Adjacent Lithosphere Research Articles

Coupling Between Lithosphere Removal and Mantle Flow in the Central Andes

AbstractThe central Andes is part of a Cordilleran orogen formed through continent‐ocean convergence. In contrast to the thickened crust, the mantle lithosphere below much of the orogen is anomalously thin. Additionally, the surface is characterized by widespread backarc magmatism and transient ∼100‐km‐wide basins that developed the over last 30 million years, with basins located systematically seaward of major backarc ignimbrite centers and basin formation predating the late Miocene magma/ignimbrite flare‐up. Using numerical models, we propose a novel mechanism whereby lithosphere removal is coupled with mantle flow. First, a small area of high‐density eclogitized lower crust initiates a gravitational instability, causing a localized basin at the surface that subsides and then uplifts. Foundering crust and adjacent lithosphere are entrained by subduction‐induced mantle flow, driving regional lithosphere removal and magmatism. The models demonstrate that mantle flow can amplify a local lithosphere instability to orogen‐wide lithosphere removal, rapidly eliminating accumulated mass in the orogen.

Crustal structure of the Murray Ridge, northwest Indian Ocean, from wide-angle seismic data

The Murray Ridge/Dalrymple Trough system forms the boundary between the Indian and Arabian plates in the northern Arabian Sea. Geodetic constraints from the surrounding continents suggest that this plate boundary is undergoing oblique extension at a rate of a few millimetres per year. We present wide-angle seismic data that constrains the composition of the Ridge and of adjacent lithosphere beneath the Indus Fan. We infer that Murray Ridge, like the adjacent Dalrymple Trough, is underlain by continental crust, while a thin crustal section beneath the Indus Fan represents thinned continental crust or exhumed serpentinized mantle that forms part of a magma-poor rifted margin. Changes in crustal structure across the Murray Ridge and Dalrymple Trough can explain short-wavelength gravity anomalies, but a long-wavelength anomaly must be attributed to deeper density contrasts that may result from a large age contrast across the plate boundary. The origin of this fragment of continental crust remains enigmatic, but the presence of basement fabrics to the south that are roughly parallel to Murray Ridge suggests that it separated from the India/Seychelles/Madagascar block by extension during early breakup of Gondwana.

Lithospheric topography, tilted plumes, and the track of the Snake River–Yellowstone hot spot

The trace of the Snake River–Yellowstone hot spot is the world's best example of a mantle plume that has been overridden by continental lithosphere. The “standard model” calls for the plume head to rise under northern Nevada and be forced northward to form basalts of the Columbia Plateau; subsequent movement of North America to the southwest over the plume tail created a hot spot trace on the surface. We present a new conceptual model for the origin of this feature that resolves inconsistencies in the current standard model and explains the recent documentation of a thermal anomaly in the mantle below Yellowstone today that plunges ∼65° WNW. Our model implies that the plume tail was forced beneath thinned cratonic lithosphere to the SE along with part of the plume head and has remained in this orientation for the last 12 Ma. We infer that almost all of the volcanism in SE Oregon and SW Idaho prior to 12 Ma results from overriding the southern extension of the plume head, not the plume tail, and that a distinct plume tail hot spot track was not established until formation of the Bruneau‐Jarbidge eruptive center around 12 Ma. The plume tail track may also be controlled by a preexisting structural boundary in lithosphere that is thinner than adjacent lithosphere. This model demonstrates the potential importance of lithospheric topography on controlling the surface manifestation of plume volcanism and the complexity that may arise when lithospheric thickness is nonuniform.

SEGMENTATION IN OMAN OPHIOLITE, IMPLICATIONS FOR FAST SPREADING RIDGES TECTONICS AND HIGH TEMPERATURE HYDROTHERMAL ACTIVITY

