Chapter 21

The Tambien Group, Northern Ethiopia (Tigre)

NATHAN R. MILLER1*, DOV AVIGAD2, ROBERT J. STERN3 & MICHAEL BEYTH4

1Department of Geological Sciences, Jackson School of Geosciences, University of Texas at Austin, Austin, TX 78712-0254, USA2Institute of Earth Sciences, Hebrew University of Jerusalem, Jerusalem 91904, Israel

3Geosciences Department, University of Texas at Dallas, Richardson, TX 75083-0688, USA4Geological Survey of Israel, Jerusalem, Jerusalem 95501, Israel

*Corresponding author (e-mail: nrmiller@mail.utexas.edu)

Abstract: The Tambien Group of northern Ethiopia (Tigre), with probable correlatives in Eritrea, is a 2–3-km-thick siliciclastic–carbonate succession that was deposited in an intra-oceanic arc platform setting within the southern Arabian–Nubian Shield (ANS)area (southern extension of the Nakfa Terrane) of the Mozambique Ocean. Its deposition occurred prior to ocean closure between con-verging fragments of East and West Gondwana and concomitant structural emergence of the East African Orogen (EAO). The TambienGroup is well exposed and best studied in the Mai Kenetal and Negash synclinoria, where litho- and chemostratigraphy (includingd13Ccarb, 87Sr/86Sr) provide the basis for a composite reference section. Two glaciogenic intervals have been suggested from exposureswithin the Didikama and Matheos Formation in the Negash Synclinorium. No reliable palaeomagnetic data exist to constrain the palaeo-latitude of Tambien Group deposition and the southern ANS, but palaeogeographic reconstructions and evaporite pseudomorphs in lowercarbonate units (Didikama Formation) imply low to intermediate latitudes (,458). Integration of available geochronological information(regional magmatism and detrital zircon) suggests c. 775–660 Ma as a plausible window constraining deposition of the prospectiveglacial intervals.

The Tambien Group appears to preserve a coherent chemostratigraphic framework that can be effectively subdivided according toshifts in d13Ccarb polarity [polarity intervals A (þ), B (–), C (þ), D (–)]. Slates underlying and interstratified with polarity interval Acarbonate preserve evidence of extreme chemical weathering that lessened prior to deposition of polarity interval B carbonate.Tambien Group carbonate units have sedimentological characteristics consistent with both shallow and deeper marine depositional set-tings. The lower prospective glacial interval lacks diagnostic sedimentological evidence of synglacial deposition, but is overlain by nega-tive d13C carbonate (polarity interval B) with sedimentological characteristics consistent with well-documented cap-carbonatesuccessions. The upper prospective glacial interval in the Negash Synclinorium (Matheos Diamictite) best exhibits characteristics con-sistent with glaciogenic deposition (matrix-supported polymictic clasts, possible dropstones, possible bullet-nosed and striated clasts). Incontrast to pericratonic rift margin settings that are common for Cryogenian glaciogenic deposits, palaeogeographic reconstructions forthe 775–660 Ma timeframe place northern Ethiopia within an intra-oceanic setting that was likely far removed from cratonic hinterlands.More work on Tambien Group sedimentology, geochronology and palaeogeography is required to better evaluate the extent and timing ofglacial conditions associated with the prospective glaciogenic intervals.

Supplementary material: Supplementary Table 21.1 of Tambien Group geochronological age constraints is available at http://www.geolsoc.org.uk/SUP18462.

The Tambien Group is exposed (Fig. 21.1) throughout portions ofnorthern Ethiopia (Tigre Province) and Eritrea (NNE extensionsfrom Figure 21.1; Bizen domain and Adobha Abi terrane ofBeyth et al. (2003) and De Souza Filho & Drury (1998)), ingreenschist-grade terranes comprising the southern portion of theArabian–Nubian Shield (ANS). It may be equivalent to similarcarbonate-rich units in NE Sudan (Bailateb Group; Stern et al.1994), SW Saudi Arabia (Hali Group; Greenwood et al. 1976),and possibly western Yemen (inferred from its Red Sea conjugateposition). Its deposition marks mainly marine siliciclastic and car-bonate sedimentation within the Mozambique Ocean, an extinctNeoproterozoic ocean basin destroyed during the late Neoprotero-zoic consolidation of Greater Gondwana with the emergence of theEast African Orogen (EAO) (Stern 1994). The Tambien Group isbest studied in northern Ethiopia from the Mai Kenetal (westernflank near 138550N, 388500E) and Negash (southern extent near138500N, 398370E) synclinoria, where much of the TambienGroup is continuously exposed, and these localities provide thesubstantial basis for regional litho- and chemostratigraphy.

Two possible glaciogenic intervals have been suggested withinthe Tambien Group (Beyth et al. 2003; Miller et al. 2003, 2009).The lower interval occurs as a greywacke-conglomerate intervalwithin slate of the lower Didikama Formation in the NegashSynclinorium (Fig. 21.2a, column D), and the top of this intervalmay correlate with the base of the Assem Limestone in the MaiKenetal Synclinorium (see discussion). The upper prospectiveglacial interval tops the Tambien Group as diamictite of the

Matheos Formation in the core of the Negash Synclinorium(Fig. 21.2a), and this interval may be equivalent to arkosic sand-stone and conglomerate of the Dugub Formation that similarlytops the Tambien Group in the western Shiraro area (Fig. 21.1,Avigad et al. 2007). The upper diamictite unit at Negash was orig-inally defined as the ‘Pebbly slate’ (Beyth 1972). It wassubsequently assigned within the Matheos Formation (Garland1980) and informally described as ‘Pebbly Slate (diamictite)’ inMiller et al. (2003), ‘Slate/Pebbly slate (Diamictite Slate)’ inAlene et al. (2006) and ‘Negash Diamictite’ in Avigadet al. (2007). Miller et al. (2009) subdivide the Matheos Formationinto three members, the youngest corresponding to the diamictitefacies. In consideration of a possible lower glaciogenic intervalin the Negash Synclinorium, the informal term ‘Matheos Diamic-tite’ is suggested for the upper prospective glacial interval. Inaddition to these northern Ethiopian localities, Eritrea hosts poss-ible glaciogenic units that have not been systematically studied.

The southern ANS is still very much a frontier region in need ofsystematic sedimentological, geochemical, and geochronologicalstudies of Neoproterozoic units. Much of what is known about theCryogenian Period for the region has been learned in only the pastdecade. The earliest suggestion of a Late Proterozoic glaciationwas by Bibolini (1920), who described faceted clasts (facce piane)within an unnamed pebbly mudstone-conglomerate unit (conglom-erati poligenici) in northern Eritrea. The basic Neoproterozoicstratigraphic framework for northern Ethiopia was established inregional mapping by Beyth (1972), including description of the

From: Arnaud, E., Halverson, G. P. & Shields-Zhou, G. (eds) The Geological Record of Neoproterozoic Glaciations. Geological Society, London,

Memoirs, 36, 263–276. 0435-4052/11/$15.00 # The Geological Society of London 2011. DOI: 10.1144/M36.21

units now posited to have glacial associations. Motivated by theSnowball Earth hypothesis, a number of reconnaissance studieshave since explored the depositional context of the TambienGroup (Beyth et al. 2003; Miller et al. 2003; Alene et al. 2006; seealso Stern et al. 2006). More comprehensive regional investigations,involving U–Pb zircon geochronology, chemical weatheringindices, and higher resolution sampling for integrated C- andSr-isotope stratigraphy, have further refined the age and range oflitho- and chemostratigraphic variations in the Tambien Group(Sifeta et al. 2005; Avigad et al. 2007; Miller et al. 2009).

Structural framework

(See additional region-specific structural information in the‘Glaciogenic deposits and associated strata’ section.) The greater

ANS consists of a patchwork of Neoproterozoic tectonostrati-graphic terranes, now bisected by the Oligocene and youngerRed Sea rift. ANS terranes include significant volumes ofjuvenile Neoproterozoic crust generated within spreading centers,arc and back-arc settings of the Mozambique Ocean (Stern 1994).Sutures between terranes, many with ophiolites and dated meta-morphic assemblages, document collisional deformation and helpconstrain the timing of terrane amalgamation. Many sutures haveappreciable strike–slip offsets, and these may broadly relate toc. 600 Ma escape tectonics (Burke & Sengor 1986), during whichANS terranes were progressively sandwiched by, and offsetbetween, obliquely converging Gondwana cratonic blocks (deSouza Filho & Drury 1998, and references therein). TambienGroup exposures in Tigre occur within the presumed southernextension of the Nakfa terrane of Eritrea. The Nakfa terrane isone of several suture-bounded low-grade volcano-sedimentary

Fig. 21.1. Location of key Tambien Group

exposures (unshaded units) (A, Shiraro area; B,

Mai Kenetal Synclinorium; C, Negash

Synclinorium; D, Samre area) within northern

Ethiopia (Tigre Province) modified from Miller

et al. (2009). Metavolcano–sedimentary block

boundaries (Tadesse et al. 1999) occur only

within the Tsaliet Group, and are inferred to

represent accreted arc terranes or slivers in a

supra-subduction zone setting. Stars show

geochronological localities for syn-tectonic

intrusives (white), post-orogenic intrusives

(black) and detrital zircons (white stars with black

dots) discussed in the text and shown in

Figure 21.5 (also in Supplementary Table 21.1).

Prospectively equivalent upper Tambien Group

strata of the Gulgula Group occur in western

Eritrea (arrow) just west of the Shiraro Area, as

well as in the Adobha Abi terrane and Bizen

Domain (arrows) to the north in Eritrea. Dashed

box east of Mai Kenetal outlines the Werii study

area of Sifeta et al. (2005). M. Alem marks the

western limb locality (Madahne Alem) within the

Negash Synclinorium bearing 774.7 + 4.8 Ma

zircons in slate c. 16 m below lowest Tambien

Group carbonate beds (Avigad et al. 2007). The

Negash Synclinorium is structurally bounded to

the east by the Atsbi Horst (AH ). Inset map

(below) shows the location of an unnamed pebbly

mudstone unit in northeastern Eritrea (Cecioni

1981) that could correlate to the Tambien Group.

N. R. MILLER ET AL.264

terranes comprising the greater Nubian Tokar superterrane(Kroner et al. 1991).

The Nakfa terrane supracrustal assemblage records a lowersequence of igneous rocks associated with arc magmatism in andaround the southern ANS sector of the Mozambique Ocean, asexemplified by calc-alkaline plutons and associated metavolcanicrocks of the Tsaliet Group. In western Tigre, the Tsaliet Group maybe preserved as a series of accreted volcanic arc terranes within asuprasubduction setting (Fig. 21.1; volcano-sedimentary blocks ofTadesse et al. 2000). Cessation of Tsaliet arc volcanism was fol-lowed by deposition of weathered arc detritus (Sifeta et al. 2005)and in turn increasing proportions of carbonate sediments, asexemplified by the Tambien Group. The contact between theTsaliet and Tambien Group is poorly understood, ranging fromseemingly conformable and gradational to fault-bounded (Sifeta

et al. 2005). An unconformable basal contact also was postulatedfrom considerable lateral variations observed among basalTambien metasediments (Arkin et al. 1971; Beyth 1972; Tadesse1999). Tambien Group deposition, including the two prospectiveglacial intervals, is thought to have occurred in an intra-oceanicplatform setting (above a substratum of consolidated arc terranes)following the main phase of Tsaliet arc magmatism in the region(c. 780 + 30 Ma; Avigad et al. 2007). The extent to whichTsaliet arc subterranes were fully accreted prior to TambienGroup deposition is uncertain, and some syndepositional reliefdifferentiation related to ongoing shortening and/or extension ispossible (Miller et al. 2009).

