Air Quality Forecasting Downwind from Fires Using RealTime MODIS Data

1
Air Quality Forecasting Downwind from Fires Using
Real-Time MODIS Data

• W. M. Hao, S. P. Urbanski, J. M. Salmon, B. L.
  Nordgren,
• and S. P. Baker
• USDA Forest Service, Fire Sciences
  Laboratory
• Missoula, Montana
• Notice The materials presented here have not
  yet been published, and are therefore not
  available for citation as a reference.

2
Fire unique source of atmospheric trace gases
and aerosols

• Major disturbance to the ecosystems
• Human activities
• - Land use (deforestation, shifting cultivation,
  fuelwood use, clearance of agricultural residues)
• - Land management (wildfires vs. prescribed
  fires)
• Natural cause lightning
• Unpredictable variability spatially and
  temporally
• Effects air quality, tropospheric chemistry,
  stratospheric chemistry, global climate, and
  public health

3
Major Factors Affecting Fire Dynamics and Smoke
Emissions

• Land use and management
• Vegetation types
• Wind, temperature and humidity
• Fuel moisture content
• Fuel elemental composition
• Topography

4
NASA Terra Satellite
NASA Aqua Satellite
Forest Service, RMRS,
Fire Sciences Laboratory,
Missoula, Montana
Forest Service, Fire Sciences Laboratory

• Forecast emission rates of PM 2.5 and 15
  pollutants
• Every 4 hrs for the following 24 hrs
• MODIS data at noon and 2 p.m. with 0.5-1 km
  resolution
• FARSITE model in real-time meteorological
  conditions

Fire Sciences Laboratory

• WRF - Smoke Dispersion Model
• Forecast PM2.5, O3 and 15 pollutant levels
• Downwind from fires with 12-32 km resolution
• Display the forecasts on FS and NIFC web sites

5
http//www.firelab.org/rsl
6
June 16th, 2002 Noon Local Time MODIS Overpass
1749 1801 UTC ADS Under flight 1754 1819 UTC
June 23rd, 2002 Noon Local Time MODIS Overpass
1755 1807 UTC ADS Under flight 1800 1812 UTC
7
Hayman Fire, Colorado
8
Hotspots as a Proxy for Monitoring Area Burned
Rodeo Fire, Arizona June 2002
Biscuit Fire, Oregon July - August 2002
Moose Fire, Montana August - September 2001
9
Comparison of MODIS Detected Hotspots with2002
Fire Occurrence Database
10
(No Transcript)
11
(A) a true color MODIS image (red 0.66 mm
green 0.55 mm blue 0.47 mm) acquired over
areas near Hayman Fire in Colorado on July 18,
2002 (B) a false color image (red 2.13 µm
green 1.24 µm blue 1.64 µm) of the same scene
(C) the same as (B), but with the burned areas
masked in yellow color and (D) a Landsat 7 image
over the same area as (A). Li et al., IEEE
Transactions of Geoscience and Remote Sensing,
2004.
12

• Burn Area Algorithm for MODIS DB System

• ? (apparent reflectance) pL / µoEo
• L measured radiance of the spectral band
• µo cosine of solar zenith angle
• Eo extra-terrestrial solar flux
• seven other tests on the 0.86, 1.24, 1.64, and
  2.13 µm bands to filter out false alarms due to
  sun glint, clouds, etc.
• 500 m x 500 m spatial resolution

13
? The convex hull is the shape produced by
stretching a rubber band around the points. ? It
consistently overestimates the burned area. ? The
degree of overestimation varies, depending upon
the concavity.
? The alpha shape is a generalization of the
convex hull. ? The parameter alpha is adjusted to
vary between the extreme cases of the point set
itself and the convex hull. ? This method is used
to map fresh burning when cloudy conditions
obscure the wavelengths sensitive to char and
ash. ? optimized distance between hotspots 1.5 km
14
Validation of Hayman Fire Burned Areas
15
(No Transcript)
16
Fire Behavior Model - FARSITE

• Fire growth and simulation model
• Simulate surface fire, crown fire, spotting,
  post-frontal combustion, fire acceleration
• Input topography, fuels, meteorological data

17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
Weather Research and Forecasting Smoke
Dispersion Model (WRF-SD)

• WRF next regional meteorological and forecasting
  model, developed by NCAR/NOAA/NCEP/AFAWA
• - Replacement for MM5 and NCEP eta
• WRF-CHEM WRF Model with chemistry, developed by
  NOAA-FSL/NCAR
• - Domain 130 x130 x 31 grid cells
• - Regional 16-km resolution for each horizontal
  grid cell
• - National 32-km resolution
• - Anthropogenic emissions EPA EM99v3
• WRF-SD WRF-CHEM emission rates from fires
  plume height model, being developed by Forest
  Service, Fire Sciences Laboratory, Missoula, MT

28
(No Transcript)
29
Surface CO Concentration, June 8, 2002, 1800
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
North Black Canyon Prescribed Burn
Vladimir A. Kovalev, Cyle Wold, Jenny Newton
and Wei Min Hao USDA Forest Service, Rocky
Mountain Research Station, Fire Sciences
Laboratory In Cooperation with The Johns Hopkins
University Desert Research Institute

• 23 April 2004

35
Lidar Horizontal Scan of Smoke Aerosols
A typical smoke plume horizontal cross section
monitored by the lidar. The colored scale shows
relative intensity of backscattering in arbitrary
units.
36
Lidar Vertical Scan of Smoke Aerosols
A typical smoke plume vertical cross section
monitored by the lidar. The colored scale shows
relative intensity of backscattering in arbitrary
units.
37
Future Challenges

• Improvement and validation of satellite
  measurements
• - Aerosol optical thickness
• - Active fires and burned areas detection limit
  of fires in different ecosystems, understory
  fires, cloud cover
• Development of real-time emission rates of
  atmospheric pollutants from fires using real-time
  MODIS data and FARSITE fire behavior model
  (12/2005)
• Forecast air quality as a result of large fires
  by assimilation of emission rates to WRFSmoke
  Dispersion Model (6/2006)

38

• Comparison of WRF-SD predicted pollutant levels
  with airborne measurements of CO and other trace
  gas concentrations (6/2007)
• - COBRA Mission Ontario and Quebec Fires, May
  and June 2003
• - NASA INTEX-NA Mission and ICART2 Alaska
  Fires, June, July
• and August 2004
• Expansion of the research region from the U.S. to
  North America (Canada, U.S., and Mexico)
• - Field experiments of biomass burning in
  temperate and tropical
• ecosystems in Mexico, 2005-2007
• - Daily emission inventories of trace gases and
  aerosols from
• biomass burning in 1-km x 1-km resolution