Sequential faulting explains the asymmetry and extension discrepancy of conjugate margins.

During early extension, cold continental lithosphere thins and subsides, creating rift basins. If extension continues to final break-up, the split and greatly thinned plates subside deep below sea level to form a conjugate pair of rifted margins. Although basins and margins are ubiquitous structures, the deformation processes leading from moderately extended basins to highly stretched margins are unclear, as studies consistently report that crustal thinning is greater than extension caused by brittle faulting. This extension discrepancy might arise from differential stretching of brittle and ductile crustal layers, but that does not readily explain the typical asymmetric structure of conjugate margins-in cross-section, one margin displays gradual thinning accompanied by large faults, and the conjugate margin displays abrupt thinning but smaller-scale faulting. Whole-crust detachments, active from early in the rifting, could in theory create both thinning and asymmetry, but are mechanically problematical. Furthermore, the extension discrepancy occurs at both conjugate margins, leading to the apparent contradiction that both seem to be upper plates to a detachment fault. Alternative models propose that much brittle extension is undetected because of seismic imaging limitations caused either by subseismic-resolution faulting, invisible deformation along top-basement 100-km-scale detachments or the structural complexity of cross-cutting arrays of faults. Here we use depth-migrated seismic images to accurately measure fault extension and compare it with crustal thinning. The observations are used to create a balanced kinematic model of rifting that resolves the extension discrepancy by producing both fault-controlled crustal thinning which progresses from a rift basin to the asymmetric structure, and extreme thinning of conjugate rifted margins. Contrary to current wisdom, the observations support the idea that thinning is to a first degree explained by simple Andersonian faulting that is unambiguously visible in seismic data.