Breccia cored columnar rosettes in a rubbly pahoehoe lava flow, elephanta island, deccan traps, and a model for their origin

Breccia cored columnar rosettes in a rubbly pahoehoe lava flow, Elephanta Island, Deccan Traps, and a model for their origin Accepted Manuscript Breccia cored columnar rosettes in a rubbly pahoehoe la[.]

Accepted Manuscript Breccia-cored columnar rosettes in a rubbly pahoehoe lava flow, Elephanta Island, Deccan Traps, and a model for their origin Hetu Sheth, Ishita Pal, Vanit Patel, Hrishikesh Samant, Joseph D’Souza PII: S1674-9871(17)30003-8 DOI: 10.1016/j.gsf.2016.12.004 Reference: GSF 525 To appear in: Geoscience Frontiers Received Date: October 2016 Revised Date: December 2016 Accepted Date: 18 December 2016 Please cite this article as: Sheth, H., Pal, I., Patel, V., Samant, H., D’Souza, J., Breccia-cored columnar rosettes in a rubbly pahoehoe lava flow, Elephanta Island, Deccan Traps, and a model for their origin, Geoscience Frontiers (2017), doi: 10.1016/j.gsf.2016.12.004 This is a PDF file of an unedited manuscript that has been accepted for publication As a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT ACCEPTED MANUSCRIPT Hetu Sheth a,*, Ishita Pal a,b, Vanit Patela, Hrishikesh Samantc, Joseph D’Souzaa RI PT Breccia-cored columnar rosettes in a rubbly pahoehoe lava flow, Elephanta Island, Deccan Traps, and a model for their origin a India b University of California San Diego, La Jolla, CA 92093-0225, USA c SC Present address: Institute of Geophysics and Planetary Physics, Scripps Institution of Oceanography, M AN U 10 Department of Earth Sciences, Indian Institute of Technology Bombay (IITB), Powai, Mumbai 400076, Department of Geology, St Xavier’s College, Mumbai 400001, India 11 12 * Corresponding author email: hcsheth@iitb.ac.in Tel: +91-22-25767264; Fax: +91-22- 13 25723480 14 ABSTRACT 16 Rubbly pahoehoe lava flows are abundant in many continental flood basalts including the 17 Deccan Traps However, structures with radial joint columns surrounding cores of flow-top 18 breccia (FTB), reported from some Deccan rubbly pahoehoe flows, are yet unknown from other 19 basaltic provinces A previous study of these Deccan “breccia-cored columnar rosettes” ruled out 20 explanations such as volcanic vents and lava tubes, and showed that the radial joint columns had 21 grown outwards from cold FTB inclusions incorporated into the hot molten interiors How the 22 highly vesicular (thus low-density) FTB blocks might have sunk into the flow interiors has 23 remained a puzzle Here we describe a new example of a Deccan rubbly pahoehoe flow with 24 FTB-cored rosettes, from Elephanta Island in the Mumbai harbour Noting that (i) thick rubbly 25 pahoehoe flows probably form by rapid inflation (involving many lava injections into a largely 26 molten advancing flow), and (ii) such flows are transitional to ‘a’ā flows (which continuously 27 shed their top clinker in front of them as they advance), we propose a model for the FTB-cored 28 rosettes We suggest that the Deccan flows under study were shedding some of their FTB in front AC C EP TE D 15 ACCEPTED MANUSCRIPT of them as they advanced and, with high-eruption rate lava injection and inflation, frontal 30 breakouts would incorporate this FTB rubble, with thickening of the flow carrying the rubble 31 into the flow interior This implies that, far from sinking into the molten interior, the FTB blocks 32 may have been rising, until lava supply and inflation stopped, the flow began solidifying, and 33 joint columns developed outward from each cold FTB inclusion as already inferred, forming the 34 FTB-cored rosettes Those rubbly pahoehoe flows which began recycling most of their FTB 35 became the ‘a’ā flows of the Deccan RI PT 29 36 Keywords: Rubbly pahoehoe; columnar jointing; flow-top breccia; volcanism; flood basalt; 38 Deccan Traps SC 37 