DONG Yuxiang 　S.L.Namikas　P.A.Hesp　MA Jun

　　1　Introduction

　　Since Bagnolds observations and experiments on Aeolian sand flow in North African desert during the 1930～40, this issue has attracted considerable attention of academic circles (Bagnold,1941; Wu et al., 2003; Zou et al., 1993; Li et al., 1998; Dong,2005; Dong et al., 2005), Numerous field observations, wind tunnel simulation experiments and numerical simulation studies have been carried out on the structure of wind-sand flow at home and abroad, including the normal research on the basic characteristics of wind-sand flow (Chepil,1945; Williams,1964; Nordstorm et al., 1990; Barndorff-Nielsen et al., 1991; Cater et al., 1992; Fryrear et al., 1993; Greeley et al., 1996) and studies of the effects of the factors such as wind velocity (Wu et al., 2003; ЗНАМЕНСКИЙ,1958; Wu et al.,1965; Ma, 1988; Butterfield,1999; Dong et al., 2002), sand transport rate (Wu et al., 2003; ЗНАМЕНСКИЙ,1958; Ma, 1988), sand particle size (Wu et al., 2003; Feng, et al., 2007), underlying surface character and shape (Wu et al., 2003; Ma, 1988; Yin, 1989; Hasi, 2004 ) etc. But there was some different idea about the effect of sand transport rate on the structure of blowing sand flow. For example, with regard to the effect of sand transport rate on the structure of wind-sand flow, it is generally accepted that as the total amount of sand transported in air flow increases, the absolute sand transport rate in various layers in the wind-sand flow increases while the relative sand transport rate (%) increases in the lower layer but decreases in the upper layer (Wu et al., 2003). Several scholars suggested that under the condition of the same wind velocity, with increase in the total sand transport rate the absolute sand transport rate in various layer increases while the relative sand transport rate decreases in the lower layer but increases in the upper layer. The reason for this is that as the total sand transport rate increases the frequency of collision and impact among sand particles increases thereby lead to the increase of the transport height of sand and dust particles (Ma, 1988). However, Wu Zheng et al. (Wu et al., 2003)hold that the increase in total sand amount causes even more frequent collision among sand particles and the energy loss in the collision processes of sand particles increases, so that the transport height of sand particles greatly lowers. However, above studies mainly focused on the inland desert dunes, and little information has been reported as to the effects of sand transport rate on the structure of wind-sand flow on coastal dunes and also no field observation and verification data are available, and similar studies in China are lacking. In view of this, the transverse ridges with largest size and most typical shape on the Changli Gold Coast in Hebei Province of China were selected to conduct the field observations on the structure of wind-sand flow over the surface of coastal dunes under different sand transport rates. Furthermore, the possible effects of sand transport rate on the structure of wind-sand flow over the surface of coastal dunes were addressed.

　　2　Method

　　2.1　Study area

　　Field observation point was selected on the Changli Gold Coast of Hebei Province, where is one of the most typical distribution regions of coastal dunes in China (Fu, et al., 1994; Wu et al., 1995; Fu et al., 1997; Dong, 2006; Hu et al., 1996). Along the coastline, 45km in length from south to north and 1.5-2km in width, there are some typical types of coastal dunes such as embryonic foredune, transverse ridge, barchan and barchan chain, and coastal sand sheet etc., of which transverse ridges have largest volume and most typical shape. Transverse ridges in the region are aligned in a NNE-SSW direction, 5-9Km in length, generally with a height of 20m or more and maximum height of 40m. Individual dunes are 150-250m wide, and their flanks are significantly asymmetric, the seaward slopes are long and gentle and dip by 8-12° but the landward slopes are short and steep, with a dip angle of 28-32°. Under the actions of strong wind, transverse ridges continuously migrate landward at an annual mean rate of 1-2m, their shapes are also constantly changing, and especially their crests change greatly due to strong blowing sand flow and its high frequency. Therefore, Changli Gold Coast of Hebei province, was selected as the field observation points to observe the structure of wind-sand flow over the crests of transverse ridges under different sand transport rates.

