Advances in analytical methods have provided new insights into the

Measuring Timescales of Magmatic Evolution Simon Turner1 and Fidel Costa2 A dvances in analytical methods have provided new insights into the timesc...
Author: Alvin Barnett
3 downloads 0 Views 3MB Size
Measuring Timescales of Magmatic Evolution Simon Turner1 and Fidel Costa2

A

dvances in analytical methods have provided new insights into the timescales of magmatic processes. Data on the abundances of U-series isotopes in bulk rocks and crystal separates indicate magma differentiation over thousands of years. Residence and differentiation times of silicic magmas based on single-crystal, in situ age data vary from 10,000 to 100,000 years, with abundant evidence for crystal recycling from previous intrusive episodes. Chemical zoning patterns in single crystals indicate that processes such as mixing and mingling of magmas and crustal assimilation may occur over much shorter timescales of months to decades. Quantifying the rates of magma generation, emplacement and differentiation constrains the processes involved and may contribute to the evaluation of volcanic hazards.

complementary contributions are improving our understanding of magmatic processes. Reid (2003) and Hawkesworth et al. (2004) recently reviewed timescale information.

METHODOLOGY

Short-Lived Radioactive Isotopes and In Situ Determinations

Advances in mass spectrometry techniques, including SIMS (secondary ion mass spectrometry), have enabled increasingly precise KEYWORDS: timescales, magmatic processes, diffusion, U-series, phenocrysts measurements of the nuclides of the U-series decay chains in bulk rocks and mineral separates and INTRODUCTION now permit in situ dating of (mainly) zircon and allanite The timescales and rates of magmatic processes are key (e.g. Bourdon et al. 2003). U-series disequilibria in parental pieces of information for understanding and modeling magmas are typically produced during mantle melting; sysmany aspects of igneous activity. For example, the rates at tems of particular interest here are 230Th–238U (half-life which magmas are transferred from the mantle to the crust 75,000 years), 226Ra–230Th (half-life 1600 years) and 210Pb–226Ra determine the types of physical and chemical processes that (half-life 22 years) (see FIG. 1). Return to secular equilibrium can occur during magma transport. Similarly, the mass and occurs over several (~5) half-lives of the daughter nuclide (see longevity of magmas below calderas or active volcanoes page 273). Lava suites from a number of different volcanoes affect both the time-integrated thermal fluxes available for show a decrease in disequilibria with increasing differentiageothermal energy and the likelihood of hazardous erup- tion (e.g. FIG. 2). Thus, U-series disequilibria can, in principle, tions. The number and quality of timescale determinations date melting and crystallization events up to several of igneous processes have increased dramatically in the last 100,000 years old and, in the case of 210Pb, which has a 15 years, mainly due to advances in experimental and ana- gaseous progenitor (222Rn), degassing up to 100 years old. lytical techniques. Elemental concentrations and isotopic At the same time, technical improvements have also ratios are now determined at unprecedented levels of preci- allowed the acquisition of in situ age data from very young sion and spatial resolution (e.g. Ginibre et al. 2007 this issue; Davidson et al. 2007 this issue). Technical advances include (1) higher-precision measurements of isotopic clocks, including U-series isotopes, on smaller amounts of material and in situ isotope determinations on very young crystals, and (2) modeling of the diffusive re-equilibration of isotopes and elements in a variety of minerals using more precise diffusion coefficients and better spatial resolution. The former allow dating and provide ages whereas the latter gives relative time, which can sometimes be transformed into an age if the age of another event is known (e.g. eruption year). By combining the two techniques, geological processes that span a few hours to millions of years can be determined (FIG. 1). Here we briefly describe the basis and the findings of the two approaches and discuss how their 1

GEMOC, Department of Earth and Planetary Sciences, Macquarie University, Sydney NSW 2109, Australia E-mail: [email protected]

2

CSIC, Institut de Ciències de la Terra 'Jaume Almera', c/ Lluís Solé i Sabarís s/n, 08028 Barcelona, Spain E-mail: [email protected]

ELEMENTS, VOL. 3,

PP.

