Explanatory Notes by Philippe Bouysse

Geological Map of the World

• At the scale of 1: 50 000 000 (2009) Sheet 1: Physiography, volcanoes, astroblemes - 1st edition Sheet 2: Geology, structure - 3rd edition

• At the scale of 1: 25 000 000 (2010) Geology, structure - 3rd edition

COMMISSION DE LA CARTE GÉOLOGIQUE DU MONDE COMMISSION FOR THE GEOLOGICAL MAP OF THE WORLD «Ce qui est simple est toujours faux. Ce qui ne l’est pas est inutilisable». Paul Valéry (Mauvaises pensées et autres, 1942)

COMMISSION FOR THE GEOLOGICAL MAP OF THE WORLD

Geological Map of the World at 1: 50 000 000

( ird edition)

2009

EXPLANATORY NOTES

By Philippe BOUYSSE (CGMW)

Translation by Philippe Bouysse & Clara Cardenas (CGMW), reviewed by Peter Miles (CGMW) and Françoise Cadet (University Paris VI). SUMMARY

• Geological Map of the World at 1: 50 000 000 (2009) Foreword, p. 4 INTRODUCTION, p. 4 SHEET 1: PHYSIOGRAPHY, VOLCANOES, ASTROBLEMES I.1 - Physiography, p. 5 I.2 - Volcanoes, p. 5 I.3 - Astroblemes, p. 5 I.4 - Additional information, p. 5 SHEET 2: GEOLOGY, STRUCTURE II.1 - ONSHORE AREAS, p. 6 II.1.1 - Chronostratigraphic units, p. 6 II.1.2 - Ophiolites, p. 6 II.1.3 - Large Igneous Provinces : traps, p. 6 II.1.4 - Glaciers, inlandsis, p. 7 II.1.5 - Structural features, p. 7 II.1.6 - e Iceland case, p. 8 II.2 - OFFSFORE AREAS, p. 8 II.2.1 - , p. 8 II.2.1.1 - / Boundary (COB) , p. 8 II.2.1.2 - Microcontinent, p. 8 II.2.1.3 - Island arcs, p. 8 II.2.1.4 - , p. 8 II.2.1.5 - Continental slope, p. 9 II.2.1.6 - Antarctic margin, p. 9 II.2.1.7 - Ice shelf, p. 9 II.2.2 - OCEANIC BASINS, p. 9 II.2.2.1 - Age of the oceanic crust, p. 9 II.2.2.2 - Abyssal plains, p. 9 II.2.2.3 - Mid-oceanic ridges, p. 10 II.2.2.4 - Axis of mid-oceanic ridges, p. 10 II.2.2.5 - Transform faults and fracture zones, p. 10 II.2.2.6 - Subduction zones, subduction trenches and other trenches, p. 11 II.2.2.7 - "Anomalous" submarine features (, oceanic plateaus, tracks) , p. 12 II.2.2.8 - Distributed or diuse plate boundaries, p. 13 II.2.2.9 - Submarine volcanism and the opening of the North , p. 13 II.2.2.10 - SDR's related to the opening of the South Atlantic Ocean, p. 13 II.3 - HOTSPOTS, p. 13 By way of conclusion…, p. 14

• Geological Map of the World at 1: 25 000 000 (2010) SINGLE SHEET: GEOLOGY, STRUCTURE (in 3 parts)

ADDITIONAL NOTE, p. 16 The Geological Map of the World at 1: 50 000 000 (2009)

Explanatory Notes by Philippe BOUYSSE (CGMW)

Foreword ese Notes presented in a somewhat heterogeneous manner, combine regular peer-reviewed information dedicated to geoscience professionals – normal users of geological maps – with more basic information intended for the wider public including high school and college students who constituted a large section of the users of the former editions of this map. It was not possible to address in these notes all the geologic, structural or geodynamic aspects that may be raised by the careful examination of the Map. e text, consisting mainly of comments on the legends, is aimed at shedding some light on a selection of examples that are, in our view, illustrative of each element of the Map. It should be noted that, in this new edition, a particular attention was given to the oceanic areas, the large magmatic events, and to the geodynamics.

INTRODUCTION

is third edition of the Geological Map of the World at the and 72°S for Sheet 2 (instead of 78°N and 65°S for the former scale of 1:50,000,000 (1:50 M) follows the rst and second two editions), and at 72°N and 70°S for Sheet 1. As a conse- editions published by the CGMW respectively in 1990 and quence, a large extent of the Antarctic continental coastline 2000. is bilingual document (English-French) is the is visible, with a better delimitation of the southern result of a highly synthetic compilation given both the small ocean. As for Greenland, only its southern half is visible. scale of the map and its educational purpose. It is a tenta- On the other hand, the Taymir peninsula has been tive and (very) simplied representation of the entire solid severed from the far north of the Eurasian continent. surface of our planet and includes both continental and e circum-polar projections extend to the 60°N and oceanic domains. 60°S parallels (instead of 70°N and 60°S for the previous editions), Greenland is now displayed in its entirety and is new edition is a completely revised concept compared the 2 circum-polar areas have the same surface area. to the map issued in 2000 and takes into account the state of eir scale was slightly enlarged to 1:46 M. the geologic knowledge at the turn of the century. For the rst time, the Map is designed in two sheets of the same For practical reasons and marketing policy, this 3rd size: edition at the scale of 1:50 M (for the Mercator projec- tion) precedes the publication at 1:25 M (original scale of – Sheet 1 (Physiography, Volcanoes, Astroblemes) revea- the dra). An interactive digital version of the Map is ling the ne-grained texture of the totality of the scheduled for the end of 2010. surface when removing the water of the . In the previous editions at 1:25 M scale, the Mercator – Sheet 2 (Geology, Structure) showing the distribution projection was printed in two parts (20°W-170°W; of the main chronostratigraphic units and the main 170°W-20°W) that allowed adjusting the center of the structural features that make up the mosaic of the Map either on the Atlantic (opening of an ocean and t of present-day surface of our planet, the result of 4.56 the conjugated ), or on the Pacic billion years of unremitting "resurfacing". Sheet 2 is the (subductions and hotspots tracks). In this new edition, equivalent of the single sheet of the second edition, the single 1:50 M sheet forced us to make a choice for the notably reworked and extended. centering. In order to overcome these inconveniences, we decided to center Sheet 1 (Physiography) on the Each sheet consists of a main map in Mercator projection, Paci c (meridian 0° for E and W sides) and Sheet 2 with the 2 polar areas in polar stereographic projections. (Geology) on the Atlantic (180° meridian for both sides). e dras have been carried out at the 1:25,000,000 scale is enables the reader to visualize both options of (1:25 M). assemblage. e Mercator projection has only a true scale representa- Scales and projections being identical, it is easy to super- tion along the equator but allows an optimal visualization pose the morphological features of the oshore areas that does not favour the continents at the expenses of the (Sheet 1) with the geological structures mapped on Sheet oceans or vice-versa, unlike many other projections used 2 using an illuminated table. for world maps. e main drawback of Mercator comes from the deformation that increases with the latitude to Nota: In the text that follows, words typed in bold become innite at the poles. For this reason, in this edition, characters correspond to the dierent items of the the "upper" and "lower" latitude limits have been set at 72°N legends.

1 e abbreviation for billion years (10 years) is Ga (from giga-annum, ocial designation of international geological bodies). e author wonders why the accusative form "annum" was chosen instead of the nominative one "annus".

