The generation and degradation of marine terraces

Marine terraces are ephemeral planar landforms. While tectonic and climatic forcings responsible for the generation of existing marine terraces have operated for at least 1 Myr, terraces have been completely removed by erosion above a given altitude (and hence above a given age). Above this altitude, the landscape has forgotten that it was once terraced. We ask what controls this characteristic time‐scale, which we term the ‘forget time’, in a landscape. We approach the problem with simple scaling arguments, and 1‐D numerical models of landscape evolution. Using a simple cliff erosion model with a realistic sea‐level history, rock uplift and a cliff retreat rule, we find that the most important means of terrace removal is through the deeper transgression of a subsequent sea cliff into the landmass. The sequence of preserved terraces depends upon the history of sea cliff incursion into the landmass. The extent of sea cliff incursion depends on the duration of the sea‐level highstand, the far‐field wave energy input and the degree to which bathymetric drag dissipates wave energy. This portion of the marine terrace survival problem is an example of a common problem in geomorphology, in which the record of past tectonic or climatic events is rendered incomplete by the potential for younger events to wipe the topographic slate clean. While sea cliffs decay through time, their form can still be recognized many hundreds of thousands of years after formation. This reflects the diffusive nature of their decay: early rapid evolution and lowering of maximum slopes yields to slower rates through time. Incision by streams, on the other hand, is rapid, as the streams respond to base‐level history driven by sea‐level changes. The rate of incision reflects the local climate conditions, and is limited by the rate of base‐level fall. The principal means of vanquishing a marine terrace is by backwearing of slopes adjacent to these incising streams. The forget time should be proportional to the spacing between major incising streams and to the angle of hillslope stability, and should be inversely proportional to the rate of channel incision. This yields an overestimate of the forget time, as the terraced interfluves are reduced as well by the headward incision of tributary streams. The resulting landscape may be viewed as a terraced fringe separating the sea from the fully channellized landscape. Over time‐scales corresponding to many glacial–interglacial sea‐level oscillations, this fringe can achieve a nearly steady width. The rate of generation of new terraced landscape, reflecting the uplift rate pattern, is then balanced by the rate at which the terraces are erased beyond recognition by channel and hillslope processes. The width of this fringe should depend upon the precipitation, and upon the distance to the nearest drainage divide, both of which limit the maximum power available to drive channel incision.