Available online at www.sciencedirect.com
Geomorphology 96 (2008) 366 – 377
www.elsevier.com/locate/geomorph
Effective mitigation of debris flows at Lemon Dam,
La Plata County, Colorado
Victor G. deWolfe a,⁎, Paul M. Santi a , J. Ey b , Joseph E. Gartner c
a
Department of Geology and Geological Engineering, Colorado School of Mines, Golden, CO 80401, United States
b
Florida Water Conservancy District, Lemon Dam, LaPlata County, CO, United States
c
U.S. Geological Survey, Golden, CO 80401, United States
Received 3 July 2006; received in revised form 28 December 2006; accepted 9 April 2007
Available online 10 May 2007
Abstract
To reduce the hazards from debris flows in drainage basins burned by wildfire, erosion control measures such as construction of check
dams, installation of log erosion barriers (LEBs), and spreading of straw mulch and seed are common practice. After the 2002 Missionary
Ridge Fire in southwest Colorado, these measures were implemented at Knight Canyon above Lemon Dam to protect the intake
structures of the dam from being filled with sediment. Hillslope erosion protection measures included LEBs at concentrations of 220–
620/ha (200–600% of typical densities), straw mulch was hand spread at concentrations up to 5.6 metric tons/hectare (125% of typical
densities), and seeds were hand spread at 67–84 kg/ha (150% of typical values). The mulch was carefully crimped into the soil to keep it
in place. In addition, 13 check dams and 3 debris racks were installed in the main drainage channel of the basin.
The technical literature shows that each mitigation method working alone, or improperly constructed or applied, was
inconsistent in its ability to reduce erosion and sedimentation. At Lemon Dam, however, these methods were effective in virtually
eliminating sedimentation into the reservoir, which can be attributed to a number of factors: the density of application of each
mitigation method, the enhancement of methods working in concert, the quality of installation, and rehabilitation of mitigation
features to extend their useful life. The check dams effectively trapped the sediment mobilized during rainstorms, and only a few
cubic meters of debris traveled downchannel, where it was intercepted by debris racks.
Using a debris volume-prediction model developed for use in burned basins in the Western U.S., recorded rainfall events
following the Missionary Ridge Fire should have produced a debris flow of approximately 10,000 m3 at Knight Canyon. The
mitigation measures, therefore, reduced the debris volume by several orders of magnitude. For comparison, rainstorm-induced
debris flows occurred in two adjacent canyons at volumes within the range predicted by the model.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Debris flow; Erosion; Mitigation; Mulch; Seeding
1. Introduction
⁎ Corresponding author. Current address: Deere & Ault Consultants,
Inc., 600 S. Airport Rd., Bldg A, Ste 205, Longmont, CO 80503,
United States. Tel.: +1 303 651 1468.
E-mail address: victor.dewolfe@deereault.com (V.G. deWolfe).
0169-555X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2007.04.008
In June and July of 2002, the Missionary Ridge Fire
burned 29,527 ha in Southwest Colorado near the city of
Durango (Fig. 1). Lemon Dam and Reservoir are located
approximately 23 km northwest of Durango, roughly in
the center of the perimeter of the fire. Because Lemon
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
367
Fig. 1. Map of the drainage basins potentially impacting Lemon Dam in Southwestern Colorado. The three basins discussed in the text are shown.
Base topographic maps are USGS Lemon Reservoir and Rules Hill 7.5 min quadrangles.
Reservoir is a critical part of the water supply system for
the city of Durango, erosion of sediment or debris flow
into the reservoir following the 2002 Missionary Ridge
Fire could have severe consequences. The Florida Water
Conservancy District (FWCD) was given the mission
and the resources to prevent significant sediment
movement into critical portions of the reservoir. They
concentrated their efforts on the drainage basin known
as Knight Canyon, which empties near the intake
structure for the water supply system. The mitigation
efforts carried out by the FWCD virtually eliminated
sedimentation into the reservoir, even following several
rainstorms that produced debris flows and sedimentladen floods in adjacent canyons. The purpose of this
paper is to document the treatment at Knight Canyon, to
demonstrate its success based on volume-predictive
modeling and comparison with adjacent canyons, and to
analyze potential effectiveness of each method of the
slope and channel treatments used at Knight Canyon
through a review of technical literature.
2. Methods
A substantial amount of research has tested the
effectiveness of typical erosion control methods at
reducing sediment movement at plot or hillslope scales.
The goal of the literature review is to summarize this
work, yet with emphasis on basin scale application, and
interpreted for reduction of the occurrence and size of
debris flows, rather than only evaluating sediment
erosion.
Next, the debris-flow mitigation methods in Knight
Canyon are described and quantified, where possible, by
comparison to typical rates of application used by the U.S.
368
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
Forest Service. Field observations provide supplemental
information on design, performance, and maintenance of
the mitigation elements.
Finally, reduction in the volume of debris flows at
Knight Canyon is demonstrated by the use of a recently
developed debris-flow volume-predictive model (Gartner, 2005). Using this model, the volume of debris
generated can be compared with an expected volume,
which is a function of basin characteristics and rainfall
magnitude. The debris-flow response in Knight Canyon
is compared to the response in two adjacent canyons:
Meyer Canyon and South Basin, where only a limited
amount of mitigation was completed.
3. Effectiveness of debris-flow mitigation methods
3.1. General
The effectiveness of erosion control treatments in
burned areas has been evaluated in the past in multiple
studies (e.g., Beyers et al., 1998; Wohlgemuth et al.,
1998, 1999; Robichaud et al., 2000; Wohlgemuth et al.,
2001; Miles, 2005; Robichaud et al., 2006). These
efforts have been primarily at plot or hillslope scales.
