INTELLIGENT COUNTERMEASURES IN UNGULATE-VEHICLE COLLISION MITIGATION by Justin Edward Farrell A professional paper submitted in partial fulfillment of the requirements for the degree of Master of Science !n Fish and Wildlife Management MONTANA STATE UNIVERSITY Bozeman, Montana April 2002 © COPYRIGHT by Justin Edward Farrell 2002 All Rights Reserved 11 APPROVAL Of a professional paper submitted by Justin Edward Farrell This professional paper has been read by each member of the graduate committee and has been found to be satisfactory regarding content, English grammar and usage, format, citations, bibliographic style, and consistency, and is ready for submission to the College of Graduate Studies. Dr. Lynn R. Irby, Chair Approval for the Department of Ecology Dr. Jay J. Rotella Date Approval for the College of Graduate Studies Dr. Bruce McLeod Date Ill STATEMENT OF PERMISSION TO USE In presenting this paper in partial fulfillment of the requirements for a master’s degree at Montana State University, I agree that the Library shall make this paper available to borrowers under the rules of the Library. If I have indicated my intention to copyright this paper by including a copyright notice page, copying this paper is allowable only for scholarly purposes, consistent with “fair use” as prescribed in the U.S. Copyright Law. Requests for permission for extended quotation from or reproduction of this paper, in whole or in parts, may be granted solely by the copyright holder. Date Signa V ACKNOWLEDGEMENTS I would like to thank the Western Transportation Institute at Montana State University for providing the funding, via the Animal-Vehicle Pooled Fund Study, for this literature synthesis. During the course of his study, both a Professional Advancement Fellowship and Graduate Research Assistantship supported the author. I sincerely appreciate the early support provided by Pat McGowen, Steve Albert, and Jodi Carson in pursuit of this interdisciplinary opportunity. Many thanks to all of the Ecology and Civil Engineering faculty and staff for their scholastic assistance. Virginia Loran provided valuable assistance with the gathering of relevant literature and deserves individual recognition. Special thanks to Dr. Lynn Irby for his guidance, mentorship, and patience. I am grateful to my committee members Dr. Robert Garrott and Pat McGowen—along with fellow Ecology graduate Jeff Fennell—for their assistance, reviews, and editorial advice on earlier drafts of this manuscript. Finally, I warmly thank my wife Jennifer Kelso Farrell for her editing, personal insight, and support throughout this endeavor. VI TABLE OF CONTENTS 1. INTRODUCTION........ ....1 2. REVIEW OF TRADITIONAL MITIGATION METHODS 8 DECREASING UNGULATE DENSITY ......8 LIMITING UNGULATE ACCESS 10 Fencing, Modified Fencing, and Grade Separation... 11 Reflectors, Repellents 13 Intercept Feeding 15 IMPROVING MOTORIST ABILITY..... 16 Speed reductions ;..17 Vegetation Removal, Roadside Clear-zones...... 18 Highway Lighting 19 Signage, Public Education 20 OVERVIEW....... 21 3. REVIEW OF NEW MITIGATION OPPORTUNITIES 22 MOOSE WARNING SYSTEM, UUSIMAA, FINLAND 23 FLASH SYSTEM, NUGGET CANYON, WYOMING 24 LASER DETECTION SYSTEM, COLVILLE, WASHINGTON ......25 DYNAMIC ELK CROSSING, SEQUIM, WASHINGTON 26 OVERVIEW ..27 4. DISCUSSION 29 CONCLUSION ...32 LITERATURE CITED . 33 Vll LIST OF TABLES Table Page 1. Published research regarding ecological relationships associated with ungulate mortality on roads, predominately for the United States (after Romin and Bissonette 1996).. 4 2. Examples of published literature assessing the efficacy of various traditional mitigation techniques in reducing large mammal-vehicle collisions (categories are not mutually exclusive) 9 Vlll LIST OF FIGURES Figure Page 1. Total highway vehicle miles traveled (VMT), in trillions, in the United States from 1960 to 1998. Data are from the Bureau of Transportation Statistics, U.S. Department of Transportation 2 IX ABSTRACT Roads affect biological systems, communities, and species in numerous ways. Wildlife mortality caused by vehicles presents a serious conservation and economic problem, as collisions with large mammals are global, pervasive, and increasing. The increasing demand for faster and more efficient transportation networks has resulted in increasing conflicts with wildlife. The combination of increasing ungulate populations combined with increasing vehicle-miles traveled has heightened the significance of this problem. I reviewed the U.S. and, secondarily, European scientific literature pertinent to mitigating the effects of ungulate-vehicle collisions. Ungulate-vehicle collisions are not randomly distributed but frequently occur in predictable locations. This review presents an analysis of ungulate movement and behavior in relation to roads to further develop general conclusions about accurately locating high frequency collision areas. There have been numerous attempts to reduce large mammal mortality due to vehicles over that past few decades. Some successes in reducing ungulate - vehicle collisions have been documented with fencing, modified fencing, and grade separation via crossing structures. However, traditional solutions to ungulate - vehicle collisions are often expensive (e.g., fencing, overpasses), have limited effectiveness (e.g., reflectors, static warning signs), or may damage the environment by furthering habitat fragmentation or creating barriers to movement (e.g., ungulate-proof fencing, vegetation clear-zones). Therefore, I also present several case studies illustrating intelligent transportation systems, animal- detection driver-warning systems, applied to the problem of large mammal - vehicle collisions. Although there is significant interest and potential in animal-detection driver- warning systems, many technical issues must be addressed before they are ready for general use. I emphasize the need for more sound statistical design in determining efficacy of treatments. With the development of new technologies and transportation agencies acknowledging the ecological problems caused by roads there is potential for increased implementation of rigorous testing of techniques for reducing large mammal - vehicle collisions. 1 CHAPTER 1 INTRODUCTION Roads affect biological systems, communities, and species in numerous ways. Some conservation scientists have identified road construction and maintenance in the United States (U.S.) as one of the most widespread forms of modification to natural ecosystems over the past 100 years (Noss and Cooperrider 1994, Trombulak and Frissell 2000). Many wildlife species depend on the preservation of large tracts of intact land, but roads often fragment these tracts. Foreman (2000) estimates that 22 percent of the contiguous U.S. is altered by the nation’s road network. Road mortality can be a serious threat to species with low population levels, such as large carnivores (Weaver et al. 1996). The ecological effects of roads have been well documented (Foreman 1998), as Trombulak and Frissell (2000) provided an excellent review of the ecological effects of roads at the taxonomic level. Collisions with large mammals are an increasing problem on the roadways of the U.S., Europe, and Japan (Groot Bruinderink and Hazebroek 1996). Mortality of animals from vehicle collisions is well documented in the literature and large mammals have been documented the most (Trombulak and Frissell 2000). Results from a survey of the nation’s natural resource agencies (n = 35 reporting mortality) indicated that deer (Odocoileus spp.) conservatively account for 538,000 collisions in the U.S. in 1991 (Romin 1994, Romin and Bissonette 1996). Conover et al. (1995) extrapolated these findings for the remaining states and estimated that ungulates account for 726,000 to 1,500,000 collisions in the U.S. annually. Groot Bruinderink and 2 Hazebroek (1996) estimate the annual number of collisions with ungulates in Europe to number 507,000. Ungulate population density is a principle factor affecting ungulate presence along roads, and increased populations have been correlated with increased ungulate - vehicle collisions (Puglisi et al. 1974, Sage et al. 1983). During the last century, many ungulate populations in the U.S. have recovered due to protection from overexploitation and the application of scientific management (Messmer 2000). For example, the nation’s white-tailed deer (O. virginianus) population is burgeoning, from about 500,000 animals at the turn of the century to more than 20 million today (Cook and Daggett 1995, Hughes et al. 1996). The combination of increasing ungulate populations combined with increasing vehicle-miles traveled has heightened the significance of this problem (Fig. 1). Figure 1. Total highway vehicle miles traveled (VMT), in trillions, in the United States from 1960 to 1998. Data are from the Bureau of Transportation Statistics, U.S. Department of Transportation. 