GEOLOGICAL SOCIETY OF AMERICA SPECIAL PAPER 151 ©1975

Plate Tectonics and the

Structural Evolution o f the

Aleutian-Bering Sea Region

DAVID W. SCHOLL U.S. Geological Survey Menlo Park, California 94025 U.S.A.

EDWIN C. BUFFINGTON Naval Undersea Center San Diego, California 92132 U.S.A.

MICHAEL S. MARLOW U.S. Geological Survey Menlo Park, California 94025 U.S.A.

ABSTRACT

The general aspects of the structural evolution of the Aleutian-Bering Sea region can be described in terms of plate tectonics. Involved in this model is the formation of the Aleutian Ridge in latest Cretaceous or earliest Tertiary time. The ridge is presumed to have formed in response to a southward relocation in the convergence zone of the Pacific oceanic plate, a shift away from the Beringian continental margin connecting Alaska and Siberia to an oceanic location at the Aleutian Trench. Prior to the formation of the ridge, Pacific crust is presumed to have directly underthrust the northeast-trending Koryak-Kamchatka coast. The middle and late Mesozoic eugeosynclinal or thalassogeosynclinal masses that underlie this segment

1 2 SCHOLL AND OTHERS of the Pacific fold belt are highly deformed, thrust faulted, and intruded by ultra- mafic bodies—characteristics that can be ascribed to the mechanical and magmatic consequence of plate underthrusting. This model implies a similar orogenic process for the formation of the stratigraphically and structurally similar Mesozoic rocks underlying the northeast-trending continental margin of southern Alaska. Less intense underthrusting may have occurred along the northwest-trending Pribilof segment of the Beringian margin connecting Alaska and Siberia. This margin may have been more parallel to the approximate direction of relative motion be­ tween the oceanic and continental plates. Nonetheless, fold belts, possibly intruded by ultramafic masses, formed along this segment of the Beringian continental margin in Late Cretaceous and perhaps earliest Tertiary time. The folds have since subsided below sea level—their eroded tops presently underlying as much as 3 km of virtually undeformed Cenozoic deposits. Our model relates pre- and postorogenic deposits underlying the Beringian margin and adjacent coast to the time of formation of the Aleutian Ridge, which marked the cessation of continental underthrusting and the beginning of island-arc under­ thrusting. Our model also implies that the ridge formed near or at its present loca­ tion and that oceanic crust of late Mesozoic age underlies the Aleutian Basin of the Bering Sea. Since formation of the ridge this basin has received from 2 to 10 km of sedimentary fill. Although the model we suggest broadly explains the observed changes in tectonic style, magmatic history, and sedimentation for the Aleutian-Bering Sea region, it also implies that thousands of kilometers of oceanic crust underthrust the Kamchat­ ka, Beringian, and Alaskan margins between Late Triassic and Late Cretaceous time, and hundreds of kilometers underthrust the Aleutian Ridge in Cenozoic time. The enormous masses of pelagic and volcanic offscrapings that would be indicative of extensive or long-term crustal underthrusting are not apparent as mappable units. Thus, while our model may be stylistically adequate, it paradoxically predicts quantities of rocks and structures that we are not able to find. Presumably they have been subducted.

INTRODUCTION

During the last decade an impressive volume of bathymetric, geologic, and geo­ physical data has been gathered in the Aleutian-Bering Sea region (Fig. 1). Emerging from this information is a partically decipherable geologic history extending back into Mesozoic time—a history that includes the formation of the Aleutian Ridge, the sedimentary infilling of the deep basin of the Bering Sea, and the structural evolution of the Beringian continental margin, which reaches in a broad northward- swinging arc from Cape Kamchatka on the west to the Alaska Peninsula on the east (Fig. 1). Although a number of geotectonic models have already been applied to the Aleutian-Bering Sea region (see Stone, 1968; Cameron and Stone, 1970; and Perry, STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 3

1971, for reviews of some of these schemes), we will attempt to relate prior and new findings within a framework of the expected consequences of sea-floor spread­ ing and interacting lithospheric plates (Isacks and others, 1968; Mitchell and Reading, 1969, 1971; Hamilton, 1969, 1970; Dewey and Horsefield, 1970; Ernst, 1970; Dewey and Bird, 1970; Oxburgh and Turcotte, 1970, 1971; Matsuda and Uyeda, 1971;Hasebe and others, 1970). Our purpose in preparing this paper has been to determine how the structural evolution of the Aleutian-Bering Sea region fits in with the suspected pattern of global tectonics, which many geologists believe is the causative mechanism behind the formation of the Pacific fold belt of late Paleozoic through Cenozoic rocks. To some extent we are molding our data to a prescribed model, but our general ability to do this, although we encounter difficulties, must not in itself be taken as proof that the plate models are valid. These models, however, are exceptionally instructive in that they allow us to treat the structural and magmatic evolution of this vast area in a unified way. This paper contains speculations about these unifying schemes. The Aleutian-Bering Sea region includes the Aleutian Trench and bordering ridge, the Bering Sea Basin (that part of the sea exclusive of its broad and shallow shelf) north of the ridge, and the Beringian continental margin sweeping northward around the basin between Alaska and Kamchatka. The submerged physiography of this vast region has been described and discussed by Gibson and Nichols (1953), Udinstev and others (1959), Gibson (1960), Gershanovich (1963), Nichols and others (1964), Kotenev (1965), Nichols and Perry (1966), Perry and Nichols (1966), Lisitsyn (1966), Scholl and others (1968), Chase and others (1970), and Perry (1971). Because the deep Bering Sea did not exist prior to the formation of the ridge, we discuss first the known geology of this ridge and its possible relation to postu­ lated movements of major lithospheric plates. Next we consider what is known about the structural, magmatic, and sedimentary histories of the deep Bering Sea and its bordering continental margin-also interpreting these findings in relation to those expected from plate interactions and the presence or absence of an outlying Aleutian Ridge. In a concluding section we outline what appears to be the best-fitting plate model, mentioning at this point important conflicts in observed and prescribed findings.

ALEUTIAN RIDGE

General Background

Based on the age of rocks actually exposed, initial growth of the ridge can with certainty be placed only in earliest Tertiary time (Fig. 2; Scholl and others, 1970a; Carr and others, 1970); an earlier or Cretaceous age is a speculation, albeit a reason­ able one (Coats, 1956a; Gates and others, 1971). Burk’s (1965) geologic study of 4 SCHOLL AND OTHERS the Alaska Peninsula and adjacent Pacific continental margin also stresses that the ridge is wholly of Cenozoic age; however, structural and stratigraphie findings from regions adjacent to the Aleutian Ridge can be interpreted to mean that it existed prior to the Tertiary Period. This evidence includes (1) the findings of Markov and others (1969) at Cape Kamchatka (Fig. 1) that structural trends including rocks of Cretaceous and Paleogene ages strike roughly toward the western end of the Aleutian Ridge; (2) the Middle Jurassic or older age of the structural and geomorphic trend of the Alaska Peninsula (Burk, 1965), which is contiguous with the opposite or eastern end of the island chain (Fig. 1); and (3) the presence of a deeply buried and thick (1 to 4 km) intermediate velocity (3.2 to 4.3 km per sec) rock layer over- lying normal oceanic crust (Shor, 1964; Kienle, 1971; Ludwig and others, 1971a) beneath the Aleutian Basin (Fig. 3) that may be a thick sequence of terrigenous deposits impounded north of a Cretaceous Aleutian Ridge (Scholl and others, 1968; Hopkins and others, 1969; Scholl and Hopkins, 1969). Viewed simply, the rocks of the Aleutian Ridge can be subdivided into four groups or series: (1) initial, (2) early, (3) middle, and (4) late. The principal distin­ guishing criterion is age, but they also differ in lithology and style of deformation. Whether or not initial growth of the Aleutian Ridge began in Mesozoic time, we presume that ridge growth was initiated by the voluminous outpouring of mafic lavas, probably of subalkaline or tholeiitic composition (Jakes and White, 1969, 1972). These rocks constitute the great bulk of the ridge (that is, its basement complex) and form the initial series. Unequivocal exposures are not known, but the youngest part of the series may be represented by thé Eocene or older Finger Bay Volcanics of Adak Island (Figs. 1,2; Coats, 1956a; Fraser and Snyder, 1959; Scholl and others, 1970a). We are unaware of exposures that suggest the initial series may include slabs of uplifted or obducted oceanic crust, which apparently form part of the basement complex of other Pacific margining island arcs (Coleman, 1966; Shiraki, 1971; Kroenke, 1972; Bryan and others, 1972; Ewart and Bryan, 1972). By at least late Eocene time the ridge had attained approximately its present size, and portions of it had built above sea level (Gates and others, 1954; Gates and Gibson, 1956; Drewes and others, 1961 ; Scholl and others, 1970a; Gates and others, 1971). Subsequently, and through early Miocene time, large volumes of sedimentary rock, chiefly beds of conglomerate, graywacke, argillite, chert, siliceous shale, diato- maceous siltstone, and volcaniclastic deposits, accumulated in basins between volcanic centers and over the flanks of the ridge. These sedimentary deposits and associated volcanic rocks form the early series (Marlow and others, 1973), a unit equivalent to the “early marine series” of Wilcox (1959). Deposition of these rocks indicates that vigorous subaerial erosion of the ridge began after about middle Eocene time. The great volume of sedimentary debris may also signify that growth of the ridge by effusive magmatism diminished at about this time. The erosional detritus attest either to an early Tertiary orogenic episode or to the constructional growth of the ridge above sea level. The volcanic rocks of the early series are basaltic and andesitic in composition, but they have been altered to greenstone and extensively albitized. Gates and others (1954) and Wilcox (1959) note that they form a spilitic- Line of gtoJogic X-section shown on Figure 5 2oO°0'—\ Depth in m eter» (Adapted from King, 1969)

Figure 1. Generalized bathymetric chart and index map o f the Aleutian-Bering Sea region. Contour interval is 400 m beginning at 200 m, which marks approximately the outer edge of the continental shelf.

SCHOLL AND OTHERS, FIGURE 1 Geological Society o f America Special Paper 151 ADAK ISLAND SOUTH NORTH

Andrew Lak« Formation j late Eocene Finger Bay Volcanic» ✓ Finger Bay Volcanic* (Early or ini ti si «cries) (E«rly S « n < l ) -Mt. Ada gdakj Granodiorlte (13m .y.) Quaternary (0.2 rn.y.)r ♦ I j t z . À -“A y- >;- - I - i f * * K ^ * £ * -Y- “ Km K m

10 Km _I XI Figure 2. Generalized north-south structure section across Adak Island, Aleutian Islands. flows o f Mount Adagdak (Cameron and Stone, 1970, p. 88). Geologic information from K-Ar analyses were employed to date plutonic body (Marlow and others, 1973a) and basal Coats (1956b), Fraser and Snyder (1959), and Scholl and others (1970a).