The generation of debris flows from recently burned
basins, however, involves runoff and erosion processes
acting throughout an entire basin and often on steep
slopes (Wells, 1987; McDonald and Giraud, 2002;
Cannon and Gartner, 2005). The shift from plot- or
hillslope- to basin-scale erosion and sediment control is
important to consider when attempting to prevent
erosion from burned basins subject to debris-flow
processes. For this reason, this study will focus on
basin-scale processes.
Santi et al. (2008-this volume) and deWolfe (2006)
demonstrated that hillside and channel treatment
methods are beneficial to reduce the volume of a debris
flow. They showed that most sediment production
comes from channel scour of moving debris (typically
over 90%), but much of the water necessary for the flow
comes from hillside runoff. Therefore, hillside treatments reduce the potential for debris flows by reducing
runoff and increasing rainfall infiltration. Channel
treatments reduced the volume of a debris flow by
reducing scour and erosion should debris flow begin to
mobilize in the channel.
3.2. Erosion control success
Robichaud et al. (2000) compiled information on the
perceived effectiveness of different erosion control
treatments. The data were derived from surveys filled
out by emergency response personnel who had
employed the various treatments. Among other items,
they were asked to rate the performance of several
treatment methods for periods of one, three, and five
years after application.
Aerial seeding is the most widely used hillslope
treatment, although the perceived effectiveness was
almost evenly distributed amongst the four categories
(excellent, good, fair, and poor). Log Erosion Barriers
(LEBs) were the next most widely used treatment,
whose effectiveness was also highly variable, although
more responses were in the “good” or “excellent”
categories. The third most common treatment was
mulching, with effectiveness ratings dominantly falling
in the “excellent” category. The fourth most common
treatment method, ground seeding, was also overwhelmingly rated as “good” for its effectiveness in
controlling erosion.
Prevalent methods of channel treatment included
three types of check dams (straw bale, log, and straw
wattle) as well as log and rock grade stabilizers. Of
these, log dams were generally considered the most
successful type of check dams, followed by straw wattle
dams. Evaluators gave the grade stabilizing treatments
mixed reviews, with a substantial number of “fair” and
“poor” evaluations.
3.3. Seeding
Seeding, frequently employed as short-term (1–
3 year) erosion control, introduces relatively fast
growing grasses into a burned area so that a vegetative
cover can be re-established as quickly as possible. The
roots of the plants will bind and stabilize soil material,
reduce the impact of raindrops, and increase infiltration
(Miles, 2005). Slopes with inclinations greater than
about 37° (75%) are considered too steep for revegetation (Chelan County Public Utility District,
2001).
Plot-scale evaluations of the effectiveness of seeding
to promote the development of vegetative cover or
reduce erosion in chaparral environments have produced
inconclusive results. Gautier (1983) measured a 31%
reduction in erosion on seeded plots compared to
unseeded plots in a chaparral forest near San Diego,
CA. Near San Luis Obispo, CA, Taskey et al. (1989),
however, found no significant difference in erosion of
seeded chaparral slopes versus unseeded ones. Vegetative cover (Beyers et al., 1998) and hillslope erosion
(Wohlgemuth et al., 1998) were studied on four hot
prescribed fires on mixed chaparral forests in southern
California. For each site, vegetative cover was higher on
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
the treated slopes, but measured rates of erosion showed
no significant difference.
Evaluations of seeding effectiveness in burned
conifer forests are also inconclusive. Miles et al.
(1989) rated aerial seeding effectiveness as moderate
for the South Fork Trinity River fires in northern
California. Amaranthus (1989), however, found that 75
to 90% of erosion occurred before annual ryegrass seeds
established a vegetative cover on a treated slope in
southern Oregon. A paired watershed study in the
northern Sierra Nevada in California found no significant difference of vegetative cover or erosion for
seeded versus unseeded plots after two years following
the fire (Roby, 1989). Geier-Hayes (1997) did not
measure erosion, but found no significant difference in
vegetative cover between seeded and unseeded plots
during five post-fire years.
A comprehensive compilation of studies quantifying
the effectiveness of seeding as an erosion control
treatment by Beyers (2004) found that less than half
showed reduced erosion from seeded slopes. From this
work we conclude that seeding is an inconsistent
mitigation method, and its effects are difficult to
distinguish from natural revegetation. For debris
flows, seeding is not expected to provide substantial
hazard reduction beyond that provided by natural
revegetation.
3.4. Mulch
Mulching is an erosion control method that seeks to
provide a suitable ground cover immediately after fires
(Miles, 2005). Areas targeted by mulching include
highly erodible soils that have been severely burned to a
degree that all ground cover is lost (Miles, 2005). The
purposes of mulching are to reduce impact of raindrops,
to hold topsoil in place, and to disperse overland flow
(Miles, 2005). Mulching is, thus, intended to provide
immediate and short-term (b 1 year) erosion control.
Mulch is often applied with seeding so that rapid
reestablishment of vegetation is facilitated. It can be
applied aerially from a helicopter over large areas or by
hand over smaller areas. While it is recommended that
the mulch cover at least 40–50% of the ground, hand
mulching tends to be more effective because 100% of
the ground can be covered (Miles, 2005). Mulch can
also be crimped into the soil to keep it from blowing or
washing away. Robichaud et al. (2000) found that the
effectiveness of mulch as erosion control is enhanced by
even application and consistent thickness. Our observations of 11 basins that received helicopter mulching in
California (Old Fire) and Utah (Mollie, Springville, and
369
Farmington Fires) include thick mounds and relatively
uneven spreading patterns of the material in at least four
of the basins. The mounds form primarily from
insufficient breakup of airborne straw bales or to wind
blowing the mulch into piles in the lee of obstacles such
as trees or rocks. During heavy rains, the burned soil
mantle was eroded from between the clumps. Furthermore, clumps prevented regrowth of vegetation for at
least two years following the fire (Fig. 2).
