Aeolian Research 17 (2015) 275–289

Contents lists available at ScienceDirect

Aeolian Research

journal homepage: www.elsevier.com/locate/aeolia

Source of the aeolian sand of Toshka area, southeastern Western Desert, ⇑ M.A. Hamdan a, A.A. Refaat a, , E. Abu Anwar b, N.A. Shallaly a a Geology Department, Faculty of Science, Cairo University, Giza, Egypt b National Research Center, Dokki, Giza, Egypt article info abstract

Article history: Sedimentological, mineralogical and geochemical investigations were carried out in order to identify the Received 1 June 2013 probable source and mode of origin of the aeolian dune sand of the Toshka area at southeastern Western Revised 26 August 2014 Desert, Egypt. A hundred and thirty sand samples were collected from the base, crest and slip face of bar- Accepted 27 August 2014 chan and linear together with windward and interdune area and from lee dunes and sand shadows. Available online 2 October 2014 Grain size analysis of the collected sediments shows that most of the aeolian sand is generally fine-grained, moderately well sorted, fine skewed and leptokurtic. The anchored dunes (lee and sand Keywords: shadows) are nearly similar and are the finest and best sorted of all the dune types of Toshka sands. Dune sand Barchans are coarser while the linear dunes represent the least sorted dune sand. The textural, mineral- Sedimentology Toshka ogical and the geochemical results supported by statistical approach indicate that the Toshka sands were Western Desert mainly derived from late Pleistocene dune sand with a minor contribution from local sources (Pleistocene Egypt alluvial sand and Holocene playa). These sands represent the proposed sources probably derived from the weathering of the Nubian sandstone since the mid-Tertiary by fluvial streams and during earlier humid periods and by aeolian processes during arid periods. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction The study area is located in the southeastern part of the Wes- tern Desert of Egypt between 22° and 24° 300 N and 30° 150 and The aeolian sands occupy significant positions in the geologic 33° E. (Fig. 1). Recently, the area attracted the attention of many history of Egypt in general and in the Quaternary in particular. scientists as well as the governmental authorities, who are inter- The way in which these sands are distributed reflects the paleocli- ested in the development of this remote area for the establishment mate of the Western Desert. Numerous geologic and geomorpho- of new communities, by increasing the cultivated areas in Egypt. logic studies were carried out concerning the dunes and aeolian The specific aim of this paper is to observe and interpret the grain sands of the Western Desert of Egypt. Among these are Embabi size attributes, mineralogical and geochemical differences between (1970), El-Baz et al. (1979), Haynes (1989), Hamdan and Refaat the different dune types in the Toshka area, southeastern Western (1999), Hamdan (2003), Stokes et al. (1998), Besler (1986, 1998, Desert of Egypt. Furthermore, this study provides information on 2000, 2008), El Gammal and Cherif (2006) and Labib and Nashed probable source and mode of origin of these dune sands. (2013). Grain size variations in coastal and desert dune sands have been widely used to infer transport and depositional mechanisms (Bagnold, 1941; Khalaf, 1989; Lancaster, 1995; Wang et al., 2003; 2. Physiography and Geology Kasper-Zubillaga and Carranza-Edwards, 2005). For example, size coarsening of the dune sands may be due to wind deflation of fine The study area covers an area of about 50,000 km2 of the south- grains leaving behind the coarse fraction in the sands (Khalaf, eastern part of the Egyptian Western Desert. Physiographically, the 1989). In addition, mineralogical and geochemical studies of dune study area could be subdivided into three regions: (a) the low relief sands provide new insights into the origin and evolution of aeolian area with elevation less than 150 m above sea level is represented sand bodies (Muhs, 2004). by the Toshka lakes and comprises about 2.1% of the whole studied area; (b) the Toshka plain delineated by the contour line 200 m;

⇑ Corresponding author. Tel.: +20 1065683141, +20 222636241 (H); represents about 26.9% and the pediments of Sinn el Kadab Plateau fax: +20 222636241. with elevation 200–300 m above sea level and represents about E-mail address: [email protected] (A.A. Refaat). 50% of the whole study area; and (c) Sinn el Kadab Plateau http://dx.doi.org/10.1016/j.aeolia.2014.08.005 1875-9637/Ó 2014 Elsevier B.V. All rights reserved. 276 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289

Fig. 1. Map of the Western Desert of Egypt showing the distribution of the major sand dune fields (modified after El-Baz, 1979) and the location of sites that are mentioned in the manuscript. The study area is located at the southeastern part of the Western Desert of Egypt.

represents the high relief area and lies more than 300 m above sea lee dunes and sand shadows accumulated in the leeward sides of level and comprises about 20.3% of the whole studied area. the Nubian sandstone hills (Hamdan et al., 2014; in prep.). Precambrian igneous and metamorphic rocks, early Cretaceous Nubian sandstone, late Cretaceous, Tertiary and Quaternary sediments are exposed in the Toshka area (Fig. 2). 3. Materials and methods Landsat imageries complemented with field studies showed that the dunes of the Toshka area are subdivided into two main The dune morphology might control the grain size parameters categories, free-moving and anchored dunes (Livingstone and (Lancaster 1983; Watson; 1986; Livingstone et al., 1999; Kasper- Warren, 1996).The barchan and linear dunes represent the former Zubillaga and Dickinson, 2001; Wang et al., 2003). The number of while lee dunes and sand shadows represent the anchored type. samples studied varies from one dune type to another. Five sam- The free-moving dunes exist mainly in the Toshka plain (Fig. 3). ples were collected from crest, slip face, windward, base and horns However, few small barchan dunes are also recorded at the top of the 10 different barchan dunes. Four samples were collected of Sinn el Kadab Plateau. The anchored dunes exist at both the from the base, crest, slip face and inter-dune of 10 different linear Toshka plain and pediments of Sinn el Kadab Plateau, where the dunes. One sample was collected from each of 25 lee dunes and 15 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 277

Fig. 2. Geological map of the Toshka area showing the exposed rock units and the main structural elements in the study area (modified after Labib and Nashed, 2013).

Fig. 3. TM Landsat image of Toshka area, showing the numbers and locations of the sampled aeolian dune. Barchan and linear dunes exist mainly in the Toshka plain. Few small barchan dunes are also recorded at the top of Sinn el Kadab Plateau. The anchored dunes exist at both the Toshka plain and the pediments of Sinn el Kadab Plateau. sand shadows. The numbers and locations of sampled dunes are samples were analysed by dry sieving technique (Carver, 1971), shown in Fig. 3 and the sampling sites and coordinates of sampled using sieves at 0.5U intervals. Mean grain size (Mz), sorting (r), dunes are described in detail in Appendix A. skewness (SkI) and kurtosis (KG) were calculated from graphic data A total of 130 samples representing different aeolian sands of following size parameters of Folk and Ward (1957). Summary of the Toshka area were subjected to textural investigations. The the average and range of grain size data of the aeolian dunes sand 278 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289

Table 1 Summary of the average and range of grain size data of the aeolian dunes sand of Toshka area, southeastern part of the Western Desert of Egypt.

Dune type Sample Mz(U) r(U) SKI KG Location Mean S.D. Range Mean S.D. Range Mean S.D. Range Mean S.D. Range Barchan (N = 50) Base 1.08 0.03 1.05–1.12 0.64 0.06 0.53–0.71 0.12 0.13 –0.09–0.38 0.97 0.06 0.88–1.07 Crest 1.46 0.06 1.38–1.55 0.69 0.06 0.61–0.77 0.06 0.01 0.05–0.06 1.01 0.02 0.98–1.05 Midslip 1.30 0.13 1.12–1.48 0.50 0.03 0.46–0.53 0.01 0.06 –0.08–0.09 0.98 0.04 0.90–1.02 Horns 1.44 0.04 1.39–1.50 0.74 0.09 0.60–0.90 0.05 0.00 0.04–0.05 1.01 0.05 0.941.10 Windward 0.83 0.03 0.78–0.88 0.52 0.02 0.50–0.54 0.11 0.05 0.07–0.21 1.12 0.03 1.08–1.16 Lee dune (N = 25) 1.72 0.4 1.28–2.52 0.55 0.23 0.20–1.16 0.17 0.14 –0.04–0.58 1.20 0.38 0.77–2.05 Sand shadow (N = 15) 1.74 0.41 0.92–2.52 0.58 0.23 0.28–1.05 0.11 0.12 –0.04–0.30 1.20 0.36 0.87–1.90 Linear (N = 40) Base 1.58 0.09 1.45–1.70 0.58 0.08 0.47–0.68 0.04 0.01 0.04–0.06 1.06 0.01 1.05–1.07 Crest 1.85 0.13 1.67–2.03 0.52 0.05 0.46–0.59 0.08 0.08 –0.01–0.27 1.34 0.25 0.95–1.70 Midslip 1.63 0.07 1.53–1.73 0.94 0.07 0.84–1.03 0.08 0.01 0.07–0.09 0.92 0.02 0.90–0.96 Interdune 1.42 0.06 1.33–1.50 1.10 0.22 0.78–1.38 0.10 0.05 0.20–0.04 0.96 0.15 0.74–1.17 Toshka sand (N = 130) 1.50 0.36 0.78–2.52 0.65 0.23 0.20–1.38 0.08 0.11 0.20–0.58 1.09 0.25 0.74–2.05

