You are on page 1of 16

Conceptual model for predicting

mudstone dimensions in
sandy braided-river reservoirs

Ranie Lynds  Geology and Geophysics,


University of Wyoming, 1000 E University
Avenue, Department 3006, Laramie, Wyoming
82071; rmlynds@yahoo.com
Ranie M. Lynds received a B.S. degree in
geology and an M.S. degree in geophysics
from Stanford University in 1998 and a Ph.D.
in geology from the University of Wyoming
in 2005.

Ranie Lynds and Elizabeth Hajek

ABSTRACT
Sandy braided-river deposits with high net-to-gross sand ratios are
commonly attractive reservoirs, yet internal lithologic heterogeneities, particularly the presence of low-permeability mudstone deposits, significantly complicate the development of such units. Previous work has focused on measuring the scale and distribution of
mudstone deposits in outcrop analogs; however, because of extreme
differences in scale, discharge, sediment load, and geologic history,
the results of these studies are difficult to apply with confidence to
a wide range of sandy braided-river reservoirs. Based on work in
modern braided rivers (Niobrara and North Loup rivers, Nebraska)
and ancient braided-river deposits (Kayenta Formation, Jurassic
and lower Castlegate Sandstone, Cretaceous, Colorado and Utah),
we propose a process-based conceptual model for understanding
and predicting the distribution and geometries of fine-grained (mudstone) intervals in sandy braided-river deposits. This model is an
idealized channel-fill unit composed of five fine-grained lithofacies
(mud plugs, channel-lining muds, interbar muds, inclined heterolithic strata, and flood-plain and overbank material) that scale
proportional to channel-thread dimensions, including depth, crossstream width, and downstream length. Each lithofacies is found in a
different region in an individual channel fill, and lithofacies found
low in a fill may be preferentially preserved. Within braided-river
deposits, extrinsic depositional factors, such as aggradation rate,
available accommodation, and avulsion-return time, produce different channel-fill stacking arrangements, preserving fine-grained
lithofacies in different, relative proportions. This conceptual model
provides an approach to reservoir characterization that deductively
constrains the dimensions and distribution of fine-grained barriers
to flow and may help account for the inherent variability in sandy
braided-river deposits.

Copyright #2006. The American Association of Petroleum Geologists. All rights reserved.
Manuscript received March 17, 2005; provisional acceptance April 28, 2005; revised manuscript received
September 8, 2005; 2nd revised manuscript received January 16, 2006; final acceptance March 8, 2006.
DOI:10.1306/03080605051

AAPG Bulletin, v. 90, no. 8 (August 2006), pp. 1273 1288

AUTHORS

1273

Elizabeth Hajek  Geology and Geophysics,


University of Wyoming, 1000 E University
Avenue, Department 3006, Laramie, Wyoming
82071; ehajek@uwyo.edu
Elizabeth A. Hajek received a B.A. degree in
geology and geography from Macalester College in 1998, an M.S. degree in geology from
the University of Wyoming in 2005, and is
currently a geology Ph.D. student at the University of Wyoming.

ACKNOWLEDGEMENTS
We gratefully acknowledge support of this
work from ConocoPhillips, Groupement
Berkine (A Sonatrach-Anadarko Association),
and the National Science Foundation (EAR0345366 to P. Heller). In addition, this work
was partially supported by the Science and
Technology Centers program of the National
Science Foundation via the National Center
for Earth-Surface Dynamics under the Agreement Number EAR-0120914. We also received
funding from the AAPG Foundation Wiemer
Family Grant and Grant-in-Aid program, Geological Society of America Graduate Student
Research Grant, International Association of
Sedimentology Student Grant, Rocky Mountain
Section SEPM Fluvial Sedimentology Award,
and Four Corners Geological Society Masters
Thesis Grant. We thank Laura Vietti and Josh
Helke for help in the field and are grateful
to Paul Heller, Trevor Elliott, Andrew Miall,
David Morse, and Ernest Mancini for insightful
reviews of the manuscript.

INTRODUCTION
Sandy braided-river deposits are commonly described
as sheetlike (high width-to-thickness ratios), with good
porosity and permeability, and particularly high netto-gross ratios (e.g., Miall, 1977, 1978; Martin, 1993;
Collinson, 1996). These characteristics make them potentially excellent oil, gas, and groundwater reservoirs.
For example, the North Rankin field in offshore Australia is a gas-condensate field with porosities ranging
from 17 to 21%, permeabilities of 500 1500 md, and
net-to-gross ratios of 0.75 0.83 (Martin, 1993). Despite these desirable properties, internal heterogeneities in braided-fluvial deposits can significantly complicate reservoir development (e.g., Weber, 1982; Brayshaw
et al., 1996). Interpreting and modeling braided-fluvial
reservoirs rely on constraining the lateral dimensions
of the reservoir, porosities and permeabilities as a function of grain size, and internal reservoir architecture
(e.g., Bridge and Tye, 2000). Within an individual reservoir, some of these properties can be measured from
cores, wire-line logs, two- and three-dimensional seismic data, and flow between injector and producer
wells. However, it is difficult to confidently extrapolate such data to new reservoirs because individual reservoirs are as unique as the braided rivers that formed
them, varying greatly in terms of scale, discharge, sediment load, and geologic history.
Within sandy braided-fluvial reservoirs, permeability variations caused by fine-grained intervals are particularly challenging complications for reservoir characterization and development. These low-permeability
units can compartmentalize reservoirs and critically
impede fluid flow and sweep efficiency (e.g., Weber,
1982; Martin, 1993; Brayshaw et al., 1996). Unlike deposits of single-thread rivers, which typically consist of
channel-sandstone bodies encased in flood-plain mudstones (Allen, 1974), fine-grained material in braidedfluvial systems is commonly deposited within the active
channel, generating sandstone-dominated successions
with small mudstone intervals intercalated throughout
the deposit. Furthermore, although braided-river deposits are typically considered sand dominated, the abundance of muddy material present in modern braided
rivers and in ancient braided-river deposits can be
high. Consequently, much work has focused on understanding the geometry and distribution of impermeable and low-permeability mudstone bodies in sandy
braided-river deposits.
Previous studies have focused on measuring the
dimensions and dispersion of fine-grained intervals
1274

to infer the distribution of such units in similar reservoirs. For example, Robinson and McCabe (1997) measured the dimensions of shale and sandstone bodies
in the Upper Jurassic Salt Wash Member of the Morrison Formation and used these empirical outcrop measurements to define a hierarchy of shale and sandstone
bodies in the unit. Moreton et al. (2002) conducted
physical experiments to better understand permeability and architecture in braided deposits and obtained
data on the geometry, shape, and spatial distribution of
fine-grained facies in experimental stratigraphy. These
and other studies use a statistical approach to generate
probability functions based on observed mudstone geometries and spatial distributions (e.g., Collinson, 1978;
numerous studies summarized by Bridge and Tye,
2000). Outcrop analogs and empirical studies provide useful information for the first attempt at a reservoir model, but because of the wide variation between braided-river deposits, the results can be difficult
to apply with confidence to new reservoirs.
Based on our observations of modern braided rivers and ancient braided-river deposits, we propose a
process-based conceptual model for understanding and
ultimately predicting the distribution and geometries
of fine-grained deposits in braided successions. Unlike
models based on empirical data, this model may be
broadly applied to sandy braided-fluvial deposits and
can help to better constrain the scale and distribution of
mudstone components in a particular reservoir without
the standard requirement of large amounts of outcrop
and subsurface data. This model is composed of five
fine-grained lithofacies, each associated with a specific
depositional setting in a braided-river channel. The distribution of these lithofacies in ancient deposits is strongly
influenced by (1) the scale of the fluvial system (i.e.,
depth, width, and length scale of channel threads) and
(2) the preservation potential. Thus, the scale and preservation history of a fluvial system can be used to constrain stochastic reservoir models and predict the occurrence and dimensions of the fine-grained lithofacies.

