Climate-Driven Changes in Wildfire Hazard and Implications for Resilient Building Design

Abstract

Wildfires pose a significant hazard to buildings and communities located at the wildland–urban interface (WUI) in Canada. Climate change is expected to intensify the duration, frequency, and severity of the wildfires. Current hazard assessments rely on historical conditions and may underestimate future hazard. This study adjusts fire intensity in national wildfire hazard maps to reflect projected changes in fire weather. Analysis is conducted for 393 National Building Code of Canada (NBC) reference locations under a 2.5 °C of global warming scenario, which corresponds to a 50-year future time-frame, a typical design life of buildings. The results show a strong upward shift in national hazard, with the number of locations in the “High” hazard class nearly doubling from 28 to 53. These findings highlight the need to integrate climate-informed hazard projections into future hazard mapping, building codes, and resilience planning. To date, no large-scale Canadian study has examined how climate-driven changes in wildfire hazard may influence the application of building design guidance at the national scale. This study provides an assessment of hazard sensitivity to climate change, highlighting the importance of considering projected fire weather conditions in national hazard assessments.

1. Introduction

Wildfires have increasingly devasted communities at the wildland–urban interface (WUI). The WUI is the area where human developments meet or are intermingled with wildland fuels or can be impacted by the heat transfer mechanisms of a wildfire [1]. In the last decade, Canada saw more than 2400 structures lost to WUI fires, including nearly 1600 homes destroyed in the 2016 Fort McMurray wildfire, which is the costliest disaster in Canadian history, with an estimated $9 billion in damages. More recently, the fires in Jasper and Lytton in 2024 and 2021 caused extensive damage to the built environment. As the climate changes, it is expected that the frequency and intensity of wildfires will increase as higher temperatures and lower precipitation lead to more favourable wildfire conditions [2]. Climate change is also projected to promote storm development, leading to a rise in lightning activity by up to 80% by the end of the 21st century [3]. This, combined with warmer and drier conditions, is expected to drive significant increases in total area burned [4,5], wildfire frequency [6], and the occurrence of high-intensity crown fires [2] across Canada. A recent study across 11 Canadian cities shows that wildfire season length is projected to increase by up to 61 days under 3.5 °C of global warming [7]. As a result, an increasing number of wildfire events are expected to exceed current suppression capabilities and pose greater hazard to the built environment.

In the face of this growing threat, there is an urgent need to enhance the resilience of buildings and communities in fire-prone areas. The National Research Council of Canada has recently developed a National Guide for Wildland–Urban Interface Fires [1] to support the integration of wildfire resilience into buildings and community design. The Guide synthesizes the latest research and lessons learned from recent WUI fire disasters. Key elements include assessing wildfire hazard and exposure levels based on national hazard maps. It also outlines strategies for vegetation management and the use of fire-resistant construction materials to reduce the likelihood of structure ignition.

The WUI guide is a framework that links the wildfire hazard level of a location to specific building requirements. Each location is assigned a hazard level based on regional fuel types, topography, historical weather, and ignition patterns. In addition, each building’s likelihood of encountering embers, radiation, or flames is assessed as the exposure. Combining the hazard and exposure, the guide specifies a construction class (CC) for the building, ranging from CC3 (least stringent) up to CC1 (most stringent), which dictates the recommended fire-resistant design measures. Higher hazard and exposure lead to more rigorous building requirements, while extensive vegetation management around a building can reduce the demands of the recommended construction measures. This approach ensures that building features (roofing, siding, vents, windows, etc.) are appropriate to the hazard and improves the chances of it surviving a fire. However, the WUI guide is currently a voluntary framework in Canada, but efforts are underway to translate it into codes and standards that provinces or municipalities could adopt. While these guidelines can help minimize current wildfire risks, changes in the climate will result in evolving fire weather patterns. Many regions that are classified as low wildfire hazard may face much higher hazard under future climate conditions. For instance, ref. [8] linked the extreme 2023 Canadian wildfire season to anomalous heat and dryness, which are likely to become more common in a warming climate. Therefore, to ensure long-term resilience, it is crucial to incorporate climate projections into the planning and design of buildings for wildfire resilience.

