Estimating wolf (Canis lupus) population size from number of packs and an individual based model

Authored by Guillaume Chapron, Camilla Wikenros, Olof Liberg, Petter Wabakken, Oystein Flagstad, Cyril Milleret, Johan Mansson, Linn Svensson, Barbara Zimmermann, Mikael Akesson, Hakan Sand

Date Published: 2016

DOI: 10.1016/j.ecolmodel.2016.08.012

Sponsors: Norwegian Research Council (NRF) Swedish Environmental Protection Agency Norwegian Environment Agency Norwegian Institute for Nature Research the County Governor of Hedmark Swedish University of Agricultural Sciences Hedmark University of Applied Sciences Swedish Association for Hunting and Wildlife Management Worldwide Fund for Nature (WWF Sweden) Swedish Research Council for Sustainable Development (FORMAS)

Platforms: C R

Model Documentation: ODD Flow charts

Model Code URLs: https://cran.r-project.org/web/packages/pop.wolf/index.html

Abstract

Estimating wildlife population-size is fundamental for wildlife management and conservation. However, making monitoring of population size less resource demanding while still keeping a high monitoring accuracy and precision remains a recurrent challenge. One proposed alternative to count individuals is to instead focus on counting a segment of the population that is easier to monitor but at the same time well informative on total population size. We show how total population size can be estimated from group counts by using an individual-based population model in a social living species. We developed a wolf (Canis lupus) specific Individual Based Model and used Approximate Bayesian Computation (ABC) to fit this population model to the time series of annual number of packs, reproductions and pairs obtained from Scandinavian monitoring data. Model informative priors were obtained with data from collared individuals by the Scandinavian wolf research project. The fitted model was then used to estimate a conversion factor from number of packs to total number of individuals and to number of reproductions. There was a good fit between the retained simulations by ABC and the observed Scandinavian wolf population trajectory. The fitted simulations returned a conversion factor of 8.0 (95\% CI = 6.62-10.07) from number of packs to total population size and of 1.0 (95\% CI = 0.93-1.12) to number of reproductions in December. A sensitivity analysis revealed that the conversion factor from packs to total population size was positively correlated with pup survival and litter size and negatively correlated with subadult, vagrant and adult survivals. Using an individual based model allowed us to model the full complexity of demographic traits of a social-living species such as the wolf. The flexibility of the model also meant that the conversion factor could be estimated for any month during the year. Our approach to estimate total population size from counts of groups requires having a population model where both individuals and groups are explicitly described and can be applied to other wolf populations and group-living species where counting all individuals over a large area is unfeasible. (C) 2016 Elsevier B.V. All rights reserved.

Tags

Dynamics Simulation-model Inference Social-structure Approximate bayesian computation Counts Indicators Abundance Viability analysis Wolves