Suckling Dome and the Australian–Woodlark plate boundary in eastern Papua: the geology of the Keveri and Ada'u Valleys

The Owen Stanley Fault Zone (OSFZ) is the low-angle thrust boundary between the Australian and Woodlark plates. The eastern extension of the OSFZ links with the Woodlark Basin spreading centre. Recent tectonic models of eastern Papua depict the OSFZ boundary passing through the Mt Suckling district, with the Keveri Fault a key component. Gravity data clearly show that the OSFZ and the Papuan Ultramafic Belt (PUB) pass north of Mt Suckling. Tectonised mafic and ultramafic rocks of the Mt Suckling district, previously referred to the PUB, are reassigned to the Awariobo Range Complex (new name). Extensive pillow basalts previously referred to the middle Eocene part of the Kutu Volcanics at the top of the PUB sequence are, in the map area, reassigned on lithological and biostratigraphic grounds to the late Oligocene–middle Miocene Wavera Volcanics. The detailed work reported here indicates that the Keveri Fault is unrelated to the OSFZ with no evidence for thrusting along the structure. The area's tectonic history has been dominated by large vertical displacements along the Keveri Fault. The commencement of late Miocene buoyant uplift of the Suckling Dome (new name), related to granite intrusion into thick crust of the eastern Papua region, marks the inception of the Keveri Fault and coincides with the initiation of Woodlark rifting. The fault facilitated much of the rapid vertical movement of the dome, with an estimated 8000 m of uplift (2.5 m/103 a) since the late Miocene. Movement on the Keveri Fault is notably different from structures flanking other metamorphic core complexes in eastern Papua. There is no field evidence for the development of a low-angle, south-dipping detachment fault along the southern margin of the Suckling Dome. The Suckling Dome is the westernmost of the eastern Papua domes, localised within a broad extensional zone that continues to propagate westward along the OSFZ plate boundary.


INTRODUCTION
The Owen Stanley Fault Zone (OSFZ) is an important tectonic element of the Papuan Orogen. It is the active plate boundary separating continental rocks of the Australian plate from ultramafic and mafic rocks of the Papuan Ultramafic Belt (PUB) on the Woodlark plate. The eastern extension of the OSFZ is thought to link with the Woodlark Basin spreading centre (Figure 1; Daczko et al. 2011;Little et al. 2011;Davies 2012;Fitz & Mann 2013). For much of its length on the Papuan Peninsula, the OSFZ consists of two or more faults. At the western and eastern ends of the fault zone, individual structures merge, but elsewhere they are separated by a fault slice some 5À10 km wide (Davies 1971;Daczko et al. 2011;Little et al. 2011;Fitz & Mann 2013). Individual shear zones may be between 100 and 500 m wide (Davies 1971). Long, straight and deeply incised valleys along the trace of the OSFZ, including the Waria River Valley, suggest a structure that is steep to vertical. Davies (1971) concluded that the OSFZ is a thrust, despite observation of only one low-angle dip (at locality 9 20ʹS), with younger vertical strike-slip normal faulting now coinciding with the thrust plane. geology and structure in the Keveri and Ada'u Valleys in the Mt Suckling district (Figures 2, 4a). The district is remote and difficult to access, and there have been no detailed published studies since the BMR reconnaissance mapping of 1965À1968 (Appendix). A well-developed sedimentary record in the map area is unlike anything else along the OSFZ and has been important in elucidating tectonic events.

TECTONIC SETTING OF EASTERN PAPUA Papuan Orogen
Rocks of the Papuan Orogen (Little et al. 2011) crop out in a 750 km belt underlying the Papuan Peninsula and islands of the Milne Bay district, including the D'Entrecasteaux Islands (Goodenough, Fergusson, Normanby), Woodlark Island and the Louisiade Archipelago (Misima, Deboyne, Rossel, Sudest, Engineer Group, Conflict Group). The orogen's geology documents a history of convergence and extension between the Australian plate and microplates including the Woodlark and Solomon plates during arcÀcontinent collision (Davies & Smith 1971;Smith & Milsom 1984;Hall 2002;Webb et al. 2008;Whattam et al. 2008;Whattam 2009;Daczko et al. 2011;Little et al. 2011;Baldwin et al. 2012;Fitz & Mann 2013). Elevations in excess of 2000 m are present on Goodenough and Fergusson islands and on the Papuan Peninsula (Mt Suckling-3676 m-is the highest mountain in Papua, and for comparison, is only 48 m below Aoraki/Mt Cook, South Island, New Zealand). Volcanism, commonly widespread, has occurred in four discrete episodes in the Papuan Orogen, viz: during the middle Eocene, late OligoceneÀmiddle Miocene, middleÀlate Miocene to Pliocene and Quaternary. The Quaternary volcanism is related to extension and sea-floor spreading in the Woodlark Basin. Present WSW convergence across the OSFZ has been estimated using GPS observations at »20 mm/a perpendicular to the fault zone north of 8 S (Wallace et al. 2004;Daczko et al. 2011).  Taylor et al. (1999). Plate motion vectors modified from Wallace et al. (2004). Simplified geological map from Davies (1980). (b) Major plates and microplates in eastern New Guinea from the plate compilation of the UTIG PLATES project (http://www.ig.utexas.edu/research/projects/ plates/).