Modeling and simulation tools are often used to estimate how wildfire behaviour may evolve under climate change. For example, a common approach is to calculate fire danger indices or fire behavior models with climate scenarios from Global Climate Models (GCMs). For example, ref. [2] employed multiple GCM projections under various emission pathways together with the Canadian Forest Fire Behaviour Prediction (FBP) System to simulate future fire environments across Canada. The FBP System is a key component of Canada’s fire danger rating system. It provides quantitative estimates of potential wildfire spread rate, fuel consumption, and intensity based on fuel type, weather, and terrain [9]. By inputting projected future weather into such models, one can estimate increases in fire behavior indices at a given location. Studies consistently show significant increases in these indices under climate change scenarios, implying that many areas will experience more frequent extreme fire behavior. For instance, ref. [10] used downscaled and bias-corrected outputs from four CMIP5 Earth System Models to evaluate future changes in fire danger and severity across the Canadian boreal forest. The study found that under RCP 8.5, fire danger metrics such as the Fine Fuel Moisture Code and Daily Severity Rating are projected to increase significantly, particularly in the second half of the century. The results suggest that large portions of the boreal forest will experience more days with high fire danger, pointing to increased risk to ecosystems and communities.

In addition, probabilistic fire simulation models are widely used to quantify wildfire hazard across large areas by estimating the likelihood and potential intensity of fire occurrence under a wide range of weather and ignition conditions. These models are designed to represent the stochastic nature of wildfire by simulating thousands of potential fire events, each initiated at random ignition locations and propagated under sampled fire weather sequences. Fire spread and behaviour are calculated as a function of key inputs, including fuel type, fuel continuity, topography, and daily fire weather variables such as fuel moisture and atmospheric conditions. The resulting simulations capture the combined influence of landscape characteristics and weather variability on fire growth, allowing for the derivation of spatially explicit metrics such as burn probability and head fire intensity. In Canada, Burn-P3 is one commonly used tool for this purpose. Burn-P3 couples a fire ignition framework with the Prometheus fire growth engine [11], which implements the FBP System to compute fire rate of spread, fuel consumption, and fire intensity for different fuel types and terrain conditions. Burn-P3 typically relies on historical fire weather records to represent the range of conditions under which fires may occur and is therefore well suited for characterizing long-term wildfire hazard rather than predicting individual fire events. The model has been applied extensively to support hazard mapping, land-use planning, and wildland–urban interface assessments. For example, ref. [12] used Burn-P3 to generate national-scale wildfire hazard maps for Canada by combining simulated burn probability and head fire intensity to characterize spatial patterns of wildfire hazard under historical climate conditions. Ref. [13] provided a comprehensive review of Burn-P3 and related simulation approaches, highlighting their application in fuels management, habitat conservation, WUI exposure analysis, and, increasingly, climate change impact assessments.

Previous studies have demonstrated that wildfire activity is influenced by interacting changes in fire weather, ignitions, and vegetation, and that fully dynamic modeling of these processes can provide detailed insights into future fire regimes [14]. However, implementing these approaches consistently at national scales remains challenging, as it requires downscaled, high-temporal-resolution climate data, future fuel maps, and extensive model calibration. Additionally, even when such models are applied, projected changes in burn probability can vary substantially across climate models and locations due to differences in the magnitude, seasonality, and interaction of temperature, precipitation, and humidity changes [15]. These complexities complicate direct comparison between historical and future simulations and highlight the uncertainty inherent in fully dynamic fire regime projections, particularly when the objective is to assess changes in wildfire hazard relevant to national-scale planning and building design. To address these challenges, some studies have adopted pseudo-global-warming approaches in which observed fire events or historical simulations are perturbed using climate-driven adjustments to isolate the influence of warming on fire behaviour while holding other components of the fire regime constant [16,17]. This approach enables the use of established wildfire hazard baselines while isolating the effect of climate-driven changes in fire intensity. Using this framework, the Senande-Rivera et al. demonstrated that anthropogenic warming has already intensified wildfire rate of spread and fire intensity, without attempting to simulate future fire occurrence or changes in burn probability.

Building on this perspective, the present study uses existing national wildfire hazard maps developed by [12] as a baseline to examine how projected changes in fire weather may influence wildfire hazard classifications at National Building Code reference locations under a 2.5 °C global warming scenario. Rather than simulating future fires directly, we apply climate-driven adjustments to historical head fire intensity to assess how increases in fire intensity alone could alter hazard levels. This approach allows for a consistent, national-scale evaluation of climate-driven hazard effects and their implications for applying the National Guide for Wildland–Urban Interface Fires. The results are used to identify locations where projected changes in wildfire hazard may influence recommendations for fire-resistant materials and vegetation management, emphasizing the national-scale application of the Guide.

