he impact of fire in Amazonian forests on the global carbon cycle is poorly understood. During El Nino periods, severe drought conditions render large areas of Amazonian forest susceptible to burning (Nepstad et al., 1999a); understory fires burned an estimated 90,000 km2 of forest during the 1998 dry season (Alencar, Moutinho, and Nepstad, unpublished data). The proposed research will augment a predictive model of forest flammability in Amazonia, RisQue, by incorporating measurements of the period between rain events and the effects of smoke on radiation load and fuel drying rates. Consecutive rainless days is an important indicator of forest flammability in different vegetation types in Amazonia (Uhl and Kauffman, 1990). Smoke from biomass burning may act as a negative feedback on further burning by reducing the radiation load by as much as 25 percent, thereby allowing air vapor pressure deficits to fall, and fine fuel moisture content to remain high (Nepstad et al., in press).
Geostationary Observational Earth Satellite (GOES) data are ideal for determining the number of rainless days and tracking smoke plumes since GOES-8 provides image data for Amazonia every half hour. Less frequent data, regardless of improved spatial resolution, would not permit an accurate estimation of these flammability components. Therefore, this study will focus on the following three challenges:
- Developing an algorithm to measure the number of rain-free days based on historic GOES-8 images;
- Calibrating the algorithm with field data from automated weather stations operated by the Brazilian Center for Weather and Climate Prediction (CPTEC) and from experimental fires (n=600) set during 1998 and 1999 (Nepstad et al., unpublished data);
- Using historical GOES-derived maps of smoke plumes and weather station data to quantify smoke effects on air vapor pressure deficit.
Predicting forest flammability in the Amazon region is a key component of responsible policy-making and management in the region, and incorporating these measurements into RisQue using remotely sensed data will permit more dynamic modeling of forest flammability during severe dry seasons.
Additional project components: Measurement and analysis of 1999 ETM+ Landsat data for the transitional forest region of the southern Amazon basin in the state of Mato Grosso (22 scenes). The correlation between burned areas and various factors of development will compliment the climatic components of flammability in this region detailed above. The goal of this project component is to assess the total area of burned transition forest in 1999, delineating changes in vegetation structure and range based on ground-truthing efforts in July/August 2001. Finally, an integrated assessment of climate-based flammability and development-induced flammability will be completed for this region of the southern Amazon. This integrated understanding of flammability will provide critical information for the municipio-level governments in this region and provide necessary insight for future development proposals in this region.
An uncontrolled pasture fire spreading to adjacent forest during the dry season of 1999 in Mato Grasso, Brazil as seen from space.