Effects of Stronach Dam removal on fluvial geomorphology in the Pine River,
Michigan, United States
Bryan A. Burroughs 1
, Daniel B. Hayes , Kristi D. Klomp, Jonathan F. Hansen, Jessica Mistak 2
Department of Fisheries and Wildlife, 13 Natural Resources Building, Michigan State University, East Lansing, MI 48824-1222, USA
a b s t r a c ta r t i c l e i n f o
Article history:
Received 7 September 2008
Received in revised form 20 March 2009
Accepted 24 March 2009
Available online 2 April 2009
Keywords:
Dam removal
Sediment incision
Sediment deposition
Fluvial processes
Impoundment
Channel evolution
Although dam removal has been increasingly used as an option in dam management, and as a river
restoration tool, few studies provide detailed quantitative assessment of the geomorphological response of
rivers to dam removal. In this study, we document the response of the Pine River, Michigan, to the gradual
removal of Stronach Dam. In 1996, prior to the initiation of removal, 31 permanent cross-sectional transects
were established in the 10-km study area. These transects were surveyed annually during the course of the
removal (1996­2003) and for the three years following removal (2004­2006). Dam removal resulted in
progressive headcutting of sediments in the former impoundment, extending upstream 3.89 km of the dam.
Over the course of the 10 years since dam removal was initiated, a net total of 92000 m3
of sediment erosion
occurred. The majority of sediments stored in the former reservoir remained in place, with only 12% of the
estimated reservoir sediment fill being eroded. Approximately 14% of the net erosion was deposited within
the stream channel 1 km downstream of the dam location, with the remainder being transported further
downstream or deposited in the floodplain. Sediment fill incision resulted in a narrower and deeper channel
upstream, with higher mean water velocity and somewhat coarser substrates. Downstream deposition
resulted in a wider and shallower channel, with little change in substrate size composition. Counterintuitively,
water velocity also increased downstream because of the increased slope that developed. Prior to
removal, bedforms in the former impoundment were dominated by runs but are showing signs of restoration
toward reference conditions. Continuing changes in river geomorphology are evident even three years
following removal and are likely to occur for years to come.
 2009 Elsevier B.V. All rights reserved.
1. Introduction
Approximately 2.5 million dams have been built in the United
States (National Research Council, 1992) on nearly every major river
system in the lower 48 states (The Heinz Center, 2002). The period
between 1950 and 1970 was marked by the most intensive dam
construction efforts (The Heinz Center, 2002) with limited understanding
of their impacts to rivers. Following this period, the body of
scientific evidence documenting the drastic effects that dams have on
river systems grew substantially. Today, a preponderance of evidence
exists describing the multitude of ways dams alter river functioning,
including alterations to the flow and temperature regimes; shifts in
sediment, nutrient, and energy transport disruption; and numerous
biological implications (e.g., Hammad,1972; Petts,1980; Williams and
Wolman, 1984; Cushman, 1985; Bain et al., 1988; Ward and Stanford,
1989; Benke, 1990; Lessard and Hayes, 2003; Ligon et al., 1995; Collier
et al., 1996; Shields et al., 2000).
Many dams continue to fulfill their intended purpose, providing
social and economic benefits. However, as dams age they require
maintenance to prolong their function and safety. Now, a large and
growing number of dams exist that no longer fulfill their intended
purpose and may not sustain sufficient benefits as to outweigh the
negative ecological impacts they cause.
Of the estimated 2.5 million dams in the U.S., 76000 are 1.83 m or
greater in height (Federal Emergency Management Agency (FEMA)
and U.S. Army Corps of Engineers (USACE), 1996). Of these 76000
dams, 80% or 60 000 are expected to be 50 years of age or older by the
year 2020 (FEMA and USACE,1996). The average design life expectancy
of dams is ~50 years, implying that a large number of dams in the U.S.
will be in need of maintenance or considered for removal (River
Alliance of Wisconsin and Trout Unlimited, 2000). Over the last several
decades, the rate at which dams have been removed in the U.S. has
risen from approximately one per year during the 1960s to approximately
20 per year during the 1990s (Pohl, 2003). The abundance of
aging dams and the increasing rate of dam removal indicate that
removal of dams will become increasingly common in the future.
Geomorphology 110 (2009) 96­107
 Corresponding author. Tel.: +1 517 432 3781; fax: +1 517 432 1699.
E-mail addresses: burrou15@msu.edu (B.A. Burroughs), hayesdan@msu.edu
(D.B. Hayes), klompkri@voyager.net (K.D. Klomp), hansen53@msu.edu (J.F. Hansen),
mistakjl@michigan.gov (J. Mistak).
1
Present address: Michigan Council of Trout Unlimited, P.O. Box 442, Dewitt, MI
48820-8820, USA.
2
Present address: Michigan Department of Natural Resources, Marquette Fisheries
Station, 484 Cherry Creek Rd., Marquette, MI 49855, USA.
0169-555X/$ ­ see front matter  2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2009.03.019
Contents lists available at ScienceDirect
Geomorphology
journal homepage: www.elsevier.com/locate/geomorph
Over 400 dams have been removed to date in the U.S. (Pohl, 2003),
but the scientific literature on the effects of dam removal is still sparse.
Much of the existing literature focuses on the administrative, legal,
and socioeconomic aspects of executing dam removals (Born et al.,
1998; River Alliance of Wisconsin and Trout Unlimited, 2000; Smith
et al., 2000, Graber et al., 2001; Trout Unlimited, 2001; Bowman,
2002; Johnson and Graber, 2002). Using analogies from various
disciplines, some researchers have developed general hypotheses of
river ecosystem responses to dam removals (Doyle et al., 2002; Gregory
et al.,2002; Pizzuto,2002; Shafrothet al., 2002; Stanleyand Doyle, 2002;
Whitelaw and MacMullan, 2002), while others have hypothesized the
outcomes of specific proposed dam removals (Shuman, 1995; Freeman
et al., 2002; The Heinz Center, 2002). Despite an emerging conceptual
basis for the effects of dam removals, this field continues to lack the
empirical information that is needed to verify these hypotheses,
calibrate preexisting models for use with dam removal, and generate
novel insights into the effects of dam removal (Bushaw-Newton et al.,
2002; Doyle et al., 2002; Graf, 2003; Hart et al., 2003). Qualitative
observations on the effects of dam removal exist for several dam
removal case studies (American Rivers et al., 1999; Smith et al., 2000),
and several quantitative studies exist on the effects of dam removal on
fluvial geomorphology (Evans et al., 2000; Wohl and Cenderelli, 2000;
Bushaw-Newton et al., 2002; Stanley et al., 2002; Chaplin, 2003;
Wildman and MacBroom, 2005), aquatic insects (Bushaw-Newton et al.,
2002; Stanley et al., 2002), and fish (Hill et al.,1994; Kanehl et al., 1997;
Bushaw-Newton et al., 2002). While these studies provide unique
Fig. 1. Photograph of Stronach Dam in 2001, during gradual removal.
Fig. 2. Location of Stronach Dam and Pine River in relation to the state of Michigan, and the location of permanent cross-sectional surveying transects within the study area of Pine
River.
97B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
insights into the outcomes of dam removals, many were relatively short
in time duration (i.e., 1­2 years post-dam removal), and the empirical
information on the effects of dam removal is still very limited (e.g., Graf,
2003; Doyle et al., 2005).
The goal of this study was to document the effects of dam removal
on fluvial geomorphology. In particular, this study was designed to
address questions such as what types of changes occur in rivers
following dam removal, what are the magnitudes of these changes,
what is the spatial extent of change, and how long do these changes
take to occur? Answers to questions such as these should lead to more
informed decision making processes regarding dam removals,
improved predictions on the outcomes of dam removal, and
improvements in how dams are removed in the future.
