Geomorphology 116 (2010) 320–329

Contents lists available at ScienceDirect

Geomorphology
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / g e o m o r p h

Mobility of large woody debris (LWD) jams in a low gradient channel

Joanna C. Curran
University of Virginia, Department of Civil and Environmental Engineering, PO Box 400742, Charlottesville, VA 22904, USA

a r t i c l e i n f o a b s t r a c t

Article history: Mobility of large woody debris (LWD) in low gradient channels is an important but often overlooked
Received 16 April 2009 transport process. The majority of studies on LWD have focused on its role in geomorphic and ecologic river
Received in revised form 9 October 2009 processes. When jams extend across the width of the channel, they have the potential to retain sediment and
Accepted 23 November 2009
alter the channel profile. When jams obstruct only a portion of the channel, they can re-direct flow, altering
Available online 4 January 2010
patterns of scour and deposition. The boundary complexity created by LWD has a recognized role in riverine
Keywords:
ecosystems which has led to programs of replacing LWD in-channel corridors where it was previously
Large woody debris removed. Although LWD jams are common in rivers around the world, they have been studied most
Morphodynamics intensely in steep, forested channel reaches where they are often found to be stable channel features. It is not
Wood jams fully known how much of the information on LWD from steep forested channels will transfer to other
Mobility channel types. Whereas it may be reasonable to assume that the ecological benefits of LWD are similar in low
gradient channels, research has shown that a much higher rate of LWD transport occurs in low gradient
channels, with jams mobilized on timescales of 100–102 years.
This study evaluates the distribution and mobility of LWD over 72 km of the San Antonio River, a low
gradient channel in southeast Texas. LWD jam locations were identified for 2003 and 2007 using a
combination of aerial photography and field mapping. Each jam was cataloged according to its location in the
channel cross-section and the amount of channel area blocked. During the four-year period, all the LWD jams
were mobilized, including those jams extending across the channel width. Although easily mobilized, 34
jams re-form in the same locations, creating 34 channel locations with persistent LWD jams. Data from the
San Antonio River are applied to two models developed to predict LWD mobility and transport distances to
assess the applicability of each model to a low gradient channel. The locations of stable (or recurring) LWD
jams were matched to model results where predicted LWD transport distances were equal to measured LWD
jam spacing. Model results showed good agreement with the mean and median spacing of LWD jams when
given input parameters specific to the channel and wood species. The ability to predict where LWD jams will
persist over time in a low gradient channel has application in watershed management. Persistent LWD jams
can exert a greater influence on channel morphology and may require active management.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction occurred in the form of bridge construction or channelization (i.e.

Swanson et al., 1982; Wohl and Goode, 2008).
Large woody debris (LWD) and woody debris jams have been LWD and wood jams were common in American rivers prior to the
documented in rivers over a number of continents and physiographic 19th century. Much of this wood was removed during the 19th and 20th
regions (i.e. Lisle, 1986; Gurnell et al., 2000; Haschenburger and centuries to improve river navigation, reduce flooding, and enhance
Wilcock, 2003; Daniels and Rhoads, 2003; Faustini and Jones, 2003; ecosystems (Montgomery et al., 2003). Many studies have focused on
Comiti and Lenzi, 2006). Woody debris has attracted the interest of rivers where the detrimental impacts of wood removal on channel
researchers for its ability to enhance aquatic ecosystems, affect morphology and aquatic ecosystem health were first documented:
channel morphology, and alter channel hydraulics. It can create flow steep channels within forested landscapes. A number of geomorphic,
resistance and influence the formation of channel bedforms, including ecologic, and hydraulic characteristics have been tied to the presence of
inducing step pool formation. Where wood jams form, upstream LWD. Research has shown a link between LWD and channel
reaches can experience backwater effects and induce sediment morphology (i.e. Robison and Beschta, 1990; Nakamura and Swanson,
deposition (Daniels, 2006). The majority of LWD studies have been 1993; Montgomery et al., 1995; Abbe and Montgomery, 1996;
conducted in forested areas where little anthropogenic influence has Wallerstein and Thorne, 2004) ecological value through enhanced
ecosystem habitat (i.e. Bilby and Likens, 1980; Smock et al., 1989; Benke
and Wallace, 1990), and hydraulics in the immediate region of LWD
E-mail address: curran@virginia.edu. jams (i.e. Magna and Kirchner, 2000; Manners et al., 2007). LWD is

0169-555X/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2009.11.027
 J.C. Curran / Geomorphology 116 (2010) 320–329 321

