United States Department of Agriculture Forest Service Technology & Development Program		Methods for Inventory and Environmental Risk Assessment of Road Drainage Crossings
7700—Transportation Systems 2500—Watershed and Air Management December 1998 9877 1809P—SDTDC Revised for Internet July 2003

Methods for Inventory and Environmental Risk Assessment of Road Drainage Crossings

Sam A. Flanagan Geologist

Michael J. Furniss, Hydrologist

Tyler S. Ledwith, Hydrologist

Stan Thiesen, Geologist

Michael Love, Engineer

Kemset Moore, Hydrologist

Jill Ory, Hydrologist

USDA Forest Service
Six Rivers National Forest
Pacific Southwest Region

• Acknowledgements
• Introduction
• The Environmental Risk of Road-Stream Crossings and the Need for Inventory and Assessment
  • What are the Environmental Risks of Road-Drainage Features?
• Review of Existing Road-Stream Crossing Assessment Techniques
• Suggestions for Inventory and Assessment
  • Materials Required
  • Time Required
  • Using a Global Positioning System for Crossing Inventory
• Data Analysis and Interpretation
  • Using a Geographic Information System in Crossing Assessment
  • Calculating Hydraulic Capacity
  • Woody Debris Capacity
  • Sediment Capacity
  • Erosional Consequences
  • Estimating Fill Volume
• Hydrologic Connectivity—Assessing Chronic Erosion Potential
  • Assessment Procedure
  • Connectivity/cross drain inventory
  • Consequences inventory
  • Hazard assessment
  • Environmental risk assessment
  • Environmental risk = Hazard^Endpoints
  • Limitations
  • Evaluating consequences by road segment
  • Failed Crossing Assessments
  • Mechanisms of Road-Stream Crossing Failure
• Literature Cited
• Appendixes
• Appendix AAn Example of Road Environmental Risk Assessment for USDA Forest Service Lands in the Upper North Fork Eel River Watershed, Six Rivers National Forest
• Appendix BSample Data Sheet for Road Drainage Inventory

Introduction
Road-stream crossings and ditch-relief culverts are commonly sites of ongoing or potential erosion. Erosion from failures of these structures can be a source of significant impacts to aquatic and riparian resources far removed from the initial failure site. The inventory and assessment of culvert installations are necessary for locating sites with potential impacts to aquatic and riparian environments for possible treatment. Locating those sites with the greatest potential for causing adverse impacts and reducing or eliminating those impacts is critical in prioritizing restricted funding available for road maintenance and repair.

Procedures for assessing the erosional hazards and risks to aquatic and riparian ecosystems of road-stream crossings, ditch-relief culverts, and other road-drainage features are discussed. Hazard assessment is the estimation of the potential physical consequences (e.g., volume of material eroded) for one or more sites. Incorporation of the potential impacts to valued resources, or endpoints, is environmental risk assessment. This approach is intended to suggest appropriate data and assessment techniques that can be adapted to a wide range of settings.

This guide is designed to:

• Review inventory and analysis methods in current or recent use in wildland settings
• Suggest various inventory procedures for the assessment of existing culvert installations and other road-drainage features, both unfailed and failed
• Provide analytical methods for predicting the potential hazards and environmental risks of road-drainage failure
• Provide background information on road-stream crossing performance.

The assessment methods discussed are useful for a screening approach where data on a large number of drainage features are collected, typically at the scale of a watershed less than 500 km². Culverts, the focus of this guide, are the most common structure employed for wildland road-stream crossings and ditch drainage. However, the procedures discussed are also adaptable to other road-drainage structures.

The following road crossing terminology is used:

Road-stream crossing is where a road crosses a natural drainage channel or unchanneled swale.

Stream crossing is used synonymously with road-stream crossing.

Cross drain is a ditch-relief culvert or other structure such as a grade dip designed to capture and remove surface water from the traveled way or other road surfaces.

Crossing is used here to cover both stream crossings and cross drains.

Hydrologically connected road is a segment of road that is connected to the natural channel network via surface flowpaths.

The Environmental Risk of Road-Stream Crossings and the Need for Inventory and Assessment
Road-stream crossings are common throughout wildland road systems. Road-stream crossings have enormous potential for erosional consequences. For federally managed lands in the Pacific Northwest, the Forest Ecosystem Management Assessment Team (FEMAT) (1993) estimated that although 250,000 road-stream crossings exist, systematic analysis of cumulative environmental risks are rare. Most of these structures can function as earthen dams, with a small hole or culvert at the base. Such configurations are rare in natural channels. In the absence of maintenance and replacement, all of these structures will eventually fail as they plug or the culvert invert deteriorates. Financial resources for maintaining or upgrading the existing network are limited. Therefore, locating sites that pose the greatest risk to aquatic and riparian resources and treating them to reduce hazards and environmental risks is vital.

