Warning: Error while sending QUERY packet. PID=38660 in /Library/WebServer/Documents/wp-includes/wp-db.php on line 1924
What Does a Magma Reservoir Look Like? The "Crystal's Eye" View - Elements
Elements Covers

What Does a Magma Reservoir Look Like? The “Crystal’s Eye” View

Crystals within volcanic rocks contain records of the changing chemical and thermal conditions within the magma reservoirs in which they resided before eruption. Observations from these crystal records place fundamental constraints on the processes operating within the reservoirs. Data from volcanic crystals are in accord with recent conceptual models of magma reservoirs being composed dominantly of crystal mushes, with small volumes and/or small fractions of melt present. The implication is that magma reservoirs have differing modes of behavior: magmas are stored over the long term in largely crystalline, quiescent, conditions, punctuated by brief episodes of intense activity during the decades to centuries immediately prior to an eruption.

DOI: 10.2113/gselements.13.1.23

Keywords: magma reservoirs, volcanic crystals, crystal mush, magmatic time scales, chemical diffusion, radiometric dating


  • Facebook
  • Twitter
  • Google+

Figure 1. (A) Artist’s conception of the magma chamber beneath Yellowstone Caldera (USA), illustrating the popular conception that a magma chamber is a “big tank” of magma. Image Credit: US National Parks Service. (B) Schematic diagram illustrating a crystal-mush dominated reservoir beneath Long Valley Caldera (California, USA). Modified from Hildreth and Wilson (2007). (C) Schematic model for the Yellowstone magma system based on seismic data. Blue and yellow regions near the surface indicate basaltic and rhyolitic eruptive deposits, respectively. Reprinted with permission of AAAS from Huang et al. (2015).

Magmas are generated deep within the Earth, and only a fraction of them eventually make it to the surface to produce eruptions. In between magma generation and eruption, all magmas must transit the crust, during which time they may stall and be stored within crustal magma reservoirs at various depths. But what do these magma reservoirs look like? What kind of conceptual model is most accurate? The most common popular conception is of a single, simple “magma chamber”: a large, homogeneous, body of completely molten magma surrounded by country rocks, with a sharp transition between the two (Fig. 1A). However, the scientific consensus has shifted in recent decades to a more complex picture of a “magma reservoir,” consisting largely of crystal mush (a mixture of crystals and liquid where the crystals make up more than ~40%–50% of the volume), grading to solid material (country rock plus solidified magmas) at the margins. In this model, only a relatively small proportion of the magma is liquid at any given time (Fig. 1B, 1C). The magma reservoir concept fits with a wide variety of observations, but it brings up additional questions. How big are these reservoirs? How long do they survive in the crust, and how much input of mass and heat is necessary to sustain them? How does their physical state (e.g. temperature, percentage of melt) vary over space and time within the reservoir? When, where, and for how long are largely liquid bodies of magma present? And what processes and timescales are necessary to accumulate the volumes of largely liquid magma that are erupted?

Four broad categories of methods can provide information about the state and evolution of magma reservoirs: geophysical imaging and remote sensing (Lees 2007; Kiser et al. 2016); theoretical or computational models (Huber et al. 2012; Annen et al. 2015; Bergantz et al. 2015); observations of plutonic rocks representing former magma reservoirs (e.g. Coleman et al. 2016); observations of volcanic rocks. Each of these four methodologies captures a different slice of a reservoir’s history and answers different parts of the question (see other articles in this issue for further discussion).

This paper focuses on crystals in volcanic rocks as an archive of the conditions and processes operating within magma reservoirs. This sample is biased, both because the magmas that have been erupted are a minority of the magmas generated in the Earth’s interior, and because each erupted magma is probably not representative of the reservoir as a whole. However, this sample bias can, in fact, be useful. The crystals directly sample the part of the reservoir that produces eruptions, providing information about the active part of the reservoir. For example, the crystals that grow within a subsurface magma can be slow to equilibrate chemically after compositional changes in the local environment and can, therefore, contain detailed, long-term records of such changes. Furthermore, these crystals provide a unique perspective on active reservoir processes because they capture the long-term (tens to hundreds of thousands of years) history of the region feeding eruptions. This then provides an important complement to other methods of studying active volcanoes.