3 For example, from 1985 to 1991, deer - vehicle collisions increased an average of 69 percent in the states of California, Illinois, Maine, Michigan, Minnesota, North Carolina, Utah, and Washington (Hughes et al. 1996). Estimates about the magnitude of damage caused by wildlife are acknowledged to be conservative and inadequate to develop accurate conclusions concerning the scale and socio-economic consequences of large mammal - vehicle collisions (Groot Bruinderink and Hazebroek 1996, Messmer 2000). Not only do large mammal - vehicle collisions affect the populations and ecosystems involved, they also result in costly and harmful affects to humans. Although the large number of smaller animals hit by vehicles may have ecological consequences, these do not typically cause a significant human safety problem (Cook and Daggett 1995). Large mammal collisions can involve safety and economic impacts, including: injuries, fatalities, property damage, increased insurance premiums, lost hunting revenue, and carcass removal expenses (Conover et al. 1995, Conover 1997). Approximately 230 fatalities and 29,000 human injuries occur annually in the U.S., while in Europe, 300 fatalities and 30,000 injuries are estimated to occur annually from ungulate - vehicle collisions (Conover et al. 1995, Groot Bruinderink and Hazebroek 1996). Conover et al. (1995) and Cook and Daggett (1995) estimated the total cost in property damage due to ungulate collisions in the U.S. to be in excess of $1.1 billion dollars annually. Estimates for Europe are similar (Groot Bruinderink and Hazebroek 1996). Many factors affect the spatial and temporal distribution of large mammal - vehicle collisions. Large mammal - vehicle collisions are not randomly distributed and 4 often occur in relation to habitat or topographical configurations, which concentrate large mammal crossings along particular sections of a roadway (Table 1). Table 1. Published research regarding ecological relationships associated with ungulate mortality on roads, predominately for the United States (after Romin and Bissonette 1996). • Reference Species Location Habitat Type Peek and Beilis 1969 O. virginianus PA Mixed hardwood Carbaugh 1970 O. virginianus PA Mixed hardwood Vaughn 1970 O. virginianus PA Mixed hardwood Beilis and Graves 1971 O. virginianus PA Mixed hardwood Puglisi et al. 1974 O. virginianus PA Mixed hardwood Reilly and Green 1974 O. virginianus MI Mixed hardwood Carbaugh et al. 1975 O. virginianus PA Mixed hardwood Mansfield and Miller 1975 O. hemionus CA Varied Allen and McCullough 1976 O. virginianus MI Mixed hardwood Goodwin and Ward 1976 O. hemionus WY Prairie Kasul 1976 O. virginianus MI Mixed hardwood Rost and Bailey 1979 O.hem/C.e.can CO Pine/Juniper/Shrub Sicuranza 1979 O. virginianus MI Mixed hardwood Kress 1980 O. virginianus PA Mixed hardwood Sage etal. 1983 O. virginianus NY Mixed Hardwood/Conifer Bashore et al. 1985 O. virginianus PA Mixed hardwood Waring et al. 1991 O. virginianus IL Mixed hardwood/Ag. Groot Bruinderink and Hazebroek 1996 Various Europe Varied Calvo and Silvy 1996 O.vir.clavium FL Varied Pafko and Kovach 1996 O. virginianus MN Mixed Hard./Conifer/Ag. Gunther etal. 1998 Various Yellowstone N.P. Varied Finder etal. 1999 O. virginianus IL Varied (GPS) Iverson and Iverson 1999 O. virginianus OH Varied Hubbard et al. 2000 O. virginianus IA Varied (GPS) Rowland et al. 2000 C. elaphus OR Pine/Bunchgrass Forest Knowledge of such factors remains critical to reducing large mammal - vehicle collisions on existing and future roads (Finder et al. 1999). Ungulate activity patterns contribute greatly to animals appearing along the roadway. Beilis and Graves (1971) found that the number of deer killed per month on Interstate 80 in central Pennsylvania was strongly 5 correlated with the numbers of deer observed grazing along the right-of-way. Generally, ungulate activity levels tend to be highest in early morning and evening, times when there is typically decreased visibility and increased commuter traffic (Putman 1997). Peek and Beilis (1969) correlated dawn and dusk peaks in collision numbers to increased deer movement during those times. Leedy (1975) noted that elk {Cervus elaphus) mortality due to vehicles occurred primarily at night. Haikonen and Summala (2001) found that in Finland, the crash rate for moose {Alces alces) and white-tailed deer was highest 1 hour after sunset. While no ungulate is strictly diurnal, crepuscular, or nocturnal, all have proven sensitive to human disturbance and tend to avoid open areas during the day (Putman 1997). Collisions between large mammals and vehicles increase when roadways are constructed through prime habitat or intersect ungulate migration routes (Reed and Woodard 1981). Vegetation and topography can work synergistically to funnel deer to predictable crossing areas. Foreman and Hersperger (1996) outline three (of six) major types of flows across landscapes that prove pertinent to highway mortality. They are surface water in streams, wildlife in major corridors, and vehicles on roads. Indeed, Hubbard et al. (2000) found that bridges in Iowa “always indicate points where major edge-creating landscape features intersect roadways” and thus provided the best indicator of high incidence areas of white-tailed deer - vehicle accidents. White-tailed deer occupy a wide variety of habitats but often prefer forested areas interspersed with agriculture (Bashore et al. 1985). When both types are available, white-tailed deer are more likely to be seen near deciduous forest types than deciduous-conifer types (Sage et al. 1983); near 6 the edge of wooded areas than in the middle of it (Puglisi et al. 1974); and in areas of topographic inclines and declines including roadway cuts and fills rather than level areas (Carbaugh et al. 1975). Being edge-adaptable, white-tailed deer stay close to forest patches, creeks, and shelterbelts to allow for ease of escape. Ungulate behavior can contribute to high incidences of collisions. In Pennsylvania, Feldhamer et al. (1986) documented that of 44 seasonal home range estimates for white-tailed deer, 16 (36.4%) included segments of 1-84 or a secondary roadway during 1 or more seasons. Ungulates can habituate to roadways and will regularly cross minor roadways during daily movements within their home ranges to reach favored foraging (Waring et al. 1991, Putman 1997) and resting areas (Carbaugh et al. 1975). Foraging can also attract ungulates to the road right-of-way because of palatable roadside plantings or vegetation succession (Case 1978, Feldhamer et al. 1986, Waring et al. 1991). Any ungulate along the right-of-way could easily wander into the roadway, substantially increasing the risk to motorists. Early green-up of right-of-way vegetation is a primary cause of sika (Cervus nippori) deer - vehicle accidents in Japan (Kaji 1996). Vehicle collisions with ungulates have also been linked with breeding and dispersal activities (Jahn 1959, Case 1978, Feldhamer et al. 1986, Groot Bruinderink and Hazebroek 1996). Studies in Pennsylvania and Michigan suggest collisions with white¬ tailed deer peak during the autumn breeding season (rut), when both females and males are more peripatetic (Puglisi et al. 1974, Allen and McCullough 1976). There is usually another small peak in spring corresponding to parturition and dispersal of young (Reilly 7 and Green 1974). For the U.S. as a whole, elk and mule deer (O. hemoinus) are most vulnerable to highway collisions in winter when driven to lower elevations by snow accumulation (Leedy 1975). Moose tolerate snow, but great depths can encumber movement and encourage moose to use plowed roads for travel (Garrett and Conway 1999). Moose are also particularly vulnerable to collisions in spring and early summer, when leeching highway salts attract them to roadside pools (Fraser 1979, Fraser and Thomas 1982). The problems associated with ungulate - vehicle collisions are global, pervasive, and increasing. Wildlife mortality caused by vehicles presents a serious conservation and animal damage problem (Haikonen and Summala 2001). The importance of accurately identifying high crash areas cannot be understated, as the success of many mitigation measures is dependant upon the accurate location of high incidence crash areas and the understanding of all the factors that contribute to them. As Putman (1997) states, “Selection of the appropriate deterrent measures in any given situation is itself dependent upon proper understanding of the actual pattern of such accidents... . Without such biological understanding, we cannot really determine where preventative measures should be concentrated, or suggest a priori which of a variety of deterrent options is likely to be most effective in given circumstances.” The remainder of this paper reviews the many research efforts, both past and present, which have attempted to reduce large mammal - vehicle collisions. 8 CHAPTER 2 REVIEW OF TRADITIONAL MITIGATION METHODS There have been numerous attempts to reduce large mammal mortality due to vehicles over the past few decades. Most authors attempt to evaluate a single mitigation technique, which makes comparisons among techniques difficult. Although good reviews do exist for Europe (Groot Bruinderink and Hazebroek 1996, Putman 1997) and the U.S. (Romin and Bissonette 1996), most of the literature suggests that many mitigation techniques have limited utility (Table 2). Researchers applying traditional countermeasures have generally approached the problem of ungulate collisions with one or more of the following goals (1) reduce ungulate density in problem areas, (2) prevent or deter animal access to the road, (3) improve the motorist’s ability to avoid a collision by elevating the awareness of the hazard. Decreasing Ungulate Density Local population density is one of the primary drivers of wildlife-traffic mortality (Finder et al. 1999, Joyce and Mahoney 2001). Allen and McCullough (1976) suggested controlling deer population numbers through harvest as one of the most effective means of reducing deer - vehicle accidents. Hunting has proven to be a fundamental and effective tool for managing ungulate populations. Sage et al. (1983) noted that hunting negatively influenced observation rates of deer along forest roads in New York, largely because of reduced deer density. In Newfoundland, the spatial distributions of moose - 9 vehicle collisions are dependent upon both traffic volume and moose densities (Joyce and Mahoney 2001). Table 2. Examples of published literature assessing the efficacy of various traditional mitigation techniques in reducing large mammal-vehicle collisions (categories are not mutually exclusive). , Reference Location Mitigation Technique Effective Reedetal. 