The AIRPACT3 Photochemical Air Quality Forecast System: Evaluation and Enhancements

1
The AIRPACT-3 Photochemical Air Quality Forecast
System Evaluation and Enhancements
Jack Chen, Farren Thorpe, Jeremy Avis, Matt
Porter, Joseph Vaughan, and Brian
Lamb Laboratory for Atmospheric
Research Washington State University
2
Tomorrows Air Quality AIRPACT-3 Daily Forecast
System

• MM5(3.7.3) numerical mesoscale meteorological
  model
• SMOKE(2.1) Sparse Matrix Operating Kernal for
  Emissions processing.
• CMAQ (4.6) Community Multi-scale Air Quality
  model
• SAPRC-99 O3 toxics chemistry
• Aerosol (ver.4) in Aitken, accumulation, coarse
  modes
• Deposition of N, S, O3, Hg species
• 12 km x 12 km grid cells, 21 layers
• Forecast to 64 hours daily

PM species Nitrates Sulfates Organic
aerosols Wind-Blown Dust (soon) PM2.5 total mass
3
AIRPACT-3 Dynamic Emissions IC/BC
2005 anthropogenic emissions (SMOKE)
Updated Biogenic Emission Model (BEIS3)
MM5/MCIP Meteorology from UW Weather Forecast
System
WSU Dairy NH3 Emissions Module
Wild and Prescribed Fire Emissions (from BlueSky
system)
Dynamic Boundary Conditions spatial temporal
variations from MOZART monthly diurnal average
profiles
CMAQ
Daisy-chain Initial Conditions
4
AIRPACT3, July, 2006 Ozone and Biogenic SOA
Formation
http//www.airpact-3.wsu.edu
5
Retrospective long term evaluationAug Nov,
2004 (coincident with Columbia Gorge Visibility
Monitoring program)

• 8 hr daily maximum O3
• 24 hr daily maximum PM2.5
• Speciated PM2.5

6
MM5 forecast performance results
GFS initialization, CCM2 radiation scheme,
Reisner-2 moisture microphysics parameterization,
Kain-Fritsch cumulus parameterization and the
MRF/Hong-Pan planetary boundary layer (PBL)
scheme
7
Daily maximum 8 hr ozone results
P/O vs Observed
Timing errors (running 8 hr means)
8
Ranked daily maximum 8 hr ozone for selected
sites (Aug/Sept, 2004)
9
Ranked speciated PM2.5 observations and
predictions
10
Enhancements to Airpact-3Wildfire emissions
exported to Airpact-3 from the Forest Service
BlueSky System
BlueSky fire locations size
Predicted PM2.5 from fires
SMOKE emissions processing for CMAQ
11
Wildfire Contributions to Regional Ozone
O3 with NOx from fires
O3 without NOx from fires
O3 difference
Daily max O3 Sept 2 6, 2006
12
Further developments

• Nested domains (4 km grid cells)
• Enhanced automatic evaluation
• NASA products

13
Collaborators Acknowledgements

• Collaborating research groups
• USDA Forest Service Don McKenzie Sim Larkin
• USDA NRCS Susan ONeill
• Funding sources
• NW-Airquest Consortium
• NASA ROSES Decision Support System grant

14
Initial evaluation results

• Ozone 8 hr daily maxima
• NMB 6 and NME 17
• Peak values correctly estimated
• PM2.5 24 hr daily maxima
• Daily FB range -75 to 75, mean FB 3
• Daily FE range 3 to 79, mean FE 50
• Speciated PM2.5
• Good match with NO3 observations
• Split decision for NH4 Gorge sites
  underestimated
• SO4 underestimated
• OC and EC overestimated at peak levels
  (associated with wildfires)

15
Enhancement of Airpact using satellite
dataOMI-Aura Satellite Retrievals and AIRPACT-3
for NO2
U.S. retrieval
AIRPACT-3
16
Urban NO2 along the I-5 corridor OMI-AURA and
Airpact

• OMI (American)

• OMI (Dutch)

AIRPACT
17
Bias and error maps for Aug-Sept daily maximum 8
hr ozone
Normalized mean bias
Normalized mean error
18
24 hr PM2.5 Performance Time Series