STUDY AREAS AND METHODS


To study the processes controlling the accumulation
and preservation of fine-grained sediments in sandy
braided rivers, we examined two modern braided rivers and two ancient sand-dominated fluvial successions
widely interpreted as braided-river deposits; the Niobrara and North Loup rivers and Kayenta Formation and
lower Castlegate Sandstone, respectively (Figure 1). To

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

Figure 1. Location map of study areas. (A) Modern data were


collected from the Niobrara River in Niobrara State Park (NSP)
and North Loup River near the town of Taylor (T), both in
Nebraska. (B) Ancient braided-fluvial deposits studied include
the Lower Jurassic Kayenta Formation and the Upper Cretaceous lower Castlegate Sandstone (CGSS). The Kayenta Formation was deposited west of the Uncompaghre uplift and Ancestral
Rocky Mountains (UU/AR), with an interpreted extent indicated
by the medium-gray region. The Kayenta Formation has an average paleotransport direction to the southwest (Luttrell, 1993;
North and Taylor, 1996; Sanabria, 2001), indicated by an arrow.
Four localities in the Kayenta Formation were studied and include the Colorado National Monument, Colorado (CNM), and
Moab (M), the San Rafael Swell (SRS), and the Capitol Reef National Park (CR), Utah. The Upper Cretaceous Castlegate Sandstone crops out in the Book Cliffs (outlined in black) in central
Utah. The lower Castlegate Sandstone is a tongue of clastic
sediment shed from the Sevier orogenic belt (SOB) transported
generally eastward (indicated by an arrow). This unit grades
from fluvial near the Sevier orogenic belt to marginal marine and
deltaic farther east (indicated by dark to light shading above
the Book Cliffs outline). The lower Castlegate Sandstone was
studied near its type section just north of Helper, Utah (H).

understand how fine-grained sediments are deposited


and preserved in modern and ancient braided rivers,
we compared localities with different discharges, flow
depths, and source-area lithology.
Modern Rivers
As modern analogs, we studied two sandy braided rivers in eastern and central Nebraska, United States: the
Niobrara River and the North Loup River (Figure 1).
Observations and measurements made in modern rivers provide a framework for interpreting ancient deposits. Cross-stream bathymetry and current direction
were measured for each reach studied, and accumulations of fine-grained sediment were mapped and
sampled.
Niobrara River
The Niobrara River originates in east-central Wyoming,
flows eastward across northern Nebraska, and terminates at the Missouri River near the study area.
Sediment sources include Miocene and Pliocene continental deposits in western Nebraska and Cretaceous
Pierre Shale and Niobrara Chalk in eastern Nebraska
(Watkins and Diffendal, 1997). The Niobrara River
maintains year-round base flow (Bristow et al., 1999)
because it derives most of its discharge from ground-

water flow (Bleed, 1990). The mean annual discharge


of the Niobrara River from 1955 to 1957 was 48 m3/s
(1695 ft3/s) (U.S. Geological Survey Water Resources,
2005a). The median grain size of active bed load and
suspended load in the Niobrara River, collected 1 km
(0.62 mi) upstream from the confluence with the Missouri River, is 0.29 mm (0.011 in.; n = 13) and 0.19 mm
(0.0075 in.; n = 19), respectively. Within the studied
reach, channel-belt width ranges from 250 to 330 m
(820 to 1083 ft), and flow depth averages 21 cm (8.3 in.),
with a maximum flow depth of 120 cm (47.2 in.)
(Figure 2A, C).
At the study locality, the Niobrara River has three
to five channel threads in a flow-perpendicular bathymetric cross section (Figure 2C). Each thread has an
associated bathymetric low (thalweg) ranging from 55 to
120 cm (21.7 to 47.2 in.) and averaging 77 cm (30.3 in.)
in depth. Channel threads range in cross-stream width
from 10 to 20 m (32.8 to 65.6 ft). Fine-grained sediment is found in spatially consistent relationships with
dominant channel threads and active bars (Figure 2A).
North Loup River
The North Loup River originates in the Sand Hills of
central Nebraska, making it a particularly sandy braided
river. The river was studied near the town of Taylor,
Nebraska (Figure 1). The mean annual discharge at the
Lynds and Hajek

1275

Figure 2. (A) Aerial photograph of the Niobrara River study area upstream of its confluence with the Missouri River in the Niobrara
State Park, Nebraska, where flow is to the north (large arrow); photo from USDA (2003) Web site. (B) Aerial photograph of the North
Loup River study area north of Taylor, Nebraska, where flow is generally to the southeast (large arrow); photo from USDA (2003)
Web site. Channel threads are noted in both photographs, and the channel belts are the full width of the rivers. White arrows indicate
some areas of active mud deposition. Cross-stream bathymetric profiles of the Niobrara (C) and North Loup (D) rivers show scaling
in each system. The profiles were constructed by measuring the flow depth along a transect perpendicular to the banks at 1-m (3.3-ft)
intervals, but do not directly correlate to the photographs because they are not time synchronous (profiles were measured in August
2004). The stippled pattern represents the channel bottom and underlying sediments, and black arrows denote the bathymetric lows
that correspond to dominant channel threads. Bar macroforms, both active and inactive, are the associated bathymetric highs. Most
barforms are compound bars with narrow second-order cross-bar channels (as described by Bridge, 1993). The channel threads are
the future sites of channel-thread fills. Note the difference in scale in the two transects. In this transect of the Niobrara River,
maximum channel-thread flow depths range from 55 to 120 cm (21.7 to 24.7 in.) and average 77 cm (30.3 in.). Channel-thread
widths range from 10 to 20 m (32.8 to 65.6 ft). Within the transect across the North Loup River, channel threads range from 5 to 10 m
(16.4 to 32.8 ft) in width with maximum flow depths ranging from 35 to 38 cm (13.8 to 15.0 in.), averaging 37 cm (14.6 in.). Two
profiles were measured on the Niobrara River, and six were taken from the North Loup River; the transects shown here are two of
several and can be considered representative.

study location from 1937 to 2002 is 14 m3/s (494 ft3/s)


(U.S. Geological Survey Water Resources, 2005b). The
median active bed-load and suspended-load grain sizes
are 0.33 mm (0.013 in.; n = 6) and 0.22 mm (0.0087 in.;
1276

n = 9), respectively. At this location, the river is 110 m


(360.9 ft) wide, with a maximum flow depth of 89 cm
(35.0 in.) and average flow depth of 15 cm (5.9 in.)
(Figure 2B, D).

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

Along the studied reach, the North Loup has three


to four channel threads per cross-stream transect
(Figure 2D). Maximum channel-thread depth ranges
from 26 to 89 cm (10.2 to 35.0 in.) and averages 34 cm
(13.4 in.). Channel threads are commonly between 5
and 10 m (16.4 and 32.8 ft) wide. As with the Niobrara
River, fine-grained sediment is deposited in spatial
association with dominant channel threads and active
bars (Figure 2B).
Ancient Deposits
We chose the Kayenta Formation and lower Castlegate Sandstone, Colorado and Utah, United States, as
ancient reservoir analogs (Figure 1). These units have
been previously interpreted as representing sandy braided
rivers primarily because they possess the typical characteristics associated with sandy braided-river deposits
(summarized by Miall, 1994); they are sand dominated
with numerous closely spaced scour surfaces and a tight
distribution of paleocurrent directions. Although this
evidence is commonly used to define ancient braidedriver deposits, we note that it is not a direct proof of
braiding because the number of active channel threads
at any given time cannot be determined from the rock
record. It is certain, however, that these are sand-bedded
river deposits that were not deposited by mud-rich
avulsion-dominated meandering rivers associated with
isolated, sandy channel fills encased in fine-grained floodplain deposits (e.g., Allen, 1974).
Kayenta Formation
The Kayenta Formation is a member of the Lower Jurassic Glen Canyon Group (Pipiringos and OSullivan,
1978), which was deposited in a topographic depression between eroded remnants of the Ancestral Rocky
Mountains to the east and the evolving continental arc
farther west (Peterson, 1986; Marzolf, 1988; Blakey,
1989). The basal eolian Wingate Sandstone grades up
into widespread sandy braided-river deposits of the
Kayenta Formation. Paleocurrent analyses show that
the Kayenta Formation was derived in the east and
gradually becomes finer grained to the west-southwest
(Luttrell, 1993; Sanabria, 2001), transporting sediment
from the Ancestral Rocky Mountains region and potentially reworking the eolian dunes of the Wingate
Sandstone. Thickness of the Kayenta Formation is typically 100200 m (328656 ft) (Harshbarger et al.,
1957) and the top grades, commonly rapidly, into the
eolian sand sea deposits of the Navajo Sandstone (Middleton and Blakey, 1983; Tuesink, 1989; Herries, 1993).

