sediments of Central Sudan; case study from the Mesolithic landscape at the Sixth Nile Cataract

Catena 149 (2017) 273–282

Contents lists available at ScienceDirect

Catena journal homepage: www.elsevier.com/locate/catena

Prehistoric dark soils/sediments of Central Sudan; case study from the Mesolithic landscape at the Sixth Nile Cataract Lenka Lisá a,⁎, Aleš Bajer b, Jan Pacina c, Jon-Paul McCool d, Václav Cílek a, Jan Rohovec a, Šárka Matoušková a, Anna Kallistova a,e, Zdeněk Gottvald b a

Institute of Geology ASCR, v. v. i., Rozvojová 269, Prague 6, Czech Republic Department of Geology and Pedology, Faculty of Forestry and Wood Technology, Mendel University in Brno, Zemědělská 3, 613 00 Brno, Czech Republic c Department of Informatics and Geoinformatics, Faculty of Environment, J. E. Purkyně University in Ústí nad Labem, Králova výšina 7, 400 96 Ústí nad Labem, Czech Republic d University of Cincinnati, Department of Geography, Braunstein Hall, Room 416, Cincinnati, OH 45209, United States e Institute of Geochemistry, Mineralogy and Mineral Resources, Faculty of Science, Charles University, Albertov 2, Prague 2 128 43, Czech Republic b

a r t i c l e

i n f o

Article history: Received 26 November 2015 Received in revised form 23 September 2016 Accepted 26 September 2016 Available online xxxx Keywords: Micromorphology Soil chemistry Saprolite Sahel Climate change

a b s t r a c t The so-called “lake or swampy” dark colored deposits along or to the west of both the White and Main Niles, which were not historically inundated by the Nile as a whole, have been recorded recently in association with Mesolithic occupation. What are the possible formation processes of these deposits and what is their potential for understanding the environmental record in relation to Mesolithic occupation? New insight into this issue might be brought to the forefront by the findings in the Rocky Cities area at the south-western edge of Jebel Sabaloka by the Sixth Nile Cataract in central Sudan. The study deposits were evaluated in terms of sedimentology, micromorphology, chemical composition, grain size and magnetic parameters. The properties detected in the study section correspond to no less than three different phases of development. The lowermost part represents a saprolite horizon of granitic rocks exposed to weathering during the wet period, which resulted in alkaline conditions (1st phase of formation process). The occurrence of shells of Bulinus forskalii retrieved from the uppermost layer suggests that there was an anoxic environment in the past, which may be linked to the conditions of the present-day Sahel and subsequent attraction of this area for occupation during the Mesolithic period. Deposition of the acidic colluvia from the surrounding granitic rocks in this environment resulted in post-depositional processes involving Fe and Mn impregnation, leading to the black coloring (2nd phase of the formation process). The third phase of the formation processes is connected with the development of recent aridisols. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Since 2011, detailed explorations by the Czech Institute of Egyptology in Prague have been conducted in the western part of Jebel Sabaloka (Suková and Cílek, 2012) with a special focus on the remains of intensive human occupation during the Mesolithic (ca. 9000–5000 cal BC) and Neolithic (ca. 5000–3000 cal BC) periods (Suková and Varadzin, 2012). The occurrence of dark brown/olive deposits detected in the close surroundings of Mesolithic sites in the Rocky Cities area in the north/western foothill zone of the mountain suggested the possible presence of an organic-rich environmental archive related to swamps, lakes or paleosols. This instigated a more detailed paleopedological investigation despite the lack of an evident interrelation between the Mesolithic remains and in situ black deposits/soils. What is really known about such types of deposits along the Nile?

⁎ Corresponding author. E-mail address: [email protected] (L. Lisá).

http://dx.doi.org/10.1016/j.catena.2016.09.023 0341-8162/© 2016 Elsevier B.V. All rights reserved.

The dark brown deposits were previously mentioned by a few authors, mainly in the context of so-called clay pans, mudflats, swamps or lakes (Williams et al., 2015; Williams and Jacobsen, 2011) at varying elevations above the modern Nile (or Blue or White Niles). Most of the sites, including Esh Shawal, Tagra, Shabona and El Khiday, are located on the White Nile, approximately 2–6 m above the recent high flood level, with the pre-dam located on sediment from White Nile inundation (Williams et al., 2015). There is one exception documented by Williams et al. (2015) and Williams and Jacobsen (2011). They introduce what they call “shell bearing clay pans” located at Wadi Mansurab, approximately 15 km west of the lower White Nile, south of Khartoum. The site is located 400–420 m above the Alexandria datum (Williams et al., 2015) or approximately 20–40 m above the unregulated presentday Nile flood level. The infill of the pans displays a layer of dark olive deposits. The shells found near the surface were dated to 9.9–7.6 ka, suggesting wetter conditions in this part of Sudan. These suggestions regarding wetter conditions near the Nile in the Early and Middle Holocene are supported by the mapping of the 450 km2 paleolake fed by an overflow channel from the Main Nile (Williams and Jacobsen,

274

L. Lisá et al. / Catena 149 (2017) 273–282

2011; Williams et al., 2015) and by the latest findings at Mograt Island (Dittrich et al., 2015) and the El Ga′ab depression (Yahia, 2012). The black deposits of Sabaloka may represent another potential environmental archive in the form of relict soils and paleosols. Soils in Sudan have largely been neglected in this study because of their lower utility for agricultural development. A unique investigation into the varying degree of pedogenesis and relative denudation was conducted by Buursing (1971) through his excavation of soil profiles on terraces of varying ages along the Nile between Khartoum and Atbara, which showed clear developmental characteristics based on the terrace ages. The youngest soils, those on Buursink's terrace I, lying 3–5 m above the low Nile water level, are clayey and do not show a significant horizonation or development of a stone pavement. Those on terrace II, approximately 11 m above the low Nile level, are sandier with a development of either a B or even Bt horizon. This terrace also begins to show evidence of significant deflation through the relatively shallow or completely absent A horizon and incipient development of a desert pavement, the stones of which lack any varnish. Terrace III is the oldest, at approximately 17 m, and contains soils developed in sandy to gravelly matrixes, the presence of Bt horizons despite a coarse texture, and well to moderately established desert pavements with a high degree of stone varnish. This allows comparative dating of the landscapes based on surficial development but also highlights the loss of information from the upper soil column due to aeolian erosion. Riverine soils in central and southern Sudan at El Bouga, near Atbara and Ed Dueim along the White Nile, were briefly discussed by Greene (1948), who showed a textural variation between the finer textured parent materials derived from the White Nile and the coarser material downstream of the confluences of the Blue Nile and Atbara with the White and Main Niles, respectively. A detailed study of the soils in Gezira was presented by Williams et al. (1982), who demonstrated the critical pedological differences between soils generated from parent material derived

