*Revised manuscript with no changes marked Click here to view linked References

1 Melanesian back-arc basin and arc development: constraints from the eastern Coral 1 2 2 Sea 3 4 5 1* 2 1 3 4 6 3 Maria Seton , Nick Mortimer , Simon Williams , Patrick Quilty , Phil Gans , Sebastien 7 8 4 Meffre3,5, Steven Micklethwaite6a, Sabin Zahirovic1, Jarrod Moore1, Kara J. Matthews1b 9 10 11 1 12 5 EarthByte Group, School of Geosciences, The University of Sydney, NSW 2006, Australia 13 14 15 6 2GNS Science, Private Bag 1930, Dunedin, 9054, New Zealand 16 17 18 3 19 7 Discipline of Earth Sciences, University of Tasmania, Private Bag 79, Hobart, TAS, 7001, 20 21 8 Australia 22 23 24 4 25 9 Department of Earth Science, UC Santa Barbara, Santa Barbara, CA, 93106-9630, USA 26 27 28 10 5ARC Centre of Excellence in Ore Deposits, University of Tasmania, Private Bag 79, Hobart, 29 30 31 11 TAS, 7001, Australia 32 33 34 12 6Centre for Exploration Targeting, University of Western Australia, Perth, 6009, Australia 35 36 37 * 38 13 Corresponding author. e-mail: [email protected] 39 40 41 14 a Present address: School of Earth, Atmosphere and Environment, Monash University, 42 43 44 15 Melbourne, VIC, 3800, Australia 45 46 47 16 b Present address: Department of Earth Sciences, University of Oxford, South Parks Road, 48 49 17 50 Oxford OX1 3AN, UK 51 52 53 18 54 55 56 19 57 58 59 60 20 Submitted to Gondwana Research 61 62 63 64 1 65 21 1 2 3 22 4 5 6 7 23 Abstract 8 9 10 24 The eastern Coral Sea is a poorly explored area at the north-eastern corner of the Australian 11 12 13 25 Tectonic Plate, where interaction between the Pacific and Australian plate boundaries, and 14 15 26 accretion of the world’s largest submarine plateau - the Ontong Java Plateau - has resulted in 16 17 18 27 a complex assemblage of back-arc basins, island arcs, continental plateaus and volcanic 19 20 28 products. This study combines new and existing magnetic anomaly profiles, seafloor fabric 21 22 23 29 from swath bathymetry data, Ar-Ar dating of E-MORB basalts, palaeontological dating of 24 25 30 carbonate sediments, and plate modelling from the eastern Coral Sea. Our results constrain 26 27 31 28 commencement of the opening of the Santa Cruz Basin and South Rennell Trough to c. 48 29 30 32 Ma and termination at 25-28 Ma. Simultaneous opening of the Melanesian Basin/Solomon 31 32 33 Sea further north suggests that a single > 2,000 km long back-arc basin, with at least one 33 34 35 34 triple junction existed landward of the Melanesian subduction zone from Eocene-Oligocene 36 37 35 times. The cessation of spreading corresponds with a reorganization of the plate boundaries 38 39 40 36 in the area and the proposed initial soft collision of the Ontong Java Plateau. The correlation 41 42 37 between back-arc basin cessation and a widespread plate reorganization event suggests that 43 44 38 45 back-arc basins may be used as markers for both local and global plate boundary changes. 46 47 48 39 Keywords 49 50 51 40 52 Eastern Coral Sea; Melanesia; SW Pacific; back-arc basin; subduction 53 54 55 41 1. Introduction 56 57 58 59 60 61 62 63 64 2 65 42 The Coral Sea lies in the north-eastern corner of the Australian Tectonic Plate, in a pivotal 1 2 43 position at the juncture between two of the world’s most tectonically complex regions: the 3 4 5 44 SW Pacific and SE Asia (Fig. 1). While the origin and evolution of the western part of the 6 7 45 Coral Sea, including the Queensland and Marion Plateaus and Coral Sea Basin is relatively 8 9 10 46 well-known (Falvey and Taylor, 1974; Gaina et al., 1999; Weissel and Watts, 1979), the 11 12 47 eastern Coral Sea, with its mosaic of ridges, plateaus and basins has a largely unknown 13 14 48 15 history. Previous plate tectonic reconstructions seeking to explain the origin and evolution of 16 17 49 the submarine features in the eastern Coral Sea have been limited. The area is either largely 18 19 50 ignored in regional plate tectonic models (Hall, 2002; Sdrolias et al., 2003; Weissel and 20 21 22 51 Watts, 1979), or overly simplified with inferred ages for many of the features (Schellart et al., 23 24 52 2006; Whattam et al., 2008; Yan and Kroenke, 1993). These poorly constrained models 25 26 27 53 reflect the scarcity of marine geophysical data and geological samples from the area, the 28 29 54 inherent complexity of the many ridges and basins, and substantial volcanic and tectonic 30 31 32 55 overprinting. Such a significant gap in our understanding of the tectonic development of the 33 34 56 Coral Sea is unsatisfactory particularly for an area that provides plate boundary continuity 35 36 57 37 between the SW Pacific and SE Asia along the western Pacific boundary, and thus is critical 38 39 58 for plate reconstructions of both areas. In addition, this area lies adjacent to the North 40 41 59 Solomon subduction zone and Ontong Java Plateau, the site of a major plateau accretion and 42 43 44 60 subduction event, which is believed to have occurred either in the Oligocene (Musgrave, 45 46 61 2013), Miocene (Petterson et al., 1999) or Pliocene (Cowley et al., 2004). This collision 47 48 49 62 induced a subduction polarity reversal in the Pliocene (Cooper and Taylor, 50 51 63 1985). Understanding the evolution of the submarine features in the eastern Coral Sea may 52 53 64 54 provide insights into the timing and processes of the Ontong Java Plateau collision and help 55 56 65 identify the effects of a major collision event on the over-riding plate. 57 58 59 60 61 62 63 64 3 65 66 We present the results of a recent research voyage (SS2012_V06) on RV Southern Surveyor 1 2 67 to the eastern Coral Sea in October-November, 2012 where 13,600 km2 of swath bathymetry 3 4 5 68 data, 6,200 km of magnetic and 6,800 km of gravity data were collected, together with 6 7 69 igneous and sedimentary samples from 14 seafloor sites (Fig. 1). Swath bathymetry and 8 9 10 70 magnetic anomaly data are used to model the timing and orientation of seafloor spreading in 11 12 71 the Santa Cruz Basin. Geochemical analysis and Ar-Ar dating of recovered igneous samples 13 14 72 15 together with paleontological age constraints from sedimentary samples at the dredge site 16 17 73 locations are used to relate spreading in the Santa Cruz Basin to activity along the South 18 19 74 Rennell Trough and the igneous activity along the Rennell Island Ridge. Combining the 20 21 22 75 results of our study with recently dated dredge samples from earlier ORSTOM voyages 23 24 76 (Mortimer et al., 2014), we develop plate tectonic reconstructions of the eastern Coral Sea 25 26 27 77 since the Cretaceous underpinned by age constraints from both marine geophysical data and 28 29 78 geological samples. Our plate reconstructions include a network of continuously closing 30 31 32 79 plate boundaries (Gurnis et al., 2012), which can be used to provide plate boundary 33 34 80 continuity between the SW Pacific and SE Asia during this time period. 35 36 37 81 2. Regional tectonic framework 38 39 40 41 82 The SW Pacific is characterised by a series of marginal basins, submerged continental slivers 42 43 83 44 and back-arc—arc—forearc complexes largely controlled by the interaction of the Australian 45 46 84 and Pacific Tectonic Plates since the Mesozoic (Crawford et al., 2003; Schellart et al., 2006; 47 48 85 Sdrolias et al., 2003; Whattam et al., 2008; Yan and Kroenke, 1993) (Fig. 1). In contrast, SE 49 50 51 86 Asia is an amalgamation of accretionary continental fragments, exotic terranes and intra- 52 53 87 oceanic island arcs, formed largely from the long-term interaction between the Australian, 54 55 56 88 Pacific and Eurasian plates and Gondwana-derived terrane accretion (Acharyya, 1998; 57 58 89 Golonka, 2004; Hall, 2002; Metcalfe, 2009; Stampfli and Borel, 2002; Zahirovic et al., 59 60 61 62 63 64 4 65 90 2014). These two tectonically complex regions are connected through the “Melanesian 1 2 91 Borderlands” region, which includes the Coral and Solomon Seas. 3 4 5 6 92 The Coral Sea is bounded by the passive margin of eastern Australian in the west, the 7 8 93 inactive Pocklington Trough and Papuan Peninsula to the north, the active South 9 10 11 94 Solomon/San Cristobal and New Hebrides Trenches to the northeast and east, and the Lord 12 13 95 Howe Rise/New Caledonia region to the south (Fig. 1). It consists of the relatively well- 14 15 16 96 explored Coral Sea Basin and Queensland and Marion plateaus in the western part and a 17 18 97 complex arrangement of poorly explored and under-sampled submarine plateaus, linear 19 20 98 21 depressions, elongated ridges, oceanic basins and seamounts in the eastern part, including the 22 23 99 Rennell Basin, Rennell Island Ridge, South Rennell Trough, West Torres Plateau and Santa 24 25 100 Cruz and d’Entrecasteaux Basins (Fig. 1). 26 27 28 29 101 Many of the features in the area are named after Rennell Island, which lies in the eastern 30 31 102 Coral Sea at 160°E, 11.5°S (Fig. 1). These include: the Rennell Island Ridge, East Rennell 32 33 34 103 Island Ridge, Rennell Trough, Rennell Basin, South Rennell Trough and Rennell Fracture 35 36 104 Zone (e.g. (Daniel et al., 1978; Landmesser et al., 1973; Larue et al., 1977)). As the region 37 38 39 105 becomes better known, we consider this preponderance of the name 'Rennell' to be unhelpful 40 41 106 in science communication, especially as some of these features are up to 700 km distant from 42 43 107 44 Rennell Island. In this paper we retain the names South Rennell Trough and Rennell Basin as 45 46 108 these are used by all aforementioned authors except Landmesser et al. (1973). Use of these 47 48 109 names supersedes Rennell Fracture Zone and Rennell Trough, respectively. We also retain 49 50 51 110 the name Rennell Island Ridge but restrict its use to the NW-SE striking part around Rennell 52 53 111 Island. We rename the East Rennell Island Ridge of Landmesser et al. (1973) as East 54 55 56 112 Lapérouse Rise (Fig. 3), and introduce the name West Lapérouse Rise, for a mirror image 57 58 113 positive bathymetric feature on the western side of the South Rennell Trough, and separated 59 60 61 62 63 64 5 65 114 from Mellish Rise by a bathymetric low (Fig. 1). Lapérouse was the French navy officer and 1 2 115 explorer whose expedition vanished in the Vanuatu area c. 1788 and was searched for by 3 4 5 116 d'Entrecasteaux. Lapérouse probably would have transited from Australia to New Caledonia 6 7 117 across this region. To our knowledge, Lapérouse has not previously been used as a name for 8 9 10 118 an undersea feature. A further complication regarding feature names in the region is the use 11 12 119 of Bellona Island and Bellona Plateau. Bellona Island lies to the north of Rennell Island (Fig. 13 14 120 15 2) at a latitude of ~11°S and is part of the Rennell and Bellona Province of the Solomon 16 17 121 Islands. The Bellona Plateau is a largely submerged plateau located at the southern tip of the 18 19 122 South Rennell Trough at a latitude of ~19°S. 20 21 22 23 123 Tectonically, the region is dominated by active subduction to the northeast and east, along the 24 25 124 South Solomon/San Cristobal Trench and New Hebrides Trench (Fig. 1), with a minor 26 27 28 125 amount (~20 mm/yr) of Pacific plate subduction along the North Solomon Trench (Tregoning 29 30 126 et al., 1998) (Fig. 1). North of the Coral Sea, seafloor spreading is active in the Woodlark 31 32 33 127 Basin. 34 35 36 128 The main bathymetric features in the eastern Coral Sea are described below and labelled in 37 38 39 129 Fig. 1. 40 41 42 130 2.1. Louisiade Plateau, Louisiade Trough and Mellish Rise 43 44 45 46 131 The submerged Louisiade Plateau and Mellish Rise (Fig. 1) are inferred to be composed of 47 48 132 continental material that rifted off the Australian margin during Tasman and Coral Sea 49 50 133 opening in the Late Cretaceous-Paleocene (Gaina et al., 1998; Gaina et al., 1999; Willcox et 51 52 53 134 al., 1980). Although no basement rocks have yet been recovered from either plateau, the 54 55 135 similarity in seismic character to the Queensland and Kenn plateaus (Exon et al., 2006; 56 57 58 136 Taylor and Falvey, 1977) and the recovery of sediments with terrigenous material from 59 60 137 basement highs on the Mellish Rise (Hoffmann et al., 2008) strongly supports a continental 61 62 63 64 6 65 138 origin. In contrast, global compilations of Large Igneous Province (LIP) locations (Bryan 1 2 139 and Ernst, 2008; Coffin et al., 2006) classify the Louisiade Plateau as an oceanic plateau 3 4 5 140 implying that it is either entirely volcanic (Cowley et al., 1998) or capped by volcanics. The 6 7 141 Louisiade Trough separates the Louisiade Plateau from the Mellish Rise and reaches depths 8 9 10 142 of over 5500 m. It accommodated seafloor spreading during the opening of the Tasman and 11 12 143 Coral seas, as evidenced by rift-related structures along the margins (Exon et al., 2006) and 13 14 144 15 the identification of magnetic anomalies 27-26 (Gaina et al., 1999). No basement samples 16 17 145 have been obtained from the Louisiade Trough. 18 19 20 146 21 2.2. Rennell Island Ridge and Basin 22 23 24 147 East of the Louisiade Plateau lies a short and narrow (50 km wide) depression with water 25 26 148 depths exceeding 5000 m (Rennell Basin) (Figs 1, 3). Seismic reflection lines show a flat- 27 28 29 149 floored basin partly filled with undeformed sediments and a trench-like morphology at the 30 31 150 eastern margin (Récy et al., 1977). The undeformed sediments are believed to be Oligocene 32 33 34 151 and younger (Récy et al., 1977). Previously-published models speculate that the Rennell 35 36 152 Basin was the site of east-dipping subduction during the Eocene (Hall, 2002; Weissel and 37 38 39 153 Watts, 1979; Yan and Kroenke, 1993) or the site of seafloor spreading (Landmesser et al., 40 41 154 1973), separating an inferred continental fragment, the Rennell Island Ridge, from the north- 42 43 155 44 eastern margin of the Louisiade Plateau. 45 46 47 156 The Rennell Island Ridge is a ~200 km long and ~70 km wide largely submerged ridge 48 49 157 structure east of the Rennell Basin (Fig. 1), with sub-aerial exposure of the Rennell and 50 51 52 158 Bellona Islands and the shallow Indispensible Reef complex (Fig. 2). Rennell and Bellona 53 54 159 Islands belong to the world’s largest exposed raised coral atoll, consisting of a thick (up to 55 56 57 160 500 m) sequence of coral reefs and dolomite (Taylor, 1973) but with no basement 58 59 161 exposures. Three alternative models for the origin of the Rennell Island Ridge have been 60 61 62 63 64 7 65 162 proposed. These include: an island arc above an east-dipping subduction zone with the 1 2 163 Rennell Basin acting as the trench (Daniel et al., 1978; Weissel and Watts, 1979; Yan and 3 4 5 164 Kroenke, 1993); a fragment of Mesozoic Australia-derived continental lithosphere with the 6 7 165 Rennell Basin acting as an abandoned basin formed by seafloor spreading (Landmesser et al., 8 9 10 166 1973) and; oceanic basalt, similar in character to the Cretaceous uplifted oceanic crust found 11 12 167 on the Malaita Islands in the southern Solomon Islands (Hughes and Turner, 1977). A 13 14 168 15 basaltic dredge sample from the western margin of the Rennell Island Ridge revealed a 16 17 169 geochemically depleted MORB- or arc-type rock with a minor degree of subduction-related 18 19 170 geochemistry of 38 ± 5 Ma age (Mortimer et al., 2014). Deciphering the origin and evolution 20 21 22 171 of the Rennell Island Ridge is further complicated by the substantial Quaternary uplift of the 23 24 172 ridge linked to the proximal northeast-dipping subduction along the South Solomon Trench 25 26 27 173 and Pliocene subduction polarity reversal. 28 29 30 174 2.3. South Rennell Trough 31 32 33 34 175 South Rennell Trough extends into the Santa Cruz Basin to the northeast, as observed in 35 36 176 bathymetry and satellite gravity maps, and terminates at the Bellona Plateau in the southwest 37 38 39 177 (Fig. 1 and 3). It is 700 km long, reaches depths of over 5000 m and includes a 30 km wide 40 41 178 central trough and two prominent, parallel ridge systems. Previous models have proposed that 42 43 179 44 the ridges and troughs formed either along a transform fault during the opening of the 45 46 180 Tasman and Coral Seas (Landmesser et al., 1973), an extinct late Oligocene spreading ridge 47 48 181 (Daniel et al., 1978; Larue et al., 1977), or a Cretaceous-Paleocene spreading ridge (Schellart 49 50 51 182 et al., 2006). Recent geochemical analysis and Ar-Ar dating of basalt samples from the 52 53 183 northern part of the South Rennell Trough (Mortimer et al., 2014) (Fig. 3) confirms MORB 54 55 56 184 geochemistry of Oligocene age and hence, a spreading centre origin for the South Rennell 57 58 185 Trough. Outstanding questions remain regarding the ridge features on either side of the 59 60 61 62 63 64 8 65 186 South Rennell Trough (the East and West Lapérouse Rises) (Fig. 3), and the relationship to 1 2 187 the oceanic crust that formed in the Santa Cruz Basin. 3 4 5 6 188 2.4. Santa Cruz Basin 7 8 9 189 The Santa Cruz Basin is located in the corner of the eastern Coral Sea between the New 10 11 12 190 Hebrides and South Solomon trenches and the West Torres Plateau and Rennell Island Ridge 13 14 191 (Fig. 1). The size of the Santa Cruz Basin has been modified due to active subduction of the 15 16 17 192 basin along the New Hebrides and South Solomon Trenches. Although there are no published 18 19 193 magnetic anomaly interpretations of the basin, it is inferred floored by oceanic 20 21 194 22 crust. However, the age of formation of the basin has been debated with models invoking 23 24 195 either Cretaceous opening (Kroenke, 1984; Schellart et al., 2006), coincident with opening of 25 26 196 the d’Entrecasteaux Basin to the south, or Oligocene opening (Larue et al., 1977; Wells, 27 28 29 197 1989; Yan and Kroenke, 1993) coincident with extension along the South Rennell 30 31 198 Trough. Although recent analysis of dredges from the ORSTOM survey revealed 28 ± 1 Ma 32 33 34 199 BABB type material from the Santa Cruz Basin (Mortimer et al., 2014) (Fig. 4), the 35 36 200 spreading evolution of the basin has yet to be deciphered. 