Written by Professor Rod Connolly,
Read Time: 467 words, about 5 minutes.
What an exciting time to be a marine scientist! Artificial intelligence (AI) is revolutionising the way we monitor coastal seas.
Smarter sensors and streaming underwater video and sound have set scientists up to measure environmental indicators in more locations, more often, at less cost. AI helps analyse big data and extract actionable data from imagery.
These amazing technological advances are just in time. We’re at the pointy end of decades of rapid loss and degradation of coastal habitats, topped with emerging climate impacts. We’re also in a Covid-induced re-set, which is focussing attention on budgets, and is increasing expectations for environmental sustainability. There is now global impetus for restoring lost habitats, as we begin the United Nations Decade for Ecosystem Restoration (2021-2030). Major new initiatives are restoring lost coral reefs, oyster reefs, seagrass meadows and mangrove forests.
As habitats are restored, monitoring of the health of plant and animal communities is essential for assessing restoration success, and for adapting where, when and how future restoration is done. This is fundamentally important to capturing the desired benefits of ecosystem restoration. And AI will help us get there.
Managing Big Data
Having massive, continuous data streams from many places can be a manager’s dream – so long as management-relevant patterns are quickly able to be determined. Alongside big data, computer intensive machine-learning (ML) methods are continuing to develop, opening up many potential applications. Working with local waterway managers, for example, Dr Ryan Pearson from the Global Wetlands Project (GLOW) team uses ML to calculate light available for seagrass growth purely from remotely collected, publically available data. This critical indicator of the health of seagrass ecosystems can now be ‘measured’ incredibly efficiently and cheaply, all day, every day.
GLOW has also developed FishID software that automatically detects, identifies and counts fish and other animals in underwater videos. The
advent of cheap, reliable underwater cameras has resulted in videos becoming the method of choice for many ecological research and monitoring tasks. Manual processing of videos, however, is slow and expensive. FishID works across diverse ecosystems, from the coral of the Great Barrier Reef, to seagrass meadows, and river estuaries. GLOW also is working with The Nature Conservancy to deploy FishID at new oyster reef restoration sites.
The astounding, rapid step-change in marine monitoring requires new skills in statistical ecology and in interpretation for management application. Students and early-career researchers are being trained in techniques that see them in demand from a wide range of employers who want to adopt new technologies, and realise they need graduate employees who are tech-ready for AI science.
NOTE: Original article from the ARI Magazine, Edition 4. You can read the full edition here – Magazine
Twitter: You can follow Professor Rod Connolly here – @ConnollyLab