Geology Reference
In-Depth Information
might produce. To overcome such limitations,
clever new strategies must be developed that
both honor everyday atmospheric physics, while
acknowledging the importance of extreme events
in geomorphology that reflect the nonlinearity of
many geomorphic processes.
We remain challenged by having to choose
proper initial conditions in our models: these
can play an important role in setting the final
look of the landscape. Their significance is
reflected by the fact that the most robust feature
of a landscape, and hence the most difficult to
change, is the planview shape of the drainage
pattern: streams most efficiently cut downward,
and, unless strongly forced, do not wander
significantly once established in a bedrock chan-
nel. Unfortunately, it is most difficult to establish
with confidence through geological evidence
what the channel pattern might have been at the
inception of an episode of mountain building.
Despite such challenges, the process of stream
capture, as well as the distortion of channel pat-
terns through crustal deformation of the rock
mass in which they are embedded, need to be
addressed in future models. These processes
may have played key roles in the evolution of
portions of several great collisional ranges.
Finally, the tectonic-geomorphic community
faces the challenge of acknowledging and some-
how addressing the role of rock type in govern-
ing the pace of erosion. More specifically, we
must learn what it is about a rock that matters in
establishing its susceptibility to erosion by spe-
cific processes. The community wallows about
in a world of “
k
”s - call it “the
k
-problem” - and
in general simply chooses a “
k
” or set of “
k
”s
that yield the pattern found in nature: the stream
profile, the hillslope profile, the glacial valley
profile, and so on. It is indeed time to attack this
issue directly. In tectonically active settings,
faults juxtapose rocks of very different charac-
ter. The tectonic activity itself can modify the
rock by causing it to crack as it is made to go
around fault bends (Molnar
et al.
, 2007).
It should not be surprising to the geomorpholo-
gist with any knowledge of specific processes
responsible for lowering either river or glacial
beds that the availability of cracks for plucking
or quarrying processes to exploit is crucial in
determining the rate at which this process can
remove rock - see Duhnforth
et al.
(2010) for
the case of the Yosemite landscape. Similarly,
whether rocks arrive at the surface pre-fractured
due to tectonics or are fractured by top-down
geomorphic processes appears to modulate
the size and depth of landslides (Clarke and
Burbank, 2010a, 2011). Given the importance
of rock type and of the roles of faults in both
juxtaposing rocks of differing types and gener-
ating cracks in rock masses, the quantification
of these effects and their incorporation into
landscape evolution models is a challenge that
should be faced head-on.
These challenges and the general importance of
modeling are now acknowledged by the full com-
munity. A recent manifestation of this is consider-
able investment in a Community Surface Dynamics
Modeling System (CSDMS) (Anderson
et al.
,
2004). This effort is designed to provide a clearing
house for existing models, to promote efficient
linkages between these models, and to make data
sets relevant to such models readily available.
Summary
We have briefly explored a variety of modeling
techniques and results that cover a range of spa-
tial and temporal scales and that allow treatment
of landscape elements from river profiles to fault
scarps to mountain ranges. Both the tectonic
forcing of the system and the suite of geomor-
phic processes acting to modify the tectonically
generated landscape must somehow be incorpo-
rated into the model. One must chose a mode-
ling strategy that suits the problem. This choice
includes decisions about whether to work in one
dimension or two, as well as the style of model
(analytical or numerical; finite difference, finite
element or boundary element). One must comb
the literature to assemble any existing data
against which the model can be tested, and
can then utilize initial model results to guide
the efficient collection of new data that will
maximally discriminate among the possible
landscape evolution scenarios.
We reiterate that the purpose of modeling in
this field is most often the development of insight
