Best Practices Manual: Wildlife Vehicle Collision Reduction Study

CHAPTER 4: DESIGN AND GUIDANCE OF WVC MITIGATIONS FOR LARGE MAMMALS

This chapter provides guidelines for the design and implementation of WVC mitigations aimed at large animals. These mitigations can be implemented in spot locations along a roadway or continuously along a corridor as part of the total WVC reduction strategy shown in figure 13.

Figure 13. Best practices for WVCs involving large animals, as part of the strategy for reducing WVCs.

4.1 WVC MITIGATIONS INCLUDED IN THIS MANUAL

The application and the characteristics of the 42 mitigation measures listed in table 2 relate to large mammals (deer size and larger) and specifically deer-vehicle collisions. For each mitigation measure, the table lists the estimated effectiveness in reducing deer-vehicle collisions, a measure of its cost effectiveness (referred to as "balance"), a recommendation for implementation, and an assessment of whether or not it qualifies as a "best practice. " More specifically:

• The estimated effectiveness of the mitigation measures in reducing deer-vehicle collisions is based on an extensive review of the literature. For further detail on the sources of information, refer to the Report to Congress. ¹
• "Balance" refers to the estimated benefits minus the estimated costs (in US dollars) of the mitigation measure per kilometer (0.62 mi) per year for a hypothetical road section that receives five deer-vehicle collisions per kilometer per year. The costs relate to the expenses associated with the design, implementation or construction, operation and maintenance, and removal over the design life of the mitigation measures. The benefits relate to the effectiveness of the mitigations in reducing collisions with deer and the costs associated with the average deer-vehicle collision. For further detail on the methods refer to the Report to Congress. ¹
• FHWA Project Committee for this study categorized each measure as either "best practice" or not (table 2) based on effectiveness in reducing wildlife-vehicle collisions, costs and benefits, and availability of alternatives. The best practices consist of:
  • Mitigations that may reduce deer-vehicle collisions by 80 percent or more and that have a positive "balance. "
  • Deer population culling, as it may be the only alternative under certain conditions, for example in certain suburban settings.
  • Mitigation measures that influence deer movements through habitat manipulation in the right-of-way or beyond the right-of-way were also selected as a "best practice" as these measures may be integrated with existing right-of-way management, and habitat alteration away from the road may be an alternative to population culling under certain conditions.

Note that future studies and more information may change which mitigation measures are considered "best practice" and which ones are not. In addition, despite the fact that "public information and education" was not labeled as a "best practice," and despite the fact that this measure is unlikely to substantially reduce collisions with large mammals, it is still considered "good practice," as the measure increases public awareness of the problem and builds public support for the implementation of mitigation measures.

Table 2. Effectiveness, benefit, and ranking of mitigation measures (see the Report to Congress).¹

¹ Determined by Project Committee for National WVC Reduction Study
² Experimental
? Unknown

Back to top

4.2 WILDLIFE FENCING

Wildlife fencing for large mammals (deer size and larger) is aimed at reducing WVCs by keeping these species from entering the road or road corridor including the right-of-way. However, wildlife fencing also increases the barrier effect of roads and traffic, not only for the target species but also for species that may not be a concern to human safety on roads. Increased barrier effect of roads and traffic may:

• Reduce or eliminate daily, seasonal and dispersal movements, and result in reducing or eliminating access to potentially critical resources.
• Reduce the long-term viability of certain species population in the region.

To minimize these negative effects, continuous wildlife fencing should typically be combined with safe crossing opportunities for wildlife species (see section 4.3 through 4.7). Safe crossing opportunities can reduce:

• Reduce the barrier effect of the wildlife fencing, roads and traffic.
• Reduce intrusions of large mammals into the road corridor as they no longer need to breach or climb the fence in order to get to the other side of the road corridor.

Figure 14. A 2.4-m-high (8-ft-high) large-mammal fence, with smaller mesh sizes toward the bottom, on U.S. Highway 93 on the Flathead Reservation in Montana (copyright: Marcel Huijser).

Figure 15. A moose fence, along the Glenn Highway (Hwy 1), northeast of Anchorage, AK (copyright: Marcel Huijser).

4.2.1 Effectiveness in Reducing Collisions with Large Mammals

The effectiveness of large-mammal fencing in combination with under- and overpasses has been estimated at 80-99 percent, with an average of 87 percent.

4.2.2 Technical Specifications

Wildlife fencing should typically be installed on:

• Both sides of the road corridor, rather than on just one side.
• Fence ends should be directly across the road from each other and not offset.

Fence Type

Wildlife fences for large mammals can be:

• Woven metal wire (figures 14 and 15). Higher gauge and galvanized wire is more durable and has a longer life span, about 20-25 years, than smaller gauge wire.
• Chain-link (figure 16).
• Electric. These may consist of several horizontal strands of rope-like material about 1.3 cm (0.5 in) in diameter. It delivers a mild electric shock to animals that touch it, discouraging them from passing through the fence.

Figure 16. A 3.4-m-high (11-ft-high) chain-link fence along SR 29 in southern Florida, designed to keep Florida panthers off the roadway and to guide them toward underpasses. Note the outriggers pointing away from the road (copyright: Marcel Huijser).

Fence Height

• Woven metal wire fences for large mammals, including deer, elk and moose, are typically 2.4 m (8 ft) high. Higher fences may be required for individual animals that are extremely motivated to get to the other side of the fence, and for species that are able to jump high (e. g. , bighorn sheep) (table 3).
• Higher fences are necessary if animals approach the fence from above, as on a sloped roadside.
• Electric fences may be lower, 1.2-2.1 m (4-7 ft), for example, depending on the species.

Table 3. Suggested fence height for selected large mammal species.

Mesh Size, Overhangs

• Woven metal wire fences for large mammals may have mesh sizes of 16 x 16 cm (6 x 6 in).
• Smaller mesh sizes may be advisable for species that can climb fences (black bears, for example, can climb woven metal fences by putting their feet in large mesh openings). In some cases, (barbed) wire overhangs are placed on the top of a fence to discourage animals (e. g. bears, cat species) from climbing over a fence (figure 17).

Figure 17. A 3.4-m-high (11-ft-high) chain-link fence along SR 29 in southern Florida equipped with three strands of outrigged barbed wire to prevent Florida panthers from climbing the fence (copyright: Marcel Huijser).

Fences for large mammals are sometimes combined with:

• Fences for smaller species, including amphibians, reptiles and small mammals. Fine mesh sizes (1 x 1 cm (0.5 x 0.5 in)) may be used at the bottom of the fence for small animals (figures 18, 19 and 20). This is usually a separate fence that is tied into the main large-mammal fence. Plastic sheets or concrete walls, as shown in figure 18, are more durable and are typically used for amphibians and reptiles in combination with a large-mammal fence. Note that fencing may be ineffective for tree frogs because of their climbing ability. Fences for small animals typically have an overhang (4-6 cm (2-3 in)) at a 45-90º or greater angle away from the large mammal fence; the bottom edge of the fence is usually buried 5-20 cm (2-8 in) in the ground (figure 20).
• Fences for medium-sized mammals. For medium-sized mammals, the woven metal wire fencing can have a smaller mesh size toward the bottom (e. g. 16 x 10 cm (6 x 4 in)) (figures 18 and 19).

Figure 18. Schematic drawing of a large-mammal fence in combination with barriers for smaller species (adapted and reprinted with permission from Kruidering et al. , 2005). 27

Figure 19. Large-mammal, medium-mammal and small-animal fence combined at the approach to the De Borkeld wildlife overpass across the A1 motorway in The Netherlands (copyright: Marcel Huijser).

Dig Barrier

If the species present in the area are unlikely to dig or are unable to dig underneath the fence, large-mammal fences can be flush with the ground. In combination with the tension on the fence, this should prevent large mammals from crawling underneath. Some wildlife fences are partially buried to reduce intrusions into the road corridor by species that can and do crawl underneath the fences (e. g. , at depressions in the terrain) or that can dig underneath the fence, such as a coyote (figure 20). The depth of the buried portion of the fence depends on the species-perhaps a few centimeters for amphibians to about 60 cm (2 ft) for coyotes. Instead of burying the main part of the large-mammal fence, a separate fence can be tied toward the bottom and then partially buried. For example, a 1-m-wide (3.2-ft-wide) chain-link fence with a 5 x 5 cm (2 x 2 in) mesh size can be attached to the main large-mammal fence and buried 60-70 cm (2-2.3 ft) in the ground. The "dig barrier" may be buried at a 45º angle away from the fence to further reduce potential for animals to dig under the fence.

Figure 20. A 2.4-m-high (8-ft-high) large-mammal fence with smaller mesh sizes toward the bottom and additional buried apron (dig barrier) along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Fence Posts

• Pressure-treated wood posts, at least 13 cm (5 in) in diameter for line posts and 16.5-18 cm (6.5-7 in) for braces and corner posts, are most commonly used for woven metal wire wildlife fences. They have a life span of about 20-30 years. Larger diameter posts make fences stronger and more durable. Posts should be placed about 70 cm (2.3 ft) into the ground, but local soil conditions may force deeper or shallower positioning of the posts. Posts may be placed at an interval of 4.2-5.4 m (14-18 ft). Longer life span metal posts set in concrete may be required in some situations, such as places where solid rock is on or close to the surface.
• Metal fence posts typically are more expensive than wood posts. Tension between posts can be achieved using reinforced cable on wooden posts or metal tubing on metal posts. Chain-link fences typically have metal posts.
• Fiberglass posts, typically 2.2 cm (7/8 in) in diameter, set 60 cm (2 ft) into the ground. Electric fences typically have fiberglass posts.

Fence Attachment to Poles

The woven metal wire fencing should be placed on the side of the poles facing away from the road (figure 21). If the wire fencing is attached on the road side of the poles, animals that repeatedly run into the fence mesh may loosen the fence from the posts.

Figure 21. Fence mesh attached to the side away from the road on wildlife fence along Interstate 90 near Bozeman, MT (copyright: Marcel Huijser).

Protective Top Cable

Nearby trees can fall and damage the fence, creating openings for large mammals to access the road corridor. A high-tensile cable on top of the fence posts (figure 22) can help break the fall of trees (up to a certain diameter and weight) and maintain the integrity of the fence as a barrier to animals. A protective top cable can also reduce fence repair costs.

Figure 22. Large-mammal fence with protective top cable along the Trans-Canada Highway in Banff National Park, Alberta (copyright: Marcel Huijser).

4.2.3 .Implementation Considerations

Human Safety

In some situations, there may be public safety concerns with having electrified fencing along public highways. Travelers (e. g. , pedestrians, cyclists) as well as recreationists (e. g. , fishermen, hikers) may be shocked when touching the fence.

Fence Length

Continuous wildlife fencing at least several kilometers in road length or longer is a more effective barrier for large mammals than relatively short sections of wildlife fencing several tens or hundreds of meters in length, mostly because of the home range size of large mammals (see appendix A).

Safe Crossing Opportunities for Wildlife

Continuous wildlife fencing should typically be combined with safe crossing opportunities for wildlife (see sections 4.3 through 4.7). One possible way to evaluate whether a road section that has been fenced is long enough to warrant safe crossing opportunities for selected species is to compare the diameter of the home range for the species concerned to the length of the road section that is mitigated. Examples of home range sizes for some species are provided in appendix A.

If a wildlife fence is targeted at large species such as deer, elk or moose but becomes a barrier for a non-target species, consider making the wildlife fence permeable for such species, particularly where crossing opportunities are far apart (e. g. , multiple times the diameter of the species' home range) (figure 23 and 24). An example is a wildlife fence in Florida aimed at Key deer where Lower Keys marsh rabbits could still access the road corridor through the 10 cm (4 in) gap between the bottom of the fence and the ground level (figure 23).

Figure 23. A 1.8.3 meter-high (6-ft-high) chain-link fence along U.S. Highway 1 on Big Pine Key, FL, with a 10-cm (4-in) gap allowing the endangered Lower Keys marsh rabbit access to the right of way (copyright: Marcel Huijser).

Figure 24. A chain-link moose fence near Kenai, AK, with a gap (and a barbed wire strand) at the bottom allowing small species to crawl underneath the fence (copyright: Marcel Huijser).

