Middleton G.V. (Ed.) Encyclopedia of Sediments and Sedimentary Rocks
Подождите немного. Документ загружается.
SURFACT FORMS
70.5
10
• Upper stage plane bed (UP)
» Transition (T)
• Dunes (D)
• Ripples (R)
- Lowef stage plane bed (LPt
d 6 8
Particle parameter, D#
39
7B
15.6
313
62 5
12b 250 500
1000
2000
4OO0
Figure S56 Graph showing existence fields of surface forms gencraled by unidirectional water flows in shallow channels, f^afa arc from
various sources. The mobility number i)' is a nonrlimensifinal measure Ihat expresses the shear stress exerted on the particles by the flowing
water, and [.y is a nondimension.il measure of the median ptirtit le ditimeter. The l{)wef vertical axis gives tbe equivalent value of aclual
mcflian particle diameter at a reference water temperature of 20 C. iFrom V^in den Berg jnd Van Celder, I99i.l
Current ripples
Over a wide range of sediment sizes, from the finest
noneohesive silts up to abottt 0.7tnm. and for Ilow sttengths
ranging upward from the threshold for seditnent movement,
the equilibrium surfaee Ibrni is eurrent ripples. For sediment
sizes less than about 0.2 mm, ripples give way to a planar
transport surfaee with inereasing flow strength: for coarser
sizes,
ripples pass abruptly into dunes with inereasing fiow
strencth.
Current ripples are rtigged snutll-seale bed forms with crests
and troughs oriented dominantly transverse to flow. Their
downcurreni surfaces have slopes usually at or near the angle
of repose ol" the seditnent (of the order of 30 ). although
tending to be less at high tlow strengths, and their upeurrent
slopes are usually mueh gentler. Spacings are usually 0.1-
0.2 m, and heights are usually no more than a few centimeters.
Height and spacing are nearly independent of flow strength,
tlow depth, itnd sediment si/e.
706
,SL)RFArF FORMS
m/s)
o
locity
L
S
i
o
E
1
equl
6
o
2.0
1.0
IIII
0.6
0.4
0.2
-
-
1
1
%
(j_
°
a/
-*
« >
% * t
^^^.^--^^^
Mil
•:
1
1
1
II 1 , _
-f
.25-.40
m
I 1
0.1
1.0
10'C-equivalent sediment size Dio
(mm)
Figure S57 Graph showing the existence fields
of
surface forms
generated
by
unidirectional water flows
in
shallow channels, with flow
depths ranging frotn 0.2.S ni
to
0.40
m.
Data are frotn various sources.
On
the
vertical axis,
t^nt is
a nondimensional measure
of
the mean
flow velocity, expressed as the equivalent value
of
the actu.il mean
flow velocity
at
a reference temperature
of
20
C. On ihe
horizontal
axi5,
D\Q
is
the corresponding measure
of
the median particle
diameter.
A,
antidunes; UP. upper-stage plane bed;
D,
dunes;
R, ripples; LP, lower-stage plane bed; NM,
no
movement. (From
Southard and Boguchwal, lyw.i
When cttrrcnt ripples tirsi tnake their appcaratice
oti a
planar scditnent surface, they are strotigly regular atid straight-
crested,
but as
they grow toward equilibrium they develop
substantial irregularity
in
geometry: crests
and
troughs tend
to
be sinuous
and
variable
in
height,
and
when traced laterally
individual ripples crests
end by
merging with another ripple
crest
of
disappearing into
a
trough.
Ripples tnovc
in the
downcurrent direction
itt
speeds orders
of magnitude less than
the
Ilow velocity
by
means
of
erosion
of
seditnent frotn
the
tipstream surface
and
deposition
on the
downstream surface. lndi\idual ripples exist
tor
titnc periods
corresponding
to
downcurrent movement typically equal
to
less than
ten
ripple spacings before merging with another
ripple. Meanwhile,
new
ripples
are
formed
by
division
of a
preexisting single ripple.
Subaqueous dunes
Over
a
range
of
sediment sizes frotn about 0.2 mm upward into
gravels,
and
over
a
wide range
of
moderate
to
high flow
strengths (Figures S56, S57) subaqueous dunes
are the
stable
surface form under unidirectional water flows.
In
sediments
finer than about 0.7 mm. dunes succeed ripples with increasing
flow strength;
in
coarser sediments, dunes make their
appearance
at
tlow strengths
not
tnuch greater than
ihe
threshold
for
sediment movement. With increasing tlow
strength, dunes
are
washed ottt
to a
planar transport surface.
In
the
past, surface fortns here called dtines have variotisly
been termed megaripples
and
sand waves
as
well
as
dunes.
In
recent years (see Ashley, 1990)
it has
become widely accepted
to describe
all
such large-scale rugged flow-transverse forms
as
dunes.
Some investigators, however, consider large-scale sand
waves
in
tidal cnvirontnents
to be
fundamentally different
frotn dunes
in
unidirectional flows (Allen. 1980).
Dunes are much like ripples
in
geometry
and
movement.
At
relatively
low
flow strengths, dunes tend
to be
regular
and
straight crested (often called "two-dimensional dunes"),
whereas
at
relatively high flow strengths they tend
to be
sinuous
and
discontinuous (often called "ihrce-dimcnsional
dunes""),
with strong variability
of
crest
and
trough heights.
For three-dimensional dunes,
the
existence
of
deep scour pits
along troughs
is
especially characteristic.
The spacing
of
dunes ranges upward from somewhat less
than half
a
meter
to
thousands
of
meters,
and the
height
of
dunes ranges upward from
a
several centimeters
to
many tens
oi tneters. The controls
on
dune size are comple,\,
but it
seetiis
clear that there
is no
fundamental litnil
to
dune size,
il" the
combination
of
tlow strength
and
seditnent
si/c are
appro-
priate
and
flow depth
is
sufficiently great. There
is a
well-
defined break
in
size between current ripples
and
dunes.
Antidunes
Over
a
wide range
of
noneohesive sediment sizes,
and
io\' high
flow strengths
and
relatively shallow depths,
the
equilibrium
surface fortn
is
antidunes. Antidunes
are
trains
of
undulatory
surface forms that develop on
a
planar transport surface.
For a
given flow depth, antidunes always develop
if
the flow strength
is increased sufficiently, although
the
common limitations
to
possible flow strengths
in
nature mean that antidunes
are
characteristic only
oi
rather shallow tiows. except perhaps
under condittons
of
catastrophic flood flows. Antidunes tend
to move slowly upstreatn, growing
in
height
and
theretbre
in
steepness
as
they move. Eventtially
the
entire train
of
atttiduiics becomes tmstable: the antidunes then break, causing
tnticli suspension ot" sediment,
and the bed
surface reverts
to
beuig nearly planar, whereupon
the
antidunes slowly reform.
