Geophys. J. Int. (1999) 139, 556–562 Numerical models of ductile rebound of crustal roots beneath mountain belts Hemin A. Koyi,1 A. Geoffrey Milnes,2,* Harro Schmeling,3 Christopher J. Talbot1, Christopher Juhlin1 and Hermann Zeyen4 1 Hans Ramberg T ectonic L aboratory, Department of Earth Sciences, V illavägen 16, S-752 36 Uppsala, Sweden. E-mail: [email protected] 2 Department of Geology, University of Bergen, N-5007 Bergen, Norway 3 Institut für Meteorologie und Geophysik, Universität Frankfurt, Feldbergstrasse 47, 60323 Frankfurt am Main, Germany 4 Département des Sciences de la T erre Université de Paris-Sud, Bât. 504, F-91405 Orsay, cedex, France Accepted 1999 July 27. Received 1999 June 14; in original form 1998 December 10 SU M MA RY Crustal roots formed beneath mountain belts are gravitationally unstable structures, which rebound when the lateral forces that created them cease or decrease significantly relative to gravity. Crustal roots do not rebound as a rigid body, but undergo intensive internal deformation during their rebound and cause intensive deformation within the ductile lower crust. 2-D numerical models are used to investigate the style and intensity of this deformation and the role that the viscosities of the upper crust and mantle lithosphere play in the process of root rebound. Numerical models of root rebound show three main features which may be of general application: first, with a low-viscosity lower crust, the rheology of the mantle lithosphere governs the rate of root rebound; second, the amount of dynamic uplift caused by root rebound depends strongly on the rheologies of both the upper crust and mantle lithosphere; and third, redistribution of the rebounding root mass causes pure and simple shear within the lower crust and produces subhorizontal planar fabrics which may give the lower crust its reflective character on many seismic images. Key words: numerical models, root rebound, Western Gneiss complex. IN TR O DU C TI O N Continental crust is thickened in collisional mountain chains by such processes as lateral shortening and partial subduction of one continent beneath another. This formation of the crustal roots is opposed by a corresponding increase of gravitational (isostatic) forces, and eventually the thermally softened root rebounds. During rebound, the deeper levels in the crust are expected to react by penetrative–ductile deformation. This has been proposed on theoretical grounds by several workers, under labels such as ‘buoyancy-driven creep’ (Gratton 1989), ‘lateral extrusion’ (Bird 1991), ‘subsurface collapse’ (Avouac & Burov 1996), ‘lower crustal spreading’ (England & Holland 1979; Platt, 199), and ‘lower crustal squeezing’ (Schmeling & Marquart 1990). Platt (1993) and England & Holland (1979) suggested the buoyant return of low-bulk-density, high-pressure metamorphic rocks as a viable mechanism for exhumation. This process has now been well documented in the field. The Western Gneiss Complex in southern Norway * Now at: GEA Consulting, Åsögatan 105, S-11829 Stockholm, Sweden. 556 is interpreted as the remnant, now well exposed, of the Caledonide crustal root (Milnes et al. 1997). The pervasive ductile deformation attributed to root rebound in the Western Gneiss Complex covers a huge area, making it clear that we are dealing with an important and widespread tectonic phenomenon. Much attention has been focused on the lateral gravity spreading of the upper levels of overthickened orogenic crust beneath its free surface (Burg et al. 1984; Burchfiel & Royden 1985; Dewey et al. 1988; Dewey et al. 1989; Buck & Sokoutis 1994). Less attention has been directed towards the ‘upward gravity spreading’ of mountain roots, our topic here. Here, we assume that mantle flow dragging continental crust attached to the subducting plate beneath the non-subducting plate eventually ceases beneath the suture (Meissner 1973; McKenzie 1984; Jones & Nur 1984; Matthews 1986; Cheadle et al. 1987; Meissner & Kusznir 1987). Phase changes increase the density of the root, but the subducted and/or thickened continental crust will still be buoyant with respect to the surrounding mantle, and will therefore rebound upwards gravitationally when continental collision and/or subduction cease and the mantle slab has been heated or detached. Recently, © 1999 RAS Ductile rebound beneath mountain belts 557 root rebound after delamination of a lithospheric root. We also use model results to define the strain patterns associated with such ductile root rebound. N UM ER I C A L M O DE LS Numerical experiments were prepared using a 2-D finite difference code (FDCONE), which solves