AGU Chapman Conference The Great Plume Debate: The Origin and Impact of LIPS and Hot Spots
Ben Nevis Hotel, Fort William, Scotland, United Kingdom 28 August – 1 September 2005
Conveners
• Ian Campbell, Research School of Earth Sciences, The Australian National University, Canberra, A.C.T. 0200, Australia, email: [email protected] • Gillian R. Foulger, Dept. Earth Sciences, Science Laboratories, University of Durham, South Rd., Durham, DH1 3LE, U.K., email: [email protected] • James H. Natland, Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker, Causeway, Miami, FL 33149, email: [email protected] • Dean C. Presnall, Geophysical Laboratory, 5251 Broad Branch Rd., NW, Washington, DC 20015, e-mail [email protected] • W. Jason Morgan, Dept. Earth Planet. Sci., Harvard Univ., 20 Oxford St., Cambridge, MA 02138, email: [email protected]
Program Committee
• J. Godfrey Fitton, University of Edinburgh, [email protected] • Brian Bell, University of Glasgow, [email protected] • C. Henry Emeleus, University of Durham, [email protected]
Sponsors
• National Science Foundation • Statoil • Department of Earth Sciences, University of Durham
Cover Image: View looking north-west from Bidean nam Bian, in the Devonian Glen Coe volcanic complex, toward Loch Leven and Ballachulish. The Ballachulish intrusive complex forms the high ground in the left middle distance (south of the Ballachulish Bridge); the lower ground to the north of the bridge is composed of Dalradian metamorphic rocks. Loch Linnhe and the Great Glen Fault run across the middle of the photograph (SW-NE). Moine metamorphic rocks form the mountains in the distance. Fieldtrip 2, the “Road to the Isles” half-day excursion, will comprise a transect of this part of the Caledonian orogenic belt.
Great Plume Debate Chapman Conference: Agenda at a Glance
Sunday 28 Monday 29 Tuesday 30 Wednesday 31 Thursday 1 Aug Aug Aug Aug Sept
Welcome and 8:00 - 8:30 introduction
Petrology & Plume theory Geochronology Field evidence 8:30 - 10:00 Temperature I Geochemistry & predictions I I I
10:00-10:30 coffee coffee coffee coffee coffee
Alternative Petrology & Temperature Geochronology Field evidence 10:30-12:00 theories & Geochemistry II II II predictions II
lunch & lunch & lunch & posters lunch & posters lunch & 12:00 - 13:30 posters posters posters
Lithosphere & 13:30-15:00 mantle physics Seismology I Discussion I & dynamics I Fieldtrip I: Fieldtrip II:
15:00-15:30 tea tea tea Ballachulish The Road to and Glen Coe the Isles Lithosphere & Discussion II 15:30-17:00 mantle physics Seismology II & Synthesis & dynamics II
17:00-18:00 posters posters posters 18:00-19:00 posters posters
19:00-20:00 dinner dinner dinner dinner dinner 20:00 posters posters posters posters posters onwards Lecture:
20:30-21:00 Scottish
Highlands geology in a
21:00-21:30 global context Planetary Scottish Ian Dalziel evening entertainment 21:30-22:00
2 Preliminary Schedule AGU Chapman Conference on the Great Plume Debate: The Origin and Impact of LIPs and Hot Spots 28 August - 01 September 2005 Fort William, Scotland, United Kingdom
Sunday 28 August 2005
Time Speaker Title
8:00 – 8:30h I. Campbell & G. Foulger Welcome and Introductions
Plume theory & predictions Conveners: Ian Campbell & W. Jason Morgan I. Campbell Testing the Plume Hypothesis 8:30 – 10:00h J. Morgan The Deep Mantle Plume Hypothesis Discussion 10:00 – 10:30h Coffee Alternative theories & predictions Conveners: Gillian Foulger & Dean Presnall G. Foulger The Generation of Melting Anomalies by Plate Tectonic Processes L. Elkins- Continental Magmatism Caused by Lithospheric Rayleigh-Taylor 10:30 – 12:00h Tanton Instabilities D. Sandwell Cracks and Warps in the Lithosphere from Thermal Contraction Discussion 12:00h-13:30h Lunch and Posters Lithosphere & mantle physics & dynamics I Conveners: John Tarduno & Erin Beutel J. Davies Mantle Convection – An Overview U. Hansen Generation and Evolution of Plumes in Mantle-Relevant Scenarios 13:30 – 15:00h J. A. Tarduno On the Motion of Hawaii and Other Mantle Plumes Discussion 15:00 – 15:30h Tea Lithosphere & mantle physics & dynamics II Conveners: Erin Beutel & John Tarduno E. Burov The Plume Head - Continental Lithosphere Interaction Using a Tectonically Realistic Formulation for the Lithosphere S. King How Many Hotspots Can be Explained by Edge Driven Convection? 15:30 – 17:00h J. Van Wijk Formation of Volcanic Rifted Margins: Influence of the Pre-Rift Lithosphere Architecture W. Stuart Hawaii Volcano Chain as a Thermoelastically-Driven Propagating Crack Discussion 17:00 – 19:00h Posters 19:00 – 20:00h Dinner 20:00h + Posters
3 Monday 29 August 2005
Temperature I Conveners: Nick Arndt & Carol Stein N. Arndt The Temperatures of Mantle Plumes M. Cheadle Komatiites and the Temperature of the Mantle: “Some Like It Hot”. 8:30 – 10:00h C. M. Lesher High-Mg Magmatism Through Time: Implications for Mantle Plumes Discussion 10:00 – 10:30h Coffee Temperature II Conveners: Carol Stein & Nick Arndt R. Harris Observations of Heat Flow on Hotspot Swells D. Presnall MORB Major-Element Systematics: Implications for Melting Models and Mantle Temperatures 10:30 – 12:00h T. Falloon Magmatic Crystallization Temperatures of Tholeiite Magmas: Implications for the Existence of Thermally Driven Mantle Plumes Discussion 12:00 – 13:30h Lunch and Posters 13:30 – 18:00h Field Trip I: Ballachulish and Glen Coe 19:00 – 20:00h Dinner 20:00h + Posters Lecture: 20:30 – 21:30h Where on Earth was Fort William during Neoproterozoic and Paleozoic times? Scottish Highlands geology in a global context. Ian Dalziel
4 Tuesday 30 August 2005
Geochronology I Conveners: Bob Duncan & Ajoy Baksi R. Duncan Timing and Duration of Volcanism at Large Igneous Provinces: Implications for Geodynamics and Links to Hotspots K. Hoernle Origin of Long-term Intraplate Volcanism in the Canaries, Madeira, Galapagos and New Zealand: Which are Consistent with 8:30 – 10:00h the Plume Hypotheses? A. Koppers The Geochronology of Hotspot Trails and the Timing of the Hawaii-Emperor Bend Discussion 10:00 – 10:30h Coffee Geochronology II Conveners: Ajoy Baksi & Bob Duncan A. Baksi Critical Assessment of Radiometric Ages for Oceanic Hotspot Tracks, Based on Statistical Analysis of Individual Ages, and Evaluation of the Alteration State of the Material Dated J. O’Connor Volcanic Imprint of Oceanic Hot Spots and LIPs: Shallow/Local 10:30 – 12:00h versus Deep/Global? D. Praeg Cenozoic Vertical Movements on the NW European ‘Passive’ Margin: Responses to Upper Mantle Convection? Discussion 12:00 – 13:30h Lunch and Posters Seismology I Conveners: Guust Nolet & Thorne Lay R. Allen Constraining the Geometry and Flow of the Iceland Mantle Upwelling D. Zhao Multiscale Seismic Tomography of Mantle Plumes and 13:30 – 15:00h Subducting Slabs G. Nolet The Role of Mantle Plumes in the Earth’s Heat Budget Discussion 15:00 – 15:30h Tea Seismology II Conveners: Thorne Lay & Guust Nolet A. Deuss Constraints on the Observation of Mantle Plumes Using Global Seismology 15:30 – 17:00h B. Julian Guided Seismic Waves: Possible Mantle-Plume Diagnostics T. Lay Is the “D” Region the Source of Mantle Plumes? Discussion 17:00 – 19:00h Posters 19:00 – 20:00h Dinner Planetary Convener: Donna Jurdy W. Hamilton The Surface of Venus Records Ancient Impacts, Not Young Plumes V. Hansen Venus’ Many Circles: Extraterrestrial Clues for the Great Plume 20:30 – 22:00h Debate A. Jones Meteorite Impacts as Triggers to LIPs and Hotspots C. Reese Impact Induced Martian Mantle Plumes: Implications for Tharsis Discussion
5 Wednesday 31 August 2005
Field evidence I Conveners: Andy Saunders & David Sandwell 8:30 – 10:00h A. Saunders Plumes and Uplift S. Jones Uplift Associated with the North Atlantic Igneous Province Y. Xu Surface Responses to Mantle Plume: Sedimentation and Lithofacies Paleogeography in SW China Before and After the Emeishan Flood Volcanism Discussion 10:00 – 10:30h Coffee Field evidence II Conveners: David Sandwell & Andy Saunders H. Sheth The Deccan Beyond the Plume Hypothesis V. Sallarès Crustal Seismology Helps Constrain the Nature of Mantle Melting Anomalies. Galápagos Volcanic Province: A Case 10:30 – 12:00h Study J. Winterer Midplate Volcanic Overprinting: New Wine in Old Bottles Discussion 12:00 – 13:30h Lunch and Posters 13:30 – 18:00h Field Trip II: The Road to the Isles 19:00 – 20:00h Dinner 20:00h + Posters
Thursday 01 September 2005
Petrology & Geochemistry I Conveners: Eiichi Takahashi & Dean Presnall C. Hawkesworth Geochemistry and Mantle Plumes J. M. Rhodes Magmatic Evolution of Mauna Loa Volcano: Implications for a Chemically and Thermally Zoned Mantle Plume 8:30 – 10:00h E. Takahashi Magma Genesis in a Mantle Plume: Based on High-pressure Melting Experiments and Growth History of Some Hawaiian Volcanoes Discussion 10:00 – 10:30h Coffee Petrology & Geochemistry II Conveners: Dean Presnall & Eiichi Takahashi G. Fitton Do Hotspot Basalts Share a Common Mantle Source? M. Keskin Eastern Anatolia: A Hot Spot in a Collision Zone Without a Mantle Plume 10:30 – 12:00h A. Schersten The Hf-W Perspective on Whether a Trace of the Earth’s Core Exists in Hot Spot Volcanic Rocks Discussion 12:00 – 13:30h Lunch and Posters 13:30 – 15:00h Discussion I 15:00 – 15:30h Tea 15:30 – 17:00h Discussion II and Synthesis 17:00 – 19:00h Posters 19:00 – 20:00h Dinner 20:00h + Posters
6
Poster Session: Sunday, 28 August 2005
First Author Title S. Goes Testing Thermal Whole Mantle Plumes Seismically Withdrawn How Geometry and Ages of Global Hotspots are Explained by Classical Hypotheses of Rigid Plate and Fixed Hotspot M. Yamamoto Plume-fed Asthenosphere Flow Model: Evidence from Isotopic Variation Along Mid-ocean Ridges Y. Niu Slab Dehydration, Subcontinental Lithosphere Thinning and Widespread Mesozoic/Cenozoic Volcanism in Eastern China: A Special Consequence of Plate Tectonics S. V. Rasskazov Late Mesozoic Through Cenozoic Magmatism in East Asia: Effect of Directly and Obliquely Subducted Slab Flexures J. W. Sears Fracture Propagation on a Sphere: Implications for Non-Plume Origin of Large Igneous Provinces on Fragmenting Supercontinents K. K. Sharma Neoproterozoic Anorogenic Magmatism Associated with Rodinia Breakup: Not a Result of Mantle Superplume A. Tiwary Are the Precambrian Magma Chambers Responsible for the Cretaceous Deccan Trap Volcanism of India? E. K. Beutel Large LIPs and the Mantle Squeeze: A Mass Balance Approach to Hotspots J. H. Davies Mantle Convection - Plumes Rooted in Mid-Mantle *A. Harris The Many Potential Faces of Buoyant Mantle Upwellings: Diversity Within the Plume Family V. C. Manea Thermal Structure Beneath Kamchatka and Plume to Arc Magmatism Transition M. Manea Thermal Structure of the Cocos Slab Beneath Southern Mexico and its Relationship with the Arc Volcanism Withdrawn Small-scale Convective Instabilities in the Upper Mantle– A Generic Class of Hotspots Linked to Recent Continental Collision in Europe and the Circum- Mediterranean Region
* Student
Poster Session: Monday, 29 August 2005
G. R. Foulger How Hot is Iceland? R. R Keays Why the High PGE Contents of Komatiites, Picrites and Allied Rocks Require Mantle Plumes H. Mashima Thermal State of NW Kyushu Mantle Suggested by Petrochemistry of Primitive Basalts J. H. Natland Influence of Eclogite in Mantle Sources on ‘Hot-spot’ Temperatures K. D. Putirka Evidence for High Temperature Mantle Plumes Based on Olivine-Liquid Thermometry C. A. Stein Does Hydrothermal Circulation Mask Anomalously High Heat Flow at Hot Spots?
7 Poster Session: Tuesday, 30 August 2005
L. A. Morgan Patterns of Rhyolitic Volcanism in the Path of the Yellowstone Hot Spot C. Tegner Timescales of Flood Volcanism Recorded by Pressure Variations in Coeval Mafic Intrusions: A Fluid Inclusion Study of the Skaergaard Intrusion, East Greenland A. B. Watts Global Gravity, Bathymetry, and the Distribution of Submarine Volcanism through Space and Time M. Widdowson Duration and Timing of the Deccan CFBP G. Laske The SWELL Pilot Experiment off Hawaii - What Can We Learn About the Hawaiian Hotspot from Surface Waves? J. R. R. Ritter Comprehensive Imaging of the Eifel Plume, Central Europe Withdrawn Seismic Evidence for a Lower Mantle Origin of the Tanzania Hotspot *M. Xue Identifying the Origin of the Newberry Hotspot Track *Withdrawn Upper Mantle Structure Beneath the Azores Hotspot From Finite Frequency Seismic Tomography D. M. Jurdy Uplift and Rifting on Venus: Role of Plumes
* Student
Poster Session: Wednesday, 31 August 2005
*D. L. Abt Delamination Origin for Columbia River Flood Basalts and Wallowa Mountains Uplift in NE Oregon, USA A. J. Breivik Continental Breakup Magmatism and Transition to Hot-Spot Influenced Seafloor Spreading From the Moere Margin to the Norway Basin R. G. Cawthorn Kaapvaal Craton, South Africa: Repeated Basic Magmatism, Diamonds and Plumes I. W. D. Dalziel The Setting of LIPS in the Lithosphere Through Time: One Test of the Plume Hypothesis B. T. Jordan Testing a Propagating Shear-Zone Hypothesis for Age-Progressive Magmatism in a Continental Setting: The Oregon High Lava Plains M. Khodayar Compressional Structures Do Not Show Regional Horizontal Compression Near the Iceland Hotspot M. Khodayar Overview of Tectonic Deformation in Past and Present Rift-Jump Blocks, West and South Iceland Withdrawn Paleogene North Atlantic Igneous Province and the Iapetus Connection K. Pierce Geologic Evidence for a Mantle Plume Origin for Yellowstone: The Pattern and Scale of Volcanism, Faulting, and Uplift Along the Yellowstone Hotspot Track I. Norton Passive Margin Evolution: Are Plumes an Integral Part of Continental Breakup? W. W. Sager Does a Fault in the Plate Circuit Ruin Intra-ocean Comparison of Hotspot Tracks? A. V. Smirnov Co-location of Eruption Sites of the Siberian Traps and North Atlantic Igneous Province: Implications for the Nature of Hotspots and Mantle Plumes F. Tsikalas NE Atlantic Breakup and Evolution of the Norwegian-Greenland Conjugate Volcanic Margins: Field Evidence to the Great Plume Debate P. R. Vogt Bermuda: Lava-lamp Plume, Edge-driven Convection, or/and Response to Distant Plate Reorganization? M. Widdowson The Deccan Basalt–Basement Contact: Evidence for a Plume-Head Generated CFBP?
* Student
8 Poster Session: Thursday, 01 September 2005
*B. C. The Evolution of Floreana Island, Galápagos Archipelago II: The Result of a Christensen Contaminated Mantle Plume C. Class Evolution of Helium Isotopes in the Earth’s Mantle E. I. Inverse Trace Element Modeling of Mantle Components from Late Cenozoic Demonterova Basalts in Central Asia M. L. Frezzotti Fluid Inclusion Evidence for Water in the Mantle Beneath Hawaii *W. R. Griffin Testing Magmatic Emplacement Mechanisms in the Balcones Igneous Province of Texas Withdrawn The Evolution of Floreana Island, Galápagos Archipelago I: The Result of Upper Mantle Heterogeneities *Withdrawn Mantle Redox Conditions in LIPs: Constraints from the North Atlantic Igneous Province M. J. Hole Plumes or Rifting?: The Mesozoic Dykes of the Falkland Islands and Their Relationship to the Break-up of Gondwana. A. V. Ivanov Pliocene-Quaternary Alkaline Basalts of the Sredinny Ridge of Kamchatka: Evidence for Melting of Recycled Oceanic Crust in Tectonic Setting of a Modern Island Arc System B. T. Jordan Communicating the Plume Debate to Undergraduate Geoscience Students S. Keshav Re-Os-Pt Partitioning in Sulfur-bearing Solid/Molten Iron Metal at 3-22 GPa and 1300-1775 C: Is the Earth’s Outer Core So Giving? H. Mashima Recycling of Archean Peridotitic Komatiite in the NW Kyushu Source *R. Meyer The Vøring Plateau Volcanic Margin: A Key Rock Succession to Understand Continental Breakup During the Initial Stages of the Opening of the NE- Atlantic J. Natland Layered Mantle Alternative to Mantle Plumes: Evidence from the Pacific Plate S. Sensarma The Dongargarh Group: A Large Igneous Province at the Archean-Proterozoic Transition in India F. M. Stuart Statistical Comparison of 3He/4He Distributions in Mid-Ocean Ridge and Ocean Island Basalts D. Pandey Applicability of Large Magma Chambers to Deccan Volcanism: A Numerical Study J. Tuff Experimental Constraints on the Role of Garnet Pyroxenite in the Genesis of High-Fe Mantle Plume Derived Melts
* Student
9 Sunday Oral Sessions instabilities that would lead to a rising plume in the mantle. A consequence of Testing the Plume Hypothesis the model is a large burst of rising material when the column reaches the surface (a I. H. Campbell (Research School of Earth flood basalt) and the eventual 'death' of a Sciences, The Australian National Univ., plume when the source layer is 'drained'. Canberra, ACT 0200, Australia; Tel: +61- (3) There is large-scale horizontal flow 2-6125-4366; Fax +61-2-61258253; E- in the asthenosphere, moving material mail: [email protected]) away from 'excesses' of asthenosphere (where a plume impinges) toward The physics of low Reynolds number deficiencies where lots of lithosphere is plumes is well understood, which allows a made (especially at mid-ocean ridges). number of testable predictions to be made (4) We know the rate asthenosphere is about mantle plumes. They are predicted being 'destroyed' (by cooling and to originate from the core-mantle attaching itself to lithospheric plates); it boundary and consist of a large head, with equals the rate that lithosphere is being a diameter of ~1000km, followed by a 'destroyed' by subduction at trenches into narrower tail. When the head reaches the the deep mantle (~ 250 km3/yr.) How is top of the mantle it flattens to form a disk the asthenosphere replenished at this rate? with a diameter that is predicted to lie (a) Heat conduction upward from within the range 2000 to 2500km. Initial deeper mantle, softening the mantle just eruption from a plume head should be below? (1/50th too small) preceded by ~1000m of domal uplift. (b) Roll-like (2-D) upwelling of mantle Picrites are expected to dominate the first from just below mid-ocean rises? melting products of a new plume and they (c) Brought up by mantle plumes, and should be concentrated near the centre of spreading out in the asthenosphere? the volcanic province. All of these If (c) occurs, and asthenosphere can easily predictions are confirmed by observations, spread out horizontally beneath the which provide strong support for the lithosphere, it can 'turn off' the other two validity of the mantle plume hypothesis. modes. The mantle does not have an adiabatic thermal temperature gradient everywhere, the hot material from CMB The Deep Mantle Plume Hypothesis depths can make the asthenosphere hotter (in a potential temperature sense) than W. Jason Morgan (Dept. Earth Planet. what's below, i.e. make the asthenosphere Sci., Harvard Univ., 20 Oxford St., stable against 'upwelling' beneath the EPR, Cambridge, MA 02138, USA; Tel: +1- etc. 508-720-1509; E-mail: (5) The downgoing slabs are continually [email protected]) carrying chemical heterogeneities into the deep mantle (seds, oceanic crust, old In this talk I present the basic ideas of the oceanic islands, depleted mantle produced plume model. There is not time to back when making oceanic crust) on scales of each of these ideas with the supporting 10 m to 10 km. Mantle convection doesn't evidence -- that should hopefully be given completely homogenize these in subsequent talks at this meeting. heterogeneities -- the mantle is 'lumpy' on (1) Near surface, near the 'rigid' top scales of 10 cm to 10 km. boundary layer, flow in the mantle has a (6) This 'lumpiness' leads to the large-scale 2-D character (mid-ocean distinction between OIB and MORB ridges, subducting slabs). But away from chemistry. The 'easy to melt' lumps, with this rigid layer the flow pattern is more more of the large-ion incompatible fluid-like; rising columns are dynamically elements in them, melt first as a plume preferred to rising sheets. nears the surface. At depths of 150 km to (2) Stacey & Loper (1984) presented a 70 km, these 'plums' are preferentially model of how a low-density, low-viscosity extracted beneath the hotspots, leaving a layer at the base of the mantle would grow residue lacking these 'plums'. This residue
10 is then the 'Depleted Mantle Material', that breakup and re-suturing results in variable after a long horizontal journey becomes lithosphere structure, composition and the source of MORB's. (The more strength. extensive, shallower melting at mid-ocean rises can melt the remainder left behind In its simplest form, the theory proposes after the first-stage of melting.) that variability in mantle source fusibility and composition result in variable The key data supporting this hypothesis magmatic productivity. Variations in are: (1) the predictability in direction and stress and lithosphere strength control the rate of hotspot tracks, (2) the fact flood location of volcanism. The theory predicts basalts have similar chemistry to OIB's that melting anomalies will tend to occur and the observable ones have age in regions of extensional stress and that progressive 'tracks' leading away from where fusible source material is available, them, (3) the basic 'age' of OIB's (~1.8 Ga) melt volumes will tend to be large. The agrees with the high overturn rate of heat needed is, in the first instance, drawn plume material at ~250 km3/yr, and (4) from the upper mantle reservoir, which many specifics of OIB and MORB redistributes its heat through conduction, chemistry are in accord with this model. convection and the advection of melt. Fusible material residing in the upper mantle may melt to a high degree as it The Generation of Melting Anomalies approaches ambient upper mantle by Plate Tectonic Processes temperatures. For example, the liquidus of eclogite may be lower than the solidus of G. R. Foulger (Dept. Earth Sciences, Univ. peridotite. of Durham, Durham DH1 3LE, U.K.; Tel: +44-191-386-4533; Fax: +44-191-334- There is considerable support for this 2301; E-mail: [email protected]) theory. One third of all "hot spots" lie at or close to spreading plate boundaries and Intraplate and on-ridge melting anomalies many lie in extensional intraplate regions may result from processes associated with such as the East African Rift, the Basin & plate tectonics. The key elements of this Range Province, and back-arc regions. general theory are: Many large igneous provinces (LIPs) are formed where continents break up along 1. Dehomogenising processes: Fusible old suture zones, in particular where three components are removed from the mantle or more branches meet. Much at ridges and re-introduced at subduction geochemical evidence is consistent with zones as oceanic crust, which transforms the presence of recycled fusible material in to eclogite at depth. Oceanic and "hot spot" and LIP lavas, including continental lithosphere are metasomatised continental crust, subducted oceanic crust, over time. Thickened continental and metasomatised lithosphere. lithosphere in collision zones may delaminate, and continental lithosphere Predictions of the theory include: may erode at the time of continental breakup and be recycled into the a) the melt volumes observed at "hot convecting upper mantle. These processes spots" and LIPs can be quantitatively result in lateral variations in fusibility, modeled by lithosphere extension and chemistry and temperature in the upper source fertility without the need for very mantle. high temperatures, 2. Processes that cause stress and b) the pattern of vertical motion lithosphere strength to vary: Lithosphere accompanying LIP emplacement may be cooling and spatial and temporal variations explained by relatively shallow tectonic in plate boundary type, tectonics and processes, configuration result in stress fields that c) the physics of the Earth permits the may be extensional in some intraplate survival of inhomogeneities in the upper regions. The history of continental mantle and continental lithosphere that
11 correspond to the ages of the recycled production can reach the scale of a large components found in "hot spot" and LIP igneous province. lavas, d) the chronology of volcanism reflects the Loss of the lower lithosphere has been migration of the locus of extension and the inferred from increases in crustal heat flow lateral variation in fertility of the source and seismic tomography in specific tapped. regions, from rapid regional uplift, and e) LIPs are not necessarily associated with from the appearance of signature high- time-progressive volcanic chains, and vice potassium magmas. A dense lower- versa, lithospheric region may develop through f) seismic tomography images reflect melt injection and transformation into lateral compositional variations and do not eclogitic phase assemblages, or through indicate temperature variations alone, thickening and cooling of a lithospheric g) "hot spot" lavas do not come from root, or through accumulation of mafic sources that are very hot compared with phases in magma chambers. Lower crustal the regional, and mantle compositions that result from h) geochemical observations be reconciled arc magmatism are likely to exceed the with fertile sources at relatively normal mantle density by 50 to 250 kg/m3, which temperatures, corresponds to about 1 to 5% density i) high 3He/4He ratios do not require a contrast. Given low enough viscosity, a lower-mantle source, density contrast as small as 1% is j) components required to come from the sufficient to drive a gravitational Earth’s core do not occur in "hot spot" and Rayleigh-Taylor instability. LIP lavas. When the Rayleigh-Taylor instability beings to sink it can pull the lithosphere Continental Magmatism Caused by downward through viscous traction as Lithospheric Rayleigh-Taylor asthenosphere flows inward along the base Instabilities of the lithosphere and down along the sides of the instability. In this phase Linda T. Elkins-Tanton (Dept. of asthenosphere may melt adiabatically as it Geological Sciences, Brown Univ. 324 rises into the space previously occupied by Brook St., Providence RI, USA; Tel: +1- the unstable lithospheric material. If the 401-863-3185; E-mail: instability contains volatiles, it may lose [email protected]) them as it sinks to higher pressures and acquires higher temperatures through Ductile removal of the lithosphere via conductive heating, much as a subducting gravitational Rayleigh-Taylor instability, slab does. Volatile content may this often referred to as delamination, can trigger melting in the surrounding mantle produce continental magmatism over a asthenosphere, or in the instability itself. range of eruptive volumes, major and trace Lithospheric instabilities may be an element compositions, volatile contents, important process for recycling crustal and accompanying topographic material and volatiles into the mantle, expressions. We investigate the process of second only to arcs on Earth, and gravitational instability using numerical potentially the most important lithospheric experiments, and find that the process can process on other terrestrial planets for produce a heterogeneous, locally hydrous recycling. Loss of the lower lithosphere, upper mantle, as well as short-duration possibly including crustal material, has eruptive episodes of hydrous, alkali-rich been suggested by other researchers as the magmas in the absence of subduction, trigger for magmatism in many terrestrial subsidence during eruption, and shallow, settings, including the southern Sierra dry melting under cratonic lithosphere. Nevada, the Tibetan Plateau, the Within certain ranges of lithospheric Altiplano/Puna region, the Leucite Hills in thicknesses, instability sizes, and mantle Wyoming, Central Italy, eastern Anatolia, potential temperatures, magmatic and the Dabie Shan. We further apply this
12 model to the Siberian flood basalts, to flexure. These cracks provide conduits for explain both the initial subsidence during the generation of volcanic ridges. the first kilometer of eruption and the volatile content of the final lavas, though we also present results from experimental Mantle Convection - An Overview petrology that the final magmas were formed from mantle material hotter than J. H. Davies (School of Earth, Ocean and the range commonly cited for oceanic Planetary Sciences, Cardiff Univ., Main mantle. Building, Park Place, Cardiff, CF10 3YE, Wales, UK; Tel: +44-29-2087-5182; Fax: +44-2087-4326; E-mail Cracks and Warps in the Lithosphere [email protected]) From Thermal Contraction I will present an overview of mantle D. T. Sandwell (Scripps Institution of convection, trying to emphasise aspects Oceanography, La Jolla, CA, 92093- that have a bearing on the plume debate. I 0225, USA ; Tel: +1-858-534-7109; Fax: will start by presenting the general physics +1-858-534-2902; E-mail: that controls mantle convection and [email protected]) mention some simplifying assumptions that are sometimes made. I will follow New marine gravity maps, based on that by briefly discussing the mantle’s retracked radar waveforms of ERS-1 and structure, and how the values of critical Geosat, show that linear volcanic ridges parameters are likely to vary with are widespread in the deep ocean basins, temperature and through the depth of the especially on seafloor that formed at mantle, and their uncertainties. higher spreading rate (full rate > 60 mm/yr). In the eastern Pacific, the ridges With this background we will go to are geometrically associated with the through a sequence of models, starting development of 150-km wavelength with the simplest model of mantle lineaments in the gravity field. Proposed convection to explain the concept of models for both the ridges and gravity thermal boundary layers, before advancing lineaments include small-scale convective through more complex examples that rolls, extension of the lithosphere, thermal show the importance of various properties. contraction, and asthenospheric return Some properties I hope to consider include flow in channels. Small-scale convection the effect of the geometry of models (2D v predicts ridges on the crests of the gravity 3D; spherical v Cartesian), the vigour of lineaments, which is in conflict with convection, surface velocity boundary observations. Plate tectonic models show conditions and the importance of plates, that extension of the Pacific plate is the mode of heating (internal v bottom), insufficient to produce gravity lineaments the introduction of chemical heterogeneity, by a boudinage mechanism although the the effect of phase transitions, morphology of the ridges seems to require compressibility, and temperature and extension. Cooling and contraction of the depth dependent viscosity, thermal lithosphere can explain both the origin of expansivity and thermal conductivity. the ridges and the gravity lineaments. Top-down cooling of the lithosphere We will briefly touch upon simple plume produces two modes of thermoelastic theory, plume heads and ‘non-lower stress - extension parallel to the ridge axis boundary layer’ convection means of and concave - down thermal bending producing ‘plumes’, e.g. lithosphere stress. The thermal bending stress is instabilities, edge convection, splash optimally released by lithospheric flexure plumes, slab breakoff etc. between regularly spaced parallel cracks. The pull of subducting slabs around the I will then consider what mantle properties Pacific plate may trigger the cracking might be required to prevent the formation which is amplified by thermoelastic of deep plumes. I will emphasise the
13 uncertainty in many parameter values, and within the Earth's mantle. The most that Earth is a Heat Engine outputting ~ prominent boundary layers are the 670km 44TW heat. There are two issues discontinuity and the Core-mantle regarding this - (i) the mantle needs to boundary at 2900 km depth. The fluid produce a large amount of heat internally dynamics of plumes, i.e. their (and we know that the MORB source spatiotemporal evolution and their region’s contribution is very low) (ii) how transport properties are only understood is this heat transported from the interior to under conditions which are oversimplified, the surface? The obvious means to as compared to the mantle. Laboratory prevent the generation of deep plumes is to experiments can hardly take into account have a very viscous lower mantle, with features which can critically influence the high thermal conductivity and low thermal formation and evolution of plumes. Mantle expansivity – current work suggests that convection is likely to be partially mantle properties have these trends. powered by internal radioactive heat Whether the likely values can prevent sources. Further, the viscosity of the plume formation though is unknown. If mantle material is known to strongly the lower mantle or core is a significant depend on temperature and pressure. Also source of the surface heat flow it is likely there is clear evidence for a decrease of that such a lower mantle could not the coefficient of thermal expansivity with transport heat sufficiently quickly by increasing pressure throughout the mantle. conduction or sluggish flow to (a) stay All those effects have an effect on the sufficiently cool to maintain its high generation and the evolution of plumes. viscosity etc, and prevent melting; and (b) By means of numerical experiments we to provide the heat flow observed at the investigate the plume evolution in surface. Therefore it is likely that a deep different mantle-relevant scenarios. It is layer that would not produce plumes demonstrated that plumes do not exist in would probably need to also have a low purely internally heated convection.as long level of heat generation and a cool core. It as constant material properties are would also be important that the assumed. A viscosity, increasing with subducting surface layer does not enter the pressure and/or a coefficient of thermal lower layer in any significant way, since expansion decreasing with pressure leads this is also likely to lead to the formation to a focusing of buoyancy into a few of upwellings. strong plumes. Such, even in internally heated systems, plume instabilities do We will conclude by emphasising that - (i) evolve. A strong temperature dependence the mantle is a heat engine that needs to of the viscosity leads to episodic plumes. transport a lot of heat, (ii) some of this Initially a massive plume heads develops flow must be by upwellings which in most and travels upwards. Subsequentla,y models to date assume a cylindrical form, pulses of hot material can rise through the (iii) but there is still a lot of uncertainty established low viscosity channel. Plumes regarding parameters, and limitations with evolving selfconsistenly from a thermal modelling (numerical and laboratory). boundary layer, do hardly entrain material during their ascent. Instead they transport mostly material from the boundary layer. Generation and Evolution of Plumes in Mantle-Relevant Scenarios
Ulrich Hansen (Institut fuer Geophysik, Westfische Willhelms-Univ. Muenster, 48165 Muenster, Germany; Tel: +49-251- 833-3592; E-mail: [email protected] muenster.de)
Plumes originate as convective instabilities from thermal boundary layers
14 On the Motion of Hawaii and Other the early history of the plume, and its rise Mantle Plumes through the mantle. Although the initiation age of the Hawaiian hotspot is unknown, J. A. Tarduno (Dept of Earth and rapid motion followed the dated initiation Environmental Sciences, Univ of of the New England and Tristan hotspots. Rochester, Rochester, NY 14627, USA; This lends some support to the idea that Tel: +1-585-275-5713; Fax:+1-585-244- hotspots can move rapidly while new 5689; E- mail: [email protected]); plume conduits adjust to the convecting P. V. Doubrovine; R. D. Cottrell mantle.