defined as logs with either a diameter of 0.1 m or a length of 1 m. Logs the channel, to evaluate jam mobility, and to apply the field data of
are supplied to channels through riparian tree mortality, bank failure, LWD mobility to models of log transport. This paper aims to fill a gap
channel incision, windthrow, and floods (Downs and Simon, 2001; concerning wood in low gradient channels. By modeling the observed
Wallerstein and Thorne, 2004). log movement, the paper evaluates the applicability of results from
The extent of the impact of LWD on channel morphodynamics is studies of LWD in steep channels to low gradient channels.
largely a function of the residence time of the logs in the channel.
Residence time measures the period of time LWD and debris jams 2. Study site
remain stationary and is often measured by indirect means, for example
as the age of young trees growing out on a stable jam (Martin, 2001). The study area encompasses 72 km along the main stem of the
LWD and debris jams in steep channels can have residence times lower San Antonio River, from below San Antonio to the crossing of
measured in decades and the effect on channel morphology can be FM541 near Poth, TX (Fig. 1). This length of river has a watershed
significant (Wallerstein and Thorne, 1996; Webb and Erskine, 2003). In contributing area of 5473 km2. The study reach begins downstream of
British Columbia, wood jams have been documented as stable for time the concrete-lined river reach within the city of San Antonio and falls
periods in excess of 200 years (Keller and Tally, 1979). Rather than entirely within the coastal plain physiographic region. Longitudinal
mobilizing during large flows, stable jams extending across the channel slope in the study reach is 5.7 × 10− 4. Land use is historically
can develop a step pool morphology as flows over the jams scour a dominated by ranches and agriculture outside of the city of San
plunge pool downstream (Thompson, 1995; Webb and Erskine, 2003). Antonio, where forested areas intersperse within the rangeland.
Channel morphodynamics adjust to the presence of the wood and the A USGS gauging station, located at Elmendorf, TX (gage 08181800),
channel planform is stabilized by the wood as the wood decays in place. coincides with the upstream end of the study reach. A continuous flow
Long residence wood is common in steep headwater channels in record occurs at this gage extending to 1962. For the period from 2003 to
forested areas where channels typically have a width that is less than or 2007 the flows measured correspond to the time of LWD transport
equal to that of the average length of log that falls in the river. The fallen observations (Fig. 2). The mean daily flow for this period is 9 m3s–1
log remains in place, sometimes spanning the river width, as channel while the range in daily flows is between 2.10 and 455.95 m3s–1. Major
flow has a limited ability to mobilize and redistribute the wood. As a floods occurred in the watershed in 1998 and 2002, both prior to the
result, accumulations of LWD and jams are common in headwater study time frame.
channels (Keller and Swanson, 1979). Riparian vegetation in the study reach is the source of LWD into
As the watershed area contributing to a channel increases, the size the channel. The entirety of the reach, in the Post Oak Savannah biotic
of the channel and rates of flow increase. The residence time of wood region, is characterized by grassy pastures with interspersed post oak
and wood jams decreases and understanding the mobility of the wood (Quercus stellata) and blackjack oak (Quercus marilandica) (Gould,
becomes increasingly important. LWD mobility focuses on how often 1962). In the immediate riparian area one large tree rarely dominates,
wood moves in a given river system. It is measured through direct enabling a larger density of trees on the edge of the river. Texas
(tagging and surveying) and indirect (repeat aerial photography) sugarberry (Celtis laevigata) dominates the riparian zone, followed by
methods (Gippel et al., 1996). The ability of the channel to transport box elder (Acer negundo), cedar elm (Ulmus crassifolia), and
the wood and the range of flows over which the wood moves increase cottonwood (Populus deltoides). Most cottonwoods (P. deltoides) and
as the channel width to log length ratio increases. In a small U.K. black willows (Salix nigra) are isolated to the inner riverbank, while
stream, 270 LWD jams changed position or character in less than Texas sugarberry (C. laevigata), cedar elm (U. crassifolia), and
12 months during which the highest recorded discharge was near American elm (Ulmus Americana) are more evenly distributed from
bankfull (Gregory et al., 1985). Channel spanning log jams are rarer as the inner bank to the outer bank (Bush and Van Auken, 1984). All of
the channel width increases, but those jams that do form are typically these species are fast growing, and indicate that the riparian zone has
larger than those found in the upstream reaches (Swanson et al., adjusted to frequent disturbances by flooding (Kochel and Baker,
1982). Mobility is often greater in low gradient channels which tend 1982) and that LWD is supplied to the river. The tallest trees within
to have larger channel widths and a higher stream order. the riparian zone are the cottonwoods and black willows, which can
Increased LWD and wood jam mobility does not necessarily grow up to 30 m and 37 m, respectively, and over a meter in diameter
correlate to reductions in the influence of wood on channel (Preston, 1976; Brockman et al., 2001). Trees of medium height
morphodynamics. Where LWD and wood jams are easily mobilized, include Texas sugarberry and American elm, both of which can reach
the frequency of wood mobility and preferred deposition locations over 20 m with diameters near 2 m. Box elder remains a small tree in
become important to understanding channel morphodynamics. When the study reach, rarely reaching 20 m tall and diameters of 1 m
wood jams recreate in the same locations after each mobilizing flood, (Preston, 1976; Brockman et al., 2001).
the effect on the channel can be similar to that of immobile jams. Low Soils adjacent to the river channel are identified for each 2 km
gradient channels are likely to include areas where the built segment within the study reach. Soil types alternate between clay and
environment intersects the river, making them subject to active clay loam with the clay soil type dominant (Wermund, 1996; Soil
channel and watershed management. It is important in these channels Survey Staff, 2008). Channel width and planform pattern were
to understand how often the large wood moves and where it tends to analyzed from historic aerial photos. Between photo years 1938 and
accumulate so that management can be targeted and construction 1948, channel width increased from an average of 25 m in 1938 to an
around channels can be more informed. average of 40 m in 1948. A large flood in 1946 is a probable catalyst for
It is not fully known how much of the information on LWD from observed bank failures and channel widening across the study reach
steep forested channels will transfer to other types of channels. during this time frame (Cawthon and Curran, 2007). Between 1948
Whereas it may be reasonable to assume that the ecological benefits and 2007, the channel width increased by an average of 9 m. The
of LWD are similar in low gradient channels, research has shown that limited amount of channel widening over the past 60 years illustrates
a much higher rate of LWD transport occurs in low gradient channels, a stable channel width during the time period over which LWD are
with jams mobilized on timescales of 100–102 years. analyzed.
This study evaluates the distribution and mobility of LWD in a low
gradient channel in southeast Texas over a four-year time span. The 3. Results: field
locations of LWD jams are documented and categorized according to
type of deposition and amount of channel coverage. The purpose of LWD jams are composed of multiple pieces of wood of different
the paper is to document and characterize the distribution of wood in lengths. Log jams form when pieces of LWD accumulate into a stable
322 J.C. Curran / Geomorphology 116 (2010) 320–329

Fig. 1. Map showing the San Antonio River watershed within Texas and the study reach within the watershed (the highlighted portion of the river).

structure. An initial log deposits when an obstacle in the channel LWD jams present in the study reach were field mapped using GPS
precludes continued downstream transport or the depth of flow reduces over the timeframe from November 2006 to February 2007. The
to the point at which the log no longer floats. This log then acts as a jam percent lateral surface channel coverage and channel cross-sectional
‘key member’ and stabilizes the smaller ‘racked members’ that comprise location were recorded for each jam (Table 1). Lateral channel
the majority of the jam (Abbe and Montgomery, 2003). A requirement coverage was estimated and those jams extending over 90% of the
in the definition used in the field during this research was that the jam channel width were considered complete jams. LWD jam in-channel
must appear stable in the channel (Fig. 3). Many locations exist where locations are described as river left, river right, or mid-channel. A total
small wood accumulated on the water surface in a defined area but did of 142 jams were identified, and of these 95 extended over more than
not include a stabilizing key member (Fig. 4). These accumulations of 5% of the channel area. Jam spacing was measured between each set of
wood create areas of congested wood transport but do not represent adjacent jams. The average and median measured spacing between
stable LWD jams and were not included in mapping as LWD jams. The jams was 430 m and 230 m, respectively. LWD jams present on
focus of the research is on the ability of LWD jams to influence channel December 7, 2003, were mapped from 30 cm resolution, low-level
morphodynamics in low gradient channels and wood accumulations in aerial photos (San Antonio River Authority, 2004). From these images
the absence of a key member are unlikely to exert a significant influence a total of 73 jams were identified with mean and median jam spacing
on the channel. of 840 m and 500 m, respectively.
Jam locations from 2003 are compared to locations mapped in the
field in 2007. The mean flow on Dec. 7, 2003 at the Elmendorf USGS
gage was 9.65 m3s–1 with a depth 4.04 m. The mean recorded flow
during the field mapping was 8.28 m3s–1 with mean water depth of
3.56 m. Approximately twice as many LWD jams were identified in
2007 when compared to 2003, and the density of jams increased from
1.0 jam/km in 2003 to 2.1 jams/km in 2007. The lower total number of
jams and larger jam spacing in 2003 when compared to 2007 may
partially result from the different methods used to locate and identify
jams. Whereas individual jams spaced within 1 m of each other can be
distinguished in the field where the river was surveyed longitudinally
by boat, on aerial photos these same jams can appear merged as a
single large jam. It is also easier to identify LWD jams that span 0–5%
of the lateral channel area when in field situation, and these smaller
Fig. 2. Discharges as measured at the USGS gage at Elmendorf, Texas during the time jams may have been present in 2003 but not identifiable from aerial
frame of the study, December, 2003–March, 2007. photos.
 J.C. Curran / Geomorphology 116 (2010) 320–329 323