the Navier–Stokes equations for an initially horizontally layered multicomponent system. The models comprised a 15-km thick upper continental crust (density r=2700 kg m−3, viscosity m=variable between 1020 and 1023 Pa s) overlying a 15-km thick lower continental crust (r=2900 kg m−3, m=1020 Pa s: Fig. 1a). A quarter-circle of lower crust (adjacent to a reflective vertical boundary) simulated an orogenic crustal root penetrating the upper mantle to a depth of 52 km (Fig. 1a). These two layers were resting on a 75-km thick layer simulating the denser mantle lithosphere (r=3300 kg m−3, m=1023 Pa s), giving a maximum lithospheric thickness of 115 km. These layers, in turn, overlay a less viscous asthenosphere (r=3300 kg m−3, m=1019 Pa s). A passive grid was introduced to all layers to act as strain markers to monitor strain during deformation. A Newtonian rheology was assumed for all four units (upper and lower crust, mantle lithosphere and asthenosphere). Since the emphasis is mainly on the first-order controls of ductile root rebound, our models did not take into account any effects of erosion or thermal disturbance, which would accelerate and localize the deformation, as would non-linear rheologies. All the models were assigned a free-slip bottom boundary and non-flexible top boundary. A series of models systematically explored the effects of changing the viscosities of the upper crust and mantle lithosphere with respect to the constant viscosities of the lower crust and the asthenosphere (both 1020 Pa s). The viscosity ratio of the upper crust relative to the lower crust (constant at 1020 Pa s) was increased to give viscosity ratios of 1, 10, 100 and 1000. For each case, four models were run, with the viscosity of the mantle lithosphere varying between 1020 and 1023 Pa s. It was assumed that the root was a 2-D feature, extending infinitely along-strike in the third dimension. In our models we assumed that, as long as a heavy lithospheric root drags the crust down, a crustal root cannot rebound. Our model simulates crustal root rebound that occurs after delamination of the lithospheric root. R ES U LTS Models with the above initial geometry were allowed to evolve under the effect of gravity (Fig. 1a). In all realizations, the Moratta et al. (1998) used numerical models to study the conditions that lead to lithospheric unrooting. Herein, we explore the effect of the viscosity variation of the upper crust and mantle lithosphere on the mechanisms and rates of crustal © 1999 RAS, GJI 139, 556–562 Figure 1. (a) Initial configuration of the model. All boundaries are free slip. Complete root rebound is shown ( b) after 390 Ma in a model with stiff upper crust (viscosity=1023 Pa s), and (c) after 135 Ma with a less stiff upper crust (viscosity=1021 Pa s). In both cases, channel flow in the distal lower crust shears both its upper and lower boundaries due to viscous drag during the lateral spreading of the rebounding root mass. In ( b), only the lower crust deforms internally since the stiff upper crust acts as a lid, whereas in (c) the less stiff upper crust is also deformed with the lower crust. The depression in the upper crust above the root area in (c), indicates isostatic readjustment of the upper crust to the diminishing root and stretching of the upper crust by the laterally spreading root mass within the lower crust. 558 H. A. Koyi et al. lower crustal root rebounds and deforms internally as it rises buoyantly. Because the mantle lithosphere has to fill the void generated by the rising lower crustal root, the rate of root rebound is controlled by the rebounding layer with highest viscosity, that is, the mantle lithosphere. When the viscosity of the mantle lithosphere is decreased by one order of magnitude, the rate of rebound increases by almost one order of magnitude (Figs 2a and b). For cases in which the root rebound takes place within a realistic time window of 10–100 Ma, the viscosity of mantle lithosphere is expected to range between 1022 and 1023 Pa s (Figs 2a and b). However, the viscosity of the upper crust has little influence on the rate of rebound if it is more viscous than the lower crust (Figs 2a and b). Model results also show that changing the rheology of the mantle lithosphere has a much greater effect on the rate of root rebound than changing its thickness (Fig. 2c) (note that the viscosity is changed by three orders of magnitude, and the thickness by a factor of <2). For example, for the same viscosity, a crustal root in a thinner lithosphere (80 km) rebounds five times faster than a crustal root in a thicker lithosphere (135 km) (Fig. 2c). As the root mass rebounds, it