Substantial increases in the likelihood of extreme fire weather events for fire-prone ecosystems in Australia

Substantial increases in the likelihood of extreme fire weather events for fire-prone ecosystems in Australia | 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 Article Substantial increases in the likelihood of extreme fire weather events for fire-prone ecosystems in Australia Ryan McGloin, Ralph Trancoso, Jozef Syktus, Rohan Eccles, Nathan Toombs, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8051769/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 13 You are reading this latest preprint version Abstract We evaluated the influence of climate change on fire weather in Australia using the McArthur Forest Fire Danger Index (FFDI) and climate simulations from an ensemble of dynamically downscaled CMIP6 projections. Rare extreme FFDI events were assessed under a range of Global Warming Levels (GWLs) using the Generalised Extreme Value (GEV) distribution, with a focus on southeast Australia’s wildfire prone eucalyptus forests. The magnitude and frequency of extreme FFDI events are projected to increase substantially, particularly in southern Australia. For the eucalyptus forests of southeast Australia, 20-year and 100-year return interval 7-day FFDI events (i.e. weekly average FFDI extremes) are projected to become approximately 2.1 and 3.0 times more likely under 3 ◦ C of global warming. The increases are most notable for the eucalyptus forests of Tasmania, with projected 20-year and 100-year return interval 7-day FFDI events becoming approximately 3.2 and 5.2 times more likely for 3 ◦ C GWL. Earth and environmental sciences/Climate sciences Biological sciences/Ecology Earth and environmental sciences/Ecology Earth and environmental sciences/Environmental sciences Earth and environmental sciences/Natural hazards Extreme fire weather FFDI Downscaled climate projections Regional climate modelling Eucalyptus forests Global Warming Levels Generalised Extreme Value distribution Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 1 Introduction In recent years, large wildfires in southeast (SE) Australia, South America, Canada, California, Siberia, and the Mediterranean region of Europe 1 – 5 have brought considerable attention to the threats wildfires pose to human life, public health, economies, ecosystems, and carbon reserves. There is also a growing interest in the role of climate change in driving the frequency and intensity of these events 6 – 10 . Due to the proximity of densely populated areas to highly flammable ecosystems with large fuel loads, SE Australia has historically been particularly vulnerable to catastrophic wildfires that have caused widespread fatalities and property damage 11 – 13 . For example, during the 2019/20 “Black Summer” fire season, over 23% of temperate forests in SE Australia were burned, an event of unprecedented scale both nationally and globally 14 . The disaster resulted in 33 lives lost and the destruction of 3,000 homes 15 . While the exact economic impact of the disaster is difficult to determine, some estimates suggest the cost could have been as high as AUD $ 100 billion 16 , 17 . Numerous studies have shown that global warming is already driving an increase in the frequency of wildfire events across various regions worldwide 1 , 6 , 18 , 19 , including Australia 10 , 20 – 22 , while extensive research also indicates that the frequency and intensity of fire weather events will accelerate further as the century unfolds 23 – 29 . Wildfires occur when three key factors align: a supply of fuel, atmospheric conditions sufficiently conducive to fire ignition and spread, and an ignition source 6 . While making projections of future fuel loads and ignition sources is complex and challenging, fire weather can be represented using an index such as the McArthur Forest Fire Danger Index (FFDI) 30 or Canadian Forest Fire Weather Index (FWI) 31 , which use air temperature, humidity, wind speed and precipitation to estimate the influence of weather on fire behaviour. While these indexes do not perfectly capture all elements that determine the intensity and impacts of wildfire events, they can be considered a useful way to combine several weather factors known to influence fire behaviour and have been shown to correlate well with attributes such as total area burned and house losses 11 , 13 , 21 . Global Climate Models (GCMs) are valuable tools for exploring how global warming may impact climate hazards such as fire weather. However, GCMs operate at coarse spatial resolutions (150–200 km), which limits their ability to accurately incorporate the influence of landscape features and capture regional-scale fire weather processes 18 . To overcome this limitation, dynamical downscaling has been employed to improve the resolution of GCM outputs, including using downscaled outputs from the 3rd and 5th phases of the Coupled Model Intercomparison Project (CMIP3 and CMIP5) to assess changes to future fire weather regimes in Australia 23 , 24 , 32 . These studies project a clear trend towards more dangerous near-surface fire weather conditions for Australia with the greatest increases occurring in spring. Categories of FFDI can be used to help specify the fire weather intensity on a given day (e.g., days with an FFDI above 50 are classified as “Severe” fire weather days by fire management agencies for communication purposes 33 ). For meaningful categorisation of FFDI calculated using downscaled GCM outputs, it is important that any biases in the underlying model data are corrected. Typically, bias correction is performed on variables that have temporally and spatially comprehensive observational records, such as temperature and precipitation. The absence of suitable gridded datasets for relative humidity and wind speed throughout Australia has historically posed a significant challenge for bias correction of FFDI, with partial bias correction (i.e. only for maximum temperature and precipitation) being the most common approach 34 , 35 . Research has shown that partial bias correction of FFDI does not consistently enhance model performance 34 , highlighting the need for robust datasets of relative humidity and wind speed to enable thorough bias correction of FFDI. Dowdy et al. 22 used an alternative approach to partial bias correction, where gridded observations of temperature, precipitation and relative humidity (derived from vapour pressure) were used in conjunction with reanalysis (NCEP–NCAR) wind speeds that were bias corrected to the Bureau of Meteorology (BoM) wind speed forecasts used to provide FFDI forecasts to fire management agencies throughout Australia. The approach resulted in daily gridded FFDI values similar in magnitude to what fire managers use for peak daily values. While it is useful to examine projected future changes in the annual number of days that fall in particular FFDI categories 21 , 36 , there is a need to focus on rarer extremes, as the majority of deaths and property destruction tend to occur during rare extreme fire weather events 37 . Studies have found that the majority of wild fire related house losses in Australia occurred on days with very rare catastrophic fire weather conditions (days with FFDI above 100) with nearly all losses occurring on days where the FFDI was above the 99.5th percentile 11 . It is also useful to analyse rare fire weather events in regions where FFDI values of moderate magnitude may still indicate dangerous fire weather, even if those same values are common and not considered dangerous in other parts of Australia 22 . Extreme values can be assessed using comparatively modest thresholds, such as the 90th, 95th, and 99th percentiles as well as rarer conditions as represented by return intervals. Return intervals can be employed to quantify changes in the frequency and duration of extreme wildfire events under global warming 26 . In the Australian context, return intervals have been used to examine the historical climatological variability of fire weather and attribute Australian wildfire risk to anthropogenic climate change 10 , 22 . Previous studies of fire weather projections in Australia have been based on older CMIP generations and have tended to use small ensembles. Therefore, there is a need to quantify future changes in extreme fire weather using the latest downscaled CMIP6 projections under different levels of global warming with larger ensembles to better account for uncertainty. Analysing future changes under global warming levels (GWLs) presents advantages over analysis using emissions scenarios due to GWLs being directly tied to international climate policy goals, such as the Paris Agreement 38 and not being tied to specific emission scenarios. While bias correction of FFDI is not required for all applications, it is essential when using FFDI thresholds to categorise fire weather intensity. Previous attempts at bias correction of FFDI have largely relied on partial bias correction methods, however, there is a need for comprehensive bias correction using high-quality datasets for all input variables. Finally, there is a particular need to focus on changes in fire weather regimes and extremes in the eucalypt forests of SE Australia, which are especially vulnerable to intense and destructive wildfires 12 , 21 . This study has three key aims which are designed to address the gaps in our understanding of fire weather in Australia and to help inform planning to build resilience: to evaluate the datasets available for bias-correcting FFDI input variables and to determine whether bias correction improves FFDI results in Australia to assess how fire weather is projected to change across Australia under different GWLs using an ensemble of dynamically downscaled CMIP6 projections to analyse projected changes in the intensity and frequency of extreme fire weather events (as represented by return intervals) across Australia, with a focused case study on the eucalypt forests of SE Australia. 2 Results 2.1 Bias correction 2.1.1 Calibration data An assessment of the similarity of two reanalysis-based datasets (MSWX-Past and BARRA2) and station observations was performed using Perkins skill scores in order to establish the most suitable dataset to be used as the calibration dataset in the bias correction process. Time series of daily relative humidity, maximum air temperature and wind speed and daily and monthly precipitation for the 1980–2020 period were extracted from the MSWX-Past and BARRA2 datasets for each station location. Note that monthly precipitation was included in this analysis as the Keetch Byram Drought Index (KBDI), used in the FFDI calculation process, is influenced both by the precipitation that falls on a given day and the cumulative effects of precipitation over the preceding 20 days, particularly during dry spells. Perkins skill scores for the entire distribution of the climate variables (Fig. 1 and Supplementary Material Figure S1 ) showed that BARRA2 performed better across most regions for the maximum temperature and wind variables. BARRA2 and MSWX-Past performed similarly for relative humidity, while MSWX-Past outperformed BARRA2 at most locations for both daily and monthly precipitation (Figure S1 ). BARRA2 tended to outperform MSWX-Past for most variables and locations when calculating the Perkins skill scores for the lower and upper tails of the distributions (Supplementary Materials Figures S1 , S2 and S3), particularly for the extreme values most relevant for fire weather (i.e. upper tails for maximum temperature and wind speed and lower tails for relative humidity and precipitation). Therefore, BARRA2 was used as the calibration dataset in the bias correction process used in this study. 2.1.2 Bias corrected data Perkins skill score results were also used to compare how well the simulations bias corrected using the BARRA2 calibration dataset and non-bias corrected historical (1981–2017) simulations of FFDI from the downscaled CMIP6 ensemble performed against FFDI calculated from observations at 39 BoM weather stations. Bias correction improved Perkins skill score results at most locations and across all seasons and percentiles (Figs. 2 and 3 ). The overall effect of the bias correction process was to reduce the magnitude of FFDI over most of Australia, particularly in Western Australia (Supplementary Material Figure S4). The exception was along the north and north-eastern coastline of Australia where the process of bias correction tended to increase FFDI slightly. The effect is important when looking at categories of FFDI using thresholds, for example, bias correcting the downscaled simulations reduced the projected average annual count of days with FFDI ≥ 50 for all of Australia by nearly 50% for both 3 ◦ C GWL and 4 ◦ C GWL (Supplementary Material Figure S4). All results presented in the following sections of this study were calculated using the bias corrected downscaled CMIP6 ensemble. The main cause of bias in the non-bias corrected FFDI is due to the Conformal Cubic Atmospheric Model (CCAM, the model used to perform the dynamical downscaling) overestimating maximum temperature and underestimating humidity when conditions are hot and dry for large parts of Australia (Fig. 4 ). Note that both BARRA2 and Australian Gridded Climate Data (AGCD) show similar results in this regard (Supplementary Material Figure S5). CCAM also tends to overestimate wind speeds in the east and north of the country, while winds in the south, west and central areas tend to be underestimated (Fig. 4 and Supplementary Material Figure S5). 2.2 FFDI severity categories Severe fire weather days (FFDI ≥ 50) are projected to increase over much of Australia with the magnitude of change increasing with the level of global warming. Increases are particularly apparent in northwestern and central Australia under 3 ◦ C and 4 ◦ C of global warming. Very High fire weather days (24 ≤ FFDI < 50) are projected to increase in the north and south of the country (Fig. 5 ). As with Severe fire weather days, noticeable increases in the number of days with FFDI above the 95th and 99.726th (equivalent to 1-in-1-year event) percentiles are projected for central and western Australia, however, other regions with distinct increases are also projected in Tasmania and some elevated regions close to the eastern coast. The greatest increases in the number of Severe fire weather days for central and western Australia and Very High fire weather days for southern Australia are projected in spring (Supplementary Material Figures S7-S10), while the greatest increases in the number of Very High fire weather days in northern Australia are projected in autumn and winter (Supplementary Material Figures S7-S10). 2.3 Extreme fire weather events in Australia The magnitude of both 7-day running mean (from here on referred to as just 7-day) and daily rare extreme FFDI events are projected to increase for most of Australia for all warming levels, with the magnitude of change increasing as the global climate warms (Fig. 6 and Supplementary Material Figures S11, S12 and S13). Increases in the magnitude of the 7-day extreme events are more noticeable in southern and central areas of Australia than in the north for the 3 ◦ C and 4 ◦ C GWLs, with prominent patches of high increases in northern NSW, southeastern Victoria and Tasmania (Fig. 6 ). Results for the daily FFDI extreme events are noisier, however, patches of large increases in northern NSW, southeastern Victoria and Tasmania are also apparent for the 3 ◦ C and 4 ◦ C GWLs (Supplementary Material Figure S11). The frequency of both 7-day and daily rare extreme FFDI events (compared to 1.2 ◦ C of global warming) are also projected to increase for all of Australia and for all warming levels, with greater increases projected for the rarest extremes (Fig. 7 , Fig. 8 and Supplementary Material Figures S14 and S15). For the 7-day extreme events, there are well defined areas of increase in central and western areas (Fig. 7 ), with nosier results once again for the daily FFDI extreme events (Supplementary Material Figure S14). On average across Australia using the ensemble median, the 20-year and 100-year return interval 7-day FFDI events are projected to be 1.7 and 2.5 times more likely at 2 ◦ C GWL, and 2.7 and 4.8 times more likely at 3 ◦ C GWL, respectively. 2.4 Extreme fire weather events in the eucalyptus forests of southeastern Australia Projections of extreme fire weather across all of Australia tend to draw the attention to the dry, sparsely vegetated regions of central and western Australia and away from SE Australia, a region with considerable coverage of dense, highly flammable forest ecosystems (Fig. 12 b). The relative increase in magnitude of both 7-day and daily rare extreme FFDI events is projected to be substantial for eucalypt forests in SE Australia (Fig. 9 and Supplementary Material Figures S16, S17 and S18), with increases in the ensemble median magnitude for 20-year and 100-year return interval 7-day FFDI events averaged for all of SE Australia’s eucalypt forests of 6.4% and 6.7% for 2 ◦ C GWL; and 10.0% and 9.2% for 3 ◦ C GWL. Particularly high increases are projected in Tasmania, where ensemble median increases in the magnitude of 20-year and 100-year return interval 7-day FFDI events are 12.1% and 11.5% for 2 ◦ C GWL, and 22.9% and 23.1% for 3 ◦ C GWL (Supplementary Material Table S1 ). The lowest increases are projected for Queensland’s eucalyptus forests with increases for 20-year and 100-year return interval 7-day FFDI events of 4.2% and 4.7% for 2 ◦ C GWL and 5.0% and 4.0% for 3 ◦ C GWL. Similar results are seen for projected increases in the magnitude of rare extreme daily FFDI events (Supplementary Material Table S2). Increases in the likelihood of both 7-day and daily rare extreme FFDI events (compared to 1.2 ◦ C of global warming) are greatest in inland areas. However, substantial increases are also projected for the eucalypt forests of SE Australia, particularly in Tasmania (Figs. 10 and 11 and Supplementary Material Figures S19 and S20), where ensemble median 20-year and 100-year return interval 7-day FFDI events increased by 2 and 2.6 times for 2 ◦ C GWL, and 3.2 and 5.2 times for 3 ◦ C GWL (Supplementary Material Table S3). As with the magnitude of the rare extreme events, lowest increases in likelihood are projected for Queensland’s eucalyptus forests with increases for 20-year and 100-year return interval 7-day FFDI events of 1.5 and 2.1 times for 2 ◦ C GWL, and 1.8 and 2.2 times for 3 ◦ C GWL, respectively. Similar results are seen for projected increases in the likelihood of rare extreme daily FFDI events (Supplementary Material Table S4). When averaged over all eucalypt forests in SE Australia, 20-year and 100-year return interval 7-day FFDI events are projected to become approximately 1.6 and 2.3 times more likely for 2 ◦ C GWL, and 2.1 and 3 times more likely for 3 ◦ C GWL (Fig. 10 and Supplementary Material Table S4). Increases in both the magnitude and likelihood of rare extreme FFDI events for eucalypt forests in Victoria and New South Wales are similar to the results for all of SE Australia, with increases tending to be slightly higher in Victoria (Supplementary Material Tables S1 - S4). It should be noted that there was generally high uncertainty associated with the ensemble median projected changes in the magnitude and frequency of the rarest extreme events (50-year and 100-year return-interval events) and for the lower GWLs, with few regions showing significant changes following our signal-to-noise ratio analysis. However, there are substantial areas of low uncertainty for the higher GWLs and the more moderate rare extreme events (5-year, 10-year and 20-year return-interval events). Note there are also patches of low uncertainty for the rarest extreme events under high levels of global warming, particularly for the 7-day FFDI events, such as in Victoria and Tasmania. The spread in the ensemble results is shown in detail in the box plots presented in Fig. 11 and Supplementary Material Figures S17, S18 and S20, and in the difference between the 10th and 90th percentile maps shown in Supplementary Material Figures S21 and S22. 