Inferring animal densities from tracking data using Markov Chains

Hal Whitehead*, Ian D. Jonsen

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

12 Citations (Scopus)
8 Downloads (Pure)

Abstract

The distributions and relative densities of species are keys to ecology. Large amounts of tracking data are being collected on a wide variety of animal species using several methods, especially electronic tags that record location. These tracking data are effectively used for many purposes, but generally provide biased measures of distribution, because the starts of the tracks are not randomly distributed among the locations used by the animals. We introduce a simple Markov-chain method that produces unbiased measures of relative density from tracking data. The density estimates can be over a geographical grid, and/or relative to environmental measures. The method assumes that the tracked animals are a random subset of the population in respect to how they move through the habitat cells, and that the movements of the animals among the habitat cells form a time-homogenous Markov chain. We illustrate the method using simulated data as well as real data on the movements of sperm whales. The simulations illustrate the bias introduced when the initial tracking locations are not randomly distributed, as well as the lack of bias when the Markov method is used. We believe that this method will be important in giving unbiased estimates of density from the growing corpus of animal tracking data.

Original languageEnglish
Article numbere60901
Pages (from-to)1-5
Number of pages5
JournalPLoS ONE
Volume8
Issue number4
DOIs
Publication statusPublished - 22 Apr 2013
Externally publishedYes

Bibliographical note

Copyright the Author(s) 2013. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

Fingerprint

Dive into the research topics of 'Inferring animal densities from tracking data using Markov Chains'. Together they form a unique fingerprint.

Cite this