267–272

FIGURE 1 267

Ranges of timescales that can be determined using radiometric dating and diffusion modeling of chemical zoning in crystals

A UGUST 2007

A

B

(226Ra/230Th) versus Th (used as an index of differentiation) plot comparing lavas from Akutan volcano (filled squares) with lavas from Aniakchak volcano (filled circles) and those from the Asal rift and Sangeang Api in the Sunda arc. After George et al. 2004

FIGURE 2

minerals; typically the U–Th system in zircon is studied using a secondary ion microprobe (SIMS). These new age determinations have provided new insights and raised new questions. The relationship between the ages of “datable” accessory minerals and main differentiation events driven by the crystallization of major rock-forming minerals is not straightforward. The distinction between entrainment of old crystals and the age of differentiation is important because it bears on the longevity of the magmatic systems, which relates in turn to the rates and types of magmatic inputs into the crust and the associated mode and frequency of volcanic activity.

C

Diffusion Modeling of Zoning Patterns in Minerals and Glasses Diffusion modeling exploits the presence in many minerals of elemental or isotopic zoning. The chemical zoning patterns in crystals and glasses provide a record of magmatic processes. These patterns dissipate or evolve toward equilibrium profiles at a rate that can be experimentally calibrated in terms of diffusion coefficients. Thus, if we know the diffusion coefficient and if we are able to measure the compositions with sufficient spatial resolution and precision, we can obtain elapsed time using Fick’s second law of diffusion (FIG. 3). The approach has been applied for some time to determine the thermal histories of metamorphic rocks. Renewed interest in igneous systems (e.g. Zellmer et al. 1999; Costa et al. 2003; Morgan et al. 2004) has resulted from the many new and precise determinations of diffusion coefficients made possible by analytical and experimental techniques. One advantage of the method is that the time information can be linked to the textural and chemical features of the rocks at the scale of a thin section (FIG. 4). The method can provide enough data to be treated statistically, because large numbers of crystals can be analysed using routine analytical techniques (e.g. electron microprobe) and because multiple elements can be determined on one or more minerals in the same rock (FIG. 4).

TIMESCALES OF MAGMATIC EVOLUTION

Magma Residence and Differentiation Times The simplest approach to constraining differentiation timescales is based on the observation that U-series disequilibrium decreases with increasing differentiation in cogenetic ELEMENTS

Schematic illustration of the principles of obtaining timescales from modeling chemical or isotopic gradients in crystals. (A) Crystal is growing from a liquid; in this case it is assumed to be unzoned. (B) A change in environmental conditions (temperature, pressure, composition) occurs in response to a process like magma mixing, and this is recorded in the zoning of the crystal. In the case shown, the system responds with the growth of a rim of a different composition (e.g. Davidson et al. 2007). (C) The measured concentration will be a combination of growth or dissolution plus diffusion. The first task is to determine the initial profile prior to diffusion (initial profile = Cini), which can be accomplished by considering elements that have different diffusion rates or by using other geological arguments about the evolution of the sample (e.g. Costa et al. 2003; Morgan et al. 2004). The zoning of the crystal is analyzed in terms of Fick’s second law; this requires knowing the diffusion coefficient (D) and the boundary conditions (e.g. if the crystal exchanges mass with the liquid, as shown in the example). The timescale of the process depends on how far the measured profile is from the initial and the equilibrium concentrations (Ceq).

FIGURE 3

suites of volcanic rocks. FIGURE 2 illustrates that, for many small- to moderate-size magmatic systems, this decrease is observed for (226Ra/230Th) (where brackets indicate activity ratios, and secular equilibrium is defined by an activity ratio of 1), from which it is inferred that the timescale of differentiation is on the order of thousands of years. The approach assumes that the time elapsed during differentiation is the only cause of a decrease in (226Ra/230Th). However, in cases where partial melts of crustal materials are in secular equilibrium (e.g. Berlo et al. 2004), the effects of assimilation will also lead to a decrease in (226Ra/230Th) and a corresponding shift in (230Th/238U) towards 1. Conversely, the common trend of decreasing disequilibrium with increasing

268

A UGUST 2007

A

the results can be more complex to interpret, not least because the crystals contained in a rock are not the ones that separated from the parental liquid to produce the bulkrock composition (Hawkesworth et al. 2004). For example, Cooper and Reid (2003) and Turner et al. (2003) analysed U–Th–Ra disequilibria in plagioclase separates from a number of volcanoes and found discrepancy between the U–Th and Ra–Th systems (e.g. FIG. 5). This was interpreted to indicate that the crystals contain recycled cumulate materials, and it is likely that individual crystals are also zoned in age. In some cases, textural information from crystal size distributions supports the notion of entrainment of cumulates (FIG. 5; Turner et al. 2003).

