Volcanoes & Climate Natural Forcing Agent Short-term Cooling, Long-term CO₂
0.3–0.5 GtCO₂
Annual volcanic CO₂ emissions
~1% of human emissions (~37 Gt/yr); not climate-significant on decadal scale
~1% of human emissions (~37 Gt/yr); not climate-significant on decadal scale
−0.5°C
Pinatubo 1991 global cooling peak
~20 Mt SO₂ injected into stratosphere; cooling lasted ~18 months
~20 Mt SO₂ injected into stratosphere; cooling lasted ~18 months
VEI 8
Largest eruption scale (Toba, ~74,000 BP)
~2,800 km³ magma; possible 6–10°C cooling; near-human-extinction debate
~2,800 km³ magma; possible 6–10°C cooling; near-human-extinction debate
~1815
Tambora — deadliest eruption in recorded history
92,000 deaths; "Year Without a Summer" 1816; global harvest failure
92,000 deaths; "Year Without a Summer" 1816; global harvest failure
1–3 years
Typical stratospheric aerosol lifetime
After VEI 6+; aerosol optical depth (AOD) radiative forcing −1 to −4 W/m²
After VEI 6+; aerosol optical depth (AOD) radiative forcing −1 to −4 W/m²
~100–200 Mt
Annual SO₂ from all volcanoes
Mostly mid/low-troposphere (no climate effect); only stratospheric injection matters
Mostly mid/low-troposphere (no climate effect); only stratospheric injection matters
★ How Volcanoes Interact with Earth's Climate
Volcanoes affect climate through two fundamentally different mechanisms that operate on opposite timescales. On short timescales (months to years), explosive eruptions that inject sulfur dioxide (SO₂) into the stratosphere create a global "veil" of sulfate aerosols that reflect incoming solar radiation, cooling the surface. The 1991 eruption of Mt. Pinatubo in the Philippines injected ~20 million tonnes of SO₂ above 25 km altitude and cooled the planet by ~0.5°C for about 18 months. On geological timescales (millions of years), sustained flood basalt volcanism — the massive outpourings that built the Siberian Traps, the Deccan Traps, and other Large Igneous Provinces — emitted enormous quantities of CO₂ and SO₂ that drove several of Earth's mass extinctions through runaway warming, ocean acidification, and toxic ash clouds.
Critically, modern human fossil fuel emissions (~37 Gt CO₂/yr) are approximately 80–100× larger than all volcanic CO₂ output (~0.3–0.5 Gt/yr). The "volcanoes emit more than humans" claim is a well-documented myth. However, volcanic forcing remains an important natural climate signal — both for understanding pre-industrial climate variability and as the direct inspiration for Stratospheric Aerosol Injection (SAI) geoengineering proposals.
Global Temperature Signal — Major Volcanic Eruptions (1850–2025)
Source: NASA GISS Surface Temperature Analysis (GISTEMP v4); Sigl et al. 2015 ice core chronology; Robock 2000 volcanic forcing database; NOAA paleoclimate; Pinatubo NASA record.
Volcanic Forcing vs. Human CO₂ — Annual Emissions Comparison
Source: Gerlach 2011 (AGU); Burton et al. 2013; IEA Global CO₂ Status Report 2024; USGS Volcano Hazards Program; Global Carbon Project 2024.
The Pinatubo Benchmark: The June 1991 eruption of Pinatubo (Philippines, VEI 6) is the best-documented large volcanic climate event in the satellite era. The eruption injected ~20 Mt SO₂ into the stratosphere, which oxidised to ~35 Mt H₂SO₄ aerosols. These aerosols increased Earth's reflectivity (albedo), reducing solar insolation reaching the surface by ~2–3 W/m². Global mean surface temperature dropped ~0.5°C by mid-1992, recovered by 1994. The event allowed scientists to validate climate models by comparing predictions to observations — and confirmed the cooling mechanism that underpins modern SAI geoengineering proposals.
Volcanic Radiative Forcing — Mechanism
Source: Ammann et al. 2003; Crowley 2000; Sigl et al. 2015; IPCC AR6 Chapter 7 (Natural Forcings); Zanchettin et al. 2016.
Forcing Mechanism Step-by-Step
1. Eruption column reaches stratosphere>12–15 km altitude
SO₂ injected above tropopauseAvoids tropospheric washout
2. SO₂ → H₂SO₄ aerosol conversion2–6 weeks
Aerosol particle size ~0.1–1 µmOptimal for SW scattering
3. Global aerosol spread1–3 months
Zonal wind transport; latitudinal mixing6–24 months full spread
4. Radiative forcing−1 to −4 W/m²
Shortwave scattering dominatesNet cooling effect
5. Surface cooling−0.1 to −1.5°C
Depends on SO₂ mass & latitudePeak ~12–18 months
6. Aerosol sedimentation1–3 year lifetime
Recovery to baseline temperature2–5 years post-eruption
Source: Robock 2000; Timmreck 2012; Kremser et al. 2016; IPCC AR6 WG1 Ch.7.