from the White versus Blue Niles. The dominant soil textures are clays with loams in areas of higher elevation, such as along levees or aeolian dunes. Differences emerge in the chemical characteristics, with the soils to the west containing higher salinity from the enriched waters of the White Nile, while those to the east, whose parent material and moisture are derived from the Blue Nile, have far lower salinity levels. Our knowledge of the depositional and pedogenic history of these black deposits is incomplete, although they may be a potential source of environmental information illuminating the occupation strategies of Mesolithic populations. The main aim of this paper is, therefore, to describe and interpret the recently identified and surprisingly extensive relics of dark deposits/soils found in the close surroundings of the Mesolithic occupation sites. The main issues addressed include determining whether this material represents relict soils or paleosols and whether these are organic rich deposits, as well as defining the formation history of these local deposits and how they may be linked with similar localities in association with Mesolithic sites along the Nile. 2. Geological and geographical setting Jebel Sabaloka is situated approximately 80 km north of the confluence of the White and Blue Nile Rivers and the capital of Sudan, Khartoum (Fig. 1). Jebel Sabaloka rises distinctively above the surrounding flat landscape, which is composed of Nubian sandstones, with a relative difference in elevation of approximately 150 m. The origin of the rounded Jebel Sabaloka structure is volcanic (Almond and Farouk, 1993; Whiteman, 1971). The Nile River has cut a deep and narrow valley into these resistant rocks with a narrow floodplain (Berry and Whiteman, 1968; Said, 1993; Lisá et al., 2012). Of the two main centers at Sabaloka, the large Cauldron Complex is composed of a subsided block of basement overlain by up to 2 km of volcanic rock

Fig. 1. The location of the study area in the area of Nile catchment, together with orthophoto map of south eastern edge of Sabaloka and highlighted concentration of Mesolithic findings (yellow circle), the appearance of dark soil (red circle) and location of the study section (black dot). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

L. Lisá et al. / Catena 149 (2017) 273–282

and circumscribed by a polygonal zone of ring-fracturing. The fracture system was intruded by a ring-dyke of porphyritic microgranite after the eruption of the volcanic rocks, and at approximately the same time, a boss of mica granite with associated tin-tungsten mineralization was injected into the subsided block. Nearly all of the volcanic and intrusive rocks of the Cauldron Complex are thoroughly acidic, but a thin group of basaltic lavas lies at the base of the volcanic succession, and a few minor intrusions are of basic and intermediate composition. Lavas dominate the lower part of the volcanic succession, whereas rhyolitic ignimbrites compose most of the upper part. The chemistry of these various rock types reflects, in many respects, their close similarity to the Younger Granite association of western Africa, although the rocks of the Cauldron Complex are somewhat poorer in soda than most analyzed acidic rocks from the Nigerian Younger Granites (Almond and Farouk, 1993). The system of hydric circulation of the study area with its distribution of potential subsurface water reserves can, without a doubt, be considered static in relation to the rate of change of other factors. Therefore, a reconstruction of the present-day situation provides a reasonable analogy of the subsurface hydric capacity during the prehistoric occupation, which might be key in the formation of the habitat, allowing for the maintenance of moister conditions, even in periods of drought. The high permeability of the rocks in the background of the black deposits is due to fracturing associated with surface weathering, which reaches a thickness of 20 m or more (Marcolongo, 1983). The area covered by the relics of dark soils is situated on the west bank of the river, approximately 1–3 km from the modern Nile and 36–25 m above its current low water level (Fig. 1), mainly around the area designated as the Rocky Cities (Suková and Varadzin, 2012). These soils are not developed on terrace sediments, but on weathered mica granites or colluvial derivatives of these rocks. The Mesolitic sites located within the study area represent the remains of intensive human occupation (Suková and Varadzin, 2012). 3. Methods 3.1. Visualization of the landscape and sampling A special chapter of this research focused on the development of the local topographic data. The lack of reference points has made measurements even more difficult. Based on a review of the data provided by the Sudan National Survey Authority in 2011 and contained in the paper titled Current Status of GIS in the Sudan (Ali, 2009), we obtained source materials at a scale of 1: 100,000 and smaller. Such small–scale maps and coarse data are insufficient for local studies, prompting our need for local topographic mapping. Moreover, it was the lack of stabilized points of the Adindan coordinate system (commonly used in Sudan) that forced us to develop our own local coordinate system. This system was first employed at the Fox Hill site (Suková and Varadzin, 2012) and subsequently applied to the entire study area. After the identification of darker soils in relation to those typical in the region, a number of cores and small trenches were excavated to facilitate research and basic mapping. One location with the best preserved section (Fig. 1–high-lighted by a black square) was chosen for a detailed study, and bulk samples (each 5 cm) as well as micromorphological samples (depth of 10 and 30 cm below recent surface) were collected. 3.2. Micromorphology Two samples from the study section were collected using small paper kubiena boxes 3 × 4 cm in size. They were taken at depths of 10 cm and 30 cm. The samples were dried in situ, impregnated with resin and thin-sectioned in the laboratory of the Institute of Geology ASCR in Prague. The processing followed that of the standards (Stoops, 2003; Bullock and Murphy, 1983). Thin sections were studied under

