The northern New England Fold Belt in central and southern Queensland comprises a collage of terranes the complexity and variety of which dominantly reflect tectonic events in the ~100my interval from mid-Carboniferous through the Middle Triassic. We present a synthesis of our current, but evolving, understanding of the Carboniferous to Late Permian tectonic evolution of this region which we argue was controlled by a transition from active continental-arc/accretionary processes to widespread crustal extension.

Contrary to earlier interpretations, we find little evidence for the existence of an Early Carboniferous magmatic "arc" (Connors-Auburn Volcanic Arc) that forms a companion to the more easterly interpreted forearc and accretionary complex elements. These elements, together, have been interpreted widely in terms of segments of a complete subduction complex. Dating of the magmatic elements of assumed Early Carboniferous age shows them to be overwhelmingly Late Carboniferous to Early Permian in age. The Connors-Auburn Volcanic Arc and the Early Permian Camboon Volcanic Arc are thus part of a single, broad magmatic cycle that we propose should be called the Connors-Camboon Province. Significant intraformational unconformities within the volcanics of this Province are interpreted as evidence for ongoing extension. The present data allow two alternative end-member interpretations, or perhaps more realistically mixtures of these end-members. These are 1) the existence of Late Carbonifereous subduction and associated magmatism that is transitional in nature to Early Permian extension, and 2) a more controversial interpretation of a passive margin within which all magmatism of Late Carboniferous to Early Permian age records a continental-scale extension and thermal event. Evidence for this thermal event is recorded much further east of the traditionally recognised magmatic province in the Connors-Camboon Province as extension and crustal melting in the earlier forearc assemblages. Irrespective of these interpretations, the calc-alkaline dominated phase of volcanism is transitional into the late Early Permian and formation of the Bowen Basin, a basin that originally extended eastward of the present structural margin to a shallow marine coastline. The extensional phase of basin formation is characterised by clastic sedimentation into developing grabens and half grabens and associated mafic-dominated volcanism. Elements of basin formation are recognised throughout the NNEFB and the distribution of "Bowen Basin" rocks differs significantly from traditional models.

The New England Fold Belt (NEFB) of eastern Australia (Fig.1) comprises a number of variably deformed terranes, ranging in age from Early Palaeozoic to Late Triassic. Each terrane is generally interpreted to reflect a different tectonic element, or record a different orogenic episode and generally has been regarded as forming within a broadly convergent continental margin setting (Day et al., 1978, 1983; Henderson, 1980). This paper, and its companion (Holcombe et al., this volume) present a synthesis of our current, but evolving, understanding of the post-Devonian time-space framework and provide constraints on tectonic models for the northern New England Fold Belt (NNEFB) in central Queensland.

The main tectonic elements of NNEFB had been recognized by the mid-1970’s. Day et al. (1978), in a major summary of the regional geological history, presented a model involving the periodic development of a magmatic arc along what is now the Connors-Auburn Arch. This was interpreted in terms of a convergent continental margin, with most of the NNEFB elements in the arc-forearc region. The tectonic history was generally thought of in terms of three arc-producing events: the Siluro-Devonian Calliope Arc, developed as an island arc then subsequently accreted; the Devonian-Carboniferous Connors-Auburn Volcanic Arc, developed along an Andean-style margin; and the Early Permian Camboon Volcanic Arc, developed essentially on top of the Connors-Auburn Volcanic Arc.

Of these separate "orogenic" events, however, only the Late Devonian-Carboniferous arc was associated with most of the tectonic elements expected at a continental margin subduction orogen (Day et al., 1978; Murray et al., 1987). These elements comprise the Wandilla Slope and Basin accretionary complex, the Yarrol forearc basin, and the coeval, possibly back-arc extensional Drummond Basin (Johnson & Henderson, 1991; Henderson et al., 1993). The forearc-setting interpretations have been supported by more recent work (e.g., Fergusson et al., 1990; Leitch et al., 1994), and Leitch et al. (1993) suggested that the final timing of active accretion persisted into the latest Carboniferous.

Several authors have argued for possible allochthonous elements within the fold belt. Both the Palaeozoic North D’Aguilar and the Permo-Triassic Gympie Blocks, in particular, are characterized by units that are dissimilar in both composition and metamorphic grade to adjacent rocks thought to be of the same age. It has also been suggested that the North D’Aguilar block, with its dominantly mafic metavolcanic units and large serpentinite bodies, was an accreted arc or back-arc (Waterhouse & Sivell, 1987a). Various origins have been suggested for the Gympie Block ranging from an accreted arc (Sivell & Waterhouse , 1987) to an exotic displaced terrane derived perhaps from New Zealand (Harrington, 1974,1983; Waterhouse & Sivell, 1987b). In the northern part of the fold belt, the origin of the serpentinites and greenschist and amphibolite facies schists of the Marlborough block remained somewhat enigmatic. The ultramafic components have generally been regarded as ophiolitic in origin (Murray, 1974; Leitch et al., 1994). The schists were regarded as distinct from the surrounding coastal accretionary terranes (Henderson et al., 1993).

The last decade has seen a resurgence of new research in the northern New England Fold Belt (NNEFB) in Queensland. Independent groups working on projects as diverse as southeast Queensland structure and tectonics, Bowen Basin sedimentological studies, and Permian volcanic petrology began to converge by the early 1990’s toward common conclusions that, in some respects, differ considerably from the earlier tectonic models. In particular, we now recognize two departures from previous models:

• we argue that there are limited data to support the existence of an Early Carboniferous volcanic arc that developed coevally with, and to the west of, the subduction complex elements.
• we also recognise a major crustal thermal event in the Late Carboniferous-earliest Permian, that arguably records a transition from a convergent to an extensional, possibly back-arc, setting.

