17 Geol.Mag. 133(1), 1996, pp. 17-31. Copyright© 1996 Cambridge University Press Pan-African volcanism: petrology and geochemistry of the Dokhan Volcanic Suite in the northern Nubian shield ABDEL-FATTAH M. ABDEL-RAHMAN Department of Geology, Concordia University, 7141 Sherbrooke St. West, Montreal, Quebec H4B 1R6, Canada (Received 11 November 1994; accepted 13 July 1995) Abstract - The Late Proterozoic Dokhan volcanic suite (620 Ma) of the northern Nubian shield is the product of Late Pan-African volcanism. The suite covers the entire spectrum from basalt to high-silica rhyolite and occurs as two units: a dark-coloured unit containing basalt-andesite-dacite, and a light-coloured unit encompassing dacite-rhyodacite-rhyolite. The latter unit is made up largely of ash flow tuffs and ignimbrites that are locally interstratified with basalt and andesite lava flows. The suite forms a continuum in composition with a wide range of SiO2 (48-77 wt%), CaO (0.1-8.9 wt%), Sr (81-906 ppm), Zr (85-340 ppm) as well as most other elements, and is moderately enriched in incompatible elements, including rare earth elements (REE). The suite exhibits fractionated, subparallel REE patterns that are similar overall to Andean andesites and ignimbrites. Well-defined major and trace element trends and fractionated REE profiles are consistent with a fractionated basalt to rhyolite calc-alkaline magma series. It is a typical calc-alkaline orogenic complex and exhibits mineralogical-geochemical traits of arc-related volcanism. The suite neither resembles products of extensional nor transitional tectonic regimes as previously thought, but was produced in a subduction-related tectonic environment. The mafic nature of the least-evolved rocks of the suite, along with its relatively low initial ^Sr/^Sr ratio (0.7039) are considered to indicate a mantle source. A mantle-derived basaltic magma fractionated, with amphibole and plagioclase dominating the fractionating assemblage, to produce the more felsic varieties, as suggested by major and trace element fractionation modelling. 1. Introduction Two major volcanic episodes have been recognized in the Precambrian crust of the Nubian shield. Volcanism associated with the first episode produced the 'Shadli metavolcanic' sequences; these are older (c. 950-750 Ma), metamorphosed, largely mafic in composition. This was succeeded by a younger Pan-African volcanic cycle (680-550 Ma) that produced voluminous volcanic assemblages, referred to as the 'Dokhan volcanic rocks' (e.g. El-Ramly, 1972; Hashad, 1980). The term 'Dokhan' describes vari-coloured volcanic rocks with a wide range of (mafic to felsic) compositions. The Dokhan volcanic assemblage, with its relatively large stratigraphic extent, represents a major crustal component in northeastern Egypt (Fig. 1), which occupies the northern part of the Late Proterozoic Nubian shield. The latter is separated by the Red Sea from its previously contiguous counterpart, the Arabian shield. The Arabian-Nubian shield consists mainly of metasedimentary rocks and migmatites, metavolcanic suites, serpentinites and dismembered ophiolite complexes, gabbro-diorite-granodiorite complexes, Dokhan volcanic sequences and younger (680-550 Ma) granitic batholiths (El-Ramly, 1972; Fleck et al. 1976; Gass, 1981; Vail, 1985; Abdel-Rahman & Doig, 1987; Kroner e?a/.1988). The Pan-African plutonic complexes of eastern Egypt have been studied in some detail (Brown, 1980; Greenberg, 1981; Gass, 1981, 1982; Stern & Gottfried, 1986; AbdelRahman & Martin, 1987, 1990a; Abdel-Rahman, 1990). http://journals.cambridge.org Downloaded: 20 Mar 2015 Products of Early Pan-African volcanism represented by the 'Shadli' metavolcanic group have been studied by Stern (1981) and Kroner et al. (1991); Stern subdivided this metavolcanic group into older metavolcanic rocks (ultramafic rocks, gabbros and pillowed basalts), and younger metavolcanic rocks (dominantly andesites and volcanogenic metasediments). However, the Dokhan volcanic sequences were largely neglected. There is limited geochemical data on Dokhan volcanic rocks near Safaga and Mount Dokhan (Fig. 1), but little work has been done on the Dokhan volcanic rocks within the area under study. The tectonic environment of emplacement of the Dokhan volcanic rocks is uncertain. The aim of this study is to use petrologicalgeochemical data to interpret the evolution and tectonic setting of the Dokhan volcanic suite, and evaluate its origin. This will shed light on Late Proterozoic crustal development in the region. The study also provides a classic example of the evolution of calc-alkaline magmas via amphibole and plagioclase fractionation. 2. Regional geology The 'Dokhan' volcanic sequences are widely distributed in northeastern Egypt (Fig. 1). The study area, the Mount El-Kharaza region, represents the largest Dokhan volcanic continuous exposures in eastern Egypt. The area was previously mapped (scale 1:100 000) by Francis (1972) and Ghanem (1972). The absence of detailed mapping combined with lack of isotopic age dating, and the IP address: 137.222.24.34 18 A.-F. \ -1 28°- Phanerozoic cover Other shield rocks ~ Dokhan volcanic rocks M. ABDEL-RAHMAN The investigated Dokhan volcanic rocks range from green, buff and brown to purple and greyish black. These rocks include porphyries of andesite, dacite and rhyolite, as well as tuffs and moderately to highly welded ignimbrites. The latter are interstratified with porphyritic lava flows. The pyroclastic (locally ignimbritic) rocks are composed mainly of lithic and pumice clasts in a tuffaceous groundmass. In hand specimen, collapsed pumice may appear as dark blebs of devitrified glass (Ross & Smith, 1961). Lapilli fragments including quartzite and granite fragments are present. The repeated eruption of mafic to felsic volcanic materials locally produced layered structures, forming horizontal to slightly-tilted sheets, commonly 3-8 m thick. These Dokhan volcanic rocks overlie gabbro-diorite-tonalite complexes (Abdel-Rahman, 1990), but are intruded by granodiorite-granite rocks, trondhjemites and dykes (Fig. 2). 3. Geochronological context 24°- Figure 1. Distribution of the Dokhan volcanic rocks in the Eastern Desert of Egypt. Note the scarcity of these rocks in the southern part of Eastern Egypt. MK - Mount El-Kharaza Dokhan volcanic suite; MD - Mount Dokhan (the type locality); and F-S, - Fatira-Safaga Dokhan volcanic rocks. The rectangle outlines the map area shown in Figure 2. presence of welded ignimbritic textures in these rocks (possibly previously misidentified as metamorphic textures), may have led to these volcanic rocks being linked with the much older Shadli metavolcanic group. However, the absence of dioritic offshoots, and the presence of granitic clasts in these extrusive rocks, in addition to their younger age (620±16Ma; Abdel-Rahman & Doig, 1987), clearly indicate that these rocks belong to the Dokhan group. Equivalents to the Shadli metavolcanic group evidently do not exist in the study area. This agrees with observations by Ries et al. (1983) and Stern, Gottfried & Hedge (1984) that older metavolcanic and metasedimentary units are characteristic of southeastern Egypt, but are missing northwards. Schumann (1966) indicated a maximum estimated thickness of 1200 m for the Dokhan volcanic rocks in eastern Egypt. Basta, Kamel & Awadallah (1979) and Basta, Kotb & Awadallah (1980) reported on some petrographic and petrochemical features of these rocks in the type locality 'Mount Dokhan', 150 km southeast of the study area (Fig. 1). Ressetar & Monrad (1983) studied the geochemistry of some Dokhan volcanic rocks in the Fatira-Safaga area (about 300 km southeast of the study area). http://journals.cambridge.org Downloaded: 20 Mar 2015 Abdel-Rahman & Doig (1987) obtained a 7-point whole rock Rb-Sr age of 620+16 Ma, with an initial "Sr/^Sr ratio of 0.7039 ±0.0004 for the Dokhan volcanic rocks investigated. Rb-Sr ages of 654 to 665 Ma were obtained (El-Shazly et al. 1973) on Dokhan volcanic rocks from the central Eastern Desert of Egypt. Stern & Hedge (1985) obtained an Rb-Sr age of 592 ± 13 Ma for Dokhan