Impacts of Automated Isoprene Chemical Mechanism Reduction on SOA Chemistry and Air Quality in GEOS-Chem

BENJAMIN YANG, Forwood Wiser, V. Faye McNeill, Arlene Fiore, Daniel Westervelt, Columbia University

     Abstract Number: 634
     Working Group: Remote and Regional Atmospheric Aerosol

Abstract
Biogenic volatile organic compounds, chiefly isoprene, are a significant source of secondary organic aerosols (SOA) and other air pollutants in the eastern U.S. during summer. The chemical mechanism is highly complex, thus difficult to manually update and often the most computationally-intensive component of global 3-D chemical transport models including GEOS-Chem. Employing our novel, graph theory-based Atmospheric Chemistry Model Reduction (AMORE) algorithm, we condense the full-chemistry GEOS-Chem mechanism involving gas-phase isoprene oxidation by 33 species (-11%) and 161 reactions (-18%). Sensitivity simulations in which isoprene emissions are artificially removed in the model are compared to a base case with both the AMORE and default mechanisms at 2° x 2.5° horizontal resolution in order to quantify changes in ozone, SOA, fine particulate matter (PM_2.5), nitrogen oxides, among others. Specifically, U.S. and global biogenic isoprene emissions are each set to zero in order to determine their fractional contributions to air pollution via both mechanisms. We utilize ground-based, airborne, and sonde measurements, such as from the Long Island Sound Tropospheric Ozone Study (LISTOS) campaign, to evaluate our base case AMORE and default GEOS-Chem mechanism simulations. Initial tests reveal that the wall time of a one-month GEOS-Chem classic simulation run on 32 cores on a single node is reduced by 1-2 hours (from 17-18 hours to 16-17 hours), while AMORE estimates slightly higher annual surface PM_2.5 by 0.2 µg m^-3 and ozone by 1.9 ppb on average in the eastern U.S. relative to the default mechanism. We aim to make AMORE open source and sufficiently flexible to allow for timely future updates and application to other chemical systems (e.g. aqueous-phase) for efficient air quality forecasting, research, and management.