TWO N E W M E T H O D S OF P A L E O C U R R E N T D E T E R M I N A T I O N F R O M T R O U G H

CROSS-STRATIFICATION l

PETER G. DeCELLES AND RICHARD P. LANGFORD

Department of Geology
Indiana University
Bloomington, Indiana 47405
AND
ROBERT K. SCHWARTZ
Department of Geology
Allegheny College
Meadville. Pennsylvania 16335

ABS-I-RACI: Trough cross-stratification is the most abundant, accurate, and efficient paleocurrent-direetion
indicator in fluvial deposits. However, exposures iri which trough axes can be measured are rare. Two
new methods are presented for the determination of paleoeurrent directions from the more common ex-
posure types--namely, the use of three-dimensional exposures of oblique cuts through individual trough
limbs, and two-dimensional exposures of sets of trough cross-strata. Three-dimensional exposures of trough
limbs can be measured and treated statistically on a stereographic plot to determine the average trough-
axis orientation. In two-dimensional exposures, characteristic asymmetry of the basal scour surface and
truncation of foreset laminae can aid in estimation, within about 25 °, of paleocun'ent direction.

INTRODUCTION lunate (or crescentic) dunes (Allen, 1963; Wil-

Purpose
The utility of trough cross-stratification in
paleocurrent studies has been greatly limited
because of the rarity of three-dimensional ex-
posures of troughs that permit accurate deter-
ruination of the orientations of trough axes. The
purpose of this paper is to present and evaluate
two new methods for the collection of paleo-
current data from more common types of ex-
posure: i) oblique, three-dimensional expo-
sures of individual trough limbs, and ii) two-
dimensional exposures oriented at various an-
gles to trough axes. Although the focus of this
paper is on fluvial cross-strata, the new meth-
ods can be applied to marine cross-strata as well.
Discussion
Trough cross-strata constitute the majority
of all stratification in fluvial deposits (Frazier
and Osanik, 1961; Harms et al., 1963; Dott,
1970; Williams, 1971; Jackson, 1980). Large-
scale (5-200 cm in thickness) trough cross-strata
are formed by the downcurrent migration of
~Manuscript received May 24, 1982; revised October
21, 1982.
liams, 1968; Coleman, 1969; Harms et al.,
1975; Jackson, 1976) or by the burial of large
(for example, 0.6 m deep, 2.4 m wide, 15 m
long), spoon-shaped scours by downstream
migrating barforms and subaqueous sand waves
(Harms et al., 1963; Williams and Rust, 1969;
McGowen and Garner, 1970). Small-scale (0.5-
5 cm in thickness) trough cross-strata are formed
by the migration of cuspate, linguoid, and sin-
uous crested ripples (Hamblin, 1961; Harms et
al., 1963; Bhattacharyya, 1969; High and Pi-
card, 1973). Trough cross-strata develop under
the influence of primary currents on virtually
all major bar types and on channel floors (Harms
and Fahnestock, 1965; McGowen and Garner,
1970; Williams, 1971; High and Picard, 1974).
Axes of large- and small-scale trough cross-
sets are generally aligned parallel to local pri-
mary flow direction and show a relatively small
degree of scatter (Dott, 1973; High and Picard,
1974). A comprehensive review of the statis-
tical characteristics of trough cross-stratifica-
tion paleocurrent data is presented by High and
Picard (1974).
Planar cross-stratification in fluvial systems
is formed by the lateral and downstream mi-
gration of major bars (especially transverse bars)
and by smaller scale, straight-crested bedforms