volcanic rocks from Mount Dokhan. A Rb-Sr age of 602 ± 13 Ma was also obtained by R. J. Stern (unpub. Ph.D. thesis, Univ. California, San Diego, 1979) for some Dokhan andesite-rhyolite rocks near Qena-Safaga road. Ressetar & Monrad (1983) reported preliminary Rb-Sr data that suggest that eruption of Dokhan volcanic rocks took place as early as 620 Ma in the Fatira-Safaga area. Similar Rb-Sr ages (633-548 Ma; Segev, 1987) were obtained on orogenic, calc-alkaline, andesite-rhyolite rocks from Sinai, such as the Kid rhyodacite and the Rutig andesite-rhyolite (609±12Ma and 587±9Ma, respectively; M. Bielski, unpub. Ph.D. thesis, Hebrew Univ., Jerusalem, 1982). In Saudi Arabia, Fleck et al. (1980) obtained Rb-Sr ages of 620 ±90 Ma on andesites (Juqjuq Formation), 612 Ma on rhyolitic ignimbrite and 593 ± 53 Ma on dacite (Halaban Formation). Darbyshire et al. (1983) reported ages of 573 ± 23 Ma (Hummah volcanic rocks; andesite-dacite), 608 ± 9 Ma (Arfan volcanic rocks; basalt-andesite) and 616±13Ma (Jahhad volcanic rocks; basalt-rhyolite), whereas Duyverman, Harris & Hawkesworth (1982) reported an age of 625 ± 16 Ma for Nuqrah volcanic rocks, and Aldrich et al. (1978) obtained an age of 633 ± 15 Ma for Al-Madenah rhyolite (Murdama suite), all from Saudi Arabia. In Sudan, ages of numerous syn-orogenic to late orogenic arc-related volcanic rocks (similar to the Dokhan volcanic rocks) and coeval plutonic complexes ranging from 660 to 550 Ma are reported by Vail (1990). 4. Petrography Several petrographic units have been recognized within the Dokhan volcanic suite. Basalt- and andesite porphyry are IP address: 137.222.24.34 19 Pan-African volcanism, Nubian shield 3 3° '00' 10 km Phanerozoic sedimentary Gabbro-diorite-tonalite cover Dokhan volcanic rocks Granodiorite-adamelliteleucogranite Dyke ^ s* Fault + + M + + Trondhjemite + + + Figure 2. Geological map of igneous complexes in Mount El-Kharaza region, northeastern Egypt. composed of phenocrysts (10-40%), mainly of plagioclase, amphibole and minor augite (partially altered to chlorite and epidote), with magnetite microphenocrysts, in a microcrystalline groundmass. Plagioclase forms subhedral zoned grains showing glomeroporphyritic texture. Dacite porphyry is made up of quartz, feldspars, chloritized amphibole and biotite phenocrysts (15-20%) in a devitrified groundmass. Opaque magnetite occurs mostly as microphenocrysts. The groundmass consists of a dense mat of anhedral quartz and feldspars, occasionally in a microspherulitic array, probably representing devitrified glass. Rhyolite porphyry consists of quartz, alkali feldspar and plagioclase phenocrysts, with some magnetite and biotite microphenocrysts, in a microcrystalline felsitic groundmass (about 70% of the rock), occasionally showing micrographic or spherulitic textures. Accessory zircon, apatite and iron oxides are present. http://journals.cambridge.org Downloaded: 20 Mar 2015 Ignimbrite is characterized by the predominance of vesiculated juvenile material (pumice and shards), showing features indicating a pyroclastic flow origin. The ignimbritic flow-unit includes these petrographic varieties: (a) Fine-grained crystal-lithic welded tuff is a moderately to highly welded, fine-grained rock showing eutaxitic texture and composed of feldspar phenocrysts, devitrified pumice and glass shards (Fig. 3), with some iron oxides, in a tuffaceous to devitrified groundmass. Fine-grained, compacted and devitrified lapilli, commonly showing flow banding, are present. These lapilli are moulded against sharp corners of the phenocrysts in the firmly welded specimens (Fig. 3). Highly flattened, stretched and devitrified pumice clasts account for a small amount (20-25 %) of the rock. In some samples, slightly coarser, lenticular pumice clasts show the filamentous IP address: 137.222.24.34 20 A.-F. M. A B D E L - R A H M A N Table I. Major element whole-rock compositions (in wt %) and trace elements (in ppm) of the Dokhan volcanic suite Sample no. 27a 37a 40a 41a 43a 71a 77a 62a 69a 85a 86a 90a SiO, TiO, A1 2 6 3 K26 P,O, LOI 75.80 0.09 14.10 1.11 0.19 0.00 0.38 3.33 3.40 0.03 1.53 69.52 0.43 15.31 2.06 0.24 0.11 1.58 4.57 4.54 0.08 1.91 73.18 0.34 14.16 1.82 0.31 0.08 0.32 2.00 6.19 0.06 1.74 76.32 0.12 12.50 1.57 0.07 0.02 0.13 3.32 4.90 0.03 0.78 68.26 0.33 16.01 3.19 0.15 0.27 1.34 6.84 1.56 0.06 2.22 76.94 0.14 12.36 1.42 0.13 0.02 0.45 3.41 3.91 0.03 1.08 67.39 0.74 15.07 4.52 1.11 0.12 2.71 4.98 2.56 0.22 0.92 71.35 0.24 14.75 2.06 0.13 0.06 0.79 3.34 6.65 0.04 1.07 69.57 0.49 15.37 2.53 0.25 0.03 0.94 4.34 4.99 0.09 1.72 74.16 0.20 13.65 1.49 0.16 0.06 0.65 4.69 4.12 0.05 0.39 69.19 0.50 16.22 2.31 0.28 0.06 0.87 5.93 3.31 0.08 0.84 67.56 0.58 15.73 3.78 0.67 0.08 1.46 6.12 2.03 0.15 1.60 Total 99.96 100.36 100.21 99.76 100.23 99.88 100.34 100.48 100.32 99.61 99.58 99.76 50 445 194 113 99 72 41 111 1241 243 139 268 273 FeO MgO MnO CaO Na,0 Rb Sr Ba Zr Hf Y Nb U Th 87 95 137 116 58 77 102 148 89 81 148 87 193 33.1 12.3 12.9 13.1 295 244 200 336 21.3 12.0 23.7 13.2 20.1 13.8 4.8 9.2 4.4 8.7 89a LOI 67.77 0.47 16.62 2.71 0.63 0.10 1.81 5.17 4.14 0.11 0.86 Total 100.38 Rb Sr Ba 107 355 163 25.9 13.0 11.1 23.2 16.6 10.5 13.4 27b 78b 84b 66.60 0.89 16.02 4.84 0.30 0.07 1.66 4.31 4.53 0.27 0.50 48.62 0.99 13.13 8.61 8.82 0.15 7.41 2.53 0.86 0.26 9.27 56.12 1.01 17.66 7.93 3.36 0.13 7.06 3.38 1.14 0.31 1.76 99.99 100.65 197 207 6.3 762 341 292 30.0 13.1 4.8 6.5 16.5 85b 82b 70b 90b 56.92 0.71 18.86 6.25 3.24 0.08 5.06 4.16 2.38 0.20 2.34 53.51 0.81 18.17 8.04 3.75 0.18 8.86 4.14 1.99 0.27 0.65 63.99 0.72 15.69 5.56 1.57 0.06 4.39 4.67 1.67 0.18 1.58 47.94 1.91 16.13 11.79 5.44 0.19 7.34 2.84 2.03 0.53 4.33 62.23 0.86 15.67 6.37 2.06 0.12 4.78 4.23 2.19 0.24 1.01 57.74 1.23 15.32 8.60 2.43 0.13 4.04 4.90 1.85 0.33 3.23 55.25 0.72 17.48 7.26 4.07 0.11 6.30 4.89 1.52 0.18 1.92 58.13 0.62 17.70 6.55 2.66 0.08 5.34 5.03 1.40 0.15 2.20 99.87 100.22 100.37 100.07 100.48 99.76 99.80 99.70 99.86 29.3 7.8 6.1 17.3 14.1 20.0 12.1 6.9 7.9 13.5 11.2 8.1 7.9 234 5.8 27.1 10.0 12.6 6.5 19.3 12.8 184 27.9 12.7 10.0 5.8 Sample no. SiO 2 TiO 2 A1 2 O 3 FeO MgO MnO CaO Na,0 K26 P,O5 Zr Hf Y Nb U Th 125a 109 368 328 531 59 722 829 32 906 875 33 392 708 46 577 50 379 34 405 754 243 117 121 163 162 162 144 201 168 3.5 20.5 1051 310 7.0 20.6 13.3 22.5 13.8 4.2 6.3 7.4 11.4 106b 3.7 3.5 3.5 16.6 16.1 15.6 20.8 9.1 8.4 8.3 3.2 8.8 8.2 2.7 9.6 11.0 10.5 11.0 13.1 1.4 2.1 11.0 3.4 22.6 10.8 23.1 10.8 22.6 10.6 9.0 1.3 7.9 3.1 7.5 2.2 225b 229b 50 407 38 430 85 89 11.9 11.2 6.5 9.1 2.0 11.6 6.5 2.2 Total iron is given as FeO. LOI - loss on ignition. Sample numbers followed by the letter 'a' represent dacite-rhyolite (light-coloured) rocks, whereas those followed by the letter 'b' represent basalt-andesite-dacite (dark-coloured) rocks. Ba and Hf were analysed by NAA technique. nature of fiamme with their characteristic tails; devitrification products are quartz and feldspars. (b) Coarse-grained pumice-rich tuff is composed of large, coarse-grained, light grey filamentous pumice clasts and shards (70 % of the rock), with some feldspar phenocrysts in a tuffaceous, brownish groundmass. Scattered grains of magnetite and oxidized mafic minerals are present. The poorly sorted nature and the predomi- http://journals.cambridge.org Downloaded: 20 Mar 2015 nance of relatively large clasts of (devitrified) pumice are remarkable features in this rock. (c) Thin-bedded graded tuff is a greenish grey, banded, clast-free tuffaceous rock. It is made up of very fine-grained, vitreous-ashy materials, with tiny particles of quartz and feldspars, showing vari-coloured banding and laminations due to gradational changes in grain size. This rock is interpreted as an air-fall tuff, or the fine dust IP address: 137.222.24.34 21 Pan-African volcanism, Nubian shield exhibit a compositional gap. It forms roughly linear arrays with relatively smooth variations showing progressive decrease in Ca, Fe, Mg and Al (Fig. 4a), and progressive increase in alkalies (Fig. 5a), with gradual increase in silica. Most rocks plot in the 'subalkaline' field of Irvine & Baragar (1971; Fig. 5a), indicating that they are not alkaline in nature. Furthermore, the rocks define a typical calc-alkaline trend on the standard AFM diagram (Fig. 5b), forming a continuous calc-alkaline magmatic series. 