100% found this document useful (1 vote)

49 views26 pages

Schmitt, A. Et Al. (2023)

The document discusses advances in zircon geochronology that allow resolving zircon age dispersions beyond analytical uncertainties. It reviews how zircon age spectra can be used to quantify parameters controlling the longevity and thermal evolution of magma bodies, such as magma recharge flux and duration. Models are presented that aim to translate zircon age spectral properties into such magmatic processes. The ability to resolve zircon age spectra offers opportunities to better understand magmatism.

Uploaded by

williams mata

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (1 vote)

49 views26 pages

Schmitt, A. Et Al. (2023)

Uploaded by

williams mata

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 26

Research Paper THEMED ISSUE: Different Personalities of Granites and Rhyolites: Silicic Magmas from the Lower Crust

ites and Rhyolites: Silicic Magmas from the Lower Crust to the Surface

GEOSPHERE Zircon age spectra to quantify magma evolution

A.K. Schmitt1,2,3, J. Sliwinski4,5, L. Caricchi6, O. Bachmann4, N. Riel2,7, B.J.P. Kaus2,7, A. Cisneros de Léon8, J. Cornet4, B. Friedrichs9, O. Lovera10,
T. Sheldrake5, and G. Weber11
GEOSPHERE, v. 19, no. X 1
Institute of Earth Sciences, Heidelberg University, Im Neuenheimer Feld 234–236, 69120 Heidelberg, Germany
2
Terrestrial Magmatic System Research Center (Universities of Mainz, Heidelberg, and Frankfurt), Germany
https://doi.org/10.1130/GES02563.1 3
John de Laeter Centre, Curtin University, Kent Street, Bentley, Western Australia 6102, Australia
4
Institute of Geochemistry and Petrology, Department of Earth Sciences, ETH Zurich, Clausiusstrasse 25, 8092 Zurich, Switzerland
5
School of Earth and Environmental Sciences, University of St. Andrews; Queen’s Terrace, St. Andrews KY16 9TS, UK
14 figures; 1 table; 1 supplemental file 6
Department of Earth Sciences, Rue des Maraichers 13, 1205 Geneva, Switzerland
7
Institute of Geosciences, Johannes-Gutenberg-University Mainz, Johann-Joachim-Becher-Weg 21, 55099 Mainz, Germany
CORRESPONDENCE: 8
School of the Earth, Ocean and Environment, University of South Carolina, 701 Sumter Street, EWS 617, Columbia, South Carolina 29208, USA
[email protected] 9
Department of Environment & Biodiversity, Paris Lodron University Salzburg, Hellbrunner Straße 34, 5020 Salzburg, Austria
10
Department of Earth, Planetary, and Space Sciences, University of California, 595 Charles Young Drive, Los Angeles, California 90095, USA
CITATION: Schmitt, A.K., Sliwinski, J., Caricchi, L.,
11
Department of Earth Sciences, University of Oxford, Oxford OX1 3AN, UK
Bachmann, O., Riel, N., Kaus, B.J.P., Cisneros de Léon,
A., Cornet, J., Friedrichs, B., Lovera, O., Sheldrake, T.,
and Weber, G., 2023, Zircon age spectra to quantify
magma evolution: Geosphere, v. 19, no. X, p. 1–26, ABSTRACT by tens of degrees. Moreover, zircon saturation temperatures for geochronology, whereby the dis-
https://doi.org/10.1130/GES02563.1. thermometry is supported by the trace element solution of zircon typically outpaces the diffusive
The past decades have seen tremendous and isotopic inventory of zircon, which records the mobilization of parent and daughter isotopes within
Science Editor: Shanaka de Silva advances in analytical capabilities regarding the thermochemical and compositional evolution of crystalline zircon at reasonable temperatures and
Guest Associate Editor: Jonathan S. Miller
sensitivity, spatial selectivity, and instrumental melts at high fidelity. Here, we first review the prop- timescales for crustal magmatic systems (Cherniak
