全球变化知识资源中心

Worldwide, fires have substantial economic, social and health-related impacts. Brazil is one of the most affected areas in the globe, particularly the Cerrado, a savanna-like biome, whose composition, structure, species abundance and diversity are shaped by recurring fires. The aim is to assess present and future trends of fire danger and burned area (BA), using the Daily Severity Rating (DSR), an extension of the Canadian Forest Fire Weather Index System, and climate outputs from a regional climate model, the RCA4 from the Rossby Centre. To that end, we validated the climate variables simulated by RCA4 and the resulting DSR, both showing consistency with observation-based datasets. We then developed a statistical model of BA using fire season averaged DSR as a predictor, found to explain 71% of the interannual variability of BA from 2003 to 2017. Using the statistical model, we projected future fire danger and BA over Cerrado for IPCC Representative Concentration Pathways 2.6, 4.5 and 8.5. Results show an increase in future BA for all scenarios, with pronounced changes for RCPs 4.5 and 8.5, where BA is expected to increase by 39% and 95% for 2100. In the case of RCP 2.6, the closest scenario to the 1.5 degrees C target established by the United Nations, results indicate an increase in mean BA up to 22% by 2050, compared to the historical period, followed by a decrease to 11% by 2100. This is especially relevant since RCP 2.6 is the only scenario where such a decrease is projected, highlighting the importance of keeping global mean surface temperature below the 1.5 degrees C warming target.