6.b. Trace elements Figure 3. A photomicrograph of crystal-lithic welded tuff showing angular feldspar phenocrysts, lithic clasts, stretched glass shards and a pumice clast. Plane-polarized light, width of view 3.3 mm. elutriated from the top of the pyroclastic flow during flowage. Sparks, Self & Walker (1973) pointed out that the billowing cloud seen rising above nuees ardentes is thought to be the source of material for such a layer. 5. Analytical techniques 5.a. Major elements Concentrations of the major elements were determined by X-ray fluorescence spectrometry (Philips PW1400) on fused discs (operating conditions: Rh radiation, 40 kV, 70 mA). Loss on ignition (LOI) was determined by heating powdered samples for 50 minutes at 1000 °C. 5.b. Trace elements Concentrations of Rb, Sr, Zr, Y, Nb, U and Th were determined on pressed pellets by X-ray fluorescence (operating conditions: Rh radiation, 70 kV, 40 mA). The analytical precision, as calculated from 20 replicate analyses of one sample, is better than 1 % for most major elements and better than 5 % for most trace elements. 5.c. Rare earth elements (REE) The neutron-activation technique of Barnes & Gorton (1984) was used to determine concentrations of Ba, Cr, Sc, Co, Hf and eight rare-earth elements (REE; La, Ce, Nd, Sm, Eu, Tb, Yb and Lu). A precision of better than 5 % was achieved for most elements. Normalization values are those of the Leedy chondrite (Masuda, Nakamura 6 Tanaka, 1973) divided by 1.2 (Taylor & Gorton, 1977). 6. Geochemistry 6.a. Major elements The Dokhan volcanic rocks cover a wide range of composition and exhibit geochemical trends spanning the entire range from basalt to high-silica rhyolite (48-77 wt% SiO • Table 1, Fig. 4a). The suite does not appear to http://journals.cambridge.org Downloaded: 20 Mar 2015 The Dokhan volcanic rocks are also characterized by a wide range of trace elements; Sr varies from 81 to 906 ppm, Rb from 21 to 193 ppm, Ba from 708 to 1241 ppm, Zr from 85 to 340 ppm, Nb from 6.5 to 16.6 ppm and Y from 11 to 33 ppm (Tables 1, 2), typical of calc-alkaline orogenic suites (Gill, 1981). Hf does not show much variation (3.5-7 ppm) as it is usually much more depleted than Zr in most calcalkaline rocks. The compatible elements Cr, Sc and Co vary from 12 to 91,3.6 to 24 and 1 to 22 ppm, respectively. A plot of SiO2 vs. selected trace elements (Fig. 4b) shows a gradual general decrease in Sr and gradual increase in Rb, Zr and Nb with increasing silica. Note that, in the case of Rb and Zr, as well as the alkalies (cf. Fig. 5a), convex trends are obtained; the trends show gradual increases in these elements with increasing silica up to a maximum value of about 70wt% SiO,, after which a sharp decrease with increasing SiO2 occurs (between 70 and 77 wt% SiO,). This feature, also observed in other fractionating calc-alkaline suites, has been documented by several authors (Dietrich, 1968; Watson, 1979; Miller & Mittlefehldt, 1984). Watson (1979) demonstrated experimentally that zircon saturation in a magma shows pronounced dependence upon the (Na2O + FC,O/A12O3) ratio of the melt; the low solubility of zircon in melts with low alkali/aluminium ratio result in an early crystallization of zircon, causing Zr depletion in the more evolved products of crystallization (high-silica rhyolites). This is consistent with the striking similarity in the variation of both Zr and alkalies vs. silica obtained in Figures 4b and 5a. When highly incompatible elements are plotted against each other, such as Zr vs. Nb (Fig. 6a), a linear trend (with some scatter) is obtained, showing gradual increase in Zr with increasing Nb. Rb also shows a similar trend. The Nb/Y vs. Zr/TiO2 diagram (Fig. 6b; Winchester & Floyd, 1977) shows the classification of this volcanic suite; it contains basalt, andesite, dacite, rhyodacite and rhyolite. It should be noted that these rock types exhibit relatively similar Nb/Y ratios. The Zr/TiO2 ratios increase gradually from basalt to rhyolite (Fig. 6b), as Zr is more incompatible than Ti. The relative concentrations of trace elements are shown in the form of a primitive mantle-normalized diagram (Fig. 7a), based on the normalization factors of Sun & McDonough (1989). In general, all rocks show moderate concentrations of most large ion lithophile elements, high IP address: 137.222.24.34 22 A.-F. 10 - 0 ID 0 8 - i L • N 200 D a • j i - • 6 - 4 - 2 - I i i n i i a 120 40 D i 160 Oo O o D D . • D D _ nlP D 1000 a 800 600 . 0 400 D * 20 12 I 80 8 16 i a 10 0 < a i 200 a - o D D 100 a cP a • D 0 U. oo 300 6 2 0) I 0 0°oOo m 0 o o£Ln , , ° cP.SP 2 - 01 4 0 o o 0 | Z -- Dacite-rhyolite 0 D° 4 - (b) 20 Basalt-andesite o D a 6 • _ 10 - 0 (a) a 8 M. ABDEL-RAHMAN - • a ada 30 QD O°o 200 OO • 0 50 ooo 60 ,OO°,QP 70 80 SIOo. CD i 50 60 70 80 SiO 2 Figure 4. Variations of SiO., vs. (a) other major elements, and (b) selected trace elements, within the Dokhan volcanic suite. field strength elements and light rare earth elements. The dacite-rhyolite rocks show pronounced Nb, Sr and Ti negative anomalies, whereas the less-evolved basalt-andesite rocks exhibit Th and Nb negative anomalies (Fig. 7a). These patterns are typical of modern arc-related calc-alkaline orogenic suites, as the rhyolite-dacite ash-flow tuffs and coeval andesite-rhyolite lavas of the Indian Peak volcanicfieldin Nevada and Utah (Best, Christiansen & Blank, 1989). REE patterns of the less-evolved rocks (53.5-64 wt% SiO2; Fig. 7b) are similar overall to those of arc-related basalts (50-53 wt% SiO2) of the southern Andes 36-41°S (Lopez-Escobar, 1984), but with relatively higher REE values. The REE profiles of the dacite-rhyolite rocks (Fig. 7c) resemble those of the Andean ignimbrites of northern Chile and northwestern Argentina (Francis et al. 1989), and are thus typical of light-rareearth-enriched calc-alkaline volcanic rocks. 6.c. The rare earth elements (REE) Rock varieties of the Dokhan volcanic suite show a general increase in REE contents toward the more-evolved rocks (Table 2), and exhibit subparallel REE chondritenormalized patterns (Fig. 7b, c). All samples show light REE enrichment over heavy REE, and display a range of negative Eu-anomalies; the more-evolved lithologies exhibiting larger anomalies compared to the less-evolved ones (Fig. 7b, c). This could be due to low-pressure fractionation of plagioclase that led to depletion in Sr and Eu in the more-evolved rocks. Higher oxidation conditions in more-evolved melts could also cause negative Euanomalies. http://journals.cambridge.org Downloaded: 20 Mar 2015 7. Discussion 7.a. Bimodality The Dokhan volcanic rocks described here form a continuum in composition from basalt and andesite to dacite and rhyolite, with no apparent compositional gap (cf. Figs 4a, b, 5b). Abdel-Rahman & Martin (1989) showed also that other Dokhan volcanic rocks from the Safaga-Fatira area which were studied by Ressetar & Monrad (1983) cover a similar wide range in composition. Thus, the basis for the inference that Dokhan volcanic rocks are generally 'bimodal suites' (Stern, Gottfried & Hedge, 1984; Stern & Gottfried, 1986; Stern, Sellers & Gottfried, 1988) is not IP address: 137.222.24.34 23 Pan-African volcanism, Nubian shield 20 18 16 14 Table 2. Rare earth and other trace element compositions (in ppm) of the Dokhan volcanic suite (a) D O Biull-andeiite Dacite-rhyolite 12 10 » 6 ® Po o° (9 C s/ DD ] 4 Suba Ikalln 2 / 35 40 45 50 55 60 65 70 75 80 85 Sample no. 62a 90a La Ce Nd Sm Eu Tb Yb Lu 46.49 86.49 34.47 6.46 0.79 1.00 2.83 0.44 27.01 53.55 24.48 6.09 1.23 0.77 2.77 0.40 Cr Sc Co 16.10 3.56 1.04 16.50 11.20 2.39 84b 85b 82b 31.72 71.06 24.73 5.12 1.28 0.71 1.95 0.37 24.97 51.02 20.97 4.69 1.31 0.70 1.47 0.30 21.58 44.53 20.43 4.50 1.20 0.46 1.57 0.26 16.96 41.29 18.62 4.39 1.31 0.80 2.21 0.39 17.64 38.70 19.96 4.92 1.15 0.64 1.79 0.32 19.90 4.35 3.02 31.80 12.60 22.33 90.70 16.30 22.34 24.20 16.90 9.17 12.10 24.32 20.75 125a 106b SiO 2 The rare earth elements as well as Cr, Sc and Co were analysed by the NAA technique. Sample numbers followed by the letter 'a' represent dacite-rhyolite (light-coloured) rocks, whereas those followed by the letter 'b' represent basalt-andesite-dacite (dark-coloured) rocks. FeO* (b) D Basalt-andesite o Dacite-rhyolite Na2O+K2O MgO Figure 5. (a) SiO2 vs. Na,0 + K.,0 variation diagram; the alkaline and subalkalinefieldsare after Irvine & Baragar(1971). (b) A-F-M diagram showing the calc-alkaline nature of the Dokhan volcanic suite. readily apparent. It should be noted that, in these three studies, the latter authors used data from unrelated rock types as dykes, 'Younger' granites, and volcanic rocks of different ages and localities (80 to 100 km apart), and considered all to be products of the same tectonic environment. Thus, the apparent bimodality obtained by those authors was in part the result of using unrelated stratigraphic units and may also be an artifact of sampling. 