depositional sequence analysis applied to late proterozoic

Australian Journal of Earth Sciences (1988) 35, 59-71

Depositional sequence analysis applied to Late Proterozoic WilpenaGroup, Adelaide Geosyncline, South Australia

C. C. von der Borch,1 N. Christie-Blick2 and A. E. Grady1

1 Flinders University, Bedford Park, SA 5042, Australia.2 Lamont-Doherty Geological Observatory of Columbia University, Palisades, New York, USA.

The initial application of depositional sequence analysis to selected stratigraphic sections throughoutcropping Late Proterozoic strata of the Adelaide Geosyncline in South Australia has identified majordepositional sequences within the several-kilometre-thick Wilpena Group. Sharp facies shifts in verticalstratigraphic sections are proposed as actual sequence boundaries which, provided they are the result ofeustatic sea level variations, may be key elements for future attempts at inter-regional chronostratigraphiccorrelation.

Two major sequence boundaries are identified, one at the base of the Nuccaleena Formation (boundaryA) and a second at the top of the Brachina Subgroup (boundary B). These are attributed to significantbasinward shifts in coastal onlap resulting in subaerial exposure and at least localized erosion, followed ineach case by establishment of relatively deepwater environments. A somewhat different boundary(boundary C) is associated with an interval of diagenetic dolostone interbeds and is interpreted either as adownlap surface within a sequence, or as a combined deepwater sequence boundary and downlap surface.It may have developed during an episode of reduced sediment input in response to a period of maximumtransgression. Alternatively it may represent a hiatus at the termination of a depositional sequence, prior tosubsequent downlap or onlap of the succeeding sequence.

Boundary C lies a few metres below the stratigraphic level from which kilometre-deep canyons haveincised underlying sequences. These canyons, which are infilled by a complex succession of carbonatebreccias, conglomerates, sandstone and mudstone, may have been eroded in a submarine setting byturbidity currents. Such a model requires a significant increase in rate of eustatic sea level fall or a decreasein the rate of tectonic subsidence, in order to move the locus of coastal onlap to the vicinity of the shelf edge.If the cause was eustatic, evidence for it should be found at an equivalent sequence boundary in LateProterozoic basins remote from the Adelaide Geosyncline. Alternatively, the canyons may have beeneroded in a subaerial setting and infilled by coastal sediments during an ensuing period of relative sea levelrise. In this model a considerably greater drop in relative sea level is required, most likely related tolocalized tectonic uplift.

Key words: Adelaide Geosyncline, depositional sequence analysis, stratigraphy, Wilpena Group.

INTRODUCTION

Most correlation in Proterozoic strata has con-ventionally been lithostratigraphic rather thanchronostratigraphic in nature because of theobvious lack of biostratigraphic resolution, thesmall number of reliable radiometric ages andlimited palaeomagnetic data. Gross stratigraphicsubdivisions of the Proterozoic have been basedon obvious angular unconformities but, with theexception of such prominent breaks, Proterozoicsedimentary rocks have conventionally beenviewed as conformable. In most cases it has notbeen possible to detect lesser hiatuses in con-cordant strata, even where their presence hasbeen suspected from radiometric dating.

Such assumed conformity and continuity ofsedimentation is not borne out by detailedsedimentological and stratigraphic observations,especially in younger rocks for which better ageresolution is possible. Although strata com-monly appear to be concordant when viewedfrom a distance, at all scales they actually consistof hierarchically arranged layers varying frommillimetres to hundreds of metres in thickness.Each layer is bounded by stratal surfaces whichrepresent discontinuities in sedimentation attime scales ranging from less than 10~7 to morethan 109 years (Campbell 1967; Dott 1983). On ascale of tens to hundreds of metres, many sedi-mentary successions are composed of 'deposi-tional sequences' (Fig. 1), essentially hierarchical

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011

60 C. C. VON DER BORCH ET AL

LANOWARD|SEAWARD HIGH LOW

SAND/MUD-RICH FACIES

TIME LINES

SEQUENCE BOUNDARY

DOWNLAP SURFACE WITHIN SEQUENCE

Fig. 1 Diagrammatic illustration of three hypothetical depositional sequences, separated by sequence boundariesA, B and C. Eustatic sea level and relative coastal onlap curves (after Vail et al 1984) illustrate simplest of severalpossible causes of sequence boundary formation. Note downlap surface, and region X-Y which is a combineddownlap surface and sequence boundary; note also erosional incision (canyon) which relates to events associatedwith sequence boundary C. The degree of vertical exaggeration is extreme; e.g. for typical second-order sequences,vertical distance between A and B could be in kilometres, whereas the lateral extent of the sequence could measuretens or hundreds of kilometres.

packages of genetically related lithofacies,bounded by regional unconformities or theircorrelative conformities (Vail et al 1977, 1984).In Mesozoic and younger deposits from whichmost seismic reflection data are available, thesesurfaces are easily recognized on seismic profiles.Mitchum et al (1977) and Vail et al (1977) havetermed the surfaces sequence boundaries, andthey have been utilized to develop time-stratigraphy in sedimentary basins subject toexploration for petroleum (the seismic strati-graphic approach). Although sequence boun-daries usually represent hiatuses of variableduration, they have chronostratigraphic signific-ance because with few exceptions it is commonfor strata above an unconformity to be every-where younger than strata below it.

The fundamental premise of this paper is thatas in Phanerozoic successions, Proterozoic rocksshould consist of hierarchical depositionalsequences bounded by sequence boundaries,

broadly comparable to the patterns shown by thesimplified model in Fig. 1. The identification ofProterozoic sequence boundaries in outcropshould improve relative time-stratigraphy andupgrade depositional models. Sequence boun-daries commonly correspond with abrupt faciesdiscontinuities in stratigraphic sections, andmodels which assume continuous sedimentationacross such boundaries may be incorrect. Thesequence analysis technique utilized in this paperhas successfully been applied to the 1.89 GaRocknest Formation in northwestern Canada byGrotzinger(1986).