4 SHEET - 1: PHYSIOGRAPHY, VOLCANOES, ASTROBLEMES

I.1- PHYSIOGRAPHY Data acquisition dates back to April 2006. e sources are: Published for the rst time by the CGMW, this Map 1/ Planetary And Space Science Centre of New Brunswick displays all of the Earth’s morphology and, in particular, the University (John Spray & Jason Hines, web site: lesser known domains of the submarine areas that represent www.unb.ca/passc/Impact.Database) with 174 structures; nearly 71% of its surface. Colour palettes are used to repre- 2/ Jarmo Moilanen, Finland: Impact structures of the sent the land topography and, ocean bathymetry, the latter World with 21 structures; including ne black lines to indicate depth contours (site : www.somerikko.net/old/geo/imp/impacts.htm); (isobaths) at every 1000 m. In order to avoid blurring of the physiographic perception of mountain chains, the equiva- 3/ NASA/Goddard Space Flight Center Scienti c Visualiza- lent for the subaerial areas (isohypses) were not plotted, tion Studio for the Araona/Iturralde crater (Bolivia), not except for the Greenland and Artarctica ice caps. validated yet; e topography was generated from a digital database of 4/ Wade S. et al., in Lunar and PlanetaryScience, 2002, land and - oor elevation (EOTPO2) on a 2-minute XXXIII, for the Velingara crater (Senegal); latitute/longitude grid resolution. Sea oor data are from 5/ Paillou Ph. et al., in C. R. Géoscience, 2004, v. 336, for the work of W. Smith & D. Sandwell (1997). ese data were the Gilf Kebir structure (Égypte). derived from satellite altimetry observations combined with carefully quality-assured shipboard echo-sounding Even though it is not an impact crater stricto sensu, the measurements. Land data were primarily from 30-second location of the Tunguska (Central ) airblast of an gridded data collected from various sources by the National asteroid (a comet?) in 1908 was identied with a red circle. Imagery and Mapping Agency (USA). e compilation of altimetry data was carried out by our I.4- ADDITIONAL INFORMATION late colleague Jacques Ségoun, who passed away on 8 is Sheet also includes information concerning the eleva- September 2008 before the release of this Map. tion of some specic locations: I.2- VOLCANOES e highest elevations (in meters) are indicated for each continent. It is noteworthy that in the particular case of On this Sheet are also plotted 1506 active or recent volca- Mount Elburs (at 5642 m, the highest peak in the noes or volcanic elds that, a priori, cannot be considered range), it is oen that taken as the geographic boundary as denitively extinct (i.e. having erupted during the last between and . Also the Puncak Jaya (4848 m) in 10,000 years, corresponding to the Holocene epoch). ese New Guinea, and geologically part of the the Australian volcanic systems exist in 1436 subaerial edi ces (red trian- continental ensemble, surpasses the highest point of Austra- gles), and 70 submarine volcanoes (blue triangles) as, for lia, Mount Kosciusko. example, the Graham (called also Julia or Ferdi- nandea) located near the southern coast of Sicily. All trian- A selection of the lowest points onshore, but below sea level, gles are bordered by a ne white line that allows dierentia- are shown. ese include the surface of the (at tion between each volcano in very active , such as in -412 m, presently in the process of drying out if no drastic the island arcs (e.g. Sunda Islands) or in arc cordilleras (e.g. measures are taken) a salt lake whose bottom is at 742 m ). Consequently volcanic gaps are visible in some under the mean ocean level!. arcs such as that stretching all along 1500 km from the Two other lakes with remarkable characteristics that set south of Ecuador to the south of Peru. world records are: ese volcanoes were extracted from the Global Volcanism - Titicaca, on the Andean Altiplano is the highest navigable Program of the Smithsonian Institution catalogue (as at lake (+3810 m; maximum depth of 284 m). March 2006) in their web site : www.si.edu/world/gvp/ . - Baikal, in Siberia is the deepest lake in the world (1642 m e ssure volcanism that characterizes the active mid- beneath its surface which is at +456 m of altitude). It also oceanic ridges (where the divergence of the lithospheric contains the largest deposit of fresh liquid water on the plates takes place) is not included here. It is however repre- planet’s surface (23 000 km). sented by the axis of active accretionary ridge (red lines) e lowest point in the ocean is located in the south of the drawn in Sheet 2 of the Map. Mariana subduction trench (–10 920 m, in the Challenger Deep). I.3- ASTROBLEMES Finally, the highest mountain on Earth is not Everest Except for one site adjacent to Chesapeake , 198 (+8848 m), but the volcanic island of Hawaii (Big Island) onshore astroblemes, or meteoritic impact craters, are with a total height of 10 239 m, if the elevation of its highest plotted on the Map. ey are divided into 2 categories of peak, the Mauna Kea volcano (+4206 m above sea level) is crater diameter: <10 km and ≥10 km, shown as small and added to the maximum depth of its submarine bottom large black asterisks respectively. (–6033 m).

2 A former researcher at the Institut de Physique du Globe of Paris, he was Secretary General of the CGMW Sub-commission for Sea oor Maps from 1983 to 2004.

5 SHEET - 2: GEOLOGY, STRUCTURE

II.1- ONSHORE AREAS tions (bright green hue) is relatively small and quite oen hardly visible. e ophiolites plotted on this map are restric- II.1.1- Chronostratigraphic units ted to the Meso-Cenozoic times (younger than 250 million e onshore areas represent 29,2% of our planet’s surface years). Particularly noticeable are the ophiolites of the and correspond mainly to the rock formations of continen- Alpine arc, the Dinarides/Hellenides, the Zagros (Iran) and tal origin (or continentalized in the case of island arcs). the . ey are classied using 8 broad chronostratigraphic As an example of an island of ophiolitic origin, it is worth units : 1= Cenozoic; 2= Mesozoic; 3= Upper Paleozoic; 4= mentioning the tiny Gorgona island located on the conti- Lower Paleozoic; 5= Neoproterozoic; 6= Mesoproterozoic; 7= nental Pacic margin of Colombia. Also Macquarie island Paleoproterozoic; 8= Archean. A number of regroupings (some thousand kilometers to the SSW of New Zealand) is were made when necessary by the geological or cartogra- the result of a tranpressive motion along the large dextral phic contexts. In comparison with the previous edition, transform fault (see note 16) that separates the and for the sake of coherence, the Quaternary and the (Indian-Australian plate) from the Pacic Ocean (Pacic Triassic Periods within the Cenozoic and Mesozoic eras plate) and uplied a slice of Cenozoic oceanic crust. Also we respectively have not been shown individually. Also the 3 note Zabargad island (formerly called St. John island) in the eras of the Proterozoic Eon have been introduced as units 5, (Egypt), known since Antiquity (Egyptians, Greeks 6 and 7. and Romans) for its peridotite intrusion containing beauti- Within these time units 3 main lithological facies ensem- ful olivines (marked by a green asterisk). bles were distinguished: • sedimentary formations or those of an undierentiated nature (uneasy to dene); • extrusive II.1.3- Large igneous provinces: the traps volcanic formations (V), corresponding to subaerial During some periods in the history of our planet large magmatism; • endogenous formations (P), representing eruptive pulses of a relatively short duration (in some cases rocks originating in the Earth’s interior at depth and having less than 1 million years) occurred in the Earth at mantle undergone signicant metamorphism or that correspond depth. ese magmatic “crises” led to the vast and volumi- to plutonic magmatic rocks. e last two rock categories are nous outpouring of basalts at the surface of the continents illustrated by a scattering of superimposed dots (blue for (traps) as well as on the ocean oor (oceanic plateaus). extrusive, red for the endogenous). ese huge lava ows are interpreted as the consequence of One exception was made for the Cenozoic volcanism (V1) the ascent of a large up to the base of the that is identied by a uniform strong blue hue. Actually, the lithosphere to produce the head of a strong “hotspot”, during volcanism of this era (which includes Quaternary and the rst phases of its life (cf. II.2.2.7 et II.3). ese surface Present times) is, in many cases (e.g. subduction volca- features are labelled “Large Igneous Provinces” nism), the consequence of on-going geodynamic activity. It (abbreviation LIP). e lavas of the traps, very uid, are also is therefore important that this volcanism be seen in termed “ood basalts”. relation to the “active” volcanoes of Sheet 1 . In the former editions of this map the traps were merged Another exception was also made for the oldest formations, into the too large time slices used to corresponding to the the Archean (“8”, older than 2,5 billion years/Ga), as here chronostratigraphic units of the legend (e.g. Upper Paleozoic they are not dierentiated for the sake of simplication. It for the Siberian traps, or the Mezosoic for the Deccan ones in should be noted that the largest Archean outcrops are India). On the other hand, a number of traps straddle the located in Canada. large main stratigraphic boundaries of these units, e.g. Upper Paleozoic/Mesozoic (250 Ma) in Siberia; II.1.2- Ophiolites Mesozoic/Cenozoic boundary (65.5 Ma, also called K/T e ophiolites are remnants of oceanic lithosphere (in boundary) for the Deccan event. is might not be coinci- increasing depth: submarine basalts, gabbros, peridotites) dental since, for a number of geologists (e.g. Courtillot and which, in a nal phase of subduction following the collision co-workers), the great mass extinctions that aected a of two continental blocks (or continentalized in the case of number of living species might be due to massive gas and island arcs), escaped from their usual recycling within the noxious aerosols produced by these cataclysmic eruptions. Earth’s mantle to become exposed inside mountain chains. is hypothesis is however in competition with (but also ey are the evidence of a “lost ocean” (Jean Aubouin) and later associated to) the big meteoritic impact thesis, exempli- punctuate large suture zones. ed by the Chicxulub crater in the north of Yucatan in Mexico, for the K/T limit (see Sheet 1). ey can also be the product of an obduction, as in Oman, where a slice of oceanic lithosphere overthrusts the edge of In order to deal with these issues, we chose for this new a continental basement. edition to assign the same color (bright red-orange) to all the traps, with an indication in black of their average age in At the scale of the Map the extent of the ophiolitic forma- Ma (e.g. “16 Ma” for the Columbia River/Snake River traps in