7.b. Tectonic environment 7.b.l. Previous models With three different tectonic models previously proposed by various authors for its emplacement (extensional, transitional and arc-related settings), the tectonic environment of the Dokhan volcanic suite is still controversial. In the opinion of Stern & Gottfried (1986), the 'Dokhan' suite in the Dokhan-Qattar area is dominantly mafic in character, defines a calc-alkaline trend on the AFM diagram and is comparable to typical orogenic http://journals.cambridge.org Downloaded: 20 Mar 2015 andesites, but with moderate enrichments in incompatible elements. However, they considered the suite to be bimodal in nature and have used this, along with its moderate enrichment in some incompatible elements to regard these volcanic rocks to be products of 'strong crustal extension' (or anorogenic). Ressetar & Monrad (1983) found the Dokhan volcanic rocks of the Safaga-Fatira area to exhibit abundances of most major and trace elements similar to subduction-related calcalkaline volcanic rocks in orogenic belts, but with slightly higher contents of alkalies, Ti, Zr and Nb than typical island arc or continental margin volcanic suites. These authors found that the analyses plot near the boundary between suites considered to be subduction-related and those considered typical of late-orogenic or extensional environments (their figs 5, 6). They interpreted these rocks to have been formed in an environment transitional between compressive and extensional tectonic settings. It should be noted, however, that slight enrichment in some incompatible elements within arc-related volcanic rocks is not uncommon. As pointed out by Saunders & Tarney (1979), dehydration of the downgoing oceanic slab enriches the overlying mantle wedge, and subsequently magmas generated from it, in large-ion-lithophile (LIL) elements. Higher Zr contents which may occasionally reach within-plate magma values (see, e.g. Gokten & Floyd, 1987) are a characteristic feature of calc-alkaline volcanic rocks formed in a continental arc (margin) environment (Pearce, 1983; Capan & Floyd, 1985). Gill (1981) and Ringwood (1990) demonstrated that the behaviour of trace elements, particularly REE and HFS elements, in subduction zone environments, will influence the trace-element signature of arc-related volcanic rocks. Recent results of Hickmott, Sorensen & Rogers (1992) showed that heavy REEs are selectively transported during metasomatism in a subduction complex, whereas Philippot & Silverstone (1991) showed that both Ti and Zr may be strongly concentrated in aqueous fluids at subduction zone P-Tconditions. IP address: 137.222.24.34 24 A.-F. 7.c. Nature of analogous volcanic and coeval plutonic rocks (a) - • Basalt-andesite o Dacite-rhyolite O 15 Q10 Z ©° o - • s^ 5 1 1 i 1 200 - i 1 0 § o 40 i i 100 i 180 1 1 260 8° I o o 1 340 (b) O Biialt-udesite O Dacite-rhyolite h .1 L N .01 Andesite/Basalt SubAlkaline Basalt .001 .01 .1 Alk-Bas Bsn/Nph" 10 Figure 6. (a) Nb and Rb vs. Zr variation diagram, (b) Log (Nb/Y) vs. log (Zr/TiO,) diagram showing the classification of the Dokhan volcanic rocks; fields are after Winchester & Floyd (1977). 7.b.2. Suggested tectonic model The Dokhan volcanic suite is calc-alkaline in nature (cf. Fig. 5), and its trace element and REE patterns resemble those typical of modern, arc-related, orogenic suites (Fig. 7). Data points of basalt-andesite rocks plot, on geochemical-tectonic discriminant diagrams (such as Fig. 8a), in the field of calc-alkaline basalts (CAB) of Pearce & Cann (1973). The dacite-rhyolite rocks plot (Fig. 8b) in the volcanic arc field of Pearce, Harris & Tindle (1984) and in the late orogenic and syn-collision fields (Fig. 8c) of Batchelor & Bowden (1985). None of the data points plot in the 'anorogenic' field. Thus, the Dokhan volcanic suite is interpreted to be the product of subduction-related volcanism produced during late stages of the Pan-African orogeny. http://journals.cambridge.org M. ABDEL-RAHMAN Downloaded: 20 Mar 2015 In Sinai, Segev (1987) showed that the 633-548 Ma rhyodacitic rocks belong to the orogenic stage. In Saudi Arabia, similar volcanic suites including the Hummah suite (573 ±23 Ma; Darbyshire et al. 1983), Shammar suite (572-557 Ma; Hadley, 1973) and Murdama suite (633±15Ma;Aldriche?a/. 1978; Roobol et al. 1983) are calc-alkaline, orogenic rocks. Schmidt & Brown (1982) and Stuckless et al. (1983) showed that the 620-560 Ma magmatism in northwestern Arabia predominantly involved calc-alkaline I-type melts. In Sudan, numerous synorogenic to late orogenic volcanic rocks (equivalent to the Dokhan volcanic rocks) and coeval plutonic complexes (660-550 Ma) are interpreted to be the result of arc-magmatism (e.g. Vail, 1990, and references therein). Thus, analogous volcanic suites in Sinai, Sudan and Saudi Arabia also have been interpreted as products of destructive plate margin magmatism. Also, coeval granodioritic batholiths covering large areas in the Arabian-Nubian shield were emplaced during the Late Pan-African magmatic episode (~650-550 Ma); these occur as volumetrically significant, I-type, calc-alkaline, Andean-type complexes (Neary, Gass & Cavanagh, 1976; Brown, 1980; Gass, 1982; Abdel-Rahman & Martin, 1987; Vail, 1990). Rift-related volcanic (and plutonic) complexes do occur within the Pan-African belt, but these are younger (mostly less than 550 Ma), post-Pan-African minor suites, locally emplaced within an otherwise calc-alkaline, Pan-African orogenic assemblage of batholithic dimensions. The bimodal volcanic-subvolcanic suite (550-540 Ma) that crops out on the eastern flank of Wadi Araba Rift Valley, southwest Jordan (Jarrar, Wachendorf & Saffarini, 1992), is a good example of rift-related, postPan-African anorogenic volcanic rocks. Unlike orogenic volcanism, products of anorogenic volcanism typically erupt along ring fractures, forming relatively small circular or elongated tabular structures and are alkaline to peralkaline in composition (see, e.g. Abdel-Rahman & Martin, 1990/?; Abdel-Rahman & Miller, 1993). In view of the lack of clear evidence for 'crustal extension', or 'transitional' tectonic environment advanced by previous authors, it is suggested that the slight enrichment in some incompatible elements (Na, K, Ti, Zr, Nb) observed in some Dokhan volcanic rocks from the Qattar-Dokhan and Safaga-Fatira areas may reflect enrichment above subduction zones, and do not reflect anorogenic emplacement. Thus, a subduction-related tectonic setting for the emplacement of the investigated suite is indicated by the petrological-geochemical features documented herein. 