Journal of Petrology Advance Access originally published online on August 12, 2004
Journal of Petrology 2004 45(10):2045-2066; doi:10.1093/petrology/egh047
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology 45(10) © Oxford University Press 2004; all rights reserved
The Petrology of the Rotoiti Eruption Sequence, Taupo Volcanic Zone: an Example of Fractionation and Mixing in a Rhyolitic System
1 DEPARTMENT OF GEOSCIENCES, BOISE STATE UNIVERSITY, BOISE, ID 83725-1535, USA
2 DEPARTMENT OF GEOLOGY, UNIVERSITY OF AUCKLAND, PB92019, AUCKLAND, NEW ZEALAND
RECEIVED MARCH 26, 2002; ACCEPTED MAY 26, 2004
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONThe Rotoiti eruption from the Taupo Volcanic Zone (TVZ) in northern New Zealand produced voluminous pyroclastic deposits. The ferromagnesian mineral assemblage in these dominantly consists of cummingtonite + hornblende + orthopyroxene with uniform magnesium/iron ratios; a second assemblage of biotite + hornblende + orthopyroxene, also with uniform Fe/Mg ratios, appears midway through the eruption sequence and, thereafter, increases in abundance. These contrasting mineral assemblages, together with pumice clast and groundmass glass compositions, provide evidence for mingling of two discrete magmas. Similarities in the chemical characteristics of the two magmas suggest that they developed from a similar source. The eruption initially tapped relatively homogeneous magma that was erupted throughout most of this phase of activity. The middle stages of the eruption included some mixed magma. The final stages of the eruption were dominated by a second magma composition, which was probably injected into the bottom of the main magma body as the eruption proceeded. The source that fed the eruption was complex, and discrete magma bodies existed and evolved separately prior to the eruption. We conclude that eruptions in the TVZ are fed from a diffuse upper-crustal zone of partially interconnected, and at times physically separate, magma bodies rather than from centralized and necessarily large long-lived magma chambers.
KEY WORDS: Taupo Volcanic Zone; Okataina Volcanic Centre; Rotoiti eruption; rhyolite system; magma mixing
|
INTRODUCTION |
|---|
Detailed studies of the petrology of silicic eruption sequences provide a window into the processes that produce and modify magmas within the Earth's crust. This has been recognized in the seminal studies of, for example, Smith & Bailey (1966)
, Hildreth (1981) and Flood et al. (1989). However, despite intense study during the past two decades, the nature of silicic magma chambers remains uncertain. The potential variety of sources and diversity of liquidus phase assemblages of magmas, their interpreted P–T–X conditions and their variable residence times within the upper crust make these silicic bodies more complicated than their mafic counterparts. Complex processes of crystal–liquid and liquid–liquid differentiation, diffusion, volatile complexing, convection, magma–magma and magma–fluid mixing and unmixing all have the potential to modify the composition of the melts. It has become clear that there is no one paradigm for silicic magma systems, occurring as they do in diverse tectonic settings throughout the geological record. Understanding silicic magma chambers depends on detailed petrological study of stratigraphically well-understood samples together with interpretation of their variations in terms of magmatic processes. Over the past 25 years, studies of rhyolitic pyroclastic deposits have recognized abundant evidence for compositional, mineralogical and thermal zoning in silicic deposits that provide a key to these processes.
The tectonic setting, and the geophysical, geochemical and volcanological characteristics of the Taupo Volcanic Zone (TVZ) of central North Island, New Zealand, have been broadly described by Wilson et al. (1984), Walcott (1989), Hochstein et al. (1993), Graham et al. (1995) and Houghton et al. (1995). The TVZ hosts a voluminous record of Quaternary volcanism. Continuing eruptive activity and excellent preservation of the recent products of volcanism provide a unique opportunity to study, in close detail, the conditions and processes within rhyolitic magma chambers prior to and during eruption. In contrast to most rhyolitic provinces, studies of the caldera volcanoes of the TVZ have found only limited geochemical variation within the voluminous pyroclastic deposits of the zone [e.g. Dunbar et al., 1989; Hochstein et al., 1993; Sutton et al., 1995, 2000; however, see Briggs et al. (1993) for exceptions in the older Mangakino volcanic centre]. This general lack of compositional zonation within TVZ eruption sequences has been interpreted as evidence for chemical and thermal homogeneity in the rhyolitic magma chambers, broadly attributed to the high heat flow in the central TVZ and short residence times of magmas in the crust (Hochstein et al., 1993; Houghton et al., 1995).
The Okataina Volcanic Centre is one of two currently active rhyolitic volcanoes in the TVZ and has been closely studied in terms of its stratigraphy and physical volcanology (Nairn, 1972, 1981, 1989, 2002; Pullar & Nairn, 1972; Nairn & Kohn, 1973). These detailed studies provide an excellent framework within which to conduct geochemical studies of the petrogenesis and evolution of rhyolitic magma chambers. Geochemical work on the deposits of the Okataina Volcanic Centre show that they are dominated by eruptions of high-silica rhyolite with less common dacite and rhyolite (Nairn, 1992; Bailey & Carr, 1994; Jurado-Chichay & Walker, 2000; Smith et al., 2002). Subtle trace element variations have been interpreted to suggest that discrete batches of magma are erupted over periods of up to 20 kyr (Hochstein et al., 1993).
The most recent of a series of caldera-forming events that defined the basic architecture of Okataina Volcano is represented by a sequence of pyroclastic deposits known as the Rotoiti pyroclastics. These were produced by multi-vent eruptions and followed the major eruptive events that produced the Rangitaiki Ignimbrite (360 ± 30 ka; Nairn, 1989) and the Matahina Ignimbrite (280 ± 30 ka; Bailey & Carr, 1994). The widely accepted age for the Rotoiti eruption is 64 ± 4 ka (Wilson et al., 1992) but this is controversial (see Charlier et al., 2003) and ages as young as 44 ka have been proposed (Shane & Sandiford, 2003). The relatively large Rotoiti eruption was contemporaneous with, or closely followed by, a relatively small caldera-forming eruption on the southern flank of the Okataina Caldera that is known from deposits termed Earthquake Flat Breccia. Subsequent eruptions in the Okataina Volcanic Centre during the last 50 kyr have produced the small-volume pyroclastic and lava complexes of the Mangaone Tephra sequence and Haroharo and Tarawera dome complexes.
Here we present a detailed study of the petrology of the Rotoiti pyroclastics. These deposits provide an instantaneous picture of a moderate-sized, upper-crustal, felsic magmatic system at the time of its eruption.
|
THE ROTOITI PYROCLASTICS |
|---|
The Rotoiti pyroclastics (Nairn, 2002
) are a diverse package of pyroclastic fall and flow deposits including the Rotoiti Breccia Formation (Healy et al., 1964), which is also referred to as the Rotoiti Ignimbrite (Nairn, 1981), the underlying and mantling Rotoehu Ash Member (Nairn, 1972), and the Matahi Basalt Tephra member (Pullar & Nairn, 1972); all of these have been grouped as the Rotoiti Tephra Formation by Froggatt & Lowe (1990). Although distinctions between pyroclastic flow and fall deposits are important in studies of eruptive style and rates of eruption they can obscure petrogenetic relationships in geochemical studies. In this paper fall and flow deposits are considered equivalent in preserving the record of subterranean magmatic processes, and the term Rotoiti pyroclastics is used to refer to the rhyolitic fall beds and flows as a group.
Outcrops of the Rotoiti pyroclastics are widespread between Rotorua and the Bay of Plenty (Fig. 1). Of the three constituent members, the basaltic Matahi Scoria has the smallest dispersal (
500 km2) and volume (<1 km3) (Froggatt & Lowe, 1990). The major part of the volume of the eruption is represented by pyroclastic flow deposits of the Rotoiti Ignimbrite, which occur over an area of >850 km2 north and east of its source in the Haroharo Caldera. Estimates of their volume converge at >50 km3 (Nairn, 1981; Froggatt & Lowe, 1990). The pyroclastic fall deposits that collectively make up the Rotoehu Tephra member occur underlying, intercalated with and mantling the Rotoiti Ignimbrite, and were originally dispersed over most of North Island. Conservative volume estimates of these range from >50 to 90 km3 (Nairn, 1981; Froggatt & Lowe, 1990). The distribution of the rhyolitic Rotoiti pyroclastics suggests that they were erupted from a vent lineament in the northern part of Haroharo Caldera from vents that are now concealed by younger lavas and pyroclastic deposits (Nairn, 1989). The basaltic Matahi Scoria member may also have been erupted from this vent lineament; however, a lack of evidence for mixing of basaltic and rhyolitic magma types strongly suggests distinct vents for the contrasting magma types.
|
Internal stratigraphy
The stratigraphy of the Rotoiti pyroclastics is well known from the work of Healy et al. (1964)
, Ewart & Healy (1965), Thompson (1968), Vucetich & Pullar (1969), Nairn (1972, 1981, 1989), Pullar & Nairn (1972), Walker (1979), Davis (1985) and Froggatt & Lowe (1990). This stratigraphic succession is here subdivided into five informal units on the basis of depositional style and mineralogy (Fig. 2).
|
The Rotoiti eruptive episode commenced with eruption of the sub-plinian Matahi Scoria Member. This tephra deposit is composed of small (<2 cm) vesicular basaltic scoria lapilli and ash conformably overlain by rhyolitic plinian and phreatoplinian fall deposits of the lower Rotoehu Tephra member (Unit 1), which mark the commencement of the felsic phases of the eruption. A lack of palaeosol development between the Matahi Scoria Member and the lowest Rotoehu Tephra member indicates that only a very short time elapsed between basalt and rhyolite eruptions. However, there is no rhyolitic ash interspersed within the basaltic ash and neither is there basaltic material found in the overlying rhyolitic ash beds. There is also no evidence for mingling of the contrasting magmas. These observations suggest a complete cessation of the basaltic eruption prior to beginning of the rhyolite eruption.