Mud waves
In many regions
in the
abyssal ocean,
at the
foot
of die
continental slope,
the
surt"aees
of
large bodies
of
fine sediment
known
as
sediment drifts
are
covered with large-scale,
low-
amplitude wavelike features:
for a
well-documented recent
example,
see
Flood
and
Shor (1988). These features have
variously been called sediment waves,
mud
waves,
or
abyssal
antidunes
(the
last because they seem
to
have shifted
upeurrent:
see
below). Sitnilar
but
smaller waves have been
observed
on the
flanks
of
channels
on
submarine fans (e.g.,
Normark
etaf,
1980: McHugh atid Ryan, 2000).
The seditnent
of
which the waves consist seems invariably
to
be fine
mud. The
spacing
of
these waves ranges from
one
kilometer
to as
much
as
twenty kilotneters,
and
heights
are in
the range
of
five meters
to one
hundred meters. Slopes
of the
flanks
of
these waves
are
seldom greater than
one
degree.
Despite their gentle slopes,
the
waves are strikingly revealed
by
subbottom echo-sounding techniques.
In
some cases these
waves
are
symtnetrical.
but
typically they have shit'ted slowly
in the direction that
is
known
or
assutned
to be
upeurrent.
The
subbottom record shov\s cIcarK ihat
in
tnost cases there
has
been deposition
on
both flanks
oi"
the wave,
but at a
lower rate
on
the
flank toward which
the
wave
has
shifted.
The
internal
stratilication
of
waves shows clearly that they existed, and were
the site
of
active continuous
or
semicontinuous deposition,
for
geologically long titncs.
In
some cases they
are
seen
to be
draped with dit"t"ercnt sediment
and
therefore inactive:
in
other
cases they seem likely
(o be
active still.
SURrACE
707
The dynamics of the mud waves are not tully imderstood.
The waves seem to have been formed by slow deposition oftine
sediment from bottom currents with moderate velocities and
presumably low sediment concentrations. In the case of the
waves on scdimenl drifts, the depositing currents seem to be
bottom-hugging contour currents cai rying low concentrations
ofsiispended sediment, which however are substantially greater
than the mueh cleaner water higher in ihe water column. In ihe
caseof the waves on submarine fans, the evidence points to mild
turbidity flows that result from overbank spillage of channe-
lized episodic turbidity currents from ehannels onto broad
levees. The similarity in features between these two different
environments suggests a common origin. The most widely
accepted theory (Normark
etaf,
1980: Flood. 19S8) is that
these waves are a manifesto tion ofa pattern of nearly stationary
inierna! waves generated in the near-botlom interval of vertical
density gradient that is occasioned by the upward decrease in
suspended-sediment concentration.
Flow-parallel surface forms
Fuvr(i\ys are a kind of flow-parallel surfaee form that is
produced by localized erosion of cohesive sediments subjected
to moderately strong unidirectional or bidirectional currents in
a generally depositional regime. They have been reported from
(he abyssal ocean (e.g., Hollister etal.. 1974), estuaries (e.g..
Flood, 1981). and lakes (e.g.. Vieknian ciai. 1992).
Furrows range in depth from a meter to more than ten
meters, and in width from a meter to as much as one hundred
meters. In cross section they often appear lo be roughly
trapezoidal, with steep and often rippled side slopes and a
nearly (lat lloor. Spacing of ihe furrows ranges from more than
ten times width to even less than width. The sediment surfaee
between furrows is generally flat. Individual furrows can
extend for distances greater than one kilometer, and perhaps
much farther, given the difficulties of continuous observation
ovei- broad areas in inaccessible environments. Furrows
commonly bifurcate in tuning-fork-like junctions, which open
toward the oncoming eurrent. Measured current strengths
in furrowed areas range frotn five centimeters per second in
abyssal examples to more than half a meter per second in
shallou conlinental shelves and large lakes. In at least some
cases there is good evidence of continuing deposition in the
areas between the furrows, and this has to be reeoncited with
the obviously erosional nature of the furrows themselves.
It is generally believed that the furrows are a manifestation
of the existence oi patterns oi helieoidal streamwise secondary
circulaiion su peri in posed upon the mean flow. The furrows
become localized along streamwise zones of flow convergence,
where low-density aggregates colled in windrows and abrade
the underlying cohesive sediment as they are moved down-
current. The sediment put in suspension then tends to be
carried upward in the ascending branch of the circulation cell.
In this view, the furrows can he maintained by oeeasional
strong erosive flows on a background of overall litie-sediment
deposition during long periods oi weaker flow. See Flood
(1983) for an exposition of this theory.
Erosional forms: flutes
When an erosive current passes over a bed of cohesive
sediment, the sediment surface tends to take on a charaeteristic
fluted geometry. The individual llutes vary in their shape, but
their distinctive characteristic is an approximately bilateral
symmetry about a vertical plane parallel to the flow but a
strong asymmetry when viewed normal to that plane, whereby
the depth of the scour and the steepness of the surface are
much greater at the upstream end than at the downstream end.
The adjective "heel-shaped" (Allen. 1984) is an apt descriptor.
The typical flute, when viewed from above, has an upeurrent
end with approximately parabolic shape, although in many
cases the upstream end of the flute is more pointed or has an
asymmetrical corkscrew shape. The geometrical pattern of the
flutes shows considerable variety, as does the density of flutes
on the sediment surfaee. Most flutes are on streamwise scales
of the order often centimeters, but they range upward in size
to at least a few meters, and there have been reports of giant
flute-shaped scours that seem to be the result of especially
powerful turbidity currents {e.g,, Shor etai. 1990). Much of
our understanding of the nature and origin of flutes has come
about through the experimental work of J.R.L, Allen (for a
summary, see Allen. 1984).
Flutes tend to be formed by powerful currents, like turbidity
currents or river floods, which form the flutes during an initial
erosive phase and then bury them during a later depositional
phase. Much oi what is known about flutes is derived from
observations of the bases of sandstone beds less resistant
overlying shales in the ancient sedimentary reeord; the
sandstones record the casts {i.e.. the negative imprints) of the
flutes. There has been much sueeess in producing realistic llutes
experimentally: see Allen (1984) for a summary of such work
(and see
Seout:
Scour Marks). It is clear that flutes are the result
of nei erosive removal of material from the sediment surface
without further interaetion with the eroded material with the
seditnent bed downstream; the eroded material goes directly
into suspension rather than being transported as bed load. The
close similarity between flutes formed on cohesive mud beds
and those formed by dissolution on soluble surfaees of rock or
ice by flowing water supports that notion.
Subaqueous surface forms generated by
oscillatory and combined flow
Oscillation ripples
Noneohesive sediments over a wide range of sediment sizes
assume surface forms in a wide range of sizes and geometries in
bidirectional purely oscillatory flow. Such forms have beeti the
subject of systematic study in both natural environments and
laboratory tanks sinee early in the twentieth century. Among
the earliest of modern laboratory studies is that of Bagnold
(1946);
the classic tield study was by Inman (1957).
