Geography Reference
In-Depth Information
In the second case, the approach can be more inductive.
One of the main benefits of images compared to field
measures is the possibility to exhaustively analyse the
longitudinal structure of corridors all along a channel or
within a network (e.g., Piegay et al., 2000; Petts et al.,
2000; Alber and Piegay, 2011), to locate specific features
or thresholds between them, to define types of biophysical
features, such as geomorphic reaches (Rollet, 2007), and
to highlight regional organisation.
In terms of modelling, the aim is to provide regional
models mainly linking width or other planimetric factors
with catchment size and combined planimetric features
to provide hydraulic meaning and process-based under-
standing. An example is provided by Lagasse et al. (2004)
who explored prediction of meander migration rates
based on GIS analysis of a vast database of single bends.
Another is provided by Constantine and Dunne (2008)
who worked with images of Google Earth to measure
channel and oxbow-lake characteristics of 30 world rivers
to identify the factors controlling oxbow lake develop-
ment (Figure 11.3). They predicted the size-frequency
distributions of lakes of five world river reaches using
only channel sinuosity and the geometric mean oxbow
lake length variance. The temporal rate of cut-off can be
estimated using channel sinuosity, the fraction by which
cut-off reduces channel length, and the rate at which the
reach lengthens by meander growth.
When working in an inter-site comparative approach,
scaling and size effects become a critical issue. It therefore
appears that many parameters are scale-dependant, as
recently stated by Dodov and Foufoula-Georgiou (2004).
These authors notably showed law dependencies between
channel morphology (sinuosity, meander wavelength,
and radius of curvature) and consequent variations
in channel cross-sectional shape with scale (e.g., mean
velocity and discharge under different flow conditions)
for at-station hydraulic geometry (HG) using between
70 and 100 stations in Nebraska, Kansas, Missouri and
Oklahoma.
Following the increasing availability of aerial photo
information, development of GIS tools to analyse geo-
graphical objects and their attributes is emerging as shown
by G uneralp and Rhoads (2008) or Heo et al. (2009) to
continuous characterisation of single-thread planform
geometry or to continuous sampling of widths (Lauer
and Parker, 2008). It is possible to draw or detect poly-
gons and describe them by their radiometric values or by
their geometry and size. The object can be basically char-
acterised by a manual measure such as length (channel
lengths mainly rated by different other lengths, braided
index, sinuosity rate) or by advanced remote sensing
procedures. Generic geographical objects can be extracted
from real ones having a biophysical meaning based on
GIS rules (e.g., a centreline, an elementary polygon of
constant length), mainly created as intermediate objects
for geomatic treatments. Longitudinal polygons or linear
features, such as the valley corridor or the active channel
area can be also represented synthetically by a single line.
These generic objects are used to provide spatial reference
to create other metrics: centrelines (e.g., valley centreline,
channel centreline, meander belt centreline, etc.), buffers
of a given width on both sides of polygons or centrelines,
or elementary segments with a systematic length that is
proportional to the catchment size (see details in Alber
and Piegay, 2011).
11.3.3 Exampleoflongitudinalgenericparameters
treatment using unorthorectified photos
The use of historical records usually requires the rectifica-
tion of images, a long and exhausting process that limits
such effort both in time (to a limited number of dates)
and space (to a limited reach). As a consequence, sev-
eral approaches have focused on extracting information
directly from the images without georeferencing them
with great precision. It can be worthwhile when studying
channel width, depth, sinuosity or planform character-
istics). Lateral channel activity plays a prominent role
in many of the classification systems described above.
Most methods for quantifying lateral channel activity are
sensitive to spatial error in the remotely sensed imagery
because rectification error will lead to apparent shift even
when the channel is laterally stable. These errors can
be addressed to some extent by considering the center-
line elongation that generally occurs as rivers migrate.
On single-thread meandering channels, lateral migration
causes the streamwise coordinate s (measured relative to
a fixed starting coordinate) to elongate according to the
equation (Seminara et al., 2001):
s
ds
dt =−
CVds
(11.1)
0
where C is the local centreline curvature d θ /ds, θ repre-
sents the angle relative to an arbitrary directional datum
of a line tangent to the centreline, and V is the local lateral
shifting rate orthogonal to the centreline (Figure 11.4A).
Note that V is negative when the channel is shifting to the
right with respect to the positive streamwise direction.
Aalto et al. (2008) showed that ds/dt was easily mea-
sured from two aerial photographs by comparing s ( t ) with
s(t
t) at corresponding points along a river centerline,
 
Search WWH ::




Custom Search