|Rural Development in White Nile Province, Sudan (UNU, 1986, 139 pages)|
|I. The physical environment of the White Nile|
R. A. Shakesby
This chapter reviews the geomorphology of White Nile Province between 14 00'N and 14°45'N, to the west of the White Nile. (This area will be refered to henceforth as the research area.) As with many other parts of the Sudan, little if any detailed geomorphological investigation has been attempted.
One fundamental problem in carrying out an analysis of the geomorphology of the research area lies in the absence of suitable topographic maps. The Quarter Million series published by the Sudan Survey provides little useful information for the geomorphologist. Furthermore, when used as a basis for locating aerial photographs and for field mapping it is clear that many wadi systems and other topographic features are generalized and many significant features of the human landscape omitted entirely.
Aerial photography at a scale of approximately 1 :40,000, carried out in the late 1950s, is available for most of the research area, and the geomorphological map (fig.1.4) and much of the interpretation of the landforms and sediment types have been based on this source. The lack of any later photography has meant that detailed comparative investigations of the extent of any sand encroachment or other changes could not be undertaken.
First of all, the underlying geological structure will be analysed and the superficial deposits described. Upon this basis the main landscape components will be examined and the palaeoand contemporary hydrology discussed. In the final section an assessment of the potential value for human use of each of the landscape components is attempted and some suggested lines for future research are indicated.
The lack of economically important mineral deposits in the research area has meant that there has been little incentive to carry out detailed geological exploration, and the area has until recently only been considered within the general synthesis of the geology of the Sudan (see, for example, Andrew, 1948; Whiteman, 1971; Vail, 1978). However, recently Vail (1982) has provided a summary of the geology of the Gezira region and its environs.
A generalized map of the geology of the research area is provided in fig.1.1. In only a small portion of the research area are solid rock formations known to crop out and these belong to either the Basement Complex or the Nubian Sandstone Series. The more recent superficial deposits in the area fall conveniently into three groups: the Umm Ruwaba Series; the qoz sands; and clay deposits.
This is a general term for the oldest rocks exposed in northeast Africa. These rocks consist of a largely undifferentiated group of gneisses and schists of probably PreCambrian age (Vail, 1978). The main outcrops in the research area are in the three main hill masses (Jebel Arashkol, Jebel Er Rereid and Jebel Tuyus) with their surrounding pediments (fig. 1.4), though, as its name implies, this series underlies all other deposits. Jebel Arashkol has a strong north-east/ southwest tectonic trend. The Basement Complex here includes porphyritic granite, biotite schist, and gneiss, according to Lotongol (1976).
Vail (1978) indicates that there is a small enclave of Nubian Sandstone in the north-east corner of the research area where this formation forms a hard pebble-strewn pediment surface. This represents the southward limit of a large, almost uninterrupted expanse of this series covering central and northern Sudan. Nubian Sandstone is thought to be of Mesozoic ( Lower? Cretaceous) age, but because of the lack of diagnostic fossils or distinctive marker beds this series has been the source of much controversy within geological circles. The Nubian Sandstone Series commonly consists of poorly sorted, coarse to medium-grained, cream or brown sandstone with quartz pebbles and mud flakes (Vail, 1978). Thick lenses of mudstones and clays also occur. The origin of the sediments is unclear, though it seems that the sandstones and mudstones represent continental deposits laid down as flow or flood-plain materials by braiding rivers.
Umm Ruwaba Series
These deposits cover wide areas of central and southern Sudan including the central part of the research area, the northernmost occurrence of the series west of the White Nile (fig. 1.1). The series comprises unconsolidated continental sands, sandy clays, and clays that vary considerably in Ethology and thickness. Since, as with the Nubian Sandstone, diagnostic fossils are uncommon, the period of formation and stratigraphic subdivision also remain uncertain, though a late Pliocene or early Pleistocene date of formation is suggested. At this time, it is believed that a number of deep crustal warps became filled with sands and clays. The existence of saline deposits suggests periodic aridity during sedimentation (Grove and Warren, 1968). These deposits are important locally because they give rise to good agricultural land and provide an important source of underground water.
In literary Arabic qoz (pi. queizan) means sand dune, but in central Sudan it can also refer to a sandy soil, areas west of the clay plain, or a particularly loose, sandy road surface (Alam el Din, 1968). In technical literature, the word has also had slightly different meanings. Sandford (1935) used the term to mean stabilized sand dunes. Edmonds (1942) referred to dune-like accumulations of sand as qoz. More recently, however, the term has come to mean any kind of sand deposit. Warren (1970) introduced the terms "high" and "low" qoz in referring to stable dunes in Kordofan Province, the latter denoting areas where dune forms are subdued or absent and the former defining areas where dune relief is marked. Work to date on the qoz of central
Sudan has adopted a broad regional approach (e.g. Edmonds, 1942; Warren, 1966, 1970; Alam el Din,1968; Grove and Warren, 1968).