Fig. 3. Example LWD jam in the study reach, located near Elmendorf on the San Antonio River. Jam spans the channel and two ‘key members’ are identified (Cawthon, 2007).

Variations and patterns in jam density in the study reach are 60 km has a tortuous meander pattern and sinuosity ratio of 2.6 rather
consistent between 2003 and 2007. Between river km 43 and river km than the gentler bends and lower sinuosity elsewhere in the study reach.
60, the LWD jam density increases dramatically with an increase in river LWD jam locations are compared to estimate the overall LWD jam
sinuosity. The channel sinuosity between river km 43 and 60 is 2.6 while mobility. Tracking devices were not used to monitor the movement of
outside this segment the average sinuosity is 1.86. In 2003, the jam individual logs, so it is not possible to distinguish with certainty
density increases from 0.9 jams/km outside of river km 43–60 to whether a jam identified in 2003 and 2007 remained stationary or re-
1.6 jams/km within the 17 km segment. In 2007, the pattern repeats formed in the same location. Field observations do not indicate that the
such that outside of river km 43–60 jam density is 1.6 and within the logs comprising the debris jams had been exposed to the weather for a
segment between km 43 and 60, jam density is 4.1 jams/km. The long period of time. Of the 73 LWD jams located on photos from 2003,
consistent pattern of LWD jam density indicates a source of wood and an 41 had mobilized and transported downstream by 2007. Neither a
increase of wood storage in jams between river km 43 and river km 60. significant increase in the size of LWD jams between 2003 and 2007
The confluence of the Seguin River with the San Antonio River occurs at nor an increase in downed wood in the riparian area was observed,
km 32, increasing the total drainage area by 49 km2 and providing LWD indicating that the wood from these 41 jams traveled out of the study
to the channel. More importantly, the channel planform between 43 and reach. Whereas tributaries to the San Antonio River are often cleared of

Fig. 4. Example of an area of wood accumulation on the San Antonio River, but not a LWD jam (Cawthon, 2007).
324 J.C. Curran / Geomorphology 116 (2010) 320–329

Table 1 Table 2
LWD locations and characteristics recorded in the field from 2006 to 2007. Details of the LWD jams that re-form in the channel.

% lateral channel Over 90% 60%–90% 40%–60% 20%–40% 5%–20% 0–5% Longitudinal Location % lateral Longitudinal Location % lateral
coverage location in-channel channel location in-channel channel
(river km) (from 2007) coverage (river km) (from 2007) coverage
19 10 10 15 41 47
(from 2007) (from 2007)

Location Complete River right Mid-channel River left 14.1 River left 0–5 45.6 River left 0–5
15.9 River right 0–5 45.8 River left 5–20
23 40 25 54
16.4 River left 0–5 46.7 River left 0–5
25.9 River left 40–60 46.75 River left 0–5
30.6 River left 0–5 46.9 Mid-channel 20–40
woody debris, the main stem of the river is not. Thus, the wood can be 31.5 River left 0–5 47.9 River right 0–5
assumed to have moved through the study reach, implying that the 33.6 Complete 90–100 48.8 River left 0–5
35 Mid-channel 60–90 52.8 Complete 90–100
other 32 LWD jams identified on the 2003 photos were also mobile.
36 River left 0–5 53.6 Mid-channel 5–20
Wood mobility is closely tied to the rates and depths of flow in the 36.5 Mid-channel 5–20 55.1 Mid-channel 60–90
channel. Mobility of wood and LWD jams in the study reach is 100% 38.5 Mid-channel 60–90 57.5 River left 20–40
over the four-year period. During this time, the largest flow in the 43.5 River right 40–60 57.9 River left 5–20
river on November 23, 2004 was 456 m3s–1, which represents a flood 43.95 Mid-channel 5–20 59.4 River left 5–20
44.1 Mid-channel 5–20 63.9 Mid-channel 60–90
with a 10% probability of exceedence. The discharge rate and channel 44.4 River left 0–5 64.9 Mid-channel 5–20
depth were large enough to float and mobilize the LWD, including the 44.5 Mid-channel 20–40 66.6 River right 5–20
complete jams. LWD jams may also mobilize at flows lower flows as Average LWD 1700 Median LWD 980
evidenced upon inspection of aerial photos from June 2, 2004, which jam spacing (m) jam spacing (m)
Average spacing between 930 Median spacing between 730
were available for limited portions of the study reach. When photos
43 and 60 km (m) 43 and 60 km (m)
pairs from December, 2003 and June, 2004 were compared, a channel Average spacing between 2550 Median spacing between 1840
spanning jam documented in 2003 was found to have mobilized by 0 and 43 km and 60 and 0 and 43 km and 60 and
June, 2004. During this period, the largest flow was 153 m3s–1, with a 67 km (m) 67 km (m)
94% exceedence, indicating that LWD and jams are easily mobilized in
the lower San Antonio River.
The comparison of 2003 and 2007 data identified 32 preferential jams form in bends and straight reaches, but the majority is associated
LWD deposition and jam formation locations (Fig. 5; Table 2). The with channel meanders. The most common repeat locations for LWD
jams are the outside of meander bends where the key member is a
large tree fallen into the channel as the meander bend eroded
outward (Table 1). Although the average channel width over the
study area increased by only 9 m between 1948 and 2007, the
majority of the widening occurred on the outside of meanders to
enlarge channel bends (Cawthon and Curran, 2007). The fallen tree
acts as the key member and traps smaller wood being transported
downstream to form a jam. The fewest persistent jams are associated
with the inside of meander bends, which build after a key member
deposits on a point bar. Mid-channel bars create an obstruction to
continued LWD transport at low flows and are also a persistent
location for LWD jams. The bed of the San Antonio River is primarily
sand with small gravel where bedforms and mid-channel bars are
easily built. Similar to the increase in LWD jam density downstream of
km 34, the number of repeat jam locations increases in the lower part
of the study reach. Complete channel jams repeatedly form at two
locations, one of which is a bridge.