- 2ND LAYCA

Deformed volcanic lastic -3 AD LA V E It •10 mudstone, insular slope s' Km deposit»; late Miocene & o ld e r

S' ALEUTIAN A 6TSSAL PeUgics; middle - Oligocene to H o lo c e n e ( ^VA* X 2 5 '“Turbidite s; late ear^y Eocene to middle Oligocene 3 .5 Acoustic velocity Approximate limit/ of in K m / 3e c A leu tian Abyssal Plain S------N —200^---^ Line of X- se ctio n depth in m e t e f s

Figure 3. Idealized crustal section from Pacific Basin to outer part of Bering conti­ across Aleutian Abyssal Plain is generalized from Hamilton (1967, 1973). Age and nental shelf. Construction is based on reflection and refraction profiles of Shor (1964), lithologic assignments are based on subsurface sampling at DSDP sites (Creager, Scholl, Ewing and others (1965), Scholl and others (1966, 1968), Scholl and Hopkins (1969^, and others, 1973) and dredging of outcrops on continental margin. Ludwig and others (1971a, 1971b), and unpublished data in authors’ files. Section S -N*

SCHOLL AND OTHERS, FIGURES 2 AND 3 Geological Society o f America Special Paper 151 STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 5 keratophyric suite of submarine effusives not unlike those typical of eugeosynclinal assemblages. Associated sedimentary rocks are slightly metamorphosed. In middle and late Miocene time the ridge was uplifted, and along its crest the sedimentary and volcanic rocks of the early series were broadly folded and exten- sionally faulted (Gates and others, 1954; Gates and Gibson, 1956; Anderson, 1970; Marlow and others, 1973). Between 10 and 15 m.y. ago, presumably during the major phase of uplift, the rocks of the early series were intruded by epizonal plutons typically of granodioritic composition (Fig. 2; Carr and others, 1970, 1971; Marlow and others, 1973; unpublished K-Ar dates, authors’ files). Anderson (1970) has speculated that a likely cause for the uplift was the emplacement of the plutons along the length of the ridge, the resulting engorgement and distension of the ridge fracturing the rocks of the early series and possibly initiating the formation of new crestal and ridge-flank basins. Evidently because the ridge had been elevated, comag- matic volcanism of andesitic magma was largely from subaerial vents. In contrast to the pervasively altered rocks of the older early series, the younger plutonic, sedi­ mentary, and felsic volcanic rocks are characteristically unaltered (Gates and others, 1954; Wilcox, 1959). These synorogenic rocks of middle and late Miocene age (5 to 15 m.y.) form the middle series. They are perhaps best exposed along the central segment of the arc in the general vicinity of its connection with Bowers Ridge (Fig. 1; Carr and others, 1970,1971). A record of widespread volcanism in the late Miocene (5 to 10 m.y. ago) is not preserved along the ridge crest. This was a period of continued uplift and intense ridge-crest erosion. The resulting erosional debris accumulated in summit basins created by crestal rifting (Marlow and others, 1970) and over structural terraces (for example, the Aleutian Terrace) along the flanks of the ridge (Marlow and others, 1973). Deep subbottom sampling at Deep Sea Drilling Project (DSDP) sites 186,187, and 189 (Fig. 1) indicates that only the Pliocene and younger beds of these basins are not significantly deformed (Scholl and Creager, 1973), under­ lying upper Miocene and older strata having been previously deformed by massive slumping, diapirism(?), thrusting, and normal faulting (Marlow and others, 1973; Giow, 1973). The undeformed or postorogenic beds that have been depositionally draped over the ridge since late Miocene time constitute the late series. This defini­ tion is at slight variance with that of Marlow and others (1973), who included the rocks of the synorogenic or middle series in their early series. Also included in the late series are the emanations from the lofty stratovolcanoes that form the arcuate chain of active and quiescent vents generally centered along the northern side of the ridge crest. All older volcanic structures of constructional origin were obliterated before or during the late Miocene and early Pliocene erosional episode. Along the ridge crest the late Cenozoic stratovolcanoes formed uncon- formably above the prominent erosional surface cut across the rocks of the early and middle series. Radiometric dating of the andesitic magmas erupted from or near the existing vents (Cameron and Stone, 1970; Marlow and others, 1973), and the stratigraphic occurrence of Neogene volcanic ash determined at DSDP drilling sites (Creager, Scholl, and others, 1973) indicates that the modern volcanic arc began to form about 3 m.y. ago, in the middle Pliocene. 6 SCHOLL AND OTHERS

In summary, our present concept of the geologic growth of the Aleutian Ridge includes an initial phase (Late Mesozoic? to earliest Tertiary) during which the bulk of the ridge formed rapidly (10 to 20 m.y. or less) by mafic submarine volcanism and plutonism; an early phase (Eocene to middle Miocene) during which volcanism was diminished greatly and subaerial erosion of a tectonically or constructionally elevated volcanic terrane filled ridge-crest and flank basins; a middle or synorogenic phase (middle Miocene to middle Pliocene) during which the ridge was plutonized, uplifted, and deeply eroded; and a final phase (middle Pliocene to present) during which an extensionally rifted ridge was depositionally overlain by postorogenic deposits and crested by a chain of andesitic stratovolcanoes. In numerous ways this history is similar to the history of other island arcs outlined by Mitchell and Bell (1970) and Mitchell and Reading (1971). Our present concept of ridge growth is not that of geosynclinal filling. While it is true that the exposed rocks of the early series are the typical spilitic, silicic, and terrigenous beds of the Pacific eugeosynclinal suite (Wilcox, 1959; King, 1969; Bogdanov, 1969), the sedimentary sequences appear to have formed in isolated or at least discontinuous basins along the summit and upper flanks of a massive volcanic and intrusive welt (Scholl and others, 1970a). Because the ridge is not con­ structed of rocks that accumulated in a classic geosynclinal trough, it is perhaps conceptually more realistic to think of the exposed rocks as constituting a “eugeo- anticlinal” complex. The thick (5 to 8 km) sedimentary deposits filling and under­ lying the Aleutian Terrace are a fundamental part of this complex (Figs. 2, 6; Grow, 1973).

Growth and Plate Tectonics

Early notions about the structural history of the Aleutian Ridge and adjacent trench attempted to synthesize geologic information gleaned from outcrops, tele- seismic data that outlined a zone of hypocenters dipping northward beneath the ridge, and bathymetric observations detailing the geomorphic shape of the ridge- trench pair (Gates and others, 1954; Gates and Gibson, 1956; Coats, 1956a, 1962). More recently these findings have been compared with seismic refraction and re­ flection results (Shor, 1964; Ewing and others, 1965; Murdock, 1967; Helmberger, 1968; Gaynanov and others, 1968; Marlow and others, 1973; Grow, 1973) and gravity data (Peter, 1966; Gaynanov and others, 1968; Malahoff and Erickson, 1969; Kienle, 1971). Relating the whole to the implications of plate tectonics (Isacks and others, 1968; Stauder, 1968), and following the lead of Coats (1962), most geologists now generally interpret the ridge as being an arcuate welt of thick­ ened crust beneath effusive and related intrusive and sedimentary rocks that accumu­ lated near the zone of convergence of a north Pacific plate underthrusting a North American-Eurasian plate (Hayes and Heirtzler, 1968; Pitman and Hayes, 1968; Malahoff and Erickson, 1969; Anderson, 1970; Scholl and Buffington, 1970; Grow and Atwater, 1970; Jones, 1971; Hayes and Pitman, 1970; Marlow and others, 1973; Grow, 1973). Conversely, Perry (1969, 1970,1971) has proposed that STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 7 spreading took place from the ridge; his model is one of emergence rather than subduction of crustal material at the Aleutian Trench. Regional structural relations (outlined elsewhere) allow us to speculate that growth of the Aleutian Ridge was initiated by the fragmentation of an oceanic plate to form a new “oceanic” convergence zone in the vicinity of the existing trench (Fig. 4). The isolated or unsubducted northern portion of this plate is now the deeply buried (by Cretaceous? and Tertiary deposits) oceanic crust of the deep Bering Sea (Shor, 1964; Ludwig and others, 1971a, 1971b). Presumably, frictional heating of the descending slab or subcrustal injection of water by it triggered crustal thickening and the production of magmas that welled up to form the massive bulk of the early Tertiary ridge (Oxburgh and Turcotte, 1970, 1971; Hasebe and others, 1970;Matsuda and Uyeda, 1971). Changes in relative plate motion described for the north Pacific area (Pitman and Hayes, 1968; Atwater, 1970; Grow and Atwater, 1970; Hayes and Pitman, 1970) can be linked only generally to changes in the tectonic and magmatic history of the ridge (Marlow and others, 1973). For example, the late Paleogene (25 to 30 m.y. ago) passage of the Kula spreading center of Grow and Atwater (1970) into the trench should be related to a major tectonic transition (Fig. 4); yet subduction of the Kula ridge occurred more than 15 m.y. after the apparent reduction of mafic volcanism that corresponded with the accumulation of the sedimentary and volcanic deposits of the early series, 15 to 20 m.y. before the emplacement of the epizonal plutons that date the Miocene orogenic event, and 25 m.y. before the growth of the Aleutian stratovolcanoes (Fig. 2). The relative motion models of Pitman and Hayes (1968) and Hayes and Pitman (1970) presume that the Kula ridge was subducted in late Paleocene or early Eocene time, a time that corresponds with our suggestion and that of Marlow and others (1973) and Clague and Jarrard (1973) that rapid submarine growth of the ridge (by the presumed emplacement of the mafic rocks of the initial series) ceased before late Eocene time (Fig. 4). Pitman and Hayes’ model further stipulates that no con­ vergence took place between the Pacific plate and the ridge until after late Miocene time. Although the rejuvenated plate subduction seems to be an adequate mechanism accounting for the intense episode of late Cenozoic volcanism that spawned the Aleutian stratovolcanoes, their model does not account for the Miocene orogenic event and the attending plutonism. Also, although the evidence is spotty, ridge volcanism seems to have been continuous even after magmatism diminished in Eocene time. Perhaps, as Marlow and others (1973) suggest, convergent motion between Pacific crust and the Aleutian Ridge did not entirely cease after the Eocene subduction of the Kula spreading center, but continued at a rate sufficient to supply a modest amount of felsic magma continuously to the ridge—magma generation that culminated in the intrusion of middle Miocene granodioritic plutons and the associated orogenic event. As noted, more rapid motion after this time can be called upon for the formation of the later Pliocene-Holocene stratovolcanoes. The existence of Alaska-derived turbidites of Eocene and Oligocene age south of the Alaska Peninsula (Fig. 1, DSDP site 183, Aleutian Abyssal Plain; Fig. 3, section S'-N') and thick, land-derived hemipelagic claystone deposits of middle and late Mio­ 8 SCHOLL AND OTHERS cene age burying the northern Emperor Seamounts (Fig. 1, DSDP site 192, Miji Guyot) also imposes constraints on the amount of underthrusting that took place between the ridge and the Pacific plate during Cenozoic time (Scholl and Creager, 1973; Hamilton, 1973; Buffington, 1973). The amount is probably less than 500 to 1,000 km since the late-early Eocene, and less than 500 km since the middle Miocene STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 9

