Future climate-driven escalation of Southeastern Siberia wildfires revealed by deep learning

By using reanalysis data and climate model output together with a deep learning model, the study explores the relationship between positive-phase North Atlantic Tripole (NAT) sea-surface temperature anomalies and Southeastern Siberia (SES) wildfire increases and project the future trend in SES wildfire intensities under climate change. The researchers found that the positive-phase April NAT enhances the Siberian anticyclone, causing increased temperatures and snowmelt via strengthened transport of warm-air advection into the SES region.

The latter process heightens the exposure of local high-density peatlands to favorable conditions for fire ignition and leads to more intensive wildfire incidents. The study further demonstrates that the projected NAT variations can drive interdecadal changes in future April SES wildfires. With future phase shifting of NAT modes under global warming, the regionally averaged burned area in SES could be increased by 47–62% under different warming scenarios from 1982–2014 to 2015–2100. Our findings reveal the climate-driven escalation of future wildfires in SES in the context of global warming and call for active and urgent fire management strategies to mitigate the fire risk.