New, fine scale mapping in the large NW-SE ridge segment formerly identified in the Oman ophiolite (Nicolas and Boudier, 1995) has revealed that this structure is composed of smaller nestled segments, each being centered on the small mantle diapirs already mapped. The contacts with adjacent lithosphere and the tips of these segments have been mapped in detail. Their nature and structure depend on the difference in age between the two lithospheres. When the difference in age is in the range of 1 Myr, strike slip shear zones, 1-2 km wide, are developed in the mantle of the new segment. When this difference drops to ~0.5 Myr, the shear zones are small and diffuse but, in the mantle wedge at the tip of the segment which penetrates the older lithosphere, spectacular deformations are observed. The mantle and lower crust of the older lithosphere near the Moho are shovelled vertically and kilometer-sized folds develop in the gabbro unit. In contrast, the lid is not affected, suggesting that, at present day fast spreading ridges, similar major tectonic structures, as seen in Oman thanks to deep sections, may also be present. Contacts and tips of new segments are invaded by mafic dikes and sills issued from the segment magmatic activity and trapped against these colder boundaries. Melt reaction in the uppermost mantle at the tip of propagators generates large volumes of dunites and wehrlites. An important contribution to this magmatism results from massive seawater penetration down to the Moho, possibly favored by the segment tectonic activity. This water is heated at very high temperatures (VHT), 1000C° and possibly above. Inside crystallizing magma chamber, the hydrous reaction (Koepke et al.,2004) rejuvenates magmatic activity and generates orthopyroxene gabbros, injected and stretched parallel to the olivine gabbros layering. Outside the magma chamber, VHT seawater produces, by hydrous anatexis, copious melts which mix and react with the indigenous segment melts and crystallize as orthopyroxene- or pargasitebearing, clinopyroxene gabbros, diorites and trondjhemites. It is suggested that the geochemical signature of such hybrid melts should be looked for in present day ridges.

The thermal effects of steady-state slab-driven mantle flow above a subducting plate: the Cascadia subduction zone and backarc

At subduction zones, geophysical and geochemical observations indicate that the arc and backarc regions are hot, in spite of the cooling effects of a subducting plate. At the well-studied Cascadia subduction zone, high mantle temperatures persist for over 500 km into the backarc, with little lateral variation. These high temperatures are even more surprising due to the juxtaposition of the hot Cascadia backarc against the thick, cold North America craton lithosphere. Given that local heat sources appear to be negligible, mantle flow is required to transport heat into the wedge and backarc. We have examined the thermal effects of mantle flow induced by traction along the top of the subducting plate. Through systematic tests of the backarc model boundary, we have shown that the model thermal structure of the wedge is primarily determined by the assumed temperatures along this boundary. To get high temperatures in the wedge, it is necessary for flow to mine heat from depth, either by using a temperature-dependent rheology, or by introducing a deep cold boundary through a thick adjacent lithosphere, consistent with the presence of a craton. Regardless of the thermal conditions along the backarc boundary, flow within an isoviscous wedge is too slow to transport a significant amount of heat into the wedge corner. With a more realistic stress- and temperature-dependent wedge rheology, flow is focused into the wedge corner, resulting in rapid flow upward toward the corner and enhanced temperatures below the arc, compatible with temperatures required for arc magma generation. However, this strong flow focusing produces a nearly stagnant region further landward in the shallow backarc mantle, where model temperatures and heat flow are much lower than observed. Observations of high backarc temperatures, particularly in areas that have not undergone recent extension, provide an important constraint on wedge dynamics. None of the models of simple traction-driven flow were able to simultaneously produce high temperatures beneath the volcanic arc and throughout the backarc. The most likely way to produce hot and isothermal conditions in the backarc is through vigorous small-scale-free convection in a low-viscosity asthenosphere.

Thin‐shell modeling of neotectonics in the Azores‐Gibraltar Region

We applied the thin‐shell neotectonic modeling method to study the neotectonics of the Africa/Eurasia plate boundary in the Azores‐Gibraltar region. The plate boundary consists of a simple fault system running from Azores to the Gorringe Bank where it branches along the Betics and Rift‐Tell thrust fronts. Major faults in west Iberia and NW Africa have also been incorporated. Results are compared with seismic strain rates, fault slip rates and stress orientations. The best estimate for the fault friction coefficient is 0.1–0.15 meaning that the plate‐boundary is only about 1/4 as strong as the adjacent lithosphere. The largest fault slip rates (>1.5 mm/yr) are obtained along the Gloria fault (strike‐slip), and the Betic (transpressive) and Rif‐Tell (compressive) thrust systems. Whereas tectonic activity in the Atlas region is comparable to that obtained along the plate boundary, the fault slip rates in the west Iberia fault systems are one order of magnitude less.