PALEOCENE COLLISION-FORMATION OF THE OWEN STANLEY FAULT ZONE AND EMPLACEMENT OF THE PAPUAN ULTRAMAFIC BELT
Emplacement of late Mesozoic mafic and ultramafic rocks (PUB) indicates commencement of arcÀcontinent collision in the Papuan Orogen and the formation of the Australian plateÀWoodlark plate boundary (OSFZ). The PUB complex crops out over a length of 375 km. The ophiolite consists of a sequence of ultramafic (4À8 km thick), gabbroic (4 km thick) and basaltic rocks (4 km thick). A sheeted dyke complex is only rarely observed in the sequence (Davies 1971). There is no direct radiometric evidence for the age of the PUB with the age of ). The dashed black line represents the OSFZ as interpreted from gravity data (this paper). The coastline is shown as a fine white line. Gravity data indicate the PUB and the bounding OSFZ passes south of Mt Trafalgar (MT) and Mt Victory (MV) volcanoes and Cape Nelson and then along the southern margin of Collingwood Bay (CB) towards Cape Vogel Peninsula. The "embayment" structure (EB) in the OSFZ between 148 E and 149 E does not reconcile with geology (the subject of this paper); the inferred low-angle structure in this region is the Keveri Fault (KF), which is shown by inversion studies of detailed low-level airborne total magnetic intensity data to be vertical to at least 7À10 km depth. Much of the geology of the Mt Suckling district and the area immediately to the west, previously mapped as part of the PUB, is indicated by both geology and geophysics to be unrelated to the ophiolite complex. Other tectonic features include Normanby Island (NI) where the OSFZ is thought to link with the Woodlark Basin spreading centre Fitz & Mann 2013), and the D'Entrecasteaux Islands (DEI), Mt Suckling (MS) and Mt Dayman (MD), dome complexes. Coverage of the Keveri and Ada'u Valleys 1:50 000 geological map sheet (Figure 4a) is shown. the ophiolite constrained to ca 71À65 Ma (Late Cretaceous) by foraminiferal dating of fine sediments associated with basalt and Paleocene tonalite stocks and plutons that intrude the gabbroic part of the complex (Davies 2012

GEOPHYSICAL SIGNATURE OF THE PAPUAN ULTRAMAFIC BELT
The PUB is associated with strong, elongate and coherent gravity and magnetic anomalies (Figure 3; Davies 1980;GSPNG & BGS 2004). By contrast, there is no gravity anomaly associated with the ultramafic rocks of the Mt Suckling district (Figure 2), suggesting they are unrelated to the PUB. East of 148 30ʹE, the gravity data are clear in indicating that the PUB and the bounding OSFZ passes south of Mt Trafalgar and Mt Victory volcanoes and Cape Nelson, before passing offshore into Collingwood Bay ( Figure 3). Extension of the OSFZ east of Collingwood Bay into the Goodenough Basin is uncertain and may be via a jog structure(s) north of and around Cape Vogel Peninsula to connect with the Goodenough Fault, one of the most active faults of the Woodlark rift ( Figure 3; Little et al. 2011). Gravity data also indicate that the easternmost limit of the PUB, as a coherent intact body, is 149 30ʹE (Figure 3). Fitz & Mann (2013) interpreted the presence of Paleocene gabbro and diabase recovered by ODP Leg 180 drilling on Moresby Seamount, NE of Normanby Island, as evidence for the eastern extension of the PUB (Figures 1, 3).

MIDDLE EOCENE EXTENSION-KUTU VOLCANICS
Middle Eocene pillowed basalt lava, with interbedded sparsely fossiliferous, fetid, flaggy deep-water limestone (Kutu Volcanics), widespread on the Papuan Peninsula and some islands, was erupted from numerous volcanic centres and marked the cessation of convergence and the commencement of a period of crustal extension. Numerous narrow (2À3 m) vertical gabbro dykes typically intrude basalt and limestone units (e.g. Pini Range quarry exposures, 150 16ʹE/10 22ʹS; Lindley 1991). These gabbro dykes are mineralogically similar to gabbros of the middle Eocene? East Cape Gabbro (150 45ʹE/10 15ʹS; Smith & Davies 1973a) and the late EoceneÀmiddle Oligocene Sadowa Intrusive Complex (Yates & de Ferranti 1967), suggesting the intrusive bodies may be subvolcanic plutons. The Sadowa gabbro is a large elongated (120 km £ 30 km) pluton extending from 147 15ʹE to 148 15ʹE and consists of olivine gabbro, pyroxe-neÀhornblende diorite, granophyre and ultramafic rocks (Yates & de Ferranti 1967;Pieters 1978;Rogerson et al. 1981). The pluton was emplaced near the margin of the Australian plate and is subparallel to the OSFZ. Both northern and southern contacts of the pluton are either fault controlled or parallel the regional strike of sedimentary units (Yates & de Ferranti 1967). The proposed genetic link between the East Cape Gabbro and Sadowa Intrusive Complex plutons, and the Kutu Volcanics suggests crustal extension and the formation of a magmatic arc in the middle Eocene. Middle Eocene extension is counter to the view of other authors who proposed plate convergence and collision continued into the early Miocene (summarised in Little et al. 2011, p. 43).

LATE OLIGOCENEÀMIDDLE MIOCENE TECTONIC STABILITY
Volcanic and tectonic quiescence prevailed throughout most the Papuan Orogen during the late OligoceneÀmiddle Miocene. Limestone, marl, tuffaceous sandstone and siltstone units with minor limestone (Ada'u Limestone, Woruka Siltstone, Castle Hill Limestone; Davies & Smith 1974;Modewa River Beds;Smith & Davies 1973a) were deposited in shallow seas across the Papuan Peninsula. Several submarine eruptive centres contributed pillowed basalt lavas on Cape Vogel Peninsula (CVP, Figure 1; Dabi Volcanics; Davies & Smith 1974) and in the Mt Suckling district (Figure 2; Wavera Volcanics, this paper). Localised occurrences of andesitic lava, agglomerate, volcanolithic conglomerate and arenite, with dense grey limestone, crop out on Normanby Island (151 00ʹE/10 00ʹS; Sewa Beds; Smith & Davies 1973a). The problematic Astrolabe Agglomerate (no younger than 5.7 Ma; Pain 1983) is a coarsely stratified, clast-supported deposit of angular and subrounded basalt with intercalated thin lenses of lithic tuff containing current and graded bedding. The formation crops out on the Sogeri Plateau and the Astrolabe Range, north of Port Moresby. Casts of tree trunks have been found in both fine and coarse units, and the deposit has no identified volcanic source (Ollier 1969;Pieters 1978;Pain 1983).

RIFTING AND THE FORMATION OF THE WOODLARK SPREADING CENTRE
Woodlark Basin rifting probably commenced in the late Miocene (Taylor et al. 1999) and continues to the present, splitting the continental crust of eastern Papua (Little et al. 2011). As rifting has continued, the Woodlark spreading centre ridge has propagated westward >500 km into the rift (Little et al. 2011).

