Wendy A. Bohrson

Chapter 13 Petrogenesis of the Campanian Ignimbrite: implications for crystal-melt separation and open-system processes from major and trace elements and Th isotopic data

Abstract The Campanian Ignimbrite is a large-volume trachytic to phonolitic ignimbrite that was deposited at ≈39.3 ka and represents one of a number of highly explosive volcanic events that have occurred in the region near Naples, Italy. Thermodynamic modeling using the MELTS algorithm reveals that major element variations are dominated by crystal-liquid separation at 0.15 GPa. Initial dissolved H 2 O content in the parental melt is ∼3 wt.% and the magmatic system fugacity of oxygen was buffered along QFM+1. Significantly, MELTS results also indicate that the liquid line of descent is marked by a large change in the proportion of melt (from 0.46 to 0.09) at ∼884°C, which leads to a discontinuity in melt composition (i.e., a compositional gap) and different thermodynamic and transport properties of melt and magma across the gap. Crystallization of alkali feldspar and plagioclase dominates the phase assemblage at this pseudo-invariant point temperature of ∼884°C. Evaluation of the variations in the trace elements Zr, Nb, Th, U, Rb, Sm, and Sr using a mass balance equation that accounts for changing bulk mineral-melt partition coefficients as crystallization occurs indicates that crystal-liquid separation and open-system processes were important. Th isotope data yield an apparent isochron that is ∼20 kyr younger than the age of the deposit, and age-corrected Th isotope data indicate that the magma body was an open system at the time of eruption. Because open-system behavior can profoundly change isotopic and elemental characteristics of a magma body, these Th results illustrate that it is critical to understand the contribution that open-system processes make to magmatic systems prior to assigning relevance to age or timescale information derived from such systems. Fluid-magma interaction has been proposed as a mechanism to change isotopic and elemental characteristics of magma bodies, but an evaluation of the mass and thermal constraints on such a process suggests large-scale interaction is unlikely. In the case of the magma body associated with the Campanian Ignimbrite, the most likely source of the open-system signatures is assimilation of partial melts of compositionally heterogeneous basement composed of cumulates and intrusive equivalents of volcanic activity that has characterized the Campanian region for over 300 kyr.

Partitioning of trace elements among coexisting crystals, melt, and supercritical fluid during isobaric crystallization and melting

Abstract The distribution of trace elements among coexisting crystals, melt, and supercritical fluid during melting and crystallization is a critical constraint for understanding the evolution of magmatic systems, including the origin and development of continental and oceanic crust. Although trace-element partitioning between crystals and melt during Rayleigh fractional crystallization or melting is well-known, partitioning among co-existing melt, crystals, and supercritical fluid during anatexis or crystallization is less explored despite the ubiquity of magmatic fluids. Here we develop the trace-element differential equations governing solid-melt-fluid equilibria for melting and crystallization under fluid-present conditions and provide analytical solutions for fractional and equilibrium crystallization and melting. A compilation of solid-fluid and melt-fluid distribution coefficients for about 30 trace elements in olivine, clinopyroxene, garnet, plagioclase, alkali feldspar, biotite, amphibole, apatite, and silicic melts is provided. Forward modeling demonstrates the conditions under which fluid-meltsolid partitioning will impact trace-element signatures in magmatic systems. We show that for trace elements soluble in aqueous fluids, the composition of a melt derived by fluid-present fractional crystallization or by fluid-present fractional melting will be significantly different than in otherwise comparable fluid-absent systems. Ignoring the partitioning of soluble elements into the fluid phase leads to large errors in concentrations (over 100%) and ratios and consequent misinterpretation of the trace-element character of source material and/or the processes of fractional crystallization and melting. Although significant in any setting involving fluid-present equilibria, this analysis may have a most profound influence on fluid-present subduction zone magma generation and the evolution of shallow level fluid-saturated silicic magmatic systems.