One of the more interesting implications of plate convergence is the possible growth of the lower flank of the Aleutian Ridge, the inner slope of the trench, by the piling up of pelagic sediment, trench deposits, and small fragments of oceanic crust (Grow, 1971). For example, Malahoff and Erickson (1969) have noted that the typical gravity minimum associated with the Aleutian Trench is displaced some­ what ridgeward of its axis, which suggested to them that the apparently offset mini­ mum may reflect sediment masses thrust beneath the ridge by the descending oceanic lithosphere. Despite the views of Holmes and others (1970) that direct evidence for under- thrusting has been found, seismic records across the northern or inner slope of the trench (that is, the lower slope of the ridge seaward of the Aleutian Terrace; Figs. 2, 3) that we and others have studied provide little evidence that offscraping has taken place (Marlow and Scholl, 1972; Scholl and Marlow, 1974a; Hamilton, 1973). More­ over, the steplike ascent of the inner trench slope, which includes frontal scarps as steep as 35° and more than 500 m high, are not the types of structures we would expect if the roughly 2,000-m-high inner slope is a crumpled mass of offscraped trench and pelagic debris (Perry, 1971). Nonetheless, Grow’s (1973) structural analysis of the Aleutian Terrace and the results of subbottom sampling at DSDP sites 186 and 187 located at the top of the inner slope (Creager, Scholl, and others, 1973), established that the inner slope is underlain by a deformed sequence of sedimentary beds many kilometers thick. Stratigraphic and lithologic information recovered at the DSDP sites indicates that these deposits are at least in part and perhaps mostly composed of terrigenous detritus and volcanic debris derived from the adjacent Aleutian Ridge (Fig. 6). Scholl and Marlow (1974a, 1974b) have argued that the thick wedge of turbidite deposits now filling the Aleutian Trench is a special feature of Pleistocene age. There is little reason, therefore, to suspect that thick terrigenous trench sequence would have been available in the past for offscraping at the base of the inner trench slope. Presumably, large volumes of pelagic deposits transported tectonically into the trench by the underthrusting Pacific plate would have been available for off­ scraping. Although only future sampling will determine if the stacking of pelagic debris has significantly built out the lower flank of the ridge, it is reasonable to speculate that sediment loss by subduction, perhaps in the manner suggested by

Figure 4. Idealized representation o f relative plate motions in northern Pacific-late Mesozoic and early Tertiary (Pitman and Hayes, 1968; Grow and Atwater, 1970; Hayes and Pitman, 1970; Larson and Chase, 1972; Larson and Pitman, 1972). Dashed lines represent trend of magnetic anomalies associated with northward-migrating Kula ridge; thin arrows show spreading and migration direction relative to Pacific plate (south of Kula ridge). Large arrows show approximate vector o f relative motion between Kula plate and the North American-Eurasian plate. Prior to the possible formation o f the Aleutian Ridge in latest Cretaceous or earliest Tertiary time, the Kula plate is presumed to have underthrust Kamchatka, the Beringian margin, and southern Alaska; trenches presumably formed in these areas. Growth of the Aleutian Ridge marked a southward shift in the convergence zone of the Kula plate to the Aleutian Trench, thereby isolating the continental margin o f the Bering Sea from the structural consequences of direct underthrusting. For simplicity, the diagram does not attempt to deal with the effects of oroclining in Alaska and the possible southward bowing o f the Beringian margin in the Late Cretaceous and early Tertiary (Hopkins and Scholl, 1970; Patton and Tailleur, 1972). 10 SCHOLL AND OTHERS the models of Malahoff (1970), rather than offscraping is the important tectonic process occurring along the Aleutian Trench (Scholl and Marlow, 1974a, 1974b). If this is true, then it is likely that tectonic erosion of the inner slope has steadily reduced rather than broadened the width of the Aleutian Ridge. Rutland (1971) has advanced this idea for the Peru-Chile Trench. Regardless of the tectonic processes taking place along the inner trench slope, there is little likelihood that the rocks underlying the crestal area of the ridge are elevated offscrapings. Oceanic deposits and ultramafic bodies typical of the ophio- litic suite are unknown as insular outcrops (Wilcox, 1959) and they have not been dredged from either its crest or upper flanks. In terms of the expected effects of plate underthrusting, it is important at this juncture to compare the known geologic aspects of the Aleutian Ridge to the paired or Pacific-type belts of igneous and metamorphic rocks (inner and outer belts) outlined by Miyashiro (1961) for the Japanese arc. Ernst (1970), Ernst and others (1970), Oxburgh and Turcotte (1971), Packham and Falvey (1971), and Matsuda and Uyeda (1971) have related this zonation to the thermal processes associated with a sinking plate and the resulting formation of magma beneath the landward or inner belt (Hasebe and others, 1970). The paired belts, which are separated by a magmatic or volcanic front approximately delineating the seaward limit of granitic intrusives, are well formed along only the eastern segment of the Aleutian Ridge— the zonation fading westward of Buldir Island (Fig. 1). Significantly, this eastern segment is a region of active underthrusting (to judge from teleseismic data), whereas the region west of Buldir Island is one of essentially strike-slip motion between the plate of the ridge and that of the north Pacific Basin (Isacks and others, 1968; Stauder, 1968; Grow and Atwater, 1970; Cormier, 1972). For the eastern segment of the ridge (which can also be extended landward to include the Alaska Peninsula), the outer belt would encompass its Pacific-facing slope, the Aleutian Terrace, and the trench, which combine to form an area of eugeosynclinal sedimentation as noted for the Japanese arc by Matsuda and Uyeda (1971). Hie outer belt is presumably underlain by a zone of underthrusting within which high P-Tmetamorphism takes place. The seaward or southern edge of the inner belt would correspond to that part of the ridge crest including and lying north of the belt of late Miocene epizonal plutons and the closely flanking line of Quaternary stratovolcanoes—the present magmatic front. Low P-T metamorphic rocks, also characteristic of the inner belt, are known along this part of the Aleutian Ridge (Coats, 1956a; Fraser and Barnett, 1959; Drewes and others, 1961). The inner belt may also include the entire area of the deep Bering Sea. The general models of Hasebe and others (1970), Oxburgh and Turcotte (1971), Packham and Falvey (1971), and Matsuda and Uyeda (1971) imply that magma generated by tectonic heating along the top of the plunging oceanic lithosphere would ascend beneath the floor of the basin to form the upper mantle there. The ascending magma could cause spreading or opening of the Bering Sea Basin in conjunction with the southward migration of the Aleutian Ridge (see also Karig, 1970,1971;Mitchell and Reading, 1969,1971). Obviously, as it relates to the origin of the Bering Sea, the structural implication of this model is profound. STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 11

BERING SEA BASIN

General Background

Two aseismic ridges, Bowers, looping north from the Aleutian Ridge, and Shir- shov, jutting southward from the Siberian mainland, cordon the Bering Sea Basin into three smaller basins—Bowers, Kamchatka, and Aleutian (Fig. 1). Refraction and reflection data (Shor, 1964; Ewing and others, 1965; Scholl and others, 1968; Ludwig and others, 1971a, 1971b) reveal that the Aleutian Basin, by far the largest of the three, is underlain by approximately 3 km of semiconsolidated sedimentary deposits (velocity 2.1 to 2.9 km per sec) overlying 1 to 6 km of lithified sedimentary (or in part volcanic) rock characterized by an acoustic velocity that is typically about 3.7 km per,sec (range is about 3.2 to 4.3 km per sec; Figs. 3, 5). Rocks dredged from the Beringian continental margin rimming the Bering Sea, and sedi­ mentary beds penetrated at DSDP sites 184, 185, 188, 189,190, and 191 (Fig. 1; Creager, Scholl, and others, 1973), indicate the thick semiconsolidated sequence is of Cenozoic age. The upper 500 to 1,500 m are turbidite beds interbedded with diatom ooze of late Pliocene and younger age. Pliocene and upper Miocene units in­ clude beds of diatom ooze, diatomaceous terrigenous clay, and mudstone. Older strata are probably dominated by mudstone and associated coarser terrigenous deposits. The lower lithified series may be a flyschlike terrigenous unit of early Tertiary or Cretaceous age (Fig. 3). Except for a thin and hypothetical layer of deep-sea pelagic deposits beneath it, this lower lithified unit rests on a normal “second” oceanic layer of velocity 4.7 to 5.5 km per sec that in turn overlies a 6.8 to 7.2 km per sec “third” crustal layer (Ludwig and others, 1971a, 1971b). The sedimentary fill of Bowers Basin is similar to that of the Aleutian Basin except that it is slightly thinner and directly overlies a thick (2 to 4 km) rock unit of about 5.8 to 6.2 km per sec velocity; this unit thickens to form much of the internal bulk of both Bowers and Aleutian Ridges (Figs. 3, 5; Ludwig and others, 1971a). Sparse data suggest that the structure of Kamchatka Basin is intermediate be­ tween that of Aleutian and Bowers Basins. Reflection records (Fig. 5) show that Kamchatka Basin is underlain by 1,000 to 2,000 m of semiconsolidated deposits overlying an irregular basement surface. A lower lithified layer of sedimentary rock characterized by acoustic velocities near or exceeding about 3.0 km per sec is evi­ dently present but thin or difficult to detect (Ludwig and others, 1971a). The velo­ city of the basement is poorly known, but near Shirshov Ridge it is 6.6 to 7.0 km per sec (Fig. 5; Gaynanov and others, 1968; Ludwig and others, 1971b). Sampling at DSDP site 191 recovered slightly altered tholeiitic basalt from the basement (Creager, Scholl, and others, 1973). Based on three closely agreeing K-Ar ages of feldspar microlites, the basalt is of middle Oligocene age (29.6 m.y.). The basalt was recovered from a basement knoll buried beneath about 910 m of late Cenozoic turbidite and mudstone beds; mudstone probably no older than late Miocene imme­ diately overlies the knoll. Older, higher velocity sedimentary units, located by sono- buoy reflection and refraction data (Ludwig and others, 1971b) in deeper basement 12 SCHOLL AND OTHERS swales, were evidently not deposited over the site 191 knoll. The basement can be traced to the base of the Siberian segment of the Beringian continental margin, which is underlain by mildly deformed early and middle Cenozoic deposits and intensely deformed and intruded (ultramafics) eugeosynclinal rocks largely of Cretaceous age.

Plate Tectonics, Formation, and Sedimentary Filling

Implications of a Late Mesozoic Ridge. The existence of an Aleutian Ridge in late Mesozoic time provides an explanation for the lower, or lithified, sedimentary rock sequence of Bowers and Aleutian Basins (the 3.7-km-per-sec unit) as a thick layer of impounded terrigenous deposits of late Mesozoic age (Fig. 3). The judgment that this basal layer may be a sedimentary sequence of late Mesozoic age is based on regional stratigraphic considerations (Scholl and others, 1966,1968) and the litho- logy and Late Cretaceous (Campanian) age of rocks dredged from the Beringian continental margin near the Pribilof Islands (Hopkins and others, 1969). This model implies a static buildup of sedimentary deposits within the Bering Sea Basin proba­ bly beginning some time in the Cretaceous, a buildup that took place well north of a convergence zone below a coeval Aleutian Trench. Gravity (Kienle, 1971) and refraction data (Ludwig and others, 1971a) also establish that the lower layer thick­ ens appreciably, to as much as 6 km, in a crescent-shaped body immediately flanking the outer or convex side of the north-looping Bowers Ridge (Figs. 1, 5). This can be interpreted to mean that a “trench” filled with late Mesozoic deposits, now deeply buried beneath 2 to 3 km of Cenozoic strata, occurs here. In simplest terms, the origin of a late Mesozoic Bering Sea Basin can be viewed as resulting from the volcanic growth of the Aleutian Ridge at its present location in late Mesozoic time—a simple blocking-off of the northern reaches of the Pacific Basin (Fig. 4; Shor, 1964). However, a more complicated model, one calling for Mesozoic ridge migration, must also be considered, as it is likely that the ridge was underthrust to the northwest by a rapidly sinking Kula plate (Fig. 4; Pitman and Hayes, 1968; Hayes and Pitman, 1970; Grow and Atwater, 1970; Larson and Chase, 1972; Larson and Pitman, 1972). Accordingly, the thermal models of Hasebe and others (1970), Oxburgh and Turcotte (1971), and Matsuda and Uyeda (1971) force us to speculate that the Bering Sea Basin, or parts of it, may have been formed in response to spreading or opening of the crust behind a southward-migrating ridge (Karig, 1970, 1971, 1972). Formation of part or all of the deep basin of the Bering Sea by late Mesozoic ridge migration would have been accompanied by either regional or ridge-centered emplacement of new oceanic crust and upper mantle material. In this regard it is interesting to note that the data assembled by Kienle (1971) suggest that the north- trending Shirshov Ridge (Fig. 1) may have been a spreading center (J. Kienle, 1971, personal commun.). Its north-south trend and off-center position relative to the perimeters of the basin do not obviously suggest a pattern of Mesozoic crustal generation behind a southward-migrating Aleutian Ridge. Also, the middle Oligocene age of tholeiitic basalt underlying Kamchatka Basin and the early Miocene age of STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 13

