Restoration of Fish Habitat in Relocated Streams
TENMILE CREEK- A CASE STUDY IN STREAM RELOCATION
West of Denver, the only practical location for interstate Highway 70 lies in narrow mountain canyons, one of which is that of Tenmile Creek. The route through Tenmile Canyon presented
difficult alternatives to the location engineers. A location on the canyon floor would involve
extensive relocation of Tenmile Creek. Alternative locations avoiding the creek required either a
costly tunnel, or deep cuts in the previously-unscarred western slopes of the canyon. Several
years of study by the Colorado Department of Highways and other agencies demonstrated that
the adverse environmental impacts on the canyon of highway construction would be least for the
canyon bottom route, and also the easiest to remedy. This route was adopted in January 1971.
An important factor in the decision to locate on the canyon floor was the poor condition of
Tenmile Creek. When the decision was made, there was very little of the original natural channel
left. Beginning in the 1880's, two narrow gage railroads and a wagon road were built through the
canyon to serve mining communities upstream and to the west. The railroads made liberal use of
channel changes and dikes to protect their roadbeds from floods. Later, after the railroads went
out of business their roadbeds were taken over by the State and County, widened, and made into
roads. These improvements were accompanied by further stream changes.
To make things worse, large amounts of pulverized rock were brought down as sediments from
mill tailing ponds at several mining communities upstream. Over the years, these sediments
covered the spawning gravels at the head of the canyon, and in places filled the natural channels,
causing the stream to spread widely in shallow braided rills that were impassable to fish at low
water stages. Toxic drainage from old mines and tailing ponds completed the destruction of the
fishery in Tenmile Creek.
Strict pollution controls were imposed on the mines in the middle 1960's, and by 1972 the water
quality of Tenmile Creek had so improved that brown trout began to move up the stream from
Dillon Reservoir. This encouraging development led to the hope that a spawning run might be
established in Tenmile Creek if the channel could be restored to a reasonably healthy condition.
Preliminary Stream Relocation Studies
The original Canyon Bottom Route required two crossings of Tenmile Creek and 7,800 ft (2380
m) of stream relocations. Further studies showed that if the length of stream relocation were
doubled bridges would not be needed. Since the existing channel was of poor quality and choked
with mine sediments, it seemed to the Colorado Department of Highways that relocation could
actually improve the quality of the stream, and the construction money saved by omitting the
bridges could be used to provide habitat improvements in the new channel. The Colorado
Division of Wildlife and the U.S. Forest Service concurred in this assessment, and the bridges
were eliminated. At the same time, it was agreed that the existing county road occupying one of
the former railroad beds along the eastern side of the canyon should be left in place for access to
the stream and as a route for utilities.
The design engineers plotted the required channel relocations on the topographic maps made for
the location studies, and these new channels were examined in the field by an interdisciplinary
team. At various times, this team included highway engineers, hydraulic engineers, hydrologists,
foresters, landscape architects, and biologists of the Colorado Department of Highways, the
Colorado Division of Wildlife, the U.S. Forest Service and the Federal Highway Administration.
In some places, the team moved the sketched tentative channel locations to avoid large trees, or
to introduce more curvature into the channel. A small grove of pines was preserved by dividing
the stream leaving the grove on an island. A few historic channels, long unused because of
previous diversions, were utilized for the new relocations. The members of the team tried to
visualize how the new channels would look when excavated, and how they would relate to their
surroundings. In the end, the team selected detailed routes for seven new channels totaling
16,130 ft (4,916 m) in length, leaving 12,430 ft (3,789 m) of stream undisturbed. A survey party
made a traverse and profile of the proposed new channels in order to delineate their locations on
the construction plans.
While the location studies were under way Department of Highways geologists determined that
the canyon floor was deeply overlaid by coarse gravel deposits, indicating a potentially stable
channel. Extensive talus deposits below the steep eastern wall of the canyon would provide an
ample supply of large angular rocks for stream habitat improvement.
The Colorado Division of Wildlife made a fishery survey of Tenmile Creek. Water quality and
temperature were found to be acceptable for trout habitat, with adequate flows and no barriers to
migration during low stages. A count of the fish population disclosed 80 brown trout and 7
brook trout living in a 1500-foot (457 m) sample of the stream-proof that the creek could sustain
a good fishery.
According to the ecological survey, the existing stream bank vegetation consisted of several
varieties of willow, mixed with river birch, alder and lesser plants and grasses. Elsewhere on the
valley floor were small stands of lodgepole pine and aspen, with numerous smaller varieties and
seedlings in the understory. The report also noted several ancient bristlecone pines and old
Douglas firs, and an "extremely aged" lodgepole pine, and cautioned against disturbing these
ecologically valuable specimens.
