Search for: All records

Much of Earth’s magma is stored as extensive crystal mush systems, yet the prevalence of physical processes operating within mushes and their importance in volcanically active regions remain enigmatic. In this Review, we explore the physical properties and key processes of crystal mush systems. The initiation, evolution and decline of volcanic systems, modulated by heat supply and loss, could generate differences in the prevalence of mush processes through space and time. Additionally, regional tectonics alter mush properties, with mushes in cool wet settings having persistent residual melt, permitting more effective melt segregation than in hot dry settings. Disaggregation of mushes results in crystal mush material being mobilized or entrained into lavas and erupted, presenting opportunities to define the timescales and chemistry of some mush processes in volcanically active regions. Mush systems can be observed on length scales ranging from kilometres (using geological mapping) to micrometres (using crystal textures). Therefore, it is difficult to integrate data and interpretations across different fields. Improved integration of thermodynamics, textural analysis, geochemistry, modelling and experiments, alongside inputs from adjacent fields such as porous media dynamics, engineering and metallurgy will help to advance understanding of mush systems and ultimately improve hazard evaluation at active and dormant volcanic systems. 

Documenting the processes and timescales of magma formation and diversification and defining the locations, shapes, volumes, and phase states of magma storage and transport zones rely on data produced by novel analytical techniques and state-of-the-art experimental methods. Computational modeling effectively links these critical tools. The Magma Chamber Simulator (MCS) is an internally consistent thermodynamic open system model that uses experimental constraints from rhyolite MELTS (Gualda et al. 2012, Ghiorso & Gualda, 2015) to compute paths of open system magmas that evolve via processes including crystallization, magma mixing, cumulate/mush entrainment, and host-rock assimilation. MCS results yield elemental, isotope, mass, and thermal characteristics of melt ± crystals ± volatiles in "resident" magma, crustal wallrock (melt and solids), recharge magma, and entrained material. To model the petrochemical evolution of igneous rocks related by open-system processes, one typically runs 200+ models that vary initial compositions, pressures, and temperatures of magma, wallrock, etc. Comparison of model results with whole rock, mineral, and melt inclusion chemical data and other constraints (e.g., thermobarometry) yield interpretations about igneous processes at a range of scales—from how crust forms and evolves to processes responsible for in situ geochemical records of crystals—and allows assessment of epistemic and aleatoric uncertainties. Two examples of computational studies will illustrate MCS's utility and flexibility. (1) Modeling of historical basalts at Mt. Etna (Italy) provides evidence for variable degrees of melting of metasomatized mantle, followed by magma recharge and assimilation of partial melts of carbonate-flysch crust. (2) MCS models reproduce whole rock and mineral data of plagioclase-rich basalts at Steens Mountain (USA) through entrainment of gabbroic mush that likely formed in early stages of Columbia River Basalt magmatism. To enhance understanding of trans-lithospheric magma systems, future work on MCS will prioritize (i) building a post-processing environment that utilizes select statistical methods to inform "best-fit" models and to quantitatively assess uncertainty, and (ii) increasing modeling efficiency by adding automated modeling capabilities. 

We investigate the implications of prolonging the equilibrium crystallization (EQX) stage of lunar magma ocean (LMO) solidification beyond the oft-modeled 50% volume solids, to 60%. Most models of two-stage LMO crystallization halt the EQX phase once 50% of a molten Moon (post-core formation) solidifies, after which the remaining 50% of the LMO solidifies via fractional crystallization (FRX). We quantitatively show through a simple scaling analysis that compares crystal settling velocity to vertical convective velocity that the early EQX regime can operate up to (and possible even slightly beyond) 60% volume solids. Phases that stabilize during the EQX and FRX regimes are then computed using Perple_X (thermodynamic calculator) along with the hp633ver database and associated activity-composition relations for solid solutions, and consider an adiabat that remains between the liquidus and solidus. Early results show two key findings: 1) only low volumes (~2%) of ilmenite form over ~50-km thick upper mantle layers for both 50% and 60% EQX regimes, suggesting that a mantle overturn may have been sluggish and/or limited in depth (dense ilmenite is thought to have been a critical driver of late-stage mantle mixing); and 2) contrary to most published two-stage LMO models, a refractory-enriched (i.e. high Al2O3) bulk silicate Moon is not required to produce garnet in the lunar mantle, assuming an Earth-like bulk silicate Moon composition with an alumina content of ~4 wt.%. To complement and test these numerical phase equilibria model results, a series of piston-cylinder experiments is underway that simulate the pressures and temperatures experienced by an FeO+TiO2-rich residual LMO in order to assess the volume and distribution of ilmenite produced during LMO solidification. These results are compared to those of the numerical phase equilibria models. Despite the model-dependent nature of these results, they provide a unique insight into potential LMO crystallization that has not been previously considered in the literature. 

