animal-vehicle accident analysis - Utah Government Digital Library
Report No. UT-03.31
ANIMAL-VEHICLE ACCIDENT ANALYSIS
Prepared For: Utah Department of Transportation Research and Development Division Submitted By: University of Utah
Authored By: Dr. Joseph Perrin Rodrigo Disegni
November 2003
UDOT RESEARCH & DEVELOPMENT REPORT ABSTRACT
1. Report No.
2. Government Accession No.
UT-03.31
4. Title and Subtitle
5. Report Date
Safety Benefits of UDOT Highway Program, Animal-Vehicle Accident Analysis 7. Author(s)
3. Recipient's Catalog No.
November 2003 6. Performing Organization Code 8. Performing Organization Report No.
Perrin, Joseph
9. Performing Organization Name and Address
10. Work Unit No.
University of Utah 122 South Central Campus Drive, Room 104 SLC, UT 84112
11. Contract No.
12. Sponsoring Agency Name and Address
13. Type of Report and Period Covered
Utah Department of Transportation 4501 South 2700 West Box 148410 SLC, Utah 84114
Research – August to November 2003 14. Sponsoring Agency Code
UDOT - 039178
15. Supplementary Notes
16. Abstract
Vehicle-animal accidents represented 4.6% of US automobile accidents with more than 1.5 million accidents a year, 150 deaths and $1.1 billion in vehicle damage. Animal related accidents in Utah represent 1.2% of statewide automobile accidents. In 2001 there were 2,688 vehicle-animal collisions, including 3 death, and 235 injuries. In Utah, animal related accidents are subdivided in wild and domestic animals. Domestic animals include livestock, such as cows and sheep or horses. Wild animal most often refer to deer, elk and moose. Using 10-years of statewide accident information, the problem locations are identified and a comparison between the domestic and wild animal accidents based on severity is examined. The accident analysis determined that while the domestic animal accidents represent only16% of the animal-vehicle accidents, they are more severe than wild animal accidents. Domestic animal accidents result in injury 23% of the time while wild animal accidents results in injury only 7% of the time. When a motorcycle is involved, it was found that 94% of the animal-motorcycle related accidents resulted in injury compared with only 11% of the non-motorcycle-animal accidents. Overall, there is a 7.9 times greater chance of a fatality with domestic animal accidents compared with wild animals. This is attributed to the height and weight of domestic animal relative to the common wild animal. While many countermeasures are attempted such as whistles and reflectors, the principal countermeasure to control animal related accidents has been the use of fences along the roads. The 4-foot high right-of-way fences are effective for the domestic animals but the wildlife animals requires the higher 8-foot fences since deer can easily circumvent the 4-foot fence heights. Alternative countermeasures such as one-way deer gates and ecopassages are also reducing wild animal hits. This study utilized UDOTs CARS accident database to identify the vehicle-animal crash problem in Utah. The study describes the extent of the problem; some literature on various countermeasures used throughout the world, and finally identifies the most dangerous section of routes between the years 1999-2001 in terms of the accidents per mile for the wild and domestic animals. 17. Key Words Animals - Accidents, wild animals, domestic animals
18. Distribution Statement
19. Security Classification (of this report)
21. No. of Pages
20. Security Classification (of this page)
35
22. Price
EXECUTIVE SUMMARY Vehicle-animal accidents represented 4.6% of US automobile accidents with more than 1.5 million accidents a year, 150 deaths and $1.1 billion in vehicle damage. Animal related accidents in Utah represent 1.2% of statewide automobile accidents. In 2001 there were 2,688 vehicle-animal collisions, including 3 death, and 235 injuries. In Utah, animal related accidents are subdivided in wild and domestic animals. Domestic animals include livestock, such as cows and sheep or horses. Wild animal most often refer to deer, elk and moose. Using 10-years of statewide accident information, the problem locations are identified and a comparison between the domestic and wild animal accidents based on severity is examined. The accident analysis determined that while the domestic animal accidents represent only16% of the animal-vehicle accidents, they are more severe than wild animal accidents. Domestic animal accidents result in injury 23% of the time while wild animal accidents results in injury only 7% of the time. When a motorcycle is involved, it was found that 94% of the animal-motorcycle related accidents resulted in injury compared with only 11% of the non-motorcycle-animal accidents. Overall, there is a 7.9 times greater chance of a fatality with domestic animal accidents compared with wild animals. This is attributed to the height and weight of domestic animal relative to the common wild animal. While many countermeasures are attempted such as whistles and reflectors, the principal countermeasure to control animal related accidents has been the use of fences along the roads. The 4-foot high right-of-way fences are effective for the domestic animals but the wildlife animals requires the higher 8-foot fences since deer can easily circumvent the 4-foot fence heights. Alternative countermeasures such as one-way deer gates and ecopassages are also reducing wild animal hits. This study utilized UDOTs CARS accident database to identify the vehicle-animal crash problem in Utah. The study describes the extent of the problem; some literature on various countermeasures used throughout the world, and finally identifies the most dangerous section of routes between the years 1999-2001 in terms of the accidents per mile for the wild and domestic animals.
i
TABLE OF CONTENTS
List of Figures ................................................................................................ iii List of Tables .................................................................................................iv 1
Introduction............................................................................................. 1
2
Literature Review ................................................................................... 1
3
Scope ..................................................................................................... 5
4
Methodology........................................................................................... 6
5
Results ................................................................................................... 8
6
Recommendations ............................................................................... 27
7
Future Work.......................................................................................... 28
8
Appendix .............................................................................................. 30
References .................................................................................................. 34
ii
LIST OF FIGURES Figure 1: Domestic Animal Accidents Trend........................................................9 Figure 2: Wild Animal Accidents Trend..............................................................10 Figure 3: Fatality Rate for Domestic and Wild Animal Accidents .......................11 Figure 4: Vehicle-Animal Accidents by Time of Day ..........................................12 Figure 5: Vehicle-Animal Accidents by Month of the Year .................................13 Figure 6: Comparison Between Domestic And Wild Animal Severity ................16 Figure 7: Motorcycle-Animal Accident Severity .................................................17 Figure 8: Passenger Car-Animal Accident Severity...........................................18 Figure 9: Routes With High Wild Animal Accident Rate ....................................19 Figure 10: Routes With High Domestic Animal Accident Rate...........................20
iii
LIST OF TABLES Table 1: Accident Classification...........................................................................7 Table 2: Percentage of Multi-Vehicle Accidents Between 1992 and 2001.........14 Table 3: Animal Accident Costs (1992-2001) ....................................................15 Table 4: Accident Severity Classification and Description .................................15 Table 5: Wild Animal Accidents Classification ...................................................19 Table 6: Domestic Animal Accident Category....................................................20 Table 7: Wild Animals High Accident Sections 1999-2001 ................................22 Table 8: Classification of Sections Wild Animals ...............................................22 Table 9: Domestic Animals High Accident Sections 1999-2001 ........................24 Table 10: Classification of Section Domestic Animals .......................................24 Table 11: Accident Rate of Sections With Fences (1999-2001) ........................26
iv
1
INTRODUCTION
The following report documents the animal-vehicle accidents problem in Utah. While the more general purpose of this study is to identify the value of the webinterface to the Centralized Accident Record Systems (CARS) database, the objectives to this study are to utilize the CARS system to identify general trends over the past 10-years and identify the “hot spots” throughout the state for the past three years. The accidents are segregated into wild and domestic and thus the analysis distinguishes between these two types of animal-vehicle accident.
2
LITERATURE REVIEW
The literature on vehicle-animal accidents begins in the 1980s. Past literature has focused on defining the problem, determining the economic impacts of the accidents for both the driver and the wildlife, and evaluating the effectiveness of the devices that keep wildlife off roads. As vehicle travel has increased and intercity travel expanded, these routes and frequency of traffic conflicts with wildlife path increasing the interaction between vehicles and animals. In 1980 the number of accidents involving deer in America’s highways was estimated at 200,000 collisions (Schafer et al. 1985). This number is estimated to be conservative, as many incidents are not reported because of the low severity. The latest number reported in 2001 shows a considerable increase at 292,000 accidents, which equates to 4.6% of the total accidents in 2001. During 2001 there were also 165 fatalities (0.4% of all fatalities) and 19,000 injuries (0.9% of all injuries) as a consequence of vehicle-animal collisions (Traffic Safety Facts 2001). In terms of the crash circumstances, most of the vehicle-animal crashes occurred on rural roads compared to urban roads, and during nighttime between 7 and 10 pm (Hughes, Sremi, and Paniati 1996). Researchers have also found that
1
October and November are the months of the year when the number of accidents peak, as those months represent the mating season for deer (Hughes et al.). This last finding, however, contradicts the conclusions of a similar study where researchers in Washington State concluded that most of the animal related accident occurred during the summer time versus late fall (Schafer et al. 1985). We estimate that there is likely a relationship between traffic volume variations and vacation travel related to the accident rates in Washington as summer is a much higher travel time within the typically rainy state. The economic value involved in vehicle-animal incidents considers both the driver and the animal costs. It was estimated in 1977 that the driver’s cost of an accident was about $730, which equates to $2,160 in 2003 dollars. The economic lost for a deer however is more difficult to quantify, but was estimated in 1976 at $350 dollars, which represents $1,100 in 2003 dollars (Schafer et al.). Other researchers, however, estimated the dollar value of a deer at twice this figure based on hunting expenditures alone (Hartman, and Norman). It can be concluded that whatever the economic value of deer is, the overall problem should be, if under reporting is considered around one billion dollars annually. Several devices have been used in the past to prevent wildlife from being on the roads. Some of the devices commonly used are: warning signs, fences, wildlife reflectors, and lately ecopassages (underpasses). The evaluation of these devices has reached different levels of success. However, an issue that is relevant to all the studies that evaluates the performances of these devices is the difficulty to control deer population during the study. Therefore a decrease in accident rate on a fenced road may be explained by a decrease in deer population. The cheapest device used to prevent vehicle-animal accidents has been the placement of wildlife warning signs. However, it has been reported that they are the least effective and its indiscriminate use has lowered the credibility of the sign. A study concluded that 60% of the drivers will not even notice the sign while driving (Ministry of BC, Hughes).