andesitic tuff from the basement of Shirshov Ridge (Fig. 5) suggest a Tertiary rather than a Mesozoic age for the volcanic crust of the western part of the Bering Sea Basin. The possibility that crustal generation and ridge migration occurred in the Cenozoic is discussed later. Geometrically, Bowers Ridge fits better as a Mesozoic spreading center for the Bering Sea Basin, but this ridge is flanked to the north by a crescent-shaped “trench-fill” that in terms of plate mechanics implies convergence rather than spreading (Kienle, 1971). Perhaps a strong argument against a late Mesozoic opening of the Bering Sea by ridge migration and, incidently, for the existence of the ridge at this time, is the absence of the paired igneous and metamorphic belts of Cretaceous age along the Aleutian Ridge, but their presence in the coastal ranges of eastern Siberia (Kamchat- ka-Koryak region) and beneath the Bering shelf. In the Koryak Mountains the outer belt comprises the thalassogeosynclinal accumulations of Bogdanov (1969, 1970), a broad zone of folded and thrusted Jurassic to Late Cretaceous volcanic and terri­ genous deposits overlying oceanic crust (Egiazarov, 1963; Krasny, 1964; King, 1969; Tilman and others, 1969; Drabkin, 1970). Similar rock assemblages underlie Kam­ chatka (Avdeiko, 1971). The geosynclinal complex includes bands of high P-T metamorphic rocks1 and tracks of ultramafic bodies that roughly parallel the present continental margin (Dobretsov and others, 1966; Irwin and Coleman, 1972). An ultramafic body also underlies St. George Island (Barth, 1956), the southern of the two Pribilof Islands rising above the Bering shelf. The corresponding inner belt can be recognized in the late Mesozoic volcanic and plutonic terrane of the Okhotsk- Chukchi volcanic zone of eastern Siberia (Krasny, 1964; Tilman and others, 1969; Drabkin, 1970), a belt that can be traced eastward to St. Lawrence and St. Matthew Islands of the Bering shelf (Patton and Csejtey, 1971; W. W. Patton, 1972, oral com- mun.). The existence of these belts can be viewed as evidence that in late Mesozoic time underthrusting north of the present Aleutian Ridge took place at the base of the Beringian continental margin, and not at the base of a southward-migrating proto-Aleutian Ridge. An additional complicating factor for a mainly Mesozoic opening of the Bering Sea Basin is the structural alignment of the eastern end of the ridge with the struc­ tural grain of the Alaska Peninsula (Burk, 1965), and possibly its western end with that of Cape Kamchatka (Fig. 1; Markov and others, 1969). It seems unlikely to us that the ridge would have migrated from its area of formation, presumably near the present Beringian continental margin, to establish an arcuate alignment with conti­ nental structures of the same or older age. We interpret these alignments to mean that if a late Mesozoic ridge had formed, it did so near its present position relative to the flanking continental areas—subsequent migration involving the entire Aleutian- Bering Sea area rather than just the ridge alone. Implications of a Cenozoic Ridge. If we assume that the Aleutian Ridge began

1It is interesting to note the findings o f Firsov and Dobretsov (1969) that some o f the high P-T metamorphic rocks along the western side o f the outer belt have radiometric ages corre­ sponding to middle Paleozoic time; yet the adjacent terrane is largely underlain by younger Mesozoic rocks. 14 SCHOLL AND OTHERS to form in earliest Tertiary time, then we must conclude that the basal sedimentary rock sequence of Aleutian and Bowers Basins is either (1) impounded terrigenous, pelagic, and volcanic debris of equivalent age, or (2) an older (Cretaceous), thick sequence of continental rise and abyssal plain deposits that accumulated in the Pacific Basin prior to the formation of the ridge (Fig. 6). The first possibility re­ quires that we abandon our previous conclusion that the basal layer may be ponded strata of late Mesozoic age. This probably also means that sedimentary rocks of Late Cretaceous (Campanian) age dredged from the continental margin correspond only in part or not at all to the thick basal layer of the adjacent Aleutian Basin. Mesozoic deposits would be represented by an undetected and presumably much thinner underlying layer of pelagic sediment resting on normal oceanic crust (Figs. 3,6). The second possibility requires the removal of terrigenous deposits of Mesozoic age from the sea floor south of the Aleutian Ridge. This is necessary because, except in the vicinity of the Aleutian Abyssal Plain, seismic reflection records south of the ridge do not reveal terrigenous deposits here (Figs. 3, 6; Hamilton, 1967; Ewing and others, 1968; Hayes and Ewing, 1970; Hamilton, 1973). In terms of plate mechanics, it is equally plausible to speculate that the missing deposits were either subducted beneath the lithosphere of the ridge or tectonically scraped off the descending Kula plate to form the rocks underlying the north wall of the trench. Offscraping is perhaps favored by most geologists investigating the tectonics of continental and insular margins (Dietz, 1963a; Hamilton, 1969, 1970;Mitchell and Reading, 1969; Dewey and Bird, 1970; Oxburgh and Turcotte, 1971); however, the rarity of pelagic offscrapings in the coastal mountains of the Pacific perimeter, which has been under­ thrust by many thousands of kilometers of oceanic crust since early Mesozoic time (Larson and Chase, 1972; Larson and Pitman, 1972), implies that subduction of trench and oceanic deposits is the dominant process (Gilluly and others, 1970; Gilluly, 1971; Scholl and Marlow, 1974a, 1974b). Regardless of what may or may not have happened to terrigenous deposits of Mesozoic age, by at least Tertiary time most of the terrigenous, pyroclastic, and pelagic debris reaching the Bering Sea Basin was trapped there. During Cenozoic time 3 to 6 km of sedimentary debris accumulated north of the ridge (Shor, 1964; Ewing and others, 1965; Ludwig and others, 1971a, 1971b). If this static buildup continues and is not terminated orogenically, simple isostatic models and Cenozoic rates of sedimentation indicate that the basin will be filled in 100 to 200 m.y.— a geologic era. At this time new continental crust 20 to 25 km in thickness (Shor, 1964) would extend southward from the present Beringian continental margin to the Aleutian Ridge, which perhaps then would be one of the numerous examples of basement highs so typical of modern continental edges (Emery, 1968a; Burk, 1968). If we return to the concept of a migrating ridge, it is possible that the Kula plate during early Tertiary time and the Pacific plates during middle and late Cenozoic time underthrust all or most of the Aleutian Ridge. At least the inner of the paired igneous and metamorphic rock belts of the Pacific type is also recognizable along the eastern segment of the ridge. Thus the models of Karig (1970,1971), Oxburgh 6ERI NGIAN MARGIN

Island a ftN LATEST CRETACEOUS

formation of Aleutian 3ubducti«n zone

(6-7 cm/yr- relative to rid ge) Motion of Pacific Plate b«9an in middlei?) Miocene after period of little or no motion since middle Eocene

Approximatc Scale X 2 5 Figure 6. Hypothetical crustal sections illustrating important stages in the structural (Fig. 1). This segment o f the margin was presumably only obliquely underthrust by the evolution o f the Aleutian-Bering Sea region. Plate convergence is presumed to be the oceanic plate; for part o f the Mesozoic it may have been a zone o f predominantly major tectonic mechanism involved. These sections are approximately along the line of strike-slip contact. Figure 3, and thus they cross the Bering margin in the vicinity of the Pribilof Islands

SCHOLL AND OTHERS, FIGURE 6 Geological Society of America Special Paper 151 B KAMCHATKA BASIN SHIRSHOV RIDGE ALEUTIAN BASIN Altered andesitic -toff overlain by nnoothed by wave erosion (?) diatom ooze; middle (?) or late Mioc Mudstone ^ fate On .younger Turbid it«s St. M¡oc.(?) Bt PIioc. diatom ooLt; late 3 - -3

Tholeiitíc. b»*altj 29.6m.y., Í Mudstone; middle -6 K-Arj middle Oligocene Mioc. (?) 8c older - X 10 Km Near (&.A-7.0) (Water) DSDP 191

c D ALEUTIAN RIDGE KAMCHATKA BASIN SHIRSHOV RIDGE ALEUTIAN BASIKI E a riy serie s > Andtsit/e toff j I6.6m.^.j K~Ar> early Miocene

X 10 DSDP 'Dcform«d mudstone, I ate M ioc. 8c. older Mudstone; middle' 169 Mioc.f?) 8c older

Figure 5. Generalized structural sections across Shirshov and Bowers Ridges. Drawings are based on data collected by the authors augmented by the geological and geophysical findings of Gaynanov and others (1968), Kienle (1971), Ludwig and others (1971a, 1971b), and Creager, Scholl, and others (1973). The lines o f the sections are shown on Figure 1. Note that the vertical scale for the upper two figures shows accurately only the water depth; the thickness o f subbottom strata is therefore slightly greater than shown. The vertical scale for the lower figure shows correctly the approximate thickness of all acoustically detected units.

