New OECD System Uses Satellite Images and AIS Data to Monitor World Trade

As global trade becomes increasingly vulnerable to wars, sanctions, shipping disruptions, and tariff battles, economists are struggling to keep pace with rapidly changing events. Official trade statistics often take weeks or months to appear, leaving governments and businesses reacting too slowly to economic shocks.

A new OECD study aims to change that. Researchers from the OECD Statistics and Data Directorate, working with Birch Analytics and supported by the United Kingdom’s Department for Business and Trade, have developed a system that can monitor global trade flows almost in real time using ship-tracking signals, satellite imagery, and artificial intelligence. The study argues that maritime data could become one of the world’s fastest economic indicators.

How Ships Became Economic Data Sources

The research is built around the Automatic Identification System (AIS), a maritime safety technology used by ships worldwide. AIS continuously sends signals containing information such as a ship’s location, speed, destination, and draught, the depth to which it sits in the water.

Since more than 80% of world merchandise trade by volume moves by sea, these signals provide a constant stream of information about the movement of goods across the global economy.

But turning ship movements into trade statistics is far from simple. AIS data are often noisy and incomplete. Signals can disappear, vessels can switch off transmitters, and ports themselves are highly complex, handling multiple products simultaneously.

To solve this problem, the researchers created a three-step system that combines maritime tracking data with satellite images and machine learning tools.

AI Learns to Read Ports from Space

One of the study’s biggest innovations is its ability to identify what products are being handled at specific parts of ports.

Instead of analysing entire ports as single units, the researchers focused on individual berths — the exact docking areas where ships load and unload cargo. Using advanced mapping techniques, they identified nearly 30,000 berths across more than 4,100 ports worldwide.

The team then downloaded high-resolution satellite images for these berths and analysed them using Google’s Gemini artificial intelligence model. The AI system was trained to recognise visual features linked to different commodities, including container cranes, oil tanks, grain silos, timber stockpiles, vehicle storage areas, and chemical pipelines.

By combining these visual clues with vessel traffic patterns and international trade databases, the researchers classified trade into 23 commodity groups, including crude oil, chemicals, coal, machinery, textiles, metals, vehicles, and agricultural products.

Tracking Wars, Tariffs, and Supply-Chain Shocks

The system proved capable of detecting major trade disruptions long before official statistics became available.

The study showed how shipping traffic through the Suez Canal collapsed after attacks in the Red Sea disrupted one of the world’s most important maritime routes. It also tracked the sharp fall in cargo movements through the Port of Baltimore after the collapse of the Francis Scott Key Bridge in 2024.

The methodology also captured the effects of geopolitical tensions. Researchers documented the steady decline in Russian oil exports to Europe after the invasion of Ukraine, alongside increasing shipments toward Asian markets.

Another striking example involved Australian coal exports to China. After Beijing imposed informal restrictions on Australian coal imports in 2020, shipments collapsed dramatically. The system later detected trade recovering once restrictions eased in 2023.

The researchers also observed companies rushing imports into US East Coast ports ahead of expected tariff increases in 2025, followed by a sharp decline once the tariffs came into effect.

Fast Signals for an Uncertain World

The study goes beyond measuring trade flows. It also introduces experimental estimates of transit trade goods that pass through countries such as Singapore, Hong Kong, Dubai, and the Netherlands without entering domestic markets.

According to the research, transit trade accounts for roughly 15–17% of trade flows in Hong Kong, around 15% in the United Arab Emirates, 14% in Singapore, and about 8% in the Netherlands.

The authors stress that AIS-based indicators are not designed to replace official trade statistics. The data still suffer from limitations, including missing signals, outdated imagery, and difficulties measuring container trade accurately.

However, the real strength of the system lies in speed. While traditional statistics may take months to compile, AIS-based indicators can reveal disruptions, turning points, and emerging trends almost immediately.

In a world where supply chains can be disrupted overnight, the ability to monitor global trade in near real time may become one of the most valuable economic tools of the decade.