2
The use of fences has been found to be successful at keeping animals off the road. The ministry of Transportation and highways in BC, Canada reported that the use of 2.4 meter (7.8 feet) high fence has been very efficient at keeping wildlife off the roads. They report that this higher fencing is 97-99% effective in preventing vehicle-animal accidents. Some of the problems of fencing are that wildlife is restricted to cross roads and in the case of fence penetration wildlife would be trapped in the road. Restricting wildlife from crossing roads can have serious consequences if food and water resources are across the road. Fence penetration is more common when the 1.2 meter high fence is used instead of the 2.4 meter high fence. The use of fencing along with under and overpasses as well as one-way gates has been reported as a better way to deal with the problem. Ludwig 1983 reports benefit cost ratios of 3.61 and reduction of deer accidents of 60% and 93% in two sites where fencing and one-way gates were installed. Wildlife reflectors are a different type of technology that has been used in the past with a wide range of success. The main advantage of reflectors is that they are cheaper and less intrusive than fencing. The theory behind reflectors is that the headlight of a vehicle is reflected into a red light that produces a freezing response in deer (Schafer et al.). Although the theory has been widely discussed and researched there are still some controversies regarding the effect that red light may have on deer. A study by Zacks (1986) tested the response of deer to the red color of the reflectors. The study found that the behavior of deer was the same under red reflectors, white reflectors and no reflectors. The conclusion of the study was that red reflectors do not modify deer’s behavior. The researchers suggested that the effect of reflectors might be on the drivers’ behavior rather than on the deer. Nonetheless success and failure of reflectors have been documented in two studies with two opposite outcomes. A study by Schafer et al., which represent one of the best studies in terms of controlling for deer population, concluded that wildlife reflectors decrease the number of vehicleanimal incidents over the study period. Another study by the Ministry of
3
Transportation and highways in British Columbia, Canada (2003) concluded that after the installation of reflectors the number of accidents involving deer increased, deer population however was not controlled in this study. Although both studies differ in terms of the methodology used, they still represent valuable research in terms of the effectiveness of wildlife reflectors. The newest device that has been used and evaluated is the ecopassages. Ecopassages consist of a wall or fence that keeps animals off the road and a set of underground passes, usually culverts, that allow animals to cross the road underneath (Dockstader and Southall 2003). The use of ecopassages in Florida reported a decreased in the number of wildlife killed of 64.2% compared to the before period. The study also reported an increase in the wildlife species that were using the underpasses to cross the road. The literature shows that the use of before and after comparison to evaluate the effectiveness of vehicle-animal mitigations leads to misleading conclusions if deer population is not controlled. The change in migration patterns and animal behavior from year to year and even from month to month make the comparison of before and after studies, where wildlife population is not controlled, difficult from a statistical scientifically standpoint, much like most accident trend analysis. .
4
3
SCOPE
3.1
Problem Definition
The vehicle-animal collision problem in Utah is of similar characteristics to that faced by other states in the US. About 5.1% of the total accidents in 2001 involved either wild or domestic animals. There were a total of 288 vehicleanimal accidents in 2001 including 3 deaths (1.2% of all fatalities) and 235 injuries (1.2% of all injuries) (Traffic Safety Facts 2001). This study identifies a methodology to determine high accident spots through the use of the Centralized Accident Record System (CARS) database. The study focuses on accidents that involve wild and domestic animals. Wild animal accidents usually represent accidents involving deer, moose, bear, wolf, and other species that are considered not domestic. Domestic animal accidents usually involve cows, sheep and horses.
3.2
Research Objectives The main objectives of this study are: •
To use the UDOT CARS database Web site to perform a safety analysis of accidents involving wild and domestic animals.
•
To define a methodology that identifies road sections that have high accident, and should be further analyzed.
5
4
METHODOLOGY
The primary result identifies routes by milepost where high animal-vehicle accidents occurr. The data analysis of this study has been divided in two parts. The first one identifies routes among all the state routes, which have a high accident per mile rate. The second one identifies route sections where accident numbers are particularly high. This study estimates high accident rate and numbers based on expected value analysis. Expected value analysis assumes that the occurrence of accidents follows a normal distribution. Under the expected value analysis the mean and the standard deviation determine the shape that the normal distribution will have. Based on the normal distribution and a 95% confidence level, it is defined that values higher than the average plus 1.96 standard deviation are consider being abnormal. Abnormal locations would have high accident rate and thus further studies should be performed to determine the reasons for the abnormal accident rate or accident number. The following two sections describe the methodology used to determine the most dangerous routes and sections for the wild and domestic animal accidents.
4.1
Estimation of the Most Dangerous Routes
This study includes all the state routes in Utah, excluding the routes that are shorter than 1 mile in length. These routes are excluded from the study because the inclusion of short routes can skew the results, as their accident per mile rate can be high with just the occurrence of 1 accident in the section. The study period is defined as 3 years between 1999 and 2001. The accident rate and average number of accidents per mile are selected as the MOE to estimate the routes with a high number of accidents. The accident per mile rate is selected because it allows the comparison between long routes that may have a large number of accidents with short routes that may have a low 6
number of accidents. This study performs a separate analysis for accidents involving domestic and wild animals, as they can be different in terms of severity and frequency. Based on the average number of accidents and the standard deviation each route can be classified in terms of a Low, Moderate, and High accident rate scale. Table 1 shows the limits of the different categories. Based on the normal distribution the mean should include 50% of the population, the mean plus one standard deviation should include 68% of the population, and finally the mean plus 1.96 standard deviation includes 95% of the population. Therefore accident rates that are higher than the 95% of the population are considered abnormal (High).
Table 1: Accident Classification Category
4.2
Limit
Low
Mean
Moderate
Mean + Standard Deviation
High
Mean +1.96*Standard Deviation
Estimation of Route Sections With High Accident Numbers
All the routes are analyzed in terms of the number of accidents per 1-mile segment increment. As several sections within a route do not have accidents during the 3-year period, all the analyses are referred to sections that have at least 1 accident during the 3-year study period (1999-2001). Using a similar approach as the one used in the high route identification, the mean and standard deviation of the 1-mile sections with more than 1 accident for all the routes are calculated. Based on the definition shown in Table 1 all the sections with an accident number higher than the mean plus 1.96 standard deviation are define as high.
7
The identification of high accidents section would indicate places where maintenance or installation of new fence should be considered. 5
RESULTS
The findings of this study are included with a discussion of the results. This section has been organized in two parts. It starts with a description of the trend of animal accidents in Utah between the 10-years from 1992 and 2001. The second part provides a more refined analysis of the animal incidents between the years 1999 and 2001 for all the state routes in Utah excluding the routes that are shorter than 1 mile.
5.1
Vehicle-Animal Accident Trends between 1992 and 2001
Figure 1 shows the accident trend of crashes involving domestic animals. The figure shows that the number of accidents has increased over the years. The number of accidents increased from 340 accidents in 1992 to over 460 accidents in 2001.
8
Number of Accidents
600 500 400 300 200 100 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Year
Figure 1: Domestic Animal Accidents Trend
Figure 2 presents the number of accident involving wild animals. The number of accidents has remained fairly constant over the years. It can be concluded that wild animal accident are much more frequent than domestic animal accidents. In average the number of wild accidents are 5 times the number of domestic animal accidents. A reason for this may be the higher population that wild animals represents compared to domestic animals. Another reason may be the fact that domestic animals are usually in a fenced area, and easier contained with smaller right-of-way fences compared to wild animals that may be running freely.
