JOURNAL OF PETROLOGY
VOLUME 51
NUMBER 12
PAGES 2465^2488
2010
doi:10.1093/petrology/egq064
The Magmatic Evolution of the Whakamaru
Supereruption, New Zealand, Constrained
by a Microanalytical Study of Plagioclase
and Quartz
K. E. SAUNDERS1*, D. J. MORGAN2, J. A. BAKER1 AND
R. J. WYSOCZANSKI3
1
SCHOOL OF GEOGRAPHY, ENVIRONMENT AND EARTH SCIENCES, VICTORIA UNIVERSITY OF WELLINGTON,
PO BOX 600, WELLINGTON 6014, NEW ZEALAND
2
SCHOOL OF EARTH AND ENVIRONMENT, EARTH SCIENCE BUILDING, THE UNIVERSITY OF LEEDS,
LEEDS LS2 9JT, UK
3
NATIONAL INSTITUTE OF WATER AND ATMOSPHERIC RESEARCH, PRIVATE BAG 14901, WELLINGTON 6041,
NEW ZEALAND
RECEIVED MARCH 23, 2010; ACCEPTED SEPTEMBER 21, 2010
ADVANCE ACCESS PUBLICATION NOVEMBER 23, 2010
The Whakamaru eruption is the largest-volume eruption known to
have originated from the hyper-productive Taupo Volcanic Zone,
New Zealand. Major, minor and trace element concentrations
of plagioclase crystals and cathodoluminescence images, used as a
proxy for Ti concentrations in quartz crystals, have been used to
explore their chemical zonation. Three plagioclase populations are
identified. Group 1 crystals are characterized by inherited cores of
composition An45^60, Ba 115^650 ppm and La 3^9 ppm, rims of c.
An30, Ba 450^800 ppm and La 7^10 ppm and the presence of
a thin overgrowth rim on several crystals cores. Group 2 crystals
are oscillatory-zoned plagioclases of composition An30^40, Ba
450^730 ppm and La 8·5^9·5 ppm. Group 3 plagioclase crystals
have cores of An25^35 and rims of An20^25 and low Sr contents
(280^480 ppm). From the chemical composition of these plagioclase
crystals, four physicochemically distinct rhyolitic melts are identified:
(1) an andesitic progenitor melt in which the cores of Group 1 crystals
crystallized; (2) a greywacke melt or greywacke protolith melt
responsible for narrow overgrowth rims on Group 1 crystal cores; (3)
melt derived from the rejuvenation of a mature crystal mush body
from which Group 3 plagioclase crystals crystallized; (4) a final,
rhyolitic melt created by the amalgamation of varying proportions
of the andesitic, greywacke-derived and rejuvenated melts with
*Corresponding author. Present address: Department of Earth
Sciences, University of Bristol, Wills Memorial Building, Queens
Road, Bristol BS8 1RJ, UK. Telephone: þ44 (0)117 3315131. Fax: þ44
(0)117 9253385. E-mail: [email protected]
subsequent, open-system fractional crystallization of a plagioclasedominant crystal assemblage. Cathodoluminescence imaging of
quartz crystals reveals complex zonation, the result of a dynamic crystallization history from potentially polygenetic sources. Diffusion
modelling of the greyscale intensity of cathodoluminescence images
(as a proxy for Ti content) for a selection of bright core^rim interfaces of quartz crystals suggests that renewed quartz growth at the
rim zones occurred5300 years (peak likelihood 50^70 years) prior
to and continued towards the climactic eruption. This is consistent
with timescales of 5280 years determined from core^rim interfaces
of Group 1 plagioclase crystals, suggesting that the magma chamber
was ephemeral, derived from mixing of magmas from multiple
sources shortly prior to eruption. This study adds to a growing
body of evidence for the ephemeral nature and geologically rapid
mixing and mobilization of liquid silicic magma bodies leading to
supereruptions, compared with the timescales of hundreds of thousands of years required to accumulate the precursor magma and
crystals.
KEY WORDS:
diffusion; plagioclase; quartz; Whakamaru ignimbrite;
timescale
ß The Author 2010. Published by Oxford University Press. All
rights reserved. For Permissions, please e-mail: journals.permissions@
oxfordjournals.org
JOURNAL OF PETROLOGY
VOLUME 51
I N T RO D U C T I O N
One of the fundamental questions that remains unanswered in relation to large-volume silicic eruptions is
the timescale of the petrogenesis and accumulation of
large volumes of eruptible magma. Differentiation processes in silicic magmas are speculated to occur over a
range of timescales. For example, deposits from Oligocene
supereruptions [defined in this study as eruptions producing 4300 km3 of magma after Sparks et al. (2005)] in
Yemen and Ethiopia post-date the flood basalts that they
are derived from by 51 Myr, indicating that the volumes
of silicic magma required to feed supereruptions can be
generated in 5106 years (Baker et al., 1996, 2000; Ukstins
Peate et al., 2005, 2007). Processes such as gravity settling
of crystals and the formation of crystal mush bodies occur
over timescales of 104^105 years (Bachmann & Bergantz,
2004) and the modelling of fractional crystallization of
both basaltic and rhyolitic magmas in a 10 km3 magma
chamber indicates that 50% crystallization could be
achieved in 53000 years (Hawkesworth et al., 2000). In
comparison, crustal anatexis instigated and driven by
mantle-derived mafic melts occurs rapidly, with the potential to form large volumes of melt in relatively short time
periods of 102^103 years (Huppert & Sparks, 1988;
Bergantz, 1989). Thus, detailing the longevity and the timescales over which large volumes of eruptible silicic magma
may accumulate sufficiently to feed potentially catastrophic supereruptions is fundamental to our understanding of
the dynamics of magmatic systems and the prediction of
future eruptions.
A contemporary view of sub-volcanic magmatic systems
is a complex one of interlinked sills, dykes and multi-level
storage areas that in a rhyolitic system culminates in a
magma reservoir composed of a crystal mush zone and
overlying melt-rich magma chamber (Annen & Sparks,
2002; Bachmann & Bergantz, 2003, 2004, 2008;
Bachmann et al., 2007; Davidson et al., 2007; Hildreth &
Wilson, 2007; Jerram & Davidson, 2007). Increasingly,
upper crustal magma chambers are being considered as
ephemeral bodies that can be replenished on much
shorter timescales than the formation and growth of the
whole magmatic system (e.g. Glazner et al., 2004) or the
remobilization of pre-existing crystal mush bodies (e.g.
Bachmann & Bergantz, 2004; Charlier et al., 2007). The
magma itself is formed from interstitial silicic melt extracted from large, underlying crystal mush bodies resident within the crust in a near-solid state, which may
represent a stage in batholith formation (Bachmann &
Bergantz, 2003; Glazner et al., 2004; Hildreth & Wilson,
2007; Wiebe et al., 2007). The derivation of large volumes
of silicic magma from semi-solid crustal bodies may explain the scarcity of melt bodies detected through geophysical techniques in active volcanic regions where
large-volume silicic eruptions are known to be common
NUMBER 12
DECEMBER 2010
(Bachmann & Bergantz, 2003, 2008; Glazner et al., 2004;
Bachmann et al., 2007). Additional constraints come from
thermal modelling of large intrusive bodies, which suggests that magma bodies with sizes of the order of
1200 km3 would solidify rapidly (103^105 years) in the absence of further thermal inputs; this leads to the conclusion that these bodies must be formed by an incremental
generation process, in which regions of partial melt are
expected to be ephemeral (Coleman et al., 2004; Glazner
et al., 2004).
Crystals hosted within volcanic rocks can record magmatic evolution within their structure, with each magmatic
process leaving a chemical or textural signature within
the crystal (Ginibre et al., 2002a, 2002b, 2004, 2007;
Davidson et al., 2007). Interpretation of the mineral chemistry provides the basis for identification of the magmatic
process(es) responsible for the chemical zonation, and the
chemical gradient between two adjacent zones within a
crystal may allow the timescales of the related processes
to be determined. Diffusion modelling, uniquely among
dating methods, has a strong temperature dependence,
which, combined with the ubiquity of zoning in crystals,
makes it possible to determine the relative timing of a
range of high-temperature magmatic processes from
magma degassing to crystal residence times (e.g. Zellmer
et al., 1999; Costa et al., 2003; Morgan et al., 2006; Kent
et al., 2007).
One of the most potentially dangerous regions of active
silicic volcanism is the highly productive Taupo Volcanic
Zone (TVZ), New Zealand. During the last 1·6 Myr, at
least 25 caldera-forming eruptions interspersed with smaller silicic eruptions are known to have originated within
the TVZ (Wilson et al., 2009). The average eruption rate
and magnitude of large-caldera eruptions from the TVZ
is comparable with that of the Southern Rocky Mountain
volcanic field and Yellowstone in the USA (Wilson et al.,
2009). Here we present major, minor and trace element
data for plagioclase crystals, imaging and diffusion studies
of Ti zoning within quartz crystals from the 330 ka
Whakamaru supereruption, currently the largest known
eruption to have originated from the TVZ. For two reasons, we have in this study focused on microanalytical
chemical characterization of plagioclase and quartz crystals from deposits of the Whakamaru supereruption. First,
plagioclase crystallizes over the entire igneous compositional spectrum from basalt to rhyolite and is usually
zoned, making it an ideal mineral to investigate the
evolution of silicic magmas (e.g. Ginibre et al., 2002a;
Berlo et al., 2007; Charlier et al., 2008). Second, recent studies have documented the utility of Ti zonation in quartz
crystals, illustrating its potential to record aspects of magmatic evolution (Liu et al., 2006; Wark et al., 2007; Shane
et al., 2008).
2466
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
W H A K A M A RU S U P E R E RU P T I O N
A N D SAMPLES
The continental TVZ in the SW Pacific is currently one of
the most active regions of silicic volcanism on Earth with
46000 km3 of dominantly rhyolitic magma erupted in the
last 1·6 Myr (Houghton et al., 1995; Wilson et al., 2009).