From a study of the U.S. Geological Survey's records of streamflow and weather, the
hydrologists predicted the following probable discharges for Tenmile Creek where it enters the
50 percent chance flood (Q2) 650 cfs (17.6 m3/sec)
2 percent chance flood (Q50)1300 cfs (35.1 m3/sec)
1 percent chance flood (Q100) 1400 cfs (37.8 m3/sec)
Design of the Relocated Channels
The Division of Wildlife requested that the new channels be as narrow as possible to concentrate
the low water flow during the fall spawning season, and to prevent the formation of anchor ice
during the winter. They also requested that the channels be excavated with vertical slopes, thus
reducing the width of clearing and enhancing the possibility that the stream would undercut the
banks and provide cover for fish. Since the new channels would lie almost entirely in coarse
gravel deposits, these steep slopes could be expected to weather down to a natural appearance
without serious bank erosion.
The Division of Wildlife also recommended that fish habitat be provided in the new channels by
adding check dams, current deflectors and large rocks, and by excavating "pockets" or "eddies" at
intervals in the banks, as shown by Figure 22. To avoid creating migration barriers, they
requested that all bridges and culverts be designed with gravel bottoms, and provide a minimum
depth of 4 inches (10 cm) of water during the flows expected in the spawning season, and a
velocity no greater than 2 ft/sec (0.6 m/sec) during such flows. All of these features were
incorporated in the final design.
A critical decision that had to be made by the Colorado Department of Highways was the
magnitude of the design flood. It was finally decided to size the new channels to carry the 50
percent chance flood, (Q,), without bank overflow, and to provide dikes to prevent greater floods
from flowing toward the road embankments.
The starting point for the hydraulic design was the profile surveyed in the field for each of the
seven relocated channels. On this profile, the hydraulic engineer laid long, uniform gradients
such that the average channel depth would be 3 to 5 ft (0.9 to 1.5 m). Where these gradients were
less than 1 percent he assumed a channel width of 30 ft (9.1 m); elsewhere, the width was 25 ft
(7.6 m). From the Manning equation, the hydraulic engineer then calculated the velocity and
depth of f low to carry the design flood on the assumed gradients. If this calculated depth of flow
was less than the depth of the proposed channel, the new channel was deemed satisfactory.
The calculated Q2 velocities were quite high for the steeper channels-around 1 1 to 14 ft/sec (3.4
to 4.3 m/sec)-indicating that erosion could be expected where the channel bed consisted of gravel
or finer materials. However, inspection of the existing natural channel suggested that the new
channel would become stable after the smaller materials had washed away, leaving a bed of
boulders and cobbles throughout most of the relocated channel.
Provisions for Habitat Improvement
The habitat improvement plan made liberal use of large rocks, of which there was an abundance
within easy haul of the creek. These were to be paid for according to the equipment hours
required to excavate, haul and place them, leaving their number and placement to the field forces.
In addition to large rocks, provision was made for 43 check dams to be placed in the channels.
These could be of logs, as shown in Figure 15(c), or of rocks, as in Figure 13. These structures
were spaced 200 to 250 ft (61 to 76 m) apart in the steeper channels, but it was understood that
these locations would be changed to meet field conditions after the channels were excavated.
The habitat plan provided for 42 deflectors, which could be of rocks, as in Figure 6, or of logs.
The latter were designed to produce a scour hole under the logs at low and medium flows.
These, also, could be shifted from the locations shown on the plans to better fit field conditions.
The Sediment Control Plan
The specifications required that channel excavation begin in the lowest channel and proceed
upstream. Seepage into the new channels would be collected into settling ponds from which,
after settlement of the solids, the trapped water would be pumped to vegetated filter areas. Cross
drainages which might discharge water into the new channels were to be temporarily diverted
through ditches or culverts, or pumped to the settling ponds. The locations of these diversions
and ponds were shown on the plans, and contract items were provided to compensate the
contractor for sediment control work. The project engineer was required to monitor water quality
during construction by testing the turbidity of samples taken daily above and below the channel
excavation operations. Diversion of Tenmile Creek from the existing channel was prohibited
from September 15 to November 30, during the annual brown trout spawning run.
Landscape Protection and Restoration
The designers estimated the surface area of the channel relocations at about 11 acres (4.5 ha).
Within this area were a number of medium sized trees, and hundreds of smaller trees and shrubs.
