Environmental Engineering Reference
In-Depth Information
(OSAT and APCA) tracks the contributions to each grid cell from emissions
source groups, emissions source regions, initial conditions, and boundary conditions
with reactive tracer species. APCA recognizes that there are source categories
such as biogenics that can not be controlled so the model only attributes ozone to
biogenics when it is due to the interaction of biogenic VOC + biogenic NOx.
Ozone and particulate matter source apportionment technology (PSAT) has been
implemented into the most recent version of the CAMx model (v4.5) and is
publicly available (ENVIRON, 2008). PSAT estimates the contribution from
specific emissions source groups, emissions source regions, initial conditions, and
boundary conditions to PM2.5 using reactive tracers.
The Ozone and Precursor Tagging Methodology (OPTM) and Particle and
Precursor Tagging Methodology (PPTM) have been implemented in CMAQ v4.6
(ICF International 2007a, b). OPTM estimates contributions from emissions source
groups, emissions source regions, initial conditions, and boundary conditions to
ozone by tracking total oxidant (defined for this purpose as {NO
2
+ NO
3
+ 2*N
2
O
5
+ O
3
}) formation from duplicate model species for each contributing source.
PPTM estimates contributions from emissions source groups, emissions source
regions, initial conditions, and boundary conditions to PM2.5 by adding duplicate
model species for each contributing source.
3. Results and Discussion
Both models tend to over-predict the lowest ozone concentrations and under-
predict the highest ozone concentrations. PM2.5 model performance is similar for
both models. CMAQ tends to predict higher concentrations of ammonium nitrate
at this location. Both models show the best agreement with observations for
PM2.5 sulfate ion.
In general, CMAQ/OPTM tends to predict higher contributions for the boun-
dary conditions and less contribution to more local regions. The CAMx/APCA
approach estimates the highest contribution from local sources. This is in part due
to APCA switching contribution to local NOX emissions during conditions where
local NOX mixes with biogenic VOC under VOC limited ozone formation.
CMAQ and CAMx contribution estimates are well correlated (r
2
= .9) and have a
RMSE less than 2.55. Most of the error between CAMx and CMAQ estimates is
from the contributions from the boundary and all other non-tagged sources.
CMAQ/OPTM tends to predict higher contributions from the boundary than the
CAMx ozone source apportionment implementations. CAMx tends to estimate
higher contributions from the non-tagged emissions inside the modeling domain
than CMAQ/OPTM. CMAQ has higher estimates of local contributions. A com-
parison of DDM sensitivity to CAMx ozone source apportionment indicates ozone
source apportionment may overstate the contribution from the boundaries (Dunker
et al., 2003). The spatial pattern of contribution from the sources regions tracked in
each model slightly different. CAMx/APCA tends to have the highest contributions
