APP下载

A mechanism for the origin and development of the large-scale dunefield on the right flank of the lower reach of Laoha River, Northeast China

2015-10-28GuangHanGuiFangZhangLiYouLiangZhouLinYangXueYongZhaoYuLinLiTongHuiZhang

Sciences in Cold and Arid Regions 2015年1期

Guang Han, GuiFang Zhang, Li You, Liang Zhou, Lin Yang,XueYong Zhao, YuLin Li, TongHui Zhang

1. Hunan Normal University, Changsha, Hunan 410081, China

2. Hunan University, Changsha, Hunan 410082, China

3. Climatic Center of Inner Mongolia, Huhhot, Inner Mongolia 010041, China

4. Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou,Gansu 730000, China

A mechanism for the origin and development of the large-scale dunefield on the right flank of the lower reach of Laoha River, Northeast China

Guang Han1*, GuiFang Zhang2, Li You3, Liang Zhou1, Lin Yang1,XueYong Zhao4, YuLin Li4, TongHui Zhang4

1. Hunan Normal University, Changsha, Hunan 410081, China

2. Hunan University, Changsha, Hunan 410082, China

3. Climatic Center of Inner Mongolia, Huhhot, Inner Mongolia 010041, China

4. Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou,Gansu 730000, China

By viewing satellite imagery, a striking large-scale dunefield can be clearly perceived, with a size of nearly 63 km long and 11 km wide, and trending NE-SW, on the right flank of the lower Laoha River, Northeast China. By means of remote sensing imagery analysis and field observation as well as a comparison with a small-scale dunefield on the right flank of the lower Xiangshui River, analogous to the case of the lower Laoha River, this paper presents a new mechanism for its origin and development. The results show that: (1) the large-scale dunefield bears a tile-style framework overwhelmingly composed of transverse barchanoid ridges perpendicular to the predominant winds, and inlaid diverse blowouts. (2) The small-scale dunefield, referred to as a primary structural unit of the large one, is typical of an incipient dunefield, following the same rules of evolution as the larger. (3) A succession of barchanoid ridge chains can steadily migrate downwind in much the same manner as surface wave propagation in air or water stimulated by an incised valley, and ultimately tend to bear roughly the same wavelength and amplitude under stable climate and hydrologic regimes. (4) The first ridge chain acquires its sand source substantially from the downwind escarpments exposing the loose Quaternary sandy sediments to the air, while the ensuing ridges derive their sands dominantly from in situ deflation of the underlain Quaternary loose sandy sediments in blowouts, partly from the upwind ridges through northern elongated horns. Theoretically, the sands from riparian escarpments can be transported by wind to the downwind distal end of a dunefield after sufficient long duration. (5) The lower Laohahe region experienced probably three significant climatic changes in the past, corresponding to the three active dune belts, suggesting that once a large-scale dunefield occurs, it is nearly impossible to be completely stabilized, at least in its central portions. At present, seasonal shrinkage and stagnation of the lower Laoha River, widespread farming and afforestation in the valley, and establishing windbreaks downwind of the valley as well as surrounding the dunefield, appear to have significantly modified local flow fields and sand sources, engendering significant degradation of the dunefield.

lower Laoha River; large-scale dunefield; wavy barchanoid ridges; topographic perturbation; blowouts

1 Introduction

Riparian source-bordering dunefields are a pronounced feature in arid and semiarid zones worldwide(Page, 1971; Bullard and Nash, 1998; Bullard and McTainsh, 2003; Ivestera and Leigh, 2003; Maroulis et al., 2007). This includes the semiarid zones in China,especially on the Xiliaohe Plain, Northeast China. In this unique region, there are some large-scale source-bordering dunefields along several river courses such as the Laoha and Wiliji Mulun rivers. The most striking is on the right flank of the lower Laoha River, a tributary of the Xiliao River, extending some 63 km long and 11 km wide, trending NE-SW and perpendicular to the local predominant northwester.

Han et al. (2007) reported several cases of actively forming riparian source-bordering dunes in the Xiliaohe Plain, including the region in question, and they thereby put forward a novel model for the origin and development of dunefields on the extensive plain. Nevertheless, the model does not offer detailed aerodynamic mechanisms and is, in fact, somewhat speculative although it is supported by field observations. Moreover, the dunefield bears not only a large spatial dimension, but complicated structures and patterns, suggesting a long evolutional history, apart from the influences from human activity and climatic change. The necessary evidence and clues to reflect the characteristics of evolution stages or processes are too rare to deduce the mechanism of its origin and development. There is thus an urgent need for pertinent examples to thereby give a reasonable explanation. Thanks to the availability of high-resolution satellite imagery (e.g., Google Earth) and better accessibility to remote core areas, where it was very difficult to access in the past, there are more possibilities to explore the intricate spatial patterns and structures with more convincing data to summarize the inherent mechanism.

