validation of these results has been attempted. Therefore, other explanations for the
lack of agreement between strikes and bears’ relative abundance could be due to the
field and/or analysis methods used. The survey design has been shown to affect
the estimation of roadkill hotspots for small-bodied species (Santos et al. 2015).
The field data collected may have failed to adequately represent the true distribution of bears along the CPR, since bears may have been moving or utilizing the
railroad without leaving scat, which would have been undetectable using these
methods. Additionally, scat sampling methods may have only represented animals
that regularly encountered the CPR and have learned to coexist, to some degree,
with the railroad. Bears that utilized grain as a food item in this study area may have
developed behaviors that allowed them to forage on grain with reduced strike risk
(learned behavior). Lastly, the analysis scale may have been too small or too large.
Bears have been documented to have large home ranges thus a larger analysis scale
may have been more appropriate. General large-scale agreement between relative
abundance and strikes are apparent in Fig. 9.3d. However, it may be difficult to
implement management solutions at this scale other than those generally aimed at
population distribution, size, or health.
Similarly, the data collected may have failed to fully describe the deer distribution. This may have been why right-of-way width (ROW-width) was positively
correlated to strikes for deer. The hypothesized influence of ROW width was that
narrower ROW’s (a negative relationship) would increase strike risk to wildlife.
This variable may have accounted for spatial or temporal variability in deer
abundance that the transect data failed to capture.
An animal’s behavior when a strike occurs is only partially influenced by
location-specific variables. Other variables that are likely to affect an animal’s
behavior are difficult to measure over long periods of time and large geographic
areas. These variables include: fine-scale temporal variables and behavioral
heterogeneity between individual animals. For this study, strikes spanning 21 years
was used to provide adequate sample sizes for comparison to railroad variables that
have not changed during that time; however, it is clear that strikes have changed over
this period (Fig. 9.4), and thus relative abundance may have also changed during this
period. To address this concern, analyses were repeated using a 10 year period.
Results showed some shifts in the locations of hotspots and high-risk segments, but
they these did not alter the general conclusions. The longer term 21-year analysis
was likely better because of the larger sample size, which provided more information
for hotspots, multivariate modeling, and more stable risk estimates.
Management Implications
This study found the wildlife abundance, train speed, barriers, bridges, and
right-of-way widths affect strike rates for ungulates and/or bears. Based on these
results, a variety of approaches could be used to reduce strike rates. These include
those that affect wildlife abundance near railroads, reduced train speeds, or modified
9 Relative Risk and Variables Associated with Bear and Ungulate …
151
lack of agreement between strikes and bears’ relative abundance could be due to the
field and/or analysis methods used. The survey design has been shown to affect
the estimation of roadkill hotspots for small-bodied species (Santos et al. 2015).
The field data collected may have failed to adequately represent the true distribution of bears along the CPR, since bears may have been moving or utilizing the
railroad without leaving scat, which would have been undetectable using these
methods. Additionally, scat sampling methods may have only represented animals
that regularly encountered the CPR and have learned to coexist, to some degree,
with the railroad. Bears that utilized grain as a food item in this study area may have
developed behaviors that allowed them to forage on grain with reduced strike risk
(learned behavior). Lastly, the analysis scale may have been too small or too large.
Bears have been documented to have large home ranges thus a larger analysis scale
may have been more appropriate. General large-scale agreement between relative
abundance and strikes are apparent in Fig. 9.3d. However, it may be difficult to
implement management solutions at this scale other than those generally aimed at
population distribution, size, or health.
Similarly, the data collected may have failed to fully describe the deer distribution. This may have been why right-of-way width (ROW-width) was positively
correlated to strikes for deer. The hypothesized influence of ROW width was that
narrower ROW’s (a negative relationship) would increase strike risk to wildlife.
This variable may have accounted for spatial or temporal variability in deer
abundance that the transect data failed to capture.
An animal’s behavior when a strike occurs is only partially influenced by
location-specific variables. Other variables that are likely to affect an animal’s
behavior are difficult to measure over long periods of time and large geographic
areas. These variables include: fine-scale temporal variables and behavioral
heterogeneity between individual animals. For this study, strikes spanning 21 years
was used to provide adequate sample sizes for comparison to railroad variables that
have not changed during that time; however, it is clear that strikes have changed over
this period (Fig. 9.4), and thus relative abundance may have also changed during this
period. To address this concern, analyses were repeated using a 10 year period.
Results showed some shifts in the locations of hotspots and high-risk segments, but
they these did not alter the general conclusions. The longer term 21-year analysis
was likely better because of the larger sample size, which provided more information
for hotspots, multivariate modeling, and more stable risk estimates.
Management Implications
This study found the wildlife abundance, train speed, barriers, bridges, and
right-of-way widths affect strike rates for ungulates and/or bears. Based on these
results, a variety of approaches could be used to reduce strike rates. These include
those that affect wildlife abundance near railroads, reduced train speeds, or modified
9 Relative Risk and Variables Associated with Bear and Ungulate …
151