JOURNAL OF SEDIMENTARYPETROLOGY, VOL. 53, NO. 2, JUNE, 1983, V. 0629--0642

Copyright © 1983, The Society of Economic Paleontologists and Mineralogists 0022-4472/83/0053-0629/$03.00
630 PETER G. DeCELLES, RICHARD P. LANGFORD, AND ROBERT K. SCHWARTZ
which are superimposed upon bar margins and
tops (Stewart, 1961; Smith, 1970; Williams,
1971; High and Picard, 1973, 1974; Harms et
al., 1975; Jackson, 1976; Levey, 1978). In
braided systems, most bar migration occurs
during flood stage, producing large-scale av-
alanche foresets (Williams and Rust, 1969;
Williams, 1971; Smith, 1972; Levey, 1978).
Local current direction is controlled by bank
slumping during rising flood stage and by sand
bar deposition and emergence during falling
stage (Coleman, 1969). As flow strength and
depth decrease during falling stage, bar tops
emerge and currents sweep around and flow
parallel to lee sides of bars, reworking bar
margins and producing smaller scale bedforms
with orientations totally unrelated to main flow
direction (Simons and Richardson, 1965; Smith,
1972). Flow over tops of bars is divided and
directed as much as 90 ° away from the main
channel flow direction. The resulting sedimen-
tary structures, and in particular planar cross-
strata, have orientations which are produced
by a complicated and not fully understood in-
terplay between overall channel morphology,
discharge, rate of stage change, flow velocity,
local bartop topography, sediment load, and
grain size (Simons and Richardson, 1965;
Coleman, 1969; Williams, 1971; Smith, 1972;
Schwartz, 1978).
Although processes may differ somewhat in
meandering streams, similar phenomena have
been reported in meandering systems by Bluck
(1971), Jackson (1976), and Schwartz (1978),
and in a sinuous braided system by High and
Picard (1974). However, the scattering effect
produced on planar foreset orientations is mag-
nified by the sinuous nature of meandering-
stream channels and the greater diversity of
major bar types, especially scroll bars and point
bars (Jackson, 1976; Smith, 1978).
Smith (1974) has argued that the preserva-
tion potential of planar cross-stratification
formed by the reworking of flood-stage de-
posits is high in braided systems, and Schwartz
(1978) found corroborative evidence for ge-
netically similar structures in a meandering
system.
As shown in studies of modern braided and
meandering streams by Frazier and Osanik
(1961), Coleman (1969), Williams (1971),
Bluck (1971), Smith (1972), High and Picard
(1974), and Jackson (1976) and in studies of
ancient fluvial deposits by Pick (1964), Meckel
(1967), Dott (1973), Michelson and Dott
(1973), and High and Picard (1974), orienta-
tions of planar foresets show a relatively high
degree of scatter and deviation from primary
flow direction. As emphasized by High and Pi-
card (1974), unless numerous data (20-25
measurements) are collected, planar cross-
stratification is not a reliable indicator of pri-
mary paleocurrent direction.
In contrast, many of the same studies noted
above determined that axial orientations of
trough cross-strata are significantly less scat-
tered and also tend to deviate less from the pri-
mary flow direction. Unfortunately, however,
poor exposure usually hinders the accurate de-
termination of trough-axis orientations. Yet, in
many outcrops three-dimensional exposures of
individual limbs, or well-exposed, two-dimen-
sional oblique cuts through troughs are com-
mon. The two methods of paleocurrent data
collection presented in this paper permit rela-
tively accurate estimation of paleoflow direc-
tion for these two types of exposures.
METHOD I: THREE-DIMENSIONAL EXPOSURES OF
T R O U G H LIMBS
Figure 1 illustrates a common exposure of
trough cross-stratification. A number of indi-
vidual troughs are shown which are cut by the
outcrop face at an oblique angle--that is, at
an angle other than fully transverse to paleo-
flow. These foresets and scour bases can be
easily measured, even along minor surfaces (for
example, a couple of centimeters), with a cir-
cular dip-direction indicator (Pryor, 1958) and
Brunton compass. Opposing trough limbs usu-
FIG. 1.--Photographof obliquely cut trough cross-sets,
showing right-hand (R) and left-hand (L) limbs. Fromthe
Lower Cretaceous Kootenai Formation of southwestern
Montana.
 MEASUREMENT OF TROUGH CROSS-STRATIFICATION
ally produce a bimodal distribution of dip-di-
rection azimuths, in contrast to a unimodal dis-
tribution for trough axes (Dott, 1970). These
individual, opposing-limb sets can usually be
recognized at the outcrop and assigned right-
or left-hand designations, measured and re-
corded as such (Fig. 1). A more distinct bi-
modality of right- and left-hand groups, re-
flecting the two limb populations at a particular
site, will result from i) measurement at points
more distal from the near-axis position of a
given trough and ii) small natural variance in
the orientations of trough axes. Between 15 and
30 measurements (depending on exposure), di-
vided approximately equally between right- and
left-hand limbs, seems to be the optimum
number per outcrop.
Each data point can be expressed in terms
of the orientation of a foreset or scour-base
plane. Poles to these planes are plotted stereo-
graphically (Wulff Net) and labelled as right-
or left-hand limbs. Because of the purposeful
sampling of opposing limb sets, two clusters
of poles typically result (Fig. 2). Overlapping
pole populations may result as a function of
sample-point proximity to the trough axis,
obliquity of the cut to the trough axis (see sec-
tion on Method II), or a large trough-axis azi-
muth variance. In the case of distinct bimo-
dality, two stereonet procedures may be used
which yield the same average azimuth for trough
axes. For either procedure, an average pole from
each cluster is chosen by visual estimation or
stereographic contouring. A ~biplanar method"
may be used in which the intersection of the
two average-limb planes gives the orientation
of the trough axis (Fig. 2). Alternatively, the
pole to the great circle which passes through
the two average-limb points gives the orien-
tation of the average trough axis (Fig. 2). The
latter approach is most useful in the case of
poor bimodality where estimation of average-
limb values is difficult and a best-fit great cir-
cle can be determined for the wider scattering
of data points.
If tectonic rotation of strata has occurred,
special caution should be exercised in the field
to accurately determine bedding orientation. The
plunge of undisturbed trough axes is com-
monly less than 10° (Slingerland and Wil-
liams, 1979). We have found that in deformed
strata, the dip of bedding and the dip of trough
limbs are commonly quite close in value. If so,
an inaccurate bedding estimate may result in a
631
FIG. 2,--Stereoplot of trough-limb data showing clus-
ters of right-hand (open circles) and left-hand (closed cir-
cles) poles. Average poles are indicated by stars. The in-
tersection of the two corresponding great circles and the
pole to the best-fit great circle both give the trough axis.
180 ° azimuth error upon stereographic resto-
ration as a consequence of the incorrect as-
signment of bedding dip as slightly greater or
less than cross-stratification dip.
If tectonic rotation about a single horizontal
axis (for example, for a non-plunging fold) or
double rotation about a set of axes (for ex-
ample, for a plunging fold) is necessary, cor-
rection can be made stereographically on an
individual basis (that is, each pole can be cor-
rected), after the average poles have been cho-
sen or after the trough-axis orientation has been
determined. Implicit in rotation procedures are
assumptions concerning the kinematics of de-
formation and a simplified model for fold ge-
ometry (for example, cylindrical shape). The