Stratospheric vs. tropospheric injection — the critical distinction: Volcanic eruptions that release SO₂ into the troposphere (below ~12 km) have negligible climate effect because rain quickly washes SO₂ out within days to weeks. Only eruptions powerful enough to punch through the tropopause — generally VEI 5 or larger — can create sustained stratospheric aerosol veils. Effusive "shield" volcanoes like Kīlauea in Hawaii emit enormous SO₂ quantities but primarily at low altitudes, creating local air quality problems (vog) but essentially no global climate forcing. This distinction is crucial for understanding why not all volcanic activity affects climate equally.
VEI Scale — Volcanic Explosivity Index & Climate Impact
| VEI | Ejecta Volume | Column Height | SO₂ Stratospheric Injection | Climate Effect | Return Period | Example |
|---|---|---|---|---|---|---|
| VEI 3 | 0.01–0.1 km³ | 3–15 km | Negligible | None | Frequent (yearly) | Stromboli (ongoing) |
| VEI 4 | 0.1–1 km³ | 10–25 km | Minimal | Local/regional only | ~10 yrs | Eyjafjallajökull 2010 |
| VEI 5 | 1–10 km³ | 20–35 km | ~1–5 Mt SO₂ | −0.1 to −0.3°C, 1–2 yr | ~50 yrs | Mt. St. Helens 1980; Hunga Tonga 2022 |
| VEI 6 | 10–100 km³ | 25–40 km | ~10–30 Mt SO₂ | −0.3 to −0.6°C, 1–3 yr | ~100 yrs | Pinatubo 1991; Krakatau 1883 |
| VEI 7 | 100–1,000 km³ | >40 km | ~50–200 Mt SO₂ | −1 to −3°C, 3–10 yr | ~500–1,000 yrs | Tambora 1815; Samalas 1257 |
| VEI 8 | >1,000 km³ | >50 km | >500 Mt SO₂ | −3 to −10°C+, decades | ~50,000–100,000 yrs | Toba ~74,000 BP; Yellowstone ~640,000 BP |
Source: Newhall & Self 1982 (VEI scale); Siebert et al. 2010; Oppenheimer 2011; Robock 2000; Metzner et al. 2012; IPCC AR6.
Historically Significant Eruptions — Climate & Human Impact
| Eruption | Date | VEI | SO₂ (Mt) | ΔT Global | Duration | Significance |
|---|---|---|---|---|---|---|
| Samalas, Indonesia | 1257 AD | VEI 7 | ~158 Mt | −1.5°C | ~3–5 yr | Largest Holocene eruption by sulphur emission; ice core signal unprecedented; linked to 1258 "mystery eruption" cold spell; contributed to medieval famines across Europe and Asia. |
| Tambora, Indonesia | April 1815 | VEI 7 | ~60 Mt | −0.5 to −1°C | ~3 yr | Deadliest eruption in recorded history. 92,000 total deaths (direct + famine). 1816 "Year Without a Summer" — crop failures across North America, Europe, Asia. Triggered mass migration, cholera pandemics, and inspired Mary Shelley's Frankenstein. |
| Krakatau, Indonesia | August 1883 | VEI 6 | ~20 Mt | −0.3°C | ~2 yr | Loudest sound in recorded history (heard 5,000 km away). 36,000 deaths mostly from tsunamis. Global "blue moon" and vivid sunsets for 2 years from aerosols — inspired Munch's "The Scream" sky. First volcanic event studied in detail with modern science. |
| Santa María, Guatemala | October 1902 | VEI 6 | ~5 Mt | −0.1°C | <1 yr | One of the largest 20th-century eruptions. ~5,000 direct deaths. Accompanied by contemporaneous Caribbean eruptions (Mont Pelée, Soufrière) in 1902 — a statistically remarkable multi-eruption year. |
| Novarupta / Katmai, Alaska | June 1912 | VEI 6 | ~5 Mt | −0.2°C | ~1 yr | Largest 20th-century eruption by volume. Created "Valley of Ten Thousand Smokes." Relatively low death toll due to remote Alaska location. Significant hemispheric aerosol loading. |
| Mt. St. Helens, USA | May 1980 | VEI 5 | ~1 Mt | −0.1°C | <1 yr | Most studied volcanic eruption in history. Lateral blast destroyed 600 km² forest. 57 deaths. Introduced millions of Americans to volcanic hazard. Minimal global climate effect due to modest SO₂ injection altitude. |