Existing Structures

Where fencing meets existing bridges, tunnels, drainage culverts, wildlife underpasses or overpasses, or animal detection systems, the wildlife fence should tightly connect to wing walls or sides of the structures or to the sensors of animal detection systems to prevent unintended animal intrusions into the road corridor. Even though a fence may be mainly targeted at large mammals, it should not make existing crossing opportunities such as drainage culverts inaccessible to small- and medium-sized species.

Fence Location

Wildlife fences are typically placed on the edge of the right-of-way at the property boundary. Many roads already have a right-of-way fence, and the wildlife fence should replace the right-of-way fence rather than be positioned parallel to it. In some cases the habitat in the right-of-way may be the only remaining habitat for certain species in an otherwise unsuitable environment (e. g. , large scale agricultural fields or areas with housing). In such situations, consider placing the wildlife fence closer to the road allowing animals to access at least a portion of the habitat in the right-of-way. For legal reasons, it may be necessary to keep an additional standard right-of-way fence at the property boundary in these cases.

Fence Ends

Fence ends can be associated with a concentration of wildlife-vehicle collisions. This situation can be mitigated by ending a wildlife fence at or close to:

• A safe crossing opportunity for wildlife (e. g. , wildlife underpasses, overpasses, animal detection systems, or existing bridges not specifically designed for wildlife-see sections 4.3 through 4.7).
• Steep, rugged terrain such as rock cuts (though not effective for bighorn sheep or mountain goats- see figures 25 and 26).
• Habitat that may limit movement, such as open areas for forest-dwelling species or open water for terrestrial species.
• Areas exposed to regular human activity and disturbance.

Figure 25. Fence end at top of cliff along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Figure 26. Fence end at bottom of cliff along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Potential "fence-end effects" can be further mitigated by:

• Terminating fences on straight sections of highway. This may allow for good visibility, and longer sight distances for drivers.
• Lighting at fence ends. This may improve visibility to drivers (though lighting may discourage certain species, especially carnivores, from crossing).
• Wildlife warning signs, reduced posted speed limits or advisory speeds. It is unknown whether these signs actually reduce vehicle speed or makes drivers more alert.

Additional measures may be required at the fence ends to discourage animals from entering the right-of-way. These measures might include:

• A fence end that is positioned close to the road. Concrete barriers or guardrails may be installed along the road side for human safety (figure 27).
• A boulder field between the fence and the road, and in the median, if applicable (figure 28). This may discourage ungulates from wandering off in the right-of-way. The boulder field may be 50-100 m (165-325 ft) long along roadway. The boulders must extend from the edge of the pavement up to the fence to preclude any path for wildlife to skirt the boulders. Boulder aprons are made of sub-angular, quarried rock, ranging in size from 20-60 cm (10-25 in), however most should be larger than 30 cm (12 in). The boulders are placed on geo-fabric on sub-excavated smoothed ground at a depth of about 40-50 cm (16-20 in) below the surface of the surrounding area. The boulders project about 20-30 cm (10-12 in) above the local ground surface. Concrete barriers or guardrails may be installed along the road side for human safety.
• Wildlife guards (similar to cattle guards) across the road (figure 29).
• Electric mats embedded in the road surface deliver a mild electric shock to animals that touch it.

Figure 27. Fence end brought close to the road with a concrete barrier for safety in Banff National Park, just west of Castle Junction, Alberta (copyright: Marcel Huijser).	Figure 28. The boulder field at the fence end at Dead Man's Flats along the Trans-Canada Highway east of Canmore, Alberta (copyright: Bruce Leeson).
Figure 29. Wildlife guard at a fence end on the two-lane U.S. Highway 1 on Big Pine Key, FL (copyright: Marcel Huijser) .

Escape Opportunities from the Right-of-Way

Animals that become trapped on the roadway between wildlife fences pose a safety risk to humans and to themselves. Therefore, wildlife fences should typically be combined with escape opportunities for wildlife. Escape opportunities may include:

• Swing gates, which are generally used (with or without ramps) in areas where highways are regularly patrolled by wardens or rangers. As part of their job, if animals are found inside the fenced road corridor, they can open the nearest gates and walk toward the animals, pushing them to the opened gate. A double gate, being wider, is more effective than a single gate, especially for larger species such as elk or moose. Swing gates are used to remove ungulates and large carnivores such as bears, as smaller wildlife species can often escape under the hinged doors.
• Earthen ramps or jump-outs allow medium and large mammals to safely exit right-of-ways on their own, without the aid of wardens or rangers (figures 30, 31, 32 and 33). Animals that are caught between the fences typically follow the wildlife fence until they find an escape opportunity.
• Small, hinged doors placed at the ground level for small- and medium-sized species, (figure 34) allow for escape from the right-of-way. However, these hinged doors may stop functioning properly over time and may eventually remain permanently open after use.
• Natural objects such as tree stumps, tree branches or brush can be staked against the fence until it reaches the top of the fence, allowing small- and medium-sized species to exit the right-of-way. Stacking of brush and woody debris against the fence line and to fence height will allow climbers to exit safely.

Figure 30. A short section of perpendicular fence to guide animals on top of a jump-out along a 2.4-meter-high (8-ft-high) fence along U.S. Highway 93 in Montana (copyright: Marcel Huijser).	Figure 31. A jump-out along a 2.4-meter-high (8-ft-high) fence along U.S. Highway 93 in Montana (copyright: Marcel Huijser).
Figure 32. A jump-out intended for Eurasian badger and roe deer along a 2-meter-high (6.6-ft-high) fence along the A73 motorway near Roermond, The Netherlands. Note the wildlife overpass "Waterloo" in the background (copyright: Marcel Huijser).	Figure 33. A jump-out along a 2.4-meter-high (8-ft-high) fence with smooth metal to prevent bears from climbing into the right of way along the Trans-Canada Highway, Banff National Park, Alberta (copyright: Marcel Huijser).
Figure 34. A one-way Eurasian Badger gate in the Netherlands (copyright: Marcel Huijser). .

Considerations for jump-outs:

• Jump-outs should be high enough to discourage large mammals from jumping up into the road corridor, and low enough so that animals caught between the fences will readily use them to exit the road corridor. Deer and elk are the most common users, but moose, bighorn sheep, bears and cougars have also used these structures. Little is known about the optimal height of jump-outs, but a height around 2-2.1 m (6.5-7 ft) is common.
• The landing spot around the outside wall should have loose soil or some other soft material to prevent animals from injuring themselves when jumping out.
• In areas frequented by bears, the outside walls of the jump-out should be smooth to prevent them or other animals from climbing up (figure 33).
• Escape ramps should be positioned in a set-back in the fence, in an area protected with dense vegetative cover so animals have time to calm down and look over the situation before deciding whether to jump off, continue walking along the fence or cross the highway.

Escape opportunities should be carefully designed for the target species. Both their location and regular maintenance are essential. There are currently no standards for the spacing of jump-outs, but distances of about 300 m (984 ft) between them have been suggested for mule deer.

Landscape Aesthetics

The visual impact of fences to humans, both looking from the road and looking to the road, may have to be addressed. The visual impacts of a fence can be lessened through:

• Using existing vegetation or structures as a background.
• Planting new shrubs or bushes in front of or behind a fence.
• Dark coating for chain-link fences to make them blend in with the background (figure 35).
• Positioning the fence at a relatively low point along the road, but avoid situations where animals may approach the fence going down slope as this may require a higher fence to be effective.
• Lower fence heights in situations where there are commercial or residential concerns about the visual impact of fencing, and where animals may not be abundant because of human disturbance in the area. Note that lower fence height in selected locations may compromise the effectiveness of the fence.

A wildlife fence may have to be made more visible in cases where:

• The fence makes a high-use area no longer accessible to certain species. For example, steep cliffs were no longer accessible to bighorn sheep after a wildlife fence was placed along the Trans-Canada Highway in Banff National Park, and the bighorn sheep ran into the fences trying to escape from predators. The problem was greatly reduced after the fence was made more visible with green mesh (figure 36).
• The fence is located in an area with low-flying birds.

Figure 35. A 1.8.3-meter-high (6-ft-high) chain-link fence along U.S. Highway 1 between Florida City and Key Largo, FL, has been coated with colored plastic (copyright: Marcel Huijser).

Figure 36. A 2.4-meter-high (8-ft-high) large-mammal chain-link fence along the Trans-Canada Highway between Canmore and Banff, Alberta, is equipped with green mesh to make the fence more visible to bighorn sheep (copyright: Marcel Huijser).

Access Points Requiring Fence Breaks

Where wildlife fences intersect with roads, accommodations must be made to:

• Avoid obstructing vehicles turning on or off the fenced road.
• Allow continued access for people to trails or other specific locations after a fence has been erected.
• Avoid blocking waterways. Fencing across waterways could cause problems for the water flow, floating debris, aquatic and semi-aquatic animals, and boats. These breaks in the fence require special modifications to prevent or reduce animal intrusions into the right-of-way.

Access Roads

• Wildlife guards: Transportation and land management agencies commonly install cattle guards ("Texas gates" in Canada) where fences intersect access roads (figures 37 and 38). Many different designs have been used, but few have been tested with regard to wildlife species. Cattle guard designs vary in dimensions, grate characteristics (flat or cylindrical steel grates), and grate adaptations for safe passage by pedestrians and cyclists. A grate pattern was recently developed that was 95 percent effective in blocking Key deer movement and was considered safe for pedestrians and cyclists. ²⁸ A similar design was implemented along U.S. Highway 93 on the Flathead Indian Reservation in Montana (figure 39). Small species can fall in the pit under a wildlife guard. Therefore, escape ramps are typically provided (figures 40 and 41).

Figure 37. A wildlife guard, similar to a standard cattle guard, at an on-ramp to Interstate 90 east of Bozeman, MT (copyright: Marcel Huijser).	Figure 38. A wildlife guard, similar to a standard cattle guard, at a forest access road connecting to the road "Hilversumsestraatweg", near Hilversum, The Netherlands (copyright: Marcel Huijser).
Figure 39. A modified wildlife guard (bridge-grate material) at an access road on U.S. Highway 93, south of Ravalli, MT (copyright: Marcel Huijser).	Figure 40. An escape ramp for small species from a wildlife guard pit near the town "De Lage Vuursche", The Netherlands (copyright: Marcel Huijser).
Figure 41. An escape ramp for small species from a wildlife guard pit along the road "Hilversumsestraatweg," near Hilversum, The Netherlands (copyright: Marcel Huijser). .

• Electric wildlife guards: Mats that are embedded in the pavement or that can be rolled across a low-volume road can deliver a mild electric shock when animals step on them. These mats are intended to discourage wildlife from crossing the gap in the fence. Pedestrians wearing shoes and bicyclists can cross the mats safely, but dogs, horses and people without shoes will receive an electric shock.
• Swing gates: Swing gates may be installed at low-volume access roads. Procedures must be in place to ensure that the swing gate is closed after use (figure 42).

Figure 42. A gate at an access road on U.S. Highway 93, north of Ravalli, MT (copyright: Marcel Huijser).

Trails

• Swing gates: Gates can be used to negotiate fences where they impede public access to popular recreation areas. Gates should have mechanisms to ensure that they close automatically (e. g. , spring-activated hinges or positioning the gates at an angle so they are closed by gravity). In areas of heavy snowfall, gates may be elevated and steps built to keep the bottom of the gate above snow (figures 43 and 44).

Figure 43. A swing gate for pedestrians and a wildlife guard at a bicycle path into an enclosure with large mammals (cattle), near the town "De Lage Vuursche", The Netherlands (copyright: Marcel Huijser).

Figure 44. Spring-loaded swing gate in fence allowing access for people along the Trans-Canada Highway in Banff National Park, Alberta (copyright: Adam Ford).

• Angled fence openings: An alternative (untested) design allows people to walk through an angled opening in the fence, but species such as large ungulates that have trouble bending their back sideways may not be able to pass through (figure 45).

Figure 45. Access point for people along U.S. Highway 93 south of Missoula, MT (copyright: Marcel Huijser).

Watercourses

Watercourses pose problems for keeping fences impermeable to large mammals, as their flow levels tend to fluctuate throughout the year. These problems include gaps that may appear under fencing across the waterways during low water levels, allowing wildlife to pass beneath. Also, fencing that extends close to or below the water surface can cause flooding problems when debris transported by the water is trapped against the fence, obstructing water flow. One solution is to elevate the fencing above the water and hang chains, or hinged rubber strips that float, from the bottom of the fence to the low waterline. This can allow passage of water and debris while maintaining a barrier to wildlife.

