Biomass burning in tropical ecosystems: An analysis of vegetation, land settlement, and land cover change to understand fire use in the Brazilian lower Amazon

Research in global environmental change emphasizes that biomass burning significantly contributes to increased atmospheric trace gases and possible climate change. Analyses of what drives anthropogenic fire is less thoroughly examined because such study involves examining the human and physical dimensions of biomass burning at local and regional scales. This dissertation uses a multi-scale approach to address fire use within local and regional contexts. It investigates dynamics and effects of fire use within four rural communities with different settlement histories, then expands these findings to understand burning patterns in a larger agricultural frontier south of Santarem, Brazil's third largest Amazon city. The aim of the dissertation is to understand how landscape environmental factors and land settlement shape land use practices and the burning patterns associated to those practices.; The dissertation integrates analyses of biomass burning at three spatial scales: regional, ecological field, and local. At the regional scale, a model of biomass change is developed from remotely sensed data (Landsat TM images path/row: 227/62; dates: Oct. 1986, Oct 1995) and used in combination with household land use information to infer extent of biomass burning in the study region over a 9 year period. At the field scale, physical evidence of slash and burn agriculture is examined through vegetation inventories and measure of post-fire fuel loads in 14 agricultural fields. At the local scale, in-depth household interviews on household history, land use strategies, and present/historical burning practices compliment physical evidence, to provide a fuller understanding of the local causes and impacts of fire use. Throughout the dissertation a geographic information system (GIS) is used to assess temporal and spatial characteristics of human settlement; and a global positioning system (GPS) is used to link vegetation information and settlement findings to land cover classifications derived from remotely sensed data.; This dissertation advocates the need for local and regional studies on environmental issues to inform global environmental change research and estimation. It provides a framework that links human dimensions of biomass burning to larger global change issues. Findings in the dissertation contribute to the discipline of geography in the area of human/environment interactions.