Here we present initial results of sequenceanalysis applied to the lower Wilpena Group ofthe Late Proterozoic Adelaide Geosyncline (vonder Borch 1980; Rutland et al 1981) in thenorthern Flinders Ranges of South Australia(Fig. 2). We set out to characterize some of themajor sequence boundaries as they appear inoutcrop and to document their sedimentology,

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011

DEPOSITIONAL SEQUENCE ANALYSIS, WILPENA GROUP, SA 61

but postpone discussion of regional variations ofthe boundaries and detailing of higher-ordersequences. The Wilpena Group has been selectedfor analysis because the approximatelyequivalent stratigraphic interval, bracketedbetween the Precambrian glaciations and thePrecambrian-Cambrian boundary, can readilybe identified on other continents and thereforecritically compared. The style of sedimentationand major cyclicity within this interval isremarkably similar throughout Australia, andbetween Australia and North America, raisingthe so far untested possibility of a eustatic controlof base level during sedimentation of some of thedepositional sequences. This opens the furtherpossibility of inter-regional chronostratigraphiccorrelation, and is one of the main reasons forinitiating the present analysis of sequenceboundaries. An important brief of a related on-going study on the Late Proterozoic and EarlyCambrian Brigham Group, Utah, USA(Christie-Blick & Levy 1985; Christie-Blick &von der Borch 1985), is to test this premise. TheWilpena Group also hosts kilometre-deeperosional canyons (von der Borch et al 1982,1985) which, although their origin is debatable,may be analogous to major incisions associatedwith prominent sequence boundaries beneathmany present-day passive margins (Vail et al1977).

SEQUENCES, SEQUENCE BOUNDARIESAND DOWNLAP SURFACES INOUTCROP

Sequences

A stratigraphic log across an individual deposi-tional sequence may show a variety ofstratigraphic patterns, depending upon the sec-tion location with respect to basin geometry. Atypical example for siliciclastic sediments couldbegin with shallow-water facies at the base(onlap) or might exhibit apparently rapid appear-ance of what may be interpreted as relativelydeepwater sediments superimposed at a sharpboundary over shallow-water or subaerial sedi-ments of the underlying sequence. In the case of asecond- or third-order sequence, relatively deep-water facies may then persist upsection for aninterval of tens to hundreds of metres, at whichpoint sedimentological evidence of shallowing

conditions may be encountered. Ultimately sucha sequence may be capped by regressivesediments in the form of hummocky cross-strati-fied sandstones overlain by mature, nearshoresandstones. These either may persist up to theoverlying sequence boundary, or could passupsection to coastal plain sediments beneath anoverlying sequence boundary.

Sequence boundaries

Sequence boundaries can be recognized in threeways: (1) from the termination of layering (Fig. 1)by onlap, downlap, toplap and erosional trun-cation (Vail et al 1977, 1984); (2) by biostrati-graphy; and (3) from facies discontinuities. Thefirst method is the one used in seismic strati-graphy, but regional stratal terminations aresubtle and difficult to observe in limited outcrop(Bjorlykke 1982). They can be detected inregional mapping, especially where the structureis relatively simple, such as in the Triassic andJurassic strata of the Colorado Plateau(Pipiringos & O'SuIli van 1978), and in his studiesof turbidites in the Eocene Hecho Group, in thesouth-central Pyrenees of Northern Spain, Mutti(1985) has successfully mapped sequenceboundaries on aerial photographs. Biostrati-graphy is one of the oldest methods of identifyinginterregional unconformities (Sloss 1963), but asstated above, is applicable only in a very broadway to Proterozoic rocks. The interpretation ofmost Proterozoic sequence boundaries, there-fore, must be based primarily on facies discon-tinuities, and this is the technique employed inthis paper.

Facies discontinuities of this type, viewed inoutcrop, vary from obvious to cryptic, dependingupon several factors. These include the degree ofshift of coastal onlap during sequence boundaryformation, presence or absence of erosionaltruncation of the underlying sequence, rates ofrelative sea level rise following sequenceboundary development, and the actual locationof the outcrop section with respect to basinpalaeogeography. At one end of the spectrum,therefore, an abrupt facies discontinuity may beobserved across such a boundary, commonlybetween nearshore marine to subaerial facies ofthe underlying sequence and what may be con-strued as relatively deepwater, low energy shales.In a more complex situation, small scale cyclicity

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


3O"2<7

138*30'139*00'

Mesozic & Cainozoic seds.

-5 Carbonates & Silici-clastics

<X

Pound Subgp.

|_z Wonoka Fm.; incl. Billy Springs£ Beds in Umberatana Syncline

Brachina & Bunyeroo Fms.,undiff.

Adelaide Geosyncline rocksbelow Wilpena Gp.; black areasindicate disrupted, partly intru~

sive Callanna Gp.

Pre-Adelaidean rocks

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


may be observed within an interval above such aboundary, ultimately grading upsection to moreuniform deeper water facies. If, on the otherhand, the transition represents a 'deepwater'sequence boundary, careful stratigraphic loggingand facies analysis may be required, and it mayeven be necessary to map such a boundary for aconsiderable distance along strike in order tocorrelate it with confidence with a more obvioustransition elsewhere.

Downlap surfaces

A downlap surface (Fig. 1) is a surface within asequence which commonly corresponds with thetime of maximum transgression (Vail et al 1984).Such a surface is a marine hiatus, and may beassociated in outcrop or drill core with a thin butcontinuous zone of lithified beds (hardgrounds)and concentrations of rare authigenic com-ponents such as glauconite or phosphorite.

SEQUENCE ANALYSIS OF WILPENAGROUP

General description

The generalized stratigraphy of the WilpenaGroup in the northern Flinders Ranges is illu-strated in Fig. 3. Two first-order coarsening-upward sequences have been proposed byPlummer (1978a) as constituting most of theGroup, the lower being the Brachina Subgroup(Nuccaleena and Brachina Formations and ABCRange Quartzite) and the upper extending fromthe base of the Bunyeroo Formation to the top ofthe Bonney Sandstone. In the present study thesefirst-order sequences have been partly sub-divided into second- or third-order sequences.