displaces and shears the lower crust on its flanks. Our models suggest the formation of two types of fabric within the lower crust during root rebound: (1) a subhorizontal fabric above the root, formed due to vertical shortening by the rebounding mass; and (2) inclined fabrics formed on either side of the root area. The latter fabrics form when the rebounding mass induces channel flow within the lower crust, resulting in viscous drag along the boundaries between the lower and upper crust and lower crust and mantle. The resultant simple shear produces inclined fabrics with opposing senses of shear along each boundary (Figs 1b and c). Away from the root, the effect of the root rebound decreases and the fabric weakens. The fabrics produced within the lower crust by the spreading root mass could explain the reflective character of the lower crust on seismic images. Deep crustal reflections elsewhere have been ascribed to lithological layering and/or overpressured fluids trapped in intergrain cracks and fractures (Cheadle et al. 1987; Jones & Nur 1984; Matthews 1986; McKenzie 1984; Meissner 1973; Meissner & Kusznir 1987), petrophysical layering related to tectonometamorphic processes and mylonitization (Fountain et al. 1984; Hurich et al. 1985; Reston 1987; Smithson et al. 1986), and, in accordance with our model results, to the development of flat-lying shear zones as a result of intense ductile deformation in the lower crust (Rey 1993; Fountain et al. 1994). Fountain et al. (1994) calculated reflection coefficients for the 30–150-m thick eclogite facies shear zones of Caledonian age in the Bergen Figure 2. Plots of crustal thickness (in the root area) versus time for models with (a) a deformable and ( b) a stiff upper crust. The two plots are extraordinarily similar, suggesting that the rate of root rebound is not governed by the rheology of the upper crust. In both cases, however, note the dramatic increase in the rate of rebound as the viscosity of the mantle lithosphere decreases. Considering the realistic time for root rebound to be between 10 and 100 Ma, the viscosity of the mantle lithosphere is expected to be between 1022 and 1023 Pa s. (c) Plot of root thickness with time for three cases showing the influence of the initial thickness of mantle lithosphere on the rate of root rebound. Note that the rate of rebound is not significantly influenced by the thickness of the mantle lithosphere. The time axis should start at time zero. However, in order to plot time logarithmically, the time axis starts at 0.1 Ma. arcs of Norway. They suggested that the ‘high calculated reflection coefficients (0.04–0.14) of these eclogitic shear zones indicate that they are excellent candidates for deep crustal reflectors in portions of crust that experienced high pressure conditions but escaped thermal reactivation’. Although the shear zones in the Western Gneiss Complex are not themselves eclogitic, the above example suggests that these shear zones may still be reflective. © 1999 RAS, GJI 139, 556–562 Ductile rebound beneath mountain belts The intensity and style of deformation within the lower crust are strongly governed by both the viscosity ratio between the upper and lower crust and the viscosity of the mantle lithosphere. In general, highly viscous (1023 Pa s) upper crust deforms too slowly to be affected by shear in the lower crust and acts as a rigid lid. In contrast, as in the case of a highly viscous mantle lithosphere (1023 Pa s), a highly viscous upper crust intensifies shearing in the lower crust by viscous drag along its boundary. When the upper crust and mantle lithosphere are both highly viscous (1023 Pa s), deformation within the lower crust extends to large distances (Fig. 1b), resulting in strong shearing near the root area which diminishes nonlinearly at short distances away from it (Fig. 3a). On the other hand, a less viscous mantle lithosphere deforms more easily so that shear strain is lower along its boundary with the lower crust, but decreases more smoothly towards the external parts of the model (Fig. 3a). In the models, cumulative shear strain in the lower crust depends on distance from the root, and time. Shear strain at 5 km from the root margin increases rapidly until 46 Ma when the root has rebounded half way. Thereafter, the accumulation of shear strain slows down and the lateral gradient decreases (Fig. 3b). In the same model, at 80 km from the root, cumulative strain flattens out as early as after 12 Ma, when only onequarter of the root has rebounded (Fig. 3b). Since such shear strain results from channel flow of the rebounded root material within the lower