The formation of the West European Rift; a new model as exemplified by the Massif Central area

Abstract In this paper, we use mainly field data from the Massif Central area, which have been presented in a companion paper [Michon and Merle, 2001], to discuss the origin and the evolution of the West European Rift system. It is shown that the tectonic event in the Tertiary is two-stage. The overall geological evolution reveal a tectonic paradox as the first stage strongly suggests passive rifting, whereas the second stage displays the first stage of active rifting. In the north, crustal thinning, graben formation and sedimentation at sea level without volcanism during the Lower Oligocene, followed by scattered volcanism in a thinned area during Upper Oligocene and Lower Miocene, represent the classical evolution of a rift resulting from extensional stresses within the lithosphere (i.e. passive rifting). In the south, thinning of the lithospheric mantle associated with doming and volcanism in the Upper Miocene, together with the lack of crustal thinning, may be easily interpreted in terms of the first stage of active rifting due to the ascent of a mantle plume. This active rifting process would have been inhibited before stretching of the crust, as asthenospheric rise associated with uplift and volcanism are the only tectonic events observed. The diachronism of these two events is emphasized by two clearly distinct orientations of crustal thinning in the north and mantle lithospheric thinning in the south. To understand this tectonic paradox, a new model is discussed taking into account the Tertiary evolution of the Alpine chain. It is shown that the formation of a deep lithospheric root may have important mechanical consequences on the adjacent lithosphere. The downward gravitational force acting on the descending slab may induce coeval extension in the surrounding lithosphere. This could trigger graben formation and laguno-marine sedimentation at sea level followed by volcanism as expected for passive rifting. Concurrently, the descending lithospheric flow induces a flow pattern in the asthenosphere which can bring up hot mantle to the base of the adjacent lithosphere. Slow thermal erosion of the base of the lithosphere may lead to a late-stage volcanism and uplift as expected for active rifting.

When the Wilson Cycle breaks down: how orogens can produce strong lithosphere and inhibit their future reworking

Abstract Although poly-cyclicity is common, many orogens show a remarkable lack of reworking. In this paper, a review of some factors that may either enhance or inhibit reworking of orogens is presented. As a general rule, orogens are unlikely to rift and rework if their lithospheric strength is higher than adjacent lithosphere. The strength of the lithosphere is strongly dependent on the geothermal gradient and the rheology of the rocks; both these factors can depend on the preceding orogenic evolution, even several hundred Ma after orogenesis. Strong orogenic lithosphere is expected if the crust is composed of material with a low radiogenic heat production capacity, such as island arcs, or if the underlying sub-continental lithosphere is still thickened, as in the Urals. Extensive dehydration metamorphism, a concentration of radiogenic heat production in the upper crust and erosional thinning of the orogenic crust can also strengthen the lithosphere and inhibit reworking. However, proximity of Archean cratons and anomalously high mantle heat flow appear to strongly enhance susceptibility to reworking.

Velocity–density relationships and modeling the lithospheric density variations of the Kenya Rift

We have devised pressure-dependent and temperature-dependent linear velocity–density relationships from the crustal rocks in the Kenya Rift and applied the relationships to the crustal velocity model derived from the KRISP90 refraction experiment to compute densities and gravity anomalies. Our results indicate that pressure and temperature considerations reduce the misfit between the observed and the calculated gravity anomalies in the areas of large lateral pressure and temperature variations. Using a separate relationship involving velocity and density contrasts for the mantle of the Kenya Rift, the KRISP teleseismic velocity model, and the variation of the observed gravity anomalies, we derive a d ρ/d V p value of ∼0.05 (Mg s m −3 km −1) which results in the density contrast of ∼0.03 Mg m −3 at ∼100 km depth beneath the rift and its flanks. Using the magnitude of the velocity and density contrasts, we conclude that the mantle at the depth of 100 km beneath the rift could be ∼200°C to 375°C hotter than the adjacent lithosphere (and much of it is considered subsolidus) and contains partial melt up to 2.5%. The amount of partial melt in the region of the Moho upwarp would be greater because both velocities and densities are lower in that region. Our results are consistent with the upwarping of the thermal boundary layer representing the lithosphere–asthenosphere boundary and the consequent partial melting of the lithosphere.

Detailed structures of the subducted Philippine Sea Plate beneath northeast Taiwan: A new type of double seismic zone