The entire Nakfa basement complex was deformed after depo-sition of the Tambien Group in conjunction with closure of theMozambique Ocean between fragments of east and west Gondwana,

Mai Kenetal Synform

Mai Kenetal LS~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

~~~~~~~~~~~~~~~

unconformity?

paraconformity?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Tsedia Slate

Assem LS

Werii Slate

C (++)

C (+)

B (-)

A (+?)

D (-)

C (++)

B (-)

A (+)

Tsaliet Metavolcanics

Negash Synform Afro-Arabian Peneplain

Matheos Fm Diamictite MbrTransition MbrBlack LS Mbr

Didikama Fm Upper (dol>slate)

Lower (slate >dol)GCI

Lower Slate (Atsbi Horst)

Tsaliet Metavolcanics

1

2

? ? ?

? ? ?

3

Mat

heos

Fm

Adigrat Ss

Algal* Ls

Shiraro Cgl

Arkosic ss

Mentebtab Ss

Bedded lsMica dolomite

Purple slate w/grnreduction spots

TsalietMetavolcanics

TsalietMetavolcanics

TsalietMetavolcanics

Undiff.Metavolc.

Adigrat SsPz & Mz

Seds Enticho Ss

Mai Kenetal Ls (200m)(Tselim Imni Ls)

Tsedia Slate (500m)(subdivided into upper Bilato and lower Logmiti Slates)

Assem LS (300m)(Filafil Ls)

Werii Slate (1000m)(Segali Slate)

Enticho Ss

Pebbly slate (200m)

Black detrital ls (300m)

Mica dolomite & slate (700m)

Slate & dolomite (200m)

Purple slate w/grnreduction spots; greywacke (300m)

Tsaliet Group

Enticho Ss

Diamictite Mbr

Transitional Mbr

Black Ls Mbr

Dolomite > Slate

Slate > Dolomite

Lower Slate(Atsbi Horst)D

idik

ama

Fm

Shira

ro G

roup

Enda

Zeb

i Fm

Dug

ub F

m -

- - M

ente

btab

Fm

Beyth 1972 (Tadesse 1999) Beyth 1972Beyth 1972

Tadesse 1999

Miller et al. 2009(Garland 1980)

C.Negash

Synclinorium

D.Negash

Synclinorium

A.Shiraro Plains

Region

B.Mai Kenetal

Synclinorium

?

?

?

Relationto syn-& post-

tectonicintrusives

----

Tam

bien

Gro

up --

--

Did

ikam

a Fm

Post

-oro

gen

ic g

ran

ito

ids

Syn-

?

? ? ?

GCI ?

d13C carb

Polarity

(a)

(b)

Fig. 21.2. (a) Lithostratigraphic subdivision of the Tambien Group in previous work. Prospective glacial intervals in columns A, C and D are superimposed from Beyth

et al. (2003) and Miller et al. (2009). The upper Algal Limestone unit in the Shiraro plains region (column A), originally termed the ‘Algal, stromatoporoidea (?)

limestone’ by Beyth (1972) was not mapped in the Axum Sheet (Tadesse 1999) and may occur in the Badme region (north of Shiraro, Fig. 21.1). (b) Chemo- and

lithostratigraphic correlations between the Mai Kenetal and Negash synclinoria used to create a composite Tambien Group stratigraphic section (see discussion).

Stratigraphic intervals in each section are alphabetized (A–D) according to the succession of associated d13Ccarb polarity (þ or –) intervals. See Figure 21.3 for associated

chemostratigraphic characteristics (circled data sets: d13CTOC, 87Sr/86Sr, [Sr] and chemical weathering index) that support the correlations. Numbers 1–3 correspond to

the different segments used in constructing the composite stratigraphy of Figure 21.6. Abbreviations (A&B): Fm, formation; GCI, Greywacke–Conglomerate Interval,

the lower prospective glacial interval in the Negash Synclinorium; grn, green; Ls, limestone; Mbr, member; Metavolc, metavolcanics; Mz, Mesozoic; Pz, Palaeozoic;

Ss, sandstone; Undiff, undifferentiated. (Modified from Miller et al. 2009.)

THE TAMBIEN GROUP, NORTHERN ETHIOPIA 265

and the concomitant structural emergence of the EAO, beginningc. 630 Ma. As a consequence, NNE-trending upright/overturnedfolds and faults dominate the regional structural grain, with TambienGroup exposures best preserved in synclinoria and grabens. Inter-nal minor or secondary folds and faults complicate measured sec-tions in the Mai Kenetal and Negash synclinoria. Restoringshortening (�200%) associated with these collisional structuresgreatly extends the minimal aerial extent of the Tambien basin orbasins. The deformed Tsaliet-through-Tambien Group successionwas subsequently intruded by c. 610 Ma (‘Mareb’) granitoidsassociated with crustal thickening and differentiation of the matur-ing EAO. Whether or not, and for how long, Tambien Group depo-sition continued prior to emplacement of the ‘Mareb’ post-orogenicgranitoids is unknown. Despite these post-depositional orogenicprocesses, temperatures and pressures did not exceed low-greenschist grade metamorphism and well-preserved primary sedi-mentary structures are common in many Tambien Group carbonatesuccessions (Alene et al. 2006; Miller et al. 2009).

Extensive erosion of the EAO resulted in cutting of a widespreadregional unconformity prior to Cambro-Ordovician time (Avigadet al. 2005), herein termed the Afro-Arabian Peneplain (AAP).Uppermost Tambien Group exposures in the core of the NegashSynclinorium (Matheos Diamictite) and possibly equivalent silici-clastic deposits of the Dugub Formation in the western Shiraroarea (Fig. 21.1) are thought to be northern Ethiopia’s youngestpreserved Neoproterozoic metasediments below the AAP.

Stratigraphy

Beyth (1972) documented Tambien Group outcrops in the Shiraroarea, in several en echelon synclinoria to the east (i.e. Mai Kenetal,Tsedia, Chemit and Negash), and the Escarpment area east ofAdigrat (Fig. 21.1, study areas A–C). Apparent regional faciesvariations within the internal stratigraphy of the Tambien Groupwere described, and a preliminary correlation was proposed forthree areas: Shiraro (west), Mai Kenetal (centre) and Negash(east). Subsequent workers have modified the stratigraphy invarious ways as described below and shown in Figure 21.2a.Figure 21.2b shows a recent correlation scheme proposed for theTambien Group based on integrated litho- and chemostratigraphyof the Mai Kenetal and Negash synclinoria (Miller et al. 2009),which is further elaborated in the discussion section.

Above the Tsaliet Metavolcanics, Beyth (1972) assigned fourformational designations within the Tambien Group based on theMai Kenetal type section (in stratigraphic order): Werii Slate,Assem Limestone, Tsedia Slate and Mai Kenetal Limestone. Inthe Negash Synclinorium, Beyth defined informal facies (in strati-graphic order: greywacke and purple slate with green reductionspots, slate and dolomite, mica dolomite and slate, black detritallimestone, and pebbly slate), which he suggested might correlatewith the Mai Kenetal section (Fig 21.2a, columns C v. B).Dolomite-slate successions similar to those of the lower TambienGroup in the Negash Synclinorium were subsequently documentedin the Escarpment east of Adigrat, as well as in synclines in theShiraro area. Garland (1980) formalized Beyth’s Tsaliet Metavolca-nics as the Tsaliet Group, and subdivided the Negash TambienGroup succession (Fig. 21.2a, column D) into the lower DidikamaFormation (after similar dolomitic rocks in the Shiraro Area) andoverlying Matheos Formation (equivalent to Beyth’s informalblack detrital limestone and pebbly slate facies). In mapping ofthe Mai Kenetal synclinorium, Tadesse (1999) used local namesto address the possibility that some of Beyth’s original type areas(i.e. Werii Slate, Tsedia Slate) might reside in different volcano-sedimentary arc terranes (Fig. 21.1). The main difference betweenthe two lithostratigraphic schemes is differentiation of Beyth’sTsedia Slate into a lower Logmiti Slate and overlying Bilato Lime-stone and Slate (Fig. 21.2a, column B). Tadesse (1999) also formal-ized a new Neoproterozoic stratigraphic scheme for the Shiraro

area, splitting Beyth’s informal Tambien Group facies into thebasal Didikama Formation and overlying Shiraro Group(Fig. 21.2a, column A). Designated within the latter were threemetasiliciclastic formations (Enda Zebi, Dugub and Mentebtabformations), each with respective lateral facies variations.

In the eastern limb of the Negash Synclinorium, Beyth (1972)mapped the base of the Tambien Group as an interval of arkosicsandstone and conglomerate, above a possible tectonic contactwith the Astbi Horst (Fig. 21.1, Fig. 21.2a column C). Thiscoarse metasiliciclastic interval corresponds to the lower prospec-tive glacial interval, the Greywacke-Conglomerate interval (GCI)of the Didikama Formation. The basal contact of the GCI has notbeen studied in detail, but the underlying depositional sequenceinvolves a minimum of several hundred metres of conformablybedded fine-grained marine metasedimentary units (mainly slatewith episodic thin carbonate interbeds up to a few centimetresthick). This lower slate succession contrasts from coarse volcanicagglomerate and conglomerate of the Tsaliet Group exposed at thetop of the Atsbi Horst (Fig. 21.1), but the contact between theseunits has not been systematically mapped or studied. Milleret al. (2003, 2009) include this lower slate interval (below theGCI) within the Tambien Group, as the informal Lower Slatemember of the Didikama Formation (Fig. 21.2a, column D).Above the GCI, the Didikama Formation transitions into varie-gated slate that includes the first prominent dolomite beds (infor-mal Slate . Dolomite member) and thicker bedded dolomite atthe top of the formation (informal Dolomite . Slate member).The GCI and overlying purple slate (with green reduction spots;the ‘Purple Egg Slate’ of Beyth 1972), with estimated combinedthickness of �300 m, forms a distinctive dark (grey-purple)marker unit that can be traced laterally throughout much of thec. 30-km-long exposed eastern synclinorial limb.

The Didikama Formation is overlain sharply, and very likelyunconformably (Beyth 1972, p. 74; Garland 1980, p. 14), by dis-tinctive black limestone of the Matheos Formation. Miller et al.(2009) subdivide the Matheos Formation into three members.The basal Black Limestone Member is well laminated and partlydetrital (intraclasts), and transitions upward into non-calcareousslate (Transitional Member) and eventually pebbly slate (Diamic-tite Member).

Prospective glaciogenic deposits and associated strata

Lower prospective glaciogenic unit

Greywacke–Conglomerate Interval, lower Didikama Formation (LowerTambien Group). Beyth (1972) described the GCI within theeastern flank of the Negash Synclinorium as poorly sorted, light-grey to black greywacke, containing mostly medium grained topebble-sized, subrounded quartz, in a mud (slate) supportedmatrix (Fig. 21.2a, column C). Thin conglomeratic layers andripple marks (exposed in bedding planes) also occur within theunit. The unit fines upward into purple slate with distinctiveovate greenish-yellow patches, interpreted as reduction spots.This lower interval (c. 300 m thick, Beyth 1972) is overlain by var-iegated slate (orange and green), which in turn grades upward(with poor outcrop) into the lowest exposed dolomite beds of theDidikama Formation. Similar purple spotted slate is reportedbelow the Didikama Formation in the Shiraro area (Beyth 1972;Fig. 21.2a, column A). The GCI is underlain by well-layeredslate (with occasional sub-decimetre-thick carbonate interbedsand minor foliation) that crops out continuously down-section(eastward) for at least several hundred metres in the hangingwall flank of the Atsbi Horst (Fig. 21.1).

