| United
States
Technology & |
![]() |
Water/Road Interaction: Examples from Three Flood Assessment Sites in Western Oregon |
![]() |
||
| 7700—Transportation
Systems 2500—Watershed and Air Management September 1998 9877 1805P—SDTDC Revised for Internet July 2003 |
||
Water/Road Interaction: Examples from Three Flood Assessment Sites in Western Oregon
Ronald L.
Copstead, P.E.
USDA Forest Service
Pacific Northwest Research Station
David Kim
Johansen, P.E.
USDA Forest Service
Willamette National Forest
Acknowledgments
Information and technical review of this manuscript were provided by the resource
staff of the Willamette National Forest. Background information regarding flood
events was provided by Larry Cronenwett, (retired), formerly of the Engineering
Staff, Pacific Northwest Region, USDA Forest Service.
ABSTRACT
Copstead, Ronald L.;
Johansen, David Kim. 1998. Water/Road Interaction: Examples from Three Flood
Assessment Sites in Western Oregon. Report 9877 1805—SDTDC. San Dimas,
CA: U.S. Department of Agriculture, Forest Service, Technology and Development
Program. 15 p.
Assessments of damage from storms in 1995 and 1996 to three forest road segments on the Detroit Ranger District of the Willamette National Forest are reported. Consequences to roads and road-related structures are discussed. Changes for the three road segments in the designs and materials used for road surfacing and road drainage structures including ditches and cross drains are suggested. Information is based on that developed and reported for other publications in the Water/Road Interaction Technology Series.
Key words: Forest roads, drainage, floods, ditches, culverts, road erosion
Summary — Learning
From Flood Events
From the fall of 1995 through the spring of 1996, intense storms caused widespread
flooding in areas of the Pacific Northwest of the United States. This flooding
provided an opportunity for U.S. Department of Agriculture (USDA) Forest Service
staff to see where the “weak links” were in the integrity of the
system of national forest roads for which they have responsibility. Following
the storms, assessments were made so that road segments could be repaired, rebuilt,
decommissioned, or obliterated. Of particular interest for the Water/Road Interaction
series of publications are examples, afforded by these events, of how forest
road drainage facilities fail when stressed. The following key observations
were made:
Providing “fail-safe” road drainage and stream crossing designs will minimize risk to downstream values during intense runoff periods.
Damage from surface runoff during storms did not tend to be initiated on dense and well-graded gravel road surfaces. Rather, sediment and rock debris from eroded, unlined ditches were initially scattered onto road surfaces. Erosion of these ditches progressed to the point where severe gullying resulted. Ditches eroded rapidly because of the erosive nature (fine-grained, noncohesive) of the soil.
Integrity of the road surface can be enhanced during high-runoff periods by using rock sources that produce well-graded material with adequate plasticity.
For the three sites surveyed, forest vegetation tended to buffer the flow of debris and sediment so that it did not reach large streams. Relatively broad, flat areas adjacent to the eroded section of road (especially at the K-Creek site) caused water to pool and sediment to drop out.
In locations where obstructed crossings carry a high risk to downstream values, it may be necessary to consider structures other than pipe crossings, such as low-water fords.
Culvert inlet areas that had an abundance of vegetation, and that were wide and shallow, had the effect of slowing stream flows, thereby causing debris and sediment to settle and accumulate at the pipe entrance. These inlets often became obstructed.
Shallow fills over bedrock surfaces that are parallel to the slope and adjacent to roads are at high risk to initiate slides during periods of intense runoff. During construction of full-bench roads, preventing these shallow fills over bedrock will reduce the risk of slide initiation. Existing steep, wet, relatively shallow fills should be watched closely by maintenance personnel.
A regular function of routine maintenance of forest roads is to assess road segments for conditions that could lead to damage during periods of high runoff. All too often there is a temptation to repair road damage by simply restoring the road and associated structures to a condition similar to what existed before the storm. In many cases this may be appropriate, but careful evaluation may also suggest improvements that could reduce risk of future damage.
Introduction
Between November 1995 and April 1996, the Pacific Northwest experienced a series
of intense storms. Some of the effects of these storms included high runoff
into drainage systems associated with forest roads. While there was widespread
and extensive property damage and thousands of people were unfortunately affected
in adverse ways, these storms provided natural resource professionals with opportunities
to see firsthand how roads and road-related structures performed in response
to high-intensity events.
