Latent Space Atmospheric Particulate Science

Abstract Airborne particulate matter is a major contributor to global illness and originates from a range of natural and anthropogenic sources. These particles can undergo agglomeration, whereby individual units cluster into larger, structurally complex assemblies. This process can span multiple spatial and temporal scales, from nanometer-sized particles interacting over seconds to minutes, to the formation of larger aggregates that form and evolve over hours or days. Understanding how particles agglomerate and transport is essential for accurately predicting their dispersion, environmental impact, and potential health risks. However, traditional analytical models struggle to capture the multivariate, nonlinear nature of these processes. Recent advances in deep learning, in particular the use of latent space representations, offer promising new tools for addressing this complexity. More specifically, latent spaces allow high-dimensional data, such as particle morphology and spatiotemporal measurements, to be projected into lower-dimensional manifolds where hidden structure and dynamics may become more interpretable. In this work, I explore how latent space methods could be applied to model particle agglomeration and its evolution over space and time, providing an alternative framework for understanding aerosol behavior beyond conventional techniques. Significance Statement This work proposes how advanced computational techniques can simplify the complex patterns of airborne particles by converting high-dimensional data into more manageable forms. By employing these methods to identify patterns in how particles group together, new insights into the behavior of air pollution within the atmosphere could be obtained. This is significant because these techniques could lead to models to be developed that can improve air quality forecasts and inform public health strategies. Additionally, this work suggests that integrating these techniques with multiple data sources could further enhance our understanding of how pollutants form and evolve over time in the atmosphere, ultimately helping society better predict and mitigate the impacts of air pollution.