Figure 2 is a Time-Space plot of the northern NEFB that summarizes our current state of knowledge of the critical terranes on which we base much of our tectonic analysis of the post-Devonian history. Elements of this framework have appeared in numerous conference proceedings and recent unpublished theses. The stratigraphic timescale used for correlation is based on Jones (1996) and future refinements of this will change some critical event boundaries, particularly in the Permian.

Southeast Queensland

The oldest rocks in the southern Queensland section of the NEFB are strongly deformed, Devonian?-Carboniferous accretionary rocks exposed within several discrete "basement blocks". The Beenleigh and South D'Aguilar blocks (Figs. 1, 3) comprise low grade, internally discontinuous (fault-bounded) units of ribbon chert, argillite, chert-argillite mélange, greywacke, and metabasaltic greenstone (including pillow basalt). The rocks (Neranleigh-Fernvale beds) are moderately to strongly deformed, generally containing a single generation, steeply west-dipping slaty cleavage accompanying a transposition layering. Remnants of similar rocks occur throughout the North D’Aguilar block (Wide Bay Creek Broken Formation, Booloumba beds, Amamoor beds: Table 1). Metamorphic grades are low: prehnite-pumpellyite or prehnite-actinolite facies assemblages occur in the basaltic rocks; anchizonal conditions, with T<250o C and moderate pressure, are inferred from illite crystallinity and bo data in the metasediments (Sliwa, 1994). These rocks, containing Late Devonian to Early Carboniferous macroinvertebrates and radiolaria (Fleming et al., 1974; Aitchison, 1988; Ishiga, 1990), are interpreted as representing upper levels of the accretionary prism, and are typical of other accretionary assemblages in the Woolomin terrane in the southern NEFB (Cawood & Leitch, 1985) and the Wandilla/Shoalwater terranes in the northern NEFB (Fleming et al., 1975; Fergusson et al., 1993).

The North D’Aguilar block (Fig.1) is a composite terrane consisting of deeply subducted (18-20 km), polymetamorphic, but originally epidote-blueschist facies, ophiolitic assemblages (Rocksberg Greenstone, Mt Mia serpentinite-matrix mélange; Table 1) juxtaposed against the low grade accretionary rocks by a complex system of low angle detachment faults, steep normal faults, and younger thrust faults (Little et al., 1992; 1993; Sliwa, 1994). The Rocksberg Greenstone outcrops as a semi-coherent, mafic metavolcaniclastic unit in the southern part of the North D’Aguilar Block, and as detached blocks and "knockers" within the serpentinite-matrix mélange in the northern part of the Block. Although strongly overprinted by later deformation, the initial fabric in the Rocksberg Greenstone is steeply dipping, and is interpreted to be an accretion-related fabric similar to that in the lower grade basement blocks. Strain is highly partitioned in this earliest deformation and the highly schistose, map-scale zones within which epidote blueschist assemblages are preferentially developed are thought to represent deep fluid-focussing shear zones within the subducting pile (Holcombe and Little, 1994). These rocks are interpreted as underplated ophiolitic components of the accretionary wedge, that were detached and subducted to depths >18 km before being exhumed in the footwall of an extensional detachment in the Late Carboniferous (see below).

Marlborough Block

The Marlborough Block near the northern end of the NEFB is a thin-skinned nappe sheet (Holcombe et al., 1995; this volume) carrying composite terranes that are very similar to those in the North D’Aguilar Block in southeast Queensland. Within the block, serpentinized ultramafic rocks, including harzburgites, cumulates, and gabbro are juxtaposed by steep (rotated?) thrusts and shear zones against chert-argillite, amphibolite, schist, and locally marble of greenschist to amphibolite facies metamorphic grade. We now recognize that the schist units include deformed equivalents of foliated S-type granitoids that occur elsewhere within the block (see below), and we believe that much of the chert-argillite component resembles the Devonian-Carboniferous accretionary rocks exposed further to the east (but see Henderson et al., 1993). Unlike the North D’Aguilar Block, serpentinite-matrix mélange is rare and most of the serpentinite bodies in the Marlborough Block are serpentinized massive ultramafic rocks. Sheared serpentinite is most commonly associated with Cretaceous or Tertiary faults.

Age constraints on accretion

The age of the first deformational fabric in the accretionary rocks, and hence on the timing of active accretion, is poorly constrained at mid-Carboniferous. An absolute upper limit is imposed by the tightly constrained ~305 Ma overprinting metamorphic and magmatic event (below). Aitchison (1988) describes Early Carboniferous (Tournaisian) radiolaria from ribbon cherts of the Neranleigh-Fernvale beds. Little et al. (1995) interpret a complex Ar⁴⁰/Ar³⁹ whole rock spectrum taken from a slate in the higher level accretionary rocks as indicating a minimum age on the subduction-related fabric of ~315 Ma but with most of the high K/Ca steps >330 Ma.

Southeast Queensland

In the far northern part of the North D’Aguilar Block (Figs. 1, 3), tectonostratigraphic relationships between the polymetamorphic, higher grade, ophiolitic rocks and the low grade, single fabric, accretionary rocks are well-defined (Little, 1992; Donchak et al., 1995). The higher grade rocks are overlain structurally by the thick Mt Mia serpentinite-matrix mélange sheet, which is separated in turn from the overlying low grade rocks by a gently arched detachment fault (Little et al., 1993). Shallowly dipping, second generation, deformation fabrics (sub-parallel to the detachment) dominate most outcrops of the higher grade rocks and these fabrics are accompanied by a regional greenschist facies overprint of the earlier epidote-blueschist facies mineral assemblages. Similar fabrics, with similar detachment-parallel orientation, occur within fault-bounded phyllitic units exposed along the detachment. These appear to represent hangingwall accretionary rocks (eg., Anderson Creek Phyllite and probably Bunya Phyllite) that were derived from intermediate depth levels of the complex, heated and recrystallized by contact with the footwall, and then accreted to the lower plate. The Greenbank Layered Sequence, with its shallowly dipping reflectors seismically imaged at depths equivalent to >1.5 seconds Two-Way Time below the Beenleigh Block (Korsch et al., 1989), probably represents these same fabrics and detachment system.

