COS 198-5 - Simulated fire spread in reburned boreal deciduous forests limited by wind speed and fuel connectivity, not fuel availability

Abstract: Fire frequency in the boreal forest has increased via longer burning seasons, dryer conditions, and higher temperatures. But fires have historically also self-regulated via fuel limitations, mediating the effects of changes in climate and fire weather. Post-fire landscapes are believed to enable negative feedbacks to future burning, via the lower flammability of regenerating, typically deciduous shrub-dominated landscapes compared to their intact, conifer counterparts. However, the strength of that self-regulation remains unclear in the context of changing fire weather and emerging novel forest communities and structures. One step in evaluating potential future fire behavior in emerging deciduous landscapes is quantifying the abundance, composition, and structure of fuels across a landscape scale to understand how these unique patterns of fuel influence the rate and sustainability of fire spread and fuel consumption. We quantified fuel composition, abundance, and structure in stands in Interior Alaska after 1, 2 or 3 fires in short-intervals, and then created virtual landscapes and simulated fire behavior using the Wildland-Urban Interface Fire Dynamics Simulator (WFDS), a physics-based fire behavior model. WFDS treats vegetation composition and structure in three-dimensions, accounting for bulk density, surface area to volume ratio and fuel moisture and allowing for explicit combustion/atmospheric interactions. Using this modeling framework, we modeled a suite of fire behavior metrics, calculating wind velocity, fire intensity, fire consumption and fire velocity. These metrics are direct links to fire behavior properties that direct subsequent total area burned, fire severity, and fire management conditions and follow directly from the structures measured in the field. After a single fire, extreme fire weather conditions allowed for sustained fire spread, suggesting that intense fire conditions can enable short-interval events. However, fire spread was not sustained in thrice-burned landscapes, where regeneration was often dense but more clumped and thus overall less connected, separated by patches of bare soil. Fire traveled an average of 50 meters into thrice-burned landscapes before dying out, even under extreme fire weather conditions. Further modeling work is needed to determine the sensitivity of the model results to the specific spatial patterns and wind conditions used, and future fieldwork needed to determine how fine fuels respond to interannual variation in temperature, precipitation, and other environmental factors. This work suggests that the negative feedbacks observed today may persist in reburned landscapes 10-15 years postfire.