275

binocular and polarizing microscopes using magnifications between 8×–800× (Stoops, 2003). 3.3. Particle size analyses The particle size distribution was determined using a laser granulometer CILAS 1190 LD, which provides a measurement range from 0.04 to 2500 μm. The first measurement was taken after the sample reacted for 10 min with KOH solution to provide sufficient dispersion. The second measurement of the same sample was made after the carbonates were removed after a reaction with 35% concentrated HCl for 10 min. The third dispersion was performed with the aim of removing any organic matter through reaction with H2O2 (Łomotowski et al., 2008). The data in this paper are reported in three fractions: clay (up to 2 μm), silt (2–63 μm), and sand (63–2000 μm) (Wentworth, 1922). 3.4. Chemical proxies Chemical analyses included ICP EOS analyses on an Intrepid DUO spectrometer (ThermoFisher) of the main macroelements detected from two different solutions. The first, identified as HCL leachate (intended for a total dissolution of carbonate and a dissolution of unstable silicates, leaving quartz grains and stable silicates intact), was prepared by the reaction of a sample with 20% aqueous hydrochloric acid, and the second, reflecting the plant available elements, was detected by Mehlich III extraction (Sparks, 1996). The content of soil organic carbon (Corg) was determined by using sulfochromic oxidation, and the absorbance of the green chromium complex was then measured spectrophotometrically using Specol 11 (Walinga et al. 1992, ISO/DIS 14235 Soil Quality 1995). The soil nitrogen content (Nt) was measured according to Dumas (Buckee, 1994). The soil samples were tested for current soil acidity (pH/H2O) and potential soil acidity (pH/KCl) using the potentiometric method with a glass electrode (soil: H2O or 1 M KCl = 1:2.5) (Grant, 1996). X-ray powder diffraction was conducted with a Bruker D8 Discover diffractometer equipped with a germanium primary monochromator providing CuKα1 radiation (λ = 1.54056 Å) to obtain a phase composition of the soil samples (measurement conditions: 3–80° 2Theta; step size–0,017°; 1 s/step). To clarify the question of whether the soil color is caused by the presence of organic compounds or by iron and manganese compounds, two experiments were performed. At first, soil samples were extracted by Tamm solution (mixture of oxalic acid and ammonium oxalate) using the procedure routinely applied to quantify the amount of amorphous iron and manganese (oxo) hydroxides present (Sparks, 1996; Schwertman, 1964). The amount of iron and manganese extracted was quantified by ICP EOS (Agilent 5100 SVDV). To exclude the presence of organic compounds as the coloring agent, the standard procedure for extraction of humic compounds, which is based on extraction of soil samples by diluted aqueous NaOH (Sparks, 1996), was used. The amount of organic carbon in the extracts was quantified on a Shimadzu automated DOC and analyzed and expressed in mg C/kg solid samples. The extracts obtained were additionally also studied by UV VIS spectroscopy. 3.5. Magnetic proxies Magnetic susceptibility was measured using an Agico MFK1-FA Kappabridge at two different operating frequencies, f1 = 976 Hz and f3 = 15,616 Hz, and an amplitude of the AC field of 200 A/m (peak value). Readings of the unconsolidated samples were taken in plastic bags. The measured susceptibility values were normalized by the mass of each sample and expressed as mass susceptibility [m3/kg]. Frequency dependent magnetic susceptibility, kFD, is characterized by the following commonly accepted parameter (Dearing et al., 1996): kFD = 100 × (kf1 − kf3)/kf1 [%], where kf1 and kf3 are the susceptibilities at frequency f1 (976 Hz) and frequency f3 (15,616 Hz), respectively.

276

L. Lisá et al. / Catena 149 (2017) 273–282

4. Results 4.1. Topography The evaluation of the relative elevation is important for understanding the formation processes connected with Nile floods. The modern alluvial plain close to the Nile is approximately 374 m above sea level. The Nile water surface is variable, but a typical low-flow level is approximately 6 m below the alluvial plain. The area around the Rocky Cities is approximately 402 m above sea level, with the bottom of wadis running through the Rocky Cities Area (Fig. 1) reaching 393 m above sea level. This suggests that the elevation difference between the current Nile and the locations at which the dark soils are exposed varies between 25 and 36 m. 4.2. Sedimentary and micromorphological records The extent of the dark soils mapped within the study area together with the location of the studied sections is presented in Fig. 1. The soil depth varied across the landscape, with the deepest sections measured at approx. 1.5 m. At some locations, the black color was even visible at the surface and recognizable in satellite images. These soils were mostly covered by unsorted colluvial deposits redeposited from the slopes of the Sabaloka Hills. These accumulation events eroded the black soil and also served as a partial protection from further erosion or pedogenic overprinting by the current arid climate. The segment studied in this paper was covered by a few cm of a desert pebble pavement. Macroscopically, the 60 cm sedimentary record of the blackish material may appear to be quite homogenous (Fig. 2), but significant differences are evident from the thin section studies (Table 1; Fig. 3). The uppermost part (first 10 cm) of the section has a granular structure and was described as a sandy silt of light gray color with occasional signs of root bioturbation. Some of the granulae are composed of salts with a white color. The transition downwards is abrupt. The horizon below is approximately 25 cm thick and silty, with the occasional presence of very

