38
small vort�ces and subject to periodic disruption and
uplift. Thi11 passes up into an outer layer containing
large breakup vortices that rise steeply to the surface,
where they produce boils. These large vortices develop mainly over the lee of dunes, where the presSure gradient encourages bursting. It has been found
that the bursts occur, on average, every five water
depths in the flow direction. As Leeder (1983, p. 7)
stated, "the feedback effect between flow and bedform is clearly of great importance.''
Leeder (1983) summarized progress in the application of these data on turbulence to the theory of
bedform generation. Ripples are controlled by flowseparation dynamics, and are not influenced by the
outer layer. They develop on hydraulically smooth
boundaries by the action of flow separation and
reattachment. Their form is insensitive to water
depth, and only loosely related to bed shear stress.
With increased shear stress turbulent vortices become larger but more varied in scale and speed,
leading to the generation of a wider range of heights,
wavelengths, and forward ripple speeds. Leeder
(1983) suggested that the larger ripples cannibalize
the smaller ones, leading to enhanced scour in the
ripple lee, and still further increases in ripple height.
Dunes, the characteristic bedform in the highervelocity ranges of subcritical flow, show strong correlation of height and wavelength with water depth
(Allen 1968; jackson 1976a), suggesting that their
form is controlled by the outer turbulent layer. It
seems likely that bursts develop from large ripples as
the ripples grow in size, and that these then domi�
nate the structure of the flow and become the principal mechanism in bedform production (Leeder
1983). Sediment that is lifted into suspension by the
turbulent bursts travel downstream a maximum of
about five flow depths before deposition takes place.
Continued bursting, erosion and deposition will
eventually cause the bedforms to adjust to a relationship between bedform wavelength, height, and flow
depth. A stable dune bed will gradually develop in
which macroseparation as well as burst macroturbulence play an important role in determining dune
morphology and magnitude (Leeder 1983).
The transition to upper-stage plane beds may
take place because the high grain concentration that
develops at high flow speeds blankets and inhibits
turbulence development. Also, the frequency of turbulent sweeps may overcome the scouring effect of
leeside separation eddies. Saunderson and Lockett
(1983) reported a series of flume experiments carried out to examine this transition. They found that a
rhixture of flat-bed and dune�like conditions ocHistorical Background
curred. The decreasing scale and erosive power of
separation eddies at higher flow speeds leads to
smaller scour poc,kets and the draping of sand over
rounded dune cr�sts, rather than the production of
angle-of-repose foreset lamination. The result is the
generation of <(humpback" dunes with low-angle,
sigmoidal cross-bedding.
Considerable problems of classification and
nomenclature were apparent by the late 1970s, as
discussed by Miall (1977) and N.D. Smith (1978).
Leeder (!983) offered a tentative genetic classification of bedforms according to the dominant physical
process controlling their size limits and stability
(Fig. 2.23). However, the problem of nomenclature
was not satisfactorily resolved until an SEPM Research Symposium was held in 1987, the results of
which were collated by Ashley (1990). This paper is
discussed in a later section. Some other recent research results have been reviewed by Fielding
(1993b).
2.4.2 The Decline and Fall of the Vertical Profile
The first fluvial symposium in 1977 was held at a
time when a great deal of new information on fluvial
facies was being published. At this time the concept
of the simplified facies model was much in vogue
(Walker 1979). Such models were commonly presented in simplified, abbreviated form, in terms of a
vertical stratigraphic profile, which sets out a typical
sequence of lithologies with their accompanying
biofacies and sedimentary structures (e.g., Visher
1965a,b ). This practice developed from the application of Walther's law, which states that only those
facies that can be found forming side by side in
nature can occur in contact with one another in
vertical succession, unless the succession contains
internal erosion surfaces (Middleton 1973). The vertical profiles are cyclic, in the sense that they are
repeated several or many times in succession, generally with some internal variation. The model may
also contain information about lateral variability,
particularly with reference to the position of a shoreline or sediment source area and the direction of
prevailing air or water currents. Most models were
based on a combination of observations from modern environments and the ancient rock record.
