Factors Influencing the Road Mortality of
Snakes on the Eastern Snake River Plain
Transportation lies at the center of our society, linking
destinations, and is ever expanding. A vast network of roads
stretches across our landscape affecting ecosystem processes in
myriad ways. Roads transform existing vegetation into a
compacted earthen surface with altered thermal and moisture
characteristics, and generate an array of ecological effects
that disrupt ecosystem processes and wildlife movement.
Researchers have conducted surveys along roads in attempts to
quantify the most conspicuous effect that roads impose on
wildlife, mortality inflicted by vehicles. In reviewing the
literature, it became apparent that rigorous studies concerning
road mortality of snakes are scarce. Furthermore, studies tend
to be focused in the southeast and southwestern US, with only
three studies conducted in northern latitudes.
However, northern temperate snakes possess several
characteristics that increase their susceptibility to road
mortality. They migrate seasonally to locate specific resources
(Gregory et al. 1987; King and Duvall 1990) such as refuge,
mates, prey and egg-laying habitat (for oviparous species).
These resources tend to be located in distinct habitats that are
patchily distributed across the landscape. Many large-bodied
snake species make a loop-like migration from a communal
hibernaculum (overwintering den site) to summer foraging
habitats (King and Duvall 1990). Seasonal movements are defined
by three distinct phases: 1) egress, or rapid movement away from
the hibernacula, 2) stationary, or periods of short-distance
movements associated with foraging, gestation, or ecdysis, and
3) ingress, or long-distance movements toward the hibernacula as
described by Cobb (1994). The overlap of these movement
corridors with the road network may result in high mortality.
Publications tend to report numbers of fatalities according to
species, but rarely explore the relationship of mortality with
season, sex, or age of individuals.
Road mortality of snakes is a conservation issue that needs
to be addressed. Future research must question if this mortality
has the potential to severely reduce snake populations to a
level where reproductive output cannot replace road-killed
individuals (Rosen and Lowe 1994; Rudolph et al. 1999). The
adverse effects of roads can be minimized, but the correct
placement of mitigation efforts is critical. Ultimately, this
research seeks to identify landscape and road variables that are
highly correlated with snake mortality. These correlations could
then be used to identify areas that may represent high risks for
snake road mortality. Studies suggest that mitigation success is
dependent on correct placement of efforts (Jackson 1999) by
identifying high-risk sites.
This study was designed to address five objectives: (1)
quantify the road mortality of snakes on the eastern Snake River
Plain; (2) identify any variation of mortality with respect to
species, season, sex, age, traffic volume; (3) examine the
spatial pattern of mortality across the survey route; (4)
evaluate the importance of various landscape factors influencing
this pattern; (5) develop a logistic regression model to predict
road sections with intense mortality.
Successful completion of the 2003 and 2004 field seasons
including 333 total road observations of snakes along the survey
route in over 10,000 kilometers driven.
- Performed spatial and statistical analyses of the data, as
well as identification of important landscape and habitat
features that influence where snakes cross roads.
- Presented general findings of this research at the
Intermountain Herpetological Rendezvous in Logan, Utah (2004),
Society for Northwest Vertebrate Biology meeting in Corvallis,
Oregon (February, 2005), and at the International conference
on Ecology and Transportation in San Diego, California
- Successfully defended a master's thesis based on this
- Generated a poster publication to be used for subsequent
Road mortality of snakes was quantified by road cruising
(driving slowly in a vehicle and recording all snakes observed
on a road surface) a 170-kilometer route from May through
October of 2003. The survey route is located within the
northeastern portion of the Snake River Plain and covers
portions of US Highways 20, 26, 20/26, 22/33, Franklin
Boulevard, and Lincoln Boulevard. Sampling consisted of 55 total
trips along this route, and resulted in 9,350 total kilometers
traveled over the 2003 field season (Table 9-4).
A total of 253 snakes were observed on roads along the survey
route and across the entire survey period; 93 percent of these
animals were found dead on the road surface (kill rate of 0.023
individuals/km surveyed). Spatial visualization and analyses
indicate that these observations are clustered along the survey
route (Figure 9-8). We documented the road mortality of 4
species belonging to families Colubridae and Viperidae. However,
the majority of observations belonged to 2 species, Pituophis
catenifer (gophersnake) and Crotalus oreganus lutosus
(Great Basin rattlesnake). We observed gophersnakes most often
on roads, comprising 74 percent of all road records, and
rattlesnakes were observed more frequently than the remaining
two species, comprising 18 percent of all road records (Figure 9-9). Furthermore, we observed more adult males dead on roads
for both these species than any other sex or age class. Juvenile
observations comprised only 28 percent of total gophersnakes,
and 17 percent of total rattlesnake road mortality.