precision of U-Th-Pb zircon geochronology. Along erties of true zircon age spectra that are defined and Watson, 2001). Hence, U-Th-Pb zircon dates can
Received 28 June 2022
Revision received 2 April 2023 with improved zircon pretreatment to mitigate by a statistically robust overdispersion relative to be straightforwardly interpreted as recording its
Accepted 1 May 2023 the effects of Pb-loss, these advancements have analytical uncertainties. Secondly, we evaluate crystallization in zircon-saturated melts (e.g., Wat-
resulted in the emergence of U-Th-Pb dating as existing models and present new models that aim son and Harrison, 1983), although zircon can also
Published online 6 July 2023 the most widely used geochronometer. In paral- to quantitatively translate the properties of zircon form in metamorphic and hydrothermal environ-
lel, it became increasingly obvious that modern age spectra into parameters controlling the longev- ments (e.g., Kohn et al., 2015; Schaltegger, 2007).
analytical techniques can resolve zircon age dis- ity and thermal evolution of crustal magma bodies Three main zircon geochronology methods are
persal beyond instrumental uncertainties and that such as magma recharge flux and duration. These used, and each has its individual strengths and
this dispersion cannot be attributed to Pb-loss or developing approaches, which aspire to capture all limitations (Table 1). Isotopic analysis of U-Th-Pb
inheritance. Hence, there is a pressing need to processes that affect the formation and dispersal in zircon by isotope dilution–thermal ionization
refine statistical procedures for displaying and of zircon in dynamic crustal magma systems, have mass spectrometry (ID-TIMS) is now achieved at
interpreting dispersed age data from volcanic and the potential to foster an improved understanding <0.1% relative age uncertainty for single crystals or
plutonic rocks, where zircon ages were traditionally of magmatism with implications for volcanic haz- µg-sized fragments (e.g., Schaltegger et al., 2015)
assigned to the quasi-instantaneous events of erup- ard assessment, geothermal energy uses, and the because of the routine pretreatment of zircon by
tion and magma emplacement, respectively. The origins of ore deposits. chemical abrasion (CA), which successfully miti-
ability to resolve zircon age spectra, which often gates the effects of Pb-loss due to metamictization
range over timescales of 103–106 years, also offers (e.g., Kryza et al., 2012; Mattinson, 2005; von Quadt
new opportunities to monitor magmatic processes, ■ 1 INTRODUCTION: SCOPE OF THIS et al., 2014). These improvements were paralleled
because zircon crystallization directly relates to REVIEW by progress in high spatial resolution analysis at
the temperature and composition of its host melt. the scale of subgrain domains (with <1 ng material
This relation is, at least for typical subalkaline melt Zircon geochronology is the gold standard for consumed) using laser ablation–inductively cou-
compositions, well calibrated by multiple zircon defining the geologic timescale (e.g., Schaltegger pled plasma–mass spectrometry (LA-ICP-MS), and
saturation experiments, although absolute satu- et al., 2015; Schoene, 2014). The U-Th-Pb system in at even higher spatial resolution with secondary
ration temperatures derived from them can vary zircon, including the chronometer based on inter- ionization mass spectrometric (SIMS) techniques
This paper is published under the terms of the mediate decay chain isotopes 230Th and its indirect (Table 1). U-Pb zircon geochronology has a lower
CC‑BY-NC license. A.K. Schmitt https://orcid.org/0000-0 002-9029-4211 parent, 238U, ranks nominally at the top of closure limit for precise age determination due to a
© 2023 The Authors