4. Models

Two models, developed to predict the circumstances under which

LWD is mobilized, transported, and deposited, are applied to the San
Antonio River data. These models were developed from flume work and
designed around small, steep gradient, forested channels. Data from the
San Antonio River will be used to assess the applicability of each model
to a low gradient channel where LWD mobility is frequent. A short
summary of each model is provided here with the details behind model
development available through the referenced literature.
The first model, developed in Braudrick et al. (1997) and Braudrick
and Grant (2000), is referred to here as the Braudrick and Grant model. It
was developed from flume experiments where the mobility of wooden
dowels, with and without disks attached to simulate rootwads, was
measured over a sand bed. Movements of dowels, including rolling,
Fig. 5. Persistent locations of LWD jams in the study reach. LWD jam locations as field sliding, and rotating, were documented along with rates of transport
mapped between 2006 and 2007 are shown as circles. Where a mapped jam is in the
same location as a jam identified from 2003 aerial photography, the location is marked
and preferential deposition locations. Movement of dowels was a
as a repeat jam and identified by a red triangle. (For interpretation of the references to function of the ability of the wood to float and, hence, a function of the
color in this figure legend, the reader is referred to the web version of this article.) type of wood. In the lab, the measured physical characteristics of the
 J.C. Curran / Geomorphology 116 (2010) 320–329 325

dowels included wood species, specific gravity, and moisture content. water depth (m), ρlog is the density of the log (kgm− 3), and ρw is the
For the San Antonio River, moisture contents are assumed to be between density of water(kgm− 3). When Y* N 1.26, the wood will float, and
either 20 and 40% or 30 and 60%, depending on tree species. At these below that value it may move by sliding or rolling on the channel bed,
levels of saturation, all of the species have specific gravities above determined as a function of the friction angle of the log on the bed, log
500 kgm− 3. At moisture contents above 60%, the log density becomes length, and log diameter. Once the forces acting on the wood are great
greater than water density and causes the log to sink (Simpson and enough to surpass the threshold necessary to float the log, the model
TenWolde, 1999). Given the physical characteristics of the wood, the predicts the probability of LWD deposition in the presence of
critical depth for floatation can be calculated using Eq. (1) (re-arranged obstacles. The model assumes that the distance traveled by a log is a
from Eq. (3) in Braudrick et al., 1997). function of the dimensionless log length, L*T = LT/L0, and the excess
2
 3 force acting on the log, X*e (Eq. (3), from Eqs. (4) and (7) in Bocchiola
ρ −1
d −r 
π ρlog − 12 − sin c et al., 2006).
26 r 7
dc = r + r 4
w
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 5 ð1Þ
2  
2dc r−d2c * 1 Uw 1 1:26
Xe = − −1 = 1:26 ð3Þ
2 gDlog ðcosα tanδ− sinαÞ Y*
where r is the log radius (m), dc is the critical water depth for
floatation (m), ρlog is the density of the log (kgm− 3), and ρw is the where LT is the length of log travel (m), L0 is the spacing between
density of water (kgm− 3). Dowels floated when the water depth was obstacles in the channel (m), Uw is the average water velocity (ms− 1),
a minimum of half the dowel diameter, which was expressed as the α is the bed slope, δ is the friction angle of the log when rolling, and g
dimensionless ratio of dowel diameter to water depth. This ratio was is gravity (ms− 2). The distance each dowel traveled during the flume
the primary determinant of mobility for dowels with and without a experiments was documented as well as the number of pieces
simulated rootwad. depositing to form jams in the channel. The results were used to
Exceeding the floating threshold for an individual log will not develop a probabilistic model of log movement, deposition rate, and
necessarily transport that log because channel morphology is an deposition location as a function of log Froude number, where
important feature in determining log mobility. A limitation of the Frlog = Uw/(gDlog)0.5. The normalized dimensionless length traveled
Braudrick and Grant model is that it was designed around shallow by the dowels and number of dowels in each jam were fit to gamma
channels and not tested in deep rivers. In all the experiments, the distributions. For each value of L*T (where L*T = LT/L0), the average
depth of the water was less than the diameter of the dowels. This is dimensionless distance traveled by a single piece of wood or jam can
often the case in steep channels, where the critical depth criteria be predicted as a function of the Frlog. Combining the average distance
works well (Wohl and Goode, 2008). Where the water depth is low in between obstacles in the river with the probability of travel distance,
large, low gradient rivers, this is likely because of channel bar the probable distance traveled by each log and LWD jam can be
formation. Instances of locally reduced water depth can be considered predicted.
in the model through inclusion of spatially varying depths of flow and
channel bathymetry when applying the model criteria of Dlog/dw N 0.5 5. Results: modeling
for floatation. Channel bars act as preferential deposition locations in
the San Antonio River, and the Braudrick and Grant model can be used Channel morphology and flow hydraulics are most influenced by
to predict log transport off the bars. Secondary in importance in LWD in locations where LWD jams persistently form. Data from the
determining log mobility but still a controlling factor over log San Antonio River are applied to the Braudrick and Grant and
transport is the dimensionless ratio, log length to channel width. By Bocchiola models to evaluate each model's ability to predict persistent
using both ratios, log mobility is predicted as a function of channel jam locations and spacing. For each wood type, three different
morphology. For a wide channel, such as the San Antonio River, the moisture contents are modeled to simulate wood of different
ratio must be greater than 0.5 for individual logs to move. The Dlog/dw residence times in the river. The rates of flow applied in the models
criterion has been shown to be a useful predictor of log mobility in correspond to the minimum, mean, and median flows over the time
shallow, steep gradient rivers (Haga et al., 2002). period of the fieldwork (2006–2007), the time period of the study
The second model examined, the Bocchiola, Rulli, and Rosso (2003–2007), and the long term gage record (1962–2007).
model, comes from two manuscripts of Bocchiola et al. (2006, 2008), The Braudrick and Grant model was used to calculate the
and examines the transport and deposition of large woody debris in minimum depth (dc) required to float LWD in the San Antonio River
the presence of obstacles. It is referred to here as the Bocchiola model. (Table 3). Calculated dc values are all much lower than the measured
The analytical model is derived from flume experiments where depths in the study reach, which range from 3.0 to 14.24 m over the
dowels were fixed into the channel bed to create obstacles to period of study, indicating that the wood is able to be mobilized at all
transport. As loose dowels were added to the flume channel, the flows. If this were the only factor governing LWD presence, the river
depositional locations and the influence of obstacles on transport would not retain any LWD. The spatial variability in the bathymetry of
were documented. The Bocchiola model extends the Braudrick and the channel bed is, however, not accounted for in the model. Where
Grant model by including explicit predictions of log movement by channel bars occur, the local depth will be shallower than that
sliding or rolling on the channel bed as a function of flow velocity, bed measured at the Elmendorf gage. LWD may accumulate locally on in-
slope, log length, and the angle of friction of the dowels on the bed. channel bedforms although the measured gage flows indicate that it
The criterion for LWD flotation is adjusted in the Bocchiola model to should float and transport.
account for the effect of the dowel on the local depth of water. The Log characteristics are important to modeling mobility. Measured
threshold for floatation is given by Y*, which includes the buoyancy of depths of flow from the USGS gage are used to calculate the Dlog/dw ratio
the dowel and density of the wood (Eq. (2); from Eq. (2) in Bocchiola in the river. This ratio must be less than 0.5 for the log to transport
et al., 2006). downstream (Fig. 6). As this ratio decreases below the 0.5 threshold, the
logs will transition from sliding and rolling to floating. LWD from
ρw dw 1:26 Cottonwoods plots closest to the threshold line and at the lowest flows
Y* = = ð2Þ
ρlog Dlog 1 + 2:49X*S;R depths, plots above the threshold. Thus, Cottonwood is the most likely
species of LWD to find in the river as it is the least mobile overall.
where X*S,R accounts for the force balance causing a dowel to move by Mobility is a function of the critical water depth required for floatation
rolling (X*R) or sliding (X*S), Dlog is the diameter of the log (m), dw is the and the ratio of log length to channel width. When the Llog/wc ratio is
326 J.C. Curran / Geomorphology 116 (2010) 320–329

