So, the sampler design captured the vertical flux profile with reasonably good accuracy. The relationship between the total sediment flux and wind speed fit Kawamura’s (1951) or White’s (1979) aeolian transport equation (Eq. (3)) very well (Fig. 5) when the proportionality coefficient C = 89.19. Q ¼ Cð1  Rt Þð1 þ R2t Þðq=gÞV3 ;

Rt ¼ Vt =V ð3Þ

Where Q is the total sediment flux (in kg m 1 s 1), V* and V*t are the friction wind speed and threshold

125

wind friction speed (in m s 1), q is the density of air (1.25 kg m 3), g is the gravitational acceleration (9.81 m s 2), C is a proportionality coefficient (89.19). So, the WITSEG sampler also captures the total flux effectively. 3.3. Sampling efficiency The most important characteristic of a sediment sampler is its efficiency and a great many efforts have been devoted to defining a sampler’s efficiency (Jones and Willetts, 1979; Shao et al., 1993; Nickling and McKenna Neuman, 1997; Goossens and Offer, 2000; Goossens et al., 2000). The efficiency of a sampler depends on the sampler design, the speed of the wind relative to the air speed at the sampler orifice, the size of airborne particles and the total time of sampling (Goossens and Offer, 2000; Goossens et al., 2000). Many calibration methods have been reported in the literature (see the literature review by Goossens et al., 2000). Goossens et al. (2000) thought the only way to measure the absolute efficiency of a sediment sampler is to compare the flux measured by the sampler to the flux measured by a nonintrusive sampler measuring at the same location and under identical conditions. However, these requirements are rarely met. Some compromises must be made.

Fig. 5. Comparison of the measured and predicted total sediment flux.

126

Z. Dong et al. / Geomorphology 59 (2004) 119–129

The efficiency is defined as the ratio of the collected total flux by the sampler to that considered being a ‘‘correct’’ value of the total flux rate: E ¼ Qs =Qc

ð4Þ

where E is the sampling efficiency, Qs is the collected total flux by the sampler, Qc is the correct value of the total sediment flux. So, the key to obtaining the sampling efficiency is to define Qs and Qc. In our calibration, Qc was obtained by: Qc ¼ ðWo  We Þ=ðL
T Þ

ð5Þ

where Wo is the total weight of the sand sample before the test, We is the total weight of the sand sample after test, L is the width of the sand tray and T is the sample time. In Eq. (5), Qc represents the average of the flux rate along the cross-section of the wind tunnel. However, in the test, the sampler was located in the middle of the cross-section. Table 2

lists Qc at different wind speeds. Constrained by the side-walls of the wind tunnel, the airflow and hence the sediment flux varies across the cross-section, as Ling (1994) suggested. To evaluate the sampling efficiency, the sampler must sample the sediment flux representative of the average flux of the crosssection. The flux measured by the sampler in the middle of the cross-section must be converted to the average flux by a correction factor. The variation across the cross-section of the collected sediment flux was measured at three wind friction velocities (0.31, 0.44 and 0.58 m s 1) by setting the sampler at different locations. Fig. 6 shows these results. The sediment flux is generally greater in the middle line. So, use of the collected sediment flux in the middle line alone will overestimate Qs. Fig. 6 reveals that the average sediment flux is 0.81, 0.82 and 0.84 of the middle-line flux, increasing slightly with the wind speed. Here, the mean value 0.823 is used as a correction factor account for the variation in flux

Table 2 The correct total sediment flux V* (m s 1)

Run

Wo (kg)

We (kg)

Wo  We (kg)

T (s)

Qc (g m 1 s 1)

Mean Qc (kg m 1 s 1)

S.D.

0.21

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

94.65 92.12 90.25 93.00 94.02 95.87 94.35 93.33 92.68 97.82 95.21 93.32 94.00 92.20 92.66 92.30 91.44 95.12 93.43 94.55 96.88 94.00 90.98 93.45

87.43 83.98 80.37 81.17 82.10 81.87 76.85 77.19 73.67 77.55 74.94 72.39 72.08 71.46 74.49 73.42 70.57 76.33 73.77 72.40 74.62 73.80 69.31 74.43

7.22 8.14 9.88 11.83 11.92 14.00 17.51 18.15 19.00 20.27 20.27 20.93 21.92 20.74 18.21 18.88 20.87 18.79 19.66 22.15 22.26 20.20 21.67 19.02

211 312 289 180 179 191 128 126 131 78 78 76 49 48 46 30 36 34 28 28 26 17 22 20

0.038 0.029 0.038 0.073 0.074 0.081 0.152 0.160 0.162 0.303 0.303 0.306 0.497 0.480 0.440 0.699 0.644 0.614 0.780 0.877 0.962 1.320 1.100 1.011

0.035

0.0042

0.076

0.0036

0.158

0.0043

0.304

0.0014

0.472

0.0239

0.652

0.0352

0.871

0.0744

1.140

0.1299

0.26

0.31

0.35

0.39

0.44

0.49

0.53

V* is the wind friction speed, Wo is the total weight of the sand sample before test, We is the total weight of the sand sample after test, T is the sample time, S.D. is the standard deviation.

