Close Window

Distinguishing Between Conversion and Maintenance Fires in the Amazon

Douglas Christopher Morton, University of Maryland, morton@geog.umd.edu (Presenting)
Ruth Sarah DeFries, University of Maryland, rdefries@geog.umd.edu
Wilfrid Schroeder, University of Maryland, schroeder@hermes.geog.umd.edu
Louis Giglio, SSAI, giglio@hades.gsfc.nasa.gov
Jim Randerson, University of California, Irvine, jranders@uci.edu
Guido Van der Werf, Vrije Uiversiteit Amsterdam, The Netherlands, guido.van.der.werf@falw.vu.nl
Ivan Csiszar, University of Maryland, icsiszar@hermes.geog.umd.edu
Jeff Morisette, NASA-GSFC, jeff.morisette@nasa.gov

Fire is a critical management tool for forest conversion and maintenance of existing pastures in Amazonia. Varying patterns of fire use control the timing and magnitude of carbon emissions from land management. Quantifying the relative contribution of conversion and maintenance fires to satellite-based fire detections has proven elusive due to uncertainties in instrument sensitivity, poor geolocation and low spatial resolution of remotely sensed data products, and the rapid pace of land cover conversion at the forest frontier. We develop a simple approach to identify conversion fires based on the frequency of fire detections from the MODIS sensors. We then combine active fire detections with data on deforestation and post-clearing land use to quantify the frequency of fire events and duration of the clearing process associated with conversions of forest to pasture and cropland in Mato Grosso State. Conversion fires account for at least 57% of the MODIS active fire detections in Mato Grosso during 2003-2005. Deforestation for cropland occurs at a higher mean fire frequency than conversions to pasture in the years following forest clearing (up to 40 fire days per year). Repeated burning of trunks, stumps, and woody roots in preparation for mechanized agriculture commonly occurs over two or more years following deforestation. Fire use for pasture creation is less intensive; only 1/3 of recent forest to pasture conversions resulted in high-frequency fire detections in more than one year. Finally, we provide basinwide estimates of conversion fire activity during 2001-2005 to highlight inter-annual and regional differences in land use fire. The dynamics of the conversion process, as evident from satellite-based detection of fires, provides critical new information for determining the nature of carbon emissions from fire activity in Amazonia.

Science Theme:  LC (Land Use and Land Cover Change)

Session:  1D: Rates and Patterns of Land Use Change

Presentation Type:  Oral

Abstract ID: 69

View presentation

Close Window