Crossings can cause both chronic sedimentation impacts during typical water years and catastrophic effects when floods trigger crossing failure (figure 1). When water overtops the road fill, it may divert down the road or ditch and onto hillslopes unaccustomed to concentrated overland flow and produce erosional consequences far removed from the crossing. Erosion from diverted stream crossings in the 197-km² lower Redwood Creek basin in northwestern California accounted for 90 percent of the total gully erosion (Weaver et al. 1995). Best et al. (1995) found that road-stream crossings accounted for 80 percent of all road-related fluvial erosion in a 10.8-km² tributary basin to lower Redwood Creek. Erosion from failed or improperly designed road-drainage structures accounted for 31 percent of the total sediment inputs (Best et al. 1995). Identifying those sites with the greatest potential of erosional consequences (e.g., potential to divert stream flow) can direct restoration efforts.

Abandoned roads represent an unknown, but potentially high, erosional hazard. Failures often go unseen in the absence of inspection or maintenance. Where stream diversions have occurred, erosion is likely to continue for decades as new stream channels are incised. Determining the extent of abandoned roads in an area is necessary for assessing the entire road-related erosional hazards.

Previous assessments of existing road-stream crossings have examined the hydraulic capacity of the culverts (Piehl et al. 1988 and Pyles et al. 1989). These results show that culverts were installed without consistent design standards and that design flow capacities were often below standards established by the Oregon Forest Practice Rules (Piehl et al. 1988). In a recent survey of road-stream crossings in Redwood Creek, CA, 60 percent of the culverts could not pass a 50-year peak flow (Redwood National Park, unpublished data).

Nonstream crossing drainage structures are also of concern. These include ditch-relief culverts, waterbars, and rolling dips. An expanded discussion on rolling dips is provided by the companion document in this series, “Diversion Potential at Road-Stream Crossings” (Furniss et al. 1997). Detailed design considerations for dips are provided by Hafterson (1973). Not only do these structures present similar hazards and environmental risks, such as those discussed for road-stream crossings, they are also capable of extending the natural drainage network, if not properly configured (Wemple 1994). Extension of the drainage network occurs when ditch flow and road-surface runoff are conveyed to a stream channel. Where roads are hydrologically connected to the drainage network, road-produced sediment and runoff are delivered directly to the channel network. Hydrologic connectivity often involves extensive gullies linking roads and channels (Wemple 1994).

Assessing the hazards and environmental risks of crossings involves substantial inventory costs. Roads are not a direct reflection of the underlying landscapes. Thus, remote sensing techniques cannot provide necessary information on existing hazards. Simply locating the installed system by using remote sensing techniques provides little information on potential erosional consequences. Several techniques are presented for inventorying roads and crossings at various levels of effort.

Figure 1—Road-stream crossings can alter channel form and processes. They are sites of both chronic and catastrophic erosion.

What are the Environmental Risks of Road-Drainage Features?
It is useful to think about the environmental risk of road-stream crossings, cross drains, and other drainage structures (e.g., waterbars and rolling dips) in four parts (figure 2). The four parts of the environmental risk assessment model used in this guide are:

Inputs: The materials presented to the crossing—water, organic debris, sediment, and fish. (This document does not discuss fish/amphibian passage.)

• Capacity: The ability of the culvert to pass the inputs.
• Consequences: The physical effects of capacity exceedance, often expressed as a volume of material eroded (or deposited).
• Endpoints: The valued resources. Some endpoints may be:
  - Fish or other species of concern
  - Domestic water supply
  - Aquatic species refugia value (e.g., cool waters in a warm water basin, access to spawning areas).

The hazard of a crossing failure is a combination of the inputs and the structure’s capacity to accommodate them, which defines the probability of exceedance, and the potential physical consequences of exceedance. Capacity must be considered separately from the potential erosional consequences. In the absence of maintenance or replacement, all crossings will eventually fail. Therefore, it is imperative that potential erosional consequences not be ignored simply because capacity has been judged “sufficient.” Alternately, a site with a very limited capacity may possess relatively little erosional hazard. Treatments to increase capacity at such a site may be impractical and ineffective in reducing the cumulative environmental risk of crossings (figure 3).

Environmental risk of a crossing is the combination of the existing hazard and endpoints. Environmental risk assessment identifies the relative likelihood of a site adversely affecting one or more endpoints.

Figure 2—Road-stream crossing and cross drain environmental risk can be expressed as a four part model consisting of inputs—those materials delivered to the culvert; capacity—the ability of the structure to maintain the natural transport regime of the delivered inputs; physical consequences—the erosional and/or depositional consequences occurring when capacity is exceeded; and endpoints—potentially affected aquatic and riparian resources, human uses, and other values.

Figure 3—The physical consequences of this road-stream crossing failure are low when compared to sites with much larger fills or where stream diversions occurs.

Broda and Shockey (1996) of the Umpqua National Forest use a “risk rating table” (table 2) to rank road-stream crossings on the Umpqua National Forest.

Using this approach, each crossing is assigned a score from 1 to 5. Factors considered in this approach are shown in table 2.

Hazards

• Hydrology/hydraulics—flows, culvert capacity
• Plugging potential—wood and debris (soil and rocks)
• Slope stability—channel slopes, road fill.

