Module for simulating composition effects on secondary organic aerosol partitioning and its evaluation in the southeastern United States

Abstract

This study systematically investigated the effects of aerosol chemical composition on SOA production in real atmosphere. The findings in this study help reveal the interactions between individual aerosol components and the subsequent effects on secondary organic aerosol (SOA) partitioning.

The aerosol code in the CMAQ model was modified to incorporate structure information and partitioning parameters of lumped SOA product groups, which are formulated directly from the corresponding properties of the individual SOA products. The updated CMAQ was evaluated against field measurements, from two monitoring networks –IMPROVE and SEARCH and one field study - Southern Oxidants Study (SOS99), in the Nashville region during the summer of 1999. It is found that POA composition representation greatly affects the quantification of the composition effects on SOA production. While assuming aerosol phase activity coefficients are 1 for all organics is a good approximation to speed up the simulation for an aerosol mixture solely composed of wood smoke and SOA components, for a mixture of diesel soot and SOA products, making such a simple assumption would result in a great overprediction of the ambient SOA concentrations.

Simulations were also conducted to study the influence of model parameters of great uncertainty such as the vaporization enthalpy of the individual SOA products and the number of lumped groups used to represent SOA production. Simulation results indicate that the vaporization enthalpy for the SOA components need more research efforts due to its significant effects on the predicted SOA concentrations. It is also worthwhile to pay attention to the number of lumped groups applied in air quality simulation. The effects of water absorbing into the aerosol phase were also studied. Allowing water absorbing into the aerosol phase improves the model prediction on SOA concentration in this study. However, the CPU time is increased by several times compared to that required for the base case simulation. This new module is not only applicable to CMAQ, but also can be incorporated into other air quality models.