2. Methods and Data

2.1. Study Area

This study was conducted across Canada at reference locations listed in the National Building Code (NBC), with analysis focused on 393 of the 680 Table C-2 sites, shown in Figure 1. These locations span a wide range of ecological zones, from coastal rainforests in British Columbia to mixed wood forests in eastern Canada and boreal forests across central and northern regions. There are a number of NBC locations that were not analyzed due to data availability constraints. Vegetation types, topography, and fire regimes vary substantially across the country, reflecting strong regional gradients in climate and land cover.

Western Canada is characterized by complex terrain and fuel-rich forests, where steep slopes, episodic drought, and wind-driven fire behaviour contribute to high-intensity fires and elevated wildfire hazard, particularly along the wildland–urban interface. Central and northern regions are dominated by boreal forest ecosystems, where large, infrequent fires account for most of the annual area burned and fire activity is strongly controlled by synoptic-scale fire weather and fuel continuity. In contrast, eastern Canada generally experiences more humid conditions and lower fire frequencies, although extended dry periods can still support high-intensity fire behaviour under extreme weather conditions. Southern Canada, where most NBC reference locations are concentrated, is influenced by warmer temperatures, higher population density, and increasing development near forested landscapes, contributing to heightened exposure even in regions with historically moderate fire activity.

While the majority of wildfire activity and area burned in Canada occurs in boreal and sub-boreal regions, these fires often occur in sparsely populated areas. In contrast, population exposure to wildfire hazard is concentrated in southern Canada, where most communities and NBC reference locations are located near forested landscapes. Western Canada, in particular, has experienced repeated high-impact wildland–urban interface fires due to the proximity of communities to fire-prone forests combined with topographic and climatic conditions conducive to extreme fire behaviour.

Case Study

To characterize wildfire hazard conditions surrounding each NBC Table C-2 location, we developed a generalizable and efficient spatial analysis approach for national-scale assessments. For each NBC Table C-2 location, we extracted geospatial data within defined buffer zones. For each site, a 5 km buffer was created around the location’s coordinates to define the region of interest and identify urban land cover cells using the FBP fuel type raster from Natural Resources Canada. This was done to identify areas which may reasonably be deemed to be a part of the same community. WUI environments are highly heterogeneous. However, the urban classification is used only to identify community extents and does not imply uniform wildfire exposure or fuel conditions. These urban cells were then used to extend a secondary 15 km buffer zone, intended to capture the surrounding area most relevant to potential wildfire assessment for the site, as shown in Figure 2. Prior studies indicate that analyzing fuel composition and slope within ~15 km provides an effective scale for representing both immediate and advancing wildfire threats [7,18]. The 5 km buffer provides a pragmatic representation of community extent at the scale of NBC reference locations, while the 15 km buffer captures the broader landscape conditions that influence the potential for large, high-intensity fires capable of generating ember exposure. This approach allows for consistent characterization of regional wildfire hazard across Canada. Within this 15 km zone, key data layers, including wildfire hazard, burn probability, fire intensity, fuel type, and slope were extracted for this area. Where no valid data were present within the buffer, the site was excluded from further analysis.

2.2. Data

This study relies on a combination of static geospatial datasets and dynamic climate model projections to evaluate current wildfire hazard conditions and estimate how these may change under future global warming scenarios. The analysis incorporates historical fire simulation outputs, landscape characteristics, and future fire weather index (FWI) derived from a multi-ensemble climate dataset.

2.2.1. Fuel and Topography

Fuel and topography characteristics were extracted from Natural Resources Canada’s Canadian Forest Fuel Type Grid (CanFG), a national-scale fuel classification dataset with a 250 m resolution. Slope data, representing the steepness of terrain in degrees, were acquired from the same NRCan dataset and are used to account for the influence of topography on fire spread and intensity.

