Role of volatiles in highly explosive basaltic eruptions

The study investigates how volatile species, primarily H2O and CO2, control magma ascent and eruptive behaviour in basaltic systems. Volatiles exsolve as pressure drops, increasing buoyancy and accelerating ascent, while also affecting crystallisation, rheology, temperature, and outgassing. A prevailing view links higher pre-eruptive volatile content to greater eruption intensity. Recent arguments suggest high CO2 facilitates additional H2O exsolution (via solubility relations), enhancing buoyancy and explosivity. However, this overlooks non-linear, coupled processes during ascent (crystallisation, outgassing efficiency, adiabatic cooling) that can counteract or amplify effects. The paper aims to quantify the respective roles of H2O and CO2 on ascent dynamics, eruption intensity (mass eruption rate), and eruptive style for shallow-sourced basaltic magmas, using the well-documented Etna 122 BCE basaltic Plinian eruption as a test case to explore parameter space rather than replicate the event.

Prior work establishes volatiles as key drivers of ascent dynamics, degassing, and eruptive style in mafic systems. H2O and CO2 dominate magmatic gases, with H2O being more soluble than CO2 at crustal pressures. Studies have linked higher water contents to higher eruption intensities and suggested CO2-rich conditions could trigger stronger explosions by promoting H2O exsolution and increasing buoyancy. Additional literature highlights the importance of degassing-driven crystallisation, adiabatic cooling from bubble expansion, rheological stiffening with crystallinity, and the role of outgassing in effusive–explosive transitions. Evidence from basaltic Plinian eruptions (e.g., Etna 122 BCE, Masaya Triple Layer, Tarawera 1886) indicates powerful eruptions can occur with relatively low water contents (~1–3 wt.%). Earlier modelling work explored conduit dynamics, permeability, and fragmentation criteria, forming the basis for the present multiphase, multicomponent approach that integrates solubility, rheology, crystallisation kinetics, and outgassing.

The authors use a 1D steady-state, multiphase, multicomponent conduit model for magma ascent in a cylindrical conduit (length 6 km), solving for coupled mass, momentum, energy, volatile exsolution, crystallisation, and gas–liquid relative motion. Application to a given system requires constitutive laws for rheology, solubility, crystallisation, outgassing/permeability, and fragmentation, along with equations of state for each component and boundary conditions. Test case and setups: The Etna 122 BCE basaltic Plinian eruption provides representative conditions: inlet pressure 160 MPa, inlet temperature 1060 °C, initial crystal fraction ~10 vol.% (phenocrysts), conduit radius 35 m. Dissolved volatile ranges for Etna (from melt inclusions) are ~1–3 wt.% H2O and 200–900 ppm CO2. The study explores several scenarios: (a) reference (3.0 wt.% H2O, 0.9 wt.% CO2); (b) excess H2O (6 and 10 wt.% total H2O, with CO2 fixed at 0.9 wt.%); (c) excess CO2 (6 and 10 wt.% total CO2, with H2O fixed at 3 wt.%). Excess volatiles represent redistribution of pre-eruptive bubbles to the chamber top without changing chamber pressure or temperature; the inlet pressure is fixed for these runs. Governing framework: The model treats magma below fragmentation as a two-phase mixture (liquid: melt + crystals + dissolved volatiles; gas: exsolved volatiles) and above fragmentation as gas–particle flow. Conservation equations for mixture mass, momentum, energy are coupled with transport equations for liquid and gas volume fractions, disequilibrium exsolution of each volatile component (characteristic time τe = 1 s), disequilibrium crystallisation for major phases (plagioclase, olivine, pyroxene; characteristic time τc = 10 s), and gas–liquid relative motion governed by a relaxation rate dependent on permeabilities and drag. Constitutive laws: Melt viscosity follows a Vogel–Fulcher–Tammann-based formulation parameterised by melt composition and dissolved H2O, with crystal-bearing mixture viscosity augmented via suspension rheology (Costa et al.). Crystallisation equilibria depend on pressure, temperature, and dissolved water, fitted with polynomials calibrated using alphaMELTS across P–T–xH2O space. Wall friction transitions from laminar (Hagen–Poiseuille) to turbulent (Fang et al.) based on Reynolds number; above fragmentation, a gas-phase friction factor is prescribed. Outgassing below fragmentation assumes a permeable gas flow regime; lateral gas escape is neglected. Solubility models: Equilibrium dissolved H2O and CO2 are functions of fugacity, using relations from Lesne et al. for H2O and Allison et al. for CO2; fugacities approximated from partial pressures (Holloway and Blank). Equations of state: Exsolved gases use a non-ideal van der Waals formulation; melt and crystals use linearised Mie–Grüneisen forms. Fragmentation criterion: A brittle criterion based on the Deborah number (De = λm/λd = μ/(γ̇ G∞)) with threshold De > 0.01, where μ is bulk viscosity and G∞ = 10^10 Pa for basaltic melts. Boundary conditions: At inlet (6 km depth), set pressure, temperature, total volatile contents, and initial crystal content; at the vent, atmospheric pressure and/or choked flow condition (Mach 1) determine exit state. Numerical solution uses a shooting method with step-size control and complex-step derivative approximation. Sensitivity analysis: 10,000 simulations varying inlet pressure (140–180 MPa), temperature (1030–1130 °C), conduit radius (10–50 m), total H2O (1–10 wt.%), total CO2 (0.1–10 wt.%), and initial crystal content (0–20 vol.%), sampled uniformly. DAKOTA toolkit provides sensitivity and Sobol indices for outputs: mass eruption rate (MER), melt exit velocity, and fragmentation depth. External volatile source experiments: Additional simulations consider adding 1 wt.% H2O or 1 wt.% CO2 to the chamber from an external source (e.g., carbonate interaction), computing chamber pressurisation and expansion from mass balance and elastic response (ΔV/V = 3ΔP/4μ, μrock = 10 GPa). In these runs, melt+crystal mass is held constant across simulations while chamber pressure and volume increase (e.g., +31 MPa for +1 wt.% H2O; +69 MPa for +1 wt.% CO2).