coarse sand to a gravel fraction of weathered granitic rock. The soil becomes solid and impregnated by a clay fraction of white color. Generally, the color of the horizon is gray and the transition downwards is diffuse. The lowermost horizon distinguished within the study section is approximately 25 cm thick. The grain size is silty clay with an abundance of very coarse sand to a gravel fraction of weathered granitic rock. The horizon is massive due to white clay impregnation, with an overall gray color. Excavation was terminated due to the material density, which graded into weathered granitic saprolite. The micromorphological samples from depths of 10 and 30 cm are generally very similar, but the pedofeatures of dusty clay coatings, neoformed nodules and development of the ped structures are better expressed in the upper part of the section. The provenance of the upper part of the section differs obviously by the presence of the Nile's alluvium material. The lower part of the sample is less silty and has a visible decrease in pores and channels (Fig. 2). A detailed micromorphological description is provided in Table 1, and examples of the microstructure and pedofeatures are shown in Fig. 3. 4.3. Grain size distribution The trend of the clay increasing with depth is visible in the upper parts of the section (measured in KOH dispersion as well as in HCl dispersion) (Fig. 3). The lack of this trend was recorded in the total dispersion when the organic fraction was removed, and therefore, the presence of the clay fraction is probably related to the presence of a very fine-grained decomposed organic matter. The variations of the silt and sand fractions in the HCl and total dispersions reflect the removed carbonates, but no trend was detected, which demonstrates that the carbonates are redistributed throughout the section. 4.4. Chemical analyses and magnetic proxies The studied material has a typically high content of amorphous Feoxohydroxides, which together with Mn, gives the soil its black color

Fig. 2. The view of study section together with positions of micromorphological samples as well as colleagues JP McCool and Ales Bajer digging the section. In the right lower corner is photo of Bulinus forskalii molusca findings.

L. Lisá et al. / Catena 149 (2017) 273–282

277

Table 1 Micromorphologial description of two samples taken from the study section. Groundmass

e

e

No

Textural pedofeatures

Nodules

Excrements

Passage featuresd

No

No

Fe/Mn smallc bige

No

Big sparitic

No

Big sparitic

e

d

Fe/Mn smalla

e

CaCOc

Quartz and feldspar of mm sizeb; angular and subangular quartz and feldspar of 0.1 mm sizeb; opaque minerals of silt sizea; opaque minerals of sand sized

Small micritic CaCO3f

c

CaCOa

d

Small micritic CaCO3a

d

CaCO3c

No

Infilling

Quasicoating

Hypocoating

Coating

Charred organic matter

Plant residues

Punctuation

organic matter

Amorphous

of organic matter

composition, anorganic residues

d

CaCO3a

Dark gray

Porphyric

(50 µm): 10:90

(500 µm) = 10:90

Moderately sorted

Cracks (10 %) and vughs (30-40%)

Mineral

c

SB

30 - 35

groundmass (XPL)

groundmass (PPL)

b-fabric of

c

Crystalline

Dark gray

Angular quartz and feldspar of mm sizeb; angular and subangular quartz and feldspar of 0.1 mm sizeb; opaque minerals of silt sizee; opaque minerals of sand sizee; rounded rutile, zircone, quartz, feldspared

Crystalline

Nature of

C/F ratio related distribution Porphyric

(10 µm) = 60:40

(50 µm) = 40:60

(500 µm) = 5:95

Unsorted

10–15

Pores–cracks (30%) and vughs (50-60 %)

SB

Fis

Pedofeatures

Organic components

C/F ratio

Texture class and sorting

Porosity

Microstructure

Sample depth (cm)

Mineral components

SB–subangular blocky; Fis–fissure; aMany (5-10%). bVery abundant (> 20 %) of visible area. cOccasional (2-5 %). dPresent in trace amounts. eRare (< 2 %). fAbundant (10-20%).

(Fig. 4). There is a considerable change at 15 cm below the surface in the elemental distribution within the observed section. The content of most plant available elements decreases with depth in the upper 15 cm of the study section (Fig. 4). A strong correlation is visible mainly in the case of Fe, P and K. The same trend was detected in terms of the behavior of several elements measured in the total dissolution sample, such as Mn, Na, and Si. The decrease of plant available elements is positively correlated with Cox and N and negatively correlated with pH and the C/N ratio. The high correlation between Mg, Ca and Sr changes only slightly in the uppermost part of the study section along with the enhancement of the magnetic signal. The pedogenic processes or variation in the sediment provenance detected within the lower part of the section are identified by slight increases in the measured elements, both in the total dissolution and Mehlich III extract. Experiments conducted to qualify the agent coloring the sedimentary body show that brown extracts with a violet tint were obtained from Tamm solution. The violet tint shaded in a span of several hours (typical behavior of manganese oxalate complexes) resulted in the formation of yellowish brown solutions (Table 2). The amount of extractable iron through the profile ranged between 274 and 388 mg Fe/kg of soil, and the amount of extractable manganese ranged between 217 and 264 mg Mn/kg of sediment. The amount of manganese, comparable to the amount of iron, points towards the importance of Mn oxohydroxides as a coloring component of all samples of the studied soil profile. Practically independent of the position of the sample in the profile, the amount of organic carbon was very low (ranging from 62 to 88 mg/kg). Additionally, in absolute value, the extractable Fe and Mn contents were 3× higher than that of carbon. The results of UV VIS spectroscopy (Table 2) show that the

absorbance of extracts is very low at both 465 nm and 665 nm; thus, the amount of organic compound extracted is practically negligible. In Fig. 6, the oxalate extracts (A) and dil. NaOH extracts (B) can be visually compared. Instructively, the amount of iron/manganese extracted is visible by the naked eye, whereas the dil. NaOH extracts targeting the organic compounds are colorless. 4.5. Mineralogy The dominant phase of all samples is quartz. The minority phases are represented by clay minerals (smectite-chlorite group; kaolinite), albite and K-feldspars, minerals from the monoclinic amphibole group, and calcite (Fig. 5). Higher contents of smectite-chlorite minerals were detected in the uppermost parts of the section, likely reflecting the differences in study material provenance. The crystalline Mn oxohydroxide content recorded by XRD is restricted only to below 15 cm. The highest crystalline Mn oxohydroxide content (1.7%) was recorded in the central part of the section and decreased with depth. The appearance of crystalline Mn oxohydroxide can be used to infer the extended stabilization of the soil. Crystalline forms of Na compounds were not detected in the study material, which excludes the presence of solid macrocrystalline NaCl salts. The leachable sodium and chloride contents are highest within the uppermost 15 cm of the study section, and these elements are positively correlated. From a stoichiometric point of view, the amount of sodium is double that of chloride. This means that Na has to also be available in other types of salts, probably CaCO3 × Na2CO3. The development of these white salts is strongly related to a more humid environment than the modern arid landscape of central Sudan.

