202
difficult to recognize accretionary geometries in
large, wide rivers (Leeder 1973). As the discussion in
the subsequent sections should make clear) recent
work has resulted in numerous documented examples of macroforms in a wide range of fluvial
settings. Nevertheless, the concerns quoted above
are real. The best architectural methods may not be
able to identify accretionary geometries where expo�
sures are inadequate, for example) where an outcrop
is oriented parallel to the strike of gently dipping
accretion surfaces. It remains very difficult to document macro forms in the subsurface.
It is important to note that almost all the developments that have taken place in the d ocumentation of
the two- and three-dimensional architecture of fluvial systems have relied on studies of the ancient
record, and several of the fl uvial styles described in
this section are basfd almost entirely on studies of
the ancient. There is a pressing need for research
into the internal architecture of modern river deposits, preferably using high-resolution seismic or
ground-penetrating radar methods, such as are described in Sect. 9.5. Only work of this type can provide the undisputed link between surface processes
and the preserved deposits that is necessary for the
reliable definition of process-response models.
What is reported here is therefore work in progress
toward that goal.
One important test of the architectural method in
a modern river was carried out by Brierley and
Hickin (1991). They documented the facies sequences and architectural elements in bank exposures of a modern gravel-sand river in British
Columbia that undergoes a downstream style change
fr om braided to wandering to meandering. Facies
successions and architectural cross sections were
similar for all three fluvial styles, and the authors
concluded that these kinds of data would not be able
to provide predictions of fluvial style in other comparable fluvial settings. It is possible to question
their results because the river is undergoing rapid
evolution, and the surface channel style may not be
an accurate representation of the fluvial style at the
time the sampled channel and floodplain assemblages were deposited. Nevertheless, their conclusion might seem disturbing. In fact, their results bear
out the point this writer has been trying to make, that
it is only by carrying out a detailed architectural
documentation that the researcher can hope to arrive at precise and accurate interpretations. In the
case ofBrierley and Hickin's (1991) study, their work
did not include paleocurrent analysis, which would
provide vital data for the interpretation of channel
Fluvial Styles and Facies Models
sinuosity and the range of channel and bar orientations - two of the key differences between meandering, wandering, and braided gravel-bed rivers. As
noted elsewhere in this book, fluvial style should not
be interpreted from single exposures of one channelfill succession, the method tested by Brierley and
Hickin (1991), but should be based on three-dimensional data from several successions. Such a data
base would provide a more quantitative measure of
the relative importance of the various facies and the
types of channel and bar present in the river.
Several example of intermediate fluvial styles are
referred to in the succeeding sections. The «wandering gravel-bed" river and the «low-sinuosity
braided-meandering river with alternate bars" are
by definition intermediate fluvial styles. Reference is
also made to transitions betw'een the "high-energy
sand-bed braided" and the "flashy, ephemeral
sheetflood" styles, and between the latter and the
"sheetflood distal braided style". Differences between some of these fluvial styles are small, and it
could be argued that no purpose is served by proposing so many "models". The point in doing so is
simply to recognize the real-life variability of natural
systems and to encourage flexibility in their interpretation. Many actual examples of fluvial successions will display characteristics of more than one of
the models described here. This is not <
<
channel dimensions, flow characteristics, etc. In this
way, each river and each fluvial deposit are treated as
an individual, possibly unique entity. End-member
force-fitting, which is what tends to occur when few
models are available, is therefore avoided.
An excellent example of an ancient fluvial system
that does not fit any of the "standard" models (including those described below), and yet contains
many fe atures tantalizingly similar to several of
them, is the Escanilla Formation (Eocene) of the
Spanish Pyrenees. A regional architectural reconN
struction is provided by Bentham et al. (1993), and
additional details of the gravel and sand bodies are
described by Dreyer et a!. (1993). This unit consists
of poorly interconnected sheet sandstones and conglomerates, separated by fine-grained sediments
constituting more than 40% by volume of the total
difficult to recognize accretionary geometries in
large, wide rivers (Leeder 1973). As the discussion in
the subsequent sections should make clear) recent
work has resulted in numerous documented examples of macroforms in a wide range of fluvial
settings. Nevertheless, the concerns quoted above
are real. The best architectural methods may not be
able to identify accretionary geometries where expo�
sures are inadequate, for example) where an outcrop
is oriented parallel to the strike of gently dipping
accretion surfaces. It remains very difficult to document macro forms in the subsurface.
