Combining the ADMM and Mathematical Homogenization for Atmospherical Pollutant Dispersion Modeling

Abstract The advection-diffusion multilayer method (ADMM) produces accurate semi-analytical solutions of initial/boundary-value problems for advection-diffusion equations with variable coefficients that model pollutant dispersion in the atmosphere, and exhibits lower computational cost in comparison to other integral transform-based methods. However, in operative situations such as natural/industrial disasters resulting in the release of pollutants to the atmosphere, it is necessary to assess rapidly and accurately the ground-level distribution of pollutant concentration in order to minimize the impact on health and economy. Here, in order to accelerate the availability of results with little loss of accuracy, the ADMM is combined with mathematical homogenization, whose use in pollutant dispersion modeling seems to be new. The proposed approach is compared with the direct application of the ADMM and to the observations of the Hanford experiment in order to access both its accuracy and computational cost, for stable atmospheric conditions and considering the influence of deposition velocity. The results show that the combination of the ADMM and mathematical homogenization reduces remarkably the computation cost with little loss of accuracy.