Abstract

New and/or enlarged datasets of U–Th disequilibrium model ages from secondary ionization mass spectrometry (SIMS) analyses of zircons in eight eruptive units from the area of Taupo volcano, New Zealand, highlight the behavioural contrasts of two closely adjacent, contemporaneous but independent magma chambers. One yielded closely similar crystal-poor (‘Oruanui-type’) rhyolites, sampled in three small precursor eruptions (Tihoi, ‘New plinian’, Okaia) from ∼45 to 30 ka, then the major 27 ka Oruanui eruption. Three of the four eruptions had vents within the modern Lake Taupo, whereas the fourth (‘New plinian’) was sourced ∼20 km NNE of the other vents, fed by lateral magma migration. Samples from all four eruptions share a common model-age peak at ∼95 ka of antecrystic zircons. However, three of the four differ in younger pre-eruptive model-age peaks that require their parental melt-dominant bodies to have been physically extracted independently from a common mush zone represented by the ∼95 ka peak. A sample from a fifth eruption (‘New phreatoplinian’, also at ∼45 ka) shares an older 80–100 ka peak but has numerous older grains and distinctly contrasting Sr-isotopic characteristics to the ‘Oruanui-type’ magmas. The 530 km3 Oruanui melt-dominant body was produced in at most ∼3000 years as shown by differences in zircon model-age spectra and average ages between it and the 30 ka Okaia eruption, despite their coincidence in vent locations. The second suite of eruptions at ∼47, 28 and 16 ka ejected moderately crystal-rich biotite rhyolites from a second source chamber, which vented over a 15 km wide area NE of Taupo (overlapping with Maroa volcano). This second chamber is inferred to have comparable horizontal dimensions to the vent spacing. The three biotite rhyolites show unimodal model-age spectra that peak at 30, 15–25 and 6 kyr prior to each eruption, respectively, and underwent single cycles of melt generation and eruption with no recycling of significantly older antecrysts or xenocrysts (< 1% equiline grains). Crystallization peaks defined by probability density function curves are not in phase between the two magma chambers and they had wholly independent thermal and chemical histories, despite their close geographical proximity. Post-Oruanui activity involved recycling of Oruanui-age zircons, but these crystals are xenocrystic, as the host melts show no lineage towards or mixing with the Oruanui compositions. Magma chambers at Taupo accumulated melt-dominant bodies as rapidly as > 5 m3/s (Oruanui) and effectively drained the mush of melt in doing so (Oruanui vs post-Oruanui activity), probably mediated by active rifting processes and tectonic disruption of the mush pile. Comparisons of ‘magma residence times’ and discussion of the growth histories of large silicic chambers represented by volcanic or plutonic rocks are self-limited by the uncertainties in the respective SIMS analyses. Growth times of Miocene and older plutons dated by SIMS U–Pb techniques are comparable with the 2 Myr lifetime of the whole Taupo Volcanic Zone, and the associated 1σ SIMS analytical uncertainties exceed the lifetime of a volcano such as Taupo. Subtle details that indicate the rapidity of magma accumulation and recycling of crystals in the young Taupo system cannot be discerned in most pre-300 ka silicic systems. Averaging of SIMS model-age data further obscures subtle details that would allow discrimination of newly crystallized versus recycled zircons. Discussions of volcano–plutonic relationships and accumulation rates for large silicic melt-dominant bodies cannot rely on age data in isolation, but require knowledge of the stratigraphic and compositional settings.

INTRODUCTION

Exceedingly large explosive silicic eruptions (sometimes for convenience or effect referred to as supereruptions i.e., discharging >1015 kg [∼450 km3] of magma or >∼1000 km3 of tephra), are inferred to occur globally on average about once every 105 years (Mason et al., 2004; Sparks et al., 2005; Miller & Wark, 2008). Many theoretical and field-focused studies on the development of the magma bodies that feed these vast eruptions have implied that, typically, such bodies take periods of the order of hundreds of thousands of years or longer to develop (Smith, 1979; Spera & Crisp, 1981; Shaw, 1985; Trial & Spera, 1990; Simon & Reid, 2005; White et al., 2006; Reid, 2008). There is, therefore, the distinct possibility that somewhere on Earth, at present, the magma chamber that will feed the next supereruption is actively in operation. What might the physical nature be of such a chamber, and are there factors about its thermal and magmatic evolution that would suggest such a large eruption is to be expected, or inevitable, or imminent?

To understand the magmatic conditions that culminate in supereruptions is a challenge from several viewpoints, among them to quantify the processes involved in the generation of silicic magmas, measure the time scales of accumulation and storage of the melt-dominant magma body, reconstruct the eruptions themselves and consider the nature and significance of post-climactic activity. Of the broad spectrum of silicic supereruption deposits, a significant proportion (the most abundant numerically, although not the largest in individual terms) are rhyolitic in bulk composition, and involve relatively crystal-poor magmas (up to ∼30% crystals) that accumulated as large bodies prior to eruptions (Hildreth, 1981; Miller & Wark, 2008). Do these melt-dominant zones (Hildreth & Wilson, 2007) or holding chambers (Wilson et al., 2006) represent the end product of prolonged steady-state or episodic growth, or can the accumulation of large volumes of crystal-poor melt from some source region (generally inferred to be a crystal-rich, melt-poor mush: e.g. Bachmann & Bergantz, 2004; Hildreth, 2004) occur on much shorter time scales than the long-term average? Do the processes that create large melt-rich bodies operate on such a scale in the crust that they constrain all magmatic activity in an area into a single focus (see Hildreth, 1981), or can other thermally and chemically independent magmatic systems also be active in close proximity?

Two lines of evidence are central to answering such questions on the timing and rates of accumulation of a large silicic magma body. The first is the record of eruptions that can be interpreted to represent precursor leaks and record any variations in mineralogical, geochemical or isotopic composition in the growing melt-dominant body. The second is the record preserved in the crystals themselves erupted in either the climactic eruption or its precursor leaks. This may be in the form of elemental or isotopic compositional diversity between crystals or zonation within crystals. In any particular case study, geochemical data can also address two other aspects: first, the sources of melt or crystal components, as shown by their chemistry and isotopic ratios; second, the time scales for growth of the magma body, as shown by the eruptive record and the ages of crystal components amenable to dating. In addition, the record of subsequent volcanism gives valuable clues to the nature of the post-climactic magma chamber, ranging from eruption of magma compositionally similar to that in the climactic event (e.g. Hildreth, 2004), through temporally evolving systems (e.g. Stix et al., 1988), to a complete resetting of the record as a result of readjustment of the magmatic system following caldera collapse (e.g. Sutton et al., 2000).

In interpreting the crystal and melt records, many complexities can arise, particularly when there is open-system behaviour and the information in any given crystal or crystal species is compromised by inheritance. For example, the isotopic ratios in feldspars in the Bishop Tuff were taken by earlier workers to indicate that the parental magma body resulted from prolonged growth over a period of >106 years (Halliday et al., 1989; Christensen & DePaolo, 1993; Christensen & Halliday, 1996; Davies & Halliday, 1998). However, this record runs counter to single-crystal model-age evidence that implies that zircons (inferred to be one of the earliest-crystallizing phases in the Bishop magma: Hildreth, 1979) crystallized for a time period up to a maximum of ∼160 kyr before the eruption (Reid & Coath, 2000; Simon & Reid, 2005; however, see also Crowley et al., 2007). Similarly, the Rb–Sr systematics in feldspars from the 1·61 Ma Otowi Member of the Bandelier Tuffs could be used to suggest a pattern of old model ages and prolonged incubation for the magma body; however, this was contradicted by evidence that showed that an accompanying diversity of Pb-isotopic ratios could not be due to ingrowth and hence that many of the feldspar crystals were inherited (Wolff et al., 1999; Wolff & Ramos, 2003). In some examples of rhyolitic supereruptions, growth of the melt-dominant zone is monitored by precursor eruptions (for Long Valley: Metz & Mahood, 1985, 1991; Metz & Bailey, 1993; for the Bandelier Tuffs: Stix et al., 1988; Spell & Kyle, 1989; Spell et al., 1990, 1996; Stix & Gorton, 1993). However, the wide temporal spacing of the precursor events (and associated uncertainties with which ages can be measured) places limits on the clarity with which the growth processes in the main chamber can be recognized.

Here we consider the magmatic and volcanic processes that operated in the area of Taupo volcano, New Zealand (Figs 1 and 2), in the build-up to, climax and aftermath of the 27 ka, ∼530 km3 Oruanui supereruption. The youthfulness of this sequence and an abundance of eruptive activity both before and after the climactic event allow for unusually fine-scale temporal resolution of the eruption sequences (Wilson, 1993; Wilson et al., 2009). Stratigraphic, petrological and isotopic studies at Taupo volcano have established four features of this eruptive sequence that are particularly relevant to disentangling the processes leading up to and following a very large rhyolitic eruption.

(1) Deposits containing pumices with closely similar mineralogical, chemical and isotopic characteristics to those analysed from the Oruanui eruption itself were erupted for about 40 kyr prior to the climactic event (Figs 3 and 4; ‘Oruanui-type magma’ of Sutton et al., 1995). In isolation, the compositional and isotopic data from these samples can be taken to suggest that a growing magma chamber was tapped by these eruptions. Such samples could, in principle, show the evolution and development of the melt-dominant zone beneath Taupo.

Fig. 1.

(a) Index map of New Zealand. (b) Index map of the central North Island, showing the rhyolitic (R) central segment of the Taupo Volcanic Zone (TVZ) and young eruptive centres (active in the last 340 ka).

Fig. 1.

(a) Index map of New Zealand. (b) Index map of the central North Island, showing the rhyolitic (R) central segment of the Taupo Volcanic Zone (TVZ) and young eruptive centres (active in the last 340 ka).

Fig. 2.

Detailed map of the Taupo–Maroa area. Surficial lava domes (all post-340 ka in age) and ‘active’ faults (i.e. with ruptures in the past 20 kyr, from the GNS Active Faults database: http://data.gns.cri.nz/af/) are marked. The dash–dot line marks the arbitrary boundary between Taupo (T) and Maroa (M) volcanoes as defined by Wilson et al. (1986). Vents for three of the fall deposits studied here are shown by open stars (see Fig. 5 for the field data used to derive these vent positions). Shapes of the domes from the ∼47 ka Ngangiho, 28 ka Trig 9471 and Rubbish Tip, and 16 ka Puketarata eruptions are shown to scale. Open ovoids NE and SW of Puketarata domes represent explosion craters from this event (from Brooker et al., 1993). Areas defined under Lake Taupo are the structural elements of the Oruanui caldera and subsequent collapse events, from Wilson (2001). Marginal ticks and numbers are those of the 10 km grid squares in the New Zealand metric map grid.

Fig. 2.

Detailed map of the Taupo–Maroa area. Surficial lava domes (all post-340 ka in age) and ‘active’ faults (i.e. with ruptures in the past 20 kyr, from the GNS Active Faults database: http://data.gns.cri.nz/af/) are marked. The dash–dot line marks the arbitrary boundary between Taupo (T) and Maroa (M) volcanoes as defined by Wilson et al. (1986). Vents for three of the fall deposits studied here are shown by open stars (see Fig. 5 for the field data used to derive these vent positions). Shapes of the domes from the ∼47 ka Ngangiho, 28 ka Trig 9471 and Rubbish Tip, and 16 ka Puketarata eruptions are shown to scale. Open ovoids NE and SW of Puketarata domes represent explosion craters from this event (from Brooker et al., 1993). Areas defined under Lake Taupo are the structural elements of the Oruanui caldera and subsequent collapse events, from Wilson (2001). Marginal ticks and numbers are those of the 10 km grid squares in the New Zealand metric map grid.

Fig. 3.

Plot of approximate eruption volumes (magma, after Wilson et al., 2009) vs time for eruptions in the Maroa–Taupo area for the past 60 kyr. Shaded bands in the post-Oruanui sequence are the magma groupings from Taupo eruptions (Sutton et al., 2000). Thicker lines with names represent the eruption units studied here, and other post-Oruanui units discussed in the text are labelled. Modified from Charlier et al. (2005) in the light of new field observations by C.J.N.W.

Fig. 3.

Plot of approximate eruption volumes (magma, after Wilson et al., 2009) vs time for eruptions in the Maroa–Taupo area for the past 60 kyr. Shaded bands in the post-Oruanui sequence are the magma groupings from Taupo eruptions (Sutton et al., 2000). Thicker lines with names represent the eruption units studied here, and other post-Oruanui units discussed in the text are labelled. Modified from Charlier et al. (2005) in the light of new field observations by C.J.N.W.

Fig. 4.

Plot of Rb/Sr vs 87Sr/86Sr to show the discrimination between the main consanguineous groups of eruptive units described in this paper. Modified from Charlier et al. (2005). (See Electronic Appendix 1 for data.)

Fig. 4.

Plot of Rb/Sr vs 87Sr/86Sr to show the discrimination between the main consanguineous groups of eruptive units described in this paper. Modified from Charlier et al. (2005). (See Electronic Appendix 1 for data.)

(2) Two examples of eruptive units with similar chemical characteristics (‘NE dome type’ of Sutton et al., 1995), and contrasting as such with the ‘Oruanui-type’, are found interfingered with Oruanui-type deposits and were erupted from vents close nearby (Figs 2 and 4). These are the Ngangiho dome and its unnamed, associated pyroclastic fall deposit, and the Rubbish Tip and Trig 9471 domes and their associated Poihipi fall deposit (Vucetich & Howorth, 1976; Wilson et al., 2009). A third small eruption (Puketarata: Lloyd, 1972; Brooker et al., 1993), which discharged magma with broadly similar characteristics to the two pre-Oruanui eruptions (Fig. 4), occurred at ∼16 ka from vents about 15 km north of the envelope enclosing the area of young (post-Oruanui) Taupo vents (Fig. 2: Wilson, 1993; Sutton et al., 2000). The stratigraphic interfingering of and vent siting for these three eruption deposits implies that magmas derived from a contrasting source could be erupted from vents positioned as close as ∼15 km to where the Oruanui system vented, and that the Oruanui magma chamber was not the only one active in the Taupo area prior to 27 ka, or subsequently. It should be noted that the fact that these eruptions are represented by lava domes at the present day has no bearing on any perceived contrast with the penecontemporaneous eruptions of the ‘Oruanui-type’ and post-Oruanui magmas that are now represented mostly by pyroclastic deposits. The closing phases of many of these latter eruptions are rich in dense juvenile material that probably accompanied contemporaneous dome extrusion (e.g. Wilson, 1993; Cole et al., 1998). Such domes were largely destroyed by engulfment in the Oruanui and 1·8 ka Taupo caldera-forming eruptions.

(3) The material erupted in the climactic Oruanui eruption shows evidence for significant mixing on several scales. Bulk pumice compositions fall along a consistent trend, but are not ordered with stratigraphic sequence, implying that any systematic zonation in the melt-dominant zone was disrupted or non-systematically tapped during the eruption (Wilson et al., 2006). Crystals extracted from single rhyolitic pumices of homogeneous appearance reflect mixed populations, not only with respect to major-element chemical compositions (plagioclase, orthopyroxene), but also in model ages (zircons), compositions and sources of melt inclusions (quartz), and Sr-isotopic characteristics (plagioclase). This is despite the same pumices having a relative uniformity of magmatic temperature as represented by Fe–Ti oxide compositions (Charlier & Zellmer, 2000; Charlier et al., 2005, 2008; Liu et al., 2006; Wilson et al., 2006).

(4) Eruptions after the Oruanui recommenced after a pause of only about 6000 years (Wilson, 1993). This break is comparable, for example, with the analytical uncertainties associated with radiometric determinations of the eruption ages of the Bishop Tuff and the oldest of the post-Bishop resurgent dome lavas in Long Valley caldera (Bailey, 1989). At Long Valley, the post-climactic crystal-poor lavas and sub-surface intrusions overlap in composition with the less evolved end of the late-stage crystal-rich pumices erupted in the Bishop Tuff (McConnell et al., 1995; Hildreth, 2004; Hildreth & Wilson, 2007). In sharp contrast, however, at Taupo the earliest post-Oruanui eruptive units were dacites with no clear compositional linkage to the Oruanui rhyolites (Fig. 4), and compositionally and isotopically distinct batches of magma have erupted up to as recently as 1800 years ago (Sutton et al., 2000). Also in this temporal sequence is the Puketarata rhyolite (Figs 2–4), the composition and existence of which implies that two different magma chambers remain active in the broader Taupo–Maroa area, probably to the present day.

To understand the relationships between these independent magmatic systems, as well as to re-examine the crystallization and accumulation history of the Oruanui magma body, we present here new or enlarged datasets of zircon model ages from eight eruption units, including the Oruanui itself (Tables 1 and 2). We couple the model-age information with consideration of eruptive compositions and timing to constrain the temporal and spatial scales over which the magma chambers were active.