To answer these questions, we monitored several aspects of fluvial
geomorphology from 1995 through 2006 before, during, and after the
gradual removal of Stronach Dam on the Pine River, Manistee County,
Michigan. The specific objectives of this study were to (i) document
the spatial and temporal extent of sediment erosion, transport, and
deposition that occurred because of the dam removal; and
(ii) document the changes in river morphology attributes (i.e., slope,
width, depth, water velocity, substrate composition, and bedform
(riffle-pool) diversity) that occurred because of the sediment erosion,
transport, and deposition. The 12 years of detailed quantitative
monitoring of the outcomes of this dam removal provides a unique
data set useful for validating and refining existing hypotheses about
the geomorphologic effects of dam removals, generating novel
hypotheses, and leading to improvements in the ongoing practice of
removing dams.
2. Regional setting
Stronach Dam was located on the Pine River, a tributary to the
Manistee River, in the northwestern Lower Peninsula of Michigan
(Fig. 1). The Pine River is a 77-km-long, fourth-order stream and
drains a 68635-ha watershed dominated by sandy glacial outwash
plains, recessional moraines, and areas of consolidated clay (Hansen,
1971; Rozich, 1998). Mean daily discharge recorded at two U.S.
Geological Survey (USGS) gaging stations on the Pine River (8 and
13.7 km upstream from Stronach Dam) averaged 8.10 m3
/s during
34 years of record, with a minimum discharge of 4.56 m3
/s, a
maximum of 69.09 m3
/s, and an average annual ratio of low to high
mean monthly flows of 2.02, indicating "stable to very stable" flows
(Rozich, 1998). The river carries a large bedload of sand because of the
local geology and extensive logging operations of the late 1800s,
which exacerbated bank instability along the river. Hansen (1971)
estimated a mean annual sediment discharge of 50 000 tons at
Stronach Dam from 1967 to 1970, which was 70­75% sand.
Stronach Dam was constructed from 1911 to 1912, 5.6 km upstream
from the confluence of the Pine River and the Manistee River (Fig. 2).
Stronach Dam was originally a hydroelectric dam with an earth
embankment and concrete corewall, a 4.57-m fixed-concrete spillway
section with 0.91 m of flashboards on top of the spillway and a
concrete and brick powerhouse with two turbine bays (Consumers
Power Company, 1994). Stronach Dam, with 5.49 m of head height
possible, was operated mostly around 5.18 m of head. This created a
26.7-ha reservoir with a 789428-m3
volume capacity (Hansen, 1971;
Consumers Power Company, 1994). Tippy Dam (17.07 m head height),
also a hydroelectric dam, was constructed in 1918 immediately
downstream of the confluence of the Pine and Manistee Rivers
(Rozich, 1998) (Fig. 2), which created a 428-ha, 48722530-m3
impoundment and impounded water upstream to Stronach Dam.
Because of the Pine River's large sediment load, the Stronach Dam
reservoir quickly filled with sediment and problems arose with the
operation of the dam's turbines. Attempts made in the 1930s to remove
the accumulation of sediment behind the dam were only marginally
successful and eventually became uneconomical (Consumers Power
Company, 1994). In 1953, 41 years after construction, Stronach Dam
was decommissioned as a hydroelectric dam and the river flow was
directed over the spillway, shifting the location of the active river
channel to the opposite side of the dam. The spillway flashboards were
removed gradually over the following years with the last removed in
1983 (Consumers Power Company, 1994).
In the early 1990s, removal of Stronach Dam was negotiated as part
of a Federal Energy Regulatory Commission (FERC) settlement
agreement in the relicensing of Tippy Dam, owned by Consumers
Power Company, the same company as Stronach Dam. A "staged," or
gradual, removal was decided upon in order to allow gradual river
channel adjustments with the least amount of environmental impact
risk, at the lowest cost, and without impacting the operation of Tippy
Dam (Battige et al., 1997). In 1996, a 3.6-m-high "stop-log" structure
was installed in the old powerhouse to allow a gradual drawdown of
the river. The stop-log structure consisted of hollow metal pipes
(15 cm diameter) stacked one on top of another, with a metal grate
called a "trash-rack" immediately upstream to protect the stop logs
from debris impingement. The original removal schedule called for
one stop-log to be removed every three months, for a total of 0.60 m/
year, over the course of six years (with corresponding trash-rack
removal). This plan was altered because of recreational safety
concerns, feasibility issues, and technical difficulties with removal
(D.S. Battige, Consumer Power Company, personal communication,
2002). The actual sequence of the staged dam removal occurred
between 1997 and 2003 and is shown in Table 1.
3. Materials and methods
In 1995, two years prior to the commencement of dam removal
activities, Pine River was assessed to document the spatial extent of
impoundment effects from Stronach Dam. This assessment involved
the surveying and description of physical characteristics, including
categorization of the stream into bedform units of runs, riffles, pools,
rapids, or complexes (a designation where more than one category
applied), following the criteria developed by Hicks and Watson (1985).
The assessment of stream bedforms allowed detection of impoundment
effects well upstream of the readily noticeable reservoir area.
This impoundment area of the river extended for 3.89 km upstream of
Table 1
Schedule of removal events during the staged removal of Stronach Dam on Pine River,
Manistee County, Michigan.
Date Number of stop logs removed Meters of trash rack removed
March 17, 1997 1 (0.15) 0 (0)
June 5, 1997 1 (0.30) 0 (0)
June 16, 1997 2 (0.61) 0 (0)
June 24, 1997 2 (0.91) 0 (0)
September 15, 1997 1 (1.07) 0 (0)
December 15, 1997 1 (1.22) 0 (0)
March 16, 1998 1 (1.37) 0 (0)
May 7, 1998 0 (1.37) 1.83 (1.83)
May 29, 1998 0 (1.37) 0.30 (2.13)
June 15, 1998 1 (1.52) 0 (2.13)
September 8, 1998 1 (1.68) 0.30 (2.44)
December 14, 1998 1 (1.83) 0.30 (2.74)
March 15, 1999 1 (1.98) 0 (2.74)
May 11, 1999 1 (2.13) 0 (2.74)
September 13, 1999 2 (2.44) 0 (2.74)
September 16, 1999 0 (2.44) 0.61 (3.35)
April 17, 2000 2 (2.74) 0 (3.35)
October 2, 2000 2 (3.05) 0 (3.35)
October 5, 2000 0 (3.05) 0.61 (3.96)
May 8, 2001 2 (3.35) 0 (3.96)
September 8, 2001 2 (3.66) 0 (3.96)
November 11, 2002 0 (3.66) 1.52 (5.49)
December 2003 Remaining spillway and dam superstructure removed
Stop logs are 15.24 cm diameter hollow metal pipes stacked on top of one another. Trash
rack removal estimates are approximate. Cumulative meters removed are in
parentheses (Dave Battige, Consumers Energy, personal communication 2003).
98 B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
Stronach Dam and was typified as being relatively wide, slow-flowing,
and sand-bottomed and generally consisted of only run bedform units.
A 3.7-km reach upstream from the impoundment, which exhibited
no effects from the impoundment, was chosen as a reference zone.
The river was narrower, faster-flowing, had coarser substrates, and
showed high bedform heterogeneity. A third study zone was chosen
downstream of Stronach Dam, where the river was wide, very slowflowing,
and sand-bottomed and consisted entirely of run bedforms.
Prior to the removal of Stronach Dam, water was impounded in this
study zone by Tippy Dam Reservoir, and the study zone extended for
only 0.63 km downstream of Stronach Dam. During the dam removal
process, slope increased in this zone and the impoundment of water
by Tippy Dam was not observed for 2.55 km downstream of Stronach
Dam. In 2002, this downstream study zone was lengthened from 0.63
to 2.55 km.
