Understanding Ice Crystal Habit Diversity with Self-Supervised Learning

2025/09/09 by Joseph Ko, Ko, Joseph, Hariprasath Govindarajan +11
Environmental Science · Earth and Planetary Sciences · #Atmospheric aerosols and clouds #Atmospheric chemistry and aerosols #Cryospheric studies and observations

paper · pdf · doi:10.48550/arxiv.2509.07688

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.

Citations

• DINOv3
• On Partial Prototype Collapse in the DINO Family of Self-Supervised Methods
• Automatic Data Curation for Self-Supervised Learning: A Clustering-Based Approach
• DINO as a von Mises-Fisher mixture model
• On Pretraining Data Diversity for Self-Supervised Learning
• Masked Siamese Networks for Label-Efficient Learning
• Emerging Properties in Self-Supervised Vision Transformers
• Emerging Properties in Self-Supervised Vision Transformers
• An Assessment of Earth's Climate Sensitivity Using Multiple Lines of Evidence
• Confronting the Challenge of Modeling Cloud and Precipitation Microphysics
• Revisiting Self-Supervised Visual Representation Learning

Related

• A Machine Learning Framework for Predicting Microphysical Properties of Ice Crystals from Cloud Particle Imagery
• Discovering How Ice Crystals Grow Using Neural ODE's and Symbolic Regression
• Gauge theory approach to describe ice crystals habit evolution in ice clouds
• Understanding cirrus clouds using explainable machine learning
• IceCoder: Identification of Ice phases in molecular simulation using variational autoencoder