It is important to note that almost all the developments that have taken place in the d ocumentation of
the two- and three-dimensional architecture of fluvial systems have relied on studies of the ancient
record, and several of the fl uvial styles described in
this section are basfd almost entirely on studies of
the ancient. There is a pressing need for research
into the internal architecture of modern river deposits, preferably using high-resolution seismic or
ground-penetrating radar methods, such as are described in Sect. 9.5. Only work of this type can provide the undisputed link between surface processes
and the preserved deposits that is necessary for the
reliable definition of process-response models.
What is reported here is therefore work in progress
toward that goal.
One important test of the architectural method in
a modern river was carried out by Brierley and
Hickin (1991). They documented the facies sequences and architectural elements in bank exposures of a modern gravel-sand river in British
Columbia that undergoes a downstream style change
fr om braided to wandering to meandering. Facies
successions and architectural cross sections were
similar for all three fluvial styles, and the authors
concluded that these kinds of data would not be able
to provide predictions of fluvial style in other comparable fluvial settings. It is possible to question
their results because the river is undergoing rapid
evolution, and the surface channel style may not be
an accurate representation of the fluvial style at the
time the sampled channel and floodplain assemblages were deposited. Nevertheless, their conclusion might seem disturbing. In fact, their results bear
out the point this writer has been trying to make, that
it is only by carrying out a detailed architectural
documentation that the researcher can hope to arrive at precise and accurate interpretations. In the
case ofBrierley and Hickin's (1991) study, their work
did not include paleocurrent analysis, which would
provide vital data for the interpretation of channel
Fluvial Styles and Facies Models
sinuosity and the range of channel and bar orientations - two of the key differences between meandering, wandering, and braided gravel-bed rivers. As
noted elsewhere in this book, fluvial style should not
be interpreted from single exposures of one channelfill succession, the method tested by Brierley and
Hickin (1991), but should be based on three-dimensional data from several successions. Such a data
base would provide a more quantitative measure of
the relative importance of the various facies and the
types of channel and bar present in the river.
Several example of intermediate fluvial styles are
referred to in the succeeding sections. The «wandering gravel-bed" river and the «low-sinuosity
braided-meandering river with alternate bars" are
by definition intermediate fluvial styles. Reference is
also made to transitions betw'een the "high-energy
sand-bed braided" and the "flashy, ephemeral
sheetflood" styles, and between the latter and the
"sheetflood distal braided style". Differences between some of these fluvial styles are small, and it
could be argued that no purpose is served by proposing so many "models". The point in doing so is
simply to recognize the real-life variability of natural
systems and to encourage flexibility in their interpretation. Many actual examples of fluvial successions will display characteristics of more than one of
the models described here. This is not <
<
channel dimensions, flow characteristics, etc. In this
way, each river and each fluvial deposit are treated as
an individual, possibly unique entity. End-member
force-fitting, which is what tends to occur when few
models are available, is therefore avoided.
An excellent example of an ancient fluvial system
that does not fit any of the "standard" models (including those described below), and yet contains
many fe atures tantalizingly similar to several of
them, is the Escanilla Formation (Eocene) of the
Spanish Pyrenees. A regional architectural reconN
struction is provided by Bentham et al. (1993), and
additional details of the gravel and sand bodies are
described by Dreyer et a!. (1993). This unit consists
of poorly interconnected sheet sandstones and conglomerates, separated by fine-grained sediments
constituting more than 40% by volume of the total