During the assessment of effects from the weather events of November 1995 and February 1996, patterns began to emerge regarding damage to roads and adjacent sites. This report describes and discusses examples of road damage caused by storms using three forest road segments in the Detroit Ranger District of the Willamette National Forest. The primary purpose of this publication is to describe the road and road-related structures that were in place at the time of the storm events, highlight what happened to them, and discuss what is needed to improve the design and maintenance of road drainage.
Background
Forest roads and surrounding areas of the Pacific Northwest experienced various
types of storm-caused damage, depending on factors such as soil type, land steepness,
vegetation, storm intensity, and road construction details. Various natural
phenomena, such as channel erosion, slope movement, snow avalanche, and surface
erosion caused damage to roads. Conditions related to activities such as timber
harvest, drainage modification, and harvest and road-related stream channel
modification also caused damage to roads. Damage to forest and streams was also
attributed to culvert plugging, stream diversion, erosion and sedimentation,
and road-fill failures.
Although damage to facilities was costly and in some cases caused inconvenience and hardship, the percentage of land area or length of road that was damaged to the point of requiring repair was relatively small (less than 1 percent of total land area or length of road). National forests that were affected by these storms concluded from their assessments that recent restoration and maintenance efforts probably reduced the number of road-related slides that deposited large amounts of sediment to streams. The flood damage, nevertheless, presented an opportunity to observe road-drainage features that were the most vulnerable to failure and to study the cause of those failures.
Surveys of damage to roads and to local areas surrounding roads on the Detroit Ranger District showed that plugged culverts and road-surface and road-fill erosion accounted for most of the damage to roads and road-stream crossings (figure 1).
Culvert plugging by stream bedload and woody debris was the most common type of failure overall (28 percent). Often a small branch is caught in the culvert inlet, resulting in stream bedload accumulation and eventual burying of the inlet. Culverts that were 600 mm (24 in) in diameter or less accounted for 81 percent of the plugged culverts.
The cause of flood damage to roads from erosion of road surfaces and stream channels adjacent to roads, fill failures, and forest slides contributed about equally to road damage, each comprising 14 to 18 percent of the number of sites that were damaged. Cutslope failures contributed only 6 percent of sites that were damaged.
Damage to roads in the form of fill failures, fill erosion, road surface erosion, and debris piles were all found at about the same frequency and contributed to 16 to 20 percent of the damage.
![]() |
| Figure 1—Road related storm damage by type (352 sites on Detroit RD 1996). |
Site Descriptions and Damage
Assessment
Road surface materials for each of the three sites considered in this report
are similar (table 1). The basic geology in the area consists of mostly andesite
and basalt lava flows intermixed with volcanic deposits of ash and tufts. Soils
are derived from this material and include coluvium and glacial deposits. In
terms of engineering soil gradation, these soils are mostly sand and silt mixtures
(Unified Soil Classifications SM and ML), containing rock fragments up to cobble
size. Low-plasticity soils derived from volcanic deposits tend to be quite erodible.
Surfacing for all roads was derived from local andesite lava flows and was crushed
to 20-mm (3/4-in) minus, dense-graded aggregate. This material also has low
plasticity. Roadside ditches were unlined and partially vegetated, primarily
with grass, ferns, or other small plants.
Breitenbush Site
The damage on this site began where two adjacent small streams carrying high
flows and sand- to cobble-sized bedload plugged road culverts (figure 2). Both
of these 1.0- to 1.5-m- (3.3- to 4.9-ft-) wide stream channels crossed the road
with 450-mm (18-in) diameter corrugated metal pipes lying on a 5 percent (stream)
gradient. The upstream end of the culverts projected into widened inlet basins.
Roadside ditches were open uphill and downhill to these inlet basins. Stream
bedload accumulated at both of the culvert inlet basins, filling them until
their inlets were buried. Storm flow was then diverted to the road ditch. The
flow and stream sediment from the first culvert flowed down to the second culvert,
where they combined and continued down the road ditch. The flow from the slope
above the road and road-surface drainage also accumulated in the ditch. The
high-volume and high-velocity flow—the road and ditch grade ranged from
8 to 11 percent—began eroding the road ditch to a depth limited only by
bedrock lying from 0.5 to 3 m (1.6 to 9.8 ft) below the originally constructed
ditch bottom (figure 3). At the entrances to ditch-relief culverts, either inlet
basins were severely eroded or the entrances were plugged by sediment so that
none of the culverts were transporting water across the road. The accumulated
flow could not escape the ditch and continued down the road for 1.2 km (0.75
mi).