Paleomagnetic analyses conducted during Ocean Drilling Program (ODP) Leg 197 The Plume Head - Continental (Tarduno et al., 2003), studies of plate Lithosphere Interaction Using a circuits (e.g. Cande et al., 1995) and Tectonically Realistic Formulation for geodynamic modeling results (e.g. the Lithosphere Steinberger and O'Connell, 1998) have all pointed toward motion of the Hawaiian E. Burov (Univ. Paris VI, Case 129, 4 hotspot during formation of the Emperor Place Jussieu, Paris, France, 75252; Tel: Seamounts, between 81 and 47 Ma. The +33-144273859; Fax: +33-144275085; E- recognition of hotspot motion from mail: [email protected]); L. paleomagnetic data follows work that has Guillou-Frottier (Mineral Resources Dept., demonstrated that prior Pacific apparent BRGM, Orleans, France; Tel: +33- polar wander paths fail internal 238644791; Fax: +33-238643518; E-mail: consistency tests and that the physical [email protected]) processes derived from these paths, including rates of true polar wander, are Debates on the existence of mantle plumes untenable. The failure of prior Pacific largely originate from interpretations of APWPs can be traced to an over-reliance supposed signatures of plume-induced on remote sensing data (modeling of surface topography that are compared with seamount anomalies and marine magnetic predictions of hydrodynamic models of anomaly skewness), and an under- plume-lithosphere interactions. Yet, these appreciation of the inherent limitations of models are not really well suited for such data. Recent attempts to prediction of surface evolution: in general, systematically change the age of hotspots, they assume a fixed upper surface and a their positions and the data used to single layer viscous lithosphere. In nature, compare the Pacific and Indo-Atlantic the surface evolution is conditioned by realms (Gordon, 2005) are ad hoc elastic - brittle - ductile deformation, by extensions of the fixed hotspot model. The free upper surface and by layered structure relevant questions today include the scale of the lithosphere. We attempt to reconcile of mantle flow recorded by hotspot track mantle- and tectonic-scale studies by segments, the underlying processes that introducing a tectonically realistic cause track segments to be dominated by continental plate model in large-scale mantle flow rather than plate motion for plume-lithosphere interaction. The model given time intervals, and what these includes (1) a free surface boundary factors tell us about the nature of plumes. condition, (2) an explicit elastic-viscous A first-order consistency between the (ductile)-plastic (brittle) rheology and (3) a Hawaiian-Emperor and Louisville tracks stratified structure of continental argues for a Pacific-basin wide component lithosphere. The numerical experiments of mantle flow. However the mismatch demonstrate important differences from between these two tracks and central predictions of conventional models. In Pacific hotspot tracks indicates that particular, this relates to plate bending, smaller scale shallow processes must also mechanical decoupling of crustal and be at work. Some efforts to model the mantle layers (followed by lateral crustal observational constraints indicating flow) and tension-compression motion of the Hawaiian hotspot highlight instabilities, which produce transient
15 topographic signatures such as uplift and discontinuity. Edge driven convection subsidence at large (> 500 km) and small does not explain long linear island chains horizontal scale (300-400 km, 200-300 or oceanic hotspots thousands of km and 50-100 km). The mantle plumes kilometers from a continent. do not necessarily produce detectable long-wavelength topographic highs but often generate alternating smaller-scale Formation of Volcanic Rifted Margins: surface features that could be otherwise Influence of the Pre-Rift Lithosphere attributed to regional tectonics. A single Architecture long-wavelength domal uplift, predicted by conventional models, develops only for Jolante van Wijk (IGPP, Scripps a very cold and thick lithospheres. Distinct Institution of Oceanography, UCSD, La topographic wavelengths or temporarily Jolla, CA 92093-0225, USA; Tel: +1-858- spaced events observed in the East African 534- 5101; Fax: +1-858-534-5332; E-mail: Rift system, as well as over French Massif [email protected]) Central, can be explained by a single plume impinging at the base of the It is known that continental rifting tends to continental lithosphere, without evoking follow weak zones in the lithosphere, such complex asthenospheric upwelling. as orogenic belts, while stronger cratonic regions are usually not significantly deformed. For example, in the northern How Many Hotspots Can be Explained North Atlantic, rifting occurred along the by Edge Driven Convection? Caledonian belt, and in eastern Africa, the East African Rift System is located within S. D. King (Dept. of Earth and a Late Proterozoic belt. We think that such Atmospheric Sciences, 550 Stadium Mall a pre-structured lithopshere not only Drive, Purdue Univ., West Lafayette, IN influences the location of rifting, but that 47907-2051, USA; Tel:+1-765-494-3696; the inherited lithosphere architecture plays Fax:+1-765-496-1210; E-mail: a role in the rifting process itself. To study [email protected]) the role of a pre-structured lithosphere in rifting and (volcanic) passive margin Fluid near a vertical wall of uniform formation we use a numerical modeling temperature is dynamically unstable and approach. Our models are designed to this geometry will drive small-scale simulate Atlantic-type passive margin convection near the wall. This is the key formation. element of the edge driven convection mechanism. Continent-ocean and craton Results from numerical experiments boundaries are ideal locations for this form suggest that the old structures not only of small scale convection. Because the influence the location of rifting, but also 660-km discontinuity is an effective lithosphere deformation during rifting, the barrier to short-wavelength flow; and thermal evolution of the lithosphere, and because convective instabilities form thus decompressional melting. The nearly unit aspect ratio cells, hotspots amount, timing and distribution of within 600-1,000 km of a continent-ocean decompressional melting are influenced by or craton boundary are prime candidates factors such as pre-rift crustal or for edge driven convection. The lifetime of lithosphere thickness, orientation of edge driven instabilities depends on the inherited structures, and lithosphere stability of the cratonic root. The most composition. The modeling results suggest unique diagnostic for edge driven that a mantle plume is not a pre-requisite convection is imaging the downwelling to form a volcanic rifted margin; dynamic limb of the edge driven convection cell. processes related to lithosphere extension Because this is an upper mantle instability can explain the sometimes enigmatic edge-driven convection would more amounts of melt observed at Atlantic-type strongly impact the 410-km discontinuity passive margins. and may have little impact on the 660-km
16 Hawaii Volcano Chain as a nearly circular oceanic plate with Thermoelastically-driven Propagating spreading at one edge and subduction at Crack the opposite edge will have a stress field favorable to tension cracks located near W. D. Stuart (U.S. Geological Survey, midplate and oriented in the plate velocity Menlo Park CA 94025, USA; Tel: +1- direction. 650-329-4648; E-mail: [email protected]); G R Foulger (Dept. Earth Sciences, Science Laboratories, Univ. of Durham, Sunday Poster Session Durham DH1 3LE, U.K.; E-mail: [email protected]); M Barall Testing Thermal Whole Mantle Plumes (U.S. Geological Survey, Menlo Park CA Seismically 94025, USA, E-mail: [email protected]) S. Goes (Dept. Earth Science and Lithosphere crack models for the Hawaii- Engineering, Imperial College London, Emperor volcano chain require a U.K.; E-mail: [email protected]); F. horizontal extensional stress normal to the Cammarano (Berkeley Seismological volcano chain for the crack tip (Hawaii) to Laboratory, UC Berkeley, CA, USA; E- propagate to the southeast. We calculate mail: [email protected]); A. the 3D thermoelastic stress rate field for a Deuss (Institute of Theoretical model Pacific plate lithosphere and find Geophysics, Univ. of Cambridge, U.K.; E- that horizontal tensional stress occurs mail: [email protected]) normal to the entire Hawaii chain, but not beyond Hawaii to the southeast. In the To prove or disprove the existence of model, the actual Pacific plate geometry is plumes, clearly formulated hypotheses that represented by a fragment of a spherical can be tested against data are required. We elastic shell with simplified lateral set up two seismic tests for the simplest boundaries. The stress field is a hypothesis of thermal, whole mantle consequence of cooling of Pacific plate plumes. If a thermal boundary layer exists material accreted at the East Pacific Rise at the base of the mantle and is not and its precedents, and the current 3D completely stabilized chemically, such temperature field is inferred from the upwellings should form. seafloor age (Mueller et el., JGR, 1997). Our first test checks whether the 1D The general horizontal stress pattern structure of single layer thermally across the plate from the East Pacific Rise convecting mantle (pyrolitic with phase spreading boundary (southeast) to the transitions, following a mantle adiabat Aleutian-Japan subduction boundary between 300 and 2500 km depth) is (northwest) is tension near the rise, then compatible with global travel-time and compression to midplate, and extension mode-frequency data. We find that, in throughout the rest. Strong extension also spite of the large uncertainties in the occurs near Samoa where the plate elastic and anelastic mineral physics data boundary is a stress-concentrating notch. required to convert physical into seismic The thermoelastic model implies that the structure, the single-layer convection plate stress field and a feature dependent structure is difficult to reconcile with the on it are approximately stationary with seismic data. The seismic data prefer a respect to the plate boundary reference slower transition zone and a slower deep frame. Thus, if part of the Hawaii section mantle. This could be an indication of of the Hawaii-Emperor chain is equivalent compositional heterogeneity in the deep to a vertical tensile crack, the southeast mantle, which would also affect plume crack tip holds a fixed position with generation. respect to the plate boundary. As the Pacific plate moves northwest through the For the second test, we made a set of stress field, the crack extends an equal dynamically self-consistent thermal plume distance to the southeast. In short, a models with parameters that are Earthlike
17 and compatible with observations. The analyzed global paleomagnetic data and temperature contrast of the basal boundary found no significant relative motions layer is chosen to yield sublithospheric betweenthe Pacific group of hotspots and plume temperature anomalies between 100 the African group of hotspots detected. We and 300K. The resulting plumes are weak revised the model of True Polar Wander and slow to form, yet they all have Path and suggest that the southward buoyancy fluxes larger than the the paleolatitude shift of Hawaiian hotspot strongest hotspots. Their seismic anomaly (Tarduno et al., 2003) is not caused by the amplitudes (strong in upper mantle, but plume drift but by the True Polar Wander. only 1-1.5% in Vs in lower mantle) and widths (500-800 km wide in the lower mantle) are compatible with surface and Plume-fed Asthenosphere Flow Model: seismic observations. Very strongly Evidence from Isotopic Variation Along temperature-, and depth-dependent Mid-ocean Ridges viscosity yields the narrowest and lowest buoyancy flux plumes, but induces M. Yamamoto (Earth and Atmospheric sublithospheric small-scale convection Sciences, Cornell Univ. Ithaca, NY 14853, which may complicate the plume’s surface USA; Tel: +1-607-255-1459; Fax: +1-607- signature. 254-4780; E-mail: [email protected]); J Phipps Morgan (Earth and Atmospheric Sciences, Cornell Univ, Ithaca, NY 14853, How Geometry and Ages of Global USA; Tel: +1-607-255-1631; Fax: +1-607- Hotspots are Explained by Classical 254-4780; E-mail: [email protected]); W Hypotheses of Rigid Plate and Fixed J Morgan (Earth and Planetary Sciences, Hotspot Harvard Univ., Cambridge, MA, 02138, USA; Tel: +1-617-496-8283; Fax: +1-617- Y. Harada (School of Marine Science and 495-8839; E-mail: Technology, Tokai Univ., Orido Shimizu [email protected]) Shizuoka, 424-8610 Japan; Tel: +81-543- 34-0411; Fax: +81-543-34-9840; E-mail: Asthenosphere plume-to-ridge flow has [email protected]); P Wessel often been proposed to explain both the (Dept. of Geology & Geophysics, SOEST, existence of geochemical anomalies at the Univ. of Hawaii at Manoa, HI 96822, mid-ocean ridge segments nearest an off- USA; Tel: +1-808-956-4774; Fax: +1-808- axis hotspot, and the existence of apparent 956-5154; E-mail: [email protected]) geochemical ‘provinces’ within the global mid-ocean spreading system. We have Since J. Tuzo Wilson and W. Jason constructed a thin-spherical-shell finite Morgan's hypothesis of moving rigid plate element model to explore the possible over fixed hotspots, quite a few geological structure of global asthenosphere flow and and geophysical observations were well to determine whether plume-fed explained by the theory for more than asthenosphere flow is compatible with thirty years. But over the last decade, there present-day geochemical and seismic are increasing arguments on discrepancies observations. In this model, lubrication between the classical theory and theory approximations are used to solve observations, and those are, they say, for the flow profile in the vertical caused by intra-plate deformations and/or direction, and a ~100-km-scale mesh is plume drift. Here we would like to point used to solve for the mean horizontal out that those discrepancies will disappear asthenosphere flow. Asthenosphere is for the most part if we apply accurate assumed to be brought up by mantle models of absolute plate motions. We plumes, with ‘sinks’ of asthenosphere at show new technique of constructing spreading centers where compositional models of absolute plate motions, and how lithosphere is made, at trenches (where geometrical positions and radiometric ages some, but not much asthenosphere is of global hotspots are consistent with the entrained and dragged down by two conventional assumptions. Also we subducting lithosphere), and also a
18 distributed sink of asthenosphere due to its transition zone beneath eastern Chinese cooling and attachment to the base of the continent as detected by seismic aging and thickening oceanic lithosphere. tomographic models [2]. The Mesozoic We also assume that the strength of all the volcanism may be genetically associated plume (hotspot) asthenosphere sources is with the lithospheric thinning because the equal to the sum of all the asthenosphere basaltic source is ancient isotopically sinks, i.e. that the asthenosphere has a enriched [3] lithosphere - being converted present-day steady-state thickness and to the asthenosphere. The NNE-SSW hotspot fluxes have remained constant Great Gradient Line (GGL) marked by the through time. In spite of these evident sharp altitude, gravity anomaly, crustal oversimplifications, the model appears to thickness, and mantle seismic velocity show considerable promise as a possible changes from the plateau in the west to the mechanism to explain observed patterns of hilly plains of eastern China is an MOR geochemical segmentation. This expression of variation in lithospheric specific prediction lends support to the thickness from probably > 150-200 km plume theory and plume-fed thick beneath the plateaus in the west to asthenosphere model proposed by Phipps the thin, probably < 80 km thick, beneath Morgan et al., JGR, 1995. eastern China. The “remote” western Pacific subduction systems (“wedge suction”[1]) induce asthenospheric flow Slab Dehydration, Subcontinental from beneath eastern China towards the Lithosphere Thinning and Widespread subduction zones, which, in turn, requires Mesozoic/Cenozoic Volcanism in asthenospheric material replenishment Eastern China: A Special Consequence from beneath the western plateaus to of Plate Tectonics eastern China. As a result, such eastward asthenospheric flow experiences upwelling Yaoling Niu (Dept. of Earth Sciences, and decompression (from beneath Durham Univ., Durham DH1 3LE; Tel: thickened to thinned lithosphere), which +44-191-334-2311; Fax: +44-191-334- causes the flowing asthenosphere 2301; E-mail: (isotopically depleted) [3] to melt and [email protected]); S. Song & L. produce Cenozoic eastern China basaltic Zhang, Peking Univ., China; J. Deng, X. volcanism. Such volcanism may have Mo, China Univ. of Geosciences, Beijing, actually begun in the late Cretaceous [4]. China; S. Su, China Univ. of Geosciences, Our proposed mechanism of lithosphere Beijing, China; Z. Guo, Chinese Academy thinning (1) does not require hot mantle of Sciences, China; J. Liu, Chinese plumes that may have not existed beneath Academy of Sciences, China eastern China; the horizontally-lain transition-zone slabs act as a cold thermal Subcontinental lithosphere “delamination” boundary layer that sucks heat from above and the widespread Mesozoic/Cenozoic and below, thus preventing hot mantle intra-plate basaltic volcanism in eastern plumes to rise from the lower mantle and China have been widely speculated as to form in the upper mantle; (2) does not resulting from ascending hot mantle require lithospheric “delamination”, which plumes or large scale lithospheric describes that deep portions of the buoyant extension. We argue that the cratonic lithosphere sink into the dense Mesozoic/Cenozoic Earth events in asthenosphere – a scenario that is eastern China is a special consequence of physically difficult; (3) does not require plate tectonics [1]. The Mesozoic lithospheric extension/stretching whose lithosphere thinning in eastern China existence and scale in the Mesozoic results from “transformation” of the basal remain elusive; (4) explains the portion of the lithosphere into lithosphere thinning in the entire eastern asthenosphere by hydration. The water China, not just the North China Craton that did so came from dehydration of (NCC); and thus (5) questions the subducted Pacific (or predecessor) oceanic significance of South China continental lithosphere that is lying horizontally in the subduction as a cause of lithosphere
19 thinning beneath the NCC. Our suggested Oceanic Hotspots, Springer-Verlag, New mechanisms for the Mesozoic/Cenozoic York, p. 257-279, 2004. volcanism in eastern China are consistent with the geochemistry of the basalts [3], physical scenarios of mantle melting [1] Late Mesozoic Through Cenozoic and geophysical observations [1,2]. These Magmatism in East Asia: Effect of principles and observations (1) disfavor Directly and Obliquely Subducted Slab the suggestion of hot mantle plume origin Flexures of eastern China volcanism; (2) disfavor the suggestion of ocean ridge-like passive S. V. Rasskazov (Institute of the Earth’s mantle upwelling and decompression Crust, Irkutsk, Russia 664033; Tel: + melting; and (3) argue that the eastern 7(3952)51-16-59; Fax: +7(3952)42-69-00; China Mesozoic/Cenozoic basins cannot E-mail: [email protected]); H. Taniguchi be used as evidence for continental (Center for Northeast Asian Studies, extension; the basins may very well be an Tohoku Univ., Kawauchi, Aoba-ku, isostatic response [1] to the horizontally- Sendai 980-8576, Japan; Fax:+ 81-22-217- lain dense slab materials in the transition 6236; E-mail: zone. It is also important to note: (1) the [email protected]) GGL is likely a young feature as a result of Indo-Asian collision since the early Mesozoic through Cenozoic evolution of Tertiary; (2) subduction-zone dehydration East Asia was related traditionally to plate is necessarily incomplete [1,5] (lawsonite, motions in Pacific Ocean and marginal which can contain ~ 11 wt % water, is seas. After Molnar and Tapponnier‘s stable up to 11 GPa [6]; serpentines within rejuvenation of the hypothesis by Argand the oceanic lithospheric mantle [5] on influence of the India-Asian collision contains up to 13 wt % water, and is stable on seismic activity and deformations in up to 7 GPa [7] before transformed to Inner Asia, numerous multi-disciplinary dense hydrous magnesium silicate phases studies supported this point. The at much greater pressures of ~ 5 to 50 GPa collisional hypothesis became dominating [6,8]), thus allowing slab water release at in explanations of tectonics not only in great depths at elevated temperatures (e.g., Inner Asia but also at the Eurasia-Pacific isobaric heating of the slabs in the convergent zone. In the early 1990-th, the transition zone) [1,5]; (3) mantle wedge plume hypothesis was applied for suction [1], while less strong than ridge explanation of volcanic activity in some suction [9], is an important driving force areas of East and Central Asia. It was for asthenospheric flow; and (4) the speculated also that diffused intraplate suggestion of more recent lithosphere Cenozoic volcanism was produced by a accretion beneath eastern China is in fact a broad asthenospheric upwelling from a straightforward consequence of conductive deeper level (“hot region”) (Tatsumi et al., cooling of the asthenospheric mantle. 1990) or by convective flow ascending References: [1] Niu, Y., Geol. J. China from the core-mantle boundary Univ., 11, 9-46, 2005; [2] Kárason, H. & (Zonenshain et al., 1991). Studies of the R. van der Hilst, Geophys. Monogr., 121, last decade showed that Cenozoic rifting 277-288, 2000; [3] Xu, Y.-G., Phys. and magmatism in this area was due to Chem. Earth (A), 26, 747-757, 2001; [4] combined effects of (1) generation of low- Yan, J., J.-F. Chen, Z. Xie & T.-X. Zhou, velocity sub-lithospheric mantle, (2) Chinese Sci. Bull., 8, 1570-1574, 2003; collisional events at plate boundaries, and [5] Niu, Y., J. Petrol., 45, 2423-2458, (3) subduction of oceanic plates beneath a 2004; [6] Williams, Q. & R.J. Hemley, continent (Rasskazov et al., 1998). A high- An. Rev. Earth Planet. Sci., 29, 365–418, velocity anomaly - “the stagnated slab” - 2001; [7] Ulmer, P. & V. Trommsdorff, at the mantle transitional zone (Fukao et Science, 268, 858-861, 1995; [8] Frost, al., 1992) and the low-velocity Transbaikal D.J., Geochem. Soc. Spec. Publ., 6, 283- mantle domain at depth of 200-350 km 297, 1999; [9] Niu, Y. & R. Hékinian, In (Rasskazov et al., 2003) were well recognized thomographically. The
20 anomalous mantle region was suggested to Gondwanan and Rodinian fractures and be a time-integrated expression of associated large igneous provinces define subduction processes (Rasskazov et al., geometrically regular, energy-minimizing 2004). The aim of this presentation is to arrays of interspersed hexagons and show spatial-temporal regularities of the pentagons. The arrays fit the spherical Late Mesozoic through Cenozoic projection of the truncated icosahedron, magmatism evolution which is indicative for which each edge-length subtends 23 for formation of the high-velocity and degrees of great-circle arc. Lithospheric low-velocity material. The latter could be fractures and giant dike swarms follow the stored firstly during closing of the edges of the tessellations, while large Mongolia-Okhotsk Ocean finalized at ca. igneous provinces and intracratonic basins 140 Ma. After terrane accretion and occupy the vertices. The tessellated structural reorganization at 113-107 Ma, distribution of these large igneous subduction of the Kula-Izanagi plate provinces abjures impingement of defined the northern margin of the randomly-dispersed mantle plumes for anomalous mantle region. Low-velocity their origins. anomalies extended from a continental margin landward over 1000 km beneath Analogous fracture tessellations the Aldan shield of the Siberian craton. spontaneously propagate in a thin, The structural reorganization between 65 spherical shell under uniform layer- and 50 Ma took place contemporaneously parallel tension. A master fracture initiates with accretion of the Okhotsk Sea plate to at a weak point and zig-zags across the Eurasia. Block rotations and extension at shell, spontaneously bending at the continental margin were accompanied approximately 120 degrees as it by formation of the oblique Sikhote Alin propagates. Secondary fractures branch slab flexure of the Pacific plate. outward from the bends to establish triple Afterwards, the slab flexure was widening junctions. The secondary fractures then to the south due to landward growing of zig-zag and branch as they propagate, the directly subducted Honshu-Khingan eventually forming a tessellated network slab fragment. The latter resulted in of fractures and tiles. development of the southern margin of the anomalous mantle region. The structural The Gondwanan and Rodinian fracture reorganization between 21 and 15 Ma was tessellations propagated after formation coeval to accretion of the Philippine Sea and stagnation of supercontinents. plate to Eurasia with formation of the Thermal expansion of insulated Japan-Korea oblique slab flexure, trench asthenosphere beneath the supercontinents rolling-back effect, block rotations, and may have generated homogenous, layer- extension at the continental margin. The parallel tension within the lithosphere. present-day subduction activity of the Separation of the tiles drove Pacific slab is focused at the oblique decompression melting and opened Japan-Korea and direct Hokkaido-Amur avenues for eruption of basalt from the flexures. asthenosphere. Large igneous provinces preferentially erupted at triple junctions because of greater local pressure relief. Fracture Propagation on a Sphere: These diachronous eruptions were Implications for Non-Plume Origin of secondary and depended on the ability of Large Igneous Provinces on the tiles to separate, controlled by plate Fragmenting Supercontinents boundary conditions.
J. W. Sears (Univ. of Montana, Missoula, MT 59812, USA; Tel: +1-406-243-2341; Fax: +1-406-243-4028; E-mail: [email protected])
21 Neoproterozoic Anorogenic Magmatism caused cracking and disintegration of Associated with Rodinia Breakup: Not a Rodinia supercontinent at 750 Ma. Result of Mantle Superplume However deep-mantle plume paradigm is seriously questioned and several alternate Kamal K. Sharma (Dept. of Geology, models are proposed to explain large scale Government College, Sirohi 307 001, magmatism and continental fragmentation. Rajasthan, India; E-mail: Geodynamicists assert that the [email protected]); Gillian R. endothermic phase boundary at 660-km Foulger (Dept. Earth Sciences, Science depth isolates lower and upper mantle Laboratories, Univ. of Durham, South Rd., convection, and that propagation of Durham, DH1 3LE, U.K.; E-mail: lithospheric cracks triggers anorogenic [email protected]) magmatism through decompression melting of ordinary asthenosphere. The splitting of Rodinia supercontinent at Similarly, elucidates a top-down tectonic 750 Ma led to the development of new viewpoint, in which a stagnant ocean floor, and cratonic fragmentation, supercontinent insulates the underlying which ultimately resulted intraplate mantle, leading to thermal expansion, anorogenic rift magmatism on the partial melting resulting anorogenic northwestern Indian shield, Madagascar, magmatism along propagating rift zones. Seychelles, Australia, central and southern The amalgamation of Rodinia continent Appalachians and other landmasses. marks the end of orogenic accretion processes during Neoproterozoic time on The anorogenic bimodal Malani volcano- the earth crust. The crust became thick and plutonic province (750 Ma) is spread over remained insulated for prolonged time. approximately 50,000 km2 in western This caused increase of thermal gradient Rajasthan, India. Besides this, equivalent and lithospheric extension and Rodinia to Malani is reported from Sind Province breakup, which ultimately led to silicic of Pakistan, Kutch, Madagascar, south magma generation from the melting of China and Seychelles. The Neoproterozoic crustal region. The Rodinia breakup Malani anorogenic magmatic province of resulted Neoproterozoic anorogenic northwestern India is an example of Silicic magmatism on most of the landmasses. Large Igneous Province (SLIP) development at a time of Rodinia break- up. The Malani magmatism is generally of Are the Precambrian Magma terrestrial in origin. The volcanism Chambers Responsible for the resulted minor initial basalt flows, Cretaceous Deccan Trap Volcanism of ignimbrite eruptions, rhyolite flows and India? ash fall eruptions through multiple fissure/rift systems that developed in the Anju Tiwary (National Centre For intraplate extensional tectonic setting. This Antarctic and Ocean Research, Headland was followed by A-type granite plutonism Sada, Vasco-da-Gama, Goa-403804, and terminal felsic/silicic dykes. India; Tel: +0091-(0)832-2520876; Fax: 0091-(0)832-2520877; E-mail: The ultimate causes and controls of [email protected]); Dhananjai Pandey and continental rifting and associated S. Rajan anorogenic magmatism remain controversial questions amongst the earth We propose a genetic relationship between scientists. It is postulated that wide spread 65-68 Ma old Deccan volcanic province magmatism and continental breakup is a (DVP), extensional tectonism and several result of impingement of hypothetical deep episodes of Precambrian magmatism. The mantle super plumes beneath Rodinia. It is model is compatible with regional hypothesized that abnormally hot mantle tectonics and provides a plausible region, having lateral dimensions of 6000- explanation of the spatial coincidence of 10000 km on the surface, originated from the DVP with Precambrian West Coast 2900 km deep near core-mantle boundary and Son-Narmada rift zones. Plume theory
22 and other alternative theories proposed for Seychelles resulted in enhancing the DVP so far ignore two factors (1) temperature for the magma stored in Reasons behind the selective spatial chambers due to factors mentioned in distribution of DVP along the West Coast point 6; 11. Stored volume of magma and Son-Narmada rifts (2) Effect of melted and due to availability of plumbing geological past and the intra-continental ways through the West Coast fault erupted stresses of the peninsular India on the mostly along the west coast of India and DVP. Our theory takes into account these partly flowed towards another adjacent factors and shows that it is the weak zone of Son-Narmada rift 12. The geodynamical evolution of the host and huge volume of this Deccan volcanism is the resultant stress vectors, which paved due to large amount of magma stored in the way for the genesis of large igneous the chambers which were located beneath province of DVP. India-Seychelles rift zone.
The proposed concept involves two stages of geological evolution of the Indian Large LIPs and the Mantle Squeeze: A peninsula which led to the genesis of Mass Balance Approach to Hotspots DVP. The brief of postulated activities in each stage is given below. E. K. Beutel (Dept. of Geology and Environmental Sciences, College of Stage 1: 1. Due to intra-continental Charleston, 66 George St Charleston, SC stresses, continental stretching along the 29424, USA; Tel: +1-843-953-5591; West Coast deep crustal fault in peninsular Fax:+1-843-953-5446; E-mail: India; 2. Decompression melting and [email protected]) magma generation in the upper lithospheric mantle below the West Coast I propose a new simplified thermodynamic fault; 3. Magma eruption where the mantle model coupled with tectonic plumbing system was available and reconstructions that provides a possible Magma ponding where plumbing system non-plume alternative to the predicted was not available; 4. This phenomena large scale doming, flood volcanism continued for almost 350 Ma i.e. 460 Ma preceded by uplift, and slightly hotter to 90 Ma. mantle temperatures of the plume hypothesis. The model indicates that Stage 2: 5. Progressive continental mantle temperature rise and large volumes stretching for almost 350 Ma resulted in of melt can be produced as a result of continental scale rifting around 90 Ma overpressuring mantle material trapped when Madagascar separated from India; 6. between the 440 km transition zone, The magma, which was solidified in the ringed subduction zones, capping chambers, heated up due to two factors (a) continents, and/or advancing cratonic roots decompress ional melting; and (b) behind regressing slabs. The amount of lowering of the solidus temperature by the melt, the degree of doming prior to magma addition of continental crust impurities to emplacement, and the relationship to the system; 7. Temperature of around of tectonic structures are determined by the 800-900oC was reached in those magma following variables; the rate and volume chambers which were affected by the of mantle overpressuring, external tectonic breakup due to their location; 8. Rhyolitic forcing mechanisms (slab-roll back, magma was produced by fractional number and magnitude of plate driving melting and felsic volcanism took place in forces), thickness of the lithosphere, and India and Madagascar; 9. From 90 to 65 the degree of penetration of the 440/660 Ma, continental scale rifting between transition zones by subducting slabs. India and Seychelles progressed and corresponding advancement in magma While this is an early model and may not melting resulted in chambers located apply to all instances of large igneous beneath the affected region; 10. At 65 Ma, province (LIP) emplacement, it explains continental scale rifting between India- much of the variation observed. For
23 example, the Afar 'hotspot' may be phase transitions. The models have depth explained by a large volume of mantle dependent viscosity; with a viscous material compressed between the slab as it lithosphere and a lower mantle rolls back and the cold, deep cratonic root approximately 40 times more viscous than of Africa. As the mantle is squeezed it the upper mantle. The plate motion may warm and rise creating a large dome history of the past 120Myrs is applied as a at the thinnest point in the lithosphere, the surface velocity boundary condition. African-Arabian suture. The same collision that is overpressuring the mantle I find passive localised and shallow also exerts pressure on the continent, upwellings are generated beneath the causing longitudinal splitting along the spreading ridges. The planform of weak African-Arabian suture and the downwellings is linear at the surface and African plate. In this case, doming become cylindrical as they descend into precedes rifting and continental rifting the mantle. Similarly upwellings at the merely enhances the decompression base of the mantle start linearly but melting of the overpressured mantle. become more cylindrical rising from the However, in other cases (e.g. Central junctions of the linear features. In the Atlantic Magmatic Province) rifting may simulations described here these features precede uplift and doming depending on are rare and weak. the forces acting on the plates, including the pressure of the mantle on the cratonic There are also upwellings that are not roots and/or subducting slabs. This model, rooted at the base of the mantle.They along with the multitude of other non- develop as follows. Regions which have plume models, suggests that plumes may not suffered recent subduction become be the exception rather than the rule when hotter than average and tend to form sheets it comes to generating voluminous melts. that rise passively and slowly. Downwellings from the surface fall onto the sheets and make them into bowls as Mantle Convection - Plumes Rooted in the hot material is forced up around the Mid-Mantle sides. The rims of the bowls can become unstable producing cylindrical upwellings J. H. Davies (School of Earth, Ocean and (plumes). Since they look a bit like water Planetary Sciences, Cardiff Univ., Main droplet splashes, I have abbreviated Building, Park Place, Cardiff, CF10 3YE, 'plumes not rooted in thermal boundary Wales, UK; Tel:+44-29-2087-5182; layers' as 'splash plumes'. The splash Fax:+44-2087-4326, E-mail: plumes originate at a range of depths. In [email protected]) fact the downwellings can push the sheets all the way to the core mantle boundary in I have discovered a new class of thermal certain cases where it is then difficult to upwellings in mantle convection tell splash plumes from 'traditional simulations. They are not rooted in plumes'. thermal boundary layers, but at varying depths in the mantle. The best test for splash plumes will be seismic imaging. Their thin plumes and These are high resolution, Earth-like narrow bowls very near fast downwellings vigour thermal convection calculations in will require high resolution. These three-dimensional spherical geometry. 'plumes' also have implications for fixity, The models are for compressible temperature contrast, and lifespan which convection, with decreasing coefficient of can all be tested. thermal expansion with depth, as observed in experiments. The models have ~ chondritic rates of internal heating; and varying bottom heating that straddle estimated Earth values. Some of the models have included transition zone
24 The Many Potential Faces of Buoyant is they include the subduction of Mantle Upwellings: Diversity Within compositionally laminated slabs that sink the Plume Family though and ultimately stall and spread within a basal (hot) TBL. Experiments A. Harris and C. Kincaid (Graduate involve either three (2 slab laminates & School of Oceanography, Univ. of Rhode ambient) or four (2 slab laminates, upper Island, Narragansett, RI 02874, USA; Tel: vs. lower ambient fluid layers) distinct +1-401-874-6571; Fax: +1-401-874-6811; chemical reservoirs. Important variables E-mail: [email protected]) include the isothermal density and viscosity contrasts between the individual A focal point of the great plume debate slab components and the ambient fluids. seems to be the large headed, small tailed Results show a wide variety in upwelling variety of buoyant mantle upwelling in morphologies evolving when such which the plume head is both hot and laminated plates are allowed to reside compositionally homogeneous. Such flow within a basal TBL. A range in morphologies occur naturally in 2-D and morphologies from large-headed, small 3-D computer and laboratory models of tailed upwellings to medium headed- convection within a temperature- medium tailed upwellings are recorded. dependent viscous fluid that is heated from Upwellings matching the description of below. They are also seen in experiments the plume in the great plume debate (big, where a hot, low density-viscosity fluid is hot, homogeneous heads) are rare. Instead injected into a cooler, denser, more the buoyant upwellings exhibit striking viscous ambient fluid. Another common levels of lateral and vertical structure in feature of viscous convection experiments both composition and temperature. For are predictions of significant interaction example, one very common feature is a between surface (cool) and basal (hot) two-faced upwelling, that is boundary layers. Slab-like downwellings morphologically symmetric about its tend to sink until they stall at a density centerline, but with very different interface, which also tend to be regions compositions/temperatures on either side that support thermal boundary layers of this dividing line. Results from these (TBLs). One of the more robust relatively simple viscous, thermal- observations for the mantle is that chemical convection experiments suggest lithospheric plates sink deeply into the that if such conditions exist in the mantle, mantle and that these plates are then we can expect a rich diversity in compositionally heterogeneous, with upwelling styles. The commonly relatively well-defined upper and lower depicted, plume that is thermally driven, layers or laminates. with a large hot head filled with uniform composition material may in fact be a rare We report on results of 3-D laboratory occurrence. Instead upwellings from such experiments of convection within a polluted TBLs may be more like viscous fluid that include, in a simplified snowflakes, where no two are expected to manner, aspects of what we think we be the same. (Warning: Application of know about subduction, and how these these model results to the mantle requires influence upwellings in a way that may be a leap of faith that, among other things, relevant to the great plume debate. The there is a deep mantle TBL, that mantle experiments utilize corn syrup for a Raleigh numbers are well above the working fluid and are designed such that critical value, that chemically the key dimensionless numbers governing heterogeneous slabs exist and make it into convection and transport fall within ranges mantle TBLs and that scaling in (or approach values) that have been geophysical fluid dynamics is valid.) suggested for the mantle in the geophysical literature. Thermal convection occurs because the fluid is heated from below and cooled from above. An important feature of these experiments
25 Thermal Structure Beneath Kamchatka Thermal Structure of the Cocos Slab and Plume to Arc Magmatism Beneath Southern Mexico and its Transition Relationship with the Arc Volcanism
Vlad Constantin Manea (Seismological Marina Manea (Seismological Laboratory Laboratory 252-21, CalTech, Pasadena, 252-21, CalTech, Pasadena, CA 91125, CA 91125, USA; Tel: +1-626-395-3382; USA; Tel: +1-626-395-3382; Fax: +1-626- Fax: +1-626-564-0715; E-mail: 564-0715; E-mail: [email protected]); Marina Manea [email protected]; Vlad Constantin (Seismological Laboratory 252-21, Manea (Seismological Laboratory 252-21, CalTech, Pasadena, CA 91125, USA; Tel: CalTech, Pasadena, CA 91125, USA; Tel: +1-626-395-3382; Fax:+1-626-564-0715; +1-626-395-3382; Fax: +1-626-564-0715; E-mail: [email protected]) E-mail: [email protected])
The Kamchatka subduction zone is one of Southern Mexico is a very interesting area the most active seismic and volcanic where the subducting Cocos slab regions in the world and located in the drastically changes its geometry: from a proximity of the Meiji Guyot mantle flat slab in Central Mexico to a ~ 45º dip plume. We propose a convection model angle beneath Chiapas. Also, the currently which shows the a hot blob rising from active volcanic arc, the modern depths grater that 1000 km would change Chiapanecan volcanic arc, is oblique and the initial spherical shape into a cylindrical situated far inland from the Middle shape deflected near surface by the Pacific America trench, where the slab depth is ~ plate movement. 200 km. In contrast, the Central America volcanic arc is parallel to the Middle Geochemical studies of volcanic rocks in America trench and the slab depth is ~ 100 Central Kamchatka show a complex km. A 2D steady state thermo-mechanical pattern, from basalts of intermediate model explains the calc-alkaline composition to alkaline basalts of plume volcanism by high temperature (~ 1300º type and adakites. Our models suggest that C) in the mantle wedge just beneath the the buoyant plume cannot penetrate the Central America volcanic arc and strong cold subducting slab in order to enrich the dehydration (~ 5 wt.%) of the Cocos slab. mantle wedge and to produce the alkaline In contrast, the thermal model for the plume type basalts. Instead, a gap in the modern Chiapanecan volcanic arc shows subduction process, likely created by high P-T conditions beneath the coast accretion of new terrains, would create an where the Miocene Chiapanecan extinct easy way for the hot plume material to arc is present, and is therefore unable to enrich the mantle wedge. offer a reasonable explanation for the origin of the modern Chiapanecan The contact between the hot plume and the volcanic arc. We propose a model in oceanic plate offshore Kamchatka which the origin of the modern produces a rejuvenation of the ~ 100 Ma Chiapanecan volcanic arc is related to the old Pacific plate, the thermal age being ~ space-time evolution of the Cocos slab in 40 Ma. 2D steady state thermal models Central Mexico. The initiation of flat with such hot incoming slab show that the subduction in Central Mexico in the oceanic crust beneath the active volcanic middle Miocene would have generated a arc has undergone melting and therefore hot mantle wedge inflow from NW to SE, adakitic volcanism. generating the new modern Chiapanecan volcanic arc. Because of the contact between the hot mantle wedge beneath Chiapas and the proximity of a newly formed cold flat slab, the previous hot mantle wedge in Chiapas became colder in time, finally leading to the extinction of the Miocene Chiapanecan volcanic arc.