4.2.4. Example Cost Estimates

Fencing

Fencing along the western end of the Trans-Canada Highway in Banff National Park (the portion of the project known as phase 3-B, constructed in 2006-2007) was estimated at $69 (in 2007 $) per meter of fencing (Can $75 in 2006) (personal communication, Terry McGuire, Parks Canada). This fence was 2.4 m (8 ft) high, had pressure-treated wooden posts and a dig barrier. The fence had smaller mesh at the bottom (16 x 10 cm (6 x 4 in)) and bigger mesh toward the top (16 x 16 cm (6 x 6 in)). The dig barrier consisted of a buried apron (1-m-wide (3.3-ft-wide) chain link with 5 x 5 cm (2 x 2 in) mesh) that extended about 30 cm (1 ft) above ground. The rest of the apron (about 60-70 cm (2 ft)) was buried at a 45º angle away from the fence.

The cost of wildlife fencing installed along U.S. Highway 93 on the Flathead Reservation in Montana varied, depending on the road section concerned, from $27 to $42 (in 2007 $) per meter ($26 to $41 in 2006) (personal communication, Pat Basting, Montana Department of Transportation). A finer mesh fence was dug into the soil and attached to the wildlife fence for some fence sections at a cost of $12 (in 2007 $) ($12 in 2006) per meter.

Jump-outs

Reported costs for one jump-out range from $13,241 (in 2007 $) ($11,000 in 2000) to $6,425 (in 2007 $) ($6,250 in 2006) (personal communication Pat Basting, Montana Department of Transportation). ²⁹

Swing Gates for Access Roads

Costs for single- and double-panel gates along U.S. Highway 93 on the Flathead Indian Reservation in Montana were $308-370 and $360-565, respectively (in 2007 $) ($300-360 and $350-550, respectively, in 2006) (personal communication Pat Basting, Montana Department of Transportation).

Wildlife Guards for Access Roads

The reported cost of a specially designed wildlife guard was ($30,840 in 2007 $) ($30,000 in 2006) (personal communication Pat Basting, Montana Department of Transportation).

4.2.5. Maintenance

Fences are subject to damage and wear from numerous sources: vehicular accidents, falling trees, human vandalism, mowing activities in the right-of-way, animals climbing the fence or repeatedly running into the fence, soil erosion, flooding, and digging by animals. Therefore, fences must be checked regularly (e. g. , every 6 to 12 months) by walking the entire fence line identifying gaps, breaks and other deficiencies and general wear and tear including:

• Holes.
• Loose fence attachment to the poles.
• Loose embedding of posts in the ground.
• Digging underneath the fence.
• Evidence of animal passage such as wildlife trails and hair caught on the fence.
• Monitor vegetation growing adjacent to the fence and observe whether it allows animals to intrude into the road corridor. This may include trees and shrubs for species that can use them to get across the fence. Smaller species (e. g. , amphibians and small mammals) may also use grasses and herbs to climb the screens or small mesh fences, and then crawl through the larger mesh sizes higher up.

Note: avoid placing the fence where it conflicts with the management of the right-of-way vegetation and/or ditches.

Back to top

4.3. WILDLIFE UNDERPASSES

Wildlife underpasses are primarily designed to provide connectivity for wildlife species, often in combination with wildlife fencing. However, underpasses are sometimes also deployed as stand-alone mitigation measures. When used in combination with wildlife fencing, they help reduce intrusions into the road corridor as animals are provided with a safe crossing opportunity. When used as a stand-alone mitigation measure, the reduction in wildlife-vehicle collisions may be limited to the immediate vicinity of the underpass.

4.3.1. Effectiveness in Reducing Collisions with Large Mammals

The effectiveness of "crossing structures" (i. e. , underpasses and overpasses combined) in reducing large-mammal WVCs has been estimated at 79-97 percent, with an average of 86 percent, when used in combination with large-mammal fencing (see section 4.2).

4.3.2. Technical Specifications

Crossing Structure Types

Four different types of crossing structures are described in table 4, with examples in figures 46-50.Note that there are many different types of crossing structures and that their dimensions may vary greatly. Nonetheless, the types of underpasses described in this section serve a wide range of mammal species from small to large. For a more exhaustive discussion on the implementation and design specifications for a large number of wildlife underpasses, please consult Clevenger and others. (4) In addition, a limited number of underpasses that can be or have been implemented for threatened and endangered species are discussed in chapter 5.

Table 4.Dimensions of the four different types of underpasses.

Figure 46. An open-span bridge over Spring Creek, along U.S. Highway 93 south of Ravalli, MT (copyright: Marcel Huijser).	Figure 47. A large-mammal underpass (7-8 m (23-26.2 ft) wide, 4-5 m (13.1-16.4 ft) high) along U.S. Highway 93 south of Ravalli, MT (copyright: Marcel Huijser).
Figure 48. A medium-mammal box culvert (1.2 m (3.9 ft) wide, 1.8.m (5.9 ft) high) along U.S. Highway 93, south of Ravalli, MT (copyright: Marcel Huijser).	Figure 49. A medium-mammal culvert (2 m (6.6 ft) wide, 1.5 m (4.9 ft) high) along U.S. Highway 93, south of Ravalli, MT (copyright: Marcel Huijser).
Figure 50. A small- to medium-mammal pipe (badger pipe) in The Netherlands (copyright: Marcel Huijser).

Table 5 provides an overview of the suitability of the four different types of underpasses for selected species. For the purpose of comparison, wildlife overpasses (see section 4.4) are also included in this table. When evaluating the suitability of the crossing structures, the authors assumed no human co-use of the crossing opportunities. The suitability of the different types of crossing opportunities is influenced not only by the size of the species and their habitat, but also by their behavior. Because suitability depends on the species, and large landscape connectors (e. g. , tunneling or elevated road sections over long distances) are rare, providing a variety of different types of safe crossing opportunities generally provides connectivity for more species than implementing only one type, even if that structure is relatively large. Thus, providing a variety in type and dimensions of crossing structures along a corridor appears advisable.

It is important to consider designing wildlife crossing structures for a wide array of species with different home range sizes, mobility and habitat requirements, rather than focusing on a few selected target species. While the immediate concern for human safety may be with large mammals, many other species may be affected by roads and traffic and mitigation measures that increase the barrier effect of the road.

Consider making the interval between wildlife crossing structures, their location, type, and dimensions interval more "attractive" if the crossing structures are to serve:

• Individuals that disperse.
• Individuals, populations, or species that display seasonal migration.
• Individuals or species that may not be habituated to roads, traffic and associated disturbances.

Individuals, populations or species that live in the immediate vicinity of the road or that are habituated to roads and traffic and associated human disturbance may have lower requirements with regard to wildlife crossing structures.

Table 5.Suitability of different types of crossing structures for selected mammal species (mostly based on Clevenger et al. , 2008). (4) For the purpose of comparison, wildlife overpasses (see section 4.4) are also included in this table.

R = Recommended/Optimum solution
P = Possible if adapted to local conditions
N = Not recommended
? = Unknown, more data are required

Foundation

A location with stable and dry soil or rock is preferred to simplify foundation requirements and reduce resulting construction costs. Much depends on the type of crossing structure, its weight, and the local soil and hydrology.

Surrounding Terrain

Wildlife underpasses are typically constructed at locations where the roadway is relatively high compared to the surrounding terrain (figure 51). This reduces the need to raise the roadbed or to lower the approaches to the underpass, meaning less overall excavation is required.

Regarding wildlife approaches to the underpass, it is advisable to use a design that:

• Allows animals to see through to the other side and not require them to descend into a "cave" or have to climb out on the other side.
• Avoids flooding of the underpasses and the associated soil erosion.

These suggestions notwithstanding, the structures should be located where animals are likely to use them; not all underpasses should be located in road fills.

Figure 51. Wildlife underpasses are often installed in low-lying areas with road fill because construction costs are lower (copyright: Marcel Huijser).

Fence

At wildlife underpasses, the wildlife fencing alongside the road corridor typically ties into the wing walls of the underpass or the structure itself (figure 52).

Figure 52. Wildlife fencing ties in with the underpass wing walls, preventing animals from entering the road corridor (copyright: Marcel Huijser).

Figure 53 shows how gaps between the fence and the crossing structure can allow animals to enter the roadway.

Figure 53. This fence post leaves more space than desired between the post and the wing walls of a crossing structure, allowing some medium- or large-sized mammals to enter the road corridor (copyright: Marcel Huijser).

With small underpasses such as small- to medium-sized culverts, the fencing can continue on top of the structure (figure 54). Naturally, the approaches to a crossing structure are not fenced.

Figure 54. The wildlife fencing continues above a box culvert for medium-sized mammals (copyright: Marcel Huijser).

Soil Depth and Substrate

Because of limited light and moisture, underpasses typically have little or no vegetation. However, natural soil is preferred over unnatural substrate such as corrugated metal or concrete for covering the bottom or floor of the underpass. In such situations, soil, preferably from the immediate surroundings, may be used in a layer greater than 15 cm (6 in) (figure 55). If water would wash the substrate away at certain times of year, consider a concrete layer on top of the corrugated metal of a culvert. Open bottom underpasses typically already have the natural substrate of the surrounding area.

Figure 55. Soil in corrugated metal culvert makes it more suitable for use by wildlife (copyright: Marcel Huijser).

Habitat Inside Underpasses

• The habitat of the approaches and inside the underpasses should reflect that of the surroundings and the habitat requirements of the target species as much as possible. Often, this requires the presence of multiple habitat types, which can also influence the width of the crossing structure as it takes space to create multiple habitat types. The different habitat types should span the entire structure, should continue in the approaches to the underpass, and should be integrated with the adjacent habitat.
• Preferably, use only native species for plantings near underpasses, and if possible use only seeds or plants from the immediate surroundings.
• Wildlife underpasses typically have insufficient light and moisture to allow for plant growth inside the structure. However, if the underpass goes under a divided highway, the underpass may have an "open roof" at the median that would allow for vegetation growth. Wildlife fencing must be installed in the median to prevent wildlife from entering the road corridor.
• Tree stumps, branches, rocks and other natural material is often positioned inside wildlife underpasses to provide cover to small- and medium-sized animals. If wet habitat or streams are present (see also section 4.5), underpasses can also provide habitat for (semi-) aquatic species including amphibians.

4.3.3. Implementation Considerations

Adjacent Land Use

• Adjacent lands, beyond the right-of-way, should be secured where possible. Land acquisition or zoning may need to be considered to protect the land that serves as an approach to the crossing structures. Note that the life span of underpasses is typically projected at about 75 years. Disturbance in the Surroundings
• Human activity and human-related disturbance should be avoided as much as possible in the vicinity of wildlife crossing structures, including the approaches.
• Depending on local conditions, also consider relocating streetlights or other light sources that may be in the immediate vicinity of a crossing structure.
• Locations with frontage roads should be avoided, or additional mitigation measures for the frontage road may need to be considered.
• Discourage livestock from accessing the wildlife underpasses for shade. The presence and smell of livestock (including their feces and urine) may discourage wildlife from using the underpass. Figure 56 shows the use of livestock fencing (four strands-the top and bottom strands are smooth rather than barbed wire) that keeps livestock from accessing the underpass.
• Consider designing underpasses to minimize noise disturbance from the road (e. g. , employing sound-attenuating walls above the entrances).

Figure 56. A wildlife underpass with livestock fencing along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Avoid Habitat Destruction During Construction

• Habitat destruction during construction activities should be avoided or minimized, especially in the areas around the crossing structures themselves. For open-span bridges, this is particularly important with regard to the area under the bridge.
• Consider erosion control and re-vegetation after construction to restore the natural conditions, and to provide cover at the approaches to the crossing structures (figures 57 and 58).

Figure 57. After construction, straw mats were used to protect slopes from erosion and rapid establishment of weeds along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Figure 58. Native shrubs and trees were planted at the approaches to crossing structures along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Baiting or Cutting Trails

• If appropriate, consider baiting crossing structures with attractants such as salt or hay immediately after completion of the mitigation measures. If crossing structures are combined with wildlife fencing, the usual routes travelled by animals may no longer be available. This may justify luring the animals to the crossing structures and facilitating the approach (e. g. , cutting trails that lead to the crossing structures) to help them learn the location of the crossing structures and that it is safe to use them.
• Consider reducing or stopping such efforts after wildlife activity at the underpasses has reached a desired level.