　　2.2　Observation method

　　The field observation of wind-sand flow structure under different sand transport rates were conducted using the field gradient anemometer and flat-orifice sand trap developed by the Institute of Wind And Sand Disaster Research at the Beijing Normal University. This gradient anemometer can be used to determine wind velocity at different height in the range of 0.3-30.0m/s. With a resolution of 0.1m/s, it is equipped with a data logger, and the results can be exported in the form of ASCII code data to make statistical analysis by directly using Excel software. The total height of the flat-orifice sand trap is 85cm, its sand and dust-collecting height is 60cm, sampling gradient consists of 30 continuous 2cm×2cm sand inlets, in total it can collect 30 layers of data, with a collecting efficiency larger than 80%. 

　　The field observations were conducted on April 2-4, 2007, the transverse sand ridge selected for this observation is 39.5m high and 143.8m wide, with typical shape and huge volume. The field observations were synchronously conducted using anemometers and sand traps, the observation heights of wind velocity were 5,15,30,60,120,180,280 and 380cm, the time to collect the sand samples in sand traps was decided by the wind velocity values and sand-collecting situations of sand traps. Then the sand samples were simply treated in the laboratory and weighted using electronic balance (accuracy 0.001g) to obtain the amount of sand and dust transported by air flow at different height. And then the particle-size analysis was made using the Malvern MS 2000 Laser Radiation. 

　　3　Results

　　At the same observation point with the same underlying surface character and consistent surface material composition and under the approximate wind velocity (Table 1), two sets of effective data of total sand transport rate with large difference were obtained, i.e. Observation D and Observation F, their total sand transport rates were 8.378 g·cm-2·min-1 and 15.695g·cm-2·min-1 respectively. Based on these observation data, the effects of sand transport rate on the structure of wind-sand flow over the surface of coastal dunes were studied.

　　Table 1　Data of vertical change of wind velocity over the crest of coastal transverse ridge

　　3.1　Changes of absolute sand transport rate in various layers

　　Observation results show that (Table 2), as the total sand transport rate in air flow increased, the absolute sand transport rate in various layers in the structure of wind-sand flow increased, and the sand transport rates in various layers were larger in Observation F with large total sand transport rate than in Observation D with small total sand transport rate, but there was a certain difference in the increase amplitude of sand transport rate at different height. So far as the increased absolute sand transport rate is concerned, the total sand transport rate of Observation F increased by 7.317 g·cm-2·min-1 compared to Observation D. The largest increase in the absolute sand transport rate in thewind-sand flow occurred at the height of 4~6cm and 6~8cm, the increased sand transport rate in these two layers accounted for 32.65% of the totally increased amount, while the absolutely increased sand transport rate in various layers above 26cm was less than 0.1 g·cm-2·min-1. So far as the relatively increased sand transport rate is concerned, the totally increased percentage of the total sand transport rate of Observation F was 87.33% compared to Observation D, in all the structure of wind-sand flow only the increased percentage in the two layers at 0~4cm height was lower than the totally increased percentage, the increased percentages of sand transport rate in 0~2cm and 2~4cm layers were 21.93% and 6.66% respectively, and except the 4-6cm layer the increased percentages in other layers were larger than 100%, or even reached up to a few ten-fold. Therefore, there existed a phenomenon that as the total sand transport rate increased, the increased percentage of sand transport rate was low in the lower wind-sand flow layer but large in the upper wind-sand flow layer.