26 The position and the distinct K-alkaline Massif (Czech Republic) where both local volcanism at El Chichón volcano are and global seismic tomographic studies proposed to be related to the arrival of the indicate the existence of localised zones of highly serpentinized Tehuantepec Ridge mantle upwelling extending to the base of beneath modern Chiapanecan volcanic arc. the upper mantle, several hundred km The deserpentinization of Tehuantepec across and up to 100-150 degrees Ridge would have released significant Centigrade hotter than ambient mantle. amounts of water into the overlying Short-wavelength uplift of the lithosphere, mantle, therefore favoring vigorous without associated volcanism, also occurs melting of the mantle wedge and probably further north in the UK, Scandinavia and of the slab. Brittany; this may also be a sign of mantle upwelling.
Small-scale Convective Instabilities in A fundamental question concerns the the Upper Mantle – A Generic Class of relationship between this distinctive Hotspots Linked to Recent Continental (short-wavelength) mode of mantle Collision in Europe and the Circum- convection and the development of the Mediterranean Region Alpine orogenic belt. Beneath Europe and the Mediterranean region the Transition M. Wilson (Institute of Geophysics & Zone is seismically fast [4] and may, Tectonics, Earth Sciences, School of Earth therefore, be cooler than the overlying & Environment; Leeds Univ., Leeds LS2 mantle and also compositionally distinct 9JT, UK; Tel: +44-113-343-5236; Fax: (containing, for example, a significant +44-113-343-5236; E-mail: component of subducted oceanic crust). It [email protected]); U Achauer has been argued that the Transition Zone is (EOST Strasbourg, UMR 7516, in fact a “slab graveyard” containing the Laboratoire de sismologie, 5, rue Rene remnants of subducted Tethyan (or older) Descartes, F-67084 Strasbourg, France; E- oceanic lithosphere [4]. If this is the case, mail: [email protected]) then the upwelling mantle diapirs are unlikely to be driven by either thermal or Many intra-plate hotspots (both compositional buoyancy. So what process continental and oceanic) appear to be or processes does drive the mantle associated with short-wavelength upwelling which triggers partial melting? convective instabilities (diapirs) within the upper mantle, originating from Transition We evaluate a range of mantle convection Zone (410-660 km) depths. We investigate models in the context of the regional plate the dynamics of such instabilities and, tectonic setting of Europe and constraints develop a model for a generic class of provided by both local and seismic hotspots which form in regions of the tomographic studies. We explore the role upper mantle which have experienced of discontinuities in the base of the recent subduction and continental lithosphere (e.g. Variscan terrane collision. boundaries) and Tertiary lithospheric extension in nucleating diapiric Paleocene-Recent volcanism within instabilities which may subsequently western and central Europe, which is propagate downwards, producing, “top spatially and temporally linked to the downwards”, small-scale structures which development of a major intra-continental look like plume stems but which may have rift system and to domal uplift of Variscan a totally different origin. basement massifs, has been attributed to the diapiric upwelling of small-scale, [1] Granet, M., Wilson, M. and Achauer, finger-like, convective instabilities from U. (1995) Imaging a mantle plume beneath the base of the upper mantle [1]. Evidence the French Massif Central. Earth Planet. for this model comes from the French Sci. Lett., 136, p. 281-296. Massif Central [1], the Eifel province of [2] Ritter, J. R. R., Jordan, M., northern Germany [2] and the Bohemian Christensen, U. R. & Achauer, U. R.
27 (2001) A mantle plume beneath the Eifel and He isotope compositions of these volcanic fields, Germany. Earth Planet. lavas provide convincing evidence that Sci. Lett. 186, 7-14. their parental magmas formed through [3] Wilson, M. & Patterson, R. (2001) extreme fractional melting and that their Intra-plate magmatism related to hot immediate source was anhydrous. In the fingers in the upper mantle: Evidence from Archean magmas were still hotter. The the Tertiary-Quaternary volcanic province most magnesian komatiites contain 30- of western and central Europe. In: R. Ernst 32% MgO and erupted at temperatures & K. Buchan (Editors) Mantle Plumes: close to 1600°C. We do not accept Their identification through time. GSA arguments that common Archean Special Paper 352, pp37-58. komatiites came from a wet subduction [4] Piromallo, C., Vincent, A. P., Yuen, D. environment (some komatiites did indeed A. & Morelli, A. (2001) Dynamics of the contain a small amount of water but these transition zone under Europe inferred from rocks have unusual compositions and are wavelet cross-spectra of seismic rare). We attribute the formation of normal tomography. Physics of the Earth and komatiites to deep melting in abnormally Planetary Interiors, 125, 125-139. hot mantle plumes.
Monday Oral Sessions Komatiites and the Temperature of the Mantle: “Some like it hot”. The Temperatures of Mantle Plumes M. J. Cheadle (Dept. of Geology and N. T. Arndt (LGCA, Univ. Joseph Fourier, Geophysics, Univ. of Wyoming, Laramie, Grenoble. France; E-mail: arndt@ujf- WY 82071 USA; Tel: +1-307-766-3206; grenoble.fr); C. Herzberg (Dept. Fax: +1-307-766-6679; E-mail: Geological Sciences, Rutgers Univ., NJ [email protected]); D. Sparks (Dept. of 08854, USA; E-mail: Geology and Geophysics, Texas A&M [email protected]) Univ., College Station, TX, 77843-3115, USA; Tel: +1-979-458-1051; Fax: +1-979- By using the compositions of whole-rock 845-6162; E-mail: samples and the Fo contents of olivine [email protected]) phenocrysts it is possible to estimate temperatures in the mantle sources of Archaean komatiites are the highest MgO mafic and ultramafic lavas. The hottest lavas recorded on the Earth. Two end- present-day volcanism is on Hawaii where member models have been suggested to picrites with estimated liquidus explain their formation: i) anomalously temperatures between 1400-1450°C high temperature, relatively ‘dry’, melting erupted. The lower temperature (in mantle plumes?) or ii) by volatile/water corresponds to a water content of 0.3- induced melting (in subduction zones?). 0.4%, the maximum likely in Hawaiian As with modern-day mafic magmas, it is picrites. These eruption temperatures are possible that komatiites were generated in some100-150 degrees higher than those of both environments; the key problem is to the source of MORB: Hawaii, the type recognise diagnostic characteristics that example of an active plume, is the hottest can be used to determine the environment of active hot spots. in which any given komatiites formed.
In the recent past, temperatures were still Field constraints are most important. higher. The 90 Ma old komatiites of Many komatiites are demonstrably Gorgona Island off the coast of Colombia, phenocryst poor lava flows that erupted erupted at temperatures only slightly lower non-explosively and crystallized spinifex than those of the hottest Hawaiian lavas, textured flow-tops above olivine-rich but Gorgona picrites erupted at cumulate lower zones at low pressures. temperatures a little over 1500°C. The Consequently, it is most unlikely that they trace-element characteristics and the Nd were magmatically wet melts. Even if they
28 were, they probably would have degassed low-Mg basalts erupted at mid-ocean water before crystallization and thus it is ridges or in volcanic arcs throughout difficult to infer magmatic water contents geological time, the most magnesian by studying the products of in-situ or magmas identified thus far in the shallow crystallization. Phanerozoic are low-Mg komatiites with up to ~20% MgO (e.g., Baffin, Gorgona), The second diagnostic characteristic is the most magnesian magmas identified whole-rock geochemistry, yet this is thus far in the Proterozoic are low-Mg fraught with the problem that komatiites komatiites with up to 22% MgO (e.g., are high degree melts (up to 50% of the Thompson), and the most magnesian mantle) and thus major and many trace magmas identified thus far in the Archean element compositions must converge are high-Mg komatiites with up to 32% regardless of whether they are formed by MgO (e.g., Abitibi, Barberton, Belingwe, volatile or high temperature induced Norseman-Wiluna). Although the melting. Interpretation of the geochemistry production and preservation of mafic- is further complicated by the fact that ultramafic magmas has not been komatiites require physical conditions of continuous, reflecting specific periods melt generation that do not occur today. with unusually high rates of production Models of komatiite generation based on and preservation (e.g., 1.9 Ga, 2.7 Ga) , geochemistry must be physically plausible there appears to have been an abrupt, not and all existing models require high continuous, change in the maximum MgO mantle temperatures. Komatiite generation content and therefore potential at subduction zones requires higher mantle temperature of mantle-derived magmas at wedge temperatures (+/- faster convection) the end of the Archean. Such an abrupt than today, and thus implicitly requires the change in mantle potential temperature, presence of high temperature mantle. We along with the plethora of other will present the arguments for komatiite fundamental changes in geological generation by anhydrous melting and a processes that occurred at the end of the viable physical model for their origin in Archean, suggests that there was a high temperature mantle plumes. We will fundamental change in the thermal use as our example, one of the best- structure of the mantle at that time. It has preserved sequences of Archean been suggested, however, that some or all komatiites from the Belingwe greenstone high-Mg komatiites have been produced belt in Zimbabwe. by hydrous melting. Although some komatiites do contain vesicles and/or igneous amphibole, they occur only High-Mg Magmatism Through Time: locally, even within individual volcanic- Implications for Mantle Plumes stratigraphic units, which suggests that water was derived locally during C. M. Lesher (Mineral Exploration emplacement. Critically, most komatiites Research Centre and Dept. of Earth (and many associated basalts) are Sciences, Laurentian Univ., Sudbury, ON moderately to strongly depleted of highly- P3E6B5, Canada; Tel: +1-705-675-1151 x incompatible lithophile elements relative 2276; Fax: +1-705-675-4898; E-mail: to moderately incompatible lithophile [email protected]) elements, have positive epsilon Nd values, and have broadly chondritic gamma Os, Of the many arguments that have been consistent with derivation by partial made for the existence of mantle plumes, melting of a long-term depleted mantle the presence of high Mg magmas such as source. Because H2O is incompatible picrites and komatiites magmas with during melting, we may infer that melting inferred potential temperatures greater was not facilitated by the addition of water than that which would be produced by or other metasomatic components. Indeed, normal decompressive or hydrous melting the strong depletion in HILE suggests that of upper mantle seems to be the most the komatiite source, which may be compelling. Compared to the ubiquitous otherwise interpreted to be representative
29 of normal convecting asthenospheric Root mean square variations along the mantle, may have contained as little as 20 Oahu and Maro Reef profiles are 15 and 5 ppm H2O. The majority of komatiites mW m**-2, respectively, and along both appear to have been generated by high- Reunion profiles are about13 mW m**-2. temperature, high-degree partial melting of Coupled heat and fluid flow models normal convecting asthenospheric mantle, demonstrate that thermal buoyancy due to most likely in a thermal plume. bathymetric relief is capable of driving significant fluid flow that may suppress the background thermal field. These Observations of Heat Flow on Hotspot models are consistent with heat flow Swells patterns observed at individual seamounts and oceanic basement highs that are more Robert N. Harris (Dept. of Geology and easily sampled and characterized than Geophysics, Univ. of Utah, Salt Lake City, large hotspot swells. We caution that the UT 84102-0111, USA; Tel: +1-801-587- ability of fluid flow to mask variations in 9366; Fax: +1-801-581-7065; E-mail: sublithospheric heat flux is under [email protected]); Marcia K. appreciated. McNutt (Monterey Bay Aquarium Research Inst., Moss Landing, CA 95039, USA; E-mail: [email protected]) MORB Major-Element Systematics: Implications for Melting Models and Heat flow data collected on hotspot swells Mantle Temperatures have been used to argue for and against sublithospheric thermal anomalies. The D. C. Presnall (Geophysical Laboratory, presence of sublithospheric thermal 5251 Broad Branch Rd., NW, anomalies has been inferred from Washington, DC 20015, USA; Tel: +1- interpretations of anomalously high heat 202-478-8905; Fax: +1-202-478-8901; E- flow determinations, whereas the mail: [email protected]); G H contention that swells result from normal Gudfinnsson (Geophysical Laboratory, melting processes within the lithosphere is 5251 Broad Branch Rd., NW, based on 'normal' heat flow values. These Washington, DC 20015, USA; Tel: +1- arguments depend in part on the choice of 202-478-8931; Fax: +1-202-478-8901; E- a thermal reference model, but more mail: [email protected]) importantly assume conductive heat transfer through the lithosphere. We Klein and Langmuir (1987, 1989) and review heat flow determinations collected Langmuir et al. (1992) (hereafter, KL87) on hotspot swells and argue that shallow argued that MORB magmas, including fluid flow is likely obscuring the thermal those from Iceland, are fractionated conditions within or at the base of the aggregates of melts produced by lithosphere. combining small basaltic melt fractions from a wide range of pressures in melting Discriminating between environments columns. A crucial aspect of their where heat is transferred conductively or modeling is their correlation of Na8 with convectively requires closely spaced heat crustal thickness. However, this hinges flow determinations (1-2 km) collocated strongly on two extrema, one for Iceland, with seismic reflection profiles. Only where the crustal thickness is Hawaii and Reunion have surveys controversial, and the other for the satisfying these criteria. The Hawaiian Cayman Trough, where the crustal survey consists of two profiles, one north thickness is based on geological inference of Oahu and one north of Maro Reef. The rather than seismology. Also, Reunion survey also consists of two experimental data show that low-pressure profiles, both north of Mauritius. These olivine-controlled fractionation is required heat flow profiles reveal greater scatter for melts produced in the KL87 model to than anticipated with spectral peaks on the reach the composition range of observed order of 10 km consistent with fluid flow. MORB glasses. Here we reexamine the
30 global correlations of KL87. To eliminate UTas, GPO Box 252-79, Hobart, interlaboratory analytical differences, we Tasmania 7001, Australia; Tel: +61-3- use the Smithsonian database of basalt 62262469; Fax: +61-3-62262547; E-mail: glasses. Also, we use a separate set of [email protected]); D.H. Green published and unpublished basalt glass (Research School of Earth Sciences, compositions from Iceland, all analyzed Australian National Univ., Canberra, on the electron microprobe at the Australian Capital Territory 0200, University of Iceland. Neither the Iceland Australia; Tel: +61-2-61252488; Fax:+61- nor MORB data sets show even a hint of 2-61250738; E-mail: the olivine-controlled fractionation [email protected]) required by the KL87 model. For the East Pacific Rise, we confirm the result of To compare magmatic crystallization KL87 that an inverse Na8-Fe8 correlation temperatures between ‘plume-related’ OIB occurs at all length scales along the ridge. and normal MORB tholeiite magmas we For the Mid-Atlantic Ridge, KL87 found have examined in detail two well studied "local" trends (< 100 km ridge length) suites of rocks where we have both whole- with positive Na8-Fe8 correlations. For rock and glass data as well as constraints the North Atlantic south of the Charlie on volatile contents. The two suites are the Gibbs Fracture Zone, we find that the data 1959 summit eruptions of Kilauea are dominated by two offset positive volcano, and MORB rocks and glasses correlations, each of which extends more from ODP Leg 148 (hole 896A). To model than 2000 km along the ridge. That is, the the magmatic crystallization of olivine we local trends of KL87 are actually regional. use the software PETROLOG The Iceland data show a third trend for (/Danyushevsky/, 2001) and the olivine Na8-Fe8 that is offset from the other two. melt model of /Ford et al./ [1983]. The Thus, the Na8-Fe8 data for the Mid- effect of H_2 O on the olivine liquidus Atlantic Ridge are best described as a surface was modelled using the model of series of offset positive trends rather than a /Falloon and Danyushevsky/ [2000]. single trend. For both the East Pacific Rise and Mid-Atlantic Ridge, the data are The 1959 summit eruption of Kilauean consistent with magma generation in the volcano provides a good example of a range 1-1.5 GPa from a heterogeneous picritic suite of rocks where whole-rock mantle, which implies magma-generation compositions are controlled by the temperatures of about 1260-1280 C. For mechanical accumulation of olivine. Glass ridges in other parts of the ocean basins, sampled from the eruptions by /Murata the amount of data in the Smithsonian and Richter/ [1966] however define a database is inadequate for robust analysis liquid-line of descent controlled by of this type, but the data give no indication equilibrium fractional crystallization of that higher magma-generation olivine (/Woronow et al.,/ 1996). Using temperatures are required. Petrolog and volatile constraints from quenched picritic glasses from Kilauean volcano (/Clague et al.,/ 1991) we Magmatic Crystallization Temperatures calculate magmatic olivine crystallization of Tholeiite Magmas: Implications for temperatures for the glass data of /Murata the Existence of Thermally Driven and Richter/ [1966] of between 1189.8- Mantle Plumes 1065.6°C for the 1959 eruptions, which matches almost exactly the range of T. J. Falloon (School of Earth Sciences observed eruption temperatures (1190- and Centre for Marine Science, UTas, 1060°C, /Ault et al.,/ 1961). Using GPO Box 252-79, Hobart, Tasmania 7001, Petrolog we calculate a parental magma of Australia; Tel: +61-3-62262270; Fax: +61- 16.5 wt% MgO, containing 0.51 wt% H_2 3-62262547; E-mail: O with an eruption temperature of 1328°C [email protected]); L.V. at 1bar and in equilibrium with the most Danyushevsky (Centre for Ore Deposit magnesian olivine phenocryst from Research and School of Earth Sciences, Kilauea volcano (Fo 90.7). The average
31 result using 5 other olivine melt models is Monday Poster Sessions a parental magma of 16.5±0.3 wt% MgO at a temperature of 1327±15°C. /Eggins/ How Hot is Iceland? [1992] studied a Kilauean parent magma of 16 wt% MgO and demonstrated G. R. Foulger (Dept. Earth Sciences, Univ. multiple saturation with peridotite of Durham, Durham DH1 3LE, U.K.; Tel: (harzburgite) at 2 GPa 1450°C under +44-191-386-4533; Fax: +44-191-334- nominally anhydrous conditions. Using 2301; E-mail: [email protected]) Petrolog our parental composition is in equilibrium with olivine at 1429°C at 2 High potential temperature compared with GPa (1485°C anhydrous). surrounding mantle is a primary characteristic of mantle plumes. A /McNeill and Danyushevsky/ [1996] temperature anomaly of the order of studied depleted MORB glasses from hundreds of Kelvin is thought to be ODP Leg 148, hole 896A. The most required, even for a weak upper-mantle magnesian glass had MgO of 9.43 wt% plume rising from the base of the mantle MgO and 0.05 wt% H_2 O. Using transition zone. (It should be mentioned in Petrolog a calculated parent in equilibrium passing that this latter model ignores the with the maximum olivine observed (Fo fact that the base of the transition zone is a 91.6) has 15.6 wt% MgO, 0.04 wt% H_2 mineralogical phase change and there is no O and a liquidus temperature of 1336°C at evidence that it is a strong temperature 1bar. The average result using 5 other discontinuity.) Many independent methods olivine melt models is a parental magma have been used to estimate the temperature of 15.8±0.2 wt% MgO at a temperature of anomaly at various depths in the crust and 1334±5°C. We have performed reaction mantle beneath Iceland. Seismic methods experiments on this calculated parental include measuring attenuation and the composition with peridotite MM-3 at 1.8 Vp/Vs ratio to place bounds on the and 2.0GPa. The reaction experiments temperature of the crust, and P- and S- demonstrate that this calculated parent is wave mantle tomography to study the in equilibrium with mantle peridotite mantle. P- and S-wave receiver functions (lherzolite) at 2 GPa 1460° C under have been applied to study the depths of nominally anhydrous conditions. Using discontinuities, including those associated Petrolog our parental composition is in with the low-velocity zone (between ~ 80 equilibrium with olivine at 1436°C at 2 and 135 km depth) and the 410- and 660- GPa (1462°C anhydrous). km discontinuities that form the upper and lower boundaries of the mantle transition The results of this study demonstrates that zone. Seismic wave travel times have been there is very little difference (<10°C) in used to estimate the velocities between the the magmatic temperatures at 1bar, and discontinuities. Temperature has also been temperatures and depths of origin, estimated using petrological methods such between two well studied OIB and MORB as olivine glass geothermometry, the study tholeiite suites, with the MORB suite of melt inclusions in basalts, CMASNF having a slightly higher temperature. The geothermometry of high-MgO glasses, results of this study support the major element systematics of Icelandic conclusions of /Green et al./ [2001] and MORB and the search for an olivine /Green and Falloon/ [2005] that their is no control line in Icelandic picrite cumulates. evidence for a temperature contrast of Additional approaches include deducing (delta)Tp ~ 200-250°C between ‘Hot-spot’ thermal history by modeling the or ‘Deep Mantle Plume’ sources and bathymetry of the north Atlantic, the ambient (MOR source) asthenospheric subsidence of the ocean crust and uplift of mantle. the Hebrides shelf. Heat flow measurements from the oceans provide additional weak constraints on temperature.
32 All results either require or are compatible Why the High PGE Contents of with a mantle potential temperature Komatiites, Picrites and Allied Rocks anomaly of no more than ~ 50-100 K Require Mantle Plumes beneath Iceland. The Icelandic crust is cooler than that beneath the East Pacific Reid R. Keays (VIEPS, School of Rise. Seismic tomography is compatible Geosciences, Monash Univ., Victoria, with temperature anomalies of up to 200 Australia, 3800; Tel: +61-3-9905-4072; K, decreasing to about 100 K at depths Fax: +61-3-9905-4903; E-mail: greater than 200 km where the seismic [email protected]) anomaly is weaker. This assumes that compositional effects are zero and partial The high concentrations of both Pd and Ir melt is absent, however. Other seismic in komatiites, komatiitic basalts, picrites results require the presence of partial melt, and some Continental Flood Basalts necessitating substantial downward- (CFB) require both high temperature and adjustment of the temperature anomaly high degree partial melting of their mantle estimate from tomography. P-wave source reservoirs. The only mechanism by receiver functions show that the 410-km which the sufficiently high temperatures discontinuity is warped downward by ~ 15 required can be achieved is through mantle km but that the 650-km discontinuity is plume processes which transfer heat from flat. A hot upwelling penetrating the entire deep in the mantle to the regions of mantle transition zone would predict magma generation. Palladium is the most upward-warping of the 650-km chalcophile metal known, and is retained discontinuity in addition to downward- in the mantle reservoir until all of the warping of the 410-km discontinuity. The sulfides in the mantle reservoir are results are consistent with a temperature exhausted by high temperature, high anomaly of about 100 K at 410 km but degrees (>25%) of partial melting. zero at 650 km. All petrological methods Although Ir is also chalcophile, it is one of suggest relatively small temperature the most siderophile elements known. anomalies unless olivine control is Iridium occurs in mantle and crustal assumed for Icelandic picrite cumulates, phases as alloys with Os and possibly Ru. an assumption whose validity is The Ir content of mantle-derived magmas questionable. Modeling of bathymetry and is directly proportional to the degree of vertical motions suggests temperature partial melting which produced them. anomalies of no more than 100 K. Hence, whereas komatiites, which are high temperature magmas produced by high A large majority of temperature estimates degrees of partial melting, have high Pd from the Iceland region are thus consistent and Ir contents, boninites, which are with a modest temperature anomaly that is moderately high temperature magmas that well within what is expected for upper- involve lower degrees of partial melting mantle variability from such processes as due to fluxing of slab-derived volatiles, continental insulation and recent slab have high Pd but low Ir contents. subduction. There is little evidence for the high temperatures required by a plume The high PGE contents and geochemical from the lower mantle rising through its signatures of some CFB are best explained own thermal buoyancy. by the interaction of high temperature, high MgO melts from mantle plumes with lithospheric mantle that has been depleted in S, but mildly enriched in the PGE, by previous melt extraction. High temperature magmas from a mantle plume are required to provide the energy required to melt the refractory lithospheric mantle. In addition, the high crustal contamination signature of the mafic and ultramafic rocks that host many Ni-Cu-PGE sulfide
33 deposits suggests that the mantle-derived Influence of Eclogite in Mantle Sources magmas that produced the ores had to be on ‘Hot-spot’ Temperatures very energetic in order for them to digest the large mass of crustal rocks that were James H. Natland (Division of Marine assimilated. Geology and Geophysics, Rosenstiel School of Marine and Atmospheric Science, Univ. of Miami, Miami, FL Thermal State of NW Kyushu Mantle 33149, USA; Tel: +1-305-421-4123; Suggested by Petrochemistry of Fax:+1-305-421-4632; E-mail: Primitive Basalts [email protected])
H. Mashima (Tono Geoscience Center, Recent studies of compositions of olivine Japan Nuclear Cycle Dev. Institute, Jorin- basalt and olivine-hosted melt inclusions ji 9598-31, Izumi-cho, Toki-shi, Gifu, from Iceland, Greenland, and Hawai`i 509-5102, Japan; Tel: +81-572-53-0211; point to the potential involvement of Fax: +81-572-55-0180; E-mail: recycled ocean crust as eclogite in mantle [email protected]) sources. The most distinctive geochemical indicators (e.g., high Y/Zr, low Zr/REE; Based on Sr-Nd-Pb isotopes and high Sr/REE) result from partial melting incompatible element compositions, of rocks that originally were lower ocean Cenozoic basalts in NW Kyushu, SW crustal gabbroic adcumulates (1) with very Japan is considered to be caused by mantle low residual melt porosities (nearly no plumes with enormously high temperature trapped intercumulus liquid). Abyssal from the 660 km discontinuity and/or core. gabbro adcumulates have significantly Nevertheless, tomographic observations higher Ca/(Ca+Na) at given Mg# = suggest that the 660 km discontinuity 100*Mg/(Mg+Fe2+) than their original beneath NW Kyushu is cold. Moreover, host MORB liquids and most have MgNo petrochemistry of primitive NW Kyushu = 70-85. Experimental studies point to basalts indicates that their source is not so derivation of primitive basaltic liquids hot. Bulk rock MgO-FeO-NiO suggests with low CaO/Al2O3 (0.5-0.9) and that the basalts with MgO = 10-6 wt. % moderate to high MgNo (55-80) from can equilibrate with the mantle. Results of eclogite with similar composition and also previous melting experiments suggest that pyroxenite (e.g., 2). Basaltic experimental these basalts were last in equilibrium with liquids have low total iron (FeOT) the source at 1200-1300 degrees contents (6-10%) at given MgO (6-14%), centigrade at 1GPa. These melting do not lie along olivine-controlled liquid temperatures are relatively low and similar lines of descent and match many inclusion to those of MORB and subduction zone compositions hosted in olivine and other basalts, which suggest that NW Kyushu minerals in picrites from the North basalts were formed by relatively shallow Atlantic Igneous Province (NAIP) and mantle upwelling. Such mantle upwelling Hawai`i (e.g., 3, 4). The crystal-liquid would be caused by lithospheric extention exchange model for Cr-spinel of (5) at the junction of the SW Japan arc and the predicts crystallization from such liquids Ryuku arc. Petrochemical similarity with of spinel that can be either more highly NW Kyushu basalts suggests that origin of chromian (Cr# = Cr/[Cr+Al] = 60-75) than intraplate basalts from other than NW in MORB spinel, or which trends to lower Kyushu can be explained by tectonic- Mg# at nearly fixed Cr#, or both. The controlled hypothesis. trends result mainly from low Mg# and Al2O3 (13-15%) in eclogitic partial melts, the latter especially at pressures >3 GPa. Such trends in spinel are common among picritic basalts from these locations and also Gorgona. Highly chromian spinel with a trend toward iron enrichment thus is a mineral proxy for the influence of
34 eclogite/pyroxenite in mantle sources at all Gronvold, K.,2000. Earth Planet. Sci. these places. Lett., 183: 15-16. 4) Sobolev, A., Hofmann, A., Sobolev, S., and The spinel compositions in turn allow Nikogosian, I., 2005. Nature, 434: 590- estimation of liquid Fe2+/Fe3+, and 597. 5)Poustovetov, A., and Roeder, P., indicate regional differences of oxidation 2000. Canad. Mineral., 39: 309-317. states of recycled crustal components in mantle sources (West Greenland and Gorgona higher than Iceland). The Evidence for High Temperature Mantle inclusions demonstrate that most picrites Plumes Based on Olivine-Liquid are hybrid compositions produced by Thermometry mixing of both primitive (including eclogite-derived) and differentiated K. D. Putirka (California State Univ., magma strains, such that neither the Fresno, Dept. of Earth and Environmental olivine nor the spinel they contain Sciences, 2345 E. San Ramon Ave., crystallized along a single (olivine- MS/MH24, Fresno, CA 93720, USA; Tel: controlled) liquid line of descent. Indeed +1-559-278-4524; Fax: +1-559-278-5980; picrites from Iceland also contain E-mail: [email protected]) phenocrysts of calcic plagioclase and chromian endiopside. Inference of Temperature is perhaps the most important crystallization temperature from assumed variable in the debate over the existence of relationships between forsteritic (Fo91) mantle plumes. If the temperature olivine and an estimated magnesian host differences between hot spots and mid- liquid (MgO = 15-20%) thus is almost ocean ridges are large (160-280 K) then always contradicted by evidence from melt the existence of thermal upwellings would inclusions,olivine and spinel for magma appear more than plausible [Sleep, 1990; mixing. Adjusting properly for oxidation Schilling, 1991]. But if such temperature states inferred from spinel compositions, differences are nil [Anderson, 1998; Green liquids with compositions of the melt et al., 1999], then the modern concept of inclusions are capable of crystallizing plumes must be abandoned. Olivine quite forsteritic olivine (Fo90-92) at low-P phenocrysts from lavas erupted in ocean and T = 1260-1325oC, the highest basins provide perhaps the most useful temperature being about the maximum means for measuring temperature indicated by olivine-liquid thermometry differences. Olivine compositions are for any melt inclusion or natural glass highly sensitive to temperature, and in found at any of Iceland, Greenland, or general, minerals provide a less Hawai`i. The presence of strongly homogenized account of the melting forsteritic olivine in picrites thus is no process compared to glass and whole rock guarantee that they crystallized at samples. temperatures in excess of 1400oC. Spinel compositions point to no higher An updated form of the Roeder and Emslie temperatures of crystallization at either [1970] olivine saturation surface is applied Skye or Gorgona than at Iceland and to olivine phenocrysts from two hot spots, indeed suggest generally similar (but not Hawaii and Iceland, and from a segment komatiitic) primitive liquid compositions of the East Pacific Rise mid-ocean ridge at all localities discussed here, with (MOR) system, the Siqueiros transform. potential temperatures only about 50-100o The Roeder and Emslie [1970] approach higher than for primitive MORB. to thermometry is applied because mineral and host rock compositions can be directly 1) Foulger, G.R., Natland, J.H. and input to estimate T, obviating the need to Anderson, D.L., 2005. J. Volcan. estimate primitive melt MgO (e.g., Green Geotherm. Res., 141: 23-44; 2) Kogiso, et al. [1999]). The updated olivine T., Hirschmann, M., and Pertermann, M., saturation surface shows that olivine- 2004, J. Petrol. 45: 2407-2022. 3) liquid equilibration temperatures at Hawaii Sigurdsson, I., Steinthorsson, S., and and Iceland are hotter by 250 and 165 K
35 respectively compared to MORs (error = different mechanisms have been proposed 50 K). These temperature differences are for hotspot formation. greater than those obtained by Green et al. [1999] for Hawaii, but are more consistent The shallow depth associated with hot with observed FeO contents for Hawaiian spots was originally thought to reflect a lavas. In addition, these estimates include hot mantle plume, rising from deep in the the effects of water, alkalis, and silica on mantle, penetrating the approximately olivine-melt equilibria, and are robust 100-km-thick oceanic lithosphere and against other variations in source or liquid causing heating to about 50 km of the composition, such as FeO and carbon surface. Although anomalously high heat dioxide. Translated to differences in flow was initially reported, consistent with mantle potential temperature, Hawaii and significant lithospheric reheating, Iceland are hotter than ambient MORs by subsequent analysis showed that most if 213-235 and 162-184 K. Absolute not all of the apparent anomalies resulted estimates of mantle potential temperatures from comparing the data to reference are less certain as they depend upon thermal models that underestimated heat estimates of F and depth-of-equilibration; flow elsewhere. The small size or absence presuming F = 0.2 and that magmas are of a heat flow anomaly at midplate delivered from the base of the lithosphere, hotspots such as Hawaii implies that uplift mantle potential temperatures are: Hawaii may result from the dynamic effects of = 1961 K; Iceland = 1910 K; MORs = rising plumes and/or the associated 1726-1748 K. Within error, estimates for compositional buoyancy, whose thermal MORs match T estimates obtained from effects are concentrated at the base of the ocean bathymetry [1723 K; Stein and lithosphere and hence would raise surface Stein, 1992].These temperatures thus show heat flow at most slightly, given that tens that at least at Iceland and Hawaii, of millions of years are required for heat volcanism is driven by high temperature conduction to the surface. thermal anomalies whose magnitudes are consistent with the existence of thermally Recently it has been proposed that high driven mantle plumes. heat flow is not measured because extensive hydrothermal circulation transfers significant amounts of heat. If Does Hydrothermal Circulation Mask so, it is possible that significant Anomalously High Heat Flow at Hot lithospheric reheating may contribute to Spots? hotspot formation. Here, I compare heat flow data and other associated local C. A. Stein (Dept. of Earth and factors from both hotspot regions and Environmental Sciences, m/c 186, Univ. areas of similar ages far from hot spots to of Illinois at Chicago, 845 W. Taylor St., examine this issue. Chicago, IL 60607-7059, USA; Tel: +1- 312-996-9349; Fax: +1-312-413-2279; E- mail: [email protected]) Tuesday Oral Sessions
Sea floor depths and heat flow data are Timing and Duration of Volcanism at two of the primary observable constraints Large Igneous Provinces: Implications on mechanisms for hot spots. A challenge for Geodynamics and Links to Hotspots is that these two data reflect not only the effect of the source of the excess magma, R. A. Duncan (College of Oceanic and but also combinations of "normal" Atmospheric Sciences, Oregon State lithospheric cooling and other local Univ., Corvallis, OR 97331, USA; Tel: phenomena. While hot spots are associated +1-541-737-5189; Fax: +1-541-737-2064; with anomalously shallow depths, it is E-mail: [email protected]) unclear if heat flow is anomalously high. Depending on the answer to this question, Estimates of timescales of volcanic activity at large igneous provinces (LIPs),
36 derived from radiometric methods and Tel: +49-431-6002642; Fax: +49-431- magneto-stratigraphic data, lead to the 6002924; E-mail: khoernle@ifm- conclusion that vast portions of these geomar.de); P vd Bogaard (IFM- enormous volumes of mantle-derived GEOMAR Leibniz Institute for Marine magma were erupted at rates far exceeding Sciences, Wischhoststr. 1-3, 24148-Kiel, any contemporary site, and are therefore Germany; Tel: +49-431-6002129; Fax:+ unrelated to any known plate tectonic 49-431-6002924; E-mail: pbogaard@ifm- melting regime. Furthermore, field geomar.de); F Hauff (IFM-GEOMAR evidence reveals that LIPs are built from Leibniz Institute for Marine Sciences, very large, discrete eruptive events (ca. Wischhoststr. 1-3, 24148-Kiel, Germany; 103-104 km3 each) over periods of ~1 Tel: +49-431-6002125; Fax: +49-431- million years, rather than continuous 6002924; E-mail: [email protected]); volcanic activity. These features are most R Werner (IFM-GEOMAR Leibniz consistent with the startup plume model. Institute for Marine Sciences, One continental LIP, the North Atlantic Wischhoststr. 1-3, 24148-Kiel, Germany; igneous province, and one oceanic LIP, Tel: +49-431-6001416; Fax: +49-431- the Caribbean plateau, have been studied 6002924; E-mail: rwerner@ifm- in particular detail. 40Ar-39Ar incremental geomar.de); J Geldmacher (IFM- heating and U-Pb methods are capable of GEOMAR Leibniz Institute for Marine providing high precision age Sciences, Wischhoststr. 1-3, 24148-Kiel, determinations for extrusive and intrusive Germany; Tel: +49-431-6002642; Fax: constructional phases. In the NAIP case, +49-431-6002924; E-mail: initial volcanic activity (~62 Ma) [email protected]) coincident with the birth of the Iceland hotspot, penetrated thick continental Intraplate volcanism in the Canary, lithosphere over a broad area (1000s of km Madeira, Galapagos and New Zealand diameter), then focused at ~55 Ma into a regions can be traced back to at least 70 narrow spreading zone that led to plate Ma and probably back more than 100 Ma. separation (with eruption rates ~1000 New age and geochemical data from km3/m.y. per km of rift). In the CLIP seamounts to the NE of the Canary and contemporaneous volcanism occurred over Madeira Islands, specifically targeted to a similarly broad area on eastern Pacific test the “hotspot (plume) hypotheses”, oceanic lithosphere at the inception of the confirms general age progressions with Galapagos hotspot (~93 Ma). A ages increasing in the direction of plate subsequent, geographically restricted motion. These age progressive island- phase of volcanic and intrusive activity seamount chains, as is also the case for the occurred 80-76 Ma in response to plateau Walvis and St. Helena chains in the extension. southern Atlantic, can be explained by rotation around a common Euler pole, Clearly defined, age-progressive hotspot consistent with the hotspot hypothesis. tracks connect these LIPs to their presently Irregularities in the chains can be active hotspots. Geochemical explained by a combination of factors, “fingerprints” also link these LIPs to the such as pulsing rather than continuous associated hotspots. plumes, interaction of rising plume pulses (blobs) with edge driven convection against the African continental root and Origin of Long-term Intraplate lithospheric control on the exact location Volcanism in the Canaries, Madeira, and morphology of individual volcanic Galapagos and New Zealand: Which structures. Geochemical and age studies of are Consistent with the Plume the Madeira-Tore Rise suggest that its Hypotheses? Cretaceous basement may have also formed from the Canary hotspot, as the K. A. Hoernle (IFM-GEOMAR Leibniz mid Atlantic Rise passed over the hotspot. Institute for Marine Sciences, Wischhoststr f1-3, 24148-Kiel, Germany;
37 In order to link the Galapagos hotspot to The Geochronology of Hotspot Trails the Caribbean LIP, we also assummed the and the Timing of the Hawaii-Emperor plume model and searched the Central Bend American forearc for accreted parts of paleo-Galapagos Hotspot tracks. Accreted A. A. P. Koppers (Institute of Geophysics terranes with Galapagos-type and Planetary Physics, Scripps Institution geochemistry and ages ranging from ca of Oceanography, La Jolla, CA 92093- 20-70 Ma were found, consistent with the 0225 USA; Tel: +1-858-534-8771; Fax: hotspot hypothesis. Older samples from +1-858-534-8090; E-mail: the Nicoya Peninsula in Costa Rica (111- [email protected]) 139 Ma) and parts of the Guatemalan forearc basement (possibly as old as 170 Recent studies of seamount trails in the Ma) also have Galapagos-type Pacific have shown that at least two geochemistry but don't fit the classical classes of hotspots are required to explain plume head hypothesis. Either mantle these topographic features in terms of their plumes don’t always begin with a single morphology and geochronology. On the plume head or subduction has resulted in one hand, the Hawaiian and Louisville the accumulation of different volcanic hotspots may be explained by deep-seated structures derived from mantle reservoirs mantle plumes that generate intra-plate with similar geochemistry. volcanism over long periods of geological time, in a rather continuous and age In the SW Pacific, new age and progressive manner. Most characteristics geochemical data show that the Hikurangi of these two unique seamount trails are oceanic plateau formed still consistent with the mantle plume contemporaneously and from a similar theory and the standard hotspot source as the Manihiki and Ontong Java hypothesis, even though new plateaus, requiring at least two Ontong paleomagnetic and 40Ar/39Ar Java-sized plateaus to have formed at the geochronology data have shown that these same time (at ca 120 Ma) several thousand mantle plumes cannot be fixed in the kilometers apart. This contemporaneous mantle (moving up to 40 mm/yr with volcanism can best be explained by the respect to the geomagnetic pole) and that plume head hypotheses. Beginning at ca these mantle plumes move slowly (but 100 Ma, HIMU-type volcanism began on significantly) with respect to each other. the Hikurangi Plateau and the neighboring On the other hand, the Pacific plate is Zealandia micro-continent. Extensive littered with short seamount trail segments sampling from submarine parts of the that have persistent isotopic fingerprints. Chatham Rise and Campbell Plateaus, as However, these short-lived seamount trails well as land-based work, show that the generally have irregular age progressions HIMU volcanism has continued until at and show bends that are asynchronous least the Pleistocene. Almost the entire with the infamous bend in the Hawaii- age range occurs on the Chatham Rise. Emperor trail at 47 Ma. These Considering that the plate has moved observations are much harder to reconcile many thousands of kilometers over this with “fixed hotspots” since (by definition) time period, the hotspot hypothesis cannot these age progressions and bends should explain this volcanism. give a consistent picture of the changes in plate motion, happening simultaneously In conclusion, although the hotspot across the entire Pacific plate. (plume) hypotheses provides a good framework to explain many areas of It thus becomes clear that plate motion and intraplate volcanism, it cannot explain all fixed hotspots alone cannot explain these intraplate volcanism. age systematics, but that both hotspot motion and changing lithospheric stress regimes may play an important role in the creation of seamount trails. The simple and elegant “hotspot” model that (almost
38 without difficulty) may explain the 40Ar/39Ar total fusion ages for the primary hotspots such as Hawaii and Samoan Shield, Cook-Austral Island Louisville, seems unsatisfactory in the Chain, New Hebrides-Samoa Lineament explanation of age distributions for short- and the Galapagos seamounts (Natland lived hotspots. and Turner, 1985; Turner and Jarrard, 982; Duncan, 1985; Sinton et al., 1996). Published 40Ar/39Ar stepheating ages for Critical Assessment of Radiometric (altered) submarine rocks from the Ages for Oceanic Hotspot Tracks, Based Western Pacific (Ozima and Saito, 1977; on Statistical Analysis of Individual Ozima et al., 1977) clearly fail the relevant Ages, and Evaluation of the Alteration statistical tests. Stepheating ages for State of the Material Dated guyots in the Northwest Pacific (Winterer et al., 1993), also generally fail the Ajoy K. Baksi (Dept. of Geology and relevant statistical tests. In tracing Easter Geophysics, Louisiana State Univ., Baton Chain volcanism (O’Connor et al., 1995) Rouge, LA 70803, USA; Tel: +1-225-578- and the Foundation Chain (O’Connor et 3422; Fax: +1-225-578-2302; E-mail: al., 1998), a few ages need to be rejected; [email protected]) statistical evaluation shows that many of the other results are “too good”, with The argon dating methods are most probabilities of occurrence generally well commonly used to determine the ages of over 0.90. The material dated was material thought to be related to hotspot relatively fresh, as based on 36Ar tracks. Rocks recovered from oceanic contents. The plateau ages may be more islands and/or the seafloor, tend to altered, precise than suggested by the authors. resulting in loss of (some of the) radiogenic argon. The literature is All ages purporting to show linear age peppered with unreliable ages. K-Ar dates progression of hotspot tracks, must be and/or 40Ar/39Ar total fusion ages, in carefully evaluated, using the techniques general, should not be treated as accurate outlined herein. In light of the work of estimates of the time of crystallization. Sharp and Clague (2002), this includes the 40Ar/39Ar stepheating ages must pass the (classical) case of the Hawaii-Emperor requisite statistical tests (Baksi, 1999, Chain. 2005) to be termed proper plateaux and/or isochron ages. Further, the data must be examined closely in light of 36Ar Volcanic Imprint of Oceanic Hot Spots contents, to eliminate “ages” obtained and LIPs: Shallow/Local versus from altered material. Deep/Global?