The Brachina Subgroup is a kilometre-thickshallowing-upward unit with a thin, persistentdolostone, the Nuccaleena Formation at its base.This grades upsection to olive-drab mudstone

Fig. 2 Locality map of the Late Proterozoic AdelaideGeosyncline and Stuart Shelf (lower right) and detailsof northern Flinders Ranges geology. Adapted from1:600 000 geological map compiled by Preiss (1983).Specific localities for boundaries A, B and C discussedin text are lettered L, M and N, respectively (cf. Figs 3,4). Note major canyons infilled by sediments ofWonoka Formation near L and M.

and immature sandstone, which in turn give wayto interbedded mudstone and hummocky cross-stratified sandstone, slumped mudstone andultimately to a mature nearshore-facies sand-rich unit, the ABC Range Quartzite.

Overlying the ABC Range Quartzite withsharp contact is the Bunyeroo Formation,dominantly a monotonous grey-red to blue-greylaminated to massive mudstone, approximately500 m thick. The Bunyeroo Formation typicallyhas basal sand-rich horizons, and is abruptlycapped by a thin (1-2 m) zone of concretionaryand intraclastic dolostones, the uppermost ofwhich has been selected by Gostin and Jenkins(1983) to define the base of the Wonoka Forma-tion. Several hundred metres of mixed silici-clastic and carbonate sediments comprising theWonoka Formation overlie the Bunyeroo For-mation. The Wonoka Formation hosts at leasttwo second- or third-order sequences, which arecurrently being analysed (di Bona & von derBorch 1985). Haines (1986a,b), from a regionalstudy of the formation in the central andsouthern Flinders Ranges, emphasizes the sig-nificance of hummocky cross-stratification incalcareous sandstones throughout much ofmiddle to upper Wonoka Formation.

A major interval of sandstone, the PoundSubgroup, overlies the Wonoka Formation,attaining a thickness in excess of 3 km in theArkaroola Syncline (Figs 2, 3). The interval iscomposed of the Bonney Sandstone andoverlying Rawnsley Quartzite. The cyclicalBonney Sandstone is characterized by felds-pathic sandstone and interbedded red-brownmudstone representing a complex of subtidal,intertidal and fluvial environments. The over-lying Rawnsley Quartzite contains only minormudstone interbeds. It embraces at least twothird- Or fourth-order depositional sequences(not shown on Fig. 3), which locally host theEdiacara Fauna (Gehling 1983; Jenkins et al1983).

Major canyon-like features of the WilpenaGroup have been eroded down in excess of 1 kminto the Bunyeroo Formation and Brachina Sub-group from a zone within the lower Wonoka For-mation (Fig. 3). One of these, 'Patsy SpringsCanyon' (von der Borch et al 1982), is exposed inthe axial part of the Angepena Syncline; a portionof a canyon, interpreted as meandering, crops outas a series of west-trending incisions along thewestern margin of the Arkaroola Syncline (Fig.

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


UMBERATANAGROUP (B)

®<Loc.N)

Fig. 3 Diagrammatic stratigraphy (not to exact scale) of Wilpena Group in northern Flinders Ranges. A, B and Crefer to major boundaries discussed in text, at localities L, M and N, respectively (cf. Figs 2 and 4); three additionalmajor sequence boundaries currently under study are indicated by unlettered arrows. Note stratigraphicrelationships of canyons, which are eroded down from a cryptic sequence boundary within lower WonokaFormation slightly above boundary C. Brachina Subgroup (see text) incorporates Nuccaleena and BrachinaFormations and ABC Range Quartzite.

2). These canyons are infilled by quartzite-pebbleconglomerate, carbonate breccia, quartzosesandstone and calcarenite, constituting a dis-tinctive facies of the lower Wonoka Formation.This is overlain by a siliciclastic mudstone,quartzose sandstone, calcarenite, calcareousmudstone and carbonate breccia.

Major sequence boundaries, lower WilpenaGroup

Stratigraphic and sedimentological aspects ofmajor sequence boundaries and a possibledownlap surface within the lower Wilpena

, Group are illustrated in Figs 3 and 4 (boundariesA, B and C) and are detailed below. Geographiclocations of the three sections containingboundaries A, B and C are shown as L, M and N,respectively, on Fig. 2. We emphasize that theboundaries we identify have previously beenrecognized as formation, member or groupboundaries, for the simple reason that sequence

boundaries commonly separate sediments ofdistinctly contrasting facies. However, dia-chronous facies boundaries (e.g. base ofhypothetical sandstone unit below sequenceboundary B, Fig. 1) are not sequence boundaries,despite the fact that they may be designated asformation boundaries. In this context, the base ofthe ABC Range Quartzite is not identified as asequence boundary in Figs 3 and 4, although it isa formal formation boundary.

SEQUENCE BOUNDARY A

Description

Sequence boundary A (Fig. 4) defines the base ofthe Wilpena Group. Variations on the themedescribed below have been noted in central andsouthern regions of the Adelaide Geosyncline,but the section (locality L, Fig. 2) is repre-sentative of the area selected for initial study anddisplays key elements of this particularboundary.

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


3-n

TILL ITE

Fig. 4 Details of stratigraphy across the three selected boundaries, A, B and C. The dolomitic unit above sequenceboundary A is the Nuccaleena Formation, which grades upsection via interbedded dolostones, dolomiticconcretions and shale into the dominantly shaly section composing much of the lower Brachina Subgroup. Thesandy unit immediately below sequence boundary B is the ABC Range Quartzite, which 'caps' the BrachinaSubgroup.