Upper Werri Slate (Lower Tambien Group). In the NW flank of theMai Kenetal Synclinorium, the Assem Limestone lies sharplyand conformably above the Werii Slate and shares

N. R. MILLER ET AL.266

sedimentological and isotopic characteristics of well-studied cap-carbonate sequences (i.e. abrupt basal lithological transition tohigh-energy carbonate facies, basal dolomite bed (c. 1 m thick)transitioning upward into limestone, consistently negatived13Ccarb, primary and/or early diagenetic (pre-compaction) pris-matic and radial fibrous cements (Miller et al. 2009, figs S2D–G; Hoffman et al. 2007)). The observation that many Cryogeniancarbonate successions with negative d13C compositions areaffiliated with glaciogenic intervals (e.g. Yoshioka et al. 2003;Halverson et al. 2005; Corsetti et al. 2007 and other arguments,see discussion) raises the possibility that the pre-Assem Limestonedepositional sequence could have a glacial affiliation. The under-lying Werri Slate is mainly finely laminated, well-foliated, non-calcareous slate that does not exhibit obvious glaciogenic sedimen-tological characteristics. However, the nature of the upper WerriSlate and its contact with the Assem Limestone has not beenstudied regionally and poorly sorted metasiliciclastic intervalsare documented, at least locally, in the Werri Slate. For example,Beyth (1972) reported occurrences of greywacke within theWerii Slate regionally, and encountered poorly sorted rhyoliticagglomerate with well-rounded metre-scale quartzitic fragmentsat the southern end of Mai Kenetal syncline (around 138470Nand 388510E). Greywacke is also observed near its gradationalcontact with Tsaliet Group (Alene et al. 2006).

Upper prospective glaciogenic unit

Diamictite Member, Matheos Formation (Upper Tambien Group). TheMatheos Formation, comprising the interior of the Negash Syncli-norium, records a conformable transition from black limestone todiamictite. Ranging from faintly layered (massive) to well-bedded,with near vertical dips, the c. 250-m-thick basal Black LimestoneMember (Fig. 21.2a, column D) forms a distinctive ridge thatrims the interior of the Negash Synclinorium. Individual bedsrange in thickness from decimetre- to metre-scale and are typicallyfinely laminated in thin section. Intraclastic grainstone intervals arecommon. Towards the synclinorial axis, limestone beds thin(becoming sub-decimetre thick) and interstratify with increasingproportions of light to dark grey phyllitic slate. The overlying c.100 m Transitional Member includes non-calcareous phylliticslate that passes upward into pebbly slate, without interbeddedcarbonate. The primary depositional fabric involves mainly fine(,1 to 3 cm scale) horizontal beds. Initial disparate sedimentaryclasts are sub-centimetre scale (compositions have not beenstudied). The overlying c. 200-m-thick Diamictite Memberoccupies the tightly folded core of the Negash Synclinorium andis distinguished by an overall upward increase in clast abundanceand size (typically ,10 cm, but up to 20 cm in diameter). Clastsare matrix-supported and the finer matrix retains horizontal layer-ing, with some lateral pinching and swelling, throughout themember. Syn-orogenic compression of the synclinal core hasdeformed the diamictite to varying extents. More pelitic parts ofthe diamictite display foliation and associated clasts are often some-what elongated with pressure shadows. Despite these superimposedfeatures, clasts that appear to preferentially deform underlyingmatrix laminae are relatively common. Diamictite clasts have sub-rounded, elongate and angular shapes (including some bullet-nosedclasts), and some planar clast surfaces may be striated (Miller et al.2003). Clast lithologies are polymictic, including felsic volcanicrocks, fine-grained black limestone and dolomite (including well-rounded clasts that retain primary sedimentary structures; e.g.oolite), low-grade semipelitic sediments, and rare volcanic con-glomerate consistent with the upper Tsaliet Group.

Dugub Formation, Shiraro Group (Upper Tambien Group). Conglo-merate and arkosic metasediments comprise youngest TambienGroup exposures in the lowland Shiraro area of western Tigre(Shiraro Group of Tadesse 1999; Fig. 21.2a, column A). These

rocks overlie carbonate units assigned to the Didikama For-mation (Tadesse 1999) and crop out within the Shiraro Block(Fig. 21.1), a graben-like depression that is fault-bounded tothe east against the Adi Hageray Block and extends westwardbeyond the studied area into Eritrea. The areally most extensiveunit is the Dugub Formation, which we summarize followingTadesse (1999) and our own studies 4.5 km ENE of the townof Shiraro (Fig. 21.1).

The Dugub Formation consists of weakly metamorphosed con-glomerate, arkosic sandstone and siltstone (Tadesse 1999), whichare regionally intercalated at different scales. Well-preservedprimary sedimentary structures are common as graded bedding,cross-lamination, ripple marks, and flame and slump structures.Elliptical (rounded to subrounded) and well-sorted metaconglome-rate clasts (�8 cm) comprise up to 40% of the rock volume. Clastcompositions include low-grade (Tsaliet-like) volcanic rocks(chlorite schist, tuffaceous metasediments), phyllite, granite,quartz pegmatite, quartzite and chert. Fine- to medium-grainedchlorite, muscovite, feldspar and quartz are notable componentsof the metaconglomerate groundmass. Carbonate is notablyabsent, but is reported (as scarce marble layers a few to tens ofmetres thick) in the underlying Enda Zebi Formation (Tadesse1999). Dugub metasandstone contains quartz pebbles (up to afew millimetres across) and grades laterally and vertically intonon-pebbly sandstone and siltstone. Cross bedding exhibits setheights ranging between 0.1 and 0.5 m; coset heights range up to1.5 m. Orientations of asymmetrical ripple marks and slumpstructures suggest west-southwestward transport and palaeoslopedirections (Tadesse 1999; Avigad et al. 2007).

Prospective glaciogenic deposits in Eritrea

The high-grade Arag terrane in NE Eritrea (c. 400 km NNW ofNegash) has a widespread unnamed pebbly mudstone unit, consist-ing of silicic clasts within an argillaceous-siliceous matrix (Verri1909; Bibolini 1920, 1921, 1922). The unit is reported as widelyexposed between 178N and 178450N along its north-northwesterlytrend (c. 90 km) and penetrated by post-orogenic granitoids (acidaplitic intrusions, Cecioni 1981), which have not been dated.Cecioni (1981) described the pebbly mudstone matrix as a hard,violet, clay-quartz cement, and the clasts as pebbles and bouldersof quartzite and silicified schist. Some of the small pebbles(,1 cm) are pitted and knotted; others are angular. Bibolini(1920, 1921) noted the occurrence of faceted clasts and postulateda glaciogenic origin. To our knowledge this is the earliest suggestionof a possible Late Proterozoic (‘algonkiano’) glaciation within theANS. Cecioni (1981) preferred a gravity flow origin for this unit,but recognized a possible glaciogenic association due to theoccurrence of faceted pebbles. This unit trends more or less alongstrike with Tambien Group exposures in northern Ethiopia andprospective Bizen Domain equivalents in southern Eritrea, but hasapparently not been further studied in relation to a glacialassociation.

The Neoproterozoic supracrustal succession of western Eritrea(just west of the Shiraro area in northern Ethiopia, Fig. 21.1) con-sists of volcano-sedimentary assemblages (Augaro Group) uncon-formably overlain by basin-fill metasediments of the GulgulaGroup (Teklay et al. 2003; Teklay 2006). The Gulgula successioninvolves greenschist-grade polymictic conglomerate, shale, phyllitewith interstratified marble lenses, in addition to carbonaceous sand-stone, quartz arenite and allodapic carbonate, which, like theTambien Group, are deformed as upright, recumbent, and isoclinalfolds. Teklay (2006) describes the polymictic conglomerates asclasts, both matrix- and clast-supported, ranging from a few centi-metres to a half-metre of granite, slate, phyllite and epidotite. TheGulgula Group merges southwards into the Shiraro block in north-ern Ethiopia, suggesting a composite Gulgula-Shiraro depositionalbasin and a lithostratigraphic affiliation with the Shiraro Group

THE TAMBIEN GROUP, NORTHERN ETHIOPIA 267

(upper Tambien Group). The origin of Gulgula Group polymictconglomerates, in addition to those described above in theShiraro Group (Dugub Formation) require further study to clarifywhether or not glacial processes were involved.

Boundary relations with overlying and underlying

non-glacial units

Lower prospective glaciogenic unit

Greywacke-Conglomerate Interval (GCI) of the Didikama Formation,Negash Synclinorium. Where examined in the eastern limb of theNegash synclinorium, the GCI occurs within the Didikama For-mation (Fig. 21.2a, columns C and D) above thick slate (LowerSlate Member) and below variegated (purple/green/orange)slate that, in turn, grades upward into lowest dolomite beds ofthe Didikama Formation (Slate . Dol Member). The continuityof the transition from the Lower Slate member to the GCI is uncer-tain due to minor folding and possible faulting associated with theeastward structural transition to the Atsbi Horst (Fig. 21.1),whereas the transition above the GCI appears to be continuousand conformable.

Upper Werri Slate, Mai Kenetal Synclinorium. Miller et al. (2009)make chemostratigraphic arguments that the equivalent of theGCI may occur in the Mai Kenetal Synclinorium below theAssem Limestone, either within the upper Werii Slate orbetween the Werii Slate and Assem Limestone as a paraconfor-mity. The abrupt Werii Slate-Assem Limestone contact can befollowed laterally in the field as well as traced (.13 km) insatellite imagery.

Upper prospective glaciogenic unit

Matheos Formation Diamictite Member, Negash Synclinorium. TheMatheos Formation Diamictite Member is mapped only withinthe folded core of the Negash Synclinorium. Within the MatheosFormation, it overlies a distinctive but gradational transitionfrom black limestone (Black Limestone Member) to non-calcareous slate (Transitional Member). Although a directcontact has not been mapped, the oldest overlying sediments areOrdovician Enticho Sandstone, the base of which marks the AAP.

Dugub Formation, Shiraro Area. The Dugub Formation in thewestern Shiraro Area (Fig. 21.1) occupies a grossly similarstratigraphic position within the Tambien Group that could beequivalent to the Matheos Diamictite, but boundary relations forthis low-lying unit are poorly known.

Chemostratigraphy