Mz: mean grain size; r:sorting; SKI: skewness; KG: kurtosis; S.D.: standard deviation.

of Toshka area is shown in Table 1. The heavy minerals were sep- The obtained textural, mineralogical and geochemical results of arated from the fine sand fractions of 40 representative samples (8 the Toshka sand are represented as variables and subjected to a of barchans, 12 of linear dunes and 10 samples for each lee and statistical analysis using SPSS programme. Descriptive statistics sand shadows). Separation was conducted using bromoform and are used to investigate the statistical and geochemical distribution heavy fractions were washed with alcohol, dried, weighed and of the examined variables. For each variable, the mean, standard their percentages were calculated. The separated fractions were deviation, maximum, minimum and range values were calculated. examined microscopically and the relative proportions of the var- Pearson’s correlation coefficient (r) was determined for each pair of ious heavy-mineral species were determined (see Table 2). variables which shows the degree to which pairs of variables are X-ray fluorescence (XRF) was used to determine the major and correlated. Correlation matrix between the most common major trace elements composition for 10 samples representing different oxides and trace elements of the aeolian sands of the Toshka area dune types. Representative sediments subsamples are initially is shown in Table 4. In order to test the hypothesis that the means wet sieved to remove silt and clay-sized materials and dry sieved of two or more variables are or are not significantly different One- to remove grains coarser than 0.25U in size. The subsamples grin- way ANOVA procedure using SPSS has been conducted in this ded to powder then mixed with lithium tetraborate to prepare study. Once we have determined that differences exist among fusion discs with a sample to flux ratio of 1:4. Major and trace ele- the means of different dune sands, SPSS provides post hoc tests ments concentration (Table 3) were determined with a Siemens and Dunnett’s T3 pairwise multiple comparisons that can deter- sequential X-ray fluorescence spectrometer, equipped with an mine which means differ. These pairwise multiple comparisons end-window Rh target tube and a 125 lm thin Beryllium window. test the difference between each pair of means and yield a matrix The analytical errors of this method are ±3–5% for majors and ±10% where asterisks indicate significantly different group means at P for traces. level (probability) of 0.05 (i.e. 95% confidence level). This test is

Table 2 Heavy minerals composition (%) of the aeolian dune sands of the Toshka area. The non-opaque heavy minerals recalculated as percentage of the total 100%.

Dune type OP ZR TO RU ST GA EP PX OT S/I G/E Barchan Base 93.2 20.0 30.0 16.4 9.2 12.1 9.2 2.1 1.0 2.72 1.32 Crest 75.0 21.6 20.3 14.0 12.0 12.4 8.2 1.4 0.1 2.53 1.51 Mid-slip 81.3 30.4 23.6 13.6 9.1 14.3 11.1 2.3 2.5 2.24 1.29 Horns 94.0 24.2 26.4 16.4 11.0 13.7 8.9 1.8 0.6 2.68 1.54 Windward 91.0 27.0 24.0 12.7 9.0 15.5 9.0 2.1 0.7 2.33 1.72 Average 86.9 24.6 24.9 14.6 10.1 13.6 9.3 1.9 1.0 2.47 1.47 S.D. 8.36 4.17 3.60 1.69 1.36 1.40 1.08 0.35 0.91 0.21 0.18 Lee dune 95.0 28.7 26.5 6.7 10.7 14.6 9.4 1.8 1.8 2.24 1.55 Sand Shd 94.0 36.5 27.7 5.8 11.1 12.1 11.0 0.8 1.0 2.81 1.10 Linear dune Base 81.7 29.9 25.9 8.3 11.3 12.8 8.0 2.0 1.8 2.61 1.60 Crest 76.3 29.3 23.6 4.5 13.0 12.2 8.0 3.0 4.4 2.08 1.53 Mid-slip 70.8 31.5 23.6 9.0 16.3 12.1 10.4 2.1 2.7 2.35 1.16 Interdune 74.7 30.7 29.9 5.2 7.9 13.5 7.9 5.5 1.4 2.33 1.71 Average 75.9 30.3 25.8 6.8 12.1 12.7 8.6 3.2 2.6 2.34 1.50 S.D. 4.52 0.96 2.97 2.23 3.50 0.65 1.22 1.63 1.33 0.22 0.24 TSA 84.3 28.2 25.6 10.2 11 13.2 9.2 2.3 1.6 2.4 1.46 S.D. 9.31 4.73 2.94 4.51 2.32 1.19 1.18 1.20 1.21 0.23 0.21

OP: opaques; RU: rutile; TO: tourmaline; ZR: zirco n; ST: staurolite; GA: garnet; EP: epidote; PX: pyroxene; OT: others; S/I: ratio of stable to unstable heavy minerals; G/E: ratio of garnet to epidote; Sand Shd: sand shadows; TSA: Toshka sand average; S.D.: standard deviation . Average values are shown in bold face type. M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 279

Table 3 Major oxides (%) and trace element (ppm) composition of some selected samples of the studied aeolian sand dunes of Toshka area, southeastern Western Desert, Egypt.

Oxides/Elements Dune Type Barchan Linear dune Lee dune Sand shadow Toshka Average Sample No. 1 13 40 31 47 51 32 20 29 39

SiO2 95.51 94.76 85.08 94.12 93.89 93.91 93.57 95.11 88.57 96.92 93.14

TiO2 0.41 0.17 0.75 0.29 0.33 0.20 0.78 0.20 0.19 0.11 0.34

Al2O3 1.04 1.16 1.82 2.08 1.94 1.43 2.54 1.22 1.65 1.15 1.60

Fe2O3 1.33 1.03 2.06 1.96 2.80 2.71 1.71 1.80 2.12 0.56 1.81 MnO 0.01 0.01 0.28 0.02 0.02 0.02 0.02 0.02 0.02 0.01 0.04 MgO 0.07 0.07 0.93 0.16 0.09 0.07 0.15 0.08 0.08 0.07 0.18 CaO 0.07 0.07 1.17 0.24 0.07 0.07 0.13 0.10 0.07 0.07 0.21

Na2O 0.10 0.06 1.55 0.19 0.06 0.06 0.07 0.06 0.06 0.06 0.23

K2O 0.08 0.14 3.71 0.28 0.15 0.09 0.37 0.11 0.09 0.13 0.52

P2O5 0.02 0.01 0.14 0.03 0.03 0.02 0.04 0.02 0.03 0.01 0.03

SO3 0.11 0.11 0.09 0.11 0.11 0.11 0.11 0.11 0.10 0.11 0.10 Cl 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 F 0.16 0.16 0.13 0.15 0.15 0.15 0.15 0.15 0.15 0.16 0.15 LOI 0.31 0.32 1.65 0.24 0.39 1.00 0.54 0.11 7.25 0.55 1.23 Sum 99.23 98.10 99.39 99.90 100.05 99.87 100.20 99.11 100.38 99.93 99.62 Ba 2015 1252 530 482 158 81 181 92 93 62 495 Ce <24 <24 285 <24 <24 <24 <24 <24 <24 26 50.3 Co <4 <4 12 <4 5 5 4 5 4 <4 4 Cr 616 536 266 616 616 616 356 616 616 282 514 Cu <10 9 18 9 <9 <9 <9 <9 <9 10 6.5 Ga 44474444 45 48.4 La <10 <10 94 18 11 16 10 <10 20 13 17 Mo 8 7 18 7 27 9 20 24 19 7 18 Ni 20 18 21 33 44 47 14 32 28 7 26 Pb 5 6 92 13 12 8 5 5 5 10 16 Rb <5 6 162 9 7 6 12 <5 <5 <5 34 Sr 17 23 153 43 35 30 38 27 35 22 42 Th 11 10 53 10 8 12 17 <8 9 <8 16 U <4<412<4<4<4<4<4<4<412 V 1213351925281229401723 W <4 9 14 <4 <4 <4 <4 <4 <4 <4 12 Y <4 <4 103 <4 <4 <4 <4 <4 <4 <4 13.9 Zn <13 <13 380 13 14 <13 16 <13 <13 <13 50.1 Zr 519 95 888 333 516 282 888 128 230 103 398.2 Fe/Ni 465 400 686 415 445 403 854 39 530 560 480 Fe/Co 2325 1801 1201 3427 3917 3791 2990 252 3707 980 2439 Fe/Cr 15 13 54 22 32 31 34 2 24 14 24 Ni/Co 5 5 2 8 9 9 4 6 7 2 6

