Hiding in Plain Sight: a Less-Explored Secret of Secondary Organic Aerosols

Shrivastava, M., Pacific Northwest National Laboratory

Aerosol Properties

Aerosol Life Cycle

Shrivastava M, C Zhao, RC Easter, Y Qian, A Zelenyuk, JD Fast, Y Liu, Q Zhang, and A Guenther. 2016. "Sensitivity analysis of simulated SOA loadings using a variance-based statistical approach." Journal of Advances in Modeling Earth Systems, 8(2), 10.1002/2015ms000554.


Pollution and particles blown from cities to the foothills of Sacramento mix with natural emissions from trees and vegetation of the forested region, forming secondary organic aerosols. In this study, scientists found a chemical process that rapidly transforms the composition, volatility, and viscosity of these particles that has until now been missing from their models.


Pollution and particles blown from cities to the foothills of Sacramento mix with natural emissions from trees and vegetation of the forested region, forming secondary organic aerosols. In this study, scientists found a chemical process that rapidly transforms the composition, volatility, and viscosity of these particles that has until now been missing from their models.

Science

More than 50% of the fine atmospheric aerosol particle mass is often comprised of organic aerosols, most of which are secondary organic aerosols (SOA), formed by atmospheric oxidation of organic gases. Recent measurements suggest that a major fraction of SOA has volatility that is orders of magnitude lower than previously assumed. It is important to understand which parameters/processes are most important for SOA loadings in models and the model sensitivity to those processes.

Impact

Researchers found that among the seven model parameters tested, rapid particle-phase oligomerization (formation of a molecular complex) is the most influential parameter affecting SOA loadings. However, this oligomerization process is largely missing in most SOA models. The statistical approach used in this study not only estimates the contribution of each parameter to model sensitivity, but provides its statistical significance, and quantifies the relative contribution of each. As the process-level understanding of SOA evolves, this variance-based statistical approach can be applied to a broader set of model parameters. Further, the analysis in this study can determine the influence of newly discovered SOA formation processes and their evolution in the atmosphere.

Summary

A research team at Pacific Northwest National Laboratory used a variance-based statistical approach to efficiently investigate seven uncertain parameters simultaneously related to SOA. They applied a modified volatility basis-set (VBS) statistical approach, using data from a 2010 research campaign called the Carbonaceous Aerosol and Radiative Effects Study (CARES), to seven model parameters in a regional transport model. Four parameters involved human-caused and natural organic compounds emissions that are SOA precursor gases and NOx emissions; two involved dry deposition of SOA precursor gases; and one involved the particle-phase transformation of SOA to a low-volatility state due to a rapid oligomerization process. This process causes smaller molecules to combine and form larger molecules in SOA particles. The rapid oligomerization process was recently discovered through measurements, and is missing in most Earth System Models that simulate SOA. The team performed 250-member ensemble simulations using the regional model, and accounting for some of the latest advances in SOA treatments based on recent work. Using a Generalized Linear Model, they estimated individual parameter and parameter interaction contributions to SOA variance, and their statistical significances.