Was Canada’s 2023 fire season a preview of things to come in North America?

Was Canada’s 2023 fire season a preview of things to come in North America? | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Physical Sciences - Article Was Canada’s 2023 fire season a preview of things to come in North America? Kaiwei Luo, Xianli Wang, Dante Castellanos-Acuna, Mike Flannigan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8779558/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Canada’s record-breaking 2023 wildfire season, followed by the severe 2024 and second-worst 2025 seasons, upended expectations of gradual boreal fire regime shifts. Together with recent Western U.S. extremes, these events raise a critical question: are they statistical outliers or the new reality in North America? Linking satellite observations with climate models to project future diurnal fire potential, we show that the climate-driven loss of diurnal firebreak can rapidly normalize the unprecedented flammability of Canada’s 2023 season by mid-century (2041–2070) even under the most ambitious mitigation scenario, while the Western U.S. 2020/2021 extremes approach typical conditions by late-century (2071-2100) but only under higher-warming scenarios. Mechanistically, we find a disproportionate escalation in overnight burning potential, concentrating fire opportunities into multi-day, round-the-clock runs when containment is least effective, and accelerating throughout the century across all warming scenarios. This shift is strongly amplified at high latitudes, with boreal overnight potential doubling to tripling by late-century under high warming, and it occurs synchronously across the entire boreal zone, challenging fire management strategies reliant on inter-regional resource sharing. The diurnally asymmetric, “locked-in” acceleration of northern fire risk requires urgent adaptation beyond current suppression capabilities. Earth and environmental sciences/Ecology/Fire ecology Earth and environmental sciences/Climate sciences/Climate change/Climate-change impacts/Environmental health Earth and environmental sciences/Climate sciences/Climate change/Projection and prediction Figures Figure 1 Figure 2 Figure 3 Figure 4 One-sentence summary The climate-driven loss of the diurnal firebreak will rapidly normalize the unprecedented flammability of Canada’s 2023 fire season by mid-century, regardless of mitigation efforts. Main Text Climate change is profoundly reshaping wildfire regimes worldwide, with boreal forests experiencing particularly remarkable transformations due to their unique vulnerability to fire-climate feedbacks 1-4 . Longstanding projections have warned of escalating fire activity in these systems under warming climates 5-7 , yet the relatively stable fire trends observed across much of Canada by the early 2000s 8-11 might have suggested that regime shifts unfold gradually over time. This expectation was upended by the record-breaking Canada's 2023 fire season, which burned approximately 15 million hectares, more than double the previous record, while releasing ~3 Gt CO₂e and forcing the evacuation of 232,000 people 12,13 . In the subsequent years of 2024 and 2025, Canada again experienced two consecutive severe fire seasons by burning over 5 million hectares, and the year of 2025 became the second-worst fire season on record. Similarly, the western U.S. endured back-to-back severe wildfire events in 2020 and 2021 14,15 . Together, these episodes raise an alarming question: are such fire seasons just statistical outliers that we can attribute to natural variability, or are they the early manifestations of a new normal for North America's fire future in a rapidly warming world? To answer this question, projecting future fire potential via linking fire activity to fire weather offers a direct, quantifiable pathway 6,16-18 . It is well-established that wildfire impacts are dominated by brief, critical windows with extreme fire weather 17,19,20 . Across North America, just ~3% of ignitions drive nearly all area burned 21 and structures lost 19 , often with a single day’s surge accounting for ~40% of a fire’s total extent 19 . These critical windows lie in the tail of the fire weather distribution; small shifts in their frequency or magnitude under climate shifts can therefore have outsized effects 17,22-24 . Yet, this sensitivity tends to be muted when projections rely on monthly or seasonal means 16,25,26 . Daily-scale projections capture more of this tail behavior, yet even they may obscure a critical dimension: the restructuring of fire opportunity within the diurnal cycle, as emerging evidence underscores these intraday surges, particularly overnight, as critical amplifiers, fueling multi-day runs while compressing operational timeframes 27,28 . At this diurnal resolution, we quantify two complementary metrics: active burning days (ABD), tallying days with viable burning hours, typically centered on daytime peaks, to gauge baseline daily potential; and overnight burning events (OBE), capturing extreme nights where fires persist through cooler, humid lulls that traditionally aid containment 29 . OBE, in particular, eliminates natural diurnal breaks, with 99% tied to fires >1,000 ha and their incidence correlating strongly with ultimate fire size and duration 29 . Together, these metrics—ABD for routine windows and OBE for their critical, intensified peaks—offer a holistic perspective on evolving diurnal fire cycles, yet their future trajectories across warming scenarios remain unknown. Here we quantify how North America's diurnal burning windows will evolve under climate change. Using 2017-2023 sub-hourly GOES-R satellite observations and high-resolution fire perimeters, we developed machine learning linking observed ABD and OBE to daily fire weather indices. Applying these relationships to four CMIP6 climate models under four warming scenarios spanning ambitious mitigation to very high warming (Shared Socioeconomic Pathway; SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5), we projected future ABD potential (ABDp) and OBE potential (OBEp) for mid-century (2041-2070) and late-century (2071-2100) relative to a 1991-2020 baseline (Methods). Our projections reveal changes in their magnitude, spatial pattern, pathway sensitivity, and timing of emergence above baseline variability across North America's diverse fire-prone biomes. Critically, we benchmark recent extreme fire years, the Canada 2023 and Western-US 2020-2021 fire seasons, against future climate distributions to assess whether today's catastrophic conditions represent tomorrow's routine fire seasons. Disproportionate increases in overnight burning under warming Across North America’s burnable landscapes (Supplementary Fig. S1), we project a pronounced lengthening of both ABDp and OBEp by late century. Under high warming scenario (SSP3-7.0), relative to the 1991-2020 baseline period, late-century annual ABDp increases by a mean of 15% continentally (≈ +14.5 d yr⁻¹; climate model range +6–32% or +6–30 d yr⁻¹; Fig. 1A, B; Supplementary Fig. S2, Table S1). The response for OBEp is even sharper, rising by 47% (≈ +14 d yr⁻¹; +30–80% or + 9–24 d yr −1 ; Fig. 1C, D; Supplementary Fig. S2, Table S1). Spatial coherence is strong: on average, 84% of grid cells show positive ABDp changes, rising to 96% for OBEp, underscoring a stronger and more widespread overnight response. These changes emerge early. By mid-century under SSP3-7.0, ABDp shows a modest mean increase (+5%, -2.6% to +15%), whereas OBEp already escalates by 30% (+19–49%; Supplementary Fig. S3). This indicates that nighttime constraints weaken earlier than the regular burning window. Responses increase with warming across scenarios for both time periods, SSP1-2.6 < SSP2-4.5 < SSP3-7.0 < SSP5-8.5, and this scenario sensitivity is steeper for OBEp than for ABDp (Fig. 1E, G; Supplementary Table S1). Notably, OBEp already shows clear mid-century increases even at the ambitious mitigation scenario, SSP1-2.6 (+21%, +14–30%), while ABDp remains comparatively stable (-0.7%, -5% to +5%; Fig 1E). The area of unanimous increase across climate models (i.e., all models agree in sign) also generally expands with warming; for example, for late-century, OBEp from 78% of the grids under SSP2-4.5 increases to 91% under SSP3-7.0, and ABDp from 53% to 66% (Fig. 1F). Across all climate models, time periods, and scenarios, OBEp increases outpace ABDp in both magnitude and spatial coherence (Fig. 1E-1H, Supplementary Table S1), highlighting a systematic weakening of the diurnal constraint and a rising nighttime fire pressure in a warming climate. Latitudinal amplification of burning-window shifts Despite continent-wide coherence, the magnitude and timing of changes vary across biomes. A pronounced north-south gradient of shifts emerges in both ABDp and OBEp across all climate models, periods, and scenarios. Boreal biomes anchor the upper end of this gradient. For ABDp, the five boreal biomes span widely different baseline regimes—from mountain systems averaging ~31 d yr⁻¹ to western coniferous forests reaching ~70 d yr⁻¹—yet all exhibit substantial mean increases of +30% to +53% (+48–93 d yr⁻¹) in late-century SSP3-7.0 projections (Fig. 2A). OBEp increases are larger, with increase means of +99-179%, no boreal biome shows <~2x increases. Individual climate models reveal substantial spread, for example, while the most conservative model still anticipates at least +124% OBEp increases, the most aggressive model projects near-quadrupling (+269%) in eastern tundra woodlands (Fig. 2A). Moving southward, percentage changes moderate but absolute budgets remain large. In temperate systems, ABDp changes are +5–14% yet amount to +7–16 d·yr⁻¹ on average at SSP3-7.0 by late-century; OBEp spans +3–119%, with steppe regions showing the highest relative gains (Fig. 2A). Subtropical regions show ABDp −3–+4% and OBEp +7–56%, while retaining large absolute burning windows. This latitudinal amplification already emerges by mid-century and persists across scenarios, with contrasts scaling with warming and arising primarily from larger boreal increases (Fig. 2, Supplementary Fig. S4). For example, mid-century boreal OBEp is projected to increase by 38–110% with ABDp by 9–27% even at SSP1-2.6 while temperate and subtropical systems show slight or minimal response. Recent extremes become future routine To contextualize these projections, we examine how recent extreme fire seasons would rank within future mid- and late-century distributions. This provides critical benchmarks for evaluating whether recent extremes represent the new normal—information essential for fire management agencies facing an uncertain future. Canada 2023 in late-century: from unprecedented to multiple-per-decade. Canada's 2023 fire season not only broke all modern records in terms of fire attributes e.g., burned area, emissions, and intensity, but also here in climatological flammability—72 ABDp and 18 OBEp d yr⁻¹ exceed every observation in the 1991-2020 baseline (100ᵗʰ percentile for both metrics; Fig. 3). Across individual boreal biomes where Canada 2023 largely occurred, 2023 is yet typically an upper-tail year rather than a wholesale departure from baseline variability (ABDp ~67–100th; OBEp ~90–100th, Supplementary Fig. S5). This national-scale extreme therefore reflects the unusual synchrony of upper-tail flammability across multiple boreal biomes. Late-century climate warming casts this benchmark as ordinary. By late century under SSP3-7.0, Canada’s 2023 ABDp and OBEp conditions drop to the 57th (range 0-97th; lower ranking = more frequent exceedance of this 2023 extreme) and 17th (0-53rd) percentiles, respectively (Fig. 3A, B). Under this high warming trajectory, 2023-level or higher conditions would occur routinely—multiple times per decade rather than once in the observational record. In the most aggressive model under SSP3-7.0, the late-century means reach 91 ABDp and 38 OBEp d yr⁻¹, placing 2023 below the modeled late-century distribution and rendering that "catastrophic" season comparatively mild. Across scenarios, the frequency of Canada-2023 conditions scales with warming; notably, under no scenario does 2023’s OBEp value exceed the late-century distribution median, indicating that amplified diurnal burning—especially sustained overnight burning—may become an unavoidable reality (Fig. 3B). Canada 2023 in mid-century: early normalization even under mitigation . By mid-century, Canada’s 2023 benchmark already shifts markedly within projected distributions, indicating a rapid normalization of extremes. Under SSP3-7.0, the mid-century distributions already reposition Canada-2023 toward frequent attainment (Fig. 3C, D): ABDp drops to the upper half of the distribution (multi-model means: 81st), and OBEp moves even further toward routine occurrence (37th). Strikingly, even under SSP1-2.6 by mid-century, Canada’s OBEp repositions to the median (50th) with ABDp remaining upper-tail (96th), further highlighting the urgency; SSP2-4.5 strengthens this signal and SSP5-8.5 accelerates it (Fig. 3D). Western US 2020/2021 in late-century: extremes edge toward the center. The western United States experienced consecutive severe fire seasons in 2020/2021 (~210 ABDp and ~125 OBEp d yr⁻¹), each above the 91st percentile in the baseline distribution (Fig. 4). Under SSP3-7.0, their late-century ranks shift toward the center for ABDp (62nd and 75th, respectively) and OBEp (94th and 78th), while multi-model means converge near 205 ABDp and 115 OBEp d yr⁻¹, comparable to the historical extremes themselves (Fig. 4A, B). Thus, even in a region already characterized by long burn seasons, 2020/2021-like conditions are expected to be more frequent by century’s end. Across SSPs, these rankings show modest variation, with SSP3-7.0 generally intermediate between SSP2-4.5 and SSP5-8.5. Western US 2020/2021 in mid-century: limited transitions. Mid-century projections show comparatively modest shifts for the western US (Fig. 4C, D). Across scenarios, the 2020/2021 benchmarks largely remain in the upper tail of the projected distributions for both ABDp and OBEp, indicating that seasons comparable to 2020/2021 are not yet representative of typical mid-century conditions. Lower-warming scenarios (SSP1-2.6/2-4.5) can slightly shorten burning windows in some models, whereas higher warming tends to push distributions upward. Discussion: North America’s fire future Our projections confirm the alarming question posed by recent catastrophic fire seasons: the events of Canada 2023 and the Western US 2020/2021 are not statistical outliers, but early manifestations of the new reality. We find this transformation is driven by three key mechanisms: (1) a disproportionate escalation of overnight burning compared to daytime potential, fundamentally a reflection of the weakening diurnal firebreak; (2) a pronounced latitudinal amplification, concentrating the greatest relative changes in high-latitude boreal ecosystems; and (3) responses scaling with warming level, revealing that while aggressive mitigation can prevent the worst outcomes, a substantial increase in fire risk is already 'locked in' for the coming decades. The intensification of critical overnight burning conditions emerges early and accelerates throughout the century. Across SSPs, time periods, and climate models, projected increases in OBEp consistently outpace those of ABDp. This divergence suggests that a growing share of season’s fire potential may concentrate into multi-day, round-the-clock runs when suppression is least effective 28-30 . Such temporal clustering, compounded by the early onset of overnight burning after ignition 29 , compresses the window for initial attack and increases the likelihood of escaped fires and subsequent extreme fire events. More broadly, this shift indicates an ongoing restructuring of diurnal fire cycles 31 . Critical burning days become more frequent, and daily burnable windows also expand alongside higher fire intensities 20 . This expansion is increasingly reinforced by weakening nighttime recovery (e.g., smaller humidity rebound and asymmetric diurnal warming 32,33 ) and climate-driven transition to flash, intensified and prolonged droughts 34 . The degradation of diurnal burning exhibits a pronounced latitudinal gradient, with the strongest shifts concentrated in boreal ecosystems. Although northern regions historically showed lagged fire responses despite faster