| El Chichón, Mexico | April 1982 | VEI 5 | ~7 Mt | −0.3°C | ~2 yr | Disproportionately large climate effect for its size due to unusually high sulphur content. Occurred simultaneously with a strong El Niño — the combined cooling/warming signals created complex attribution challenges. 2,000 deaths. |
| Pinatubo, Philippines | June 1991 | VEI 6 | ~20 Mt | −0.5°C | ~2 yr | Best-studied large eruption in history (satellite era). Injected stratospheric aerosol veil visible from space. 800 deaths (despite 58,000 evacuated). Validated climate model predictions. Directly inspired SAI geoengineering research. Temporarily masked ~2 years of anthropogenic warming. |
| Hunga Tonga-Hunga Ha'apai | January 2022 | VEI 5 | ~0.4 Mt SO₂ but ~150 Mt H₂O | +0.05°C? (H₂O warming) | ~5–7 yr (water vapour) | Unique in volcanic history: injected enormous quantities of water vapour (a greenhouse gas) into the stratosphere rather than cooling SO₂. May have caused a slight temporary warming signal, complicating the 2022–2024 temperature anomaly interpretation. Triggered catastrophic tsunamis reaching Tonga, Fiji, New Zealand, Peru. |
Source: Smithsonian GVP Eruption Catalog; Robock 2000; Oppenheimer 2011; Sigl et al. 2015; Self 2006; Millán et al. 2022 (Hunga Tonga water vapour); IPCC AR6 Ch.7.
Setting the record straight — volcanism vs. human CO₂: The myth that "volcanoes emit more CO₂ than humans" is one of the most persistent climate misconceptions. The best estimates of global volcanic CO₂ output are 0.26–0.54 Gt CO₂/yr from all sources (subaerial + submarine), with a central estimate of ~0.3–0.4 Gt/yr. Human fossil fuel combustion, cement production, and deforestation collectively emit ~37–40 Gt CO₂/yr. The ratio is approximately 80–100:1. Even considering all volcanic-related degassing (including hydrothermal systems, volcanic lakes, CO₂-rich springs), the total does not exceed ~1 Gt/yr — still 37× lower than human emissions.
Volcanic CO₂ vs. Human Emissions (2023, GtCO₂/yr)
Source: Gerlach 2011 (AGU Focus); Burton et al. 2013; Werner et al. 2019; Global Carbon Project 2024 (37.4 Gt human CO₂); USGS Volcanic CO₂ fact sheet.
Long-term Volcanic Carbon Role — The Geological Carbon Cycle
Over geological timescales (millions to billions of years), volcanism plays a fundamentally important role in Earth's carbon cycle — just not in the way climate deniers suggest. The long-term carbon cycle (operating over millions of years) involves:
Degassing — volcanic CO₂ outgassingWarms climate
Weathering — silicate rock CO₂ drawdownCools climate
Subduction — carbon returned to mantleGeological timescale
Berner GEOCARB model (500 Ma CO₂ history)Governed by this cycle
Flood basalt events (LIPs) — episodic CO₂ spikesMass extinction driver
Siberian Traps (252 Ma) → End-Permian extinction~96% marine species lost
Deccan Traps (66 Ma) → End-Cretaceous contextCoincident with K-Pg impact
The key distinction: geological volcanic CO₂ acts over millions of years; human emissions are compressing millions of years of fossil carbon release into decades — orders of magnitude faster than natural volcanic forcing can operate.
Source: Berner 2004 GEOCARB III; Wignall 2001 (LIPs); Self et al. 2005 (flood basalts); Rothman 2001; IPCC AR6 Ch.5.
Stratospheric Aerosol Injection (SAI) — Pinatubo as Blueprint
The apparent success of the Pinatubo eruption in temporarily cooling the globe by ~0.5°C with ~20 Mt of SO₂ injection directly inspired proposals to deliberately inject sulphate aerosols into the stratosphere to offset anthropogenic warming — a technique known as Stratospheric Aerosol Injection (SAI) or, more colloquially, "solar geoengineering." SAI is the most technically mature and cost-competitive of the proposed solar geoengineering approaches, but it carries profound risks, governance challenges, and side-effects that make it deeply controversial.
