[ivory-search id="3463" title="Default Search Form"]
Search
Register here
Search
Type below to search:
OnlineDiscovery - AI and Climate Science Badge available

Physics-informed machine learning of cloud microphysical processes

  • Register

    * Register (or log in) to the Neural Network to add this session to your agenda or watch the replay

    • Date
    1 April 2026
    Timeframe
    17:00 - 18:00 CET
    Duration
    60 minutes
    • Days
      Hours
      Min
      Sec

    Earth System Models (ESM) encode our knowledge about the physical world, enabling both short-term weather and long-term climate prediction; yet they cannot explicitly resolve many key processes such as clouds, turbulence, and convection. Instead, these processes are represented using simplified parameterizations, which remain a major source of uncertainty in climate projections. While machine learning has recently enabled powerful emulators of model components, physically interpretable parameterizations remain essential for scientific insight, robustness, and reliable prediction in a changing climate.

    This seminar explores how physics-informed machine learning can bridge laboratory measurements, field observations, and high-resolution simulations to improve representations of cloud processes in ESMs. The speaker will present recent work combining neural ordinary differential equations and symbolic regression to discover interpretable laws governing ice crystal growth from sparse laboratory observations. The speaker will also discuss data-driven reduced-order modeling approaches to derive simplified microphysics schemes directly from high-fidelity atmospheric simulations. Finally, the speaker will describe recent work applying generative AI to infer probabilistic atmospheric histories of individual ice crystals observed during airborne campaigns, providing new observational constraints on cloud formation pathways.

    These approaches illustrate how interpretable machine learning can accelerate the discovery, development, and evaluation of physically grounded parameterizations, enabling more accurate and reliable climate prediction.

     

    Session Objectives:

    By the end of this session, participants will be able to:

    • Understand the limitations of current cloud microphysical parameterizations in Earth System Models and why they remain a major source of uncertainty in climate prediction.
    • Explore how physics-informed machine learning methods can uncover interpretable physical relationships from atmospheric data.
    • Identify pathways for integrating AI-driven discovery with Earth System Model development, enabling faster, more robust workflows for improving climate model parameterizations.

     

    Recommended Mastery Level/Prerequisites:

    • Graduate-level understanding of atmospheric science or Earth system modeling.
    • Familiarity with Python and scientific machine learning is recommended but not mandatory.
    • Basic knowledge of numerical weather prediction and climate modeling concepts is helpful.
    Share this session with your network
    • Session starts in
      Days
      Hours
      Min
      Sec
    Add to Calendar
    Discover more from programme stream
    Discovery - AI and Climate Science
    Physics-informed machine learning of cloud microphysical processes

    Physics-informed machine learning of cloud microphysical processes

    1 April 2026 - 17:00 to 18:00

    Related stories

    Are you sure you want to remove this speaker?