warming 35-39 , our projections indicate that accumulated warming will reverse this pattern, with boreal OBEp potentially doubling or tripling by late-century under high warming. This is consistent with projections of increasing burned area, spread rates, and extreme fire years in northern forests 5,7,17,40-43 and with recent record-breaking polar and tundra fires 44,45 . In this context, Canada’s 2023 fire season represents a prototype of the emerging baseline that conditions outside the historical range shift toward routine mid-century occurrence, even under the most ambitious scenario. This "locked-in" nature signals an urgent need to prepare for a new era of extreme boreal fire activity, regardless of near-term mitigation success. Moreover, we found the unprecedented 2023 fire season arose from the simultaneous, near-maximum fire weather across the entire Canadian boreal shield. Projections further suggest that traditionally less fire-prone eastern boreal regions 24,38 may experience increases comparable to, or greater than, those in the west. Such spatial coherence challenges fire management strategies that rely on inter-regional resource sharing during peak seasons 9,30,46 , designed for localized hotspots rather than nation-wide synchronized extremes. In contrast to this rapid and synchronized boreal transition, the Western US exhibits a delayed shift toward typical conditions. While the 2020/2021 fire seasons remain high-percentile events through mid-century, this apparent stability reflects an already elevated baseline rather than lower risk 18,43,47 . The convergence of these recent extremes toward the mean occurs later in the century under higher warming scenarios, which suggests the Western US faces a sustained persistence of its current high-severity regime before further intensification. While our analysis quantifies the climatic envelope of fire potential, its translation into realized fire activity will be modulated by dynamic fuels, ignition regimes, and human intervention. Dynamic fuels can constrain and amplify the expression of longer potential burning windows through biomass availability, post-disturbance recovery, and vegetation transitions 11,48-50 . Ignition regimes may shift with changing convective environments and continued wildland-urban interface expansion 51,52 . Human actions, including suppression legacies, land use, infrastructure, and treatments, can likewise influence the translation of potential 53-56 . These factors add uncertainties but are unlikely to negate the strong forcing of the expanding diurnal fire niche; instead, they determine where and how rapidly projected potential is expressed. Future work should therefore couple diurnal fire-weather potential with dynamic fuel and ignition modules, and evaluate how management interventions alter the probability that projected overnight windows translate into large-fire growth. The convergence of evidence indicates that the fundamental transformation of North America's fire regime is not a distant possibility but an ongoing reality. Recent extreme seasons preview a future where overnight burning becomes commonplace, fire seasons extend substantially, and suppression resources designed for 20th-century fire regimes face mounting pressure. The choice between aggressive mitigation and high emissions will determine whether future fire regimes remain marginally manageable or face more substantial challenges. As Canada's 2023 fire season demonstrated, when boreal systems approach critical thresholds, the wildfires impact the whole planet through widespread smoke, carbon emissions, and economic disruption 12,13,57 . Our projections suggest such fire seasons will intensify and regularize in the coming decades, underscoring the urgent need for accelerated adaptation strategies, especially for northern regions. Methods Study Area and Biome Classification Our study area covers the continental United States and Canada, employing a detailed biome classification derived from ref. 58 and extending approaches from ref. 39 . The continental domain is divided into 20 distinct biomes, each defined by distinctive climatic and vegetative characteristics. The classification system includes five primary Boreal biomes (western and eastern coniferous forests, western and eastern tundra woodlands, plus mountain systems), six Temperate biomes (continental forest, oceanic forest, western and eastern mountain systems, steppe, and desert), and five Subtropical zones (humid forest, dry forest, mountain system, steppe, and desert). The classification is completed with tropical, polar, and water biomes for comprehensive continental representation. Based on satellite-derived observations of active burning days (ABD) and overnight burning events (OBE) from 2017-2023 across North America, we identified 11 biomes as burnable and fire-prone areas: Subtropical steppe, Subtropical desert, Subtropical mountain system, Temperate steppe, Temperate desert, Temperate mountain system west, Boreal coniferous forest east, Boreal coniferous forest west, Boreal tundra woodland east, Boreal tundra woodland west, and Boreal mountain system. For western U.S. 2020-2021 analysis, the area is specifically defined as areas overlapped between fire-prone biomes and Washington, Oregon, California, Nevada, Idaho, Montana, Wyoming, Utah, Colorado, Arizona, and New Mexico. Wildland Fire Geospatial Databases This study incorporates wildland fire geospatial data from three comprehensive databases that document fire perimeters and other important information including e.g., burned areas and ignition date throughout North America. We utilized the Canadian National Burned Area Composite (NBAC) 59 as our source for Canadian fire data, which serves as a key component within the Fire Monitoring, Accounting and Reporting System (FireMARS). Developed collaboratively by the Canada Centre for Mapping and Earth Observation and the Canadian Forest Service, NBAC delivers burned area mapping spanning 1972 to 2023 (version 20240530). The dataset maintains high spatial precision by integrating 30-meter Landsat imagery with higher-resolution agency imagery (<30-meter spatial resolution). U.S. fire information was obtained through the Monitoring Trends in Burn Severity (MTBS) 60 database as the primary source, with additional coverage from the Interagency Wildland Fire Perimeter History (IWFPH). MTBS represents a collaborative effort between the U.S. Geological Survey (EROS) and the USDA Forest Service (GTAC), documenting burn severity and extent throughout the United States from 1984 to 2022 (released August 22, 2024). The database applies regionally-specific inclusion criteria: fires ≥1,000 acres (~405 ha) in the western U.S. and ≥500 acres (~202 ha) in the eastern U.S. To achieve complete temporal coverage through 2023, we supplemented MTBS records with fire perimeter data from the IWFPH (version 20240825), which is maintained by the Wildland Fire Management Research, Development, & Application program data team and distributed through the National Interagency Fire center (NIFC). Geostationary Active Fire Detections Following ref. 29 , we used the Fire/Hot Spot Characterization Full Disk (FDCF) products from the GOES-R series (GOES-16/17/18) to obtain sub-hourly active-fire detections across North America for 2017–2023 61 . GOES-16 (GOES-East) has operated near 75.2°W since 2017. GOES-17 (GOES-West) observed near 137.2°W from late 2018 until early 2023, and GOES-18 has occupied ~137.2°W since late 2022, replacing GOES-17 as the West satellite in 2023. Accordingly, we used GOES-16 from its first available date through 2023, GOES-17 from its first available date through 2022, and GOES-18 for 2023. Together, GOES-East and GOES-West provide geostationary coverage of burnable land across North and South America. FDCF is derived from the Advanced Baseline Imager (ABI) visible and infrared bands and reports sub-pixel fire characteristics at a 5–15-min temporal resolution, with ~2 km nominal pixel size at the sub-satellite point and coarser effective resolution at larger view-zenith angles (VZA). The availability, sampling frequency, and overall data quality of GOES-R FDCF vary across space and time during 2017–2023 due to the staggered satellite deployments, changes in imaging frequency, differences in coverage and VZA between GOES-East and GOES-West, and occasional instrument outages. Given these inconsistencies, we used FDCF to infer binary hourly burning status (active vs. inactive) for fires with well-defined spatiotemporal boundaries in wildland fire databases, rather than relying on hotspot counts. As the available data provide at least four observations per hour at any location across the study area, we considered these data fit for this purpose. Spatial coverage differs between the East and West platforms. Northwestern North America (including Alaska and Yukon) is not visible to GOES-East; therefore, detections there are only available from August 2018 onward, when GOES-17 (and later GOES-18) became operational. In contrast, northeastern North America is only observed by GOES-East. In the central region where coverage overlapped, the data volume approximately doubled after GOES-17/18 was active. Imaging frequency changed during the study period as satellites transitioned among operational scan modes, producing full-disk imagery (and thus FDCF) at 15 min (Mode 3), 10 min (Mode 6), or 5 min (Mode 4). In the early mission phase (2017–2018), GOES-16 and GOES-17 were predominantly in Mode 3, whereas Mode 6 became the dominant configuration from 2019 onward for GOES-16/17 and subsequently for GOES-18. As a result, later years provide denser FDCF sampling. For more information on GOES-R scanning mode, please refer to https://www.goes-r.gov/. We pooled detections from GOES-16/17/18 to create a single continental-scale dataset, aiming to maximize the completeness of the geostationary record for each fire location rather than to compare satellites. Several technical considerations arise from the fixed geostationary viewing geometry. For any ground point, VZA is effectively constant for a given satellite, and larger VZA degrades detection performance in multiple ways 62,63 . In particular, ABI pixel size and omission errors generally increase with VZA, especially for small/cool fires and in high-latitude or complex-terrain settings. These limitations further motivate our focus on binary burning status (rather than hotspot counts). Geometric effects and terrain can also introduce parallax-related geolocation offsets; to reduce mismatches between hotspot locations and mapped fire perimeters, we applied a 2 km buffer (approximately one nominal ABI fire-pixel width) around each perimeter when associating detections with individual fires (see “Active Burning Days and Overnight Burning Events”). Finally, nighttime backgrounds are typically cooler and more spatially homogeneous, improving contrast for active-fire pixels; consequently, the nighttime detection algorithm is generally more sensitive than the daytime algorithm, particularly for smaller and/or cooler fires 64 . Active Burning Days and Overnight Burning Events We established a framework to assess hourly burning status (active or inactive) across fire events. The methodology commenced by extracting GOES-R active fire hotspots that spatially intersected with fire perimeters from our integrated database (NBAC and MTBS-IWFPH). To associate hotspots with events, we intersected GOES-R hotspots with perimeters buffered by 2 km (~nominal FDCF pixel width), which accommodates ABI geolocation error, parallax at oblique angles, and perimeter uncertainty. Fire event temporal boundaries were determined through a dual strategy: incorporating documented start and end dates where available, or deriving these dates from GOES-R hotspot observations based on initial and final instances of consecutive active burning hours. Fire events from 2017-2023 lacking corresponding hotspot data were excluded from analysis, recognizing that such omissions may stem from detection algorithm constraints under specific conditions 64 . For temporal precision, all fire activity data were transformed from Coordinated Universal Time (UTC) to local time zones, utilizing fire perimeter spatial centroids and accounting for day of year variations. We implemented exact daylight delineation using precise sunrise and sunset calculations, with sunrise defined as the moment the sun's upper edge becomes visible on the horizon and sunset as its disappearance below the horizon. This framework enabled categorization of each hour into four distinct states: active daytime, non-active daytime, active night-time, or non-active night-time. For analytical consistency, we defined a day as spanning from the first hour following sunrise to the final hour preceding the subsequent sunrise and defined each hour as a half-open interval [hh:00:00, hh:59:59)—daytime spans from the first hour after sunrise to the last hour before sunset and hours outside this range are nighttime. The event temporal window is therefore from the first hour of the start date to the last hour of the end date. Fire activity classification criteria were established whereby a fire was deemed active during any hour containing at least one detected hotspot within its perimeter. Non-active hours were those with no detected hotspots within the fire perimeter. We derived ABD as days containing at least one active hour; non-ABD were days with no active hours. For OBE, nights were classified when all night-time hours registered as active, indicating continuous burning throughout the entire night 29 . Non-OBE included nights with incomplete fire activity across night-time hours. OBE represents fire persistence during the diurnally least favorable burning conditions. We selected GOES-R geostationary satellite products as they provide the only consistent, high-frequency (≤15-minute) fire detection capability for North America, uniquely enabling capture of hourly fire patterns. However, the burning hours and thus ABD and OBE likely represent underestimates. This is because, within the abovementioned spatio-temporal window, the physical fire state is theoretically either flaming or smouldering at the most of time; however, when status needs to be determined from the geostationary perspective, “non-active” hour may reflect non-flaming/smoldering combustion below the detection threshold, very small flaming area, cloud/canopy obstruction, unfavorable viewing geometry, or genuine inactivity 64 . Fire Weather Calculation and Extraction Fire weather variables were derived based on ERA5 reanalysis 65 and the Canadian Forest Fire Weather Index System (CFWIS), a widely used operational framework for assessing fire danger that calculates six components, including three fuel moisture codes (Fine Fuel Moisture Code (FFMC), Duff Moisture Code (DMC), and Drought Code (DC)) and three fire behavior indices (Initial Spread Index (ISI), Buildup Index (BUI), and Fire Weather Index (FWI)) 66 . ERA5 provides global atmospheric fields on a 0.25° horizontal grid with hourly resolution. From ERA5 reanalysis for the period 1991-2023, we obtained 2-m temperature, 2-m dewpoint temperature, 10-m u-component of wind, 10-m v-component of wind, and 24-hour precipitation to derive the required inputs at noon local standard time for CFWIS calculations: temperature, relative humidity, wind speed, and accumulated precipitation. Daily CFWIS components were calculated using the cffdrs R package 67 following established overwintering procedures 68 . For each fire event occurring during 2017-2023, the extracted CFWIS variables were spatially averaged across all grid cells that intersected the fire perimeter. The extraction period spanned the entire lifetime of each fire, with a 1-day buffer added at the start to capture early fire conditions. Finally, these daily fire weather metrics were time-matched to the corresponding fire dynamics used in our analyses. Link ABD and OBE to Daily Fire Weather To link daily fire weather conditions with ABD and OBE occurrence, we developed machine learning models using the matched fire dynamics and fire weather data from 2017-2023. The feature set included all six CFWIS components (FFMC, DMC, DC, ISI, BUI, FWI) along with categorical variables for biome type and month. Prior to model training, numerical features were standardized and categorical variables were one-hot encoded. We used random forest because it is an ensemble