Modelling multi-component aerosol transport problems by the efficient splitting characteristic method

In this paper, a splitting characteristic method is developed for solving general multi-component aerosol transports in atmosphere, which can efficiently compute the aerosol transports by using large time step sizes. The proposed characteristic finite difference method (C-FDM) can solve the multi-component aerosol distributions in high dimensional domains over large ranges of concentrations and for different aerosol types. The C-FDM is first tested to compute the moving of a Gaussian concentration hump. Comparing with the Runge-Kutta method (RKM), our C-FDM can use very large time step sizes. Using Δt = 0.1, the accuracy of our C-FDM is 10−4, but the RKM only gets the accuracy of 10−2 using a small Δt = 0.01 and the accuracy of 10−3 even using a much smaller Δt = 0.002. A simulation of sulfate transport in a varying wind field is then carried out by the splitting C-FDM, where the sulfate pollution is numerically showed expanding along the wind direction and the effects of the different time step sizes and different wind speeds are analyzed. Further, a realistic multi-component aerosol transport over an area in northeastern United States is studied. Concentrations of PM2.5 sulfate, ammonium, nitrate are high in the urban area, and low in the marine area, while sea salts of sodium and chloride mainly exist in the marine area. The normalized mean bias and the normalized mean error of the predicted PM2.5 concentrations are −6.5% and 24.1% compared to the observed data measured at monitor stations. The time series of numerical aerosol concentration distribution show that the strong winds can move the aerosol concentration peaks horizontally for a long distance, such as from the urban area to the rural area and from the marine area to the urban and rural area. Moreover, we also show the numerical time duration patterns of the aerosol concentration distributions due to the affections of the turbulence and the deposition removal. The developed splitting C-FDM algorithm can be applied to model spatial multi-component aerosol transport problems in large domains in atmosphere.