Paleomagnetic behavior of volcanic rocks from Isla Socorro, Mexico

The direction and magnitude of the geomagnetic field vary both spatially and temporally and undergo significant departures from that of a geocentric axial dipole. In order to properly characterize persistent behaviors, time-averaged field models must be based on the highest quality data. Here we present full-vector paleomagnetic data for volcanic units exposed in the southeast quadrant of the island of Socorro, Mexico. We carried out a joint expedition between the Scripps Institution of Oceanography and the Universidad Nacional Autónoma México to Isla Socorro in January of 2005 during which we collected oriented paleomagnetic samples from 21 sites, representing as many as 10 different volcanic units (the oldest of which is ~540 ka). We subjected over 100 specimens to the most up-to-date paleointensity methods, and included the standard reliability checks. In an earlier study, Bohrson et al. (1996) proposed a series of widespread eruptive events, based on similarities of argon/argon dates. Paleointensity from specimens that conform to the strictest acceptance criteria are available from both the (unoriented) original sample collection and our fully oriented (but as yet undated) new collection. Correlation between the two collections is however problematic. The time-averaged direction from Socorro is consistent with that expected from a geocentric axial dipole, and the time-averaged intensity is 30.0±7.1 μT, equivalent to a virtual axial dipole moment (VADM) of 67.6±16.0 ZAm2.

Dynamics and thermodynamics of magma mixing: Insights from a simple exploratory model

Abstract The mixing of magmas of distinct temperature, bulk composition, mineralogy, and physical properties plays a central role in explaining the diversity of magma types on Earth and in explaining the growth of continental and oceanic crust. Magma mixing is also of practical importance. For example, the mixing of distinct magmas has been cited as an important process in creation of economically important horizons in layered intrusions as well as a triggering mechanism for initiation of volcanic eruptions. The motivation for better quantifying the dynamics and thermodynamics of magma mixing and its attendant plutonic and volcanic products is clear. The degree of magma mixing, which spans a continuum from mingling to complete hybridization, depends upon initial and boundary conditions, magma properties, driving forces, and time available for mixing. Magma mingling produces a heterogeneous mixture of discrete clumps of the end-member magmas, whereas complete hybridization involves the thermodynamic equilibration of two distinct magmas to form a third. Qualitatively, mixing occurs via reduction in the size of compositional heterogeneities (i.e., clumps) through stretching and folding by viscous flow, followed by homogenization, once shear has reduced the size of compositional anomalies to diffusive length scales. Quantification of this process relies on two statistical measures: the linear scale of segregation (Λ) defined as the spatial integral of the compositional correlation function related to the size-distribution of the segregated clumps within the mixture, and the intensity of segregation (I) a measure that quantifies how much the composition at each location differs from the average. The mixing dynamics of a layered system are analyzed in terms of the parameters governing mixing (Rayleigh, Lewis, and buoyancy numbers and viscosity ratio) to estimate how the timescale for magma hybridization, τH, compares to solidification, recharge, diffusive, and assimilation timescales. This analysis illustrates that hybridization times can be shorter than or comparable to thermal, solidification, and replenishment timescales; thus, formation of hybridized or nearly hybridized magmas is one anticipated outcome of mixing. The machinery of thermodynamics can be used to compute the hybrid magma state. An exploratory model for the thermochemistry of hybridization is developed based on binary eutectic phase relations and thermodynamics. Eight thermodynamic parameters define the phase diagram and associated energetics, and six parameters (initial temperatures, compositions, mass ratio of mixing magmas, and an enthalpy parameter) are necessary and sufficient to determine the state of hybrid magma uniquely. While relevant combinations of 14 thermodynamic and mixing parameters might suggest that the number of mixing outcomes (i.e., products) is too high to systematize, Monte Carlo simulations using the exploratory model document how millions of arbitrary initial states evolve into five possible final (mixed) states. Such an analysis implies that a magma mixing taxonomy that defines possible mixed product states can be developed and tied to petrologic indicators of mixing. Additional insights gained from this exploratory model that are supported by independent results from a multicomponent, multiphase thermodynamic model of magma mixing (Magma Chamber Simulator) include: (1) the proclivity of invariant point hybrid states, which may explain some instances of compositionally monotonous melts associated with mixed magma eruptions; (2) a surprising thermal effect such that the temperature of hybridized magma can be significantly less than the initial temperature of either of the mixing magmas. This type of magma mixing may result in crystal resorption, thus invalidating an assumption that resorption textures in crystals are typically the result of a magma heating event; (3) illustration of the differing effects of stoped block temperature and composition on hybrid magma temperature and phase state; and (4) illustration of a cessation of crystallization effect that may pertain to the MORB pyroxene “paradox.” Differences between adiabatic or R-hybridization and diabatic or RFC-hybridization are also explored. The model can be used to elucidate the thermodynamic principles underlying magma mixing in the hybridization limit. These principles are of general applicability and carry over to more compositionally complicated systems.