Ooz covers some 97,000 km2, from 10°N in the south to about 16°N in the north. The sand can vary in thickness from a thin covering to 50 m. The sands are composed almost exclusively of frosted quartz grains (Andrew, 1948). Under the scanning electron microscope (SEM), this frosting is seen to form the typical surface texture of upturned plates which are characteristic of aeolian environments (Bull et al., in press) (fig. 1.2 A and B).
Particle size curves for 12 surface samples of qoz in the research area are shown in fig.1.3 (sites located on fig.1.4). It is readily apparent that the qoz deposits vary appreciably in grain size within the range 4 to 2 ? (where ? represents -logz of the diameter of a particle in millimetres). Three samples (1,4 and 5) stand out as having particularly good size-sorting characteristics, with a large proportion of sand grains (about 62 per cent for sample 1) between 3 and 2 ?. This size grade forms the middle of the range of particle sizes that are particularly prone to saltation by wind (Bagnold,1941). Saltation involves a bounding movement of grains and normally accounts for 95 per cent of the bulk transport of sand in dunefields (Mabbutt, 1977). All three samples are derived from currently active wind-blown sand complexes. Sample 1 was collected from a small active crescentic dune near Et Tur'a; sample 4 from a shrub coppice dune near La'ota; and sample 5 from sand on the surface of the pediment near Jebel Arashkol. The high proportion of sand ranging from 3 to 2 ? in size in this sample is indicative of recently winddeposited sand forming a thin layer on the pediment surface. The remaining nine samples have a more even spread of particle sizes, with most particles between 3 ? and-1?? in diameter. Any movement by wind of these generally coarser sands would be by surface creep, a considerably slower movement process than saltation.
The source of the sand making up the qoz has been the subject of much debate. Grabham (1935) believed that granite was the source of the sands and clays found in Sudan. Whilst Edmonds (1942), Andrew (1948), and Tothill (1948b) favoured direct weathering of Nubian Sandstone, Kleinsorge and Zscheked (1958) maintained that other rocks, such as Basement Complex, have undoubtedly contributed much material. Vail (1978) notes that in northeast Africa qoz-type sands overlie a wide range of rock types of all ages and susceptibility to weathering. On the basis of dune trend patterns, he argues that the ultimate source area of the sand would have been the Sirte Basin in Libya, but that undoubtedly the southward-blowing winds would have entrained and deposited other suitable sandy materials as they swept across Egypt and Sudan.
In attempting to explain sand dunes on the east side of the White Nile, mostly situated between Hashaba and El Geteina, Williams (1968b) thought it more likely that these sands had been derived from local sandy alluvial deposits, because the dunes are associated with fossil meanders and the lightcoloured sand has not been ferruginized, as has much of the qoz. This process imparts the characteristic red colour to much of the surface qoz sands and is caused by the gradual acquisition of haematitic coatings on individual grains.
The conditions under which the reddening of qoz sands can arise have been debated. Grabham (1926) regarded it as the result of a "lateritic" climate, "a good deal more rainy than at present," whereas Edmonds (1942) favoured a period of "maximum aridity" to account for the colour. More recently, several workers have noted the prevalence of reddened dune sand in present-day and fossil desert environments and have attempted to derive a general theory relating dune reddening with age (Norris, 1969; Mabbutt, 1977). Norris (1969) considered that weathering leads to an increasing thickness of the haematitic coating and affects increasing numbers of grains with time. He argued that the basic requirements for the development are warm temperatures, oxidizing conditions, and the presence of moisture, and that these conditions are commonly met with even in very dry environments.
Grabham (1926), using the evidence of wells in Kordofan, found that the red coloration was confined to the upper 6 to 7 m of the sands and generally did not extend more than 3 m below the surface. Edmonds (1942) regarded 1.3 to 1.7 m as the general depth at which the redness begins to disappear. It has been suggested by some that the absence of red staining on certain desert sands indicates that the red coating has been eroded off the grains as a result of grain-to grain impact during aeolian transport, though Norris ( 1969) has shown that this is not necessarily true.
The qoz sands of the research area reflect many of these findings. The red colouring appears to be largely confined to surface layers and disappears with depth as conditions apparently become less conductive for iron staining to take place. The active sand samples ( 1 and 4) have amongst the lightest colours (10 yr 7/4 and 10 yr 614), and the sands broadly become increasingly red from east to west as the apparently older and stabler dunes are approached. Sands of the qoz on or adjacent to the clay plain (fig. 1.4) are yellowish in colour (e.g. sample 1,10 yr 7/4; sample 7, 10 yr 6/3; sample 12, 10 yr 5/3; Munsell Color dry), compared with the distinctly more reddish hue of sands further west (sample 6, 7.5 yr 4/6; sample 8, 7.5 yr 6/6; sample 10, 5 yr 5/8). Some of the paler coloured sands along the White Nile clay plain would appear to have been derived from local alluvial deposits
However, in some locations in the research area (e.g. 1 km west of Bileila) bedding was observed in slightly indurated reddish sand below the top 30 cm or so of loose sand. This has not been commonly reported and field work by Warren (1970) in Kordofan revealed no evidence of dune bedding.