2.2.2. Wildfire Hazard

Baseline wildfire hazard data were sourced from the national-scale outputs developed by [12], who applied the Burn-P3 fire simulation model to assess wildfire hazard across Canada under historical climate conditions. Burn-P3 simulates thousands of stochastic fire events for each location, incorporating random ignition points, fire weather patterns, fuel types, and terrain conditions to estimate potential fire behavior over time. From these simulations, ref. [12] derived three primary outputs used in this study: burn probability, head fire intensity, and wildfire hazard.

Burn probability represents the proportion of simulations in which a given pixel was burned, quantifying the likelihood of fire occurrence. Head fire intensity (HFI) is expressed in kilowatts per meter, capturing how severe a fire is likely to be at its leading edge, based on local fuels, slope, and fire weather conditions. Higher HFI values indicate greater potential for extreme fire behavior. The wildfire hazard combines both burn probability and fire intensity to assign a hazard level reflecting the combined likelihood and severity of fire occurrence.

Erni et al. conducted these simulations using historical climate data from 1970 to 2017 and assumed static fuel conditions based on observed vegetation and land cover. By running thousands of simulations across Canada’s diverse fire regimes, the resulting maps offer a consistent baseline wildfire hazard at the national scale. In this study, Burn-P3 outputs are used solely as a spatially consistent baseline representation of relative wildfire hazard, rather than as a predictive model of future fire behaviour.

2.2.3. Future Projections of Wildfire Weather

Climate projection data were obtained from the Canadian Large Ensemble Adjusted Dataset version 1 [19], which provides daily climate projections from the Canadian Regional Climate Model v4 across Canada [20]. The data are bias-corrected using a multivariate approach which adjusts the marginal distributions of each variable and preserves realistic inter-variable relationships [21]. Bias adjustments are calibrated based on a high-quality reanalysis dataset, EWEMBI, ensuring the modeled data closely reflects real-world historical climate. The climate variables include temperature, precipitation, relative humidity, wind speed, surface pressure, and radiation. These projections were used by [22] to produce future Canadian Fire Weather Index (CanLEAD-FWI) values under different warming scenarios. From this 50-member ensemble, we compared the average FWI from 1950–2100 and selected one ensemble that represents the most severe fire weather conditions across Canada to illustrate our approach. The analysis compares a historical reference period (1986–2016) with a future period (2054–2084), corresponding to approximately 2.5 °C of global mean warming under RCP8.5. This warming level was selected because it aligns with the typical 50-year design life of buildings and is consistent with time horizons commonly used in the National Building Code of Canada to assess climate change impacts on design-relevant hazards. Framing the analysis in terms of a warming level, rather than a specific calendar year, also allows the results to remain relevant across different emissions pathways. The selected ensemble member represents a high-impact but plausible future fire-weather scenario and is intended to support stress-testing of wildfire hazard classifications relevant to long-term building resilience, rather than to provide probabilistic forecasts of future wildfire occurrence.

2.3. Climate-Driven Changes in Wildfire Hazard

To estimate climate-driven changes in wildfire hazard classification, we used a scaling approach that applies a Climate Change Factor (CCF) to baseline data developed by [12]. This method adjusts baseline head fire intensity values to reflect changes in fire intensity associated with projected fire weather conditions, rather than simulating future fire occurrence or spread. The overall methodology used to derive climate-adjusted wildfire hazard is illustrated in Figure 3. Burn probability was held constant to isolate the influence of climate-driven changes in fire intensity on hazard classification, rather than to project future wildfire occurrence or frequency.

CCFs were calculated for each location by comparing HFI values derived from climate-informed fire weather conditions between historical and future periods. Specifically, daily fire weather indices from the Canadian Forest Fire Weather Index System, including the Fine Fuel Moisture Code (FFMC), Initial Spread Index (ISI), and Buildup Index (BUI), were computed using bias-corrected climate projection data. These indices are driven by underlying climate variables, including daily temperature, precipitation, relative humidity, and wind speed. The resulting FBP outputs provide gridded HFI time series, from which CCFs were computed as the ratio of the 30-year mean HFI under future and historical conditions. These factors were interpolated to a 250 m resolution corresponding to the fuel grid. For each NBC location, CCFs were spatially matched to the surrounding area, and climate-adjusted HFI was calculated as the product of baseline HFI and the corresponding CCF. Consistent with perturbation-based approaches commonly applied in regional climate modeling [23,24], this method estimates the sensitivity of fire intensity to climate change while preserving the spatial structure of historical hazard patterns.