The region is characterized by thin (15^25 km), extending
(8^10 mm a1) continental crust, resulting in a graben
packed structure of volcanic infill with a high heat flow of
700 mW m2 (Cole, 1981; Stern, 1987; Darby et al., 2000;
Harrison & White, 2004). Recent silicic volcanism is confined to the central area of the TVZ, with the northern
and southern extremes characterized by andesitic volcanism (Wilson et al., 1995). One of the largest eruptive events
known to originate from this region is the 320^340 ka
Whakamaru eruptions, which occurred in two distinct periods [termed Whakamaru 1 and 2 by Wilson et al. (2009)]
following a 350 kyr hiatus in large-volume explosive rhyolitic volcanism (Houghton et al., 1995; Brown et al., 1998;
Wilson et al., 2009). The earlier of these two eruptive periods (c. 335 ka) deposited the Whakamaru Group ignimbrites, a large-scale crystal-rich, and in places highly
welded, group of ignimbrites blanketing a large area of
the central North Island of New Zealand with estimated
volumes exceeding 1500 km3 of magma (Briggs, 1976;
Brown et al., 1998; Wilson et al., 2009). The Whakamaru
Group ignimbrites consists of the Whakamaru, Manunui,
Rangitaiki, Te Whaiti, Wairakei and Paeroa Range Group
ignimbrites, which are dated to within 10 kyr of each
other and share similar petrographic features, cropping
out over an area of 13 000 km2 (Brown et al., 1998). The largest of these ignimbrites is the Whakamaru ignimbrite;
this is currently dated at c. 335 ka and is the focus of
this study (Fig. 1). A previous whole-rock study of the
Whakamaru group ignimbrites (Brown et al., 1998) indicated that at least four distinct rhyolitic magmas and a
high-alumina basalt were involved in the generation of
the magma body; a conclusion that our work refines and
broadly supports. Zircon geochronology has potentially
indicated that the Whakamaru magmatic system was
active for at least 250 kyr preceding the final catastrophic
eruption, which deposited the Whakamaru ignimbrite
(Brown & Fletcher, 1999; Charlier et al., 2005).
The mineral assemblage within the Whakamaru
ignimbrite consists of quartz, plagioclase, orthopyroxene amphibole, ilmenite, magnetite, K-feldspar, biotite, and
zircon (Brown et al., 1998; Table 1). The matrix consists of
highly welded glass at the based of the ignimbrite to consolidated ash at the top. Plagioclase crystals are generally euhedral to sub-euhedral and in some instances are
fragmented and broken where they fractured during eruption. A study of quartz-hosted melt inclusions (Saunders
et al., 2010) indicated that at least a proportion of the
quartz crystals could be considered as antecrysts, derived
from an evolved crystal mush or near-solid plutonic body.
Plagioclase and quartz crystals studied in this work were
sampled from the accessible parts of a c. 100 m section of
the ignimbrite at Maraetai Dam (Figs 1 and 2; Table 1).
Samples were taken from a discontinuous crystal lag at
the base of the ignimbrite (WH1) and at higher stratigraphic levels (WH2, WH3, WH4, WH5, WH6).
A N A LY T I C A L M E T H O D S
Sample preparation
Samples were lightly crushed using a mortar and pestle
and dry sieved to extract material 51mm. Fine glass was
floated off and the remaining material dried at 408C for
12^24 h and then magnetically separated. Plagioclase,
quartz and orthopyroxene crystals were hand picked
under a binocular microscope from the 500^1000 mm size
fraction for analysis. These crystals were mounted in
epoxy blocks and polished with 600 and 2400 grit SiC
paper and then finely polished using 3 mm and 0·25 mm
diamond paste. In addition, thick (100 mm) polished sections of Whakamaru pumices were also prepared and
used for plagioclase analyses.
Electron microprobe analyses
Major element compositions of plagioclase and Fe^Ti
oxides were determined on a JEOL 733 SuperProbe
Electron probe microanalyser (EPMA) at Victoria
University of Wellington (VUW). All crystals were analysed at 15 kVaccelerating voltage and 12 nA beam current.
Plagioclase crystals were analysed with a 10 mm spot size
at 10 mm intervals. Fe^Ti oxides were analysed at the same
conditions with a focused beam. ZAF corrections (Bence
& Albee, 1968) were used for data reduction. Primary calibrations used a mixture of natural and synthetic standards.
Secondary standards of basaltic glass A99 (USNM
113498/1) for plagioclase analyses and ilmenite standard
USNM 96189 for Fe^Ti oxides analyses (Jarosewich et al.,
1980) were analysed at the beginning and end of each analytical session (Table 2). Quality control was applied to
the plagioclase data and only analyses with totals between
97^102 wt% are reported. Cathodoluminescence (CL)
images of quartz crystals were obtained using a photomultiplier attached to the electron microprobe at the
same analytical conditions as described above. The relatively long decay time of the CL process can in theory
lead to blurry images under rastered-beam imaging
(Reed, 1996). However, examination of the images revealed
that distinct, sharp boundaries are visible, suggesting that
this effect was not significant.
Laser ablation inductively coupled plasma
mass spectrometry
Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) analyses of plagioclase crystals were
2467
JOURNAL OF PETROLOGY
VOLUME 51
NUMBER 12
DECEMBER 2010
Fig. 1. Map of the Taupo Volcanic Zone, New Zealand, showing the location of the Whakamaru caldera, the extent of the Whakamaru ignimbrite and sample location (white circle) of this study. The extent of the Whakamaru ignimbrite is taken from Brown et al. (1998). Inset shows
the regional setting of the Taupo Volcanic Zone with the outline of the Taupo Volcanic Zone shown by a bold dashed line.
Table 1: Sample locations and mineralogy of the studied samples
Sample no.
Latitude (S)
Longitude (E)
Altitude (m)
Mineralogy
WH1
38821·016’
175831·020’
WH2
38820·849’
175844·206’
170
plag; qtz; opx; Fe–Ti oxides
WH3
WH4
38820·953’
175844·034’
250
plag; opx; qtz; amp; bt; Fe–Ti oxides
38820·782’
175844·033’
246
plag; qtz; opx; amp; Fe–Ti oxides
WH5
38820·782’
175844·033’
246
plag; qtz; opx; amp; Fe–Ti oxides
WH6
38820·737’
175844·110’
211
plag; qtz; opx; amp; K-spar; Fe–Ti oxides
WH7
38820·799’
175844·144’
196
plag; qtz; opx; amp; Fe–Ti oxides
plag; qtz; opx; amp; bt; K-spar; Fe–Ti oxides
Mineral phases are listed in order of decreasing abundance. plag, plagioclase; qtz, quartz; opx, orthopyroxene; amp,
amphibole; bt, biotite; K-spar, K-feldspar.
2468
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 2. Schematic stratigraphic section through the ignimbrite indicating the location of samples with respect to the Briggs (1976) subdivisions
(D, E, F) and the abundance of the different plagioclase groups (this study) with stratigraphic height (in metres).
Table 2: Standard calibrations and reference values used for EPMA in this study
A99, Basalt glass
Mean
2 SD
Ilmenite standard USNM 96189
Ref.
Mean
n ¼ 19
% 2SD
% diff.
values
n¼6
values
(wt %)
(wt %)
(wt %)
(wt %)
50·9
1·44
2·8
0·1
50·9
Al2O3
12·4
0·57
4·6
0·4
12·4
FeO
13·3
0·63
4·7
0·5
13·3
0·11
2·1
1·6
SiO2
TiO2
MgO
5·16
5·08
MnO
CaO
9·15
0·38
4·2
1·6
Na2O
2·60
0·19
7·2
2·2
2·66
0·84
0·07
8·2
2·7
0·82
94·5
47·6
1·55
46·1
2·32
0·32
0·07
4·61
0·22
% 2SD
% diff.
Ref.
3·2
4·3
45·7
5·0
0·9
46·5
22
4·7
3·1
0·31
3·4
4·77
9·30
K2O
Total
2 SD
94·5
92·5
Reference values of basaltic glass A99 and ilmenite standard USNM 96189 (Jarosewich et al., 1980) for electron microprobe analyses. n, number of analyses averaged.
performed in the Geochemistry Laboratory at VUW with
a NewWave deep UV laser (193 nm solid state) coupled to
an Agilent 7500cs ICP-MS system using helium as the carrier gas. CaO was determined prior to LA-ICPMS analysis
by EPMA, allowing 43Ca to be used as the internal standard. Samples were analysed for Sr, Ba and Mg with
20 mm laser spot diameters at 20 mm intervals along the
EPMA profiles. Core and rim analyses of rare earth elements (REE), Y and Pb were also carried out using 35 or
50 mm laser spot diameters, a repetition rate of 5 Hz and
power of 5·6^6·7 J cm2. The NIST SRM 610 glass
standard was used as the primary standard and for signal
optimization at the start and throughout each analytical
session. NIST SRM 610 reference values are taken as the
2469
JOURNAL OF PETROLOGY
VOLUME 51
NUMBER 12
DECEMBER 2010
Table 3: Repeat analyses of the MPI-DING standard glass STH-G at 20 mm and 35 mm laser spot diameters
Isotope
Element
Spot size
Mean (n ¼ 5)
2 SD
24
Mg
20
12643
155
1·2
6·4
11879
25
Mg
20
11234
393
3·5
5·4
11879
88
Sr
20
486
16
3·2
0·8
482
137
Ba
20
306
23
7·4
2·7
298
89
Y*
35
10·9
0·3
2·7
4·5
11·4
139
La*
35
11·8
0·6
4·7
1·5
12·0
140
Ce*
35
25·5
0·5
2·1
2·4
26·1
141
Pr*
35
0·13
4·5
7·1
146
Nd*
35
1·0
7·7
0·9
2·97
13·1
% 2SD
% diff.
Ref. values
3·20
13·0
147
Sm*
35
2·76
0·39
2·8
0·6
153
Eu*
35
0·917
0·069
7·5
1·0
157
Gd*
35
2·26
0·29
12·9
12·6
2·59
163
Dy*
35
2·16
0·29
13·3
2·8
2·22
166
Er*
35
1·10
0·08
7·4
7·0
1·18
172
Yb*
35
1·02
0·11
10·3
9·8
1·13
175
Lu*
35
0·157
0·027
17·4
208
Pb*
35
8·29
0·42
5·1
0·17
19·5
2·78
0·950
0·170
10·3
Reference values (NIST SRM 610 or 612) are taken from GeoRem preferred values (http:///georem.mpch-mainz
.gwdg.de/).
*Unpublished data of J. A. Baker at 35 mm laser spot diameters and NIST SRM 612 as primary standard.
GeoRem preferred reference values (http:///georem
.mpch-mainz.gwdg.de/) (Table 3). Repeat analyses of
MPI-DING glass STH-G at 20 mm laser spot diameters at
the same analytical conditions as described above are
given in Table 3. This indicates an elemental uncertainty
(2s) of 6·5% for Sr, Ba and Mg at 20 mm laser
spot diameters. We estimate an elemental uncertainty of
510% (2s) for REE, Yand Pb when analysed at 35 and
50 mm laser spot diameters from repeat analysis of STH-G
at 35 mm laser spot diameters with NIST SRM 612 as the
primary standard at the analytical conditions described
above (Table 3). Correlation between EPMA and
LA-ICPMS analytical areas was achieved through detailed
backscattered electron images taken before and after both
EPMA and LA-ICPMS analyses.