Field studies of oseillatory-flow surface forms sutler from
the difflculties of observation under conditions of heavy seas,
and laboratory studies have concentrated mainly on the fairly
short oseillation periods to which small apparatus are limited.
More recent laboratory studies at mueh longer oscillation
periods in a large oscillatory-flow tunnel are reported by
Southard etal. (1990).
A useful way to deseribe the conditions of existence of
oscillatory-flow surface forms is to consider them as a function
of three leading variables: median sediment size, maximum
oscillatory velocity, and oscillation period. As with unidirec-
tional-flow surface forms in the preceding section, the ranges
of existence of the various oseillatory-flow surface forms can
708
SURFAa FORMS
L_L
1 2 4 6 8 10
10°C-equivalent oscillation period Tio (s)
20
Figure S58 Relationships of oscilkitory-flow surface forms i/i a graph
ol nondimensional maximum oscillatory flow speed U,,\o versus
nondimensional oscillation period 7",,), expressed as equivalent values
at a reference temperature of 10'^C, for particle slices of about 0.1 mtn.
Symbols for spacing of surface forms: solid diamonds, <O.U)Omm;
open circles, 0,10-0.175 mm; solid circles,
0.1
7,'j-O,,3Om; open
triangles, O,,iO-O.55 m; solid triangles, 0.5,5-1.Om; open squares, 1.0-
1,7,5
01;
solid squares, >1.75m. The dashed curves are approximate
contours of surface-form spacing. Horizontal and vertical tick marks
represent three-dimensional configurations; symbols without tick
marks represent two-dimensional configurations. (From Southard,
1991.)
be viewed in a three-dimensional graph with nondimensiona-
lized versions of ihesc three variables along the a.xcs.
Figure S58 is a veloeity period for a sediment size of about
0.1mm, with axes expressed as IOC-equivalent maximum
oscillalory velocity and oseillation period. For a wide range of
oscillatory velocities above the threshold for sediment move-
ment, ripples are the equilibrium surface form. At even higher
oscillatory velocities, the ripples are washed out to a planar
transport surface. Over most of the ripple
field,
ripple spaeing
inercases upward to the right, with increasing velocity and
period.
The spacing of the ripples seales closely with the orbital
diameter (the distance traversed by water partieles in the
course of a single o.seillation). For relatively small velocities
and periods the ripples are geometrically regular and straight-
crested,
but with increasing veloeity and pei'iod there is a
well-
defined change to less reguiar geometry. The graph for coarser
sand sizes is similar, but the transition to plane-bed transport
is at a higher veloeity. and the ripples are reguiar and straight-
crested over the entire existence
tield.
The nature of oscillation ripples formed by more cotnplex
multidirectional oscillatory Hows is poorly itnderstood. In the
case of what are called ladderback ripples, for which two
different sets of oscillation ripples are superimposed, it is
generally assumed that a second set of ripples become
superimposed on an preexisting set without, however, com-
pletely obliterating the earlier set. Geometrically complex
small-scale ripples showing mixtures of hexagonal and even
0,04 coa 0,12 D.I 6
UNtOIRECTIONAL VELOCITY
Uu C
Figure S59 Relationships among combined-flow surface forms in a
plot of nondimensional maximum oscillatory flow speed
U..n)
versus
nondimensional maximum unidirectional flow speed fuio; expressed
as equivalent values at a reference temperature of 10"C, for particle
sizes of around 0.1 mm and oscillation periods of around 9.5 second.
Solid circles, combined-flow ripples; open circles, plane-bed sediment
movemen!
(for large U,io), no sediment movement (for small U,n)).
Data arc from various sources. (From Arnotl and Southard, J99U.1
pentagonal shapes might be generated by multidirectional
oseillatory flow, inasmuch as normal-to-erest profiles are close
to being symmetrical, but observational data are not available
to confirm sueh a view. Hummocky forms with rounded crests
interspersed with swales, in a seemingly random pattern on
lateral scales of up to a meter, are known from bedding planes
in aneient shallow-marine deposits. It is not clear whether sueh
forms are generated by strong bidirectional or strong multi-
direetional oscillatory Hows, because of the difficulties of
observation in modern field and laboratory flow environments
under such conditions.
Combined-flow forms
Combined flows are common in natural flow environments,
especially shallow marine and lacustrine settings, where strong
winds assoeiated with storms cause both utiidirectional and
o.scillatory water motions. The range of
such
cotnbined flows is
enormous, and it should be expected that the surface forms
generated by those flows should aecordingly show
a
wide range
of scales and geometries.
Much less is known about the nature oi combined-flow
surface forms than about those generated by unidirectional
and oscillatory flows. There have been few recent tield studies:
that by Amos
eial.
(1988) is representative of the problems and
possibilities of making such observations in shallow marine
settings. Moreover, there have been few laboratory studies,
and even an ambitious laboratory program can hope to cover
only a narrow subset of combined-flow surface forms there
have been few laboratory studies. Among the few recent
studies. Arnott and Southard (1990) made a systematic survey
of combined-flow surface features in a large combined-flow
duct but for only a single oscillation period, a single sediment
size,
and a wide range of oscillatory flow velocities but only a
narrow range of unidirectional flow veloeities (Figure S59).
SURFACE FORMS
709
It is clear from studies of combined-flou surface forms up to
now that only a small unidirectional flow component is needed
to make the forms strongly asymmetrical. The eonsequence is
that combined-flow fortns, at least in line sands, look
supcrflciall\ like current ripples. A broad range of larger-scale
combined-flow forms generated by stronger oscillatory flows
with a subordinate unidirectional component look supcrhcially
like small unidirectional-flow dunes, although the sediment
size is much flner than can support unidirectional-flow dunes.
This may be the main reason why reports of combined-flow
surface forms preserved in the sedimentary record ha\e been so
much scarcer than their importance uould seem to imply.
Much more observation is needed before even a flrst-order
picture of the range of combined-flow forms is available.
Subaerial surface forms
When the wind blows over a surface of loose sediment at a
speed sufticicnt to move the particles, the surface is molded
into a variety of conligurations with characteristic scale and
geometry. The smallest-scale features are called ripples, and
larger-seale features are called dunes, although with much
variation in terminology. Modern studies of subaerial surfaee
forms date from the classic field and laboratory work by
Bagnold (1941). For recent summaries of the nature.
classilieation. and dynamics of subaerial surface forms, sec
Greeley and Iversen (1985), Pye and Tsoar (1990), Niekling
(1994),
and Lancaster (1994),
It is generally agreed that there is a fundamental dynamical
distinetion between subaerial ripples and dunes. This is
reflected in a well-deflned break or gap in the range of sizes
of the two kinds of features. It is less clear whether there are
natural breaks in the range of dune sizes. Wilson (1972)
reported a clear break between two classes of large-scale
subaerial surfaee forms, which he called dunes and draa, but
later workers (e.g,, Ellwood et af. 1975) have been more
inclined lo believe that there is a smooth variation in large-
.scale eolian tbims. without any natural breaks.