Previous work done on sedimentary environments of


the Kayenta Formation has interpreted the unit to be
the product of low-sinuosity, sandy braided-river deposition, both ephemeral and perennial, in our study areas
in the eastern part of the basin (Miall, 1988; Bromley,
1991; Luttrell, 1993; North and Taylor, 1996).
The Kayenta Formation was analyzed using measured sections and photomosaic panels from the Colorado National Monument, Colorado, and near Moab,
the San Rafael Swell, and Capitol Reef National Park,
Utah (Figure 1). For comparisons within this study
and to account for potential problems associated with
downstream fining, all grain size and dimensional information for the Kayenta Formation is taken only from
the section exposed at the Colorado National Monument.
Within the Kayenta Formation, channel scours several tens of meters (less than 30 m [100 ft]) wide have
been reported (Stephens, 1994) and bar-scale crossbeds, or clinoforms, with both upstream and downstream accretion are well preserved and suggest paleoflow depths that average approximately 1 m (3.3 ft).
Typical channel fills in the Kayenta Formation include
coarse-grained (medium sand to, uncommonly, gravel)
clinoforms (interpreted as barforms), fine- to very finegrained planar laminations (planar bedding), fine- to
medium-grained, inclined, and trough cross-beds (dune
foresets and troughs), silty climbing ripples (ripples),
and massive mudstone (slackwaters).
Castlegate Sandstone
The Castlegate Sandstone (Campanian) is part of the
Mesaverde Group (Upper Cretaceous) in the Book Cliffs
of east-central Utah. The Mesaverde Group includes
marine and terrestrial sedimentary rocks deposited in
a foreland basin during the Sevier orogeny (Spieker,
1946; Van de Graaff, 1972; Fouch et al., 1983; Lawton, 1985). The Castlegate Sandstone has been interpreted as a progradational clastic wedge ranging from
fully fluvial in the proximal part of the basin to marginal marine in the distal portion. Our study focuses on
the type section for the Castlegate Sandstone exposed
near Helper, Utah (Figure 1), where the Castlegate
Sandstone comprises sandy fluvial deposits. The basal,
cliff-forming 80 m (262.5 ft) of the Castlegate Sandstone (informally designated the lower Castlegate Sandstone, Fouch et al., 1983) has been interpreted as
braided-fluvial deposits (e.g., Fouch et al., 1983; Miall,
1993; Olsen et al., 1995), representing approximately
2 m.y. of deposition (Olsen et al., 1995; Robinson and
Slingerland, 1998). Fouch et al. (1983) describe the depositional environment of the lower Castlegate Sandstone
Lynds and Hajek

1277

as a broad braid plain transporting recycled passivemargin sedimentary rocks shed eastward from the Sevier orogenic belt into the Western Interior seaway.
Detailed lithologic information was collected in measured sections through the lower Castlegate Sandstone,
and reservoir architecture was obtained from highresolution ground-based light detection and ranging
(LIDAR) images of cliff-face surface topography.
Based on fully preserved clinoforms, interpreted as
barform accretion sets, typical paleoflow depths of lower
Castlegate rivers ranged from 1.5 to more than 8 m (4.9
to more than 26.2 ft), with average depth of 4.1 m
(13.4 ft), and channel scours are typically several tens to
hundreds of meters in width (Miall, 1993; this study).
Individual channel-fill deposits generally include a basal
scour overlain by trough cross-bedded or soft-sedimentdeformed, fine- to lower medium-grained sandstone
(interpreted as a basal thalweg), fine-grained sandstone
clinoforms with tabular cross-beds, and ripple crosslaminae (barforms), capped by rippled siltstone and mudstone (slackwaters) and/or structureless mudstone and
claystone (overbanks). Some authors (e.g., Olsen et al.
1995; Yoshida et al., 1996; McClaurin and Steel, 2000)
have suggested that the lower Castlegate Sandstone represents a time of low accommodation as evidenced by
apparently truncated channel-fill deposits. However,
fully preserved clinoforms topped with structureless
mudstone and claystone can be routinely observed
throughout the entire section (Hajek and Heller, 2004).

STUDY AREA OBSERVATIONS


The cross sectional profiles from the Niobrara and North
Loup rivers (Figure 2C, D, respectively) show main channel threads (i.e., major bathymetric lows or thalwegs)
and bar macroforms (bathymetric highs). These dominant channel threads represent the second-order channels of Bristow (1987) or the first-order channels of
Bridge (1993). The width and depth of channel threads
are likely related to bar-forming (Bridge, 2003) or dominant (Thorne et al., 1993) discharges; thus, the number
and dimensions of channel threads vary among braided
rivers (e.g., Figure 2B, D). Principal channel threads are
probable sites for preservation of future bar deposits
because they occupy the topographically low areas that
have the highest preservation potential (e.g., Paola,
1989). The bathymetric highs in Figure 2C and D can be
considered amalgamated or compound bars (defined
by Bridge, 1993) that are significantly wider than associated channel threads. Small channel threads traverse
compound bars (Figure 2C, D); however, because of
their size and position high in the flow, these have relatively low preservation potential.
Figure 3A and B show channel-fill units representative of those typically found in the Kayenta Formation and lower Castlegate Sandstone, respectively. We
interpret these deposits as individual channel-thread
(and not channel-belt) fills, based on our observations
in modern rivers. It is important to stress the distinction

Figure 3. (A) Kayenta Formation channel fill (above) at the Colorado National Monument, Colorado (Figure 1), with definition sketch
(below). Photograph shows scoured base with bar macroforms accreting toward the left and maximum barform height of 1.2 m
(3.9 ft). (B) LIDAR surface topography image of lower Castlegate Sandstone channel fill north of Helper, Utah (Figure 1), with definition
sketch (below). Channel fill has scoured base with right-accreting bar macroforms, approximately 5 m (16.4 ft) high. Bar macroforms
showing evidence of rollover at the top are interpreted as fully preserved barforms, indicative of local paleoflow depth. Note the
similarity in geometry despite the difference in scale between the Kayenta Formation (A) and lower Castlegate Sandstone (B).
1278

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

Table 1. Median Grain Sizes* for Bed Load and Four Fine-Grained Lithofacies in the Modern Niobrara and North Loup Rivers and
the Kayenta Formation**

Bed load
Channel-lining muds
Interbar muds
IHS muds
Mud plugs

North Loup River

Niobrara River

Kayenta Formation

0.35 (1.4e 2)

0.18 (7.0e 3)
0.063 (2.5e 3)

0.25 (9.8e3)

0.13 (4.9e3)
0.031 (1.2e 3)

0.18 (7.0e 3)
0.0055 (2.2e 4)
0.074 (2.9e3)

0.037 (1.5e2)

*In millimeters (and inches).


**Dashes indicate no data. A consistent relationship exists between maximum grain size (bed load) and each fine-grained lithofacies in the study areas. This
relationship is also consistent with qualitative observations made in the lower Castlegate Sandstone. All samples were processed for grain size using a Horiba LA300 laser particle-size analyzer (LPSA) and Retsch Technology Digital, image-processing particle-size analyzer (Camsizer). The LPSA measures the fine-grained
fraction ( < 0.09 mm; < 0.0035 in.), and the Camsizer measures the coarse-grained fraction ( > 0.09 mm; > 0.0035 in.). Samples were not analyzed for the
Castlegate Sandstone.

between channel threads and channel belts; active deposition and erosion processes are not occurring collectively across an entire channel belt, but are localized
within active channel threads. Therefore, within ancient
braided-river deposits, it is possible to constrain local
dimensional information in an individual channel thread
without having to identify the width of the full channel belt. We suggest that individual channel-fill deposits
generated by barforms can be considered discrete depositional units in the modern braided-river deposits and
are easily identifiable in ancient braided-river deposits.