model, and it is built on multiple decision tree classifiers and can capture complex, non-linear relationships between fire weather and fire activity with robust performance. We employed a hierarchical random forest modeling framework with two stages: (1) a binary classifier predicting whether a day was an ABD or non-ABD, and (2) a conditional model predicting OBE binary occurrence given ABD conditions. To address class imbalance in the training data, we applied random undersampling to create balanced datasets while preserving all positive instances. Model development utilized 5-fold cross-validation repeated three times, with hyperparameters including 500 trees and square root of features for node splitting. Model performance was evaluated using area under the ROC curve (AUC), F1-score, and precision-recall metrics. AUC summarizes the model’s ability to discriminate between positive and negative classes across all possible classification thresholds. The F1 score is the harmonic mean of precision and recall (sensitivity), where precision is the fraction of predicted positives that are true positives, and recall is the fraction of true positives that are correctly identified. We selected the classification threshold by maximizing F1 along the precision–recall curve, and report the resulting AUC and F1 for both the ABD and OBE models. The ABD prediction model achieved AUC = 0.814 and F1 = 0.750. The OBE prediction model demonstrated stronger performance with AUC = 0.908 and F1 = 0.841. These metrics indicate robust predictive capability for both fire activity measures. Future Fire Weather Projections To assess future changes in fire weather conditions, we developed daily fire weather projections for two future periods (mid-century 2041-2070 and late-century 2071-2100) using a delta change downscaling approach. Future climate projections 69 were obtained from four CMIP6 General Circulation Models (GCMs): CanESM, UKESM, EC-Earth, and GFDL, selected based on guidelines in ref. 70 and their demonstrated good performance in evaluation by ref. 71 . We analyzed four Shared Socioeconomic Pathways (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) from ambitious mitigation to very high warming scenarios. Specifically, with the raw outputs from the GCMs, we calculated monthly 30-year averages for the periods 1990-2020 (GCM baseline) and the two future periods. We then subtracted the monthly GCM baseline averages from the monthly GCM future averages to obtain the change in climate conditions between time periods. Temperature and wind speed were calculated as simple delta changes (e.g., Tfuture - Tbaseline), while precipitation and relative humidity were calculated as ratio changes (e.g., RHfuture - RHbaseline)/RHbaseline). Since the output of GCMs varies in spatial resolution, all of the anomalies were downscaled by bilinear interpolation to the resolution of the daily baseline weather from ERA5 (0.25°). Finally, we applied the monthly delta changes from each future period/GCM/SSP combination to the baseline 1991-2020 daily variables from ERA5 in the same way as the deltas were calculated: temperature and wind speed as simple additions, and precipitation and relative humidity as ratio changes. This process generated 32 sets of future daily meteorological variables (4 GCMs × 4 SSPs x 2 periods). Using these datasets, we then calculated the corresponding daily CFWIS components detailed in ‘Fire Weather Calculation and Extraction’ section, producing comprehensive daily fire weather index projections that serve as inputs for our ABDp and OBEp predictions. Notably, as the delta approach does not change the frequency of precipitation events, projected changes may be conservative 72 . Projecting Future ABDp and OBEp We applied the trained models to our 32 sets of future daily fire weather projections to generate gridded estimates of daily potential for active burning days (ABDp) and overnight burning events (OBEp). For each grid cell and day in the future scenarios, we: (1) extracted the corresponding fire weather indices, biome type, and month; (2) applied the same standardization and encoding transformations used during training; (3) generated probability estimates using the hierarchical model structure; and (4) classified each day based on whether its probability exceeded the optimal threshold, then aggregated daily probabilities to annual sums, representing the expected number of days per year with conditions conducive to active burning or overnight events. This approach produced spatially explicit projections of ABDp and OBEp at 0.25° resolution across North America for both mid-century (2041-2070) and late-century (2071-2100) periods under each GCM-SSP combination. The resulting maps provide quantitative assessments of how fire potential may shift across different regions and time periods under varying climate scenarios. Analysis of Projected Changes and Spatial Agreement To quantify continental-scale changes in fire potential, we calculated statistics across all grid cells within burnable and fire-prone biomes. For each GCM–SSP combination, we first aggregated daily ABDp and OBEp predictions to annual totals for each grid cell and for each year within a 30-year period (baseline: 1991–2020; mid-century: 2041–2070; late-century: 2071–2100). We then computed period means as the average of the 30 annual totals at each grid cell. Absolute changes were calculated as (future period mean − baseline period mean), and relative changes as ((future period mean − baseline period mean) / baseline period mean × 100%) for both ABDp and OBEp at each grid cell. Continental means were derived by averaging grid-cell period means across all valid grid cells, with model ranges reported as minimum to maximum values across the four GCMs. To assess spatial coherence of projections, we identified grid cells where all four models (or three out of four models) agreed on the direction of the period-mean change (positive or negative). Biome-Level Analysis and Latitudinal Patterns For biome-specific assessments, we aggregated grid-cell ABDp and OBEp annual totals by biome type for baseline and future periods. We computed biome-level statistics including 30-year biome mean changes and percentile shifts to quantify distribution transformations. The latitudinal gradient was examined through the inherent latitudinal zonation of biome types (boreal, temperate, and subtropical), allowing assessment of the latitudinal fire potential changes. Contextualization of Recent Extreme Fire Seasons To place recent extreme fire seasons in the context of future projections, we compared observed ABDp and OBEp values from specific events against the 30-year distributions of area-averaged annual values in both baseline and two future periods. For Canada's 2023 season and Western US 2020/2021 seasons, we extracted area-averaged ABDp and OBEp values using the same spatial domains and methodologies applied to the projections. We then calculated the percentile rank of these observed values within: (1) the 1991-2020 baseline distribution, and (2) the 2041-2070 and 2071-2100 projected distributions under each GCM-SSP combination. Distributions were constructed from the 30 annual area-averaged values within each period (i.e., one value per year), and percentile calculations used empirical cumulative distribution functions; ranges across the four GCMs were reported to show projection spread. This approach quantified how frequently future climate conditions would produce fire weather matching or exceeding recent extremes. For Canada 2023, we additionally repeated this contextualization within each boreal biome (where most burned area occurred), computing biome-specific area-averaged annual values and percentiles to capture how the national extreme reflects a sum of biome-level extremes. The area definition of the Western US can be found in the section "Study Area and Biome Classification". Declarations Data availability The datasets for conducting the analysis presented here are all publicly available. The NBAC, MTBS and IWFPH wildland fire datasets are respectively available from the Canadian Forest Service (https://cwfis.cfs.nrcan.gc.ca/datamart/metadata/nbac), https://www.mtbs.gov/ and the National Interagency Fire Center (https://data-nifc.opendata.arcgis.com/datasets/nifc::interagencyfireperimeterhistory-all-years-view/about). The GOES-16, GOES-17, and GOES-18 full disk active fire products are available on Amazon Web Service S3 Explorer (https://registry.opendata.aws/noaa-goes/). The hourly ERA5 climate data are available at https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview. The CMIP6 climate projections are available at https://cds.climate.copernicus.eu/datasets/projections-cmip6?tab=overview. The biome categorizations used in this study are available at https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world. Code availability Codes used to analyse the data are available from https://github.com/KaiweiLL/Canada-2023-diurnal-projections. Author contributions Conceptualization: M.F., X.W. and K.L. Methodology: K.L., X.W., and M.F. Data processing: K.L. and D.C. Investigation: K.L., X.W., and M.F. Visualization: K.L. Funding acquisition: K.L., X.W., and M.F. Project administration: M.F. Supervision: M.F. and X.W. Writing – original draft: K.L. Writing – review & editing: K.L., X.W., and M.F. Competing interests The authors declare that they have no competing interests. References Flannigan, M., Stocks, B., Turetsky, M. & Wotton, M. Impacts of climate change on fire activity and fire management in the circumboreal forest. Global change biology 15 , 549–560 (2009). Rogers, B. M., Soja, A. J., Goulden, M. L. & Randerson, J. T. Influence of tree species on continental differences in boreal fires and climate feedbacks. Nature Geoscience 8 , 228–234 (2015). Wotton, B. M., Flannigan, M. D. & Marshall, G. A. Potential climate change impacts on fire intensity and key wildfire suppression thresholds in Canada. Environ Res Lett 12 , 095003 (2017). Flannigan, M. D., Krawchuk, M. A., de Groot, W. J., Wotton, B. M. & Gowman, L. M. Implications of changing climate for global wildland fire. International journal of wildland fire 18 , 483–507 (2009). Flannigan, M. D., Logan, K. A., Amiro, B. D., Skinner, W. R. & Stocks, B. J. Future area burned in Canada. Climatic change 72 , 1–16 (2005). Wang, X., Swystun, T. & Flannigan, M. D. Future wildfire extent and frequency determined by the longest fire-conducive weather spell. Science of the total environment 830 , 154752 (2022). Wang, X. et al. Projected changes in fire size from daily spread potential in Canada over the 21st century. Environ Res Lett 15 , 104048 (2020). Coops, N. C., Hermosilla, T., Wulder, M. A., White, J. C. & Bolton, D. K. A thirty year, fine-scale, characterization of area burned in Canadian forests shows evidence of regionally increasing trends in the last decade. Plos one 13 , e0197218 (2018). Scholten, R. C., Veraverbeke, S., Chen, Y. & Randerson, J. T. Spatial variability in Arctic–boreal fire regimes influenced by environmental and human factors. Nature Geoscience 17 , 866–873 (2024). Amiro, B., Stocks, B., Alexander, M., Flannigan, M. & Wotton, B. Fire, climate change, carbon and fuel management in the Canadian boreal forest. International Journal of Wildland Fire 10 , 405–413 (2001). Hanes, C. C. et al. Fire-regime changes in Canada over the last half century. Canadian Journal of Forest Research 49 , 256–269 (2019). Jain, P. et al. Drivers and impacts of the record-breaking 2023 wildfire season in Canada. Nature Communications 15 , 6764 (2024). Byrne, B. et al. Carbon emissions from the 2023 Canadian wildfires. Nature 633 , 835–839 (2024). Safford, H. D., Paulson, A. K., Steel, Z. L., Young, D. J. & Wayman, R. B. The 2020 California fire season: A year like no other, a return to the past or a harbinger of the future? Global Ecology and Biogeography 31 , 2005–2025 (2022). Jain, P., Sharma, A. R., Acuna, D. C., Abatzoglou, J. T. & Flannigan, M. Record-breaking fire weather in North America in 2021 was initiated by the Pacific northwest heat dome. Communications Earth & Environment 5 , 202 (2024). Quilcaille, Y., Batibeniz, F., Ribeiro, A. F., Padrón, R. S. & Seneviratne, S. I. Fire weather index data under historical and SSP projections in CMIP6 from 1850 to 2100. Earth System Science Data Discussions 2022 , 1–31 (2022). Wang, X. et al. Critical fire weather conditions during active fire spread days in Canada. Science of the total environment 869 , 161831 (2023). Abatzoglou, J. T. & Williams, A. P. Impact of anthropogenic climate change on wildfire across western US forests. Proceedings of the National Academy of sciences 113 , 11770–11775 (2016). Balch, J. K. et al. The fastest-growing and most destructive fires in the US (2001 to 2020). Science 386 , 425–431 (2024). Kaiwei Luo, X. W., Dante Castellanos-Acuna, Mike Flannigan. A weakened diurnal weather cycle leads to longer burning hours in North America (Under Review). (2025). Stocks, B. et al. Large forest fires in Canada, 1959–1997. Journal of Geophysical Research: Atmospheres 107 , FFR 5–1–FFR 5–12 (2002). Jolly, W. M. et al. Climate-induced variations in global wildfire danger from 1979 to 2013. Nature communications 6 , 7537 (2015). Cunningham, C. X., Williamson, G. J. & Bowman, D. M. Increasing frequency and intensity of the most extreme wildfires on Earth. Nature ecology & evolution 8 , 1420–1425 (2024). Wang, X. et al. Increasing frequency of extreme fire weather in Canada with climate change. Climatic Change 130 , 573–586 (2015). Bedia, J. et al. Global patterns in the sensitivity of burned area to fire-weather: Implications for climate change. Agr Forest Meteorol 214 , 369–379 (2015). Matteo, A. et al. Challenges in assessing Fire Weather changes in a warming climate. arXiv preprint arXiv:2503.01818 (2025). Freeborn, P. H., Jolly, W. M., Cochrane, M. A. & Roberts, G. Large wildfire driven increases in nighttime fire activity observed across CONUS from 2003–2020. Remote Sens Environ 268 , 112777 (2022). Balch, J. K. et al. Warming weakens the night-time barrier to global fire. Nature 602 , 442–448 (2022). Luo, K., Wang, X., de Jong, M. & Flannigan, M. Drought triggers and sustains overnight fires in North America. Nature 627 , 321–327 (2024). Tymstra, C., Stocks, B. J., Cai, X. & Flannigan, M. D. Wildfire management in Canada: Review, challenges and opportunities. Progress in Disaster Science 5 , 100045 (2020). Kreye, J. K. et al. Fire behavior in masticated fuels: A review. Forest Ecology and Management 314 , 193–207 (2014). Davy, R., Esau, I., Chernokulsky, A., Outten, S. & Zilitinkevich, S. Diurnal asymmetry to the observed global warming. International Journal of Climatology 37 , 79–93 (2017). Chiodi, A. M., Potter, B. E. & Larkin, N. K. Multi‐decadal change in western US nighttime vapor pressure deficit. Geophys Res Lett 48 , e2021GL092830 (2021). Yuan, X. et al. A global transition to flash droughts under climate change. Science 380 , 187–191 (2023). Bush, E. & Lemmen, D. S. Canada's changing climate report. (No Title) (2019). Serreze, M. C. & Barry, R. G. Processes and impacts of Arctic amplification: A research synthesis. Global Planet Change 77 , 85–96 (2011). Pithan, F. & Mauritsen, T. Arctic amplification dominated by temperature feedbacks in contemporary climate models. Nature geoscience 7 , 181–184 (2014). Bergeron, Y., Gauthier, S., Kafka, V., Lefort, P. & Lesieur, D. Natural fire frequency for the eastern Canadian boreal forest: consequences for sustainable forestry. Canadian journal of forest research 31 , 384–391 (2001). Jain, P., Castellanos-Acuna, D., Coogan, S. C., Abatzoglou, J. T. & Flannigan, M. D. Observed increases in extreme fire weather driven by atmospheric humidity and temperature. Nature Climate Change 12 , 63–70 (2022). Zhao, H. et al. Future enhanced threshold effects of wildfire drivers could increase burned areas in northern mid-and high latitudes. Communications Earth & Environment 6 , 224 (2025). Flannigan, M. D., Stocks, B. J. & Wotton, B. M. Climate change and forest fires. Science of the total environment 262 , 221–229 (2000). de Groot, W. J., Flannigan, M. D. & Cantin, A. S. Climate change impacts on future boreal fire regimes. Forest Ecology and Management 294 , 35–44 (2013). Abatzoglou, J. T. et al. Climate change has increased the odds of extreme regional forest fire years globally. Nature Communications 16 , 6390 (2025). Hu, F. S. et al. Arctic tundra fires: natural variability and responses to climate change. Frontiers in Ecology and the Environment 13 , 369–377 (2015). Descals, A. et al. Unprecedented fire activity above the Arctic Circle linked to rising temperatures. Science 378 , 532–537 (2022). Erni, S., Arseneault, D., Parisien, M. A. & Bégin, Y. Spatial and temporal dimensions of fire activity in the fire‐prone eastern Canadian taiga. Global Change Biology 23 , 1152–1166 (2017). Abatzoglou, J. T. et al. Projected increases in western US forest fire despite growing fuel constraints. Communications Earth & Environment 2 , 227 (2021). Andela, N. et al. A human-driven decline in global burned area. Science 356 , 1356–1362 (2017). Jones, M. W. et al. Global and regional trends and drivers of fire under climate change. Reviews of Geophysics 60 , e2020RG000726 (2022). Baltzer, J. L. et al. Increasing fire and the decline of fire adapted black spruce in the boreal forest. Proceedings of the National Academy of Sciences 118 , e2024872118 (2021). Romps, D. M., Seeley, J. T., Vollaro, D. & Molinari, J. Projected increase in lightning strikes in the United States due to global warming. Science 346 , 851–854 (2014). Radeloff, V. C. et al. Rapid growth of the US wildland-urban interface raises wildfire risk. Proceedings of the National Academy of Sciences 115 , 3314–3319 (2018). Parisien, M.