GEOSPHERE | Volume 19 | Number X

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/doi/10.1130/GES02563.1/5898376/ges02563.pdf
Schmitt et al. | Zircon age spectra 1
by guest
Research Paper

TABLE 1. COMPARISON OF ROUTINE GEOCHRONOLOGICAL METHODS APPLIED TO ZIRCON IN ORDER OF DECREASING SAMPLE CONSUMPTION AND HENCE RISK FOR CONVOLUTION OF AGE DOMAINS
Isotope dilution–thermal ionization Laser ablation–inductively coupled plasma– Secondary ion mass spectrometry (SIMS)
mass spectrometry (ID-TIMS) mass spectrometry (LA-ICP-MS)
Relative analytical precision and achievable absolute U–Pb ≤0.1%; ~103–104 yr at 1–10 Ma; U–Th U–Pb and U–Th ~2%; ~103–104 yr for U–Th at U–Pb and U–Th ~1–2%; ~103–104 yr for U–Th at
age resolution (2σ) impractical due to multi-grain sample requirement <<300 ka <<300 ka
Spatial resolution Single crystals or fragments analyzed with mixing of Spot analysis with ~20–30 μm diameter and ~10–20 Spot analysis with ~20–30 μm diameter and ~1–5 μm
age domains μm depth depth; sub-μm resolution in depth profiling
Theoretically resolvable age difference* ~3×104 yr for radial growth of 30 µm ~9 × 103 yr for 10 µm depth ~9 × 102 yr for 1 µm depth
Sample preparation Grain separates Mounts with sectioned or unsectioned grains; in-situ Mounts with sectioned or unsectioned grains; in-situ
Targeting aids †
CL imaging possible, but results in loss of crystal Targeting after CL imaging Targeting after CL imaging (except for depth profiling)
available for bulk analysis
Time requirements Time-consuming digestion and chemical separation Fast volumetric excavation rate (~10 μm3/s at ~2 J/cm–2 Slow volumetric excavation rate (~0.1 μm3/s/nA
prior to mass spectrometric analysis fluence); ~2–3 min per analysis primary beam); ~10–20 min per analysis
Notes: Table modified from Schaltegger et al. (2015).
*Based on a radial zircon growth rate of 3.6 × 10–15 cm/s (Storm et al., 2012); faster growth rates and therefore briefer resolvable age differences are possible, e.g., in high-temperature or alkaline melts.
†
CL—cathodoluminescence; note that other non-destructive imaging methods are potentially useful, such as backscattered electron, Micro-Raman, or Time-of-Flight (TOF) SIMS mapping.

combination of low radiogenic signals and required emplacement of a plutonic rock, with these events methods, however, zircon dates even from hand
corrections for initial disequilibrium in the U decay being thought to have happened faster than the sample-sized volcanic or plutonic rocks often dis-
chains (e.g., Schaltegger et al., 2015). Extending timespan indicated by stated uncertainties. Uncer- play a spectrum of ages rather than defining a point
accessory mineral geochronology into the Holo- tainties can be further reduced by averaging results in time. Zircon age spectra thus result from dates
cene requires application of U-series methods, in from replicate analyses of individual crystals and/or being distributed over a time interval in which at
which disequilibrium between 230Th and its indirect spots from the same sample or closely related least the extremes are distinctly resolved beyond
parent, 238U, is measured (U-Th geochronology). samples. Normally distributed ages from replicate analytical uncertainties. Age overdispersal rela-
The absolute age resolution for the U-Th method, analyses are generally vetted using the reduced tive to analytical uncertainties ranging between
however, diminishes when approaching secular chi-square statistics, with the mean square of 103 years and >106 years has been documented
equilibrium, which effectively limits its applicability weighted deviates (MSWD) expressing the degree for both U-Th disequilibrium and U-Pb geochro-
to zircon of <300 ka (e.g., Schmitt, 2011). Applying of dispersion of analyses around an average rela- nometers at different timescales, and it is found
these methods, sometimes in combination, has tive to the magnitude of the analytical errors. Data regardless of the analytical method applied (TIMS,
increasingly helped to resolve complex crystalli- sets in which the MSWD is within acceptable limits LA-ICP-MS, and SIMS) or the geological environ-
zation histories and inheritance in zircon, whereby of unity are reasonably interpreted to be derived ment of the sample (e.g., for volcanic and plutonic
the highest absolute age precision is achievable in from the same normal distribution. In this case, zircon). Examples are detailed in section 2.6 below.
samples of young age and/or by the analytically the uncertainty for the age is the standard error, The veracity of zircon age spectra is also upheld if
most precise method (Table 1). i.e., the standard deviation divided by the square stringent criteria for identifying potential Pb-loss
The general expectation of most geochronology root of the number of analyses. Averaging repli- or unrecognized inheritance (i.e., the incorporation
users is that applying any of these different and cate analyses results in much improved precision of recycled zircon typically from wall rocks) are
often complementary methods to zircon will lead compared to the individual ages (e.g., Crowley et upheld, although both effects are certainly relevant
to a well-defined age—the equivalent of a point on al., 2007). Data sets with MSWD values >>1 indi- for ancient zircon, or for some TIMS dates due to
the time axis (Fig. 1)—where assigned uncertainties cate data dispersion exceeding what is explainable the larger crystal volume sampled using TIMS as
only reflect the analytical and systematic uncertain- by analytical uncertainties alone. These are often compared to LA-ICP-MS or SIMS (e.g., Schaltegger
ties of the method (e.g., uncertainties in isotope interpreted as either affected by Pb-loss or zircon et al., 2015). Furthermore, smoking-gun evidence
ratio measurements, and decay constant errors, inheritance, which leads to younger and older ages against zircon age spectra being simply caused by
respectively). This age may then be interpreted as in a population, respectively. unrecognized Pb-loss or inheritance comes from
corresponding to a specific geologic event, such as With increased analytical precision and improved the U-Th zircon geochronology of late Pleistocene–
the eruption of a volcanic rock (Fig. 1) or magma spatial selectivity of available geochronological Holocene volcanic rocks, which are too young to