crust, its intensity depends on the thickness of the ductile lower crust. In the current models, the lowviscosity lower crust is relatively thick (15 km); a thinner lower crust (5–7 km) would act as a narrow corridor for the spreading root material and intense shearing. While our fluid dynamical model has a non-flexible top, the dynamic topography (h) can be calculated a posteriori by equating the vertical stress at the top with an equivalent lithostatic stress produced by the topography h. It implicitly includes the isostatic contribution. Immediately after starting the model calculation the maximum topography is reached as a consequence of the instantaneously acting stress field. We thus neglect the build-up stage of topography during root growth, as it takes place on the timescale of the isostatic relaxation time (104 yr) which is an insignificant time in orogeny. The subsequent dynamic topography above the root area is shown in Fig. 4 for various cases. Viscous stresses retard the dynamic topography beneath the purely isostatic value (which for a 22-km thick root would be 3.26 km). The dynamic uplift (=1600 m, Fig. 4a) above the rebounding root has a positive correlation with the viscosity of the upper crust. A less viscous upper crust (1021 Pa s) accommodates root rebound and lateral spreading of the lower crust by internal deformation and is lifted less than a more viscous upper crust (1023 Pa s), which instead accommodates root rebound by dynamic uplift (Fig. 4b). We emphasize here that changing the boundary condition or using other rheolgies may change the dynamic topography formed by root rebound. Using numerical models of lithosphere under a convergence regime, Moratta et al. (1998) suggested that the order of amplitude of dynamic topography is controlled by the width of the deformation region and by the convergent velocity. The amplitude of dynamic uplift in our models correlates negatively with the viscosity of the mantle lithosphere. For a constant viscosity of the upper crust (1023 Pa s), root rebound causes up to 1600 m of dynamic uplift when the mantle © 1999 RAS, GJI 139, 556–562 559 Figure 3. (a) Plot showing the change in the intensity of shear strain within the lower crust relative to the distance from the root. In the model with a stiff upper crust (viscosity=1023 Pa s), shear strain is more intense and covers a larger lateral distance from the root (see also Figs 1b and c), whereas in a model where the upper crust is less stiff (viscosity=1021 Pa s), shear strain is less intense since root rebound is taken up by internal deformation of both the lower and upper layers. ( b) Plots of shear strain with time at two different locations: 5 and 80 km from the root area, for a model with a stiff upper crust (viscosity=1023 Pa s). The area close to the root undergoes severe shear strain, whereas the area 80 km away from the root undergo only minor amounts of shear strain. lithosphere is less viscous (1021 Pa s), whereas it causes only about 500 m of dynamic uplift when the mantle lithosphere is more viscous (1023 Pa s). The viscosity of the mantle lithosphere and upper crust govern the dynamic topography by controlling the rebound rate of the crustal root. In the case of a stiffer mantle lithosphere relative to the upper crust, the mantle lithosphere controls the rate of rebound and takes most of the buoyancy force, not allowing the buoyancy forces to push up the topography. In the case of a stiffer upper crust 560 H. A. Koyi et al. stiff, thickening of the upper crust will modify our model results only at very late stages. Thermal effects play a significant role in crustal and lithospheric deformation (Marotta et al. 1998; Schott & Schmeling 1998), although it is beyond the scope of this paper to fully incorporate solving the heat equation and include temperaturedependent rheologies. Furthermore, this paper does not address destabilization effects caused by a cold mantle lithosphere. To a first approximation, however, we account for the thermal effect on the rheology by choosing low viscosities for the lower crust and the asthenosphere. CA LE DO N ID ES Figure 4. Plots of dynamic topography above the root axis versus time (a) for models with constant upper crust viscosity (1023 Pa s) but variable mantle lithosphere viscosity, and (b) for models with constant mantle lithosphere viscosity (1022 Pa s) but variable upper crust viscosity. (a) shows that the dynamic topography is inversely proportional to mantle rheology: the higher the viscosity of the mantle lithosphere, the smaller the dynamic topography. In contrast, ( b) shows that the dynamic uplift is directly