The relationship between morphology and surficial geology is used to quantify the denudation that has occurred across southwestern Africa since the fragmentation of Gondwana during the Early Mesozoic. Two main points emerge. Significant denudation, of the order of kilometers, is widespread except in the Kalahari region of the continental interior. The denudation is systematically distribution so that the continental exterior catchment, draining directly to the Cape basin, is denuded to a greater depth than the interior catchment inland of the Great Escarpment. The analysis also implies that the majority of the denudation occurred before the beginning of the Cenozoic for both the exterior and interior catchments. Existing models of landscape development are reviewed, and implications of the denudation chronology are incorporated into a revised conceptual model. This revision implies that the primary effect of rifting on the subsequent landscape evolution is that it generates two distinct drainage regimes. A marginal upwarp, or rift flank uplift, separates rejuvenated rivers that drain into the subsiding rift from rivers in the continental interior that are deflected but not rejuvenated. The two catchments evolve independently unless they are integrated by breaching of the marginal upwarp. If this occurs, the exterior baselevel is communicated to the interior catchment that is denuded accordingly. Denudation rates generally decrease as the margin evolves, and this decrease is reinforced by the exposure of substrate that is resistant to denudation and/or a change to a more arid climate.

The observations do not reveal a particular style of smaller-scale landscape evolution, such as escarpment retreat, that is responsible for the differential denudation across the region. It is proposed that numerical model experiments, which reflect the observational insights at the large scale, may identify the smaller-scale controls on escarpment development if the model and natural systems are analogous. Four numerical experiments are presented in which the roles of antecedent topography, resistant substrate, climate change, and lowering the baselevel of the interior catchment are investigated for an initially high elevation margin bordered by an escarpment. The model results suggest several styles of landscape evolution that are compatible with the observations. Escarpments may retreat in a regular manner, but they also degrade and are destroyed, only to reform at the drainage divide between exterior and interior catchments. ¿ American Geophysical Union 1994