Abstract Magmatic systems below volcanoes are often dominated by partially crystalline magma over the long term. Rejuvenation of these systems during eruptive events can impact lava composition and eruption style—sometimes resulting in more violent or explosive activity than would be expected, as was the case at Fissure 17 during Kīlauea’s 2018 eruption. Here, we explore how the crystallinity of unerupted intrusion magmas affect hybrid magma compositions and petrological signatures by constructing phase-equilibria models to evaluate mineral and melt compositions of low-MgO lavas erupted along the East Rift Zone of Kīlauea volcano on 30 to 31 January 1997 (Episode 54, Fissures A-F). We then compare calculated mixing proportions and petrologically derived magma volumes to GPS-based geodetic inversions of ground deformation and intrusion growth in an attempt to reconcile geodetic and petrologically estimated magma volumes. Open-system phase-equilibria thermodynamic models were used to constrain the composition, degree of differentiation, and thermodynamic state of a rift-stored, two pyroxene + plagioclase saturated low-MgO magma body immediately preceding its mixing with high-MgO recharge and degassed drainback (lava lake) magma from Pu‘u‘ō‘ō‘, shortly before fissure activity within Nāpau Crater began on 29 January 1997. Mixing models constructed using the Magma Chamber Simulator reproduce the mineralogy and compositions of Episode 54 lavas within uncertainties and suggest that the identity of the low-MgO magma body may be either variably differentiated remnants of un-erupted magmas intruded into Nāpau Crater in October 1968, or another spatially and compositionally similar magma body. We find that magmas derived from a single, compositionally stratified magma emplaced beneath Nāpau Crater in 1968 can mix with mafic Kīlauea magmas to reproduce average Episode 54 bulk lava, mineralogy and mineral compositions without necessitating the interaction of multiple, low-MgO rift-stored magma bodies to produce Episode 54 lava compositions. Further, by constructing phase equilibria-based mixing models of Episode 54, we can better define the pre-eruptive state of the magmatic system. The resultant mineral assemblages and compositions are consistent with the possibility that the now-fractionated, rift-stored magma body was compositionally stratified and ~ 40% to 50% crystalline at the time of mixing. Finally, we estimate the volume of the low-MgO magma body to be ~7.51 Mm3. Phase-equilibria model results corroborate field and geochemical relationships demonstrating how shallow intrusions at intraplate shield volcanoes can crystallize, evolve, and then be remobilized by new, later batches of mafic magma. Most notably, our MCS models demonstrate that the pre-eruptive conditions of an intrusive body may be recovered by examining mineral compositions within mixed lavas. Discrepancies between the geodetic constraints on volumes of stored rift versus newly intruded (recharge) magma and our best-fit results produced by MCS mixing models (which respectively are mmafic:mlow-MgO ≈ 2 vs. mmafic:mlow-MgO ≈ 0.75) are interpreted to highlight the complex nature of incomplete mixing on more localized scales as reflected in erupted lavas, compared to geodetically constrained volumes that likely reflect large spatial scale contributions to a magmatic system. These dissimilar volume relationships may also help to constrain eruptive versus unerupted volumes in magmatic systems undergoing mixing. By demonstrating the usefulness of MCS in modeling past eruptions, we highlight the potential to use it as a tool to aid in petrologic monitoring of ongoing activity. 