We studied the detailed structure of the subducted Philippine Sea plate beneath northeast Taiwan where oblique subduction, regional collision, and back arc opening are all actively occurring. Simultaneous inversion for velocity structure and earthquake hypocenters are performed using the vast, high‐quality data recorded by the Taiwan Seismic Network. We further supplement the inversion results with earthquake source parameters determined from inversion of teleseismic P and SH waveforms, a critical step to define the position of plate interface and the state of strain within the subducted slab. The most interesting feature is that relocated hypocenters tend to occur along a two‐layered structure. The upper layer is located immediately below the plate interface and extends down to 70–80 km at a dip of 40°–50°. Below approximately 100 km, the dip increases dramatically to 70°–80°. The lower layer commences at 45–50 km and stays approximately parallel to the upper layer with a separation of 15±5 km in between down to 70–80 km. Below that the separation decreases and the two layers seem to gradually merge into one Wadati‐Benioff Zone. We propose to term the classic double seismic zones observed beneath Japan and Kuril as “type I” and that we observed as “type II,” respectively. A global survey indicates that type II double seismic zones are also observed in New Zealand near the southernmost North Island, Cascadia, just north of the Mendocino triple junction, and the Cook Inlet area of Alaska. All of them are located near the termini of subducted slabs in a tectonic setting of oblique subduction. We interpret the seismogenesis of type II double seismic zones as reflecting the lateral compressive stress between the subducted plate and the adjacent lithosphere (originating from oblique subduction) and the downdip extension (from slab pulling force). The upper seismic layer represents seismicity occurring in the upper crust of a subducted plate and/or along the plate interface, whereas the lower layer is associated with events in the uppermost mantle.

Oceanic transform earthquakes with unusual mechanisms or locations: relation to fault geometry and state of stress in the adjacent lithosphere

Oceanic transform earthquakes with unusual mechanisms or locations: relation to fault geometry and state of stress in the adjacent lithosphere

Oceanic transform earthquakes with unusual mechanisms or locations: Relation to fault geometry and state of stress in the adjacent lithosphere

On oceanic transforms, most earthquakes are expected to occur on the principal transform displacement zone (PTDZ) and to have strike‐slip mechanisms consistent with transform‐parallel motion. We conducted a search for transform earthquakes departing from this pattern on the basis of source mechanisms and locations taken from the Harvard centroid moment tensor catalogue and the bulletin of the International Seismological Centre, respectively. Events with unusual mechanisms occur on several transforms. We have determined the source mechanisms and centroid depths of 10 such earthquakes on the St. Paul's, Marathon, Owen, Heezen, Tharp, Menard, and Rivera transforms from inversions of long‐period body waveforms. Relative locations of earthquakes along these transforms have been determined with a multiple‐event relocation technique. Much of the anomalous earthquake activity on oceanic transforms is associated with complexities in the geometry of the PTDZ or the presence of large structural features that may influence slip on the fault. Reverse‐faulting earthquakes occur at a compressional bend in the Owen transform in the area of Mount Error and at the St. Paul's transform near St. Peter's and St. Paul's Rocks. A normal‐faulting earthquake on the Heezen transform is located at the edge of a pull‐apart basin marking an extensional offset of the fault. Normal‐faulting earthquakes along the Tharp, Menard, and Rivera transforms may also be related to extensional offsets. Some events with unusual mechanisms occur outside of the transform fault zone, however, and do not appear to be related to fault zone geometry. For instance, earthquakes with mechanisms indicating reverse‐faulting on ridge‐parallel fault planes are located near the ridge‐transform intersections of the St. Paul's and the Marathon transforms. Possible additional contributors to the occurrence of anomalous earthquakes include recent changes in plate motion, differential lithospheric cooling, and the development of a zone of weakness along the fault zone, but we do not find strong evidence to confirm the influence of these processes.

Microearthquake evidence for extension across the Kane Transform Fault

We report the results of a microearthquake experiment conducted in 1987 in the active transform portion of the Kane Fracture Zone. The Kane is a slow slipping (25 mm/yr), large‐offset (150 km) transform delineated by a pronounced transform valley. The experiment site lies in the eastern half of the transform, 40 km west of the eastern ridge‐transform intersection, in a region of marked transform‐parallel topography. The network consisted of 12 ocean bottom hydrophones and seismometers and extended for 25 km along the transform and 10 km in width, spanning the transform valley and a narrow median ridge to the north. Hypocentral parameters were determined for 86 earthquakes. Epicenters within the network define a narrow transform‐parallel trend along the steep southern wall of the transform valley. Maximum focal depths increase from 6 to 9 km from east to west across the network. Significant activity was also detected outside the transform tectonized zone. To the southeast this activity may be associated with the inside corner of the ridge transform intersection, while to the north about 20 events occurred in 7‐m.y.‐old crust in a region of ridge‐parallel bathymetry. Six focal mechanisms were obtained from P wave first motions for events within the transform valley. The best constrained solutions, for four earthquakes within the network, show normal faulting on fault planes subparallel to the trend of the transform. All the mechanisms indicate a tension axis perpendicular to the trend of the transform. These results together with a significant historical record of large earthquakes near the experiment site lead us to conclude that the principal transform displacement zone was inactive during our experiment and that the activity we recorded is the result of extension in the adjacent lithosphere. The observed focal mechanisms and the inference that the axis of least compressive stress is approximately perpendicular to the transform provide direct evidence that the transform fault is mechanically weak relative to the surrounding lithosphere. Potential sources of extension across the transform include thermal stress in the young oceanic lithosphere, topographic loading, a small component of plate divergence normal to the transform, and northward motion of the asthenosphere relative to the surface plate.