M AN U 39 40 41 Introduction Many flood basalt provinces of the world contain abundant and voluminous lava flows of 43 rubbly pahoehoe, i.e., flows with extensively brecciated upper crusts (Keszthelyi and 44 Thordarson, 2000, 2001) Rubbly pahoehoe flows have been reported from the mid-Cambrian 45 Kalkarindji flood basalts (Marshall et al., 2016), the Triassic/Jurassic Boundary CAMP flood 46 basalts (El Hachimi et al., 2011), the Early Cretaceous Kerguelen oceanic plateau (Keszthelyi, 47 2002), the Late Cretaceous–Palaeocene Deccan Traps (Duraiswami et al., 2008), the Miocene 48 Columbia River flood basalts (e.g., Swanson and Wright, 1981; Self et al., 1997; Bondre and 49 Hart, 2008; Reidel et al., 2013), the Pliocene-Pleistocene South Caucasus flood basalts (Sheth et 50 al., 2015), < 4500 yr old flood basalt fields in Saudi Arabia (Murcia et al., 2014), and the 1783– 51 1784 Laki eruption in Iceland (Guilbaud et al., 2005) Rubbly pahoehoe lavas are also 52 recognized on the planet Mars (Keszthelyi et al., 2006) Given the significant environmental 53 impact of the historical Laki eruption (Guilbaud et al., 2005), and noting that many prehistoric 54 flood basalt events closely correlate with biological mass extinctions (e.g., Rampino and 55 Stothers, 1988; Wignall, 2001), topics such as the physical emplacement of flood basalt lava 56 flows, their emplacement duration, and volatile release are of major interest (e.g., Self et al., 57 1997, 2014; Parisio et al., 2016) AC C EP TE D 42 58 The Deccan Traps currently occupy ~500,000 km2 in western and central India, and in 59 the Western Ghats escarpment (Fig 1a) they attain a stratigraphic thickness of ~3.4 km over a ACCEPTED MANUSCRIPT ~500 km distance (e.g., Beane et al., 1986) Walker (1971) described many individual lava flows 61 of the Deccan Traps as “compound”, made up of numerous constituent flow units or lobes He 62 described other Deccan flows, which are single, thick and areally extensive, typically columnar- 63 jointed flow units, as “simple” flows Compound flows (dominantly pahoehoe, with minor ‘a’ā) 64 are abundant in the lower part of the Western Ghats stratigraphic sequence (Walker, 1971; 65 Keszthelyi et al., 1999; Bondre et al., 2004; Sheth, 2006; Brown et al., 2011) Simple flows 66 dominate the upper parts of the sequence and many of them are pahoehoe and rubbly pahoehoe, 67 with some ‘a’ā (Bondre et al., 2004; Duraiswami et al., 2008, 2014) It is generally assumed that 68 during the emplacement and cooling of thick rubbly pahoehoe flows, their crusts and interiors 69 not undergo large-scale remixing However, Sheth et al (2011) presented field, petrographic and 70 geochemical evidence from some rubbly pahoehoe flows of the Deccan Traps showing that their 71 broken flow-top breccia (FTB) crusts became incorporated into the flows’ molten interiors 72 where, strongly affecting the internal temperature distribution, they led to radial cooling joint 73 columns growing outwards from them To our knowledge, these Deccan “FTB-cored rosettes” 74 (Fig 2) have not been described from any other flood basalt province, though many examples 75 are known worldwide of columnar rosettes without obvious cores, such as the “war bonnet” 76 structures of the Columbia River flood basalt province (Waters, 1960; Spry, 1962; Scheidegger, 77 1978; De, 1996) TE D M AN U SC RI PT 60 A persisting problem, however, is how blocks of the FTB forming the upper crust were 79 incorporated into the molten flow interiors Sheth et al (2011) thought that pieces and blocks of 80 the FTB upper crust (highly vesiculated and therefore low-density) might have sunk into the 81 molten interior owing to temporary gravity instabilities, or perhaps limited lava convection and 82 crustal overturn, but noted that these explanations were unsatisfactory In this paper, we describe 83 a new occurrence