The pyroclastic fall beds of Unit 1 accumulated to a thickness of 4 m in proximal localities, accompanied by intraplinian fine-grained pyroclastic flows of limited (proximal) extent. In addition to plagioclase, quartz and Fe–Ti oxides, these fall and flow deposits contain a ferromagnesian assemblage of cummingtonite, hypersthene and hornblende. At the top of Unit 1 lie very coarse crystal-rich fall beds whose lithology and wide dispersal have been described by Walker (1979) and Nairn (1981).
The Unit 1 pyroclastic fall deposits are overlain by thick (>50 m) multiple unwelded pyroclastic flows of Rotoiti Ignimbrite (informally termed Unit 2). These are pink or buff deposits consisting of unsorted pumice clasts, crystals and lithic fragments in a fine-grained ash matrix. The pumice clasts are variable in size (<1 to >40 cm diameter) and are concentrated in discontinuous layers and pods that probably signify the tops of individual flows. In proximal exposures, fall beds similar to those underlying the flows are interbedded with Unit 2. A 1–2 m horizon of lithic-rich co-ignimbrite lag breccia near the top of Unit 2 is also exposed in proximal outcrops on the northern shores of Lake Rotoiti and nearby Lake Rotoehu. This bed contains a diverse population of sub-angular to rounded, large (up to 1 m in diameter) lithic clasts including rhyolite, vein quartz and welded vitric tuffs; it is interpreted to reflect vent widening associated with caldera collapse during the last phase of Unit 2. These lithic fragments have been described in detail by Brown et al. (1998) and Burt et al. (1998).
The upper units of the Rotoiti sequence (Units 3–5) are distinguished from preceding units by the occurrence of biotite in the phenocryst assemblage. Unit 3 consists of multiple flows separated by pumice concentration zones and co-ignimbrite ashes. Mantling the uppermost Unit 3 pyroclastic flows are crystal-rich pyroclastic fall beds similar to Unit 1 beds except for the notable presence of biotite and an increase in the content of hornblende at the expense of cummingtonite. The change in ferromagnesian mineralogy culminates in the thin uppermost bed of Unit 4, which consists of biotite, hornblende and hypersthene without cummingtonite.
Unit 5 of the Rotoiti pyroclastics consists of finely bedded well-sorted, often cross-bedded ash and pumice beds. They clearly represent water reworked Unit 3 and 4 material as evidenced by cut and fill structures, heavy mineral accumulations, trains of pumice clasts and their finely laminated nature.
|
SAMPLING, ANALYTICAL METHODS AND CONVENTIONS |
|---|
Bulk samples of flow and fall deposits were collected from outcrops throughout the Bay of Plenty region (Fig. 1). In addition, medium to large pumice clasts (>15 cm) were collected from pyroclastic flow deposits for whole-clast analysis. All samples were well known stratigraphically. Sample numbers (prefixed AU) refer to material housed in the petrological research collection at The University of Auckland. Representative analytical data are presented in Tables 1 and 2. A complete dataset is available at http://www.petrology.oupjournals.org.
|
|
Pumice clasts were prepared for analysis in a uniform manner to ensure internal consistency. Pumices ranging between 15 and 40 cm diameter were dried at 80°C for 6 h prior to rough splitting down to 10 cm fragments to remove outer weathered material. The fresh pumice fragments were soaked in distilled water for 6 h, rinsed and dried for 4 days at 80°C. These clean fragments were crushed between tungsten carbide plates and a 100 g aliquot of each sample was ground to <200 mesh in a tungsten carbide ring grinder. Major element concentrations were determined by X-ray fluorescence (XRF) using standard techniques using a PW1410 spectrometer. Matrix correction procedures for major elements followed Norrish & Hutton (1969)
. Trace element concentrations were determined using a Seimens SRS 3000 spectrometer. A suite of 36 international standards were used for calibration. Data reduction used Siemens SPECTRA 3000 software; the Compton scatter of X-ray tube line Rh Kß1 was used to correct for mass attenuation and appropriate corrections were used for those elements analysed at energies below the Fe absorption edge. Precision is better than 1% for Sr and Zr, 1–3% for V, Cr, Zn and Y, 3–5% for Ba, 5–10% for Rb and Nb, 10–20% for Pb, and >20% for La and Th. Detection limits are <2 ppm for Rb, Sr, Y, Zr and Nb, 2–5 ppm for V, Cr, Zn, La, Pb and Th, and 5–10 ppm for Ba.
Mineral and glass components were separated from crushed pumice clasts and matrix samples, collected from fall and flow deposits, using standard heavy liquid and magnetic separation techniques. Chemical analyses were determined by energy dispersive electron microprobe. The system used at the University of Auckland was a JEOL JXA-5A instrument using a LINK systems LZ-5 detector, QX-2000 pulse processor and ZAF-4/FLS matrix correction software. An accelerating voltage of 15 kV, beam current of 0·5 nA, beam diameter of 5 µm and live count times of 100 s were used for all mineral analyses. Analytical precision was estimated by replicate analyses of mineral standards as (
) 
3% for elements present in abundances >1 wt %. For glass analyses conditions were modified by the use of a defocused (10 µm diameter) beam to minimize time-dependent migration of sodium. Melt inclusion glasses were analysed by wavelength-dispersive methods using the JEOL JXA733 Superprobe at the Analytical Facility in the Department of Geology, Victoria University of Wellington. Procedures in this case followed those recommended by Froggatt (1992), using a reduced beam current of 8 nA, initial counting of sodium for 3 x 10 s intervals and a defocused 10 µm diameter electron beam. Extensive collection and comparison of analyses of natural and synthetic glass standards between the electron microprobes at the Victoria University of Wellington and the University of Auckland were made to ensure equivalence of data.
Minor trace element concentrations were obtained by solution inductively coupled plasma mass spectrometry (ICP-MS) on the VG PQ 2+S at the Massachusetts Institute of Technology (MIT); 50 mg of each sample powder were completely dissolved in a mix of HF–HNO3 in pressure vessels at 220°C for 5 days, dried and redissolved in 7 M HNO3, spiked with internal standards of Se, In, and Bi, diluted by a factor of 1000, and analysed in triplicate. Following internal and external instrumental drift corrections, blank and isobaric interference corrections, concentrations were calibrated against US Geological Survey (USGS) whole-rock standards (BIR-1, BHVO-1, BCR-1, AGV-1) and are inferred to be accurate and precise to ±2–5% based on replicate analyses of blind standards.
Additional rare earth element (REE) concentrations of selected samples were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) in the Department of Geology at the University of Auckland. Approximately 2 g of rock powder were digested in several stages using a combination of acids in a sealed bomb. A REE concentrate was produced using cation exchange chromatography and the resulting solution analysed using an ARL 3410 spectrometer. Comparisons between the MIT and Auckland datasets shows consistent abundances and trends.
For isotopic analyses, 100 mg of each sample powder were completely dissolved in 3 ml of a 2:1 mix of HF–HNO3 in pressure vessels at 220°C for 5 days, dried and redissolved in 5 ml of 6 M HCl at 120°C overnight. The clear solutions were aliquoted 2:1:1 for Nd, Sr and Pb analysis, respectively. For Pb separation, the sample aliquot was dried and redissolved in 1·1 M HBr for purification of Pb by conventional HBr–HCl-based anion-exchange chemistry [modified from Strelow & von Toerien (1966)]. The separation and purification of Nd was accomplished with a standard two-stage dilute HCl cation exchange–HDEHP (hydrogen-diethyl-hexyl-phosphate) reverse chromatography procedure [modified from Richard et al. (1976)]. The Sr aliquot was dried and redissolved in 3·5 M HNO3 for purification on Sr-specific crown ether resin [after Pin & Bassin (1992)]. Nd, Sr and Pb isotopes were analysed by conventional thermal-ionization mass spectrometry on the MIT VG Sector 54 multicollector mass spectrometer. Total procedural blanks for Nd, Sr, and Pb were all measured as <20 pg, and thus represented a negligible contribution to the sample isotopes. Lead was loaded with a silica gel–phosphoric acid mixture (Gerstenberger & Haase, 1997) on previously degassed single Re filaments, and its isotopes were measured in static mode on Faraday cups, with 207Pb ion beams <10–11 A. Pb isotopic fractionation [0·1 ± 0·03%/a.m.u. (2)] was monitored and corrected throughout the course of the study by daily analysis of the NBS-981 common Pb standard. Purified Sr was loaded with 1 µl 1 M H3PO4 + 1 µl TaCl5 solution onto single degassed Re filaments, and its isotopes were analysed in dynamic mode. The Sr isotopic composition was fractionation corrected with an exponential law relative to 86Sr/88Sr = 0·1194, and is reported bias corrected relative to the accepted value of the NBS-987 standard (0·71025). The external reproducibility of the NBS-987 standard over the course of the study was 0·71026 ± 3 (2). Purified Nd was loaded on triple Re filaments with 1 µl of 0·1 M H3PO4 and analysed as metal ions in dynamic multicollector mode with a 144Nd ion beam of 1·5 x 10–10 A. Nd isotope ratios were fractionation corrected with an exponential law, normalizing to 146Nd/144Nd = 0·7219. The long-term reproducibility of Nd isotopic standards at MIT is 20 ppm (2); multiple analyses (n = 12) of USGS standard BCR-1 during the course of the study yielded 143Nd/144Nd = 0·512643 ± 9 (2 S.D.); present-day 
Nd, tCHUR and tDM were calculated with (147Sm/144Nd)CHUR = 0·1967, (143Nd/144Nd)CHUR = 0·512638, (147Sm/144Nd)DM = 0·2137, and (143Nd/144Nd)DM = 0·513151.