Table 1:

Summary of characteristics of rhyolites studied

Sample Age Volume Magma Wt% Rb/Sr 87Sr/86Sr Wt % Major References 
 (ka ± 1 SD) (km3group SiO2 ratio ratio xtls xtl phases  
Puketarata 16 ± 1 0·14 74·5 0·86– 0·70521 16–20 Pl, Q, Bi, 1, 2 
     1·04   Hb, Hy, Ox  
Oruanui 27 ± 1 530 71·8– 0·44– 0·70551– 6–13 Pl, Q, Hb, 
    76·7 0·98 0·70568  Opx, Ox  
Rubbish Tip 28 ± 1 1·5 73·0– 0·60– 0·70527– 23 Pl, Q, Bi, 1, 2, 4, 5, 6 
dome    75·0 0·90 0·70530  Hb, Opx, Ox  
Okaia 30 ± 1 1·5 73·4– 0·62– 0·70556 Pl, Q, Opx, 1, 3 
    75·2 0·88   Hb, Ox  
‘New (45) ± 2 0·5 n.d. 0·60 0·70570 n.d. not studied 2, 5 
phreatoplinian’          
‘New plinian’ (45) ± 2 0·05 71·6 0·54– 0·70542 n.d. Pl, Q, Hb, 1, 2, 5 
     0·62   Hy, Ox  
Tihoi (45) ± 2 0·5 72·0– 0·51– 0·70554 10 Pl, Q, Opx, 1, 5 
    74·1 0·76   Hb, Ox  
Ngangiho (47) ± 2 0·15 74·8– 0·93– 0·70530 12 Pl, Q, Hb, Opx, 1, 2, 5 
    75·2 1·25   Bi, Cpx, Ox  
Sample Age Volume Magma Wt% Rb/Sr 87Sr/86Sr Wt % Major References 
 (ka ± 1 SD) (km3group SiO2 ratio ratio xtls xtl phases  
Puketarata 16 ± 1 0·14 74·5 0·86– 0·70521 16–20 Pl, Q, Bi, 1, 2 
     1·04   Hb, Hy, Ox  
Oruanui 27 ± 1 530 71·8– 0·44– 0·70551– 6–13 Pl, Q, Hb, 
    76·7 0·98 0·70568  Opx, Ox  
Rubbish Tip 28 ± 1 1·5 73·0– 0·60– 0·70527– 23 Pl, Q, Bi, 1, 2, 4, 5, 6 
dome    75·0 0·90 0·70530  Hb, Opx, Ox  
Okaia 30 ± 1 1·5 73·4– 0·62– 0·70556 Pl, Q, Opx, 1, 3 
    75·2 0·88   Hb, Ox  
‘New (45) ± 2 0·5 n.d. 0·60 0·70570 n.d. not studied 2, 5 
phreatoplinian’          
‘New plinian’ (45) ± 2 0·05 71·6 0·54– 0·70542 n.d. Pl, Q, Hb, 1, 2, 5 
     0·62   Hy, Ox  
Tihoi (45) ± 2 0·5 72·0– 0·51– 0·70554 10 Pl, Q, Opx, 1, 5 
    74·1 0·76   Hb, Ox  
Ngangiho (47) ± 2 0·15 74·8– 0·93– 0·70530 12 Pl, Q, Hb, Opx, 1, 2, 5 
    75·2 1·25   Bi, Cpx, Ox  

Ages given and uncertainties are discussed in text. Volumes are magma; for Rubbish Tip dome, the value includes the co-eruptive Poihipi Tephra and Trig 9471 dome. Numbers in the column ‘Magma group’ refer to the broad magma groupings of Sutton et al. (1995) and discussed in the text: 1, ‘NE-dome type’ or similar; 2, ‘Oruanui-type’; 3, other. References: 1, Sutton (1995); 2, newly determined for this paper; 3, Wilson et al. (2006); 4, Lowe et al. (2008); 5, C. J. N. Wilson, unpublished field observations; 6, SiO2 values are for samples from both the dome and the juvenile component in the Poihipi Tephra. n.d., not determined.

Table 2:

Summary of zircon samples used in this paper

Sample Nature Eruption unit Grains analysed No. of analyses with finite ages 
Puketarata (16 ka) 
P1835 Multiple pumices Fall deposit 50 49 
Oruanui eruption products (27 ka) 
P1520 Single pumice Unit 10 ignimbrite 50 44 
P1373 Single pumice Unit 10 ignimbrite 54 50 
P1634 Multiple pumices Unit 1 fall deposit 103 100 
Rubbish Tip dome (28 ka) 
R559 Glassy lava Dome carapace 53 53 
Okaia fall deposit (30 ka) 
P1195 Multiple pumices Plinian fall deposit 60 55 
‘New phreatoplinian’ fall deposit (∼45 ka) 
P1836 Bulk sample Phreatoplinian fall deposit 54 40 
‘New plinian’ fall deposit (∼45 ka) 
P1834 Multiple pumices Plinian fall deposit 54 52 
Tihoi fall deposit (∼45 ka) 
P433 Multiple pumices Plinian fall deposit 69 69 
Ngangiho dome (∼47 ka) 
R727 Glassy lava Dome carapace 51 51 
Sample Nature Eruption unit Grains analysed No. of analyses with finite ages 
Puketarata (16 ka) 
P1835 Multiple pumices Fall deposit 50 49 
Oruanui eruption products (27 ka) 
P1520 Single pumice Unit 10 ignimbrite 50 44 
P1373 Single pumice Unit 10 ignimbrite 54 50 
P1634 Multiple pumices Unit 1 fall deposit 103 100 
Rubbish Tip dome (28 ka) 
R559 Glassy lava Dome carapace 53 53 
Okaia fall deposit (30 ka) 
P1195 Multiple pumices Plinian fall deposit 60 55 
‘New phreatoplinian’ fall deposit (∼45 ka) 
P1836 Bulk sample Phreatoplinian fall deposit 54 40 
‘New plinian’ fall deposit (∼45 ka) 
P1834 Multiple pumices Plinian fall deposit 54 52 
Tihoi fall deposit (∼45 ka) 
P433 Multiple pumices Plinian fall deposit 69 69 
Ngangiho dome (∼47 ka) 
R727 Glassy lava Dome carapace 51 51 

Details of samples and eruption units are given in the text. Crystals with ‘finite ages’ denote those that plotted at least 1 SD below the equiline.

In our descriptions here, we use the term ‘magma chamber’ in the sense of Hildreth (2004) and Hildreth & Wilson (2007, fig. 17) to refer to the total entity that contains at least some melt, or may be wholly crystallized but is still in thermal and spatial continuity with melt-bearing material. Within the chamber, the melt-dominant zone (or holding chamber) represents the material of which part or all is erupted in any single event. In many cases (notably the Bishop Tuff), the melt-dominant zone has gradational characteristics in chemical composition, crystal content or volatile content downwards into the crystal-rich, melt-subordinate mush zones forming basal zones to the chamber. However, in the examples studied at Taupo, the absence of systematic gradients in composition or crystal content implies that the melt-dominant zones appear to have sharply defined borders against any walls, roof or floor (Blake et al., 1992; Sutton et al., 1995, 2000; Wilson et al., 2006). For the purposes of describing the magmas erupted at Taupo, we use the terms crystal-poor, crystal-moderate and crystal-rich to refer to materials with <5, 5–15 and >15 wt % crystals, respectively.

TAUPO VOLCANO

Eruptive history

Taupo volcano is situated at the southern end of the central Taupo Volcanic Zone (TVZ) in New Zealand (Wilson et al., 1995; Fig. 1) and is one of two highly active rhyolitic centres in the present-day TVZ (Wilson et al., 2009). Taupo has been active since eruption of the voluminous Whakamaru group of ignimbrites and associated Rangitawa Tephra at 340–320 ka (Wilson et al., 1986; Houghton et al., 1995; Pillans et al., 1996; Brown et al., 1998), but its history until about 65 ka is known only in outline because of poor age control and burial by younger deposits. From ∼300 to ∼100 ka, activity at Taupo was accompanied by numerous explosive and effusive eruptions of a wide diversity of compositions from the Maroa volcano to the north (Wilson et al., 1986; Leonard, 2003). From about 100 ka onwards, activity at Maroa declined in intensity and the main focus of activity shifted south into Taupo. From 61 ka onwards, as defined by the regionally extensive marker horizon of the Rotoehu Tephra from Okataina volcano (Nairn, 1972; Wilson et al., 2007), voluminous explosive activity has occurred at Taupo with lesser volumes of lava (most of which is inferred to have been engulfed in 27 ka and 1·8 ka caldera collapses: e.g. Cole et al., 1998). In contrast, minor activity at Maroa over the same time period has almost always been accompanied by lava extrusion. Both Taupo and Maroa volcanoes have erupted a variety of rhyolites and trivial volumes of less evolved compositions (Sutton, 1995; Sutton et al., 1995; Leonard, 2003). Up to the time of the 27 ka Oruanui eruption, the diverse magma compositions in the Taupo area were erupted in alternating succession from vents that were separated geographically over an area some 20 km across (Figs 2–4). Since the Oruanui eruption, 28 eruptions have occurred from Taupo (Wilson, 1993), but only one (Puketarata) from the southernmost part of Maroa.

Eruptive units studied

Eight eruptive units form the focus of this study (Table 1; Fig. 3), four of which were also considered by Charlier et al. (2005). The published datasets in that earlier paper have been supplemented here by new secondary ionization mass spectrometry (SIMS) analyses. Six of the eight eruptions occurred during a ∼20 kyr period prior to the climactic Oruanui eruption, four of them in relatively close succession. Below we summarize these eruptions, from oldest to youngest, and the samples studied.

The Ngangiho eruption is largely represented by a small joined pair of domical lava extrusions (Fig. 2), with intact moderately fresh pumiceous glassy carapaces. The stratigraphic position of the Ngangiho extrusions is established from a correlative fall deposit, identified by its coarse grain size and similar mineralogy and chemistry, which is exposed in a nearby road-cut exposure, underlying the Tihoi fall deposit with an intervening palaeosol. The fall deposit is not very widespread or well preserved, and the eruption is inferred to be of minor volume (no more than ∼0·15 km3). From the degree of palaeosol development below the Tihoi deposit, and stratigraphic relationships with respect to the 61 ka Rotoehu fall deposit exposed at the same location, an age of ∼47 ka is adopted here. A sample (R727) was collected from the glassy pumiceous carapace breccia from the eastern dome at U17/761867 (grid references are to the nearest 100 m in the New Zealand metric grid).

The Tihoi fall deposit was erupted from a source now concealed beneath northern Lake Taupo (Figs 2 and 5a–c). It is a bedded, normally graded pumice fall deposit, of wide (plinian) dispersal but modest volume (∼0·5 km3, magma: Vucetich & Howorth, 1976; C. J. N. Wilson, unpublished data). The age of this unit is not accurately established, but in most of its exposures around the Taupo area this deposit is separated from the 61 ka Rotoehu fall deposit by a substantial palaeosol that is comparable in thickness with that which separates the Tihoi from the 30 ka Okaia fall deposit. An eruption age of ∼45 ka is thus inferred, as adopted by Charlier et al. (2005). A new ion probe mount was made up using zircons from the same sample (P433) as used by Charlier et al. (2005). The chemistry, mineralogy and bulk isotopic characteristics of pumices from this deposit link it in with the Oruanui-type magma of Sutton et al. (1995).

Fig. 5.

Series of maps showing data values, isopachs (lines connecting equal thicknesses of deposited material) and isopleths (lines connecting fragments of the same size) for the ∼45 ka Tihoi, ∼45 ka ‘New plinian’ and 30 ka Okaia fall deposits, used to infer their vent localities. Values for data points and contours are in millimetres; isopleths represent the average lengths of the five largest fragments found at each location. (a–c) Tihoi, overall deposit; (d–f) ‘New plinian’, overall deposit (abs = absent; tr = trace); (g–l) Okaia. Panels (g)–(i) are for a lower, uniformly to normally graded bed in the Okaia that has a distinctive appearance, a markedly bilobate distribution and contains lithic fragments that are dominated by old lithologies. Panels (j)–(l) are for the upper, volumetrically dominant part of the Okaia deposit, which is coarser and has lithic fragments that include a high to dominant proportion of grey, glassy, non-vesicular material that is inferred to be juvenile. Vent sites inferred from these field data are marked with open star symbols. Shaded areas defined under Lake Taupo represent the structural elements of the Oruanui caldera and subsequent collapse events, from Wilson (2001). Marginal ticks and numbers are those of the 10 km grid squares in the New Zealand metric map grid.

Fig. 5.

Series of maps showing data values, isopachs (lines connecting equal thicknesses of deposited material) and isopleths (lines connecting fragments of the same size) for the ∼45 ka Tihoi, ∼45 ka ‘New plinian’ and 30 ka Okaia fall deposits, used to infer their vent localities. Values for data points and contours are in millimetres; isopleths represent the average lengths of the five largest fragments found at each location. (a–c) Tihoi, overall deposit; (d–f) ‘New plinian’, overall deposit (abs = absent; tr = trace); (g–l) Okaia. Panels (g)–(i) are for a lower, uniformly to normally graded bed in the Okaia that has a distinctive appearance, a markedly bilobate distribution and contains lithic fragments that are dominated by old lithologies. Panels (j)–(l) are for the upper, volumetrically dominant part of the Okaia deposit, which is coarser and has lithic fragments that include a high to dominant proportion of grey, glassy, non-vesicular material that is inferred to be juvenile. Vent sites inferred from these field data are marked with open star symbols. Shaded areas defined under Lake Taupo represent the structural elements of the Oruanui caldera and subsequent collapse events, from Wilson (2001). Marginal ticks and numbers are those of the 10 km grid squares in the New Zealand metric map grid.

The Tihoi eruption was succeeded by three eruptions, two magmatic and one hydrothermal, that followed so closely that there is only trivial development of palaeosols at most between the relevant deposits where the contacts are best preserved. As a result, the nominal ages are given here as the same. The first of these is a small pumice fall deposit of plinian dispersal that is intercalated with beds of clay-rich material from large-scale hydrothermal eruptions from a different source. This deposit was not recognized by Vucetich & Howorth (1976), and we use the informal term ‘New plinian’ for it here. Although of similar chemistry and mineralogy to the Oruanui and other eruptive units of the ‘Oruanui-type magma’ (Sutton, 1995; Sutton et al., 1995), its bulk 87Sr/86Sr ratio is somewhat lower (Fig. 4; Table 1). In addition, isopach and isopleth data from this fall deposit (Fig. 5d–f) demonstrate that it was erupted from a source that was well north of Lake Taupo. The vent site is inferred to be adjacent to domes of contrasting chemistry forming part of the SW corner of the central Maroa dome complex (Leonard, 2003). The eruption is of small volume, estimated at 0·05 km3 of magma. A sample of multiple pumices (P1834) from a near-source exposure at T17/690962 was taken and crushed for zircon extraction, whereas the whole-rock U–Th values and 87Sr/86Sr ratios were measured on a multiple-pumice sample (P1300) from an exposure at U17/757883.

Another short period of time, with some erosion but very little weathering, then occurred before a phreatomagmatic fall deposit was erupted that is widely distributed (of phreatoplinian dispersal: Self & Sparks, 1978) but very poorly preserved. This deposit also was not recognized by Vucetich & Howorth (1976), and we use the informal term ‘New phreatoplinian’ for it here. The volume is uncertain, but is inferred to be roughly 0·5 km3 of magma from the distribution mapped by C. J. N. Wilson (unpublished data). No clear indication is available for the vent position, but thickness variations suggest a source in or close to northeastern Lake Taupo. A bulk sample (P1836) was taken from the deposit at T17/612837 and wet sieved: zircons were extracted from the 125 and 63 μm sieve fractions, whereas the sparse pumice lapilli (to 1–2 cm across) were used to obtain whole-rock values for the U/Th and Sr isotopic data.

The Okaia eruption followed the ‘New phreatoplinian’ eruption after at least one other substantial eruption from an uncertain source (the Tahuna Tephra of Smith & Shane, 2002, not dealt with here) and a prolonged time period represented by soil formation. The Okaia is a bedded plinian fall deposit of mixed magmatic and phreatomagmatic styles, with a total inferred magma volume of 1·5 km3. A source below the northern part of Lake Taupo is indicated from isopach and isopleth data (Fig. 5g–l). An age of 30 ka is inferred on the basis of a single radiocarbon age from below the distal fall deposit in the Auckland area (Sandiford et al., 2002; Lowe et al., 2008). The characteristics of the Okaia link it strongly to the earlier Tihoi and (although to a lesser extent) ‘New plinian’ deposits as part of the ‘Oruanui-type magma’ of Sutton et al. (1995). A new ion probe mount was made up using zircons from the same crystal separate (P1195) as used by Charlier et al. (2005).

Overlying the Okaia deposits after a period of weathering and soil formation is a widespread, fine-grained phreatoplinian fall deposit known as the Poihipi Tephra (Vucetich & Howorth, 1976; Self & Sparks, 1978). Radiocarbon dates on organic material enclosing the distal fall deposit in the Auckland area have been used to derive a calendar age estimate of 28 ka (Sandiford et al., 2001; Lowe et al., 2008). Isopach mapping (C. J. N. Wilson, unpublished data) and geochemical data (Sutton, 1995; Sutton et al., 1995) link the source of the Poihipi Tephra to be closely adjacent to an area NE of Lake Taupo now marked by two lava extrusions, Trig 9471 dome and Rubbish Tip dome (RTD). The chemistry and mineralogy of these domes and the juvenile component in the Poihipi Tephra are closely similar (Sutton, 1995). However, the dense rock equivalent (DRE) volume of the fall deposits (1·4 km3) is at least an order of magnitude greater than those of the domes (∼0·1 km3; Wilson et al., 2009). Fresh pumiceous dome carapace material from the RTD (sample R559) was used for zircon extraction by Charlier et al. (2005) and a new mount was made up using zircons from the same crystal separate.

The Oruanui eruption was the largest eruption at Taupo, or in the TVZ as a whole for the last ∼320 kyr. The age for this unit is based on the radiocarbon determinations by Wilson et al. (1988), with various calibrations summarized by Lowe et al. (2008) yielding mean calendar-age estimates between 26·5 and 27 ka; the latter value is used here for consistency with the ages adopted for the RTD and Okaia. The physical volcanology, bulk geochemistry and zircon model-age systematics of the eruption products have been detailed in earlier studies (Wilson, 2001; Charlier et al., 2005; Wilson et al., 2006). The zircon model-age spectra reported by Charlier et al. (2005) came from three large pumices collected from the latest parts of the Oruanui ignimbrite. Using the zircon separates from these pumices, several multi-grain samples were analysed by thermal ionization mass spectrometry (TIMS) (Charlier & Zellmer, 2000), and material from two pumices (P1373 and P1520) was analysed by SIMS. Because these pumices were all from the later parts of the eruption, for this study we sampled a suite of pumices (P1634) from the basal fall deposit (fall unit 1 of Wilson, 2001) at a locality at U18/737784 to see if the zircon age spectrum in the earliest-erupted material differed in any way.

Puketarata (Fig. 2) is the youngest eruption from Maroa volcano, and generated a moderately widespread fall deposit, a proximal tuff ring and two small lava domes, totalling 0·14 km3 of magma, of which ∼90% was erupted as pyroclastic material (Lloyd, 1972; Brooker et al., 1993). The eruption occurred at ∼16 ka; it is not directly dated, but is bracketed by older and younger deposits from other sources that are accurately radiocarbon dated (Lowe, 1988; Lowe et al., 2008). Multiple dense pumice clasts (P1835) were collected from fall unit B1 of Brooker et al. (1993) at a road-cut exposure at U17/752903, and a whole-rock value was obtained from a sample of the pumiceous dome carapace (P1773) collected at U17/759902. We studied Puketarata because in its mineralogical characteristics (crystal-rich and with biotite in the ferromagnesian fraction) it has affinities with the older nearby Ngangiho and more distant Rubbish Tip domes (Fig. 2), although its 87Sr/86Sr ratio is slightly lower (Fig. 4; Table 1). In addition, Puketarata compositions present a marked contrast in all respects to the bracketing eruption units from Taupo, the earlier dacites (∼21·5–17 ka) and their linked derivatives, the rhyolites of Subgroup 1 (11·8–10 ka: Wilson, 1993; Sutton et al., 2000; Figs 3 and 4).