Chemostratigraphic data for Tambien Group exposures in Ethiopia(Alene et al. 1999, 2006; Miller et al. 2003; Sifeta et al. 2005) andlikely equivalents in Eritrea (Beyth et al. 2003) have becomeavailable in only the last decade (Fig. 21.3). For the most part, che-mostratigraphic surveys have been limited in terms of types of ana-lyses performed and stratigraphic sampling frequency within agiven lithological unit, with fewer than 30 sample intervalsconsidered in any one study (Fig. 21.3b). Miller et al. (2009) sub-stantially expanded the regional chemostratigraphic database,assessing 100 carbonate and 30 slate intervals from the Shiraro,Samre, Mai Kenetal and Negash regions. The integrated TambienGroup record now constitutes a significant chemostratigraphic dataset for evaluating Cryogenian marine secular variations (including71 stratigraphic intervals with 87Sr/86Sr measurements onleast-altered units), with the Mai Kenetal and Negash synclinoria

data comprising most (c. 80%) analyses (Fig. 21.3). This compi-lation reveals significant chemostratigraphic trends, from whichregional correlations are proposed (Fig. 21.2b, discussion).

Carbonate d13C compositions for the Tambien Group rangebetween –8 and þ8‰ and can be characterized as fluctuatingrepeatedly upsection between positive and negative polarity(Fig. 21.3). Low-grade Bizen Domain stromatolitic carbonatesin SE Eritrea, considered to be Tambien Group equivalents(Beyth et al. 2003), are among the most negative d13Ccarb valuesreported (range, –7.2 to –4.8‰; 3 cal, 5 dol) but sample preser-vation was not critically assessed. Highest d13Ccarb compositions,with maxima near þ7‰, derive from the Mai Kenetal Limestoneand Matheos Formation Black Limestone Member. The negatived13Ccarb compositions of the Assem Limestone (c. –1 to –4‰)are similar to those comprising the lower negative d13Ccarb intervalin the Negash Synclinorium (c. –1 to –3‰) (despite their contrast-ing mineralogies), as well as negative d13Ccarb intervals mapped asAssem Limestone in the Chemit (–4.4 to –3.0‰) and Tsedia(–4.5 to –0.7‰) synclinoria (Alene et al. 1999, 2006; Milleret al. 2009). Bizen Domain stromatolitic dolomites with negatived13Ccarb (Beyth et al. 2003) may correlate with the negatived13Ccarb interval of the Didikama Formation. The Negashcarbonate sequence is so far unique in recording two negatived13Ccarb excursions.

Carbonate d18O compositions range between 0 and –16‰, withdolomite typically enriched (by 2–5‰) relative to stratigraphi-cally proximal (or prospectively equivalent) limestone. Althoughsubstantial scatter is apparent, median compositions are mainlybetween –4 and –10‰, similar to other Cryogenian datasets(Fig. 21.3; Jacobsen & Kaufman 1999; Robb et al. 2004; Halver-son et al. 2005, supplementary information). Values more negativethan –11‰ may be below the modal range of Cryogenian samples(Kaufman et al. 1993), and could be particularly altered. The MaiKenetal Limestone and Matheos Formation Black Limestone(units considered best preserved for d13Ccarb and 87Sr/86Sr, seediscussion) have median compositions between –7.6 and–3.5‰, with Matheos Black Limestones underlying the transitionto diamictite deposition most enriched.

Carbon-isotopic compositions of organic matter (d13CTOC) inthe Tambien Group, principally from the Mai Kenetal andNegash synclinoria, are mainly in the range of –20 to –30‰,with extremely light values (e.g. , –40‰) suggesting contri-butions from methanogenic biomass (Fig. 21.3). Both MaiKenetal and Negash successions show similar stratigraphic enrich-ment trends in d13CTOC with pre-diamictite values in uppermostMatheos Formation limestones about 3‰ heavier than uppermostMai Kenetal Limestone exposures.

The Tambien Group has a large compilation of 87Sr/86Sr com-positions in studies by Miller et al. (2003, n ¼ 9; 2009, n ¼ 71)and Alene et al. (2006, n ¼ 5). Results obtained from the sameunits are complementary among the studies despite contrastsin sample processing approaches. Least-altered 87Sr/86Sr compo-sitions (normalized to SRM 987 ¼ 0.710240) are mainlybetween 0.7055 and 0.7068 (Fig. 21.3). The Mai Kenetal succes-sion shows a stratigraphic enrichment trend that compares with amore extensive enrichment trend in the Negash succession.The lower negative d13Ccarb interval at Negash has 87Sr/86Srcompositions (avg: 0.706178 + 40, n ¼ 2) within error of thosefor Assem Limestone (avg: 0.706175 + 14, n ¼ 14)(Miller et al. 2009). Negative d13Ccarb compositions in theAssem Limestone may initiate with 87Sr/86Sr compositions near0.70597 (Miller et al. 2009).

Tambien Group carbonate units have Sr compositions rangingup to about 4000 ppm, with pronounced stratigraphic enrichmentevident in both the Mai Kenetal and Negash synclinoria. Sr com-positions less than 1000 ppm typify limestone and dolostone inthe Assem Limestone and Didikama Formation (also DidikamaFormation in Shiraro and Samre areas), whereas much higher con-centrations averaging 2000–3000 ppm occur in the Mai Kenetal

N. R. MILLER ET AL.268

(a)

(b)

Fig

.21.3

.S

um

mar

yof

chem

ost

rati

gra

phic

dat

ase

tsfo

rth

eT

ambie

nG

roup.

(a)

Sta

tist

ical

sum

mar

ies

of

chem

ost

rati

gra

phic

(d13C

carb

,d

18O

carb

,d

13C

TO

C,

87S

r/86S

r,S

rco

nce

ntr

atio

nan

dch

emic

alw

eath

erin

gin

dex

)dat

a

for

Tam

bie

nG

roup

lith

ost

rati

gra

phic

unit

sby

loca

lity

and

inst

rati

gra

phic

ord

er.L

ette

rsA

–D

mar

kst

rati

gra

phic

unit

sof

alte

rnat

ingd

13C

carb

pola

rity

;co

rres

pondin

gci

rcle

ddat

ase

tsin

dic

ate

poss

ible

corr

elat

ions

bet

wee

nth

e

Mai

Ken

etal

and

Neg

ash

syncl

inori

a(b

lack

ver

tica

lbar

s)an

doth

erT

ambie

nG

roup

loca

liti

es(s

eedis

cuss

ion);

line

types

of

circ

led

dat

ase

tscl

arif

yver

tica

lco

rrel

atio

ns.

Bla

cktr

iangle

sin

dic

ate

pro

spec

tive

gla

ciogen

icin

terv

als.

Dat

aco

mpil

edfr

om

Mil

ler

etal.

(2009)

and

Ale

ne

etal.

(2006,

vei

nca

lcit

esa

mple

sex

cluded

)ar

ere

stri

cted

tosa

mple

sco

nsi

der

edle

ast

alte

red

for

87S

r/86S

r.C

hem

ical

Index

of

Wea

ther

ing

(CIA

or

PIA

)dat

aex

clude

sam

ple

sco

nta

inin

gli

thic

frag

men

tsor

carb

onat

em

iner

als.

(b)

Dis

trib

uti

on

of

chem

ost

rati

gra

phic

anal

yse

sfo

rea

chst

udy

use

dto

com

pil

eth

eupper

gra

ph

stat

isti

cal

sum

mar

ies.

The

his

togra

m(t

op-t

o-b

ott

om

)bar

ord

er

corr

esponds

toan

alysi

sty

pe,

asli

sted

from

left

tori

ght

under

the

‘Tota

lA

nal

yse

s/L

east

Alt

ered

Anal

yse

s’co

mpil

atio

n.

THE TAMBIEN GROUP, NORTHERN ETHIOPIA 269

Limestone and Matheos Formation Black Limestone Member.Intermediate Sr compositions (avg: 1198 + 91 ppm, n ¼ 2,Miller et al. 2009) occur in the Tsedia Slate in the Mai Kenetal Syn-clinorium. Similar transitional compositions are missing in theNegash Synclinorium and instead there is a sudden stratigraphicjump in Sr concentrations between the Didikama Formation dolo-mite and directly overlying Matheos Formation Black Limestone.A range of Sr concentrations have been measured for matrix car-bonate (980 + 712 ppm, n ¼ 11; Miller et al. 2003) and carbonateclasts in the Matheos Formation Diamictite Member (512 ppm,oolite cobble; Miller et al. 2009).

The petrological and chemical transition from largely metavol-canic and metasiliciclastic (slate) units in the Tsaliet Group andlower Tambien Group (Werri Slate) to predominant carbonatedeposition in the higher Tambien Group (Assem Limestone andDidikama Formation) was investigated by Sifeta et al. (2005) inthe Werii area east of the Mai Kenetal Synclinorium (Fig. 21.1).Metavolcanic rocks are sub-alkaline, with chemical fingerprintsthat are compatible with island arc (primitive or evolved) and/or MORB settings. Lower Werri Slate compositions indicate

derivations from comparable juvenile crustal sources. Associatedchemical weathering indices (Fig. 21.3), which proxy the degreeof weathering from fresh rock (50) to full clay conversion (100),are generally between 70 and 90 for the Chemical Index of Altera-tion (CIA, Nesbitt & Young 1982) and .75 for the PlagioclaseIndex of Alteration (PIA, Fedo et al. 1995).

Palaeolatitude and palaeogeography

There are no reliable palaeomagnetic constraints for TambienGroup deposition within the southern ANS. Strike–slip displace-ment and structural shortening associated with closure of theMozambique Ocean and formation of the EAO constitutesignificant challenges for early Cryogenian palaeogeographicreconstructions within the ANS.

Palaeogeographic reconstructions spanning the 775–660 Mainterval (Fig. 21.4a–c) generally place the ANS (as inferredoceanic arcs) within the greater Mozambique Ocean between sub-sequently flanking Gondwana fragments (e.g. between India and

After Meert 2003; Meert & Torsvik 2003 Collins & Pisarevsky 2005

++

++

+

++

++

++

++

+

+++

+

+

+

++

+

+++

++

+

++

++

++

++

++

++

+

++

+

+

+

+

++

+ ++

+

+

++

+

+ +

Subduction Zone

Southern limit of Early Palaeozoic sandstonemarking African-Arabian peneplain (AAP)

AmazonianCraton

CongoCraton

EastAntarctic

Shield

IndianShield

Pz-Mz Orogen

Pz Orogen

Mz-Cz Orogen

1000 km

KalahariCraton

SaharanMeta-craton

RP

SFMad

SL

ANS

W.

NS

Australia

East Gondwana

West Gondwana?

Tambien Group ( )c. 550 Ma

~ A

nta

rcti

c ~

EA

O

WestAfricanCration

Neoproterozoic Orogen

750 Ma

Mozambique

Ad

ola

Brazilia

no

Pacific

Adam

aste

r

Az

Au - M

a

Co

In

WA Am

RPSF

La

Ka

630 Ma

Brazilia

no

Mozambique

Adam

aste

r

Az

CoSF

Sah

WA

Am

RP

La

In Ka

Au - M

a

CoKa

M

RP

Au SSC

WA

Am

B

LEA

MozambiqueOcean

BrasilianoOcean

800 Ma

SF

EAO

Pacific

(a)

(d)

(b) (c)

InIn

Continents withc. 750 Ma

palaeomagneticdata

Postulatedsubduction

zone

Schematiccontinental

collision

AAP

AAP

Modified: Meert & Lieberman 2008

Fig. 21.4. Global palaeogeographic

Colour

online=colour

hardcopy

reconstructions for (a) 800 Ma, (b) 750 Ma, (c) 630 Ma and (d) 550 Ma, showing the inferred palaeogeographic location and

structural-tectonic setting of the ANS/Tambien Group (black star) during the Cryogenian Period. Prospective glaciogenic intervals in the Tambien Group were likely

deposited during the earlier Cryogenian (.630 Ma) prior to the emergence of the East African Orogen (EAO). (a, b) Rifting and break-up of Rodinia (c. 900–750 Ma)

was associated with sea-floor spreading, arc and back-arc basin formation, and terrane accretion in the Mozambique Ocean. (c) Accommodation space in the southern

ANS basin likely inverted by 630 Ma in response to closure of the Mozambique Ocean between converging elements of West and East Gondwana and emergence of the

EAO. Occurrence of undeformed post-orogenic intrusives of this age or older that puncture the deformed Neoproterozoic supracrustal sequence in Ethiopia and Eritrea,

suggests that the Tambien Group at 630 Ma was likely deformed and uplifted within the EAO. (d) Location of the Tambien Group at c. 550 Ma within the context of

Gondwana amalgamation and emergence of the Antarctic-EAO. The modern southern limit of Early Palaeozoic sandstone (superimposed from Avigad et al. 2005)

documents the minimal extent of the Afro-Arabian Peneplain (AAP) and extent of regional uplift and Cambro-Ordovician erosion associated with the emergent EAO