Physical consequences/endpoints

• Sediment delivery (amount)
• Resource damage—trees, fish, habitat (cost)
• Capital investment losses (cost).

Hanek (1996), also of the Umpqua National Forest, uses a similar approach (table 3). Factors considered in this approach in order of decreasing priority are:

Effects

• High
  - Large drainages (1.2 km²+) with diversion potential
  - Medium drainages (0.4 to 1.2 km²) with diversion potential.
• Medium
  - Large drainages without diversion potential but with larger fills (i.e., over 1.5 m from inlet invert to grade)
  - Medium drainages without diversion potential but with larger fills.
• Low
  - Everything else (i.e., large and medium drainages without diversion and small fills and all small drainages (< 0.4 km²).

Hazards

• Hydraulic capacity of the culvert and associated fill prior to overtopping or diversion (using 100-year design storm)
• Culvert plug potential
• Existing culvert condition
• Embankment stability under no-pool, full-pool, and half-pool conditions
• Diversion potential.

An approach used by Weinhold (1996) of the Siskiyou National Forest involving potential eroded fill volume is shown in table 4.

Here, high failure potential is represented by an undersized culvert, evidence of past plugging by debris, and slope and channel having the potential to generate debris flows.

Stuber et al. (1994), working in the Huron-Manistee National Forest in Michigan, assigns each crossing a severity ranking based on a point system (table 5).

Scores over 30 points are placed in the severe category. Scores under 15 points are considered to be minor.

The Mt. Baker-Snoqualmie National Forest (1997) uses a road segment approach where crossings are one component of the method (table 6). An effects-of-failure score (K) is a combination of sediment delivery and valued resources. It functions as a multiplier to the failure potential score.

Determining K involves the use of a short dichotomous key (table 7). The score is based on the proximity of the road to streams, wetlands, infrastructure, or other valuable natural resources.

Inglis et al. (1995) of the Payette National Forest ranks the condition of stream crossings as either high, medium, or low. Crossings are rated high priority if three or more of the following criteria are met:

• The stream at the crossing is fish bearing;
• Road use is heavy;
• There is a chance of road loss at the crossing; or
• The culvert at the crossing is failing.

Medium priority crossings have to meet at least two of the following criteria:

• The stream at the crossing is nonfish bearing;
• Road use is low;
• There is little to no chance of road loss at the stream crossing, and
• The culvert at the crossing is not failing.

All other crossings needing work are listed as low priority.

The following matrix (table 8) is part of a larger road assessment package developed by Kennard (1994) for the Weyerhauser Corporation. A “weight of evidence” approach shown below is taken to assess the hazard from combinations of indicators.

Table 10-—Summary of suggested data fields for various levels of road-stream crossing and cross drain inventory and assessment.

Suggestions for Inventory and Assessment
The road-drainage network can be inventoried at various intensity levels, depending on objectives of the inventory, and on time and monetary constraints. A baseline inventory will, at a minimum, provide the locations of the installed drainage system. At the other end of the spectrum is a complete crossing environmental risk assessment that addresses all components displayed in figure 2. Four increasingly intensive levels of inventory and assessment are presented. These suggested inventory techniques are synthesized from the review of existing techniques presented in the previous section and from other studies examining the performance of road-stream crossings.

• Consequences inventory—Designed to locate and document the installed system and identify the occurrence of diversion potential over the area of inventory, this is the quickest inventory technique. Because remediation of diversion potential is often inexpensive and straightforward, it is meant to identify sites where large erosional consequences can be easily minimized.
• Connectivity/cross drain inventory—Determines the amount of road hydrologically connected to the natural channel network. It is meant to 1) identify areas of chronic erosion where sediment is being delivered directly to stream channels and 2) identify opportunities to reduce sediment impacts to aquatic ecosystems by “disconnecting” roads from streams.
• Hazard assessment—Addresses the likelihood and potential erosional volume of crossing failure based on a more extensive data collection effort.
• Environmental risk assessment—Building on the results produced from a hazard assessment, resources of concern (endpoints) are incorporated to locate sites that have the greatest probability of impacting those resources.
  A summary of these data fields is presented in table 10 for the four levels of inventory. For cross drains, a smaller data set is suggested (table 10). Note that these data sets can be tailored to suit specific needs and to satisfy time and financial constraints.

Materials Required
All that is required for a consequences inventory is a tape for measuring culvert diameter and a means of locating the site (e.g., a topographic map or aerial photo).

For more intensive hazard or environmental risk assessments, a clinometer or hand level and stadia rod is required to measure slopes of fills and culverts. A rangefinder is useful for obtaining potential diversion distances or for very large fills.

The drainage area is most easily determined by using a topographic map, however, locating the site on aerial photos as well will facilitate potential geographic information system (GIS) applications. Site location can also be accomplished with a global positioning system (GPS), and this approach is discussed on page 18.

Time Required
The time required for the various levels of inventory and analysis will vary depending on the access, frequency of drainage structures, and method of data collection (i.e., use of a global positioning system, which is discussed in the next section). Estimates presented in table 11 assume vehicular access, an automated data logger, and a GPS.