SCHOLL AND OTHERS, FIGURE 5 Geological Socicty of America Special Paper 151 STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 15 and Turcotte (1971), Packham and Falvey (1971), and Matsuda and Uyeda (1971) for a southward migrating ridge must be considered for the Cenozoic. Except by analogy with the migrating and remnant arcs and ridges of the western Pacific (Karig, 1972), we are unable to identify evidence that any significant migra­ tion has occurred along the Aleutian Ridge since its formation. Arguing against ridge migration is the previously mentioned problem of a resulting fortuitous alignment with continental structural elements, in this case with elements of equal or older age. As Burk (1965) suggests, it seems more likely that the Aleutian Ridge formed in the early Tertiary along the long-established Mesozoic trend of the Alaska Penin­ sula (Figs. 1,7, 8). The extraordinarily thick and generally uniform fills of Aleutian and Bowers Basins argue for simple damming north of a generally static ridge. De- positional units extend completely across the larger Aleutian Basin, from continental margin to ridge flank, with no obvious extensional rifting, or depositional wedging or shingling suggestive of infilling behind a moving ridge (Shor, 1964; Ewing and others, 1965; Scholl and Marlow, 1970; Ludwig and others, 1971a, 1971b; unpub­ lished records in authors’ files). However, as emphasized earlier, mafic volcanic crust has formed in Kamchatka Basin, the western part of the Bering Sea Basin, at least as recently as middle Oligo- cene time—a circumstance that may in part account for its relatively thin (1,000 to 2,000 m) sedimentary fill. Kamchatka basin is flanked to the west by Shirshov Ridge, a sediment-covered mountain range largely formed by volcanism in early Tertiary time; this ridge connects with Late Cretaceous and early Tertiary structural trends at Cape Oliutorsky (Fig. 1). To the south the basin is bordered by the Koman- dorsky segment of the 2,200-km-long and smoothly arcuate Aleutian Ridge, and to the west and north by Cretaceous and early Tertiary structural trends of the north­ ern part of Kamchatka (Krasny, 1964; Vlasov, 1964; Drabkin, 1970; Avdeiko, 1971). The basin appears to be nearly enclosed by structural trends, connections, and alignments that are as old or older than its Oligocene basaltic crust sampled at DSDP site 191. This arrangement of structural blocks and one possible subduction zone (Aleutian Trench) does not suggest basin formation by crustal spreading behind either a southward-migrating Aleutian Ridge or an eastward-shifting Shirshov Ridge, which Karig (1972) speculates may be a remnant arc. For Shirshov Ridge, the thick blanket of undisturbed sedimentary beds of Neogene age lying at the base of its eastern or Aleutian Basin side (DSDP site 189, Creager, Scholl, and others, 1973) stipulates that any eastward motion ceased by the end of Paleogene time. Early Tertiary migration of this ridge to form Kamchatka Basin would have been possible only if a subduction zone formed along its eastern side, or at the base of the Ber- ingian continental margin in the vicinity of the Pribilof Islands (Fig. 1). We know of no evidence that subduction took place in these areas in early Tertiary time (Scholl and others, 1968; Kienle, 1971). One further point, reflection profiles and dredging reveal tfyat a thick sequence of Miocene and older Cenozoic beds underlies the con­ tinental margin of northern Kamchatka. Geologic mapping in the coastal area further demonstrates that great thicknesses of highly deformed eugeosynclinal deposits of Cretaceous age (thought by Avdeiko, 1971, to include trench deposits) form the 16 SCHOLL AND OTHERS underlying structural framework of this margin. The implication is that a continental margin and adjacent deep-water basin of some type have existed in this region for at least the last 150 m.y. Heat flow over the floor of Kamchatka Basin is higher than normal, a circum­ stance that may indicate subbasin injection of basaltic magma (Langseth and von Herzen, 1971; M. G. Langseth, 1972, written commun.; Cormier, 1972). Cormier (1972) further notes that spreading, which need not involve migration of ridges, may be associated with magma emplacement, and that evidence for this is reported compressional structures in the adjacent coastal area of northern Kamchatka. The last significant deformation of Mesozoic and Tertiary deposits underlying the coastal area was in the middle Miocene through early Pliocene; folding was not intense, and the associated tectonic motion may have been dominantly uplift rather than com­ pression (Tilman and others, 1969; Drabkin, 1970). Reflection records reveal no regional compressional folding in thick (1 to 2 km) late Miocene through Holocene beds deposited at the base of the continental margin, which is underlain by deformed Cenozoic beds thought to be mostly of pre-Pliocene age. Also, the sedimentary section of late Miocene through Holocene beds that forms the upper 1 to 2 km of Kamchatka Basin is a continuous blanket unbroken by major extensional ruptures. We interpret these observations to mean that the basin floor has undergone no signi­ ficant internal spreading or rifting during at least the last 5 m.y. This conclusion also means that subbasin intrusion of late Cenozoic magma, as suggested by heat- flow data, was not associated with significant crustal dilation or underthrusting. Cormier’s (1972) interesting idea of spreading within Kamchatka Basin should also be tested for the Paleogene and early Neogene. Of special importance to this possibility was the discovery at DSDP site 191 (Figs. 1, 5) that virtually the whole of the middle Tertiary is missing (30 to 5 m.y. ago). Because the drilling site was located over the summit of a basement knoll, where the overlying late Cenozoic deposits were exceptionally thin, it cannot be concluded that older Tertiary deposits are therefore absent from the basin. For example, the wide-angle reflection data of Ludwig and others (1971b) can be interpreted as showing that older, higher velocity (2.5 to 3.0 km per sec) sedimentary deposits overlie the igneous crust in swales between basement highs; nonetheless, the geological and geophysical data that we have examined indicate that the bulk of the basin’s sedimentary fill is of late Miocene and younger age, and that older Tertiary deposits are either thin, sparsely distributed, or perhaps absent. The basin’s thin or missing middle Tertiary section requires either a very slow sedimentation rate, sweeping of a normal thickness of these deposits into a subduction zone, or in situ engulfment by contemporaneous or younger basalt. In the Paleogene and early Neogene approximately 10,000 m of dominantly volcanic and volcaniclastic beds accumulated in Oliutorsky Depression, a géosyn­ clinal trough underlying the coastal and inner shelf regions of northern Kamchatka (Tilman and others, 1969; Drabkin, 1970; Avdeiko, 1971). This downwarp, which lies along the western side of Kamchatka Basin, was last significantly deformed in middle Miocene through early Pliocene time. This depression conceivably trapped sufficient sediment to starve the adjacent Kamchatka Basin until post-late Miocene STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 17 time, when uplift shunted terrigenous detritus to the basin./The sediment-limiting hypothesis means that most of its basaltic crust is of Cretaceous age, although limited magmatism must have occurred in the early Tertiary. Formation of the downwarp and its great masses of volcanic rock can also be linked to magmatism and spreading in Kamchatka Basin, a process of crustal genera­ tion that caused westward subduction of oceanic plate until the late Miocene or early Pliocene. Westward underthrusting requires that Shirshov Ridge, on the oppo­ site side of the basin, was either situated above a complimentary subduction zone dipping eastward beneath it, or that an eastward-migrating spreading center occupied the basin floor. Shirshov may also have been a stationary or eastward-migrating spreading ridge characterized by asymmetric crustal generation, that is, chiefly to the west. For reasons we discuss later, we do not believe that Shirshov Ridge was ever a spreading center, let alone a migrating one. However, in the early Tertiary it was a linear volcanic “arc,” and conceivably the magmatic product of a subducted oceanic slab generated in Kamchatka Basin. Although we do not dismiss this model, the occurrence of middle Oligocene basalt at DSDP site 191, which is near the western side of the ridge, makes it difficult to believe that sufficient plate could have been created by a spreading center in this relatively small basin to have spawned the magmatism required to construct Shirshov Ridge and the thick, mafic, and inter­ mediate volcanic masses of Oliutorsky Depression and adjacent areas. Kienle (1971) has also observed that a major gravity low suggestive of a filled trench does not flank the ridge. Considerations of the several geological and geometrical difficulties involved in forming Kamchatka Basin by either crustal spreading or ridge migration led Scholl and Creager (1973) to suggest that emplacement of the basaltic crust may have been a nonspreading process of in situ sea-floor magmatism. They hypothesize that the basin has maintained much of its present size since the earliest Tertiary, when Shirshov Ridge formed, but that the bulk of early Tertiary and older deposits over- lying still older oceanic crust was in part or wholly engulfed by the subsequent emplacement of Cenozoic basalt. We can suggest further that the principal episode of subbasin magmatism may have been contemporaneous with the volcanic growth of Shirshov Ridge and the emplacement of the voluminous Paleogene and early Neogene effusives of Oliutorsky Depression. It should be pointed out that Cenozoic volcanism was by no means restricted to the Oliutorsky sector of the Koryak-Kam- chatka region, which borders the whole of the western side of the Bering Sea Basin. Kamchatka Basin can therefore be viewed as originally underlain by Pacific crust of Mesozoic age that was isolated north of the Pacific region by the formation of the Aleutian Ridge in earliest Tertiary time. Subsequent or contemporaneous vol­ canic growth of Shirshov Ridge in turn isolated Kamchatka Basin from the Aleutian Basin. Subbasin injection of basaltic magma ingested much or all of the underlying Mesozoic crust and superimposed sedimentary cover. Magmatism presumably con­ tinued throughout Paleogene and early Neogene time but diminished greatly after or in the late Miocene. This model of in situ magmatism implies steady crustal thickening during most of the Tertiary, a factor that can be evaluated by crustal refraction studies, and that the gentle basement knoll sampled at DSDP site 191 is 18 SCHOLL AND OTHERS a basaltic “nunatak” of Oligocene age rising above younger subbasin flood basalt, a relation that can be tested by future drilling. Subduction models have proved successful for explaining the emplacement of Cenozoic basalt beneath marginal seas of the western Pacific (Karig, 1970, 1971, 1972; Oxburgh and Turcotte, 1970, 1971; Matsuda and Uyeda, 1971). If the under­ lying cause of Cenozoic magmatism in Kamchatka Basin is also linked to oceanic plate subduction along the Aleutian Trench, then a number of problems are en­ countered. When underthrusting occurred the relative direction beneath the Aleutian Ridge was to the north or northwest. Since the Eocene (Pitman and Hayes, 1968; Hayes and Pitman, 1970), or at the latest the middle or late Oligocene (Grow and Atwater, 1970), subduction has not taken place west of about 176° E., which includes the segment of the Aleutian Ridge that forms the southern boundary of Kamchatka Basin. Thus, none of the several models of plate motion proposed for this region isolates Kamchatka Basin as the sole or even the principal area north of the ridge that would be magmatized by a subducted oceanic lithosphere. Perplex- ingly, subbasin inflow of basaltic lava should have occurred beneath the Aleutian Basin, if anywhere. Perhaps widespread magmatism did take place there, in the early Tertiary, but most certainly significant spreading or ridge migration was not a result. Equally puzzling, magmatism in Kamchatka Basin presumably diminished mark­ edly after the late Miocene, a time during which relative motion between the Pacific crust and that of Bering Sea Basin was rapid if not accelerated over middle Tertiary rates (Atwater, 1970; Grow and Atwater, 1970; Hayes and Pitman, 1970; Marlow and others, 1973). It is important to note here that the existence of the eastward- flowing mantle current suggested by Nelson and Temple (1972) might possibly explain a concentration of magma upwelling over the western part of the Bering Sea Basin, that is, the Shirshov Ridge and Kamchatka Basin areas. Their model interests us, but its implications for a marginal sea lying north of an east-trending arc-trench system are difficult to assess. Obviously, only additional geophysical studies and deep subbottom sampling will correctly outline the history of volcanism in the Bering Sea Basin and provide an opportunity to understand the underlying magmatic mechanism.