proportional to the rheology of the upper crust. In nature, these effects would be additional to any gravity spreading of the mountain mass. The time axis should start at time zero. In order to plot time logarithmically, however, the time axis starts at 0.1 Ma. relative to the mantle lithosphere, the faster rebound results in a higher topography as stronger buoyancy forces are transmitted to the upper crust. In the latter case, the rebound rate is governed by the combined system of upper and lower crust. In all models, however, the topography diminishes rapidly to subside above the root area during root rebound (Fig. 4). In models with a deformable upper crust (viscosity= 1020–21 Pa s), the later stages of root rebound induce a phase of extension and isostatic subsidence of the upper crust (Figs 1c and 4b). This extension is caused by lateral spreading of the root mass beneath the upper crust. In nature, stages of postorogenic extension and basin formation may indicate the lateral spreading of a crustal root beneath a weak upper crust. This work considers only a thickened lower crust; orogenic crustal stacking might also involve a thickening of the upper crust. Because the density contrast between the upper and lower crust is relatively small and the upper crust is relatively The fabric produced in the models corresponds to the subhorizontal foliation and stretching lineation in the Western Gneiss Complex of southern Norway. Recently, Milnes et al. (1997) and Milnes (1998) described in detail the structural history along a transect through a collisional orogenic belt (southern Norwegian Caledonides), where subsequent tectonic processes, uplift and erosion have exposed an almost complete cross-section through part of the orogenic root (Western Gneiss Complex). Retro-deformation of the various structural events enables an approximate reconstruction of the root at its deepest level. Eclogites formed at depths of around 100 km in the root, whereas the crustal thickness under the orogenic foreland (the Baltic Shield) remained near 40 km. Evidence from higher levels shows that upper crustal contraction changed to extension soon after or at about the same time. Under these conditions, the root became gravitationally unstable, buoyancy forces no longer being overcome by subduction-related forces. The root rebounded upwards against an upper crustal ‘lid’, which in the southern Norwegian Caledonides, was the unusually stiff and coherent Jotun nappe. Crustal root rebound raised the presently exposed eclogites more or less isothermally, from depths of 60 to 40 km. The surrounding felsic gneisses underwent ductile penetrative ‘pure shear’-type deformation (subvertical shortening, subhorizontal extension) under decreasing pressure, mainly within the amphibolite facies. The results of kinematic analysis suggest that the main Caledonian deformation over a large area of the felsic gneisses was purely the result of gravitational rebound in a process which can be described as ‘inverted gravity spreading’ of the root mass within the lower crust. U R A LS The present-day root observed below the Urals (Thouvenot et al. 1995; Ryzhiy et al. 1992; Juhlin et al. 1996) is likely to have formed during the Palaeozoic, when the East European Craton collided with Asian island-arc terranes and formed the 2500-km long near-linear Uralian orogenic belt (Hamilton 1970; Zonenshain et al. 1990). The collision was followed by some extension. Only minor extension appears to have taken place in the Southern Urals (Brown et al. 1998), while in the Middle Urals the amount of extension appears to have been greater (Juhlin et al. 1998). Post-collisional extension lasted for about 50 Myr and preceded renewed subduction below the eastern boundary of the continent and subsequent deformation of the accreted island-arcs. Reflection seismic experiments over the Middle Urals (Juhlin et al. 1998) reveal a bivergent structural geometry in the upper © 1999 RAS, GJI 139, 556–562 Ductile rebound beneath mountain belts 561 Figure 5. Combined migrated line across the Middle Urals (Juhlin et al. 1998) showing the rebounded root and the internal reflectors above and on either side of the root, which is now only 6 km thick. Note the subhorizontal reflectors above the root within the lower crust and also the gently westward-dipping (towards the root) reflectors on the eastern shoulder of the root within the lower crust. These reflectors could be the result of shearing during lateral spreading of the root mass within the lower crust as the root rebounded. Kilometre ( km) scale refers to distance from the Main Uralian Thrust Fault (MUTF) (positive