Wildfires Earth Systems Atmospheric Chemistry & Climate Feedback

~4.5 million km² burned globally per year — an area larger than the European Union Fire emissions: ~2.0–2.5 Gt CO₂/yr plus black carbon, PM2.5, ozone precursors, and hundreds of VOCs Sources: GFED4, NASA FIRMS, IPCC AR6, WHO, Copernicus CAMS, USFS, Global Wildfire Information System
~4.5M km²
Global average annual burned area
Satellite era (GFED4, 1997–present); area larger than EU-27 burned every year
~2.2 Gt CO₂/yr
Mean annual fire CO₂ emissions
~6% of total global fossil fuel CO₂; net effect depends on regrowth carbon uptake
~339,000
Premature deaths per year from fire smoke
Primarily PM2.5 exposure; sub-Saharan Africa and Southeast Asia bear greatest burden
≥420 Ma
Age of Earth's oldest charcoal evidence
Charcoalified plant tissue from Silurian Period; fire co-evolved with terrestrial plant life
10,000+ km
Maximum smoke transport distance
Australian 2019–20 smoke reached South America; circumnavigated the globe in 3 weeks
+25%
Increase in extreme fire weather days since 1980
Longer fire seasons on every continent; driven by warming, drought, and wind changes

★ Fire: Earth's Ancient Recycler and Modern Climate Driver

Fire is not a disturbance to Earth's ecosystems — it is one of their fundamental processes. Combustion has shaped landscapes, atmospheric composition, and evolutionary trajectories for hundreds of millions of years. The Silurian and Devonian periods saw the first charcoal preserved in geological strata, testament to wildland fires burning across early vascular plant communities. By the Carboniferous (359–299 Ma), when atmospheric oxygen peaked near 35% (vs. 21% today), fires burned so frequently and intensely that coal seams of that era are riddled with fossilised charcoal — "inertinite" — accounting for up to 70% of some coal deposits.

In the modern era, the interplay between wildfire and climate has intensified dramatically. Global warming drives longer droughts, earlier snowmelt, reduced soil moisture, and more extreme wind events — all of which prime landscapes for catastrophic fire. At the same time, fire emissions inject greenhouse gases, black carbon, and ozone precursors into the atmosphere, creating a reinforcing feedback loop. The 2019–20 Australian "Black Summer" fires emitted approximately 0.9 Gt CO₂ — more than Australia's entire annual anthropogenic emissions — while blanketing New Zealand in smoke and measurably increasing global mean CO₂ for several months. The 2023 Canadian wildfire season broke all records, burning 18.4 million hectares and sending smoke across the eastern United States and Atlantic Ocean.

Unlike volcanic eruptions or industrial emissions, wildfires straddle a complex carbon accounting question: because forests regrow and reabsorb CO₂ over decades, fires in stable ecosystems are roughly carbon-neutral on a centennial timescale. But when climate change-driven fires kill forests that do not regrow — converting boreal forest to grassland or shrubland, or killing tropical forest that turns to savanna — the carbon balance shifts permanently toward net emissions. This "committed" carbon loss is one of the most concerning and least-modelled aspects of the current wildfire crisis.

Global Burned Area — Satellite Era 1997–2023 (GFED4)

Source: Global Fire Emissions Database v4s (GFED4s), van der Werf et al. 2017 updated; NASA FIRMS MODIS/VIIRS active fire data; Copernicus CAMS Global Fire Assimilation System (GFAS).