Storm flow eventually eroded through softer road-fill material, forming a 1.3-m-(4.3-ft-) deep gully. Debris was carried into a small, intermittent drainage where some of it was deposited. The remainder was carried further downhill into a larger stream drainage. This larger drainage crossed the road near the junction via the forest highway with a 900-mm (36-in) culvert. The entrance basin to this crossing was filled by debris and the pipe entrance was buried. The stream flow overtopped the road and eroded through the fill approximately 1 m (3.3 ft) down to bedrock, carrying the material 100 m (328 ft) to the forest highway below. Debris deposited on the highway, and the flow spread over the road surface and was dispersed into the forest below.
The dense, well-graded gravel road surface received little damage from surface runoff during the storm. Sediment and rock debris were scattered onto the road surface at some locations. A summary of drainage features and observations of what happened during and after storm events is shown in table 2. Virtually all of the material that eroded and moved offsite came from the unlined ditches and resulting gullying that occurred during the storm (figure 3).
K-Creek Site
Damage at this site began at the 2.7-km (1.7-mi) stream crossing, where debris
filled a stream crossing culvert inlet basin, plugging the 600-mm (24-in) corrugated
metal pipe and diverting the stream into the road ditch (figure 4). As the road
ditch was eroded, debris was scattered on the road surface in numerous places.
Ditch relief culverts were plugged by coarse-grained sediment. A portion of
the water was deflected off the road by this debris, while the remainder flowed
down the road ditch and traveled on the surface, causing additional damage.
Eventually the stream eroded through the road template, creating a gully up
to 3-m (9.8-ft) deep by 3-m (9.8-ft) wide (figure 5). This gully followed the
road until it eroded through the fill, diverting the stream and depositing sediment
in a broad, shallow basin adjacent to the road. The forest vegetation served
as a buffer strip to help capture sediment before it reached stream channels.
The basin allowed water to pool and drop suspended sediment. Location of cross
drains and stream crossing culverts are shown in table 3.
Debris generated from the diverted stream 0.5 km (0.3 mi) away was deposited in the inlet basin for a 600-mm (24-in) diameter culvert near the bottom of the site (at location 3.2 km), plugging the inlet. The stream flow pooled and overtopped the road, eroding the fill and exposing the culverts. A summary of drainage features and observations of what happened during and after storm events at this site is shown in table 3.
Devil’s Creek Site
This site involved the failure of a thin sidecast road fill that scoured the
soil off a drainage headwall on the hillside and deposited it on a lower segment
of the road, within the lower segment of the drainage and within Devil’s
Creek (figure 6). The failure was caused by water being diverted by a small
cutslope slide (estimated volume was 5 m3) that dammed a ditch, causing ditch
water to run out onto the road surface, down wheel tracks, and eventually over
the fill slope edge of the road (figure 7). The fill slope apparently failed
as a result of saturation caused by rainfall combined with the overtopping ditch
water eroding the toe of the fill. The failed fill slope became a fast-moving
earth flow that combined with wet 1- to 5-m- (3.3- to 16.4-ft-) thick surface
soil in the drainage headwall, gathering volume as it moved and not stopping
until it encountered the road below. The debris then filled the inlet basin
of the road culvert, covered the road with about 800 m3 (28,251.7 ft3) of debris,
and sent the bulk of the debris into the drainage channel below the road and
into Devil’s Creek. The channel debris swept up a forest slope, knocking
down trees and burying the channel. The drainage eventually eroded through the
channel debris to reconnect with Devil’s Creek. It was estimated that,
of the approximately 4,200 m3 (148,321.6 ft3) that were eroded and subsequently
slid from the site, about 25 percent was stopped by the road, with the remaining
volume ending up within the stream bed below. The road fill was partially eroded,
damaging the outlet end of two 750-mm- (30-in-) diameter stream culverts located
at the site. Apparently, one of the culverts had been damaged in the past by
a debris slide, left in place, and replaced with a new culvert of the same size.