Eruption age uncertainties

Uncertainties on these eruption age estimates need to be quantified in order to compare with the zircon model-age spectra. However, these uncertainties reflect the influence of a combination of factors, which in themselves are poorly constrained. For deposits where radiocarbon ages have been determined [Okaia, Poihipi (for RTD), Oruanui], these include errors on relevant radiocarbon age determinations, and errors in the calibration curves used for conversion to calendar ages (e.g. Lowe et al., 2008). For those deposits where direct age data are lacking, the uncertainties are kept in check by knowledge of the stratigraphic relationships between the relevant undated deposit and dated units above and below. Thus for the Puketarata eruption, its age is constrained by bracketing units with multiple age determinations indicating eruption ages of 13·8 and 17 calibrated ka (Lowe, 1988; Lowe et al., 2008). For the Ngangiho, Tihoi, ‘New plinian’ and ‘New phreatoplinian’ eruptions, their relative ages and breaks between them are established from field observations but not their absolute ages. The estimate of 45 ka for the age of the Tihoi eruption is based on degrees of soil development (see previous section), and observations of the presence of numerous other fall deposits and soils between it and overlying units directly dated at 37 ka (Unit E of Jurado-Chichay & Walker, 2000) and the underlying 61 ka Rotoehu fall deposit (Wilson et al., 2007). From these constraints we adopt 1 SD errors of ±1 kyr for the Okaia, Poihipi (for the RTD), Oruanui and Puketarata eruption ages (see Lowe et al., 2008), and ±2 kyr for the older deposits. All reasonable ways of estimating the eruption ages are associated with uncertainties that are smaller than those for the SIMS single-grain model ages and larger than those for the TIMS multiple-grain analyses for zircons for the respective units.

ANALYTICAL TECHNIQUES

Samples used for zircon separation are listed in Table 1. Zircon extraction and cleaning techniques for the new samples were the same as described by Charlier et al. (2005). For SIMS U–Th analyses by sensitive high-resolution ion microprobe (SHRIMP) using the USGS–Stanford SHRIMP-RG instrument, techniques used were based on those of Lowenstern et al. (2000) and Charlier et al. (2005). Zircons were mounted in epoxy resin, polished, photographed in reflected light, and imaged by cathodoluminescence (CL). The mount was acid rinsed, coated with 100 nm of Au and left in the SHRIMP-RG sample chamber overnight to reach full vacuum. Using a 34 nA 16O or 16O2- primary ion beam, a 50 μm × 50 μm square region was rastered for 2 min to remove the Au coat and any surface contamination. A flat-floored elliptical pit 2 μm × 25 μm × 37 μm was then excavated into the zircon. This liberated ∼4–6 ng of sample that was sent as positive secondary ions to the mass spectrometer. Data were collected in 10 scans per point for 90Zr216O, 230Th16O, 232Th, 232Th16O, 238U and 238U16O. Dwell times ranged from 2 to 40 s for each peak. In particular, to allow for the generally low U contents of the zircons, 230Th16O and the background (at mass 246·16) were measured for 40 s each. Additionally, a 4 s measurement at mass 244 (‘ThC’) was used as a check for the beam impinging on the epoxy resin in the mount. If the count rate for mass 244 significantly exceeded the background count rate the analysis was discarded, as inevitably the resulting meaningless data lay in the sector of Th-excess [i.e. with a (238U/230Th) value of >1] on the equiline diagram (see Schmitt et al., 2006, appendix 1).

A U–Th fractionation factor was empirically determined through the repeated analysis of several zircon standards run from the same mounts as those for the unknowns. These included, at various stages: (1) MAD (concentration standard), a 555 Ma gem-quality zircon from Madagascar with U = 4196 ppm and Th = 1166 ppm (F. K. Mazdab, personal communication); (2) CZ3 (concentration standard: 550 ppm U); (3) R33; (4) VP10; (5) AS57 (see Ireland & Williams, 2003; Charlier et al., 2005; Lowenstern et al., 2006). Given their antiquity, 238U and 230Th activities in these zircons are at secular equilibrium and therefore, after the application of a U–Th fractionation factor, the calculated activity of (230Th/238U) should equal unity. This was determined on a mount-by-mount basis using the measured 230Th16O+/238U16O+ ratios. To achieve a weighted mean (230Th/238U) of unity for the standards, fractionation factors varied between 1· 04 and 1·14 over the analysis period, the reciprocals of which were applied in the calculation of the (238U/232Th) of the Taupo zircons for the relevant day of their analysis. Repeated analysis of the standard allowed us to arrive at a conservative best-estimate 1 SD error of ±3% on the fractionation factor, which translates into a corresponding ±3% error on the (238U/232Th) values. This uncertainty is substantially larger than that calculable from the count statistics alone and hence is inferred to give a more robust limit to total errors in the (238U/232Th) values (see Charlier et al., 2005). The uncertainty on the (230Th/232Th), however, was derived from the counting statistics.

Whole-rock values of the (238U/232Th) and (230Th/232Th) ratios for samples from Tihoi, Okaia, Rubbish Tip Dome and samples from the late-erupted Oruanui pumices had already been obtained by TIMS techniques (reported by Charlier et al., 2005), whereas new whole-rock determinations for the Ngangiho, ‘New plinian’, ‘New phreatoplinian’ and Puketarata samples were obtained by a different approach, as follows. The (230Th/232Th) was determined on unspiked dissolutions by multicollector inductively coupled plasma mass spectrometry (MC-ICP-MS) using a Thermo-Finnigan Neptune instrument equipped with a retarding potential quadropole, whereas the (238U/232Th) was calculated from U and Th concentration analyses by quadrupole (Q)-ICP-MS; both instruments were at The Open University. Repeat analysis of one whole-rock powder (P1209) using both TIMS, and MC- and Q-ICP-MS approaches yielded data that were essentially identical within analytical uncertainty.

230Th–238U isochron ages were calculated as two-point model ages by referencing each of the zircon analyses in turn to the respective whole-rock (WR) analysis on the equiline diagram (e.g. Dickin, 1995). In this study, the zircons are considered to have the same initial (230Th/232Th) as the host rocks (i.e. they plot in a horizontal array on the equiline diagram at the time of their crystallization), and because the (238U/232Th) values of the zircons are so high, the calculated slope of the zircon–WR regression lines (and thus the age), are relatively insensitive to small variations in the initial (230Th/232Th). Samples that are older than five times the half-life of 230Th (i.e. ∼350 kyr) plot within error of the equiline at a (230Th/232Th) value that reflects the (238U/232Th) value. The only age information that can be obtained from these samples is that they are older than 350 ka.

Sr isotope ratio determinations were carried out on new samples separated using the techniques detailed by Charlier et al. (2006). The Neptune MC-ICP-MS instrument was operated in static collection mode using a normalizing value of 86Sr/88Sr = 0·1194 and the exponential law relationship. Repeated analysis of NBS 987 on the day of analysis gave 87Sr/86Sr = 0·710239 ± 13 (18·8 ppm) 2 SD (n = 9). To allow direct comparison with data previously presented by Wilson et al. (2006), the new Sr isotope data were normalized to 87Sr/86Sr = 0·710250 for NBS 987.

RESULTS

Table 2 summarizes the numbers of zircon model ages obtained from each unit. Raw data are presented in Electronic Appendices 3–10, which may be downloaded from the Journal of Petrology website at http://www.petrology.oxfordjournals.org. To compare the SIMS model-age data with average ages (where previously obtained by TIMS), and with the peaks in the probability density function (PDF) curves defined from Isoplot (Ludwig, 2003), we calculated concentration-weighted mean values for the SIMS datasets (Table 3).

Table 3:

Summary of concentration-weighted mean model ages from zircons in the samples used for this study

Material (238U/232Th) Error (230Th/232Th) Error Slope Error slope Age (ka) Error + age Error − age 
Puketarata (Fig. 14) 
zircons 2·89 0·09 1·46 0·05      
whole-rock 0·72 0·01 0·75 0·00 0·33 0·03 43 
All Oruanui (Fig. 12) 
zircons 3·33 0·10 2·02 0·14      
whole-rock 0·72 0·01 0·74 0·01 0·49 0·06 73 13 12 
Rubbish Tip dome (Fig. 10) 
zircons 2·23 0·07 1·38 0·06      
whole-rock 0·73 0·00 0·75 0·00 0·42 0·05 59 
Okaia (Fig. 11) 
zircons 2·80 0·08 1·87 0·12      
whole-rock 0·74 0·01 0·75 0·01 0·54 0·06 85 16 14 
‘New phreatoplinian’ (Fig. 9) 
zircons 2·95 0·09 2·49 0·07      
whole-rock 0·71 0·01 0·71 0·00 0·79 0·05 170 26 21 
‘New plinian’ (Fig. 8) 
zircons 2·51 0·08 1·79 0·06      
whole-rock 0·72 0·01 0·74 0·01 0·59 0·04 96 11 10 
Tihoi (Fig. 7) 
zircons 2·41 0·07 1·66 0·11      
whole-rock 0·65 0·01 0·67 0·01 0·56 0·07 90 18 16 
Ngangiho (Fig. 6) 
zircons 2·09 0·06 1·47 0·05      
whole-rock 0·73 0·01 0·74 0·00 0·54 0·05 84 11 10 
Material (238U/232Th) Error (230Th/232Th) Error Slope Error slope Age (ka) Error + age Error − age 
Puketarata (Fig. 14) 
zircons 2·89 0·09 1·46 0·05      
whole-rock 0·72 0·01 0·75 0·00 0·33 0·03 43 
All Oruanui (Fig. 12) 
zircons 3·33 0·10 2·02 0·14      
whole-rock 0·72 0·01 0·74 0·01 0·49 0·06 73 13 12 
Rubbish Tip dome (Fig. 10) 
zircons 2·23 0·07 1·38 0·06      
whole-rock 0·73 0·00 0·75 0·00 0·42 0·05 59 
Okaia (Fig. 11) 
zircons 2·80 0·08 1·87 0·12      
whole-rock 0·74 0·01 0·75 0·01 0·54 0·06 85 16 14 
‘New phreatoplinian’ (Fig. 9) 
zircons 2·95 0·09 2·49 0·07      
whole-rock 0·71 0·01 0·71 0·00 0·79 0·05 170 26 21 
‘New plinian’ (Fig. 8) 
zircons 2·51 0·08 1·79 0·06      
whole-rock 0·72 0·01 0·74 0·01 0·59 0·04 96 11 10 
Tihoi (Fig. 7) 
zircons 2·41 0·07 1·66 0·11      
whole-rock 0·65 0·01 0·67 0·01 0·56 0·07 90 18 16 
Ngangiho (Fig. 6) 
zircons 2·09 0·06 1·47 0·05      
whole-rock 0·73 0·01 0·74 0·00 0·54 0·05 84 11 10 

Errors on the slopes are calculated as 1 SD a priori uncertainties. Uncertainties on the activity ratios are 3% for (238U/232Th) and 1 SD internal error (within-run) for (230Th/232Th) for the single data points as well as the calculated weighted means and calculated slopes. Uncertainties on the ages are also 1 SD. It should be noted that rounding of the numbers has resulted in some mismatch between the isochron slopes and their errors and their corresponding age values.

Ngangiho dome

Fifty-one grains were analysed from sample R727; all yielded finite ages in the U–Th system. The full dataset is given in Electronic Appendix 3. The model-age spectrum (Fig. 6) is unimodal, with a peak in the PDF curve at 79 ka. The SIMS data yield a concentration-weighted isotopic mean age of 84 + 11/–10 ka.

Fig. 6.

Zircon model-age data from the Ngangiho dome sample R727. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. Reference isochrons indicate the eruption age (∼47 ka), 100, 200 and 300 ka. Error ellipses represent 1 SD analytical uncertainties on (230Th/232Th) and a standard 3% error on (238U/232Th) (see Charlier et al., 2005). An isochron slope and age was determined by referencing this isotopic average data point to the whole-rock values (Electronic Appendix 2) to generate a two-point model age. (b) Probability density function (PDF) curve (from Isoplot: Ludwig, 2003) and histogram based on isochron slopes derived from two-point whole-rock–zircon SIMS determinations. The PDF line is based on the two-point isochron slopes rather than the ages determined from them, as the slope uncertainty is symmetrical with respect to the slope value whereas the absolute age is not. A concentration-weighted isotopic mean age (84 ka: see Table 3) was determined by weighting all the SIMS (230Th/232Th) and (238U/232Th) values according to the U and Th concentrations of each spot analysis (to the nearest 100 ppm). The dashed line represents the value and the grey band the width of the ±1 SD errors on the concentration-weighted mean. Analytical data and model ages are given in Electronic Appendix 3.

Fig. 6.

Zircon model-age data from the Ngangiho dome sample R727. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. Reference isochrons indicate the eruption age (∼47 ka), 100, 200 and 300 ka. Error ellipses represent 1 SD analytical uncertainties on (230Th/232Th) and a standard 3% error on (238U/232Th) (see Charlier et al., 2005). An isochron slope and age was determined by referencing this isotopic average data point to the whole-rock values (Electronic Appendix 2) to generate a two-point model age. (b) Probability density function (PDF) curve (from Isoplot: Ludwig, 2003) and histogram based on isochron slopes derived from two-point whole-rock–zircon SIMS determinations. The PDF line is based on the two-point isochron slopes rather than the ages determined from them, as the slope uncertainty is symmetrical with respect to the slope value whereas the absolute age is not. A concentration-weighted isotopic mean age (84 ka: see Table 3) was determined by weighting all the SIMS (230Th/232Th) and (238U/232Th) values according to the U and Th concentrations of each spot analysis (to the nearest 100 ppm). The dashed line represents the value and the grey band the width of the ±1 SD errors on the concentration-weighted mean. Analytical data and model ages are given in Electronic Appendix 3.

Tihoi fall deposit

The earlier compilation of Charlier et al. (2005) reported 40 grains with finite ages and an additional one that lay on the equiline within error. We have supplemented this dataset with 29 new analyses, all of which yielded finite ages, and the two datasets are plotted in Fig. 7. The enlarged dataset (Electronic Appendix 4) has a pronounced PDF peak at 96 ka, and a secondary peak at 64 ka, identical within error to the corresponding values (104 and 61 ka, respectively) reported from the smaller dataset by Charlier et al. (2005). Values of average zircon ages derived from TIMS analysis of multiple-grain separates by Charlier et al. (2005) are between 63·5 ± 1· 4 ka and 75·9 ± 1·5 ka, whereas the concentration-weighted mean SIMS age from the enlarged dataset is 90 + 18/–16 ka.

Fig. 7.

Data from the Tihoi eruption sample P433. These combine our new data (29 points) with 40 points from Charlier et al. (2005). Model ages and analytical data are in Electronic Appendix 4. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (90 ka) from SIMS data. Other details as in Fig. 6.

Fig. 7.

Data from the Tihoi eruption sample P433. These combine our new data (29 points) with 40 points from Charlier et al. (2005). Model ages and analytical data are in Electronic Appendix 4. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (90 ka) from SIMS data. Other details as in Fig. 6.

‘New plinian’ fall deposit

Fifty-four grains were analysed, of which 52 returned finite ages. The overall model-age spectrum (Fig. 8; Electronic Appendix 5) is closely similar to that from the Tihoi eruption, with major and minor PDF peaks at 94 and 67 ka, respectively, and a concentration-weighted mean of 96 + 11/–10 ka. The bulk chemistry and mineralogy of pumices from this eruption are also similar to those of the ‘Oruanui-type magma’ (Tihoi, Okaia, Oruanui) reported by Sutton (1995) and Sutton et al. (1995). The 87Sr/86Sr ratio, however, for multiple pumices from this deposit (Table 1, Fig. 4) that we measured from sample P1300 is somewhat lower than the values reported from the earlier work.

Fig. 8.

Data from the ‘New plinian’ eruption sample P1834. Model ages and analytical data are in Electronic Appendix 5. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (96 ka) from SIMS data on grains that returned finite ages. The approximate eruption age (∼45 ka; see text) is marked. Other details as in Fig. 6.

Fig. 8.

Data from the ‘New plinian’ eruption sample P1834. Model ages and analytical data are in Electronic Appendix 5. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (96 ka) from SIMS data on grains that returned finite ages. The approximate eruption age (∼45 ka; see text) is marked. Other details as in Fig. 6.

‘New phreatoplinian’ fall deposit

Fifty-four grains were analysed, of which 40 yielded finite ages (Electronic Appendix 6). The resulting model-age spectrum (Fig. 9) differs from that of any of the other units reported here, with a pair of peaks at 83 and 101 ka and a substantial tail of older grains that generate a PDF peak at 295 ka. The concentration-weighted mean is between the two main peaks, at 170 + 26/–21 ka. None of the other samples analysed for this study yielded as high a proportion of ‘clean’ grains (i.e. without any obvious signs of older cores from CL imagery) with old or unresolvable ages. The 87Sr/86Sr ratio obtained from multiple small pumices hand-picked from the sample is substantially higher than for any of the other eruptions analysed from this time period (Electronic Appendix 1; Fig. 4).

Fig. 9.

Data from the ‘New phreatoplinian’ eruption sample P1836. Model ages and analytical data are in Electronic Appendix 6. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (170 ka) from SIMS data on grains that returned finite ages. The approximate eruption age (∼45 ka; see text) is marked. Other details as in Fig. 6.

Fig. 9.

Data from the ‘New phreatoplinian’ eruption sample P1836. Model ages and analytical data are in Electronic Appendix 6. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (170 ka) from SIMS data on grains that returned finite ages. The approximate eruption age (∼45 ka; see text) is marked. Other details as in Fig. 6.

Okaia fall deposit

Charlier et al. (2005) reported 26 grains with finite ages and a further two that lay on the equiline within error. We have supplemented this dataset with 32 new analyses, 30 of which yielded finite ages, and the two datasets are plotted in Fig. 10. The enlarged dataset (Electronic Appendix 7) has a distinctly bimodal model-age distribution, with PDF peaks at 95 and 34 ka, closely similar in value to those seen in the Oruanui dataset. However, the relative abundance of zircons with model ages defining the younger peak is much smaller than in the Oruanui dataset, as evidenced by the numerical abundances of grains of given ages, the average zircon model ages from multiple-grain TIMS analyses (47·3 ± 1·2 to 69·3 ± 1·1 ka: Charlier et al., 2005) and the concentration-weighted mean model age of 85 + 16/–14 ka for the SIMS dataset.