A complicating factor in classifying subaerial surfaee forms
and understanding their dynamics is that in most settings wind
speed and direetion are at least to some degree variable. This is
less ofa problem for ripples than (br dunes, however, because
typieal small-scale ripples respond to sand-moving winds on
timescales of minutes to hours, whereas dunes, with their far
greater bulk relative to rates of sand transport, require much
longer times to become equilibrated to a steady wind.
Another reason behind controversy over the dynamics and
elassification of subaerial surface fbrms is that nature tends to
give us biased realizations of the phenomena. Fractionation oi
particle si/e along transport paths causes particle-size distribu-
tions to evohc in partieular ways, and this has a tendency to
obscLue the picture of dynamics and classification by making it
difficult obtain an c\en view of the process and features
invoked.
Comprehensive data on equilibrium subaerial forms gener-
ated by steady winds, of the kind that have been obtained for
subaqueous forms in laboratory water ehannels and tanks,
would eonstitute a natural basis for evaluating the funda-
mental nature of subaerial forms. Only for small-scale
subaerial ripples, however, have studies been carried out in
large wind tunnels in which not only wind strength but also
particle si/e and sorting are varied over a wide and unbiased
range. Such studies are \irtually impossible for large-scale
subaerial features.
The dynamics oi subaerial surface forms and subaqueous
surfaee forms should be expeeted to be at least in part
fundamentally differcnl, beeause of the great difference in the
ratio of sediment density to fluid density in the two cases, ln
transport of mineral-density particles by water, the density
ratio is less than about three, exeept for local and uneommon
concentrations of higher-density mineral particles, whereas the
density ratio in most eolian settings is of the order of l()\ in
eonsequence. particle transport by wind is much less affected
by fluid turbulence: the sediment particles, once set into
motion, tend to follow trajectories, usually described as
ballistic, that are governed much more by particle inertia Ihan
by fluid turbulence. It should also be expeeted. however, that
this difference is much more significant for subaerial ripples
than for subaerial dunes. The scale of ripples is of the same
order as the typical excursion lengths of the transported
partieles. whereas dunes are far larger. The dynamies oi wind
ripples are in some way governed by inst:ibilities associated
with the nature and seale of particle excursions, whereas the
dynamics oi dunes is in some way related to instabilities
associated with larger-scale streamwise variations in sediment
transport rate, in ways that are likely to be broadly similar to
subaqueous dunes.
Mineral-density particles in water and in air, the two cases
of density ratio that are inipoitant on Earth, represent only
two points along the spectrum of density ratios. It seems likely
to many if not most astronomers that there are planets in other
solar systems with surface conditions represented by a far
larger range of sediment-to-fluid density ratios. Very little is
known about such cases; in particular, the question arises: do
the surfaee forms we know as current ripples and wind ripples
grade contitmously from one to the other, or is there a break in
existence in some intermediate range.' For some comments on
ripples on olher planets, see Greeley and Iversen (1985).
Subaerial ripples
Subaerial ripples are small-seale features that are oriented
transverse to the wind. They range in spacing from less than a
centimeter to as much as a few tens oi meters, and in height
from a few millimeters to as much as a meter. The most
common ripples range in spacing from a few centimeters lo
betv\een ten and fifteen centiineters. These are variously called
impact ripples or ballistic ripples, which are gcnetie terms that
reflect commonly held views on their dynamics (see below).
Some investigators refer to these simply as "normar" ripples.
They are strikingly regular in their geometry, with long,
straight crests and troughs that commonly extend for some
meters before cither ending or bifureating. Crest height and
trough depth vary little in the transverse direction. In vertical
profile the ripples are noticeably asymmetrie, with upwind
(stoss) slopes of 8 l()- and downwind (lee) slopes of 20-30 .
For an especially illuminating field study of wind ripples, see
Sharp (1963).
From a number of field and laboratory studies, summarized
by Pye and Tsoar (1990) and Lancaster (1994). it is known that
ripple size increases with both wind velocity and particle size.
There is some evidence as well thai ripple siT-e increases as the
sorting of the sediment decreases (i.e.. the spread of sizes
around the mean increases).
710
StJRFACE FORMS
Many theories have been proposed for the dynamics of wind
ripples, and there is as yet no general agreement. The classical
viewpoint, dating from the pioneering work of Bagnold (1941).
is that the ripple spaeing is governed by the saltation distanec
of the moving particles. This view has been challenged by
Sharp (1963) and more recently by .Anderson (1987),
Larger subaerial ripples in coarser sediment, well into the
pebble size range (>2mm). although less common that the
smaller ripples in sands, are well-known. These have variously
been called granule ripples or megaripples. Sediment size
distributions are usually bimodal, and the ripples tend to be
less regular in the crest pattern. There is disagreement about
whether these are dynamically distinct from (he more common
smaller ripples. Also reported in the literature are small-scale
but much less regular ripples, variously ealled fluid drag ripples
or aerodynamic ripples, which are formed by strong winds in
very fine sand or coarse silt.
When saltating sand particles move across a damp sand
surface, as on a beach at low tide during a strong offshore
wind that transports sand from coastal dunes onto the beach,
some oi the particles adhere to the surface and remain there,
held by surface tension. They arc wetted by capillarity, and in
turn capture more saltating particles. In some cases the surfaee
remains approximately planar as it aggrades: in other cases, a
pattern of irregularly transverse ridges, called adhesion ripples,
develops. Adhesion ripples commonly have spacings of the
order ofa centimeter and heights ofa fraction ofa millimeter
to a few millimeters. Less regular patterns without any
appreciable transverse orientation have been ealled adhesion
warts.
Subaeriai dunes
Subaerial dunes are mounds or ridges of sand that are shaped
by the
wind.
In a narrower definition, dunes are independent
of surrounding topography: in that respect, dunes are features
that develop in broad areas that are underlain by loose sand
and are subjected to sand-moving winds. In addition, however,
certain features that are commonly termed dunes are
aeeumulations of sand that form to the windward or leeward
sides of cliffs, hilts, or other topographic features, large and
small.
One of the most extensive studies of dunes in the broad
regions of dunes in subtropical deserts, known as sand seas, is
by McKee(1979),
An often bewildering variety of terms have been used to
describe or elassify dunes. Many of these terms arc partly or
wholly synonymous, and many have been used only in local
areas.
This diversity in terminology is in great part a refleetion
of the great variability in conditions under which dunes form,
as noted below. Breed and Grow (1979) provide an extensive
eorrelation among the various classifications schemes.