FINE-GRAINED LITHOFACIES
The finest grained, least permeable lithofacies in sandy
braided rivers are uniquely defined by the specific subenvironment in which they are deposited. Each lithofacies (mud plugs, channel-lining muds, interbar muds,
inclined heterolithic strata, and flood-plain and overbank material) is associated with distinct grain sizes
(Table 1) and geometry normalized by channel-thread
dimensions (length, width, and thickness; Table 2) that

vary predictably within a particular fluvial system. Numerous studies have classified these fine-grained lithofacies collectively as slackwater or fine-grained channelfill deposits (e.g., Cant and Walker, 1976; Miall, 1988).
However, our results suggest that it may be useful to
distinguish between the different types of fine-grained
lithofacies and classify them based on the distinct processes and locations of origin.
We stress that paleoflow depths are the principal
measurements to be taken when studying fluvial deposits because they enable scaling comparisons between
fluvial systems of different sizes. At the outcrop scale,
flow depths are best approximated by fully preserved
bar clinoforms or barforms (e.g., Bridge and Tye, 2000;
Mohrig et al., 2000), formed under dominant discharge
(Thorne et al., 1993) or channel-forming discharge
(Bridge, 2003). The complete height of a barform is
preserved if the top shows rollover of the clinoform as seen
at the crests of modern river bars, such as in Figure 3A.
Whereas complete barform height is the most accurate
measurement of local paleoflow depth (Bridge and
Tye, 2000; Mohrig et al., 2000), it can be difficult to
measure clinoform height from cores or wire-line logs,

Table 2. Channel-Thread Scaling of Fine-Grained Lithofacies*

Channel-lining muds
Interbar muds
IHS muds
Mud plugs

Maximum Thickness

Maximum Width

Maximum Length

Maximum scour depth


Local flow depth
Local flow depth
Maximum scour depth

Thread width
Thread width
Bar-front width
Thread width

Thread length
Downstream bar spacing
Local bar height and flow depth
Thread length

*The maximum thickness, width, and length dimensions of channel-lining muds, interbar muds, IHS (inclined heterolithic stratification), and mud plugs are set by
aspects of the active channel thread in which they are deposited. Overbank deposits (not listed in this table) are not deposited within active channels and do not
scale directly with channel dimensions.

Lynds and Hajek

1279

Figure 4. (A) Mud plug (outlined) from the Kayenta Formation at the Colorado National Monument, Colorado. The paleoflow depth averages 1 m (3.3 ft) in this locality, and the mud plug
is 65 cm (25.6 in.) thick. (B) Representative mud plug (outlined)
from the lower Castlegate Sandstone near Helper, Utah. The
mud plug is 1.8 m (5.9 ft) high; in this part of the lower Castlegate Sandstone, average paleoflow depth is 2.5 m (8.2 ft).

and only exceptional reflection seismic data clearly show


clinoform features. In such cases where clinoform height
cannot be determined from cores, paleoflow depth can
be estimated, although with more uncertainty, from dune
cross-bed set height (e.g., Paola and Borgman, 1991;
Bridge, 1997; Leclair et al., 1997). Paleoflow depth is
a very important outcrop measurement because local
flow depth directly affects the deposition and preservation of fine-grained sediment in a braided river.

Abandonment Facies: Mud Plugs and Channel-Lining Muds


We recognize two types of abandonment facies: mud
plugs and channel-lining muds. These types of finegrained deposits are the result of channel-thread aban-

donment or avulsion. Avulsions and abandonment in


braided rivers occur over days to tens of years (e.g.,
Smith et al., 1998; Bristow et al., 1999; Skelly et al.,
2003). During abandonment, a channel thread carries
less water, resulting in a decrease in local shear stress,
leading to deposition of fine-grained suspended load.
This sediment fills the abandoned channel thread, producing muddy deposits with geometries that approximate the shape of the original channel thread.
Mud plugs are commonly described in fluvial deposits and are commonly recognized by their characteristic channel shape (Figure 4). Mud plugs are always
finer grained than typical bed load for a river because
abandoned channel threads tend to trap the finest grained
component of the total sediment load, which, in sandy
rivers, commonly consists of very fine sand, silt, and
clay (Table 1). Mud plugs are typically structureless
when composed of clays and contain climbing ripples
when composed of silt and/or fine sand. In addition,
they commonly contain plant fragments (transported
organics). Because mud-plug deposits form in abandoned channel-thread reaches, their thickness cannot
exceed the maximum scour depth of the particular
fluvial system, which is controlled by flow depths. Likewise, the cross-stream length of mud plugs cannot exceed the width of a channel thread. The downstream
length of mud plugs is based on the length of the abandoned channel threads and is thus somewhat more variable (Table 2).
Channel-lining muds (Figure 5) are uncommonly
noted in fluvial deposits, but can be found throughout much of the Kayenta Formation. Channel-lining
muds are similar to mud plug deposits and form on
channel-thread abandonment from suspension settling.
Channel-lining muds differ from mud plugs only in
that instead of representing complete abandonment of
a channel thread (suggested by complete preservation
of a channel-shaped mud plug), they are formed when
muds accumulate in temporarily abandoned reaches

Figure 5. Channel-lining mud in Kayenta Formation at the Colorado National Monument, Colorado. The dark layer of mud lines the
entire channel-thread scour. Channel thread is approximately 18 m (59.1 ft) wide and was subsequently filled by sandy barforms
accreting to the left.
1280

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

that are later reoccupied by active channel threads. This


reoccupation excavates some of the fine-grained material, leaving behind a relatively thin layer of mudstone
below a sandy channel fill (Figure 5). Grain sizes range
from clay to silt (Table 1), and sedimentary structures
are similar to those found in mud plugs. Channel-lining
muds are typically thin, on the order of tens of centimeters (a few inches), and their thickness is a function
of the duration of abandonment and amount of scour
upon reoccupation. The maximum width of channellining muds scales proportionally to the local width of
channel threads, and the length scale is related to the
length of a channel-thread segment (Table 2).
Mud plugs and channel-lining muds represent two
different avulsion and abandonment scenarios in braided
rivers. Sometimes, reaches within a channel belt are abandoned and later reoccupied, generating channel-lining
muds. In other cases, channel threads are completely
abandoned, evidenced stratigraphically by a fully preserved mud-plug deposit capped by overbank and floodplain deposits. Abandonment facies are deposited within an active channel belt and are therefore limited in
scale by the stochastic range of channel-thread scales
inherent to a particular system. For example, fully preserved mud plugs measured in the Kayenta Formation
range in thickness from 55 to 141 cm (1.8 to 4.6 ft), and
measured paleoflow depths range from 50 to 170 cm
(1.6 to 5.6 ft).
Slackwater Deposits: Inclined Heterolithic Strata
and Interbar Muds
In the Niobrara and North Loup rivers, active midchannel bars regularly build to the water surface, effectively sheltering the lee side of the bar (similar to that
interpreted in the Jamuna River, Bangladesh, by Best
et al., 2003). In this protected or slackwater area, suspended load is commonly deposited, yielding interbar
muds (Figure 6) and inclined heterolithic strata (IHS).
Interbar muds consist of transported plant fragments and very fine-grained sand, silt, and clay (Table 1).
Sedimentary structures seldom occur in these deposits
because velocities in the slackwater areas are nearly
zero. However, low-velocity currents near the channel
bed sometimes rework fine-grained interbar deposits
into trains of climbing ripples. It is not uncommon for
interbar muds to interfinger with bed load at the downstream toes of bars where current velocities increase
as flow confluence is encountered downstream of the
bar. Interbar muds are frequently deposited immediately downstream (<1 m;<3.3 ft) of the crest of actively

Figure 6. (A) Active interbar mud deposition in the North Loup


River near Taylor, Nebraska. The sandy barform on bottom and
left side of the photograph is migrating downstream to the
upper right. The lee side of the bar is sheltered from strong
currents, resulting in extremely low to zero flow velocities and
the deposition of the suspended load into the interbar area
(1.7-m [5.5-ft]-tall person is standing in center of interbar area).
(B) Interpreted interbar mudstone (outlined in black) in the
Kayenta Formation at the Colorado National Monument, Colorado. This interbar mudstone is approximately 40 cm (15.7 in.)
thick, and local paleoflow depth is approximately 1 m (3.3 ft).
Paleocurrent indicators on the associated barform suggest that
the flow direction was into the cliff face.