SAI — Potential Benefits
Rapid deployment (months vs. decades for mitigation)Key advantage
Low direct cost (~$2–8B/yr for 1°C of cooling)Technologically accessible
Proven analogue (Pinatubo, El Chichón)Physical mechanism validated
Reduces extreme heat events during transitionBuys time
Modelling suggests agricultural benefit in some regionsCrop stress reduction
SAI — Known Risks & Concerns
"Termination shock" — abrupt end → rapid rebound warmingMajor risk
Monsoon disruption (South Asia, Africa)Uneven regional effects
Stratospheric ozone depletion (SO₄ + ClO₃ chemistry)UV increase
Diffuse light reduces direct-beam solar power outputEnergy transition conflict
Governance: no treaty; unilateral deployment riskGeopolitical flashpoint
Does not address ocean acidification (CO₂ remains)Treats symptom not cause
Moral hazard — reduces mitigation urgencyPolitical economy risk
Source: Robock et al. 2009 (20 reasons against SAI); NAS 2021 Reflecting Sunlight; IPCC AR6 Ch.4; Lawrence et al. 2018; Irvine et al. 2019.
The "Termination Shock" problem: If SAI were deployed at scale to offset 2–3°C of warming and then suddenly halted — due to war, political change, or economic collapse — the masked warming would manifest within a few years, potentially triggering a temperature spike of 1–3°C in a decade. This rate of warming would far exceed anything in the geological record except asteroid impacts, and would be catastrophic for ecosystems adapted to stable temperatures over decades. This "termination shock" scenario is the single most-cited technical risk of SAI deployment.
Notable Currently-Active & Historically Significant Volcanoes
| Volcano | Location | Type | Status | Climate Relevance |
|---|---|---|---|---|
| Kīlauea | Hawaii, USA | Shield (effusive) | Continuously active | Minimal — low-altitude SO₂; local vog only |
| Stromboli | Italy (Aeolian Islands) | Strombolian | Persistently active ("Lighthouse of the Med") | None — very low ejecta volume |
| Etna | Sicily, Italy | Stratovolcano | Frequently active | Minimal — largest SO₂ emitter in Europe but tropospheric |
| Sakurajima | Japan | Stratovolcano | Continuously active | None — frequent small eruptions |
| Merapi | Java, Indonesia | Stratovolcano | Frequently active (VEI 3–4 cycles) | Minimal — 3 million people live on flanks |
| Popocatépetl | Mexico | Stratovolcano | Persistently active; elevated 2023–2024 | Minimal currently; VEI 5+ potential |
| Yellowstone | Wyoming, USA | Supervolcano (caldera) | Dormant (not overdue) | VEI 8 potential — last eruption 640,000 BP; recurrence interval ~600,000–800,000 yr |
| Campi Flegrei | Naples, Italy | Caldera | Unrest/uplift since 2020; most recent VEI 7 eruption ~39,000 BP (Campanian Ignimbrite) | VEI 7+ potential — under monitoring; 500,000 in red zone Naples |
| Taal | Philippines | Complex | Active; VEI 4 eruption Jan 2020 | Moderate — 2020 eruption disrupted Manila air traffic |
| Hunga Tonga | Tonga, Pacific | Submarine/island | Post-2022 rebuilding | 2022 eruption — anomalous H₂O injection; climate significance debated |
Source: Smithsonian GVP Volcano Activity Reports 2024; USGS Volcano Hazards Program; INGV (Italy) monitoring data; PHIVOLCS (Philippines) 2024.
Supervolcano Threat Assessment
A supervolcano is informally defined as a volcano capable of a VEI 8 eruption (ejecting >1,000 km³ of material). Known supervolcanoes include Yellowstone (USA), Campi Flegrei (Italy), Toba (Indonesia), Taupo (New Zealand), and Long Valley (USA). The last VEI 8 event was Toba ~74,000 years ago.
Toba eruption (~74,000 BP) cooling estimate−6 to −10°C globally
Duration of Toba winter5–10+ years
Human population bottleneck hypothesis~3,000–10,000 individuals survived?
Yellowstone — annual probability of VEI 8~1 in 730,000
Time since last Yellowstone VEI 8~640,000 years ago
Campi Flegrei — last VEI 7+ event~39,000 BP (Campanian Ignimbrite)
Annual probability of a VEI 6+ globally~1–2% per decade
Perspective: While supervolcano eruptions represent existential-level natural hazards, their probability in any given century is extremely low. The more pressing volcanic climate concern is a VEI 6–7 event occurring during a period of high anthropogenic warming — the combined signal of post-eruption cooling followed by rapid rebound could confuse public and policymaker understanding of climate change.
Source: Rampino & Self 1992; Ambrose 1998 (Toba bottleneck); USGS Yellowstone Volcano Observatory; Mastin et al. 2014; Rougier et al. 2018 (annual probabilities).