2.4. Construction Classes

Construction classes define a tiered set of wildfire-resistant building features, ranging from CC3 (least stringent) to CC1 (most stringent), and are intended to align the level of construction mitigation with a building’s wildfire risk (

Table A1. Determination of construction classes. [FR1] and [FR2] denote a variation in CC1 as outlined in the Guide.

Exposure Level	Recommended Construction Classes for Use with Mitigation Measures Applied in the Listed Priority Zones
	None	1A	1A and 1	1A to 2	1A to 3
Ember-Only or Low	CC1 [FR1]	CC1	CC3	CC3	CC3
Moderate	CC1 [FR1]	CC1 [FR2]	CC2	CC3	CC3
High	CC1 [FR1]	CC1 [FR2]	CC1	CC2	CC3

). These classes guide both new builds and retrofits in fire-prone areas, serving as a tool to reduce structure ignition potential and improve survivability during wildfire events. According to

Table A2. Determination of exposure level from hazard level and exposure, cells color coded as green, yellow, orange, and red for Nil–Very Low, Nil, Low, and High classes.

Hazard Level	Exposure
	Nil	Low	Moderate	High
Nil–Very Low	Nil	Nil	Nil	Nil
Low	Nil	Low	Low	Moderate
Moderate	Nil	Low	Moderate	High
High	Nil	Low	Moderate	High

, exposure level is determined by cross-referencing a site’s hazard level with its assessed exposure. For example, assuming high exposure, a site that transitions from low to moderate hazard would move from moderate to high exposure level. Since the CC depends on the highest exposure level, this could elevate the building’s construction requirements from CC2 to CC1, triggering more stringent and costly fire-resistant design elements such as non-combustible cladding, enhanced vent protection, and tempered glazing.

Determining a building’s construction class follows a stepwise process outlined in the National Guide for Wildland–Urban Interface Fires [1]. First, the wildfire hazard level of the surrounding landscape is established based on modeled fire behavior, local fuel types, and topography. Next, the exposure level of the building is evaluated, reflecting its likelihood of encountering embers, radiant heat, or direct flame, factors that depend on site conditions, such as vegetation clearance, slope, and the proximity of other structures. The intersection of these two factors, hazard and exposure, determines the building’s overall exposure level, as summarized in Table A1. This exposure level is then translated into a CC (Table A1), which specifies the corresponding wildfire-resistant building features conditioned on the application of mitigation measures.

3. Results and Discussion

3.1. Case Study

To illustrate the application of this approach, we highlight Lytton, British Columbia as a case study. Lytton was devasted by a wildfire in 2021 and is currently undergoing reconstruction, providing a valuable case study for evaluating how climate-informed hazard projections can support resilient building efforts. Wildfire probability, intensity, and hazard values for this site were extracted from the national Burn-P3 wildfire hazard dataset [12]. Figure 4 shows the historical fire behavior in the region during the baseline period (1970–2017). In the Lytton area, there is considerable spatial variation in modeled fire intensity, with many areas over 4000 kW/m due to dense fuels and steep topography. There are pockets in the area estimated to be of higher hazard, but overall, the hazard in Lytton is mostly moderate.

During the future period (2054–2084), we applied climate-adjusted head fire intensity (HFI) values derived from the CanLEAD-FWI ensemble to assess how wildfire hazard may evolve in a warming climate. The scenarios correspond to a global warming level of approximately 2.5 °C and captures the high end of future fire weather conditions across Canada. As shown in Figure 5, wildfire intensity is projected to increase substantially, with large portions of the buffer zone exhibiting HFI values above 10,000 kW/m. Consequently, much of the region surrounding Lytton transitions to high wildfire hazard.

Table 1. Number of grid cells in each wildfire hazard level across the entire study area during the baseline and future climate change scenario; brackets denote the percent increase in grid cells compared to the baseline scenario.