R E S U LT S
Plagioclase
Backscattered electron images of 4100 plagioclase crystals
from seven samples of the Whakamaru ignimbrite were obtained and revealed subtle, yet ubiquitous, zoning of crystals (Electronic Appendices 1 and 2, available at http://
www.petrology.oxfordjournals.org). Of these, 61 crystals
were analysed for core and rim compositions, and major
element profiles were measured across 22 crystals
(Electronic Appendix 3). Representative compositions are
given in Table 4. Sr, Ba and Mg profiles were obtained for
eight of these 22 crystals, and spot analyses of REE, Y
and Pb were obtained for a representative selection of
crystals based on textural appearance and anorthite content (Electronic Appendix 4). Representative analyses are
given in Table 5.
On the basis of textural observations and anorthite (An)
content (Fig. 3), plagioclase crystals fall into three groups,
denoted 1, 2, and 3 (Fig. 4). Most plagioclase crystals
belong to Group 2 (77%), with only relatively minor quantities of Group 1 (11%) and Group 3 (12%) crystals being
observed, although the abundance of Group 1 crystals
may be biased if the sectioning of crystals was off-centre
and the sectioning process did not intersect the core (see
Electronic Appendices 1^3). There is no observed systematic variation in the abundance of plagioclase populations
with stratigraphic height in the ignimbrite
Group 1
Group 1 crystals are characterized by cores of An45^60
mantled by concentric, relatively homogeneous rims of c.
An30 (Fig. 4a). Trace element concentrations in the cores
and rims are distinct, with elevated Ba (c. 400^800 ppm),
La (7^10 ppm) (Fig. 5) and Ce (9^13 ppm) concentrations
in the rims compared with the cores at similar Sr
2470
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Table 4: Representative major element analyses (wt %) of plagioclase crystals
Sample:
WH1h_8
WH1h_38
WH6da_2
WH6da_38
WH2a_3
WH3e_4
WH2g_x2
WH3b_4
Group:
1 rim
1 core
1 rim
1 core
2
2
3
3
SiO2
60·8
53·3
59·9
56·4
60·6
60·5
61·4
60·5
Al2O3
25·4
29·9
24·8
28·8
24·2
24·7
24·0
24·5
FeO
0·31
CaO
6·67
Na2O
7·82
K2O
Total
XAn
0·72
101·8
0·31
0·24
10·8
5·16
0·30
0·21
0·20
0·30
0·28
0·24
6·68
10·46
6·37
6·65
5·74
6·06
7·29
5·59
7·49
7·35
8·33
7·98
0·64
99·7
0·25
99·6
0·54
0·71
101·7
0·32
0·65
99·7
0·50
0·71
100·2
0·31
0·17
100·4
0·32
0·67
99·9
0·27
0·28
XAn calculated on a mole fraction basis.
Table 5: Representative REE, Sr, Yand Pb analyses of Whakamaru plagioclases
Sample:
WH1h_32
WH1h_34
WH1h_35
WH3e_45
WH3e_49
WH6b_50
WH6b_52
Position:
rim
core
core
rim
core
core
rim
Group:
1
1
1
2
2
3
3
XAn
0·30
CaO
6·28
Sr
537
0·54
10·9
697
0·56
11·4
768
0·32
0·29
0·28
6·51
5·94
5·86
547
590
485
0·21
4·27
332
Y
0·2
0·7
0·2
0·1
0·1
0·2
0·1
La
9·0
4·0
3·4
9·6
9·2
12·5
7·7
Ce
10·1
6·2
4·8
10·7
9·4
12·8
8·8
Pr
0·79
0·63
0·45
0·79
0·61
0·91
0·63
Nd
2·4
2·4
1·5
2·3
1·7
2·6
2·2
Sm
0·34
0·47
0·27
0·23
0·15
0·15
0·16
Eu
1·59
1·74
1·94
1·51
1·4
2·65
1·43
Gd
0·10
0·43
0·26
0·06
0·15
b.d.
0·02
Tb
b.d.
0·02
b.d.
b.d.
0·02
0·01
0·01
Dy
0·11
0·27
0·09
b.d.
0·04
b.d.
0·03
Ho
b.d.
0·02
0·01
b.d.
b.d.
0·01
0·01
0·02
Er
b.d.
0·03
0·04
b.d.
b.d.
b.d.
Tm
b.d.
b.d.
b.d.
b.d.
b.d.
b.d.
0·01
Yb
b.d.
0·09
0·11
b.d.
b.d.
b.d.
0·06
Lu
b.d.
0·01
0·01
b.d.
0·01
b.d.
0·01
Pb
6·28
4·4
4·89
6·31
6·42
11·02
6·56
b.d., below detection limit.
concentrations. Texturally, two distinct core types are
observed: (1) crystals with a diffuse core^rim interface
such as WH1_h and WH2_b; (2) crystals with partially resorbed cores and a sharp core^rim interface, such as
WH2_d (Electronic Appendices 1 and 2).
Group 2
Group 2 plagioclase crystals exhibit pronounced oscillatory zoning with a limited range in composition An30^40
(Figure 4). High An spikes (c. An36^40) are observed and
these correspond to micro-inclusion trails within the
2471
JOURNAL OF PETROLOGY
VOLUME 51
NUMBER 12
DECEMBER 2010
Fig. 3. Partial ternary feldspar diagram for Whakamaru plagioclase crystals showing the anorthite composition of the three plagioclase groups.
Number of analyses ¼1246.
crystal, which could indicate that resorption and/or regrowth of the crystal had occurred. Based on textural appearance and An composition, crystal WH3e appears to
be representative of Group 2 crystals and was analysed
for trace elements. With the exception of four analyses
from the crystal core region of WH3e, Sr (dominantly
550^630 ppm), Ba (dominantly 450^730 ppm), Pb and
REE concentrations in Group 2 crystals overlap those of
Group 1 rims, which also have similar An compositions.
The core region analyses exhibit anomalously high Sr
(625^815 ppm) and Ba (1010^1325 ppm) concentrations
that are not reflected in the Pb and REE analyses (Fig. 5).
Group 3
These crystals are characterized by cores of An25^35
and rims of An20^25, and low Sr (280^480 ppm) and Mg
(19^32 ppm) concentrations compared with Group 1 and
2 plagioclase crystals (Fig. 4). These low abundances of Sr
and Mg, together with low Ba abundances, indicate that
the melt from which these crystals originated was not the
same as the melt that Group 1 or 2 crystals crystallized
from and was a separate melt from the one in which the
Group 3 crystals were erupted. Pb and REE concentrations in the cores of Group 3 plagioclase crystals show the
most extreme enrichments observed amongst the plagioclase crystal populations, with elevated Pb, Eu, and La
concentrations relative to Group 1, Group 2 and Group 3
rim concentrations (Fig. 5).
Quartz
Cathodoluminescence (CL) images reveal complex zonation within Whakamaru quartz crystals (Fig. 6;
Electronic Appendix 5), demonstrating a growth history
just as dynamic as that of the plagioclase crystals.
Oscillatory zoning is ubiquitous and in some cases is
truncated by later overgrowths (e.g. WH1_39), demonstrating the presence of distinct core regions and multiple resorption events. The nature of the zoning in the crystals
shows the presence of several distinct zoning patterns,
which we interpret as indicating populations comparable
with those seen in plagioclase. Several quartz crystals
(e.g. WH1_37, WH2_3, WH2_6) display a bright-CL rim,
implying that these crystals experienced a common final
stage of magmatic evolution. However, the presence of a
bright-CL core in crystal WH1_37 in comparison with
those of WH2_3 and WH2_6 indicates that the early
period of magmatic evolution represented by crystal cores
is unique to each crystal. In contrast, crystals WH1_40,
WH2_4, WH2_5 and WH2_8 exhibit strongly oscillatory
zoned rims indicative of fluctuating magmatic conditions
during the final stages of quartz growth. The intensity of
CL zoning can be used as a proxy for Ti concentrations,
with darker regions indicating lower Ti concentrations
that result from a change in either the magmatic temperature or the Ti activity of the melt (Wark & Spear, 2005).
We exploit the variations in CL intensity as a proxy for Ti
content in the diffusion modelling section that follows, in
a manner analogous to that of Wark et al. (2007).
Fe^Ti oxides
Fe^Ti oxides are present both as free crystals and hosted
within orthopyroxene phenocrysts. Free Fe^Ti oxides were
not analysed, as it is impossible to determine the relationship of these crystals to others prior to the crushing of
pumices and separation of crystals. Fe^Ti oxide crystals
hosted within the same orthopyroxene crystals were analysed from sample WH1 (Table 6); Fe^Ti oxides in only six
of these crystals were fully enclosed in the orthopyroxene
host, and we have assumed that this allowed the majority
of the Fe^Ti oxide pairs to equilibrate with the external
2472
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 4. Representative backscattered electron images and XAn, Sr and Mg concentrations along rim^rim profiles for each of the plagioclase
populations. The start (S) and end (E) of the EPMA and LA-ICPMS profiles are shown.
melt prior to eruption. Equilibrium between oxides was
tested using the method of Bacon & Hirschmann (1988),
with temperatures and oxygen fugacity (fO2) calculated
using the method of Sauerzapf et al. (2008) (Table 6).
Geothermometry
Fe^Ti oxides are useful probes of magmatic temperature
with the ability to record heating events on the timescales
of hours to days prior to eruption (Venezky &
2473
JOURNAL OF PETROLOGY
VOLUME 51
Fig. 5. Measured Sr, Ba, La and Pb concentrations of plagioclase crystals.
2474
NUMBER 12
DECEMBER 2010
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 6. CL images for a selection of quartz crystals from WH1 and WH2. White boxes denote areas selected for diffusion modelling with the respective timescales indicated in years.