Sum of major oxides is shown in bold face type. appropriate when the variances are unequal. The results of ANOVA and post hoc tests are presented in Tables 5 and 6. The means of Table 4 textural, mineralogical and geochemical variables that fall within Correlation matrix (N = 10) between the most common major oxides and trace elements of the aeolian sands of the Toshka area. the estimated 95% confidence interval (CI) for each variable were plotted for each dune type (Figs. 4 and 6). Thus, we have only used Correlation Variables: SiO2 TiO3 Al2O3 Fe2O3 the variable values in the graphs that lie within this confidence SiO2 1.00 ** interval. The 5% extreme variable values that lie at the tails of TiO2 0.50 1.00 * ** the distribution and do not include the true mean were excluded Al2O3 0.35 0.65 1.00 Fe O 0.40* 0.46* 1.00 from the illustrated diagrams. 2 3 MgO 0.80** 0.66** 0.26 CaO 0.78** 0.63** 0.24 ** ** Na2O 0.79 0.61 ** ** 4. Results and Discussion K2O 0.79 0.65 0.22 Ba 0.42* 0.39* ** * 4.1. Textural characteristics Cr 0.27 0.51 0.20 0.44 Ni 0.24 0.88** Pb 0.77** 0.57** Significant variations in grain size parameters are recorded Rb 0.99** 0.61** from one dune type to another (Table 1). The aeolian sediments Sr 0.64** 0.33 0.24 ** ** of the Toshka area are represented mainly by medium to fine- V 0.72 0.57 Zr 0.51** 0.97** 0.67** 0.31 grained sand (Mean = 1.5U) with an average value varying between 0.78U (windward sample of barchan dunes) and 2.52U (lee dune). * Above 95% significance level. ** These values are comparable with the world average (2.23U) mean Above 99% significance level. grain size of most of the aeolian dune sand (Goudie et al., 1987). Performing one-way ANOVA test for the textural results shows sig- the Kharga Oases (1.2U; Hamdan, 1987) and slightly coarser than nificant differences between means of grain size parameters the barchan dune sands of the Dakhla Oases (2.19U; Sharaky, (P < 0.001) of different dune sands of the Toshka area (Table 5). 1990). The anchored dunes (lee and sand shadows) are similar Among the different dune types of Toshka sands, the barchan and are the finest and best sorted of all the dune types of Toshka dunes are coarser (average 1.2U) than other dune types. The mea- sands (Fig. 4). However, they are only slightly coarser than the sured Mz values of the studied barchan sand are close to those of average values of the lee and sand shadows of the northern 280 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289

Western Desert (2.55U and 2.09U, respectively; Hamdan and ples (1.46U). The same character is also observed in linear dunes, Refaat, 1999). where the crest is characterised by the finest size (1.85U) among As the ANOVA test proved that differences exist among the the other dune sites. This is also recorded in the linear dunes of means of the studied dune sands, post hoc tests and Dunnett’s T3 the Namib Desert (Livingstone, 1987; Lancaster, 1989). pairwise multiple comparisons provided by SPSS were used to The post hoc tests for sorting (r) distinguish linear dunes from determine which of these means differ. The results of these tests other dunes types as a separate group (P = .004; Table 5). The sands are shown in Table 5. Barchan sands are differentiated in a separate of linear dunes shows the least sorted values of the all studied dune group among the other Toshka dune types, where significant dif- types (Table 1), with an average value of 0.78U (moderately sorted) ferences (a less than 0.05) in the mean grain size of Toshka sands and ranging from 0.52U (crest samples) to 1.10U (inter-dune sam- exist. Within a single barchan dune (Table 1), the pattern of mean ples). The mean sorting of the studied aeolian sands is about 0.65U, grain size shows fining up in mean grain size from the base and which is less sorted than the world average value (0.53U; Goudie windward samples (1.08 and 0.83U, respectively) to the crest sam- et al., 1987). The average sorting values of the barchan sands sys- tematically vary from one site to another of the same dune; for example, mid-slip and windward samples are the best sorted sands Table 5 of 0.50U and 0.52U, respectively whereas sands of horns of the bar- Average values, ANOVA test (P-values) and Post Hoc Test P-values (Dunnett’s T3 chan dunes show the lowest sorting value (0.74U). pairwise comparisons) for grain size parameters of different dune types of the Toshka Skewness is significant between lee dunes and both barchan sand. and linear dunes (Table 1). Most of the studied dune sands are pos-

Dune type NMzr SkI KG itively skewed, or fine skewed to nearly symmetric. The mean Average values skewness of all samples is about 0.08 (nearly symmetric). They Barchan 50 1.20 0.62 0.07 1.02 range between 0.03 (nearly symmetric) for linear and 0.17 (fine Lee dune 25 1.72 0.55 0.17 1.20 skewed) for lee dunes samples. However, kurtosis is less significant Sand shadow 15 1.74 0.58 0.11 1.20 in differentiation between the different dune types of Toshka area. Linear dune 40 1.62 0.79 0.03 1.07 Kurtosis measures the degree to which the grain size distribution is ANOVA test (P-values) peaked. Almost all studied samples of the aeolian dune sands are <0.001 <0.001 <0.001 .006 extremely leptokurtic. Source (I) Source (J) Post Hoc Test P-values The interrelation between mean grain size (Mz) and sorting (r) Barchan Lee dune <0.001 .701 .007 .141 is shown in (Fig. 4). The graph clearly separates the anchored dunes Sand shadow .001 .979 .651 .324 (lee and sand shadows) from barchans and linear dunes. Barchans Linear dune <0.001 .004 .103 .605 Lee dune Sand shadow 1.000 1.000 .602 1.000 are coarser and better sorted when compared with the linear dunes. Linear dune .743 .003 <0.001 .550 Sand shadow Linear dune .822 .046 .084 .714 4.2. Heavy mineral characteristics N: number of samples; Mz: mean graphic size (phi units); r: sorting (phi units); SKI: skewness; KG: kurtosis. P-values < 0.05 indicate significant differences among The heavy-mineral assemblages provide valuable information sources (Bold face type). about the aeolian history of sands (processes of accumulation,

Table 6 Average values, ANOVA test (P-values) and Post Hoc Test P-values (Dunnett’s T3 pairwise comparisons) for grain size, heavy mineral and geochemical ratios of the Toshka sand and their proposed sources of aeolian dune sands in the Western Desert of Egypt. Cretaceous Nubian Sandstone (Assaad, 1988; Koeberl, 1997; Knox et al., 2009); Lower Miocene Moghra Sandstone (Philip and Darwish, 1977); Pleistocene dune Sand (Besler, 2008); Pleistocene alluvial sand and Holocene playa (Hamdan, 1987).

Aeolian dune NMzr SkI KG S/I G/E Fe/Cr Co/Ni Average values TS 130 1.5 0.65 0.08 1.09 2.5 1.5 24 6 NS 18 1.0 1.21 0.06 0.52 71.5 3.7 105 0.9 MS 41 2.14 0.65 0.12 0.57 0.73 0.54 170 9.9 PD 20 3.12 1.98 0.52 2.67 1.5 5.1 51 3.7 AS 25 2.25 1.64 0.71 1.8 1.35 8 88 1.9 HP 13 1.82 0.73 0.04 0.95 1.2 23.7 64 3.1 ANOVA test (P-values) <0.001 <0.001 <0.001 <0.001 <0.001 .042 <0.001 .003 Source (I) Source (J) Post Hoc Test P-values TS NS .061 <0.001 .018 <0.001 .818 .999 .272 .001 MS <0.001 1.000 .997 <0.001 <0.001 <0.001 .008 .002 PD .002 .978 1.000 .009 .020 .742 .001 .055 AS <0.001 <0.001 .005 .034 .004 .769 .001 .006 HP <0.001 <0.001 <0.001 .003 .001 .368 .001 .295 NS MS <0.001 <0.001 .024 .086 .308 .515 .583 .042 PD .001 <0.001 .585 <0.001 .317 .995 .461 .013 AS <0.001 .075 .001 <0.001 .315 .988 .690 .028 HP <0.001 .010 <0.001 <0.001 .313 1.000 .384 .077 MS PD .063 .992 .980 <0.001 .002 .001 .042 .048 AS 1.000 <0.001 .010 <0.001 .003 .618 .052 .043 HP .007 <0.001 .001 <0.001 .092 .030 .071 .020 PD AS .035 <0.001 .003 .008 .973 .885 .148 .017 HP .001 <0.001 <0.001 .002 .610 .999 .285 .285 AS HP .024 .828 .972 .282 .996 .928 .118 .101

Mz: mean graphic size (phi units); r: sorting (phi units); SKI: skewness; KG: kurtosis; S/I: stable to unstable; G/E: garnet to epidote heavy mineral ratios; TS: Toshka Sand; NS: Nubian Sandstone; MS: Moghra Sandstone; PD: Pleistocene Dune sand; AS: Alluvial Sand; HP: Holocene Playa. P-values <0.05 indicate significant differences among sources (Bold face type). M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 281