The Mount Curtis Tillite, at the top of theunderlying Umberatana Group, is a several-metre-thick polymict pebbly sandstone ofpossible glacial-marine origin. It is composed ofa variety of poorly sorted basement-derivedclasts of up to boulder size, in a sandy matrix.Overlying the Mount Curtis Tillite, or locally atrough cross-bedded feldspathic sandstone(Balparana Sandstone), with a sharp contact, is aseveral-metre-thick dolostone, the NuccaleenaFormation. A distinct disconformity is presentlocally between the base of the NuccaleenaFormation and the underlying Balparana Sand-stone or Mount Curtis Tillite. The basal fewcentimetres of the dolostone at locality L, over-lying the Mount Curtis Tillite, contain reworkedsand and gravel probably derived from theunderlying pebbly sandstone. The lowermost fewmetres of the Nuccaleena Formation consist of aflaggy pink dolomicrite which ranges in texturefrom massive to diffusely banded and possiblyalgal laminated. In nearby areas stromatolites,intraclast breccias and teepee structures havebeen reported (Plummer 1978b; Williams 1979).Towards its top the Nuccaleena Formationgrades into grey-green mudstone via inter-bedded thin dolomite layers and concretions and

siliciclastic mudstone. Interbeds of siliciclasticmudstones then increase in thickness at theexpense of layered and nodular dolostones of theupper Nuccaleena Formation. Above this levelthe dominantly siliciclastic Brachina Subgroupbegins what normally is a kilometre-thickupwards-coarsening succession. At locality L,subsequent canyon erosion related to WonokaFormation deposition has removed much of1 theSubgroup.

Interpretation

The lower portion of the Nuccaleena Formationin the study area is interpreted as peritidal inorigin. Dolostones of this type typically form insea-marginal saline and alkaline ephemeral lakesand associated sabkhas. These environments candevelop approximately at sea level immediatelylandward of the shoreline, where continentalgroundwaters become buoyed upwards over themore dense marine porewaters. Chemical evolu-tion of evaporitically modified groundwatersand admixed seawater provides appropriateconditions for genesis of'penecontemporaneous'microcrystalline dolostones and associatedsupratidal dolomitic crusts and teepees, a

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


modern example being the Coorong region ofSouth Australia (von der Borch & Lock 1979).The Nuccaleena Formation is most likely trans-gressive in origin and consequently a dia-chronous unit. The upper portion of the forma-tion may have been partly derived by reworkingof the lower portion, with transgressivemudstone of the overlying lower BrachinaSubgroup ultimately overwhelming the peritidalcarbonates to dominate the sequence.

The mappable boundary separating the MountCurtis tillite and the Nuccaleena Formation(Figs 3 and 4) was previously denned as a groupboundary (Thompson et al 1964). The location ofthe lowermost sequence boundary defined in thispaper is placed at this distinct facies break.

SEQUENCE BOUNDARY B

Description

The stratigraphic section is situated in a region(locality M, Fig. 2) where a thick succession ofgreen argillaceous sandstone and siltstone of theBrachina Subgroup is overlain by several centi-metres of a cross-bedded quartzite, which maycorrelate to the south with the ABC RangeQuartzite. The Bunyeroo Formation lies withsharp contact over this thin quartzite, and has atits immediate base a 20-30 cm thick cupriferousunit composed of trough cross-bedded andrippled coarse grained sandstone lenses, andinterbedded millimetre- to centimetre-thick redmudstone. Towards the top of this thin basal unitthe'sands decrease in grain size and are asso-ciated with Baser and wavy bedding. Severalcentimetres farther upsection the sand contentdiminishes and a monotonous massive to finelylaminated red-brown to green silicic mudstone,typical of the bulk of the Bunyeroo Formation,dominates the sequence. The basal contact of theBunyeroo Formation is likewise sharp in all areasof the Flinders Ranges, either directly over well-developed ABC Range Quartzite or, where this isabsent, argillaceous sandstones of the uppermostBrachina Subgroup. In the western Bayley Rangearea (Copley 1:250 000 geological sheet; Coats1973) typical Bunyeroo Formation mudstoneoverlies several metres of interbedded sandstoneand reddish-brown mudstone with intercalateddolomitic quartzite-pebble breccias, calcitepseudomorphs after discoidal gypsum, andcarbonate clast breccias. These are assigned to

the lowermost Bunyeroo Formation and overliemature sandstone of the ABC Range Quartzite.

Interpretation

The ABC Range Quartzite is interpreted as theregressive nearshore facies 'cap' of the lowermostmajor depositional sequence of the WilpenaGroup. Breccias of obviously pre-lithifiedquartzite pebbles, locally intercalated within thebasal Bunyeroo Formation, imply that indur-ation of the ABC Range Quartzite was well ad-vanced prior to Bunyeroo Formation deposition.The basal portion of the Bunyeroo Formation isbest interpreted as a transgressive nearshoredeposit, possibly of tidal origin and locally withevaporitic affinities. The sharp contact betweenthe ABC Range Quartzite and basal BunyerooFormation is interpreted as a sequence boundary(boundary B) at the base of a second majordepositional sequence within the WilpenaGroup.

SEQUENCE BOUNDARY OR DOWNLAPSURFACE C

Description

A sharp-based and locally channelled intraclasticdolostone, about 15 cm thick, occurs at localityN (Fig. 2), at the stratigraphic level of boundary C(Figs 3, 4). The dolostone separates metallic-greylaminated siliciclastic mudstone of the re-defined (Jenkins & Gostin 1983) upper BunyerooFormation from overlying siliciclastic mudstoneand flaggy calcarenite of the Wonoka Formation.Locally this dolostone exhibits several thin(about 5 cm) graded beds of mixed carbonate andsiliciclastic sandstone and mudstone withcarbonate intraclasts. Immediately below theintraclastic dolostone, the upper BunyerooFormation contains two thin (about 5-10 cm)concretionary beds of diagenetic dolomite,similar to the thin interbeds of dolomite presentat the same stratigraphic level in the BunyerooGorge area (Haines 1986a). Dolostones of thistype are widespread at this stratigraphic levelthroughout the central and northern AdelaideGeosyncline. In the region of section C, theseveral-hundred-metre-thick Bunyeroo For-mation below these dolostones is composedentirely of monotonously laminated mudstone.In the vicinity of St Ronan's Well (Copley