It has been shown (Baksi, 1999, 2005) that J. M. O’Connor (Dept. of Isotope there is no evidence of a linear progression Geochemistry, Faculty of Earth Sciences, of hotspot track ages from the Indian and Vrije Univ., De Boelelaan 1085, 1081 HV Atlantic Oceans. Specifically for the Amsterdam, The Netherlands; Tel: Ninetyeast Ridge (Kerguelen Hotspot), the +31(0)20-598-7289; Fax: +31(0)20-598- Chagos-Laccadive and Mascarene 9942; E-mail: Plateaus (Reunion Hotspot), New England [email protected]); P Stoffers Seamount (Great Meteor Hotspot) and the (Institute for Geosciences, Christian- Rio Grande Rise and Walvis Ridge Albrechts-Univ., Olshausenstraße 40-60, (Tristan da Cunha Hotspot). It will be D-24098 Kiel, Germany; Tel: +49-(0)431- shown that the ages purported to be linked 880-2850; Fax: +49(0)431-880-4376; E- to the St. Helena and Walvis hot spot- mail: [email protected]); J R. Wijbrans plume systems (O’Connor and LeRoex, (Dept. of Isotope Geochemistry, Faculty 1992), are without merit. of Earth Sciences, Vrije Univ., De Boelelaan 1085, 1081 HV Amsterdam, For the Pacific Ocean, I will illustrate the The Netherlands; Tel: +31(0)20-598- dangers of using K-Ar dates and 7296; Fax: +31(0)20-598-9942; E-mail:
39 [email protected]); T A Worthington with sampling and analyzing seawater (Institute for Geosciences, Christian- altered seamount and aseismic ridge Albrechts-Univ., Olshausenstraße 40-60, dredge samples. D-24098 Kiel, Germany; Tel: +49 (0)431- 880-2854; Fax: +49(0)431-880-4376; E- The widely accepted explanation for the mail: [email protected]) origin of Galapagos hotspot volcanism is motion of the Cocos and Nazca plates over A fundamental assumption of the mantle a small narrow plume 'tail' located at the plume paradigm for the past 10-15 years western end of the Archipelago (e.g. Geist has been that massive LIPs are formed by et al., 1988, White et al., 1993, Kurz & a short lived 'starting' plume head, Geist, 1999, Harpp & White, 2001, Harpp followed by plume 'tail' rising from the & Geist, 2002). Extensive volcanism deep mantle to create long-lived hotspots. extending across the Galapagos Time-progressive lines of islands, Archipelago/Platform is explained by seamounts and ridges are elegantly interaction between this plume 'tail' and explained by the migration of tectonic the Cocos-Nazca spreading center (e.g. plates over hotspots created by these Geist et al., 1988, White et al., 1993, Kurz narrow plume conduits. Thus, inferring & Geist, 1999, Harpp & White, 2001, narrow, underlying mantle plume 'tails' Harpp & Geist, 2002, Werner et al., 2003, from comparably small regions of active Schilling et al., 2003). Although the hotspot volcanism has become an essential Galapagos Archipelago is one of the best aspect of explaining the origin and studied active hotspot systems, very little characteristics of hotspot volcanism in age control is available for the far more terms of the mantle plume hypothesis. extensive aseismic ridges and associated This has led to a situation where a major a seamounts extending away from the priori question that must be addressed Galapagos Archipelago on the Nazca and when interpreting age, geochemical and Cocos tectonic plates. We therefore geophysical data for oceanic hotspot reconnaissance-sampled these volcanic volcanism and nearby spreading centers is: structures during a 1999 cruise of the R/V "where is the mantle plume"? SONNE. Measured Ar/Ar ages suggest that the Galapagos hotspot influences a However, the fundamental, simple much larger region than inferred from assumption now inherent in the plume small clusters of active ocean islands and hypothesis of small hotspots marking the seamounts (O'Connor et al., 2004, location of narrow plume 'tails' seems not submitted 2005). Broad zones of to be holding up in light of age dating of overlapping hotspot volcanism are created new dredge samples from seamounts and on both the Cocos and Nazca plates. The aseismic ridges linked to the Galapagos complex tectonic history of the Cocos- (O'Connor et al., 2004, submitted 2005) Nazca spreading center has controlled how and Foundation (O'Connor et al., 1998, this volcanism is apportioned between the 2001, 2002, 2004) hotspots. We review Nazca and Cocos plates. However, plume here these examples of how substantially 'tail' - spreading center interaction does not improved age control for the volcanic seem to be the primary process controlling imprint of key oceanic hotspots can point the volume and distribution of Galapagos the way forward to testing both the plume hotspot volcanism. Thus our new data are paradigm and new alternatives. Such new not compatible with the widely held notion information raises the issue of whether we of a small Galapagos hotspot marking a are over-interpreting ocean island and comparably narrow plume conduit - with spreading center data at the expense of most hotspot volcanism reflecting fundamental information from older more interaction between the Galapagos plume widespread seamounts and aseismic and the Cocos-Nazca spreading center ridges? Nonetheless, this limitation on (O'Connor et al., 2004, submitted 2005). data required to better test the plume Although evidence for a very broad hypothesis and alternatives is likely to Galapagos hotspot melting anomaly can continue due to the challenges associated be interpreted as favouring the non-plume
40 alternative models, another key prediction continental hotspot activity than predicted of the plume theory - age-progressive by the 'starting' plume head model. volcanism, the main observation leading to the birth of the plume model in the first One of the main reasons for investigating place - has been reinforced for Galapagos hotspots and LIPs is the quest for a deep, hotspot volcanism on both the Cocos and global mantle process? Evidence for Carnegie plates. correlations between synchronous hotspot events or processes on a global scale and Measured ages for samples recovered major changes in plate tectonic might well during our 1998 sampling of the support some of the shallow/local Foundation Seamount Chain, SE Pacific alternative models for the origin of provides supporting evidence for the hotspots? However, identification of such notion of broad hotspot melting anomalies synchronous events/processes that do not associated with age progressive volcanism. correlate with major changes in plate The Foundation Chain, a narrow line of tectonics might better reflect a deep/global seamounts extending from the Pacific- mantle process? Our recent studies are Antarctic spreading center, is demonstrating that the distribution, fundamentally age-progressive. However, composition and history of the volcanic broadening of the chain linked to changes imprint of oceanic hotspots must first be in lithospheric structure suggest that the resolved in sufficient detail to identify underlying hotspot melting anomaly is far shallow/local lithospheric processes before broader than indicated by the width of the it can become possible to test for the chain (O'Connor et al., 2001, 2002, 2004). signal of any possible deeper/global mantle process. If the notion of plume 'tails' is not finding support from new Galapagos and Cenozoic Vertical Movements on the Foundation age data then what is the NW European ‘Passive’ Margin: situation regarding the equally important Responses to Upper Mantle and elegant idea of 'starting' plume heads? Convection? Most studies tend to focus on direct sampling and geophysical profiling of D. Praeg (Istituto Nazionale di LIPs. However, relying exclusively on this Oceanografia e di Geofisica Sperimentale approach could be problematic given the - OGS - Trieste, Italy; Tel: +39-040-214- great difficulty of sampling these massive 0340; Fax: +39-040-214-327307; E-mail: structures. A complementary approach is [email protected]); M S Stoker to look at the temporal relationship (British Geological Survey, BGS, between the volcanic traces of proposed Edinburgh, EH9 3LA, UK; Tel: +44- 'starting' plume heads (LIPs) and 'tails' (0)131-650-0374; Fax: +44(0)131-668- (seamount chains and aseismic ridges). 4140; E-mail: [email protected]); P M Age-progression along the Walvis Ridge, Shannon (Dept. of Geology, Univ. College South Atlantic, has suggested a very close Dublin, Ireland; Tel: +353-1-7162331; temporal link between rapid Continental Fax: +353-1-2837733; E-mail: Flood Basalt emplacement and continental [email protected]); STRATAGEM rifting followed by the onset of Walvis Partners Ridge formation. But more recent age data for the St Helena Seamount Chain The North Atlantic continental margins are indicates that the established migration archetypally passive, yet they contain rate for volcanism along the Walvis Ridge evidence of post-rift vertical movements is not yet sufficiently well constrained of up to km-scale. The Cenozoic history of (O'Connor et al., 1999) to be cited as such movements on the NW European supporting the 'starting' plume head margin, from Ireland to mid-Norway, has model. This opens up the possibility of been examined by integrating published testing both the 'starting' plume head analyses of tectonic subsidence and/or model and important alternatives that uplift with higher resolution tectono- propose a much longer/earlier history of stratigraphic evidence of sedimentary
41 responses (including results from the EC and tilting is proposed to record changing STRATAGEM project). Three episodes of forms of small-scale convection in the differential vertical or epeirogenic upper mantle of the NE Atlantic region movement are identified, in the early, mid- during its opening. The evolution of upper and late Cenozoic, distinct from at least mantle convection during ocean opening one phase of compressive tectonism. Two has been proposed to account for episodic types of epeirogenic movement are post-rift tectonism on other Atlantic recognised, referred to as tilting (coeval ‘passive’ margins and to be the underlying uplift and subsidence over 100s of cause of plate reorganisations. kilometres, rotations <1 degree) and sagging (strongly differential subsidence over <100 km, rotations of several Constraining the Geometry and Flow of degrees). Each epeirogenic episode the Iceland Mantle Upwelling involved relatively rapid (<5-10 Ma) km- scale tectonic movements that drove major Richard M. Allen (Dept. Earth & changes in the patterns of offshore Planetary Science, Univ. of California sedimentation to find expression in Berkeley, 307 McCone Hall, Berkeley CA regional unconformity-bounded 94720, USA; Tel: +1-510-642-1275, Fax: successions. Thus tilting in the early +1-510-643-5811; E-mail: Cenozoic (c. 60-50 Ma) and late Cenozoic [email protected]); Mei Xue (Dept. (from the early Pliocene, 4 ± 1 Ma) Earth & Planetary Science, Univ. of resulted in the basinward progradation of California Berkeley, 307 McCone Hall, shelf-slope wedges from uplifts along the Berkeley CA 94720, USA; E-mail: inner continental margin and offshore [email protected]) highs, as well as late Cenozoic changes in oceanographic current circulation; sagging The Great Plume Debate is to some extent in the mid-Cenozoic (c. 35-25 Ma) ended fueled by incomplete consideration of the progradation of early Cenozoic shelf- uncertainties in constraining datasets slope wedges and outstripped which allow inconsistencies between sedimentation to dreve the onset of hypothetical models and constraining data contourite systems in underfilled basins. to be excused. We will present our constraints on the range of low velocity The early, mid- and late Cenozoic geometries in the mantle beneath Iceland epeirogenic episodes coincided with and on the geometry of flow beneath the Atlantic plate reorganisations, but the region which, contrary to previous studies, observed km-scale tectonic deflections are does show evidence for flow related to the too large to be explained as flexure in upwelling. response to intra-plate stress variations. Mantle-lithosphere interactions are The low-velocity structure beneath Iceland implied, but the succession of transient is constrained by a suite of resolution tests epeirogenic movements, of differing form, designed to determine the range of are difficult to reconcile with the various velocity structures that satisfy the data. syn- to post-rift mechanisms of permanent These include ray-theoretical squeezing uplift (e.g. underplating, delamination) or experiments which attempt to force variations in regional dynamic support velocity anomalies into specific proposed in the hypothetical context of a geometries while still satisfying the plume beneath Iceland. The epeirogenic dataset, and finite-frequency experiments movements can be explained as dynamic which use the Spectral-Element Method topographic responses to different forms (SEM) to simulate full waveform of upper mantle convective flow: tilting as propagation through various 3D velocity coeval upwelling and downwelling above models. We find that the width of the an edge-driven convection cell; sagging as upwelling conduit beneath Iceland must lie a loss of dynamic support above a former in the range of 100 to 200 km which is convective upwelling. The observed generally narrower than ray-theoretical succession of epeirogenic tilting, sagging tomography and broader then preferred
42 geodynamic models. Separate tests on the Africa have lateral extensions of over minimum depth extent of the anomaly 1000 km and exist in the entire mantle, show that significant low velocities are representing two superplumes. The Pacific required to 350 km depth using this superplume has a larger spatial extent and regional dataset. Should the true conduit stronger slow anomalies than that of the be at the narrower end of the possible Africa superplume. The Hawaiian plume range, both compressional and shear wave is not part of the Pacific superplume. The perturbations greater than 10% would be slow anomalies under hotspots usually do required to depths of at least 350 km. not show a straight pillar shape, but exhibit winding images, suggesting that Two end-member geometries, radial flow plumes are not fixed in the mantle but can and ridge-channeled flow, have been be deflected by the mantle flow. As a proposed for the dispersion of material consequence, hotspots are not really fixed upwelling beneath Iceland. We use but can wander on the Earth's surface, as teleseismic shear-wave splitting evidenced by the recent paleomagnetic and observations to constrain the flow numeric modeling studies. Wider and geometry beneath the region. The more prominent slow anomalies are observed anisotropy pattern is inconsistent visible at the core-mantle boundary with radial flow away from the upwelling. (CMB) than most of the lower mantle, and Instead we propose a ridge-channeled flow there is a good correlation between the model in which there is horizontal flow of distribution of slow anomalies at the CMB material away from the upwelling axis and that of hotspots on the surface, beneath southeast Iceland toward the suggesting that most of the strong mantle southern end of the Kolbeinsey Ridge and plumes under the hotspots originate from the northern end of the Reykjanes Ridge, the CMB. However, there are some small- both of which are west of the upwelling. scaled, weak plumes originating from the This geometry is similar to the ridge transition zone or mid mantle depths. perpendicular flow predicted for offridge Clear images of subducting slabs and hotspots towards the ridge. We mgma chambers in the upper mantle hypothesize that upwelled material then wedge beneath active arc volcanoes are feeds ridge parallel asthenospheric obtained, indicating that geodynamic channels beneath the North Atlantic systems associated with arc magmatism Ridge. Our interpretation is thus consistent and back-arc spreading are related to deep with generation of V-shaped ridges by processes, such as convective circulation channeling of upwelling material down the in the mantle wedge and dehydration Reykjanes and Kolbeinsey ridges. reactions of the subducting slab. Evidence also shows that arc magma and slab dehydration may also contribute to the Multiscale Seismic Tomography of generation of various types of earthquakes Mantle Plumes and Subducting Slabs in subduction zones. Most of the slab materials are stagnant in the mantle Dapeng Zhao (Geodynamics Research transition zone before finally collapsing Center, Ehime Univ., Matsuyama 790- down to the core-mantle boundary as a 8577, Japan; Tel: +81-89-927-9652; Fax: result of large gravitational instability +81-89-927-8167; E-mail: from phase transitions. [email protected])
Local, regional and global tomographic studies are made to image mantle plumes and subducting slabs. Plume-like slow anomalies are clearly visible under the major hotspot regions in most parts of the mantle, in particular, under Hawaii, Iceland, South Pacific and Africa. The slow anomalies under South Pacific and
43 The Role of Mantle Plumes in the Earth's Heat Budget Histograms of accepted models show a 'maximum likelihood' convective regime Guust Nolet (Dept. of Geosciences, where plume fluxes are of the order of 1 Princeton Univ., Princeton NJ 08544, W/m2 at the center of the plume, with rise USA; E-mail:[email protected]); Shun velocities of the order of 1 cm/yr. These Karato (Dept. of Geology and Geophysics, estimates show large uncertainty, but, Yale Univ., New Haven, CT 06520, USA; when extrapolated to all plumes, would be E-mail: [email protected]); consistent with a situation in which the Raffaella Montelli (Dept. of Geosciences, volume flux of downgoing slabs is exactly Princeton Univ., Princeton NJ 08544, balanced by that in upgoing plumes, and in USA; E-mail: [email protected]) which the Earth has two well-separated convective regimes in which mass The seismic tomography images obtained exchanges are limited by break-throughs by Montelli et al. (Science, 2004) from a of slabs and plumes at the 670 km combination of high- and low frequency P discontinuity. delay times show very wide, diapir-like plumes instead of the classical large- head/thin-tail structure, suggesting that Constraints on the Observation of these observed plume 'tails' play an Mantle Plumes Using Global important role in transporting heat through Seismology the 670 km discontinuity. Because high- frequency delay times are proportional to A. Deuss (Dept. Earth Sciences, Univ. of the plume radius R, whereas low Cambridge, Cambridge CB3 0EZ, U.K.; frequency times have a quadratic Tel:+44-1223-337101; E-mail: dependence R2, the finite-frequency [email protected]) analysis used by Montelli is a powerful method to constrain the width of the lower The existence of mantle plumes has been mantle plumes. highly debated in the last few years. Constraints come from many disciplines, New, dedicated, resolution analysis of the but seismology is the only method that plume images confirms this notion and allows us to 'see through' the Earth and shows a number of lower mantle plume provides us with a snapshot of the Earth's segments where both the observed plume deep interior. Here, I will review the radii and velocity anomalies are different tools that are used in global sufficiently well resolved to directly seismology to image mantle plumes and, estimate the heat and volume flux, if the most importantly, their limitations. viscosity and the chemical density difference were known. To deal with the In seismology, two different data types are large uncertainties in those (and several used which have very different sensitivity less influential) model parameters, we and resolution. Body waves are high randomly generate 560,000 models and frequency data which are studied using ray compute plume fluxes. For commonly theory and can be used to image small accepted viscosities of the order of 1022 scale structure. These data have been used Pa s, the plume flux easily exceeds that in global tomography, but also to image deduced from the observed buoyancy flux sharp features such as mantle by an order of magnitude. We conclude discontinuities and they are very suitable that the buoyancy flux may be a severe to study the possible existence of narrow underestimate of the lower mantle flux, plumes. However, data coverage is limited unless the viscosity is of the order 1023 Pa by the distribution of earthquakes and s and/or the plumes suffer loss of receivers and this is a problem in buoyancy through the effects of iron particular for oceanic areas, where most of enrichment. the plumes are located. In addition, it is important to realise that body waves are only sensitive to velocity perturbations.