4.3.4. Example Cost Estimates

Open-span Bridges

The costs for an open-span bridge along the two-lane U.S. Highway 93 on the Flathead Reservation in Montana (across the Jocko River seasonal side canal) were estimated at $435,340 (in 2007 $) ($423,483 in 2006) (personal communication, Pat Basting, Montana Department of Transportation). The bridge measured 30 m (98.4 ft) in width (road width) and 12 m (39.4 ft) in length (road length). Because of slopes, the effective width of the underpass was less than 12 m (39.4 ft).

Underpasses under open-span bridges across the four-lane Trans-Canada Highway in Banff National Park (Phase 3-A) measure about 12 m (39.4 ft) in width and about 5 m (16.4 ft) in height. Costs were estimated to be between $675,597 and $965,139 (in 2007 $) Can$700,000-1 million (in about 1996) (personal communication, Terry McGuire, Parks Canada). In 2007, based on the construction on the Trans-Canada Highway in the Lake Louise area, the costs for a 16 to 25 m-wide (52.5 to 82.0 ft wide) underpass structure across a two-lane road, including traffic control and detour, was estimated at $2,350,000 (in 2007 $) (Can$2.5 million in 2007) (personal communication, Terry McGuire, Parks Canada)

Large-Mammal Underpasses

In this guide, large-mammal underpasses are defined as structures that are not bridges but, for instance, box culverts or arched culverts that are at least 7-8 m (23.0-26.2 ft) wide and 4-5 m (13.1-16.4 ft) high. Large-mammal underpasses along the four-lane Trans-Canada Highway in Banff National Park (Phase 3-A) measure about 7 m (23.0 ft) in width and 4 m (13.1 ft) in height. Costs were estimated between $217,156 and $241,285 (in 2007 $) (Can$225,000 to Can$250,000 in about 1996) (personal communication, Terry McGuire, Parks Canada).

Three large-mammal wildlife underpasses, all arched culverts, along the two-lane U.S. Highway 93 on the Flathead Reservation in Montana (south of Ravalli) measure about 7-8 m (23.0-26.2 ft) in width and about 5 m (16.4 ft) in height. The length (road width) varies between 18.3 and 21.9 m (60.0-71.8.ft). The costs were estimated at about $223,076 (in 2007 $) ($217,000 in 2006) (personal communication, Pat Basting, Montana Department of Transportation).

In The Netherlands, large-mammal underpasses 7-10 m (723.0-32.8.ft) wide and about 4 m (13.1 ft) high were estimated at $38,192 to $63,654 (in 2007 $) per meter of road width (€30,000-€50,000 in 2005). (27) Assuming a road width of 20 m (65.6 ft), the costs were $763,845-$1,273,075 (in 2007 $) (€600,000-€1,000,000 in 2005).

Medium-Mammal Underpasses

Medium-mammal underpasses are defined herein as box culverts or culverts that are between 0.8.and 3 m (2.6-9.8.ft) wide, and 0.5 and 2.5 m (1.6-8.2 ft) high.

Medium-mammal box culverts under the four-lane Trans-Canada Highway in Banff National Park (Phase 3-A) measure about 3 m (9.8.ft) in width and 2.5 m (8.2 ft) in height. Costs were estimated at $173,725 (in 2007 $) (Can$180,000 in about 1996) (personal communication, Terry McGuire, Parks Canada). In 2007, based on construction on the Trans-Canada Highway in the Lake Louise area, the costs for a box or elliptical culvert 3-4 m (9.8.-13.1 ft) wide and high across a two-lane road were estimated at approximately $940,000 (in 2007 $) (Can$1 million), including traffic control and detour (personal communication, Terry McGuire, Parks Canada).

Two medium-mammal box culverts 1.2-1.8.m (3.9-5.9 ft) wide, 1.2-1.8.m (3.9-5.9 ft) high and 27.5 m (30.1 yd) long, and one medium-mammal culvert about 2 m (6.6 ft) wide, 1.5 m (4.9 ft) high and 27.5 m (90.2 ft) long along the two-lane U.S. Highway 93 on the Flathead Reservation in Montana (south of Ravalli) were estimated at about $70,932 each (in 2007 $) ($69,000 each in 2006) (personal communication, Pat Basting, Montana Department of Transportation).

In The Netherlands, medium-mammal box culverts 0.8.-1.3 m (2.6-4.3 ft) wide, 0.5-0.75 m (1.6-2.5 ft) high were estimated at $1,528-$3,183 (in 2007 $) per m (road width) (€1,200-2,500 in 2005). (27) Assuming a road width of 20 m (65.6 ft), the costs were $30,554-$63,654 (in 2007 $) (€24,000-50,000 in 2005).

Small- and Medium-Mammal Pipes

Small- and medium-mammal pipes are defined for the purposes of this guide as those that measure about 0.3-0.6 m (1.0-2.0 ft) in diameter. In The Netherlands, small- and medium-mammal pipes, or "badger pipes," 0.6 m (2.0 ft) in diameter, were estimated at $891-$1,528 (in 2007 $) per m (€700-€1,200 in 2005). (27) Assuming a road width of 20 m (65.6 ft), the costs were $17,823-$30,554 (in 2007 $) (€14,000-€24,000 in 2005).

4.3.5. Maintenance

• It is important to ensure that diversity in vegetation and habitat is maintained in later years as the vegetation matures and succession may make an approach to a crossing structure less open than desirable.
• Human use of the area around the crossing structures and of the crossing structures themselves should be monitored. If necessary, consider taking measures to discourage human use.
• If wildlife crossing structures are not being monitored on a regular basis, specific periodic visits should be made to ensure that there are no obstacles or foreign matter in or near the crossing structures that might affect wildlife use.

Back to top

4.4. WILDLIFE OVERPASSES

Wildlife overpasses are primarily designed to provide connectivity for wildlife species, often in combination with wildlife fencing. However, overpasses are sometimes also deployed as stand-alone mitigation measures, with no or limited fencing. When used in combination with wildlife fencing, they help reduce intrusions into the road corridor as animals are provided with a safe crossing opportunity. When used as a stand-alone mitigation measure, the reduction in wildlife-vehicle collisions may be limited to the immediate vicinity of the overpass.

4.4.1. Effectiveness in Reducing Collisions with Large Mammals

The effectiveness of "crossing structures" (i. e. , underpasses and overpasses combined) in reducing large-mammal WVCs has been estimated at 79-97 percent, with an average of 86 percent, when used in combination with large-mammal fencing (see section 4.2).

4.4.2. Technical Specifications

Crossing Structure Types

For the purpose of this handbook one type of wildlife overpass is discussed-those that are 50-70 m (164.0-229.7 ft) wide (figure 59).combined with the wildlife underpasses described in section 4.3, this handbook describes five types of crossing structures that serve a wide range of mammal species from small to large. For a more exhaustive discussion on the implementation and design specifications for a large number of wildlife crossing structure types, please consult Clevenger and others. (4) In addition, a limited number of crossing structures that can be or have been implemented for threatened and endangered species are discussed in chapter 5.See Table 5 in section 4.3 (Wildlife Underpasses) for the suitability of wildlife overpasses for different species and species groups.

Figure 59. Red Earth Overpass on the Trans-Canada Highway in Banff National Park, Alberta (copyright: A. P. Clevenger).

Foundation

A location with stable and dry soil or rock is preferred to simplify foundation requirements and reduce resulting construction costs. Much depends on the type of crossing structure, its weight, and the local soil and hydrology.

Surrounding Terrain

Wildlife overpasses are typically constructed at a location where the terrain on either side of the road is higher than the road (figure 60). This situation allows for:

• A more gradual approach (ideally, a 5: 1 slope or less) to the overpass and allows the animals to see across to the other side when deciding whether to cross.
• Reduced amount of fill material needed for the approaches.

These suggestions notwithstanding, the structures should be located where animals are likely to use them; not all overpasses should be located in road cuts and not all underpasses should be located in road fills.

Figure 60. A wildlife overpass in Germany. The terrestrial habitat connected by the overpass is at a higher elevation than the roadway. (Source: Mary Gray, FHWA)

Fence, Screens and Berms

A wildlife overpass typically has wildlife fencing on both sides to prevent animals from jumping off the overpass into the traffic below. This fence is seamlessly connected to the wildlife fencing along the road corridor (figure 61).

Figure 61. This fence for amphibians and medium- and large-sized mammals connects well to the fence on the wildlife overpass "Waterloo" near the town "Roermond," The Netherlands. Note that the fence on the wildlife overpass also acts as a light and sound barrier (copyright: Marcel Huijser).

Wildlife overpasses usually have berms or screens greater than 2.5 m (8 ft) attached to the side to reduce the exposure of animals on the overpass to the sound and light from the traffic below (figures 62, 63 and 64).

• If a sound and light barrier is high and strong enough it can replace a wildlife fence on top of wildlife overpasses.
• On relatively narrow overpasses care must be taken that these barriers do not create a tunnel effect. To reduce the tunnel effect, the sound and light barrier can be directed slightly outwards.
• Note that lightweight materials other than soil may be used for the core of berms to reduce the weight and thus the required load-bearing capacity of the structure.
• Screens should be placed on the outer edges of a structure as they would otherwise reduce the effective width of the overpass for wildlife.

Figure 62. The wildlife overpass "Waterloo" with wildlife fencing across the A73 motorway near the town "Roermond," The Netherlands, consists of planks that also act as a light and sound barrier (copyright: Marcel Huijser).	Figure 63. The wildlife fencing on the wildlife overpass "Waterloo" across the A73 motorway near the town "Roermond," The Netherlands, consists of planks that also act as a light and sound barrier (copyright: Marcel Huijser).
Figure 64. The wildlife fencing on the wildlife overpass "Waterloo" across the A73 motorway near the town "Roermond," The Netherlands. Note that the concrete barrier prevents small animal species from falling off the overpass (copyright: Marcel Huijser).

Soil Depth and Substrate

Depending on the target species, different types of habitat may need to be created on top of a wildlife overpass or inside a wildlife underpass. The overpass should be designed accordingly and have sufficient soil depth, sufficient soil fertility, and suitable hydrology if, for example, trees or shrubs are to grow on top of the overpass. The following soil depth is recommended for overpasses in a temperate climate with annual precipitation of approximately 800 mm (31 in): (27)

• Grasses and herbs: greater than 0.3 m (1 ft).
• Shrubs: greater than 0.6 m (2 ft).
• Trees: greater than 1.5 m (5 ft).

Soil depth can be varied to promote variation in vegetation and physical conditions on overpasses. Use soil from the immediate surroundings as much as possible. On some overpasses concrete beams have been attached across the width of the overpass to keep moisture from running off too quickly. In addition, impermeable soil can be used to create a wetter zone that holds water longer than the adjacent soil to create attractive habitat for (semi-)aquatic species, including amphibians.

Habitat on Overpasses

The habitat on top of overpasses should reflect that of the surroundings and the habitat requirements of the target species. Often, this requires the presence of multiple habitat types, which can also influence the width of the crossing structure, as it takes space to create multiple habitat types. The habitat on top of wildlife overpasses may include:

• Open habitat (grasses, herbs).
• Cover (shrubs, trees, tree stumps, logs, branches, rocks) (figures 65, 66, and 67).
• Ditches or depressions, and berms (on the sides).
• Wet areas or (artificial) streams.

Figure 65. Small trees, shrubs and grass vegetation on top of one of the overpasses along the Trans-Canada Highway in Banff National Park, Alberta (copyright: Marcel Huijser).	Figure 66. A row of tree stumps leading animals to and across the wildlife overpass "Waterloo" across the A73 motorway near the town "Roermond", The Netherlands (copyright: Marcel Huijser).
Figure 67. Newly planted shrubs (foreground) and a row of tree stumps on the De Borkeld wildlife overpass across the A1 motorway in The Netherlands (copyright: Marcel Huijser).