The crystals within volcanic rocks can be broadly divided into two categories: major phases and accessory phases. The major phases are the major rock-forming minerals (e.g. olivine, feldspars, pyroxenes, amphiboles) that would make up the bulk of the solidified rock were it to crystallize to completion. The proportions of these major phases present are functions of variables such as the major-element composition of the bulk magma, pressure–temperature conditions of crystallization, and volatile contents of the magma. In contrast, accessory phases are present in abundances of much less than 1% of the total volume of the rock, and their presence in the magma often reflects (at least in part) the abundances of trace elements in the melt, in addition to those variables controlling the major-element composition. For example, zircon saturation in magmas is sensitive to the zirconium concentration in the melt as well as the temperature and bulk melt composition (Boehnke et al. 2013). Major phases form the network of a crystal mush, but the accessory phases are much less abundant and are smaller. Therefore, during processes that extract interstitial liquid from a crystal mush, the accessory phases may travel with the melt rather than remaining with the crystal network (Claiborne et al. 2010). As a result, the major and accessory phases may be sensitive to different variables within the reservoir, may be segregated over time, and may, therefore, capture different parts of the history of magma.

Accessing the records of magmatic processes that are contained within the crystals, and placing the changes in crystal chemistry within a temporal context, requires connecting information about time with information about processes operating within a magma reservoir, such as injections of new magma, cooling, degassing, crystallization of resident magma, and movement of magmas within a reservoir (including differential crystal–liquid movement such as segregation of liquid-rich bodies from a mush). Many of these processes leave chemical “fingerprints” in growing crystals that are reflected in changes in the major- or trace-element or isotopic composition of the minerals. Time information can be obtained through radiometric dating (which provides an absolute age of crystallization for a mineral or a population of minerals sampled by a bulk mineral separate) or through kinetic clocks, such as diffusion (Costa et al. 2008) or faceting of inclusions (Pamukcu et al. 2015), which provide information about the duration of the kinetic processes. Modern analytical techniques allow researchers to measure compositional and age variations in crystals at the micron scale. Major elements can be analyzed by electron microprobe, trace elements can be analyzed to submicron scale by nanoscale secondary ion mass spectrometry (NanoSIMS), and trace elements and Pb and Hf isotopic compositions can be analyzed to tens of microns scale by laser-ablation multicollector inductively coupled plasma mass spectrometry or by SIMS. By combining crystallization ages, kinetic ages, and records of compositional changes, a surprisingly detailed and informative history of past magma reservoir processes can be reconstructed from the crystal record.

Longevity of Magma Reservoirs

  • Facebook
  • Twitter
  • Google+

Figure 2. Compilation of U-series crystal age data for accessory and major phases from a wide variety of erupted magmas. (A) Compilation of model 238U–230Th ages for individual spots (~40 μm in diameter) on zircon and other accessory phase grains from volcanic rocks. Ages are expressed as preeruptive crystal residence (i.e. crystal age minus the eruption age). Each vertical array shows data for a single sample (or a group of closely related samples), where lines indicate the range of ages for individual spot analyses; solid lines for interior analyses and dashed lines for unpolished surface analyses. The mean ages, or the dominant peaks on probability density functions (pdf), are shown as solid circles; secondary pdf peaks are indicated by open circles. Upward-pointing arrowheads indicate analyses within error of secular equilibrium (older than ~400 ka); downward-pointing arrowheads indicate analyses within error of eruption age. ‘CA’ is California. Modified from Cooper (2015). (B) Compilation of 238U–230Th (light blue circles) and 230Th–226Ra (dark blue diamonds) ages of bulk mineral separates of major phases, expressed as preeruptive residence age. Vertical lines show maximum and minimum residence ages calculated by combining the 2s uncertainties on the eruption age and the crystal age. Arrows on the error bars represent samples with ages within error of eruption age (downward-pointing arrowheads) or secular equilibrium (upward-pointing arrows). Mean or best estimate ages are shown by the symbols; lines without symbols indicate samples for which only a maximum and minimum age estimate were available. ‘OVC’ is Okataina Volcanic Center. Modified from Cooper and Kent (2014) and Cooper (2015).