Thirty-one permanent cross-sectional transects were established
in 1996 to document changes in river channel morphology over the
course of dam removal. Ten additional transects were added in 2002
to fill in longitudinal gaps in areas of geomorphologic interest and to
extend the downstream zone. The vertical and horizontal location of
all cross-sectional transect endpoints were permanently referenced
using rebar rods and spikes in witness trees. Thirteen transects were
located in the upstream reference reach, 21 transects were located in
the impoundment, and seven transects were placed in the downstream
zone (Fig. 2). Photographs, site descriptions, and latitudelongitude
coordinates for all transects are archived at Michigan State
University, Department of Fisheries and Wildlife. A typical photograph
and stream cross-section for each of the study zones is provided in
Fig. 3. All transects were linked to elevations above sea level based on
a USGS monument. The distance of each transect from Stronach Dam
was determined by floating the river in a canoe using the trip-log
feature of handheld Garmin GPS units on several occasions and
averaging the results. All transects were surveyed annually from 1996
through 2006, during June to early July of each year. At each transect,
streambank elevations were recorded at points of slope change to
document streambank morphology. Streambed elevations were
Fig. 3. Photographs and channel cross-section diagrams illustrating typical channel forms within each of the three study zones. Dashed lines indicate streambed and water surface
level in 1996, prior to initiation of dam removal; solid lines represent post-dam removal, in 2006.
99B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
recorded every 0.61 m across the wetted width of the stream, and
water surface elevation was measured once per transect. All
measurements were taken at the same locations each subsequent
year of the study. Elevations were recorded to the nearest 3 mm.
Channel geometries were expected to change following dam
removal because of sediment incision or deposition processes. Therefore,
at each cross-sectional transect we defined a constant baseline
elevation above which no channel changes occurred (i.e., an arbitrary
plane above the top of the bank prior to removal) in order to estimate
overall changes in channel cross-sectional areas. The trapezoidal rule
for numerical integration (Press et al.,1992) was used to estimate areas
between each pair of cross-sectional points surveyed in a transect, and
the sum of these was the transect cross-sectional area estimate.
Change in transect cross-sectional area reflects the net amount of
erosion or deposition that occurred between surveying events.
Estimates of sediment volume transported during dam removal were
calculated using the trapezoidal rule for integration (given the
amounts of cross-sectional area change at transects) and the distances
between transects. Water slope was calculated as the change in water
surface elevation between two transects over the river distance
between those two transects. Width (W) and mean depth (D) were
calculated for each transect from the wetted channel dimensions. Bank
slopes for each side of each transect were calculated as the difference in
elevations from the top bank to the bank toe, divided by the lateral
distance between those two points (e.g., Duan, 2005). These points,
including the top bank elevations, were allowed to change each year as
channel profile changed. The change in bank slope, compared to the
initial pre-dam removal slopes, was calculated and averaged for each
transect.
Water velocity was measured at 10 of the permanent transects
(Fig. 2) annually from 1996 to 2006. From 1996 to 2000 a MarshMcBirney
Model 201 portable current meter was used and from 2001
to 2006 water velocity was measured using a Global Flow Probe Model
FP101. Water velocity was measured at 0.61-m intervals across the
wetted width of the stream. If water depth was b0.75 m, water
velocity was measured at 60% of the water depth from the water
surface. If water depth was N0.75 m, water velocity was measured at
20% and 80% of the water depth from the water surface, and the two
measurements were averaged (Gallagher and Stevenson, 1999).
Water velocities were measured annually in June during transect
surveying. In 1996, the average discharge during these measurements
was 7.70 m3
/s. In 2006, because of continuing precipitation events,
measurements were taken at an average discharge of 9.29 m3
/s,
significantly higher than 1996. Because of the influence of discharge
on water velocity, we chose to use measurements taken in 2005, when
discharge was 8.07 m3
/s, for comparison with 1996 data. Water
velocity was not surveyed in all downstream zone transects in 1996 or
1997, so data from 1998 (when discharge was 6.94 m3
/s) were used
for comparison purposes.
The Kolmogorov­Smirnov (K­S) two-sample test (Steel and Torrie,
1980) was used to test for differences between water velocity
frequency distributions between years. The K­S test was used because
it evaluates the observed difference between two distributions, not
just differences in means.
Streambed substrate size composition was measured at 10 of the
permanent transects in 1996, at each of the 31 original permanent
transect sites annually from 1997­2006, and also at the 10 additional
sites from 2002 through 2006. A modified pebble count method was
used (Wolman, 1954; Kondolf and Li, 1992) to determine substrate
size composition. We sampled 100 streambed particles systematically
along each transect, measuring the intermediate axis and assigning a
size class code to each particle (from a modified Wentworth scale)
(Wentworth, 1922; Cummins, 1962). Systematically sampling the
substrate allowed linkage to other morphological data (e.g., size of
substrate at places of erosion or deposition within a transect cross
section) in addition to providing a measurement of the size structure
of the substrate. Median substrate size (D50) for a transect was
calculated after excluding "organic" or "trash" designations that did
not have corresponding size classes. The K­S two-sample test (Steel
and Torrie, 1980) was used to test for differences in substrate size
frequency distributions between years.
In 1995, two years prior to dam removal, bedforms (also commonly
referred to as "mesohabitat" or "channel unit types") were mapped for
all three study zones. This assessment involved the categorization of
the stream into bedforms, following the criteria developed by Hicks
and Watson (1985). The length (along the thalweg) and width of each
bedform were measured, as well as periodically recording latitude and
longitude of selected bedforms for mapping purposes. In 2004, the
year following complete removal of the dam, this assessment was
repeated to document changes in the frequencies of bedforms in each
study zone. This assessment was conducted at low flow levels to aid in
the delineation of bedform units.
4. Results
4.1. Sediment transport
During and following the staged removal of Stronach Dam,
significant sediment transport occurred in both the impoundment and
the downstream zones, while the upstream reference reach remained
remarkably stable. Typical of a stable stream, the upstream reference
reach experienced net erosion some years while net deposition occurred
in others (Fig. 4). The average annual net sediment transport in the
reference was 289 m3
of erosion, with a standard deviation of 1946 m3
.
This translates to only ~4 mm average annual vertical change for each
square meter of streambed in the upstream reference reach.
The total volume of sediment eroded from the impoundment
during the staged removal through three years post-removal was
~92000 m3
. The annual volume of sediment eroded from the
impoundment averaged 9159 m3
and was quite variable with a
standard deviation of 6947 m3
and a range from ~0­21 000 m3
(Fig. 4).
In 2006, no net erosion occurred in the impoundment. Although the
amount of mean annual sediment erosion was large, it was still
substantially less than the mean annual bedload estimated by Hansen
in 1971 at the Stronach Dam site. Hansen estimated that over three
years, the mean annual bedload was 50 000 tons, or ~28000 m3
(using
a density of 1.8 g/cm3
for sand; Morris and Fan, 1998). This makes the
total net amount of sediment erosion from the former impoundment
over the 10 years during and post-dam removal roughly equivalent to
about 3.5 years of annual sediment bedload during the time of
Hansen's estimates.
The annual volume of sediment eroded from the impoundment
was not correlated to the amount of dam removed between sampling
events (R2
=0.04), annual mean flows (R2
=0.07), annual peak flows
Fig. 4. Net annual sediment erosion or deposition volumes during and after the staged
removal of Stronach Dam (1997­2003). Positive values correspond to net erosion;
negative values correspond to net deposition.
100 B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
(R2
=0.00), days at or above 1.5-year recurrence discharge (R2
=0.01),
or approximate stream power in the impoundment reach (factoring in
cumulative mean flows and changing stream slopes) (R2
=0.12).