of FTB-cored rosettes in a rubbly pahoehoe flow on the island of Elephanta in 84 the Mumbai harbour (Fig 1a,b), in the western Deccan Traps We provide field, petrographic 85 and geochemical data on this flow, and present a model for the FTB-cored rosettes that is based 86 on well-understood mechanisms of emplacement of such flows and does not require sinking of 87 blocks of the FTB upper crust into the flow interior AC C EP 78 88 89 Geology of Elephanta Island and its rubbly pahoehoe flow ACCEPTED MANUSCRIPT Elephanta Island in the Mumbai harbour rises 168 m above sea level and is covered in 91 large part with dense jungle (Fig 1b) It is however well known for the ca mid-6th century A.D 92 Hindu rock-cut caves, a World Heritage Site of the UNESCO (The United Nations Educational, 93 Scientific and Cultural Organization) since 1987 The caves are carved into small-scale 94 (Hawaiian-size) compound pahoehoe flows (in the terminology of Walker, 1971) A detailed 95 description of these compound flows can be found in Sheth et al (in press) The lava flows dip 96 west-northwest by ~12° due to the Panvel flexure, a late-stage tectonic megastructure along the 97 western Indian rifted margin (Sheth, 1998; Samant et al., in press) The southeastern part of the 98 island (Fig 1b) exposes a 40 m thick lava flow of rubbly pahoehoe which underlies the 99 compound flows of the Elephanta Caves The rubbly pahoehoe flow was quarried during the 100 early to mid-seventies to provide construction material for the then upcoming major port of 101 Nhava–Sheva km east of the island (Fig 1b), but the quarrying was stopped in a few years as it 102 was found detrimental to the historical monument The rubbly pahoehoe flow is traversed by two 103 subparallel, oblique-slip normal faults with well-developed slickensides and easterly downthrows 104 (Samant et al., in press) M AN U SC RI PT 90 Observations of the rubbly pahoehoe flow made on the eastern fault surface, outside the 106 abandoned quarry, show an upper tier of joint columns that dip in various orientations, and are 107 overlain by FTB upper crust, whereas the lower part of the flow is massive and structureless 108 (Fig 3) Fans of well-developed, long and slender, subvertical columns are seen in outcrops 109 along the trace of the eastern fault (Fig 3a,b) When followed southwards, almost horizontal 110 columns are seen (Fig 3c) in immediate lateral contact with subvertical columns (Fig 3d), which 111 are immediately juxtaposed against a meters-thick FTB crust (Fig 3e) The boundary between 112 the FTB upper crust and the columnar tier below is thus highly irregular An outcrop only meters 113 from the outcrops in Fig 3c–e shows fans of short and thick joint columns diverging and 114 widening from the FTB upper crust (Fig 3f) Note that the outcrops shown in Fig all contain a 115 FTB upper crustal zone in place, and fans of the columns diverging from the locally depressed 116 lower boundary of that zone, rather than rosettes of joint columns around loose and suspended 117 FTB blocks in the flow interior as seen in the Deccan examples of Fig AC C EP TE D 105 118 Observations inside the abandoned quarry also show an uppermost zone of FTB, 119 followed downwards by a zone of very chaotic columnar jointing patterns, followed downwards 120 by a massive and structureless zone (Fig 4a) It is hazardous to try to climb to the upper FTB ACCEPTED MANUSCRIPT crust but large blocks of the FTB upper crust, left by the quarrying, are found at the bottom of 122 the quarry near its entrance (Fig 4b) The thick flow core underlying the FTB crust shows a 123 large radial jointing structure with no obvious FTB core (circled in Fig 4c), and adjacent to this, 124 random and widely spaced columns (Fig 4d) The field appearance of this flow is somewhat 125 reminiscent of that with the abundant FTB-cored rosettes