3 Discussion The results from this study indicate that BARRA2 is generally more appropriate than MSWX-Past for bias correcting FFDI inputs from downscaled GCMs in Australia. This is logical considering that BARRA2 is a reanalysis product specifically developed for the Australian region 39 . Other studies have also shown that BARRA2 outperforms other reanalysis products when simulating climate variables in the Australian region, such as wind power 40 . However, the results from this study suggest that MSWX-Past may be a more appropriate calibration source for bias correction if looking at precipitation specific indices such as extreme rainfall. Previous studies have typically relied on partial bias correction to bias correct FFDI and these studies have found that partial bias correction of FFDI does not consistently improve model performance and can lead to a greater spread in model simulations, particularly for extreme values 34 . Results from this study suggest that there are benefits to bias correcting each FFDI input variable separately using the Quantile Matching for Extremes (QME) method and BARRA2 data as calibration data, resulting in improved performance with historical FFDI values calculated using observations. However, the best approach for bias correction of FFDI may be dependent on the application. For example, a previous study 22 that bias corrected wind data to match the BoM operational forecast wind speeds used for FFDI forecasts noted that the FFDI values tended to be higher than the station-based FFDI 41 used to assess the impact of bias correction in this study. Ultimately, while improvements have been made in recent years, there are still limitations associated with the available meteorological datasets required for complete bias correction of FFDI. This is particularly the case for wind speed given the relatively limited number of stations that have wind data of suitable quality and issues associated with spatial interpolation using sparse datasets 42 , 43 and also because FFDI was designed to use 10-minute averages of wind speed 44 rather than the instantaneous values provided by most climate models and reanalysis products 39 . It should be noted that bias correction in this study only substantially affected results when FFDI thresholds were used to categorise fire weather intensity (e.g. days with FFDI ≥ 50), with bias corrected vs non-bias corrected results being very similar when analysing changes in percentiles and return intervals (please compare Fig. 10 and Supplementary Materials Figure S23 for an illustration of this). This study found that Severe fire weather (FFDI ≥ 50) days are projected to increase substantially in north-western and central Australia, while Very High fire weather days (24 ≤ FFDI < 50) are projected to increase in the north and south of the country. Interpretation of the increases in Severe fire weather days for central and western Australia should be done with caution as these regions are not typically dominated by the forested ecosystems for which the FFDI was designed 30 and are far from major population centres where wildfires are likely to have the greatest impact. In contrast, the region in southern Australia where there are projected increases in Very High fire weather conditions (including the areas surrounding Adelaide and southwestern Victoria) is an area that has been greatly affected by devastating wildfires in the past, such as the Ash Wednesday and Black Friday events. The increases in Very High fire weather conditions in northern Australia could also have implications due to the high number of fires that already occur in the regions’ tropical savannas, which account for approximately 68% of the national fire extent annually 13 , 45 . Note that for more accurate simulation of savanna fire weather conditions, the Grassland Fire Danger Index (GFDI) 46 is recommended instead of the FFDI presented here, noting that GFDI includes factors such as grass curing that have substantial uncertainties when estimated using climate model data. The largest increases in Severe and Very High fire weather category days were projected in spring and autumn, this agrees with findings from other studies in Australia. Both observations and projections show a trend of the largest increases in fire weather conditions occurring in spring 22 , 23 , 41 , while an exponential increasing trend in total burned forest area during autumn and winter has been suggested 21 . Therefore, the findings of this and other studies 47 indicate an increasing trend in fire season length across Australia. The magnitude and frequency of extreme FFDI events were projected to increase substantially in the eucalypt forests of SE Australia. Observational studies have shown that this is a region that is already experiencing large increases in both extreme fire weather conditions and the area burned during wildfires 21 , 47 , 48 . Recent increases in fire weather magnitude in this region are associated with increases in temperature and decreases in precipitation and humidity linked to climate change 19 , 47 . Shifts in large-scale teleconnection modes likely played a role in these changes with evidence suggesting that fire promoting phases of tropical Pacific and Indian Ocean variability (+ ENSO, +IOD, and/or − SAM) are becoming more frequent 20 . In addition, an increase in anticyclonic blocking events may also have contributed to more frequent days with coinciding low relative humidity and high wind speeds 47 . The particularly high increases in the magnitude and frequency of rare extreme events projected for Tasmania in this study complements the findings of observation-based studies which have shown that Tasmania has seen particularly large increases in the area burned during wildfires over the past 30 years compared to other regions in Australia 21 . Tasmania has experienced several devastating wildfires in the past. For example, 62 people were killed and nearly 3,000 structures were destroyed in southern Tasmania during the 1967 Black Tuesday wildfires, while in January 2013, wildfires caused widespread destruction of infrastructure, including 203 homes, in the village of Dunalle 49 . Tasmanian vegetation is characterised by mosaics of fire-dependent and fire sensitive vegetation types, where large areas of flammable eucalyptus forests surround more fire-sensitive vegetation types such as wet and dry rainforests and alpine shrublands 49 . Fire frequency has been found to be the dominant driver of future fire activity in Tasmania because of frequent fire leading to shifts in vegetation type away from fire-sensitive types towards drier, flammable, fire-adapted vegetation 50 . This suggests that impacts from the projected increases in frequency of extreme fire weather conditions in Tasmania could be amplified by a shift in vegetation towards more flammable fire-adapted vegetation types. This study represents one of the few that has analysed future downscaled simulations of rare extreme fire weather events in Australia and the only study, that these authors are aware of, that has used the latest CMIP6 projections under different levels of global warming to analyse rare extremes for SE Australia’s vulnerable eucalypt forests. These results complement the findings of previous studies that have generally found a clear trend towards more dangerous fire weather conditions for SE Australia 23 , 24 , 32 . This includes a previous study using ensembles of bias corrected downscaled projections that found increases in extreme FFDI that scaled approximately linearly with GWL 51 , while noting that was based on the older set of CMIP5 GCMs. It also accompanies other studies from around the world that have identified projected increases in the frequency and magnitude of rare wildfire extremes in fire prone regions in North America and Europe 18 , 26 and others that have analysed the impacts of non-global warming related factors on extreme fire weather, such as deforestation in Borneo 52 . There would be value in repeating the analysis of this study to consider multiple ensembles of projections derived from multiple downscaling methodologies to account for uncertainty arising from using downscaled projections from a single modelling group 53 , 54 . Additional sources of uncertainty include natural variability, scenario uncertainty, and epistemic uncertainty 55 which have not been explicitly considered in this study. Natural variability has been shown to be the largest source of uncertainty in near-term projections of climate extremes globally 54 , including over Australia which is subject to very high inter-annual variability. There is a degree of epistemic uncertainty inherent in climate models, while uncertainty is also introduced into this study via the estimates of extremes, particularly for the rarest extremes 56 , 57 . These are the result of fitting distributions to a relatively limited dataset (20 years), though the grid cell pooling approach adopted in this study may go some way to helping reduce this 58 . It should also be highlighted that the projected changes of FFDI presented in this study solely represent a measure of how conducive the atmospheric conditions are for wildfire ignition and spread and are not projections of wildfire risk, as this would require information on ecological and human factors beyond the reliability of climate projections 27 . To conclude, we assessed the impact of climate change on fire weather in Australia using high-resolution climate simulations from an ensemble of downscaled CMIP6 GCMs. The study combines dynamical downscaling, GEV distribution and GWL analysis to better understand the potential impacts of climate change on rare extreme fire weather events across Australia, including the vulnerable eucalypt forests in the southeast of the country. It was found that bias correcting each of the FFDI input variables using a BARRA2 calibration dataset resulted in improvements over non-bias corrected data when compared with FFDI calculated using historical observations. Severe fire weather (FFDI ≥ 50) days are projected to increase considerably in north-western and central Australia under 3 ◦ C and 4 ◦ C of global warming, while Very High fire weather days (24 ≤ FFDI < 50) are projected to increase in the north and south of the country. The magnitude and frequency of extreme FFDI events are expected to increase substantially, particularly in the southern half of Australia, with 20-year and 100-year return interval 7-day FFDI events averaged for all of Australia projected to become approximately 1.7 and 2.5 times more likely for 2 ◦ C GWL, and 2.7 and 4.8 times more likely for 3 ◦ C GWL. For the wildfire prone eucalyptus ecosystems of SE Australia, the increases are most notable for Tasmania with 20-year and 100-year return interval 7-day FFDI events projected to become approximately 2 and 2.6 times more likely for 2 ◦ C GWL, and 3.2 and 5.2 times more likely for 3 ◦ C GWL. Queensland’s eucalyptus forests had the lowest projected increases in frequency; however, the changes were not immaterial, with the likelihood of 20-year and 100-year return interval 7-day FFDI events increasing by approximately 1.5 and 2.1 times for 2 ◦ C GWL, and 1.8 and 2.2 times for 3 ◦ C GWL, respectively. 4 Methodology 4.1 Study Area This study evaluated changes to fire weather across the Australian continent, which encompasses a range of climate regions, including arid, equatorial, grassland, sub-tropical, temperate, and tropical regions (Fig. 12 a). To support this analysis, observational datasets from selected weather stations across Australia were used to validate the data inputs and outputs of the FFDI bias correction process (see section 4.2.2 for details). This study includes a case study focussed on SE Australia, a region with considerable coverage of dense, highly flammable forest ecosystems and that has experienced several large-scale wildfires with catastrophic impacts during the last century (Fig. 12 b) 4.2 Data 4.2.1 Modelling We used an ensemble of 15 downscaled climate model simulations derived from 11 different CMIP6 GCMs (Table 1 ), with some model variants downscaled multiple times. The ensemble consists of historical simulations up to 2014 and future simulations under a high emissions pathway (SSP370) up to 2100. The SSP370 pathway was selected as this was the highest emissions pathway available and enabled analysis up to 4 ◦ C of global warming. The GCMs included in this study were selected to best represent the future spread in the climate change signals from the GCMs, while prioritising models that performed well in simulating the Australian climate 61 , 62 . Dynamical downscaling was performed for the full ensemble of GCMs using CCAM 63 developed by CSIRO 64 . CCAM was run using a stretched C288 grid, providing a model resolution of approximately 10 km for Australia. The downscaling approach used involved correcting for model biases in sea surface temperatures and sea ice 65 , 66 . This approach has been found to improve the simulations of climate from CCAM and other regional climate models 65 , 67 , 68 . Five of the CCAM simulations were run using dynamic atmosphere-ocean coupling as shown in Table 1 . The downscaling approach adopted has been shown to significantly improve the performance over the host GCMs for precipitation and temperature in all seasons, with the largest improvements noted for climate extremes 61 . Table 1 Details of the 15 climate model simulations downscaled from 11 CMIP6 GCMs considered in this study. CMIP6 Model Model full name Resolution Ensemble member CCAM setup ACCESS-ESM1.5 Australian Community Climate and Earth System Simulator, version 1.5 1.875 x 1.25° r6i1p1f1 atmospheric r20i1p1f1 atm-ocean coupled r40i1p1f1 atm-ocean coupled ACCESS_CM2 Australian Community Climate and Earth System Simulator, version 2 1.875 x 1.25° r2i1p1f1 atm-ocean coupled CMCC-ESM2 Centro Euro-Mediterraneo sui Cambiamenti Climatici 0.9 x 1.25° r1i1p1f1 atmospheric CNRM-CM6-1-HR Centre National de Recherches Météorologiques Coupled Global Climate Model, version 6.1, high-resolution 0.5 x 0.5° r1i1p1f2 atmospheric r1i1p1f2 atm-ocean coupled EC-Earth3 European Community Earth-System Model, version 3 0.8 x 0.8° r1i1p1f1 atmospheric FGOALS-g3 Flexible Global Ocean-Atmosphere-Land System Model, grid point version 3 2.5 x 2.5° R4i1p1f1 atmospheric GFDL-ESM4 Geophysical Fluid Dynamics Laboratory Earth System Model, version 4 1 x 1° r1i1p1f1 atmospheric GISS-E2-2-G Goddard Institute for Space Studies Model E2.2G 2. x 2.5° r2i1p1f2 atmospheric MPI-ESM1-2-LR Max Planck Institute Earth System Model, version 1.2, low resolution 1.9 x 1.9° r9i1p1f1 atmospheric MRI-ESM2-0 Meteorological Research Institute Earth System Model, version 2.0 1.125 x 1.125° r1i1p1f1 atmospheric NorESM2-MM Norwegian Earth System Model, version 2, 1 degree resolution 1 x 1° r1i1p1f1 atmospheric r1i1p1f1 atm-ocean coupled 4.2.2 Reanalysis and observations Bias correction of FFDI when using downscaled simulations as input variables requires a calibration dataset of gridded meteorological variables at an appropriate resolution and temporal scale. For FFDI, this necessitates sub-daily wind speed and humidity data, as the index is typically calculated using mid-afternoon (e.g., 3 pm) or daily maximum measurements of these variables (see section 4.5). However, this has historically posed a challenge, as sub-daily observations, particularly for wind speed, are sparse in Australia 34 . To address this limitation, high resolution reanalysis datasets can be used in place of observations, providing long-term, spatially complete records of historical climate variables. Two reanalysis-based products were identified as suitable for this purpose, offering high resolution, sub-daily data across Australia for all required FFDI input variables: The Multi-Source Weather product (MSWX-Past) 69 , and Bureau of Meteorology Atmospheric high-resolution Regional Reanalysis for Australia (BARRA2) 39 MSWX is a global gridded meteorological product featuring 3-hourly 0.1° resolution forecasts and historical records. The historical part of the record (MSWX- Past) starts from 1 January 1979 and is based on ERA5 reanalysis data bias corrected and downscaled using high-resolution reference climatologies. The reference climatologies are based on data derived from station observations, satellite imagery, and/or model output, depending on the variable 69 . BARRA2 is the second generation of Australian reanalysis and covers the Australian continent and the surrounding region. BARRA2 reanalysis is produced by running the limited-area Australian Community Climate and Earth-System Simulator (ACCESS) model 70 . The ACCESS model is forced with ERA5 reanalyses boundary conditions and adjusted to better reflect observations via data assimilation 39 . Both datasets were re-gridded to match the resolution of the downscaled climate simulations using distance-weighted interpolation for precipitation and bilinear interpolation for all other variables. These datasets were evaluated against station observations in section 2.1.1 of this study, with the best performing dataset selected for use as the calibration dataset in the bias correction process. Observational datasets from selected weather stations (Fig. 12 a) were used to validate both the reanalysis data used to bias correct the FFDI input variables and the final bias corrected FFDI data. Daily timeseries of mean relative humidity (%), wind speed (km/hr), precipitation (mm) and daily maximum surface air temperature (°C) were extracted for the 1980–2020 period from Australian BoM stations, which are marked in blue and black in Fig. 12 a. Stations were selected based on whether all four input variables were available and according to the length and completeness of their data records. Note that ideally 3 pm (Local Time: LT) relative humidity and wind speed would have been used instead of daily values, since the FFDI calculation and bias correction use the 3 pm variables (see section 4.5). However, sub-daily relative humidity and wind speed were not available for the BoM stations and therefore, we decided that the product that best simulated the daily relative humidity and wind speed was also highly likely to be superior in simulating the sub daily situation. We also used a historical dataset of FFDI 41 , 71 calculated from observations from 39 high quality BoM stations (red and black stations in Fig. 12 a) to validate the final bias corrected FFDI data. This dataset consists of 97th, 95th, 90th, 75th and 50th percentiles of daily FFDI estimated over standard meteorological seasons from 1981–2017. 4.3 Bias correction Bias correction is a common tool used to correct climate model data and involves removing biases in model outputs by calibrating them against reliable observational data 72 . After establishing the reanalysis dataset that showed the best agreement with station observations (see section 4.2.2), the downscaled FFDI input variables were bias corrected using the QME method 73 , which applies quantile-quantile matching to the input data histograms. A calibration period of 1981–2020 was used in this process. Correction functions derived from the calibration period were then applied to future climate projections for each climate model on an individual grid cell basis. 4.4 Evaluation We evaluated the performance of the reanalysis-based datasets used to bias correct the FFDI and the resulting bias corrected FFDI using the Perkins skill score 74 . The Perkins skill score evaluates data based on similarity between the modelled (in this case reanalysis data and modelled FFDI data) and observed probability density functions (PDFs). The binning of data to construct histograms was based on the distribution of the observed data. The Perkins skill score was then calculated as follows: $$\:{S}_{score}=\sum\:_{1}^{n}minimum({Z}_{m},{Z}_{0})$$ where n is the number of bins used to calculate the histogram, Z m , is the frequency of values in each bin for the model, and Z o is the frequency of values in each bin for the observations. If a model simulates the observed PDF poorly, the Perkins skill score will be close to zero, while if the PDF is well simulated the score will approach a maximum of 1. The Perkins skill score was used to assess the entire distribution of the climate variables as well as the lower and upper tails of the distributions (0–5th and 95–100th percentiles). Stippling on the maps presented in this study represents where the signal-to-noise ratio was greater than 1, indicating where the climate change signal from the ensemble emerges from the noise of the ensemble. Here, the signal is represented by either the ensemble mean or median change and the noise is the standard deviation from all 15 models in the ensemble as in Chapman et al. 66 . 4.5 FFDI The McArthur FFDI was calculated following 75 : $$\:FFDI=2exp\left(0.0338T+0.0234W-0.0345RH+0.987\text{ln}\left(DF\right)-0.45\right)$$ where DF is the drought factor, T the daily surface air temperature (°C), W the wind speed at a height of 10 m (km h − 1 ) and RH the relative humidity (%). In this study we have followed the common approach of using the maximum daily temperature as T , and 3 pm values of wind speed and relative humidity as W and RH (%) to represent mid-afternoon FFDI, when values are typically near the daily maximum 76 . The drought factor is an estimate of fuel dryness and is computed using the KBDI 77 and a function to expresses the influence of the past precipitation amount and the time since it fell 78 . Daily bias corrected and non-bias corrected FFDI were calculated for the downscaled CMIP6 ensemble with projected changes in FFDI presented for different GWLs. GWLs were derived from CCAM using an 11-year running average of global surface temperatures 79 . Warming levels were calculated relative to 1995–2014 with an offset of 0.85°C 80 added to derive the warming from preindustrial levels. Twenty-year time slices, centred on the target GWL (midpoint − 9 years to midpoint + 10 years), were used to extract the FFDI data for that GWL. Changes to future fire weather were assessed in terms of counts of days within specific FFDI categories (i.e. Severe, defined as FFDI > 50, Very High, defined as 24 ≤ FFDI < 50, and High, defined as 12 ≤ FFDI < 24), percentiles (i.e. FFDI greater or equal to the 95th percentile and 99.726th percentile) and return intervals (see section 4.6). 