Evaluations of the effectiveness of mulch treatments
indicate that at a plot scale, mulch can play an important
role in decreasing rates of erosion. Bautista et al. (1996)
found that straw mulch, applied at a rate of 2 metric tons/
hectare, reduced sediment yield from slopes in Spain by
about 50 to 60%. Kay (1983) measured erosion from areas
treated with jute excelsior (a woven netting) and straw
mulch, and found straw mulch to be cheaper and more
effective. Buxton and Caruccio (1979) found similar
results. Miles et al. (1989) considered mulching to have a
low risk of failure, and to be highly effective in controlling
erosion when used to treat burned soils at the South Fork
Trinity River fires in California. Robichaud (2005)
reported that straw mulch could reduce post-wildfire
erosion rates by 50–94%. A comparative study, performed after the 2000 Cerro Grande Fire in New Mexico,
found that during the first post-fire year, plots treated with
a combination of aerial seed and mulch produced 70%
less sediment from erosion than control plots, and 95%
less sediment during the second post-fire year (Dean,
2001). During the second post-fire year after the 2000
Bobcat Fire in Colorado, Wagenbrenner et al. (2006)
reported a reduction in sediment yields of at least 95%
from slopes treated with mulch compared to those treated
only with seeds and compared to untreated slopes.
Fig. 2. Photograph illustrating the clumping character of heli-mulch
dropped in 2003 and suppressed vegetation growth in 2005; location
near Devore, California.
370
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
These results demonstrate that, unlike seeding alone,
properly-applied mulch is effective at reducing sediment
erosion. Therefore, one may assume that runoff and
sediment supply for debris flows will also be reduced.
3.5. Log erosion barriers
LEBs are installed to provide mechanical barriers to
runoff (thereby reducing the potential of rill erosion),
while increasing infiltration potential (Robichaud et al.,
2000). They also provide a small basin that catches eroded
soil. LEBs consist of felled and limbed trees, aligned
along contour, and held in place with stumps or wooden
stakes. Logs are typically placed in small trenches to
ensure direct contact with the ground and to prevent
undercutting by runoff water (Moench and Fusaro, 2002).
LEBs are staggered in a brick-like pattern, so water has no
direct line of flow down the slope.
Robichaud et al. (2000) recommend installation of
LEBs on hillsides mantled with fine soil and with
gradients of less than 40%. Slopes with thin soil, high
rock content, and gradients greater than 75% are not
appropriate for LEBs. Additionally, they report that
mobilization of highly erosive soils, such as those
derived from glacial till or highly weathered granitic
rock, can overwhelm smaller LEBs.
Gartner (2003) conducted an evaluation of the
effectiveness of LEBs as erosion control treatment for the
2000 Hi Meadow Fire in Colorado. The study compared
erosion and sedimentation characteristics from two
adjacent watersheds, one treated with LEBs and the other
untreated. He concluded that LEBs did not significantly
affect either infiltration or overland flow based on
observations of water flowing over and around LEBs,
measurements indicating that less than 4% of the treated
area showed increased infiltration, and observation of flood
waves from treated and untreated watersheds arriving at a
common location at the same time (Gartner, 2003).
A similar study by Wohlgemuth et al. (2001) in southern
California after a fire in 1999 found that 157 LEBs stored a
total of only 4 m3 (5.2 yd3) of sediment after the first postfire year, and 9 m3 (11.8 yd3) of sediment after the second
post-fire year. Sediment yield data from the first post-fire
year showed that the treated basin yielded 14 times more
sediment than the untreated basin, while data from the
second post-fire year showed that the sediment yield from
the untreated basin was 18 times higher that from the
treated basin (Wohlgemuth et al., 2001). The conclusion
was that the two basins had too many inherent differences
to adequately evaluate the effectiveness of LEBs.
LEBs were shown to be effective in a study by Dean
(2001), who found that erosion from plots treated with a
combination of LEBs, mulch, and seed produced 77%
less sediment during the first post-fire year and 96% less
during the second year.
LEBs have also been evaluated for effectiveness
relative to rainfall of varying intensities (Robichaud,
2005). Robichaud (2005) found that LEBs are more
effective during low to moderate intensity rainfall, and
that the effectiveness decreases sharply with high
intensity events and when the basins become filled
with sediment. On the basis of these observations, one
may assume that LEBs are effective for reducing
hazards from debris flow only for smaller rainfall events
and only when installed in appropriate terrain.
3.6. Debris racks
VanDine (1996) describes debris racks as debrisstraining structures. The general principal is to provide
debris-resisting barriers that are designed to trap and
induce deposition of coarse debris, thereby allowing
fine material and water to pass (Fig. 3). Debris racks,
also referred to as trash racks or steel rail debris
deflectors, are often positioned at the fronts of culverts
or bridges to keep them free of debris and to minimize
structural damage (Reihsen and Harrison, 1971).
While most debris racks are rigid structures, some are
designed to be flexible if cables are used to absorb the
energy of flowing debris. Recently, flexible structures,
such as ring-nets (ROCCO ®) (Duffy and DeNatale,
1996; Thommen and Duffy, 1997), have been used. For
either rigid debris racks or flexible nets, the structures
must be properly sized and engineered.
We could find no literature on the effectiveness of debris
racks in controlling sediment transport from burned basins.
Fig. 3. Debris rack (constructed in South Basin, below Lemon Dam)
that has intercepted 103 m3 of coarse debris and wood. Fine material
and water are allowed to pass through the rack.