Subsequent maintenance at the site removed debris from the road surface, but did not re-establish flow into the cross drain. Instead, water flowed into the road ditch and continued about 70 m (229.7 ft) down to the next culvert. The sediment carried by the diverted drainage eventually plugged the next culvert, then diverted over the road surface, eroding approximately 3,000 m3 (105,944 ft3) of road fill. The fill material was deposited in a flattened area below the road and within a forested buffer strip area below. Very little road fill reached Devil’s Creek drainage below. The road culvert was severely damaged and the road subgrade was destroyed.
| Table
1—General description of three flood-assessment sites on
the Detroit Ranger District. Willamette National Forest, Oregon. |
|||
|---|---|---|---|
Site |
Breitenbush |
K-Creek |
Devil's
Creek |
| Road segment location | FS road
4600-040, km post 0.0 to 1.55 |
FS road
1003-416, km post 2.7 to 3.2 |
FS road 2231-870, |
| Type of construction and surface drainage features | Cut and
fill Section 8-11% single lane crown surface v-ditch, 450 mm culverts aggregate surfacing |
Cuts and
fills 4-10% single lane crown surface v-ditch, 450 mm culverts aggregate surfacing |
Full bench 2-4% single lane crown surface v-ditch, 450 mm culverts aggregate surfacing |
| Position on slope | Lower 1/3 |
Upper middle
1/3 |
Upper 1/3 |
| Elevation (m) | 670-850 |
825-850 |
1,340 |
| Typical overland flow path length above top culvert inlet (m) | 975 |
1,585 |
[Did not
involve a culvert failure] |
| Typical overland flow path slope above top culvert inlet (percent) | 28 |
23 |
[Did not
inovlve a culvert failure] |
| Two-year, 24-hour rainfall intensity (mm) | 89 |
89 |
102 |
| Predominant vegetation above the road | 20 to 30
year-old second-growth forest |
5 to 10
year-old second-growth forest |
Clear-cut
(1990) |
| Estimate of material eroded (m³) | 1,900-2,700 |
1,250 |
4,200 |
| Material estimated that entered stream (m³) | 1,700-2,500 |
< 50 |
3,500 |
![]() |
Figure
2—Breitenbush site map. |
![]() |
Figure
3—Severe gullying at Breitenbush flood-assessment site.
|
| Table
2—Cross drain culvert locations and characteristics on
Breitenbush Road prior to April 1996 and after 1996 storm events. |
||||||
|---|---|---|---|---|---|---|
Culvert
size & type |
Distance
to next culvert |
Road grade |
Road grade |
Drainage
area |
Condition
after storm events |
Apparent
cause |
meters |
percent |
hectares |
||||
| 450 mm relief culvert | 232 |
12 |
27 |
open |
||
| 450 mm relief/stream culvert | 137 |
12 |
21 |
plugged |
bedload |
|
| 450 mm relief/stream culvert | 107 |
11 |
2 |
plugged |
bedload |
|
| 450 mm relief/stream culvert and slope drain | 174 |
11 |
2 |
eroded around |
large ditch
flow |
|
| 450 mm relief/stream culvert | 137 |
9 |
2 |
plugged |
sediment |
|
| 450 mm relief culvert | 265 |
9 |
2 |
open |
ditch water
eroded through fill before pipe |
|
| 450 mm relief culvert | 70 |
11 |
2 |
plugged |
sediment |
|
| 450 mm relief culvert | 143 |
12 |
2 |
plugged |
sediment |
|
| 450 mm relief culvert | 107 |
13 |
2 |
open |
culvert
inlet overtopped |
|
| 600 mm stream culvert | 183 |
13 |
70 |
plugged |
stream bedload
and ditch sediment |
|
| 450 mm relief culvert | -- |
-- |
2 |
plugged |
ditch sediment |
|
![]() |
Figure
4—K-Creek site map. |
![]() |
Figure
5—Severe gullying at K-Creek flood-assessment site. |
| Table
3—Cross drain culvert locations and characteristics on
K-Creek road prior to April 1996 and after 1996 storm events |
|||||
|---|---|---|---|---|---|
Culvert |
Distance
to next culvert |
Road
grade |
Drainage
area |
Condition
after storm events |
Apparent
cause |
meters |
percent |
hectares |
|||
| 600 mm stream and ditch relief culvert | 85 |
9 |
63 |
plugged |
stream debris
and bedload, flow ran past here and into ditch |
| 450 mm ditch relief culvert | 43 |
8 |
3 |
plugged |
bedload,
ditch and gully sediment |
| 450 mm ditch relief culvert | 311 |
4 |
8 |
plugged |
bedload,
ditch and gully sediment |
| 450 mm ditch relief culvert | 7 |
4 |
2 |
plugged |
bedload,
ditch sediment |
| 600 mm stream and ditch relief culvert | -- |
-- |
193 |
plugged |
bedload,
ditch and stream sediment, overtopped the road surface |
![]() |
Figure
6—Devil's Creek site map. |
![]() |
Figure
7—Initiation of debris slide at Devil's Creek flood-assessment
site. |
Discussion
Plugged Stream Culverts
The site damage at the Breitenbush and K-Creek sites started from stream crossings
that were overwhelmed with eroded bedload and debris, causing the crossings
to become plugged. In general, a culvert stream crossing must meet hydraulic
requirements and accommodate bedload and debris. This can be done by sizing
the culvert to meet hydraulic requirements and then either increasing capacity
to accommodate bedload and debris, or providing a means for accumulating this
material at the inlet so that it can be removed by maintenance crews. It is
preferable to provide a means for the material to pass through the crossing.