-A. et al. Fire deficit increases wildfire risk for many communities in the Canadian boreal forest. Nature communications 11 , 2121 (2020). Prichard, S. J. et al. Adapting western North American forests to climate change and wildfires: 10 common questions. Ecological applications 31 , e02433 (2021). Kreider, M. R. et al. Fire suppression makes wildfires more severe and accentuates impacts of climate change and fuel accumulation. Nature communications 15 , 2412 (2024). Fernandes, P. M. & Botelho, H. S. A review of prescribed burning effectiveness in fire hazard reduction. International Journal of wildland fire 12 , 117–128 (2003). Chen, H., Zhang, W. & Sheng, L. Canadian record-breaking wildfires in 2023 and their impact on US air quality. Atmospheric Environment 342 , 120941 (2025). Olson, D. M. et al. Terrestrial Ecoregions of the World: A New Map of Life on Earth: A new global map of terrestrial ecoregions provides an innovative tool for conserving biodiversity. BioScience 51 , 933–938 (2001). Hall, R. et al. Generating annual estimates of forest fire disturbance in Canada: the National Burned Area Composite. International journal of wildland fire 29 , 878–891 (2020). Eidenshink, J. et al. A project for monitoring trends in burn severity. Fire Ecol 3 , 3–21 (2007). Schmit, T. J. et al. A closer look at the ABI on the GOES-R series. Bulletin of the American Meteorological Society 98 , 681–698 (2017). Hall, J. V., Zhang, R., Schroeder, W., Huang, C. & Giglio, L. Validation of GOES-16 ABI and MSG SEVIRI active fire products. Int J Appl Earth Obs 83 , 101928 (2019). Wooster, M. et al. Meteosat SEVIRI Fire Radiative Power (FRP) products from the Land Surface Analysis Satellite Applications Facility (LSA SAF)--Part 1: Algorithms, product contents and analysis. Atmospheric Chemistry & Physics Discussions 15 (2015). Wooster, M. J. et al. Satellite remote sensing of active fires: History and current status, applications and future requirements. Remote Sens Environ 267 , 112694 (2021). Hersbach, H. et al. The ERA5 global reanalysis. Quarterly Journal of the Royal Meteorological Society 146 , 1999–2049 (2020). Wagner, C. V. Development and structure of the Canadian forest fire weather index system . (1987). Wang, X. et al. cffdrs: an R package for the Canadian forest fire danger rating system. Ecological Processes 6 , 1–11 (2017). McElhinny, M., Beckers, J. F., Hanes, C., Flannigan, M. & Jain, P. A high-resolution reanalysis of global fire weather from 1979 to 2018–overwintering the Drought Code. Earth System Science Data 12 , 1823–1833 (2020). Copernicus Climate Change Service, C. D. S., (2021): CMIP6 climate projections. Copernicus Climate Change Service (C3S) Climate Data Store (CDS). DOI: 10.24381/cds.c866074c. Mahony, C. R., Wang, T., Hamann, A. & Cannon, A. J. A global climate model ensemble for downscaled monthly climate normals over North America. International Journal of Climatology 42 , 5871–5891 (2022). Jeong, D. I. & Cannon, A. J. An approach for selecting observationally-constrained global climate model ensembles for regional climate impacts and adaptation studies in Canada. Atmosphere-Ocean 61 , 335–351 (2023). Clarke, H. et al. Gazing into the flames: A guide to assessing the impacts of climate change on landscape fire. Science Advances 11 , eadz2429 (2025). Additional Declarations There is NO Competing Interest. Supplementary Files NaturesupplementarymaterialsfuturenormalizationofNorthAmericanfireextremes.docx Supplementary materials Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8779558","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Physical Sciences - Article","associatedPublications":[],"authors":[{"id":590955724,"identity":"b294d25e-c41d-4c6e-9a5d-218a226d835f","order_by":0,"name":"Kaiwei Luo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2ElEQVRIie3SMQrCMBTG8VccXKJdE8Se4YmgS+lZEgp1dRUHA0K7KK6d9Ap6AyWDi9hV6GLxAropdLC6Ck27OeS3//leIACG8YcoWBIAXYAmqZcEAI1PwislX6pGwiIV0vs4Eat563CbPD2wo3150iEiZDGmIlZtv3fiPtCTZsoBEaakSKQiAyZ5A1B3nWNnYZrjWWwUGb4knwHaV81htFgB3IttsWJJrgCpZoXFWfRaoN/fFW9hMjgSetGs0GR0w2fuddfH5eEh3aljrzQrP6r+AcMwDKPMG9BnQolIAGMVAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0009-0001-1921-5429","institution":"University of Alberta; Northern Forestry Centre, Canadian Forest Service, Natural Resources Canada","correspondingAuthor":true,"prefix":"","firstName":"Kaiwei","middleName":"","lastName":"Luo","suffix":""},{"id":590955725,"identity":"69f861f6-871f-4eab-8ace-1e3932bf29fe","order_by":1,"name":"Xianli Wang","email":"","orcid":"https://orcid.org/0000-0001-9873-9767","institution":"Northern Forestry Centre, Canadian Forest Service, Natural Resources Canada","correspondingAuthor":false,"prefix":"","firstName":"Xianli","middleName":"","lastName":"Wang","suffix":""},{"id":590955726,"identity":"bd2a0441-8759-41c0-af95-fcf11e3a0d24","order_by":2,"name":"Dante Castellanos-Acuna","email":"","orcid":"","institution":"Northern Forestry Centre, Canadian Forest Service, Natural Resources Canada; University of Alberta","correspondingAuthor":false,"prefix":"","firstName":"Dante","middleName":"","lastName":"Castellanos-Acuna","suffix":""},{"id":590955727,"identity":"30fcbf3d-5c80-4788-92a5-8d3101cf5aa2","order_by":3,"name":"Mike Flannigan","email":"","orcid":"","institution":"Thompson Rivers University","correspondingAuthor":false,"prefix":"","firstName":"Mike","middleName":"","lastName":"Flannigan","suffix":""}],"badges":[],"createdAt":"2026-02-03 19:50:44","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8779558/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8779558/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":103507237,"identity":"dfce31ff-e4d2-4ea4-a5e6-298a822d5b3e","added_by":"auto","created_at":"2026-02-26 13:40:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":696937,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eProjected changes and spatial consistency of ABDp and OBEp in North America.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eA\u003c/strong\u003e–\u003cstrong\u003eD\u003c/strong\u003e) Maps of late-century (2071–2100) multi-model mean absolute (days year⁻¹) and relative (%) changes in annual potential active burning days (ABDp, \u003cstrong\u003eA\u003c/strong\u003e, \u003cstrong\u003eB\u003c/strong\u003e) as well as the potential overnight burning events (OBEp, \u003cstrong\u003eC\u003c/strong\u003e, \u003cstrong\u003eD\u003c/strong\u003e) relative to the 1991–2020 baseline under high warming scenario SSP3-7.0 in North America. The multi-model mean is calculated across four climate models (CanESM, EC-Earth, GFDL and UKESM; see Methods for model selection). Red indicates increases, blue indicates decreases. White areas represent areas with insufficient data. Maps for individual models and scenarios are provided in Supplementary Figs. S2–S3.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(\u003cstrong\u003eE\u003c/strong\u003e, \u003cstrong\u003eG\u003c/strong\u003e) Multi-model mean relative changes in ABDp and OBEp across four SSPs (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5) for late-century (\u003cstrong\u003eE\u003c/strong\u003e) and mid-century (2041–2070,\u003cstrong\u003e G\u003c/strong\u003e) in North America. Vertical lines show the inter-model range (min–max) among the four GCMs. Details in each model can be found in Supplementary Table S1. (\u003cstrong\u003eF\u003c/strong\u003e, \u003cstrong\u003eH\u003c/strong\u003e) Spatial consistency of projected changes for late-century (\u003cstrong\u003eF\u003c/strong\u003e) and mid-century (\u003cstrong\u003eH\u003c/strong\u003e), quantified as the fraction of grid cells where at least three quarters of the models (solid lines) or all models (dashed lines) agree on the sign of change.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8779558/v1/931a17171372329262fdfb5b.png"},{"id":103431598,"identity":"7417c7cf-4aae-432f-a8ed-86b1a6c68c86","added_by":"auto","created_at":"2026-02-25 15:37:00","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":342954,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eBiome-specific ABDp and OBEp projections under high warming scenario for mid- and late-century.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eHeatmaps show projected changes across biomes in annual potential active burning days (ABDp) and overnight burning events (OBEp) under SSP3-7.0 relative to the baseline period (1991–2020). Panel \u003cstrong\u003eA\u003c/strong\u003e presents late-century (2071–2100) projections, and panel \u003cstrong\u003eB\u003c/strong\u003e shows mid-century (2041–2070) projections. Each cell represents output from an individual climate model (CanESM, ECEarth, GFDL, UKESM; see Methods for model selection), with relative changes (%) and absolute changes (days yr⁻¹, in parentheses). Colors indicate the magnitude of relative change from decreases (blue) to increases (red). A strong latitudinal gradient emerged in ABDp and OBEp intensification, with boreal regions exhibiting the largest increases, temperate biomes showing moderate increases, and subtropical regions displaying more mixed responses, and with changes generally amplifying from mid- to late-century. Results of other scenarios can be found in Supplementary Fig. S4.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8779558/v1/e327c458f63472c1175cb69d.png"},{"id":103431601,"identity":"f92e8fb6-734d-4617-aa9f-5c9259e56039","added_by":"auto","created_at":"2026-02-25 15:37:01","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":307608,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eCanada 2023: from record-breaking to routine exceedance in future ABDp and OBEp.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e–\u003cstrong\u003eD\u003c/strong\u003e | Canada 2023 vs. historical baseline and future projections of annual potential active burning days (ABDp) and potential overnight burning events (OBEp). Gray violins show the 1991–2020 baseline distribution; colored violins show pooled model–year distributions for each SSP, combining annual values from four climate models (CanESM, ECEarth, GFDL, UKESM) for late-century (2071–2100; \u003cstrong\u003eA\u003c/strong\u003e–\u003cstrong\u003eB\u003c/strong\u003e) and mid-century (2041–2070; \u003cstrong\u003eC\u003c/strong\u003e–\u003cstrong\u003eD\u003c/strong\u003e) under SSP1-2.6/2-4.5/3-7.0/5-8.5. Black markers denote climate model means, and smaller gray markers indicate the responding individual annual samples (model–year values).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8779558/v1/86b8c6484d7be7555f8e53d3.png"},{"id":104397770,"identity":"b9683478-2c2e-46fd-a1a7-cd40feaf60c8","added_by":"auto","created_at":"2026-03-11 11:56:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":353165,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eWestern US 2020/2021 relative to historical baseline and future projections of ABDp and OBEp.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e–\u003cstrong\u003eD\u003c/strong\u003e| Western US 2020/2021 vs. historical baseline and future projections of annual potential active burning days (ABDp) and potential overnight burning events (OBEp). Gray violins show the 1991–2020 baseline distribution; colored violins show pooled model–year distributions for each SSP, combining annual values from four climate models (CanESM, ECEarth, GFDL, UKESM) for late-century (2071–2100; \u003cstrong\u003eA\u003c/strong\u003e–\u003cstrong\u003eB\u003c/strong\u003e) and mid-century (2041–2070; \u003cstrong\u003eC\u003c/strong\u003e–\u003cstrong\u003eD\u003c/strong\u003e) under SSP1-2.6/2-4.5/3-7.0/5-8.5. Black markers denote climate model means, and smaller gray markers indicate the responding individual annual samples (model–year values). Numbers below each scenario report the percentile ranks of the 2020 and 2021 benchmarks within the projected distribution (mean; min–max across models; top line = 2020, bottom line = 2021); lower percentiles indicate more frequent exceedance (i.e., 2020/2021-like or worse conditions become more routine).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8779558/v1/36bc34e89f1d29037314a5cf.png"},{"id":104407382,"identity":"a81e261d-e48c-45b5-8506-0b78993f7ba5","added_by":"auto","created_at":"2026-03-11 12:37:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2581258,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8779558/v1/bbe669b8-101c-402c-8e7c-f422ab8fa309.pdf"},{"id":103431600,"identity":"969a8ac4-0bef-4358-9445-14518f1df848","added_by":"auto","created_at":"2026-02-25 15:37:00","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4554153,"visible":true,"origin":"","legend":"Supplementary materials","description":"","filename":"NaturesupplementarymaterialsfuturenormalizationofNorthAmericanfireextremes.docx","url":"https://assets-eu.researchsquare.com/files/rs-8779558/v1/fc32f313d4e786dddfd4d0de.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Was Canada’s 2023 fire season a preview of things to come in North America?","fulltext":[{"header":"One-sentence summary","content":"\u003cp\u003eThe climate-driven loss of the diurnal firebreak will rapidly normalize the unprecedented flammability of Canada\u0026rsquo;s 2023 fire season by mid-century, regardless of mitigation efforts.\u003c/p\u003e"},{"header":"Main Text","content":"\u003cp\u003eClimate change is\u0026nbsp;profoundly reshaping wildfire regimes worldwide, with boreal forests experiencing particularly remarkable transformations due to their unique vulnerability to fire-climate feedbacks\u003csup\u003e1-4\u003c/sup\u003e. Longstanding projections have warned of escalating fire activity in these systems under warming climates\u003csup\u003e5-7\u003c/sup\u003e, yet the relatively stable fire trends observed across much of Canada by the early 2000s\u003csup\u003e8-11\u003c/sup\u003e might have suggested that regime shifts unfold gradually over time. This expectation was upended by the record-breaking Canada's 2023 fire season, which burned approximately 15 million hectares, more than double the previous record, while releasing ~3 Gt CO₂e and forcing the evacuation of 232,000 people\u003csup\u003e12,13\u003c/sup\u003e. In the subsequent years of 2024 and 2025, Canada again experienced two consecutive severe fire seasons by burning over 5 million hectares, and the year of 2025 became the second-worst fire season on record. Similarly, the western U.S. endured back-to-back severe wildfire events in 2020 and 2021\u003csup\u003e14,15\u003c/sup\u003e. Together, these episodes raise an alarming question: are such fire seasons just statistical outliers that we can attribute to natural variability, or are they the early manifestations of a new normal for North America's fire future in a rapidly warming world?