The longevity of a magma reservoir depends on how it is defined. Life spans of volcanic systems (i.e. multiple volcanoes or vents active over time in a broad area) are similar to the total duration of magmatism recorded in plutonic systems (~10 My) (Grunder et al. 2006; Coleman et al. 2016). However, the duration of activity at any given volcanic edifice is shorter, typically ~1 My (Grunder et al. 2006). Taking one further step down in scale, individual volcanic eruptions contain crystals with ages that typically span tens to hundreds of thousands of years (Reid 2003; Zellmer et al. 2005; Turner and Costa 2007; Schmitt 2011; Cooper 2015) (Fig. 2). Thus, the life span of activity at a given magmatic system seems to be on the order of millions to ~10 million years, of which a “moving window” of tens to hundreds of thousands of years of activity is sampled by any given eruption. But the residence time of eruptible magma itself seems much shorter—decades or centuries up to a few millennia. These apparently contradictory estimates of storage timescale can, however, be reconciled if the conditions and timescales of long-term storage within the “background” state of the reservoir are different from the conditions and timescales of accumulation and storage of magma bodies just before eruption.

Physical State of Magma Reservoirs

The replacement of the “big tank” model of a magma chamber with a “crystal mush” model (e.g. Hildreth and Wilson 2007; Bachmann and Bergantz 2008) has been driven by the following observations:
Seismic imaging of magma reservoirs is, typically, consistent with the presence of a few tens of percent of melt at most (Lees 2007; Huang et al. 2015; Kiser et al. 2016).

Crystals within volcanic rocks typically show considerable diversity in composition (ranging from major- and trace-element zoning to differences in isotopic composition both within and between crystals) and in age, inconsistent with a simple cooling and crystallization history for the reservoir as a whole (e.g. Reid 2003; Schmitt 2011; Storm et al. 2014; Cooper 2015; Stelten et al. 2015).

Numerical models of periodic injection of magma recharge into a reservoir are consistent with maintaining some melt-dominated zones within a larger, partially molten region over long periods of time (e.g. Huber et al. 2012; Annen et al. 2015).

Furthermore, liquid-dominated bodies of magma demonstrably exist, because they are erupted. But for how long are eruptible magma bodies stored in the reservoir prior to eruption, and how are they produced within an overall mushy reservoir? The question of what a magma reservoir looks like becomes dependent on its size and on the timescales being considered. For example, records from volcanic crystals can provide a much more detailed picture than the reservoir-scale averaging inherent in most numerical models or geophysical imaging. Critically, the ability to date individual crystals (or, in the case of major phases, crystal populations) and to compare ages to crystal-scale chemical information provides constraints on the temporal evolution of the chemical environment within the reservoir.

Some important observations can be drawn from the age data alone. For example, compiled age data for zircons (and other accessory minerals) in volcanic rocks show a large range of ages in single eruptions, within a single hand sample, and even, in some cases, within a single crystal (e.g. Reid 2003; Schmitt 2011; Cooper 2015) (see Fig. 2A, where the vertical lines encompass the range of individual spot ages). This has two implications: First, eruptions can tap parts of the magma reservoir that have existed for a significant part of a volcano’s lifetime; second, each eruption taps material of diverse ages and, therefore, samples parts of the reservoir that may have experienced different compositional and thermal histories. There are also hints that average ages of bulk mineral separates for the major phases may be younger than the averages ages for zircon data (in Fig. 2B, roughly half of the bulk separate ages are <10 ka, whereas in Fig. 2A a smaller proportion of the main probability density function peaks for accessory phases are <10 ka). Maybe the major and accessory minerals are sampling different parts of the reservoir that have different histories. Studies utilizing both types of minerals will, therefore, be needed to provide more information about the evolution of a magma reservoir.

  • Facebook
  • Twitter
  • Google+

Figure 3. Summary of U-series age data for Yellowstone caldera (USA) case study. (A) 238U–230Th zircon spot age vs. Eu/Eu* for the same spots. Orange squares represent spots on polished zircon interiors, white circles represent spots on unpolished zircon surfaces. Eu/Eu* is the europium anomaly (the deviation of Eu measured from that expected based on neighboring rare earth element concentrations), which is related to feldspar fractionation and magma differentiation. The upper and lower blue bands encompass ages of the two most recent episodes of postcaldera volcanism. (B) Average ages and Eu/Eu* of all zircon interior (orange squares) and surface spot ages (white circles) for individual eruptions studied, compared to glass data for the same eruptions (blue pentagons). Arrows indicate temporal trends in glass and sanidine compositions. (C) Average zircon spot ages for crystal interiors (orange squares) and surfaces (white hexagons) compared to sanidine ages (purple circles) and eruption ages (green bars) for the three individual samples. Sanidine and zircon surface average ages are coeval and within error of eruption age (except for zircon surfaces for Grants Pass Flow), whereas zircon interiors are, on average, tens of thousands of years older than eruption. (D) Schematic model for the Yellowstone upper-crustal reservoir showing (1) growth of zircon interiors in a long-lived, heterogeneous crystal mush, followed by (2) melt (+zircon) extraction from the mush (white arrows) and accumulation in a separate body. All diagrams modified from Stelten et al. (2015).