The amount of erosion, or change in transect cross-sectional area
that occurred in the impoundment varied spatially with distance from
the dam (Fig. 5). In general, greater amounts of erosion occurred
closer to the dam site with the magnitude of erosion attenuating
upstream. During the first several years of the removal, erosion
progressed upstream only through the easily recognizable former
reservoir (1.21 km), and it was not until 2001­2002 that net erosion
was documented at the farthest upstream extent of the original
impoundment, 3.89 km from the dam. A pair of transects, 0.39 and
0.50 km upstream of the dam, provide an exception to the longitudinal
pattern of diminishing erosion with distance upstream of the dam.
These two sites experienced greater amounts of erosion than other
sites closer to the dam because of large amounts of lateral erosion in
addition to vertical erosion.
Each year of the dam removal, with the exception of 2000, some
amount of sediment moving downstream from the impoundment was
deposited and retained in the first 0.63 km downstream of the dam
(Fig. 4). The volume of sediment that was retained and not transported
farther downstream varied considerably between years (average=
1360 m3
, standard deviation=1518 m3
), totaling 13599 m3
by 2006.
The remainder of the 92000 m3
of sediment eroded from the
impoundment was either transported farther downstream in the river,
eventually being deposited in Tippy Dam Reservoir or onto the
floodplain downstream of the dam during high flow events.
4.2. Channel geometry
In the upstream reference reach, width and width-to-mean-depth
ratio (w/d) of the wetted stream channel remained stable (Figs. 6
and 7). In the impoundment, the width of the wetted channel
generally decreased (Fig. 6), with the magnitude of this narrowing
generally corresponding to the amount of erosion that occurred at a
transect (greatest closest to the dam and diminishing upstream
through the impoundment). Localized differences in geology and
slope influenced this general pattern at some transects, however.
Change in the w/d ratio of the impoundment did not show any easily
discernable patterns (Fig. 7). In the downstream reach, width and w/d
ratio both increased (Figs. 6 and 7).
In the upstream reference reach, bank slopes at some transects
remained stable while several others were quite dynamic, showing
both increases and decreases in slope (Fig. 8). Because the streambed
in this reach was stable during the study period, these changes are
likely due to lateral erosion processes common to streams in this area.
Bank slopes in the impoundment gradually increased during the dam
removal, with the greatest increases seen closest to the dam (Fig. 8). In
2006, bank slopes farther upstream in the impoundment reach began
to decrease, while those closer to the dam site remained steepened. In
the downstream reach, bank slopes remained relatively stable (Fig. 8).
Fig. 5. Longitudinal pattern of change in cross-sectional area during the staged removal
of Stronach Dam (1997­2003) compared to before dam removal (1996). Positive values
correspond to deposition within a transect; negative values correspond to erosion at a
transect.
Fig. 6. Longitudinal pattern of change in wetted stream width from pre-dam removal
conditions in 1996. Only 2001and 2006, 5 and 10 years after the beginning of dam
removal, are shown for simplicity and clarity.
Fig. 7. Longitudinal pattern of change in wetted stream width to mean depth ratio (w/d)
from pre-dam removal conditions in 1996. Only 2001 and 2006, 5 and 10 years after the
beginning of dam removal, are shown for simplicity and clarity.
Fig. 8. Longitudinal pattern of change in bank slopes from pre-dam removal conditions
in 1996.
101B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
4.3. Water slope and velocity
As a result of the sediment erosion and deposition processes that
occurred, water slope in the impoundment and downstream reaches
increased during dam removal (Fig. 9). Increases in slope were
greatest in the first 1.59 km upstream of the dam removal, but were
observed to a lesser degree in most of the former impoundment. Slope
increased in the entire impoundment from 0.13% in 1996 to 0.21% in
2006 (with 0.26% slope in the first 1.59 km upstream of the dam site in
2006). The slope in the downstream zone also increased from 0.06% in
1996 to 0.10% in 2006. Slope in the upstream reference reach remained
at 0.16% in both 1996 and 2006.
Prior to dam removal, mean water velocities generally decreased in
a downstream direction through the impoundment and downstream
zone (Fig.10). With the dam removal and increased slope, mean water
velocity generally increased in both the lower impoundment and the
downstream zone, with some of the highest mean water velocities
found in the impoundment after dam removal. Because of localized
differences in slope and channel morphology, mean water velocities in
the impoundment were more variable after dam removal compared to
before dam removal.
The frequency distributions of water velocities were compared for
each zone in both the first and last year of sampling. The water
velocity frequency distribution for the upstream reference reach was
not significantly different in 2005 compared with 1996 (reference: KS
test Dmax =0.227, n=77, pN0.01). Water velocity frequencies in the
impoundment were significantly faster in 2005 than in 1996
(impoundment: K­S test Dmax =0.247, n=137, pb0.01). The water
velocity frequency distribution for the downstream zone also was
significantly faster in 2005 than in 1998 (downstream: K­S test
Dmax =0.235, n=107, pb0.01). In the impoundment and downstream
zones, frequencies of water velocities b0.3 m/s were relatively
unchanged (representing the slower water velocities found at the
stream margins), while the magnitude of thalweg water velocities
increased (Fig. 11). The former impoundment is now the only study
reach to contain water velocities N1.2 m/s.
4.4. Substrate
Median (D50) substrate sizes and substrate size frequency distributions
were compared for each zone in both 1997­1998 and 2005­2006,
the first and last two years of sampling. The first and last two years were
averaged to reduce the influence of annual variability in substrate
composition measurements in comparisons. Before, during, and after
the removal of Stronach Dam, substrate size composition in the Pine
River showed considerable spatial variability and patchiness (Fig. 12).
Prior to removing the dam, substrates were generally coarser and more
heterogeneous in the upstream reference reach than in the impoundment
and downstream. The impoundment was mostly fine sediments
(such as sand) with a few patches of small gravel, while the downstream
zone was almost totally dominated by sand. During the dam removal
process, median substrate size in the reference decreased slightly
(reference: D50 1997­1998=34.8 mm="very coarse gravel," 2005­
2006=29.5 mm="coarse gravel"), but the overall substrate size
frequency distribution did not change significantly (reference: K­S test
Dmax =0.063, n1 =989, n2 =1122, pN0.01) (Fig. 13). Median substrate
size increased slightly in the impoundment (Impoundment: D50 1997­
1998=7.9 mm="fine gravel," 2005­2006=9.7 mm="medium
gravel") (Fig. 12). The substrate size frequency distribution for the
impoundment was significantly coarser in 2005­2006 than for 1997­
1998 (impoundment: K­S test Dmax =0.087, n1 =1574, n2 =1567,
pb0.01), with increased frequencies of large gravel (12­48 mm)
(Fig. 13). Median substrate size increased slightly in the downstream
zone (downstream: D50 1997­1998=1.0 mm="sand," 2005Fig.
9. Longitudinal profile of water slopes or gradients. Vertical scale exaggerated
relative to horizontal scale.
Fig. 10. Longitudinal pattern of mean water velocity (m/s) in Pine River.
Fig. 11. Cumulative percent frequency distributions for water velocities in each study
zone of Pine River.
Fig. 12. Longitudinal pattern of median substrate size in Pine River, comparing 1997­
1998 to 2005­2006.
102 B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
2006=2.3 mm="very fine gravel"), but the overall substrate size
frequency distribution for the downstream zone did not change
significantly (downstream: K­S test Dmax =0.119, n1 =191, n2 =184,
pN0.01) (Figs. 12 and 13).
4.5. Bedform
In 1995, the upstream reference reach contained the highest
heterogeneity of bedforms, and both the impoundment and downstream
zones were comprised almost exclusively of run bedforms. In
2004, the reference reach had changed little from the 1995 survey, but
some differences in the percentages of the pool/complex and rapid
designations and minor differences in the percentages of run
bedforms occurred (Table 2). These changes were most likely due to
differences between sampling crews in the designations of complex
bedforms versus run or pool bedforms and do not likely represent any
real changes. In 2004, the impoundment had higher percentages of
riffles and pools and had lower percentages of runs than in 1995
(Table 2). The downstream zone remained overwhelmingly run
bedform in the 2004 survey, gaining only one pool unit (Table 2).
