Air Quality Remote Sensing Program
GOES Smoke Products

NOAA / NWS requested NESDIS to provide observational support to verify interim HYSPLIT smoke (PM2.5) forecasts. NESDIS initiated the algorithm development work to derive smoke concentrations from GOES AOD measurements. This algorithm uses source apportionment technique by fusing GOES observations of fire hot spots and AOD maps at 30-minute interval. Plume direction and extent from all observed fire sources are first determined. AOD values in the map not associated with the fires are dropped. For plumes transported long distance from fire sources, we use pattern recognition technique. Since we are working with 30-minute AOD imagery, this algorithm provides us with wind speed and direction associated with the plumes. However, the vertical location of these plumes is not known.

The figure below shows the sequential steps involved in the algorithm. Top figure shows original GOES AOD image with clouds and other gray areas where retrieval was not possible. Middle figure shows AOD associated with fires, and the bottom figure shows smoke concentrations derived from AOD. This smoke concentration file will be used by NWS in its forecast verification process.

GOES-12 ch1+Aerosol Optical Thickness 20050908 2145Z

GOES-12 Detected Smoke Aerosol Optical Thickness (20050908 2145Z)

Smoke plumes from all fires. To get from the first image to the Aerosol Optical Thickness image, all other data not associated with fires has been automatically dropped.

GOES-12 Detected Smoke Concentration (2005 09 08 2145Z)

AOD is converted to smoke concentrations by using aerosol mass extinction coefficient of 7.8m²/g.