Current Status of Air Quality Models for Particulate Matter

Abstract

The current status of the mathematical modeling of atmospheric particulate matter (PM) is reviewed in this paper. Simulating PM requires treating various processes, including the formation of condensable species, the gas/ particle partitioning of condensable compounds, and in some cases, the evolution of the particle size distribution. The algorithms available to simulate these processes are reviewed and discussed. Eleven 3-dimensional (3-D) Eulerian air quality models for PM are reviewed in terms of their formulation and past applications. Results of past performance evaluations of 3-D Eulerian PM models are presented. Currently, 24-hr average PM 2.5 concentrations appear to be predicted within 50% for urban-scale domains. However, there are compensating errors among individual particulate species. The lowest errors tend to be associated with SO 4 2-, while NO 3 -, black carbon (BC), and organic carbon (OC) typically show larger errors due to uncertainties in emissions inventories and the prediction of the secondary OC fraction. Further improvements and performance evaluations are recommended.

Integrated Assessment Modeling of Atmospheric Pollutants in the Southern Appalachian Mountains: Part II. Fine Particulate Matter and Visibility

“…16,17 This is especially encouraging, because the duration of the SAMI episodes were typically longer than those of previous studies and cover a wide range of meteorological conditions including winter, spring, and summer episodes. The fall season was not modeled directly but was represented by days chosen from other seasons with similar meteorology and air quality.…”

“…Aerosol NO 3 , formed by gas-to-particle conversion of nitric acid, is usually found in low levels (Ͻ1 g/m 3 ); therefore, the normalized error can be high, although the absolute error is quite low. However, it should be noted that even these large errors associated with NO 3 are much lower compared with 295% reported by Seigneur 16 4 when the observations are assumed to be in the form of (NH 4 )HSO 4 is only 2.6%.…”

“…Current EPA Draft Guidance 8 for modeling regional haze and PM does not address the question of when the model performance is "acceptable" for regulatory applications. Seigneur 16 suggests that current air quality models can achieve normalized errors of Յ50% for PM 2.5 and SO 4 for urban-scale domains. An error of 50% may seem to be large; however, for the purposes of developing control strategies, air quality models are typically used in a relative sense rather than an absolute sense.…”

“…Finally, an episode selection program was run to optimize across the pollutants and Class I areas in an effort to minimize the errors associated with using a set of episode days to represent annual or seasonal air quality metrics. The end product consisted of six multiday episodes (46 modeling days) representing the range of meteorological conditions that frequently occur in the region: February 8 -13, 1994; March 23-31, 1993; May 11-17, 1993; May 24 -29, 1995; July 23-31, 1991; and July [11][12][13][14][15][16][17][18][19]1995. These episodes also represent a range of O 3 , PM, and acid deposition levels with the highest PM levels found in the summer episodes and the lowest in the winter and early spring episodes.…”

Technical Note: A new SIze REsolved Aerosol Model (SIREAM)

“…Historically, CTMs focused on ozone formation or acid deposition and did not include a detailed treatment of aerosol. Several models have been developed that include a very thorough treatment of aerosol processes Gong et al (2003); Adams and Seinfeld (2002); Spracklen et al (2005), but there are still many limitations Seigneur (2001).…”