37 38 39 40 201 2.5. West Torres Plateau 41 42 43 202 The West Torres Plateau is a submarine plateau, located south of the Santa Cruz Basin and 44 45 46 203 north of the d’Entrecasteaux Basin, which interacts with the New Hebrides Trench on its 47 48 204 eastern border (Fig. 1). It has a depth range of ~1000-3500 m and a width of approximately 49 50 205 51 200 km (Fig. 5). The origin of the plateau is unknown, with no available dredge 52 53 206 samples. Two alternative models exist for the nature of basement: a continental origin 54 55 207 (Schellart et al., 2006), perhaps as a continental raft from eastern Gondwana, or a mantle 56 57 58 208 plume origin, with the eruption of the associated LIP in the Oligocene (Yan and Kroenke, 59 60 209 1993). A continental origin for the West Torres Plateau would require an approximately east- 61 62 63 64 9 65 210 west spreading direction in the neighbouring d’Entrecasteaux Basin, unless significant north- 1 2 211 south translation occurred subsequently. A LIP origin would require a connection to a past or 3 4 5 212 present-day hotspot and the presence of an associated trail. 6 7 8 213 2.6. d’Entrecasteaux Basin 9 10 11 12 214 The d’Entrecasteaux Basin (sometimes referred to as the North d’Entrecasteaux Basin) is a 13 14 215 fan-shaped basin bounded by the South Rennell Trough to the north, the arcuate-shaped 15 16 17 216 d’Entrecasteaux Zone to the south, the West Torres Plateau to the east and the New 18 19 217 Caledonia Basin and South Rennell Trough to the west (Fig. 1). A tentative interpretation 20 21 218 22 based on limited magnetic anomaly profiles suggested a Cretaceous age (between anomalies 23 24 219 30-34; 65-84 Ma) (Lapouille, 1982) from roughly NW-SE directed spreading (Fig. 25 26 220 5). Although this interpretation has been questioned, no alternative interpretation of the 27 28 29 221 magnetic lineations has been published and no in-situ samples have been collected. The 30 31 222 relationship between the d’Entrecasteaux Basin and the surrounding area has been suggested 32 33 34 223 in three ways: the d’Entrecasteaux Basin and North Loyalty Basin were originally one basin 35 36 224 (Lapouille, 1982); the d’Entrecasteaux Basin, South Rennell Trough and Santa Cruz Basin 37 38 39 225 share a common Cretaceous opening history (Schellart et al., 2006); and the d’Entrecasteaux 40 41 226 Basin is the north-western extent of the New Caledonia Basin (Schellart et al., 2006). 42 43 44 227 45 The d’Entrecasteaux Basin terminates in the south at the d’Entrecasteaux Zone, a ~100 km 46 47 228 wide arcuate structure consisting of various horsts and graben (Maillet et al., 1983) and 48 49 229 partitioned into the North d’Entrecasteaux Ridge and the South d’Entrecasteaux Chain (see 50 51 52 230 Mortimer et al. (2014) for description). Steep slopes mark its boundary into the North 53 54 231 Loyalty Basin to the south and the West Santo and d’Entrecasteaux basins to the north (Fig. 55 56 57 232 5). Recent analysis and dating of samples from the North d’Entrecasteaux Ridge indicate 58 59 233 primitive arc tholeiites of Eocene age. One sample yielded a 49 ± 10 Ma age for E-MORB 60 61 62 63 64 10 65 234 back-arc basin type rocks (Mortimer et al., 2014), although they caution that more samples 1 2 235 need to be collected to confirm the interpretation. 3 4 5 6 236 3. Tectonic features and bathymetric interpretation 7 8 9 237 New swath bathymetry data were collected during the SS2012_V06 research voyage using a 10 11 12 238 30 kHz Simrad EM300 multibeam echo sounder system (see Figs. S1-3). The data were 13 14 239 processed using Caris software and were combined with regional, publicly available swath 15 16 17 240 bathymetry data available through the NGDC, JAMSTEC and Geoscience Australia 18 19 241 databases. Where swath coverage was unavailable, global bathymetry compilations were 20 21 242 22 used. Seismic reflection profiles from the area (e.g. (Landmesser et al., 1973)) are typically 23 24 243 old and crude but were useful to help reveal large-scale basement structure and faulting 25 26 244 patterns. 27 28 29 30 245 3.1. Rennell Island Ridge and Rennell Basin 31 32 33 246 The Rennell Island Ridge is characterised by the sub-aerial exposure of Rennell and Bellona 34 35 36 247 Islands and the shallow Indispensible Reef complex (Fig. 2). Regional bathymetry 37 38 248 compilations and seismic reflection profiles (Landmesser et al., 1973) show thickened crust 39 40 41 249 with variation in structure and morphology along strike. There are two main structural 42 43 250 trends: a NW-SE orientation, similar to the south-western margin of the West Torres Plateau 44 45 46 251 and parallel to Indispensible Reef and; a WNW-ESE orientation, similar to part of the trend 47 48 252 of the South Solomon subduction zone and parallel to Rennell and Bellona Islands (Fig. 49 50 253 2). Between the two main structural trends in the southern part of the Rennell Island Ridge 51 52 53 254 complex lies a small < 40 km wide, 1800 m deep trough with smooth seafloor (Fig. 2 and 54 55 255 S1), which likely contains thick sediment accumulations that obscure the underlying smaller- 56 57 58 256 scale structure. The Rennell Island Ridge is bounded by normal faults along its western 59 60 257 margin (see Daniel et al. (1978) and Recy et al. (1977) where it rises from depths of over 61 62 63 64 11 65 258 3500 m at the edge of the slope to 1200 m at the shelf edge (Fig. 2). At its southern 1 2 259 termination, the Rennell Island Ridge does not share structural affinities with the South 3 4 5 260 Rennell Trough. 6 7 8 261 No available swath profiles exist across the Rennell Basin but previous seismic reflection 9 10 11 262 lines show a flat-floored basin partly filled with undeformed sediments, the thickness of 12 13 263 which increases southward (Daniel et al., 1978). In addition, seismic reflection profiles show 14 15 16 264 a morphology and structure characteristic of an inactive subduction zone along the eastern 17 18 265 margin (Daniel et al., 1978; Récy et al., 1977). However, this interpretation remains tentative 19 20 266 21 and cannot be supported structurally without the addition of high-resolution bathymetry 22 23 267 mapping from the Rennell Basin paired with the recovery of a suite of subduction-related 24 25 268 rocks from the Rennell Island Ridge. 26 27 28 29 269 3.2. South Rennell Trough 30 31 32 270 Swath bathymetry profiles collected during SS2012_V06 along the South Rennell Trough 33 34 35 271 were combined with profiles from the RV Melville (BMRG07MV) and RV Mirai (MR0707) 36 37 272 (Fig. 3). The South Rennell Trough is an extinct spreading centre, based on the E-MORB 38 39 40 273 geochemistry and 28-29 Ma Ar-Ar dates of the recovered basalts from the edge of a rift 41 42 274 graben (Mortimer et al., 2014). The broad scale structure is that of a deep (up to 5400 m), 43 44 275 ° 45 steep-sided (gradients up to 30 ), wide (10-30 km) rift valley with a series of 2-3 sub-parallel 46 47 276 abyssal hills/ridge lines on either side that become narrower southward as they transition in to 48 49 277 the Bellona Plateau. These abyssal hills have elevations of between 1500-2000 m (highest 50 51 52 278 along the outer ridge), giving a total relief, from rift valley to axial ridge, of ~3000-4000 53 54 279 m. The rift valley floor of the South Rennell Trough contains NE-SE-directed seafloor fabric 55 56 57 280 (Fig. 3a). The similarity in trend between the abyssal hill fabric and the broad scale structure 58 59 60 61 62 63 64 12 65 281 of the ridge suggests no reactivation of the plate boundary after formation, as suggested by 1 2 282 Coleman and Packham (1976)). 3 4 5 6 283 The morphology and structure of the South Rennell Trough share similar characteristics to 7 8 284 the ultra-slow spreading Gakkel Ridge in the Arctic Ocean. The Gakkel Ridge is 9 10 11 285 characterised by a deep axial valley (depth between 4700-5300 m), a width of ~15-30 km, a 12 13 286 continuous ridge axis with no transform offsets and linear rift-parallel ridges, and fault- 14 15 16 287 bounded troughs with up to 2000 m elevation (Cochran et al., 2003). In this scenario, the 17 18 288 elevated ridges on either side of the South Rennell Trough, the East and West Lapérouse 19 20 289 21 Rises, can be explained by the complex interplay between magma supply and plate stretching 22 23 290 (Buck et al., 2005). The limits on seafloor spreading rates (maximum of 20 mm/yr; (Dick et 24 25 291 al., 2003)) and the general morphology and structure of the South Rennell Trough complex 26 27 28 292 can be used as a constraint on plate tectonic reconstructions of this seafloor spreading system. 29 30 31 293 3.3. Santa Cruz Basin 32 33 34 35 294 Three full and four partial swath profiles collected across the Santa Cruz Basin during voyage 36 37 295 SS2012_V06 were combined with a swath profile from voyage MGLN06MV on the RV 38 39 40 296 Melville (Fig. 4 and Fig. S2). 41 42 43 297 The most pronounced feature in the Santa Cruz Basin is a wide (average width 20-30 km), 44 45 46 298 >1000 m deep trough that is continuous with the South Rennell Trough (Fig. 4). Axial 47 48 299 troughs of this depth and width are characteristic of slow (20-55 mm/yr) or ultra-slow (< 20 49 50 300 mm/yr) seafloor spreading where magma is supplied slowly enough for the crest of the ridge 51 52 53 301 to subside as the oceanic plates cool. Faults on either side of the axial trough are numerous in 54 55 302 slow spreading rifts and create typically rough topography, as is observed on either side of 56 57 58 303 the axial trough in the Santa Cruz Basin (Fig. 4). 59 60 61 62 63 64 13 65 304 Away from the inferred extinct mid-ocean ridge, the swath data show clear abyssal hill fabric 1 2 305 indicative of seafloor spreading. The abyssal hill relief is typically 100-300 m and rough, 3 4 5 306 again suggesting a slow seafloor-spreading rate. The seafloor spreading fabric displays two 6 7 307 main structural trends: 30° (from N) adjacent to the extinct ridge, similar to the fault trends in 8 9 ° 10 308 the South Rennell Trough; and a trend at 70 (from N) away from the extinct ridge in the east 11 12 309 of the basin, beyond ~164° longitude (Fig. 4 and Fig. S2). This indicates a change from 13 14 310 15 NNW-SSE oriented to NW-SE oriented spreading during the formation of the Santa Cruz 16 17 311 Basin. Fracture zones were not identified in any of the profiles across the Santa Cruz 18 19 312 Basin. The lack of fracture zones cannot be related to obstruction by sediment infilling as 20 21 22 313 Luyendyk et al. (1974) and De Broin et al. (1977) described a subhorizontal, well-bedded 23 24 314 sedimentary section less than 1 km thick in the basin. Instead, this may be related to the 25 26 27 315 spreading rate, as ultra-slow and some slow spreading ridges characteristically lack transform 28 29 316 segments (Dick et al., 2003). 30 31 32 33 317 Although the South Rennell Trough appears to extend into the Santa Cruz Basin, the elevated 34 35 318 nature of the East and West Lapérouse Rises suddenly disappears at the point where the 36 37 319 38 South Rennell Trough intersects the Rennell Island Ridge and West Torres Plateau (i.e., into 39 40 320 the Santa Cruz Basin). This may reflect a difference in spreading processes, such as rate of 41 42 321 seafloor spreading, between the South Rennell Trough (ultra-slow) and Santa Cruz Basin 43 44 45 322 (slow to ultra-slow). However, the spatial relationship to the Rennell Island Ridge and West 46 47 323 Torres Plateau suggests that there may be further differences, such as the initial rifting 48 49 50 324 process or thermal structure. 51 52 53 325 3.4. West Torres Plateau 54 55 56 57 58 59 60 61 62 63 64 14 65 326 Two profiles were collected during the voyage SS2012_V06 in a small section of the western 1 2 327 margin of the West Torres Plateau (Fig. 5), which were combined with a swath profile from 3 4 5 328 voyage KIWI10RR on the RV Roger Revelle. 6 7 8 329 The West Torres Plateau is steep-sided, oriented NW-SE, similar to the main structural trend 9 10 11 330 of the Rennell Island Ridge (Fig. 2 and 5), and faulted along the southwestern margin (slopes 12 13 331 greater than 30o). Gentle slopes extend into the Santa Cruz Basin and New Hebrides Trench 14 15 16 332 (Fig. 5). An arm of the West Torres Plateau protrudes westward near the South Rennell 17 18 333 Trough. A detailed swath survey along this arm reveals N-S structural trends, which may be 19 20 334 21 lava flow fields rather than reflecting underlying structural fabric (Fig. S1, C). 22 23 24 335 In the southwest corner of the West Torres Plateau lies a small basin, which has previously 25 26 336 been included as part of the d’Entrecasteaux Basin. Two swath profiles that cross this basin 27 28 29 337 from voyage KIWI10RR on the RV Roger Revelle and BMRG07MV on the RV Melville (Fig. 30 31 338 5) reveal NE-SW oriented seafloor fabric in the north, which changes to E-W oriented fabric 32 33 34 339 in the south adjacent to the E-W trending features of the d’Entrecasteaux Zone. The fabric 35 36 340 identified in these profiles confirms that this basin has a discrete opening history from the 37 38 39 341 main part of the d’Entrecasteaux Basin. 40 41 42 342 3.5. d’Entrecasteaux Basin 43 44 45 46 343 Structural trends in both existing and newly collected swath bathymetry data reveal that the 47 48 344 d’Entrecasteaux Basin consists of three discrete zones (Fig. 5). Zone 1 (Fig. 5) is the largest 49 50 345 and forms the d’Entrecasteaux Basin proper. While three swath profiles collected during 51 52 53 346 voyage SS2012_V06 failed to yield useful structural information related to seafloor (Fig. S3), 54 55 347 a swath profile that was collected adjacent to the South Rennell Trough by the RV Mirai 56 57 58 348 during voyage MR0707 reveals NE-SW oriented structural trends in the basin (Fig. 59 60 349 5). Whether this fabric is indicative of seafloor fabric or late-stage faulting during South 61 62 63 64 15 65 350 Rennell Trough activity is unknown. Zone 2 (Fig. 5) is separated from the third by a narrow 1 2 351 band of about 3000 m high seafloor. A single swath profile from voyage BMRG07MV on 3 4 5 352 the RV Melville reveals roughly N-S oriented seafloor structures that may represent abyssal 6 7 353 hill fabric (Fig. 5). TZone 3) is bounded by the West Torres Plateau to the north and east and 8 9 10 354 the d’Entrecasteaux Zone to the south. In this basin, the structural trends are NE-SW 11 12 355 oriented in the north, changing to E-W oriented in the south adjacent to the E-W trending 13 14 356 15 features of the d’Entrecasteaux Zone (Fig. 5). The fabric to the north could be indicative of 16 17 357 abyssal hill fabric with the E-W fabric likely related to later deformation along the 18 19 358 d’Entrecasteaux Zone. 20 21 22 23 359 4. Rock samples 24 25 26 360 Samples described in this paper were obtained from seven dredge sites on the Rennell Island 27 28 29 361 Ridge, West Torres Plateau, East Lapérouse Rise and the South Rennell Trough (Table 1 and 30 31 362 Fig. S1 and S4). Rocks obtained from other dredge sites during voyage SS2012_V06 32 33 34 363 comprise intra-plate lavas from the Lord Howe Seamount chain and the Zealandia continent 35 36 364 and are presented in Morimer et al. (in review). 37 38 39 40 365 A subset of sorted dredge samples was given GNS Science “P numbers” and catalogued in 41 42 366 GNS Science’s National Petrology Reference Collection. Location, rock type and analytical 43 44 367 45 data are lodged in the open file PETLAB database 46 47 368 (http://pet.gns.cri.nz/result_list.jsp?Type=Ext&Collector=ECOSAT). 48 49 50 369 A range of volcanic and sedimentary rocks was dredged on voyage SS2012_V06, mainly 51 52 53 370 basalts, limestones and calcareous mudstones. Summary rock type and age data are given in 54 55 371 Table 1. Photographs of analysed rocks are shown in Fig. S4 with results plotted in Figs. 6-9 56 57 58 372 and S5. A detailed description of the methods undertaken can be found in the Supplementary 59 60 373 Section with raw data given in the Supplementary Data tables. 61 62 63 64 16 65 374 4.1. Rennell Island Ridge 1 2 3 375 4.1.1. South-west side (DR2) 4 5 6 7 376 Only sedimentary rocks were recovered from this dredge site on the southwest side of 8 9 377 Indispensable Reefs. Sample DR2C is a typical rock from this location. It consists of 10 11 12 378 moderately hard olive green calcareous mudstone containing radiolaria and benthic 13 14 379 foraminifera giving a late Early Miocene (N7/8) age. Most species groups have wide depth 15 16 17 380 ranges but the low planktonic content and lack of diversity combined with diverse benthics 18 19 381 suggests shelf conditions. Sample DR2E is a grey-cream bored mudstone containing a 20 21 382 22 reasonable fauna of planktonic and benthic foraminifera of Middle Eocene (E10-12) age. 23 24 383 Judging by the planktonic percentage and type of benthics present, deposition appears to have 25 26 384 taken place in continental slope depths (200-2000 m). DR2H is a hard and bioturbated cream- 27 28 29 385 grey mudstone with a prominent 1 cm thick graded sandy siltstone bed. On disaggregation, 30 31 386 the mudstone yielded common radiolaria along with a few sponge spicules, echinoid spine 32 33 34 387 fragments and low diversity foraminifera. The latter are of Middle-Late Eocene (E10-E15) 35 36 388 age. Thin section contains some simple, large agglutinated benthic foraminifera 37 38 39 389 (Rhizammina/Hyperammina). The benthic foraminifera in DR2E and DR2H are similar, 40 41 390 suggesting the material is from the same, continental slope sequence. 