　　3.2　Changes of relative sand transport rate in various layers

　　As viewed from the relative sand transport rate, i.e. the percentage of sand transport rate in the air flow layer at different height in the total sand transport rates in Observation F and Observation D (Table 3), in Observation D, 98.17% of sand and dust transported by wind was distributed within the 20cm height above the surface, of which 89.58% of sand amount was mainly concentrated in the vertical range below 10cm, the sand transport rate in 0-4cm height accounted for 61.46%, while above 10cm and 20cm height the amount of sand and dust transported by air flow only accounted for 10.42% and 1.87%. In Observation F, the sand transport rate in 0-4cm height only accounted for 38.02%, the sand transport rate below 10cm and 20cm height accounted for 74.54% and 93.09% respectively. But above 10cm and 20cm height, the amount of sand and dust transported by wind accounted for 25.46% and 6.91% respectively. Therefore, as the total sand transport rate increased in 0-60cm height, the 4cm height was a turning point for the direction changes in relative sand transport rate of wind-sand flow. With 4cm height as a boundary, the relative sand transport rate of wind-sand flow decreased in the lower layer (0-4cm) but increased in the upper layer (4-60cm). This result is basically consistent with that observed by Ma (Ma, 1988).

　　Table 2　Absolute Sand transport rate and its change at the different height layer in wind-sand flow (g·cm-2·min-1)

　　Table 3　Ratio of sand transport rate at different heights at different total sand transport rates  (%)

　　3.3　Changes of the structure model of wind-sand flow

　　For the analysis of the effects of the changes in sand transport rate on the structure of wind-sand flow over the crest of coastal transverse ridge, a bulk and stepwise correlation analyses were conducted on the amount of sand transported by air flow in relation to the height in accordance with the observed results of the Observation D and Observation F (Fig.1). From the data comparison of Observation D and Observation F, it can be seen that under the influence of different sand transport rate, the structural pattern of wind-sand flow occurred come changes, which were mainly manifested in:

　　Within 60cm height (Fig.1a and Fig.1b), both the exponential function and power function fittings on the sand transport rates in various layer of wind-sand flow and heights yielded a high correlation coefficient. The correlation coefficient (R2) at 0.01 significance level was larger than 0.83. However, comparatively the power function relation of Observation D was slightly stronger, while the Observation F basically exhibited an exponential function relation with a correlation coefficient of 0.94.

　　The optional fitting function for the sand transport rate and height in 0-40cm height was an exponential function and the correlation coefficient (R2) at 0.01 significance lever was 0.98 or more(Fig.1c and Fig.1d), suggesting that the sand transport rate of wind-sand flow in 0-40cm height over the crest of transverse ridge exhibited an exponential decrease with height.

　　Fig.1　Correlation formulas between sand transport rate and height in the structure of wind-sand flow at different total sand transport rates

　　The relation between sand transport rate and height in 40-60cm height was complex(Fig.1e and Fig.1f). The optional fitting function of Observation D was a polynomial and the correlation coefficient (R2) was only 0.36 while the optional fitting curve of Observation F was entirely consistent with a polynomial function and its correlation coefficient (R2) was 0.95.

　　4　Conclusion and Discussion

　　Due to the limitation of field observation conditions, there was nearly no way to obtain the observation data concerning the structure of wind-sand flow of the same wind velocity, and the bulk sample under different sand transport rates, but at the same observation point with same underlying surface character and consistent surface material composition and under the approximate wind velocity, the observation results of the effects of sand transport rate on the structure of wind-sand flow over the surface of coastal dunes show that as the total amount of sand transported by air flow increased, the absolute sand transport rates in various layers of wind-sand flow over the crest of transverse sand ridges in Changli Gold Coast of Hebei Province increased, of which the largest increase occurred in 4~8cm height, while the relative sand transport rate decreased in lower layer (0-4cm) but increased in upper layer (4-60cm). The structure model of wind-sand flow exhibited an exponential distribution in 0~40cm height, but in 0~60cm height it changed from power function distribution into exponential function distribution with increase in total sand transport rate, and in 40~60cm height it changed from insignificant relation into polynomial function distribution with increase in total sand transport rate. The reason for this change is closely related to the sand-carrying limit of different-velocity airflow and the different composition of sand grain size in the wind-sand flow, the vertical distribution characters of different sand particles moving in different ways in wind-sand flow were different.

