A  Lecture Supported by the Hilton Johnson Endowment

Erosion is a vector: What makes lateral cutting by landslides and meandering rivers so important?

Colin Stark
Columbia University

 

All current landscape evolution models treat erosion as a strictly vertical process with topography rendered as a 2+1D surface (like a DEM). In reality, many important erosion processes have a significant horizontal velocity component which we ignore at our peril. For example, river channel meandering is largely a lateral erosion process (whether the channels are net-aggradational or incised into bedrock) which gives channels a mutability that is fundamental to catchment evolution. Another example is horizontal erosion by landsliding along ridges, which underpins drainage divide migration. In this talk I will describe a couple of simple models of landscape evolution that explicitly incorporate lateral erosion processes, and I will show how, in so doing, such models capture key aspects of catchment dynamics - some of which are rather unexpected. In particular I will: (1) show how drainage divides may behave as low-dimensional non-linear oscillators; (2) argue that divide migration by landsliding is the key to network "optimization"; (3) document (from field, DEM and satellite data) how lateral erosion of bedrock channels is a useful recorder of bedrock erodibility and extreme rainfall/flood frequency.