In addition to providing information about a magma reservoir’s longevity, crystals also act as archives for how a magma reservoir’s composition may have changed. Variations in trace-element and isotopic composition of crystals, especially when combined with age data for the same crystals, can provide information about the degree to which different parts of a reservoir demonstrate coherent compositional evolution. A comparison of combined age and trace-element data for zircon surfaces (white symbols in Fig. 3A) with age and trace-element data for zircon interiors (orange symbols in Fig. 3A) in a recent study of post-caldera volcanic activity at Yellowstone Caldera (USA) (Stelten et al. 2015), shows that the zircon surfaces are relatively compositionally restricted at any given time but that surfaces and interiors both show systematic variations in composition over time (Fig. 3B). This pattern of data indicates that the zircon crystals began growing in a long-lived, compositionally diverse part of the reservoir, then moved to a region that was more restricted in composition where the final stages of zircon growth (the surfaces) occurred. At the same time, the average compositions of zircon interiors and surfaces mimic temporal changes in glass composition (Fig. 3B), suggesting a coherent change in the average chemical composition of the reservoir over time. Sanidine ages (Fig. 3C) and lead isotopic compositions (not shown) from the same Yellowstone samples (purple symbols in Fig. 3C) document that the sanidine also crystallized shortly prior to eruption from a compositionally restricted part of the reservoir. Because sanidine would be expected to crystallize from rhyolitic magma during the entire interval of zircon crystallization, this suggests that the sanidine that crystallized coevally with the zircon interiors was not erupted. Considering all this information together produces a model of the magma reservoir below Yellowstone (Fig. 3D) where melt plus zircon interiors are extracted from a long-lived, compositionally diverse, crystal mush, leaving behind the sanidine and, presumably, other major phases. The melt plus zircon is then transported to a different, melt-dominated, region of the reservoir where it is stored for a comparatively short time (less than a few thousand years) during which time zircon surfaces and additional sanidine crystallize from the melt prior to eruption (Stelten et al. 2015).

  • Facebook
  • Twitter
  • Google+

Figure 4. How to constrain a magma’s thermal history. (A) Trace-element concentrations measured in a profile across a crystal. In this example, measured Sr concentrations (red line) across a plagioclase crystal (Inset) are far from concentrations expected for equilibrium with host liquid (purple line). Thin dashed lines show modeled evolution of the Sr concentration profile from initial concentrations toward equilibrium, calculated at 850 °C. (B) Diagram illustrating how a crystal’s thermal history can be constrained using modeled diffusion durations combined with absolute crystal ages. The blue line shows a schematic temperature–time history of crystals within a magma reservoir, where high temperatures could be related to recharge events (or other heat input events to the reservoir), followed by cooling. The arrowed blue line indicates the total time recorded by the crystals, as captured by the U-series crystallization ages. The red regions (and red segments in top section) indicate the maximum duration of storage at or above a specified temperature that could preserve the measured profile. Temperatures of liquidus (the temperature above which the magma is entirely liquid; Tliq), solidus (the temperature below which the magma is entirely solid; Tsol) and an arbitrary modeled diffusion temeperature (Td) are shown as dashed lines. Modified from Cooper and Kent (2014).

The concept that much of a magma reservoir exists as a crystal mush (or even near-solidus) for much of the time is supported by constraints on the thermal history of crystals from within magma reservoirs. Trace-element concentration variations within crystals are a natural consequence of crystal growth within a compositionally evolving magma reservoir and/or of crystals being transported from one region of a compositionally diverse reservoir to another. Trace-element variations are commonly observed in crystals from arc volcanoes, such as Mount Hood (Oregon, USA) (Cooper and Kent 2014) (see red line in Fig. 4A).