3 i.e. geologic time slices. In the corresponding legend’s table as well as in the oceanic crust ages (cf. II.2.2.1), the dates indicated are those validated by the International Commission on Stratigraphy in the Geologic Time Scale 2008 published in 2008. e margin of error (2σ) was not mentioned for the sake of simplication. . 4 e abbreviation for million years (10 years) is Ma from the “mega-annum”, see note 1. 5 is term and its abbreviation LIP are currently used in the international geoscience community and were coined in 1994 by Millard Con and Olav Eldholm. Rev. Geophysics, 32 :1-36. 6 "K/T" for Cretaceous/Tertiary. e use of the term « Tertiary » that corresponded to the Cenozoic without the Quaternary, should be avoided from now on. 7 e large chronostratrigraphic delimitations (eras, periods, epochs) were created in the XIX century aer the observation of sudden, very important and generalized changes, in the association of fossiles and micro-fossiles contained in the sedimentary deposits, mainly marine facies.

6 north-western USA). It is to be noted that the Parana traps II.1.4- Glaciers, inlandsis in southern Brazil have the same age (133 Ma, earliest part Glaciers of some importance were mapped in the far south of Cretaceous) as the less extensive Etendeka traps in Nami- of the Andes, along with those covering islands of the far bia. Initially, these two features formed a single entity, but North Canada and . ey were assigned the same are now separated by several thousand of kilometres of color as the Greenland and Antarctica ice caps (light grey). ocean oor. ey were originally produced by the Tristan For the latter inlandsis, the zero meter level contour (sea da Cunha “hotspot” (identied as HG in the inset at the level) was drawn. e areas oulined by these contours repre- bottom of Sheet 2) and separated during the opening of sent the subglacial bedrock lowered by the ice loading were South Atlantic which started shortly aer, during the Early distinguished from the ice caps using a darker hue (light Cretaceous. Not too far from the Etendeka traps exists purple). another slightly older ensemble of traps (183 Ma, Early Jurassic), the Karoo, that outcrop in southern and II.1.5- Structural features were subsequently dismantled by erosion. A third large “LIP” in Africa are the Ethiopian traps (30 Ma, Oligocene) With the exception of Iceland (cf. II.1.6), the Afar (II.2.2.4) including also those of SW Yemen that are only separated and the Makran (II.2.2.6), the onshore areas show only two by the narrow entrance of the Red Sea (Bab el Mandeb structural features: the large normal faults and those of straits). Almost coeval with the Karoo traps, the remnants undertermined nature (black line); the large thrust fronts of the Ferrar traps (175 Ma) are associated with the sills of (jagged black line) curving round the large orogenic belts; same age (marked on the Antarctic Polar projection by a Alpine (Alpes-Carpathian Mountains, Caucasus, Himalayas, red asterisk of the same color as the traps). ese are scatte- Maghrebides, Rocky Mountains, Andes) or the older Hercy- red along the large Transantarctic Mountains range. e nian (Variscan, Urals, etc.), Caledonian (Appalachian, temporal and geographic proximity of these two ensembles, northern British Isles, western , ...) and even the when part of the Gondwana supercontinent, might indicate roots of Precambrian belts (Canadian shield, etc.). that they were generated by the same hotspot. Two small Among the many large structural lineaments on the map, it traps located to the NE of the Deccan traps do not belong to is worth noting the following: the latter; in the NE corner of the Indian shield is the – A line extending from the south of Norway to the Black Rajmahal (118 Ma, Early Cretaceous) and slightly to the Sea (Tornquist-Teisseyre line) that separates the “Precam- east, Sylhet (116 Ma), near the Assam/Bangladesh boun- brian Eo-Europe”, including the Baltic shield (more appro- dary. e source of these two traps is thought to be the priately called Fenno-Scandian shield), and the Archean and (HI). In later time this may have also generated the Ninetyeast Ridge (or 90° E Ridge, cf. II.2.2.7). Proterozoic outcrops in Ukraine, from the pattern seen in e Emeishan traps formed in China towards 260 Ma more recent European structures (Paleo-, Meso-, Neo- (Paleozoic, at the limit Middle Permian/Late Permian). Europe). e huge Siberia traps mentioned above presently outcrop – e continental ri system emplaced since the Oligocene over the majority of the eastern part of the Siberian craton. which stretches across from the northern Some remnants are found further to the north in the to the Gulf of Lion via the Rhine valley and the southern part of the Taymir peninsula (only visible in the Rhodanian corridor. It is punctuated locally by volcanic Artic map in polar projection). Originally, these traps complexes (i.e. Vogelsberg and Eifel in Rhineland-Hesse, covered a much larger area (some authors give an estimate Cantal and Chaîne des Puys in Auvergne). of about 4 million km, or even more). e red dashed- – e large Amazon graben which isolates the two Guianan dotted line drawn on the West Siberian plain corresponds shields from the Brasilian cratons (Central Amazonas and to a minimal estimate of their western extension beneath São-Francisco) to the south. the Meso-Cenozoic sedimentary deposits (Reichow et al., – A large and old SW-NE fracture cutting the Africa in two 2002). from the across to the middle part of the Red Finally, a large red dashed line gures the boundary (that Sea. one can follow from the east of and the NE – e great East-African ri valley system, emplaced during of to the west of Africa and Europe, drawn the Cenozoic, and its relationship to the Afar hotspot (H1) aer J.G. McHone, 2003) of a sole large magmatic province. and the opening of the and the Red Sea. e is boundary outlines the traps of the CAMP (Central ris are oen occupied by great lakes. From north to south Atlantic Magmatic Province) generated by a hotspot 200 are: Turkana, Albert, Edward, Kivu, Tanganyika and Malawi million years ago (limit Triassic/Jurassic) shortly before the lakes oen punctuated by important volcanism. Should this opening of the Central Atlantic dislocated this ensemble. continental ri and spreading persist, the East-African Ri Although the erosion caused the disappearance of piling-up will progressively become an oceanic lineament similar to of lava ows, the CAMP was reconstructed thanks to the the Red Sea and eventually be of the form of the Gulf of occurence of related sills and dykes (volcanic intrusive Aden and separate the “Somalia” plate from the rest of bodies), that underlaid the surface outpourings. Africa, named “” plate by some geologists. A last point to explain concerning the continental LIP: the – e large faults that extend from the Pamir in a fan-like Seychelles Islands are made of Neoproterozoic (P5) grani- pattern between China and SE Asia. ese wrench faults tes marked by an arrow because these islands are hardly worked in response to the continuous push that the Indian distinguishable on the Map. ese granites are intruded by sub-continent has been exerting against the east of the 65 Ma old dykes (gured also by an arrow and a red Eurasian continent for some 50 million years. Faults such as asterisk). is is the evidence that the Seychelles micro- Altyn-Tagh (SW-NE) and Kunlun (W-E) carved out great continent was part of India, or very close to it, during the basins such as Tarim (in the Xin Jian or Chinese Turkestan). times of the Deccan traps eruption. – Again in Africa, it is worthy noting the existence of the “Great Zimbabwe Dyke”, a narrow strip of intrusive Paleo- proterozoic, stretching N-S for 550 km, whose width does not exceed a dozen kilometers.

