S U M M A R YBoth extension rate and rheology have been shown to influence the style of continental rifting. So far, their effect on the dynamics of extension has been discussed separately in the literature, leading to diverging conclusions. The rate of extension mostly affects the dynamics of rifting through the occurrence of syn-rift cooling at low extension rates. This cooling results in an increase of strength if the lithosphere has a temperature-dependent rheology. England (1983) first suggested that this effect will limit the amount of extension that an area may undergo. Kusznir & Park (1987) argued more generally that it will induce a widening of the rift, and concluded that extension rate is the major control on wide versus narrow rifting. Later studies (Bassi 1991; Bassi, Keen & Potter 1993). however, have demonstrated the importance of rheological layering within the lithosphere for the mode of continental extension by showing that a plasticity-dominated rheology will lead to rapid and narrow necking while a creep-dominated rheology results in wider rifts and margins. Gravitational stresses arising from density anomalies due to lithosphere thinning also affect the style of rifting (Buck 1991). In this paper, we examine the combined effect of these different parameters in order to reconcile earlier studies and assess the importance of syn-rift cooling for the process of lithospheric necking. The approach, introduced in a previous study (Bassi 1991), uses a finite-element technique to simulate extension of a vertical cross-section of the lithosphere. The effect of cooling on the necking pattern depends strongly on the rheology of the thinned lithosphere when it starts to cool. If heat diffusion always induces strengthening and delays rupture in time, it only modifies significantly the geometry of the rift when the upper mantle is weak and viscous. In this case, the locus of maximum strain rate migrates laterally once the central area starts to cool and harden, as hypothesized by Kusznir & Park (1987); eventually, strain concentrates at the transition between deformed and undeformed lithosphere. We predict that this pattern will induce asymmetric breakup and produce a very narrow margin and a wide, uniformly thinned, conjugate margin. In all other cases, i.e. when the upper mantle is initially in the low-plasticity regime, or if it reaches the yield stress during extension, the rate of extension appears to have little effect on the rifting pattern.

No abstract

We present the results of a dynamical model of lithospheric rifting and rupture which show that a wide range of crustal thinning patterns across rifted passive margins can be produced by varying the steady state geotherm, lithospheric composition (dry versus wet materials), and strain rate. The basic mechanism of continental rupture is assumed to be passive rifting and necking. We use a numerical thermomechanical model of lithosphere extension based on a finite element approach. When plasticity is significant (i.e., at lower temperatures or for "harder" materials) deformation is unstable and thinning takes place abruptly, over a narrow area. Conversely, a progressive thinning across the margin is observed when creep is dominant (i.e., in warm or ductile conditions). Cooling and associated hardening of the thinned area can occur during extension and cause the locus of extension to migrate laterally. In these circumstances, rupture is likely to take place asymmetrically along one edge of the thinned area, producing a narrow margin and a very wide conjugate. The eastern margins of Canada and their conjugates across the North Atlantic provide examples which cover this range of theoretical profiles. The crustal thinning patterns, inferred from deep seismic data, and the duration of rifting compare well with model results. We discuss also the constraints that these geodynamical models provide on such current issues as the seismic reflectivity of the lower crust, or the location of the ocean-continent boundary in wide areas supposedly underlain by 5-km thin continental crust. controlled the distribution of deformation across and along the margin. WIDE VERSUS NARROW BASINS: CONTROLLING FACTORS Let us review the processes and parameters that may influence the crustal thinning pattern and margin width. Most authors have emphasized the importance of one factor while overlooking the others. Indeed, the number of parameters involved in any modeling of lithospheric rifting and the complexity of its theology make it very difficult to examine the effect of each parameter individually and systematically, in a dynamical model. The most extensively discussed factor has been the effect of heat diffusion during extension. England [1983] first pointed out that cooling during extension may cause the vertically averaged strength of the lithosphere to increase because of the temperature dependence of rock rheology. This strengthening will eventually inhibit further deformation, thus limiting the extensional strain attainable in a given area. This happens when the replacement of crust by stronger mantle dominates weakening of the lithosphere due to thinning. Such Planet. Sci. Lett., 40, 25-32, 1978. Nichols, B.C., and G. Bassi, Use of ADINA to model large deformation of the Earth's lithosphere, Comoeut. Struct., 32 (3/4), 761-777, 1989. Pinheiro, L.M., R.B. Whitmarsh and P.R. Miles, The ocean-continent boundary off the western continental margin of Iberia. II, Crustal structure in the Tagus Abyssal Plain, Geophys. J. Int., in press, 1993....

Continental break-up, which precedes oceanic accretion, probably results from an unstable extension of the lithosphere, analogous to necking of metals when they are submitted to tension. By reason of complexity of the rheology, no conclusion about lithospheric extension stability may be reached by an apriori analysis. We thus examine directly the evolution, when the lithosphere is stretched, of lateral inhomogeneities, represented in our example by small-scale variations of thickness. The rheological model is derived from the hypotheses of Brace & Kohlstedt (1980) and is consistent with the results of rock mechanics. The lithosphere consists of three or four layers of varying thicknesses and mechanical properties. The brittle upper crust and, eventually, the brittle part of the mantle are assimilated to perfectly plastic media and are described, in a state of uniform extension, by a constant viscosity. In the lower crust and ductile mantle lithosphere, the effective viscosity is supposed to be exponential. The mechanical model relies on a perturbation method developed by Fletcher & Hallet (1983), among others. Contrary to previous published results, no unstable behaviour of the lithosphere is observed unless the latter is more dense than the asthenosphere, in which case a gravitational instability may develop. This discrepancy can be explained by differences in assumptions concerning the variation of strength in the lithosphere, as yet poorly constrained by the data. We observe a great sensitivity of the results to the strength stratification and to the artificial discontinuities of density or viscosity implied by the models.