• The different habitat types should span the entire structure, should continue in the approaches to the crossing structure, and should be integrated with the adjacent habitat.
• Consider planting higher shrubs and trees on the north or east side of an overpass to avoid shading out the overpass entirely.
• Tree species that grow tall and that have large and deep root systems should be avoided on an overpass because of concerns for the integrity of the structure and the potential for the trees to fall on the traffic below. Limit tree height to about 2.5-4 m (8-12 ft).
• Many overpasses have artificially created ponds and attractive vegetation (e. g. , berry producing shrubs) on at least one side of wildlife overpasses to encourage animals to visit the location and use the crossing structure (figure 68).
• Preferably, use only native plant species, and if possible use only seeds or plants from the immediate surroundings.

Figure 68. An artificial pond on the approach of the wildlife overpass "Waterloo" across the motorway A73, near the town "Roermond," The Netherlands (copyright: Marcel Huijser).

4.4.3. Implementation Considerations

Design Overpass

Different constructions and designs may be considered when designing wildlife overpasses, including steel or concrete arches or bridge spans. A wildlife overpass may be rectangular or hourglass shaped (figure 69), each of which come with their own construction and design considerations and costs. Note that the dimensions in table 6 relate to the minimum width of crossing structures-i. e. , the middle of the hourglass.

Figure 69. A rectangular and an hourglass-shaped wildlife overpass design (reprinted with permission from Kruidering et al. , 2005). (27)

Adjacent Land Use

Adjacent lands, beyond the right-of-way, should be secured for at least the life span of the crossing structures (about 75 years). Land acquisition or zoning may need to be considered to protect the land that serves as an approach to the crossing structures.

Disturbance in the Surroundings

• Human activity and human-related disturbance should be avoided as much as possible in the vicinity of wildlife crossing structures, including the approaches.
• Depending on local conditions, also consider relocating streetlights or other light sources that may be in the immediate vicinity of a crossing structure.
• Locations with frontage roads should be avoided, or additional mitigation measures for the frontage road may need to be considered.
• Livestock should be discouraged or prevented from accessing the wildlife overpasses through the use of livestock fencing or other means. The presence and smell of livestock (including their feces and urine) may discourage wildlife from using the overpass.

Habitat Destruction during Construction

• Habitat destruction during construction activities should be avoided or minimized, especially in the areas around the crossing structures themselves.
• Consider re-vegetation after construction to restore the natural conditions, and to provide cover at the approaches of the crossing structures.

Baiting or Cutting Trails

• If appropriate, consider baiting crossing structures with attractants such as salt or hay immediately after completion of the mitigation measures. If crossing structures are combined with wildlife fencing, the usual routes travelled by animals may no longer be available. This may justify luring the animals to the crossing structures and facilitating the approach (e. g. , cutting trails that lead to the crossing structures) to help them learn the location of the crossing structures and to understand that it is safe to use them.
• Consider reducing or stopping such efforts after wildlife activity at the overpasses has reached a desired level.

4.4.4. Example Cost Estimates

Wildlife overpasses across the four-lane Trans-Canada Highway in Banff National Park (Phase 3-A) were estimated to cost $1,688,993 each (in 2007 $) (Can$1.75 million each in about 1996) (personal communication, Terry McGuire, Parks Canada). The overpasses were 52 m wide and 70 m long (170.6 ft x 229.7 ft), crossing four lanes of traffic. In 2007, based on construction on the Trans-Canada Highway in the Lake Louise area, the costs for a 60-m-wide (196.9 ft) overpass across a two-lane road was estimated at $3,290,000 to $3,760,000 (in 2007 $) (Can$3.5 million to Can$4 million in 2007), including traffic control and detour (personal communication, Terry McGuire, Parks Canada).

A proposed overpass across Montana Highway 83 near Salmon Lake (a two-lane road) was estimated to cost $1,542,000 to $2,467,200 (in 2007 $) ($1.5 million to $2.4 million in 2006) (personal communication, Pat Basting, Montana Department of Transportation). The costs for six wildlife overpasses (30-50 m wide (98.4-164.0 ft)) across four-lane roads in The Netherlands ranged between $4,684,045 and $19,408,640 (in 2007 $) (€3,500,000 and €14,750,000 in about 2005) (table 6).

Table 6.Characteristics of wildlife overpasses in The Netherlands (Partially based on Kruidering et al. (2005) and personal communication, Hans Bekker, Ministerie van Verkeer en Waterstaat, the Netherlands). (27)

* cost in year of completion
Note that some of the wildlife overpasses are actually multi-functional overpasses (see section 4.6)

4.4.5. Maintenance

• Vegetation maintenance on top of wildlife overpasses (e. g. , supply water and fertilizer, weed control) may be required in the first few years after planting. The same applies to plantings in the approaches to the crossing structures.
• It is important to ensure that diversity in vegetation and habitat is maintained in later years, as the maturing vegetation and succession may make an overpass or an approach to an overpass less open than desirable.
• Human use of the area around the crossing structures and of the crossing structures themselves should be monitored. If necessary, consider taking measures to discourage human use.
• If wildlife crossing structures are not being monitored on a regular basis, specific periodic visits should be made to ensure that there are no obstacles or foreign matter on or near the crossing structures that might affect wildlife use.

Back to top

4.5. MULTIPLE USE UNDERPASSES

Multiple use underpasses are not only designed for use by wildlife, but also for water flow, roads or railroads that cross under the main road (including farm roads), and bike or pedestrian paths. Because of this multiple use concept, compromises are inevitable. With regard to connectivity for wildlife, multiple use crossing structures may be exposed to more disturbance than structures that are strictly designed for wildlife. On the other hand, species associated with specific habitat-e. g. , streams, rivers, riparian habitat, or vegetation associated with streams-may benefit more from a wildlife crossing structure when it is combined with a stream or river crossing. Similarly, if a crossing structure is constructed because of human transportation needs (e. g. , a road or railroad crossing, or a bike or pedestrian path), wildlife may benefit if the crossing structure is not solely designed for human transportation but is modified to encourage co-use by wildlife. A wildlife crossing structure that is adapted for human transportation needs, may affect use by certain wildlife species. Note that many of the considerations for "wildlife only" underpasses (section 4.3) also apply to multiple use underpasses.

4.5.1. Effectiveness in Reducing Collisions with Large Mammals

The effectiveness of underpasses in reducing large-mammal WVCs has been estimated at 79-97 percent, with an average of 86 percent, when used in combination with large-mammal fencing (see section 4.2).

4.5.2. Technical Specifications

Stream Sizes and Seasonal Dynamics in Water Flow

• Stream characteristics and stream dynamics must be carefully studied to ensure that the conditions inside the crossing structure are and remain similar to that of the stream up- and downstream of the structure. Parameters of importance can include:
  • Water velocity.
  • Water depth.
  • Turbulence.
  • Variability in water velocity.
  • Sediment.
  • Debris blockage.
  • Erosion of substrate inside the crossing structure or upstream and downstream of the structure (figure 70).
  • Implications of high- and low-water events, including debris and maintenance issues.
• In general, wide crossing structures with open bottoms are preferable to narrow structures with bottoms that are part of the structure (unless the structure bottom is well embedded).

Figure 70. Structures built to pass water typically should be designed to avoid erosion and allow passage of aquatic species. The structure shown has no adaptations for terrestrial species and is also likely to be impassable for many aquatic species (copyright: Matt Blank).

Width and Clearance of Walkways

Streams and rivers vary greatly in size and the relative space for the stream or river on one hand, and the terrestrial habitat on the other can also vary greatly. Figure 71 shows a box culvert for a stream that has been modified to allow for the passage of small- and medium-sized mammals along walkways inside the culvert. The walkways can be made out of different materials (e. g. , wood, steel, or concrete), and may be integrated in the original construction or added as a retrofit. It is important that the walkways are accessible to the terrestrial animals and that slightly lower or higher water levels do not immediately make the walkways inaccessible.

Dimensions of Walkways:

• For small- and medium-sized species (up to marten and rabbit size), a minimum walkway width of 0.5-0.7 m (1.6-2.3 ft) is recommended. Preferred walkway width is about 1 m (3.3 ft). Minimum clearance between the walkway and the ceiling of an underpass is about 0.6 m (2 ft) or greater.
• For large mammals the minimum width of the walkway is about 2 m (6.5 ft). Preferred walkway width is greater than 3 m (10 ft). For large mammals, minimum clearance between the walkway and the ceiling of an underpass is about 3 m (10 ft). Preferred clearance is about 4 m (13 ft) or greater.

Walkways in Culverts Modified for Wildlife Use

The top side of the walkways should consist of material that has a relatively rough surface that animals will not slip on. In some situations, where erosion danger is low, soil may be placed on the walkways. Some walkway designs have a hollow beam or tube integrated into the walkway design or attached to the walk way to encourage small mammals (mice, voles) to use the walkway.

Figure 71. Different walkway designs for small- and medium-sized mammals (e. g. , up to marten and rabbit size) in box culverts (retrofitted or integrated design). A = minimum 0.5-0.7 m (1.6-2.3 ft), preferred 1m (3.3 ft), B= 0.6 m (2 ft) (reprinted with permission from Kruidering et al. , 2005). (27)

It is important that walkways are well connected to the banks and that small- and medium-sized mammals can access the walkway regardless of stream level (figure 72). The angle of the walkway where it connects to the bank should be 30-45º at a maximum.

Figure 72. Connection of walkway to adjacent bank (reprinted with permission from Kruidering et al. , 2005). (27)

Wildlife Passage in Stream Underpasses with Variable Water Flow

When underpasses are designed specifically for streams, particularly streams with variable flow rates, they can often easily incorporate wildlife passage. Figure 73 shows the bed of a seasonal creek that is relatively small. Most of the space in the underpass is accessible for terrestrial species.

Figure 73. Wildlife underpass (large-mammal culvert) along U.S. Highway 93 in Montana with a ditch that allows water to flow (seasonally) (copyright: Marcel Huijser).

Figure 74 shows a creek of a similar size, but this underpass also provides a connection between a larger river and an old arm of the river that will flow when water levels in the main channel rise above a certain level. Note that the straw bundles along the creek were placed to reduce erosion just after construction. Also note the slight berm to the left side of the creek, to allow it to rise some before it overflows, keeping the underpass dry and accessible to terrestrial species.

Figure 74. Wildlife underpass (over-span bridge) in combination with a creek crossing along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Figure 75 shows the same underpass as in figure 74, except during high water flow. Although the underpass remains mostly unusable for terrestrial animals during high flow, these slight design modifications can make it usable for most of the rest of the year.

Figure 75. The same wildlife underpass as in figure 74 during spring runoff (copyright: Tiffany Holland).

Figures 76 and 77 show larger creeks that flow year-round. The banks have been stabilized with larger rocks, and walkways have been provided to allow terrestrial species to cross under the over-span bridge.

Figure 76. Over-span bridge across a stream, with a bank and walkway for terrestrial large-mammal species along the Trans-Canada Highway, Banff National Park, Alberta (copyright: Marcel Huijser).

Figure 77. Concrete walkways for wildlife on both banks of a creek at an over-span bridge along U.S. Highway 93, south of Missoula, MT (copyright: Marcel Huijser).

The walkways in figure 77 are made of concrete, which some animals may avoid. Finding a balance between material that will not wash out during high flows but is desirable for animals can be a challenge. One solution is to use concrete covered with several centimeters of topsoil. This topsoil may wash out every few years, but it can be easily replaced. Note that the creek is starting to undercut the concrete path on the left side of figure 77.

Figure 78 shows a bridge across a river where the banks and slope are protected with large rocks. However, a path has been cleared to allow people and large mammal species to pass under the bridge.

Figure 78. Walkway cleared of large rocks, for people and large mammals, along State Highway 75 between Ketchum and Hailey, ID (copyright: Marcel Huijser).

Figure 79 shows a bridge across a river that also spans several tens of meters of bank on either side of the river. This approach allows for some changes in the location of the river bed, it allows for passage of large terrestrial mammals, and it allows for flooding events.

Figure 79. Bridge across the Jocko River, along U.S. Highway 93 in Montana (copyright: Marcel Huijser).

Roads, Farm Roads and Recreational Bike and Pedestrian Paths

Underpasses constructed to provide access for farmers and livestock may also be used by wildlife (figure 80). To encourage wildlife use, access to the underpass by livestock may be limited to occasional movement between pastures. Gates that can be closed can be installed to prevent livestock from accessing the underpass, except when changing pastures. The gate is not a substantial barrier for most wildlife species, but the gates should be left open if there are no livestock in the area.