Table 3 graphs where model results cluster near a field measured value.
Details of San Antonio data for use with the Braudrick and Grant model. Despite scatter in the model results, groupings are identifiable (Fig. 7).
Species Moisture content (%) ρlog (kgm− 3) dc (m) Llog (m) Llog/wc When the obstacle spacing in the Bocchiola model is set at 500 m to
simulate the tight meander bends, the modeled LWD travel distances
Cottonwood 20 516 0.50 21.8 1.43
Cottonwood 40 770 0.75 21.8 1.43 cluster around 400 and 950 m. The cluster at 950 m closely matches
Cottonwood 60 880 0.87 21.8 1.43 the median persistent jam spacing across the study reach (980 m) and
Willow 20 708 0.21 8.8 0.58 the mean persistent jam spacing between river km 43 and 60 (930 m).
Willow 25 813 0.25 8.8 0.58
The predicted cluster of LWD deposition around 400 m is more
Willow 30 988 0.31 8.8 0.58
Box elder 20 516 0.24 5.5 0.36 representative of the spacing of the temporary LWD in the channel,
Box elder 40 770 0.33 5.5 0.36 jams that were mapped but did not persist in the same location over
Box elder 60 880 0.38 5.5 0.36 time. When all LWD locations, temporary and persistent, are
Sugarberry 20 708 0.30 7.3 0.48 considered, the median jam spacing between 43 and 60 m in 2003
Sugarberry 25 813 0.35 7.3 0.48
was 300 m and in 2007, the total mean spacing was 430 m. When the
Sugarberry 30 988 0.44 7.3 0.48
obstacle spacing was set to 1000 m in the Bocchiola model, the model
predicts preferred LWD deposition after travel distances of 1000 m
greater than 0.5, logs are expected to lodge in the channel where they and 1900 m. These predictions replicate the mean and median spacing
can act as key pieces in jam formation. The average channel width of of all persistent LWD jams as well as the median jam spacing when
15.24 m was used to calculate the Llog/wc ratio for each tree species. only those jams outside the sinuous segment are considered (Table 4).
Although the channel width can vary locally, the standard deviation The 1000 m spacing prediction also agrees well with the measured
across the study reach is 3.4 m. At the average width, the longest LWD mean spacing for all jams, temporary and persistent, in 2003. The
from cottonwoods and willows will lodge in the channel, with Llog/wc largest obstacle spacing used with the model, 2500 m, represents the
ratios of 1.43 and 0.58 respectively. Willows were often recorded at least sinuous segments of the study reach. Model results using a
lengths less than the maximum possible and the Llog/wc reduced.
Characteristics of the San Antonio River are used in the Bocchiola
model to evaluate predicted log transport distances against field
identified persistent LWD jam locations. The Y* value was calculated
to be between 1.83 and 18.36 for all combinations of wood species and
flow rates, which is above the 1.26 threshold value necessary to float
and transport LWD. The relationship between Frlog and the average
distance traveled by LWD in the presence of obstacles that is part of
the Bocchiola model is extended here to predict the distance traveled
by LWD in the San Antonio River. Obstacles to downstream transport
in the San Antonio River include in-channel bars, channel bends, tree
falls, and bridges. Channel bends are the most prevalent and
permanent obstacles to continued downstream transport of LWD,
and three values are modeled: 500 m to represent the tightly
meandering section in the study reach between river km 43 and 60,
1000 m to represent the average spacing between meander bends
throughout the study reach; 2500 m to represent the straight sections
in the study reach. The frequency function for the average distance
traveled by LWD was held constant at 50%.
Model results show an increase in travel distance as discharge and
obstacle spacing increase (Fig. 7; Table 4), illustrating a correlation
between persistent jam locations and the presence of obstacles in the
channel. To evaluate the correlation between model results and
measured LWD jam spacing, groupings are identified by eye on the

Fig. 7. Results of Bocchiola, Rulli, and Russo model applied to data from the San Antonio
River. Discharges and flow depths correspond to the minimum, mean, and median
flows over the time period of the fieldwork, study period, and long term record at the
USGS gage at Elmendorf, Texas. (A) has the obstacle spacing set at 500 m to replicate
the spacing of channel bends between river km 43 and 60. (B) has the obstacle spacing
Fig. 6. Results of Braudrick and Grant model applied to data from the San Antonio River. set at 1000 m to replicate the spacing of the more gentle meanders in the study reach.
Discharges and flow depths correspond to the minimum, mean, and median flows over (C) has the obstacle spacing set at 2500 m to replicate the spacing of the straight
the time period of the fieldwork, study period, and long term record at the USGS gage at sections in the study reach. Groupings of LWD transport distances that correspond to
Elmendorf, Texas. The dashed line represents the value of the Dlog/dw ratio that is an measured LWD spacing are marked with circles on each graph. Please note the change
upper bound on log transport. in scale on the y-axis of the graphs.
 J.C. Curran / Geomorphology 116 (2010) 320–329 327