7.d. Petrogenesis of the Dokhan volcanic suite l.d.l. Fractional crystallization vs. other processes Rocks of the Dokhan volcanic suite cover the entire spectrum from basalt to rhyolite. The smooth variations in IP address: 137.222.24.34 Pan-African volcanism, Nubian shield 25 Basalt-Andesite Dacite-Rhyolite I «• a: o O o 1 - i 1 : 1 (b) Basalt-Andesite I ROCK /CHONDRI1 UJ IAB (S. Andes) i i i i La Co Pr Nd i i i i I 1 1 1 1 1 1 Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Dacite-Rhyolite (c) Basalt-Andesite La Co Pr Nd I I I Pm Sm Eu Gd Tb Dy Ho Er TmYb Lu Figure 7. (a) Primitive mantle-normalized trace element plot showing comparison of element distributions for the lessevolved basalt-andesite rocks (patterns with a square symbol), superimposed on a dotted envelope representing the moreevolved dacite-rhyolite rocks of the Dokhan volcanic suite, (b) Chondrite-normalized plot showing four rare-earth patterns for less-evolved Dokhan volcanic rocks (basalt-andesite; 53.5-64 wt% SiO2), and an envelope (in diagonal lines) representing REE patterns for typical arc-related basalts from Laguna del Maule, Antuco and Cordon Cenizos volcanic complexes of the http://journals.cambridge.org Downloaded: 20 Mar 2015 major and trace elements with well-defined geochemical trends (Figs 4,6), and the general increase in the contents of REE accompanied by larger negative Eu-anomalies in the more evolved rocks, may suggest that the different lithologies of the suite are co-magmatic, that is, derived from a single magma body, and were possibly evolved by fractional crystallization. There is no field, isotopic or any other evidence to suggest processes such as magma mixing or crustal assimilation for the evolution of this suite. Furthermore, the Sr-isotope data indicate that all analysed samples fall on a linear array, defining a exceptionally coherent trend with a positive correlation between 87Sr/86Sr and Rb/Sr ratios (Fig. 9). The 87Sr/86Sr ratios occupy a wide range from 0.7053 to 0.7436 (Abdel-Rahman & Doig, 1987), reflecting the wide compositional range of the suite, and further support the concept that its various rock types were derived from the same source. The mafic nature of the less-evolved members of the suite, along with its low initial 87Sr/86Sr ratio (0.7039), which is considered lower than contemporary values in a source region containing ancient sialic basement, preclude an origin by melting of older (e.g. Archean) crustal material. Moreover, there is no evidence to suggest the presence of Archean crust at least in the northern Nubian shield (see, e.g. Kroner et al. 1988). Several genetic models have been proposed for the origin of calc-alkaline magmas. The partial melting of water-bearing upper-mantle and lower-crust rocks provides a general working hypothesis for calc-alkaline magma genesis. Subducted oceanic crust, modified by exchange with seawater, is found to be an important source material for such magma (Gill, 1981; DeVore, 1983). But partial melting of such subducted oceanic crust (at depth 100-200 km) would produce magmas of andesitic-dacitic compositions (Green & Ringwood, 1968; Marsh & Carmichael, 1974; Green, 1980). Judging from the mafic nature of the magma of the Dokhan volcanic suite (calc-alkaline basalt; 48wt% SiO2) and the low initial Sr-isotopic ratio, partial melting of a mantle material (e.g. hydrated lherzolite) is a more likely source. The relative enrichment of incompatible elements including light-rare-earth elements (cf. Tables 1, 2; Fig. 7), is considered to manifest incorporation of slab-derived incompatible components (via a fluid phase) into a relatively undepleted mantle wedge above a subduction zone. Partial melting of hydrated mantle rocks southern Andes 3 6 ^ 1 °S, taken from Lopez-Escobar (1984). (c) Chondrite-normalized plot showing three REE patterns for the more-evolved Dokhan volcanic rocks (dacite-rhyolite; 67.6-71.4 wt% SiO2), and an envelope (in diagonal lines) representing the less-evolved Dokhan volcanic rocks. These patterns are superimposed on an envelope (defined by thick lines) representing REE profiles of typical Andean ignimbrites of northern Chile and northwestern Argentina, taken from Francis etal. (1989); see text for details. IP address: 137.222.24.34 26 A.-F. M. ABDEL-RAHMAN Ti/100 D Basalt-andesite - O Dacite-rhyolite (a) o L D Basalt-andesite (/} is a 0.730 - i 0.720 _ o 0 w » o° 0.710 DO i 0.0 i 0.4 I I I 0.6 I 1.2 I I 1-6 Rb/Sr Sr/2 Zr 1000 D Basalt-andesite ; o Dacite-rhyolite Figure 9. Rb/Sr vs. "Sr/^Sr variation diagram for the Dokhan volcanic suite; data points exhibit an exceptionally coherent trend. would produce calc-alkaline basaltic magma, from which this basalt-rhyolite series could have been produced, possibly by fractional crystallization. (b) 7.d.2. Fractionation modelling: major elements 100 * 2000 0) 2 + 1-Mantle Fractionates 2—Pre-plate Collision 3-Post-collision Uplift 4—Late-orogenic 5-Anorogenic 6-Syn—collision 7—Post-orogenic (c) O Dacite-rhyolite Q 1000 (0 II W o: 3000 1000 2000 R 1 =4Si-1 1 (Na+K)-2(Fe+Ti) Figure 8. (a) Ti-Zr-Sr triangular diagram showing that the Dokhan basalt-andesite rocks (squares) plot in the field of calcalkaline basalt (CAB) of Pearce & Cann (1973). The other two fields are island-arc basalt (IAB) and ocean floor basalt (OFB). (b) Plot of Rb vs. (Y + Nb); Fields VAG, volcanic arc granite; Syn-COLG - syn-collision granite; WPG - within-plate granite; and ORG - oceanic ridge granite, are after Pearce, Harris & Tindle (1984). Circles represent dacitic-rhyolite rocks and squares generally represent andesitic rocks and plotted here only for reference, (c) R1 [4Si - 11 (Na + K) - 2(Fe + Ti)] vs. R2 [6Ca + 2Mg + Al] diagram showing that the Dokhan dacitic-rhyolitic rocks are mostly late-orogenic; none of the samples plot in the anorogenic field (fields are after Batchelor & Bowden, 1985). http://journals.cambridge.org Downloaded: 20 Mar 2015 Modelling of closed-system fractional crystallization by least-squares methods (Wright & Doherty, 1970) can be used to approximate the amounts and proportions of fractionating phases (Perry et al. 1990). Due to changes in the phenocryst assemblages and in the compositions of solidsolution minerals within this compositionally broad basalt-rhyolite series, modelling in three stages using representative compositions of fractionating phases from three parental compositions is more appropriate. Parental compositions were chosen at similar intervals of SiO2 contents along the 'liquid line of descent', starting from the least silica-rich basalt (sample 70b), to the most silica-rich rhyolite (71a). The composition of the mineral phases used are the actual microprobe data of observed phenocrysts. Fractionation models (Tables 3a, b, c) show that the most evolved (daughter) compositions can be produced from the least-evolved (parent) magmas. Thus, in stage 1, andesite (daughter, 106b) was produced from a basaltic parent (70b) by fractional crystallization (with a 54.2 % residual liquid) of 26.4% amphibole, 13.4% bytownite and 6.1 % magnetite (normalized to 100; 57.5 %, 29.2 %, 13.3%, respectively). In stage 2, dacite (daughter, 90a) was produced from the andesitic parent by fractionation (with a 63.8% residual liquid) of 11.2% amphibole, 19.9% andesine, and 5.1% magnetite (normalized to 100; 30.9 %, 55 %, 14.1 %, respectively). In stage 3, highsilica rhyolite (daughter, 71a) was produced from the dacitic parent by fractionation (with a 51.4% residual liquid) of 2.2% biotite, 43.8% oligoclase and 2.6% magnetite (normalized to 100; 4.5%, 90.1%, 5.3%, respectively). The low values of the sum of the squares of residuals (ZR2: 0.87, 0.20 and 0.26, for stages 1 to 3, respectively) indicate a reasonable fit (Tables 3a, b, c). Several authors (e.g. Green & Ringwood, 1968; Gill, 1981; Anderson, 1980; Kokelaar, 1986) have demonstrated the role of amphibole fractionation to produce IP address: 137.222.24.34 Pan-African volcanism, Nubian shield 27 Table 3a. Results of crystal fractionation least-squares modelling from basalt to andesite Fractionating phases amphibole 42.34 1.37 14.72 9.21 14.83 0.05 12.92 0.96 1.11 SiO 2 TiO 2 A12O3 FeO MgO MnO CaO Na,0 K,6 plagioclase 49.73 29.52 0.41 0.09 16.30 3.56 0.66 magnetite 0.68 11.41 2.32 77.96 3.00 _ — - Observed daughter (no. 106b) Observed Dtircnt (no. 70b) Calculated parent 60.00 1.28 15.92 8.94 2.52 0.14 4.20 5.09 1.92 50.14 2.00 16.87 12.33 5.70 0.20 7.68 2.97 2.12 50.15 1.90 16.62 12.35 5.65 0.22 7.98 3.58 1.54 Weighted residual -0.01 0.10 0.25 -0.02 0.05 -0.02 -O.30 -0.61 0.58 The daughter (sample no. 106b) and the parent (no. 70b) are recalculated to 100% and the total iron is given as FeO. The compositions