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
3.7. Check dams
Check dams are called by various names based on the
design and function, including consolidation dams,
Sabo dams, grid dams, slit dams, steel cell dams,
retention dams and retarding dams. Check dams are
constructed in series in channels to decrease steep
channel gradients by encouraging deposition of debris
and to minimize scour along the channel bed and
channel margins (VanDine, 1996). Accumulated debris
is typically not removed from behind check dams. Other
dams, referred to as “debris basins,” are usually not
constructed in series, but as much larger single dams,
and require access to periodically clean out debris.
Most check dams are highly designed and engineered
structures constructed from concrete or reinforced
concrete, but are also commonly constructed as timber
and steel rock-filled cribs, or as stone masonry and
gabion structures (VanDine, 1996). Excavated pit and
berm earthen check dams have also been installed in
burned channels to control erosion and the generation of
debris flows at Farmington, Utah; Lemon Dam, Colorado; at Piru, California; and the Hayman Fire in
Colorado (Robichaud, 2006).
We could find no literature on the effectiveness of
channel check dams in controlling sediment transport
specifically from burned basins, although a significant
amount of information about the function and design for
unburned settings does exist (Leys and Hagen, 1971;
Eisbacher and Clague, 1984; Government of Japan, 1984;
371
Thurber Consultants Ltd., 1984; Heierli and Merk, 1985;
Whittaker et al., 1985; Chatwin et al., 1994).
Regardless of the effectiveness, failure of check dams is
a significant issue, typically resulting from improper
engineering design of the dam, such as a lack of spillway,
insufficient keying into the bank(s), undersized for the
volume of material, or construction with materials of
insufficient strength. Pit and berm check dams are
particularly prone to failure (Robichaud, 2006). At failure,
the debris flow will incorporate the materials stored behind
the dams into the flow, thus increasing its volume and
erosive potential. If one check dam fails, it is likely that the
resulting debris flow will gain sufficient energy and
volume to damage the remaining check dams in the series.
Failure of log-barrier check dam installations in
Dominguez Canyon following the 2003 Piru Fire in
southern California resulted in significant degradation
of the health of the riparian zone and exacerbation of
sedimentation problems in the canyon (Hubbert, 2005).
Spillways were not constructed in the dams, many of
which failed by undercutting and side cuts into channel
banks. Robichaud (P.C. 2006) also reported repeated
failure of earthen check dams installed in the Hayman
Fire in Colorado.
4. Knight Canyon treatment methods
Whereas the technical literature shows inconsistent
effectiveness of many of the potential methods for the
prevention and mitigation of debris flows for burned
Table 1
Comparison of debris-flow mitigation measures in Knight Canyon to two adjacent control canyons (Meyer Canyon and South Basin) to standard
Forest Service Practice (from Wright Water Engineers, 2005; BAER, 2002a,b)
Mitigation
Knight Canyon (Florida
Water Conservancy Dist.)
Meyer Canyon (USFS)
220–620 (approximately
100 a (approximately 33% of the basin was
50% of the basin was treated) treated, but 83% of the LEBs failed by
undercutting (Smith, 2004)
Mulching
1.72–5.6 (higher in critical
None
(metric tons/hectare) areas, hand spread, crimped)
South Basin (USFS)
USDA FS Practice
Seeding (kg/hectare)
67–84 (hand spread)
None
Debris racks
Check dams
Measured debris-flow
volume (m3)
Predicted debris-flow
volume using
Eq. (1) (m3)
3
13
13
None
None
3464
100 a(approximately
100
75% of the basin was
treated)
None
4.5 b (helicopter
spread, not
crimped)
None
45 (helicopter
spread)
2
Not addressed
None
Not addressed
445
9854
5345
837
LEBs (#/hectare)
a
b
Assumed density, as treatment was coordinated by USDA FS.
Recommended by NRCS.
372
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
Fig. 4. Timeline of the erosion and sediment control program implemented at Lemon Dam listed in order of priority (information from FWCD).
areas, the combination of multiple methods was
effective at Knight Canyon above Lemon Dam and
reservoir. The density of application of each method is
detailed below (summarized in Table 1) and a timeline
of the implementation is shown in Fig. 4.
4.1. Seeding
Typical concentrations of reseeding efforts for large
areas are approximately 45 kg/ha (BAER, 2002a,b;
Wright Water Engineers, 2005). The erosion control
program at Lemon Dam used 67–84 kg/ha, which is
more than 150% the typical levels. Observations of
those slopes show that spreading this concentration
among crimped mulch helped reestablish a vegetative
cover during the first growing season (Fig. 5).
4.2. Mulch
area) and crimped by shovel (Fig. 6) (Wright Water
Engineers, 2005). While this is less than usual rates
of approximately 4.5 metric tons/hectare (as recommended in typical NRCS-USDA data sheets), 30 ha
in critical locations had application densities of
5.6 metric tons/hectare (Wright Water Engineers,
2005). Crimping helps mulch to maintain contact
with the soil and if water repellant layers are detected
in burned soils, crimping can break through the layer,
allowing water to infiltrate rather than run off the
surface. Un-crimped hand spread mulch is vulnerable
to clumping effects because wind can blow it around
and deposit it in piles that can inhibit the reestablishment of vegetation.
Observations of those slopes show that the crimped
mulch, combined with an intensive seeding program
and LEBs, resulted in a substantial and extensive
vegetative cover during the first growing season
(Fig. 5).
At Lemon Dam, straw mulch was spread by hand
at a rate of 1.7 metric tons/hectare (over a 100 ha
Fig. 5. Photograph showing the amount of vegetative cover in Knight
Canyon by August, 2003 from hand spread seeding and natural growth.
Fig. 6. Condition of hand spread mulch at Lemon Dam. Notice the
vertical straws, products of crimping. Foreground of photo spans
approximately 2 m.