7 II.1.6- e Iceland case case: (1) for the Seychelles platform (granites of 750 Ma) in e entirely volcanic island of Iceland covers a signicant the Indian Ocean; (2) of the Jan Mayen microcontinent in area (103 000 km) and has an exclusively oceanic origin. It the far North Atlantic; (3) of the Bollons (60° S, was built on a substratum of oceanic crust modied by a 177° W) close to the New Zealand continental margin in the powerful hotspot (marked HD on Sheet 2) and is linked to Pacic; (4) the South Orkneys microcontinent detached the opening of the North Atlantic (north of 60°N). e axis from the tip of the Antarctic Peninsula among others. of the Mid-Atlantic (spreading) Ridge runs across the On the contrary, in this edition the Agulhas Bank (25°E, island to separate two distinct geodynamic domains; the 40°S) to the south of South Africa, has no longer been to the east, and the to assigned a continental nature. is is on the basis of recent the west. Instead of mapping this island in the same way as works that suggest a volcanic origin of this quite large the rest of the onshore areas (i.e. in “V1”), as in the former morphostructure built up on oceanic crust, as with the other editions, it was decided to represent it as a surface of large submarine reliefs of the SW Indian Ocean. oceanic crust where Plio-Quaternary and Miocene basalts are distinguished from each side of the spreading axis. II.2.1.3- Island arcs e island arcs follow the same mapping principle used for II.2- OFFSHORE AREAS the continents and are bounded by the same medium blue e world ocean represents more than two thirds of our line. It is known that they are the product of magmatic planet’s surface (70,8%). It covers, on one hand, the processes peculiar to the subduction events that lead to the submerged edges of the continents, the continental margins, formation of a “continentalized” crust (becoming thicker and also the deep seaoor whose substratum consists of and lighter than the oceanic crust). It is probable that in a oceanic crust “produced” at the axes of the spreading ridges, number of cases, such as in the Japanese archipelago, their also called “Mid-Oceanic Ridges”. e average depth of the substratum was detached from the nearby continent. is ocean is 3 680 m, a value much higher than the 840 m occurs through a general characteristic of the subduction average elevation of the continents. e drawing of the mechanism known as “slab roll-back” that initiates the oshore part of Sheet 2 was constructed, for some elements opening of a back-arc basin (or marginal basin; cf. II.2.2.4 et (spreading axes, transform faults/fracture zones, subduc- II.2.2.6). tion zone axes, oceanic plateaus, hotspot tracks and other anomalous reliefs), by superposing the tracing dra of this II.2.1.4- Continental shelf sheet over the “Physiography” sheet. e continental shelves (or “continental platform”, or “conti- nental terrace”) represent the innermost part of the conti- II.2.1- CONTINENTAL MARGIN nental margins. ey extend from the coastline to the shelf break which tops the continental slope. e external limit of II.2.1.1- Continent/Ocean Boundary (COB) this shelf has an average depth of –132 m. For practical e boundary between the continental crust and the reasons, and given the scale of the Map, the commonly oceanic crust (COB) is shown by a blue line. is outlines assigned –200 m isobath is used here to delineate the conti- the passive continental margins generated by the riing of nental shelf since this depth is generally close to the shelf two separating continental blocks to form an ocean. break. On this Sheet, and from a mapping point of view, the Actually, this boundary is not that precise and one should continental shelf was considered only from a morphologic include a transitional zone (OCT) between a well identied point of view ( a terrace) and conceals all other cartographic continental crust and a “normal” oceanic crust characteri- units it might overlay. us, the “continental” shelf of the zed by well identied magnetic anomalies. e transition Niger delta obliterates the oceanic nature of the underlying zone oen displays a stretched and thinned continental oceanic crust upon which the sedimentary fan of this large crust intruded by peridotites that rise from the underlying African river is prograding (i.e. it builds up seawards). e mantle (exhumation). same applies to the “continental” shelf of Iceland, actually an Along the active continental margins, characterized by a island entirely generated by oceanic volcanism (cf. II.1.6). subduction zone, the COB is well dened (corresponding All the continental (and island arc) shelf areas represents to the subduction trench axis) and the above mentioned about 7.5% of the oceans surface. On the Map, the continen- COB blue line is completely overwritten in this cartography tal shelf is characterized by a light beige color. For practical by the specic line depicting the subduction (cf. II.2.2.6). reasons this cartographic unit also encompasses the shelves or terraces of atolls and volcanic islands that are not “geneti- Considering the legal (and therefore political and econo- mic) implications arising from the delimitation of the COB cally” continental but of oceanic origin (e.g. the Tuamotu in the frame of the United Nations Convention on the Law archipelago). Indeed, the term “continental shelf” has a of the Sea (UNCLOS), it is expressly stated that the drawing broader meaning in the formulation of the Law of the Sea of the COB limit on this Map is only approximative and (UNCLOS). sometimes conjectural, and that it does not have any legal e continental platform is very narrow along many sectors status and neither is any implied. of the African coast (only a few kilometers o Mogadishu, Somalia) and along the Brazilian margin south of the II.2.1.2- Microcontinents equator. On the island arcs, it is not well developed either. Some tiny « ras » of continental crust (therefore encircled On the contrary, it is very wide o the coast of SE Asia (East by a specic blue line) are shown on this Sheet isolated China Sea, Sunda shelf), o Argentine (up to 600 within an oceanic basin. ey are named microcontinents km wide) and a maximum extension can be observed along and result from the complex history of the break-up and the front of Northern Eurasia (up to some 900 km on seaoor spreading in the formation of an ocean. is is the the continental shelf of Eastern Siberia).

8 e mapping of the continental shelf is one of the innova- that of the continents. Oceanic basins cover about 59% of tions of this third edition of the Map. It is an important the planet surface. Five main types of morphostructures are element when considering the Quaternary palaeogeogra- to be distinguished: • abyssal plains; • mid-oceanic ridges; • phy of the world. It allows us to consider the withdrawal of large fracture zones; • subduction trenches; • “anomalous” sea level that occurred during the great Würm regression oceanic features, i.e. structure of volcanic origin whose (ca. 20,000 years ago), the Last Glacial Maximum during genesis postdates the age of the oceanic crust on which they which the sea level dropped by about 130 m. During this have been built up. event, the volume of water removed from the oceans was transferred to build up the huge glacial ice caps in northern II.2.2.1- Age of the oceanic crust North America (up to 4 km thick above the ) In comparison with the age of the continents, whose oldest and NW Eurasia. At that time, the and its outcropping nuclei have been dated at some 4 Ga (billion Western Approaches were completely emerged. It was also years), the age of the oceanic basins substratum never possible to travel overland from the the far north of the Gulf exceeds 200 Ma (million years). In the present state of of Siam to Bali, and from New Guinea to . knowledge the oldest ages are Middle Jurassic (starting at 175.6 Ma). ese are found o the eastern margin of the II.2.1.5- Continental slope United States and o its conjugated margin of Western e part of the continental margin located seaward of the Africa (both margins tted into each other before the shelf break and extending down to the contact with the opening of the Central Atlantic). ey also exist in the oceanic crust (i.e. COB) is called continental slope. is Central part of Western Pacic. As Earth volume is term applies also to the island arc margins, as explained constant, every piece of oceanic crust formed at the axes of above. is element of the oshore morphology is repre- mid-oceanic ridges prior to this limit of 200 Ma was necessa- sented in a yellowish green. e continental slope can be rily entirely “swallowed” by the subduction process, trapped quite extensive, such as o southern South America where as slivers within continental collision or overthrust during the spur bearing the Falkland Islands projects itself to the an obduction (cf. II.1.2). east towards the South Sandwich island arc over more than e mapping of the age of the oceanic crust was made by 1 500 km. interpolation of the position of the magnetic anomalies II.2.1.6- e Antarctic margin generated by the eect of periodic inversion of the Earth magnetic eld on newly formed crust (cf. Müller et al., e continental margin of Antarctica presents specic 1997). In this we have displayed only the limits of the morphological characteristics owing to the isostatic loading chronostratigraphic units shown: Plio-Quaternary – excerted on the continental lithosphere for some 30 Ma by Miocene – Oligocene – Eocene – Paleocene – Upper Creta- its huge ice cap. e most salient characteristics are: the ceous – Lower Cretaceous – Upper Jurassic – Middle Jurassic frequent presence of a nearshore depression (down to 1 000 (cf. the relevant legend). e colours for the dierent m deep) and a continental terrace abnormally lowered oceanic units are those currently used for CGMW seaoor (from –400 to –700 ) in front of the shelf break. erefore maps. the continental shelf and continental slope have been merged into a single map unit shown in a light yellowish For the enclosed basins such as the Arctic basins, the green to dierentiate it from that of the continental slope. remnants of the ancient Tethys Ocean (Eastern Mediterra- nean) and back-arcs basins, where the age of the crust is II.2.1.7- Ice-shelf sometimes not precisely known, a larger age range was used For glaciologists, an ice-shelf is a thick volume of ice (e.g. Undierentiated Jurassic- Cretaceous for the Eastern creeping (“owing” slowly) from the ice cap to beyond the Mediterranean, or Neogene for the marginal basin located to coast and has the form of a glacial sheet oating above a the south of the , Indonesia). Moreover in some continental terrace. Its thickness varies from 100 to 1000 m. sectors, where the colors might not be clearly discernible, the ese platforms are characterized on the Map by a blueish age is also given by the corresponding symbol in the legend grey color. e ice-shelves of Greenland and Canadian (e.g. “j3” for the South Caspian basin, or “g” for the Celebes Arctic Islands, too small at the scale of the Map, were not basin). plotted. e Antarctic ice-shelves have a total surface of Finally, shown in grey are a number of oceanic areas where about 1.5 million km and could be highly aected by the the magnetic anomalies have so far not been identied by ongoing climate change. e largest ice-shelves are the geophysicists and where the age of the crust remains unde- Ronne to the “north”, and the Ross to the “south”, the latter termined. ey are to be found mainly around Antarctica, partially encircling Ross Island, location of the active and to the east and SE of Australia. volcano Erebus (cf. Sheet 1). Ice-shelves should not be mistaken for pack ice, the latter II.2.2.2- Abyssal plains being a thin ice sheet (few meters thickness) of frozen e abyssal plains are characterized by a very at sea bed seawater. eir size signicantly changes during the with a, sometimes quite thick, sedimentary cover that seasons of the year. exends to both sides of the mid-oceanic ridges. eir depth (blue hues on the physiography of Sheet 1) increases imper- II.2.2- OCEANIC BASINS ceptibly from some 4000 m to a little over 6 000 m. Schema- Oceanic basins are that part of the seaoor whose basaltic tically, the age, the density and the depth of the basaltic substratum is made up of oceanic crust. ey are overlain substratum increase with the distance from the axis of the by sediments, except in the axial zones of the mid-oceanic mid-oceanic ridge. Likewise, the thickness of the sedimen- ridges. eir history and structure dier drastically from tary cover increases with the distance from the mid-oceanic

