Kristin and AI: the Difference Between “It Happened” and “It Was Prevented”

The passage of Storm Kristin through Portugal is yet another clear sign of a trend that has ceased to be the exception and has become the norm. Extreme weather events such as storms, floods, heatwaves and wildfires are occurring with greater frequency, severity and impact, reflecting the cumulative effects of climate change.

The figures confirm this reality unequivocally. According to United Nations data, over the past 50 years approximately 11,000 climate-related disasters have occurred globally, resulting in more than 2 million deaths and economic losses exceeding USD 3.6 trillion. In the European Union, between 1980 and 2024, economic losses associated with extreme weather events amounted to approximately EUR 822 billion, with the particularly striking fact that 25% of that total occurred between 2021 and 2024 alone. On average, annual losses increased from EUR 8.6 billion in the 1980s to nearly EUR 45 billion per year between 2020 and 2024, according to the European Environment Agency.

This worsening trend is not merely statistical; it is structural. The current pattern of climate risk exposes deep vulnerabilities in critical infrastructure and in traditional planning and governance models. Power grids, transport systems, communications networks and water supply systems are now subject to stress levels far beyond those for which they were originally designed. Continuing to react only after a disaster has occurred is no longer a viable option. To protect communities and economic assets in a context of growing climate risk, it is imperative to adopt new approaches and tools.

It is in this context that Artificial Intelligence (AI) emerges as a crucial ally in the transition from a reactive logic to a genuinely predictive approach to climate risk management.

Thanks to its ability to process large volumes of heterogeneous data — including meteorological, geospatial and historical data, as well as real-time operational conditions — AI has demonstrated significant gains in forecasting extreme events and strengthening early warning systems, providing authorities and populations with more time to prepare (ECMWF).

For example, advanced deep learning models can anticipate severe storms, floods or wildfires earlier and with greater accuracy than traditional methods, giving authorities and communities more time to activate emergency plans more effectively.

However, the potential of AI goes beyond weather forecasting. One of its most relevant contributions lies in the detection of anomalies and latent failures in critical infrastructure. Through sensors installed on bridges, power lines, substations, gas pipelines or railway systems, intelligent models continuously monitor critical variables such as vibration, temperature, pressure or load. Small statistical deviations can be identified as early warning signs of imminent failure, enabling preventive interventions before an extreme event turns a vulnerability into a collapse.

Moreover, AI techniques are being used to map climate risk zones in urban environments and transport networks by cross-referencing historical data on extreme events with the physical and operational characteristics of infrastructure. These models make it possible to prioritise investments, guide adaptation strategies and strengthen resilience where risk is effectively greatest.

In summary, whether through advanced forecasting models or predictive monitoring and maintenance systems, AI now offers concrete tools to anticipate extreme events and significantly reduce their human and economic impacts. Some estimates suggest that the systematic application of AI to infrastructure resilience could reduce annual losses associated with natural disasters by approximately 15%, representing global savings on the order of USD 70 billion per year by 2050 (Deloitte).

The lesson left by Storm Kristin is clear: climate risk is no longer a future scenario, nor an eccentricity confined to other geographies. It is a real risk in our country. For that reason, integrating AI into systems has become a strategic decision and, increasingly, a collective responsibility to protect lives, territories and economies.

Original portuguese article published in Jornal Económico