Frequently, these changing environments during growth lead to different zones within a crystal that are not in chemical equilibrium with each other, or with the host liquid (compare purple line with red line in Fig. 4A). As a consequence, the trace elements will diffuse between zones within the crystal, moving towards equilibrium. The rate of this diffusive reequilibration within a crystal is exponentially dependent on temperature. If initial conditions and diffusion rates for a particular element in a given crystal are known or can be estimated, trace-element profiles can be modelled to constrain the duration of crystal residence at a particular temperature (Costa et al. 2008). Even if all of these parameters are not fully constrained, the time over which disequilibrium between crystal and host liquid can be preserved is a strong function of temperature. This means that some order-of-magnitude information about storage durations can be obtained even in the absence of complete information. By combining the crystal ages (blue line at the top of Fig. 4B) with diffusion modelling, assuming a particular temperature (Td in Fig. 4B), it is possible to constrain the maximum percentage of the total crystal lifetime that could have been spent at that temperature (red line segments and areas in Fig. 4B).

In the context of the mush model of magma reservoirs, one temperature that is of particular interest is the “­rheological lock-up” temperature—the temperature at which a particular magma composition reaches a crystallinity of ~40%–50%, causing a transition from crystals suspended in a liquid to liquid existing within a locked crystal network—a crystal mush. In the case of Mount Hood, preservation of Sr disequilibria within plagioclase crystals limits the time spent at 750 °C (the lock-up temperature for this magma composition) to a few centuries up to 2,800 years, corresponding to a maximum of 12% of the minimum age of the crystal cores (21 ka), and <<1% of the average crystal age (126 ka)(Cooper and Kent 2014). Of course, more complex thermal histories than storage at a single temperature are possible (and even likely), but because diffusion operates more quickly at higher temperatures, any time spent above 750 °C would shorten the duration of storage implied at 750 °C. Work in progress on other volcanic systems suggest that this is a common phenomenon. Thus, the long-term storage conditions of magma reservoirs—or at least the parts of the reservoirs sampled by the crystals in volcanic rocks—likely exist as a crystal mush or in a near-solidus state. This is consistent with the lack of geophysical data indicating high percentages of melt present beneath volcanoes (e.g. Lees 2007). Although some numerical models suggest that maintaining magma reservoirs in a partially molten state for long periods of time requires unusually high rates of input of new, hot magma (e.g. Annen et al. 2015), others support the concept of long-term storage of magmas in a crystal mush (e.g. Huber et al. 2012). Therefore, multiple lines of evidence are consistent with a dominantly crystalline (>50%) reservoir over most of a volcanic system’s history, although this model has not been universally accepted (e.g. Barboni et al. 2016).

How Long Does it Take to “Prime” a Reservoir for Eruption?

The mush model of magma reservoirs does present something of a paradox: in order to erupt, magmas must be above the rheological lock-up temperature, yet much of a reservoir likely exists for most of the time as a crystal mush below the rheological lock-up temperature. Therefore, an important implication of the thermal history data is that all of the “action” in terms of rejuvenation of a crystal mush, accumulation of erupted bodies of magma, mixing between resident and recharge magmas, growth of crystals or rims of crystals that are in equilibrium with the erupted melt composition, must take place within a comparatively short period of time. Indeed, the Yellowstone study discussed above is but one example of a growing body of evidence suggesting that erupted magma bodies are assembled rapidly—within decades to centuries of eruption, or at most a few millennia—from compositionally diverse and longer-lived magma reservoirs (e.g. Costa et al. 2008; Druitt et al. 2012, 2013). Further evidence for this idea comes from timescales of diffusion of trace elements across compositional boundaries within crystals (Costa et al. 2008), and from the kinetics of growing crystal faces from initially rounded melt inclusions (Pamukcu et al. 2015).