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


1:250 000 geological sheet; Coats 1973), the up-permost few metres of the Bunyeroo Formationhost several thin (10-20 mm) sandstones and athin intraclastic calcareous sandstone, inter-bedded with grey-red mudstone. This interval isoverlain abruptly by a succession of sharp-basedsiliciclastic sandstone with flute casts and inter-bedded grey-red mudstone. Hummocky cross-stratification is present within many of thesandstone beds. This unit is typical of the lowerWonoka Formation in the central and southernFlinders Ranges and correlates with WonokaFormation Unit 2 of Haines (1986a). Gostin andJenkins (1983) suggested that sandstones, nowdesignated as belonging to Unit 2 and interpretedas turbidites, are genetically related to the processwhich caused erosion of the canyons. Dolostonesof Wonoka Formation Unit 1 of Haines (1986a)are not represented at this locality.

The uppermost level of incision of the canyonsterminates at an interval which lies severalmetres above boundary C (Fig. 3). At PatsySprings Canyon (Fig. 2, near locality M), thecanyon shoulder lies a few metres above a thindolostone, which most likely correlates with theintraclastic dolostone defining the base of theWonoka Formation. The interval between thisintraclastic dolostone and the canyon shoulder iscomposed of 2-3 m of interbedded hummockycross-stratified sandstone and mudstone of thelower Wonoka Formation, which are correlatedin a lithostratigraphic sense with Unit 2 ofHaines (1986a). These sediments are overlain'bya second thin dolostone and a metre-thick pink,flaggy calcarenite, which can be traced into a pinkcalcarenite and carbonate breccia mantling thecanyon wall.

Interpretation

The monotonous fine grained and commonlyfinely laminated nature of Bunyeroo Formationmudstones implies relatively quiet deepwaterconditions. Lack of associated sandy material,except locally near the base and top, may imply abypass-slope setting. The thin sandy interbeds insome localities near the top of the formationsuggest the beginning of a subtle upwards-shoaling event which, in the region of accessibleoutcrop, clearly did not progress to subaerialconditions.

Concretionary dolostones at the top of theBunyeroo Formation are interpreted as having

formed during diagenesis, analogous to dolomiteformation in organic-rich continental-marginsediments described by Baker and Burns (1985).Dolomites of this type can form where organiccarbon content exceeds 0.5 wt% and wheresedimentation rates are relatively low. If such amodel applies to the uppermost BunyerooFormation, then a condensed interval is impliedfor this portion of the section, perhaps related toa downlap surface or a combined sequenceboundary and downlap surface (e.g. X-Y; Fig. 1).A surface of this nature would have experienced aprolonged period of exposure on the seafloor inrelatively deep water, prior to burial. Subsequentburial may have been caused by the downlappingor onlapping of succeeding units on to thesequence boundary. The intraclastic dolostones(Wonoka Formation Unit 1 of Haines 1986a)which overlie this boundary most likelyrepresent lag deposits, developed on the seafloorduring a period of low sediment input by theerosion and winnowing of the concretionarydolostones. The overlying sandstone (WonokaFormation Unit 2 of Haines 1986a), commonlywith hummocky cross-stratification, is ofuncertain derivation. It may imply lower shore-face sedimentation under the influence of storm-developed oscillations (Walker 1984), or mayrepresent deepwater deposition, imprinted bydeepwater hummocky cross-stratification of thetype described by Prave (1985). Further work isobviously required at this critical stratigraphiclevel, particularly in view of the major canyonerosion event that appears to coincide with orimmediately post-date deposition of thesesandstones.

DISCUSSION AND CONCLUSIONS

Late Proterozoic strata at the approximate strati-graphic level of the Wilpena Group exhibitremarkable similarities over a broad region.Within the Australian continent, for example,Preiss et al (1978), Coats and Preiss (1980) andPreiss and Forbes (1981) demonstrate simil-arities of lithology and sedimentary style withinseveral major cycles between the KimberleyRanges of northwestern Australia, the AmadeusBasin of central Australia, and the AdelaideGeosyncline. These cycles, in the context of thispaper, would be described as depositionalsequences. Although further research is required,some of the sequence boundaries between the

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011

68 C. C. VON DERBORCH ET AL

large-scale (first-, second- and third-order)sequences may be correlative in a chronostrati-graphic sense. The present authors also notestriking similarities between several sequences inapproximately time-equivalent strata of theNorth American Cordillera and those of theAdelaide Geosyncline, and Young (1981) hintsthat Proterozoic rocks of the western Pacific con-tinents may represent a rifted portion of theNorth American Cordillera. It is thereforepossible, although by no means proven, thateustatically-driven base level control may haveplayed a major role in producing some of thesesimilarities.

In order to seek underlying reasons for theabove observations, a fundamental step isprecisely to identify and characterize the majorsequence boundaries — the purpose of thispaper. The next step, briefly covered in thisdiscussion, is to seek evidence of what actuallytriggered sequence boundary development. Thismost likely was due to significant lowerings ofrelative sea level. The important question iswhether localized tectonism was responsible forthe observed relative sea level changes, orwhether eustatic sea level variations, induced bychanges in the geometry of the ocean basins orthe volume of continental ice, may have beenresponsible.

Evidence of significant seaward migrations ofcoastal onlap

Sequence boundary A (Fig. 4) is associated with asharp facies break between the Mount CurtisTillite and Nuccaleena Formation. While it is notyet possible to propose a specific depositionalenvironment for pebbly and bouldery sand-stones of the Mount Curtis Tillite, this is locallyoverlain by a texturally mature sandstone unit,the Balparana Sandstone, which has possiblefluvial affinities and which clearly was truncatedand locally removed by erosion prior toNuccaleena Formation deposition. The abruptlithological change to the interpreted 'supratidal'dolostone, which grades upsection to mudstoneand fine grained sandstone, implies coastal onlapon to the sequence boundary during trans-gression, followed by overall deepening.