Fire Emissions by Region — Annual CO₂ Equivalent

Source: GFED4s regional emissions by biome; van der Werf et al. 2017; Giglio et al. 2013. Regions: NHAF=Northern Hemisphere Africa, SHAF=Southern Hemisphere Africa, BONA=Boreal North America, TENA=Temperate North America, CEAM=Central America, NHSA=Northern Hemisphere South America, SHSA=Southern Hemisphere South America, EURO=Europe, MIDE=Middle East, BOAS=Boreal Asia, CEAS=Central Asia, SEAS=Southeast Asia, EQAS=Equatorial Asia, AUST=Australia.
The 2023 Canadian Fire Season — A New Benchmark: In 2023, Canadian wildfires burned 18.4 million hectares (184,000 km²) — nearly 7× the 10-year average and more than any previously recorded season. Smoke from these fires reached AQI "hazardous" levels across New York City, Washington D.C., and Philadelphia — cities that typically experience some of the best air quality in the United States. More than 200,000 Canadians were evacuated. The fires emitted an estimated 0.48 Gt CO₂, roughly equivalent to Canada's entire annual fossil fuel output. Satellite imagery tracked smoke plumes crossing the Atlantic and reaching northern Europe within 10 days.
Why Burned Area Is Decreasing Globally While Fire Risk Is Increasing: Paradoxically, total global burned area has declined slightly since 2000 in satellite-era records — primarily because of land use change in tropical Africa and South America, where agricultural expansion has fragmented fire-prone savanna and cerrado. However, the intensity of fires in mid-latitude and boreal forests has increased dramatically. The fires that matter most for climate, carbon, and health — large, intense, long-duration forest fires — are becoming more frequent and severe on every continent. Lower overall burned area masks a sharp increase in forest-type fires that emit far more carbon per hectare than savanna fires.

Deep Time: Fire Through 420 Million Years

The history of wildfire is inseparable from the history of atmospheric oxygen and terrestrial plant life. Fire requires three elements — fuel, heat, and oxygen — and all three varied dramatically over geological time. The paleofire record is read primarily through charcoal (known as "inertinite" in coal, or "fusain" in geological literature) preserved in sedimentary rocks. These charcoal fragments are essentially fire's fossil record, surviving diagenesis for hundreds of millions of years.

Silurian-Devonian (~420–360 Ma): The first convincing charcoal evidence appears in rocks ~420 million years old, coinciding with the colonisation of land by vascular plants. Early fires were likely small, fuelled by low-growing mosses and early tracheophytes. As larger plants evolved — ferns, seed ferns, early trees — fuel loads increased and fires became more widespread. Atmospheric O₂ at this time was close to modern levels (~18–21%).

Carboniferous (~359–299 Ma): This period witnessed Earth's most intense fire history. Atmospheric O₂ rose to ~30–35%, driven by the massive burial of organic carbon that would eventually become coal. At these oxygen levels, wet vegetation could ignite and fire spread through even humid environments. Inertinite content in Carboniferous coal measures averages 10–20% and can reach 70% in some seams — direct evidence of pervasive wildfire. The evolution of "fire-resistant" bark and seed coats in Carboniferous plants reflects evolutionary pressure from chronic fire.

Permian–Triassic (~299–200 Ma): Declining O₂ toward the Permian-Triassic boundary (~252 Ma, Earth's worst mass extinction) reduced fire frequency. The Triassic recovery involved sporadic high-fire episodes punctuated by low-oxygen intervals. The Permian-Triassic extinction itself was associated with Siberian Traps volcanism and catastrophic warming — not primarily fire — but fire played a role in post-extinction landscape restructuring.

Cretaceous (~145–66 Ma): Warm, moist conditions and high plant productivity made the Cretaceous a moderate fire period. The end-Cretaceous impact (Chicxulub, 66 Ma) ignited global wildfires on a scale never since repeated — some models suggest the impact-related fires consumed much of Earth's above-ground biomass in a matter of months. The global soot layer in K-Pg boundary sediments is direct evidence of this event.

Cenozoic (66 Ma–present): Grassland expansion from ~25 Ma onward created fire-adapted ecosystems (savanna, prairie, steppe) that transformed global fire regimes. C4 grasses that dominate modern savannas evolved alongside fire, using dry-season burns to outcompete trees. Human fire use added a transformative new ignition source approximately 1 million years ago (Homo erectus), and deliberate landscape burning by modern humans reshaped fire regimes globally beginning ~70,000–100,000 years ago.

Paleofire Activity — Atmospheric O₂ & Charcoal Record (400 Ma – Present)

Source: Belcher & McElwain 2008; Scott & Glasspool 2006; Berner 2004 GEOCARB III O₂ model; Marynowski & Filipiak 2007; Glasspool et al. 2015; He et al. 2021.