1975 CO Highway Underpasses Ward 1982 WY Highway Fencing and Underpasses Ludwig and Bermicker 1983 MN Highway Fencing and One-way Gates Schafer and Penland 1985 WA Swareflex Reflectors Wood and Wolfe 1988 UT Intercept Feeding Jarenetal. 19911 Norway Vegetation Removal Lavsund and Sandegren 1991 Sweden Highway Fencing, Vegetation Removal Foster and Humphrey 1995 FL Highway Underpasses Messmer et al. 1999 UT Temporary, Seasonal Signage Clevenger et al. 2001 Alberta, Canada Highway Fencing Apparently Ineffective Woodward et al. 1973 CO Swareflex Reflectors Pojar et al. 1975 CO Lighted, Animated Deer Crossing Signage Falk et al. 1978 PA Highway Fencing Reed and Woodard 1981 CO Highway Lighting Feldhamer et al. 1986 PA Highway Fencing Lavsund and Sandegren 1991 Sweden Repellents (light, sound, and scent) Ford and Villa 1993 CA Swareflex Reflectors Reeve and Anderson 1993 WY Swareflex Reflectors Ujvarietal. 1998 Denmark WEGU Reflectors Inconclusive Beilis and Graves 1971 PA Highway Fencing Puglisietal. 1974 PA Highway Fencing Gilbert 1982 ME Deer Mirrors Pafko and Kovach 1996 MN Deer Reflectors Lehnert and Bissonette 1997 UT Highway Crosswalk Structures 1 This study assessed efficacy in reducing moose - train collisions. 10 Both Michigan and Illinois have used harvest in an attempt to reduce local deer populations and decrease deer - vehicle collisions (Romin and Bissonette 1996). Michigan indicated that hunting was successful (Romin and Bissonette 1996), whereas despite local population declines in Illinois, deer - vehicle collisions did not subsequently decrease (Waring et al.1991). Inconsistencies such as these suggest that the frequencies of ungulate - vehicle collisions are not simply density dependent. Similarly, the utility of highway mortality as an index of species population trends has been debated (Jahn 1959, McCaffery 1973, Loughry and McDonough 1995). According to Case (1978), ungulate - vehicle collisions are the function of the following parameters: population densities, seasonal behavior, traffic speed, traffic volume, and roadside vegetation. Limiting Ungulate Access Management of ungulates on roads often consists of countermeasures designed to reduce crossing or change the pattern of crossing activity (Putman 1997). The goals of many countermeasures include altering, limiting, or preventing animal access to the roadway in areas exhibiting frequent collisions. Traditional countermeasures attempting to accomplish these goals include: (1) fencing, modified fencing, and grade separation through overpasses and underpasses to prevent animals from entering the roadway; (2) reflectors, scent repellents or sound signals that temporarily arrest ungulate movement; and (3) vegetative plantings to alter ungulate movement patterns or the relative attractiveness of right-of-way versus non right-of-way vegetation. 11 Fencing, Modified Fencing, Grade Separation Building barriers, such as fences, is the most common approach to prevent ungulate - vehicle collisions (Cook and Daggett 1995). A variety of fences exist to address the problem and they vary in cost and effectiveness (Clevenger et al. 2001). Most of the fencing used to limit human access to high capacity freeways is 1.22 m woven or barbed wire (Cook and Daggett 1995). However, ungulates can readily jump such fences making ungulate-proof fencing necessary. Ungulate-proof fencing, generally 2.2 to 2.7 m high, is considered an effective restraint and is typically used to channel ungulates to crossing structures (Falk et al. 1978, Ward 1982, Cook and Daggett 1995). Although the literature offers no clear guidance on the length of ungulate-proof fencing (Foster and Humphrey 1995) - ungulate-proof fencing must be of sufficient length so as to not encourage end-runs (Ward 1982, Feldhamer et al. 1986). End runs occur when ungulates travel to the end of the fence and become trapped in the road corridor, often re¬ concentrating collisions. Because of this phenomenon, ungulate fencing is sometimes modified by additional one-way gates, which allow ungulates caught within the paved area to escape through the gate (Reed et al. 1974). Fencing is only effective when designs take into account local topography, snow accumulation, and when they are well maintained (Ward 1982). White-tailed deer have been documented crawling through fence openings less than 23 cm wide (Falk et al 1978). Overpasses, underpasses, and crosswalks are sometimes used in combination with fences to increase permeability across, over, or under the roadway. Grade separation is the process of channeling ungulate movement toward crossing structures, mainly through 12 fencing, so that they pass over or under the highway rather than walking across it at grade (Cook and Daggett 1995). Several studies demonstrate that grade separation, through the use of overpasses and underpasses, is an effective measure to increase permeability of roads for many species of wildlife (Foster and Humphrey 1995,Yanes et al. 1995, Clevenger 1998, Clevenger and Waltho 2000, Gloyne and Clevenger 2001). However, target species can be initially reluctant to use crossing structures (e.g. mule deer; Reed et al. 1975); therefore, it is important to determine the design features of crossing structures that increase efficacy (Rodriguez et al. 1996). Several studies demonstrate that large mammal use of any crossing structure is influenced by structure dimension and location, nearby cover, and human activities (Reed et al. 1975, Singer and Doherty 1985, Clevenger and Waltho 2000). Generally, the larger and more open crossing structures are the most effective. Reed et al. (1975) recommend a height and width of 4.3 m or larger for ungulate underpasses with the shortest practical length. Reed et al. (1975) found that neither artificial lighting nor skylights increased the use of underpasses by mule deer in Colorado. Recent work shows that carnivores were less likely to use crossing structures that exhibited high levels of human activity (Clevenger and Waltho 2000, Gloyne and Clevenger 2001). A relatively inexpensive alternative to grade separation are crosswalks, which consist of a break in fencing (at grade) accompanied by signs that warn motorists of crossing animals. Lehnert and Bissonette (1997) estimate the cost of crosswalks for a 2 and 4-lane highway to be $15,000 and $28,000 respectively, as compared to retrofitting underpasses on those same highways to be $92,000 and $173,000 (US). 13 Reflectors, Repellents Wildlife reflectors do not physically block animals entering the roadway, but they purport to discourage animals from entering the road by creating a visual barrier via incident light reflected by headlights until vehicles have passed (Gilbert 1982). Typical systems consist of a series of reflectors mounted on posts installed at regular intervals along the roadside. Reflector systems are relatively inexpensive, estimated to cost between $8,000 and $10,000 per mile (Gilbert 1982). Several states have experimented with reflective devices, though results were often mixed (Romin and Bissonette 1996, Putman 1997). Three types of reflectors exist: polished metal mirrors and WEGU reflectors (Walter Drabing KG, Kassel, Germany) that reflect incident light from headlights (e.g. Gilbert 1982 and Ujvari et al. 1998, respectively), and Swareflex reflectors (D. Swarovski and Company, Tirol, Austria), which transmits incident light as a continuous visual barrier of red or blue-green light (e.g. Schafer and Penland 1985). Gilbert (1982) noted that regular deer mirrors were ineffective in reducing deer - vehicle collisions in Maine, even though small sample size limited any formal conclusions. Swareflex reflectors reduced deer mortalities in Iowa (Gladfelter 1984) and Washington (Schafer and Penland 1985) but were unsuccessful in Colorado (Woodard et al.1973), Illinois (Waring et al. 1991), California (Ford and Villa 1993), and Wyoming (Reeve and Anderson 1993). Fallow deer {Cervus dama) in Denmark exhibited increasing indifference to WEGU reflectors, suggesting that they too are ineffective at reducing deer - vehicle accidents (Ujvari et al. 1998). It is worth noting that because 14 reflectors depend on incident light from vehicles, their effective operation in any capacity is limited to nighttime or other low light conditions (Pafko and Kovach 1996, Putman 1997). Furthermore, Zacks (1986) questioned the notion that ungulates avoid the color red when the results from his experiment provided