Fig. 10.

Data from the Okaia fall deposit sample P1195. These combine our new data (33 points, 30 of which yielded finite ages) with 27 points (25 of which yielded finite ages) from Charlier et al. (2005). Model ages and analytical data are in Electronic Appendix 7. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (85 ka) from SIMS data on grains that returned finite ages. The multiple-grain TIMS ages from Charlier et al. (2005) are also shown as a hatched bar; errors on single points are contained within the symbol size. Other details as in Fig. 6.

Fig. 10.

Data from the Okaia fall deposit sample P1195. These combine our new data (33 points, 30 of which yielded finite ages) with 27 points (25 of which yielded finite ages) from Charlier et al. (2005). Model ages and analytical data are in Electronic Appendix 7. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (85 ka) from SIMS data on grains that returned finite ages. The multiple-grain TIMS ages from Charlier et al. (2005) are also shown as a hatched bar; errors on single points are contained within the symbol size. Other details as in Fig. 6.

Rubbish Tip dome

The earlier dataset reported by Charlier et al. (2005) included 21 grains, and we have analysed a further 32 grains (full dataset in Electronic Appendix 8). All grains analysed yielded finite ages, with the oldest being 188 + 32/–26 ka. The original and new, combined datasets (Fig. 11) show a unimodal distribution with a peak in the PDF curve at 45–54 ka that is closely matched by the previously published multiple-grain TIMS data, which yielded model ages between 40·1 ± 0·6 and 51·6 ± 0·8 ka. The concentration-weighted mean value is somewhat older, at 59 + 9/–8 ka.

Fig. 11.

Data from the Rubbish Tip dome sample R559. These combine our new data (32 points, all of which yielded finite ages) with 21 points (all of which yielded finite ages) from Charlier et al. (2005). Model ages and analytical data are in Electronic Appendix 8. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maximum marked), histogram and concentration-weighted mean value (59 ka) from SIMS data. Other details as in Fig. 6.

Fig. 11.

Data from the Rubbish Tip dome sample R559. These combine our new data (32 points, all of which yielded finite ages) with 21 points (all of which yielded finite ages) from Charlier et al. (2005). Model ages and analytical data are in Electronic Appendix 8. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maximum marked), histogram and concentration-weighted mean value (59 ka) from SIMS data. Other details as in Fig. 6.

Oruanui eruption products

Previously published SIMS data (Charlier et al., 2005) showed a strongly developed bimodal distribution in model ages, with an older peak identical within error to that in the Tihoi, ‘New plinian’ and Okaia fall deposits. Multiple-grain TIMS data showed a relationship between crystal sizes and model ages (Charlier & Zellmer, 2000). Larger size-fractions of crystals returned older model ages, which could be interpreted as reflecting either prolonged growth or the mixing of two populations of crystals. All these SIMS data were from two large pumices (‘Oruanui 2’ = P1373 and ‘Oruanui 3’ = P1520) collected from ignimbrite erupted in the late stages of the eruption (Phase 10 of Wilson, 2001). In the light of other evidence from crystal populations in Oruanui pumices for intensive mixing in the magma body prior to eruption (Liu et al., 2006; Wilson et al., 2006), we collected a new suite of SIMS data on zircons from the 250 and 125 μm sieve fractions of a multiple-pumice sample (P1634) from the earliest Oruanui deposits (fall unit 1 of Wilson, 2001). The full dataset for all three samples is given in Electronic Appendix 9. The model-age spectrum from P1634 (Fig. 12) is closely similar to the combined spectra of model ages collected from the two late-erupted pumices (Fig. 13), with PDF peaks of 86 and 41 ka. (The position of the 86 ka PDF peak is controlled by an 84 + 7/–8 ka grain with 7300 ppm U and 9700 ppm Th. Omission of this one grain alone moves the older peak in the PDF curve to 94 ka.) In sharp contrast to the data from the Okaia, the average ages from the TIMS multiple-grain analyses of zircons from late-erupted pumices lie close to or coincident with the younger peak, and show a grain-size dependence (Charlier & Zellmer, 2000), thus:

Fig. 12.

SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for all zircons analysed from the Oruanui samples. Existing data from two pumices recovered from deposits erupted late in the sequence (104 grains, 94 of which yielded finite age estimates: Charlier et al., 2005) are supplemented by 103 spots from 125 μm and 250 μm grains in the earliest erupted pumice clasts collected as P1634 (see Fig. 13, below). Analytical data and model ages are in Electronic Appendix 9. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (73 ka; but see text) from SIMS data on grains that returned finite ages. Other details as in Fig. 6.

Fig. 12.

SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for all zircons analysed from the Oruanui samples. Existing data from two pumices recovered from deposits erupted late in the sequence (104 grains, 94 of which yielded finite age estimates: Charlier et al., 2005) are supplemented by 103 spots from 125 μm and 250 μm grains in the earliest erupted pumice clasts collected as P1634 (see Fig. 13, below). Analytical data and model ages are in Electronic Appendix 9. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve (with maxima marked), histogram and concentration-weighted mean value (73 ka; but see text) from SIMS data on grains that returned finite ages. Other details as in Fig. 6.

Fig. 13.

Cumulative probability density curves and histograms from SIMS data for data from the Oruanui deposits to show the identical nature of the age spectra from early vs late-erupted material. Analytical data and model ages are in Electronic Appendix 9. (a) All samples; (b) two late-erupted pumices (P1373 = Oruanui 2; P1520 = Oruanui 3) previously published by Charlier et al. (2005); (c) a collection of earliest-erupted pumices (P1634).

Fig. 13.

Cumulative probability density curves and histograms from SIMS data for data from the Oruanui deposits to show the identical nature of the age spectra from early vs late-erupted material. Analytical data and model ages are in Electronic Appendix 9. (a) All samples; (b) two late-erupted pumices (P1373 = Oruanui 2; P1520 = Oruanui 3) previously published by Charlier et al. (2005); (c) a collection of earliest-erupted pumices (P1634).

125–250 μm sieve fraction: 39·6 ± 1·0 ka; 38·2 ± 0·7 ka;

63–125 μm sieve fraction: 37·2 ± 0·8 ka; 37·0 ± 0·8 ka; 35·4 ± 0·8 ka;

<63 μm sieve fraction: 32·0 ± 0·5 ka;

bulk zircon fraction (all sizes): 38·5 ± 1·0 ka.

The concentration-weighted mean from the SIMS data of 73 + 13/–12 ka for all Oruanui zircon samples (Fig. 12) is significantly older than the TIMS bulk ages.

Puketarata eruption products

Fifty grains were analysed, of which all but one yielded finite model ages. The full dataset is given in Electronic Appendix 10. The SIMS analyses define a strongly unimodal model-age spectrum, with a peak in the PDF distribution at 21 ka (Fig. 14). In part, the sharpness of this peak is a function of the relatively small uncertainties associated with these young zircons. However, even if the data are treated with an artificially higher estimate for the errors on single ages, the closeness of the PDF peak to the eruption age contrasts with the situation for the RTD and Ngangiho samples. The concentration-weighted mean from all grains from the SIMS dataset is 43 ± 4 ka.

Fig. 14.

Data from the Puketarata eruption sample P1835. Model ages and analytical data are in Electronic Appendix 10. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve, histogram and concentration-weighted mean value (43 ka) from SIMS data on grains that returned finite ages. Other details as in Fig. 6.

Fig. 14.

Data from the Puketarata eruption sample P1835. Model ages and analytical data are in Electronic Appendix 10. (a) SIMS (230Th/232Th) vs (238U/232Th) equiline diagram for single zircons. (b) Cumulative probability density curve, histogram and concentration-weighted mean value (43 ka) from SIMS data on grains that returned finite ages. Other details as in Fig. 6.

DISCUSSION

Information from Taupo zircons

Model-age data

The model-age data presented here represent a mixture of cores and rims in grains where the U contents (as indicated qualitatively by the intensity of CL: e.g. Dobson et al., 2008) were judged to be adequate to obtain an analysis with reasonably small uncertainties and the grains themselves (or growth zones of uniform CL intensity) were large enough to place the beam. With the exception of uncommonly large zircons from the 250 μm sieve fraction of Oruanui sample P1634 (data to be presented elsewhere) we have at no stage found any significant, consistent age zonation in the crystals we have analysed for U–Th disequilibrium model ages from rocks in the TVZ. Both cores and rims in the 125 μm size fraction (which were analysed for preference, because of the relative abundance of grains) of sample P1634 yielded ages at seemingly random points within the overall model-age spectrum. For six grains from this size fraction we analysed cores and rims (the latter being as close to the true rim as we could get without impinging on the epoxy mount, i.e. within ∼10 μm) and found only one grain with significant differences in age (Table 4). We infer that the crystals analysed grew relatively quickly (when compared with the total spectrum of model ages within each sample) and afterwards did not grow to any great extent that could be seen from zonation visible under CL or analysed with the SHRIMP-RG spot placed as close to the rim as possible. Surface profiling might well reveal thin rims with model ages closer to the eruption age (see Reid & Coath, 2000; Vazquez & Reid, 2004), but this would not affect our conclusions. We would thus interpret the multiple-grain TIMS results for the Oruanui zircons reported by Charlier & Zellmer (2000) as dominantly reflecting mixing of two zircon populations with contrasting grain-size distributions and mean model ages (see Figs 12 and 13), rather than interrupted growth of crystals with older cores and younger rims. The model-age variations are such that any thin eruption-age outer rims (i.e. too thin to be visible on the CL images) could not in themselves provide the age leverage reported by Charlier & Zellmer (2000). Comparably similar core–rim model-age relationships were reported for zircons analysed in cross-section by SIMS for U–Pb in the Bishop Tuff (Reid & Coath, 2000).

Table 4:

Summary of model ages from cores and rims of 125 μm sieve-fraction zircons in Oruanui sample P1634

Grain Spot Core  Rim  Core  Rim  
 number Age 1 SD Age 1 SD Th Th 
  (ka) error (ka) error (ppm) (ppm) (ppm) (ppm) 
P1634-39 P1634-39·1 32 ±4   2200 3100   
 P1634-39·2   39 +18/−15   200 100 
P1634-42 P1634-42·1 40 ±4   3000 4100   
 P1634-42·3   54 ±9   700 700 
P1634-43 P1634-43·1 43 ±4   1900 1700   
 P1634-43·3   38 +10/−9   400 400 
P1634-44 P1634-44·1 66 +22/−18   200 200   
 P1634-44·2   23 +10/−9   400 300 
P1634-46 P1634-46·1 84 +31/−24   200 100   
 P1634-46·2   42 +13/−12   300 300 
P1634-49 P1634-49·1 41 ±4   3100 4500   
 P1634-49·2   35 +9/−8   600 500 
Grain Spot Core  Rim  Core  Rim  
 number Age 1 SD Age 1 SD Th Th 
  (ka) error (ka) error (ppm) (ppm) (ppm) (ppm) 
P1634-39 P1634-39·1 32 ±4   2200 3100   
 P1634-39·2   39 +18/−15   200 100 
P1634-42 P1634-42·1 40 ±4   3000 4100   
 P1634-42·3   54 ±9   700 700 
P1634-43 P1634-43·1 43 ±4   1900 1700   
 P1634-43·3   38 +10/−9   400 400 
P1634-44 P1634-44·1 66 +22/−18   200 200   
 P1634-44·2   23 +10/−9   400 300 
P1634-46 P1634-46·1 84 +31/−24   200 100   
 P1634-46·2   42 +13/−12   300 300 
P1634-49 P1634-49·1 41 ±4   3100 4500   
 P1634-49·2   35 +9/−8   600 500 

There is a useful contrast to be made in the Taupo datasets between SIMS and TIMS model-age information. The TIMS analyses reported by Charlier et al. (2005) included many hundreds to thousands of grains, and thus represent a better estimate for the average ages in the relevant size fractions of crystals. The SIMS data, however, yield a much clearer picture of the total model-age spectrum. In Table 3, we summarize the SIMS concentration-weighted mean values for the samples presented here. In general, the concentration-weighted means from SIMS data are significantly older than the TIMS values. We do not infer that this difference merely reflects a sample-size issue (i.e. the number of grains analysed), but instead suggest that it reflects three factors, two of which reflect the nature of the SIMS analytical techniques. First, we are not sampling the smaller grains (rarely the 63 μm and never the <63 μm fractions) that would be expected to be growing actively in the magma at the time of eruption. Second, we are sampling grains with mid- to dark grey or better hues under CL to ensure that the U concentration is adequate to ensure a viable analysis (see Dobson et al., 2008). Many grains that are bright under CL thus have to be avoided, and in our experience on the 250 μm Oruanui P1634 grains (Charlier et al., in preparation) the low-U grains are skewed toward younger ages. Third, the concentration-weighting process can be skewed by single grains with very high U and Th concentrations, in some cases changing the mean by 5–10 kyr with the addition or subtraction of a single grain from the dataset.

Model-age spectra

The model-age spectra illustrated in Figs 6–14 can be grouped into three suites (Fig. 15).

(1) The distributions shown by samples from the ‘Oruanui-type’ magma (Sutton et al., 1995); that is, the Oruanui eruption itself plus the Okaia and Tihoi, and possibly the ‘New plinian’. These show a consistent peak that resolves itself in the PDF function in Isoplot (Ludwig, 2003) at 86–95 ka, and a variably developed younger peak that is poorly resolved in the Tihoi and ‘New plinian’ samples, small but clearly defined in the Okaia, then dominant in the Oruanui itself. This bimodality is not as prominent in the model-age spectra from the post-Oruanui units studied by SIMS, but the younger peak is seen in rhyolitic units B (11·8 ka) and G (6·7 ka: Fig. 15). The SIMS data for zircon model ages for the Oruanui in Charlier et al. (2005) were obtained from two late-erupted pumices. To compare with these data, we measured the model-age spectrum in a multiple-clast sample from fall unit 1, the earliest deposit of the eruption. From comparisons between the early and late-erupted samples (Fig. 13) it is apparent that the age spectra are identical. This indicates that whatever process created the bimodal model-age distribution, and regardless of the spectrum of Oruanui pumice bulk compositions, this distribution was present throughout virtually the entire erupted volume of magma.

Fig. 15.

Summary diagram to show overall features of the zircon model-age spectra for the units detailed in this paper, plus the 11·8 ka Unit B and 6·7 ka Unit G rhyolites (Wilson, 1993; Sutton et al., 2000). PDF curves from each of Figs 6–14 and the Units B and G data from Charlier et al. (2005) have been taken and scaled to a common maximum peak height. Shaded grey bands highlight the 86–95 ka and 34–41 ka peaks in PDF curves for the model-age spectra shown by the Oruanui, its precursors and some successors. Dashed lines indicate the eruption age. (See text for discussion.)

Fig. 15.

Summary diagram to show overall features of the zircon model-age spectra for the units detailed in this paper, plus the 11·8 ka Unit B and 6·7 ka Unit G rhyolites (Wilson, 1993; Sutton et al., 2000). PDF curves from each of Figs 6–14 and the Units B and G data from Charlier et al. (2005) have been taken and scaled to a common maximum peak height. Shaded grey bands highlight the 86–95 ka and 34–41 ka peaks in PDF curves for the model-age spectra shown by the Oruanui, its precursors and some successors. Dashed lines indicate the eruption age. (See text for discussion.)

(2) A more or less unimodal spectrum, with very few or no crystals returning ages that overlap with the equiline. This spectrum is seen in the samples from Ngangiho dome, Rubbish Tip dome and Puketarata eruption deposits. Although the difference in age between the PDF peak and inferred eruption ages is not the same across the samples (Ngangiho ∼30 kyr; Rubbish Tip ∼15–25 kyr; Puketarata ∼5 kyr), all three are moderately crystal rich, contain biotite in the ferromagnesian fraction and are unusual as such in young eruptive units in the Taupo–Maroa area (Ewart, 1968; Sutton et al., 1995; Leonard, 2003). The Ngangiho and Rubbish Tip domes (together with the Poihipi phreatoplinian fall deposit that was co-eruptive with but preceded the latter) are similar chemically and isotopically, and are linked as the ‘NE dome’ magma type by Sutton et al. (1995). The Puketarata deposits are slightly different on both those counts (Fig. 4), but are similar enough to suggest a consanguineous relationship to the other two eruptive units and a common source chamber.

(3) A broad bimodal spectrum, with a 100–86 ka peak (like the older peak in the Oruanui-type) and a substantial tail of older model ages, some finite, and others on or within error of the equiline, as shown by P1836 from the ‘New phreatoplinian’ fall deposit. This contrast in the model-age spectrum is matched by a substantial difference in Sr-isotopic ratio measured in the pumices (Electronic Appendix 1, Fig. 4). The peak of zircon model ages in the 200–300 ka range (Fig. 9) is matched closely by the age of surficial rhyolitic volcanism at Maroa volcano (Leonard, 2003), although field data on the thickness and distribution of the ‘New phreatoplinian’ fall deposit suggest that it was not erupted through or even close to the dated surficial domes or deposits themselves.

Relationships between magmas in the Taupo–Maroa area

The contrasting suites of model-age data suggest that there were two very contrasting types of magma chamber in the Taupo area up to ∼27 ka, yielding the ‘Oruanui-type’ and the biotite-bearing rhyolite compositions. These occur in interbedded deposits and so both magma chambers were active over the same time period.

‘Oruanui-type’ magmas

The melt-rich aliquots from the volumetrically dominant ‘Oruanui-type’ magma chamber are inferred to have incorporated antecrystic zircons from an older crystal mush (ages peaking at 86–95 ka in all four samples), with variable amounts of crystallization occurring as the relevant melts crystallized and/or cooled to some extent prior to eruption. The two older eruptions (Tihoi and ‘New plinian’) show only minor younger crystallization (Figs 6 and 7), and in turn also extend only to somewhat less-evolved compositions than the Oruanui itself (Table 1; Fig. 4; Sutton, 1995). The Okaia shows a modest peak of young crystallization, centred on 34 ka, whereas the total Oruanui dataset has a dominant younger peak also at 40 ka. The relative importance of the older and younger peaks in model ages from zircons in the four deposits (as estimated by concentration-weighted mean ages from the SIMS data) is reflected by the average ages derived from TIMS data (Charlier et al., 2005).