Historical Megafires — Notable Events in the Instrumental & Historical Record

YearEventArea (Mha)Est. CO₂Deaths
1871Peshtigo Fire, Wisconsin, USA0.5~1,500–2,500 (deadliest US fire)
1910The Big Blowup, Northern Rockies, USA1.285 firefighters; shaped US fire policy for a century
1915–1916Taiga fires, Siberia (Russian Far East)~12Minimal human toll; enormous ecological impact
1987Black Dragon Fire, China/USSR7.3~0.3 Gt~200; largest 20th-century forest fire
1997–98Indonesia peatland fires (El Niño)~2.5~0.8–2.6 Gt~500 directly; 100,000+ from haze exposure
2003Siberian fires~2.4~0.2 Gt
2010Russia steppe & forest fires~1.0~0.06 Gt~55,000 heat-related deaths linked to smoke
2015–16Indonesia peatland fires (El Niño)~2.6~0.9–1.5 Gt~100,000 excess deaths from haze
2019–20Australian "Black Summer"18.6~0.9 Gt~33 direct; ~417 smoke-related; 3B animals affected
2020Western United States fires1.7~0.13 Gt33 direct; widespread smoke AQI impacts
2021Siberia / Yakutia fires~1.8~0.97 Gt CO₂eq—; largest Siberian fire season on record at the time
2023Canadian wildfire season18.4~0.48 GtWidespread evacuations; record AQI impacts on US East Coast
Source: GFED4s; Cochrane 2003; Page et al. 2002; van der Werf et al. 2008; Kiely et al. 2021; CWFIS Canada; NSW RFS; CalFire; Copernicus CAMS.
Human Fire Use — From Homo erectus to Deliberate Landscape Burning: Evidence of controlled fire use by human ancestors extends back at least 1 million years at Wonderwerk Cave (South Africa) and ~400,000 years at multiple European sites. But deliberate, large-scale landscape burning — fire as an agricultural and hunting tool — reshaped global fire regimes beginning with anatomically modern humans (~100,000–70,000 BP). Aboriginal Australians maintained a "fire-stick farming" tradition for at least 50,000 years, creating the mosaic of grassland and woodland visible in pre-colonial Australia. Indigenous burning practices in North America, Sub-Saharan Africa, and Amazonia similarly modified vegetation patterns still visible today. Colonial suppression of these practices, combined with 20th-century fire exclusion policies, created the "fire debt" of accumulated fuels now feeding megafires across North America and Australia.
Fire and Human Evolution: The control of fire may be the most consequential technology in human evolutionary history. Cooking with fire — a behaviour traceable to at least 780,000 years ago at Gesher Benot Ya'aqov (Israel) and possibly 1 million years ago — dramatically increased the energy yield of food, enabling the energy-intensive growth of the human brain. Richard Wrangham's "cooking hypothesis" argues that fire-cooked food was the caloric prerequisite for the evolution of Homo sapiens brain size. Warming, protection from predators, and extended social time around hearths also shaped human psychology and social structures. In this sense, wildfire was not only part of our evolutionary environment — we domesticated it, and it helped domesticate us.

Active Fire Detection — Current Season (NASA FIRMS)

Source: NASA Fire Information for Resource Management System (FIRMS) MODIS Collection 6.1 and VIIRS 375m active fire detections; Global Wildfire Information System (GWIS) — Copernicus Emergency Management Service. Data represent typical seasonal fire detection density (7-day rolling average, climatological mean 2003–2023).

Burned Area by Biome — Annual Average (GFED4s, 1997–2023)

Source: GFED4s (Global Fire Emissions Database v4 small fires included), van der Werf et al. 2017; Giglio et al. 2013. Biome categories: savanna/grassland includes tropical and temperate; forest includes boreal, temperate, and tropical; shrubland includes chaparral, fynbos, heath; peatland data from Huijnen et al. 2016.