4.6 Extreme value analysis Extreme value analysis was applied to assess changes to the probability distribution of rare extreme events. Here these events represent the 1-in-5, 1-in-10, 1-in-20, 1-in-50, and 1-in-100 Annual Exceedance Probabilities (AEPs) which approximately correspond to events with annual return intervals of 5, 10, 20, 50, and 100 years respectively. We sampled the daily timeseries of FFDI at each grid cell using the block maxima approach to derive annual maxima (AM) and then pooled together data from nearby cells using a 5x5 box centred on each grid cell to extend the data series used for the extreme event analysis as per Eccles et al. 57 . Typically, annual maximum daily values are used when using the block maxima approach, however, some studies have found that 7-day running mean FWI (which typically behaves similarly to FFDI 33 ) has shown a good correlation with area burned during wildfires, including during the 2019/20 “Black Summer” fire season in Australia 10 , 81 . Therefore, in this study, AM series were calculated using both the annual maximum daily FFDI and annual maximum 7-day running mean FFDI. The Generalised Extreme Value (GEV) distribution was then fitted to the AM series using the l-moments method for parameter estimation. The GEV distribution is a generalised expression combining the Gumbel, Fréchet, and the Weibull distributions and is given by: $$\:G\left(x\right)=exp\left\{-{\left[1+\xi\:\left(\frac{x-\mu\:}{\sigma\:}\right)\right]}^{-1/\xi\:}\right\},\:for\:\left\{x:1+\xi\:\left(\frac{x-\mu\:}{\sigma\:}\right)>0\right\}$$ where, µ , σ , and ξ are the location, scale, and shape parameters, respectively. Here, the location parameter is a measure of the central tendency and is loosely linked to the mean, the scale parameter is a measure of variance, and the shape parameter describes the tail behaviour. We fit the GEV distribution to each 20-year time slice of FFDI AM. These distributions were then used to determine the FFDI values corresponding to particular return intervals. Changes in the magnitude of return interval events were calculated relative to the values under 1.2°C of global warming. We also present results of changes in the frequency of particular return interval events (with the magnitude of the events taken from the 1.2°C GWL). Declarations Competing interests All authors declare no financial or non-financial competing interests. Funding Declaration This research did not receive funding. Author Contribution R.M.: Writing – Original draft preparation, Conceptualization, Methodology, Formal analysis. R.T.: Conceptualization, Methodology, Writing. - Review & Editing. J.S.: Data Curation, Methodology, Writing. - Review & Editing. R.E.: Conceptualization, Methodology, Writing - Review & Editing. N.T.: Data Curation, Methodology. A.D.: Conceptualization, Writing - Review & Editing. Acknowledgement We acknowledge the World Climate Research Programme, which, through its Working Group on Coupled Modelling, coordinated and promoted CMIP6. We thank the climate modelling groups for producing and making available their model output, the Earth System Grid Federation (ESGF) for archiving the data and providing access, and the multiple funding agencies who support CMIP6 and ESGF. Data are available through the ESGF at: http://esgf.llnl.gov/. We also acknowledge Lindsay Brebber from Information and Digital Science Delivery at Queensland Government for support with high performance computing and data storage. Data Availability All data used in this study are publicly available. The downscaled climate projections that are used in the FFDI calculations can be accessed through the National Computer Infrastructure at: [https://dx.doi.org/10.25914/8fve-1910](https:/dx.doi.org/10.25914/8fve-1910) . The CMIP6 global climate model data are available through the Earth System Grid Federation at: [http://esgf.llnl.gov/](http:/esgf.llnl.gov) . Gridded BARRA2 reanalysis data can be accessed through the National Computer Infrastructure at: [https://dx.doi.org/10.25914/90rq-d839](https:/dx.doi.org/10.25914/90rq-d839) , while MSWX data can be accessed via request at: [https://www.gloh2o.org/mswx/](https:/www.gloh2o.org/mswx) . Seasonal McArthur Forest Fire Danger Index data for Australia41 is available from [https://data.mendeley.com/datasets/xf5bv3hcvw/1](https:/data.mendeley.com/datasets/xf5bv3hcvw/1) . References Jain, P. et al. Drivers and Impacts of the Record-Breaking 2023 Wildfire Season in Canada. Nat. Commun. 15, 6764 (2024). Jones, M. W. et al. State of Wildfires 2023–2024. Earth Syst. Sci. Data 16, 3601–3685 (2024). Keeley, J. E. & Syphard, A. D. Large California wildfires: 2020 fires in historical context. Fire Ecol. 17, 22 (2021). Squire, D. T. et al. Likelihood of unprecedented drought and fire weather during Australia’s 2019 megafires. Npj Clim. Atmospheric Sci. 4, 1–12 (2021). Wang, Z. et al. Severe Global Environmental Issues Caused by Canada’s Record-Breaking Wildfires in 2023. Adv. Atmospheric Sci. 41, 565–571 (2024). Jones, M. W. et al. Global and Regional Trends and Drivers of Fire Under Climate Change. Rev. Geophys. 60, e2020RG000726 (2022). Madadgar, S., Sadegh, M., Chiang, F., Ragno, E. & AghaKouchak, A. Quantifying increased fire risk in California in response to different levels of warming and drying. Stoch. Environ. Res. Risk Assess. 34, 2023–2031 (2020). Senande-Rivera, M., Insua-Costa, D. & Miguez-Macho, G. Spatial and temporal expansion of global wildland fire activity in response to climate change. Nat. Commun. 13, 1208 (2022). United Nations Environment Programme. Spreading like Wildfire: The Rising Threat of Extraordinary Landscape Fires, A UNEP Rapid Response Assessment . https://www.unep.org/resources/report/spreading-wildfire-rising-threat-extraordinary-landscape-fires (2022). van Oldenborgh, G. J. et al. Attribution of the Australian bushfire risk to anthropogenic climate change. Nat. Hazards Earth Syst. Sci. 1–46 (2020) doi: 10.5194/nhess-2020-69 . Blanchi, R., Lucas, C., Leonard, J. & Finkele, K. Meteorological conditions and wildfire-related houseloss in Australia. Int. J. Wildland Fire 19, 914 (2010). Fletcher, M.-S., Romano, A., Connor, S., Mariani, M. & Maezumi, S. Y. Catastrophic Bushfires, Indigenous Fire Knowledge and Reframing Science in Southeast Australia. Fire 4, 61 (2021). Wang, B. et al. Extreme fire weather is the major driver of severe bushfires in southeast Australia. Sci. Bull. 67, 655–664 (2022). Boer, M. M. & de Dios, V. R. Unprecendent burn area of Australian mega forest fires. Nat. Clim. Change 10, 170 (2020). Filkov, A. I., Ngo, T., Matthews, S., Telfer, S. & Penman, T. D. Impact of Australia’s catastrophic 2019/20 bushfire season on communities and environment. Retrospective analysis and current trends. J. Saf. Sci. Resil. 1, 44–56 (2020). Quiggin, J. Opinion: Australia is promising $ 2 billion for the fires. I estimate recovery will cost $ 100 billion. CNN https://www.cnn.com/2020/01/10/perspectives/australia-fires-cost (2020). Read, P. & Denniss, R. With costs approaching $ 100 billion, the fires are Australia’s costliest natural disaster. The Conversation http://theconversation.com/with-costs-approaching-100-billion-the-fires-are-australias-costliest-natural-disaster-129433 (2020). Brown, P. T. et al. Climate warming increases extreme daily wildfire growth risk in California. Nature 621, 760–766,766A (2023). Jolly, W. M. et al. Climate-induced variations in global wildfire danger from 1979 to 2013. Nat. Commun. 6, 7537 (2015). Abram, N. J. et al. Connections of climate change and variability to large and extreme forest fires in southeast Australia. Commun. Earth Environ. 2, (2021). Canadell, J. G. et al. Multi-decadal increase of forest burned area in Australia is linked to climate change. Nat. Commun. 12, 6921 (2021). Dowdy, A. J. Climatological variability of fire weather in Australia. J. Appl. Meteorol. Climatol. 57, 221–234 (2018). Di Virgilio, G. et al. Climate Change Increases the Potential for Extreme Wildfires. Geophys. Res. Lett. 46, 8517–8526 (2019). Dowdy, A. J. et al. Future changes in extreme weather and pyroconvection risk factors for Australian wildfires. Sci. Rep. 9, 1–11 (2019). Dupuy, J. et al. Climate change impact on future wildfire danger and activity in southern Europe: a review. Ann. For. Sci. 77, 1–24 (2020). El Garroussi, S., Di Giuseppe, F., Barnard, C. & Wetterhall, F. Europe faces up to tenfold increase in extreme fires in a warming climate. Npj Clim. Atmospheric Sci. 7, 1–11 (2024). He, Y., Zhou, Z., Im, E.-S. & Kwon, H.-H. Wildfire risk in a changing climate: Evaluating fire weather indices and their global patterns with CMIP6 multi-model projections. Weather Clim. Extrem. 48, 100751 (2025). Pimont, F. et al. Future expansion, seasonal lengthening and intensification of fire activity under climate change in southeastern France. Int. J. Wildland Fire 32, 4–14 (2022). Singo, M. V. et al. Projections of future fire risk under climate change over the South African savanna. Stoch. Environ. Res. Risk Assess. 37, 2677–2691 (2023). McArthur, A. G. Fire behaviour in Eucalypt forests. in Leaf. No 107 (Forestry and Timber Bureau, Canberra, ACT, 1967). Van Wagner, C. E. Development and structure of the Canadian Forest Fire Weather Index System, Report No. 35. Canadian Forestry Service . https://ostrnrcan-dostrncan.canada.ca/handle/1845/228434 (1987). Fox-Hughes, P. et al. A review of early severe weather applications of high-resolution regional reanalysis in Australia. Meteorol. Appl. 29, e2087 (2022). Dowdy, A. J., Mills, G. A., Finkele, K. & de Groot, W. Australian fire weather as represented by the McArthur Forest Fire Danger Index and the Canadian Forest Fire Weather Index, The Centre for Australian Weather and Climate Research . (2009). Clarke, H. & Evans, J. P. Exploring the future change space for fire weather in southeast Australia. Theor. Appl. Climatol. 136, 513–527 (2019). Fox-Hughes, P., Harris, R., Lee, G., Grose, M. & Bindoff, N. Future fire danger climatology for Tasmania, Australia, using a dynamically downscaled regional climate model. Int. J. Wildland Fire 23, 309–321 (2014). Dowdy, A. J. Seamless climate change projections and seasonal predictions for bushfires in Australia. J. South. Hemisphere Earth Syst. Sci. 70, 120–138 (2020). McAneney, J., Chen, K. & Pitman, A. 100-years of Australian bushfire property losses: Is the risk significant and is it increasing? J. Environ. Manage. 90, 2819–2822 (2009). IPCC. Global Warming of 1.5°C: IPCC Special Report on Impacts of Global Warming of 1.5°C above Pre-industrial Levels in Context of Strengthening Response to Climate Change, Sustainable Development, and Efforts to Eradicate Poverty, Cambridge University Press . https://www.cambridge.org/core/product/identifier/9781009157940/type/book (2018). Su, C.-H. et al. BARRA2: Development of the next-generation Australian regional atmospheric reanalysis, Report No. 067. Bureau of Meteorology . http://www.bom.gov.au/research/publications/researchreports/BRR-067.pdf (2022). Palmer, G. et al. Validation of BARRA2 and comparison with MERRA-2 and ERA5 using historical wind power generation. J. South. Hemisphere Earth Syst. Sci. 75, ES24028 (2025). Harris, S. & Lucas, C. Understanding the variability of Australian fire weather between 1973 and 2017. PLOS ONE 14, e0222328 (2019). Jakob, D. Challenges in developing a high-quality surface wind-speed data-set for Australia. Aust. Meteorol. Oceanogr. J. 60, 227–236 (2010). Lucas, C. On developing a historical fire weather data-set for Australia. Aust. Meteorol. Oceanogr. J. 61, 1–13 (2010). Clarke, H., Lucas, C. & Smith, P. Changes in Australian fire weather between 1973 and 2010. Int. J. Climatol. 33, 931–944 (2013). Russell-Smith, J. & Yates, C. P. Australian Savanna Fire Regimes: Context, Scales, Patchiness. Fire Ecol. 3, 48–63 (2007). McArthur, A. G. Weather and grassland fire behaviour. in Leaf. No 100 (Forestry and Timber Bureau, Canberra, ACT, 1966). Richardson, D. et al. Increased extreme fire weather occurrence in southeast Australia and related atmospheric drivers. Weather Clim. Extrem. 34, 100397 (2021). Lindenmayer, D. B. & Taylor, C. New spatial analyses of Australian wildfires highlight the need for new fire, resource, and conservation policies. Proc. Natl. Acad. Sci. 117, 12481–12485 (2020). Bowman, D. M. J. S., Kolden, C. A. & Williamson, G. J. Bushfires in Tasmania, Australia: An Introduction. Fire 5, 33 (2022). Harris, R. M. B., Remenyi, T., Fox-Hughes, P., Love, P. & Bindoff, N. L. Exploring the Future of Fuel Loads in Tasmania, Australia: Shifts in Vegetation in Response to Changing Fire Weather, Productivity, and Fire Frequency. Forests 9, 210 (2018). Dowdy, A. & King, A. Novel climate analysis methods applied to the Australian ESCI projections data. Front. Clim. 6, (2025). Trancoso, R. et al. Converting tropical forests to agriculture increases fire risk by fourfold. Environ. Res. Lett. 17, 104019 (2022). Lafferty, D. C. & Sriver, R. L. Downscaling and bias-correction contribute considerable uncertainty to local climate projections in CMIP6. Npj Clim. Atmospheric Sci. 6, 1–13 (2023). Wootten, A., Terando, A., Reich, B. J., Boyles, R. P. & Semazzi, F. Characterizing sources of uncertainty from global climate models and downscaling techniques. J. Appl. Meteorol. Climatol. 56, 3245–3262 (2017). Narsey, S. et al. Disentangling the uncertainties in regional projections for Australia. J. South. Hemisphere Earth Syst. Sci. https://www.publish.csiro.au/ES/justaccepted/ES25015 (in press). Balbi, M. & Lallemant, D. C. B. The Cost of Imperfect Knowledge: How Epistemic Uncertainties Influence Flood Hazard Assessments. Water Resour. Res. 59, e2023WR035685 (2023). Eccles, R. et al. Substantial increases in future precipitation extremes—insights from a large ensemble of downscaled CMIP6 models. Npj Nat. Hazards 2, 60 (2025). Li, C. et al. Changes in Annual Extremes of Daily Temperature and Precipitation in CMIP6 Models. J. Clim. 34, (2021). DCCEEW. National Vegetation Information System - Present Major Vegetation Groups and Subgroups - Version 7.0. https://fed.dcceew.gov.au/datasets/5e70b5afc 36a4c458a2cceb313eb3889/about (2024). Digital Atlas of Australia. Bushfire Historical Extents. https://digital.atlas.gov.au/datasets/912492b1683e494e8efbe84573ce6d70/about (2024). Chapman, S. et al. Evaluation of Dynamically Downscaled CMIP6-CCAM Models Over Australia. Earths Future 11, e2023EF003548 (2023). Trancoso, R., Syktus, J., Toombs, N. & Chapman, S. Assessing and selecting CMIP6 GCMs ensemble runs based on their ability to represent historical climate and future climate change signal, Report No. EGU23-11412. Copernicus Meetings . https://meetingorganizer.copernicus.org/EGU23/EGU23-11412.html (2023). Thatcher, M. Scientific description - CCAM. https://research.csiro.au/ccam/ (2020). McGregor, J. L. & Dix, M. R. An updated description of the conformal-cubic atmospheric model. in High Resolution Numerical Modelling of the Atmosphere and Ocean 51–75 (Springer New York, 2008). Hoffmann, P., Katzfey, J. J., McGregor, J. L. & Thatcher, M. Bias and variance correction of sea surface temperatures used for dynamical downscaling. J. Geophys. Res. 121, 12,877 – 12,890 (2016). Chapman, S., Syktus, J., Trancoso, R., Toombs, N. & Eccles, R. Projected changes in mean climate and extremes from downscaled high-resolution CMIP6 simulations in Australia. Weather Clim. Extrem. 46, 100733 (2024). Kim, Y., Rocheta, E., Evans, J. P. & Sharma, A. Impact of bias correction of regional climate model boundary conditions on the simulation of precipitation extremes. Clim. Dyn. 55, 3507–3526 (2020). Lim, C. M., Yhang, Y. B. & Ham, S. Application of GCM Bias Correction to RCM Simulations of East Asian Winter Climate. Atmosphere 10, 382 (2019). Beck, H. E. et al. MSWX: Global 3-Hourly 0.1° Bias-Corrected Meteorological Data Including Near-Real-Time Updates and Forecast Ensembles. Bull. Am. Meteorol. Soc. 103, E710–E732 (2022). Puri, K. et al. Implementation of the initial ACCESS numerical weather prediction system. Aust. Meteorol. Oceanogr. J. 63, 265–284 (2013). Lucas, C. Seasonal McArthur Forest Fire Danger Index (FFDI) data for Australia: 1973–2017. https://data.mendeley.com/datasets/xf5bv3hcvw/1 (2019). Zhang, H., Chapman, S., Trancoso, R., Toombs, N. & Syktus, J. Assessing the impact of bias correction approaches on climate extremes and the climate change signal. Meteorol. Appl. 31, e2204 (2024). Dowdy, A. A bias correction method designed for weather and climate extremes, Report No. 087. Australian Bureau of Meteorology . (2023). Perkins, S. E., Pitman, A. J., Holbrook, N. J. & McAneney, J. Evaluation of the AR4 climate models’ simulated daily maximum temperature, minimum temperature, and precipitation over Australia using probability density functions. J. Clim. 20, 4356–4376 (2007). Noble, I. R., Gill, A. M. & Bary, G. a. V. McArthur’s fire-danger meters expressed as equations. Aust. J. Ecol. 5, 201–203 (1980). Lucas, C., Hennessy, K., Mills, G. & Bathols, J. Bushfire Weather in Southeast Australia: Recent Trends and Projected Climate Change Impacts, Bushfire Cooperative Research Centre . (2007). Keetch, J. J. & Byram, G. M. A Drought Index for Forest Fire Control. US Dep. Agric. For. Serv. Res. Pap. 038, (1968). Griffiths, D. Improved Formula for the Drought Factor in McArthur’s Forest Fire Danger Meter. Aust. For. 62, 202–206 (1999). Tebaldi, C. et al. Climate model projections from the Scenario Model Intercomparison Project (ScenarioMIP) of CMIP6. Earth Syst. Dyn. 12, 253–293 (2021). IPCC. Climate Change 2021 – The Physical Science Basis: Working Group I Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press . https://www.cambridge.org/core/books/climate-change-2021-the-physical-science-basis/415F29233B8BD 19FB55F65E3DC67272B (2021). Krikken, F., Lehner, F., Haustein, K., Drobyshev, I. & van Oldenborgh, G. J. Attribution of the role of climate change in the forest fires in Sweden 2018. Nat. Hazards Earth Syst. Sci. 21, 2169–2179 (2021). Additional Declarations No competing interests reported. Supplementary Files Fireweathersupplementarynaturalhazardsformat.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 26 Nov, 2025 Reviews received at journal 26 Nov, 2025 Reviews received at journal 26 Nov, 2025 Reviews received at journal 22 Nov, 2025 Reviewers agreed at journal 17 Nov, 2025 Reviewers agreed at journal 16 Nov, 2025 Reviewers agreed at journal 16 Nov, 2025 Reviewers agreed at journal 16 Nov, 2025 Reviewers agreed at journal 12 Nov, 2025 Reviewers invited by journal 12 Nov, 2025 Editor assigned by journal 12 Nov, 2025 Submission checks completed at journal 12 Nov, 2025 First submitted to journal 06 Nov, 2025 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-8051769","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":549029043,"identity":"3cf75b27-3b1e-4da7-bba5-d281bf9536ba","order_by":0,"name":"Ryan McGloin","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABE0lEQVRIie2RsWrDMBCGTxSkRdQdr7TkGQKBpAVjv4qFwFlc6ORZJaAsoXMeo5NnG68ls8eWgqcM6VJCCbSySkkHGTwGqg+Bfh36dBwC8HhOmXMgCna0i2cKBynUKGRtFTJYMc/zIcpsLdsXkkWRZgv9FuZhDEw8NPAZCsWexy7luklnY1JIqXm1nGSbVCj+urghj13InApiMkVSlJKi0Fd3uk7ABCSregLQp8w/jsqt/oqPSrDtUTLbJbIK0aWZ3Siwr0eAPV34NkdRyITySl+uNlLobhah0hHF9t6psHmB70UUB2zZ4j7vgqya3SHkQSCfXApcJABmCfVboPaof37KSVDaLf5TMgoc+u57PB7PP+Qb3RxXijdLyH0AAAAASUVORK5CYII=","orcid":"","institution":"Queensland Future Climate Science Program, Queensland Government","correspondingAuthor":true,"prefix":"","firstName":"Ryan","middleName":"","lastName":"McGloin","suffix":""},{"id":549029044,"identity":"e4bbfd8c-b068-4263-8635-18f1d592b4ab","order_by":1,"name":"Ralph Trancoso","email":"","orcid":"","institution":"Queensland Future Climate Science Program, Queensland Government","correspondingAuthor":false,"prefix":"","firstName":"Ralph","middleName":"","lastName":"Trancoso","suffix":""},{"id":549029045,"identity":"138c2f04-5c88-4942-a15b-b14a7e5e648a","order_by":2,"name":"Jozef Syktus","email":"","orcid":"","institution":"School of The Environment, The University of Queensland","correspondingAuthor":false,"prefix":"","firstName":"Jozef","middleName":"","lastName":"Syktus","suffix":""},{"id":549029046,"identity":"9fe43dd1-de75-4926-97a6-e9aed2ce59ab","order_by":3,"name":"Rohan Eccles","email":"","orcid":"","institution":"Queensland Future Climate Science Program, Queensland Government","correspondingAuthor":false,"prefix":"","firstName":"Rohan","middleName":"","lastName":"Eccles","suffix":""},{"id":549029047,"identity":"be719cc7-8250-427a-b30c-d0d1e79ed110","order_by":4,"name":"Nathan Toombs","email":"","orcid":"","institution":"Queensland Future Climate Science Program, Queensland Government","correspondingAuthor":false,"prefix":"","firstName":"Nathan","middleName":"","lastName":"Toombs","suffix":""},{"id":549029048,"identity":"473b4a8e-7e0e-48db-a518-e9be01015e99","order_by":5,"name":"Andrew Dowdy","email":"","orcid":"","institution":"School of Geography, Earth and Atmospheric Sciences, University of Melbourne","correspondingAuthor":false,"prefix":"","firstName":"Andrew","middleName":"","lastName":"Dowdy","suffix":""}],"badges":[],"createdAt":"2025-11-07 