These observations, however, raise major implications about the nature of the ‘Oruanui-type’ magma chamber. The volume of the melt-dominant zone tapped in the Oruanui eruption (∼530 km3) dwarfs that of all the other eruptions, and so it would naturally be assumed that the earlier tappings should record the progressive growth of such a body, but this is demonstrably not the case from the model-age spectra. The Okaia model-age spectrum contains few crystals with model ages corresponding to the pre-eruption minor growth peaks reflected in PDF curves in data from the Tihoi (64 ka) or ‘New plinian’ (67 ka) fall deposits (see Figs 7, 8 and 10), implying that the Okaia was independently generated from the common ‘95 ka mush zone’, with no mixing in of any residual zircon-bearing crystal-rich or melt-rich material from these intervening two events. The contrast between the Okaia and Oruanui spectra is equally clear cut. Despite being erupted only ∼3 kyr apart, the relative dominance of the young ∼40 ka peak in zircon model ages in the Oruanui (Fig. 12) is striking when compared with its importance in the Okaia (Fig. 10). However, several lines of evidence effectively preclude the Okaia being an earlier tapping off of a subsequently erupted Oruanui melt-dominant zone, with an intervening episode of large-scale zircon crystallization.

(1) The TIMS model ages for the smallest (<63 μm) size fractions of zircons from the Okaia (50·4 ± 1·3; 47·9 ± 1·3 ka) and Oruanui (32·0 ± 0·5 ka; errors are 2 SD) are different. Given the errors on the eruption ages (Lowe et al., 2008), it would seem unlikely that the same melt-dominant zone could have been tapped at 30 ka in the Okaia eruption and 27 ka in the Oruanui eruption when the youngest TIMS zircon model ages in the Oruanui pre-date the Okaia event.

(2) The younger Oruanui SIMS model-age peak should lie centred between 30 and 27 ka, whereas in fact it is centred on 40 ka.

(3) Any suggestion that the Okaia might simply have tapped magma that contained a higher proportion of older zircons (for example, from a shallower, cooler portion of the melt-dominant zone) than the Oruanui is opposed by two factors. First, the crystallinities, Fe–Ti oxide temperatures and crystal assemblages of the Okaia and the Oruanui rhyolite pumices sampled for this work are the same (Sutton, 1995). Second, zircon yields from Okaia and Oruanui pumices were the same (Charlier et al., 2005), so precluding any notion that the young population of zircons was somehow ‘missing’ in the Okaia, but present in the Oruanui.

Although their vent sites are closely coincident, the erupted Okaia magma (i.e. the melt-dominant body from which it was erupted) was generated independently of the Oruanui magma, even though they were derived from the same mush zone in the same chamber.

What caused the greater degree of young zircon crystallization in the Oruanui samples, given that both the Okaia and Oruanui compositions are both zircon saturated (Charlier et al., 2005)? We infer that this reflects two interconnected processes. First, the bulk pumices analysed in the Oruanui have a range in major- and trace-element compositions consistent with the most-evolved being derived from the least-evolved compositions by fractionation and removal of ∼28 wt % of the crystal phases present in the pumices (Sutton, 1995; Wilson et al., 2006). If, instead of the crystals being sequestered from a melt-dominant magma, the easier process of having the liquid abstracted from a crystal-dominant mush was to operate (Bachmann & Bergantz, 2004; Hildreth, 2004) then it seems likely that some zircons could have been elutriated along with the melt. Second, in addition to the ∼28% of fractionation inferred to have occurred from the bulk pumice data there is evidence for further degrees of crystal fractionation in the compositions of quartz-hosted melt inclusions. The compositions of these melt inclusions range from being similar to the surrounding host melt to compositions that are consistent with up to 40–50% equilibrium crystal fractionation involving the phases that are present in the bulk pumices (Liu et al., 2006).

If it is thus inferred that some significant fraction of the quartz crystals in the Oruanui pumices were derived from a crystal mush, then it would be inevitable (as shown by the trace-element characteristics of the same glass inclusions: Liu et al., 2006) that such a mush would also be crystallizing zircon (along with other mineral phases). These zircons would also be introduced to any melt-dominant zone at some stage during the evolution of the chamber (in addition to any zircons already growing in the melt-dominant zone itself). From these observations we conclude that the major differences in the zircon model-age spectra between the Okaia (and ‘New plinian’ and Tihoi) and the Oruanui itself reflect the abundance in the Oruanui of crystals derived from mush material formed after the Tihoi and ‘New plinian’ eruptions as the counterpart to the crystal fractionation processes reflected in the bulk pumice chemical variations and quartz-hosted melt inclusions.

In this scenario, the magmas erupted in the Tihoi, ‘New plinian’ and Okaia events represent material derived by remobilization of older (95 ka peak model age) mush material, which then underwent limited crystallization (and proportionally limited major- and trace-element evolution) some time prior to eruption. The inferred source of the Okaia eruption is within the geographical area of the Oruanui melt-dominant zone, with its footprint reflected in the Oruanui structural caldera collapse area (Davy & Caldwell, 1998; Wilson, 2001; Fig. 2). We discuss above why the magma parcel erupted in the Okaia event is inferred to have been independently generated from the Oruanui melt-dominant zone. This in turn dictates that at least a major proportion of the Oruanui melt-dominant body was physically generated during the ∼3 kyr time gap between the Okaia and Oruanui eruptions by extraction and accumulation of crystal-bearing liquids that had evolved in a common mush zone. The peak of the chemical generation of these liquids and hence the accompanying crystallization of mineral phases in the Oruanui magma is indicated by the younger model-age zircon peak around 41–35 ka and the 33 + 18/–16 ka average age of the major mineral phases (Charlier et al., 2005, fig. 7).

Biotite-bearing magmas

Of the three biotite-bearing eruptive units, Ngangiho and Rubbish Tip share unimodal zircon model-age spectra (Figs 6 and 11) and contrast as such with the Oruanui-type magma, even though both the ‘Oruanui’ and ‘NE dome’ magma chambers were active at the same time. Each unit shows a peak in model age significantly prior to the inferred eruption age, and all the grains analysed yielded finite ages. Each magma appears to have been generated, matured by crystallizing various minerals, including zircon, and then been released in a single cycle, with crystallization peaking at ∼30 kyr (Ngangiho dome) and ∼15–25 kyr (Rubbish Tip dome) prior to eruption.

Although post-dating the Oruanui eruption, Puketarata is similarly crystal-rich and biotite-bearing to the earlier two units, but has some chemical and isotopic differences (e.g. Fig. 4: Sutton, 1995). Its stratigraphic position places it between the early post-Oruanui crystal-rich dacites (Units Ψ, Ω and A of Wilson, 1993) and the earliest of the crystal-poor Subgroup 1 rhyolites erupted at 11· 8 ka (Unit B of Wilson, 1993; Sutton et al., 2000). However, there are no links between the Puketarata composition and either of the bracketing compositional groupings at Taupo. The peak in Puketarata crystallization is more narrowly defined (43 of 49 model ages between 15 and 40 ka) and is centred closer to the eruption age than the corresponding peaks in the RTD and Ngangiho examples.

In all three examples, the model-age spectra suggest that the magmas concerned were derived from some kind of progenitor magma via a process that effectively stripped out any inherited old zircons, such that only one spot analysis out of 155 yielded a model age that lay within 1 SD. of the equiline (as compared, for example, with the ‘New phreatoplinian’ unit). If some proportion of the crystals analysed represents antecrysts or xenocrysts (which seems likely, from the shape of the older tail to the age distribution) this cannot be quantified from the data available. However, there is no bimodality to the model-age spectra such as that seen in the Oruanui-type magmas (Figs 12 and 13) that would suggest large-scale remobilization of a progenitor crystal mush of narrowly defined age. Crystallization processes operated in all three cases to generate biotite-bearing magmas with moderate to high crystal contents [11% (Ngangiho); 23% (RTD); 16–20% (Puketarata)] but low Fe–Ti oxide temperatures (745°C for RTD and Ngangiho, 725°C for Puketarata: Sutton, 1995; Sutton et al., 1995).

Laterally extensive sources or lateral magma transport?

The zircon model-age spectra and apparent eruption ages for the Tihoi and ‘New plinian’ deposits are closely similar, yet they were erupted from sources ∼20 km apart (Figs 2 and 5). Does this spatial separation indicate that the deep magma chamber mushy roots (with a 95 ka peak in zircon model ages) for the ‘Oruanui-type’ system extended over a total north–south distance of ∼30 km, or that the ‘New plinian’ eruption was fed by magma that had undergone lateral transport in the crust? We consider the latter interpretation to be more likely, in part from the general TVZ setting, where lateral magma transport has been shown to occur in association with single or clustered eruptions that accompany regional rifting events (e.g. Kohn & Topping, 1978; Nairn et al., 1998, 2005), and in part from the apparent coincidence of three magmatic and one phreatic eruptions occurring so close together in time from widely separated vent areas (C. J. N. Wilson, unpublished data). In addition, a line between the inferred vent sites for the Tihoi and ‘New plinian’ eruptions is parallel to the general NNE–SSW tectonic grain of the TVZ (Fig. 2). Conventionally, the vents for small rhyolite eruptions are considered to directly overlie their source magma chambers, but we suggest that this view may be misleading in areas such as the TVZ where magma-assisted extension is significant.

For the three biotite-bearing magmas, an opposite conclusion is reached. For the Ngangiho and Rubbish Tip domes, the vent sites are ∼15 km apart (Fig. 2), yet the similarities in chemistry, mineralogy and isotopic ratios (Fig. 4; Sutton et al., 1995) between the two eruptive units suggest that they come from a common magma chamber. The vent alignments for the two lobes of the Ngangiho domes and also the Rubbish Tip and its inferred co-eruptive Trig 9471 domes are NNE–SSW, the same as the regional tectonic grain for the TVZ. However, in both cases these local alignments are nearly perpendicular to a line joining the two eruptive centres that cuts across this tectonic grain. We therefore infer that these two eruptions are unlikely to have been fed from a common geographically confined focus (unless the feeder system was a sill at depth) but, instead, that the source region for these two domes has a width comparable with the spacing of the vent sites.

Active magma chambers in the Taupo area

In classical treatments of large silicic magmatic systems (e.g. Smith, 1979; Hildreth, 1981; Shaw, 1985), particularly those of a size capable of producing a supereruption, emphasis is given to three aspects. First, the density trap formed by the mid-crustal accumulation of partially molten mush and silicic fractionates, and the thermal trap formed by hydrothermal systems in the shallow crust, are thought to trap and focus hotter, denser, less-evolved magmas rising from the mantle and lower crust, limiting their interaction with the melt-dominant silicic body above (e.g. Hildreth, 1981, 2004; Annen et al., 2006). The mature host system will largely buffer its melt-dominant zone from the extremes both of interaction with surrounding crustal rocks and direct input from less-evolved melts. Second, the melt-dominant eruptible magma is in physical and thermal continuity with its mushy roots, and typically any large eruption will tap down into the crystal-richer transition zone and into the mush zone (e.g. Smith, 1979; Hildreth & Wilson, 2007). Third, eruptions subsequent to the climactic event will involve mixing between the magma compositions associated with the climactic event and new compositions (e.g. Hildreth, 2004).

In the case of the TVZ in general and the Taupo area in particular, these generalizations do not appear to hold true for the ≤61 ka period of activity, and also for some older eruptions, as shown by a variety of evidence.

(1) Mafic compositions, of broadly basaltic to andesitic compositions have penetrated to interact with the rhyolitic melt-dominant zone of TVZ magma chambers (Blake et al., 1992; Brown et al., 1998; Leonard et al., 2002; Schmitz & Smith, 2004; Wilson et al., 2006; Shane et al., 2007, 2008a). Although the presence of a hot melt-bearing low-density zone in the crust is inferred from geophysical evidence in addition to petrology (e.g. Heise et al., 2007), the presence of active rifting appears to be a key factor that permits mafic to intermediate magmas to rise to high levels in the crust and be occasionally erupted along with the silicic magmas.

(2) Isotopic data suggest that some degree of interaction with non-igneous lithologies may continue at all stages of magma generation from basalt to rhyolite. At Taupo, dacites and rhyolites show evidence of crustal greywacke components (zircons, feldspars crystallized from melts with unusually high radiogenic Sr-isotopic compositions) being incorporated into the magmas, in some cases only shortly prior to eruption (Charlier et al., 2005, 2008).

(3) No eruptions show a systematic zonation from crystal-poorer to crystal-richer compositions that would indicate a stable zonation in the parental melt-dominant bodies (compare the Bishop Tuff: Hildreth, 1979; Hildreth & Wilson, 2007). Crystal contents (and bulk-rock chemistries and other features) can be uniform (e.g. the 1·8 ka Taupo eruption: Dunbar & Kyle, 1993; Sutton et al., 1995), non-systematically variable (e.g. the Oruanui eruption: Wilson et al., 2006), or variable as a result of multiple magma batches, each distinctive in character and inferred to be from separate melt-poor bodies, with only limited syn-eruptive mingling (Brown et al., 1998; Nairn et al., 2004; Smith et al., 2004; Shane et al., 2007, 2008a, 2008b).

At Taupo, available data from field, geochemical and model-age studies suggest that the build-up to the large Oruanui eruption saw the activity of two independent magma chambers only ∼15 km apart. Cycles of crystallization in the melt-dominant zones formed from these two chambers are independent, such that if the peaks in crystallization age are considered to represent thermal events, the two chambers did not experience the same thermal histories. Thus, development of the Oruanui magma chamber did not stop or divert magmatic activity in areas that were very close nearby.

Following the Oruanui eruption, about 6000 years elapsed before volcanism below the Taupo area resumed with three small dacite eruptions, which were followed by numerous rhyolite eruptions (Wilson, 1993). Compositions of these rhyolites cluster into three subgroups on the basis of subtle but distinctive differences in bulk chemistry, mineral chemistry and isotopic ratios (Sutton et al., 2000). None of these rhyolitic or dacitic compositions or isotopic ratios fall along any mixing trend with the Oruanui-type magmas (Fig. 4), but the zircon age spectra show significant degrees of temporal overlap with the Oruanui and common PDF peaks (Fig. 15). The average model ages for post-Oruanui rhyolites from TIMS data (Charlier et al., 2005), however, range widely in apparent response to the degrees of zircon saturation in the magmas. Undersaturated (Unit S: 3·6 ka) to just-saturated [Units D (11·4 ka) and G (6·7 ka)] examples show average model ages that pre-date the Oruanui eruption, whereas oversaturated examples [Units B (11· 8 ka) and E (10·0 ka)] have average model ages that are younger than 27 ka and reflect the dominance of post-Oruanui crystallization accompanying the generation of rhyolite from the consanguineous dacitic progenitors erupted from 17 to 21·5 ka (Sutton et al., 2000). The post-Oruanui data collectively imply that the average zircon model-age patterns in large part reflect the local crystallization conditions (namely, the state of zircon saturation in the melts) and not necessarily the longevity of the magma source zone for any single eruption. However, the SIMS age spectrum (Fig. 15) for Unit B has a 35 ka peak in the PDF curve closely similar to that of the Oruanui and Okaia deposits, and it is apparent that zircons from the earlier crystallization episode were remobilized and incorporated from a common source zone ‘chamber’. This chamber in post-Oruanui activity, however, did not retain enough melt to influence significantly the compositional characteristics of the dacites or rhyolites. In addition, the presence of greywacke-derived zircons in two of the deposits (dacitic Unit Ω and rhyolitic Unit G) investigated by Charlier et al. (2005) indicates that influxes of melts from old metasedimentary crustal protoliths must also have been involved very shortly prior to eruption (see also Charlier et al., 2008).

At the same time that post-Oruanui volcanic activity was active at Taupo, the 16 ka Puketarata eruption occurred from vents only ∼15 km north of Lake Taupo (Fig. 2), showing contrasting chemical, mineralogical and isotopic characteristics (Fig. 4) and SIMS model-age spectrum (Fig. 14) with respect to either the Oruanui or any other magma-type vented in the Lake Taupo area. The independence of magmatic systems in the broader Taupo area seen in the build-up to the Oruanui eruption has thus continued (and may be extant).

Age information from zircons in young volcanic rocks

Model ages

There is now a substantial body of data on U–Th and U–Pb model-age systematics from zircons in young (Quaternary) volcanic rocks, and a vigorous debate on the significance of the age spectra and what they tell us about the timing of ‘crystal residence’ and growth of silicic melt-rich bodies in particular or magma chambers in general (among many papers: Reid et al., 1997; Brown & Fletcher, 1999; Reid & Coath, 2000; Lowenstern et al., 2000, 2006; Bindeman et al., 2001, 2008; Bacon & Lowenstern, 2005; Simon & Reid, 2005; Charlier et al., 2005; Bachmann et al., 2007a; Crowley et al., 2007; Reid, 2008; Simon et al., 2008). In comparing our work with these published papers and their conclusions, there are three questions that arise from any suite of model ages. (1) What sort of sample is required for an accurate assessment of the crystal age spectra? (2) What are the single model ages telling us? (3) What do the spectra of crystal model ages reveal about processes in silicic magma chambers?

In U–Pb studies of ancient detrital (and many magmatic) zircons, a (the?) primary aim is to identify and date single model-age populations within a complex spectrum of ages over (usually) tens to hundreds to thousands of millions of years. In such cases the analytical uncertainties on the single model ages are much smaller than the timings between peaks. The questions that then arise about how many populations are present and whether multiple populations are resolvable can be addressed with statistical techniques for assessing how many grains are required to be analysed (e.g. Dodson et al., 1988), and defining and separating various sub-populations (e.g. Sambridge & Compston, 1994).

With U–Th disequilibrium model ages, in contrast, the analytical uncertainties on single SIMS analyses (typically thousands to tens of thousands of years) are much larger relative to the time window over which such age estimates are valid (i.e. ∼300 kyr), and so identification of grain sub-populations cannot be done in isolation from consideration of analytical uncertainties. With a longer count time for 230Th16O for our recent results when compared with the earlier data, analytical precision has been improved. This raises an important issue in considering the nature of the peaks shown in Figs 6–14, especially the form of the PDF curve yielded by Isoplot (Ludwig, 2003). Depending on the uncertainties associated with each analysis, Isoplot will generate a single smooth PDF curve where errors are larger, but tends to split the curve into multiple smaller spikes as the range of model ages forming a particular mode or cluster of ages exceeds the analytical precision (Fig. 16). Variations in analytical precision in these young samples replaces the choice of band width in controlling the shape of PDFs in old zircon analyses (see Rudge, 2008). Thus it is important to consider what is being measured in the spectrum of ages obtained. In this study, we are interested in assessing the timing of zircon crystallization, as represented by model ages, on a time scale of the order 5–20 kyr in order to look at the large-scale crystallization or thermal histories of the magma chambers in the Taupo area. Analytical uncertainties that are larger than we have obtained do not significantly affect the ages of peaks in the PDF curve for unimodal samples (Fig. 16), but remove the bimodality in model-age suites (and hence the model-age basis for the distinction between antecrysts and phenocrysts: Hildreth, 2004; Charlier et al., 2005; Bryan et al., 2008).