Regional Fire Hotspots — Key Statistics by Region

RegionAvg Burned Area/yrPeak SeasonPrimary BiomeCO₂ Emissions/yrKey DriversTrend
Sub-Saharan Africa~2,000,000 km²Nov–MarSavanna / tropical woodland~0.9 GtSeasonal agricultural burning, pastoral traditions Stable / declining
South America (Amazon + Cerrado)~400,000 km²Jul–OctTropical forest, cerrado~0.35 GtDeforestation clearing fires, drought, El Niño Increasing
Boreal Asia (Siberia)~200,000 km²Jun–AugBoreal/taiga forest, tundra~0.25 GtRecord heat, permafrost-peat ignition, lightning Rapidly increasing
Boreal North America (Canada/Alaska)~100,000 km²Jun–AugBoreal forest~0.10 GtWarming, drought, fuel accumulation from fire suppression Rapidly increasing
Western United States~30,000 km²Jul–OctChaparral, mixed conifer, shrubland~0.04 GtDrought, heat, fuel load, WUI expansion, diablo winds Increasing
Southeast Asia (incl. Indonesia)~100,000 km²Aug–Nov (ENSO yr)Tropical forest, peatland~0.3–1.5 Gt (El Niño yr)Peatland drainage, oil palm expansion, El Niño Variable / high in El Niño
Australia~400,000 km²Nov–MarSclerophyll forest, eucalypt, spinifex~0.10 Gt (record yr: ~0.9 Gt)IOD, drought, heat, fuel accumulation Increasing in extreme years
Mediterranean Basin (EU + MENA)~5,000 km²Jun–SepMaquis, pine forest, garrigue~0.005 GtSummer drought, land abandonment, climate warming Increasing
Source: GFED4s regional time series; Copernicus CAMS GFAS; van der Werf et al. 2017; Global Wildfire Information System 2024; IPCC AR6 WG1 Chapter 12 regional assessments; EC JRC Global Forest Watch; WWF Fire Monitor.

Fire Perimeter Growth Rate — Extreme Events

Average wildfire (US)~0.5 ha/hr
Fast-moving surface fire1–5 ha/hr
Wind-driven chaparral fire10–50 ha/hr
2018 Camp Fire (Paradise, CA)~80 ha/hr at peak
Crown fire (extreme conditions)100–300+ ha/hr
Australian ember attack rangeUp to 35 km ahead of front
Source: Andrews 2018; USFS Fire Research; Gill et al. 2013; CAL FIRE incident reports.

Global Fire Season Length — Change Since 1979

Global average fire season+18.7% longer
Western North America+84 days longer
Mediterranean Europe+50 days longer
Southern Australia+46 days longer
Boreal Asia (Siberia)+30 days longer
Amazon / Brazil+24 days longer
Source: Jolly et al. 2015 (Nature Comms); Jones et al. 2022 (IPCC Atlas); Copernicus C3S FWI analysis 1979–2023.

Fire Weather — Key Indices

Fire Weather Index (FWI)Composite danger rating
Keetch-Byram Drought IndexSoil moisture deficit (0–800)
Energy Release Component (ERC)Available energy in fuel complex
Haines IndexAtmospheric instability + dryness
Fine Fuel Moisture Content (FFMC)1-hr fuel moisture (twigs, grasses)
100-hr / 1,000-hr fuel moistureLarger fuel response to humidity
Vapor Pressure Deficit (VPD)Most predictive modern metric
Source: USFS NFDRS; Canadian FWI System; Copernicus CAMS; Williams 2019 (drought-fire VPD relationship).

★ What Fire Releases: A Chemical Portrait of Wildfire Smoke

Wildfire smoke is one of the most chemically complex aerosol mixtures in the atmosphere — a shifting cocktail of thousands of compounds produced across a spectrum of combustion temperatures and fuel types. Understanding this chemistry is essential not only for human health, but for quantifying fire's impact on climate, atmospheric oxidation capacity, ozone formation, and the global carbon cycle.

Biomass combustion occurs in three phases: flaming combustion (temperatures 600–1,000°C, produces CO₂, NOx, black carbon, and ozone precursors); smouldering combustion (300–500°C, produces large amounts of CO, CH₄, particulate organic matter, and volatile organic compounds); and pyrolysis (200–400°C ahead of the fire front, releasing hundreds of VOCs including carcinogenic compounds). Each phase produces a distinct chemical signature, and real fires cycle through all three simultaneously across their perimeter.

A key metric is the Modified Combustion Efficiency (MCE) — the ratio of CO₂ to total carbon emitted. High MCE (≥0.95) indicates predominantly flaming combustion; low MCE (≤0.85) indicates smouldering, which produces far more CO, CH₄, and organic aerosols. Peatland fires have among the lowest MCE values (~0.80–0.88), explaining their outsized atmospheric chemistry impact relative to burned area.