in the Koyna quarry (Fig 2a-d) RI PT 121 At one place along the eastern fault, a large landslide has produced much bouldery rubble 127 including boulders with slickensides and boulders of FTB upper crust (Fig 5a) The top of this 128 landslide face shows a FTB inclusion within a complete columnar rosette (Fig 5b), whereas a 129 large boulder several meters in size shows a core of FTB upper crust to which columnar lava is 130 welded on two opposite sides (Fig 5c) SC 126 132 M AN U 131 Samples, petrography, and geochemistry We sampled the FTB upper crust (sample IP1) and an underlying column (sample IP2) 134 from the outcrop in Fig 3f We also sampled the FTB core (sample IP3) and one of the attached 135 columns (sample IP4) from the large landslide boulder shown in Fig 5c Inside the quarry, we 136 collected a sample from the fresh, fine-grained, non-vesicular and massive lower part of the flow 137 (sample ELF1) inside the quarry, and another (sample ELF2) from a loose block of joint columns 138 on the path outside the quarry that came from a nearby outcrop Sample ELF1 thus represents the 139 relatively fresh, massive lower part of the flow, whereas samples ELF2, IP2 and IP4 represent 140 various columnar zones in the flow The base of the flow is marked by a red bole horizon partly 141 exposed at the quarry entrance, possibly representing alteration of the flow’s glassy lower chilled 142 margin as shown by Duraiswami et al (2008) The red bole is affected by the eastern fault and 143 shows well-developed slickensides as reported by Samant et al (in press) TE EP AC C 144 D 133 In thin section, sample ELF1 shows a very fine-grained texture with microphenocrysts of 145 olivine and clinopyroxene (Fig 6a), whereas the joint column samples also all show a very fine- 146 grained groundmass with olivine, plagioclase and clinopyroxene microphenocrysts (Fig 6b) 147 which sometimes form clots comprising many individuals (Fig 6c) 148 Small, fresh chips of these samples were cleaned in an ultrasonic bath and ground to 149 powders of < 75 µm grain size using a Retsch PM-100 planetary ball mill and stainless steel 150 grinding balls Solutions of the sample powders were prepared following the methods described 151 in Vijayan et al (2016), and analyzed for the major and a few trace elements on a SPECTRO ACCEPTED MANUSCRIPT ARCOS inductively coupled plasma atomic emission spectrometer at the Sophisticated 153 Analytical Instrumentation Facility (SAIF), IIT Bombay Several U S Geological Survey rock 154 standards covering a large compositional range were dissolved along with the samples The 155 standards DNC-1, BIR-1, BCR-2 and BHVO-2 were used for calibrating the instrument, whereas 156 the standard W-2a was analyzed as an unknown to estimate the analytical accuracy Loss on 157 ignition (LOI) values were determined by heating the rock powders at 1000°C in platinum 158 crucibles, after overnight drying in an oven at 110 °C to drive away adsorbed moisture (H2O−) 159 The geochemical data are presented in Table 1, along with the CIPW norms and Mg# (Mg 160 Number) values obtained after adjusting the data on an LOI-free basis with the program 161 SINCLAS (Verma et al., 2002) The three samples have closely similar compositions They are 162 moderately evolved with MgO contents of 6-7 wt.