Search

The processes that control the crystallization of magmas, namely nucleation and growth, remain poorly constrained. Classical nucleation theory (CNT) has so far not provided a unified framework to explain crystal number density in laboratory experiments and the field. We use numerical experiments to study the influence of different cooling rates and CNT parameters on the crystal number density measured under constrained conditions in laboratory experiments. With varying the magnitude of pre‐exponential factor (J0) of nucleation rate, which represents the number of cluster as heterogeneous nucleation sites, we find that the cooling rate exponent (ξ) of the crystal number density varies from 3/2 to −1, depending on the magnitude of J0, cooling rate and surface tension. Using the regime maps of ξ as functions of J0 and surface tension, we can successfully interpret the diversity of cooling rate exponents reported by laboratory experiments in terms of J0 and surface tension. For extreme case of a negative cooling rate exponent ξ = −1, we propose a new crystallization model with constant rates of nucleation and crystallization, which reproduces a log‐linear crystal size distribution. In this condition, it is interesting that the crystal growth rate is inversely proportional to time, even if the diffusion‐limited growth is inversely proportional to the square root of time. For the case study of zoning profiles in microlites formed by decompression‐induced crystallization during the Shinmoedake 2011 subplinian eruptions, the comparison between theoretical predictions and natural volcanic products supports that our model is well applicable to magma ascent processes. Plain Language Summary: Crystallization of magmas denotes the formation of crystals in cooling magmas owing to heat loss during their ascension to the Earth surface or to the decompression and vesiculation occurring during an eruption. Since it strongly depends on cooling rate, the crystal number density is the most relevant textural index of the final products of the ascending magmas, the erupted rocks. Continuous efforts have been made in theory, laboratory experiments and observation of natural rock samples to understand how the cooling rate of magmas controls the texture of the erupted materials. Yet the role that different factors play on the texture of rocks during crystallization remains poorly constrained. For instance, the crystal number density has been observed to increase with cooling rate in some laboratory experiments and show no dependence or even decrease in others. The model proposed here identifies the nucleation rate as the main parameter controlling the formation of crystals. This study provides a consistent framework to relate the crystal number densities measured in laboratory experiments and observed in natural samples to their cooling conditions. Key Points: Pre‐exponential factor of nucleation rate in classical theory has the greatest effect on number density of crystals formed by coolingThe model can explain the diversity of the experimental and natural dataConstant rates of nucleation and crystallization yield log‐linear CSDs [ABSTRACT FROM AUTHOR]

Altered oceanic crust (AOC) plays a critical role in geochemical recycling in subduction zones. However, identifying contributions of subducted AOC to arc magmas remains a conundrum due to the lack of effective tracers. Here, we investigate the Ba‐Sr‐Nd isotopic compositions of lavas from the Mariana arc and back‐arc. Based on a statistical analysis of the Sr‐Nd isotopes for global arc volcanoes, we confirm that AOC‐derived fluid (or hydrous melt), rather than sediment‐derived melt or fluid, is responsible for the Sr‐Nd isotope decoupling (i.e., 87Sr/86Sr is "excessively" enriched relative to 143Nd/144Nd when compared to the "normal" mantle derivates) observed in island arc lavas. We show that the arc lavas with increasingly decoupled Sr‐Nd isotopes generally have heavier Ba isotope ratios, which is also a characteristic feature of AOC‐derived fluids. Thus, these results establish an intimate link between subducted AOC, heavy Ba isotope compositions, and Sr‐Nd isotope decoupling signature in island arcs, which provides a powerful tool to trace the AOC recycling in subduction zones. Furthermore, a similar correlation is observed between Sr‐Nd isotope decoupling and heavy B isotope ratios for global arc lavas, implying that the recycling of AOC component is generally linked to serpentinite dehydration in subduction zones. Plain Language Summary: Altered oceanic crust (AOC) is an important subduction component that contributes to the generation of arc magmas. However, identifying this contribution to arc magmas is challenging for the lack of effective tracers. Nonetheless, AOC has two characteristic features: (a) Sr‐Nd isotope decoupling (i.e., enriched in 87Sr/86Sr relative to 143Nd/144Nd in contrast to the "normal" mantle array); and (b) relatively heavy Ba isotope compositions. Here, we analyze the Ba‐Sr‐Nd isotopic compositions of lavas from the Mariana arc and back‐arc, which are combined with a statistical analysis of the Sr‐Nd isotopes for global arc volcanoes. Our results confirm that fluid (or hydrous melt) derived from subducted AOC, rather than sediment, is responsible for the Sr‐Nd isotope decoupling observed in island arc lavas. Strikingly, the magnitude of the Sr‐Nd isotope decoupling correlates well with the Ba isotope compositions of the arc lavas. Thus, our results establish an intimate link between subducted AOC, heavy Ba isotope ratios, and Sr‐Nd isotope decoupling signature in island arcs. This combination of Ba‐Sr‐Nd isotopes can therefore now be used as a powerful tracer to identify the contributions of subducted AOC to arc magmas. Our results also suggest that AOC and serpentinite show synergistic behavior during subduction recycling. Key Points: Global arc data confirm that the Sr‐Nd isotope decoupling of arc lavas is caused by AOC‐derived fluidDecoupled Sr‐Nd isotopes plus heavy Ba isotope is a good tracer for recycled altered oceanic crust (AOC) in arc magmasAOC and serpentinite show synergistic behaviors during subduction recycling [ABSTRACT FROM AUTHOR]

Dehydration of subducting oceanic lithosphere (slab) induces Li‐isotope fractionation between the fluid and the slab, suggested by the δ7Li variation (∼10‰) in exhumed subduction complexes. Given that arc magmas represent melt of the supraslab mantle, a large δ7Li variation is anticipated for arc volcanic rocks. However, the δ7Li values in these rocks are mostly homogeneous within the range of mid‐ocean ridge basalts (+1.6 to +5.6‰). The lack of a subduction‐related δ7Li signature has been explained by (a) homogenization by mixing of different magma sources, (b) loss of Li from the slab via dehydration, or (c) homogenization by diffusive exchange of slab‐derived Li and the mantle. The Chugoku district in SW Japan is an ideal place to study the process responsible for Li‐isotope variation in arc magmas, since the Chugoku volcanic rocks show large δ7Li variation (−1.9 to +7.4‰). High δ7Li values (+6.3 to +7.4‰) are found in some high‐Sr andesites and dacites (adakites) whereas low δ7Li values (−1.0 to −0.1‰) are found in high‐Mg andesites. The parental magmas of these rocks have been sourced from subducted oceanic crust and sediments, respectively, with various extents of the interaction with wedge mantle. The limited extents of Li isotope modification are indicated by the similarity of the δ7Li values of these rocks and their supposed sources. The models for a slab dehydration and a diffusive exchange between slab‐derived melt and mantle demonstrate that the δ7Li signatures of the sources can be preserved in the adakites if they ascent rapidly in mantle. Plain Language Summary: Many of Earth's volcanoes occur near the trenches, where one plate subducts beneath another. These volcanoes are formed as a result of the melting of mantle that contains materials of subducted plates and the transportation of melts through the mantle. A quantitative understanding of these important processes has been hampered by the lack of geochemical tools to examine consecutive chemical reactions beneath the volcanoes. Here we conducted a lithium isotope investigation of the volcanic rocks from the Chugoku district in Southwest Japan. This district hosts the volcanoes of andesites and dacites which have been derived from the melting of materials in the subducted plate. We found that these volcanic rocks preserve Li‐isotope compositions of the original materials when they melted. This in turn suggests that these magmas ascent rapidly (<300,000 years) in the mantle. Key Points: Volcanic rocks in the Chugoku district in Southwest Japan show large δ7Li variation (−1.9‰ to +7.4‰)Li isotopes fractionate to the limited extent during arc magma productionSlab‐derived δ7Li features can be preserved if slab‐derived fluids/melts have distinct δ7Li values from mantle and ascend rapidly [ABSTRACT FROM AUTHOR]

Biotite (sensu lato) is a widespread rock‐forming mineral in magmatic rocks that can be stable in a broad range of pressure and temperature, but appropriate biotite thermometers or barometers are lacking. Based on a collected experimental dataset (n = 839, T = 625–1,325°C, P = 1–48 kbar) containing biotites that span a wide compositional range [e.g., Mg/(Mg + Fe) = 0–1, TiO2 = 0–9 wt%], we have trained several machine learning algorithms for calibrating a biotite thermobarometer. Our evaluation on model performance reveals that the thermobarometry derived from extremely randomized trees is the best option, which returns coefficients of determination (R2) ≥0.97 for estimating both temperature and pressure using either biotite‐only or biotite + melt model. The model reliability were evaluated using three different approaches for the biotite‐only and biotite + melt thermobarometers respectively, including Monte Carlo cross‐validation (RMSEs are 65°C and 4.7 kbar, 38°C and 3.2 kbar, respectively), testing with independent test set (RMSEs are 54°C and 4.4 kbar, 35°C and 2.4 kbar, respectively), and error propagation from assumed analytical uncertainty (2*MAD are 54°C and 1.27 kbar, 10°C and 1.26 kbar, respectively). Quantified relative importance of involved components in the thermobarometers supports an intrinsic control of thermodynamics in the stability of biotite as a function of pressure and temperature. We applied the new biotite thermobarometer for biotite‐bearing andesitic, phonolitic and rhyolitic volcanic systems provide reliable constraints of temperature and pressure for magma storage, ascent, and evolution. We also offer a user‐friendly webpage for online performance of the thermobarometers (https://lixiaoyan.shinyapps.io/Biotite_thermobarometer/). Plain Language Summary: Biotite is a common rock‐forming mineral in a variety of igneous rocks, and its chemical composition has been widely used to retrieve associated magma properties and magmatic processes. Temperature and pressure are the two most important intensive parameters of magmatic processes, but a reliable biotite‐based thermobarometry is lacking in literature, in contrast to the considerable attentions that have been paid to other rock‐forming rocks such as clinopyroxene and amphibole. In this study, we provide a robust biotite thermobarometer based on machine learning methods. Several potentially useful algorithms have been tested on a carefully collected experimental data that include full information of temperature, pressure, and biotite and coexisting melt compositions. Our evaluations demonstrate that an optimized thermometry model comes from the method called extremely randomized trees. We also applied the proposed biotite thermobarometer to natural biotites from volcanic eruptions, including andesite (Central Volcanic Zone of Andes), phonolite (Canary Islands), and rhyolite (Bishop Tuff), and the estimated temperature and pressure are informative and complementary to the pre‐eruption conditions estimated from other thermobarometers. In addition, we also provide a webpage that can be easily used for applying our biotite thermobarometry. Key Points: Experimental data of biotite‐bearing magmas are used to train machine learning algorithms for establishing thermobarometerThe method of Extremely Randomized Trees is demonstrated to provide an optimal biotite‐based thermobarometerApplication of the thermobarometer to biotite‐bearing volcanic systems provides reliable constraints on magmatic processes [ABSTRACT FROM AUTHOR]

The effects of buoyant ridge subduction have been researched for decades. However, it remains unclear how this process influences magma chemistry. Here we use a compilation of geochemical data, well‐established geochemical proxies (i.e., Ba/Nb, Th/Nb) and mantle redox modeling (i.e., V/Sc, Cu) to propose that the subduction of a ridge underneath the central portion of the Vanuatu arc causes shallow‐angle subduction and the development of a slab tear. We suggest that the shallow slab pinches out the asthenospheric mantle and bulldozes ancient metasomatized lithospheric mantle from the forearc toward the main‐arc. Slab‐fluxed melting of this bulldozed material could account for the along‐arc 87Sr/86Sr‐143Nd/144Nd variations of the Vanuatu magmas. The influx of hot sub‐slab material into the Vanuatu arc mantle wedge through a slab tear produces magmatism within the forearc. Modeling of V/Sc and Cu systematics suggest that the mantle source of the forearc magmas has a higher fO2 and Cu content than the source of the main‐arc and rear‐arc samples. The main‐arc and rear‐arc mantles were metasomatized by both slab‐derived fluids and melts. Whereas release of high Cu‐SO2 slab‐derived fluids caused oxidation and Cu enrichment of the forearc mantle. These systematics indicate a decrease in the fO2 of slab fluxes with increasing depth‐to‐slab and distance‐from‐trench. Our findings highlight the role of ridge subduction in controlling the along‐arc and across‐arc variations in the chemistry of Vanuatu arc magmas. This updated geodynamic model, based on geochemistry, is consistent with recent geophysical constraints and 3D numerical modeling. Plain Language Summary: Subduction of oceanic plates (slabs) at convergent plate margins (trenches) causes the slab to heat up and release fluids, which causes the overlying mantle wedge to melt and produce magmatism. Some parts of the slab (e.g., ridges) are less dense than others and therefore subduct at shallower angles. This can cause (a) the subducting slab to tear and (b) the material above the subducting slab to be pushed away from the trench toward the main‐arc (a process termed "bulldozing"). It remains unclear whether the effects of ridge subduction can be detected using the geochemistry of subduction‐related arc magmas. We use a combination of geochemical data and modeling to suggest that the latitudinal variations in the chemistry of Vanuatu arc samples results from "bulldozing" effects associated with shallow angle ridge subduction. We suggest that the mantle near the Vanuatu trench might be a copper‐rich reservoir, which would not normally melt because it is too cold. However, the inflow of hot sub‐slab mantle through a slab tear appears to provide the heat necessary to melt this source and generate Cu‐rich magmas anomalously close to the trench. Hence, ridge subduction appears to have a profound influence on arc magma chemistry. Key Points: Buoyant ridge subduction causes shallow‐angle subduction and a slab tear to develop beneath the Vanuatu arcAlong‐arc changes in Sr‐Nd isotopes of Vanuatu lavas are linked to slab‐fluxed melting of bulldozed keel as a result of shallow subductionInflux of hot sub‐slab material through a slab tear into the cold and oxidized forearc mantle produces forearc magmas with high fO2 and Cu [ABSTRACT FROM AUTHOR]