Fire Emission Compounds — Relative Quantities and Climate/Health Significance

Source: Andreae & Merlet 2001; Akagi et al. 2011; Andreae 2019 fire emission factor update; GEOS-Chem global fire tracer transport; Keywood et al. 2013 (IGAC).

Fire Aerosol Optical Depth (AOD) — Radiative Forcing

Source: MODIS Terra/Aqua AOD Collection 6.1 (2001–2023); Copernicus CAMS reanalysis fire aerosol optical depth; AERONET sun photometer network validation; Liu et al. 2021 fire AOD climatology; Tosca et al. 2013.

Key Atmospheric Compounds Emitted by Wildfires — Detailed Chemistry

CompoundTypeAnnual Fire EmissionAtmospheric LifetimeClimate EffectHealth Effect
CO₂ (Carbon Dioxide)GHG ~2.0–2.5 Gt C/yrCenturies (net, via regrowth) Warming; net ~neutral if regrowth occurs; net positive if forest converted Minimal direct; indirect via climate change
CO (Carbon Monoxide)Indirect GHG / pollutant ~0.4–0.6 Gt C/yr~1–3 months Indirect warming: oxidises OH (→ prolongs CH₄ lifetime), produces tropospheric ozone Toxic at high concentrations; reduces O₂ carrying capacity of blood
CH₄ (Methane)GHG ~20–30 Tg CH₄/yr (~6% of global sources)~9–12 years GWP₁₀₀ = 28–34; significant warming contribution from peatland fires Not directly toxic at ambient concentrations
Black Carbon (BC) / Soot Absorbing aerosol ~3–5 Tg BC/yr (~40% of global BC)Days–1 week (troposphere) Strong direct warming (GWP₂₀ ~3,200); darkens snow/ice (albedo effect); net forcing +0.4 to +0.9 W/m² PM2.5; deep lung penetration; cardiovascular and respiratory disease; carcinogen
Organic Carbon (OC) / Brown Carbon (BrC) Scattering/absorbing aerosol ~30–40 Tg OC/yrDays–weeks Complex: OC primarily scatters (cooling); BrC absorbs UV (modest warming); total net slight cooling Major PM2.5 component; strong association with respiratory and cardiovascular mortality
NOx (Nitrogen Oxides) Ozone precursor ~5–8 Tg N/yrHours–days Produces tropospheric ozone (warming); in low-NOx environments may destroy ozone Respiratory irritant; contributes to smog formation; PM2.5 secondary formation
VOCs (Volatile Organic Compounds) Ozone/SOA precursors Hundreds of species; ~10–30 Tg NMHC/yrHours–weeks Ozone production; secondary organic aerosol (SOA) formation; complex radiative effects Many are toxic/carcinogenic (benzene, formaldehyde, acrolein, polyaromatic hydrocarbons)
PM2.5 (Fine Particulate Matter) Aerosol particle ~30–50 Tg PM2.5/yr (fires ~40% of global biomass burning PM2.5)Days–weeks Direct (scattering/absorption) and indirect (cloud condensation nuclei) radiative effects Leading fire health impact; 339,000 premature deaths/yr (Marlier et al.); penetrates alveoli; no safe threshold
N₂O (Nitrous Oxide) GHG / ozone depleter ~2–3 Tg N₂O/yr (~10% of global sources)~120 years GWP₁₀₀ = 265; destroys stratospheric ozone; long-term climate impact Not directly toxic at ambient fire concentrations
Ozone (O₃) — secondary Secondary pollutant / GHG Formed in plumes from NOx + VOC + UVHours–months (depending on altitude) Tropospheric O₃ is 3rd largest GHG forcing; stratospheric O₃ depletion from N₂O Respiratory damage; reduces lung function; agricultural crop damage (yield loss)
Mercury (Hg) Toxic heavy metal ~175–800 Mg Hg/yr; fires ~30% of natural Hg sources6–12 months (elemental Hg) Deposited to oceans → methylmercury in marine food chains Neurotoxin; bioaccumulates in fish; pregnancy risk; linked to high-latitude boreal fire events
Source: Andreae 2019 (updated emission factor database); Akagi et al. 2011; Kroll et al. 2015; Liu et al. 2017; IPCC AR6 Chapter 6 (Short-Lived Climate Forcers); WHO Air Quality Guidelines 2021; Marlier et al. 2013 (premature deaths); Obrist et al. 2018 (mercury).
Black Carbon — Fire's Hidden Climate Agent: Black carbon (BC), commonly called "soot," is produced during the flaming phase of combustion and is the most potent light-absorbing aerosol in the atmosphere. Biomass burning (wildfires + agricultural burning) contributes roughly 40% of global BC emissions. BC's direct radiative forcing — absorbing solar radiation in the atmosphere — is estimated at +0.4 to +0.9 W/m², making it the second or third most powerful climate forcing agent after CO₂ and CH₄. When black carbon deposits on snow and ice (glaciers, Arctic sea ice, Greenland ice sheet), it dramatically reduces albedo, accelerating melting. Fire events in boreal Canada and Siberia deposit measurable BC on Arctic snow within days. BC's short atmospheric lifetime (days to weeks in the troposphere) means that reducing fire frequency could provide rapid, meaningful climate benefits on timescales far shorter than CO₂ mitigation.
Pyroconvection — When Fire Creates Its Own Weather: The most intense wildfires generate sufficient heat to create self-sustaining convective columns — pyrocumulus (PyroCu) and pyrocumulonimbus (PyroCb) clouds that can penetrate the tropopause and inject smoke, black carbon, and water vapour directly into the stratosphere. PyroCb events were documented during the 2019–20 Australian fires, which injected smoke plumes to 35 km altitude — deep into the stratosphere, well above the typical injection height of even VEI 4 volcanic eruptions. Stratospheric smoke aerosols have an atmospheric lifetime of months to years (similar to volcanic aerosols), and the Australian PyroCb smoke caused measurable global radiative perturbations, a temporary reduction in stratospheric ozone, and observable stratospheric heating detected by satellite. Fire-generated stratospheric injection is now recognised as a non-volcanic pathway for stratospheric aerosol loading.
Ozone Chemistry in Fire Plumes — A Complex Story: Fire smoke simultaneously produces and destroys ozone depending on conditions. In the immediate fire plume and in NOx-rich environments, VOC + NOx + UV photochemistry generates large quantities of tropospheric ozone (a greenhouse gas and respiratory irritant). Downwind of fires, models and observations show ozone enhancements of 10–80 ppb in aged smoke plumes — well above the EPA 8-hour standard of 70 ppb. However, in remote, low-NOx environments (open ocean, polar regions), fire smoke can actually destroy ozone through halogen chemistry and HOx reactions. The 2019–20 Australian fires were associated with the largest Antarctic ozone depletion event in 20 years, partly attributed to stratospheric PyroCb smoke that activated halogen chemistry on aerosol surfaces.