Figure 80. An underpass designed for wildlife and human access, farm machinery and livestock along U.S. Highway 93, south of Missoula, MT (copyright: Marcel Huijser).

• Underpasses for roads designed to encourage co-use by wildlife may only be feasible if traffic intensity is relatively low.
• Cover such as tree stumps or rocks may be provided on one or both sides of an underpass (figure 81 and 82).
• If a median and an "open roof" is present in the underpass, trees, shrubs and other vegetation may grow halfway through the underpass (figures 81 and 82).
• In some cases, light from cars may be shielded to minimize disturbance (figures 83 and 84).
• For roads that receive little to no use, consideration may be given to change a paved road into a gravel or dirt road, or remove the road altogether.

Figure 81. An underpass for a road (Lage Vuurscheweg) under the motorway A27 near the town "Hilversum," The Netherlands, that was modified with soil and tree stumps to encourage co-use by wildlife (target species included invertebrates, amphibians, reptiles, and small- and medium-sized mammals). Note that light and moisture allow for shrubs in the gap in the median (foreground) and that tree stumps provide cover under the bridge for northbound traffic (background) (copyright: Marcel Huijser).

Figure 82. An underpass for a road (Hilversumsestraatweg) under the motorway A27 near the town "Hilversum," The Netherlands, that was modified with soil and tree stumps to encourage co-use by wildlife. Note that light and moisture allow for shrubs and trees in the gap in the median (background) and that tree stumps provide cover under the bridge for northbound traffic (foreground) (copyright: Marcel Huijser).

Figure 83. An underpass for a road (Hilversumsestraatweg) under the motorway A27 near the town "Hilversum," The Netherlands, that was modified with soil and tree stumps to encourage co-use by wildlife to the right of the fence. The purpose of the fence, which is painted black, is to reduce light and noise disturbance from traffic (copyright: Marcel Huijser).

Figure 84. Close-up of the fence, painted black, aimed at reducing light and sound disturbance from traffic (copyright: Marcel Huijser).

4.5.3. Implementation Considerations

Stream and River Crossings

Adapting stream and river crossings in such a way that they become more passable by terrestrial animal species requires a structure larger than the stream itself, as the banks of the stream or river have to be included to provide a dry passage. Additional benefits of large structures consist of:

• More space for potential ecosystem processes (e. g. , room for the channel dynamics in the river or stream).
• Reduced maintenance costs related to removal of debris in undersized structures that may hinder the water flow.

Note: If wildlife crossing opportunities are only provided at stream or river crossings, species of drier habitat away from rivers and streams may not benefit sufficiently from these crossing opportunities.

Farm Roads and Recreational Bike and Pedestrian Paths

• Suitable habitat must be available on both sides of the structure.
• Paths or riding trails intended for human use should be confined to one side of a crossing structure rather than in the middle, leaving greater space for wildlife use. Vegetation or other cover such as tree stumps, rocks or screens can be used to reduce the impact of human use on wildlife. While co-use opportunities (by people and wildlife) are inherently attractive, several items require careful evaluation before moving ahead with them.
• When the crossing structure is in an important or sensitive ecological area or if the target species for the crossing structure are sensitive to human disturbance, wildlife and human use should probably not be combined in the same structure. Two separate structures should be considered in this case.
• When the crossing structure is located in a multifunctional landscape with considerable human disturbance, and if the target species for the crossing structure are not very sensitive to human disturbance, and perhaps even thrive with a certain level of human disturbance (for example, raccoons thrive in an agricultural landscape), wildlife and human use can probably be combined in the same structure.

4.5.4. Example Cost Estimates

If the main function of a crossing structure is to pass water or traffic, the extra costs for wildlife only apply to the extended length of a bridge or the increased size of an underpass. See section 4.3 for costs for structures that are specifically designed for wildlife.

4.5.5. Maintenance

Inspect the underpasses regularly for problems, including debris (especially with streams and rivers), garbage, and the condition and dimensions of the banks, paths and walkways.

Back to top

4.6. MULTIPLE USE OVERPASSES

Multiple use overpasses are not only designed for use by wildlife, but also for water flow, roads or railroads that cross over the main road (including farm roads), and bike or pedestrian paths. Because of this multiple use concept, compromises are inevitable. With regard to connectivity for wildlife, multiple use crossing structures may be exposed to more disturbance than structures that are strictly designed for wildlife. On the other hand, species associated with specific habitat-e. g. , streams, rivers, riparian habitat, or vegetation associated with streams-may benefit more from a wildlife crossing structure when it is combined with a stream or river crossing. Similarly, if a crossing structure is needed because of human transportation needs (e. g. , a road or railroad crossing, or a bike or pedestrian path), wildlife may benefit if the crossing structure is not solely designed for human transportation but is modified to encourage co-use by wildlife. Adapting a wildlife crossing structure for human transportation needs may affect use by certain wildlife species. Note that many of the considerations for "wildlife only" overpasses (section 4.4) also apply to multiple use overpasses.

4.6.1. Effectiveness in Reducing Collisions with Large Mammals

The effectiveness of under- and overpasses in reducing large-mammal WVCs has been estimated at 79-97 percent, with an average of 86 percent, when used in combination with large-mammal fencing (see section 4.2).

4.6.2. Technical Specifications

Stream Sizes and Seasonal Dynamics

While most stream crossings involved underpasses rather than overpasses, some overpasses include stream crossings or wetland habitat. Stream characteristics and stream dynamics must be carefully studied to ensure that the conditions inside the crossing structure are and remain similar to that found upstream and downstream of the structure. Parameters of importance can include:

• Water velocity.
• Water depth.
• Turbulence.
• Variability in water velocity.
• Sediment.
• Debris blockage.
• Erosion of substrate inside the crossing structure or up- and downstream of the structure.
• Implications of high- and low-water events, including debris and maintenance issues.

Width of Walkways

Streams and rivers vary greatly in size, and the relative space for the stream or river on one hand and the terrestrial habitat on the other can also vary greatly. However, it is important that the terrestrial space for the walkways is accessible to the terrestrial animals, and that slightly lower or higher water levels do not immediately make the walkways inaccessible.

• For small- and medium-sized species (up to marten and rabbit size), a minimum walkway width of 0.5-0.7 m (1.6-2.3 ft) is recommended. Preferred walkway width is about 1 m (3.3 ft).
• For large mammals the minimum width of the walkway is about 2 m (6.5 ft). Preferred walkway width is greater than 3 m (10 ft).

Walkways for Wildlife Use

• The space for terrestrial mammals may have actual soil and vegetation. However, some overpasses primarily designed for water flow may consist of a concrete structure that may have walkways for terrestrial mammals attached to it. In such cases, the top side of the walkways should consist of material that has a relatively rough surface that animals will not slip on. In some situations, where erosion danger is low, soil may be placed on the walkways. Some walkway designs have a hollow beam or tube integrated into the walkway design or attached to the walkway to encourage small mammals (mice, voles) to use the walkway.
• It is important that walkways are well connected to the terrestrial habitat adjacent to the overpass and that small- and medium-sized mammals can access the walkway regardless of stream level.

Roads, Farm Roads and Recreational Bike and Pedestrian Paths

• Overpasses constructed to provide access for farmers and livestock may also be used by wildlife. To encourage wildlife use, access to the overpass by livestock may be limited to occasional movement between pastures. Gates may be installed that can be closed to prevent livestock from accessing the overpass, except when changing pastures. Typically, gates are not a substantial barrier for most wildlife species, but the gates should be left open if there are no livestock in the area.
• Overpasses for roads can be modified to encourage co-use by wildlife. Such modifications may only be feasible if traffic intensity is relatively low.
• Soil depth of greater than 0.3 m (1 ft) is recommended for structures in a temperate climate with annual precipitation of approximately 800 mm (31 in).
• Cover such as tree stumps or rocks may be provided on one or both sides of an overpass (figures 85 and 86). An overpass may also allow for grass-herb vegetation. The strip should have a minimum width of 1.5-2 m (5.0-6.5 ft) for soil only and a minimum width of 2.5 m (8.2 ft) if combined with a row of tree stumps.
• Depending on the target species (e. g. , for amphibians), a physical barrier, or a light and sound screen may be placed on the outer side of a bridge to reduce noise and light from the road below (figures 87 and 88).
• In general there should be no barrier between the strip intended for wildlife use and the area intended for human use so that wildlife experience the maximum dimensions of the crossing structure.
• Naturally, an overpass should be carefully evaluated for the impact of the weight of added soil and other material on its structural integrity.
• Tree stumps or other large or heavy material should be firmly attached on overpasses to prevent them from falling off or being thrown or pushed onto the road.
• The drainage of the crossing structure should not be blocked or hindered by the soil or cover on or inside the crossing structure.
• In dry areas and in areas where vandalism (e. g. , setting tree stumps on fire) may be a concern, consider the use of rocks or boulders instead of tree stumps.
• For roads that receive little to no use, consideration may be given to changing a paved road into a gravel or dirt road, or remove the road altogether (figure 89).

Figure 85. A row of tree stumps, grass-herb-shrub vegetation, and a road (road name: Ericaweg, overpass name: Wallenburg) on top of an overpass across the A28 motorway near the town "Zeist", The Netherlands. The overpass was originally designed for a road only (copyright: Marcel Huijser).	Figure 86. The same overpass as figure 85 showing the gentle grade of the concrete curb, which allows small animals to access the vegetated strip (copyright: Marcel Huijser).
Figure 87. Another view of the same overpass shows the black screen that was attached to the fence to reduce light disturbance from the A28 motorway (copyright: Marcel Huijser).	Figure 88. A row of tree stumps, grass-herb-shrub vegetation, and a road (overpass name: "Mauritskamp") on top of an overpass across the A28 motorway near the town Zeist, The Netherlands. The overpass was originally designed for a road only. Note the black screen attached to a chain link fence to reduce light disturbance from the A28 motorway (copyright: Marcel Huijser).
Figure 89. Road removed on bridge over motorway in Belgium and replaced with soil for vegetation and bike path (copyright: Bethanie Walder). (copyright: Bethanie Walder).

4.6.3. Implementation Considerations

Farm Roads and Recreational Bike and Pedestrian Paths

• Suitable habitat must be available on both sides of the structure.
• Paths or riding trails intended for human use should be confined to one side of a crossing structure rather than placed in the middle, leaving greater space for wildlife use. Vegetation or other cover such as tree stumps, rocks or screens can be used to reduce the impact of human use on wildlife.

While co-use opportunities (by people and wildlife) are inherently attractive, several items require careful evaluation before moving ahead with them.

• When the crossing structure is in important or sensitive ecological areas or if the target species for the crossing structure are sensitive to human disturbance, wildlife and human use should not be combined on or in the same structure. Two separate structures should be considered in this case. The width of the hourglass-shaped wildlife overpass on the right side of figure 90 is only 16 m (52.5 ft) in the middle and co-use by humans would have jeopardized its functioning for large species or other species that are susceptible to human disturbance. Building a second bridge specifically for bicyclists and pedestrians was preferred over a wider wildlife overpass that would have allowed human co-use.
• When the crossing structure is located in a multifunctional landscape with considerable human disturbance, and if the target species for the crossing structure are not very sensitive to human disturbance, and perhaps even thrive with a certain level of human disturbance (for example, raccoons thrive in an agricultural landscape), wildlife and human use can be combined in or on the same structure (figure 91).

Figure 90. A bike/pedestrian bridge adjacent to the De Borkeld wildlife overpass (right) across the A1 motorway in The Netherlands (copyright: Marcel Huijser).

Figure 91. . A wildlife overpass combined with a bike/pedestrian path in Germany. The roadway is underneath and not visible in the photograph. (Source: Mary Gray, FHWA)

4.6.4. Example Cost Estimates

If the main function of a crossing structure is to pass water or traffic, the extra costs for wildlife only apply to the increased width of an overpass. See section 4.4 for costs for overpass structures that are specifically designed for wildlife.

4.6.5. Maintenance

• The soil and other material on top of an overpass or inside an underpass should not hinder inspection of the critical elements of the structure.
• Inspect overpasses regularly for problems, including debris (especially with streams and rivers), garbage, and the condition and dimensions of the banks, paths and walkways.
• During the first few years following construction it may be necessary to provide water to irrigate vegetation on the structure, particularly if there are extended periods with little rainfall. Sufficient water will allow vegetation to settle and take root.