of amphibole, plagioclase and magnetite are the actual electron microprobe analyses of phenocryst phases. Residual liquid (no. 106b): 54.2%. Fractionating phases: 45.8 % (amphibole 26.4 %, bytownite 13.4 %, magnetite 6.1 %). Sum of the squares of residuals (ER 2 ) is 0.87. Table 3b. Results of crystal fractionation least-squares modelling from andesite to dacite Fractionating phases amphibole SiO 2 TiO 2 A1,O3 FeO MgO MnO CaO Na,O K.6 53.13 0.94 2.60 14.20 14.92 0.18 11.98 0.08 0.11 plagioclase 56.92 27.12 0.23 9.05 5.57 0.82 magnetite 0.78 5.34 0.26 88.53 - Observed daughter (no. 90a) Observed parent (no. 106b) Calculated parent 68.93 0.59 16.05 3.86 0.68 0.08 1.50 6.24 2.07 60.00 1.28 15.92 8.94 2.52 0.14 4.20 5.09 1.92 60.00 1.07 15.82 8.95 2.40 0.39 4.33 5.27 1.77 Weighted residual 0.00 0.21 0.10 -0.01 0.12 -0.25 -0.13 -0.18 0.15 The daughter (sample no. 90a) and the parent (no. 106b) are recalculated to 100 % and the total iron is given as FeO. The compositions of amphibole, plagioclase and magnetite are the actual electron microprobe analyses of phenocryst phases. Residual liquid (no. 90a): 63.8 %. Fractionating phases: 36.2 % (amphibole 11.2 %, andesine 19.9 %, magnetite 5.1 %). Sum of the squares of residuals ( I R 2 ) is 0.20. Table 3c. Results of crystal fractionation least-squares modelling from dacite to high-silica rhyolite Fractionating phases biotite SiO 2 TiO 2 A12O3 FeO MgO MnO CaO Nafi K,0 36.41 2.97 12.25 20.49 11.96 0.12 0.33 9.28 plagioclase 65.61 21.50 0.12 0.16 2.36 10.47 0.14 magnetite 0.51 2.14 0.32 90.71 - Observed daughter (no. 71a) Observed Durcnt (no. 90a) Calculated parent 77.89 0.14 12.51 1.44 0.13 0.02 0.46 3.45 3.96 68.93 0.59 16.05 3.86 0.68 0.08 1.50 6.24 2.07 68.91 0.32 16.06 3.87 0.53 0.13 1.37 6.40 2.41 Weighted residual 0.02 0.27 -0.01 -0.01 0.15 -0.05 0.13 -0.16 -0.34 The daughter (sample no. 71a) and the parent (no. 90a) are recalculated to 100% and the total iron is given as FeO. The compositions of biotite, plagioclase and magnetite are the actual electron microprobe analyses of phenocryst phases. Residual liquid (no. 71a): 51.4%. Fractionating phases: 48.6 % (biotite 2.2 %, oligoclase 43.8 %, magnetite 2.6 %). Sum of the squares of residuals (SR 2 ) is 0.26. andesite from a basaltic magma. Cawthorn & O'Hara (1976) have shown that in mafic and intermediate calcalkaline magmas, amphibole is an important nearliquidus phase over the range PHlO = 2-\Q kb. More recently, Romick, Kay & Kay (1992) studied calc-alkaline andesite and dacite tephra from the central Aleutian http://journals.cambridge.org Downloaded: 20 Mar 2015 arc (Alaska) and concluded that trace element modelling supports an open system origin for the andesitic lavas involving amphibole fractionation, whether or not amphibole is a modal mineral. In view of the results of fractionation modelling, it is clear that amphibole dominates the early fractionating mineral assemblage from basalt to IP address: 137.222.24.34 28 A.-F. andesite. Plagioclase minerals were the dominant fractionating phases from andesite to dacite to high-silica rhyolite. This is consistent with the abundance of plagioclase as a phenocryst phase in these rocks, and with the progressive decrease in Sr with increasing SiO r The continuous Fe depletion (Fig. 4a) and suppression of relative Fe enrichment that produced the calc-alkaline trend (Fig. 5b) indicate the continuous removal of magnetite which, with amphibole, is also responsible for the depletion of Ti (cf. Table 1). The relatively well-defined geochemical trends obtained for the Dokhan volcanic suite, along with the progressive increase in Eu-anomalies toward moreevolved compositions, combined with results of the geochemical modelling, support the idea that a fractional crystallization process has played a major role during the evolution of the Dokhan volcanic magma series. Table 4. Partition coefficients used in the trace-element fractional crystallization modelling (after Arth, 1976, and Gill, 1981) La Ce Nd Sm Eu Yb Lu Nb Rb Sr Plagioclase Magnetite 0.40 0.510 1.20 2.0 1.70 2.0 1.7 1.3 0.05 0.23 0.35 0.24 0.17 0.13 2.11 0.077 0.062 0.025 0.041 4.4 0.098 0.11 0.13 0.15 0.1 0.17 0.19 0.90 0.0 0.0 Table 5. Results of trace element fractional crystallization models at 70 % fractionation 7.d.3. Trace element modelling La Ce Nd Sm Eu Yb Lu Nb Rb Sr where C is the initial concentration of the trace element in the source, F is the fraction of liquid remaining, and Ds is the bulk distribution coefficients of the fractionating minerals. Modelling was carried out in one stage using available trace element data. The calculations were done using trace element values (Cj) of the least fractionated sample for which rare earth data are available (85b; 53.5 wt% SiO,). The percentage of fractionating phases used are those obtained from the least-squares fractionation modelling (Table 3b); amphibole 12.1, plagioclase 22.5 and magnetite 4.4 (normalized to 100: 31 % amphibole, 58 % plagioclase, 11 % magnetite). The distribution coefficients (KDs) used are taken from Arth (1976) and Gill (1981) and are listed in Table 4. The results (Table 5), indicate a close match between calculated REE and other trace element values in the differentiated liquid at 70 % crystallization and those in the most fractionated sample measured (62a; 71.4 wt % SiO,). It should be noted that calculated trace element abundances are considered acceptable if model values are within 20 % of measured values (Graham & Hackett, 1987; Patterson & Graham, 1988). However, the calculated Rb values differ significantly from those measured. This is due, in part, to the wide range of KDs of some elements (some may differ from actual KD), changes in the phenocryst assemblage and changes in the compositions of solid-solution minerals during the course of such fractional crystallization process. 8. Summary and conclusions (1) The volcanic history of the Pan-African belt in eastern Egypt encompasses two major magmatic episodes, an ear- http://journals.cambridge.org Amphibole Source (no. 85b) To further constrain this fractional crystallization hypothesis, trace element modelling using the equation of Arth (1976) was applied. This equation was solved for CL (the concentration of the trace element in the differentiated liquid: Downloaded: 20 Mar 2015 M. ABDEL-RAHMAN 21.58 44.53 20.43 4.5 1.2 1.57 0.26 8.8 32.2 906.0 Fractionated (no. 62a) 46.49 86.49 34.47 6.46 0.79 2.83 0.44 13.1 194.0 110.5 Fractionated liquid (calculated) 48.34 102.75 38.04 6.36 0.52 2.29 0.43 15.26 102.53 150.44 Her episode (950-750 Ma) produced the Shadli metavolcanic assemblages, and a younger episode (680-550 Ma) produced the Dokhan volcanic rocks. (2) The Dokhan volcanic suite (620 Ma) includes basalt, andesite, dacite (dark-coloured rocks), and dacite, rhyodacite, rhyolite ignimbrite (light-coloured rocks). (3) Geochemically, the suite forms a continuum in composition with a wide range of SiO2 (48-77 wt %; Fig. 4a), and all lithologies are considered co-magmatic. The lack of a compositional gap precludes its identification as a bimodal suite. The rocks are clearly orogenic calc-alkaline in nature (Fig. 5). (4) Rb and Nb correlate positively and linearly with Zr (Fig. 6a). The sum of REE increases from basaltic to rhyolitic compositions, and the patterns are subparallel but with increasing Eu-anomalies toward rhyolites (Fig. 7b, c). Trace element patterns generally resemble those of typical Andean orogenic andesites and ignimbrites. (5) All lithologies exhibit geochemical traits of arcrelated volcanism (Fig. 8). The suite resembles products of neither extensional nor transitional tectonic regimes as previously thought, but was produced in a subductionrelated tectonic environment. (6) The low initial '"Sr/^Sr ratio (0.7039) of the suite precludes a crustal ensialic origin. The rocks are considered to be derived from a mantle source produced above a subducted slab of lithosphere. Melting of such a hydrated mantle wedge produced a calc-alkaline basaltic magma. The latter fractionated (11-26% amphibole, 13-44% IP address: 137.222.24.34 Pan-African volcanism, Nubian shield 29 