Dunes vary extremely widely in seale and geometry. This
should not come as a surprise, inasmuch as several faetors
exert important controls: the size, sorting, and thickness oi
loose sediment available to be formed into dunes; wind speed
and its variability: wind direction, and its variability: vegeta-
tion:
and downwind change in time-mean sediment transport
rate,
whieh determines whether the dunes are experiencing net
aggradation, net degradation, or a neutral condition. Leaving
aside the stabilizing effect of vegetation, it is generally
considered that the most important of these are the nature of
the wind regime and the thiekness of available sediment.
Many classifications of dunes have been proposed. These
classification arc of two kinds: morphologic or morphody-
namic. A morphologic classification is based solely on
observable dune geometry, whereas a morphodynamic classi-
fication takes account of assumed eharactciistics of wind and
transport regimes. One of the niosl common morphodynamic
classifications divides dunes into longitudinal, transverse, and
oblique with respect to the direction of net sand transport.
This classification is not without problems, however, beeause it
is commonly difficult or even impossible to establish the
direction of net sand transport, owing to the \ariability of the
wind and the long timescales over which the wind and
transport must be averaged.
In some regions with dunes, the sand-moving winds are very
consistently from a narrow range of directions, so that there is
a simple relationship between the dominant wind and the net
sediment transport direction. In other eases, however, tbe
sediment-transporting wind is complexly multidirectional, and
the relationship between the wind and the net transport
direction is less obvious. In the case of star dunes, the dunes
are formed in areas with little or no net transport.
Much of the variability in dune characteristics is a
eonsequence of the effective average thiekness of transportable
sand in a given area in relation to the potential size of the
dunes. For a given dune size, some minimum thickness of sand
is neeessary to assure full coverage of movable sand even in
dune troughs; eonversely, for a given thickness of movable
sand,
there is an upper limit to dune size without "starvation"
(i.e..
exposure of rigid substrate) in dune troughs.
The effect is especially relevant to variations in geometry of
transverse dunes. When only a minimum of movable sand is
available, barchans are the typical dune form. Barchans are
isolated,
crescent-shaped dunes with sharp horns pointing
downwind.
They have a gentle upwind slope and generally an
angle-of-repose downwind slope. With increasing availability
of movable sand, the barehans begin to join at the horns, and
they eventually form continuous but sinuous and irregular
transverse ridges. Eventually even troughs are occupied by
movable sand.
Linear dunes, often strikingly regular in their geometry and
commonly oriented nearly parallel to what is thought to be the
direction of net sand transport, are common in regions of
abundant sand supply. Such dunes are especially characteristic
of the world's great subtropical deserts. In most such regions,
the troughs between the linear dunes are nonetheless starved of
sand.
In such areas individual linear dunes can be continuous
for distances of
as
great as two hundred kilometers. The origin
of these linear dunes continues to be controversial. In one
view, they are formed imder the influence oi helical secondary
circulations superimposed on the mean wind in the atmo-
spheric boundary layer. Such circulations are known to exist,
but no convincing direct evidence has yet to be adduced that
they are instrumental in the generation and maintenance of
linear dunes.
Smaller-scale dunes are commonly seen to be superimposed
on larger-scale dunes. Dunes with superimposed smaller dunes
are termed compound dunes. There is often uncertainty about
whether the superimposition represents an equilibrium condi-
tion or a manifestation of adjustment of the dunes to changing
conditions. In at least some cases it seems clear that the former
is the case.
The fundamental question as to why dunes form is an
enduring one. Two approaches, whieh arc not fundamentally
SUKhACE FORMS
711
different, are possible. In one approach, which is relevant to
conditions of minimum sand supply, one ean think in terms oi
the conditions needed to localize a palch of sand tnoving over
a rigid substrate and then cause the patch to grow into one or a
series oi dunes. This approaeh was first developed system-
alically by Bagnold (1941). In the other approach, one can
assume a thick continuous bed of sand and think in terms of
how the sand-moving wind generates one or another dune
type.
In the latter approach, whieh is the one usually taken in
the case of subaqueous surfaee forms, the problem of the
dynamics and classification of dunes might be considered f>om
the perspective ofa "thought experiment" in whieh a full bed
of sand is subjected to a wide range of wind eonditions but also
a gradually shrinking thickness of the sand layer, to reveal
changes as sand starvation makes its appearance and increases
in itnportance. For subaerial dunes, of course, the feasibility of
an actual experiment of that kind is extremely limited, owing
to the large seales involved.
John B, Southard
Bibliography
.-Mien. J,R,L,. I'JSO,
SLIDII
waves: a mode! of origin and internul
sliLiLltJrc. Sedimentary Geology. 26: 281-328.
Allen. .I.R.L,. 1984, Sedimentary Structures: Their Charaeter and Physi-
cal
Basis.
\olume I
Volume
II. Amslerdam: Elsevier,
Amos. CL,, Bowen, A,J,, Huniley, D,A,. and Lewis. CF.M., l^SS,
Ripple generation tinder the combined influenee of Wiive^ ;ind
currenis,
Continentat .Shelf
Reseittch,
S:
1129
II.Sl,
Anderson, R,S,, 1987, A iheoretical model for aeolian impact ripples,
Sedimetilology, 34: 44.1 956.
Arnoll, R.W,, ;md .Southard. J.B,, 1990, Expcrimentiil study of
combincd-flou bed conligurations in fine sands, and some
iniplicalions foe slrutificiUion. Jonrnal of Sedimentarv Pelrologv.
60:211-219,
.Ashley. G.M,. IWO, Classitieatlon of large-scale subiiqucous bed
forms:
a new look al an old problem, SF.PM Bedforms and
Beddimi Slruclures Research Sympositim, ./otirnalo/ .Sedimentary
Petrology. 60: 160 172,
Bagnold. R.A., I^ML The
Phy^^ics
of Blown Sand and Desert Dunes.
London: Melluien,
Bagnold, R,A,, 1946, Motions of waves in shallow water, hueraction
of waves and sand bottoins. Royal
S<icieiy
I London
j.
Proceedings.
A187:
1-15,
Breed, CS,. and Grow, T,. !979, Morphology and disiribuiion of
dunes in sand seas observed by remoie sensing. In McKee. t,D,.
(ed.),
A Study of Global Sand Seas, US
Geological
Survey. Profes-
.sional Paper
1052,
pp, 305 }97.
Eiiwood. J,M,. Evans, P,D,. and Wilson, 1,G,. 1975, Small-seale
Lteolian bedforms, JtmrnalofSedinu-niary
Petrology,
45: 554 561,
Flood. R,D,,I981, Distribiuion. morphology, and origin of sedimen-
tary f~urrows in cohesive sediments, Souihamplon Water. Sedi-
menlology, 28: 511 529.