Fig. 16.

Plot showing histograms of raw slopes and model ages, and associated PDF curves from (a) Puketarata (Fig. 14) and (b) the Oruanui total dataset (Fig. 12), to show the effects of varying analytical precision on the shape and quality of the PDF curve yielded in Isoplot (Ludwig, 2003). The values in each panel refer to an artificial 1 SD uncertainty on the equiline slope that defines the two-point isochron for each data point, and the age of the resulting peak in the PDF curve. The uncertainties for the ‘actual data’ are the real values given in Electronic Appendices 9 and 10. (See text for discussion.)

Fig. 16.

Plot showing histograms of raw slopes and model ages, and associated PDF curves from (a) Puketarata (Fig. 14) and (b) the Oruanui total dataset (Fig. 12), to show the effects of varying analytical precision on the shape and quality of the PDF curve yielded in Isoplot (Ludwig, 2003). The values in each panel refer to an artificial 1 SD uncertainty on the equiline slope that defines the two-point isochron for each data point, and the age of the resulting peak in the PDF curve. The uncertainties for the ‘actual data’ are the real values given in Electronic Appendices 9 and 10. (See text for discussion.)

Furthermore, any degree of bimodality to the extent shown by our U–Th data and inferences about recycled antecrysts versus newly grown phenocrysts from the Oruanui samples and its precursors will not easily be recognized in analogous studies of older systems that are reliant on U–Pb age systematics determined by SIMS. This is because the associated age uncertainty from SIMS analysis (typically 50–200 ka in zircons in TVZ rocks: Brown & Fletcher, 1999; Brown & Smith, 2004; Wilson et al., 2008) will cause the PDF function to show only one peak when the modes are less than ∼100 kyr apart. In isotope dilution (ID)-TIMS studies, however, although analytical precision is of the order of a few thousand years (e.g. Crowley et al., 2007), the time-consuming nature of analytical techniques restricts the number of samples that can be reasonably processed, and the consequently limited size of the sample suite may cause one or more modes to be missed completely. At its most extreme, for both U–Th and U–Pb dating, achievement of very high precision (i.e. to the level where the single age uncertainties are much smaller than the variation of single grain ages, such as is usually the case with U–Pb dating of detrital zircons in sedimentary rocks) causes breakdown of the age spectra into multiple modes. The significance of any clustering is then obscured by variability related to other causes, such as an inadequate number of grains to represent the age spectrum, and intra-grain variability in model ages.

Model-age spectra

What do the spectra of model ages tell us about the conditions accompanying growth of silicic magma bodies? A common feature in Quaternary examples is that the peak in PDF curves from the model ages occurs some time prior to quenching—between 4 and 40 kyr in the examples documented here. Simon et al. (2008) reviewed the literature on Quaternary zircon model-age data for many examples (including the Taupo and Okataina data of Charlier et al., 2003, 2005) and cited values of 33–247 kyr between the mean model ages of zircons and eruption, with a median of 92 kyr (70 kyr for ‘large-volume events’). In contrast, Crowley et al. (2007) presented single-crystal ID-TIMS data from eight zircons from the Bishop Tuff that yielded U–Pb model ages that in essence were identical within error to a plausible eruption age estimate. These contrasting views reflect two fundamental issues that are rarely addressed in discussions of zircon age spectra.

(1) Why should the maximum apparent rates of crystallization in magma bodies, as reflected in the zircon model-age spectra (whether as maxima in PDF curves, or as mean values; Simon et al., 2008), occur some significant time prior to the eruption and not at the point of eruption? There could be several reasons for this that relate to analytical techniques, not to natural variations introduced by the crystallization history of the magma. One reason may be that SIMS measurements for U–Th disequilibrium and U–Pb determinations typically have to deal with crystals of a certain size, in our experience mostly from the 125 μm sieve fraction (although age variations within such crystals are limited: see Table 4). Smaller crystals are less often measured, and multiple-crystal TIMS determinations show that the smaller size fractions that can be recovered during sample processing may be preferentially younger on average (Charlier & Zellmer, 2000). In turn, very small (<63 μm) crystals that were actively nucleating and growing at the time of quenching may not be measured at all. Thus the age determinations for many deposits are biased by what sizes of zircon crystals have been analysed. A second reason reflects the processes that actually lead to eruption. At Taupo, the eruption record is such that eruptions appear to occur chaotically, with no systematic relationships between repose periods between the eruptions and the size of the preceding or following eruptions (Fig. 3; Wilson, 1993). Various studies in the TVZ (e.g. Nairn et al., 1998; Gravley et al., 2007; Berryman et al., 2008) have demonstrated a close temporal relationship between rifting processes and volcanic events, such that, in several cases, regional tectonic faulting can be shown to have occurred at the time of eruptions from vents over 10–20 km away. Modern geodetic observations and studies of ancient shorelines at Taupo volcano itself (Otway et al., 2002; Manville & Wilson, 2003; Smith et al., 2007) show deformation patterns that suggest that there are regular inputs of magma into the crust below the Taupo area but, clearly, only sometimes is this magma triggered into eruption. From these observations and the longevity of and timing of major zircon crystallization episodes in the Taupo magmas (Fig. 15), it seems likely that melt-dominant bodies may be being generated or present virtually continuously from an evolving deeper source, and that the eruption timing is decided by rifting processes, not by any age-related state (or consequent physical condition) of the melt-dominant body or its parental mush zone. The gap between eruption and peak model ages would then reflect the chance tectonically triggered tapping of a melt-dominant body in a chamber where numerous other pulses of melt-dominant material might have been generated, then largely to wholly crystallized again without eruption.

(2) Do average ages (cumulative weighted means) for zircon model-age spectra from SIMS or other single-crystal analyses have any meaning? The model-age spectra at Taupo show that taking the average model ages, or the spread in model ages for zircons derived from an eruptive rock may be misleading in assessing the pre-eruption residence time of crystals and the longevity of the magmatic system. There are several reasons for this. First, as we show above, the average model-age value of a zircon population (a) may conceal the bimodality or polymodality suggestive of recycling of antecrysts or xenocrysts, and (b) depends strongly on the size and nature of each grain analysed, with inherent biases in SIMS studies towards larger and/or more U-rich grains. Second, taking the average conceals the fact that the crystals measured might be phenocrysts, antecrysts, or xenocrysts unrelated to the activity associated with the magmatic system that generated the rock sampled (Charlier et al., 2005; Bryan et al., 2008). Such components cannot be defined without further information. For example, at Taupo, the Unit Ω dacite erupted at 20 ka contains zircons that yield ages between 41 + 34/–26 or 45 + 10/–9 ka and 524 ± 10 Ma (Charlier et al., 2005). The host melt is strongly undersaturated in zircon, so all the zircons in this magma are antecrysts or xenocrysts, inherited during the magma generation process immediately prior to eruption (within years or less, on the basis of their euhedral morphologies and estimates of the time for total dissolution: Charlier et al., 2005). None of them are phenocrysts. The rhyolite of Unit B (erupted at 11·8 ka, and plausibly derived from a dacitic parent of a composition similar to Unit Ω: Sutton et al., 2000) has a model-age spectrum that includes ages that pre- and post-date the Unit Ω eruption age. The post-20 ka zircons are thus plausibly phenocrysts in Unit B, but pre-20 ka crystals have then by definition to be inherited antecrysts. Age data alone thus cannot determine where the phenocryst–antecryst–xenocryst divisions lie and the zircon model-age spectrum (whether U–Th disequlibrium or U–Pb) cannot be used in isolation to infer how long a magma chamber has been active or what the residence time of a magma was (see Brown & Fletcher, 1999; Simon & Reid, 2005; Simon et al., 2008). Of course, approaches that employ a combination of dating of accessory minerals and microanalytical techniques on these and other crystal phases may be able to unambiguously resolve patterns of inheritance and magma chamber evolution.

Mush zones, pluton growth and the dynamics of silicic magma chambers

Many papers have debated the connections and relationships between silicic plutonic systems and their volcanic counterparts (for review, see Bachmann et al., 2007b). Several elements of the debate arise, however, because of a lack of clarity and understanding of the time and length scales that can be discerned from the different records available. The relative youth and precision of the age record at Taupo in particular, and the central TVZ in general, can help sharpen the focus of such debates. Four points stand out.

(1) In the lifetime of the central TVZ, silicic volcanism has occurred over a ∼6000 km2 area over about a 1· 6 Myr period. The spatial overlap of volcanic sources (e.g. Wilson et al., 1995, 2009), and the present-day distribution of mid-crustal melt-bearing zones (e.g. Heise et al., 2007) imply that there exists a ‘central TVZ Batholith’. This batholith is of equivalent areal extent to the envelope enclosing silicic vents and is visible from gravity anomaly data (Bibby et al., 1995; Stagpoole & Bibby, 1999). It extends from depths as shallow as ∼4 km in places (Liu et al., 2006; Shane et al., 2008b) to about 16 km depth in its silicic to intermediate roots, as defined by crustal velocity data (Harrison & White, 2004, 2006; Stratford & Stern, 2004, 2006). The deeper mafic roots extend for another ∼15 km below the intermediate to silicic zone, from the same seismic evidence. With the information currently available, this batholith is likely to be exceptionally complex in structure and composition, yet would be manifested as a single batholith in an area such as the Sierra Nevada (see Coleman et al., 2004; Miller et al., 2007) or Colorado River Extension Corridor (see Walker et al., 2007), where the uncertainties in single age determinations match the longevities of entire volcanic centres in the TVZ, Taupo included. All indications are that single active portions of this central TVZ Batholith (whether temporary melt-dominant bodies, consanguineous mush zones of narrowly defined ages associated with single eruptive periods, or residual sub-solidus zones) have horizontal dimensions (defined by vent envelopes) that exceed their vertical extent (defined by melt-inclusion volatile contents) and are thus horizontal sheet-forming.

(2) There is a close relationship in the central TVZ between volcanism, plutonism and tectonism (Wilson et al., 2009; Rowland et al., in preparation), but this relationship is expressed in several different and contrasting ways. At the present day, there is an intense zone of geothermal fluid and heat flux in an area situated between where the loci of silicic volcanism (Okataina, Taupo) have been for the last ∼60 kyr (Hunt & Glover, 1994; Bibby et al., 1995; Wilson et al., 1995). This spatial separation implies that magmatic intrusion (to power the geothermal systems) can occur wholly independently of, and in the absence of, surface volcanism for time scales of 104–105 years or longer. In contrast, the closely related timing of volcanic episodes and tectonism (faulting and extension associated with the secular widening of the TVZ) implies that the two processes are intimately linked (e.g. Nairn & Kohn, 1973; Nairn et al., 1998, 2005; Gravley et al., 2007; Berryman et al., 2008; Rowland et al., in preparation). On a longer time scale also, Okataina and Taupo volcanoes appear to be linked by large-scale tectonic processes; for example, the period of volcanic quiescence at Taupo between ∼45 and 30 ka was when intensive activity occurred at Okataina (most of the pyroclastic deposits of the Mangaone Subgroup: Jurado-Chichay & Walker, 2000). In addition, the styles of activity and periodicity at Okataina changed at the time of the Oruanui eruption from Taupo, 80 km away. Post-Oruanui activity at Okataina has occurred at more regular intervals than prior to 27 ka, with a more uniform range of volumes, together with more consistent styles involving fissure feeders and a dominance of effusive activity (Nairn, 2002). In the presence of active rifting, and the likelihood of tectonic disruption of the mushy roots to the silicic volcanoes, time scales and rates of separation of melt-dominant bodies in the TVZ are significantly faster than standardized models of melt extraction that are limited by passive or shear-induced melt–crystal separation from a static mush (Vigneresse et al., 1996; Petford & Koenders, 2003; Bachmann & Bergantz, 2004).

(3) Episodes of intense surface volcanicity do not necessarily date intense episodes of plutonic magmatism, and vice versa, at least on the scale of tens to hundreds of thousands of years that can be discerned in the young Taupo systems (see Bachmann et al., 2007b). The ∼95 ka peak of zircon model ages common to the ‘Oruanui-type magma’ eruption units at and around Taupo volcano logically represents tappings from a substantial body of crystal mush, yet there were only minor surficial dome-building and no major pyroclastic eruptions around this time period (Houghton et al., 1991; Leonard, 2003; G. S. Leonard, personal communication, 2008). The period from ∼45 to 30 ka saw no recorded activity at Taupo, whereas at the same time large-scale generation was occurring of the Oruanui rhyolite, indicated by the 35–41 ka peak in SIMS model ages and similar values for TIMS bulk model ages. Conversely, zircons in the mush residuum of the erupted Oruanui magma, if exposed as a pluton and analysed by techniques with the analytical uncertainties of typical SIMS U–Pb age determinations, would give an average age that bore no resemblance to either the actual inferred maxima in crystallization intensity (seen from the U–Th disequilibrium model-age peaks), or the eruption age itself (see Fig. 16). Knowledge of the dynamics of the ‘volcanic–plutonic connection’ cannot be divorced from the limitations of the dating methods, and >500 km3 melt-dominant bodies that feed eruptions like the Oruanui can be generated on time scales that are too rapid for U–Pb SIMS measurements to discern.

(4) Interpretations of whether a given zircon (or any other crystal phase) within a suite of crystals analysed is phenocrystic, antecrystic or xenocrystic (Hildreth, 2004) cannot be made in isolation on model-age data alone. In essence, perceptions of ‘comagmatic’ versus ‘recycled’ zircon grains in such circumstances depend entirely on the uncertainties associated with the model-age determinations. The combination of the consistent ∼95 ka model-age populations in the ‘Oruanui-type magma’ eruptions with variably developed younger model-age peaks is used (Charlier et al., 2005) to suggest that the former crystals are recycled antecrysts, and that only the younger peak may represent phenocrysts. This is despite the fact that the model-age relationships in the Okaia versus Oruanui (Figs 9, 12 and 15) show that the younger (45–41 ka) model-age peaks in pumices from those two eruptions must have developed separately in the mush zone and not in a single melt-dominant body. In other eruptions in the Taupo area sparse zircons recovered from the strongly zircon-undersaturated 20 ka Unit Ω dacite return model ages back to 0·51 ± 0·11 Ma (Charlier et al., 2005), and the dataset from the 340 ka Whakamaru Ignimbrite (Brown & Fletcher, 1999) has grains as old as 0·52 ± 0·03 Ma (euhedral cores) and 0·61 ± 0·02 Ma (resorbed cores). The Whakamaru zircon record was interpreted (and has since been widely cited) to reflect a >250 kyr magma residence time, yet the euhedral zircons returning the same maximum ages in the Unit Ω dacite can only be xenocrystic. The term ‘magma residence time’ thus has no meaning without the broader context of eruption histories and details of chemical and isotopic relationships between magmas and mineral phases concerned.

CONCLUSIONS

A combination of detailed field and geochemical information provides a framework within which the zircon model-age spectra from eight eruption units in the Taupo area can be interpreted. Two distinct magma chambers (sensu Hildreth, 2004) were active in close proximity for at least ∼20 kyr prior to the Oruanui supereruption.

The larger and more productive chamber (‘Oruanui-type magma’) underwent a major period of crystallization centred around 95 ka, then in the period considered in this study erupted four melt-rich aliquots at ∼45 ka (twice in close succession from vents ∼20 km apart), 30 ka and 27 ka. One other eruption (‘New phreatoplinian’, also at ∼45 ka) shows links to, but distinct differences from, the closely contemporaneous pair, and cannot be from the same source. Each erupted magma contained zircons from a younger period of crystallization but that pre-dated the eruption age by 5–20 kyr. There was no continuously growing, periodically tapped melt-rich magma body, only single bodies of crystal-poor melt that accumulated shortly prior to each eruption and contained newly grown phenocrysts and entrained antecrysts from the mush zone below. By far the largest body, discharged in the 530 km3 Oruanui eruption, accumulated over only ∼3000 years (i.e. >5 m3/s); that is, in the time gap following the Okaia precursor. These two eruptions are consanguineous and share closely coincident vent areas, but have markedly different zircon model-age spectra. Most of the zircon crystals measured for this study grew in the crystal-rich mush zone, not in the melt-dominant body that eventually erupted, although the systematically younger model ages of finer size fractions of zircons in the Oruanui samples (Charlier & Zellmer, 2000), as well as the differences between TIMS mean ages and SIMS concentration-weighted means, imply that nucleation and growth of small (<63 μm) zircons continued in that body up to the point of eruption. Following the Oruanui event, renewed activity at Taupo was of distinctly different dacite and rhyolite compositions but the Oruanui zircon model-age modes remain discernible in the younger rocks, even where accompanied by abundant post-27 ka crystallization.

A smaller chamber to the NE of Lake Taupo vented at ∼47, 28 and 16 ka, producing biotite-bearing rhyolites that had undergone single cycles of crystallization, peaking 5–30 kyr prior to the eruption. All three rhyolites are crystal-moderate to crystal-rich (12–23%) and relatively cool (725–745°C: Sutton, 1995), and were generated by a process that did not incorporate large numbers of old antecrystic or xenocrystic grains. These magmas hence plausibly formed in large part by fractionation from a higher-temperature intermediate magma (see McCulloch et al., 1994; Price et al., 2005) where any inherited grains were largely stripped out.

