Abstract The thick sedimentary sequence (2-12 km [6500-39,400 ft]) underlying the abyssal floors (3-4 km [9800-13,100 ft]) of the Bering Sea Basin is shal-lowly (<360 m [<1181 ft]) underlain by large (>2 km [>6500 ft] in diameter, ~200 m [~656 ft] thick) deposits of concentrated methane hydrate. Mound-shaped bodies of hydrate are displayed on seismic reflection records as velocity-amplitude anomaly (VAMP) structures imaged as velocity pull-ups overlying pushdowns. The VAMPs are numerous (hundreds to thousands) and occur across an area of approximately 250,000 km 2 (96,525 mi 2 ). The abundance of VAMP structures is conjectured to be a consequence of high rates of basinwide planktonic productivity; of preservation of organic matter; biosiliceous sedimentation; of silica diagenesis; and of high heat flow; and deposition of a thick (700-1000 m [2296-3281 ft]), upper section of perhaps latest Miocene but mostly glacial-age (early Pliocene and Quaternary) turbidite beds and diatom ooze. Stacking of this upper Cenozoic sequence of water-rich beds heated underlying diatomaceous deposits of Miocene and older age and enhanced the generation of thermogenic methane and the diagenetic conversion of the opal A of porous diatom beds to the denser and contractionally fractured opal-cristobalite tridymite phase of porcellaneous shale. Silica transformation expelled large volumes of interstitial and silica-bound water that, with methane, ascended through the shale via chimneys of fracture pathways to enter the porous (~60%) upper Cenozoic section of diatom ooze and turbidite beds. Ascending methane entered the hydrate stability field at approximately 360 m (1180 ft), above which concentrated deposits of methane hydrate formed as either pore-filling accumulations or more massive lenses. The deposition of high-velocity methane hydrate above a multitude of chimney structures transporting low-velocity, gas-charged fluids toward the sea floor is posited to account for the widespread recording of VAMP structures in the Bering Sea Basin.

Abstract Velocity-amplitude anomalies (VAMPs), comprising coincident seismic traveltime anomalies and gas bright spots, are features widely identified in seismic reflection images from the deep-water Bering Sea basins. Interval traveltime anomalies are used to develop a method for the objective detection and quantification of these features. The approach selected uses relative traveltime variation in the sedimentary intervals above and below the gas-hydrate bottom-simulating reflector (BSR) as a diagnostic, measuring pull-up in the hydrate stability zone and push-down in the underlying gas zone, relative to a 400-common-depth-point (CDP) running-average interval reference. The method is used to explore the distribution of gas and hydrate indicators within a 120-km (74-mi) reflection profile segment in the central Aleutian Basin. This study segment includes 17 detected VAMPs, only 6 of which appear to contain significant quantities of stored hydrate. The total estimated volume of natural gas stored within the hydrate caps of these VAMPs is about 4 tcf (0.1 tcm). The largest three VAMP features contain greater than 85% of that total. Not all of the most visually obvious VAMPs are important hydrate contributors. We suggest that VAMPs are fluid-expulsion features that have become involved in the transport of natural gas. As such, the VAMP systems should have been more active in the past. The VAMPs with significant hydrate present are most likely to be active today. The largest VAMPs, including all of those associated with hydrate indicators, are located above prominent basement highs. This association suggests that fluid-migration patterns in these undeformed deep-water basins were originally established in response to sedimentation and compaction over basement topography, and that those ancient patterns have never been superceded. It also suggests that to make an informed estimate of gas-hydrate total volumes for the deep-water Bering Sea, the regional relationship between VAMP hydrate concentrations and basement topographic highs needs to be considered.