To pass bedload and debris, inlet basins should be narrow and straight, with
the culvert and stream at the same grade. The culvert should be as large as
possible, up to the stream channel width and depth. If this is not possible,
a storage inlet basin should be considered. In locations where obstructed crossings
carry a high risk of damage to downstream values, it is necessary to consider
structures other than culvert crossings, such as low-water fords. Even when
culvert crossings are used, an armored overtopping path for diverted flow should
be considered.
For some cases, stream culvert crossings are needed where downstream values are at low risk, upstream geometry is broad and relatively flat, or the potential for debris and bedload flow at the crossing is determined to be small, and, therefore, that storage inlet basins will provide adequate protection from culvert plugging. In these cases, there must be reasonable assurance that inlet basins can be cleaned out on regular maintenance schedules. Because the size and quantity of bedload and woody debris that may impinge on stream crossings during storms often is not predictable, it is difficult to determine minimum standard dimensions for storage inlet basins, and often they are simply made as large as possible.
Small streams and drainages should be studied to determine drainage area and expected flow for storm events. Many small but recognizable drainages have enough volume to require culverts much larger than the 450-mm- (18-in-) diameter pipes that were prevalent at the three subject sites. For example, the small drainages at the top of the Breitenbush site have drainage areas of 32 ha (79.1 a) each. The size of a stream culvert capable of passing a 100-year storm flow of 1.11 m3 (39.2 ft3) per sec, with a headwater height equal to the culvert diameter at a gradient of 5 percent is 900 mm (36 in). The culverts installed were 450 mm (18 in) and, thus, substantially undersized for the stream’s hydraulic requirements. Debris considerations would have indicated the need for an even larger culvert that could pass flood flows and debris at the same time. A more appropriate size would have been 1,200 mm (48 in). Another approach to consider in this case is a crossing design that allows water, debris, and bedload to flow over the top of the traveled-way surface.
The stream crossing at the lowest elevation at Breitenbush was plugged from
stream bedload and ditch sediment. The estimated peak 100-year flood discharge
for this site was 3.45 m3 (121.8 ft3) per sec (Harris, et al. 1979). Based on
accepted practice for calculating culvert size for a 5-percent culvert gradient,
a 1,200-mm- (48-in-) diameter culvert would have been required to prevent overtopping
the road surface, which was located only 1.5 m (59 in) above the existing 600-mm
(24-in) culvert inlet bottom (Normann, et al. 1985). Therefore, this culvert
was considerably undersized for its drainage. The culvert inlet was wide, shallow,
and heavily vegetated, which had the effect of slowing stream flows and causing
debris to settle and accumulate at the culvert inlet.
Table 4 shows a comparison of a sample of the installed culverts with sizes
calculated to meet 100-year flood discharge. It is likely that if the stream
crossings had been sized to accommodate a 100-year storm flow, the larger culvert
inlets would not have become plugged, and less water would have been diverted
from upper stream crossings to lower crossings.