EXTENSIONAL TECTONICS
Three graben-like depressions are located at the eastern end of the Papuan Peninsula (Smith & Milsom 1984). Mullins Harbour is considered to be the oldest graben and is almost entirely filled by sediment (Smith & Milsom 1984). Milne Bay is an eastÀwest-oriented graben varying in width from 12 to 14 km. It has subsided 775 m in the last 18 000 years with the accumulation of 450 m of sediment (Jongsma 1972). The youngest graben is the Goodenough Basin (GB, Figure 1). All three basins are thought to be the result of rift propagation associated with the Woodlark Basin rifting (Smith & Milsom 1984). The Goodenough Fault, bounding the southern edge of the Goodenough Basin, has estimated slipping rates of 10À20 mm/a (Little et al. 2011). A Quaternary peralkaline rhyolite suite is presently being erupted from volcanic centres at the eastern end of Fergusson Island and on Dobu Island (Smith 1976). These peralkaline rocks are characteristic of a tensional environment (Smith 1976). The post-middle Miocene evolution of the Papuan Orogen is characterised by two phases of magmatic-volcanic activity: (1) middleÀlate Miocene to Pliocene magmatism and volcanism; and (2) Quaternary volcanism related to extension and sea-floor spreading in the Woodlark Basin. These episodes are separated by a Plio-Pleistocene period of volcanic and tectonic quiescence (Smith & Milsom 1984). MiddleÀlate Miocene to Pliocene volcanic rocks include the Cloudy Bay Volcanics (Smith & Davies 1973b), Fife Bay Volcanics, Normanby Volcanics (Smith & Davies 1973a) and Amphlett Volcanics (Davies 1973 1990). The volcanic activity is thought by many workers to have been associated with the development of the Trobriand Trough during the Miocene and magmatism generated by the southward subduction of the Solomon plate Little et al. 2011). This is despite the Trobriand Trough being an aseismic structure .

METAMORPHIC CORE COMPLEXES AND QUATERNARY VOLCANISM
Metamorphic core complexes on the D'Entrecasteaux Islands (Goodenough, Mailolo, Oitabu, NW Normanby Domes) and the Papuan Peninsula (Dayman Dome) have heights in excess of 2000 m and are being exhumed at rates of >20 mm/a (Ollier & Pain 1980Hill et al. 1992;  Geological, geophysical and platinum group metal (PGM) geochemistry observations indicate that "ophiolite" in the Mt Suckling district is not part of the PUB: (1) The Awariobo Range Complex does not have a sheeted dyke zone or a basalt zone, typical of ophiolite complexes such as Troodos Complex of Cyprus.
(2) There is no metamorphic sole to the complex (cf. Worthing 1988). The Awariobo Range Complex is noteworthy for PtCPd enrichment, with 258 unmineralised rocks having PtCPd enrichment is associated with layered mafic deposits formed in intracontinental tectonic settings (e.g. the Bushveld or Great Dyke), with thick crust playing a key role in facilitating enrichment (W. D. Maier pers. comm.), and arc-type layered igneous complexes (Christie et al. 2006). Although there is an apparent association of PGM mineralisation with chromite, typical of the Bushveld and Great Dyke, the Awariobo Range Complex is interpreted to have formed in a magmatic arc setting.
Tectonite ultramafic rocks comprise more than 95% of the ultramafics mapped in the Awariobo Range Complex (Smith & Davies 1976). Cumulate ultramafic rocks were noted by Smith & Davies (1976) to be restricted to a small outcrop east of Dori Creek in the upper Ada'u Valley. Common tectonite ultramafic lithologies include harzburgite, peridotite and orthopyroxenite. Serpentinisation of these rocks is common with anastomosing and wispy bands of serpentinite imparting an obvious waxy appearance to rocks. Tectonite fabric and serpentinisation of ultramafic rocks may be related to tectonisation during the rapid uplift of the Suckling Dome. The presence of wehrlite is considered notable by Smith & Davies (1976) as this rock type is not known from the PUB.
A distinctive dunite comprising vitreous olivine crystals with disseminations of fineÀcoarse chromite is sourced only from two adjacent tributaries of upper Dimidi Creek (Bilia'e No. 1 and No. 2 streams), suggesting a localised dunite body or pipe-like feature with an estimated diameter of 1000À1500 m (Kd: Figures 4a, b). Dunite is associated with thick chromite seams, with chromitite float boulders and individual boulders of massive chromite up to 50 cm (attesting to the thickness of seams) present in Dimidi Creek. Individual chromite seams may show alteration selvages top and bottom ( Figure 6f). Arumba (1993, plate 1a) described an apparently localised chromitite occurrence from the Sadowa Intrusive Complex at Mt Lawes, NE of Port Moresby that is visually similar to the Dimidi Creek chomitite. Specimens from both localities have lensoidal chromite seams (with distinctive pale alteration selvages) set in vitreous, coarsely crystalline dunite. Rhythmic layering of chromite seams, graded sedimentary settling, pinching of chromite seams and sedimentary scour-and-fill structures indicate that currents swept the floor of the dunite pipe, in what was a relatively stable crustal region.
Gabbroic rocks appear to be restricted to the Doriri Creek area. Reconnaissance mapping summarised in Smith & Davies (1976) suggested gabbro was present in a 1 km-wide strip along the southern edge of the Awariobo Range in the Ada'u Valley. Based on an assumed dip of 45 N, they calculated a thickness for the gabbro zone of 0.5 km. My mapping in the Ada'u Valley in consecutive streams including (west to east) Doriri Creek, Oiso Creek, Urua Creek and Dimidi Creek does not indicate simple layered ultramafic/gabbro geology. At Doriri Creek, norite with orthopyroxenite and minor dunite is host to the Doriri NiÀPGM lode (Gonz alez-Alvarez et al.