migrating barforms, resulting in an abrupt downstream


drop in grain size. To accumulate, interbar muds must
be protected by an existing barform. Consequently,
the local flow depth limits the maximum thickness of
interbar mud accumulations. In our study areas, interbar muds were commonly only one-half to two-thirds
the local flow depth in thickness. The downstream
length and cross-stream spacing of bars (i.e., channelthread spacing) set the maximum length and width of
the interbar muds, respectively (Table 2).
Dipping, alternating layers of sand and mud (inclined heterolithic stratification or IHS) are found in
modern braided-river deposits and in the lower Castlegate Sandstone. These deposits represent locally fluctuating flow conditions and have been documented in
several low-sinuosity and multithread river settings
Lynds and Hajek

1281

(Thomas et al., 1987). Inclined heterolithic strata can


form on laterally accreting bars (Thomas et al., 1987),
as well as when intermittent bed-load sedimentation occurs on the lee sides of midchannel bars. Inclined heterolithic strata are sometimes considered tidal indicators
in fluvial settings (e.g., Yoshida et al., 1996; Yoshida,
2000). Yoshida (2000) presents evidence for tidal influence in the distal parts of the Castlegate Sandstone near
Green River, Utah; however, no tidal indicators have
been documented in our study area near Helper, Utah. In
the absence of other tidal indicators in the study area
and because they are similar to IHS observed in modern
rivers, we believe IHS in the lower Castlegate Sandstone
exposed in our study area are wholly fluvial in origin.
In our study areas, the coarse-grained component
of IHS is generally composed of alternating layers of
ripple-laminated, very fine-grained sand or silt and finegrained intervals of clay and/or silt (Table 1). These
fine-grained intervals are commonly structureless, with
abundant carbonaceous material and plant fragments.
Inclined heterolithic strata sets typically range in thickness from 5 to 20 cm (2.0 to 8.9 in.), with the coarser
strata thickening downdip into thalweg deposits. As
with interbar mud deposits, maximum IHS thickness is
limited by local flow depth. Maximum downdip length
and cross-stream width are set by bar height and width,
respectively (Table 2).

Interbar muds and IHS are formed within active


channel threads in active channel belts. Because they
are deposited near and/or on active bar fronts, their
thickness is limited by local flow depth. In general,
slackwater deposits in the Kayenta Formation (mean
paleoflow depth = 0.7 m or 2.3 ft) tend to be much
smaller than those measured in the Castlegate Sandstone (mean paleoflow depth = 4.1 m or 13.4 ft).
However, when slackwater deposit thickness is normalized by local paleoflow depth, we observed an average ratio (slackwater-deposit thickness to local paleoflow depth) of 0.60 in these systems. More data are
needed to validate this relationship.
Overbank and Flood Plain
Overbank deposits represent flood plains formed
adjacent to the active channel belt (Figure 7). These
deposits consist of fine-grained sand, silts, and clays
and, in exceptional cases, can locally represent more
than 40% by volume of the (interpreted) braided-river
deposit (Bentham et al., 1993). Overbank units can be
rippled or massive and commonly contain mud cracks,
root casts, and occasional paleosols. Unlike other finegrained lithofacies, overbank deposits do not scale in
thickness with the averaged flow depth of the channel.
Instead, thickness of overbank mudstones is primarily

Figure 7. (A) Overbank and


flood plain with weak paleosol
development in the Kayenta
Formation near Moab, Utah.
The base of the overbank is not
exposed and does not show
a full thickness. (B) Overbank
and flood-plain deposits from
the lower Castlegate Sandstone
near Helper, Utah. The flood
plains are not laterally extensive
because of erosional truncation.

1282

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

Figure 8. Idealized channel-fill unit model that provides a conceptual basis for understanding the depositional relationships
and preservation potential of the five fine-grained lithofacies in
one individual channel-fill deposit, which can be scaled to paleoflow depth (pfd). (A) channel-lining muds, (B) interbar muds,
(C) inclined heterolithic strata, (D) mud plugs, and (E) flood-plain
and overbank deposits.

a function of proximity to the active channel belt


(closer = thicker). In the Kayenta Formation and lower
Castlegate Sandstone, we observed overbank deposits
no greater than three flow depths in thickness. Lateral
thickness is highly variable, but commonly not more
than several tens of channels threads in width.

MODEL
Based on our observations, we suggest that consistent
relationships exist between fine-grained lithofacies and
flow depth in a fluvial system. We propose a processbased conceptual model for understanding and predicting the geometry and distribution of fine-grained
deposits in sandy braided-river reservoirs. Five finegrained lithofacies are incorporated into a generic
channel-fill unit (Figure 8) that can be tailored to include subtleties of individual reservoirs. This model
provides a systematic basis for understanding the threedimensional distribution of low-permeability units and
populating subseismic-scale reservoir models without
requiring large amounts of outcrop data. Instead of relying on frequency distributions of, for example, widthto-thickness ratios, dimensional information about finegrained deposits in a reservoir may be extrapolated by
constraining the range of flow depths.
Idealized Channel-Fill Unit
An idealized channel-fill unit summarizes the depositional relationships of all potential fine-grained deposits in a sandy braided river (Figure 8). Each lithofacies is associated with a distinctive location in the
channel-fill unit. Channel-lining muds line the channelfill unit and represent the basal fine-grained deposit.

Interbar muds correspond to the next phases of filling


and are found low in the deposit. Mud-plug deposits,
which form during later stages of an individual channelfill event, are found near the top of the channel-fill unit.
Last, flood-plain muds are the final capping lithofacies of
the idealized channel-fill unit. A particular flood-plain
deposit is not genetically related to the individual channel fill it overlies (because flood-plain sediments are laterally, not vertically, associated with a channel belt), but
it represents the last possible lithofacies that can be deposited before the next channel-filling event. The relative position of each lithofacies also reflects the order
of deposition in a channel fill. For instance, within one
single channel-filling event, an interbar mud cannot be
deposited after the channel thread is abandoned (resulting in a mud plug).
Note that the channel-fill unit is idealized and does
not correspond to any specific modern river or ancient
outcrop. Figure 3A and B are examples of channel fills
in the Kayenta Formation and lower Castlegate Sandstone, respectively. To interpret these channel fills in
the framework of the idealized model, it is necessary
to tailor the generic channel-fill unit to the specific fluvial system. In the case of the Kayenta Formation, IHS
is uncommon and, therefore, need not be included in
the modeled channel-fill unit. Likewise, channel-lining
muds are unusual in the lower Castlegate Sandstone
and should not be included in a model of this system.
The channel unit provides a generic model that must
be modified to fit the depositional organization of a
specific reservoir.

DISCUSSION
The deposition and preservation of muddy intervals
in sandy braided-river systems are not random. Active fluvial processes control the accumulation of finegrained components in channel systems, and long-term
responses of a river influence the preservation of finegrained deposits. Within the framework of the idealized channel-fill unit model, depositional and preservation relationships can be evaluated in a specific
sandy braided-river reservoir and compared between
reservoirs.
Scaling the Model
We have emphasized that the dimensions of mudstone deposits, in most cases, are proportional to
the size (flow depth, width, and length) of individual
Lynds and Hajek

1283

channel threads in braided rivers. As a result, the


idealized channel-fill unit can be scaled to fit different
reservoirs once the size range of channel threads is
determined. This approach can potentially constrain
stochastic reservoir models better than using large
amounts of empirical data obtained from unrelated
reservoirs and outcrops. For example, obtaining dimensional data from the lower Castlegate Sandstone
outcrops might seem an appropriate statistical basis
for estimating mudstone distribution in a reservoir
model of the Kayenta Formation. However, lower Castlegate Sandstone river channel threads were, on average, almost 3 m (9.8 ft) deeper (based on observed
average paleoflow depths of 4.1 m [13.4 ft]) than those
of Kayenta Formation rivers, which were typically
less than 1 m (3.3 ft) deep. Applying dimensional data
from the lower Castlegate Sandstone would result in
overestimating the fine-grained lithofacies dimensions
and sand-body thicknesses by a factor of 4 in a Kayenta
Formation reservoir model. Without scaling empirical
data to fluvial processes, there is no means of evaluating the applicability of a data set to a particular reservoir. Alternatively, the channel-fill unit model can be
used to constrain the maximum dimensions of finegrained sediment in a reservoir based on in-situ channelthread measurements.
To best apply the channel-fill unit model, the range
of paleoflow depths, and if possible, channel-thread
cross-stream widths and downstream lengths should be
well constrained within a particular deposit. Paleoflow
depth can be estimated from core using cross-bed set
thickness (e.g., Paola and Borgman, 1991; Bridge, 1997;
Leclair et al., 1997), but such data are commonly difficult to obtain. In the absence of direct bar clinoform or
cross-bed set measurements, channel-fill units themselves can be used to constrain the range of channelthread dimensions in a reservoir. By carefully identifying the succession of fine- and coarse-grained facies in
a core, it is possible to identify complete and partial
channel-fill units. These observations provide information about the range of channel-fill thicknesses in a
reservoir (which are a function of paleoflow depths in
a fluvial system) and the degree of preservation characteristic of the particular system (see the section on
Stacking the Model).
With the present knowledge, channel-thread width
and downstream length can only be obtained from
outcrop. Numerous empirical studies relate channel
width to depth in modern rivers (e.g., Leopold and
Wolman, 1957; Schumm, 1960; Williams, 1986) and
ancient examples (e.g., Robinson and McCabe, 1997,
1284

