Modelling Amazon fire regimes under climate change scenarios

Abstract

Fire is one of the most important disturbances of the earth-system, shaping the biodiversity of ecosystems and particularly forests. Climatic change and other anthropogenic drivers such as deforestation and land use change could produce abrupt changes in fire regimes, potentially triggering transition from forests to savannah or grasslands ecosystems with large accompanying biodiversity losses. The interplay between climate change and deforestation might intensify fire ignition and spread, potentially giving rise to more extensive, intense, and frequent fires, but this is highly uncertain. We use a simple forest-fire model to analyze the possible changes in the Amazon region's fire regime that depend on climate change-related variables. We first explored the model behavior and found that there are two possible regime changes: a critical regime that implies high variability in fire extent and mega-fires, and an absorbing phase transition which would produce the extinction of the forest and transition to a different vegetation state. We parameterize the model using remote sensing data on fire extent and temperature, and show that it demonstrates proficiency in predicting past fires. Upon considering 21st-century climate projections and deforestation scenarios, our findings suggest that the Amazon region is not currently nearing any of these regime changes but predict a consistent increase in fire extent mainly induced by deforestation. Therefore, stopping deforestation could be an important factor in reducing the potential for drastic alterations in tropical forests of the Amazon region.