Deriving Biomass Burning Emissions from GOES WildFire Products

1
Deriving Biomass Burning Emissions from GOES
WildFire Products

• P.I Shobha Kondragunta
• NOAA/NESDIS/ORA
• Co-I Chris Schmidt
• UW-Madison

2
NWS Air Quality Forecast Model Schematic
Emissions
Initial Conditions Boundary Conditions Meteorologi
cal Fields
Anthropogenic
Biogenic/Biomass Burning
Community Multiscale Air Quality (CMAQ) Model
Air Quality Forecast (O3, PM25,)
3
Problem

• PM25 emissions from biomass burning (forest
  fires) are currently not included in CMAQ
  simulations
• This leads to uncertainties in PM25 forecasts
  during long range transport of forest fire smoke

Long range transport of smoke (PM25) to US from
forest fires in Canada during July 16-18, 2004
4
Satellite Observations vs CMAQ PredictionsJuly
20 2004 16Z
Difference
Histograms
CMAQ model probably underestimating AOD fields
due to absence of PM25 emissions from biomass
burning
5
GASP-CMAQ ComparisonsJuly 20 2004 16Z
Difference
Histograms
6
Satellite vs Model (AOD Correlation)
7
Model AOD vs Model PM25

• Model internally consistent
• Minimum threshold in CMAQ AOD

8
Satellite AOD vs Model PM25
No correlation between satellite observed AOD and
Model predicted PM25 Most of the observed smoke
not represented in the model
9
Solution Incorporate PM25 emissions into CMAQ

• Approach 1
• Scale emission climatologies with fire counts
  (e.g., GOCART model)

10
Solution Incorporate PM25 emissions into CMAQ

• Approach 1
• Scale emission climatologies with fire counts
  (e.g., GOCART model)
• Approach 2
• Assume each fire point corresponds to 10 ha
• burning and emitting at a rate of 15 kg/ha/hr
• (e.g., HYSPLIT model)

11
Solution Incorporate PM25 emissions into CMAQ

• Approach 1
• Scale emission climatologies with fire counts
  (e.g., GOCART model)
• Approach 2
• Assume each fire point corresponds to 10 ha
• burning and emitting at a rate of 15 kg/ha/hr
• (e.g., HYSPLIT model)
• Approach 3
• Derive emissions using fire points, fuel load,
• and emission factors information
• EBA X FL X FF X EF

E Emissions FF Fuel
Fraction BA Burned area EF Emissions
Factors FL Fuel Loading
12
Approach 3 (Schematic)
Fire Counts (e.g., from GOES
Weather
Fuel moisture
Vegetation
Fuel type
Emissions Factors
Fuel Fraction Consumed
Emissions Estimates
Fuel loading
Static
CMAQ
Dynamic
Air Quality Forecast Model
13
Summary and Work Plan

• Summary
• CMAQ simulations during a known biomass burning
  event underestimate PM25 and AOD fields. This is
  due to the absence of smoke emissions in the
  model
• Satellite-derived PM25 emissions in near real
  time will be developed to be incorporated into
  CMAQ
• Work Plan
• Collect existing static (fuel load, emission
  factors) data bases from USFS and other sources
• Analyze and assess the databases
• Derive emissions for a known fire episode (use
  GOES fire products and other satellite
  information for fuel moisture and so forth)
• Conduct impact studies in collaboration with USFS
  and EPA
• Transition technology to NWS

14
Acknowledgements

• CMAQ runs (Pius Lee and Jeff McQueen)
• GOES AOD product (Ana Prados)
• GOES re-gridded AOD product (Chieko Kittaka)
• HYSPLIT forecast map (Roland Draxler)