In September 2023, large parts of Europe experienced severe flooding that caught many by surprise despite the availability of sophisticated weather forecasting technology enhanced by artificial intelligence (AI). The floods, which devastated numerous regions, were forecasted accurately in terms of rainfall amounts by advanced models, yet the impact and the subsequent response revealed gaps in preparation and communication.
AI has become a significant tool in climate forecasting, deploying various statistical mechanisms to analyse historical data and predict weather patterns more cost-effectively than traditional methods. Techniques employed include machine learning models such as Google-funded GraphCast, which outperform conventional models by using historical data reanalysis. These methods re-evaluate past forecasts with contemporary models, providing a comprehensive historical climate picture.
Challenges remain, however, in the data's utilization, particularly regarding extreme weather events which are becoming more frequent. Such events require accurate prediction and effective communication to mitigate their impacts. Andrew Charlton-Perez, a professor of meteorology at the University of Reading, emphasises that AI models are as effective as the data they are fed. Sparse data input or unusual timing of extreme events can complicate accurate predictions.
Since January 2023, the European Centre for Medium-Range Weather Forecasts (ECMWF) has integrated AI into its forecasting framework, using the Artificial Intelligence/Integrated Forecasting System (AIFS). AIFS enables rapid predictions of significant weather events such as cyclones and heatwaves. The ECMWF's predictions ahead of September's floods proved accurate, anticipating 300-400 millimetres of rainfall according to Thomas Wostal of GeoSphere Austria. Despite this, effectively communicating the severity of these predictions remains crucial. Shruti Nath from Oxford University highlights the importance of setting clear, communicative warnings to help communities appreciate the need for preparedness.
The phenomenon underscores broader concerns about climate readiness in Europe. A report from the European Environment Agency (EEA) warns of the escalating risks of extreme weather impacting lives across Europe, even under optimistic climate scenarios. The European Commission's Janez Lenarcic noted that such severe weather events, previously rare, are becoming increasingly common. This new reality poses significant challenges to policies and adaptation measures within Europe.
Jonas Torland, co-founder of 7Analytics, advocates for improved readiness, noting that European governments lag behind their U.S. counterparts in assessing environmental hazards. He states that substantial investments are often made without robust data-supported decision-making. His views encapsulate a growing concern within the tech community about the readiness of European governments to embrace advanced AI solutions. Torland points out that the processing demands of complex AI models—requiring frequent updates to remain accurate—are significant, both in terms of computing power and environmental impact.
Top tech firms like Microsoft and Google are exploring nuclear power alternatives to reduce the energy footprint of their data centres. Other experts, such as Imperial College London's Friederike Otto, suggest that beyond technological advancements, physical adaptations need investment. Initiatives like creating floodwater storage areas, implementing early warning systems, and reducing development in vulnerable areas are crucial steps. Otto argues that these measures, combined with a commitment to reducing fossil fuel use in favour of renewable energy, are essential for limiting the negative impacts of climate change.
The calls for increased action and investment highlight ongoing challenges in aligning advanced technological capabilities with practical and political frameworks to effectively mitigate extreme weather consequences in Europe.
Source: Noah Wire Services