We integrate textural and in situ compositional information from plagioclase and clinopyroxene (Cpx) phenocrysts together with groundmass compositions in early Cretaceous andesite dykes within the Sulu belt of China to propose a new petrogenetic model for andesite. Plagioclase phenocrysts are mostly andesine; they are depleted in high field strength elements (HFSE). However, Cpx phenocrysts are either reversely zoned (type I) or homogeneous (type II), with the zoned Cpx divided into subtypes IA and IB. All Cpx has high Mg#, low Na2O and generally low Al2O3, with depletions in HFSE and variably high 87Sr/86Sr ratios, suggesting crystallization above the Moho from magmas derived from enriched lithospheric mantle. The cores of type IA/IB and type II Cpx have normal major‐ and trace‐element compositional variations and similar 87Sr/86Sr ratios to each other and to plagioclase, consistent with fractional crystallization from a common magma (magma 1). The rims of type IA and IB Cpx also have normal major‐ and trace‐element compositional variations, but these are not as evolved as the cores, and the rims have lower 87Sr/86Sr ratios, demonstrating crystallization from an isotopically distinct magma (magma 2). Based on modeled major and rare earth element compositions of magmas inferred to have been in equilibrium with different Cpx (±plagioclase) domains, the measured groundmass compositions can be reproduced by variable mixing between the two magmas. Our study demonstrates for the first time that andesite magma can be made through fractionation and shallow hybridization of magmas derived from variably enriched lithospheric mantle. Plain Language Summary: Andesite is the characteristic volcanic rock that is formed where Earth's tectonic plates are thrust one under another, as occurs along the Andes, the Cordillera of Central and North America, and the western margin of the Pacific Ocean, but how the magmas form remains controversial. In the mantle model, an enriched source is thought to form by reaction between mantle peridotite and metamorphic fluids derived from seafloor sediments and/or altered oceanic lithosphere during under‐thrusting of an ocean plate beneath a continental margin. By contrast, the shallow model proposes that andesite magmas are formed by differentiation of basaltic magmas or mixing of magmas from mantle and crustal sources in the lithosphere of an overriding continental plate. In either case, the composition of the magma may evolve due to crystal fractionation or be modified by crustal assimilation. Here, we use petrographic observations and in situ major‐ and trace‐element, and Sr isotope compositions of plagioclase and clinopyroxene phenocrysts and groundmass in early Cretaceous andesite dykes from the Sulu belt of China, to show for the first time that andesite can be made through fractionation and shallow hybridization of two different magmas derived from variably enriched lithospheric mantle of the underlying continental plate. Key Points: Using compositions of plagioclase/clinopyroxene phenocrysts and the groundmass in andesite we develop a new model of andesite petrogenesisThis model requires polybaric crystallization and hybridization between 2 enriched lithospheric mantle‐derived magmas at crustal depthWe demonstrate that andesite of similar composition to continental crust can be made through mixing magmas derived from the enriched subcontinental lithospheric mantle [ABSTRACT FROM AUTHOR]

Axial Seamount is a submarine volcano on the Juan de Fuca Ridge with enhanced magma supply from the Cobb hotspot. We compare several deformation model configurations to explore how the spatial component of Axial's deformation time series relates to magma reservoir geometry imaged by multi‐channel seismic (MCS) surveys. To constrain the models, we use vertical displacements from seafloor pressure sensors and repeat autonomous underwater vehicle (AUV) bathymetric surveys between 2016 and 2020. We show that implementing the MCS‐derived 3D main magma reservoir (MMR) geometry with uniform pressure in a finite element model with uniform elastic host rock properties poorly fits the geodetic data. To test the hypothesis that there is compartmentalization within the MMR that results in heterogeneous pressure distribution, we compare analytical models using various horizontal sill configurations constrained by the MMR geometry. Using distributed pressure sources significantly improves the Root Mean Square Error (RMSE) between the inflation data and the models by an order of magnitude. The RMSE between the AUV data and the models is not improved as much, likely due to larger uncertainty of the AUV data. The models estimate the volume change for the 2016–2020 inter‐eruptive inflation period to be between 0.054 and 0.060 km3 and suggest that the MMR is compartmentalized, with most magma accumulating in sill‐like bodies embedded in crystal mush along the western‐central edge of the MMR. The results reveal the complexity of Axial's plumbing system and demonstrate the utility of integrating geodetic data and seismic imagery to gain insights into magma storage at active volcanoes. Plain Language Summary: Axial Seamount is a submarine volcano on the Juan de Fuca Ridge (NE Pacific Ocean) with enhanced magma supply from the Cobb hotspot. Its frequent activity and long‐term deformation time series covering eruptions in 1998, 2011, and 2015 make it an ideal place to study volcanic processes. Improved magma reservoir modeling at Axial will aid in understanding how magma transport and storage are related to surface deformation, seismicity, and eruption timing. Here we compare several models of Axial's magma reservoir to explore how the spatial component of the observed deformation at Axial compares to seismically imaged magma reservoir geometry. To constrain the models, we use vertical displacements covering an inflation period between 2016 and 2020, derived from pressure measurements collected at seafloor benchmarks and repeated bathymetric surveys. The models estimate the volume change for the 2016–2020 inflation period to be between 0.054 and 0.060 km3. Our results suggest that Axial's magma reservoir is compartmentalized, with most magma accumulating in sill‐like bodies embedded in crystal mush. The results reveal the spatial complexity of Axial's plumbing system and demonstrate how deformation data and seismic imagery can be used together to gain deeper insights into magma storage at active volcanoes. Key Points: Uniform pressurization of Axial Seamount's seismically imaged magma reservoir does not adequately fit the observed geodetic dataOur models estimate that Axial's magma reservoir inflated by 0.054–0.060 km3 during the inter‐eruptive recharge period between 2016 and 2020Axial's magma reservoir is likely compartmentalized, with magma accumulating in sills along the western‐central edge of the magma reservoir [ABSTRACT FROM AUTHOR]

In regions with small ocean plates and continental blocks, such as the Mediterranean in the Cenozoic and modern Indonesia, many blueschists may be stored at shallow depth in the mantle lithosphere that is newly formed during the amalgamation of small continental blocks and oceans. Melting of these blueschists at low pressures in the mixed lithosphere of this type during postcollisional relaxation may be involved in the generation of postcollisional orogenic magmas. Here, we have conducted a series of melting experiments on three blueschist samples with differing protoliths at 800–850°C, 2 GPa to investigate phase relations, mineral compositions, and trace element partitioning (using laser ablation‐inductively coupled plasma‐mass spectrometry [LA‐ICP‐MS]/nanoscale secondary ion mass spectrometry [NanoSIMS]), including the first results for clinozoisite. Dacitic/rhyolitic melts (SiO2 60 to 66 wt%) containing ~13 wt% volatiles are produced in all runs. Mn‐rich and Mn‐poor almandine garnets exhibit distinct rare earth element (REE) distribution patterns, while Mn‐rich garnets are effective preconcentrators of heavy REEs (HREEs) for later metamorphic processes. Melts produced from fluid‐affected blueschists of terrigenous origin show the most large‐ion lithophile element (LILE) enrichment, whereas those from terrigenous blueschists with no fluid history exhibit the lowest K/Th and Ba/Th and strongest Ce/Pb and Nb/U fractionations. Melts of blueschists derived from mid‐ocean ridge basalt (MORB) have relatively consistent, MORB‐like Ba/Th and Nb/U (~3) but the lowest Ce/Pb ratios (3.3). V/Sc and Zr/Hf ratios remain constant at ≥850°C regardless of the origin of the blueschist. Dehydration melting in new mantle lithosphere where blueschist is stored at shallow depth will significantly affect trace element behavior and should be considered carefully when examining chemical cycling and magmatism in postcollisional orogenic settings. Key Points: Mn‐rich garnets are effective preconcentrators of HREEsProtolith origin affects melt compositions and trace element partitioningBlueschist could be stored at shallow depth in the lithospheric mantle [ABSTRACT FROM AUTHOR]

The thermal properties of the Earth's primordial magma are the key factors that constrained crystallization and other thermo‐chemical processes in Earth's primitive magma ocean and therefore controlled the Earth's long‐term evolution. Thermal conductivity of the primordial magma is conventionally assumed to be a constant of about 4 W m−1 K−1 under the high pressure‐temperature conditions of the primitive magma ocean. Here we measured the lattice thermal conductivity of a variety of basaltic and silicate glasses at high pressures and a wide range of temperatures. Our results suggest that the primordial magma, if it is indeed represented by basaltic melts, had a thermal conductivity of ∼1.0–1.9 W m−1 K−1, much lower than previously thought. Such low thermal conduction reduced heat loss and thus prolonged the cooling time of the early magma ocean, promoting convection in the solidifying mantle and preventing a global overturn. Moreover, if the seismic ultralow velocity zones presently observed in the lowermost mantle are made of basaltic melts, originating either from remnants of the primitive magma ocean or pieces of subducted crust, the material in these zones must have an ultralow thermal conductivity, which would reduce cooling and thus influence the thermo‐chemical evolution of the present day core‐mantle boundary. Plain Language Summary: It is believed that a global magma ocean existed in the early Earth's interior. The thermo‐chemical evolution of the primitive magma ocean holds the key to understand the Earth's subsequent evolution and even its present day structure. Thus, knowledge of the fundamental physical properties of the primitive magma ocean, in particular, the thermal properties, would significantly advance our comprehension of our planet's history. Here, we precisely determined the thermal conductivity of a series of basaltic and silicate glasses at the high pressure‐temperature conditions assumed to be characteristic of the primitive magma ocean. Our results suggest that the primitive magma ocean, if it indeed consisted of basaltic melts, had a thermal conductivity much lower than previously assumed, which substantially slowed down cooling and allowed convection of the magma ocean before it entirely solidified. Furthermore, if the ultralow seismic velocity zones observed at present are remnants of the primitive magma ocean, their exceptionally low thermal conductivity should hinder heat transfer and thus strongly affect the thermo‐chemical state of the core‐mantle boundary. Key Points: We measured lattice thermal conductivity of basaltic and silicate glasses at high pressures and wide range of temperaturesIf primordial magma can be modeled as basaltic melts, its thermal conductivity in the mantle is much lower than previously thoughtSuch poorly conductive primitive magma ocean has a longer lifetime, facilitating solid‐state convection in the solidified mantle [ABSTRACT FROM AUTHOR]

Magma supply likely exerts primary control on seafloor morphology of oceanic crust, but most studies have related morphology to spreading rate. Here we examine global patterns of morphology on mid‐ocean ridge (MOR) flanks in relation to magma supply derived from residual mantle Bouguer gravity anomaly (proxy for relative crustal thickness) and spreading rate. We use multibeam bathymetry to characterize morphology using both qualitative (descriptive) and quantitative approaches, and we compare results to both magma supply and spreading rate. Morphology becomes more isotropic and abyssal hills are more irregular and discontinuous as magma supply decreases, while roughness, area of steeper slopes, and anomalous fabric orientation increase. We interpret these changes to reflect changing magma distribution along‐axis, from large‐volume and spatially extensive to progressively reduced, increasingly localized, and more irregularly emplaced. Observed relations between crustal thickness and morphology imply that average thickness of purely magmatic crust in the Atlantic and parts of the Indian ridge system is significantly less than average seismically determined crust. Thus seismically defined crustal thickness in those regions likely includes significant non‐magmatic components such as serpentinized mantle. Excepting regions of extensive mantle exposure, most morphologic parameters that we examined are sensitive to estimated magma supply but not necessarily to spreading rate alone. We summarize our results in schematic models that relate morphologic variations to changes in magma supply and mantle serpentinization throughout the global MOR system. Finally, we note that combined qualitative and quantitative results of our study may be useful for developing automated morphologic classification schemes. Plain Language Summary: Mid‐ocean ridges encircle the globe and spread apart along their axes to form ocean crust. Seafloor morphology of crust on the flanks of different ridges can be highly variable. The variations classically have been attributed to differences in spreading rate but there are notable exceptions to this correlation. Here we use changes in gravity over different mid‐ocean ridges (MORs) together with spreading rate to infer magma supply to the ridges, and we examine how morphology varies with changes in the magma supply. We find that features such as long, linear, and structurally continuous abyssal hills correlate well with high magma supply while progressively rougher, steeper, and more irregular morphology forms as magma supply is reduced; this is observed irrespective of spreading rate, indicating that magma supply better predicts morphologic style. Observed relations between crustal thickness and morphology imply that seismically determined crustal thickness in the Atlantic and parts of the Indian Ocean MOR system includes significant components of non‐magmatic crust (e.g., altered mantle). Our study characterizes crustal morphology both qualitatively (visually) and quantitatively; correlations between these characterizations may be useful for developing automated morphologic classification schemes. Key Points: Variable morphology of global mid‐ocean ridge flanks is examined qualitatively and quantitatively using multibeam bathymetric dataChange in magma supply as derived from gravity and spreading rate, plus mantle serpentinization, explains observed morphologic variationsMorphologic changes indicate thinner magmatic crust than seismically determined crust in the Atlantic and parts of the Indian Ocean [ABSTRACT FROM AUTHOR]

Recent geodetic data has clearly imaged significant contraction of rift zones during effusive eruptions, which is attributed to deformation of active, elongated feeders, challenging the rigid conduit paradigm. In this work we develop a physical model to understand the impact of conduit deformation on the eruptive dynamics. Numerical calculations show that magma overpressure increases the width of the conduit, resulting in higher discharge rates than would be expected from the rigid case. At the same time, conduits with high aspect ratio, have larger compressibility and can store significant amount of magma, thereby acting as secondary reservoirs. The net result is that rift zones can maintain high fluxes over prolonged periods of time, leading to large volume eruptions and biasing magma compressibility estimates. We apply our findings to the 2018 Kīlauea eruption where episodic collapse of the summit led to pressure pulses that propagated down‐rift and that were recorded by both tiltmeters and peaks in the effusion rates. Inversion of the data indicates a conduit with a height of 700–800 m and a maximum opening of 4 m, located at a depth of 2.5 km in the Upper East Rift Zone, becoming shallower in the Puʻu ʻŌʻō region and propagating sub‐horizontally in the Middle and Lower East Rift Zone. Based on these properties we infer that up to 30% of the erupted volume can be attributed to magma stored in the rift zone. In agreement with recent studies, we find that magma over‐pressure was low at the end of the 2018 eruption. Plain Language Summary: Deformation of the volcanic edifice shows a contraction of the conduits feeding eruptions. We have developed a physical model to understand how such deformation affects the amount of magma extruded and for how long the eruption will last. The results of the model indicate that elastic conduit deformation allows for larger volumes of magma to surface more quickly, compared to cases where the conduit behaves rigidly. One important finding was that elastic conduits act as secondary magma reservoirs. We applied the model to estimate the properties of the storage and transport system of the 2018 Kīlauea eruption. We found that the conduit had a height of about 750 m and a maximum opening of 4 m. The conduit is located at a depth of about 2.5 km in the Upper East Rift Zone, becoming shallower in the Pu'u 'Ō'ō region and propagating sub‐horizontally in the Middle and Lower East Rift Zone. As much as 30% of the total erupted volume could have originated from the conduit itself. Indeed, the pressure that drove the flow of the magma was low at the end of the eruption. Key Points: Geodetic data show significant deformation of rift zones during eruptionsA physical model of elastic conduit is derived to understand deformation effects on eruption dynamicsThe model is applied to 2018 Kīlauea eruption to invert the parameters of its plumbing system [ABSTRACT FROM AUTHOR]