00:23:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8051769/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8051769/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96620389,"identity":"5f34b1a0-9a40-4341-927c-f477b48e9bfe","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14122611,"visible":true,"origin":"","legend":"","description":"","filename":"Fireweatherarticlenaturalhazardsformat.docx","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/bcec3956d4af69d0550a633b.docx"},{"id":96620431,"identity":"5f7fed33-c8b8-45f9-9a8d-dd67de395482","added_by":"auto","created_at":"2025-11-24 10:54:09","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8333,"visible":true,"origin":"","legend":"","description":"","filename":"b2a2b061a1c14bdaab6979b1bd146cdd.json","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/270f151f534b7a677676d92f.json"},{"id":96620382,"identity":"c76b2c1e-1be9-433c-bf85-08a4346dbac3","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23362644,"visible":true,"origin":"","legend":"","description":"","filename":"Fireweathersupplementarynaturalhazardsformat.docx","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/0317be371c263d1a926f8dcd.docx"},{"id":96620409,"identity":"6efb5e80-fecb-45ed-9a0b-5ae735cf6c16","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":173937,"visible":true,"origin":"","legend":"","description":"","filename":"b2a2b061a1c14bdaab6979b1bd146cdd1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/9aaa62f8bafd3a48f50e7b9d.xml"},{"id":96620394,"identity":"2fb3a00e-020f-4a6e-ad14-1bc31b688752","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1276038,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/b10f65c35a5c0bbbcc4eaf16.png"},{"id":96620401,"identity":"01f0e5e0-9a6f-4eb7-997b-e72adc840344","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2076799,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/7390d9e507e42b9b8cd6c6c4.png"},{"id":96620418,"identity":"8ad9d5d6-e672-4e17-b529-2d30856f567c","added_by":"auto","created_at":"2025-11-24 10:54:08","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":217843,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/15badc47816bf9794bddbb7a.png"},{"id":96620405,"identity":"0664ddb5-a8dd-43dd-afd0-14259f71c75c","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":695871,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/254564da36bf6987e7589c87.png"},{"id":96620387,"identity":"47cb6897-c38e-4c1c-ba7c-10a46bc4d988","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1081032,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/07ff3cb61eb60154e3c29093.png"},{"id":96709044,"identity":"e9316982-0f1e-4ba8-96c2-071676b95baf","added_by":"auto","created_at":"2025-11-25 10:07:17","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":139598,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/5ac62c85b62d3801ca74d74c.png"},{"id":96708502,"identity":"20ed3935-cad5-4807-a7ae-c1c5710c46e7","added_by":"auto","created_at":"2025-11-25 10:03:56","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1419749,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/d69c1b3e943255ca12247d53.png"},{"id":96620423,"identity":"bea6353f-125f-408b-88de-b9cf39297401","added_by":"auto","created_at":"2025-11-24 10:54:09","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1779505,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/08493194aa71c9d555f43b84.png"},{"id":96620432,"identity":"944eb5c6-5e1e-4055-92e5-dbbca88d69d2","added_by":"auto","created_at":"2025-11-24 10:54:09","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1710335,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/034a0e7d051acfc79c7d8ced.png"},{"id":96620377,"identity":"f0ee2a70-1de7-4a55-a073-695b0533b027","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1760722,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/3420222c6b13a7ba25fe7100.png"},{"id":96709035,"identity":"6c729344-ebe6-4b8a-9820-5902603fe99b","added_by":"auto","created_at":"2025-11-25 10:07:14","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93066,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/f1c3cba2cef2f80c9d771acf.png"},{"id":96620370,"identity":"cff60b61-0e42-4ce5-b865-491afb9be7a9","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1639814,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/2c6bd8cd6d30cb3f4c8a8087.png"},{"id":96620395,"identity":"cb63bb45-7635-4372-b2cb-a0c3975300ce","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":306383,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/dfeb8c3494279a58653921d8.png"},{"id":96620374,"identity":"5eb981fe-5f62-43e1-9094-51307e4d58a9","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":400066,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/b21087c40db4f562e7a00763.png"},{"id":96620402,"identity":"c5134ac2-294c-4ec2-ac00-5f794dd9950b","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":71260,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/7e820e17eb9cf5a488854527.png"},{"id":96620308,"identity":"9c7a66e1-0c12-47b8-ab1f-31f1b2d75b9f","added_by":"auto","created_at":"2025-11-24 10:54:04","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":148861,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/f984891a0fe879f2ddde861a.png"},{"id":96708508,"identity":"bdead67e-431b-481d-a361-2bbf044db76e","added_by":"auto","created_at":"2025-11-25 10:04:02","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":250114,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/b5f064fd083d2a12c119daf5.png"},{"id":96708740,"identity":"6ee3f918-49bd-4239-b396-0a4cca270ea5","added_by":"auto","created_at":"2025-11-25 10:05:19","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39533,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/b4dff84dfc2aa37a5ee837e9.png"},{"id":96620412,"identity":"a884ac06-4d06-4607-9469-97fadec37322","added_by":"auto","created_at":"2025-11-24 10:54:08","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":335868,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/950a1dd0684e76918fe6adcb.png"},{"id":96620411,"identity":"a84ecb66-8a46-4272-ac1a-37b319bfdaf5","added_by":"auto","created_at":"2025-11-24 10:54:08","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":418385,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/83f564de3d89481ab5de3b52.png"},{"id":96620429,"identity":"a1de8696-e4ba-4f84-81b1-5097a1d7b56b","added_by":"auto","created_at":"2025-11-24 10:54:09","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":444441,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/7b69b868120525e576442b9a.png"},{"id":96620428,"identity":"4d1448b3-d446-4918-87c5-5250a1b7e9c6","added_by":"auto","created_at":"2025-11-24 10:54:09","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":457270,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/71d58ed9a347ee4fd5e2d11a.png"},{"id":96620324,"identity":"34913eb3-8ef7-41c4-8747-adeaa3ccd9c9","added_by":"auto","created_at":"2025-11-24 10:54:04","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40188,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/67b9f30654a859b3ab6773c8.png"},{"id":96620392,"identity":"e7ae6229-1d2c-4be7-83ec-2a3ed6c30faa","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":359462,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/c8904979a0b3df82d8641a37.png"},{"id":96620390,"identity":"fde64a13-10bd-4bd1-8cf2-2dbcd84adfa3","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"xml","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174828,"visible":true,"origin":"","legend":"","description":"","filename":"b2a2b061a1c14bdaab6979b1bd146cdd1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/4ab75c7475f14228bccc24a6.xml"},{"id":96620398,"identity":"ecce262b-0ab9-4a60-9712-d4c3351dc983","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"html","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":187974,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/02adae924d08b1e490da723d.html"},{"id":96620372,"identity":"64d8b074-8b79-4bb7-b5a2-d078fd9f4c6f","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2024341,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHeat map of Perkins skill scores calculated for the MSWX-Past and BARRA2 datasets using station observations for the 1980-2020 period\u003c/strong\u003e. Perkins skill scores are provided for each station and variable (daily relative humidity, maximum air temperature and wind speed and daily and monthly precipitation). The entire distributions were used in the Perkins skill score calculations. The higher of the MSWX-Past and BARRA2 Perkins skill scores at each station is highlighted in bold italics. The climate regions where the stations are located are indicated on the lefthand side of the figure using the following abbreviations: Temp = Temperate, Subt = Subtropical, Trop = Tropical, Equa = Equatorial, and Gras = Grassland.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/bb06fa9c27f9c727c0d7cf6b.png"},{"id":96620410,"identity":"ffa47fff-368a-42d7-b0a9-ad0875f0729a","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1744557,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHeat map of Perkins skill scores calculated for the bias corrected and non-bias corrected simulations of FFDI using a historical dataset of FFDI calculated from observations\u003c/strong\u003e. The FFDI data used in this analysis consists of the 50\u003csup\u003eth\u003c/sup\u003e, 75\u003csup\u003eth\u003c/sup\u003e, 90\u003csup\u003eth\u003c/sup\u003e, 95\u003csup\u003eth\u003c/sup\u003e and 97\u003csup\u003eth\u003c/sup\u003e percentiles of daily FFDI calculated for each summer in the 1981-2017 period. The mean Perkins skill score from the bias corrected and non-bias corrected downscaled CMIP6 ensembles is presented for each station and percentile. The higher of the bias corrected and non-bias corrected scores at each station is highlighted in bold italics.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/9cae3e6ee4862b9d2f39358e.png"},{"id":96620407,"identity":"696aa5d2-864d-4eed-87a7-4b3f16d398c9","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":347414,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCounts of stations where the Perkins skill score was higher for either the non-bias corrected or bias corrected downscaled ensembles\u003c/strong\u003e. Results are presented for Perkins skill scores calculated using seasonal FFDI percentiles (50\u003csup\u003eth\u003c/sup\u003e, 75\u003csup\u003eth\u003c/sup\u003e, 90\u003csup\u003eth\u003c/sup\u003e, 95\u003csup\u003eth\u003c/sup\u003e and 97\u003csup\u003eth\u003c/sup\u003e) for a) spring, b) summer, c) autumn and d) winter across the 1981-2017 period.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/a40377051c10e3a993a98e8a.png"},{"id":96708674,"identity":"b5e1fa7d-9fc9-4295-adba-ea16892c1bc5","added_by":"auto","created_at":"2025-11-25 10:05:03","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2049225,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison between percentiles of CCAM and BARRA2 FFDI input variables\u003c/strong\u003e. The ensemble average 95\u003csup\u003eth\u003c/sup\u003e percentile (a) bias corrected and (b) non-bias corrected FFDI is shown alongside the ensemble average CCAM and BARRA2 (d and e) 95\u003csup\u003eth\u003c/sup\u003e percentile daily maximum temperature, (g and h) 5\u003csup\u003eth\u003c/sup\u003e percentile 3pm (Local Time: LT) relative humidity, (j and k) 95\u003csup\u003eth\u003c/sup\u003e percentile 3pm (LT) wind speed, and (m and n) 5\u003csup\u003eth\u003c/sup\u003e percentile daily precipitation. Differences between the bias corrected and non-bias corrected FFDI, and CCAM and BARRA2 percentiles are also displayed (c, f, i, l and o). All percentiles were calculated for the 1981-2014 period. Red values in the top right corner show the averages over Australia.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/9e37e95b14319d5351bb6017.png"},{"id":96709234,"identity":"4a440fd0-515c-46a8-ad49-7655bbadd369","added_by":"auto","created_at":"2025-11-25 10:08:20","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2892235,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of ensemble mean change in the mean annual count of days with high (12 ≤ FFDI \u0026lt; 24), Very High (24 ≤ FFDI \u0026lt; 50) and Severe (FFDI ≥ 50) fire weather conditions for different levels of global warming\u003c/strong\u003e. Changes are calculated compared to the values under 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming (as shown in the top row). Changes in the counts of days with FFDI above the 95\u003csup\u003eth\u003c/sup\u003e and 99.726\u003csup\u003eth\u003c/sup\u003e percentiles calculated under 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming are also shown. Red values in the top right corner show the average changes over Australia while stippling shows where the signal to noise ratio \u0026gt; 1.0. Boxplots showing the variations in mean changes over Australia across the model ensemble are provided in Supplementary Material Figure S6.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/8579e60baaafae92fa8a7836.png"},{"id":96620417,"identity":"4cc83841-5bd3-42c6-9b12-6d5825efb23d","added_by":"auto","created_at":"2025-11-24 10:54:08","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":3270688,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of ensemble median change in the magnitude of 5-, 10-, 20-, 50-, and 100-year return interval 7-day running mean FFDI events for different levels of global warming\u003c/strong\u003e. Changes are calculated as percentage change compared to the values under 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming (as shown in the top row). Red values in the top right corner show the average changes over Australia while stippling shows where the signal to noise ratio \u0026gt; 1.0. Boxplots showing the variations in median changes over Australia across the model ensemble are provided in Supplementary Material Figure S12.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/fce1f689ad9d8eceabaae2c7.png"},{"id":96620388,"identity":"b24f9e55-3e85-470e-9335-78fb57e5d555","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2350271,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of ensemble median change in the likelihood of 5-, 10-, 20-, 50-, and 100-year return interval 7-day running mean FFDI events for different levels of global warming\u003c/strong\u003e. Changes are calculated as change in the likelihood of 7-day running mean FFDI events with magnitudes equal to those of 5-, 10-, 20-, 50-, and 100-year return interval events under 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming. Red values in the top right corner show the average changes over Australia while stippling shows where the signal to noise ratio \u0026gt; 1.0.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/48cb9f29753a20c2a8ca96d9.png"},{"id":96620426,"identity":"eca864f3-1585-4f93-b1ab-0b9d71a28cb2","added_by":"auto","created_at":"2025-11-24 10:54:09","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":799673,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eChanges in the likelihood of 5-, 10-, 20-, 50-, and 100-year return interval 7-day running mean FFDI events for different levels of global warming compared to 1.2 \u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e◦\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eC GWL\u003c/strong\u003e. Boxplots show variability in the median change for Australia across the model ensemble.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/e362740fa7fad26a662e8003.png"},{"id":96620375,"identity":"6075132e-1017-4530-ba20-074e0f0979ab","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":2795956,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of ensemble median change in the magnitude of 5-, 10-, 20-, 50-, and 100-year return interval 7-day running mean FFDI events for different levels of global warming\u003c/strong\u003e. Changes are calculated as percentage change compared to the values under 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming (as shown in the top row). Data is displayed for SE Australia with the dense eucalyptus forests in the region outlined in green. Red values in the top left corner show the average changes for all SE Australia while the green values are the average for the eucalypt forests. Stippling shows where the signal to noise ratio \u0026gt; 1.0. Boxplots showing the variations in median changes across the model ensemble for southeast Australia’s eucalypt forests are provided in Supplementary Material Figure S17.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/47f80321dbd858996bb007ac.png"},{"id":96620400,"identity":"3ef5023f-8cad-4c08-ad92-bd99faf9e71b","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":2598205,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of ensemble median change in the likelihood of 5-, 10-, 20-, 50-, and 100-year return interval 7-day running mean FFDI events for different levels of global warming\u003c/strong\u003e. Changes are calculated as change in the likelihood of 7-day running mean FFDI events with magnitudes equal to those of 5-, 10-, 20-, 50-, and 100-year return interval events under 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming. Data is displayed for SE Australia, with the dense eucalyptus forests in the region outlined in green. Red values in the top left corner show the average changes for all SE Australia while the green values are the average for the eucalypt forests. Stippling shows where the signal to noise ratio \u0026gt; 1.0.\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/abd9eeb65f473f7052fadba0.png"},{"id":96620406,"identity":"3111fdaf-dec1-4636-8201-6ca56aa12fc4","added_by":"auto","created_at":"2025-11-24 10:54:07","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":1665576,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eChanges in the likelihood of 5-, 10-, 20-, 50-, and 100-year return interval 7-day running mean FFDI events for southeast Australia’s eucalypt forests under different levels of global warming compared to 1.2 \u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e◦\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eC GWL\u003c/strong\u003e. Boxplots show variability in the median change for the different regions across the model ensemble.\u003c/p\u003e","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/7adbf84c5034bb77f24355ed.png"},{"id":96620385,"identity":"65723223-08a5-43e1-9373-9812f3999fbf","added_by":"auto","created_at":"2025-11-24 10:54:06","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":889872,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMaps of study area, and eucalyptus forest and historical wildfire extents in southeast Australia\u003c/strong\u003e.\u003cem\u003e a) \u003c/em\u003eExtent of study area with major climate regions and the location of the observational weather stations used and b) map of SE Australia’s dense (\u0026gt; 50% surface coverage per 10 by 10 km cell) eucalyptus forests (layer created using data from the Department of Climate Change, Energy, the Environment and Water\u003csup\u003e59\u003c/sup\u003e) and the combined area burned during five of the most destructive wildfire events in Australia’s history: the Black Summer (2019-2020), Black Saturday (2009), Black Friday (1939), Black Tuesday (1967) and Ash Wednesday (1983) fires (layer created using data from the Digital Atlas of Australia\u003csup\u003e60\u003c/sup\u003e).\u003c/p\u003e","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/436ec0594d9c2a95b571a94c.png"},{"id":96913545,"identity":"8a1cab1c-5bd9-4bbf-9677-aeb06fa1d845","added_by":"auto","created_at":"2025-11-27 14:02:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":25048943,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/bc409bcb-dbc2-4eb7-83d7-56f5b11fd276.pdf"},{"id":96620379,"identity":"f63229d3-2733-4b9b-918b-377f98b85e2e","added_by":"auto","created_at":"2025-11-24 10:54:05","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":23362644,"visible":true,"origin":"","legend":"","description":"","filename":"Fireweathersupplementarynaturalhazardsformat.docx","url":"https://assets-eu.researchsquare.com/files/rs-8051769/v1/7c23c1c15e92a39654017119.