1998). These may provide a rough constraint on channel width based on flow-depth estimates. However,
there are drawbacks to using these empirical relationships for quantitative reservoir models. The definition of a channel varies between studies. In most
modern studies, river width is taken to be the width
of the entire active channel belt (first-order channels of Bristow, 1987). Within sand-dominated deposits, it is difficult to define the margins of an entire
channel belt in outcrop; consequently, many ancient
channel width measurements are made from individual channel-thread scours. Additionally, the relationship between flow depth and channel width is empirical and not based on a defined process. This makes
it difficult to apply consistently across widely variable
braided-river systems.
Although river processes such as flow depth clearly
define maximum dimensions for fine-grained lithofacies (e.g., interbar muds cannot be thicker than local
flow depth), more data are needed to understand typical relative thicknesses of various fine-grained lithofacies. For example, evidence from our study suggests
that interbar mud thicknesses normally reach onehalf to two-thirds the height of local flow depth
and uncommonly reach the maximum possible thickness of a full flow depth. The channel-fill unit model
defines the maximum possible thickness of all finegrained lithofacies (except overbank muds) as one
flow depth, but more data are needed to establish robust stochastic ranges for the geometry of each lithofacies. Dimensional data from studies of reservoirs,
outcrops, or experimental deposits (e.g., Robinson and
McCabe, 1997; Moreton et al., 2002) are extremely important for characterizing this relationship, but need
to be defined relative to paleochannel-thread dimensions. We suggest that geometric data from lithofacies associated with all types of fluvial deposits should
be collected relative to channel and/or channel-thread
dimensions.
We also note that there is an apparent scaling relationship between bed-load particle diameter and the
grain size of each distinct fine-grained lithofacies in
the individual braided-river system (Table 1). Not only
are the fine-grained deposits predictably finer than the
bed load; our preliminary analysis suggests a systematic relative relationship between the grain size of each
lithofacies. In this study, the finest grained lithofacies
are always IHS muds, followed by mud plugs and interbar muds. This potential relationship is consistent
within our study areas, but needs to be verified with
more data.

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

Stacking the Model


Individual channel-fill units represent discrete depositional units, but can be stacked or considered in groups
to help understand the implications of long-term depositional conditions on the distribution of fine-grained
lithofacies and, thus, reservoir quality. On the time
scale of the recurrence interval for flood events (years to
hundreds of years), new channel deposits can be generated, and older ones can be completely removed. On
the time scales required to generate reservoir-scale deposits (tens of thousands to millions of years), three
primary factors influence the preservation of channelfill deposits: aggradation rates, available accommodation, and avulsion-return time (Bristow and Best, 1993).
Increased preservation of channel-fill deposits is associated with high aggradation rates, high accommodation,
and long avulsion-return times. The interplay of these
longer term aspects of fluvial sedimentation affects how
much of the older channel deposits are removed by the
active channel and, thus, how much of the older, underlying deposits are preserved. For example, in a setting with high aggradation rates but low relative accommodation, such as the active Niobrara River (Skelly
et al., 2003), available accommodation has been filled
because of high sedimentation rates, and the tops of
barforms are repeatedly truncated, preferentially preserving the lower parts of channel fills (e.g., Figure 9A).
Conversely, even in a river with low aggradation rates,
full channel-fill units may be preserved if either avulsionreturn times are long or accommodation is high (e.g.,
Figure 9B). Mud plugs and overbank deposits are routinely preserved in the Kayenta Formation and lower
Castlegate Sandstone, suggesting that during deposition, these systems had high aggradation rates, high
accommodation, and/or long avulsion-return times. Further study is needed to determine the relative interaction of these factors that enabled complete preservation of channel-fill units in each formation.
For an individual braided-river system, regional
history, as it relates to aggradation, accommodation,
and avulsion-return time, directly influences the type
and distribution of fine-grained lithofacies preserved
in the deposit. This is an important factor in reservoir
modeling. For example, overbank deposits can be laterally continuous, impermeable units that significantly impede vertical flow in a reservoir. However, they
are deposited over the top of individual channel fills,
making them more susceptible to removal by erosion in
low aggradation, low accommodation, or rapidly avulsing fluvial systems. Figure 9C conceptually illustrates

Figure 9. Hypothetical stacking scenarios of idealized channelfill units. (A) Idealized channel-fill units stacked assuming either
low aggradation, low accommodation, or high avulsion-return
rate. For this case, subsequent channels consistently remove the
upper parts of the channel-fill units, which include the upperparts
of IHS muds, mud plugs, and overbank deposits. (B) Idealized
channel-fill units stacked assuming either high aggradation, high
accommodation, or low avulsion-return rate. In this instance, the
upper part of the channel fill is more commonly preserved than
example (A). (C) Bar graph showing the relative abundance of
each fine-grained lithofacies for cases A and B. This hypothetical
example illustrates how varied depositional conditions (including
aggradation, accommodation, and avulsion history) can significantly affect the types and relative abundance of fine-grained
lithofacies preserved within a reservoir.

the effect of high aggradation rates versus low aggradation rates (assuming all else are equal) on hypothetical reservoir deposits. In this hypothetical example,
flood plains are uncommonly preserved in the lowaggradation case (representing only 3% of total finegrained deposits), but are relatively common in the
high-aggradation setting (comprising 20% of total mud
content). Conversely, interbar muds compose 35% of
the fine-grained lithofacies when aggradation rates are
low, but only 17% with high aggradation rates. In this
hypothetical example, the relative amount of each
Lynds and Hajek

1285

specific fine-grained lithofacies is a direct result of


preservation potential caused by vertical stacking. The
total amount of preserved, fine-grained sediment is
related to stacking and the number of channels with
fine-grained sediment. This conceptual model may have
important applications to actual reservoirs. Understanding how aggradation, accommodation, and avulsionreturn time influenced preservation of channel-fill units
in a particular reservoir may help adapt production
strategies to the specific types of lithofacies characteristic of that reservoir.

ervoir and may be used as a basis for constraining


the stochastic range of mudstone dimensions in reservoir models.
Analyzing braided-river reservoirs using this conceptual model requires no new data acquisition for
most reservoirs, only a new way of approaching reservoir characterization. Fine-grained sediments intercalated throughout sand-dominated braided-river
deposits pose significant problems for reservoir development. This model provides a paradigm for populating reservoir models that is efficient, potentially
more accurate, and can easily be adjusted to account for
the inherent variability of sandy braided-river deposits.