42 43 44 391 45 The thin sandstone bed in DR2H is a heavy mineral layer consisting of detrital plagioclase, 46 47 392 chlorite, titanomagnetite, ilmenite and chromite. No quartz, mica or zircon was present. 48 49 393 Chromite chemistry indicates derivation from an eroding volcanic, rather than an exhumed 50 51 52 394 peridotite, source (Fig. 7). 53 54 55 395 4.1.2. North-east side (DR3) 56 57 58 59 60 61 62 63 64 17 65 396 In contrast, at the DR3 dredge site on the northeast side of Indispensable Reefs, basalts were 1 2 397 recovered along with sedimentary rocks. DR3C is a friable cream-coloured calcareous 3 4 5 398 mudstone coating on the outside of basalt DR3Aiv. It consists dominantly of planktonic 6 7 399 faunas of Middle Eocene (E13) and Quaternary (N22/23) age, deposited in a bathyal 8 9 10 400 environment. DR3F comprises separate pieces of soft, friable muddy micritic limestone, not 11 12 401 in physical contact with basalt. Excellent Middle Eocene (E11) planktonic fauna were 13 14 402 Nuttallides truempyi 15 obtained and also the deep-water benthic form . DR3H mainly comprises 16 17 403 hard coral fragments; a non-coral piece of the sample contained foraminifera along with 18 19 404 fragments of corals and bryozoans. The faunas are mixed and of Pliocene (N21) and 20 21 22 405 Pleistocene (N22) age. Foraminifera types present are consistent with a shallower depth of 23 24 406 deposition than the ooze samples of DR3C and DR3F. 25 26 27 28 407 The basalts from DR3 are mainly grey, sparsely plagioclase- and augite-porphyritic pillow 29 30 408 lavas with glassy rinds (e.g., P82182). Some microdoleritic textured samples are present (e.g., 31 32 33 409 P82183) and may be hypabyssal intrusions or the coarse interiors of lava flows. Chemical 34 35 410 analyses reveal them to be similar, but not identical, subalkaline basalts. Multi-element 36 37 411 38 normalised patterns (Fig. S5) show good coherency between incompatible element 39 40 412 concentrations with P82182 being light rare-earth element (REE) enriched and heavy REE 41 42 413 depleted with respect to P82183. P82182 is also more oceanic-island basalt-like than P82183. 43 44 45 414 Nonetheless, both samples have compositions within the overall range of enriched mid-ocean 46 47 415 ridge basalt (E-MORB). 48 49 50 51 416 Plagioclase from the same two basalts was dated by Ar-Ar methods. The gas release 52 53 417 spectrum from P82182 plagioclase (Fig. 8A) is slightly U-shaped, but generally flat. K/Ca is 54 55 56 418 low (0.002 to 0.005); radiogenic yields are 30-60% then drop. A pseudoplateau age based on 57 58 419 the two steps at the bottom of the 'U' gives an age of 42.3 ± 1.2 Ma (2 sigma). This is 59 60 61 62 63 64 18 65 420 indistinguishable from an inverse isochron age of 41.6 ± 4.6 Ma based on the 750-1040°C 1 2 421 steps (see Supplementary File). Plagioclase from the P82183 microdolerite has a similar, 3 4 5 422 slightly U-shaped spectrum but higher K/Ca ratios (Fig. 8B). A weighted mean plateau age 6 7 423 (WMPA) using 57% of the gas is 43.15 ± 1.0 Ma. This is within error of the isochron age of 8 9 10 424 43.5 ± 2.2 Ma using the same gas steps. Our preferred interpreted ages for P82182 and 11 12 425 P82183 plagioclases are 42.0 ± 2.0 Ma and 43.2 ± 1.5 Ma, respectively. These are arrived at 13 14 426 15 by rounding the plateau ages and by conservatively lowering the precision. A weighted 16 17 427 average of these ages is 42.8 ± 1.2 Ma. 18 19 20 428 4.2. West Torres Plateau (DR4) 21 22 23 24 429 More than 99% of the dredge consisted of one rock type: amygdaloidal pillow basalt with 25 26 430 glassy rinds and sparse, clay-altered olivine phenocrysts (e.g., P82188). Colour ranges from 27 28 29 431 grey to dark orange depending on alteration. Amygdules are filled with calcite and chlorite. 30 31 432 Some lavas are slightly more porphyritic with up to 10% clay-altered olivine phenocrysts 32 33 34 433 (e.g., P82189). Manganese crusts and rinds up to 9 cm thick are present on all samples. 35 36 37 434 On the largest boulder some white, non-calcareous material forms matrix to hyaloclastite 38 39 40 435 breccia and was subsampled as DR4Aiv. The few microfossils in it are dominantly in the 41 42 436 form of chamber internal moulds. A few echinoid spines and coelenterate fragments are 43 44 437 45 present. Only two identifiable specimens are noted. Both belong to the robust genus 46 47 438 Catapsydrax also known to resist dissolution and diagenesis. If they are C. stainforthi then 48 49 439 the sample is Early Miocene (N4-N7) but if they are C. unicavus then a broader, Late Eocene 50 51 52 440 to Early Miocene (E13-N6) age range is indicated. Planktonic foraminifera in separate lumps 53 54 441 of soft, pale brown ashy mudstone (DR4E) give a Quaternary (N22) age. 55 56 57 58 442 Glass from P82188 and holocrystalline whole rocks from P82188 and 82189 were analysed 59 60 443 for whole rock geochemistry. We attribute the small compositional variation between glass 61 62 63 64 19 65 444 and whole rock (e.g., separation between the two red dots in Fig. 9) to element mobility in the 1 2 445 slightly clay-altered whole rock. The DR4 West Torres Plateau samples are quite different 3 4 5 446 from those at DR3 and have the overall geochemistry of subalkaline basalts from normal 6 7 447 mid-ocean ridge basalt (N-MORB) suites. An N-MORB rather than arc-related origin for the 8 9 10 448 DR4 samples is corroborated by the chrome spinel microphenocryst chemistry (Fig. 7). 11 12 13 449 Four magnetic splits of plagioclase and groundmass separates were made from P82189. Here 14 15 16 450 we report the results for the analytically best (the least magnetic, most plagioclase-rich) 17 18 451 fraction. The argon release spectrum climbs then flattens into broad plateau (Fig. 8C). K/Ca 19 20 452 21 ratios are fair, as are radiogenic yields. As with the samples from DR3, the weighted mean 22 23 453 plateau of 26.2 ± 0.8 Ma and inverse isochron age of 27.0 ± 1.7 Ma for the same gas release 24 25 454 steps both agree within error. We report a preferred age of 26.2 ± 0.8 Ma for the eruption and 26 27 28 455 cooling of this West Torres Plateau basalt. 29 30 31 456 4.3. East Lapérouse Rise 32 33 34 35 457 4.3.1. Northern part (DR5) 36 37 38 458 No igneous rocks were obtained from this dredge, which consisted of ~300 kg of limestones. 39 40 41 459 DR5Aiii is a subsample of hard, cream-coloured recrystallised and cemented biocalcarenite. 42 43 460 Thin section examination reveals common poorly preserved, thin-walled globigerinid 44 45 46 461 planktonic foraminifera, lack of keeled globorotalids, and a sphaeroidinellid. Small benthic 47 48 462 forms include miliolids and Globocassidulina. Large forms include Amphistegina (consistent 49 50 463 with A. hauerina), and large, thin lepidocycline species. There are some echinoid spines. 51 52 53 54 464 DR5Av is another limestone subsample, of cavitated calcirudite containing broken coral 55 56 465 clasts. A variety of planktonic and large (photic zone) foraminifera are present. Following 57 58 59 466 Chaproniere (1984) an Oligocene-Early Miocene (N2-4) age estimate is based on the benthic 60 61 62 63 64 20 65 467 foraminifera Lepidocyclina (Eulepidina) ?ephippioides, Heterostegina borneensis and 1 2 468 Amphistegina hauerina. The presence of Miogypsinoides in DR5Aiii reinforces the DR5Av 3 4 5 469 age estimate and the presence of the sphaeroidinellid planktonic form in DR5Aiii suggests a 6 7 470 Miocene rather than Oligocene age for the faunas. Planktonic ooze of Quaternary (N22) age 8 9 10 471 is present on both DR5 samples. 11 12 13 472 4.3.2. Central part (DR6) 14 15 16 17 473 The only sample obtained from this dredge is a single Mn crusted cobble of orange-grey 18 19 474 plagioclase- and olivine-porphyritic basalt (P82193). A 5-10 mm thick glass rind adheres to 20 21 475 22 one side of the rock. The glass has the geochemistry of a typical E-MORB (Fig. 9). This is 23 24 476 supported by the chrome spinel microphenocryst chemistry that straddles the MORB and 25 26 477 OIB fields (Fig. 7). 27 28 29 30 478 Plagioclase from P82193 gives an excellent flat argon release spectrum although K/Ca ratios 31 32 479 are very low (Fig. 8D). Radiogenic yields are good for most of the spectrum (60-88%). With 33 34 35 480 the well-defined plateau and isochron, the biggest uncertainty stems from large Ca 36 37 481 corrections. The WMPA (92% of gas) yields 39.3 ± 0.6 Ma and inverse isochron age of 39.1 38 39 40 482 ± 1.2 Ma. Conservatively, we choose the isochron age to be representative of the cooling and 41 42 483 eruption of this basalt lava. 43 44 45 46 484 4.4. South Rennell Trough 47 48 49 485 4.4.1. West scarp (DR13) 50 51 52 53 486 A moderate quantity of limestone was obtained from the dredge, but no other rock types. A 54 55 487 hard, bored, pale, cream-coloured foram limestone with a 5 mm think manganese oxide rind, 56 57 488 lacking corals or shell fragments, (DR13Ai) was chosen for micropaleontological study. The 58 59 60 489 disaggregated residue consists of planktonic foraminifera ooze. The moderately diverse 61 62 63 64 21 65 490 species indicate a late Middle to middle Late Miocene age (N14-N17A) and a -deep-water 1 2 491 environment. 3 4 5 6 492 4.4.2. East Scarp ‘Ski Jump’ (DR14) 7 8 9 493 More than 99% of this dredge consisted of basalts and associated volcaniclastic breccia, 10 11 12 494 sandstone and mudstone. Some slabs of the finer grained sedimentary lithologies were 13 14 495 studied to try and date the volcaniclastic rocks. DR14Ai is a yellow-brown calcareous 15 16 17 496 volcaniclastic mudstone, penetrated by manganese oxide nuclei and stringers. Sponge 18 19 497 spicules, radiolaria, diatoms, echinoid spines, minute bivalves and ostracods are present. 20 21 498 22 Planktonic foraminifera are quite common and are dominated by large simple globigerinid 23 24 499 species. They are of Pliocene (N19-21) age. The diversity of accompanying invertebrates 25 26 500 suggests moderately shallow water but this is normally in conflict with the presence of 27 28 29 501 radiolaria. 30 31 32 502 Of the many individual basalt samples recovered from this 400 kg dredge, at least 15 separate 33 34 35 503 volcanic lithotypes were identified on the basis of phenocryst content and grain size. 36 37 504 Geochemical analyses were made of five representative samples: P82204 is a dark grey to 38 39 40 505 orange olivine-phyric basalt with a glassy rind, P82206 is a non-glassy basalt with 15% clay- 41 42 506 altered olivine phenocrysts, P82212 is an aphyric basalt, P82214 is a highly amygdaloidal (c. 43 44 507 45 30%) basalt and P82217 is a possible olivine-, augite and plagioclase-phyric basalt. 46 47 48 508 Geochemically, P82206 and 82214 are picritic (>12 wt% MgO and >180 ppm Ni). However, 49 50 509 these two are not more depleted in any trace elements than the other three and, together, all 51 52 53 510 five samples appear to be from the same igneous suite. Element ratios and concentrations 54 55 511 seem typical of E-MORB compositions and transitional to ocean island basalt (OIB) (Fig. 9). 56 57 58 512 Chrome spinel compositions from P82206 (DR14Eiii) are also transitional between these lava 59 60 513 types (Fig. 7). 61 62 63 64 22 65 514 Two of the basalts were selected for dating. P82206 was fresh and unaltered enough for 1 2 515 groundmass to be selected (the only groundmass dated in this sample set). It has a moderately 3 4 5 516 flat argon release spectrum. Ages monotonically decrease during the first 75% indicating 6 7 517 recoil in this part of the spectrum, K/Ca ratios are reasonably high (0.02 to 0.9), but 8 9 10 518 radiogenic yields are very low (12-25%). Three steps (670-770°C) yield a small weighted 11 12 519 pseudoplateau of age 30.1 ± 0.5 Ma with an accompanying isochron age of 30.8 ± 1.6 Ma. 13 14 520 15 However, given the fine grain size of the groundmass and the clear evidence for recoil makes 16 17 521 us wary of over interpreting it. Our preferred age is 30.0 ± 2.0 Ma, a combination of the 18 19 522 central age with decreased precision. 20 21 22 23 523 Plagioclase from basalt P82217 gave an excellent flat spectrum, with reasonable K/Ca ratios 24 25 524 (0.005 to 0.063) and radiogenic yields (60-95%). Analytically, this was one the highest 26 27 28 525 quality samples in our dataset. Weighted mean plateau and isochron ages are both well 29 30 526 defined at 34.7 ± 0.4 Ma (79% of gas) and 34.3 ± 1.2 Ma (100% of gas), respectively (Fig. 31 32 33 527 8F). Our preferred interpreted age for eruption and cooling of this lava is the weighted mean 34 35 528 plateau age with a slightly decreased precision, namely 34.7 ± 0.5 Ma. 36 37 38 39 529 5. Magnetic anomaly interpretation 40 41 42 530 Much of the eastern Coral Sea suffers from a paucity of magnetic anomaly profiles. During 43 44 531 SS2012_V06 45 voyage , a total of 6,200 km of magnetic anomaly data were collected using a 46 47 532 SeaSPY high sensitivity total field magnetometer. Profiles focused in the Santa Cruz and 48 49 533 d’Entrecasteaux Basins. These data were combined into an in-house ship track marine 50 51 52 534 geophysical database maintained at the University of Sydney (primarily contains data from 53 54 535 the NGDC (National Geophysical Data Centre) database) and compared to the geomagnetic 55 56 57 536 reversal timescale. The structural interpretation, in particular the abyssal hill fabric imaged 58 59 537 by swath profiles, was used to guide the orientation and direction of spreading. We 60 61 62 63 64 23 65 538 interpreted a sequence of magnetic anomalies using the timescale of Gee and Kent (2007), 1 2 539 and created a set of magnetic anomaly identifications using the picking convention as 3 4 5 540 described in Seton et al. (2014). The magnetic anomaly identifications were used as the 6 7 541 primary constraint in developing a tectonic history (and associated tectonic rotation 8 9 10 542 parameters) for basin opening. 11 12 13 543 5.1. Santa Cruz Basin 14 15 16 17 544 Previous estimates of the age of the oceanic lithosphere within the Santa Cruz Basin have 18 19 545 ranged from Cretaceous (Kroenke, 1984; Schellart et al., 2006) to Eocene-Oligocene (Larue et 20 21 22 546 al., 1977; Mortimer et al., 2014; Wells, 1989; Yan and Kroenke, 1993) (see Section 2). 23 24 547 Recent dating of E-MORB samples from the South Rennell Trough and Santa Cruz Basin 25 26 548 (Mortimer et al., 2014) supports the Oligocene interpretation. However, no attempt has been 27 28 29 549 made to interpret the magnetic anomaly data in the Santa Cruz Basin to place constraints on 30 31 550 the basin evolution. 32 33 34 35 551 We collected three full and four partial profiles over the Santa Cruz Basin during voyage 36 37 552 SS2012_V06 to supplement existing magnetic data (Fig. 10). We identify magnetic anomalies 38 39 40 553 13o-20o (33.5-43.8 Ma) in the Santa Cruz Basin (Fig. 10 and 11), with younger anomalies 41 42 554 present close to the ridge crest and potentially older anomalies in the northwest corner of the 43 44 555 45 basin, perhaps up to anomaly 21o (47.9 Ma). Our interpretation is consistent with the age 46 47 556 constraints provided by the dated dredge samples from the area (see Section 4 and Mortimer 48 49 557 et al. (2014)). The seafloor-spreading direction, as inferred from tracing the magnetic 50 51 52 558 anomalies between profiles, is ESE-WNW close to the ridge, consistent with the abyssal hill 53 54 559 trends identified in the swath bathymetry data. The change in spreading direction is dated to 55 56 57 560 ~36-35 Ma. We model three changes in spreading rate during the opening of the Santa Cruz 58 59 561 Basin, with a slow rate of approximately 45 mm/yr during the initial stages of seafloor 60 61 62 63 64 24 65 562 spreading from anomalies 20o-18o (43.8-40.1 Ma) and a decrease to 30 mm/yr from 1 2 563 anomalies 18o-13y (40.1-33.1 Ma) (Fig. 11). Close to the ridge, where the identification of 3 4 5 564 magnetic anomalies is problematic, we model a further decrease in seafloor-spreading rate to 6 7 565 around 10 mm/yr, in line with estimates along ultra-slow seafloor-spreading ridges (Dick et 8 9 10 566 al., 2003). 11 12 13 567 While no fracture zones were identified in the swath bathymetry or gravity anomalies, the 14 15 16 568 abyssal hill fabric observed in the swath data and the correlation of the magnetic anomalies 17 18 569 across profiles leads us confidently to interpret the spreading direction as being 19 20 570 21 predominantly NW-SE directed. 22 23 24 571 The timing of spreading cessation in the Santa Cruz Basin inferred from the magnetic 25 26 572 anomaly data is difficult to quantify. Approximately 30-40 km of crust exists between the 27 28 29 573 youngest identified anomaly and the axial ridge of the mid-ocean ridge (clearly identified in 30 31 574 swath data). The difficulty in interpreting the data here may reflect a combination of a 32 33 34 575 decrease in seafloor spreading rate to ultra-slow and increasing tectonic and magmatic 35 36 576 complexity during final stages of seafloor spreading. Reducing the seafloor-spreading rate to 37 38 39 577 around 10 mm/yr predicts a minimum date for cessation of spreading in the Santa Cruz Basin 40 41 578 to 28-29 Ma, consistent with the basalt ages (28 Ma, (Mortimer et al., 2014)) from the crest 42 43 579 44 of the South Rennell Trough. However, as spreading may have waned for several million 45 46 580 years, we infer that 28 Ma is the minimum time of spreading cessation. 