It is generally agreed that the qoz deposits are Quaternary in age (e.g. Tothill, 1948b; Vail, 1978), but since neither marker horizons nor datable material has yet been found in association with the sands, no precise age can be assigned. Some workers have, however, attempted to determine broad phases of qoz formation. Warren (1970) considered that there had been dry and wet phases during the Quaternary and that most of the sand-blowing, and hence dune formation, took place during two dry phases, the second and main one dating from c. 25,000 to 12,000 B.P. Whiteman (1971) suggests a rather later dating for Warren's two main periods of qoz formation, namely 25,000 B.P. and 3,000-0 B.P. Where artefacts, pottery fragments, and shells are associated with dunes on or near the clay plains, it has been possible to indicate a minimum age of stabilization of these features. Williams (1966) noted that some dunes on the east bank of the White Nile have been stabilized for over 6,000 years.
In the study area, alluvial "clay" plains, as distinct from the old alluvial deposits of the Umm Ruwaba or Gezira formations, form a belt of land bordering the White Nile. The uppermost, grey alluvium is remarkably uniform, comprising clay with sand and silt but little coarser material apart from small weathered carbonate concretions. These clays are light gray in colour (typically 2.5 yr 5/2-10 yr 5/1) when dry, and prone to dessication cracking, but become dark gray, nearly black, when wet, creating a striking contrast with the reddish sands of the qoz at all times of the year. The clay deposits of the White Nile plains are known to vary in thickness from about 3 to 20 m, though exceptionally depths of over 40 m have been recorded (Ruxton and Berry, 1978).
Whilst the upper few metros of the clay plains are almost uniform, there are marked variations in stratigraphy with depth. Williams and Adamson (1980) found that in certain locations on the east bank of the White Nile the grey clay is underlain respectively by green clays, sands, quartz gravels and layers of calcite and dolomite, all within the top 5 m. In general, the clay plains become less clay-rich with depth.
The clay plains of the White Nile are broadly analogous to the upper few metres of clay deposits covering the vast triangular-shaped expanse of the Gezira, which lies east and north-east of the research area, between the White and Blue Niles.
There have been a number of suggestions as to the origin of the clay plains of the Gezira and White Nile. Grabham considered that the clay had an aeolian origin and that it was more or less contemporaneous with the qoz formation of Kordofan (Grabham, 1909, 1917). He considered the haboobs (dust storms) which occur in this area as a factor, and there can be little doubt that these are responsible for a certain amount of redistribution of fine particulate matter (Smalley and Vita Finzi,1968; Ruxton and Berry, 1978).
Tyler (1932) was of the opinion that the clays in the Gedaref region had been derived from the weathering of basaltic lavas. Ball (1939) believed in a lacustrine and Edmonds (1942) favoured an alluvial origin over long periods of time. Tothill (1946, 1948) attempted to date the fluvial activity responsible for the Gezira clays as between 50,000 and 10,000 B.P.
Recent work using radio-carbon dating and other modern techniques have elucidated the problem considerably (Williams et al., 1973; Williams et al., 1975; Williams and Adamson, 1973, 1980; Adamson et al., 1980). It seems clear that the upper few metros of the White Nile clay plains are indeed of alluvial origin and underwent inundation and seasonal flooding at various times during the late Pleistocene and Holocene. Sub-fossil molluscan assemblages, archaeological evidence, and radio-carbon dating all support this view. Analysis of the quartz grain surface textures from the clays of the research area clearly show features associated with fluvial transport (fig. 1.2 C). However, Ruxton and Berry (1978) call for great care in interpreting the age of these "aggradational" clays from the occurrence of sub-fossil mollusca. They point out that the plain could have been in existence prior to its being populated by molluscs, noting that these may become incorporated into lower layers by falling down cracks. Thus shells may indicate only a minimum age for the clay plains and for the particular layer in which they occur.
The stratigraphical relationship between qoz sand and the clay plains is not unequivocal. Sandford (1935) claimed that "cotton soil" (clay) north of Mellit (14°40'N, 26°E) was overlain by typical qoz sand. Venzlaff (reported by Ruxton and Berry, 1978) found qoz sand overlying up to 30 m of unconsolidated clayey sediments at Sodiri. Near Hashaba, on the east bank of the White Nile, dunes overlie clays, whereas on the west bank in the research area Williams and Adamson (1976) maintained that in the Arashkol palaeochannel (fig. 1.4) sands and clays interdigitate. Williams and Adamson (1980) argue that isolated qoz dunes in the research area have clearly been subject to scalloping and marginal trimming by White Nile palaeochannels and that many of the dunes are flanked and partially submerged by stratigraphically younger clays.