\u003c/p\u003e\n\u003cp\u003eTo answer this question, projecting future fire potential via linking fire activity to fire weather offers a direct, quantifiable pathway\u003csup\u003e6,16-18\u003c/sup\u003e. It is well-established that wildfire impacts are dominated by brief, critical windows with extreme fire weather\u003csup\u003e17,19,20\u003c/sup\u003e. Across North America, just ~3% of ignitions drive nearly all area burned\u003csup\u003e21\u003c/sup\u003e and structures lost\u003csup\u003e19\u003c/sup\u003e, often with a single day’s surge accounting for ~40% of a fire’s total extent\u003csup\u003e19\u003c/sup\u003e. These critical windows lie in the tail of the fire weather distribution; small shifts in their frequency or magnitude under climate shifts can therefore have outsized effects\u003csup\u003e17,22-24\u003c/sup\u003e. Yet, this sensitivity tends to be muted when projections rely on monthly or seasonal means\u003csup\u003e16,25,26\u003c/sup\u003e. Daily-scale projections capture more of this tail behavior, yet even they may obscure a critical dimension: the restructuring of fire opportunity within the diurnal cycle, as emerging evidence underscores these intraday surges, particularly overnight, as critical amplifiers, fueling multi-day runs while compressing operational timeframes\u003csup\u003e27,28\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eAt this diurnal resolution, we quantify two complementary metrics: active burning days (ABD), tallying days with viable burning hours, typically centered on daytime peaks, to gauge baseline daily potential; and overnight burning events (OBE), capturing extreme nights where fires persist through cooler, humid lulls that traditionally aid containment\u003csup\u003e29\u003c/sup\u003e. OBE, in particular, eliminates natural diurnal breaks, with 99% tied to fires \u0026gt;1,000 ha and their incidence correlating strongly with ultimate fire size and duration\u003csup\u003e29\u003c/sup\u003e. Together, these metrics—ABD for routine windows and OBE for their critical, intensified peaks—offer a holistic perspective on evolving diurnal fire cycles, yet their future trajectories across warming scenarios remain unknown.\u003c/p\u003e\n\u003cp\u003eHere we quantify how North America's diurnal burning windows will evolve under climate change. Using 2017-2023 sub-hourly GOES-R satellite observations and high-resolution fire perimeters, we developed machine learning linking observed ABD and OBE to daily fire weather indices. Applying these relationships to four CMIP6 climate models under four warming scenarios spanning ambitious mitigation to very high warming (Shared Socioeconomic Pathway; SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5), we projected future ABD potential (ABDp) and OBE potential (OBEp) for mid-century (2041-2070) and late-century (2071-2100) relative to a 1991-2020 baseline (Methods). Our projections reveal changes in their magnitude, spatial pattern, pathway sensitivity, and timing of emergence above baseline variability across North America's diverse fire-prone biomes. Critically, we benchmark recent extreme fire years, the Canada 2023 and Western-US 2020-2021 fire seasons, against future climate distributions to assess whether today's catastrophic conditions represent tomorrow's routine fire seasons.\u003c/p\u003e\n\u003ch1\u003eDisproportionate increases in overnight burning under warming\u003c/h1\u003e\n\u003cp\u003eAcross North America’s burnable landscapes (Supplementary Fig. S1), we project a pronounced lengthening of both ABDp and OBEp by late century. Under high warming scenario (SSP3-7.0), relative to the 1991-2020 baseline period, late-century annual ABDp increases by a mean of 15% continentally (≈ +14.5 d yr⁻¹; climate model range +6–32% or +6–30 d yr⁻¹;\u0026nbsp;Fig. 1A, B;\u0026nbsp;Supplementary\u0026nbsp;Fig. S2, Table S1). The response for OBEp is even sharper, rising by 47%\u0026nbsp;(≈ +14 d yr⁻¹;\u0026nbsp;+30–80% or + 9–24 d yr\u003csup\u003e−1\u003c/sup\u003e;\u0026nbsp;Fig. 1C, D; Supplementary\u0026nbsp;Fig. S2, Table S1).\u0026nbsp;Spatial coherence is strong: on average, 84% of grid cells show positive ABDp changes, rising to 96% for OBEp, underscoring a stronger and more widespread overnight response.\u003c/p\u003e\n\u003cp\u003eThese changes emerge early. By mid-century under SSP3-7.0, ABDp shows a modest mean increase (+5%, -2.6% to +15%), whereas OBEp already escalates by 30% (+19–49%; Supplementary Fig. S3). This indicates that nighttime constraints weaken earlier than the regular burning window.\u003c/p\u003e\n\u003cp\u003eResponses increase with warming across scenarios for both time periods, SSP1-2.6 \u0026lt; SSP2-4.5 \u0026lt; SSP3-7.0 \u0026lt; SSP5-8.5, and this scenario sensitivity is steeper for OBEp than for ABDp (Fig. 1E, G; Supplementary Table S1). Notably, OBEp already shows clear mid-century increases even at the ambitious mitigation scenario, SSP1-2.6 (+21%, +14–30%), while ABDp remains comparatively stable (-0.7%, -5% to +5%; Fig 1E). The area of unanimous increase across climate models (i.e., all models agree in sign) also generally expands with warming; for example, for late-century, OBEp from 78% of the grids under SSP2-4.5 increases to 91% under SSP3-7.0, and ABDp from 53% to 66% (Fig. 1F).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAcross all climate models, time periods, and scenarios, OBEp increases outpace ABDp in both magnitude and spatial coherence (Fig. 1E-1H, Supplementary Table S1), highlighting a systematic weakening of the diurnal constraint and a rising nighttime fire pressure in a warming climate.\u003c/p\u003e\u003ch1\u003eLatitudinal amplification of burning-window shifts\u003c/h1\u003e\n\u003cp\u003eDespite continent-wide coherence, the magnitude and timing of changes vary across biomes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA pronounced north-south gradient of shifts emerges in both ABDp and OBEp across all climate models, periods, and scenarios. Boreal biomes anchor the upper end of this gradient. For ABDp, the five boreal biomes span widely different baseline regimes\u0026mdash;from mountain systems averaging ~31 d yr⁻\u0026sup1; to western coniferous forests reaching ~70 d yr⁻\u0026sup1;\u0026mdash;yet all exhibit substantial mean increases of +30% to +53% (+48\u0026ndash;93 d yr⁻\u0026sup1;) in late-century SSP3-7.0 projections (Fig. 2A). OBEp increases are larger, with increase means of +99-179%, no boreal biome shows \u0026lt;~2x increases. Individual climate models reveal substantial spread, for example, while the most conservative model still anticipates at least +124% OBEp increases, the most aggressive model projects near-quadrupling (+269%) in eastern tundra woodlands (Fig. 2A).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMoving southward, percentage changes moderate but absolute budgets remain large. In temperate systems, ABDp changes are +5\u0026ndash;14% yet amount to +7\u0026ndash;16 d\u0026middot;yr⁻\u0026sup1; on average at SSP3-7.0 by late-century; OBEp spans +3\u0026ndash;119%, with steppe regions showing the highest relative gains (Fig. 2A). Subtropical regions show ABDp \u0026minus;3\u0026ndash;+4% and OBEp +7\u0026ndash;56%, while retaining large absolute burning windows.\u003c/p\u003e\n\u003cp\u003eThis latitudinal amplification already emerges by mid-century and persists across scenarios, with contrasts scaling with warming and arising primarily from larger boreal increases (Fig. 2, Supplementary Fig. S4). For example, mid-century boreal OBEp is projected to increase by 38\u0026ndash;110% with ABDp by 9\u0026ndash;27% even at SSP1-2.6 while temperate and subtropical systems show slight or minimal response.\u003c/p\u003e\n\u003ch1\u003eRecent extremes become future routine\u0026nbsp;\u003c/h1\u003e\n\u003cp\u003eTo contextualize these projections, we examine how recent extreme fire seasons would rank within future mid- and late-century distributions. This provides critical benchmarks for evaluating whether recent extremes represent the new normal\u0026mdash;information essential for fire management agencies facing an uncertain future.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCanada 2023 in late-century: from unprecedented to multiple-per-decade.\u003c/strong\u003e Canada\u0026apos;s 2023 fire season not only broke all modern records in terms of fire attributes e.g., burned area, emissions, and intensity, but also here in climatological flammability\u0026mdash;72 ABDp and 18 OBEp d yr⁻\u0026sup1; exceed every observation in the 1991-2020 baseline (100ᵗʰ percentile for both metrics; Fig. 3). Across individual boreal biomes where Canada 2023 largely occurred, 2023 is yet typically an upper-tail year rather than a wholesale departure from baseline variability (ABDp ~67\u0026ndash;100th; OBEp ~90\u0026ndash;100th, Supplementary Fig. S5). This national-scale extreme therefore reflects the unusual synchrony of upper-tail flammability across multiple boreal biomes.\u003c/p\u003e\n\u003cp\u003eLate-century climate warming casts this benchmark as ordinary. By late century under SSP3-7.0, Canada\u0026rsquo;s 2023 ABDp and OBEp conditions drop to the 57th (range 0-97th; lower ranking = more frequent exceedance of this 2023 extreme) and 17th (0-53rd) percentiles, respectively (Fig. 3A, B). Under this high warming trajectory, 2023-level or higher conditions would occur routinely\u0026mdash;multiple times per decade rather than once in the observational record. In the most aggressive model under SSP3-7.0, the late-century means reach 91 ABDp and 38 OBEp d yr⁻\u0026sup1;, placing 2023 below the modeled late-century distribution and rendering that \u0026quot;catastrophic\u0026quot; season comparatively mild. Across scenarios, the frequency of Canada-2023 conditions scales with warming; notably, under no scenario does 2023\u0026rsquo;s OBEp value exceed the late-century distribution median, indicating that amplified diurnal burning\u0026mdash;especially sustained overnight burning\u0026mdash;may become an unavoidable reality (Fig. 3B).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCanada 2023 in mid-century: early normalization even under mitigation\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e By mid-century, Canada\u0026rsquo;s 2023 benchmark already shifts markedly within projected distributions, indicating a rapid normalization of extremes. Under SSP3-7.0, the mid-century distributions already reposition Canada-2023 toward frequent attainment (Fig. 3C, D): ABDp drops to the upper half of the distribution (multi-model means: 81st), and OBEp moves even further toward routine occurrence (37th). Strikingly, even under SSP1-2.6 by mid-century, Canada\u0026rsquo;s OBEp repositions to the median (50th) with ABDp remaining upper-tail (96th), further highlighting the urgency; SSP2-4.5 strengthens this signal and SSP5-8.5 accelerates it (Fig. 3D).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWestern US 2020/2021 in late-century: extremes edge toward the center.\u0026nbsp;\u003c/strong\u003eThe western United States experienced consecutive severe fire seasons in 2020/2021 (~210 ABDp and ~125 OBEp d yr⁻\u0026sup1;), each above the 91st percentile in the baseline distribution (Fig. 4). Under SSP3-7.0, their late-century ranks shift toward the center for ABDp (62nd and 75th, respectively) and OBEp (94th\u003csup\u003e\u0026nbsp;\u003c/sup\u003eand 78th), while multi-model means converge near 205 ABDp and 115 OBEp d yr⁻\u0026sup1;, comparable to the historical extremes themselves (Fig. 4A, B). Thus, even in a region already characterized by long burn seasons, 2020/2021-like conditions are expected to be more frequent by century\u0026rsquo;s end. Across SSPs, these rankings show modest variation, with SSP3-7.0 generally intermediate between SSP2-4.5 and SSP5-8.5.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWestern US 2020/2021 in mid-century: limited transitions.\u0026nbsp;\u003c/strong\u003eMid-century projections show comparatively modest shifts for the western US (Fig. 4C, D). Across scenarios, the 2020/2021 benchmarks largely remain in the upper tail of the projected distributions for both ABDp and OBEp, indicating that seasons comparable to 2020/2021 are not yet representative of typical mid-century conditions. Lower-warming scenarios (SSP1-2.6/2-4.5) can slightly shorten burning windows in some models, whereas higher warming tends to push distributions upward.\u003c/p\u003e"},{"header":"Discussion: North America’s fire future","content":"\u003cp\u003eOur projections confirm the alarming question posed by recent catastrophic fire seasons: the events of Canada 2023 and the Western US 2020/2021 are not statistical outliers, but early manifestations of the new reality. We find this transformation is driven by three key mechanisms: (1) a disproportionate escalation of overnight burning compared to daytime potential, fundamentally a reflection of the weakening diurnal firebreak; (2) a pronounced latitudinal amplification, concentrating the greatest relative changes in high-latitude boreal ecosystems; and (3) responses scaling with warming level, revealing that while aggressive mitigation can prevent the worst outcomes, a substantial increase in fire risk is already 'locked in' for the coming decades.\u003c/p\u003e\n\u003cp\u003eThe intensification of critical overnight burning conditions emerges early and accelerates throughout the century. Across SSPs, time periods, and climate models, projected increases in OBEp consistently outpace those of ABDp. This divergence suggests that a growing share of season’s fire potential may concentrate into multi-day, round-the-clock runs when suppression is least effective\u003csup\u003e28-30\u003c/sup\u003e.\u0026nbsp;Such temporal clustering, compounded by the early onset of overnight burning after ignition\u003csup\u003e29\u003c/sup\u003e, compresses the window for initial attack and increases the likelihood of escaped fires and subsequent extreme fire events.\u0026nbsp;More broadly, this shift indicates an ongoing restructuring of diurnal fire\u0026nbsp;cycles\u003csup\u003e31\u003c/sup\u003e. Critical burning days become more frequent, and daily burnable windows also expand alongside higher fire\u0026nbsp;intensities\u003csup\u003e20\u003c/sup\u003e. This expansion is increasingly reinforced by\u0026nbsp;weakening nighttime recovery (e.g., smaller humidity rebound and asymmetric diurnal warming\u003csup\u003e32,33\u003c/sup\u003e) and climate-driven transition to flash, intensified and prolonged droughts\u003csup\u003e34\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe degradation of diurnal burning exhibits a pronounced latitudinal gradient, with the strongest shifts concentrated in boreal ecosystems. Although northern regions historically showed lagged fire responses despite faster warming\u003csup\u003e35-39\u003c/sup\u003e, our projections indicate that accumulated warming will reverse this pattern, with boreal OBEp potentially doubling or tripling by late-century under high warming. This is consistent with projections of increasing burned area, spread rates, and extreme fire years in northern forests\u003csup\u003e5,7,17,40-43\u003c/sup\u003e and with recent record-breaking polar and tundra fires\u003csup\u003e44,45\u003c/sup\u003e. In this context, Canada’s 2023 fire season represents a prototype of the emerging baseline that conditions outside the historical range shift toward routine mid-century occurrence, even under the most ambitious scenario. This \"locked-in\" nature signals an urgent need to prepare for a new era of extreme boreal fire activity, regardless of near-term mitigation success.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMoreover, we found the unprecedented 2023 fire season arose from the simultaneous, near-maximum fire weather across the entire Canadian boreal shield. Projections further suggest that traditionally less fire-prone eastern boreal regions\u003csup\u003e24,38\u003c/sup\u003e may experience increases comparable to, or greater than, those in the west. Such spatial coherence challenges fire management strategies that rely on inter-regional resource sharing during peak seasons\u003csup\u003e9,30,46\u003c/sup\u003e, designed for localized hotspots rather than nation-wide synchronized extremes. In contrast to this rapid and synchronized boreal transition, the Western US exhibits a delayed shift toward typical conditions. While the 2020/2021 fire seasons remain high-percentile events through mid-century, this apparent stability reflects an already elevated baseline rather than lower risk\u003csup\u003e18,43,47\u003c/sup\u003e. The convergence of these recent extremes toward the mean occurs later in the century under higher warming scenarios, which suggests the Western US faces a sustained persistence of its current high-severity regime before further intensification.