B

D

C

E

Example of zoning patterns and diffusion modeling of multiple elements in olivine from a basalt of the Tatara–San Pedro volcanic complex. (A) Photomicrograph (partially crossed polars) of olivine crystals in a plagioclase-rich matrix. The positions of two electron microprobe traverses are shown as arrows labeled T6 and T7. The rectangle delimits the area of the X-ray map. (B) X-ray map of Mg concentration. Red is high concentration and blue is low. (C) Crystallographic axis orientations for the same crystal and electron microprobe traverses projected onto the lower hemisphere. Note that T6 is almost parallel to the a-axis and T7 to the c-axis. The orientation of the olivine was obtained in situ by electron backscatter diffraction. (D) and (E) Profiles of traverses T6 and T7 showing the measured concentrations, initial conditions, model profiles, and calculated times for 1125°C and oxygen fugacity at the Ni–NiO buffer. The diffusion model includes a compositional dependence for Fe–Mg and Ni, the effect of oxygen fugacity, and the effect of anisotropy. The results of the two traverses for all elements overlap at 5.4 ± 2.5 years. This suggests that the zoning patterns are due to diffusive exchange, that the assumption of initial profile is appropriate, and that the diffusion coefficients are correct. In this case the times reflect the duration between assimilation of gabbroic plutons by basaltic magma and eruption. Modified after Costa and Dungan (2005).

FIGURE 4

differentiation can be taken as evidence that crustal partial melts are typically close to secular equilibrium. Nevertheless, because the effects of assimilation are likely to reduce the disequilibria inherited from parental magmas, the timescales inferred should be viewed as maxima, and in the extreme case of mixing of two end-members, differentiation could occur instantaneously with respect to the halflife of 226Ra. A complementary approach is to obtain ages from crystals in volcanic rocks. In some cases good mineral isochrons appear to provide robust information on the timescales of crystallization (e.g. Heumann and Davies 2002). However,

ELEMENTS

The previous section highlights some of the limitations of analyses of bulk mineral separates but, unfortunately, measurement of U–Th disequilibria in individual crystals is not yet possible for most mineral species. High-U–Th accessory minerals such as zircon are exceptions to this generalization, and these have provided insight into the evolution of several large-volume rhyolite systems. Studies of large silicic deposits (>100 km3) have provided evidence that some magma residence times, including the time taken for differentiation, are one to two orders of magnitude longer than the ~1000 years inferred above (FIG. 6). Some of the most spectacular results arise from the in situ age determination of zircon and allanite. Brown and Fletcher (1999) dated zircon crystals from Whakamaru Group ignimbrites (Taupo Volcanic Zone) and found crystal cores 250,000 years older than the rims. Similarly, Vazquez and Reid (2004) reported U-series age data from allanite from the Toba eruption; there, crystal cores are 160,000 years older than the rims. In both examples the range of ages between core and rim suggests that the crystals might have grown uninterrupted for long periods of time. Charlier and Zellmer (2000) reported data from different size fractions of bulk zircon separates from the Taupo Oruanui eruption and showed that a correlation exists between crystal size and age. These ages range from 6000 to 12,000 years, which is broadly consistent with differentiation timescales inferred from island arc wholerock 226Ra data. However, models allowing for continuous zircon growth imply much older ages of ~90,000 years (Charlier and Zellmer 2000), similar to in situ zircon data (see below). In other examples it has not been possible to spatially resolve more than ‘cores and rims’, and in most SIMS studies only a single point on each crystal has been analysed. Ages from the Bishop Tuff (Reid and Coath 2000; Simon and Reid 2005) and the Rotoiti and Oruani tuffs in Taupo (Charlier et al. 2003, 2005) indicate residence times of a few thousand to several hundred thousand years (FIG. 6). However, it is not always clear whether the dated crystals grew from the host magma or were recycled from pre-existing intrusions (e.g. Charlier et al. 2005). In a few studies, diffusion has been used to obtain timescales for magmatic differentiation. Morgan and Blake (2006) used Sr and Ba concentrations in sanidine from the Bishop Tuff and obtained timescales of ca. 100,000 years for magma differentiation. This result is in agreement with the higher end of the age range obtained using SIMS (FIG. 6). In contrast, Zellmer et al. (1999) used Sr zoning in plagioclase crystals from St. Vincent (Lesser Antilles) and Kameni (Aegean) volcanoes and determined that the time elapsed between shallow-level crystallization and eruption was of the order of 100 to 450 years. This range is shorter than those inferred from U–Th–Ra studies. These age differences may be due to the recycling of crystals which were cooled before they could equilibrate by diffusion (see Turner et al. 2003 for further discussion).