8 From the end of Fi ies onward, the Australian geologist Samuel Warren Carey proposed the theory of the expanding Earth where the surface of our planet must have been increasing for the last 200 Ma which is correlative to the break-up of the supercontinent Pangea and to the continental dri . Consequently, he dismissed the existence of subduction zones. is theory was (almost) denitively abandonned.

9 ridge. A good example of a well individualized abyssal this represents a "triple junction" where the Gulf of Aden (a plain, free of anomalous reliefs, is the Argentine basin – continuation of the active Carlsberg ridge in the northern whose centre deepens to more than 6 000 m depth – half of Indian Ocean), the Red Sea oceanic ri and the Great surrounded by the South Atlantic mid-oceanic ridge, the East African Ri converge. Although the Afar is still largely Falkland spur and the Argentine continental margin. of continental nature, three small segments of oceanic accre- tionary axes are plotted somewhat schematically to gure II.2.2.3- Mid-oceanic ridges the (possible) beginnings of a future oceanization (if the e mid-oceanic ridges (or oceanic accretionary ridges) form present geodynamic context remains unmodied, cf. II.1.5). the largest mountain range in the world with a total length of nearly 80 000 km that extends through the four oceans. Concerning the back-arc basins (or "marginal basins) that Starting at the base of the continental margin of the open "behind" an island arc (i.e. on the opposite side to the river delta (Eastern Siberia) in the Arctic, this system runs subduction trench), a micro-ocean forms and therefore the through the Atlantic from north to south, enters the Indian oceanic accretionary axis is represented by the same red line Ocean (with a northern branch running up to the Red as for the oceans. is can be seen in the marginal basin of Sea, generates a "triple junction" of ridges to the SE of the Mariana island arc (Western Pacic), the Lau Basin that Reunion island), then rounds the southern tip of New opens behind the Tonga island arc (SW Pacic) and the Zealand continental margin to step into the Pacic. In this South Sandwich back-arc basin (formerly named Southern latter ocean, the oceanic ridge is not in a mid-oceanic Lesser Antilles), a part of the loop linking southernmost position but is largely oset to the east (justifying thereby Andes to the Antarctic Peninsula. An incipient stage of its East-Paci c Ridge/Rise or EPR label) before "dying" in back-arc spreading is occurring in the Okinawa Basin to the the (or Cortes Sea). From this long ridge, west of the Ryukyu island arc (southernmost Japan archipe- originate two branches extending to South America: the lago), with a series of small active en echelon segments that South Chile Ridge and the Galapagos Ridge. Farther to the begin to cut out the continental margin of the East China north, the Juan de Fuca Ridge is located at some distance Sea. A more advanced stage is found in the (very narrow) from the coast extending from the north of California to Branseld marginal basin located at the rear of the South British Columbia. is small oceanic ridge is linked to the Shetland subduction zone within the Antarctic Peninsula. Gulf of California along the San Andreas transform fault e extinct axes of oceanic accretion are gured like the system (cf. II.2.2.5). e Juan de Fuca and the EPR formed active axes (by way of a red dashed line), as e.g. in the Scotia a single continuous ridge before the now missing segment Sea (between South America and Antarctica), or in the was “swallowed” beneath present-day California by the east of Australia. ese are zones where the former subduction zone. divergence stopped inside an ocean or a back-arc basin. One With a width varying from 1 000 to 3 000 km, the oceanic of the most interesting examples is that in an area of the ridges rise 2 500 to 3 000 m above the abyssal plains. e North Atlantic where the spreading process began between mean depth of the crest of these ridges is about 2 500 m Canada and Greenland in the Paleocene, then hesitated beneath the sea level. ey occupy nearly a third of the between west and east of Greenland in the Eocene. Even- surface of the seaoor. tually, this divergence ceased in the and the Ban Basin, and the opening jumped east to separate e Mid-Atlantic Ridge, with its winding outline similar to Greenland from . e Labrador Sea is that of the two sets of conjugated continental margins, is the an aborted ocean with Greenland, aer a stage of dissocia- type example of seaoor spreading and related continental tion from the North American plate, reintegrated with the t. latter. II.2.2.4- Axis of mid-oceanic ridges II.2.2.5- Transform faults and fracture zones e axis of active mid-oceanic ridges marks the boundary One of the salient characteristics of the morphology of the between two divergent lithospheric plates. is boundary is oceanic basins is their sectioning, or slicing, by a set of long characterized by seismic activity. e axes are represented faults (black lines on the Map) that cut perpendicular to the by a continuous red line, a color that recalls the fact that mid-oceanic ridges. Between the ends of two successive they are a key element of the Earth volcanism since they are, segments of active axes, the fault undergoes strike-slip geologically speaking, continuously providing magma. motion and is seismically active. is part is called a trans- Depending on whether the divergence rate is low or high, form fault. Beyond and along the fault, there is no longer the morphology of the ridge diers. At low spreading rates any lateral displacement between the two sides of the fault (2 to 3 cm/year), as in the Atlantic, the topography is rough and it becomes a seismically inactive (F.Z.) and shows a deep axial valley (ri). At high spreading representing the "scar" of the transform fault. is type of velocities (about 15 cm/year), as in the East-Pacic Rise, the complex fault frequently reaches a length of several topography is smoother without deep axial valleys. is thousand kilometers. As one might expect, the largest contrast is strikingly noticeable on Sheet 1 (Physiography). fracture zones (some 6 000 km) are located in the Pacic e particular case of Iceland, with its subaerial oceanic Ocean: the Mendocino F.Z. (touching the eponymous cape, accretionary ris, was mentioned above (II.1.6). As for the near the border between California and Oregon), the Afar "triangle" (also located above a hotspot, marked HA), Clipperton F.Z., the Eltanin F.Z. system (between the Antarc-

9 eir thickness can reach several thousand meters when approaching the foot of certain continental margins, in particular those where the terrigenous supply comes from the high sedimentary input of very large river systems (Amazon, Ganges/Brahmaputra, Indus, ...). 10 e length goes down to some 60 000 km when taking into account only the cumulated length of the segments of oceanic accretionary axes. 11 is branch heads rst northwards with the Central Indian Ridge, then turns o north-westward with the Carlsberg Ridge, then runs westwards with the Gulf of Aden oceanic Ridge before connecting in a complex way, via the Afar zone, to the Red Sea. 12 e gures correspond to average values over a certain time lapse; they don’t necessarily mean the spreading occurs regularly every year. 13 For English-speaking authors, the terms of transform fault and fracture zone seem somewhat synonymous. 14 In spherical geometry, any mouvement corresponds to a rotation movement whose axis passes through the center of the Earth. e fracture zones consequently follow small circles centered on the plates’ rotation poles (which are to be distinguished from the planet’s rotation axis poles).

