Comprehensive Surface-Based Performance Evaluation of a Size- and Composition-Resolved Regional Particulate-Matter Model for a One-Year Simulation

A comprehensive performance evaluation has been carried out for the first annual simulation made with AURAMS, a size- and composition-resolved, off-line, regional particulate-matter (PM) modelling system. The year simulated was 2002, the modelling domain covered most of North America, and the horizontal grid size was 42 km. The large evaluation data set consisted of filter-based and continuous surface air-chemistry measurements made by five Canadian and U.S. networks and precipitation-chemistry measurements made by seven Canadian and U.S. networks. Completeness criteria were used to exclude stations with incomplete records, and units conversions were performed to maximize uniformity and conparability. Quantities used in the performance evaluation included annual air concentrations of SO2, NO2, O-3, HNO3, PM2.5, PM10, PM2.5-SO4, PM2.5-NO3, PM2.5-NH4, PM2.5-CM, PM2.5-EC, and PM2.5-TOM, and annual concentrations in precipitation of SO4=, NO3-, and NH4+. The extensive evaluation has allowed inferences about factors contributing to some model weaknesses.