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
373
4.3. Log erosion barriers
LEBs at Lemon Dam appeared to have reduced
hillslope erosion and enhanced runoff infiltration
because they were applied in heavy concentrations and
with other erosion control measures. At Lemon Dam,
93 ha of a severely burned basin above the spillway and
intake structures of the dam was treated with concentrations between 220 and 620 LEBs/hectare (Table 1,
Fig. 7). This is over two to six times the normal
concentration of about 100 LEBs/hectare (Table 1). The
LEBs were emptied of soil each time they were filled
following rainstorms, and the removed material was
packed along the front of the LEB.
For comparison, observations made of LEBs constructed in the Mollie burn area above Santaquin, Utah
are quite different from those made at Lemon Dam. In
this case, fewer logs were placed and LEBs covered a
total of 74 ha in the upper reaches of three basins,
covering 8 to 15% of the basin areas. LEBs were
observed two to three years after installation. While
most logs were observed to be on contour with true
ground contact, some were severely undercut. Substantial volumes of debris flows were produced by these
canyons, ranging from 3000 to 9000 m3 (Santi et al.,
2008-this volume).
4.4. Debris racks
Five debris racks were constructed in the Lemon
Dam sediment control program between October and
December 2002. Three racks were built in Knight
Canyon and two in South Canyon (discussed in Section
3.6). One of the racks in South Canyon intercepted the
Fig. 8. Check dam full of mud and debris constructed in the upper
reaches of Knight Canyon above Lemon Dam.
coarsest part (130 m3 in volume) of a debris-flow
measured to be about 445 m3 in total volume (Fig. 3).
The design of this debris rack prevented failure and
allowed the fine material (315 m3 in volume) to
continue on down channel, where part of it was held
back at a second debris rack. Only muddy water was
passed on to the Florida River. No debris flows
developed in Knight Canyon, so the debris racks built
there were never tested.
The effectiveness of debris racks at reducing the
volume of a debris flow depends on the strength, design
to pass fines, position in a channel and the size of that
channel. The debris racks installed at Lemon Dam were
constructed across the center of the channel thalweg
from 15-cm diameter steel pipe with 13-mm thick walls
driven into the soil using a vibro-hammer (Fig. 3). Highstrength welding and 1.5 m3 of concrete were used to
reinforce the structure.
4.5. Check dams
Fig. 7. LEBs installed on slopes above Lemon Dam in November 2002
after the Missionary Ridge Fire.
Thirteen earthen check dams were constructed on 6
to 10% gradients in the main channel of Knight Canyon.
Some, but not all, of the dams had logs on the cores
keyed into the channel banks. The dams also had a
spillway notch cut into the crest (Wright Water
Engineers, 2005). Fig. 8, taken on September 9, 2003
(14 months after the fire), shows a check dam filled with
an ash/mud deposit after a heavy rain in the basin. In the
background, another check dam can be seen in the
series. These dams prevented sediment-laden runoff
from developing into debris flows. The dams were
monitored by the FWCD and cleaned out after such
erosional events.
374
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
5. Volume prediction and comparison with adjacent
basins
The effectiveness of debris flow and erosion control
mitigation at Knight Canyon can also be demonstrated
by showing that the volume of debris produced is
substantially less than the volume predicted, especially
when adjacent canyons which were also burned and
experiencing similar rainfall produced volumes of
debris comparable to predicted values.
5.1. Predicted vs. actual volume of debris flows
Gartner (2005) used multiple regression analysis to
develop an equation to predict the volume from debrisflows in burned areas in the Western U.S., using a
database of 56 debris flows from Colorado, Utah, and
California (nine of the basins in the database had some
form of erosion control treatment, although Knight
canyon was not included in the database):
Fig. 9. Relationship between measured and predicted volumes for
several debris flows in the Western U.S. (from deWolfe, 2006). The
solid line represents a best fit regression and dashed lines represent one
standard deviation from the mean (68% confidence interval). The
measured volume for Knight Canyon is several orders of magnitude
less than the predicted volume.
September 9, 2003. It appears that the concentrated
hillslope and channel treatment applications prevented
this volume of material from mobilizing as a debris flow
in Knight Canyon.
V ¼ EXPð0:65ðlnSÞ þ 0:86ðB1=2 Þ þ 0:22ðR1=2 Þ þ 6:64Þ
ð1Þ
Where:
V
S
B
R
Volume (m3)
Area with slopes z 30% (km2)
Area burned at moderate and high severity
(km2)
Storm rainfall total (mm).
This model was considered the best of several
generated, based on the R2 of 0.83, the residual standard
error of 0.90, the ease in measuring the input parameters,
and support from independent validation with data
points outside the set used in the regression (Gartner,
2005).
Eq. (1) was applied to Knight Canyon to predict the
amount of debris expected from various recorded
rainfalls, and to compare the prediction to actual
volumes of debris produced. Fig. 9 is a plot that
shows the linear relationship between measured and
predicted volumes for the burned basins in Gartner's
dataset (deWolfe, 2006). The most extreme data point on
Fig. 9 is Knight Canyon. The measured volume of
13 m3 was estimated after cleaning out check dam
basins, which is likely a low estimate because some
material may have gone over the spillway and down the
channel. The true volume is certainly substantially less,
however, than the predicted volume of 9854 m3 that
could have mobilized from the 62 mm rainstorm on
5.2. Comparison with Meyer Canyon
After the Missionary Ridge Fire, runoff from four
separate rainstorms produced debris from Meyer
Canyon, which is located adjacent to Knight Canyon
(Cannon et al., 2003). The two basins have almost
identical input parameters for Eq. (1), except that Meyer
Canyon has approximately one-fourth the “S” value
(area with slopes N30%) as Knight Canyon. While
LEBs were constructed in Meyer Canyon, approximately 83% of them were undercut, as shown on Table 1
(Smith, 2004), so the basin is essentially unmitigated.
