Epsilon Archive for Student Projects

Abstract

Large carnivores cause many conflicts in Scandinavia but depredation on livestock is one of
the most central ones. Free ranging semi domestic reindeer is an important prey for all large
carnivores within the reindeer husbandry area causing a complex carnivore-livestock conflict.
In Sweden the state compensates reindeer owners who suffer depredation caused by protected
large carnivores through a compensation system based upon presence of large carnivores. To
understand the effects a predator has on their prey, estimates of individual kill rates are
needed. This study had two main objectives; 1) to test the GPS-technique as a method for
identifying kill sites and 2) to estimate lynx kill rate on reindeer. We used GPS-collared lynx
within the reindeer husbandry area to evaluate the possibilities to use the GPS-technique to
find kill sites by lynx and to develop a model based on GPS-data to estimate kill rate on semidomestic
reindeer in northern Sweden during winter. The GPS-collars on lynx gave a mean
success rate of GPS positions over the study period of 82.5 % (range=76.5-87.8) and the
proportion of 3D positions (the highest quality) averaged 85.3% (range=81.1-89.8),
suggesting that the GPS-collars provided reliable data for studying kill rate by lynx. Using
GPS-locations and movement patterns for lynx around visited kill sites and non-kill sites, I
developed statistical models to estimate the probability that an unvisited cluster of lynx GPSlocation
was a kill site or not. Kill rate for 3 individuals, during 7.5, 7 and 4.5 months
respectively, were estimated by combining the data from clusters visited in the field and the
statistical model for unvisited clusters. The estimated kill rate including these statistical
models gave a mean kill rate of 4.6 reindeer per month, ranging from 3.1 to 7.1 reindeer per
month. This type of study can contribute to setting appropriate levels of compensation and
thereby reduce conflict levels between reindeer owners and large carnivores.