(Avigad et al. 2005). Reconstructions modified from (a) Meert 2003; Meert & Torskvik 2003; (b, c) Collins & Pisarevsky 2005; (d) Meert & Lieberman 2008; Grey et al.

2008. Abbreviations for cratons and continents: Am, Amazonia; AuMa, Australia/Mawson Block; Az, Azania; B, Baltica; Co, Congo/Tanzania/Bangweulu Block; EA,

East Antarctica; Ka, Kalahari Block; La/L, Laurentia; In, India; Mad, Madigascar; RP, Rio de la Plata; Sah, Saharan Metacraton; SC, South China; SF, Sao Francisco;

S, Siberia; WA, West Africa; Adola, Adamastor, Braziliano, Mozambique, Pacific, oceanic basins; Mz, Mesozoic, Cz, Cenozoic; Pz, Palaeozoic.

N. R. MILLER ET AL.270

Congo cratons). For example, Collins & Pisarevsky (2005) placeophiolite-bearing strata of the Adola Belt (southern Ethiopia) out-board of the Congo Craton (Fig. 21.4b). As the Adola Belt is gener-allyconsidered tomark thesouthernextentof theEAO, it is reasonableto infer a similar outboard location for the Tambien Group.

The earliest reliable regional palaeomagnetic data derive fromthe time of Gondwana amalgamation (Fig. 21.4d); late Cryogenian(593 + 15 Ma) Dokhan volcanics of Egypt (Davies et al. 1980;Wilde & Youssef 2000), interpreted to have a subtropical palaeo-latitude (P-lat: 20.6 + 5.08, A95: 10.08, Trindade & Macouin2007; but see Nairn et al. 1987). The Dokhan palaeopole hasbeen widely used in subsequent palaeogeographic reconstructions(e.g. Meert 2003; Meert & Torsvik 2003; Macouin et al. 2004;Trindade & Macouin 2007).

Low subtropical palaeolatitudes (9–138) were also reported forHuqf Supergroup units in Oman thought to represent depositionduring and after an upper Cryogenian (Fiq Formation) glaciationbefore 544 Ma (Kempf et al. 2000; Kilner et al. 2005), whenOman was likely amalgamating with the ANS (detrital zircondata in Rieu et al. 2007). Subsequent work (Rieu et al. 2006;Allen 2007, and references therein) demonstrates that MirbatGroup localities in both palaeomagnetic studies correspond to anolder Cryogenian (Ghubrah-Ayn Formations) glaciation, con-strained radiometrically to be ,722 Ma and ongoing at711.8 + 1.6 Ma. The fact that palaeolatitudes for Oman unitsassociated with both glacial intervals (between c. 722 and544 Ma) are similarly low, and also close to published c. 550 Mapalaeopoles from Gondwana, has yet to be fully reconciled withthe available geologic data (Allen 2007). However, even if lowpalaeolatitudes are confirmed, Oman had yet to accrete with theANS at the time of the older Cryogenian (Ghubrah-Ayn For-mations) glaciation, and therefore these constraints could not beprecisely extended to the Tambien Group.

To the extent that the ANS occupied a gross latitudinal rangesimilar to the Congo-Sao Francisco and/or East-Sahara cratonsduring the early Cryogenian (c. 750 Ma), as implied by variouspalaeogeographic reconstructions (e.g. Fig. 21.4a,b; Trindade &Macouin 2007, fig. 3a), the Tambien Group may have occupiedlow to intermediate latitudes (,458). Occurrence of evaporitepseudomorphs in lower Didikama Formation dolomite (interpretedas ,774.7 + 4.8 Ma, Miller et al. 2009) may support a subtropicalpalaeolatitude, as palaeomagnetically constrained evaporite basinsas old as 2.25 Ga have statistically significant volume-weightedconcentrations in the palaeo-subtropics (Evans 2006).

Geochronological constraints

The age of Tambien Group deposition is bracketed by magmatismassociated with formation of the underlying Nakfa basementcomplex (Tsaliet Group) and orogenic thickening and differen-tiation related to maturation of the EAO (‘Mareb’ granitoids).Affiliated magmatism is widely dated throughout the ANS (e.g.Meert 2003; Johnson & Kattan 2007, and references therein).Our geochronological review is restricted to southern ANSstudies in eastern Sudan (Kroner et al. 1991), Eritrea (Teklay1997; Teklay et al. 2001, 2003; Andersson et al. 2006), northernEthiopia (Tadesse et al. 1997, 2000; Miller et al. 2003; Asratet al. 2004; Avigad et al. 2007) and western Ethiopia (Ayalewet al. 1990) closest to Tambien Group exposures and its prospec-tive equivalents (Figs 21.1, 21.5, Supplementary Table 21.1).

Detrital zircon age spectra from mature Ordovician sand (EntichoSandstone) capping the AAP in Tigre are interpreted by Avigad et al.(2007) to reflect the magmatic history of the regional Neoprotero-zoic–Ordovician basement complex, as it was uplifted and erodedwithin the EAO. The Enticho Sandstone zircon age distribution(Fig. 21.5) reinforces that Neoproterozoic magmatism occurred intwo main episodes: (i) arc-related calc-alkaline magmatismc. 850–740 Ma (c. 780 Ma peak), largely related to petrogenesisof the Tsaliet Group, and (ii) post-orogenic magmatism c. 660–580 Ma (c. 630 Ma peak) after Tambien Group deposition.Additional geochronological constraints elaborated below andshown in Figure 21.5 further suggest that the lower and upperprospective glacial intervals within the Tambien Group were depos-ited during the lull between these two magmatic episodes, with�775to .660 Ma as a plausible depositional window (see Discussion).

Pre-Tambien Group magmatism

Deformed (syn-tectonic) plutons and metavolcanic rocks of theTsaliet Group (lower calc-alkaline magmatic components of theNakfa terrane) have been dated by a number of techniques (e.g.Pb/Pb zircon, Rb–Sr, Sm–Nd, CHIME zircon, U–Pb zircon,zircon evaporation) with varying associated uncertainties (Sup-plementary Table 21.1 – Geochronological Data). These agesprovide an indirect lower age limit for Tambien Group carbonatedeposition. The oldest rocks in this region include an862 + 6 Ma granite clast within the Gulgula Group (Fig. 21.1,Teklay et al. 2003), 854 + 3 Ma deformed volcanic rocks and

900

850

800

750

700

650

600

550

500

Post-orogenic plutons

Tsaliet Group syn-tectonic granites and metavolcanics

TambienGroupCarbonateDeposition

Ag

e (M

a)

Relative W-to-E position (Tigre)

E. Sudan

Eritrea

Eritrea

5 43

10

22

2

2

2

1

1

6 6 7

NegashHawzien

AzehoHawzien

MaiKenetal

Chila

Deset

Sibta Shire

Rama

Mareb R.

Enticho Ss (Ord)detrital zirconage distribution

1

BSS

Ton

.C

ryo

gen

ian

NEO

PRO

TERO

ZOIC

PHZ

Edia

cara

nC

Youngest detrital zircons in upper prospective glacial intervals

Zircons below 1st dolomite beds

89

9

E. Sudan

1

MadahneAlem

(Negash)

2

1

Dugub Fm(Shiraro)

MatheosDiamictite(Negash)

21

1. Avigad et al. (2007)2. Tadesse et al. (2000)3. Tadesse et al. (1999)4. Teklay (1997)5. Teklay et al. (2002)

6. Miller et al. (2003)7. Asrat et al. (2004)8. Teklay et al. (2001)9. Kröner et al. (1991)10. Teklay et al. (2003)

STUDY

Pb-Pb zircon

Rb-Sr

Sm-Nd

CHIME zircon

U-Pb zircon

Zircon evaporation

Youngest detrital zircon(SHRIMP U-Pb)

DATE TYPE/METHOD

Fig. 21.5. Radiometric age constraints

bearing on the age of the Tambien Group

within the magmatic evolution of the

southern ANS and EAO. Tsaliet arc

magmatic products and ‘Mareb’

post-orogenic pluton ages arranged

according to their relative west-to-east

positions in Tigre. Magmatic equivalents

for localities in Eritrea and Sudan (open

squares) are shown to the left. The Enticho

Sandstone detrital zircon age spectrum

(Avigad et al. 2007) shown at the left of the

figure derives from the Enticho area as

shown in Figure 21.1. The shaded

horizontal bar shows the age range (827+6

to 770 + 7 Ma) for the Bitter Springs stage

(BSS) negative d13Ccarb excursion of

Australia (Halverson et al. 2005, 2007; and

references therein). Note that this range

overlaps with the main phase of Tsaliet

syn-tectonic arc magmatism; see text for

discussion.

THE TAMBIEN GROUP, NORTHERN ETHIOPIA 271

811 + 11 Ma granite from Eritrea (Teklay 1997). In the Axumarea (western Tigre), syn-tectonic pluton ages range between806 + 21 Ma and 756 + 33 Ma (Tadesse et al. 2000). A784 + 14 Ma syn-tectonic granitoid occurs SSW of Hauzien(Fig. 21.1; Avigad et al. 2007). A conformable felsic intervalshortly (c. 17 m) below initial dolomite beds of the Didikama For-mation, in the western limb of the Negash Synclinorium (MadahneAlem), produced a zircon U–Pb date of 774.7 + 4.8 Ma (Avigadet al. 2007; Figs 21.4 and 21.5).

Post-Tambien Group magmatism

Post-orogenic ‘Mareb’ granitoid plutons (Fig. 21.1) penetrate theolder deformed Neoproterozoic complex (including the TambienGroup) throughout northern Ethiopia and Eritrea. In Tigre, agesfor these commonly rounded and undeformed intrusives rangefrom 613.4 + 0.9 Ma (Mai Kenetal Granite, Avigad et al.2007) to 545 + 24 Ma (Mareb Granite, Tadesse 1997). TheNegash Pluton, just west of the Negash syncline, is c. 606 Ma(606.0 + 0.9 Ma, Miller et al. 2003; 607 + 7 Ma, Asrat et al.2004). A small granitoid body puncturing the western Negashlimb (lower Didikama Formation) is interpreted to have a compar-able age, but did not render suitable zircons for geochronology(Miller et al. 2009). Ages of post-orogenic intrusives are somewhatolder in Eritrea (628 + 4 Ma, 622 + 1 Ma; Teklay et al. 2001) andSudan (SE of Tokar: 652 + 14 Ma, Kroner et al. 1991). The goodagreement between individually dated syn-tectonic and post-orogenic magmatic products with corresponding modal peaks inthe Enticho Sandstone zircon age distribution (Fig. 21.5), andthe lack of detrital zircons younger than c. 739.2 + 6.3 Ma ineither of the prospective glaciogenic units capping the TambienGroup (Dugub Fm, Matheos Diamictite) are strong evidence thatthe Tambien Group is older than EAO post-tectonic graniticrocks (Avigad et al. 2007). In the vicinity of Mai Kenetal andNegash, this constraint is c. 610 Ma, but its upper age is likelymuch older considering that regional deformation and crustalthickening must have preceded the ‘Mareb’ intrusives. Theoldest dated post-tectonic pluton in the region (652 + 14 Ma;Kroner et al. 1991) and Enticho sandstone detrital zircon record(initiation of the second magmatic phase at c. 660 Ma; Avigadet al. 2007) suggest that the Tambien Group may have been depos-ited prior to c. 660 Ma (Fig. 21.5).

Discussion

Regional chemostratigraphic context of the prospective

glacial intervals

The observation that many Tambien Group exposures in Tigreexhibit a similar succession of lithofacies that have comparable che-mostratigraphic characteristics suggests that the Tambien Group pre-serves a regionally coherent chemostratigraphic framework (Beyth1972; Alene et al. 2006; Miller et al. 2009). Tambien Group deposi-tional history, including the depositional context of the prospectiveglaciogenic intervals, can be effectively considered relative tostratigraphic changes in the polarity of associated d13Ccarb. Theserelationships are best demonstrated in the Mai Kenetal and Negashsynclinoria (Figs 21.2b, 21.3) but are consistent with the stratigraphicsuccession in other areas (e.g. Shiraro and Samre).

The carbonate successions in the Mai Kenetal and Negash syn-clinoria both initiate above thick marine slate sequences; however,the nature of this transition to carbonate deposition differs in eachlocality. In Mai Kenetal the transition from Werii Slate to theAssem Limestone is a sharp conformable or paraconformablecontact, the Assem Limestone beginning with and maintainingnegative d13Ccarb compositions. In Negash, faulting in the lowerportion of each synclinorial limb interrupts the stratigraphic