The Nuccaleena Formation is strikinglysimilar, both in lithology and stratigraphicposition, to other 'cap dolomites' which overlieLate Proterozoic glaciogene successions in theKimberley Ranges (Williams 1979; Coats &

Preiss 1980) and Amadeus Basin (Preiss et al1978). The widespread presence of such thin(several metres thick) dolostone units, at the baseof a clearly transgressive succession, impliestransgression of a broad, low gradient coastalplain, following emergence and erosion. Theclose association of such dolostones withunderlying glaciogene strata argues strongly forglacio-eustatic control, possibly due to LateProterozoic deglaciation, as previously suggestedby Preiss and Forbes (1981). Although furtherverification is required, it is tempting to proposethat sequence boundary A in the Adelaide Geo-syncline is indeed related to a eustatic event, withthe corollary that boundary A is of global chrono-stratigraphic significance.

Sequence boundary B (Fig. 4) is taken at asharp contact between what is interpreted as anearshore marine sandstone unit (ABC RangeQuartzite) and thin overlying sandstone andinterbedded mudstone of the basal BunyerooFormation. The ABC Range Quartzite here isonly centimetres in thickness, compared withseveral hundred metres farther south. Althoughequivocal, it is possible that subaerial erosion hasremoved much of the ABC Range Quartzite atlocality M (Fig. 2). However, the thinning is morelikely to be depositional and to relate to lateralfacies changes. Basal sediments ;;f the BunyerooFormation which overlie boundary B demon-strate a distinct upwards-fining trend, and in anearby locality contain rare calcite pseudo-morphs after discoidal gypsum, suggesting atleast localized subaerial exposure. The evidenceagain points to a significant seaward migration ofcoastal onlap associated with boundary B,followed by onlap on to the sequence boundaryof basal Bunyeroo Formation and subsequentdevelopment of relatively deepwater mudstone.Sequence boundary B exhibits a similar stylethroughout widespread areas of the AdelaideGeosyncline. Such a lack of regional variation isdifficult to reconcile with localized changes intectonism or sedimentation rates. It is proposedthat a eustatic sea level control best accounts forthe evidence.

Boundary C (Fig. 4) is interpreted either as adownlap surface within a sequence, or as a com-bined sequence boundary and downlap surface.Whatever its character in seismic stratigraphicterms, it is considered to represent a relatively'deepwater' boundary, at least in the vicinity oflocality N (Fig. 2) and possibly over a large part ofthe central and northern Adelaide Geosyncline.

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


If this interpretation is correct, it contrasts withthe inferred subaerial nature of boundaries A andB at localities L and M. Boundary C is of addedinterest because it lies just below the stratigraphiclevel from which the large canyons were incisedinto underlying sequences. Because of theobvious significance of the canyon-formingprocess in the context of global or local base levelchanges, some aspects of the canyons arediscussed below.

Canyon erosion

Large scale erosional incisions or canyons, withvertical dimensions measured in hundreds ofmetres to in excess of one kilometre, within theWonoka Formation (von der Borch et al 1982,1985) have been described as submarine can-yons, with the inference that much of theirerosion and infilling was submarine in origin andrelated to turbidity current activity on basinslopes. Another possibility is that erosion ofthese features could in part have been subaerial(Eickhoff et al 1986). Infilling sediments mayeven be of estuarine origin. Each interpretationhas important implications for possible eustaticsea level changes, which in turn have a bearing onglobal correlation of Late Proterozoic deposi-tional sequences. A key element in the debaterelates to the interpreted environments ofdeposition of the canyon-filling sediments. Thelack of faunal remains makes this a difficult task,with often equivocal sedimentary structures pro-viding the only available data. In addition, thereis a lack of reliable facies models both forestuaries and for submarine canyons.

Canyon-filling sediments of the Wonoka For-mation constitute an upwards-fining succession.Basal sediments are typically breccias andconglomerates, with clasts composed of diversecarbonates and quartzites as well as obviouscanyon-wall lithologies derived from theBrachina Subgroup, Bunyeroo Formation andUnit 2 (Haines 1986a) of the WonokaFormation. Interbedded with and overlyingbasal breccias are mixed carbonate and silici-clastic sandstone and interbedded mudstone.Sandstone beds, which vary in thickness fromcentimetres to in excess of one metre, oftendisplay sharp bases, well-defined flute casts andabundant climbing ripples. Some sandstonesdisplay features which could be ascribed to aturbidity current origin, typically Tbc, Tbcd andmore rarely Tabcd of the Bouma sequence.

However, many throughout the canyon-fill arecharacterized by hummocky cross-stratification,which is normally interpreted as havingdeveloped in lower shoreface environments (seeWalker 1984), but which has also been describedin deepwater sediments (Prave 1985).

The canyons may have been excavated in asubmarine setting, either by turbidity currents,mass-movements, or a combination of theseprocesses. In this model, a significantly seawardshift in coastal onlap to the vicinity of the shelfedge is required prior to canyon excavation inorder to permit canyon-cutting elastics to betransported to a slope environment. The verticaldrop in coastal onlap in this model needs to be ofthe order of 200 m. According to field strati-graphic relationships, the stratigraphic positionof this event coincides with a cryptic sequenceboundary situated several metres above Unit 2(Haines 1986a) of the Wonoka Formation.

Another possibility, also related to a sub-marine erosion model, places the canyon-triggering seaward shift in coastal onlap at alower stratigraphic level, possibly coincidingwith sequence boundary B. In this model a majorsequence (Bunyeroo Formation), and possiblylesser sequences wjthin lowermost WonokaFormation), were deposited on the shelf andslope around shoulders of previously incisedcanyons, similar to the situation described for aQuaternary canyon offshore from New Guinea(von der Borch 1969). Mass movements duringsubmarine cutting and subsequent infilling of thecanyons could then have caused widening andlateral wasting of the Bunyeroo Formation andoverlying sequences which draped canyon inter-fluves, thereby accounting for clasts derivedfrom these units in canyon-wall breccias.Canyons could subsequently have becomeinfilled by marine onlap prior to deposition ofmiddle to upper Wonoka Formation sedi-ments.