amount of error resulting from use of a single-
rotation versus double-rotation technique is de-
pendent upon the amount of fold plunge and
location of cross-stratification measurement
within the fold (that is, proximal or distal to
the nose of the fold). For the most commonly
used stereonet procedure, which assumes a cy-
lindrical model, no error results from single
rotation of points at the nose of a fold, whereas
error increases outward along the limbs with
the increasing angle of deviation between the
trend of the fold axis and dip direction of the
632 PETER G. DeCELLES, RICHARD P. LANGFORD, AND ROBERT K. SCHWARTZ
outcrop. The maximum possible error (equal
to the plunge value) occurs at dip-direction po-
sitions 90 ° from the nose. In general, it is sug-
gested that for low angles of plunge (<25°), a
single rotation about bedding can be made for
any point on the fold without significant error
in the restored trough azimuth ( < 10° error at
limb positions up to 650-75 ° away from the
nose; maximum error 20°-25°). A double ro-
tation is suggested for more steeply plunging
folds, particularly at limb positions where higher
angles of dip-direction deviation from the fold
axial trend occur and where plunge of strata is
still suspected or can be determined.
Discussion
The advantages of this approach are several.
First, trough foresets and scour bases are dis-
tinguished from planar foresets. Erroneous pa-
leocurrent interpretations can result if trough
limbs are treated as planar foresets. For ex-
ample, if it is not recognized that the foresets
are trough limbs, the resulting bimodal distri-
bution could be incorrectly interpreted as a bi-
modal paleocurrent pattern, when in fact it is
merely a reflection of trough geometry. Fur-
thermore, the interpreted paleocurrent direc-
tions would be nearly 90 ° different from the
true direction.
Second, the orientation of the average trough
axis is used as a paleocurrent-direction indica-
tor. This is a more accurate and efficient ap-
proach than just measuring as many foresets as
possible, regardless of type, and plotting them
on a rose diagram as if they were planar fore-
sets.
Finally, the product of the analysis is one
paleocurrent arrow per outcrop. This is pre-
ferred to other methods that leave the inves-
tigator with a sometimes puzzling frequency
distribution or rose diagram that must in turn
be deciphered. Use of Miall's (1974) "weight-
ing by volume" method is not applicable here,
because it is only appropriate in situations in
which one paleocurrent arrow is obtained per
structure.
One potential source of error in this method
is that it assumes symmetrical fdling of troughs.
If this were not the case, however, it should
show up on the stereographic plot as a skewed
distribution of poles or in the field as an excess
of either left-hand or fight-hand limbs. If only
scour bases are measured, asymmetric filling
should not affect the results.
An estimate of axis orientation is still pos-
sible for more widely spread, skewed unimod-
al data by assuming that an entire trough shape
can be approximated by a great circle fit (Fig.
2) defined with data from laminae on just one
side of the scour base. Such an estimate should
be considered tentative and used only in lieu
of more geometrically definitive bimodal data.
A second potential problem is that the method
assumes that only one population of troughs is
present. This problem is most acute in marine
sediments, where strong polymodal currents
exist. In fluvial systems the primary flow is
always unimodal, thus minimizing the possi-
bility that more than one trough population is
present.
The method has been successfully tested on
trough cross-stratification in the Lower Cre-
taceous Kootenai Formation (DeCelles, 1982)
and the Upper Cretaceous Blackleaf Formation
(Schwartz, in press) in southwestern Montana.
METHOD II: TWO-DIMENSIONAL EXPOSURES OF
TROUGH CROSS-SETS
It is often the case, especially with well-ce-
mented, contact-metamorphosed and very
coarse-grained material, that exposure of trough
cross-stratification is limited to two dimen-
sions. One example of the latter situation is the
Fountain Formation (Pennsylvanian) at Mani-
tou Springs, central Colorado. Very coarse to
granular sandstones typify many outcrops
wherein large-scale trough cross-strata are
present (Fig. 3). These troughs, however, have
not proven useful as paleocurrent indicators
because their two-dimensional outcrop geom-
etry prevents measurement of either trough axes
or limbs. It appears that the coarse grain size
is at least partially responsible for the paucity
of three-dimensionally exposed troughs.
It is possible, however, to gain some insight
toward paleocurrent direction from two-di-
mensional exposures without the usual mea-
surements. Slingerland and Williams (1979)
developed a quantitative approach, but it is only
applicable to troughs in tectonically unde-
formed, near-horizontal strata. The problem is
here addressed from a qualitative, theoretical
viewpoint, and sketches of two-dimensional cuts
at various angles through different trough types
are presented. It is hoped that these sketches
will be useful for field estimation, within about
25 °, of paleocurrent direction from two-di-
 MEASUREMENT OF TROUGH CROSS-STRATIFICATION
I~G. 3.--Photograph of a two-dimensionally exposed
trough cross-set in the Pennsylvanian Fountain Formation
near Manitou Springs, Colorado.
mensional exposures, regardless of tectonic
deformation of the strata.
Geometry and Terminology
On the basis of data from the literature, Slin-
gerland and Williams (1979) reasoned that about
80 percent of the time a random cut through a
large-scale trough cross-set will transect the
portion of the trough base which approximates
a gently plunging (<10 °) semicylinder. The
remaining 20 percent of the time a cut will ex-
pose portions of the more steeply plunging (20°-
25°), scoop-shaped, upcurrent end of the trough.
The former will be referred to as cylindrical
troughs, and the latter as scoop-shaped troughs
(Fig. 4). Various cuts will be referred to by a
combination of two hyphenated adjectives as
shown in Figure 5.
~TROUGH AXIS
FIG. 4,--Idealized trough with cylindrical (a) and scoop-shaped(b) portions.
633
Cylindrical Troughs
Transverse-vertical and Oblique-vertical
Cuts.--Vertical cuts through cylindrical
troughs, regardless of the degree of obliquity
with respect to the axial plane, always show
symmetrical basal scour surfaces. Transverse-
vertical cuts (Fig. 5a) show the true trough
cross-section, along the trough axis (Fig. 6).
Infilling foreset laminae appear essentially
concentric and conformable with the scour base.
A current sense (not direction) along the line
of sight, roughly normal to the cut, is all that
can be obtained from transverse-vertical ex-
posures.
An oblique-vertical cut is shown in Figure
5b. Any obliquity between the cut and the
transverse-vertical plane causes an apparent
widening of the trough base and asymmetric
truncation of foreset laminae (Fig. 7). As long
as the cut plane is vertically oriented, the scour
base will be symmetrical. Paleoflow direction
with respect to the cut now has two compo-
n e n t s - o n e in or out of the cut, and one in the
plane of the cut. The latter component is in the
direction of dip of the foresets. The directional
component normal to the cut is not discernible.
However, the closer the cut approximates a true
longitudinal section, the greater the apparent
widening of the trough and number of foresets
that are truncated. This means that a relative
sense of component strengths can be estimated
(Fig. 8).
Longitudinal cuts (Fig. 5e), whether vertical
or nonvertical, centered or offcentered with re-
spect to the trough axis, were discussed by
634 PETER G. DeCELLES. RICHARD P. LANGFORD. AND ROBERT K. SCHWAR1Z