To evaluate the effect of melt viscosity on bubble nucleation, we formulated the homogeneous nucleation rate of water bubbles to explicitly include melt viscosity. The viscosity coefficient appears in the preexponential factor of the nucleation rate in terms of the Péclet number: the ratio of the bubble growth timescale by molecular diffusion and the viscous relaxation timescale. The preexponential factor is almost constant when viscosity is low (or a high Péclet number), whereas it linearly decreases with increasing viscosity (or a decreasing Péclet number) exceeding the crossover value of viscosity, under a given supersaturation. The crossover point depends on whether homogeneous or heterogeneous nucleation takes place. We numerically solved the evolution of bubble nucleation and growth processes in ascending magmas by using the new nucleation rate formula and a precise approximation of moment equations of the bubble size distribution function. The resultant bubble number density has two regimes, similar to the previous study, but the transition point between the diffusion‐controlled regime and the viscosity‐controlled regime moves to higher viscosity or higher decompression rates by 0.6 log units at the maximum. In the viscosity‐controlled regime, the effect of the better approximation of bubble size distribution moment equations reduces bubble number density by a few orders of magnitude compared with the previous study. As a result of compiling the past laboratory experimental data, it turned out that all the experiments are conducted under the conditions equivalent to the diffusion‐controlled regime. We propose an experimental condition to confirm the presence of the viscosity‐controlled regime. Key Points: We theoretically derived the formula of the homogeneous nucleation rate of water bubbles explicitly including melt viscosity effectThe exactly approximated bubble size distribution moment equations reduced the bubble number density in highly viscous regionsWe compiled past laboratory experimental data and showed that they are outside the viscosity‐controlled regime [ABSTRACT FROM AUTHOR]

Previous work at Fernandina, the most active volcano of the Western Galápagos (Ecuador), revealed evidence for both a shallow and a deep magma reservoir, but the relative contribution of the two reservoirs to eruptions remains unclear. Here we investigate the September 2017 circumferential eruption and the June 2018 radial eruption using interferometric synthetic aperture radar data and geodetic modeling. Our results show that during the 2017 eruption magma was simultaneously withdrawn from the deep reservoir, injected upwards through the shallow reservoir, and then fed the circumferential feeder dike to the SW of the caldera. Two episodes of inflow of new magma occurred in both the deep and shallow magma reservoirs in the inter‐eruptive period from December 2017 to May 2018. During the 2018 eruption, both reservoirs fed two radial feeder dikes below the north flank, probably interacting with an underlying peripheral melt pocket, and an inclined sheet below the NW sector of the caldera. Our results highlight the primary role of the deeper reservoir which accumulates most of the magma before eruptions. Both eruptions were characterized by rapid magma transfer from the deeper to the shallower reservoir. This is similar to what is observed at the nearby Wolf volcano, but unlike nearby Sierra Negra, where a shallower reservoir accumulates higher volumes of magma before eruptions. These differences in the pre‐eruptive role of the deeper and shallower reservoirs might be related to the different evolutionary stages of Fernandina and Wolf with regard to the more mature Sierra Negra. Plain Language Summary: Fernandina is the most active volcano of the Western Galápagos (Ecuador). Two magma reservoirs at different depths below the volcano trigger eruptions both near the caldera rim (circumferential eruptions) and lower on the volcano flanks (radial eruptions). The role played by the two reservoirs in the eruptions is poorly understood. Here we studied the 2017 and the 2018 eruptions that occurred at Fernandina by using deformation data derived from radar satellites to investigate the inputs and outputs of new magma into the two reservoirs when eruptions occurred. The 2017 eruption had similar characteristics to the 2005 eruption, while the 2018 eruption had more complex mechanisms of shallow magma transfer, with magma transported in multiple directions below the north flank of Fernandina. We found that rapid magma transfer from the deep to the shallow reservoir provided most of the magma for both eruptions, similar to the nearby Wolf volcano, and unlike nearby Sierra Negra, a feature that may be due to the evolutionary stage of the volcanoes. Key Points: The 2017 circumferential eruption at Fernandina shows similar characteristics to the 2005 eruption, occurred in the same areaThe 2018 eruption is connected to the propagation of multiple dikes and sills, with most of the magma remaining intrudedBoth eruptions were characterized by rapid magma transfer from the deeper to the shallower reservoir, similar to what is observed at Wolf [ABSTRACT FROM AUTHOR]

Resolving the geochemical discrepancies in Nb/Ta between the bulk continental crust and its parental mantle melts can provide insights into the processes by which basaltic magmas differentiated into the felsic crust. Deep‐seated garnet pyroxenite cumulates (i.e., arclogite) have high Nb/Ta values due to the presence of high Nb/Ta‐rutile. For this reason, rutile‐bearing arclogites having Nb‐rich composition were proposed to complement the low Nb/Ta nature of the continental crust and evolved arc magmas, that is to say, continental arc magma evolution dominantly occurred in the stability field of garnet. However, this suggestion is challenged by the incomplete understanding of the role of rutile in the Nb/Ta fractionation during arc magma evolution due to insufficient partitioning data at low temperatures (T ≤ 1,000°C). Here, we determined Nb and Ta partitioning ratios in rutile‐felsic melt systems at 2.0 GPa, 850–1,000°C and ∼4–20 wt.% H2O. Our results confirm that equilibrium crystallization of rutile with respect to Nb and Ta will not impart lower Nb/Ta values to any derivative liquids, whereas diffusive fractionation of Nb/Ta in siliceous melts during the growth of rutile can reduce Nb/Ta ratios in melts. However, a positive correlation between Nb/Ta and Dy/Yb for mature continental arc magmas, incipient continental arc magmas, island arc magmas, and global arc magmas indicates that amphibole rather than rutile‐bearing arclogite was responsible for causing low Nb/Ta ratios of many evolved intermediate rocks. Accordingly, the chemical differentiation from primary basaltic to average andesitic compositions of the continental crust mainly occurred in the stability field of amphibole. Key Points: Equilibrium crystallization of rutile will not reduce melt Nb/Ta, whereas Nb/Ta diffusive fractionation in melt is a viable mechanismNb/Ta correlates positively with Dy/Yb for arc magmas, suggesting fractionation of amphiboles, rather than rutile‐bearing arclogitesThe chemical differentiation of continental arc magmas mainly occurred in the amphibole stability field [ABSTRACT FROM AUTHOR]

Towada volcano is an active volcano in northeast Japan, that caused two large caldera‐forming eruptions at 36 ka (episode N) and 15.5 ka (episode L). Petrological analyses and phase equilibrium experiments were performed on silicic endmember rhyolitic pumices from these two eruptions to constrain the pre‐eruptive magma storage conditions. A common mineral assemblage of plagioclase + orthopyroxene + clinopyroxene + magnetite + ilmenite was observed in both eruptions, whereas amphibole (hornblende) was observed only in the episode L pumice. Mineral thermometers presented temperatures of 832–883°C and 802–870°C for the episode N and L pumices, respectively, with an oxygen fugacity of ∼nickel–nickel oxide (NNO) + 1 in log units. The water‐saturation pressures evaluated using the plagioclase‐melt hygrometer were in the range of ∼100–200 MPa. Experiments at 100–350 MPa and 825–900°C under water‐saturated and NNO‐buffered conditions successfully reproduced the mineral assemblages in both pumices, except for magnetite. Further constraints by phase compositions and phase proportions resulted in the preferred storage conditions of 840–850°C and 150–170 MPa for both eruptions. The emergence of hornblende in the episode L magma was likely caused by the higher CaO content compared to the episode N magma and not by the difference in the storage conditions. The storage depths at ∼5–7 km coincide with the depth of the low seismic velocity beneath the present Towada volcano, suggesting possible magma accumulation at the same depths as in the past caldera‐forming eruptions. Plain Language Summary: Experimental constraints on the pre‐eruptive storage conditions of voluminous SiO2‐rich magma are fundamentally important to mitigate the risks of serious future volcanic hazards. Towada volcano is an active volcano located in northeast Japan, that caused two large caldera‐forming eruptions 36 and 15.5 kyr ago. Hornblende is present only in the later eruption, potentially reflecting differences in magma storage conditions and/or bulk compositions. We conducted water‐saturated phase equilibrium experiments on the most SiO2‐rich pumices derived from two caldera‐forming eruptions at Towada volcano to find the pressure and temperature conditions that best explain the mineral assemblage, phase compositions, and phase proportions recorded in the natural pumice. Our preferred pre‐eruptive conditions were 840–850°C and 150–170 MPa for both eruptions under water‐saturated conditions at oxygen fugacity buffered by nickel–nickel oxide. We emphasize the importance of subtle differences in the whole rock compositions to crystallize hornblende in the later eruption. The pre‐eruptive storage depths of 5–7 km coincide with the depth of the low seismic velocity signal beneath the present Towada volcano. This suggests that magma, if present, is stored at the same depth as in the caldera‐forming stage. Key Points: Water‐saturated phase equilibrium experiments successfully reproduced the mineral assemblages in the two magmas, except for magnetitePre‐eruptive depths at 5–7 km for the two magmas coincide with the depth of the low seismic velocity signal beneath the volcanoHornblende crystallization in the later eruption was likely due to the difference in bulk compositions, but not in storage conditions [ABSTRACT FROM AUTHOR]

Simulating magma propagation pathways requires both a well‐calibrated model for the stress state of the volcano and models for dike advance within such a stress field. Here, we establish a framework for calculating computationally efficient and flexible magma propagation scenarios in the presence of caldera structures. We first develop a three‐dimensional (3D) numerical model for the stress state at volcanoes with mild topography, including the stress induced by surface loads and unloading due to the formation of caldera depressions. Then, we introduce a new, simplified 3D model of dike propagation. Such a model captures the complexity of 3D magma trajectories with low running time, and can backtrack dikes from a vent to the magma storage region. We compare the new dike propagation model to a previously published 3D model. Finally, we employ the simplified model to produce shallow dike propagation scenarios for a set of synthetic caldera settings with increasingly complex topographies. The resulting synthetic magma pathways and eruptive vent locations broadly reproduce the variability observed in natural calderas. Plain Language Summary: Understanding the pathways that bring magma from an underground chamber to the surface helps to prepare for future eruptions in volcanic areas. Dikes are fractures filled with magma and represent the most common mechanism of magma transport in the Earth's crust. Their trajectories may be curved if the Earth's crust is deformed by the load of topography or by tectonic forces. Here we first discuss a model of such deformation processes in volcanic regions with complex but mild topography. Then, we develop a simplified dike propagation model that we compare to a more sophisticated one. Next, we combine our models and simulate magma pathways in artificially‐generated scenarios. Key Points: We present numerical models of crustal stress state in the presence of caldera structuresWe develop a fast dike propagation model and validate it on a previous numerical modelWe combine our stress and dike models to simulate magma pathways at synthetic calderas [ABSTRACT FROM AUTHOR]

We introduce a new approach, based on machine learning, to estimate pre‐eruptive temperatures and storage depths using clinopyroxene‐melt pairs and clinopyroxene‐only chemistry. The model is calibrated for magmas of a wide compositional range, it complements existing models, and it can be applied independently of tectonic setting. Additionally, it allows the identification of the main chemical exchange mechanisms occurring in response to pressure and temperature variations on the base of experimental data without a priori assumptions. After the validation process, performances are assessed with test data never used during the training phase. We estimate the uncertainty using the root‐mean‐square error (RMSE) and the coefficient of determination (R2). The application of the best performing algorithm (trained in the range 0–40 kbar and 952–1882 K) to clinopyroxene‐melt pairs from primitive to extremely differentiated magmas of both subalkaline and alkaline systems returns a RMSE on the order of 2.6 kbar and 40 K for pressure and temperature, respectively. We additionally present a melt‐ and temperature‐independent clinopyroxene barometer in the range 0–40 kbar, characterized by a RMSE of the order of 3 kbar. Tested for tholeiitic compositions in the range 0–10 kbar, the melt‐ and temperature‐independent clinopyroxene barometer has a RMSE of 1.7 kbar. We finally apply the proposed approach to clinopyroxenes from Iceland, providing new, independent, insights about pre‐eruptive storage depths of Icelandic volcanoes. The general applicability of this model will promote the comparison between the architecture of plumbing systems across tectonic settings and facilitate the comparison between petrologic and geophysical studies. Key Points: Machine learning thermo‐barometry is a powerful tool in petrology and volcanologyMachine learning thermometers and barometers work in a wide compositional range and can be applied independently of tectonic settingThe most represented magma storage depths in Icelandic volcanoes are in the range 1–5 kbar, with modal values at about 2–3 kbar [ABSTRACT FROM AUTHOR]