Topographic and volcanic asymmetry around the Red Sea: Constraints on rift models

The Red Sea rift is asymmetric, with the locus of uplift and Tertiary volcanism in Saudi Arabia, 200–400 km east of the present rift axis. One model for this asymmetry involves simple shear extension on an east dipping, low‐angle master fault that penetrates the lithosphere. Uplift and volcanism would mark the location where the fault enters the asthenosphere. However, observed seismicity, regional heat flow, and sedimentation are not in accord with this model's predictions. An alternate model involves interaction of upwelling asthenosphere in the early rifting stage with a nearby crustal weak zone such as a suture that controls rift location. In general, the location of the weak zone and subsequent rift will not coincide with the locus of mantle upwelling, leading to several asymmetric rift features. In this model, Tertiary volcanism in Saudi Arabia marks the location of initial upwelling, and uplift is due to crustal thickening associated with magmatic underplating and crustal intrusion. The incipient crustal rift and the locus of mantle upwelling will tend to align as rifting continues and stable seafloor spreading develops, implying relative migration of the lithosphere and asthenosphere. Absolute plate motion models for Africa and strongly asymmetric spreading in the Red Sea are consistent with northeast migration of the incipient rift and adjacent lithosphere with respect to a zone of asthenospheric upwelling.

Small-scale convection induced by passive rifting: the cause for uplift of rift shoulders

Areas adjacent to rifts, or rift shoulders, are often observed to be uplifted as much as a kilometer or more. In some of these regions geologic data indicate a passive origin for the rifting itself (i.e. there was no anomalous heating of the regions before rifting). Purely conductive heat transport between the rift, where the lithosphere has been thinned, and the rift flanks cannot account for the magnitude of the uplift. Small-scale convection will be induced in the mantle beneath a rift due to the lateral temperature gradients there. Numerical experiments show that convection increases the amount of heat transported vertically into the rift and laterally out of it. In these calculations, the viscosity is taken to be dependent on temperature and pressure and, in some cases, stress. The mantle flow results in thinning of the adjacent lithosphere causing flanking uplift as well as slowing of the subsidence of the middle of the rift. The magnitude of the uplift is dependent on the geometry of the rift and the importance of stress-dependence in the rheology of the mantle. For viscosity parameters which are consistent with the pre-rift temperature structure small-scale convection can produce uplift at least twice as great as would be produced by lateral conduction alone.

On the mechanism of magma injection and plate divergence during the Krafla Rifting Episode in NE Iceland

Geodetic observations of horizontal and vertical point displacements in a ∼100‐km‐wide network across the Krafla fissure swarm of NE Iceland are analyzed with simple elastic models. The observed displacements accumulated from 1975 to 1980 in several events of magma injection into fissures that were widened. The fissure swarm was extended, and it subsided while the adjacent lithosphere contracted or was compressed accompanied by surface uplift. We applied two models to fit the far‐field deformation: an infinite horizontal pressure line source or a pipe and an infinite, vertically limited, vertical dyke source loaded by normal stress at the walls, both embedded in an elastic half‐space. In the far‐field inversion the two models cannot be distinguished, but geology and the observed axial graben subsidence clearly discriminate against the pipe and for the dyke. The inversion suggests that the widened dykes cut through the whole lithosphere which in the fissure swarm is only 4–6 km thick. During the rifting episode the general magma level in the fissure swarm rose to near the surface by 1980, but the loading increments appear to have decreased whith time from 1975 to 1980.