The main objectives of this paper are to describe firstly the primary characteristics of the dunefield, its spatial pattern, composition and fine textures; then,analyze an analogue case, in which a small-scale riparian dunefield is actively forming and developing,with special respect to the spatial pattern, composition and structure; and finally, provide a comparison between the two riparian dunefields, and consequently draw conclusions.

2 Natural context of study area

2.1Physiography

The Laoha River, a principal tributary of the Xiliao River, rises on the eastern slopes of the Qilaotu Mountains, runs through the southwestern parts of the Xiliaohe Plain, eastern Inner Mongolia, China, and finally enters into the Xiliao River at the location of 43°25′57″N and 120°46′41″E, extending some 425 km and having the drainage of roughly 29,709 km2. Its mean annual discharge attains 4.03×108m3. Because of the controls by tectonic faults, the river has several acute turns, of which the last one is the starting point of the dunefield studied (Figure 1).

The study area is located mainly in the lower reaches of the river, approximately from 42°56′24″N and 120°14′02″E to 43°20′54″N and 120°45′57″E. The section of the valley adjacent to the large dunefield trends NE-SW, largely following a tectonic fault with the identical dip to the northeast. The stream in the valley floor takes a sinuous, somewhat meandering form. The river has incised into the surrounding plain ca. 10 m, underlain by Quaternary mediateand fine-grained sediments. The widths of the valley vary from the narrowest of 760 m at the last acute turn in the south to the widest of 6,700 m near to the confluence with the Xila Mullen River in the north (Figure 1b).

In the valley floor, there are few low vegetated dunes, some oxbow lakes, stretches of croplands and windbreaks. In addition, west of the valley are extensive dunefields as well, either active or partial stabilized by vegetation, some of which have encroached into the valley and buried portions of the floodplain.

2.2Geology

The central parts of the Xiliaohe Plain are situated in the second subsidence zone of Neo-Cathaysian Tectonic System, and is therefore abundant in Quaternary loose sandy sediments. It is bounded by the Great Tsingan Ranges to the northwest, trending NE-SW, and low mountains and loess hills to its south,opening to the east and forming approximately a triangle shape.

The study area is mantled dominantly by Quaternary sediments, derived essentially from fluvio-lacustrine and aeolian deposits with pale gray color. Their depths attain generally 100 m, with a maximum of 200 m, and the grain size is 0.01-0.25 mm, pertaining fine to very fine grains, very susceptible to wind deflation. The minerals overwhelmingly consist of quartz and feldspars in weight with a few dark minerals, and the sands are well-sorted and well-rounded, reflecting long-periods of aeolian action(Qiu, 1989). Generally, they are overlain extensively by Holocene dunefields and current fluvial deposits along the river valley.

2.3Climate

A semiarid monsoon temperate climate controlsthe area. The principal synoptic influences on rainfall in the drainage basin are in the form of subtropical air incursions from the Pacific in summer, which pass over the broad Yanshan Complex and the Jibei-Liaoxi Mountains in the south, and dry-frigid strong winds resulting from the burst of Mongolian High in winter blow over the Great Tsingan Ranges. The former can bring unstable but mild rainfall, and the latter frequent high winds.

Mean annual temperature at Daqintala Town,Naiman Banner, Inner Mongolia, 22 km southwest of the study area, is 6.0-6.5 °C, and mean annual precipitation is about 366 mm, concentrated in June to August. The variability of rainfall is rather high, larger than 45%, and consequently annual and inter-annual water levels and discharges of the stream vary dramatically,and mean annual evaporation reaches about 2,300 mm/a, much more than the precipitation. Mean annual wind velocity ranges from 3.6 to 4.1 m/s, and the strongest winds come largely from the northwest, but those from the south and southwest are by no means negligible, especially in spring. Local threshold velocity for entraining sand grains by wind is 4.9 m/s at standard 10 m height determined by field observation. The sum of gales (larger than 17 m/s) reaches 23.2 days and dust storms occur frequently in spring.

Figure 1 Sketch map of study area

2.4Vegetation and soils

The native typical zonal vegetation is primarily steppe with sparse trees, predominantly being Ulmus pumila, Salix matsudana, and Populus davidiana. The common grasses include Stipa grandis, Cleistogenes squarosa, Agropiron cristatum, and Aneurolepidium chinense. There are a few native shrubs, e.g.,Caragana microphylla, Lespedeza davurica, and Hedysarum fruticosum var. lignosum. There are also some psammophytes on the dunes, such as Salix gordejevii, Artemisia halodendron, and Agriophyllum squarrosa, some aquatic and limnetic species on the floodplains of valley, such as Phragmites communis,Typha minima, and Plantago depressa, and some halophytes on or around playas and pans, such as Achnatherum splendens, Iris ensata, and Puccinellia tenuiflora. In fact, because of widespread desertification and extensive farming, the natural vegetation in the area is decreasing, largely substituted by dunefields,interdune flats with sparse shrubs, and croplands.