Bulk rock composition and geochemistry of olivine-hosted melt inclusions in the Grey Porri Tuff and selected lavas of the Monte dei Porri volcano, Salina, Aeolian Islands, southern Italy

The Aeolian Islands are an arcuate chain of submarine seamounts and volcanic islands, lying just north of Sicily in southern Italy. The second largest of the islands, Salina, exhibits a wide range of compositional variation in its erupted products, from basaltic lavas to rhyolitic pumice. The Monte dei Porri eruptions occurred between 60 ka and 30 ka, following a period of approximately 60,000 years of repose. The bulk rock composition of the Monte dei Porri products range from basaltic-andesite scoria to andesitic pumice in the Grey Porri Tuff (GPT), with the Monte dei Porri lavas having basaltic-andesite compositions. The typical mineral assemblage of the GPT is calcic plagioclase, clinopyroxene (augite), olivine (Fo72−84) and orthopyroxene (enstatite) ± amphibole and Ti-Fe oxides. The lava units show a similar mineral assemblage, but contain lower Fo olivines (Fo57−78). The lava units also contain numerous glomerocrysts, including an unusual variety that contains quartz, K-feldspar and mica. Melt inclusions (MI) are ubiquitous in all mineral phases from all units of the Monte dei Porri eruptions; however, only data from olivine-hosted MI in the GPT are reported here. Compositions of MI in the GPT are typically basaltic (average SiO2 of 49.8 wt %) in the pumices and basaltic-andesite (average SiO2 of 55.6 wt %) in the scoriae and show a bimodal distribution in most compositional discrimination plots. The compositions of most of the MI in the scoriae overlap with bulk rock compositions of the lavas. Petrological and geochemical evidence suggest that mixing of one or more magmas and/or crustal assimilation played a role in the evolution of the Monte dei Porri magmatic system, especially the GPT. Analyses of the more evolved mineral phases are required to better constrain the evolution of the magma.

Rejuvenation of crustal magma mush: A tale of multiply nested processes and timescales

