Lithospheric necking and magmatic intrusion are thought to be the primary processes responsible for narrow continental rifts. There is much less agreement as to the cause of wide continental rifts. We attempt to shed light on this problem using three approaches. First, we derive approximate relations between the change in force needed for extension and parameters that should control a number of tectonic processes. Processes resulting in a decrease in tectonic force: and so localization of rifting, include: lithospheric necking, magmatic intrusion and loss of brittle layer cohesion. Processes leading to delocalization include viscous flow, local compensation of crustal thickness variations, and regional compensation of extension-related relief. Second, we review the work of others on linear stability analysis applicable to the earliest phase of lithospheric extension. This approach predicts that all conditions lead to wide rift formation and never predicts narrow rift development. Our scaling relations suggest that the pattern of strain for finite extension could be different than that predicted by linear stability theory. Lastly, to test this idea: we carried out a series of numerical experiments for finite-amplitude extension. We find that when strains are quite small the numerical model matches the prediction of linear stability analysis. However, when strains are large a narrow rift can develop. For reasonable model parameters, we get wide rifting only when local crustal buoyancy effects dominate localizing processes. Consistent with our simple scaling relations, this only occurs when the crust is much thicker than the brittle lithosphere. This agrees with observations that wide rifts only occur in regions of higher than normal heat flow with relatively thick crust.
199EDTimes Cited:16Cited References Count:55