The concept of rapid remobilization of a mush is supported by numerical models that explore the conditions required for remobilization, although different models have widely varying predictions about reactivation times (Burgisser and Bergantz 2011; Huber et al. 2012; Parmigiani et al. 2014; Bergantz et al. 2015). In addition, numerical models that track individual crystals during reactivation support the idea that crystals from chemically diverse layers can be intimately mixed during the process of remobilization (Bergantz et al. 2015). Finally, there is growing evidence that eruptions (even very large ones) can be fed by multiple smaller bodies that may be distributed within a crystal mush (Allan et al. 2013; Storm et al. 2014). In some cases, it appears that multiple magma bodies were tapped during the course of a single eruption without having been assembled into a larger body prior to eruption (e.g. Eyjafjallajökull volcano in Iceland; see Cashman and Giordano 2014 and references therein). Thus, the emerging picture of what a magma reservoir looks like is highly dependent on when you are looking. Reservoirs appear to spend most of their existence in a quiescent, largely crystalline state (where crystallization, diffusion, and compositional evolution operate slowly) that is punctuated by brief periods of intense activity during which diverse crystals are gathered together and are mixed along with their host liquids, crystallization occurs rapidly, and the magmas become primed for eruption.

Future Directions

A growing body of evidence supports the idea of dominantly cool and mushy storage conditions and the rapid assembly of erupted magma bodies. Yet, much remains to be explored. Are storage conditions different for the magmas that feed eruptions of different sizes? Do different ­compositions of magmas have different storage histories? What are the conditions required to maintain a magma reservoir as a relatively cold crystal mush for long periods of time, and how do these conditions change when “priming” the system for eruption? Are the liquid and crystal components of erupted magmas stored separately and only combined immediately before eruptions? What geophysical or other observational signals should we expect to see when a reservoir is in a “normal” background state compared to being in an active state that may precede eruption?

Three themes of ongoing research encourage optimism that we are poised to make progress toward answering these questions. First, efforts to directly constrain the thermal histories of magma storage are increasing in number. The fruits of these efforts will provide much-needed observational evidence about how thermal conditions of magma storage evolve over space and time. Having this kind of data for eruptions for a range of sizes and compositions will allow us to look for commonalities or differences in behavior that will be key to understanding the variables that drive storage and eruption. Second, modelling of processes in magma reservoirs has been rapidly advancing in the sophistication of situations that can be simulated. This increasingly provides modelling output that can be directly compared to the observational records, which should allow us to explore the physical processes of storage and mobilization. Third, there is an increasing ability to compare petrologic data to geophysical data in order to better understand the deformation signals (Kahl et al. 2013).

In short, there is closer communication than ever between petrological, geophysical, and modelling approaches to understand magma storage and mobilization, with enormous potential to make fundamental advances in our understanding of magma storage systems and eruptions.


The ideas presented here were developed over a number of years and under the auspices of many NSF awards including, in particular, EAR-0738749, EAR-0838389, EAR-1144945 and EAR-1426858, and have relied heavily on the work of former and present graduate students Gary Eppich, Mark Stelten, Allison Rubin, and Kevin Schrecengost. I thank Mary Reid, Olivier Bachmann, Fidel Costa, Adam Kent, George Bergantz, Jorge Vazquez, and Christy Till for many fruitful discussions. Thanks also to Keith Putirka, Gordon Brown, and Jodi Rosso for their editorial handling and thoughtful analysis of the manuscript, and to reviewers Olivier Bachmann, Kathy Cashman, and Georg Zellmer for comments and suggestions that led to substantial improvements in the manuscript.


Allan ASR, Morgan DJ, Wilson CJN, Millet M-A (2013) From mush to eruption in centuries: assembly of the super-sized Oruanui magma body. Contributions to Mineralogy and Petrology 166: 143-164

Annen C, Blundy JD, Leuthold J, Sparks RSJ (2015) Construction and evolution of igneous bodies: Towards an integrated perspective of crustal magmatism. Lithos 230: 206-221

Bachmann O, Bergantz GW (2008) Rhyolites and their source mushes across tectonic settings. Journal of Petrology 49: 2277-2285

Barboni M and 6 coauthors (2016) Warm storage for arc magmas. Proceedings of the National Academy of Sciences of the United States of America, doi: 10.1073/pnas.1616129113

Bergantz GW, Schleicher JM, Burgisser A (2015) Open-system dynamics and mixing in magma mushes. Nature Geoscience 8: 793-796

Boehnke P, Watson EB, Trail D, Harrison TM, Schmitt AK (2013) Zircon saturation re-revisited. Chemical Geology 351: 324-334

Burgisser A, Bergantz GW (2011) A rapid mechanism to remobilize and homogenize highly crystalline magma bodies. Nature 471: 212-215