The first three events occurred in the late summer of
2002 and produced debris flows with boulders up to 1 m
in diameter within a matrix consisting of ash and finegrained material. The flows were deposited on La Plata
County Road 243 and in the Florida River. The
following year, on September 9, 2003, a storm in the
area dropped 62 mm of rain and produced another
sedimentation event from Meyer Canyon. No indication
was found as to whether it was actually a debris flow or
a highly sediment charged flood. While each of these
events occurred independently of one another, the total
volume produced by Meyer Canyon was measured in
the summer of 2004 and calculated to be approximately
3464 m3 (deWolfe, 2006). For comparison, the volume
of debris predicted by Eq. (1) for a 61.75 mm rainfall in
Meyer Canyon is 5345 m3. While the actual volume of
debris was less than the predicted volume, it is within
the expected error range for Eq. (1). The effectiveness of
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
375
On July 31, 2003, an isolated storm produced a
debris flow in South Basin near Lemon Dam, with a
measured volume of 445 m3. Mitigation for debris flows
consisted of two debris racks and LEBs. No check dams
were installed, and the basin was too steep for mulching
or seeding. Eq. (1) predicts 837 m3 for the measured
34 mm rainfall. South Basin also shows an actual
volume of debris less than the predicted volume, which
might be attributable to the LEBs, yet the values are
within the expected error range for Eq. (1). As with the
comparison to Meyer Canyon, this rainfall event had
negligible effect on Knight Canyon.
performance. A third factor contributing to the success at
Knight Canyon was the quality of work for each method:
mulch was spread evenly by hand and crimped into the
ground, LEBs were installed to be in good contact with
the ground, and check dams and debris racks were
carefully engineered for strength, resistance to movement and erosion, and to be large enough to handle the
anticipated volume of debris. Finally, the rehabilitation
of LEBs on several occasions and the cleaning out of
check dams extended the lifespans and improved the
capacities to intercept eroded sediment or moving debris.
Admittedly, the mitigation program at Knight
Canyon was driven by the severe impacts a debris
flow could have on the Durango water supply in Lemon
Reservoir. The finances and manpower committed to the
project are seldom available for many burned areas. The
success of the program shows, however, that the size
and likelihood of a debris flow can be reduced.
6. Discussion and conclusions
Acknowledgements
The debris-flow mitigation methods used at Knight
Canyon virtually eliminated sedimentation into the
reservoir. The produced volume of debris in Knight
Canyon after rainstorms was several orders of magnitude less than what would be predicted without any
mitigation work, as the response of two adjacent
canyons established. South Canyon, where treatment
included only LEBs and debris racks, developed a
debris flow comparable in size to the predicted value.
Meyer Canyon, with a failed treatment program,
experienced three debris flows and one event that was
either a debris-flow or a sediment-laden flood.
The technical literature shows that seeding has
inconsistent effectiveness as a remediation method,
and that mulch and LEBs can be effective, but success is
sensitive to installation methods, basin slope, and
rainfall intensity. At Knight Canyon, enhancement of
methods working in concert was shown. For example,
seeding probably was successful because the mulch held
it in place and retained moisture. These two methods
then reduced runoff and enhanced infiltration, as did
LEBs, which subsequently reduced the mobilization of
sediment, allowing check dams and debris racks to work
without being overwhelmed.
In addition to synergistic effects, the density of application of each method improved its effectiveness: seed
was applied at a density in excess of 150% of typical
values, LEBs in excess of 200%, and mulch up to 125%
in critical areas. It is unclear whether a threshold of
application density exists at which performance becomes
effective or whether increasing density steadily increases
This work was conducted under a grant from the Joint
Fire Science Program, contract 03-1-4-14, as part of a
larger study entitled “Evaluation of Post-Wildfire DebrisFlow Mitigation Methods and Development of DecisionSupport Tools.” The authors are grateful for field data
collection by Morgan McArthur, Adam Prochaska, and
Nate Soule. Jeff Coe and two anonymous reviewers
provided comments that greatly improved the manuscript.
mitigation of debris flows in Knight Canyon is demonstrated as it received similar rainfall amounts, yet
produced almost no debris.
5.3. Comparison with South Basin
References
Amaranthus, M.P., 1989. Effect of grass seeding and fertilizing on
surface erosion in two intensely burned sites in southwest Oregon.
In: N.H. Berg (Technical coordinator), Proceedings of the
Symposium on Fire and Watershed Management, General
Technical Report PSW-GTR-109, U.S. Department of Agriculture
Forest Service, Pacific Southwest Forest and Range Experiment
Station. Berkeley, California.
BAER (Burned Area Emergency Response) Team, 2002a. Coal Seam
Fire: Burned Area Emergency Stabilization and Rehabilitation
Plan. USDA Forest Service, Bureau of Land Management, U.S.
Geological Survey, 250 pp.
BAER (Burned Area Emergency Response) Team, 2002b. Missionary
Ridge Complex, Durango, Colorado June–July 2002: Burned Area
Emergency Stabilization and Rehabilitation Plan. USDA Forest
Service, Bureau of Land Management, U.S. Geological Survey,
251 pp.
Bautista, S., Bellot, J., Vallejo, V.R., 1996. Mulching treatment for
postfire soil conservation in a semiarid ecosystem. Arid Soil
Research and Rehabilitation 10, 235–242.
Beyers, J.L., 2004. Postfire seeding for erosion control: effectiveness
and impacts on native plant communities. USDA Forest Service.
Conservation Biology 18 (4), 947–956.
