Project information
Project aim
Knowledge of an animal’s large scale movements through the landscape is essential to understanding its ecology. Many far-ranging Antarctic species can only be tracked with electronic tags. These small data loggers either directly record the animal’s location through satellite tracking, or record ambient light levels and the animal’s path is determined from the recorded pattern of day and night when the tag is recovered. Early generations of tags were bulky and expensive, and could only be deployed on larger animals. Technological advances have seen tags decrease in both size and cost, and these tags are playing an increasingly important role in Antarctic Science. This project will develop an integrated suite of tools for estimating animal tracks from tag data, in particular, developing improved methods capable of fully leveraging the data collected by the latest generations of tracking tags.
Project milestones
Summary of the project achievements 2017/18
Knowledge of an animal's large scale movements through the landscape is essential to understanding its ecology, and many far-ranging Antarctic species can only be tracked by attaching electronic tags that record ambient light and other environmental data. The animal's path is then reconstructed from the times at which the animal experienced twilight. This analysis is often made more difficult where the light record is incomplete because the animal has entered a burrow, or has sat or lain on the tag obscuring the sensors. In these cases, it is essential that all sources of data collected by the tag be exploited to pin down the animal's location.
This project has developed an integrated suite of tools for estimating animal tracks from electronic tag data. These new tools can leverage environmental data (such as ambient temperatures) collected by the latest generation of tracking tags to further improve location accuracy, can impose simple geographic constraints on the estimated tracks to match known habitat ranges and migratory behaviours, and can cope with to high latitude tracks where the animal may experience extended periods of day or night.
In particular, a new "twilight free" method has been developed that does not require the times of twilight to be explicitly determined, and is much more robust to interruptions in the light record where the tag's sensors have been obscured.
Summary of the season 2016/17
In this season the analysis methods and software tools developed in the first season have been further refined. The pre-processing tools have been expanded to accept data from a wider range of GLS tags and are now more robust to a range of common problems, and the major analysis methods are now better integrated and simpler to apply in general, making it easier to derive meaningful tracks from GLS data.
Development has also continued on an entirely new method that does not require times of twilight to be identified in the light record, and can produce estimates for data that is so noisy the other methods fail. While not as accurate as the more sophisticated techniques, it has proven to be easier to apply and more reliable than originally anticipated.
There continues to be significant world wide interest in the project, with both national and international research groups using the new methods to analyse tracking data.
Summary of the season 2015/16
This season's work has focused on the development and implementation of several new methods for the analysing tracking data, and we have released two software packages for use with the R statistical computing environment, one for preprocessing data from GLS tags, and the other for deriving tracks from the data collected by the tag.
There has been significant world wide interest in the project, with both national and international research groups using the new methods to analyse their tracking data.