The tectonic history of Kohistan, northern Pakistan, involves two collisional events. Cleavage and folding developed at 90-100 Ma along the northern suture between the Kohistan island arc and the Asian plate. At the same time there was major folding and shearing of the lower part of the Kohistan arc, approximately 100 km south of the suture. This deformation was followed by ocean subduction south of the Kohistan arc, generating the Kohistan calc-alkaline batholith, with subsequent ocean closure during the Eocene and obduction of the Kohistan arc, together with the adjacent part of the Asian plate, over the Indian continental crust. The construction of balanced cross-sections through the imbricated upper part of the Indian continental crust, in the footwall to this southern suture indicates a minimum displacement of 470 km, requiring the western Himalayan hinterland to be underlain by a large wedge of Indian middle to lower crust. There is some shortening of the overriding Kohistan and Asian plates by thrusts and shear zones, but it is insufficient to satisfy the palaeomagnetic data; there must be major crustal shortening, involving thrusts, in the Hindu Kush and Pamirs north of Kohistan. The post-Eocene thrust direction, which for most of Pakistan is towards 160°, is almost perpendicular to that immediately to the east in the Himalayan belt, generating complex refolded thrust patterns in the Hazara syntaxis and large scale folding and rapid uplift with associated brittle faulting and seismic activity adjacent to the Nanga Parbat syntaxis. These different thrust trends indicate that major thrust movement as well as the folds and deformation fabrics, cannot always be related to plate movement vectors, but are modified by, or develop from, complex rotations during place collision or from the gravitational spreading of a thickened crust. A regional approach is required to recognize and correctly attribute the various components in thrust displacements.

Summary: Collision mountain belts have recently been interpreted as being large scale stacked sedimentary basins, where the wholesale reactivation of pre-existing normal faults exerts the primary control on orogenic style. This model is explored here using examples from the Western Alps. The use of stratigraphic separation across faults, in the absence of corroborative sedimentological data, provides an ambiguous criterion to determinate fault reactivation. Variations between ‘compressional’ and ‘extensional’ stratigraphic separations across the Frontal Pennine thrust are more likely to be due to regional unconformity patterns developed in lower Tertiary times rather than normal fault reactivation along the eastern margin of the external zones. Mesozoic-Tertiary basin structure has been important in controlling the sites of major thrust ramps where stratigraphic horizons suitable for activation as thrust detachment zones have been offset. Examples include the thrust front in the Jura (Triassic detachment dropped to W by faults bounding the Bresse basin) and Vercors (basement uplifted to W at Faille de l’Isere, bounding the Mesozoic sub-Alpine basin to W). Buttressing structures, where thrust displacement is converted into vertical pure shear, may indicate buried normal faults but can occur at either hinterland or foreland-directed basin faults. The overall structural style in the sub-Alpine thrust belt indicates detachment between basement and cover with basement involvement along the Sole thrust which carries the external basement massifs. There is no indication from surface geology that this thrust originates from reactivation of a Mesozoic basin fault although this process may have occurred at depth. Similar styles of detachment-dominated thrusting with large horizontal relative to vertical displacements also characterize internal Alpine structure. Preliminary results from the ECORS-CROP deep seismic reflection profile across the Western Alps support the detachment thrusting model although they do not rule out minor normal fault reactivation.

The structural restoration of collision orogenic belts onto crustal templates provides important insights into the tectonic evolution, deep structure and amounts of plate convergence after the initial contact between two continental masses. Balanced cross-sections have been constructed, parallel to the local displacement directions, across the western Alps and western Himalayas and demonstrate very large amounts of crustal shortening above intra-crustal detachments. To achieve a balance, substantial volumes of lower crust must have been subducted beneath the two tectonic hinterlands. A model of eclogite metamorphism is invoked to facilitate this subduction and to explain the varying isostatic responses of the Alpine and Himalayan hinterlands. Patterns of eclogite metamorphism are controlled by the geometry of thrust profiles on a lithospheric scale: the development of crust-mantle detachments being of crucial importance. Such a profile is proposed for the Himalayas, suggesting that relatively small volumes of the footwall crust succumbed to eclogite metamorphism. In the Alps however, a steeper thrust profile apparently developed, emplacing mantle onto crust, causing wholescale eclogite metamorphism in the underlying Franco–Swiss crust. The resultant density increase would provide a mechanism of isostatic collapse and flexural subsidence in the Po plain region.