Uncertain Tectonic Significance of Bowers and Shirshov Ridges

The bedrock core of Shirshov Ridge is in part composed of lithified andesitic tuff (about 60 percent Si02) and albitized palagonite tuff. The K-Ar age of un­ altered tuff collected from the summit of the southern terminus of the ridge is 16.8 m.y. (early Miocene), an age based on the average of whole-rock (15.6 m.y.) and plagioclase (18 m.y.) analyses. Albitized volcanic and thoroughly lithified volcaniclastic rocks have been dredged from the summit of Bowers Ridge, where its bedrock core is exposed. These rocks are similar in all respects to those typical of the early series underlying the connecting Aleutian Ridge; this series is typically of early Miocene and older age. Bathymetrically, the distal ends of Bowers and Shirshov Ridges nearly meet (Fig. 1). The western segment of Bowers Ridge is low lying, narrow, and convexed STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 19 toward the south. Similarly, the eastward-curving southern end of Shirshov Ridge, where andesitic tuff of early Miocene age occurs, is narrow, deeply submerged, and convexed toward the south. Our reflection records and those published by Ludwig and others (1971b) reveal that beneath the flat floor of the Aleutian Basin an equally narrow ridge connects Bowers and Shirshov Ridges. The connecting subsurface ridge has deformed overlying turbidite and pelagic beds of late Cenozoic age. These same beds have been uplifted against the flanks of the two bathymetric ridges (Fig. 5, profile C-D). The main parts of Bowers and Shirshov Ridges are therefore struc­ turally connected by a narrow, arcuate ridge—a ridge that is in part composed of andesitic rocks of early Miocene age and a ridge that in the late Cenozoic rose tectonically relative to the adjacent abyssal floor of the Aleutian Basin. The bedrock cores of Shirshov and Bowers Ridges are overlain by stratified deposits as thick as 2,000 m (Fig. 5; Ludwig and others, 1971b). Our dredge sam­ ples, those of Hanna (1929), and drilling at DSDP sites 188 and 190 (Figs. 1, 5) indicate that the upper part of this stratified sequence is composed largely of diato- maceous beds (pelagic deposits) of late Miocene and younger age. These deposits overlie mudstone of middle(?) and late Miocene age. The sedimentary sequence is little deformed, but in perched or crestal basins along the summits of both ridges, the basal beds have been broadly folded and offset by differential movements of the underlying basement (Fig. 5). Basement highs along the crests of both ridges have been beveled to relatively smooth surfaces of low relief. These guyotlike areas are now submerged beneath 500 to 1,000 m of water; they apparently signify crestal erosion followed by or in conjunction with ridge subsidence (Ludwig and others, 1971b). This implication was realized earlier by Udinstev and others (1959), who (on the basis of bathymetric data) thought these erosionally flattened areas were very extensive over the north­ ern reaches of Shirshov Ridge. However, reflection records show that in most areas the ridge’s broad, smooth summit has resulted from the smothering of basement relief by the accumulation of a thick pile of pelagic and hemipelagic deposits (Fig. 5). The relations outlined above suggest a geologic history for both Shirshov and Bowers Ridges that begins with formation by volcanic processes probably in early Tertiary time, uplift, cessation of volcanism, and erosion in the middle and late Miocene, and subsequent and contemporaneous accumulation of a thick blanket of hemipelagic and pelagic deposits over a differentially subsiding bedrock core. Except for the notable absence of volcanism in the late Neogene, the eastern or main part of Bowers Ridge appears to have had a geologic history similar to that of the Aleutian Ridge, to which it is structurally connected. In terms of the expected magmatic consequences of plate tectonics, it is difficult to generate Shirshov and Bowers Ridges by a simple scheme of plate interactions. Karig (1972), by analogy with volcanic lineaments in the western Pacific, has specu­ lated that Bowers and Shirshov Ridges are remnant arcs. In the western Pacific remnant arcs are associated with ridge migration and crustal spreading. Bowers Ridge is certainly a type of volcanic arc, but one that is associated with a sediment-filled trough (trench) below its northern or convex slope (Ludwig and others, 1971a; Kienle, 1971). As Kienle (1971) notes, an arc-trench pair implies convergence be­ tween lithospheric plates. Thus, either low-angle reverse faulting, which Nichols and 20 SCHOLL AND OTHERS others (1964) previously considered likely, or underthrusting, probably from the northeast (Kienle, 1971), of the subduction type formerly took place along the northern or convex side of the ridge. Pitman and Talwani (1972) offer the intriguing speculation that underthrusting from the northeast was related to spreading in the North Atlantic that caused a southward (transcurrent) motion of Alaska relative to the Eurasian plate. Under­ thrusting during the Cenozoic resulted from a 500-km displacement between the two continental areas. We are unaware of geologic evidence for this displacement, which would have taken place principally beneath the Bering Sea. Also, we believe that Bowers Ridge and the Aleutian Ridge, which are structurally connected, are part of the American plate that comprises Alaska. Southward drift of Alaska would have been accommodated by subduction of Pacific lithosphere at the Aleutian Trench rather than Bering Sea crust along the northern side of Bowers Ridge. We prefer to speculate that Bowers Ridge formed as an arcuate outgrowth of the Aleutian Ridge and that the magma-producing mechanism for both ridges was much the same. Perhaps slight spreading in Bowers Basin during either the Late Cretaceous or earliest Tertiary forced the growing Bowers Ridge generally northward over the oceanic crust of the Aleutian Basin, thereby creating a short-lived subduction zone and corresponding trench. Magnetic data suggests to Kienle (1971; 1971, personal commun.) that the nearly linear Shirshov Ridge may be a fossil-spreading center. We have already dis­ cussed the implications of this important idea in accounting for the Paleogene age of the basaltic crust underlying Kamchatka Basin. If it were a site of crustal genera­ tion, the substantial thickness of Cenozoic deposits burying the bedrock core of its summit and eastern (Aleutian Basin) flank stipulates that spreading ceased in or prior to the middle Miocene (Fig. 5). Three geomorphic and geologic aspects of this ridge signify to us that it was not a spreading ridge even before this time. First, the ridge, as noted by many workers, is the geomorphic extension of a belt of on­ shore volcanic rocks of Late Cretaceous age. These severely deformed eugeosynclinal rocks underlie the southwestern coast of the Koryak Mountains and strike seaward at Cape Oliutorsky (Fig. 1; Tilman and others, 1969; Drabkin, 1970; Avdeiko, 1971). The close relation of the two volcanic trends, one apparently slightly older than the other, requires a fortuitous alignment of greatly differing structures if the ridge is a former spreading center. Second, within a short distance, generally less than 150 km, the ridge’s basement core rises 3 to 5 km above that of the basaltic crust of the flanking Kamchatka and Aleutian Basins, a relief that is uncharacteristic of active spreading centers, let alone one that could not have been the site of crustal generation during at least the last 15 to 20 m.y. And third, all samples of the bed­ rock core that we have collected are fragmental volcanic rocks of andesitic composi­ tion, and most of these are albitized; these altered intermediate rocks are more typical of island arcs than the tholeiitic magma extruded at spreading centers. We speculate that Shirshov Ridge is a former volcanic arc, now deeply submerged, that probably represents an early Tertiary continuation of the Late Cretaceous volcanic trend established in the Oliutorsky region. In terms of plate tectonics, linear volcanic welts near continental edges are generally interpreted as the mag- STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 21 matic consequences of crustal subduction (Hasebe and others, 1970; Matsuda and Uyeda, 1971; Oxburgh and Turcotte, 1970,1971), a model that is awkward but not impossible to apply to Shirshov Ridge (Kienle, 1971). Like Bowers Ridge, to which it is structurally connected, the pattern of early Tertiary plate tectonics associated with the magmatic growth of Shirshov Ridge is difficult to deduce.

BERINGIAN CONTINENTAL MARGIN

General Background

Sweeping in a 2,400-km-long arc from Cape Kamchatka to near the tip of the Alaska Peninsula, the Beringian continental margin separates the deeps and shelves of the Bering Sea (Fig. 1). The term “margin” is meant to include only the outer part of the shelf, the flanking continental slope, and the upper reaches of the con­ tinental rise. The height, from continental rise to shelf break, of the steeper part of the margin is about 3,000 m. It has the typical average steepness (3° to 7°) of most continental slopes. In many areas it is deeply incised by submarine canyons, some of which number among the largest in the world (Scholl and others, 1970b). Bathy­ metric descriptions of the Bering margin have been presented by Udinstev and others (1959), Gershanovich (1963), Kotenev (1965), Lisitsyn (1966), and Scholl and others (1968). Eastern or Pribilof Segment. The structure of the eastern or Pribilof segment of the Beringian margin, which is its northwest-trending segment connecting Siberia and Alaska (Fig. 1), consists of two relatively low-velocity rock units overlying a seaward-thinning continental crust (Fig. 3; Shor, 1964). The upper unit (1.7 km per sec), informally termed the “main layered sequence” by Scholl and others (1968), consists of semiconsolidated terrigenous and diatomaceous deposits. Although beds as old as middle Oligocene have been dredged, the sequence is chiefly of Neogene age (Hopkins and others, 1969). They are little deformed except near major sub­ shelf faults or where downslope slumping has occurred. Over the continental slope the main layered sequence is typically less than 500 to 1,000 m in thickness. Be­ neath the outer shelf the sequence thickens to more than 3,000 m in depressions associated with faulting or warping in the underlying and far more lithified second rock layer (3.2 to 3.6 km per sec). The surface of this latter unit commonly corre­ sponds to the acoustic basement on seismic reflection profiles (Scholl and Hopkins, 1969; Fig. 7). These basins are associated with negative free-air gravity anomalies (Gaynanov and others, 1968; Kienle, 1971). Rocks dredged from exposures of the lithified layer on the continental slope are fractured and sheared flysch-type deposits of mudstone, siltstone, and sandstone of Late Cretaceous (Campanian) age (Hopkins and others, 1969). Reflection records reveal that these rocks are tightly folded beneath the outer shelf (Scholl and Buff­ ington, 1970), where they are apparently intruded by ultramafic masses as well (Barth, 1956). The width of the fold belt is unknown, but presumably it underlies much of the Bering shelf south of Nunivak Arch (Fig. 7). The trend (or trends) of 22 SCHOLL AND OTHERS this obviously complex fold belt is also unknown, but seismic profiles suggest that it is roughly parallel to the margin. This is also implied by the northwest trend of the isopach lines on Figure 7, which show that Cenozoic basins are elongated ap­ proximately parallel to the margin. It seems reasonable to suspect that this Cenozoic fabric is in part inherited from that of the structural framework of the underlying Mesozoic rocks. Prior to their burial beneath the Oligocene and younger beds of the main layered sequence, the tops of the folds involving Cretaceous (and earliest Tertiary?) beds were eroded off (Figs. 3, 6). The eastern segment of the Beringian margin is there­ fore underlain by a prominent angular unconformity—an unconformity that sepa­ rates pre- and postorogenic deposits (that is, presumably Mesozoic and Cenozoic beds, respectively). Near the shelf edge, the unconformity flexes downwarp, roughly parallel to the inclination of the continental slope, and crops out in canyon bottoms at depths between 500 and 2,500 m (Fig. 3). At greater depths the unconformity, recognized as an acoustic basement, passes beneath 2,000 to 3,000 m of virtually undeformed deposits underlying the continental rise. It is unlikely that an angular unconformity or any significant depositional hiatus separates Mesozoic and Ceno­ zoic beds underlying the central area of the Aleutian Basin. Western or Siberian Segment. The Siberian segment of the Beringian continental