10 tica Peninsula and the continental margin of New Zealand) km, where it begins to become dehydrated. A case is the among others. Fracture zones are the markers of the Pacic ‘’. rotation between two divergent plates controlled by plate e total length of the subduction zones is approximatively tectonic geometries. e most remarkable example is 55 000 km, a size comparable to that of the mid-oceanic provided by the Agulhas-Falkland F.Z. joining the tip of ridges (cf. note 10). Southern Africa to the southern extremity of South. is F.Z. traces a near perfect small circle arc that aids recons- e active subduction zones are shown by a green line with truction of the fanlike opening of the South Atlantic. solid triangles whose tops are situated on the leading A good example of an important transform fault is the (overlying) plate to indicate the direction of the subduction. Owen FZ, in the NW Indian Ocean. is osets the active e convexity of island arcs is always facing the subduction ridge of the Gulf of Aden relative to that of Carlsberg trench (e.g. Lesser Antilles in the Atlantic, Mariana in the Ridge (located in the middle of the northern area of this Western Pacic), but a rectilinear shape may occur (e.g. the ocean), and then links this accretionary system to the Tonga-Kermadec in the SW Pacic). On the concave side of Makran subduction zone along the Pakistani and Iranian the island arcs, a back-arc basin opens by separating itself Baluchistans. is fault "transforms" therefore a divergent either from a continent (as in the case of Japan where a small movement into a convergent one (cf. also II.2.2.6). is oceanic basin within the Japan Sea partially separates it from SW-NE fracture ends up in front of Karachi, directly facing the eastern continental margin of Asia), or from another the thrust front of the orogenic belts bordering the west of island arc that has become a remnant arc, i.e. extinct. e the Indus valley and connecting to the Himalayan collision latter is illustrated in the case of the Mariana active arc belts. e Owen transform fault with its dextral motion, West Mariana (opening marginal) basinWest Mariana constitutes the boundary between the Indo-Australian and (remnant arc) ridge, or the activeTonga arcopening Lau the Arabian plates. Basinremnant Lau Ridge. On Sheet 2, only 22 examples of movements of large trans- e convergence zones are generally characterized in the form faults (or simply large wrench faults) are plotted submarine morphology by a subduction trench, a long and (double half black arrows in opposite directions) either in narrow depression normally delineated by the 5 000 or 6 000 an oceanic or continental domains. Only 3 examples are m isobath. e greatest depth recorded is 10 920 m in the mentioned here: 1- the transform faults that constitute the southern part of the Mariana Trench (Sheet 1). Trenches are northern (le-lateral) and southern (right-lateral) bounda- not always visible because, in some areas, voluminous ries of the Plate; 2- the right-lateral transform sedimentary input are released into the ocean by large river fault of San Andreas (sensu lato) linking the opening system systems that ll up part of the trench lengthwise. e upper of the Gulf of California, cutting through all the west of part of the sedimentary cover of the dipping plate abuts California and ending up at Mendocino Cape where it againtst the backstop (i.e. the rim of the leading plate) connects to the axis of the Juan de Fuca oceanic ridge; and instead of being swallowed by the subduction, and is hence 3- the le-lateral Levant fault joining the Red Sea to the "scraped o" (thus escaping absorption into the Earth’s collision zone of the Arabian plate with , and where mantle). is becomes piled up as imbricate thrust slices in its locally step-like shape opens the small Dead Sea and Sea front of the arc. Hence an accretionary sedimentary prism of Galilee basins. forms, the deformation front of which is indicated on the Map by a symbol similar to that of the subduction, but II.2.2.6- Subduction zones, subduction trenches and colored light blue with open triangles. In the between other trenches this thrust front and the axis of the lled part of the subduc- Like all plate boundaries subduction zones are seismically tion trench it was decided to show the age of the underlying active. However, in the tectonics of convergence, the oceanic crust, as yet to be subducted but, concealed by the (heavier) oceanic lithosphere of a subducting plate dips as a sedimentary prism that would have otherwise been shown more or less slanting slab beneath the edge of the overlying as part of the arc margin. e most remarkable illustration plate whose lithosphere is made of either the lighter conti- is provided by the Barbados accretionary prism, located in nental crust (case of a arc-cordillera) or continentalized front of the southern half of the Lesser Antillas arc. As a crust (case of an island arc behind which one nds a back- consequence of a huge sedimentary input coming from the arc basin, or marginal basin, of oceanic origin). is is the Amazon and Orinoco rivers, the maximum thickness of this reason why subduction zones are also denominated as accretionary sedimentary prism reaches some 20 km active margins, in contrast to nonseismically active passive beneath the island of Barbados. Two other prisms are drawn margins which result from the driing of two continental on the Map: the Mediterranean complex located to the south blocks from either side of an initial ri (as in the case of the of Calabria and Greece, and the Makran. An interesting Atlantic). e subduction of oceanic crust generally produ- feature of the latter is that the inner part of the prism has ces a volcanic line that is at the origin of island or cordillera emerged and constitutes the coastal region of Baluchistan. arcs (cf. also II.2.1.3). ese volcanoes (characterized by is is the reason why, in this case, the axis of the subduction explosive, hence dangerous, eruptions) are located above a is plotted onshore and indicates the contact between, on one strip of the subduction slab, starting at a depth of some 100 side, the "backstop" represented by the lithosphere of the

15 See note 11. 16 e strike-slip movement is dened by standing on either side of the fault and observing the direction of the motion of the opposite side. If it moves to the right, the movement is dextral or right-lateral; if it moves to the le, it is sinistral or le -lateral. 17 ese small basins generated by strike-slip faults are more commonly called "pull-apart basins". 18 A part of the earthquakes generated by the subduction are distributed along the dipping lithosphere; is seismic slab is called "Wadati- Benio zone", aer the name of the two geophysicists who discovered this phenomenon. 19 e part of the dipping "slab" generating the subduction volcanoes is located at a depth rarely exceeding 150 km.