269

A UGUST 2007

A

Magma Assimilation, Magma Mixing and Pluton Remobilisation

B

Diffusion studies have provided significant constraints on the nature of open-system processes. The time required for magma mixing or mingling has been investigated using Fe–Ti gradients in oxides and major- plus trace-element zoning in olivine and plagioclase (see data compilation in Costa and Chakraborty 2004). The pattern that has emerged from these studies is that mixing between end-members that are compositionally similar (e.g. two mafic magmas) requires less time—only a few months—than the years to decades estimated for mixing between dacite and basaltic andesite. Comparison of these data with measured timescales for deformation of volcanic edifices (Costa and Chakraborty 2004) suggests that it may be possible to correlate the behaviour of the igneous system at surface with its behaviour at depth. Similarly, constraints have been placed on the amount of time elapsed between assimilation of crustal wall rocks and eruption in mafic and silicic systems (FIG. 6; e.g. Wolff et al. 2002; Costa and Dungan 2005). Costa et al. (2003) determined timescales on the order of 100 years for melt and fluid percolation and metasomatism of gabbroic xenoliths using Ca–Na, Mg, Fe, K and La zoning in plagioclase. The time required for remobilisation of completely or partially crystallized rocks to yield silicic magmas has also been explored in ways other than those used for obtaining the SIMS zircon ages discussed above. Based on oxygen isotope disequilibrium between different minerals, Bindeman and Valley (2001) calculated that between 500 and 5000 years elapsed between melting and eruption of the post-caldera lavas in Yellowstone. Similar time periods (10 to 1200 years) were obtained by Zellmer et al. (2003) using Sr and Ba zoning in plagioclase from the Soufrière Hills volcano (FIG. 6). These periods are, however, much longer than the days to weeks determined for the same rocks by Devine et al. (2003) using Fe–Ti zoning in magnetite. The different results obtained by the two approaches probably reflect the much faster Fe–Ti diffusion rates in oxides compared to those for Sr and Ba in plagioclase, such that the former record only the last reheating event.

C

Magma Transport Rates D

(A) Photomicrograph of a Tongan andesite (plag = plagioclase, cpx = clinopyroxene, mt = magnetite). (B) Crystal size distribution with crystal growth times for linear sections of the plots based on a plagioclase growth rate of 10-11 cm·s-1 (n = the number of crystals measured). (C) U–Th equiline diagram. (D) 226Ra/Ba versus time evolution diagram for late groundmass and calculated liquid in equilibrium with the plagioclase separate (after Turner et al. 2003).

FIGURE 5

ELEMENTS

Magma transport rates can be constrained by U-series disequilibria if the site of origin and/or the magnitude of initial disequilibria are known. For example, the positive correlation between 226Ra excess and slab fluid indices in arc lavas (e.g. Ba/Th or Sr/Th ratios) suggests that these magmas ascend through the mantle wedge at ≥100–1000 m/yr (Turner et al. 2001). Similarly, data from ocean island lavas have been interpreted to require melt ascent at ≥10–100 m/yr (Stracke et al. 2006), and recent 210Pb data suggest that melt formed beneath mid-ocean ridges may rise at more than 1000 m/yr (Rubin et al. 2005). These results have led to the consensus that melt is extracted from the mantle via high-porosity channels. Magma transfer times and histories have also been obtained using a diffusion approach. Kelley and Wartho (2000) used 40Ar/39Ar age data from phlogopite to infer transport times of hours to days from the mantle to the crust. Several hours to days were also obtained by Demouchy et al. (2006) and Peslier and Luhr (2006) using zoning in H content in olivine from mantle xenoliths. Notwithstanding these data, Klügel (2001) and Shaw et al. (2006) used Fe–Mg zoning in olivine to constrain the duration of two different events: fast transport of magma between reservoirs and the surface in hours to days, and storage of xenoliths in different crustal reservoirs for years to decades (up to a hundred years or more) before finally reaching the surface. These longer timescales are in agreement with the U-series data discussed above (FIG. 6). 270

A UGUST 2007

A

B

Summary of the timescales determined from radioactive isotopes and from diffusion modeling. See text and references for details.