278

L. Lisá et al. / Catena 149 (2017) 273–282

Fig. 3. The microphotographs of different features observed in study sections (all of them taken in plane polarised light); A–Fe/Mn nodule attached to the micritic accumulation (upper sample); B–unsorted material and subangular microstructure of the sample taken from the depth of 10–15 cm (upper sample); C–highly weathered mineral with Fe staining on the surface (lower sample); D–moderately sorted material and fissure microstructure of the sample taken in the depth of 30–35 cm (lower sample); E–partly decomposed fragment of organic matter (upper sample); F–rounded nodule composed of sparitic CaCO3.

5. Discussion The greatest surprise from our research was the fact that seemingly organic sediment does not contain substantial organic carbon, but the gray to black color is caused more likely by manganese oxides, even in places where the present morphology indicates the existence of marshes or intermittent lakes. Thus we are faced with at least two interconnected questions–how was the probable organic compound removed and what is the origin of the abundant manganese and ferrous oxides and hydroxides? The occurrence of dark soils in a modern mostly arid location is striking and can be explained only as a likely relict of some former soil reflecting a past humid environment, perhaps such as that in the modern Sahel. We know that during the Holocene (Kuper and Kröpelin, 2006), at approximately 8500 BCE, precipitation amounts increased in North Africa during a more humid phase, with conditions similar to those in the Sahel shifting to the north. With the subsequent return of these humid conditions to the south ca. 5300 BCE, northern and subsequently central Sudan saw progressive aridification. At present, the Sahel's northernmost border is situated approximately 150 km to the south of the study area. The sporadic relicts of former soils scattered across the desert that developed under a more humid environment are referred to by different authors as relict soils, fossil soils, and paleosols (Williams, 2014) and may be considered to be similar to the dark soils observed at the edge of Sabaloka Mountain. The occurrence

of these relict soils was also described in connection with the Mesolithic archaeological record along or to the west of the White and Main Niles and was recently revised by Williams et al. (2015). Beyond just being a peculiarity, does the soil record in the study area possess a sufficient environmental value for the interpretation of the past environment? We know that soils reflect the environmental conditions under which they develop. Not all soils developed under drylands are currently active and in balance with the present bioclimatic environment. Proxies measured within the study section show that the process of formation was more complex, with a minimum of three main formation phases. 5.1. Saprolite formation The origin of the black deposits in the Rocky Cities area is probably closely connected with the upper part of less compact saprolite developed in metabazite background rocks (Zauyah et al., 2010) with signs of pedoplasmation (Stoops and Schaefer, 2010). In arid zones, where weathering is slower than pedoturbation, no pedoplasmation zone is observed, except on paleosaprolites, which formed during periods of wetter climatic conditions (Stoops and Schaefer, 2010). In the case of granitic rocks, the upper part of the saprolite displays fractures, which may break the crystals into smaller angular grains. This seems to be the case of the Sabaloka's dark deposits. Relating these deposits to a saprolite pedoplasmation zone would explain the increased Mn and Fe contents, pore type and lack of some features reflecting pedogenic

L. Lisá et al. / Catena 149 (2017) 273–282 279

Fig. 4. The chemical composition of study material, together with grain size parameters and magnetic proxies.

280

L. Lisá et al. / Catena 149 (2017) 273–282

Table 2 Data obtained by experiments in order to quantify the possible coloring agent; a 1 g of sample extracted with 25 ml 0.1 M NaOH; b extract in 0.1 M NaOH (1 cm polypropylene cuvette, against 0.1 M NaOH). Active manganese and iron oxohydroxides

UV VIS

Depth (cm)

Fe (259.940 nm) mg/kg solid sample

Mn (257.610 nm) mg/kg solid sample

Organics mg C/kg solid sample

A at 665 nm

A at 465 nm

A at 410 nm

0–10 10–20 20–30 30–40 40–50 50–60 60–70

274 260 300 317 354 388 281

256 228 225 264 212 225 217

86.4 81.9 74.1 61.8 88.3 64.0 77.8

0.0028 0.0021 0.0007 0.0011 0.0027 0.0006 0.0018

0.0045 0.0020 0.0041 0.0032 0.0074 0.0042 0.0055

0.0081 0.0048 0.0086 0.0061 0.0144 0.0089 0.0099

processes. Illite, vermiculite and chlorite are usually reported from sapropels originating in a temperate climate (Sequeira Braga et al., 2002), whereas in a Mediterranean climate, feldspar is transformed to kaolinite and illite in densely jointed rocks and to interstratified illite/ smectite in less fractured rocks (Jiménez-Espinosa et al., 2007). These findings are consistent with our results, whereas weathering products typical of more humid subtropical or tropical environments were not reported. Impregnative nodules of iron oxide and chalcedony were reported by Stoops and Dedecker (2006) in Thailand. Similar iron nodules are described in the study section. The presence of crystalline Mn oxohydroxides in the lower part of the section reflects long-term stabilization. The pedogenic processes or variation in sediment provenance detected within the lower part of the section are identified by slight increases in measured elements both in the total dissolution and Mehlich III extract. These increased values are likely associated with the unaltered or slightly weathered parent material (Almond and Farouk, 1993). 5.2. Possible reflections of the Sahel environment The occurrence of white salts within the study section (Fig. 4) is an important climatic indicator. These salts seem to be composed of CaCO3 × Na2CO3, and their origin is tightly connected with a definite humidity of the environment that is not comparable with the modern arid landscape of this part of central Sudan. The natron samples from the Darfur natron basins that are commonly sold in markets in Sudan have calcium carbonate contents up to 40%, with some carbonates composed of aragonite, which may explain the elevated Sr levels. The