44 Thus, we do not know if the imaged should be very careful in only using the features are caused by temperature observations within their limitations and it anomalies only, or also compositional might be too early at present to draw any differences. Normal mode data, on the strong conclusions for or against the other hand, are sensitive to both velocity existence of mantle plumes. and density perturbations and can in principle provide us with temperature and composition varations in the mantle. Guided Seismic Waves: Possible However, normal modes only image the Mantle-Plume Diagnostics large scale features of the Earth and may not be able to show narrow plumes. B. R. Julian (U.S. Geological Survey, 345 Middlefield Rd. MS 977, Menlo Park, CA Seismologists have been very succesful in 94025, USA; Tel: +1-650-329-4797; Fax: making global tomographic velocity +1-650-329-5163; E-mail: models using body waves, which in some [email protected]); J R Evans (U.S. models are combined with normal mode Geological Survey, 345 Middlefield Rd. data. Different models have very similar MS 977, Menlo Park, CA 94025, USA; long-period structure and all show large Tel: +1-650-329-4753; Fax: +1-650-329- areas of low velocities in the upper and 5163; E-mail: [email protected]) lower mantle, which are often interpreted as super plumes (see for example Ritsema Seismic waves potentially provide by far it et al., 1999). But, from velocity alone the highest resolution view of the three- we do not know their temperature dimensional structure of the mantle, and perturbation and we have to be careful in the hope of detecting wave-speed interpreting these features. Indeed, the first anomalies caused by mantle plumes has tomographic model showing temperature been a major incentive to the development and composition variations was published of tomographic seismic techniques. recently (Trampert et al., 2004) which Seismic tomography is limited, however, suggested that the super plumes are by the uneven geographical distribution of compositional in origin and dense rather earthquakes and seismometers, which can than thermally bouyant. produce artificial tomographic wave-speed anomalies that are difficult to distinguish In addition, seismologists have used from real structures in the mantle. transition zone discontinuities to determine if their topography agrees with An alternate approach may be possible, plumes going trough the transition zone, because low-wave-speed channels would or not. In particular SS-precursor studies act as waveguides that could transmit show areas of thinner transition structure, seismic waves efficiently over great depth which agrees with a thermal interpretation ranges in the mantle. The same of a hot plume going through the transition phenomenon causes fault zones to act as zone. However, this relies on an waveguides and provides a useful method interpretation of the transition zone for studying faults in the crust. Plume- discontinuities in terms of olivine phase guided waves would be little affected by transitions only and recent studies suggest bends or other geometric complexities in that the phase transitions in garnet and the waveguides (think of a French horn or ilmenite have an important influence on a fiber-optic cable), and their dispersion the transition zone discontinuities as well would make them distinctive on (Deuss & Woodhouse, 2001). This implies seismograms and would provide that we cannot use simple discontinuity information on the size and structure of the topography observations to determine the waveguide. origin of mantle plumes. The main unanswered question is whether Thus, in priciple seismology will be able guided waves in plumes could be excited to provide us with strong contraints on the sufficiently to be observable. Earthquakes observation of mantle plumes. But, we do not occur in the deep mantle, but two
45 other possible sources of excitation can be output of heat from the core of from 0.5 to imagined: (1) shallow earthquakes at or 15 TW. The large uncertainty in near plume-fed hotspots; and (2) coupling geodynamo heat output gives large to seismic body waves in heterogeneous uncertainty in estimates of the degree of regions of the deep mantle. In the first heating from below experienced by the case, downward-traveling guided waves mantle. Nonetheless, there is general transformed to seismic body waves at the agreement that a thermal boundary layer is bottom of the waveguide would have to be present in D", and efforts have been made detected at teleseismic distances. In the to study the detailed seismological second case, upward-traveling guided structure of the lowermost mantle to assess waves generated by teleseismic body the processes occuring in this region. D" waves would be detected on seismometers structure has several characteristics that at hotspots. Qualitative reasoning based indicate that the boundary layer is more on considerations of reciprocity suggests complex than a simple thermal boundary that the signals in these two situations layer. Large-scale patterns of seismic should be similar in size and appearance. heterogeneity exist in D", with two massive low shear velocity provinces A failure to find these guided waves beneath Africa and the south Pacific, both experimentally could mean either that the of which appear to extend upward 500- waveguides (plumes) do not exist or that 1000 km from the CMB. While the excitation mechanisms and/or commonly described as 'superplumes', this seismometer networks are inadequate. appears to be a misnomer. Both regions Distinguishing these two possibilities have very sharp lateral gradients on their would require careful analysis. margins; both have unusual Vs/Vp Anticipated major improvements in velocity ratio perturbations, with much seismic instrumentation, such as the stronger decreases in Vs than in Vp; and EarthScope initiative, make this a there is preliminary evidence that these propitious time to undertake such a regions are anomalously dense. It appears project. that these are not large thermal plumes, but are instead large chemical heterogeneities. Other regions of D" have Is the “D" Region the Source of Mantle relatively high seismic velocities and Plumes? abrupt increases in Vs and Vp about 250 km above the CMB. If relatively low Thorne Lay (Earth Sciences Dept., Univ. temperatures are responsible for the high of California, Santa Cruz, CA 95064, seismic velocities, it is plausible that these USA; Tel: +1-831-459-3164; Fax:+ 1-831- regions involve the post-perovskite phase 459-3074; E-mail: [email protected]) transition that was discovered in 2004. The strong positive Clapeyron slope of The mantle plume hypothesis requires this phase transition predicts existence of at least one boundary layer in destabilization of the boundary layer, but the deep interior that can experience ironically, this involves regions where thermal instabilities that rise to the surface mantle down-wellings are believed to as concentrated upwellings. While the exist, not where plumes are hypothesized. possibility of mid-mantle thermal A summary of these and other boundary layers has not been conclusively seismological observations and the ruled out, there is little direct observational challenges they present to the notion of D" support for their existence. As a result, the as a source of mantle plumes will be D" region in the lowermost mantle, a given. several hundred kilometer thick layer above the core-mantle boundary (CMB), is widely invoked as the putative mantle plume source. A thermal boundary layer is expected above the CMB based on theories of the geodynamo, which predict
46 The Surface of Venus Records Ancient how plumes, antiplumes, and related Impacts, Not Young Plumes endogenic processes might form these old structures have addressed their consistent Warren B. Hamilton (Dept. of Geophysics, circularity, morphology, and Colorado School of Mines, Golden, CO, superpositions. Four of the youngest of the 80401, USA; E-mail: old structures have circular rims 800 to [email protected]) 2000 km in diameter and, if indeed of impact origin, are, by analogy with dated The assumptions and speculations that late-stage lunar Imbrium Basin, about 3.9 comprise terrestrial plumology have Ga. Unearthlike broad, yet very low, become ever more convoluted as their volcanoes in, and overflowing ("volcano- predictions have been falsified. Although corona hybrids"), ancient impact basins do Venus has no features like those attributed not resemble terrestrial volcanoes or large to terrestrial plumes, plumology was igneous provinces and likely are of impact exported to Venus, even before high- melts, as are pancake-spreading magma- resolution Magellan radar imagery became lake(?) plateaus 1000-2000 km in available in 1991-94, to explain an diameter. Venus apparently preserves an assumed need for loss of heat by a planet eroded and partly buried landscape of late- that obviously lacks plate tectonics. stage planetary accretion because (as is Venusian, like terrestrial, plumology also consistent with all available geophysical became dogma by repetition and self- and compositional data) its mantle is too citation, but its specific rationalizations are cold and dry for earthlike circulation. quite different. Plumology may be inapplicable also to The Venusian features conventionally Mars. Half of the surface is of eroded attributed to plumes have the impact-saturated rubble, and most of the characteristics expected of variably eroded other half has a similar landscape buried and buried impact structures and impact- beneath thin sediments. Rationales that melt constructs, and likely record late- local volcanic highlands were formed by stage planetary accretion older than 3.9 plumes are strained, and their magmatism Ga. Accepted by all as of impact orign are may be of impact origin. 1000 small (maximum rim diameter 275 km, but mostly <50 km) little-modified impact craters. The maximum age of these Venus’ Many Circles: Extraterrestrial is widely assumed to be <1 Ga although Clues for the Great Plume Debate size-frequency distribution, and consideration of dense-atmosphere effects V. L. Hansen (Dept. of Geological and bolide strengths, permit much greater Sciences, Univ. of Minnesota Duluth, age. These are gradationally distinguished, Duluth, MN 55812, USA; Tel: +1-218- via increasing surface smoothing and 726-6211; Fax: +1-218-726-8275; E-mail: degradation in highlands and burial by [email protected]) sediments in lowlands, from several thousand older circular structures Venus has long been considered Earth’s commonly assumed to be endogenic. (800 sister based on solar distance, size, or so of the large old structures are density, and presumed composition. Yet, assigned to pigeonholes such as "coronae", like many siblings, Venus and Earth took whereas many other large ones, and nearly quite different evolutionary paths. Venus all small ones, are ignored in conventional lacks evidence of plate tectonics, marked analysis.) Many of these old structures are on Earth by linear global features. Rather, much larger than any of the young impact Venus displays a wide range of circular structures, and most have the rimmed- features: impact craters (1-270 km), basin morphology, and often the cookie- coronae (60-2600 km, 200-km mean), cutter superpositions, expected of impact volcanic rises (1500-2400 km), and crustal structures. None of the scores of plateaus (1500-2600 km); the latter three conflicting published conjectures as to are variably interpreted as the surface
47 expressions of diapiric structures, be they Quetzalpetlatl), best explained by a thermal (i.e., plume), compositional, or thermal plume hypothesis. undefined. Some workers suggest that all coronae represent impact features, and others propose that all of Venus’ circular Meteorite Impacts as Triggers to LIPs features resulted from bolide impact. and Hotspots
Venus provides an excellent opportunity A. P. Jones (Dept. of Earth Sciences, to test plume and alternate plume Univ. College London, Gower St., hypotheses. 1) Venus is expected to have a London, WC1E 6BT, UK; Tel: +0207- heat budget similar to Earth, and as such 679-2415; Fax: +0207-679-4166; E-mail: requires a mechanism by which heat is [email protected]); G D Price (Dept. transferred from core (or mantle) to crust. of Earth Sciences, Univ. College London, 2) Venus lacks plate-tectonic processes, Gower St., London, WC1E 6BT, UK; Tel: thus plate tectonic-related alternate +0207-679-7083; Fax: 0207-679-4166; E- hypotheses are not viable. 3) Venus mail: [email protected]) preserves an excellent surface record (albeit remote) unmodified by erosion and The combined effects of impact melting, plate-tectonic processes. 4) Widely enhanced by sub-crater decompression available NASA Magellan data provide melting in computer simulations of ~200 3D surface views at high resolution, km craters produce characteristic high allowing first-order geologic analysis, melt volumes of approximately 106 km3 coupled with subsurface constraints based derived from the uppermost ~150 km of on gravity-topography data. 5) Venusian the Earth’s mantle. The coincidence ‘fieldwork’ only requires computer access, between expected frequency of such making data collection, data analysis, and impact events combined with the hypothesis testing accessible to all. similarity in magma volumes of large igneous provinces (LIPs) suggests that Morphologically, structurally, and large meteorite impacts may be capable of geologically distinct features on Venus triggering LIPs and mantle hotspots from a likely formed by a variety of mechanisms. point source which is subsequently buried, Volcanic rises appear most consistent with as eg: oceanic plateaus. There are two plume-lithosphere signatures based on main aspects to test this idea; firstly what size, morphology, ADC (apparent depth of are the distinctive macroscopic criteria compensation) and geologic history. predicted from an impact model, and Although we previously argued that secondly, how may these be recognised in crustal plateaus (marked by shallow ADC) the geological record of the Earth. We represent the surface signature of thermal illustrate the impact melting potential with plumes on thin lithosphere, recent detailed a detailed model for the Ontong Java structural analysis of plateau surfaces Plateau, and use this as a discussion point appears more consistent with progressive to assess predicted geochemical, crystallization of huge lava ponds, perhaps geophysical and geological characteristics the result of bolide impact on thin of such a process. These might include; lithosphere and massive partial melting of geotherms and melting; mixed volcanic the mantle. Coronae seem to represent at versus impact signals; paucity of shock least three types of genetically unrelated relative to thermal features; distal deposits, features: 1) domical features (~60-500 km) spherule beds, ejecta and ash layers; where marked by radial fractures and extensive to look?; proximal-distal-global markers; flows likely representing (compositional) ocean drill cores and sequence startigraphy diapiric structures; 2) circular topographic etc. lows (~100-400 km) that lack radial fractures and extensive associate flows, possibly recording bolide impact on locally weak crust; and 3) huge features (~800-2600 km; e.g., Artemis, Hengo,
48 Impact Induced Martian Mantle Tuesday Poster Sessions Plumes: Implications for Tharsis Patterns of Rhyolitic Volcanism in the C. C. Reese (Dept. of Earth and Planetary Path of the Yellowstone Hot Spot Sciences, Washington Univ., Saint Louis, MO 63130-4899, USA; Tel: +1-314-935- L. A. Morgan (U.S. Geological Survey, 4921; Fax: +1-314-935-7361; E-mail: Federal Center, Box 25046, MS 966, [email protected]); V S Solomatov (Dept. Denver, CO 80225, USA; E-mail: of Earth and Planetary Sciences, [email protected]); Kenneth L. Pierce Washington Univ., Saint Louis, MO (USGS, Northern Rocky Mountain 63130-4899, USA; Tel:+1-314-935-7882; Science Center, Bozeman, MT, USA; E- Fax:+1-314-935-7361; E-mail: mail: [email protected]); William C. [email protected]) McIntosh (New Mexico Bureau of Mines and Mineral Resources, 801 Leroy Place, Tharsis province is a major center of Socorro, NM 87801, USA; E-mail: Martian volcanism and [email protected]) tectonismcharacterized by large gravity and topography anomalies. Initial The Snake River Plain-Yellowstone development of Tharsis occurred rapidly Plateau (SRP-YP) developed over the last during the earliest epoch of Mars history. 16 Ma as a bimodal volcanic province in The distribution of volcanic and tectonic response to the southwest movement of activity associated with Tharsis led to the the North American plate over a fixed suggestion that the planform of Martian melting anomaly. Volcanism along the mantle convection is dominated by a SRP-YP province is dominated by single, large, long-lived upwelling. The eruptions of explosive high-silica rhyolites origin of the upwelling is debated. One and represents some of the largest hypothesis is that the upwelling is a eruptions known (Mason and others, thermal plume originating at the core- 2004). Basaltic eruptions represent the mantle boundary similar to a terrestrial final stages of volcanism, forming a thin plume but much larger. An alternative cap above voluminous rhyolitic deposits. hypothesis is that the upwelling is Volcanism progressed episodically from associated with late stage evolution of a southwest to northeast along the plain in local Martian magma ocean produced by a successive volcanic fields comprised of large impact during planetary formation. nested caldera complexes. Most major The physical processes controlling caldera-forming eruptions within a evolution of an impact induced melt particular field are separated by about 0.2 region suggest that, within the to 1 m.y., similar to the present-day uncertainties of various parameters, in situ Yellowstone Plateau volcanic field. Major crystallization to a partially molten solid ignimbrite eruptions from volcanic fields state and a relatively long adjustment are separated in time by as much as 2 m.y. timescale are a possibility. Fully three from the eruption of the youngest caldera- dimensional spherical shell simulations of forming event in one field to the next thermochemical mantle convection caldera-forming event in the younger, indicate that such an initial state generates adjacent field. Spatially, adjacent volcanic a localized mantle upwelling - an impact fields are separated by 50-150 km from induced plume. A scenario consistent center to center. Passage of the North with observational constraints on the American plate over the melting anomaly volume of magmatic material and at a particular point in time and space emplacement rate is suggested. results in uplift, regional tectonism, massive explosive eruptions and caldera subsidence, resurgence in selected areas, and finally basaltic volcanism and subsidence (Pierce and Morgan, 1992; Pierce and others, 2002). Smaller rhyolitic
49 lava flows occur during both pre- and Secular variations in the pressure of mafic post-caldera phases. intrusions emplaced into the basal portion of flood basalt successions provide The Heise volcanic field in the eastern constraints on the timescale of volcanic SRP, Idaho immediately to the southwest built-up and subsidence. Because mafic of the Yellowstone Plateau volcanic field, intrusions solidify in a few hundred provides a definitive example of the SRP- thousand years or less, changes in the YP volcanic field cycle. Five large- pressure of such intrusions provide succint volume rhyolitic ignimbrites constitute a constraints on the duration of volcanism time-stratigraphic framework of late and is, typically, on a finer timescale than Miocene to early Pliocene volcanism in radiometric age determinations. As an the eastern SRP. Field relations and high- example, we have determined the pressure precision 40Ar/39Ar age determinations of the formation of fluid inclusions in establish that four of these regional quartz from granophyre bodies, that is cm- ignimbrites were erupted from the Heise to metre-sized pockets dominated by volcanic field forming the framework of intergrown quartz and alkali feldspar, in the Heise Group; the source of the fifth the Skaergaard Intrusion emplaced into the unit is outside the Heise field. The Heise basal portion of the East Greenland flood units include the Blacktail Creek Tuff basalt succession. The results demonstrate (6.62± 0.03 Ma), Walcott Tuff (6.27± 0.04 that the pressure for different zones of the Ma), Conant Creek Tuff (5.51± 0.13 Ma), intrusion, corrected to the pressure at the and Kilgore Tuff (4.45± 0.05 Ma; all position of the roof in contact with the errors reported at ± 2 sigma). Facies and overlying volcanic rocks, increased from magnetic fabric data indicate that these 0.8 ± 0.5 to 3.0 ± 0.5 kbar during units erupted from separate calderas, each solidification. This pressure increase can with a set of discrete vents. Major be explained by the emplacement of 7.3 ± eruptions in this province are separated in 1.5 km of flood basalts over the intrusion. time and space while less voluminous The cooling model of Norton and Taylor minor rhyolitic eruptions fill in the (Journal of Petrology 20, p. 421-486, spectrum. Large caldera forming 1979) shows that the time elapsed between eruptions and development of individual the formation of the low and high pressure volcanic fields are related to rate of rocks in the intrusion was c. 260,000 movement of NA plate, amount of years. This is, however, most likely a material available in crust for melt, the minimum value because they assumed a degree of depletion which has occurred in constant pressure of 1 kbar. We therefore the crust, and flow dynamics in the upper surmise that 7.3 ± 1.5 km of flood basalts crust. were emplaced over the Skaergaard Intrusion in less than 3-400,000 years. These flood basalts erupted immediately Timescales of Flood Volcanism landward of anomalously thick oceanic Recorded by Pressure Variations in crust of the Greenland-Iceland Ridge. The Coeval Mafic Intrusions: A Fluid volcanic productivity over the Skaergaard Inclusion Study of the Skaergaard Intrusion is similar to that calculated for Intrusion, East Greenland the oldest and most landward seafloor of the Greenland-Iceland Ridge (Holbrook et C. Tegner (Dept. of Earth Sciences, Univ. al., Earth and Planetary Science Letters of Aarhus, C.F. Mollers Alle 110, DK- 190, p. 251-266, 2001). 8000, Aarhus C, Denmark; Tel: +45- 89422530; Fax: +45-89422525; E-mail: [email protected]); R. B. Larsen (Dept. of Geology and Mineral Resources Engineering, Norwegian Univ. of Science and Technology, N-7491 Trondheim, Norway; Tel:+47 73594968; E-mail: [email protected])
50 Global Gravity, Bathymetry, and the comparing predicted and observed Distribution of Submarine Volcanism bathymetry in the vicinity of each feature through Space and Time in the Wessel (2001) data base, we have been able to estimate Te at > 10,000 A. B. Watts (Dept. of Earth Sciences, submarine volcanoes. We present here the Univ. of Oxford, Oxford, OX1 3PR, UK; results of our analysis and examine its Tel: +01865-272032; E-mail: implications for the tectonic setting, age, [email protected]); D T Sandwell and distribution of submarine volcanism (Scripps Institution of Oceanography, La through space and time. Jolla, CA, 92093-0225, USA; Tel: +1-858- 534-7109; E-mail: [email protected]); W H F Smith (Laboratory for Satellite Duration and Timing of the Deccan Altimetry, NOAA, Silver Spring, MD, CFBP 20910-3282, USA; Tel: +1-301-713-2857; E-mail: [email protected]); P M. Widdowson (Dept. of Earth Sciences, Wessel (Dept. of Geology & Geophysics, The Open Univ., Walton Hall, Milton SOEST, Univ. of Hawaii, Honolulu, HI Keynes, MK7 6AA, UK; Tel: +01908 96822, USA; Tel: +1-808-956-4778; E- 652986; Fax: +01908-655151; E-mail: mail: [email protected]) [email protected]); S.P. Kelley (Dept. of Earth Sciences, The Open Univ., The seafloor is characterised by numerous Walton Hall, Milton Keynes, MK7 6AA, seamounts and oceanic islands which are UK; Tel: +01908-653009; Fax: +01908- mainly volcanic in origin. Relatively few 655151; E-mail: [email protected]) of these features (<~ 0.1%) have been dated and so there is little known about Previous dating studies of the Deccan either their tectonic setting or age. One CFBP have concentrated upon the ‘classic’ parameter that is sensitive to whether a sections exposed along the Western Ghats submarine volcano formed on, near, or far escarpment. These data have provided from a mid-ocean ridge is the elastic crucial preliminary insight regarding the thickness, Te, which is a proxy for the timing and duration of the CFBP, and have flexural strength of the lithosphere. Most indicated a geologically rapid eruption at, previous estimates of Te are based on or near to, the Cretaceous-Tertiary using the bathymetry to calculate the boundary (KTB; 65.0 0.1 Ma). This gravity anomaly for different values of Te argument has remained pivotal to and then, selecting the Te that best fits the arguments citing Deccan volcanism as a observed anomaly. The problem with this primary or contributory cause to global approach is that, for most seamounts, biotic extinction, and to the wider models bathymetric data are usually limited to of CFBP generation. However, whilst single-beam echo-sounder data acquired ‘classic’ sections provide easy access to along individual ship tracks. These data do parts of the Deccan succession, they do not define the shape of the seamount. not offer an entire chronological record of Therefore, gravity modelling usually eruptions. assumes either that the seamount has infinite extent perpendicular to a ship track High-precision 40Ar/39Ar ages,taken or is circular in planform. Either of these within a well-defined and detailed assumptions may bias Te. The gravity stratigraphical context are presented for anomaly derived from satellite altimeter the Deccan. Data from the data, in contrast, has furnished information stratigraphically highest and lowest basalts on the shape of seamounts, as has been bracket the entire Deccan lava succession, shown most recently by Wessel (2001). and thus provide a comprehensive estimate We have therefore used a technique, for the timing and duration of this volcanic developed originally by Smith and episode. Analyses were performed on Sandwell (1994) and Lyons et al. (2000), more than 20 selected samples, using to predict bathymetry directly from the incremental laser heating on plagioclase satellite-derived gravity anomaly. By
51 separates. Results are reported relative to The pilot experiment used differential the GA1550 biotite standard. pressure gauges (DPG) as sensors. For the deep-ocean environment off Hawaii, it The new data suggest a duration of 2 – 3 remains to be seen whether ocean bottom Myr for the eruption of the main Deccan seismometers can reach similar signal-to- province (MDP). The earliest flows (66.8 noise levels. 0.4 Ma) occur in the north-western corner of the MDP where they were erupted onto the Late Cretaceous, dinosaur-bearing, Comprehensive Imaging of the Eifel Lameta beds. Importantly, these basal Plume, Central Europe lavas indicate that the onset of flood volcanism began some 1.5 – 2 Myr before J. R. R. Ritter (Geophysical Institute, the (Cretaceous-Tertiary boundary (KTB). Univ. of Karlsruhe, Hertzstr. 16, 76187 The final Deccan flood basalt eruptions Karlsruhe, Germany; Tel: +49-(0)721- are of Paleocene age (63.6 0.4 Ma) and are 608-4539; Fax: +49-(0)721-71173; E- preserved in the south-western Deccan. mail:[email protected]); The onset of Deccan CFB volcanism Eifel Plume Team began in the north of the province during the late Maastrictian, and rapidly reached Volcanic eruptions occurred in the Eifel an acme of eruption during Chron 29R. mountains in the western part of Germany This eruptive acme continued across the since Mesozoic time. Two new volcanic KTB and into the Danian, with fields evolved in the last 600 ka and the successively younger lava fields building latest eruptions occurred only 11-12 ka on the southern flank of the evolving B.P. At the same time strong uplift (up to Deccan volcanic edifice. 250 m in 600 ka) occurred in the region. To study the deep structure of the Eifel region 10 European institutions shared The SWELL Pilot Experiment off their facilities to operate a network with 84 Hawaii - What Can We Learn About permanent and 158 mobile stations the Hawaiian Hotspot from Surface including 32 broadband instruments Waves? during an 8 months field experiment. The network had a 500 km by 500 km aperture G. Laske (IGPP, Scripps Institution of and the mobile stations were deployed Oceanography, UC San Diego, 9500 between Nov-3-1997 to June-23-1998. Gilman Dr., La Jolla, CA 92093-0225, USA; Tel: +1-858-534-8774; Fax: +1-858- The 3D P-wave model contains a column- 534-5332; E-mail: [email protected]) like low-velocity anomaly (LVA, -1% to - 3%) in the upper mantle underneath the The SWELL Pilot Experiment was one of Eifel volcanic fields reaching down to at the first long-term deployments of ocean least 400 km depth. The S-wave model has bottom seismic equipment to record long- a prominent LVA of up to -5% in the period surface waves to a period of 90s. upper 100 km of the mantle. At 200 ± 50 The superior quality of the dataset allowed km depth there is no clear S-wave velocity us to image the southwestern edge of the anomaly. Below the S-wave velocity Hawaiian swell. A surprisingly deep reduction is about -1%, and it extends at situated low-velocity anomaly suggests least to the transition zone. Receiver that the cause for the swell relief is sub- functions also find the positive S-wave lithospheric. We have also embedded the anomaly at 200 ± 50 km depth. The 410 dataset in our global database. Initial km discontinuity is depressed by 20±5 km research suggests that the low-velocity indicating an increased temperature, while anomaly is not centered beneath the the 660 km discontinuity appears to be Hawaiian island chain but slightly off- unperturbed. The top of the low-velocity center to the southwest where the swell anomaly is found at about 50-60 km depth bathymetry is shallower than in the by receiver functions as well as Rayleigh northeast. and Love wave dispersion models.
52 Teleseismic P-wave attenuation shows a beneath the Kaapvaal craton and becomes strong absorption anomaly in the broader beneath Tanzania. At shallower lithosphere and a weaker anomaly in the depth and in the mantle transition zone, mantle. The lithospheric damping anomaly the low-velocity anomaly becomes is interpreted as a magmatic intrusion zone localized, forming a cylindrical feature that scatters the wave amplitudes. In the beneath Tanzania, which is consistent with asthenosphere temperature-induced solid- an anomalously thin mantle transition in state anelastic attenuation is assumed. A the area. study of the teleseismic shear-wave splitting (SKS) indicates a local anomaly Our observations have important in the Eifel which may be due to plume- implications for the deep earth engine. related asthenospheric flow. The P- and S- Models of mantle convection driven by a velocity anomalies in the lithosphere and combination of basal and internal heating upper asthenosphere can be explained by predict a transition in the shape of a an increase of temperature by about 100- buoyant upwelling from tabular to 150 K plus 1% melt. In the lower cylindrical near the basal boundary layer. asthenosphere, above the transition zone, The ridge-like low-velocity anomaly in the the excess temperature of the plume is at upper part of the lower mantle beneath least 70 K, because the velocity anomalies southern Africa may thus represent in the tomography model are upwelling from the instability of a underestimated. boundary layer in the mid-mantle, possibly above the top of the African superplume. Furthermore, the ridge-like low-velocity Seismic Evidence for a Lower Mantle anomaly beneath southern African cratons Origin of the Tanzania Hotspot suggests that volcanism in eastern Africa is not the result of passive upwelling Yang Shen (Graduate School of driven by plate separation but mainly the Oceanography, Univ. of Rhode Island, consequence of deep, buoyant mantle Narragansett, RI 02882, USA; Tel: +1- convection. 401-874-6848; Fax: +1-401-874-6811; E- mail: [email protected]); Shu-Huei Hung (National Taiwan Univ.; Tel: +1-886-2- Identifying the Origin of the Newberry 2364-9352; Fax: +1-886-2-2363-6095; E- Hotspot Track mail: [email protected]) Mei Xue (Dept. of Earth and Planetary Global seismic tomographic imaging has Science, Univ. of California, 289 McCone shown a large and coherent low-velocity Hall, Berkeley, CA 94720-4767, USA; anomaly in the lower part of the mantle Tel: +1-510-643-5450; Fax: +1-510-643- beneath Africa. It has been suggested that 9980; E-mail: this anomalous structure may be [email protected]); Richard responsible for the unique geological M. Allen (Dept. of Earth and Planetary history of Africa. But the link between the Science, Univ. of California, 279 McCone structure in the deep mantle and surface Hall, Berkeley, CA 94720-4767, USA; tectonics remains unclear. In this study we Tel: +1-510-642-1275; Fax: +1-510-643- carry out tomographic inversions for the 5811; E-mail: seismic velocity structure beneath [email protected]) southern Africa, utilizing “banana- doughnut” traveltime sensitivity kernels of Located in the northwestern United States, body waves recorded by broadband the Newberry hotspot track consists of a temporary and permanent seismic stations sequence of age-progressive volcanic in the region. Preliminary results show a domes and lava flows, showing a ridge-like low-velocity feature in the upper monotonic age progression from east to part of the lower mantle (900 – 1200 km west ending at the Newberry Caldera. The depth) beneath eastern Africa. The low- northwest trending Newberry track velocity feature is 300-400 km wide crudely mirrors the better-known
53 Yellowstone hotspot track. While located the mantle structure beneath the hotspot on the North American Plate, the from teleseismic body waves recorded by Newberry track cannot be the product of 5 portable seismic stations on the Azores plate motion over a stationary mantle islands and the Global Seismic Network source as its orientation is oblique to plate station CMLA. Unlike in conventional motion. Instead three end-member tectonic tomography based on ray theory, the three- models have been proposed: (1) dimensional sensitivity kernels of body subduction counterflow, (2) gravitational wave travel times are used to account for flow along lithospheric topography, and waves front healing, scattering, and other (3) lithospheric faulting. Our preliminary diffraction effects of realistic seismic SKS splits imply anisotropy primarily waves. Our inversion from 186 P-wave comes from the asthenosphere and that the relative travel times revealed a pronounced anisotropy orientation does not vary with low velocity anomaly in the shallow depth beneath the Newberry track. The mantle (less than 200 km depth) along the first order observation is that the SKS Azores islands and beneath the center of splits are not aligned with the Newberry the Azores plateau (38.5N, 29W). From hotspot track as the subduction ~250 km to at least the middle of counterflow model or the gravitational transition zone, the low velocity anomaly flow model would require, indicating forms a plume-like column beneath either the splits are not sensitive to mantle northeast of Terceira (~200 km away from flow oriented along the track or the track the center of plateau), bending toward is not the product of asthenospheric flow. southwest in the shallow mantle. These Our splitting observations strongly argue results are consistent with other for one layer of anisotropy, making the geophysical and geochemical observations second explanation preferable. This leaves and provide seismic evidence for the the lithospheric faulting model as the most proposed hotspot-ridge interaction model, likely causal process. in which the magma plumbing system bends towards the Azores triple junction, supplying melt preferentially Upper Mantle Structure Beneath the southwestward along a lateral Azores Hotspot From Finite Frequency sublithospheric channel. Seismic Tomography
Ting Yang (Graduate School of Uplift and Rifting on Venus: Role of Oceanography, Univ. of Rhode Island, Plumes South Ferry Rd., Narragansett, RI 02882, USA; Tel: +1-401-874-6828; E-mail: Donna M. Jurdy (Dept. of Geological Sci., [email protected]); Y Shen (Graduate Northwestern Univ., Evanston, IL 60208- School of Oceanography, Univ. of Rhode 2150, USA; Tel: +1-847-491-7163; Fax: Island, South Ferry Rd., Narragansett, RI +1-847-491-8060; E-mail: 02882, USA); S-H Hung (Dept. of [email protected]); Paul R. Geosciences, National Taiwan Univ., Stoddard (Dept. of Geology and Taipei, Taiwan) Environmental Geosciences, Northern Illinois Univ., DeKalb, IL 60115, USA; The Azores archipelago represents one of Tel: +1-815-753-7929; Fax: +1-815-753- the classic hotspots that interact with the 1945; E-mail: [email protected]) mid-ocean ridges. While geochemical studies and a variety of geophysical Venus, although Earth's "twin" in size and observables have provided clear evidence composition, lacks plate tectonics. for the influence of the deep mantle However, Venus does show evidence of materials on the formation of the plateau, recent uplift as well as both tectonic and the seismic structure beneath Azores has volcanic activity. The planet's most never been imaged directly at regional and pronounced geoid and topographic highs, local scales. Here we construct the first Atla and Beta Regiones, lie at nodal points high-resolution P-wave velocity models of in a network of chasmata. The chasmata,
54 linear to arcuate troughs, extend 1000's of on Venus, and in turn to link the process to km and may be sites, of current rifting. the global system of rifting. Great circle arcs fit the chasmata system at the 89.6% level. These arcs total 54,464 km in length, when corrected for Venus' Wendesday Oral Sessions smaller size this comes within 2.7% of the 59,200 km estimate for Earth's spreading Plumes and Uplift ridges. Hundreds of coronae, enigmatic circular features, adorn Venus' surface, Andy Saunders (Dept. of Geology, Univ. ranging in diameter from 60-2600 km with of Leicester, Leicester, LE1 7RH, U.K; a median of 250 km. Coronae have been Tel: +44-116-2523923; Fax:+44-116- modeled and attributed to diapiric- 2523918; E-mail: [email protected]) upwellings; however, ancient impacts have been proposed as an alternative Hotspots are so-called for good reason. explanation. The distribution of coronae Not only is there excess magmatism, but over Venus' surface is highly non-uniform, there is also increased heat flow, and with an excess near chasmata. Indeed, surface uplift due to thermal and dynamic some coronae have rims that intertwine buoyancy. Surface uplift persists where a with chasmata walls, suggesting hotspot is active e.g. Hawai'i and Iceland contemporaneous formation. The raised and decays away from the hotspot, areas of both of Atla and Beta Regiones consistent with predictions from plume are ringed by numerous coronae, but models. During the early stages of hotspot neither has coronae at or near their crests (and plume) development, where a large at the highest geoid levels. Using an igneous province may be formed, uplift evolutionary classification of coronae, we may be both rapid and large. Here, model use inferred coronae stage to assess the predictions are more contentious, but relative activity at Atla and Beta, and we uplifts of the order of one or more conclude that Atla has been active more kilometres, radiating over a distance of recently. several thousand kilometres, are expected. Are they found? Venus' craters, unlike coronae, have nearly a random distribution over the planet's In this session specific case studies will be surface, with a slight, but statistically used to illustrate the consistency of the significant deficit close to chasmata. A plume model with predicted and observed minority of craters are modified, but an uplift patterns. I shall focus specifically excess of the tectonically or volcanically on the Siberian LIP which, others have disturbed craters occurs near chasmata and argued, has no associated uplift and, by the uplifted regiones. Impact craters on implication, no associated plume. This Venus can independently determine suggestion has been made despite the relative age, degree of modification, and rapid eruption of between 3 and 4 million reorientation of an area. Low crater cubic kilometres of basalt, making it the densities exist with coronae and largest continental flood basalt province surprisingly, even out to 4 radii, whereas on Earth. I shall summarise data which we would expect locally higher densities indicate that the main uplift occurred not of craters if coronae were themselves very in the region of Norilsk, but in the West old impact features. Furthermore, Siberian Basin-Khatanga Trough region, modification of craters adjacent to and to the northwest of Norilsk. Uplift could within coronae argues for the existence of have exceeded 1 km, but unfortunately we long-term tectonic and volcanic activity, have no way of determining the total as opposed to formation of coronae by extent of subaerial uplift in this region. instantaneous, ancient impacts. We attempt to relate coronae formation to Alternative models, such as the EDGE the existence of plumes on Venus. Our model, have been used to explain the goal is to clarify the mechanism of uplift apparent lack of uplift associated with LIPs generally, and the Siberian Traps in
55 particular. However, this is arguably false before and peaks coeval with Europe- logic. Any realistic EDGE model, where Greenland break-up and associated hot mantle flows from beneath thick development of volcanic passive margins, cratonic lithosphere into an adjacent so that uplift and magmatism are probably region with thin lithosphere, and where the related. Significant uplift of Britain and mantle is sufficiently hot to generate the Ireland over 1000 km away from the large volumes of magma, would cause break-up zone means that plate boundary uplift in the overlying lithosphere. Thus, processes such as edge-driven convection ironically, the magnitude of uplift cannot alone have been responsible for predicted for the EDGE model could be uplift. Phase 2 uplift of Britain and Ireland similar to that predicted for a plume. is therefore best interpreted as the south- eastern quadrant of an uplifted swell, which may have had a diameter of around Uplift Associated with the North 2000-3000 km and a maximum amplitude Atlantic Igneous Province of over 2 km. In both the first and second uplift phases, onset of major uplift Stephen M. Jones (Dept. of Geology, preceded maximum uplift by at most 1 Trinity College, Dublin 2, Ireland; Tel: Myr and possibly by significantly less. +35316081235; Fax: +35316711199; E- These rapid timescales can be explained mail: [email protected]) by injection of hot mantle into a low viscosity asthenosphere channel. The magnitude, planform extent and rates of uplift related to North Atlantic Igneous Province (NAIP) activity are estimated Surface Responses to Mantle Plume: using a combination of stratigraphical, Sedimentation and Lithofacies sedimentological and paleontological Paleogeography in SW China Before information from outcrop and well logs and After the Emeishan Flood across Britain, Ireland and the surrounding Volcanism continental shelf. Two principal phases of uplift are recognized. The first uplift phase Y. G. Xu (Key Laboratory of is associated with emplacement of the Geochronology and Geochemistry, British part of the NAIP (61--58 Ma) and Guangzhou Institute of Geochemistry, is relatively limited in both spatial extent Chinese Academy of Sciences, 510640 (few 100 km) and magnitude (few 100 m). Guangzhou, RP China; Tel: +86-20- A significant proportion of this uplift was 85290109; Fax: +86-20-85290130; E- permanent, associated with igneous mail: [email protected]); B He, Key addition to the crust in the vicinity of the (Laboratory of Geochronology and surface outcrop. These observations can be Geochemistry, Guangzhou Institute of explained by upwelling of relatively small Geochemistry, Chinese Academy of volumes of unusually hot and/or fertile Sciences, 510640 Guangzhou, RP China; mantle directly beneath the region. A Tel: +86-20-85290133; Fax: +86-20- second phase of more widespread transient 85290130; E-mail: [email protected]) uplift was initiated in the latest Paleocene (around 56 Ma), peaked coeval with the The Middle Permian Maokou limestone Late Paleocene Thermal Maximum event that immediately underlies the Emeishan (around 55 Ma) and decayed through the flood basalts in southwest China are Eocene. Sedimentary basins adjacent to variably thinned/eroded. Its isopachs the west of Britain and Ireland were delineate a subcircular uplifted area, uplifted by up to 1 km, while regions 500 suggesting a rapid, kilometer-scale crustal km to the SE were uplifted by less than doming prior to the Emeishan volcanism. 100 m. These large amounts of uplift Unusual depositions of Permian age are require unusually hot mantle beneath the also present in this region. Specifically, plate and the lack of local magmatic carbonate gravity flows, submarine incised activity means that this hot mantle was canyon fillings were developed along the moving laterally. Phase 2 uplift begins just western margin of the postulated uplifted
56 area, and rifting trenches, alluvial fan The widely accepted mantle plume model deposits were formed on the eastern postulates that (i) the currently margin at the boundary between the inner volcanically active Réunion Island, in the and intermediate zones. These deposits all Indian Ocean, is fed by the narrow “tail” rest on the Maokou Formation and are in of a mantle plume that rises from the core- turn covered by the Emeishan basalts, mantle boundary, (ii) the Deccan implying the synchronism between crustal continental flood basalt province, of India, uplift and depositional events. They likely originated from the “head” of the same resulted from differential uplift of the plume during its early eruptive phase near inner zone relative to other zones. the end of the Cretaceous, and (iii) the Comparison of paleogeography lithofacies Lakshadweep-Chagos Ridge, an important before and after the Emeishan flood linear volcanic ridge in the Indian Ocean, volcanism highlights the determinant role is a product of the plume. It is not of mantle plume activity in the geologic generally appreciated, however, that this evolution in SW China and furthermore so-called “classic” case of a plume reveals a two-phase crustal uplift in this contradicts the plume model in many region. The rapid (< 3 M.y.), differential ways. For example, there is little erosion of the Maokou Formation was petrological evidence as yet that the likely related to plume-induced dynamic Deccan source was “abnormally hot”, and uplift. This uplift was apparently followed the short (~1.0-0.5 m.y.) duration claimed by subsidence resulting in deposition of by some for the eruption of the Deccan is the marine clastic rocks sandwiched in conflict with recent Ar-Ar age data that between the basalts and Maokou suggest the total duration to have been at Formation in the east, and submarine least ~8 m.y. The Deccan CFB was basalts along the western margins of the associated with the breakup of the province. A second phase uplift, attributed Seychelles microcontinent from India. to underplating of plume-derived melts at Geological and geophysical data from the the crust-mantle boundary, is characterized Deccan provide no support to the plume by a plateau-type, prolonged (~ 45 Ma) model and arguably undermine it uplift and was responsible for the altogether. The interplay of several formation of the Chuandian “old land”. intersecting continental rift zones in India Prevolcanic lithospheric doming also is apparently responsible for the roughly provides a new framework to evaluate the circular outcrop of the Deccan. The geology, geochemistry, and geophysics of Lakshadweep-Chagos Ridge, and the the Emeishan LIP. Systematic spatial islands of Mauritius and Réunion, are variations are observed across the domal located along fracture zones, and the structure in the distribution and thickness apparent systematic age progression along of clastic and carbonate sediments, the the Ridge may be a result of southward extent of erosion, thickness, and chemistry crack propagation through the oceanic of volcanic rocks, and the crust-mantle lithosphere. This idea avoids the problem structure. All these features, which are of a 10-degree palaeolatitude discrepancy best explained by a mantle plume, may be which the plume model can only solve used to track older plume sites in the with the ad hoc inclusion of mantle roll. geologic record. Published Ar-Ar age data for the Lakshadweep-Chagos Ridge basalts have been seriously questioned, and The Deccan Beyond the Plume geochemical data suggest that they likely Hypothesis represent post-shield volcanism and so are unsuitable for hotspot-based plate H C Sheth (Dept. of Earth Sciences, reconstructions. “Enriched” isotopic ratios Indian Institute of Tech. Bombay, Powai, such as higher-than-N-MORB values of Bombay 400076 India; Tel: + 91-22- 87Sr/86Sr, observed in basalts of the 25767264; Fax: +91-22-25723480; E- Ridge and the Mascarene Islands may mail: [email protected]) mark the involvement of delaminated enriched continental mantle instead of a