47 48 49 581 The initiation of seafloor-spreading in the basin is also difficult to determine because much of 50 51 52 582 the northwestern flank of the basin and a portion of the eastern flank have been lost to 53 54 583 subduction. The oldest confidently identified magnetic anomaly is 20o (43.8 Ma) but there is 55 56 57 584 the potential for anomaly 21o (47.9 Ma) in the northwest. One limiting factor on the age of 58 59 585 initiation of spreading in the Santa Cruz Basin may be whether or not space is available 60 61 62 63 64 25 65 586 between the oldest mag netic anomaly on the south-eastern flank and the West Torres Plateau 1 2 587 to accommodate anomaly 21o. With this consideration and an assumption of slow spreading 3 4 5 588 rates, we assume that the initiation of spreading could have occurred from around 48 Ma or 6 7 589 slightly earlier. However, this is only a valid constraint if the West Torres Plateau was a pre- 8 9 10 590 existing feature and not erupted on top of older seafloor. If the West Torres Plateau is 11 12 591 confirmed to be a volcanic feature that formed after or during the opening of the Santa Cruz 13 14 592 15 Basin, then the maximum age of formation for the Santa Cruz Basin can be extended by an 16 17 593 additional 4-5 million years. 18 19 20 594 21 Our interpretation of the magnetic data from the Santa Cruz Basin confirms an Eocene- 22 23 595 Oligocene age for the basin with the inferred spreading centre a continuation of the South 24 25 596 Rennell Trough. This is in agreement with a recent reinterpretation and age dating of several 26 27 28 597 ORSTOM samples from the South Rennell Trough as Eocene-Oligocene (Mortimer et al., 29 30 598 2014). 31 32 33 34 599 5.2. d’Entrecasteaux Basin 35 36 37 600 The existing magnetic anomaly profiles in the d’Entrecasteaux Basin have been used to 38 39 40 601 interpret the basin as having formed in the Cretaceous (83-65 Ma; anomalies 30-34) with 41 42 602 three different spreading orientations: E-W, NE-SW and WNW-ESE (Lapouille, 1982) (Fig. 43 44 603 45 12). According to Lapouille (1982), the d’Entrecasteaux Basin has a shared spreading history 46 47 604 with the North Loyalty Basin to the south, separated by the d’Entrecasteaux Zone after the 48 49 605 cessation of seafloor-spreading. Other studies have instead suggested a shared Cretaceous 50 51 52 606 spreading history between the d’Entrecasteaux and Santa Cruz Basins (Kroenke, 1984; 53 54 607 Schellart et al., 2006). 55 56 57 58 608 We collected four profiles across the d’Entrecasteaux Basin during voyage SS2012_V06 to 59 60 609 supplement existing magnetic data (Fig. 12). An interpretation of the age of the basin via 61 62 63 64 26 65 610 magnetic anomalies remains difficult due to the lack of magnetic anomaly pattern continuity 1 2 611 between profiles (Fig. 12). However, a combination of structural data from the basin, 3 4 5 612 integrated with new seafloor-spreading interpretations and age-dating from the surrounding 6 7 613 areas allows us to discount several previous scenarios for basin formation. 8 9 10 11 614 Firstly, the prevailing interpretation of Lapouille (1982) (Fig. 12) can no longer be used to 12 13 615 date basin opening as the magnetic lineations cross-cut the structural trends from the three 14 15 16 616 zones that make up the d’Entrecasteaux Basin (Fig. 5 and 12). The N-S and NE-SW 17 18 617 structural trends cover the area interpreted by Lapouille (1982) as having formed via NW-SE 19 20 618 21 and ENE-WSW spreading. In addition, anomaly 34 of Lapouille (1982) cross-cuts the South 22 23 619 Rennell Trough where a clear NW-SE spreading direction and Oligocene opening age has 24 25 620 been established (see previous section). 26 27 28 29 621 Secondly, the Santa Cruz Basin and d’Entrecasteaux Basin cannot share the same Cretaceous 30 31 622 opening history, as suggested by Kroenke (1984) and Schellart et al. (2006). A Cretaceous 32 33 34 623 opening history would require that the age of the Santa Cruz Basin is also Cretaceous but this 35 36 624 basin has unequivocally been dated to Eocene-Miocene with a shared spreading history along 37 38 39 625 the South Rennell Tough (Mortimer et al., 2014). An alternative possibility that the 40 41 626 d’Entrecasteaux and Santa Cruz basins still shared a common spreading history but during 42 43 627 44 the Eocene-Oligocene, with the d’Entrecasteaux Basin reflecting one flank of the oldest 45 46 628 portion of the South Rennell Trough, can also be discounted as this interpretation would 47 48 629 require that oceanic crust of equivalent extent was present to the west of the West Lapérouse 49 50 51 630 Rise (assuming spreading symmetry) or that a period of short-lived subduction occurred close 52 53 631 to it. There is no geological or geophysical evidence to suggest either scenario is possible. 54 55 56 57 632 Thirdly, the d’Entrecasteaux Basin and the North Loyalty Basin cannot have a shared 58 59 633 spreading history, as suggested by Lapouille (1982), as the structural trend in the 60 61 62 63 64 27 65 634 d’Entrecasteaux Basin , which lies adjacent to the North Loyalty Basin is almost 1 2 635 perpendicular. Age dates from along the d’Entrecasteaux Zone, which range from ~50 to 30 3 4 5 636 Ma (Mortimer et al., 2014) indicate that there was a tectonic boundary between these two 6 7 637 basins during the opening of the North Loyalty Basin (Sdrolias et al., 2003). 8 9 10 11 638 6. Discussion 12 13 14 639 6.1. Opening history of the South Rennell Trough and Santa Cruz Basin 15 16 17 18 640 Combining the petrological and microfossil analyses of dredge samples, structural 19 20 641 interpretation of the region and magnetic anomaly interpretation of the Santa Cruz Basin, we 21 22 23 642 create a rotation model for the opening of the South Rennell Trough and Santa Cruz Basin as 24 25 643 a single spreading system (Fig. 13). Our finite rotations for the opening of the Santa Cruz 26 27 644 28 Basin and South Rennell Trough, computed using the visual-fitting technique (i.e. manual- 29 30 645 fitting of data) in the plate reconstruction software, GPlates (Boyden et al., 2011), can be 31 32 646 found in the Supplementary Data. 33 34 35 36 647 We model the initiation of basin opening to have begun at 48 Ma in the Santa Cruz Basin 37 38 648 based on the possibility of magnetic anomaly 21o being present in the northwest of the Santa 39 40 41 649 Cruz Basin, with southward propagation occurring along the South Rennell Trough. The 42 43 650 agreement in ages of the independently-dated volcanic and sedimentary rocks at site DR3 on 44 45 46 651 the south-eastern edge of the Rennell Island Ridge lends strong support to the idea that the 47 48 652 age of formation of igneous crust at these sites is at least middle Eocene. A middle Eocene 49 50 653 age for the East Lapérouse Rise (DR6) and a late Eocene to early Oligocene age for the 51 52 53 654 southern end of the South Rennell Trough is also clear from the data and consistent with the 54 55 655 magnetic anomaly data for basin opening from the Santa Cruz Basin. 56 57 58 59 60 61 62 63 64 28 65 656 The southward propagating ridge stops in the south where it intersects with the Bellona 1 2 657 Plateau, along the Lord Howe Seamount chain (one of several southward-younging hotspot 3 4 5 658 trails in the SW Pacific). The pole of rotation for the initial opening was centred near the 6 7 659 Bellona Plateau at 17.52°S, 158.63°E and migrated to 17.18°S, 158.82°E at the time of 8 9 10 660 spreading cessation. As such, seafloor spreading rates vary quite significantly along strike 11 12 661 from between 30-45 mm/yr in the Santa Cruz Basin to 10-30 mm/yr in the southern part of 13 14 662 15 the South Rennell Trough during the main period of basin opening. This results in a 16 17 663 characteristic fan-shaped basin opening pattern. 18 19 20 664 21 Cessation of seafloor spreading is primarily constrained from the dating of MORB at the 22 23 665 ridge crest at 28 Ma (Mortimer et al., 2014) and is also consistent with the magnetic anomaly 24 25 666 interpretation in the Santa Cruz Basin (see Section 5.1). Spreading rates may have decreased 26 27 28 667 to as low as 10 mm/yr in the Santa Cruz Basin at around this time, consistent with typical 29 30 668 <12 mm/yr rates for other ultra-slow spreading systems (Dick et al., 2003). Ultra-slow 31 32 33 669 spreading systems report E-MORB glass compositions from these ridge axes (e.g. Mühe et al. 34 35 670 (1997)), similar to the compositions we report, further supporting an ultra-slow spreading 36 37 671 38 regime. However, it is important to note that a wider range of N-MORB and K-rich 39 40 672 compositions have also been sampled in ultra-slow spreading regimes (Michael et al., 2003; 41 42 673 Nauret et al., 2011). The final timing of cessation of basin opening can be constrained at 28 43 44 45 674 Ma as a minimum age, but may have waned for several million years until the complete 46 47 675 depletion of the magma supply. We therefore suggest an age range for basin cessation 48 49 50 676 between ~28-25 Ma. Other ultra-slow spreading systems also exhibit a gradual decrease in 51 52 677 spreading rate until final termination of spreading (e.g. Labrador Sea (Delescluse et al., 53 54 55 678 2015)). 56 57 58 59 60 61 62 63 64 29 65 679 Our model for the South Rennell-Santa Cruz spreading system is of an ultra-slow, back-arc 1 2 680 spreading system. We infer this spreading to have taken place in a back-arc basin setting 3 4 5 681 because, although exact SW Pacific Paleogene paleogeography is speculative, arc-related 6 7 682 igneous rocks of Eocene age have been reported from more Pacific-ward positions than the 8 9 10 683 South Rennell Trough. These include the d’Entrecasteaux Ridge (Mortimer et al., 2014), 11 12 684 New Caledonia (Cluzel et al., 2006), Fiji (Todd et al., 2012), Vanuatu (Buys et al., 2014) and 13 14 685 15 the Tonga forearc (Meffre et al., 2012). 16 17 18 686 Compositionally, all lavas thus far sampled (at five sites) in and around the South Rennell 19 20 687 21 Trough spreading system are distinctly E-MORB-like (Fig. 9). These plot further along the 22 23 688 mantle array towards OIB than any lavas from other SW Pacific back-arc basins including 24 25 689 the South Fiji, North Loyalty and Coral Sea basins. The possible Lord Howe plume influence 26 27 28 690 on the South Rennell Trough lavas is explored in a companion paper (see Williams et al. 29 30 691 (2014)). Alternatively, E-MORBs may be indicative of an Indian mantle-derived MORB 31 32 33 692 source rather than Pacific mantle, an explanation used by Mortimer et al. (2012) for Tasman 34 35 693 Sea Basin MORBs. Whatever the explanation, the E-MORB lava types of the South Rennell 36 37 694 38 Trough do seem geographically unique. 39 40 41 695 An intriguing aspect of the South Rennell-Santa Cruz spreading system is the nature of the 42 43 696 44 elevated and bathymetrically symmetrical East and West Lapérouse Rises. These rises are 45 46 697 modelled to be the magmatic footprints of an ultra-slow spreading ridge and may be 47 48 698 analogous to the magmatic or amagmatic ridge structures, a key component of ultra-slow 49 50 51 699 spreading systems, as reported by Dick et al. (2003). In our model for basin opening, we 52 53 700 assume that the East and West Lapérouse Rises were formed during the opening of the South 54 55 56 701 Rennell Trough rather than being pre-existing ridge features that rifted apart with the 57 58 702 establishment of the South Rennell spreading ridge. This does not preclude that some 59 60 61 62 63 64 30 65 703 inherited continental or other older volcanic material may have been isolated during the 1 2 704 initial stages of basin opening. 3 4 5 6 705 6.2. Tectonic evolution of the eastern Coral Sea 7 8 9 706 The tectonic evolution of the eastern Coral Sea encompasses the history of the Melanesian 10 11 12 707 and New Hebrides Arcs, the northeast Australian and Papuan margins and the tectonic 13 14 708 processes occurring north of New Caledonia. We develop our plate tectonic model within the 15 16 17 709 context of the results presented in this paper, recognising that there are many alternative 18 19 710 models for the wider SW Pacific and SE Asian region. Our model is merged with the global 20 21 711 22 plate tectonic model of M l ler et al. (2016) with modifications in Zahirovic et al. (in review). 23 24 25 712 We initiate our model at 85 Ma (late Cretaceous), where we firstly consider the 26 27 713 28 characteristics of the eastern Gondwana margin, which form the western and southern passive 29 30 714 margins of the present-day eastern Coral Sea and provide a basis for interpreting material 31 32 715 dispersed throughout the SW Pacific. 33 34 35 36 716 6.2.1. Late Cretaceous-Paleocene 37 38 39 717 A series of recent studies on the island arcs of the SW Pacific have identified continental- 40 41 42 718 derived zircons from the Australian margin sitting within the island arcs of the Tonga Arc 43 44 719 (Falloon et al., 2014), Vanuatu (Buys et al., 2014) and the Solomon Islands (Tapster et al., 45 46 47 720 2014) suggesting that many of these island arcs may have origins from Gondwana rather than 48 49 721 being purely intra-oceanic in nature. The Melanesian Arc, for example, has previously been 50 51 722 52 modelled as an intra-oceanic arc formed by spontaneous subduction initiation sometime in 53 54 723 the Paleocene (e.g. Gaina and Müller (2007); (Hall, 2002); Yan and Kroenke (1993)), which 55 56 724 conflicts with the new evidence from Tapster et al. (2014) and may support the model for 57 58 59 725 subduction initiation at relic island arcs proposed by Leng and Gurnis (2015). Although 60 61 62 63 64 31 65 726 recognised, the timing for rifting of these fragments from the Gondwana margin remains 1 2 727 unconstrained. In our model, we follow a simple scenario whereby portions of the 3 4 5 728 Melanesian and New Hebrides Arc (Vanuatu) rifted from the Gondwana margin (adjacent to 6 7 729 present-day Queensland, see Buys et al. (2014)) during the Cretaceous (Fig. 14A-B). We 8 9 10 730 model this rifting phase to be contemporaneous with a major change in the tectonic regime 11 12 731 along eastern Gondwana in the Cretaceous, which eventually led to the dispersal of 13 14 732 15 continental blocks and the subsequent opening of the Tasman and Coral Seas (Gaina et al., 16 17 733 1998; Gaina et al., 1999) (Fig. 14C). We follow the model of Gaina et al. (1998) for the 18 19 734 opening of the Tasman Sea at 90 Ma (late Cretaceous) and Gaina et al. (1999) for the opening 20 21 22 735 of the Coral Seas and Louisiade Trough (and initiation of the Australian-Lord Howe- 23 24 736 Louisiade triple junction) at 61 Ma (Early Paleocene). Our model additionally includes the 25 26 27 737 contemporaneous opening of a series of back-arc basins oceanward of the Tasman and Coral 28 29 738 Seas, such as the South Loyalty Basin adjacent to the Lord Howe Rise (Cluzel et al., 2001), 30 31 32 739 according to the model of Matthews et al. (2015), and a new model for the opening of the 33 34 740 newly named Emo Basin adjacent to the Coral Sea and PNG (Fig. 14B). 35 36 37 741 38 Our model for the Emo Basin is based on trace element, rare earth element and isotopic 39 40 742 signatures of the Emo Metamorphics (part of the Owen Stanley Metamorphic Belt, Papuan 41 42 743 Peninsula) that indicate tholeiitic influence with normal mid-ocean ridge basalt and enriched 43 44 45 744 mid-ocean ridge basalt compositions that formed in a back-arc setting (Worthing and 46 47 745 Crawford, 1996). While the age of the Emo Metamorphics is poorly constrained, 48 49 50 746 Maastrichtian (~71-65 Ma) aged foraminifera are found in carbonate lenses within the 51 52 747 structurally equivalent Goropu Metabasalt (part of the Milne Terrain) suggesting that the 53 54 55 748 Emo Metamorphics protolith likely existed at this time (Smith and Davies, 1976; Worthing 56 57 749 and Crawford, 1996). A Maastrichtian age is also proposed for crystallisation of the Papuan 58 59 750 Ultramafic Belt (PUB) Ophiolite based on dating of foraminifera found in sediments 60 61 62 63 64 32 65 751 intercalated in the basalts (Belford, 1976; Smith and Davies, 1976), suggesting that the basin 1 2 752 that formed the underlying crust is at least this old. The Late Cretaceous crystallisation age of 3 4 5 753 the supra-subduction oceanic crust that forms the ophiolites is consistent with our interpreted 6 7 754 age of north-dipping subduction initiation along the Sepik Arc to consume the latest Jurassic 8 9 10 755 to Early Cretaceous Sepik Basin (Zahirovic et al., in review). The Sepik composite terrane is 11 12 756 interpreted to be a composite terrane with continental and intra-oceanic crust, which is 13 14 757 15 separated from New Guinea by the Central Ophiolite Belt, and separated from the younger 16 17 758 (South) Caroline Arc to the north by the strike-slip faults that make up the Papuan Mobile 18 19 759 Belt, following the terminology of Hall (2002) and Zahirovic et al. (2014). Lus et al. (2004) 20 21 22 760 dated hornblende samples from the metamorphic sole of the PUB to determine the timing of 23 24 761 cooling of the sole, and obtained a mean age of 58.3 ± 0.4 Ma using 40Ar/39Ar dating, 25 26 27 762 providing a minimum age of formation of the Emo Basin. The Milne Terrain, to the southeast 28 29 763 of the PUB, consists of upper Cretaceous and Eocene rocks of MORB-type affinity (Smith, 30 31 32 764 2013). Although these ages are consistent with those of the Emo Metamorphics, Smith (2013) 33 34 765 argued that geochemical data indicate that the Milne Terrain formed by the opening of the 35 36 766 37 Coral Sea Basin rather than a back-arc basin such as those of the Emo Metamorphics. 