continuity of the slate-to-carbonate transition, leading to negatived13Ccarb compositions above the lower prospective glacial interval(GCI) in the eastern limb and positive d13Ccarb compositionsabove variegated slate in the western limb (see Miller et al. 2009for details). The available regional data suggest that TambienGroup carbonate deposition began with positive d13Ccarb compo-sitions (interval A). For Negash this interval (c. 250 m thick) inthe western limb may crop out within an east-verging thrustblock, whereas the interval in the eastern limb may be faultedout or yet unrecognized below the GCI. The nature and continuityof the transition from Tsaliet Group agglomerates to lower Didi-kama Formation slate and dolomite remains poorly understoodand a systematic stratigraphic transition from lower positive (inter-val A) to higher negative (interval B) d13Ccarb compositions has yetto be documented in a conformable sequence. Based on similarlithostratigraphic position and character (slate with negligible car-bonate content and advanced chemical weathering indices;Fig. 21.3), the informal lower slate member of the DidikamaFormation could be equivalent to the Werii Slate and these unitsare tentatively included in polarity interval A.

The lower prospective glacial interval in both sequencesunderlies a lower carbonate interval with negative d13Ccarb (intervalB) that is in turn overlain by a distinctive black limestone successionwith the most enriched d13Ccarb compositions of the entire TambienGroup (interval C; Fig. 21.2b). Polarity interval B and C units inboth localities have highly complementary d13CTOC, 87Sr/86Sr andSr compositions that support their regional correlation (Fig. 21.3).The lithological and chemostratigraphic transition from interval Bto C appears to be continuous in Mai Kenetal, whereas this transitionin Negash is abrupt and likely associated with an unconformity(Miller et al. 2009). Above the black limestone interval in theNegash Synclinorium is the second ‘upper’ negative d13Ccarb excur-sion (interval D) associated with a conformable transition to theupper prospective glacial unit (Matheos Diamictite).

Composite chemostratigraphic reference section

Figure 21.6 presents a composite chemostratigraphic referencesection for the Tambien Group based on the correlations betweenthe Mai Kenetal and Negash synclinoria shown in Figure 21.2band the reasoning above. The Mai Kenetal section above the WeriiSlate (Assem Limestone, Tsedia Slate and Mai Kenetal Limestone)essentially replaces the Negash sequence above the basal positived13Ccarb interval of the lower Didikama Formation and below(or within) the lower portion of the Matheos Formation BlackLimestone Member. This composite effectively replaces dolomitelithologies below a probable unconformity (Negash) with a continu-ous and conformable succession of limestone lithologies; these seemto show gradational chemostratigraphic trends into highly enrichedvalues that correlate well with the lower Matheos Formation BlackLimestone Member (Miller et al. 2009). The abrupt transitionfrom positive (polarity interval A) to negative (polarity intervalB) d13Ccarb compositions may be a structural contact, but wesuggest that the relative stratigraphic order of d13Ccarb polarity ismaintained in the composite. The 87Sr/86Sr evolution of Cryogen-ian seawater (albeit poorly delineated) involves a general trend ofincreasing 87Sr/86Sr (Halverson et al. 2007). That least alteredsamples from polarity interval A dolomite beds have lowest87Sr/86Sr compositions is consistent with early Cryogenian depo-sition within the Tambien Group carbonate platform. The evolutionof Tambien Group depositional environments is now consideredbased on the composite reference section (Fig. 21.6).

Palaeoenvironmental changes

The Tambien Group involves mainly a transition from weatheredTsaliet arc detritus (lower slate intervals; Sifeta et al. 2005) to

N. R. MILLER ET AL.272

predominant carbonate deposition in a marine arc-accretion plat-form setting. The lack of obvious ash beds, volcaniclastic intervalsand syn-tectonic intrusive rocks in the main carbonate succession(polarity interval B and higher) suggests deposition during a mag-matically quiet interval (Avigad et al. 2007), and the upwardwaning of slate deposition in favour of carbonate may representa phasing out of arc magmatic activity in the region.

Chemical weathering indices of slates in polarity interval Asuggest an overall stratigraphic trend of increasing chemical weath-ering of source areas to extreme levels, which lessened somewhatprior to deposition of polarity interval B (Fig. 21.6). The associationof extreme weathering indices with increasing 87Sr/86Sr is consist-ent with an interval of rapid chemical weathering of arc terranes inthe southern ANS portion of the Mozambique Ocean. The finegrained nature of slates and absence of shallow water indicatorssuggests a moderately deep-water depositional setting.

Above the lower slate intervals, the Tambien Group carbonatepile is broadly differentiable between lower carbonate units(polarity intervals A–B) suggestive of shallow marine deposi-tional settings and upper carbonate units (polarity interval C) sug-gestive of deeper marine environments. Based on the occurrenceof evaporite pseudomorphs, sheetcrack cements, microbialami-nates and stromatolites (Miller et al. 2009, fig. S5), carbonates

associated with polarity interval A are interpreted to have beendeposited in high alkalinity tidal flat and intertidal settings. Mod-erate energy shallow intertidal or subtidal settings are interpretedfor the Assem Limestone and upper Didikama Formation(polarity interval B) on the basis of prominent domal and interdi-gitate stromatolites, coarse rip-ups, and cross-bedded grainstones(Miller et al. 2009, figs S2, S3). The base of the Assem Lime-stone occurs abruptly above the Werii Slate as a metre-thickdolomite bed, with sedimentary features similar to thosedescribed for well-studied transgressive cap-carbonate sequences.Lower carbonate units of the Tambien Group (polarity intervalsA–B) have low TOC contents with highly variable d13CTOC,including light compositions, which together with the sedimen-tary characteristics are consistent with nearshore environments(Miller et al. 2009, figs 7B,C and 9).

Deeper subtidal environments with lower (but still variable)energy levels and ubiquitous micrite are interpreted for uppercarbonate units of the Tambien Group (polarity interval C).The transition appears to begin at the base of the Tsedia Slate inMai Kenetal (equivalent to the proposed unconformity betweenthe Didikama and Matheos Formations in Negash), and this bound-ary could mark the base of a transgressive system followingregression. Both the Mai Kenetal Limestone and the Matheos

d18Ocarb d13CTOC

Stra

tig

rap

hic

Hei

gh

t (m

)

87Sr/86Sr

d13Ccarb Sr (ppm) TOC

0.705 0.706 0.707 -15 -10 -5 -45 -35 -250

-8

2000

1800

1600

1400

1200

1000

800

600

400

200

0

80 0 0.0 0.1 1.02000 4000-4 4

1

2

3

774.7±4.7 Ma

EntichoSandstone

Tsaliet Gp(Meta-

volcanics)

Tam

bie

n G

rou

p

Low

erSl

ate

Ass

emLS

MK

LSTs

edia

Sla

teM

ath

eos

FmLo

gm

iti

Bla

ck L

SD

iam

.Tr

Bila

to

Evaporite pseudomorphs StromatolitesSheet crack cements Internal slump Probable dropstones Cap carbonate-like texturesCC

CC ?

MTS

Molar tooth structuresMTS

Slat

e>

Do

l ?

(a)

(b)

(c)

(d) Interval devoid of bedded carbonateInterval devoid of bedded carbonate

Extreme weathering indices (PIA: 92-99)Extreme weathering indices (PIA: 92-99)

mat

rix

oolitecobble

Fig. 21.6. Composite chemostratigraphic reference sequence for the Tambien Group based on the Mai Kenetal and Negash synclinoria. Segments 1–3 denote the

basis for constructing this composite sequence (cf. Fig. 21.2b) and the black triangles denote the prospective glacial intervals. The composite sequence is differentiable by

four changes (a–d, separated by horizontal dashed lines) in the polarity of associated d13Ccarb and consists of limestone units except for segment 1 (dolomite comprising

the lower portion of the Didikama Formation in the western limb of Negash). The lower horizontal shaded interval indicates the stratigraphic range of slates having

extreme chemical weathering indices (PIA, plagioclase index of alteration of Fedo et al. 1995). The upper horizontal shaded interval corresponds to the upper portion of

the Matheos Formation lacking bedded carbonate; the darker grey d13Ccarb field in the Matheos Diamictite shows the range for diamictite matrix carbonate (Miller et al.

2003), whereas the uppermost data point corresponds to an oolite cobble (Miller et al. 2009). Abbreviations: Ls, Limestone; Dol, dolomite; Tr, transitional,

Diam-Matheos diamictite; MK, Mai Kenetal.

THE TAMBIEN GROUP, NORTHERN ETHIOPIA 273

Formation Black Limestone are distinctive dark grey to black lime-stone units characterized by fine horizontal layering and high lateralcontinuity (Miller et al. 2009, figs S3 and S6). The lack of stroma-tolites and high-energy sedimentary structures (coarsely gradedbeds, cross-bedding) but occurrence of dark intraclastic intervalsand intraformational slumping may indicate slope depositionbelow storm wave base. These upper carbonate units are relativelyenriched in TOC that has less variable d13CTOC, consistent withmore open marine environments (Miller et al. 2009, fig. 7B,C).Upper carbonate units are also most enriched in d13Ccarb andd13CTOC, particularly in the upper part of the Matheos FormationBlack Limestone Member. This mutual stratigraphic enrichmentpattern, together with distinct stratigraphic increases in 87Sr/86Srand Sr concentrations, is consistent with high rates of organicmatter burial and related fractionation of the upper (photic)marine d13CDIC pool in the lead up to Matheos Diamictite depo-sition. Although d18Ocarb is the marine proxy most likely toundergo post-depositional alteration, it is notable that the intervalswith most enriched compositions (consistent with cryogenicsequestering of d16O) occur in association with the two prospectiveglacial intervals (Fig. 21.6). The possibility that d18O records fluc-tuations in ice volume is more likely for the upper prospectiveglacial interval because the lower prospective glacial interval isunderlain by dolomite, which is typically enriched by 2–4‰over coeval limestone (Jaffres et al. 2007, and references therein).

Timing of the prospective glacial intervals

