Methods for Estimating the Wolf Population in Minnesota: Are They Reasonable?
by Bill Route
Estimating Wolf Numbers in Small Study
Areas
Most wolves live in packs and roam an area called a territory. To
count wolves in relatively small study areas, say 1,000 square miles,
biologists keep at least one wolf radio-collared in each pack over
winter. This allows the biologist to fly over the study area to
locate the collared wolf and count the members of its pack. Because
some wolves disperse (leave or join other packs) and some die, it
usually requires collaring two or more wolves in each pack to maintain
radio-contact. This takes time and money.
Winter is best for counting wolves in the north because there
are no leaves on the trees and the background is white with snow.
It takes several flights to get a good count of the wolves, usually
one or two flights each week, because packs sometimes split up and
not all members are observed each time. But eventually, a good count
of the number of "pack wolves" in the study area is obtained. The
biologist must then add an estimate of the number of lone wolves
in the area by calculating the proportion of collared wolves that
are not associated with a pack.
It takes a lot of time, money, and effort, but the procedures
are straightforward and result in very accurate estimates. In fact,
estimates for wolves in small areas are among the most accurate
of counts when compared to other species, primarily because wolves
live in packs and are territorial.
Extrapolating Abundance Over Large Areas
It is not reasonable to radio-collar wolves in every pack in Minnesota,
so biologists have used a three step process. The steps have varied
a little over the years, but the 1997-98 survey is a good example
of the process.
First, state biologists obtained estimates from all areas where
wolves were studied using radio-collars as explained above. Five
different studies involving 36 wolf packs provided estimates of
average pack size and average territory size.
Second, they needed to estimate the area of the state occupied
by wolves. To estimate this "wolf range", 464 natural resource
personnel from 179 field stations mapped where they observed wolves,
saw wolf tracks or scats, or heard wolves howling during the winter.
The observers were also asked to note whether their observations
were of packs or single wolves. During the 1997-98 survey, 3451
observations were mapped. Many who worked in the north or northeastern
part of the state saw sign of wolves often. Those who lived in
the southern or far western part of the state rarely or never
saw sign of wolves.
All of the mapped sightings were combined with United States
Department of Agriculture's wolf depredation data and the telemetry
study data to create a single map. This map provided a picture
of where wolves were known to exist and where wolf sign was
never observed. But, even with hundreds of people in the field,
portions of the state were left uncovered. To compensate for
the uncovered areas, biologists used a computer to identify
the townships that had low human- and road-densities, and therefore
probably contained wolves. The human- and road-density threshold
they used was based on previous research. The map showing low
human and road density merely filled in the holes where observers
had not been able to go. Finally, the area covered by major
roads, large lakes, towns and cities was subtracted to provide
an estimate of the total area believed to be occupied by wolves.
The third and final step, was to calculate the average wolf
density found in study areas (from data in step one) and then
multiply this number by the estimate of wolf range (step two)
for an estimate of the state's total population. The method was
slightly more complicated because statistical "confidence intervals"
were calculated to provide a range of estimates, but these are
the basic steps for arriving at the 1997-98 estimate of 2,450
wolves (range 1,995 to 2,905).
Potential Problems
One potential for for error is that some observers could have misidentified
coyote or dog sign for wolf sign resulting in an overestimate of
the wolf range. The authors of the 1988-89 study stated that this
was indeed a concern, but believed most of the people who worked
on the survey were knowledgeable enough to provide good data.
A second potential for error is the use of the human- and road-density
model to evaluate areas not traversed by observers. This model identified
the Superior National Forest as wolf habitat in 1988-89, and nearly
three decades of wolf research proves it is inhabited by wolves.
However, observers were unable to traverse the more remote areas,
so that without the human- and road-density model, they would have
been omitted. In this case, and in many others, it seems reasonable
to include the unobserved areas.
Comparisons With Estimates Done Elsewhere
Wolf populations are estimated each year in Wisconsin and Michigan
using similar methods. Wolf range is relatively small in those states,
so they can afford to do it annually and can radio-collar a greater
proportion of the wolf packs. This results in less reliance on sign
surveys and probably greater accuracy. Minnesota's wolf population
is roughly five times larger than these two states combined, that
level of effort is probably not feasible for Minnesota.
Similarly, many packs are radio-monitored in the reintroduced or
recolonizing wolf populations in the western United States and can
be followed more frequently. Biologists there have already indicated
they will resort more and more to sign surveys as their populations
grow.
In 1997 the International Wolf center evaluated the status of wolves
around the world and obtained wolf population estimates from biologists
in 34 countries. Of the 28 countries that reported methods, 24 (86%)
used sign surveys as a primary method, while three simply used "expert
estimates" and one used wolf-harvest data.
What Do You Think?
When designing a study to estimate the abundance of any species,
a biologist must take into account:
the resources available (staff time and money);
the nature of the beast (some species are more difficult to
count than others);
the size of the area that needs to be studied (in this case
over 30,000 square miles in just the northeastern area
of the state);
and how critical the estimate is for management of the species
(in this case a species being examined for delisting from the
federal endangered species list).
OK, you are the wildlife manager. Do you think the methods used
for estimating wolf numbers in Minnesota are reasonable?