38 39 40 767 We model the opening of the Emo Basin to occur from 85 Ma (late Cretaceous), coincident 41 42 768 with the opening of the South Loyalty Basin further south (Matthews et al., 2015). The 43 44 45 769 obduction of the Emo Metamorphics and Papuan Ultramafic Belt are modelled to occur at 58 46 47 770 Ma (late Paleocene), induced by the opening of the Coral Sea basin at 61 Ma (Fig. 14C). As 48 49 50 771 these back-arc basins opened, elements of the Melanesian and New Hebrides Arc were 51 52 772 isolated from the Gondwana margin (Fig. 14A-B). 53 54 55 56 773 Previous models suggest southward-dipping subduction along the Melanesian and New 57 58 774 Hebrides Arcs and Fiji, which together formed the Vitiaz Arc in the Eocene, with a 59 60 61 62 63 64 33 65 775 connection to the proto-Tonga-Kermadec Ridge further south (Gaina and Müller, 2007; Hall, 1 2 776 2002; Schellart et al., 2006; Sdrolias et al., 2003; Yan and Kroenke, 1993). The presence of 3 4 5 777 arc-related volcanism from the Melanesian Arc (Jajao Igenous Suite on Malaita and Santa 6 7 778 Isabel islands, Solomon Islands) between ~62-46 Ma (Tejada et al., 1996), evidence for late 8 9 10 779 Eocene volcanism in New Britain (Lindley, 1988) and subsequent U-Pb dating of this 11 12 780 volcanism (Holm et al., 2013), Ar-Ar and U-Pb dating of arc-rocks from Vanuatu indicating 13 14 781 15 late Eocene-Miocene arc volcanism (Buys et al., 2014), and Eocene to early Miocene arc- 16 17 782 volcanism in Fiji (see Taylor et al. (2000) for references) all support subduction occurring 18 19 783 along this arc system at this time. In addition, the supra-subduction zone nature of the Jajao 20 21 22 784 Igneous Suite (similar to BABB) may constrain the timing for back-arc basin formation 23 24 785 behind the Melanesian Arc (Tejada et al., 1996). Although limited arc-related rocks exists 25 26 27 786 that record active subduction prior to the Eocene, our model predicts that southward-dipping 28 29 787 subduction along this arc system has occurred since at least the Late Cretaceous (Fig. 14), 30 31 32 788 with a likely increase in subduction roll-back after the initiation of seafloor spreading in the 33 34 789 Coral Sea at 61 Ma (Gaina et al., 1999), cessation of spreading in the Emo Basin and the 35 36 790 37 initiation of its closure. Extending the subduction record back to the Cretaceous is based on a 38 39 791 combination of the Late Cretaceous-aged arc rocks of the Jajao Igneous Suite (Tejada et al., 40 41 792 1996); ages and interpretations of South Loyalty Basin crust obducted onto New Caledonia 42 43 44 793 based on the extensive review of Matthews et al. (2015), which supports Cretaceous 45 46 794 subduction along eastern Gondwana; and plate kinematic and geodynamic considerations 47 48 49 795 such as the requirement of a subduction zone to drive back-arc opening in the Emo Basin. 50 51 796 The phase of arc volcanism recorded in the rocks from the Vitiaz Arc system from the 52 53 797 54 Eocene may reflect the plate boundary readjustments that would have occurred synchronous 55 56 798 with these events. After the initiation of Emo Basin closure, we model an oceanward jump in 57 58 799 the location of back-arc basin opening behind the Melanesian portion of Vitiaz subduction 59 60 61 62 63 64 34 65 800 zone, forming the oldest crust in the Melanesian Basin (its present-day remnants residing in 1 2 801 the Solomon Sea) (Fig. 14C). Further south, spreading continued in the South Loyalty Basin 3 4 5 802 behind a west-dipping subduction zone (Matthews et al., 2015). 6 7 8 803 6.2.2. Early Eocene-Late Oligocene 9 10 11 12 804 A change in plate boundaries in the SW Pacific between New Caledonia and New Zealand 13 14 805 occurred in the Early Eocene at 55 Ma (Cluzel et al., 2006; Crawford et al., 2003; Lagabrielle 15 16 17 806 et al., 2013; Matthews et al., 2015; Ulrich et al., 2010; Whattam et al., 2008), initiating a 18 19 807 period of unambiguous subduction east of the Lord Howe Rise (Matthews et al., 2015) (Fig. 20 21 808 22 14C). This subduction zone initiated the closure of the South Loyalty Basin behind an east- 23 24 809 dipping subduction zone possibly along the Loyalty Arc (Matthews et al., 2015), while 25 26 810 southwest-dipping subduction was continuing further north along the Vitiaz Trench (Fig. 27 28 29 811 14C). We connect these two opposite-verging subduction zones via a right-lateral transform 30 31 812 boundary along the newly established d’Entrecasteaux Zone (Fig. 14C). This new plate 32 33 34 813 boundary isolated the northern part of the South Loyalty Basin north of the d’Entrecasteaux 35 36 814 Zone from the rest of the basin, which was being consumed behind an east-dipping 37 38 39 815 subduction zone along the Loyalty Ridge. As there is no geological or geophysical evidence 40 41 816 for subduction north of New Caledonia adjacent to the northern Lord Howe Rise, Kenn 42 43 817 44 Plateau and Mellish Rise, we propose that the trapped piece of the South Loyalty Basin north 45 46 818 of the d’Entrecatseaux Zone may underlie the present-day d’Entrecasteaux Basin (zone 1; see 47 48 819 Fig. 5 and 8 and 14C). Although we are unable to resolve the age of basin formation from the 49 50 51 820 magnetic anomaly data, this assumption does resolve the difficulty in opening the 52 53 821 d’Entrecasteaux Basin based on our understanding of the surrounding plate boundaries, the 54 55 56 822 geometry of the basin itself and the lack of an identifiable extinct ridge. However, fully 57 58 823 testing this hypothesis requires age dating of the crust and further magnetic anomaly and 59 60 61 62 63 64 35 65 824 swath bathymetry profiles. The ~55 Ma subduction initiation event may have led to the 1 2 825 waning of seafloor spreading in the Tasman and Coral Sea, where spreading ceased by the 3 4 5 826 Early Eocene, at 52 Ma (Gaina et al., 1998; Gaina et al., 1999). 6 7 8 827 Seafloor-spreading in the Santa Cruz Basin and South Rennell Trough began in the Early 9 10 11 828 Eocene, at ~48 Ma (see section 6.1) (Fig. 13 and 14D). Prior to the opening of this spreading 12 13 829 system, we assume that the Rennell Island Ridge and West Torres Plateau formed a 14 15 16 830 continuous ridge, likely volcanic or arc-volcanic in nature. The structural trends observed in 17 18 831 the bathymetry and gravity data indicate NW-SE directed motion along the western margin of 19 20 832 21 the West Torres Plateau, consistent with the structural trends (trend T1; Fig. 2) along the 22 23 833 Rennell Island Ridge. If this hypothesis is correct, samples from the core of the West Torres 24 25 834 Plateau should have similar ages and geochemical affinities to the Rennell Island Ridge. 26 27 28 835 However, as no samples from the West Torres Plateau proper are available, this remains an 29 30 836 open question. We model a connection between the NW-SE-directed spreading in the Santa 31 32 33 837 Cruz-South Rennell system with the NE-SW directed spreading in the Melanesian Basin 34 35 838 (opening behind the Melanesian Arc) in a triple junction setting, forming a basin that exceeds 36 37 839 38 2000 km in length (Fig. 14D-E). Triple junctions and spreading centre complexity are 39 40 840 common features in back-arc basins (e.g. North Fiji Basin (Auzende et al., 1988), South Fiji 41 42 841 Basin (Herzer et al., 2011; Sdrolias et al., 2003), Lau Basin (Taylor et al., 1996), Parece Vela 43 44 45 842 and Shikoku Basins (Sdrolias et al., 2004)). We model the crust forming along the north- 46 47 843 western arm of this triple junction (Melanesian Basin component) as being the crust that now 48 49 50 844 underlies the Solomon Sea. The Solomon Sea has been interpreted as opening either from 60- 51 52 845 40 Ma (Davies and Price, 1986; Joshima et al., 1986) or 45-30 Ma (Falvey and Pritchard, 53 54 55 846 1982), with chrons 19-15 (45-35 Ma) identified (Gaina and Müller, 2007). These ages are 56 57 847 consistent with the timing of Santa Cruz-South Rennell opening (Fig. 14D-E). 58 59 60 61 62 63 64 36 65 848 To accommodate the NW-SE-directed opening of the Santa Cruz-South Rennell spreading 1 2 849 system, compression or convergence would have occurred either to the northwest or 3 4 5 850 southeast. Indications of subduction activity along the d’Entrecasteaux Zone during the time 6 7 851 of basin opening (48-25 Ma) are sparse but arc/BAB type basalts with age ranges of around 8 9 10 852 30-38 Ma have been reported (Mortimer et al., 2014). We suggest that the pre-existing short- 11 12 853 lived right-lateral transform boundary along the d’Entrecasteaux Zone was a pre-existing 13 14 854 15 zone of weakness, which may have allowed for the localization of subduction during the 16 17 855 opening of the Santa Cruz-South Rennell spreading system (Fig. 13 and 14D-E). In our 18 19 856 model, approximately 300 km of the trapped South Loyalty Basin crust that now underlies 20 21 22 857 the d’Entrecasteaux Basin was consumed during this southward-dipping subduction phase 23 24 858 and may also have induced back-arc spreading in the North Loyalty Basin, dated to have 25 26 27 859 initiated at around 44 Ma (Sdrolias et al., 2003). The southward-dipping subduction along the 28 29 860 d’Entrecasteaux Zone connected with the east-dipping subduction along the Loyalty Ridge 30 31 32 861 (Meffre et al., 2012), which was consuming crust formed during South Loyalty Basin 33 34 862 spreading (Fig. 14D-E). It is important to note that the concept of this period of subduction 35 36 863 37 along the d’Entrecasteaux Zone is model-dependent. 38 39 40 864 While spreading continued along the Melanesia-South Rennell-Santa Cruz triple junction 41 42 865 between 44-25 Ma (Eocene-Oligocene), subduction was ongoing to the west of the region, 43 44 45 866 north of PNG along the Sepik Arc, consuming back-arc basin crust from the Cretaceous 46 47 867 Sepik Basin (Zahirovic et al., in review). We extend this northward-dipping subduction zone 48 49 50 868 towards the southeast, consuming the remainder of the Emo Basin crust (Fig. 14E). The 51 52 869 closure and obduction of the Sepik and Emo Basin occurred in the Late Oligocene, at ~30-25 53 54 55 870 Ma (Zahirovic et al., in review; Zahirovic et al., 2014) synchronous with other collisional 56 57 871 events in the area. 58 59 60 61 62 63 64 37 65 872 6.2.3. Late Oligocene-Miocene 1 2 3 873 A major change in plate boundary configurations occurred at around 25 Ma in the SW 4 5 6 874 Pacific-SE Asia. The Ontong Java Plateau, the world’s largest LIP, is believed to have 7 8 875 arrived at the Vitiaz Trench (Fig. 14F) at 25 Ma (Holm et al., 2013; Petterson et al., 9 10 11 876 1999). The arrival is assumed to be marked by a soft collision with no major associated 12 13 877 uplift (Petterson et al., 1999) and it has been proposed that it affected Australian plate motion 14 15 16 878 (Knesel et al., 2008). Although other interpretations for the timing of collision exist (e.g., 17 18 879 Oligocene based on paleomagnetic constraints (Musgrave, 2013) and Pliocene based on 19 20 880 21 marine geophysical data (Cowley et al., 2004), we adhere to the 25 Ma age as consistent with 22 23 881 the geological record from the Melanesian Arc (Petterson et al., 1999). The youngest rocks 24 25 882 dredged from the ridge axis along the South Rennell Trough and Santa Cruz Basin have ages 26 27 28 883 of ~28 Ma (Middle Oligocene) (Mortimer et al., 2014). We suggest that complete cessation 29 30 884 of activity along the Melanesia-Santa Cruz-South Rennell system may not have occurred 31 32 33 885 until 25 Ma (Late Oligocene), at which point the collision of the Ontong Java Plateau with 34 35 886 the Melanesian Arc would have induced a reconfiguration of plate boundaries on the over- 36 37 887 38 riding plate, leading to a termination of magma supply to the ridge. Southward-dipping 39 40 888 subduction, albeit at a reduced rate, continued along the Melanesian subduction zone as the 41 42 889 plateau was being subducted. To accommodate continued convergence between the Pacific 43 44 45 890 and Australian plates, we support the concept of southward-directed subduction landward of 46 47 891 the Melanesian subduction zone along the Trobriand Trough, as previously suggested by 48 49 50 892 Petterson et al. (1999), Hall (2002), Hill and Hall (2003) and Schellart et al. (2006), and 51 52 893 related igneous rock emplacement in the Late Oligocene in the New Guinea Mobile Belt 53 54 55 894 (Cloos et al., 2005; Hill and Raza, 1999). This subduction zone was established after the 56 57 895 accretion of the Sepik composite terrane to PNG (Zahirovic et al., in review; Zahirovic et al., 58 59 896 2014) and consumed crust that formed during the earliest stage of Melanesian Basin opening 60 61 62 63 64 38 65 897 (Fig. 14F). Within the eastern Coral Sea, we model ~200 km of extension occurring in Zone 1 2 898 2 of the d’Entrecasteaux Basin, translating the West Torres Plateau eastward via right-lateral 3 4 5 899 transform motion along the d’Entrecasteaux Zone (Fig. 13 and 14F). The enigmatic 26 Ma 6 7 900 (Late Oligocene) basalt sample from a spur of the West Torres Plateau (Fig. 8) may reflect 8 9 10 901 the initial stages of extension in this small basin. To the southeast of the area, it has been 11 12 902 proposed that cessation of South Fiji Basin opening occurred at 25 Ma (Sdrolias et al., 2003). 13 14 903 15 Although recent redating of magnetic anomalies with additional geophysical data has placed 16 17 904 the cessation of South Fiji Basin opening at 15 Ma (Herzer et al., 2011), this model implies a 18 19 905 change in spreading direction at ~25 Ma. 20 21 22 23 906 Continued convergence subsequently caused major uplift and compression along the 24 25 907 Melanesian Arc. By the late Miocene (~12 Ma), a reversal in the polarity of subduction 26 27 28 908 occurred (Cooper and Taylor, 1985; Petterson et al., 1999) as continued subduction of the 29 30 909 Ontong Java Plateau choked the subduction system during hard collision. Subduction was re- 31 32 33 910 initiated along the present-day South Solomon/San Cristobal Trench at 12 Ma, leading to the 34 35 911 subduction of crust from the Melanesian-Santa Cruz-South Rennell spreading system beneath 36 37 912 38 the Pacific Plate (Fig. 14G). Subduction along the Papuan Peninsula continued, forming the 39 40 913 Maramuni Arc (Holm et al., 2015) and may have coincided with a subduction zone between 41 42 914 the Louisiade Arc and Woodlark Island at 15 Ma (Webb et al., 2014) (Fig. 14G). To the north 43 44 45 915 of the area, subduction along the New Britain trench is believed to have begun in the late 46 47 916 Miocene (Cooper and Taylor, 1985; Johnson, 1979) and north-dipping subduction in New 48 49 50 917 Ireland at 12 Ma based on subsidence and local volcanism (Stewart and Sandy, 1988), 51 52 918 coincident with the subduction polarity reversal along the Melanesian Arc. To the south, 53 54 55 919 paleomagnetic evidence exists for substantial clockwise rotation of the New Hebrides Arc 56 57 920 since the late Miocene (Falvey, 1975; Musgrave and Firth, 1999) and for a large 58 59 921 anticlockwise rotation of Fiji starting at or prior to 10 Ma and ceasing at 3 Ma (Taylor et al., 60 61 62 63 64 39 65 922 2000). We model the late stage evolution of New Hebrides and Fiji based on these 1 2 923 paleomagnetic results but suggest that the initiation of their rotation was at 12 Ma, consistent 3 4 5 924 with the subduction polarity reversal event in Melanesia (Fig. 14G). We model the 6 7 925 abandonment of the Maramuni Arc (related to Trobriand subduction) due to the collision of 8 9 10 926 the Torricelli-Finisterre Terranes and South Caroline Arc with New Guinea from ~15 Ma to 11 12 927 form the Papuan Mobile Belt, with a diachronous collision younging eastwards. 13 14 15 16 928 6.2.4. Late Miocene-present day 17 18 19 929 To the north of the eastern Coral Sea, the opening of the Woodlark Basin between the 20 21 930 22 remnant Melanesian Basin (Solomon Sea) and the Pocklington Trough (Fig. 1) occurred from 23 24 931 around 7 Ma (Late Miocene) (Fig. 14H). The opening of the Woodlark Basin is modelled 25 26 932 using the rates of rotation of 2-4°/Myr (Taylor et al., 1999) from 3.6 million years to present 27 28 29 933 but extended to 7 Ma. These rates of seafloor-spreading may have decreased to 1.9°/Myr in 30 31 934 the recent past (Wallace et al., 2014). The Solomon Sea (containing a remnant piece of the 32 33 34 935 Melanesia Basin) lies to the north of the Woodlark Basin. As there is no present-day evidence 35 36 936 for subduction between the Woodlark Basin and Solomon Sea (Bird, 2003), our model 37 38 39 937 implies that the Solomon Sea underwent minor rotation tied to the opening of the Woodlark 40 41 938 Basin. Subduction is accommodated to the north and west of the Solomon Sea, along the 42 43 939 44 New Britain Trench. 45 46 47 940 The present-day plate boundary configuration in the area is complex, with some plate 48 49 941 boundaries still poorly-defined (Bird, 2003) (Fig. 1 and 14I). The eastern Coral Sea sits at the 50 51 52 942 north-eastern corner of the Australian Plate, where the crust of the Santa Cruz Basin is being 53 54 943 actively subducted along the South Solomon/San Cristobal Trench and New Hebrides 55 56 57 944 trenches (Fig. 1, 13 and 14I). The south, west and northwest of the region are tectonically 58 59 945 quiescent. Further north, seafloor spreading is active in the Woodlark Basin, where incipient 60 61 62 63 64 40 65 946 continental rifting is occurring as the rift intersects with the PNG mainland to the west 1 2 947 (Taylor et al., 1995), and ridge subduction/slab window formation where the ridge intersects 3 4 5 948 with the South Solomon/San Cristobal Trench to the east (Chadwick et al., 2009; Crook and 6 7 949 Taylor, 1994). 