Comparing secondary organic aerosols schemes implemented in current chemical transport models and the policy implications of uncertainties

Abstract

Secondary organic aerosol (SOA) constitutes a major component of fine particulate matter (PM2.5) that models must account for to assess how human activities influence air quality, climate, and public health. We characterize the current state of SOA modeling by analyzing eight SOA schemes implemented in five widely used air quality models: CAMx, CMAQ, GEOS-Chem, WRF-Chem and CHIMERE. We performed offline calculations to compare non-aged SOA yields, the effects of SOA aging processes, and the influence of NOx conditions on yields. Our objective is to understand variation rather than to identify a superior scheme. We find significant discrepancies in SOA yields with the ratio of maximum to minimum non-aged yield spans from 1.8 to over 1000, depending upon precursor. The impact of nitrogen oxide (NOx) conditions on SOA yields is also highly variable among schemes. While some schemes include SOA aging, their treatments differ substantially, with some schemes showing large increases in SOA mass, while others exhibit minimal changes. Box model simulations confirmed the substantial discrepancies in predicted SOA concentrations and their responses to precursor emission changes. The substantial differences among current SOA schemes highlight a lack of consensus within the air quality modelling community. Evaluating model simulation results using ambient measurements is unlikely to resolve these discrepancies because uncertainties in SOA formation and precursor emissions are deeply intertwined. The limitations of current SOA schemes should be recognized and acknowledged because model choice can greatly influence predicted SOA concentrations and their evolution, ultimately impacting air quality forecasts, assessments, and regulatory decisions.