Climate Change Scenario	Nil–Very Low Hazard (% Increase)	Low Hazard (% Increase)	Moderate Hazard (% Increase)	High Hazard (% Increase)
Baseline	917	2724	7594	1999
High	755 (−18%)	1998 (−27%)	4594 (−40%)	5887 (294%)

summarizes the number of grid cells falling within each wildfire hazard class under the baseline and future scenario. The number of high hazard cells increase from 1999 to 5887. These changes are accompanied by large decreases in the number of cells classified as low or moderate hazard, indicating a strong shift in hazard severity across the landscape. While there is inherent uncertainty in projected fire weather due to natural climate variability, the trend clearly shows that the potential intensity of future wildfires, and the associated hazard, is expected to increase as the climate warms. However, factors not explicitly addressed in this study, such as changes in vegetation and ignition patterns driven by shifting climatic conditions could potentially lead to even greater hazard. Ignition sources may evolve under a changing climate. Lightning currently accounts for approximately 45–55% of wildfire ignitions in Canada but is responsible for nearly 80–90% of the total area burned, particularly across boreal regions where large fires dominate annual burned area [6,25]. Observational and modeling studies suggest that lightning activity may increase by roughly 10–30% per degree of warming, with the strongest increases projected for western and northern Canada under high-emission scenarios [3,25]. In addition, long-term analyses indicate that fire weather indices and area burned have already increased in several regions since the 1980s, with recent extreme seasons such as 2016, 2021, and 2023 occurring under anomalously warm and dry conditions consistent with broader warming trends [5,8,26]. Climate-driven vegetation shifts are also expected to alter fuel availability and structure, including northward forest expansion, increased drought stress in southern forests, and changes in fuel continuity following insect disturbance or prolonged drying [26,27]. While these processes vary regionally and were not explicitly modeled in this study, they may influence both ignition likelihood and burn probability in addition to fire intensity. Consequently, the projected hazard increases presented here, which isolate climate-driven changes in fire weather and intensity, likely represent a conservative estimate of future wildfire hazard.

3.2. National Scale

To illustrate the implications of failing to consider climate change in wildfire hazard assessments, we performed a national-scale analysis to identify locations where climate-driven changes in wildfire hazard may influence building design requirements. For each location, wildfire hazard was summarized across the 15 km buffer using the mode hazard class, representing the most prevalent hazard condition in the surrounding landscape. While this approach does resolve directional fire spread pathways or spatial connectivity, the dynamics are event-specific and depend on ignition location, wind direction, and suppression response, which are beyond the scope of a national-scale assessment. Instead, the objective is to characterize dominant background hazard conditions that define the broader fire regime to which communities are exposed over long time horizons. Using the mode avoids over-interpreting localized spread features that may be relevant for individual fire events but are not transferable indicators for national building design guidance. Figure 4 and

Table 2. The number of locations and their overall wildfire hazard at each NBC site during the baseline and projected climate change period calculated using the mode value in a 15 km radius.

Wildfire Hazard	Nil–Very Low	Low	Moderate	High
Baseline	168	83	114	28
Projected	158	76	106	53

provide a national summary of how wildfire hazard classifications change between the baseline and climate-adjusted scenarios across all NBC locations. Under baseline conditions, 114 locations are classified as Moderate hazard and 28 as High. Under climate-adjusted conditions, there is a clear upward shift in hazard, with the number of locations classified as High increasing to 53, while those classified as Nil–Very Low, Low, and Moderate decline.

Figure 6 shows the geographic distribution of wildfire hazard transitions across Canada, comparing baseline to projected hazard levels. Increases in hazard are consistent with warmer, drier summers, declining snowpacks, and conditions conducive to more intense fire behaviour [25,26,27]. This region also faces compounding factors from fuel changes to prolonged drought stress, and a growing wildland–urban interface, which create conditions for more intense and higher impact fires [28,29,30,31,32].

Table 3. The number of locations and their prescribed construction class during the baseline and projected period (in brackets). [FR1] and [FR2] denote a variation in CC1 as outlined in the Guide.

Exposure	Priority Zone Mitigation	CC1	CC2	CC3
Nil–Low	None	393 (393) [FR1]
Moderate	None	393 (393) [FR1]
High	None	393 (393) [FR1]
Nil–Low	1A	393 (393)
Moderate	1A	393 (393) [FR2]
High	1A	393 (393) [FR2]
Nil–Low	1A and 1			393 (393)
Moderate	1A and 1		142 (159)	251 (234)
High	1A and 1	142 (159)	83 (76)	168 (158)
Nil–Low	1A to 2			393 (393)
Moderate	1A to 2			393 (393)
High	1A to 2		142 (159)	251 (234)
Nil–Low	1A to 3			393 (393)
Moderate	1A to 3			393 (393)
High	1A to 3			393 (393)

illustrates the number of NBC locations assigned to each CC under different exposure assumptions and mitigation strategies (

Table A3. Number of locations in each exposure level given the mode hazard across the site, cells color coded as green, yellow, orange, and red for Nil–Very Low, Nil, Low, and High classes.