Table 5 shows other features that should be considered for incorporation into crossing culvert designs in this publication such as those at the three sites considered.
| Table
4—Comparison of the size of installed culverts with sizes
calculated to meet 100-year flood discharge. |
|||
|---|---|---|---|
Road |
Purpose |
Existing
Culvert Diameter |
Culvert
Diameter for Q100 |
millimeters |
millimeters |
||
| Breitenbush | stream | 450 |
900 |
| Breitenbush | stream | 450 |
900 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | ditch relief | 450 |
450 |
| Breitenbush | stream | 600 |
1,350 |
| Breitenbush | ditch relief | 450 |
450 |
| K-Creek | stream | 600 |
1,200 |
| K-Creek | ditch relief | 450 |
450 |
| K-Creek | ditch relief | 450 |
600 |
| K-Creek | ditch relief | 450 |
450 |
| K-Creek | stream | 600 |
1,650 |
| Table
5—Features to consider when designing small stream crossings,
including ditch-relief culverts. |
|---|
|
Culvert Inlet Design
As discussed above, inlet design can have an effect on how bedload and debris
behave at crossings. Inlets should be designed to direct water into culverts
without slowing it down. Matching the culvert diameter to the cross-section
area of the stream channel can minimize changes to flow velocity at the approach
to the inlet. An inlet basin that is much larger than a culvert inlet allows
settling and collection of material at the inlet during storms because it allows
the water to slow down, reducing its transporting ability. The inlet area should
be shaped to match the stream channel, or slowly tapered to the culvert diameter,
to keep water velocity sufficient to help debris flow through the culvert.
Culvert Spacing
The cross drains at all three sites were not close enough to each other to carry
the flow contributed by their respective drainage areas. For example, at Breitenbush,
based on the average cross-drain drainage area of 16 ha (39.5 a) at this site,
the diameter of a cross-drain culvert capable of passing a 100-year storm flow
without overtopping the road surface would be 600 mm (24 in). By spacing cross
drains closer together, the pipe diameter required to meet hydraulic requirements
could be reduced to 450 mm (18 in).
During the storm, even if debris had not plugged culvert inlets, cross-drain
culverts at Breitenbush and K-Creek would have overtopped the road because flow
would have backed up and pooled at the inlet. The rise in water would redirect
a portion of the flow across the road or down the ditch, similar to what happened
during the storm.
The average spacing of cross drains for Breitenbush, including stream-crossing
culverts (which also function as ditch relief) is greater than 155 m (508.5
ft) (see table 2). If cross drains had been spaced according to published guides
(Baeder and Christner 1981), spacing would have been between 25 and 60 m (82.0
and 196.9 ft), and the storm flow that each of the 450-mm (18-in) culverts would
have to have carried would have been reduced by 70 percent. Had they been installed
at a minimum 5-percent gradient, they probably would not have been overtopped.
Ditch erosion, gully formation, and culvert inlet plugging would probably have
been reduced with closer spacing.
The cost for the additional 26 culverts needed to achieve an average cross-drain spacing of 42 m (137.8 ft) would be about $19,500. The marginal cost for the larger-sized stream culverts would be approximately $4,000. The combined total of $23,500 is 21 percent of the estimated $110,000 needed to restore the transportation function of the road.
Ditch Erosion
The soils at both Breitenbush and K-Creek are easily eroded. Ditches would have
benefited from either more frequent relief culverts or from erosion protection
that could have been provided by grass or small riprap. Natural ditch vegetation
was insufficient and natural armoring through erosion of fines, leaving coarse
rock fragments for protection, did not occur at either site. The high-energy
conditions and high-volume flow at the Breitenbush site produced very high erosional
forces. It would be expensive and difficult to provide adequate protection for
those conditions. At the K-Creek site, the high-flow conditions resulted in
erosion of the unprotected ditch. In both cases, the road surfacing provided
some protection from erosion and gully formation into the road. The key to preventing
ditch erosion at these sites would appear to be in preventing plugging of stream
culverts and diversion (see table 5). Other important road-drainage features,
such as closer spacing of cross drains could have reduced overall road damage,
since they would have provided additional places for diverted stream flow and
drainage to escape. Ditch dams would have been helpful in getting ditch water
into ditch relief culverts, especially at steeper grade sites.
Fill Failure
Damage to the Devil’s Creek site resulted from a failed cut bank plugging
a ditch. Wheel tracks carried the diverted water down the road until a surface
depression was reached. The effect was to saturate a fill, which then initiated
a slide at the toe of a slope. This small slide developed into a debris chute.