- Under fixed chamber pressure (160 MPa) and geometry, varying CO2 content from 0.9 to 10 wt.% has negligible impact on MER (4×10^7–5×10^7 kg s^-1), despite increasing gas volume fraction and buoyancy, because CO2-rich conditions increase mixture viscosity via enhanced crystallisation and reduced dissolved volatiles. - Increasing H2O content lowers viscosity and increases ascent velocity. MER rises from ~5×10^7 kg s^-1 at 3 wt.% H2O to ~3×10^8 kg s^-1 at 10 wt.% H2O (just under an order of magnitude). - Excess H2O simulations yield very high ascent velocities below ~500 m depth, which suppress elongational strain rates in low-viscosity basaltic magma and favour lava fountaining rather than Plinian/sub-Plinian explosivity. - Excess CO2 simulations produce higher crystallinity and viscosities (up to ~10× the reference at depth), leading to similar ascent velocities at depth compared to the reference case, and fragmentation depths that tend to be deeper on average as CO2 increases. - Representative values (Table 2): at 3000 m depth, gas volume fraction increases from 0.16 (reference) to 0.54 (H2O 10 wt.%) and 0.52 (CO2 10 wt.%); viscosity increases markedly with CO2 (e.g., 5.3×10^3 Pa s at 3000 m for CO2 10 wt.% vs 2.3×10^2 Pa s for H2O 10 wt.%); velocities at 3000 m range from 5 m s^-1 (reference) to 58 m s^-1 (H2O 10 wt.%) and 8 m s^-1 (CO2 10 wt.%). At 500 m depth, velocities are ~7 (ref), 92 (H2O 10 wt.%), and 11 m s^-1 (CO2 10 wt.%). - Sensitivity analysis (10,000 runs) shows eruptive style is most sensitive to inlet temperature and initial crystal content: lower temperature and higher crystallinity increase viscosity and Deborah number, promoting explosive fragmentation. Conduit radius and inlet pressure have weaker effects, with larger radius and higher pressure favouring fountaining by reducing shallow acceleration. - H2O content strongly influences ascent rate and exit velocity (40–300 m s^-1 overall), promoting fountaining at higher H2O by suppressing shallow strain rates. Powerful explosive eruptions remain possible at relatively low H2O (~1–3 wt.%). CO2 content shows minimal impact on MER and style over the explored range. - MER across the sensitivity space is controlled primarily by conduit radius and inlet temperature rather than volatile contents; both explosive and high-fountain regimes span ~10^7–10^9 kg s^-1. - External volatile source: Adding only 1 wt.% external H2O or CO2 increases chamber pressure by ~31 or ~69 MPa and chamber volume by ~0.2% or ~0.5%, respectively, significantly increasing ascent rates and MER (about an order of magnitude), and shifting style from explosive to lava fountaining due to high ascent rates reducing shallow strain rates.

