An Air Quality Proving Ground (AQPG) for GOES-R

1
An Air Quality Proving Ground (AQPG) for GOES-R

• R. M. Hoff (UMBC GEST/JCET), S. A. Christopher
  (UAH), F. Moshary (CCNY), S. Kondragunta (STAR),
  R. B. Pierce (NESDIS/CIMSS),
• M. Green (DRI), A. Huff (Battelle)
• GOES-R Proving Ground January 2010 Call

2
IDEA (http//star.nesdis.noaa.gov/smcd/spb/aq/)
3
GOES Aerosol and Smoke Product (GASP)
GASP is derived from a single visible channel and
from a 28 day tracking of the darkest pixel in a
scene Cannot do what MODIS and other
multiwavelength sensors can do!
4
GOES lt---gt GOES - R

• Single wavelength
• 1/2 hourly scenes
• Requires 28 day spin-up
• Has a known diurnal bias
• Less precise than MODIS AOD

• Advanced Baseline Imager (ABI) MODIS at GEO
• 16 spectral channels
• Full disk, CONUS, and special scans
• 5 minute images
• AOD should be as good as MODIS

5
Aerosol Detection Physical Description

• Spectral (wavelength dependent) thresholds can
  separate thick smoke, light smoke, and clear sky
  conditions

6
Air Quality Proving Ground

• Using MODIS Models Ground data in hand, can
  we create cases that look interesting enough to
  train users?
• NOAA is creating proxy data sets from model data
• UMBC/UAH identifying cases which impact multiple
  areas and stations (UMBC, UAH, UW, CCNY, ..?)

7
AQPG Workflow
8
AQPG Case 1 - Aug 20-24, 2006

• Mark Green of DRI is working on a case study
  which exercises the AQPG
• This is a case with smoke in the US Northwest and
  sulfate haze in the east
• Period chosen in part because it occurred during
  the Second Texas Air Quality Study (TexAQS II)
• We have a proxy GOES-R product for this case
  produced by Brad Pierce
• A model is guilty until proven innocent- Bill
  Ryan

9
Evaluation of the Case

• Use GOCART aerosol module - predicts
  concentrations of seven aerosol species (SO4,
  hydrophobic OC, hydrophilic OC, hydrophobic BC,
  hydrophilic BC, dust, sea-salt) other
  pm2.5(p25)
• Output at 15 minute intervals
• Model PM2.5 calculated as pm2_5_dryp25bc1bc2o
  c1oc2dust1dust20.286ssalt1ssalt20.942sulfa
  te
• NH4 not included so added 0.375SO4 to account
  for ammonium in ammonium sulfate
• Added larger dust and sea salt categories to
  obtain PM10

10
Contour map of IMPROVE network particulate sulfur
(8/24/06)
11
Contour map of IMPROVE organic carbon for 8/24/06
12
GOES and WRF-Chem AOD show similar patterns
WRF-chem.gif
13
Results
WRF-Chem does a good job predicting SO4
Good correlation for OC, but WRF-Chem biased
factor of 3 low - not surprising as sources are
not inventoried
14
The overall WRF-Chem PM2.5 prediction is
dominated by this under-prediction of OC
15
Impact of speciation on AOD
Bondville- WRF-Chem AOD close to AERONET AOD
except when WRF-Chem predicts clouds- much higher
SO4 AOD predicted
Howard- Increase in SO4 and OC AOD with WRF-Chem
clouds (growth of hydrophilic OC and well as SO4)
16
Next Steps

• Several more case studies have been identified
• Amy Huff of Battelle Memorial Institute will be
  forming a user group at the EPA National Air
  Quality Conference in March
• We will have a workshop in August to start
  training users on the case studies
• Funding has been provided by GOES-R program
  office (Steve Goodman) under cooperative
  agreement number NA09NES4400022 and through the
  CREST Cooperative Agreement