The road is a full-bench design that had a shallow layer of soil sidecast as
waste during construction (a “sliver fill”). The sidecast soil lies
on a thin mat of soil and organic matter (partially decomposed by now) such
as stumps and brush. This created a zone of weakness that failed more readily
than other soil in the area.
The topography and management regime below the road probably contributed to the failure. The area below the road was a 55 to 70 percent side slope, 10-year-old, clear-cut unit. Typically, tree roots remaining in the soil in this type of area have significantly rotted, losing shear strength that helps hold soil on the slope. The absence of large trees also meant there was little to stop a slide once it started. The underlying bedrock mostly parallel to the slope, acted as a slip zone for the soil. This situation suggests that maintenance personnel should carefully monitor roads in similar areas with steep, wet, relatively shallow soils with thin fills. Unstable fills should be pulled back before they can fail. Also, during full-bench construction, efforts to pull back fill spilled over the side may eliminate potential failures.
The $5,000 to fix the upper road (clean the ditch and the ditch relief culvert) was small compared to the value of other resources that could have been damaged. In contrast to the other two sites, the role of the culvert/drainage design for this road was incidental to the initiation of the failure. The relatively small cut-slope failure that initiated this damage is an example of how seemingly isolated and innocuous events can trigger much more catastrophic results when “fail-safe” drainage designs are not used.
Traveled-way Surface and Ditch Shape
Traveled-way surface and ditch shape can play an important part in mitigating
damage caused by floods of this magnitude. The Breitenbush and K-Creek sites
both had crowned road surface shape. The advantage of this with respect to runoff
is that the effective capacity of the inside ditch is increased somewhat because
it includes the half of the road that is insloped. Half the surface water is
directed to and over the outside shoulder of the road instead of along the direction
of travel or to the inside ditch, as on an insloped road surface. While water
tended to spread from the ditch over the road where the road grade flattened,
dips in the road surface would have prevented excessive concentration of runoff
and made the design more fail-safe by directing the water over and off the road
at predetermined locations. Also, none of the three sites had ditches that incorporated
ditch dams, which could have directed water into relief culverts, reducing the
accumulation of ditch flow during moderate storms.
At the Devil’s Creek site, the upper road damage would probably have been
prevented if the traveled-way surface had been insloped rather than crowned.
With an insloped surface, all the surface water on the road is directed toward
the cutslope, unless a slide covers the entire road. In this case, the diverted
water would have flowed around the debris slide into wheel tracks and eventually
run into the roadside ditch. In the case of the Devil’s Creek site, however,
the crowned surface caused the flow diverted by the cutslope slide to spill
onto a fill slope that was not armored to prevent erosion, and that probably
would have been susceptible to saturation (and failure) regardless of any measures
that were taken.
Road-surfacing Materials
The properties of road-surfacing materials are often a factor in how well a
road withstands severe flood events. The road segments considered at the three
sites were all aggregate surfaced, but the surfacing on top (the most recently
applied) was not capable of withstanding the velocity of the flows during these
storm events. In several locations, older aggregate surfacing that became exposed
only after the newer, top layer of surfacing was washed away, stayed intact.
This older surfacing tended to have more fines, greater plasticity, was better
consolidated, and was consequently able to hold up better to the flood flows.
The newer surfacing had fewer fines and less plasticity. These differences could
be a result of differences in the materials as they were initially installed,
or of changes in characteristics that develop in any surfacing material as it
becomes buried under newer material. Typically, surface material is subject
to the deconsolidating influence of vehicle traffic and maintenance equipment,
which tends to reduce the shear-resisting properties of any surfacing material.
To protect the integrity of the road surface during storms (which may result
in failed drainage systems), rock sources that produce well-graded material
with adequate plasticity should be chosen. Management of the quantity and characteristics
of vehicle traffic should also be considered. For example, reducing vehicle
tire pressures has been shown to reduce the degradation of traveled-way surfaces.
Conclusions, Summary, and Recommendations
Many lessons can be learned regarding forest roads from flood events such as
those that occurred on the Detroit Ranger District during 1995 and 1996. While
it is difficult to predict where damage will occur, risk can be assessed, and,
in these particular cases, could have resulted in preventive measures that would
have reduced damaging effects.
While it is not expected that all damage from intense storms will be avoided, damage can be minimized. The three sites featured in this report provided good examples of situations where relatively low-cost changes in drainage design and maintenance practices could pay large dividends in reducing flood-repair costs.