2475
JOURNAL OF PETROLOGY
VOLUME 51
NUMBER 12
DECEMBER 2010
Table 6: Fe^Ti oxides major element compositions and calculated temperature and oxygen fugacity
Ilmenite
SiO2
TiO2
Al2O3
FeO
WH1_O12-1
0·05
47·9
WH1_O18-1*
0·05
48·7
0·11
49·8
WH1_O21-2*
0·00
49·03
WH1_O23-2*
0·042
49·02
WH1_O24-1*
0·042
49·23
WH1_O26-2*
WH1_21-O1
0·05
0·07
49·37
48·30
0·13
0·10
0·18
0·14
0·07
0·16
48·79
50·53
49·57
49·13
48·61
48·26
0·83
MnO
0·85
0·77
0·83
0·87
0·93
0·89
MgO
1·85
1·97
1·76
1·81
1·78
1·83
1·73
Total
100·61
100·42
102·26
101·45
101·20
100·77
99·28
Fe2O3
12·10
10·20
11·60
10·60
49·20
9·20
9·60
FeO
39·00
39·60
40·10
40·00
37·90
40·30
39·60
WH1_O24-2
WH1_O26-1
WH1_21-O2
Magnetite
WH1_O12-2
WH1_O18-2
WH1_O21-1
WH1_O23-1
SiO2
0·06
0·14
0·08
0·076
0·07
0·14
0·09
TiO2
8·80
9·31
9·30
9·57
9·32
9·47
9·22
Al2O3
1·41
1·55
1·44
1·45
1·43
1·50
81·52
81·22
81·72
82·28
FeO
82·1
81·5
1·50
81·9
MnO
0·49
0·52
0·62
0·522
0·58
0·52
0·63
MgO
0·75
0·95
0·88
0·89
0·91
0·82
0·86
Total
92·96
93·54
93·96
94·72
94·32
93·79
94·13
Fe2O3
49·20
48·30
48·90
48·90
9·90
48·20
49·10
FeO
37·30
37·70
37·70
38·30
40·20
38·10
37·80
T (8C)
752
744
753
749
737
735
735
fO2
–14·1
–14·5
–14·2
–14·4
–14·7
–14·8
–14·7
Ilmenite
WH1_24-O2
WH1_26-O2
WH1_27-O2
WH1_30-O2*
WH1_31a-2
WH1_32-O2*
WH1_32-O3*
SiO2
0·00
0·05
0·07
0·00
0·05
0·01
0·00
TiO2
45·30
48·65
45·59
45·44
45·20
45·73
45·31
Al2O3
0·11
0·08
0·12
0·1
0·16
0·11
0·14
FeO
49·96
49·38
50·09
49·86
50·19
50·02
50·80
MnO
0·86
0·94
0·82
0·80
0·87
0·97
0·84
MgO
1·75
1·89
1·80
1·68
1·76
1·69
1·65
Total
97·98
100·94
98·48
97·88
98·23
98·53
98·74
Fe2O3
14·70
11·00
14·50
14·20
15·00
14·30
15·40
FeO
36·80
39·50
37·10
37·10
36·70
37·20
37·00
Magnetite
WH1_24-O3
WH1_26-O3
WH1_27-O1
WH1_30-O1
WH1_31a-1
WH1_32-O1
WH1_32-O4
SiO2
0·02
0·08
0·05
0·01
0·04
0·06
0·07
TiO2
8·63
9·23
8·68
8·50
8·80
8·43
8·69
Al2O3
1·40
1·50
1·49
1·38
1·50
1·51
1·43
84·33
83·42
83·61
0·51
FeO
82·8
81·5
83·4
83·5
MnO
0·53
0·56
0·44
0·56
0·59
0·67
MgO
0·76
0·87
0·86
0·83
0·87
0·85
0·87
Total
94·16
93·69
94·98
94·77
96·14
94·93
95·17
Fe2O3
50·50
48·70
50·90
51·30
51·50
51·30
51·00
FeO
37·40
37·70
37·70
37·30
38·00
37·30
37·70
T (8C)
772
757
770
763
775
762
776
fO2
13·5
14·1
13·5
13·7
13·4
13·7
13·4
*Fe–Ti oxide pairs exposed to the external melt.
2476
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Rutherford, 1999). Coexisting Fe^Ti oxides from WH1 were
used to infer a final mean magmatic temperature of
756 308C (2SE) (Table 6).
A longer record of magmatic temperature is potentially
preserved in plagioclase and quartz crystals as the major
element diffusion is slow, retaining temperature evidence
for earlier periods of the magmatic history compared with
the Fe^Ti oxides (Grove et al., 1984; Morse, 1984; Giletti &
Casserly, 1994; Venezky & Rutherford 1999; Cherniak
et al., 2007). Application of the plagioclase^melt equilibrium approach of Putirka (2008) has the potential to
determine magmatic temperatures for different zones of
the plagioclase crystal if a suitable melt composition can
be constrained. The presence of three distinct plagioclase
populations indicates the existence of multiple magma
batches. Therefore, to employ a single constant melt composition in association with the plagioclase^melt equilibria
may provide erroneous temperature estimates. As Group 1
rims and Group 2 plagioclase crystals appear to be in equilibrium with the groundmass glass (see below) and make
up the majority of the crystals, a mean groundmass glass
composition (Brown, 1994) is taken as the liquid composition and this is used to calculate magmatic temperatures.
A summary of calculated magmatic temperatures is given
in Table 7, and indicates a mean temperature of 8308C for
Group 1 rims and Group 2 crystals. Furthermore, this
would indicate that Group 1 cores fractionated from
magma with a temperature of c. 8508C or higher if, as
discussed below, the cores of Group 1 crystals fractionated
from an andesitic precursor melt. Magmatic temperatures of Group 3 plagioclase crystals are more equivocal.
A groundmass glass composition yielded magmatic
temperatures similar to Group 1 rims and Group 2 crystals, although it is unlikely that these crystals were in equilibrium with the final groundmass glass composition (see
below), indicating that these temperature estimates are
misleading. Thus, a magmatic temperature of 8308C was
used for the diffusion modelling of Group 1 plagioclase
crystals.
DI FFUSION MODELLI NG
Plagioclase crystals
We have assumed that during plagioclase crystallization
the crystal rim maintained local equilibrium with
the surrounding melt. As a result of the slow, coupled
inter-diffusion of NaSi^CaAl, the original anorthite composition of Whakamaru plagioclase crystals would have
been retained at length scales 45 mm (Grove et al., 1984;
Morse, 1984). The incorporation of trace elements
into the plagioclase crystals would have been governed by
well-constrained partitioning relationships (Blundy &
Wood, 1991, 1994; Bindeman et al., 1998). As a result of fluctuating melt conditions, disequilibrium can arise between
adjacent crystal zones. Changes in the magma conditions
Table 7: Summary of magmatic temperature from plagioclase^melt equilibria
Group
Sample
n
Minimum
Maximum
Average
temperature
temperature
temperature
(8C)
(8C)
(8C)
Rims
1
WH1_h
33
831
834
832
1
WH2_b
59
830
836
833
1
WH2_d
52
830
838
834
1
WH6_da
36
831
835
833
1
WH6_db
29
860
807
836
1
WH6_f
52
831
837
834
1
WH1_h
28
849
856
853
1
WH2_b
30
836
850
844
1
WH2_d
24
832
860
855
1
WH6_da
10
837
854
848
1
WH6_db
7
838
860
850
1
WH6_f
10
850
860
855
2
WH1_b
98
829
846
833
2
WH2_a
151
830
842
834
2
WH2_f
69
832
841
834
2
WH3_c
57
831
838
834
2
WH3_d
58
828
841
836
2
WH3_e
88
831
840
834
2
WH5_d
66
832
844
837
2
WH5_f
61
830
838
834
2
WH6_e
47
831
838
834
2
WH6_g
65
830
843
834
3
WH2_g
42
830
840
833
3
WH3_b
51
826
842
831
3
WH6_b
42
825
832
828
3
WH6_c
90
825
840
833
Cores
n, number of plagioclase analyses from each crystal used
to calculate average. Water content of melt taken as
3·42 wt % (Saunders, 2009). Full temperature determinations are given in Electronic Appendix 3.
(pressure, temperature, H2O budget or melt composition)
result in subsequent plagioclase growth at different
anorthite concentrations (e.g. Bowen, 1913; Johannes, 1989;
Putirka, 2008; Ginibre et al., 2007) with concomitant stepchanges in trace element concentrations. With a stepchange in anorthite content, a step-change in trace element
concentrations at equilibrium is also likely, as a result of
composition-dependent partitioning. However, following
a magma mixing event, or change in temperature, the
pre-existing crystal composition is no longer in equilibrium with the melt or, by extension, with material
2477
JOURNAL OF PETROLOGY
VOLUME 51
crystallized from that melt. This produces a chemical potential gradient within the crystal, which drives subsequent
diffusion. The induced trace-element disequilibrium of
this step-change between two adjacent zones is then gradually erased by a subsequent diffusion-governed homogenization. If equilibrium is not reattained by the time of
eruption, then a compositional profile incorporating some
extent of disequilibrium will be ‘frozen’ into the crystal by
the rapid drop in diffusion rates as the crystal temperature
drops following eruption.
The chemical potential gradient that is the driving force
for diffusion can be approximated by using standard partition coefficients to translate element abundance into a
melt equivalent. The partition coefficients of adjacent crystal zones govern the relative equilibrium concentrations of
trace elements between the two regions of the crystal.
Therefore, two adjacent zones of plagioclase with differing
An contents will not necessarily possess the same trace
element concentrations when fully equilibrated (e.g.
Zellmer et al., 1999; Costa et al., 2003; Zellmer & Clavero,
2006) although, locally, the chemical potential will be in
equilibrium. If equilibrium has not been attained, the
length of time a crystal interface resided at magmatic temperature can be modelled with the appropriate solution to
Fick’s second law and calculated partition coefficients (K)
(e.g. Crank, 1976; Zellmer et al., 1999; Costa et al., 2003;
Morgan et al., 2004, 2006; Zellmer & Clavero, 2006).
Partition coefficients for Sr and Ba in plagioclase crystals
can be calculated from the relationships determined by
Blundy & Wood (1991); these relate the trace element
concentration of the plagioclase to that of the melt from
which they fractionated:
RT ln KSr ¼ 26 800 26 700XAn
ð1aÞ
RT ln KBa ¼ 10 200 38 200XAn
1
ð1bÞ
1
where R is the gas constant (8·3145 J mol K ), T is the
temperature in Kelvin and XAn is the An content on a
mole fraction basis. These partition coefficients can then
be used to calculate equilibrium melt concentrations of Sr
and Ba that would be in instantaneous equilibrium with
that crystal zone, following the method described by
Costa et al. (2003) and the coexisting melt compositions at
the time the plagioclase crystallized.
Diffusion modelling based on Sr profiles
in plagioclase
The length of time the core^rim interface in Group 1 crystals resided at magmatic temperatures can be estimated
through analysis of Sr diffusion profiles. Sr diffusion modelling is considered appropriate because the diffusivity of
Sr in plagioclase is well constrained (Giletti & Casserly,
1994). Diffusivity of Sr (D) is assumed to be isotropic
within analytical error (Giletti & Casserly, 1994; Zellmer
et al., 1999) and is calculated using the pre-exponential
NUMBER 12
DECEMBER 2010
factor, D0 ¼ 10ð41XAn þ408Þ, with an activation energy of
276 kJ mol1 and a magmatic temperature of 8308C.
Two crystals, WH6_da and WH6_f, were selected for
modelling using a finite-difference model (FINDIF,
Martin et al., 2008) similar to that used by Costa et al.
(2003). This allows consideration of the initial condition,
with full treatment of the variable diffusion coefficient as
a function of mol % An. This model shows that much of
the variation in the crystals has to be due to growth zonation. The spatial variation in diffusivity caused by the
variation in An content indicates that the variation in the
data cannot be reconciled with a simple model with a
single jump in composition. It is also clear that the short
timescales returned indicate that diffusion of the Sr signal
in these crystals is therefore close to the spatial resolution
of the LA-ICPMS data available, and is probably less
than 60 mm. Thus the conclusion is drawn that these
crystal rims were generated in 5280 years, equivalent
to the resolution of our measurement technique. The development of Group 1 rims therefore shows that remobilization of Group 1 crystal cores occurs relatively rapidly
before eruption, within a time interval short enough for
diffusion of Sr to be at the limit of resolution of our
LA-ICPMS technique. To take this study further, we have
therefore investigated Ti diffusion in quartz.