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Substantial increases in the likelihood of extreme fire weather events for fire-prone ecosystems in Australia","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eIn recent years, large wildfires in southeast (SE) Australia, South America, Canada, California, Siberia, and the Mediterranean region of Europe\u003csup\u003e\u003cspan additionalcitationids=\"CR2 CR3 CR4\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e have brought considerable attention to the threats wildfires pose to human life, public health, economies, ecosystems, and carbon reserves. There is also a growing interest in the role of climate change in driving the frequency and intensity of these events\u003csup\u003e\u003cspan additionalcitationids=\"CR7 CR8 CR9\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Due to the proximity of densely populated areas to highly flammable ecosystems with large fuel loads, SE Australia has historically been particularly vulnerable to catastrophic wildfires that have caused widespread fatalities and property damage\u003csup\u003e\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. For example, during the 2019/20 \u0026ldquo;Black Summer\u0026rdquo; fire season, over 23% of temperate forests in SE Australia were burned, an event of unprecedented scale both nationally and globally\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. The disaster resulted in 33 lives lost and the destruction of 3,000 homes\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. While the exact economic impact of the disaster is difficult to determine, some estimates suggest the cost could have been as high as AUD\u003cspan\u003e$\u003c/span\u003e100 billion\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eNumerous studies have shown that global warming is already driving an increase in the frequency of wildfire events across various regions worldwide\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e, including Australia\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan additionalcitationids=\"CR21\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, while extensive research also indicates that the frequency and intensity of fire weather events will accelerate further as the century unfolds\u003csup\u003e\u003cspan additionalcitationids=\"CR24 CR25 CR26 CR27 CR28\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. Wildfires occur when three key factors align: a supply of fuel, atmospheric conditions sufficiently conducive to fire ignition and spread, and an ignition source\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. While making projections of future fuel loads and ignition sources is complex and challenging, fire weather can be represented using an index such as the McArthur Forest Fire Danger Index (FFDI)\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e or Canadian Forest Fire Weather Index (FWI)\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e, which use air temperature, humidity, wind speed and precipitation to estimate the influence of weather on fire behaviour. While these indexes do not perfectly capture all elements that determine the intensity and impacts of wildfire events, they can be considered a useful way to combine several weather factors known to influence fire behaviour and have been shown to correlate well with attributes such as total area burned and house losses\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eGlobal Climate Models (GCMs) are valuable tools for exploring how global warming may impact climate hazards such as fire weather. However, GCMs operate at coarse spatial resolutions (150\u0026ndash;200 km), which limits their ability to accurately incorporate the influence of landscape features and capture regional-scale fire weather processes\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. To overcome this limitation, dynamical downscaling has been employed to improve the resolution of GCM outputs, including using downscaled outputs from the 3rd and 5th phases of the Coupled Model Intercomparison Project (CMIP3 and CMIP5) to assess changes to future fire weather regimes in Australia\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. These studies project a clear trend towards more dangerous near-surface fire weather conditions for Australia with the greatest increases occurring in spring.\u003c/p\u003e\u003cp\u003eCategories of FFDI can be used to help specify the fire weather intensity on a given day (e.g., days with an FFDI above 50 are classified as \u0026ldquo;Severe\u0026rdquo; fire weather days by fire management agencies for communication purposes\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e). For meaningful categorisation of FFDI calculated using downscaled GCM outputs, it is important that any biases in the underlying model data are corrected. Typically, bias correction is performed on variables that have temporally and spatially comprehensive observational records, such as temperature and precipitation. The absence of suitable gridded datasets for relative humidity and wind speed throughout Australia has historically posed a significant challenge for bias correction of FFDI, with partial bias correction (i.e. only for maximum temperature and precipitation) being the most common approach\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Research has shown that partial bias correction of FFDI does not consistently enhance model performance\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e, highlighting the need for robust datasets of relative humidity and wind speed to enable thorough bias correction of FFDI. Dowdy et al.\u003csup\u003e22\u003c/sup\u003e used an alternative approach to partial bias correction, where gridded observations of temperature, precipitation and relative humidity (derived from vapour pressure) were used in conjunction with reanalysis (NCEP\u0026ndash;NCAR) wind speeds that were bias corrected to the Bureau of Meteorology (BoM) wind speed forecasts used to provide FFDI forecasts to fire management agencies throughout Australia. The approach resulted in daily gridded FFDI values similar in magnitude to what fire managers use for peak daily values.\u003c/p\u003e\u003cp\u003eWhile it is useful to examine projected future changes in the annual number of days that fall in particular FFDI categories\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e, there is a need to focus on rarer extremes, as the majority of deaths and property destruction tend to occur during rare extreme fire weather events\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Studies have found that the majority of wild fire related house losses in Australia occurred on days with very rare catastrophic fire weather conditions (days with FFDI above 100) with nearly all losses occurring on days where the FFDI was above the 99.5th percentile\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. It is also useful to analyse rare fire weather events in regions where FFDI values of moderate magnitude may still indicate dangerous fire weather, even if those same values are common and not considered dangerous in other parts of Australia\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Extreme values can be assessed using comparatively modest thresholds, such as the 90th, 95th, and 99th percentiles as well as rarer conditions as represented by return intervals. Return intervals can be employed to quantify changes in the frequency and duration of extreme wildfire events under global warming\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. In the Australian context, return intervals have been used to examine the historical climatological variability of fire weather and attribute Australian wildfire risk to anthropogenic climate change\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003ePrevious studies of fire weather projections in Australia have been based on older CMIP generations and have tended to use small ensembles. Therefore, there is a need to quantify future changes in extreme fire weather using the latest downscaled CMIP6 projections under different levels of global warming with larger ensembles to better account for uncertainty. Analysing future changes under global warming levels (GWLs) presents advantages over analysis using emissions scenarios due to GWLs being directly tied to international climate policy goals, such as the Paris Agreement\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e and not being tied to specific emission scenarios. While bias correction of FFDI is not required for all applications, it is essential when using FFDI thresholds to categorise fire weather intensity. Previous attempts at bias correction of FFDI have largely relied on partial bias correction methods, however, there is a need for comprehensive bias correction using high-quality datasets for all input variables. Finally, there is a particular need to focus on changes in fire weather regimes and extremes in the eucalypt forests of SE Australia, which are especially vulnerable to intense and destructive wildfires\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. This study has three key aims which are designed to address the gaps in our understanding of fire weather in Australia and to help inform planning to build resilience:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eto evaluate the datasets available for bias-correcting FFDI input variables and to determine whether bias correction improves FFDI results in Australia\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eto assess how fire weather is projected to change across Australia under different GWLs using an ensemble of dynamically downscaled CMIP6 projections\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eto analyse projected changes in the intensity and frequency of extreme fire weather events (as represented by return intervals) across Australia, with a focused case study on the eucalypt forests of SE Australia.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e"},{"header":"2 Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Bias correction\u003c/h2\u003e\u003cdiv id=\"Sec4\" class=\"Section3\"\u003e\u003ch2\u003e2.1.1 Calibration data\u003c/h2\u003e\u003cp\u003eAn assessment of the similarity of two reanalysis-based datasets (MSWX-Past and BARRA2) and station observations was performed using Perkins skill scores in order to establish the most suitable dataset to be used as the calibration dataset in the bias correction process. Time series of daily relative humidity, maximum air temperature and wind speed and daily and monthly precipitation for the 1980\u0026ndash;2020 period were extracted from the MSWX-Past and BARRA2 datasets for each station location. Note that monthly precipitation was included in this analysis as the Keetch Byram Drought Index (KBDI), used in the FFDI calculation process, is influenced both by the precipitation that falls on a given day and the cumulative effects of precipitation over the preceding 20 days, particularly during dry spells.\u003c/p\u003e\u003cp\u003ePerkins skill scores for the entire distribution of the climate variables (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Supplementary Material Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e) showed that BARRA2 performed better across most regions for the maximum temperature and wind variables. BARRA2 and MSWX-Past performed similarly for relative humidity, while MSWX-Past outperformed BARRA2 at most locations for both daily and monthly precipitation (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). BARRA2 tended to outperform MSWX-Past for most variables and locations when calculating the Perkins skill scores for the lower and upper tails of the distributions (Supplementary Materials Figures \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e, S2 and S3), particularly for the extreme values most relevant for fire weather (i.e. upper tails for maximum temperature and wind speed and lower tails for relative humidity and precipitation). Therefore, BARRA2 was used as the calibration dataset in the bias correction process used in this study.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\u003ch2\u003e2.1.2 Bias corrected data\u003c/h2\u003e\u003cp\u003ePerkins skill score results were also used to compare how well the simulations bias corrected using the BARRA2 calibration dataset and non-bias corrected historical (1981\u0026ndash;2017) simulations of FFDI from the downscaled CMIP6 ensemble performed against FFDI calculated from observations at 39 BoM weather stations. Bias correction improved Perkins skill score results at most locations and across all seasons and percentiles (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). The overall effect of the bias correction process was to reduce the magnitude of FFDI over most of Australia, particularly in Western Australia (Supplementary Material Figure S4). The exception was along the north and north-eastern coastline of Australia where the process of bias correction tended to increase FFDI slightly. The effect is important when looking at categories of FFDI using thresholds, for example, bias correcting the downscaled simulations reduced the projected average annual count of days with FFDI\u0026thinsp;\u0026ge;\u0026thinsp;50 for all of Australia by nearly 50% for both 3 \u003csup\u003e◦\u003c/sup\u003eC GWL and 4 \u003csup\u003e◦\u003c/sup\u003eC GWL (Supplementary Material Figure S4). All results presented in the following sections of this study were calculated using the bias corrected downscaled CMIP6 ensemble.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe main cause of bias in the non-bias corrected FFDI is due to the Conformal Cubic Atmospheric Model (CCAM, the model used to perform the dynamical downscaling) overestimating maximum temperature and underestimating humidity when conditions are hot and dry for large parts of Australia (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Note that both BARRA2 and Australian Gridded Climate Data (AGCD) show similar results in this regard (Supplementary Material Figure S5). CCAM also tends to overestimate wind speeds in the east and north of the country, while winds in the south, west and central areas tend to be underestimated (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e and Supplementary Material Figure S5).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.2 FFDI severity categories\u003c/h2\u003e\u003cp\u003eSevere fire weather days (FFDI\u0026thinsp;\u0026ge;\u0026thinsp;50) are projected to increase over much of Australia with the magnitude of change increasing with the level of global warming. Increases are particularly apparent in northwestern and central Australia under 3 \u003csup\u003e◦\u003c/sup\u003eC and 4 \u003csup\u003e◦\u003c/sup\u003eC of global warming. Very High fire weather days (24\u0026thinsp;\u0026le;\u0026thinsp;FFDI\u0026thinsp;\u0026lt;\u0026thinsp;50) are projected to increase in the north and south of the country (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). As with Severe fire weather days, noticeable increases in the number of days with FFDI above the 95th and 99.726th (equivalent to 1-in-1-year event) percentiles are projected for central and western Australia, however, other regions with distinct increases are also projected in Tasmania and some elevated regions close to the eastern coast. The greatest increases in the number of Severe fire weather days for central and western Australia and Very High fire weather days for southern Australia are projected in spring (Supplementary Material Figures S7-S10), while the greatest increases in the number of Very High fire weather days in northern Australia are projected in autumn and winter (Supplementary Material Figures S7-S10).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Extreme fire weather events in Australia\u003c/h2\u003e\u003cp\u003eThe magnitude of both 7-day running mean (from here on referred to as just 7-day) and daily rare extreme FFDI events are projected to increase for most of Australia for all warming levels, with the magnitude of change increasing as the global climate warms (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e and Supplementary Material Figures S11, S12 and S13). Increases in the magnitude of the 7-day extreme events are more noticeable in southern and central areas of Australia than in the north for the 3 \u003csup\u003e◦\u003c/sup\u003eC and 4 \u003csup\u003e◦\u003c/sup\u003eC GWLs, with prominent patches of high increases in northern NSW, southeastern Victoria and Tasmania (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Results for the daily FFDI extreme events are noisier, however, patches of large increases in northern NSW, southeastern Victoria and Tasmania are also apparent for the 3 \u003csup\u003e◦\u003c/sup\u003eC and 4 \u003csup\u003e◦\u003c/sup\u003eC GWLs (Supplementary Material Figure S11). The frequency of both 7-day and daily rare extreme FFDI events (compared to 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming) are also projected to increase for all of Australia and for all warming levels, with greater increases projected for the rarest extremes (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e and Supplementary Material Figures S14 and S15). For the 7-day extreme events, there are well defined areas of increase in central and western areas (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e), with nosier results once again for the daily FFDI extreme events (Supplementary Material Figure S14). On average across Australia using the ensemble median, the 20-year and 100-year return interval 7-day FFDI events are projected to be 1.7 and 2.5 times more likely at 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 2.7 and 4.8 times more likely at 3 \u003csup\u003e◦\u003c/sup\u003eC GWL, respectively.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Extreme fire weather events in the eucalyptus forests of southeastern Australia\u003c/h2\u003e\u003cp\u003eProjections of extreme fire weather across all of Australia tend to draw the attention to the dry, sparsely vegetated regions of central and western Australia and away from SE Australia, a region with considerable coverage of dense, highly flammable forest ecosystems (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eb). The relative increase in magnitude of both 7-day and daily rare extreme FFDI events is projected to be substantial for eucalypt forests in SE Australia (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e and Supplementary Material Figures S16, S17 and S18), with increases in the ensemble median magnitude for 20-year and 100-year return interval 7-day FFDI events averaged for all of SE Australia\u0026rsquo;s eucalypt forests of 6.4% and 6.7% for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL; and 10.0% and 9.2% for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL. Particularly high increases are projected in Tasmania, where ensemble median increases in the magnitude of 20-year and 100-year return interval 7-day FFDI events are 12.1% and 11.5% for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 22.9% and 23.1% for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL (Supplementary Material Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). The lowest increases are projected for Queensland\u0026rsquo;s eucalyptus forests with increases for 20-year and 100-year return interval 7-day FFDI events of 4.2% and 4.7% for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL and 5.0% and 4.0% for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL. Similar results are seen for projected increases in the magnitude of rare extreme daily FFDI events (Supplementary Material Table S2).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eIncreases in the likelihood of both 7-day and daily rare extreme FFDI events (compared to 1.2 \u003csup\u003e◦\u003c/sup\u003eC of global warming) are greatest in inland areas. However, substantial increases are also projected for the eucalypt forests of SE Australia, particularly in Tasmania (Figs.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e and \u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e and Supplementary Material Figures S19 and S20), where ensemble median 20-year and 100-year return interval 7-day FFDI events increased by 2 and 2.6 times for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 3.2 and 5.2 times for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL (Supplementary Material Table S3). As with the magnitude of the rare extreme events, lowest increases in likelihood are projected for Queensland\u0026rsquo;s eucalyptus forests with increases for 20-year and 100-year return interval 7-day FFDI events of 1.5 and 2.1 times for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 1.8 and 2.2 times for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL, respectively. Similar results are seen for projected increases in the likelihood of rare extreme daily FFDI events (Supplementary Material Table S4).\u003c/p\u003e\u003cp\u003eWhen averaged over all eucalypt forests in SE Australia, 20-year and 100-year return interval 7-day FFDI events are projected to become approximately 1.6 and 2.3 times more likely for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 2.1 and 3 times more likely for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e and Supplementary Material Table S4). Increases in both the magnitude and likelihood of rare extreme FFDI events for eucalypt forests in Victoria and New South Wales are similar to the results for all of SE Australia, with increases tending to be slightly higher in Victoria (Supplementary Material Tables S1 - S4). It should be noted that there was generally high uncertainty associated with the ensemble median projected changes in the magnitude and frequency of the rarest extreme events (50-year and 100-year return-interval events) and for the lower GWLs, with few regions showing significant changes following our signal-to-noise ratio analysis. However, there are substantial areas of low uncertainty for the higher GWLs and the more moderate rare extreme events (5-year, 10-year and 20-year return-interval events). Note there are also patches of low uncertainty for the rarest extreme events under high levels of global warming, particularly for the 7-day FFDI events, such as in Victoria and Tasmania. The spread in the ensemble results is shown in detail in the box plots presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e and Supplementary Material Figures S17, S18 and S20, and in the difference between the 10th and 90th percentile maps shown in Supplementary Material Figures S21 and S22.