CONCLUSIONS
Braided rivers vary widely in terms of scale, sediment
supply, discharge, aggradation rate, accommodation
history, and avulsion-return time. Because of these different factors, all braided-river reservoirs are not created
equal. To account for this variability, empirically derived reservoir models based on outcrop measurements
must be carefully crafted to include only data from
observed deposits that are similar both in scale and
stacking to the reservoir under analysis. By studying
fine-grained deposits in two modern and two ancient
braided-river systems, we have reached the following
conclusions:
1. Identifying fine-grained lithofacies is important because each type of fine-grained material may present different challenges to reservoir development.
We identify five lithofacies (mud plugs, channellining muds, interbar muds, IHS, and flood-plain
and overbank material) that formed in three distinct depositional settings in sandy braided channel
belts.
2. Because each fine-grained lithofacies is associated
with a specific fluvial process, the potential thickness and distribution of each type is limited by the
scale of the fluvial system. Flow depth is a firstorder control on the maximum thickness of most
fine-grained lithofacies.
3. The scaling and preservation relationships we observed are summarized by an idealized channel-fill
unit model. Ancient braided-river deposits can be
described as a series of stacked channel-fill units
that are scaled to the range of paleochannel-thread
dimensions of a particular system.
4. The channel-fill unit model can be tailored to account for the unique history of an individual res1286

REFERENCES CITED
Allen, J. R. L., 1974, Studies in fluviatile sedimentation: Implications of pedogenic carbonate units, lower Old Red Sandstone,
Anglo-Welsh outcrop: Geological Journal, v. 9, p. 181 208.
Bentham, P. A., P. J. Talling, and D. W. Burbank, 1993, Braided
stream and flood-plain deposition in a rapidly aggrading basin:
The Escanilla Formation, Spanish Pyrenees, in J. L. Best and
C. S. Bristow, eds., Braided rivers: Geological Society (London)
Special Publication 75, p. 177 194.
Best, J. L., P. J. Ashworth, C. S. Bristow, and J. Roden, 2003,
Three-dimensional sedimentary architecture of a large, midchannel sand braid bar, Jamuna River, Bangladesh: Journal of
Sedimentary Research, v. 73, no. 4, p. 516 530.
Blakey, R. C., 1989, Triassic and Jurassic geology of the southern
Colorado Plateau, in J. P. Jenney and S. J. Reynolds, eds.,
Geological evolution of Arizona: Arizona Geological Society
Digest, v. 17, p. 369 396.
Bleed, A., 1990, Groundwater, in A. Bleed and C. Flowerday, eds.,
An atlas of the Sand Hills: Conservation and Survey Division,
Institute of Agriculture and Natural Resources, University of
Nebraska Lincoln, v. 5a, p. 67 114.
Brayshaw, A. C., G. W. Davies, and P. W. M. Corbett, 1996, Depositional controls on primary permeability and porosity at the
bedform scale in fluvial reservoir sandstones, in P. A. Carling
and M. R. Dawson, eds., Advances in fluvial dynamics and stratigraphy: Chichester, United Kingdom, Wiley, p. 373 394.
Bridge, J. S., 1993, The interaction between channel geometry,
water flow, sediment transport and deposition in braided
rivers, in J. L. Best and C. S. Bristow, eds., Braided rivers:
Geological Society (London) Special Publication 75, p. 13 71.
Bridge, J. S., 1997, Thickness of sets of cross strata and planar strata
as a function of formative bed-wave geometry and migration,
and aggradation rate: Geology, v. 25, no. 11, p. 971 974.
Bridge, J. S., 2003, Rivers and floodplains: Forms, processes, and
sedimentary record: Oxford, Blackwell Publishing, 491 p.
Bridge, J. S., and R. S. Tye, 2000, Interpreting the dimensions of
ancient fluvial channel bars, channels, and channel belts from
wire-line logs and cores: AAPG Bulletin, v. 84, no. 8, p. 1205
1228.
Bristow, C. S., 1987, Brahmaputra River: Channel migration and
deposition, in F. G. Ethridge, R. M. Flores, and M. D. Harvey,
eds., Recent developments in fluvial sedimentology: SEPM
Special Publication 39, p. 63 74.

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

Bristow, C. S., and J. L. Best, 1993, Braided rivers: Perspectives and


problems, in J. L. Best and C. S. Bristow, eds., Braided rivers:
Geological Society (London) Special Publication 75, p. 1 11.
Bristow, C. S., R. L. Skelly, and F. G. Ethridge, 1999, Crevasse
splays from the rapidly aggrading, sand-bed, braided Niobrara
River, Nebraska: Effect of base-level rise: Sedimentology, v. 46,
no. 6, p. 1029 1047.
Bromley, M. H, 1991, Variations in fluvial style as revealed by
architectural elements, Kayenta Formation, Mesa Creek, Colorado, U.S.A.; evidence for both ephemeral and perennial fluvial
processes, in A. D. Miall and N. Tyler, The three-dimensional
facies architecture of terrigenous clastic sediments and its implications for hydrocarbon discovery and recovery: Concepts
in Sedimentology and Paleontology 3, p. 94 102.
Cant, D. J., and R. G. Walker, 1976, Development of a braided-fluvial
facies model for the Devonian Battery Point Sandstone, Quebec:
Canadian Journal of Earth Science, v. 13, no. 1, p. 102 119.
Collinson, J. D., 1978, Vertical sequence and sand body shape in
alluvial sequences, in A. D. Miall, ed., Fluvial sedimentology:
Canadian Society of Petroleum Geologists Memoir 5, p. 577
586.
Collinson, J. D., 1996, Alluvial sediments, in H. G. Reading, ed.,
Sedimentary environments: Processes, facies and stratigraphy:
Oxford, United Kingdom, Blackwell Science, p. 37 82.
Fouch, T. D., T. F. Lawton, D. J. Nichols, W. B. Cashion, and W. A.
Cobban, 1983, Patterns and timing of synorogenic sedimentation in Upper Cretaceous rocks of central and northeast Utah,
in M. W. Reynolds and E. D. Dolly, eds., Mesozoic paleography of west-central United States: Rocky Mountain Section,
SEPM, p. 305 336.
Hajek, E. A., and P. L. Heller, 2004, Determining fluvial stacking
patterns in the lower Castlegate Sandstone (Campanian,
Helper, Utah) using LIDAR imaging: Geological Society of
America Abstracts with Programs, v. 36, no. 5, p. 462.
Harshbarger, J. W., C. A. Repenning, and J. H. Irwin, 1957, Stratigraphy of the uppermost Triassic and the Jurassic rocks of the
Navajo Country: U.S. Geological Survey Professional Paper 291,
74 p.
Herries, R. D., 1993, Contrasting styles of fluvial-aeolian interaction
at a downwind erg margin; Jurassic Kayenta Navajo transition,
northeastern Arizona, U.S.A., in C. P. North and D. J. Prosser,
eds., Characterization of fluvial and aeolian reservoirs: Geological Society (London) Special Publication 73, p. 199 218.
Lawton, T. F., 1985, Style and timing of frontal structures, thrust
belt, central Utah: AAPG Bulletin, v. 69, no. 7, p. 1145 1159.
Leclair, S. F., J. S. Bridge, and F. Wang, 1997, Preservation of crossstrata due to migration of subaqueous dunes over aggrading
and non-aggrading beds; comparison of experimental data with
theory: Geosciences Canada, v. 24, no. 1, p. 55 66.
Leopold, L. B., and M. G. Wolman, 1957, River channel patterns:
Braided, meandering and straight: U.S. Geological Survey
Professional Paper 282-B, 85 p.
Luttrell, P. R., 1993, Basinwide sedimentation and the continuum
of paleoflow in an ancient river system: Kayenta Formation
(Lower Jurassic), central portion Colorado Plateau: Sedimentary Geology, v. 85, p. 411 434.
Martin, J. H., 1993, A review of braided fluvial hydrocarbon
reservoirs: The petroleum engineers perspective, in J. L. Best
and C. S. Bristow, eds., Braided rivers: Geological Society
(London) Special Publication 75, p. 333 368.
Marzolf, J. E., 1988, Controls on late Paleozoic and early Mesozoic
eolian deposition of the western United States: Sedimentary
Geology, v. 56, p. 167 191.
McClaurin, B. T., and R. J. Steel, 2000, Forth-order nonmarine to
marine sequences, middle Castlegate Formation, Book Cliffs,
Utah: Geology, v. 28, no. 4, p. 359 362.

