The new National Ambient Air Quality Standards (NAAQS) far particulate matt
er (PM), promulgated by the U.S. Environmental Protection Agency, include 2
4-hour and annual average standards for fine particles (PM2.5), in addition
to the previous PM10 standards. Numerical models are needed to develop the
emission control strategies that will bring polluted areas into attainment
of the standards. Because the fraction of PM material that is formed in th
e atmosphere (secondary PM) is more significant in PM2.5 than in PM10, the
numerical models required to develop reliable source receptor relationships
must take this secondary PM into account. We review numerical modeling tec
hniques in terms of their ability to address the PM standards. We recommend
that various techniques be used (sometimes in combination) to address the
different PM standards. Further model development and evaluation, additiona
l field data collection, and training of agency staff in the use of more ad
vanced modeling techniques are recommended.