30:00

AI Climate Academy - Belém 2025

AI for Disaster Management & Relief


Edier Aristizábal

Departamento de Geociencias y Medio Ambiente, Universidad Nacional, Medellín, Colombia


Disasters

Cause?

Disasters

Weather and climate-related disasters

Floods

Floods impact more people worldwide than any other disaster, and the economic, social, and environmental impacts are getting worse.

Tornado

Hurricanes

Climate change is increasing ocean and atmospheric temperatures and increasing the frequency, duration, and intensity of hurricanes.

Tornado

Landslides

By altering precipitation patterns, leading to more intense rainfall events that saturate soil and increase runoff.

Tornado

Wildfires

Climate change intensifying, hotter temperatures, more intense and longer dry seasons.

Tornado

The number of weather, climate and water extremes are increasing and will become more frequent and severe in many parts of the world as a result of climate change.

Cascade hazards & Multi-risk scenarios

Source: modified after Jakob (2005)

50%

Weather & climate-related hazards

From 1970 to 2019, weather, climate and water hazards accounted for 50% of all disasters.

x5

Number of disasters

It has increased by a factor of five over the 50-year period, driven by climate change and improved reporting.

45%

Deaths & economic losses

45% of all reported deaths and 74% of all reported economic losses.

91%

Deaths

More than 91% of disaster casualties occurred in developing countries.

Source: WMO (2021)

Risk management framework

Applications of AI in disaster management

Source: Sun, Bocchini & Davison (2020)
  • 01

    AI for data: monitoring and detecting

    There are new opportunities for AI in Space and related technologies that can also have an actual application in disaster risk detection, analysis and then reduction.

  • 02

    AI for modelling: forecasting, and projecting

    AI helps accelerate data acquisition and analysis, taking into account the spectrum of heterogeneity owned by each natural hazard and situations.

  • 03

    AI for communication

    AI’s deployment is at the utmost priority to enhance the understanding of all phases of disasters, and this can be done by accelerating the development of algorithms that are reliable for our safety.

Source: EGU blogs, NEXCOM

AI Tools and Example Uses

Tools Description Examples of Commercial Systems Uses in Emergency & Disaster Management
Predictive Analytics Finds patterns in data and forecasts future outcomes. Salesforce Risk modeling; disease outbreak spread prediction; flood/wildfire spread prediction; dashboards and situational awareness
Generative AI and Natural Language Processing Understands and translates human language and creates new text, images, or video ChatGPT, Claude, DALL·E Drafting emergency communication templates; creating scenarios for training; Multilingual crisis communication; rumor detection
Robotics & Automation Performs physical tasks with or without human control, including operating vehicles. iRobot Roomba, Da Vinci Surgical System; Boston Dynamics robots; Waymo Search-and-rescue in dangerous areas; supply delivery; debris clearing
Computer Vision Identifies and interprets objects, people, and activities images/video. Google Photos, Clearview AI; Tesla Autopilot Damage assessment via drones/satellites; search-and-rescue; wildfire smoke mapping
Speech Recognition & Generation Converts speech to text and produces human-like speech from text. Siri, Alexa Voice-to-text for field reporting; hands free operations
Recommendation Systems Suggests products, content, or actions based on user behavior. Netflix, Spotify, Amazon Resource allocation; shelter options; individual risk alerts
Fraud Detection & Security Identifies anomalies to call attention to risks. Mastercard AI Security, Darktrace, PayPal Detecting fraud in payments; cybersecurity
Source: Patrick S. Roberts (2025)

The AI in Disaster Risk Market

Risk Intervention Matrix

Prospective Intervention Actions (Future risk prevention) Corrective Intervention Actions (Current risk mitigation) Vulnerability Reduction Actions Hazard Reduction Actions Inventory Hazard zoning Education and training Vulnerability and risk zoning Policies Early warning systems Response and relief Resettlement Land Use Plans Protection soils Expansion zones Urban control Mitigation measures

AI & Remote sensing data

AI has the potential to speed up our understanding of natural hazards, analysing large volumes of data (and images) from different sources and improve proactive rather than reactive actions for disaster risk reduction (DRR).

Source: Lechner et al. (2020)

Landslide mapping

Source: Lu et al. (2019)

Machine learning techniques

Source: Gudiyangada et al. (2020)

Land use planning

Early warning systems

Source: Reichstein et al. (2025)

Global Observing System

Taken from https://community.wmo.int/en/observation-components-global-observing-system

Interferometric Synthetic Aperture Radar (InSAR)

Source: Geoscience Australia portal

Effectiveness of communication

Taken from https://resiliencechallenge.nz/opportunities-for-early-warning-systems-a-review/

Impact-based forecast and warnings

Taken from https://community.wmo.int/en/impact-based-forecast-and-warning-services

Damage and strategize rescue efforts

Source: xView2, Tate Ryan-Mosley (2023)

Key takeaways

Social Construction of Risk

Disasters are the result of scenarios created by society.

Sustainable Land Use

The best measure is sustainable occupation of the territory in harmony with nature.

AI for hazard mapping

AI tools are essential for inventorying past events and creating more accurate hazard maps with ML.

The Need for EWS

Since risk will always exist, Early Warning Systems are fundamental for managing current scenarios.

AI-Powered EWS

EWS are based on AI-enhanced observation and impact-focused alert dissemination tools.

¡Gracias!

https://edieraristizabal.github.io/Presentaciones/AIclimate_Disaster.html


evaristizabalg@unal.edu.co