Back to top

4.7. ANIMAL DETECTION SYSTEMS

Animal detection systems use sensors to detect large animals (e. g. , deer, elk, moose) as they approach the road. When an animal is detected, signs are activated to warn drivers that large animals may be on or near the road at that time. Drivers may then respond through:

• A higher state of alertness.
• Lower vehicle speed.
• A combination of the two.

Driver response should then result in reduced risk of a collision with large animals and less severe collisions. When used as a stand-alone mitigation measure, animal detection systems do not increase the barrier effect of the road because they do not restrict animals in their movements across the landscape or the road.

4.7.1. Effectiveness in Reducing Collisions with Large Mammals

The effectiveness of animal detection systems in reducing large-mammal WVCs has been estimated at 82-91 percent, with an average of 87 percent. Note that this measure should still be considered experimental and that the estimated effectiveness of this mitigation measure may be adjusted as more data become available.

4.7.2. Technical Specifications

System Types

The technology used for most animal detection systems falls within one of two groups (see Huijser and others for more detail on animal detection systems): ³⁰

• "Area-cover" sensors: Area-cover sensors detect large animals within a certain range of a sensor (figure 92). Area-cover systems can be passive or active. Passive systems detect animals only by receiving signals from the surrounding environment. The two most common passive area-cover systems are passive infrared and video detection. These systems require algorithms that distinguish between, for example, moving vehicles with warm engines or moving pockets of hot air and movements of large animals. Active systems send a signal over an area and measure its reflection. The primary active area-cover system uses microwave radar.
• "Break-the-beam" sensors: Break-the-beam sensors (figure 93) detect large animals when their body blocks or reduces a beam of infra-red, laser, or microwave radio signals sent by a transmitter and received by a receiver.

Figure 92. An area-cover animal detection system (passive infrared, manufactured by ADPRO (Xtralis, USA)), designed to detect large mammals on both sides of the pole, at a WTI-MSU test facility near Lewistown, MT (copyright: Marcel Huijser).

Figure 93. An infrared break-the-beam animal detection system (manufactured by Calonder Energy, Switzerland) at a gap in a wildlife fence near the town 't Harde, The Netherlands (copyright: Marcel Huijser).

Other detection techniques include geophones that record vibrations in the ground as large animals approach, buried sensors that record changes in the electromagnetic spectrum as a large mammal walks by above ground, and radio-collared animals combined with receivers located in the right-of-way. Figure 94 shows a system where beacons are activated when radio-collared elk come within 400 m (1/4 mi) from the receivers placed in the right-of-way.

Figure 94. Activated warning signal and sign in Sequim, WA (copyright: Marcel Huijser).

Site Suitability

Suggested parameters for evaluating the suitability of a site for installation of an animal detection system are partially based on Huijser and others: (30)

• Wildlife-vehicle collisions. The site should have a history of a relatively high number of WVCs with large animals, especially ungulates (e. g. , deer, elk or moose). This is for two reasons: 1) The costs associated with the purchase, installation, operation and maintenance of an animal detection system may be compensated by the savings associated with reduced WVCs; and 2) If an animal detection system is evaluated for its effectiveness in reducing WVCs, historic data on WVCs should be available. In addition, historic WVCs from control sites are helpful.
• Animal movements. The site should be located in an area where many large animals are known to cross the road (daily movements or seasonal migration). Note: not all animal movements across a road result in WVCs.
• Traffic volume. As traffic volume increases it becomes less desirable to have large animals cross at grade. In addition, above a certain traffic volume the barrier effect of the road may be close to absolute with few animals that even try to cross the road. In that type of situation, the problem of collisions has been largely replaced with the problem of barriers to animal movement.
• Terrain. The terrain must allow for the installation of an animal detection system. For example, an abundance of ridges, gullies and rocky outcrops may make a location less suitable for a detection system, especially a break-the-beam system. Difficult terrain may also require more sensors and other equipment than relatively flat areas would require.
• Access roads. The number of access roads should be kept to a minimum to avoid gaps (blind spots) or excessive false positives caused by traffic turning on or off the road, depending on what sets off the sensors.
• Vegetation. The vegetation should allow for the installation of an animal detection system. For example, bushes and trees that grow up to the edge of the pavement increase the chances for triggering the system-i. e. , they would cause excessive false positives for most area-cover or break-the-beam systems.
• Length of road section. If an animal detection system is deployed as a stand-alone mitigation measure, the road section must be at least 805-1,609 m (0.5-1.0 mi) long to be able to accommodate for spatial errors in historic road kill data used to select the site. If an animal detection system is installed in a gap in a wildlife fence, the gap width can be variable, but a gap is typically between 30 and 200 m (98-656 ft) wide, depending on the range of the sensors and the local conditions.
• Changes in road or landscape. The road and surrounding landscape should not be scheduled to undergo major changes within the life span of the mitigation measure, which, for animal detection systems, is about 10 years. However, should changes in the landscape occur and result in changes where animals cross the road and where animals are hit by vehicles, then consideration can be given to relocating the animal detection system. Of course, there are relocation costs involved with such an effort. In addition, major changes, other than the installation of the animal detection system, would confound the results of any study into the effectiveness of the animal detection system in reducing WVCs.
• Power. The site should allow for either solar power or a connection to a 110 V power source.
• Controlled access. The site should have a low risk of theft and vandalism, such as a location on a controlled access road.

Implementation and Research Environment

• Project partners. All the organizations and individuals that have jurisdiction or that are stakeholders in activities at the study site should support the project. This includes support for installation, operation and maintenance.
• Travel costs. The site should be close to where operation and maintenance personnel are stationed to reduce travel costs.
• Pull-out. The site should have a safe pull-out location for vendors and maintenance and research personnel.

4.7.3. Implementation Considerations

The advantages and disadvantages of animal detection systems when compared to wildlife fencing in combination with wildlife underpasses and overpasses are discussed below (see also Huijser and others). (30)

Advantages of Animal Detection Systems

• Animal detection systems have the potential to provide wildlife with safe crossing opportunities anywhere along the mitigated roadway, but wildlife crossing structures are usually limited in number and they are rarely wider than about 50 m (164 ft).
• Animal detection systems are less restrictive to wildlife movement than fencing or crossing structures. They allow animals to use existing paths to the road or to change them over time.
• Animal detection systems can be installed without major road construction or traffic control for long periods.
• Animal detection systems are likely to be less expensive than wildlife crossing structures, especially once they are mass produced.

Disadvantages of Animal Detection Systems

• Although the available data on the effectiveness of animal detection systems with regard to collision reduction are encouraging, animal detection systems should currently be considered experimental as opposed to the more robust performance of wildlife crossing structures in combination with wildlife fencing.
• Currently, animal detection systems only detect large animals (e. g. , deer, elk, or moose). Small animals are hard to detect, and drivers are not warned about their presence on or near the road.
• Wildlife crossing structures can provide cover (e. g. , vegetation, living trees, tree stumps) and natural substrate (e. g. , sand, water) allowing better continuity of habitat.
• Above a certain traffic volume, perhaps around 15,000-20,000 vehicles per day, animal detection systems may be less desirable, as animals may shy away from crossing the road at grade and road mortality may be increasingly overshadowed by the barrier effect of the road.
• Some types of animal detection systems are only active in the dark and animals that cross during the daylight may not be protected.
• Animal detection systems usually require the presence of poles and equipment in the right of way, sometimes even in the clear zone, presenting a safety hazard of their own.
• Animal detection systems may substantially reduce the number of WVCs, but since they allow large animals to cross the road at grade, they will never completely eliminate WVCs. Nonetheless, the available data suggest that animal detection systems may reduce collisions with large mammals to a level that is similar to wildlife crossing structures in combination with wildlife fencing.
• Animal detection systems can be aesthetically displeasing because of the equipment placed along the road, including solar panels.
• Wildlife crossing structures and wildlife fencing are likely to have greater longevity (possibly 75 years for crossing structures and 20-30 years for wildlife fencing) than animal detection systems (possibly 10 years) and lower maintenance and monitoring costs.

The choice between:

• Animal detection systems (with or without wildlife fencing), or;
• Wildlife crossing structures in combination with wildlife fencing; currently depends on whether the success of the project is defined as:
  • Accomplishing a certain minimum result in terms of WVC reduction and/or safe crossing opportunities for wildlife (i. e. select wildlife crossing structures in combination with wildlife fencing).
  • Acquiring data that helps to further evaluate the effectiveness of other mitigation measures (i. e. select animal detection systems (with or without wildlife fencing)).

The choice also depends on:

• The problem at hand (WVCs and/or lack of safe crossing opportunities for wildlife).
• The species or species groups concerned.
• The local situation, including road, right-of-way, and landscape characteristics.

For additional considerations see Huijser and others. (30)

Potential Applications

Animal detection systems may be deployed as a stand-alone mitigation measure or in combination with other mitigation measures. Figure 95 features a schematic representation of potential applications of animal detection systems along a road:

1. A system installed over a relatively long road section without wildlife fencing.
2. A system installed in a gap with extensive wildlife fences on either side.
3. A system installed in a gap with limited wildlife fences on either side aimed at funneling the animals toward the road section with the system.
4. A system installed at the end of extensive wildlife fencing.
5. A system installed at the end of extensive wildlife fencing aimed at funneling the animals through an underpass.
6. A system installed along a low-volume road that parallels a high-volume road with an underpass.

Figure 95. Potential applications for animal detection systems.

Important issues to remember when implementing animal detection systems:

• Engineering plan: It is advisable to prepare an engineering plan that shows where and how the equipment will be installed and how the different components (e. g. , the detectors and the warning signs) will be integrated.
• Signage: Signage will have to be standardized. Signing may have to either target vehicle speed or driver alertness.
• Technological difficulties and substantial delays: Prepare for technological problems and delays following the installation of an animal detection system. It may take many months or even several years before an animal detection system may become operational.
• Monitoring and maintenance: Even systems that are initially successful will fail without proper monitoring and maintenance.

4.7.4. Example Cost Estimates

Cost depends on many parameters, including system type, the range of the sensors, and the local conditions. However, costs are about:

• $65,000 (in 2007 $) per 1.6 km (1 mi) for a system that is deployed as a stand-alone mitigation measure along both sides of a road (excluding installation and signage).
• $20,000 (in 2007 $) for a system that covers a gap in a fence on both sides of the road (excluding installation, signs, and wildlife fence).

4.7.5. Maintenance

Many of the systems that were installed at about 30 sites throughout Europe and North America experienced one or more problems during installation, operation and maintenance (table 7). The most common issues fall within four categories (table 7):

• False positives: These occur if the system is triggered by causes other than the presence of large animals (target species).
• False negatives: False negatives occur if a large animal is present, but the system fails to detect it.
• Maintenance: All systems can suffer a variety of maintenance issues. In addition, most systems require a period during which major technical problems are identified and hopefully solved. It is important to recognize this and to treat an animal detection system project as an experiment rather than the implementation of a robust mitigation measure for which most of the issues have been addressed.
• Landscape, ecology and animals: Some animal detection systems are considered to affect landscape aesthetics. In addition, some animal detection systems (e. g. radio-collar systems) require capture and handling of animals. Finally, animals may wander and loiter in the right-of-way after the warning signs have already turned off.

Table 7 shows that area-cover and break-the-beam systems seem to be particularly vulnerable to false positives and false negatives.

Table 7. Summary of issues, problems and experience with operations.

x = problem has been reported or issue applies
(x) = problem has not been reported, but it could occur
¹For Swedish system that illuminates the road and right-of-ways once an animal is detected.

Back to top

4.8. VEGETATION MANAGEMENT IN THE RIGHT-OF-WAY

This section provides a brief overview of mitigations and methods that can be incorporated into roadside vegetation management that may reduce WVCs. For more detail refer to Harper-Lore and Wilson. ³¹

4.8.1. Shrub and Tree Removal

Shrubs and trees that grow close to the edge of the pavement appear to be associated with deer- and moose-vehicle collisions. This may be due to:

• Browse and cover sought by deer and moose.
• Limited the sight distance for drivers.

Removal or trimming of shrubs and trees may increase the sight distance for drivers while reducing cover for deer and moose (figure 96). However, care should be taken that re-growth of shrubs and trees, or the grass-herb vegetation that now may grow, does not result in attracting large herbivores due to its accessibility (close to the ground, within reach of the animals) and nutritional value.

