Seismic savanna: machine learning for classifying wildlife and behaviours using ground‐based vibration field recordings. Issue 2 (9th November 2021)

• Record Type: Journal Article
• Authors: Szenicer, Alexandre; Reinwald, Michael; Moseley, Ben; Nissen‐Meyer, Tarje; Mutinda Muteti, Zachary; Oduor, Sandy; McDermott‐Roberts, Alex; Baydin, Atilim G.; Mortimer, Beth
• Editors: Lecours, Vincent; Rovero, Francesco
• Is Part Of: Remote sensing in ecology and conservation. Volume 8:Issue 2(2022)
• Page Start: 236
• Page End: 250
• Publication Date: 2021-11-09
• Subjects: African elephant -- machine learning -- seismic waves -- wildlife monitoring; Remote sensing -- Periodicals; Ecology -- Research -- Periodicals; Ecology -- Methodology -- Periodicals; Ecology -- Remote sensing -- Periodicals; Nature conservation -- Methodology -- Periodicals; 577.0723
• DOI: 10.1002/rse2.242
• Languages: English
• ISSNs: 2056-3485

Abstract

Abstract: We develop a machine learning approach to detect and discriminate elephants from other species, and to recognise important behaviours such as running and rumbling, based only on seismic data generated by the animals. We demonstrate our approach using data acquired in the Kenyan savanna, consisting of 8000 h seismic recordings and 250 k camera trap pictures. Our classifiers, different convolutional neural networks trained on seismograms and spectrograms, achieved 80%–90% balanced accuracy in detecting elephants up to 100 m away, and over 90% balanced accuracy in recognising running and rumbling behaviours from the seismic data. We release the dataset used in this study: SeisSavanna represents a unique collection of seismic signals with the associated wildlife species and behaviour. Our results suggest that seismic data offer substantial benefits for monitoring wildlife, and we propose to further develop our methods using dense arrays that could result in a seismic shift for wildlife monitoring. Abstract : Our planet is facing its sixth mass extinction, with hundreds of species disappearing largely because of human activity. To mitigate this existential threat of biodiversity loss, we must monitor and protect endangered species, from the small golden frog of Panama, to elephants and other megafauna. We present a novel approach using deep learning to detect elephants and classify their behaviours with high accuracy by using seismic field data acquired from a multisensor deployment in Kenya. Our results showcase the promising potential of seismic data to be used in conservation to monitor animals in near real‐time, and alert to behaviours of concern. Crucially, we open‐source SeisSavanna: a unique dataset of seismic signals from wildlife, curated specifically for machine learning applications.