Air Pollution as a Climate Forcing: A Workshop

Day 1 Presentations

Estimating Fire Emissions from Boreal Forest Fires

Zhanqing Li
Dept. of Meteorology and ESSIC, University of Maryland, College Park, MD

You may download a MS PowerPoint version (7.2 MB) of this presentation.


The boreal forest zone accounts for about 25% of global forest area. On average, wild land fires affect some 14-15 million ha of boreal forest annually. Due to large biomass content, boreal forest fires emit large quantities of trace gases (e.g., CO2, CO and CH4) and smoke aerosol into the atmosphere, both of which influence the earth's ecosystem, environment and climate. Fire activity and the ensuing quick growing dictate, to a large extent, whether the boreal ecosystem serves as the source or sink in the global carbon cycle. During the fire season, biomass burning in North America could account for 60% of variability in CO concentration, and may travel across the Atlantic to influence air quality in Europe (Spichtinger et al. 2001). Reduction of the surface solar radiation due to smoke may account for one third of total energy depletion due to both clouds and smoke during an active fire season in Canada (Li and Kao, 1998).

To estimate the potential impacts of forest fire on environment and climate, techniques are required to determine emissions of various gases (some of which are major air pollutants) that vary with the extent of burning, fuel density (loading), emission efficiency and emission ratios of different species:

Emissions = Fuel Loading × Burned Area × Emission Efficiency × Emission Ratio

Due to the widespread extent and remoteness of boreal fires, satellite remote sensing plays an indispensable role. At present, remote sensing algorithms for mapping burned area are relatively mature (Li et al. 2001a), but the quality of various fire products still varies considerably. In general, detection and mapping of forest fires are more robust and accurate than is the case for savanna fires due to their stronger signal and longer duration. For example, the algorithms of Li et al. (2000) and Fraser et al. (2000) can detect and map over 95% of forest fires with very low commission errors. They are applied to a 15-year archive of AVHRR satellite data to generate a long-term fire climatology over all forest regions across North America. Fuel loading refers to the biomass both above and below ground. Above-ground biomass varies with fuel type, age, density etc. Fuel type information is available over limited regions that were usually compiled by forest agencies. There is an increasing use of satellite-based classification of land cover types to generate high-resolution fuel type data, although large uncertainties may result from such conversion unless extensive in-situ measurements are available to assist the conversion. Tree density and age are related to the greenness of vegetation that can be measured by the combination of visible and NIR reflectance. The difficulty using this approach lies in the easy saturation of the signal, yet understory vegetation could distort the signal significantly. SWIR measurement offers some possibility of circumventing the problem (Fraser and Li 2002). Below-ground biomass is most difficult to obtain, which is a primary source of uncertainties in estimating fire emissions. Estimates of emission efficiency (total emission per unit biomass burned) and emission ratio (fraction of emission for a particular gas) have relied mostly on in-situ measurements. The former is usually much more variable than the latter. Total emission could be inferred from smoke plume that is visible from visible satellite observations (Li et al. 2001b). However, this requires extensive field campaigns, which have not yet been attempted, to establish the relationships.


  • Fraser, R., Li, Z., and J. Cihlar, 2000, Hotspot and NDVI differencing synergy: A new method for mapping fire scars, Rem. Sens. Environ., 74, 362-375, 2000.
  • Fraser, R., and Z. Li, 2001, SPOT VEGETATION for Estimating parameters related to boreal forest fire: burned area, regeration age, and biomass; Rem. Sens. Environ., in press.
  • Li, Z., and L. Kou, 1998, Atmospheric direct radiative forcing by smoke aerosols determined from satellite and surface measurements, Tellus (B), 50, 543-554.
  • Li, Z., S. Nadon, J. Cihlar, Satellite detection of Canadian boreal forest fires: Development and application of an algorithm, Int. J. Rem. Sens., 21, 3057-3069, 2000.
  • Li, Z., Y. J. Kaufman, C. Ithoku, R. Fraser, A. Trishchenko, L. Giglio, J. Jin, X. Yu, A review of AVHRR-based active fire detection algorithms: Principles, limitations, and recommendations, in Global and Regional Vegetation Fire Monitoring from Space, Planning and Coordinated International Effort (Eds. F. Ahern, J.G. Goldammer, C. Justice), pp 199-225, 2001a.
  • Li, Z., A. Khananian, R. Fraser, J. Cihlar, Detecting smoke from boreal forest fires using neural network and threshold approaches applied to AVHRR imagery, IEEE Tran. Geosci. & Rem. Sen., 39, 1859-1870, 2001b.
  • Spichtinger, N., M.Wenig, P. James, T. Wagner, U. Platt, A. Stohl, Satellite detection of a continental-scale plume of nitrogen oxides from boreal forest fires, G. Res. Let., 28, 4759-4782, 2001.

Workshop Homepage * Background
Summaries: Overview, Gases, Aerosols, Tech., Health, Agri./Eco.
Abstracts: Day 1, Day 2, Day 3, Day 4, Day 5 * Participants