IAM ‐ FIRE : A Climate Emulator–Based Framework to Project Wildfire Impacts and Risks for Integrated Assessment Models

ABSTRACT

Most Integrated Assessment Models (IAMs) underrepresent dynamic feedbacks from climate‐driven disturbances such as wildfires, potentially overestimating the permanence of land‐based carbon sinks. In particular, representing the impacts of forest fires is becoming increasingly important, as these are expected to intensify in the future. We introduce IAM‐FIRE (Integrated Assessment Model—Fire Impacts & Risks Emulator), a novel framework that enables the projection of wildfire burned area (BA) and carbon emissions (CE) directly from IAM outputs. IAM‐FIRE combines a spatial climate emulator, land‐use downscaling, vegetation productivity modelling, and an empirical fire model to generate global annual wildfire impacts for arbitrary socioeconomic and emissions scenarios at 0.5° resolution for the period 2020–2100. Calibrated against GFEDv5 observations and using inputs from the Global Change Analysis Model (GCAM), BA and CE projections are reported for four scenarios: SSP1‐2.6, SSP2‐4.5, SSP3‐6.6 and SSP5‐7.6. The model reproduces historical global trends for both total BA—including the observed global decline since the early 2000s—and forest BA. Projected fire trajectories differ strongly among scenarios: by 2100, total BA range from 441 Mha year ⁻¹ under SSP1‐2.6 (decline of −2.22 Mha year ⁻² relative to 2020) to 794 Mha year ⁻¹ under SSP3‐6.6 (increase of +2.3 Mha year ⁻² relative to 2020). Corresponding total CE show a similar divergence by 2100 ranging from 1.8 PgC year ⁻¹ in SSP1‐2.6 (decline of −9.15 TgC year ⁻² relative to 2020) to 3.6 PgC year ⁻¹ in SSP5‐7.6 (increase of +12.70 TgC year ⁻² relative to 2020). Socioeconomic development exerts a dominant suppressing effect on wildfire impacts while climate change and CO ₂ ‐driven increases in vegetation productivity amplify fire risk. Compared to CMIP6 and FireMIP, IAM‐FIRE exhibits greater sensitivity to radiative forcing and a stronger role for human‐driven fire suppression, highlighting structural uncertainties in fire projections. IAM‐FIRE enables systematic exploration of fire–climate–land feedbacks and supports improved assessments of mitigation permanence and climate risks in future integrated scenarios.