8 9 10 11 950 7. Conclusions 12 13 14 951 We have combined a structural interpretation of swath bathymetry data, petrological and 15 16 17 952 microfossil analysis of dredge samples and an interpretation of marine geophysical data from 18 19 953 the eastern Coral Sea to propose a new model for the opening of the Santa Cruz Basin and 20 21 954 22 South Rennell Trough and the wider Melanesian arc and back-arc region. We constrain the 23 24 955 opening of the Santa Cruz Basin and South Rennell Trough to between ~48-25 Ma based on 25 26 956 a combination of magnetic anomaly interpretations, where chrons 20o-13y have been 27 28 29 957 identified, and age-dating of dredge samples. Geochemical analysis of the basalts from the 30 31 958 Rennell Island Ridge and South Rennell Trough shows they resemble E-MORB basalts, 32 33 34 959 similar in composition to the recently re-analysed and dated ORSTOM dredges (Mortimer et 35 36 960 al., 2014). The cessation of spreading along the South Rennell Trough, Santa Cruz Basin and 37 38 39 961 Melanesia Basin/Solomon Sea occurred by at least 28 Ma with possible complete termination 40 41 962 by 25 Ma, corresponding to a regional reorganization of the plate boundaries in the area, 42 43 963 44 possibly attributed to the initial soft collision of the Ontong Java Plateau. 45 46 47 964 Our model proposes that the Melanesian Basin (Solomon Sea) began at ~61 Ma and by 48 48 49 965 Ma was simultaneously opening with the Santa Cruz-South Rennell spreading system in a 50 51 52 966 triple junction configuration. The 55 Ma major change in plate boundary configurations in 53 54 967 the SW Pacific south of New Caledonia led to the establishment of the d’Entrecasteaux Zone 55 56 57 968 as a plate boundary, separating the SW Pacific from the Melanesian-SE Asia region and may 58 59 969 have trapped South Loyalty Basin crust in the present-day d’Entrecasteaux 60 61 62 63 64 41 65 970 Basin. Complexity in the plate boundaries in the area started from around 48 Ma, with the 1 2 971 establishment of the Melanesia-Santa Cruz-South Rennell spreading system, with two 3 4 5 972 subduction zones in the area by 40 Ma, one along the Melanesian/Vitiaz Arc bordering the 6 7 973 Pacific Plate and another close to PNG along the Sepik Arc, with the latter likely active since 8 9 10 974 the Late Cretaceous. This initiated a period of subduction zone complexity, whereby crust 11 12 975 created in the Melanesian Basin was being actively subducted. The 25 Ma regional plate 13 14 976 15 reorganization event is reflected in the cessation of back-arc basin activity, the establishment 16 17 977 of the Maramuni subduction zone along PNG and the isolation of the Solomon Sea 18 19 978 Plate. This was followed by a reorganization of plate boundaries at 12 Ma, leading to the 20 21 22 979 establishment of northward-dipping subduction along the South Solomon and New Hebrides 23 24 980 subduction zones and setting the scene for the present-day configuration of plate boundaries. 25 26 27 28 981 Our plate reconstructions, which include a network of continuously closing plate boundaries 29 30 982 can form the basis of a model of plate boundary continuity between the SW Pacific and SE 31 32 33 983 Asia throughout the Cenozoic, and can be used to address questions such as the behaviour of 34 35 984 intra-oceanic subduction zones and their relationship to back-arc processes and major oceanic 36 37 985 38 plateau accretion events. 39 40 41 986 Acknowledgements 42 43 44 987 R/V Southern Surveyor 45 We thank the Captain and crew of and Australia’s Marine National 46 47 988 Facility (MNF) for the success of voyage SS2012_V06. Mineral separation was done by John 48 49 989 Simes and Belinda Smith Lyttle. We thank Hugh Davies for his attempts to locate old 50 51 52 990 Solomon Sea dredge samples. We thank Geoscience Australia and New Caledonia agencies 53 54 991 for voyage support. We would like to acknowledge funding from the Australian Research 55 56 57 992 Council through grants DP0987713 and FT130101564 (MS), FL0992245 (SEW), 58 59 60 61 62 63 64 42 65 993 IH130200012 (SZ) and DP130101946 (KJM). Support from the New Zealand Government 1 2 994 through core funding to GNS Science (NM) is also acknowledged. 3 4 5 6 995 Figures 7 8 9 996 Figure 1: Regional bathymetry compilation of the eastern Coral Sea from Geoscience 10 11 12 997 Australia, with hill-shade using gravity gradients from Sandwell et al. (2014). Red stars mark 13 14 998 the location of dredges from voyage SS2012_V06 used in this study, white stars mark the 15 16 17 999 location of dredges from voyage SS2012_V06 not used in this study and grey line denotes the 18 191000 ship track from voyage SS2012_V06. Rennell Island is coloured white as opposed to other 20 211001 22 land coloured light yellow. BP = Bellona Plateau, CSB = Coral Sea Basin, DEZ = 23 241002 d’Entrecasteaux Zone, ELR = East Lapérouse Rise, LHR = Lord Howe Rise, NCB = New 25 261003 Caledonia Basin, NCR = New Caledonia Ridge, NLB = North Loyalty Basin, RI = Rennell 27 28 291004 Island, WLR = West Lapérouse Rise, WSB = West Santo Basin. 30 31 321005 Figure 2: Regional Geoscience Australia bathymetry compilation with gravity gradients 33 34 351006 (Sandwell et al., 2014) for the Rennell Island Ridge and Rennell Basin. Plate boundaries are 36 371007 from Bird (2003). The two main structural trends in the region are marked as black dotted 38 39 401008 lines. Localised structural interpretation from bathymetry and gravity data shown as orange 41 421009 lines, dredge locations from voyage SS2012_V06 marked as red stars, other dredge locations 43 441010 45 marked as green stars. ELR = East Lapérouse Rise, LP = Louisiade Plateau, LT = Louisiade 46 471011 Trough, SCB = Santa Cruz Basin, SRT = South Rennell Trough, WLR = West Lapérouse 48 491012 Rise. 50 51 52 531013 Figure 3: Basemap and key same as in Fig. 2 but for the South Rennell Trough. BP = 54 551014 Bellona Plateau. 56 57 58 59 60 61 62 63 64 43 65 1015 Figure 4: Basemap and key same as in Fig. 2 but for the Santa Cruz Basin. A detailed view 1 21016 of the swath bathymetry data can be found in Fig. S2. 3 4 5 61017 Figure 5: Basemap and key same as in Fig. 2 but for the West Torres Plateau and 7 81018 d’Entrecasteaux Basin. A detailed view of the swath bathymetry data can be found in Fig. S3. 9 10 11 121019 Figure 6: Summary of new rock types with microfossil and Ar-Ar ages from samples 13 141020 dredged on voyage SS2012_V06. 15 16 17 181021 Figure 7: An Al2O3 vs TiO2 bivariate plot with chrome spinel geochemistry from four 19 201022 dredge sites and reference fields from Kamenetsky et al. (2001). Sample DR2H has affinities 21 22 231023 with arc-related basalts, indicating the sediments around Rennell Island derived from an 24 251024 eroding volcanic source. In contrast, DR4B from the West Torres Plateau is of an N-MORB 26 271025 28 rather than arc-related origin. Samples DR6 and DR14iii, from the East Lapérouse Rise and 29 301026 the South Rennell Trough, are both transitional to arc-related basalts, consistent with an E- 31 321027 MORB chemistry. Abbreviations of locations in key same as Fig. 2. 33 34 35 361028 Figure 8: Argon-argon degassing spectra for six plagioclase and groundmass separates. 37 381029 Corresponding K/Ca ratios are in green. Uncertainties shown in this figure are ±2σ. 39 40 411030 TFA=total fusion age, WMPA=statistically valid weighted mean plateau age, 42 431031 PseudoPA=subjectively selected pseudoplateau age. 44 45 46 471032 Figure 9: Binary diagrams illustrating the enriched mid-ocean ridge basalt (E-MORB) nature 48 491033 of lavas from the South Rennell Trough spreading system (green symbols). A. Ti vs V after 50 511034 B 52 Shervais (1982); . Nb/Yb vs Th/Yb after Pearce (2008). New data shown in symbols with 53 541035 thick black borders. Other data from Davies and Price (1986), Peate et al. (1997) and 55 561036 (Mortimer et al., 2012; Mortimer et al., 2014; Mortimer et al., 2007). N-MORB-normal mid- 57 58 591037 ocean ridge basalt, BABB=back-arc basin basalt, OIB=ocean island basalt. 60 61 62 63 64 44 65 1038 Figure 10: Magnetic anomaly profiles over the Santa Cruz Basin. Pink profiles are from 1 21039 existing publicly available datasets, blue profiles are from the ECOSAT survey presented in 3 4 51040 this paper. Shiptrack profile names used in Figure 11 are annotated. Dredge samples are 6 71041 marked as red stars (this study) and green stars from Mortimer et al. (2014). Dredge samples 8 9 101042 (this study) with reported Ar-Ar ages are annotated. Plate boundaries are from Bird (2003). 11 121043 Magnetic anomaly identifications are presented as coloured circles (see key for chrons and 13 141044 15 ages). 16 17 181045 Figure 11: Stacked magnetic anomaly profiles for the Santa Cruz Basin. Ship profiles used 19 201046 21 as annotated and labelled on Figure 10. Synthetic profile was created using the Modmag 22 231047 software (Mendel et al., 2005) assuming a magnetic source layer at a depth of 4 km, with a 24 251048 thickness of 0.5 km, magnetization of 10 A/m, and a contamination coefficient of 0.5. 26 27 281049 Modelled spreading rates of 45 mm/yr (chrons 20-18) and 30 mm/yr (chrons 18-13) were 29 301050 used for the main period of basin opening, while 10 mm/yr was used to model the waning of 31 32 331051 the spreading system to the extinct ridge at 25 Ma. 34 35 361052 Figure 12: Magnetic anomaly profiles over the d’Entrecasteaux Basin. Pink profiles are from 37 38 391053 existing publicly available datasets, blue profiles are from voyage SS2012_V06 presented in 40 411054 this paper. Dredge samples are marked as red stars (this study) and green stars from Mortimer 42 431055 44 et al. (2014). Dredge samples (this study) with reported Ar-Ar ages are annotated. Plate 45 461056 boundaries are from Bird (2003). Magnetic anomaly interpretation of Lapouille (1982) 47 481057 plotted as light blue dotted lines. 49 50 51 521058 Figure 13: Plate tectonic reconstructions in a fixed Australia reference frame for the Santa 53 541059 Cruz Basin/South Rennell Trough region starting at 52 Ma (marks the end of spreading in the 55 56 571060 Tasman and Coral Seas and Louisiade Trough) to the present-day. Coloured polygons denote 58 591061 different plates involved in the system (see key for plate names). Plate boundaries are 60 61 62 63 64 45 65 1062 marked in black (refer to Fig. 14 for names of plate boundary features), with open triangles 1 21063 for the subduction zones denoting model-dependent convergence. Main bathymetric features 3 4 51064 shown as thin black outlines. Green line denotes the South Rennell spreading axis: active 6 71065 (solid), waning (dashed), extinct (dotted). ELR = East Lapérouse Rise, KP = Kenn Plateau, 8 9 101066 LP = Louisiade Plateau, MR = Mellish Rise, NBP = North Bellona Plateau, RIR = Rennell 11 121067 Island Ridge, WLR = West Lapérouse Rise, WTP = West Torres Plateau. 13 14 15 161068 Figure 14: Regional plate tectonic reconstructions in a mantle frame of reference for the SW 17 181069 Pacific from 85 Ma to the present-day. Plate boundaries are marked in black, with open 19 201070 21 triangles for subduction zones denoting model-dependent convergence. Basemap shows the 22 231071 age of the oceanic crust in the region. Purple polygons are volcanic related products from the 24 251072 region, including Large Igneous Provinces and hotspot trails (note that the Louisiade Plateau 26 27 281073 is tentatively marked as a LIP, see Section 2). Main plates are labelled in bold black text: 29 301074 AUS = Australia, CS = Caroline Sea, PAC = Pacific, PMP = Proto-Molucca, PSP = 31 32 331075 Philippine Sea, SFP = South Fiji, SSP = Solomon Sea. Features mentioned in text are 34 351076 annotated in grey text: CSB = Coral Sea Basin, DEB = d’Entrecasteaux Basin, DEZ = 36 371077 38 d’Entrecasteaux Zone, EM = Emo Basin, ESZ = Emo subduction zone, LHR = Lord Howe 39 401078 Rise, MA = Melanesia Arc, MB = Melanesia Basin, MSZ = Maramuni subduction zone, NFB 41 421079 = North Fiji Basin, NHA = New Hebrides Arc, NHSZ = New Hebrides subduction zone, 43 44 451080 NLB = North Loyalty Basin, OJP = Ontong Java Plateau, SCSR = Santa Cruz-South Rennell 46 471081 spreading system, SLB = South Loyalty Basin, SLSZ = South Loyalty subduction zone, 48 49 501082 SOSZ = South Solomon subduction zone, SSZ = Sepik subduction zone, TSZ = Tonga- 51 521083 Kermadec subduction zone, VSZ = Vitiaz subduction zone, WB = Woodlark Basin. 53 54 55 561084 Table Captions 57 58 591085 Table 1. Summary of dredge localities, rock types and analyses. 60 61 62 63 64 46 65 1086 Table 2. Summary of Ar-Ar dating 1 2 31087 References 4 5 6 71088 Acharyya, S.K., 1998. Break-up of the greater Indo-Australian continent and accretion of 81089 blocks framing South and East Asia. J. Geodyn. 26, 149-170. 91090 Auzende, J.-M., Lafoy, Y., Marsset, B., 1988. Recent geodynamic evolution of the north Fiji 10 111091 basin (southwest Pacific). Geology 16, 925-929. 121092 Belford, D., 1976. Appendix: Foraminifera and age of samples from south-eastern Papua. 131093 Geology of the South-eastern Papuan Mainland, Bull. Bur. Miner. Resour., Geol. Geophys., 141094 Aust 165, 73-82. 15 161095 Bird, P., 2003. An updated digital model of plate boundaries. Geochemistry, Geophysics, 171096 Geosystems 4, 1027, doi:1010.1029/2001GC000252. 181097 Boyden, J.A., Müller, R.D., Gurnis, M., Torsvik, T.H., Clark, J.A., Turner, M., Ivey-Law, H., 191098 Watson, R.J., Cannon, J.S., 2011. Next-generation plate-tectonic reconstructions using 20 211099 GPlates, In: Keller, G.R., and Baru, C (Ed.), Geoinformatics: Cyberinfrastructure for the 221100 Solid Earth Sciences. Cambridge University Press, pp. 95-114. 231101 Bryan, S.E., Ernst, R.E., 2008. Revised definition of large igneous provinces (LIPs). Earth- 241102 Science Reviews 86, 175-202. 25 261103 Buck, W.R., Lavier, L.L., Poliakov, A.N., 2005. Modes of faulting at mid-ocean ridges. 271104 Nature 434, 719-723. 281105 Buys, J., Spandler, C., Holm, R.J., Richards, S.W., 2014. Remnants of ancient Australia in 291106 Vanuatu: Implications for crustal evolution in island arcs and tectonic development of 30 311107 the southwest Pacific. Geology 42, 939-942. 321108 Chadwick, J., Perfit, M., McInnes, B., Kamenov, G., Plank, T., Jonasson, I., Chadwick, C., 331109 2009. Arc lavas on both sides of a trench: Slab window effects at the Solomon Islands 341110 triple junction, SW Pacific. Earth and Planetary Science Letters 279, 293-302. 351111 Chaproniere, G.C., 1984. The Neogene larger foraminiferal sequence in the Australian 36 371112 and New Zealand regions, and its relevance to the East Indies letter stage classification. 381113 Palaeogeography, palaeoclimatology, palaeoecology 46, 25-35. 391114 Cloos, M., Sapiie, B., van Ufford, A.Q., Weiland, R.J., Warren, P.Q., McMahon, T.P., 2005. 401115 Collisional delamination in New Guinea: The geotectonics of subducting slab breakoff. 41 421116 Geological Society of America Special Papers 400, 1-51. 431117 Cluzel, D., Aitchison, J.C., Picard, C., 2001. Tectonic accretion and underplating of mafic 441118 terranes in the Late Eocene intraoceanic fore-arc of New Caledonia (Southwest Pacific): 451119 geodynamic implications. Tectonophysics 340, 23-59. 46 471120 Cluzel, D., Meffre, S., Maurizot, P., Crawford, A.J., 2006. Earliest Eocene (53 Ma) 481121 convergence in the Southwest Pacific: evidence from pre‐obduction dikes in the 491122 ophiolite of New Caledonia. Terra Nova 18, 395-402. 50 511123 Cochran, J.R., Kurras, G.J., Edwards, M.H., Coakley, B.J., 2003. The Gakkel Ridge: 521124 Bathymetry, gravity anomalies, and crustal accretion at extremely slow spreading rates. 531125 Journal of Geophysical Research: Solid Earth (1978–2012) 108. 541126 Coffin, M., Duncan, R., Eldholm, O., Fitton, J., Frey, F., Larsen, H., Mahoney, J., Saunders, A., 551127 Schlich, R., Wallace, P., 2006. Large igneous provinces and scientific ocean drilling: 56 571128 Status quo and a look ahead. Oceanography 19, 150-160. 581129 Coleman, P.J., Packham, G.H., 1976. The Melanesian Borderlands and India—Pacific 591130 plates' boundary. Earth-Science Reviews 12, 197-233. 60 61 62 63 64 47 65 1131 Cooper, P.A., Taylor, B., 1985. Polarity reversal in the Solomon Islands arc. 11132 Cowley, S., Mann, P., Coffin, M., Shipley, T.H., 2004. Oligocene to Recent tectonic history 21133 of the Central Solomon intra-arc basin as determined from marine seismic reflection 3 41134 data and compilation of onland geology. Tectonophysics 389, 267-307. 51135 Cowley, S., Mann, P., Gahagan, L., 1998. Tectonic history of the northern Louisiade 61136 Plateau based on deep penetration seismic reflection lines. Eos Trans. AGU 79, 870. 71137 Crawford, T., Meffre, S., Symonds, P.A., 2003. Tectonic Evolution of the SW Pacific: 8 91138 Lessons for the Geological Evolution of the Tasman Fold Belt System in eastern 101139 Australia from 600 to 220 Ma., In: Hillis, R., Müller, R.D. (Eds.), Evolution and Dynamics 111140 of the Australian Plate. 121141 Crook, K.A., Taylor, B., 1994. Stucture and Quaternary tectonic history of the Woodlark 13 141142 triple junction region, Solomon Islands. Marine Geophysical Researches 16, 65-89. 151143 Daniel, J., Jouannic, C., Larue, B.M., Recy, J., 1978. Marine geology of eastern Coral Sea. 161144 South Pacific Marine Geological Notes 1, 81-94. 171145 Davies, H., Price, R., 1986. Basalts from the Solomon and Bismarck seas. Geo-marine 18 191146 letters 6, 193-202. 201147 De Broin, C., Aubertin, F., Ravenne, C., 1977. Structure and history of the Solomon-New 211148 Ireland region, International Symposium on Geodynamics in the Southwest Pacific. 221149 Editions Technip. Paris, pp. 37-50. 23 241150 Delescluse, M., Funck, T., Dehler, S.A., Louden, K.E., Watremez, L., 2015. The oceanic 251151 crustal structure at the extinct, slow to ultra‐slow Labrador Sea spreading center. 261152 Journal of Geophysical Research: Solid Earth. 271153 Dick, H.J., Lin, J., Schouten, H., 2003. An ultraslow-spreading class of ocean ridge. Nature 28 291154 426, 405-412. 