Pumice clasts within the Rotoiti deposits are rhyolite to high-Si rhyolite with a small range in silica content (72–80 wt %). Because of the tendency of the glass component of the pumices to hydrate with time they show variably high loss on ignition (LOI = 1·36–4·31 wt %). Major element compositions are presented normalized to 100% volatile free in all diagrams to compensate for the diluting effect of hydration.
|
MINERAL AND GLASS CHARACTERISTICS AND COMPOSITIONS |
|---|
Matahi Scoria Member
The Matahi Scoria is composed of basaltic andesite ash and scoriaceous clasts. Included particles of rhyolitic pumice and feldspar xenocrysts represent underlying material through which the magma passed and not Rotoiti material. The scoria clasts comprise small (<0·5 mm) plagioclase (An46–73), olivine (Fo74–76) and clinopyroxene (En45–Fs14–Wo41) phenocrysts set in a dense micro-vesicular, glassy, groundmass. Plagioclase occurs as euhedral laths up to 1 mm long, often in aggregates with clinopyroxene.
Units 1–4: Rotoehu Ash and Rotoiti Ignimbrite
Most of the Rotoiti deposits consist of relatively crystal-rich pyroclastics. Nairn (1981) determined crystal concentrations in a number of pyroclastic fall beds. These concentrations are highly variable as a result of syn- eruptive crystal-glass sorting and range from 15 to 75% crystals by volume, averaging about 55% crystals. Davis (1985) determined crystal concentrations in the pumice clasts from the four main units and they range between 16 and 28 vol. % crystals for the Unit 1 and Unit 4 fall deposits and between 18 and 27 vol. % crystals in the Units 2 and 3 pyroclastic flow deposits. These modal proportions are presented in Table 3 and indicate that there is no systematic difference in crystal concentrations between the four main units.
|
The composition and classification of the mineral phases in the Rotoiti deposits was studied in detail to investigate variations in composition through the stratigraphy. The mineral phases present are felsic silicates (quartz, plagioclase), ferromagnesian silicates (orthopyroxene, cummingtonite, hornblende, biotite), iron–titanium oxides (titanomagnetite, ilmenite) and accessory minerals (zircon, apatite, pyrrhotite). The pumice glass from throughout the stratigraphic sucession was also analysed and these data provide a framework against which the compositions of glass inclusions within the phenocryst phases may be compared.
Felsic silicates
Quartz occurs throughout the Rotoiti Formation as euhedral to subhedral bipyramidal crystals. It is rarely in growth contact with other minerals. These large crystals (1–3 mm) are often embayed and contain abundant inclusions of clear to brown glass as well as small tabular and octahedral Fe–Ti oxides and apatite needles. The glass inclusions, notable for their size (up to 100 µm diameter), are visible in hand specimen.
Plagioclase is the only feldspar in the Rotoiti pyroclastics and it occurs as euhedral 1–2 mm diameter crystals. The large size, abundance, and evident zonation of plagioclase crystals suggest that they were one of the earliest of the liquidus phases in the melt. Some large complexly zoned plagioclases have smaller euhedral plagioclase crystals at their cores that may have acted as seed crystals. Included within plagioclase phenocrysts are Fe–Ti oxides, apatite needles and small zircon grains. Significantly, plagioclase crystals partially enclose ferromagnesian crystals only at their outer margins.
Complex optical and compositional zoning of the feldspar is a characteristic of these rocks. The zoning exhibits an oscillatory pattern superimposed over a broad normal zoning trend from calcic core to more sodic rim. The cores of plagioclase crystals are often patchy or mottled in appearance or contain thick, diffuse, optically distinct bands. Plagioclase cores are typically bounded by a strongly corroded margin and immediately outside this is a thin, reversely zoned, band followed without break by a return to normal zonation. A second thin, reversely zoned, band is commonly seen following another corroded margin. The outermost portion of all crystals consists of thin oscillatory zones superimposed on a normal zonation trend.
Figure 3 illustrates plagioclase phenocryst compositions. Plagioclase is dominantly andesine within a total compositional range of An24–55. Crystal cores are for the most part more calcic than coexisting rims. Samples from Unit 3 and the lower part of Unit 4 contain plagioclase phenocrysts with significantly more calcic cores (An45–55) than those of Units 1 and 2. The bed at the top of Unit 4, represented by sample AU46253, is distinctive in that plagioclase phenocrysts are reversely zoned with sodic cores (An24–28) and slightly more calcic rims (An25–35). Potassium substitution in the plagioclase phenocrysts mirrors the potassium content of the magma from which they crystallized. The orthoclase content of phenocrysts is generally lower in cores than in the rims, reflecting their normal zonation in anorthite content. Unit 4 fall beds contain rare, reversely zoned, high-Or (Or3·5–4·5) crystals. The uppermost bed of Unit 4 contains only this high-Or plagioclase in contrast to the bulk of the Rotoiti pyroclastics.
|
) and rims (
) plotted in relative stratigraphic order. Sample numbers are given in the left-hand column.Ferromagnesian silicates
The nature, composition and relative proportions of the ferromagnesian silicate assemblages vary with stratigraphic height through the Rotoiti sequence (Table 4; Figs 4 and 5).
|
|
|
Cummingtonite dominates the assemblage in Unit 1 and Unit 2 beds. From the lowest Unit 3 beds small quantities of biotite enter the assemblage and the amount of hornblende increases. Marked changes in mineral proportions occur in Unit 4, where biotite, hornblende and orthopyroxene increase at the expense of cummingtonite. This trend culminates in the final ash bed of Unit 4 (AU46253), where cummingtonite is absent from the assemblage and orthopyroxene, hornblende and biotite are abundant.
Orthopyroxene is the only anhydrous ferromagnesian silicate phase in the Rotoiti pyroclastics (a single ragged crystal of clinopyroxene observed is probably xenocrystic and is compositionally distinct from augite in the Matahi Scoria). The orthopyroxene typically forms small (<1 mm) blunt prismatic crystals that are slightly pleochroic (pale pinkish brown to pale green). Crystals contain abundant inclusions of magnetite, ilmenite and glass, together with rarer zircon and apatite. The composition of analysed crystals ranges from Mg40 to Mg65 [Mgx refers to the value of cation % Mg/(Mg + Fe) in the crystal]. Relatively magnesian orthopyroxene (Mg57–65) occurs throughout the sequence (although analyses are not available for Unit 2) and a relatively iron-rich orthopyroxene occurs in Unit 2 and more abundantly in the upper part of the sequence.
Cummingtonite typically occurs as large (1–3 mm) prismatic slightly pleochroic (light green–light brown–blue green) optically positive crystals with an extinction angle of 16–20°. Exsolution lamellae of brown calcic amphibole on {110} exist in many crystals and mixed cummingtonite–hornblende compositions were detected by electron microprobe. Cummingtonite crystals contain ubiquitous inclusions of oxides and glass. The composition of the cummingtonite is nearly constant throughout the Rotoiti sequence at Mg61–63 (Figs 5 and 6) and individual crystals lack compositional zonation. Cummingtonite is absent from the uppermost bed of Unit 4.
|
Hornblende is distinguished from cummingtonite by darker colour and stronger pleochroism (dark green to brown). Crystals range up to 3 mm in the upper units and are smaller in the lower parts of the sequence. Small oxides and orthopyroxene crystals are common as inclusions. Individual hornblende crystals show no compositional zoning but there is wide compositional variation between crystals from Mg37Fe37Ca26 to Mg56Fe32Ca12. Two compositional groups are recognized (Figs 5 and 6), a magnesian hornblende (>Mg58) observed in contact with cummingtonite, which is present through Units 1–3, and an iron-rich hornblende (<Mg56), which first appears in Unit 2, is found in the upper beds of the Rotoiti sequence and is the only calcic amphibole in Unit 4 fall beds. In the Unit 4 beds, iron-rich hornblende is the only phase that contains inclusions of small euhedral biotite crystals.
Biotite occurs as dark reddish brown euhedral to subhedral tabular crystals (1–3 mm diameter). It enters the phenocryst assemblage in the Unit 3 beds. It is sparsely present in both pyroclastic flows and pyroclastic fall beds of Units 3 and 4 but increases sharply in abundance in the youngest beds of Unit 4. Biotite crystals include Fe–Ti oxides and small quartz crystals. Tiny euhedral biotite grains occur in iron-rich hornblende in the upper Unit 4 beds. Biotite has a composition of Mg48–53 similar to that in both the iron-rich orthopyroxene and the iron-rich hornblende.