Accordingly, the local dominant soils are zonalustic isohumosols (previous 'chestnut soils') with thick humus layers, mainly occurring in flat interdunes and riverine terrains. The soils exhibit a light texture that is vulnerable to wind erosion. Also, sandic primosols,alluvic primosols, orthic Halosols and orthic gleyosols cover diverse aeolian sediment, floodplains, playas or pans, and shallow waters, respectively.

3 Primary characteristics of the dunefield

3.1Spatial extent

By examining satellite imagery, it would be perceived that active aeolian sand movement and dune dynamics, manifested by a light color, are characteristic of the large-scale downwind riverine dunefield on the right flank of the lower Laoha River. It spans nearly 63 km from the southwestern end (site A) to the northeastern end (site B), and some 11 km in the widest transects, the central section (marked by C in Figure 1c) of which is well-developed and extend farthest downwind. The southern section (S) begins shrinking to 6.5 km, while the northern one is the narrowest, approximately 1.5-2.0 km. The whole dunefield covers an area of roughly 368 km2(Figure 1c). In terms of spatial size, this is the largest dunefield in the whole Korqin Sandy Land, and activity or mobility is the highest.

3.2Spatial patterns and structure

At first glance, the extensive dunefield seems to be uniform stretches of mobile dunes, nevertheless there are significant differences in its spatial structure and pattern in a fine resolution. The dunefield is mainly composed downwind by three dune belts:

1) The first belt (B1), immediately in the vicinity of the valley margins, which the extending distance along the river is the longest among the three belts,whereas its width is variable and the minimum is about 1.2 km. Its sinuosity is largely governed by the river valley. However, the upwind margins have, to a very large extent, been stabilized due to a withered stream,widespread farming and afforestation.

2) The central belt (B2) is transversely continuous,but merely extends from southern end to the location of Ln, approximately 2/3 of the full length of the dunefield, up to 43 km long. The widths range from 2.0 km to 5.0 km, and at some sites it is fused into the third belt. Some small-scale stabilized dunefields separates the two belts.

3) The third belt (B3) bears the largest width of 1.2-6.1 km, extending nearly 41 km long. Its span length and location are the same as B2. Further downwind, stretches of well- and semi- stabilized dunes are scattered on the extensive sandy plains with broad interdune flats generally occupied by farmlands and villages.

3.3Dune morphology and typology

As described above, the overall pattern of the dunefield is composed of three belts, extending in the direction of NE-SW paralleling to the valley, at a global scale. When viewing at an intermediate scale, it become immediately clear that the dunefield shows a fish-scale form, constituted by various dunes, and both interdune blowouts and coppice dunes.

3.3.1Dunes as depositional forms

The family of dunes of the dunefield encompasses mainly barchan, barchanoid ridge, and coppice dune. Barchans bear generally normal morphology, e.g., a stoss slope upwind, a slip face and two horns. The majority of barchans have a single sinuous steep crest,which coincides with the brink line. Some manifest the departure between crest and brink, that is to say, they have a rounded crest. In principle, separate or isolated barchans are uncommon and not typical.

Barchanoid ridges have the most outstanding characteristics in the whole dunefield, trending approximately the direction of NE-SW perpendicular to the local dominant winds. In light of their appearance and spatial pattern, they resemble the arrays of wave chains on coastal shores, where wave chains occur on the water surface, while subaqueous ripple and ridges steadily form and develop. Meanwhile, they essentially derive from interlocking and overlapping lateral barchanoid dunes, extending as far as 800 m. Their spatial arrangement reflects a definite adaptation to the local predominant wind direction of NW with a slight variation (Figure 2).

Ridge height is of spatial differentiation, i.e., the southern, central and northern sections range over 11-23 m, 13-30 m and 10-23 m, respectively. Accordingly, the angle of slip faces attains a maximum of 34° in spring, almost the natural repose angle of dry sands, while in summer it decreases slightly to 28°-32°. The spacing (the distance between two ridges)varies widely, and has spatial discrepancy, which the southern, central and northern sections range over 100-270 m, 140-290 m and 100-200 m, respectively.

In addition, some barchanoid ridges take the direction of nearly east-west with a narrow steep ridge,and could attain a maximal distance of 475 m. For instance, there are at least four, one, four and three ridges of the type in the right section of Figure 2a, the center of Figure 2b, the central and bottom right parts of Figure 2c, and the central right in Figure 2d, respectively. In terms of their morphology, they are similar to those described and following the samemechanism expounded by Bagnold (2005) which fundamentally result from the interactions between two dominant wind flows, one from NW and another from SW, thereby the resultant direction being E-W.