Flood. R,D,. lys.V Classification of sedimentary furrows and a model
for fiiriow initiation and evoltition,
Geoloqical
Soeiety of America
Bulletin, 94: 63(1 639,
Flood. RD,. I9SS, A lee wave model for deep-sea miidwave activity,
Deep-Sea Reseiireh. 35: 973 983.
Flood, R.[),. and Shor. A.N., 1988, Mud waves in the Argentine Basin
and their relationship to regional bottom eireulation patterns.
Deep-Sea
Re.seareh.
35: 943 971.
Gilbert. G.K,, 1914, The Tninsporiation
oI"
Debris by Running Waier,
I'S
Geologieal
Siiryey Professional
Paper
86,
Greeley, R,. and Iversen. J,D,, 1985, Wiitd as a Geologieal
Proeess
on
Earth.
Mars.
\'emis
and Titan. Cambridge University Press.
Guy. IIP,. Simons, D,B,. and Riehardson, F,V,. 1966, Summary of
Alluvial Cluinnel Data from Flume E,\perinients, 1956 1961, US
GeologicalSi/ryey, Professional
Paper
462-i.
l-follister. CD., Flood, R.D,, Johnson. D.A., Lonsdale. P.. and
.Southard. J.B,. 1974. Abyssal furrows and hypeibolie echo traces
on tite Bahama Outer Ridge. Geology. 2: 395'-4()(),
Inman, D,L,. 1957. Waye-Generaled Ripples in Nearshore Sands. U.S.
,Army Corps of Engineers, Beach Erosion Board, Technical
Memorandum KKI.
Lancaster. N,. 1994, Dune morphology and dynatnles. In Abrahams,
A,D,.
and Parsons. A.J,. (eds), Geonunphology of
Desert
Environ-
ments. London: Chapman and Hall,
McHugh. C,M,G,. and Ryan. W,B,F,. 2000. Sedimentary features
assoeiated witb ehannel overbiink flow: examples from ihe
Monterey Fan, Marine
Geology.
163: 199-215,
MeKee. H.D.. 1979, A Sludy of Global Sand Seas. USGeologtealSm-
vey Professional
Paper
1052.
Niekling, W,G.. 1994. Aeolian sedimenl transport and deposition. In
Pye,
K. {ed,). Sediment Traitsporl and Deptisiiiomtl Proces.ses.
Blaekwell. pp. 293-350,
Normark. W,R,. Hess, G,R,. Stow. D,A,V,. and Bowen, A.J., 1980,
Sediment waves on Ihe Monterey Fan Levee: a preliminary
physical inlerprelation. Marine Geology, 37: 1-18,
Pye,
k,. and Tsoar. H., 1990, Aeolian Sand and Sand Dunes. London:
linuin Hyman,
Richards. K,J,, 1980. The formation of ripples and diiiies on an
erodible
bed,
Journalof
FluidMeehanies. 99: 597
6IK,
Sharp, R,P,. 1963. Wind lipples,
.Journal
of Geology. 7L 617 636,
Shor. A.N,, Piper. D,J,W,. Huglies Clarke, J.L,.' and Mayer. L.A,,
1990,
Giant llute-like scour and other erosional features formed
by the 1929 Grand Banks turbidity current. Sedimentology. 37:
631-645.
Simons. D.B,, and Richardson. E.V., 1963, Forms of bed roughness in
alltivial channels.
.American
Soeiely of Civil
Engineers.
Transaetions,
128:
2K4 302,
Soulhard, J.B,. 1991, Fxperimental determination of bed-form
stability. Annual Review of Earth and Planetary Sciences. 19: 423-
455.
Southard, J.B,, and Boguchwal, L,A., 1990. Bed configuralions in
steady unidirectional water (lows. Part 2: Synthesis ol flume clala.
.fournal of Sedimentarv Petrology. 60: 658 679,
Southard, J.B,. Lambie. JM,. Federico, D.C, Pyie. H.T,, and
Weidttian, CR,. 1990. Experiments on bed eonligurations in fine
sands under bidirectional purely oscillalory Ilow. and the origin of
bummocky eross-stialificalion, Jottrnal of Sedimentary Petrology,
60:
I 17,'
Van den Berg, J,I-L. and Van Cielder. A,. 1993, A new bedform
stabilily diagram, with emphasis on the transition of ripples to
plane bed over fine sand and silt. In Mar7o, M,, and Puigdefab-
regas,
C (eds). Atluvitd Sedimentation. Inlernalional Association
of Sedimentologists, Special Publication. Volume 17. pp, 11 21,
Vanont, V.A., 1975, Sedimenialion Engineering. New York: American
Society of Civil Engineers,
Viekman.'B,F,, Flood, R,D,. Wimbtish, M,. Faghri, M,, Asako. Y..
and Van
l,,eer,
J.C. 1992, Sedimeniary furrows and organized flow
structure: a study in Lake Superior. Limnology and Oceanoi-raphy.
37:
797 812.
Wilson, LG,, 1972, Universal discontiiuiilies in bedforms produced by
[he wind. Journal of Sedimentary
Pelrology.
42: 667 669.
Cross-references
Bedding and Inlernal Slriietures
Bedset and Laminaset
Cross-Stratification
Dtines. Kolian
Eolian Transport and Deposilion
Flaser
Flow Resistance
Flume
Gutters and Gutter Casts
Hummocky and Swaley Cross-Slratification
Microbially Indueed Sedimentary Structures
Oflsliore Sands
Paleoeurrent Analysis
Parting Lineations and Current Crescents
712
SURFACF TEXTUKbS
Physics of Sedimenl Transport: Tbe Contributions oi' R.A, Bagnold
Ripple, Ripple Mark. Ripple Structure
Scour. Scotir Marks
Sedimentary Structures as Way-up Indicators
Tool Marks
SURFACE TEXTURES
Introduction
The advent of the scanning electron microscope (SEM)
brought ii new tool to the sedimentologists itrmory. In
particular, it allows images to be obtained at a variety of
magnilk-ations (10
10
k times) together with a depth of field
unobtainable with a light microscope, especially at high
magnifications. Despite the high cost and considerable
complexity of the instruments introduced in the late 1960s a
number of workers exploited the relalivc ease of geologieal
sample preparation for the SEM. It was, (br example, no
longer necessary to produce surface replicas, necessary with
transmission electron microscope techniques, as the surface
itself was imaged. Subsequent advances in ease of use of the
SEM has, in effect, meant that bench top instruments can now
be used as "super mieroseopes" for a wide range oi'
sedimentological purposes and that even casual users can
obtain a high degree of proficiency at a basie level.
One early geological use of the SRM. by Krinsley and co-
workers, was to view surface textures of individual sand-size
grains as an extension of optical tnicroscopic examination oi
detrital grains. An outline description of the SEM itself and a
recent review of surface textures together with a general
overview of the use of the SEM in glacial geology is given in
Whalleyl 1996) and a wide-ranging review of the technique and
assoeiated methods can be fotind in varions papers in Marshall
(1987) and the more general books on the use of SEM in
sediments by Stuart and Tovey
(1981.