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"3 Discussion","content":"\u003cp\u003eThe results from this study indicate that BARRA2 is generally more appropriate than MSWX-Past for bias correcting FFDI inputs from downscaled GCMs in Australia. This is logical considering that BARRA2 is a reanalysis product specifically developed for the Australian region\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Other studies have also shown that BARRA2 outperforms other reanalysis products when simulating climate variables in the Australian region, such as wind power\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. However, the results from this study suggest that MSWX-Past may be a more appropriate calibration source for bias correction if looking at precipitation specific indices such as extreme rainfall. Previous studies have typically relied on partial bias correction to bias correct FFDI and these studies have found that partial bias correction of FFDI does not consistently improve model performance and can lead to a greater spread in model simulations, particularly for extreme values\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. Results from this study suggest that there are benefits to bias correcting each FFDI input variable separately using the Quantile Matching for Extremes (QME) method and BARRA2 data as calibration data, resulting in improved performance with historical FFDI values calculated using observations.\u003c/p\u003e\u003cp\u003eHowever, the best approach for bias correction of FFDI may be dependent on the application. For example, a previous study\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e that bias corrected wind data to match the BoM operational forecast wind speeds used for FFDI forecasts noted that the FFDI values tended to be higher than the station-based FFDI\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e used to assess the impact of bias correction in this study. Ultimately, while improvements have been made in recent years, there are still limitations associated with the available meteorological datasets required for complete bias correction of FFDI. This is particularly the case for wind speed given the relatively limited number of stations that have wind data of suitable quality and issues associated with spatial interpolation using sparse datasets\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e and also because FFDI was designed to use 10-minute averages of wind speed\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e rather than the instantaneous values provided by most climate models and reanalysis products\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. It should be noted that bias correction in this study only substantially affected results when FFDI thresholds were used to categorise fire weather intensity (e.g. days with FFDI\u0026thinsp;\u0026ge;\u0026thinsp;50), with bias corrected vs non-bias corrected results being very similar when analysing changes in percentiles and return intervals (please compare Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e and Supplementary Materials Figure S23 for an illustration of this).\u003c/p\u003e\u003cp\u003eThis study found that Severe fire weather (FFDI\u0026thinsp;\u0026ge;\u0026thinsp;50) days are projected to increase substantially in north-western and central Australia, while Very High fire weather days (24\u0026thinsp;\u0026le;\u0026thinsp;FFDI\u0026thinsp;\u0026lt;\u0026thinsp;50) are projected to increase in the north and south of the country. Interpretation of the increases in Severe fire weather days for central and western Australia should be done with caution as these regions are not typically dominated by the forested ecosystems for which the FFDI was designed\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e and are far from major population centres where wildfires are likely to have the greatest impact. In contrast, the region in southern Australia where there are projected increases in Very High fire weather conditions (including the areas surrounding Adelaide and southwestern Victoria) is an area that has been greatly affected by devastating wildfires in the past, such as the Ash Wednesday and Black Friday events. The increases in Very High fire weather conditions in northern Australia could also have implications due to the high number of fires that already occur in the regions\u0026rsquo; tropical savannas, which account for approximately 68% of the national fire extent annually\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. Note that for more accurate simulation of savanna fire weather conditions, the Grassland Fire Danger Index (GFDI)\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e is recommended instead of the FFDI presented here, noting that GFDI includes factors such as grass curing that have substantial uncertainties when estimated using climate model data.\u003c/p\u003e\u003cp\u003eThe largest increases in Severe and Very High fire weather category days were projected in spring and autumn, this agrees with findings from other studies in Australia. Both observations and projections show a trend of the largest increases in fire weather conditions occurring in spring\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e, while an exponential increasing trend in total burned forest area during autumn and winter has been suggested\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Therefore, the findings of this and other studies\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e indicate an increasing trend in fire season length across Australia.\u003c/p\u003e\u003cp\u003eThe magnitude and frequency of extreme FFDI events were projected to increase substantially in the eucalypt forests of SE Australia. Observational studies have shown that this is a region that is already experiencing large increases in both extreme fire weather conditions and the area burned during wildfires\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. Recent increases in fire weather magnitude in this region are associated with increases in temperature and decreases in precipitation and humidity linked to climate change\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. Shifts in large-scale teleconnection modes likely played a role in these changes with evidence suggesting that fire promoting phases of tropical Pacific and Indian Ocean variability (+\u0026thinsp;ENSO, +IOD, and/or \u0026minus;\u0026thinsp;SAM) are becoming more frequent\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. In addition, an increase in anticyclonic blocking events may also have contributed to more frequent days with coinciding low relative humidity and high wind speeds\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eThe particularly high increases in the magnitude and frequency of rare extreme events projected for Tasmania in this study complements the findings of observation-based studies which have shown that Tasmania has seen particularly large increases in the area burned during wildfires over the past 30 years compared to other regions in Australia\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Tasmania has experienced several devastating wildfires in the past. For example, 62 people were killed and nearly 3,000 structures were destroyed in southern Tasmania during the 1967 Black Tuesday wildfires, while in January 2013, wildfires caused widespread destruction of infrastructure, including 203 homes, in the village of Dunalle\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Tasmanian vegetation is characterised by mosaics of fire-dependent and fire sensitive vegetation types, where large areas of flammable eucalyptus forests surround more fire-sensitive vegetation types such as wet and dry rainforests and alpine shrublands\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Fire frequency has been found to be the dominant driver of future fire activity in Tasmania because of frequent fire leading to shifts in vegetation type away from fire-sensitive types towards drier, flammable, fire-adapted vegetation\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. This suggests that impacts from the projected increases in frequency of extreme fire weather conditions in Tasmania could be amplified by a shift in vegetation towards more flammable fire-adapted vegetation types.\u003c/p\u003e\u003cp\u003eThis study represents one of the few that has analysed future downscaled simulations of rare extreme fire weather events in Australia and the only study, that these authors are aware of, that has used the latest CMIP6 projections under different levels of global warming to analyse rare extremes for SE Australia\u0026rsquo;s vulnerable eucalypt forests. These results complement the findings of previous studies that have generally found a clear trend towards more dangerous fire weather conditions for SE Australia\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. This includes a previous study using ensembles of bias corrected downscaled projections that found increases in extreme FFDI that scaled approximately linearly with GWL\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e, while noting that was based on the older set of CMIP5 GCMs. It also accompanies other studies from around the world that have identified projected increases in the frequency and magnitude of rare wildfire extremes in fire prone regions in North America and Europe\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e and others that have analysed the impacts of non-global warming related factors on extreme fire weather, such as deforestation in Borneo\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eThere would be value in repeating the analysis of this study to consider multiple ensembles of projections derived from multiple downscaling methodologies to account for uncertainty arising from using downscaled projections from a single modelling group\u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e,\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e. Additional sources of uncertainty include natural variability, scenario uncertainty, and epistemic uncertainty\u003csup\u003e\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e which have not been explicitly considered in this study. Natural variability has been shown to be the largest source of uncertainty in near-term projections of climate extremes globally\u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e, including over Australia which is subject to very high inter-annual variability. There is a degree of epistemic uncertainty inherent in climate models, while uncertainty is also introduced into this study via the estimates of extremes, particularly for the rarest extremes\u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e,\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. These are the result of fitting distributions to a relatively limited dataset (20 years), though the grid cell pooling approach adopted in this study may go some way to helping reduce this\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. It should also be highlighted that the projected changes of FFDI presented in this study solely represent a measure of how conducive the atmospheric conditions are for wildfire ignition and spread and are not projections of wildfire risk, as this would require information on ecological and human factors beyond the reliability of climate projections\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eTo conclude, we assessed the impact of climate change on fire weather in Australia using high-resolution climate simulations from an ensemble of downscaled CMIP6 GCMs. The study combines dynamical downscaling, GEV distribution and GWL analysis to better understand the potential impacts of climate change on rare extreme fire weather events across Australia, including the vulnerable eucalypt forests in the southeast of the country. It was found that bias correcting each of the FFDI input variables using a BARRA2 calibration dataset resulted in improvements over non-bias corrected data when compared with FFDI calculated using historical observations.\u003c/p\u003e\u003cp\u003eSevere fire weather (FFDI\u0026thinsp;\u0026ge;\u0026thinsp;50) days are projected to increase considerably in north-western and central Australia under 3 \u003csup\u003e◦\u003c/sup\u003eC and 4 \u003csup\u003e◦\u003c/sup\u003eC of global warming, while Very High fire weather days (24\u0026thinsp;\u0026le;\u0026thinsp;FFDI\u0026thinsp;\u0026lt;\u0026thinsp;50) are projected to increase in the north and south of the country. The magnitude and frequency of extreme FFDI events are expected to increase substantially, particularly in the southern half of Australia, with 20-year and 100-year return interval 7-day FFDI events averaged for all of Australia projected to become approximately 1.7 and 2.5 times more likely for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 2.7 and 4.8 times more likely for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL. For the wildfire prone eucalyptus ecosystems of SE Australia, the increases are most notable for Tasmania with 20-year and 100-year return interval 7-day FFDI events projected to become approximately 2 and 2.6 times more likely for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 3.2 and 5.2 times more likely for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL. Queensland\u0026rsquo;s eucalyptus forests had the lowest projected increases in frequency; however, the changes were not immaterial, with the likelihood of 20-year and 100-year return interval 7-day FFDI events increasing by approximately 1.5 and 2.1 times for 2 \u003csup\u003e◦\u003c/sup\u003eC GWL, and 1.8 and 2.2 times for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL, respectively.\u003c/p\u003e"},{"header":"4 Methodology","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e4.1 Study Area\u003c/h2\u003e\u003cp\u003eThis study evaluated changes to fire weather across the Australian continent, which encompasses a range of climate regions, including arid, equatorial, grassland, sub-tropical, temperate, and tropical regions (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003ea). To support this analysis, observational datasets from selected weather stations across Australia were used to validate the data inputs and outputs of the FFDI bias correction process (see section 4.2.2 for details). This study includes a case study focussed on SE Australia, a region with considerable coverage of dense, highly flammable forest ecosystems and that has experienced several large-scale wildfires with catastrophic impacts during the last century (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eb)\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e4.2 Data\u003c/h2\u003e\u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\u003ch2\u003e4.2.1 Modelling\u003c/h2\u003e\u003cp\u003eWe used an ensemble of 15 downscaled climate model simulations derived from 11 different CMIP6 GCMs (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), with some model variants downscaled multiple times. The ensemble consists of historical simulations up to 2014 and future simulations under a high emissions pathway (SSP370) up to 2100. The SSP370 pathway was selected as this was the highest emissions pathway available and enabled analysis up to 4 \u003csup\u003e◦\u003c/sup\u003eC of global warming. The GCMs included in this study were selected to best represent the future spread in the climate change signals from the GCMs, while prioritising models that performed well in simulating the Australian climate\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e,\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e. Dynamical downscaling was performed for the full ensemble of GCMs using CCAM\u003csup\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/sup\u003e developed by CSIRO\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. CCAM was run using a stretched C288 grid, providing a model resolution of approximately 10 km for Australia. The downscaling approach used involved correcting for model biases in sea surface temperatures and sea ice\u003csup\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e,\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e. This approach has been found to improve the simulations of climate from CCAM and other regional climate models\u003csup\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e,\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e,\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e. Five of the CCAM simulations were run using dynamic atmosphere-ocean coupling as shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The downscaling approach adopted has been shown to significantly improve the performance over the host GCMs for precipitation and temperature in all seasons, with the largest improvements noted for climate extremes\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDetails of the 15 climate model simulations downscaled from 11 CMIP6 GCMs considered in this study.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCMIP6 Model\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eModel full name\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eResolution\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEnsemble member\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eCCAM setup\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003eACCESS-ESM1.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003eAustralian Community Climate and Earth System Simulator, version 1.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003e1.875 x 1.25\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er6i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er20i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatm-ocean coupled\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er40i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatm-ocean coupled\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eACCESS_CM2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAustralian Community Climate and Earth System Simulator, version 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.875 x 1.25\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er2i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatm-ocean coupled\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCMCC-ESM2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCentro Euro-Mediterraneo sui Cambiamenti Climatici\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.9 x 1.25\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eCNRM-CM6-1-HR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eCentre National de Recherches M\u0026eacute;t\u0026eacute;orologiques Coupled Global Climate Model, version 6.1, high-resolution\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e0.5 x 0.5\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatm-ocean coupled\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEC-Earth3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEuropean Community Earth-System Model, version 3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.8 x 0.8\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFGOALS-g3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFlexible Global Ocean-Atmosphere-Land System Model, grid point version 3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2.5 x 2.5\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eR4i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGFDL-ESM4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGeophysical Fluid Dynamics Laboratory Earth System Model, version 4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1 x 1\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGISS-E2-2-G\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGoddard Institute for Space Studies Model E2.2G\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2. x 2.5\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er2i1p1f2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMPI-ESM1-2-LR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMax Planck Institute Earth System Model, version 1.2, low resolution\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.9 x 1.9\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er9i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMRI-ESM2-0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMeteorological Research Institute Earth System Model, version 2.