Diffusion in quartz crystals
Several of the imaged quartz crystals display a pronounced
zone of high CL emission intensity at or near the rim of
the crystals (Fig. 6). Irrespective of whether the higher Ti
growth zones result from an increase in magmatic temperature, a change in the Ti activity of the melt or a pressure change, the interfaces between low- and high-Ti
zones are distinct. Using the CL intensity as a proxy for
Ti concentration, Ti diffusion across this interface can be
modelled, allowing the timescale between the renewed
growth of the quartz crystal and eruption to be estimated
(e.g. Wark et al., 2007). The rounded nature of the inner
crystal cores indicates a period of resorption that occurred
before the final overgrowth and therefore this boundary
cannot be used to assess the timing of the final recharge
event, representing as it does an earlier stage of the system.
As diffusion spans only a small fraction of the two adjacent regions, a simple, one-dimensional diffusion model is
appropriate. The diffusion model presented by Morgan
et al. (2004, 2006) has been applied. Ti diffusivity in
quartz is calculated using a D0 of 7·01 108 m2 s1 (log
D0 ¼ ^7·154 0·525), an activation energy of 273 12 kJ
mol1 (Cherniak et al., 2007) and a temperature of 7568C
derived from Fe^Ti oxides, as this best represents the magmatic temperature immediately prior to eruption when
this final growth would have occurred. A step-function is
assumed at the interface and therefore this results in the
calculation of maximum timescales for each boundary
modelled. Uncertainties are estimated by propagating the
2478
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 7. Probability distributions of each diffusional timescale calculated in this study as a function of the 2s temperature uncertainty for quartz
crystals. Fine lines denote the probability distribution for each timescale and the bold line denotes the entire population. The regions modelled
are highlighted by the white boxes in Fig. 6.
2s temperature uncertainty of our Fe^Ti oxide temperatures determinations of 30 K onto the calculated diffusivities. Propagating the measured analytical errors of the
pre-exponential factor and activation energy onto the calculated diffusivities results in errors that are relatively
small (e.g. 0·03 log10 units for 1s) compared with the determined errors owing to the temperature uncertainties
(e.g. 0·20 log10 units for 1s) and this yields a combined calculated 1s uncertainty of 0·20 log10 units. The results are
illustrated as probability distributions owing to the temperature uncertainty in Fig. 7. This suggests that renewed
quartz growth occurred less than 200 years prior to eruption, with a peak likelihood in the 50^70 years region but
up to 170 years. These results are potentially similar to the
timescales determined from plagioclase crystals if better
analytical resolution could be attained (Figs 6 and 7;
Table 8). The better time resolution of quartz in this
sample is a direct consequence of the higher spatial resolution of CL imaging compared with LA-ICPMS.
However, because of the disparity in calculated magmatic
temperatures of the various mineral phases and the
Table 8: Summary of diffusional timescales calculated from
quartz crystals
Sample
Diffusional
Residence
half-width (mm)
time (years)
WH1_37
4·60
160
WH2_3
5·92
172
1
2·97
79
2
2·38
46
1
2·97
77
2
4·16
137
2·97
52
WH2_4
WH2_5
WH2_8
Boundaries modelled are highlighted in Fig. 6.
potential differences in timescales (quartz crystals c. 50^70
years and plagioclase crystals of 5300 years), it is suggested that although these mineral phases were integrated
into the final magma body during the 300 years preceding
2479
JOURNAL OF PETROLOGY
VOLUME 51
eruption, they may not have been incorporated during the
same magmatic event.
DISCUSSION
None of the plagioclase or quartz crystals examined yield
identical values for the calculated timescales, indicating
that, within the scope of the study, each of these crystals
experienced a unique history. However, each crystal population possesses common features that allow the successive
magmatic evolution as recorded from the core to the rim
of the crystals to be unravelled. The compositional variability of the melt bodies from which the plagioclase crystals fractionated is explored by determining the Sr and Ba
melt compositions in equilibrium with the crystal throughout its growth [using equations (1a) and (1b); Blundy &
Wood (1991)] and comparing these with known melt compositions (Fig. 8). However, caution must be applied as
some diffusion has inevitably modified the Sr and Ba concentrations of inherited plagioclase cores. Nevertheless,
these calculations can be interpreted to indicate the presence of at least four, compositionally distinct rhyolitic
melts during the formation of the Whakamaru magma
(Figs 8 and 9). Three of these four rhyolitic melts can be
equated with the melt batches identified in the earlier
whole-rock pumice study by Brown et al. (1998) and a new,
previously unknown, rhyolitic melt composition is revealed
(Figs 8 and 9) consistent with quartz-hosted melt inclusion
compositions from the cores of the crystals (Saunders
et al., 2010). Furthermore, the location of each of the identified melt compositions within single plagioclase crystals
permits the timing and mutual relationships of these melts
to be unravelled.
Origins of Group 1 plagioclase crystals
The documented compositional variability of Group 1
cores and rims points to at least two distinct periods of
crystallization. Importantly, the change in composition
from core to rim can be used to trace the successive magmatic evolution. The relatively wide variability and high
Sr and Eu, and low Ba, La and Ce core concentrations
compared with the rims suggests that the cores formed
from a more mafic magma (Fig. 5). Further evidence for
this is provided by the calculated Sr and Ba concentrations
in the melt. These intersect with the defined composition
of andesitic groundmass glasses (Price et al., 2005) at
8308C (Fig. 8). However, the calculated range of Sr and
Ba melt compositions would encompass the andesitic
groundmass glass compositions if the core of the plagioclase crystals crystallized from the magma at an earlier
stage of petrogenetic evolution when the magma was
hotter [e.g. c. 940^9508C, the estimated magmatic temperature calculated from plagioclase^melt equilibria
(Putirka, 2008) when the andesitic groundmass glass
composition is taken as the melt composition] prior to the
NUMBER 12
DECEMBER 2010
cooling of magma. The region defined by this melt
composition is hereafter denoted as Melt A (Fig. 8).
Unfortunately, the limited data for the measured Sr and
Ba concentrations of andesitic melts from the TVZ restrict
our evaluation of whether the wide variability of Sr and
Ba melt concentrations observed in the cores characterizes
that of the natural melt composition or is instead the
result of diffusional modification of the Sr and Ba concentrations of the crystal cores.
It should be noted that the composition of Group 1 cores
is very similar to the composition of plagioclase crystals
derived from metasedimentary xenoliths from Ruapehu
(Price et al., 2005). This could suggest that these plagioclase
cores are derived from the disaggregation of crustal xenoliths. However, there is no evidence in the Whakamaru
data for the involvement of a melt composition similar to
that associated with the crustal xenoliths, although this
chemical signature may have been masked by subsequent
magmatic processes. Without Sr^Pb isotopic information,
it is not possible to determine if Group 1 plagioclase crystal
cores are derived directly from a less evolved silicic
magma or from the disaggregation of crustal xenoliths.
In contrast, Group 1 rims have restricted Sr
(465^630 ppm), Ba (450^575 ppm), La, Ce and Eu concentrations, and the calculated Sr and Ba melt compositions
span the known compositional range of Whakamaru
group ignimbrite groundmass glasses (Brown, 1994). This,
therefore, provides compelling evidence for the crystallization of Group 1 rims from the final host melt (hereafter
termed Melt D) prior to eruption. Furthermore, a third
minor melt composition (hereafter denoted Melt B) is present in some Group 1 crystals (e.g. WH1_h, WH6_da) as a
thin overgrowth surrounding the crystal core. It is most
readily distinguished on the basis of high calculated Ba
melt concentrations.
The core and the rim melts of Group 1 crystals can be
petrologically linked through a fractional crystallization
model using the observed crystal assemblage; the trends
are similar to those described by Berlo et al. (2007) for
Mount St. Helens. The crystal assemblage is dominated
by plagioclase but contains smaller proportions of
K-feldspar, biotite, amphibole, and Fe^Ti oxides that may
influence the behaviour of Sr and Ba in the melt. The fractionation of quartz and orthopyroxene, with KSr and KBa
0, will cause an increase in the Sr and Ba abundance in
the melt. Beginning with a parental melt composition of
176 ppm and 728 ppm of Sr and Ba respectively, calculated
from the core region with the highest An content, and
therefore slowest Sr diffusion (Giletti & Casserly, 1994;
Giletti & Shanahan, 1997; Zellmer et al., 1999), and a magmatic temperature of 8308C, the fractionation trend of a
crystal assemblage consisting of purely plagioclase with
XAn of 0·3 and 0·4 has been modelled (Fig. 10). Fractional
crystallization modelling of the observed crystal
2480
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 8. (a) Calculated coexisting Sr and Ba melt compositions for the three defined plagioclase populations calculated at 8308C, using the partitioning relationships [equations (1a) and (1b)] of Blundy & Wood (1991). The melt composition of Group 1 cores at 9508C is denoted by a
fine grey dotted line. (b) Schematic illustration of the Sr and Ba concentrations of the four melt compositions discussed in the text. The
shaded regions represent melts of known composition: dark grey, Whakamaru groundmass glasses (Brown, 1994); medium grey, andesitic glasses
(Price et al., 2005); black, quartz-hosted melt inclusions (Saunders et al., 2010). Inset indicates the location of melt compositions in relation to
the three plagioclase populations.
assemblage of plagioclase, quartz, orthopyroxene,
K-feldspar, biotite, amphibole and Fe^Ti oxides (Table 9)
to attain a Melt D composition results in a trend parallel
to that calculated using the pure plagioclase assemblage,
but requires greater degrees of fractional crystallization
(50^60%) compared with 10^20% as indicated by partial
melting (Fig. 10). All of the calculated fractionation trends
project through the calculated rim melt composition,
providing strong evidence for the formation of Melt D
by fractional crystallization of the observed crystal
2481
JOURNAL OF PETROLOGY
VOLUME 51
NUMBER 12
DECEMBER 2010
Fig. 9. Comparison of pumice compositions as defined by Brown et al. (1998) with the calculated coexisting Sr and Ba melt compositions
for the three defined plagioclase populations calculated at 8308C, with the partitioning relationships [equations (1a) and (1b)] of Blundy &
Wood (1991).
assemblage from a progenitor andesitic melt (Melt A)
(Fig. 10). Furthermore, the modelled decrease in Ba concentrations at low Sr concentrations (550 ppm) of the
observed crystal assemblage hints at a possible origin of
Melt C from extreme fractional crystallization of progenitor andesitic melts involving either (1) the observed crystal
assemblage as modelled or (2) crystallization of a dominantly plagioclase, quartz and orthopyroxene Fe^Ti
oxides assemblage to generate Melt D, but late-stage crystallization of K-feldspar, amphibole and biotite significantly reducing the Sr and Ba concentrations of the residual
melt, to produce Melt C. This model could therefore indicate that over time it could have been possible for Melt D
to evolve into a mature crystal mush zone, if it was
retained within the crust and not erupted. Potentially, a
proto-mush zone could have been produced during the generation of the final melt body, but as the crystal content at
the time of eruption would have been insufficient to form a
rigid framework in the magma body, it was instead erupted.