Using a Global Positioning System for Crossing Inventory
The advantages of using a GPS include:

• Accurate location of drainage structures on unmapped or poorly mapped roads
• Site information directly input to a data logger without having to transfer the information later
• GPS points and associated data can go straight into a GIS.

The disadvantages include:

• Cost of equipment
• Less than complete reliability of equipment
• Potential for operator error
• Subject to satellite availability, which can be affected by timing, canopy, and topography
• Need for specialized training of personnel in field use, data reduction, and analysis.

A crossing inventory can be conducted with or without the use of a GPS. The crossings must be accurately located so that the true drainage area can be calculated. This can be aided by using a GPS, although it is still worthwhile to have the location plotted on a topographic map and an aerial photo as backup and confirmation. If a GPS is not available, the topographic map and aerial photo should be sufficient to locate the crossing. Unfortunately, some roads are not shown on topographic maps or are only approximately located. Without a GPS, this situation is probably best addressed by plotting the location on an aerial photo and transferring that point onto an orthophoto or a digital orthoquad.

Data Analysis and Interpretation
Using a Geographic Information System in Crossing Assessment

The advantages of using a GIS for a crossing assessment include:

• The ability to combine crossing information with other coverages for analysis and display
• Ease of updating the information as upgrades, decommissions, and failures occur
• Calculation of the contributing basin area for the crossing by using scanned contour lines to draw the area on a screen
• Sharing of information, which can be enhanced by the transfer of electronic data.

The primary disadvantage is that GIS experience is required to manipulate the data.

Once the crossing locations (points), drainage areas (polygons), and associated information (attributes) are in a GIS system, they are much easier to compare spatially. The crossing points can have attributes added from existing polygon coverages such as: bedrock geology, geomorphology, hillslope gradient, slope position, soils, vegetation, and precipitation. These attributes may help to predict characteristics of uninventoried crossings. Crossings can be checked for spatial accuracy on the screen by using digital orthoquads or by comparing the crossing locations to road-stream or road crenulation (the declivity where a channel may occur as expressed by contour lines) intersections. The crossing locations and attributes such as the hazard rating can be plotted in combination with other coverages like roads, crenulation, and slope position to show where the crossings are located in relation to these features. This can be useful for spatially analyzing the information. For instance, a plot may indicate that most of the higher hazard crossings are located in lower slope positions on certain geologic types in a specific portion of a watershed. On the other hand, a higher slope position on a different geologic type may have very few crossings, with those being of low hazard.

Calculating Hydraulic Capacity
The hydraulic capacity of a culvert is the design flow it can accommodate at a specified headwater depth (the depth of water at the inlet with respect to the base of the culvert inlet). Capacity can be determined from nomographs presented in Normann et al. (1985) for a given headwater depth. For dented inlets, the diameter should be adjusted accordingly. Piehl et al. (1988a) used an equation to approximate the nomograph for inlet-controlled, circular, and corrugated metal culverts.

With this equation, inventory data incorporated into a computer spreadsheet allows rapid computing of culvert capacities. Based on hydrologic data collection for each site, plotting flows for various recurrence intervals can generate a flood frequency curve and express culvert hydraulic capacity as an exceedance probability or a recurrence interval (T) (figures 4 and 5). This method is not suitable for cross drains or small drainages where the drainage area cannot be accurately delineated. For relatively large culverts in small drainages, calculation and extrapolation can produce unreasonably large recurrence intervals (or improbably small exceedance probabilities). For convention, hydraulic capacity has a maximum value of T = 250 years (p = 0.004).

Caution must be exercised when interpreting results. Discharges will vary depending on the method used to estimate peak flows, and the error for individual design flow estimates can be large. However, if considered as a relative ranking for all the road-stream crossings in an assessment area, the results can suggest possible high-priority sites based on probability of exceedance. Because hydraulic exceedance may not be the principle mechanism of crossing failure (figure 5), hydraulic capacity assessment should be characterized as a minimum screening level for hazard assessment. If the culvert cannot pass the design peak flow, it is likely that associated debris and sediment cannot be passed either.

Increasing culvert hydraulic capacity typically requires either increasing pipe size or adding culverts or end treatments such as side tapered inlets (refer to AISI 1994 for a listing of end treatments).

Woody Debris Capacity
Plugging of culverts by organic debris is a common failure mechanism. Debris lodged at the culvert inlet reduces hydraulic capacity and promotes further plugging by organic debris and sediment. Although sediment accumulation is often deemed the cause of failure, excavation sometimes reveals blockage by one or more pieces of wood. Furthermore, plugging may not be caused by impeded transport of large debris, but by small limbs and twigs, often not much longer than the culvert diameter (figure 6), readily transported by frequently occurring storms (Flanagan in review).