The transcrustal, mush‐dominated magma storage paradigm, which posits liquid melt is heterogeneously distributed within a vertically extensive magma mush, differs significantly from classical geodetic models, where melt is stored within liquid‐dominant chambers within an elastic crust. Here, we present mechanical models consistent with transcrustal melt storage by separating the magmatic system into three domains: liquid melt lenses, surrounding crystal‐dominated poroelastic magma mush, and elastic crust. Our results indicate that pressure changes within the melt lens may induce surface displacements that approximate the displacements predicted by spheroidal pressure sources that mimic the geometry of the mush zone. Adopting constitutive parameters of the mush dependent on mush melt fraction, we show that a magma storage system will have an effective geometry inferred from surface displacements that smoothly transitions from the geometry of the melt lens to the geometry of the mush as mush melt fraction increases. This holds true across multiple storage zone geometries, including a "transcrustal" storage zone with a magma mush that extends deep in the crust. Accounting for the presence of a magma mush can lead to an increase in the estimated volume of injected or withdrawn magma (by several multiples) compared to values obtained using fully elastic models. Comparing erupted magma volumes to source volume changes allows for an estimation of magma compressibility; we show the presence of a mush can increase this estimated magma compressibility by up to approximately 50%, suggesting magmas may have higher bubble fraction than previous geodetically derived estimates. Plain Language Summary: Different kinds of geophysical observations provide differing views into the nature of magma storage regions within the crust. Many geophysical observations, such as seismic data, indicate that melt is stored within discrete, horizontally elongate pockets that are contained within a "mush" of solid crystals and liquid melt. However, surface deformations (geodetic data) are typically explained with models that ignore the presence of a magma mush. In this paper, we attempt to reconcile these views of magma storage regions, by incorporating the effects of a magma mush into models of ground deformation. Our results show that ground deformation measurements are potentially more sensitive to the magma mush than the discrete pockets of melt within the mush, and that appropriately accounting for the presence of a magma mush can change estimates of important quantities, such as the volume of injected magma and magma compressibility. Key Points: We develop a numerical poroelastic model for surface displacements induced by melt migration within a mush‐dominated melt storage regionSurface displacements may be more sensitive to the geometry of a low melt content mush than a melt‐rich lens within the mushAccounting for magma mush zones leads to important differences in estimation of quantities of interest, such as inferred volume change [ABSTRACT FROM AUTHOR]

We investigated essential factors controlling magma fragmentation during the 2011 sub‐Plinian eruptions of Kirishima Shinmoe‐dake volcano, Japan, based on observational data and the analysis of the conduit flow system. Key observational data of the Shinmoe‐dake eruptions are geophysically estimated magma discharge rate and petrologically measured crystallinity and porosity at the time of fragmentation. Using formulations of fragmentation criterion and conduit flow model, we derived analytical expressions describing the relationship among magmatic and geological parameters at the time of strain rate‐induced or stress fragmentation. The combination of the observational data and the analytical expressions revealed the relationship among weakly constrained parameters at the fragmentation, such as pressure, gas permeability in the magma, conduit radius, and maximum packing fraction of crystals, β*. From this relationship, we found that a smaller β*, less than approximately 0.5, is necessary for the fragmentation to occur for a wide range of the conduit radii under realistic pressure and gas permeability values. This reflects that in the case of the sub‐Plinian eruption with a moderate magma discharge rate, a drastic increase in magma viscosity induced by the smaller β* plays a crucial role in the occurrence of fragmentation. The strong dependence of the fragmentation process on β* indicates that the suspension rheology of crystal‐bearing magma is an essential factor controlling the magma ascent process during the moderate explosive eruption. Plain Language Summary: Magma ascent from the magma chamber to the surface through a volcanic conduit is an important process that controls the style and intensity of volcanic eruptions. In explosive eruptions, magma fragmentation during the magma ascent is a crucial process causing high‐speed ejection of gas and pyroclasts, leading to hazardous surface phenomena such as eruption clouds and pyroclastic flows. Since direct observation of the magma ascent is impossible, it is necessary to integrate indirect observations, models, and experiments for detailed understanding. We investigated essential factors controlling magma fragmentation during moderate‐intensity explosive eruptions by combining observational data and newly developed analytical expressions describing the parameter relationship at fragmentation. The target of our study is the 2011 explosive eruption of Shinmoe‐dake volcano, in which a typical moderate‐intensity eruption was densely observed by various methods. By utilizing these observations, we revealed that during moderate‐intensity explosive eruptions, a drastic increase in magma viscosity induced by magma crystallization plays a crucial role in the occurrence of fragmentation. The viscosity increase with crystallization is closely related to the microscopic structure of crystal‐bearing magma. Therefore, combining petrological information and conduit flow modeling is essential for a detailed analysis of the magma ascent process during moderate‐intensity explosive eruptions. Key Points: Observations of the Shinmoe‐dake eruptions, which provide constraints on magma fragmentationDerivation of analytical expressions describing the relationship among parameters at fragmentationStrong effects of magma rheology on fragmentation during moderate‐intensity explosive eruptions [ABSTRACT FROM AUTHOR]

The Longgang volcanic field (LVF) is one of the largest intraplate monogenetic volcanic regions in China, and eruption lasted from the early Pleistocene to the Holocene. Previous research shows that the LVF is not only experiencing rapid surface uplift but also significant seismic activity. The magmatic system under the LVF poses a hazard of potential eruption. The three‐dimensional electrical resistivity model developed by magnetotelluric data shows that the upper crust of the LVF and the surrounding areas hosts a high‐resistivity body composed of volcanic rocks produced by massive eruptions during past volcanic episodes. There are two main high‐conductivity structures below the depth of 10 km in the southwest and northeast of the LVF. The two structures merge in the lower crust and extend to the mantle. We interpret these high‐conductivity structures to be magmatic channels (the resistivity of C1 is 20–50 and C2 is 5–20 Ωm). Magmatic channel upwelling along the C1 may have caused the eruption of the Jinlongdingzi volcano 1,700 years ago. The C2 magmatic channel in the northeast of the LVF is larger and lower resistivity than C1. The modern surface uplift and seismicity in the LVF is a result of upwelling and intrusion along C2 magmatic channel. The Hunjiang fault passing through the LVF may also influence the upwelling of material along the C2 magmatic channel, posing a hazard of future eruption in the LVF. The high‐conductivity structures in the LVF pass through the Moho and extend to the upper mantle and are connected from the top to bottom. Combined with geochemical data, we speculate that the magmatic system in the LVF may directly come from the upper mantle and the magma has experienced rapid ascent with limited crustal internal evolution. Plain Language Summary: The Longgang volcanic field (LVF) is one of the main active volcanic regions in Jilin Province, northeast China. It is densely populated and economically developed, and it is also the most important food production place and tourist region in China. Therefore, the risk of volcanic and seismic activity in this area has always been the focus of attention and research. Through a geophysical method called Magnetotellurics, we develop a three‐dimensional electrical resistivity model of the LVF. The model reveals two crustal and mantle scale magmatic channels C1 and C2. The modern crustal uplift and seismicity are result of upwelling and intrusion of the C2 magmatic channel, posing a hazard of future eruption in the LVF. The results of this paper can provide deep geophysical basis for the prediction of volcanic magma system and eruption risk. Key Points: Broadband magnetotelluric data were used to generate a three‐dimensional electrical structure of the Longgang volcanic field (LVF) and adjacent areasThe modern crustal uplift and seismicity in the LVF is a result of upwelling and intrusion of the electrically conductive zone interpreted as a magmatic channelThe magmatic system in the LVF is sourced from the deep mantle and has experienced rapid uplift and limited crustal internal evolution [ABSTRACT FROM AUTHOR]

As volcanoes undergo unrest, understanding the conditions and timescales required for magma reservoir failure, and the links to geodetic observations, are critical when evaluating the potential for magma migration to the surface and eruption. Inferring the dynamics of a pressurized magmatic system from episodes of surface deformation is heavily reliant on the assumed crustal rheology, typically represented by an elastic medium. Here, we use Finite Element models to identify the rheological response to reservoir pressurization within a temperature‐dependent Standard Linear Solid viscoelastic ("thermo‐viscoelastic") domain. We assess the mechanical stability of a deforming reservoir by evaluating the overpressures required to initiate brittle failure along the reservoir wall, and the sensitivity to key parameters. Reservoir inflation facilitates compression of the ductile wall rock, due to the non‐uniform crustal viscosity, impacting the temporal evolution of the induced tensile stress. Thermo‐viscoelasticity enables a deforming reservoir to sustain greater overpressures prior to failure, compared to elastic analyses. High‐temperature (e.g., mafic) reservoirs fail at lower overpressures compared to low‐temperature (e.g., felsic) reservoirs, producing smaller coincident displacements at the ground surface. The impact of thermo‐viscoelasticity on reservoir failure is significant across a wide range of overpressure loading rates. By resisting mechanical failure on the reservoir wall, thermo‐viscoelasticity impacts dyke nucleation and formation of shear fractures. Numerical models may need to incorporate additional processes that act to promote failure, such as regional stresses (e.g., topographic and tectonic), external triggers (e.g., earthquake stress drops), or pre‐existing weaknesses along the reservoir wall. Plain Language Summary: Understanding the conditions required for magma reservoir failure, which can result in the transport of magma to the ground surface, and its links to ground deformation is a persistent challenge when assessing the eruption potential of a volcano. Numerical models are commonly used to account for different complexities observed in nature, which can significantly affect the ambient subsurface stress field or the threshold for reservoir failure. Most numerical models do not consider the impact of crustal behavior; whilst the results can be instructive, an accurate model is of paramount importance when modeling reservoir failure. Here we investigate the conditions for failure within a thermo‐viscoelastic crust, accounting for the increased temperatures and ductility surrounding the reservoir. Our results highlight that an inflating reservoir results in the compression of the surrounding crustal rock due to the non‐uniform viscosity, which is not observed in simpler elastic models. This hinders the development of the tensile stress, and may require overpressures greater than usually considered to initiate reservoir failure. As the behavior of the crust is controlled by its temperature, and can strongly affect the conditions for reservoir failure, we suggest that thermo‐viscoelasticity should be incorporated in deformation models and failure analyses. Key Points: Reservoir failure within a thermo‐viscoelastic crust is controlled by thermal heterogeneity and the associated viscous timescalesNon‐uniform crustal viscosity enables compression of the surrounding host rock, altering the tensile stress evolution around the reservoirLow‐temperature reservoirs sustain elevated overpressures prior to failure, due to long viscous timescales, increasing surface deformation [ABSTRACT FROM AUTHOR]

Crustal magmatic systems likely consist of magmatic reservoirs dominated by crystal mush. Recent studies suggest that the physical processes occurring in crystal mush could alter the response of magmatic reservoirs during volcanic unrest. Here, we present a magma chamber deformation model that incorporates two new aspects in crystal mush: heat and exsolved gas. The model is based on earlier studies by Liao et al. (2021, https://doi.org/10.1029/2020JB019395) with additional processes including thermal‐mechanical coupling, dependence of material properties on gas content, and temperature evolution following an injection of hotter magma. The post‐injection time‐dependent evolution of the system can be grouped into three periods, which are dominated by poroelastic diffusion (short term), viscoelastic relaxation (mid term), and thermal equilibration (long term). All three time‐regimes are strongly affected by gas distribution, which alters the relative compressibility of the crystal‐rich and crystal‐poor regions in the chamber. The contribution of thermal evolution emerges during the mid‐term evolution. The time‐dependent evolution of the system highlights the intrinsic ability of a gas‐bearing mushy magma chamber to generate non‐monotonic time series of stresses, deformation, and magma transport. Plain Language Summary: The processes that occur in magmatic reservoirs and their surrounding crustal rocks can modulate the triggering, duration, and style of volcanic eruptions. Although magmatic reservoirs are typically modeled as cavities filled with fluid magma, geophysical and petrological observations have long suggested that they likely contain crystal mush, an ensemble of solid crystals and fluid (gas/melt) that reside in interstitial pore spaces. Some recent studies show that a mushy magma chamber could behave differently than a fluid‐filled chamber, leading to different interpretation of observations such as ground deformation. Here, we extend these studies to incorporate two new aspects that are typical for crustal magmatic systems: gas phase (in the form of disconnected gas bubbles) and nonuniform temperature. The incorporation of these two new aspects allows for more processes that can be examined quantitatively, and a more realistic depiction of crustal magmatic reservoirs. We find that when gas and heat distribute unevenly, a simple magma injection event could result in complex time‐dependent changes in pressure, stress, and magma flows in the chamber. Key Points: A physics‐based model explores the influence of gas and heat in the a magma chamber containing crystal mush upon magma injectionThe non‐uniform distributions of gas and heat cause non‐monotonic time evolution of deformation, crustal stresses, and magma transportPoroelastic diffusion, viscous relaxation, and thermal equilibration lead to three time regimes in the post‐injection evolution [ABSTRACT FROM AUTHOR]

Olivine‐hosted melt inclusions have the potential to provide direct evidence about the nature of the mantle sources of magmas. In this study, we report the major and trace element and Sr–Nd–Pb–Hf isotope compositions of the Lijiang picrites from the Emeishan large igneous province, as well as major and trace element and Pb isotopic compositions of olivine‐hosted melt inclusions in the picrites. Using these data, we infer the nature of the mantle sources for the Lijiang high‐Ti picrites (Ti/Y = 645–654), and compare them with the Dali low‐Ti picrites (Ti/Y = 351–408). Lijiang melt inclusions have highly heterogeneous Pb isotope compositions (208Pb/206Pb = 2.015–2.201, 207Pb/206Pb = 0.812–0.882, 260 Ma age‐corrected) that overlap the majority of the fields from FOZO to enriched mantle 1 (EM1). The primitive mantle‐normalized trace element patterns of these melt inclusions reveal an enriched composition with depletions in Ba, Nb, Ta, Zr, and Hf. We suggest that an old recycled GLOSS II‐like sediment was probably the main component that imparted the EM1 signature into the sources of the Lijiang picrites. Furthermore, we reverse calculated the mantle potential temperature (Tp) beneath Lijiang and Dali using the Al‐in‐ol temperature. The estimated mantle Tp beneath Lijiang is lower than Dali. We infer that the Lijiang picrites might have originated from the mantle plume with lower thermal energy, corresponding to the waning stage of the axis of the Emeishan mantle plume. Our study suggest that high‐Ti magmas of the Emeishan large igneous province were derived from those parts or stages of the mantle plume with lower thermal energy. Key Points: Highly heterogeneous Pb isotope compositions of olivine‐hosted melt inclusions are found in the Lijiang picritesGLOSS II‐like subduction sediments are responsible for the enriched mantle 1‐like component in the source of the Emeishan large igneous province (LIP)Thermal variation of the mantle plume controlled the types and distribution of basaltic magmas in the Emeishan LIP [ABSTRACT FROM AUTHOR]