The pressure gap across the detachment fault between the epidote-blueschist facies rocks (>0.6 GPa) and the anchizonal accretionary sediments is ~0.3GPa, representing a depth difference of ~9 km (Little et al., 1995) . The geometry is that of a core complex where a crustal scale extensional detachment has exhumed a lower plate of deep-seated, ductilely deformed rocks against an upper plate of high level, more brittley deformed rocks (Little et al., 1993). Similar extensional core complexes occur in other convergent margin settings (e.g. Lee and Lister, 1992; Wijbrans et al., 1993).

There is some question as to whether the arched nature of the detachment system in the northern North D’Aguilar Block is a function of the original core complex geometry as suggested by Little et al. (1993) or has been overprinted by later flexuring. Recent mapping in the area indicates multiple thrust imbrication of some units (Donchak et al., 1995), suggesting that this later contraction (see Holcombe et al., this volume) may control the overall geometry. Broad folding of shallowly dipping thrust fabrics also occurs in the central part of the block (Sliwa, 1994).

A suite of Late Carboniferous, mildly S-type, granodiorite and diorite plutons are exposed within a thrust sheet (Claddagh Thrust) in the northern North D’Aguilar Block (Little et al., 1993). The plutons intrude the lower plate, polymetamorphic, rocks and include the Claddagh and Gallangowan granodiorites and several smaller unnamed plutons (Sliwa, 1994). Granodiorite intrusion was synkinematic with the deformation forming the detachment-parallel fabrics in the higher grade rocks (Little et al., 1993, 1994, 1995). The granodiorite bodies are variably foliated, and xenoliths of more deformed (mylonitic) phases are common. The regional greenschist facies thermal metamorphic overprint increases to amphibolite facies around the igneous bodies, and the foliation within the granodiorite is parallel to the detachment-parallel fabrics in the aureole rocks.

Because of the demonstrable structural relationship between magmatism, metamorphism, and deformation, the timing of the deformation fabrics accompanying exhumation can be reliably dated. 40Ar/39Ar dating of hornblende and white mica from the Claddagh granodiorite and its aureole indicate syntectonic crystallization at ~306 Ma (306.5±0.6; 306.9±1.2 Ma ages on hornblende), followed by continued exhumation and cooling until all rocks had passed through the ~350°C white mica blocking temperature by ~296 Ma (see Little et al., 1995 for details). K/Ar ages on these foliated granodiorites are regarded as less precise, and yield slightly older ages of ~310 Ma Sliwa (1994; and Table 1).

The precise dating of a significant crustal extension, and crustal melting, event at ~305 Ma is a major constraint on our tectonic models and poses a significant problem when related to the coeval peak of apparent arc magmatism occurring to the west.

Other foliated S-Type granitoid terranes

Foliated S-type granitoids similar to the Claddagh-Gallangowan bodies described above occur sporadically throughout the NEFB. They are generally intrusive into the earlier accretionary rocks and are locally associated with metamorphic culminations in the fold belt. In the southern NEFB, these include the Hillgrove igneous suite which is considered to have an emplacement age of ~300 Ma but for which locally other origins have been proposed (e.g., thrust model: Dirks et al., 1993; diapiric models: Shaw and Flood, 1981).

A widespread, poorly exposed, area of garnet-(cordierite)-bearing gneissic granitoids occurs about 90 km west of the North D’Aguilar Block bodies. These metagranitoids, that we have termed the Chahpingah Complex (Fig. 1), are problematic in terms of their regional affinities, and have been undescribed previously except in unpublished Honours theses (Arthur, 1969; Virisheff, 1974). They have a gently folded, shallowly dipping, locally shear-dominated, foliation. Preliminary interpretation of 40Ar/39Ar step-heating spectra of biotite and muscovite from the metagranitoids indicates distinct groupings of cooling ages between 250-225 Ma as these rocks passed through the ~300° C isotherm. We surmise that they were emplaced syntectonically, probably before ~250 Ma and remained at depth until emplaced to their present structural position during the Permo-Triassic thrust event, and then subsequently exhumed.

We have mapped similar foliated garnet-bearing granitoid terranes within the Marlborough Block thrust nappe in the northern NEFB. Locally we correlate the gneissic metagranite and their amphibolite facies schistose aureole with the Broome Head Metamorphics (Morand, 1993) to the east. Isotopic ages derived from these rocks are ambiguous. Leitch et al. (1993) reported a Rb/Sr age of ~255 Ma from the Broome Head Metamorphics and we have derived a biotite K/Ar date of 268±3 Ma from one of the least deformed foliated granites in the Marlborough block (Muscio,1994). Preliminary interpretation of 40Ar/39Ar dating of biotite from other granitoids in this block yield cooling ages of 248.8±0.5 Ma from an unfoliated granodiorite and 242.9±0.4 Ma from a strongly foliated phase (Table 1). We suspect that the intrusion age of these S-type granitoids is probably older than Late Permian-Early Triassic and that the younger ages reflect exhumation and cooling during the massive thrust event in this region.

Calc-alkaline rocks of the Connors and Torsdale Volcanics are considered to represent the Early to mid-Carboniferous arc and the Early Permian Camboon Volcanics are the basis for postulation of an Early Permian arc (e.g., Day et al., 1978). Field and geochronologic studies, directed towards the distinction between the two groups of calc-alkaline rocks, indicate major problems with the previous arc models.

