The Community Multi-Scale Air Quality (CMAQ) modelin
gsystem was used to investi
gate ozone and aerosolconcentrations in the Pacific Northwest (PNW) durin
g hotsummertime conditions durin
g July 1-15, 1996. Twoemission inventories (EI) were developed: emissions forthe first EI were based upon the National Emission Trend1996 (NET96) database and the BEIS2 bio
genic emissionmodel, and emissions for the second EI were developedthrou
gh a "bottom up" approach that included bio
genicemissions obtained from the GLOBEIS model. The twosimulations showed that elevated PM
2.5 concentrationsoccurred near and downwind of the Interstate-5 corridoralon
g the foothills of the Cascade Mountains and in forestedareas of central Idaho. The relative contributions ofor
ganic and inor
ganic aerosols varied by re
gion, but
generally or
ganic aerosols constituted the lar
gest fractionof PM
2.5. In wilderness areas near the I-5 corridor,or
ganic carbon from anthropo
genic sources contributedapproximately 50% of the total or
ganic carbon withthe remainder from bio
genic precursors, while in wildernessareas in Idaho, bio
genic or
ganic carbon accounted for80% of the total or
ganic aerosol. Re
gional analysis of thesecondary or
ganic aerosol formation in the ColumbiaRiver Gor
ge, Central Idaho, and the Olympics/Pu
get Soundshowed that the production rate of secondary or
ganiccarbon depends on local terpene concentrations and thelocal oxidizin
g capacity of the atmosphere, which was stron
glyinfluenced by anthropo
genic emissions. Comparison withobservations from 12 IMPROVE sites and 21 ozone monitorin
gsites showed that results from the two EI simulations
generally bracketed the avera
ge observed PM parametersand that errors calculated for the model results werewithin acceptable bounds. Analysis across all statisticalparameters indicated that the NW-AIRQUEST EI solutionperformed better at predictin
g PM
2.5, PM
1, and
![](/ima<font color=)
ges/
gifchars/beta2.
gif" BORDER=0 ALIGN="middle">
ext eventhou
gh or
ganic carbon PM was over-predicted, and theNET96 EI solution performed better with re
gard to the inor
ganicaerosols. For the NW-AIRQUEST EI solution, the normalizedbias was 30% and the normalized absolute error was49% for PM
2.5 mass. The NW-AIRQUEST solution sli
ghtlyoverestimated peak hourly ozone downwind of urban areas,while the NET96 solution sli
ghtly underestimated peakvalues, and both solutions over-predicted avera
geO
3 concentrations across the domain by approximately 6ppb.