no evidence that white-tailed deer responded any differently to the presence of red Swareflex reflectors, white reflectors of the same geometry, or a headlight beam without reflectors. Ujvari et al. (1998) notes that reflectors are not a reliable method to reduce ungulate - vehicle collisions on a long-term basis, due to technical limitations and ungulate propensity to habituate to reflectors. Wildlife repellents exist in many forms and with many different repelling principles, but most applied to ungulate - vehicle crashes utilize high frequency sound waves or odors that are either unpleasant to the animal or frighten them. Sound repellents may be stationary or installed as ultrasonic whistles on vehicles (Romin and Dalton 1992). When motorists reach certain speeds, the whistles produce frequencies of 16 to 20 kHz (Romin and Dalton 1992). In theory, the tone warns animals of approaching traffic. However, Romin and Dalton (1992) failed to detect behavioral response differences in 150 groups of mule deer that were exposed to whistles in Utah. In Sweden, stationary sounds of 70 dB and frequencies up to 50 kHz were employed, yet moose failed to respond to sounds less than 21 kHz (Lavsund and Sandegren 1991). Bomford and O’Brien (1990), in a comprehensive review of sonic deterrents in animal damage management, state “devices producing sounds other than communicative signals (alarm or distress) have no persistent effect on animals’ space use or food intake.” There is also 15 evidence of habituation to sonic repellents with prolonged or frequent exposure (Bomford and O’Brien 1990, Lavsund and Sandegren 1991). Scent appears to be a better deterrent for animals than sound, but Lavsund and Sandegren (1991) noted that scents have had limited effectiveness in reducing moose - vehicle collisions in Sweden. A research team at the University of Umea (Sweden) synthetically produced a substance that resembled the component smells in wolf urine. The motivating principal behind this development is that all ungulates possess a natural instinctive fear of predators (Koehne 1991). Testing in Sweden is ongoing, and preliminary results indicate it may be effective. Fraser and Hristienko (1982) demonstrated that putrescent material (putrescent egg and cattle manure) and certain volatile compounds (isobutyric acid and creosote) were effective in repelling moose from salty roadside pools in Ontario. However, some researchers question the long-term utility of scent deterrents because the substances tend to deteriorate over time (Fraser and Hristienko 1982). Intercept Feeding Understanding why an ungulate approaches the roadway is also a way to understanding how to prevent collisions. In some areas, lack of quality forage in roadside forests caused deer to use the right-of-way as a food source (Feldhamer et al. 1986, Waring et al. 1991). For white-tailed deer, the highway right-of-way is an “increasingly common, if not ‘natural,’ aspect of their environment” (Carbaugh et al. 1975). Planting unpalatable species within the right-of-way or creating alternate feeding areas away from 16 the roadway can discourage ungulate use of roadside habitat or intercept ungulates moving toward the road. Indeed, Leopold (1933) attributed reduced deer - vehicle accidents in Michigan to intercept supplementing of salt away from the highway. Similarly, Fraser and Thompson (1982) showed that alternative salt sources could be established to lure moose away from the highway. Wood and Wolf (1988) showed intercept feeding of mule deer to be useful at reducing deer - vehicle crashes in Utah. They further suggest that intercept feeding might reduce deer - vehicle collisions by <50 percent over the short term. However, as intercept feeding is labor intensive and ungulates may become dependent upon supplemental food, it is not recommended for long-term ungulate - vehicle crash reductions (Wood and Wolfe 1988). Improving Motorist Ability Several measures exist that attempt to improve a driver’s ability to react should they encounter a large mammal in the roadway (Koehne 1991). By improving the driver’s ability to react, both the severity and the frequency of large mammal - vehicle crashes can be reduced (Koehne 1991). Some measures for improving the driver’s ability to react have included: (1) reducing vehicle speeds in high crash areas to allow the driver more time to react after spotting an animal; (2) removal of vegetation adjacent to the roadway to allow the driver to see the animal before it enters the roadway; (3) installing additional roadway lighting to improve nighttime visibility; and (4) through signing and public education programs. With the exception of habitat changes associated with areas cleared of vegetation, measures included in this category are considered less ecologically 17 severe than those discussed previously because they do not restrict or hamper large mammal movements. Instead, these countermeasures attempt to give vehicle drivers an early warning of a large mammal’s presence. Speed Reductions Early reports on road-killed wildlife implicated increasing traffic speeds as a potential factor in increasing collisions (Stoner 1925, Haugen 1944). Since then, high vehicle speeds have commonly been considered one of the central causes of large mammal - vehicle collisions (Pojar et al. 1975, Case 1978, Groot Bruinderink and Hazebroek 1996). Bashore et al. (1985) found that the probability of deer - vehicle collisions decreased with decreases in speed limits. Gunther et al. (1996) concluded that vehicle speeds are the primary factor contributing to large mammal - vehicle collisions in Yellowstone National Park. In Yellowstone, large mammal - vehicle collisions occurred more than expected on roads with posted speeds of 88.5 km/h (p<0.10) and less than expected on roads with posted speeds of 72.4 km/h or less (p<0.10; Gunther et al. 1996). By reducing vehicle speeds through high incident locations, motorists are potentially given a greater opportunity to avoid large mammal - vehicle collisions. Lavsund and Sandegren (1991) demonstrate that reduced speed limits at least reduced the severity of moose - vehicle collisions in Sweden. However, deer - vehicle accidents increased with increased vehicle speeds only to an asymptote (88 km/h), after which they decreased (Allen and McCullough 1976). While speed reductions are regarded as a solution among many natural resource agencies, reductions in posted speed have not been thoroughly 18 evaluated with regard to frequency of large mammal - vehicle collisions (Romin and Bissonette 1996). While speed is linked to the probability of being in a large mammal - vehicle collision, such events prove to be complex, which are seldom attributable to a single factor (see Transportation Research Board 1998). Beside the obvious tradeoff between speed reductions and travel time, other more subtle factors such as speed distribution (range of speed) contribute to collision involvement. Reducing the posted speed below highway design speed has been shown to increase speed distribution and collision rates can be higher on roads with wider ranges of speed (TRB 1998). Indeed, slow drivers can be just as dangerous as fast drivers (TRB 1998). Although speed reductions are a commonly suggested option to reducing the probability of a large mammal - vehicle collision, such reductions may cause other safety and economic (enforcement) problems. Vegetation Removal Roadside Clear-zones An important design feature in roadside safety is the provision of an unobstructed space alongside the roadway for errant vehicles to recover and stop without striking a hazard (e.g., trees, power lines, etc.; Ray 1998). Because these unobstructed spaces, referred to as clear-zones, must allow sufficient time for vehicles to recover or stop, appropriate width is determined by the design speed of the roadway and the average daily traffic (ADT). For example, the U.S. Roadside Design Guide recommends a roadside clear-zone of 9 m for roadways with an ADT of 6,000 vehicles per day and a design speed of 100 km/h (AASHTO 1996; 3-3). Researchers have suggested that extending the 19 clear-zone out even further improves visibility for motorists and allows a measure of advanced warning, giving the driver a greater scanning area and more time to reduce speed or avoid a crash (Lavsund and Sandegren 1991). For example, Jaren et al. (1991) found that spraying vegetation with the herbicide glyphosate (Roundup®) in a 20-30 m section on each side of two railway lines caused a 56 percent (±16 %) reduction in the number of moose-train collisions. However, vegetation clear-zones may also attract animals to the roadside where early successional vegetation is exposed, and the method is expensive and must be well maintained due to regrowth (Lavsund and Sandegren 1991). Highway Lighting Because the preponderance of ungulate - vehicle collisions occur in the hours from sunset to sunrise, the installation of roadway lighting was thought to improve motorist visual acuity (Reed and Woodard 1981). It is hypothesized that with increased roadside lighting, animals can be more easily sighted prior to entering the roadway, thus reducing the probability of an ungulate - vehicle collision. Reed and Woodard (1981) noted that highway lighting was successful at reducing serious vehicular accidents in urban settings. However increased highway lighting did not affect motorist speeds, deer crossings, or crossings-per-accident ratios in Colorado, and thus was not effective at reducing deer - vehicle collisions (Reed and Woodard 1981). 