% and Mg Numbers of 52.3 to 55.3, contain a 163 little or no normative olivine or quartz, and are classified as subalkalic basalts by SINCLAS M AN U SC RI PT 152 164 165 Discussion 166 4.1 Current understanding The FTB-cored rosettes of Sheth et al (2011) shown in Fig are in a lava flow of the 168 Ambenali Formation, exposed at Koynanagar and near Sajjangad (Fig 1a) which are 32 km from 169 each other Another example described by them comes from near Burhanpur (Fig 1a) in the 170 northern Deccan For our samples ELF1 and ELF2, the TiO2 vs Zr/Y plot of Peng et al (2014), 171 involving three alteration-resistant elements, indicates a character transitional between the 172 Poladpur and Ambenali Formations (both of which are in the upper part of the stratigraphically 173 3.5 km thick Western Ghats sequence), but more data including Sr–Nd isotopic data would be 174 needed to establish their exact geochemical correlations (if any) to that sequence (a topic beyond 175 the scope of this study) TE EP AC C 176 D 167 The field observations of Sheth et al (2011) on the Deccan FTB-cored rosettes can be 177 summarized as follows: The flows hosting these structures are thick (> 20 m) and typically 178 chaotically jointed, and the many FTB-cored rosettes they contain are generally unconnected to 179 the upper FTB crust and occur suspended at various heights in the flow The FTB crusts and 180 cores are highly vesicular (with the vesicles now filled by quartz and zeolites) The basalt lava 181 forming the columns is quite non-vesicular and dense, and it was chilled against the FTB core as 182 inferred from its glassy margin The joint columns, narrow close to the FTB core because of ACCEPTED MANUSCRIPT faster cooling (Grossenbacher and McDuffie, 1995), become progressively wider away from the 184 core before finally merging with the massive flow interior (Sheth et al., 2011) Because joint 185 columns should develop perpendicular to isotherms (parallel to the highest thermal gradient, e.g., 186 Budkewitsch and Robin, 1994; Lyle, 2000), Sheth et al (2011) inferred that each FTB core in the 187 Deccan FTB-cored rosettes had acted as a cold inclusion that strongly warped the isotherms 188 around itself, and the columns grew outwards from this inclusion RI PT 183 Sheth et al (2011) showed that past interpretations of these features as breccia-filled 190 volcanic vents (Agashe and Gupte, 1971) and progressively inward-closing lava tubes (Waters, 191 1960; Spry, 1962; Misra, 2002) were untenable These features are too small to be vents, they are 192 present within individual thick flows, being exposed in vertical sections they would imply 193 horizontal feeder conduits, and the eruptive material they should have produced is missing Sheth 194 et al (2011) also pointed out that in a progressively cooling lava tube the central core should be 195 the last to solidify, whereas in the FTB-cored rosettes the cores (the FTB inclusions) had been 196 cold to begin with, and the columnar jointing had propagated outwards from them The “war 197 bonnet” structures of the Columbia River flood basalts (so named because of their resemblance 198 to the radial feathers on the headdress of a native American) are large radial jointing structures 199 tens of meters across, have no obvious cores, and were interpreted as filled lava tubes by Waters 200 (1960) and Spry (1962), with the radial arrangement of the columns due to the isotherms being 201 concentric around their centers Greeley et al (1996) noted the absence of concentric 202 arrangement of vesicle zones in the war bonnet structures, the textural uniformity of the 203 columnar basalt with the rest of the flow, and their continuous transition from one to the other 204 with the columns widening radially outwards (Waters, 1960) Greeley et al (1996) concluded 205 that the war bonnet structures could not be lava tubes, and Sheth et al (2011) concluded the 206 same for the FTB-cored rosettes M AN U D TE EP AC C 207 SC 189 Simple ingress of meteoric water into the flow interior along cooling joints (e.g., Long 208 and Wood, 1986; DeGraff and Aydin, 1987; Lyle, 2000) could well explain a highly irregular 209 and uneven boundary between FTB upper crust and the lava interior, and the resultant distorted 210 isotherms would explain the highly variable joint column orientations from near-horizontal to 211 vertical just below this boundary as governed everywhere