GEOSPHERE | Volume 19 | Number X

pre-eruption post-eruption involve high-temperature, mafic-to-intermediate

s
magma compositions, where eruption tempera-

de
ys

s
th
a

ca
da
ni

on
tures are close to zircon saturation conditions

de
n

m
ille

s
ia

s
a

to
de

to
nn
M
(e.g., in mid-oceanic-ridge settings; Hayman et

s
s

s
ca

ys
ille
-1

ur
ar

ar
nt
’s
10

da
ce

ye
Time al., 2019; Schmitt et al., 2011). Nonetheless, even

m
such zircon age data sets, when viewed system-
collapse and venting wide instead of at the scale of an individual
crystal entrainment, rim growth hand-sample, can be complex and indicative of
protracted magmatic evolution (e.g., Schmitt et
mafic magma injection
Process al., 2010).
melt extraction and assembly soil and water deterioration, Here, we review challenges in zircon geochro-
evolved melt segregation climate effects nology to reliably interpret zircon age spectra in the
large-volume eruption repose pyroclastic flows, tsunamis, ash-fall context of the different methods by which they were
life-span of magmatic arcs earthquakes generated. As our focus is primarily on magmatic
human observation systems, we largely disregard the possibility of
subsolidus (“hydrothermal”) zircon crystallization
Observable zircon age spectra eruption chronometers
14
C, 40Ar/39Ar, (U-Th)/He etc. in this review (e.g., Troch et al., 2018) and refrain
intra-crystalline diffusion
from any discussion of zircon crystallization during
Figure 1. Schematic compilation contrasting a geological event illustrated by the eruption of an island arc volcano with regional metamorphism; instead, we refer readers
the processes occurring before and after this event (pre- and post-eruptive) and their respective timescales (modified to excellent recent reviews (Kohn, 2016; Rubatto,
after Flaherty et al., 2018). Potential observables for reconstructing these processes are shown with their respective tem- 2017). Because zircon crystallization in magmas is
poral resolution. Zircon age spectra in volcanic and plutonic rocks are useful for processes operating at 103 to >106 year
directly linked to the thermochemical conditions
timescales, whereas diffusion modeling of crystals can bridge these timescales with those accessible by direct human
observation (e.g., through geophysical monitoring of active volcanoes). Instead of an eruption, magma extraction and of the melt that are faithfully recorded by zircon
intrusion would be an alternative scenario for geologically fast events, although a more inconspicuous one. (e.g., Claiborne et al., 2006, 2010), zircon age spec-
tra systematics, when properly understood, can
reveal processes of magma migration, accumula-
be affected by metamictization, and where U decay 40
Ar/39Ar (Fig. 1; e.g., Oberli et al., 1990; Keller et tion, and cooling timescales for natural systems at
products, including 230Th, are essentially immobile al., 2018). In this regard, the resulting zircon age timescales unattainable by other methods (Fig. 1).
in the crystal lattice (Cherniak et al., 1997). Thus, spectra in magmatic rocks are analogous to those Calvin Miller, to whom this themed issue is ded-
post-crystallization disturbance can be ruled out found in sedimentary rocks, and for interpreting icated, pioneered the investigation of accessory
for young zircon crystals, and spatially selective the geologic significance of these heterogeneous minerals to gain insights into magmatic processes
analysis (e.g., through SIMS depth profiling) can ages, igneous petrologists face problems similar that operate at scales ranging from individual volca-
distinguish growth domains at the micrometer to those of sedimentologists attempting to extract noes or plutons to planetary crusts. In this tradition,
scale that show significant rimward younging. Zir- depositional ages and information on provenance zircon age spectra are increasingly harnessed to
con growth rates (Table 1) have been determined from detrital zircon data (e.g., Coutts et al., 2019). constrain the thermal history of magmatic sys-
by this method (e.g., Storm et al., 2012; Friedrichs Notable exceptions are (per-)alkaline magmas, tems and the long-term magma fluxes that sustain
et al., 2021). where zircon solubility is extremely high (e.g., them (e.g., Caricchi et al., 2014, 2016; Ratschbacher
Thus, it has become evident over the past 25 Gervasoni et al., 2016) and zircon can crystallize et al., 2018; Friedrichs et al., 2021; Liu et al., 2021;
years or so that zircon crystallization in magmas rapidly (and sometimes even as millimeter-sized Lukács et al., 2021; Tierney et al., 2016; Weber et al.,
must be conceptualized as prolonged rather than megacrysts) within the resolution of U-Th and 2020; Melnik et al., 2021; Angeles-De La Torre et al.,
instantaneous, and this may significantly predate U-Pb chronometers (Table 1). Such rocks display 2023). Used in this sense, they are complementary
a volcanic eruption or emplacement of a shal- comparatively uniform zircon age distributions to crystal-scale diffusion chronometers that provide
low intrusion (Fig. 1; Reid et al., 1997). This has in individual samples, but not necessarily for magmatic timescale information if diffusion param-
severe implications for interpreting the timing of different samples from the same volcano (e.g., eters are known, and which—unlike zircon—also
a geological event such as a volcanic eruption Campi Flegrei: Fedele et al., 2006; Gebauer et can be applied to mafic rocks where zircon is com-
from zircon age data, especially when compared al., 2014; Laacher See: Schmitt et al., 2010, 2017). monly absent, but with absolute time information
to low closure temperature chronometers such as Other cases where zircon crystallizes rapidly (Fig. 1; see Costa et al., 2020, for a recent review).