20 Signage, Public Education Conventional warning signs have been widely used to alert both frequent and infrequent motorists of dangers along the roadway (Pojar et al. 1975). Forty of 43 states (93 %) surveyed by Romin and Bissonette (1996) used static deer-waming signs. However, static deer-waming signs have been shown to have a limited effect on driver behavior but may be useful for public relations and liability considerations (Pojar et al. 1975). Warning signs could be effective if they require a specific driving modification (i.e. reduced speed). However, signs are common and not necessarily predictable of ungulate presence, thus motorists become complacent to the warning (Romin and Bissonette 1996, Putman 1997). Messmer et al. (1999) show seasonal warning signs to be effective in reducing motorist speeds on Highway 89 in Utah, but cautions “drivers may initially slow down because of the flashing lights and signs, but if they do not encounter deer, their speeds may increase.” Studies have shown that drivers base their behavior on what they see on the road in front of them and not necessarily on the signing (e.g. Aberg 1981). Even seasonally lighted, animated deer-crossing signs failed to elicit enough of a motorist response to reduce the number of deer killed attempting to cross State Highway 82 in Colorado (Pojar et al. 1975). Similarly, Lehnert and Bissonette (1997) found a lack of motorist response to crosswalk warning signs and surmised that they may have been mistaken for standard deer-waming signs. Public education programs inform motorists of potential dangers in the roadway environment. The intent is to alter driving behavior and improve alertness levels. Public awareness programs were used by 22 of 43 states (51 %) reporting to Romin and 21 Bissonette’s (1996) survey. Although, 24 percent believed the programs successful (62 % inconclusive) these programs have not been rigorously evaluated (Romin and Bissonette 1996). Potential information distributed through a local public education effort should include statistics showing magnitude and severity of the problem, high crash locations, the times at which the risk is highest, what a driver can do to minimize risk, and what a driver should do if a crash does occur. In spite of inconclusive results, educating motorists of the risks of large mammal - vehicle collisions remains a fundamental recommendation of several authors (Pojar et al. 1975, Lavsund and Sandegren 1991, Groot Bruinderink and Hazebroek 1996). Overview Some successes in reducing ungulate - vehicle collisions have been documented with ungulate-proof fencing, modified fencing, and grade separation via crossing structures. However, traditional solutions to ungulate - vehicle collisions are often expensive (e.g., fencing, overpasses), have limited effectiveness (e.g., reflectors, static warning signs), or may damage the environment by furthering habitat fragmentation or creating barriers to movement (e.g., ungulate-proof fencing, vegetation clear-zones). For most techniques, adequate tests have not been conducted, or when tests were conducted the sample size was too small to validate effectiveness of most treatments. Few studies incorporated any statistical design in the planning process (Gilbert 1982, Romin and Bissonette 1996). 22 CHAPTER 3 REVIEW OF NEW MITIGATION OPPORTUNITIES In a search for more sophisticated ways to reduce animal - vehicle collisions, transportation agencies have turned to advanced technology solutions implemented as Intelligent Transportation Systems (ITS; Hughes et al. 1996). Such systems detect large mammal presence on the roadside and provide an active, dynamic warning to the driver. ITS pilot systems focus on two aspects of the problem: (1) the ability to detect ungulate presence on or approaching the roadway; and (2) the driver’s response to the dynamic warning signs. Animal detection can be accomplished through a single method or a combination of methods. Currently, vendors are promoting microwave radar, passive and active infrared, fiber-optic grating, seismic sensors, or thermal imaging technologies to detect large mammals. Image recognition software can be used to identify animal presence in video or infrared images. Buried seismic sensors may detect ground vibrations caused by animal presence. A beam may be broken such as microwave, laser or other light-wave sent between a transmitter and a receiver. Microwave radar detection systems have the highest potential for success because the systems are capable of reducing false detections caused by blowing vegetation or other small animal intrusions (Taskula 1997). A disadvantage of complex systems of this type is the need to use advanced software packages, which require the ability to process many algorithms. Once animals have been detected in the right-of-way, ITS systems can warn drivers of the ungulate presence via dynamic signing, flashing beacons, or audible 23 warnings. Given the limited effectiveness of static signing at eliciting a motorist response, dynamic signing is perceived as more appropriate and can range from a static sign with a flashing beacon to a full matrix variable message sign (Pojar et al. 1975, Cook and Daggett 1995). Recently, there have been several pilot animal-detection, driver- warning systems installed, including: (1) Moose Warning System, Finland; (2) FLASH System, Wyoming; (3) Laser Detection System, Washington; and (4) Dynamic Elk Crossing, Washington. I have reviewed each of these systems. Two additional systems are under development, one from the Minnesota Department of Transportation and another planned by the Western Transportation Institute at Montana State University (14- state, pooled-fund study), although no data were available on the efficacy of either system. • Moose Warning System. Uusimaa. Finland In Finland, moose account for about 1,300 collisions annually, costing about $10 million (US) in human injury and property damage (Taskula 1997). In 1995 on all public roads in the Uusimaa region, there were 435 moose collisions reported to police. On Highway 7, a moose-detection driver-warning system was installed to increase motorist awareness of the hazard and to alleviate vehicle damage. Here, moose were funneled by 1,650 m of fence into a predictable 220-m opening that allowed moose to cross the road (Taskula 1997). A motion-detecting-system configuration, utilizing microwave radar sensors (two per pole, 50 m apart), spanned the 220-m crossing in the right-of-way. Positive detections triggered fiber-optic moose-warning signs located approximately 150- 24 200 m upstream of the crossing/detection zone on both sides of the road (4 in all). Minor adjustments concerning moose movement rates were necessary to avoid false detections due to blowing grass and small birds (Taskula 1997). In order to reduce false detections caused by rain and air pressure fluctuations, passive infrared detectors and a rain detector were also built into the system. Driver reaction, in the form of reduced speed, was measured during periods of sign activation using inductive loop traffic detectors. When encountering the activated signs (versus control periods), motorists decreased speed in rainy conditions (14.0 to 15.6 km/h) and at night (1.6 to 2.6 km/h), yet there was little impact on motorist speed during daylight periods with good visibility (increase of 0.4 to 0.5 km/h; Sabik Oy, unpublished report). Flashing Light Animal Sensing Host (FLASH) System. Nugget Canyon. Wyoming On U.S. Highway 30 between Kemmerer and Cokeville in Wyoming (milepost 30.5) hundreds of mule deer are killed annually during seasonal migrations (Gordon and Anderson 2002). The extensive road crossings by migrating mule deer, along with occasional crossings by elk, pronghorn (Antilocapra americand), and moose prompted officials to install 11.3 km (7 mi) of deer-proof fencing in 1989 (see Reeve and Anderson 1993), with one opening for ungulate crossings. Additionally, an ungulate-detection driver-warning system was installed at each side of the fence openings. The detection system consisted of two passive infrared radar sensors detecting deer body heat and a backup system of 10 buried geophone sensors detecting ground vibrations caused by ungulates. The infrared detection system coupled with flashing beacons and signing to 25 form the driver-warning system, while the geophone system served as a partial backup and gathered data on deer crossings. Standard signage was modified to read “Deer on Road when Lights are Flashing.” The infrared system turns the lights on only when an animal is detected in the crossing zone. Additionally, highway advisory radio plays a 30- second informative message about the crossing zones and why drivers should reduce speed. Initially, technical issues such as detection zone layout, sensor alignment, and optimal positioning of signs hampered the evaluation of the effectiveness of this system. Researchers found that more than 50 percent of the detections registered by the FLASH system were false detections, although the backup geophone system functioned near perfectly (Gordon and Anderson 2002). Data collected to gauge driver reaction to the system revealed that passenger