Abstract ID: 530

Analysis of Long-Term Fire Dynamics and Impacts on the Amazon Using Integrated Multi-Source Fire Observations

This LBA-ECO (Large-Scale Biosphere-Atmosphere Experiment in Amazonia - Ecosystems) Phase III study was designed to delineate long-term (1995-present) fire dynamics in the Amazon Basin and to produce associated biomass burning emissions estimates. The data record is based on active fire detections by the GOES East geostationary satellite imager data, processed by the Wildfire Automated Biomass Burning Algorithm (WF_ABBA). Fire detections from the WF_ABBA product were validated by coincident 30 m resolution ASTER and ETM+ data, which allowed for the calculation of commission and omission error rates, and probabilities of detection for a range of vegetation and background conditions. Improvements in the WF_ABBA algorithm, including, but not limited to, the provision of information on cloud obscuration have enabled the creation of a climate quality long-term data record of fire activity. At the current stage of the project hotspots from the GOES data record are adjusted for rates of commission and omission from the validation study and for cloud obscuration by a statistical approach that estimates the probability of burning based on fuel type, fuel conditions, recent precipitation and prior fire occurrence. The corrected hot spots are used as source terms to generate a long-term record of biomass burning emissions in the region. GOES data are also processed by operational algorithm used by CPTEC/INPE (Weather Forecasting and Climate Analysis Center / National Center for Space Research) and differences from the two algorithms are analyzed. This overview presentation describes the main elements of this collaborative project, linkages and dependencies between the various research components and initial results over selected samples of the data record.

Session:  Fire - Fire, drought, and changes in vegetation structure and composition

Presentation Type:  Oral