The source of Melt B is somewhat ambiguous, but lies on
a vector parallel to a possible greywacke partial melt, calculated from the greywacke partial melt model of Reid
(1982) or at least a melt generated from a progenitor
greywacke partial melt (Fig. 10). An alternative origin for
this melt is through the fractional crystallization from a
progenitor melt A composition of a crystal assemblage
dominated by plagioclase at high (9508C) magmatic temperatures (Fig. 10).
Irrespective of the source of Melt B, the presence of this
melt as a thin overgrowth mantling the core in several
crystals suggests that a magma mixing event occurred,
perhaps not necessarily chamber-wide, prior to the substantial crystallization of Group 1 rims.
Origins of Group 2 plagioclase crystals
The limited compositional variability of Group 2 crystals
implies crystallization from a relatively homogeneous melt.
The small oscillations in An can be accounted for by thermal
perturbations and/or fluctuations in the water content of
the magma (Alle'gre et al.,1981; Ginibre et al., 2002a; Putirka,
2008). Furthermore, the consistency between the composition of the majority of Group 2 crystals and Group 1
rims (Figs 4 and 5, Electronic Appendices 3 and 4) signifies
crystallization from a melt body with the same or a similar
composition. The coherency between calculated Sr and Ba
melt compositions and groundmass glass (Brown, 1994)
compositions is interpreted as an indication that the crystallization of Group 2 crystals and Group 1 rims was from
the final melt composition. Allied with the evidence
from the Sr diffusion modelling of Group 1 crystals, the
2482
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 10. Fractional crystallization model for the generation of Group 1 rim Sr and Ba melt compositions (Melt D) from the calculated Sr and Ba
melt composition of Group 1 cores (Melt A) using WH1_h as an example. A crystal assemblage consisting of pure An30 or An40 plagioclase is
modelled and the observed phenocryst assemblage (Table 7). Partition coefficients for Sr and Ba are calculated from the relationships of
Blundy & Wood (1991) and a magmatic temperature of 8308C for the An30 and An40 assemblage. Partition coefficients for the observed assemblage are given in Table 8. Crosses mark every 10% of fractional crystallization from the parental composition of 176 ppm Sr and 728 ppm Ba.
A plagioclase-dominated crystal assemblage of An30^40 could generate the Group 1 rim compositions after only 10^20% crystallization, in comparison with the observed crystal assemblage, which requires 40^55% crystallization. Also shown is a possible fractional crystallization trend
illustrating the generation of Melt B compositions from a parental Group 1 core melt composition of 204 ppm and 1094 ppm of Sr and Ba
respectively at a high magmatic temperatures of 9508C. The crystal assemblage is dominated by plagioclase with a composition of An55 (as
observed in Group 1 cores) and partition coefficients are calculated from the relationships of Blundy & Wood (1991). A greywacke partial
melt model (dark grey dashed line) from Reid (1982) is also shown and suggests that Melt B could be produced by a 40^65% greywacke
partial melt.
Table 9: Crystal assemblage and partition coefficients of
WH1 used in modelling
plag opx
Crystal proportion 24
qtz
ilm mag k-feld amp bt
17·3 20·6 1·7 1·7
7
Bulk Ki
20·7 7
Sr
7·41
0·1 0·01 1·2
0·4
0·3 2·23
Ba
0·84
0·1
0·3
7
4·9
fractionation of the majority of Group 2 crystals and Group1
rims is inferred to have occurred in the 300 years preceding
the final eruption. The only exception to this is the four analyses at the core of WH3_e, which result in calculated Sr and
Ba melt compositions that correspond to a Melt B composition (Fig. 8), implying that they may have fractionated from
the melt prior to this.
Origins of Group 3 plagioclase crystals
1·09
Crystal assemblage recalculated to 100% for use with the
fractional crystallization model of Group 1 plagioclase crystals. KSr and KBa of plagioclase crystals were calculated
from the measured plagioclase rim compositions and
groundmass glass concentrations (Brown, 1994). Sr and
Ba are assumed perfectly incompatible where no partition
coefficients are shown. KSr and KBa of amphibole and biotite are from Nash & Crecraft (1985), amphibole and magnetite are from Bacon & Druitt (1988) and ilmenite from
Ewart & Griffin (1994). plag, plagioclase; opx, orthopyroxene; qtz, quartz; ilm, ilmenite; mag, magnetite; k-feld,
K-feldspar; amp, amphibole; bt, biotite.
Group 3 crystals are distinctive with lower Sr
(280^480 ppm), Mg (19^32 ppm) and An concentrations
and similar REE and Pb abundances relative to those
observed in Group 1 and Group 2 crystals (Fig. 5). The
core analysis of crystal WH6_b exhibits elevated La, Eu
and Pb concentrations compared with the rim, consistent
with equilibration with a melt derived from either a
highly evolved igneous or greywacke protolith. Reported
An compositions of greywacke plagioclase crystals indicate
two plagioclase populations of An40^50 and An0^10 in the
Waipapa terrane and a single plagioclase population of
An0^10 in the Torlesse terrane (Reid, 1982), precluding a
2483
JOURNAL OF PETROLOGY
VOLUME 51
greywacke protolith origin. Moreover, an igneous origin is
probable as the calculated Sr and Ba melt compositions of
Group 3 crystals overlap with the Sr and Ba concentrations
of quartz-hosted melt inclusions located in the core of crystals that potentially originated from a mature crystal
mush body (Saunders et al., 2010) and the calculated melt
trend overlaps and trends towards granitic compositions
(e.g. Walker et al., 2007; Wiebe et al., 2007). Thus, it is interpreted that Group 3 plagioclase crystals are antecrysts
derived from a mature and highly evolved crystal mush or
plutonic body, potentially formed through the fractional
crystallization of progenitor andesitic melts (see above discussion on fractional crystallization of Group 1 core and
rim melts) at an early stage of magmatic evolution of the
Whakamaru magmatic system. They are therefore unlikely
to be cogenetic with the final melt composition. Insights
into the timing of crystallization of Group 3 plagioclase
crystals can be obtained from the observation that the calculated Sr and Ba melt compositions are similar to the
compositions of quartz-hosted melt inclusions located in
the core of crystals. Diffusion modelling of the core^rim
interfaces of quartz crystals indicates that the final rim
growth occurred 5300 years prior to eruption. Therefore,
the core of the quartz crystals and thus Group 3 plagioclase crystals crystallized prior to this (Fig. 11). Zircon
ages determined through U^Pb geochronology extend to
250 kyr prior to eruption (Brown & Fletcher, 1999).
Charlier et al. (2005) reinterpreted these data to show that
the older zircon populations are not cogenetic with the
final melt and are inherited into the Whakamaru magma.
This provides evidence for the presence of a crystal mush
body from an early stage within the evolution of the
Whakamaru magmatic system and imposes a maximum
time constraint for the crystallization of Group 3 plagioclase crystals and the cores of quartz crystals. Because of
the low abundance of K-feldspar and biotite crystals
observed in the Whakamaru magma, it is speculated that
these crystals may also be potentially inherited from this
crystal mush body.
Origin of quartz crystals
The origin of quartz crystals is more difficult to decipher,
as they have a more restricted chemistry. Texturally, there
is evidence (bright-CL rims and oscillatory zoned rims)
for the existence of at least two populations or two discrete
evolutionary pathways for quartz crystals. However, little
can be inferred about the early history of these crystals,
except as discussed in the previous section. The observed
bright-CL rims (e.g. WH1_37, WH2_3) mantling dissolution
surfaces indicates mixing into a hotter, possibly more
Ti-rich, magma prior to the renewed growth. One possible
source for the cores of these crystals is a mature crystal
mush body. In contrast, the second group of quartz crystals
with dominantly oscillatory-zoned rims may indicate
growth from a magma that experienced multiple, small
NUMBER 12
DECEMBER 2010
recharge events leading to renewed quartz growth, possibly related to the oscillatory-zoned plagioclase of Group
2. The close correspondence in calculated timescales for
rim growth of quartz and plagioclase provides further evidence for a common stage of late growth.
P E T RO G E N E S I S O F T H E
W H A K A M A RU M A G M A
There is evidence from zircon geochronology that the formation of the Whakamaru magma system commenced at
least 250 kyr prior to the catastrophic eruption (Brown &
Fletcher, 1999; Charlier et al., 2005). However, diffusion
modelling conducted in this study indicates that the final
body of eruptible magma was assembled in the 300 years
preceding the catastrophic eruption. The assembly of this
final Whakamaru magma body was not simple but entailed the mixing and mingling of magmas from multiple
temporally and spatially separated sources within the magmatic system. The earliest history of the magmatic system
involved fractional crystallization of and assimilation of
greywacke country rock by mantle-derived magmas that
eventually formed mature, highly evolved crystal mush
bodies that would ultimately supply Melt C compositions
to the Whakamaru magma (Fig. 11a). However, it is
suggested that with continued magma supply further
generation of crystal-rich magmas occurred, generating
proto-mush bodies (Melt A) that may be represented at
the surface by high-Si, crystal-rich andesites (e.g. Price
et al., 2005). It is such magma bodies that provide a ‘crystal
nursery’, generating the cores observed in Group 1 plagioclase crystals (Fig. 11b and c).
At the magmatic conditions observed in the evolution of
the Whakamaru magma (relatively low temperature, high
Si), pressure exerts little control over the An composition
of plagioclase crystals (Putirka, 2008) and is insufficient to
account for the change in An composition between Group
1 plagioclase cores and rims. However, decompression
during ascent could explain the resorbed and/or rounded
texture of Group 1 cores (e.g. Nelson & Montana, 1992;
Berlo et al., 2007). Plagioclase stability is controlled by the
water saturation of the melt (Johannes, 1989), and it has
been shown that the degree of plagioclase crystallization
increases at low partial pressures of water at shallow crustal levels (e.g. Blundy et al., 2006; Berlo et al., 2007). We interpret the overwhelming dominance of Group 2 crystals
combined with the coherent Group 1 rim compositions as
evidence for a dramatic increase in the amount of plagioclase crystallization (Berlo et al., 2007). This increase in
plagioclase fractionation is driven by the intersection at
shallow levels in the P^T space of the magma ascent path
with the plagioclase liquidus (Berlo et al., 2007), probably
as a result of degassing of water. This may have occurred
as the melt ascended into the magma chamber from the
underlying mush region (Fig. 11e). Prior to this, the
2484
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
Fig. 11. Schematic illustrations summarizing the timing and growth of the studied Whakamaru plagioclase and quartz crystals and the
formation of the Whakamaru magmatic system over time. Initiation of the magmatic system is assumed to occur 250 kyr prior to the eruption
based on zircon chronology (Brown & Fletcher, 1999). Letters (a^f) on the timeline at the top correspond to the figures below; full explanation
is given in the text. Depths of magma reservoir are estimated from Harrison & White (2004).