Crystals retain an imprint of the dynamic changes within a magma reservoir and hence contain invaluable information about the evolving conditions inside volcanic plumbing systems. However, instead of telling a single, simple story, they comprise overprinted evidence of numerous processes relating to temperature, pressure and composition that drive crystal precipitation and dissolution in magmatic systems. To decipher these different elements in the story that crystals tell, we attempt to identify the observational signatures of a simple, yet ubiquitous process: crystal precipitation and dissolution during magma cooling. To isolate this process in a complex magmatic system with intricate dynamic feedbacks, we assume that synthetic crystals precipitate and dissolve rapidly in response to deviations from thermodynamic equilibrium. In our crystalline‐scale simulations, synthetic crystals drag along the cooler‐than‐ambient melt in which they precipitated and can drive a temperature‐dependent, crystal‐driven convection. We analyze the non‐dimensional conditions for this coupled convection and record the heterogeneous thermal histories that synthetic crystals in this flow regime experience. We show that many synthetic crystals dissolve, loosing their thermal record of the convection. Based on our findings, we suggest that heterogeneity in the thermal history of crystals is more indicative of local, crystal‐scale processes than the overall, system‐wide cooling trend. Plain Language Summary: Similar to tree rings, crystals tell a story of their past through their sequential growth records. However, unlike trees, crystals are dynamic and can travel throughout the magmatic domain. We hypothesize that this mobility may result in a bias in the crystal record and prevent it from accurately recording the magma reservoir history. In order to understand what patterns natural sample crystals would record during cooling, we built a simulator that tracks individual crystals and the thermal environment they experience. We simulate how crystals drive flow, but also precipitate and dissolve in response to temperature change. Our results show that each synthetic crystal has a unique record of its past that is dependent on the temperature distribution within its vicinity. Additionally, most of the synthetic crystals dissolve in the convection, so we lose their record. Overall, the synthetic crystals in our simulation can reproduce some of the variable patterns that are often found in natural samples, suggesting that real crystals are imperfect witnesses of the magmatic dynamics they are exposed to. Key Points: Crystals in convection preserve the least information about the most dynamic flow domainCrystals in convection show signs of dissolution, suggesting a slower than actual cooling rateCrystals can record cooling, heating, and complex thermal histories even during steady cooling [ABSTRACT FROM AUTHOR]

The temporal evolution of effusion rate is the main controlling factor of lava spreading and emplacement conditions. Therefore, it represents the most relevant parameter for characterizing the dynamics of effusive eruptions and thus for assessing the volcanic hazard associated with this type of volcanism. Since the effusion rate curves can provide important insights into the properties of the magma feeding system, several efforts have been performed for their classification and interpretation. Here, a recently published numerical model is employed for studying the effects of magma source and feeding dike properties on the main characteristics (e.g., duration, erupted mass, and effusion rate trend) of small‐volume effusive eruptions, in the absence of syn‐eruptive magma injection from deeper storages. We show that the total erupted mass is mainly controlled by magma reservoir conditions (i.e., dimensions and overpressure) prior to the eruption, whereas conduit processes along with reservoir properties can significantly affect mean effusion rate, and thus, they dramatically influence eruption duration. Simulations reproduce a wide variety of effusion rate trends, whose occurrence is controlled by the complex competition between conduit enlargement and overpressure decrease due to magma withdrawal. These effusion rate curves were classified in four groups, which were associated with the different types described in the literature. Results agree with the traditional explanation of effusion rate curves and provide new insights for interpreting them, highlighting the importance of magma reservoir size, initial overpressure, and initial width of the feeding dike in controlling the nature of the resulting effusion rate curve. Key Points: The effects of magma source and feeding dike properties on the evolution of small‐volume effusive eruptions are studied numericallyTotal erupted mass is mainly controlled by magma reservoir conditions (in particular, dimensions and overpressure) prior to the eruptionDifferent types of effusion rate curves were modeled, classified, and compared with the effusion rate trends defined in the literature [ABSTRACT FROM AUTHOR]

Crystal‐hosted melt embayments and melt inclusions partially record magmatic processes at depth, but it is not always obvious how to interpret this record. One impediment is our incomplete understanding of how embayments and melt inclusions form. In this study, we investigate the formation mechanism of embayments and melt inclusions during quartz growth to quantify the relationship between the compositions of the entrapped and average melt. We study the growth of embayments and inclusions through direct numerical simulations that couple the growth of a crystal surface with the evolution of the concentrations of incompatible components in the surrounding melt. We find that H2O is more enriched in the interior of defects on crystal surface compared to the exterior. The resultant lower disequilibrium in the defect interior causes lower growth rate than in the exterior, elongating the defect into an embayment. If crystal growth stops, the composition in the embayment equilibrates with the average melt within days to months. If crystal growth continues until the embayment neck closes, a melt inclusion forms. The melt entrapped by both embayments and melt inclusions is enriched in incompatible components, such as H2O and CO2. In addition to inclusion size, the enrichment of incompatible components in melt inclusions also depends on component diffusivity and the crystal growth regime. High‐diffusivity components like H2O have similar enrichment levels in all scenarios, while lower‐diffusivity components like CO2 are more enriched in melt inclusions with smaller sizes or formed in continuous crystal growth. Plain Language Summary: Embayments and melt inclusions are small droplets of melt entrapped by crystals. These droplets may contain valuable clues for understanding processes inside volcanoes inaccessible to direct observation, because they typically form at depth before the onset of eruptions. Embayments are partially connected to exterior melt and can record processes during eruption. Melt inclusions are fully isolated and can record magma compositions at depth. However, we do not yet understand how they form. Our lack of understanding hinders our ability to interpret this data. Here, we use numerical simulations to quantify how the enrichment of dissolved water in the melt near the crystal affects the formation of embayments and melt inclusions during quartz growth. There is a positive feedback between the water enrichment in the embayment interior and embayment elongation, because water is not integrated into the quartz as it grows. Our simulations put constraints on the application of embayments and melt inclusions to constrain volcanic processes: For embayments, component enrichment likely equilibrates with surrounding melt prior to eruption. For melt inclusions, high‐diffusivity components like water are similarly enriched in all scenarios, while lower‐diffusivity components like carbon dioxide are more enriched in smaller inclusions formed in continuous crystal growth. Key Points: Embayments can form from crystal defects because water is more enriched inside the defect than outsideEnrichment of incompatible components in embayments likely equilibrates in pre‐eruptive magma storageCarbon dioxide is more enriched in smaller melt inclusions formed in continuous crystal growth but water has similar enrichment in all cases [ABSTRACT FROM AUTHOR]

All instrumented basaltic caldera collapses have generated Mw > 5 very long period earthquakes. However, previous studies of source dynamics have been limited to lumped models treating the caldera block as rigid, leaving open questions related to how ruptures initiate and propagate around the ring fault, and the seismic expressions of those dynamics. We present the first 3D numerical model capturing the nucleation and propagation of ring fault rupture, the mechanical coupling to the underlying viscoelastic magma, and the associated seismic wavefield. We demonstrate that seismic radiation, neglected in previous models, acts as a damping mechanism reducing coseismic slip by up to half, with effects most pronounced for large magma chamber volume/ring fault radius or highly compliant crust/compressible magma. Viscosity of basaltic magma has negligible effect on collapse dynamics. In contrast, viscosity of silicic magma significantly reduces ring fault slip. We use the model to simulate the 2018 Kı̄lauea caldera collapse. Three stages of collapse, characterized by ring fault rupture initiation and propagation, deceleration of the downward‐moving caldera block and magma column, and post‐collapse resonant oscillations, in addition to chamber pressurization, are identified in simulated and observed (unfiltered) near‐field seismograms. A detailed comparison of simulated and observed displacement waveforms corresponding to collapse earthquakes with hypocenters at various azimuths of the ring fault reveals a complex nucleation phase for earthquakes initiated on the northwest. Our numerical simulation framework will enhance future efforts to reconcile seismic and geodetic observations of caldera collapse with conceptual models of ring fault and magma chamber dynamics. Plain Language Summary: Caldera collapse manifests as the rapid subsidence of a kilometer‐scale block of crust circumscribed by a near‐circular fault on top of a volcano. The subsidence of the caldera block is caused by the eruption‐induced withdrawal of magma and reduction in pressure in the underlying magma chamber. All scientifically instrumented caldera collapses at volcanoes with low‐viscosity magma are accompanied by earthquakes of magnitude 5 and above. How do magma viscosity and the seismic wave radiation influence the amount of slip per earthquake on the fault? What can we learn about the dynamics of these earthquakes from seismic records? We address these questions by performing computer simulations of caldera collapse earthquakes and compare the results to the seismic records from the Kı̄lauea caldera collapse of 2018. Key Points: Seismic wave radiation through ring fault and magma, as well as high magma viscosity, reduce fault slip by up to half during collapseRupture propagation, downward momentum transfer via magma pressure waves, and chamber pressurization are identified in unfiltered seismogramsComparison between simulated and observed near‐field seismograms from Kı̄lauea 2018 reveals complex nucleation phase on the ring fault [ABSTRACT FROM AUTHOR]

Recent multi‐channel seismic studies of fast spreading and hot‐spot influenced mid‐ocean ridges reveal magma bodies located beneath the mid‐crustal Axial Magma Lens (AML), embedded within the underlying crustal mush zone. We here present new seismic images from the Juan de Fuca Ridge that show reflections interpreted to be from vertically stacked magma lenses in a number of locations beneath this intermediate‐spreading ridge. The brightest reflections are beneath Northern Symmetric segment, from ∼46°42′‐52′N and Split Seamount, where a small magma body at local Moho depths is also detected, inferred to be a source reservoir for the stacked magma lenses in the crust above. The imaged magma bodies are sub‐horizontal, extend continuously for along‐axis lengths of ∼1–8 km, with the shallowest located at depths of ∼100–1,200 m below the AML, and are similar to sub‐AML bodies found at the East Pacific Rise. At both ridges, stacked sill‐like lenses are detected beneath only a small fraction of the ridge length examined and are inferred to mark local sites of higher melt flux and active replenishment from depth. The imaged magma lenses are focused in the upper part of the lower crust, which coincides with the most melt rich part of the crystal mush zone detected in other geophysical studies and where sub‐vertical fabrics are observed in geologic exposures of oceanic crust. We infer that the multi‐level magma accumulations are ephemeral and may result from porous flow and mush compaction, and that they can be tapped and drained during dike intrusion and eruption events. Key Points: Multi‐channel seismic data reveal stacked magma lenses within the crystal mush zone at the Juan de Fuca RidgeStacked lenses underlie the Axial Magma Lens within the melt‐rich mid crust and span depths of the foliated gabbros in crustal exposuresMid‐crust magma lenses are interpreted to be ephemeral, formed by mush compaction and contributing magmas during eruption events [ABSTRACT FROM AUTHOR]

Persistent volcanic activity is thought to be linked to degassing, but volatile transport at depth cannot be observed directly. Instead, we rely on indirect constraints, such as CO2‐H2O concentrations in melt inclusions trapped at different depth, but this data is rarely straight‐forward to interpret. In this study, we integrate a multiscale conduit‐flow model for non‐eruptive conditions and a volatile‐concentration model to compute synthetic profiles of volatile concentrations for different flow conditions and CO2 fluxing. We find that actively segregating bubbles in the flow enhance the mixing of volatile‐poor and volatile‐rich magma in vertical conduit segments, even if the radius of these bubbles is several orders of magnitude smaller than the width of the conduit. This finding suggests that magma mixing is common in volcanic systems when magma viscosities are low enough to allow for bubble segregation as born out by our comparison with melt‐inclusion data: Our simulations show that even a small degree of mixing leads to volatile concentration profiles that are much more comparable to observations than either open‐ or closed‐system degassing trends for both Stromboli and Mount Erebus. Our results also show that two of the main processes affecting observed volatile concentrations, magma mixing and CO2 fluxing, leave distinct observational signatures, suggesting that tracking them jointly could help better constrain changes in conduit flow. We argue that disaggregating melt‐inclusion data based on the eruptive behavior at the time could advance our understanding of how conduit flow changes with eruptive regimes. Plain Language Summary: Persistently degassing volcanoes typically erupt multiple times a day, emitting copious gas and thermal energy but relatively little magma. The ascent of the erupted magma is likely a byproduct of degassing, but it is unclear how exactly eruptive behavior and degassing are related. Without the ability to observe degassing processes at depth, we rely on erupted samples to derive constraints on pre‐eruptive flow conditions. Some of these samples seal in magma droplets named melt inclusions, which represent valuable snapshots of volatile concentrations at depth. Here, we use numerical simulations to study how magma flow in vertical conduit‐like structures at depth alters the volatile concentrations in melt inclusions. We demonstrate that volatile‐rich, ascending magma mixes with volatile‐poor, descending magma that has lost volatiles at the surface. The degree of mixing depends on the physical properties of magma and bubbles. This magma mixing, together with the carbon dioxide fluxing, can significantly shift the water and carbon dioxide concentrations in melt inclusions. Our study suggests that some degree of magma mixing is almost inevitable in persistently degassing volcanoes, but that mixing may vary considerably with depth. We suggest that melt inclusion data could potentially help tracking the evolving flow conditions in volcanic conduits. Key Points: Magma mixing occurs commonly at the interface between volatile‐rich and volatile‐poor magma in persistently degassing volcanoesBubble speed, magma viscosities, and bubble volume fraction control magma mixingMagma mixing and carbon dioxide fluxing leave distinct signatures in melt‐inclusion data [ABSTRACT FROM AUTHOR]