The histories of magmatism and volcanism in the Taupo area can be seen in proper context only when a suite of stratigraphic, compositional and isotopic data are used to constrain the information from zircon model-age spectra, and vice versa. Even with the limitations imposed by SIMS analytical uncertainties, the details of crystallization histories available from the young Taupo deposits through U–Th disequilibrium model ages show that magma chambers can generate eruptible bodies at extremely high rates, plausibly related in the Taupo case to the active rift setting. The commonly held notion that geophysical data do not indicate the presence of any large bodies of melt-dominant magma below large silicic calderas worldwide (Bachmann et al., 2007b, for review) may be accurate, but of little relevance in forecasting a future eruptive event from a magma chamber, regardless of the size of either the chamber or the eruption. Meaningful discussion of the volcanic versus plutonic histories of large silicic systems may be problematic because the uncertainties associated with dating of zircons from the relevant period will strongly affect the time resolution with which comparisons can be made. For the same reason, consideration of the ‘average’ model age of zircons in a given eruption as a guide to either the lead-in time for development of the erupted magma body or for relating zircon ages to eruptive ages is not possible, unless other data are used to delineate phenocrysts from antecrysts from xenocrysts. For example, averaging the model ages of zircons we have analysed by SIMS from the Oruanui pumices generates a real but irrelevant model age, whether done by calculating a concentration-weighted mean (Table 3) or increasing the uncertainties on single data points to mimic the uncertainties associated with typical U–Pb SIMS analyses (Fig. 16).

Precursor eruptions of the ‘Oruanui-type’ magma (Tihoi, ‘New plinian’, Okaia and one other earlier event, not detailed here) do chemically and isotopically reflect the evolution of the overall Oruanui chamber but demonstrably do not represent aliquots from a growing unitary melt-dominant body because of contrasts in the model-age spectra we present here. Thus studies that envisage a supereruption as the culmination of development of a unitary melt-dominant magma body (most notably the Bishop Tuff; e.g. Hildreth, 2004; Simon & Reid, 2005; Hildreth & Wilson, 2007) and that rely on SIMS U–Pb model-age data may be missing the subtleties of the age spectra seen at Taupo that allow the physical growth of the melt-dominant body to be accurately defined, independently of the chemical processes that have led to the development of its compositional characteristics over a longer time scale. Any future study that can combine the analytical precision of ID-TIMS single-crystal ages (e.g. Crowley et al., 2007) with the utility and spatial precision of SIMS (e.g. Reid & Coath, 2000; Simon & Reid, 2005) will allow these age contrasts to be addressed.

SUPPLEMENTARY DATA

Supplementary data for this paper are available at Journal of Petrology online.

ACKNOWLEDGEMENTS

We, as always, are very grateful to Joe Wooden, Frank Mazdab, Bettina Weigand and Brad Ito for their help and guidance with U–Th dating studies at the USGS–Stanford ion probe facility. The Marsden Fund administered by the Royal Society of New Zealand (C.J.N.W.) and the UK Natural Environment Research Council (B.L.A.C.) provided financial support, and the University of Auckland Research Committee (C.J.N.W.) underwrote the ion-probe costs. Insightful reviews from Olivier Bachmann, Scott Bryan and Axel Schmitt helped clarify our thoughts, and editorial handling from John Gamble and the eagle eye of Monica Handler are much appreciated.

REFERENCES

Annen
C
Blundy
JD
Sparks
RSJ
The genesis of intermediate and silicic magmas in deep crustal hot zones
Journal of Petrology
 , 
2006
, vol. 
47
 (pg. 
505
-
539
)
Bachmann
O
Bergantz
GW
On the origin of crystal-poor rhyolites: extracted from batholithic crystal mushes
Journal of Petrology
 , 
2004
, vol. 
45
 (pg. 
1565
-
1582
)
Bachmann
O
Charlier
BLA
Lowenstern
JB
Zircon crystallization and recycling in the magma chamber of the rhyolitic Kos Plateau Tuff (Aegean arc)
Geology
 , 
2007
, vol. 
35
 (pg. 
73
-
76
)
Bachmann
O
Miller
CF
de Silva
SL
The volcanic–plutonic connection as a stage for understanding crustal magmatism
Journal of Volcanology and Geothermal Research
 , 
2007
, vol. 
167
 (pg. 
1
-
23
)
Bacon
CR
Lowenstern
JB
Late Pleistocene granodiorite source for recycled zircon and phenocrysts in rhyodacite lava at Crater Lake, Oregon
Earth and Planetary Science Letters
 , 
2005
, vol. 
233
 (pg. 
277
-
293
)
Bailey
RA
Geologic map of the Long Valley caldera, Mono–Inyo Craters volcanic chain, and vicinity, eastern California
US Geological Survey Map
 , 
1989
, vol. 
I-1933
  
scale 1:62 500
Berryman
KR
Villamor
P
Nairn
IA
Van Dissen
RJ
Begg
JG
Lee
J
Late Pleistocene surface rupture history of the Paeroa Fault, Taupo Rift, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
2008
, vol. 
51
 (pg. 
135
-
158
)
Bibby
HM
Caldwell
TG
Davey
FJ
Webb
TH
Geophysical evidence on the structure of the Taupo Volcanic Zone and its hydrothermal circulation
Journal of Volcanology and Geothermal Research
 , 
1995
, vol. 
68
 (pg. 
29
-
58
)
Bindeman
IN
Valley
JW
Wooden
JL
Persing
HM
Post-caldera volcanism: in situ measurement of U–Pb age and oxygen isotope ratio in Pleistocene zircons from Yellowstone caldera
Earth and Planetary Science Letters
 , 
2001
, vol. 
189
 (pg. 
197
-
206
)
Bindeman
IN
Fu
B
Kita
NT
Valley
JW
Origin and evolution of silicic magmatism at Yellowstone based on ion microprobe analysis of isotopically zoned zircons
Journal of Petrology
 , 
2008
, vol. 
49
 (pg. 
163
-
193
)
Blake
S
Wilson
CJN
Smith
IEM
Walker
GPL
Petrology and dynamics of the Waimihia mixed magma eruption, Taupo Volcano, New Zealand
Journal of the Geological Society, London
 , 
1992
, vol. 
149
 (pg. 
193
-
207
)
Brooker
MR
Houghton
BF
Wilson
CJN
Gamble
JA
Pyroclastic phases of a rhyolitic dome-building eruption: Puketarata tuff ring, Taupo Volcanic Zone, New Zealand
Bulletin of Volcanology
 , 
1993
, vol. 
55
 (pg. 
395
-
406
)
Brown
SJA
Fletcher
IR
SHRIMP U–Th 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
 , 
1999
, vol. 
27
 (pg. 
1035
-
1038
)
Brown
SJA
Smith
RT
Crystallisation history and crustal inheritance in a large silicic magma system: 206Pb/238U ion probe dating of zircons from the 1·2 Ma Ongatiti ignimbrite, Taupo Volcanic Zone
Journal of Volcanology and Geothermal Research
 , 
2004
, vol. 
135
 (pg. 
247
-
257
)
Brown
SJA
Wilson
CJN
Cole
JW
Wooden
J
The Whakamaru group ignimbrites, Taupo Volcanic Zone, New Zealand: evidence for reverse tapping of a zoned silicic magma system
Journal of Volcanology and Geothermal Research
 , 
1998
, vol. 
84
 (pg. 
1
-
37
)
Bryan
SE
Ferrari
L
Reiners
PW
Allen
CM
Petrone
CM
Ramos-Rosique
A
Campbell
IH
New insights into crustal contributions to large-volume rhyolite generation in the mid-Tertiary Sierra Madre Occidental province, Mexico, revealed by U–Pb geochronology
Journal of Petrology
 , 
2008
, vol. 
49
 (pg. 
47
-
77
)
Charlier
BLA
Zellmer
GF
Some remarks on U–Th mineral ages from igneous rocks with prolonged crystallisation histories
Earth and Planetary Science Letters
 , 
2000
, vol. 
183
 (pg. 
457
-
469
)
Charlier
BLA
Peate
DW
Wilson
CJN
Lowenstern
JB
Storey
M
Brown
SJA
Crystallisation ages in coeval silicic magma bodies: U–Th disequilibrium evidence from the Rotoiti and Earthquake Flat eruption deposits, Taupo Volcanic Zone, New Zealand
Earth and Planetary Science Letters
 , 
2003
, vol. 
206
 (pg. 
441
-
457
)
Charlier
BLA
Wilson
CJN
Lowenstern
JB
Blake
S
van Calsteren
PW
Davidson
JP
Magma generation at a large, hyperactive silicic volcano (Taupo, New Zealand) revealed by U–Th and U–Pb systematics in zircons
Journal of Petrology
 , 
2005
, vol. 
46
 (pg. 
3
-
32
)
Charlier
BLA
Ginibre
C
Morgan
D
Nowell
GM
Pearson
DG
Davidson
JP
Ottley
CJ
Methods for the microsampling and high-precision analysis of strontium and rubidium isotopes at single crystal scale for petrological and geochronological applications
Chemical Geology
 , 
2006
, vol. 
232
 (pg. 
114
-
133
)
Charlier
BLA
Wilson
CJN
Davidson
JP
Rapid open-system assembly of a large silicic magma body: time-resolved evidence from cored plagioclase crystals in the Oruanui eruption deposits, New Zealand
Contributions to Mineralogy and Petrology
 , 
2008
, vol. 
156
 (pg. 
799
-
813
)
Christensen
JN
DePaolo
DJ
Time scales of large volume silicic magma systems: Sr isotopic systematics of phenocrysts and glass from the Bishop Tuff, Long Valley, California
Contributions to Mineralogy and Petrology
 , 
1993
, vol. 
113
 (pg. 
100
-
114
)
Christensen
JN
Halliday
AN
Rb–Sr ages and Nd isotopic compositions of melt inclusions from the Bishop Tuff and the generation of silicic magma
Earth and Planetary Science Letters
 , 
1996
, vol. 
144
 (pg. 
547
-
561
)
Cole
JW
Brown
SJA
Burt
RM
Beresford
SW
Wilson
CJN
Lithic types in ignimbrites as a guide to the evolution of a caldera complex, Taupo volcanic centre, New Zealand
Journal of Volcanology and Geothermal Research
 , 
1998
, vol. 
80
 (pg. 
217
-
237
)
Coleman
DS
Gray
W
Glazner
AF
Rethinking the emplacement and evolution of zoned plutons: geochronologic evidence for incremental assembly of the Tuolomne Intrusive Suite, California
Geology
 , 
2004
, vol. 
32
 (pg. 
433
-
436
)
Crowley
JL
Schoene
B
Bowring
SA
U–Pb dating of zircon in the Bishop Tuff at the millenial scale
Geology
 , 
2007
, vol. 
35
 (pg. 
1123
-
1126
)
Davies
GR
Halliday
AN
Development of the Long Valley rhyolitic magmatic system: strontium and neodymium isotope evidence from glasses and individual phenocrysts
Geochimica et Cosmochimica Acta
 , 
1998
, vol. 
62
 (pg. 
3561
-
3574
)
Davy
BW
Caldwell
TG
Gravity, magnetic and seismic surveys of the caldera complex, Lake Taupo, North Island, New Zealand
Journal of Volcanology and Geothermal Research
 , 
1998
, vol. 
81
 (pg. 
69
-
89
)
Dickin
AP
Radiogenic Isotope Geology
 , 
1995
Cambridge
Cambridge University Press
Dobson
KJ
Stuart
FM
Dempster
TJ
EIMF
U and Th zonation in Fish Canyon Tuff zircons: implications for a zircon (U–Th)/He standard
Geochimica et Cosmochimica Acta
 , 
2008
, vol. 
72
 (pg. 
4745
-
4755
)
Dodson
MH
Compston
W
Williams
IS
Wilson
JF
A search for ancient detrital zircons in Zimbabwean sediments
Journal of the Geological Society, London
 , 
1988
, vol. 
145
 (pg. 
977
-
983
)
Dunbar
NW
Kyle
PR
Lack of volatile gradient in the Taupo plinian–ignimbrite transition: evidence from melt inclusion analysis
American Mineralogist
 , 
1993
, vol. 
78
 (pg. 
612
-
618
)
Ewart
A
The petrography of the central North Island rhyolitic lavas. Part 2—regional petrography including notes on associated ash-flow pumice deposits
New Zealand Journal of Geology and Geophysics
 , 
1968
, vol. 
11
 (pg. 
478
-
545
)
Gravley
DM
Wilson
CJN
Leonard
GS
Cole
JW
Double trouble: paired ignimbrite eruptions and collateral subsidence in the Taupo Volcanic Zone, New Zealand
Geological Society of America Bulletin
 , 
2007
, vol. 
119
 (pg. 
18
-
30
)
Halliday
AN
Mahood
GA
Holden
P
Metz
JM
Dempster
TJ
Davidson
JP
Evidence for long residence times of rhyolitic magma in the Long Valley magmatic system: the isotopic record in precaldera lavas of Glass Mountain
Earth and Planetary Science Letters
 , 
1989
, vol. 
94
 (pg. 
274
-
290
)
Harrison
AJ
White
RS
Crustal structure of the Taupo Volcanic Zone, New Zealand: stretching and igneous intrusion
Geophysical Research Letters
 , 
2004
, vol. 
31
  
L13615, doi:10.11029/12004GL019885
Harrison
AJ
White
RS
Lithospheric structure of an active backarc basin: the Taupo Volcanic Zone, New Zealand
Geophysical Journal International
 , 
2006
, vol. 
167
 (pg. 
968
-
990
)
Heise
W
Bibby
HM
Caldwell
TG
Bannister
SC
Ogawa
Y
Takakura
S
Uchida
T
Melt distribution beneath a young continental rift: the Taupo Volcanic Zone, New Zealand
Geophysical Research Letters
 , 
2007
, vol. 
34
  
L14313, doi:10.11029/12007GL029629
Hildreth
W
Chapin
CE
Elston
WE
The Bishop Tuff: evidence for the origin of compositional zonation in silicic magma chambers
Ash-flow Tuffs. Geological Society of America, Special Papers
 , 
1979
, vol. 
180
 (pg. 
43
-
75
)
Hildreth
W
Gradients in silicic magma chambers: implications for lithospheric magmatism
Journal of Geophysical Research
 , 
1981
, vol. 
86
 (pg. 
10153
-
10192
)
Hildreth
W
Volcanological perspectives on Long Valley, Mammoth Mountain, and Mono Craters: several contiguous but discrete systems
Journal of Volcanology and Geothermal Research
 , 
2004
, vol. 
136
 (pg. 
169
-
198
)
Hildreth
W
Wilson
CJN
Compositional zoning of the Bishop Tuff
Journal of Petrology
 , 
2007
, vol. 
48
 (pg. 
951
-
999
)
Houghton
BF
Lloyd
EF
Wilson
CJN
Lanphere
MA
K–Ar ages from the Western Dome Belt and associated rhyolitic lavas in the Maroa–Taupo area, Taupo Volcanic Zone, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
1991
, vol. 
34
 (pg. 
99
-
101
)
Houghton
BF
Wilson
CJN
McWilliams
MO
Lanphere
MA
Weaver
SD
Briggs
RM
Pringle
MS
Chronology and dynamics of a large silicic magmatic system: central Taupo Volcanic Zone, New Zealand
Geology
 , 
1995
, vol. 
23
 (pg. 
13
-
16
)
Hunt
TM
Glover
RB
Waimangu, Waiotapu, and Waikite geothermal systems. New Zealand
Geothermics
 , 
1994
, vol. 
23
 (pg. 
373
-
658
)
Ireland
TR
Williams
IS
Considerations in zircon geochronology by SIMS
Reviews in Mineralogy and Geochemistry
 , 
2003
, vol. 
53
 (pg. 
215
-
241
)
Jurado-Chichay
Z
Walker
GPL
Stratigraphy and dispersal of the Mangaone Subgroup pyroclastic deposits, Okataina Volcanic Centre, New Zealand
Journal of Volcanology and Geothermal Research
 , 
2000
, vol. 
104
 (pg. 
319
-
383
)
Kohn
BP
Topping
WW
Time–space relationships between late Quaternary rhyolitic and andesitic volcanism in the southern Taupo Volcanic Zone, New Zealand
Geological Society of America Bulletin
 , 
1978
, vol. 
89
 (pg. 
1265
-
1271
)
Leonard
GS
The evolution of Maroa Volcanic Centre, Taupo Volcanic Zone, New Zealand
 , 
2003
Ph.D. thesis, University of Canterbury
Christchurch
Leonard
GS
Cole
JW
Nairn
IA
Self
S
Basalt triggering of the c. AD 1305 Kaharoa rhyolite eruption, Tarawera Volcanic Complex, New Zealand
Journal of Volcanology and Geothermal Research
 , 
2002
, vol. 
115
 (pg. 
461
-
486
)
Liu
Y
Anderson
AT
Wilson
CJN
Davis
AM
Steele
IM
Mixing and differentiation in the Oruanui rhyolitic magma, Taupo, New Zealand: evidence from volatiles and trace elements in melt inclusions
Contributions to Mineralogy and Petrology
 , 
2006
, vol. 
151
 (pg. 
71
-
87
)
Lloyd
EF
The Geology and Hot Springs of Orakeikorako. New Zealand Geological Survey Bulletin
 , 
1972
, vol. 
85
 
Lowe
DJ
Stratigraphy, age, composition, and correlation of late Quaternary tephras interbedded with organic sediments in Waikato lakes, North Island, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
1988
, vol. 
31
 (pg. 
125
-
165
)
Lowe
DJ
Shane
PAR
Alloway
BV
Newnham
RM
Fingerprints and age models for widespread New Zealand tephra marker beds erupted since 30,000 years ago: a framework for NZ-INTIMATE
Quaternary Science Reviews
 , 
2008
, vol. 
27
 (pg. 
95
-
126
)
Lowenstern
JB
Persing
HM
Wooden
JL
Lanphere
MA
Donnelly-Nolan
J
Grove
TL
U–Th dating of single zircons from young granitoid xenoliths: new tools for understanding volcanic processes
Earth and Planetary Science Letters
 , 
2000
, vol. 
183
 (pg. 
291
-
302
)
Lowenstern
JB
Charlier
BLA
Clynne
MA
Wooden
JL
Extreme U–Th disequilibrium in rift-related basalts, rhyolites and granophyric granite and the timescale of rhyolite generation, intrusion and crystallization at Alid volcanic center, Eritrea
Journal of Petrology
 , 
2006
, vol. 
47
 (pg. 
2105
-
2122
)
Ludwig
KR
Isoplot/Ex Version 3.41, A Geochronological Toolkit for Microsoft Excel. Berkeley Geochronology Center Special Publications
 , 
2003
, vol. 
4
 
