Fire and Savannization

Slow-moving fireline on forest floor

Fire is an important agent of transformation in the Amazon landscape. The role of fire in deforestation is widely known, but the impact of escaped fires is less well documented.

Every year, accidental understory fires damage a large percentage of Amazon forest. This is the phenomenon we have called 'cryptic deforestation' (Nepstad 1999). These escaped fires travel slowly (10-20 meters an hour) and only reach a few inches in height, but they can be quite destructive to understory vegetation and kill many larger trees. Once a forest has burned, increased leaf shedding, an abundance of branches and other coarse fuel enhance its fuel load, and an open canopy ensures the fuel will be dry enough to burn. Since settlement is typically a one-way process, the ignition source - humans - is there to stay, and successive burns tend to be more intense and destructive.

Cryptic deforestation due to forest fires and logging may affect as much forest area as deforestation in most years — and even more during periods of prolonged drought. For instance, during El Niño years, up to 25,000 km² of Brazilian forests may be affected by fire. Nevertheless, the overall importance of these forest fires is still poorly understood because it is difficult to detect fire scars using satellite imagery analysis even a short time after a burn, and even more challenging to evaluate the extent of the damage. We know a bit more about the impacts of fire from field studies.

Accidental fire advances through a transitional forest in Mato Grosso, 2004

For years we have conducted research on the impact of fires and fire frequency in Amazon forests. We have quantified the environmental and economic costs of accidental fires and have studied the relationship between drought and the susceptibility of Amazon forests. We have recently begun a field experiment to learn about the impact of repetitive fires on forest, and the likelihood of forests being replaced by savanna-like vegetation along the edges of the Amazon basin, where forests give way to a shrubby savanna-like region known in Brazil as the Cerrado. We hope this experiment will help us to understand the fragile balance of the savanna/forest boundary as well as the mechanisms that may trigger a runaway expansion of the savanna, and retraction of the edges of the Amazon rainforest.

©Woods Hole Research Center, 2007