Figure 96. Recently cleared shrubs to improve driver visibility and to reduce browse for moose along the George Parks Highway (Hwy 3) in Alaska (copyright: Marcel Huijser).

4.8.2. Nutritional Value of Vegetation

Roadside vegetation or vegetation in the right-of-way may be an attractant to large mammal species, including large ungulates and bears:

• In relatively dry areas runoff from the road may allow for greener and more palatable vegetation immediately adjacent to the road compared to vegetation farther from the road (figure 97).
• In cold climates, snow melt may happen faster immediately adjacent to roads (snow plows, south facing slope of road bed), allowing for access to the vegetation, as well as earlier growth of the vegetation compared to vegetation farther from the road.
• The vegetation in the right-of-way typically differs from the surrounding vegetation because of the species that were planted or seeded, non-native invasive species that may have spread along the road corridor, or because maintenance of the right-of-way vegetation differs from that directed at surrounding areas.
• Mowing favors grass-herb vegetation over shrubs and trees and can be an attractant to large herbivores depending on the climate and the surrounding habitat. For example, in forested environments or in areas of mixed forest and croplands a grass-herb vegetation may be an attractant to grazers including deer and elk, at least during certain times of the year. Accessible and palatable browse (e. g. , re-growth from mowed or clipped shrubs and trees) may be an attractant to deer and moose in an environment dominated by mature forests with low light conditions close to the ground.
• Planted, seeded, or non-native species may be more palatable or nutritional than native species that may dominate outside of the road corridor. For example, black bears are known to feed on dandelions in the right-of-way. Right-of-way vegetation may be made less attractive to large herbivores by planting unpalatable species or reducing forage quality through vegetation maintenance (mowing, cutting, noxious chemicals).
• The nutritional value of the right-of-way vegetation may not be an attractant to large herbivores everywhere, and it may be difficult to change the species composition of the right-of-way vegetation. It is also difficult to time mowing and cutting operations such that it minimizes the nutritional value of the right-of-way vegetation and, as a result, reduces collisions with large mammals.

Figure 97. Pronghorn eating roadside vegetation (copyright: Marcel Huijser)

4.8.3. Effectiveness in Reducing Collisions with Large Mammals

Little has been published on the effectiveness of vegetation clearance or mowing and clipping of vegetation in right-of-ways and its effects on collisions with large mammals. However, clearing vegetation from roadsides resulted in a 20 percent reduction in moose-vehicle collisions in Sweden. ³²

In a study examining the effect of scent-marking, intercept feeding and forest clearing, analyses demonstrated that forest clearing resulted in a 49 percent reduction in collisions. ³³ While it is recognized that the results may not translate to a highway setting, the clearing of vegetation across a 20-30 m (70-100 ft) swath on each side of a Norwegian railway reduced moose-train collisions by 56 percent (+/-16 percent). ³⁴

4.8.4. Technical Specifications

• Depending on the species and the local situation, including land ownership, shrubs and trees may need to be removed or trimmed up to 20-40 m (66-131 ft) from the edge of the pavement. For moose this distance may need to be up to 100 m (328 ft).
• Thomas provides a summary of a variety of vegetation clearing methods used in Alaska, including hydroaxing, hand clearing, steam clearing, and spray inhibitors. ³⁵
• The timing and effect of vegetation clearing depends on the vegetation and climate at the roadway site and needs to be carefully evaluated and is likely to be highly dependent on local conditions. A detailed literature review on roadside vegetation management, plant response to tissue removal, and ungulate foraging behavior yielded recommendations for more carefully designed cutting regimes as a countermeasure for reducing moose-vehicle collisions. ³⁶ Willows cut in mid-July were found to be high in digestible energy and protein compared to plants cut at other times of the year and to uncut controls, suggesting that summer brush cutting regimes may inadvertently be attracting moose with nutritious re-growth. (36,37) Cutting in early June results in browse with significantly less nutritional value for the first two years after cutting compared to plants cut later in the growing season and uncut controls. (36,37) Rea recommends cutting roadside brush in early spring soon after leaves develop to keep nutritional value and palatability to a minimum, but recognizes operational challenges and limitations (e. g. , ground too wet for tractor use, different ungulate species-specific responses to same management regime, etc. ) and cautions that this countermeasure may not be suitable for all management areas. ³⁷

4.8.5. Implementation Considerations

• Removal of brush or trees may result in fresh growth of attractive forage (re-growth from shrubs or trees or grass-herb vegetation) that may draw herbivores (grazers and browsers) to the right-of-way, counteracting the safety gains of better visibility with increased probability of drivers encountering wildlife (figure 98).

Figure 98. Re-growth of cleared shrubs along the George Parks Highway (Hwy 3) in Alaska (copyright: Marcel Huijser).

• Shrub and tree removal must be carefully evaluated for its potential effect on the landscape (e. g. , historic value, regional landscape characteristics), as well as its effect on the vegetation's desirable function in limiting sound, light and pollutants near roadways.
• Reducing or removing large trees near roads may result in fewer collisions with these stationary objects. The removal may extend out further than mandatory clear zones.
• Shrubs and trees that grow close to the road limit sight distance for drivers and may therefore be associated with lower vehicle speed. Removal of trees and shrubs may result in higher operating speeds.
• Removing (semi-)natural vegetation from the right-of-way or making the right-of-way less attractive to large mammals may conflict with conservation interests in landscapes where the few remaining patches of (semi-)natural vegetation may be one of the most important refugia for plant and animal species in an otherwise hostile environment. This may include rare, threatened or endangered plant and animal species (figure 99).

Figure 99. Vegetation management in right-of-way may conflict with conservation interests (copyright: Marcel Huijser).

4.8.6. Example Cost Estimates

Vegetation removal requires a long-term maintenance commitment and may involve expenses to acquire right-of-way in order to manage vegetation as desired. Jaren and others calculated that if collisions are reduced by at least 50 percent as a result of removing vegetation, then the costs of vegetation removal treatment would be economically beneficial if applied in areas where more than 0.3 moose-train collisions occur per km (0.48 per mi). ³⁸ Andreassen and others estimated forest clearing for 18 km (11.2 mi) costs $530 per km ($853 per mi) or $9,548 per year (in 2007 $) ($500 per km ($800 per mi) or $9,000 per year in 2005). ³³ He estimated that based on the reduction in moose-vehicle collisions, this technique had a net benefit of $1,146 (in 2007 $) per year ($1,080 in 2005). Andreassen and others state that forest clearing may be more economical than scent-marking and supplemental feeding, and that the initial cutting is the main expense. ³³

An experimental study of vegetation removal along a railway line (20-30 m (66-98 ft) on each side) in Norway caused a 56 percent (+/- 16 percent) reduction in moose-train collisions. ³⁴ The researchers concluded that there would be an economic benefit to vegetation removal treatments in areas with more than 0.3 collisions per km (0.48 per mi), but that local evaluations are necessary to confirm that vegetation cover is the main contributor to collisions in specific sections. ³⁴ It is possible, however, that the experimental design may have overstated the collision reduction potential of vegetation removal. (34,38)

Forage repellents, planting or seeding unpalatable species, and roadside brush removal have been used with limited effectiveness or are not cost-efficient when broadly applied.³⁶

4.8.7. Maintenance

Brush or tree removal or clipping, and mowing of grass-herb vegetation in the right-of-way is a long term effort that, depending on the climate, soil and hydrology, may have to be repeated several times per year or once every few years. In Alaska, shrub vegetation in right-of-ways is typically cleared once every two to three years.

Back to top

4.9. WILDLIFE CULLING

Wildlife culling involves a substantial reduction in the population by eliminating a large number of individual animals over a short period of time. This measure is typically applied to or proposed for white-tailed deer, mule deer or elk in a (sub)urban setting where other measures to reduce human-wildlife conflicts have failed or are not considered feasible (figure 100). In some (sub)urban settings in North America, the presence of white-tailed deer, mule deer and elk has led to a variety of other wildlife-human conflicts including:

• Threats to human safety as a result of a loss of fear of humans.
• Damage to gardens and parks.
• Collisions with vehicles.
• Attraction of large predators that can also be a threat to human safety.

In more natural areas or (sub)urban areas that border relatively natural areas, damages to the ecosystem, for example as a result of overpopulation or concentration in small areas, can also be one of the reasons to consider population culling. Finally, population culling may be considered for non-native species that cause damage to the ecosystem. Culling of the target species leads, at least theoretically, to fewer individuals and reduced wildlife-human conflict, including reduced ungulate-vehicle collisions.

Figure 100. White-tailed deer can cause wildlife-human conflicts, including deer-vehicle collisions, especially in urban and suburban settings (copyright: Marcel Huijser).

4.9.1. Effectiveness in Reducing Collisions with Large Mammals

Actual data on the effectiveness of population reduction programs on wildlife-vehicle collisions are few. A field test showed that a deer population reduction program in Minnesota reduced winter deer densities by 46 percent and deer-vehicle collisions by 30 percent.³⁹ However, reductions in population size of 50 percent or more may be hard to achieve, perhaps capping the potential reduction in deer-vehicle collisions at 50 percent or less.⁴⁰

Because of the highly variable relationship between population size, population size reduction, and ungulate-vehicle collisions, this measure should be considered experimental.

4.9.2. Technical Specifications

The fertility of white-tailed deer is weakly density-dependent for adult does. However, the fertility of first-year and yearling females is strongly density-dependent, with very low fertility when population densities exceed 30 deer per km2 (78 deer per mi2). (40) These observations suggest that as population density is reduced, increased effort is needed to keep the deer density at the lower level. This phenomenon needs to be addressed in potential population size reduction programs.

The killing of does (females) is more effective for reducing the population size than the killing of bucks (males), not only because the reproductive potential of the herd is more effectively reduced, but also because does tend to stay in their existing home range while bucks have a greater tendency to disperse. The does are less likely to migrate and establish new populations elsewhere.

A modeling project by Porter and others showed that if female dispersal (i. e. , animals that leave the area) was 8 percent, culling would have to reduce annual survival to 58 percent to maintain a population just under ecological carrying capacity (the maximum sustainable population size). ⁴¹ A further reduction of the annual survival to 42 percent would keep the population at half the carrying capacity.

Baiting in order to facilitate wildlife culling increases efficacy but is illegal in some areas, and it can lead to undesirable side effects such as increased risk for spreading diseases, reduction in the consumption of natural foods and consequent changes in the ecosystem, population increase and consequent starvation, crowding, fighting and injuries of deer, deer domestication and habituation to unnatural foods and humans, decrease in hunter satisfaction, and increase in concerns of the non-hunting public.

4.9.3. Implementation Considerations

Wildlife culling can be met by strong public opposition, possibly causing delay or abandonment of the effort. A public relations campaign should be considered before a culling effort begins, especially in (sub)urban areas.

Culling efforts are more likely to result in a substantial reduction in deer population size if:

• The herd size is relatively small to begin with.
• It is a closed population that does not allow influx of animals from nearby places.

Culling may not be possible or effective:

• On private lands.
• In remote locations.
• In urban and suburban areas.

If refugia are present, more intensive effort will have to be undertaken at locations that are accessible to hunters or wildlife managers.

Sharp shooting by professionals over bait was deemed to be the most effective and adaptable culling method in an urban setting, as opposed to controlled hunts in large parks and refuges or opportunistic sharp shooting by professionals. ³⁹ If the public is to participate in the culling, consider modifying hunting regulations to stimulate hunters to target younger animals and does rather than bucks. But with a decline in the number of hunters, a shift from recreational hunting to professional culling may need to be considered.

4.9.4. Example Cost Estimates

The costs for a controlled hunt were estimated at $137 (in 2007 $) ($117 in 2001) per deer killed. The cost of using professional sharpshooters was $126, $142, and $227 (in 2007 $) ($108, $121, and $194 in 2001) per deer for conservation officers, park rangers, and police officers, respectively.³⁹ Others estimated these costs at $110-$373 (in 2007 $) ($91-310 in 2000) per deer. ⁴²

4.9.5. Maintenance

The culling effort has to be repeated periodically as the deer population will rebound to the same levels if the habitat conditions remain unchanged. In addition, the effort involved for population size reduction programs increases disproportionately with higher population size reduction goals, and substantial reductions (for example greater than 50 percent) may be hard to obtain, perhaps capping the potential reduction in deer-vehicle collisions at 50 percent.

Back to top