The Camboon Volcanics and Torsdale Beds (Jones, 1994; Jones et al., 1996) occur on the western side of the Auburn Arch. The Torsdale Beds are dominantly silicic ignimbrites that are intruded by numerous small granite bodies. Recent K/Ar dating of hornblende from a sample collected from near the unconformity with the overlying Camboon Volcanics gave an age of 312±4 Ma (Jones et al., op cit.). The unconformity is marked by coarse basal conglomerates composed predominantly of granitoid and felsic volcanic clasts. The Camboon Volcanics consists of a lower bimodal mafic lava/felsic ignimbrite unit and an upper terrestrial andesitic sequence from which a K/Ar age on plagioclase of 282.8±5 Ma was obtained at the top of the sequence (Jones et al., op cit.). The Camboon Volcanics are unconformably overlain by the late Artinskian (c. 275-280 Ma, using the AGSO 1995 time scale, Jones, 1996) marine Buffel Formation of the Bowen Basin.

Inliers of similar volcanics are exposed within the Gogango Overfolded Zone (GOZ) along the trend of the Connors-Auburn Arch. These were interpreted on 1: 250 000 maps as Camboon Volcanics in the south and as components of the Rannes Beds in the north (Malone et al., 1969; Kirkegaard et al., 1970; Dear et al., 1971). Our mapping shows that in both areas, these volcanics are the structurally deepest rocks exposed at the base of imbricate thrust sheets or within antiformal cores (Holcombe et al., 1995, this volume). Rocks mapped as Camboon Volcanics are variably deformed intermediate volcanic and volcaniclastic rocks. These rocks are predominantly subaerial, although intermediate-composition pillow lavas are locally developed. In the north, rocks mapped as Connors Volcanics or Rannes Beds consist predominantly of silicic volcanic and volcaniclastic rocks with relatively minor intermediate and mafic volcanic rocks.

Within the Connors Arch, the Connors Volcanics, and the lower parts of units mapped as Lizzie Creek Volcanics and Carmila Beds, consist of a stratigraphically complex sequence dominated by dacitic to felsic ignimbrite, rhyolite and volcaniclastic rocks, with lesser andesite and basalt. This sequence has an upper contact with a predominantly fine-grained, lacustrine sedimentary succession that typically is associated with basaltic volcanism, and which we equate on lithologic and structural grounds with the Early Permian extension phase of the Bowen Basin (see below, and Fielding et al., this volume). In rocks mapped as Carmila Beds north of the Marlborough Block (Kirkegaard, 1970), we have observed a mixed sequence of fine-grained sedimentary, rhyolitic and coarse-grained dacitic volcaniclastic rocks containing Early Permian Fauna 1 marine fossils occurring within 10’s of metres of the contact with the overlying fine-grained succession.

Bowen Basin elements in the Gogango Overfolded Zone

Following on from the work of others, Fielding et al. (1990, 1995) and Baker et al. (1993) presented a three-part model for the Bowen Basin, based on stratigraphic correlations, interpreted sediment dispersal patterns and sandstone petrology. According to this model, the stratigraphic record of the Bowen Basin may be divided into three suites, reflecting 1) an Early Permian phase of extensional subsidence and magmatic activity, 2) an early Late Permian period of mainly passive, thermal subsidence, and 3) a late Late Permian to early Late Triassic phase of foreland loading and contractional deformation. The last phase is associated with the onset of the Hunter-Bowen contractional event (see Holcombe et al., this volume) and is not treated here.

Tectonic interpretation of deformed Permian rocks within the GOZ has been hampered by a stratigraphic nomenclature based in part on deformation style rather than stratigraphic criteria (eg. Rannes Beds: Holcombe et al., this volume), and inconsistent stratigraphic ordering and correlation of units particularly between adjacent mapsheets. A critical step in our understanding of the Permian framework of the NNEFB has been regional stratigraphic correlations based on recognition of distinct assemblages of sedimentary lithofacies and associated igneous rock types, sediment dispersal patterns and interpreted palaeoenvironments, that are characteristic of the three phase evolution of the Bowen Basin. The issues summarised below are developed in more detail by Fielding et al. (this volume).

Phase 1 rocks are a diverse assemblage of fine- and coarse-grained sedimentary rocks, interbedded with a mafic-dominated bimodal igneous suite of mafic lavas and sills, minor ignimbrite and associated tuffs. Palaeocurrent distributions and sediment compositions indicate mainly local sediment derivation from multiple sources. Sequences in the NNEFB that we correlate with Phase 1 are the Youlambie Conglomerate and upper, fine-grained sedimentary parts of both the Carmila Beds and Lizzie Creek Volcanics. Locally, the marine Rookwood Volcanics are laterally transitional to the upper part of the Youlambie Conglomerate. Except for the Rookwood Volcanics, depositional environments during much of this phase were non-marine, mainly alluvial and lacustrine. Sediment accumulation is interpreted to have occurred within active grabens and half-grabens, separated by uplifted basement blocks.

A major change in depositional environment occurred towards the close of the extensional phase involving a diachronous marine transgression which flooded much of the basin, and led to the widespread establishment of coastal to marine shelf conditions. Bioclastic limestones were sporadically (but extensively) developed on remnant basement highs. The top of this succession is a basin-wide hiatal surface that almost certainly relates to tectonic processes within or marginal to the basin.

Phase 2 sediments form a continuous sheet across the basin, in contrast to underlying Phase 1 deposits, and sediment dispersal was westward (towards the Bowen Basin) on both sides of the "Connors Arch". Resulting, mainly fine-grained, clastic sediments are now represented by the Undifferentiated Back Creek Group ("Pb" of the local mapsheets) and parts of the Rannes Beds, Moah Creek Beds and Boomer Formation.

Rookwood Volcanics

The marine Rookwood Volcanics has been one of the most enigmatic units in the NNEFB. The unit was ascribed a late Lower Permian age on the basis of a single fossil locality (Briggs, 1993). The stratigraphic context of the unit, nonetheless, is poorly constrained on published map sheets.