There are a number of coppice dunes in the interdune areas, vegetated generally by native bushes such as Salix gordejevii, Caragana microphylla, Artemisia halodendron, and Agriophyllum squarrosa. In fact,they are, in origin, derived from either interdune flats or stabilized antecedent dunes. They, in the most part,suffer strong deflation upwind by wind, and unavoidably receive some deposits at the lee sides or trap some sands in bushes. The sand sources are principally from upwind blowouts.

Figure 2 Spatial pattern and structures of the dunefield. The right two images magnified (b and d) correspond to the two rectangles in the left two images (a and c), respectively. The top two were photographed in spring, the bottom two in summer. BR = barchanoid ridge, bl = blowout, er = eroded ridge, tr = trough, cd = coppice dune. The central positions of images a, b, c, d are 42°59′22″N and 120°20′47″E, 42°59′22″N and 120°20′57″E, 43°14′02″N and 120°36′42″E, 43°14′05″N and 120°36′25″E, respectively

3.3.2Blowouts as erosional forms

In the vast dunefield, blowouts are generated and ubiquitously well-developed in interdune areas. Their size, morphology and arrangement are variable, primarily depending on local landform, microclimate,vegetation and soil. They are separated by either longitudinal residual flats vegetated or dune ridges, which could indicate the local dominant wind direction. Most are juxtaposed laterally (Figure 2d), whereas the rest are arrayed in a cascade form (top left portion of Figure 2b), longitudinally. The lowest points lie upwind side, close to the windward slip faces, which the blowouts as such are progressively buried by fallout and avalanching of aeolian sands. Such blowouts usually show irregular outlines.

In addition, many blowouts are in the form of troughs (Figure 2d), thereby the ridges between the troughs are well-developed, largely perpendicular to the predominant barchanoid ridges. The unique ridges are largely bounded by linear shrubs and have a bare crest, extending as far as 265 m long and 7-15 m wide, the axes of which indicate the dominant wind direction.

The appearance, form and outline of blowouts vary substantially with seasons. In both winter and spring,in which violent winds deflate interdune blowouts and drive dunes steadily forward, the appearance is of a grey tint such that the eroded ridges in interdune areas become unclear. Due to a full growth of plants insummer and a weak wind regime, the residue eroded ridges become recognizable owing to the flourishing shrubs aligning along the ridge flanks.

Since all characteristics, spatial patterns and structures described above are insufficient to sum up and deduce their processes and inherent mechanisms,we therefore resorts to some analogous cases surrounding the extensive complex dunefield.

4 A small-scale dunefield from the lower Xiangshui River as an optical analogous case to that of the lower Laoha River

Roughly 85 km west of the study area lies the lower reaches of the Xiangshui River, a tributary of the Xila Mullen River, running north obliquely, where a small-scale riparian dunefield in its infant stage has formed and at present is steadily developing. Considering that it bears almost the same structure and spatial pattern as those downwind of the lower Laoha River, it can be thought as an optical prototype which can be applied to the Laoha River's case so that an ideal mechanism can be deduced.

4.1Distribution and spatial pattern

The lower reaches of the Xiangshui River are located in the northern central parts of Wengniute Banner, a neighbor of Naiman Banner, eastern Inner Mongolia, and a small-scale dunefield just on the central part of the right flank of the river downwind is in progress, central position being 43°08′04″N and 119°18′54″E. It covers an area of some 5.17 km2, and extends nearly 3 km downwind from the riparian escarpment to the distal sand sheets.

By examining the color, fine-scale features and structures, the whole dunefield can be divided into three zones: the deflated ramps with a gentle slope just adjacent to the valley, the barchanoid ridge chain array,and the distal sand sheets. The first zone experiences long deflation by sustained high winds, and consequently the sand mass is transported downwind and deposited on the stoss slopes of the first ridge chains. The second zone consists of a few arrays of barchanoid ridge chains, some of which are nascent and short of typical morphology. The third zone comprises patches of sand sheets, extending lingulately downwind,sometimes like streaks.

There are 6-7 chains of transverse ridges, attenuating in length downwind till disappearance. Distinct blowouts and swales scatter randomly between the ridges, many of which are carved into the Quaternary pale sediments and showing the contour-like lines,suggesting that the ensuing ridges seem to obtain their sand supply mainly from these blowouts. Maybe because of the limited sand sources for each unit of the barchanoid ridges, the ridge heights vary, the maximum reaching 17.7 m and the minimum being 7.8 m for the second row of ridges (Figure 3). The spacing between barchanoid ridges is generally in the range of 250-330 m. Moreover, owing to the influence of valley morphology and consequently modified airflow field, the ridges on the two downwind flanks are not well-developed as those along the central line.

