PNNL

Abstract

This paper describes the development and evaluation of a new Model for Simulating Aerosol Interactions and Chemistry (MOSAIC), with a special focus on addressing the long-standing issues associated with solving the dynamic partitioning of semi-volatile inorganic gases (HNO3, HCl, and NH3) to size-distributed atmospheric aerosol particles. The coupled ordinary differential equations (ODE) for dynamic gas-particle mass transfer are extremely stiff, and the available numerical techniques are either too expensive or produce oscillatory and/or inaccurate steady-state solutions. These limitations are overcome in MOSAIC, which couples an accurate and computationally efficient thermodynamic module [Zaveri et al., 2005a,b] with a new dynamic gas-particle partitioning module described here. The algorithm involves time-split integrations of non-volatile and semi-volatile species, and a new concept of “dynamic pH” and an adaptive time-stepping scheme hold the key to smooth, accurate, and efficient solutions over the entire relative humidity range. MOSAIC is found to be in excellent agreement with a benchmark version of the model that uses LSODES (a Gear solver) for rigorously integrating the stiff ODEs. The steady-state MOSAIC results for monodisperse aerosol test cases are also in excellent agreement with those obtained with the benchmark equilibrium model AIM. MOSAIC is also evaluated within a 3-D model, and the average CPU speed is estimated to be over 100 times faster than the dynamic aerosol model MADM [Pilinis et al., 2000]. These results suggest that MOSAIC is highly attractive for use in 3-D aerosol and air quality models in which both accuracy and efficiency are critically important.