Quantifying Fire Behavior Prediction Uncertainty Associated with User-Defined Variables in WFDS

Coupled fire–atmosphere models (CFAMs) are increasingly proposed as an important tool for investigating a range of scientific and management questions, including the design of fuel management strategies. Uncertainty in CFAM outputs arises from environmental and fuel inputs, and a host of user-defined simulation choices such as fire approach angle and ignition timing. In this study, we used the Wildland–Urban Interface Fire Dynamics Simulator (WFDS) to quantify uncertainty in rate of spread and canopy consumption across pre- and post-restoration ponderosa pine stands. Two ensembles were conducted: (1) varying fire approach angles across 12 rotations and (2) varying ignition time in 80 simulations (five per plot, with 0 s delay, 250 s delay, and three random intervals in between). Metrics evaluated were rate of spread and percent canopy consumption. Variability was quantified using coefficients of variation (CVs). Approach-angle variation produced a mean CV of 5.46% for rate of spread, with treated stands exhibiting reduced variability relative to untreated stands. Canopy consumption showed a mean CV of 6.16%, with treatment having no effect. Ignition-time variation produced smaller CVs (rate of spread: 2.47%; canopy consumption: 3.60%) with no differences between management conditions. These results indicate that user-defined configuration choices contribute measurable but modest uncertainty, and that structural modifications from management can reduce sensitivity of fire spread. Incorporating these sources of uncertainty into formal frameworks will improve interpretation of CFAM outputs for operational and research applications.