The plans called for transplanting 200 of the smaller trees and shrubs to areas such as new dikes
and channel banks; and for stripping and storing all available topsoil within the construction
limits for placement on disturbed areas such as excavation, disposal areas, dikes and haul roads.
Finally, all disturbed areas were required to be seeded with a mixture of native and pasture
grasses, followed by hydro-mulching.
It was anticipated that the contractor would do most of his hauling in the completed channels,
using them as roads. He was prohibited from operating outside of construction limits except
where temporary haul roads and turning places were designated for his use.
Stream Relocation By Separate Contract
Most road projects involve little or no stream relocation work. Where some relocation is
required, it is generally a minor part of the total construction effort-so small in fact that it is
seldom economical or practical to do the work by a separate contract.
The situation at Tenmile Creek was quite different. Here was a proposal to relocate 3 miles (4.8
km) of stream involving 77,000 cubic yards (569200 m3) of excavation and estimated to cost
$443,000, a sum large enough to justify construction as a separate contract. By isolating the
stream relocation from other road work the Colorado Department of Highways hoped to acquire
reliable information on the cost of this highly specialized work. The Division of Wildlife
thought a better job would result from the contractor's applying his undivided efforts to building
a trout stream rather than a highway.
When the relocation of Tenmile Creek was advertised for bids in the spring of 1976, five
contractors submitted bids. The wide spread between the high and low bids reflected the unusual
character of the work. For example, the bid price for channel excavation ranged from $2.80 to
$7.50 per cu yd ($3.80 to $10.50 per m3). Log dams ranged from $160 each to $800 each, over
and above the cost of the logs, which were supplied free by the State. Rock dams varied from a
low of $150. each to $700. The high bid was about double the low bid. A contract for the work
was awarded on the low bid of $493,467. Of this amount, approximately 78 percent was for
mobilization, channel excavation, concrete box culverts and sediment control, and 22 percent for
habitat rocks, check dams, deflectors and landscaping.
Staking the Relocated Channels for Construction
The long, straight channel grades shown on the plans were not followed exactly when the
relocated channels were staked for the contractor. Instead, the field party varied the gradient
practically every 100 ft (30 m), departing as much as 1 or 2 percent from the average plan
gradients in places. They also made minor deviations from the channel alignments sketched on
the plans. As a result, the staked channels had many minor irregularities in both line and grade.
As the staking proceeded, station by station, the party chief checked the probable design
discharge depth of flow for each segment of channel from discharge curves prepared by the
hydraulic designer. The channel depth was adjusted by raising or lowering the bed elevation to
insure a freeboard of at least 1 ft (30 cm) above the design discharge water surface. The
landscape architect followed the stakeout party, marking trees and shrubs to be saved for
transplants. Finally, each channel was reviewed by the resident engineer, hydrologist and fishery
biologist before excavation began.
Channel Excavation Operations
Before any channel excavation was done the Department of Highways obtained a water quality
certificate from the State Department of Health.
Channel excavation began at the downstream end of the northernmost channel, and proceeded
upstream. The contractor first removed the transplants to temporary storage areas and stripped
and stockpiled the topsoil. A heavy bulldozer equipped with a ripper and rockrake loosened the
dense gravel deposit underlying the new channel, and a 51/2 cu yd (4 m3 ) rubber tired loader
loaded the gravel on 31 ton (28 t) end dump trucks for transport to the dikes and other fill areas.
Boulders large enough to be used as habitat rocks were cast aside for later placement in the
channel. Most of the excavation was disposed of in the dikes; but some of the spoil was used to
improve braided sections of the original creek bed by confining the stream to a narrower channel.
The installation of habitat structures followed closely after the channel excavation. The locations
for these structures were selected by the fishery biologist and hydrologist, both of whom had
studied and observed fish habitat in trout streams for many years. Practically all of the check
dams were located where the stream gradient was 2 percent or steeper, while the deflectors were
placed in the flatter sections.
One of the problems in check dam installation is how to make them passable for fish during low
water. This problem was solved for the log dams by cutting a notch in the top log to concentrate
the low flows in the middle of the dam. It was much more difficult to concentrate the flow over
the rock dams as it was nearly impossible to make an impervious crest with loose rocks. The
design finally adopted was to split the dam into two segments, positioning one below the other as
in Figure 14. At high flows this structure performs as a straight dam, but at low water the flow is
concentrated in a diagonal chute for easy passage by migrating fish.