5. Discussion
5.1. Impoundment
The removal of each stop log during the dam removal process
resulted in an abrupt increase in water slope and water velocity at the
dam. The increase in shear stress created by stop log removal resulted
in erosion of the impoundment sediments immediately upstream of
the dam. As sediments were eroded adjacent to the dam and transported
downstream, the process of sediment fill incision progressed
upstream. Our onsite experience suggested that most of the sediment
fill incision occurred rapidly, within hours to days, similar to
observations by Wildman and MacBroom (2005). Each stage in the
removal process pushed sediment fill incision progressively farther
upstream, eventually reaching the upstream boundary of the original
impoundment.
Relatively small amounts of sediment eroded during the first two
years of the removal, likely because of the low slope of the impoundment
during that period. As the removal progressed, the slope increased
in the impoundment and, on average, more sediment erosion resulted
each year -- with the second and third highest amounts of erosion
coming in the first two years following completion of the dam removal.
These amounts did not correlate well with simple mean flows, peak
flows, time duration over bankfull flow, approximate stream power
(discharge and slope), or even the height of dam removed during each
stage. Further explanation of the temporal variability of sediment
erosion would likely require a site-specific approach, incorporating
transect level differences both in space and time.
Despite the variability and complexity in estimating yearly
amounts of erosion, the total volume of sediment that is likely to
erode following a dam removal seems to be fairly easily estimated. Not
all of the sediment in a reservoir will be mobilized. In Pine River, the
size of the reservoir that was filled with sediment was 789428 m3
.
However, as the stream channel eroded vertically through this
sediment fill, the width of the wetted channel decreased and became
very close to the average width of the upstream reference. The
underlying slope of the reservoir was only slightly higher than the
upstream reference and, consequently, the mean width of the stream
channel in the impoundment (17.6 m) became similar to the mean
width of the stream in the reference (16.9 m). Therefore, the volume
of sediment to be eroded because of dam removal could be estimated
by (HL/2)W, where H is maximum height of the sediment fill, L is
the longitudinal distance of the sediment fill, and W is the average
width of stream immediately upstream of impoundment effects
(Fig.14). For reservoirs such as Stronach Dam, where the reservoir was
completely filled with sediment, H would equal the dam height and L
would equal the length of the impoundment (delineated with a
bedform survey). For reservoirs not completely filled with sediment,
H would equal the maximum height of the sediment fill (usually at
the downstream or leading edge of the sediment delta) and L would
equal the length of the sediment fill (upstream extent delineated with
bedform survey and downstream extent with a bathymetric survey).
To produce an area estimate for the triangular-shaped sediment fill,
HL is divided by 2. In the case of Stronach Dam, this yields an
estimate of (3.66 m3800 m/2)16.9 m=117522 m3
. Through 2006,
~92000 m3
eroded from the impoundment with no new net erosion
occurring during 2006. Assuming that no additional erosion will occur
after 2006, this estimation method overpredicted the amount of
sediment eroded, partly from the imperfect triangular shape of the
impoundment and partly from the rectangular cross-sectional shape
our approximation assumes (Fig. 14). However, this simple estimation
method provides a useful and precautionary estimate of the volume of
sediment that is likely to erode. If significant tributaries entered the
impoundment, this method would have to be applied to these as well
and the results combined with the estimates from the main river
channel.
Prior to the removal of Stronach Dam, the former impoundment
was estimated to contain approximately 789000 m3
of sediment. Of
this stored sediment, we estimated a total of 92000 m3
was eroded in
the 10 years since the initiation of dam removal. Thus, only 12% of the
reservoir sediment was mobilized and transported downstream.
Similarly small percentages of reservoir sediment fill mobilization
were found by Evans et al. (2000) in an Ohio dam failure (9­13%) and
Doyle et al. (2003) in a Wisconsin dam removal (8­14%). The amount
of sediment to be mobilized with a dam removal is an important
consideration during the planning and decision making process in
removing dams (Doyle et al., 2002; Randle, 2003; Rathburn and Wohl,
2003, Wildman and MacBroom, 2005). In situations where sediment
transport downstream from dam removals is undesirable, sediment
removal or management can add considerably to the expense of
removing dams. Given the small percentage of the reservoir sediment
fill that will likely be transported following a dam removal, managing
sediment transport downstream of a dam removal (e.g., through the
use of sediment traps or collection devices) may be more cost effective
than removing larger amounts of sediment from the impoundment
prior to dam removal (e.g., through dredging).
As the streambed incised in the impoundment and slope increased,
the wetted width of the stream also generally decreased. Differences
in the localized geology and slope between transects affected how
much stream width decreased, but the average stream width in the
impoundment became remarkably similar to the average stream
width of the reference. These localized differences in geology and
Fig. 13. Cumulative percent frequency distributions for substrate size compositions in
each study zone of Pine River, comparing beginning size distribution to final size
distribution.
103B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
slope exerted greater influence on water depth at transects and led to
large differences in w/d ratio changes between transects.
Changes in the slopes of stream banks following dam removal have
been predicted to follow Channel Evolution Models (CEMs) (Simon,
1989; Pizzuto, 2002; Doyle et al., 2003; Doyle et al., 2005; Wildman
and MacBroom, 2005). These models were developed from incising
channels and predict that, following dam removal, bank slopes in
former impoundments should increase along with vertical incision
(and so should be steeper, initially, closest to the dam). Banks should
continue to steepen with further incision until a point is reached
where the slope is too great for the cohesive forces of the sediment or
vegetation to continue holding it together, causing slumping and a
reduction in bank slope and allowing for the development of
equilibrium channel dimensions. In the Stronach Dam impoundment,
bank slopes did increase gradually during dam removal, with the
initial and greatest increases occurring closest to the dam. Bank slopes
continued to increase during the last year of the study, and decreases
in bank slopes were observed only at sites farther upstream in the
impoundment where streambed erosion had ceased. However, bank
slopes in the reference reach also exhibited changes of similar
magnitude (albeit different direction) during the study, indicating
bank slopes in Pine River may be naturally dynamic and variable.
While some patterns consistent with the CEM seem discernible from
these data, the natural variability in bank slopes makes interpreting
the significance of those patterns difficult.
As stream slope increased in the impoundment following dam
removal, so did mean water velocities. Comparison of water velocities
across years is partially confounded with differences in discharge.
Although we were able to sample during periods when differences in
discharge were small (i.e., b5%), this still could have resulted in
changes to mean water velocity. Of more interest, however, is how the
frequencies of water velocities changed in the impoundment. The
frequency of slower water velocities (i.e., b0.30 m/s), located primarily
near the stream margins, did not change substantially despite
increasing mean velocities and increased thalweg velocities. Water
velocities therefore not only increased but became more diverse.
Changes of this nature will have important implications for sediment
dynamics and should be beneficial in providing diverse habitat
conditions for different species and life stages of aquatic invertebrates
and fish.
Average substrate size increased throughout most of the impoundment
in response to higher slope and water velocities. Larger transient
changes in median substrate size were seen during dam removal,
however. Apparently during years when less than roughly 15000 m3
of sediment eroded from the impoundment, substrate coarsening
progressed in both the impoundment and downstream reaches
(Fig. 15). However, during years when large volumes of sediment
(N15000 m3
) were eroded from the impoundment (1999­2000 and
2003­2004), fine sediment was transported through these reaches,
covering up previously coarsened substrate and decreasing median
substrate size (Fig. 15). An exception to this trend occurred in 2006,
three years after the removal, when no net sediment erosion occurred
and median substrate size still decreased.