In subsurface analysis, application of Walther's
law enables us to interpret lateral facies relationships
from their vertical sequence as expressed in a facies
model. Several textbooks appeared during this period which provided catalogues of vertical proflles as
small vort�ces and subject to periodic disruption and
uplift. Thi11 passes up into an outer layer containing
large breakup vortices that rise steeply to the surface,
where they produce boils. These large vortices develop mainly over the lee of dunes, where the presSure gradient encourages bursting. It has been found
that the bursts occur, on average, every five water
depths in the flow direction. As Leeder (1983, p. 7)
stated, "the feedback effect between flow and bedform is clearly of great importance.''
Leeder (1983) summarized progress in the application of these data on turbulence to the theory of
bedform generation. Ripples are controlled by flowseparation dynamics, and are not influenced by the
outer layer. They develop on hydraulically smooth
boundaries by the action of flow separation and
reattachment. Their form is insensitive to water
depth, and only loosely related to bed shear stress.
With increased shear stress turbulent vortices become larger but more varied in scale and speed,
leading to the generation of a wider range of heights,
wavelengths, and forward ripple speeds. Leeder
(1983) suggested that the larger ripples cannibalize
the smaller ones, leading to enhanced scour in the
ripple lee, and still further increases in ripple height.
Dunes, the characteristic bedform in the highervelocity ranges of subcritical flow, show strong correlation of height and wavelength with water depth
(Allen 1968; jackson 1976a), suggesting that their
form is controlled by the outer turbulent layer. It
seems likely that bursts develop from large ripples as
the ripples grow in size, and that these then domi�
nate the structure of the flow and become the principal mechanism in bedform production (Leeder
1983). Sediment that is lifted into suspension by the
turbulent bursts travel downstream a maximum of
about five flow depths before deposition takes place.
Continued bursting, erosion and deposition will
eventually cause the bedforms to adjust to a relationship between bedform wavelength, height, and flow
depth. A stable dune bed will gradually develop in
which macroseparation as well as burst macroturbulence play an important role in determining dune
morphology and magnitude (Leeder 1983).
The transition to upper-stage plane beds may
take place because the high grain concentration that
develops at high flow speeds blankets and inhibits
turbulence development. Also, the frequency of turbulent sweeps may overcome the scouring effect of
leeside separation eddies. Saunderson and Lockett
(1983) reported a series of flume experiments carried out to examine this transition. They found that a
rhixture of flat-bed and dune�like conditions ocHistorical Background
curred. The decreasing scale and erosive power of
separation eddies at higher flow speeds leads to
smaller scour poc,kets and the draping of sand over
rounded dune cr�sts, rather than the production of
angle-of-repose foreset lamination. The result is the
generation of <(humpback" dunes with low-angle,
sigmoidal cross-bedding.
Considerable problems of classification and
nomenclature were apparent by the late 1970s, as
discussed by Miall (1977) and N.D. Smith (1978).
Leeder (!983) offered a tentative genetic classification of bedforms according to the dominant physical
process controlling their size limits and stability
(Fig. 2.23). However, the problem of nomenclature
was not satisfactorily resolved until an SEPM Research Symposium was held in 1987, the results of
which were collated by Ashley (1990). This paper is
discussed in a later section. Some other recent research results have been reviewed by Fielding
(1993b).
2.4.2 The Decline and Fall of the Vertical Profile
The first fluvial symposium in 1977 was held at a
time when a great deal of new information on fluvial
facies was being published. At this time the concept
of the simplified facies model was much in vogue
(Walker 1979). Such models were commonly presented in simplified, abbreviated form, in terms of a
vertical stratigraphic profile, which sets out a typical
sequence of lithologies with their accompanying
biofacies and sedimentary structures (e.g., Visher
1965a,b ). This practice developed from the application of Walther's law, which states that only those
facies that can be found forming side by side in
nature can occur in contact with one another in
vertical succession, unless the succession contains
internal erosion surfaces (Middleton 1973). The vertical profiles are cyclic, in the sense that they are
repeated several or many times in succession, generally with some internal variation. The model may
also contain information about lateral variability,
particularly with reference to the position of a shoreline or sediment source area and the direction of
prevailing air or water currents. Most models were
based on a combination of observations from modern environments and the ancient rock record.
In subsurface analysis, application of Walther's
law enables us to interpret lateral facies relationships
from their vertical sequence as expressed in a facies
model. Several textbooks appeared during this period which provided catalogues of vertical proflles as