FIGURE 6

In several recent studies, (210Pb/226Ra) measurements have been used to constrain magma degassing rates. Gauthier and Condomines (1999) presented data from Stromboli and Merapi volcanoes and a model to determine the time required to produce 210Pb deficits, i.e. [(210Pb/226Ra) < 1], by degassing. A survey of island arc volcanic rocks (Turner et al. 2004) showed that many have 210Pb deficits consistent with degassing over a few decades prior to eruption. However, many also contain 210Pb excesses, and this would appear to require gas fluxing from fresh magma at depth (Berlo et al. 2004; Turner et al. 2004). Such signals could potentially be monitored to assess the relative amounts of fresh magma beneath a volcano, and thus they may contribute to eruption forecasting models (Berlo et al. 2006; Reagan et al. 2006). Interestingly, these degassing times are longer than those obtained from diffusion studies. Castro et al. (2005) determined times for bubble nucleation, degassing and quenching using compositional profiles of H2O and found that the processes required between 0.4 and 15 days. In this case, the difference in time between the two approaches may highlight their complementary nature (FIG. 6). The diffusion data refer to the last degassing event and thus are relevant to bubble formation and magma fragmentation processes that happen during eruption, whereas the isotope data relate to degassing of magma at depth and are thus applicable to understanding the longer-term behavior of shallow magma reservoirs.

DISCUSSION The timescales determined by radiometric clocks can complement those deduced from measured diffusion profiles (FIG. 6). In some cases the radiometric clocks yield significantly longer times, but this difference can sometimes be attributed to the inferred temperature–time path of the crystals. This is schematically illustrated in FIGURE 7, in which the U–Th-series ages for some crystals are on the order of 50,000 years whereas the time obtained from diffusion is about 100 years. Crystal ages could record the total time since crystallization, cumulate formation and remobilization by intrusions, whereas the diffusion approach may record only the time since the last replenishment event. This can be visualized by considering the studied crystals as complexly zoned and made of an old core and a much younger rim. The cores ELEMENTS

Interpretation of difference in time information obtained from radioactive isotope and diffusion clocks in the same mineral. (A) The system crystallizes and accumulates minerals for 50,000 years and the radiometric clock is started if the mineral and system are close to diffusive exchange. (B) A large input of new magma disrupts cumulates and partially reacts with them creating a new rim much younger than the core (e.g. Fig. 2). During this event the diffusion clock starts, but shortly after, magma reaches the surface (time = t2 = less than a hundred years). Then, the time obtained from the radiometric clock (50,000 years + t2, but depends on the mass proportions of the old cores and young rims) is probably much longer than that of the diffusion time t2.

FIGURE 7

Magma Degassing

might be from cumulates that were stored for a long time before they were disaggregated by the powerful intrusion of a batch of magma that also triggered the eruption. The time information obtained from the radiometric clock will be a mixed record from the old cores and the rims of crystals formed from this last batch of magma. In contrast, the diffusion clock will record only the time elapsed since formation of the rims and, thus, the time since the last replenishment event that created the driving force for diffusion. Therefore, it is expected that the time obtained from the diffusion clock will be shorter than that derived using the radiogenic approach. Another aspect to note is that the error associated with the U-series ages can be on the order of 1000 years, which is less than the time range for processes that happen just prior to eruption (e.g. a replenishment event). In this case, a U-series clock might indicate the presence of cumulates as old as 50,000 years and allow inferences about the overall longevity of the system, whereas the diffusion data record information related to the processes that lead to eruption. Future progress will rely on further improvements in analytical techniques, better diffusion coefficient data and detailed case studies.

ACKNOWLEDGMENTS FC acknowledges many discussions with Sumit Chakraborty about diffusion in igneous and metamorphic rocks. FC is funded by a Ramon y Cajal Fellowship from the Ministerio de Educación y Ciencia de España and by the DFG (SFB 526, project B7). ST acknowledges a Federation Fellowship from the Australian Research Council. We thank Kari Cooper and Georg Zellmer for helpful reviews and Jon Davidson and Dougal Jerram for their suggestions and their efforts in bringing together this issue. !