The most striking feature of the landscape in the research area is its almost flat, peneplaned appearance. Here, apart from three isolated jebels (with Jebel Tuyus reaching 657 m), the ground surface probably nowhere exceeds 420 m and falls to 377.8 m for the surface of the White Nile upstream of the Jebel Aulia Dam, giving a range of some 40 m. The origin of the vast peneplain of which the research area forms a part attracted the interest of geomorphologists during the 1950s and 1960s. L. C. King (1962) maintained that the pediplanation and deposition that operated during Jurassic times continued into the Cretaceous, leading to a unicyclic surface characterized by low relief. According to him, uplift occurred during the late Cretaceous, leading to erosion of much of the Nubian Sandstone in the research area and farther south. Crustal warping in the late Pliocene and early Pleistocene then gave rise to deep hollows in which the Umm Ruwaba beds were deposited. The qoz and clay deposits can thus be regarded merely as a blanketing of this peneplaned surface.
As a result of these events, three main landscape components can be identified in the research area: hills and pediments; qoz; and clay plains (fig.1.4).
Hills and Pediments
The three Basement Complex hill masses in the area (Jebel Arashkol, Jebel Tuyus and Jebel Er Rereid) are associated with gently sloping pediment surfaces. Other areas of pediment occur in the north-east of the research area and in the vicinity of Esh Shuqeiq. The pediments associated with the hill masses are characteristically mantled in debris with a greater range of particle size than on either the qoz or the clay. Coarse, angular, and weathered gravel and pebbles are abundant close to the hill slopes, but material becomes generally finer the farther it is from the base of the hill slope. This fining with distance might reflect the decreasing competence of khors in this direction or it may indicate that material farther away from the hill slopes has undergone greater weathering.
The pediment associated with Jebel Arashkol (557 m) lies on its west and north-west sides and is about 10 m above ground level immediately south-east of the jebel. Three main khors drain through the south-west/north-east aligned tectonic structure and have formed small alluvial fans on its east and south-east sides. As the alignment of these khors bears little relationship to the present-day topography, it seems that, as erosion proceeded on the pediment surface, the gradual emergence of the more resistant strata of the present jebel failed to disrupt the existing south-east and east drainage pattern. The fans formed by these khors are currently undergoing entrenchment which Lustig (1965) regarded as indicative of changes in discharge conditions associated with either climatic change or removal of vegetation by overgrazing. An alternative view of fan entrenchment put forward by Denny (1967) suggests that it is "the normal consequence of large variations in flood discharge." The material forming the alluvial fans may be distinguished from that on the pediment surface in that it has a generally narrower range of particle size. Both these sediment types contain a variety of minerals reflecting their origin from the Basement Complex, and are generally brown or gray-brown in colour. Wind-blown sands of the area are by contrast paler and contain a predominance of quartz grains.
There is clear evidence that Jebel Arashkol itself is subject to a considerable amount of weathering, with both massive exfoliation and spheroidal onion-skin weathering of individual blocks taking place. The outcrop is also undergoing block and granular disintegration with flaking and splitting. In addition, the effects of erosion by flowing water can be seen even on bare rock surfaces, where depressions have been at least partially eroded by episodic rill action. (Current research on the weathering and erosion of Jebel Arashkol and its surroundings is being undertaken by M. Campbell of University College, Swansea.)
Jebel Tuyus (657 m) is similar to Jebel Arashkol in that it comprises a group of isolated hills that rise abruptly from a pediment surface. In contrast, however, the slopes of the four hill masses of Jebel Tuyus are composed in large part of rectilinear, debris-mantled slopes with an angle of up to 20°. The entrenched alluvial fans that radiate generally west, east, and south-east from the individual hill masses on to the pediment surface can be distinguished from the latter by their darker colour.
Jebel Er Rereid is different. It comprises two low, adjacent knolls rising only a few tens of metros above the surrounding pediment. No bedrock is exposed on these knolls, which are entirely covered with rock fragments.
In the north east of the research area, patches of pediment also seem to be present. In the extreme north-east corner the pediment is developed on Nubian Sandstone rather than Basement Complex.
This section deals with the bedforms developed on the qoz in the research area. The areas of qoz stand out as slightly undulating but distinct red or yellow-brown low rises, above the flat monotonous topography of the grayish clay plain. Where the qoz is not cultivated it is frequently characterized by a scanty vegetation cover of marakh shrub (Leptadenia pyrotechnica) and grasses. The research area was categorized by Warren (1970) as "high" qoz (with welldeveloped dune formations), but there are large areas where dune forms are extremely denuded or even absent. These areas might be more correctly designated as "low" qoz. There are two main types of dune form in the research area, longitudinal and transverse, both of which are clearly related to winds from the north which today blow predominantly during the winter months.
