Using AI to map ocean activities

The global economy is increasingly reliant on ocean-based industries, generating over $1.5 trillion annually and supporting 31 million jobs. This economic impact, stemming from sectors like fishing, shipping, and energy production, has grown exponentially in the past 50 years and is projected to double by 2030. However, managing this “blue acceleration” is challenging due to the ocean’s vastness and the complexity of monitoring industrial activities at scale.

A groundbreaking study, involving researchers from Global Fishing Watch, Duke University, the University of California, Santa Barbara, and SkyTruth, addresses this challenge. Published in “Nature,” the study utilizes satellite imagery, vessel GPS data, and artificial intelligence to map human industrial activities across the oceans over five years. This comprehensive public map and dataset reveal a significant amount of activity that goes unmonitored, particularly in fishing.

Key findings include the discovery that about 75% of fishing vessels identified were not tracked by public Automatic Identification System (AIS) monitoring, with a high concentration of such activity around Africa and South Asia. This new data contradicts previous assumptions based on AIS data alone, especially in the distribution of fishing activities between Asia and Europe, and the Mediterranean’s European and African sides.

The researchers also uncovered gaps in AIS tracking, noting that not all vessels are mandated to use AIS, regions with poor reception, and instances where vessels engaged in illegal activities might disable or tamper with their AIS devices. To overcome these limitations, the study integrated artificial intelligence models with satellite-based radar and optical images, and 53 billion AIS vessel position reports.

In addition to fishing vessels, the study also mapped other vessels and offshore structures like oil platforms and wind turbines. It identified about 28,000 offshore structures, noting a modest increase in oil infrastructure and a significant rise in wind turbines, particularly in northern Europe and China.

This data, freely available through the Global Fishing Watch data portal, has several potential applications. It can assist in monitoring fishing activities in data-poor regions, identifying illegal, unreported, and unregulated fishing, and detecting sanction-busting trade. For instance, the data revealed extensive undisclosed fishing in North Korean waters, likely of foreign origin, violating United Nations sanctions.

Furthermore, this dataset can aid in climate change mitigation by quantifying greenhouse gas emissions from maritime activities and evaluating the environmental impact of offshore energy development. It promises to support evidence-based decision-making, making ocean management more equitable, effective, and sustainable. Overall, this research marks a significant advancement in our ability to monitor and manage the ocean’s resources responsibly.

https://theconversation.com/we-used-ai-and-satellite-imagery-to-map-ocean-activities-that-take-place-out-of-sight-including-fishing-shipping-and-energy-development-219367