57 plume. High values of the 3He/4He ratio profiles suggest that all the volcanic also do not represent a deep mantle edifices of the GVP are the product of a component or plume. The three Mascarene single, long-lasting, stable mantle melting Islands (Mauritius, Réunion, and anomaly (i. e., the Galápagos hotspot, Rodrigues) are not related to the Deccan GHS), in conjunction with a regular but reflect the recent (post-10 Ma) process of ridge edification, which has tectonic-magmatic development of the been active for at least the last 20 m.y. The African Plate. I relate CFB volcanism to velocity-derived crustal density models continental rifting, which often (but not account for the gravity and depth always) evolves into full-fledged sea-floor anomalies considering uniform and normal spreading. I ascribe the rifting itself not to mantle densities, indicating that the ridges mantle plume heads but to large-scale are isostatically compensated at the base plate dynamics themselves, possibly aided of the crust. A remarkable feature that is by long-term thermal insulation beneath a systematically observed is an overall supercontinent which may have surface anticorrelation between lower crust effects similar to those predicted for velocities (and densities) and total crustal “plume incubation” models. Non-plume, thickness. A 2-D steady-state mantle plate tectonic models are capable of melting model is developed and used to explaining the Deccan in all its greatness, illustrate that it is very difficult to account and there is no trace of a mantle plume in for the seismic structure of the ridges if it this vast region. is assumed that the source of the GHS is a thermal anomaly, even if vigorous mantle upwelling coupled with deep damp Crustal Seismology Helps Constraining melting is included in the model. It seems the Nature of Mantle Melting easier to account for the estimated crustal Anomalies. Galápagos Volcanic structure if a more drastic, major element Province, A Case Study heterogeneity, is also considered. Based on these results, we suggest that the primary Valentí Sallarès (Unitat de Tecnologia source of the GHS is a compositional Marina, CSIC, Passeig Maritim de la heterogeneity, which may represent a Barceloneta, 37-49, 08003 Barcelona, mixture of depleted mantle and recycled Spain) oceanic crust. Such a mantle source explains well the isotope and trace element Wide-angle seismics provides accurate patterns showed by GVP basalts. information on the crustal velocity structure and the geometry of the crust- mantle boundary, and velocity-derived Midplate Volcanic Overprinting: New density models can be subsequently used Wine in Old Bottles to estimate mantle density required to explain observed gravity and topography E. L. Winterer (Scripps Institution of anomalies. A connection between the Oceanography, La Jolla, CA 92093, USA; seismic parameters and mantle melting Tel: +1-858-534-2360; Fax: +1-858-534- parameters can be established based on 0784; E-mail: [email protected]) existing empirical relationships between crustal thickness and seismic velocity of In addition to fresh fracture systems, igneous rocks and the mean pressure and younger midplate volcanism commonly fraction of melting within the mantle occupies older lines of weakness, such as melting region. older fracture zones, ridge-parallel faults, abandoned spreading centers and In this work, we compare seismic velocity pseudofaults created during original and density models of the crust and spreading-ridge propagation. Ages of uppermost mantle along five transects emplacement of such overprinted crossing the Cocos, Carnegie, and Malpelo volcanoes are commonly not a linear ridges in the Galapagos Volcanic Province function of distance, and may be entirely (GVP). The similar results obtained in all out of order or nearly simultaneous over
58 long distances. Directions and rates of swarms that produced 90% of these basalts plate motion cannot safely be inferred over 1.5 million years, with extreme uplift from the orientations of lines of present in several granite-cored mountains. overprinted volcanoes: crack propagation In addition, from a conventional, fixed is not everywhere the main control. deep mantle plume source reference frame Ponded broad lenses of fertile magmas for Yellowstone, North American plate beneath the base of the lithosphere can be motion predicts a hotspot location at 17 tapped by the opening of new fissures, as Ma that is 400 km south of this area. in the Hawaiian and Samoan chains, or Though some volcanism (e.g. Steens and through older fissures, especially where Malheur Gorge basalt) was present here at the lithosphere is in tension. that time, the majority of flood basalt volcanism (e.g. Columbia River Basalts) occurred nearly 400 km away in northeast Wednesday Poster Sessions Oregon, southeast Washington, and western Idaho. Delamination Origin for Columbia River Flood Basalts and Wallowa With these simple, well constrained Mountains Uplift in NE Oregon, USA contradictions to the standard plume head hypothesis for hot spot initiation in mind, D. L. Abt (Dept. of Geological Sciences, we deployed the first seismic array Brown Univ., Providence, RI 02912, focused on imaging the upper mantle in USA; Tel: +1-401-863-1965; Fax: +1-401- northeast Oregon. Using ~150 teleseismic 863-2058; E-mail: P-wave arrivals collected over 5 months at [email protected]); T C Hales the six-station array of broadband, three (Dept. of Geological Sciences, Univ. of component seismometers, we find high Oregon, Eugene, OR 97403, USA; Tel: velocity mantle, +4% Vp relative to +1-541-346-4354; Fax: +1-541-346-4692; IASPI91, beneath northeast Oregon at 70- E-mail: [email protected]); E 150 km depth and interpret this anomaly D Humphreys (Dept. of Geological as the melt-depleted mantle source region Sciences, Univ. of Oregon, Eugene, OR for the Columbia River Basalts. 97403, USA; Tel: +1-541-346-5575; Fax: Assuming an experimentally based density +1-541-346-4692; E-mail: change for depletion of garnet peridotite, [email protected]); J J Roering this volume of residuum could provide (Dept. of Geological Sciences, Univ. of sufficient isostatic support for the total Oregon, Eugene, OR 97403, USA; Tel:+1- volume of crustal uplift in the broad 541-346-5574; Fax:+1-541-346-4692; E- region. Yet, when considering the local mail: [email protected]) areas possessing excessive uplift (i.e., the Wallowa, Cuddy, and Elkhorn Mountains) The Columbia River Basalts, erupting 17- within this broadly uplifted region, it is 6 Ma and covering ~175,000 km2 in the apparent that a different mechanism for U.S. Pacific Northwest, represent the most uplift is necessary. Mechanical recent flood basalt event on Earth. The foundering of the lithosphere would allow well-mapped, large, continuous areal for large-scale mantle upwelling and extent of most early flows allows for decompression melting to occur prior to detailed flow interface analysis. As uplift. Furthermore, the removal of demonstrated in a companion abstract [T compositionally dense, eclogitic roots to C Hales et al.], significant crustal uplift the granitic plutons would decrease preceding flood basalt eruption, a feature density locally, allowing for anomalous typically associated with impingement of a areas of uplift. Therefore, we propose that rising mantle plume head, was not present lithospheric delamination provides a better here. Rather, mild pre-eruptive subsidence explanation for uplift history and followed by syn-eruptive and post- Columbia River flood basalt volcanism eruptive uplift of 300 m, on average, than a standard mantle plume. occurred over a broad, 200 km diameter region to the south and west of the dike
59 Continental Breakup Magmatism and 11 km thick, tapering down to thin oceanic Transition to Hot-Spot Influenced crust at A23 time, the increased melt Seafloor Spreading From the Moere potential was thus spent ~2.5 Ma after Margin to the Norway Basin continental breakup. There is a conspicuous correlation between half A. J. Breivik (Dept. of Geosciences, Univ. spreading rate and igneous crustal of Oslo, P.O. Box 1047 Blindern, N-0316 thickness. As this is not observed in a Oslo, NORWAY; Tel:+47-22-856676; normal seafloor spreading environment, Fax:+47-22-854215, E-mail: both plate spreading velocity and magma [email protected]); R. Mjelde (Dept. production should be governed by a of Earth Science, Univ. of Bergen, Allegt. common cause, presumably hot 41, N-5007 Bergen, NORWAY; Tel: +47- asthenosphere restricted to the rift zone 55-582879; E-mail: causing both increased melting and extra [email protected]); J I Faleide buoyancy driving force on the plate edges. (Dept. of Geosciences, Univ. of Oslo, P.O. Later seafloor spreading in the Norway Box 1047 Blindern, N-0316 Oslo, Basin resulted in both thin oceanic crust NORWAY; Tel: +4722-856677; E-mail: (4-5 km) typical for slow spreading ridges, [email protected]) and a complete lack of fracture zones along its 500 km length, similar to the Being part of the North Atlantic Igneous ultra-slow spreading Gakkel Ridge. A V- Province (NAIP), the Moere margin off shaped pattern seen in the gravity field mid-Norway is a volcanic passive margin, around the northern part of the Aegir offset from the volcanic Voering margin to Ridge corresponds to increased crustal the north by the East Jan Mayen Fracture thickness in the OBS model, Zone (EJMFZ). Both margin segments demonstrating the northeast migration of were created during the earliest Eocene asthenosphere zones with increased melt (~54Ma), associated with emplacement of production at a speed of 0.3-0.6 cm/a. large volumes of igneous rocks. In year These observations indicate that once the 2000, an ocean bottom seismometer breakup magmatism abated, the influence (OBS) profile was acquired across the on the seafloor spreading from the Iceland Moere margin to the Aegir Ridge, a hot-spot was present but very weak in the spreading axis dying in the Late Paleogene. Also, the maximum magma Oligocene. The P- and S-wave data were productivity was estimated to be just modeled by a combined ray-tracing and above 250 cubic-km/Ma/km, considerably inversion into a 2D velocity model. Due to lower than seen in other parts of the NAIP. low magnetic data coverage, a satellite Comparing igneous thickness and P-wave derived gravity map was used to velocity with mantle melting models does reinterpret the EJMFZ system, but no not support active mantle upwelling as an other proposed fracture zones could be important melt-producing process during identified. The revised EJMFZ trace was early seafloor spreading. used to re-evaluate spreading direction in the Norway Basin, which is quite asymmetric as it is condensed on the Kaapvaal Craton, South Africa: southwestern side. The magnetic track Repeated Basic Magmatism, Diamonds recorded along the OBS profile was used and Plumes to identify magnetic seafloor spreading anomalies by forward magnetic modeling, R. G. Cawthorn (School of Geosciences, and projected onto synthetic flow lines Univ. of the Witwatersrand, P O Wits, half spreading rates were derived along- 2050, South Africa; Tel: +27-11-717- profile. Maximum rate was above 3 cm/a 6557; Fax: +27-11-717-6579; E-mail: between A24A and A24b. The OBS [email protected]) transect shows a rapid transition from continental to oceanic crust with little The Kaapvaal craton in southern Africa tectonic thinning of the continent. Breakup contains remarkably well-preserved suites magmatism created igneous crust up to 10- of igneous rocks that provide major
60 constraints on plumes through time. Some The Setting of LIPS in the Lithosphere of the oldest komatiites (Barberton-type Through Time: One Test of the Plume locality) occur here, suggesting that Hypothesis keel/lithosphere formation began 3.5 Ga ago in this region. The best-preserved I. W. D. Dalziel (Institute for Geophysics, Proterozoic flood basalt sequence in the Jackson School of Geosciences, Univ. of world is found in the Ventersdorp Texas at Austin, 4412 Spicewood Springs Sequence (2.7 Ga). Lavas range from high Road (Building 600), Austin, TX 78759- magnesian to normal basalt. Feeder dykes 8500, USA; Tel: +1-512-471-0431; Fax: are known, demonstrating that magmas +1-512-471-0348; E-mail: were emplaced vertically and did not flow [email protected]) large distances laterally from outside the craton margin. The Bushveld Complex A fundamental prediction of the plume (2.06 Ga) is the largest known layered hypothesis is that hot plumes rising from intrusion. It formed from magmas that the mantle can impinge anywhere on the were both high magnesian and normal lower surface of the rigid lithospheric shell basalt. Importantly, 1 million cubic km of of the Earth. There can be no first-order magma was emplaced within a period of ‘top down’ control. Therefore, if large 65,000 years. It was immediately preceded igneous provinces, LIPS, result from by the largest known felsic volcanic event, impingement on the base of the the Rooiberg Group, presumed to be a lithosphere of upwelling mantle plumes, crustal melt. Once again, this feature however modified by interaction with that demonstrates that the basic magmas were upper thermal boundary layer during emplaced vertically, not laterally. Finally, emplacement, they should occur in the Drakensberg lavas of the Karoo geological provinces of all ages, and in all Supergroup (0.18 Ga) represent another tectonic settings. I believe that a strong major flood basalt outpouring with early case can be made that LIPS are indeed high magnesian basalt followed by normal geologically and tectonically ubiquitous in basalt. this sense, they formed in cratonic interiors, cratonic margins, oceanic Despite these major basic magmatic lithosphere, and active orogenic belts. episodes (and many smaller events LIPS may therefore be the result of a between) a stable keel survived. Through deep-seated process such as plume this keel has erupted major entrainment that is ‘blind’ to a mantle- diamondiferous kimberlite swarms at 1.2, enveloping shell. 0.5 and 0.1 Ga. Mantle nodules and inclusions in diamonds reveal ages from The heterogeneity of the lithosphere with over 3.0 Ga and assorted younger ages that regard to both strength and state of stress may reflect the above-mentioned major appears to control its response to the LIP- basic magmatic events. Thus, a significant generating process. Thus two such proportion of the 200 km-thick Archaean magmatic episodes within the same craton lithosphere survived these magmatic at different times, may in one case be events. Pathways for magma through this associated with continental break-up, and lithosphere probably caused the resetting in another only with rifting. Compare ages, suggesting that the magma source within Laurentia, for example, the (plume) did not continue to ascend above a association in space and time of the ~200 depth of 200 km. The similarity in Ma central Atlantic magmatic province magmatism for the last 2.7 Ga suggests with the initial opening of a major ocean that similar processes have operated for basin, and the ~1100 Ma Keeweenawan this time period. province and its association with the aborted mid-continent rift system. Again, LIP formation is totally independent of the lithospheric setting, taking place both within an assembled supercontinent in the former case and during supercontinental
61 assembly in the latter. This is also is 120 degrees off of the back-azimuth of consistent with a plume origin. plate motion. One model proposed to explain the volcanic trend is a propagating It is in active convergent margin orogenic shear zone, the Brothers fault zone. zones that LIPS are perhaps least common. Whether or not this fault zone has Subducting slabs may inhibit their propagated is testable by evaluation of the emplacement in the overriding lithosphere, time-space patterns of faulting. Recent although along the Transantarctic and Ar/Ar dating provides the necessary age Andean margins emplacement of constraints. Seven domains with rocks of Mesozoic LIPS appears to be controlled known age were chosen to cover the by supra-subduction zone extensional province spatially and temporally. Faults regimes. It has been suggested that a in these domains were digitized in a GIS; buoyant plume beneath a subducting slab 3,973 points were entered on 321 faults. may lower its angle of descent, resulting in Old rocks (~7.5 Ma) in the part of the deformation in the interior of an province where the rhyolitic volcanic overriding continent, for example in the system recently arrived (~1 Ma) are cases of the Laramide and Gondwanide significantly more deformed than younger orogenies. ‘Flat-slab’ zones may therefore rocks. This is part of a pattern that is provide as yet untapped opportunities for inconsistent with propagation of Brothers seismic tests of the existence of plumes. fault zone over the time-frame of age- progressive volcanism. Alternative models for the High Lava Plains relate it to the Testing a Propagating Shear-Zone Yellowstone melting anomaly, whether Hypothesis for Age-Progressive that be the result of a mantle plume or Magmatism in a Continental Setting: shallow mantle upwelling. The Oregon High Lava Plains
B. T. Jordan (Dept. of Geology, Whitman Compressional Structures Do Not Show College, Walla Walla, WA 99362, USA; Regional Horizontal Compression Near Tel: +1-509-527-4934; Fax: +1-509-527- the Iceland Hotspot 5904; E-mail: [email protected]) Maryam Khodayar (ÍSOR, Iceland A frequently proposed alternative to the Geosurvey, Grensásvegur 9, 108 mantle plume hypothesis for age- Reykjavík, Iceland; Tel: +354-5281522; progressive volcanic systems is E-mail: [email protected]); Páll Einarsson propagating rifts or shear zones. In the (Institute of Earth Sciences, Univ. of oceanic setting this may be difficult to test Iceland, Sturlugata 7, 101 Reykjavík, by study of the structures themselves Iceland; Tel: +354-5254816; E-mail: because the voluminous volcanic systems [email protected]) overwhelm pre-existing structures. In the continental setting it will more commonly Oceanic crust is generally formed at be possible to test the propagating shear diverging plate boundaries under condition zones model by careful evaluation of the of rifting and extension where normal time-space patterns of faulting. The faulting predominates within rift zones. Oregon High Lava Plains province is used However, reverse fault plane solutions are to illustrate an approach. occasionally found for earthquakes along the Mid-Atlantic Ridge. Furthermore, field Rhyolitic volcanism of the Oregon High observations of striae and in situ Lava Plains is age-progressive, younging measurements in places show compression to the west from 10 Ma to recent along a and reverse faulting. 250 km long trend from the Owyhee Plateau to Newberry volcano. The system Does horizontal compression play an can not be explained as the direct result of important role at oceanic diverging plate the motion of a plate over a mantle plume boundaries? This question is particularly because the azimuth of the trend (N75W) intriguing in areas above hot spots where
62 the high tectono-magmatic activitiy may, Overview of Tectonic Deformation in locally and temporarily, cause Past and Present Rift-Jump Blocks, compression. Horizontal compression West and South Iceland induces shortening in the rocks and closes fractures, and thus has implications for the Maryam Khodayar (ÍSOR, Iceland kinematics of the plate boundaries, the Geosurvey, Grensásvegur 9, 108 magmatism, and the fluid circulation along Reykjavík, Iceland; Tel: +1-354-5281522; the fractures. We have systematically E-mail: [email protected]); Páll Einarsson looked for evidence of such compression. (Institute of Earth Sciences, Univ. of One of the best ways is to study older and Iceland, Sturlugata 7, 101 Reykjavík, accessible structures in eroded crust that Iceland; Tel: +354-5254816; E-mail: was once formed at the plate boundaries [email protected]); Sveinbjörn Björnsson but shifted away. (Orkustofnun, Grensásvegur 9, 108 Reykjavík, Iceland; Tel: +354-5696028; We have found several cases of reverse- E-mail: [email protected]); Hjalti Franzson slip motion in our study of thousands of (ÍSOR, Iceland Geosurvey, Grensásvegur fractures in two areas in Iceland. We 9, 108 Reykjavík, Iceland; Tel: +354- selected eight examples from the eroded 5281523; E-mail: [email protected]) Tertiary and Quaternary crust of the Borgarfjörður and Hreppar rift-jump Plate boundaries become unstable above blocks, respectively, in West and South hotspots when they drift over these areas Iceland. These examples are among the of high volcanic production. The few structures were both striae and marker configuration of the plate boundary may horizons could be used to deduce the sense suddenly change as rift segments and of reverse-slip motion. These structures transform zones connecting propagating are commonly found in the field and used and receding rifts are relocated. Blocks of as indicators of compression. We find that crust may be transferred from one plate to in our cases the structures have variable the other by these rift jumps. These rift- kinematic origins and are not the result of jump blocks are characterised by intense regional compression. Vertical fracturing and a complex fracture pattern displacement along these structures is reflecting a time-varying stress field. always small, i.e. less than 10 m, and their Extinct rifts, rift-jumps and rift strike is inconsistent with stress fields propagation are identified by their derived from other fractures in the structural relationships, regional tilt, surrounding area. We believe reverse-slip unconformities, paleomagnetic and motions occur: (a) In association with petrological signatures. In this dykes, sills, or cone-sheets without presentation we focus on the fracture evidence of shortening across intrusions; pattern of rift-jump blocks. (b) Due to local bends and irregularities of steeply-dipping regional and secondary Several rift-jumps have occurred in the normal faults. (c) At the tips of steeply- Iceland area since the initiation of dipping normal faults uplifted by spreading between Greenland and Europe. underlying propagating dykes. Two such rift-jumps have been identified If horizontal compression acts in Iceland, in Western and Southern Iceland, south of its magnitude is not sufficient to shorten 65° N. The Borgarfjörður Block of West the rocks on a regional scale, to close the Iceland is located between the Snæfellsnes fractures, or to cause folding. Most Rift Zone (active during 15-5 Ma) and the reverse-slip motions are local adjustments Reykjanes-Langjökull Rift Zone-RLRZ to companion structures, without regional (active since about 6 Ma). Traces of an old compressional stress or shortening, and transform zone or an oblique rift exist in can lead to misinterpretation if not put into this block. The intra-plate Snæfellsnes the correct context. Volcanic Zone has been active since 2 Ma, and the 1974 Borgarfjörður earthquakes and prominent geothermal manifestations lie near the eastern tip of this chain. The
63 Hreppar Block is located between the Paleogene North Atlantic Igneous RLRZ and the Eastern Rift Zone (active Province and the Iapetus Connection since about 3 Ma), and lies north of the presently active transform zone, the South C. E. Lesher (Dept. of Geology, Univ. of Iceland Seismic Zone (SISZ). A long-term California, Davis, CA 95616, USA; Tel: plan to study the tectonics of these rift- +1-530-752-9779; Fax: +1-530-752-0951; jump blocks has been set up in order to E-mail: [email protected]); E better understand the mechanisms of L Brown (Dept. of Geology, Univ. of unstable plate boundaries in time and California, Davis, CA 95616, USA; Tel: space. Below is an overview of a few +1-530-752-5041; Fax: +1-530-752-0951; interesting results to date. E-mail: [email protected]); L E Heister (Dept. of Geology, Univ. of In both blocks the fracture density changes California, Davis, CA 95616, USA; Tel: as a function of stratigraphical level within +1-530-752-5041; Fax: +1-530-752-0951; the crustal section. The fracture pattern, E-mail: [email protected]) tectono-magmatic activity and stress field are complex and change with time. The The prevailing paradigm for the early fracture population consists of six fracture Paleogene North Atlantic igneous families striking parallel, but also oblique province holds that the large volumes of and perpendicular to the rift zones. In the basalt erupted during continental breakup Borgarfjörður Block, the rift-parallel were generated by decompression of the fractures constitute only 1/3 of the total. ancestral Iceland plume arriving at the Dominant fracture directions vary with base of Pangaea ~60 mys ago. Based on a rock age, but the six fracture families comparative study of the geochemistry of appear from the time when the two rifts early Paleogene basalts from the conjugate and a transform zone connecting them margins of the North Atlantic Ocean basin, acted together. Among the six families in we explore an alternative model. In this the younger Hreppar Block, four are most model, mantle sources tapped by partial prominent in the late Tertiary and melting reflect ambient upper mantle Quaternary rocks. A simpler pattern of conditions at that time, specifically lateral conjugate strike-slip faults appears in the differences in sublithospheric mantle Holocene lavas of the presently active composition coupled with modest SISZ. Some of the active faults of this temperature anomalies (<100 K). Basalts zone can be traced into the older crust from central east Greenland are dominated where they are injected by dykes. by tholeiites enriched in iron and titanium, strongly fractionated rare earth element Fault types are similar in both blocks, with ratios, low Zr/Nb ratios, and radiogenic both normal faulting and strike-slip isotope compositions corresponding to faulting of opposite senses occurring along immature HIMU values. These all six directions. At the southern characteristics are not shared by early boundary of the Hreppar Block, the Paleogene basalts erupted in southeast and locations of geothermal manifestations are west Greenland that are low in titanium, strongly influenced by the tectonics of the depleted in incompatible trace elements, transform zone. In the Borgarfjörður and possess radiogenic isotope Block, on the other hand, the distribution compositions similar to depleted OIB and of geothermal activity is influenced by MORB. The isotopic compositions of intraplate deformation, characterised by basalts from central east Greenland reflect normal faulting earthquakes. Despite the a mantle source with a mean age of ~600 role of earthquakes, the local openings Ma, while those from southeast and west leading to geothermal manifestations at the Greenland are derived from significantly surface occur along the six fracture older mantle (>2 Ga). Such regional families that are inherited and reactivated differences are not readily explained by from the underlying crusts of the Hreppar current plume models, but can be related and Borgarfjörður rift-jump blocks. to features of the regional basement geology. It is probably no coincidence
64 that the southwestern boundary of the tectonics, particularly extension). A Caledonian front and zone of closure of plume mechanism is supported by (1) the Iapetus ocean correspond closely to the youthful uplift and dissection of as much southwestern limit of the geochemical as 1 km NE of Yellowstone that precedes anomaly associated with the central east volcanism, (2) crustal tilting away from Greenland volcanic province. We Yellowstone on leading margin of attribute the distinctive basalt Yellowstone hotspot, and (3) the highest geochemistry and high melt productivity geoid anomaly in the US that is centered accompanying rifting within the on and extends NE (ahead) of Caledonian suture zone to the presence of Yellowstone. The following large-scale subducted Iapetus crust in the upper processes associated with Yellowstone are mantle along this portion of the rifted more directly explained by an active, margin. Given the depleted nature of deep-sourced, rather than passive, mantle basalts from southeast and west upwelling: 1) Yellowstone has the largest Greenland, similar "recycled" material was concentration of geysers and other not available south and west of the geothermal features in the world and heat Caledonian suture zone. Likewise, flow 50 times normal, 2) large-volume comparisons between central east volcanism and intrusions occur with Greenland and modern ridge system Yellowstone and the YHT, and 3) 1) the suggest that source heterogeneities north scale of uplift and faulting extending south of the Caledonian front have changed and west from Yellowstone span a width through time. The connection between of ~400 km,. The eastern Snake River subduction processes involved in Plain (ESRP), in the younger part of the supercontinent assembly and mantle YHT, has active faults on its margin and heterogeneities sampled during breakup volcanic rift zones across the ESRP that reduces the necessity to derive recycled are at high angles to its trend, inconsistent material from a lower mantle plume early with formation of the ESRP by in the development of the North Atlantic lithospheric rifting driving mantle Ocean basin, although it may be important upwelling. After 10 Ma, both caldera- at Iceland today. forming volcanism and high rates of faulting have migrated NE along the YHT at the same rate and direction as predicted Geologic Evidence for a Mantle Plume by North American plate motion over a Origin for Yellowstone: The Pattern mantle plume. The YHT is parallel to but and Scale of Volcanism, Faulting, and failed to follow a major crustal flaw (the Uplift Along the Yellowstone Hotspot Great Falls tectonic zone in the larger Track Madison mylonite zone) even though the zone is only 40 km N of the YHT, Pierce, K. L. (U.S. Geological Survey, favoring a fixed mantle plume rather the P.O. Box 173492, Montana State Univ., lithospheric pull apart (rift) origin. The Bozeman, MT 59717-3492, USA; Tel: +1- YHT started ~16 Ma with widespread 406-994-5085; E-mail: [email protected]) flood basalt volcanism (Columbia River L. A. Morgan (U.S. Geological Survey, Basalts and correlative flood basalts in PO Box 25046, 966 Federal Center, Oregon and N California) and formation Denver, CO 80226, USA; Tel: +1-303- of the N-S, 1,100-km-long Nevada- 273-8646; E-mail: [email protected]) Washington rift zone, similar to other plume-head events. Modeled A NE-migrating pattern of volcanism, asthenosphere parallel to this N-S rift is faulting, and uplift define the 750-km-long buoyant and consistent with a relict plume Yellowstone hotspot track (YHT). Based head. The simultaneous NW progression on the geology associated with the YHT, of rhyolite volcanism across southern we favor formation of the YHT by mantle Oregon may involve plume-head plume (active mantle, deep-sourced) over spreading, slab subduction, and strike-slip that by lithospheric-tectonics (a passive faulting and not necessarily negate a mantle responding to lithospheric plume origin of the YHT. Finally, recent
65 tomographic studies have imaged a 100- Does a Fault in the Plate Circuit Ruin km-wide low velocity zone to at least 500 Intra-ocean Comparison of Hotspot km beneath Yellowstone (Yuan and Tracks? Dueker, in press 2005). William W. Sager (Dept. of Oceanography, Texas A&M Univ., Passive Margin Evolution: Are Plumes College Station, TX 77843-3146 USA; an Integral Part of Continental Tel: +1-979-845-9828; Fax: +1-979-845- Breakup? 6331; E-mail: [email protected]); Richard G. Gordon (Dept. of Earth Ian Norton (ExxonMobil Upstream Science, Rice Univ., Houston TX 77005, Research, P.O. Box 2189, Houston, TX USA; Tel: +1-713-348-5279; Fax:+1-713- 77252, USA; Tel:+1-713-431-4240; Fax: 348-5214; E-mail: [email protected]) +1-713-431-6193; E-mail: [email protected]) Whether hotspots are plumes that are relatively fixed in the mantle is an When continental crust undergoes rifting important question of modern and extension that results in formation of geodynamics because of its implications new oceanic crust, a transition zone for modes of mantle convection and the between undeformed continental crust and widespread use of the hotspots as a oceanic crust is formed. This zone reference frame for plate reconstructions. becomes a locus of later sedimentary Although once widely accepted, the deposits and, as long as it remains "fixed" hotspot approximation has come tectonically quiescent, is termed a ‘passive under recent scrutiny as some results that margin’. Traditionally, passive margins apparently contradict this assumption have have been classified as ‘volcanic’ or ‘non- been reported. One source of doubt about volcanic’. This nomenclature is fixed hotspots is the finding of a misleading, however, as all passive discrepancy between hotspot tracks margins are ultimately volcanic, i.e. ocean predicted in different oceans. Typically, crust formation is a volcanic process. We investigators have tried to predict the path propose a classification based on timing of of the Hawaiian-Emperor seamount chain onset of volcanism, with ‘volcanic’ from the motion of Indo-Atlantic plates margins simply a result of early onset of over the hotspots transferred to the Pacific volcanism and ‘non-volcanic’ margins using a plate circuit through Antarctica. resulting from continental stretching Several such studies have found that the without volcanism before intiation of sea predicted Hawaiian-Emperor track has floor spreading. With this scheme, the little or no bend and diverges significantly role of plumes as sources of early from the trend of the older, Emperor volcanism is less clear. This poster will Seamounts. One interpretation is that the discuss these points, with a focus on the Indo-Atlantic and Pacific hotspots South Atlantic. experienced significant and potentially rapid relative motion prior to 40 to 50 Ma. An alternative interpretation is that plates in the plate motion circuit may not have been rigid throughout the time interval encompassed by the reconstructions. Both Antarctica and the south Pacific plate are poorly explored areas that have been postulated as the locus of this missing motion. In the Pacific, paleomagnetic data from Ontong Java Plateau display a discrepancy that suggests ~15 degrees less northward motion than the rest of the Pacific, approximately the same offset indicated by intra-ocean hotspot
66 comparisons. Furthermore, Jurassic [email protected]); J P Morgan basalts from the plateau edge are out of (Dept of Earth and Atmospheric Sciences, place compared with other coeval Pacific Snee Hall, Cornell Univ., Ithaca, NY crust by the same amount. Other south 14853, USA; Fax: 607-254-4780; E-mail: Pacific paleomagnetic data may support [email protected]) this finding and suggest that there is an unrecognized break in the southern Pacific One of the striking exceptions to the plate. Within Antarctica, Late Cretaceous mantle plume head-tail hypothesis that and early Cenozoic relative motion seeks to explain the exceptional between east and west Antarctica has been magmatism of large igneous provinces shown in several studies; although, the (LIPs) and hotspot tracks is the ~250 amount, timing, and rotation poles are million-year-old Siberian Traps. The lack uncertain. It seems clear that a break in of a clear hotspot track linked to this LIP the plate circuit is not only plausible, but has been one motivation to explore non- likely. What is more, recent plume alternative mechanisms. Here, we reconstructions of Pacific and Indo- use paleomagnetic Euler pole analysis to Atlantic paleomagnetic data into the constrain the location of the Siberian hotspot reference frame show good Traps at the time of their eruption. The agreement whereas reconstructions using reconstructed position coincides with the plate motions fail to bring the region that also saw eruption of the ~61-58 paleomagnetic data into coincidence million-year-old North Atlantic Volcanic [Andrews, D. L., R. G. Gordon, and B. C. Province (NAVP). Together with LIP Horner-Johnson, EOS, Trans. AGU, 85, volume estimates, this reconstruction Fall Mtg. Suppl., Abstr. V51B-0540, poses a dilemma for some non-plume 2004; Gordon, R. G., D. L. Andrews, B. models: the partial-melts needed to C. Horner-Johnson, and R. R. Kumar, account for the Siberian Traps should have EOS, Trans. AGU, Jt. Assem. Suppl., depleted the enriched upper mantle source Abtr. GP22A-06, 2005 ]. This result that is in turn crucial for the later suggests that the discrepancy between formation of the NAVP. These analyses observed and predicted hotspot tracks suggest the existence of a long-lived (>250 results from one or more flaws in the m.y.) lower mantle chemical and/or trans-Antarctic plate circuit, rather than thermal anomaly, in the region of the significant intra-hotspot motion. Whether present-day Iceland hotspot. These hotspots are have little motion relative to observational constraints also suggest that one another is still an open question; a nascent plume is not needed to produce however, the failure of efforts to flood basalt volcanism - plume material reconstruct hotspot tracks using plate filling a rifting lithospheric thinspot may circuits is not a compelling argument be sufficient. against the fixed hotspot hypothesis.
NE Atlantic Breakup and Evolution of Co-location of Eruption Sites of the the Norwegian-Greenland Conjugate Siberian Traps and North Atlantic Volcanic Margins: Field Evidence to the Igneous Province: Implications for the Great Plume Debate Nature of Hotspots and Mantle Plumes F. Tsikalas (Dept. of Geosciences, Univ. A. V. Smirnov (Dept of Geology and of Oslo, Norway; Tel: +4722856678; Fax: Geophysics, Yale Univ., New Haven, CT +4722854215; E-mail: 06511 USA; Tel: +1-203-432-1269; Fax: [email protected]); J I Faleide +1-203-432-3134; E-mail: (Dept. of Geosciences, Univ. of Oslo, [email protected]); J A Tarduno Norway; Tel: +4722856677; Fax: (Dept. of Earth and Environmental +4722854215; E-mail: Sciences, Univ. of Rochester, Rochester, [email protected]); A J Breivik, NY 14627, USA; Tel: +1-585-275-5713; (Dept. of Geosciences, Univ. of Oslo, Fax: +1-585-244-5689; E-mail: Norway; Tel: +4722856676; Fax:
67 +4722854215; E-mail: highly intruded lower crust. Therefore, [email protected]); R. Mjelde (Dept. improved constraints can be reached of Earth Science, Univ. of Bergen, through modelling of various syn-breakup Norway; Tel: +4755582879; Fax:+ 47- processes, such as: regional extension and 55583660; E-mail: magmatism (including estimates of [email protected]); O. Eldholm volumes and rates); extent and timing of (Dept. of Earth Science, Univ. of Bergen, the subsequent continental heating; and the Norway; Tel: +4755583281; Fax: +47- interplay of sedimentation, magmatism 55583660; E-mail: and vertical motion. Seismic mapping [email protected] shows a distinct lateral distribution of breakup-related rocks revealing the A set of regional transects across the influence of the along-strike segmentation conjugate Norwegian-Greenland margins and its role in the tectono-magmatic in NE Atlantic integrated with potential margin evolution. In particular, the field data and modelling reveal important Norwegian margin is divided into a series vertical and lateral variations in crustal of rifted, sheared, and oblique-segments configuration and composition associated that appear to have experienced different with continental breakup near the structural, magmatic and temperature Paleocene-Eocene transition and its histories. Details on margin physiography preceding phase of extension in Late during early opening are crucial for Cretaceous-Early Tertiary times. The understanding the driving force for vertical Norwegian-Greenland volcanic margins movements at a local and regional scale. belong to the North Atlantic Large Igneous Province, and several observations substantiate their formation Bermuda: Lava-lamp Plume, Edge- by impingement of the Iceland plume on a driven Convection, or/and Response to lithosphere under extension. In particular, Distant Plate Reorganization? preceding extension culminated with regional uplift and subsequent erosion P. R. Vogt (Dept. of Geology, Univ. of towards the end of Paleocene, and California, Santa Barbara, CA 93106- lithospheric breakup was accompanied by 3060, USA; Tel: +1-410-586-0067; E- a massive, regional magmatic event, mail: [email protected]) resulting in thick seaward dipping reflector sequences that manifest massive eruptions The ca. 600x900km Bermuda Rise (BR) of lavas covering large areas along the represents old (ca. 150-90Ma) oceanic continent-ocean transition. Refinement of lithosphere currently elevated up to ca. the continent-ocean boundary and seafloor 800-1000m (residual depth anomaly). The spreading anomalies along the conjugate rise, compensated at ca. 50km depth, is margins provide improved geometrical accompanied by a 5-10m geoid high and a and azimuthal constraints on early opening small (5 to 10 mW/m2), poorly resolved plate reconstructions that, in turn, confirm heat flow high. The uplift began in Middle high early opening spreading rates and Eocene times but continued at reduced contribute to a better understanding of the rates into the Miocene. Igneous activity coupling between volcanism and early sea was confined to the rise summit, where floor spreading. Furthermore, a denser pillow basalts began to build a 100km OBS data coverage contributes to a more long chain of four volcanoes (Bermuda is detailed (3D) mapping of magmatic the largest) during or immediately after continental underplating (7+ km/s lower rise uplift began. crustal body) volumes and distribution, constrained further by potential field Geophysical modeling, largely in the modelling. From high-quality OBS data 1970s-1980s, has been reasonably the Vp/Vs ratio can provide constraints on successful in matching observations with physical properties of the underplated some combination of distributed heating in material to enlighten its nature as either a the lower lithosphere, and/or shallow compact melt body or most probably
68 mantle convection involving the lower part kind of local or regional "pre-conditioning of the thermal lithosphere. anomaly" in the subjacent mantle lithosphere/asthenosphere below Bermuda However, the lack of the expected hotspot is then required (e.g.,location relative to trace, among other factors, led Sclater and continental edges and MOR axis Wixon (1986), Vogt (1991) and King and (Vogt,1991); fertile patch or zone of Anderson (1998) to consider shallow weakness) to explain how and why mantle convection that is attached to and Bermuda happened WHERE it did. moving with the lithosphere. A plume type model for Bermuda could be Hawaii-Bermuda differences (Both BR supported (or not) if 1) future seismic and the short volcanic chain on its crest experiments resolved velocity anomalies are perpendicular, not (sub-)parallel to in the lithosphere, asthenosphere, or putative plate motion; absence of a deeper below the predicted present continuous hotpot "trace" plus evidence location of the Bermuda hotspot, ca. 500 against migration of Bermuda uplift with km east of Bermuda; 2) seismic time; lack of subsidence of rise and experiments resolved anomalies in the volcanic pedestal) have been attributed by lithosphere between the BR and plume advocates to the slower motion (15- Mississippi; and/or 3) geochemical or 30 mm/a at Bermuda vs 90mm/a at isotope characteristics of the Mississippi- Hawaii) over the subjacent mantle, and/or Kansas igneous bodies in some significant a "weak", lava-lamp-like plume under way resembled those of Bermuda lavas Bermuda. The latter behavior is required and lamprophyre sheets. Non-plume to account for the lack of a hotspot trace formation models would be supported by east of Bermuda, and the 20-30my gap in 1) pre-existing (i.e., pre-Eocene) "pre- igneous activity between Bermuda and the conditioning" anomalies in the lithosphere youngest (65Ma) postulated Bermuda- within which the BR and volcanoes hotspot-generated igneous bodies in developed and 2) turbidites offlap the BR Mississippi. However, as noted by Cox in a way to suggest simultaneous uplift and van Arsdale (2002) particularly the over the entire BR, vs outflow from a Duncan (1984) fixed-hotpot model does narrow plume. crudely predict the directions and ages of igneous activity from Mississippi to Kansas (115Ma), observations which other The Deccan Basalt – Basement Contact: models must write off as coincidental. Evidence for a Plume-Head Generated CFBP? A new compilation of unusual coeval tectonic and magmatic events around the M. Widdowson (Dept. of Earth Sciences, world suggests Bermuda was not a The Open Univ., Walton Hall, Milton temporally isolated event, but one of many Keynes, MK7 6AA, UK; Tel: +01908 local/regional responses (e.g. Rona and 652986; Fax: +01908-655151; E-mail: Richarson, 1978) to global plate [email protected]) reorganization occasioned by the Eocene replacement of rapid (ca 100 mm/a) The widely accepted ‘plume head’ model northward subduction of Tethyan oceanic of continental flood basalt (CFB) genesis lithosphere by slower(50mm/a) collision predicts a pre-volcanic, domal uplift. This of Indian with Eurasian continental crust. should be accompanied by associated Because plate reorganizations in general regional erosion immediately prior to, and require reorganizations of passive during the onset of the main eruptive asthenosphere return flow ,global phase. However, once initiated, rapid lava synchronism might be explained effusion generates a swiftly-growing (following Anderson, 2002) without volcanic edifice that may effectively bury appealing to global reorganization of evidence of these early uplift and erosion plume-type mantle stages. By contrast, non-plume, ‘edge- convection(Vogt,1975). However, some driven convection’ (EDC) models do not
69 predict significant pre-volcanic uplift; an The notion of mantle plumes has been absence of regional uplift or erosional variously invoked in the geochemical indicators in the pre-volcanic geological literature, and there is some danger that record has thus been used to help the existence of mantle plumes is debated substantiate such models. The current on the basis of the weaker, rather than the work synthesizes evidence gathered from stronger arguments for their support. It extensive Deccan fieldwork, conducted remains difficult to identify mantle plumes over a period of 20 years, in an effort to on the basis of geochemistry, both because resolve whether or not models predicting mantle plumes are physical phenomena, ‘plume head’ uplift and erosion actually and because it is only occasionally satisfy the field observations. possible to infer the depth of origin of intraplate material on the basis of The nature of the basement-basalt contact geochemistry. Geochemistry is used to is examined at key localities around the distinguish intraplate basalts from those periphery of the Deccan Volcanic generated in other tectonic settings, but Province (DVP), India. Each locality only some of those basalts may be offers a different geological setting and, associated with mantle plumes. Thus, the when considered together, reveals the isotope and trace element geochemistry of nature of the pre-, syn-, and late-stage basalts from Auckland, the Cape Verdes, volcanic environments, and the growth of certain seamounts, and Hawaii, have many the DVP edifice. High precision similarities implying that they are derived 40Ar/39Ar data confirms that the lavas from non-MORB relatively enriched comprising the DVP become younger both mantle. However, it is most unlikely that southward and south-eastward, thus they are all associated with mantle plumes, supporting models invoking a southward irrespective of how mantle plumes are migrating locus of volcanism. In the defined. The key aspect of magmas that northern DVP, early lavas include K-rich have been associated with mantle plumes picrites which lie with angular is that they have relatively high volumes, unconformity upon Late Cretaceous and melt generation rates – they are marine limestone and fluvial sandstone volume anomalies that necessarily reflect successions. In some instances thick some form of physical anomaly in the melt conglomerate sheets containing both generation zone, whether that is basement and basalt clasts intervene temperature or volatiles. At present there between the tilted sedimentary succession is little evidence that they are and overlying lavas. These provide characterised by relatively high volatile evidence of significant erosion having contents. These high volume magmas occurred in the earliest phases of DVP share some geochemical similarities that eruption. Only at a much later stage did are in turn therefore associated with the lava fields arrive at the eastern and melting within upwelling material, widely southern periphery of the DVP. Here, they termed mantle plumes. Finally, the links quiescently over-ran pre-existing between mantle components, recognised topography and contemporaneous primarily on the basis of their radiogenic lacustrine environments. isotope ratios, and mantle plumes will be explored.