sources of the carbonates are as follows: local rocks influenced by post-volcanic changes, i.e., by the solution related to it or from the decomposition of plagioclases contained in the rhyolites and metabasic rocks. Another source might be the Nile alluvium and groundwater (Kerpen et al., 1960; Blokhuis et al., 1968). In all cases, most of the soils in Sudan are rich in carbonates, with soils without carbonates detected only in the areas where precipitation exceeds 750 mm (Buursing, 1971). These wetter periods may correspond to the timespan during which this part of Sudan was influenced by the Sahel environment. A shell of Bulinus forskalii retrieved from one sample suggests the presence of an anoxic environment corresponding with a phase in which the precipitation lasted longer, and small pans may have formed above the non-permeable background in the area. The increasing aridity after 5300 BCE (Kuper and Kröpelin, 2006) changed the environmental conditions, with new formation processes overprinting those of the past. The basic environment of “natron” lakes and their surroundings inhibits the migration of Mn and Fe ions. We therefore propose that the local volcanic and granitoid rock weathering at elevation during warm humid phases under the usual slightly acidic conditions were then mobilized by rain events down to swampy or shallow lacustrine basins. Dark minerals, such as amphiboles and pyroxenes, provided Fe, Mg, Mn, Na, other ions, and plagioclases were the sources of Ca, Na, K, Sr and other elements that helped to form the natron-calcium carbonate (sometimes with chalcedony nodules) environment. Thus, the abrupt change of pH between the slopes and basins resulted in dark Mn and Fe precipitation that is still indicative of the former lake and swampy

Fig. 5. Section of selected XRD patterns showing the overall phase composition of studied samples.

L. Lisá et al. / Catena 149 (2017) 273–282

Fig. 6. Visual comparison of the oxalate extracts (A) and dil. NaOH extracts (B). Instructively, the amount of iron/manganese extracted is visible by naked eye (A), while the dil. NaOH extracts targeting the organic compounds are colorless (B). Numbers 1–7 follow sampling depths 0–10; 10–20; etc.

conditions, while the organic compounds were nearly completely oxidized. The only remaining or formerly organic materials that can be found now are mollusk shell fragments and we also can't exclude a small amount of stable dark organic compounds (also charred organic matter) which are not alkali soluble and belong to so-called non-hydrolyzed residue, therefore difficult to recognize chemically. There might be similar analogues for the small shallow depression environments resulting in dark colored sediments along the Nile. The most macroscopically similar sites seem to be El Khiday and Wadi Mansurab in Central Sudan. In the case of El Khiday, Zerboni (2011) and repeatedly Williams et al. (2015) reported the presence of highly organic black deposits of a former swampy environment. Unfortunately, the presence of organic material in these situations was only investigated by micromorphological observations of decomposed organic material, without further analyses. However, the decomposed organic matter can be misleading for fine grained iron and manganese accumulations. The black colored sediments interpreted as the result of a small water pan environment were also recorded by Williams and Jacobsen (2011), but again without any further analyses concerning the blackening. In both cases, the presence of a water environment was interpreted based on the occurrence of water Mollusca shells. With respect to the original interpretations, it should be noted that the black color of such environments is more likely caused by the increased presence of Mn and Fe rather than organic matter and humic acids, as confirmed in the case of the Sabaloka experiments (Fig. 5, Table 2).

5.3. Modern pedogenic processes The movement of elements dependent on modern pedogenic processes (pedoplasmation in the sense of Stoops and Schaefer, 2010), which mainly reflect the short duration of drying and wetting cycles, is quite visible. These processes influence only the uppermost 15 cm of the surface and are well detected by sediment lithology and micromorphology (Fig. 2) and have an exceedingly low amount of organic matter and high salinity (increased content of Na) and alkalinity (increased Ca content) (Fig. 4). Sodium shows a rapid decrease below

281

the surface, with Calcium showing an approximate 1% increase at 10 cm. This likely represents the modern aeolian ionic contribution in conjunction with the clays, as seen in the decreasing clay percentage with depth within the top 30 cm of the section. The presence of middepth crystalline Mn oxohydroxides and near-surface non-crystalline Mn compounds (Fig. 5) further supports this two-phase climatic interpretation. During wetter periods, Mn was mobilized and moved down through the profile until reaching saturation and precipitating in a crystalline form at approximately 30 cm. The high concentration of non-crystalline amorphous Mn compounds at and near the surface is indicative of the continued availability of the source Mn, but a lack of moisture capability. The high correlation between Mg, Ca and Sr changes only slightly in the uppermost part of the study section. This inhomogeneity reflects the occurrence of recent plants removing available Ca and Mg from the soil to a different extent. Another proxy reflecting either biological activity or a different sediment provenance is the enhanced signal of magnetic susceptibility or frequency dependent magnetic susceptibility (Lisá et al., 2012). Additionally, the pedogenic accumulation of carbonates in the uppermost part of the study section (Courty et al., 1987) reflects the alternation of arid and more humid seasons. Mobilization occurs during periods of precipitation within arid phases when vegetation is sparse (Khormali et al., 2006; Moazallahi and Farpoor, 2009). Such soil properties are quite typical of arid soils (Amit et al., 1996) and reflect modern conditions in the study area. Physical soil processes are stimulated by frequent and pronounced changes in temperature (day and night, before and after rains), whereas chemical processes progress at increased speed under high temperature conditions, unless other factors interfere, such as a water deficiency (Buursing, 1971). 5.4. Human occupation and past climate The occurrence of Mesolithic occupation at most Mesolitic sites in central Sudan seems to be connected with the presence of gray to blackish–colored sediments (Marks and Mohamed, 1991; Caneva, 1983, 1988), which may be derivatives of the black deposits described above. A detailed map of the Mesolithic/Neolithic occupation of the western part of Jebel Sabaloka was published by Suková and Varadzin (2012). It is obvious that the characteristic dark soils in the Sabaloka region are confined to the surroundings of the Rocky Cities, where the remains of occupation only during the Mesolithic period have been detected. Sedimentary material containing the cultural record of the Mesolithic period might have provenance in these dark soils, with its gray color constituting the result of a subsequent influx of carbonates (Varadzinová Suková et al., 2015). Paleopedological mapping has a fundamental importance for the environmental interpretation of the study area. The recent landscape differs significantly from that of the past occupied by Mesolithic communities not only because of different climate but also because of different hydrological conditions. These conditions are strongly influenced by different precipitation rates (movement of the monsoon belt and, as a consequence, the vegetation belts) as well as by the downcutting of the Nile riverbed into the geological background. The elevation difference between the study area and recent Nile equates to meters, so we may presume that the widely distributed Mesolithic occupation in the area of the Rocky Cities, located approximately 3600 m from the Nile, originated in a more suitable humanfriendly environment compared to that available today. Black deposits have been detected in close surroundings of rocky outcrops and formations of the Rocky Cities area, whereas the terraces and platforms that are situated on or within these outcrops and formations that contain the remains of the Mesolithic occupation mainly consist of gray-colored soil material derived from the black deposits. It is probable that these black deposits had already developed before the intensive occupation of this area as Mesolithic communities began. The soil material was mechanically disturbed by humans, whereas leaching constituted the main pedological process that occurred in the already developed soil material.