values to the east and negative values to the west). CUZ: Central Uralian Zone; EUZ: East Uralian Zone; MUFZ: Main Uralain Fault Zone; NF: Normal Fault; PSZ: Prianitchnikova Shear Zone; SMF: Serov-Mauk Fault; TMZ: Tagil-Magnitogorsk Zone; TUTZ: Trans-Uralian Thrust Zone; TZ: Thrust Zone. crust (Fig. 5). A zone of increased seismic reflectivity is observed in the middle crust (23–32 km) along the central 80 km of the profile ( km −10 to 70, Fig. 5). On the eastern shoulder of the root, at about km 55, lower crustal reflectivity becomes prominent at depths of 36–44 km (Fig. 5). This reflectivity has been interpreted by Juhlin et al. (1998) to be different in origin from the middle crustal reflectivity further west. The lower crust below the central portion of the section lacks strong reflectivity down to the Moho depth at about 53 km (Juhlin et al. 1996). Most of the dipping upper crustal reflections can be correlated to geological features such as thrust zones or normal faults. Although the lower crustal reflectivity to the east could be attributed to compression during the westward subduction at the end of the Palaeozoic or extension during the opening of the West Siberian Basin, the middle and lower crustal reflectivity could be due to root rebound. If the observed middle and lower crustal reflectivity in the Middle Urals is a result of extension and shearing during root © 1999 RAS, GJI 139, 556–562 rebound, their spatial and temporal formation is constrained by the following criteria. (1) Roots reached depths of 25–35 km along an 80-km long strip centred below the Tagil Oceanic and Volcanic-Arc Complex. (2) The brittle upper crust must have extended by normal faulting. (3) The reflectivity must have developed after arc accretion, but before the onset of westward subduction: a time window of approximately 50 Ma. The observation of normal faults (some with apparently large throws) on geological and geophysical data in the Middle Urals (Juhlin et al. 1998) is consistent with brittle deformation in the upper crust and ductile deformation in the middle and lower crust. CON CLU SION S Our models suggest intense ductile deformation of the lower crust during root rebound as a mechanism for the formation 562 H. A. Koyi et al. of subhorizontal structures which could give the lower and middle crust its reflective character on seismic images. It should be noted that, in nature, root rebound takes place by ductile flow in the lower crust and surface erosion. Therefore our models only represent an end member, giving the maximum possible deformation in the lower crust and the maximum rebound times compared to cases with erosion. Furthermore, our models show that the strength of the mantle lithosphere dominates the timescale of crustal rebound. The rate of root rebound in the models decreases with increasing viscosity of the mantle lithosphere. Cases of weak mantle lithosphere with fast crustal root rebound may be associated with orogenies that experienced delamination or break-off of strong mantle lithospheres (Schott & Schmeling 1998), or in which the mantle lithosphere has been weakened by the release of fluids during the collision stage. The dynamic topography due to root rebound is directly proportional to the viscosity of the mantle lithosphere. AC KN O WL ED GM E NTS We thank Drs Sadoon Morad and Alasdair Skelton for reviewing this manuscript. Special thanks to Dr Bertram Schott for help with transferring the data and computer manipulation. HAK and CJ are supported by the Swedish Natural Sciences Research Council (NFR). RE FE R ENC ES Avouac, J.P. & Burov, E.B., 1996. Erosion as a driving mechanism of intracontinental mountain growth, J. geophys. Res., 101, 17 747–17 769. Bird, P., 1991. Lateral extrusion of lower crust from under high topography, in the isostatic limit, J. geophys. Res., 96, 10 275–10 286. Brown, D., Juhlin, C., Alvarez-Marron, J., Pérez-Estaún, A. & Oslianski, A., 1998. Crustal scale structure and evolution of an arccontinent collision zone in the southern Urals, Russia, T ectonics, 17, 158–171. Buck, W.R. & Sokoutis, D., 1994. Analogue model of gravitational collapse and surface extension during continental convergence, Nature, 369, 737–740. Burchfiel, B.C. & Royden, L.H., 1985. North–south extension within the convergent Himalayan region, Geology, 13, 679–682. Burg, J.