by the local shape of the isotherms 212 (Fig 3) However, this mechanism would not explain the sizeable (several meters) FTB cores ACCEPTED MANUSCRIPT 213 which were undoubtedly derived from the highly vesicular FTB upper crust of the same flow as 214 shown by trace element matches and, in particular, Nd isotopic data (Figs and 5b,c) Sheth et al (2011) noted that though liquid basalt is about 10% lower in density than 216 solid basalt (Philpotts and Ague, 2009), the highly vesicular and therefore lower-density FTB 217 crust of their rubbly pahoehoe flows should have floated on the molten interior, rather than 218 having sunk into it Nevertheless, they thought that temporary gravitational instabilities or even 219 limited lava convection and crustal overturn (perhaps due to new injection of hotter and therefore 220 lower-density lava into the flow), may have caused the broken FTB crust to sink into the molten 221 interior Whereas bodily sinking of gravitationally stable crystal mush through the molten 222 interior of the flow was suggested by Philpotts and Dickson (2002) for the Holyoke flood basalt 223 flow (representing the 200 Ma Central Atlantic Magmatic Province, CAMP, in the northeastern 224 USA), no physical explanation has been available for the sinking of FTB crust into a flow 225 interior as apparently required by the Deccan FTB-cored rosettes M AN U SC RI PT 215 226 227 4.2 A model We propose a model for the origin of these FTB-cored rosettes in which the incorporation 229 of FTB upper crust into the molten flow interior does not require sinking of the FTB crust into 230 the molten interior TE D 228 We note that pahoehoe flows of all sizes typically grow by inflation (e.g., Hon et al., 232 1994; Self et al., 1997; Reidel, 1998) A key aspect of inflation is that initial flow thickness 233 during emplacement is much less than the flow thickness during final solidification (Fig 7) A 234 lava lobe (single cooling unit of size ranging from centimeters to meters) would swell (inflate) as 235 lava that continued to be supplied lifted the upper crust (Fig 7a) A viscoelastic upper crustal 236 layer under the top brittle crustal layer underwent ductile stretching and accommodated the 237 incoming lava The lobe, inflating like a water-filled balloon, eventually burst and a new lobe 238 emerged and itself grew by inflation, this process repeating as long as lava continued to be 239 supplied from the vent If fresh lava inputs into the flow lobes were very rapid, the various lobes 240 would coalesce and a lava flow that was a single thick cooling unit would form AC C EP 231 241 Shearing of the viscoelastic crust could occur if the flux of the lava continuing to flow 242 under the crust rose significantly leading to rapid flow inflation, or during large surges in the 243 lava supply, as when lava temporarily ponded in a part of the lava flow was suddenly released in ACCEPTED MANUSCRIPT Parisio, L., Jourdan, F., Marzoli, A., Melluso, L., Sethna, S F., Bellieni, G., 2016 40Ar/39Ar ages 431 of alkaline and tholeiitic rocks from the northern Deccan Traps: implications for 432 magmatic processes and the K-Pg Boundary Journal of the Geological Society of 433 London 173, 679–688 RI PT 430 434 Peng, Z X., Mahoney, J J., Vanderkluysen, L., Hooper, P R., 2014 Sr, Nd and Pb isotopic and 435 chemical compositions of central Deccan Traps lavas and relation to southwestern 436 Deccan stratigraphy In: Sheth, H C., Vanderkluysen, L (Eds.), Flood Basalts of Asia 437 Journal of Asian Earth Sciences 84, 83–94 Peterson, D W., Tilling, R I., 1980 Transition of basaltic lava from pahoehoe to ‘a’ā, Kilauea 439 volcano, Hawaii: field observations and key factors Journal of Volcanology and 440 Geothermal Research 7, 271–293 442 M AN U 441 SC 438 Philpotts, A R., Ague, J J., 2009 Principles of Igneous and Metamorphic Petrology, 2nd Edn