In Fig. 6 the compositions of ferromagnesian minerals are plotted in the system Fe–Mg–Ca (cation %). The magnesium ratios of cummingtonite and the more magnesian hornblende and orthopyroxene phenocrysts are similar at Mg60 and overlap with the fields of cummingtonite-bearing assemblages from other lavas of the Okataina Volcanic Centre (Ewart et al., 1975). In contrast, the similar Mg50 ratios of the iron-rich hornblende and orthopyroxene overlap with the fields of that same assemblage in other rhyolites from the Okataina Volcanic Centre.
Oxides and accessory minerals
Euhedral octahedra of titanomagnetite and tabular ilmenite crystals are the main oxide phases. Although common as free phases, they also occur as inclusions within ferromagnesian silicate phases. Ilmenite is approximately an order of magnitude less abundant and crystals are normally smaller than those of coexisting titanomagnetite.
Zircon and apatite are common accessory phases and occur mainly as inclusions in all phenocryst phases. Apatite is roughly twice as abundant as zircon. Both are particularly abundant as inclusions in titanomagnetite and ilmenite, and can make up as much as 1 vol. % of individual host crystals. Small anhedral crystals of pyrrhotite are also observed as inclusions in Fe–Ti oxide crystals.
Coexisting titanomagnetite and ilmenite from the Rotoiti sequence are plotted in the ternary system Ti–Mg–Mn (cation %) in Fig. 7. Titanomagnetites from Units 1–4 plot as a field away from the Ti apex whereas ilmenites from the same units form a linear trend parallel to the Mg–Ti join. A group of the titanomagnetites in the final Unit 4 bed (AU46253) form a separate field toward higher Ti contents; however, ilmenites from this bed plot at various points on the Rotoiti array. All coexisting phases were assessed for equilibrium using the empirical Mg/Mn partitioning relationships of Bacon & Hirschmann (1988). Crystals that failed this equilibrium test are from the Unit 4 beds.
|
Glass inclusions
All phenocryst phases in the Rotoiti pyroclastics contain small inclusions (<1 to >100 µm) of glass commonly referred to as melt inclusions. Inclusions within the ferromagnesian phases tend to be elongate parallel to growth planes and commonly intersect the outside of fragmental crystals. However, inclusions within plagioclase and quartz phenocrysts are larger and vary in terms of their size, shape, colour, degree of crystallinity, ratio of vapour bubble to inclusion volume, and connection to the crystal surface either through cracks or thin ‘necks’.
In plagioclase phenocrysts there are two populations of inclusion based on size and mode of occurrence. The first group consists of very small inclusions (<10 µm) that are concentrated in crystal growth zones. Most of these contain shrinkage vapour bubbles that consistently occupy between 10 and 20% of the volume of the inclusion. They are generally somewhat prismatic, rather than rounded, which suggests that there has been some post-entrapment crystallization at the inclusion walls. This group of inclusions is probably the result of rapid skeletal growth of the plagioclase spurred by a change in the magma chamber environment. The second group of plagioclase-hosted inclusions are consistently larger (10–50 µm) and are most commonly observed within phenocryst cores, although they also occur in other zones. They are not found in groups and are more randomly distributed than the first type. Inclusions of this second group often have no shrinkage bubble and are round, irregular or prismatic in shape.
Glass inclusions in quartz phenocrysts differ from those in plagioclase primarily in their larger size. They commonly exceed 100 µm and are nearly always single phase, being composed of a homogeneous clear to pale brown glass. However, small, multi-phase inclusions are also observed in quartz phenocrysts, but the relative volume of vapour bubbles rarely exceeds 10% of the inclusion volume. Inclusions within quartz often have a rounded to prismatic habit and, although obviously forming at growth surfaces, they are randomly distributed within crystals.
Microprobe analyses of glass inclusions from Rotoiti phenocrysts indicate that both exsolution of vapour bubbles and crystallization of microlites have affected their chemical composition. Table 5 compares the major element compositions of single- and multi-phase inclusions within quartz. Compared with the analysis of an isolated inclusion with no vapour bubble or crystallites, the ‘necked’ or cracked inclusion with a large vapour bubble has a much higher total, near 100%, which suggests that it has lost all of its volatile component. The ‘devitrified’ isolated inclusion has a total less than 100% and is interpreted to have preserved its volatile component, although its lower Fe, Ca and Mg content suggests that these elements have been preferentially incorporated into crystallites. Some elements may also have been preferentially incorporated into host material crystallized onto the inner walls of the inclusion after entrapment but prior to quenching (e.g. Ca and Na into plagioclase). The prismatic shape of many Rotoiti melt inclusions is evidence for post-entrapment crystallization of host material.
|
Selected compositional parameters for glass inclusions in quartz and feldspar that are considered not to have suffered significant post-entrapment modification on the basis of mode of occurrence and consistent K/Na ratios are plotted in Fig. 8. Two compositionally distinct groups of inclusion are apparent. The compositions of pumice glass also define two fields. A relatively low-K field represents compositions that are found throughout the Rotoiti sequence and a relatively high-K field represents samples from the upper part of Unit 4. Most inclusion compositions overlap those of pumice glass representing the liquid phase in the magma prior to eruption. Inclusion glasses are slightly lower in Na and are more Q-normative, and are interpreted as less evolved (earlier) components of the Rotoiti magma. The second group of inclusions shows compositions that are clearly more potassic and these are from crystals in the upper beds of Unit 4. They are comparable with the second group of pumice glass compositions found in Unit 4.
|
|
INTENSIVE PARAMETERS OF THE MAGMA |
|---|
The equilibrium thermodynamics of TVZ rhyolites has been discussed by Ewart et al. (1971
, 1975). The Rotoiti eruption in particular has attracted interest because of its ferromagnesian mineral assemblage, which includes cummingtonite and constrains determinations of water fugacity and total pressure (Wood & Carmichael, 1973; Oba & Nicholls, 1986; Nicholls et al., 1992). The results of these previous studies indicate that the Rotoiti magmas equilibrated under conditions of low temperature (700–800°C), oxygen fugacity (14–15 log units) and pressure (
3 kbar), with the partial pressure of water approaching the total pressure (PH2O Ptotal).
Indications of chemical and mineralogical variation, particularly at the end of the Rotoiti eruption sequence, have prompted a reassessment of the Fe–Ti oxide geothermometry. An assumption of equilibrium between coexisting titanomagnetite–ilmenite pairs was checked by Mg/Mn partitioning (Bacon & Hirschmann, 1988). Magnetite–ilmenite pairs from pumice clasts have been used to estimate temperature and oxygen fugacity for the Rotoiti magma using the algorithm of Ghiorso & Sack (1991). The estimates of T and fO2 by this method are limited mainly by the uncertainties in the electron microprobe measurements to ±30°C and ±0·5 log units fO2.
The results of the geothermometric calculations are illustrated in Fig. 9. The T–fO2 data from coexisting oxides from Unit 1 to Unit 4 define a linear array from 780°C (13·3 log units fO2) to 840°C (12·0 log units fO2). This trend closely follows, but lies below, the empirically derived hbl–opx buffer of Ewart et al. (1975) but lies above the trend defined by post-Rotoiti Okataina eruptives. T–fO2 data for coexisting oxides from the uppermost Unit 4 bed plot separately from the main Rotoiti array at lower fO2 for a given temperature. Many titanomagnetites from this bed have higher TiO2 contents; however, coexisting ilmenites are low-Ti phases similar to those in the rest of the Rotoiti deposits. These observations are interpreted to indicate disequilibrium conditions in the last of the Rotoiti material erupted.
|
An upper limit to the pressure and inferred depth of the Rotoiti magma chamber has been estimated at <3 kbar and 10 km by the presence of cummingtonite in the mineral assemblage (Nicholls et al., 1992
). An independent estimate of the pressure of crystallization is given by the concentration of water dissolved in the glass of the deposits. The solubility of water in silicate magmas has been studied in the experiments of Burnham & Jahns (1962), Hamilton et al. (1964) and Burnham (1979). These studies have shown that water solubility is dependent on melt composition, pressure and, to a lesser extent, temperature. An estimate of the original water content of the Rotoiti glasses has been determined indirectly through the ‘water by difference’ method using electron microprobe analyses of inclusion glasses. The difference of the analysis total from 100% is assumed to be the concentration of volatiles (predominantly water) in the glass. This assumption was validated by comparison with known anhydrous glasses whose compositions consistently totalled 100 ± 1%.
The average water by difference concentration of the Rotoiti inclusions was 5·62 ± 0·9 wt %. There was no significant difference in the average concentration for inclusions from different mineral phases or throughout the deposit. This value is taken as an approximation of the concentration of water in the Rotoiti magma prior to eruption. The solution model of Burnham (1979) was used to calculate the total pressure of the Rotoiti magma at vapour saturation; as the bulk composition of the Rotoiti magma is similar to the felsic composition studied by Burnham & Jahns (1962) and Burnham (1979), the pressure was directly determined from the experimental solubility curve for the Harding pegmatite. The pressure of entrapment is estimated by this procedure as 2 kbar. The Rotoiti magma chamber is thus constrained to be >6 km below the surface.
|
WHOLE-CLAST CHEMICAL COMPOSITIONS |
|---|
Representative chemical analyses of whole pumice clasts from the Matahi Tephra, the Rotoiti pyroclastics and the Earthquake Flat ignimbrite are presented in Table 1.