★ The Fire-Climate Feedback Loop

Wildfire and climate are locked in a reinforcing feedback cycle that is accelerating as the planet warms. Rising temperatures drive increased evapotranspiration (moisture loss from soil and vegetation), more frequent and severe droughts, earlier snowmelt (reducing summer soil moisture), and more extreme heat events — all of which prime landscapes for catastrophic fire. When those fires burn, they release stored carbon, inject black carbon into the atmosphere, and remove the forest cover that previously provided local cooling, moisture cycling, and albedo. This creates conditions that make subsequent fires more likely — a self-reinforcing cycle with no natural brake.

The boreal zone is of particular concern. Permafrost soils beneath boreal and tundra ecosystems contain an estimated 1,500 Gt of carbon — nearly twice the current atmospheric carbon stock. As warming thaws permafrost and fire frequencies increase in Siberia and Alaska, this carbon becomes available for combustion and microbial decomposition. The 2021 Siberian fire season emitted an estimated 970 Mt CO₂-equivalent — and much of that came from peat and permafrost-underlain soils that do not regrow their carbon stores for centuries. This is the definition of a committed, irreversible emission.

Fire-Climate Feedback Pathways

Source: Bowman et al. 2009 (Science); Flannigan et al. 2013; Abatzoglou & Williams 2016 (climate change doubled burned area in western US); Veraverbeke et al. 2017 (lightning ignitions in boreal); Walker et al. 2019 (tundra fire); IPCC AR6 Chapter 5 (carbon feedbacks).