301155 Exon, N., Hill, P., Lafoy, Y., Heine, C., Bernardel, G., 2006. Kenn Plateau off northeast 311156 Australia: a continental fragment in the southwest Pacific jigsaw. Australian Journal of 321157 Earth Sciences 53, 541-564. 33 341158 Falloon, T.J., Meffre, S., Crawford, A.J., Hoernle, K., Hauff, F., Bloomer, S.H., Wright, D.J., 351159 2014. Cretaceous fore-arc basalts from the Tonga arc: Geochemistry and implications 361160 for the tectonic history of the SW Pacific. Tectonophysics 630, 21-32. 371161 Falvey, D., 1975. Part 2. Younger Pacific arcs and the eastern marginal seas: Arc 38 391162 reversals, and a tectonic model for the North Fiji Basin. Exploration Geophysics 6, 47-49. 401163 Falvey, D., Pritchard, T., 1982. Paleomagnetic Evidence for Large Microplate Rotations in 411164 the Southwest Pacific: ABSTRACT. AAPG Bulletin 66, 966-966. 421165 Falvey, D.A., Taylor, L.W., 1974. Queensland Plateau and Coral Sea Basin: structural and 43 441166 time-stratigraphic patterns. Exploration Geophysics 5, 123-126. 451167 Gaina, C., Müller, D.R., Royer, J.Y., Stock, J., Hardebeck, J., Symonds, P., 1998. The tectonic 461168 history of the Tasman Sea: a puzzle with 13 pieces. Journal of Geophysical Research 103, 471169 12413-12433. 481170 Gaina, C., Müller, R.D., 2007. Cenozoic tectonic and depth/age evolution of the 49 501171 Indonesian gateway and associated back-arc basins. Earth-Science Reviews 83, 177-203. 511172 Gaina, C., Müller, R.D., Royer, J.-Y., Symonds, P., 1999. The tectonic evolution of the 521173 Louisiade Triple Junction. Journal of Geophysical Research 104, 12973-12939. 531174 Gee, J., Kent, D., 2007. Source of Oceanic magnetic anomalies and the geomagnetic 54 551175 polarity timescale. Treatrise Geophys 5, 455ñ507. 561176 Golonka, J., 2004. Plate tectonic evolution of the southern margin of Eurasia in the 571177 Mesozoic and Cenozoic. Tectonophysics 381, 235-273. 58 59 60 61 62 63 64 48 65 1178 Gurnis, M., Turner, M., Zahirovic, S., DiCaprio, L., Spasojevic, S., Müller, R.D., Boyden, J., 11179 Seton, M., Manea, V.C., Bower, D.J., 2012. Plate tectonic reconstructions with 21180 continuously closing plates. Computers & Geosciences 38, 35-42. 3 41181 Hall, R., 2002. Cenozoic geological and plate tectonic evolution of Southeast Asia and the 51182 SW Pacific: computer-based reconstructions, models and animations. Journal of Asian 61183 Earth Sciences 20, 353-431. 71184 Herzer, R., Barker, D., Roest, W., Mortimer, N., 2011. Oligocene‐Miocene spreading 8 91185 history of the northern South Fiji Basin and implications for the evolution of the New 101186 Zealand plate boundary. Geochemistry, Geophysics, Geosystems 12. 111187 Hill, K.C., Hall, R., 2003. Mesozoic-Cenozoic evolution of Australia's New Guinea margin 121188 in a west Pacific context. Geological Society of America Special Papers 372, 265-290. 13 141189 Hill, K.C., Raza, A., 1999. Arc‐continent collision in Papua Guinea: Constraints from 151190 fission track thermochronology. Tectonics 18, 950-966. 161191 Hoffmann, K., Exon, N., Quilty, P., Findlay, C., 2008. Mellish Rise and adjacent deep water 17 181192 plateaus off northeast Australia: new evidence for continental basement from Cenozoic 191193 micropalaeontology and sedimentary geology, PESA Eastern Australian Basins 201194 Symposium III, pp. 317-323. 211195 Holm, R.J., Spandler, C., Richards, S.W., 2013. Melanesian arc far-field response to 22 231196 collision of the Ontong Java Plateau: geochronology and petrogenesis of the Simuku 241197 Igneous Complex, New Britain, Papua New Guinea. Tectonophysics 603, 189-212. 251198 Holm, R.J., Spandler, C., Richards, S.W., 2015. Continental collision, orogenesis and arc 261199 magmatism of the Miocene Maramuni arc, Papua New Guinea. Gondwana Research 28, 27 281200 1117-1136. 291201 Hughes, G.W., Turner, C.C., 1977. Upraised Pacific Ocean floor, southern Malaita, 301202 Solomon Islands. Geological Society of America Bulletin 88, 412-424. 311203 Johnson, R., 1979. Geotectonics and volcanism in Papua New Guinea: a review of the late 321204 Cainozoic. BMR J. Aust. Geol. Geophys 4, 181-207. 33 341205 Joshima, M., Okuda, Y., Murakami, F., Kishimoto, K., Honza, E., 1986. Age of the Solomon 351206 Sea Basin from magnetic lineations. Geo-Marine Letters 6, 229-234. 361207 Knesel, K.M., Cohen, B.E., Vasconcelos, P.M., Thiede, D.S., 2008. Rapid change in drift of 371208 the Australian plate records collision with Ontong Java plateau. Nature 454, 754-757. 38 391209 Kroenke, L., 1984. Solomon Islands: San Cristobal to Bougainville and Buka, In: Kroenke, L. (Ed.), 401210 Cenozoic tectonic development of the southwest Pacific. U.N. ESCAP, pp. 47-61. 411211 Lagabrielle, Y., Chauvet, A., Ulrich, M., Guillot, S., 2013. Passive obduction and gravity- 421212 driven emplacement of large ophiolitic sheets: The New Caledonia ophiolite (SW 43 441213 Pacific) as a case study? Bulletin de la Société Géologique de France 184, 545-556. 451214 Landmesser, C., Andrews, J., Packham, G., 1973. Aspects of the geology of the eastern 461215 Coral Sea and the western New Hebrides Basin. Initial Reports of the Deep Sea Drilling 471216 Project 30, 647-662. 48 491217 Lapouille, A., 1982. …TUDE DE~ BASSINSMARGINAUXFOSSILES DUSUD- 501218 OUESTPACIFIQUE: BASSINNORD-D'ENTRECASTEAUX, BASSIN NORD-LOYAUT…, 511219 BASSIN SUD-FIDJIEN. 521220 Larue, B., Daniel, J., Jouannic, C., Recy, J., 1977. The South Rennell Trough: Evidence for a 53 541221 fossil spreading zone, p. 51ñ61. 551222 Leng, W., Gurnis, M., 2015. Subduction initiation at relic arcs. Geophysical Research 561223 Letters 42, 7014-7021. 571224 Lindley, D., 1988. Early Cainozoic stratigraphy and structure of the Gazelle Peninsula, 58 591225 east New Britain: An example of extensional tectonics in the New Britain arc-trench 601226 complex. Journal of the Geological Society of Australia 35, 231-244. 61 62 63 64 49 65 1227 Lus, W.Y., McDougall, I., Davies, H.L., 2004. Age of the metamorphic sole of the Papuan 11228 Ultramafic Belt ophiolite, Papua New Guinea. Tectonophysics 392, 85-101. 21229 Luyendyk, B.P., Bryan, W., Jezek, P., 1974. Shallow structure of the New Hebrides island 3 41230 arc. Geological Society of America Bulletin 85, 1287-1300. 51231 Maillet, P., Monzier, M., Selo, M., Storzer, D., 1983. The d'Entrecasteaux zone (southwest 61232 Pacific). A petrological and geochronological reappraisal. Marine Geology 53, 179-197. 71233 Matthews, K.J., Williams, S.E., Whittaker, J.M., Müller, R.D., Seton, M., Clarke, G.L., 2015. 8 91234 Geologic and kinematic constraints on Late Cretaceous to mid Eocene plate boundaries 101235 in the southwest Pacific. Earth-Science Reviews 140, 72-107. 111236 Meffre, S., Falloon, T.J., Crawford, T.J., Hoernle, K., Hauff, F., Duncan, R.A., Bloomer, S.H., 121237 Wright, D.J., 2012. Basalts erupted along the Tongan fore arc during subduction 13 141238 initiation: Evidence from geochronology of dredged rocks from the Tonga fore arc and 151239 trench. Geochemistry, Geophysics, Geosystems 13. 161240 Mendel, V., Munschy, M., Sauter, D., 2005. MODMAG, a MATLAB program to model 171241 marine magnetic anomalies. Computers & geosciences 31, 589-597. 18 191242 Metcalfe, I., 2009. Late Palaeozoic and Mesozoic tectonic and palaeogeographical 201243 evolution of SE Asia. Geological Society, London, Special Publications 315, 7-23. 211244 Michael, P., Langmuir, C., Dick, H., Snow, J., Goldstein, S., Graham, D., Lehnert, K., Kurras, 221245 G., Mühe, R., Edmonds, H., 2003. Magmatic and amagmatic seafloor spreading at the 23 241246 slowest mid-ocean ridge: Gakkel Ridge. Arctic Ocean, Nature 423, 956-961. 251247 Morimer, N., Gans, P., Meffre, S., Martin, C., Seton, M., Williams, S., Turnbull, R., Quilty, P., 261248 Micklethwaite, S., Sutherland, R., Bache, F., Timm, C., Collot, J., Maurizot, P., Rollet, N., in 271249 review. Regional volcanism of northern Zealandia: post-Gondwana breakup magmatism 281250 on an extended, submerged continent, In: Sensarma, S., Storey, B. (Eds.), Geological 29 301251 Society of London Special Publications. 311252 Mortimer, N., Gans, P., Hauff, F., Barker, D., 2012. Paleocene MORB and OIB from the 321253 Resolution Ridge, Tasman Sea. Australian Journal of Earth Sciences 59, 953-964. 331254 Mortimer, N., Gans, P.B., Palin, J.M., Herzer, R.H., Pelletier, B., Monzier, M., 2014. Eocene 34 351255 and Oligocene basins and ridges of the Coral Sea‐New Caledonia region: Tectonic link 361256 between Melanesia, Fiji, and Zealandia. Tectonics 33, 1386-1407. 371257 Mortimer, N., Herzer, R., Gans, P., Laporte-Magoni, C., Calvert, A., Bosch, D., 2007. 38 391258 Oligocene–Miocene tectonic evolution of the South Fiji Basin and Northland Plateau, SW 401259 Pacific Ocean: evidence from petrology and dating of dredged rocks. Marine Geology 411260 237, 1-24. 421261 Mühe, R., Bohrmann, H., Garbe-Schönberg, D., Kassens, H., 1997. E-MORB glasses from 43 441262 the Gakkel Ridge (Arctic Ocean) at 87 N: evidence for the Earth's most northerly 451263 volcanic activity. Earth and Planetary Science Letters 152, 1-9. 461264 Mu lle , R., Seton, M., Zahirovic, S., Williams, S., Matthews, K., Wright, N., Shephard, G., 471265 Maloney, K., Barnett-Moore, N., Hosseinpour, M., Bower, D., Cannon, J., 2016. Ocean 481266 basin evolution and global-scale plate reorganization events since Pangea breakup. 49 501267 Annual Review of Earth and Planetary Sciences 44. 511268 Musgrave, R.J., 2013. Evidence for Late Eocene emplacement of the Malaita Terrane, 521269 Solomon Islands: Implications for an even larger Ontong Java Nui oceanic plateau. 531270 Journal of Geophysical Research: Solid Earth 118, 2670-2686. 54 551271 Musgrave, R.J., Firth, J.V., 1999. Magnitude and timing of New Hebrides Arc rotation: 561272 paleomagnetic evidence from Nendo, Solomon Islands. Journal of Geophysical Research: 571273 Solid Earth (1978–2012) 104, 2841-2853. 581274 Nauret, F., Snow, J., Hellebrand, E., Weis, D., 2011. Geochemical composition of K-rich 59 601275 lavas from the Lena Trough (Arctic Ocean). Journal of Petrology, egr024. 61 62 63 64 50 65 1276 Pearce, J.A., 2008. Geochemical fingerprinting of oceanic basalts with applications to 11277 ophiolite classification and the search for Archean oceanic crust. Lithos 100, 14-48. 21278 Peate, D.W., Pearce, J.A., Hawkesworth, C.J., Colley, H., Edwards, C.M., Hirose, K., 1997. 3 41279 Geochemical variations in Vanuatu arc lavas: the role of subducted material and a 51280 variable mantle wedge composition. Journal of Petrology 38, 1331-1358. 61281 Petterson, M., Babbs, T., Neal, C., Mahoney, J., Saunders, A., Duncan, R., Tolia, D., Magu, R., 71282 Qopoto, C., Mahoa, H., 1999. Geological–tectonic framework of Solomon Islands, SW 8 91283 Pacific: crustal accretion and growth within an intra-oceanic setting. Tectonophysics 101284 301, 35-60. 111285 Récy, J., Dubois, J., Daniel, J., Dupont, J., Launay, J., 1977. Fossil subduction zones: 121286 examples in the South West Pacific. 13 141287 Sandwell, D.T., Müller, R.D., Smith, W.H., Garcia, E., Francis, R., 2014. New global marine 151288 gravity model from CryoSat-2 and Jason-1 reveals buried tectonic structure. science 346, 161289 65-67. 171290 Schellart, W., Lister, G., Toy, V., 2006. A Late Cretaceous and Cenozoic reconstruction of 18 191291 the Southwest Pacific region: tectonics controlled by subduction and slab rollback 201292 processes. Earth-Science Reviews 76, 191-233. 211293 Sdrolias, M., R.D., M., Gaina, C., 2003. Tectonic Evolution of the SW Pacific Using 221294 Constraints from Back-arc Basins, In: R.R., H., R.D., M. (Eds.), Evolution and Dynamics of 23 241295 the Australian Plate. Geological Society of Australia Special Publication 22 and 251296 Geological Society of America Special Paper. 261297 Sdrolias, M., Roest, W.R., M√ºller, R.D., 2004. An expression of Philippine Sea plate 271298 rotation: the Parece Vela and Shikoku basins. Tectonophysics 394, 69-86. 28 291299 Seton, M., Whittaker, J.M., Wessel, P., Müller, R.D., DeMets, C., Merkouriev, S., Cande, S., 301300 Gaina, C., Eagles, G., Granot, R., 2014. Community infrastructure and repository for 311301 marine magnetic identifications. Geochemistry, Geophysics, Geosystems 15, 1629-1641. 321302 Shervais, J.W., 1982. Ti-V plots and the petrogenesis of modern and ophiolitic lavas. 33 341303 Earth and planetary science letters 59, 101-118. 351304 Smith, I.E., 2013. The chemical characterization and tectonic significance of ophiolite 361305 terrains in southeastern Papua New Guinea. Tectonics 32, 159-170. 371306 Smith, I.E., Davies, H., 1976. Geology of the southeast Papuan mainland. Australian 38 391307 Government Pub. Service. 401308 Stampfli, G.M., Borel, G.D., 2002. A plate tectonic model for the Paleozoic and Mesozoic 411309 constrained by dynamic plate boundaries and restored synthetic oceanic isochrons. 421310 Earth and Planetary Science Letters 196, 17-33. 43 441311 Stewart, W., Sandy, M., 1988. Geology of New Ireland and Djaul Islands, northeastern 451312 Papua New Guinea. 461313 Tapster, S., Roberts, N., Petterson, M., Saunders, A., Naden, J., 2014. From continent to 471314 intra-oceanic arc: Zircon xenocrysts record the crustal evolution of the Solomon island 481315 arc. Geology 42, 1087-1090. 49 501316 Taylor, B., Goodliffe, A., Martinez, F., Hey, R., 1995. Continental rifting and initial sea- 511317 floor spreading in the Woodlark basin. Nature 374, 534-537. 521318 Taylor, B., Goodliffe, A.M., Martinez, F., 1999. How continents break up: insights from 531319 Papua New Guinea. Journal of Geophysical Research: Solid Earth (1978–2012) 104, 54 551320 7497-7512. 561321 Taylor, B., Zellmer, K., Martinez, F., Goodliffe, A., 1996. Sea-floor spreading in the Lau 571322 back-arc basin. Earth and Planetary Science Letters 144, 35-40. 58 59 60 61 62 63 64 51 65 1323 Taylor, G.K., Gascoyne, J., Colley, H., 2000. Rapid rotation of Fiji: paleomagnetic evidence 11324 and tectonic implications. Journal of Geophysical Research: Solid Earth (1978–2012) 21325 105, 5771-5781. 3 41326 Taylor, G.R., 1973. Preliminary observations on the structural history of Rennell Island, 51327 south Solomon Sea. Geological Society of America Bulletin 84, 2795-2806. 61328 Taylor, L., Falvey, D., 1977. Queensland Plateau and Coral Sea Basin: stratigraphy, 71329 structure and tectonics. The APEA Journal 17, 13-29. 8 91330 Tejada, M., Mahoney, J., Duncan, R., Hawkins, M., 1996. Age and geochemistry of 101331 basement and alkalic rocks of Malaita and Santa Isabel, Solomon Islands, southern 111332 margin of Ontong Java Plateau. Journal of Petrology 37, 361-394. 121333 Todd, E., Gill, J.B., Pearce, J.A., 2012. A variably enriched mantle wedge and contrasting 13 141334 melt types during arc stages following subduction initiation in Fiji and Tonga, southwest 151335 Pacific. Earth and Planetary Science Letters 335, 180-194. 161336 Tregoning, P., Tan, F., Gilliland, J., McQueen, H., Lambeck, K., 1998. Present-day crustal 171337 motion in the Solomon Islands from GPS observations. Geophysical Research Letters 25, 18 191338 3627-3630. 201339 Ulrich, M., Picard, C., Guillot, S., Chauvel, C., Cluzel, D., Meffre, S., 2010. Multiple melting 211340 stages and refertilization as indicators for ridge to subduction formation: The New 221341 Caledonia ophiolite. Lithos 115, 223-236. 23 241342 Wallace, L.M., Ellis, S., Little, T., Tregoning, P., Palmer, N., Rosa, R., Stanaway, R., Oa, J., 251343 Nidkombu, E., Kwazi, J., 2014. Continental breakup and UHP rock exhumation in action: 261344 GPS results from the Woodlark Rift, Papua New Guinea. Geochemistry, Geophysics, 271345 Geosystems 15, 4267-4290. 28 ‐ 291346 Webb, L.E., Baldwin, S.L., Fitzgerald, P.G., 2014. The Early Middle Miocene subduction 301347 complex of the Louisiade Archipelago, southern margin of the Woodlark Rift. 311348 Geochemistry, Geophysics, Geosystems 15, 4024-4046. 321349 Weissel, J.K., Watts, A.B., 1979. Tectonic evolution of the Coral Sea Basin. Journal of 33 341350 Geophysical Research 84, 4572-4582. 351351 Wells, R.E., 1989. Origin of the oceanic basalt basement of the Solomon Islands arc and 361352 its relationship to the Ontong Java Plateau--insights from Cenozoic plate motion models. 371353 Tectonophysics 165, 219-235. 38 391354 Whattam, S.A., Malpas, J., Ali, J.R., Smith, I.E., 2008. New SW Pacific tectonic model: 401355 Cyclical intraoceanic magmatic arc construction and near‐coeval emplacement along 411356 the Australia‐Pacific margin in the Cenozoic. Geochemistry, Geophysics, Geosystems 9. 42 431357 Willcox, J., Symonds, P., Hinz, K., Bennett, D., 1980. Lord Howe Rise, Tasman Sea- 441358 Preliminary geophysical results and petroleum prospects. BMR J. Aust. Geol. and 451359 Geophys. 5, 225-236. 461360 Williams, S., Gans, P., Mortimer, N., Meffre, S., Seton, M., 2014. Age-progressive 47 481361 volcanism in the Tasman and Coral seas, AGU Fall Meeting Abstracts, p. 4692. 491362 Worthing, M., Crawford, A., 1996. The igneous geochemistry and tectonic setting of 501363 metabasites from the Emo Metamorphics, Papua New Guinea; a record of the evolution 511364 and destruction of a backarc basin. Mineralogy and Petrology 58, 79-100. 521365 Yan, C.Y., Kroenke, L.W., 1993. A plate tectonic reconstruction of the Southwest Pacific, 0 53 541366 100 Ma, In: Berger, W.H., Kroenke, L.W., Mayer, L.A. (Eds.), Proceedings of the Ocean 551367 Drilling Program, Scientific Results, College Station, TX, p. 697. 561368 Zahirovic, S., Matthews, K.J., Flament, N., Muller, R., KC, H., Seton, M., Gurnis, M., in 571369 review. Tectonic evolution and deep mantle structure of eastern Tethyan domain since 58 591370 the latest Jurassic. Earth Science Reviews. 60 61 62 63 64 52 65 1371 Zahirovic, S., Seton, M., Müller, R., 2014. The Cretaceous and Cenozoic tectonic evolution 11372 of Southeast Asia. Solid Earth 5, 227-273. 21373 3 41374 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 53 65 Figure 1