GEOSPHERE | Volume 19 | Number X

■■ 2 REPRESENTATION AND PROPERTIES Moreover, analytical uncertainties for young zircon 2.3 Kernel Density Estimates (KDEs)
OF OVERDISPERSED ZIRCON AGES are primarily controlled by low signal intensities,
(AGE SPECTRA) unlike older zircon crystals, where other sources of KDEs are similar to PDPs in basic principle: Each
uncertainty affecting the results dominate, such as measurement is arranged into a distribution (referred
2.1 Basic Considerations and Histograms correction factors for isotopic mass fractionation in to here as a “kernel,” for which different shapes can
ID-TIMS or relative sensitivities in LA-ICP-MS and be prescribed, including Gaussian), where the width
Zircon crystallizes primarily upon cooling SIMS (e.g., Schaltegger et al., 2015; Schoene, 2014). of the distribution is not determined by the analytical
and differentiation of magma as first experimen- Hence, uncertainties for young high-U zircon are uncertainty (as in PDPs), but by a “bandwidth” that is
tally quantified by Watson and Harrison (1983). often much smaller than those for low-U zircon. algorithmically adjusted based on point density and
Because magmatic systems evolve dynamically, One should note that as the number of analyses, therefore variable across the age range. As noted by
with magma reservoirs undergoing episodes of n, approaches infinity in a data set, the shape of Vermeesch (2012), KDEs have advantages for prove-
recharge and eruptive venting (e.g., Annen et al., a histogram converges with that of a continuous nance analysis using detrital zircon age spectra where
2006), thermal variability in these reservoirs over curve. However, time and cost are limiting factors analytical uncertainties and ages vary considerably.
time and space is expected. At low temperature and for generating such high-n data sets; so, for most Drawbacks of KDEs are the complexity of the
in subalkaline melts, the kinetics of zircon dissolu- samples, an approach is required where an accept- formulas used and the potential for the arbitrary
tion are sluggish; short-lived prograde pulses and able approximation can be extrapolated from a selection of bandwidths. Whereas histograms
the associated chemical changes resulting from limited number of dates (e.g., n = 20–100). The gen- and PDPs are easily produced with minimal input
magma recharge only cause incomplete resorption eration of such an approximation can be achieved parameters and subsequently easily reproduced by
(Watson, 1996). The propensity of zircon to survive in multiple ways, and the resulting representation other researchers, KDE formulas are typically hid-
such heating events is underscored by the pres- of zircon age spectra is dependent on several key den from view and can vary in how users of the
ervation of rounded crystals in arc-volcanic rocks assumptions (see section 2.5). program (e.g., DensityPlotter; Vermeesch, 2012)
with bulk compositions and eruptive temperatures adjust the bandwidth. For example, consider a syn-
at which zircon is undersaturated (e.g., Weber et thetic bimodal data set with two prescribed peaks
al., 2020). If not erupted, preexisting zircon, albeit 2.2 Probability Density Plots (PDP) at 5 Ma and 7 Ma (7000 ages; Fig. 2A). A 250-point
potentially partially resorbed, can become the random sample of the population is selected, rep-
nucleus for renewed zircon crystallization during PDPs are another straightforward way to resenting the collection of data from a much larger
retrograde episodes (Szymanowski et al., 2020; generate age spectra in addition to histograms. population. The algorithm of Botev et al. (2010)
Bindeman and Melnik, 2016). The shape of the The age and uncertainty of each measurement yields multiple peaks that actually represent noise