Thursday Oral Sessions Magmatic Evolution of Mauna Loa Geochemistry and Mantle Plumes Volcano: Implications for a Chemically and Thermally Zoned Mantle Plume Chris Hawkesworth (Dept. of Earth Sciences, Univ. of Bristol, Wills Memorial J. M. Rhodes (Dept. of Geosciences, Univ. Building, Queens Road, Bristol, BS8 1RJ, of Massachusetts, Amherst, MA 01003, UK; Tel: +0117-954- 5425; Fax: +0117- USA; Tel:+1-413-545-2841; Fax 413-545- 954-5236; E-mail: 1200; E-mail [email protected]); [email protected]) D. Weis (Dept. of Earth and Ocean
70 Sciences, Univ. of British Columbia, one should expect to see differences in Vancouver, BC, V6T_1Z4, Canada; M. O. these variables over time as the volcano Garcia, Dept. of Geology-Geophysics, transits the plume. This is the case for Univ. of Hawaii, Honolulu, HI 96822, Mauna Kea. The normalized SiO2 content USA) of the lavas decreases over time, along with eruption rates, as the volcano Hawaii is regarded as the archetypal approaches the post-shield stage. expression of a mantle plume. The Paradoxically, there are no such temporal Hawaiian plume is thought to be zoned in changes of composition in Mauna Loa both temperature and composition, and the lavas. With the exception of some alkalic magmatic evolution of Hawaiian radial vents, all Mauna Loa lavas follow volcanoes, and their inter-volcano well-defined olivine-control trends that are compositional differences, are widely indistinguishable from modern historical interpreted in terms of melting within the lavas. The implications are that although plume. Our studies of samples from the plume may be compositionally zoned Mauna Loa volcano include historical on a scale of ~ 30 km, either melting and lavas (0.2 ka), 14C-dated prehistoric lavas melt segregation processes have remained (0.2 - 36 ka), lavas recovered by the remarkably constant over a broad region Hawaii Scientific Drilling Project (10 - of the plume, or alternative processes are 100 ka) and lavas from Mauna Loa’s at work. submarine southwest rift zone (<100 - >300 ka). This magmatic record is comparable with that of the ~ 400 ka Magma Genesis in a Mantle Plume: record for Mauna Kea volcano obtained by Based on High-pressure Melting the Hawaii Scientific Drilling Project, and Experiments and Growth History of provides another excellent opportunity for Some Hawaiian Volcanoes testing and/or modifying the plume model. During this period, Mauna Loa will have Eiichi Takahashi (Dept. of Earth and transited >30 km to the NE and should Planetary Sciences, Tokyo Institute of record differences in both source Technology, 2-12-1 Ookayama, components and melting processes in the Meguroku, Tokyo 152-8551, Japan; Tel: composition of its lavas if the plume +81-3-5734-2338; Fax+81-3-5734-3538; model is tenable. High precision Pb and E-mail: [email protected]) Sr isotopic ratios, in combination with X/Nb ratios, clearly indicate that lavas The significance of recycled oceanic crust older than 100 ka are distinct from (eclogite) component in generation of hot younger lavas (<36 ka). The younger spot magmas has received much attention lavas are more variable, and tend to have [e.g., Hauri, 1996; Campbell, 1998; higher Sr and lower Pb isotopic ratios than Takahashi et al., 1998; 2002]. The the older lavas. The older lavas have Pb mechanism of magma production in the and Sr isotopic ratios that approach those hybrid partial melt zone (consisting of of Loihi, the youngest volcano along the eclogite and peridotite), however, is not so-called Loa trend. These data are in well understood. The role of the interface accord with a zoned plume model, but that separates these lithologically different require bi-lateral asymmetric zoning and domains in the source region is an heterogeneities, rather than concentric essential but unknown factor in plume zoning. magma genesis. The goal of this study is to model magma generation in the Hawaii The major elements, and most of the trace plume based on new experiments by and data, are another matter. Melting the growth history of Hawaiian volcanoes experiments on peridotites indicate that discovered by studies based on JAMSTEC changes in the extent of melting, or in the Hawaii cruises (e.g. Koolau, Tanaka et al., depth of melt segregation, will result in 2002; Haleakala, Ren et al. 2004). differences in SiO2, MgO and FeO in the magmas. Given a thermally-zoned plume,
71 The basalt/peridotite sandwich melting plume based on melting model of simple experiments were conducted at 2.5 to 3.5 peridotite source may be up to 100 C GPa and 1400-1550 C for 10-100 hrs in overestimated in the case where melting of order to study the reaction kinetics of the entrained eclogite component is prevailing basalt/peridotite interface and the in magma genesis. chemistry of melts produced by reactions between layered basalt and peridotite. Both melts and solid were analyzed with Do Hotspot Basalts Share a Common EPMA and LA-ICPMS. It is found that Mantle Source? melts formed above the peridotite solidus (>1475 C at 3 GPa) are silica under- Godfrey Fitton (School of GeoSciences, saturated alkalic basalts (alkalic picrite) Univ. of Edinburgh, EH9 3JW, UK; Tel: similar to OIBs (Loihi). As temperature +44-131-650-8529; Fax:+44-131-668- decreases, the melt composition becomes 3184; E-mail: [email protected]) olivine tholeiites (picrite) similar to those in Kilauea in the temperature range where Iceland is the ideal place in which to study basalt interacts with peridotite wall rocks the mantle source(s) for hotspot basalt by melt infiltration (1475-1450 C at 3 because its location on the Mid-Atlantic GPa). At temperatures below the solidus Ridge causes the source to melt to large of most fertile peridotite the melts formed degrees on-axis and small degrees off- in the basalt layer are isolated from the axis. Icelandic basalt ranges in peridotite matrix by opx reaction band and composition from highly depleted picrite have composition of silica-saturated to moderately enriched mildly alkaline basaltic andesite (<1425 C at 3 GPa). The basalt; a much larger range than at any basaltic andesite melt could be the source other hotspot. It forms a single array on a component for the silica-rich Makapuu logarithmic plot of Nb/Y vs. Zr/Y that is stage of Koolau volcano. distinct and parallel to the array defined by N-MORB, thus supporting the conclusion Based on these experiments, temperature of Pb-isotope studies that the N-MORB profile of the Hawaiian mantle plume is source is not a significant component in constructed. Magma feeding zone under the source of Icelandic basalt. The most Loihi may be highest in temperature (1475 depleted Icelandic basalts are more C) and decrease as the volcano migrated depleted in incompatible elements than are away from the plume axis. At the most most N-MORB and yet they are still voluminous olivine tholeiite stage in relatively rich in Nb. This relative Kilauea volcano (1450 C) as much as 50 enrichment in Nb can be expressed vol.% of melts are supplied from eclogite numerically as delta-Nb, which is the source. As the volcanoes migrate further deviation from a reference line separating away from the plume axis, melting in the the Iceland and N-MORB arrays. Thus peridotite matrix diminishes and the silica- virtually all N-MORB has delta-Nb < 0 rich melts formed in the eclogite domain and all Icelandic basalt has delta-Nb > 0. becomes prevailing (Mauna Loa to Koolau stage, 1425 C). Based on this model Nearly all OIB, and basalt from most LIPs absence of tholeiite magma in most hot (e.g. Ontong Java Plateau; Columbia River spots may be explained by the absence of Plateau) plot in the Iceland array and entrained eclogite blocks or very small in therefore have positive delta-Nb, size if any. Independent analyses based on suggesting that this parameter reflects a Fe-Mg of magma source indicate that fundamental characteristic of the source of entrained eclogite blocks are most hotspot basalt. Most basalt erupted in abundant in Hawaii and Iceland plumes. continental rift systems also has positive Magma production rate in mantle plume delta-Nb, irrespective of whether rifting is may be controlled by 1) plume flux, 2) accompanied by uplift (e.g. East Africa; potential temperature, 3) thickness of the Basin and Range) or not (e.g. Scottish plate and 4) amount of the entrained Midland Valley; North Sea). This suggests eclogite blocks. Potential temperature of a that the mantle component with positive
72 delta-Nb (probably subducted oceanic experienced an abrupt block uplift, crust) is present on all scales, from large consequently being elevated above sea mantle upwellings as inferred to be level. This was followed by a widespread present, for example, beneath Iceland and subaerial volcanism in Eastern Anatolia Hawaii, to streaks of enriched material and the surrounding regions (e.g. Georgia, present in the upper mantle and Armenia, Azerbaijan and Iran). Volcanic preferentially sampled through small- activity produced a wide range of volcanic degree melting beneath passive continental products (i.e. lavas and pyroclastic units), rift systems. Rare, near-axis MOR spanning the whole compositional range seamounts with positive delta-Nb from basalts to rhyolites. Great volumes of demonstrate the existence of small-volume volcanic material reaching over 1 km in streaks or blobs of this enriched material thickness in places erupted onto the in the upper mantle. Although positive surface, covering almost two-thirds of the delta-Nb is a characteristic feature of region. The region gradually gained a basalt at most hotspots, magmatism on the regional domal shape comparable to that Siberian Platform and in the early phase of of the Ethiopian High Plateau, although activity of the North Atlantic Tertiary the Anatolian dome has a north-south province is dominated by basalt with shortened asymmetrical shape possibly negative delta-Nb (i.e. it resembles N- due to collision (Sengor et al., 2003). MORB). The existence of a regional dome structure Thus, there is no simple correspondence and widespread magmatism are both between OIB-like basalt (with positive regarded as evidence for the existence of a delta-Nb) and postulated deep-mantle mantle plume and these do indeed exist in upwelling (plumes). Small seamounts and Eastern Anatolia. By these properties, the OIB-like basalt formed as a passive Eastern Anatolia region can be regarded as response to lithospheric extension clearly the site of a "melting anomaly" or do not involve mantle plumes. On the "hotspot" resembling closely the setting other hand, some LIPs are composed of proposed for mantle plumes. However, basalt that is not OIB-like and more geologic and geochemical data provide closely resembles N-MORB. The evidence against a plume origin. The geochemical diversity of hotspot basalts results of the Eastern Turkey Seismic requires that if mantle plumes do exist Experiment Project (Sandvol et al., 2003), then they cannot all share a common coupled with geological and geochemical mantle source. findings, support the view that both domal uplift and extensive magma generation can be linked to the mechanical removal of a Eastern Anatolia: A Hot Spot in a portion or the whole thickness of the Collision Zone Without a Mantle Plume mantle lithosphere, accompanied by passive upwelling of normal-temperature M. Keskin (Dept. of Geological asthenospheric mantle to a depth as Engineering, Faculty of Engineering, shallow as 38 to 50 km. This process Istanbul Univ., 34850 Avcilar, Istanbul, might have taken place either by Turkey; Tel: +90-212-473-7070 ext: delamination (Pearce et al., 1990; Keskin 17887; Fax: +90-212-473 7179; E-mail: et al., 1998), slab-steepening & breakoff [email protected]) (Keskin, 2003), or a combination of both. The Eastern Anatolian region is The Eastern Anatolian example is considered to be one of the best examples important in showing that not only plumes of a continental collision zone in the but also shallow plate tectonic processes world. It contains one of the highest have the potential to generate regional plateaus of the Alpine-Himalayan domal structures in the Earth's lithosphere mountain belt with an average elevation of as well as large volumes of magma. 2 km above sea level. Geological records indicate that the region was beneath sea level until Serravalian (~13 Ma), and then
73 The Hf-W Perspective on Whether a arc related non magmatic) processes and Trace of the Earth’s Core Exists in Hot display a mean silicate Earth ratio of 0.19 Spot Volcanic Rocks ±0.06 [4]. Using Hawaiian Th concentrations as a proxy for W and the A Scherstén (Faculteit der Aard- en petrogenetic model of Norman & Levenswetenschappen,Vrije Univ., De Garcia[5], a source around 10 ppb W is Boelelaan 1085, 1081 HV Amsterdam, estimated; on the same order as previously Netherlands; Tel:+31-20598-7335; E-mail: modelled [2]. Importantly, such a [email protected]); T. Elliott moderately depleted mantle source and C. Hawkesworth (Dept. of Earth indicates that recycling contamination of Sciences, Univ. of Bristol, Will’s high [W] crust to the Hawaiian source was Memorial Building, Queen’s Road, Bristol insignificant, and the sensitivity of the W BS8 1RJ, UK) isotope tracer is maintained. While Hf-W was only extant during a brief initial Radiogenic 187 Os/188 Os in ocean island period in Earth history, Re-Os and Pt-Os basalt is traditionally interpreted to are long-lived decay systems, and e.g. represent recycled crust, whereas coupled sulphides or metalliferous sediments may 186 Os/188 Os and 187 Os/188 Os fractionate Pt-Re-Os in a way that the enrichments were interpreted as a small observed Os-correlations can be generated. core contribution in e.g. Hawaiian lavas It thus appears as if the contact with the [1]. A core origin for radiogenic Os would core is lost, and that constraints on the establish that the Hawaiian mantle source origin of the at least the Hawaiian source originated at the core-mantle boundary. cannot be made from Os-isotopes. Importantly, coupled Os isotope enrichments hitherto only been found in 1. Brandon, et al., Science, 1998, hot-spot related lavas, possibly arguing 280, 1570-1573. against general core-mantle mixing and for 2. Schersten, et al., Nature, 2004. mantle plumes. The extinct Hf-W isotope 427, 234-237. system enables an independent test of the 3. Brandon & Walker, EPSL, 2005. core contribution interpretation because W 232, 211-225. partitions into the core while Hf remains 4. Newsom, et al., GCA, 1996. 60, in the mantle. Both elements are 1155-1169. refractory, and the bulk Earth Hf/W ratio 5. Norman, & Garcia, EPSL, 1999. is well established. A two parts in ten 168, 27-44. thousand 182 W/184 W ratio difference (i.e. 2 e182 W) between chondrite meteorites and silicate Earth [2 & refs. Thursday Poster Sessions therein] translates into a similar difference between the Earth’s core and mantle. The Evolution of Floreana Island, Because W is strongly enriched in the core Galápagos Archipelago II: The Result and incompatible during mantle melting of a Contaminated Mantle Plume and 182 Hf-decay to 182 W only occurred during the first ~50 Myr of solar system B. C. Christensen (Dept. of Geology, history, W isotope may be a very sensitive Colgate Univ., Hamilton, NY 13346, tracer of small core contributions to the USA; Tel: +1-845-242-8893; Fax: +1-315- Earth’s mantle. Hawaiian lavas do not 228-7187; E-mail: have any resolvable W isotope anomalies, [email protected]); K S which puts the core contribution Harpp (Dept. of Geology, Colgate Univ., hypothesis into question [2]. However, it Hamilton, NY 13346, USA; Tel:+1-315- was pointed out by Brandon & Walker [3] 228-7211; Fax:+1-315-228-7187; E-mail: that a recycled component may mask the [email protected]); A Koleszar, W isotope signature because of the Dept. of Geological Sciences, Brown significant W enrichment that e.g. Univ., Providence, RI 02912,USA; E- sediments may show. Th and W follow mail: [email protected]) each other very closely in magmatic (and
74 The Galápagos Archipelago has been isotopic ratios. Floreana’s mantle source attributed to a mantle plume since the has to have been enriched in Rb/Sr, U/Pb, hypothesis was proposed in the early and Th/Pb for much longer than12 m.y., 1970s. The plume is rooted ~250 km south the age of the underlying lithosphere. The of the Galápagos Spreading Center (GSC) anomalous characteristics of Floreana’s in the western archipelago, beneath melts can be explained simply with a Fernandina and southwestern Isabela heterogeneous mantle plume, enriched in Islands. Evidence for the archipelago’s elements commonly associated with plume origin includes: 1) islands and metasomatic activity. Recent models of seamounts exhibit an age progression plume ascent indicate that heterogeneities broadly consistent with eastward motion can be spatially preserved spatially during of the Nazca plate (e.g., White et al., ascent from the core-mantle boundary in 1993); 2) regional tectonic reconstructions filament-like structures. We propose that indicate that the GSC has migrated the source of Floreana magmas is northward over a stationary hotspot over metasomatized ancient lithosphere that has the past 8 m.y. (Wilson and Hey, 1995); 3) been recycled to the lower mantle. recent s-wave tomographic studies suggest the presence of a steeply-dipping, well- defined low velocity zone extending 400 Evolution of Helium Isotopes in the km beneath the western archipelago Earth’s Mantle (Toomey et al., 2001); 4) thinning of the transition zone beneath the western C. Class (Lamont-Doherty Earth archipelago suggests a thermal plume Observatory of Columbia Univ., Palisades, traverses through it (Hooft et al., 2004); 5) N.Y. 10964, USA; Tel: +1-845-365-8712; the Galápagos platform is underlain by Fax: +1-845-365-8155; E-mail: crust more than twice as thick as normal [email protected]); S L Goldstein, oceanic crust; 6) variation in geochemical Lamont-Doherty Earth Observatory of and isotopic signatures of Galápagos lavas Columbia Univ., Palisades, N.Y. 10964, can be explained as the result of mixing USA; Tel: +1-845-365-8787; Fax: +1-845- between a heterogeneous, three- 365-8155; E-mail: component plume and the depleted upper [email protected]) mantle (Harpp and White, 2001); 7) 3He/4He ratios vary up to nearly 30 times The presence of primordial 3He in ocean atmospheric values and are highest where island basalts (OIB) has been considered the plume is rooted (and there is no the primary evidence for the existence of a correlation between 3He/4He and [He]; primitive, unmelted reservoir in the deep Kurz and Geist, 1999; Graham et al., Earth. A new global data compilation 1993); and 8) geochemical, gravity, and relating helium isotopes of OIB and mid- bathymetric anomalies along the GSC ocean ridge basalts (MORB) to Sr-Nd-Pb peak when the ridge is closest to the isotopes and trace element abundances islands (e.g., Schilling et al., 2003). yields new insights into mantle dynamics. The compilation is based on data from the Floreana Island is located 150 km east online GEOROC and PetDB databases. (i.e., downstream in terms of plate motion) Oceanic islands are divided into groups of the root of the Galápagos plume. based on the highest 3He/4He ratios Floreana lavas have distinct geochemical measured in mineral separates: (1) “low and isotopic signatures from those of the 3He/4He” (3He/4He <7 RA); (2) “MORB- western shield volcanoes, and Floreana like 3He/4He” (8±1 RA); (3) “moderately magmas originate from shallower depths high 3He/4He” (9-15 RA); (4) “high as well. Floreana basalts are enriched in 3He/4He” (>15 RA). These designations radiogenic Sr and Pb, and depleted in enable us to follow the geochemical 3He/4He ratios relative to the western characteristics of OIB sources based on shield volcanoes, which are interpreted to their 3He/4He ratios despite the global represent the pristine plume signature; paucity of combined He-Sr-Nd-Pb isotope significantly less variation exists in Nd data on individual samples. The
75 geochemical compositions of the four 3952-511659; Fax: +7-3952-426900; E- groups of OIB are not distributed mail: [email protected]); S V randomly within the global range but show Rasskazov (Institute of Earth's Crust SB the following systematics: (i) lower RAS, Lermontov st., 128, Irkutsk, 664033, 143Nd/144Nd ratios are associated with Russia; Tel: +7-3952-511659; Fax: +7- lower 3He/4He, (ii) for a constant 3952-426900; E-mail: [email protected]) 143Nd/144Nd ratio, higher 206Pb/204Pb ratios are associated with lower 3He/4He Analyses of Sr-Nd isotopes in the Late groups, and (iii) low 3He/4He ratios are Cenozoic basaltic lavas of a vide territory accompanied by low 206Pb/204Pb, of the Central Asia (e.g. Hangai, East 207Pb/204Pb, 208Pb/204Pb ratios and Sayan, Khamar-Daban) revealed three high Th, U abundances independent of distinct components (end-members) partial melting effects. referred to as A, B and C (Rasskazov et al., 1999). The component A is moderately The new compilation shows there is a depleted (87Sr/86Sr ~0.704, direct relationship between the 3He/4He 143Nd/144Nd ~0.5128), whereas groups and Th+U contents of OIB (, with components B (87Sr/86Sr ~0.7045, only Samoa as an exception). Thus, 143Nd/144Nd ~0.5122) and C (87Sr/86Sr 3He/4He in OIB appear to reflect the ~0.7053; 143Nd/144Nd ~0.5127) are production rates of 4He from recycled enriched (Rasskazov et al., 2002; Barry et oceanic crust plus sediment variably al., 2003; Yarmolyuk et al., 2003). Despite enriched in (Th+U) in plume sources. OIB the good concordance of the published displaying the strongest primordial data, the interpretations between the 3He/4He signal are chemically and different studies are remarkably different. isotopically (i.e. Sr, Nd, Pb) most like Rasskazov et al. (2002) and Barry et al. mid-ocean ridge basalts. Thus the global (2003) suggest that component A belongs compilation confirms that OIB and MORB to convective sublithospheric mantle and sample mantle regions with different components B and C reflect composition helium isotope signatures. However, the of lithospheric mantle. Acknowledging compilation indicates a common melting lithospheric features of the B and C history for the sources of MORB and OIB components, Yarmolyuk et al. (2003) over geological time. We show through propose that these are recycled lithosphere modeling that helium isotopes in the that has resided long term in the deep mantle can be explained through mantle (lower-upper mantle transition and continuous, incomplete degassing by D’’ layer) and that these components are continent and ocean crust formation. It sampled by a plume. For the component A appears that the high 3He/4He component they suggest a deep source of the lower in plume sources does not represent mantle. In this presentation we apply unmelted primitive mantle but rather “old” inverse trace element modelling to depleted mantle, isolated from convection distinguish mineralogy of the mantle and upper mantle degassing for 1-2 billion sources for these isotopic components. We years. use Rb as the most suitable highly incompatible trace element and plot individual datasets for each component on Inverse Trace Element Modeling of Rb/Ci…n – Rb diagrams, where Ci…n are Mantle Components from Late concentrations of different trace elements. Cenozoic Basalts in Central Asia Slope and intercept for individual datasets on such diagrams provide information on E. I. Demonterova (Institute of Earth's concentrations of elements in their sources Crust SB RAS, Lermontov st., 128, and bulk distribution coefficients for these Irkutsk, 664033 Russia; Tel: +7-3952- elements (Ormerod et al., 1991). For 511659; Fax: +7-3952-426900; E-mail: component A we find significant [email protected]); A V Ivanov (Institute correlations for Ci…n elements; Ba, K, Sr, of Earth's Crust SB RAS, Lermontov st., Pb, Hf, Ti, Y and REE (except La and Ce) 128, Irkutsk, 664033 Russia; Tel: +7- and for component C – for U, Ta, Sr, P,
76 Zr, Ti, Y and REE. For component B we Mantle components in Late Cenozoic find only limited number of correlations volcanics, East Sayan (from Pb, Sr, and for Ci…n elements; Sr, Pb, P, Ti and Y, Nd isotopes). Russian Geology and showing that basalts with such Sr-Nd Geophysics, v. 43, p. 1065-1079.; isotopic features do not belong to the same Yarmolyuk, V.V., Ivanov, V.G., uniform dataset (therefore, do not belong Kovalenko, V.I. And Pokrovskii B.G., to the same mantle source). Component C 2003, Magmatism and geodynamics of the is characterized by gradual increase of Southern Baikal volcanic region (Mantle slopes and decrease of intercepts in order Hot Spot): results of geochronological, of REE compatibility. Heavy REE and Ti geochemical, and isotopic (Sr, Nd, and O) exhibit negative intercepts, implying investigations. Petrology, v. 11, p. 1-30. distribution coefficients above unity for these elements. Component A shows gradual increase of slopes with almost the Fluid Inclusion Evidence for Water in same positive intercepts for most of REE. the Mantle Beneath Hawaii Exception is a negative intercept for Yb. Intercept for Ti is positive. On basis of the Maria Luce Frezzotti (Dipartimento di inverse trace element modeling we suggest Scienze della Terra, Univ. di Siena, Via that component A is within garnet-spinel Laterina 8, 53100 Siena, Italy; Tel: +39- transition stability field (about 55-65 km 0577-233929; Fax: +39-0577-233938; E- depths) without Ti-bearing phases. The mail: [email protected]); Angelo component B is shallower component of Peccerillo (Dipartimento di Scienze della heterogeneous lithosphere. Component C Terra, P.za Univ. 1, 06100 Perugia, Italy; is deeper sublithospheric mantle Tel: +39-075-5852608; Fax: +39- 075- component within garnet stability field 5852603; E-mail: [email protected]) with Ti-bearing phases. The work is supported by RFBR grant 05-05-64477. Post-erosional alkalic lavas from Salt Lake Crater, on the apron of the Koolau volcano References: Barry, T.L., Saunders, A.D., on Oahu (Hawaii) contain a rather unique Kempton, P.D., Windley, B.F.,Pringle, group of mantle xenoliths, representing the M.S., Dorjnamjaa, D. and Saandar, S., oceanic lithosphere. In order to have 2003, Petrogenesis of Cenozoic basalts information on the composition of free from Mongolia: evidence for the role of fluids present in the Hawaiian solid mantle asthenospheric versus metasomatised at depths > 50-60 km, we have analyzed lithospheric mantle sources. Journal of fluid inclusions in garnet pyroxenites, Petrology, v. 44, p. 55-91.; Ormerod, D.S., which are part of a collection on loan from Rogers, N.W., Hawkesworth, C.J., 1991, the Smithsonian Institutions. Melting in the lithospheric mantle: Inverse modelling of alkali-olivine basalts from Studied pyroxenites are dry and consist of the Big Pine Volcanic Field, California. clinopyroxene, orthopyroxene, olivine and Contributions to Mineralogy and garnet. Abundant fluid inclusions are Petrology, v. 108, p. 305-317; Rasskazov, present in all mineral phases, mostly S.V., Bowring, S.A., Harris, N., Brandt, contained within intergranular and I.S., Brandt, S.B., Ivanov, A.V., Lhur, J.F., intragranular trails, indicating late Coleman, D. and Housh, T., 1999. trapping. In ortho- and clinopyroxene, Lithospheric domains in East Sayan: Pb, however, a few early fluid inclusions are Sr, and Nd isotope constraints on sources preserved; these are present isolated, or of the Late Cenozoic magmatism in the form small clusters within single grains. In northern segment of the Riphean Tuva- clinopyroxene, early inclusions predate Mongolian massif. In: Rifting in orthopyroxene and spinel exsolutions. intracontinental setting: Baikal Rift Most early inclusions are decrepitated; System and other Continental Rifts, previous studies (Frezzotti et. al., 1992) Irkutsk-Tervuren, p. 163-166.; Rasskazov, indicated that a few small (< 8 µm) early S.V., Saranina, E.I., Demonterova, E.I., inclusions still preserve high-density to Maslovskaya, M.N., Ivanov, A.V., 2002, superdense CO2 fluids, with densities up
77 to 1.21 g/cm3 - the highest density ever Various intraplate volcanic fields, large recorded in CO2 mantle fluids - igneous provinces, and continental flood corresponding to pressures of 1.8-2 GPa, basalt provinces have long been at the inferred mantle temperatures of considered to be the result of mantle 1100°C. plumes interacting with the lithosphere. This idea has been strengthened by the A new detailed Raman study in selected recognition that the geochemical and high-density and superdense early CO2 isotopic compositions of intraplate inclusions reveals vibrational bands at magmatism differ significantly from those 3638 cm-1 and 2609 cm-1, which are of mid-ocean ridge and arc settings. characteristic of OH- stretching vibrations Regional and global tomographic evidence for isolated water molecules, and of H2S, of the presence of mantle plumes in a respectively. A few inclusions additionally number of locations is mounting, however, show the typical vibration for diamond. not all intraplate magmatism fits the mantle plume model. Ad-hoc refinements Fluids contained within inclusions are to the original models of Wilson and CO2-rich, and contain H2O and H2S, ± Morgan to explain anomalous diamond, defining a complex fluid mixture characteristics has recently prompted a in the C-O-H-S system in the mantle global debate within the geoscience beneath Hawaii, at P = 2 GPa. The non community ranging from the number of systematic presence of diamonds within mantle plumes that exist, to the very fluid inclusions, strongly suggest that existence of mantle plumes at all. An these phases did not form within fluid outgrowth of this ongoing debate has been inclusions. Diamonds were already the proposal of a number of alternative present in CO2-H2O-H2S mantle fluids, models that attempt to explain the indicating that inclusions trapped a deep existence of intraplate magmatism in the fluid, which originated in the absence of a mantle plume source. The asthenosphere. Balcones Igneous Province (BIP) of south central Texas provides an excellent We argue that such fluids, migrating opportunity to investigate magmatic within the asthenospheric mantle represent processes that occur in a continental a strong metasomatic agent able to induce intraplate setting. An integrated field, partial melting at the base of the geochemical, and geochronologic study is lithosphere at normal mantle temperatures, in progress in the BIP to understand the obviating the need for concentrated hot petrogenetic processes and magmatic jets localized under “hotspot” volcanoes, emplacement mechanisms responsible for and deep mantle reservoirs. its formation. The BIP is an arcuate zone of Late Cretaceous (86-77 Ma) intraplate volcanic and intrusive igneous bodies Testing Magmatic Emplacement approximately 400 km in length by 100 Mechanisms in the Balcones Igneous km in width. Initial geochemical analyses Province of Texas (n=12; major element, trace element, rare- earth element, radiogenic isotope (Sr, Nd, W. R. Griffin Geosciences Dept., Univ. of Pb) and mineral composition) from the Texas at Dallas, Richardson, TX 75083, BIP suggest magmatism resulted from USA; Tel: +1-972-883-2401; Fax:+1-972- small degrees of partial melting in the 883-2537; E-mail: [email protected]; S garnet stability zone of a depleted mantle C Bergman, Geosciences Dept., Univ. of source that had experienced re-enrichment Texas at Dallas, Richardson, TX 75083, in incompatible trace elements. Although USA; Tel: +1-972-883-2401; Fax: +1-972- previous workers have suggested BIP 883-2537; E-mail: [email protected]; magmatism may be related to OIB type M I Leybourne, Geosciences Dept., Univ. sources, it is likely that several upper of Texas at Dallas, Richardson, TX 75083, mantle magmatic processes were involved. USA; Tel: +1-972-883-2403; Fax:+1-972- Initial geochronological analyses (n=7; 883-2537; E-mail: [email protected]) U/Pb SHRIMP, 40Ar/39Ar) from the BIP
78 demonstrate magmatism occurred over a The Galápagos Archipelago poses a much shorter time-period than previously number of challenges to the mantle plume thought. Regional geophysical data hypothesis: 1) volcanism is not indicate the BIP coincides with the inner consistently time transgressive, with many margin separating Neoproterozoic (1.3 Ga) exposed lavas exhibiting ages younger cratonic lithosphere, from much younger than those predicted by plate motion (Paleozoic to Mesozoic) transitional vectors (e.g., White et al., 1993). This lithosphere of the southern margin of the pattern of volcanism that is “too young” North American continent. Uncertainties extends up to 500 km away from the exist regarding the lithospheric transition hotspot, along the Cocos and Carnegie zones to the west into Mexico and Trans- Ridges; 2) a complex array of volcanic Pecos (TP) Texas, because Mesozoic and alignments throughout the region indicates Cenozoic tectonic events have effectively control by lithospheric stresses, likely masked much of the earlier evidence in the related to the transform fault at 91°W on region. Our work has produced new the Galapagos Spreading Center (GSC; geochronological evidence from TP Texas Harpp and Geist, 2002); 3) the Cocos and that demonstrates mafic magmatism was Carnegie aseismic ridges that emanate present in the region at 73 Ma (U/Pb from the archipelago do not conform to SHRIMP), similar in age to that of the subsidence trends predicted by normal BIP, suggesting the BIP may extend 400 cooling of a reheated lithosphere; 4) km further west that previously several Galápagos volcanoes erupt lava recognized. Alkaline silica undersaturated that is isotopically indistinguishable from igneous complexes in northern Coahuila, MORB produced at the GSC, some with Mexico appear to provide a spatial link isotopic signatures similar to those between the south Texas and TP Texas observed at the GSC over 800 km away segments. In a broader regional context, from the archipelago; 5) much of the the BIP may represent a small portion of a geochemical variation in the archipelago Middle to Late Cretaceous Gulf Coast can be attributed to low degree, deeper diffuse igneous province, as large as 1400 melts of enriched upper mantle x 200 km, comprising igneous centers in heterogeneities; and 7) 3He/4He ratios central Arkansas, northeast Louisiana and resemble those of MORB (or lower) in central Mississippi. Similar geochemical many parts of the archipelago. compositions and emplacement ages across the region support this idea. Floreana Island is located 150 km east (i.e., downstream in terms of plate motion) of the most active volcanoes, and exposed The Evolution of Floreana Island, lavas are up to 1.1 Ma. Eruptive vents on Galápagos Archipelago I: The Result of Floreana are aligned parallel to lineaments Upper Mantle Heterogeneities observed throughout the eastern archipelago. Floreana lavas have K. S. Harpp (Dept. of Geology, Colgate geochemical and isotopic signatures Univ., Hamilton, NY 13346, USA; Tel: distinct from those erupted elsewhere in +1-315-228-7211; Fax: +1-315-228-7211; the archipelago, and unlike the other E-mail: [email protected]); B.C. Galapagos volcanoes, lack a garnet Christensen (Dept. of Geology, Colgate signature. IFloreana basalts are enriched in Univ., Hamilton, NY 13346, USA; Tel: radiogenic Sr and Pb ratios relative to the +1-845-242-8893; Fax: +1-315-228-7211; western shield volcanoes. Floreana’s E-mail: [email protected]); mantle source has been enriched in Rb/Sr, D J Geist (Dept. of Geological Sciences, U/Pb, and Th/Pb for much longer than 12 Univ. of Idaho, Moscow, ID, 83844, USA; Ma, the age of the underlying lithosphere. E-mail: [email protected]); A Koleszar Contributions from this enriched source (Dept. of Geological Sciences, Brown decrease systematically to the north, west, Univ., Providence, RI 02912, USA; E- and east of Floreana. The anomalous mail: [email protected]) characteristics of Floreana’s melts can be explained as the result of melting of an
