Unprecedented recent burning of the Alaskan boreal forest points to shifting controls on fire: A paleoecological data-model investigation

Fire is the primary disturbance in the boreal forest, and variations in fire regime have important ecological, biogeochemical, and socioeconomic implications. Climate has been a dominant control on boreal wildfire throughout the brief observational record, and recent warming is associated with widespread burning. However, fire-regime dynamics are highly scale-dependent, and it is unclear whether the observed climate-fire relationship is applicable to long-term projections. Paleo-fire reconstructions offer a valuable extension to historical fire records, and simulation models allow prescribed ecosystem processes to be played out through hypothetical scenarios. Together, the two approaches offer unique insight: reconstructions can validate long-term model behavior, which then becomes a credible basis for inference on the unobserved drivers of past and potential future fire-regime shifts. To infer fire-regime variations over the past 10,000 years, we created a composite charcoal record from eleven lakes in the Alaskan boreal forest. We compared the record to fires documented in the 60-year observational fire history, and to a 500-year fire history simulated by ALFRESCO, a stochastic model of boreal-forest fire dynamics. This paleodata-observation-model comparison allows us to (1) evaluate recent burning in the context of Holocene variability, (2) solidify the link between proxy charcoal data and actual fire regime characteristics, (3) validate centennial-scale model behavior, and (4) explore nonlinear fire-climate-vegetation interactions over decadal to millennial time scales. Our composite record reveals that recent charcoal production is unprecedented in the context of Holocene variability. Agreement among simulated, observed, and reconstructed fire histories implies that this charcoal signal is indicative of peak burning rates in the past several decades, likely driven by warm and dry summer conditions. The simulation results also reveal emergent long-term system dynamics not captured in direct observations to date. When the climatic warming of the past 50-100 years is prescribed to ALFRESCO, extensive burning eventually depletes standing late-successional forest, reducing flammability at the landscape scale. This vegetation feedback dampens the direct effect of rising temperature on regional fire-regime characteristics. Thus, although boreal-forest burning is demonstrably climate-driven in both the recent historical record and ALFRESCO simulations, continued warming may promote a shift to a substantially fuel-limited system. The composite charcoal record shows that such a status has not been reached previously, suggesting additional novelty of modern conditions in the boreal forests of Alaska. Our results highlight the importance of historical perspective and integrative understanding of processes and feedbacks in long-range forecasts of ecosystem dynamics.