Z. Dong et al. / Geomorphology 59 (2004) 119–129

127

Fig. 6. Variation of the sand transport across the cross-section of the wind tunnel.

across the cross-section and convert the measured sediment flux at the middle line to Qs. So, E ¼ Qs =Qc ¼ 0:823Qo =Qc

ð6Þ

Where Qo is the total flux collected by the sampler. The sampling efficiency at different wind speeds is obtained from the data of Qo and Qc listed in Tables 1 and 2. The sampling efficiency is less than 1, ranging from 0.87 to 0.96, with an average of 0.91 (Fig. 7). The factors responsible for the differences in the efficiency of our sampler from others reported are

complex. According to Fryrear (1986), Shao et al. (1993), Nickling and McKenna Neuman (1997), Goossens and Offer (2000) and Goossens et al. (2000), the efficiency of a sediment trap may be above or below unity, depending on the frictional losses (caused by the trap) in the fluid and the particle’s motion, and on the intensity of airflow acceleration or deceleration near the trap’s inlet. In the wind tunnel evaluation of several samplers, Shao et al. (1993) found that active samplers had higher efficiency than the passive samplers and these samplers were usually more efficient when sampling particles with greater

Fig. 7. Sampling efficiency of the WITSEG sampler.

128

Z. Dong et al. / Geomorphology 59 (2004) 119–129

size. For example, the overall efficiency of an active vertically integrating trap was 105 F 5% for an aeolian sand, while that of the passive Leach trap (White, 1982) was 85% and 70% for an aeolian sand and particles less than 10 Am, respectively. Fryrear’s BSNE trap had an efficiency of 90 F 5% for an aeolian sand and 40% for particles less than 10 Am, respectively. Goossens and Offer (2000), and Goossens et al. (2000) found that the same sampler (BSNE) had higher efficiency when sampling sand than sampling dust and generally the greater the particle size, the higher the efficiency. Different calibration methods can produce different sampling efficiencies for the same sampler. For example, Goossen et al.’s (2000) calibration indicated that the BSNE sampler had an efficiency ranging from 100% to 120%, while Shao et al.’s (1993) calibration found that the BSNE’s sampling efficiency was about 90% for a similar particle size. So the method used to calibrate the sampler itself is an issue still open to argument.

4. Conclusions A segmented aeolian sampler for studying the flux profile of blowing sand in a wind tunnel (WITSEG sampler) has been developed. The sampler can measure the sediment flux at 60 heights with 10-mm intervals and provide detailed data for developing the function of blown sand flux profile. The WITSEG sampler is passive type but each sand chamber is well vented. It has been evaluated in a wind tunnel using typical aeolian sand. The measured flux profiles at different wind speed show an exponential decay with height, in agreement with many previously reported flux profile functions. The variation of total flux with the wind speed fits very well the Kawamura (1951) or White (1979) equation. The overall efficiency is 0.91. In the wind tunnel, there is simplest condition of uni-directional wind. Nickling and McKenna Neuman (1997) pointed out that wind direction was a significant factor influencing the efficiency of a sampler. A rotating mechanism is needed for the WITSEG sampler to be used in the field. In this paper, we have presented a method to evaluate the efficiency of the WITSEG sampler in a wind tunnel. As many pioneer workers have shown,

the efficiency of a sampler is influenced by many factors and the evaluation method for estimating sampling efficiency is open to argument. More attention is needed to not only develop samplers but also the standard methods to calibrate them.

Acknowledgements We gratefully acknowledge the funding from the National Science Fund for Distinguished Young Scholars of the National Natural Science Foundation of China (40225003) and the Knowledge Innovation Project of the Chinese Academy of Sciences (KZCX3-SW-324). We wish to thank Mr. G. Wang for his help in the wind tunnel experiment and Dr. G. Wu and Mr. B. He for preparing the illustrations.