cut

t b.

1 \

t rough - "~
axis

do

i "j ll,.~if i~

FIG. 5.--Various types of idealized planar cuts through a given trough, a) transverse-vertical, b) oblique-vertical,
c) transverse-oblique, d) oblique-oblique; e) longitudinal cuts: 1) centered vertical, 2) off-centered vertical, and 3)
centered nonvertical.

Slingerland and Williams (1979, p. 726). With
these types of cuts one is hard-pressed to even
be sure that one is dealing with trough cross-
stratification, unless numerous troughs of sim-
ilar scale and positive identification are also
present. In some instances a longitudinal sec-
tion can be recognized and the downcurrent di-
rection can be assumed to be in the plane of
the cut and in the direction of dip of the fore-
sets.
Transverse-oblique C u t s . - - C u t s that are trans-
verse but nonvertical (Fig. 5c) produce char-
actefistic apparent deepening of troughs and
truncations of foreset laminae. There are es-
sentially three types of transverse-oblique cuts
(Fig. 9), all of which show symmetrical basal-
scour surfaces.
A cut which passes through the trough at an
angle near, but not less than, the dip angle of
the foresets is illustrated in Figure 9a. Appar-
ent deepening of the trough results and foresets
are truncated downward (Fig. 9b). If a true
transverse-oblique cut exists, foreset trunca-
tion is symmetrical with respect to the trough
 MEASUREMENT OF TROUGH CROSS-STRATIFICATION 635

LEGEND
current
into c u t

j~ into o r
A B out of cut

current
out o f Cut

current
in p l a n e oF c u t

FIG. 6.--Transverse-vertical cut through a cylindrical trough, showing authentic cross section (A-B) of the trough,
transverse to the trough axis. Interpretable paleocurrent component is along the axis but cannot be assigned a direction
in or out of the cut. Legend of paleocurrent symbols is used in some of the ensuing figures.

axis. The closer the cut is to the dip angle o f
the foresets, the more laminae that are trun-
cated and the greater the apparent deepening
of the trough. F r o m such exposures, one can
assume that the foresets are dipping near the
angle of the cut. This observation can be com-
bined with the bedding orientation to deter-
mine whether the paleocurrent direction is in
or out of the cut.
If the b e d d i n g / c u t relationship is like that
shown in Figure 9e and a pattern similar to
Figure 9b is present, one can assume that the
trough foresets are inclined roughly parallel to
the cut. Therefore, the direction o f paleocur-
rent flow is toward the o b s e r v e r - - t h a t is, out
of the plane of the cut. The symmetry in the
shape of the trough base and truncation of

