Growth of Present-Day Concepts, up to 1977
"If a stream has a large load of coarse rock waste, its graded
flood plain must be relatively steep (a descent offrom 5 to
20 ft or more a mile). In this case the stream does not turn
far aside from a direct course along the flood plain; but is
constantly embarassed by the formation of bars and islands of gravel and sand, splitting its current into many
shifting channels."
Chamberlin and Salisbury (1909) stated:
"In general, anything which greatly increases the load of a
stream near its head is likely to cause deposition, and so
the development of a flood plain, at some point farther
down the valley. Streams which are actively aggrading
their valleys are likely to anastomose. This results from the
fll ling of the channels until they are too small to
accomodate all the water. The latter then break out of the
channel at few or many points. The new channels thus
established suffer the same fate."
Nowadays it is recommended that the term anastomosing be restricted to rivers characterized by a
network of stable channels of low to high sinuosity,
and that the term braided should be used for the type
of river described by Davis (1898b) and Chamberlin
and Salisbury (1909), which show unstable, lowsinuosity channel patterns, a wide, shallow cross
section, and a large bedload (Mialll977).
The fourth approach to river classification is the
one of most interest to fluvial sedimentologists. Its
development overlaps with that of the morphological classification. The link between channel cross
section and sediment type was first suggested by
Griffith (1927), who observed that natural streams
with heavy bedload tend to fl ow in broad, shallow
channels, and concluded that a river must tend to
adopt the cross-section shape which gives it the
maximum sediment-carrying capacity. Actual classification began with the work of Melton ( 1936). This
was an important geomorphological contribution,
in which aerial photographs were first used extensively to illustrate fluvial landforms, but it was
largely ignored by geologists (with the exception of
Fisk). The reason may be that, although the paper
contains numerous useful observations regarding
channels, bars, and floodplains, the classification
proposed by Melton is complex and uses terminoM
logy that is difficult to interpret. The word «plain" is
used to refer to the bar-covered channel (as in braidplain, meander-plain) and, in a different context, its
use implies the entire width of the alluvial valley.
Melton's primary subdivision is into "single-crest"
and "double-crest» streams. Double-crest streams
are those which "possess two phases of geological
activity. For example, many streams build meanders
during high water, and also deposit an alluvial cover
15
on the flood-plain surface during extreme stages".
Single-crest streams are those which lack this second
depositional stage. Melton distinguished between
vertical-accretion and lateral-accretion deposits; he
discussed the difference between braided and meandering rivers and described some of the characteristics of intermediate morphological types; but all this
information was assigned a secondary importance in
his classifi cation scheme.
An important advance came with the extensive
series of flume experiments carried out by Friedkin
(1945), who investigated the effects of altering discharge, bed materials, and slope on meander patterns and the stability of bars. Friedkin also
produced braided channel patterns by increasing the
load and slope, and concluded that an important
contributary factor in the development of braided
patterns was easily erodible banks. This work was
much quoted by Fisk (1947) in his description of the
Mississippi alluvial deposits.
Leopold and Wolman (1957, 1960) used both field
and flwne experiments to study bar and channel
behavior. They demonstrated that channel patterns
form a continuum resulting from variations in discharge, load, and slope, and showed how channel
patterns change downstream in response to variations in these controls.
Schumm (1963) divided rivers into three types:
bedload-, suspended-load-, and mixed-loadstreams, and showed how load characteristics are
related to sinuosity and channel cross-section shape,
and to the tendencies toward vertical or lateral erosion and sedimentation within the channel. Several
workers, notably Galloway (1981), later used this
classification as a basis for a basic subdivision of
fluvial styles (Sect. 2.4.4).
Galay et a!. (1973) and Mollard (!973) carried
. these studies further, showing, in particular, the
wide range of variability that can exist in channel
plan and cross-section shapes and the sediment
characteristics associated with them. Research is
needed to relate these features in more detail to the
sedimentary record, there being a particular need for
quantitative studies. Rust (1978a) developed a new
braiding index and proposed coupling this with a
measure of sinuosity in an inclusive channel classification.
The implications of much of this recent work for
studies . of ancient fluvial sediments have yet to be
explored in detail. (See Sect. 2.3.6 for additional
discussion of this area, where its relevance for facies
analysis is explored.)