Following Warren's (1970) nomenclature, the qoz bedforms can be classified according to their dune type.
These are short sinuous ridges and vary in trend from westsouth-west/east-north -east to west- north -west/east-south east. They are generally less than 10 m in height. The crest lengths of all the identifiable transverse dunes in the area were measured and are represented as percentage frequency histograms in figure 1.5. This shows that the transverse dunes in areas A and B are much smaller than in the rest of the research area. Here, besides a goodly proportion of small transverse dunes, some may exceed 3,000 m in crest length.
Apart from those in area A on figure 1.4, the distribution of transverse dunes in the research area suggests a broad association with the drainage lines. The dune forms are best represented where they occur in the broad depressions of the western part of the area with drainage lines directed eastwards towards Esh Shugeiq. For many of these depressions, individual khor courses do not apparently exist, but the pattern of transverse dunes suggests links with the welldefined khors such as Wadi El Khidir Hamid, Khor If aetus and Wadi Shaqq er Ramad. Other observers investigating the qoz of central Sudan have noted the association of dunes with khor deposits and basins (Edmonds, 1942; Alam el Din, 1968). Hack (1941), considering dunes in Navajo Country, USA, argued that transverse dunes occurred where the supply of sand was large enough to destroy all or nearly all the vegetation. Such conditions are likely to be satisfied along the wadi courses of the research area. Warren (1970) carried out particle size analyses on different types of sand dunes in central Kordofan, and concluded that transverse dune fields are formed initially from poorly sorted sands and that further aeolian winnowing and transportation of the remaining fine well-sorted sands leads to the formation of longitudinal dunes.
These can be distinguished in the research area by their virtual meridional alignment and generally greater length and wavelength. Whereas the transverse dunes appear to be associated with drainage lines, the longitudinal dunes tend to be located on the margins of the qoz and on the clay plains. The longitudinal dunes can be distinguished into three orders of magnitude.
The first-order, and largest, longitudinal dunes occur sometimes as isolated and sometimes as connected, irregular, low, sinuous mounds of pale sand on the clay plain. These dunes rise no more than about 10 m above the surrounding clay and have slopes of less than 5 . These dunes have been subject to trimming by White Nile palaeochannels, as suggested by Williams and Adamson (1980). In the research area, partial burying has been observed, but in the form of clay deposited for some distance up the flanks of some dunes where the usual thin surface cover of wind-blown sand has been removed.
The longitudinal dunes north of Jebel Arashkol can be distinguished from those located to the south of it. The latter comprise a series of connected mounds, some of which have a transverse element together with well-defined, narrow, sinuous clayey interdune areas. This transverse element and the narrow interdune areas are not characteristic of the dunes found farther north. Grove and Warren (1968) and Warren (1970) noted this change in the form of the dunes in the Arashkol area. They considered that the northern meridionally aligned dunes were typical "shoreline" dunes to the former "Nile Lake" at 382 m suggested by Berry (1959), and that where this shoreline swings from north/south to become north-west/ south-east south of Arashkol, the dune pattern changes to one of "remnant arms of parabolic dunes into which the shore ridges are often broken" (Warren, 1970, p.164).
This explanation for the changing pattern of these first order longitudinal dunes and their partial submergence beneath alluvial clays laid down 7,000 to 8,000 years ago suggests that these dunes are stable features. It has already been noted that the dunes are paler than the qoz to the west and that there may be a connection between degree of redness and age of sand. The change in colour towards the west in the research area, already discussed, might also suggest that the qoz on the clay plains is younger than the qoz farther west. However, on the northernmost mound at Shatawi, gullying has revealed red-brown, slightly indurated sand beneath the ubiquitous pale surface sand. Similarly, pale, sinuous sand dunes on the east bank of the White Nile around Hashaba are thought to have been derived from local alluvial sands (Williams, 1968a; and Smith, 1965, for analogous situation in Nevada). It seems feasible therefore that these first-order dunes may have a core of older red-brown qoz, and that later, when the White Nile was characterized by a sandy bedload, this alluvial sand was reworked by the wind to cover and modify pre-existing dunes.
Second-order longitudinal dunes occur south-south-west of Sheikh el Hasin. These are revealed on aerial photographs as long, narrow, north/south trending, slightly sinuous dunes. Their alignment is made clear by the darker sediments of the inter-dune areas. South of Jebel Arashkol, they become fainter. Of the longitudinal dunes in the research area, these bear the closest resemblance in plan to self dunes, but their degraded nature suggests fossil forms. Warren (1970) considered that the formation of longitudinal dunes in Kordofan required fine, well-sorted sand, noting that the Umm Ruwaba formation contained large quantities of such fluvially sorted sediment, and that the main field of longitudinal dunes was associated with this formation. It may therefore be significant that these second-order dunes overlie the Umm Ruwaba Series in the research area. Hack (1941), on the other hand, argued that a plentiful supply of sand favoured transverse dunes and meagre quantities favoured longitudinal ones.