Abstract Arc magmatism at subduction zones (SZs) most voluminously supplies juvenile igneous material to build rafts of continental and intra-oceanic or island arc (CIA) crust. Return or recycling of accumulated CIA material to the mantle is also most vigorous at SZs. Recycling is effected by the processes of sediment subduction, subduction erosion, and detachment and sinking of deeply underthrust sectors of CIA crust. Long-term (>10–20 Ma) rates of additions and losses can be estimated from observational data gathered where oceanic crust underruns modern, long-running (Cenozoic to mid-Mesozoic) ocean-margin subduction zones (OMSZs, e.g. Aleutian and South America SZs). Long-term rates can also be observationally assessed at Mesozoic and older crust-suturing subduction zone (CSSZs) where thick bodies of CIA crust collided in tectonic contact (e.g. Wopmay and Appalachian orogens, India and SE Asia). At modern OMSZs arc magmatic additions at intra-oceanic arcs and at continental margins are globally estimated at c . 1.5 AU and c . 1.0 AU, respectively (1 AU, or Armstrong Unit,=1 km 3 a −1 of solid material). During collisional suturing at fossil CSSZs, global arc magmatic addition is estimated at 0.2 AU. This assessment presumes that in the past the global length of crustal collision zones averaged c . 6000 km, which is one-half that under way since the early Tertiary. The average long-term rate of arc magmatic additions extracted from modern OMSZs and older CSSZs is thus evaluated at 2.7 AU. Crustal recycling at Mesozoic and younger OMSZs is assessed at c . 60 km 3 Ma −1 km −1 ( c . 60% by subduction erosion). The corresponding global recycling rate is c . 2.5 AU. At CSSZs of Mesozoic, Palaeozoic and Proterozoic age, the combined upper and lower plate losses of CIA crust via subduction erosion, sediment subduction, and lower plate crustal detachment and sinking are assessed far less securely at c . 115 km 3 Ma −1 km −1 . At a global length of 6000 km, recycling at CSSZs is accordingly c . 0.7 AU. The collective loss of CIA crust estimated for modern OMSZs and for older CSSZs is thus estimated at c . 3.2 AU. SZ additions (+2.7 AU) and subtractions (−3.2 AU) are similar. Because many uncertainties and assumptions are involved in assessing and applying them to the deep past, the net growth of CIA crust during at least Phanerozoic time is viewed as effectively nil. With increasing uncertainty, the long-term balance can be applied to the Proterozoic, but not before the initiation of the present style of subduction at c . 3 Ga. Allowing that since this time a rounded-down rate of recycling of 3 AU is applicable, a startlingly high volume of CIA crust equal to that existing now has been recycled to the mantle. Although the recycled volume ( c . 9×10 9 km 3 ) is small ( c . 1%) compared with that of the mantle, it is large enough to impart to the mantle the signature of recycled CIA crust. Because subduction zones are not spatially fixed, and their average global lengths have episodically been less or greater than at present, recycling must have contributed significantly to creating recognized heterogeneities in mantle geochemistry.

A new model is proposed for the problematic preservation of an Ordovician forearc basin, which records a complete sedimentary record of arc-continent collision during the Grampian (Taconic) orogeny in the west of Ireland. The South Mayo Trough represents an arc and forearc complex developed above a subduction zone in which the slab dipped away from the Laurentian passive margin. The collision of this arc with Laurentia caused the Middle Ordovician Grampian orogeny. However, the South Mayo Trough, in the hanging wall of this collision zone, remained a site of marine sedimentation during the entire process. Early sediments show derivation from an island-arc complex, an ophiolitic backstop, and polymetamorphic trench sediments. These are conformably overlain by marine deposits derived from a more evolved arc complex and an emerging juvenile orogen. This transition is dated as being coeval with the Grampian metamorphism of the Laurentian footwall. The problem remains as to why subsidence continued in a basin on the hanging wall. It is proposed that the suppression of the expected topography is due to the nature of the Laurentian continental margin. Geophysical and geological evidence suggests that this was a volcanic margin during Neoproterozoic rifting. It is argued that the subduction of this margin caused the formation of eclogites, which reduced its buoyancy. Simple numerical models are presented which show that this is a viable mechanism for the suppression of topography during early stages of arc-continent collision and hence for the preservation of forearcs.

This paper quantifies the flexural subsidence expected from loading by a volcanic arc. The resulting mathematical model shows that the arc width should grow with time and that the subsidence beneath the load can be estimated from the observed arc width at the surface. Application of this model to the Halmahera Arc in Indonesia shows an excellent fit to observations if a broken-plate model of flexure is assumed. The model also gives an excellent fit to data from East Java, also in Indonesia, where it is possible to forward model gravity anomalies. In particular, the depth, location, and width of the depocenter-associated gravity low are accurately reproduced, although the model does require a high density for the volcanic arc (2900 kg m −3 ). This may indicate additional buried loads due, for example, to magmatic underplating. Our main conclusion is that loads generated by the volcanic arc are sufficient to account for much, if not all, of the subsidence in basins within ∼100 km of active volcanoes at subduction plate boundaries, if the plate is broken. The basins will be asymmetrical and, close to the arc, will contain coarse volcaniclastic material, whereas deposits farther away are likely to be volcaniclastic turbidites. The density contrast between arc and underlying crust required to produce the Indonesian arc basins means that they are unlikely to form in young intraoceanic arcs but may be common in older and more mature arcs.