Figure 7. Trends o f late Mesozoic geosynclines, Alaska and Koryak regions, and thickness of Cenozoic deposits over outer part o f Bering shelf. Isopachs are based chiefly on seismic- reflection data and were computed as if the entire section had an acoustic velocity o f 2.0 km per sec. Data in the Bristol Basin area were taken from Hatten (1971). Isopachs can also be read as approximately equivalent to structural contours on rocks o f Mesozoic age. A prominent angular unconformity separates Mesozoic and Cenozoic deposits. STRUCTURAL EVOLUTION, A LEUTLA N-B ER ING SEA REGION 23 margin, from Cape Navarin to Cape Kamchatka (Fig. 1), is closely flanked by a topographically high coastal area, an area underlain by deformed eugeosynclinal deposits chiefly of Jurassic to Late Cretaceous or possibly early Tertiary age (Egia- zarov, 1963; Vlasov, 1964; Gladenkov, 1964; King, 1969; Tilman and others, 1969; Bogdanov, 1969, 1970; Drabkin, 1970; Avdeiko, 1971). The geomorphic setting of this western margin is therefore considerably different from that of the eastern segment, which is as far as 550 km from the nearest continental shore. Both areas are geologically similar in that a major angular unconformity separates Mesozoic and Cenozoic beds. But the Siberian segment has undergone repeated uplift and mild deformation in the Cenozoic; marine beds of this age are therefore exposed in coastal outcrops. The strongest episode of Cenozoic tectonism occurred in the middle Miocene through the early Pliocene (Gladenkov, 1964; Tilman and others, 1969; Drabkin, 1970), a time that corresponds closely to the uplift, mild deforma­ tion, and plutonism of the Aleutian Ridge. Seaward of the Koryak coast, which flanks the northwestern corner of the Aleu­ tian Basin, reflection profiles of the continental margin and samples of submarine outcrops indicate that generally undeformed terrigenous and diatomaceous strata of early Pliocene and younger age, the main layered sequence, drape this margin and bury the upper surface of more deformed units. Only seaward of Capes Oliutorsky and Navarin (Fig. 1) was the main layered sequence found resting unconformably on complexly folded volcanic and terrigenous rock of Mesozoic age. Along the remainder of the Koryak margin the main layered sequence is about 500 m thick and typically rests on internally deformed masses (probably by slumping and large- scale gravity sliding) of older Tertiary strata. These predominantly pre-Pliocene beds can be traced to the floor of the Aleutian Basin, where they are several kilometers thick and presumably overlie Cretaceous beds of unknown thickness and lithology. Beneath the shelf, the main layered sequence thickens to more than 1 km and rests on an unknown but presumably great (>2 km) thickness of Paleogene and early Neogene terrigenous deposits. These deposits can be correlated with uplift and gently deformed neritic and coastal deposits of equivalent age underlying the Opukhsko-Pekulneisky downwarp of Gladenkov (1964). Southwest of the Koryak coast the northern Kamchatka section of the Siberian margin borders the Kamchatka Basin between Cape Oliutorsky and Cape Kamchatka (Fig. 1). This margin is characterized by an exceptionally steep continental slope (10° to 12° overall, but with segments exceeding 25° to 30°; Udinstev and others, 1959) that is underlain by deformed Tertiary beds. Sampling of submerged out­ crops and coastal mapping suggests that these beds are largely of pre-Pliocene age. Younger, generally undeformed deposits typifying the main layered sequence are absent or thin, possibly because of extensive canyon cutting in the Pleistocene and the basinward slumping of superficial deposits. At the base of the continental slope, seismic reflection records indicate that the older, more deformed slope beds either pass beneath or merge with the basal part of an undeformed sequence of basin deposits as much as 2,500 m thick. This sequence can be correlated with late Mio- cene(?) and younger terrigenous beds penetrated at DSDP site 191 (Creager, Scholl, and others, 1973). Near the base of the Kamchatka margin, these deposits overlie 24 SCHOLL AND OTHERS an acoustic basement exhibiting only minor relief. This basal reflector is either underlain by thoroughly lithified sedimentary beds of Miocene and older age, or, as we previously remarked, basalt of Cenozoic age that presumably has buried older deposits.

Plate Tectonics and Marginal Deformation

Implications of a Cretaceous Aleutian Ridge. Geologic studies in the Koryak- Kamchatka region have demonstrated that it is underlain by large eugeosynclinal masses mostly of Late Jurassic and Cretaceous age (Egiazarov, 1963; Gladenkov, 1964; Krasny, 1964; Vlasov, 1964; Tilman and others, 1969; Bogdanov, 1969, 1970; Drabkin, 1970; Avdeiko, 1971; Gnibidenko and others, 1973). These highly deformed assemblages of volcanic, terrigenous, and ultramafic rocks were deposited on oceanic crust (Avdeiko, 1971) and form the ensimatic or thalassogeosyncline of Bogdanov (1969). The late Mesozoic rocks of this area are much like those of the Franciscan Formation of California and Oregon, an assemblage of rocks that most geologists believe accumulated in a subduction zone located over and perhaps at the base of a continental margin (Hamilton, 1969; Dewey and Bird, 1970; Ernst, 1970; Page, 1970a, 1970b; Hsu, 1971). Burk (1965) has previously noted the lithologic and stratigraphic similarity be­ tween the deformed eugeosynclinal masses of the Koryak-Kamchatka region and those of southern Alaska, suggesting that the two regions were formerly a continu­ ous region of geosynclinal sedimentation associated with a continental margin and perhaps an adjacent deep-sea trench. His conclusion has certainly been strengthened by the studies of Moore (1972,1973) and the discovery of an ultramafically intrud­ ed fold belt of Late Cretaceous rocks connecting Alaska and Siberia beneath the Pribilof segment of the Beringian continental margin (Hopkins and others, 1969; Scholl and Buffington, 1970). Viewed from the standpoint of plate tectonics, during the Late Jurassic and Cretaceous, the entire Beringian margin and the connecting Pacific fold belts of southern Alaska and Kamchatka were a continuous zone of eugeosynclinal sedimentation and magmatism, a zone associated with subduction or underthrusting of oceanic lithosphere. The existence of an Aleutian Ridge of Cretaceous age rules out the possibility that a Pacific plate underthrust any part of the Beringian margin; therefore, forma­ tion and deformation of the Mesozoic eugeosynclinal masses that underlie it were not associated with an underthrusting Kula plate. Instead, these fold belts must be ascribed to either the movements of smaller plates generated and subducted within the Bering Sea Basin or to the combined effects of spreading in the Arctic Ocean and continental drift of North America that may have formed the Alaskan oroclinal complex (Hopkins and Scholl, 1970; Coney, 1971). Because the Mesozoic eugeo­ synclinal masses ring the Bering Sea Basin only to the east, north, and west (unless the Aleutian Ridge is regarded as a southern segment of this ring), then long- continued subduction along this perimeter requires asymmetric intrabasin spreading generally away from the Aleutian Ridge, an awkward if not impossible arrangement of spreading centers and subduction zones. The double oroclinal flexing of the STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 25

Alaskan-Siberian continental crust may explain the fold belts of the Beringian mar­ gin, as related to its steady closing or collapsing upon the oceanic crust of the basin (Hopkins and Scholl, 1970); however, their structural connection with the Pacific fold belts of equivalent age that lie outside the oroclinal area is evidence that hori­ zontal crustal flexing is not the principal orogenic mechanism that formed the Bering Sea belts. This argument is equally applicable against the likelihood that intrabasin spreading formed them. Implications of a Cenozoic Ridge. A major change in the style of marginal tectonism, from orogenic folding attended by ultramafic intrusions to marginal subsidence unmarked by significant deformation or plutonic invasion, can be re­ lated to a change in relative plate motion at the base of the continental margin (Fig. 6). If we assume that the Aleutian Ridge did not exist prior to latest Creta­ ceous or earliest Tertiary time, then in the middle and late Mesozoic the Kula plate extended into the area of the Bering Sea and terminated against the North American- Eurasian plate along the line of the then-existing continental margin (Fig. 6). In the general manner that Bailey and Blake (1969), Hamilton (1969), Ernst (1970), Page (1970a, 1970b), Bailey and others (1970), and Hsu (1971) view the tectonic history of the Mesozoic margin of western North America, we can ascribe the formation and the subsequent folding and intrusion of the Late Jurassic through Late Creta­ ceous geosynclinal complex of the Koryak-Kamchatka region to the consequences of rapid plate convergence here. By arranging the vector of relative motion toward the northwest (Pitman and Hayes, 1968; Grow and Atwater, 1970; Larson and Pitman, 1972), a high rate of underthrusting, probably exceeding 6 cm per yr, can be achieved (Fig. 4). Similarly, orogenesis along the southern Alaskan margin, which parallels the Koryak-Kamchatka region and outwardly appears to have had a similar tectonic history (Burk, 1965; Jones and others, 1970; Berg and others, 1972; Plaf- ker, 1972), can also be ascribed to the effects of a rapidly underthrust continental edge. Because the northwestward trend of the eastern or Pribilof segment of the Bering­ ian margin may have been more nearly parallel to the prescribed plate motion, less plate would have been subducted here. We might expect, therefore, that Mesozoic eugeosynclinal processes along the Pribilof segment were less complete than those affecting the adjoining but more directly underthrust Alaskan and Siberian margins. A thinner geosynclinal mass may have resulted, a circumstance that could account for the lack of any subaerial remnant of the fold belts formed here in the Mesozoic. Avdeiko (1971) has been able to identify a seaward younging of deformed eugeo­ synclinal systems underlying the Koryak-Kamchatka region (from Late Jurassic to Paleogene), and Payne (1955) has identified a generally analogous trend in southern Alaska. It seems likely that a similar pattern of marginal growth, or continental accretion, has also taken place beneath the Pribilof segment of the Beringian margin. The amount of accretion is unknown, but we speculate it could be as much as 200 km (Fig. 8). The last episode of significant growth took place in latest Cretaceous or earliest Tertiary time (Laramide?), and involved the deformation of sedimentary deposits at least as young as Campanian. It is of interest to note that the southwest trend of the Cretaceous geosynclinal 26 SCHOLL AND OTHERS rocks of southwest Alaska is directly toward the Pribilof segment of the Beringian margin (that is, the Koyukok and Kuskokwim geosyncline; Payne, 1955; Miller and others, 1959; Hoare, 1961; Gates and Gryc, 1963). These more interior geosynclines, which could not have formed at a continental margin, must therefore either narrow and vanish before reaching the margin, perhaps against Nunivak and Goodnews Archs (Fig. 7), or turn to the northwest and merge with the Pacific fold belts under­ lying and paralleling the existing margin. The initiation and rapid growth of the Aleutian Ridge in latest Cretaceous or early Paleogene time are presumed to imply that at this time the convergence or subduction zone for the Kula plate shifted southward to the Aleutian Trench (Fig. 4) For whatever reason the relocation took place, the new convergence zone aligned itself with an older one (that of Alaska Peninsula) off southern Alaska and thereby tectonically “short circuited” the entire Bering continental margin from further interaction with the oceanic plate. Perhaps because of the relaxation of compres- sional stresses, marginal fold belts of Mesozoic and possibly early Tertiary age sub­ sided below sea level, and a regional unconformity was cut across their tops (Figs. 6, 7). Cenozoic collapse or subsidence of the outer Bering shelf to form part of the existing margin can also be linked to sedimentary loading of the adjacent deep basin of the Bering Sea.

SUMMARY AND CONCLUDING REMARKS

Favored Plate Model

Believing that nature abhors complexity, we favor the simplest scheme of plate tectonics that can account for the major aspects of the structural and magmatic evolution of the Aleutian-Bering Sea region. Our favored model therefore calls for the formation of the Aleutian Ridge in latest Cretaceous or earliest Tertiary time. Although the ideas presented here were outlined earlier by Scholl and Buffington (1970) and Hopkins and Scholl (1970), a conceptually similar model has been devel­ oped independently by Jones (1971). His model includes the interesting speculation that the northward-migrating Kula spreading ridge may have been transformed into a subduction zone. Such a transformation would mean that the ridge, or fragments of it, could have served as the nucleus for the formation of the Aleutian Ridge. Principally because we find it difficult to imagine how to sink (subduct) a freshly created and therefore “hot” oceanic lithosphere, we have not used this aspect of Jones’ (1971) model on Figure 6, which summarizes our concept of the geologic development of the Aleutian-Bering Sea region. Prior to formation of the Aleutian Ridge, the lithospheric plate of the North Pacific (Kula) collided with that of the North American-Eurasian crust along a line approximately paralleling the position of the existing continental margin connecting these two continents. We further postulate that the relative motion between these plates was generally similar (that is, to the north or northwest) to that proposed by Pitman and Hayes (1968), Atwater (1970), Grow and Atwater (1970), and Larson STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 27