FiG. 7.--Oblique-vertical cut through a cylindrical

trough (top), showing the resolution of a single paleo-
current vector into its components normal to (n) and par-
A B allel to (p) the plane of the cut. Foresets are truncated
asymmetrically on the left side of the trough, which in-
dicates a paleocurrent component to the left in the plane
of the cut (bottom). As shown by the paleocurrent com-
ponent symbol, the direction of the component normal to
r ee • the cut is not discernible (see Figure 6 legend).
636 PETER G. DeCELLES, RICHARD P. LANGFORD, AND ROBERT K. SCHWARTZ
j4
/3
12
FIG. 8.--Diagram illustrating how the apparent width
of trough and number of truncated foresets increases as
the cut more closely approaches a longitudinal cut. Note
that the traces of trough basal surfaces remain symmetri-
cal, which prevents the determination of the paleocurrent
directional component normal to the cut. Also, the rela-
tive strength of the component in the plane of the cut in-
creases as the cut approaches a longitudinal orientation.
For explanation of paleocurrent symbols, see legend in
Figure 6.
foreset laminae indicate that the paleocurrent
component in the plane of the cut is negligible.
If the bedding/cut relationship is as illustrated
in Figure 9f and a pattern like Figure 9b is
present, one can agam assume that the trough
foresets are inclined roughly parallel to the cut,
and therefore the direction of paleocurrent flow
is into the plane of the cut.
A transverse-oblique cut oriented at 50 ° to
100 ° from the angle of foreset dip is depicted
in Figure 9a. Apparent deepening of the trough
results; foreset laminae are truncated upward
and symmetrically arranged about the trough
axis (Fig. 9c). Again this information can be
combined with the bedding/cut relationship to
give the paleocurrent direction either in or out
of the cut. If the bedding/cut relationship is
like that in Figure 9e, one can assume that the
foresets are inclined at a high angle to the cut;
hence the paleocurrent direction is into the cut.
If the same pattern is present but with the op-
posite bedding/cut relationship (Fig. 9f), the
paleocurrent direction is out of the cut.
A transverse-oblique cut oriented at an angle
approximately equal to the foreset dip angle
(Fig. 9a) shows apparent deepening of the
trough and a quasicircular pattern of foreset
laminae (Fig. 9d), for the cut passes through
both upcurrent and downcurrent ends of the
concave foresets. It can be understood by
imagining the appearance of the trace of a
spherical or ellipsoidal body on an obliquely
intersecting plane. Such a trace will always be
circular or elliptical, respectively. The direc-
tion of paleocurrent flow is down the axis, or
toward the base of the trough.
Oblique-oblique Cuts.--Oblique-oblique cuts
are nonvertical cuts whose strikes are not per-
pendicular to the trough axis (Fig. 5d). These
cuts produce asymmetry in both the basal scour
surface and the truncation of foresets. Figure
10 illustrates the key concept in understanding
how the asymmetry of the trough base origi-
nates and how, when combined with the bed-
ding/cut relationship, the asymmetry indicates
the approximate orientation of the trough axis.
Again, the paleocurrent component in the plane
of the cut can be determined by observing which
way the foresets dip. The paleocurrent com-
ponent normal to the cut depends on whether
the foresets dip toward or away from the steep
side of the trough (Fig. 11).
In the authors' experience, oblique-oblique
cuts are the most common type of two-dimen-
sional exposure. Occasionally, a trough may
be exposed in three dimensions. In such in-
stances, the method can be tested on the near-
transverse exposure and checked against the
near-longitudinal exposure.
 MEASUREMENT OF TROUGH CROSS-STRATIFICATION 637

ii

a.

~_~//./~ fo resets

// \
b.

cut

/ bedding

C. / e.

f.

d.

f trough axis
1
Fie. 9 . - - a ) Sketch depicting side view of three types of transverse-oblique cuts through a trough. Cut i dips near
but not less than the angle of foreset dip. Cut ii dips in the opposite direction to the foresets at a relatively high angle
(50°-100°). Cut iii is oriented at an angle approximately equal to that of foreset dip. b) Two-dimensional exposure
of cut i, showing apparent deepening of trough base and symmetrically arranged foresets which have been truncated
downward, c) Two-dimensional exposure of cut ii, showing symmetrically arranged, upward-truncated foresets, d)
Two-dimensional exposure of cut iii, with position of trough axis shown. For discussion of e and f, see text.
638 PETER G. DeCELLES, RICHARD P. LANGFORD, AND ROBERT K. SCHWARTZ

