Tracking marine mammals in 3D using electronic tag data

Research output: Contribution to journalArticlepeer-review

24 Citations (Scopus)


1. Information about at-depth behaviour of marine mammals is fundamental yet very hard to obtain from direct visual observation. Animal-borne multisensor electronic tags provide a unique window of observation into such behaviours.

2. Electronic tag sensors allow the estimation of the animal's 3-dimensional (3D) orientation, depth and speed. Using tag flow noise level to provide an estimate of animal speed, we extend existing approaches of 3D track reconstruction by allowing the direction of movement to differ from that of the animal's longitudinal axis.

3. Data are processed by a hierarchical Bayesian model that allows processing of multisource data, accounting for measurement errors and testing hypotheses about animal movement by comparing models.

4. We illustrate the approach by reconstructing the 3D track of a 52-min deep dive of a Blainville's beaked whale Mesoplodon densirostris adult male fit with a digital tag (DTAG) in the Bahamas. At depth, the whale alternated regular movements at large speed (>1·5 m s-1) and more complex movements at lower speed (<1·5 m s-1) with differences between movement and longitudinal axis directions of up to 28°. The reconstructed 3D track agrees closely with independent acoustic-based localizations.

5. The approach is potentially applicable to study the underwater behaviour (e.g. response to anthropogenic disturbances) of a wide variety of species of marine mammals fitted with triaxial magnetometer and accelerometer tags.

Original languageEnglish
Pages (from-to)987-996
JournalMethods in Ecology and Evolution
Issue number9
Early online date10 Apr 2015
Publication statusPublished - Sept 2015


  • Animal movement modelling
  • Dead reckoning
  • Electronic tag
  • Flow noise
  • Hierarchical Bayesian modelling
  • Track reconstruction
  • Triaxial magnetometer and accelerometer


Dive into the research topics of 'Tracking marine mammals in 3D using electronic tag data'. Together they form a unique fingerprint.

Cite this