2013), whereas in nearby Oiso
Creek and Urua Creek the lowermost 2À3 km intervals of each of these streams contain outcrops of pillow basalt of the Wavera Volcanics ( Figure 4a). Peridotite with minor orthopyroxenite crops out along the entire length of Dimidi Creek.
There is no radiometric evidence for the age of the Awariobo Range Complex. Late MioceneÀearly Pliocene KÀAr ages for the Suckling Granite and Mai'iu Monzonite, which intrude the complex, help constrain the unit's age to the Eocene based on similarities with the East Cape Gabbro and the Sadowa Intrusive Complex.

GOROPU METABASALT (Kg)
The Goropu Metabasalt (Smith & Davies 1976) is exposed in the upper Ada'u River in a fault slice between the Nonia Fault and the Keveri Fault. The formation may be intruded by ultramafic rocks of the Awariobo Range Complex east of the Nonia Fault in Ioleu Creek and Bonua River (east of the map area). Interpretation of detailed low-level total count radiometrics with field observations has been useful in distinguishing the Goropu Metabasalt from ultramafic rock of the Awariobo Range Complex and pillow basalt of the Wavera Volcanics. Metamorphosed basalts are typically massive, dense black rocks lacking any structure (e.g. pillows, columnar jointing). Metabasalt in Ioleu Creek is strongly jointed and locally intensely sheared, and includes finegrained microdiorite. The Goropu Metabasalt is assigned a Late Cretaceous age using planktonic foraminifera from a calcareous schist member (Bonenau Schist Member) mapped by Smith & Davies (1976) to the NE and east of the map area near Mt Suckling and Mt Dayman, respectively.

OWEN STANLEY TERRANE
The Owen Stanley Metamorphics (Ko) typically consists of low-grade greenschist pelitic and psammitic metasediments and subordinate metavolcanics (Pigram & Davies 1987). Although the formation is not present in the map area, it crops out only 2 km to the west in Pigleg Creek, a tributary of the upper Domara River (Figure 2; Lindley & Tamu 1988). From here, the formation extends as a fault slice to the NW along the south bank of the Domara River ( Figure 2; Davies & Smith 1974). Outcrop at Pigleg Creek consists of pelitic metasediments (phyllite and mica schist) and metavolcanics (amphibolite). Quartz veins crosscut the greenschist facies rocks and contain pyrite and chalcopyrite. The Owen Stanley Metamorphics is largely unfossiliferous, and the age of the formation is poorly constrained (Pigram & Davies 1987). The formation is assigned a Cretaceous age (Smith & Davies 1976).
Late Oligocene to Pliocene Davies & Smith (1974) and Smith & Davies (1976) used limited reconnaissance traverses and aerial photographic interpretation to map an Eocene calcilutite and limestone unit (Godaguina Beds) restricted to the headwaters of the E'au River (D Godaguina River) and in Urua Creek (Smith & Davies 1976, p. 25); the unit was interpreted as a lenticular body or bodies within the Kutu Volcanics (Smith & Davies 1976).
Float from the upper Ea'u River type area of the Godaguina Beds was inspected and on lithological grounds is referred to the late OligoceneÀmiddle Miocene Wavera Volcanics. This is in accord with Macnab's (1967) mapping of these rocks as the Mount Clarence Calcilutite Member of the Wavera Volcanics. The Godaguina Beds are considered redundant. Macnab (1967) originally proposed the Wavera Volcanics for predominantly marine volcanic (pillow basalt) rocks cropping out in the headwaters of the Wavera River, east and west of the Keveri Valley and on the north and south fall of the Main Range. He nominated the Wavera River headwaters (148 45ʹE/9 55ʹS) as the type area. Macnab (1967) also mapped and described two limestone members within the Wavera Volcanics (Ada'u Limestone Member and Mount Clarence Calcilutite Member). Macnab (1967) assigned an early Miocene age to the Wavera Volcanics. In contrast, Davies & Smith (1974, p. 33) considered the Wavera Beds [sic.] to represent the Eocene part of the Kutu Volcanics with their map assigning all the pillow basalt sequence south of the Keveri Fault to the Upper CretaceousÀEocene Kutu Volcanics. The Ada'u Limestone is a readily mappable unit and Davies & Smith (1974) elevated the limestone to formation status. With additional foraminiferal information available to Davies & Smith (1974), the age of the Ada'u Limestone was amended to earlyÀmiddle Miocene. Davies & Smith's (1974) revised status for the Ada'u Limestone has been adopted herein.
The Wavera Volcanics are recognised north of the Ada'u River and the Keveri Fault in the area between Doriri Creek and Urua Creek, and east from the Ea'u River into the Liba River catchment (Figure 4a), a significant difference from the mapping of Davies & Smith (1974). Good continuous exposure of the Wavera Volcanics is present along Oiso Creek and Urua Creek (Figures 4a,  5a). The volcanic formation has a distinctive total count radiometric signal, readily distinguishing it from mafic and ultramafic rocks of the Awariobo Range Complex and adjacent formations (such as the Domara River Conglomerate). The Ada'u Limestone crops out as an 11 km-long fault sliver bounded to the north by the Keveri Fault and an unnamed fault to the south (Figure 4a). It underlies a prominent ESE-striking range of hills, with distinctive pinnacles and karst features especially to the west of old Ba'u village (Figure 5c). Shallow south dipping to vertical dips indicate localised tectonic disruption of the formation. The Mount Clarence Calcilutite Member crops out as a single stratigraphic horizon along the north fall of Mt Clarence (Macnab 1967), SW of the map area, and in the upper Godaguina Valley.
The Wavera Volcanics consist of fine-grained, reddish and dark brown basalt. Pillow structure is common in many outcrops throughout the Wavera, Ada'u, Liba, Amin and Ea'u valleys (Figure 5a, b). Terrestrial volcanic rocks, which include accretionary lapilli tuff, vesicular basalt and amygdaloidal basalt, have been identified in the Ada'u Valley at Urua Creek and Waki Creek. These units include accretionary lapilli tuff, vesicular basalt and amygdaloidal basalt. Outcrop may be fractured and jointed with calcite and zeolite in-fillings. The formation is occasionally indurated, especially in the Ea'u River. The Ada'u Limestone consists of massive, creamycoloured limestone that is locally fossiliferous. G. C. H. Chaproniere (in Lindley et al. 2010b) described the limestone in Urua Creek as a moderately recrystallised bioclastic packstone dominated by shallow-water bioclasts. The recrystallisation is thought to be associated with the mingling of limestone with hot basalt pillow lava.
The Wavera Volcanics unconformably overlies rocks of the Awariobo Range Complex north of the Keveri Fault. South of the Keveri Fault in the Pigleg Creek area, the formation unconformably overlies phyllite of the Owen Stanley Metamorphics (Lindley & Tamu 1988). The volcanic rocks have an estimated thickness >730 m based on the difference in altitude between the highest and lowest exposures of the apparently horizontally bedded unit in the unfaulted block in the Liba River area. The base of the Ada'u Limestone is not exposed. The formation has a minimum 440 m thickness based on the topographic limits of exposure of gently south-dipping beds west of old Ba'u village.
Breccias formed during mingling of pillow lavas of the Wavera Volcanics and the Ada'u Limestone are common in outcrop and float in the Wavera, Ea'u, Amin and Ada'u valleys (see Figure 4a for occurrences), but neither Macnab (1967) nor Davies & Smith (1974) (Figure 5e). Field-rock relationships and biostratigraphy clearly indicate that the Wavera Volcanics and Ada'u Limestone are lateral equivalents. Foraminiferal biostratigraphic data indicate a late OligoceneÀmiddle Miocene age for both formations.
The distribution of basaltÀlimestone breccias about a topographically anomalous 2 km 2 plateau, consisting of steep-sided semi-circular depressions filled with basalt scree and swamps, semicircular drainage patterns and basalt scree-covered steep escarpments on the Main Range between Waki Creek and Amin River (148 50.5ʹE/ 9 58ʹS), suggests this area may have been an eruptive vent for the Wavera Volcanics.