Abstract: We present K‐Ar ages, major and trace element concentrations, and Sr‐Nd‐Pb isotope data for late Cenozoic volcanic rocks from the Chugoku district, southwest Japan arc. Andesite and dacite lavas in this region are enriched in Sr (mostly >800 μg g−1) and show geochemical characteristics of volcanic rocks commonly referred to as “adakite.” K‐Ar dating of these lavas revealed that the eruption of high‐Sr andesitic to dacitic magmas occurred during the last 2 Myr, following or concurrent with the eruption of basalt in adjacent regions. Trace‐element characteristics of high‐Sr andesites and dacites are consistent with the formation of their parent magmas by partial melting of the basaltic layer of the subducting Shikoku Basin Plate. Mass balance modeling of trace element concentrations and isotopic compositions suggests that the parental magmas of high‐Sr andesites and dacites are best explained by mixing of partial melts from oceanic crust (F = 5–15%) and sediment (F = 30%) at 80:20 to 55:45 ratios. Spatial coincidence of the occurrences of high‐Sr andesites and dacites and seismic gaps of the subducting slab demonstrates the causal link between slab melting and mantle upwelling at slab tears. We speculate that these tears could have been formed by subduction of ridges on the plate. A warm mantle upwelled through tears, preventing the solidification of the siliceous slab melts in the mantle and facilitating the transportation of these melts to the surface. [ABSTRACT FROM AUTHOR]

Abstract: Mixing of felsic magmas from different sources may be common in granite petrogenesis, but challenging to unravel. Nevertheless, this is identified by a combined study of in situ zircon and garnet geochemistry with whole‐rock geochemistry for the Jiuling batholith in South China. The Jiuling batholith is composed of low‐Si granitoids and high‐Si granites. Although they show similar whole‐rock Sr‐Nd‐Hf isotope compositions, the low‐Si granitoids exhibit higher MgO, FeO, TiO2, Al2O3, and Middle rare earth elements contents, and higher Zr saturation temperature than the high‐Si granites. Two groups of zircons and garnets are identified in the form of either individual grains or different domains of a grain. Although they exhibit highly variable O and Hf isotope compositions, Group I zircons show higher δ18O and lower εHf(t) values than Group II zircons. Group I garnets are higher in FeO and MgO but lower in CaO and MnO than Group II garnets. Group I garnets show more remarkable negative Eu anomalies than Group II garnets. Groups I and II garnets exhibit distinct δ18O values, in equilibrium with Groups I and II zircons, respectively. These differences indicate that the two groups of zircons and garnets were derived from two different batches of felsic magmas. Batch I magma may be derived from partial melting of ancient sedimentary rocks at higher temperature, whereas Batch II magma would probably form by partial melting of relatively juvenile crust at lower temperature. Therefore, the Jiuling batholith is produced by mixing between two batches of granitic magmas with different trace element and isotope compositions. [ABSTRACT FROM AUTHOR]

We present a 2D thermo‐elastic model of a magma body formation in a granitic crust by injection of rhyolitic or basaltic dikes and sills. Elastic analytical solutions enable the computation of rock displacement in response to magma intrusion and evacuation during volcanic eruptions. Phase diagrams for magma and rocks govern melting/crystallization behavior and temperature evolution in 16 million computational cells. Calculated temperature histories are used to predict zircon crystallization/dissolution, their ages, and isotopic ratios within individual batches of magma and rocks. Incremental dike injection naturally generates magma batches of melt that form kilometer‐wide clusters appearing in different parts of the domain with diverse shapes, horizontal and vertical interconnectivity. The volume of eruptive melt strongly depends on magma influx rates Q, the width of the injection region W, and eruptions. For example, rhyolitic dike injection with Q = 0.125 m3/s with W = 5 km during 100 ka generates ∼50 km3 of eruptive melt while no significant melt forms if W = 10 km. Injection of basaltic dikes into the granitic crust generates comparable amounts of rhyolitic melt from dike‐residual and country rocks. High injection intensities produce magma reservoirs capable of large eruptions, while repetitive eruptions lead to shrinkage of magma bodies. High heat input causes host rock zircons to lose a significant portion of their old cores. Magmatic zircons in the periphery form quickly due to rapid cooling of intruded dikes while in the central part crystals can grow during several hundreds of ka. The model predicts highly heterogeneous O and Hf isotopic ratios recorded in zircons. Plain Language Summary: The formation of large magmatic eruptible bodies beneath active volcanoes or plutons in the Earth's crust involves complicated thermo‐mechanical processes related to magma injection, host rock melting, solid‐melt separation, etc. Revealing enigmas requires joint efforts of geologists, geophysics, and geomechanics. With the development of modern computer architecture and numerical methods, modern models can capture details of these processes with increasing levels of detail. In this paper, we present a new model of magma chamber formation based on the intrusion of dikes and sills in 2D geometry with an extremely high spatial and temporal resolution that can capture individual batches of injected magma. We use temperature‐time histories recorded in more than 10,000 individual batches of magma with zircon crystallization software previously published by us. We are now able to describe in detail magma chamber growth and formation and link it to the history of zircon dissolution and growth. Key Points: A 2D thermo‐elastic model produces vertically extended interconnected magma system in 104 years after basaltic and rhyolitic dike injectionMagma influx rate, the width of injection zone and repetitive eruptions determine the thermal and petrological evolution of a magma chamberZircons record history of magma accretion and crustal melt/intruded magma proportions in their age and O and Hf isotopes [ABSTRACT FROM AUTHOR]

Magma migration and differentiation processes are key to understanding the development and evolution of oceanic magma reservoirs. To provide new quantitative geochemical constraints on these processes, we applied a high‐resolution approach to study an interlayered section of the lower oceanic crust sampled at Atlantis Bank, on the (ultra)slow‐spreading Southwest Indian Ridge. The section is characterized by sharp grain‐size layering between fine‐ and coarse‐grained olivine gabbros that is representative of other layered structures described at the Atlantis Bank oceanic core complex. The textures and fabrics of the layers and the nature of their contacts indicate formation by intrusion of a magma (i.e., crystal‐bearing) into an almost solidified coarse‐grained mush. Petrographic observations and in situ incompatible trace element signatures indicate that the fine‐ and coarse‐grained layers record reactive porous migration of melts. Widespread reactive porous flow occurred prior to intrusion within the coarse‐grained gabbro, producing mineral compositions enriched in incompatible elements. The intrusive fine‐grained lithology records a late stage event of localized reactive melt percolation in cm‐scale structures, which lead to strong light rare earth elements depletion relative to heavy rare earth elements. In addition, we highlight the occurrence of interactions at the contacts between layers and partial modification in compositions of the intruded lithology. This layered section likely represents a contact between two larger magma bodies emplaced within the lower crust during accretion, where the type of melt migration (intrusion or porous flow) and the modalities of melt percolation (widespread or localized) strongly govern the composition of the crustal lithologies. Plain Language Summary: We present the study of a 1m35 borehole section of oceanic crust exhumed at Atlantis Bank on the Southwest Indian Ridge. This section presents characteristic layering defined by successive fine‐ and coarse‐grained olivine gabbro layers, which recorded the conditions and modalities of magma emplacement and evolution during oceanic crust accretion. We conducted a high‐resolution study of samples collected continuously along the section, including a description of the petrography, analysis of microstructures, and geochemistry of both rocks and minerals. This high‐resolution approach allowed us to reconstruct the intricate formation steps of the section, which involves the two main melt migrations processes at depth: an intrusion, represented by the fine‐grained layers, and porous melt flow, which occurs at all steps of formation of the lithologies. The latter process strongly impacted the textures and composition of the lithologies. These compositions provide information on the modalities of the porous melt migration, which can be pervasive or focused in discrete structures. The geometry of the layer contacts together with the emplacement model suggests that this layering likely represents an irregular contact between two larger intrusions or magma reservoirs emplaced within the oceanic crust. Key Points: At slow‐spreading ridges, igneous layering in the lower crust records intrusive events of crystal‐bearing magmasAt lower crustal levels, melt migration proceeds by intrusion and reactive porous flowMelt migration modes strongly influence the chemical compositions of melts, crystal matrices, and crustal lithologies [ABSTRACT FROM AUTHOR]

The 2004–2009 caldera uplift is the largest instrumentally recorded episode of unrest at Yellowstone caldera. We use GPS and Interferometric Synthetic Aperture Radar (InSAR) time series spanning 2004–2015, with a focus in the aforementioned event to understand the mechanisms of unrest. InSAR data recorded ∼25 and ∼20 cm of uplift at the Sour Creek (SCD) and Mallard Lake (MLD) resurgent domes during 2004–2009, and ∼8 cm of subsidence at the Norris Geyser Basin (NGB) during 2004–2008. The SCD/MLD uplift was followed by subsidence across the caldera floor with a maximum at MLD of ∼1.5–2.5 cm/yr, and no deformation at NGB. The best‐fit source models for the 2004–2009 period are two horizontal sills at depths of ∼8.7 and 10.6 km for the caldera source and NGB, respectively, with volume changes of 0.354 and −0.121 km3, and an overpressure of ∼0.1 MPa. The InSAR and GPS time series record exponentially increasing followed by exponentially decreasing uplift between 2004 and 2009, which is indicative of magma injection into the caldera reservoir, with no need for other mechanisms of unrest. However, magma extraction from NGB to the caldera is unable to explain the subsidence coeval with the caldera uplift. Models of magma injection can also explain other episodes of caldera uplift like that in 2014–2015. Distributed sill opening models show that magma is stored across the caldera source with no clear boundary between MLD and SCD. Since the magma overpressure is orders of magnitude below the tensile strength of the encasing rock, historical episodes of unrest like these are very unlikely to trigger an eruption. Plain Language Summary: Resurgent calderas such as Yellowstone, Long Valley and Campi Flegrei show ground uplift that lasts years to decades. Despite decades of observations, in general the mechanisms of unrest at these type of volcanoes are not well understood. Using a combination of geodetic data from Interferometric Synthetic Aperture Radar (InSAR) and Global Position System (GPS), kinematic source models based on linear elasticity, and fluid mechanic models of magma flow, we constrain the mechanism of ground uplift at Yellowstone caldera between 2004 and 2009 — the best studied episode of uplift at this volcano. The caldera uplift can be entirely explained by the injection of basalt into a pressurized sill that lies ∼8.7 km below the caldera. However, subsidence in the neighboring Norris Geyser Basin cannot be explained by magma flow from this region towards the caldera. Despite the good fit of the models to the deformation data, more realistic mechanisms of fluid flow such a pressurization from exsolved volatiles should be explored at Yellowstone. Key Points: We reanalyze all the Interferometric Synthetic Aperture Radar (InSAR) and GPS data that span the 2004–2009 episode of unrestThe GPS and InSAR time series record uplift with an exponential increase followed by an exponential decrease indicating magma injectionMagma migration cannot explain subsidence at the Norris Geyser basin [ABSTRACT FROM AUTHOR]

Improvements in geochemical, petrologic, and geophysical methods and data have led to a paradigm‐shift in understanding magmatic system architecture away from isolated magmatic chambers toward transcrustal "mushy" magmatic systems. Geochemical and petrologic studies indicate that magma storage and differentiation occurs over a range of depths throughout the crust, while geophysical techniques delineate in situ spatial distributions of melt at similar scales. However, seismic properties are nonunique with respect to composition and melt percentage, and must be integrated with geochemical observations to better understand the architecture and chemical evolution of magmatic systems. The southern Puna Plateau represents a locality where extensive geochemical and seismic data can be combined to investigate the role of depth‐dependent processes in generating compositionally diverse volcanism. By combining geochemical modeling with seismic images of shear‐wave velocity and attenuation, we find that magma storage and differentiation occurs predominantly at two depth ranges: (a) near the base of the crust (>40 km), where shear‐wave velocities are anomalously slow compared to what is expected (<4.0 km/s) and geochemical signatures indicate fractional crystallization and crustal assimilation in the garnet stability field, and (b) in the mid‐crust (∼20 km) beneath the Cerro Galan Caldera, where shear‐wave velocities necessitate the presence of melt (∼2.7 km/s) and europium (Eu) anomalies indicate low‐pressure differentiation. These geochemical and seismic results provide evidence for multi‐level melt evolution and are consistent with petrological features of exposed arcs, whose transcrustal architecture is characterized by discrete and compositionally evolved intrusive bodies in the mid‐crust and a compositionally heterogeneous lower crust. Plain Language Summary: The inferred architecture of the magmatic plumbing beneath volcanoes can vary widely depending on the discipline and/or methodology used to investigate them. Recent studies have worked to integrate these different perspectives into a holistic model, but this has proven difficult as some data sets provide very different information than others. In this study, we outline the apparently contrasting viewpoints of magma system architecture based on the geophysical and geochemical characteristics of active magmatic systems, and field studies of exposed and extinct magmatic systems. We then investigate a region in the central Andes of South America with good seismic data coverage and compositionally diverse volcanics that has erupted over a relatively small area in the last ∼8 Ma. These data sets reveal evidence for melt stagnation and evolution at two main locations: near the crust‐mantle transition and in mid‐crustal magma reservoirs. Contextualizing these results with the apparently continuous magmatic structure seen in exposed sections indicates that these contrasting perspectives can be reconciled by understanding the sensitivities and limitations of the individual data sets. Thus, the evidence that has led to the discrete conceptualizations of magma systems often inferred from seismic data or geochemistry are consistent with more‐or‐less continuous magmatic systems throughout the crust. Key Points: We integrate geophysical and geochemical results to understand the architecture of a continental arc systemMagmas stagnate and differentiate predominantly at two depths: at or near the crust‐mantle transition (melting, assimilation, storage, and homogenization zone) and in the mid‐crustThe discrete nature of magmatic systems from geochemistry and seismology are reconcilable with the transcrustal petrology of exposed arcs [ABSTRACT FROM AUTHOR]