282

L. Lisá et al. / Catena 149 (2017) 273–282

6. Conclusions The dark deposits recorded at the southwestern edge of the Sabaloka Mountains by the Sixth Nile Cataract are described as a relict of pedoplasmated saprolite reflecting past environments. Pedoplasmation partially occurred during the wet phase of the Holocene when a wetland setting probably corresponding to that of the present-day Sahel existed there. The shells of Bulinus forskalii retrieved from one sample suggest the presence of an anoxic environment. Another record of the past humid phase may be retrieved from the occurrence of white salts, which also suggests the presence of alkaline conditions. The deposition of the acidic colluvia from the surrounding granitic rocks in this environment resulted in post-depositional processes involving Fe and Mn impregnation, resulting in the black coloring of the studied deposits. The uppermost part of the study section contains signs of recent arid pedogenic processes that are mainly related to short-lasting wetting and drying phases as well as the occurrence of aeolian material deflated from the recent Nile alluvium and homogenized within the uppermost tens of centimeters of the study section. Therefore, the Mesolithic occupation likely developed under a more humid environment, mainly based on the geographic proximity and occurrence of interpreted black deposit derivatives within archaeological contexts. Acknowledgement This project was funded by the internal programme of the Institute of Geology ASCR in Prague no. RVO 67985831. The preparation of this paper was supported by the Univerzita Karlova v Praze Scientific development programme No. 14: Archaeology of non-European areas, Subproject: Ancient Egyptian civilisation research: cultural and political adaptations of the North African civilisations in Antiquity (5000 BCE– 1000 CE), and by “PAPAVER–Centre for Human and Plant Studies in Europe and Northern Africa in the postglacial period”, reg. no. CZ.1.07/ 2.3.00/20.0289. We would like to thank for important comments to the manuscript to Lenka Varadzinova, Ladislav Varadzin and to Lucie Juříčková for the determination of Bulinus forskalii shell. References Ali, A.E., 2009. Current status of GIS in the Sudan. Eighteenth United Nations Regional Cartographic Conference for Asia and the Pacific, Bangkok. Almond, D.C., Farouk, A., 1993. Field Guide to the Geology of the Sabaloka Inlier. Khartoum University Press, Central Sudan, Khartoum. Amit, R., Harrison, J.B.J., Enzel, Y., Porat, N., 1996. Soils as a tool for estimating ages of Quaternary fault scarps in a hyperarid environment–the southern Arava valley, the Dead Sea Rift, Israel. Catena 28, 21–45. Berry, L., Whiteman, A.J., 1968. The Nile in the Sudan. Geogr. J. 134 (1), 1–133. Blokhuis, W.A., Pape, T., Slager, S., 1968. Morphology and distribution of pedogenic carbonate in some Vertisols of the Sudan. Geoderma 2, 173–200. Buckee, G.K., 1994. Determination of total nitrogen in barley, malt and beer by Kjeldahl procedures and the Dumas combustion method–collaborative trial. J. Inst. Brew. 100, 57–64. Bullock, P., Murphy, C.P., 1983. Soil micromorphology. – Berkhamsted, AB Academic. Buursing, J., 1971. Soils of Central Sudan. Utrecht, p. 265. Caneva, I., 1983. Pottery using gatherers and hunters at Saggai (Sudan): preconditions for food production. Origins 12, 7–271. Caneva, I., 1988. El Geili, the history of a Middle Nile Environment 7000 BCE–1500 CE. BAR International Series 424. British Archaeological Reports, Oxford. Courty, M.A., Dhir, R.P., Raghavan, H., 1987. Microfabrics of calcium carbonate accumulations in arid soils of western India. In: Fedoroff, N., Bresson, L.M., Courty, M.A. (Eds.), Soil Micromorphology. AFES, Plaisir, pp. 227–234. Dearing, J.A., Hay, K., Baban, S., Huddleston, A.S., Wellington, E.M.H., Loveland, P.J., 1996. Magnetic susceptibility of topsoils: a test of conflicting theories using a national database. Geophys. J. Int. 127, 728–734. Dittrich, A., Gessner, K., Neogi, S., Ehlert, M., Nolde, N., 2015. Holocene stratigraphies and sediments on Mograt Island (Sudan)–the second season of the Late Prehistoric Survey 2014/2015. Der Antike Sudan 26, 123–144. Grant, W.T., 1996. Soil pH and soil acidity. In: Sparks (Ed.), Methods of Soil Analysis; Part 3 - Chemical Methods. SSSA Book, Madison, pp. 475–490 series 5.