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.125 x 1.125\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eNorESM2-MM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eNorwegian Earth System Model, version 2, 1 degree resolution\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e1 x 1\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatmospheric\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003er1i1p1f1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eatm-ocean coupled\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section3\"\u003e\u003ch2\u003e4.2.2 Reanalysis and observations\u003c/h2\u003e\u003cp\u003eBias correction of FFDI when using downscaled simulations as input variables requires a calibration dataset of gridded meteorological variables at an appropriate resolution and temporal scale. For FFDI, this necessitates sub-daily wind speed and humidity data, as the index is typically calculated using mid-afternoon (e.g., 3 pm) or daily maximum measurements of these variables (see section 4.5). However, this has historically posed a challenge, as sub-daily observations, particularly for wind speed, are sparse in Australia\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. To address this limitation, high resolution reanalysis datasets can be used in place of observations, providing long-term, spatially complete records of historical climate variables. Two reanalysis-based products were identified as suitable for this purpose, offering high resolution, sub-daily data across Australia for all required FFDI input variables:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eThe Multi-Source Weather product (MSWX-Past)\u003csup\u003e\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e, and\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eBureau of Meteorology Atmospheric high-resolution Regional Reanalysis for Australia (BARRA2)\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003eMSWX is a global gridded meteorological product featuring 3-hourly 0.1\u0026deg; resolution forecasts and historical records. The historical part of the record (MSWX- Past) starts from 1 January 1979 and is based on ERA5 reanalysis data bias corrected and downscaled using high-resolution reference climatologies. The reference climatologies are based on data derived from station observations, satellite imagery, and/or model output, depending on the variable\u003csup\u003e\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e. BARRA2 is the second generation of Australian reanalysis and covers the Australian continent and the surrounding region. BARRA2 reanalysis is produced by running the limited-area Australian Community Climate and Earth-System Simulator (ACCESS) model\u003csup\u003e\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e\u003c/sup\u003e. The ACCESS model is forced with ERA5 reanalyses boundary conditions and adjusted to better reflect observations via data assimilation\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Both datasets were re-gridded to match the resolution of the downscaled climate simulations using distance-weighted interpolation for precipitation and bilinear interpolation for all other variables. These datasets were evaluated against station observations in section 2.1.1 of this study, with the best performing dataset selected for use as the calibration dataset in the bias correction process.\u003c/p\u003e\u003cp\u003eObservational datasets from selected weather stations (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003ea) were used to validate both the reanalysis data used to bias correct the FFDI input variables and the final bias corrected FFDI data. Daily timeseries of mean relative humidity (%), wind speed (km/hr), precipitation (mm) and daily maximum surface air temperature (\u0026deg;C) were extracted for the 1980\u0026ndash;2020 period from Australian BoM stations, which are marked in blue and black in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003ea. Stations were selected based on whether all four input variables were available and according to the length and completeness of their data records. Note that ideally 3 pm (Local Time: LT) relative humidity and wind speed would have been used instead of daily values, since the FFDI calculation and bias correction use the 3 pm variables (see section 4.5). However, sub-daily relative humidity and wind speed were not available for the BoM stations and therefore, we decided that the product that best simulated the daily relative humidity and wind speed was also highly likely to be superior in simulating the sub daily situation.\u003c/p\u003e\u003cp\u003eWe also used a historical dataset of FFDI\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u003c/sup\u003e calculated from observations from 39 high quality BoM stations (red and black stations in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003ea) to validate the final bias corrected FFDI data. This dataset consists of 97th, 95th, 90th, 75th and 50th percentiles of daily FFDI estimated over standard meteorological seasons from 1981\u0026ndash;2017.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e4.3 Bias correction\u003c/h2\u003e\u003cp\u003eBias correction is a common tool used to correct climate model data and involves removing biases in model outputs by calibrating them against reliable observational data\u003csup\u003e\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u003c/sup\u003e. After establishing the reanalysis dataset that showed the best agreement with station observations (see section 4.2.2), the downscaled FFDI input variables were bias corrected using the QME method\u003csup\u003e\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e\u003c/sup\u003e, which applies quantile-quantile matching to the input data histograms. A calibration period of 1981\u0026ndash;2020 was used in this process. Correction functions derived from the calibration period were then applied to future climate projections for each climate model on an individual grid cell basis.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003e4.4 Evaluation\u003c/h2\u003e\u003cp\u003eWe evaluated the performance of the reanalysis-based datasets used to bias correct the FFDI and the resulting bias corrected FFDI using the Perkins skill score\u003csup\u003e\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e\u003c/sup\u003e. The Perkins skill score evaluates data based on similarity between the modelled (in this case reanalysis data and modelled FFDI data) and observed probability density functions (PDFs). The binning of data to construct histograms was based on the distribution of the observed data. The Perkins skill score was then calculated as follows:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:{S}_{score}=\\sum\\:_{1}^{n}minimum({Z}_{m},{Z}_{0})$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cem\u003en\u003c/em\u003e is the number of bins used to calculate the histogram, \u003cem\u003eZ\u003c/em\u003e\u003csub\u003e\u003cem\u003em\u003c/em\u003e\u003c/sub\u003e, is the frequency of values in each bin for the model, and \u003cem\u003eZ\u003c/em\u003e\u003csub\u003e\u003cem\u003eo\u003c/em\u003e\u003c/sub\u003e is the frequency of values in each bin for the observations. If a model simulates the observed PDF poorly, the Perkins skill score will be close to zero, while if the PDF is well simulated the score will approach a maximum of 1. The Perkins skill score was used to assess the entire distribution of the climate variables as well as the lower and upper tails of the distributions (0\u0026ndash;5th and 95\u0026ndash;100th percentiles).\u003c/p\u003e\u003cp\u003eStippling on the maps presented in this study represents where the signal-to-noise ratio was greater than 1, indicating where the climate change signal from the ensemble emerges from the noise of the ensemble. Here, the signal is represented by either the ensemble mean or median change and the noise is the standard deviation from all 15 models in the ensemble as in Chapman et al.\u003csup\u003e66\u003c/sup\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003e4.5 FFDI\u003c/h2\u003e\u003cp\u003eThe McArthur FFDI was calculated following\u003csup\u003e\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e\u003c/sup\u003e:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:FFDI=2exp\\left(0.0338T+0.0234W-0.0345RH+0.987\\text{ln}\\left(DF\\right)-0.45\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cem\u003eDF\u003c/em\u003e is the drought factor, \u003cem\u003eT\u003c/em\u003e the daily surface air temperature (\u0026deg;C), \u003cem\u003eW\u003c/em\u003e the wind speed at a height of 10 m (km h\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e) and \u003cem\u003eRH\u003c/em\u003e the relative humidity (%). In this study we have followed the common approach of using the maximum daily temperature as \u003cem\u003eT\u003c/em\u003e, and 3 pm values of wind speed and relative humidity as \u003cem\u003eW\u003c/em\u003e and \u003cem\u003eRH\u003c/em\u003e (%) to represent mid-afternoon FFDI, when values are typically near the daily maximum\u003csup\u003e\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e\u003c/sup\u003e. The drought factor is an estimate of fuel dryness and is computed using the KBDI\u003csup\u003e\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e\u003c/sup\u003e and a function to expresses the influence of the past precipitation amount and the time since it fell\u003csup\u003e\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e\u003c/sup\u003e. Daily bias corrected and non-bias corrected FFDI were calculated for the downscaled CMIP6 ensemble with projected changes in FFDI presented for different GWLs. GWLs were derived from CCAM using an 11-year running average of global surface temperatures\u003csup\u003e\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e\u003c/sup\u003e. Warming levels were calculated relative to 1995\u0026ndash;2014 with an offset of 0.85\u0026deg;C\u003csup\u003e80\u003c/sup\u003e added to derive the warming from preindustrial levels. Twenty-year time slices, centred on the target GWL (midpoint\u0026thinsp;\u0026minus;\u0026thinsp;9 years to midpoint\u0026thinsp;+\u0026thinsp;10 years), were used to extract the FFDI data for that GWL. Changes to future fire weather were assessed in terms of counts of days within specific FFDI categories (i.e. Severe, defined as FFDI\u0026thinsp;\u0026gt;\u0026thinsp;50, Very High, defined as 24\u0026thinsp;\u0026le;\u0026thinsp;FFDI\u0026thinsp;\u0026lt;\u0026thinsp;50, and High, defined as 12\u0026thinsp;\u0026le;\u0026thinsp;FFDI\u0026thinsp;\u0026lt;\u0026thinsp;24), percentiles (i.e. FFDI greater or equal to the 95th percentile and 99.726th percentile) and return intervals (see section 4.6).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003e4.6 Extreme value analysis\u003c/h2\u003e\u003cp\u003eExtreme value analysis was applied to assess changes to the probability distribution of rare extreme events. Here these events represent the 1-in-5, 1-in-10, 1-in-20, 1-in-50, and 1-in-100 Annual Exceedance Probabilities (AEPs) which approximately correspond to events with annual return intervals of 5, 10, 20, 50, and 100 years respectively. We sampled the daily timeseries of FFDI at each grid cell using the block maxima approach to derive annual maxima (AM) and then pooled together data from nearby cells using a 5x5 box centred on each grid cell to extend the data series used for the extreme event analysis as per Eccles et al.\u003csup\u003e57\u003c/sup\u003e. Typically, annual maximum daily values are used when using the block maxima approach, however, some studies have found that 7-day running mean FWI (which typically behaves similarly to FFDI\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e) has shown a good correlation with area burned during wildfires, including during the 2019/20 \u0026ldquo;Black Summer\u0026rdquo; fire season in Australia\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e\u003c/sup\u003e. Therefore, in this study, AM series were calculated using both the annual maximum daily FFDI and annual maximum 7-day running mean FFDI. The Generalised Extreme Value (GEV) distribution was then fitted to the AM series using the l-moments method for parameter estimation. The GEV distribution is a generalised expression combining the Gumbel, Fr\u0026eacute;chet, and the Weibull distributions and is given by:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:G\\left(x\\right)=exp\\left\\{-{\\left[1+\\xi\\:\\left(\\frac{x-\\mu\\:}{\\sigma\\:}\\right)\\right]}^{-1/\\xi\\:}\\right\\},\\:for\\:\\left\\{x:1+\\xi\\:\\left(\\frac{x-\\mu\\:}{\\sigma\\:}\\right)\u0026gt;0\\right\\}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere, \u003cem\u003e\u0026micro;\u003c/em\u003e, \u003cem\u003eσ\u003c/em\u003e, and \u003cem\u003eξ\u003c/em\u003e are the location, scale, and shape parameters, respectively. Here, the location parameter is a measure of the central tendency and is loosely linked to the mean, the scale parameter is a measure of variance, and the shape parameter describes the tail behaviour. We fit the GEV distribution to each 20-year time slice of FFDI AM. These distributions were then used to determine the FFDI values corresponding to particular return intervals. Changes in the magnitude of return interval events were calculated relative to the values under 1.2\u0026deg;C of global warming. We also present results of changes in the frequency of particular return interval events (with the magnitude of the events taken from the 1.2\u0026deg;C GWL).\u003c/p\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eCompeting interests\u003c/h2\u003e\u003cp\u003eAll authors declare no financial or non-financial competing interests.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eDeclaration\u003c/p\u003e\u003cp\u003eThis research did not receive funding.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eR.M.: Writing \u0026ndash; Original draft preparation, Conceptualization, Methodology, Formal analysis. R.T.: Conceptualization, Methodology, Writing. - Review \u0026amp; Editing. J.S.: Data Curation, Methodology, Writing. - Review \u0026amp; Editing. R.E.: Conceptualization, Methodology, Writing - Review \u0026amp; Editing. N.T.: Data Curation, Methodology. A.D.: Conceptualization, Writing - Review \u0026amp; Editing.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe acknowledge the World Climate Research Programme, which, through its Working Group on Coupled Modelling, coordinated and promoted CMIP6. We thank the climate modelling groups for producing and making available their model output, the Earth System Grid Federation (ESGF) for archiving the data and providing access, and the multiple funding agencies who support CMIP6 and ESGF. Data are available through the ESGF at: http://esgf.llnl.gov/. We also acknowledge Lindsay Brebber from Information and Digital Science Delivery at Queensland Government for support with high performance computing and data storage.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data used in this study are publicly available. The downscaled climate projections that are used in the FFDI calculations can be accessed through the National Computer Infrastructure at: [https://dx.doi.org/10.25914/8fve-1910](https:/dx.doi.org/10.25914/8fve-1910) . The CMIP6 global climate model data are available through the Earth System Grid Federation at: [http://esgf.llnl.gov/](http:/esgf.llnl.gov) . Gridded BARRA2 reanalysis data can be accessed through the National Computer Infrastructure at: [https://dx.doi.org/10.25914/90rq-d839](https:/dx.doi.org/10.25914/90rq-d839) , while MSWX data can be accessed via request at: [https://www.gloh2o.org/mswx/](https:/www.gloh2o.org/mswx) . Seasonal McArthur Forest Fire Danger Index data for Australia41 is available from [https://data.mendeley.com/datasets/xf5bv3hcvw/1](https:/data.mendeley.com/datasets/xf5bv3hcvw/1) .\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eJain, P. \u003cem\u003eet al.\u003c/em\u003e Drivers and Impacts of the Record-Breaking 2023 Wildfire Season in Canada. \u003cem\u003eNat. Commun.\u003c/em\u003e 15, 6764 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJones, M. W. \u003cem\u003eet al.\u003c/em\u003e State of Wildfires 2023\u0026ndash;2024. \u003cem\u003eEarth Syst. Sci. Data\u003c/em\u003e 16, 3601\u0026ndash;3685 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKeeley, J. E. \u0026amp; Syphard, A. D. Large California wildfires: 2020 fires in historical context. \u003cem\u003eFire Ecol.\u003c/em\u003e 17, 22 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSquire, D. T. \u003cem\u003eet al.\u003c/em\u003e Likelihood of unprecedented drought and fire weather during Australia\u0026rsquo;s 2019 megafires. \u003cem\u003eNpj Clim. Atmospheric Sci.\u003c/em\u003e 4, 1\u0026ndash;12 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang, Z. \u003cem\u003eet al.\u003c/em\u003e Severe Global Environmental Issues Caused by Canada\u0026rsquo;s Record-Breaking Wildfires in 2023. \u003cem\u003eAdv. Atmospheric Sci.\u003c/em\u003e 41, 565\u0026ndash;571 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJones, M. W. \u003cem\u003eet al.\u003c/em\u003e Global and Regional Trends and Drivers of Fire Under Climate Change. \u003cem\u003eRev. Geophys.\u003c/em\u003e 60, e2020RG000726 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMadadgar, S., Sadegh, M., Chiang, F., Ragno, E. \u0026amp; AghaKouchak, A. Quantifying increased fire risk in California in response to different levels of warming and drying. \u003cem\u003eStoch. Environ. Res. Risk Assess.\u003c/em\u003e 34, 2023\u0026ndash;2031 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSenande-Rivera, M., Insua-Costa, D. \u0026amp; Miguez-Macho, G. Spatial and temporal expansion of global wildland fire activity in response to climate change. \u003cem\u003eNat. Commun.\u003c/em\u003e 13, 1208 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUnited Nations Environment Programme. Spreading like Wildfire: The Rising Threat of Extraordinary Landscape Fires, \u003cem\u003eA UNEP Rapid Response Assessment\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.unep.org/resources/report/spreading-wildfire-rising-threat-extraordinary-landscape-fires\u003c/span\u003e\u003cspan address=\"https://www.unep.org/resources/report/spreading-wildfire-rising-threat-extraordinary-landscape-fires\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Oldenborgh, G. J. \u003cem\u003eet al.\u003c/em\u003e Attribution of the Australian bushfire risk to anthropogenic climate change. \u003cem\u003eNat. Hazards Earth Syst. Sci.\u003c/em\u003e 1\u0026ndash;46 (2020) doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5194/nhess-2020-69\u003c/span\u003e\u003cspan address=\"10.5194/nhess-2020-69\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBlanchi, R., Lucas, C., Leonard, J. \u0026amp; Finkele, K. Meteorological conditions and wildfire-related houseloss in Australia. \u003cem\u003eInt. J. Wildland Fire\u003c/em\u003e 19, 914 (2010).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFletcher, M.-S., Romano, A., Connor, S., Mariani, M. \u0026amp; Maezumi, S. Y. Catastrophic Bushfires, Indigenous Fire Knowledge and Reframing Science in Southeast Australia. \u003cem\u003eFire\u003c/em\u003e 4, 61 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang, B. \u003cem\u003eet al.\u003c/em\u003e Extreme fire weather is the major driver of severe bushfires in southeast Australia. \u003cem\u003eSci. Bull.\u003c/em\u003e 67, 655\u0026ndash;664 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBoer, M. M. \u0026amp; de Dios, V. R. Unprecendent burn area of Australian mega forest fires. \u003cem\u003eNat. Clim. Change\u003c/em\u003e 10, 170 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFilkov, A. I., Ngo, T., Matthews, S., Telfer, S. \u0026amp; Penman, T. D. Impact of Australia\u0026rsquo;s catastrophic 2019/20 bushfire season on communities and environment. Retrospective analysis and current trends. \u003cem\u003eJ. Saf. Sci. Resil.\u003c/em\u003e 1, 44\u0026ndash;56 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQuiggin, J. Opinion: Australia is promising \u003cspan\u003e$\u003c/span\u003e2 billion for the fires. I estimate recovery will cost \u003cspan\u003e$\u003c/span\u003e100 billion. \u003cem\u003eCNN\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cnn.com/2020/01/10/perspectives/australia-fires-cost\u003c/span\u003e\u003cspan address=\"https://www.cnn.com/2020/01/10/perspectives/australia-fires-cost\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRead, P. \u0026amp; Denniss, R. With costs approaching \u003cspan\u003e$\u003c/span\u003e100 billion, the fires are Australia\u0026rsquo;s costliest natural disaster. \u003cem\u003eThe Conversation\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://theconversation.com/with-costs-approaching-100-billion-the-fires-are-australias-costliest-natural-disaster-129433\u003c/span\u003e\u003cspan address=\"http://theconversation.com/with-costs-approaching-100-billion-the-fires-are-australias-costliest-natural-disaster-129433\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrown, P. T. \u003cem\u003eet al.\u003c/em\u003e Climate warming increases extreme daily wildfire growth risk in California. \u003cem\u003eNature\u003c/em\u003e 621, 760\u0026ndash;766,766A (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJolly, W. M. \u003cem\u003eet al.\u003c/em\u003e Climate-induced variations in global wildfire danger from 1979 to 2013. \u003cem\u003eNat. Commun.\u003c/em\u003e 6, 7537 (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbram, N. J. \u003cem\u003eet al.\u003c/em\u003e Connections of climate change and variability to large and extreme forest fires in southeast Australia. \u003cem\u003eCommun. Earth Environ.\u003c/em\u003e 2, (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCanadell, J. G. \u003cem\u003eet al.\u003c/em\u003e Multi-decadal increase of forest burned area in Australia is linked to climate change. \u003cem\u003eNat. Commun.\u003c/em\u003e 12, 6921 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowdy, A. J. Climatological variability of fire weather in Australia. \u003cem\u003eJ. Appl. Meteorol. Climatol.