The smallest, third-order longitudinal dunes occur as a cluster in association with the small transverse dunes of area A (fig. 1.4). These dunes have an average of 970 m and standard deviation of 640 m. Their wavelength varies from 200 m to 600 m. Individual dunes are orientated less consistently than the second-order dunes, though their general trend is similar. Both these third-order and the second-order dunes, at their windward ends, lead to areas lacking a significant sand source (Basement Complex pediment and clay plain respectively). This would lend support to Hack's (1941) suggestion that longitudinal dunes are associated with a scant sand supply.
Minor Qoz Sand Formations
Lee and Foredunes These are of limited importance. The distinct mound of pale sand immediately to the west of Arashkol village and banked against the south-east-facing slopes of the jebel represents a lee dune. Other small lee dunes occur around Jebel Tuyus. The tongue of sand declining in height northwards from the northernmost limit of this hill mass is an example of a large stable fore dune.
Crescentic Dunes These were found at only one locality, on the qoz adjacent to the clay plain some 5 km north-east of Et Tur'a. Here, on the broad top of the qoz, several small barchan-like dunes about 3 m high were active in January 1981. They were formed of well-sorted sands and contrasted markedly with the coarser sands of the underlying qoz.
Sand Sheets Much of the qoz of the research area presents a gently undulating or almost flat sand sheet, a feature of much of the other qoz of central Sudan (slam el Din, 1968; Warren, 1970). Much of it is characterized by an apparently random pattern of rises and hollows which gradually merge into areas of well-defined transverse or longitudinal dunes. These random patterns suggest degraded dunes from a rather earlier dune-forming period.
The clay plain of the research area can be regarded as flat. From Khanoum to Juba, a distance of some 1,200 km, the clay plains only increase in altitude by some 80 m. Thus to "the observer on the ground probably the most striking feature of the clay plains is their utter monotony" (Barbour, 1961). Almost the only detectable change of relief visible to the ground observer on the clay plain of the research area is the Arashkol palaeochannel (fig. 1.4). All but the keenest eye would regard the rest of the clay plain as featureless. From the air many other features become apparent: in particular an intricate pattern of palaeochannels, some curving across the plain, others sweeping against and clearly responsible for trimming individual qoz mounds. The trends of the more prominent of these channels are shown in fig. 1.4. All these channels represent former courses of the White Nile.
Three main groups may be distinguished. First, there are two main bands of similarly curving channels. One band sweeps around from the present White Nile west of Ed Dueim and turns again towards the river north of Shabasha. This coincides with the Selati-Shabasha Basin. The other band curves around Qoz Umm Sufi in the north and forms the Sufi Basin. These clearly represent major incursions of White Nile waters onto the clay plain during former higher discharges. The Dueim-Shabasha and Qoz Umm Sufi areas would doubtless have become alluvial islands during such times. Second, there are single, well-defined channels that wind between the dunes on the clay plain immediately west of Ed Dueim, including the Arashkol palaeochannel. LANDSAT imagery of the area to the south of 14 00'N suggests that these channels acted as drainage routes for White Nile waters flowing onto the clay plain from the south and becoming constricted by the dunes. Third, there are two broad, roughly oval areas, one south-east of Arashkol, the other between Et Tur'a and the Sufi Basin, with a more varied pattern of curving channels. Since they are partly enclosed by dune mounds, it seems possible that these areas were lakes when the clay plains were flooded.
Hydrology and Palaeohydrology
A fourth element of the landscape is the White Nile itself, which forms the easternmost boundary of the research area. The history of this river and its present-day behaviour are both essential elements to understanding the details of the region's geomorphology and are basic to an understanding of man's use of the area.
A distinguishing and highly significant feature of the river is its very low gradient of 1:101,000 between Malakal (386 m) and Khartoum (378 m), a distance of 809 km (Shukri, 1949). Surprisingly, the river has few meanders, and only two major curves in this distance. For most of its course it is a braiding river with many alluvial islands (Berry, 1961). Under natural conditions the river is influenced chiefly by the flood from the Sobat, the more constant flow of water through the Sudd and the ponding back caused by the 5.5 m rise of the Blue Nile in flood. The Jebel Aulia Dam, built in 1937, has complicated this pattern and now forms a seasonal reservoir extending upstream for 500 km as far as Renk. This reservoir rises to a maximum of 377.8 m. Seasonal khors that enter the river are locally important but do not affect appreciably its discharge.
The White Nile contributes some 27.5 km3 per year of water to the main Nile, compared with 51.0 km3 per year for the Blue Nile. The Blue Nile carries almost twenty times the sediment load of the White Nile but is so highly seasonal that 83 per cent of the minimum monthly flow of the main Nile comes from the White Nile.