Stream channel width influences the size distribution of transported woody debris (e.g., Lienkaemper and Swanson 1987, Nakamura and Swanson 1994, Braudrick et al. 1997). In low-order channels of northwest California, 99 percent of transported wood greater than 300-mm long was less than the channel width (Flanagan in review). These findings suggest that culverts sized equal to the channel width will pass a significant portion of potentially pluggable wood. However, the remaining 1 percent of the pieces remain a hazard. Thus, a wood-plugging hazard can be reduced, but not eliminated. The woody debris capacity of a crossing can be assessed by taking the ratio of the culvert diameter to the channel width (w*). Crossings with low values of w* are more prone to debris plugging. Using the northwest California coast region as an example, sizing culverts equal to the channel width will, in most cases, satisfy a 100-year design peak flow (figure 7). However, on wider channels (e.g., > 2 m), the cost of employing this strategy can be prohibitive.

The configuration of the inlet basin will also influence wood plugging. Inlet basin design should strive to maintain the preexisting channel cross section, planform, and stream gradient. Channel widening upstream of the inlet is typically undesirable (figure 8). During ponded conditions (HW/D > 1), debris in transport accumulates in the eddies formed by the widened channel. Piece rotation in the eddies promotes a perpendicular alignment to the culvert inlet. Furthermore, when the inlet is fully submerged, wood accumulates in the pond. When the inlet is reexposed, it is often presented with an enormous, often interlocking, raft of debris (figure 9).
Channel approach angle, or culvert skew, also influences debris lodgment (figure 8). Where the channel enters the culvert at an angle, debris lodgment is increased (Weaver and Hagans 1994). During runoff events, wood in transport cannot rotate parallel to the culvert and pass through it. Cross drains are susceptible to this constriction because they often possess high approach angles (Garland 1983 and Piehl et al. 1988b).

Figure 4—Determining the design storm capacity of existing culvert installations requires the use of a flood estimator and the hydraulic capacity of hte culvert adjusted for any denting or crushing of the inlet. Using an equation presented by Piehl et al. (1988), this procedure is easily automated in a spreadsheet or similar application to assess a large number of culverts. Refer also to figure 5.

Figure 5—Culvert hydraulic capacity can be expressed as a recurrence interval (T). In this example, two design discharges are calculated for a 900-mm culvert. The discharge at HW/D = 1 is assigned a recurrrence interval of 31 years (exceedence probabiligy = 0.032). The discharge necessary to overtop the road (in this case, the fill height above the inlet invert is 1.5 m) is assigned a recurrence interval of 149 years (exceedence probability = 0.0067). This example flood frequency was generated using a regional flood estimator for northwest California (Waananen and Crippen 1978).

Figure 6—Woody debris lodged at culvert inlets is often only slightly longer than the culvert diameter. Note that wood length is expressed as a ratio to culvert diameter (n = 50).

Sediment Capacity
Three types of sediment inputs are discussed here: fluvial transport, sediment “slugs,” and debris flows (a viscous, flowing mass of water, sediment, and woody debris).

In general, culverts are efficient conveyors of sediment because of their narrow and relatively smooth, uniform cross section. Fluvially transported sediments usually present little hazard to stream crossing installations. Burial of the inlet by fluvially transported sediment is often the result of woody debris lodged at the inlet. Thus, sediment deposition is often a consequence, not a cause, of the failure. Culvert design and assessment of the maximum potential particle size likely to be mobilized during peak flows have been incorporated to pass loads, ensuring sufficient diameter and slope. In order to pass sediment, culverts should be set at a grade related to the stream channel (Weaver and Hagans 1994). An index value for the sediment plugging hazard is the ratio of culvert slope to channel slope (s*). This assumes that relatively flat culverts on steep channels are more prone to sediment accumulations than steeper culverts.

Rapid, more catastrophic inputs of sediment (“sediment slugs”) are typically responsible for sediment-caused failure. Small failures from over-steepened cutslopes near the inlet may bury the entrances of cross drains and small stream crossings. Small culverts with relatively small inlet basins adjacent to steep cutslopes are most prone to this. However, predicting where such failures are likely to occur is difficult. If evidence suggests that failure of adjacent slopes is likely, the site should be further inspected. During the design phase, inlets should be located away from unstable cut slopes. While enlarging the inlet basin to capture debris is an option, it should be implemented only where no other treatment is possible because enlarged basins may alter stream hydraulics and promote woody debris and sediment deposition.

Assessment of debris flow hazard is difficult. Morphometric characteristics of debris flows and initiation areas are discussed by Costa and Jarret (1981) and Benda (1990). Debris flow hazard can be estimated from aerial photographs, digital terrain data, and evidence at the site (e.g., Chatwin et al. 1994). Debris flows interact with road-stream crossing fill prisms either by removing or impounding against the fill. Fill volumes should be reduced to minimize replacement or excavation costs. Vented fords can effectively handle debris flows by minimizing the intervention in natural stream processes from road fill. Another alternative is “hardening” the fill to minimize the chance of washout.

Figure 7—Hydraulic capacity (expressed as a recurrence interval) of 22 culverts with diameters equal to the channel width. Such a strategy to facilitate woody debris passage must be weighed with the costs of installing a large culvert on a wide channel. Costs are of less concern on smaller channels. Inlet modifications to promote passage can reduce the need for large culvert diameters. Data are from northwestern California.

