Soaring with Eagles, Swimming with Sharks: Measuring Animal Behaviour with Hidden Markov Models


Around the world there are concerns over the impacts of land use change and the developments (such as wind farms). These concerns have led to the implementation of tracking studies to better understand movement patterns of animals. Such studies have provided a wealth of high-resolution data and opportunities to explore sophisticated statistical methods for analysis of animal behaviour.

We use accelerometer data from aerial (Verreaux’s eagle in South Africa) and marine (blacktip reef shark in Hawai’i) systems to demonstrate the use of hidden Markov models (HMMs) in providing quantitative measures of behaviour. HMMs work really well for analysing animal accelerometer data because they account for serial autocorrelation in data. They allow for inferences to be made about relative activity and behaviour when animals cannot be directly observed too, which is very important.

In addition to this, HMMs provide data-driven estimates of the underlying distributions of the acceleration metrics – and the probability of switching between states – possibly as a function of covariates. The framework that we provide in ‘Analysis of animal accelerometer data using hidden Markov models‘ can be applied to a wide range of activity data. It opens up exciting opportunities for understanding drivers of individual animal behaviour.

The following images provide an inside view into the ecosystems in which the Verreaux’s eagle and blacktip reef shark reside.

Soaring with Veraux’s Eagles

Swimming with Blacktip Reef Sharks

To find out more, read our Methods in Ecology and Evolution article ‘Analysis of animal accelerometer data using hidden Markov models’.

New podcast and video

In case you haven’t seen them, this month we have published a new podcast and video so far.

In our latest video, David Warton, The University of New South Wales, Australia, presents his ‘mvabund’ package on multivariate analysis. What makes this software different from other ones on multivariate analysis, is that it’s all about models that you can fit to your data. David explains how to look at the properties of your data and the common pitfalls in modelling multivariate data. He also goes through how to fit generalised linear models to your data. Do check David’s dancing!

Mvabund is a free application.

Movement ecology and habitat selection in human resource users

In their podcast with slideshow, Sarah Papworth and Nils Bunnefeld, Imperial College London, applied ecological methods and principles to GPS data on human movement to investigate the differences in movement ecology and habitat selection in human hunters and non hunters who return to a central place. Please note this is an mp4 file, to listen or download the mp3 file of the podcast click here.


Latest papers online

In the past week, MEE has been at the ITN Speciation conference in Jyväskylä. As a result, journal updates have been slower than usual. So here is a quick overview of the new papers available online during the past week:

Research papers:
Movement ecology of human resource users: using net squared displacement, biased random bridges and resource utilization functions to quantify hunter and gatherer behaviour
Sarah K. Papworth, Nils Bunnefeld, Katie Slocombe and E. J. Milner-Gulland
This paper is accompanied by a podcast. Follow this link if you use a Mac to access the podcast.

Modelling dispersal: an eco-evolutionary framework incorporating emigration, movement, settlement behaviour and the multiple costs involved
Justin M. J. Travis, Karen Mustin, Kamil A. Bartoń, Tim G. Benton, Jean Clobert, Maria M. Delgado, Calvin Dytham, Thomas Hovestadt, Stephen C. F. Palmer, Hans Van Dyck and Dries Bonte

Designing a benthic monitoring programme with multiple conflicting objectives
Allert I. Bijleveld, Jan A. van Gils, Jaap van der Meer, Anne Dekinga, Casper Kraan, Henk W. van der Veer and Theunis Piersma

Application – as you know, all our applications are free:
mvabund– an R package for model-based analysis of multivariate abundance data
Yi Wang, Ulrike Naumann, Stephen T. Wright and David I. Warton