Miall, A. D., 1977, A review of the braided-river depositional environment: Earth Science Reviews, v. 13, p. 1 62.
Miall, A. D., 1978, Lithofacies types and vertical profile models in
braided river deposits: A summary, in A. D. Miall, ed., Fluvial
sedimentology: Canadian Society of Petroleum Geologists
Memoir 5, p. 597 604.
Miall, A. D., 1988, Reservoir heterogeneities in fluvial sandstones:
Lessons from outcrop studies: AAPG Bulletin, v. 72, p. 682
697.
Miall, A. D., 1993, The architecture of fluvial-deltaic sequences in
the upper Mesaverde Group (Upper Cretaceous), Book Cliffs,
Utah, in J. L. Best and C. S. Bristow, eds., Braided rivers: Geological Society (London) Special Publication 75, p. 305 332.
Miall, A. D., 1994, Reconstructing fluvial macroform architecture
from two-dimensional outcrops: Examples from the Castlegate
Sandstone, Book Cliffs, Utah: Journal of Sedimentary Research, v. B64, p. 146 158.
Middleton, L. T., and R. C. Blakey, 1983, Processes and controls on
the intertonguing of the Kayenta and Navajo formations, northern Arizona; eolian-fluvial interactions, in M. E. Brookfield and
T. S. Ahlbrandt, eds., Eolian sediments and processes: Developments in Sedimentology, v. 38, p. 613 634.
Mohrig, D., P. L. Heller, C. Paola, and W. J. Lyons, 2000, Interpreting avulsion process from ancient alluvial sequences: Guadalope
Matarranya system (northern Spain) and Wasatch Formation
(western Colorado): Geological Society of America Bulletin,
v. 112, no. 12, p. 1787 1803.
Moreton, D. J, P. J. Ashworth, and J. L. Best, 2002, The physical
scale modeling of braided alluvial architecture and estimation
of subsurface permeability, in P. M. Burgess and C. Paola, eds.,
Numerical and physical experimental stratigraphy: Basin Research, v. 14, p. 265 285.
North, C. P., and K. S. Taylor, 1996, Ephemeral-fluvial deposits;
integrated outcrop and simulation studies reveal complexity:
AAPG Bulletin, v. 80, p. 811 830.
Olsen, T., R. Steel, K. Hogseth, T. Skar, and S. L. Row, 1995,
Sequential architecture in a fluvial succession; sequence stratigraphy in the Upper Cretaceous Mesaverde Group, Prince
Canyon, Utah: Journal of Sedimentary Research, Section B:
Stratigraphy and Global Studies, v. 65, no. 2, p. 265 280.
Paola, C., 1989, Topographic sorting: Eos, Transactions, American
Geophysical Union, v. 70, no. 15, p. 323.
Paola, C., and L. Borgman, 1991, Reconstructing random topography from preserved stratification: Sedimentology, v. 38, p. 553
565.
Peterson, F., 1986, Jurassic paleotectonics in the west-central part
of the Colorado Plateau, Utah and Arizona, in J. A. Peterson,
ed., Paleotectonics and sedimentation in the Rocky Mountain
Region, United States: AAPG Memoir 41, p. 563 596.
Pipiringos, G. N., and R. B. OSullivan, 1978, Principal unconformities in Triassic and Jurassic rocks, Western Interior United
States: A preliminary survey: U.S. Geological Survey Professional Paper 1035-A, 29 p.
Robinson, J. W., and P. J. McCabe, 1997, Sandstone-body and shalebody dimensions in a braided fluvial system: Salt Wash Member
(Morrison Formation), Garfield County, Utah: AAPG Bulletin,
v. 81, no. 8, p. 1267 1291.
Robinson, J. W., and P. J. McCabe, 1998, Evolution of a braided river
system: The Salt Wash Member of the Morrison Formation
(Jurassic) in southern Utah, in K. W. Shanley and P. J. McCabe,
eds., Relative role of eustasy, climate, and tectonism in
continental rocks: SEPM Special Publication 59, p. 93 107.
Robinson, R. A. J., and R. L. Slingerland, 1998, Grain-size trends,
basin subsidence and sediment supply in the Campanian
Castlegate Sandstone and equivalent conglomerates of central
Utah: Basin Research, v. 10, p. 109 127.

Lynds and Hajek

1287

Sanabria, S. A., 2001, Sedimentology and sequence stratigraphy of the


Lower Jurassic Kayenta Formation, Colorado Plateau, U.S.A.:
Ph.D. dissertation, Rice University, Houston, Texas, 245 p.
Schumm, S. A., 1960, The shape of alluvial channels in relation to
sediment type: U.S. Geological Survey Professional Paper 352B, p. 17 30.
Skelly, R. L., C. S. Bristow, and F. G. Ethridge, 2003, Architecture
of channel-belt deposits in an aggrading shallow sandbed
braided river: The lower Niobrara River, northeast Nebraska:
Sedimentary Geology, v. 158, p. 249 270.
Smith, N. D., R. L. Slingerland, M. Perez-Arlucea, and G. S.
Morozova, 1998, The 1870s avulsion of the Saskatchewan
River: Canadian Journal of Earth Sciences, v. 35, no. 4, p. 453
466.
Spieker, E. M., 1946, Late Mesozoic and early Cenozoic history of
central Utah: U.S. Geological Survey Professional Paper 0205D, p. 117 161.
Stephens, M., 1994, Architectural element analysis within the
Kayenta Formation (Lower Jurassic) using ground-probing
radar and sedimentological profiling, southwestern Colorado:
Sedimentary Geology, v. 90, p. 179 211.
Thomas, R. G., D. G. Smith, J. M. Wood, J. Visser, E. A. CalverlyRange, and E. H. Koster, 1987, Inclined heterolithic stratification; terminology, description, interpretation and significance: Sedimentary Geology, v. 53, no. 1 2, p. 123 179.
Thorne, C. R., A. P. G. Russell, and M. K. Alam, 1993, Planform
pattern and channel evolution of the Brahmaputra River, Bangladesh, in J. L. Best and C. S. Bristow, eds., Braided rivers: Geological Society (London) Special Publication 75, p. 257 276.
Tuesink, M. F., 1989, Depositional analysis of an eolian-fluvial
environment; the intertonguing of the Kayenta Formation and
Navajo Sandstone (Jurassic) in southwestern Utah: Masters
thesis, Northern Arizona University, Flagstaff, Arizona, 189 p.
USDA (U.S. Department of Agriculture) Farm Service Agency,
2003, Data bank of rectified digital imagery: http://www.dnr

1288

.state.ne.us/databank/fsa03.html (accessed December 18,


2004).
U.S. Geological Survey Water Resources, 2005a, Calendar year
streamflow statistics for Nebraska, U.S. Geological Survey
06466000 Niobrara River at Niobrara, Nebraska: http://nwis
.waterdata.usgs.gov/ne/nwis/annual/?site_no=06466000&
agency_cd=USGS (accessed February 28, 2005).
U.S. Geological Survey Water Resources, 2005b, Calendar year streamflow statistics for Nebraska, U.S. Geological Survey 06786000
North Loup River at Taylor, Nebraska: http://nwis.waterdata.usgs
.gov/ne/nwis/annual/?site_no=06786000&agency_cd=USGS
(accessed February 28, 2005).
Van de Graaff, F. R., 1972, Fluvial-deltaic facies of the Castlegate
Sandstone (Cretaceous), east-central Utah: Journal of Sedimentary Petrology, v. 42, no. 3, p. 558 571.
Watkins, D. K., and R. F. Diffendal Jr., 1997, Geology of Niobrara
State Park Area, in C. A. Flowerday and R. F. Diffendal Jr.,
eds., Geology of Niobrara State Park, Knox County, Nebraska,
and adjacent areas With a brief history of the park, Gavins
Point Dam, and Lewis and Clark Lake: Conservation and Survey Division, Institute of Agriculture and Natural Resources,
University of Nebraska Lincoln, p. 6 12.
Weber, K. J., 1982, Influence of common sedimentary structures on
fluid flow in reservoir models: Journal of Petroleum Technology, v. 34, p. 665 672.
Williams, G. P., 1986, River meanders and channel size: Journal of
Hydrology, v. 88, p. 147 164.
Yoshida, S. A., 2000, Sequence and facies architecture of the upper
Blackhawk Formation and the lower Castlegate Sandstone
(Upper Cretaceous), Book Cliffs, Utah, U.S.A.: Sedimentary
Geology, v. 136, no. 3 4, p. 239 276.
Yoshida, S., A. Willis, and A. D. Miall, 1996, Tectonic control of
nested sequence architecture in the Castlegate Sandstone
(Upper Cretaceous) Book Cliffs, Utah: Journal of Sedimentary
Research, v. 66, p. 737 748.

Conceptual Model for Predicting Mudstone Dimensions in Braided-River Reservoirs

You might also like