"If a stream has a large load of coarse rock waste, its graded
flood plain must be relatively steep (a descent offrom 5 to
20 ft or more a mile). In this case the stream does not turn
far aside from a direct course along the flood plain; but is
constantly embarassed by the formation of bars and islands of gravel and sand, splitting its current into many
shifting channels."
Chamberlin and Salisbury (1909) stated:
"In general, anything which greatly increases the load of a
stream near its head is likely to cause deposition, and so
the development of a flood plain, at some point farther
down the valley. Streams which are actively aggrading
their valleys are likely to anastomose. This results from the
fll ling of the channels until they are too small to
accomodate all the water. The latter then break out of the
channel at few or many points. The new channels thus
established suffer the same fate."
Nowadays it is recommended that the term anastomosing be restricted to rivers characterized by a
network of stable channels of low to high sinuosity,
and that the term braided should be used for the type
of river described by Davis (1898b) and Chamberlin
and Salisbury (1909), which show unstable, lowsinuosity channel patterns, a wide, shallow cross
section, and a large bedload (Mialll977).
The fourth approach to river classification is the
one of most interest to fluvial sedimentologists. Its
development overlaps with that of the morphological classification. The link between channel cross
section and sediment type was first suggested by
Griffith (1927), who observed that natural streams
with heavy bedload tend to fl ow in broad, shallow
channels, and concluded that a river must tend to
adopt the cross-section shape which gives it the
maximum sediment-carrying capacity. Actual classification began with the work of Melton ( 1936). This
was an important geomorphological contribution,
in which aerial photographs were first used extensively to illustrate fluvial landforms, but it was
largely ignored by geologists (with the exception of
Fisk). The reason may be that, although the paper
contains numerous useful observations regarding
channels, bars, and floodplains, the classification
proposed by Melton is complex and uses terminoM
logy that is difficult to interpret. The word «plain" is
used to refer to the bar-covered channel (as in braidplain, meander-plain) and, in a different context, its
use implies the entire width of the alluvial valley.
Melton's primary subdivision is into "single-crest"
and "double-crest» streams. Double-crest streams
are those which "possess two phases of geological
activity. For example, many streams build meanders
during high water, and also deposit an alluvial cover
15
on the flood-plain surface during extreme stages".
Single-crest streams are those which lack this second
depositional stage. Melton distinguished between
vertical-accretion and lateral-accretion deposits; he
discussed the difference between braided and meandering rivers and described some of the characteristics of intermediate morphological types; but all this
information was assigned a secondary importance in
his classifi cation scheme.
An important advance came with the extensive
series of flume experiments carried out by Friedkin
(1945), who investigated the effects of altering discharge, bed materials, and slope on meander patterns and the stability of bars. Friedkin also
produced braided channel patterns by increasing the
load and slope, and concluded that an important
contributary factor in the development of braided
patterns was easily erodible banks. This work was
much quoted by Fisk (1947) in his description of the
Mississippi alluvial deposits.
Leopold and Wolman (1957, 1960) used both field
and flwne experiments to study bar and channel
behavior. They demonstrated that channel patterns
form a continuum resulting from variations in discharge, load, and slope, and showed how channel
patterns change downstream in response to variations in these controls.
Schumm (1963) divided rivers into three types:
bedload-, suspended-load-, and mixed-loadstreams, and showed how load characteristics are
related to sinuosity and channel cross-section shape,
and to the tendencies toward vertical or lateral erosion and sedimentation within the channel. Several
workers, notably Galloway (1981), later used this
classification as a basis for a basic subdivision of
fluvial styles (Sect. 2.4.4).
Galay et a!. (1973) and Mollard (!973) carried
. these studies further, showing, in particular, the
wide range of variability that can exist in channel
plan and cross-section shapes and the sediment
characteristics associated with them. Research is
needed to relate these features in more detail to the
sedimentary record, there being a particular need for
quantitative studies. Rust (1978a) developed a new
braiding index and proposed coupling this with a
measure of sinuosity in an inclusive channel classification.
The implications of much of this recent work for
studies . of ancient fluvial sediments have yet to be
explored in detail. (See Sect. 2.3.6 for additional
discussion of this area, where its relevance for facies
analysis is explored.)