Hazard Level	Exposure
	Nil	Low	Moderate	High
	Exposure Level
Nil–Very Low	168 (158)	168 (158)	168 (158)	168 (158)
Low	83 (76)	83 (76)	83 (76)	83 (76)
Moderate	114 (106)	114 (106)	114 (106)	114 (106)
High	28 (53)	28 (53)	28 (53)	28 (53)

), comparing baseline and future hazard scenarios. Without any mitigation, all locations remain in CC1 regardless of hazard changes, demonstrating that under those configurations, CC assignment is unaffected. However, with increasing implementation of more comprehensive mitigation, the distribution of CCs becomes more variable. Under high exposure and mitigation in zones “1A and 1,” for instance, the number of locations requiring CC1 increases from 142 to 159 in the future, whereas if the fuel mitigation measures outlined in the Guide are taken up to Zone 2, then no locations need to build as stringently as CC1. This shift reflects the ability of mitigation actions to offset rising hazard and avoid escalation to more extreme construction requirements. These results highlight the economic potential of applying WUI mitigation strategies early on. By investing in vegetation management, defensible space, or other site-level measures, Canada may be able to limit the number of locations that require CC1 construction standards, consequently avoiding the highest costs to construct or retrofit buildings.

Our results underscore the importance of incorporating climate-informed hazard projections into the WUI risk assessment process. As the WUI Guide is translated into codes and standards, reliance on current hazard maps may underestimate future risk in areas where hazard levels are expected to increase. Our results demonstrate that many locations across Canada currently classified as low or moderate hazard may shift to higher levels and therefore require more stringent construction classes to maintain the same level of protection as climate change progresses. Integrating future climate projections into the hazard classification process would ensure that buildings constructed today remain appropriately protected over their full lifespan.

4. Conclusions

This study examined how climate-driven changes in fire weather may influence wildfire hazard classification, relevant to climate-resilient building design. Current wildfire hazard assessments, including those referenced in the National Guide for Wildland–Urban Interface Fires, are relying on historical conditions and may not fully reflect changing wildfire hazard. Our results show that wildfire hazard is projected to increase across many locations in Canada. These shifts have direct implications for exposure classification and construction class (CC) requirements under the Guide. In many communities, maintaining the current levels of fire resilience may require upgrading to more stringent CCs. For example, under high exposure with priority zone mitigation in 1A and 1, the number of locations requiring CC1 rises from 142 to 159, reflecting a tangible shift in construction requirements that would trigger additional fire-resistant design measures such as non-combustible cladding, enhanced ventilation protection, and tempered glazing.

While the results offer useful insights, several key assumptions should be noted. The projected hazard estimates are based on a pseudo-global-warming approach intended to assess climate sensitivity in wildfire hazard classification, rather than to simulate future wildfire regimes or predict fire occurrence. Additionally, this method does not account for future changes in the WUI, vegetation, ignition likelihood, or suppression strategies, and it does not incorporate dynamic fire behavior besides those assumed in the baseline case. Future work could refine these estimates by incorporating dynamic fuel transitions, expanding geographic coverage, and coupling with socio-economic or land use change scenarios. Although dynamic fire growth models provide detailed representations of fire spread, they remain challenging to apply consistently at national scales. In contrast, the approach adopted here offers a scalable means of evaluating how climate-driven changes in fire weather may alter hazard, highlighting locations where climate sensitivity is likely to be most relevant for resilience planning and building design. The results should be interpreted as an assessment of climate sensitivity in wildfire hazard, rather than as forecasts of future wildfire occurrence, burn probability, or fire perimeters. The approach provides a practical way to examine how climate-driven increases in fire intensity may influence the application of wildfire-resilient building guidance over a typical building lifespan. By identifying locations where climate-adjusted hazard differs from current levels, the findings help inform discussions on the robustness of existing design assumptions and support climate-informed planning for codes, standards, and long-term resilience across Canada’s fire-prone regions.