Table 4 channel by as much as 25% (Wohl and Goode, 2008). This research
LWD jam spacing from 2003 and 2007. focuses on the jams because they could be confidently identified on
Both persistent and Mean spacing (m) Median spacing (m) aerial photos. Mobility predicted by the models is applied to the
temporary LWD Jams persistent LWD jams in the San Antonio River.
2003 Total 840 500 The Braudrick and Grant model predicts wood mobility as a function
km 43–60 590 300 of channel depth. The model can be applied to predict jam formation on
Outside km 43–60 1015 700 in-channel bars where a detailed river bathymetry is known. This project
2007 Total 430 230
did not include a high-resolution map of the channel bed, therefore, the
km 43–60 260 165
Outside km 43–60 601 375 model was most useful in predicting the trees most likely to act as a key
member in jam formation. Using the physical characteristics of the
Persistent LWD jams between Mean spacing (m) Median spacing (m) different tree species in the San Antonio River, Cottonwoods were
2003 and 2007 predicted to be the least mobile species in the channel and, therefore, the
Total 1700 980 most likely to be found in a LWD jam. Field observations confirmed that
km 43–60 930 730 the Cottonwoods, while not the most prevalent species in the riparian
Outside km 43–60 2550 1840
zone, were common as the key member in a LWD jam. The implication of
this result is that riparian areas with a large amount of Cottonwood trees
can act as a source of LWD to a channel indicate a region where a higher
2500 m obstacle spacing agreed with the measured mean and median number of jams may be expected.
persistent LWD jams spacing outside of the sinuous region. The Bocchiola model predicts LWD transport distances in the
presence of obstacles. The spacing of the repeated LWD jams in the
6. Discussion San Antonio was used to evaluate the ability of the model to predict
jam locations. Required input to the model includes the spacing of
Studies of LWD mobility where individual pieces of wood are obstacles in the channel that prevent continued downstream
tagged and tracked over a range of flow rates have produced direct transport of LWD. The San Antonio River is meandering, with
measurements of LWD movement and a range of LWD residence sinuosity ratios between 1.86 and 2.60. Channel meanders are
times (Diez et al., 2001; Haga et al., 2002; Wohl and Goode, 2008). recognized areas of preferential wood deposition and jam formation
Recent temporal studies of the longitudinal distribution of LWD in (Piegay and Gurnell, 1997). The obstacle spacing used for the model
rivers have allowed for process linkages to begin to be made between was set to represent the distance between channel bends. Where the
LWD movement and channel morphology. In an evaluation of meanders are close together, the model simulated the spacing of the
11 years of data tracking LWD and LWD jam movement in the persistent LWD jams well. Jams had an average measured spacing of
Colorado Rocky Mountains, Wohl and Goode (2008) found that the 930 m and the model predicted a group of LWD deposition locations
dimensionless variables L*log (where L*log = Llog/wc) and D*log (where after 950 m of transport. In the areas of the study reach with greater
D*log = dpeak/Dlog) exert the greatest influence over LWD mobility. This distance between channel bends, the model again performed well in
is in agreement with observations and results from other field and predicting persistent LWD jam spacing. Modeling results predict a
flume research, including the models employed in this research group of LWD travel distances near 1900 m while the field measured
(Braudrick et al., 1997; Hilderbrand et al., 1998; Andreoli et al., 2007; median LWD jam spacing was 1940 m. The model also predicted LWD
Bocchiola et al., 2008). with a shorter travel distance and, therefore, a shorter spacing. Two
All of the LWD and LWD jams identified from 2003 aerial photos possible explanations exist for the extra model predictions and the
over a 72 km stretch of the San Antonio River had mobilized by 2007, scatter in the modeled results. First, more LWD jams occur in the
when field mapping was conducted. Upon further inspection of the channel than accounted for by the persistent jams. These jams are
data, it was found that 32 of the LWD jams identified in 2007 were in more mobile, do not necessarily re-form between flows, and have a
the same channel locations as jams were in 2003. When the jams re- shorter average spacing. Second, only the meander bends were
form in the same channel locations, a greater influence by the LWD considered in setting the distance between obstacles. In the field,
exists on local channel hydraulics and morphology. The combination channel bends were one of many potential obstacles that served as
of LWD and a mobile sediment bed can create a reinforcing pattern deposition locations and sites of jam formation. Channel bars,
where wood deposited on the channel bed increases sediment bedforms, and localized bank erosion can all be considered obstacles
deposition around it. Then, as more sediment deposits around the to transport but were not included as part of the model input.
wood, a mid-channel bar develops which acts as a preferential Inclusion would likely have improved the model predictions,
deposition location for more wood. The wood helps stabilize the particularly in areas where channel bends are less likely to be the
channel bar, increasing flow resistance and deflecting flow around the dominant obstacle to log transport.
bedform. If flow is diverted toward channel banks, local bank scour The balance of forces acting on LWD has come under scrutiny as a
occurs and the channel widens to flow around the wood jam and bar means for predicting LWD movement in a channel. When hydrodynamic
(Keller and Swanson, 1979). Alternatively, the LWD may accumulate forces acting on the LWD are greater than resisting forces, LWD will
along the channel boundary and act as a deterrent to increased bank move by rolling, sliding, or floating. The variables required to predict the
erosion, deflecting flow away from the bank. Because of the potential initiation of log movement include wood density, angle of the log with
interaction between LWD, channel form, and channel hydraulics, it is respect to the channel, channel slope, friction between channel bed and
useful to be able to predict stable or recurrent LWD jam locations. log, and the drag coefficient of the log (Haga et al., 2002). The apparent
Two models that were designed to predict LWD mobility were drag acting on an individual log is a function of the depth of submergence
explored with the San Antonio data in an attempt to model the of the log (Hygelund and Manga, 2003). Forces acting on LWD are, in
persistent LWD jam spacing and the flows that would mobilize these turn, related to how the LWD affects the hydrodynamics of a river reach.
jams. Both models were developed from flume research that Some amount of the total shear stress in a channel flow will be
examined the mobility of individual logs. Conflicting evidence exists attributable to LWD and LWD jams (Hygelund and Manga, 2003; Daniels
of whether or not the rate of change of LWD movement occurs when and Rhoads, 2003; Manners et al., 2007). Where logs and jams are stable
the logs are part of the jam. Field evidence has shown both that single in the channel or persistently form in the same locations, the LWD will
pieces of LWD have the same mobility as logs in groups (Hilderbrand have more effect on shear stresses in the channel which in turn influence
et al., 1997) and that jams can increase the residence time of LWD in a the channel morphology and decisions related to channel management.
328 J.C. Curran / Geomorphology 116 (2010) 320–329

