What the Strah Polls say about roadkill
From ANIMAL PEOPLE, March 1997:
MENTOR, Ohio––Shocked at the
carnage and curious about the impact on local
wildlife, transportation department employee
Cathy Strah some years back began counting
the roadkills collected for disposal by the
town crews of Mentor, Ohio. In 1993 Strah
began sending her data to ANIMAL PEOPLE,
as a participant in a single-year national
roadkill census we were doing, concurrent
with the separate start of the now nationally
recognized Dr. Splatt counts. The latter are
done by middle school students across the
U.S., coordinated by Brewster Bartlett, a
science teacher at Pinkerton Academy in
Essex, New Hampshire.
The Strah Polls have since accumulated
into a unique data base affording some
of the best information available about roadkill
frequency. The Strah Polls cover about
3,000 roadkills to date, a third as many as the
Dr. Splatt counts, but have been compiled in
a more consistent manner, and come from a
suburb of 47,358 residents which may be as
representative of the U.S. as any one community
could be. Cleveland is to the south,
Lake Erie to the west, and open countryside
to the north and east. As Strah Polls cover a
single locale and the Dr. Splatt counts mainly
cover similar habitat, projecting national
roadkill estimates from them is problematic,
but the Strah numbers are easier to use: just
multiply her totals by .0002, Mentor’s share
of the total U.S. population, to get a rough
idea of the percentage of driving done to produce
the roadkill count.
The Strah Polls have the further
virtue of counting roadkills year-round. Until
last year the Dr. Splatt data came only from
the last week of March through the last week
of May. Bartlett recently led an experimental
fall count, but multi-year fall Dr. Splatt data
is still unavailable.
Listed below are all the species that
account for at least 1% of cumulative roadkills
in the Strah counts, with percentages of
their deaths stated for each peak month. The
percentages for any month not listed fall
within the range of probable random variance.
The percentages of total animals killed
in each month, the bottom line, add up to
111% because of the cumulative effect of
rounding off numbers.
The “CS” column states Strah’s
findings as to the percentage of total roadkills
accounted for by each of the most-hit species.
Except for geese in July/August 1993 and
opossums in September 1994, these percentages
have been remarkable steady. If roadkills
of one species decline, all decline moreor-less
in proportion. To some extent this
reflects the availability of shared plant food
sources, and the role of raccoons and opossums
as roadkill scavengers. But the severity
of winter seems to be the strongest factor.
The Strah Poll projected 105 million roadkills
in 1993, 137 million in 1994, and 156 million
in 1995, plunging to 91 million in 1996.
Bartlett has observed similar fluctuations.
A harsh winter, like the winter of
1995-1996, knocks down wildlife populations,
and roadkills decline accordingly.
The “DS” column states the percentage
of total roadkills each species made
up in the 1993-1995 Dr. Splatt roadkill
counts. As the largest body of roadkill data
ever compiled, recording 10,510 roadkills
during the three years for which we have
final totals, and perhaps as many as 15,000
overall, the Dr. Splatt counts form a useful
crosscheck of the reliability of the Strah
Polls. Evident discrepancies probably reflect
the absence of the fall peaks in roadkills of
raccoons, skunks, and woodchucks from the
Dr. Splatt findings; the greater number of
squirrels in the Dr. Splatt data may reflect the
more wooded habitat of New England.
ROADKILL DISTRIBUTION, 1993-1996
Species CS DS Jan Feb Mar Apr May Jun Jly Aug Spt Oct Nov Dec
Cats 4% 5% 14% 10% 10% 16%
Deer 1% 1% 15% 15% 15% 15% 15%
Geese 4% 37% 27%
Opossums 7% 6% 20%
Rabbits 8% 3% 19% 15% 16%
Raccoons 19% 9% 24% 14% 18% 16%
Skunks 12% 6% 16% 20% 20% 14%
Squirrels 25% 34% 18% 18% 11%
Woodchucks 3% 1% 38% 24%
83% 68% 3% 5% 14% 15% 13% 10% 12% 13% 14% 7% 2% 3%
Species observations
Cats have three peaks of vulnerability
to cars: March, coinciding with spring
mating; June, when “kitten season” peaks
and second mating occurs; and September/
October, when second litters are born.
Deer are most often roadkilled in
early summer, when fawns begin following
their mothers, and in hunting season, coinciding
with the rut. Hunters like to pretend
that the rut rather than their own pursuit is
the greater factor in bringing about nearly
half of all deer/car collisions during hunting
season, but the Mentor numbers show that
deer/car collisions continue at a high level
into December, after the rut is over and an
average of 65% of the Ohio buck population
has already been shot, but various special
deer seasons remain open. Deer data provides
another crosscheck of the Strah Polls,
since the Ohio Division of Wildlife also
keeps a count of deer/car collisions. For any
one year, the Strah Poll deer projection to all
of Ohio is far off, obviously due to random
fluctuation in relatively small totals––but for
1994 and 1995 combined, the Strah Poll
projection is only 9% below the ODoW
count. The 1996 ODoW count isn’t available
yet.
Canada geese, who in Mentor are
residents rather than migrants, get thwacked
during the midsummer dispersal of families––and
this shows up even when the
abnormally high death tolls from
July/August 1993 are subtracted from the
data.
Opossums had an exceptionally
hard month around Mentor in September
1994, for no apparent reason. Otherwise,
opossum roadkill mortality seems evenly distributed
throughout most years.
Rabbit roadkills peak early, as
rabbits are among the first young mammals
to leave their dens. Roadkills are only a
minor source of rabbit mortality, however:
foxes, typically born in April, and coyotes,
typically born in May, soon push rabbit populations
into steep decline.
Raccoons are most heavily hit in
July, when mothers escort their litters on
their first ventures from the den. This coincides
with corn ripening, a favorite raccoon
food, and, as raccoons avidly scavenge
roadkills, with the peak in availability of
woodchuck carcasses. Scavenging in turn
puts raccoons at risk of becoming roadkills.
As raccoon families tend to remain together
when a member is hit, whole families are
often killed in one place. Raccoon roadkill
mortality stays high through the fall dispersal
of litters, coinciding with peak availability
of roadkilled squirrels, but falls abruptly in
winter. As raccoons do not hibernate, it
may be that only the brightest raccoons survive
through the first snowfall. Raccoon
numbers might also be depressed by then by
the summer/fall mortality. Fur trapping further
depresses raccoon numbers each winter.
Skunks are most often hit in the
four months between the late summer exodus
of almost mature skunk kittens from dens to
forage alone, and the beginning of hibernation.
Although skunk families also tend to
remain together even after one member is hit,
there is no visible spike in roadkills during
the time that kittens are with their mothers.
Squirrel roadkill mortality seems
to coincide entirely with the availability of
food on the ground, and not at all with
reproductive cycle. Unfortunately, the data
collection method doesn’t permit us to distinguish
roadkill patterns by squirrel species,
but there also seems to be no reason to
believe that the patterns pertaining to grey
squirrels and red squirrels might be different.
Woodchuck roadkill mortality
peaks first with dispersal of litters