vehicles and tractor-trailers significantly reduced their speed, by 18.7 and 10.1 km/h respectively (11.6 and 6.3 mph), when the signs were animated and a mule deer decoy was deployed in the crossing (Gordon and Anderson 2002). Other treatments resulted in decreases in vehicle speeds of 8 km/h or less (<5 mph) which were not deemed sufficient enough to reduce the likelihood of a deer - vehicle collision by the authors. These results showed that speed reduction was generally higher for passenger cars than tractor-trailers. Very few large trucks responded with any reduction in speed. Laser Detection System, Colville, Washington The Washington Department of Transportation identified MP 290 on US Highway 395, south of Colville near Chewelah, Washington, as a high deer - vehicle 26 collision area (J. Schafer, Washington Department of Transportation, personal communication). The highway segment is 402 m (!4 mi) in length with the necessary clear line-of-sight along the right-of-way to support a simple broken-beam detection system. The system consisted of two lasers (one on each side of the road); two standard deer warning signs with supplemental plaques, which read “When Flashing” and red beacons. The system was partially solar-powered and activated the warning beacons when the detection beam was broken. Unfortunately, the system experienced numerous technical and maintenance problems. Sighting the laser proved difficult, as proper alignment at threshold distances (400 m for most beam technologies) can be difficult to obtain and sustain. Distortion of the laser via direct solar radiation disrupted sensor alignment and lead to detection failures and false detections without shade hoods. Theft of the solar power units has also been a problem. Dynamic Elk Crossing, Sequim. Washington On the Olympic Peninsula near the city of Sequim Washington, approximately 10,000 vehicles pass through on Highway 101 per day. From 1994 to 2000 despite standard crossing signage installed in 1996, vehicles killed 12 resident Roosevelt elk (C. elaphus roosevelti) whose home range was bisected by the road. Collisions between vehicles and elk presented a safety concern for the region, which was likely to increase when the new Sequim Bypass, completed in the Fall 2000, produced increased traffic volumes and road density. To address the problem, local officials installed an ungulate- detection driver-warning system in December 2001. Eight adult elk from the 81-member 27 elk herd were radio collared. The VHP signal transmitted from their collars triggers warning signs located along 4.8 km of highway the elk frequently cross to reach the northern portion of their range. The 6 signs were standard elk crossing signs (with “ELK X-ING” supplemental plaques) modified with flashing beacons. When the collared elk moved within 402 m of the highway right-of-way, the 360-degree whip antenna detected their proximity and the radio-activated signs began flashing to warn motorists to reduce speed. Since the installation, one elk mortality due to traffic has been documented. The limited data available suggest that the system decreased mortality from 1.7 elk/yr to 0.5 elk/yr. Overview Although there is significant interest and potential in ITS systems, many technical issues must be addressed before they are ready for general use. Critical parameters which affect the feasibility of animal-detection driver-warning systems include: detection zone layout, differentiation of large mammals from smaller objects, duration of warning signal, motorist reaction time, and local climatic conditions (Taskula 1997). Other problems include inherent range limitations, coverage limitation within detection zones, and impacts of background influencing animal-detection efficiency. False detections are a common problem among most of the systems reviewed. One of the leading theories is that multiple detection systems, where two or more detectors must be triggered to verify animal presence, would reduce or eliminate false detections (Taskula 1997). Any ITS system will carry substantial development costs; have the potential for considerable 28 maintenance costs (e.g., aligning and replacing sensors); and multiple systems will cost more than single systems. Until technology is improved to reduce the cost, it is likely that multiple detection systems should only be placed in areas of high crash occurrence. My review of the literature indicates there is a paucity of clear information on the accuracy and reliability for the different sensors available in detecting large mammals. This information can be collected and would aid transportation professionals make more informed decisions relative to deployment of ITS systems. Even if detection technologies work flawlessly, motorists may not respond enough to dynamic signing to significantly reduce the probability of ungulate - vehicle collisions (e.g. Gordon and Anderson 2002). 29 CHAPTER 4 DISCUSSION The increasing demand for faster and more efficient transportation networks has resulted in conflicts with wildlife. To date, the problem of large mammal - vehicle collisions has been underestimated (Groot Bruinderink and Hazebroek 1996). While traffic mortality may not be imperiling most large mammal populations, the increasing danger to motorists and associated property damage costs justify further research and additional mitigation measures (Cook and Daggett 1995, Conover et al. 1995). The identification of one or more proven animal-detection driver-warning systems that successfully mitigate large mammal - vehicle crashes will be directly beneficial to transportation departments worldwide. Toward that end, many countries have begun mitigating the effects or roads on animal populations (Hourdequin 2000). Effective testing of ungulate - vehicle collision mitigation measures has not kept pace with development of alternative methodologies. Many evaluations have been short¬ term tests of commercially developed and marketed products (e.g. Swareflex reflectors). Evaluations that compared large mammal mortality before and after installation yielded confounded results of efficacy (e.g. Pafko and Kovach 1996), because many studies recognized that the large mammal - vehicle collisions vary temporally with respect to topography, habitat, behavior, local population concentrations, and traffic volume. Some early evaluations lacked experimental controls, which precluded robust conclusions about expected collision numbers in the absence of countermeasures (Gilbert 1982). Where 30 experimental controls have been used, they are often merely adjacent roadway sections (e.g. Lehnert and Bissonette 1997). Independence can be compromised by control sections proximity to treatment sections (see Bomford and O’Brien 1990). In such cases, countermeasures in treatment sections may displace ungulates onto control sections, potentially enhancing the treatment’s effect. Groot Bruinderink and Hazebroek (1996) conclude that the surest way to make large mammal - vehicle collision studies more rigorous is to more effectively monitor large mammal mortality statistics, preferably on a national level. However, economic and technical difficulties inherent in monitoring large mammal - vehicle collisions make nation-wide systems unlikely. Rather, I recommend systematic, well-designed tests of different countermeasures in lieu of a national monitoring system. Equal emphasis should be placed on increasing motorist response to animal-detection driver-warning systems. If motorists do not respond by reducing speed or increasing vigilance, the best animal detection system will be ineffective. Rather than attempting to reduce animal - vehicle collisions by focusing on animal-detection driver-warning systems, engineers should consider highway design in an ecological context to reduce the interactions between mammals and vehicles. Transportation and natural resource departments need to work together to identify and protect wildlife movement corridors. This is fundamental to any attempt to mitigate the problem and may itself require a major effort, as broad scale studies of landscape features contributing to large mammal - vehicle collisions are generally lacking (Hubbard et al. 2000). Finder et al. (1999) demonstrated that deer - vehicle accident statistics, along with remotely sensed habitat and highway data might be used to predict high incidence deer - 31 vehicle collision locations. New road designs and reconstruction plans should include wildlife passage at critical locations. Fencing, one of the few effective methods for reducing traffic mortality for large mammals, can be used to direct movement to these passages (Cook and Daggett 1995, Putman 1997). When considering fencing projects, engineers and biologists should realize that barrier fencing profoundly affects animal movement and is not always feasible or acceptable (Clevenger and Waltho 2000, Hourdequin 2000). An example of the potential for agency cooperation is the improvement of U.S. Highway 93 on the Flathead Reservation from Evaro to Poison (90.6 km) in northwest Montana. Recently, a memorandum of agreement was signed by the Confederated Salish and Kootenai Tribes, Montana Department of Transportation, and Federal Highway Administration allowing for the expansion of the highway from 2-lanes to a combination of 2-lanes, 4-lanes, and passing sections. This document further