The Indonesian region includes several volcanic island arcs that are highly active at the present day, and also contains a record of Cenozoic volcanic activity owing to subduction of oceanic lithosphere at the margins of SE Asia. As a result of long-term subduction, there is a high regional heat flow, and a weak crust and lithosphere, as identified in other subduction zone backarcs. The stratigraphic record in the Indonesian region reflects a complex tectonic history, including collisions, changing plate boundaries, subduction polarity reversals, elimination of volcanic arcs, and extension. The arcs have not behaved as often portrayed in many arc models. They mark subduction but were not continuously active, and it is possible to have subduction without magmatism. Subduction hinge retreat was accompanied by significant arc volcanism, whereas periods of hinge advance were marked by reduction or cessation of volcanic activity. Growth of the region occurred in an episodic way, by the addition of ophiolites and continental slivers, and as a result of arc magmatism. In Indonesia, relatively small amounts of material were accreted from the downgoing plate during subduction, but there is also little evidence for subduction erosion. During collision the arc region may fail, resulting in thrusting, and the weakest point is the position of the active volcanic arc itself. Volcanic arcs shift position suddenly, and arcs can disappear during collision by overthrusting. Arcs are geologically ephemeral features and may have very short histories in comparison with most well-known older orogenic belts. The stratigraphic record of the basins within arc regions is complex. Because of a weak lithosphere the character of sedimentary basins may be unusual, and basins are commonly very deep and subside rapidly. There is a high sediment flux. The volcanic arc itself influences the stratigraphic record and basin development. The load imposed by the volcanic arc causes flexure and provides accommodation space. The volcanic arc thus can form the basin and supply most of its sediment. Tropical processes influence the mineralogy and apparent maturity of the sediment, especially volcanogenic material. A complex stratigraphy will result from the waxing and waning of volcanic activity.

Shallow-marine carbonate facies from volcanic-arc settings provide an important, but commonly overlooked, record of relative sea-level change, differential subsidence-uplift, paleoclimate trends, and other environmental changes. Carbonate strata are thin where volcanic eruptions are frequent and voluminous, unless shallow, bathy-metric highs persist for long periods of time and volcaniclastic sediment and erupted materials are trapped in adjacent depocenters. Carbonate platforms and reefs can attain significant thickness, however, if subsidence continues after volcanic activity ceases or the volcanic front migrates. The areal extent of shallow-marine carbonate sedimentation is likewise affected by differential tectonic subsidence, although carbonate platforms are most laterally extensive during transgressive to highstand conditions and when arc depocenters are filled with sediment. Tectonic controls on shallow-marine carbonate sedimentation in arc depocenters include (1) coseismic fault displacements and associated surface deformation; (2) long-wavelength tectonic subsidence related to dynamic mantle flow, flexure, lithospheric thinning, and thermal subsidence; and (3) large-scale plate deformation related to local conditions of subduction. Depositional controls on carbonate sedimentation in arc depocenters include (1) the frequency, volume, and style of volcanic eruptions; (2) accumulation rates for siliciclastic-volcaniclastic sediment; (3) the frequency, volume, and dispersal paths of erupted material; (4) (paleo)wind direction, which influences both carbonate facies development directly and indirectly by controlling the dispersal of volcanic ash and other pyroclastic sediment, which can bury carbonate-producing organisms; (5) the frequency and intensity of tsunami events; and (6) volcanically or seismically triggered mass-wasting events, which can erode or bury carbonate strata. Regarding platform morphologies in arc-related settings, (1) fringing reefs or barrier reef systems with lagoons may develop around volcanic edifices throughout the long-term evolution of volcanic arcs; (2) local reefs and mounds may build on intrabasinal, fault-bounded highs within underfilled forearc, intra-arc, and backarc basins; (3) isolated platforms with variable platform margin-to-basin transitions are common in “underfilled” and tectonically active depocenters; and (4) broad ramps and rimmed carbonate shelves are typically found in tectonically mature and sediment-filled depocenters.