Warming → Fire → Warming: Quantified Feedbacks

Climate change contribution to western US fire 1984–2015+2× burned area
Fraction of global warming attributable to fire feedbacks~0.1–0.4°C by 2100 (SSP5-8.5)
Projected increase in global burned area by 2100 (2°C warming)+25% to +50%
Projected increase at 4°C warming+60% to +100%+
Siberian permafrost carbon potentially mobilised by fire by 21005–60 Gt CO₂eq (wide uncertainty)
Black Summer Australia: fire-driven SST anomaly (Indian Ocean)+0.08°C (aerosol dimming effect)
Lightning ignitions in boreal: change per °C warming+12% per °C (US CONUS)
Source: Abatzoglou & Williams 2016; Knorr et al. 2016; Settele et al. IPCC SR1.5; Veraverbeke et al. 2017; Flannigan et al. 2013; Tosca et al. 2013; IPCC AR6 WG1 Ch.5.

Amplifying Feedbacks by System

Boreal / Permafrost

Fire removes insulating moss and organic layer → soil warming → permafrost thaw → CO₂ and CH₄ release from microbial decomposition → more warming. Return fire interval in Alaska has shortened from ~150 years to ~80 years since 1980. Reburn of recently-burned areas (which have abundant dead fuel) is increasing.

Tropical Forest (Amazon)

Deforestation + drought + fire create "savannification" — irreversible conversion of rainforest to savanna. Amazon deforestation is estimated to have pushed 15–17% of the forest past its local tipping point already. Once savannified, areas are drier, hotter, and more fire-prone — precluding rainforest re-establishment even if fire is excluded.

Mid-latitude Forest (Western US)

Bark beetle outbreaks (driven by warm winters) kill trees → dead fuel accumulation → catastrophic fire → type conversion to shrubland. Post-fire conifer regeneration is failing across much of the American West due to warmer, drier conditions — converting fire-cleared areas to persistent shrub or grass states with higher fire frequency.

The "Zombie Fires" of the Arctic: A newly documented fire phenomenon — overwintering fires or "zombie fires" — has been observed in Siberia and Alaska since the early 2010s. These fires smoulder underground in peat deposits through the Arctic winter (at temperatures as low as −40°C), surviving as ember sources beneath snow cover, then re-emerge in spring to ignite the following season's fire. In 2020, the Siberian fires that blazed through summer were partly re-ignitions of 2019 peat fires that had survived the winter. As Arctic warming continues — the Arctic is warming 4× faster than the global mean — the conditions for zombie fire persistence are expanding. Overwintering fires bypass the normal spring ignition window, allowing earlier and more extensive fire seasons.
Does Fire Feedback Appear in IPCC Models? IPCC AR6 includes fire-carbon feedbacks in Earth System Models (ESMs), but with substantial uncertainty. Most ESMs represent fire through simplified fire modules (SPITFIRE, JSBACH-SPITFIRE, CLM-Li) that capture broad burned area trends but underestimate extreme fire events and fail to adequately represent peatland combustion, pyroconvection, and post-fire vegetation type conversion. As a result, IPCC projections may underestimate both future fire frequency and the carbon cycle amplification from fire feedbacks — a significant source of upside risk in climate projections. The 2023 Canadian fires, for example, exceeded the 95th percentile of ESM projections for that region decades ahead of schedule.

Global Health Burden of Wildfire Smoke (PM2.5 Exposure)

Source: Marlier et al. 2013 (Nature Climate Change) — 339,000 premature deaths/yr from landscape fire smoke; Johnston et al. 2012; Chen et al. 2021 (updated global estimate); GBD 2019 (Global Burden of Disease Study) fire smoke attributable deaths; WHO Air Quality Guidelines 2021; Analitis et al. 2012 (European fire-health dose-response).

Economic Costs — Direct & Indirect

Source: Thomas et al. 2017 (USFS economic costs); Beland Kingsley et al. 2019 (Australia); OECD ENV/WKP(2019)15 — Global economic cost of wildfire; Nauslar et al. 2018; CAL FIRE annual reports; Insurance Information Institute (III) Catastrophe data 2024; IPCC AR6 WGII Chapter 7 (economic impacts).