South Solomons Trench Woodlark Basin −10˚

Rennell Island Ridge

Poklington Trough San Cristobal Trench Rennell Basin RI

New Hebrides Trench Santa Cruz Louisiade DR3 −12˚ Plateau DR2 Basin DR1 DR5 West Torres Plateau WLR DR4 −14˚ DR6 CSB Louisiade Trough ELR WSB DR7 South Rennell Trough Mellish Rise d’Entrecasteaux DEZ −16˚ Basin DR13 SE DR14 NLB Asia DR15 DR11 DR8-10 SW −18˚ Kenn Pacific Plateau DR16 BP LHR NCB NCR 154˚E 156˚E 158˚E 160˚E 162˚E 164˚E 166˚E 168˚E

m

−5000 −4000 −3000 −2000 −1000 0 Bathymetry

Seton et al. Figure 1 Figure 2

−10˚

T1 T1 Bellona T2 −11˚ T2 Island

Rennell GO301 GO301 Island Rennell Basin Rennell Basin

LP LP T2 −12˚ T2 DR3 DR3 DR2 DR2

T1 SCB T1 SCB Indispensable Reefs −13˚ LT Mellish LT Mellish Rise Rise WLR SRT WLR SRT

159˚E 160˚E 161˚E 159˚E 160˚E 161˚E

KEY T1 ECOSAT dredge −6000 −4000 −2000 0 T2 ORSTOM dredge Bathymetry [m] Swath & gravity lineations Tectonic block outline Figure 3

SCB −12˚ Rennell Basin DR3 Rennell Basin DR3 DR2 DR2 SCB SCB

−13˚ Mellish Mellish Rise GO309 Rise GO309 GO308 GO308 GO307 DR5 GO307 DR5 −14˚ WLR SRT DR6 DR6 ELR West Lapérouse Rise

−15˚

East Lapérouse Rise South Rennell Trough DR7 DR7

−16˚ d’Entrecasteaux d’Entrecasteaux DR12 Basin DR12 Basin DR13 DR13

−17˚ DR14 DR14 BP BP

159˚E 160˚E 161˚E 162˚E 159˚E 160˚E 161˚E 162˚E

KEY Tectonic block outline ECOSAT dredge −6000 −4000 −2000 0 Swath & gravity lineations ORSTOM dredge Bathymetry [m] Figure 4

SI SI SI −11˚

−12˚ DR3 GO310 DR3 GO310

−13˚ WLR WLR West Torres West Torres SRT ELR SRT ELR GO308 Plateau GO308 Plateau GO307 DR4 GO307 DR4 −14˚

161˚E 162˚E 163˚E 164˚E 165˚E 161˚E 162˚E 163˚E 164˚E 165˚E

KEY Ridge flank Tectonic block outline ECOSAT dredge −6000 −4000 −2000 0 Swath & gravity lineations ORSTOM dredge Bathymetry [m] Figure 5

Santa Cruz Basin Santa Cruz Basin −13˚ Mellish Mellish SRT West Torres SRT West Torres Rise GO309 GO308 DR5 Rise GO309 GO308 DR5 DR4 Plateau DR4 Plateau −14˚ GO307 GO307 WLR ZONE 2 WLR ZONE 2 DR6 ZONE 3 DR6 ZONE 3 −15˚ ELR WSB ELR WSB GO316 GO315 GO316 GO315 DR7 GO323 GO317 GO314 DR7 GO323 GO317 GO314 −16˚ GO322 GO322 Basin Basin DR12 DR12 D’Entrecasteaux D’Entrecasteaux ZONE 1 DR13 ZONE 1 DR13 GO324 GO324 −17˚ Zone Zone DR14 D’EntrecasteauxNorth DR14 D’Entrecasteaux North Loyalty Basin Loyalty Basin −18˚ Lord Howe Lord Howe Rise LR Rise LR −19˚

160˚E 162˚E 164˚E 166˚E 160˚E 162˚E 164˚E 166˚E

KEY Tectonic block outline T1 ECOSAT dredge −6000 −4000 −2000 0 Swath & gravity lineations ORSTOM dredge Bathymetry [m]

Seton et al. Fig. 5 Figure6

Mellish Rennell Island West Lapérouse S. Rennell Rise Ridge Torres Rise Trough 0 Quat. Plio. N21 N17 L 10 N16 M N10 N8 E N6 20 Miocene N5 N4

L P22 v P21 v P19 O2 v 30 ? P82189 v Olig. E P18 O1 L E15 v P15 E14 P82206 P14 E13 v

v P82217 40 E11 v P12 v v M E10 v v

P11 E9 P82193

P10 E8 P82182 P9 E7 P82183 50 Eocene E P7 E5 1 DR2 DR3 DR4 DR5 6 13 DR14 Ma P6 E4 A C E H C F HAiAiv Aiv EBi Aiii v A Ai Ai Eiii G

Ooze Limestone Mudstone Sandstone v v Basalt

Fig. 6 Seton et al. Figure 7

DR2H detrital, RIR DR4B basalt, WTP DR6A basalt, ELR (centre) DR14iii basalt, SRT (east)

Fig. 7. Seton et al. Figure8

60 0.01 60 0.01 A. Rennell Island Ridge DR3Ai B. Rennell Island Ridge DR3Aiv P82182 plagioclase P82183 plagioclase 50 42.0±2.0 Ma 43.2±1.5 Ma 0.005 50 0.005 40

0.002 40 30 TFA = 45.8±0.8 Ma TFA = 45.5±0.9 Ma PseudoPA = 42.3±1.2 Ma WMPA = 43.2±1.0 Ma 20 0.001 30 0.002

40 0.05 0.005 C. West Torres Plateau DR4Bi D. East Lapérouse Rise DR6A P82189 plagioclase 60 P82193 plagioclase 26.2±0.8 Ma 39.3±1.2 Ma 0.02 0.004 30 50 K/Ca 0.01 0.003 40 20 0.005 apparent age (Ma) TFA = 25.2±0.8 Ma TFA = 39.9±0.6 Ma 30 WMPA = 26.2±0.8 Ma WMPA = 39.3±0.6 Ma 10 0.002 0.002

40 45 E. South Rennell Trough DR14Eiii F. South Rennell Trough DR14G 0.01 P82206 groundmass P82217 plagioclase 35 30.0±2.0 Ma 1 34.7±0.5 Ma 40 0.005 30 0.1 35 25 0.01 0.002 30 20 TFA = 30.8±0.3 Ma TFA 34.8±0.4 Ma PseudoPA = 30.1±0.5 Ma WMPA = 34.7±0.4 Ma

15 0.001 25 0.001 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 fraction 39 Ar released

Fig. 8. Seton et al. Figure9

600 A. B.

Subduction- related MORB & OIB BABB 1 S 400 S S N Subduction- S F related T TT N Th/Yb T ppm V N E-MORB C TF TT T 0.1 F 200 Intraplate- N F OIB NN C

N-MORB

0 0.01 0 4000 8000 12000 16000 0.1 1 10 ppm Ti Nb/Yb

Rennell spreading system Other S Solomon Sea Basin 12°S Indispensable (DR3) 12°S (ORSTOM) West Torres (DR4) N North Loyalty Basin F South Fiji Basin 14°S (ORSTOM) Rennell Is. Ridge 14°S E Lapérouse (DR6) C Coral Sea Basin 16°S Ski Jump (DR14) T Tasman Sea Basin

Fig. 9. Seton et al. Figure 10

−10˚

L682SP EW9511 GEO100 87001411 −11˚ 82031604

CH100L08

−12˚ DR3 GO310 42.8 ± 1.2 Ma ECOSAT03

ECOSAT02 −13˚ ECOSAT01 KH68D GO309 GO308 DR4 26.2 ± 0.8 Ma −14˚

160˚E 161˚E 162˚E 163˚E 164˚E 165˚E 166˚E

KEY Magnetic picks 18o (40.1 Ma) Tectonic block outline 13o (33.5 Ma) 19o (41.5 Ma) ECOSAT dredge 16o (36.3 Ma) 20o (43.8 Ma) ORSTOM dredge 17o (37.5 Ma) 21o (47.9 Ma)

Seton et al. Fig. 10 Figure11

Spreading Axis

L682SP

EW9511

87001411

ECOSAT03

ECOSAT02

ECOSAT01

GEO100

82031604

CH100L08

KH68D

200

0

−200 20n 19n 18n 17n 16n 15n 13n 18n17n16n15n13n 19n 20n Magnetic Anomaly (nT)

45 mm/yr 30 mm/yr 30 mm/yr 45 mm/yr

−200 −100 0 100 200 Distance [km]

Seton et al. Fig. 11 Figure 12

−13˚

GO309 DR5 33 GO308 DR4 −14˚ GO307 ZONE 2 26.2 ± 0.8 Ma 34 32 DR6 32 39.1 ± 1.2 Ma ZONE 3

−15˚ 33 30 31 GO316 GO315 DR7 GO314 ZONE 1 GO317 −16˚ 31 GO323 30 GO322 DR12 DR13 GO324 −17˚ DR14 30.0 ± 2.0 Ma 34.7 ± 0.5 Ma

−18˚

−19˚

159˚E 160˚E 161˚E 162˚E 163˚E 164˚E 165˚E 166˚E

Seton et al. Fig. 12 Figure 13

52 Ma 10˚S 48 Ma 10˚S 35 MaMa35 RIR RIR RIR LP LP LP WTP WTP WTP

15˚S 15˚S MR MR MR

KP KP KP

155˚E 160˚E 165˚E 155˚E 160˚E 165˚E 155˚E 160˚E 165˚E

25 Ma 10˚S 0 Ma KEY LP RIR LP RIR Australia Pacific WLR WLR WTP WTP ELR Coral Sea Solomon Sea ELR 15˚S MR MR Lord Howe Rise South Rennell

Peripheral back-arc basin plates KP KP

Fig. 13 Seton et. al. Figure 14

A B C PSP PAC PAC PAC PMP PMP PMP MB EB ESZ DEZ MA VSZ SLSZ SLB NHA AUS AUS AUS CSB LHR DEB

D E PSP F PSP CS PAC PAC OJP PAC MSZ PMP SSP SSZ MB NLB SCSR TSZ SFP

AUS AUS AUS

G H CS I CS CS PAC PAC PAC SSP SSP

NHSZ WB SOSZ NFB AUS AUS AUS

0 20 40 60 80 100 120 140 160 180 200 Age of Oceanic Crust [m.yrs.] Table1

Table 1. Summary of dredge localities, rock types and analyses

Dredge Location Lat (°S) Long (°E) Depth (m) Approx. Rocks (excl. Mn nodules Ar-Ar & Micropal- Spinels wt (kg) geochem entology DR1 Louisiade Plateau, SE side 13.2012 157.9813 3000-3500 30 Pale brown soft mudstone Y DR2 Rennell Ridge, SW side 12.2042 159.7074 2800-3500 150 Cream-grey calc-mudstones, some sandy Y Y DR3 Rennell Ridge, NE side 12.0251 160.8317 2200-2800 50 Basalts, some jointed and pillowed, Y Y with rare lst-cemented volcanic breccia DR4 West Torres Plateau, W tip 13.8801 163.3189 2600-2900 50 Basalt, some with glassy rinds Y Y Y DR5 East Lapérouse Rise, NE end 13.6510 162.2192 1200-1600 300 Cream-coloured varitextured limestones Y DR6 East Lapérouse Rise, mid part 14.4534 161.1168 2600-3000 1 Basalt Y Y DR13 South Rennell Trough, W scarp 16.6396 158.8356 1530-2000 3 Pale cream-coloured limestone Y DR14 South Rennell Trough, E scarp 17.0755 159.1991 2600-3200 400 Basalt lavas and associated volcanic Y Y breccia, sandstone and mudstone Table2

Table 2. Summary of Ar-Ar dating

Dredge Site GNS # UCSB # Material wt. (mg) TFA (Ma) WMPA (Ma) ISOA (Ma) % gas K/Ca J Pref. age DR3Ai Rennell Ridge, NE side P82182 SB66-7,8,10,11 Plagioclase 25.9 45.8±0.8 42.3±1.2 42.5±15 32 0.001-0.003 0.0041962 42.0 ± 2.0 DR3Aiv Rennell Ridge, N side P82183 SB66-13,14 Plagioclase 14.1 45.5±0.8 43.2±1.0 42.8±2.2 57 0.004-0.007 0.0041830 43.2 ± 1.5 DR4Bi West Torres Plateau P82189 SB66-15 Plagioclase 21.1 25.2±0.8 26.2±0.8 25.5±5.0 69 0.005-0.030 0.0041784 26.2 ± 0.8 DR6A East Lapérouse Rise P82193 SB66-43 Plagioclase 25.6 39.9±0.6 39.3±0.6 39.1±1.2 92 0.002-0.005 0.0041987 39.3 ± 1.2 DR14Eiii South Rennell Trough P82206 SB66-45 Groundmass 26.7 30.8±0.4 30.1±0.5 30.8±1.6 28 0.003-0.970 0.0041991 30.0 ± 2.0 DR14G South Rennell Trough P82217 SB66-51 Plagioclase 23.9 34.8±0.4 34.7±0.4 34.3±1.2 79 0.005-0.006 0.0041887 34.7 ± 0.5

GNS#=Institute of Geological and Nuclear Sciences Petrology Collection number, UCSB#=University of California Santa Barbara irradiation number, TFA=total fusion age, WMPA=weighted mean plateau age, ISOA=inverse isochron age, J=dimensionless irradiation parameter, Pref. age= preferred interpreted age. e-component: supplementary material Click here to download e-component: Seton++_Supplementary_Material_Summary_revision_unmarked.docx Supplementary Figure 1 Click here to download e-component: SupplFig1_Seton++.pdf Supplementary Figure 2 Click here to download e-component: SupplFig2_Seton++.pdf Supplementary Figure 3 Click here to download e-component: SupplFig3_Seton++.pdf e-component: supplementary Figure 4 Click here to download e-component: SupplFig4_Seton++_NEW.pdf e-component: supplementary Figure 5 Click here to download e-component: SupplFig5_Seton++_NEW.pdf Supplementary File A Click here to download Raw research data (under CC BY license; see above): SupplFileA_Seton++.pdf Supplementary File B Click here to download Raw research data (under CC BY license; see above): SupplFileB_Seton++.pdf Supplementary File C Click here to download Raw research data (under CC BY license; see above): SuppFileC_Seton++.pdf Supplementary File D Click here to download Raw research data (under CC BY license; see above): SupplFileD_Seton++.pdf e-component: animation - Plate_Reconstruction_Animation_Seton++.mov

This file could not be included in the PDF because the file type is not supported. Data directory Click here to download Raw research data (under CC BY license; see above): Seton++_Data.zip