plagioclase, 3 - 6 % magnetite; normalized to 100: 3 1 - 5 8 % , 2 9 - 9 0 % , 5-14%, respectively) to produce a continuous basalt to high-silica rhyolite calc-alkaline series, as suggested by well-defined geochemical trends and documented by major and trace element fractionation modelling. Further investigations are warranted in view of the significance of the Dokhan volcanic sequences that occur on a regional scale in northeastern Egypt, Sinai, Sudan and Saudi Arabia, and the potential for increased insight into Late Proterozoic crustal development in the northern Arabian-Nubian shield. Acknowledgements. I thank Dr R. F. Martin for valuable comments and for reading an earlier version of this work. I have benefitted from discussions with Drs D. Francis, A. Hynes, A. Lalonde and S. Nadeau. I am grateful to many geologists at the Geological Survey of Egypt, particularly Dr A. Hussein and M. Said for their help with logistics which facilitated field work. I thank A. Ahmadali and G. Piori who facilitated the acquisition of chemical analysis. This study was funded through a Natural Sciences and Engineering Research Council of Canada (NSERC) research grant to the author. Careful reviews by two anonymous referees and by the editors substantially improved the manuscript and are greatly appreciated. results for international Newsletter^, 17-23. standards. Geostandards BASTA, E. Z., KAMEL, O. A. & AWADALLAH, M. F. 1979. Petrography of Gabal Dokhan volcanics, Eastern Desert, Egypt. Egyptian Journal of Geology 22, 145-71. BASTA, E. Z., KOTB, H. & AWADALLAH, M. F. 1980. Petrochemical and geochemical characteristics of the Dokhan formation at the type locality, Jabal Dokhan, Eastern Desert, Egypt. Institute of Applied Geology Bulletin, Jeddah 3, 122-40. BATCHELOR, R. A. & BOWDEN, P. 1985. Petrogenetic interpreta- tion of granitoid rock series using multicationic parameters. Chemical Geology 48,43-55. BEST, M. G., CHRISTIANSEN, E. H. & BLANK, R. H. 1989. Oligocene caldera complex and calc-alkaline tuffs and lavas of the Indian Peak volcanic field, Nevada and Utah. Geological Society ofAmerica Bulletin 101, 1076-90. BROWN, G. C. 1980. Calc-alkaline magma genesis: the Pan-African contribution to crustal growth? Institute of Applied Geology Bulletin, Jeddah 3, 19-29. CAPAN, U. Z. & FLOYD, F. A. 1985. Geochemical and petro- graphic features of metabasites within units of the Ankara melange, Turkey. Ofioliti 10, 3-18. CAWTHORN, R. G. & O'HARA, M. J. 1976. Amphibole fractiona- tion in calc-alkaline magma genesis. American Journal of Science 276, 309-29. DARBYSHIRE, D. P. F, JACKSON, N. J., RAMSAY, C. R. & ROOBOL, References ABDEL-RAHMAN, A. M. 1990. Petrogenesis of early-orogenic diorites, tonalites and post-orogenic trondhjemites in the Nubian shield. Journal of Petrology 31, 1285-1312. ABDEL-RAHMAN, A. M. & DOIG, R. 1987. The Rb-Sr geochronological evolution of the Ras Gharib segment of the northern Nubian shield. Journal of the Geological Society, London 144, 577-86. ABDEL-RAHMAN, A. M. & MARTIN, R. F. 1987. Late Pan-African magmatism and crustal development in northeastern Egypt. Geological Journal 22,281-301. ABDEL-RAHMAN, A. M. & MARTIN, R. F. 1989. Late Pan-African magmatism and crustal development in northeastern Egypt: Reply. Geological Journal 24,375-81. N. J. 1983. Rb-Sr isotope study of latest Proterozoic volcano-sedimentary belts in the Central Arabian shield. Journal of the Geological Society, London 140,203-13. DEVORE, G. W. 1983. The influence of submarine weathering of basalts on their partial melting during subduction. Lithos 16,203-13. DIETRICH, R. V. 1968. Behavior of zirconium in certain artificial magmas under diverse P-Tconditions. Lithos 1, 20-9. DUYVERMAN, H. J., HARRIS, N. B. W. & HAWKESWORTH, C. J. 1982. Crustal accretion in the Pan-African: Nd and Sr isotope evidence from the Arabian shield. Earth and Planetary Science Letters 59, 315-26. EL-RAMLY, M. F. 1972. A new geological map for the basement rocks in the Eastern and South-Western Desert of Egypt. Annals of the Geological Survey of Egypt 2, 1-18. ABDEL-RAHMAN, A. M. & MARTIN, R. F. 1990a. The Mount EL-SHAZLY, E. M., HASHAD, A. H., SAYYAH, T. A. & BASSYUNI, F. Gharib A-type granite, Nubian shield: petrogenesis and role of metasomatism at the source. Contributions to Mineralogy and Petrology 104,173-83. A. 1973. Geochronology of Abu Swayel area, South Eastern Desert, Egypt. Egyptian Journal of Geology 17,1-18. ABDEL-RAHMAN, A. M. & MARTIN, R. F. 1990b. The Deloro FLECK, R. J., COLEMAN, R. G., CORNWALL, H. R., GREENWOOD, W. R., HADLEY, D. G., PRINZ, W. C , RATTLE, J. S. & anorogenic igneous complex, Madoc, Ontario. II. Evolution and Post-eruption metasomatism of the volcanic units. Canadian Mineralogist 28,267-85. SCHMIDT, D. L. 1976. Potassium-argon geochronology of the Arabian Shield, Kingdom of Saudi-Arabia. Geological Society ofAmerica Bulletin 87,9-21. ABDEL-RAHMAN, A. M. & MILLER, R. R. 1993. The Flowers River FLECK, R. J., GREENWOOD, W. R., ANDERSON, R. E. & SCHMIDT, anorogenic caldera complex, Labrador: stratigraphy and evolution. Geological Association of Canada-Mineralogical Association of Canada, Abstract with Programs 18, A1. ALDRICH, L. T., BROWN, G. F , HEDGE, C. & MARVIN, R. F. 1978. Geochronologic data for the Arabian shield. U.S. Geological Survey Open-file Report, 78-75. ANDERSON, A. T. 1980. Significance of hornblende in calc-alkaline andesites and basalts. American Mineralogist 65, 837-51. ARTH, J. G. 1976. Behaviour of trace elements during magmatic processes - a summary of theoretical models and their applications. Journal of Research of the U.S. Geological Survey 4,41-7. BARNES, S. J. & GORTON, M. P. 1984. Trace element analysis by neutron activation with a low flux reactor SLOWPOKE II.: http://journals.cambridge.org Downloaded: 20 Mar 2015 D. L. 1980. Rubidium-strontium geochronology and plate tectonic evolution of the southern part of the Arabian Shield. U.S. Geological Survey Professional paper 1131. FRANCIS, M. H. 1972. Geology of the basement complex in the North Eastern Desert between latitudes 27° 30' and 28° 00' N. Annals of the Geological Survey of Egypt 2,161 -80. FRANCIS, P. W., SPARKS, R. S. J., HAWKESWORTH, C. J., THORPE, R. S., PYLE, D. M., TAIT, S. R., MANTOVANI, M. S. & MCDERMOTT, F 1989. Petrology and geochemistry of volcanic rocks of the Cerro Galan caldera, northwest Argentina. Geological Magazine 126,515-47. GASS, I. G. 1981. Pan-African Upper Proterozoic. plate tectonics of the Arabian-Nubian Shield. In Precambrian Plate Tectonics (ed. A. Kroner), pp. 387-405. Amsterdam: Elsevier. IP address: 137.222.24.34 30 A.-F. M. ABDEL-RAHMAN GASS, I. G. 1982. Upper Proterozoic Pan-African calc-alkaline magmatism in north-eastern Africa and Arabia. In Andesites (ed. R. S. Thorpe), pp. 591-609. New York: John Wiley & Sons. GHANEM, M. 1972. Geology of the basement rocks north of latitude 28°N, Eastern Desert, Ras Gharib area. Annals of the Geological Survey of Egypt 2,181 -97. GILL, J. B. 1981. Orogenic Andesites and Plate Tectonics. New York: Springer-Verlag, 390 pp. MARSH, B. D. & CARMICHAEL, I. S. E. 1974. Benioff zone mag- GOKTEN, E. & FLOYD, P. A. 1987. Geochemistry and tectonic NEARY, C. R., GASS, I. G. & CAVANAGH, B. J. 1976. Granitic environment of the Sarkisla area volcanic rocks in central Anatolia, Turkey. Mineralogical Magazine 51, 553-9. association of northeastern Sudan. Geological Society of America Bulletin 87, 1501-12. GRAHAM, I. J. & HACKETT, W. R. 1987. Petrology of calc-alka- PATTERSON, D. B. & GRAHAM, I. J. 1988. Petrogenesis of line lavas from Ruapehu Volcano and related vents, Taupo Volcanic Zone, New Zealand. Journal of Petrology 28, 531-67. GREEN, T. H. 1980. Island arc and continent-building magmatism - a review of petrogenetic models based on experimental petrology and geochemistry. Tectonophysics 63, 367-85. andesitic lavas from Mangatepopo Valley and upper Tama Lake, Tangariro volcanic centre, New Zealand. Journal of Volcanology and Geothermal Research 35, 17-29. PEARCE, J. A. 1983. Role of sub-continental lithosphere in magma genesis at active continental margins. In Continental basalts and mantle xenoliths (eds C. J. Hawkesworth and M. J. Norry), pp. 230-49. Cheshire: Shiva. PEARCE, J. A. & CANN, J. R. 1973. Tectonic setting of basic volcanic rocks determined using trace element analyses. Earth and Planetary Science Letters 19,290-300. GREEN, T. H. & RINGWOOD, A. E. 1968. Genesis of the calc- alkaline igneous rock suite. Contributions to Mineralogy and Petrology 18, 163-74. GREENBERG, J. K. 1981. Characteristics and origin of Egyptian Younger Granites. Geological Society of America Bulletin 92, 749-840. HADLEY, D. G. 1973. Geology of the Sahl Al-Matran quadrangle, northwestern Hijaz, Kingdom of Saudi Arabia. Saudi Arabian Directorate General Mineral Resources Geological Map, Gm-6. HASHAD, A. H. 1980. Present status of geochronological data on the Egyptian basement complex. Institute of Applied Geology Bulletin, Jeddah 3,31-46. HlCKMOTT, D. D., SORENSEN, S. S. & ROGERS, P. S. Z. 1992. Metasomatism in a subduction complex: constraints from microanalysis of trace elements in minerals from garnet amphibolite from the Catalina Schist. Geology 20,347-50. IRVINE, T. N. & BARAGAR, W. R. A. 1971. A guide to the chemi- cal classification of the common volcanic rocks. Canadian Journal of Earth Sciences 8, 5 2 3 ^ 8 . JARRAR, G., WACHENDORF, H. & SAFFARINI, F. 1992. A Late Proterozoic bimodal volcanic/subvolcanic suite from Wadi Araba, southwest Jordan. Precambrian Research 56, 51-72. KOKELAAR, P. 1986. Petrology and geochemistry of the Rhobell volcanic complex: Amphibole dominated fractionation at an Early Ordovician arc volcano in north Wales. Journal of Petrology 27,887-914. matism. Journal of Geophysical Research 79, 1196-1206. MASUDA, A., NAKAMURA, N. & TANAKA, T. 1973. Fine structure of mutually normalized rare-earth patterns of chondrites. Geochimica Cosmochimica Acta 37,239-48. MILLER, C. F. & MITTLEFEHLDT, D. W. 1984. Extreme fractiona- tion in felsic magma chambers: a product of liquid-state diffusion or fractional crystallization? Earth and Planetary Science Letters 68, 151-8. PEARCE, J. A., HARRIS, N. B. W. & TINDLE, A. G. 1984. Trace element discrimination diagrams for the tectonic interpretation of granitic rocks. Journal of Petrology 25, 956-83. PERRY, F. V, BALDRIDGE, W. C, DEPAOLO, D. J. & SHAFIQULLAH, M. 1990. Evolution of a magmatic system during continental extension: the Mount Taylor volcanic field, New Mexico. Journal of Geophysical Research 95, 19327^8. PHILIPPOT, P. & SILVERSTONE, J. E. 1991. Trace-element-rich brines in eclogitic veins: Implications for fluid composition and transport during subduction. Contributions to Mineralogy and Petrology 106,417-30. RESSETAR, R. & MONRAD, J. R. 1983. Chemical composition and tectonic setting of the Dokhan volcanic formation, Eastern Desert, Egypt. Journal ofAfrican Earth Sciences 1, 103-12. RIES, A. C , SHACKLETON, R. M., GRAHAM, R. H. & FITCHES, W. R. 1983. Pan-African structures, ophiolites and melange in the Eastern Desert of Egypt: a traverse at 26°N. Journal of the Geological Society, London 140, 75-95. RINGWOOD, A. E. 1990. Slab-mantle interactions 3. Petrogenesis of intraplate magmas and structure of the upper mantle. Chemical Geology 82, 187-207. ROMICK, J. D., KAY, S. M. & KAY, R. W. 1992. The influence of KRONER, A., LINNEBACHER, P., STERN, R. J., REISCHMANN, T , MANTON, W. & HUSSEIN, I. M. 1991. Evolution of amphibole fractionation on the evolution of calc-alkaline andesite and dacite tephra from the central Aleutians, Alaska. Contributions to Mineralogy and Petrology 112, 101-18. Pan-African island arc assemblages in the southern Red ROOBOL, M. J., RAMSAY, C. R., JACKSON, N. J. & DARBYSHIRE, Sea Hills, Sudan, and in southwestern Arabia as exemplified by geochemistry and geochronology. In Proterozoic D. P. F. 1983. Late Proterozoic lavas of the central Arabian Crustal Evolution in the Late Proterozoic (eds R. J. Stern and W. R. Van Schmus), pp. 99-118. Precambrian Research 53. KRONER, A., REISCHMANN, T., WUST, H.-J. & RASHWAN, A. A. 1988. Is there any Pre-Pan-African > 950 Ma. Basement in the Eastern Desert of Egypt? In The Pan-African Belt of Northeast Africa and Adjacent areas (eds S. El-Gaby and R.O. Greiling), pp. 95-119. Braunschweig/Wiesbaden: Friedrich Vieweg & Sohn. LOPEZ-ESCOBAR, L. 1984. Petrology and chemistry of volcanic rocks of the southern Andes. In Andean Magmatism, Chemical and Isotopic Constraints (eds R. S. Harmon and B. A. Barreiro), pp. 47-71. Cheshire: Shiva. http://journals.cambridge.org Downloaded: 20 Mar 2015 Shield -evolution of an ancient volcanic arc system. Journal of the Geological Society, London 140, 185-202. Ross, C. S. & SMITH, R. L. 1961. Ash-flow tuffs, their origin, geologic relations and identification. U.S. Geological Survey Professional Paper 366,81 pp. SAUNDERS, A. D. & TARNEY, J. 1979. The geochemistry of basalts from a back-arc spreading centre in the East Scotia Sea. Geochimica Cosmochimica Acta 43,555-72. SCHMIDT, D. L. & BROWN, G. F. 1982. Chemistry of volcanic and plutonic rocks of the Precambrian shield, Kingdom of Saudi Arabia. Precambrian Research (Abstracts) 16, A37. SCHURMANN, H. M. 1966. The Pre-Cambrian Along the Gulf of Suez and the Northern Part of the Red Sea. Leiden: E. J. Brill, 186 pp. IP address: 137.222.24.34 31 Pan-African volcanism, Nubian shield SEGEV, A. 1987. The age of the latest Precambrian volcanism in southern Israel, northeastern Sinai and southwestern Jordan - a re-evaluation. Precambrian Research 36, 277-85. SPARKS, R. S. J., SELF, S. & WALKER, G. P. L. 1973. Products of ignimbrite eruptions. Geology 1, 115-18. STERN, R. J. 1981. Petrogenesis and tectonic setting of Late Precambrian ensimatic volcanic rocks, central eastern desert of Egypt. Precambrian Research 16, 195-230. STERN, R. J. & GOTTFRIED, D. 1986. Petrogenesis of a Late Precambrian 575-600 Ma bimodal suite in northeast Africa. Contributions to Mineralogy and Petrology 92, 492-501. STERN, R. J., GOTTFRIED, D. & HEDGE, C. E. 1984. Late Precambrian rifting and crustal evolution in the northeastern Desert of Egypt. Geology 12, 168-72. STERN, R. J. & HEDGE, C. E. 1985. Geochronologic and isotopic constraints on Late Precambrian crustal evolution in the Eastern Desert of Egypt. American Journal of Science 285, 97-127. STERN, R. J., SELLERS, G. & GOTTFRIED, D. 1988. Bimodal Dyke swarms in the North Eastern Desert of Egypt: Significance for the origin of Late Precambrian 'A-type' granites in Northern Afro-Arabia. In The Pan-African Belt of Northeast Africa and adjacent areas (eds S. El-Gaby and R. O. Greiling), pp. 147-79. Braunschweig/Wiesbaden: Friedrich Vieweg & Sohn. STUCKLESS, J. S., NAKOMO, I. T., WENNER, D. B. & VAN TRUMP, G. 1983. Geochemistry and Uranium favourability of the post orogenic granites of the northwestern Arabian shield, http://journals.cambridge.org Downloaded: 20 Mar 2015 Kingdom of Saudi Arabia. Institute of Applied Geology Bulletin, Jeddah 3, 195-209. SUN, S.-S. & MCDONOUGH, W. F. 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In Magmatism in the Ocean Basins (eds A. D. Saunders and M. J. Norry), pp. 313-45. Journal of the Geological Society, London, Special Publication no. 42. TAYLOR, S. R. & GORTON, M. P. 1977. Geochemical application of spark source mass spectrography. III. Element sensitivity, precision and accuracy. Geochimica Cosmochimica Acta 41,1375-80. VAIL, J. R. 1985. Pan African Late Precambrian tectonic terrains and the reconstruction of the Arabian-Nubian shield. Geology 13, 839^2. VAIL, J. R. 1990. Geochronology of the Sudan. Overseas Geology and Mineral Resources no. 66. London: British Geological Survey; HMSO, 58 pp. WATSON, E. B. 1979. Zircon saturation in felsic liquids: experimental results and applications to trace element geochemistry. Contributions to Mineralogy and Petrology 70, 407-19. WINCHESTER, J. A. & FLOYD, P. A. 1977. Geochemical discrimi- nation of different magma series and their differentiation products using immobile elements. Chemical Geology 20, 325-43. WRIGHT, T. L. & DOHERTY, P. C. 1970. A linear programming and least squares computer method for solving petrologic mixing problems. Geological Society of America Bulletin 81, 1995-2008. IP address: 137.222.24.34