Modelling Landscape Evolution - Environmental Modelling: Finding Simplicity in Complexity - page 296

19

Modelling Landscape
Evolution

Peter van der Beek
Universite Joseph Fourier, Grenoble, France

scales, the development of geochronological techniques
such as cosmogenic nuclide dating (e.g. Cerling and Craig,
1994; Bierman and Nichols, 2004) or low-temperature
thermochronology (e.g. Brown et al ., 1994; Braun et al .,
2006) allowed denudation rates to be constrained on
temporal scales of thousands to millions of years, and
increased computer power permitted the development of
numerical models of landscape development.
This chapter is dedicated to reviewing the development,
application and validity of such 'landscape-evolution'
or 'surface-process' models. Several such reviews have
been published throughout the last decade (Beaumont
et al ., 2000; Coulthard, 2001; Martin and Church, 2004;
Willgoose, 2005; Codilean et al ., 2006; Tucker and Han-
cock, 2010), each from its own perspective. Others have
discussed the use of such models within an epistemolog-
ical context (Kerr, 1997; Bras et al ., 2003; Dietrich et al .,
2003). Here, I will discuss the construction of landscape-
evolution models and the philosophy underlying them,
as well as the prime ingredients of such models: the
numerical algorithms describing different geomorphic
processes. I will subsequently discuss different attempts
at, and approaches to, model validation and calibration.
I will review the coupling of surface process models to
other numerical models, in particular those predicting
tectonic motions in the crust and lithosphere, as well as a
number of studies applying landscape-evolution models
to specific geomorphic/tectonic contexts. I conclude with
a brief perspective of where the field stands today and
what directions might be taken in the future.

19.1 Introduction

The relationships between landscape form and the pro-
cesses that shape it lie at the heart of the science of geomor-
phology. Whereas these relationships can be observed or
inferred in a relatively straightforward manner on small
spatial and temporal scales, they have proven elusive on
the spatial scale of an entire landscape and on geologi-
cal timescales. Landscapes evolve through a multitude of
interacting processes that act on scales that are very much
smaller than those on which the resulting landscape and
its evolution can be observed. Nevertheless, theories for
landscape evolution on geological timescales (e.g. Davis,
1899; Penck, 1953; Hack, 1960) have strongly influenced
geomorphological studies throughout the early twentieth
century. In the second half of the last century, however,
many workers turned away from these 'grand' but inher-
ently untestable theories to study processes on smaller
spatial and temporal scales (see reviews by Summerfield,
1991; Burbank and Anderson, 2001).
The study of landscape evolution on large spatial and
temporal scales has attracted regained interest since the
early 1990s. This interest has come to a large extent from
the geodynamics community, which started to appreciate
the importance of variations in the free upper surface
of the earth to processes within it (Beaumont et al .,
1992; Hoffman and Grotzinger, 1993). By that time,
new data and methods were also available: widespread
digital topographic data of sufficient resolution allowed
landscape form to be analyzed on unprecedented spatial