1. How Effective Is Road Mitigation at Reducing Road-Kill? A Meta-Analysisby
Trina Rytwinski
Carleton University, Ottawa, Ontario
Co-authors: Kylie Soanes, Jochen A. G. Jaeger, Lenore Fahrig, C.
Scott Findlay, Jeff Houlahan, Rodney van der Ree, Edgar A van der Grift

3. Acknowledgements
This study was supported by the SAFEROAD project, part of CEDR Transnational Road Research Programme Call 2013: Roads and Wildlife, which is funded by the national road administrations of Austria, Denmark, Germany, Ireland, Norway, Sweden, Netherlands and United Kingdom
Natural Sciences and Engineering Research Council of Canada
The Baker Foundation

4. Objective
Conduct a meta-analysis of studies that quantified the relationship between road-kill and a mitigation measure designed to reduce road-kill

5. Why use meta-analysis?

6. Why use meta-analysis?
Overall effect size (i.e., effectiveness)

7. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?

8. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
To what extent do taxa differ in the effectiveness of particular road mitigation measures?

9. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
To what extent do taxa differ in the effectiveness of particular road mitigation measures?
To what extent does study design influence the estimated effectiveness of road mitigation measures?

10. Methods

11. Identification 1st level Screening 2nd level Screening Data extraction
1,497 articles found with initial search
652 articles excluded on title & abstract; 6 missing abstracts
616 articles screened at full-text
1,274 articles screened on title & abstract
223 duplicates removed
476 articles excluded on full-text
140 articles included for extraction
Identification
1st level Screening
2nd level Screening
Data extraction
90 articles excluded at data extraction
50 articles included in meta-analysis
Included

12. Identification 1st level Screening 2nd level Screening Data extraction
1,497 articles found with initial search
652 articles excluded on title & abstract; 6 missing abstracts
616 articles screened at full-text
1,274 articles screened on title & abstract
223 duplicates removed
476 articles excluded on full-text
140 articles included for extraction
Identification
1st level Screening
2nd level Screening
Data extraction
90 articles excluded at data extraction
50 articles included in meta-analysis
Included

13. Identification 1st level Screening 2nd level Screening Data extraction
1,497 articles found with initial search
652 articles excluded on title & abstract; 6 missing abstracts
616 articles screened at full-text
1,274 articles screened on title & abstract
223 duplicates removed
476 articles excluded on full-text
140 articles included for extraction
Identification
1st level Screening
2nd level Screening
Data extraction
90 articles excluded at data extraction
50 articles included in meta-analysis
Included

14. Methods Studies included employed one of three study designs:
Before-After (BA)
Control-Impact (CI)
Before-After-Control-Impact (BACI)

15. Methods Studies included employed one of three study designs:
Before-After (BA)
Control-Impact (CI)
Before-After-Control-Impact (BACI)
Converted estimates into a common measure of association, standardized mean difference (Hedges' d)
positive d = a reduction in road-kill with the road mitigation
negative d = an increase in road-kill with the road mitigation

16. Description of Studies

17. Description of studies # studies
# effect sizes
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

18. Description of studies # studies
# effect sizes
82% from NA
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

19. Description of studies # studies
# effect sizes
90% of studies
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

20. Description of studies # studies
# effect sizes
59% of effect sizes large mammals
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

21. Description of studies # studies
# effect sizes
67% of studies
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

22. Description of studies # studies
# effect sizes
54% of studies were grey literature
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

23. Description of studies # studies
# effect sizes
41% of studies
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

24. Description of studies # studies
# effect sizes
79% of studies
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

25. Description of studies # studies
# effect sizes
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

26. Description of studies # studies
# effect sizes
72% of studies
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

27. Overall trend
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

28. Overall trend
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

29. Overall trend
Overall, mitigation measures reduce road-kill by approximately 40% compared to controls.
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

30. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?

31. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
Are fences with crossing structures more effective than fences or crossing structures alone?

32. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

33. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
86% road-kill
51% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

34. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
No detectable additional reduction in road-kill afforded by adding crossing structures to fencing
86% road-kill
51% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

35. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

36. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
51% road-kill
23% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

37. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
Additional reduction in road-kill afforded by adding fencing to crossing structures
51% road-kill
23% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

38. (a) Are fences with crossing structures more effective than fences or crossing structures alone?
Quantitative evidence that to reduce road-kill, mitigation should include wildlife fencing.
Installing crossing structures alone is not effective for mitigating road- kill.
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

39. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
What mitigation measures are most effective for a given taxon?

40. (b) What mitigation measures are most effective for large mammals?
Crossing structures with fencing and animal detection systems reduced large mammal road-kill much more than wildlife reflectors
83% road-kill
57% road-kill
1% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

41. (b) What mitigation measures are most effective for large mammals?
Current animal detection systems can reduce road-kill, though not as effectively as wildlife fencing.
Little evidence that reflectors reduce road-kill.
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

42. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
Which attributes (if any) of the most common measures are associated with effectiveness?

43. (c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of crossing structures are associated with effectiveness?

44. (c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of crossing structures are associated with effectiveness?

45. Which attributes of fencing are associated with effectiveness?
(c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of fencing are associated with effectiveness?

46. Which attributes of fencing are associated with effectiveness?
(c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of fencing are associated with effectiveness?
Fence length

47. Which attributes of fencing are associated with effectiveness?
(c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of fencing are associated with effectiveness?
Fence length
But not associated with large mammal or amphibian and reptile fence effectiveness in reducing road-kill

48. Which attributes of fencing are associated with effectiveness?
(c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of fencing are associated with effectiveness?
Fence length
But not associated with large mammal or amphibian and reptile fence effectiveness in reducing road-kill
not due to insufficient variation or confounding factors
larger sample size necessary?

