A path reconstruction method integrating dead-reckoning and position fixes applied to humpback whales

Area of Science:

• Ecology
• Marine Biology
• Animal Behavior

Background:

• Accurate animal location and movement data are vital for understanding behavior and responses to human activities.
• Dead-reckoning alone leads to increasing uncertainty over time.
• Fastloc-GPS technology offers potential for precise georeferenced tracking, especially for marine animals.

Purpose of the Study:

• To develop an efficient Bayesian state-space model for estimating humpback whale locations.
• To integrate dead-reckoning data with Fastloc-GPS and visual observations for fine-scale track reconstruction.
• To quantify positional uncertainty in reconstructed animal movement tracks.

Main Methods:

• A computationally efficient Bayesian state-space modeling technique was employed.
• Dead-reckoning data from on-animal sensors were integrated with Fastloc-GPS and visual position fixes.
• Positional observation error models were developed based on calibration data.

Main Results:

• High-resolution 3D movement tracks were generated for 13 whales.
• Fastloc-GPS significantly reduced positional uncertainty compared to visual data alone.
• Incorporating dead-reckoning data greatly enhanced the detail of reconstructed humpback whale tracks.
• The model showed improved predictability of Fastloc-GPS data via cross-validation.

Conclusions:

• The developed model provides quantitative estimates of location uncertainty for animal movement analyses.
• Systematically accounting for observation errors improves the reliability of movement data.
• This generic method is applicable to various marine species and data recording systems.