614.4 Branch Seminar - Doug Morton

University of Maryland, College Park
Fire is the dominant means of clearing Amazon forest for agricultural uses. Deforestation fires often escape their intended boundaries and burn into standing forest, killing between 5 and 50% of canopy trees, thereby altering forest structure, species composition, and the likelihood of futuredisturbances. Previous estimates of the extent of fire-damaged forests in Amazonia are poorly constrained due to large inter-annual variability in forest fire activity and frequent confusion among deforestation, selective logging, and forest fires in satellite image analyses. We build a more complete understanding of the source, size, and frequency of forest fire events in southern Amazonia using time-series methods to characterize anthropogenic fire use and isolate fire-damaged forest from logging and deforestation. Results from the satellite-based assessment of burned forest are used to calibrate a new fire model in Ecosystem Demography (ED), a height-structured ecosystem model. Model testing demonstrates that frequent fire exposure can trap a forested ecosystem in a low-biomass state dominated by short vegetation. Future modeling work will evaluate the contribution of anthropogenic fires to changes in Amazon forest structure under a range of climate scenarios to generate insight into how different land use possibilities alter the long-term response to climate change.