9
Number of Accidents
3000 2500 2000 1500 1000 500 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Year
Figure 2: Wild Animal Accidents Trend
Figure 3 shows the fatality rate for domestic and wild animal accidents. Please note that during the years 1993 and 1995 there were no fatal accidents in both categories. The data indicates that the fatality rate for domestic animal accidents is almost 8 times higher than for wild animals. If we consider that the domestic animal accidents represents fewer accidents compared to domestic animal accidents, then we can conclude that the chances of having an injury or even a fatal accident is higher for domestic animals than wild animals. A reason for this is the type of animals included under each category. While a wild animal represents usually a deer or a moose, a domestic animal many times is a cow or horse. These animals are very different in terms of weight. A cow can easily be more than 5 times heavier than a deer; in average a cow weighs 1,000 pounds while a deer can weight 140 pounds (9,10).
10
10 9
Fatality Rate
8 Wild Domestic
7 6 5 4 3 2 1 0 1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Year
Figure 3: Fatality Rate for Domestic and Wild Animal Accidents
5.2
Vehicle-Animal Accident Analysis between 1999 and 2001
Assessing when and where the accidents occur sheds insight into the circumstances surrounding the accidents. Figure 4 shows the percentage of animal related accident occurrence by time of day. The figure shows that most of the accidents occur during darkness between 6 PM and 10 PM. There is also another peak of accidents that happen early in the morning at 6 AM, but representing a smaller proportion of accidents.
11
% of Total Accidents
14 12 10 8 6 4 2 0 0
2
4
6
8
10
12
14
16
18
20
22
24
Time of Day
Figure 4: Vehicle-Animal Accidents by Time of Day
Figure 5 shows the percentage of accident related animals for the different month in the year. It is concluded from the figure that animal related accidents peak during the fall season. October, November, and December are the months with the highest percentage of animal related accident. It is described in the literature that this period of the year (fall) coincides with the mating season of deer and migration as weather changes and more animals move from higher elevations to the lower elevations where there is also a higher likelihood of vehicle–animal interaction.
12
14 % of Total Accidents
12 10 8 6 4 2 0 Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov Dec
Month
Figure 5: Vehicle-Animal Accidents by Month of the Year
Most of the animal related accidents involve only one vehicle in the collision. However there are cases when there is more than one vehicle involved. Table 2 shows the percentage of multi-vehicle accidents for wild and domestic animal incidents. It is concluded that domestic animals have a higher percentage of accidents involving 2 or more vehicles. A reason for this can be the fact that domestic animals are heavier compared to wild animals and result in more severe accidents. Therefore after a car has hit a domestic animal, the car may be forced to stop abruptly causing more vehicles to get involved in the incident or the body of the animal may still be blocking the road, therefore increasing the probability of more vehicles involved.
13
Table 2: Percentage of Multi-Vehicle Accidents Between 1992 and 2001 Vehicles Involved
Wild
Domestic
1
99.5%
96.5%
2
0.5%
3.2%
3
0.0%
0.2%
4+
0.0%
0.1%
The economic cost involved in animal related accidents is difficult to determine since the cost of wildlife is difficult to quantify. Past studies have attempted to estimate the cost involved with the lost of wildlife reaching widely different numbers. This study does not attempt to estimate the economic loss of wildlife, but recognizes that the cost of wildlife constitute an important factor in the economic evaluation of animal related accidents. In terms of the driver’s cost involved in animal related accidents, this study uses the current values of accidents based on severity as defined by Federal Highways. Table 3 presents the estimated cost for accidents according to the severity of the incident. The cost ranges from $2,300 dollars for an incident with no injuries (severity 1) up to 3 million dollars for a fatal accident (severity 5). Table 4 shows a description of the different severity types, categories 1 through 5, and the characteristics of the accident. Table 3 also shows the number of wild and domestic animal accidents between the years 1992 and 2001 for each severity type. Based on the cost of each type of accident and the number of accidents during the 10-year period, this study estimates that the overall cost involved in animal related accidents is approximately $469 million dollars or an average $46 million per year.
14
Table 3: Animal Accident Costs (1992-2001) Wild Cost per
Domestic Cost
Cost
Total
Severity Accident Accidents Thousand Accidents Thousand Thousand 1
2,300
20,629
$47,446
3,367
$7,744
$55,190
2
6,000
582
$3,492
328
$1,968
$5,460
3
45,000
418
$18,810
294
$13,230
$32,040
4
565,000
5
3,000,000 Total
293 $165,545 10
242 $136,730 $302,275
$30,000
21,932 $265,293
15
$45,000
$75,000
4,246 $204,672 $469,965
Table 4: Accident Severity Classification and Description Accident Severity Number
Description
1
No Injury
2
Possible Injury
3
Bruises and Abrasions
4
Broken Bones or Bleeding Wounds
5
Fatal
Accident severity differs between domestic and wild animals. Figure 6 shows a comparison between accident severity for wild and domestic animals between 1999 and 2001. Please note that accident severity 1 (no injury) is not included in the graph as there are a large percentage of accidents that follow under that category. No injury accident represents 77% and 93% of domestic and wild animal accidents respectively. The inclusion of no injury accidents in the figure will make the proportion of the other columns too small to see. It can be concluded from the figure that domestic animal accidents have a higher percentage of resulting on a high severity accident. This means that the 15
probabilities of getting injured or killed in an incident involving a domestic animal are higher compared to wild animals. Wild animals accidents result in 7% injury or death accidents while domestic animal accidents result in injury or death 23% of the time.
9
% of Total Accidents
8 7 6 5
Wild Domestic
4 3 2 1 0 5
4
3
2
Severity
Figure 6: Comparison Between Domestic And Wild Animal Severity
Another factor that is important when analyzing animal related accidents is the type of vehicle involved in the collision. Between the years 1999 and 2001 there were 5 deaths related to wild animal accidents, 3 of them involved a motorcycle and the other 2 involved passenger cars. Figure 7 shows the percentage of accident severity for motorcycle- and passenger car-animal related accidents. The figure shows that almost 50% of the accidents involving motorcycle will results in broken bones or bleeding wounds (severity 4). The figure also shows that just a small percentage of motorcycle accidents results in no injury. In terms of the type of animal involved both categories wild and domestic are of similar characteristics. On the other hand accident involving passenger cars are much 16
more likely to result in no injury accidents. Figure 8 shows the percentage of accident severity for accidents involving passenger cars. The figure shows that a large percentage of accidents result in no injury accidents and also a low percentage of accidents result in a fatal accident.
% Motorcycle Accidents
60 Wild Domestic
50 40 30 20 10 0 5
4
3 Severity
Figure 7: Motorcycle-Animal Accident Severity
17
2
1
100 % Pass. Car Accidents
90 80
Wild Domestic
70 60 50 40 30 20 10 0 5
4
3
2
1
Severity
Figure 8: Passenger Car-Animal Accident Severity
Figure 9 and Figure 10 present the results related to the analysis of the accident rate per mile for the most dangerous routes in relation to the occurrence of wild and domestic animal accidents. Figure 9 shows the routes with the highest wild animal accidents per mile. Their accident rate was estimated high based on the expected value analysis considering the accidents between 1999 and 2001. Table 5 shows the range of accident rate for each of the categories: low, moderate, and high. The average expected rate is 0.7 accidents per mile and the standard deviation is 0.9 accidents per mile. All the routes in Figure 9 have an accident per mile rate higher than 2.5 accidents per mile.
18
6
Accidents/Mile
5 4 3 2 1 0 146
189
38
248
40
52
111
91
203
224
Routes
Figure 9: Routes With High Wild Animal Accident Rate
Table 5: Wild Animal Accidents Classification Category
Acc/mile range
Low
0.0 – 1.6
Moderate
1.6 – 2.5
High
2.51 – more
Figure 10 shows the routes with high domestic animal accident per mile rate. They were selected, as their rate was higher based on the expected value analysis. The average number of accidents per mile rate for the domestic animals accidents is 0.2 accidents per mile and the standard deviation is 0.2 accidents per mile. It should be noted that the accident rate for domestic animals is significantly lower than the accident rate for the wild animal accident rate.
19
Acdidents/Mile
2
1
0 116
138
51
20
61
240
186
121
Routes
Figure 10: Routes With High Domestic Animal Accident Rate
Table 6: Domestic Animal Accident Category Category
Upper Limit (Acc/mile)
Low
0.0 – 0.4
Moderate
0.4 – 0.6
High
0.61 - more
20
104
Table 8 shows the route section with the highest accident number during the 3year period (1999-2001). Figure 11 shows the section of road with high wild animal accident number. The expected value analysis for the wild animal accidents indicates that 1-mile sections of the road that have more than 10 accidents during the 3-year period are considered to be high. Table 7 shows the classification for the 1-mile sections in terms of low, moderate, and high based on the expected value analysis. The numbers were based on the average and the standard deviation of1-mile sections with 1 or more accident during the 3year period. The average for the selected routes within the wild animals is 3.9 accidents and the standard deviation is 3.2 accidents per 1-mile section. This study found 55 1-mile sections that have high accident value. An additional 44 sections were included because when inspected they were found to be spatially located between other high accident sections. Table 8 presents a summary of the routes and the corresponding mile points for the high accident number sections. Table 8 shows the accidents in order of accident rate. A map number is also included which identifies the location of the accident within the State as shown on Figure 11.