trace of troughbase~
in plane of

trough axis~ - - ~ - ~ --~

BASAL ASYMMETRY SIDE VIEWS PLAN VIEWS

/ //
\, //
\,, \\
FIG. 10.--Conceptual diagram illustrating the origin of basal asymmetry in two-dimensional exposures of trough
cross-stratification and how it can indicate the approximate orientation of a cylindrical trough, a) Oblique-oblique cut
showing the trace of the scour base in the plane of the cut. Distance AC is shorter than distance BC, which causes
the left side of the trace to appear steeper than the right side. b) Oblique-oblique cuts; stars mark steep sides of trough
bases in cuts. c) Diagrams depicting the asymmetry of the traces of basal surfaces in oblique-oblique cuts; side views
indicate approximate bedding/cut relationships, and plan views look down on troughs. If the basal asymmetry and
bedding/cut relationships are as shown in the first two columns, the respective approximate trough orientations relative
to the cut are as depicted in third column.
 MEASUREMENT OF TROUGH CROSS-STRATIFICATION
A stratification. The sketches are not meant to
Fio. 11.--Oblique-oblique cut through a cylindrical
trough and accompanying two-dimensional exposure.
Foresets are truncated at the trough base in the plane of
the cut. In this instance, they dip toward the steep side
of the trough in the cut. This indicates that the paleocur-
rent direction components are to the left and out of the
cut (compare basal asymmetry with Fig. 10c). Had the
paleocurrent flowed in the opposite direction, the foresets
would have dipped toward and been truncated against the
right-hand side.
Scoop-shaped Troughs
Paleocurrent directional components in two-
dimensional exposures of s c o o p - s h a p e d
troughs--that is, upstream ends of troughs (Fig.
4)--can be determined in the same manner as
cylindrical troughs.
Transverse-vertical and Transverse-oblique
Cuts.--Transverse-vertical and transverse-
oblique cuts through scoop-shaped troughs show
patterns which are essentially identical to
transverse-vertical and transverse-oblique cuts
through cylindrical troughs. Paleocurrent di-
rectional components can be ascertained using
the same principles used with cylindrical
troughs.
Oblique-vertical and Oblique-oblique C u t s . -
Because of the geometry of the bases of scoop-
shaped troughs, all nontransverse cuts show
inherent basal asymmetry. Unfortunately, this
makes it theoretically impossible to determine
the paleocurrent directional component normal
to the cut because the type of basal asymmetry
is not characteristic of any one trough/cut ori-
entation, as is the case with cylindrical troughs.
It is still possible, however, to determine the
component in the plane of the cut. As with cy-
lindrical troughs, the paleocurrent component
639
in the plane of the cut is in the direction of
foreset dip.
Measurement Procedure
A compilation of sketches of the various types
of two-dimensional exposures of trough cross-
stratification, possible bedding/cut relation-
ships, and their respective paleocurrent inter-
pretations is presented in Figure 12. They are
given in a condensed, reproducible form and
should be useful in making field interpretations
of two-dimensional exposures of trough cross-
represent exact replicas of what one might see
in the field. Rather, they are intended to illus-
trate the key characteristics involved in deter-
mining the paleocurrent components--namely,
the distortion of the basal-scour surface and the
truncation of foreset laminae. It is suggested
that one first understand the geometrical con-
cepts utilized in producing the sketches. Then
the sketches can be used in the field as a quick
method for estimation of paleocurrent direc-
tion.
The following procedures should be exe-
cuted in the field when interpreting two-di-
mensional exposures. First, determine if the
trough base is symmetrical or asymmetrical. If
it is asymmetrical, the cut is oblique-oblique.
Second, note whether the foresets are asym-
metrically truncated or symmetrically disposed
about the trough axial plane. If they are sym-
metrically disposed the cut is either transverse-
vertical or transverse-oblique. Third, deter-
mine the bedding/cut relationship. Then use
the idealized sketches in Figure 12 to estimate
the orientation of the trough and the dip-direc-
tion components of the foresets. To measure
the approximate orientation of the trough axis,
stand back a few meters from the outcrop and
sight with a compass along the estimated di-
rection obtained from the sketches. This al-
lows recording of an actual measurement, as
approximate as it may be. The chief advantage
is that correction for tectonic rotation becomes
unnecessary. As long as the direction of the
paleocurrent is obtained with respect to the
outcrop face, it will not change drastically upon
rotational correction.
Discussion
Complications in Method II do exist. First,
if oblique-vertical and oblique-oblique cuts
640 PETER G. DeCELLES, RICHARD P. LANGFORD, AND ROBERT K. SCHWARTZ