Beyers, J.L., Wakeman, C.D., Wohlgemuth, P.M., Conard, S.G., 1998.
Effects of postfire grass seeding on native vegetation in southern
376
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
California chaparral. Proceedings, 19th Annual Forest Vegetation
Management Conference: Wildfire Rehabilitation Forest Vegetation Management Conference, Redding, California, pp. 52–64.
Buxton, H., Caruccio, F.T., 1979. Evaluation of selective erosion
control techniques, Piedmont Region of S.E. United States.
Municipal Environmental Research Laboratory Office of Research
and Development U.S. Environmental Protection Agency EPA600/2-79-124, Cincinnati, OH. 107 pp.
Cannon, S.H., Gartner, J.E., 2005. Runoff and erosion generated debris
flows from recently burned basins. Geological Society of America
Abstracts with Program, Salt Lake City, UT.
Cannon, S.H., Gartner, J.E., Holland-Sears, A., Thurston, B.M.,
Gleason, J.A., 2003. Debris flow response of basins burned by the
2002 Coal Seam and Missionary Ridge fires, Colorado. In: Boyer,
D.D., Santi, P.M., Rogers, W.P. (Eds.), Engineering Geology in
Colorado-Contributions, Trends, and Case Histories, AEG Special
Publication 15 Colorado Geological Survey Special Publication,
vol. 55. 31 pp., on CD-ROM.
Chatwin, S.C., Howes, D.E., Schwab, J.W., Swanston, D.N., 1994. A
Guide for Management of Landslide-Prone Terrain in the Pacific
Northwest, 2nd ed. Land Management, Handbook, vol. 18. B.C.
Ministry of Forestry, Victoria, B.C. 220 pp. http://www.for.gov.bc.
ca/hfd/pubs/Docs/Lmh/LMH-09.pdf. Accessed on 2/12/2005.
Chelan County Public Utility District, 2001. Erosion Control Techniques and Permitting, Lake Chelan Hydroelectric Project. FERC
Project No. 637, P.U.D #1, Chelan County Power, Wenatchee,
Washington (revised draft).
Dean, A.E., 2001. Evaluating effectiveness of watershed conservation
treatments applied after the Cerro Grande Fire, Los Alamos, New
Mexico. MS thesis, University of Arizona, Tucson.
deWolfe, V.G., 2006. An evaluation of erosion control methods after
wildfire in debris-flow prone areas. M.S. thesis, Colorado School
of Mines, Golden, CO.
Duffy, J.D., DeNatale, J.D., 1996. Debris flow mitigation using flexible
barriers. Proceedings of the 47th Annual Highway Geology
Symposium. Wyoming Department of Transportation, Cody, WY,
pp. 243–252.
Eisbacher, G.H., Clague, J.J., 1984. Destructive mass movements in
high mountains: hazard and management. Geological Survey of
Canada Paper, vol. 84–16. 230 pp.
Gartner, J.D., 2003. Erosion after wildfire: the effectiveness of log erosion
barriers. M.S. Thesis. University of Colorado, Boulder, CO, 70 pp.
Gartner, J.E., 2005. Relations between debris-flow volumes generated
from recently burned basins and basin morphology, triggering
storm rainfall and material properties. M.S. thesis, University of
Colorado, Boulder, CO, 87 pp.
Gautier, C.R., 1983. Sedimentation in burned watersheds: is emergency
revegetation justified? Water Resources Bulletin 19, 793–802.
Geier-Hayes, K., 1997. The impact of post-fire seeded grasses on
native vegetative communities in central Idaho. In: Greenlee, J.M.
(Ed.), Proceedings of the First Conference on Fire Effects on Rare
and Endangered Species and Habitats. International Association of
Wildland Fire, Fairfield, Washington, pp. 19–26.
Government of Japan, 1984. Basics of planning the measures against
debris flows and planning countermeasure facilities against debris
flow. Ministry of Construction. 9 pp.
Heierli, W., Merk, A., 1985. Barriers for erosion control of mountain
torrents. Proceedings of the International Symposium on Erosion,
Debris Flow and Disaster Prevention, Tsukuba, Japan, pp. 375–380.
Hubbert, K.R., 2005. Treatment effectiveness monitoring For Southern
California Wildfires: 2003 to 2004. The Cedar, Grand Prix/Old,
Piru, and Padua Fires. Hubbert and Associates. 216 pp.
Kay, B.L., 1983. Straw as an erosion control mulch. Agronomy
Progress Report No. 140. University of California Agricultural
Experiment Station, Davis, CA. 11 pp.
Leys, E., Hagen, G., 1971. Design of gravity dams against water
pressure and earth pressure. Imst Regional Construction Office,
Torrent and Avalanche Control, Austria. 39 pp. (in German).
McDonald, G.N., Giraud, R.E., 2002. Fire-related debris flows east of
Santaquin and Spring Lake, Utah County, Utah. Utah Geological
Survey Technical Report 02-09. 15 pp.
Miles, S.R., 2005. Listing of BAER treatments: cost, risk, effectiveness, production rates and applications. Burned Area Emergency
Response Lesson. http://www.fs.fed.us/r5/baer/4g.html, Accessed
on 2/10/2005.
Miles, S.R., Haskins, D.M., Ranken, D.W., 1989. Emergency burn
rehabilitation; cost, risk, and effectiveness. In: Berg, N.H.
(Technical Coordinator), Proceedings of the Symposium on Fire
and Watershed Management, October 26–28, 1988, Sacramento,
California. Gen. Tech. Rep. PSW-109. Berkeley, CA: U.S.
Department of Agriculture, Forest Service, Pacific Southwest
Forest and Range Experiment Station, pp. 97–102. http://www.fs.
fed.us/psw/publications/documents/psw_gtr109/psw_gtr109b.pdf.