79 upper mantle heterogeneity, enriched in basalt province, which is commonly elements commonly associated with attributed to arrival of the ancestral metasomatic activity; metasomatism was Iceland plume at ~55-57 Ma. To test the ancient, however, and cannot be contained relative importance of melting systematics, within the young lithosphere. Regional source composition, and oxygen fugacity lithospheric stresses initiated by the 91°W on the Sc/V systematics for NAIP basalts transform fault initiate melting of the in both space and time, we incorporated enriched asthenospheric material beneath both Sc and oxygen-fugacity-dependent V Floreana, which erupt along lines of mineral-melt partitioning data of [5] into tensional stresses. The enriched anomaly the polybaric decompression melting has a low solidus, which causes melting at model REEBOX [6]. Our model captures shallow levels, where it would not the salient compositional features of CEG ordinarily occur otherwise. and Iceland basalts, which are related to the role of residual garnet in the mantle source and differences in fO2. To model Mantle Redox Conditions in LIPs: the entire range of east Greenland lavas Constraints from the North Atlantic requires that the Paleogene mantle source Igneous Province was ~0.5 log units more oxidized than the Iceland source. These differences may be L. E. Heister (Dept of Geology, Univ. of attributed to a change in the composition California, Davis, CA 95616, USA; of the Iceland plume or reflect the Tel:+1-530-752-5041; Fax:+1-530-752- involvement of metasomatized upper 0951; E-mail: mantle associated with Iapetus subduction [email protected]); C E Lesher in the formation of the east Greenland (Dept. of Geology, Univ. of California, basalts. We further examine the Davis, CA 95616, USA; Tel: +1-530-752- compositional variability in mantle redox 9779; Fax: +1-530-752-0951; E-mail: conditions throughout the NAIP during [email protected]) continental break-up. Regions distal to the axis of the proposed Iceland plume and The North Atlantic igneous province Caledonian front appear to be (NAIP) has long been viewed as a region fundamentally different from central east of anomalous mantle upwelling related to Greenland, suggesting that the upper plume activity, continental rifting, and a mantle in these areas was either more heterogeneous mantle source. Prior to oxidized or markedly different in continental rifting in the Tertiary, the composition. This study of the NAIP northern portion of the region was the site provides the basis for comparisons of of closure of the Iapetus ocean basin. This redox conditions in other LIPs. tectonic event may have contributed to heterogeneities within the upper mantle [1] Canil, D., 1997, [2] Canil, D. 1999, and altered its oxidation state relative to [3] Li and Lee, 2004 [4] Lee et al., 2005 the ambient mantle. Vanadium has been [5] Canil and Fedortchouk, 2000, [6] Fram shown to be a useful indicator of redox and Lesher, 1993 conditions due to its multiple valence states (e.g. [1-2]), and the vanadium to scandium ratio (V/Sc) for basalts from Plumes or Rifting? : The Mesozoic mid-ocean ridge and arc environments has Dykes of the Falkland Islands and Their been proposed as a useful proxy for fO2 Relationship to the Break-up of conditions during partial melting (e.g. [3- Gondwana. 4]). We compare V/Sc ratios from basalts from early Tertiary successions in the Malcolm J. Hole (Dept. of Geology & NAIP with those from Iceland, and relate Petroleum Geology, Univ. of Aberdeen, the variations to differences in both redox King’s College, Aberdeen AB24 3UE, conditions and melting systematics. We UK; Tel: +1224 273434; Fax:+1224 examine these relationships in detail for 273488; E-mail: [email protected]); the central east Greenland (CEG) flood R.M. Ellam (Isotope Geosciences Unit,
80 Scottich Universities Environmental Gonwana break-up rather than a mantle Research Centre, East Kilbride, G75 plume of apparently diverse composition. 0QF,UK; E-mail: [email protected]); S.P. Kelly (Dept. of Earth Siences, Open Univ., Pliocene-Quaternary Alkaline Basalts of Walton Hall, Milton Keynes, MK77,UK; the Sredinny Ridge of Kamchatka: E-mail: [email protected]); D.I.M Evidence for Melting of Recycled Macdonald (Dept. of Geology & Oceanic Crust in Tectonic Setting of a Petroleum Geology, Univ. of Aberdeen, Modern Island Arc System King’s College, Aberdeen AB9 2UE, UK; Tel: +1-224-273434; Fax:+1-224-273488; A. V. Ivanov (Institute of the Earths Crust, E-mail: [email protected]) Siberian Branch, Russian Academy of Sciences, Lermontov st. 128, 664033 Many researchers consider mantle plumes Irkutsk, Russia;Email: as being the driving force behind [email protected]); A B Perepelov continental fragmentation. Here we report (Vinogradov Institute of Geochemistry, new major and trace element, Sr-, Nd- and Siberian Branch, Russian Academy of Pb-isotope data a Ar-Ar geochronogical Sciences, Favorsky st. 2, 664033 Irkutsk, data for mafic Mesozoic dykes of the Russia); M Yu Puzankov (Institute of Falkand Islands. All analysed samples are Volcanic, Geology, and Geochemistry, of the low TiO2 magma type recognized Far-East Branch, Russian Academy of throughout Gondwana. New 40Ar-39Ar Sciences, Boulevard Piipa 1, data on plagioclase phenocrysts separates Petropavlovsk-Kamchatsky, Russia) from three intrusions yield a single statistically viable age of 182.3±1.5 Ma, Volcanic rocks of island-arc tectonic concordant with ages reported from the settings are characterized by definitive low-Ti Karoo lavas of Lesotho and trace element features such as depletion in Lebombo. Within the Falkland Islands, high field strengths (HFS) elements four distinct magma types are recognized. relative to large ion lithophile (LIL) One group of intrusions exhibit Sr-, Nd-, elements. This is interpreted as evidence and Pb- isotope ratios (87Sr/86Sr = 0.707 for transport of LIL elements from to-0.714; eNdt -5 to -12; 206Pb/204Pb subducting slab by aqueous fluids and 17.75-18.25), high La/Ta (44-56), and low enrichment of overlying mantle wedge by Nb/Zr ratios (0.03-0.07) which are these elements. HFS elements are not consistent with significant interaction with soluble in the aqueous fluids and therefore Gondwana lithosphere by the process of mantle wedge is relatively depleted by AFC, and are typical of many of the these elements. Alkaline nepheline- Jurassic magmatic recognized throughout normative volcanic rocks without HFS- Gondwana (e.g Ferrar Province). The depletion pattern, though often found in remaining three groups, however, exhibit the island arc tectonic setting, are considerably more depleted Sr- and Nd- considered as unusual or atypical. Here we isotope ratios (87Sr/86Sr 0.703-0.706; investigate the atypical Pliocene- eNdt +4 to-2) indicating more limited, or Quaternary alkaline basalts of the indeed, negligible, interaction with Sredinny Ridge of Kamchatka to reveal Gondwana lithosphere. This lack of their structural position. We also apply lithospheric input to magmatism gives us a inverse element modeling to show clear window into the mantle source difference in source of melting between region from which the magmas were typical island-arc HFS-element depleted derived. The MORB-like geochemistry and atypical alkaline magmas. In the of these isotopically depleted magmas Kamchatka there are two (frontal and rear) does not necessarily require a plume Pliocene-Quaternary volcanic belts with origin. It is therefore conceivable that plate typical HFS-element depleted rocks. These boundary forces were the most important volcanic belts are subparallel to the driving force for magmatism during seismofocal zone and are conventionally considered as being related with two
81 regions of the mantle wedge above water- basaltic volcanism of the Sredinny Ridge, bearing mineral decomposition fields Kamchatka: case study of the within the modern subducting slab. Coeval Payalpanvolcano-tectonic structure. In: alkaline volcanic rocks were found mainly Metallogeny of the Pacific Northwest: within the rear volcanic belt. These tectonics, magmatism and metallogeny of alkaline rocks are characterized by trace active continental margins, edited by A.I. element pattern similar to continental rift Khanchuk et al. Dalnauka, Vladivostok, p. basalts and oceanic island basalts. They 345-349. Portnyagin, M., Hoernle, K., are referred to as within-plate, riftogenic Avdeiko, G., Hauff, F., Werner, R., or oceanic type basalts by different authors Bindeman, I., Uspensky, V. and Garbe- (e.g. Volynets, 1994; Churikova et al., Schonberg, D., 2005, Transition from arc 2001; Ivanov et al., 2004; Portnyagin et to oceanic magmatism at the Kamchatka- al., 2005). Using TiO2 content in basalts Aleutian junction. Geology, v. 33, p. 25- as a proxy for the within-plate- 28. Volynets, O.N., 1994, Geochemical geochemical type, we distinguish about types, petrology, and genesis of Late fifty within-plate-type volcanic centers in Cenozoic volcanic rocks from the Kurile- the Sredinny ridge (used authors data Kamchatka island-arc system. among geological survey and literature International Geology Review, v. 36, p. data). These volcanoes form five linear 373-405. chains, which are oblique to the general trend of the rear volcanic belt. These volcanic chains are subperpendicular to Communicating the Plume Debate to the axis of Pacific plate movement (hence Undergraduate Geoscience Students they are subparallel to magnetic anomalies of the Pacific plate crust). еДХМХВМШЕ B. T. Jordan (Dept. of Geology, Whitman within-plate type volcanoes in south College, Walla Walla, WA 99362, USA (Bakening, Taburetka) and north Tel: +1-509-527-4934; Fax: +1-509-527- (Nachikinsky), which do not belong to 5904; E-mail: [email protected]) these linear chains of the Sredinny Ridge, situate above surface projection of The mantle plume model is presented as transform faults. Using new and published the explanation for age-progressive trace element data obtained for basalts intraplate volcanism (e.g. Hawaii) and along the rear volcanic belt we distinguish anomalous plate margin volcanism (e.g. two pure end-member types; the island-arc Iceland) in most introductory geology type and within-plate type with spectrum textbooks. This model is also presented in of transitional compositions. Applying igneous petrology and tectonics textbooks. inverse trace element modeling for the Should alternatives be presented to end-member types, we found that typical undergraduate geoscience students? island-arc HFS-element depleted magmas were generated from a spinel-bearing Dutch (1980; J. Geo. Ed., v. 30, p. 6-13) mantle source (40-60 km depths), whereas described a three-tiered hierarchical atypical within-plate magmas originated classification of theories: central, frontier, from a deeper garnet-rich source. Eclogite and fringe. All theories begin as fringe (recycled oceanic crust) is the best theories; most will be discarded. Some are candidate for this garnet-rich source. The supported but have some unresolved work is supported by RFBR grants 04-05- inconsistencies and/or serious alternatives; 64800 and 05-05-64477. References: these are frontier theories which constitute Churikova, T, Dorendorf, F. and Worner the mainstream thinking of a scientific G., 2001, Sources and fluids in the mantle discipline. Central theories are no longer wedge below Kamchatka, Evidence from seriously disputed, and form the across-arc geochemical variation. Journal foundation of a discipline. Plate tectonics of Petrology, v. 42, p. 1567-1593.; Ivanov, is a central theory in geology and mantle A.V., Perepelov, A.B., Puzankov, M.Yu., plume theoryis a frontier theory. Yasnygina, T.A., Malykh, Yu.M. and Rasskazov, S.V., 2004, Rift- and arc-type
82 A quick review of introductory and Earth’s surface is by transfer of material physical geology textbooks reveals that from the core-mantle boundary, a process most treat mantle plume theory as a central that provides evidence for the existence of theory. Only Press et al. (2004; deep mantle plumes. In this contribution, Understanding Earth 4th edition; WH we test the above geochemical hypothesis Freeman Co.) suggest a by performing Re-Os-Pt partitioning controversy. Other authors should follow experiments between sulfur-bearing their lead. Some might argue that solid/molten iron metal. Pressure and describing the uncertainty of such a well temperature, besides bulk composition, are known theory could weaken public potentially important factors in controlling perception of the state of the science. the partitioning systematics, and Others might argue that the foundations of experiments were designed to cover a the debate are too sophisticated for this wide, and so far the largest (3-22 GPa) level. Students are introduced to the idea pressure and temperature (1300-1775 C) that anomalous temperature is just one of range. It must also be noted that the three ways to cause melting, so that hot highest-pressure datum in this work is still spots might not be hot and may be the more than an order of magnitude less than result of something other than a plume the pressure at the outer-inner core should be comprehensible and reinforce boundary (330-360 GPa). The starting mix other learning. These ideas can be was prepared from high-purity Fe-metal explored in greater depth in upper division and FeS (7 wt% S in the bulk), and courses. That there is debate about the metallic Re-Os-Pt (each 1 wt%), and basic workings of the earth should be finely ground in an agate mortar. exciting to students; there are big Experiments were designed to have 8-21 questions that remain to be addressed in wt% sulfur in the liquid metal (Sliq), their careers. within the range proposed for the outer core. Experiments were performed using MA6/8 multianvil modules with 18/11 Re-Os-Pt Partitioning in Sulfur-bearing (Cast-MgO cell; graphite furnace; zirconia Solid/Molten Iron Metal at 3-22 GPa insulator) and 8/3 (Cr-MgO cell; Re and 1300-1775 C: Is the Earth’s Outer furnace; LaCrO3 insulator) pressure cells, Core So Giving? Type C thermocouples, and MgO capsules. To ensure bulk homogeneity, Shantanu Keshav (Geophysical Lab, each charge was taken above the liquidus Carnegie Institute, Washington, DC temperature, kept there for at least 30 20015, USA; Tel: +1-202-478-8929; Fax: mins, and then brought to the target +1-202-478-8901; E-mail: temperature at 5-10 C/min. At the target [email protected]); James A. Van temperature, run duration was 3-4 hours. Orman (Geol. Sc. Case Western Reserve After each experiment, the sample was Univ., Cleveland, OH 44106, USA) mounted in epoxy and longitudinally ground for optical and electron microprobe Recently, some high-MgO basaltic lavas analyses. Concentrations of all the have been shown to exhibit coupled elements were determined using JEOL- enrichments in 186 Os/188 Os and 187 8800 JXA microprobe with WD Os/188 Os. One way to have this spectrometry. Owing to slow diffusive enrichment is by strong fractionation of Re equilibrium in metal-bearing systems, and Pt (preferring liquid metal) from Os analyses were done within 10-15 micron (retained more in the solid metal) during of the interface, and 8-35 analyses were crystallization of the inner core. By obtained. Spot and raster modes were fractionating Pt and Re from Os, the utilized to analyze solid and liquid metal, Earth’s outer core would develop highly respectively, with analyses covering radiogenic 186 Os/188 Os and 187 Os/188 almost the entire portion of the liquid part Os ratios, and it has been suggested that of the charge. All the charges were free of perhaps the only way to have the lavas metallic nuggets. The liquid metal phase with this enriched component on the was identified by its dendritic texture, and
83 the solid and liquid metal phases have a ways to do post-mortem on tiny, diamond- sharply defined boundary. The retrieved anvil cell samples to properly understand partition coefficients (Dsolid-metal/liquid- the partitioning systematics at ultra-high metal) show a strong positive dependence pressures, narrowing the huge uncertainty on the sulfur content (9-20 wt%) of the in the nature of light element(s) in the liquid metal. Between 3-22 GPa, and at outer core by combining experiments with similar Sliq (12 wt%), values of DOs, cosmochemistry, and the effect of a DRe, and DPt range from 6-8.5, 4.6-5.9, multitude of light dilutant on D and 3.2-3.3, respectively, and thus it systematics. It would also be beneficial to appears that pressure does not have a define the possible reservoirs of Re, Os, remarkable effect on the D values. It is and Pt in the most accessible upper mantle important to note that even though it is rocks and the factors controlling the necessary to know the absolute D values, distribution and fractionation of these it is the relative fractionation, that is Pt/Re, elements in a variety of tectonic as well as Pt/Os, and Os/Re that we are most rock settings, to ultimately help us concerned with, as it is the fractionation of understand the Os signals. Os from Pt and Re that is necessary to generate enriched 186 Os/188 Os and 187 Os/188 Os ratios. In the pressure range Recycling of Archean Peridotitic studied, with increasing Sliq, DOs/DRe Komatiite in the NW Kyushu Source remains virtually unchanged. Also, at a given Sliq, DOs/DRe seems to be pressure H. Mashima (Tono Geoscience Center, independent, implying that with increasing Japan Nuclear Cycle Development pressure, Os and Re do not readily Institute, Jorin-ji 9598-31, Izumi-cho, fractionate from each other. At 3 GPa, Toki-shi,Gifu, 509-5102, Japan; Tel: +81- DPt/DOs and DPt/DRe actually decrease 572-53-0211; Fax: +81-572-55-0180; E- with increasing Sliq, indicating, that Pt mail: [email protected]) significantly fractionates from both Re and Os. Importantly, however, at a given Sliq, Based on Sr-Nd-Pb isotope and trace DPt/DOs and DPt/DRe increase with element features of the NW Kyushu pressure, indicating, that Pt is actually basalts, their sources were considered to strongly becoming more compatible in the be mixtures of depleted upper mantle and solid metal. The results above are exactly recycled components such as ancient the opposite to what is envisaged in the oceanic crust. plume scenario, we are testing here. Thus, if the trend of increasing DPt/DOs and Nevertheless, recent high-pressure melting DPt/DRe and the relative constancy of experiments of basalt/peridotite hybrids DOs/DRe described above continues to have revealed that partial melting of the depths of outer-inner core levels in the basalt/peridotite hybrids cannot form NW planet, then it becomes even more difficult Kyushu basalts since these basalts are to suitably, and rather strongly, fractionate enriched in Fe and depleted in Al Pt and Re from Os, as needed in compared with experimental partial melts geochemical models. Hence, on the basis of basalt/peridotite hybrids. Results of of the experimental data presented here, it previous high pressure melting seems impossible to generate the required experiments indicate that cpx in the NW fractionation of Re, Os, and Pt from each Kyushu source is the first disappearing other (in a core that contains sulfur as the phase at low pressure, and that the NW principal light element), and an outer core, Kyushu source has high FeO* compared Os isotopic component in some basaltic with typical peridotites. Among terrestrial lavas can be dismissed. Possible future materials used as starting materials in endeavors to more severely address the Os experiments, a peridotitic komatiite is only signals should focus more on cautiously one with these characteristics. Model pushing the limits of the multianvil calculations suggest that partial melting of modules to higher pressures by using Archean peridotitic komatiite can sintered diamond cubes, perhaps more essentially explain major element, trace
84 element and Sr-Nd-Pb isotope features of lithosphere and a geologically young NW Kyushu basalts. Thus, the simplest Cenozoic oceanic lithosphere. explanation for the NW Kyushu source is recycling of Archean oceanic crust with The ODP Leg 104, Site 642E cores of the peridotitic komatiite composition. Vøring Plateau consist of a Lower Series (LS) of basic dykes, glassy andesite/dacite lava flows, and interbedded volcaniclastics The Vøring Plateau Volcanic Margin : which underlies a 770 m thick Upper A Key Rock Succession to Understand Series (US) of transitional-type mid-ocean Continental Breakup During the Initial ridge basalts. The resampling strategy of Stages of the Opening of the NE- core 642E in the framework of a Atlantic EUROMARGINS sub-project aimed to attain an improved sampling density of the R. Meyer (Afd. Fysico-Chemische heterogeneous LS and the transition to the Geologie, Katholieke Universiteit Leuven, homogeneous US. The degree and Celestijnenlaan 200C, 3001 Heverlee, processes of mantle-crust interaction are Belgium; Tel: +32-16-327593; Fax: +32- investigated by ICP-MS, TIMS and multi- 16-327981; E-mail: collector ICP-MS trace element and [email protected]); J radiogenic isotope geochemical analyses. Hertogen (Afd. Fysico-Chemische The US tholeiitic MORB flows are Geologie, Katholieke Universiteit Leuven, characterized by slightly LREE enriched Celestijnenlaan 200C, B-3001 Heverlee, and consistently parallel convex C1- Belgium; Tel: +32(0)16327587; Fax: normalized REE patterns. The REE +32(0)16327981; E-mail: compositional differences are due to [email protected]); R B different degrees of partial melting of a Pedersen (Institutt for geovitenskap, Univ. uniform source, coupled to variable extent i Bergen, Allegaten 41, N-5007 Bergen, of fractional crystallisation of the mantle Norway; Tel: +47-555-83-517; Fax: +47- melts. On the contrary, the tholeiitic basalt 555-89-416; E-mail: dykes of the LS have somewhat concave [email protected]); L Viereck- MORB-like REE patterns. The LS basaltic Götte, Institut für Geowissenschaften, andesitic and dacitic flows are more Friedrich-Schiller-Universität Jena, enriched in LREE, forming steep concave Burgweg 11, D-07749 Jena, Germany; C1-normalized REE patterns. The samples Tel: +49-3641-948720; Fax: +49-3641- have a negative Eu anomaly and can be 948662; E-mail: lothar.viereck- interpreted as resulting from interaction of [email protected]); M Abratis (Institut mantle melts with crustal material and/or für Geowissenschaften, Friedrich-Schiller- of significant crustal melting by crustal Universität Jena, Burgweg 11, D-07749 underplating. Isotope data of Leg 104 Jena, Germany; Tel: +49-3641-948721; samples from the LS and US show a Fax: +49-3641-948662; E-mail: dichotomy in terms of Sr and Nd isotopic [email protected]) composition. The US is isotopically relatively homogeneous with typical The Vøring rifted passive margin is mantle-source derived isotope ratios located at the central segment of the (87Sr/86Sr ± 0.703) while the LS shows Norwegian Margin and forms part of the more radiogenic ratios and a higher North Atlantic Igneous Province. Drilling variability (e.g. 87Sr/86Sr > 0.710). at the Vøring Plateau and the SE Greenland margin during the Ocean The marked diversity of the US and LS Drilling Program, recovered volcanic rock samples from Site 642E offers successions that erupted during the initial opportunities to distinguish geochemical stages of breakup of Greenland from signatures related to crustal contamination Fennoscandia. Hence, the mid-Norwegian from those related to intrinsic mantle margin now represents the transition source variations of the plume component. region between an old Proterozoic cratonic
85 Project supported by EUROMARGINS concentrated, rather than mantle plumes. CRP-01-LEC-EMA13F. Every concession to source heterogeneity (enriched, volatile-charged, pyroxenitic, or eclogitic components in the source) Layered Mantle Alternative to Mantle reduces the likelihood that average Plumes: Evidence from the Pacific Plate primitive magmas have a high potential temperature that can only be explained by James H. Natland (Division of Marine a mantle plume. Geology and Geophysics, Rosenstiel School of Marine and Atmospheric The term “ocean-island basalt” (OIB) is a Science, Univ. of Miami, Miami, FL misnomer, since geochemically similar 33149,USA;Tel:+1-305-421-4123; lavas occur on many thousands of widely Fax:+1-305-421-4632; E-mail: dispersed seamounts and submarine [email protected]) volcanic ridges, and even in small percentages on the crests of spreading The features of central volcanic vents do ridges. Many seamounts and islands in the not readily allow perception of the Pacific could not have formed above arrangement of enriched mantle sources narrow vertical mantle conduit systems, beneath the active volcanoes of linear whether fixed or moving, and indeed the island chains. From experimental few strongly co-linear age-progressive petrology, all basaltic magmas derive from chains are only prominent during the last fairly shallow depths in the mantle at or third of the history of the Pacific plate. near the bases of lithospheric plates, and The majority of places have widely they carry with them the geochemical distributed volcanoes, cross-trend ridges, attributes of magma sources at those and/or repetitions of volcanism over wide depths and no deeper. Whether even regions that have spanned several tens of several central conduits in a 100-km million years; these favor persistent radius, such as at Hawaii, penetrate the top tapping of a widely-distributed enriched of a plume or the top of a widespread layer and shallow asthenospheric layer rather of enriched mantle is topologically than numerous narrow mantle plumes that impossible to determine. Inferred patterns originated deep in the mantle. The layer of “plume zoning” can also be interpreted appears to accumulate over time from in terms of a layered mantle, and in any aggregation of dispersed or strongly case are not consistent among volcanic concentrated enriched materials in the chains. The volcanoes themselves do not asthenosphere. The lithosphere acts as a reveal whether their sources have cycled permeability barrier to the ascent of down to the base of the mantle and then buoyant material derived from such upward again in a plume conduit through sources. Thus seamounts on older the entire mantle. Nor do the diverse lithosphere tend to be more consistently geochemical attributes of the basalts, and strongly enriched than those near the however they are distributed in time and ridge axis. The thickness and composition space, divulge deep mantle sources. of the layer likely derive from the Instead, most geochemists argue that their arrangement and concentration of low- enriched characteristics ultimately derive density materials implanted in the depleted from contact with the atmosphere, the upper mantle during ancient episodes of oceans, the hydrothermal processes that subduction. Xenoliths show that the drive differentiation in the ocean crust, lithosphere itself can become infused with weathered continental crust, and the enriched veins and cumulates charged with waters of rivers and streams. Thus the volatile-rich inclusions, which in turn can source of basalts from seamounts and be carried away for long distances, and islands may be a shallow layer beneath the even to subduction, by plate motion. lithosphere within which low-degree melts Shallow contractive and tectonic forces or fluids derived from buoyant subducted transmitted across the Pacific plate from materials or delaminated subcontinental subduction boundaries plus buoyancy lower crust and upper mantle have forces from the irregular enriched layer
86 determine where mid-plate fissure systems CaO,Sr.The high-Mg basalts in the PV develop and volcanism occurs. These (P1) and SV (S1) have MgO (7.5-10.5 have changed as the Pacific plate has wt%) at SiO2 (50-54 wt%), with grown and become more strongly directed TiO2(~0.5 wt%), Ni (~200 ppm) and Cr in its direction and rate of subduction over (>300 ppm), chondritic Al2O3/TiO2 (20- the past 160 million years. 25) and CaO/Al2O3 (~0.8), and nearly flat HREE (10xchondritic), and are of Munro- type komatiite affinity. The S1 rocks The Dongargarh Group: A Large occasionally show pyroxene-spinifex Igneous Province at the Archean- texture. The P1 and S1 have elevated FeOt Proterozoic Transition in India than the peridotite mantle (9.5 wt% vs.8 wt%).The S1, however, show depletion in S. Sensarma (Dept. of Geology, St. incompatible elements (e.g.,Rb,Sr,Ba,La) Anthony’s College, Shillong 793001, compared to P1. The tholeiites in the PV India;Tel:+91-364-2222558; Fax: (P2) and SV (S2), on the other hand, have +2537766; E-mail: MgO (5-6 wt%) at SiO2 (47-48 wt%),with [email protected]) TiO (~1 wt%), Ni (<100ppm) and Cr (100ppm), flat HREE (20xchondritic), The palaeoproterozoic (~2.5 Ga) sub-chondritic Al2O3/TiO2 (10-15), and Dongargarh Group in the Central Indian CaO/Al2O3 (<0.7). The P2 and S2 have Craton is a continuous ~10 km.thick FeOt (13-14 wt%). volcano-sedimentary sequence with large and sub-equal volumes of coeval mafic The olivine fractionation from a komatiitic and felsic lavas occurring in a bimodal parental melts,and interaction with bulk distribution for a length of ~250 km and Bijli melts gave rise to the high-Mg width of ~90 km. The rocks are weakly basalts, with decline in both olivine deformed into a regional syncline and fractionation and rhyolite contributions in metamorphosed at low-grade green schist the second pulse (S1). The P2 and S2 are facies.The felsic rocks (Bijli Rhyolite) possibly related to parental picro-tholeiites occur at the base of the succession.The by olivine and/or pyroxene fractionations. mafic volcanism is pulsatory,and mainly The komatiitic melts may have been represented by the Pitepani (PV) and derived from Fe-rich peridotite and/or Sitagota Volcanics (SV) successively harzburgite at 3-4 GPa.The picro- holeiites towards top. The minor andesitic were generated at 1-2 Gpa from a source Mangikhuta Volcanics (MV) denote the with variable but elevated FeOt/MgO than terminal volcanicity.The SV is thicker (~3 that of komatiite.The upper andesites are km) than PV (~1 km) and MV(<1 km).The related to the CFB-type tholeiites by 50% high-Mg basalts–Fe-rich tholeiites fractionation. The high temperature Bijli assemblage with common intrinsic melts (950°c) is hybrid having crustal as geochemical characteristics form the PV well as material contributions from high- and SV,and are repeated in the succession. Mg basalts.The pulsatory eruptions of Because of the great volumes of coeval high-Mg basalts and tholeiites in the belt mafic and felsic lavas,pulsatory eruptions could reasonably be explained in terms of of high-Mg basalts/basaltic komatiites and anomalously hot plume head melting in an CFB-type tholeiitic basalts with melts extensional environment. Direct volume in second pulse exceeding that of involvement of Fe-rich deeper mantle in the first,dominance of thick tholeiites over production of juvenile crust was important komatiites in the sequence, the at the A-P transition in Central India, Dongargarh Group may be characterized similar to many cratons in the erstwhile as a large igneous province,one of the few Gondwanaland. of its kind known at or near the Archean- Proterozoic (A-P) transition.
The rhyolites have high silica, high alkali, FeO/MgO, Ga/Al, REE, and low
87 Statistical Comparison of 3He/4He isotope data for MORB and OIB. MORB Distributions in Mid-Ocean Ridge and values do cluster around 8 RA (e.g., the Ocean Island Basalts median value for all ocean basins is 8.1 RA). A minority of moderately high F. M. Stuart (Isotope Geosciences Unit, 3He/4He (up to 14.2 RA) exist, usually in SUERC, East Kilbride, G75 0QF, UK; MORB from spreading ridges near ocean Tel: +01355-270133; Fax: +01355- islands or bathymetric highs. OIB 229898; E-mail: [email protected]); 3He/4He range from somewhat below the D Porcelli (Dept. of Earth Sciences, Univ. MORB range to 50 RA. An undeniable of Oxford, Oxford OX13PR, UK; Tel: feature of the OIB data is that while there +01865-282121; Fax: +01865-272072; E- is a concentration of values lower than mail: [email protected]); D. MORB, there is a preponderance of values Graham (College of Oceanic and well above the MORB range. This is true Atmospheric Sciences, Oregon State regardless of how the data are weighted, Univ., Corvallis, OR 97331-550, USA; whether equally, by hotspot size, or by Tel: +1-541-737- 4140; Fax: +1-541-737- region. In particular, both large and small 2064; E-mail: hotspots have high 3He/4He, making it [email protected]); D R unlikely that the MORB-OIB distinction Hilton (Geosciences Research Div., can be explained by differences in depth or Scripps Institution of Oceanography, La extent of melting. The most plausible Jolla, CA 92093-0244, USA; Tel: +1-858- explanation is that the mantle source of 822-0639; Fax:+1-858-822-3310; E-mail: OIB with high 3He/4He is distinct from [email protected]) the mantle source of MORB.
Helium isotope variations in mantle- Applicability of Large Magma derived materials provide a central Chambers to Deccan Volcanism: A constraint on models of mantle structure Numerical Study and evolution. Mid-ocean ridge basalts (MORB), generally interpreted to be Dhananjai Pandey (National Centre for derived from the upper mantle, are Antarctic and Ocean Research, Headland typically described as clustering around a Sada, Vasco-da-Gama, Goa - 403804, 3He/4He value of 8 RA, representing a INDIA; Tel:+91 832 2520863; Fax:+91 mixture of radiogenic He (<0.01 RA) and 832-2520877, E-mail: He initially trapped within the Earth (120 [email protected]); Anju Tiwary; S. RA). Ocean island basalts (OIB) from a Rajan variety of hotspot locations (in particular Hawaii, Iceland, Samoa and Galapagos) Deccan Volcanic Province (DVP) is one exhibit values ranging up to ~50 RA, and of the largest continental flood basalt this has been widely interpreted as (CFB) provinces. One of the most evidence for a mantle source having a important questions connected with the much higher 3He/(U+Th) ratio over Earth CFBs is to understand the enormous history. This in turn suggests isolation volume of volcanism. Explanations for from the main convective flow, likely such massive scale of magmatism over deeper within the Earth. However, several apparently short duration of time have recent studies claim that both OIB and been sought over the years through various MORB have similar average values for proposed processes. However, a lot more 3He/4He. This, in turn, could indicate information is required to understand the derivation from a common, heterogeneous dynamics involved in such complex reservoir, with OIB exhibiting a larger phenomenon. spread due to more limited averaging of small-scale mantle domains by the melt We suggest a mechanism through which generation process. large volumes of magma, sufficient to generate DVP, can be stored at shallow We have made a comprehensive crustal depths and can be erupted in compilation of all the published He relatively short duration of time. This
88 concept is based on the physical and last equilibrated either at ca.2.2 GPa with thermodynamic properties of accumulated an olivine clinopyroxene residue, or at ca. magma in an extensional environment 5 GPa with a garnet clinopyroxene leading to continental rifting. Our model is residue. The low bulk-rock Al2O3 content derived from the published numerical data, (9 wt %) and high [Gd/Yb]n ratio (3.1) are dealing with the possibilities of immense consistent with the presence of residual amount of magma generation and storage. garnet in the ferropicrite melt source and Further, the model incorporates the favour high-pressure melting of a garnet geological/geochemical characteristics of pyroxenite source. The garnet pyroxenite the DVP to assess the applicability of may represent subducted oceanic these numerical models on this CFB. lithosphere entrained by the upwelling Tristan starting mantle plume head. During adiabatic decompression, Experimental Constraints on the Role intersection of the garnet pyroxenite of Garnet Pyroxenite in the Genesis of solidus at ca. 5 GPa would occur at a High-Fe Mantle Plume Derived Melts mantle potential temperature of 1550 C and yield a ferropicrite primary magma. Eiichi Takahashi (Dept. of Earth and Subsequent melting of the surrounding Planetary Sciences, Tokyo Institute of peridotite at ca. 4.5 GPa may be restricted Technology, 2-12-1 Ookayama, by the thickness of the overlying sub- Meguroku, Tokyo 152-8551, Japan; Tel: continental lithosphere, such that dilution +81-3-5734-2338; Fax: +81-3-5734-3538; of the garnet pyroxenite melt component E-mail: [email protected]); James would be significantly less than in intra- Tuff (Dept. of Earth Sciences, Univ. of oceanic plate settings (where the Cambridge, Downing Street, Cambridge, lithosphere is thinner). This model may CB2 3EQ, UK; Tel: +44-1223-333400; explain the limited occurrence of Fax: +44-1223-333450; E-mail: ferropicrites at the base of continental [email protected]); Sally Gibson flood basalt sequences and their apparent (Dept. of Earth Sciences, Univ. of absence in ocean-island basalt successions. Cambridge, Downing Street, Cambridge, CB2 3EQ, UK; Tel: +44-1223-333400; Fax: +44-1223-333450, E-mail: [email protected])
The anhydrous phase relations of a primitive, ferropicrite lava from the base of the Early Cretaceous Parana´ Etendeka continental flood basalt province have been determined between 1 atm and 7 GPa. The sample has high contents of MgO (14.9 wt %), FeO* (14.9 wt %) and Ni (660 ppm). Olivine phenocrysts have maximum Fo contents of 85 and are in equilibrium with the bulk rock, assuming a K Ol_liquid (Fe/Mg) of 0.32. A comparison of our results with previous experimental studies of high-Mg rocks shows that the high FeO content of the ferropicrite causes an expansion of the liquidus crystallization field of garnet and clinopyroxene relative to olivine; orthopyroxene was not observed in any of our experiments. The high FeO content also decreases solidus temperatures. Phase relations indicate that the ferropicrite melt
89