LEAP Wallerstein Panel on AI + Extreme Weather Preparedness

LEAP’s new Panel Series on AI and Extreme Weather Preparedness will foster a continuous exchange of knowledge, providing a space for interdisciplinary dialogue on improving disaster preparedness globally through the use of AI in climate forecasts and modeling, monitoring, and improved communication and dissemination of advisories. The goal of this dialogue is to provide an opportunity for the broader climate and disaster community to come together with climate and data scientists to identify the emerging scientific advances in early warning systems and to learn best practices to improve decision-making and facilitate effective anticipatory action and response. 

Key themes will include:

  • Effective Use of Climate Forecasts. How can forecasts – from weather to seasonal time scales, as well as the emerging role of AI in climate projections – be best utilized for predicting and responding to extreme events?
  • Decision-Making Tools. How can existing tools and data, including AI-backed real-time monitoring, be employed more effectively to protect people and enhance disaster preparedness?
  • Action. What are some strategies for improving communication and engagement to ensure that forecasts are understood and acted upon for real-world disaster mitigation?
  • Long-term Planning. What role does AI play in improving climate prediction, and how can it be harnessed for long-term resilience building and infrastructure planning?

Format: Virtual
Moderated by: Geneva List (Columbia Climate School)

This Panel Series is made possible by generous support from David Wallerstein.

photo credit: Eva Manez, Reuters

Panelists
Gustau Camps-Valls
(University of Valencia)
Félix Francés 
(Polytechnic University of Valencia)
Luis Gómez-Chova 
(University of Valencia)
Ana Ruescas 
(University of Valencia)
*   *   *
Introductory remarks from
Pierre Gentine (Columbia University) and panel moderation by Geneva List
(Columbia Climate School).

FRIDAY || FEBRUARY 28, 2025

FLOODING IN VALENCIA, SPAIN
On October 29th, 2024, Spain’s Valencia region received one year’s worth of rain in eight hours. The resulting flash floods killed more than 220 people and highlighted critical gaps in disaster preparedness—despite having an advanced national weather agency. Key challenges stemmed from deficiencies in local warning communications and urban development on floodplains.

The inaugural event of the new LEAP Panel Series: AI + Extreme Weather will examine this flood event, assessing what disaster preparedness and response mechanisms proved effective and where improvements could have been made—particularly through AI-driven solutions. The discussion will explore the role of AI in enhancing early warning systems, flood mapping, and forecasting to mitigate the impact of future extreme weather events.

Early Warning

photo credit: Erwan Hesry

Panelists
Shruti Nath
(University of Oxford)
Isaac Obai
(World Food Programme)
Grey Nearing
(Google Research)
Josh DeVincenzo
(Columbia University)
*   *   *
Panel moderation by Geneva List
(Columbia Climate School).

FRIDAY || MARCH 28, 2025 

AI FOR EARLY WARNING SYSTEMS AND ANTICIPATORY ACTION
Machine learning and artificial intelligence are ushering in advances in weather and hazard forecasting, resulting in more accurate predictions with longer lead times. Harnessing these improved forecasts can also enhance disaster preparedness and mitigation through improved early warning systems and anticipatory action. Anticipatory action approaches offer a proactive alternative to traditional response efforts that can save lives, livelihoods, and money.

This panel will examine current and future applications of AI in Early Warning Systems and anticipatory action. We will delve into the practical challenges of integrating AI-driven forecasts, addressing key considerations such as integration, trustworthiness, and collaboration. Our discussion will focus on bridging the gap between AI-powered predictions and actionable responses.

Policy

photo credit: Scott Graham

Panelists
Terri Adams-Fuller
(Howard University)
Equisha Glenn
(NYC Metropolitan Transportation Authority)
Miguel-Ángel Fernández Torres
(Universidad Carlos III Madrid)
*   *   *
Panel moderation by Geneva List
(Columbia Climate School).

FRIDAY || April 25, 2025 || 10:00 a.m. EST

AI + DATA FOR DECISION-SUPPORT AND POLICY 
Effective climate disaster management requires access to accurate, localized data and clear, actionable communications. Yet in many regions, high-quality data for downscaled climate and impact modeling remains limited. Even when such data exists, it is often not translated into accessible formats or user-friendly interfaces that support life-saving decisions or long-term resilience planning and adaptation policies.

This panel will explore how data is currently used in disaster management and climate resilience, and examine the emerging role of AI-informed real-time monitoring, predictive models, digital twins and other operational tools to enhance decision-making and planning. Exploring the challenges of data accessibility and usability, we will also discuss how the research community and data providers can better tailor their questions, data, and tools to be more effectively used for decision support.

Fire Hurricane

photo credit: Tyler Rutherford

Panelists
Ali Ahmadalipour
(Google X)
Elena Xoplaki
(Justus-Liebig-University Giessen)
Jorge Peréz Aracil
(Universidad de Alcalá)
Jatan Buch
(Columbia University)
*   *   *
Panel moderation by Geneva List
(Columbia Climate School).

FRIDAY || May 16, 2025 || 10:00 a.m. EST

Register to attend

AI FOR WILDFIRES & HEATWAVES
With boreal summer on the horizon, wildfires and heatwaves are poised to once again dominate headlines. In recent years, forest fires have become more widespread, now consuming twice as much tree cover as they did two decades ago. At the same time, heatwaves—the leading cause of weather-related fatalities in the U.S.—are growing hotter and more prevalent.

This panel will explore emerging AI research and tools aimed at improving our ability to predict, understand, and respond to these events. We’ll highlight Bellwether, a Google X Moonshot project harnessing AI and Earth data to assess severe weather and predict their impacts. We’ll also explore the MedEWSA project, a pan-European initiative advancing integrated multi-hazard early warning and decision support systems, and CLINT, which is developing an AI framework to analyze large-scale climate data for better detection and attribution of extreme events. Join us for a discussion at the intersection of AI, climate science, and disaster resilience.