Figure 8. Late Mesozoic structural trends, Alaska and southeastern Siberia, and the trend of possible connecting fold belts beneath the Pribilof segment o f the Beringian continental margin. Dotted lines approximately delimit the inner edge of Pacific fold belts, which are under­ lain by eugeosynclinal rocks that formed in close association with a tectonically active conti­ nental margin. The interior belts o f geosynclinal rocks, those o f western Alaska and the Bering shelf, either merge with or are cut off by the trends of the Pacific belts. and Pitman (1972). During middle and late Mesozoic time, this margin and the contiguous continental edges of southern Alaska and eastern Kamchatka were the sites of eugeosynclinal sedimentation and repeated orogenic pulses involving folding, faulting, and the emplacement of ultramafic masses—an idea advanced originally by Burk (1965). We visualize eugeosynclinal sedimentation and magmatism as taking place over a continental or insular slope descending seaward of a volcanic arc, or in an inner-arc basin associated with a volcanic arc rising seaward of the then-existing continental margin. Trench sedimentation may also have been involved, but for reasons outlined elsewhere (Scholl and Marlow, 1974a, 1974b), we do not believe that any significant part of the eugeosynclinal deposits accumulated in trenches. As outlined by Dietz (1963a), marginal orogenesis is presumed to have resulted from the mechanical and magmatic interactions of the converging lithospheric plates. Marginal orogenesis virtually ceased north of the Aleutian Ridge in early Tertiary time. This change in tectonic mechanism, from predominantly laterally directed to vertically directed stresses, is thought to correlate with the initial formation of the Aleutian Ridge and, most importantly, to a southward relocation of the convergence zone to the Aleutian Trench seaward of the ridge. Formation of a new and oceanic convergence zone at the Aleutian Trench tectonically isolated or “short circuited” the Bering continental margin from further interaction with the underthrusting edge of the oceanic plate. This initiated a nonorogenic episode of marginal collapse and fragmentation, save for modest uplift and folding in the Kamchatka-Koryak region that was most pronounced in middle Miocene through early Pliocene time. We therefore propose that pre- and postorogenic deposits of the continental margin corresponded, respectively, to pre- and post-Aleutian Ridge time. 28 SCHOLL AND OTHERS

The above generalization is not completely applicable to the northern part of Kamchatka, the area of Oliutorsky Depression. Eugeosynclinal sedimentation and magmatism did not terminate here in the Late Cretaceous but continued through the Oligocene and perhaps the early Miocene. This “anomalous” segment of con­ tinental margin borders Kamchatka Basin, the “anomalous” western region of the Bering Sea Basin characterized by a thin sedimentary fill and Cenozoic magmatism. We suppose that a common or closely related lithospheric mechanism is responsible for the Cenozoic magmatic histories of these two unusual but adjoining regions, and that a related mechanism formed Shirshov Ridge, which separates them from a more “normal” Bering Sea Basin to the east and continental margin to the northeast. Diminishing of coastal and submarine volcanism toward the end of Miocene time may have been associated with coastal uplift in middle Miocene through early Plio­ cene time. It must be emphasized that Cenozoic volcanism was not restricted to the Oliutorsky area but was general throughout the Koryak-Kamchatka region, as was the mild Miocene-Pliocene tectonism. In contrast, Cenozoic uplift and volcanism affected only small areas of the broad Bering Sea shelf that borders Alaska, where coastal exposures of marine beds of Tertiary age are virtually unknown. Although Pacific lithospheric plate ceased to underthrust the Beringian margin after about Late Cretaceous-early Tertiary time (that is, 60 to 65 m.y. ago), the Kula plate underthrust the growing Aleutian Ridge until sometime in the Eocene (about 45 m.y. ago). At this time the Kula ridge was subducted, and underthrusting either ceased or greatly diminished (Pitman and Hayes, 1968; Hayes and Pitman, 1970; Marlow and others, 1973; Clague and Jarrard, 1973). Correspondingly, volcanic growth of the ridge diminished, and a great volume of erosional debris accumulated over its summit and flanks. Important volcanism may not have begun again until the middle Miocene, when the ridge was plutonized and uplifted. We can associate these events with the slow beginning of a new episode of late Cenozoic underthrusting (Marlow and others, 1973; Clague and Jarrard, 1973), an episode that has culminated in the formation of the magnificent stratovolcanoes of Kam­ chatka and of the Aleutian Islands. Nonetheless, as discussed below, many problems and uncertainties are associated with plate tectonics and the history of magmatism in the Aleutian-Bering Sea region.

Unresolved Problems and Other Complications

The plate model outlined above accounts for major changes in tectonic style observed in the Aleutian-Bering Sea area during most of late Mesozoic and Cenozoic time. However, our model and the plate reconstruction schemes of Larson and Chase (1972) and Larson and Pitman (1972) imply that crustal underthrusting during Meso­ zoic and Cenozoic time must have amounted to thousands of kilometers; yet strati- graphic and structural evidence suggest that the amount of underthrusting was not more than several hundred kilometers, at least an order of magnitude less than expected. Our model, therefore, deals most effectively with the qualitative aspects of structure and lithostratigraphic units actually observed, and least effectively with the quantities of these units that should be present. These and related problems requiring further inquiry are briefly sketched below. STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 29

1. After magmatic growth of the Aleutian Ridge greatly diminished in the Eocene, a significant new episode of ridge-crest volcanism may not have begun again until the middle Miocene, when the ridge was plutonized and uplifted. Uplift and folding of the distant Koryak-Kamchatka coastal area and flanking shelf also began in the middle Miocene; Miocene volcanism was also widespread in this area (Drabkin, 1970). It is tempting to associate these tectonic and magmatic events with a new episode of underthrusting at the Aleutian Trench, probably one that began in the late Oligo- cene (Marlow and others, 1973; Clague and Jarrard, 1973). But this solution is awkward because volcanic rocks of late Miocene and early Pliocene age are not especially abundant along the Aleutian Ridge, and ash falls of this age are not abun­ dant in the sedimentary deposits penetrated at nearby DSDP drilling sites (Creager, Scholl, and others, 1973). Also, at this time, or slightly before, volcanism on Shir- shov Ridge and beneath the neighboring Kamchatka Basin markedly lessened or ceased. The slacking of magmatism in these areas suggests a quick end to the new episode of subduction. The subsequent construction of the magnificent Aleutian and Kamchatka stratovolcanoes after the early Pliocene also seems best related to the effects of rapid subduction. The initiation of subduction could have been as early as the middle or late Miocene, but seemingly was much too late to account for the middle Miocene plutonism, volcanism, and uplift. It is evident that the history of plate tectonics in the Aleutian-Bering Sea area cannot be correctly deduced from our present sketchy understanding of the history of magmatism of this vast region. 2. Conservative estimates indicate that thousands (at least 7,000 km; Larson and Pitman, 1972) of kilometers of oceanic crust must have underthrust the Koryak- Kamchatka margin during middle and late Mesozoic time. A conservative estimate of the volume of deep-sea pelagic deposits that could have been scraped off the descending plate is approximately 30 percent of the estimated volume of the eugeo- synclinal complexes of this area. Although siliceous and calcareous deposits of probable pelagic origin are known from this area, their volume is not great. More­ over, they are stratigraphically associated with terrigenous deposits, indicating that they accumulated near or at the base of a continental margin and not in an open- ocean environment. Hence, pelagic offscrapings indicative of thousands of kilo­ meters of underthrusting are not present. Scholl and Marlow (1974a, 1974b) have suggested that pelagic debris swept into the trenches of subduction zones are thrust beneath the adjacent crust rather than incorporated as offscrapings in the eugeo- synclinal mass. This is a difficult concept to imagine, and one that stipulates that trench deposits do not form a significant part of eugeosynclinal accumulations. 3. Bogdanov (1969) views the eugeosynclinal or “thalassogeosynclinal” rocks of the Koryak region as the structural and stratigraphic counterpart of the thorough­ ly studied Franciscan assemblage of California. Mapping has shown that the Fran­ ciscan rock mass has been driven landward and downward along several major thrust zones (Bailey and others, 1964; Bailey and Blake, 1969; Bailey and others, 1970; Ernst, 1970). In the Koryak region the amount of mappable telescoping is on the order of several hundred kilometers (Tilman and others, 1969; Bogdanov, 1970; Avdeiko, 1971) but probably not the many thousands of kilometers implied by the plate-tectonic mechanism. Hsu (1971) presumes that tectonic churning of the Fran­ ciscan rocks between major thrust slices (melanges) is a measure of the remaining 30 SCHOLL AND OTHERS amount of underthrusting. The geosyncJinal rocks of the Koryak region also include chaotic structures; yet major structural and stratigraphic units can be distinguished (Egiazarov, 1963; Gladenkov, 1964; Bogdanov, 1969, 1970; Drabkin, 1970; Avdei- ko, 1971)—a fact that does not imply extensive churning attributable to thousands of kilometers of underthrusting. 4. The plate model that we favor for the structural evolution of the Bering con­ tinental margin requires that its present Z-shaped path between Alaska and Kam­ chatka was established in pre-Tertiary time. This configuration allows us to speculate that the most rapid plate convergence, and presumably the greatest structural mobility, occurred beneath the thick geosynclinal masses of the Koryak-Kamchatka and the southern Alaska regions, and that the least rapid or even strike-slip motion took place along the now completely submerged and presumably structurally thinner Mesozoic rocks underlying the Pribilof segment of the Beringian margin that connects them. However, this model does not by itself account for the origin of the Z-shaped configuration of the Beringian margin. We can surmise that it is an arrangement of structural blocks that developed predominantly during (perhaps prior to) Cretaceous and early Tertiary times as a result of continental drift of North America, formation of the Alaskan orocline, spreading in the Atlantic and Arctic Oceans, or a combination of one or more of these mechanisms (Churkin, 1969, 1970; Tailleur, 1970; Tailleur and Brosge, 1970; Vogt and Ostenso, 1970; Hopkins and Scholl, 1970; Hamilton, 1970; Perry, 1971; Coney, 1971; Patton and Tailleur, 1972; Pitman and Talwani, 1972). 5. During Mesozoic time the northwest-trending Pribilof segment of the Bering­ ian margin connecting Alaska and Siberia may have been in part situated along a strike-slip (transform) boundary between oceanic and continental plates. If part of the continental plate became attached to the oceanic plate, then a boundary fault system similar to that of the San Andreas (Atwater, 1970) may have developed along the Beringian margin. It is possible, therefore, that tectonic slices of structur­ ally and lithologically dissimilar terranes may have been juxtaposed in this area— perhaps in a manner similar to that proposed for California (Hamilton, 1969; Page, 1970b; Atwater, 1970) and southern Alaska (Jones and others, 1970; Richter and Jones, 1970; Berg and others, 1972). Accordingly, the structural unraveling of the preorogenic rocks underlying the eastern segment of the Beringian continental margin may prove to be an exceptionally difficult task.

ACKNOWLEDGMENTS

We thank Arthur Grantz, Michael Churkin, Jr., Edgar H. Bailey, David M. Hopkins, and Eli A. Silver, U.S. Geological Survey, and Benjamin M. Page, Stanford University, for helpful discussions and suggestions. Critical readings of the manu­ script were contributed by Ernest H. Lathram and William W. Patton, Jr., U.S. Geological Survey. Paleontological age assignments of rocks dredged from the floor of the Bering STRUCTURAL EVOLUTION, ALEUTIAN-BERING SEA REGION 31

Sea were generously provided by Lloyd H. Burckle, Lamont-Doherty Geological Observatory (diatoms), Weldon W. Rau, U.S. Geological Survey (Foraminifera), and Warren O. Addicott, U.S. Geological Survey (mollusks). Brief petrologic examina­ tions of many of the recovered rocks were carried out by Norman S. MacLeod, U.S. Geological Survey. We are grateful to these people and numerous others for their contributions to this manuscript. Printed in U.S.A.