Fig. 4. Mean grain size (Mz) versus sorting (in U units) of the aeolian sand of Toshka area using only the grain size and sorting values that lie within the 95% CI for means. deflation and reactivation) and their sources (Besler, 2008). Forty also rich in both the bedrock Nubian sandstones and the ancient representative samples of aeolian sands of the Toshka area have aeolian deposits of the Toshka area (Besler, 2008). been subjected to a detailed mineralogical investigation for heavy The second abundant mineral group is the stable heavy miner- mineral compositions (Table 2). als, which are represented mainly by zircon, tourmaline and rutile. The microscopic investigations revealed that the heavy mineral Zircon is recorded in all studied samples and is represented by assemblage of the studied aeolian sands is represented by opaques euhedral prismatic colourless crystals with less abundant yellow- and non-opaques (Table 2). The non-opaque heavy mineral assem- ish, greenish grey and pale brown crystals. It varies between 20% blages (Fig. 5) are represented in decreasing order of abundance by (base of barchan dune) and 36.5% (sand shadows) with an average stable minerals (zircon, tourmaline and rutile), metamorphic min- value of 28.2% of all the non-opaque heavy minerals. Zircon con- erals (garnet and staurolite), and unstable minerals (pyroxene, epi- tent is nearly similar to that of the aeolian sands of the Gilf-Uwinat dote and others). Opaque minerals are the most abundant mineral area (Hamdan, 2003). Tourmaline is recorded in all studied sam- group of the heavy minerals (about 87%) and are represented ples as well rounded pleochroic brown varieties. It varies between mainly by iron oxide minerals (Table 2). They vary between 20.3% and 30% (crest and base of barchan dune, respectively) with 70.8% (slip-face of linear dunes) and 95% (lee dune sand). This aver- an average value of 25.6% of the total non-opaque heavy mineral age value is higher than that of the proposed sources of sand in the content. Rutile occurs as deep red and reddish brown prismatic idi- Western Desert of Egypt; the Moghra Formation, Pleistocene and omorphic grains with rounded edges. It varies between 4.5% (crest Holocene fluvial and lacustrine sediments. Opaque minerals are of linear dune) and 16.4% (base and horns of barchan dunes), with

Fig. 5. Heavy mineral distribution of the studied aeolian sand of Toshka area, compared with other sandstones in the Western Desert of Egypt. Opaque minerals are not considered here. Cretaceous Nubian Sandstone (Shukri and Ayouty, 1953); Lower Miocene Moghra Sandstone (Philip and Darwish, 1977); Gilf-Uwinat, Pleistocene alluvial sand and Holocene sand (Hamdan, 2003); Great Sand Sea (Besler, 2008). 282 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 average value of 10.2% (Table 2). The stable minerals usually con- area (4.7%; Hamdan, 2003) and the world value of sandstones stitute the main component of the heavy mineral content in both (2.50%; Turekian and Wedepohl, 1961). The relatively fair positive the Nubian sandstone bedrock and the ancient aeolian sands correlation between Al2O3 and K2O(r = 0.22) in the studied sam- (Hamdan, 2003; Besler, 2008). ples probably indicates that Al2O3 content is related to the K-feld- Metamorphic minerals in the studied Toshka sand are repre- spars, whereas the positive correlation between Al2O3 with Fe2O3 sented mainly by staurolite and garnet with very low percentages and TiO2 may also suggest the existence of detrital biotite grains. of sillimanite. Staurolite is represented by golden yellow and red- Cr and Zr contents are the most abundant trace elements of the dish brown paleochroic varieties. Inclusions are common and studied sands. The Cr content varies between 266 and 616 ppm grains with pitted surfaces are observed. It varies between 7.9% with an average value of 514 ppm. This relatively high enrichment (inter-dune) and 16.3% (slip-face of linear dunes) with an average of Cr possibly related to the opaque minerals. Zr content ranges value of 11% of the total content of non-opaque minerals. Garnet from 95 to 888 ppm; averaging 240 ppm. Zr shows a significant ranges from about 12% (mid-slip of linear dune) to 15.5% (wind- correlation with Ti (r = 0.97) reflecting its detrital origin. Other ward of barchan dune), averaging 13.2%. Metamorphic mineral trace elements of the studied sands as Ni, Sr, and V are lower than grains are also abundant in the Nubian sandstone bedrock, but those recorded in the Gilf-Uwinat area studied by Hamdan (2003), they are less abundant in ancient alluvial deposits. but the Toshka sands are higher in their Ba content. Ni is strongly

The unstable minerals are represented mainly by epidote, positively correlated with Fe2O3 (r = 0.88), which occurs mainly as pyroxene and to a lesser degree by amphiboles. The former are ferruginous materials. Most of the geochemical ratios of the stud- recorded in all studied samples averaging 9.2% and ranges from ied sand as Fe/Ni, Fe/Co and Fe/Cr shown in Table 3, are lower than 7.9% (interdune) to 11.1% (slip-face of barchan dune). Pyroxene is those recorded in the Gilf-Uwinat area (Hamdan, 2003). mainly represented by augite. The interdune area of linear dune is actually enriched in pyroxene, which may reach 5.5% (Table 2). The studied aeolian sand in Toshka area seems to be more 5. Discussion enriched in stable heavy minerals than other localities in the Wes- tern Desert of Egypt. The ratios of stable to unstable heavy miner- 5.1. Source of aeolian sand als of the Toshka area range from 2.1 (crest of linear dune) to 2.81 (sand shadows) with a mean value of 2.4 for all studied samples There are three theories to explain the source of aeolian sand of compared to less than 2 at most other localities in the Western Western Desert of Egypt. The earlier theory of Beadnell (1910), Desert of Egypt. The ratio of garnet to epidote in the Toshka area which considered the sands of the dunes of the Western Desert, (average 1.46) is comparatively less than garnet to epidote ratios must have been derived from the arenaceous formation in north- in most other localities in the Western Desert of Egypt (Table 6). ern Egypt. Ball (1927) agreed with the views of Beadnell (1910) that the dunes originate from the great depressions that stretch 4.3. Geochemical analysis from Siwa to Wadi El- Natrun, where the supply of sands is from the loosely compacted Miocene and Oligocene fluvio-marine beds. The major elements’ values of the studied aeolian sand samples The second theory was introduced by El-Baz and his collaborative are nearly similar (Table 3). However, there are higher values of (El-Baz, 1988; El-Baz and Wolfe, 1982), who mentioned that the trace elements like Ba, Rb, Sr, Y, Co, Ni, Cu and Zn for barchan dune dune sand originated by fluvial erosion of sandstone rocks such sands compared to other dune sands. SiO2 is the main oxide form- as those of the ‘‘Nubian sandstone’’ to the south of the dune fields ing the dune sand samples, which ranges from 85.1% (barchan of the eastern . The third theory was given by Said (1998) dune) to 96.9% (sand shadows), averaging 93.1% of the whole com- who believed that dune fields of the Egyptian Western Desert position. This value seems close to the average world value of started to accumulate during the last glacial period when northern sandstones (Turekian and Wedepohl, 1961) and is relatively higher Africa was arid and the sea-level lower by at least 120 m. The than the average SiO2 values of the dune sand of the southwestern exposed continental shelf which probably extended into the Med- Western Desert (85.6%; Hamdan, 2003) and dunes from north- iterranean Sea at places for more than forty kilometres made a western Mexico (Kasper-Zubillaga et al., 2007a). The interrelation- ready source for all the sand needed to build the huge dunes of ships between elements (bivariate statistics) showed positive the eastern Sahara. correlations between Al vs. Ti (r = 0.65) and Fe (r = 0.46), while In the present study, we have used the textural, mineralogical negative correlations exist between Si vs. Fe, Ti and Al, (r = 0.4, and geochemical published data of different sandstones and allu- 0.5 and 0.35, respectively; Table 4) which probably indicates vial, lacustrine and ancient aeolian sands in the Western Desert different sources of the studied aeolian sands. of Egypt in order to depict the possible source of the studied Tosh-

Fe2O3 content ranges from 0.56% to 2.8% with an average value ka sand. Among these are: Shukri and Ayouty (1953), Philip and of 1.8%, which records the second abundance in the studied sand Assaad (1969), Assaad (1988), Koeberl (1997) and Knox et al. samples (Table 3). It is slightly higher than the world value of sand- (2009) that dealt with the proposed pre-Cretaceous Nubian sand- stones (1.38%; Turekian and Wedepohl, 1961). Iron oxide exists in stone source; Philip and Darwish (1977) for the early Miocene different forms; a thin hematite coating on the quartz grains as Moghra sandstone; Besler (2008) for the late Pleistocene dune multiple layers of clay stained by iron oxides (Besler, 1998). It also and Hamdan (1987) for the proposed Pleistocene alluvial sand fills the cavities on the eroded quartz and feldspar grains as and Holocene playa sand sources. This data together with the reported in quartz and feldspar grains from beach and dune sands obtained results were processed statistically and comparisons of New Zealand (Kasper-Zubillaga et al., 2007b). Iron oxide occurs between different types of sand dunes located in the Western Des- also as detrital opaque grains, which are represented mainly by ert of Egypt were carried out. The results of these statistical analy- magnetite and sometimes by ilmenite. The latter is indicated by ses are shown in Table 6. the presence of considerable amounts of TiO2 in some samples. Aeolian sands in the Western Desert of Egypt differ in their tex- In contrast to the behaviour of silica, Al2O3 exhibits positive cor- tural, mineralogical and geochemical variables (Table 6). Correla- relations with most other major cations. It represents the third tions were established between their textural, mineralogical and abundant oxide of the studied samples and ranges from 1.04% to geochemical parameters. ANOVA and Dunnett’s T3 pairwise com- 2.54% with a mean value of 1.60% (Table 3). The average value parisons for the Toshka sand and their proposed sources of aeolian seems lower than those of the aeolian deposits of the Gilf-Uwinat dune sands in the Western Desert of Egypt yielded significant dif- M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 283