The Rookwood Volcanics consist almost entirely of pillow lavas and high level intrusives with minor associated volcaniclastic sediments. Sediment intercalations are rare within the Rookwood Volcanics, but locally form accumulations ranging from a few centimetres up to at least 300 metres thick. Slices of massive rhyodacite and pillow basalt occur together within several thrust imbricates at Comanche Station west of Rockhampton but their contact relationships are unclear.

Most contacts between the Rookwood Volcanics and other units are either faulted or obscured by poor outcrop but the Rookwood Volcanics commonly structurally overlie the Early Permian Youlambie Conglomerate. In a complete section that we have mapped through the Rookwood Volcanics near the Fitzroy River, west of Rockhampton (and see O’Connell, 1995) the Rookwood Volcanics both overlie and pass laterally into the Youlambie Conglomerate. The unit here is overlain by rocks we equate with Phase 2. Early Permian (Artinskian) fossils have been found within both the uppermost Youlambie Conglomerate and near the top of the Rookwood Volcanics. K/Ar dating of two dolerites within the Rookwood Volcanics yield ages of 253±4 Ma and 258±5 Ma (O’Connell, op cit.;Table 1) but we interpret these to be ages that have been reset during the subsequent contractional event.

Locally abundant trace fossils (Cruziana ichnofacies), macroinvertebrate fragments and the one foraminifera locality, indicate a marine origin for the sequence, with water no deeper than shelfal (~200m). Small-scale, soft-sediment deformation is common, and attests to the likely unstable nature of the depositional basin. The abrupt lateral thinning of the Rookwood Volcanics and equivalence with the uppermost Youlambie Conglomerate at this locality is interpreted to define the hinge margin of a developing half-graben.

In a regional sense, the Rookwood Volcanics are petrographically indistinguishable from other Phase 1 lavas. Malone et al. (1969) interpreted the textures and petrographic characteristics of the Rookwood Volcanics as due to spilitisation, and this interpretation has consistently been raised as evidence for a deep marine eruptive environment. The identical secondary mineralogy and textures observed within all Phase 1 lavas, some of which were erupted subaerially, are clearly due to burial metamorphism during later basin evolution.

Berserker Beds

The Berserker Beds in the Rockhampton area are another marine volcanic unit within the NNEFB for which regional correlations are poorly constrained. The unit is composed of intermediate and felsic volcanic rocks and volcaniclastic sedimentary rocks, apparently passing upward into fine grained marine, commonly bioturbated, sedimentary units containing abundant trace fossils (Cruziana ichnofacies) and Early Permian shallow marine macrofossils (upper Sakmarian to lower Artinskian: Kirkegaard et al., 1970; Sainty, 1992; Briggs, 1993). The unit is host to the Mount Chalmers VHMS-style deposit and, as with the Rookwood Volcanics, there is a contradiction between the shallow marine conditions inferred from the fossil assemblage, and the deep marine (>1000m) conditions commonly inferred for exhalative sulphides (Hunns, 1994).

The internal structure is one of gently undulating dips, but the unit is entirely exposed between two northwest-trending thrust faults (Wilmott et al., 1986). The southern end of this belt passes offshore, and the northern end is covered by the Marlborough Thrust nappe, where several isolated klippen of serpentinite overlying the Berserker Beds attest to the sub-horizontal nature of the floor thrust. The trace fossil assemblage and character of the sedimentary units in the upper part of the stratigraphy are similar to those within sedimentary members within the nearby Rookwood Volcanics, and the fossil age ranges may overlap.

The geochemical character of the two sequences of volcanics is markedly dissimilar. An area of felsic volcanic, volcaniclastic and sedimentary rocks that we correlate with Connors Volcanics/Camboon Andesite (see above and Fielding et al., this volume) is exposed to the north of the Marlborough Block, at Charon Point. We suggest this area and the Berserker Beds may be contiguous, and have been overridden by the Marlborough Block nappe. If this is the case, then the Berserker Beds are most likely the regional equivalents of the Carboniferous-Early Permian volcanics, and the later Early Permian sediments equivalent to Phase 1 sedimentary strata further west.

Southeast Queensland basins

Several small, fault-bounded basins and basin remnants, characterized by indurated, labile sedimentary rocks, overlie the exhumed accretionary rocks of the North D’Aguilar Block. The principal unit is the Marumba Beds. Rock types include lithic sandstone, clast and matrix-supported conglomerate, megabreccia (clasts up to 10 metres diameter), massive argillite (locally calcareous), and minor basalt (Sliwa, 1994). Internal disruption of bedding, consistent with syn-sedimentary deformation in a local fault basin, is common. Clasts are dominantly intermediate but also include chert-argillite lithologies similar to those in the low grade accretionary rocks and mainly unfoliated granitoid. No blue amphibole-bearing rocks or serpentinite, rock types characteristic of the deepest structural levels of the adjacent metamorphic basement, have been found as clasts, although abundant foliated S-type granodiorite clasts similar to the adjacent Gallangowan Granodiorite have been found at one locality. The sequence is moderately to steeply dipping and relatively uncleaved, and exposed within the same thrust sheet that contains Late Carboniferous foliated granitoid.

Similar Early Permian basins, containing abundant mud-matrix conglomerate and breccia and interpreted as extensional in origin, occur extensively within the NEFB (eg., Barnard Basin, Leitch, 1988; Reids Dome Beds, Draper and Beeston, 1985; Fitzroy region, Fielding et al., this volume). We propose that the Marumba Beds may represent an early extensional manifestation of the Esk Trough. We correlate andesitic rocks of the Cedarton Volcanics with the Marumba Beds on the basis of the similar degree of metamorphism and presence of Early Permian macrofossils (Hill et al., 1972; Murphy et al., 1976; Murray et al., 1979). The Marumba Beds may also be equivalent to the seismically-imaged "deep rift-fill" interpreted by Korsch et al. (1989) to underlie the Esk Trough.