In the Bismarck Volcanic Arc in Papua New Guinea, six fields of sediment waves were imaged with sonar. Sediment structures observed in seismic data and swath bathymetry are not unique and can result from predominantly continuous (bottom) currents, or episodic (turbidity) currents, or from deformation of sediment. Two of these wave fields overlap and appear to be of turbidity-current origin and modified by bottom currents, with one field unconformably overlying the other field. A field off the coast of Dakataua caldera displays an arcuate morphology, and a series of enclosed depressions within the field suggests creation by extensional deformation of rapidly deposited sediment. Scour features in side-scan imagery suggest turbidity-current activity, which also likely modifies the sediment waves. The wave field is isolated from hyperpycnal currents, however, suggesting that in the absence of a shelf, coastal erosion and small landslides can produce semiregular gravity-driven sediment flows that deposit in deep (>1400 m) water. In Kimbe Bay a fourth sediment-wave field also displays arcuate morphology and enclosed depressions within the field. This wave field is found within a bay >40 km from shore and also appears to have been formed by a combination of extensional deformation of sediment and energetic current activity. Two additional fields in Hixon Bay are fed by small and medium rivers (<∼450 m 3 /s mean annual discharge) draining volcanoes and mountainous regions. One small field appears within a slide scar, suggesting that the initial topography of the scar provided the conditions for early sediment-wave growth. A much larger field is best explained by repeated hyperpycnal currents originating from the Pandi River. We cored a series of upward-fining, graded sequences consistent with a turbidity-current origin. Ages from these cores and measurements of relative thickness in sub-bottom imagery of the field constrain deposition rates for the field and suggest that a large part of the Pandi River discharge must be bypassing the shelf and depositing on the sediment-wave field in deep water (>1200 m). These findings suggest that the sedimentary record in arc collision zones will be dominated by mass-wasting deposits very close to volcanoes, and by river discharge depositing in select, extent regions far from shore. Because sedimentation rates can vary by a factor of 2 between the two flanks of a sediment wave, care must be taken when comparing bed thickness across an entire sedimentary section.

Marine surveys show that the submarine Huatung Ridge extends northward to the Lichi Mélange in the southwestern Coastal Range, suggesting that formation of the Lichi Mélange is related to arcward thrusting of the forearc strata in the western part of the North Luzon Trough during active arc-continent collision off southern Taiwan. A new seismic survey along the 21° N transect across the North Luzon Trough in the incipient arc-continent collision zone further reveals that deformation of the Huatung Ridge occurred soon after sedimentation in the western forearc basin, whereas sedimentation was continuous in the eastern part of the remnant North Luzon Trough until the complete closure of the forearc basin approaching SE Taiwan. This suggests that the sequence in the Huatung Ridge can be coeval with just the lower sequence of the remnant-forearc-basin strata. Multiple lines of new evidence, including micropaleontology, clay mineralogy, and fission track analyses along the Mukeng River and its tributary key sections, are used to test this thrusting-forearc-origin hypothesis of the Lichi Mélange. In the SW Coastal Range the Lichi Mélange lies between the collision suture of Longitudinal Valley to the west and the Taiyuan remnant forearc basin to the east. A field survey indicates that the Taiyuan forearc-basin sequence and its volcanic basement were thrust westward over the Lichi Mélange along the east-dipping Tuluanshan Fault. The Lichi Mélange shows varying degrees of fragmentation of strata, mixing, and shearing. An apparently wide range of facies is present, from the weakly sheared broken formation facies, with discernible relict sedimentary structures, to the intensely sheared block-in-matrix mélange facies, with pervasively scaly foliation dipping to the SE. Sedimentological study reveals that the subangular to subrounded, fractured, matrix-supported metasandstone conglomerates in the pebbly mudstone layers are repeatedly found in the broken formation facies of the Lichi Mélange. Their composition and occurrence are identical to the deep-sea-fan conglomerate beds in the Taiyuan remnant-forearc-basin strata to the east. Benthic foraminiferal faunas are similar in the Lichi Mélange, regardless of the varying intensity of shearing and strata disruptions, and are compatible with the benthic foraminiferal fauna in the Taiyuan remnant-forearc-basin turbidites, supporting the interpretation that the protolith of the Lichi Mélange was originally deposited in the North Luzon Trough. Age determination of planktic microfossils further demonstrates that the Lichi Mélange is early Pliocene (3.5–3.7 Ma), implying that this mélange was deposited in a short time and that deformation occurred soon after its deposition. The early Pliocene age of the Lichi Mélange is coeval with just the lower part of the Taiyuan remnant forearc strata, and is much younger than the upper forearc sequence (3–1 Ma). Thus the Taiyuan coherent-forearc-basin strata (3.7–1 Ma) were deposited continuously in the remnant North Luzon Trough regardless of the deformation in its western part (the protomélange). This scenario is an analogue for the modern configuration of the Huatung Ridge–remnant North Luzon Trough off the southern Coastal Range in the active arc-continent collision zone north of lat 21° N. In addition to its kaolinite content (11–15%), the clay mineral composition of the Lichi Mélange is compatible with the Taiyuan remnant forearc turbidites. In the Coastal Range, kaolinites are found only in the volcanic rocks of the Tuluanshan Formation. This additional kaolinite in the Lichi Mélange could not have been derived from the exposed accretionary prism to the North Luzon Trough by sedimentary mass slumping, because no such volcanic rocks are now exposed in the accretionary prism west of the Coastal Range. Instead, they could have been derived from the Tuluanshan Formation when it was emplaced into the Lichi Mélange by thrusting during the last 1 Ma when the Luzon arc-forearc was accreted to form the southern Coastal Range. Thus the kaolinites of the volcanic arc rocks were redistributed into the Lichi Mélange by fluid flows along the ubiquitous geological fractures in the mélange, consistent with the field occurrences of the large, rootless, fault-bounded volcanic rocks of andesitic breccia, tuff, and agglomerates that were floating in the intensely sheared block-in-matrix mélange facies of the Lichi Mélange. Mélange is commonly considered to develop in the accretionary prism of a subduction zone. However, the Lichi Mélange in the SW Coastal Range originated from the thrust forearc strata, representing a unique forearc mélange for orogenic belts worldwide. The young age and wide distribution—especially the continuous offshore-onshore connection—of the Lichi Mélange provides a unique example for further research into active modern mélange-forming processes by forearc thrusting during progressive closure of the forearc basin in this active region of arc-continent collision.