4.2Structure and aeolian landforms

The barchanoid ridges are composed of multiple elements which essentially are barchans, and yet northern or upper horns could extend a long distance of 200-300 m (Figure 3c). The extent of lateral coalescence of the barchanoid ridges is not high, i.e., the ridges develop imperfectly mainly due to their short history. Among all barchanoid ridge chains, the fourth is typical, well-interlinked laterally. It can be expected that over time the continuity and consistency of certain ridge chains can develop to a perfect form.

Between ridges are interdune blowouts, flats, and sometimes coppice dunes. There are a large amount of blowouts and swales between the ridge chains, some of which immediately lie on the toes of slip faces, manifesting the pale color and contour lines. At present, the blowouts and swales are being buried by sand with the advance of ridges (Figures 3b, 3c). Nevertheless, their downwind ends and the toes of ensuing stoss slopes are exposed to the robust wind experiencing violent deflation, indicating that the blowout will migrate downwind with time. As a result of this process, the ridge system would encroach step by step following the dominant wind. Some existing blowouts behind the upwind slip faces are progressively buried, while others downwind are being developed by strongly deflating the flats or rudimentary blowouts, either juxtaposed transversely or arranged in a cascade.

It is worth noticing that there are a large amount of eroded ridges from interdune flats between blowouts,which can attain up to 200 m long and 5-10 m wide(Figure 3c). They largely trend the local predominant wind of NW-SE, and some troughs could develop when two neighboring eroded ridges are quite close and the paths of sand transportation are rather narrow(Figure 3b). Nevertheless, there are few transverse eroded ridges on the downwind ends of blowouts. Overall, longitudinal eroded blowouts prevail in the interdune areas.

The swales are generally transient ponds after heavy rains in summer, while those with standing water are scarce in the region. Interdune flats have survived up to the present largely because of the younger ages and better vegetation cover, but at some sites they are already sculptured into quasi-yardang landscapes or mesas.

Also, there are coppice dunes between the barchanoid ridges, either large or small, either separate or clustered, usually covered by local native shrubs such as Salix gordejevii, Caragana microphylla, Artemisia halodendron, and Agriophyllum squarrosa. They generally have a height of 5-8 m and 5-12 m across.

Figure 3 Spatial pattern and structures of the small-scale dunefield. The images in Figure 3 show the synoptic spatial pattern, and its fine-scale structures. (a) shows the overall view of the small dunefield, while (b) and (c) are magnified portions of the blue rectangles(T1 and T2) in (a), respectively. All three images manifest distinctly the spatial correlations between barchanoid ridges, blowouts,diverse eroded ridges and coppice dunes. Here, br=barchanoid ridge, bl=blowout, er=eroded ridge, tr= trough, cd= coppice dune, and uh= upper horn or northern horn. The central positions of images (a), (b) and (c) are 43°07′57″N and 119°18′50″E, 43°08′13″N and 119°18′36″E, 43°08′08″N and 119°18′57″E, respectively

5 Discussion

5.1Similarity assessment between the two cases

As mentioned above, the two dunefields are substantially identical irrelevant of origin, spatial structure or pattern, although they have significant differences in size, composition and spatial structures. However,the small one can supply evidence and details to sum up the mechanisms for dunefield history.

To begin with, the two dunefields are situated on the downwind flanks of lower streams, merely with differences in discharge and valley dimensions. The Laoha River has a discharge of 4.03×108m3/a and a valley 3 km wide (in its central portion), while the Xiangshui River is only 4.73×107m3/a and 295-450 m,respectively, which the former is 8.52-fold and 8.33-fold compared to the latter, respectively.

Secondly, their natural settings are essentially consistent, i.e., similar Quaternary sediments, wind regime, and landform as mentioned above.

Thirdly, with special relation to the fine-scale features, the spatial patterns and structures are nearly identical, dominated by transverse barchanoid ridges and interdune blowouts, whereas the remarked difference lies merely in the degree of complexity.

However, it must be emphasized that the dunefield of the lower Laoha River bears a more large-scale and high-order spatial pattern, i.e., spatial differentiation downwind into three belts, while that of the Xiangshui River only one. The difference might be ascribed to the long history and multi-phase climatic variations of the large dunefield, even though they are both very similar and obey fully the scaling rule.

Thus, it can be inferred that the small-scale dunefield is experiencing its incipient stage of a large-scale dunefield, in the same mechanism. A large-scale dunefield is primarily derived from elementary small-scale forms, thereafter acquiring stable growth in dimension by its own dilation and coalescence of surrounding small dunefields, and the former will experience a longer history than the latter.