The Nubian sandstone and the Moghra sandstone are consid- ered to be the two principal sources for aeolian dune sands in the Western Desert of Egypt. The Nubian sandstone is a complex of sandstones, deposited from the Carboniferous to Cretaceous that is widely exposed in the southern Western Desert of Egypt. It seems to be the probable ‘‘Mother Source’’ for most of alluvial, lacustrine and aeolian dunes (including the Toshka sand) in this desert. Although most of these sands differ in grain size parameters from the Nubian sandstone, statistical studies have shown no sig- nificant difference between the Nubian sandstone, Pleistocene dune and the studied Toshka sand in their textural, heavy mineral compositions and geochemical data; which may indicate that they are genetically related (Table 6). The early Miocene Moghra sand- stone exposed in northern Egypt represents another possible source for aeolian dune sands in the Western Desert of Egypt. It Fig. 6. The stable/unstable heavy minerals versus garnet/epidote ratios of the is characterised by fine to very fine, well-rounded quartz grains Toshka area (using only the values lie within the 95% CI for means) in comparison and existence of thick shale layers (Hassan et al., 2012). Based on with the proposed sources in the Western Desert of Egypt. The Nubian sandstone the textural and mineralogical criteria (Table 6), the Moghra sand- and Toshka dune samples are more enriched in stable heavy minerals than other dune sands. Filled rhombs: Toshka samples; filled squares: Nubian sandstone; filled stone seems unlikely to be a probable source of Toshka dune sands. triangles: Moghra sandstone; filled circles: Pleistocene dune; open triangle: However, the Moghra sandstone represents the main source of Pleistocene alluvial sand; open circles: Holocene playa. aeolian sand at Qattara-Siwa area (Hamdan and Refaat, 1999). On the other hand, the Pleistocene alluvial sand and Holocene playa are not significantly different from the Toshka sand in gar- ferences in mean grain size. The exception is the Nubian sandstone, net/epidote ratio reflecting possible local sources for the studied which consists of coarse-grained quartoze sand similar in mean Toshka sand. The three more recent sands; the Pleistocene dune; grain size to the Toshka sands. Post hoc tests revealed that the alluvial sand and Holocene playa are not significantly different Nubian sandstone and Toshka sand are not significantly different from each other, which implies that they are mineralogically and in mean grain size, heavy mineral composition and Fe/Cr ratio. geochemically related. Moreover, the Toshka dune sands contain Although, the Pleistocene sand is slightly finer in grain size, post relatively high percentages of pyroxene and epidote minerals hoc tests revealed that the Pleistocene dune and Toshka sand are (Table 2), which are very rich in the Quaternary Nile sediments not significantly different in the other textural parameters, G/E (Shukri, 1950), which suggests a contribution from local sources. heavy mineral and Co/Ni ratios, which support the existence of a The Fe/Cr vs. Ni/Co plot (Fig. 7) shows that the studied aeolian good relationship between them (Table 6). The stable/unstable Toshka sands are placed far from the Moghra sandstone. They heavy minerals versus garnet/epidote diagram using the values are plotted closer to Pleistocene dune, alluvial sand and Holocene that lie within the 95% CI for means, shows that Pleistocene dunes playa samples. These Pleistocene sands represent proposed sources samples are placed closer to Toshka samples, indicating that they probably derived from the weathering of the Nubian sandstone are a possible source for the Toshka sand. The Nubian sandstone since the mid-Tertiary by fluvial streams and lakes during earlier and Toshka dune samples are both more enriched in stable heavy humid periods and by aeolian processes during arid periods. The minerals than other dune sands (Fig. 6). The Moghra sandstone is close relationship between the Toshka sand with its probable significantly different from the Toshka sand in mean grain size, source (the Pleistocne dune) and the Mother Source (the Nubian heavy mineral composition and geochemical ratios (Table 6). The sandstone) was emphasized as shown in Fig. 8. This figure shows samples of Pleistocene dune, alluvial sand and Holocene playa the systematic derivation of these Pleistocene sands from the overlap, which reflects a similarity in their mineralogical composi- Nubian sandstone through alternating cycles of humid and arid tion and perhaps derivation from each other. conditions.

Fig. 7. Fe/Cr versus Ni/Co ratios of the Toshka aeolian sand and other proposed source sands: Nubian sandstone (Koeberl, 1997); Moghra sandstone (Philip and Darwish, 1977); Pleistocene dune (Besler, 2008); Pleistocene alluvial sand and Holocene Playa sand (Hamdan, 2003). 284 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289

5.2. History (Mode of origin) of Toshka sand (2) The importance of the aeolian activity and dune migration began with the increased length and severity of the arid period The obvious textural and mineralogical similarities between the during the Pleistocene. During the early Pleistocene semi-arid cli- studied aeolian dune sands and the Nubian sandstone indicate that mate, wadi-spring activity led to winning and reworking of Radar they are genetically related. Dune sands are mainly multicyclic, River sediments and the formation of topographic inversions left being derived from pre-existing recycled sediments as evidenced many alluvial surfaces armord by lag gravels (Fig. 9b). Inverted by their markedly high textural and mineralogical maturity. The wadis are elongated sinuous and branching gravel ridges which textural, mineralogical and geochemical aspects revealed that the stand out above the desert surface (Giegengack, 1986; Said, Nubian sandstone is the most likely source of the studied Toshka 1990). They are made of alternating poorly bedded gravel beds sands. The derivation of Toshka sands from the Nubian sandstone and thin beds of coarse sand. The gravel is composed mainly of must have been carried out through different cycles of alluvium Nubian sandstone fragments embedded in a pale brown sandy and wind deflation since the mid-Tertiary (Fig. 9). The current matrix. The advent of middle Pleistocene arid conditions led to model implies that climatic cycling was most important and the deflation of the Radar River sediments into a vast flat sandy region climate change alternately disintegrated the bedrock surface locally termed ‘‘Selima Sand Sheet’’ (Fig. 9c). It occupies more than during humid periods that made them susceptible to aeolian 120,000 km2 of the hyper-arid area in southwest Egypt and north- erosion during arid cycles; as follows: west Sudan and is characterised by a featureless surface of lag (1) The reworking of the Nubian sandstone bedrock was granules with dune fields and giant ripples of varying heights resumed in the mid-Tertiary by the Radar Rivers (McCauley and wavelengths (Maxwell and Haynes, 2001). It seems that the et al., 1982). These Radar Rivers were a part of continental wide sand sheet owes its flatness to the wind erosion which was limited Trans-Africa Drainage System (TADS) that originated in the Red by the depth of the water table (Haynes, 1989). The Selima Sand Sea hills (McCauley et al., 1986). More recently, Ghoneim and Sheet is dated from Middle Paleolithic to Neolithic time (Said, El-Baz (2007) considered that the sediments of the Radar Rivers 1990). in the south Western Desert were formed by the Toshka Hydro- (3) During the late Pleistocene hyper-arid conditions, at least System which consists of four main tributaries that would have two generations of aeolian sand accumulation were recorded in collected water from a vast catchment region and drain eastward the south Western Desert. The oldest aeolian sediments of the from the north, west and southwest, starting at highland areas. Toshka area are aeolian sand accumulations dated from 65 to The textural and mineralogical compositions of the sands of Radar 20 ka, as indicated archaeologically by the lack of Upper Paleolithic Rivers are similar to those of the Nubian sandstone bedrock habitation of the Eastern Sahara (Wendorf et al., 1987) and radio- exposed in southwestern part of Egypt (Fig. 9a). metric dating of lacustrine deposits of this area (Szabo et al., 1995).

Fig. 8. Schematic diagram illustrating the probable sources of the aeolian dune sand at Toshka area. The systematic derivation of the Toshka sand from its probable source (the Pleistocne dune) and the Mother Source (the Nubian sandstone) was achieved through alternating cycles of humid and arid conditions. M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 285

The second stage of late Pleistocene aeolian dunes (ca.13,000 BP) is (4) Under humid conditions of the early Holocene (11,000– represented by a longitudinal draa (whaleback dunes) at the south- 7000 BP), the late Pleistocene sand dunes were subjected to a per- ern part of the Great Sand Sea, which was formed under the effects iod of stabilization as indicated by existence of paleosol and vege- of trade winds (Besler, 1998). The genetic relationship between the tation remains (Besler, 1998; Hamdan and Lacurini, 2013). Nubian sandstone and late Pleistocene dune sand could be evi- Moreover, the surfacial and slope sediments of the weathered denced mineralogically; where the content of opaque minerals Nubian sandstone were reworked into playa lakes and alluvial increased from 74% in the Nubian sandstone to 84% in the Pleisto- channels (Fig. 9d). During that time, a branch of the River Nile cene dune sand due to additional weathering either before or after passed through the Toshka basin and deposited channel and flood the draa formation (Besler, 1998). plain sediments (El-Baz, 1998).