Table 7: Classification of Sections Wild Animals Category Low Moderate High
Accidents in 1-Mile Section 0.0 – 7.2 7.21 – 10.3 10.31 – more
21
Table 8: Wild Animals High Accident Sections 1999-2001 Map #
Route
MP Start
MP End # of Acc.
19
89
371
373
61
30.5
3
6
234
235
14
14.0
15
89
235
238
42
14.0
18
89
340
344
56
14.0
25
210
0
1
14
14.0
4
15
121
126
68
13.6
2
6
222
228
79
13.2
6
40
6
13
92
13.1
13
73
30
31
13
13.0
11
70
3
6
38
12.7
10
68
35
39
48
12.0
16
89
249
250
12
12.0
20
92
1
4
36
12.0
21
130
4
5
12
12.0
22
146
2
3
12
12.0
7
40
76
78
23
11.5
24
191
62
74
133
11.1
5
15
142
143
11
11.0
9
40
122
123
11
11.0
26
224
10
11
11
11.0
14
80
133
141
79
9.9
1
6
192
195
29
9.7
17
89
288
290
19
9.5
12
70
74
77
24
8.0
23
189
17
26
72
8.0
8
40
85
89
27
6.8
22
Acc./Mile
High Wild Animal Accident Sections
¯
19
18 14 25
26 24
6 20
10
7 13
23
9
22
8
1 17 2 3
16 15 12 5 4 11
21
Figure 11: High Wild Animal Accident Sections
23
Table 9 shows the sections of road with high and moderate domestic animal accidents based on the expected value analysis for the domestic animal accidents. The high and moderate locations are selected because of the higher severity associated with the domestic animal accidents. Table 10 shows the values for the different low, moderate, and high categories. The average and standard deviation are 1.7 and 1.4 accidents per 1-mile section, respectively. Domestic animals have fewer locations of concern. The occurrence of domestic animals is less frequent and also occurs randomly based on fence maintenance rather than being concentrated in certain potential migration locations. Figure 12 shows domestic animal accident hot spots. Table 9: Domestic Animals High and Moderate Accident Sections 1999-2001
Map # Route
MP
MP
# of
Start
End
Acc.
Acc./Mile Class.
3
20
6
7
9
9.0
High
2
138
6
7
6
6.0
High
1
116
4
5
5
5.0
High
5
40
124
125
4
4.0
Moderate
7
89
398
399
4
4.0
Moderate
8
91
22
23
4
4.0
Moderate
4
10
5
7
7
3.5
Moderate
6
89
178
180
7
3.5
Moderate
Table 10: Classification of Section Domestic Animals Category Low Moderate High
Accidents in 1-Mile Section 0.0 – 3.1 3.11 – 4.4 4.41 – more
24
High Domestic Animal Accident Sections
¯
8 7
2
5
1
4
6
3
Figure 12: High Domestic Animal Accident Sections
25
Table 11 shows a description of where fences are located as provided by UDOT. The data included route and milepost. Table 11 shows the accident rate per mile and the classification based on the expected value analysis considered for the wild animals. It can be concluded from the table that some of the sections with fence have a low accident rate, but some section have high accident rate. It cannot be concluded that fences have been beneficial. Further study needs to be done in order to determine if fence is effective. Data regarding fence condition were not available, thus this study is not able to reach a conclusion. More information about the number of accident at these locations can be found in appendix 1. Table 11: Accident Rate of Sections With Fences (1999-2001) Route Mp end Mp start Wild Accidents Accidents per mile Classification 15
26.9
34.0
12
1.71
Moderate
15
61.9
82.9
37
1.76
Moderate
15
94.5
108.8
5
0.35
Low
15
170.0
187.0
7
0.41
Low
15
193.0
194.0
1
1.00
Low
15
263.6
271.8
3
0.37
Low
15
273.0
277.0
10
2.50
High
15
279.8
284.2
7
1.59
Low
89
162.0
163.7
0
0.00
Low
70
7.6
57.0
40
0.81
Low
91
3.8
16.0
43
3.54
High
248
3.6
12.0
26
3.11
High
40
2.6
10.1
74
9.91
High
26
6
RECOMMENDATIONS
The historic accident trend, severity, cost to society and general vehicle-animal accident problem is identified in this study. While the end result is locations of specific concern related to both domestic and wild animal accidents, the study identifies the power of the CARS database in its ability to query accident records for the past 10 years. This research and investigation of animal accidents and including location, severity and distinction between domestic and wild animals would not have been available without the access to the historic accident records. Having identified the scope of the problem and location of accident hotspots regarding vehicle-animal accidents, more focus and efficient decision making can now be brought onto solution options and thereby allow the limited economic resources available to be most effectively utilized. The CARS value as a decision-support tool is obvious and should be promoted among the different UDOT division. While this study focused in the analysis of all the routes in the state, the use of CARS database can be applied to smaller, more focused studies providing even more detailed analysis. 7
FUTURE WORK
Now that the locations are identified, site investigation and migration patterns would allow a better understanding of common features that may create a more hazardous crossing than others. Future work should be focused on the evaluation of mitigation measures and their effectiveness including fence and overpasses/underpasses. Some of the variables that should be considered are deer population, fence condition, date of installation and maintenance, as well as wildlife behavior and ways to reduce accidents. In relation to domestic animals, work should focus on ways to eliminate the problem since these are much more dangerous based on severity than the wild animal accidents. A better understanding of cause of domestic accidents and the
27
legal ramifications for the State and domestic animal owner would be an area of research to explore.
REFERENCES 1. Schafer, James A., Stephen Penland, and William P. Carr. “Effectiveness of Wildlife Warning Reflectors in Reducing Deer-Vehicle Accidents in Washington State.” Transportation Research Record 1010 (1985): 85-88 2. Darryll J. Dockstader and Peter D. Southall. “Ecopasage Reduces Roadkills” TR News 227 July-August 2003. pp 38-39. 3. Ludwig, John, and Timothy Bremicker. “Evaluation of 2.4 m Fences and One-way Gates for Reducing Deer-Vehicle Collisions in Minnesota. ” Transportation Research Record 913 (1983): 19-22. 4. Schafer, James A., Stephen Penland, and William P. Carr. “Effectiveness of Wildlife Warning Reflectors in Reducing Deer-Vehicle Accidents in Washington State.” Transportation Research Record 1010 (1985): 85-88 5. Zacks, James L. “Do White-Tail Deer Avoid Red? An Evaluation of the Premise Underlying the Design of Swareflex Wildlife Reflectors.” Transportation Research Record 1075 (1986): 35-43. 6. Ministry of Transportation and Highways of British Columbia, Canada. “Wildlife Fencing Program”. Retrieved from on October 6th 2003. 7. F. E. Hartman. “Hunting is Big Business. Pennsylvania Game News, Vol. 44, 1973, pp. 26-31. 8. R. Norman. “Using Wildlife Values in Benefit-Cost Analysis and Mitigation of Wildlife Losses. Proc., International Association of Game, Fish, and Conservation Commissioners, Vol. 56, 1975, pp. 119-128. 9. North Dakota State University * Dickinson Research Extension Center. Retrieved on October 9, 2003 from http://www.ag.ndsu.nodak.edu/dickinso/research/1975/straw75.htm 10. Great southern outdoors. “Alabama Cooperative Deer Management 28
Program & Deer Harvest Report 2002-2003 . Retrieved on October 9, 2003 from http://www.greatsouthernoutdoors.com/Results/results.html
29
8
APPENDIX
Accidents per year at fenced locations (1996-2001)
Num ber of Accidents
Wild Animal Accidents US40 (mp 2-11) 40 35 30 25 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents US89 (mp162-164) 2.5 2 1.5 1 0.5 0 1996
1997
1998
1999 Year
30
2000
2001
Num ber of Accidents
Wild Animal Accidents US70 (mp7-57) 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents US91 (mp 3-17) 50 40 30 20 10 0 1996
1997
1998
1999 Year
31
2000
2001
Num ber of Accidents
Wild Animal Accidents US248 (mp3-13) 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Wild Animal Accidents I-15 (mp 263.5-272) Num ber of Accidents
Num ber of Accidents
Wild Animal Accidents I-15 (mp 26.8-34.2)
2.5 2 1.5 1 0.5 0 1996
1997
1998
1999 Year
32
2000
2001
Wild Animal Accidents I-15 (mp 279.5-284.5) Num ber of Accidents
6 5 4 3 2 1 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents I-15 (mp 61.8-83) 25 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents I-15 (mp 273-277) 5 4 3 2 1 0 1996
1997
1998
1999 Year
33
2000
2001
Num ber of Accidents
Wild Animal Accidents I-15 (mp 94-109) 10 8 6 4 2 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents I-15 (mp 170-187) 3.5 3 2.5 2 1.5 1 0.5 0 1996
1997
1998
1999 Year
34
2000
2001
ANIMAL-VEHICLE ACCIDENT ANALYSIS
Prepared For: Utah Department of Transportation Research and Development Division Submitted By: University of Utah
Authored By: Dr. Joseph Perrin Rodrigo Disegni
November 2003
UDOT RESEARCH & DEVELOPMENT REPORT ABSTRACT
1. Report No.
2. Government Accession No.
UT-03.31
4. Title and Subtitle
5. Report Date
Safety Benefits of UDOT Highway Program, Animal-Vehicle Accident Analysis 7. Author(s)
3. Recipient's Catalog No.
November 2003 6. Performing Organization Code 8. Performing Organization Report No.
