The competition between the impact of inherited weaknesses and plate kinematics determines the location and style of deformation during rifting, yet the relative impacts of these ‘internal’ and ‘external’ factors remain poorly understood, especially in 3D. In this study, we used brittle‐viscous analogue models to assess how multiphase rifting, that is changes in plate divergence rate or direction, and the presence and orientation of weaknesses in the competent mantle and crust, influences rift evolution. We find that the combined reactivation of mantle and crustal weaknesses without any kinematic changes already creates complex rift structures. Divergence rates affect the strength of the weak lower crustal layer and hence the degree of mantle‐crustal coupling; slow rifting decreases coupling, so that crustal weaknesses can dominate deformation localisation and surface structures, whereas fast rifting increases coupling and deformation related to mantle weaknesses can have a dominant surface expression. Through a change from slow to fast rifting mantle‐related deformation can overprint structures that previously formed along (differently oriented) crustal weaknesses. Conversely, a change from fast to slow rifting may shift deformation from mantle‐controlled towards crust‐controlled. When changing divergence directions, structures from the first rifting phase may control where subsequent deformation occurs, but only when they are sufficiently well developed. We furthermore place our results in a larger framework of brittle‐viscous rift modelling results from previous experimental studies, showing the importance of general lithospheric layering, divergence rate, the type of deformation in the mantle, and finally upper crustal structural inheritance. The interaction between these parameters can produce a variety of deformation styles that may, however, lead to comparable end products. Therefore, careful investigation of the distribution of strain localisation, and to an equal extent of basin depocenter locations over time is required to properly determine the evolution of complex rift systems, providing an incentive to revisit various natural examples.

We investigate lithosphere necking using two‐dimensional thermo‐mechanical numerical simulations without strain softening or weakening mechanisms. The models have an initial small sinusoidal perturbation of the Moho depth, whose wavelength corresponds to the model width. Applied boundary conditions (constant extension velocity or bulk extension rate) and initial model width significantly impact the necking dynamics. For constant bulk extension rates, wider models generate more intense necking with locally higher strain rates, whereas for constant velocity extension, models evolution is similar independent on their initial width. However, the width of the final necking zones ranges consistently between 45 and 105 km, independent on the type of applied boundary conditions and the initial Moho wavelength. The modeled widths are similar to along dip necking zones widths of natural rifted margins that formed during a single, unidirectional, and relatively continuous extensional event (e.g., Iberia‐Newfoundland margins, Porcupine Basin, Gulf of Aden). When the crust is mechanically decoupled from the mantle by a weak ductile lower crust, models exhibit three characteristic stages: (1) distributed thinning and extension associated with progressive subsidence; (2) upper mantle necking compensated by flow of the weak lower crust, which hampers both crustal thinning and subsidence at the rift center; and (3) crustal necking associated with fast subsidence after the mantle has necked. Decoupled models display regions of relatively thick crust on one or both sides of the rift center, comparable to the Galicia, Rockall, Hatton, and Porcupine Banks along the North Atlantic rifted margins.

During lithospheric extension, localization of deformation often occurs along structural weaknesses inherited from previous tectonic phases. Such weaknesses may occur in both the crust and mantle, but the combined effects of these weaknesses on rift evolution remain poorly understood. Here we present a series of 3D brittle–viscous analogue models to test the interaction between differently oriented weaknesses located in the brittle upper crust and/or upper mantle. We find that crustal weaknesses usually express first at the surface, with the formation of grabens parallel to their orientation; then, structures parallel to the mantle weakness overprint them and often become dominant. Furthermore, the direction of extension exerts minimal control on rift trends when inherited weaknesses are present, which implies that present-day rift orientations are not always indicative of past extension directions. We also suggest that multiphase extension is not required to explain different structural orientations in natural rift systems. The degree of coupling between the mantle and upper crust affects the relative influence of the crustal and mantle weaknesses: low coupling enhances the influence of crustal weaknesses, whereas high coupling enhances the influence of mantle weaknesses. Such coupling may vary over time due to progressive thinning of the lower crustal layer, as well as due to variations in extension velocity. These findings provide a strong incentive to reassess the tectonic history of various natural examples.

Although the Wilson cycle is usually considered in terms of wide oceans floored with normal oceanic crust, numerous orogens result from the closure of embryonic oceans. We discuss how orogenic and post-orogenic processes may be controlled by the size/maturity of the inverted basin. We focus on the role of lithospheric mantle in controlling deformation and the magmatic budget. We describe the physical properties (composition, density, rheology) of three types of mantle: inherited, fertilized and depleted oceanic mantle. By comparing these, we highlight that fertilized mantle underlying embryonic oceans is mechanically weaker, less dense and more fertile than other types of mantle. We suggest that orogens resulting from the closure of a narrow, immature extensional system are essentially controlled by mechanical processes without significant thermal and lithological modification. The underlying mantle is fertile and thus has a high potential for magma generation during subsequent tectonic events. Conversely, the thermal state and lithology of orogens resulting from the closure of a wide, mature ocean are largely modified by subduction-related arc magmatism. The underlying mantle wedge is depleted, which may inhibit magma generation during post-orogenic extension. These end-member considerations are supported by observations derived from the Western Europe–North Atlantic region.