Fig. 9. Schematic diagram (not to scale) showing the mode of evolution of Toshka dune sands, previously sourced from the Nubian sandstone through complex history of alluviation and wind deflation during alternating humid and arid climatic conditions since the mid-Tertiary. (a) Mid-Tertiary (humid climate); (b) early Pleistocene (arid climate); (c) middle-late Pleistocene (hyper-arid climate); (d) early Holocene (humid climate); (e) late Holocene (hyper-arid climate). 286 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289

(5) It is now generally agreed that arid to hyper-arid condi- Si vs. Fe, Ti and Al, which probably indicates different sources of tions in the Western Desert of Egypt began about 4500 BP the studied aeolian sands. (Kuper, 2006). This desiccation phase led to the formation of Comparing Toshka sands with other possible sand sources in the modern active dune sand in the whole Western Desert includ- the Western Desert revealed no significant difference between ing the Toshka area (Fig 9e). The major parts of the Toshka dune the Nubian sandstone, Pleistocene dune and the studied Toshka sands were derived mainly from older (late Pleistocene) dunes sand in their textural, heavy mineral compositions and geochemi- where contributions from alluvial and lacustrine sediments local cal data; which may indicate that they are genetically related. sources are significant. These contributions are evidenced by the Pleistocene alluvial sand and Holocene playa sand are not signifi- increase in mean grain size and the nature and distribution of cantly different in the garnet/epidote ratio from the Toshka sand the heavy-mineral assemblage present which shows a marked reflecting other possible local sources for the Toshka sand. Based abundance of opaques and ultrastable heavy minerals with some on the textural and mineralogical criteria, the Moghra sandstone pyroxene, which contributed from the Holocene Nile sediment seems unlikely to be a probable source of Toshka dune sands. terraces in Wadi Toshka. Textural, mineralogical and geochemical proxies indicate that the Toshka aeolian sands were derived mainly from older (late Pleistocene) dune sand to the north of study area. These dune 6. Conclusions sands were previously sourced from the mother source (the Nubian sandstone) through complex history of alluviation and wind defla- The aeolian sand dunes of the Toshka area are represented by tion corresponding to alternating humid and arid climatic condi- two types, free-moving (barchan and linear) and anchored (lee tions since the mid-Tertiary. Contributions from local sources in and sand shadows) dunes. Sedimentological investigations the Toshka area added several textural and mineralogical revealed that the aeolian dune sands of the Toshka area are fine elements. to medium-grained, moderately well sorted, fine skewed and lep- tokurtic. Statistically significant differences are recorded between Acknowledgements means of grain size parameters of dune sands of Toshka area. Barchan dunes are the coarser while the linear dunes represent Field work of this paper was supported by the Ministry of State the less sorted. The anchored dunes are the finest and best sorted For Environmental Affairs. We would like to express our thanks to dune types of Toshka sands. Prof. Bayern (Erlangen University) for facilitating the XRF for the The heavy mineral assemblage of the studied aeolian sands is analyzed samples of Toshka sands. We are also indebted to Prof. characterised by predominance of opaques and stable minerals Kasper-Zubillaga, Prof. M. El Sharkawi and an anonymous reviewer together with subordinate metamorphic minerals and few unsta- for their helpful comments that have improved the final version of ble pyroxene and epidote minerals. The interrelationships between the manuscript. Special thanks are also due to Prof. Ola Hafez; Pro- elements (bivariate statistics) showed positive correlations fessor of English language, Faculty of Arts, Cairo University, who between Al vs. Ti and Fe, while negative correlations exist between edited this manuscript for grammar and spelling.

Appendix A

Sample no. Sampling site Locality Lat. Long. Mz(U) r(U) SKI KG 1 Bb Toshka plain 23° 100 0500N32° 29’ 05’’E 1.05 0.53 0.38 0.96 Bc 1.4 0.75 0.06 1.01 Bms 1.21 0.52 0.06 1 Bh 1.4 0.75 0.05 1 Bw 0.78 0.54 0.17 1.16 2 Lb Toshka plain 23° 18’ 05’’N 32° 34’ 10’’E 1.5 0.6 0.04 1.05 Lc 1.7 0.59 0.1 1.6 Lms 1.6 0.9 0.08 0.92 Id 1.35 1 0.04 0.9 3 Sand Shd Toshka plain 23° 02’ 00’’N 32° 29’ 45’’E 1.6 0.5 0.15 1 4 SandShd Toshka plain 23° 02’ 15’’N 32° 25’ 12’’E 1.5 0.6 0.01 0.9 5 Lee Dune Toshka plain 23° 02’ 30’’N 32° 18’ 20’’E 1.3 0.5 0.25 0.9 6 SandShd Toshka plain 22° 48’ 03’’N 32° 25’ 15’’E 1.4 0.7 0.11 0.97 7 SandShd Toshka plain 22° 58’ 20’’N 32° 18’ 12’’E 1.6 0.5 0.3 1.1 8 SandShd Toshka plain 23° 01’ 00’’N 32° 07’ 25’’E 1.9 1 0.04 1.5 9 SandShd Toshka plain 22° 50’ 15’’N 32° 15’ 00’’E 1.6 0.6 0.21 1.6 10 Lee Toshka plain 22° 46’ 52’’N 32° 22’ 14’’E 1.7 0.4 0.3 0.87 11 Lee Dune Toshka plain 22° 47’ 50’’N 32° 12’ 25’’E 1.4 0.4 0.4 0.88 12 Lee Dune Toshka plain 22° 46’ 55’’N 32° 06’ 32’’E 2.5 0.2 0.15 0.96 13 Bb Toshka plain 22° 38’ 17’’N 32° 15’ 12’’E 1.12 0.6 0.01 0.93 Bc 1.38 0.77 0.05 0.99 Bms 1.2 0.49 0.05 0.9 Bh 1.45 0.68 0.04 1.1 Bw 0.82 0.54 0.08 1.08 (continued on next page) M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 287

Appendix A (continued)

Sample no. Sampling site Locality Lat. Long. Mz(U) r(U) SKI KG 14 SandShd Toshka plain 22° 46’ 30’’N 31° 59’ 03’’E 1.5 0.7 0.3 0.9 15 Bb Toshka plain 22° 45’ 05’’N 31° 57’ 37’’E 1.09 0.65 0.2 1.07 Bc 1.45 0.7 0.06 1.02 Bms 1.33 0.47 0.04 0.92 Bh 1.47 0.65 0.05 1 Bw 0.87 0.53 0.07 1.09 16 Lee Dune Toshka plain 22° 38’ 45’’N 31° 46’ 04’’E 1.8 0.4 0.14 1.34 17 SandShd Toshka plain 22° 38’ 15’’N 31° 50’ 12’’E 1.9 0.4 0.04 0.88 18 Lee Dune Toshka plain 23° 00’ 10’’N 31° 59’ 23’’E 1.7 0.37 0.09 0.83 19 Lee Dune Sinn el Kadab 23° 27’ 25’’N 32° 14’ 05’’E 1.4 0.5 0.04 0.77 20 Lee Dune Sinn el Kadab 23° 36’ 05’’N 32° 16’ 07’’E 1.4 0.2 0.15 0.95 21 Lee Dune Sinn el Kadab 23° 46’ 05’’N 32° 19’ 11’’E 2.4 0.9 0.18 1.6 22 Lee Dune Sinn el Kadab 24° 06’ 32’’N 32° 29’ 53’’E 2.2 0.8 0.2 1.2 23 Lee Dune Sinn el Kadab 24° 10’ 05’’N 32° 13’ 12’’E 1.28 0.5 0.04 1.8 24 Lee Dune Sinn el Kadab 23° 57’ 27’’N 32° 15’ 03’’E 1.3 0.4 0.05 1.4 25 Lee Dune Sinn el Kadab 23° 45’ 10’’N 32° 13’ 57’’E 1.6 0.36 0.2 0.8 26 Lee Dune Sinn el Kadab 23° 31’ 05’’N 32° 06’ 33’’E 1.8 0.6 0.12 1.2 27 Lee Dune Sinn el Kadab 23° 34’ 20’’N 32° 00’ 04’’E 2 0.8 0.16 2.05 28 Lee Dune Sinn el Kadab 23° 28’ 12’’N 32° 05’ 17’’E 1.6 0.4 0.08 1.5 29 SandShd Sinn el Kadab 23° 28’ 57’’N 31° 57’ 44’’E 0.92 1.05 0.15 1.8 30 SandShd Sinn el Kadab 23° 27’ 50’’N 31° 48’ 13’’E 1.92 0.5 0.03 1.2 31 Lb Toshka plain 23° 14’ 36’’N 31° 57’ 54’’E 1.55 0.65 0.04 1.06 Lc 1.77 0.58 0.15 1.3 Lms 1.59 1 0.08 0.9 Id 1.4 1.3 0.09 0.8 32 Lee Dune Toshka plain 23° 15’ 05’’N 31° 48’ 06’’E 1.3 0.49 0.34 1.45 33 Lee Dune Toshka plain 23° 15’ 07’’N 31° 45’ 01’’E 1.5 0.4 0.11 0.8 34 Lee Dune Toshka plain 23° 14’ 42’’N 31° 36’ 09’’E 1.8 0.6 0.25 1.23 35 Lee Dune Toshka plain 23° 07’ 13’’N 31° 31’ 13’’E 2.52 0.8 0.2 1.8 36 Lee Dune Sinn el Kadab 23° 16’ 37’’N 31° 21’ 22’’E 2.3 0.93 0.15 0.99 37 SandShd Sinn el Kadab 23° 19’ 02’’N 31° 16’ 04’’E 2.52 0.28 0.14 1.9 38 Bb Sinn el Kadab 24° 14’ 51’’N 30° 45’ 01’’E 1.12 0.71 0.1 0.95 Bc 1.39 0.76 0.05 0.98 Bms 1.45 0.48 0.02 1.01 Bh 1.39 0.76 0.05 1.1 Bw 0.88 0.5 0.09 1.1 39 SandShd Toshka plain 22° 45’ 37’’N 31° 08’ 01’’E 1.6 0.7 0.17 0.89 40 Bb Toshka plain 22° 28’ 13’’N 31° 30’ 23’’E 1.55 0.63 0.06 1.02 Bc 1.55 0.63 0.06 1.02 Bms 1.12 0.53 0.05 1.02 Bh 1.49 0.78 0.05 1 Bw 0.82 0.53 0.1 1.16 41 Lee Dune Toshka plain 22° 32’ 03’’N 31° 28’ 34’’E 2.1 1.16 0.18 1.4 42 Lee Dune Toshka plain 22° 30’ 01’’N 31° 25’ 11’’E 1.4 0.6 0.03 0.78 43 Lee Dune Toshka plain 22° 29’ 55’’N 31° 13’ 22’’E 1.4 0.5 0.16 0.81 44 SandShd Toshka plain 23° 13’ 04’’N 30° 50’ 01’’E 2.4 0.3 0.21 0.87 45 Lb Sinn el Kadab 23° 59’ 22’’N 30° 59’ 27’’E 1.57 0.68 0.04 1.05 Lc 1.72 0.57 0.03 0.95 Lms 1.68 0.89 0.071 0.91 Id 1.37 1.3 0.1 1.11 46 Lb Toshka plain 23° 31’ 56’’N 30° 43’ 17’’E 1.57 0.65 0.041 1.07 Lc 1.82 0.51 0.01 1.6 Lms 1.64 0.9 0.08 0.92 Id 1.39 1.2 0.2 1.17 47 Lb Toshka plain 23° 12’ 07’’N 30° 41’ 51’’E 1.64 0.56 0.043 1.07 Lc 1.88 0.5 0.03 0.99 Lms 1.7 1.02 0.07 0.91 Id 1.47 0.78 0.15 0.82 48 Lb Toshka plain 22° 53’ 11’’N 30° 50’ 12’’E 1.68 0.53 0.044 1.06 Lc 1.9 0.46 0.01 1.27