5.2The issue of sand sources

In the light of field observations, no bare sand deposits are seen in the whole valley, where floodplains are covered by channels, swamps, lakes and meadows with dense grasses and some trees. The escarpments on the right margins have suffered from violent deflation so that the striking rims have come into being. And, the rims and first ridge chain, even the second ridge chain at some sites, manifest pale grey color on satellite images, suggesting that the dark minerals from underlain sediments have no enough time to be decomposed and stained by ferric substances under surface environments. Meanwhile, there are always some interdune blowouts or swales, if any,suffering the burial of upwind slip faces, from the first ridge chain onwards. Thus, it can be concluded that the sand is derived essentially from downwind escarpments on the margins of the river as an initial source,and from local underlain Quaternary sandy sediments from the second ridge chain on, which the transporting distance of these dune sands is short and local. There is no evidence to suggest direct fluvial sources on floodplains for both cases. This conclusion is consistent with that drawn by Muhs and Holiday (2001) in southwestern U.S.A..

However, it should be pointed out that with sufficient time, a great deal of sand can be transported downwind through either the nearly easterly secondary ridges (upper horns or northern horns) between any pair of dominant ridges (Figures 2 and 3), or the slow advance of barchanoid ridges. This implies that with sufficient time, escarpment sand could be transported to the distal extremes downwind. Probably, it is the time factor that causes some sharp disputes on whether the main provenance is actually autochthonous or allochthonous (Fryberger and Ahlbrandt, 1979;Mainguet and Vimeux-Richeux, 1981; Wasson, 1983;Pell and Chivas, 1995).

Overall, substantial sources of sand are derived from local underlying Quaternary sediments, either fluvial, aeolian or lacustrine, or their combinations,and certainly modern soil or dune sand.

5.3Impetus and valley's functions for dunefield development

As for the origin and development of a dunefield,two factors should be taken into consideration, i.e.,powerful aerodynamic driving actions and necessary initial sand sources. An abundance of field observations have revealed that gentle or mild wind, even intermediate-intense winds, larger than the local thresholds for drifting-sand, exert generally on ripple and ridge on the surface of dunes, and consequently their transportation rates of sand are not large enough so as to deliver a large volume of sand mass. Also, only the winds pertaining to high wind events larger than 17 m/s, e.g., gales or gusts, can undoubtedly transport sufficient volume of sand such that dune morphology is substantially modified.

According to the formula given by Bagnold (2005),the transport rate q is proportional to the term (v-vt)3,therefore, the ratio of q for gale of 22 m/s to mild wind of 8 m/s can attain nearly 182, which the local threshold wind is 5 m/s in the standard height of 10 m. This suggests that the large-scale primary spatial pattern and structure of a dunefield entail sustained high wind events, reflecting an equilibrium between aerody-namics and bedforms, whereas the other drifting winds are by no means negligible, especially for local finescale features.

Viewed from the entire Xiliaohe Plain, direct sand sources on the ground for aeolian actions are not abundant and efficient, deriving either from modern fluvial deposits or existing dunes, ancient or modern. The problem is how to determine the sand source when a new small-scale dunefield initiates its primary pattern and structure. A river incised into the extensive fluvial plain is the most reliable impetus to render the necessary conditions for a new dunefield in light of intensive field investigations and in situ observations. An active river enables the exposure of rich sandy sediments underlying the downwind escarpments, and suitable flow field, where the upper escarpment could receive the largest shear stress due to topographic forcing. Furthermore, the active lateral erosion by streams continuously delivers bare ramps with abundant sands by way of river swing in the valley floor.

Aerodynamically, there are wave motions within near-surface boundary layers when windflow passes a sharp or significant topographic change on an extensive plain, specifically analogous to the wave action on coastal shores. In fact, a river valley is a distinct linear depression, which would give rise to remarked perturbation to the dominant windflow and therefore induce strong surface waves in the lower atmospheric boundary layers whenever a high wind event occurs(Andersen et al., 2001; Elbelrhiti et al., 2005; Andreotti et al., 2009). Valley perturbation is variable with the valley's dynamics, and generally becomes more significant with widening of the valley, albeit the thickness of turbulent boundary-layer induced by the valley is merely up to several ten meters or so. The valley effect can be sustained and propagated steadily downwind, one barchanoid ridge after another, but its intensity, i.e., amplitude and spacing (or wavelength) of ridge chains will attenuate downwind just as the usual damping waves in certain medium(e.g., water, air). It is quite difficult for wind to dig into the ground of interdune blowouts over a given depth, where underground water levels dictate strongly downward deflation.

Thereby, it can be concluded that, to a very large extent, the history and characteristics of a stream bears a close correlation to those of a downwind dunefield,and even exert a decisive impact on the latter.

5.4The episodes of dunefield evolution

A separate dune, say a barchan, needs not only an appropriate flow field and sufficient sands, but also subsequently a certain period of time to form and sustain its essential morphology. Thus, a large-scale dunefield will definitely take a longer time than a small one, with disparate spatial dimensions and patterns. This evidently demonstrates that in the dunefields of the lower Laoha and Xiangshui rivers, the former is of a pattern of the superior order of three-belt structure,and the latter lacks such belt structure.