271

A UGUST 2007

REFERENCES Berlo K, Turner S, Blundy J, Hawkesworth C (2004) The extent of U-series disequilibria produced during partial melting of the lower crust with implications for the formation of the Mount St. Helens dacites. Contributions to Mineralogy and Petrology 148: 122-130 Berlo K, Turner S, Blundy J, Black S, Hawkesworth C (2006) Tracing pre-eruptive magma degassing using (210Pb/226Ra) disequilibria in the volcanic deposits of the 1980–1986 eruption of Mount St. Helens. Earth and Planetary Science Letters 249: 337-349 Low-δ18O

Bindeman IN, Valley JW (2001) rhyolites from Yellowstone: Magmatic evolution based on analyses of zircons and individual phenocrysts. Journal of Petrology 42: 1491-1517

Bourdon B, Henderson G, Lundstrom C, Turner S (2003) Uranium Series Geochemistry. Reviews in Mineralogy & Geochemistry 52, Mineralogical Society of America, Chantilly, VA, 656 pp Brown SJA, Fletcher IR (1999) SHRIMP UPb dating of the preeruption growth history of zircons from the 340 ka Whakamaru Ignimbrite, New Zealand: Evidence for >250 k.y. magma residence times. Geology 27: 1035-1038 Castro JM, Manga M, Martin MC (2005) Vesiculation rates of obsidian domes inferred from H2O concentration profiles. Geophysical Research Letters 32: doi 10.1029/2005GL024029 Charlier B, Zellmer G (2000) Some remarks on U–Th mineral ages from igneous rocks with prolonged crystallization histories. Earth and Planetary Science Letters 183: 457-469 Charlier BLA, Peate DW, Wilson CJN, Lowenstern JB, Storey M, Brown SJA (2003) Crystallisation ages in coeval silicic magma bodies: 238U–230Th disequilibrium evidence from the Rotoiti and Earthquake Flat eruption deposits, Taupo Volcanic Zone, New Zealand. Earth and Planetary Science Letters 206: 441-457 Charlier BLA, Wilson CJN, Lowenstern JB, Blake S, Van Calsteren PW, Davidson JP (2005) Magma generation at a large, hyperactive silicic volcano (Taupo, New Zealand) revealed by U–Th and U–Pb systematics in zircons. Journal of Petrology 46: 3-32 Cooper KM, Reid MR (2003) Re-examination of crystal ages in recent Mount St. Helens lavas: implications for magma reservoir processes. Earth and Planetary Science Letters 213: 149-167 Costa F, Chakraborty S (2004) Decadal time gaps between mafic intrusion and silicic eruption obtained from chemical zoning patterns in olivine. Earth and Planetary Science Letters 227: 517-530 Costa F, Dungan M (2005) Short time scales of magmatic assimilation from diffusion modeling of multiple elements in olivine. Geology 33: 837-840 Costa F, Chakraborty S, Dohmen R (2003) Diffusion coupling between trace and major elements and a model for calculation of magma residence times using plagioclase. Geochimica et Cosmochimica Acta 67: 2189-2200

ELEMENTS

Davidson JD, Morgan DJ, Charlier BLA (2007) Isotopic microsampling of magmatic rocks. Elements 3: 253-259 Demouchy S, Jacobsen SD, Gaillard F, Stern CR (2006) Rapid magma ascent recorded by water diffusion profiles in mantle olivine. Geology 34: 429-432 Devine JD, Rutherford MJ, Norton GE, Young SR (2003) Magma storage region processes inferred from geochemistry of Fe–Ti oxides in andesitic magma, Soufrière Hills volcano, Montserrat, W.I. Journal of Petrology 44: 1375-1400 Gauthier P-J, Condomines M (1999) 210Pa–226Ra radioactive disequilibria in recent lavas and radon degassing: inferences on the magma chamber dynamics at Stromboli and Merapi volcanoes. Earth and Planetary Science Letters 172: 111-126 George R, Turner S, Hawkesworth C, Bacon C, Nye C, Stelling P, Dreher S (2004) Chemical versus temporal controls on the evolution of tholeiitic and calcalkaline magmas at two volcanoes in the Alaska–Aleutian arc. Journal of Petrology 45: 203-219 Ginibre C, Wörner G, Kronz A (2007) Crystal zoning as an archive for magma evolution. Elements 3: 261-266 Hawkesworth C, George R, Turner S, Zellmer G (2004) Time scales of magmatic processes. Earth and Planetary Science Letters 218: 1-16 Heumann A, Davies GR (2002) U–Th disequilibrium and Rb–Sr age constraints on the magmatic evolution of peralkaline rhyolites from Kenya. Journal of Petrology 43: 557-577 Kelley SP, Wartho J-A (2000) Rapid kimberlite ascent and the significance of Ar-Ar ages in xenolith phlogopites. Science 289: 609-611 Klügel A (2001) Prolonged reactions between harzburgite xenoliths and silicaundersaturated melt: implications for dissolution and Fe-Mg interdiffusion rates of orthopyroxene. Contributions to Mineralogy and Petrology 141: 1-14 Morgan DJ, Blake S (2006) Magmatic residence times of zoned phenocrysts: introduction and application of the binary element diffusion modelling (BEDM) technique. Contributions to Mineralogy and Petrology 151: 58-70 Morgan DJ, Blake S, Rogers NW, DeVivo B, Rolandi G, Macdonald R, Hawkesworth CJ (2004) Time scales of crystal residence and magma chamber volume from modelling of diffusion profiles in phenocrysts: Vesuvius 1944. Earth and Planetary Science Letters 222: 933-946 Peslier AH, Luhr JF (2006) Hydrogen loss from olivines in mantle xenoliths from Simcoe (USA) and Mexico: Mafic alkalic magma ascent rates and water budget of the sub-continental lithosphere. Earth and Planetary Science Letters 242: 302-319 Reagan MK, Tepley FJ III, Gill JB, Wortel M, Garrison J (2006) Timescales of degassing and crystallization implied by 210Po–210Pb–226Ra disequilibria for andesitic lavas erupted from Arenal volcano. Journal of Volcanology and Geothermal Research 157: 135-146

