Understanding Ice Crystal Habit Diversity with Self-Supervised Learning

2025-09-09

AI Summary

Ice crystal morphology exhibits high variability, impeding accurate modeling of cloud microphysical processes and limiting the fidelity of climate system simulations. To address this, we propose the first vision Transformer-based self-supervised pretraining framework specifically designed for ice crystal imagery—requiring no manual annotations and capable of learning robust, interpretable morphological representations from large-scale unlabeled ice crystal images. Our method enables, for the first time, quantitative characterization of ice crystal morphological diversity and significantly enhances representation capability for scientific downstream tasks, including radiative transfer modeling and cloud parameterization. Experimental results demonstrate that the learned representations effectively constrain ice crystal optical properties and scattering characteristics. This advances the physical interpretability and predictive accuracy of ice cloud modules in climate models, providing a transferable, data-driven tool for improving cloud–radiation interaction schemes in Earth system modeling.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Unsupervised & Self-Supervised Learning
• Natural Language Processing: Lexical Semantics and Morphology

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Ice-containing clouds strongly impact climate, but they are hard to model due to ice crystal habit (i.e., shape) diversity. We use self-supervised learning (SSL) to learn latent representations of crystals from ice crystal imagery. By pre-training a vision transformer with many cloud particle images, we learn robust representations of crystal morphology, which can be used for various science-driven tasks. Our key contributions include (1) validating that our SSL approach can be used to learn meaningful representations, and (2) presenting a relevant application where we quantify ice crystal diversity with these latent representations. Our results demonstrate the power of SSL-driven representations to improve the characterization of ice crystals and subsequently constrain their role in Earth's climate system.

Problem

• Modeling ice crystal habit diversity in clouds
• Learning latent representations from ice imagery
• Improving characterization of crystals' climate role

Innovation

• Self-supervised learning for ice crystal imagery
• Vision transformer pre-trained with particle images
• Latent representations quantify crystal morphology diversity