The small dunefield of the lower Xiangshui River can be viewed as an elementary order structural unit,corresponding to a 250 m wide valley and a dunefield of roughly 3 km propagation distance, whereas the dunefield of the lower Laoha River is a complex wind action system, with an average valley width of 2.5 km,and roughly 11 km propagation distance. Their spatial patterns and structures show that they largely attain equilibrium between the bedforms and regional aerodynamic regimes, even though the small-scale dunefield is still expanding downwind and laterally its spatial dimension.

As to the large dunefield, each dune belt represents a given episode in the history of whole dunefield,where the riparian belt immediately adjacent to the valley should have come into being firstly as initial sources of dunefield, followed by the central belt,finally the downwind distal belt. This is a propagation process of whole topographic wave series in space, not atmospheric wave series. In terms of spatial extent, the distal belt has perhaps experienced the longest transportation distance, which was not completely stabilized during the return of a humid climate, and might be the results of the maximum aeolian actions in the past. The riparian belt experienced the shortest time according to its width, where numerous sections of the frontal rims having been stabilized mainly because of concurrent interactions of seasonal shrinkage and stagnation of the lower Laoha River, widespread farming and afforestation in the valley, and establishing windbreaks downwind of valley as well as surrounding the dunefield.

Concerning semi- and fully- stabilized dunefield outsides of the distal belt, it is quite possible that they formed from a more distant past as a regional background for present-day active dunefields, and not the product of the lower Laoha River.

Moreover, past studies undertaken in this region show that extensive active dunefields formed largely in the Holocene, especially since 5,000 B.C. (Hu, 1989;Qiu, 1989; Zhang, 1989; Qiu et al., 1992; Yang et al.,2011). Still, in the regions of Korqin with large-scale active dunefields, there are few dunes stabilized prior to the 5,000 a B.C. and survived till now. Under the circumstances of dune ridge propagation as wave,aeolian sands are transported steadily downwind, and leeside soils would be definitely buried by upwind dunes or destroyed by deflation in blowouts. As a result, remnant palaeosols are too sparse to be encountered, and even if some can be found the age data are rather diverse, either by C14 or OSL techniques,suggesting the problem of multi-generations. It implies that the vast dunefield on the right flank of the Laoha River almost impossibly allows us a clear and definitive timetable of its development history under current conditions.

6 Conclusion

The large-scale dunefield on the right flank of the lower Laoha River, Northeast China, is the most striking feature on the extensive Xiliaohe Plain for its attractive colour and shape. By means of remote sensing imagery analysis and field observation as well as a comparison with a small-scale dunefield on the right flank of the lower Xiangshui River, this paper puts forward a new mechanism for its origin and development. The tile-style framework is overwhelmingly composed of transverse barchanoid ridges perpendicular to the predominant winds, and encircled diverse blowouts demonstrates an essentially stable drifting-sand wind regime. The small-scale dunefield of the Xiangshui River is typical of an incipient dunefield, with the same spatial pattern and structure and following the same rules of development as the large-scale one, therefore it is able to provide valuable evidence and clues for summarizing a general origin and development rule. Moreover, a succession of barchanoid ridge chains can steadily migrate downwind in much the same manner as surface wave propagation in air or water stimulated by an incised valley, and ultimately tend to bear roughly the same wavelength and amplitude under stable optimal climate and hydrologic regimes.

With respect to the provenance of aeolian sand, it can be inferred that the first ridge chain acquires its direct sand source substantially from downwind escarpments as outcrops of the loose Quaternary sandy sediments, while the ensuing ridges derive their sands dominantly from in situ deflation of the underlain loose Quaternary sandy sediments in blowouts, partly from the upwind ridges through northern elongated ridges. Chronologically, the vast dunefield experienced three significant climatic changes in the past,corresponding to the three active dune belts, suggesting that once a large-scale dunefield occurs, it is nearly impossible to be completely stabilized under natural conditions. At present, seasonal shrinkage and stagnation of the lower Laoha River, widespread farming and afforestation in the valley, and establishment of windbreaks downwind of the valley as well as surrounding the dunefield, appear to have significantly altered local flow fields and sand sources, engendering significant degradation of the dunefield.

Acknowledgments:

The study is funded by NSFC (Grant No. 41271025) and the Construct Program of the Key Discipline in Hunan Province, China (2012001). We would like to sincerely thank Prof. XueYong Zou and Prof. Hong Cheng of the State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Nornal University, for their support in instrumentation and vehicles. Meanwhile, we are grateful to the staff of the Experimental Station for Sandy Desertification Research of Naiman, CAS, and of the Forestry Bureau of Wengniute, Inner Mongolia, for their warm and generous aid during field work.

