AMORE V2.0: A Comprehensive Algorithm for the Automated Model Reduction of Atmospheric Oxidation Mechanisms
FORWOOD WISER, V. Faye McNeill, Siddhartha Sen, Daniel Westervelt, Daven Henze, Arlene Fiore, Julia Lee-Taylor, Kelley Barsanti, John Orlando, Zhizhao Wang, Columbia University
Abstract Number: 307
Working Group: Aerosol Chemistry
Abstract
The Automated Model Reduction (AMORE) version 2.0 Algorithm is a comprehensive algorithm for the reduction of atmospheric oxidation mechanisms of organic species. This algorithm is a follow up to the AMORE v1.0 algorithm presented in Wiser et al., GMD (2023). Unlike AMORE v1.0, AMORE v2.0 creates reduced mechanisms of any size and does not require any additional manual input. The algorithm employs graph theory techniques to measure the yields of all species in the mechanism, sort them by importance, group them when applicable, and remove species while maintaining mechanism accuracy. Reduced mechanisms can be generated within minutes with a specified size for the given application. Accuracy of mechanism properties, such as secondary organic aerosol formation and concentration of low volatility products can be preserved as well. Example reductions are demonstrated for the isoprene oxidation mechanism (430 species) and the larger GECKO-generated camphene oxidation mechanism (180,000 species). This algorithm has broad potential for the reduction of oxidation mechanisms to allow for their use in transport models, which can only utilize smaller mechanisms due to computational constraints.