DOMARA RIVER CONGLOMERATE (Tpd)
The Domara River Conglomerate crops out extensively in an approximately 400 km 2 area to the west of the map area in the Domara River, Foasi River and Musa Valley (Figure 2; Smith & Green 1961;Macnab 1967;Davies & Smith 1974). The conglomerate formation underlies much of the Keveri Valley, forming a distinctive topography of low hills moderately dissected by numerous small streams. The unit does not extend north of the Keveri Fault (Figure 4a). The Domara River Conglomerate has a particularly distinctive intense total count radiometric airborne geophysical signature, making it easy to identify the outcrop limits of the formation. The formation in the Keveri Valley is a consolidated, coarse, poorly sorted polymictic conglomerate, with minor sandstone and grey, weakly consolidated mudstone noted in Wavera River outcrops (Figure 6aÀc). Conglomerate clasts include fine-grained mafic volcanic rocks, gabbro, monzonite, granite and harzburgite, apparently derived from erosion of the Awariobo Range Complex and other units to the north of the Keveri Fault. Clast size ranges up to 0.5 m. Clasts are set in a red-purplish lithic matrix, typical of non-marine deposition.
Rapid along-strike lateral facies changes are a feature of the formation in the Musa Valley (Smith & Green 1961). Macnab (1967) also noted cyclical deposition of conglomerate, sandstone and mudstone, probably a response to periodic source area uplift facilitated by movement along the Keveri Fault. Smith & Green (1961) and Macnab (1967) both noted carbonised wood up to 1 m long in poorly sorted boulder conglomerate. Smith & Green (1961) collected four freshwater molluscan faunas (thin-shelled gastropods and occasional pelecypods) in calcareous mudstone in the Foasi River and Ikumu River about 20 km NW of the Keveri Valley. N. H. Ludbrook (in Smith & Green 1961) studied the collections and concluded a probable Pleistocene age for the Domara River Conglomerate. Ludbrook also concluded that the formation was deposited in lacustrine and piedmont environments. Macnab (1967) also collected freshwater thin-shelled gastropod faunas from thin carbonaceous and calcareous beds in the upper Domara River but none proved useful for dating.
Smith & Green (1961) mapped three basalticÀandesitic volcanic members near the base of the Domara River Conglomerate in the Musa Valley. They also noted interbedded lava flows at several localities in the Domara River and Foasi River and the presence of porphyritic andesitic dykes. Macnab (1967) also noted the basal volcanic phase and rare lava flows and/or sills. Davies & Smith (1974) used KÀAr ages of a related dyke (Bonua Porphyry?-5.3 Ma) and a whole-rock determination on a volcanic clast (2.36 Ma) to conclude that the Domara River Conglomerate is Pliocene. Thus, paleontological and radiometric dating suggests the Domara River Conglomerate is of late Pliocene age, probably extending into the Pleistocene.
The Domara River Conglomerate is typically flat lying to gently dipping in the Keveri Valley. However, moderate 40À45 S dips are typical of outcrop adjacent to the Keveri Fault (Figures 6a, 8b) with tilting of these rocks probably related to drag on the Keveri Fault. Similar moderately south-dipping beds are also very obvious in the Domara River area to the west of the map sheet (