1982).
Principles of the technique
Earge energy inputs into transportive sedimentary systems
usually produce atirition
oC
particles as well as sorting. It is the
small scale, intcr-particlc, attrition, which produces changes in
shape and rounding, as well as impacts oC surfaces, that are
revealed by SEM examination. Thus, the fundamental concept
behind the examination of surface textures is that detrital
grains, undergoing transport, will abrade and that the core (or
original) grain will reflect the process of sediment movement
by the development of a distinctive surface texture. This
texture is in addition to, but assoeiated with, any changes
produced in the grain rounding or shape modification that
may occur. Chemical action on grain surfaces may be lound
under snitable environments and surface textures are also
produced in as a result of sueh modifications. Eor many years
it was recognized that "frosting" was a typical textnre seen on
detrital grains from both present-day deserts as well as grains
from the geological record. Krinsley's exploitation of the St"M
was to show that a range oi distinctive microscale features
could be identified from various sedimentological environ-
ments (e.g., Krinsley and Marshall (1987) for an overview).
The technique thus associates specific surface textures (often a
range of features) on the surfaces of grains with corresponding
sedimentological events, thereby giving clues as to the
sediment's history. Quart?, most usually sand-sized, grains
are generally used for identification of surface textures because
of their ubiquity, although garnets or heavy minerals (e.g.,
Sticglitz, 1969) have also been used. There are however
dangers in using surface textures as a unique method of
sediment fingerprinting.
Sampling and preparation of material
Cost, bolh "beam time'" aiid subsequent viewing time, is a
problem with SEM examination. As SEM surface texture
examination is an individual rather than a bulk teehnique. the
representativeness of a sample is. or may be, problematical.
Even a minimum of 30 grains ean be expensive if multiple
samples are to be examined. As the technique is largely
dependent upon visual characterization (see below) a bias may
be introduced so blind sampling and sediment preparation is
recommended. Indeed, to be as objective as possible, the
person viewing grains and identifying the textures should not
be the person who needs the data or interprets the findings.
Such precautions may not be necessary if the technique is used
as check or as an adjunct to other methods of characterization.
Grains for mounting should be cleaned by gentle washing.
adhering material being removed with hydrochloric acid and
stannous chloride. Ultrasonic treatment has been discouraged
by some workers although it may be useful unless grains are
very delicate. Sieving shotild be avoided at any stage of the
sampling/selection process as (his may introduce new surface
textures. It is usual to examine grains in the 25O-lOOOj,tm
range. At the simplest level, preparation involves sticking
grains to an SEM specimen stub and grains of this size can be
picked up by micro-forceps under a binocular microscope and
arranged on the specimen stub. This allows easy identification
of quartz (or other mineral type) but may introduce a sampling
bias.
Under the SEM it may not be possible to identify the
mineralogy without the use of an elemental detector (usually
an energy-dispersive spectrometer. EDS/EDX) which,
although useful, is still an expensive ancillary item for desk-
top SEMs. The specimen stub surface is best coated with
double-stick tape as grains placed carefully on this usually
adhere well. Coating by gold (sputtering) or carbon (evapora-
tion) provides good electrical continuity over the surface, poor
contact gives eharging under the electron beam which degrades
the image. Grains may be placed on stubs in rows or in a spiral
to allow subsequent "revisiting". Typically, rather more than
30 grains should be placed to allow for poor photography, etc
at the observation stage. Sprinkling is an alternative way oi
providing many grains for examination but experience shows
this frequently gives poor electrical connection for larger
grains and less easy identification due to overlapping.
However, sprinkling may be the only way to mount small
grains (<25()(.im). An electronic spatula is one way to
distribute grains carefully and a gentle jet of clean dry air
both removes surplus grains and improves adhesion.
Observations under the SEM
"Ihe normal (secondary/emitted) mode of SEM observation is
generally employed for visual eharacterization. There may be
advantages in cathodo-Iuminescence (Smart and Tovey. 1981)
and back-scatlcred images tor certain specialized needs but
are rarely necessary. The traditional way of observing has
SURFACE TEXTUKfS
713
been to identify a row of grains and move progressively along,
identifying grains and recording on a lab sheet, and to take a
low power photograph of each with one or more higher power
(200 lOOOx ) photographs of each grain and its surface
textures. Because of the expense and/or pressure on beam
time,
photographs are best examined tor subsequent feature
recognition rather than when observing by SEM. A certain
amount of experience is necessary to identify useful features
and this tends to slow the direct examination process.
Conventional photography may take a minute or more to
obtain an image. With the advent oi quality video photo-
graphy it has become mueh easier to take many images, and at
a \ariety of scales, so it is most convenient to carry out all the
detailed examination "off-line". This reduces the cost of both
beam time and the time needed to perform conventional
photography, Ideniificalion of grains on lab sheets is
important as it may be necessary to revisit grains to obtain
quality photography for publication purposes.
Identification of textures
That it should be teasible io dilTerentiate between the basie
cmironments of eolian. beach, iluvial atirition, etc. is perhaps
remarkable. However, the early experiments did show that
grains tVom an undisputed, single component provenance
examined by SEM did provide unique and different textures. A
"surface texture" is often identified by a suite of microfealures
and early papers listed a few of these components and showed
representative micrographs. To identify provenance therefore,
grain textures are identified from photographs of sample
grains which are then compared with catalogued images. The
"SEM atlas" of Krinsley and Doornkamp (1973) is still widely
used for this purpose. A few example illustrations are given in
this article. There has been discus.sion about the degree to
whieh interpretation can be made. Care must be taken to avoid
over-interpretation, especially where grains have undergone
mtiltiple transport cycles. F^urthcr discussion on this can be
found in Whalley (19%).
An apparently iTiore quantitative method than texture
comparison was subsequently evolved whereby the individual
features or texture components were recorded and then
compared with lists of features categorized from known
emironments. These lists were built up from both original
and previously published papers. The number of texture
components may vary tVom paper to paper bul percentages/
histograms or presence/absence have been used to provide
assessment oi a suite oi textures (Bull and Culver. 1979). The
paper by Higgs (1979) is useful as it provides both a semi-
quantitative assessment and a selection of micrograph.s. It
should be noted that the range of environments which ean he
identified uniquely is somewhat limited, although various
authors have tried to extend the possibilities. Some oi these
possibilities are mentioned below (see also Whalley. 1996).
Sedimentological information abotit environments is im-
printed upon the tabula
rasa
of an original grain. This is easily
appreciated when grains are tVcshly weathered but it is less
easy lo identify environments unambiguously if grains are part
ofa multi-stage cycle. Eurther, it should be noted that many of
the features or texture components are not independent. Eor
example, a grain released from a rock mass by weathering or
by impact, such as a block falling on talus, bolh produce sharp
edges and corners, stepped and rippled surfaees ("Wallner
lines") as a direct produci oi brittle fracture (see Gallagher.