Cambridge University Press, 667 pp 443 Philpotts, A R., Dickson, L D., 2002 Millimeter–scale modal layering and the nature of the 444 upper solidification zone in thick flood–basalt flows and other sheets of magma Journal 445 of Structure Geology 24, 1171–1177 448 449 Science 241, 663–668 D 447 Rampino, M R., Stothers, R B., 1988 Flood basalt volcanism during the past 250 million years TE 446 Reidel, S P., 1998 Emplacement of Columbia River flood basalt Journal of Geophysical Research 103, 27393–27410 Reidel, S P., Camp, V E., Tolan, T L., Martin, B S., 2013 The Columbia River flood basalt 451 province: stratigraphy, areal extent, volume, and physical volcanology In: Reidel, S P., 452 Camp, V E., Ross, M E., Wolff, J A., Martin, B S., Tolan, T L., Wells, R E (Eds.), 454 AC C 453 EP 450 The Columbia River Flood Basalt Province Geological Society of America, Special Papers 497, 1–43 455 Samant, H., Pundalik, A., D’Souza, J., Sheth, H., Carmo Lobo, K., D’Souza, K., Patel, V., in 456 press Geology of the Elephanta Island fault zone, western Indian rifted margin, and its 457 significance for understanding the Panvel flexure Journal of Earth System Science, doi: 458 10.1007/s12040-016-0793-8 459 460 Scheidegger, A E., 1978 The tectonic significance of joints in the Canary Islands Rock Mechanics 11, 69–85 15 ACCEPTED MANUSCRIPT Self, S., Thordarson, Th., Keszthelyi, L., 1997 Emplacement of continental flood basalt lava 462 flows In: Mahoney, J J., Coffin, M F (Eds.), Large Igneous Provinces: Continental, 463 Oceanic, and Planetary Flood Volcanism American Geophysical Union Geophysical 464 Monograph 100, p 381–410 RI PT 461 465 Self, S., Schmidt, A., Mather, T A., 2014 Emplacement characteristics, time scales, and 466 volcanic gas release rates of continental flood basalt eruptions on Earth In: Keller, G., 467 Kerr, A C (Eds.), Volcanism, Impacts, and Mass Extinctions: Causes and Effects 468 Geological Society of America, Special Papers 505, 319–337 Sen, B., Sabale, A B., 2011 Flow–types and lava emplacement history of Rajahmundry Traps, 470 west of River Godavari, Andhra Pradesh Journal of the Geological Society of India 78, 471 457–467 473 474 475 M AN U 472 SC 469 Sheth, H C., 1998 A reappraisal of the coastal Panvel flexure, Deccan Traps, as a listric fault– controlled reverse drag structure Tectonophysics 294, 143–149 Sheth, H C., 2006 The emplacement of pahoehoe lavas on Kilauea and in the Deccan Traps Journal of Earth System Science 115, 615–629 Sheth, H C., Ray, J S., Senthil Kumar, P., Duraiswami, R A., Chatterjee, R N., Gurav, T., 477 2011 Recyling of flow–top breccia crusts into molten interiors of flood basalt lava flows: 478 field and geochemical evidence from the Deccan Traps In: Ray, J., Sen, G., Ghosh, B 479 (Eds.), Topics in Igneous Petrology Springer, p 161–180 TE D 476 Sheth, H., Meliksetian, K., Gevorgyan, H., Israyelyan, A., Navasardyan, G., 2015 Intracanyon 481 basalt lavas of the Debed River (northern Armenia), part of a Pliocene–Pleistocene 482 continental flood basalt province in the South Caucasus Journal of Volcanology and 483 Geothermal Research 295, 1–15 485 486 487 AC C 484 EP 480 Sheth, H., Samant, H., Patel, V., D’Souza, J., in press The volcanic geoheritage of the Elephanta Caves, Deccan Traps, western India Geoheritage, doi: 10.1007/s12371-016-0214-z Spry, A., 1962 The origin of columnar jointing, particularly in basalt flows J Austr Geol Soc 8, 191–216 488 Swanson, D A., Wright, T L., 1981 The regional approach to studying the Columbia River 489 Basalt Group In: Subbarao, K V., Sukheswala, R N (Eds.), Deccan Volcanism and 490 Related Basalt Provinces in Other Parts of the World Geological Society of India 491 Memoir 3, p 58–80 16 ACCEPTED MANUSCRIPT 492 Verma, S P., Torres–Alvarado, I S., Sotelo–Rodriguez, Z T., 2002 SINCLAS: standard 493 igneous norm and volcanic rock classification