2485
JOURNAL OF PETROLOGY
VOLUME 51
crystallization of plagioclase was suppressed owing to the
high H2O concentrations in the andesitic proto-mush
body, which were generated during the extreme fractional
crystallization required to produce this magma body.
Both prior to and concurrently, the crystal mush body provided heat to the surrounding greywacke country rock,
assimilating wall-rocks and generating the third melt composition (Melt B) (Fig. 11d).
It was the amalgamation of these melts in varying proportions, combined with fractional crystallization, that
generated the final Whakamaru melt compositions. The
zonation of plagioclase and quartz crystals, and the diversity of residence times of the quartz crystal rims imply
that mixing of these magmas occurred continually until
the eruption. Intrusion of mafic magmas into silicic
magma bodies is commonly invoked to trigger volcanic
eruptions (e.g. Pallister et al., 1992; Synder, 2000; Ginibre
et al., 2007; Martin et al., 2008). We speculate that an intrusion of hot mafic magma into the base of the magma
chamber occurred prior to eruption; evidence for this is
the rounded nature of plagioclase crystals, indicating that
they were at the time actively resorbing as the eruption
commenced. Additionally, Brown et al. (1998) observed
clasts of high-alumina basalt, both as mixed pumices and
as basaltic scoria, indicating the presence of a more mafic
magmatic component in the Whakamaru magma system
during eruption.
CONC LUSIONS
Chemical and textural zonation of plagioclase and quartz
crystals provides evidence for four distinct rhyolitic melts
of different origins feeding the generation of the final
Whakamaru supereruption. Magmatic evolution occurred
over a period exceeding 250 kyr (Brown & Fletcher, 1999).
However, Sr diffusion modelling of core^rim interfaces of
Group 1 plagioclase crystals and diffusion modelling of the
bright-CL rims of quartz crystals indicates that the final
pre-eruption assembly of melt-dominant magma chambers
bodies may be a transient, ephemeral stage; this may have
occurred shortly prior to eruption in several discrete steps
on timescales of 5300 years. This contrasts with estimates
of timescales derived from radiometric methods (e.g.
zircon populations), which suggest that magma accumulation can take place over periods of up to 100 or 200 kyr.
This difference in timescales is fundamental, as it can provide important insights into the magmatic processes that
occur in the build-up to eruptions and could potentially
be important for future hazard prediction of large-volume
silicic eruptions. Thus, this study adds to a growing body
of evidence that silicic magma bodies are the result of
long periods of build-up, followed by geologically rapid
re-mobilization and homogenization processes.
NUMBER 12
DECEMBER 2010
AC K N O W L E D G E M E N T S
Stewart Bush is thanked for help with sample preparation,
and Tod Waight, Ian Smith, Julie Vry and Zoe Laing for
discussions. Richard Price, Phillip Leat and John Wolff are
thanked for their helpful reviews.
FU NDI NG
This work was supported by a Commonwealth Scholarship
and a Victoria University of Wellington Small Research
Grant to K.E.S. D.J.M. would like to thank the University
of Leeds for financial support through the provision of
start-up funds during this work.
S U P P L E M E N TA RY DATA
Supplementary data for this paper are available at Journal
of Petrology online.
R E F E R E NC E S
Alle'gre, C. J., Provost, A. & Jaupart, C. (1981). Oscillatory zoning: a
pathological case of crystal growth. Nature 294, 223^228.
Annen, C. & Sparks, R. S. J. (2002). Effects of repetitive emplacement
of basaltic intrusions on thermal evolution and melt generation in
the deep crust. Earth and Planetary Science Letters 203, 937^955.
Bachmann, O. & Bergantz, G.W. (2003). Rejuvenation of the Fish
Canyon magma body: A window into the evolution of large-volume
silicic magma systems. Geology 31, 789^792.
Bachmann, O. & Bergantz, G.W. (2004). On the origin of crystal-poor
rhyolites: extracted from batholithic crystal mushes. Journal of
Petrology 45, 1565^1582.
Bachmann, O. & Bergantz, G. (2008). The magma reservoirs that feed
supereruptions. Elements 4, 17^21.
Bachmann, O., Miller, C. F. & de Silva, S.L. (2007). The volcanic^
plutonic connection as a stage for understanding crustal magmatism. Journal of Volcanology and Geothermal Research 167, 1^23.
Bacon, C. R. & Druitt, T. H. (1988). Compositional evolution of the
zoned calc-alkaline magma chamber of Mount Mazama, Crater
Lake, Oregon. Contributions to Mineralogy and Petrology 98, 224^256.
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.
Baker, J. A., Snee, L. & Menzies, M. A. (1996). A brief Oligocene
period of flood volcanism: implications for the duration and rate
of continental flood volcanic at the Afro-Arabian triple junction.
Earth and Planetary Science Letters 138, 39^55.
Baker, J. A., MacPherson, C. G., Menzies, M. A., Thirlwall, M. F.,
Al-Kadasi, M. & Mattey, D. P. (2000). Resolving crustal and
mantle contributions to continental flood volcanism, Yeman; constraints from mineral oxygen isotope data. Journal of Petrology 41,
1805^1820.
Bence, A. E. & Albee, A. L. (1968). Empirical correction factors for
the electron microprobe analysis of silicates and oxides. Journal of
Geology 76, 382^403.
Bergantz, G. W. (1989). Underplating and partial melting:
Implications for melt generation and extraction. Science 245,
1093^1095.
Berlo, K., Blundy, J., Turner, S. & Hawkesworth, C. (2007). Textural
and chemical variation in plagioclase phenocrysts from the 1980
2486
SAUNDERS et al.
PLAGIOCLASE ZONATION, WHAKAMARU IGNIMBRITE
eruptions of Mount St. Helens, USA. Contributions to Mineralogy and
Petrology 154, 291^308.
Bindeman, I. N., Davis, A. M. & Drake, M. J. (1998). Ion microprobe
study of plagioclase^basalt partition experiments at natural concentration levels of trace elements. Geochimica et Cosmochimica Acta
62, 1175^1193.
Blundy, J. D. & Wood, B. J. (1991). Crystal-chemical controls on the
partitioning of Sr and Ba between plagioclase feldspar, silicate
melts, and hydrothermal solutions. Geochimica et Cosmochimica Acta
55, 193^209.
Blundy, J. D. & Wood, B. J. (1994). Prediction of crystal^melt partition
coefficients from elastic moduli. Nature 372, 452^454.
Blundy, J., Cashman, K. & Humphreys, M. (2006). Magma heating
by decompression-driven crystallization beneath andesite volcanoes. Nature 443, 76^80.
Bowen, N. L. (1913). The melting phenomena of plagioclase feldspars.
AmericanJournal of Science 35, 577^599.
Briggs, N. D. (1976). Recognition and correlation of subdivisions
within the Whakamaru ignimbrite, central North Island, New
Zealand. New Zealand Journal of Geology and Geophysics 19, 463^501.
Brown, S. J. A. (1994). Geology and geochemistry of the Whakamaru
group ignimbrites, and associated rhyolite domes, Taupo Volcanic
Zone, New Zealand, PhD thesis, University of Canterbury,
Christchurch.
Brown, S. J. A. & Fletcher, I. R. (1999). SHRIMP U^Pb dating of
the preeruption growth history of zircons from the 340 ka
Whakamaru Ignimbrite, New Zealand: Evidence for 4250 k.y.
magma residence times. Geology 27, 1035^1038.
Brown, S. J. A., Wilson, C. J. N., Cole, J. W. & Wooden, J. (1998).
The Whakamaru group ignimbrites, Taupo Volcanic Zone, New
Zealand: evidence for reverse tapping of a zoned silicic magmatic
system. Journal of Volcanology and Geothermal Research 84, 1^37.
Charlier, B. L. A., Wilson, C. J. N., Lowenstern, J. B., Blake, S., Van
Calsteren, P. W. & Davidson, J. P. (2005). Magma generation at a
large, hyperactive silicic volcano (Taupo, New Zealand) revealed
by U^Th and U^Pb systematics in zircons. Journal of Petrology 46,
3^32.
Charlier, B. L. A., Bachmann, O., Davidson, J. P., Dungan, M. A. &
Morgan, D. J. (2007). The upper crustal evolution of a large silicic
magma body: Evidence from crystal-scale Rb^Sr isotopic heterogeneities in the Fish Canyon magmatic system, Colorado. Journal of
Petrology 48, 1875^1894.
Charlier, B. L. A., Wilson, C. J. N. & Davidson, J. P. (2008). Rapid
open-system assembly of a large silicic magma body: time-resolved
from cored plagioclase crystals in the Oruanui eruption deposits,
New Zealand. Contributions to Mineralogy and Petrology 156, 799^813.
Cherniak, D. J., Watson, E. B. & Wark, D. A. (2007). Ti diffusion in
quartz. Chemical Geology 236, 65^74.
Cole, J. W. (1981). Genesis of lavas of the Taupo Volcanic Zone, North
Island, New Zealand. Journal of Volcanology and Geothermal Research
10, 317^337.
Coleman, D. S., Gray, W. & Glazner, A. F. (2004). Rethinking the
emplacement and evolution of zoned plutons: Geochronologic
evidence for incremental assembly of the Tuolumne Intrusive
Suite, California. Geology 32, 433^436.
Costa, F., Chakraborty, S. & Dohmen, R. (2003). Diffusion coupling
between trace and major elements and a model for calculation
of magma residence times using plagioclase. Geochimica et
Cosmochimica Acta 67, 2189^2200.
Crank, J. (1976). The Mathematics of Diffusion. Oxford: Oxford
University Press.
Darby, D. J., Hodgkinson, K. H. & Blick, G. H. (2000). Geodetic
measurements of deformation in the Taupo Volcanic Zone, New
Zealand: the North Taupo Network revisited. New Zealand Journal
of Geology and Geophysics 43, 157^170.
Davidson, J. P., Morgan, D. J., Charlier, B. L. A., Harlou, R. &
Hora, J. A. (2007). Microsampling and isotopic analysis of igneous
rocks; Implications for the study of magmatic systems. Annual
Review of Earth and Planetary Sciences 35, 273^311.
Ewart, A. & Griffin, W. L. (1994). Application of proton-microprobe
data to trace element partitioning in volcanic rocks. Chemical
Geology 117, 251^284.
