Research

Climate Modeling with Generative AI

Climate Modeling with Generative AI
Prof. Tian Zheng | Columbia University
Developing high-resolution climate models using AI to predict local impacts such as flooding and extreme heat.

Challenge

Climate change is already driving significant impacts across New York, with projections estimating $10 billion in annual damages by 2050. Communities are increasingly vulnerable to extreme heat, flooding, and other climate-related risks.

However, existing climate models often operate at a coarse geographic scale, making them insufficient for local decision-making. City planners, emergency responders, and policymakers need detailed, neighborhood-level forecasts to effectively prepare for and mitigate these risks.

Generating such high-resolution models requires enormous computational power and the ability to integrate diverse datasets — capabilities that have historically been limited.

Researcher’s Approach Using Empire AI

The team combines AI with climate simulations and real-world data.
Empire AI enables large-scale simulations and data integration.

Empire AI Enables
  • Neighborhood-level forecasts
  • Long-term projections
  • Faster simulations

Without Empire AI, Models would remain too coarse for practical use.

Potential Impacts
  • Improves disaster preparedness and emergency response
  • Reduces economic losses from climate-related events
  • Supports smarter infrastructure and urban planning decisions
  • Enhances long-term climate resilience across New York