SUCKLING GRANITE (Tmk)
The Suckling Granite crops out in two stocks at the NE corner of the map area, about 5À10 km south and SW of the summit of Mt Suckling. These stocks cover 20 km 2 and include medium-and coarse-grained granite and adamellite (Davies & Smith 1974). The granite intrudes ultramafic rocks of the Awariobo Range Complex. The Sucking Granite is dated as late MioceneÀearly Pliocene by KÀAr determinations on hornblende (10.8, 9.43 Ma) and biotite (3.32, 3.24 Ma) (Davies & Smith 1974). Wholerock analyses of the Suckling Granite indicate it to be calc-alkaline (Smith & Davies 1976).
The unroofing of Suckling Granite stocks in the summit area of Mt Suckling is evidence for very rapid rates of late PlioceneÀQuaternary uplift and erosion for the Suckling Dome. Granites are rare in Papua New Guinea. Granite crops out on Mabaduan Hill near the southern coast of Western Province and on nearby small islands of the Torres Strait, and decomposed and light grey to white granite was intersected in oil wells at depths of 1988 m and 2971 m in the Aramia River and Aworra River, respectively, both in Western Province (APC 1961). These occurrences are part of the crystalline Australian plate basement. The Mai'iu Monzonite crops out in the NE corner of the map area. Davies & Smith (1974) mapped a single large stock of 150 km 2 in size SE of Mt Suckling (Figure 2). The monzonite stock consists of xenolithic granodiorite, biotite monzonite and biotite hornblendite (Davies & Smith 1974). The monzonite is dated as late MioceneÀearly Pliocene on the basis of KÀAr age determinations on hornblende (6.26, 6.03, 4.37 Ma; Davies & Smith 1974). The stock intrudes ultramafic rocks of the Awariobo Range Complex and, east of the map area, the Goropu Metabasalt (Davies & Smith 1974). Whole-rock analyses of the Mai'iu Monzonite indicate it to be calcalkaline (Smith & Davies 1976).
Two monzonite stocks intersected in deep drilling beneath diatreme breccia at Urua prospect, 21 km SW of the Mai'iu Monzonite stock, are considered related. Two distinct monzonite bodies were intersected in drilling, one an unaltered even-grained monzonite, the other an intensely epidote-altered monzonite with AuÀCu anomalous veins and stockwork veinlets of chalcopyriteÀ pyriteÀmagnetite. An unaltered, even-grained grey monzonite dyke was intersected during drilling of the footwall of the Doriri NiÀPGM lode.   (Figure 4a). The Urua and Araboro diatreme breccias are associated with distinctive bright-red airborne radiogenic thorium anomalies. Interpretation of thorium imagery suggests another diatreme body is present 1.5 km SE of the Urua diatreme.
A swarm of NNE-trending porphyritic felsic dykes is present in Dimidi Creek, the largest measuring 300À500 m wide with a strike length of 2500 m (Figure 4a). These dykes have strong airborne radiogenic potassium anomalies. The dykes lie within the NNE-trending trans-island Dimidi structural trend, which passes along the length of Dimidi Creek.
Several small gabbro stocks (Tg) have been mapped at Urua Creek and Ioleu Creek. Mafic dykes have been mapped in the Ada'u River upstream of Doriri Creek where they intrude ultramafic rocks of the Awariobo Range Complex and, as noted above, post-date diatreme rocks at Urua. These dykes appear to be related to faults.

QUATERNARY
Quaternary deposits are widespread in lower elevations to the west and NW of the map area ( Figure 2) and include the Ibau Breccia, Silimidi Conglomerate (with the Sivai Breccia Member) (all Pleistocene) and the Ubo and Wakioki fanglomerates (Holocene) (Smith & Green 1961;Davies & Smith 1974). Ultramafic breccia equivalents of the Ibau Breccia are recognised in the map area and floodplain, and terrace deposits of gravel, sand and mud are a feature of the broad, braided Ada'u River Valley. Gold-bearing gravels of the old Keveri Goldfield at Paiwa village are recognised as a new unit.

ULTRAMAFIC BRECCIA DEPOSITS (Qu)
Ultramafic breccias in the map area are restricted to the Ada'u Valley and the upper slopes of Mt Suckling. They are typically consolidated coarse to very coarse-grained and chaotically sorted deposits (Figure 7b). Clasts are angular to subangular, may be up to 0.75 mÀ1.0 m in size and are dominated by ultramafic rock. Breccias on the upper slopes of Mt Suckling have a sheet form. Valley-confined breccias have been mapped in Urua Creek and are particularly well exposed along the length of Dimidi Creek (Figure 4a). In some outcrops, cycles of sorting of very coarse breccia with mediumÀcoarse breccia are obvious and attest to pulsating phases of uplift and vigorous erosion (Figure 7a). Interbedded planarbedded sandstone is also present in the Dimidi Creek breccias.

ALLUVIAL TERRACE AND FLOODPLAIN DEPOSITS (Qa)
Floodplain and gravel deposits of the Ada'u River and its northern tributaries are dominated by poorly to well- sorted gravels with sand and mud. The Ada'u River is a typical braided stream with active and inactive channels developed on a floodplain up to 400À500 m wide (Figure 8a). Voluminous quantities of gravel are sourced from the numerous northern tributaries of the Ada'u River that drain the rugged Awariobo Range (and flanks of the Suckling Dome), where elevations reach 2542 m (a local relief of 2100 m). These tributaries have very steep gradients and short fetches. Rainfall tends to be localised and is characterised by extreme precipitation over a short time period. Accordingly, stream flows rapidly fluctuate, and floods are of short duration.
Older terraces are preserved at progressively higher levels above the Ada'u floodplain along the entire length of the river. Terraces are particularly prominent on the north bank of the Ada'u River at Doriri Creek, immediately downstream of Uruna Creek and the Urua Creek flats near old Ba'u village.

LANDSLIDE DEPOSITS
Landslide deposits are widespread on the over-steepened slopes of the Awariobo Range, particularly in the upper reaches of the Urua and Dimidi valleys and are too numerous to be shown on the geological map ( Figure 4a).