-P., Brunel, M., Gapais, D., Chen, G.M. & Liu, G.H., 1984. Deformation of leucogranites of the crystalline Main Central Sheet in southern Tibet (China), J. struct. Geol., 6, 535–542. Cheadle, M.J.S., Greary, M., Varner, D.H. & Matthews, 1987. Extensional structures on the western UK continental shelf; a review of evidence from deep seismic profiling, in Continental Extensional T ectonics, eds Coward, M.P., Dewey, J.F. & Hancock, L., Geol. Soc. Lond. Spec. Publ. 28, 445–465. Dewey, J.F., Ryan, P.D. & Andersen, T.B., 1988. Orogenic uplift and collapse, crustal thickness, fabrics and metamorphic phase changes: the role of eclogites, in Magmatic Processes and Plate T ectonics, eds Prichard, H.M., Alabaster, T., Harris, N.B.W. & Neary, C.R., Geol. Soc. Lond. Spec. Publ. 76, 325–344. Dewey, J.F., Cande, S. & Pitman, W.C., 1989. Tectonic evolution of the India/Eurasia Collision Zone, Eclog. geolog. Helvitiae, 82, 717–734. England, P.C. & Holland, T.J.B., 1979. Archimedes and the Tauern eclogites; the role of buoyancy in the preservation of exotic eclogite blocks, Earth planet. Sci. L ett., 44, 287–294. Fountain, D.M., Hurich, C.A. & Smithson, S.B., 1984. Seismic reflectivity of mylonite zones in the crust, Geology, 12, 195–198. Fountain, D.M., Boundy, T.M., Austrheim, H. & Rey, P., 1994. Eclogite facies shear zones—deep crustal reflectors, T ectonophysics, 232, 411–424. Gratton, J., 1989. Crustal shortening, root spreading, isostacy, and the growth of orogenic belts, J. geophys. Res., 94, 15 624–15 634. Hamilton, W., 1970. The Uralides and the motion of the Russian and Siberian platforms, Geol. Soc. Am. Bull., 81, 2553–2576. Hurich, C.A., Smithson, S.B., Fountain, D.M. & Humphreys, M.C., 1985. Seismic evidence of mylonite reflectivity and deep structure in the Kettle Dome metamorphic core complex, Geology, 13, 577–580. Jones, T.D. & Nur, A., 1984. The nature of seismic reflections from deep crustal fault zones, J. geophys. Res., 89, 3153–3171. Juhlin, C., Knapp, J.H., Kashubin, S. & Bliznetsov, M., 1996. Crustal evolution of the Middle Urals based on seismic reflection and refraction data, T ectonophysics, 264, 21–34. Juhlin, C., Friberg, M., Echtler, H.P., Hismatulin, H., Rybalka, A., Green, A.G. & Ansorge, J., 1998. Crustal structure of the Middle Urals: results from the Europrobe Seismic Reflection Profiling in the Urals (ESRU) experiments, T ectonics, 17, 710–725. McKenzie, D.P., 1984. A possible mechanism for epeirogenic uplift, Nature, 307, 616–618. Matthews, D.H., 1986. Seismic reflections from the lower crust around Britain, in T he Nature of the L ower Continental Crust, eds Dawson, J.B., Carswell, D.A., Hall, J., Wedepohl, K.H., Geol. Soc. Lond. Spec. Publ. 24, 11–21. Meissner, R., 1973. Moho as a transition zone, Geophys. Surv., 1, 195–216. Meissner, R. & Kusznir, N., 1987. Crustal viscosity and the reflectivity of the lower crust, Ann. Geophys., Ser. B5, 4, 365–373. Milnes, A.G., 1998. Alpine and Caledonide tectonics: a brief comparative study, Geologiska Förenings Förhandlingar (GFF), 120, 237–247. Milnes, A.G., Wennberg, O.P., Skår, Ø & Koestler, A.G., 1997. Contraction, extension and timing in the south Norwegian caledonides: the sognefjord transect., in Orogeny T hrough T ime, eds Burg, J.-P. & Ford, M., Geol. Soc. Lond. Spec. Publ. 121, 123–148. Moratta, A.A., Fernandéz, M. & Sabadini, R., 1998. Mantle unrooting in collisional settings, T ectonophysics, 296, 31–46. Platt, J.P., 1993. Exhumation of high-pressure rocks; a review of concepts and processes, T erra Nova, 5, 119–133. Reston, T.J., 1987. Spatial interference, reflection character and the structure of the lower crust under extension; results from 2-D seismic modelling, Ann. Geophys., 5B (4), 339–347. Rey, P., 1993. Seismic and tectono-metamorphic characters of the lower continental crust in Phanerozoic areas: a consequence of post-thickening extension, T ectonics, 12, 580–590. Ryzhiy, B.P., Druzhinin, V.S., Yunusov, F.F. & Ananyin, I.V., 1992. Deep structure of the Urals region and its seismicity, Phys. Earth planet. Inter., 75, 185–191. Schmeling, H. & Marquart, G., 1990. A mechanism for crustal thinning with lateral extension, Geophys. Res. L ett., 17, 2417–2420. Schott, B. & Schmeling, H., 1998. Delamination and detachment of a lithospheric roof, T ectonophysics, 296, 225–247. Smithson, S.B., Johnson, R.A. & Hurich, C.A., 1986. Crustal reflections and crustal structure, in Reflection Seismology: the Continental Crust, eds Barzangi, M. & Brown, L., Geodynamics Series 14, 21–32, AGU, Washington. Thouvenot, F., Kashubin, S.N., Poupinet, G., Makovsky, V.V., Kashubina, T.V., Matte, P. & Jenatton, L., 1995. The root of the Urals: evidence from wide-angle reflection seismics, T ectonophysics, 250, 1–13. Zonenshain, L.P., Kuzmin, M.I., Natapov, L.M. & Page, B.M., 1990. Geology of the USSR; a Plate-T ectonic Synthesis, ed. Page, B.M., Am. geophys. Un., 21, 27–54. © 1999 RAS, GJI 139, 556–562
© Copyright 2026 Paperzz