AI to the rescue for Great Barrier Reef

Developed by the University of South Australia (UniSA) with input from Queensland and Victorian scientists, the platform will absorb data from photographs, videos, satellite images, and marine sensors. Machine learning and AI will be used to interrogate the data, providing information to a central dashboard for real-time monitoring of the world’s coral reefs. 

The UNESCO World Heritage-listed Great Barrier Reef (GBR) – the world’s best known reef and one of Australia’s most heralded ecological and tourism sites - has been decimated by severe bleaching events since 2016, exacerbated by ongoing crown-of-thorns starfish outbreaks and coastal development. Worldwide, 75 per cent of reefs experiencing bleaching-level heat stress in the past two years, driven primarily by the climate crisis and its warming effect on the oceans.

The technology will bring together datasets from organisations like the National Oceanic and Atmospheric Administration (NOAA), the Monterey Bay Aquarium Research Institute (MBARI), the Hawaii Undersea Research Laboratory (HURL) and Australia’s CSIRO. According to the researchers, having a centralised platform that can provide environmental scientists with real-time predictions could be an important tool in the fight to stall the degradation of the world’s reefs.

“At the moment we have separate models that analyse substantial data on reef health – including bleaching levels, disease incidence, juvenile coral density and reef fish abundance – but these data sets are not integrated, and they exist in silos,” said UniSA data analyst Dr Abdullahi Chowdhury, lead researcher on the study, which is published in Electronics.

“Consequently, it is challenging to see the ‘big picture’ of reef health or to conduct large scale, real-time analyses.”

The researchers say an integrated system will track bleaching severity and trends over time; monitor crown-of-thorns starfish populations and predation risks; detect disease outbreaks and juvenile coral levels; and assess reef fish abundance, diversity, length, and biomass.

“By centralising all this data in real-time, we can generate predictive models that will help conservation efforts, enabling earlier intervention,” said Central Queensland University PhD candidate Musfera Jahan, a geographic information systems (GIS) data expert.

“Our coral reefs are dying very fast due to climate change – not just in Australia but across the world – so we need to take serious action pretty quickly.”