49. Which attributes of fencing are associated with effectiveness?
(c) Which attributes (if any) of the most common measures are associated with effectiveness?
Which attributes of fencing are associated with effectiveness?
Fence length
But not associated with large mammal or amphibian and reptile fence effectiveness in reducing road-kill
not due to insufficient variation or confounding factors
larger sample size necessary?
Insufficient evidence to draw conclusions on the association of fence length and effectiveness in reducing road-kill.

50. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
To what extent do taxa differ in the effectiveness of particular road mitigation measures?

51. 2. To what extent do taxa differ in the effectiveness of particular road mitigation measures?
Crossing structures with fencing were most effective at reducing road-kill for large mammals than for other taxa
71% road-kill
74% road-kill
4% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

52. Research Questions
To what extent does road-kill mitigation effectiveness differ among measures?
To what extent do taxa differ in the effectiveness of particular road mitigation measures?
To what extent does study design influence the estimated effectiveness of road mitigation measures?

53. BA and BACI studies are much better able to detect effectiveness of mitigation measures on road-kill than CI studies
3. To what extent does study design influence the estimated effectiveness of road mitigation measures?
54% road-kill
44% road-kill
25% road-kill
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

54. Review limitations Many unanswered questions Why?
Studies lacked suitable data for extraction
Sample size too small
Poor reporting e.g., distance between crossing structures was reported in 32% of studies and length of crossing structures in 24% of studies
Not enough variation in predictor (attributes) variables e.g., fence height

55. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:

56. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:
Mitigation for road-kill should include wildlife fencing.

57. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:
Mitigation for road-kill should include wildlife fencing.
For large mammals, current animal detection systems can reduce road-kill, though not as effectively as wildlife fencing.

58. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:
Mitigation for road-kill should include wildlife fencing.
For large mammals, current animal detection systems can reduce road-kill, though not as effectively as wildlife fencing.
If the goal of a crossing structure includes reducing road-kill, fences must be included.

59. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:
Mitigation for road-kill should include wildlife fencing.
For large mammals, current animal detection systems can reduce road-kill, though not as effectively as wildlife fencing.
If the goal of a crossing structure includes reducing road-kill, fences must be included.
There is little evidence that other measures aimed at affecting driver or animal behaviour (e.g., reflectors) reduce road-kill.

60. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:
Mitigation for road-kill should include wildlife fencing.
For large mammals, current animal detection systems can reduce road-kill, though not as effectively as wildlife fencing.
If the goal of a crossing structure includes reducing road-kill, fences must be included.
There is little evidence that other measures aimed at affecting driver or animal behaviour (e.g., reflectors) reduce road-kill.
Studies should incorporate data collection before the mitigation is applied.
a minimum study duration of 4 years for Before-After, and a minimum of either 4 years or 4 sites for Before-After-Control-Impact designs

61. Conclusions
Key points of consideration when at least one of the goals of mitigation is to reduce road-kill:
Mitigation for road-kill should include wildlife fencing.
For large mammals, current animal detection systems can reduce road-kill, though not as effectively as wildlife fencing.
If the goal of a crossing structure includes reducing road-kill, fences must be included.
There is little evidence that other measures aimed at affecting driver or animal behaviour (e.g., reflectors) reduce road-kill.
Studies should incorporate data collection before the mitigation is applied.
a minimum study duration of 4 years for Before-After, and a minimum of either 4 years or 4 sites for Before-After-Control-Impact designs

62. Insufficient information 36.5%
1,497 articles found with initial search
652 articles excluded on title & abstract; 6 missing abstracts
616 articles screened at full-text
1,274 articles screened on title & abstract
223 duplicates removed
476 articles excluded on full-text
140 articles included for extraction
Identification
1st level Screening
2nd level Screening
Data extraction
90 articles excluded at data extraction
50 articles included in meta-analysis
Included
Duplicate data sets 46.5%
Too small sample size 17%
Insufficient information 36.5%

63. Methods: Data analysis
For Hedges' d:
positive d = a reduction in road-kill with the road mitigation
negative d = an increase in road-kill with the road mitigation
% road-kill decrease:
Note: since percent change cannot be computed when means of group 1= 0, we added a small constant q = 0.01 to group 1 means for each effect size estimate within the dataset

64. Results: Description of studies
(Rytwinski et al Reprinted with permission from © 2016 PLOS)

65. (Rytwinski et al. 2016 - Reprinted with permission from © 2016 PLOS)

66. To better facilitate quantitative reviews
Make all monitoring data available e.g., raw data in an appendix or data archiving site
Road-kill data should be reported for each year before and after implementation, and for each control and impact site separately
Authors should clearly distinguish before, during, and after mitigation implementation periods
Provide comprehensive information on:
study locations
study designs
road(s) and traffic
attributes of mitigation measures

67. Fence-end issue
Of the 25 studies that involved fencing (with or without crossing structures), only six (corresponding to 22 of 58 effect size estimates) measured road-kill beyond the fence-ends.
Our results suggest that studies not accounting for fence-end issues may overestimate effectiveness:
We observed larger average effect sizes for studies that did not measure road- kill beyond the fence-ends [1.14 (95% CI: 0.70, 1.58), n = 36; 59%] compared to studies that did [0.73 (95% CI: −0.03, 1.49), n = 22; 46%]