PAIWA GRAVELS (NEW NAME; Qp)
The Paiwa Gravels crop out at Paiwa village on the lower Wavera River. It was at this locality that the Keveri Goldfield was proclaimed on 6 August 1904, following Frank and Dan Pryke's discovery of gold in 1902. Most of the mining activity had subsided by 1907, and since 1926 there has been little recorded production from the field. The field produced a total of 5903 ounces of gold most of which was recovered during 1904À1905. The resurgence in the gold price during the past 10 years has seen the reoccupation of Paiwa village and a resumption of alluvial mining by Keveri landowners from Amau village using portable gold dredges. Supplies are carried in from the south coast along the Babauguina track (Figure 4a).
The Paiwa Gravels consist of a 0.5 km 2 area of perched Holocene gravels between Paiwa village and Oiyaku Creek that are unconformable on the Domara River Conglomerate and the Wavera Volcanics (Figure 4a). The unit consists of unconsolidated coarse, poorly sorted, matrix-and clast-supported gravels; clasts range to 0.25 m size and are sub-to well rounded (Lindley 1988). The fineÀmedium-grained matrix is weakly consolidated with blueÀgrey clay an important but variable component of the matrix. The gravels contain a distinctive coarse (1À3 mm) detrital mineral assemblage of hematite pseudomorphs, augite and hornblende with gold, allowing ready recognition of the gravels from the underlying Domara River Conglomerate (Lindley 1988). The average thickness of the Paiwa Gravels is 4 m (Lindley & Tamu 1988). Locally overlying the Paiwa Gravels, iron-cemented scree of basaltic gravels and clay, approximately 8À9 m thick, are derived from weathering of the Wavera Volcanics, probably related to another phase of Holocene uplift along faults immediately south of Paiwa village.
The Paiwa Gravels were deposited in an alluvial fan bordering the frontal ranges of an actively uplifting source area of Wavera Volcanics. Variations in the unit's distinctive detrital mineral assemblage, clay content and coarseness indicate a source area immediately SW of the Keveri Valley. The presence of gravels with a clay-rich matrix indicates poor sorting and rapid sedimentation. The coarseness of the brittle detrital augite and hornblende grains indicates source-area proximity. The composition of both the detrital assemblage and gravel clasts indicates a basalt volcanic provenance (Wavera Volcanics) with minor diorite intrusive rock (Bonua Porphyry). An unusual detrital assemblage in the gravel fan north of Oiyaku Creek suggests the development of a separately sourced gravel sheet adjacent to the main Paiwa basin. The northern extent of this fan is unknown.

Inversion modelling of total magnetic intensity data
The Keveri Fault, Nonia Fault (new name) and several unnamed flanking faults of the Wavera Valley cross the map area. A 2010 helicopter-borne detailed low-level airborne geophysical survey covered an 8 km swathe along a 30 km interval of the Keveri Fault (Figure 9a). Inversion modelling of total magnetic intensity (TMI) data is a computer-based analysis aimed to focus on deeper magnetic bodies (up to 7À10 km depth) in preference to smaller shallower sources. Surface traces of significant faults and formational boundaries are superimposed on the inversion model (Figure 9b). The traces of the Keveri and Nonia faults when projected vertically closely coincide with the boundaries of deep magnetic bodies identified by modelling, indicating these faults are vertical.
Keveri Fault (Synonymy: Keveri Fault System; Davies & Smith 1974) The Keveri Fault is marked by the aligned streams and broad valley floors of the braided Ada'u River and the Domara River. The fault's trace in the Ada'u Valley is marked by an abrupt step in elevation immediately north of the river. The land surface rises sharply to the 3676 m high Mt Suckling. Truncated ridges and terraces in Holocene gravels in the Ada'u Valley and the deflection of drainage channels in Holocene gravels (e.g. the Ada'u River channel and Dimidi Creek) also mark the fault and indicate dextral strike-slip movement (Figure 8a, c). Fault-bounded blocks of the Wavera Volcanics and the Ada'u Limestone occur in the trace of the fault, especially in the Urua Creek area. Tremolite within yellowish serpentinite boulders in I'eve Creek at the western end of the Keveri Valley (148 43ʹE/9 51ʹS; Figure 4a) indicate localised high temperature and pressure conditions along the fault.
The Keveri Fault has been interpreted as a thrust fault, at least during its early history, based one or more of the following: (1) theoretical grounds; (2)  Uplift rates on the Keveri Fault can be estimated from the Suckling Granite, which crops out near the summit of Mt Suckling. The granite is dated between 9.4 and 3.3 Ma and, assuming a depth of emplacement of around 4 km (Lynn et al. 1981), an uplift rate along the Keveri Fault of 8000 m during the past 5 Ma (2.5 m/10 3 a) is indicated. High rates of erosion associated with the uplift of Mt Suckling are manifest by braided streams, alluvial terraces, ultramafic breccias, and landslide and scree deposits. For comparison, rates of uplift along the Alpine Fault in New Zealand are estimated at between 9000 and 14 000 m during a 6 Ma period (1.5À2.3 m/10 3 a; Suggate 1963). Moderately tilted, south-dipping units of the Domara River Conglomerate adjacent to the Keveri Fault are a response to drag on the fault (Figures 6a, 8b). Strike-slip rates can be estimated from the dextral offset of the Dimidi Creek channel at its confluence with the Ada'u River (Figure 8c). If the 300 m offset occurred during the Holocene (past 0.01 Ma), the Keveri Fault has slipped dextrally at a rate of 30 m/10 3 a. A similar rate of movement is also obtained for the dextral offset of the Ada'u River channel at 148 47ʹE/9 52.5ʹS.
The map area has experienced low levels of seismicity (Brooks 1965;Ripper & Letz 1991;ISC 2013), but in the period 1964À1989 two shallow (0À39 km depth) M5.0À5.9 events originated within the map area (Ripper & Letz 1991; Figure 4g). Both events plot along the Keveri Fault indicating the fault remains active and may present a seismic hazard.

Nonia Fault (New name)
The Nonia Fault is a prominent northern splay of the Keveri Fault. The fault has a length of 21 km, and its trace is obvious on total count radiometric imagery. For much of its length, it separates Goropu Metabasalt from ultramafic rocks of the Awariobo Range Complex. At the fault's western end, it forms the northern truncated margin of the Urua diatreme, indicating post-early Pliocene movement (Figure 4a). In Dimidi Creek, the lack of any significant lateral offset of the NNE-striking felsic dyke swarm suggests post-early Pliocene vertical movement along the Nonia Fault.
A 16 m wide serpentinite zone is exposed within the Nonia Fault in Dimidi Creek (Figure 10). Wallrock to the zone is a weakly serpentinised peridotite, with massive/ blocky serpentinite with brittle fracture, sheared serpentinite and talc zones mapped (Figure 6d, e). Large blocks of weakly serpentinised pyroxenite, interpreted as tectonic inclusions, are present in the fault zone. The serpentinite zone carries anomalous Ni, Cr and Pt contents.