Overall increases in the proportions of coarse substrate were
similarly observed following the removal of Woolen Mills Dam on the
Milwaukee River, Wisconsin, USA (Kanehl et al., 1997). However,
substrate size frequencies also showed changes similar to those of
water velocities. In the impoundment where substrates had been
dominated by sand before dam removal, frequencies of sand decreased
because of the dam removal, but frequencies of silt did not decrease.
Similar changes were reported by Stanley et al. (2002) shortly after the
removal of two low-head dams on the Baraboo River, Wisconsin, USA.
The frequency of slower water velocities at the stream margins stayed
constant, allowing the retention of finer substrates such as silt. As sand
decreased in frequency, so did several size classes of the smaller gravels
that were later replaced by larger size gravels. This shift corresponds
with the thalweg velocities increasing in magnitude but not frequency.
This also has important implications on stream biota because the
homogenous sand substrate prior to dam removal was not replaced
with homogenous larger substrate but with a greater diversity of
substrates. At the conclusion of this study, we predict substrate size
composition will continue to coarsen in the impoundment area. While
substrates are already somewhat coarser, they are still smaller than
predicted to be stable, even under typical flows, based on calculations
of relative bed stability. This means that while substrate is coarser it is
still considered unstable and not as beneficial to stream biota as
possible (Gordon et al., 2004).
The alternating patterns of riffles, runs, and pools in mixed gravel
streams are seen as a way rivers self-adjust to regulate energy
expenditure and are very important to the biological productivity of
streams (Gordon et al., 2004). These bedforms can be created by
Table 2
Percentages of bedform types found in each study zone in 1995 (prior to dam removal)
and 2004 (after dam removal); pool and complex bedforms were aggregated to insure
compatibility of complex bedform delineation.
Study zone and year Run Riffle Pool/complex Rapid
Upstream reference 1995 44.0 32.9 16.3 6.8
2004 41.5 32.8 23.6 2.1
Impoundment 1995 96.4 1.4 2.2 0
2004 68.3 13.8 17.9 0
Downstream 1995 100 0 0 0
2004 96.9 0 3.1 0
Fig. 14. Longitudinal profile of the Pine River streambed before and after dam removal.
The triangle approximates the area of the sediment fill and, along with average stream
width upstream of the impoundment, can be used to estimate the volume of sediment
that can be mobilized at a dam removal.
Fig. 15. Annual median substrate sizes in the impoundment and downstream study
zones shown in relation to the annual amounts of sediment erosion occurring in the
impoundment zone each year of the staged dam removal.
104 B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
localized scour during normal flows, at river bends, or by wood debris,
but are normally formed by high flow events with recurrence intervals
of roughly 5­20 years (Knighton,1984; Petts and Foster,1985; Beschta
and Platts, 1986, as cited in Pizzuto, 2002; Gordon et al., 2004). During
the period of the Stronach Dam removal, a 1-in-5-year flood occurred,
but a 1-in-10-year flood did not. Despite this, some new riffle and pool
bedforms formed in the impoundment reach. In 2004, the diversity of
these bedforms was not as high as seen in the reference reach and may
not be realized in the impoundment until very high flows are
experienced. Long time durations for the complete recovery of
bedforms were predicted by Pizzuto (2002) based on short-term
empirical studies and laboratory flume experiments. Bushaw-Newton
et al. (2002) also noted that riffle-pool bedforms had not reformed in
the impoundment within one year of the removal of Manatawny Creek
Dam. This has important implications for the functioning of streams as
bedform diversity influences sediment transport and sorting, influences
nutrient cycling, and is crucially important to the habitat
suitability of stream biota (Gordon et al., 2004).
Based on the results we observed in Pine River that indicate
restoration of bedform heterogeneity may take a long time to develop,
we suggest that innovative ways of actively helping bedform
reformation following dam removal may be valuable to restore river
function. For example, if another water control structure existed
upstream of a dam removal site, water releases could be negotiated to
allow pool-riffle forming flows; or various structures such as wood
debris or gravel bars could be added to the stream to aid bedform
formation. Managing for bedform reformation may lead to a faster
realization of the full benefits of stream rehabilitation achievable
through dam removal.
Another insight derived from the consideration of bedform
diversity concerns the delineation of the impoundment effects
boundary. If the top height elevation of the dam was followed
upstream, it would correspond to the boundary of the formal reservoir
where water widths would likely be very wide and water impoundment
would be most noticeable. In the case of Stronach Dam, this
corresponded to a point approximately 1.21 km upstream of the dam.
However, through our method of delineating bedform types, low
bedform diversity was apparent for a considerable distance upstream
(3.89 km upstream of the dam) of the formal reservoir. In addition, the
substrate in this area of low bedform diversity was largely sanddominated.
Therefore, the farthest upstream extent of this sandy run
habitat became our upstream boundary of impoundment effects and
the extent of where we expected to see changes from the dam removal.
This method, along with the use of aerial photos, proved quite accurate
in delineating the farthest upstream extent of impoundment effects
and changes from the dam removal. We recommend this as an easy and
cost-effective technique to predict how far upstream changes may take
place following dam removal. This information can be important in the
early dam removal planning process for assessing possible impacts
such as infrastructure concerns, mitigation measures, and landowner
impact assessments.
5.2. Downstream
As each stage of the dam was removed, the drastic difference in
elevation from the top of the sediment fill to the downstream side of
the dam created a situation where water velocities were extremely
high and flow became supercritical. Transport of the sandy substrate
proceeded with antidune formation. Farther downstream where slope
and water velocities decreased, sediment transport continued with
dune and ripple formations (Gordon et al., 2004). The transport of this
sediment aggraded the streambed by increasing the slope of this
section of stream downstream of the dam by nearly 100%, decreasing
the water depth, and slightly increasing the stream width. These
changes would have led to higher sediment transport ability, but the
sediment eroded from the impoundment was in excess of the transport
capacity of this stretch of stream. At the end of this study, ~14% of the
sediment eroded from the impoundment was retained and stored in
the first 1 km of river downstream from the dam. The rest of the
sediment was either transported farther downstream, forming a
sediment delta at the confluence of Tippy Dam Reservoir, or deposited
on the floodplain during high flows. As the stream channel downstream
from the dam aggraded, the elevation difference between the
stream and the floodplain decreased. During high flows, suspended
sediments were deposited onto the floodplain, vertically raising the
top bank by as much as 0.50 m at one of the transects.
The implications of these changes could be important in considering
the impacts from dam removals. During the erosion process
upstream of dams, the streambed is lowered reducing connectivity
with adjacent floodplain wetlands. In some circumstances, regulatory
agencies responsible for dam removal permitting may request
remediation for any lost wetlands (even if the wetlands were created
by dam construction). However, if a river valley downstream of a dam
removal is not steep and narrowly confined, floodplain connection
and recharge (frequent overbank flooding) in this stretch of stream
could be enhanced, leading to the recharge of historic wetlands or the
creation of new ones. Frequent overbank flooding was observed in the
downstream reach of Pine River following dam removal and has been
predicted to occur following the removal of other dams (Stoker and
Harbor, 1991; Randle, 2003).
Sediment deposition in the downstream reach resulted in a
substantial decrease in water depth and an increase in width, together
greatly increasing the w/d ratio of this reach. This increase in w/d
ratio reached a peak during the later stages of dam removal and began
to decrease after the removal was completed. At the conclusion of this
study, the w/d ratios were only slightly higher than pre-removal
levels.
With the increased slope in the downstream zone, water velocities
also increased. Average water velocities increased in this section and
frequency distributions changed significantly. As with the impoundment,
water velocity frequencies in the downstream zone increased in
variability. Frequencies of slower water velocities (b0.3 m/s) increased
as the stream channels became wider and shallower. At the same time,
the fastest water velocities became faster, leading to an overall greater
diversity of water velocities.