Manville
V
Wilson
CJN
Interactions between volcanism, rifting and subsidence: implications of intracaldera palaeoshorelines at Taupo volcano, New Zealand
Journal of the Geological Society, London
 , 
2003
, vol. 
160
 (pg. 
3
-
6
)
Mason
BG
Pyle
DM
Oppenheimer
C
The size and frequency of the largest explosive eruptions on Earth
Bulletin of Volcanology
 , 
2004
, vol. 
66
 (pg. 
735
-
748
)
McConnell
VS
Shearer
CK
Eichelberger
JC
Keskinen
MJ
Layer
PW
Papike
JJ
Rhyolite intrusions in the intracaldera Bishop Tuff, Long Valley caldera, California
Journal of Volcanology and Geothermal Research
 , 
1995
, vol. 
67
 (pg. 
41
-
60
)
McCulloch
MT
Kyser
TK
Woodhead
J
Kinsley
L
Pb–Sr–Nd–O isotopic constraints on the origin of rhyolites from the Taupo Volcanic Zone of New Zealand: evidence for assimilation followed by fractionation from basalt
Contributions to Mineralogy and Petrology
 , 
1994
, vol. 
115
 (pg. 
303
-
312
)
Metz
JM
Bailey
RA
Geologic map of Glass Mountain, Mono County, California
US Geological Survey Map
 , 
1993
, vol. 
1995
  
scale 1:24 000
Metz
JM
Mahood
GA
Precursors to the Bishop Tuff eruption: Glass Mountain, Long Valley, California
Journal of Geophysical Research
 , 
1985
, vol. 
90
 (pg. 
11121
-
11126
)
Metz
JM
Mahood
GA
Development of the Long Valley, California, magma chamber recorded in precaldera rhyolite lavas of Glass Mountain
Contributions to Mineralogy and Petrology
 , 
1991
, vol. 
106
 (pg. 
379
-
397
)
Miller
CF
Wark
DA
Supervolcanoes and their explosive supereruptions
Elements
 , 
2008
, vol. 
4
 (pg. 
11
-
16
)
Miller
JS
Matzel
JEP
Miller
CF
Burgess
SD
Miller
RB
Zircon growth and recycling during the assembly of large, composite arc plutons
Journal of Volcanology and Geothermal Research
 , 
2007
, vol. 
167
 (pg. 
282
-
299
)
Nairn
IA
Rotoehu Ash and the Rotoiti Breccia Formation, Taupo Volcanic Zone, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
1972
, vol. 
15
 (pg. 
251
-
261
)
Nairn
IA
Geology of the Okataina Volcanic Centre, scale 1:50,000
Institute of Geological & Nuclear Sciences geological map 25
 , 
2002
Lower Hutt: Institute of Geological & Nuclear Sciences Limited
1 sheet abd
pg. 
156
 
Nairn
IA
Kohn
BP
Relation of the Earthquake Flat Breccia to the Rotoiti Breccia, Central North Island, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
1973
, vol. 
16
 (pg. 
269
-
279
)
Nairn
IA
Kobayashi
T
Nakagawa
M
The ∼10 ka multiple vent pyroclastic eruption sequence at Tongariro Volcanic Centre, Taupo Volcanic Zone, New Zealand: Part 1. Eruptive processes during regional extension
Journal of Volcanology and Geothermal Research
 , 
1998
, vol. 
86
 (pg. 
19
-
44
)
Nairn
IA
Shane
PR
Cole
JW
Leonard
GJ
Self
S
Pearson
N
Rhyolite magma processes of the ∼AD1315 Kaharoa eruption episode, Tarawera volcano, New Zealand
Journal of Volcanology and Geothermal Research
 , 
2004
, vol. 
131
 (pg. 
265
-
294
)
Nairn
IA
Hedenquist
JW
Villamor
P
Berryman
KR
Shane
PA
The ∼AD1315 Tarawera and Waiotapu eruptions, New Zealand: contemporaneous rhyolite and hydrothermal eruptions driven by an arrested basalt dike system?
Bulletin of Volcanology
 , 
2005
, vol. 
67
 (pg. 
186
-
193
)
Otway
PM
Blick
GH
Scott
BJ
Vertical deformation at Lake Taupo, New Zealand, from lake levelling surveys, 1979–99
New Zealand Journal of Geology and Geophysics
 , 
2002
, vol. 
45
 (pg. 
121
-
132
)
Petford
N
Koenders
MA
Shear-induced pressure changes and seepage phenomena in a deforming porous layer—I
Geophysical Journal International
 , 
2003
, vol. 
155
 (pg. 
857
-
869
)
Pillans
BJ
Kohn
BP
Berger
GW
Froggatt
PC
Duller
GAT
Alloway
BV
Hesse
PP
Multi-method dating comparison for mid-Pleistocene Rangitawa Tephra, New Zealand
Quaternary Science Reviews
 , 
1996
, vol. 
15
 (pg. 
641
-
653
)
Price
RC
Gamble
JA
Smith
IEM
Stewart
RB
Eggins
SM
Wright
IC
An integrated model for the temporal evolution of andesites and rhyolites and crustal development in New Zealand's North Island
Journal of Volcanology and Geothermal Research
 , 
2005
, vol. 
140
 (pg. 
1
-
24
)
Reid
MR
How long does it take to supersize an eruption?
Elements
 , 
2008
, vol. 
4
 (pg. 
23
-
28
)
Reid
MR
Coath
CD
In situ U–Pb ages of zircons from the Bishop Tuff: no evidence for long crystal residence times
Geology
 , 
2000
, vol. 
28
 (pg. 
443
-
446
)
Reid
MR
Coath
CD
Harrison
TM
McKeegan
KD
Prolonged residence times for the youngest rhyolites associated with Long Valley caldera: 230Th–238U ion microprobe dating of young zircons
Earth and Planetary Science Letters
 , 
1997
, vol. 
150
 (pg. 
27
-
39
)
Rudge
JF
Finding peaks in geochemical distributions: a re-examination of the helium–continental crust correlation
Earth and Planetary Science Letters
 , 
2008
, vol. 
274
 (pg. 
179
-
188
)
Sambridge
MS
Compston
W
Mixture modeling of multi-component data sets with application to ion-probe zircon ages
Earth and Planetary Science Letters
 , 
1994
, vol. 
128
 (pg. 
373
-
390
)
Sandiford
A
Alloway
BV
Shane
P
A 28,000–6600 cal yr record of local and distal volcanism preserved in a paleolake, Auckland, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
2001
, vol. 
44
 (pg. 
323
-
336
)
Sandiford
A
Horrocks
M
Newnham
RM
Ogden
J
Alloway
BV
Environmental change during the Last Glacial Maximum (ca. 25,500–16,500 years BP) at Mt Richmond, Auckland Isthmus, New Zealand
Journal of the Royal Society of New Zealand
 , 
2002
, vol. 
32
 (pg. 
155
-
167
)
Schmitt
AK
Stockli
DF
Hausback
BP
Eruption and magma crystallization ages of Las Tres Vírgines (Baja California) constrained by combined 230Th/238U and (U–Th)/He dating of zircon
Journal of Volcanology and Geothermal Research
 , 
2006
, vol. 
158
 (pg. 
281
-
295
)
Schmitz
MD
Smith
IEM
The petrology of the Rotoiti eruption sequence, Taupo Volcanic Zone: an example of fractionation and mixing in a rhyolitic system
Journal of Petrology
 , 
2004
, vol. 
45
 (pg. 
2045
-
2066
)
Self
S
Sparks
RSJ
Characteristics of widespread pyroclastic deposits formed by the interaction of silicic magma and water
Bulletin Volcanologique
 , 
1978
, vol. 
41
 (pg. 
196
-
212
)
Shane
P
Martin
SB
Smith
VC
Beggs
KF
Darragh
MB
Cole
JW
Nairn
IA
Multiple rhyolite magmas and basalt injection in the 17·7 ka Rerewhakaaitu eruption episode from Tarawera volcanic complex, New Zealand
Journal of Volcanology and Geothermal Research
 , 
2007
, vol. 
164
 (pg. 
1
-
26
)
Shane
P
Nairn
IA
Smith
VC
Darragh
M
Beggs
KF
Cole
JW
Silicic recharge of multiple rhyolite magmas by basaltic intrusion during the 22·6 ka Okareka eruption episode, New Zealand
Lithos
 , 
2008
, vol. 
103
 (pg. 
527
-
549
)
Shane
P
Smith
VC
Nairn
IA
Millennial timescale resolution of rhyolite magma recharge at Tarawera volcano: insights from quartz chemistry and melt inclusions
Contributions to Mineralogy and Petrology
 , 
2008
, vol. 
156
 (pg. 
397
-
411
)
Shaw
HR
Links between magma–tectonic rate balances, plutonism, and volcanism
Journal of Geophysical Research
 , 
1985
, vol. 
90
 (pg. 
11275
-
11288
)
Simon
JI
Reid
MR
The pace of rhyolite differentiation and storage in an ‘archetypical’ silicic magma system, Long Valley, California
Earth and Planetary Science Letters
 , 
2005
, vol. 
235
 (pg. 
123
-
140
)
Simon
JI
Renne
PR
Mundil
R
Implications of pre-eruptive magmatic histories of zircons for U–Pb geochronology of silicic extrusions
Earth and Planetary Science Letters
 , 
2008
, vol. 
266
 (pg. 
182
-
194
)
Smith
EGC
Williams
TD
Darby
DJ
Principal component analysis and modeling of the subsidence of the shoreline of Lake Taupo, New Zealand, 1983–1999: evidence for the dewatering of a magmatic intrusion?
Journal of Geophysical Research
 , 
2007
, vol. 
112
  
B08406, doi:10.01029/02006JB004652
Smith
RL
Chapin
CE
Elston
WE
Ash-flow magmatism
Ash-flow Tuffs. Geological Society of America, Special Papers
 , 
1979
, vol. 
180
 (pg. 
5
-
27
)
Smith
V
Shane
P
Geochemical characteristics of the widespread Tahuna Tephra
New Zealand Journal of Geology and Geophysics
 , 
2002
, vol. 
45
 (pg. 
103
-
107
)
Smith
VC
Shane
P
Nairn
IA
Reactivation of a rhyolitic magma body by new rhyolitic intrusion before the 15·8 ka Rotorua eruptive episode: implications for magma storage in the Okataina Volcanic Centre, New Zealand
Journal of the Geological Society, London
 , 
2004
, vol. 
161
 (pg. 
757
-
772
)
Sparks
RSJ
Self
S
Grattan
JP
Oppenheimer
C
Pyle
DM
Rymer
H
Super-eruptions: global effects and future threats: Report of a Geological Society of London Working Group
 , 
2005
2nd
London
Geological Society
pg. 
24
 
Spell
TL
Kyle
PR
Petrogenesis of Valle Grande Member rhyolites, Valles caldera, New Mexico: implications for evolution of the Jemez Mountains magmatic system
Journal of Geophysical Research
 , 
1989
, vol. 
94
 (pg. 
10379
-
10396
)
Spell
TL
Harrison
TM
Wolff
JA
40Ar/39Ar dating of the Bandelier Tuff and San Diego Canyon ignimbrites, Jemez Mountains, New Mexico: temporal constraints on magmatic evolution
Journal of Volcanology and Geothermal Research
 , 
1990
, vol. 
43
 (pg. 
175
-
193
)
Spell
TL
McDougall
I
Doulgeris
AP
Cerro Toledo Rhyolite, Jemez Volcanic Field, New Mexico: 40Ar/39Ar geochronology of eruptions between two caldera-forming events
Geological Society of America Bulletin
 , 
1996
, vol. 
108
 (pg. 
1549
-
1566
)
Spera
FJ
Crisp
JA
Eruption volume, periodicity and caldera area: relationships and interferences on development of compositional zonation in silicic magma chambers
Journal of Volcanology and Geothermal Research
 , 
1981
, vol. 
11
 (pg. 
169
-
187
)
Stagpoole
VM
Bibby
HM
Residual gravity anomaly map of the Taupo Volcanic Zone, New Zealand, 1:250 000
Lower Hutt: Institute of Geological & Nuclear Sciences, Geophysical Map
 , 
1999
, vol. 
13
 
Stix
J
Gorton
MP
Replenishment and crystallization in epicontinental silicic magma chambers: evidence from the Bandelier magmatic system
Journal of Volcanology and Geothermal Research
 , 
1993
, vol. 
55
 (pg. 
201
-
215
)
Stix
J
Goff
F
Gorton
MP
Heiken
G
Garcia
SR
Restoration of compositional zonation in the Bandelier silicic magma chamber between two caldera-forming eruptions: geochemistry and origin of the Cerro Toledo Rhyolite, Jemez Mountains, New Mexico
Journal of Geophysical Research
 , 
1988
, vol. 
93
 (pg. 
6129
-
6147
)
Stratford
WR
Stern
TA
Strong seismic reflections and melts in the mantle of a continental back-arc basin
Geophysical Research Letters
 , 
2004
, vol. 
31
  
L06622, doi:10.01029/02003GL019232
Stratford
WR
Stern
TA
Crust and upper mantle structure of a continental backarc: central North Island, New Zealand
Geophysical Journal International
 , 
2006
, vol. 
166
 (pg. 
469
-
484
)
Sutton
AN
Evolution of a large silicic magma system: Taupo volcanic centre, New Zealand
 , 
1995
Ph.D. thesis, The Open University
Milton Keynes
Sutton
AN
Blake
S
Wilson
CJN
An outline geochemistry of rhyolite eruptives from Taupo volcanic centre, New Zealand
Journal of Volcanology and Geothermal Research
 , 
1995
, vol. 
68
 (pg. 
153
-
175
)
Sutton
AN
Blake
S
Wilson
CJN
Charlier
BLA
Late Quaternary evolution of a hyperactive rhyolite magmatic system: Taupo volcanic centre, New Zealand
Journal of the Geological Society, London
 , 
2000
, vol. 
157
 (pg. 
537
-
552
)
Trial
AF
Spera
FJ
Mechanisms for the generation of compositional heterogeneities in magma chambers
Geological Society of America Bulletin
 , 
1990
, vol. 
102
 (pg. 
353
-
367
)
Vazquez
JA
Reid
MR
Probing the accumulation history of the voluminous Toba magma
Science
 , 
2004
, vol. 
305
 (pg. 
991
-
994
)
Vigneresse
JL
Barbey
P
Cuney
M
Rheological transitions during partial melting and crystallization with application to felsic magma segregation and transfer
Journal of Petrology
 , 
1996
, vol. 
37
 (pg. 
1579
-
1600
)
Vucetich
CG
Howorth
R
Late Pleistocene tephrostratigraphy in the Taupo District, New Zealand
New Zealand Journal of Geology and Geophysics
 , 
1976
, vol. 
19
 (pg. 
51
-
69
)
Walker
BA
Miller
CF
Claiborne
LL
Wooden
JL
Miller
JS
Geology and geochronology of the Spirit Mountain batholith, southern Nevada: implications for timescales and physical processes of batholith construction
Journal of Volcanology and Geothermal Research
 , 
2007
, vol. 
167
 (pg. 
239
-
262
)
White
SM
Crisp
JA
Spera
FJ
Long-term volumetric eruption rates and magma budgets
Geochemistry, Geophysics, Geosystems
 , 
2006
, vol. 
7
  
Q03010, doi:10.01029/02005GC001002
Wilson
CJN
Stratigraphy, chronology, styles and dynamics of late Quaternary eruptions from Taupo volcano, New Zealand
Philosophical Transactions of the Royal Society of London, Series A
 , 
1993
, vol. 
343
 (pg. 
205
-
306
)
Wilson
CJN
The 26·5 ka Oruanui eruption, New Zealand: an introduction and overview
Journal of Volcanology and Geothermal Research
 , 
2001
, vol. 
112
 (pg. 
133
-
174
)
Wilson
CJN
Houghton
BF
Lloyd
EF
Smith
IEM
Volcanic history and evolution of the Maroa–Taupo area, central North Island
Late Cenozoic Volcanism in New Zealand. Royal Society of New Zealand, Bulletin
 , 
1986
, vol. 
23
 (pg. 
194
-
223
)
Wilson
CJN
Switsur
VR
Ward
AP
A new 14C age for the Oruanui (Wairakei) eruption, New Zealand
Geological Magazine
 , 
1988
, vol. 
125
 (pg. 
297
-
300
)
Wilson
CJN
Houghton
BF
McWilliams
MO
Lanphere
MA
Weaver
SD
Briggs
RM
Volcanic and structural evolution of Taupo Volcanic Zone, New Zealand: a review
Journal of Volcanology and Geothermal Research
 , 
1995
, vol. 
68
 (pg. 
1
-
28
)
Wilson
CJN
Blake
S
Charlier
BLA
Sutton
AN
The 26·5 ka Oruanui eruption, Taupo volcano, New Zealand: development, characteristics and evacuation of a large rhyolitic magma body
Journal of Petrology
 , 
2006
, vol. 
47
 (pg. 
35
-
69
)
Wilson
CJN
Rhoades
DA
Lanphere
MA
Calvert
AT
Houghton
BF
Weaver
SD
Cole
JW
A multiple-approach radiometric age estimate for the Rotoiti and Earthquake Flat eruptions, New Zealand, with implications for the MIS 4/3 boundary
Quaternary Science Reviews
 , 
2007
, vol. 
26
 (pg. 
1861
-
1870
)
Wilson
CJN
Charlier
BLA
Fagan
CJ
Spinks
KD
Gravley
DM
Simmons
SF
Browne
PRL
U–Pb dating of zircon in hydrothermally altered rocks as a correlation tool: application to the Mangakino geothermal field, New Zealand
Journal of Volcanology and Geothermal Research
 , 
2008
, vol. 
176
 (pg. 
191
-
198
)
Wilson
CJN
Gravley
DM
Leonard
GS
Rowland
JV
Thordarson
T
Larsen
G
Self
S
Rowland
SK
Hoskuldsson
A
Volcanism in the central Taupo Volcanic Zone, New Zealand: tempo styles and controls
Studies in Volcanology: The Legacy of George Walker. Special Publications of IAVCEI
 , 
2009
, vol. 
2
 
London
Geological Society
(pg. 
225
-
247
)
Wolff
JA
Ramos
FC
Pb isotope variations among Bandelier Tuff feldspars: no evidence for a long-lived silicic magma chamber
Geology
 , 
2003
, vol. 
31
 (pg. 
533
-
536
)
Wolff
JA
Ramos
FC
Davidson
JP
Sr isotope disequilibrium during differentiation of the Bandelier Tuff: constraints on the crystallization of a large rhyolitic magma chamber
Geology
 , 
1999
, vol. 
27
 (pg. 
495
-
498
)