Comparison of two numerical schemes for integrating the atmospheric diffusion equation: evaluation with atmospheric data

Abstract

Several recent studies have examined the accuracy of alternative numerical schemes for solving the species conservation equation applied to tropospheric air pollution problems. These studies have centered mainly on the use of simplified numerical experiments involving the advection of different initial concentration distributions (e.g. a cosine hill). Few studies report comparative evaluation of alternative numerical schemes using established photochemical modeling data bases for realistic urban settings. This study presents a comparison of two contemporary schemes—SHASTA and the multidimensional flux correction method of Zalesak—applied to the Los Angeles, Calif. and Philadelphia, Pa urban areas. The results indicate that, for the conditions examined, the Zalesak method gives slightly more accurate peak ozone predictions than the SHASTA method in applications to Los Angeles (where peak ozone levels are three times the federal standard). For the more moderate ozone conditions found in Philadelphia, very little difference in accuracy is seen between the two methods.