Dimidi Trend (New name)
The Dimidi Trend is a previously unrecognised transisland structure (Figure 2)

Other faults to the south of the Keveri Fault
Keveri Fault-parallel structures are developed along the southern edge of the Wavera Valley (Figure 4a). The formation of a succession of intermontane basins was a response to uplift of the Main Range (south of the Wavera faults) and the Suckling Dome (north of the Keveri Fault). Piedmont and lacustrine sediments filled the intermontane basins (late PlioceneÀearly Pleistocene? Domara River Conglomerate; Holocene Paiwa Gravels). Well-developed horizontal slickensides exposed in the Wavera River, 1.5 km south of old Mar e village, indicate that strike-slip movement also occurred along the Wavera Valley faults. Post-depositional strike-slip offset of the Domara River Conglomerate in the map area suggests some of this dextral strike-slip movement has occurred since the early Pleistocene.

SYNTHESIS AND CONCLUSIONS
Suckling Dome (New name) BUOYANT UPLIFT MODEL Ollier & Pain (1980 proposed a model of buoyant uplift triggered by granite pluton emplacement in regions of thick crust undergoing extension to account for the origin of eastern Papuan metamorphic core complexes including the Dayman Dome. Their model readily accounts for the geology and tectonics of the Suckling Dome. The presence on the summit of the Suckling Dome of granite (Suckling Granite) and monzonite (Mai'iu Monzonite) stocks (Figures 2, 4b) that have intruded strongly fractured and cleaved metamorphosed basalts (Goropu Metabasalt; Lindley 1992) is compelling evidence for buoyancy-driven uplift. I propose that the late Miocene intrusion of a granite diapir (precursor to the Suckling Granite) into the thick crust of the eastern Papua region (Milsom & Smith 1975) and the initiation of buoyant uplift of the Suckling Dome was the beginning of a cycle of tectonic instability that continues to the present. The rising granite body was responsible for metamorphism and foliation of overlying rock, a feature noted in other eastern Papuan domes (Ollier & Pain 1980). In turn, the foliated and brittle carapace was thrust upwards by the buoyant driving force of a much larger granite batholith, eventually breaking through ground surface as a surficial dome. The initiation of buoyant uplift of the Suckling Dome coincided with the commencement of Woodlark rifting (Taylor et al. 1999).

TOPOGRAPHY
The broad 3676 m high Suckling Dome (Figures 4b, 11, 12) is well dissected. Northern and NW flanks are relatively intact, but faulting and erosion have obscured the western and southern slopes ( Figure 11). The contrasting topography of the northern slopes of the Suckling Dome and Dayman Dome (evident on the SAFIA and DAYMAN 1:100 000 topographic maps) suggests the former to be an older structure. Many subparallel streams drain the slopes of the Dayman Dome (Ollier & Pain 1980Daczko et al. 2011). Increased dissection of higher slopes of the Dayman Dome, the result of more erosion, is evident while the lower slopes are less weathered (Ollier & Pain 1980. By contrast, well-developed network drainages have dissected both the lower and upper slopes of the Suckling Dome (Figure 11), indicating considerably more erosion and the longevity of the dome structure. I. D. Lindley TAJE_A_965980.3d (TAJE) (210£274mm) 10-11-2014 14:53

DOME DEVELOPMENT
The Keveri Fault facilitated most of the rapid vertical movement on the southern margin of the Suckling Dome, with an estimated 8000 m of uplift since the late Miocene, at a rate of around 2.5 m/10 3 a. The vertical movement on the Keveri Fault is notably different from that described for structures flanking other eastern Papua metamorphic core complexes. Detachment faults are present along the northern margin of the Suckling Dome, Dayman Dome and the D'Entrecasteaux domes (Ollier & Pain 1980Hill et al. 1992;Hill & Baldwin 1993;Daczko et al. 2011;Little et al. 2011). There is no field evidence for the development of a low-angle, south-dipping detachment fault along the southern margin of the Suckling Dome.
Episodic and very energetic uplift of the Suckling Dome continued during the late PlioceneÀ?early Pleistocene. Flanking subsidiary faults to the Keveri Fault in the Wavera Valley were activated, forming an intermontane basin in the Keveri Valley. Rapid uplift and erosion resulted in the cyclical deposition of poorly sorted gravel and sandstone (Domara River Conglomerate) in the intermontane basin and the development of extensive piedmont fan complexes covering a 400 km 2 area west and NW of Mt Suckling (Figure 2). Estimated thicknesses of non-marine sediment deposited in the fan complexes vary between 1500 and 1600 m. Rapid uplift during the Quaternary, with movement continuing to be accommodated along the Keveri Fault, resulted in localised tilting and displacement of the Domara River Conglomerate. Dextral strike-slip motion along the Wavera Valley faults displaced the Domara River Conglomerate and, along the Keveri Fault, offset tributaries of the Ada'u River. Rapid erosion associated with the vigorous uplift of Mt Suckling to its present height resulted in the unroofing of high-level granite (Suckling Granite) and monzonite (Mai'iu Monzonite) apophyses connected at depth to the granite diapir(s) that initiated buoyant uplift (Figures 2, 4b). Valley-confined ultramafic breccia deposits in Dimidi Creek, on the SW flanks of the Suckling Dome, attest to repeated cycles of vigorous uplift.

Keveri Fault and the Owen Stanley Fault System
Geophysical data, and in particular gravity, are very clear in showing that the OSFZ (and the PUB) passes north of the Mt Suckling district into Collingwood Bay. Gravity data also indicates that the PUB as an intact coherent body is not present east of 149 30ʹE. Movement on the Keveri Fault commenced in the late Miocene with the initiation of buoyant uplift of the Suckling Dome. Regional BMR mapping indicates that the western extension of the Keveri Fault in the 148 15ʹE to 148 40ʹE interval is represented by a series of splays (Figure 2; Davies & Smith 1974;Pieters 1978), suggesting that maximum movement occurred along the Ada'u Valley interval of the fault as uplift of the Suckling Dome progressed. There is no obvious connection between the western extensions of the Keveri Fault and the OSFZ. Both structures are subparallel and may be part of a broad zone of shearing. East of the Ada'u Valley, the trace of the Keveri Fault is prominent in the Bonua Valley, underlying the broad sediment filled valley (149 08ʹE/9 56ʹS) and