The Matahi Tephra is a high-Al basaltic andesite chemically similar to other small-volume basalts erupted in the TVZ over the past 2 Myr. Although evidence for magmatic mingling or mixing is lacking, the basaltic clasts contain abundant pumiceous particles and plagioclase xenocrysts, suggesting selective, if not bulk, contamination to produce the observed compositions, despite efforts to mechanically separate the felsic contaminant. Our interpretation of the data is that the analyses of Matahi clasts in Table 1 represent basaltic magma with relatively increased Si, Al and large ion lithophile elements (LILE), and decreased Ca, Na, Mg, high field strength elements (HFSE), V and Cr as a result of felsic pumice and feldspar contamination. Similar trends are seen in other TVZ basalt samples (Cole, 1979).
Pumice clasts collected to represent the stratigraphic sequence through the Rotoiti deposits are rhyolite to high-Si rhyolite with a small range in silica content (72–76 wt % on a loss-free basis). The data plot as well-correlated linear arrays (Fig. 10). Although subtle, there is some correlation of composition with stratigraphy, with samples from Units 2 and 3 showing a greater range in composition than samples from Units 1 and 4. The compositional trends of depletion (Ti, Al, Fe, Ca, Mg) and enrichment (Na, K) with increasing SiO2 are characteristic of rhyolitic suites from the TVZ, as are the very low abundances of Mg and P. Most compatible trace element concentrations are below XRF detection limits, exceptions being Zn and V, both of which decrease with increasing SiO2. The usually incompatible trace elements Sr and Zr do not correlate positively with SiO2 (Fig. 10). Zr shows a moderate spread of abundances (150–190 ppm), whereas other incompatible trace elements have similar ranges (Rb 56–70 ppm; Sr 140–170 ppm).
|
Major element compositions of groundmass glass from clasts and lapilli are also plotted in Fig. 10 and these lie on the trends defined by whole clast compositions but at higher SiO2 contents (77·5–78·5 wt %). Some groundmass glass compositions from Unit 3 and particularly Unit 4 are markedly higher in K2O and lower in Na2O than the general trend of the Rotoiti suite.
REE abundances of representative samples are presented in Table 2 and plotted in Fig. 11. The chondrite-normalized REE patterns are typical of TVZ rhyolites and show light REE (LREE) enrichment relative to heavy REE (HREE) and the development of a small negative Eu anomaly in most samples. There is a consistent trend of decreasing overall REE abundance from Unit 1 to Unit 4 (also correlated with increasing SiO2) but no change in relative light/heavy fractionation.
|
The patterns of element depletion and enrichment shown by the Rotoiti eruptives suggest the operation of crystal–liquid fractionation processes. Fractional crystallization can explain the negative correlations of Sr and Zr with SiO2. The compatibility of Zr and Y in zircon, P in apatite, Ti, Nb and V in hornblende and Fe–Ti oxides, and Sr in plagioclase can explain the relatively low REE concentrations and negative Eu anomalies correlated with higher SiO2. In contrast, Rb and K appear incompatible with the fractionating assemblage.
Although the relatively low saturation limits of Zr and P in silicate melts result in the formation of zircon and apatite early in the crystallization sequence, their small size inhibits their ability to separate from the liquid. However, both of these phases are characteristically included within major phenocryst phases, particularly Fe–Ti oxides. In Rotoiti samples titanomagnetite crystals can contain >1% by volume of apatite together with less abundant zircon. Fractionation of oxides from the magma can thus affect fractionation of Zr, Y and HREE. The observed covariance between Zr and Ti in the Rotoiti samples supports this fractionation model.
Quantitative modelling of the proposed fractionation process is illustrated in Fig. 12. The starting composition is the most mafic pumice composition from the Rotoiti sequence (AU46227), a clast from the uppermost Unit 2 pyroclastic flows. The fractionating assemblage is that observed as phenocrysts, and partition coefficients were taken from studies involving rocks of similar bulk composition, temperature and mineral assemblage (Table 6). The calculations indicate that the compositional range observed in the Rotoiti suite can be explained by 25–30% fractional crystallization from a starting material with the composition of AU46227. The dominant controls on this model are the crystallization of plagioclase and Fe–Ti oxides with their included apatite and zircon.
|
|
Nd–Sr–Pb isotopic compositions of a small number of representative samples from the Rotoiti sequence are presented in Table 7 and plotted in Fig. 13. These compositions are comparable with those of other groups of TVZ rhyolites. Isotopic data for three samples from the Earthquake Flat deposits are also presented. The rationale for analysing Earthquake Flat samples is to test the possibility that there is a genetic relationship between the Rotoiti and Earthquake Flat magmas because they were erupted at essentially the same time, and because the mineral assemblage in the Earthquake Flat samples resembles that of assemblage B (see below) in the upper levels of the Rotoiti deposits. Although subtle, there is a distinct isotopic difference between the two magmas, which suggests that there is no direct petrogenetic relationship.
|
|
). In the Nd–Sr diagram the fields for Okataina, Maroa and Taupo volcanic centres are drawn from data presented by Graham et al. (1992) |
ZONING IN THE ROTOITI ERUPTIVES |
|---|
The combined mineralogical and compositional variations within the Rotoiti sequence can be summarized as follows.
- Whereas the total phenocryst abundance within the different units of the sequence remains relatively consistent at around 25%, the modal proportions of ferromagnesian phenocrysts vary with stratigraphic position. Cummingtonite (Mg62) dominates (to <99%) the mafic assemblage in early Unit 1 beds, whereas orthopyroxene and hornblende become increasingly abundant toward the top of the sequence (<20% and <7%, respectively).
- Two compositions of orthopyroxene and hornblende are present in the sequence including a magnesian (Mg63) composition and a more iron-rich composition (Mg50). The magnesian phenocrysts are more common in the lower Rotoiti pyroclastics and have a Mg/Fe ratio similar to that of coexisting cummingtonite. Iron-rich crystals increase in abundance upwards through the younger units of the Rotoiti pyroclastics and these have Mg/Fe ratios similar to those of coexisting biotite phenocrysts.
- Units 3 and 4 are defined by the occurrence of biotite in their mafic modal assemblage. Biotite (Mg50) is present as free crystals, enclosed within pumice clasts and as inclusions within Fe-rich hornblende; its abundance also increases upward in the sequence.
- Plagioclase crystals in most of the Rotoiti pyroclastics are typically normally zoned, with the exception of crystals in the uppermost bed of Unit 4, which are more potassic and are reversely zoned. Mixing of these two populations occurs in the other Unit 4 beds.
- Pumice glass, melt inclusions, and whole-clast compositions show parallel trends, with a general decrease in incompatible element concentrations (including K2O) from Unit 1 to Unit 3 followed by an increase through Unit 4. A bimodal population of low- and high-K2O glass is also observed in the upper beds of Unit 4.
- The mineralogy of the uppermost bed of Unit 4 contrasts with the rest of the Rotoiti sequence in containing: reversely zoned potassic plagioclase, potassic pumice glass and melt inclusion compositions, and abundant iron-rich orthopyroxene, hornblende and biotite to the exclusion of cummingtonite and magnesian orthopyroxene and hornblende.
The Mg/Fe ratios of the coexisting ferromagnesian phases in the Rotoiti deposits clearly divide them into two groups, one comprising cummingtonite + orthopyroxene + hornblende of roughly Mg63 and the other comprising biotite + orthopyroxene + hornblende of roughly Mg50. The more magnesian minerals are interpreted to represent an equilibrium assemblage, henceforth referred to as assemblage A. It is similar to other cummingtonite-bearing assemblages in rhyolites subsequently erupted from the Okataina Volcanic Centre (Ewart et al., 1975; Nairn, 2002) The more iron-rich minerals probably also represent an equilibrium assemblage (assemblage B) based on their similar Mg/Fe ratios and observed growth relationships (e.g. biotite is observed as inclusions within the iron-rich hornblende). This second mineral assemblage overlaps the compositions of similar biotite-bearing assemblages in TVZ rhyolites.
The presence of two distinct ferromagnesian mineral assemblages in the Rotoiti magma is an important petrogenetic constraint. Because of the overwhelming abundance of cummingtonite in the Rotoiti deposits, assemblage A is considered to be the crystallization product of the Rotoiti magma chamber. The accessory assemblage B may be explained as a xenocrystic assemblage entrained in the magma immediately prior to eruption, as an assemblage that developed within the Rotoiti magma but at different P–T–fO2–fH2O conditions, or as an assemblage that developed in a separate magma that later mixed with the Rotoiti magma.
Assemblage B has been been considered xenocrystic in origin and derived from plutonic bodies or older ignimbrite deposits by Davis (1985) and, more recently, by Burt et al. (1998). Xenocrysts derived from a plutonic rock or welded ignimbrite would be expected to be fragmentary and to be associated with an enclosing matrix; selective incorporation of an equilibrium assemblage of minerals without any sign of other minerals or matrix fragments is unlikely. The euhedral nature of the iron-rich phenocrysts in the Rotoiti deposits indicates that they crystallized with entirely free surfaces, and this strongly supports a direct magmatic origin. The uppermost bed of Unit 4 (represented by sample AU46253) provides the best approximation to the end-member composition of the magma from which assemblage B could have crystallized.
