We developed models of suppression expenditures for individual extended attack fires in British Columbia using parametric and nonparametric machine-learning (ML) methods. Our models revealed that suppression expenditures were significantly affected by a fire’s size, proximity to the wildland–urban interface (WUI) and populated places, a weather based fire severity index, and the amount of coniferous forest cover. We also found that inflation-adjusted individual fire suppression expenditures have increased over the 1981 to 2014 study period. The ML and parametric models had similar predictive performance: the ML models had somewhat lower root mean squared errors but not on mean average errors. Better specification of fire priority as well as resource constraints might improve future model performance.

Global climate change is leading to increased wildfire activity in many parts of the world, and with increasing development, a heightened threat to communities in the wildland urban interface. Evaluating the potential for fire to affect communities and critical infrastructure is essential for effective response decision-making and resource prioritization, including evacuation planning, with changing weather conditions during the fire season. Using a receptor–pathway–source assessment framework, we estimate the potential source area from which a wildfire could spread to a community in British Columbia by projecting fire growth outward from the community’s perimeter. The outer perimeter of the source area is effectively an evacuation trigger line for the forecast period. The novel aspects of our method are inverting fire growth in both space and time by reversing the wind direction, the time course of hourly weather, and slope and aspect inputs to a time-evolving fire growth simulation model Prometheus. We also ran a forward simulation from the perimeter of a large fire that was threatening the community to the community edge and back. In addition, we conducted a series of experiments to examine the influence of varying environmental conditions and ignition patterns on the invertibility of fire growth simulations. These cases demonstrate that time-evolving fire growth simulations can be inverted for practical purposes, although caution is needed when interpreting results in areas with extensive non-fuel cover or complex community perimeters. The advantages of this method over conventional simulation from a fire source are that it can be used for pre-attack planning before fire arrival, and following fire arrival, it does not require having an up-to-the-minute map of the fire location. The advantage over the use of minimum travel time methods for inverse modeling is that it allows for changing weather during the forecast period. This procedure provides a practical tool to inform real-time wildfire response decisions around communities, including resource allocation and evacuation planning, that could be implemented with several time-evolving fire growth models.

Burned area, fire severity, and suppression expenditures have increased in British Columbia in recent decades with climate change. Approximately 80% of suppression expenditures are attributable to wildfires near the Wildland–Urban Interface (WUI). Evaluating the potential for fuel management to reduce suppression expenditures is essential to mitigating demands on fire response resources and reducing impacts on communities. One management approach is to increase the proportion of deciduous tree species, which have a lower propensity for crown fire. Using fire suppression expenditure data from 1981 to 2014, we applied the machine learning method causal forests (CFs) to estimate the effect of the proportion of conifer forest cover on suppression expenditures for WUI fires and how these effects varied with other influential factors (i.e., heterogenous treatment effects). Across all fires, the effect of conifer cover on suppression expenditures was stronger on private land compared to public land, under high fire danger measured by daily severity ratings (DSRs), which reflect wind speed and fuel moisture, and for fires igniting earlier in the calendar year, based on Julian day. These findings provide insights into prioritizing wildland fuel treatment when budgets are limited. The CFs approach demonstrates potential for broader applications in fire risk mitigation and analysis beyond the scope of the current data. CFs may also be valuable in other areas of forest research where heterogenous treatment effects are common.

Surface fire frequently occurs beneath the canopy of North American forests under moderate wind speed and moisture-deficit conditions. Surface rate of spread (sROS) models can provide guidance for suppression operations and can be incorporated into fire growth modelling systems and other tools. We used a database of primarily Canadian experimental surface fires in conifer and deciduous stands from multiple sites to fit empirical sROS models for operational use and compare with pre-existing models. Various predictor combinations represented fires in boreal conifer (BOCON), deciduous, and Ponderosa pine-dominated stands, the latter analyzed to estimate grass-curing influence. The main predictors were wind speed (WS₁₀), estimated fuel moisture, and Canadian Fire Weather Index (FWI) System components (original and stand-adjusted). The ensuing fitted models (N = 51–93) were evaluated using standard metrics and tested using an independent conifer dataset (N = 26). The simplest model finds BOCON sROS to be equal to 1.2% of the WS₁₀, 1/7th the speed of crown fire spread under similar conditions; it is easily calculated as 20% of WS₁₀ using a common unit conversion (WS₁₀ in km h⁻¹, sROS in m min⁻¹). The best-performing sROS models displayed nonlinear-sigmoidal responses to wind and litter moisture variables, including the Initial Spread Index (ISI), and improved upon pre-existing models. Estimated accuracy was mostly +/− 2–4 m min⁻¹ within the range of the data in both training and validation datasets. These models reflect a dataset gathered from multiple sites using varying experimental methods. While imprecise, they are suitable for many applications, including operational forecasting and designing hazard reduction treatments.