EAGER: Exploring Applications of Graph Theory for Improved Understanding and Predictability of Atmospheric Chemistry

University Of Southern California, Los Angeles CA

Investigators

Abstract

This EAGER project focuses on the development of tools for simplifying the representation of complex atmospheric chemical reactions in air quality and climate models. Several mathematical techniques will be used to develop and assess reduced-complexity representations of atmospheric chemical mechanisms to enhance the research community’s ability to study atmospheric chemistry and the role it plays in causing air pollution and climate change. This effort is expected to provide new tools and approaches to characterizing, utilizing, and understanding complex chemical systems in the atmosphere. The two science objectives for this proposed work are: (1) to quantify the structural and dynamical properties of atmospheric chemical reaction mechanisms using novel graph theoretical techniques to enable new scientific insights; and (2) to apply these graph theoretical techniques to support the development and assessment of reduced-form models of atmospheric chemistry, leveraging both traditional graph clustering methods and modern graph machine learning. The first objective will use techniques including motif analysis, path and cycle analysis, and network robustness metrics, while the second objective will use Louvain clustering and graph machine learning techniques. This proposal meets the EAGER criteria because the application of methods derived from graph theory to atmospheric chemical mechanisms development and evaluation is largely untested and has the potential to transform our ability to understand complex and reduced-form chemical mechanisms. The current process of developing and evaluating atmospheric chemical mechanisms is complex, time consuming, and contains substantial subjective decision making. By taking established methods from graph theory, this work could provide new tools and new approaches to characterizing, utilizing, and understanding complex chemical systems. This project is a high-risk, high-reward project as the methods proposed have only been applied to chemical mechanisms in preliminary work and so the success of the proposed work is difficult to predict. This effort has the potential to provide transformative tools that could further our understanding of existing and to-be-developed atmospheric chemical mechanisms. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.