CUT TYPE I EXPOSURE tCUT R E L A T I O N S H I P S & R E S P E C T I V E

PALEOCURRENT C O M P O N E N T INTERPRETATIONS
TRANSVERSE
-VERTICAL Im ®

my
o>

//" @
/
uJ
O'
m

.J /
m
O
/ ®
i
uJ
(n
uJ

I-
J/ ® \
\\ @

/
xxNr
/ \

O |

1,1,1
O,
m

m
o
/ \

FIG. 12.--Compilation of the various types of two-dimensional exposures of trough cross-stratification. Depending
on the symmetry of the basal scour surface, the dip of the foresets, and the approximate bedding/cut relationship as
seen from the side, one can make the accompanying paleocurrent-direction interpretations. As an example of field
interpretation, suppose that the foresets dip toward and are truncated against the shallow side of the trough. Finally,
supppose the cut dips back into the b e d d i n g I t h a t is, away from the observer. The appropriate exposure type is
illustrated in row 9 under oblique-oblique exposures. The appropriate bedding/cut relationship is the second one in
row 9. Beside the bedding/cut diagram is the paleocurrent-direction-component symbol (see legend in Fig. 6), which
in this example indicates that the direction of paleoflow is toward the right and out of the cut. This approximate
direction can then be measured (see text).
 MEASUREMENT OF TROUGH CROSS-STRATIFICATION
through scoop-shaped troughs can appear iden-
tical to similar cuts through cylindrical troughs,
how can they be distinguished? The distinction
is important because if the trough is not cylin-
drical the method does not work on oblique-
vertical and oblique-oblique cuts. The problem
can only be resolved statistically; as noted ear-
lier, Slingerland and Williams have concluded
that approximately 80 percent of all trough ex-
posures are cylindrical. Thus, the probability
of encountering scoop-shaped troughs is rela-
tively low.
Second, the method assumes that troughs are
symmetrically filled. This problem must also
be approached statistically, and Slmgerland and
Williams determined from a review of the lit-
erature that 50 to 70 percent of foreset laminae
are disposed symmetrically about trough axes.
Therefore, it is believed that in most cases the
method is applicable.
The utility of the technique is predicated on
the belief that rigorous accuracy in paleocur-
rent documentation is not generally attainable.
As stated in the introduction, studies of both
braided and meandering modem streams dem-
onstrate variability of at least 90 ° in local cur-
rent directions. How much more is the poten-
tial for precise accuracy in studies of ancient
fluvial systems destroyed by differences in
preservation potential and incomplete outcrop
distribution?
Strictly speaking, the method can estimate
the paleocurrent direction from a two-dimen-
sional exposure of a trough cross-set within
-+45 ° . For transverse-oblique exposures, the
degree of accuracy is increased to within about
10° of the true paleocurrent direction. A com-
puter program was developed by the authors
to model trough cross-stratification and gen-
erate plots of scour-base and foreset geometry
on variously oriented cuts through a given
trough. From examination of these plots, inter-
pretations of oblique-oblique exposures fall
within about 25 ° of the true paleocurrent di-
rection. This cannot be more precisely quan-
tified because the geometry of a given com-
puter plot is dependent on the values chosen
for trough dimensions. Perhaps with more in-
formation on the shapes and proportions of in-
dividual trough foresets and scour b a s e s - - f o r
example, of the type presented by Pick (1964)--
a more rigorous program could be developed
to produce quantitatively accurate sketches.
In any event, it is suggested that 25 ° of pos-
641
sible error is not enough to render the data
worthless, especially if enough data are col-
lected. The two-dimensional method has been
tested on two-dimensional exposures of large-
scale trough cross-stratification in the Fountain
Formation at Manitou Springs, Colorado. These
troughs were previously considered largely
useless as paleocurrent direction indicators, even
though they are very common and three-di-
mensional exposures are rare.
ACKNOWLEDGMENTS
Method I was originated by Schwartz,
Method II by DeCelles and Langford. Field
work was supported in part by the Geological
Society of America, U.S. Geological Survey,
SOHIO Petroleum Company, Sigma Xi, Al-
legheny College, and the Indiana University
Geologic Field Station. D. E. Wuellner as-
sisted in the field and J. Brothers helped with
the mathematics of the computer program used
for Method II. W. Brewer provided key in-
sight, thought, and criticism regarding the
treatment of trough and fold geometry. We
gratefully acknowledge R. L. Slingerland, D.
Swift, and R. E. Hunter for reviewing the
manuscript. Finally, we thank L. J. Suttner for
encouragement, criticism, and helpful discus-
sions.
REFERENCES
ALLEN, J. R. L., 1963, The classification of cross-strat-
ified units with notes on their origin: Sedimentology,
v. 2, p. 95-114.
BHATrACHARYYA,A., 1969, Applicability of cross-beds in
determining paleocurrent direction: Quarterly Jour.
Geological, Mining and Metallurgical Soc. India, v. 41,
p. 45-49.
BLUCK,B. J., 1971, Sedimentation in the meandering River
Endfick: Scottish Jour. Geol., v. 3, p. 93-138.
COLEMAN,J. M., 1969, Brahmaputra River: channel pro-
cesses and sedimentation: Sed. Geology, v. 3, p. 129-
239.
DECELLES, P. G., 1982, Sediment dispersal directions in