The findings challenge the notion that higher pre-eruptive CO2 directly yields more intense basaltic eruptions. While greater CO2 lowers dissolved volatile content and increases crystallisation-driven viscosity (which can deepen fragmentation), it does not appreciably increase ascent rates or MER under fixed chamber pressure and geometry. In contrast, higher H2O reduces viscosity and increases buoyancy, accelerating ascent and increasing MER, but the resulting low-viscosity, fast-flow regime suppresses elongational strain rates in the shallow conduit, biasing towards lava fountaining rather than Plinian behaviour. Thus, eruptive style and intensity decouple: MER may rise with H2O while explosivity weakens, whereas lower H2O can increase viscosity and strain rates, enabling fragmentation and highly explosive eruptions even at modest water contents (~1–3 wt.%). Sensitivity analysis underscores the dominant roles of pre-eruptive temperature and phenocryst content in setting viscosity and fragmentation propensity, and of conduit radius and temperature in controlling MER. A key exception occurs when external volatiles (especially CO2) pressurise the chamber; the increased driving pressure substantially boosts ascent rates and MER and can alter eruptive style. These results apply to shallow-sourced basaltic systems (<6 km) where CO2 is largely exsolved prior to ascent; different behaviour may arise in deeper systems (>10 km) where CO2 exsolution initiates and H2O remains mostly dissolved, warranting further investigation. Overall, the study provides a holistic, process-coupled framework linking volatile budgets, rheology, ascent dynamics, fragmentation depth, and eruption intensity/style.

Volatiles fundamentally control ascent dynamics in basaltic magmas. Water-rich magmas ascend faster due to enhanced buoyancy and reduced viscosity, increasing MER by up to roughly an order of magnitude over the explored range, but often favouring lava fountaining through suppression of shallow strain rates. Elevated CO2 increases viscosity through reduced dissolved volatiles and enhanced crystallisation, counterbalancing buoyancy such that MER and ascent rates remain largely unaffected by CO2 content alone. The most impactful role for CO2 in intensifying eruptions arises when external sources add CO2 (or H2O) to a closed system, pressurising the chamber and substantially increasing MER and ascent rates, with potential shifts in eruptive style. Powerful explosive basaltic eruptions can occur at relatively low water contents (<3 wt.%), consistent with historical basaltic Plinian events. Future work should extend these analyses to deeper storage systems (>10 km), explore transient, non-steady ascent, incorporate lateral outgassing pathways, and further constrain kinetic parameters for exsolution and crystallisation.

- The model is 1D and steady-state, not resolving transient dynamics or conduit heterogeneity. - Simulations emphasise shallow reservoirs (<6 km); results may not generalise to deeper systems where CO2 exsolution initiates and H2O remains largely dissolved. - Lateral gas escape is neglected; only vertical outgassing is included. - Disequilibrium kinetics are represented by constant characteristic times (τe = 1 s for exsolution; τc = 10 s for crystallisation) due to limited experimental constraints for basalts. - Excess volatile scenarios assume redistribution of pre-eruptive bubbles without changing chamber pressure; real systems may experience pressure variations. - Sensitivity analysis treats scenarios as independent with uniform parameter distributions; it does not simulate temporal evolution of a single magmatic system. - Conduit radius is fixed with depth and wall roughness set; natural conduits may vary. - The Etna 122 BCE eruption is used as a test case for parameter selection; simulations are not tuned to reproduce the event in detail.