Substrate size increased only slightly in the downstream zone. This
section of stream, while having faster velocities and higher slopes than
prior to dam removal, continued to receive sediment from the eroding
impoundment. The median size of substrate increased very slightly, but
the frequencies of substrate did not change significantly. The substrate
composition of this zone will likely stay dominated by sand until the
former impoundment section reaches an equilibrium. At that time, the
downstream reach should experience some substrate coarsening
immediately downstream of the dam. However, this section overall
may stay relatively sand dominated because of the close proximity of the
next downstream reservoir and its impoundment effects.
Changes in bedform diversity downstream of Stronach Dam may
never occur. This section is a much lower gradient because of the
Tippy Dam impoundment, and even during high flows the stream may
not have the power to scour pools. This section was all run bedform
before dam removal and remained largely run bedform throughout
the course of this study. Even if pools were scoured and until the
substrate becomes more diverse, including more gravel and cobbles,
those pools may not be easily maintained in a stretch of stream with
easily movable sand.
6. Conclusions
This study achieved its objectives of documenting the spatial and
temporal dynamics of sediment erosion, transport, and deposition
following the removal of Stronach Dam. These processes and the subsequent
changes in river morphology were spatially and temporally
105B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
variable in magnitude and extent but, in general, were clearly understandable
using the principles of fluvial geomorphology. At the end of
this study, changes are still occurring in Pine River because of the
removal of Stronach Dam, even 10 years after its initiation. While
sediment incision in the impoundment seems to have finally slowed
three years after dam removal, lateral erosion, substrate coarsening, and
bedform formation will likely continue in Pine River for many years.
Although dam removal has the potential to be an effective tool for
stream rehabilitation, many of the outcomes may take years to decades
to be fully realized. We hope that this research serves as a valuable
starting point for future research on the effects of dam removal and as a
tool to improve the effectiveness and efficiency of future dam removals.
Acknowledgements
The authors gratefully acknowledge the support of the Michigan
Department of Natural Resources, Consumer's Energy, and Michigan
State University. We thank Joanna Lessard, Jeffry Spoelstra, Bradley
Thompson, Jeff Leighton, Nolan Banish, Dean Burdett, Edward McCoy,
Ben Nessia, Tim Riley, Matt Klungle, Mike Fulk, Mark Monroe, Scott
Hughes, Kelly DeGrandchamp, Kevin Mann, Brian Bellgraph, Ryan
Mann, Brent Newell, Mike Shoemaker, William Alguire, Colby Bruchs,
Jon Wagner, John Matousek, Joel Berry, and Mart Williams for their
assistance with field work. We also thank Richard Merritt, Thomas
Burton, Thomas Coon, Michael Jones, Michael Wagner, and Phanikumar
Mantha for their insightful discussions and reviews of previous
drafts of this work.
References
American Rivers, Friends of the Earth, Trout Unlimited, 1999. Dam Removal Success
Stories: Restoring Rivers Through Selective Removal of Dams that Don't Make
Sense. American Rivers, Washington, DC.
Bain, M.B., Finn, J.T., Booke, H.E., 1988. Streamflow regulation and fish community
structure. Ecology 69 (2), 382­392.
Battige, D.S., Fields, B.L., Sowers, D.L.,1997. Removal of Stronach Dam. Proceedings of the
International Conference on Hydropower (WaterPower 1997) 2, 1341­1350.
Benke, A.C., 1990. A perspective on America's vanishing streams. Journal of the North
American Benthological Society 9 (1), 77­88.
Beschta, R.L., Platts, W.S.,1986. Morphological features of small streams significance and
function. Water Resources Bulletin 22, 369­379.
Born, S.M., Genskow, K.D., Filbert, T.L., Hernandez-Mora, N., Keefer, M.L., White, K.A.,
1998. Socioeconomic and institutional dimensions of dam removals: the Wisconsin
experience. Environmental Management 22 (3), 359­370.
Bowman, M., 2002. Legal perspectives on dam removal. Bioscience 52 (8), 739­747.
Bushaw-Newton, K.L., Hart, D.D., Pizzuto, J.E., Thomson, J.R., Egan, J., Ashley, J.T.,
Johnson, T.E., Horwitz, R.K., Keeley, M., Lawrence, J., Charles, D., Gatenby, C., Kreeger,
D.A., Nightengale, T., Thomas, R.L., Velinsky, D.J., 2002. An integrative approach
towards understanding ecological responses to dam removal: the Manatawny
Creek study. Journal of the American Water Resources Association 38 (6),
1581­1599.
Chaplin, J.J., 2003. Framework for monitoring and preliminary results after removal of
Good Hope Mill Dam. In: Graf, W.L. (Ed.), Dam Removal Research: Status and
Prospects. The H. John Heinz III Center for Science, Economics and the Environment,
Washington, DC.
Collier, M., Webb, R.H., Schmidt, J.C., 1996. Dams and Rivers: Primer on the Downstream
Effects of Dams. U.S. Geological Survey, Denver, CO. Circular 1126, 94 pp.
Consumers Power Company, 1994. Historical perspective: Stronach Dam, Pine River,
Michigan. Consumers Power Company, Jackson, Michigan. February.
Cummins, K.W., 1962. An evaluation of some techniques for the collection and analysis
of benthic samples with special emphasis on lotic waters. American Midland
Naturalist 67, 477­504.
Cushman, R.M., 1985. Review of ecological effects of rapidly varying flows downstream
from hydroelectric facilities. North American Journal of Fisheries Management 5,
330­339.
Doyle, M.W., Stanley, E.H., Harbor, J.M., 2002. Predicting channel response to dam removal
using geomorphic analogies. Journal of American Water Resources Association, 38,
1567­1579.
Doyle, M.W., Stanley, E.H., Harbor, J.M., 2003. Channel adjustments following two dam
removals in Wisconsin. Water Resources Research 39,1011 doi:10.1029/2002WR001714.
Doyle, M.W., Stanley, E.H., Orr, C.H., Selle, A.R., Sethi, S.A., Harbor, J.M., 2005. Stream
ecosystem response to small dam removal: lessons from the heartland. Geomorphology
71, 227­244.
Duan, J.G., 2005. Analytical approach to calculate rate of bank erosion. Journal of
Hydraulic Engineering 131, 980­990.
Evans, J.E., Mackey, S.D., Gottgens, J.F., Gill, W.M., 2000. Lessons from a dam failure. Ohio
Journal of Science 100 (5), 121­131.
Federal Emergency Management Agency and U.S. Army Corps of Engineers. 1996. Water
control infrastructure -- National Inventory of Dams, Updated data 1995 ­ 1996.
Freeman, R., Bowerman, W., Grubb, T., Bath, A., Dawson, G., Ennis, K., Giesy, J., 2002.
Opening rivers to trojan fish: the ecological dilemma of dam removal in the Great
Lakes. Conservation in Practice 3 (4), 35­40.
Gallagher, A.S., Stevenson, N.J., 1999. Streamflow. In: Bain, M.B., Stevenson, N.J. (Eds.),
Aquatic Habitat Assessment: Common Methods. American Fisheries Society,
Bethesda, Maryland, pp. 149­158.
Gordon, N.D., McMahon, T.A., Finlayson, B.L., Gippel, C.J., Nathan, R.J., 2004. Stream
Hydrology: An Introduction for Ecologists, Second Edition. J. Wiley & Sons Ltd.,
England.
Graber, B.E., Bowman, M., Carney, R.S., Doyle, M.W., Fisher, M., Mackey, S.D., Wildman, L.,
2001. Technical issues in small dam removal engineering. The Future of Dams and
their Reservoirs, 21st Annual USSD Lecture Series, Denver, CO: July 30­August, vol. 3,
p. 2001.