Implications of volcano and swell heights for thinning of the lithosphere by hotspots

The heights of hotspot volcanoes are modeled by assuming isostatic equilibrium between the magma column and the adjacent lithosphere. The depth to the level of compensation is assumed to be related to the thickness of the lithosphere after it has been reheated and thinned by a hotspot. There is a square‐root relationship between volcano height and lithospheric age. Relationships between volcano height and lithospheric thickness and between reheated thickness and age of the lithosphere can be constrained within the limits imposed by uncertainties in lithospheric and magma densities. If the reheated thickness determined from the isostatic model can be compared with lithospheric thickness determined from thermal models of the lithosphere, then a relationship between the thickness of the lithosphere before and after reheating can be derived. Combining the upper limit on the reheated thickness/age relationship with a theoretical expression, derived by S. T. Crough, which relates swell height to lithospheric age, reset thickness, and physical constants, yields a swell‐height/age relationship that is in good agreement with empirical swell‐height data. This agreement supports the assumption that the lithospheric thickness defined isostatically by volcano heights is closely related to the thermal thickness of the lithosphere, a result that, in turn, supports the thermal origin of hotspot swells.

Laramide Interactions of Structural Elements in Southwestern Montana: ABSTRACT

The present geologic framework of southwestern Montana is the result of the interactions of several structural units during the Laramide orogeny. The Lewis and Clark lineament is the oldest of these, having been in existence since the Precambrian when some of its elements bounded the Belt embayment. It is composed of five or six parallel features with differing geologic histories. During the Laramide orogeny, horizontal compression initiated deep-seated wrench faulting along the Lewis and Clark lineament. This faulting allowed the transition from the continued Sevier End_Page 1348------------------------------ style of thin-skinned overthrust deformation, which occurred north of the lineament, to the foreland basement-involved thrust deformation, which occurred to the south. The differential motion of the basement was absorbed by small amounts of left-lateral transform motion. This lineament may also have created a weak zone in the crust into which the Boulder batholith intruded during the early stages of the Laramide. The voluminous volcanic material associated with the batholith created a supracrustal load which downwarped the adjacent lithosphere. If the batholith itself slid eastward, as advocated by Hyndman in 1979, the load was enhanced. Decoupling of the lithosphere along the southernmost elements of the Lewis and Clark lineament localized and accentuated the load-induced subsidence, creating the Crazy Mountains basin and localizing the accumulation of the thick volcaniclastic sediments of the Livingston Group. End_of_Article - Last_Page 1349------------

Subsidence of the Paris Basin

We discuss the main results of a systematic quantitative analysis of the subsidence of the Paris basin made on 100 wells. Backstripping techniques with decompaction as well as sea level and water depth corrections are used to derive tectonic subsidence curves (in the absence of water and sediment loading) assuming local as well as regional compensation. A best fitting sea level variations curve is derived for the deepest central part on the basis of the resulting tectonic subsidence curve. The adopted amplitude of the regression since Upper Cretaceous is 330 m. The spatial and temporal distribution of tectonic subsidence is reconstructed and discussed. It is shown that rigidity of the lithosphere (with an adopted 50‐km flexural parameter) plays a significant role. These results are then quantitatively interpreted in terms of the lithospheric stretching model. This interpretation is reasonable provided it is made in terms of volumetric subsidence over the whole basin, because of the thermal conduction and the rigidity low‐pass filtering effects. The subsidence can be explained if the axial zone was thinned during Permo‐Triassic by a factor of about 1.3 with respect to the adjacent lithosphere. It is shown that either the initial altitude was more than 500 m above sea level or the lower part of the lithosphere was lost at the beginning of the Permo‐Triassic phase of extension, resulting in a rapid uplift of several hundreds of meters. The similarity of the Permo‐Triassic initiating extensional phase with active continental rifting is shown.

Seismotectonics of the northern Aegean area

Data concerning the focal mechanism and the spatial distribution of earthquakes have been used to investigate the active tectonics of the northern Aegean and the surrounding area. A thrust region, which includes the northernmost part of the Aegean and at least part of the Marmara Sea, has been defined. An amphitheatrical Benioff zone dipping towards the thrust region from south, east and probably from west, at a mean angle of about 30°, has been detected. The thrust region is surrounded by a region of normal faulting. An eastward progression of the seismic activity in this normal faulting region between 1954 and 1971 has been observed. A correspondence between the earthquake occurrence in the thrust and normal faulting regions has also been observed. Each large shock produced by tensional mechanism in the region of normal faulting is preceded or followed by one or more shocks of compressional mechanism in the thrust region. The focal mechanism, the distribution of the earthquake foci with intermediate focal depth, as well as some magnetic and gravimetric observations can be interpreted by assuming that dense oceanic crust sinks in the northern part of this area and that the adjacent lithosphere moves by segmentation to fill the void with the consequence of producing tensile stresses associated with normal faulting. Such a mechanism of lithospheric interaction suggests that accretion probably takes place in this area.