　　The capacity of the largest sand particles transported by the airflow at the given velocity has a clear limit value. As wind velocity is approximate but the total sand transport rate increases, the increased sand materials by airflow transport should be smaller than the sand particles with the largest particle size that can be transported at that wind velocity. As for the composition of sand materials on dune surface and wind velocity in the study region, the increased sand material by airflow transport mainly were the fine sand and medium sand. Therefore, as the total sand transport rate greatly increased, the transport rate of medium sand and fine sand in airflow increased but the sand transport rate of coarse sand showed no increase. According to the sand grain size of Observation D and Observation F (Table 4), the grain size of Observation D has more coarse and bad sorting features, the content of coarse sand and medium sand in the wind-sand flow of Observation D is higher than in Observation F, but the content of fine sand in the wind-sand flow of Observation D is lower than in Observation F, specially the content of coarse sand and medium sand at the 0~2cm and 20~60cm height in the wind-sand flow of Observation D is higher than in Observation F.

　　Table 4　Parameters and compositions of grain size at different heights in the structure of wind-sand flow at the different total sand transport rates

　　The coarse-grained sands mainly occurred in the surface layer of sand dunes and moved in creep way, which mostly could be affected by the micro-features of dune surface. Turbulence mainly developed in the near-surface layer and the leading action of turbulence caused the changes in sand transport rate with height to satisfy the negative power-law relation (Li et al., 1993). Furthermore, due to strong saltation of coarse sands, they reached a higher height and thus caused the result that the content of coarse sand in the layer above 20cm in the wind-sand flow increased rather than decreased. 

　　Fine sands mostly move in saltation way, many studied results demonstrate that the vertical distribution of the sand transport rate of saltation sand particles is a single exponential decrease function (Wu et al., 2003; Dong et al., 2002; Feng et al., 2007; Ni et al., 2002; Li et al., 2004). Hence the vertical distribution of wind-sand flow in Observation F exhibits a typical exponential function distribution pattern because of its higher fine sand content, and the increase in saltating component inevitably results in the decrease in relative sand transport rate in the lower layer of wind-sand flow and the increase in the upper layer. 

　　Medium sands move in creep or saltation way depending on the wind velocity values. As wind velocity is enough high, it is dominated by saltation, when wind velocity is small it is dominated by creep. In this observation the content of saltating component was high due to high wind velocity and strong wind force, which led to its vertical distribution basically coincided with an exponential function distribution pattern, especially in 0-40cm height. It was because the increase in sand transport rate that led to the increase in saltation strength of medium sand and fine sand, and due to the influence of saltating height, the increase in sand transport rate in wind-sand flow mainly occurred within 20cm height or even 10cm height, coupling with the effect of “elephant nose”, 4-6cm and 6-8cm heights became two airflow layers occurred the largest increase in sand transport rate.

　　With regard to the changes of sand transport rate in 40-60cm airflow layer, it was inferred that due to very low sand and dust content in Observation D, the factors and probability the sand transport rate increased, coupling with reduced influential degree of height, it showed no obvious function relation. But in Observation F, the height influence became significant as the total sand transport rate increased; combined with the effects of other factors it appeared as a polynomial function distribution. However, further study is needed to examine this problem.

　　Acknowledgments

　　The authors would like to thank National Natural Science Foundation of China for the support in this research. Here we also extend our sincere thanks to Prof. Hasi, Prof. Zou Xueyong and Prof. Liu Lianyou of Beijing Normal University for their great help for the field work. We express our gratitude to Huang Dequan, Zhang Xiaoxiao, Zheng Yinghua, Xia Xiandong, Fu Chuan and Ni Shaochun, who did the field work with authors. And we also thank Prof. Libaosheng and Dr. Wen Xiaohao for their help for sand grain size analysis.

　　原載：Chinese Geographical Science,2008，18（3）：255-261.