Graf, W.L., 2003. Summary and perspective. In: Graf, W.L. (Ed.), Dam removal research:
status and prospects. The H. John Heinz III Center for Science, Economics and the
Environment, Washington, D.C.
Gregory, S., Li, H., Li, J., 2002. The conceptual basis for ecological responses to dam
removal. Bioscience 52 (8), 713­723.
Hammad, H.Y., 1972. River bed degradation after closure of dams. American Society of
Civil Engineers. Journal of the Hydraulics Division 98, 591­607.
Hansen, E.A., 1971. Sediment in a Michigan trout stream, its source, movement, and
some effects on fish habitat. U.S. Forest Service Research Paper NC-59.
Hart, D.D., Johnson, T.E., Bushaw-Newton, K.L., Horwitz, R.J., Pizzuto, J.E., 2003.
Ecological effects of dam removal: an integrative case study and risk assessment
framework for prediction. In: Graf, W.L. (Ed.), Dam Removal Research: Status and
Prospects. The H. John Heinz III Center for Science, Economics and the Environment,
Washington, D.C.
Hicks, B.J., Watson, N.R.N., 1985. Seasonal changes in abundance of brown trout (Salmo
trutta) and rainbow trout (S. gairdnerii) assessed by drift diving in the Rangitkei river,
New Zealand. New Zealand Journal of Marine and Freshwater Research 19, 1­10.
Hill, M.J., Long, E.A., Hardin, S., 1994. Effects of dam removal on Dead Lake Chipola River,
Florida. Proceedings of the Annual Conference SEAFWA, pp. 512­523.
Johnson, S.E., Graber, B.E., 2002. Enlisting the social sciences in decisions about dam
removal. Bioscience 52 (8), 731­738.
Kanehl, P., Lyons, D.J., Nelson, J.E., 1997. Changes in the habitat and fish community of
the Milwaukee River, Wisconsin, following removal of the Woolen Mills Dam. North
American Journal of Fisheries Management 17, 387­400.
Knighton, D., 1984. Fluvial Forms and Processes. Edward Arnold, London.
Kondolf, G.M., Li, S., 1992. The pebble count technique for quantifying surface bed
material size in instream flow studies. Rivers 3 (2), 80­87.
Lessard, J.L., Hayes, D.B., 2003. Effects of elevated water temperature on fish and
macroinvertebrate communities below small dams. River Research and Applications
19, 721­732.
Ligon, F.K., Dietrich, W.E., Trush, W.J., 1995. Downstream ecological effects of dams.
Bioscience 45 (3), 183­192.
Morris, G., Fan, J.,1998. Reservoir Sedimentation Handbook: Design and Management of
Dams, Reservoirs and Watersheds for Sustainable Use. McGraw-Hill Co.
National Research Council, 1992. Restoration of Aquatic Ecosystems. National Academy
Press, Washington, D.C.
Petts, G.E., 1980. Long-term consequences of upstream impoundment. Environmental
Conservation 7 (4), 325­332.
Petts, G.E., Foster, I., 1985. Rivers and Landscape. Edward Arnold, London.
Pizzuto, J., 2002. Effects of dam removal on river form and process. Bioscience 52 (8),
683­691.
Pohl, M., 2003. American dam removal census: available data and data needs. Chapter 2
In: Graf, W.L. (Ed.), Dam Removal Research: Status and Prospects. The H. John Heinz
III Center for Science, Economics and the Environment, Washington, D.C.
Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.,1992. Numerical Recipes in C:
the Art of Scientific Computing, Second edition. Cambridge Univ. Press, pp.131­132.
Randle, T.J., 2003. Dam removal and sediment management. Chapter 6 In: Graf, W.L.
(Ed.), Dam Removal Research: Status and Prospects. The Heinz Center for Science,
Economics, and the Environment, Washington, D.C.
Rathburn, S.L., Wohl, E.E., 2003. Sedimentation hazards downstream from reservoirs.
In: Graf, W.L. (Ed.), Dam Removal Research: Status and Prospects. The Heinz Center
for Science, Economics, and the Environment, Washington, D.C, pp. 105­118.
River Alliance of Wisconsin and Trout Unlimited. 2000. A citizen's guide to restoring
rivers.
Rozich, T.J. 1998. Manistee River assessment. Michigan Department of Natural
Resources Fisheries Division Special Report Number 21, Lansing, Michigan.
Shafroth, P.B., Friedman, J.M., Auble, G.T., Scott, M.L., Braatne, J.H., 2002. Potential
responses of riparian vegetation to dam removal. Bioscience 52 (8), 703­712.
Shields, F.D., Simon, A., Steffen, L.J., 2000. Reservoir effects on downstream river channel
migration. Environmental Conservation 27 (1), 54­66.
Shuman, J.R., 1995. Environmental considerations for assessing dam removal alternatives
for river restoration. Regulated Rivers: Research and Management 11, 249­261.
Simon, A., 1989. A model of channel response in disturbed alluvial channels. Earth
Surface Processes and Landforms 14, 11­26.
Smith, L.W., Dittmer, E., Prevost, M., Burt, D.R., 2000. Breaching of a small irrigation dam in
Oregon: a case history. North American Journal of Fisheries Management 20, 205­219.
Stanley, E.H., Doyle, M.W., 2002. A geomorphic perspective on nutrient retention
following dam removal. Bioscience 52 (8), 693­701.
Stanley, E.H., Luebke, M.A., Doyle, M.W., Marshall, D.W., 2002. Short-term changes in
channel form and macroinvertebrate communities following low-head dam
removal. Journal of the North American Benthological Society 21 (1), 172­187.
106 B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107
Steel, R.G.D., Torrie, J.H., 1980. Principles and Procedures of Statistics: a Biometrical
Approach, Second Edition. McGraw-Hill Book Company. New York.
Stoker, B., Harbor, J., 1991. Dam removal methods, Elwha River, Washington. In: Shane,
R.M. (Ed.), Hydraulic Engineering: Proceedings of the 1991 National Conference on
Hydraulic Engineering, Nashville, Tennessee. American Society of Civil Engineers,
NY, pp. 668­673.
The Heinz Center, 2002. Dam removal: Science and Decision Making. The H. John Heinz
III Center for Science, Economics and the Environment, Washington, D.C.
Trout Unlimited, 2001. Small Dam Removal: a Review of Potential Economic Benefits.
Ward, J.V., Stanford, J.A., 1989. Riverine ecosystems: the influence of man on catchment
dynamics and fish ecology. In: Dodge, D.P. (Ed.), Proceedings of the International
Large River Symposium. Canadian Special Publication of Fisheries and Aquatic
Sciences, vol. 106, pp. 56­64.
Wentworth, C.K., 1922. A scale of grade and class for elastic sediments. Journal of
Geology 30, 377­392.
Whitelaw, E., MacMullan, E., 2002. A framework for estimating the costs and benefits of
dam removal. Bioscience 52 (8), 724­730.
Wildman, L.A.S., MacBroom, J.G., 2005. The evolution of gravel bed channels after dam
removal: case study of the Anaconda and Union City Dam removals. Geomorphology
71, 245­262.
Williams, G.P., Wolman, M.G., 1984. Downstream effects of dams on alluvial rivers.
Professional paper 1286. U.S. Geological Survey, Washington, D.C.
Wohl, E.E., Cenderelli, D.A., 2000. Sediment deposition and transport patterns following
a reservoir sediment release. Water Resources Research 36 (1), 319­333.
Wolman, M.G., 1954. A method of sampling coarse river-bed material. Transactions of
the American Geophysical Union 35 (6), 951­956.
107B.A. Burroughs et al. / Geomorphology 110 (2009) 96­107