Prof. Paul Roe1, A/Prof Susan Fuller1, Dr Kellie Vella1, Dr Anthony Truskinger1
1QUT, Brisbane, Australia
Nature is under enormous pressure from climate change, land clearing, intense bushfires and invasive species. Monitoring and accounting for biodiversity is difficult, yet vital for conservation. Acoustic monitoring is set to revolutionise fauna monitoring by capturing a permanent, direct, scalable and objective record of the environment. However, managing and analysing big acoustic data is difficult and local data collections prevent aggregated continental-scale analysis. Data, analyses, analysis tools and effort need to be shared. However, unlike other big data sciences such as bioinformatics, astronomy and remote sensing, there is no open platform in Australia or elsewhere for ecoacoustic data, and traditional databases and fileservers do not suffice. We are developing an open-source platform, supported by standards and training, so that data can be aggregated, discovered, analysed, shared and mined into the future.
This talk will present the Open Ecoacoustics platform (www.openecoacoustics.org): an Australian Research Data Commons (ARDC) project supporting the collection and analysis of large ecoacoustics datasets. Data can be uploaded manually or by sensors and automatically analysed using high performance computers to detect species or events of interest. Data can be browsed and explored through novel visualisations, and citizen scientists can validate analyses. Analysed data can downloaded and ingested into downstream workflows platforms such as Atlas of Living Australia (ALA) and Ecocommons. We will present the challenges and our solutions to managing and analyzing large ecoacoustics data, which enables new scales of biodiversity monitoring.
Biography:
Paul is a professor of computer science at Queensland University of Technology. For the past 12 years he has been researching how sound can be used for biodiversity monitoring: ecoacoustics. His research group is a world leader in ecoacoustics, it is developing novel human-in-the-loop machine learning techniques to analyse data and innovative data visualisations and summaries. He works with ecologists, data scientists and citizen scientists. He leads the Australian Acoustic Observatory project, a network of 400 acoustic sensors which is collecting sound from all over Australia. This will result in a unique sound archive of over 2 millennia of sound data. Paul has published over 150 papers and received over $12M in competitive research funding.