An alternative model relates to subaerialerosion and subsequent infilling by shallowmarine sediments during coastal onlap. In itssimplest form, this requires a relative base levelfall of the same order of magnitude as that of theactual canyon depth — about 1 km. Such a fallcould permit fluvial erosion to cut the canyons tomarine base level, while a subsequent relative sealevel rise of the same order of magnitude isrequired to account for onlap-fill by shallowmarine sediments.

Whichever of these models proves correct, a

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


noteworthy increase in rate of fall of eustatic sealevel, a decrease in rate of tectonic subsidence, orlocalized tectonic uplift, is required to initiatecanyon erosion. If eustatic changes are involved,a signal of this type should leave its record inother Late Proterozoic basins. If such evidencecan be found elsewhere at the equivalentsequence boundary, then the case for a eustaticcontrol may be strong. Absence of such evidencemay argue for more local base level control,perhaps related to tectonic uplift during sedi-mentation. The problem in this case is the largeamount of uplift required, which would be inexcess of 1 km. Uplifts of this magnitude arepossible in tectonic settings such as the ColoradoPlateau, but such a situation is not obviouslycompatible with the observed tectonic setting ofthe Adelaide Geosyncline during Wilpena Groupdeposition. Research is continuing on sequencesand sequence boundaries of the Wilpena Groupand its stratigraphic equivalents in other basins,in order to clarify some of these enigmaticobservations.

ACKNOWLEDGMENTS

Fruitful discussions and field studies withAdelaide Geosyncline coworkers Ian Dyson,Karl EickhofF, Pete DiBona, Jonathan Clarke,Ukat Sukanta, Steve Abbott and Chris Hill aregratefully acknowledged. Esso Australia Ltd, theAustralian Research Grants Scheme and theFlinders University Research Budget supportedfieldwork by von der Borch and Grady. Travel toAustralia by Christie-Blick was supported in partby the Donors of the Petroleum Research Fund,administered by the American Chemical Society(PRF 16042-G2 to N.C.B.) and by Lamont-Doherty Geological Observatory. Typing wasdone by Inara Stuart and drafting by GailJackson. We thank David McCormick andCharlotte Schreiber for their helpful reviews. Weare also indebted to Vic Gostin, Larry Frakes,Barry Webby and an anonymous reviewer fortheir critical comments. We also appreciatedcritical comments in the field from members ofI.A.S. excursion No. 27b (1986). Lamont-Doherty Geological Observatory Reference No.4209 (Christie-Blick).

REFERENCESBAKER P. A. & BURNS S. J. 1985. Occurrence and

formation of dolomite in organic-rich continentalmargin sediments. American Association ofPetroleum Geologists Bulletin 69, 1917-1930.

BJORLYKKE K. 1982. Correlation of late Precambrianand early Palaeozoic sequences by eustatic sealevelchanges and the selection of the Precambrian-Cambrian boundary. Precambrian Research 17,99-104.

CAMPBELL C. V. 1967. Lamina, laminaset, bed andbedset. Sedimentology 8, 7-26.

CHRISTIE-BLICK N. & LEVY M. 1985. A new approachto time correlation in Proterozoic rocks: sequenceboundaries in the Brigham Group, Utah. Geo-logical Society of America, Abstracts withPrograms 17, 546.

CHRISTIE-BLICK N. & VON DER BORCH C. C. 1985.Sequence boundaries in Proterozoic successions: acomparison of the Brigham Group (Utah) andWilpena Group (South Australia), with impli-cations for sea-level change and global correlation.Geological Society of America, Abstracts withPrograms 17, 546.

COATS R. P. 1973. Copley. 1:250,000 Geological Sheet.Geological Survey of South A ustralia, ExplanatoryNotes 1:250 000 Geological Series SH/54-9.

COATS R. P. & PREISS W. V. 1980. Stratigraphic andgeochronological reinterpretation of late Protero-zoic glaciogenic sequences in the Kimberleyregion, Western Australia. Precambrian Research13, 181-208.

DI BONA P. A. & VON DER BORCH C. C. 1985. WonokaFormation sedimentation, relative sealevel, andbasin analysis in the northern Flinders Ranges,South Australia. Geological Society of Australia,Abstracts 15, 228.

DOTT R. H. JR. 1983. 1982 SEPM Presidential address:Episodic sedimentation — How normal isaverage? How rare is rare? Does it matter? Journalof Sedimentary Petrology 53, 5-23.

EICKHOFF K. H., VON DER BORCH C. C. & GRADY A. E.1986. Incised meandering canyon in the lateProterozoic Wilpena Group; problems of timingand thalweg migration. International Sedi-mentological Congress, Canberra, Abstracts 92.

GEHLING J. G. 1983. The Ediacara Member: ashallowing upward submarine fan sequence withinthe Pound Subgroup. Geological Society ofAustralia, Abstracts 10, 52-54.

GOSTIN V. A. & JENKINS R. F. 1983. Sedimentation ofthe early Ediacaran, Flinders Ranges, SouthAustralia. Geological Society of Australia,Abstracts 9, 196-197.

GROTZINGER J. P. 1986. Cyclicity and paleoenviron-mental dynamics, Rocknest Platform, northwestCanada. Geological Society of America Bullet in 97,1208-1231.

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


HAINES P. W. 1986a. The upper Wilpena Group atBunyeroo Gorge, a one day excursion. In DalgarnoC. R. compiler. Proterozoic to Cambriansedimentary environments and resource potential,Flinders Ranges, South Australia. Guidebook,Eighth Australian Geological Convention,Geological Society of Australia.