Some relatively crystal rich silicic volcanic deposits, including large volume ignimbrites, preserve evidence of a history where a rheologically locked high crystallinity magma was rejuvenated (unlocked) through enthalpy (± mass) exchange with newly injected recharge magma of higher specific enthalpy. This is one event in a possible complex history. That is, the volcanic product of an eruption reflects an array of sequential and nested processes including melt formation and segregation, ascent, cooling, crystallization, crustal assimilation, magma recharge, unlocking, shallow ascent, fluid exsolution, and eruption. Deconvolution of these these nested processes and concomitant timescales is complicated and relies on a multidisciplinary approach; studies that do not clearly associate process and correlated timescale have the potential to provide misleading timescale information. We report the results of thermodynamic and heat transfer calculations that document mass, energy, and phase equilibria constraints for the unlocking of near-solidus rhyolite mush via magma mingling (heat exchange only) with basaltic recharge magma of higher specific enthalpy. To achieve unlocking, defined as the transition from near-solidus to ∼50 percent melt of the host silicic magma, phase equilibria computations provide (1) the enthalpy required to unlock mush, (2) the mass ratio of recharge magma to mush (MR/MM) when the two magmas achieve thermal equilibrium, and (3) the changes in melt, mineral, and fluid phase masses, compositions, and temperatures during the approach to unlocking. The behavior of trace elements is computed with knowledge of mineral, fluid, and melt proportions and solid-fluid and solid-melt partition coefficients. Evaluation of unlocking for relatively ‘dry’ (0.5 wt. % H2O) and ‘wet’ (3.9 wt. % H2O) rhyolitic mushy (locked) magma by basaltic recharge at upper crustal pressures indicates minimum values of MR/MM can be significantly less than 1, assuming the mingling process is isenthalpic with no ‘waste’ heat. For active volcanic systems estimates of MR may be tested using geodetic data. Wet mush has lower energy requirements for unlocking and thus requires lower MR/MM than dry mush. Wet rejuvenated magmas therefore may be more abundant in the volcanic rock record, and unlocked dry mushes may be restricted to extensional tectonic settings with high recharge flux. Temperature changes in dry mush as it unlocks are pronounced (greater than 150 °C) compared to those in wet mush, which are smaller than the resolution of classical geothermometry (∼15 °C). Phase equilibria calculations show that, as required, the net volume of crystals decreases during unlocking. Interestingly, calculations also indicate reactive crystal growth by chemical re-equilibration at the crystal-size scale during unlocking may also take place. In either dissolution by unidirectional resorption or reactive dissolution/ new growth, the chemical signatures of unlocking, potentially preserved in crystals or parts of crystals (for example, rims), are predictable and hence testable. Independent of unlocking thermodynamics, the phase equilibria and elemental consequences of isentropic magma ascent, a transport event that follows unlocking, can also be predicted; detailed examination of several canonical cases reveals a marked contrast with isenthalpic unlocking, thereby providing a means of process deconvolution. Unlocking timescales are estimated by two methods, one that calculates the time to reach thermal equilibrium for recharge magma dispersed in mush as ‘clumps’ of fixed size, and the second where the required volume of recharge magma is initially a single clump and evolves to smaller size through clump stretching and folding. For a range of magma volumes from 0.1 to 5000 km3, unlocking times range from 10−2 to 106 years. The shorter timescales for any magma volume requires a large number of relatively small clumps (n >106), which implies that large volumes of mush purported to unlock over short timescales (102–103 years and less) should preserve and exhibit evidence of intimate magma mingling. The key result of our analysis is that multiple timescales are operative during the potentially long and complex history of silicic mush formation, rejuvenation, and ascent. To correctly ascribe timescales to unlocking requires a holistic understanding of the myriad processes that affect the magma before, during, and after enthalpy exchange. In the absence of this context, unlocking timescales may be incorrectly constrained which, in turn, may hinder eruption forecasting and associated hazard mitigation.

Origin of primitive ocean island basalts by crustal gabbro assimilation and multiple recharge of plume-derived melts

Chemical Geodynamics relies on a paradigm that the isotopic composition of ocean island basalt (OIB) represents equilibrium with its primary mantle sources. However, the discovery of huge isotopic heterogeneity within olivine-hosted melt inclusions in primitive basalts from Kerguelen, Iceland, Hawaii and South Pacific Polynesia islands implies open-system behavior of OIBs, where during magma residence and transport, basaltic melts are contaminated by surrounding lithosphere. To constrain the processes of crustal assimilation by OIBs, we employed the Magma Chamber Simulator (MCS), an energy-constrained thermodynamic model of recharge, assimilation and fractional crystallization. For a case study of the 21 – 19 Ma basaltic series, the most primitive series ever found among the Kerguelen OIBs, we performed sixty-seven simulations in the pressure range from 0.2 to 1.0 GPa using compositions of olivine-hosted melt inclusions as parental magmas, and metagabbro xenoliths from the Kerguelen Archipelago as wallrock. MCS modeling requires that the assimilant is anatectic crustal melts (P2O5 ≤ 0.4 wt.% contents) derived from the Kerguelen oceanic metagabbro wallrock. To best fit the phenocryst assemblage observed in the investigated basaltic series, recharge of relatively large masses of hydrous primitive basaltic melts (H2O = 2 – 3 wt%; MgO = 7 – 10 wt.%) into a middle crustal chamber at 0.2 to 0.3 GPa is required. Our results thus highlight the important impact that crustal gabbro assimilation and mantle recharge can have on the geochemistry of mantle-derived olivine-phyric OIBs. The importance of crustal assimilation affecting primitive plume-derived basaltic melts underscores that isotopic and chemical equilibrium between ocean island basalts and associated deep plume mantle source(s) may be the exception rather than the rule.