the lower Kootenai Formation (Lower Cretaceous),
southwestern Montana: Geol. Soc. America Rocky Mtn.
Section Abstracts with Programs, v. 14, p. 309.
D o T , R. H., JR., 1970, Limitations in paleocurrent anal-
ysis of trough cross-stratification (abstract): Am. As-
soc. Petroleum Geologists Bull., v. 54, p. 844-845.
- - , 1973, Paleocurrent analysis of trough cross-
stratification: Jour. Sed. Petrology, v. 43, p. 779-783.
I~.AZER, D. E., AND OSAN[K,A., 1961, Point-bar Depos-
its, Old River Locksite, Louisiana: Trans. Gulf Coast
Assoc. Geol. Soc., v. 1 I, p. 121-137.
HAMBL~, W. H., 1961, Micro-cross lamination in Upper
Keweenawan sediments of northern Michigan: Jour. Sed.
Petrology, v. 31, p. 390-401.
642 PETER G. DeCELLES, RICHARD P. LANGFORD, AND ROBERT K. S C H W A R T Z
HARMS,J. C., MACKENZIE,D. B., ANDMcCuBBIN, D. S.,
1963, Stratification in modem sands of the Red River,
Louisiana: Jour. Geology, v. 71, p. 566-580.
HARMS, I. C., AND FAHNFESTOCK,R. K., 1965, Stratifi-
cation, bedforms, and flow phenomena (with an ex-
ample from the Rio Grande), in Middleton, G. V., ed.,
Primary Sedimentary Structures and Their Hydrody-
namic Interpretation: Soc. Econ. Paleontologists Min-
eralogists Spec. Pub. No. 12, p. 84-115.
HARMS, J. C., SOUTHARO,J. B., SPEARING, D. R., AND
WALKER, R. G., 1975, Depositional environments as
interpreted from primary sedimentary structures and
stratification sequences: Soc. Econ. Paleontologists
Mineralogists Short Course No. 2, 161 p.
HtGH, L. R., JR., AND PICARD,M. D., Sedimentary Struc-
tures of Ephemeral Streams: Amsterdam, London, New
York, Elsevier, 223 p.
HIGH, L. R., JR., AND PICARD, M. D., 1974, Reliability
of cross-stratification types as paleocurrent indicators in
fluvial rocks: Joar. Sed. Petrology, v. 44, p. 158-168.
JACKSON, R. G., 1976, Large-scale tipples of the lower
Wabash River: Sedimentology, v. 23, p. 593-623.
, 1980, Sedimentology of the Vermilion River (Il-
linois and Indiana, U.S.A.) and its significance for
lithofacies models of meandering alluvial streams, in
Fraser, G. S., ed., Sediments of the Modern Meander
Belt of the Vermilion River near Eugene, Indiana: Soc.
Econ. Paleontologists Mineralogists Great Lakes Sec-
tion Tenth Ann. Mtg. Guidebook, Danville, Illinois, p.
17-43.
LEVEY, R. A., 1978, Bedform distribution and internal
stratification of coarse-grained point bars, upper Con-
garee River, S.C., in Miall, A. D., ed., Fluvial Sed-
irnentology: Canadian Soc. Petroleum Geologists Mere.
5, p. 105-127.
McGowEN, J. H., AND GARNER, L. E., 1970, Physio-
graphic features and stratification types of coarse-grained
point bars: modem and ancient examples: Sedimentol-
ogy, v. 14, p. 77-111.
MECKEL, L. D., 1967, Tabular and trough cross-bedding
comparison of dip azimuth variability: Jour. Sed. Pe-
trology, 37, p. 80-86.
MIALL, A. D., 1974, Paleocurrent analysis of alluvial sed-
iments: a discussion of directional variance and vector
magnitude: Jour. Sed. Petrology, v. 44, p. 1174-1185.
MICHELSON, P. C., AND Do'l-r, R. H., JR., 1973, Orien-
tation analysis of trough cross-stratification in upper
Cambrian sandstones of western Wisconsin: Jour. Sed.
Petrology, v. 43, p. 784-794.
PICK, M. C., 1964, The stratigraphy and sedimentary lea-
tures of the Old Red Sandstone, Portishead coastal sec-
tion, north-east Somerset: Proc. Geol. Assoc., v. 75,
p. 199-221.
PRYOR, W. A., 1958, Dip direction indicator: Jour. Sed.
Petrology, v. 28, p. 230.
SCHWARTZ,D. E., 1978, Hydrology and current orienta-
tion analysis of a braided to meandering transition: the
Red River in Oklahoma and Texas, U.S.A., in Miall,
A. D., ed., Fluvial Sedimentology: Canadian Soc. Pe-
troleum Geologists Mere. 5, p. 231-256.
SCHWARTZ, R. K., in press, Broken foreland basin in
southwestern Montana: tectonism and sedimentation with
implications toward hydrocarbon occurrence, in The
Western Overthrust Belt: Geology and Hydrocarbon
Potential: Rocky Mountain Assoc. Geologists Sympos-
ium Volume.
SIMONS, D. B., RICHARDSON,E. V., AND NORDIN, C. F.,
JR., 1965, Sedimentary structures generated by flow in
alluvial channels, in Middleton, G. V., ed., Primary
Sedimentary Structures and Their Hydrodynamic Inter-
pretation: Soc. Econ. Paleontologists Mineralogists Spec.
Pub. No. 12, p. 34 52.
SLLNGERLAND,R. L., AND WILLIAMS,E. G., 1979, Paleo-
current analysis in light of trough cross-stratification
geometry: Jour. Geology, v. 87, p. 724-732.
SMITH, N. D., 1972, Some sedimentological aspects of
planar cross-stratification in a sandy braided river: Jour.
Sed. Petrology, v. 42, p. 624-634.
- - , 1974, Some sedimentological aspects of planar
cross-stratification in a sandy braided river, a reply to
N. L. Banks and J. D. Collinson: Jour. Sed. Petrology,
v. 44, p. 267-269.
- - , 1978, Some comments on terminology for bars
in shallow rivers, in Miall, A. D., eds., Fluvial Sedi-
mentology: Canadian Soc. Petroleum Geologists Mem.
5, p. 85-88.
STEWART, J. H., 1961, Origin of cross-strata in fluvial
sandstone layers in the Chinle Formation (Upper Trias-
sic) on the Colorado Plateau: U.S. Geological Survey
Prof. Paper 424-B, p. 127-129.
WtLUAMS, G. E., 1968, Formation of large-scale trough
cross-stratification in a fluvial environment: Jour. Sed.
Petrology, v. 38, p. 136-140.
- - , 1971, Flood deposits of the sand-bed ephemeral
streams of central Australia: Sedimentology, v. 17, p.
1-40.
WILLIAM, G. E., AND RUST, B. R., 1969, The sedimen-
tology of a braided river: Jour. Sed. Petrology, v. 39,
p. 649-679.