can be very short

John F. Dewey*

Department of , University of California, Davis, CA 95616; and University College, Oxford OX1 4HH, United Kingdom

This contribution is part of the special series of Inaugural Articles by members of the National Academy of Sciences elected on April 29, 1997.

Contributed by John F. Dewey, July 12, 2005

In contrast to continent͞continent collision, arc– collision margin, common during the early stages of generates very short-lived orogeny because the buoyancy-driven when, driven by roll-back, arcs bend and wrap impedance of the subduction of continental , accom- themselves into remnant oceanic holes (7). Arc͞rifted margin panied by arc͞suprasubduction-zone , is re- collision involves the blocking of a subduction zone with buoy- lieved by subduction polarity reversal (flip). This tectonic principle ant, weak, continental , followed by subduction polarity is illustrated by the early Grampian Orogeny in the reversal͞flip (8, 9). To achieve a flipped polarity, the preflip British and Irish Caledonides, in which a wealth of detailed sedi- subducted lithosphere must be removed by slab break-off to mentologic, heavy mineral, and geochronologic data pin the Orog- allow postflip subduction (10, 11). This process has occurred on eny to a very short Arenig͞Llanvirn event. The Orogeny, from the many active continental margins, including New Guinea during initial subduction of continental margin to the end of the (9, 12), north Borneo (13), New Caledonia during postflip shortening, lasted Ϸ18 million (my). The collisional the Eocene (14), the during the Jurassic (15), the shortening, prograde-metamorphic phase of the Orogeny lasted 8 Urals during the late (16, 17), the Appalachian͞ my, extensional collapse and of midcrustal rocks Caledonian System during the Ordovician (10, 18), and in the lasted 1.5 my, and postflip shortening lasted 4.5 my. Strain rates northern midcontinent USA during the Paleoproterozoic (19). ؊15 were a typical plate-boundary-zone 10 . , to the The subduction-flip mechanism generates a very short-lived second sillimanite isograd, with extensive partial melting, occurred orogeny because the resistive forces of collisional shortening are within a few my after initial collision, indicating that conductive relieved by reversed subduction. Thus, the collision of short-lived models for metamorphic heat transfer in Barrovian are oceanic arcs with rifted continental margins, commonly involv- incorrect and must be replaced by advective models in which large ing suprasubduction zone (SSZ) ophiolite oduction, is followed volumes of mafic͞ultramafic magma are emplaced, syn-tectoni- by the longer subduction of a postflip ocean and the protracted cally, below and into evolving stacks. Arc͞continent collision events leading to continental collision, with which ophiolite generates fast and very short orogeny, regional metamorphism, and exhumation. obduction is rarely associated (20, 21). The timing and duration of an orogeny or orogenic cycle is notoriously difficult to define, 8 million orogeny ͉ continent͞arc collision ͉ orogenic mechanisms stratigraphically, by the youngest deformed and metamorphosed sediments, and the oldest sediments that rest unconformably on the denuded orogen, which yield, generally, long outside esti- rogeny is continental deformation resulting from plate- mates. Integrated , sedimentary facies package boundary-scale shortening, which usually generates moun- O analysis, and heavy mineral analysis enable much tighter and tain belts and occurs in several ways. First, continent͞continent finer definitions (10). collision may cause doubling of the thickness of the continental I report here on a short-lived early Ordovician orogeny (Gram- crust over a very wide (1,000 km) zone, with far-field continental shortening, and last for 50 million years (my). Such zones are pian; roughly equivalent to the Taconic Orogeny of the Appala- characterized by wide orogenic plateaux, with surrounding and chians) in the Caledonian Belt from to , in which internal basins bordered by thrust belts, such as the Himalayan͞ geologic data tightly constrain the orogenic sequence from the Tibetan͞Kunlun͞Tienshan͞Altai system. Second, there is the initial subduction of rifted margin sediments beneath a continent- ͞ spectrum from the progressive, mono-vergent, shortening col- ward-facing oceanic arc [478 mega-annum (Ma)], arc ophiolite lapse, over 40 my, of narrow ‘‘para-oceans’’ like the and the obduction, through 8 my of dextral transpressional (22) collision, Dinarides, to the far-field intracratonic shortening and crustal thickening and collisional tightening (475–467 Ma), sub- of complexes like the Amadeus Basin in central , the duction polarity flip (467 Ma), extensional collapse (467–464.5 Gamburtsev of Antarctica (1), and the Ancestral Ma), renewed shortening and retrocharriage during postsubduction Rockies (2) during the , the La Salle deformation flip shortening (464.5–460 Ma) to postorogenic (443 Ma) uncon- belt of Illinois (3), the Yanshan Orogen in China (4), the formity (10, 23, 24). Kennedy (25) recognized that the Grampian Palmyride Belt of Syria (5), the late Paleozoic Cuvette Centrale Orogeny must have been an Ordovician (26) rather than a Pre- of Zaire (6), and, probably, the Knoydartian Orog- or pre-Ordovician event. Dewey and Mange (10) recog- eny in Scotland. Third, Andean-style orogeny is generated by nized that the Orogeny must have occurred during a very short time episodic shortening of the leading edge of a plate overriding a interval during the deposition of the conformable stratigraphic subduction zone, with short-lived ignimbrite ‘‘flare-ups’’ and sequence of the South Mayo Trough (SMT), supported by geo- shortening events punctuating longer periods of ‘‘stasis.’’ Fourth, chronology (27) and geology (28). Orogeny is, perhaps, best un- decollement at the Moho may allow the crust to detach and be derstood by studying the facies and detrital mineralogy of coeval fed through a thickening wedge into a continuously and rapidly sedimentary basins; the Grampian Orogen permits the integration eroded crustal , as in the transpressive of these data with metamorphic͞geochronologic data. of . These four orogenic mechanisms involve long-lived, continuous, or episodic events. Fifth, locking͞ restraining bends in large intracontinental transform faults, such Abbreviations: Ma, mega-annum; my, million years; CSB, Clifden Steep Belt; RBS, Renvyle͞ as the Transverse Ranges in California and the Anti-Lebanon Bofin Slide; SMT, South Mayo Trough; SSZ, suprasubduction zone. Mountains, generate zones of ephemeral crustal shortening. All See accompanying Profile on page 15283. these forms are distinct from the sixth style of orogeny, gener- *E-mail: [email protected]. ated by the collision of an with a rifted continental © 2005 by The National Academy of Sciences of the USA

15286–15293 ͉ PNAS ͉ October 25, 2005 ͉ vol. 102 ͉ no. 43 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505516102 Downloaded by guest on October 2, 2021 The Western Irish and Scottish Caledonides: ophiolite sheet with increasing amounts of detrital chromite (10) The Grampian Orogen and Cr͞Ni͞Mg and decreasing Fe͞Ti in the sediments (36). The The Grampian Orogen consists of the deformed and metamor- Glensaul Group is roughly correlative with the Derrymore͞ INAUGURAL ARTICLE phosed 15-km Dalradian sequence, exposed in the Caledonian Sheeffry and consists of a silicic arc sequence, indicating the orthotectonic zone (29) from Connemara, in western Ireland, to subduction of rifted continental margin sediments (37). Silicic the Grampian Highlands of Scotland (Fig. 1A). It continues into tuffs and lahars in the Sheeffry were derived from the Glensaul the Appalachian Taconic Orogen. Barrovian metamorphic as- arc to the south. The lower Derrylea is similar to the Sheeffry semblages, in and thrust nappe stacks, include garnet, with thicker turbidites; as in the Sheeffry, west- and south- staurolite, kyanite, sillimanite, andalusite, and chloritoid. In flowing, dark green turbidites contain abundant chrome spinel Scotland, the Grampian Zone has Siluro-Devonian sinistral and rounded purple zircon, whereas north-flowing, paler, tur- boundaries, the Great Glen and Highland Boundary bidites contain silicic arc debris including euhedral colorless Faults, both with displacements equal, at least, to their lengths zircons. In the upper Derrylea, a major change occurs at a (30). In Scotland, a direct stratigraphic lower age limit to massive, thick, turbidite horizon in which staurolite, almandine, Grampian deformation is given by unconformable Lower De- and chloritoid appear with floods of large muscovite grains; the vonian strata (417 Ma), a certain upper age limit by the topmost zircon population shows a diminution of rounded purple in favor Lower Cambrian Dalradian Leny (525 Ma), and a of subhedral and euhedral. The Glenummera is pelitic with thin, probable upper age limit by the putatively early Ordovician fine-grained turbidites and a sparse heavy mineral population Dalradian McDuff (485 Ma). In western Ireland (Fig. 2), similar to the upper Derrylea. The is, mainly, a coarse, a lower age limit is given, uncertainly, by the unconformable, red-brown, fluviatile arkosic group with six ignimbrite Lower Llandovery sequence (443 Ma) and, horizons, a greatly reduced rounded purple zircon and chromite population, and common almandine, rounded apatite, and eu- certainly, by the Upper Llandovery Lettergesh sequence (433 hedral volcanic zircon. The SMT was shortened and eroded Ma). The direct stratigraphic evidence, therefore, constrains the before the deposition of Llandovery sediments. sedi- Grampian Orogeny, certainly, to between 525 and 433 Ma and, ments contain volcaniclastic detritus and a diverse heavy mineral less certainly, to between 485 and 443 Ma. assemblage of Grampian metamorphic and Precambrian base- In western Ireland (Fig. 1 B and C), the Dalradian outcrops in ment , testifying to the deep of the Grampian Connemara and northwest Mayo, between which lie the SMT Orogen by the Llandovery. and the Clew Bay Complex. The Clew Bay Complex is mainly an Geochronologic data (see Supporting Appendix, which is pub- accretionary complex that shares a Grampian deformation se- lished as supporting information on the PNAS web site) allow quence with the Dalradian (31, 32), contains glaucophane, Grampian structural, metamorphic, and igneous events to be pyrope, and phengite and represents the collisional zone dated with great precision and tied into the Ordovician bios- between the Laurentian rifted continental margin and an Or- tratigraphic sequence (Fig. 2). References are given in Support- dovician oceanic arc to the south. The SMT was the precolli- ing Appendix, and the corresponding numbers are cited in square sional fore-arc, and syn to postcollisional successor sedimentary brackets in the main text. Detrital ages in the SMT [38] are basin that records the generation and unroofing of the Grampian illuminating; the Killadangan and Letterbrock have only Ar- Orogen (10). Post-Grampian terrane repetition occurred by chean to Proterozoic ages, confirming their pre-Grampian age. dextral transtension, then dextral transpression, to detach, move, The Sheeffry and Derrylea contain a mixture of Precambrian and dock the Connemara Terrane, before the Late Llandovery, and Grampian ages, confirming their syn to post-Grampian age. which placed the Connemara Terrane ‘‘outboard,’’ to the south, The Rosroe, , Mweelrea, and Derryveeny contain of the SMT; Connemara Terrane docking is recorded by the only Grampian ages. Detrital mica ages in the Southern Uplands unfossiliferous Derryveeny Formation, which contains boulders of Scotland [39] occur in turbidites in an accretionary of of certain Connemara provenance. The , between a south-facing postsubduction-flip arc and record a spread from the Connemara Dalradian and the Lettergesh Sequence, reworks 505 to 433 Ma. a low-angle normal zone that records the extensional Structural, metamorphic, and igneous events in the Grampian detachment and docking of the Connemara Terrane (33). The Orogen are well-dated by reliable methods and show a great late Silurian (M. lundgreni zone) continental collision, resulting internal consistency. The precollisional oceanic (infant) arc is from closure of the , was sinistrally transpressive; represented by [2–5, 11–16, 21–24, 41–45, and 48]. Subduction the time of transition from dextral transpression cannot be beneath the infant arc is recorded by with ages from determined any more precisely than between Caradoc and 505 to 490 Ma (18). SSZ ophiolite complexes of the oceanic arc Wenlock. in , Ballantrae, and Newfoundland yield ages from 501 The Ordovician stratigraphic sequence of the SMT (Fig. 2, [30] to 475 Ma [40] and illustrate the dextral transtensional GEOLOGY columns K and L) has been interpreted as the fore-arc basin of disruption of the precollisional oceanic arc. In Newfoundland, the precollisional oceanic arc, which became a syn-orogenic the Twillingate Trondjemite [2, 3] cuts dextrally transtensionally basin on the southern margin of the evolving Grampian Orogen deformed mafic pillow lavas (Sleepy Cove) and is transtension- (10, 24). The oldest sediments, the Tremadoc͞Arenig Letter- ally deformed (40). The precollisional arc and its SSZ brock Formation, contain a heavy mineral assemblage domi- contain a complex range of polyphase-deformed mafic and nated by rounded purple zircons, with quartz͞feldspar͞lithic ultramafic rocks, which wholly predate the Grampian Orogeny ratios nearly identical to those of the Killadangan accretionary (41–43). The beginning of subduction of Laurentian rifted complex to the north from which, probably, they were derived. margin sediments beneath the infant oceanic arc is recorded by The Lough Nafooey Group and Bohaun Formation to the south the change from mafic to silicic volcanism, and in the La͞Sm are roughly correlative with the Letterbrock; they represent a ratio, in the Lough Nafooey͞Glensaul transition (37). The mafic, high-Mg, oceanic arc with calc-alkaline basalts, boninites, obduction of the infant oceanic arc and its SSZ ophiolites is and andesites, an infant arc like those of the west Pacific (34). recorded, in Newfoundland, by the mafic-protolith two- The Letterbrock also received ophiolitic detritus (epidote in the pyroxene͞garnet granulite sole beneath the Bay of Islands lower, chromite in the upper), probably from an ophiolitic, Complex, whose 40Ar͞39Ar ages [17, 18] are a little young for , backstop like those of the west Pacific (10, 35); the the likely initial obduction͞detachment age. Derrymore Formation continues the Letterbrock pattern. The In Connemara, North Mayo, and the Grampian Highlands of Sheeffry Formation records the erosional unroofing of an Scotland, the Grampian deformation͞metamorphic sequence is

Dewey PNAS ͉ October 25, 2005 ͉ vol. 102 ͉ no. 43 ͉ 15287 Downloaded by guest on October 2, 2021 Fig. 1. Geology of the Grampian Zone. (A) Schematic tectonic map of the Irish and British Caledonides. (B) Schematic tectonic map of the western Irish Caledonides. DD, Delaney ; MT, Mannin Thrust. (C) Schematic geologic section along line AB on B above. CG, Currywongaun ; GS, Gowlaun Slide; M, Mweelrea; SH, Sheeffry; SBS, South Achill Beg Slide; SCG, South Connemara Group, 1 and 2-first and second movement sense on SBS; all other abbreviations, symbols, and ornament are as in B.

15288 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505516102 Dewey Downloaded by guest on October 2, 2021 INAUGURAL ARTICLE

Fig. 2. Grampian and Taconic stratigraphic and geochronologic data. Ages [24, etc.] are defined in Supporting Appendix. Key to columns: A, Western GEOLOGY Newfoundland Ordovician Shelf; B, Notre Dame Bay arc stratigraphy; C, West Newfoundland ophiolites; D, Notre Dame arc ages; E, Quebec͞New England; F, ; G, Achill; H, Connemara; I, Clew Bay Complex; J, Scottish ophiolites; K, north limb of SMT; L, south limb of SMT (thicknesses in K and L in meters), detrital mica ages [71] in SMT; M, Derryveeny; N, Mweelrea; O, Derrylea; P, Rosroe; Q, Maumtrasna; R, Sheefry; S, Southern Uplands accretionary prism; T, detrital mica ages [70] in Southern Uplands accretionary prism. CO, Corsewall; CP, Croagh Patrick; DD, Delaney Dome; DL, Derrylea; DM, Derrymore; DV, Derryveeny; GE, Glenlee; GL, Glenummera; GO, Goldson; GS, Glensaul; GT, Goose Tickle; KC, Kirkcolm; KD, Killadangan; LB, Letterbrock; LN, Lough Nafooey; LP, Long Point; MB, Marchburn; MN, Maumtrasna; MT, Mannin Thrust; MW, Mweelrea; NB, New Bay; PL, Point Leamington; PP, Portpatrick; RR, Rosroe; SG, Saint George; SH, Sheeffry; SI, Shinnel; TH, Table Head. Key to timescale: A, Arenig; AS, Ashgill; C, Caradoc; L, Llanvirn; LC, Late Cambrian; LL, Lower Llandovery; ML, Middle Llandovery; T, Tremadoc; UL, Upper Llandovery. Key to Ordovician graptolite zones: an, Anisograptus; ap, Tetragraptus approximatus; ar, Didymograptus artus; cl, Dicranograptus clingani; de, Didymograptus deflexus; gi, Isograptus gibberulus; gr, Nemagraptus gracilis; hi, Didymograptus hirundo; ms, Didymograptus murchisoni; mu, Diplograptus multidens; ni, Didymograptus nitidus; te, Glyptograptus teretiusculus.

tightly constrained by high-precision geochronologic data (Fig. D4 Mannin Thrust, which places garnet-sillimanite-cordierite 2). North of the Clifden Steep Belt (CSB), Grampian D1 to D3 metamorphics over silicic volcanic rocks of the Delany Dome structures are north-vergent and folded by the gently east- Formation (474.6 Ϯ 5.5) [48] correlative with the Glensaul silicic plunging D4 Connemara Antiform, whose southern limb be- arc. The Grampian deformation and metamorphism, in Conne- comes vertical in the CSB and inverted to the south, in a typical mara, is precisely dated by the 474–470 Ma, calc-alkaline ‘‘retrocharriage’’ relationship, above the south-southeast verging mafic͞ultramafic, igneous Cashel Suite that, syn-kinematically,

Dewey PNAS ͉ October 25, 2005 ͉ vol. 102 ͉ no. 43 ͉ 15289 Downloaded by guest on October 2, 2021 injected the north-vergent nappe pile during D2͞D3 (44). Gar- ably a result of the oceanic arc colliding obliquely with an net, staurolite, and kyanite growth during D1͞D2 was followed irregular rifted Laurentian margin. In Newfoundland, volcanism by a sillimanite peak during D3. To the south of the CSB, the in the continentward-facing arc lasted until Ϸ460 Ma in contrast Cashel Suite is developed pervasively in, and provided the heat to 467 Ma in western Ireland. The postflip, continental margin for, enveloping sillimanite͞cordierite . On the northern arc in Newfoundland is manifest as a granodiorite͞tonalite ‘‘sea’’ flank of the Connemara Antiform, north of the Renvyle͞Bofin of plutons [25–27]; possibly, some of the bodies classified above Slide (RBS), the Cashel Suite is represented by the syn-D3 as precollisional arc plutons may be syn-collisional or postflip. Currywongaun͞Doughrough Gabbro, Dawros (44), and sparse satellite minor intrusions. North of the RBS, the The Model regional metamorphic grade is low with widespread andalusite. The basic arc͞continent collision model for the Grampian From late D3 times, a permitted interpretation is that the Cashel Orogeny, outlined in Fig. 3 and illustrated in Fig. 4 (10, 23, 24), Suite is a calc-alkaline partial melt from the shallow astheno- is modified here in several substantial ways, notably by postsub- sphere of an obducting oceanic arc͞SSZ ophiolite, which later- duction-flip extension (47) and a renewed phase of middle ally injected the upper parts of north-vergent beneath the Ordovician postflip shortening. It is a strictly kinematic model obducting arc; at the present erosion level, deeper and hotter based on field relationships from geologic mapping, sedimento- (sillimanite) passed northwards into shallower and cooler rocks logic and structural field data, and geochronologic (Fig. 2) and (andalusite). The RBS may have originated as a north-vergent heavy mineral data. Fig. 3 summarizes the bulk sedimentary thrust that rooted in the CSB and carried the Cashel Suite facies packages of the SMT, Grampian structural and metamor- northwards. During and after D3, quartz-, , and phic events in the Dalradian rocks, and their tectonic interpre- pegmatites invaded the high-temperature region south of the tation. The transition from the mafic͞andesitic oceanic arc to the CSB [58–62], probably during the 466–465 Ma period of exten- silicic collisional arc occurred at 477 Ma, indicating melting of sional collapse. As yet, there are no zircon ages from the subducted Laurentian rifted margin sediments (37). Grampian of northwest Mayo, where Rb͞Sr and 40Ar͞39Ar give subduction to 30 km beneath accretionary prisms is well- younger ages [53, 54] than in Connemara and may be extensional documented. Woodford and Fraser (48) argued for sediment and erosional cooling ages. subduction to melting depths from Pb, Sr, and Be isotopes in In Scotland, the more tholeiitic Insch Gabbro Suite [50, 57, 73, volcanics of the Marianas Arc; similarly, Palmer (49) argued, 79] pins the D2͞D3 shortening interval as 472–468 Ma, identical from B isotope systematics, for the involvement of both oceanic to that of the calc-alkaline Cashel Suite in Connemara. Meta- crust and subducted sediments in volcanic rocks of the Halma- morphics, early granites, and gneisses [47, 56, 67–69, hera Arc. At (an arbitrary) 170 mm͞aϪ1 (per year), sediment 74–78] give an age range from 475 to 467 Ma, almost identical would take 1 my to reach a melting depth of 120 km at 650°C͞40 to Connemara. Scotland differs from Connemara in the south- kb, at a subduction angle of 45° (50). This mafic to silicic volcanic ward of the Iltay Nappe, a huge recumbent fold (45). transition is a very important harbinger of the earliest stages of I suggest that the Iltay Nappe developed during the south- arc͞rifted margin collision (37). It is seen in the Triassic͞Jurassic vergent extensional collapse phase as a high-level ophiolite sheet accretionary arcs of the Cordilleran System of western North was withdrawn southwards. The monoclinal downturn of the America (Howard Day, personal communication) where silicic southern margin of the Iltay Nappe to generate the downward- tholoids cap mafic pillowed sequences, in the late facing structures of the Highland Border (46) was probably Tsumba volcanics, and in the early Miocene Finisterre volcanics developed during the post-465-Ma phase shortening in the at the earliest stages of collision of the Ramu, Erap, Finisterre, postflip arc. and Ontong Java with the north Australian rifted The phase of postflip shortening is well-expressed in western continental margin (10). Similarly, Elburg, van Bergen, and Ireland. Retrocharriage of the north-verging nappe pile in Foden (51) have shown that anomalously radiogenic 3He͞4He Connemara caused south-verging dextrally transpressive rota- ratios and higher 208Pb͞204Pb for a given 207Pb͞204Pb are char- tion of the nappes into the CSB and south-southeastward acteristic of collisional sectors, contrasted to noncollisional transport of the Cashel Suite and its high-grade metamorphic sectors, of the extant collision zone between the northern rifted envelope along the Mannin Thrust over Ordovician silicic vol- margin of Australia and the Sunda͞Banda Arc. As described canics (Delaney Dome Formation at 460 Ma [64]). The D4 above, the precollisional arc already possessed, before collision, Connemara Antiform probably grew and was intruded by the a very complex deformation͞igneous sequence developed over Oughterard Granite [63, 65] at this time. Synchronously, the Ϸ30 my; this trantensional deformation (40) wholly predates the Mweelrea was deposited in a tightening SMT. Grampian and Taconic . Transtensional polyphase In the northern Appalachians, Taconic metamorphic rocks in deformation sequences, in oceanic arcs, constitute one kind of Quebec [1] yield a spread of ages from typical Grampian ‘‘spoof orogeny,’’ which is well-documented in southern Mexico numbers of 473 and 468 with a spread down to 453 Ma. In (52) and by Cook et al. (53) in the Devonian (54) Lizard Newfoundland, Taconic metamorphic rocks have yielded enig- Ophiolite Complex in southwest England. matic late Ordovician [19] and mid-Silurian [20] ages. The Progressive obduction of the oceanic arc and its SSZ ophiolite destruction of the rifted continental margin and continental complexes onto the rifted continental margin caused the stack- shelf by the collision of the oceanic arc is poorly expressed in the ing of subjacent, north-vergent, fold nappes (D1-D3), during by the termination of Durness Limestone deposi- which rapid prograde Barrovian metamorphism was caused by tion. No is preserved; the Moine Series of the the advection of heat from the syn-kinematic injection of mafic, Northwest Highlands, although containing a Grampian imprint, calc-alkaline, partial melts from the obducting subarc astheno- were thrust across the Durness Foreland in mid-Silurian times sphere. This obduction of a hot arc͞SSZ ophiolite advecting heat during the Laurentia͞Baltica collision. In western Newfound- into a subjacent metamorphic complex contrasts to obduction of land, the St. George carbonate shelf was terminated by uplift the cold Josephine Ophiolite over a hot subjacent arc (55). The caused by a westward-migrating flexural peripheral bulge. This earliest phases of obduction may have been below sea level but, uplift was followed by subsidence, from the Table Head Lime- as crustal thickening proceeded, the SSZ ophiolite rose above stone into the Goose Tickle foreland basin turbidites. The sea level, and erosion bit deeper into the high-level obducted Humber Arm , including the high-level Bay of Islands ophiolite complex giving a detrital heavy mineral sequence of Ophiolite Complex, was emplaced at Ϸ460 Ma. The Taconic decreasing epidote͞increasing chromite (10) and decreasing Fe Orogeny is slightly younger than the Grampian Orogeny, prob- and Ti͞increasing Cr and Ni (36). Rapid erosion of the uplifted

15290 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505516102 Dewey Downloaded by guest on October 2, 2021 INAUGURAL ARTICLE

Fig. 3. Phases of the Grampian Orogeny and their geologic signature. SUT, Southern Uplands Terrane. Numbers indicate ages in millions of years.

collided arc is expressed in the Rosroe Group in which large collision zone were exhumed. The rapid exhumation of high- tonalite boulders sit in channel and fan deposits, which coincides grade metamorphic rocks occurs in a variety of other ways, with a 467-Ma biotite-cooling- in the Dalradian rocks of including by late orogenic transtension (57), intracontinental Connemara. plate boundary transtension (40), subducted wedge exhumation, Eventually, buoyancy forces, resulting from collisional short- late orogenic collapse in continent͞continent collision (56), ening and crustal thickening, caused a flip in subduction polarity rapid erosion of a shortening monocline as in the dextrally from toward to away from the continent. Slab break-off is a transpressive Southern Alps of New Zealand, or postsubduction necessary precondition to allow the flipped subducting slab free flip extension as in the Grampian Orogen. Ϸ access into the (10, 11). Subduction flip led, At 465 Ma, a new phase of silicic volcanism and shortening rapidly, to a phase of orogenic extension (47), probably driven by in a continental margin Andean-type arc developed. In this Ϸ subduction roll-back, and buoyancy forces. Major extensional phase, from 465 to 460 Ma, the Mweelrea, west-flowing, detachments formed (e.g., RBS, South Achill Beg Slide) in fluviatile sequence with its ignimbritic volcanism was depos- ited in the tightening SMT. Simultaneously, retrocharriage whose footwalls midcrustal rocks were uplifted rapidly at Ϸ466 GEOLOGY occurred in Connemara to rotate the nappes of the Twelve Ma to provide, suddenly, detrital staurolite, garnet, and silli- Bens into the CSB and cause major south-southeast thrusting manite to the upper Derrylea͞Rosroe͞Maumtrasna. Apart from along the Mannin Thrust and the development of the D4 the SMT and the Southern Uplands, the British and Irish Connemara Antiform and the Delaney Dome. In South Con- Caledonides have little or resulting from the nemara, a new, south-vergent, accretionary subduction com- erosion of the Grampian Orogen, which is testimony to exten- plex developed (58), into which was delivered coarse meta- sional collapse as an important, perhaps dominant, morphic detritus of clear Connemara provenance (10). Before mechanism in the Grampian as in other orogens (56). This the Upper Llandovery, the Connemara Grampian Terrane extensional phase culminated in major subsidence, represented detached, in dextral transtension, from the Laurentian margin by the slates and thin distal turbidites of the Glenummera, during and docked, in dextral transpression, to generate the Gram- which the extensionally denuded Grampian Orogen may have pian detritus of the Derryveeny. This episode of Grampian D4 sunk, mostly, below sea level. The preservation of the anchi- and SMT shortening, and Derryveeny transpressional docking, metamorphic Clew Bay Complex, SMT, and dismembered Deer generated the subSilurian unconformity of western Ireland Park Ophiolite Complex is further witness to the extensional and represents the last vestiges of the Grampian Orogeny and unroofing of the Grampian Orogen; as part of the highest-level provided the beginning of a new template for the weak, late north-vergent nappe, these rocks were preserved by slipping Silurian, collisional orogeny. back along an earlier thrust below which glaucophane͞phengite Geochronologic data indicate a long Grampian ‘‘tail’’ (10) metamorphics that had been generated in the arc͞continent extending into the Silurian, which can be interpreted as resulting

Dewey PNAS ͉ October 25, 2005 ͉ vol. 102 ͉ no. 43 ͉ 15291 Downloaded by guest on October 2, 2021 Fig. 4. Schematic Ordovician sectional evolution of the Grampian Orogen.

from slow cooling (10, 59, 60) as the Grampian Orogen was expected in a 2,000-km arc͞continent collision, and was chopped up slowly denuded and cooled by erosion. This interpretation of the and variably obscured by Silurian left-lateral terraning. The exten- tail contrasts with interpretations (61) that it is the result of sional collapse phase is recorded by the footwall uplift and erosion partial and variable resetting by a late Silurian overprint. This of staurolite-grade metamorphics, the RBS, and the exhumation of interpretation cannot be correct because Silurian deformation Achill blue- (47). Strain-rates to generate this short-lived and metamorphism, resulting from the sinistral transtensional orogen were a ‘‘normal’’ plate boundary rate of 10Ϫ15 (10). How- closure of the Iapetus Ocean, is restricted to areas of Ordovi- ever, the Grampian Orogen, especially in western Ireland, is a cian͞Silurian rocks in the paratectonic areas of the British and warning for simple, conductive, thermal relaxation models for Irish Caledonides and is largely absent from the orthotectonic regional metamorphism (62, 63). Such models may be appropriate Grampian areas (29). for long-lived metamorphism in continent͞continent collision, but they cannot explain the short-lived (5 my), type-locality, Barrovian Conclusions: Rates and Mechanisms of Orogenic, Structural, metamorphism of the Grampian Orogen. In Connemara, meta- and Metamorphic Events morphism to the second sillimanite isograd and extensive partial The key to understanding the Grampian Orogeny is the synchro- melting was achieved, rapidly, in demonstrable connection with the neity of the Grampian structural͞igneous͞metamorphic sequence injection of a ‘‘sea’’ of calc-alkaline mafic͞ultramafic intrusions, with the stratigraphic evolution of the SMT (10). The SMT records which were derived, probably, from the obducting arc , the faithfully, in its sedimentary detritus and facies, events in the base of which was at a minimum of 900°C, and during the resulting evolving Grampian Orogen, from the initial ‘‘soft touch’’ between postflip slab break-off. the oceanic arc and the Laurentian rifted margin (478 Ma), through Erosional detritus, at least in Ordovician sedimentary basins hard collision, SSZ ophiolite obduction, and nappe formation, to now within or adjacent to the Grampian Orogen, is insufficient 467-Ma subduction flip, roll-back, and extensional collapse until to account for the denudation of the Orogen solely by erosion. 464.5 Ma, then renewed shortening and retrocharriage until 460 Crustal thinning by postflip extensional collapse is the likely Ma, for an ‘‘orogenic’’ total of 18 my. The preflip shortening phase explanation for the deficit and is probably typical of arc͞ lasted only 8 my; then, high-grade metamorphic rocks were almost continent collision orogeny, where collapse, aided by subduction immediately exposed and eroded. Events in the colliding arc͞ zone roll-back, is promoted toward the new subduction zone. continental margin also are recorded in the Newfoundland shelf, High-grade metamorphic detritus suddenly appears in the SMT, where a westward-migrating flexural peripheral bulge is seen in the immediately after the erosion of an SSZ ophiolite nappe, without sub-Table Head unconformity and the subsidence recorded by the the progressive erosion of a nappe stack through the lower and transition into the Goose Tickle shales and ophiolite-bearing flysch. middle metamorphic grades. Footwall uplift beneath extensional This Ordovician orogeny affected the Laurentian rifted margin detachments, immediately after subduction polarity flip, seems from Scotland to Alabama, albeit slightly diachronously as would be the likely explanation (47). However, this type of very short

15292 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505516102 Dewey Downloaded by guest on October 2, 2021 arc͞continent collision orogeny probably never builds high I acknowledge, with deep gratitude, the help and advice, over many mountains like those of protracted continent͞continent collision years, of those senior doyens of Appalachian͞Caledonian geology, such as the and Tibet. It seems to be a feature of John Graham, Tony Harris, Ben Kennedy, Maria Mange (who drafted INAUGURAL ARTICLE periods of continental dispersal, high sea level, carbonate plat- the figures), Stuart McKerrow, Adrian Philips, Paul Ryan, John Ramsay, Nick Rast, Jack Soper, Rob Strachan, Brian Sturt, Janet forms, high faunal diversity, metamorphism, and SSZ Watson, and Hank Williams, and of my 56 graduate students. I thank ophiolite obduction, typical of the early Ordovician and mid- Maria Mange and Eldridge Moores for their comprehensive and (21, 64). detailed reviews.

1. Veevers, J. J. (1994) Geology 22, 593–596. 33. Williams, D. M. & Rice, A. H. N. (1989) 8, 417–428. 2. Dewey, J. F. & Pitman, W. C. (1982) Philos. Trans. R. Soc. London A 305, 34. Bloomer, S. H., Taylor, B., MacLeod, J., Stern, R. J., Fryer, P., Hawkins, J. W. 145–148. & Johnson, L. (1995) Am. Geophys. Un. Mon. 88, 1–30. 3. McBride, J. H. (1997) Bull. Geol. Soc. Am. 108, 1502–1513. 35. Fryer, P., Mott, M., Johnson, L., Haggerty, J., Phipps, S. & Maekawa, H. (1995) 4. Chen, A. (1998) Tectonophysics 292, 17–42. Am. Geophys. Un. Mon. 88, 259–279. 5. Al Saad, D., Sawaf, T., Gebran, A., Barazangi, M., Best, J. A & Chaimov, T. A. 36. Wrafter, J. P. & Graham, J. R. (1989) J. Geol. Soc. London 146, 213–215. (1992) Tectonophysics 207, 345–358. 37. Draut, A. E. & Clift, P. D. (2001) Geology 29, 543–546. 6. Daly, M. C., Lawrence, S. R., Kimun’a, D. & Binga, M. (1991) Nature 350, 38. Idleman, B., Dewey, J. F. & Mange, M. A. Irish J. Sci., in press. 605–607. 39. Kelley, S. P. & Bluck, B. J. (1992) Chem. Geol. 101, 143–156. 7. Dewey, J. F., Helman, M. L., Turco, E., Hutton, D. H. W. & Knott, S. D. (1988) 40. Dewey, J. F. (2002) Int. Geol. Rev. 44, 402–439. Geol. Soc. London Spec. Publ. 45, 265–283. 41. Karson, J. & Dewey, J. F. (1978) Bull. Geol. Soc. Am. 89, 1037–1049. 8. McKenzie, D. P. (1969) Geophys. J. R. Astron. Soc. 18, 1–32. 42. Suhr, G. & Cawood, P. A. (2001) Bull. Geol. Soc. Am. 113, 1025–1038. 9. Dewey, J. F. & Bird, J. M. (1970) J. Geophys. Res. 75, 2625–2647. Geol. Soc. Am. Spec. Paper 349, 10. Dewey, J. F. & Mange, M. A. (1999) Geol. Soc. London Spec. Publ. 164, 55–107. 43. Suhr, G. & Edwards, S. J. (2000) 432–442. 11. Teng, L. S., Lee, C. T., Teal, Y. B. & Hsiao, L. Y. (2000) Geology 28, 155–158. 44. Wellings, S. A. (1998) J. Geol. Soc. London 155, 25–37. 12. Hill, K. C. & Raza, A. (1999) Tectonics 18, 950–966. 45. Krabbendam, M, Leslie, A. G., Crane, A. & Goodman, S. (1997) J. Geol. Soc. 13. Rangin, C., Jolivet, L. & Pubellier, M. (1990) Bull. Geol. Soc. France 6, 889–905. London 154, 15–24. 14. Aitcheson, J. C., Clarke, G. J., Meffre, S. & Cluzel, D. (1995) Geology 23, 46. Shackleton, R. M. (1958) Q. J. Geol. Soc. London 113, 361–392. 161–164. 47. Clift, P. D., Dewey., Draut, A. E., Chew, D. M., Mange, M. & Ryan, P. D. (2004) 15. Day, H. W. & Bickford, M. E. (2004) Bull. Geol. Soc. Am. 116, 1515–1528. Tectonophysics 384, 91–113. 16. Brown, D., Hetzel, R. & Scarrow, J. H. (2000) J. Geol. Soc. London 157, 48. Woodford, J. D. & Fraser, D. G. (1985) Geochim. Cosmochim. Acta 49, 901–904. 1925–1930. 17. Ayarza, P., Brown, D., Alvaez-Marron, J. & Juhlin, C. (2000) J. Geol. Soc. 49. Palmer, M. R. (1991) Geology 19, 215–217. London 157, 1065–1076. 50. Nicholls, G. T., Wyllie, P. J. & Stern, C. R. (1994) Nature 371, 785–788. 18. Van Staal, C. R., Dewey, J. F., McKerrow, W. S. & MacNiocaill, C. (1998) Geol. 51. Elberg, M. A., van Bergen, M. J. & Foden, J. D. (2004) Geology 32, 41–44. Soc. London Spec. Publ. 143, 199–242. 52. Ratschbacher, L., Riller, U., Meschede, M., Herrmann, U. & Frisch, W. (1991) 19. Holm, D. K., Van Schmus, W. R., Boerboom, T. J., Schweitzer, D. & Schneider, Geology 19, 1233–1236. D. (2005) Bull Geol. Soc. Am. 117, 259–275. 53. Cook, C. A., Holdsworth, R. E., Styles, M. T. & Pearce, J. A. (2000) J. Geol. 20. Dewey, J. F. (1976) Tectonophysics 31, 93–120. Soc. London 157, 1049–1064. 21. Dewey, J. F. (2003) Geol. Soc. Am. Spec. Paper 373, 153–158. 54. Clark, A. H., Scott, D. J., Sandeman, H. A., Bromley, A. V. & Farrar, F. (1998) 152, 22. Harris, D. H. M. (1995) J. Geol. Soc. London 797–806. J. Geol. Soc. London 155, 595–598. 23. Dewey, J. F. & Shackleton, R. M. (1984) Nature 312, 115–121. 55. Harper, G. D., Grady, K. & Coulton, A. J. (1996) Tectonics 15, 296–613. 24. Dewey, J. F. & Ryan, P. D. (1990) Tectonics 9, 887–901. 56. Dewey, J. F. (1988) Tectonics 7, 1123–1139. 25. Kennedy, W. Q. (1958) Trans. Geol. Soc. Glasgow 23, 106–133. 57. Krabbendam, M. & Dewey, J. F. (1998) Geol. Soc. London Spec. Publ. 135, 26. Hutchison, A. R. & Oliver G. J. H. (1998) J. Geol. Soc. London 155, 541–550. 159–181. 27. Friedrich, A. M., Hodges, K. V., Bowring, S. A. & Martin, M. W. (1999) J. Geol. Soc. London 156, 1217–1230. 58. Ryan, P. D. & Dewey, J. F. (2004) J. Geodyn. 37, 513–529. 28. Soper, N. J., Ryan, P. D. & Dewey, J. F. (1999) J. Geol. Soc. London 156, 59. Dewey, J. F. & Pankhurst, R. J. (1970) Trans. R. Soc. Edinburgh 68, 1231–1236. 361–369. 29. Dewey, J. F. (1969) Nature 222, 124–129. 60. Dunlap, W. J. (2000) Geology 28, 139–142. 30. Dewey, J. F. & Strachan, R. A. (2003) J. Geol. Soc. London 160, 219–229. 61. Fitch, F. J., Miller, J. A. & Brown, P. E. (1964) Nature 203, 275–278. 31. Chew, D. M. (2003) Geol. Mag. 140, 53–78. 62. England, P. C. & Thompson, A. B. (1984) J. Petrol. 25, 894–928. 32. Chew. D. M., Daly, J. S., Page, L. M. & Kennedy, M. J. (2003) J. Geol. Soc. 63. Thompson, A. B. & England, P. C. (1984) J. Petrol. 25, 929–955. London 160, 911–924. 64. Dewey, J. F. (1988) Geol. Soc. London Spec. Publ. 37, 23–40. GEOLOGY

Dewey PNAS ͉ October 25, 2005 ͉ vol. 102 ͉ no. 43 ͉ 15293 Downloaded by guest on October 2, 2021