References Anderson, R.S., Hallet, B., 1986. Sediment transport by wind: toward a general model. Geological Society of America Bulletin 97, 523 – 535. Arens, S.M., van der Lee, G.E.M., 1995. Saltation sand traps for the measurement of aeolian transport into the foredunes. Soil Technology 8, 61 – 74. Bagnold, R.A., 1941. The Physics of Blown Sand and Desert Dunes Methuen, London. 265 pp. Chen, W., Yang, Z., Zhang, J., Han, Z., 1996. Vertical distribution of wind-blown sand flux in the surface layer, Taklimakan Desert, Central Asia. Physical Geography 17, 193 – 218. Dong, Z., Wang, H., Liu, X., Wang, X., 2003. The blown sand flux over a sandy surface: a wind tunnel investigation on the fetch effect. Geomorphology 57, 117 – 127. Fryrear, D.W., 1986. A field dust sampler. Journal of Soil and Water Conservation 41, 117 – 120. Fryrear, D.W., Saleh, A., 1993. Field wind erosion: vertical distribution. Soil Science 155, 294 – 300. Fryrear, D.W., Stout, J.E., Hagen, L.J., Vories, E.D., 1991. Wind erosion: field measurement and analysis. Transactions of the ASAE 34, 155 – 160. Gerety, K.M., Slingerland, R., 1983. Nature of saltation population in wind tunnel experiments with heterogeneous size-density sands. In: Brookfield, M.E., Ahlbrand, T.S. (Eds.), Eolian Sediments and Progresses: Developments in Sedimentology. Elsevier, Amsterdam, pp. 115 – 131. Goossens, D., Offer, Z.Y., 2000. Wind tunnel and field calibration of six aeolian dust samplers. Atmospheric Environment 34, 1043 – 1057. Goossens, D., Offer, Z.Y., London, G., 2000. Wind tunnel and field calibration of five sand traps. Geomorphology 35, 233 – 252. Greeley, R., Iversen, J.D., 1985. Wind as a Geological Process on

Z. Dong et al. / Geomorphology 59 (2004) 119–129 Earth, Mars, Venus and Titan Cambridge Univ. Press, Cambridge. 333 pp. Greeley, R., Williams, S.H., Marshall, J.R., 1983. Velocities of wind-blown particles in saltation: preliminary laboratory and field measurements. In: Brookfield, M.E., Ahlbrandt, T.S. (Eds.), Eolian Sediments and Progresses: Developments in Sedimentology. Elsevier, Amsterdam, pp. 133 – 148. Ha, S., Wang, G., 2001. Characteristics of grain size over crescent dunes in Shapotou, southeastern Tengger desert. Journal of Desert Research 21, 271 – 275. Janssen, W., Tetzlaff, G., 1991. Entwicklung und Eichung einer registrierenden Suspensionsfalle. Zeitschrift fu¨r Kulturtechnik und Landentwicklung 32, 167 – 180. Jones, J.R., Willetts, B.B., 1979. Errors in measuring uniform aeolian sand flow by means of an adjustable trap. Sedimentology 26, 463 – 468. Kawamura, R., 1951. Study of sand movement by wind. Hydraulics Engineering Laboratory Report HEL-2-8, University of California, Berkeley. Lancaster, N., 1995. Geomorphology of Desert Dunes Routledge, London. 290 pp. Leatherman, S.P., 1978. A new aeolian sand trap design. Sedimentology 25, 303 – 306. Ling, Y.Q., 1994. The distributive heterogeneity of sand transporting quantity (rate) along horizontal direction. Experimental Mechanics 9, 352 – 356. Liu, T., 1985. Loess and Environment Science Press, Beijing. 481 pp. Livingstone, I., Warren, A., 1996. Aeolian Geomorphology: An Introduction Longman Singapore Publishers, Singapore. 211 pp. Nickling, W.G., McKenna Neuman, C., 1997. Wind tunnel evaluation of a wedge-shaped aeolian sediment trap. Geomorphology 18, 333 – 345.

129

Pollet, I., Gabrie¨ls, D., Cornelis, W., 1998. The catch efficiency of a windblown sand collector: a wind tunnel study. Pedologie Themata 1998 (5), 75 – 78. Pye, K., Tsoar, H., 1990. Aeolian Sand and Sand Dunes Unwin Hyman, London. 396 pp. Shao, Y., Mctainsh, G.H., Leys, J.F., Raupach, M.R., 1993. Efficiency of sediment samplers for wind erosion measurement. Australian Journal of Soil Research 31, 519 – 532. Spaan, W.P., van den Abeele, G.D., 1991. Wind borne particle measurements with acoustic sensors. Soil Technology 4, 51 – 63. Stout, J.E., 1990. Wind erosion within a simple field. Transactions of the ASAE 33, 1597 – 1600. Takeuchi, M., 1980. Vertical profile and horizontal increase of driftsnow transport. Journal of Glaciology 26, 4492 – 4498. White, B.R., 1979. Soil transport by wind on Mars. Journal of Geophysical Research 84, 4643 – 4651. White, B.R., 1982. Two-phase measurements of saltating turbulent boundary layer flow. International Journal of Multiphase Flow 8, 459 – 473. Williams, G., 1964. Some aspects of aeolian transport load. Sedimentology 3, 257 – 287. Wilson, S.J., Cook, R.U., 1980. Wind erosion. In: Kirkby, M.J., Morgan, R.P.C. (Eds.), Soil Erosion. Wiley, Chichester, pp. 217 – 251. Wu, Z., 1987. Aeolian Geomorphology (in Chinese). Science Press, Beijing. 316 pp. Wu, Z., Lin, Y., 1965. A preliminary study on blown sand movement and its control. Blown Sand Control Research (in Chinese) 7, 7 – 14. Znamenskii, A.I., 1960. Experimental Investigation of Wind Erosion of Sandland and the Control of Drifting Sand (Chinese version translated by Yang, Y.) Science Press, Beijing. 120 pp.