It is apparent from the study of these three sites that seemingly small details can lead to unexpectedly large and catastrophic failures. Some examples became apparent during the assessment of damage at the three sites covered in this report:
When these roads were built (approximately 1960), it was assumed that maintenance would be adequate. Regular, properly performed maintenance is intended to ensure that details such as road surface and ditch condition do not degrade to the point that catastrophic damage occurs during storm periods.
Plans for reconstruction, maintenance, or new road construction should consider the risk and consequence of culvert failures, implementing measures to ensure against failures that pose risk to resource values and fail-safe road drainage design. Specifically, this means analyzing the potential for diversion of stream water to the road-drainage system; looking for and avoiding the possibility of progressive failure of down-grade culverts; planning for buffer strips; using appropriate surfacing and ditch materials; designing adequate culvert spacing; and considering local slope stability when specifying traveled way surface shape.
On these sites, the planning for drainage should have included:
The three flood-assessment sites considered illustrate the need for planning and designing road-related drainage according to the following general guides:
Literature
Cited
Baeder, Larry and
Jere Christner. 1981. Revision of the guide for spacing relief culverts for
the Willamette National Forest. Eugene, OR: USDA, Forest Service, Pacific Northwest
Region, Willamette National Forest; 18 p.
Copstead, Ronald, David Kim Johansen, and Jeffry Moll. Water/Road Interaction: Introduction to surface cross drains. In: Water/Road Interaction Technology Series, Publication 9877 1806. San Dimas, CA.: USDA, Forest Service, San Dimas Technology & Development Center. 15 p.
Furniss, Michael J., Michael Love, and Sam A. Flanagan. 1997. Diversion potential at road-stream crossings. In: Water/Road Interaction Technology Series, Publication 9777 1814. San Dimas, CA.: USDA, Forest Service, San Dimas Technology & Development Center. 12 p.
Harris, D. D, Larry L. Hubbard, and Lawrence E. Hubbard. 1979. Magnitude and frequency of floods in western Oregon. Open-File Report 79-553. Portland, OR: U.S. Department of the Interior, U.S. Geological Survey. 29 p.
Normann, Jerome M., Robert J. Houghtalen, and William J. Johnson. 1985. Hydraulic design of highway culverts. Hydraulic Design Series No. 5, FHWA-IP-85-15. McLean, VA: Turner-Fairbank Highway Research Center, Office of Implementation. 272 p.
Multiply |
by |
To get |
|---|---|---|
mm (millimeters) |
0.0394 |
in (inches) |
cm (centimeters) |
0.394 |
in (inches) |
m (meters) |
39.4 |
in (inches) |
m (meters) |
3.28 |
ft (feet) |
hectares |
2.47 |
ac (acres) |
m³ (cubic
meters) |
1.31 |
yd³
(cubic yards) |
Metric |
English |
|---|---|
450 mm |
18 in |
600 mm |
24 in |
750 mm |
30 in |
900 mm |
36 in |
1,200 mm |
48 in |
1,500 mm |
60 in |
Project Leader San Dimas Technology & Development Center 444 East Bonita Avenue, San Dimas CA 91773-3198 Phone 909-599-1267; TDD: 909-599-2357; FAX: 909-592-2309 E-mail: mailroom_wo_sdtdc@fs.fed.us Information contained in this document has been developed for the guidance of employees of the U.S. Department of Agriculture (USDA) Forest Service, its contractors, and cooperating Federal and State agencies. The USDA Forest Service assumes no responsibility for the interpretation or use of this information by other than its own employees. The use of trade, firm, or corporation names is for the information and convenience of the reader. Such use does not constitute an official evaluation, conclusion, recommendation, endorsement, or approval of any product or service to the exclusion of others that may be suitable. The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, sex, religion, age, disability, political beliefs, sexual orientation, or marital or family status. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communication of program information (Braille, large print, audiotape, etc.) should contact USDA’s TARGET Center at (202) 720-2600 (voice and TDD). To file a complaint
of discrimination, write USDA, Director, Office of Civil Rights, Room
326-W, Whitten Building, 1400 Independence Avenue, SW, Washington, D.C.
20250-9410 or call (202) 720-5964 (voice and TDD). USDA is an equal opportunity
provider and employer. |