Figure 8—Inlet basin plan view. Inlet basins that maintain the natural channel configuration promote debris transport and passage through the culvert. Where the flow is allowed to spread laterally, debris can accumulate and increase the chance of plugging. Furthermore, debris rotation is promoted in the turbulent eddies of the widening flow. Similarly, where the channel abruptly changes direction, wood lodgment is enhanced. This is a common configuration for cross drains.

Erosional Consequences
When crossing capacity is exceeded, water, debris, and sediment accumulate in the inlet basin. If capacity exceedance is of sufficient magnitude, overtopping of the fill will occur with associated erosional consequences. Estimating the potential erosional consequences is relatively straightforward. The path the overflowing water takes during capacity exceedance will affect the magnitude of consequences. Erosion from flows that overtop the fill and reenter the channel near the outlet is constrained by the amount of road fill material spanning the channel. Calculating fill volume is discussed below. Where the road slopes away from the crossing in at least one direction, overtopping flows can be diverted out of the channel and away from the site via the road or ditch (figure 10). Recognizing diversion potential is important; large volumes of material may be eroded as flows enlarge the inboard ditch, overwhelm adjacent crossings, enlarge receiving channels and/or flow onto hillslopes unaccustomed to concentrated overland flows. For a more complete discussion of diversion potential, refer to the companion document in this series, “Diversion Potential at Road-Stream Crossings” (Furniss et al. 1997).

Figure 9—In areas where woody debris is entrained in streamflow, crossings that pond water will pose a greater chance of plugging than crossings that do not pond water. Wood accumulates in the ponded area and plugging is likely when flows drop and expose the inlet to the mass of debris.

Figure 10—Diversion of streamflows at stream crossings can have large erosional consequences far removed from the initial failure site. Here, water diverted along the inboard ditch for several hundred meters initiated additional crossing failures along the way. Photo courtesy of the Siskiyou National Forest.

Estimating Fill Volume
Fill volume calculations are required for hazard assessments and environmental risk assessments. The procedure for calculating fill volume described is only appropriate for estimating the quantity of potentially erodible material or the amount of material to be excavated for treatment. As it is assumed that crossing failure removes the entire fill prism, and an approximation of this value is used for assessing erosional consequences, this method should not be used for contract specifications, which require more detailed surveys. This volume is designed to help with comparisons among sites. It does not include the potentially erodible aggraded stream reach upstream of crossings set above grade or the volumes associated with offsite impacts should the crossing fail. Further, the single measurement of fill width (W_f) along the road centerline will often overestimate the upstream portion of the fill (V_c) and underestimate the downstream portion (V_d). A more accurate volume can be calculated if the fill width is taken at both the inboard and outboard edge of the road. The average of these two measurements is used for calculating the volume under the road surface (V_r).

To calculate fill volume (see figure 11) measure:

W_f = Width of the fill along the road centerline and perpendicular to the culvert axis

W_c = Width of flood prone channel

L_u = Fill slope length from inboard edge of road to inlet invert

S_u = Slope of L_u (in degrees). If field data are collected in percentages, conversion to degrees is accomplished by using the arc tangent function.

The subscript “d” in the following equations refers to the above measurements taken on the downstream portion of the road fill.

(1) Upstream prism volume (V_u):
V_u= 0.25(W_f + W_c)( L_u cos S_u)( L_u sin S_u)

(2) Downstream prism volume (V_d):
V_d= 0.25(W_f + W_c)( L_d cos S_d)( L_d sin S_d)

(3) Volume under road surface (V_r):
V_r= (H_u + H_d / 2) (W_f+ W_c / 2) Wr

where:

H_u= L_u sin S_u
and
H= L sin S_u

(4) Total fill volume (V):
V= V_u+V_d+V_r

Figure 11—Crossing fill measurements—Solid lines are measured values, dashed lines are calculated. Note that Ld often extends below the culvert outlet. The method presented here is intended for estimating fill volume. Some underestimation will occur on the downstream side while the inlet portion will be overestimated.

Hydrologic Connectivity—Assessing Chronic Erosion Potential
Cross drains represent a special case for erosional consequences. In addition to being subjected to the inputs discussed above, cross drains concentrate water draining from the road surface and intercept groundwater at the cut slope. This water is often conveyed to unchanneled hillslopes where a gully may form. This newly formed channel may connect to the natural channel network, extend it, and contribute road-derived sediment and additional surface runoff to the aquatic ecosystem. During the inventory process, cross drains with gullies or sediment plumes extending to a natural channel should be documented for possible treatment.

Length of road segments connected to the natural channel network are summed, and the proportion of the road network that is hydrologically connected is determined. Individual roads with high connectivity can be targeted for “disconnecting” by the appropriate treatment (e.g., grade dips or outsloping).

Channel network extension caused by roads can also be calculated if the ditch length and road-caused gully/plume extent below the road is known. The proportional extension of the drainage network may be used to locate areas where the natural hydrologic regime is most affected.