Giletti, B. J. & Casserly, J. E. D. (1994). Strontium diffusion kinetics in
plagioclase feldspars. Geochimica et Cosmochimica Acta 58, 3785^3793.
Giletti, B. J. & Shanahan, T. M. (1997). Alkali diffusion in plagioclase
feldspar. Chemical Geology 139, 3^20.
Ginibre, C., Kronz, A. & Wo«rner, G. (2002a). High-resolution quantitative imaging of plagioclase composition using accumulated
backscattered electron images: new constraints on oscillatory
zoning. Contributions to Mineralogy and Petrology 142, 436^448.
Ginibre, C., Wo«rner, G. & Kronz, A. (2002b). Minor- and traceelement zoning in plagioclase: implications for magma chamber
processes at Parinacota volcano, northern Chile. Contributions to
Mineralogy and Petrology 143, 300^315.
Ginibre, C., Wo«rner, G. & Kronz, A. (2004). Structure and dynamics
of the Laacher See magma chamber (Eifel, Germany) from major
and trace element zoning in sanidine: a cathodoluminescence and
electron microprobe study. Journal of Petrology 45, 2197^2223.
Ginibre, C., Wo«rner, G. & Kronz, A. (2007). Crystal zoning as an
archive for magma evolution. Elements 3, 261^266.
Glazner, A. F., Bartley, J. M., Coleman, D. S., Gray, W. & Taylor, R.
Z. (2004). Are plutons assembled over millions of years by amalgamation from small magma chambers? GSAToday 14, 4^11.
Grove, T. L., Baker, M. B. & Kinzler, R. J. (1984). Coupled
CaAl^NaSi diffusion in plagioclase feldspar: Experiments and applications to cooling rate speedometry. Geochimica et Cosmochimica
Acta 48, 2113^2121.
Harrison, A. J. & White, R. S. (2004). Crustal structure of the Taupo
Volcanic Zone, New Zealand: Stretching and igneous intrusion.
Geophysical Research Letters 31, L13615.
Hawkesworth, C. J., Blake, S., Evans, P., Hughes, R., MacDonald, R.,
Thomas, L. E., Turner, S. P. & Zellmer, G. (2000). Time scales of
crystal fractionation in magma chambersçintegrating physical,
isotopic and geochemical perspectives. Journal of Petrology 41,
991^1006.
Hildreth, W. & Wilson, C. J. N. (2007). Compositional zoning of the
Bishop Tuff. Journal of Petrology 48, 951^999.
Houghton, B. F., Wilson, C. J. N., McWilliams, M. O., 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, New Zealand. Geology 23, 13^16.
Huppert, H. E. & Sparks, S. J. (1988). The generation of granitic
magmas by intrustion of basalt into continental crust. Journal of
Petrology 29, 599^624.
Jarosewich, J. A., Nelen, J. A. & Norberg, J. A. (1980). Reference samples for electron microprobe analysis. Geostandards Newsletter 4,
43^47.
Jerram, D. A. & Davidson, J. P. (2007). Frontiers in textural and
microgeochemical analysis. Elements 3, 235^238.
Johannes, W. (1989). Melting of plagioclase^quartz assemblages at
2 kbar water pressure. Contributions to Mineralogy and Petrology 103,
270^276.
Kent, A. J. R., Blundy, J., Cashman, K. V., Cooper, K. M.,
Donnelly, C., Pallister, J. S., Reagan, M., Rowe, M. C. &
Thornber, C. R. (2007). Vapor transfer prior to the October 2004
eruption of Mount St, Helens, Washington. Geology 35, 231^234.
2487
JOURNAL OF PETROLOGY
VOLUME 51
Liu, Y., Anderson, A. T., Wilson, C. J. N., Davis, A. M. & Steele, I. M.
(2006). 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
151, 71^87.
Martin, V. M., Morgan, D. J., Jerram, D. A., Caddick, M. J., Prior, D.
J. & Davidson, J. P. (2008). Bang! Month-scale eruption triggering
at Santorini volcano. Science 321, 1178.
Morgan, D. J., Blake, S., Rogers, N. W., DeVivo, B., Rolandi, G.,
Macdonald, R. & Hawkesworth, C. J. (2004). Time scales of crystal
residence and magma chamber volume of diffusion profiles in
phenocrysts: Vesuvius 1944. Earth and Planetary Science Letters 222,
933^946.
Morgan, D. J., Blake, S., Rogers, N. W., DeVivo, B., Rolandi, G. &
Davidson, J. P. (2006). Magma chamber recharge at Vesuvius in
the century prior to eruption of A.D. 79. Geology 34, 845^848.
Morse, S. A. (1984). Cation diffusion in plagioclase feldspar. Science
225, 504^505.
Nash, W. P. & Crecraft, H. R. (1985). Partition coefficients for trace
elements in silicic magmas. Geochimica et Cosmochimica Acta 49,
2309^2322.
Nelson, S. T. & Montana, A. (1992). Sieve-textured plagioclase in
volcanic rocks produced by rapid decompression. American
Mineralogist 77, 1242^1249.
Pallister, J. S., Hoblitt, R. P. & Reyes, A. G. (1992). A basalt trigger for
the 1991 eruption of Pinatubo volcano? Nature 356, 426^428.
Price, R. C., Gamble, J. A., Smith, I. E. M., Stewart, R. B., Eggins, S.
& Wright, I. C. (2005). 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 140, 1^24.
Putirka, K. D. (2008). Thermometers and barometers for volcanic
systems. In: Putirka, K. D. & Tepley, F. J., III (eds) Minerals,
Inclusions and Volcanic Processes. Mineralogical Society of America
and Geochemical Society, Reviews in Mineralogy and Geochemistry 69,
61^120.
Reed, S. J. B. (1996). Electron Microprobe Analysis and Scanning Electron
Microscopy in Geology. Cambridge: Cambridge University Press, 201 p.
Reid, F. E. (1982). Geochemistry of Central North Island greywackes
and genesis of silicic magmas, PhD thesis, Victoria University of
Wellington, Wellington.
Sauerzapf, U., Lattard, D., Burchard, M. & Englemann, R. (2008).
The titanomagnetite^ilmenite equilibrium: new experimental data
and thermo-oxybarometric application to the crystallization of
basic to intermediate rocks. Journal of Petrology 49, 1161^1185.
Saunders, K. E. (2009). Micro-analytical studies of the petrogenesis of
silicic arc magmas in the Taupo Volcanic Zone and southern
Kermadec Arc, New Zealand, PhD thesis, Victoria University of
Wellington, Wellington.
Saunders, K. E., Baker, J. A. & Wysoczanski, R. J. (2010).
Microanalysis of large volume silicic magma in continental and
oceanic arcs: Melt inclusions in Taupo Volcanic Zone and
Kermadec Arc rocks, south West Pacific. Journal of Volcanology and
Geothermal Research 190, 203^218.
Shane, P., Smith, V. C. & Nairn, I. (2008). Millennial timescale resolution of rhyolite magma recharge at Tarawera volcano: insights
NUMBER 12
DECEMBER 2010
from quartz chemistry and melt inclusions. Contributions to
Mineralogy and Petrology 156, 397^411.
Sparks, R. S. J., Self, S., Grattan, J., Oppenheimer, C., Pyle, D. &
Rymer, H. (2005). Super-eruptions: global effects and future threats.
Report of a Geological Society of London Working Group, 25 p.
Stern, T. A. (1987). Asymmetric back-arc spreading, heat flux and
structure beneath the Central Volcanic Region of New Zealand.
Earth and Planetary Science Letters 85, 265^267.
Synder, D. (2000). Thermal effects of the intrusion of basaltic magma
into a more silicic magma chamber and implications for eruption
triggering. Earth and Planetary Science Letters 175, 257^273.
Ukstins Peate, I., Baker, J. A., Al-Kadasi, M., Al-Subbary, A.,
Knight, K. B., Riisager, P., Thirlwall, M. F., Peate, D. W., Renne, P.
R. & Menzies, M. A. (2005). Volcanic stratigraphy of large-volume
silicic pyroclastic eruptions during Oligocene Afro-Arabian flood
volcanism inYemen. Bulletin ofVolcanology 68,135^156.
Ukstins Peate, I., Kent, A. J. R., Baker, J. A. & Menzies, M. A. (2007).
Extreme geochemical heterogeneity in Afro-Arabian Oligocene
tephras: Preserving fractional crystallisation and mafic recharge
processes in silicic magma chambers. Lithos 102, 260^278.
Venezky, D. Y. & Rutherford, M. J. (1999). Petrology and Fe^Ti oxide
reequilibration of the 1991 Mount Unzen mixed magma. Journal of
Volcanology and Geothermal Research 89, 213^230.
Walker, B. A., Jr, Miller, C. F., Lowery Claiborne, L., Wooden, J. L. &
Miller, J. S. (2007). 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 167, 239^262.
Wark, D. A. & Spear, F. S. (2005). Ti in quartz: Cathodoluminescence
and thermometry. Geochimica et Cosmochimica Acta 69, A592.
Wark, D. A., Hildreth, W., Spear, F. S., Cherniak, D. J. & Watson, E.
B. (2007). Pre-eruption recharge of the Bishop magma system.
Geology 35, 235^238.
Wiebe, R. A., Wark, D. A. & Hawkins, D. P. (2007). Insights from
quartz cathodoluminescence zoning into crystallisation of the
Vinalhaven granite, coastal Maine. Contributions to Mineralogy and
Petrology 154, 439^453.
Wilson, C. J. N., Houghton, B. F., McWilliams, M. O., Lanphere, M.
A., Weaver, S. D. & Briggs, R. M. (1995). Volcanic and structural
evolution of Taupo Volcanic Zone, New Zealand: a review. Journal
of Volcanology and Geothermal Research 68, 1^18.
Wilson, C. J. N., Gravley, D. M., Leonard, G. S. & Rowland, J. V.
(2009). d. Volcanism in the central Taupo Volcanic Zone, New
Zealand: tempo, styles and controls. In: Thordarson, T.,
Larsen, G., Self, S., Rowland, S. & Hoskuldsson, A. (eds) Studies
in Volcanology: The Legacy of George Walker. Special Publications of
IAVCEI 2, 225^247.
Zellmer, G. F. & Clavero, J. E. (2006). Using trace element correlation
patterns to decipher a sanidine crystal growth chronology: An
example from Taapaca volcano, Central Andes. Journal of
Volcanology and Geothermal Research 156, 291^301.
Zellmer, G. F., Blake, S., Vance, D., Hawkesworth, C. & Turner, S.
(1999). Plagioclase residence times at two island arc volcanoes
(Kameni Islands, Santorini, and Soufrie're, St. Vincent) determined by Sr diffusion systematics. Contributions to Mineralogy and
Petrology 136, 345^357.
2488