How AI is Being Used to Combat Extreme Weather

The global shipping industry faces mounting pressure from climate-related disruptions as extreme weather events threaten supply chains worth trillions of pounds every year.

Technology companies including Google and Microsoft are responding with machine learning (ML) weather prediction systems designed to help maritime operators navigate increasingly unpredictable conditions.

Sea transport carries 80% of global trade by volume, making weather forecasting critical for maintaining supply chain operations. Yet recent incidents demonstrate the scale of potential disruption when weather prediction fails.

For instance, the 2015 sinking of cargo vessel El Faro, which could not avoid a tropical storm, resulted in 33 fatalities. The 2021 Suez Canal blockage occurred when container ship Ever Given became grounded during a sandstorm with strong winds.

The Suez incident prevented passage for more than 300 vessels and delayed delivery of approximately 16.9 million tonnes of goods – showing the economic consequences of weather-related shipping problems for international commerce.

Climate data shows the financial impact of extreme weather continues to grow. The US sustained approximately US$182bn in weather-related damage during 2024, with 568 fatalities recorded by the National Oceanic and Atmospheric Administration.

UK heatwaves also contributed to 1,311 excess deaths during the same period.

But how is AI helping?

Google and Microsoft challenging traditional forecasting

Technology companies have developed ML alternatives to conventional weather prediction methods.

Google’s GraphCast system, Microsoft’s Aurora platform and the European Centre for Medium Range Weather Forecasting’s AIFS model demonstrate improved accuracy compared to traditional benchmark forecasts.

These systems use AI techniques to process vast datasets of atmospheric conditions and generate predictions faster than conventional numerical weather models.

ML approaches can identify patterns in historical weather data that traditional physics-based models might miss.

The Met Office, the UK’s national weather service responsible for maritime forecasting, acknowledges the potential of AI-powered prediction systems.

James Shapland, Head of Regulated Transport Services at the Met Office, states the organisation is “investing in next-generation capabilities such as advanced satellite data, innovative AI models, and better ways to share vital safety information with people at sea.”

The Met Office has produced the Shipping Forecast for maritime navigation for 100 years.

James indicates the service plans to supplement traditional text-based forecasts with visual formats: “We have started the journey towards producing visualisable, graphical weather warnings and forecasts to accompany the current textual suite of forecasts and warnings, such as the Shipping Forecast,” he says.

Met Office combining traditional and AI approaches

The integration of AI with established forecasting methods represents a hybrid approach to weather prediction.

Professor Kirstine Dale, Chief AI Officer at the Met Office, explains the organisation’s strategy for combining different forecasting techniques: “I think we’ll have traditional models running alongside AI models so that we are drawing on their combined strengths to enable hyper-localised accurate forecasts, delivered fast, when you need them.”

This approach acknowledges that whilst AI weather models show promise, they have not yet completely replaced conventional forecasting systems.

Some ML models still lag behind traditional methods in certain conditions, though continued development may address these limitations.

Furthermore, the maritime industry is a significant economic sector for the UK.

James says: “Marine services are a cornerstone of the UK’s blue economy and with smarter navigation, more efficient logistics, and better environmental stewardship, we are helping to unlock new opportunities for innovation, trade and sustainability,” he states.

Weather prediction accuracy affects multiple aspects of shipping operations including route planning, cargo handling procedures and crew safety protocols.

Whereas enhanced forecasting capabilities allow shipping companies to make informed decisions about voyage timing and routing to avoid severe weather systems.

The applications extend beyond immediate safety concerns to operational efficiency and supply chain management.

Companies can adjust delivery schedules proactively when forecasts indicate potential weather-related delays. Rerouting decisions based on improved weather intelligence can prevent costly disruptions to international trade flows.

Port operations also benefit from enhanced weather prediction. Extreme weather events including earthquakes can cause billions of dollars in port infrastructure damage, creating bottlenecks that affect global trade networks.

Better forecasting allows port authorities to implement protective measures and coordinate with shipping companies to minimise disruption.

The development of AI weather models reflects broader trends in applying machine learning to complex prediction problems.

As these systems continue to develop, their integration with existing forecasting infrastructure may become standard practice across the maritime industry.

James adds: “This advancement would represent a major leap forward in how we create critical weather information for mariners.”