The average number of large rocks placed in the new channels exceeded 300 per mile (180 per
km) of channel. These ranged from 2 ft (0.6 m) in diameter up to 5 ft (1.5 m), with a few even
larger. One rock was 8 ft (2.4 m) long. Most of these rocks were placed where they would cause
scour holes to form in the bed, but some were placed in straight steep sections to reduce water
The check dams and deflectors were embedded deeply in the new channel banks to avoid
undercutting and washout. For this work the contractor's equipment operated from the excavated
channel in order to avoid damage to bank vegetation. Logs, large habitat rocks, topsoil and
transplants were brought in by the same route.
When the habitat structures were in place, the contractor removed the sediment pond at the lower
end of each channel and turned the stream into its new bed. The first channel was filled in the
latter part of June 1976 when the spring runoff was nearly at its peak and the stream was flowing
at about 350 cfs (9.8 m-1/sec).* In less than 2 hours the fine materials were swept out of the new
channel and the stream was running clear.
* Runoff was slightly below the average in 1976., The long term average flow for June for the period 1961-1975, was about 370 cfs (10.3 m3/sec).
The last new channel was opened three days after the September 15 deadline with special
permission from the Division of Wildlife. By this time, the stream flow had dropped to about 32
cfs (0.9 m3/ sec.). The project was substantially completed by November 15, 1976 at a final cost
A Rejuvenated Trout Stream
As each section of old channel was dewatered, the Colorado Division of Wildlife captured the
resident trout and released them in the new channels. In all, 1033 fish were salvaged, mostly of
small size but a few 12 inches (305 mm) or larger. A survey of the modified stream showed that
the total channel (new and old sections) as finally excavated was 5.7 miles (9.2 km) long as
compared to 5.5 miles (8.8 km) for the original channel. The surface area of the stream was
reduced from 32 acres (13 ha) to 28 acres (11.3 ha), mostly by drainage of beaver ponds and
elimination of many wide silt-choked braided reaches in the old stream. This survey was part of
a 5-year research project to evaluate the changes in channel structure and fish habitat and the
improvement in the fishery. Under this program, water quality will be monitored monthly and
sample counts of fish and insect populations will be made at regular intervals.
Response of the Relocated Channels to Floods
A year after completion, it was generally agreed that the new channel was a better trout stream
than the original Tenmile Creek. It was free of mine sediment. Instead of disappearing in broad
flats at the upper end of the canyon, the creek was concentrated in one channel with adequate
flow and fish habitat. The steeper channels had suffered little from scour, and the dams,
deflectors and habitat rocks had functioned as expected. Despite two large spills of mine tailings
and gasoline into the creek, the fish population was increasing, and a promising spawning run
Up to this time, the creek had not experienced a severe flood. Runoff for 1976 and 1977 was
below normal-less than half of the 650 cfs (17.6 m3/ sec) design flood assumed by the Colorado
Department of Highways. In June, 1978, Ten Mile Creek experienced a 4 percent chance flood,
in which the discharge reached 1200 cfs (32.5 m3/ sec). The results of this flood have not been
completely evaluated, but it appears that the channel relocations with flat gradients-0.5 to 1.5
percent-have suffered little damage from erosion, and in fact have been improved by the flood,
the sediments having been arranged into excellent riffle-pool sequences. The check dams and
habitat rocks are in place, although some are partly buried by sediment.
Reaches with medium gradients-1.6 to 2.5 percent-have suffered considerable bank scour,
especially on the outsides of bends. However the beds have not seriously degraded, and most of
the rocks and structures are in place. There has however been considerable filling of the
preformed scour holes below the log dams with coarse sediments, up to 18 inches (458 mm) in
diameter. Velocities in these channels probably reached 12 ft/sec (3.7 m/sec).
The steepest channel, of gradient 3.6 to 5.0 percent has suffered severe bank erosion, and in one
place the channel has doubled its original width. A number of log dams partly washed out, and
the downstream plunge pools generally filled with coarse sediment. Sediment from this steep
channel has caused channel build-up in about 1000 feet (305 m) of the next channel downstream,
partly or completely covering the habitat rocks and check dams. The calculated velocities in the
steep section were about 16 ft/sec (4.9 m/sec).
It is virtually certain that habitat rocks up to 4 ft (1.2 m) diameter were moved from their original
positions by flood velocities of 12 to 14 ft/sec. (3.7 to 4.3 m/sec). One such rock was observed
resting on a log dam.
A very steep section of the original stream was improved by placing many large rocks in the
channel to form stair step pools and cascades. This section came through the flood without
damage. This may be the best treatment for channel relocations with steep grades.
Despite the flood-induced alterations to the manmade channel and its structures, Tenmile Creek
remains an excellent trout stream. The relocation of 3 miles of its channel has been
accomplished without loss of scenic or ecological values.