7. Conclusion work on the dataset, this analysis would not have been possible. Frank
Engel is also thanked for his help during the fieldwork. The author is
Once wood has been input into a channel, it either remains in place grateful for the thoughtful comments of an anonymous reviewer and
or is transported downstream. When LWD accumulates and develops of Jack Vitek, which have helped improve this manuscript.
into stable jams, it has the potential to affect the channel morphology
and sediment transport processes (Daniels, 2006). Transport of wood
is dependent on physical properties of the wood, the quantity of References
supplied wood, channel hydraulics, and channel shape (Gurnell et al.,
2002). LWD physical characteristics that influence transport include Abbe, T.B., Montgomery, D.R., 1996. Large woody debris jams, channel hydraulics and
habitat formation in large rivers. Regul. Rivers: Res. Manage. 12 (2–3), 201–221.
wood length, density, diameter, orientation relative to flow, and the Abbe, T.B., Montgomery, D.R., 2003. Patterns and processes of wood debris
presence/absence of a rootwad. The two most important parameters accumulation in the Queets river basin, Washington. Geomorphology 51, 81–107.
are ratios of log length to channel width, and channel depth to log Andreoli, A., Comiti, F., Lenzi, M.A., 2007. Characteristics, distribution and geomorphic
role of large woody debris in a mountain stream of the Chilean Andes. Earth Surf.
diameter. A log floats when the moisture content and wood species
Process. Landforms 32 (11), 1675–1692.
combine for a wood density less than water. When a floating log is Benke, A.C., Wallace, J.B., 1990. Wood dynamics in coastal plain blackwater streams.
shorter than the channel width and has a diameter less than half the Can. J. Fish. Aquat. Sci. 47, 92–99.
channel depth, it is likely to transport downstream. Logs and jams that Bilby, R.E., Likens, G.E., 1980. Importance of organic debris dams in the structure and
function of stream ecosystems. Ecology 61 (5), 1107–1113.
transport easily downstream are unlikely to exert influence over Bocchiola, D., Rulli, M.C., Rosso, R., 2006. Transport of large woody debris in the
channel morphodynamics. Those jams that have long residence times presence of obstacles. Geomorphology 76 (1–2), 166–178.
in a channel or re-form at the same location in the channel have the Bocchiola, D., Rulli, M.C., Rosso, R., 2008. A flume experiment on the formation of wood
jams in rivers. Water Resour. Res. 44 (W02408).
potential to alter channel hydraulics and morphology. Braudrick, C.A., Grant, G.E., 2000. When do logs move in rivers? Water Resour. Res. 36 (2),
This project demonstrates the applicability of two models 571–583.
designed around steep, forested channels to low gradient, wide Braudrick, C.A., Grant, G.E., Ishikawa, Y., Ikeda, H., 1997. Dynamics of wood transport in
streams: a flume experiment. Earth Surf. Process. Landforms 22, 669–683.
channels. The distribution and mobility of LWD over 72 km of the San Brockman, F.C., Brockman, C.F., Merrilees, R.A., Latimer, J.P., Challinor, D., Nolting, K.S.,
Antonio River, a low gradient channel in southeast Texas, were 2001. Trees of North America: a Field Guide to the Major Native and Introduced
evaluated through fieldwork, aerial photography, and predicted from Species North of Mexico. Macmillan. revised edition.
Bush, J.K., Van Auken, O.W., 1984. Woody-species composition of the upper San
mobility models. LWD jam locations were identified for 2003 and Antonio River gallery forest. Texas J. Sci. 36, 139–148.
2007. During the four-year period, all the LWD jams were mobilized, Cawthon, T.M. 2007. Log jam characterization, distribution, and stability on the San
including those extending across the channel width. Although easily Antonio River, Texas. Master's Thesis. Texas State University, San Marcos, TX. 73p.
Cawthon, T., Curran, J. C. 2007. Channel change on the San Antonio River, Texas Water
mobilized, 34 LWD jams re-formed in the same locations. Data from
Development Board: Austin, TX. Final report on project 060483068. 52 p.
the San Antonio River are applied to the models of Braudrick and Comiti, F., Lenzi, M.A., 2006. Dimensions of standing waves at steps in mountain rivers.
Grant and Bocchiola, which were developed to predict LWD mobility Water Resour. Res. 42, W03411.
and transport distances. The results indicate that the Braudrick and Daniels, M.D., 2006. Distribution and dynamics of large woody debris and organic
matter in a low-energy meandering stream. Geomorphology 77 (3–4), 286–298.
Grant model was most useful in predicting the trees most likely to act Daniels, M.D., Rhoads, B.L., 2003. Influence of large woody debris obstruction on three-
as a key member in jam formation. When combined with information dimensional flow structure in a meander bend. Geomorphology 51, 159–173.
on the tree distribution of the riparian and source areas for LWD, the Diez, J.R., Elosegi, A., Pozo, J., 2001. Woody debris in North Iberian streams: influence of
geomorphology, vegetation, and management. Environ. Manage. 28 (5), 687–698.
approximate locations of increased wood loading in the river could be Downs, P.W., Simon, A., 2001. Fluvial geomorphological analysis of the recruitment of
predicted. The Bocchiola model uses obstacle spacing to predict the large woody debris in the Yalobusha River network, Central Mississippi, USA.
transport distance of logs in the channel, which corresponds to Geomorphology 37, 65–91.
Faustini, J.M., Jones, J.A., 2003. Influence of large woody debris on channel morphology
locations of persistent LWD jams within the area of the study reach. and dynamics in steep, boulder-rich mountain streams, western Cascades, Oregon.
Obstacle spacing was set to mimic the distance between channel Geomorphology 51, 187–205.
bends, which assumes that meander bends are the dominant obstacle Gippel, C.J., Finlayson, B.J., O'Neill, I.C., 1996. Distribution and hydraulic significance of
large woody debris in a lowland Australian river. Hydrobiologia 318, 179–194.
to log transport. When model results were compared to the measured Gould, F.W., 1962. Texas Plants: a Checklist and Ecological Summary. Texas Agricultural
spacing of LWD jams not in repeat locations, the results agreed less Experiment Station, College Station, TX.
closely with the measured LWD jam spacing. This can be at least Gregory, K.J., Gurnell, A.M., Hill, C.T., 1985. The permanence of debris dams related to
river channel processes. Hydrol. Sci. 30 (3), 371–381.
partially attributed to the temporary nature of these jams and the lack
Gurnell, A.M., Petts, G.E., Hannah, D.M., Smith, B.P.G., Edwards, P.J., Kollmann, J., Ward, J.V.,
of information on the spacing of obstacles other than channel bends. Tockner, K., 2000. Wood storage within the active zone of a large European gravel-bed
Results indicate the Bocchiola model can predict the locations of river. Geomorphology 34, 55–72.
stable (or recurring) LWD jams when the spacing of obstructions Gurnell, A.M., Piegay, H., Swanson, F.J., Gregorys, S.V., 2002. Large wood and fluvial
processes. Freshw. Biol. 47, 601–619.
impeding transport is known. Haga, H., Kumagai, T., Otsuki, K., Ogawa, S., 2002. Transport and retention of coarse
Locating stable and recurring LWD jams is important in low woody debris in mountain streams: an in situ field experiment of log transport and
gradient channels, which are common in agricultural and urban areas. a field survey of coarse woody debris distribution. Water Resouc. Res. 38 (8), 16.
Haschenburger, J.K., Wilcock, P.R., 2003. Partial transport in a natural gravel bed
These channels are often actively managed and the built environment channel. Water Resour. Res. 39, 1020.
intersects the river through roads and bridges. Studies of LWD loading Hilderbrand, R.H., Lemly, A.D., Dolloff, A.l., Harpster, K.L., 1997. Effects of large woody
and jam distribution within channels can inform channel restoration debris placement on stream channels and benthic macroinvertebrates. Can. J. Fish.
Aquat. Sci. 54, 931–939.
projects where LWD are added to a stream to improve or restore Hilderbrand, R.H., Lemly, A.D., Doloff, C.A., Harpster, K.L., 1998. Design considerations for
aquatic habitat. The ability to identify reaches where LWD jams are large woody debris placement in stream enhancement projects. North Am. J. Fish.
more likely to form naturally will result in successful projects that Manage. 18, 161–167.
Hygelund, B., Manga, M., 2003. Field measurements of drag coefficients for model large
simulate natural conditions. If natural distribution in relation to woody debris. Geomorphology 51, 175–185.
stream width, depth, and gradient is not taken into account, then LWD Keller, E.A., Swanson, F.J., 1979. Effects of large organic material on channel form and
could be placed in a reach of the channel where retention and habitat fluvial processes. Earth Surf. Process. Landforms 4, 361–380.
Keller, E.A., Tally, T., 1979. Effects of large organic material on channel form and fluvial
formation (Richmond and Fausch, 1995) will not be successful.
processes in the coastal redwood environment. In: Rhodes, D.D., Williams, G.P.
(Eds.), Adjustments of Fluvial Systems. Kendall/Hunt Publishing Co., Binghamton,
Acknowledgements NY, pp. 167–197.
Kochel, R.C., Baker, V.R., 1982. Paleoflood hydrology. Science 215, 353–361.
Lisle, T.E., 1986. Stabilization of a gravel channel by large streamside obstructions and
The author thanks Tim Cawthon, who identified the large woody bedrock beds, Jacoby Creek, northwestern California. Geol. Soc. Am. Bull. 97,
debris on the aerial photos and through the fieldwork. Without his 999–1011.
 J.C. Curran / Geomorphology 116 (2010) 320–329 329