(continued on next page) 288 M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289

Appendix A (continued)

Sample no. Sampling site Locality Lat. Long. Mz(U) r(U) SKI KG Lms 1.53 1.03 0.086 0.9 Id 1.42 0.9 0.05 0.74 49 Lee Dune Toshka plain 22° 45’ 04’’N 30° 45’ 01’’E 1.5 0.6 0.58 1.7 50 Lc Toshka plain 22° 39’ 13’’N 30° 49’ 53’’E 1.7 0.5 0.043 1.07 Lms 1.96 0.47 0.1 1.7 Id 1.58 1 0.09 0.94 Lb 1.33 1.38 0.11 0.95 51 Lc Toshka plain 22° 33’’ 04’’N 30° 43’ 06’’E 1.45 0.47 0.06 1.05 Lms 1.67 0.55 0.27 1.23 Id 1.73 0.84 0.09 0.96 Lb 1.5 0.86 0.09 0.97 52 Lc Toshka plain 22° 35’ 06’’N 30° 38’ 24’’E 1.68 0.5 0.04 1.06 Lms 2 0.47 0.02 1.33 Id 1.71 0.86 0.08 0.9 Lb 1.48 0.99 0.06 1.15 53 Bb Toshka plain 22° 29’ 52’’N 30° 27’ 21’’E 1.52 0.65 0.05 1.01 Bc 1.52 0.65 0.05 1.01 Bms 1.28 0.52 0.08 0.99 Bh 1.5 0.84 0.04 1 Bw 0.85 0.52 0.21 1.12 54 Lb Toshka plain 22° 18’ 03’’N 30° 31’ 34’’E 1.49 0.65 0.04 1.06 Lc 2.03 0.46 0.1 1.43 Lms 1.55 1 0.073 0.94 Id 1.49 1.3 0.11 0.99 55 SandShd Toshka plain 22° 15’ 01’’N 30° 45’ 03’’E 2.2 0.3 0.04 1 56 Bb Toshka plain 22° 11’ 14’’N 30° 48’ 22’’E 1.07 0.64 0.12 0.96 Bc 1.49 0.7 0.06 0.99 Bms 1.48 0.46 0.07 0.94 Bh 1.39 0.9 0.05 0.94 Bw 0.86 0.5 0.08 1.12 57 SandShd Toshka plain 22° 27’ 22’’N 30° 45’ 51’’E 1.6 0.5 0.04 1.5 58 Bb Toshka plain 22° 16’ 08’’N 30° 46’ 34’’E 1.06 0.71 0.2 0.99 Bc 1.44 0.61 0.06 1.05 Bms 1.47 0.49 0.07 1 Bh 1.44 0.6 0.05 0.98 Bw 0.78 0.5 0.17 1.1 59 Bb Toshka plain 22° 11’ 17’’N 30° 50’ 08’’E 1.08 0.66 0.1 1.07 Bc 1.53 0.62 0.05 0.99 Bms 1.18 0.51 0.02 1 Bh 1.46 0.7 0.05 1 Bw 0.81 0.52 0.07 1.14 60 Bb Toshka plain 22° 05’ 32’’N 30° 29’ 03’’E 1.11 0.59 0.02 0.89 Bc 1.48 0.69 0.06 1.04 Bms 1.28 0.53 0.09 1.02 Bh 1.41 0.75 0.05 0.98 Bw 0.83 0.52 0.08 1.13

Mz: mean grain size; r:sorting; SKI: skewness; KG: kurtosis;Bb: barchan base; Bc: Barchan crest; Bms: Barchan mid-slip; Bh: Barchan horns; Bw: Barchan windward; Lb: Linear base; Lc: Linear crest; Lms: Linear mid-slip; Id: Interdune; Sandshadow: SandShd.

References Besler, H., 2000. Modern and paleo-modeling in the Great Sand Sea of Egypt (initial results from the Cologne Cooperative Research Project 389). In: Kröpelin, S., Petit-Maire, N. (Eds.), Palaeomonsoon Variations and Global Change during the Assaad, F.A., 1988. Hydrogeological aspects and environmental concerns of the New Late Quaternary. Global and Planetary Change, vol. 26. Elsevier, Amsterdam, pp. Project, Western Desert, Egypt, with special emphasis on the southern 13–24. area. Environ. Geol. Water Sci. 12, 141–161. Besler, H., 2008. The Great Sand Sea in Egypt. Developments in Sedimentology, 59. Bagnold, R.A., 1941. The Physics of the Blown Sand and Desert Dunes. Chapman and Elsevier Publ., Amsterdam, p. 250. hall, London, p. 265. Carver, R.E., 1971. Procedures in Sedimentary Petrology. Wiley Interscience, New Ball, J., 1927. Problems of the Libyan Desert. Geogr. J. 70, 21–28 (105-129, 209-225). York, p. 653. Beadnell, L., 1910. The sand dunes of the Libyan Desert. Geogr. J. 35, 379–395. El-Baz, F., 1979. Aeolian features in the Western Desert of Egypt and some Besler, H., 1986. The Toshka-Canal dune: analysis of development and dynamics. In: applications to Mars. J. Geophys. Res. 84, 8205–8221. Nickling, W.G. (Ed.), Aeolian geomorphology, The Binghamton Symposia in El-Baz, F., 1988. Origin and evolution of the desert. Interdisc. Sci. Rev. 13, 331–347. Geomorphology, International Series, vol. 17. Allen & Unwin, Boston, pp. 113– El-Baz, F., 1998. Aeolian deposits and palaeo-rivers of the eastern Sahara. 130. Significance to archaeology and groundwater exploration. Sahara 10, 55–66. Besler, H., 1998. Periods of aeolian activity and stability in the Great Sand Sea in El-Baz, F., Slezak, M.H., Maxwell, T.A., 1979. Preliminary analysis of color variations Egypt. Zentralblatt für Geologie und Paläontologie, Teil, 1 1997, 23–39. of sand deposits in the Western Desert of Egypt. In: El-Baz, F., Warner, D.M. M.A. Hamdan et al. / Aeolian Research 17 (2015) 275–289 289