Integration of available geochronological information suggest c.775–660 Ma as a plausible window for Tambien Group depositionbracketing the two prospective glacial intervals. This depositionalwindow is not definitive, but corresponds to the lull betweenTsaliet arc magmatism and later magmatism associated with matu-ration of the EAO, on the basis of regional stratigraphy and cross-cutting relationships. Direct dates on magmatic products from eachmagmatic episode vary locally within Tigre but concur with thedetrital zircon age distribution from the immediately overlyingOrdovician Enticho Sandstone (Fig. 21.5). Somewhat older arcmagmatism is indicated in Eritrea and eastern Sudan.

The 774.7 + 4.8 Ma date obtained from zircons recovered fromvariegated (tuffaceous?) slate, c. 17 m below the first positived13Ccarb (polarity interval A) dolomite beds of the Didikama For-mation, is a possible maximum age constraint for significant(bedded) carbonate deposition in the Tambien Group (Fig. 21.5).The interval was originally interpreted as either a volcaniclasticinterval or sill injected into a volcaniclastic interval. Subsequentpetrographic analysis of this interval revealed probable sedimen-tary textures, including rounded quartz grains and a large (3 cm)rounded quartzose clast. If this date holds, the lower prospectiveglacial interval occurring at least c. 250 m higher is likely to beappreciably younger. The upper prospective glacial interval isolder than deformation and later magmatism associated with theEAO. The oldest indication of EAO magmatic activity in thesouthern ANS is from a 652 + 14 Ma post-orogenic pluton ineastern Sudan (southern Red Sea Hills, c. 500 km NNW ofNegash; Kroner et al. 1991), whereas locally dated ‘Mareb’ gran-itoids in Tigre are c. 610 Ma (avg: 609.8 + 3.8 Ma, Mai Kenetal,Negash, Hauzien plutons).

The dated Red Sea Hills locality in eastern Sudan occursc. 50 km NW of the area of prospective glacial deposits (pebblymudstones with flattened clasts) in northern Eritrea (Fig. 21.1)(Cecioni 1981). If the pebbly mudstones are glaciogenic andcorrelative with the upper prospective glacial interval of theTambien Group, the Red Sea Hills age offers an importantminimum age constraint for the Tambien Group. The mostprecise age estimates on older (pre-pebbly mudstone) associatedarc metavolcanic and plutonic rocks in the Red Sea Hills area,obtained by single zircon 207Pb/206Pb methods (c. 870–

827 Ma), are substantially older than ages determined by Rb–Srwhole rock methods (c. 770–670 Ma). The younger ages havebeen interpreted to reflect isotopic disturbance due to regionaldeformation of the arc complex, possibly as it accreted with theAfrican continent (Kroner et al. 1991; Saharan Metacraton?).Structural deformation of the orogen may thus have begunbefore 670 Ma. As this deformation was post-depositional and pre-ceded the first post-orogenic intrusives (652 + 14 Ma) by tens ofmillions of years, the Tambien Group minimum age, and the ageof the upper prospective glacial interval, could be appreciablyolder than 652 + 14 Ma.

Evidence for glacial influence on sedimentation

The lower prospective glacial interval of the Tambien Group,as the GCI in Negash and a possibly equivalent interval or paracon-formity at the top of the Werri Slate is most controversial.Direct sedimentologic evidence of syn-glacial deposition (i.e.definitive dropstones, striated clasts) is so far lacking, with thepossible exception of matrix supported horizons in the GCI(Beyth 1972). On the other hand, both of these intervals are over-lain by negative d13Ccarb carbonates with similar chemostrati-graphic characteristics. Where the base of this carbonate intervalis well exposed (Assem Limestone, Mai Kenetal) it exhibits fea-tures consistent with well-documented Cryogenian post-glacialcap carbonates.

Alene et al. (2006) discounted a glacial association for theAssem Limestone, but did not document or sample the base ofthis unit above its abrupt contact with Werri Slate. They suggestedthe Assem Limestone may correspond to a non-glacial negatived13Ccarb interval possibly equivalent to the Bitter Springs stage(BSS) of Australia (as suggested by Halverson et al. (2005,2007) for similar negative d13Ccarb intervals within early Cryogen-ian sections from NW Canada (Little Dal Group) and Svalbard(Akademikerbreen Group)). The age for the BSS in Australiarequires interbasinal correlations that are difficult to confirm, andthe age range has a corresponding large degree of uncertaintybetween 827 + 6 Ma (Gairdner Dyke Swarm) and 777 + 7 Ma(Boucat Volcanics) among various studies (Halverson et al.2005, 2007). Although a correlation with the BSS is possible, thec. 827–770 Ma age range would correspond to a time of activearc magmatism in the southern ANS (Fig. 21.5). The Assem Lime-stone and higher Tambien Group succession lacks evidence ofactive magmatism (ash beds/slate intervals/syn-tectonic intru-sives) suggesting it was deposited after the c. 780 Ma acme ofarc magmatism in the region. The lower Negash negatived13Ccarb interval occurs several hundred metres above the varie-gated slate interval dated at 774.7 + 4.8 Ma (zircon). Thissuggests that the lower negative d13Ccarb interval could be appreci-ably younger than c. 775 Ma. As there are definitively youngerpost-glacial cap carbonates (e.g. above the c. 755 Ma Kaigas For-mation, southern Namibia, Gariep Belt, Hoffmann et al. 2006)we suggest that a lower prospective glacial association withinthe Tambien Group remains a viable interpretation. It is conceiva-ble that the glaciogenic association could be indirect. For example,the cap-carbonate features and negative d13Ccarb excursion couldbe products of enhanced upwelling associated with globalcooling and/or vigorous circulation associated with post-glacialclimate change in a setting that did not earlier accumulate glaciallytransported clastic sediments. Both of these intervals in the MaiKenetal and Negash synclinoria warrant additional study toevaluate glacial associations.

Of the proposed upper glaciogenic intervals for the TambienGroup, the Matheos Formation Diamictite Member best exhibitssedimentological characteristics consistent with glacially influ-enced deposition. In addition to matrix supported polymictclasts, clasts that appear to preferentially deform underlyingbedding are interpreted as ice-rafted debris. Some bullet-nosed

N. R. MILLER ET AL.274

and possibly striated clasts are consistent with mechanical abrasionduring glacial entrainment (Kuhn et al. 1993). This intervalis further significant because of its gradational contact with theunderlying Matheos Black Limestone Member, which records adecline in d13Ccarb and possibly also 87Sr/86Sr, prior to diamictitedeposition. If the latter is not the product of diagenetic alteration,the declining Sr-isotope compositions could signal a strongdecrease in continental weathering input relative to oceanichydrothermal input, as consistent with increasing ice cover.

The possibly equivalent Dugub Formation in the Shiraro areaoccupies a gross stratigraphic position similar to the MatheosDiamictite as well as similar detrital zircon age distributions(Avigad et al. 2007), but the underlying stratigraphy is poorlyknown. Palaeocurrent proxies suggest a north-northeasternsource area for clasts, which considering the prospective glacialintervals in northern and eastern Eritrea could be consistent witha regional phase of glaciation.

Tectonic and palaeogeographic setting

The intra-oceanic setting of the Tambien Group contrasts distinctlyfrom pericratonic rift margin settings that are often associated withCryogenian glaciogenic units. Later Cryogenian intervals are unli-kely to be directly recorded throughout much of the ANS becausethe emergence of the EAO (by 660 Ma) would likely have eitherdestroyed accommodation space or uplifted related strata tolevels planed by subsequent erosion. The EAO could have hostedalpine/continental glaciation, however, perhaps even instigatinginitial downcutting of the vast AAP (Fig. 21.4d; Stern et al. 2006).

The US–Israel Binational Science Foundation (grant no. 2002337) supported this

study. We thank K. Mehari, D. Kuster and T. Tadesse for field expertise and

assistance, A. Abraham (Chief Geologist, Geological Survey of Ethiopia) and

Mekele University for logistical support during field seasons, and M. Abdel-

Salem for helpful feedback on geological aspects of the ANS. B. Schilman and

A. Ayalon of the Geological Survey of Israel provided stable isotope expertise.

The manuscript was improved by thoughtful reviews from P. Johnson, M. Pope

and E. Arnaud. This represents a contribution of the IUGS- and UNESCO-

funded IGCP (International Geoscience Programme) Project #512.

References

Alene, M., Conti, A., Sacchi, R. & Zuppi, G. 1999. Stable isotope com-position (13C and 18O) of Neoproterozoic limestones and dolomitesfrom Tigre, North Ethiopia. Bollettino della Societa GeologicaItaliana, 118, 611–615.

Alene, M., Jenkin, G. R. T., Leng, M. J. & Darbyshire, D. P. 2006. TheTambien Group, Ethiopa: an early Cryogenian (ca. 800–735 Ma)Neoproterozoic sequence in the Arabian–Nubian Shield. Precam-brian Research, 149, 79–89.

Allen, P. A. 2007. The Huqf Supergroup of Oman: basin developmentand context for Neoproterozoic glaciation. Earth Science Reviews,84, 139–185.

Andersson, U. B., Ghebreab, W. & Teklay, W. 2006. Crustal evolutionand metamorphism in east-central Eritrea, south-east Arabian-NubianShield. Journal of African Earth Sciences, 44, 45–65.

Arkin, Y., Beyth, M., Dow, D. B., Levitte, D., Haile, T. & Hailu, T.1971. Geological map of Mekele Sheet area ND 37-11 Tigre Pro-vince. Imperial Ethiopian Government, Ministry of Mines, Geologi-cal Survey of Ethiopia, scale 1:250,000.

Asrat, A., Barbey, P., Ludden, J. N., Reisberg, L., Gleizes, G. &Ayalew, D. 2004. Petrology and isotope geochemistry of thePan-African Negash Pluton, northern Ethiopia: mafic-felsic magmainteractions during the construction of shallow-level calc-alkalineplutons. Journal of Petrology, 45, 1147–1179

Avigad, D., Sandler, A., Kolodner, K., Stern, R. J., McWilliams,M. O., Miller, N. & Beyth, M. 2005. Mass-production of Cambrianquartz-rich sandstone as a consequence of chemical weathering of

Pan-African orogens: implications for global environment. Earthand Planetary Science Letters, 240, 818–826.

Avigad, D., Stern, R. J., Beyth, M., Miller, N. & McWilliams, M.2007. Detrital zircon U–Pb geochronology of Cryogenian diamictitesand Lower Palaeozoic sandstone in Ethiopia (Tigrai): age constraintson Neoproterozoic glaciation and crustal evolution of the southernArabian–Nubian Shield. Precambrian Research, 154, 88–106.

Ayalew, T., Bell, K., Moore, J. M. & Parrish, R. R. 1990. U-Pb andRb-Sr geochronology of the Western Ethiopian Shields. GeologicalSociety of America Bulletin, 102, 1309–1316.

Beyth, M. 1972. The geology of central western Tigre, Ethiopia. PhDthesis, Bonn, University of Bonn.

Beyth, M., Avigad, D., Wetzel, H. U., Matthews, A. & Berhe, S. M.2003. Crustal exhumation and indications for snowball Earth in theEast African Orogen: North Ethiopia and East Eritrea. PrecambrianResearch, 123, 187–201.

Bibolini, A. 1920. Risultali preliminary delle osservazioni faite nelNord-est della Colona Eritrea. Asmara.

Bibolini, A. 1921. Sui conglomerati di Rore Babla e dei Monti Haggarin Colonia Eritrea. Bollettino della Societa Geologica Italiana, 40,169–176.