\u003c/p\u003e\n\u003cp\u003eWhile our analysis quantifies the climatic envelope of fire potential, its translation into realized fire activity will be modulated by dynamic fuels, ignition regimes, and human intervention. Dynamic fuels can constrain and amplify the expression of longer potential burning windows through biomass availability, post-disturbance recovery, and vegetation transitions\u003csup\u003e11,48-50\u003c/sup\u003e. Ignition regimes may shift with changing convective environments and continued wildland-urban interface expansion\u003csup\u003e51,52\u003c/sup\u003e. Human actions, including suppression legacies, land use, infrastructure, and treatments, can likewise influence the translation of potential\u003csup\u003e53-56\u003c/sup\u003e. These factors add uncertainties but are unlikely to negate the strong forcing of the expanding diurnal fire niche; instead, they determine where and how rapidly projected potential is expressed. Future work should therefore couple diurnal fire-weather potential with dynamic fuel and ignition modules, and evaluate how management interventions alter the probability that projected overnight windows translate into large-fire growth.\u003c/p\u003e\n\u003cp\u003eThe convergence of evidence indicates that the fundamental transformation of North America's fire regime is not a distant possibility but an ongoing reality. Recent extreme seasons preview a future where overnight burning becomes commonplace, fire seasons extend substantially, and suppression resources designed for 20th-century fire regimes face mounting pressure. The choice between aggressive mitigation and high emissions will determine whether future fire regimes remain marginally manageable or face more substantial challenges. As Canada's 2023 fire season demonstrated, when boreal systems approach critical thresholds, the wildfires impact the whole planet through widespread smoke, carbon emissions, and economic disruption\u003csup\u003e12,13,57\u003c/sup\u003e. Our projections suggest such fire seasons will intensify and regularize in the coming decades, underscoring the urgent need for accelerated adaptation strategies, especially for northern regions.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy Area and Biome Classification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur study area covers the continental United States and Canada, employing a detailed biome classification derived from ref.\u003csup\u003e58\u003c/sup\u003e and extending approaches from ref.\u003csup\u003e39\u003c/sup\u003e. The continental domain is divided into 20 distinct biomes, each defined by distinctive climatic and vegetative characteristics. The classification system includes five primary Boreal biomes (western and eastern coniferous forests, western and eastern tundra woodlands, plus mountain systems), six Temperate biomes (continental forest, oceanic forest, western and eastern mountain systems, steppe, and desert), and five Subtropical zones (humid forest, dry forest, mountain system, steppe, and desert). The classification is completed with tropical, polar, and water biomes for comprehensive continental representation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBased on satellite-derived observations of active burning days (ABD) and overnight burning events (OBE) from 2017-2023 across North America, we identified 11 biomes as burnable and fire-prone areas: Subtropical steppe, Subtropical desert, Subtropical mountain system, Temperate steppe, Temperate desert, Temperate mountain system west, Boreal coniferous forest east, Boreal coniferous forest west, Boreal tundra woodland east, Boreal tundra woodland west, and Boreal mountain system. For western U.S. 2020-2021 analysis, the area is specifically defined as areas overlapped between fire-prone biomes and Washington, Oregon, California, Nevada, Idaho, Montana, Wyoming, Utah, Colorado, Arizona, and New Mexico.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWildland Fire Geospatial Databases\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study incorporates wildland fire geospatial data from three comprehensive databases that document fire perimeters and other important information including e.g., burned areas and ignition date throughout North America. We utilized the Canadian National Burned Area Composite (NBAC)\u003csup\u003e59\u003c/sup\u003e as our source for Canadian fire data, which serves as a key component within the Fire Monitoring, Accounting and Reporting System (FireMARS). Developed collaboratively by the Canada Centre for Mapping and Earth Observation and the Canadian Forest Service, NBAC delivers burned area mapping spanning 1972 to 2023 (version 20240530). The dataset maintains high spatial precision by integrating 30-meter Landsat imagery with higher-resolution agency imagery (\u0026lt;30-meter spatial resolution).\u003c/p\u003e\n\u003cp\u003eU.S. fire information was obtained through the Monitoring Trends in Burn Severity (MTBS)\u003csup\u003e60\u003c/sup\u003e database as the primary source, with additional coverage from the Interagency Wildland Fire Perimeter History (IWFPH). MTBS represents a collaborative effort between the U.S. Geological Survey (EROS) and the USDA Forest Service (GTAC), documenting burn severity and extent throughout the United States from 1984 to 2022 (released August 22, 2024). The database applies regionally-specific inclusion criteria: fires ≥1,000 acres (~405 ha) in the western U.S. and ≥500 acres (~202 ha) in the eastern U.S. To achieve complete temporal coverage through 2023, we supplemented MTBS records with fire perimeter data from the IWFPH (version 20240825), which is maintained by the Wildland Fire Management Research, Development, \u0026amp; Application program data team and distributed through the National Interagency Fire center (NIFC).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGeostationary Active Fire Detections\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing ref.\u003csup\u003e29\u003c/sup\u003e, we used the Fire/Hot Spot Characterization Full Disk (FDCF) products from the GOES-R series (GOES-16/17/18) to obtain sub-hourly active-fire detections across North America for 2017–2023\u003csup\u003e61\u003c/sup\u003e. GOES-16 (GOES-East) has operated near 75.2°W since 2017. GOES-17 (GOES-West) observed near 137.2°W from late 2018 until early 2023, and GOES-18 has occupied ~137.2°W since late 2022, replacing GOES-17 as the West satellite in 2023. Accordingly, we used GOES-16 from its first available date through 2023, GOES-17 from its first available date through 2022, and GOES-18 for 2023. Together, GOES-East and GOES-West provide geostationary coverage of burnable land across North and South America. FDCF is derived from the Advanced Baseline Imager (ABI) visible and infrared bands and reports sub-pixel fire characteristics at a 5–15-min temporal resolution, with ~2 km nominal pixel size at the sub-satellite point and coarser effective resolution at larger view-zenith angles (VZA).\u003c/p\u003e\n\u003cp\u003eThe availability, sampling frequency, and overall data quality of GOES-R FDCF vary across space and time during 2017–2023 due to the staggered satellite deployments, changes in imaging frequency, differences in coverage and VZA between GOES-East and GOES-West, and occasional instrument outages. Given these inconsistencies, we used FDCF to infer binary hourly burning status (active vs. inactive) for fires with well-defined spatiotemporal boundaries in wildland fire databases, rather than relying on hotspot counts. As the available data provide at least four observations per hour at any location across the study area, we considered these data fit for this purpose.\u003c/p\u003e\n\u003cp\u003eSpatial coverage differs between the East and West platforms. Northwestern North America (including Alaska and Yukon) is not visible to GOES-East; therefore, detections there are only available from August 2018 onward, when GOES-17 (and later GOES-18) became operational. In contrast, northeastern North America is only observed by GOES-East. In the central region where coverage overlapped, the data volume approximately doubled after GOES-17/18 was active. Imaging frequency changed during the study period as satellites transitioned among operational scan modes, producing full-disk imagery (and thus FDCF) at 15 min (Mode 3), 10 min (Mode 6), or 5 min (Mode 4). In the early mission phase (2017–2018), GOES-16 and GOES-17 were predominantly in Mode 3, whereas Mode 6 became the dominant configuration from 2019 onward for GOES-16/17 and subsequently for GOES-18. As a result, later years provide denser FDCF sampling. For more information on GOES-R scanning mode, please refer to https://www.goes-r.gov/. We pooled detections from GOES-16/17/18 to create a single continental-scale dataset, aiming to maximize the completeness of the geostationary record for each fire location rather than to compare satellites.\u003c/p\u003e\n\u003cp\u003eSeveral technical considerations arise from the fixed geostationary viewing geometry. For any ground point, VZA is effectively constant for a given satellite, and larger VZA degrades detection performance in multiple ways\u003csup\u003e62,63\u003c/sup\u003e. In particular, ABI pixel size and omission errors generally increase with VZA, especially for small/cool fires and in high-latitude or complex-terrain settings. These limitations further motivate our focus on binary burning status (rather than hotspot counts). Geometric effects and terrain can also introduce parallax-related geolocation offsets; to reduce mismatches between hotspot locations and mapped fire perimeters, we applied a 2 km buffer (approximately one nominal ABI fire-pixel width) around each perimeter when associating detections with individual fires (see “Active Burning Days and Overnight Burning Events”). Finally, nighttime backgrounds are typically cooler and more spatially homogeneous, improving contrast for active-fire pixels; consequently, the nighttime detection algorithm is generally more sensitive than the daytime algorithm, particularly for smaller and/or cooler fires\u003csup\u003e64\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eActive Burning Days and Overnight Burning Events\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe established a framework to assess hourly burning status (active or inactive) across fire events. The methodology commenced by extracting GOES-R active fire hotspots that spatially intersected with fire perimeters from our integrated database (NBAC and MTBS-IWFPH). To associate hotspots with events, we intersected GOES-R hotspots with perimeters buffered by 2 km (~nominal FDCF pixel width), which accommodates ABI geolocation error, parallax at oblique angles, and perimeter uncertainty. Fire event temporal boundaries were determined through a dual strategy: incorporating documented start and end dates where available, or deriving these dates from GOES-R hotspot observations based on initial and final instances of consecutive active burning hours. Fire events from 2017-2023 lacking corresponding hotspot data were excluded from analysis, recognizing that such omissions may stem from detection algorithm constraints under specific conditions\u003csup\u003e64\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eFor temporal precision, all fire activity data were transformed from Coordinated Universal Time (UTC) to local time zones, utilizing fire perimeter spatial centroids and accounting for day of year variations. We implemented exact daylight delineation using precise sunrise and sunset calculations, with sunrise defined as the moment the sun's upper edge becomes visible on the horizon and sunset as its disappearance below the horizon. This framework enabled categorization of each hour into four distinct states: active daytime, non-active daytime, active night-time, or non-active night-time. For analytical consistency, we defined a day as spanning from the first hour following sunrise to the final hour preceding the subsequent sunrise and defined each hour as a half-open interval [hh:00:00, hh:59:59)—daytime spans from the first hour after sunrise to the last hour before sunset and hours outside this range are nighttime. The event temporal window is therefore from the first hour of the start date to the last hour of the end date.\u003c/p\u003e\n\u003cp\u003eFire activity classification criteria were established whereby a fire was deemed active during any hour containing at least one detected hotspot within its perimeter. Non-active hours were those with no detected hotspots within the fire perimeter. We derived ABD as days containing at least one active hour; non-ABD were days with no active hours. For OBE, nights were classified when all night-time hours registered as active, indicating continuous burning throughout the entire night\u003csup\u003e29\u003c/sup\u003e. Non-OBE included nights with incomplete fire activity across night-time hours. OBE represents fire persistence during the diurnally least favorable burning conditions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe selected GOES-R geostationary satellite products as they provide the only consistent, high-frequency (≤15-minute) fire detection capability for North America, uniquely enabling capture of hourly fire patterns. However, the burning hours and thus ABD and OBE likely represent underestimates. This is because, within the abovementioned spatio-temporal window, the physical fire state is theoretically either flaming or smouldering at the most of time; however, when status needs to be determined from the geostationary perspective, “non-active” hour may reflect non-flaming/smoldering combustion below the detection threshold, very small flaming area, cloud/canopy obstruction, unfavorable viewing geometry, or genuine inactivity\u003csup\u003e64\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFire Weather Calculation and Extraction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFire weather variables were derived based on ERA5 reanalysis\u003csup\u003e65\u003c/sup\u003e and the Canadian Forest Fire Weather Index System (CFWIS), a widely used operational framework for assessing fire danger that calculates six components, including three fuel moisture codes (Fine Fuel Moisture Code (FFMC), Duff Moisture Code (DMC), and Drought Code (DC)) and three fire behavior indices (Initial Spread Index (ISI), Buildup Index (BUI), and Fire Weather Index (FWI))\u003csup\u003e66\u003c/sup\u003e. ERA5 provides global atmospheric fields on a 0.25° horizontal grid with hourly resolution.\u003c/p\u003e\n\u003cp\u003eFrom ERA5 reanalysis for the period 1991-2023, we obtained 2-m temperature, 2-m dewpoint temperature, 10-m u-component of wind, 10-m v-component of wind, and 24-hour precipitation to derive the required inputs at noon local standard time for CFWIS calculations: temperature, relative humidity, wind speed, and accumulated precipitation. Daily CFWIS components were calculated using the cffdrs R package\u003csup\u003e67\u003c/sup\u003e following established overwintering procedures\u003csup\u003e68\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eFor each fire event occurring during 2017-2023, the extracted CFWIS variables were spatially averaged across all grid cells that intersected the fire perimeter. The extraction period spanned the entire lifetime of each fire, with a 1-day buffer added at the start to capture early fire conditions. Finally, these daily fire weather metrics were time-matched to the corresponding fire dynamics used in our analyses.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLink ABD and OBE to Daily Fire Weather\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo link daily fire weather conditions with ABD and OBE occurrence, we developed machine learning models using the matched fire dynamics and fire weather data from 2017-2023. The feature set included all six CFWIS components (FFMC, DMC, DC, ISI, BUI, FWI) along with categorical variables for biome type and month. Prior to model training, numerical features were standardized and categorical variables were one-hot encoded. We used random forest because it is an ensemble model, and it is built on multiple decision tree classifiers and can capture complex, non-linear relationships between fire weather and fire activity with robust performance.