Abstract The nearly continuous volcanic eruption record at Mt. Etna dating back ~700 years provides an excellent opportunity to investigate the geochemical evolution of a highly active volcano. Of particular interest is elucidating the cause of selective enrichment in alkali elements (K and Rb) and 87Sr/86Sr observed in various episodes of past activity. More recently, this alkali enrichment trend started to manifest in the 17th century and accelerated after 1971, and was accompanied by an increase in the volume, frequency, and explosivity of eruptions. Explanations for this signature include recharge of alkali-enriched magmas and/or crustal contamination from the subvolcanic basement. This study quantitatively examines the role of crustal contamination in post-1971 Etnean magma compositions via hundreds of open-system phase equilibria and trace element calculations based upon whole-rock major oxides, trace elements, 87Sr/86Sr ratios, and mineral compositional data. Available pre-1971 petrochemical data are satisfactorily reproduced by fractional crystallization of a high whole-rock MgO (12–17 wt.%), Ni (135–285 ppm), and Cr (920–1330 ppm) parental magma composition that is documented in Etna's ~4-ka fall-stratified deposit. Observed post-1971 whole-rock and glass trends and phase equilibria are reproduced via modeled assimilation of a skarn and flysch mixture, lithologies that represent the uppermost 10 to 15 km of sedimentary rocks beneath Etna. Notably, models show that K2O (wt.%) and Rb (ppm) behave incompatibly during partial melting of skarn/flysch. Additionally, the observed elevation of 87Sr/86Sr in post-1971 samples is consistent with the addition of radiogenic Sr from wallrock partial melts. In best-fit models, which yield observed post-1971 K2O, Rb, and 87Sr/86Sr trends, ~17% anatectic melt is assimilated and there may be a subordinate stoped wallrock component of ≤2% (percentage is relative to the starting mass of pristine magma). Previous work has shown that metasomatized spinel lherzolite and garnet pyroxenite can be melted in different proportions to reproduce long- and short-term changes observed in Etna’s geochemical products. We propose that the alkali enrichment signature observed after 1971 can be fully explained through the combination of mantle heterogeneity and crustal contamination. In particular, up to ~20% crustal input coupled with mantle heterogeneity of primitive melts explains the geochemical signals quite well. The influence of crustal contamination on post-1971 lavas is, in part, the result of frequent recharge of magmas that thermally primed the middle to upper crust and enhanced its partial melting. 

Several times in the past 60 ka, Mount Etna has erupted lavas with variable alkaline character. The most recent chemical excursion began in 1971, accompanied by an increase in explosivity and eruption frequency. The origin of the alkaline signature remains enigmatic, with endmember hypotheses involving dominant contributions from mantle vs. crust. For lavas that erupted between 1329 and 2016, we used thermodynamic modeling to test if post-1971 anomalous alkalinity is dominated by mantle processes. First, we assessed mantle melting conditions required to reproduce the chemistry of potential parental magmas. Second, we examined the differentiation of partial melts as they underwent closed-system crustal storage and ascent. The mantle melting conditions explored via ~300 models include source compositions (peridotite + pyroxenite ± metasomatic phases), pressure, extent of melting, and fO2. Best-fit models that reproduce the major element chemistry of volcanics interpreted as parental magmas involve 20 to 30% melting of a peridotite-pyroxenite mantle source that contained phlogopite, pargasite, and CO2 (~1 wt.%) between ~1-1.5 GPa (~30-45 km) along the QFM+1 buffer. Subsequent isentropic decompression (adiabatic and reversible) of mantle partial melts + crystallization at shallower pressures (0.8-0.2 GPa) were modeled to test the effects of closed system ascent and storage. Isentropic decompression models yield no crystallization although temperature decreases ~3°C/0.1 GPa. Decompression + closed system crystallization fail to replicate post-1971 glass samples and do not explain observed post-1971 alkali enrichments. We conclude that partial melting of a metasomatized source produced Etna parental magmas, but closed system crustal ascent and storage cannot fully account for alkali enrichment highlighted in post-1971 products at Etna. Open system modeling suggests that assimilation and crystallization (e.g., Takach et al. 2024) play a critical role, and ongoing modeling is testing the contributions of recharge (magma replenishment) and entrainment of previously formed mushes to the high alkalinity excursions. 

Key Points The 15 January 2022 Hunga Tonga‐Hunga Ha'apai eruption had four episodic seismic subevents with similar waveforms within ∼300 s An unusual upward force jump‐started each subevent A magma hammer explains the force and estimates the subsurface magma mass flux which fits the vent discharge rate based on satellite data 

Abstract Lavas erupted at hotspot volcanoes provide evidence of mantle heterogeneity. Samoan Island lavas with high⁸⁷Sr/⁸⁶Sr (>0.706) typify a mantle source incorporating ancient subducted sediments. To further characterize this source, we target a single high⁸⁷Sr/⁸⁶Sr lava from Savai’i Island, Samoa for detailed analyses of⁸⁷Sr/⁸⁶Sr and¹⁴³Nd/¹⁴⁴Nd isotopes and major and trace elements on individual magmatic clinopyroxenes. We show the clinopyroxenes exhibit a remarkable range of⁸⁷Sr/⁸⁶Sr—including the highest observed in an oceanic hotspot lava—encompassing ~30% of the oceanic mantle’s total variability. These new isotopic data, data from other Samoan lavas, and magma mixing calculations are consistent with clinopyroxene⁸⁷Sr/⁸⁶Sr variability resulting from magma mixing between a high silica, high⁸⁷Sr/⁸⁶Sr (up to 0.7316) magma, and a low silica, low⁸⁷Sr/⁸⁶Sr magma. Results provide insight into the composition of magmas derived from a sediment-infiltrated mantle source and document the fate of sediment recycled into Earth’s mantle.