mandated retrofitting the highway with 42 fish and wildlife crossing structures and 23.7 km of ungulate-proof fencing for a total estimated cost of just over $9 million (CSKT et al. 2000). While the costs of many preventive measures are likely high, the benefits resulting from a reduction in accidents to the motoring public and the benefits to wildlife need to be adequately addressed via cost-benefit analysis (Reed et al. 1982). High mitigation costs may only be justified for major roadways or interstates (Putman 1997). For primary roads that combine high speed and high traffic volumes across important wildlife habitat, the most effective approach to large mammal - vehicle mitigation is to combine barrier fencing with wildlife crossing structures to provide large mammal 32 permeability (Groot Bruinderink and Hazebroek 1996). In instances where fencing costs or effects are prohibitive, as on secondary roadways, animal detection-driver warning systems are recommended (Groot Bruinderink and Hazebroek 1996). Here the goal of mitigation may be to delay crossings rather than prevent them (Putman 1997). A monitoring program using track counts or infrared detection technologies to assess large mammal use and mitigation efficacy is critical to the long-term success of any management action (Groot Bruinderink and Hazebroek 1996). Conclusion Clearly, there is no quick fix to the problem of large mammal - vehicle collisions. However this is an exciting time for wildlife and natural resource professionals working on this problem. With the development of new technologies, acknowledgement by transportation agencies of the ecological problems caused by roads, and new funding initiatives such as the Transportation Equity Act for the 21st Century, there is potential for increased implementation of rigorous testing of techniques for reducing animal - vehicle collisions. There is also a greater awareness that countermeasures must be utilized only within the context of a large-scale strategy to reduce problems within road corridors. 33 LITERATURE CITED Aberg, L. 1981. The human factor in game-vehicle accidents: a study of driver’s information acquisition. Acta Universitatis Upsaliensis, Studia Psychologica Upsaliensia 6, Uppsala, Finland. Allen, R. E. and D. R. McCullough. 1976. Deer-car accidents in southern Michigan. Journal of Wildlife Management 40:217-325. American Association of State Highway and Transportation Officials. 1996. Roadside design guide. Task Force for Roadside Safety of the Standing Committee on Highways Subcommittee on Design, Washington D.C., USA. Bashore, T. L., W. M. Tzilkowski, and E. D. Beilis. 1985. Analysis of deer - vehicle collision sites in Pennsylvania. Journal of Wildlife Management 49:769-774. Beilis, E. D, and H. B. Graves. 1971. Deer mortality on a Pennsylvania interstate highway. Journal of Wildlife Management 35:232-237. Bomford, M. and P. H. O’Brien. 1990. Sonic deterrents in animal damage control: a review of device tests and effectiveness. Wildlife Society Bulletin 18:411-422. Calvo, R. N. and N. J. Silvy. 1996. Key deer mortality, U.S. 1 in the Florida Keys. Pages unnumbered in G. Evink, P. Garrett, D. Zeigler, and J. Berry editors. Proc. of the Transportation Related Wildlife Mortality Seminar. FL-ER-58-98, Florida Department of Transportation, Tallahassee, FL, USA. Carbaugh, B. T. 1970. Activity and behavior of white-tailed deer (Odocoileus virginianus) along an interstate highway in a forest region of Pennsylvania. PhD dissertation. Pennsylvania State University, University Park, Pennsylvania, USA. Carbaugh, B. J., J. P. Vaughan, E. D. Beilis, and H. B. Graves. 1975. Distribution and activity of white-tailed deer along an interstate highway. Journal of Wildlife Management 39:570-581. Case, R. M. 1978. Interstate highway road-killed animals: a data source for biologists. Wildlife Society Bulletin 6:8-13. Clevenger, A. P. 1998. Permeability of the trans-Canada highway to wildlife in Banff National Park: the importance of crossing structures and factors influencing their effectiveness. Pages 109-119 in G. Evink, P. Garrett, D. Zeigler, and J. Berry editors. Proceedings of the International Conference on Wildlife Ecology and 34 Transportation. FL-ER-68-98, Florida Department of Transportation, Tallahassee, FL, USA. Clevenger, A. P., and N. Waltho. 2000. Factors influencing the effectiveness of wildlife underpasses in Banff National Park, Alberta, Canada. Conservation Biology 14:47-56. Clevenger, A. P., B. Chruszcz, and K. E. Gunson. 2001. Highway mitigation fencing reduces wildlife-vehicle collisions. Wildlife Society Bulletin 29:646-653. Confederated Salish and Kootenai Tribes, Federal Highway Administration, and Montana Department of Transportation. 2000. Memorandum of agreement: US 93 Evaro to Poison. Skillings-Connolly, Inc., Missoula, MT, USA. Conover, M. R., W. C. Pitt, K. K Kessler, T. J. DuBow, and W. A. Sanborn. 1995. Review of human injuries, illnesses, and economic losses caused by wildlife in the United States. Wildlife Society Bulletin. 23:407-414. Conover, M. R. 1997. Monetary and intangible valuation of deer in the United States. Wildlife Society Bulletin 25:298-305. Cook, K. E., and P. M. Daggett. 1995. Highway roadkill: associated issues of safety and impacts on highway ecotones. Task force on natural resources, Transportation Research Board. National Research Council, Washington, D.C., USA. Falk, N. W., H. B. Graves, and E. D. Beilis. 1978. Highway right-of-way fences as deer deterrents. Journal of Wildlife Management 42:646-650. Feldhammer, G. A., J. E.Gates, D. M. Harman, A. J. Loranger, and K. R. Dixon. 1986. Effects of interstate highway fencing on white-tailed deer activity. Journal of Wildlife Management. 50:497-503. Finder, R. A., J. L. Roseberry, and A. Woolf. 1999. Site and landscape conditions at white-tailed deer/vehicle collision locations in Illinois. Landscape and Urban Planning 44:77-85. Ford, S. G. and S. L. Villa. 1993. Reflector use and the effect they have on the number of mule deer killed on California highways. California Department of Transportation, Rep. No. FHWA/CA/PD-94/01. Sacramento, CA, USA. Foreman, R. T. T. 2000. Estimate of the area affected ecologically by the road system in the United States. Conservation Biology 14:31-35. 35 Foreman, R. T. T., and R. D. Deblinger. 2000. The ecological road-effect zone of a Massachusetts (U.S.A.) suburban highway. Conservation Biology 14:36-46. Foreman, R. T. T. and L. E. Alexander. 1998. Roads and their major ecological effects. Annual Review of Ecological Systems 29:207-231. Foreman, R. T. T. and A. M. Hesperger. 1996. Road ecology and road density in different landscapes, with international planning and mitigation solutions. Pages unnumbered in G. Evink, P. Garrett, D. Zeigler, and J. Berry editors. Proceedings of the Transportation Related Wildlife Mortality Seminar. FL-ER-58-98, Florida Department of Transportation, Tallahassee, FL, USA. Foster, M.L., and S.R. Humphrey. 1995. Use of highway underpasses by Florida panthers and other wildlife. Wildlife Society Bulletin 23:95-100. Fraser, D. 1979. Sightings of moose, deer, and bears on roads in northern Ontario. Wildlife Society Bulletin 7:181-184. Fraser, D., and E. R. Thomas. 1982. Moose - vehicle accidents in Ontario: relation to highway salt. Wildlife Society Bulletin 10:261-265. Fraser, D., and H. Hristienko. 1982. Moose - vehicle accidents in Ontario: a repugnant solution? Wildlife Society Bulletin 10:266-270. Garrett, L. C., and G. A. Conway. 1999. Characteristics of moose - vehicle collisions in Anchorage, Alaska, 1991-1995. Journal of Safety Research 30:219-223. Gilbert, J. R. 1982. Evaluation of deer mirrors for reducing deer/vehicle collisions. Federal Highway Administration/RD-82/061. Washington, D.C., USA. Gladfelter, L. 1984. Effect of wildlife highway warning reflectors on deer - vehicle accidents. Iowa Department of Transportation, Iowa Highway Research Board Project HR-210. Iowa, USA. Gloyne, C. G. and A. P. Clevenger. 2001. Cougar Puma concolor use of wildlife crossing structures on the trans-Canada highway in Banff National Park, Alberta. Wildlife Biology 7:117-124. Goodwin, G. A. and A. L. Ward. 1976. Mule deer mortality on interstate 80 in Wyoming: causes, patterns, and recommendations. U.S. Department of Agriculture, Forest Service Research Note RM-332. Rocky Mountain Forest and Range Experimental Station, Fort Collins, Colorado, USA. 36 Gordon, K. M, and S. H. Anderson. 2002. Motorist response to a deer-sensing warning system in western Wyoming. Pages 549-558 in Proceedings of the International Conference on Ecology and Transportation, Keystone, CO, September 24-28, 2001. Center for Transportation and the Environment, North Carolina State University, Raleigh, NC, USA. Groot Bruinderink, G. W. T. A., and E. Hazebroek. 1996. Ungulate traffic collisions in Europe. Conservation Biology 10:1059-1067. Gunther, K.A., M.J. Biel, and H.L. Robison. 1998. Factors influencing the frequency of road-killed wildlife in Yellowstone National Park. Pages 32-42 in G. Evink, P. Garrett, D. Zeigler, and J. Berry editors. Proceedings of the International Conference on Wildlife Ecology and Transportation. FL-ER-68-98, Florida Department of Transportation, Tallahassee, FL, USA. Haikonen, H., and H. Summala. 