Cashman KV, Giordano G (2014) Calderas and magma reservoirs. Journal of Volcanology and Geothermal Research 288: 28-45

Claiborne LL, Miller CF, Flanagan DM, Clynne MA, Wooden JL (2010) Zircon reveals protracted magma storage and recycling beneath Mount St. Helens. Geology 38: 1011-1014

Coleman DS, Mills RD, Zimmerer MJ (2016) The pace of plutonism. Elements 12: 97-102

Cooper KM, Kent AJR (2014) Rapid remobilization of magmatic crystals kept in cold storage. Nature 506: 480-483

Cooper KM (2015) Timescales of crustal magma reservoir processes: insights from U-series crystal ages. In: Caricchi L, Blundy JD (eds) Chemical, Physical and Temporal Evolution of Magmatic Systems. Geological Society, London, Special Publication 422. Geological Society of London, pp 141-174

Costa F, Dohmen R, Chakraborty S (2008) Time scales of magmatic processes from modeling the zoning patterns of crystals. Reviews in Mineralogy and Geochemistry 69: 545-594

Druitt TH, Costa F, Deloule E, Dungan M, Scaillet B (2012) Decadal to monthly timescales of magma transfer and reservoir growth at a caldera volcano. Nature 482: 77-80

Grunder AL, Klemetti EW, Feeley TC, McKee CM (2006) Eleven million years of arc volcanism at the Aucanquilcha Volcanic Cluster, northern Chilean Andes: implications for the life span and emplacement of plutons. Earth and Environmental Science Transactions of the Royal Society of Edinburgh 97: 415-436

Hildreth W, Wilson CJN (2007) Compositional zoning of the Bishop Tuff. Journal of Petrology 48: 951-999

Huang H-H and 5 coauthors (2015) The Yellowstone magmatic system from the mantle plume to the upper crust. Science 348: 773-776

Huber C, Bachmann O, Dufek J (2012) Crystal-poor versus crystal-rich ignimbrites: A competition between stirring and reactivation. Geology 40: 115-118

Kahl M and 5 coauthors (2013) Compositionally zoned crystals and real-time degassing data reveal changes in magma transfer dynamics during the 2006 summit eruptive episodes of Mt. Etna. Bulletin of Volcanology 75, doi: 10.1007/s00445-013-0692-7

Kiser E and 8 coauthors (2016) Magma reservoirs from the upper crust to the Moho inferred from high-resolution Vp and Vs models beneath Mount St. Helens, Washington State, USA. Geology, doi: 10.1130/G37591.1

Lees JM (2007) Seismic tomography of magmatic systems. Journal of Volcanology and Geothermal Research 167: 37-56

Pamukcu AS, Gualda GAR, Bégué F, Gravley DM (2015) Melt inclusion shapes: timekeepers of short lived giant magma bodies. Geology 43: 947-950

Parmigiani A, Huber C, Bachmann O (2014) Mush microphysics and the reactivation of crystal-rich magma reservoirs. Journal of Geophysical Research: Solid Earth 119: 6308-6322

Reid MR (2014) Timescales of magma transfer and storage in the crust. In: Rudnick RL (ed) The Crust. Treatise on Geochemistry Vol. 4. 2nd Edition. Elsevier, pp 181-201

Schmitt AK (2011) Uranium series accessory crystal dating of magmatic processes. Annual Review of Earth and Planetary Sciences 39: 321-349

Stelten ME, Cooper KM, Vazquez JA, Calvert AT, Glessner JJG (2015) Mechanisms and timescales of generating eruptible rhyolitic magmas at Yellowstone Caldera from zircon and sanidine geochronology and geochemistry. Journal of Petrology 56: 1607-1642

Storm S, Schmitt AK, Shane P, Lindsay JM (2014) Zircon trace element chemistry at sub-micrometer resolution for Tarawera volcano, New Zealand, and implications for rhyolite magma evolution. Contributions to Mineralogy and Petrology 167, doi: 10.1007/s00410-014-1000-z

Turner S, Costa F (2007) Measuring timescales of magmatic evolution. Elements 3:267-272

Zellmer GF and 5 coauthors (2005) Magma evolution and ascent at volcanic arcs: constraining petrogenetic processes through rates and chronologies. Journal of Volcanology and Geothermal Research 140: 171-191

Share This