Perrin, Joseph
9. Performing Organization Name and Address
10. Work Unit No.
University of Utah 122 South Central Campus Drive, Room 104 SLC, UT 84112
11. Contract No.
12. Sponsoring Agency Name and Address
13. Type of Report and Period Covered
Utah Department of Transportation 4501 South 2700 West Box 148410 SLC, Utah 84114
Research – August to November 2003 14. Sponsoring Agency Code
UDOT - 039178
15. Supplementary Notes
16. Abstract
Vehicle-animal accidents represented 4.6% of US automobile accidents with more than 1.5 million accidents a year, 150 deaths and $1.1 billion in vehicle damage. Animal related accidents in Utah represent 1.2% of statewide automobile accidents. In 2001 there were 2,688 vehicle-animal collisions, including 3 death, and 235 injuries. In Utah, animal related accidents are subdivided in wild and domestic animals. Domestic animals include livestock, such as cows and sheep or horses. Wild animal most often refer to deer, elk and moose. Using 10-years of statewide accident information, the problem locations are identified and a comparison between the domestic and wild animal accidents based on severity is examined. The accident analysis determined that while the domestic animal accidents represent only16% of the animal-vehicle accidents, they are more severe than wild animal accidents. Domestic animal accidents result in injury 23% of the time while wild animal accidents results in injury only 7% of the time. When a motorcycle is involved, it was found that 94% of the animal-motorcycle related accidents resulted in injury compared with only 11% of the non-motorcycle-animal accidents. Overall, there is a 7.9 times greater chance of a fatality with domestic animal accidents compared with wild animals. This is attributed to the height and weight of domestic animal relative to the common wild animal. While many countermeasures are attempted such as whistles and reflectors, the principal countermeasure to control animal related accidents has been the use of fences along the roads. The 4-foot high right-of-way fences are effective for the domestic animals but the wildlife animals requires the higher 8-foot fences since deer can easily circumvent the 4-foot fence heights. Alternative countermeasures such as one-way deer gates and ecopassages are also reducing wild animal hits. This study utilized UDOTs CARS accident database to identify the vehicle-animal crash problem in Utah. The study describes the extent of the problem; some literature on various countermeasures used throughout the world, and finally identifies the most dangerous section of routes between the years 1999-2001 in terms of the accidents per mile for the wild and domestic animals. 17. Key Words Animals - Accidents, wild animals, domestic animals
18. Distribution Statement
19. Security Classification (of this report)
21. No. of Pages
20. Security Classification (of this page)
35
22. Price
EXECUTIVE SUMMARY Vehicle-animal accidents represented 4.6% of US automobile accidents with more than 1.5 million accidents a year, 150 deaths and $1.1 billion in vehicle damage. Animal related accidents in Utah represent 1.2% of statewide automobile accidents. In 2001 there were 2,688 vehicle-animal collisions, including 3 death, and 235 injuries. In Utah, animal related accidents are subdivided in wild and domestic animals. Domestic animals include livestock, such as cows and sheep or horses. Wild animal most often refer to deer, elk and moose. Using 10-years of statewide accident information, the problem locations are identified and a comparison between the domestic and wild animal accidents based on severity is examined. The accident analysis determined that while the domestic animal accidents represent only16% of the animal-vehicle accidents, they are more severe than wild animal accidents. Domestic animal accidents result in injury 23% of the time while wild animal accidents results in injury only 7% of the time. When a motorcycle is involved, it was found that 94% of the animal-motorcycle related accidents resulted in injury compared with only 11% of the non-motorcycle-animal accidents. Overall, there is a 7.9 times greater chance of a fatality with domestic animal accidents compared with wild animals. This is attributed to the height and weight of domestic animal relative to the common wild animal. While many countermeasures are attempted such as whistles and reflectors, the principal countermeasure to control animal related accidents has been the use of fences along the roads. The 4-foot high right-of-way fences are effective for the domestic animals but the wildlife animals requires the higher 8-foot fences since deer can easily circumvent the 4-foot fence heights. Alternative countermeasures such as one-way deer gates and ecopassages are also reducing wild animal hits. This study utilized UDOTs CARS accident database to identify the vehicle-animal crash problem in Utah. The study describes the extent of the problem; some literature on various countermeasures used throughout the world, and finally identifies the most dangerous section of routes between the years 1999-2001 in terms of the accidents per mile for the wild and domestic animals.
i
TABLE OF CONTENTS
List of Figures ................................................................................................ iii List of Tables .................................................................................................iv 1
Introduction............................................................................................. 1
2
Literature Review ................................................................................... 1
3
Scope ..................................................................................................... 5
4
Methodology........................................................................................... 6
5
Results ................................................................................................... 8
6
Recommendations ............................................................................... 27
7
Future Work.......................................................................................... 28
8
Appendix .............................................................................................. 30
References .................................................................................................. 34
ii
LIST OF FIGURES Figure 1: Domestic Animal Accidents Trend........................................................9 Figure 2: Wild Animal Accidents Trend..............................................................10 Figure 3: Fatality Rate for Domestic and Wild Animal Accidents .......................11 Figure 4: Vehicle-Animal Accidents by Time of Day ..........................................12 Figure 5: Vehicle-Animal Accidents by Month of the Year .................................13 Figure 6: Comparison Between Domestic And Wild Animal Severity ................16 Figure 7: Motorcycle-Animal Accident Severity .................................................17 Figure 8: Passenger Car-Animal Accident Severity...........................................18 Figure 9: Routes With High Wild Animal Accident Rate ....................................19 Figure 10: Routes With High Domestic Animal Accident Rate...........................20
iii
LIST OF TABLES Table 1: Accident Classification...........................................................................7 Table 2: Percentage of Multi-Vehicle Accidents Between 1992 and 2001.........14 Table 3: Animal Accident Costs (1992-2001) ....................................................15 Table 4: Accident Severity Classification and Description .................................15 Table 5: Wild Animal Accidents Classification ...................................................19 Table 6: Domestic Animal Accident Category....................................................20 Table 7: Wild Animals High Accident Sections 1999-2001 ................................22 Table 8: Classification of Sections Wild Animals ...............................................22 Table 9: Domestic Animals High Accident Sections 1999-2001 ........................24 Table 10: Classification of Section Domestic Animals .......................................24 Table 11: Accident Rate of Sections With Fences (1999-2001) ........................26
iv
1
INTRODUCTION
The following report documents the animal-vehicle accidents problem in Utah. While the more general purpose of this study is to identify the value of the webinterface to the Centralized Accident Record Systems (CARS) database, the objectives to this study are to utilize the CARS system to identify general trends over the past 10-years and identify the “hot spots” throughout the state for the past three years. The accidents are segregated into wild and domestic and thus the analysis distinguishes between these two types of animal-vehicle accident.
2
LITERATURE REVIEW
The literature on vehicle-animal accidents begins in the 1980s. Past literature has focused on defining the problem, determining the economic impacts of the accidents for both the driver and the wildlife, and evaluating the effectiveness of the devices that keep wildlife off roads. As vehicle travel has increased and intercity travel expanded, these routes and frequency of traffic conflicts with wildlife path increasing the interaction between vehicles and animals. In 1980 the number of accidents involving deer in America’s highways was estimated at 200,000 collisions (Schafer et al. 1985). This number is estimated to be conservative, as many incidents are not reported because of the low severity. The latest number reported in 2001 shows a considerable increase at 292,000 accidents, which equates to 4.6% of the total accidents in 2001. During 2001 there were also 165 fatalities (0.4% of all fatalities) and 19,000 injuries (0.9% of all injuries) as a consequence of vehicle-animal collisions (Traffic Safety Facts 2001). In terms of the crash circumstances, most of the vehicle-animal crashes occurred on rural roads compared to urban roads, and during nighttime between 7 and 10 pm (Hughes, Sremi, and Paniati 1996). Researchers have also found that
1
October and November are the months of the year when the number of accidents peak, as those months represent the mating season for deer (Hughes et al.). This last finding, however, contradicts the conclusions of a similar study where researchers in Washington State concluded that most of the animal related accident occurred during the summer time versus late fall (Schafer et al. 1985). We estimate that there is likely a relationship between traffic volume variations and vacation travel related to the accident rates in Washington as summer is a much higher travel time within the typically rainy state. The economic value involved in vehicle-animal incidents considers both the driver and the animal costs. It was estimated in 1977 that the driver’s cost of an accident was about $730, which equates to $2,160 in 2003 dollars. The economic lost for a deer however is more difficult to quantify, but was estimated in 1976 at $350 dollars, which represents $1,100 in 2003 dollars (Schafer et al.). Other researchers, however, estimated the dollar value of a deer at twice this figure based on hunting expenditures alone (Hartman, and Norman). It can be concluded that whatever the economic value of deer is, the overall problem should be, if under reporting is considered around one billion dollars annually. Several devices have been used in the past to prevent wildlife from being on the roads. Some of the devices commonly used are: warning signs, fences, wildlife reflectors, and lately ecopassages (underpasses). The evaluation of these devices has reached different levels of success. However, an issue that is relevant to all the studies that evaluates the performances of these devices is the difficulty to control deer population during the study. Therefore a decrease in accident rate on a fenced road may be explained by a decrease in deer population. The cheapest device used to prevent vehicle-animal accidents has been the placement of wildlife warning signs. However, it has been reported that they are the least effective and its indiscriminate use has lowered the credibility of the sign. A study concluded that 60% of the drivers will not even notice the sign while driving (Ministry of BC, Hughes).
