Simulation of secondary organic aerosol over the Yangtze River Delta region: The impacts from the emissions of intermediate volatility organic compounds and the SOA modeling framework

Abstract Secondary organic aerosols (SOA) are an important component of fine particulate matter (PM2.5). However, photochemical models often have difficulty capturing the observed magnitudes of SOA due to uncertainties in precursor emissions and modeling approach. In this study, we conducted a modeling study of SOA in the Yangtze River Delta (YRD) region of China to investigate the impact of emissions of intermediate-volatility organic compounds (IVOC) and the SOA modeling schemes. IVOC emissions, which are important SOA precursors but are missing from most emission inventories, are estimated for the YRD region based on two methods. First, scaling IVOC from emissions of primary organic aerosols (POA) estimates 730 Gg with on-road and industry sectors being the main contributors. Second, scaling IVOC emissions using emission factors and activity data estimates 313 Gg, 57% lower than the first method, with industry and off-road sectors being the major contributors. A photochemical model simulation of SOA for July 2018, conducted using standard emission inventories for the YRD region, significantly underestimates SOA by 61% at the Dianshan Lake monitoring site (DSL). A series of simulations with two different SOA modeling schemes found that with the traditional two-product modeling framework, IVOC emissions enhance simulated SOA concentrations by 8%–27% at DSL and 5%–26% over the YRD region in July 2018. Increasing SOA mass yields from IVOC by a factor of five leads to better agreement with observations at the DSL site. Switching to the 1.5-Dimension Volatility Basis Set (VBS) approach increases simulated SOA by 76% at DSL but reduced POA concentration by 72%, leading to an overall decrease of organic aerosol (OA) by 12%. Unlike the two-product scheme where POA is entirely inert, the VBS scheme includes POA in the OA chemistry. The SOA:OA ratio and the anthropogenic-to-biogenic SOA ratio (ASOA:BSOA) also vary systematically with SOA modeling schemes suggesting that detailed measurements of SOA composition could further constrain modeling methods.