Iron-rich minerals occur not only as free crystals but are physically included in large pumice clasts (40 cm) from the Rotoiti deposits, which suggests that they were included at a magmatic stage. Further, there are crystals that span part of the compositional gap between the two assemblages that may represent crystals formed or modified as the two magmas mixed. Crystallization in a magma-mixing environment may explain the reverse zonation patterns in the plagioclase of the Rotoiti Unit 4 bed. There is also subtle change in the pumice glass compositions of the Unit 4 beds toward higher K2O contents that may reflect mixing of Rotoiti and another magma. Finally, there is the disequilibrium between coexisting oxides observed in the Unit 4 beds.
Following from our argument that the evidence from the Rotoiti sequence indicates the presence of two distinct magmas that interacted prior to eruption, the question that arises is whether these two magmas formed in, and occupied, the same chamber but were segregated into different domains by boundaries created through P–T–fO2–X differences, or whether they existed as discrete magma bodies.
All samples of the Rotoiti sequence that are dominated by assemblage A, including Unit 3 and 4 beds, which contain accessory amounts of assemblage B, were at similar conditions of PH2O Ptotal from 2 to 3 kbar as constrained by the crystallization of cummingtonite. However, the deposits that are dominated by assemblage B, including the uppermost bed of the Rotoiti sequence, do not contain cummingtonite and thus may have formed at higher pressures. Nicholls et al. (1992) demonstrated that the crystallization of cummingtonite is limited at pressures >3 kbar by the crystallization of biotite and orthopyroxene in peraluminous systems. Thus the high-Fe assemblage B could represent the higher-pressure equilibrium assemblage of a magmatic system that at low pressures is crystallizing assemblage A.
The scenario of one magmatic system partitioned into distinct domains by P–T–fO2–fH2O constraints is not supported by temperature arguments. On the basis of coexisting oxide equilibration temperatures of 780–840°C and fO2 –13·3 to –12·0 log units for assemblage A, compared with cooler and more reducing conditions (770°C, –14·5 log units fO2) for assemblage B, it is unlikely that the latter can represent deeper levels of the same magma chamber, where hotter and more oxidizing conditions would be expected. Further, the contrasting Mg/Fe ratios of the two assemblages cannot be due to realistic temperature, pressure or compositional gradients within a single magma chamber. Experimental studies (Conrad et al., 1988; Nicholls et al., 1992) have demonstrated that more magnesian mineral assemblages are favoured by higher pressures and temperatures, and thus should form deeper in the magma chamber, whereas the less magnesian assemblage should form in the cooler upper portion. According to accepted models of magma withdrawal from a storage chamber (Blake, 1981; Blake & Ivey, 1986; Spera et al., 1986; Blake et al., 1992), material from the upper levels of a magma chamber is the first erupted and deposited. However, the first deposits of the Rotoiti sequence contain the magnesian mineral assemblage A, indicating that this material resided in the upper part of the system.
The bulk compositions of the two magmas also do not support their existence as different levels of a common magma chamber. On the basis of bulk major and trace element concentrations and mineral and glass compositions, the magma containing assemblage B is the more evolved, and, in conventional models, should have been at the top of the system and hence be represented in the first material to be erupted. The fact that this trend is actually observed within the units containing assemblage A provides further support for the two-magma model.
|
THE ROTOITI ERUPTION |
|---|
A model for the Rotoiti eruption is constrained by the following observations.
- The eruption was immediately preceded by the comparatively small-volume basaltic event that produced the Matahi Scoria. There is no evidence for mixing of basalt and rhyolite magmas.
- The Rotoiti deposits are dominated by crystal-rich (25%) high-Si rhyolites containing a relatively magnesian mafic mineral assemblage that varies systematically from cummingtonite dominated upward to orthopyroxene + hornblende dominated. A subordinate iron-rich assemblage of biotite + hornblende + orthopyroxene occurs late in the sequence and increases in abundance upward to the Unit 4 beds, where bimodal low- and high-K plagioclase and glass compositions are also found.
- The Rotoiti event was immediately followed by eruption of pyroclastic flows and fall deposits of the Earthquake Flat Tephra Formation. The Earthquake Flat magma was an even more crystal-rich high-Si rhyolite (35%) with a bulk composition more evolved than that of the Rotoiti deposits. However, modelling indicates that an Earthquake Flat-type magma could have been derived by fractionation from a parent similar to that which gave rise to the Rotoiti magma. The Earthquake Flat mineral assemblage and glass composition are similar to assemblage B and the high-K components in the upper Rotoiti units.
- Xenocrystal contamination and airborne mixing cannot explain all features of the mixing of contrasting mineral assemblages and glass compositions in the Rotoiti deposit; neither can compositional gradients within the same magma chamber generate the relationships observed between the two magmas.
The observed variations in the upper units of the Rotoiti eruption sequence can be explained by mingling of two rhyolitic magmas immediately prior to eruption. With the plumbing established between the two magma bodies, the relative volume of the secondary magma and point of injection into the primary Rotoiti chamber is constrained by the abundance of the two mineral assemblages and the compositional trends through the Rotoiti stratigraphic sequence. The volume of the secondary magma is requisitely much smaller than that of the primary Rotoiti magma, as assemblage B (Mg50) is roughly an order of magnitude less abundant than assemblage A (Mg62), even in the upper Rotoiti beds. Assemblage B magma was probably injected near the bottom of the main Rotoiti magma chamber as the magma rose through the crust. The sequence of products erupted during the Rotoiti event conforms to proposed models of magma withdrawal and mixing with a basal injection of assemblage B magma (Blake, 1981; Blake & Ivey, 1986; Spera et al., 1986; Blake et al., 1992). Evolved, relatively homogeneous, cummingtonite-bearing assemblage A magma from the top of the chamber was the initial material withdrawn during the first stages of the eruption and continued to be erupted during most of the event. Very small amounts of the second magma may have been drawn up quickly in the early stages of the eruption, resulting in the rare early occurrence of assemblage B in stratigraphically lower units. However, discrete compositional domains of the second magma would probably have dispersed during early ascent through the chamber to the surface.
Middle stages of the eruption tapped, in addition to upper zone material, less evolved assemblage A magmas from middle to lower zones of the magma. The first significant quantities of introduced magma also began to be tapped and erupted, as evidenced by the occurrence of assemblage B minerals (e.g. biotite) in the materials deposited. As the Rotoiti chamber progressively emptied, the proportion of erupted mingled magmas increased, resulting in the observed increase in the proportion of assemblage B through Units 3 and 4.
Near emptying of the chamber in the final stages of eruption resulted in the withdrawal of a small volume of material dominated by the assemblage B as represented by the final bed of Unit 4. This assemblage B magma may have occupied the bottom of the primary Rotoiti chamber for a period of days or weeks, and thus would have had time to mix with the Rotoiti magma and partially equilibrate to the conditions of that chamber. Mixing of the two magmas may have effectively reversed the compositional trend (of more mafic compositions with time) observed in the pyroclastic flow pumices, and more evolved magmas were erupted as indicated by higher potassium glass compositions in the final Unit 4 fall deposit. A short residence time in the bottom of the Rotoiti chamber may also explain the disequilibrium and mixing populations in the oxides of the Rotoiti Unit 4 deposits.
The evidence for the involvement of two magmas in the Rotoiti eruption, the conclusion that they existed as separate physical entities until the eruption and the suggestion that they were ultimately linked to a common parent is based on petrological observations and their logical interpretation. A discussion of the physical nature of the plumbing system connecting the two magma bodies is speculative because little is known of the sub-surface structure of Okataina Volcanic Centre. The petrological evidence for magma mixing is compelling and points to a physical link between magma bodies within the upper crust. The Okataina Volcanic Centre characteristically produces multi-vent eruptions from SE-oriented lineaments and a consistent scenario is that the Rotoiti eruption ‘unzipped’ its magmatic system and ultimately encountered and entrained a second magma body.
Recent studies (Nairn et al., 2004; Smith et al., 2004) have demonstrated the presence of at least two distinct magmas within a single eruption sequence in several of the younger (<50 ka) eruptions from Okataina Volcano. This suggests a model in which discrete magmas are assembled in the upper crust immediately prior to eruption and do not have time to equilibrate chemically or thermally. Our data from the Rotoiti eruption support this model, although it is also possible that the magmas were assembled during the course of the eruption. Eruptions from Okataina appear not to have been fed from relatively long-lived magma chambers, rather they have tapped a heterogeneous upper-crustal plexus of magma bodies. Evidence from zircon crystal ages for long (50 kyr) magma residence times (Charlier et al., 2003) is probably a reflection of multiple components (xenoliths, entrained crystals, discrete liquids) within magmatic systems rather than evidence for the prolonged existence of large magma bodies in the upper crust.
The TVZ is a complex tectono-magmatic system that at present is known only in broad outline. A major and unresolved question is the source of the felsic magmas and the nature of the processes that generate them. Upper-crustal processes in TVZ felsic systems include mixing and mingling with mafic magmas; in some cases these interactions have led immediately to eruptions (Wilson et al., 1984; Blake et al., 1992). The Rotoiti eruption was preceded by a basaltic event, but in this case if one triggered the other the relationship was indirect, perhaps through tectonic instability. The petrology of the Rotoiti deposits shows that the melt zone beneath the Okataina Volcano at the time of the eruption was complex and that discrete felsic magma bodies existed and had evolved separately prior to the eruption. An implication is that eruptions at least in the northern part of the TVZ are fed from a diffuse upper-crustal zone of partially interconnected and at times physically separate magma bodies rather than from a centralized and necessarily large long-lived magma chamber. Logically this zone lies below the 6 km depth brittle–ductile transition identified by Bryan et al. (1999) on the basis of seismicity. The ultimate source of the felsic magmas is probably within the zone of relatively high seismic conductivity at depths of 10–15 km identified by Henrys et al. (2003).