2
The use of fences has been found to be successful at keeping animals off the road. The ministry of Transportation and highways in BC, Canada reported that the use of 2.4 meter (7.8 feet) high fence has been very efficient at keeping wildlife off the roads. They report that this higher fencing is 97-99% effective in preventing vehicle-animal accidents. Some of the problems of fencing are that wildlife is restricted to cross roads and in the case of fence penetration wildlife would be trapped in the road. Restricting wildlife from crossing roads can have serious consequences if food and water resources are across the road. Fence penetration is more common when the 1.2 meter high fence is used instead of the 2.4 meter high fence. The use of fencing along with under and overpasses as well as one-way gates has been reported as a better way to deal with the problem. Ludwig 1983 reports benefit cost ratios of 3.61 and reduction of deer accidents of 60% and 93% in two sites where fencing and one-way gates were installed. Wildlife reflectors are a different type of technology that has been used in the past with a wide range of success. The main advantage of reflectors is that they are cheaper and less intrusive than fencing. The theory behind reflectors is that the headlight of a vehicle is reflected into a red light that produces a freezing response in deer (Schafer et al.). Although the theory has been widely discussed and researched there are still some controversies regarding the effect that red light may have on deer. A study by Zacks (1986) tested the response of deer to the red color of the reflectors. The study found that the behavior of deer was the same under red reflectors, white reflectors and no reflectors. The conclusion of the study was that red reflectors do not modify deer’s behavior. The researchers suggested that the effect of reflectors might be on the drivers’ behavior rather than on the deer. Nonetheless success and failure of reflectors have been documented in two studies with two opposite outcomes. A study by Schafer et al., which represent one of the best studies in terms of controlling for deer population, concluded that wildlife reflectors decrease the number of vehicleanimal incidents over the study period. Another study by the Ministry of
3
Transportation and highways in British Columbia, Canada (2003) concluded that after the installation of reflectors the number of accidents involving deer increased, deer population however was not controlled in this study. Although both studies differ in terms of the methodology used, they still represent valuable research in terms of the effectiveness of wildlife reflectors. The newest device that has been used and evaluated is the ecopassages. Ecopassages consist of a wall or fence that keeps animals off the road and a set of underground passes, usually culverts, that allow animals to cross the road underneath (Dockstader and Southall 2003). The use of ecopassages in Florida reported a decreased in the number of wildlife killed of 64.2% compared to the before period. The study also reported an increase in the wildlife species that were using the underpasses to cross the road. The literature shows that the use of before and after comparison to evaluate the effectiveness of vehicle-animal mitigations leads to misleading conclusions if deer population is not controlled. The change in migration patterns and animal behavior from year to year and even from month to month make the comparison of before and after studies, where wildlife population is not controlled, difficult from a statistical scientifically standpoint, much like most accident trend analysis. .
4
3
SCOPE
3.1
Problem Definition
The vehicle-animal collision problem in Utah is of similar characteristics to that faced by other states in the US. About 5.1% of the total accidents in 2001 involved either wild or domestic animals. There were a total of 288 vehicleanimal accidents in 2001 including 3 deaths (1.2% of all fatalities) and 235 injuries (1.2% of all injuries) (Traffic Safety Facts 2001). This study identifies a methodology to determine high accident spots through the use of the Centralized Accident Record System (CARS) database. The study focuses on accidents that involve wild and domestic animals. Wild animal accidents usually represent accidents involving deer, moose, bear, wolf, and other species that are considered not domestic. Domestic animal accidents usually involve cows, sheep and horses.
3.2
Research Objectives The main objectives of this study are: •
To use the UDOT CARS database Web site to perform a safety analysis of accidents involving wild and domestic animals.
•
To define a methodology that identifies road sections that have high accident, and should be further analyzed.
5
4
METHODOLOGY
The primary result identifies routes by milepost where high animal-vehicle accidents occurr. The data analysis of this study has been divided in two parts. The first one identifies routes among all the state routes, which have a high accident per mile rate. The second one identifies route sections where accident numbers are particularly high. This study estimates high accident rate and numbers based on expected value analysis. Expected value analysis assumes that the occurrence of accidents follows a normal distribution. Under the expected value analysis the mean and the standard deviation determine the shape that the normal distribution will have. Based on the normal distribution and a 95% confidence level, it is defined that values higher than the average plus 1.96 standard deviation are consider being abnormal. Abnormal locations would have high accident rate and thus further studies should be performed to determine the reasons for the abnormal accident rate or accident number. The following two sections describe the methodology used to determine the most dangerous routes and sections for the wild and domestic animal accidents.
4.1
Estimation of the Most Dangerous Routes
This study includes all the state routes in Utah, excluding the routes that are shorter than 1 mile in length. These routes are excluded from the study because the inclusion of short routes can skew the results, as their accident per mile rate can be high with just the occurrence of 1 accident in the section. The study period is defined as 3 years between 1999 and 2001. The accident rate and average number of accidents per mile are selected as the MOE to estimate the routes with a high number of accidents. The accident per mile rate is selected because it allows the comparison between long routes that may have a large number of accidents with short routes that may have a low 6
number of accidents. This study performs a separate analysis for accidents involving domestic and wild animals, as they can be different in terms of severity and frequency. Based on the average number of accidents and the standard deviation each route can be classified in terms of a Low, Moderate, and High accident rate scale. Table 1 shows the limits of the different categories. Based on the normal distribution the mean should include 50% of the population, the mean plus one standard deviation should include 68% of the population, and finally the mean plus 1.96 standard deviation includes 95% of the population. Therefore accident rates that are higher than the 95% of the population are considered abnormal (High).
Table 1: Accident Classification Category
4.2
Limit
Low
Mean
Moderate
Mean + Standard Deviation
High
Mean +1.96*Standard Deviation
Estimation of Route Sections With High Accident Numbers
All the routes are analyzed in terms of the number of accidents per 1-mile segment increment. As several sections within a route do not have accidents during the 3-year period, all the analyses are referred to sections that have at least 1 accident during the 3-year study period (1999-2001). Using a similar approach as the one used in the high route identification, the mean and standard deviation of the 1-mile sections with more than 1 accident for all the routes are calculated. Based on the definition shown in Table 1 all the sections with an accident number higher than the mean plus 1.96 standard deviation are define as high.
7
The identification of high accidents section would indicate places where maintenance or installation of new fence should be considered. 5
RESULTS
The findings of this study are included with a discussion of the results. This section has been organized in two parts. It starts with a description of the trend of animal accidents in Utah between the 10-years from 1992 and 2001. The second part provides a more refined analysis of the animal incidents between the years 1999 and 2001 for all the state routes in Utah excluding the routes that are shorter than 1 mile.
5.1
Vehicle-Animal Accident Trends between 1992 and 2001
Figure 1 shows the accident trend of crashes involving domestic animals. The figure shows that the number of accidents has increased over the years. The number of accidents increased from 340 accidents in 1992 to over 460 accidents in 2001.
8
Number of Accidents
600 500 400 300 200 100 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Year
Figure 1: Domestic Animal Accidents Trend
Figure 2 presents the number of accident involving wild animals. The number of accidents has remained fairly constant over the years. It can be concluded that wild animal accident are much more frequent than domestic animal accidents. In average the number of wild accidents are 5 times the number of domestic animal accidents. A reason for this may be the higher population that wild animals represents compared to domestic animals. Another reason may be the fact that domestic animals are usually in a fenced area, and easier contained with smaller right-of-way fences compared to wild animals that may be running freely.
