Investigating dominant characteristics of fires across the Amazon during 2005–2014 through satellite data synthesis of combustion signatures

• Published in: Journal of geophysical research. Atmospheres 2017-01, Vol.122 (2), p.1224-1245
• Main Authors: Tang, W., Arellano, A. F.
• Format: Article
• Language: English
• Subjects: Aerosol optical depth; Aerosols; Carbon; Carbon monoxide; Combustion; Combustion efficiency; Computing time; Data; Deforestation; Depth; Dioxides; Distribution; Drought; Drought index; Droughts; Efficiency; Emission inventories; Emissions control; Estimates; Fires; fires in the Amazon; Forest & brush fires; Geophysics; Land cover; Mathematical models; Meteorology; MODIS AOD; MOPITT CO; Nitrogen; Nitrogen dioxide; OMI NO2; Optical analysis; Parameterization; Parametrization; Phases; Rainforests; Satellite data; satellite retrievals; Satellites; Scale (ratio); Seasons; Signatures; Smoke; Smoldering; Synthesis; Understory
• ISSN: 2169-897X; 2169-8996
• DOI: 10.1002/2016JD025216

Estimates of fire emissions remain uncertain due to limited constraints on the variations in fire characteristics. Here we demonstrate the utility of space‐based observations of smoke constituents in addressing this limitation. We introduce a satellite‐derived smoke index (SI) as an indicator of the dominant phase of large‐scale fires. This index is calculated as the ratio of the geometric mean of observed fractional enhancements (due to fire) in carbon monoxide and aerosol optical depth to that of nitrogen dioxide. We assess the usefulness of this index on fires in the Amazon. We analyze the seasonal, regional, and interannual joint distribution of SI and fire radiative power (FRP) in relation to fire hotspots, land cover, Drought Severity Index, and deforestation rate estimates. We also compare this index with an analogous quantity derived from field data or emission inventories. Our results show that SI changes from low (more flaming) to high (more smoldering) during the course of a fire season, which is consistent with the changes in observed maximum FRPs from high to low. We also find that flaming combustion is more dominant in areas where deforestation fires dominate, while smoldering combustion has a larger influence during drought years when understory fires are more likely enhanced. Lastly, we find that the spatiotemporal variation in SI is inconsistent with current emission inventories. Although we recognize some limitations of this approach, our results point to the utility of SI as a proxy for overall combustion efficiency in the parameterization of fire emission models. Key Points A new satellite‐derived smoke index is introduced to indicate combustion phases Multispecies analysis reveals variations in combustion phases within the fire season in the Amazon Flaming combustion is more dominant in deforestation fires, while drought likely enhances more the spread of smoldering understory fires