Greene, H., 1948. Soils of the Anglo-Egyptian Sudan. In: Tothill, J.D. (Ed.), Agriculture in the Sudan. Oxford University Press, London, pp. 144–175. Jiménez-Espinosa, R., Vázquez, M., Jiménez-Millán, J., 2007. Differential weathering of granitic stocks and landscape effects in a Mediterranean climate, Southern Iberian Massif (Spain). Catena 70, 243–252. Kerpen, W., Gewehr, G., Scharpenseel, H.W., 1960. Zur Kenntnis der ariden Irrigationsböden des Sudan, Teil. II. Mikrosk. Untersuchungen. Pedologie. 10, 303–323. Khormali, F., Abtahi, A., Stoops, G., 2006. Micromorphology of calcitic features in highly calcareous soils of Fars Province, Southern Iran. Geoderma 132, 31–46. Kuper, R., Kröpelin, S., 2006. Climate-controlled Holocene occupation in the Sahara: motor of Africa's evolution. Science 313 (5788), 803–807. Lisá, L., Lisý, P., Chadima, M., Čejchan, P., Bajer, A., Cílek, V., Suková, L., Schnabl, P., 2012. Microfacies description linked to the magnetic and non-magnetic proxy as a promising environmental tool: case study from alluvial deposits of the Nile river. Quat. Int. 266, 25–33. Łomotowski, J., Burszta-Adamiak, E., Kęszycka, M., Jary, Z., 2008. Metody i techniki optyczne w badaniach zawiesin. Monografie Instytutu Badań Systemowych PAN, Seria Badania Systemowe, tom 58, Warszawa. Marcolongo, B., 1983. Late Quaternary Nile and hydrology of the Khartoum- Sabaloka region (Sudan). Origini 7, 39e46. Marks, E.A., Mohamed, A.A., 1991. The Late Prehistory of the Eastern Sahel; the Mesolithic and Neolithic of Shaqadud, Sudan. Southern Methodist University Press, p. 292. Moazallahi, M., Farpoor, M.H., 2009. Soil micromorphology and genesis along a climotoposequence in Kerman Province, Central Iran. Aust. J. Basic Appl. Sci. 3, 4078–4084. Said, R., 1993. The Nile River: Geology, Hydrology and Utilization. Pergamon Press, Tarrytown, New York, p. 320. Schwertman, U., 1964. The differentiation of iorn oxide in soils by a photochemical extraction with acid ammonium oxalate.- Z. Pflanzenernaehr. Dueng. Bodenkund 105, 194–201. Sequeira Braga, M.A., Paquet, H., Begonha, H., 2002. Weathering of granites in a temperate climate (NW Portugal): granitic saprolites and arenization. Catena 49, 41–56. Sparks, D.L., 1996. Methods of soil analysis; part 3 - chemical methods. SSSA Book, Madison series 5. Stoops, G., 2003. Guidelines for Analyses and Descriptions of Soil and Regolith Thin Sections. Science Society of America, Madison WI, p. 184. Stoops, G., Dedecker, D., 2006. Microscopy of undisturbed sediments as a help in planning dredging operations. A case study from Thailand. International Conference ‘Hubs, Harbours and Deltas in Southeast Asia: Multidisciplinary and Intercultural Perspective’. Royal Academy of Overseas Sciences, Brussels, pp. 193–211. Stoops, G., Schaefer, C.E.G.R., 2010. Pedoplasmation: formation of soil material. In: Stoops, G., Marcelino, V., Mees, F. (Eds.), Interpretation of Micromorphological Features of Soils and Regoliths. Elsevier, Amsterdam, pp. 69–79. Suková, L., Cílek, V., 2012. The landscape and archaeology of Jebel Sabaloka and the Sixth Nile Cataract, Sudan. Interdisciplinaria Archaeologica–Nat. Sciences in Arch. III 2, 189–201. Suková, L., Varadzin, L., 2012. Sabaloka project: exploration of Jebel Sabaloka (West Bank), 2009-2012. Sudan & Nubia 16, 118–131. Varadzinová Suková, L., Varadzin, L., Bajer, A., Lisá, L., Pacina, J., Pokorný, P., 2015. Tracing post-depositional processes at Mesolithic occupation sites in Central Sudan: view from the site of Sphinx (SBK·W-60) at Jebel Sabaloka. IANSA 2, 133–150. Walinga, I., Kithome, M., Novozamsky, I., Houba, V.J.G., van der Lee, J.J., 1992. Spectrophotometric determination of organic carbon in soil. Comm. Soil Sci. Plant Anal. 23, 1935–1944. Wentworth, C.K., 1922. A scale of grade and class terms for clastic sediments. J. Geol. 30, 377–392. Whiteman, A.J., 1971. The Geology of the Sudan Republic. Clarendon Press, Oxford. Williams, M.A.J., 2014. Climatic Change in Deserts: Past, Present, Future. Cambridge University Press, Cambridge. Williams, M., Jacobsen, G.E., 2011. A wetter climate in the desert of northern Sudan 9900– 7600 years ago. Sahara 22, 7–14. Williams, M.A.J., Adamson, D.A., Abdulla, H.H., 1982. Landforms and soils of the Gezira: A Quaternary legacy of the Blue and White Nile rivers. In: Williams, M.A.J., Adamson, D.A. (Eds.), A Land Between Two Niles: Quaternary Geology and Biology of the Central Sudan. Balkema, Rotterdam, pp. 111–142. Williams, M.A.J., Usai, D., Salvatori, S., Williams, F.M., Zerboni, A., Maritan, L., Linseele, V., 2015. Late Quaternary environments and prehistoric occupation in the lower White Nile, central Sudan. Quat. Sci. Rev. 130, 72–88. Yahia, F.T., 2012. A Holocene Palaeolake in El-Ga′ab depression, Western Desert, Northern Sudan. Sahara 23, 99–112. Zauyah, S., Schaefer, C.E.G.R., Simas, F.N.B., 2010. Saprolites. In: Stoops, G., Marcelino, V., Mees, F. (Eds.), Interpretation of Micromorphological Features of Soils and Regoliths. Elsevier, Amsterdam, pp. 49–68. Zerboni, A., 2011. Micromorphology reveals in situ Mesolithic living floors and archaeological features in multiphase sites in Central Sudan. Geoarchaeology 26 (3), 365–391.