(Eds.), 1987, Apollo-Souez Test Project- Earth observations and photography. Kuper, R., 2006. After 5000 BC: The Libyan Desert in transition. CR Palevol 5, 409– NASA, Washington, D.C., USA, pp. 237–262. 419. El-Baz, F., Wolfe, R.W., 1982. Wind patterns in the Western Desert. In: El-Baz, F., Labib, M., Nashed, A., 2013. GIS and geotechnical mapping of expansive soil in Maxwell, T.A. (Eds.), Desert Landforms of Southeast Egypt: A Basis for Toshka region. Ain Shams Eng. J. 4, 423–433. Comparison with Mars. NASA, CR-3611, pp. 119–139. Lancaster, N., 1983. Controls on dune morphology in the Namib Sand Sea. In: El Gammal, E.A., Cherif, O.H., 2006. Use of Remote Sensing for the study of the Brookfield, M.E., Ahlbrandt, T.S. (Eds.), Eolian Sed. Proc.. Elsevier, Amsterdam, hazards of Ghard Abu Muharik Sand Dune Field, Western Desert, Egypt. The 2nd pp. 261–289. International Conf. on Water Resources and Arid Environment. NARS. Cairo, 1– Lancaster, N., 1989. The Namib Sand Sea: Dune Forms, Processes, and Sediments. 20. A.A, Balkema, Rotterdam, p. 200. Embabi, N.S., 1970. Structures of barchan dunes at the Kharga Oases Depression, the Lancaster, N., 1995. Geomorphology of desert dunes. Routledge, London, New York, Western Desert, Egypt (and a comparison with structures of two aeolian p. 290. microforms from Saudi Arabia). Bull. Soc. Geogr. d’ Egypte 43 (44), 53–71. Livingstone, I., 1987. Grain-size variation on a ‘complex’ linear dune in Namib Folk, R.L., Ward, W.L., 1957. Brazos River bar, a study in the significance of grain Desert. In: Frostick, L.E., Reid, I. (Eds.), Desert Sediments, Ancient and Modern, parameters. J. Sed. Petrol. 27, 3–27. Special Publication, Geol. Soc. London, vol. 35. Blackwell, Oxford, pp. 281–291. Ghoneim, E., El-Baz, F., 2007. The application of radar topographic data to mapping Livingstone, I., Warren, A., 1996. Aeolian geomorphology: an introduction. of mega-paleodrainage in the Eastern Sahara. J. Arid Environ. 69, 658–675. Longman, Harlow, Essex, p. 211. Giegengack, R.F., 1986. Pleistocene History of the Nile Valley in Egyptian Nubia.. Livingstone, I., Bullard, J.E., Wiggs, G.F.S., Thomas, D.S.G., 1999. Grain-size variation Yale University, New Haven (Ph.D. Thesis). on dunes in the Southwest Kalahari, Southern Africa. J. Sed. Res. 69, 546–552. Goudie, A., Warren, A., Jones, D., Cooke, R., 1987. The character and possible origins Maxwell, T.A., Haynes, C.V., 2001. Sand sheet dynamics and Quaternary landscape of the aeolian sediments of the Wahiba sand sea. Oman. Geogr. J. 154, 231–256. evolution of Selima Sand Sheet, southern Egypt. Quatern. Sci. Rev. 20, 1623– Hamdan, M.A., 1987. Geomorphology and Quaternary Geology of Umm El Dabadib 1647. area, , Western Desert Egypt (M. Sc. Thesis), Fac. of Science, Cairo McCauley, J.F., Schaber, G.G., Breed, C.S., Grolier, M.J., Haynes, C.V., Issawi, B., Elachi, University. C., Blom, R., 1982. Subsurface valleys and geoarchaeology of the eastern Sahara Hamdan, M.A., 2003. Textural, mineralogical and geochemical characteristics of the revealed by Shuttle radar. Science 218, 1004–1020. aeolian dune sand of the Gilf-Uwinat area, South Western Desert. 5th McCauley, J.F., Breed, C.S., Schaber, G.G., McHugh, W.P., Issawi, B., Haynes, C.V., International conf. of the geology of the Middle East, pp. 171–187. Grolier, M.J., El Kilani, A., 1986. Paleodrainages of the Eastern Sahara – the Radar Hamdan, M.A., Lacurini, G., 2013. Holocene paleoenvironmental, paleoclimatic and Rivers revisited (SIR - A/B implications for a mid-Tertiary Trans-African geoarchaeological significance of the Sheikh El-Obeiyid area (Farafra Oasis, Drainage system). IEEE Transactions on Geoscience and Remote Sensing, GE- Egypt). Quatern. Int. 302, 154–168. 24, pp. 624–648. Hamdan, M.A., Refaat, A.A., 1999. Textural and mineralogical characteristics and Muhs, D.R., 2004. Mineralogical maturity in dune fields of North America, Africa and distributions of aeolian sand deposits at Qattara-Siwa area, northwestern Australia. Geomorphology 59, 247–269. desert, Egypt. Int. Conf. on Geol. of the Arab World, GAW4, 1998, Cairo Univ., Philip, G., Assaad, A.F., 1969. Mechanical analysis, porosity, and permeability Egypt, pp. 684–706. studies on Nubian Sandstone, South of Bein Kharga Oasis. Western Desert. Hamdan, M.A., Refaat, A.A., Abu Anwar, E., Shallaly, N.A., 2014. Aeolian dune sand of Egypt. J. Geol. 13, 37–48. Toshka area, southeastern Western Desert, Egypt, and their environmental Phillip, G., Darwish, M., 1977. Mechanical analysis and mineralogical composition of hazards. Aeolian Research (in prep). the sandy sediments of the Moghra Formation in south of Alaen area, Western Hassan, S., Steel, R., El Barkooky, A., Hamdan, M., Olariu, C., Helper, M., 2012. Stacked Desert. Egypt. Bulletin of Fac. Sci. Cairo University 46, 507–531. Lower Miocene tide-dominated estuary deposits in a transgressive succession, Said, R., 1990. The Geology of Egypt. Balkema, Rotterdam, p. 734. Western Desert. Egypt. Sed. Geol. 282, 241–255. Said, R., 1998. Sand accumulation and groundwater in the eastern Sahara: A Haynes, C.V., 1989. Bagnold‘s Barchan: A 57-Yr record of dune movement in the rebuttal to El-Baz F. (Episodes, Vol. 21, No. 3). Episodes 21, 287–289. Eastern Sahara and implications for dune origin and paleoclimate since Sharaky, A.M., 1990. Geomorphological studies on sand dunes and ridges in some Neolithic times. Quatern. Res. 32, 153–167. African deserts (M.Sc. thesis). Institute of African Research and studies, Cairo Kasper-Zubillaga, J.J., Carranza-Edwards, A., 2005. Grain size discrimination University. between sands of desert and coastal dunes from northwestern Mexico. Shukri, N.M., 1950. The mineralogy of some Nile sediments. Quatern. J. Geol. Soc. Revista Mexicana de Ciencias Geológicas 22 (3), 383–390. London 105, 511–534. Kasper-Zubillaga, J.J., Dickinson, W.W., 2001. Discriminating depositional Shukri, N.M., Ayouty, M.K., 1953. The mineralogy of Nubian Sandstone in Aswan, environments of sands from modern source terranes using modal analysis. Bull. Inst. Desert d, Egypte 3, 65–88. Sed. Geol. 143, 149–167. Stokes, S., Maxwell, T.A., Haynes, C.V., Horrocks, J., 1998. Latest Pleistocene and Kasper-Zubillaga, J.J., Zolezzi-Ruiz, H., Carranza Edwards, A., Girón García, P., Ortiz- Holocene sand-sheet construction in the Selima Sand Sea, Eastern Sahara. In: Zamora, G.V., Palma, M., 2007a. Sedimentological, modal analysis and Alsharhan, A.S., Glennie, K.W., White, G.L., Kendall, C.G. (Eds.), Quaternary geochemical studies of desert and coastal dunes, Altar Desert, NW México. desert and climate changes. Balkema, Rotterdam and Brookfield, pp. 175–183. Earth Surf. Proc. Land. 32, 508–598. Szabo, B.J., Haynes, C.V., Maxwell, T.A., 1995. Ages of Quaternary pluvial episodes Kasper-Zubillaga, J.J., Ortiz-Zamora, G.V., Dickinson, W.W., Urrutia-Fucugauchi, J., determined by uranium-series and radiocarbon dating of lacustrine deposits of Soler-Arrechalde, A.M., 2007b. Textural and compositional controls on modern Eastern Sahara. Paleogeography, Paleoclimatology, Paleoecology 113, 227–242. beach and dunesands, New Zealand. Earth Surf. Proc. Land. 32, 366–389. Turekian, K.H., Wedepohl, K.H., 1961. Distribution of elements in some major units Khalaf, F., 1989. Textural characteristics and genesis of the aeolian sediments in the of the earth’s crust. Bull. Geol. Soc. Amer. 72, 175–192. Kuwaiti desert. Sedimentology 36, 253–271. Wang, X., Dong, Z., Zhang, J., Qu, J., Zhao, A., 2003. Grain size characteristics of dune Knox, R.W.O’B., Stadelmann, R., Harrell, J.A., Heldal, T. and Sourouzian, H., 2009. sands in the central Taklimakan Sand Sea. Sed. Geol. 161, 1–14. Mineral fingerprinting of Egyptian siliceous sandstones and the quarry source Watson, A., 1986. Grain-size variations on a longitudinal dune and a barchan dune. of the Colossi of Memnon. In: Abu-Jaber, N., Bloxam, E.G., Degryse, P., and Sed. Geol. 10, 77–106. Heldal, T. (Eds.), Quarry Scapes: ancient stone quarry landscapes in the Eastern Wendorf, F., Close, A.E., Schild, R., 1987. Recent work on the middle Paleolithic of Mediterranean, vol. 12. Geological Survey of Norway Special Publication, the Eastern Sahara. African Archaeol. Rev. 5, 49–63. pp. 77-85. Koeberl, C., 1997. Libyan Desert Glass: geochemical composition and origin. In: de Michele, V. (Ed.), Silica ’96. Proceedings of Meeting on Libyan Desert Glass and related desert events, Milan, Italy, Pyramids, pp. 121–131.