74
Purpose
A variety of means have been developed to detect lynx and other carnivores. The purpose of this
protocol is to add reliability, efficacy, and representativeness to the process of lynx detection. Each
element of the protocol has been designed to achieve this end.
Representativeness. While it can be argued that selective sampling (where one goes to the "best" places
and samples) may provide detections at lower cost, the data generated using these methods is much less
valuable. Non-representative surveys at best can provide simple occurrence data. Other more
meaningful metrics: where lynx are present and absent, the habitat relationships of lynx, minimum
viable population estimates, and current range all require representative sampling. Hair-pad methods
were chosen because they allow sampling during the snow-free period, are durable, inexpensive, and
lightweight. A lightweight, inexpensive sampling scheme which could be implemented in the summer
was a necessity for representative sampling. Areas that are dangerous or away from roads will not be
representatively sampled in the winter, and very expensive or high-maintenance detection stations can
only be placed at a few locations. Representative sampling requires unbiased and uniform placement
rules for the sample points. To this end, the protocol is grid-based and uses simple placement rules
which can be applied to most landscapes.
Efficacy. Even if sampling is representative, if detection rates are too low, the method will fail the test of
efficacy. To address this, we tested 5 commercial scent lures on wild lynx in Canada to determine which
lure produced the highest detection rate. While all lures were "hit" by lynx, one lure, a combination of
beaver castorium and catnip oil was twice as effective as the others. Additionally, we made use of
transects to sample lynx in Canada. Over a 2-4 week period, we had hits on nearly half (35/78 = 45%) of
these 5-station line transects. Based on these results, we use line transects and the most effective lure.