9
Number of Accidents
3000 2500 2000 1500 1000 500 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Year
Figure 2: Wild Animal Accidents Trend
Figure 3 shows the fatality rate for domestic and wild animal accidents. Please note that during the years 1993 and 1995 there were no fatal accidents in both categories. The data indicates that the fatality rate for domestic animal accidents is almost 8 times higher than for wild animals. If we consider that the domestic animal accidents represents fewer accidents compared to domestic animal accidents, then we can conclude that the chances of having an injury or even a fatal accident is higher for domestic animals than wild animals. A reason for this is the type of animals included under each category. While a wild animal represents usually a deer or a moose, a domestic animal many times is a cow or horse. These animals are very different in terms of weight. A cow can easily be more than 5 times heavier than a deer; in average a cow weighs 1,000 pounds while a deer can weight 140 pounds (9,10).
10
10 9
Fatality Rate
8 Wild Domestic
7 6 5 4 3 2 1 0 1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
Year
Figure 3: Fatality Rate for Domestic and Wild Animal Accidents
5.2
Vehicle-Animal Accident Analysis between 1999 and 2001
Assessing when and where the accidents occur sheds insight into the circumstances surrounding the accidents. Figure 4 shows the percentage of animal related accident occurrence by time of day. The figure shows that most of the accidents occur during darkness between 6 PM and 10 PM. There is also another peak of accidents that happen early in the morning at 6 AM, but representing a smaller proportion of accidents.
11
% of Total Accidents
14 12 10 8 6 4 2 0 0
2
4
6
8
10
12
14
16
18
20
22
24
Time of Day
Figure 4: Vehicle-Animal Accidents by Time of Day
Figure 5 shows the percentage of accident related animals for the different month in the year. It is concluded from the figure that animal related accidents peak during the fall season. October, November, and December are the months with the highest percentage of animal related accident. It is described in the literature that this period of the year (fall) coincides with the mating season of deer and migration as weather changes and more animals move from higher elevations to the lower elevations where there is also a higher likelihood of vehicle–animal interaction.
12
14 % of Total Accidents
12 10 8 6 4 2 0 Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov Dec
Month
Figure 5: Vehicle-Animal Accidents by Month of the Year
Most of the animal related accidents involve only one vehicle in the collision. However there are cases when there is more than one vehicle involved. Table 2 shows the percentage of multi-vehicle accidents for wild and domestic animal incidents. It is concluded that domestic animals have a higher percentage of accidents involving 2 or more vehicles. A reason for this can be the fact that domestic animals are heavier compared to wild animals and result in more severe accidents. Therefore after a car has hit a domestic animal, the car may be forced to stop abruptly causing more vehicles to get involved in the incident or the body of the animal may still be blocking the road, therefore increasing the probability of more vehicles involved.
13
Table 2: Percentage of Multi-Vehicle Accidents Between 1992 and 2001 Vehicles Involved
Wild
Domestic
1
99.5%
96.5%
2
0.5%
3.2%
3
0.0%
0.2%
4+
0.0%
0.1%
The economic cost involved in animal related accidents is difficult to determine since the cost of wildlife is difficult to quantify. Past studies have attempted to estimate the cost involved with the lost of wildlife reaching widely different numbers. This study does not attempt to estimate the economic loss of wildlife, but recognizes that the cost of wildlife constitute an important factor in the economic evaluation of animal related accidents. In terms of the driver’s cost involved in animal related accidents, this study uses the current values of accidents based on severity as defined by Federal Highways. Table 3 presents the estimated cost for accidents according to the severity of the incident. The cost ranges from $2,300 dollars for an incident with no injuries (severity 1) up to 3 million dollars for a fatal accident (severity 5). Table 4 shows a description of the different severity types, categories 1 through 5, and the characteristics of the accident. Table 3 also shows the number of wild and domestic animal accidents between the years 1992 and 2001 for each severity type. Based on the cost of each type of accident and the number of accidents during the 10-year period, this study estimates that the overall cost involved in animal related accidents is approximately $469 million dollars or an average $46 million per year.
14
Table 3: Animal Accident Costs (1992-2001) Wild Cost per
Domestic Cost
Cost
Total
Severity Accident Accidents Thousand Accidents Thousand Thousand 1
2,300
20,629
$47,446
3,367
$7,744
$55,190
2
6,000
582
$3,492
328
$1,968
$5,460
3
45,000
418
$18,810
294
$13,230
$32,040
4
565,000
5
3,000,000 Total
293 $165,545 10
242 $136,730 $302,275
$30,000
21,932 $265,293
15
$45,000
$75,000
4,246 $204,672 $469,965
Table 4: Accident Severity Classification and Description Accident Severity Number
Description
1
No Injury
2
Possible Injury
3
Bruises and Abrasions
4
Broken Bones or Bleeding Wounds
5
Fatal
Accident severity differs between domestic and wild animals. Figure 6 shows a comparison between accident severity for wild and domestic animals between 1999 and 2001. Please note that accident severity 1 (no injury) is not included in the graph as there are a large percentage of accidents that follow under that category. No injury accident represents 77% and 93% of domestic and wild animal accidents respectively. The inclusion of no injury accidents in the figure will make the proportion of the other columns too small to see. It can be concluded from the figure that domestic animal accidents have a higher percentage of resulting on a high severity accident. This means that the 15
probabilities of getting injured or killed in an incident involving a domestic animal are higher compared to wild animals. Wild animals accidents result in 7% injury or death accidents while domestic animal accidents result in injury or death 23% of the time.
9
% of Total Accidents
8 7 6 5
Wild Domestic
4 3 2 1 0 5
4
3
2
Severity
Figure 6: Comparison Between Domestic And Wild Animal Severity
Another factor that is important when analyzing animal related accidents is the type of vehicle involved in the collision. Between the years 1999 and 2001 there were 5 deaths related to wild animal accidents, 3 of them involved a motorcycle and the other 2 involved passenger cars. Figure 7 shows the percentage of accident severity for motorcycle- and passenger car-animal related accidents. The figure shows that almost 50% of the accidents involving motorcycle will results in broken bones or bleeding wounds (severity 4). The figure also shows that just a small percentage of motorcycle accidents results in no injury. In terms of the type of animal involved both categories wild and domestic are of similar characteristics. On the other hand accident involving passenger cars are much 16
more likely to result in no injury accidents. Figure 8 shows the percentage of accident severity for accidents involving passenger cars. The figure shows that a large percentage of accidents result in no injury accidents and also a low percentage of accidents result in a fatal accident.
% Motorcycle Accidents
60 Wild Domestic
50 40 30 20 10 0 5
4
3 Severity
Figure 7: Motorcycle-Animal Accident Severity
17
2
1
100 % Pass. Car Accidents
90 80
Wild Domestic
70 60 50 40 30 20 10 0 5
4
3
2
1
Severity
Figure 8: Passenger Car-Animal Accident Severity
Figure 9 and Figure 10 present the results related to the analysis of the accident rate per mile for the most dangerous routes in relation to the occurrence of wild and domestic animal accidents. Figure 9 shows the routes with the highest wild animal accidents per mile. Their accident rate was estimated high based on the expected value analysis considering the accidents between 1999 and 2001. Table 5 shows the range of accident rate for each of the categories: low, moderate, and high. The average expected rate is 0.7 accidents per mile and the standard deviation is 0.9 accidents per mile. All the routes in Figure 9 have an accident per mile rate higher than 2.5 accidents per mile.
18
6
Accidents/Mile
5 4 3 2 1 0 146
189
38
248
40
52
111
91
203
224
Routes
Figure 9: Routes With High Wild Animal Accident Rate
Table 5: Wild Animal Accidents Classification Category
Acc/mile range
Low
0.0 – 1.6
Moderate
1.6 – 2.5
High
2.51 – more
Figure 10 shows the routes with high domestic animal accident per mile rate. They were selected, as their rate was higher based on the expected value analysis. The average number of accidents per mile rate for the domestic animals accidents is 0.2 accidents per mile and the standard deviation is 0.2 accidents per mile. It should be noted that the accident rate for domestic animals is significantly lower than the accident rate for the wild animal accident rate.
19
Acdidents/Mile
2
1
0 116
138
51
20
61
240
186
121
Routes
Figure 10: Routes With High Domestic Animal Accident Rate
Table 6: Domestic Animal Accident Category Category
Upper Limit (Acc/mile)
Low
0.0 – 0.4
Moderate
0.4 – 0.6
High
0.61 - more
20
104
Table 8 shows the route section with the highest accident number during the 3year period (1999-2001). Figure 11 shows the section of road with high wild animal accident number. The expected value analysis for the wild animal accidents indicates that 1-mile sections of the road that have more than 10 accidents during the 3-year period are considered to be high. Table 7 shows the classification for the 1-mile sections in terms of low, moderate, and high based on the expected value analysis. The numbers were based on the average and the standard deviation of1-mile sections with 1 or more accident during the 3year period. The average for the selected routes within the wild animals is 3.9 accidents and the standard deviation is 3.2 accidents per 1-mile section. This study found 55 1-mile sections that have high accident value. An additional 44 sections were included because when inspected they were found to be spatially located between other high accident sections. Table 8 presents a summary of the routes and the corresponding mile points for the high accident number sections. Table 8 shows the accidents in order of accident rate. A map number is also included which identifies the location of the accident within the State as shown on Figure 11.