The Cambroon Beds (formerly included as part of the Amamoor Beds), are a deformed sequence of sedimentary rocks within the North D’Aguilar Block containing mud-matrix conglomerate and Early Permian macrofossils (Murray et al., 1979) and radiolaria (Ishiga, 1990). In contrast to the other Early Permian units in southeast Queensland, the Cambroon Beds are strongly cleaved, and locally polydeformed. Their juxtaposition against accretionary-complex rocks (Bouloomba Beds) that preserve an older, less complex deformational history is an important element in assessing the tectonics of the nearby Gympie Basin (see below).

Gympie Block

The Early Permian units of the Gympie Block (Highbury Volcanics, Rammutt Fm) are dominated by basaltic and andesitic volcanics and volcaniclastics, an association quite distinct in character from the mud-matrix conglomerate-bearing units of the Marumba and Cambroon Beds. Both sequences, nonetheless, contain abundant volcanic detritus of basic to intermediate composition and there is evidence in the nearby Cressbrook Creek Block for similar age marine fossil-bearing assemblages to the overlying middle Permian South Curra Limestone. There are no recognised equivalents to the Permian and Triassic sequences that overlie the limestone.

Permian sequences in the Gympie Block are not unlike those of similar age elsewhere in the NNEFB. The lower volcanic and sedimentary units are similar to Phase 1 extensional lithlogies although, like the Beserker Beds, may contain older volcanic units. The northeasterly trending normal growth faults within the Rammutt Formation that are important in localising gold mineralisation in the black slate facies at Gympie, are also consistent with formation in an extensional basin (Cuneen, 1994). The overlying South Curra Limestone is similar in age and lithology to the late Phase 1 and Phase 2 marine sequences in the GOZ to the north, and and the Late Permian transition into the overlying clastic sequences may include the Phase 2-Phase 3 thrust loading transition.

The structural style of the Gympie Block rocks has contributed as much as anything else to misconceptions and speculation about the origin of the sequence. The (apparently) most deformed rocks are those of the strongly-cleaved Triassic Kin Kin Phyllite at the top of both the structural and stratigraphic sequence. However, the metamorphic grade (very low greenschist facies) and degree of cleavage in the Kin Kin Phyllite is similar in all units. Both the intensity and orientation of cleavage vary considerably, a feature consistent with development in an easterly dipping shear regime with local strain partitioning. Strain is distributed more uniformly in the Kin Kin Phyllite, the weakest unit in the sequence. The large kink folds that are common in this unit represent very low strain magnitudes and may have formed as late as the Cretaceous deformation in the nearby Maryborough Basin.

The lithologies and deformation history of the rocks of the Gympie block are very similar to those of the Bowen Basin terranes within the northernmost part of the NNEFB. We interpret the Gympie Basin as a fragment of the Bowen Basin terranes from the northern part of the fold belt that has been displaced southward into its present position during the latter part of the Triassic thrust contraction (Holcombe et al., this volume).

A major constraint of previous tectonic interpretations of the NEFB has been the accepted existence of a complete convergent margin architecture in the northern NEFB in the Early to mid-Carboniferous, with a well-defined accretionary complex, forearc basin (Yarrol Basin), and magmatic arc system (Connors-Auburn Volcanic Arc). The Camboon Volcanics are widely considered to represent a superimposed, Early Permian subduction-related episode, although any associated forearc basin or accretionary wedge elements have not been recognised. Our field studies (in both the Marlborough and Cracow areas), in combination with re-interpreted and new age data (Gust et al., 1993, Allen pers.comm., 1996) suggest that both of these volcanic episodes are part of a broad, Late Carboniferous through Early Permian magmatic event. Furthermore, we now question whether any arc-signature volcanic rocks can be recognized that relate to the interpreted Early-Carboniferous forearc elements.

The concept of the Connors-Auburn Volcanic Arc was introduced by Day et al. (1978) who suggested the existence of an Andean-style volcanic arc along the site of the present Connors-Auburn Arch. The Yarrol Province was interpreted as an unstable forearc basin, with a complementary accretionary complex (Wandilla Terrane) originally attributed by Day et al. to be slope and abyssal plain deposits. The age of the elements of the subduction complex was most known from the Yarrol Province shelfal sequence, where oolitic limestone was widely developed. A unifying theme in associating the shelf and slope deposits was the presence of detrital oolites in the slope sequences. The age and existence of the volcanic arc was poorly constrained, based on: 1) the presence of volcanics intruded by Late Carboniferous granites, and thus broadly interpreted as Devonian/Carboniferous in age (Malone et al., 1966, Jensen et al., 1966), within the Connors Arch; and 2) a single 343 Ma K/Ar determination from an outlier of felsic volcanics in the southern Auburn Arch (Whitaker et al., 1974) correlated with the Torsdale Beds of the Auburn Arch.

Volcanism was interpreted to have ceased during the Late Carboniferous, coincident with the intrusion of voluminous granite batholiths, and sedimentation limited to relatively quartz-rich sediments on the shelf and slope. The principal metamorphism in the accretionary complex rocks was attributed a Late Carboniferous age associated with an orogeny that varied widely in age and intensity. During the Early Permian, the Camboon Volcanic Arc was interpreted to have developed over the prior Connors-Auburn Volcanic Arc, although there was no identified corresponding accretionary complex.