HAINES P. W. 1986b. Late Proterozoic carbonate shelfto shale basin transition, Wonoka Formation,Flinders Ranges, S.A. Geological Society ofAustralia, Abstracts 15, 92.

JENKINS R. F., FORD C. H. & GEHLING J. G. 1983. TheEdiacara Member of the Rawnsley Quartzite: thecontext of the Ediacaran assemblage (latePrecambrian, Flinders Ranges). Journal of theGeological Society of Australia 30, 101-119.

JENKINS R. F. & GOSTIN V. A. 1983. Marinoan andEdiacaran type sections in the context of tectoniccycles in the Adelaide Geosyncline. GeologicalSociety of Australia, Abstracts 10, 39-44.

MITCHUM R. M. JR, VAIL P. R. & THOMPSON S. III1977. Seismic stratigraphy and global changes ofsealevel, Part 2: the depositional sequence as abasic unit for stratigraphic analysis. AmericanAssociation of Petroleum Geologists Memoir 26,53-62.

MUTTI E. 1985. Turbidite systems and their relations todepositional sequences. In Zuffa G. G. ed.Provenance of arenites, pp. 65-93. D. ReidelPublishing Co., Boston.

PIPIRINGOS G. N. & O'SULLIVAN R. B. 1978. Principalunconformities in Triassic and Jurassic rocks,western interior United States — a preliminarysurvey. United States Geological SurveyProfessional Paper 1035-A.

PLUMMER P. S. 1978a. Stratigraphy of the lowerWilpena Group (late Precambrian, FlindersRanges, South Australia). Transactions of theRoyal Society of South Australia 102, 25-38.

PLUMMER P. S. 1978b. Note on the palaeoenviron-mental significance of the Nuccaleena Formation(upper Precambrian), Central Flinders Ranges,South Australia. Journal of the Geological Societyof Australia 25, 395-402.

PRAVE A. R. 1985. Can hummocky cross-stratificationbe formed below effective wave base? GeologicalSociety of America, Abstracts with Programs 117,693.

PREISS W. V. 1983. Adelaide Geosyncline and StuartShelf: Precambrian and Palaeozoic geology.1:600,000 scale geological map, Department ofMines & Energy, Adelaide.

PREISS W. V. & FORBES B. G. 1981. Stratigraphy,correlation and sedimentary history of Adelaidean(late Proterozoic) basins in Australia. PrecambrianResearch 15, 155-304.

PREISS W. V., WALTER M. R., COATS R.- P. &WELLS A. T. 1978. Lithological correlations of

Adelaidean glaciogenic rocks in parts of theAmadeus, Ngalia and Georgina Basins. BMRJournal of Australian Geology & Geophysics 3,43-53.

RUTLAND R. W. R., PARKER A. J., PITT G. M., PREISSW. V. & MURRELL B. 1981. The Precambrian ofSouth Australia. In Hunter D. R. ed. Precambrianof the Southern Hemisphere. Developments inPrecambrian Geology 2, 309-360. Elsevier,Netherlands.

SLOSS L. L. 1963. Sequence in the cratonic interior ofNorth America. Geological Society of AmericaBulletin 74, 93-113.

THOMPSON B. P., MIRAMS R. C., FORBES B. G.,DALGARNO C. R. & JOHNSON J. E. 1964. Pre-cambrian rock groups in the Adelaide Geo-syncline: a new subdivision. Quarterly GeologicalNotes, Geological Survey of South Australia 9,1-10.

VAIL P. R., HARDENBOL J. & TODD R. G. 1984. Jurassicunconformities, chronostratigraphy, and sea-levelchanges from seismic stratigraphy and bio-stratigraphy. In Schlee J. S. ed. Interregionalunconformities and hydrocarbon accumulation.American Association of Petroleum GeologistsMemoir 36, 129-144.

VAIL P. R., MITCHUM R. M. JR & THOMPSON S. III1977. Seismic stratigraphy and global changes ofsealevel, Part 3: Relative changes of sealevel fromcoastal onlap. American Association of PetroleumGeologists Memoir 26, 63-81.

VON DER BORCH C. C. 1969. Submarine canyons ofsoutheastern New Guinea: seismic andbathymetric evidence for their modes of origin.Deep Sea Research 16, 323-328.

VON DER BORCH C. C. 1980. Evolution of lateProterozoic to early Paleozoic Adelaide Foldbelt,Australia: comparisons with post-Permian riftsand passive margins. Tectonophysics 70,115-134.

VON DER BORCH C. C., GRADY A. E., ALDAM R., et al1985. A large-scale meandering submarinecanyon: outcrop example from the late Protero-zoic Adelaide Geosyncline, South Australia. Sedi-mentology 32, 507-518.

VON DER BORCH C. C. & LOCK D. 1979. Geologicalsignificance of Coorong dolomites. Sedimentology26, 813-824.

VON DER BORCH C. C., SMIT R. & GRADY A. E. 1982.Late Proterozoic submarine canyons of AdelaideGeosyncline, South Australia. American Asso-ciation of Petroleum Geologists Bulletin 66,332-347.

WALKER R. G. 1984. Shelf and shallow marine sands.In Walker R. G. ed. Facies Models, second edition,pp. 141-170. Geoscience Canada ReprintSeries 1.

WILLIAMS G. E. 1979. Sedimentology, stable-isotope

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011


geochemistry and palaeoenvironment of dolo-stones capping late Precambrian glacial sequencesin Australia. Journal of the Geological Society ofAustralia 26, 337-386.

YOUNG G. M. 1981. The Amundsen Embayment,Northwest Territories: relevance to the upper

Proterozoic evolution of north America. InCampbell F. H. A. ed. Proterozoic basins ofCanada. Geological Survey of Canada Paper81-10.

(Received 16 February 1987; accepted 12 June 1987)

Dow

nloa

ded

by [

Col

umbi

a U

nive

rsity

] at

15:

44 0

9 O

ctob

er 2

011

depositional sequence analysis applied to late proterozoic

Documents