The Tengchong volcano (TCV) is a large active volcano system in the southeastern Tibetan Plateau. It is characterized by large‐volume magmatic gas emission, active hydrothermal circulation, and intense volcanic and earthquake activities, posing a threat of near future eruptions. However, there is still no available model of the magmatic plumbing system beneath the volcano system, limiting the quantitative assessments of the eruption hazards. Here, we present a high‐resolution 3D model of the TCV constructed using ambient noise adjoint tomography. Our 3D model reveals a large basaltic magma reservoir with a volume of ∼7,000 km3 at depths of 20–35 km, which has a melt fraction of ∼2%–4.5%. Our results suggest that the deep crustal magma reservoir is fed by partial melting in the uppermost mantle and is recharging the shallow magma chambers beneath the TCV. Our results are key to understanding the volcanic activities and assessing future eruption hazards. Plain Language Summary: The Tengchong volcano (TCV) is one of the largest active intraplate volcanoes in southwest China. The unrests of the TCV in recent years have attracted a lot of attention and aroused concerns for the future explosive eruptions. In this study, we present a high‐resolution 3D model of the TCV using an innovative method of ambient noise adjoint tomography. Our results reveal a multi‐layered magmatic plumbing system in the lower crust and uppermost mantle beneath the TCV. The large basaltic magma reservoir in the lower crust has a volume of ∼7,000 km3 at depths of 20–35 km and melt fraction of ∼2%–4.5%. Thus, the total of melt volume reaches 140–315 km3. The deep crustal magma reservoir is fed by partial melting in the uppermost mantle and is recharging the shallow magma chambers. If such a large volume of melt gets erupted in the future, the lava could cover a huge area, threatening the livelihood of the half million people living nearby. It, therefore, becomes urgent to quantitatively assess and model the potential of future eruptions. Key Points: We present a high‐resolution 3‐D velocity model of the Tengchong volcano (TCV) using ambient noise adjoint tomographyWe reveal a multi‐level magma system in TCV, that is, a low velocity zone in the uppermost mantle feeds the crustal magma reservoirThe lower crustal basaltic magma reservoir has a volume of ∼7,000 km3 at depths of 20–35 km with melt fraction of ∼2%–4.5% [ABSTRACT FROM AUTHOR]

The "wet" silicate solidus of mantle peridotite defines the initial melting temperature of Earth's mantle under water‐saturated conditions and the second critical endpoint (SCEP) marks the high P‐T end of the wet solidus. However, the location of the wet solidus has remained an outstanding issue for over 50 years and the position of the SCEP is hotly debated. Published wet solidi show a difference of 200–600°C at a given pressure while reported SCEPs range from <4 to >6 GPa. Using a large‐volume multianvil apparatus, we investigated the water‐saturated melting behavior of a fertile peridotite at 3–6 GPa, 950–1200°C, and obtained well‐preserved quenched materials. On the basis of textures and compositions of the quenched materials, we bracket the wet solidus to between 950°C and 1000°C at 3 GPa and the SCEP between 3 and 4 GPa. Combining our experimental results with seismologic and petrologic observations, we propose that the lithosphere‐asthenosphere boundary in subduction zones should be constrained by the wet solidus and emphasize the role of a deep hydrous partial‐melting zone (DHPMZ) on magma genesis within the mantle wedge. We suggest that the DHPMZ is a source of hydrous melts to the primary melting zone in the mantle wedge and that the position of the volcanic front and its magma production rate may largely be controlled by melting and melt segregation processes within the DHPMZ. Our experimental results also suggest that high‐magnesian magmas (e.g., boninite, picrite, and komatiite) could be formed at conditions representative of subduction zones. Key Points: We investigated H2O‐saturated melting behavior of peridotite with a large‐volume multianvil apparatus and obtained well‐preserved quenched materialsWe determined the water‐saturated solidus and SCEP of peridotite from the textures and compositions of the quenched materialsThe experimental results provide new constraints on the melting processes and magma genesis within the mantle wedge [ABSTRACT FROM AUTHOR]

Surface deformation accompanying dike intrusions is dominated by uplift and horizontal motion directly related to the intrusions. In some cases, it includes subsidence due to associated magma reservoir deflation. When reservoir deflation is large enough, it can form, or reactivate preexisting, caldera ring‐faults. Ring‐fault reactivation, however, is rarely observed during moderate‐sized eruptions. On February 21, 2015 at Ambrym volcano in Vanuatu, a basaltic dike intrusion produced more than 1 m of coeruptive uplift, as measured by InSAR, synthetic aperture radar correlation, and Multiple Aperture Interferometry. Here, we show that an average of ∼40 cm of slip occurred on a normal caldera ring‐fault during this moderate‐sized (VEI < 3) event, which intruded a volume of ∼24 × 106 m3 and erupted ∼9.3 × 106 m3 of lava (DRE). Using the 3D Mixed Boundary Element Method, we explore the stress change imposed by the opening dike and the depressurizing reservoir on a passive, frictionless fault. Normal fault slip is promoted when stress is transferred from a depressurizing reservoir beneath one of Ambrym's main craters. After estimating magma compressibility, we provide an upper bound on the critical fraction (f = 7%) of magma extracted from the reservoir to trigger fault slip. We infer that broad basaltic calderas may form in part by hundreds of subsidence episodes no greater than a few meters, as a result of magma extraction from the reservoir during moderate‐sized dike intrusions. Plain Language Summary: Many volcanoes feature large depressions, called calderas. Calderas form when enough magma leaves a deep reservoir, and the solid rock lid above this reservoir can no longer support its own weight. Caldera faults, or cracks surrounding the reservoir which extend from the reservoir to Earth's surface, form as the lid collapses. Ambrym volcano (Vanuatu) has a 12‐km wide caldera, and researchers propose it formed during an explosive eruption 2,000 years ago. However, in 2018, Ambrym's caldera sunk along caldera faults during a nonexplosive eruption. This observation questions whether an explosive eruption was necessary to form Ambrym's caldera in the first place. Furthermore, in February 2015, an eruption 10 times smaller than in 2018 also caused the ground to sink along caldera faults. Utilizing ground motion data obtained from satellite radar systems to model magma reservoir outflow and fault displacement, we conclude that, in 2015, the ground sank along caldera faults. This sinking is explained by the removal of as little as <7% of the stored magma from the reservoir. We therefore propose that Ambrym's wide caldera may have formed as a result of many frequently occurring medium‐sized eruptions. This challenges the thought that wide calderas mainly form as a result of large eruptions. Key Points: Ground displacement at Ambrym in February 2015 was caused by a dike intrusion, deflating reservoir, and normal slip on a caldera ring‐faultExtracting at most 7% of the magma from Ambrym's reservoir suffices to reactivate the caldera ring‐faultsNormal slip along Ambrym's ring‐fault can occur during moderate‐sized eruptions, resulting in subsidence and further caldera development [ABSTRACT FROM AUTHOR]

Field observations and unmanned aerial vehicle surveys from Caldera Taburiente (La Palma, Canary Islands, Spain) show that pre‐existing dykes can capture and re‐direct younger ones to form multiple dyke composites. Chill margins suggest that the older dykes were solidified and cooled when this occurred. In one multiple dyke example, an 40Ar/39Ar age difference of 200 kyr was determined between co‐located dykes. Petrography and geomechanical measurements (ultrasonic pulse and Brazilian disc tests) show that a microscopic preferred alignment of plagioclase laths and sheet‐like structures formed by non‐randomly distributed vesicles give the solidified dykes anisotropic elastic moduli and fracture toughness. We hypothesize that this anisotropy led to the development of margin‐parallel joints within the dykes, during subsequent volcanic loading. Finite element models also suggest that the elastic contrast between solidified dykes and their host rock elevated and re‐oriented the stresses that governed subsequent dyke propagation. Thus, the margin‐parallel joints, combined with local concentration and rotation of stresses, favored the deflection of subsequent magma‐filled fractures by up to 60° to form the multiple dykes. At the edifice scale, the capture and deflection of active intrusions by older ones could change the organization of volcanic magma plumbing systems and cause unexpected propagation paths relative to the regional stress. We suggest that reactivation of older dykes by this mechanism gives the volcanic edifice a structural memory of past stress states, potentially encouraging the re‐use of older vents and deflecting intrusions along volcanic rift zones or toward shallow magma reservoirs. Key Points: Dykes form significant and highly oblique mechanical discontinuities in volcanic edificesLocal stresses and mechanical anisotropy within solidified dykes can capture and deflect subsequent intrusions to form a multiple‐dykeReactivation of magma pathways by these mechanisms influences the emergent organization of magma plumbing systems [ABSTRACT FROM AUTHOR]

Degassing processes occurring within silicic magma, such as bubble growth, bubble resorption, the welding of magma fragments, and open‐system gas loss are crucial in the control of volcanic eruption and lava emplacement, yet their details are still debated. To examine the possibility that these degassing processes are recorded in volcanic glass as heterogeneous Cl distribution patterns, we experimentally simulated these processes by heating rhyolitic obsidian and analyzed the distribution of Cl content in the recovered sample. The results showed that, for bubble growth, Cl diffused toward the bubble interface, leading to Cl depletion around the bubble. For bubble resorption, Cl was discharged from the bubble to the melt, leading to Cl enrichment in the ambient melt. For welding of magma fragments, Cl was depleted near the welded interface because each fragment had degassed Cl at the surface before the welding took place. For open‐system gas loss, Cl exsolved at the bubble interface while the bubble itself was being collapsed into a chain of small bubbles and a Cl‐depleted tail. These results indicate that Cl distribution is a reliable record of the experienced degassing processes. We then analyzed the Cl distribution in silicic lava from Naruko volcano, Japan, to study the gas flow mechanism. We observed that the glassy matrix was progressively corroded into porous crystalline material. The interface of the glass was highly enriched in Cl. We conclude that a Cl‐rich gas fluxed through hot lava and corroded the glass, developing a porous, gas‐permeable region. Key Points: We experimentally simulated degassing processes occurring in silicic magma and analyzed Cl content distribution in glassThe Cl‐content distribution was heterogeneous and the distribution pattern was specific to the degassing process experiencedWe analyzed Cl‐content in natural silicic lava and revealed that a Cl‐rich corrosive gas fluxed and developed permeable flow systems [ABSTRACT FROM AUTHOR]

At oceanic spreading centers, the interactions between the igneous system that builds the crust, and the hydrothermal system that cools it govern the plumbing system architecture and its thermokinetic evolution. At fast‐spreading centers, most of those interactions occur around the axial magma lens (AML) that feeds the upper crust, and possibly part of the underlying mushy igneous reservoir. Heat extracted from crystallizing AML is transferred through a conductive boundary layer to the overlying hydrothermal system. Quantifying the AML physical and thermal evolutions and its interactions with hydrothermal system is therefore essential to understand oceanic accretion. Those general issues were the rationale of drilling ICDP OmanDP Hole GT3A, and we present herein the geological, structural, and petrological data that were used as a site survey to select its location. GT3 area enables observations in three dimensions of fossilized AMLs and overlying dikes. The new field data and corresponding mineral compositions are used together with thermokinetic and thermodynamic models to deliver an integrated dynamic model for the AML/hydrothermal system interactions. Results attest that the isotropic gabbro interval is composite, with gabbro bodies intruding and reheating both gabbros and dikes (up to 1,040°C). We show that AMLs should be considered as transient igneous bodies that likely crystallize from primitive MORBs in decades, releasing heat to the intruded hosts, and feeding high temperature vents on the seafloor. We show for the first time that the thermal gradient recorded in AML roof is consistent with the heat fluxes reported at active hydrothermal vents. Key Points: At fast‐spreading ridges magma‐hydrothermal interactions are extensive at roof of axial magma lenses (AML) that are fed by primitive MORBAML are transient igneous bodies that provide heat to hydrothermal vents on the seafloor through conductive boundary layersAML crystallize to form the isotropic gabbros in decades, and released heat modifies the intruded dike and gabbro hosts [ABSTRACT FROM AUTHOR]

Delineation of the characteristics of magmatic fluid such as melt and water beneath volcanoes has been challenging. In this study, we obtained a detailed picture of the magma reservoir beneath the felsic caldera of Hakone volcano, central Japan using a velocity model obtained from a dense seismic observation network. In the shallow region of the caldera, a high‐velocity region associated with a solidified magma where many earthquake swarms occur was estimated, and a significantly low‐velocity region exists beneath the lower depth limit of seismicity, deeper than 6 km. The upper part of this low‐velocity region is interpreted as a water‐rich region established by the dehydration from the magma reservoir located beneath it. The temperature in this water‐rich zone was estimated to be above 350°C based on its location below the seismogenic depth. The upper depth of the magma reservoir was determined to be 9 km deep, where the high‐ratio of P‐ to S‐wave velocity was estimated. The volume fraction of melt was estimated as less than 10%, suggesting a highly crystallized mush structure. The low‐velocity zone was estimated beneath the shallow magma reservoir at a depth of 15 km and deep extension of the source region of deep low‐frequency earthquakes. The results could suggest the pathway of magma transport from the deep part of the volcano to this shallow reservoir. Key Points: A highly resolved velocity model estimated a picture of the magma reservoir beneath felsic volcanoThe upper part of the magma reservoir is at a 9‐km depth with a highly crystallized mush structureWater‐rich zone dehydrated from the magma reservoir was developed at upper part of the reservoir [ABSTRACT FROM AUTHOR]

The thermal structure of mid‐ocean ridge axes is a critical control on the mechanical properties of young lithosphere and modulates the faulting styles that shape the Earth's seafloor. Models that balance a steady input of magmatic heat with a vigorous hydrothermal output are generally successful at explaining the thermal state of magmatically robust mid‐ocean ridges. However, the magma supply of ultraslow spreading ridges is often subdued and highly episodic, and the depth‐extent and vigor of hydrothermal circulation in these settings remain poorly known, requiring modifications to the standard magmatic‐hydrothermal paradigm. We develop models that couple repeated magmatic intrusions with hydrothermal convection in a permeable domain whose boundaries are controlled by the extent of grain‐scale thermal cracking. Our simulations reveal decreasing trends of venting temperatures with increasing intrusion periodicity, as well as with increasing permeability. Our model explains the seismically inferred depth (∼10 km) to the brittle‐ductile transition and the depth of MORB crystallization at the Mid‐Cayman Spreading Center by invoking repeated intrusion of sills every 20–50 kyrs, depending on their size. Doubling this periodicity predicts a thicker (∼15 km) seismogenic layer and lower‐temperature venting, as observed at the Southwest Indian Ridge (SWIR) at 64˚E. However, if fluid convection is not possible below ∼6 km, our model requires that high‐temperature venting at ultraslow ridges (e.g., Longqi at SWIR 49.7˚E) be fueled by shallow episodic magma intrusions into the long‐term convective region, which maintain high‐output circulation for at most a few tens of kyrs. Key Points: New model for the thermal structure of ultraslow ridges couples hydrothermal circulation with episodic sill emplacementHydrothermal convection in uppermost 6 km of lithosphere with thick boundary layer can explain cold thermal regimeHigh‐output venting at ultraslow ridges likely requires transient shallow magmatic intrusions [ABSTRACT FROM AUTHOR]