The NW corner of the Pacific Ocean is a place of unique Tertiary tectonism, which provides one of the clearest examples of arc-arc collision. Voluminous Cretaceous rhyolitic-granitic magmatism along the continental margin continues into the Paleogene. In contrast, Miocene island arc volcanism follows Eocene boninitic magmatism in the Izu-Mariana Arc, in association with the opening of backarc basins, including those in the Philippine and Japan Seas. The triple junction between the Eurasian, Philippine Sea, and Pacific plates arrived in the area south of Tokyo during the Miocene, just as the Japan Sea was opening. After the beginning of Philippine Sea plate subduction to the north, the Izu Island Arc began to collide obliquely with the Honshu Arc. As a result, this unique tectonic setting in the NW Pacific has produced a miniature Alpine-type orogenic belt (Tanzawa) in the collisional center, whereas in the eastern part of the Izu Arc sediment has been actively accreting in that forearc. Such settings have resulted in systematic accretionary prism formation from the early Miocene in the Boso-Miura peninsular area to the present in the Sagami Trough area. We modeled the tectonics by a simple sandbox experiment. Systematic fault and fracture patterns of the oblique subduction type are predicted to occur during arc-arc collision.

An integrated structural, stratigraphic, and sedimentological analysis of the West Crocker formation in northwest Borneo suggests that it is best interpreted as an accretionary prism. The structural geology provides clear evidence of at least two episodes of syndepositional folding and thrust faulting. A probable Eocene age, indicated by foraminiferal and palynological assemblages, differs from the generally accepted Oligocene to early Miocene age and is consistent with deposition of the West Crocker formation during a phase of tectonism at the northwest Borneo margin. Sandstones within the West Crocker formation were deposited by high-density turbidity currents that constructed relatively small, progradational lobes in a slope apron environment, and trace fossil assemblages confirm bathyal water depths of ∼1000 m or more. The composition of the sandstones, which contain abundant feldspars and lithic fragments, suggests that their provenance was the first-cycle product of an eroded orogenic belt, whereas immature textures indicate a short distance of transport.

The Izu Arc has been colliding with the Honshu Arc in central Japan since ca. 15 Ma. In order to understand the provenance changes related to this collision, we studied lower to middle Miocene sandstones in and around the collision zone by analyzing their framework composition and the chemistries of detrital clinopyroxene, garnet, and chromian spinel. Sandstone deposited in the trench and forearc basin of the Honshu Arc prior to collision includes grains of detrital garnet and chromian spinel, which originated mainly from granites and low pressure-temperature (P-T) metamorphic rocks, and forearc peridotite, respectively, parts of the Honshu Arc. The forearc and trench-fill sandstones differ in terms of their framework composition; sedimentary lithics are more abundant in the forearc sandstone than in the trench. The two groups of sediments were supplied from different parts of the Honshu Arc. The lower part of the clastic sequence deposited within the Izu Arc is composed mainly of volcaniclastic rocks and yields detrital clinopyroxenes that originated from the Izu Arc. In contrast, the upper part is similar to the lower Miocene trench-fill deposits in terms of its framework composition and the chemistry of detrital garnet and chromian spinel. This reflects a change in provenance triggered by the initial contact of the Izu Arc and the trench between the Eurasian and Philippine Sea plates. The lower part of the middle Miocene trench-fill that was deposited following initial contact is also similar to the lower Miocene trench-fill. The upper part, however, resembles lower Miocene sedimentary rocks of the forearc basin. This suggests that the transport path was changed by collision. During the initial stages of collision between the Honshu and Izu Arcs, the Honshu Arc was preferentially uplifted, and therefore supplied most of the detritus to the collision zone.