This GitHub repository accompanies our paper 'Novel application of a conservation detection dog to recover wildlife telemetry equipment' (2026). Here you can find:

• The data generated and analysed in the study
• An R markdown script containing the analyses and figure creation - also available as a html document
• The supplementary videos referenced in the paper

To access the supplementary videos, please navigate to the folder '/Supplementary_videos/' and download the files. The videos are too large to display in the browser. If you click into the file you will find an option to 'Download raw file'.

For some additional footage of Koda working, please see this video by the Fenner School at the ANU: https://youtu.be/WlaTaq45NWE?si=SU1sMLGPscJ1qirQ

Please direct any questions to the corresponding author at shoshana.rapley@anu.edu.au

Abstract:

Wildlife tracking technologies are increasingly used to study animal behaviour, inform conservation management, and answer fundamental questions in ecology. However, a drawback to wildlife telemetry is that equipment is expensive and can be difficult to retrieve when lost in the field, leading to loss of data and inefficient use of project resources. As a potential solution, we began training dogs to search for and locate GPS tracking devices by scent. This approach targets the odour of the device itself, rather than the animal carrying the device. Here, we report on the training, deployment and evaluation of a wildlife telemetry detection dog.

During field deployments, the detection dog successfully recovered $11,200 AUD worth of lost GPS devices over a 16-month period. In a controlled field evaluation, the detection dog located 100% of targets compared to 43% by the human-only search. The detection dog was significantly faster, with an average search time of 6.4 minutes (±1.5 SE, n = 17) compared to 43.9 minutes (±4.9 SE, n = 14) for human-only searches. Combining detection rate and search speed, the detection dog was 16 times more efficient.

This is the first reported use of a detection dog to recover wildlife telemetry devices. The detection dog significantly improved device recovery efficiency, with benefits including reduced project costs and improved data completeness. This approach has broad applicability across diverse species and contexts, particularly when visual searching is inadequate such as in dense vegetation or when equipment is buried underground. This novel method offers a practical solution to equipment recovery challenges facing wildlife tracking programs worldwide.

Citation:

(to come)