Andersen KH, Chabanol M, van Hecke M, 2001. Nonlinear hydrodynamics and turbulence: Dynamical models for sand ripples beneath surface waves. Physical Review (Section E), 63:66308-66315.

Andreotti B, Fourrière A, Ould-Kaddour F, et al., 2009. Giant aeolian dune size determined by the average depth of the atmospheric boundary layer. Nature, 457: 1120-1123. DOI:10.1038/nature07787.

Bagnold RA, 2005. The Physics of Blown Sand and Desert Dunes. New York: Dover Publications, Inc..

Bullard JE, McTainsh GH, 2003. Aeolian-fluvial interactions in dryland environments: examples, concepts and Australia case study. Progress in Physical Geography, 27(4): 471-501. DOI:10.1191/0309133303pp386ra.

Bullard JE, Nash DJ, 1998. Linear dune pattern variability in the vicinity of dry valleys in the southwest Kalahari. Geomorphology, 23: 35-54. DOI: 10.1016/S0169-555X(97)00090-1.

Elbelrhiti H, Claudin P, Andreotti B, 2005. Field evidence for surface-wave-induced instability of sand dunes. Nature, 437:720-723. DOI: 10.1038/nature04058.

Fryberger SG, Ahlbrandt TS, 1979. Mechanisms for the formation of eolian sand seas. Z. Geomorph. N. F., Bd., 23: 440-460.

Han G, Zhang GF, Dong YX, 2007. A model for the active origin and development of source-bordering dunefields on a semiarid fluvial plain: A case study from the Xiliaohe Plain, Northeast China. Geomorphology, 86(3-4): 512-524. DOI: 10.1016/j. geomorph.2006.10.010.

Hu M, 1989. The Holocene environmental changes in Kerqin Sandy Land. Journal of Arid Land Resources and Environment,8: 51-58.

Ivestera AH, Leigh DS, 2003. Riverine dunes on the Coastal Plain of Georgia, USA. Geomorphology, 51(4): 289-311. DOI:10.1016/S0169-555X(02)00240-4.

Mainguet M, Vimeux-Richeux M, 1981. Autochtonie et allochtonie des sables éoliens: Etude des minéraux lourds de la couverture sableuse du Niger. Comptes Rendues de l'académie des Sciences (Série II), 292: 1437-1440.

Maroulis JC, Nanson GC, Pricea DM, et al., 2007. Aeolian-fluvial interaction and climate change: source-bordering dune development over the past 100 ka on Cooper Creek, central Australia. Quaternary Science Reviews, 26(3): 386-404. DOI: 10.1016/j. quascirev.2006.08.010.

Muhs DR, Holiday VT, 2001. Origin of late Quaternary dune fields on the Southern High Plains of Texas and New Mexico. GSABulletin, 113(1): 75-87. DOI: 0.1130/0016-7606(2001)113<0075:OOLQDF>2.0.CO;2.

Page KJ, 1971. Riverine source bordering sand dune. Australian Geographer, 11(6): 603-605. DOI: 10.1080/0004918710 8702599.

Pell SD, Chivas AR, 1995. Surface features of sand grains from the Australian Continental Dunefield. Palaeogeography, Palaeoclimatology, Palaeoecology, 113(1): 119-132. DOI: 10.1016/ 0031-0182(95)00066-U.

Qiu S, 1989. Study on the formation and development of Korqin Sandy Land. Scientia Geographica Sinica, 9: 317-328.

Qiu SW, Li QS, Xia YM, 1992. Paleosols of sandy lands and environmental changes in the western plain of Northeast China during Holocene. Quaternary Sciences, (3): 224-232.

Wasson RJ, 1983. The Cainozoic history of the Strzelecki and Simpson dunefields (Australia), and the origin of the desert dunes. Z. Geomorph. N. F., Suppl. Bd., 45: 85-115.

Yang LH, Lai ZP, Zhou J, et al., 2011. Dune evolution in Horqin Dunefield of China since Late Glacical Period inferred from luminescence dating and pollen. Scientia Geographica Sinica,31(6): 695-701.

Zhang B, 1989. The changes of the Horqin Sandy Land before the North Wei Dynasty (before 534 AD). Journal of Desert Research, 9: 37-45.

Han G, Zhang GF, You L, et al., 2015. A mechanism for the origin and development of the large-scale dunefield on the right flank of the lower reach of Laoha River, Northeast China. Sciences in Cold and Arid Regions, 7(1): 0029-0039. DOI: 10.3724/SP.J.1226.2015.00029.

*Correspondence to: Dr. Guang Han, Prof. of Hunan Normal University. No. 36, Lushan Road, Yuelu District,Changsha, Hunan 410081, China. Tel: +86-731-88872752; E-mail: hanguang@hunnu.edu.cn

August 16, 2014 Accepted: November 21, 2014