Accessed on 10/24/2005.
Moench, R., Fusaro, J., 2002. Soil erosion control after wildfire.
Colorado State University Cooperative Extension, no. 6.308. http://
www.ext.colostate.edu/PUBS/NATRES/06308.html. Accessed on 2/
10/2005.
Reihsen, G., Harrison, L.J., 1971. HEC 9 debris control structures.
Federal highway administration. Hydraulics Branch. http://www.
bossintl.com/products/dpdf/files/ftp/dir/01383/pdf/hec09.pdf.
Accessed on 6/30/2006.
Robichaud, P.R., 2005. Measurement of post-fire hillslope erosion to
evaluate and model rehabilitation treatment effectiveness and
recovery. International Journal of Wildland Fire 14, 475–485.
Robichaud, P.R., 2006. Personal Communication. USDA Forest
Service, Moscow, ID.
Robichaud, P.R., Beyers, J.L., Neary, D.G., 2000. Evaluating the
Effectiveness of Postfire Rehabilitation Treatments. General
Technical Report RMRS-GTR-63. Fort Collins, CO, U.S.
Department of Agriculture, Forest Service, Rocky Mountain
Research Station. 85 pp. http://www.fs.fed.us/rm/pubs/rmrs_gtr63.
pdf. Accessed on 10/14/2005.
Robichaud, P.R., Lillybridge, T.R., Wagenbrenner, J.W., 2006. Effects
of postfire seeding and fertilizing on hillslope erosion in northcentral Washington, USA. Catena 67 (1), 56–67.
Roby, K.B., 1989. Watershed response and recovery from the Will
Fire: ten years of observation. In Berg, N.H. (Technical
coordinator), Proceedings of the Symposium on Fire and
Watershed Management. General Technical Report PSW-GTR109, U.S. Department of Agriculture Forest Service, Pacific
Southwest Forest and Range Experiment Station, Berkeley,
California, pp. 131–136.
Santi, P.M., deWolfe, V.A., Higgins, J.D., Cannon, S.H., Gartner, J.E.,
2008. Sources of debris flow material in burned areas. Geomorphology 96, 310–321 (this volume).
Smith, R., 2004. Personal Communication. USDA Forest Service,
Pagosa Springs, CO.
Taskey, R.D., Curtis, C.L., Stone, J., 1989. Wildfire, ryegrass seeding,
and watershed rehabilitation. In Berg, N.H. (Technical coordinator),
Proceedings of the Symposium on Fire and Watershed Management.
General Technical Report PSW-GTR-109, U.S. Department of
Agriculture Forest Service, Pacific Southwest Forest and Range
Experiment Station, Berkeley, California, pp. 149–161.
V.G. deWolfe et al. / Geomorphology 96 (2008) 366–377
Thommen, R.A., Duffy, J.D., 1997. Testing and application of flexible
wire rope netting barriers to retain mud and debris flow, Geobrugg
Technical Bulletin #10. Geobrugg Protection Systems, Romanshorn, Switzerland. 8 pp.
Thurber Consultants Ltd., 1984. Debris torrents, a review of mitigative
measures. Report to B.C. Ministry of Transportation and Highways, Victoria, B.C. 32 pp.
VanDine, D.F., 1996. Debris Flow Control Structures for Forest
Engineering. Working Paper 08/1996. British Columbia Ministry
of Forests Research Program, Victoria, B.C. 68 pp. www.for.gov.
bc.ca/hfd/pubs/Docs/Wp/Wp22.htm. Accessed on 2/8/2005.
Wagenbrenner, J.W., MacDonald, L.H., Rough, D., 2006. Effectiveness of three post-fire rehabilitation treatments in the Colorado
Front Range. Hydrological Processes 20 (14), 2989–3006.
Wells, W.D., 1987. The effects of fire on the generation of debris flows
in Southern California. In: Costa, J.E., Wieczorek, G.G. (Eds.),
Debris Flow/ Avalanches: Process, Recognition, and Mitigation.
Reviews in Engineering Geology, vol. VII. Geological Society of
America, Boulder, CO, pp. 105–114.
Whittaker, J.G., Schaelchli, U., Jaeggi, M.N.R., 1985. Design
problems with torrent check dams in Switzerland. Proceedings of
377
the International Symposium: Erosion, Debris Flow and Disaster
Prevention, Tsukuba, Japan, pp. 387–394.
Wohlgemuth, P.M., Beyers, J.L., Wakeman, C.D., Conard, S.G., 1998.
Effects of fire and grass seeding on soil erosion in southern California
chaparral. Proceedings, of the 19th Annual Forest Vegetation Management Conference: Wildfire Rehabilitation, Redding, CA, pp. 41–51.
Wohlgemuth, P.M., Beyers, J.L., Conard, S.G., 1999. Postfire hillslope
erosion in southern California chaparral: a case study of prescribed
fire as a sediment management tool. In: Gonzalez-Caban, A., Omi,
P.N. (Technical coordinators), Proceedings of a Symposium on
Fire Economics, Planning, and Policy: Bottom Lines, General
technical report PSW-GTR-173, U.S. Department of Agriculture
Forest Service, Pacific Southwest Research Station, Albany, CA,
pp. 269–276.
Wohlgemuth, P.M., Hubbert, K.R, Robichaud, P.R., 2001. The effects
of log erosion barriers on post-fire hydrologic response and
sediment yield in small forested watersheds, southern California.
Hydrological Processes 15, 3053–3066.
Wright Water Engineers, Inc., 2005. Florida Watershed Post-Fire
Rehabilitation Project, Section 319 Grant Report. Wright Water
Engineers, Inc., Durango, CO.