Health Impacts — Mechanisms and Evidence

Respiratory Effects

PM2.5 from wildfire smoke penetrates deep into the alveoli, triggering oxidative stress, inflammation, and impaired gas exchange. Short-term exposure is associated with asthma exacerbation, bronchitis, pneumonia, and COPD episodes. Studies from western US fire seasons show emergency department visits for respiratory illness increase 3–7% per 10 µg/m³ increase in fire-attributable PM2.5. Long-term exposure to elevated smoke PM2.5 is associated with accelerated lung function decline and increased lung cancer risk (though dose-response quantification remains challenging due to exposure assessment complexity).

Cardiovascular Effects

Inhaled fine particles and toxic gases (CO, acrolein, benzene) enter the bloodstream, promoting systemic inflammation, oxidative stress, endothelial dysfunction, and increased coagulability — risk factors for myocardial infarction, stroke, and arrhythmia. Population studies in Australia, California, and Brazil show hospital admissions for cardiovascular events increase significantly during high-smoke periods. Vulnerable populations (elderly, those with pre-existing disease, pregnant women) face disproportionate risk.

Neurological and Developmental Effects

Emerging evidence links wildfire smoke exposure to neuroinflammation, cognitive decline, and increased dementia risk in older adults. A 2023 study (Wettstein et al.) found wildfire-day PM2.5 exposure in California was associated with increased emergency department visits for neurological conditions. Prenatal smoke exposure is associated with preterm birth, low birth weight, and altered neurodevelopment — with Indigenous and low-income communities (living in higher fire-risk areas with less capacity to shelter in place or evacuate) bearing disproportionate burden.

Mental Health Effects

Wildfire events are among the most traumatic natural disasters, with high rates of PTSD, depression, anxiety, and complicated grief in affected communities. The "eco-grief" and chronic stress associated with repeated fire threat — particularly in communities facing annual fire danger — is an emerging public health concern. The 2019–20 Australian Black Summer was associated with a measurable increase in population-wide psychological distress, with first responders and rural communities most severely affected. Children in fire-affected regions show elevated rates of anxiety disorders and sleep disruption years post-event.

Source: Reid et al. 2016 (comprehensive fire-health review); Wettstein et al. 2023 (neurological); Balmes 2022 (lung effects, NEJM review); Finlay et al. 2012; Borchers Arriagada et al. 2020 (Australian smoke health burden); Bryant et al. 2014 (mental health); Sneeuwjagt et al. 2013.

Direct Suppression Costs

US federal fire suppression (2024)~$4.1B
Australia Black Summer (2019–20)AUD $10B total cost
California 2018 Camp Fire (alone)$16.5B insured losses
Canada 2023 wildfire seasonCAD $1.1B suppression
EU 2022 fire season~€2B suppression + damages
Global OECD wildfire cost estimate (annual)$50–150B+
Source: CAL FIRE; USFS; III; OECD 2019; NSW RFS.

Infrastructure & Agricultural Losses

Structures destroyed — US annually (avg 2014–2023)~40,000 structures/yr
WUI (Wildland-Urban Interface) homes at risk (US)~44 million homes
Agricultural losses from smoke (CA Central Valley)$1–3B in high-smoke years
Tourism losses (AU, US, Mediterranean)Hundreds of millions/event
Power grid wildfire liability (CA utilities 2017–18)~$30B
Source: Radeloff et al. 2018 (WUI); USDA ERS; III; PG&E bankruptcy filings; Arent et al. IPCC AR6.

Long-Term Ecosystem Services Lost

Watershed function loss (post-fire flooding/erosion)High; multi-year impairment
Carbon stored in boreal/tropical forest (at risk)100s Gt CO₂eq
Biodiversity — endemic species at fire-risk (Australia)~113 species range impacted >50%
Air quality mortality cost (VSL-based, US annual)~$450B/yr (Marshall et al. 2020)
Soil carbon loss from peatland fires (irreversible)Centuries to millennia to recover
Source: Marshall et al. 2020; Ward et al. 2020 (AU biodiversity); Page et al. 2002 (peatlands); Brown et al. 2015 (watershed).