Learning to Simulate Aerosol Dynamics with Graph Neural Networks

TL;DR

This paper introduces GLAD, a graph neural network-based surrogate model for aerosol microphysics that accelerates particle-resolved simulations by learning from detailed models and accurately predicting aerosol dynamics.

Contribution

The paper presents a novel graph neural network framework, GLAD, for efficiently simulating aerosol microphysics, enabling faster predictions while maintaining accuracy compared to traditional models.

Findings

GLAD accurately learns aerosol chemical dynamics.

The model generalizes well across different scenarios.

It achieves faster simulation times than traditional particle-resolved models.

Abstract

Aerosol effects on climate, weather, and air quality depend on characteristics of individual particles, which are tremendously diverse and change in time. Particle-resolved models are the only models able to capture this diversity in particle physiochemical properties, and these models are computationally expensive. As a strategy for accelerating particle-resolved microphysics models, we introduce Graph-based Learning of Aerosol Dynamics (GLAD) and use this model to train a surrogate of the particle-resolved model PartMC-MOSAIC. GLAD implements a Graph Network-based Simulator (GNS), a machine learning framework that has been used to simulate particle-based fluid dynamics models. In GLAD, each particle is represented as a node in a graph, and the evolution of the particle population over time is simulated through learned message passing. We demonstrate our GNS approach on a simple aerosol system that includes condensation of sulfuric acid onto particles composed of sulfate, black carbon, organic carbon, and water. A graph with particles as nodes is constructed, and a graph neural network (GNN) is then trained using the model output from PartMC-MOSAIC. The trained GNN can then be used for simulating and predicting aerosol dynamics over time. Results demonstrate the framework's ability to accurately learn chemical dynamics and generalize across different scenarios, achieving efficient training and prediction times. We evaluate the performance across three scenarios, highlighting the framework's robustness and adaptability in modeling aerosol microphysics and chemistry.