Assessment Procedure
Assessments may proceed once the inventory and the calculations required for the level of inventory are complete. Techniques for each level of inventory are described below.

Connectivity/cross drain inventory
A connectivity inventory will result in a list of road segments that are hydrologically connected to the channel network. This level of inventory does not prioritize segments. Treatments to “disconnect” individual road segments will be based on availability of funds and transportation needs.

Consequences inventory
A consequences inventory produces a list of sites with diversion potential. Treatments to eliminate diversion potential will be based on availability of funds and transportation needs. Prioritizing sites further can be based on potential length of diversion and on receiving features.

Hazard assessment

For hazard assessment, sites are assigned a score based on several elements. For hazard data, which have been compiled into a spreadsheet or database, scores can be assigned by sorting on each of the elements or using an if-then command. The scoring system suggested below is meant to provide a flexible means of evaluating crossings. The user can adjust scores and add other factors to suit the inventory needs.

Hazard Score = Inputs + Capacity + Consequences

The inputs and capacity scoring elements are:

• T—an expression of hydraulic capacity (lowest T values are the greatest hazard)
  - T < 10 years: 3
  - T 10 to 100 years: 2
  - T > 100 years: 1
  - Site does not have definable drainage area or is an unchanneled swale: 0
• w*—an expression of woody debris capacity—culvert diameter/width of channel (lowest values are greatest hazard)
  - w* < 0.5: 3
  - w* 0.5 to 1.0: 2
  - w* > 1.0: 1
  - No definable channel: 0
• Skew angle—additional factor for debris capacity
  - Site with definable channel and skew angle > 45 degrees: 1
  - Site with no definable channel or skew angle ≤ 45 degrees: 0
• s* (slope of culvert/slope of channel)—may be used to assess the ability of the culvert to transport sediment (lowest values are highest hazard)
  - s* < 0.3: 2
  - s* 0.3 to 0.6: 1
  - s* > 0.6: 0
• R—if upslope roads are present
  - Upslope roads are present: 1
  - If no roads are present upslope: 0

For consequence scores, crossings with stream diversion potential take priority over nondiversion potential culverts because stream diversion typically results in much greater eroded volumes.

• Diversion distance (highest values are greatest hazard)
  - Distance > 300 m: 5
  - Distance 100 to 300 m: 3
  - Distance 50 to 100 m: 2
  - Distance < 50 m: 1
  - No diversion potential: 0
• Diversion onto unstable receiving feature (predetermined categories from existing data, e.g., earthflow, over-steepened sidecast)
  - Diversion onto unstable landform (may be the same as unstable geology described
  below): 3
  - Diversion onto sidecast fill: 2
  - Diversion onto valley floor: 1
• Fill volume (largest volumes are greatest hazard)
  - Volume > 1000 m³: 4
  - Volume 500 to 1000 m³: 3
  - Volume 100 to 500 m³: 2
  - Volume < 100 m³: 1
• Unstable geology (predetermined categories from existing geologic data)
  - Site is located in unstable terrain: 3
  - Site is located in moderately unstable terrain (where applicable): 2
  - Site is located in relatively stable terrain: 0
  
• Hydrologically connected
  - Site is a road-stream crossing or a cross drain that is hydrologically connected to the drainage network: 1
  - Site is a cross drain that is not connected to the drainage network: 0

Sites are prioritized based on their relative scores. However, significant features can become lost in the tallying of a hazard score. For example, a site with a high diversion potential score may have very low scores for the other elements. In this instance, the user may wish to automatically rank the site as high priority for treatment.

Environmental risk assessment

For environmental risk assessment, the score for a crossing uses predetermined endpoints in conjunction with the previously described hazard scores. Environmental risk is expressed as:

Environmental risk = Hazard^Endpoints

The endpoints value is the sum of individual endpoints values for a site.

A site is assigned a value of one if it affects an endpoint. Some endpoints may be:

• Fish (or other species of special concern) at the site—this should also generate a separate fish passage assessment.
• Fish (or other species of special concern) within the basin being inventoried.
• Sub-watershed containing the crossing has refuge value (e.g., cold water tributary).
• Domestic water supply in the basin.
• Special management area—special use areas may warrant assigning extra weight to those sites.

The number of endpoints is not fixed and should be discussed and agreed upon through interdisciplinary analysis. For meaningful results, at least one endpoint should be identified.

Table 12 presents an example of a preponderance table from an environmental risk assessment conducted on 383 road-stream crossings within a 295-km² area. Observations demonstrate that individual roads tend to have a majority of sites scoring similarly. This is favorable for upgrading or decommissioning programs where it is most practical to treat a single road segment or specific area.

Limitations
Final scores require careful interpretation. Significant features (e.g., fish at the site) can be lost if no significant hazard elements exist at the site. Although the approach is designed to make endpoints reflect the level of environmental risk, the user must assess whether the environmental risk scenario is realistic. This approach is meant to suggest sites for further inspection.