\u003c/p\u003e\n\u003cp\u003eWe employed a hierarchical random forest modeling framework with two stages: (1) a binary classifier predicting whether a day was an ABD or non-ABD, and (2) a conditional model predicting OBE binary occurrence given ABD conditions. To address class imbalance in the training data, we applied random undersampling to create balanced datasets while preserving all positive instances. Model development utilized 5-fold cross-validation repeated three times, with hyperparameters including 500 trees and square root of features for node splitting.\u003c/p\u003e\n\u003cp\u003eModel performance was evaluated using area under the ROC curve (AUC), F1-score, and precision-recall metrics. AUC summarizes the model’s ability to discriminate between positive and negative classes across all possible classification thresholds. The F1 score is the harmonic mean of precision and recall (sensitivity), where precision is the fraction of predicted positives that are true positives, and recall is the fraction of true positives that are correctly identified. We selected the classification threshold by maximizing F1 along the precision–recall curve, and report the resulting AUC and F1 for both the ABD and OBE models. The ABD prediction model achieved AUC = 0.814 and F1 = 0.750. The OBE prediction model demonstrated stronger performance with AUC = 0.908 and F1 = 0.841. These metrics indicate robust predictive capability for both fire activity measures.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFuture Fire Weather Projections\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo assess future changes in fire weather conditions, we developed daily fire weather projections for two future periods (mid-century 2041-2070 and late-century 2071-2100) using a delta change downscaling approach. Future climate projections\u003csup\u003e69\u003c/sup\u003e were obtained from four CMIP6 General Circulation Models (GCMs): CanESM, UKESM, EC-Earth, and GFDL, selected based on guidelines in ref.\u003csup\u003e70\u003c/sup\u003e and their demonstrated good performance in evaluation by ref.\u003csup\u003e71\u003c/sup\u003e. We analyzed four Shared Socioeconomic Pathways (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) from ambitious mitigation to very high warming scenarios.\u003c/p\u003e\n\u003cp\u003eSpecifically, with the raw outputs from the GCMs, we calculated monthly 30-year averages for the periods 1990-2020 (GCM baseline) and the two future periods. We then subtracted the monthly GCM baseline averages from the monthly GCM future averages to obtain the change in climate conditions between time periods. Temperature and wind speed were calculated as simple delta changes (e.g., Tfuture - Tbaseline), while precipitation and relative humidity were calculated as ratio changes (e.g., RHfuture - RHbaseline)/RHbaseline). Since the output of GCMs varies in spatial resolution, all of the anomalies were downscaled by bilinear interpolation to the resolution of the daily baseline weather from ERA5 (0.25°).\u003c/p\u003e\n\u003cp\u003eFinally, we applied the monthly delta changes from each future period/GCM/SSP combination to the baseline 1991-2020 daily variables from ERA5 in the same way as the deltas were calculated: temperature and wind speed as simple additions, and precipitation and relative humidity as ratio changes. This process generated 32 sets of future daily meteorological variables (4 GCMs × 4 SSPs x 2 periods). Using these datasets, we then calculated the corresponding daily CFWIS components detailed in ‘Fire Weather Calculation and Extraction’ section, producing comprehensive daily fire weather index projections that serve as inputs for our ABDp and OBEp predictions. Notably, as the delta approach does not change the frequency of precipitation events, projected changes may be conservative\u003csup\u003e72\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProjecting Future ABDp and OBEp\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe applied the trained models to our 32 sets of future daily fire weather projections to generate gridded estimates of daily potential for active burning days (ABDp) and overnight burning events (OBEp). For each grid cell and day in the future scenarios, we: (1) extracted the corresponding fire weather indices, biome type, and month; (2) applied the same standardization and encoding transformations used during training; (3) generated probability estimates using the hierarchical model structure; and (4) classified each day based on whether its probability exceeded the optimal threshold, then aggregated daily probabilities to annual sums, representing the expected number of days per year with conditions conducive to active burning or overnight events.\u003c/p\u003e\n\u003cp\u003eThis approach produced spatially explicit projections of ABDp and OBEp at 0.25° resolution across North America for both mid-century (2041-2070) and late-century (2071-2100) periods under each GCM-SSP combination. The resulting maps provide quantitative assessments of how fire potential may shift across different regions and time periods under varying climate scenarios.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAnalysis of Projected Changes and Spatial Agreement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo quantify continental-scale changes in fire potential, we calculated statistics across all grid cells within burnable and fire-prone biomes. For each GCM–SSP combination, we first aggregated daily ABDp and OBEp predictions to annual totals for each grid cell and for each year within a 30-year period (baseline: 1991–2020; mid-century: 2041–2070; late-century: 2071–2100). We then computed period means as the average of the 30 annual totals at each grid cell. Absolute changes were calculated as (future period mean − baseline period mean), and relative changes as ((future period mean − baseline period mean) / baseline period mean × 100%) for both ABDp and OBEp at each grid cell. Continental means were derived by averaging grid-cell period means across all valid grid cells, with model ranges reported as minimum to maximum values across the four GCMs. To assess spatial coherence of projections, we identified grid cells where all four models (or three out of four models) agreed on the direction of the period-mean change (positive or negative).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBiome-Level Analysis and Latitudinal Patterns\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor biome-specific assessments, we aggregated grid-cell ABDp and OBEp annual totals by biome type for baseline and future periods. We computed biome-level statistics including 30-year biome mean changes and percentile shifts to quantify distribution transformations. The latitudinal gradient was examined through the inherent latitudinal zonation of biome types (boreal, temperate, and subtropical), allowing assessment of the latitudinal fire potential changes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eContextualization of Recent Extreme Fire Seasons\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo place recent extreme fire seasons in the context of future projections, we compared observed ABDp and OBEp values from specific events against the 30-year distributions of area-averaged annual values in both baseline and two future periods. For Canada's 2023 season and Western US 2020/2021 seasons, we extracted area-averaged ABDp and OBEp values using the same spatial domains and methodologies applied to the projections. We then calculated the percentile rank of these observed values within: (1) the 1991-2020 baseline distribution, and (2) the 2041-2070 and 2071-2100 projected distributions under each GCM-SSP combination. Distributions were constructed from the 30 annual area-averaged values within each period (i.e., one value per year), and percentile calculations used empirical cumulative distribution functions; ranges across the four GCMs were reported to show projection spread. This approach quantified how frequently future climate conditions would produce fire weather matching or exceeding recent extremes. For Canada 2023, we additionally repeated this contextualization within each boreal biome (where most burned area occurred), computing biome-specific area-averaged annual values and percentiles to capture how the national extreme reflects a sum of biome-level extremes. The area definition of the Western US can be found in the section \"Study Area and Biome Classification\".\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch1\u003eData availability\u003c/h1\u003e\n\u003cp\u003eThe datasets for conducting the analysis presented here are all publicly available. The NBAC, MTBS and IWFPH wildland fire datasets are respectively available from the Canadian Forest Service (https://cwfis.cfs.nrcan.gc.ca/datamart/metadata/nbac), https://www.mtbs.gov/ and the National Interagency Fire Center (https://data-nifc.opendata.arcgis.com/datasets/nifc::interagencyfireperimeterhistory-all-years-view/about). The GOES-16, GOES-17, and GOES-18 full disk active fire products are available on Amazon Web Service S3 Explorer (https://registry.opendata.aws/noaa-goes/). The hourly ERA5 climate data are available at https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview. The CMIP6 climate projections are available at https://cds.climate.copernicus.eu/datasets/projections-cmip6?tab=overview. The biome categorizations used in this study are available at https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world.\u0026nbsp;\u003c/p\u003e\n\u003ch1\u003eCode availability\u003c/h1\u003e\n\u003cp\u003eCodes used to analyse the data are available from https://github.com/KaiweiLL/Canada-2023-diurnal-projections.\u0026nbsp;\u003c/p\u003e\n\u003ch1\u003eAuthor contributions\u003c/h1\u003e\n\u003cp\u003eConceptualization: M.F., X.W. and K.L. Methodology: K.L., X.W., and M.F. Data processing: K.L. and D.C. Investigation: K.L., X.W., and M.F. Visualization: K.L. Funding acquisition: K.L., X.W., and M.F. Project administration: M.F. Supervision: M.F. and X.W. Writing – original draft: K.L. Writing – review \u0026amp; editing: K.L., X.W., and M.F.\u0026nbsp;\u003c/p\u003e\n\u003ch1\u003eCompeting interests\u003c/h1\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eFlannigan, M., Stocks, B., Turetsky, M. \u0026amp; Wotton, M. Impacts of climate change on fire activity and fire management in the circumboreal forest. \u003cem\u003eGlobal change biology\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 549\u0026ndash;560 (2009). \u003c/li\u003e\n\u003cli\u003eRogers, B. M., Soja, A. J., Goulden, M. L. \u0026amp; Randerson, J. T. Influence of tree species on continental differences in boreal fires and climate feedbacks. \u003cem\u003eNature Geoscience\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 228\u0026ndash;234 (2015). \u003c/li\u003e\n\u003cli\u003eWotton, B. M., Flannigan, M. D. \u0026amp; Marshall, G. A. Potential climate change impacts on fire intensity and key wildfire suppression thresholds in Canada. \u003cem\u003eEnviron Res Lett\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 095003 (2017). \u003c/li\u003e\n\u003cli\u003eFlannigan, M. D., Krawchuk, M. A., de Groot, W. J., Wotton, B. M. \u0026amp; Gowman, L. M. Implications of changing climate for global wildland fire. \u003cem\u003eInternational journal of wildland fire\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 483\u0026ndash;507 (2009). \u003c/li\u003e\n\u003cli\u003eFlannigan, M. D., Logan, K. A., Amiro, B. D., Skinner, W. R. \u0026amp; Stocks, B. J. Future area burned in Canada. \u003cem\u003eClimatic change\u003c/em\u003e \u003cstrong\u003e72\u003c/strong\u003e, 1\u0026ndash;16 (2005). \u003c/li\u003e\n\u003cli\u003eWang, X., Swystun, T. \u0026amp; Flannigan, M. D. Future wildfire extent and frequency determined by the longest fire-conducive weather spell. \u003cem\u003eScience of the total environment\u003c/em\u003e \u003cstrong\u003e830\u003c/strong\u003e, 154752 (2022). \u003c/li\u003e\n\u003cli\u003eWang, X.\u003cem\u003e et al.\u003c/em\u003e Projected changes in fire size from daily spread potential in Canada over the 21st century. \u003cem\u003eEnviron Res Lett\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 104048 (2020). \u003c/li\u003e\n\u003cli\u003eCoops, N. C., Hermosilla, T., Wulder, M. A., White, J. C. \u0026amp; Bolton, D. K. A thirty year, fine-scale, characterization of area burned in Canadian forests shows evidence of regionally increasing trends in the last decade. \u003cem\u003ePlos one\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, e0197218 (2018). \u003c/li\u003e\n\u003cli\u003eScholten, R. C., Veraverbeke, S., Chen, Y. \u0026amp; Randerson, J. T. Spatial variability in Arctic\u0026ndash;boreal fire regimes influenced by environmental and human factors. \u003cem\u003eNature Geoscience\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 866\u0026ndash;873 (2024). \u003c/li\u003e\n\u003cli\u003eAmiro, B., Stocks, B., Alexander, M., Flannigan, M. \u0026amp; Wotton, B. Fire, climate change, carbon and fuel management in the Canadian boreal forest. \u003cem\u003eInternational Journal of Wildland Fire\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 405\u0026ndash;413 (2001). \u003c/li\u003e\n\u003cli\u003eHanes, C. C.\u003cem\u003e et al.\u003c/em\u003e Fire-regime changes in Canada over the last half century. \u003cem\u003eCanadian Journal of Forest Research\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 256\u0026ndash;269 (2019). \u003c/li\u003e\n\u003cli\u003eJain, P.\u003cem\u003e et al.\u003c/em\u003e Drivers and impacts of the record-breaking 2023 wildfire season in Canada. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 6764 (2024). \u003c/li\u003e\n\u003cli\u003eByrne, B.\u003cem\u003e et al.\u003c/em\u003e Carbon emissions from the 2023 Canadian wildfires. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e633\u003c/strong\u003e, 835\u0026ndash;839 (2024). \u003c/li\u003e\n\u003cli\u003eSafford, H. D., Paulson, A. K., Steel, Z. L., Young, D. J. \u0026amp; Wayman, R. B. The 2020 California fire season: A year like no other, a return to the past or a harbinger of the future? \u003cem\u003eGlobal Ecology and Biogeography\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 2005\u0026ndash;2025 (2022). \u003c/li\u003e\n\u003cli\u003eJain, P., Sharma, A. R., Acuna, D. C., Abatzoglou, J. T. \u0026amp; Flannigan, M. Record-breaking fire weather in North America in 2021 was initiated by the Pacific northwest heat dome. \u003cem\u003eCommunications Earth \u0026amp; Environment\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 202 (2024). \u003c/li\u003e\n\u003cli\u003eQuilcaille, Y., Batibeniz, F., Ribeiro, A. F., Padr\u0026oacute;n, R. S. \u0026amp; Seneviratne, S. I. Fire weather index data under historical and SSP projections in CMIP6 from 1850 to 2100. \u003cem\u003eEarth System Science Data Discussions\u003c/em\u003e \u003cstrong\u003e2022\u003c/strong\u003e, 1\u0026ndash;31 (2022). \u003c/li\u003e\n\u003cli\u003eWang, X.\u003cem\u003e et al.\u003c/em\u003e Critical fire weather conditions during active fire spread days in Canada. \u003cem\u003eScience of the total environment\u003c/em\u003e \u003cstrong\u003e869\u003c/strong\u003e, 161831 (2023). \u003c/li\u003e\n\u003cli\u003eAbatzoglou, J. T. \u0026amp; Williams, A. P. Impact of anthropogenic climate change on wildfire across western US forests. \u003cem\u003eProceedings of the National Academy of sciences\u003c/em\u003e \u003cstrong\u003e113\u003c/strong\u003e, 11770\u0026ndash;11775 (2016). \u003c/li\u003e\n\u003cli\u003eBalch, J. K.\u003cem\u003e et al.\u003c/em\u003e The fastest-growing and most destructive fires in the US (2001 to 2020). \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e386\u003c/strong\u003e, 425\u0026ndash;431 (2024). \u003c/li\u003e\n\u003cli\u003eKaiwei Luo, X. W., Dante Castellanos-Acuna, Mike Flannigan. A weakened diurnal weather cycle leads to longer burning hours in North America (Under Review). (2025). \u003c/li\u003e\n\u003cli\u003eStocks, B.\u003cem\u003e et al.\u003c/em\u003e Large forest fires in Canada, 1959\u0026ndash;1997. \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e \u003cstrong\u003e107\u003c/strong\u003e, FFR 5\u0026ndash;1\u0026ndash;FFR 5\u0026ndash;12 (2002). \u003c/li\u003e\n\u003cli\u003eJolly, W. M.\u003cem\u003e et al.\u003c/em\u003e Climate-induced variations in global wildfire danger from 1979 to 2013. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 7537 (2015). \u003c/li\u003e\n\u003cli\u003eCunningham, C. X., Williamson, G. J. \u0026amp; Bowman, D. M. Increasing frequency and intensity of the most extreme wildfires on Earth. \u003cem\u003eNature ecology \u0026amp; evolution\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 1420\u0026ndash;1425 (2024). \u003c/li\u003e\n\u003cli\u003eWang, X.\u003cem\u003e et al.\u003c/em\u003e Increasing frequency of extreme fire weather in Canada with climate change. \u003cem\u003eClimatic Change\u003c/em\u003e \u003cstrong\u003e130\u003c/strong\u003e, 573\u0026ndash;586 (2015). \u003c/li\u003e\n\u003cli\u003eBedia, J.\u003cem\u003e et al.\u003c/em\u003e Global patterns in the sensitivity of burned area to fire-weather: Implications for climate change. \u003cem\u003eAgr Forest Meteorol\u003c/em\u003e \u003cstrong\u003e214\u003c/strong\u003e, 369\u0026ndash;379 (2015). \u003c/li\u003e\n\u003cli\u003eMatteo, A.\u003cem\u003e et al.\u003c/em\u003e Challenges in assessing Fire Weather changes in a warming climate. \u003cem\u003earXiv preprint arXiv:2503.01818\u003c/em\u003e (2025). \u003c/li\u003e\n\u003cli\u003eFreeborn, P. H., Jolly, W. M., Cochrane, M. A. \u0026amp; Roberts, G. Large wildfire driven increases in nighttime fire activity observed across CONUS from 2003\u0026ndash;2020. \u003cem\u003eRemote Sens Environ\u003c/em\u003e \u003cstrong\u003e268\u003c/strong\u003e, 112777 (2022). \u003c/li\u003e\n\u003cli\u003eBalch, J. K.\u003cem\u003e et al.\u003c/em\u003e Warming weakens the night-time barrier to global fire. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e602\u003c/strong\u003e, 442\u0026ndash;448 (2022). \u003c/li\u003e\n\u003cli\u003eLuo, K., Wang, X., de Jong, M. \u0026amp; Flannigan, M. Drought triggers and sustains overnight fires in North America. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e627\u003c/strong\u003e, 321\u0026ndash;327 (2024). \u003c/li\u003e\n\u003cli\u003eTymstra, C., Stocks, B. J., Cai, X. \u0026amp; Flannigan, M. D. Wildfire management in Canada: Review, challenges and opportunities. \u003cem\u003eProgress in Disaster Science\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 100045 (2020). \u003c/li\u003e\n\u003cli\u003eKreye, J. K.\u003cem\u003e et al.\u003c/em\u003e Fire behavior in masticated fuels: A review. \u003cem\u003eForest Ecology and Management\u003c/em\u003e \u003cstrong\u003e314\u003c/strong\u003e, 193\u0026ndash;207 (2014). \u003c/li\u003e\n\u003cli\u003eDavy, R., Esau, I., Chernokulsky, A., Outten, S. \u0026amp; Zilitinkevich, S. Diurnal asymmetry to the observed global warming. \u003cem\u003eInternational Journal of Climatology\u003c/em\u003e \u003cstrong\u003e37\u003c/strong\u003e, 79\u0026ndash;93 (2017). \u003c/li\u003e\n\u003cli\u003eChiodi, A. M., Potter, B. E. \u0026amp; Larkin, N. K. Multi‐decadal change in western US nighttime vapor pressure deficit. \u003cem\u003eGeophys Res Lett\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, e2021GL092830 (2021). \u003c/li\u003e\n\u003cli\u003eYuan, X.\u003cem\u003e et al.\u003c/em\u003e A global transition to flash droughts under climate change. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e380\u003c/strong\u003e, 187\u0026ndash;191 (2023). \u003c/li\u003e\n\u003cli\u003eBush, E. \u0026amp; Lemmen, D. S. Canada\u0026apos;s changing climate report. \u003cem\u003e(No Title)\u003c/em\u003e (2019). \u003c/li\u003e\n\u003cli\u003eSerreze, M. C. \u0026amp; Barry, R. G. Processes and impacts of Arctic amplification: A research synthesis. \u003cem\u003eGlobal Planet Change\u003c/em\u003e \u003cstrong\u003e77\u003c/strong\u003e, 85\u0026ndash;96 (2011). \u003c/li\u003e\n\u003cli\u003ePithan, F. \u0026amp; Mauritsen, T. Arctic amplification dominated by temperature feedbacks in contemporary climate models. \u003cem\u003eNature geoscience\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 181\u0026ndash;184 (2014). \u003c/li\u003e\n\u003cli\u003eBergeron, Y., Gauthier, S., Kafka, V., Lefort, P. \u0026amp; Lesieur, D. Natural fire frequency for the eastern Canadian boreal forest: consequences for sustainable forestry. \u003cem\u003eCanadian journal of forest research\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 384\u0026ndash;391 (2001). \u003c/li\u003e\n\u003cli\u003eJain, P., Castellanos-Acuna, D., Coogan, S. C., Abatzoglou, J. T. \u0026amp; Flannigan, M. D. Observed increases in extreme fire weather driven by atmospheric humidity and temperature. \u003cem\u003eNature Climate Change\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 63\u0026ndash;70 (2022). \u003c/li\u003e\n\u003cli\u003eZhao, H.\u003cem\u003e et al.\u003c/em\u003e Future enhanced threshold effects of wildfire drivers could increase burned areas in northern mid-and high latitudes. \u003cem\u003eCommunications Earth \u0026amp; Environment\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 224 (2025). \u003c/li\u003e\n\u003cli\u003eFlannigan, M. D., Stocks, B. J. \u0026amp; Wotton, B. M. Climate change and forest fires. \u003cem\u003eScience of the total environment\u003c/em\u003e \u003cstrong\u003e262\u003c/strong\u003e, 221\u0026ndash;229 (2000). \u003c/li\u003e\n\u003cli\u003ede Groot, W. J., Flannigan, M. D. \u0026amp; Cantin, A. S. Climate change impacts on future boreal fire regimes. \u003cem\u003eForest Ecology and Management\u003c/em\u003e \u003cstrong\u003e294\u003c/strong\u003e, 35\u0026ndash;44 (2013). \u003c/li\u003e\n\u003cli\u003eAbatzoglou, J. T.\u003cem\u003e et al.\u003c/em\u003e Climate change has increased the odds of extreme regional forest fire years globally. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 6390 (2025). \u003c/li\u003e\n\u003cli\u003eHu, F. S.\u003cem\u003e et al.\u003c/em\u003e Arctic tundra fires: natural variability and responses to climate change. \u003cem\u003eFrontiers in Ecology and the Environment\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 369\u0026ndash;377 (2015). \u003c/li\u003e\n\u003cli\u003eDescals, A.\u003cem\u003e et al.\u003c/em\u003e Unprecedented fire activity above the Arctic Circle linked to rising temperatures. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e378\u003c/strong\u003e, 532\u0026ndash;537 (2022). \u003c/li\u003e\n\u003cli\u003eErni, S., Arseneault, D., Parisien, M. A. \u0026amp; B\u0026eacute;gin, Y. Spatial and temporal dimensions of fire activity in the fire‐prone eastern Canadian taiga. \u003cem\u003eGlobal Change Biology\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1152\u0026ndash;1166 (2017). \u003c/li\u003e\n\u003cli\u003eAbatzoglou, J. T.\u003cem\u003e et al.\u003c/em\u003e Projected increases in western US forest fire despite growing fuel constraints. \u003cem\u003eCommunications Earth \u0026amp; Environment\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 227 (2021). \u003c/li\u003e\n\u003cli\u003eAndela, N.\u003cem\u003e et al.\u003c/em\u003e A human-driven decline in global burned area. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e356\u003c/strong\u003e, 1356\u0026ndash;1362 (2017). \u003c/li\u003e\n\u003cli\u003eJones, M. W.\u003cem\u003e et al.\u003c/em\u003e Global and regional trends and drivers of fire under climate change. \u003cem\u003eReviews of Geophysics\u003c/em\u003e \u003cstrong\u003e60\u003c/strong\u003e, e2020RG000726 (2022). \u003c/li\u003e\n\u003cli\u003eBaltzer, J. L.\u003cem\u003e et al.\u003c/em\u003e Increasing fire and the decline of fire adapted black spruce in the boreal forest. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e118\u003c/strong\u003e, e2024872118 (2021). \u003c/li\u003e\n\u003cli\u003eRomps, D. M., Seeley, J. T., Vollaro, D. \u0026amp; Molinari, J. Projected increase in lightning strikes in the United States due to global warming. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e346\u003c/strong\u003e, 851\u0026ndash;854 (2014). \u003c/li\u003e\n\u003cli\u003eRadeloff, V. C.\u003cem\u003e et al.\u003c/em\u003e Rapid growth of the US wildland-urban interface raises wildfire risk. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e115\u003c/strong\u003e, 3314\u0026ndash;3319 (2018). \u003c/li\u003e\n\u003cli\u003eParisien, M.-A.\u003cem\u003e et al.\u003c/em\u003e Fire deficit increases wildfire risk for many communities in the Canadian boreal forest. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 2121 (2020). \u003c/li\u003e\n\u003cli\u003ePrichard, S. J.\u003cem\u003e et al.\u003c/em\u003e Adapting western North American forests to climate change and wildfires: 10 common questions. \u003cem\u003eEcological applications\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, e02433 (2021). \u003c/li\u003e\n\u003cli\u003eKreider, M. R.\u003cem\u003e et al.\u003c/em\u003e Fire suppression makes wildfires more severe and accentuates impacts of climate change and fuel accumulation. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 2412 (2024). \u003c/li\u003e\n\u003cli\u003eFernandes, P. M. \u0026amp; Botelho, H. S. A review of prescribed burning effectiveness in fire hazard reduction. \u003cem\u003eInternational Journal of wildland fire\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 117\u0026ndash;128 (2003). \u003c/li\u003e\n\u003cli\u003eChen, H., Zhang, W. \u0026amp; Sheng, L. Canadian record-breaking wildfires in 2023 and their impact on US air quality. \u003cem\u003eAtmospheric Environment\u003c/em\u003e \u003cstrong\u003e342\u003c/strong\u003e, 120941 (2025). \u003c/li\u003e\n\u003cli\u003eOlson, D. M.\u003cem\u003e et al.\u003c/em\u003e Terrestrial Ecoregions of the World: A New Map of Life on Earth: A new global map of terrestrial ecoregions provides an innovative tool for conserving biodiversity. \u003cem\u003eBioScience\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 933\u0026ndash;938 (2001). \u003c/li\u003e\n\u003cli\u003eHall, R.\u003cem\u003e et al.\u003c/em\u003e Generating annual estimates of forest fire disturbance in Canada: the National Burned Area Composite. \u003cem\u003eInternational journal of wildland fire\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 878\u0026ndash;891 (2020). \u003c/li\u003e\n\u003cli\u003eEidenshink, J.\u003cem\u003e et al.\u003c/em\u003e A project for monitoring trends in burn severity. \u003cem\u003eFire Ecol\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 3\u0026ndash;21 (2007). \u003c/li\u003e\n\u003cli\u003eSchmit, T. J.\u003cem\u003e et al.\u003c/em\u003e A closer look at the ABI on the GOES-R series. \u003cem\u003eBulletin of the American Meteorological Society\u003c/em\u003e \u003cstrong\u003e98\u003c/strong\u003e, 681\u0026ndash;698 (2017). \u003c/li\u003e\n\u003cli\u003eHall, J. V., Zhang, R., Schroeder, W., Huang, C. \u0026amp; Giglio, L. Validation of GOES-16 ABI and MSG SEVIRI active fire products. \u003cem\u003eInt J Appl Earth Obs\u003c/em\u003e \u003cstrong\u003e83\u003c/strong\u003e, 101928 (2019). \u003c/li\u003e\n\u003cli\u003eWooster, M.\u003cem\u003e et al.\u003c/em\u003e Meteosat SEVIRI Fire Radiative Power (FRP) products from the Land Surface Analysis Satellite Applications Facility (LSA SAF)--Part 1: Algorithms, product contents and analysis. \u003cem\u003eAtmospheric Chemistry \u0026amp; Physics Discussions\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e (2015). \u003c/li\u003e\n\u003cli\u003eWooster, M. J.\u003cem\u003e et al.\u003c/em\u003e Satellite remote sensing of active fires: History and current status, applications and future requirements. \u003cem\u003eRemote Sens Environ\u003c/em\u003e \u003cstrong\u003e267\u003c/strong\u003e, 112694 (2021). \u003c/li\u003e\n\u003cli\u003eHersbach, H.\u003cem\u003e et al.\u003c/em\u003e The ERA5 global reanalysis. \u003cem\u003eQuarterly Journal of the Royal Meteorological Society\u003c/em\u003e \u003cstrong\u003e146\u003c/strong\u003e, 1999\u0026ndash;2049 (2020). \u003c/li\u003e\n\u003cli\u003eWagner, C. V. \u003cem\u003eDevelopment and structure of the Canadian forest fire weather index system\u003c/em\u003e. (1987).\u003c/li\u003e\n\u003cli\u003eWang, X.\u003cem\u003e et al.\u003c/em\u003e cffdrs: an R package for the Canadian forest fire danger rating system. \u003cem\u003eEcological Processes\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 1\u0026ndash;11 (2017). \u003c/li\u003e\n\u003cli\u003eMcElhinny, M., Beckers, J. F., Hanes, C., Flannigan, M. \u0026amp; Jain, P. A high-resolution reanalysis of global fire weather from 1979 to 2018\u0026ndash;overwintering the Drought Code. \u003cem\u003eEarth System Science Data\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 1823\u0026ndash;1833 (2020). \u003c/li\u003e\n\u003cli\u003eCopernicus Climate Change Service, C. D. S., (2021): CMIP6 climate projections. Copernicus Climate Change Service (C3S) Climate Data Store (CDS). DOI: 10.24381/cds.c866074c.\u003c/li\u003e\n\u003cli\u003eMahony, C. R., Wang, T., Hamann, A. \u0026amp; Cannon, A. J. A global climate model ensemble for downscaled monthly climate normals over North America. \u003cem\u003eInternational Journal of Climatology\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, 5871\u0026ndash;5891 (2022). \u003c/li\u003e\n\u003cli\u003eJeong, D. I. \u0026amp; Cannon, A. J. An approach for selecting observationally-constrained global climate model ensembles for regional climate impacts and adaptation studies in Canada. \u003cem\u003eAtmosphere-Ocean\u003c/em\u003e \u003cstrong\u003e61\u003c/strong\u003e, 335\u0026ndash;351 (2023). \u003c/li\u003e\n\u003cli\u003eClarke, H.\u003cem\u003e et al.\u003c/em\u003e Gazing into the flames: A guide to assessing the impacts of climate change on landscape fire. \u003cem\u003eScience Advances\u003c/em\u003e\u003cstrong\u003e11\u003c/strong\u003e, eadz2429 (2025). \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8779558/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8779558/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Canada’s record-breaking 2023 wildfire season, followed by the severe 2024 and second-worst 2025 seasons, upended expectations of gradual boreal fire regime shifts. Together with recent Western U.S. extremes, these events raise a critical question: are they statistical outliers or the new reality in North America? Linking satellite observations with climate models to project future diurnal fire potential, we show that the climate-driven loss of diurnal firebreak can rapidly normalize the unprecedented flammability of Canada’s 2023 season by mid-century (2041–2070) even under the most ambitious mitigation scenario, while the Western U.S. 2020/2021 extremes approach typical conditions by late-century (2071-2100) but only under higher-warming scenarios. Mechanistically, we find a disproportionate escalation in overnight burning potential, concentrating fire opportunities into multi-day, round-the-clock runs when containment is least effective, and accelerating throughout the century across all warming scenarios. This shift is strongly amplified at high latitudes, with boreal overnight potential doubling to tripling by late-century under high warming, and it occurs synchronously across the entire boreal zone, challenging fire management strategies reliant on inter-regional resource sharing. The diurnally asymmetric, “locked-in” acceleration of northern fire risk requires urgent adaptation beyond current suppression capabilities.","manuscriptTitle":"Was Canada’s 2023 fire season a preview of things to come in North America?","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-25 15:36:56","doi":"10.21203/rs.3.rs-8779558/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-climate-change","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nclimate","sideBox":"Learn more about [Nature Climate Change](http://www.nature.com/nclimate/)","snPcode":"","submissionUrl":"","title":"Nature Climate Change","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c82d235e-24b1-461b-85f7-d3c77dba5af0","owner":[],"postedDate":"February 25th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":62890857,"name":"Earth and environmental sciences/Ecology/Fire ecology"},{"id":62890858,"name":"Earth and environmental sciences/Climate sciences/Climate change/Climate-change impacts/Environmental health"},{"id":62890859,"name":"Earth and environmental sciences/Climate sciences/Climate change/Projection and prediction"}],"tags":[],"updatedAt":"2026-02-25T15:36:56+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-25 15:36:56","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8779558","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8779558","identity":"rs-8779558","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}