272

Reid M (2003) Timescales of magma transfer and storage in the crust. In: Holland HD, Turekian KK (eds) Treatise on Geochemistry, Volume 3: The Crust, Elsevier, The Netherlands, pp 167-193 Reid MR, Coath CD (2000) In situ U-Pb ages of zircons from the Bishop Tuff: No evidence for long crystal residence times. Geology 28: 443-446 Rubin KH, van der Zander I, Smith MC, Bergmanis EC (2005) Minimum speed limit for ocean ridge magmatism from 210Pb–226Ra–230Th disequilibria. Nature 437: 534-538 Shaw CSJ, Heidelbach F, Dingwell DB (2006) The origin of reaction textures in mantle peridotite xenoliths from Sal Island, Cape Verde: the case for “metasomatism” by the host lava. Contributions to Mineralogy and Petrology 151: 681-697 Simon JI, Reid MR (2005) The pace of rhyolite differentiation and storage in an ‘archetypical’ silicic magma system, Long Valley, California. Earth and Planetary Science Letters 235: 123-140 Stracke A, Bourdon B, McKenzie D (2006) Melt extraction in the Earth’s mantle: Constraints from U–Th–Pa–Ra studies in oceanic basalts. Earth and Planetary Science Letters 244: 97-112 Turner S, Evans P, Hawkesworth C (2001) Ultrafast source-to-surface movement of melt at island arcs from 226Ra-230Th systematics. Science 292: 1363-1366 Turner S, George R, Jerram DA, Carpenter N, Hawkesworth C (2003) Case studies of plagioclase growth and residence times in island arc lavas from Tonga and the Lesser Antilles, and a model to reconcile discordant age information. Earth and Planetary Science Letters 214: 279-294 Turner S, Black S, Berlo K (2004) 210Pb–226Ra and 228Ra– 232Th systematics in young arc lavas: implications for magma degassing and ascent rates. Earth and Planetary Science Letters 227: 1-16 Vazquez Ja, Reid MR (2004) Probing the accumulation history of the voluminous Toba magma. Science 305: 991-994 Wolff JA, Balsley SD, Gregory RT (2002) Oxygen isotope disequilibrium between quartz and sanidine from the Bandelier Tuff, New Mexico, consistent with a short residence time of phenocrysts in rhyolitic magma. Journal of Volcanology and Geothermal Research 116: 119-135 Zellmer GF, Blake S, Vance D, Hawkesworth C, Turner S (1999) Plagioclase residence times at two island arc volcanoes (Kameni Islands, Santorini, and Soufriere, St. Vincent) determined by Sr diffusion systematics. Contributions to Mineralogy and Petrology 136: 345-357 Zellmer GF, Sparks RSJ, Hawkesworth CJ, Wiedenbeck M (2003) Magma emplacement and remobilization timescales beneath Montserrat: Insights from Sr and Ba zonation in plagioclase phenocrysts. Journal of Petrology 44: 1413-1431 !

A UGUST 2007