Bibolini, A. 1922. Contributions a l’etude de la geologi de l’Afriqueorientale Italienne. 13th International Geological Congress (1922,Brussels, Belgium), Title Comptes rendus de la XIIIe session, enBelgique, parts 1–3, 797–814.

Burke, K. & Sengor, C. 1986. Tectonic escape in the evolution ofcontinental crust. In: Reflection seismology – The continental crustAmerican Geophysical Union, Geodynamic series, 14, 41–53.

Cecioni, G. 1981. Precambrian pebbly mudstones in Eritrea, northeasternEthiopia. In: Hambrey, M. J. & Harland, W. B. (eds) Earth’s Pre-Pleistocene Glacial Record. Cambridge University Press, A24, 150.

Collins, A. S. & Pisarevsky, S. A. 2005. Amalgamating eastern Gond-wana: the evolution of the Circum-Indian Orogens. Earth-ScienceReviews, 71, 229–270.

Corsetti, F. A., Stewart, J. H. & Hagadorn, J. W. 2007. Neoprotero-zoic diamictite-cap carbonate succession and d13C chemostratigraphyfrom eastern Sonora, Mexico. Chemical Geology, 237, 129–142.

Davies, J., Nairn, A. E. M. & Ressetar, R. 1980. The palaeomagnetismof certain late Precambrian and early Palaeozoic rocks from the RedSea Hills, eastern desert, Egypt. Journal of Geophysical Research,85, 3699–3710.

De Souza Filho, C. R. & Drury, S. A. 1998. A Neoproterozoic supra-subduction terrane in northern Eritrea, NE Africa. Journal of theGeological Society, London, 155, 551–566.

Evans, D. 2006. Proterozoic low orbital obliquity and axial-dipolar geo-magnetic field from evaporite palaeolatitudes. Nature, 444, 51–55.

Fedo, C. M., Nesbitt, H. W. & Young, G. M. 1995. Unraveling theeffects of potassium metasomatism in sedimentary rocks and palaeo-sols, with implication for palaeoweathering conditions and prove-nance. Geology, 23, 921–924.

Garland, C. R. 1980. Geology of the Adigrat area. Ministry of MinesMemoir No. 1, 51, Addis Ababa, Ethiopia. 1:250,000 map.

Gray, D. R., Foster, D. A., Meert, J. G., Goscombe, B. D., Armstrong,R., Truow, R. A. J. & Passchier, C. W. 2008. A Damaran Perspec-tive on the Assembly of Southwestern Gondwana. Geological Society,London, Special Publications, 294, 257–278.

Greenwood, W. R., Hadley, D. G., Anderson, R. E., Fleck, R. J. &Schmidt, D. L. 1976. Late Proterozoic cratonization in southwesternSaudi Arabia. Philosophical Transactions of the Royal Society ofLondon, 280, 517–527.

Halverson, G. P., Hoffman, P. F., Schrag, D. P., Maloof, A. C. &Rice, A. H. N. 2005. Toward a Neoproterozoic composite carbon-isotope record. Geological Society of America Bulletin, 117,1181–1207.

Halverson, G. P., Dudas, F. O., Maloof, A. C. & Bowring, S. A. 2007.Evolution of the 87Sr/86Sr composition of Neoproterozoic seawater.Palaeogeography, Palaeoclimatology, Palaeoecology, 256,103–129.

Hoffman, P. F., Halverson, G. P., Domack, E. W., Husson, J. M.,Higgins, J. A. & Schrag, D. P. 2007. Are basal Ediacaran(635 Ma) post-glacial ‘cap dolostones’ diachronous? Earth andPlanetary Science Letters, 258, 114–131.

THE TAMBIEN GROUP, NORTHERN ETHIOPIA 275

Hoffmann, K.-H., Condon, D. J., Bowring, S. A., Prave, A. R. &Fallick, A. 2006. Lithostratigraphic, carbon (d13C) isotope andU–Pb zircon age constraints on early Neoproterozoic (ca. 755 Ma)glaciation in the Gariep Belt, southern Namibia, Snowball EarthConference, 16–21 July 2006, Ascona, Switzerland [abstract], 51.

Jacobsen, S. B. & Kaufman, A. J. 1999. The Sr, C and O isotopic evol-ution of Neoproterozoic seawater. Chemical Geology, 161, 37–57.

Jaffres, J. B. D., Shields, G. A. & Wallmann, K. 2007. The oxygenisotope evolution of seawater: a critical review of a long-standingcontroversy and an improved geological water cycle model for thepast 3.4 billion years. Earth-Science Reviews, 83, 83–122.

Johnson, P. R. & Kattan, F. H. 2007. Geochronologic dataset forPrecambrian rocks in the Arabian peninsula. Saudi GeologicalSurvey Open-File Report SGS-OF-2007-3, 21, tables.

Kaufman, A. J., Jacobsen, S. B. & Knoll, A. H. 1993. The Vendianrecord of C- and Sr-isotopic variations: implications for tectonicsand paleoclimate. Earth and Planetary Science Letters, 120,409–430.

Kaufman, A. J., Knoll, A. H. & Narbonne, G. M. 1997. Isotopes,ice ages, and terminal Proterozoic earth history. Proceedings of theNational Academy of Sciences USA, 94, 6600–6605.

Kempf, O., Kellerhals, P., Lowrie, W. & Matter, A. 2000. Palaeo-magnetic directions in late Precambrian glaciomarine sediments ofthe Mirbat Sandstone Formation, Oman. Earth and PlanetaryScience Letters, 175, 181–190.

Kilner, B., Conall, M. N. & Brasier, M. 2005. Low-latitude glaciationin the Neoproterozoic of Oman. Geology, 33, 413–416.

Kroner, A., Llinnebacher, P., Stern, R. J., Reischmann, T.,Manton, W. & Hussein, I. M. 1991. Evolution of the Pan-Africanisland arc assemblages in the southern Red Sea Hills, Sudan, and insouthwestern Arabia as exemplified by geochemistry and geochronol-ogy. Precambrian Research, 53, 99–118.

Kuhn, G., Melles, M., Ehrmann, W. U., Hambrey, M. J. & Schmiedl,G. 1993. Character of clasts in glaciomarine sediments as an indicatorof transport and depositional processes, Weddell and Lazarev Seas,Antarctica. Journal of Sedimentary Petrology, 63, 477–487.

Macouin, M., Besse, J., Ader, M., Gilder, S., Yang, Z., Sun, Z. &Agrinier, P. 2004. Combined palaeomagnetic and isotopic datafrom the Doushantuo carbonates, South China: implications for the‘snowball Earth’ hypothesis. Earth and Planetary Science Letters,224, 387–398

Meert, J. G. 2003. A synopsis of events related to the assembly of easternGondwana. Tectonophysics, 362, 1–40.

Meert, J. G. & Lieberman, B. S. 2008. The Neoproterozoic Assemblyof Gondwana and its relationship to the Ediacaran–Cambrianradiation. Gondwana Research, 14, 5–21.

Meert, J. G. & Torsvik, T. H. 2003. The making and unmaking of asupercontinent: Rodinia revisited. Tectonophysics, 375, 261–288.

Miller, N. R., Alene, M., Sacchi, R., Stern, R., Conti, A., Kroner, A.& Zuppi, G. 2003. Significance of the Tambien Group (Tigre,N. Ethiopia) for snowball Earth events in the Arabian–NubianShield. Precambrian Research, 121, 263–283.

Miller, N. R., Stern, R. J., Avigad, D., Beyth, M. & Schilman, B.2009. Neoproterozoic carbonate-slate sequences of the TambienGroup, N. Ethiopia (I): pre-‘Sturtian’ chemostratigraphy and regionalcorrelation. Precambrian Research, 170, 129–156.

Nairn, A. E. M., Perry, T. A., Ressetar, R. & Rogers, S. 1987. Apalaeomagnetic study of the Dokhan volcanic formation and youngergranites, eastern desert of Egypt. Journal of African Earth Sciences,6, 353–365.

Nesbitt, H. W. & Young, G. M. 1982. Early Proterozoic climates andplate motions inferred from major element geochemistry of lutites.Nature, 299, 715–717.

Rieu, R., Allen, P. A., Etienne, J. L., Cozzi, A. & Wiechert, U. 2006.A Neoproterozoic glacially influenced basin margin succession and‘atypical’ cap carbonate associated with bedrock palaeovalleys,Mirbat area, southern Oman. Basin Research, 18, 471–496.

Rieu, R., Allen, P. A., Cozzi, A., Kosler, J. & Bussy, F. 2007. Acomposite stratigraphy for the Neoproterozoic Huqf Supergroup ofOman: integrating new litho-, chemo- and chronolstratigraphic data

of the Mirbat area, southern Oman. Journal of the GeologicalSociety, London, 164, 997–1009.

Robb, L. J., Knoll, A. H., Plumb, K. A., Shields, G. A., Strauss, H. &Veizer, J. 2004. The Precambrian: the Archean and ProterozoicEons. In: Gradstein, F., Ogg, J. & Smith, A. G. (eds) AGeologic Time Scale 2004. Cambridge University Press, Cambridge,129–140.

Sifeta, K., Roser, B. P. & Kimura, J. I. 2005. Geochemistry, provenance,and tectonic setting of Neoproterozoic metavolcanic and metasedi-mentary units, Werri area, Northern Ethiopia. Journal of AfricanEarth Sciences, 41, 212–234.

Stern, R. J. 1994. Arc assembly and continental collision in the Neopro-terozoic East African Orogen: implications for the consolidation ofGondwanaland. Annual Reviews of Earth and Planetary Sciences,22, 319–351.

Stern, R. J., Kroner, A., Bender, R., Reischmann, T. & Dawoud,A. S. 1994. Precambrian basement around Wadi Halfa, Sudan: anew perspective on the evolution of the East Saharan Craton. Geolo-gische Rundschau, 83, 564–577.

Stern, R. J., Avigad, D., Miller, N. R. & Beyth, M. 2006.Geological Society of Africa Presidential Review: Evidence forthe Snowball Earth Hypothesis in the Arabian–Nubian Shield andthe East African Orogen. Journal of African Earth Sciences, 44,1–20.

Tadesse, T. 1997. The Geology of Axum Area (ND 37-6). Ethiopian Insti-tute of Geological Surveys, Addis Ababa (Memoir No. 9).

Tadesse, T. 1999. Axum sheet geological map. Geological Survey ofEthiopia, Addis Ababa, Ethiopia. 1:250,000 map.

Tadesse, T., Hoshino, M. & Sawada, Y. 1999. Geochemistry of low-grade metavolcanic rocks from the Pan African of the Axum area,northern Ethiopia. Precambrian Research, 99, 101–124.

Tadesse, T., Hoshino, M., Suzuki, K. & Izumi, S. 2000. Sm–Nd, Rb–Srand U–Pb zircon ages of syn and post-tectonic grantoids from theAxum area of northern Ethiopia. Journal of African Earth Sciences,30, 313–327.

Teklay, M. 1997. Petrology, geochemistry and geochronology of Neopro-terozoic magmatic arc rocks from Eritrea: implications for crustalevolution in the southern Nubian Shield. Department of Mines,Asmara, Eritrea, Memoir 1, 125.

Teklay, M. 2006. Neoproterozoic arc–back-arc system analog tomodern arc–back-arc systems: evidence from tholeiite–boniniteassociation, serpentinite mudflows and across-arc geochemicaltrends in Eritrea, southern Arabian–Nubian shield. PrecambrianResearch, 15, 81–92.

Teklay, M., Kroner, A. & Metzger, K. 2001. Geochemistry, geochro-nology and isotope geology of Nakfa intrusive rocks, northernEritrea: products of a tectonically thickened Neoproterozoic arccrust. Journal of African Earth Sciences, 33, 283–301.

Teklay, M., Kroner, A. & Mezger, K. 2002. Enrichment from plumeinteraction in the generation of Neoproterozoic arc rocks in northernEritrea: implications for crustal accretion in the southern Arabian-Nubian Shield. Chemical Geology, 184, 167–184.

Teklay, M., Haile, T., Kroner, A., Asmerom, Y. & Watson, J. 2003. Aback-arc palaeotectonic setting for the Augaro Neoproterozoic mag-matic rocks of western Eritrea. Gondwana Research, 6, 629–640.

Trindade, R. I. F. & Macouin, M. 2007. Palaeolatitude of glacial depos-its and palaeogeography of Neoproterozoic ice ages. Comptes RendusGeoscience, 339, 200–211.

Verri, P. 1909. Contributo allo studio geografio della Colonia Eritrea. Bol-lettino della Societa Geografica Italiana, 10, 251–320. Carta Geolo-gica 1:1,500,000.

Wilde, S. A. & Youssef, K. 2000. Significance of SHRIMP dating of theImperial porphyry and associated Dokhan volcanics, Gebel Dokhan,northeastern desert, Egypt. Journal of African Earth Science, 31,403–413.

Yoshioka, H., Asahara, Y., Tojo, B. & Kawakami, S. 2003.Systematic variations in C, O, and Sr isotopes and elementalconcentrations in Neoproterozoic carbonates in Namibia: impli-cations for a glacial to interglacial transition. Precambrian Research,124, 69–85.

N. R. MILLER ET AL.276