Magna, M., Kirchner, J.W., 2000. Stress partitioning in streams by large woody debris. Simpson, W., TenWolde, A. 1999. Physical properties and moisture relations of wood.
Water Resouc. Res. 36 (8), 2373–2379. In: Forest Products Laboratory (Ed.), Wood Handbook—Wood as an Engineering
Manners, R.B., Doyle, M.W., Small, M.J., 2007. Structure and hydraulics of natural woody Material, Gen. Tech. Rep. FPL-GTR-113 ed., U.S. Department of Agriculture, Forest
debris jams. Water Resour. Res. 43 (W06432), 17. Service, Madison, WI, pp. 1–24.
Martin, D.J., 2001. The influence of geomorphic factors and geographic region on large Smock, L.A., Metzler, G.M., Gladden, J.E., 1989. Role of debris dams in the structure and
woody debris loading and fish habitat in Alaska coastal streams. North Am. J. Fish. functioning of low-gradient headwater streams. Ecology 70 (3), 764–775.
Manage. 21, 429–440. Soil Survey Staff 2008. Natural Resources Conservation Service, United States
Montgomery, D.R., Buffington, J.M., Smith, R.D., Schmidt, K.M., Press, G., 1995. Pool Department of Agriculture. Web Soil Survey. Available online at websoilsurvey.
spacing in forest channels. Water Resour. Res. 31 (4), 1097–1105. nrcs.usda.gov/ [accessed 07/20/2008].
Montgomery, D.R., Collins, B.D., Buffington, J.M., Abbe, T.B., 2003. Geomorphic effects of Swanson, F.J., Gregory, S.V., Sedell, J.R., Campbell, A.G., 1982. Land–water interactions: the
wood in rivers. American Fisheries Society Symposium, pp. 1–27. riparian zone. In: Edmonds, R.L. (Ed.), Analysis of Coniferous Forest Ecosystems in the
Nakamura, F., Swanson, F.J., 1993. Effects of coarse woody debris on morphology and Western United States. Hutchinson Ross Publishing Co., Stroudsburg, PA, pp. 267–291.
sediment storage of a mountain stream system in western Oregon. Earth Surf. Thompson, D.M., 1995. The effects of large organic debris on sediment processes and
Process. Landforms 18, 43–61. stream morphology in Vermont. Geomorphology 11, 235–244.
Piegay, H., Gurnell, A.M., 1997. Large woody debris and river geomorphological pattern: Wallerstein, N.P., Thorne, C.R., 1996. Impacts of woody debris on fluvial processes and
examples from S.E. France and S. England. Geomorphology 19, 99–116. channel morphology in stable and unstable streams. Project no. WK2Q6C-7592-EN09.
Preston, R.J., 1976. North American Trees. The Iowa State University, Ames, Iowa. Wallerstein, N.P., Thorne, C.R., 2004. Influence of large woody debris on morphological
Richmond, A.D., Fausch, K.D., 1995. Characteristics and function of large woody debris evolution of incised, sand-bed channels. Geomorphology 57, 53–73.
in subalpine Rocky Mountain streams in northern Colorado. Can. J. Fish. Aquat. Sci. Webb, A.A., Erskine, W.D., 2003. Distribution, recruitment, and geomorphic significance
52, 1789–1802. of large woody debris in an alluvial forest stream: Tonghi Creek, southeastern
Robison, E.G., Beschta, R.L., 1990. Coarse woody debris and channel morphology Australia. Geomorphology 51, 109–126.
interactions for undisturbed streams in southeast Alaska, U.S.A. Earth Surf. Process. Wermund, E. G. 1996. Physiographic Map of Texas. 1996. 01.
Landforms 15, 149–156. Wohl, E.E., Goode, J.R., 2008. Wood dynamics in headwater streams of the Colorado
San Antonio River Authority, 2004. Tri-county Flood Study. Photo Science Geospatial Rocky Mountains. Water Resour. Res. 44 (W09429).
Solutions, Lexington, KY.