Dear (1994) described three sequences (his cycles 1-3) of predominantly felsic volcanic and volcaniclastic rocks in the southern Connors Arch, the lowest two of which are intruded by 316-287 Ma granites and are thus at least early Late Carboniferous in age. Dear correlated his Cycle 2 rocks with the Torsdale Beds on the basis of the intrusive ages, and his Cycle 3, which unconformably overlies a 316 Ma granite, with the Camboon Volcanics. The 312±4 Ma whole rock date from the Torsdale Beds (Jones et al., 1996; Table 1) and unpublished palaeomagnetic data (CSIRO report to Queensland Metals Corporation) suggest that the oldest sequences mapped by Dear are not older than mid-Carboniferous in age. Recent SHRIMP dating of zircons from granites of the Urannah Suite at the northern end of the Connors Arch (Allen, 1994; Allen, pers.comm.,1996) suggests that the main phase of activity is about 305 Ma, and that batholith development ceased by about 286 Ma. Thus, although we recognize that there are significant internal unconformities within the volcanics, we consider this entire suite, comprising Torsdale Volcanics, Connors Volcanics, Auburn Complex, Urannah Complex, and Camboon Volcanics, to be part of a broad magmatic system. We propose that these elements be collectively called the Connors-Camboon Province, to differentiate it from the structural entity that is the Connors-Auburn Arch. This episode spanned about 40 my, peaking at 305 Ma and ranging from the early Late Carboniferous, at about ~320 Ma, and continuing into the Early Permian until about ~280 Ma.

The single 343 Ma K-Ar age from the outlier of Torsdale Beds is thus the only evidence in the NNEFB for an Early Carboniferous magmatic episode that may be coeval with development of the recognised accretionary complex and fore-arc basin.

Coeval, but contrasting, magmatic suites dominate the Late Carboniferous at ~305 Ma. The interpretation of Day et al. (1978) that the rocks of the Connors-Camboon Province represent a continental margin arc is widely accepted. The igneous rocks of the Province are broadly calc-alkaline in character (Jones, 1994; Dear, 1994) and Allen (1994) argued that the batholith that underlies these volcanics was similar to other interpreted continental arc batholiths. These data are consistent with the notion that a north-south-trending magmatic arc of Late-Carboniferous to Early Permian age developed during this time. At the same time that calc-alkaline magmatic activity peaks in the Connors-Camboon Province a suite of syntectonic S-type granites overprint the older accretionary terranes and must represent a major crustal melting event associated with a mid-crustal deformation.

In the North d'Aguilar Block the metamorphic rocks and syntectonic granitoids are exposed in an extensional core complex. Metamorphic culminations containing similar syntectonic granites occur sporadically along the whole length of the orogen, from the Tia and Wongwibinda Complexes of the Late Carboniferous Hillgrove suite in the southern NEFB in NSW, to the Broome Head Metamorphics, and their correlatives described above in the Marlborough Block, at the northern end. We would suggest a common extensional origin for all, although this viewpoint is not widely accepted for the southern NEFB, where a thrust-related origin has been postulated (eg, Dirks et al., 1993). We have tightly constrained the timing of syntectonic intrusion in the North D’Aguilar Block at ~305 Ma, and a recent zircon age from the Tia Complex has given a preliminary age of ~300 Ma (Dirks et al., op cit.). The crystallization ages of the other complexes is currently unknown as most age dates have been derived by K-Ar dating and have been disrupted by more recent tectonic events. We believe that the widespread occurrence of ages at 300-305 Ma indicates a non-diachronous extensional event rather than with oblique ridge subduction and translation of a triple junction southwards along the orogen as has been proposed elsewhere (e.g., Murray et al, 1987; Fergusson et al., 1993).

The age range of the Connors-Camboon Province, ~320-280 Ma and peaking at ~305 Ma, is essentially identical to the range of ages recorded from the Bulgonunna Volcanics and associated granites/volcanic complexes to the west of the Bowen Basin and throughout north Queensland (Webb and McDougall, 1968; Branch, 1972; Oversby et al., 1980; Black et al., 1981; Allen pers. comm., 1996). Allen (in prep.) has correlated the Urannah and Bulgonunna Suites on the basis of similar geochemistry. The development of large scale calderas, batholith emplacement and ring complexes that characterises these events is characteristic of continental extensional environments and record a significant episode of crustal melting.

Significant factors bear upon any tectonic interpretation for this Connors-Camboon Province: 1) the present linear geometry of this Province is largely an artifact of later structuring and basin formation (Holcombe et al., this volume); 2) active extension and crustal melting occurred in the potential position of any forearc; 3) widespread continental crustal melting, probably accompanying extension, occurred to the west and north; and 4) there is no clearly recognised forearc complex rocks that may have developed during this time. Points 1-3 have been addressed previously in this paper, and point 4 is addressed below.

The Neerkol Formation and equivalents, in the Yarrol Province, comprise Late Carboniferous shallow marine sediments that are broadly coeval with both the Connors-Camboon suite and the forearc S-type magmatic terranes. Leitch et al. (1994) speculated that the Neerkol Formation could be a Late Carboniferous forearc basin, paired with the Shoalwater accretionary terrane, on the basis that: 1) units correlated with the Neerkol Formation on the coast east of Marlborough have a strong volcaniclastic component and 2) the Shoalwater terrane is possibly younger than the other coastal accretionary terranes based on its structural position. Although these sedimentary rocks now occupy a position that could have represented a Late Carboniferous forearc basin, they generally contain only a limited component of volcaniclastic detritus, inconsistent with the presence of a nearby active volcanic province. They are dominated by a shallow water marine mud facies with widespread, and characteristic, development of bryozoa implying quiet conditions; thus they are not typical of extensional basins that might be expected to develop within a forearc environment. The coastal rocks that Leitch et al. (op. cit.) describe are highly atypical of the Neerkol Formation regionally, and contain no fossils (elswhere a characteristic of the unit). Interpretation of the Neerkol Formation, and equivalents, is problematic and may be a key to future tectonic analyses.