1987.
for a detailed discussion). These texture suites are typical
of mechanically-weathered sediments but also of some glacial
sediments. Thus, although there has been considerable energy
involved in the creation of new surfaces, the surface texture
is no different. This has produced some problems in the
recognition of a truly "glacial" surface texture. Although
modified grain edges, gouges and chatlermarks may be found
and which may be the result oi true subglacial action, their
significance has been debated. Orr and 1-olk (1983) suggested
that chattermarks are not necessarily environmentally con-
trolled, rather that the visibility of ehattermarks trails is related
to mineralogy and weathering. It is probable that there be
several ways to produee sueh teatnres.
Modifications to an original grain surfaces are of two basic
kinds;
mechanical and chemical. As well as the complete
fracture through a grain, the former also relates lo modifica-
tion of surfaces or edges by impacting, usually spinning, grains
in a tltiid medium. Edges and eorners arc abraded first because
spinning grains catch each other (Kigure S6(l), This action also
Figure S60 (A) A gr<iin of crushed Br.izilian qu^irt? from .i s sdmpie which has been abraded in
A
closed ,iir jel system for 4 hours. Picture
width idOnm. iB) Detail of lA) showing ihe jbrasion of (he edge and corners although ihc main fates are almost devoid of any impdcl textures.
Picture width = 2,!()
714
SURFACE TEXTURES
produces a gradual modification in grain angularity to produce
more rounded grains. Thus, the basic shape may also be
modified. Surfaces, being less exposed to spinning grain
protuberances, are less prone to modification than edges. In
experiments of rounding of sand grains in eolian environments
it has been shown (Whalley elal., 1982) that textures are of
meehanical formation even though they appear to be similar to
some textures resulting from chemical action.
Chemical action on grains is usually found in the tbrm of
surface etching, or as silica re-precipitation: some crystal-
lographic orientations are particularly susceptible to etching.
The typical forms are varieties of etching and diagenetic
features (Krinsley and Doornkamp. 1973). Crystailographie
orientation with respect to etched surfaces may also play a part
in determining the size and prominence of etching and silica re-
deposition. Weathering on mineral grains has been used as a
means of dating deposits. Eor example. Black and Dudas
(1987) used a method of estimating surface weathering of soils
of glacial origin in Alberta, although others have found no
clear association between post-deposilional diagenesis and
time.
It would appear that there needs to be further research in
elueidating the actual formative mechanisms for chemical as
well as mechanical surface texture formation.
It should be noted that identification of these surface
textures, both as individual components and suites, is
ultimately the result of a pattern matching processes in the
brain. This is a surprisingly effective and discriminating
process but is very difficult to make quantitative. Only the
feature counting method allows for some quantification but it
too is applied after the pattern recognition process has been
accomplished and is still subject to vagaries of feature
identification. Categories may overlap or result from differ-
ences in crystailographie orientation or of applied stress field.
Only one published experiment (Culver ciaL. 19t^3). using the
feature counting method, shows a measure of the operator
variance of the surface texture method. It may be signifieant
however that (from unpublished data) live experienced users of
SEM lexture analysis, vs'lten given a set of micrographs
"blind", correctly and uniformly idcnlitied the provenance/
transport history of six samples in less than five minutes. The
conclusion suggests that using the brain as a classifier may
work at least as well as lengthy and tedious feature counting
methods.
Particle form: surface texture, shape,
angularity, and roundness
Traditionally, "size'" is measured and interpreted as a bulk
attribute whereas eomponents of a particle's "form"" are
analyzed individually and then aggregated. The components of
form E (effectively in 2 dimensions, the observation plane, if
the sphericity is ignored) can, following Whalley (1972), be
related by:
E = f(sh.sp,a. r. t)
Where sh = shape. sp = sphericity, a = angularity,
r = roundness, and t —surface texture. These inter-related
concepts are associated with the attrition of grains under-
going transport processes. Shape/sphericity, angularity and
roundness are numerically ascribed but the chart comparisons
for angtilarity-roundness are of only ordinal measure. True
quantitative measures, sueh as Eourier and fractal methods.
have been employed by perimeter quantifieation (see Orford
and Whalley. 1991. for review and papers in Marshall. 1987),
However, there are still very few published studies v\hich
incorporate qtiantitative assessment of form components
together with surface texture analysis. Automatic image
analysis methods exist for some simple shapes but these have
not, to date, been used with the rather complex surface
textures found on the three dimensions of quarU grains.
However, it is possible that in the future neural networks and
high powered computing may produce some degree of feature
extraction which can aid quantification and improve objectiv-
ity as well as speed analysis.
Uses and limitations of the technique
After the first liush of papers using SEM to examine surface
textures there were some sceptics about the viability of the
method. This was less a failure of the technique but rather that
it was sometimes used to "over-interpret" samples. Subse-
quently, despite recognition of the technique's utility, it has
not developed into a necessary form of sediment analysis.
However, it is a useful adjunct, in association with techniques
such as size analysis, to assist identification of sedimentary
environments. It can be especially useful where there are only a
few grains available such as from an ice core or wind-blown
collector or in tbrcnsic sedimentology.
The following sections summarize the main "texture-
environments" which have been used with SEM surface
texture analysis and add to comments made above. Photo-
micrographs are provided to show some of the ideal te.xtures.
However, published books and papers must be used to show
the range of textures for any analyzes. The main works which
show good examples of textures and the use of the technique
include: Krinsley and Doornkamp (1973). Marshall (1987).
Whalley (1978. 1996), a bibliography of older references is
given in Bull efaf (1986),
Weathered and glacial textures
As noted above, "weathered" (i.e.. untranspoited) grains have
a similar, if not identical, surface texture to certain grains of
glacial provenance. In particular, grains transported on a
glacier surface as well as perhaps some englacially-transportcd
grains have not been subjected to any tbrm of modification.
Thus,
it is not possible to use surface textures to differentiate
these environments without considerable care. It is not
possible to say that a lexture is definitively glacial without
other, supporting, evidence. However, it has been possible to
show that certain modifications to the angular characteristics
can be given by subglacial grinding or fluvio-glacial action.
Mahaney (1995) has suggested that glacial grains can be used
as indicators of age and ice sheet eharacteristics. The transport
mode by glaciers must be clearly differentiated: supraglacial
transport is unlikely to produce any textures which are
different from weathered grains or grus. subglacial action
may produce distinctive abrasion of some edges and faees.
Glacio-marine deposits
It has been suggested that glacio-marine environments can be
identified by SEM texture evaluation. However. Ihis has
usually necessitated the identification of "glacial" textures.