Deepening our knowledge of the ocean with AI and machine learning

The LoVe Ocean Observatory partners with Capgemini to launch the 4th Global Data Science Challenge and develop an AI-based solution that can more quickly review and analyze large amounts of data drawn from underwater sensors

Client: The Lofoten-Vesterålen Ocean Observatory

Region: Norway

Industry: Life Sciences

Client challenge: The LoVe Ocean Observatory wanted to more quickly sift through the massive amounts of data it gathers from oceanic sensors to better inform the conversation around marine conservation

Solution: The observatory partnered with Capgemini and Amazon Web Services as part of the latest iteration of the Global Data Science Challenge, through which an AI- and machine learning-based solution was developed

Benefits:

  • Faster data processing
  • Identification of additional key points to expand understanding of marine environment
  • Greater ability to quickly provide real-time data

When the state of environmental preservation is considered, it’s essential to remember that around 70% of Earth’s total surface area is covered by water. Of this, the majority is made up by oceans, which play an incredibly significant role in the maintenance of the global ecosystem. In fact, the oceans generate 50% of the planet’s oxygen, absorb 25% of all CO2 emissions, and capture 90% of the additional heat generated from those emissions. Despite their massive importance, the oceans still remain largely unexplored and, as a result, insufficiently understood.

Source: Image courtesy of the Institute of Marine Research and the Lofoten-Vesterålen Ocean Observatory

“The ocean is a key part of our global ecosystem,” says Sissel Rogne, Managing Director of the Institute of Marine Research (IMR). “It is changing rapidly, and we do not always understand how or why. According to the World Ocean Assessment, due to the multitude and complex nature of stressors on the ocean environment, the world is running out of time to save and sustainably manage it.”

The Lofoten-Vesterålen Ocean Observatory (LoVe Ocean), located in northern Norway, seeks to shed light on the marine ecosystem by streaming real-time data to the National Marine Data Center.

“We monitor crucial processes that provide new insights into marine life – like the timing of the spring bloom, the migration of ecological and commercially important fish stocks, and climate change-related parameters including ocean temperature, salinity, pH and CO2,” explains Bjørn Arvid Sætren, Digital Director of the IMR. “This data is used to improve our ability to forecast the overall state of the ocean, the variability in the ocean environment, and human impact.”

Throughout a ten-year period, LoVe Ocean has gathered over 100 terabytes of data from a system consisting of fiber optic subsea cables and sensor nodes that cover a large area from land to deep sea. But while the amount of information has grown, manual review techniques have struggled to keep up with the sheer volume of incoming data. In response, LoVe Ocean decided to utilize automation in order to keep pace and ensure that it provided a more comprehensive picture of the oceanic environment.

Competition leads to innovation

Fortunately, the observatory found a partner with similar values and an aligned mission statement in Capgemini. The partners agreed to use the project as the subject of the annual Global Data Science Challenge (GDSC), an event managed by Capgemini in collaboration with Amazon Web Services (AWS) that sees thousands of experts around the world being set to the task of creating a solution for a data-related challenge.

When the 4th edition of the GDSC launched with the goal of enabling LoVe Ocean to more quickly review and analyze data, teams comprising nearly 1200 total participants competed to develop a concept for an AI-based solution leveraging AWS Technology such as Amazon S3, Amazon SageMaker, and Amazon ECR.

Source: Image courtesy of the Institute of Marine Research and the Lofoten-Vesterålen Ocean Observatory

At the end of the planning and conceptualization period, a panel made up of senior stakeholders from both the Observatory and Capgemini reviewed the presented solutions and determined the final results. The winning team’s proposal included a set of solutions focused on using AI and machine learning to automatically extract data pertaining to a wide variety of unique elements of the marine ecosystem.

“This helped with the overall management of the observatory’s data and the extraction of valuable information that had been impossible to do manually,” says Zhiwei Jiang, CEO of Insights & Data at Capgemini. We gained insights in how to process data collection under extreme circumstances, how to analyze a vast array of time-series data, and how the patterns of bio-marine migration interact with various features of undersea life.”

In addition, the team picked out a number of key points to help further refine the way that LoVe Ocean and its researchers observe changes in the ocean.

“There were several things the team picked out that we’re now looking into,” Sætren says. “For example, whale vocalization, oceanographic and tidal features we hadn’t yet recognized, and the strong influence of weather on the environment.”

Providing oceanic insights

With the solution chosen, LoVe Ocean and Capgemini will now shift their attention from development to implementation. Once the AI- and machine learning-based solutions are successfully rolled out, the organizations expect the data to become far more accessible and to represent a more comprehensive view of the world that lies underwater, as well as humanity’s impact upon it. This comprehensive information will be streamed in real time so that a large community of biologists and researchers can benefit from a broader understanding of the oceanic ecosystems.

Source: Image courtesy of the Institute of Marine Research and the Lofoten-Vesterålen Ocean Observatory

Now empowered to more quickly and thoroughly provide critical data to the marine biologist and conservationist communities, LoVe Ocean will be able to lead the conversation around how the oceans are changing and can be protected. With automation providing the key lever to transform data into a vision of this expansive, largely unexplored part of the planet, researchers will be able to better predict how human behavior and a changing climate will impact marine ecosystems.

As the work done by LoVe Ocean and the IMR are already connected to the UN, this data will go on to play a role in shaping international policy and conservation efforts. By utilizing AI and machine learning, the observatory has actively shaped the future it wants and will continue to deliver a positive impact on the world.

“It’s our purpose to support the sustainable development of the ocean economy,” says Rogne. “It’s the data that helps us achieve this goal. Through this challenge, we hope to ensure comprehensive capacity development and equitable access to data and insights across all aspects of ocean science, and for all stakeholders.”

Featured Testimonials

Zhiwei Jiang, CEO of Insights & Data at Capgemini

This helped with the overall management of the observatory’s data and the extraction of valuable information that had been impossible to do manually. We gained insights in how to process data collection under extreme circumstances, how to analyze a vast array of time-series data, and how the patterns of bio-marine migration interact with various features of undersea life.

Bjørn Arvid Sætren, Digital Director of the IMR

There were several things the team picked out that we’re now looking into. For example, whale vocalization, oceanographic and tidal features we hadn’t yet recognized, and the strong influence of weather on the environment.