11 leading plate (Eurasia), and on the other side the lithos- hotspot (whatever the signication attached to this concept, phere of the subducting plate (Arabia). cf. II.3) having, with some exceptions, a relatively stationary position. ey are of three types: ere are very few places where an incipient subduction presently occurs. It represented on the Map by the symbol - submarine seamounts, relatively small, mainly covered by of active subduction (but with green open triangles). is sediments, and whose summit is sometimes at (in this case is the case of the Mussau Trench (about 149° E, 05° N) called "guyots"), resulting from the erosion of a subaerial where the Caroline plate begins to dip beneath the large volcano sinking progressively beneath the sea under the Pacic plate. is also occurs the north of the Lesser eect of normal thermal subsidence. Sunda island arc (subducting southwards) in order to - oceanic plateaus (cf. also II.1.3). accommodate the docking of the Australian continental - hotspots tracks (or trails)¸ formerly denominated "aseismic margin as a result of its northward convergence toward ridges" because these ridges lack of seismic activity compa- the Sunda arc. red to the mid-oceanic ridges located at plate boundaries. It is worth noting a case of extinct subduction Geologically speaking an is generally built (represented by a similar symbol, but in purple up during a short period of time from a pulse of intense dashed/dotted line) in the Vitiaz Trench whose maximum hotspot activity. A number in red followed by Ma indicates depth is only 5 600 m and stretches from the Solomon the average age of the plateau (e.g. "123 Ma" for the age of Archipelago to the northern tip of the Tonga arc. During the Manihiki plateau, to the NE of Samoa), or as two the Miocene, the arrival of the Ontong Java oceanic numbers separated by & when the build-up is believed to plateau blocked the whole system (cf. II.2.2.7) because it have occurred in 2 main pulses. When the age is uncertain was not dense enough (buoyancy eect) to be absorbed by inside a time lapse, the range is given by two hyphenated the subduction of the Pacic Plate which was previously numbers. e age, sometimes quite approximate, is given subducting southwards under the Indian-Australian plate. only for 10 oceanic plateaus: in the Indian Ocean, the Maud is caused reorganization of the subduction as it now Rise (73 Ma; 0°E/W, 65,5° S), the Kerguelen Plateau (119 Ma dips in an opposite direction under the New Hebrides arc. & 100 Ma), Broken Ridge Plateau (95 Ma; 95°E, 30°S); in the Pacic Ocean: the Shatsky Rise (145 Ma; 160°E, 35°N), the e subduction zones are mainly concentrated around the Hess Rise (99 Ma; 180° E/W, 35°N), the Manihiki Plateau Pacic rim and are the modern expression of the old (123 Ma; 165°W, 10°S), the (121 Ma & fashioned term "Paci c ring of re". In this ocean, there is 90 Ma), the Hikurangi Plateau (120-100 Ma, close to the east a striking contrast between the island arcs (active or of New Zealand); in the Atlantic Ocean, the Caribbean remnant) and their marginal basins which are exclusively Plateau (90 Ma & 76 Ma), and the Sierra Leone Rise (73 Ma; distributed to the west, while to the east the subduction 20°W, 05°N). zones are only dominated by volcanic cordilleras (Andes, volcanic ranges of , Rocky Mountains). e Ontong-Java oceanic plateau, named aer the atoll Outside the Pacic only two subduction systems exist in located north of the Solomon Archipelago, is by far the most the Indian Ocean, those in front of the Sunda Islands and remarkable. It has the largest surface area, estimated at that of the Makran, and two in the Atlantic, with the some 2 million km, and a volume of some 40-45 million subduction of the Lesser Antilles and in the Scotia arc km3 with an anomalous crust whose thickness can reach (between South America and Antarctica). more than 30 km. It was formed in the middle of the Creta- ceous Period, ca. 122 Ma, and probably also during a second Not all submarine trenches are exclusively related to magmatic pulse around 90 Ma. Some authors believe that subduction. Some exist along transform faults that cut this plateau was generated by the "plume" of the Louisville across the axis of mid-oceanic ridges, particularly when hotspot (marked HE on this Sheet) situated in the south of the spreading rate is low. e Romanche Trench, in the Pacic (140° W, 50°S). As mentioned above (II.2.2.6), Central Atlantic (centered on the equator by 18° W, 300 this plateau reached the former Vitiaz subduction zone km long) has the record for this type of feature with 7 758 some 20 Ma ago, then collided with the Solomon island arc m depth. about 4 Ma ago. is caused the blocking of the subduction II.2.2.7- "Anomalous" submarine features (seamounts, because of its lower density compared to the normal oceanic oceanic plateaus, hotspots tracks) crust (buoyancy eect). ese constitute on Sheet 2 a large ensemble of all sized According to the classical theory, a hotspot is located more reliefs that aects the oceans and is represented by the same or less deep beneath a lithospheric plate that is moving over orange-brown hue recalling, in a subdued way, the colour it with a velocity and direction controlled by the accretio- of the traps on continents. Actually, all these features nary axis where that plate is being generated. In its early result from a generally powerful magmatic activity postda- existence, the hotspot generates a large plume. When ting the age of the "normal" oceanic crust. is magma- reaching the overlying lithosphere, it produces a volumi- tism aects the oceanic crust initially produced at the axis nous and relatively uid volcanism at the surface in quite a of the mid-oceanic ridges. If the structure of the oceanic short time lapse in geological terms. Subsequently, large basins were controlled only by the plate tectonics princi- outpourings are formed in contained, but somehow large, ples, the ocean would only display mid-oceanic ridges, geographic areas: traps, onshore, and oceanic plateaus, fracture zones, abyssal plains, and subduction trenches oshore. Aer dissipation of the plume, only the "tail" of with their associated island arcs. All these features, thus the hotspot remains active, evidently at a lower rate but for a being of volcanic origin, are generated by the activity of a much more extended period of time. is activity is recorded

20 Violet contours drawn inside these structures correspond to second order reliefs. 21 See the reservations to be considered in relation to this statement in II.3. 22 e formation of the Ontong Java plateau by a hotspot was recently questioned by the hypothesis of a very large meteoritic impact triggering a cataclysmic magmatic output (cf. Ingle S. & Con M., 2004, E.P.S.L., 218 :123-134). 23 At present, there are not known examples of trap or oceanic plateau in formation.

12 in the moving overlying plate by a chain of volcanoes that whole width of the so-called Indian-Australian plate. dri away from the feeding hotspot, rst as an active Actually, it is not yet a true boundary showing a clear volcanic center, then extinct, and nally subsiding below separation between an Indian plate and an Australian plate, the ocean surface. e links in this volcanic chain become but a zone where the basaltic substratum is deformed by a progressively older with the distance from the hotspot. compressive stress (in response to the collision of India e whole set of this linear chain forms the hotspot track against Tibet) and where diuse seismicity also occurs. (or hotspot trail). e most illustrative example is given by II.2.2.9- Submarine volcanism related to the opening of the (code HC on Sheet 2) where the volca- the North Atlantic Ocean nic activity is at present located beneath e Big Island (Mauna Loa and Kilauea shield-volcanoes, and the subma- A red hatching overprint shows the presence of SDRs rine Loihi volcano marked by a small blue triangle on (Seaward Dipping Reector sequences), located from Sheet 1). e oldest part of this hotspot track still visible is seismic reection surveys, or submarine basalt bodies. e the Meiji seamount (dated at 85 Ma), located just in front latter can be both outcropping or buried and all provide of the Kuriles subduction zone which will eventually evidence of an extensive volcanic province related to the subduct it. Notice that at halfway along its length (ca. 40 opening of the North Atlantic Ocean during the Paleogene Ma), the orientation of the hotspot track changes from (cf.II.2.2.4), and to the activity of the powerful Iceland (HD) SE-NW to S-N direction, evidence of the reorientation of hotspot. ese dynamics had an eect on the conjugate the motion of the Pacic Plate at that time. continental margins of Greenland (and sometimes beyond), on one hand, and of the British Isles and Norway, In addition to the above mentioned Hawaii hotspot track, on the other. is eruptive activity is known onshore the age of some di erent progression steps for 5 other (volcanism "V 1" in the legend) in Greenland, as well as in trails is indicated in the map with a red number without the Faroe Islands and Ireland (Giant's Causeway). It is the "Ma": interpreted that the SDRs correspond to a series of strata - La Réunion (HF). A trail that links this island to the with alternated volcanic ows (lava and pyroclastic depo- Deccan traps via Mauritius Island, Nazareth Bank, Chagos sits) and non-volcanic sedimentary layers. Bank, Maldives and Laccadives ridge. e subsequent opening of the from the creation of the Carls- II.2.2.10- SDRs related to the opening of the South berg mid-oceanic ridge has cut this trail in two and oset Atlantic Ocean the original alignment that also included the Saya de In the South Atlantic Ocean, oil exploration has more Malha and Seychelles banks (cf. also II.I.3). recently located SDRs (blue hatching) on the conjugated - Kerguelen hotspot (Hi) probably at the origin of the continental margins of Argentina and Namibia-South Broken Ridge Plateau and the Ninetyeast Ridge and Africa. e presence of these reectors is related to the perhaps of the Rajmahal and Sylhet remnant traps. opening of the South Atlantic Ocean and the presence of - Louisville hotspot (HE) whose trail () the Tristan da Cunha (HG) hotspot. ends up at the Tonga-Kermadec subduction zone (and e two examples of these Atlantic basins show that the maybe at the origin of the Ontong Java Plateau, as seen passive continental margins (i.e. generated by an earlier above). continental ri and no longer constituting a plate boun- - Tristan da Cunha hotspot (HG), at the origin of the Rio dary) are not solely "non volcanic", as previously presumed Grande Rise to the west, and of the to the before the discovery of SDRs. It might give some evidence east, that are connected to the Parana and Etendeka traps for the presence of a hotspot being required in the initial respectively that, as seen before (cf. II.1.3), formed a single riing of a continental block and the subsequent opening of LIP unit 133 Ma years ago, before the opening of the South an ocean. Atlantic. - Easter Island hotspot (HB) that produced the Sala y II.3 – HOTSPOTS Gomez Ridge continued by the Nazca Ridge whose eastern e hotspot theory (cf. II.1.3; II.2.2.7) was proposed by the extremity is subducted into the Peru Trench. Canadian geophysicist John Tuzo Wilson who rst II.2.2.8- Distributed or di use plate boundaries proposed it in 1963 (two years before he developed the transform fault theory) taking Hawaii as a base model, and A grey hatching covers some oceanic areas where the improved by the American W. Jason Morgan in 1971. is transform boundary (strike-slip motion) between two attractive theory had enormous success in consistently lithospheric plates is ill-dened. It is distributed over an explaining the distribution of specic volcanism generally area of variable width, e.g. between the North America seen outside the plate boundaries (hence its name of intra- and South America plates, or on a part of the transform plate volcanism) and is particularly evident in the oceanic fault to the east of Azores separating Eurasia from the domain. e initial hotspot list included a score of cases, . but its number rapidly expanded to 130 units, even more e largest region displaying this kind of diuse boundary indeed (about 5 200 according to Malamud and Turcotte in is located in the middle of the Indian Ocean where it links 1999). However at this point, quoting Don Anderson and a segment of the Central Indian (accretionary) mid- Kimberly Schramm, "this brings up the question of oceanic ridge to the Sunda subduction zone (from the semantics" (see footnote 28). Today, the list has been brou- north of Sumatra to the middle of Java). is crosses the ght down to a more reasonable number varying between 40

24 e Loihi, located 34 km to the SE of Big Island and culminating at a depth of –1000 m (at the "Pelé Pit"), is the most recent expression of the Hawaii hotspot. 25 e name of the ridge was coined aer its specic geographic position located along meridian 90° E. 26 In this case, the missing segment would have been progressively absorbed by this subduction, since the motion of the Pacic Plate was westwards. 27 Cf. D. Anderson and J. Natland, (p.134), see complete reference in foot note 28.