The rhyolite‐producing Laguna del Maule volcanic field (LdMVF), Chile, has had numerous post‐glacial eruptions that produced large explosions and voluminous lava flows. During the Holocene ∼60 m of surface uplift is recorded by paleo‐shorelines of the fresh‐water Laguna del Maule, with an inflation source near the Barrancas volcanic complex. Rhyolites from the Barrancas complex erupted over ∼14 ka including some of the youngest (1.4 ± 0.6 ka) lava flows in the field. New gravity data collected on the Barrancas complex reveals a residual gravity low (−6 mGal, "Barrancas anomaly") that is distinct from the pronounced gravity low (−19 mGal; "Lake anomaly") associated with present‐day ground uplift to the northwest. Three‐dimensional inversion of the Barrancas anomaly indicates the presence of a magma body with a maximum density contrast with the host rock of −250 kg/m3 centered at a depth of ∼3 km below surface. Nearby Miocene high‐silica granites represent frozen remnants of highly evolved rhyolitic magma. Comparison of the densities measured from samples of these plutons with the geophysical model densities, and integration of thermodynamic modeling of silicic melt evolution, provide constraints on our interpretation. We propose a magma body, containing <30% melt phase and low volatile content, exists beneath Barrancas. The Barrancas and Lake gravity lows represent magma in different physical states, associated with past and present‐day storage beneath LdMVF. The gravity model mirrors geochemical observations which independently indicate that at least two distinct rhyolites were generated and stored as discrete magma bodies within the broader LdMVF. Plain Language Summary: The Laguna del Maule volcanic field (LdMVF) in Chile, has a long history of eruptions that produced large explosions and voluminous lava flows. Lava emitted during these eruptions comes from shallow sources of partially molten rock (magma) located a few kilometers below the surface. Magma pushing from below and into these shallow regions has caused the ground around the lake to rise over the last 10,000 years. To understand the causes of deformation and assess the likelihood of future eruptions, it is essential to know the location and present‐day state of magma within Earth's crust. Magma is typically less dense than the surrounding rock and these density variations cause small changes in the local gravity field, detectable with geophysical instruments. We collected a new set of gravity measurements at the Barrancas volcanic complex to determine if magma is present beneath the volcano that last erupted about 2000 years ago. Our data reveal a shallow magma body at depth beneath the Barrancas complex. This new magma body is separate from another magma body located in a different part of the LdMVF, which has important implications for the assessment of future volcanic hazards in the region. Key Points: At least two separate magma storage regions in different physical states exist beneath the Laguna del Maule volcanic fieldA −6 mGal gravity anomaly below the Barrancas complex is close to a postulated inflation source which produced ∼62 m of Holocene upliftComparing model densities to nearby plutons and seismic models indicates Barrancas magma is above solidus with a small melt proportion [ABSTRACT FROM AUTHOR]

Liquid immiscibility is a critical mechanism to diversify magma compositions. The physical separation of exsolved melt droplets is an essential process in generating new magmas. However, little attention has been paid to this physical process. In this study, we present a new model for segregation of immiscible melt droplets in which exsolution, settling, and coalescence are all considered. The separation of immiscible droplets is similar to that of crystals in magma when the discrete melt exsolves as large (millimeter‐size) droplets and/or when the magma cools slowly. However, when immiscible melt droplets are small (micrometer‐size) and/or magma cools rapidly, coalescence can significantly enhance their separation. The low interfacial tension between coexisting silicate melts leads to the exsolution of extremely small melt droplets. Furthermore, the high viscosity of silica‐rich melt suppresses the coalescence of droplets. Consequently, the separation of highly viscous silica‐rich melt droplets is slow but could have occurred in a slowly cooling magma such as the Skaergaard intrusion. By contrast, the coalescence rate of melt droplets with low viscosity is high. Iron‐rich melt droplets, which have low viscosities, could have been separated from hydrous andesitic melts to form magnetite‐apatite ore deposits at El Laco and Marcona. Furthermore, the viscosities of sulfide and carbonatitic melts are low. Therefore, immiscibility between sulfide and silicate melts may lead to the formation of magmatic sulfide deposits, and immiscibility between carbonatitic and silicate melts can support periodical eruptions of carbonatitic lava. Key Points: A new model for separation of immiscible melt droplets in magma is developedA dimensionless parameter is proposed to assess the contribution of droplet coalescence to immiscible melt separationThe separation efficiency of immiscible melt droplets in natural magma is evaluated [ABSTRACT FROM AUTHOR]

Active magma chambers are periodically replenished upon a combination of buoyancy and pressure forces driving upward motion of initially deep magma. Such periodic replenishments concur to determine the chemical evolution of shallow magmas, they are often associated to volcanic unrests, and they are nearly ubiquitously found to shortly precede a volcanic eruption. Here, we numerically simulate the dynamics of shallow magma chamber replenishment by systematically investigating the roles of buoyancy and pressure forces, from pure buoyancy to pure pressure conditions and across combinations of them. Our numerical results refer to volcanic systems that are not frequently erupting, for which magma at shallow level is isolated from the surface ("closed conduit" volcanoes). The results depict a variety of dynamic evolutions, with the pure buoyant end‐member associated with effective convection and mixing and generation of no or negative overpressure in the shallow chamber, and the pure pressure end‐member translating into effective shallow pressure increase without any dynamics of magma convection associated. Mixed conditions with variable extents of buoyancy and pressure forces illustrate dynamics initially dominated by overpressure, then, over the longer term, by buoyancy forces. Results globally suggest that many shallow magmatic systems may evolve during their lifetime under the control of buoyancy forces, likely triggered by shallow magma degassing. That naturally leads to long‐term stable dynamic conditions characterized by periodic replenishments of shallow partially degassed, heavier magma by volatile‐rich fresh deep magma, similar to those reconstructed from petrology of many shallow‐emplaced magmatic bodies. Plain Language Summary: Periodic injections of magma into shallow magma chambers are a fundamental process known to govern the evolution of magmatic systems. According to our overall understanding of magmatic systems, we simulate the time‐space dynamics in a composite magmatic system under the action of buoyancy and pressure forces. The formers are due to different chemical compositions of the shallow and deep magmas. The latter may result from chemical reactions, tectonic stress accumulation, or other deep processes. The numerical results illustrate an ample variety of situations mostly controlled by buoyancy or pressure forces. It is found that the "normal" condition for many volcanic systems is likely characterized by periodic arrivals of light batches of magma giving rise to efficient convection and mixing at shallow level, and parallel sinking of partially degassed, dense magma. This ensures essentially stable conditions unlikely to give origin to any pressure buildup leading to an eruption; it provides a mechanism for long life times of relatively small, shallow magmatic systems, and a framework for interpreting petrologic observations suggesting long sequences of mixing events at magmatic systems worldwide. Key Points: A numerical model for time‐space dynamics in shallow magma chambers replenished by deep magma is presentedEvolution under pressure and buoyancy forces is initially dominated by overpressure, and over the long term by buoyant convection and mixingLifetime evolution of volcanic systems is likely toward stable conditions with no pressure buildup leading to an eruption [ABSTRACT FROM AUTHOR]

Cyclical ground deformation, associated seismicity, and elevated degassing are important precursors to explosive eruptions at silicic volcanoes. Regular intervals for elevated activity (6–30 hr) have been observed at volcanoes such as Mount Pinatubo in the Philippines and Soufrière Hills in Montserrat. Here, we explore a hypothesis originally proposed by Michaut et al. (2013, https://doi.org/10.1038/ngeo1928) where porosity waves containing magmatic gas are responsible for the observed periodic behavior. We use two‐phase theory to construct a model where volatile‐rich, bubbly, viscous magma rises and decompresses. We conduct numerical experiments where magma gas waves with various frequencies are imposed at the base of the model volcanic conduit. We numerically verify the results of Michaut et al. (2013, https://doi.org/10.1038/ngeo1928) and then expand on the model by allowing magma viscosity to vary as a function of dissolved water and crystal content. Numerical experiments show that gas exsolution tends to damp the growth of porosity waves during decompression. The instability and resultant growth or decay of gas wave amplitude depends strongly on the gas density gradient and the ratio of the characteristic magma extraction rate to the characteristic magma degassing rate (Damköhler number, Da). We find that slow degassing can lead to a previously unrecognized filtering effect, where low‐frequency gas waves may grow in amplitude. These waves may set the periodicity of the eruptive precursors, such as those observed at Soufrière Hills Volcano. We demonstrate that degassed, crystal‐rich magma is susceptible to the growth of gas waves which may result in the periodic behavior. Key Points: A model for crystal‐rich magma with exsolution of water vapor is developed to explain long period eruptive precursors at silicic volcanoesSluggish kinetics for the exsolution of water favors the formation of porosity waves which may contribute to cyclical unrestDegassed crystal‐rich magma is most susceptible to porosity waves [ABSTRACT FROM AUTHOR]

In the continental crust, the probability of dike propagation out of magma chambers is governed by thermal, rheological, and pressure conditions of magma chamber‐wall rock systems. Incremental injection of melt into an average‐size, laccolith‐shaped, midcrustal magma chamber produces a volume of mobile magma at the bottom of the chamber that has the potential to escape as dikes through the upper, immobile portion of the chamber and the roof. Here we numerically model the conditions needed for dike propagation out of a magma chamber during continuous and episodic injections of melt into the chamber. The roles of magma buoyancy and overpressure from melt injections in generating dikes are explored within 1.78 × 104 to 1.78 × 108 Pa·s range of magma viscosities (μmag), 10 to 40 GPa range of elastic moduli (E) of the immobile top portion of the magma chamber, and 10 and 20 kyr durations of chamber growth. During episodic, high‐flux melt injections (tens of km3/yr), magma overpressure can reach >100 MPa and initiate dike propagation even when μmag and E are near the high ends of the examined ranges. The probability of generating dikes diminishes when the injection flux is lower. Continuous low‐flux injections favor magma accumulation because injection overpressure never exceeds 20 MPa. During either continuous or episodic growth of magma chamber, there is never a sufficient amount of mobile magma in the chamber for dikes to be induced by magma buoyancy alone. Key Points: Dike propagation through ductile upper portion of a growing midcrustal laccolith and its roof is modeledGiven likely ranges of magma viscosities and elastic moduli of roof, critical overpressure for dike propagation is 10 to 200 MPaChamber growth is favored by continuous injections, whereas overpressures by high‐flux injections exceed the highest critical overpressures [ABSTRACT FROM AUTHOR]

A key role of water has been emphasized in the generation of Early Cretaceous high‐Mg andesites (HMAs) and lithospheric thinning of the eastern North China Craton (NCC). However, the source of water and when it hydrated the NCC lithospheric mantle are still unclear. To address these issues, here we investigate the petrography, mineral chemistry, and the chemical and Sr‐Nd‐Pb isotopic compositions of two suites of Early Cretaceous HMAs from Western Liaoning, China, combined with new Sr‐Nd‐Pb isotopic data of the Early Cretaceous primitive basalts from the same area. Petrographic features and variations of chemical and isotopic compositions suggest a petrogenesis of mixing between mantle‐derived mafic and crust‐derived felsic melts for these HMAs and most of other Early Cretaceous HMAs in the NCC. The mantle‐derived primary melts are indicated to have been derived from the refractory lithospheric mantle beneath the NCC. Fluid‐rich melt inclusions in olivines, absence of plagioclase phenocryst, enrichments of fluid‐indicative trace elements (e.g., Rb, Ba, K, Pb, and Sr), high water content in clinopyroxenes, and low Ca in olivines indicate high water contents in the magmas and their mantle sources. The EM1‐like Sr‐Nd and unradiogenic Pb isotopic compositions of these lavas unravel a Mantle Transition Zone source for the fluids that hydrated the NCC lithospheric mantle. The wet upwelling of the Mantle Transition Zone components would be triggered by the penetration of deep‐subducted Paleo‐Pacific slab. Our study provides key evidence for the transport of water from deep Earth into the cratonic lithosphere and its role in lithospheric thinning of cratons. Key Points: A hydrated and refractory lithospheric mantle source is illustrated for the Early Cretaceous high‐Mg andesites in the North China CratonThese lavas reveal wet upwelling of the Mantle Transition Zone components triggered by the penetration of deep‐subducted Paleo‐Pacific slabKey evidence is provided for the role of Earth's deep volatile cycling in lithospheric thinning of cratons [ABSTRACT FROM AUTHOR]

Picrites are potentially near‐primary melts that offer rare insights into the earliest stages of magmatic evolution. However, without detailed investigation of magmatic processes, robust records of early melt evolution and mantle melting behavior cannot be acquired. Here petrological and geochemical interrogations of the Dali picrites were conducted to investigate the magmatic processes. Compositional variation observed in the Dali picrites is dominated by variable accumulation of a wehrlitic assemblage, transported by a low MgO carrier melt. Groundmass clinopyroxenes were produced by direct crystallization of the carrier melt. In contrast, compositions of olivine‐hosted melt inclusions and clinopyroxene macrocrysts indicate that olivine and clinopyroxene macrocrysts crystallized from diverse melt batches that are genetically related to the carrier melt, which formed as a consequence of magma mixing. The compositional range of melts in equilibrium with clinopyroxene macrocrysts with Mg# < 88 (> 88) is similar to (notably smaller than) the range of melt inclusions hosted in olivine macrocrysts with equivalent forsterite values. This observation can be explained by coupling between major and trace elements during partial melting in the mantle. Overall, a process involving periodic magma replenishment, tapping, and fractional crystallization in deep magma chamber(s) can explain the compositional variations recorded in the Dali macrocryst assemblage. Close genetic relationships between macrocrysts and carrier melts appear to be common to many primitive mafic rocks, which may reflect common deep magma chamber processes whereby diverse mantle‐derived melts are injected and evolved, and the primitive macrocrysts thus formed are subsequently transported by mixed liquids, producing macrocryst‐bearing primitive mafic rocks. Key Points: Compositional variation in the Dali picrites is dominated by variable accumulation of olivine and clinopyroxene macrocrystsThe macrocrysts crystallized from multiple parental magmas and are genetically related to the carrier meltPeriodic RTX processes in the magma chamber(s) dominate the compositional evolution of the diverse parental melts [ABSTRACT FROM AUTHOR]