In the research area, the flood plain of the White Nile is characterized by alluvial islands with the main and deepest channel near its eastern margin. The palaeochannels discussed in the previous section suggest that the pattern has not always been as it is at present.
Quaternary History of the White Nile
The existence of huge, almost flat clay plains in central and southern Sudan led early workers to postulate the former existence of lakes during wetter, pluvial periods. Arldt (1918) and Lawson (1927) envisaged large lakes (Lawson's Lake Sudd), and Ball (1939) extended his lake from Sabaloka (6th Cataract) in the north to Shambe (near Juba) in the south, a distance of 1,055 km. In more recent times workers dealing with the White Nile have postulated more realistic and smaller lakes. Berry and Whiteman (1968) envisaged the former existence of a small lake up to 382 m impounded by a clay plug deposited by the Blue Nile at its confluence with the White Nile at Khartoum. Berry (1961) considered that the large alluvial islands along the course of the White Nile had once been large mid-channel bars, estimating that discharges had formerly been ten times greater than at present. He also thought that during high flow conditions the White Nile took on the lake-like characteristics of a very gently flowing river.
Tothill (1946) showed that subfossil mollusca occur widely in the upper clay deposits of the Gezira, and it is upon the analysis of the occurrence of such fossils that current interpretations of the late Quaternary history of the White Nile are based. Such sites are indicated on figure 1.6. Near Es Samra, Williams et al. (1973) found shell-bearing sediments in a series of broad, shallow, highly localized depressions. Radio-carbon dating and a comparison of the subfossil mollusca found at these sites with their present-day distribution led them to conclude that between 7,000 and 8,500 years ago, when the White Nile itself was more extensive, small permanent lakes occupied these shallow depressions. This suggests that the climate in this area at this time was perhaps two to three times wetter than now, for Limicolaria flammata shells were found in the upper shellbearing deposits; this land snail now occurs only south of Sennar, which has a mean annual rainfall of 460 mm compared with 150 mm for Khartoum
At Esh Shawal, evaporite deposits of microcrystalline dolomite and calcite have been found. These ranged in date from about 2,500 to >40,000 B.P. and suggest late Pleistocene evaporation of a body of still, saline water along the course of the White Nile (Williams and Adamson,1980). Above these evaporites lie 4 m of progressively more clayrich river sands. At Tagra, fluviatile fine sands and clayey sands were deposited up to 4 m above the normal minimum water level (i.e. before the Jebel Aulia Dam was built). At Esh Shawal, freshwater shells were found at 5 m above the mean minimum and 2.5 m above the mean maximum water level. The base of the top 1 m of dark alluvial clays at this site gave a date of 11,000-11,500 B.P., indicating inundation at this time up to at least 380 m. Near Shabona, archaeological investigations indicate that groups of prehistoric hunters lived on sand dunes overlooking Nile swamps. It has been shown that the Holocene White Nile attained a level of at least 379 m, which is 3 m above a contemporary uncontrolled flood level (Williams and Adamson, 1980; Adamson et al., 1980). Conclusions based on radio-carbon dating of shells indicate that the White Nile levels were high around 12,500-11,400, 8,400-8,100, 7,000,5,500, 3,000-2,700, and 2,000-1,500 B.P. (fig 1.6).
Adamson et al. (1980, 1982) have attempted to draw together the evidence for White Nile river change over the past years. Between 20,000 and 12,500 B.P., cold dry conditions in the headwaters had profound effects on water and sediment supply to the White Nile. During this period, the river is thought to have been seasonal and intermittent with the bedload of sands reworked into aeolian dunes during the winter (Adamson et al., 1980). Around 12,500 B.P. Lake Victoria is thought to have overflowed, leading to major flooding along the White Nile. The Sudd did not then act as an effective regulator of flow, so that vast quantities of water were released (Williams and Adamson, 1973). River levels dropped after about 11,000 B.P. but rose again from about 8,000 to 7,000 B.P., when, as archaeological evidence shows, dunes on the clay plain on the east bank were in existence and were lapped by floodwaters which trimmed the dunes and deposited alluvial clays.
During the last 8,000 years, the White Nile has apparently incised some 2 m in response to downcutting by the Blue and main Niles (Williams and Adamson, 1980). This incision was interrupted by further periods of unusually high flood levels. Radio-carbon dates on Pila wernei(formerly Ampullaria wernei) shells at Guli, Togra, and Shabona substantiate the view that there were high levels at around 5,500, 3,000, and 2,700 B.P. Fish bones and Pi/a shells at the Jebel et Tomat prehistoric site confirm seasonally swampier conditions than today between 2,000 and 1,500 B.P. (fig.1.6).