Monitoring data indicate that rattlesnakes are the most
abundant species based on hand and drift fence captures at dens.
In fact, rattlesnakes made up 85 percent of captured snakes
(n=2,459), with gophersnakes representing most of the remaining
percentage of snakes (n=372) over a ten-year sampling period.
This raises an interesting question, are gophersnakes more
susceptible to road mortality on the Eastern Snake River Plain?
This species is a habitat generalist and is perhaps more vagile
than rattlesnakes, indicating that individuals would encounter
roads more often, exposing them to the risk of road mortality.
The road mortality of snakes was documented in all months
surveyed and seasonal patterns were evident. The mean number of
snakes observed per route while road cruising was highest during
the fall season, with a secondary peak in spring. These
differences were significant (analysis of variance [ANOVA], F =
3.638, P = 0.033). The total number of sampling days without
snake observations (11 total) was highest in late July and early
August. There were also significant differences across season
based on sex and age in gophersnakes (Figure 9-10).
Specifically, more males were killed in spring, whereas more
subadults were dead in fall (results based on Kruskal-Wallis
test). The higher numbers of certain age and sex classes with
respect to seasons indicates that individuals may be more
susceptible to road mortality during specific movements. Methods
designed to ameliorate the road mortality of snakes should
therefore coincide with these activity periods to be effective.
In addition to the systematic surveys, a 10 km segment of the
route along State Highway 22/33 (running north/south) located on
the western most edge of the study area was road-cruised in
2004. These surveys were designed to assess the probability of a
snake successfully crossing the road. In attempts to address
this, the shortened segment was driven between June and October
in 2004, during periods of peak snake activity. Twelve of these
routes were surveyed, covering 746 km and 80 snakes were
observed (rate 0.107 snakes/km surveyed). Of these 80 snakes, 59
were observed killed. Similar to the 2003 surveys, 74 percent of
observations were gophersnakes and 23 percent were rattlesnakes.
This high mortality rate occurred in low traffic, and it appears
that a traffic volume of less than ten vehicles per hour was
sufficient to cause 100 percent mortality on some nights.
To assess the effect of road and landscape variables on snake
mortality, several relevant variables were measured at each
observation location from 2003, as well as an equal number of
randomly chosen non-crossing points along the route. The
variables measured were road slope, percent vegetation and major
cover type within 10 meters of the road, distance to nearest
vegetation, distance to nearest shrub, presence of burrows,
presence of basalt, mean distance to dens (including those
identified in this research), solar radiation, and major cover
type at 50, 100, and 500 meters from the road (based on a
geographic information system [GIS] coverage). A multiple
logistic regression analysis was used to determine those
variables significantly associated with road crossing locations.
Finally, only gophersnake locations were used in this analysis
as this was the only species with enough sample locations. Four
variables were consistently included in each significant model.
These were the grass major cover type (positively associated
with snake crossing), percent vegetation cover within 10 meters
(positively associated), presence of basalt (positively
associated), and mean distance to den (positively associated).
These results were surprising because it was expected that snake
presence would be correlated with shrubs and would be negatively
associated with den distance (i.e. the closer a den was to the
road, the more mortality would be expected). A possible
explanation for the positive association of grass cover type
with snakes on roads is that this habitat is less suitable. If
this is unsuitable habitat, individual snakes may move more to
find more suitable habitat, and this increased movement would
increase the likelihood of encountering a road. Second, if
increased proximity of dens to a road actually reduces the
probability of a snake encounter, then this likely indicates a
population effect of road mortality. Specifically, this means
that dens near roads either have reduced numbers of individuals
or that snakes from those dens are not moving towards roads.
Either way, this could influence population connectivity and
ultimately, population persistence.
Investigators and Affiliations
Jochimsen, Graduate Student, Herpetology Laboratory, Department
of Biological Sciences, Idaho State University, Pocatello, ID.
Peterson, Professor, Herpetology Laboratory, Department of
Biological Sciences, Idaho State University, Pocatello, ID.
ISU Biological Sciences Department
ISU Graduate Student Research Committee
BBW Bechtel and the INEEL - ISU Education Outreach Program
ISU Biology Youth Research Program
National Science Foundation (NSF) GK-12 project