Table 7: Classification of Sections Wild Animals Category Low Moderate High
Accidents in 1-Mile Section 0.0 – 7.2 7.21 – 10.3 10.31 – more
21
Table 8: Wild Animals High Accident Sections 1999-2001 Map #
Route
MP Start
MP End # of Acc.
19
89
371
373
61
30.5
3
6
234
235
14
14.0
15
89
235
238
42
14.0
18
89
340
344
56
14.0
25
210
0
1
14
14.0
4
15
121
126
68
13.6
2
6
222
228
79
13.2
6
40
6
13
92
13.1
13
73
30
31
13
13.0
11
70
3
6
38
12.7
10
68
35
39
48
12.0
16
89
249
250
12
12.0
20
92
1
4
36
12.0
21
130
4
5
12
12.0
22
146
2
3
12
12.0
7
40
76
78
23
11.5
24
191
62
74
133
11.1
5
15
142
143
11
11.0
9
40
122
123
11
11.0
26
224
10
11
11
11.0
14
80
133
141
79
9.9
1
6
192
195
29
9.7
17
89
288
290
19
9.5
12
70
74
77
24
8.0
23
189
17
26
72
8.0
8
40
85
89
27
6.8
22
Acc./Mile
High Wild Animal Accident Sections
¯
19
18 14 25
26 24
6 20
10
7 13
23
9
22
8
1 17 2 3
16 15 12 5 4 11
21
Figure 11: High Wild Animal Accident Sections
23
Table 9 shows the sections of road with high and moderate domestic animal accidents based on the expected value analysis for the domestic animal accidents. The high and moderate locations are selected because of the higher severity associated with the domestic animal accidents. Table 10 shows the values for the different low, moderate, and high categories. The average and standard deviation are 1.7 and 1.4 accidents per 1-mile section, respectively. Domestic animals have fewer locations of concern. The occurrence of domestic animals is less frequent and also occurs randomly based on fence maintenance rather than being concentrated in certain potential migration locations. Figure 12 shows domestic animal accident hot spots. Table 9: Domestic Animals High and Moderate Accident Sections 1999-2001
Map # Route
MP
MP
# of
Start
End
Acc.
Acc./Mile Class.
3
20
6
7
9
9.0
High
2
138
6
7
6
6.0
High
1
116
4
5
5
5.0
High
5
40
124
125
4
4.0
Moderate
7
89
398
399
4
4.0
Moderate
8
91
22
23
4
4.0
Moderate
4
10
5
7
7
3.5
Moderate
6
89
178
180
7
3.5
Moderate
Table 10: Classification of Section Domestic Animals Category Low Moderate High
Accidents in 1-Mile Section 0.0 – 3.1 3.11 – 4.4 4.41 – more
24
High Domestic Animal Accident Sections
¯
8 7
2
5
1
4
6
3
Figure 12: High Domestic Animal Accident Sections
25
Table 11 shows a description of where fences are located as provided by UDOT. The data included route and milepost. Table 11 shows the accident rate per mile and the classification based on the expected value analysis considered for the wild animals. It can be concluded from the table that some of the sections with fence have a low accident rate, but some section have high accident rate. It cannot be concluded that fences have been beneficial. Further study needs to be done in order to determine if fence is effective. Data regarding fence condition were not available, thus this study is not able to reach a conclusion. More information about the number of accident at these locations can be found in appendix 1. Table 11: Accident Rate of Sections With Fences (1999-2001) Route Mp end Mp start Wild Accidents Accidents per mile Classification 15
26.9
34.0
12
1.71
Moderate
15
61.9
82.9
37
1.76
Moderate
15
94.5
108.8
5
0.35
Low
15
170.0
187.0
7
0.41
Low
15
193.0
194.0
1
1.00
Low
15
263.6
271.8
3
0.37
Low
15
273.0
277.0
10
2.50
High
15
279.8
284.2
7
1.59
Low
89
162.0
163.7
0
0.00
Low
70
7.6
57.0
40
0.81
Low
91
3.8
16.0
43
3.54
High
248
3.6
12.0
26
3.11
High
40
2.6
10.1
74
9.91
High
26
6
RECOMMENDATIONS
The historic accident trend, severity, cost to society and general vehicle-animal accident problem is identified in this study. While the end result is locations of specific concern related to both domestic and wild animal accidents, the study identifies the power of the CARS database in its ability to query accident records for the past 10 years. This research and investigation of animal accidents and including location, severity and distinction between domestic and wild animals would not have been available without the access to the historic accident records. Having identified the scope of the problem and location of accident hotspots regarding vehicle-animal accidents, more focus and efficient decision making can now be brought onto solution options and thereby allow the limited economic resources available to be most effectively utilized. The CARS value as a decision-support tool is obvious and should be promoted among the different UDOT division. While this study focused in the analysis of all the routes in the state, the use of CARS database can be applied to smaller, more focused studies providing even more detailed analysis. 7
FUTURE WORK
Now that the locations are identified, site investigation and migration patterns would allow a better understanding of common features that may create a more hazardous crossing than others. Future work should be focused on the evaluation of mitigation measures and their effectiveness including fence and overpasses/underpasses. Some of the variables that should be considered are deer population, fence condition, date of installation and maintenance, as well as wildlife behavior and ways to reduce accidents. In relation to domestic animals, work should focus on ways to eliminate the problem since these are much more dangerous based on severity than the wild animal accidents. A better understanding of cause of domestic accidents and the
27
legal ramifications for the State and domestic animal owner would be an area of research to explore.
REFERENCES 1. Schafer, James A., Stephen Penland, and William P. Carr. “Effectiveness of Wildlife Warning Reflectors in Reducing Deer-Vehicle Accidents in Washington State.” Transportation Research Record 1010 (1985): 85-88 2. Darryll J. Dockstader and Peter D. Southall. “Ecopasage Reduces Roadkills” TR News 227 July-August 2003. pp 38-39. 3. Ludwig, John, and Timothy Bremicker. “Evaluation of 2.4 m Fences and One-way Gates for Reducing Deer-Vehicle Collisions in Minnesota. ” Transportation Research Record 913 (1983): 19-22. 4. Schafer, James A., Stephen Penland, and William P. Carr. “Effectiveness of Wildlife Warning Reflectors in Reducing Deer-Vehicle Accidents in Washington State.” Transportation Research Record 1010 (1985): 85-88 5. Zacks, James L. “Do White-Tail Deer Avoid Red? An Evaluation of the Premise Underlying the Design of Swareflex Wildlife Reflectors.” Transportation Research Record 1075 (1986): 35-43. 6. Ministry of Transportation and Highways of British Columbia, Canada. “Wildlife Fencing Program”. Retrieved from on October 6th 2003. 7. F. E. Hartman. “Hunting is Big Business. Pennsylvania Game News, Vol. 44, 1973, pp. 26-31. 8. R. Norman. “Using Wildlife Values in Benefit-Cost Analysis and Mitigation of Wildlife Losses. Proc., International Association of Game, Fish, and Conservation Commissioners, Vol. 56, 1975, pp. 119-128. 9. North Dakota State University * Dickinson Research Extension Center. Retrieved on October 9, 2003 from http://www.ag.ndsu.nodak.edu/dickinso/research/1975/straw75.htm 10. Great southern outdoors. “Alabama Cooperative Deer Management 28
Program & Deer Harvest Report 2002-2003 . Retrieved on October 9, 2003 from http://www.greatsouthernoutdoors.com/Results/results.html
29
8
APPENDIX
Accidents per year at fenced locations (1996-2001)
Num ber of Accidents
Wild Animal Accidents US40 (mp 2-11) 40 35 30 25 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents US89 (mp162-164) 2.5 2 1.5 1 0.5 0 1996
1997
1998
1999 Year
30
2000
2001
Num ber of Accidents
Wild Animal Accidents US70 (mp7-57) 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents US91 (mp 3-17) 50 40 30 20 10 0 1996
1997
1998
1999 Year
31
2000
2001
Num ber of Accidents
Wild Animal Accidents US248 (mp3-13) 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Wild Animal Accidents I-15 (mp 263.5-272) Num ber of Accidents
Num ber of Accidents
Wild Animal Accidents I-15 (mp 26.8-34.2)
2.5 2 1.5 1 0.5 0 1996
1997
1998
1999 Year
32
2000
2001
Wild Animal Accidents I-15 (mp 279.5-284.5) Num ber of Accidents
6 5 4 3 2 1 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents I-15 (mp 61.8-83) 25 20 15 10 5 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents I-15 (mp 273-277) 5 4 3 2 1 0 1996
1997
1998
1999 Year
33
2000
2001
Num ber of Accidents
Wild Animal Accidents I-15 (mp 94-109) 10 8 6 4 2 0 1996
1997
1998
1999
2000
2001
Year
Num ber of Accidents
Wild Animal Accidents I-15 (mp 170-187) 3.5 3 2.5 2 1.5 1 0.5 0 1996
1997
1998
1999 Year
34
2000
2001