13 and 50 hotspots. But not all of them meet the basic criteria is current controversy is hosted by the very interesting ome. of the original model (without addressing the web site: < www.mantleplumes.org >, managed by the geochemical domain). ese are: a deep origin for the British geophysicist Gillian R. Foulger, who also published mantle plume and a long duration of the activity (several the authoritative book Plates vs Plumes. A Geological tens of million years) which determines the progression of Controversy (2010, Wiley-Blackwell). a volcanic track in surface. ose cases, disagreeing with Whatever it might be, it was considered of informative the classic model, are labelled shallow, weak hotspots or interest to plot the exact or inferred position of 45 hotspots hotlines, etc. e latter is exemplied by the NE-SW on the Sheet 2 (list given in the inset placed in the bottom Cameroon volcanic line where the age of the volcanism is of the Map). not distributed according to a regular migration throu- ghout time. It shows a more or less random mode, with the ey are categorized in 4 types of hotspots, taking into currently most active volcano being the coastal Mount consideration the criteria of Vincent Courtillot and co- Cameroon (+4 095 m) half-way between the extremities of workers (2003) in particular: the line located one at the north of the Cameroon Repu- 1/ "primary" hotspots interpreted to correspond to a power- blic, and the other beyond the small Pagalu Island (ex- ful plume, deeply rooted in the lower mantle and with a long Annobon). duration, marked HA to HG (large red continuous circle); e polemics around the hotspot concept has been harde- 2/ hotspots that might be considered as primary, shown Hh ning since the early 2000s, when some researchers (anti- to Hi (large red dotted circle); plumers, see e.g. the recent works of Don L. Anderson) 3/ less characteristic, problematic or controversial hotspots, denied the existence of a number of plumes. ey noted H1 to H34 (small red circle); proposed an explanation for the origin of LIP (Large 4/ hotspots supposed to has been extinct since much over 1 Igneous Provinces, cf. II.1.3) mainly attributed to dyna- Ma, but which would have le traces in the seaoor mics related only to the plate tectonics sensu stricto, which morphology (small blue circle). is would include the induce shear stress in the lithosphere favoured by pre- Great Meteor Bank (eH1) to the south of Azores that would existing lines of weakness such as fracture zones. is have built the New England seamount alignment and Saint case seems to apply to the Central Pacic –see in particular Helena (eH2). the works of IRD/IPGP (Valérie Clouard and Alain e rst three categories, considered as "alive" with an active, Bonneville) and USGS (Marcia McNutt and collabora- or recent (as in Hoggar) volcanism are mainly located at one tors)– with the hotspot track segments of Samoa (H27)- extremity of the trail. Most hotspots are to be found in the Rarotonga (H25)-Arago (H1)-Mcdonald (H21)-Founda- oceans. Only 6 are onshore: Afar (HA), Cameroon (H17), tion (H15), and with the Tahiti (H30)-Pitcairn (H24) Darfur/Djebel Marra (H13), Hoggar (H17), Tibesti (H32), segment. Yellowstone (H34).

***************

By way of conclusion…. … It is to be noted that Sheet 2 can be used as a basis to explicitly trace the contours of the dierent lithospheric plates, sub-plates, and micro-plates that make up the present surface of our planet through a relentless confrontation between creation dynamics and destructive processes. Two maps formerly published at the same 1:50,000,000 scale by CGMW usefully supplement the reading of this Map:

Plate tectonics from space (2006, N. Chamot-Rooke & A. Rabaute) displaying the present-day motions of the lithospheric plates, one in respect to the others;

Seismotectonic Map of the World (2002, A. Haghipour and coll.) showing the distribution of the earthquakes, particularly along plate boundaries, with dierent categories of magnitudes and focal depths of earthquakes.

28 D. Anderson & K. Schramm use in their paper « e complete hotspot catalogue » in: Plates Plumes & Paradigms (Geol. Soc. Amer., Special Paper no. 558, 2005, p. 19-29), with some humour, the neologisms « Notspots » and « Crackspots » to refer to these "dethroned" hotspots.

14 COMMISSION FOR THE GEOLOGICAL MAP OF THE WORLD

Geological Map of the World at 1: 25 000 000

( ird edition)

2010

ADDITIONAL NOTE

By Philippe BOUYSSE (CGMW)

Translation by Philippe Bouysse & Clara Cardenas (CGMW), reviewed by Françoise Cadet (University Paris VI). Geological Map of the World at 1: 25 000 000 (2010) SINGLE SHEET : GEOLOGY, STRUCTURE (3 sheets)

ADDITIONAL NOTE

As mentioned in page 4, the Geological Map of the World at 1: 25 000 000, which is the scale of the original dra s, was published in early 2010, some months a er the release (in 2009) of the Geological Map of the World at 1: 50 000 000, which is half the wall map scale. For cost reasons, this edition at 1: 25 000 000 consists of a single map (Geology, structure) that corresponds to Sheet 2 of the reduced scale version. Owing to its nal size, the wall map is split in 3 adjoining sheets: - e legends with the texts and the maps of the polar projections (at the 1: 23 000 000 scale) - e “Western part” of the Mercator projection map (the “”), from 180° to 20° W. - e “Eastern part” of the Mercator projection map (the “”), from 20° W to 180°. is multi-sheet geological wall map provides an alternative: either centering the assembled Mercator map on the Atlantic Ocean or on the Pacic Ocean (cf. p. 4). Both maps are completely identical and their geological content is therefore the same, except for some details. Obviously, the legibility of the map at the greater scale is improved, in particular as concerns the toponymy (islands, rivers, cities, ….). e wall map at 1: 25 000 000 scale diers from its reduced version in the following items: - the lettering and layout of the wall map are adapted to its nal size and cut. - the blue colour of the line displaying the boundary between the continental crust and the oceanic crust (COB) is darker; - the green colour of the ophiolites is brighter; - the size of the green solid triangles showing the active subduction zones is larger than those of the other thrust fronts; - the direction of the thrust fronts between the Banda Sea and the , which due to a technical incident was inverted in the 1 : 50 000 000 map, is correctly displayed. e text of the Explanatory Notes from page 6 to page 14 fully applies to the Map at the 1: 25 000 000 scale.

16 ERRATUM Concerning the motion of strike-slip faults

An incident occurred in the graphic treatment of the motion of strike-slip faults (double half black arrows in opposite directions) generated confusion in the representation of these symbols in some areas (5 cases), and also an irregular display of the arrows in others (2 cases : Philippines and New Zealand), the sense of the latter being nevertheless correct. Herea er are given the correct motions and positions.

1- Northern boundary of the Caribbean plate (3 cases: le -lateral)

2- N, W, and E boundaries of the Scotia plate (3 cases: le -lateral)

17 ERRATUM Concerning the motion of strike-slip faults

5- Northern New Guinea fault

3- Levant fault 4- Strike-slip fault in Afghanistan related to the India vs. Eurasia collision

6- Central Philippine Fault Zone 7- New Zealand Alpine Fault (relocation of the right-lateral symbol)

18 ERRATUM

Applies only to the map at the scale of 1 : 50 000 000

An incident occurred in the graphical treatment which inverted the direction of the triangles of the thrust front and of the facing incipient subduction, between the north of Flores and Australia. The rectification is given below.

Reproduction of the original map showing the exact direction of the incipient subduction (green open triangles) oriented southwards, located to the north of the Lesser Sunda Islands (Flores, etc.), and the thrust front (black open triangles, oriented northwards) both located between the Banda Sea and Australia.

COMMISSION DE LA CARTE GÉOLOGIQUE DU MONDE COMMISSION FOR THE GEOLOGICAL MAP OF THE WORLD 77, rue Claude-Bernard 75005 Paris, Phone +33 (0)1 47 07 22 84 Fax +33 (0)1 43 36 95 18 http://www.ccgm.org [email protected]

©CCGM/CGMW, 2010 Paris, France

With the support of: