Continental Tectonics (1980)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

II PLATE- B O UN D ARY TECTON ICS

INTRO D UCTI ON Complexities of Modern ant! Ancient Subcluction Systems WARREN lIAMILTON U.S. Geological Survey Until the advent of plate-tectonic concepts in the late 1960's, geologists and geophysicists studying the compo- nents of ancient erogenic and magmatic terranes on the continents had no actualistic frarneworlc within which to comprehend the origins and interrelationships of those components. We now see that modern analogs for many features of the ancient terranes form primarily in conti- nental-margin arcs and island arcs, above oceanic litho- spheric plates sliding beneath continental plates or other oceanic plates. These active temnes can be understood in a plate-tectonic framework, and many relationships ex- plained on a genetic basis. Ancient tenants are being increasingly comprehended in teens of analogy with the active ones. The first decade of plate-tectonic explanations of conti- nental geology has been enormously fruitful. Increas- ingly, however, our early explanations appear overly sim- plistic; we have seen the broad relationships but have missed many ofthe detailed ones because we have visual- ized plate interactions as much less complex than has been the actual case. Correspondingly, recent studies of 33 active continental margins and island arcs show them to undergo rapid and complex changes in response to fast- evolving plate interactions. Too little communication takes place in either direction I,etueen the scientists pri- marily studying the modern marine systems and their d~y- land counterparts studying ancient systems. Some pur- portedly plate-tectonic explanations published recently for ancient systems are mere ad hoc conjectures, incom- patible with the known features of active systems. Con- ~ersely, the active-margin students could infer much about their domains by analogy with deeply eroded equivalent tracts exposed on land. A cast amount of work is required to increase over ~~n- derstanding of the products and processes of plate tec- tonics. Much more integration of data already obtained is possible. More data need to be gathered in both field and laboratory, with emphasis on the petrological. structural. and geophysical criteria for identifying, distinguishing, and characterizing the component products of plate inter- actions. on paleontologic and radiometric criteria Or defining their ages, and on paleobiogeographic, paleocli- rnatic, and paleomagnetic studies to constrain their global wanderings.

34 The brief summary here of complexities of modern arc systems is adapted primarily from my study (Hamilton, 1979) of the onshore and offshore tectonics of the Indone- sian, western Melanes~an, and southern Philippine re- gions. The discussion of ancient arcs in western North America Is derived mostly from another report (Hamilton, 19781. The reader is referred to those works for elabora- tion and documentation of the concepts summarized here and for references to the work, by hundreds of other geo- scientists. from which those concepts are derived. ACTIVE CONTINENTAL.UARGINS Where an oceanic lithospheric plate is subducted beneath a continental plate, a tectonic and magmatic system like WARREN HAMILTON that of modern Sumatra or the Andes is developed. \tost early continental plate-tectonic papers (including my owns, and unfortunately also many current papers. as- sumed erroneously that in such settings the oceanic plate tips abruptly down at a trench to an inclined trajectory, along which the plate dips directly into the mantle. Rather it is now obvious that the usual case is that exem- plif~ed by Sumatra ~ Figure 3. 1). The subducting plate dips very gently beneath the continental slope, landward of a trench with gently sloping sides, for a distance of 100 km or more; only beyond this distance does the plate roll toward the steeper dip of the inclined Benioff zone as deduced from mantle earthquakes. The continental slope typically rises gently to an outer-arc ridge, landward of which is an outer-arc basin. Along some active margins, the basin is full of sedimentary straw and is expressed i>JDIA.N JAVA TER RC C)UT=R-ARC BASIN SULFA . RA Cc =~ TR E.NC~ RIDGE _ Pre-subduct i on s trata, _ beneath teas i n sedi meets, ~ ~ A Water I .04 1 ie on ocean; c crust ~ O / ~ ~ a __ =` ~ .~eiange wedge 2.2 2.4 \~ Edge of trench is ~~ lower continental crust 2.9 1 i mi t of downbow i rig by `me I ange wedge ~ Water . ._ - - O 50 ~ - ._ - c, P.hase-transi tion z~ 5 US D L'C IliNG P 4~ T ~ \\ 0V ERRIDING PLA T E ~ Mantle 3.4 -_ 100 ~ .ASTHE~C)SPHERE Hor i zon ta I and ver t i ca i s ca i e Horizontal and vertical scale 0 50 KM 1 1 ! 1 ~ 1 Sof t trench and pelagic sediments SC raped c>Ff a t toe of wedge 7 Gravitational flow i rub r i ca tes me I ange wedge - .~ —Wedge s tands h i ghes t "here me I ange b u t t res sed age i ns t leading edge of i i thosphere p T a te G rev i ta t i one I f I ow deforms outer teas; n S trata \= _ in E o ~ 20 . _ 30~ l 40 - \ - Subduc t i nc ocean i c p 1 ate d rags and th i ckens me I ange wedge FIGURE 3.1 Section through the subduction system of the Java Trench, off southern Sumatra The Cretaceous system of California presumably was similar. The configuration of the melange wedge and the structure of the outer-arc basin are constrained by reflection profiling, and the shape of the subdue tiny plate is (onstrained primanly by the location of the inclined Benioff zone of mantle earthQualces. L'pper diagram: major components; densities in g/cm3. Lower diagram: mechanism of deformation of accretionary wedge `>t melange and imbricated materials.

Complexities of Modern and Ancient Subduction Systems bathymetucallY as ~ subhonzontal continental shelf. seismic-reflection profiles and some drilling show that between the continental-slope seafloor and the top of the subducting oceanic plate there is commonly an accre- tionary wedge fondled largely of sediments scraped off the undersliding plate and imbricated ant] converted to melange along shear structures that dip moderately land- `~vard, sharply discordant to the gentle dip of the oceanic plate beneath. Large ;~`retionaIy wedges develop where voluminous elastic sediments are conveyed into the s~b- duction system. Such sediments are deposited mostly longitudinally relative to the trench, as turbidites in the trench or as abyssal fans. Some recent papers have interpreted voluminous com- plexes of melange and sheared rocks in ancient sr~bduc- tion settings to be conned primarily by submarine slump- ing rather than by shearing within the ;ac~cretionary wedge and its tectonic basement. Such conjecture receives no support from studies of modem trench systems, within which large submarine slides are uncommon. Available data snuggest that an outer-arc basin. like that of Figure 3.1, den elops only where the leading edge ofthe continental plate consists of a strip of oceanic angst, either formed in contact w ith the continental crust by a previous spreading event or sutured to the continent following pre- ~ious subduction. The basin is formed by the raising of the leading edge of this strip, not by depression fit the center. The raising is presumably a result of the stoning beneath the strip of accretionary melange. Where the ac- cretionary wedge is buttressed directly against ~ leading edge of cont~riental crust, no basin is present. The active volcanism of the Sumatra system, like that of most s~bd~ction systems, is concentrated in a belt above that part of the seismic Benioff zone that is about 125 km deep. The volcanic rocks are of silicic and intel~`lediate compositions and, in part, fonn calderas developed by Option from large, shallow magma chambers. It is rea- sonable to assume that granitic batholiths moist now be forming beneath the `~,lcanic belt. The magmatic belt is Imposed on ~ regional anticline consisting of pre~ol- canic roclcs bulged broadly upward; apparently the angst has been thickened by ~olr~mino~s magmatic intrusions. BE HAN I()R O F MODERN ISLAND ARCS IsIand arcs do not sit stably fin tenderly ing mantle for an era, constantly Naming inc`'ming manic lith`~phere, although many c`'ntinenta1 geologists `~me them to be- ha`e Ail. Rather, arcs cl~aracteristicall~ migrate relative to the plates behind them; they reverse their s`~bdr~ction ~~>larit`, fir g<, deacl; they c`'llicle with each <'ther and ~ ith c`,ntinents; and tines are del`,rrned by strike-sli~' tilt and `'r`~clinal holds Tale flits <'t these ~'r`~s care e.Y- ceeclinglv c`>mple.x aggregates an(1 s~1~er~siti`'nx `'t airy materials. Oceanic crust ~lisal'pear.s It <urn arcs fit r``tes reac h ing at least 13 ~ mix ear fir 1;30 kiwi 1(P ~ ear; fin So arc cannot long face anything except.a spreading center without major changes in geography and geometry. Consider the southern Philippine Islands, an aggregate of the products of six Cenozoic subduction systems that are clearly identifiable and, undoubtedly, others that can- not yet be distinguished. The six recognizable island-arc systems diverge southward from the main st~uthe~n Philippines and, hence, preserve their individual trenches, ridges. and magmatic arcs. On the west is the extinct Palawan arc, trending northeast from the northern tip of Borneo. active during middle Tertiary time; subduc- tion stopped at about the end of Miocene time, without any collision or other obvious external cause. A fossil trench lies along the northwest side of Palawan; the island is the top of the melange wedge (formed mostly from quartzose elastic sediments derived from Asia and Bomeo and deposited longitudinally in the trench); and an extinct volcanic arc lies beneath the sea surface to the southeast. The next eastward subduction system is the Sulu island arc, active with southeastward subduction from at least late Miocene into Pleistocene time and possibly still sub- ducting at a very slow rate. The Sulu arc connects north- east Borneo to the Zamboanga Peninsula of .Uindanao, and the geology of the onshore continuations of the arc suggests a reversal of subduction polarity, from previous northwestward subduction beneath the arc, in Miocene time. Two sectors ofthe northern extension ofthe Sulu arc swung together in a Y configuration, intersecting between Zamboanga Peninsula and Negros Island. (Such Y's are evidence for a subduction mechanism by vertical sinking, rather than by inclined injection, of He downgoing plate. ! Central and eastern .\Iindanao consists in part ofthe fused north-trending Sangihe and Halmahera arcs, which col- lided as the intervening Molucca Sea floor was subducted beneath both of them, and in part of the products of the Cotabato and Philippine subduction systems, which have subduction polarities opposite to those of the precollision arcs. Each of these identified arc systems includes melange wedges, magmatic-arc complexes, and sedimen- tarv basin fills. Further complications has e been added by oroclinal folding and strike-slip faulting. Western Melanesia displays the products of Cenozoic interaction between n~mero~s small lithospheric plates of tast-changing character, formed between the obliq~elv converging Pac iliac and Indian-.Australian megaplates. The Solomon island arc now faces southwest. c~ompli- cated by ~ sector in ``hich ~ spreading ridge intersects its trench, bitt the arc t:aced northeast lentil middle Tertiary time. The northwest extension ot the Solomon arc. the !`' em Ireland-!~an~s are, is sliding northwest past trench-trench-transh3rm triple junction at the east end of the New Britain arc; new subduction breaks through fin the north side off the projection to define another ret ersa1 `,t sedation polarity so that s~bd~ction is again so`~th- `` estward. The t:ast-mo~ ing transform. offset along short xl)re~lding centers ( firs es `` es`ard from the ~'e`` Britain trickle junction t`, intersect the north margin fit central

36 New Guinea; west of that intersection, subduction south- west beneath New Guinea is very rapid, whereas east of the intersection subduction is slow. The slow-subduction Mussau arc system trends southward to a trench~ench- trench triple junction against the curving New Ireland- Manus arc, which continues past the junction to another trench~ench-trench junction off the north coast of central New Guinea. Northern New Guinea records the collision dttring Pliocene time of a south-facing island arc (itself likely a composite) with a continental margin, mostly stable since the Jurassic ricing away of another continental mass. The collision progressed eastward with time, and the east, or unblocked, part of the arc advanced past the end of the New Guinea continent to form the Papuan Peninsula. Subduction, previously northward beneath the island arc, reversed to southward beneath the continent, as enlarged by the addition of the arc and un- derlying lithosphere, after the collision; only subse- quently did subduction reverse beneath the peninsular extension. Strike-slip faulting, orocl~nal folding, and other collisions have been additional complications. Large accretionary wedges of melange and imbricated rocks are developed in front of only those island arcs along which voluminous sediments are available on the ocean floor. Abyssal-fan and longitudinal-trench sedimentation can, however, extend several thousand kilometers from a major continental source. Immature oceanic island arcs, represented above the surface, if at all, only by small volcanic islands, lack such sediment supply; these arcs face steep-sided trenches, at which the subducting oce- anic plate tips sharply downward and dips steeply beneath the arc, without any broad melange wedge and underlying gently dipping oceanic plate. Ophiolitic melange~pelagic sediments and mafic oceanic crustal rocks, in a matrix of serpentinite disrupted and hydrated from oceanic mantle rocks~re known in many erogenic assemblages that are now parts of continents, where they are associated with magmatic assemblages analogous to modern island arcs. These ophiolitic melanges may mostly represent the small accretionary wedges fanned along oceanic island arcs, and contest win melanges, dominated by te~Tigenous elastic sediments that fanned along active continental margins. .Uodern arcs provide no support for the various hy- potheses of back-arc abduction, or flake tectonics. Where slabs of oceanic crust and mantle have ridden onto other rock assemblages and are exposed in young arc terranes, dips and transport directions are as required for those slabs to be the fronts of overdying plates. ^\lECHANISM OF SUBDUCTION WARREN HA.MILTON tory. Neither mechanism of subduction can explain the geometry of many modern systems. The Banda Antill~c and Scotia arcs have subducting trenches that are C- shaped in plan and Benioff zones, hence subducted plates, that are spoon-shaped. Opposed arcs are in the process of colliding in the Slolucca Sea and Solomon Sea regions, over subducted plates that have an inverted V configuration in section. Such geometry suggests that the subduction process below a depth of 100 or 150 km may consist primarily of the vertical sinking of the downgoing plate and that the apparent dip of a subducting plate is a Unction of the rate of advance of the plate hingeline over the mantle beneath the subducting plate. This rate would be equal to the convergence rate only where the subducting plate is fixed with regard to the mantle beneath it and where no mar- ginal sea spreading occurs behind the advancing arc. The sinking velocity of the subducted lithosphere must be a function of the density contrast between subducting slab and surrounding mantle and, hence, must increase, at an exponentially decreasing rate, with age ofthat slab. Given a constant rate of hinge advance, the dip of a subducted slab should be a function of its age. . PRE-CRETACEOUS ISLAND-ARC AND MELANGE ASSEMBLAGES OF THE WE STE RN U N ITE D S TATE S The Cretaceous and early Tertiary development of the western conterminous United States was dominated by subduction of large plates of oceanic lithosphere directly beneath the continent. Similar activity of Andean type occurred during parts of Triassic and Jurassic time also, but much ofthe subduction recorded dunng those periods occurred beneath oceanic island arcs. The arcs consist dominantly of submarine basaltic and andesitic lavas, breccias, and tufts; their intrusive and volcaniclastic equivalents; carbonate strata; and diverse accretionary- wedge materials including ophiolitic melanges. These assemblages resemble those that characterize modem oceanic arcs. These arcs were added to the west edge of the continent when intervening oceanic lithosphere was wholly subducted beneath either the continent or arcs advancing toward it. Complexities ire these ancient, ac- creted arc tenanes are still poorly understood but are numerous enough to indicate that complex histories. com- parable with those of modern Indonesia, .~Ielanesia, and the Philippines, are recorded. We must be dealing with the products of consumption of dozens of large and small lithosphenc plates and with complex, jumbled aggregates of arc components that were in part pre-assembled far from their final resting place along the continental mar- gin. After Precambrian rifting and the formation of a new ocean in the direction that is now west, the western mar- gin of North America during early and middle Paleozoic time trended from west-central Idaho to east~entral Cali- The inclined descent of subducted lithospheric plates, down to depths of at least 700 km, is commonly visualized as representing either the injection of a slab down an inclined trajectory that is fixed with regard to the mantle or the gravitational sliding of a slab down such a trajec-

Complexities of.~fc~dern ctnd.Ancient Subduction Systems Doria. This margin was stable tectonically and on it de- ~eloped an oceanward-thickening wedge of continental shelf strata. In Late Devonian and Early Mississippian time, creep-water strata from the west, with some of their underlying oceanic crust, were shoved onto the shallow- water shelf strata, and both were pushed eastward above the continental basement in the Nevada sector (Figure 3.2). This event may record the collision of an island arc with the continent; but the arc itself did not remain at- tached to the continent. tor another period of stable- margin sedimentation followed, with deep water again to the west. Presumably the polarity of subduction reversed beneath the collided arc, which migrated away from the continent, opening an ocean basin behind it. Another similar collision occurred against Nevada near the middle of Triassic time. Again, dee~water sedimentary rocks and oceanic basement were driven eastward upon shallow- water strata, and both moved eastward over the conti- nental crust; but this time, the causative island arc re- mained attached to the continent. The Permian and Triassic marine faunas of this island arc differ only moder- ately from those of mainland North America, so the pre- collision wandering of the arc need not have been at vast distances from the continent. The collision was followed by a reversal of subduction polarity, and Upper Triassic arc magmatic rocks were erupted through the old conti- nental crust, above the subduction system dipping east- ward from the Pacific edge of the added island arc. Still greater motions of lithospheric plates are required by juxtapositions of other arc and melange assemblages that yield fossil or paleomagnetic evidence for distant sites of origin. One belt of such assemblages occurs in the tectonic-accretion terrane from the southern Sierra Ne- ~ada of Califomia to the Yukon Ternto~y in Canada; it is characterized by shallow-water Upper Permian lime- stones bearing tropical Asian fossils. Particularly distinc- tive among these fossils are fusulinids aIgal-grazing, calcareous-test protozoa~of the verbeekinid family. These are widespread in correlative deposits of paleo- tropical Eurasia yet are completely unknown in paleo- tropical mainland North America. (The equator of the time trended through Texas, in the direction now north- eastward, so appropriate habitats were available; but the American fusulinid assemblages are quite different from the Eurasian ones.) These fossils occur in limestone blocks in melanges, in reef limestones capping atolls, in the scraped-off complexes along the paleomargin of North America, and in part between island-arc belts. Subduction of more than 10,000 km of oceanic lithosphere beneath North America during Late Triassic and Jurassic time, and beneath island arcs that were themselves added to North America in Jurassic time by subduction of intervening oceans, is required. Oceanward of this tract in British Columbia and southern Alaska is still another tract of isIand-arc and related rocks, characterized by southem- hemisphere paleomagnetic latitudes. The Cretaceous and early Tertiary pattern of relatively simple subduction of oceanic lithosphere beneath the continent, with minimal .~_ ,~ ~ involvement of island arcs. followed the addition of these various terranes to North America. C R ETAC E O ~ S AN' D C E NO ZO I C CALI FORN {A SU BDUCTION CONFIGURATION Both the geology of Califomia and the seafloor-spreading history of the northeast Pacific Ocean require that during Cretaceous and early Tertiary time oceanic lithosphere was being subducted beneath Califomia. The Cretaceus components of most of Califomia fit the model of Figure 3.1, when palinspastic restoration is made to reverse the disruption and offset by late Cenozoic strike-slip faults and other structures. In the west, forming most of the Coast Ranges (Figure 3.~, is the accretionary u edge, the Franciscan complex of melange and imbricated rocks- mostly of abyssal uppermost Jurassic, Cretaceous, and lower Tertiary elastic sediments but including fragments of oceanic crust and mantle and of seamounts, and masses of high-pressure, low-temperature metamorphic rocks. Shear structures in the chaotic Franciscan complex typi- cally dip steeply to moderately eastward, yet gravity stud- ies indicate that the oceanic lithosphere beneath the wedge dips only gently eastward. Along the east side of the Coast Ranges, coeval outer-arc-basin strata, the Great Valley facies, lie with depositional contact upon oceanic crust and upper mantle, which in turn is in thrust contact with the tectonically underlying Franciscan accretiona~y wedge. Lower Great Valley strata were deposited at abyssal depth on oceanic cmst, a narrow strip of which remained attached to the continent while Cretaceous sub- duction proceeded. The strata lap eastward across the oceanic crustal strip and in the subsurface butt against what was in Cretaceous time the moderately sloping edge of continental crust. The leading edge ofthe oceanic strip was rotated upward concurrent with Cretaceous subduc- tion. Farther inland, the SielTa Nevada batholith, granitic rocks mostly of Cretaceous age, followed synchronously with the formation of the Franciscan accret~onary wedge and with the development of the Great Valley outer-arc basin. The batholith represents the subjacent equivalent of a volcanic arc such as that of modern Sumatra. Batho- lithic rocks of any one age display a cross-strike increase in the ratio of potassium to silicon. characteristic of action e volcanic terranes foxed above subducting plates. and the distribution of the rocks defining the gradient indi- cates moderately steep subduction dips compatible with the Franciscan and Great Valley geometry in a system like that illustrated in Figure 3.1. The magmatic record of Cretaceous and early Tertiary time In the western states indicates that the dip of the subducting slab was moderately steep and relatively con- stant during Early and early Late Cretaceous time but that the dip was markedly gentler during late Late Cretaceous

38 WARREN H~ILTO! 120° l ~ ~ ~ ~ volcanic ~ coLuMl~la PLATeAi} . )- ~ (~} ~ ~ ~ arc ,,:._ ~ '' I'd ~ 1 (_ `~_~c~~ ~ ,,, ~ ~ / Us ! 'I 9 .r-~,o.~ == // ~~: An eland of . cane c / ~ ,; ~~Devor'ion to To,. .~7orc ~ i= ''''a ~ ~ . ~ ~ . _ . . I ~ ~ ~ ~ / ITS. · 1 At- it_ . ~ J Carboniterous- ~ ~ ~ ~~~ ( Jura.ssic ore, , ~ Earty \ | oph~olite, ond ~ , Pa's02sic \ : f \me'~nge /f me~coge \ mater~ole, ~ / ~ \ J idt' :: ,_ : :? -. -. Jurassic ana Eari, Crctaccous:, ~ `~; ~~ ''~43 pl u tons in nor t h east Oregon . ~_ ;' `',: k1 ~_' postdcte most subduction ~ ~t i accretion \'( >"~-"~ ~Al :2`.. .::. NEOGENE BASALT Fills th~nned and r~fted crust I t ^~\~ ?~;~ !~.ndecino~ ~ ic,'; 40° ~ ~.i~, ~ ~ ~ ~ ~ ~ nm~Cr9tm~rt.t ~ ~ - tit ~ c ~ , . ~, , §' .'~k ~, ~ `^, I 's / PROv'NCE 1 l ~i~.\,\ ~ 0tOut 173 of wl~th I of tasin-rOnge proYinCe i repreSents Neogene crustal extens~on, northwesteard re~ct~ve tO croton P~poroz imCt. rt~ngel~ne De,~een /j P~160201C She1~/ and crotCnic ~ j \ st~atc ~ _ ~ `/ i ~Scn Gregorio fcult ~r,,cssic o, ~urc,s,c r,~t ot old continental morg'n I ~ angul ~ aC b~ you n ger p I u rons 200 1< M 0 f00 FIGI:RE 3.2 Selected tectonic elements ot the west-central United States. The present dis~iblltion of these elements was atTected greatly by disruption during middle and late Cenozo~c time.

Complexities of .~odern arid Ancient Subduction Sy.stem.s time, perhaps as the resmelt of an increased convergence rate between the .Nrorth American and subducting plates. Nlagmatic-arc patterns changed complexly during early Tertiary time, indicating corresponding variation in sub- duction dips. In Oligocene time, for example, the sub- ducting slab was apparently segmented: the southern seg- ment decreased in dip northward Prom moderately gentle to very gentle. there was an abrupt change across a line trending northeast Tom southern Califomia to northern Colorado, and the northern segment decreased in dip northward from moderately gentle to moderately steep (see Chapter 14). .~ch ofthe Tertiary complexity may be due to the subduction of oceanic lithosphere of varying age and density as the East Pacific Rise. offset along trans- ~nn faults, approached the continental margin and young lithosphere of variable temperatures was being subducted. Conversely, the spreading history and geom- etm of eastern Pacific plates, since wholly s~bd~cted beneath North America. can perhaps be inferred from the magmatic-arc migrations. In the Klamath Contains region of northwest Califor- nia and southwest Oregon, the Franciscan melange wedge lies directly beneath crystal assemblages that were part `~t the continental crust in Cretaceus time (Figure 3.~). No `~ter-arc-basin strata lie to the east. Here, the leading edge of the continental plate was ~~..'ed of conti- nental cutest; the narrow strip of attached oceanic crust is missing, and no o~ter-arc basin developed. LATE CENOZO}C DISRUPTIO.N During Cretaceus and early Tertiary time, the Pacific margin ot Califomia was ~ relatively simple system of parallel tectonic and magmatic belts, controlled by the igoro~s s~bd~ction then occurring beneath all of Califior- nia The subduction was more rapid than the spreading of the East Pacific Rise offsl?ore to the west, so the rise missed progressively closer to the trench despite the ac- ti~e spreading. The oceanic plate east of the rise was mov ing northeastward relative to North .\merica, whereas the plate west of the rise was moving northwestward. Then the rise itself hit the s~bd~ction system oboist 30 million years (m.~.) ago. the western oceanic plate came into direct contact with ~ segment of the continental mar- Kin. and strike-slip m<,tion resmelted. The strike-slip bo~ndarv lengthened faith time. and its north limit mi- gr;Ated northwestward, as spreading ridge and trench col- lided along progressively more of the c<,ntine~ta1 margin. Tile manic lithosphere. `` hich had liven Updated shortly before the transition. `~as ·~ng and hilt. sit the change from .~bd~cti`'n to strike-slip regimes Erred `~hen the continental margin `~`I'; l~nderl.lin b~ sort. ~ eak mantle. The results `~t these changes and c`,n~liti<,ns elan be seen in s`~them California and north est \~.YiC() ( Figure 3.3). Tale Cretace`~s m`~gmatic ``rc (S it rra .N ex `~1~. Peninsr~l.~r Ranges and Baja C.~lit`~rni`~ t'~th`,lith~ii. `~ter-~`rc basin. 39 and Franciscan melange wedge are all present blat their distribution is scrambled. The simplest and ingest major disruptive structure is the San Andreas Fault. which is now the dominant break along which the western Pacific plate is moving northwest past the continent, carrying Baja California and coastal California along. The San Andreas offset of a well-documented 300 km and the concurrent opening of the Gulf of California hare oc- curred within the past 5 m.y. Lotion patterns during the transition period. between 30 m.y. and 5 m.y. ago, from subduction to strike-slip regimes were much more complex. In the southern Coast Ranges (northwest part of Figure 3.3), strike-slip faults underwent pre-San Andreas offsets of at least 200 km; yet these faults probably do not cut the Transverse Ranges, w hich trend across them in the south. In the Coast Ranges, the westward progression from magmatic arc to `'uter-arc basin to Franciscan melange wedge is present, variably truncated by Cenozoic faults; but then more outer-arc-basin strata were present beyond the Francis- can, in the Santa Ynez and Santa Monica Mountains of the western Transverse Ranges. The complete progression is present also in the Vizcaino Peninsula sector of Baja Cali- tomia. Along southern California south of the Transverse Ranges, however, the sequence is partly repeated: off- shore in the San Nicolas Island~anta Rosa Island sector, outer-arc basin strata are again present, out of place to the west of Franciscan; and Franciscan is present west of these strata' apparently in proper succession. One expla- nation possible for some of these and other relationships is that a western sliver of outer-arc basin and Franciscan terraces slid northwestward from Baja California. along a right-lateral strike-slip fault, around the Franciscan ter- rane now enclosed to the east; the leading edge of the sliver impinged on the continent, and the rest rolled clockwise past it, forming the western Transverse Ranges by rotation. The required rotation and northward transla- tion are consistent with the paleomagnetic orientations of lower Miocene basalts in the Santa Monica .\Iountains and islands to the west. Other styles of crystal disruption affected the western states further inland. Distributed crustal extension pro- d~ced the Basin and Range province (see Chapters 8 and 9) of blocks bounded by downward-flattening nodal tilts. The amount of total late Cenozoic extension is in dispute; my interpretation, more extreme than most. is that bet`` een 30 and o0 percent of the present width of the turbine represents extension. The NetWare of the transition from brittle fracturing and block rotations ot the tipper crust to hypothetical uniform extension in the deep. hid- den loner angst is also in dispute. Steadies in progress of deeply eroded basin-range terranes in southeast Calitor- nia and southern Arizona snuggest that the level of detor- mati`,n beneath the horizontal bases of rotated fault blocks is characterized by giant bodies or mullions, hating lengths <~l as milch as several tens of kilometers. and thus b` delonnation that is partly ductile and partly brittle. In

40 3so 30 WARREN H~!~{ILTON COAST RANGES 120° "S~ ~~\ ElASliV- RANGE 115° ~)PROV 1 N C£ I \aclo: 5: ! ~ N \~: He \ ~ \ ~ W~ ~: i ~ I, <o,S~-~ ~~ — AS \~! Up~rCr - Em- Eocene outer -arc - Bin stroto Tic - ~~s in Sorderlend shone __ O t00 200 300 tam I'd: if, ~ ·, i ~ ~^ <A ~ ,~,;~* G 0., ~ ~ ~ \ ~ ~ ~ \t ~ / COLORADO / PL AT E ~ U Little -d'5rupted proton em, r~toc~o\~ ^\ °~> - l 1 _~. ~` -_ ~ ~ __ \ ~00%~ a~e SON ~ ~ ~'t ;o~<,~ '^ ^~\ \ G :1\(: ~ ~.\ 6~A \ ` .: \ FIG~'RE 3.3 Selected tectonic elements of the southwestem t;nited States and northwestern \1exico.

Cvmplexitie.s of Modern and Patient Subduct:<'n System.s the northwest states. Mom northwest Nevada and south- en, Idaho northwestward' continental crust older than middle Tertiary was thinned by extension in some areas and completely rifled in others, the volumes has ing been filled subsequently mostly by middle and upper Ceno- zoic basalts (Figure 3.21. Other aspects of Cenozoic dis- ruption are discussed in Chapter 8. ~1 REFERE!.'CES Hamilton. lo. (1978). Mesozoic tectonics of the western ~ noted States. in Pacific Cubist Pale<~raphy Symp`'s2um A. Pacific Section. Society <'f Ec`'nt~mit Paler~nt<'l<>yists and Mineralogists. pp. 33~70 Hamilton At. t19~9). Tectonics of the Indonesian region C.S. Geol. Surr. Proof. Paper 10~-8, .~34.o pp.

Intracontinental Rifts anc3 Aulacogens 4 INTRODUCTION KEVIN BURKE State University of Neu, York at Albany The word rift is used in a variety of ways in geology, most of which hare nothing to do with major structures. Gregory (1921) was the first to use and later to popularize the word nEt for a large-scale structure when he described the East African rift. He defined a rift valley as a long depression between parallel normal &tilts. Here a modified version of Gregory's definition is used, and rifts are defined as "elongate depressions overlying places where the entire thickness ofthe lithosphere has ruptured in extension." This omits reference to parallel faults be- cause many tarniliar rifts, for e.,cample. the Connecticut rift, have a major fault on only one side, and in some rifts it is hard to be sure whether a boundary is a fault or a steep monoclinal flexure. By referring to "the entire thickness of the lithosphere" emphasis Is given to rifts as large-scale structures, and small depressions are excluded. The refer- ence to extension is to distinguish rifts from the other major fractures penetrating the whole lithosphere such as transform faults and those at convergent plate boundaries. Rifts are the commonest major lithospheric fractures because the strength ofthe lithosphere is least in tension. 42 Since rupture of the lithosphere is involved in rift forrna- tion, they are commonest where the lithosphere is thinnest, that is, in the oceans. This review is solely con- cemed with continental rifts that are now and probably always have been much less abundant than oceanic rids. Two great changes in continental-rifle studies over the last 10 years are the recognition that intracontinental rifts are numerous and that riRs have developed within con- tinents in a range of structural environments. These are all ens ironments ~ ith dominantly extensional tectonics, and active and ancient rift occurrences are associated with a number of different extensional plate-tectonic regimes. The Wilson cycle is the concept that permits application of plate-tectonic principles to the record of historical geology within the continents. R! FTS I ~N T H E \V I LSO N CYC LE Our discussion of continental structure and evolution comes 10~ years after J. Tuzo Wilson first drew attention to the best way in which to apply the revolutionary recogni-

Intracontinental Rifts and Aulacogens tion of the plate structure of the lithosphere to an under- standing of continental geology. Wilson ( 1968) argued, before the American Philosophical Society, that plate tec- tonics showed that on the surface of the earth, oceans are opening in some places and closing in others. The history of the earth's surface can therefore, he suggested, be con- sidered as a record of the opening and closing of oceans. Since ocean floor disappears by subduction, the record of these cycles has to be sought within the continents, and looking for records of these cycles is the most powerful way of interpreting continental geology. Wilson analyzed the cycles in terms of oceanic evolution from youth (continental rupture) to old age (continental collision) and elaborated his analysis in the second edition of his text- book (Jacobs et al., 1972). Dewey and Burke (1974) later suggested that complex interwoven cycles of ocean open- ing and closing of the kind recorded within the continents be called Wilson cycles (Figure 4.1). RIFTS AT CONTINE.N'TAL BREAKE'P Rifts occur at all stages ofthe Wilson cycle, because exten- sional tectonics can develop at all stages, but those asso- ciated with continental rupture are both particularly well developed and particularly accessible to study. The East African riPr system is the best known continental rift system, and progress in its understanding has accelerated since appreciation that its present activity dates only from the beginning of the Neogene when the African plate appears to have come to rest with respect to the uncler- lying mantle convection pattern (Burke and Wilson, 19721. Distinction of the current episode of riding from a very similar episode in Africa 100 200 million years (m.y.) ago that was associated with the breakup of Condwana has emerged. BiPt studies that define the timing of events accurately have become of particular importance in recent years and seem likely to remain important in the fixture because the establishment of how rifting relates to other tectonic events will only be possible if the relative timing `~t the phenomena are accurately known. T``o other rift properties that have been tingly estab- fished in the East Atrican rift system are that rid faults commonly follo`v and reactivate old stn~ct~res (NIc- Connell, 19~) and that rift igneous rocks are almost wholly mantle derived. Detailed studies of igneous rocks AK _ AV]K o 15 `4C ~enzlt CItO R,tt 43 associated with the East African rift system have empha- sized the widespread occurrence c~falkaline rocks, most of which appear to have last equilibrated with the mantle at depths of 60 100 km, and many of which appear to have suffered complex subsequent histories. Comparable al- kaline rocks occur in many rift systems, but so do tholeiitic rocks, suggestive of equilibration at much higher mantle levels. Studies of sediments in the East African rift system are numerous but unsystematic. They come Mom such di- verse fields as petroleum exploration and fossil hominid research, but there has as yet been little attempt to make tectonic inferences from these sediment studies or to study the riR-f~lling process as a whole. On the other hand, ~ olcanology and geomorphology have been in- tensely pursued in East Africa and tectonic inferences made from them. Teleseismic studies have been among the most fruitful of geophysical investigations in the East African rid system and have shown the peculiar nature of the mantle below the rift system as well as the extensional character of most riR earthquakes. Although local networks hare yielded important results, the limited number of modem seismic stations on the African continent has prosed somewhat of a barrier to research. Gravity studies are well advanced in East Africa and have proved most susceptible to tectonic analysis in which gravity information has been coupled with other data. Seismic refraction studies, for example, have indicated the presence of high-`elocity dike-like objects in some rifts, and this has helped to con- strain gravity interpretations (Figure 4.2~. The East African rift system has been, since its discov- e~y, one of the most studied of continental rift systems, and it is likely to continue to yield important insights in the understanding of rids. Further seismic and riR- sediment studies appear particularly promising. If the East African system continues to propagate and en entually becomes a system extending across the African continent, then a new ocean can develop and the African plate may split in two. If this happens, many currently .~cti`e rids ``ill tie OR in one fir the tither of the two new continents. Such rids will hare failed to develop into ocean. Failed rifts of this kind stretching into continents at a variety of angles Tom Atlantic-t~pe ocean margins are numerous and are rapidly becoming the best known ol all h'ssi! lifts (Burke. 19,6~..~nd Figure 4.3). This is t'ec.`u-e 3CW—~—C _~CA0O 28 >2S ,2 t2 WOllCs~on N- - cross Ke~eenC - On '18 ~cwceniCwon Suture t 7 5utu,zs R.~' Rcr~o~con Q,t? - ;~t \ suture or(nec&,c~? Bit' ~ 2 ~ "S to ~ 3 R.tt; AopalacM.cn Su~urcs;, ~ _ 1 2 ' 8 C ooDcrm~nc ~Voomcy Rlvcr R:tt S~turc 2: ~ ', At~CO~SCOW ~ - 1~- IVY ; -/ ., , Pc~OKEAN 5rcnv~t1c -PENVILLE ~P~4LACk4AN ?R0V'NCE -ror,'t ~RCVINC: Now York- ~QCV!~- l'BCCk] /";/ ' '8EP.R ~ " ~ ' SLAVE -UNCOIL' S~P~1OR PQOvirMCE at ME BEAR BRIM FIGL RE 4.l Sim~)litied cross section across North .Alneric~l tram the Burt Sea to George s Bank showing numerous rifts and stlt~lre!i :,rod~l~rd through ol)eration 1,t8the Wilson cute over the last ~ x 10~ Fears. Embers are approximate ages in years times 105.

44 FIGS RE 4.2 Sketch (by A. M. C. Sengor) of structure of typical, active, intracontinental no showing thinned lithosphere, an axial dike emplaced at a shallow level, listric faults. volcanic rocks, and sediments. the sediment fill of the failed rifts is oil-bearing in some places. Oil is produced from three major ens ironments in rifts ofthis class: nonmarine graben facies at the bottom of the rifts, marine limestones formed during rapid subsi- dence of the rifts a few tens of millions of years after continental rupture, and younger elastic material progracI- ing off the continent along or across the rift. The last environment can be subdivided into shallow, often deltaic, and deeper water subenvironments. Reflection seismic studies and drill holes associated with petroleum exploration can yield accurate infonnation on how rifts subside over intervals oftens of millions of years. Charac- terization of individual rids, or parts of rifts, as behaving thermally like continents, oceans, or in some intermediate state may prove possible. The widespread occurrence of failed rifles trending at varied angles to Atlantic continental margins may provide some explanation of the difficulty that has faced attempts in many areas to locate precisely the boundary between continent and ocean (Figure 4.4~. Although the general behavior of failed riRs at Atlantic margins seems to follow a regular pattern, there are excep- tions and much remains to be learned about this economi- cally important class of rifts. For example, the South Atlantic opened about 125 m.y. ago, and many failed ridges associated with this event have passed through normal sequences of events becoming occupied by marine sedi- ments in the Aptian about 15 m.y. later. However, the Rio Salado and Colorado rifts (L'rrien and Zambrano, 1973) continued to be filled with nonmarine sediments until much later (about 80 m.y. ago), and the Benue Trough in Nigeria behaved like a small ocean and appears to have opened and shut in a small-scale Wilson cycle of its own over the integral between 195 m.y. and 80 m.y. ago (Burke and Dewey, 1974~. A small population of rifts appears to has e developed at Atlantic-type ocean margins some time after initial con- KEVIN BE'REE tinental rupture. The Sirte rift in Libya. formed during the Lower Cretaceous facing a Tethyan margin that ruptured 100 m.y. earlier in Triassic time, is the best documented example. it appears to owe its origin to intraplate strains established during the events that led to opening of the South Atlantic. This is again a case for which accurate timing of structural events is essential in understanding. The Sirte rid is one of many rids without reported contem- porary volcanism, which. was at one time widely thought to be an essential feature of riPc-valley development. The rapid subsidence rates revealed by studies of rid sedi- k~ ''at 6D51 _.~ em' ~_%~5 __ r d \ ok, hi: . 1 ..~ .: FIGURE 4.3 Map of a closed Atlantic Ocean showing areas ( in black) in which rift-valley lithosphere, formed mainly in associa- tion with continental rupture, is well developed.

lntracontinental Rifts and Aulacogens A CONTINENT OC EAN BOUNDARY CON=NENSA~ WATER LITH=PHERE ~ ` 68ndi_— OCEANIc LITHOSPHERE B BUDDY NOVA SC="SHELF RIFT-VALLEY OCEANIC LITHC6PHERE LITHOSPHERE FIGURE 4.4 Section view comparing an idealized continent- ocean boundary (A) with an interpretation of the boundary off `Nova Scotia with net valleys and transitional lithosphere in a zone about 500 km wide between continent and ocean (B). meets (rates as high as 0.5 km per m.y. are reported by BRGM et al., 1975) indicate that fast contemporary cooling rates existed beneath the rifts. Whether igneous rocks are actually erupted at the surface during rift development probably depends on a variety of factors of less than fun- damental importance. CONVERGENT BOUNDARY RIFTS RiR developments at Andean continental margins and in island arcs are most prominent at the place where the lithosphere is thinnest, that is, along the line of the vol- canoes themselves. These rids, known in such active An- dean arcs as New Zealand and Sumatra-lava, achieve their greatest tectonic significance only when the vol- canic arc splits to form a marginal basin as, for example, happened in the Japan Sea at the beginning of the Mio- cene (Sillitoe, 1977~. Rifts of Basin and Range type appear associated with imperfect transforms motion along con- tinental boundaries, but few fossil examples have, as yet, been recognized. COLLISIONAL RIFTS Perhaps the greatest variety of rif;r development is asso- ciated with continental collisions. Not only do the failed rift systems formed at rupture become reactivated as rifts striking into the collisional mountain belts forming the population of objects that Shatski (1947) called aulaco- gens, but a whole new set of rids, of which the Lalce Baikal and Rhine rifts seem the best active examples, are set up as a result of ineracontinental strains established during the collision. These objects have been called impacto- 45 gens, and they can be readily distinguished by their geo- logical history Tom aulacogens. Besides these two major classes of collision-related rifts, final closing of an ocean is accompanied by complex riR production. This is well illustrated in the .Mediter- ranean, where formation of the rifts of Corsica and Sardinia 20 m.y. ago accompanied rotation of these islands away from France (Dewey et al., 1973) and where forrna- tion of the Aegean rifts within the last 20 m.y. has been related to the westward motion of Turkey with respect to the Black Sea and the Mediterranean (Dewey and jengor, in press). If the closure of the Mediterranean continues, rifts of the Aegean and Corsican-Sardinian types are likely to be compressed out of recognition, and such struc- tures are probably preserved only as obscure areas within mountain belts that show signs of late-stage extension before final compression. For this reason, rifts of this kind are of less general interest in continental geological his- tory than the four major rid classes: continental rupture rifts, failed rifts at Atlantic margins, aulacogens, and im- pactogens. A generalization of this type always has excel tions, and the Devonian rift basins of Nonvay (Horn, Solur~d, and Hornelen) form such an exception (Steel, 1976~. They appear to have fanned within the Caledo- nides at continental collision and closely resemble the Thakkola rift of the high Himalaya today. AUL^C OGE N S Although the aulacogen concept is over 30 years old, widespread use ofthe term among scientists in the United States dates only from attempts to interpret the aulacogen concept in plate-tectonic terms (Burke, 1977a). Shatski used the word only for rifts striking into fold belts, and it seems unwise to use it for all ancient rifts that have failed to develop into oceans. Shatski also emphasized that the depositional and structural history of the aulacogen begins at the same time as that of the fold belt into which it leads, thus eliminating the structures now being called impactogens. An unfortunate fact in the history of the use of the word aulacogen is that one of Shatski's type exam- ples, the Dneiper-Donetz structure, is a failed rift that strikes not into a fold belt but into the North Caspian depression, an area probably underlain by ocean floor formed in the Devonian that has been covered with a great thickness ~—14 km) of sediments but has, so far, escaped both subduction and abduction. Although aulacogen histories can be diverse, a twofold division can often be made into an earlier continental rupture and ocean-opening phase with typical rift fea- tures, including alkaline igneous rocks, basal coarse clas- tics, and stratigraphically higher evaporites and a distinct postcollisional phase with coarse elastic sediments de- rived from the collisional mountain belt and str~ke-slip faulting. Structurally, the outstanding features of many aulacogens are the enonnous thicknesses of sediments and volcanics that they contain. The best-clocumeIlted thickness is that of the southern Oklahoma aulacogen.

46 where more than 13 km of rock accumulated during the Paleozoic (Ham and Wilson, 1967), roughly two thirds being deposited in the first rifting event and the re- mainder in association with collision. Rift reactivation is a widespread phenomenon, and the deposition of a thick sequence at collision over the initial rift sequence, as happened in Oklahoma, indicates that the first event mod- ified the local lithosphere in a way that made it behave abnormally even hundreds of millions of years later. Shatski pointed out that the upper member of an aulaco- gen sediment pair commonly extended over an area wider than that of the initial rift, and this seems a particular case of the more general observation that numerous major in- tracontinental basins, such as the Chad, Paris, and ~Michi- gan basins, overlie older rifts (Burke, 1976b). IMPACTOGENS In map view an impactogen resembles an aulacogen as it is also a rift structure striking into a fold belt. Distinction between them starts from ascertaining whether rift history goes back as far as the opening of the ocean the closure of which is represented in the fold belt. If it does not, and riding dates only from the time of ocean closure in the fold belt, the rift is an impactogen. For example, both the Rhine graben and the Polish trough strike from northern Europe into the Alpine fold belt. The start of the geo- logical history of the former is contemporary with the mid-Eocene Nleso-Alpine collisional event (5engoret al., 1978), while that of the latter started in the Triassic. Thus the Rhine graben is an impactogen, while the Polish trough is an aulacogen. Although the Rhine graben reaches to the edge of the Alpine system in the fura, impactogenal rifts associated with the Himalayan collision, the Baikal and Shansi graben develop far from the main collision zone. Molnar and Tapponnier (1975) suggested that these rifts formed as a consequence of the eastward escape of China Tom crushing between the viselike jaws of India and Asia. The Baikal and Shansi rifts are linked to the main collisional mountains in Tibet and the Himalaya through a system of strike-slip faults and folds. Timing of the Himalayan collision is not yet as well established as that of the main Alpine collision so that it is hard to know how closely the start of faulting in Shansi and Baikal coincides with the collision marked by the Indus suture. Impactogens seem particularly well developed in asso- ciation with the Hercynian fold belt of Europe. The Her- cynian Ocean appears to have opened in two stages: one in medial Devonian times along a riR extending from west- enn Britain to the Urals—the timing ofthis event indicates that the rift may hare been formed as an impactogen related to collision in the Caledonides and the other in fairly CarI~oniferous time!) along a riR with a very differ- ent strike in Iberia. This second riR could have been an Acadian impactogen (Burke and Sawkins, 1977~. Although KEVIN BURKE evidence of the formation of He Hercynian Ocean is in- complete. evidence of its closure at the beginning of the Permian is much stronger. Rifts developed over much of northwestern Europe at the beginning ofthe Permian (the Oslo graben being the best known such rift), and these rifts are strong candidates for interpretation as impacto- gens. The rifts of northwest Europe established In this Permian event have been repeatedly reactivated in Triassic, Jurassic, Cretaceous, and Cenozoic times (Whiteman et al., 1974~. This is most obvious in the North Sea, where geological development of the huge oil prov- ince has been dominated by rift reactivation (Woodland 1975~. A possible explanation of renewed subsidence in a long inactive rift, so typical of the North Sea, is that a thermal pulse in the mantle induces transition from a basaltic to an eclogitic mineralogy in a subrift axial dike system. NORTH AMERICAN IN1'RACONTINENTAL RIFTS AND AULACOGENS ACT I VE RI FT S Active rifts in North America are dominated by those of the Basin and Range, discussed in Chapters 8 and 9, but the Rio Grande rift is a structure whose development al- though closely linked to that of the Basin and Range is sufficiently isolated to be considered on its own (Figure 4.5~. Recent studies are showing that the Rio Grande is an intracontinental riPc with many typical features perhaps including an asocial dike system. Because volcanic rocks with a wide range of ages are relatively abundant in the riR and because sediment-filled basins with determined stratigraphy are developed at various places along the 1000 km length of the system, an unusually good oppor- tunity exists in the Rio Grancle no for establishing a de- velopmental history through much of Neogene time, and this may help to elucidate the way in which a nit system can develop through a few million years for comparison with the only other well-clocumented history, that of the 4~m.y.~1d Rhine rift. North America's major active rid, the Rio Grande, is clearly a prime target for many types of geological and geophysical studies that will help to improve understand- ing of its deep structure and that of continental rifts in general. REACTIVATED RIFTS A number of ancient rift systems are at present showing signs of reactivation. This reactivation is most obvious in seismic activity (e.g., in the Reelfoot rift) or in seismicity and the occurrence of youthful faulting [e.g., the Lake George-Lake Champlain rift (Burke, 1977b; Isachsen et al., 1978~. Studies in these systems may help to illu- minate the mechanisms by which intraplate stresses are

Intracontinental Rifts and Aulacogens / SNAKE PIER DIN ) 1RANGE' PRINCE ;1 R10 ^'=N~ 3_~: ;~GRENLANO GRANDS FlJH0Y~ s~d CONN.,~' NEWARK Id, ' GUATE - : I YALLAHS FIGURE 4.5 Major Mesozoic and Cenozoic rifts of North Amer- ica. Most are associated with the development of Atlantic-type ocean margins (based on Burke et al., 1978). relieved on pre-existing structures and because the oc- currence of earthquakes in these isolated tectonic areas may constitute environmental hazards (Sykes, in press). FAILEI) BIFTS AT ATLANTIC-TYPE MARGINS An Atlantic-type margin extends from the Yucatan Penin- sula around the Gulf of Mexico, along the eastern sea- board of the United States and Canada to Baffln Bay and continues past the Arctic Islands of Canada to the norm slope of Alaska. The northern margin was formed in a rupture ev,ent of roughly medial Jurassic age, and the At- lantic and Gulf of Mexico boundaries in an event of Late Triassic age. The Labrador Sea-Baff~n Bay boundary dates from Late Cretaceous time. Failed rift systems asso- ciated with these various events show great diversity, and comparative studies are likely to be fiuitfi~1 because dif- ferent rifts show different features. Thus the McKenzie River delta is the best example in North America of a great river flowing down a failed rift to build a prograding delta over ocean floor (the Mississippi, although an example of this phenomenon, occupies a more complex structural environment). The Labrador and Baff~n failed rift systems are largely offshore, and knowledge of them depends mainly on pe- 47 troleum exploration, although the rift margin of Disko Island has received attention especially for its unusual metallic iron-bearing basalts. The best-exposed failed rifts at the Atlantic margin are the Triassic rifts of the Appalachian region. Detailed structural studies of some areas are completed, and rip sedimentation is being investigated. The opportunities for an improved understanding of rift structure and devel- opment are great. A peculiarity is the lateness of a major (tholeiitic) basaltic episode in these rifts which tools place when rifle activity had progressed for about 20 m.y. and was apparently almost contemporary with the beginning of Atlantic ocean-floor formation. Failed riRs of Late Triassic age around the Gulf of Mexico are deeply buried under the salt deposited soon after ocean opening in that area and few research results have, as yet, been published on these relatively inaccessi- ble features. AULACO GENS The early development of aulacogens associated with the opening ofthe lapetus Ocean along the site ofthe Appala- chians in late Precambrian times has been considered by Rankin (1976), and Mere is clearly an opportunity in many of these structures for integrated stratigraphic, structural, petrological, gravity, magnetic, reflection, refraction, and teleseismic studies (Figure 4.6~. The progress of research has been uneven in riRs of this age. For example, the southern Oklahoma aulacogen is structurally and strati- graphically weI1 known (it contains over 75,000 oil and gas wells), but Me petrology of igneous rocks at its base is poorly known and its deep structure has not been ana- Iyzed geophysically. Proterozoic nits that are widely in- terpreted as aulacogens are known from various places along the western Cordillera between southern California and Alaska. The Amaragosa feature in Death Valley (Wright et al., 1974) is perhaps the best described. Em- phasis in this review has been on the importance of understanding the relative timing of events in riR studies. In the Precambrian the poor resolution of most dating is a serious drawback in riPc studies, and many ambiguities are at present irresolvable. The Keweenawan rid system provides a good example. The Keweenawan rift is linked westward to the feature causing the midcontinent gravity and magnetic anoma- lies, which are of typical rift type (Halls, 1978~. At its eastern encl, the Keweenawan rift can be linked to the riPc known from gravity, magnetics, and drilling to underlie the lower peninsula of Michigan. Farther south this fea- ture becomes diffuse within the Grenville province and may have been damaged in the Tibetan-style events (Dewey and Burke, 1973) accompanying the Grenville collision. It seems possible that the Keweenawan rip system was formed in association with the opening of an ocean that closed in formation of the Grenville province, and occurrence of the Seal Lake and Gardar rids to the

48 ~ ' \ =~ ~ WHO ( 0> ~ Ultra ~ ~fl ~ ,~ B6rHURST~ J ATHAPUSCOW ~ ~ ~ RICH~ CARDER <:,GUr ~ COWMEN / MOLT :~ ~S.OKLA~ PENS. <~& `` ~ ~ REELf=; \ MA 'it ~ 0 1000 KM FIGURE 4.6 Major Proterozoic and Paleozoic rifts of North Amenca. Stippled areas are riDcs fanned et the end ofthe Protero" zoic in association with opening of the Iapetus Ocean, most of which became aulacogens during the Paleozoic (based on Burke et al., 1978). northeast of the Keweenawan system is perhaps support- ing evidence. However, resolution of these varied Pre- cambrian events in time is too poor to warrant detailed attempts at historical analysis, and the Keweenawan could, on present evidence, be an ~mpactogenal system fanned at the Grenville collision. In spite of this limitation, the Keweenawan riR and its subsurface extensions are features of peculiar interest. Their width is noticeably greater than that of most rifts that fail before the ocean-opening stage, and a great variety of volcanic, intrusive, and sedimentary rocks occupy these wide structures. Analysis of relationships within and between these units by as many geological and geophysical techniques as possible is likely to be very rewarding. Geochemistry of Keweenawan igneous rocks is already thoroughly investigated, and no doubt further refinements are possible. The oldest well-described aulacogens in North America are those of the Canadian Shield, especially the Athapu- scow structure of Great Slave Lake (HofEnan, 1973), which is now being recognized to show strong sibs of convergence and is perhaps a Proterozoic Benue trough. Less well known are rifts striking into the circum-Ungava collisional fold belt near Richmond Gulf and Cambrien KEVIN BURKE Lake, but limited evidence indicates that these are aula- cogens nether than impactogens. With recognition that the Penokean orogeny embodies collision of continental objects and shows reactivation to the south rather than to the norm, features of the southern Canadian Shield (perhaps including the Huronian) may prove to have been deposited in aulacogens. CONCLUSIONS Active intracontinental rifles and riRs with ages extending back as far as Early Proterozoic occur in North America, and study of them should help to show how the conti- nental lithosphere ruptures in extension and how, once ruptured, the lithosphere continues to develop in places where oceans fail to form. Relating these numerous rids to the various stages of the Wilson cycle appears feasible, and analysis in these terms seems the most potent way of applying the revolutionary understanding of plate tec- tonics to the tectonics of the continents. Rifles are so varied that almost all display some features that are absent or poorly developed in others, so that what is worm studying varies greatly Mom no to nR. However, in most places an integrated approach to rifle studies in- volving thorough Slid or subsurface geological assess- ment followed by application of selected appropriate geo- physical and geochemical techniques appears likely to be most rewarding. RE FERENCES BElGM, ELF, Ed, SHPA (197S). Prolog du basin d Aquiline (Atlas with 24 plates), Paris. Burke, K. (1976a). Development of glen associated with the initial ruptures of the Atlantic Ocean, Tectonophysics 36, 9~112. Burke, IC. (1976b). The Chad Basin: an active intracontinenta1 basin, Tectonophysics 36, 197-206. Burke, K. (1977a). Aulacogens and continental breakup, Ann. Rev. Earth Planet. Sci. 5, 371~96. Burke, K. (1977b). Are Lakes George and Champlain in Neogene grabens reactivating early Paleozoic nits?, Geol. Soc. Am. Abate. Programs 9, 247. Burke, K, and J. F. Dewey (1974). Two plates in Africa during the Cretaceous?, Nature 249, 313 316. Burke, K., and F. J. Sawkins (1977). Were the Rammelsberg, Meggen, Rio Tinto and related ore deposits fanned in a Devm nian rifting event?, Geol. Soc. Am. Abstr. Programs 9, 916~317. Burke, K., ant J. T. Wilson (1972). Is the African Plate stationary?. Nature 239, 387~90. Burke, K., L. Delano, J. F. Dewey, A. Edelstein, W. S. F. Kidd, K. D. Nelson, A. M. C. jengor, and J. Stroup (1978). RiRs and Sutures of the World. Unpubl. Rep. to Geophysics Branch, ESA Div. NASA/GSFC, Greenbelt, Md. Dewey, J. F., and K. Burke ( 1973). Tibetan, Variscan, and Precambrian basement reactivation: products of continental collision,J. Geol. 81, 68~92.

Intracontinental Rifts and Aulacogens Dewey, I. F., and K. Burlce (1974). Hot spots and continental breakup: some implications for collisional orogeny, Geology 2, 57 60. Dewey. J. F., and A. A. C. Sengor (in press). Aegean and sur- rounding regions: Geol. Soc. Am. Bull. Dewey, J. F., W. C. Pitman, W B. F. Ryan, and J. Bonnin (1973). Plate tectonics and the evolution of the Alpine system, Geol. Soc. Am. Bull. 84, 3137~180. Gregory, J. W. (1921). The Rift Valleys and Geology of East Africa, .\Iacmillan, New York, 479 pp. Halls, H. C. (1978). The late Precambrian geology of the central North American rid system, in Tectonics and Geophysics of Continental Rifts, I. B. Bamberg and E. R. Neumann, eds., D. Reidel,.~^To, Dordrecht, Holland, pp. 111-121. Ham, W. E., and J. L. Wilson (1967). Paleozoic epeirogeny and orogeny in the central United States, Am. 1. Sci. 265, 332~07. Hoffman' P. F. (1973). Evolution of an early Proterozoic conti- nenta1 margin, Phil. Trans. Roy. Soc. London A273, 547~81. Isachsen, Y. W., E. P. Geraghty, and S. F. Wright ( 1978). Investi- gation of Holocene deformation in the Adirondacks Mountain dome, Geol. Soc. Am. Abstr. Programs 10, 49. Jacobs, J., R. D. Russell, and J. T. Wilson (1972). Physics and Geology, 2nd ea., .McGraw-Hill, New York, 622 pp. McConnell, R. B. (1974). Evolution oftaphrogenic lineaments in continental platfonns, Geol. Rundsch. 63, 389~30. Molnar, P., and P. Tapponnier ( 1975). Cenozoic tectonics of Asia: effects of a continental collision. Science 189 419~26. Rankin, D. W. (1976). Appalachian salients and recesses: late Precambrian continental breakup and the opening of the Iapetus Ocean,J. Ceophys. Res. 81. 5605 5619. Sengor, A. A. C., K. Burke, and J. F. Dewey (1978). Rifts at high 49 angles to erogenic belts: tests for their origin and the tipper Rhine graben as an example. Am. J. Sci. 278, 2~0. Shatski, N. S. ( 1947). Structural correlations of platforms and geo- synclinal folded regions, SSFt Akad. Nauk Len. Geol. Ser. a, _ _^ Ado. Sillitoe, R. D. (1977). .Uetallogeny of an Andean continental margin in South Korea: implications for opening of the Japan Sea, in Island Arcs, Deep Sea Trenches and Back-Arc Basins, M. Talwani and W. C. Pitman, eds., Staunch Ewing Series 1, Am. Geophys. Union, Washington, D.C., pp. 303~310. Steel, R. J. ( 1976). Devonian basins of western Norway- sedimenta~y response to tectonism and to varying tectonic con- text, Tectono physics 36, 207-2~4. Sykes, L. R. (in press). Intra-plate seismicity, reactivation of pre- existing zones of weakness, alkaline magmatism and other tec- tonism post-dating continental Eagrnentation, Rec. Ceophys. Space Phys. L'rrien, C. At., andJ. Zambrano (1973). The geology ofthe basins of the Argentine continental margin, in The Ocean Basins and Their Margins, A. E. M. Nairn and F. G. Stehli, eds., Plenum, New York, pp. 13~166. Whiteman, A. J., G. Bees, D. Naylor, and R. M. Pegrum (1974). North Sea troughs, Tectono physics 26, 39 54. Wilson, J. T. (1968). Static or mobile earth, the current scientific revolution, in Gondwanaland revisited, Am. Philos. Soc. Proc. 112, 309 320. Woodland, A. W., ed. (1975). Petroleum and the Continental Shelf of North-West Europe, Halsted-Wiley, London, 501 pp. Wright, L. A., B. W. Troxel, E. G. Williams, M. T. Roberts, and P. E. Dichl (1974). Precambrian sedimentary units ofthe Death Valley Region, in Guidebook, Death Valley Region, Death Valley Publishing Co., Shoshone, Casio, pp. 27~5.

Evolution of Outer Highs on Divergent Continental Margins INTRODUCTION MARTIN A. SCHUE PBACH and PETER R. VAIL Exxon Production Research Company The development and use of multichannel seismic profil- ing and deep-drilling technology have led to an increased understanding of the earth's crust. This paper discusses the application of these developments to the nature and evolution of the boundary between continental and oceanic crusts at divergent (passive) margins. Seismic profiles across divergent continental margins and regional geological studies indicate the occurrence of highs near the boundary that probably fortned during the late rift phase of divergent margin evolution. Our seismic and regional geological studies have provided us with suffi- cient data to develop a model for the evolution of these outer highs. In order to test the mode} on the evolution of outer highs on divergent margins, extensive geophysical surveys, emphasizing regional multichannel seismic pro- files, need to be undertaken to investigate the tectonic and stratigraphic variations visible on geophysical data. Near-surface data from outcrops, piston cores, and dredge samples can provide additional important constraints for more accurate geophysical interpretations. From the com- 50 bination of these geophysical and geological data, opti- mum drill sites can be located to test the geophysical interpretations and to obtain samples and logs of the eatth's crust and overlying sedimentary section. The results of such studies on the outer portions of divergent margins should provide us with a knowledge of (1) the character, age, and position of the continental- oceanic crystal boundary for better plate reconstructions; (2) the existence, nature, and evolution of the outer high for a better understanding of mantle processes active dur- ing the formation of a divergent margin; (3) the nature of the overthickened oceanic crust commonly present in the area adjacent to the oceanic~ontinental crustal boundary, for better understanding the role of volcanism in seafloor spreading; and (4) the type and controls of late-rift-phase sedimentation, for determining the reasons for the wide- spread occurrence of salt and organic-rich shales during this phase of seafloor spreading. A knowledge of the above factors in a number of different areas would provide the needed information to understand accurately the nature and evolution ofthe continental~ceanic boundary on divergent margins.

Evolution of Outer Highs on Divergent Continental .\Iargins STRUCTURAL-STRATIGRAPH IC RELATIONS AT DIVERGENT CONTIN E.NT.\L-OCE.ANIC CRUSTAL BOUND.\RI ES GEOLOGICAL OBSERVATIONS The outer high is defined as the relatively high area be- tween the oldest oceanic crust and the onlapping late rift sediments deposited on adjacent continental crust. Based on a number of geological and seismic profiles across divergent continental margins, the outer high seems to be a common feature formed during the late-ricing phase of pull-apart before the onset of"passive" driving. In many cases, the outer high does not exist as a present-day high, because of subsidence and tilting during the driR phase. The widespread occurrence of outer highs leads to the assumption that they occur along most pull-apart diver- gent margins and thus are a part of their tectonic evolr~- tion. Divergent margins that evolved through shearing, such as the Ivory Coast, Africa, do not have similar appear- ing outer highs. Figure 5.1 shows the location of regions studied for this report. Observations of cross sections across divergent margins indicate that outer highs are variable in size, elevation. and amount of faulting. Figure 5.2 shows three diagram- matic examples that illustrate some of the variations of ouster highs. They range from a very-low-relief outer high [Figure 5.2(a)] that is only recognizable by subtle onlap on 51 the landward side and the presence of oceanic crust on the seaward side to a large high [Figure o.~(c)] that is easily recognizable by the large tilted fault blocks. Outer highs result from broad implies or arching along zones 10 to 20 km wide in areas having earlier horst-and- graben (and graben-fill) topography. We believe that the outer highs are more than single horsts formed during the early-riR phase, for three reasons: (1) Present-day aborted rifts do not have throughgoing horsts in their centers, but outer highs are continuous. (2) If the outer high were an initial horst formed during the early-rift stage, continental separation would most likely take place along one side of the initial horst, thus leaving an outer high on only one margin; however, we observe outer highs on both mar- gins. (3) The top of the outer high is higher than the top of (other) horst blocks at the end of the late-riR phase, as evidenced by the onlapping or pinching out of late-riR- phase sedimentary rocks. In most examples studied where extensive extrusive volcanics are lacking, the top of the outer high is approxi- mately at the same level as the top of the oceanic crust. This indicates that the oldest oceanic crust was formed near sea level, since the shallow marine late-ri~c-phase sediments typically onlap the continental outer high. Venters (1977) came to a similar conclusion. In many areas, the outer highs are complicated by ex- tens~ve volcanic extrusives and intrusions. The extensive ~olcanics may overlie the outer high, the adjacent oceanic crust, or both. The distribution of volcanic extrusives is _\ , hi'' ~ .~°: ~0 I'm. Cot a · RIFTS FIGL BE .~.1 .\re.~s con,;idered in this ins ~ stig.ltio~l. l a G - DIVERGENT MARGINS

52 FIGURE 5.2 Diagrammatic exam- ples of outer highs on divergent conti- nental margins. very irregular. For example, areas with extensive volt canics, such as the Faeroe Island area (Talwani and Eldholm, 1977) and eastern Greenland (Hailer, 1971) con- trast with areas having virtually no extrusive volcanics in the area of the outer high, such as the west Africa exam- ples discussed later in this chapter. The distribution of volcanics in presently rifts is similarly irregular. For example, the Rhine Graben has two large volcanic centers, Vogelsberg and Kaisers~hl, separated by areas of limited volcanic activity (Rhine Graben Research Group for Explosion Seismology, 1974~. In general, divergent margins have a characteristic pro- gression of sediment types that vary as an area progres- sively undergoes rifting and then driRing. Rifts that form in areas of thick continental crust, such as cratonic shields and areas that have undergone an erogenic event, are in most cases first filled with nonmarine sediments that grade upward to shallow marine sediments that com- monly include evapontes, as shown in the West Africa example discussed later in this chapter. RiRs that form In areas of thinned continental crust, such as older inactive rids, commonly contain marine or interbedcled marine and coastal deposits. Figure 5.3 Tom offshore east Green- land (Hailer, 1971) is an example of divergent margin evolution in which the crust has been thinned by earlier riding. Post-pull-apart or driR sediments are usually marine, except where large prograding delta complexes are pres- ent. In many areas, the marine sediments of the drift phase are prograding carbonate banks and reefs. Marine temgenous elastics commonly overlie the carbonates or extend throughout the entire driR phase, where car- bonates are absent. GEOPHYSICAL EVIDENCE FOR OUTER HIGHS Outer highs are best recognized on reflection seismic data by looking for the relatively high area in between the MARTIN A- SCHUEPBACH and PETER R. VAIL _ - _ ~ - 83 U4TE Rl~ SEDIMENTS - — _ ~ il ~NLAPPNG 0N10 DRIFT SEDIU0iig— ~ __== , " n ~ Elton u=u I ~ OCEANIC . ' CRUST _ / \ '\ ~ . / ! 0~ 8~ - SALT\ 7D ~ 1 ~ seems ted 'I LATE RlfT SEDIIUENTS / , I, _: ~6 ONTO oUrER ~ 5 ~ ()—it —ORFT SEDIMENTS MINES ___ ~ ~ TWO WAY O ;Alri /~ /oUT~8 HIGH _ ~ _ /, /, / ' / - , ' t, Il CRUST' / / \ / \, i\ `// ~ \ _ ~ 1 - ,, seismic expression of oceanic crust and seaward onlap of the reflectors over the predrift unconformity. The reflec- tor at the top of the outer high may be continuous or highly faulted. In addition, reflectors and faults may or may not be observable below the reflector at the top of the outer high. The diagrams shown in Figure 5.2 are patterned after seismic examples and thus illustrate typical reflec- tion patterns of outer highs on divergent margins. An example ofthe seismic expression of oceanic crust is shown in Figure 5.4. Note the high amplitude with nu- merous diffractions that characterize this reflection. In this example, no good reflectors occur below it. This is the common case. In other examples, however, dipping re- flectors occur below the top of the oceanic crust. Areas where these reflectors occur ale commonly paleotopo- graphic highs, which may indicate local thickening of oceanic crust by volcanic extrusives. Two multichannel seismic lines hum offshore West Af- rica (A, Figure 5.5, and B. Figure 5.7) are presented to illustrate the seismic expression of outer highs. Both lines extend from the present-day shelfto near the continental- oceanic crustal boundary. This paper only discusses the interpretations in the vicinity of the outer highs. A de- tailed discussion of seismic line A (Figures 5.5 and 5.6) is presented in Vail et al. (in press). A dashed line indicates the interpreted boundary between rift and drift reflectors' and a heavy solid line indicates the top of the outer highs On both seismic lines the interpreted presalt early-riD reflectors are bordered on the seaward side by outer highs consisting ofblocks bounded by steep faults. Seismic line A (Figure 5.5) shows a relatively smooth surface at the top of the outer high with underlying faults and reflectors Seismic line B (Figure S.7) shows large tilted fault blocks Late-rift salt reflectors onlap or pinch out against the outer highs. Onlap patterns on seismic line A indicate that the paleotopographic high point of the outer high was on its eastern side and the top of the interpreted oceanic crust actually was higher than the western portion of the outer

o C, .— 4— {: en · - ~ ~ · - - a: ca · - ~ .—,4 Q ~ ~ ._ ~ En ~ Cal ._ ~ m ~ ~ _ _ _ ._ O m~ _ ~ _— was 3wOSdO 9NIN30lM 1~101 - ~ _ _ _ J - z cr 1 ~ 1 ~ 1 L L: 1 a I 2 <_— ___ - Ct: Z ~ ~ ~ ~ ·` C) Il~ Z to ' ~ cr _=

54 .UARTIN A. SCHUEPBACH and PETER R. VAIL Am_ ~ _, OCEANIC CRUSTS , , __ !~ FIGURE 5.4 Example of the seismic expression of oceanic crust from the Blake-Bahama Basin (from Shipley asked Buffler, 1978, courtesy American Association of Petroleum Geologists). high in the early drifting phase. Unfortunately, seismic line B did not extend across the continental~ceanic crystal boundary, but the magnetic quiet zone, inter- preted as oceanic crust, is known to exist just west of this seismic line. The interpreted salt interval thins, but does not completely pinch out at the western end of this line. The outer high on seismic line A is bordered by what we interpret to be oceanic crust. The boundary is placed between reflection patterns similar to those characteris- tics of oceanic crust and to the outer high. Both seismic lines are controlled by drilling on the shelf and by regional outcrop geology. Based on these data, line A rift sediments include Triassic and Lower Jurassic (Hettangian) red beds and Lower Jurassic carbonates and anhydrites. Based on its stratigraphic position, we inter- pret the thick salt on line A to be Early Sinemurian in age, but it could be significantly older or younger. The rift sediments are overlain by drift sediments consisting of a Middle Jurassic to Cretaceous Berriasian prograding car- bonate bank and a Cretaceous-Tertiary prograding elastic wedge (see Vail et al., 19771. Wells penetrating the shelfnear seismic line B indicate that the rift sediments are Neocomian-Aptian elastics overlain by late Aptian salt. DriR sediments are post- Aptian Cretaceous carbonates and elastics and Tertiary elastics. No good evidence for volcanics is shown on these sec- tions in the vicinity of the outer highs. However, on seismic sections from other areas, the reflection paKem of volcanics is common and quite variable. Volcanics exhibit considerable variation in paleotopography at the top and/ or base of the unit as evidenced by onlap. Interval reflec- tions within the unit commonly dip at a greater angle than the top or base. In many cases these dipping reflections are high amplitude and continuous for short distances. In other examples, extrusives appear as lens-shaped lobes with a relatively high-amplitude reflection at the top and little to no reflections within the lobe. Interval velocities it, tend to be intermediate to high in the first example and high in the second. We interpret the first example to be largely interbedded volcanics and the second example to be massive flows. Intrusives commonly appear on seismic data as irregular high-amplitude reflectors, sometimes crossing and sometimes subparallel to reflections from stratal surfaces. Reflections originating Mom intnlsives commonly occur sporadically in the vicinity of outer highs. Only limited magnetic and gravity data were available over the outer highs described in this study. Where magnetic data were available, the presence of magnetic stripe anomalies correlates well with the typical seismic expression of oceanic crust discussed earlier. However, the steep magnetic gradients oRen associated with the continental~ceanic crustal boundary commonly occur well landward of the outer high. In areas where extensive volcanics are present, the relation of the magnetic pat- terns to the continental-oceanic crust boundary was vari- able. In some cases, where volcanics were present, the steep magnetic gradient was seaward of the outer high. Gravity data commonly show positive isostatic gravity anomalies extending along the interpreted outer edge of the continental margin (Emery et al., 1970; Talwani and Eldholm, 1972; Rabinowitz, 1972~. Figure 5.8 is an exam- ple. A gravity low is commonly present landward of the outer high' indicating a deep sedimentary basin. The as- sumed shallow basement of the outer high causes a rather strong gravity anomaly but not a large magnetic anomaly (Talwani and Eldholm, 1972~. SUMMARY OF EVOLUTION OF DIVERGENT (PAS S IVE ) CONTINENTAL MARGIN S If the formation of outer highs during the late rid stage is as prevalent as we believe, it would add another phase to the evolution of divergent (passive) continental margins.

up- on ~ YoL~ ~ 1 .;1r,.~ ~ ~ ~- ~ \:: -. ~. . ~.~1~.~N ~; ~ ''' + + + 1 ~~ ~ ~ + ~ + + ~ ~ ~ + L" C`3 .~~ . #I.! - ~ \ ~ +\ ~ · i, ~ F~,~ t I ~ - \ / ~ ~ / ~ ~ ' " 1~_~\~; I ~ ~ _ / _ / _ .~:': l 0 0 0 0 _ ~ ~ ~.D ~ ~ ~ — CO ~ ~ ~ I i ~ —— ~ _ I 0 0 0 0 0 15) Cal SC1NO:)3S- 3W11 ~ 1 ~ r rat ~ ~ ~ ~ ~ r ', ~ r L [' ~ L A )' 7 L 1 A ~ >` r rid 1 I C L ~ 55 - 3 - .;~ - ,, ._ :~ ,

56 AGES ~= TO RTONIAN SER RAVALLlAN BU RDIGALIAN CHATTIAN RUPELIAN PR IA80 N IAN 8ARTONIAN LUTETIAN YPR ESIAN THANETIAN OAN IAN MAASTRICHTlAN CAMPANIAN SANTONIAN _ CONIACIAN TU RONIAN CENOMANIAN ALBIAN APTIAN BAR REMIAN HAUTER IVIAN VALANGINIAN z BER RIASIAN _ PO R r LAN D IAN TITH O N - , ~ ~, KIMMERIOGIAN ~ JAI' OXf O RDIAN CAL LOVIAN BATHONIAN BAJOCIAN AAUENIAN TOARCIAN PLI ENS8ACHIAN Sl NEMU RIAN HEUANGIAN R HAETIAN NORIAN ~ Il . -=-—~ T. ~ . T~T2 T Te _ :T02.2 Td — TO 1 Tc _ E2 2 T TE3; Ta — T92j 1: _ ~,-~ _ j K2 ~ j ~ _ — ~/'~ ~ K~ _ Q K~ I Ka - _ J22 -t50 TR2 3 T R —208 c FIGURE 5.6 Stratigraphic col~nn represented in seismic line A (Figure 5.~). MARTIN A. SCHUEPBACH and PETER R. VAIL This section summarizes the evolution of divergent conti- nental margins as a basis for placing the formation of the outer high as an evolut~onary phase in divergent continental-margin development A vast amount of literature has contributed to the un- derstanding of clivergent margins. Just a few publications are citecl here: Drake et al. (1959~; Belmont' et al. (1965~; Burk (1968~; Emery et al. (1970~; Talwani asld Eldholm ( 1972, 1977~; Rabinowitz ( 1974~; Boeuf and Ooust ( 19751; Exon and W~llcox (19781. A summary of the evolution of divergent (passive) con- tinental margins (Figure S.9) is presented by the IPOD Committee on Passive Continental Margins (CulTay, in press). In that paper, the evolution of divergent margins is subdivided into the following phases: (1) variable degrees of doming, (2) niting, and (3) driflting. Doming of graben margins, such as occurs in the East African riR (Pilger and Rosler, 1976) and the Rhine Graben (Illies, 1977) might not always be present and in some cases may occur while the riRing was taking place (Carey, 19581. RiRing occurs as a single graben or as a complicated subparallel system of grabens and horsts. The driRing is characterized by the fonnation of oceanic crust, seafloor spreading, and subsidence. TECTONIC-SEDIMENTARY MODEL FOR THE EVOLUTION OF OUTER HIGHS ON DIVERGENT MARGINS The tectonic-sedimentary model presented here (Figure 5.10) for the evolution of the outer high differs fiom the Curray (in press) model for divergent margins by adding a late-riR phase consisting of a longitudinal predriR upliR within the rifle graben or graben complex. Our model shows two phases of rifting and two phases of clriPcing. The initial nflcing phase fonns a single graben in ~is model [Figure 5.10(a)~. However, as discussed in the previous section, the initial nPring may be an intricate graben system. The initial grabening phase may be caused by crustal thinning over a zone several times wider than the graben itseIf (Rhine Graben Research Group for Explosion Seismoiogy, 1974~. Figure 5.11, re- produced £rom the Rhine Graben Research Group report, shows a cross section of the Rhine Graben. This section is an example of the relationship between early ri:°cing and crus~l thinning. Whether the sediments deposited during the early-riR phase are manne or nonmarine depends primarily on ~e thickness of the continental crust in which the nflc forms. Where the crust is thick, the early rifle tends to be topo- graphically high and nonmarine sediments tend to be deposited. Where the crust is thinner, ~e level of the rifle floor tends to be close to or below sea level, resulting in marine or interbedded marine and nonmarine sediments. Sea-level fluctuations as described in Vail et al. (1977) and access to marine waters are important factors in con- trolling the sediments deposited in these types of rifts.

:. 1 ' : l : 1 ' ! en SON003S - 3W11 o 1 1 1 1 o ~ o lo. en c" ~ In SON003S- ~W11 , _ ~ . . . _ . . , , , o o o C" ~ HA sONO03S- 3W11 _ E Y U.

58 5a BURIED RIDGE ,~\ 50 ~INTOUR INTERVAL 20mgal Ad\\ ~ JO , . , , , I'\ \ , to 10° 15° FIGURE 5.8 Isostatic gravity anomaly map (from Rabinowitz, 1972) of Me Angola Continental margin showing the outer high as a positive anomaly. Dunng the late-nRing phase [Figure 5.10(b)], a narrow irregular longitudinal uplift fortes within the graben. In cross-sectional view, the narrow uplifted strip has the am pearance of a basement high. Also, during the late-nE phase, the area of thinned continental crust on both sides of the basement high continues to subside. Our data in- dicate that the late-nh phase sedimentary sequences in these two subsiding basins commonly consist of facies varying from moderately dee~water sediments to evade orites. In general, erosion removes the early-rifle-phase nonmarine sediments Dom the uplifted median basement high, and late-rifle-phase marine strata onlap it. Our geo- physical studies indicate that the outer high is commonly injected by basic intrusions and is in many cases covered by basic extrusives. The upward movement ofthe median basement high appears to us to be caused by ~ thermal- MARTIN A. SCHUEPBACH and PETER R. VAIL DON15116 R~6 ~ - ~5~ 1 ~EDU~ l l |2,W'~ ~ ~ MAN FIGURE 5.9 Passive margin model Tom Curray (in press). ]5D mechanic event Mat immediately precedes We fo~a- tion of oceanic crust. The process of the therrnal- mechanical event is not well understood and is an area recommended for further study. At the beginning of the drift phase [Figure 5.10(c)], the outer high splits longitudinally into two outer highs sepa- rated by oceanic crust The outer highs subside rapidly from close to sea level to We presently depths following the subsidence curves for thinned continental crust and oceanic crust cliscussed by Parsons and Sclater (1977~. Cooling ofthe mantle below the thinned continental crust and the newly formed oceanic crust and sedimentary C ~ ~ LEO - PROGRADHG SWAMPS OF DREG PHASE i, ~ MARE ~M~ 2- ~ / _ ', --, / `'' \ _ I,—' BROW ~ - Rem OF MOW ~ _' ~ ~—~ 1__ _\ —_ ~U RR ~ I -.-1 SEDIMENTS OF EARLY '' '---"aid-,-' ~ PRERH~M~ FIGURE 5.10 Tectonic sedimentary model showing the evolu- tion of outer lights on divergent margins.

Evolution of Outer Highs on Divergent Continental Margins w FIGURE 5.11 Crustal cross section through Rhine Graben (from Rhine Graben Research Group for Explosion Seismology, 1974). Thinning of the continental crust and mantle upwelling occur in a zone several times wider than the width of the graben. loading results in tilting and subsidence of Me continental margin. In the proper climate, the rapid subsidence and initially small Drainage areas for elastics provide a deposi- tional site ideally suited for the development of Dick pro- HYPOTHEllCAL DEPTH CONTOUR UN~ \ MEDLAN OCEANS m=1 rat 4. + 1 EAflLY RlR WAGE ONMM.E SWIMS) LATE B" n"E MARIIIE SEDINIEIJ'S) PROGRAD.16 MARIIIIE SE0 - Em FIGURE 5.12 Schematic evolution of divergent margin open- ing in a pivotlike fashion. All stages of evolution, from initial rifting to the final drift stage, may occur simultaneously along a margin opening in a pivotlike fashion. The marine basins of the late-rift stage may be restricted and completely separated From the early-riPc and driR phase by the outer high and transform faulting. 59 grading carbonate banks and reefs. As subsidence slows down througl, time, the drainage area caused by subsi- dence expands. This tends to cause tenigenous elastics to build seaward, commonly covering the carbonate bank margins. Full development of the features associated with riding is in part dependent on the total extension across the rift. A divergent margin that undergoes a pivotlilce opening (Figure 5.12) will fully develop its characteristic features sooner at large distances from the pivot point than at small distances, because large extensions are achieved sooner there. Thus, even though extension along tlie length of the opening is synchronous, basin formation and filling are diachronous. Also, the early-driR basin may be sub- divided into subbasins by highs caused by transform fault- ing. Subbasins thus formed can restrict the circulation within the late-rift- and early-driR-phase basins, thus causing conditions favorable to the deposition of black shales and/or evapontes. PROBLEMS AND RECOMMENDATIONS The combination of regional geological studies and multi- channel seismic profiles used in conjunction with other geophysical data has provided the information for the definition of some key problems related to the evolution of outer highs on divergent continental margins. We rec- ommend that a long-term program consisting of geo- physical studies, emphasizing regional multichannel seismic lines, and regional geological studies, including outcrop, piston, and dredge samples, be undertaken to locate optimum drill sites for the best possible confinna- tion ofthe geophysical interpretations and understanding of the geological problems. Such work is proposed for the 1980's by the JOIDES program (White, 1979~. The key geo- logical problems that we have defined concerning the outer highs on divergent continental margins are the following: 1. The character, age, and location of the continental- oceanic crusted boundary; 2. The existence, nature, size variations, and evolution of median highs within grabens to outer highs on diver- gent continental margins; 3. The nature and evolution of the overthickened oceanic crust and volcanics commonly present in We area adjacent to the continental-oceanic crustal boundary; 4. The type, age, and controls of late-riR-phase sedi- mentation, particularly organic-nch shales and evap- orites; 5. The relation of subsidence and heat flow to crustal type and thickness; 6. The paleoenvironmental history of the sedimentary section through the drift phase including age, paleontol- ogy, hiatuses, unconformities, lithofacies, paleobathym- etry depth, physical properties, and other significant parameters.

60 Venfication of the model relies on extensive geophys- ical surveys involving multichannel seismic, magnetic, and gravity data and testing of optimum sites by drilling. If the mode! is verified, it will better explain many of the geological problems associated win divergent conti- nental margins, including a better understanding of the nature of the oceanic~ontinental crusted transition, asso- ciatec3 deposition including salt and organic-nch shales, and Me relation of volcanism to the onset of seafloor spreading. SUMMARY Observation of a large number of geological and seismic profiles across divergent continental margins has demon- strated the widespread occurrence of structural highs that were formed during the late-nflc phase of divergent mar- gin evolution and are now the boundary between conti- nental and oceanic crust. This feature originates as a median high, which divides the rift zone into two subsid- ing basins. This median high later evolves into an outer high along doe newly forrneil condnental~ceanic bound- ary during the drift phase. The outer high commonly ex- tends the whole length of a divergent margin but can be segmented by such factors as transform faults andJor vol- canics. In many cases, the outer highs are complicated by extensive intrusives and extrusive volcanics. Early-riP: sediments are commonly nonmarine, but in areas of Tinned continent crust they may be manne. Manne sediments or volcanics are almost Sways deposited in the late-nflc basins on opposite sides of the median high. Salt and organic-rich shales are commonly deposited during this phase. This chapter has presented a mode! for the develop ment of an outer high on clivergent continental margins that relates the major structural episodes with the asso-. ciated sedimentation. It consists of (1) early-nflc-phase graben formation, ~a' late-nR-phase development of a median high subdividing the graben complex into two subsiding basinal areas, and (3) driflc-phase Connation of oceanic crust and He outer high as a remnant of the median high and subsidence. ACKNOWLE OGM ENTS The writers are indebted to E. B. Nelson of Esso Explora- tion, Inc., anil I. Viellart of Esso aEP for Heir seismic interpretations; to I. P. Verdier and R. I~ehmar~n of Esso Production Researc~European Laboratones for their paleontological contributions; to H. R. Gould of Esso Pro- duction Research Company for his guidance; to C. G. Zinkan, D. R. Seely, R. W. Murphy, I. K. Davidson, D. B. Macurda, K. T. Biddle, R. C. Vierbuchen, R. P. George, Ir., all of Exxon Production Research Company, who re- viewed the manuscnpt; and to Esso Exploration, Inc., for permission to publish He seismic lines. MARTIN A. SCHUEPBACH and PETER R. vAIL RE FE RE NC I: S Belmonte, Y., P. Hirtz, and R. Wenger (1965). The salt basins of the Gabon and He Congo (Brazzaville): a tentative paleogeo- graphic interpretation, in Salt Basins Around Africa, Elsevier, Amsterdam, pp. 5~74. Boeuf, M. G., and H. Doust (1975). Stnlcture and development of the southern margin of Australia, Aust. Petrol. Explor. Assoc. J. 15, 3~43. Bulk, C. A. (1968). Buried ridges within continental margins, N.Y. Acad. Sc'. Trans. al:), 135 160. Carey, S. W. (1958). A tectonic approach to continental driR, in Continental Drift, a Symposium, S. W. Carey, ea., Geology Department, University of Tasmania, Hobart, Tasmania, Australia, pp. 177~49. Curray, J. R. (in press). The I~D P=g~me on Passive Conti- nental Margins, in The Evolution of Passive Continental Mar- gins in the Light of Becent Dnlling Results, Philosophical Transactions of the Royal Society, London, England. D'alce, C. L., M. Ewing, and G. H. Sutton (1959). Continental mans and geosynclines: the east coast of North Amenca, north of Cape HaKeras, in Physics and Chemistry of the Earth, Vol. 3, L. H. Ahrens, ea., Pergamon, New Yorlc, pp. 11~158. EmeIy, K O., E. Uchupi, J. D. Phillips, C. O. Bowin, E. T. Bunce, and I. T. Knott (1970). Continental rise oReasten~ North Amer- ica, Am. Assoc. Petrol. Geol. Bull. 54, 99~108. Exon, N. F., and T. B. Wilicox (19~78). Geology and petroleum potential of Exmouth Plateau and off western Australia, Am. Assoc. Petrol. Ceol. Bull. 641, 4~72. Hailer, J. (1971). Geology of the East Greenland Caledonides, Interscience, New York, 413 pp. Illies, J. H. (1977~. Ancient and recent niting in the Rhine graben, Geol. Midnbouux 56, 32~. . Pawns, B., and J. G. Sclater (1977). Analysis of *me variation of ocean floor bathymetry and heat flow with age, J. Ceophys. Res. 82, 80~7. Pilger, A., and A. Roster (1976). Temporal relationships in die tectonic evolution of the Afar Depression (Ethiopia) and the adjacent Arabian riR system, in Afar Between Continental and Oceanic Rifting, Interunion Commission on Geody- namics, Rep. No. 16, E. Schweizerbart'sche Verlagsbuchhand- lung, Stuttgart, pp. 1-25. Rabinowitz, P. D. (1972). Gravity anomalies on the continental margin of Angola, Afiica,J. Geophys. Res. 77, 6327 6357. Rabinowitz, P. D. (1974). The boundary between oceanic and continental crust in the we stem North Atlantic, in The Geology of Continental Margins, C. A. Burls and C. L. Dralce, eds., Spnuger-Verlag, New York, pp. 67~84. Rhine Graben Research Group for Explosion Seismology (1974). lithe 1972 seismic refraction experiment in the Rhine graben first results, in Approach to Taphrogenesis, J. H. Illies and K. Fuchs, eds., Schweizerbart'sche Verlagsbuchbandlung, Stuttgart, pp. 122 137. Shipley, T. H., and R. T. Buffler (1978). Seismic stratigraphy and geologic history of Blalce Plateau and adjacent western Atlantic continental margin,Am. Assoc. Petrol. Geol. Bull. 62, 792~19. Talwani, M., and O. Eldholm (1972). Continental margin offNor- way: a geophysical study, Geol. Soc. Am. Bull. 83, 3S75~606. Talwani, \1., and O. EIdholm (1977). Evolution of Norwegian Greenland Sea, Ceol. Soc. Am. Bull. 88, ~999. Vail, P. R., R. M. Mitchum, Jr., R. G. Todd, I. M. Widmier, S. Thompson III, J. B. Sangree, Jr., J. N. Bubb, and W. G. Hatlelid ( 1977). Seismic stratigraphy and global changes of sea

Evolution of Outer Highs on Divergent Continental Margins level, in Seismic Stratigraphy~pplications to Hydrocarbon Exploration, C. E. Payton, ea., Am. Assoc. PetTol. Geol. Stem. 26, Tulsa. Oklahoma, pp. 4~212. Vail, P. R., R. M. Mitchum, Jr., T. H. Shipley, and R. T. Buffler (in press). Unconformities of Me North Atlantic, in The E?;olu- tion of Passive Continental Margins in the Light of Recent Drilling Results, Philosophical Transactions of the Royal Society, London, England. 61 Veevers, J. J. ( 1977). Paleobathymetry of the crest of spreading edges related to the age of ocean basins, Earth Planet. Sci. Lett. 34, 100 106. White, 5. M. (1979). Deep-sea drilling, Ceotimes, January 1979,