Users may wish to adjust the scoring system to reflect local settings. The scoring system presented is intended to assign higher hazard values to crossings with diversion potential. Elements not playing a large role in adding to the hazard of a site based on observations in mountainous regions (e.g., s*) add less to the overall hazard score. Again, it is up to the user to assess the magnitude of the scores and to determine whether they represent the relative magnitude of hazard elements in the area of inventory.

Evaluating consequences by road segment
An alternative to evaluating site-scale results is to generate results for individual road segments. This approach is particularly appropriate as transportation planning and decommissioning efforts use road segments as the unit of analysis. Hydrologic connectivity is well suited to this approach. A similar approach can be used for a hazard inventory as shown in table 13.

Data appropriate for assessing road segments in addition to those listed in table 13 are:

• Ditch length/km of road
• Mean ditch length/cross drain (or crossing)
• Mean hazard score
• Mean risk score.

Failed Crossing Assessments
Crossing failures provide a unique opportunity to assess design and installation procedures. After large storm events, storm damage reports are often generated to help assess the magnitude and extent of damage. Such efforts should provide the opportunity for adaptive design. Discussion of adaptive features designed to address failed culvert installations follows.

Mechanisms of Road-Stream Crossing Failure
Determining the mechanism of failure at a site is important because failure is likely to recur unless designs are implemented to reduce future hazard. Table 14 lists the mechanisms, the post-failure evidence for the mechanisms, and potential design criteria to reduce the hazards. In general, failure results from:

• Culvert capacity exceedance—Water, wood and sediment in excess of capacity can cause water to pond at the inlet. Most road-stream crossing fills function as, but are not designed as dams. The increased saturation of the fill may initiate a fill failure. When water overtops the road surface, some degree of fill erosion is likely (figure 12). However, ponded conditions or an inlet plugged with debris or sediment also promotes deposition within the inlet basin, which often requires costly excavation. Determining the initial mechanism may be difficult, when the inlet is buried or the fill has been completely removed, leaving little evidence (figure 13).
• Fill saturation—Saturation of the fill by ponding, overtopping flows, or a rusted culvert invert
  increases the hazard of fill failure.

Table 12—Example of a preponderance table from an environmental risk assessment of 382 road-stream crossing in the Upper North Fork Eel river, Northwest California. Note the tendency for particualr roads to have clusters of either high or low risk sites (e.g. 3S09 -3S10, and 5S30). One limitation of this approach, however, is that sites with diversion potential may be overlooked as is the case of several sites with 'low' rankings. This data set consists only of road-stream crossings and does not include cross drains (thus, hydro. conn).

Figure 12—Overtopping of the fill resulting in erosion of the roadway. Understanding the mechanisms that caused failure is useful for upgrading sites to avoid similar consequences in the future.

Figure 13—Burial of culvert inlets is often due to woody debris lodged at the inlet. However, determining this often requires excavation (a) or examination from the outlet (b).

Table 14—Types and physical evidence of culvert crossing failures and potential design criteria to reduce the hazard of such a mechanism occurring again.

Appendixes

Appendix A
An example of road environmental risk assessment for USDA Forest Service lands in the Upper North Fork Eel River Watershed, Six Rivers National Forest

Following are the results from inventory and assessment of road-stream crossings and cross drains on U.S. Department of Agriculture (USDA), Forest Service lands in the upper North Fork Eel River. Inventory was conducted on the 296 km² (114 mi2)portion of USDA Forest Service land. Culvert diameters at road-stream crossings (308) were typically 450 mm or 600 mm. Drainage area was defined on a 7.5 min topographic map for 207 (67 percent) of the road-stream crossings. For those sites, 63 percent are unable to pass a 100-year flood without submerging the culvert inlet, and 43 percent overtop the road fill for a 100-year peak flow.

Potential physical consequences include fill erosion and stream diversion. Median fill volume is 141 m³ per crossing with 15 percent having a greater volume than 500 m³. Forty-five percent of the road-stream crossings and 67 percent of the cross drains have diversion potential. Influencing potential physical consequences was an unstable geologic unit. Inspection of roads and crossings within this unit revealed that past failures were of much greater erosional consequence, and ongoing chronic erosion was higher than surrounding areas within relatively stable geologic settings.

The following endpoints were identified in the analysis area:

• Anadromous fish (this applied to all sites as all failures are assumed to have an impact)

• Fish at crossing (sites overlapping with the distribution of anadromous fish)

• Cold water refuges (the North Fork Eel River watershed is thermally impaired. Three subwatersheds were identified in the analysis area as important cold water tributaries and refuge areas).

An environmental risk score was assigned to each crossing. Maps on the following pages (figures A-1 and A-2) display the distribution of road-stream crossings, cross drains, road-stream crossings with high diversion potential, and high-risk road-stream crossings. In this example, sites were assigned into the high-risk category if the score was > 100. This value should be adjusted to reflect watershed conditions and the number of endpoints considered. Note the clustering of high-risk sites along road segments in the eastern portion of the basin. This clustering effect allowed for efficient treatment efforts.

Figure A-1

Figure A-2

For Additional Information Contact: 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.