In the research area, a site with several species of subfossil mollusca was found in the surface layers of clay near Khor el Mutraq,15 km from the present White Nile. The modern village is located some 500 m westwards of a gravel mound that rises about 3 m above the surrounding ground, on which were found scattered potsherds and bivalve shell fragments in large numbers. No bone artefacts were found nor was there any other indication of early human occupancy. Indeed, oral tradition in the village of Khor el Mutraq suggests that the gravel mound may have been occupied until 500 years ago. Immediately eastwards of the gravel mound lies the cracking clay of the Arashkol palaeochannel. It was from the surface layers of this clay that specimens of Pila ovate (Olivier), Pila wernei (Philippi), and Lanistes carinatus (Olivier) were collected (and kindly identified by Dr. D. S. Brown). All these gastropods are truly aquatic, though amphibious to the extent that they can survive for periods of several months in the absence of surface water. They respire by means of gills and lungs and now live on the tall grass plains of the Sudd region, which are flooded for three to four months each year (Tothill, 1948). When the floods subside they burrow into the wet clay until the following wet season. The species are characteristic of seasonally fluctuating ponds, lakes, or rivers with a clay bottom (Williams et al., 1973), and only occur in the parts of the acacia tall grass zone with heavy rainfall amounts similar to those received by Malakal today (840 mm; see Tothill, 1946). Lanistes carinatus and Pila wernei have been reported from the Holocene lake beds west of Jebel Aulia (Williams et al., 1973), from Tagra (Adamson et al., 1974), from other sites along the east bank (Williams, 1968a), from much of the Gezira clay (Tothill, 1946), and from north of Khartoum (Ruxton and Berry, 1978). At other sites only Pila wernei shells have been found and subfossil Pila ovate shells have not been reported. The specimens collected included a virtually complete shell. However, Omer el Badri (1972; reported by Williams and Adamson, 1973) recorded Pila ovate and Lanistes carinatus shells from a terrace 6 m above the present flood level.
To summarize this section on the subfossil mollusca, the species identified are consistent with collections made from several sites along the White Nile clay plains and on the Gezira. They indicate a much enlarged White Nile which must have flooded the clay plain of the research area seasonally. Furthermore, since collection of shells from the present position of the White Nile would involve a day's round trip, it seems probable that the potsherds and shell fragments on the mound near Khor el Mutraq indicate a site of prehistoric occupancy, close to the shores of the flooded White Nile.
Hydrology in the Rasearch Area, West of the White Nile
This can best be summarized in outline by considering the three landscape components: hill masses and associated pediments; qoz sands; and clay plains.
It is only on the pediments adjacent to hill masses that well developed dendritic, ephemeral stream patterns have been established, Typically these khors are flat-floored and have high width-depth ratios and loose sandy beds. The prominent khors that run south-eastwards from Jebel Arashkol provide vital sources of water for human and animal consumption throughout the year, since groundwater lies only a short depth beneath the khor bed. These khors have formed fans of sand where they meet the clay plain. The well-developed alluvial fans at the bases of the hill-mass slopes of Jebel Tuyus have already been briefly described.
For most of the qoz the high infiltration capacity of the sandy deposits means that surface run-off is a rarity. Generally, it is only where there are dune forms to create relief on the qoz that run-off can occur. In such circumstances the interdune areas may become temporarily swampy and as a result of evaporation these "run-on" areas are often more saline than the "run-off" areas on the dunes (Williams, 1968b). One clear exception to this generalization is the khor system which dominates the central part of the qoz of the research area. This comprises two welldefined khor courses that meet in the Esh Shuqeiq area and continue eastwards towards the clay plain as a single khor (Wadi El Khidir Hamid), which is incised into the qoz up to an estimated 20-30 m. The apparent anomaly of this large drainage system compared with the surrounding qoz, where precipitation rapidly infiltrates and virtually no drainage lines exist, is explicable in terms of the large area that the Esh Shuqeiq systems appear to drain. Ground rises gently westwards away from the White Nile beyond the western margin of the research area. Although the actual khor courses in figure 1.4 do not reveal it, nearly all the drainage from the west of the research area and further west joins up with this system. Elsewhere on the qoz in the research area, any drainage is of a purely local and deranged nature. It may be that under present-day climatic conditions such a large khor network could not have been initiated and that this process took place in wetter conditions during the late Quaternary.
On the clay plain, due in part to the almost imperceptible gradients, little run-off occurs. Infiltration is extremely slow, being confined in particularly clayey areas to cracks, and it is estimated that virtually no percolation takes place beyond a depth of about 2 m (Barbour, 1961). The result is that, following rain, sheet flow may occur across the clay plain and it becomes dotted with pools and swampy meres. These eventually evaporate and result in salt concentration.
Implications for Rural Development and Further Research
An understanding of the nature of the sediments, the landforms, and the Quaternary history of an area helps in assessing its potential for rural development. The White Nile area is no exception to this rule. Some of the geomorphological characteristics of this area, which may inhibit or encourage rural change and development, are now examined on the basis of the landscape units already identified.