\u003c/em\u003e 57, 221\u0026ndash;234 (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDi Virgilio, G. \u003cem\u003eet al.\u003c/em\u003e Climate Change Increases the Potential for Extreme Wildfires. \u003cem\u003eGeophys. Res. Lett.\u003c/em\u003e 46, 8517\u0026ndash;8526 (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowdy, A. J. \u003cem\u003eet al.\u003c/em\u003e Future changes in extreme weather and pyroconvection risk factors for Australian wildfires. \u003cem\u003eSci. Rep.\u003c/em\u003e 9, 1\u0026ndash;11 (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDupuy, J. \u003cem\u003eet al.\u003c/em\u003e Climate change impact on future wildfire danger and activity in southern Europe: a review. \u003cem\u003eAnn. For. Sci.\u003c/em\u003e 77, 1\u0026ndash;24 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEl Garroussi, S., Di Giuseppe, F., Barnard, C. \u0026amp; Wetterhall, F. Europe faces up to tenfold increase in extreme fires in a warming climate. \u003cem\u003eNpj Clim. Atmospheric Sci.\u003c/em\u003e 7, 1\u0026ndash;11 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHe, Y., Zhou, Z., Im, E.-S. \u0026amp; Kwon, H.-H. Wildfire risk in a changing climate: Evaluating fire weather indices and their global patterns with CMIP6 multi-model projections. \u003cem\u003eWeather Clim. Extrem.\u003c/em\u003e 48, 100751 (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePimont, F. \u003cem\u003eet al.\u003c/em\u003e Future expansion, seasonal lengthening and intensification of fire activity under climate change in southeastern France. \u003cem\u003eInt. J. Wildland Fire\u003c/em\u003e 32, 4\u0026ndash;14 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSingo, M. V. \u003cem\u003eet al.\u003c/em\u003e Projections of future fire risk under climate change over the South African savanna. \u003cem\u003eStoch. Environ. Res. Risk Assess.\u003c/em\u003e 37, 2677\u0026ndash;2691 (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcArthur, A. G. Fire behaviour in Eucalypt forests. in \u003cem\u003eLeaf. No 107\u003c/em\u003e (Forestry and Timber Bureau, Canberra, ACT, 1967).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVan Wagner, C. E. Development and structure of the Canadian Forest Fire Weather Index System, Report No. 35. \u003cem\u003eCanadian Forestry Service\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ostrnrcan-dostrncan.canada.ca/handle/1845/228434\u003c/span\u003e\u003cspan address=\"https://ostrnrcan-dostrncan.canada.ca/handle/1845/228434\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (1987).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFox-Hughes, P. \u003cem\u003eet al.\u003c/em\u003e A review of early severe weather applications of high-resolution regional reanalysis in Australia. \u003cem\u003eMeteorol. Appl.\u003c/em\u003e 29, e2087 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowdy, A. J., Mills, G. A., Finkele, K. \u0026amp; de Groot, W. Australian fire weather as represented by the McArthur Forest Fire Danger Index and the Canadian Forest Fire Weather Index, \u003cem\u003eThe Centre for Australian Weather and Climate Research\u003c/em\u003e. (2009).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eClarke, H. \u0026amp; Evans, J. P. Exploring the future change space for fire weather in southeast Australia. \u003cem\u003eTheor. Appl. Climatol.\u003c/em\u003e 136, 513\u0026ndash;527 (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFox-Hughes, P., Harris, R., Lee, G., Grose, M. \u0026amp; Bindoff, N. Future fire danger climatology for Tasmania, Australia, using a dynamically downscaled regional climate model. \u003cem\u003eInt. J. Wildland Fire\u003c/em\u003e 23, 309\u0026ndash;321 (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowdy, A. J. Seamless climate change projections and seasonal predictions for bushfires in Australia. \u003cem\u003eJ. South. Hemisphere Earth Syst. Sci.\u003c/em\u003e 70, 120\u0026ndash;138 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcAneney, J., Chen, K. \u0026amp; Pitman, A. 100-years of Australian bushfire property losses: Is the risk significant and is it increasing? \u003cem\u003eJ. Environ. Manage.\u003c/em\u003e 90, 2819\u0026ndash;2822 (2009).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIPCC. Global Warming of 1.5\u0026deg;C: IPCC Special Report on Impacts of Global Warming of 1.5\u0026deg;C above Pre-industrial Levels in Context of Strengthening Response to Climate Change, Sustainable Development, and Efforts to Eradicate Poverty, \u003cem\u003eCambridge University Press\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cambridge.org/core/product/identifier/9781009157940/type/book\u003c/span\u003e\u003cspan address=\"https://www.cambridge.org/core/product/identifier/9781009157940/type/book\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSu, C.-H. \u003cem\u003eet al.\u003c/em\u003e BARRA2: Development of the next-generation Australian regional atmospheric reanalysis, Report No. 067. \u003cem\u003eBureau of Meteorology\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.bom.gov.au/research/publications/researchreports/BRR-067.pdf\u003c/span\u003e\u003cspan address=\"http://www.bom.gov.au/research/publications/researchreports/BRR-067.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePalmer, G. \u003cem\u003eet al.\u003c/em\u003e Validation of BARRA2 and comparison with MERRA-2 and ERA5 using historical wind power generation. \u003cem\u003eJ. South. Hemisphere Earth Syst. Sci.\u003c/em\u003e 75, ES24028 (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHarris, S. \u0026amp; Lucas, C. Understanding the variability of Australian fire weather between 1973 and 2017. \u003cem\u003ePLOS ONE\u003c/em\u003e 14, e0222328 (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJakob, D. Challenges in developing a high-quality surface wind-speed data-set for Australia. \u003cem\u003eAust. Meteorol. Oceanogr. J.\u003c/em\u003e 60, 227\u0026ndash;236 (2010).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLucas, C. On developing a historical fire weather data-set for Australia. \u003cem\u003eAust. Meteorol. Oceanogr. J.\u003c/em\u003e 61, 1\u0026ndash;13 (2010).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eClarke, H., Lucas, C. \u0026amp; Smith, P. Changes in Australian fire weather between 1973 and 2010. \u003cem\u003eInt. J. Climatol.\u003c/em\u003e 33, 931\u0026ndash;944 (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRussell-Smith, J. \u0026amp; Yates, C. P. Australian Savanna Fire Regimes: Context, Scales, Patchiness. \u003cem\u003eFire Ecol.\u003c/em\u003e 3, 48\u0026ndash;63 (2007).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcArthur, A. G. Weather and grassland fire behaviour. in \u003cem\u003eLeaf. No 100\u003c/em\u003e (Forestry and Timber Bureau, Canberra, ACT, 1966).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRichardson, D. \u003cem\u003eet al.\u003c/em\u003e Increased extreme fire weather occurrence in southeast Australia and related atmospheric drivers. \u003cem\u003eWeather Clim. Extrem.\u003c/em\u003e 34, 100397 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLindenmayer, D. B. \u0026amp; Taylor, C. New spatial analyses of Australian wildfires highlight the need for new fire, resource, and conservation policies. \u003cem\u003eProc. Natl. Acad. Sci.\u003c/em\u003e 117, 12481\u0026ndash;12485 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBowman, D. M. J. S., Kolden, C. A. \u0026amp; Williamson, G. J. Bushfires in Tasmania, Australia: An Introduction. \u003cem\u003eFire\u003c/em\u003e 5, 33 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHarris, R. M. B., Remenyi, T., Fox-Hughes, P., Love, P. \u0026amp; Bindoff, N. L. Exploring the Future of Fuel Loads in Tasmania, Australia: Shifts in Vegetation in Response to Changing Fire Weather, Productivity, and Fire Frequency. \u003cem\u003eForests\u003c/em\u003e 9, 210 (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowdy, A. \u0026amp; King, A. Novel climate analysis methods applied to the Australian ESCI projections data. \u003cem\u003eFront. Clim.\u003c/em\u003e 6, (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTrancoso, R. \u003cem\u003eet al.\u003c/em\u003e Converting tropical forests to agriculture increases fire risk by fourfold. \u003cem\u003eEnviron. Res. Lett.\u003c/em\u003e 17, 104019 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLafferty, D. C. \u0026amp; Sriver, R. L. Downscaling and bias-correction contribute considerable uncertainty to local climate projections in CMIP6. \u003cem\u003eNpj Clim. Atmospheric Sci.\u003c/em\u003e 6, 1\u0026ndash;13 (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWootten, A., Terando, A., Reich, B. J., Boyles, R. P. \u0026amp; Semazzi, F. Characterizing sources of uncertainty from global climate models and downscaling techniques. \u003cem\u003eJ. Appl. Meteorol. Climatol.\u003c/em\u003e 56, 3245\u0026ndash;3262 (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNarsey, S. \u003cem\u003eet al.\u003c/em\u003e Disentangling the uncertainties in regional projections for Australia. \u003cem\u003eJ. South. Hemisphere Earth Syst. Sci.\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.publish.csiro.au/ES/justaccepted/ES25015\u003c/span\u003e\u003cspan address=\"https://www.publish.csiro.au/ES/justaccepted/ES25015\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (in press).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBalbi, M. \u0026amp; Lallemant, D. C. B. The Cost of Imperfect Knowledge: How Epistemic Uncertainties Influence Flood Hazard Assessments. \u003cem\u003eWater Resour. Res.\u003c/em\u003e 59, e2023WR035685 (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEccles, R. \u003cem\u003eet al.\u003c/em\u003e Substantial increases in future precipitation extremes\u0026mdash;insights from a large ensemble of downscaled CMIP6 models. \u003cem\u003eNpj Nat. Hazards\u003c/em\u003e 2, 60 (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi, C. \u003cem\u003eet al.\u003c/em\u003e Changes in Annual Extremes of Daily Temperature and Precipitation in CMIP6 Models. \u003cem\u003eJ. Clim.\u003c/em\u003e 34, (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDCCEEW. National Vegetation Information System - Present Major Vegetation Groups and Subgroups - Version 7.0. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://fed.dcceew.gov.au/datasets/5e70b5afc\u003c/span\u003e\u003cspan address=\"https://fed.dcceew.gov.au/datasets/5e70b5afc\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e36a4c458a2cceb313eb3889/about (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDigital Atlas of Australia. Bushfire Historical Extents. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://digital.atlas.gov.au/datasets/912492b1683e494e8efbe84573ce6d70/about\u003c/span\u003e\u003cspan address=\"https://digital.atlas.gov.au/datasets/912492b1683e494e8efbe84573ce6d70/about\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChapman, S. \u003cem\u003eet al.\u003c/em\u003e Evaluation of Dynamically Downscaled CMIP6-CCAM Models Over Australia. \u003cem\u003eEarths Future\u003c/em\u003e 11, e2023EF003548 (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTrancoso, R., Syktus, J., Toombs, N. \u0026amp; Chapman, S. Assessing and selecting CMIP6 GCMs ensemble runs based on their ability to represent historical climate and future climate change signal, Report No. EGU23-11412. \u003cem\u003eCopernicus Meetings\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://meetingorganizer.copernicus.org/EGU23/EGU23-11412.html\u003c/span\u003e\u003cspan address=\"https://meetingorganizer.copernicus.org/EGU23/EGU23-11412.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eThatcher, M. Scientific description - CCAM. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://research.csiro.au/ccam/\u003c/span\u003e\u003cspan address=\"https://research.csiro.au/ccam/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcGregor, J. L. \u0026amp; Dix, M. R. An updated description of the conformal-cubic atmospheric model. in \u003cem\u003eHigh Resolution Numerical Modelling of the Atmosphere and Ocean\u003c/em\u003e 51\u0026ndash;75 (Springer New York, 2008).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHoffmann, P., Katzfey, J. J., McGregor, J. L. \u0026amp; Thatcher, M. Bias and variance correction of sea surface temperatures used for dynamical downscaling. \u003cem\u003eJ. Geophys. Res.\u003c/em\u003e 121, 12,877\u0026thinsp;\u0026ndash;\u0026thinsp;12,890 (2016).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChapman, S., Syktus, J., Trancoso, R., Toombs, N. \u0026amp; Eccles, R. Projected changes in mean climate and extremes from downscaled high-resolution CMIP6 simulations in Australia. \u003cem\u003eWeather Clim. Extrem.\u003c/em\u003e 46, 100733 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim, Y., Rocheta, E., Evans, J. P. \u0026amp; Sharma, A. Impact of bias correction of regional climate model boundary conditions on the simulation of precipitation extremes. \u003cem\u003eClim. Dyn.\u003c/em\u003e 55, 3507\u0026ndash;3526 (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLim, C. M., Yhang, Y. B. \u0026amp; Ham, S. Application of GCM Bias Correction to RCM Simulations of East Asian Winter Climate. \u003cem\u003eAtmosphere\u003c/em\u003e 10, 382 (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBeck, H. E. \u003cem\u003eet al.\u003c/em\u003e MSWX: Global 3-Hourly 0.1\u0026deg; Bias-Corrected Meteorological Data Including Near-Real-Time Updates and Forecast Ensembles. \u003cem\u003eBull. Am. Meteorol. Soc.\u003c/em\u003e 103, E710\u0026ndash;E732 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePuri, K. \u003cem\u003eet al.\u003c/em\u003e Implementation of the initial ACCESS numerical weather prediction system. \u003cem\u003eAust. Meteorol. Oceanogr. J.\u003c/em\u003e 63, 265\u0026ndash;284 (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLucas, C. Seasonal McArthur Forest Fire Danger Index (FFDI) data for Australia: 1973\u0026ndash;2017. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.mendeley.com/datasets/xf5bv3hcvw/1\u003c/span\u003e\u003cspan address=\"https://data.mendeley.com/datasets/xf5bv3hcvw/1\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang, H., Chapman, S., Trancoso, R., Toombs, N. \u0026amp; Syktus, J. Assessing the impact of bias correction approaches on climate extremes and the climate change signal. \u003cem\u003eMeteorol. Appl.\u003c/em\u003e 31, e2204 (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDowdy, A. A bias correction method designed for weather and climate extremes, Report No. 087. \u003cem\u003eAustralian Bureau of Meteorology\u003c/em\u003e. (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePerkins, S. E., Pitman, A. J., Holbrook, N. J. \u0026amp; McAneney, J. Evaluation of the AR4 climate models\u0026rsquo; simulated daily maximum temperature, minimum temperature, and precipitation over Australia using probability density functions. \u003cem\u003eJ. Clim.\u003c/em\u003e 20, 4356\u0026ndash;4376 (2007).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNoble, I. R., Gill, A. M. \u0026amp; Bary, G. a. V. McArthur\u0026rsquo;s fire-danger meters expressed as equations. \u003cem\u003eAust. J. Ecol.\u003c/em\u003e 5, 201\u0026ndash;203 (1980).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLucas, C., Hennessy, K., Mills, G. \u0026amp; Bathols, J. Bushfire Weather in Southeast Australia: Recent Trends and Projected Climate Change Impacts, \u003cem\u003eBushfire Cooperative Research Centre\u003c/em\u003e. (2007).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKeetch, J. J. \u0026amp; Byram, G. M. A Drought Index for Forest Fire Control. \u003cem\u003eUS Dep. Agric. For. Serv. Res. Pap.\u003c/em\u003e 038, (1968).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGriffiths, D. Improved Formula for the Drought Factor in McArthur\u0026rsquo;s Forest Fire Danger Meter. \u003cem\u003eAust. For.\u003c/em\u003e 62, 202\u0026ndash;206 (1999).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTebaldi, C. \u003cem\u003eet al.\u003c/em\u003e Climate model projections from the Scenario Model Intercomparison Project (ScenarioMIP) of CMIP6. \u003cem\u003eEarth Syst. Dyn.\u003c/em\u003e 12, 253\u0026ndash;293 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIPCC. Climate Change 2021 \u0026ndash; The Physical Science Basis: Working Group I Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, \u003cem\u003eCambridge University Press\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cambridge.org/core/books/climate-change-2021-the-physical-science-basis/415F29233B8BD\u003c/span\u003e\u003cspan address=\"https://www.cambridge.org/core/books/climate-change-2021-the-physical-science-basis/415F29233B8BD\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e19FB55F65E3DC67272B (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKrikken, F., Lehner, F., Haustein, K., Drobyshev, I. \u0026amp; van Oldenborgh, G. J. Attribution of the role of climate change in the forest fires in Sweden 2018. \u003cem\u003eNat. Hazards Earth Syst. Sci.\u003c/em\u003e 21, 2169\u0026ndash;2179 (2021).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [npj Natural Hazards](https://www.nature.com/npjnathazards/)","snPcode":"44304","submissionUrl":"https://submission.springernature.com/new-submission/44304/3","title":"npj Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Extreme fire weather, FFDI, Downscaled climate projections, Regional climate modelling, Eucalyptus forests, Global Warming Levels, Generalised Extreme Value distribution","lastPublishedDoi":"10.21203/rs.3.rs-8051769/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8051769/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWe evaluated the influence of climate change on fire weather in Australia using the McArthur Forest Fire Danger Index (FFDI) and climate simulations from an ensemble of dynamically downscaled CMIP6 projections. Rare extreme FFDI events were assessed under a range of Global Warming Levels (GWLs) using the Generalised Extreme Value (GEV) distribution, with a focus on southeast Australia\u0026rsquo;s wildfire prone eucalyptus forests. The magnitude and frequency of extreme FFDI events are projected to increase substantially, particularly in southern Australia. For the eucalyptus forests of southeast Australia, 20-year and 100-year return interval 7-day FFDI events (i.e. weekly average FFDI extremes) are projected to become approximately 2.1 and 3.0 times more likely under 3 \u003csup\u003e◦\u003c/sup\u003eC of global warming. The increases are most notable for the eucalyptus forests of Tasmania, with projected 20-year and 100-year return interval 7-day FFDI events becoming approximately 3.2 and 5.2 times more likely for 3 \u003csup\u003e◦\u003c/sup\u003eC GWL.\u003c/p\u003e","manuscriptTitle":"Substantial increases in the likelihood of extreme fire weather events for fire-prone ecosystems in Australia","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-24 10:53:44","doi":"10.21203/rs.3.rs-8051769/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-27T00:23:38+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-26T20:32:15+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-26T06:07:39+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-22T05:56:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"204298721094674629557593902906373320825","date":"2025-11-17T05:13:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"300802015037150897006161715526816016476","date":"2025-11-17T03:24:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"183515844847741383132584943855798158223","date":"2025-11-16T13:09:49+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"276564713501200887834848371975546547792","date":"2025-11-16T10:06:23+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"250706154217689605770713773367546572413","date":"2025-11-12T07:38:02+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-12T07:30:46+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-12T06:51:09+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-12T05:49:53+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Natural Hazards","date":"2025-11-07T00:07:43+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [npj Natural Hazards](https://www.nature.com/npjnathazards/)","snPcode":"44304","submissionUrl":"https://submission.springernature.com/new-submission/44304/3","title":"npj Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c150bc76-c2bb-46e7-8055-c99f5d510836","owner":[],"postedDate":"November 24th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":58424279,"name":"Earth and environmental sciences/Climate sciences"},{"id":58424280,"name":"Biological sciences/Ecology"},{"id":58424281,"name":"Earth and environmental sciences/Ecology"},{"id":58424282,"name":"Earth and environmental sciences/Environmental sciences"},{"id":58424283,"name":"Earth and environmental sciences/Natural hazards"}],"tags":[],"updatedAt":"2026-02-22T13:54:19+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-24 10:53:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8051769","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8051769","identity":"rs-8051769","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}