We recognize two interpretations of these observations dependent on whether the Connors-Auburn-Urannah-Camboon Province is interpreted as a continental arc, or whether it is part of a broader province of crustal melting associated with continental extension:

• If the calc-alkaline rocks represent a magmatic arc, then tectonic scenarios must account for the coeval forearc extension and strong regional thermal perturbation. The implication is that an underlying fertile slab must persist below the arc and yet at the same time, magmatic production related to the downgoing slab must be overtaken by a regional thermal pertubation. In earlier papers (Little et al., 1993; Stephens et al., 1994) we have suggested that synchronous slab roll-back to the east, with extension driven by the trench-suction force, may have initiated just prior to ~ 305 Ma and, during retreat, hot mantle has been introduced beneath this region and induced mantle and crustal melting. The fact that arc magmatism does not migrate eastward with the rollback requires that melt production from the slab continues in situ and we interpreted this in terms of either a bent slab geometry (as for example in the present Tonga-Kermadec slab), or a detached slab remnant. In either case, the slab must ultimately become detached and sink into the mantle because arc magmatism continues in place and wanes slowly into the Early Permian (Camboon Volcanics time). Note that since we have linked the Early Permian Camboon Volcanics with the Late Carboniferous magmatism, we no longer require a separate discrete Permian arc episode ("Camboon Arc") as is commonly envisaged.
• If the Connors-Urannah-Auburn-Camboon Province is part of a broad extensional event then the isotopic compositions and calc-alkaline signature of the magmatism (Allen pers. comm. 1996) must be interpreted in terms of re-melting of relatively young crust that is calc-alkaline in composition. This model has been argued for Late Triassic and Lower Cretaceous volcanism in the NEFB (Stephens, 1991, Ewart et al., 1992; Stephens et al., 1993) but requires the existence of earlier calc-alkaline magmatism beneath the region. Contrary to the earlier interpretation of Allen (1994), the Urannah Suite is most likely related to extension at a time when subduction has slowed or stopped (Allen, pers.comm., 1996). Given the regional extensional signature of surrounding Permo-Carboniferous magmatic terrains, and the continuation of extension through the Early Permian (Fielding et al., this volume), we are gradually being drawn to the interpretation that the Connors-Camboon Province represents an extensional event that is not necessarily related in any respect to active subduction. We regard the position and nature of the plate boundary at this time as uncertain, although the foregoing discussion of the Beserker Beds suggests that increasingly marine conditions may have existed progressively to the east. The key to both tectonic and age interpretations of the Connors-Camboon Province lies in further, detailed research into the Connors volcanics.

The nature of the driving force for such a broad regional extension is also uncertain. Simple slab retreat, as suggested above, would seem to be insufficient as a mechanism for such a major crustal and thermal perturbation and is unlikely to have occurred synchronously over such a large length of subduction margin. This event, and the ensuing Early Permian extensional period, may mark a major tectonic event occurring along a passive margin. Certainly coeval events of the same magnitude, such as termination of the Alice Springs Orogeny in Central Australia, must be considered when determining the fundamental tectonic dynamics.

The Permian sedimentary units exposed to the east of the Connors Arch, within the Gogango Overfoded Zone were previously interpreted as having been deposited in a deep marine basin called the Grantleigh Trough that had opened within the forearc (e.g., Malone et al., 1966; Kirkegaard et al., 1970, Day et al., 1983). The Connors Arch has been widely regarded as representing the eastern margin of the depositional Bowen Basin (see Fielding et al., this volume for discussion).

Fergusson (1991) put forward a structural model suggesting that the Connors Arch is a structural high developed above an antiformal duplex stack in the basement rocks, and produced during the Permo-Triassic thrust event. The depositional Bowen Basin was shown continuous, but deepening, into the Grantleigh Trough. We agree broadly with this view and interpret the "Connors Arch" northwest of Rockhampton as the exposed core of a simple antiformal flexure that is the westernmost fold in a series of low amplitude regional folds extending to the east (Apis Creek Syncline/Strathmuir Synclinorium).

Fielding et al. (1994, this volume) argue on sedimentological and structural grounds that there is no need to invoke a separate, or deep water, sedimentary basin in order to explain observed regional stratigraphic variations, and indeed no evidence to support such an interpretation. A critical aspect of interpretations invoking the existence of the Grantleigh Trough has been the nature and age of the Rookwood Volcanics, addressed in this paper, and the internally, structurally complex Rannes Beds. The thickness of sediment originally estimated by Kirkegaard et al. (1970) for the Rannes Beds (3000 metres) is grossly exaggerated by structural thickening (Holcombe et al., this volume). In addition, the depositional characteristics of the less deformed Rannes Beds are identical to those of sediments in the adjacent Bowen Basin west of the GOZ. The sedimentary rocks of the Rannes beds are clearly not deep-water turbidites as stated by a variety of authors (Kirkegaard et al., 1970; Day et al., 1983; Fergusson, 1991), but are tectonically deformed shallow marine sediments. They are widely, and inappropriately, described in the literature as "flysch".

This paper incorporates the results of several independent ARC-funded projects (#A38830041; #A39130279; #A39331366; # A39232338) over the past ten years. We are indebted to the many students who have contributed as part of the various projects. In particular, aspects of the PhD work of Terry Harbort, Amanda Jones, and Paul Messenger and of the Honours work of Terry Harbort, Mac Denton, Vanessa Muscio, Murray Patterson, Shane O’Connell, and Eris O’Brien have added to our understanding of the region. Discussions over the years with Cec Murray, Russell Korsch, Chris Fergusson, John Draper, and others have refined our ideas. In a review paper like this it is likely that we have missed proper acknowledgement of some sources. Constructive reviews by Vince Morand and Peter Cawood have considerably strengthened the paper.

Table 1. Compilation of previously unpublished K/Ar isotopic data used in these studies, or data previously published only informally in conference abstracts. A correction has been applied to all data derived from the UQ lab between 1991 and 1995 and arises from a rounding error in data calculation. Corrected data are shown by *. The dated mineral is shown in brackets after the sample number (m = muscovite; wm = white mica; b = biotite; I = illite; p = plagioclase; k = potassium feldspar; a = amphibole; h = hornblende; wr = whole rock). References marked ^QUT are unpublished theses at the Queensland University of Technology; all other single author references are to unpublished theses at the University of Queensland. Thesis references are not repeated in the References for this paper unless they are specifically mentioned in the text.