Finally, the behaviour of the Okataina magmatic system during the last 50 kyr (this study, Nairn et al., 2004; Smith et al., 2004) appears very different from that of large silicic magmatic systems (e.g. Yellowstone, Long Valley) in western North America. We speculate that in the relatively old, thick and coherent crust of the North American continent silicic magmas can accumulate in long-lived magma chambers over periods of 103–104 years. In contrast, the crust of central North Island, New Zealand, is thin, hot and undergoing extension, and in this environment large magma chambers cannot develop and magmas are erupted relatively rapidly after their formation.
|
SUPPLEMENTARY DATA |
|---|
Supplementary data for this paper are available on Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
We would like to acknowledge the help and support of Ian Nairn, thoughtful reviews by Bruce Charlier, Olivier Bachmann and Barbara Nash, and the patience of George Bergantz.
|
FOOTNOTES |
|---|
* Corresponding author. Telephone: (649)373 7599 x 7416. Fax: (649)3737435. E-mail: ie.smith{at}auckland.ac.nz
|
REFERENCES |
|---|
Bacon, C. R. & Hirschmann, M. M. (1988). Mg/Mn partitioning as a test for equilibrium between coexisting Fe–Ti oxides. American Mineralogist 73, 57–61.[Abstract]
Bailey, R. A. & Carr, R. G. (1994). Physical geology and eruptive history of the Matahina Ignimbrite, Taupo Volcanic Zone, North Island, New Zealand. New Zealand Journal of Geology and Geophysics 37, 319–344.
Blake, S. (1981). Eruptions from zoned magma chambers. Journal of the Geological Society, London 138, 281–287.
Blake, S. & Ivey, G. N. (1986). Magma mixing and the dynamics of withdrawal from stratified reservoirs. Journal of Volcanology and Geothermal Research 27, 153–178.[CrossRef][Web of Science]
Blake, S., Wilson, C. J. N., Smith, I. E. M. & Walker, G. P. L. (1992). Petrology and dynamics of the Waimahia mixed magma eruption, Taupo Volcanic Zone, New Zealand. Journal of the Geological Society, London 149, 193–207.
Briggs, R. M., Gifford, M. G., Moyle, A. R., Taylor, S. R., Norman, M. D., Houghton, B. F. & Wilson, C. J. N. (1993). Geochemical zoning and eruptive mixing in ignimbrites from Mangakino volcano, Taupo Volcanic Zone, New Zealand. Journal of Volcanology and Geothermal Research 56, 175–203.[CrossRef][Web of Science]
Brown, S. J. A., Burt, R. M., Cole, J. W., Krippner, S. J. P., Price, R. C. & Cartwright, I. (1998). Plutonic lithics in ignimbrites of Taupo Volcanic Zone, New Zealand; sources and conditions of crystallisation. Chemical Geology 148, 21–41.[CrossRef][Web of Science]
Bryan, C. J., Sherburn, S., Bibby, H. M., Bannister, S. C. & Hurst, A. W. (1999). Shallow seismicity of the central Taupo Volcanic Zone, New Zealand: its distribution and nature. New Zealand Journal of Geology and Geophysics 42, 533–542.
Burnham, C. W. (1979). The importance of volatile constituents. In: Yoder, H. S. (ed.) The Evolution of the Igneous Rocks: Fiftieth Anniversary Perspectives. Princeton, NJ: Princeton University Press, pp. 439–482.
Burnham, C. W. & Jahns, R. H. (1962). A method for determining the solubility of water in silicate melts. American Journal of Science 260, 721–745.
Burt, R. M., Brown, S. J. A., Cole, J. W., Shelly, D. & Waight, T. E. (1998). Glass bearing plutonic fragments from ignimbrites of the Okataina Caldera complex, Taupo Volcanic Zone, New Zealand: remnants of a partially molten intrusion associated with preceding eruptions. Journal of Volcanology and Geothermal Research 84, 209–237.[CrossRef][Web of Science]
Carmichael, I. S. E. (1967). Iron–titanium oxides of salic rocks. Contributions to Mineralogy and Petrology 14, 36–64.[CrossRef]
Charlier, B. L. A., Peate, D. W., Wilson, C. J. N., Lowenstern, J. B., Story, M. & Brown, S. J. A. (2003). Crystallisation ages in coeval silicic magma bodies: 238U–230Th disequilibrium evidence from the Rotoiti and Earthquake eruption deposits, Taupo Volcanic Zone, New Zealand. Earth and Planetary Science Letters 206, 441–457.[CrossRef][Web of Science]
Cole, J. W. (1979). Structure, petrology, and genesis of Cenozoic volcanism, Taupo Volcanic Zone, New Zealand—a review. New Zealand Journal of Geology and Geophysics 22, 631–657.
Conrad, W. K., Nicholls, I. A. & Wall, V. J. (1988). Water-saturated and undersaturated melting of metaluminous and peraluminous crustal compositions at 10 kb: evidence for the origin of silicic magmas in the Taupo Volcanic Zone, New Zealand and other occurrences. Journal of Petrology 29, 765–803.
Davis, W. J. (1985). Geochemistry and petrology of the Rotoiti and Earthquake Flat pyroclastic deposits. M.Sc. thesis, University of Auckland.
Dunbar, N. W., Kyle, P. R. & Wilson, C. J. N. (1989). Evidence for limited zonation in silicic magma systems, Taupo Volcanic Zone, New Zealand. Geology 17, 234–236.
Ewart, A. & Healy, J. (1965). Rotorua—volcanic geology. New Zealand Volcanology: Central Volcanic Region. New Zealand Department of Scientific and Industrial Research Information Series 50, 110–126.
Ewart, A., Green, D. C., Carmichael, I. S. E. & Brown, F. H. (1971). Voluminous low temperature rhyolitic magmas in New Zealand. Contributions to Mineralogy and Petrology 33, 128–144.[CrossRef][Web of Science]
Ewart, A., Hildreth, W. & Carmichael, I. S. E. (1975). Quaternary acid magma in New Zealand. Contributions to Mineralogy and Petrology 51, 1–27.[CrossRef][Web of Science]
Flood, T. P., Vogel, T. A. & Schuraytz, B. C. (1989). Chemical evolution of a magmatic system: the Paintbrush Tuff, southwestern Nevada volcanic field. Journal of Geophysical Research 94, 5943–5960.
Froggatt, P. C. (1992). Standardisation of the chemical analysis of tephra deposits: Report of the ICCT Working Group. Quaternary International 13/14, 93–96.[CrossRef]
Froggatt, P. C. & Lowe, D. J. (1990). A review of late Quaternary silicic and some other tephra formations from New Zealand: their stratigraphy, nomenclature, distribution, volume and age. New Zealand Journal of Geology and Geophysics 33, 89–109.
Gerstenberger, H. & Haase, G. (1997). A highly effective emitter substance for mass spectrometric Pb isotope ratio determinations. Chemical Geology 136, 309–312.[CrossRef][Web of Science]
Ghiorso, M. S. & Sack, R. O. (1991). Fe–Ti oxide geothermometry: thermodynamic formulation and estimation of intensive variables in silicic magmas. Contributions to Mineralogy and Petrology 108, 485–510.[CrossRef][Web of Science]
Graham, I. G., Cole, J. W., Briggs, R. M., Gamble, J. A. & Smith, I. E. M. (1995). Petrology and petrogenesis of volcanic rocks from the Taupo Volcanic Zone: a review. Journal of Volcanology and Geothermal Research 68, 59–87.[CrossRef][Web of Science]
Graham, I. J., Gulson, B. L., Hedenquist, J. W. & Mizon, K. (1992). Petrogenesis of late Cenozoic volcanic rocks from the Taupo Volcanic Zone, New Zealand, in the light of new Pb isotope data. Geochimica et Cosmochimica Acta 56, 2797–2819.[CrossRef][Web of Science]
Green, T. H. (1994). Experimental studies of trace-element partitioning applicable to igneous petrogenesis; Sedona 16 years later. Chemical Geology 117, 1–36.[CrossRef][Web of Science]
Hamilton, D. L., Burnham, C. W. & Osborn, E. F. (1964). The solubility of water and effects of oxygen fugacity and water content on crystallisation in mafic magmas. Journal of Petrology 5, 21–39.
Healy, J., Schofield, J. C. & Thompson, B. N. (1964). Rotorua. New Zealand Geological Survey, Department of Scientific and Industrial Research. Geologic Map of New Zealand. Map Sheet 5. Scale 1:250 000.
Henrys, S., Reyners, M. & Bibby, H. (2003). Exploring the plate boundary structure of the North Island, New Zealand. EOS Transactions, American Geophysical Union 84, 289–295.
Hildreth, W. (1981). Gradients in silicic magma chambers: implications for lithospheric magmatism. Journal of Geophysical Research 86, 10153–10192.
Hochstein, M. P., Smith, I. E. M., Regenauer-Leib, K. & Ehara, S. (1993). Geochemistry and heat transfer processes in Quaternary rhyolitic systems of the Taupo Volcanic Zone, New Zealand. Tectonophysics 223, 213–235.[CrossRef][Web of Science]
Houghton, B. F., Wilson, C. J. N., McWilliams, M., Lanphere, M. A., Weaver, S. D., Briggs, R. M. & Pringle, M. S. (1995). Chronology and dynamics of a large silicic magmatic system: Central Taupo Volcanic Zone. Geology 23, 13–16.