2001. Deer - vehicle crashes: extensive peak one hour after sunset. American Journal of Preventative Medicine 21:209-213. Haskell, D. G. 2000. Effects of forest roads on macroinvertebrate soil fauna of the southern Appalachian Mountains. Conservation Biology 14:57-63. Haugen, A. O. 1944. Highway mortality of wildlife in southern Michigan. Journal of Mammalogy 25:177-184. Hourdequin, M. 2000. Special section: ecological effects of roads. Conservation Biology 14:16-17. Hubbard, M. W., B. J. Danielson, and R. A. Schmitz. 2000. Factors influencing the location of deer - vehicle accidents in Iowa. Journal of Wildlife Management 64:707-713. Hughes, W. E., A. R. Saremi and J. F. Paniati. 1996. Vehicle-animal crashes: an increasing safety problem. Institute of Transportation Engineers Journal 66:24- 28. Iverson, A. L. and L. R. Iverson. 1999. Spatial and temporal trends of deer harvest and deer-vehicle accidents in Ohio. The Ohio Journal of Science 99:84-94. Jahn, L. R. 1959. Highway mortality as an index of deer-population change. Journal of Wildlife Management 12:187-196. Jaren, V., R. Andersen, M. Ulleberg, P. H. Pedersen, and B. Wiseth. 1991. Moose - train collisions: the effects of vegetation removal with a cost-benefit analysis. Alces 27:93-99. 37 Joyce, T. L. and S. P. Mahoney. 2001. Spatial and temporal distributions of moose - vehicle collisions in Newfoundland. Wildlife Society Bulletin 29:281-291. Kaji, K. 1996. The distribution of sika deer and present status of cervid-vehicle accidents in Hokkaido. Pages 50-71 in N. Othtaisha, editor. Proceedings of the Symposium on Wildlife-Traffic Collisions, January 23-28. University of Sapporo, Japan. Kasul, R. L. 1976. Habitat factors associated with mortality of southern Michigan wildlife on an interstate highway. M.S. thesis. Michigan State University, East Lansing, Michigan, USA. Koehne, J.L. 1991. Wildlife-related traffic accidents: a comparison of existing and proposed mitigation measures. Prepared for the County Board of Sddermanland, Nykoping, Sweden. Kress, M. J. 1980. The effects of habitat on the distribution of white-tailed deer (Odocoileus virginianus) along a Pennsylvania interstate highway. PhD thesis. Pennsylvania State University, University Park, Pennsylvania, USA. Lavsund, S. and F. Sandegren. 1991. Moose-vehicle relations in Sweden: a review. Alces 27:118-126. Leedy, D. L. 1975. Highway-wildlife relationships, Vol. 1. A state of the art report. Federal Highway Administration/RD-76-4. Washington, D.C., USA. Lehnert, M. E. and J. A. Bissonette. 1997. Effectiveness of highway crosswalk structures at reducing deer - vehicle collisions. Wildlife Society Bulletin 25:809- 818. Leopold, A. 1933. Game management. Charles Scribner’s Sons, New York, NY, USA. Loughry, W. J. and C. M. McDonough. 1995. Are road kills valid indicators of armadillo population structure? American Midland Naturalist 135:53-59. Ludwig, J. and T. Bremicker. 1983. Evaluation of 2.4-m fences and one-way gates for reducing deer - vehicle collisions in Minnesota. Transportation Research Record 913:19-22. Mansfield, T. M. and B. D. Miller. 1975. Highway deer-kill district 02 regional study. Caltrans Internal Report, Sacramento, California, USA. 38 McCaffrey, K. R. 1973. Road-kills show trends in Wisconsin deer populations. Journal of Wildlife Management 37:212-216. Messmer, T. A. 2000. The emergence of human-wildlife conflict management: turning challenges into opportunities. International Biodeterioration and Biodegredation 45:97-102. Messmer, T. A., C. W. Hendricks, and P. W. Klimack. 1999. The effect of temporary signing on reducing deer - vehicle collisions during seasonal mule deer migrations. Internal Report, Jack H. Berryman Institute of Wildlife Damage Management, Utah State University, Logan, Utah, USA.. Noss, R. F., and A. Y. Cooperrider. 1994. Saving nature’s legacy. Island Press, Washington, D.C., USA. Pafko, F. and B. Kovach. 1996. Minnesota experience with deer reflectors. Pages unnumbered in G. Evink, P. Garrett, D. Zeigler, and J. Berry editors. Proceedings of the Transportation Related Wildlife Mortality Seminar. FL-ER-58-98, Florida Department of Transportation, Tallahassee, FL, USA. Peek, F. W., and E. D. Beilis. 1969. Deer movements and behavior along an interstate highway. Highway Research News 36:36-42. Pojar, T. M., R. A. Prosence, D. F. Reed, and T. N. Woodard. 1975. Effectiveness of a lighted, animated deer crossing sign. Journal of Wildlife Management 39:87-91. Puglisi, M. J., J. S. Lindzey, and E. D. Beilis. 1974. Factors associated with highway mortality of white-tailed deer. Journal of Wildlife Management 38:799-807. Putman, R. J. 1997. Deer and road traffic accidents: options for management. Journal of Environmental Management 51:43-57. Ray, M. H. 1998. Safety and cost-effectiveness of clearzones in rockcuts. Journal of Safety Research 29:285-293. Reed, D. F. 1981. Mule deer behavior at a highway underpass exit. Journal of Wildlife Management 45:542-543. Reed, D. F., and T. N. Woodard. 1981. Effectiveness of highway lighting in reducing deer - vehicle collisions. Journal of Wildlife Management 45:721-726. Reed, D. F., T. M. Pojar, and T. N. Woodward. 1974. Use of one-way gates by mule deer. Journal of Wildlife Management 38:9-15. 39 Reed, D. F., T. N. Woodard, and T. M. Pojar. 1975. Behavioral response of mule deer to a highway underpass. Journal of Wildlife Management 39:361-367. Reed, D. F., T. D. I. Beck, and T. N. Woodard. 1982. Methods of reducing deer - vehicle accidents: benefit-cost analysis. Wildlife Society Bulletin 10:349-354. Reeve, A. F. and S. H. Anderson. 1993. Ineffectiveness of Swareflex reflectors at reducing deer - vehicle collisions. Wildlife Society Bulletin 21:127-132. Reilly, R.E., and H.E. Green. 1974. Deer mortality on a Michigan interstate highway. Journal of Wildlife Management 38:16-19. Rodriguez, A., G. Crema, and M. Delibes. 1996. Use of non-wildlife passages across a high speed railway by terrestrial vertebrates. Journal of Applied Ecology 33:1527-1540. Romin, L. A. 1994. Factors associated with the highway mortality of mule deer at Jordanelle Reservoir, Utah. M.S. Thesis, Utah State University, Logan, Utah, USA. Romin, L. A., and J. A. Bissonette. 1996. Deer - vehicle collisions: status of state monitoring activities and mitigation efforts. Wildlife Society Bulletin 24:276- 283. Romin, L. A. and L. B. Dalton. 1992. Lack of response by mule deer to wildlife warning whistles. Wildlife Society Bulletin 20:382-384. Rost, G. R. and J. A. Bailey. 1979. Distribution of mule deer and elk in relation to roads. Journal of Wildlife Management 43:634-641. Rowland, M. M., M. J. Wisdom, B. K. Johnson, and J. G. Kie. 2000. Elk distribution and modeling in relation to roads. Journal of Wildlife Management 64:672-684 Sage, Jr., R. W., W. C. Tierson, G. F. Mattfeld, and D. F. Behrend. 1983. White-tailed deer visibility and behavior along forest roads. Journal of Wildlife Management 47:940-953. Schafer, J. A. and S. T. Penland. 1985. Effectiveness of Swareflex reflectors in reducing deer - vehicle accidents. Journal of Wildlife Management 46:774-776. Sicuranza, L. P. 1979. An ecological study of motor vehicle-deer accidents in southern Michigan. M.S. thesis. Michigan State University, East Lansing, Michigan, USA. 40 Singer, F. J. and J. L. Doherty. 1985. Managing mountain goats at a highway crossing. Wildlife Society Bulletin 13:469-477. Stoner, D. 1925. The toll of the automobile. Science 61:56-57. Taskula, K. 1997. The moose ahead. Traffic Technology International 42:170-173. Transportation Research Board. 1998. Managing speed: review of current practice for setting and enforcing speed limits. Committee for Guidance on Setting and Enforcing Speed Limits, National Research Council, National Academy Press, Washington, D.C., USA. Trombulak, S. C., and C. A. Frissell. 2000. Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14:18-30. Ujvari, M., H. J. Baagoe, and A. B. Madsen. 1998. Effectiveness of wildlife warning reflectors in reducing deer - vehicle collisions: a behavioral study. Journal of Wildlife Management 62:1094-1099. Vaughn, J. P. 1970. Influence of environment on the activity and behavior of white¬ tailed deer (Odocoileus virginianus). PhD dissertation. Pennsylvania State University, University Park, Pennsylvania, USA. Ward, L. A. 1982. Mule deer behavior in relation to fencing and underpasses on interstate 80 in Wyoming. Transportation Research Record 859:8-13. Waring, G. H., J. L. Griffis, and M. E. Vaughn. 1991. White-tailed deer roadside behavior, wildlife warning reflectors, and highway mortality. Applied Animal Behaviour Science 29:215-223. Weaver, J. L., P. C. Paquet, and L. F. Ruggiero. 1996. Resilience and conservation of large carnivores in the Rocky Mountains. Conservation Biology 10:964-976. Wood, P. and M. L. Wolfe. 1988. Intercept feeding as a means of reducing deer - vehicle collisions. Wildlife Society Bulletin 16:376-380. Woodard, T. N., D. F. Reed, and T. M. Pojar. 1973. Effectiveness of Swareflex wildlife warning reflectors in reducing deer - vehicle accidents. Internal Report, Colorado Division of Wildlife, Denver, Colorado, USA. Yanes, M., J. M. Velasco, and F. Suarez. 1995. Permeability of roads and railways to vertebrates: the importance of culverts. Biological Conservation 71:217-222. 41 Zacks, J. L. 1986. Do white-tailed deer avoid red? An evaluation of the premise underlying the design of Swareflex wildlife reflectors. Transportation Research Record 1075:35-43.