system Computers & Geosciences 28, 494 711–715 Vijayan, A., Sheth, H., Sharma, K K., 2016 Tectonic significance of dykes in the Sarnu– 496 Dandali alkaline complex, Rajasthan, northwestern Deccan Traps Geoscience Frontiers 497 7, 783–791 RI PT 495 Walker, G P L., 1971 Compound and simple lava flows and flood basalts In: 499 Aswathanarayana, U (Ed.) Deccan Trap and Other Flood Eruptions Bulletin of 500 Volcanology 176, 599–610 502 503 504 505 506 Waters, A C., 1960 Determining directions of flow in basalts American Journal of Science 258A, 350–366 M AN U 501 SC 498 Wignall, P B., 2001 Large igneous provinces and mass extinctions Earth-Science Reviews 53, 1–33 Wilson, S A., 1977 Data compilation for USGS reference material BHVO–2, Hawaiian basalt U S Geological Survey Open File Rep AC C EP TE D 507 17 ACCEPTED MANUSCRIPT 508 Figure captions 509 Figure (a) Sketch-map of the Deccan Traps (shaded) showing the Western Ghats escarpment 511 (WGE, heavy dashed line) and the region with abundant rubbly pahoehoe flows documented 512 (enclosed within the thin line, Duraiswami et al., 2008) Some important localities exposing 513 these flows (Duraiswami et al., 2008; Sheth et al., 2011) are marked Rubbly pahoehoe flows are 514 also found in Saurashtra in the northwestern Deccan Traps (R Duraiswami and H Sheth, 515 unpubl data) (b) Google Earth image of Elephanta Island in the Mumbai harbour, with parts of 516 Nhava–Sheva port on the Indian mainland immediately to the east Box with black boundary 517 shows the area of present study SC RI PT 510 M AN U 518 Figure (a–d) Outcrop photographs from the Koyna quarry at Koynanagar showing the various 520 FTB-cored rosettes in a thick rubbly pahoehoe flow Note the radial arrangement and gradual 521 widening of columns with distance away from the FTB cores The rosette in (c) lacks a scale as it 522 is exposed high up on a vertical face, but it is many meters in size (e) Large FTB-cored rosette 523 in the same flow 32 km away (as identified from the geochemical-isotopic data of Sheth et al., 524 2011), exposed in a road cut near Sajjangad Geologists providing a scale are Rudranarayan 525 Chatterjee (a) and Cliff Ollier (b, d, e) TE 526 D 519 Figure Structures in the Elephanta Island rubbly pahoehoe flow, viewed outside the 528 abandoned quarry and along the eastern fault (Samant et al., in press) (a) Columns fanning from 529 the contact with the FTB upper crust;Geologist for scale is Keegan Carmo Lobo A part of the 530 columnar tier is enlarged in (b); Panels (c) to (e), in that order, show outcrops adjacent or only a 531 few meters apart along a south-north direction and at almost the same height; subhorizontal 532 columns (c); subvertical columns (d); FTB upper crust (e), and the contact between FTB crust 533 and the columnar zone (f) Note widening of the columns downwards AC C 534 EP 527 535 Figure Structures in the Elephanta Island rubbly pahoehoe flow, viewed inside the abandoned 536 quarry (a) Section through the flow with the FTB crust at the top, a thick zone of random and 537 chaotic jointing below, and massive, largely joint-free lava at the base; (b) a large boulder of 538 FTB upper crust at the entrance to the quarry; (c) a large rosette without any obvious FTB core 18 ... in a rubbly pahoehoe lava flow, Elephanta Island, Deccan Traps, and a model for their origin a India b University of California San Diego, La Jolla, CA 92093-0225, USA c SC Present address: Institute... (FTB) and a continuously thickening lobe of rubbly pahoehoe 249 (Fig 7c) Slabby and rubbly pahoehoe are transitional lava types between pahoehoe and ? ?a? ??ā, the 250 end member forms of basaltic lava. .. Current understanding The FTB -cored rosettes of Sheth et al (2011) shown in Fig are in a lava flow of the 168 Ambenali Formation, exposed at Koynanagar and near Sajjangad (Fig 1a) which are 32 km