Jurado-Chichay, Z. & Walker, G. P. L. (2000). Stratigraphy and dispersal of the Mangaone Subgroup pyroclastic deposits, Okataina Volcanic Centre, New Zealand. Journal of Volcanology and Geothermal Research 104, 319–383.[CrossRef][Web of Science]
Mahood, G. & Hildreth, W. (1983). Large partition coefficients for trace elements in high-silica rhyolites. Geochimica et Cosmochimica Acta 47, 11–30.[CrossRef][Web of Science]
McCulloch, M. T., Kyser, T. K., Woodhead, J. & Kinsley, L. (1994). Pb–Sr–Nd–O isotopic constraints on the origin of rhyolites from the Taupo Volcanic Zone of New Zealand: evidence for assimilation by fractionation from basalt. Contributions to Mineralogy and Petrology 115, 303–312.[CrossRef][Web of Science]
Nagasawa, H. & Schnetzler, C. (1971). Partitioning of rare earth, alkali and alkaline earth elements between phenocrysts and acidic igneous magma. Geochimica et Cosmochimica Acta 35, 963–968.
Nairn, I. A. (1972). Rotoehu Ash and the Rotoiti Breccia Formation, Taupo Volcanic Zone, New Zealand. New Zealand Journal of Geology and Geophysics 15, 251–261.
Nairn, I. A. (1981). Some studies of the geology, volcanic history and geothermal resources of the Okataina volcanic centre, Taupo Volcanic Zone, New Zealand. Ph.D. thesis, Victoria University of Wellington.
Nairn, I. A. (1989). Mount Tarawera. Geological Society of New Zealand. Geological Map of New Zealand 1:50 000. Map Sheet V16 AC.
Nairn, I. A. (1992). The Te Tere and Okareka eruptive episodes—Okataina Volcanic Centre, Taupo Volcanic Centre, New Zealand. New Zealand Journal of Geology and Geophysics 35, 93–108.
Nairn, I. A. (2002). Geology of the Okataina Volcanic Centre, scale 1:50 000. Institute of Geological and Nuclear Sciences Geological Map 25. Lower Hutt, New Zealand: Institute of Geological and Nuclear Sciences Limited, 1 sheet + 156 pp.
Nairn, I. A. & Kohn, B. P. (1973). Relation of the Earthquake Flat Breccia to the Rotoiti Breccia, central North Island, New Zealand. New Zealand Journal of Geology and Geophysics 16, 269–279.
Nairn, I. A., Shane, P., Cole, J. W., Leonard, G. J., Self, S. & Pearson, N. (2004). Rhyolite magma processes of the AD1315 Kaharoa eruptive episode, Tarawera volcano, New Zealand. Journal of Volcanology and Geothermal Research 131, 265–294.
Nash, W. P. & Crecraft, H. R. (1985). Partition coefficients for trace elements in silicic magmas. Geochimica et Cosmochimica Acta 49, 2309–2322.[CrossRef][Web of Science]
Nicholls, I. A., Oba, T. & Conrad, W. K. (1992). The nature of primary rhyolitic magmas involved in crustal evolution: evidence from an experimental study of cummingtonite bearing rhyolites, Taupo Volcanic Zone, New Zealand. Geochimica et Cosmochimica Acta 56, 955–962.[CrossRef][Web of Science]
Norrish, K. & Hutton, J. T. (1969). An accurate X-ray spectrographic method for the analysis of a wide range of geological samples. Geochimica et Cosmochimica Acta 33, 431–454.[CrossRef][Web of Science]
Oba, T. & Nicholls, I. A. (1986). Experimental study of cummingtonite and Ca–Na amphibole relations in the system Cum–Act–Pl–Qtz–H2O. American Mineralogist 71, 1354–1365.[Abstract]
Pin, C. & Bassin, C. (1992). Evaluation of a strontium-specific extraction chromatographic method for isotopic analysis in geological materials. Analytica Chimica Acta 269, 249–255.[CrossRef][Web of Science]
Pullar, W. A. & Nairn, I. A. (1972). Matahi basaltic tephra member, Rotoiti Breccia Formation. New Zealand Journal of Geology and Geophysics 15, 446–450.
Richard, P., Shimizu, N. & Allègre, C. J. (1976). 143Nd/144Nd, a natural tracer: an application to oceanic basalts. Earth and Planetary Science Letters 31, 269–278.[CrossRef][Web of Science]
Shane, P. & Sandiford, A. (2003). Paleovegetation of marine isotope stages 4 and 3 in Northern New Zealand and the age of the widespread Rotoehu tephra. Quaternary Research 59, 420–429.
Smith, R. L. & Bailey, R. A. (1966). The Bandelier Tuff: a study of ash-flow eruption cycles from zoned magma chambers. Bulletin of Volcanology 29, 83–104.
Smith, V. C., Shane, P. & Smith, I. E. M. (2002). Tephro stratigraphy and geochemical fingerprinting of the Mangaone Subgroup tephra beds, Okataina Volcanic Centre, New Zealand. New Zealand Journal of Geology and Geophysics 45, 207–220.
Smith, V. C., Shane, P. & Nairn, I. A. (2004). 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 (in press).
Spera, F. J., Yuen, D. A., Greer, J. C. & Sewell, G. (1986). Dynamics of magma withdrawal from stratified magma chambers. Geology 14, 723–726.
Strelow, F. W. E. & von Toerien, F. S. (1966). Separation of lead (II) from bismuth (III), thallium (III), cadmium (II), mercury (II), gold (II), platinum (IV), palladium (II), and other elements by anion exchange chromatography. Analytical Chemistry 38, 545–548.
Sutton, A. N., Blake, S. & Wilson, C. J. N. (1995). An outline geochemistry of rhyolite eruptives from Taupo Volcanic Centre, New Zealand. Journal of Volcanology and Geothermal Research 68, 153–175.[CrossRef][Web of Science]
Sutton, A. N., Blake, S., Wilson, C. J. N. & Charlier, B. L. A. (2000). Later Quaternary evolution of a hyperactive rhyolite magmatic system: Taupo volcanic centre, New Zealand. Journal of the Geological Society, London 157, 537–552.
Thompson, B. N. (1968). Age of the Rotoiti Breccia. New Zealand Journal of Geology and Geophysics 11, 1189–1191.
Vucetich, C. G. & Pullar, W. A. (1969). Stratigraphy and chronology of late Pleistocene volcanic ash beds in central North Island, New Zealand. New Zealand Journal of Geology and Geophysics 12, 784–837.[Web of Science]
Walcott, R. I. (1989). Paleomagnetically observed rotations along the Hikurangi margin of New Zealand. In: Kissel, C. & Laj, C. (eds) Paleomagnetic Rotations and Continental Deformation. Dordrecht: Kluwer Academic, pp. 459–471.
Walker, G. P. L. (1979). A volcanic ash generated by explosions where ignimbrite entered the sea. Nature 281, 642–646.[CrossRef][Web of Science]
Wasson, K. & Kallemeyn, G. W. (1988). Compositions of chondrites. Philosophical Transactions of the Royal Society of London, Series A 325, 535–544.[CrossRef]
Wilson, C. J. N., Rogan, A. M., Smith, I. E. M., Northey, D. J., Nairn, I. A. & Houghton, B. F. (1984). Caldera volcanoes of the Taupo Volcanic Zone, New Zealand. Journal of Geophysical Research 89, 8463–8484.
Wilson, C. J. N., Houghton, B. F., Lamphere, M. A. & Weaver, S. D. (1992). A new radiometric age estimate for the Rotoehu Ash from Mayor Island volcano, New Zealand. New Zealand Journal of Geology and Geophysics 35, 371–374.
Wood, B. J. & Carmichael, I. S. E. (1973). Ptotal, PH2O, and the occurrence of cummingtonite in volcanic rocks. Contributions to Mineralogy and Petrology 40, 149–158.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. J. N. Wilson and B. L. A. Charlier Rapid Rates of Magma Generation at Contemporaneous Magma Systems, Taupo Volcano, New Zealand: Insights from U-Th Model-age Spectra in Zircons J. Petrology, May 13, 2009; (2009) egp023v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Deering, J. W. Cole, and T. A. Vogel A Rhyolite Compositional Continuum Governed by Lower Crustal Source Conditions in the Taupo Volcanic Zone, New Zealand J. Petrology, December 20, 2008; (2008) egn067v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. MOLLOY, P. SHANE, and I. NAIRN Pre-eruption thermal rejuvenation and stirring of a partly crystalline rhyolite pluton revealed by the Earthquake Flat Pyroclastics deposits, New Zealand Journal of the Geological Society, January 1, 2008; 165(1): 435 - 447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ban, H. Sagawa, K. Miura, and S. Hirotani Evidence for a short-lived stratified magma chamber: petrology of the Z-To5 tephra layer (c. 5.8 ka) at Zao volcano, NE Japan Geological Society, London, Special Publications, January 1, 2008; 304(1): 149 - 168. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Pichavant, F. Costa, A. Burgisser, B. Scaillet, C. Martel, and S. Poussineau Equilibration Scales in Silicic to Intermediate Magmas Implications for Experimental Studies J. Petrology, October 1, 2007; 48(10): 1955 - 1972. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||