natural distribution of zircon ages in individual provides the mean and standard deviation of a in the original data set (Fig. 2B). The estimate of
samples thus can be highly variable, and it also Gaussian distribution, and the sum of these over- DensityPlotter (Vermeesch, 2012), with an adaptive
will depend on the method applied to determine lapping Gaussian distributions generates the age kernel bandwidth, avoids such overinterpretation
zircon ages (see section 3). Different approaches spectrum; alternatively, the probabilities can be and faithfully reproduces the original population
to visualizing zircon ages, especially those that are summed into a cumulative PDP. The utility of this shape, which indicates that this is the algorithm of
over-dispersed relative to analytical uncertainty, approach, however, has been called into question choice for KDEs (Fig. 2C). However, the most notable
have been developed, and these are described in if analytical uncertainties are invariable and small results of this demonstration are that (1) KDEs can be
the next paragraph and sections 2.2–2.4. compared to the range of values in the data set. Ver- strongly model-dependent; (2) it is essential to report
Most intuitively, ages can be displayed in a his- meesch (2012) noted that PDPs can “punish” less metadata with KDEs (e.g., constant versus adaptive
togram with individual analytical results assigned precise measurements by under-representing such bandwidths); and (3) the algorithm of Vermeesch
to discrete bins. The appeal of a histogram is its measurements in detrital spectra. Namely, the sam- (2012) appears to be suitable for modeling zircon age
simplicity (the shape is only affected by the bin pling of a zircon population with higher inherent spectra with realistic uncertainties, provided that no
width) and the faithful reproduction of the data error (e.g., due to lower U concentration) will gen- major changes are introduced to the source code.
with a low risk of over-interpretation. However, erate a broader Gaussian distribution. Then, the
the use of a single bin width (e.g., equivalent to PDP of this distribution will be correspondingly
the analytical uncertainty) for an entire distribution smoothed due to the higher uncertainties of the 2.4 Radial and Abanico Plots
is precisely what makes data visualization diffi- individual data points. For these reasons, PDPs
cult for large data sets that cover a comparatively are sometimes foregone in favor of kernel density Another method of data visualization that is cur-
large age range relative to analytical uncertainties. estimates (KDEs). rently underutilized in U-Th-Pb zircon geochronology

GEOSPHERE | Volume 19 | Number X

12 12
Figure 2. (A) A 7000-point synthetic

10
A synthetic population
10
B Botev et al. (2010)
15 C Vermeesch (2012)
data set with prescribed age peaks
at 5 Ma and 7 Ma; (B) histogram of a
250-point random sample and a kernel
8 8
density estimate (KDE) generated by

Frequency
10
the algorithm of Botev et al. (2010);
6 6
(C) histogram of the same 250-point
random sample as in part B, but us-
4 4
5 ing the algorithm of Vermeesch (2012)
to generate a KDE that considers an
2 2
adaptive kernel bandwidth. Note the
0 differences in curve smoothing.
0 0 ||| || | ||||||| |||||||
||||||||||||||
|||||||||||
||||||||||||||||||||||||||||||||||||||| |||| | ||||||| || |||||| || | ||||||||| ||||| ||||| | |
3 4 5 6 7 8 9 3 4 5 6 7 8 9
3 4 5 6 7 8 9
Age [Ma]

is the radial plot, which was introduced by Galbraith al., 2007, for definitions of antecrystic and xenocrys- data point, decreasing the risk of overinterpreta-
(1990) to represent fission-track data. Here, an addi- tic as used here). A modified version of this plot, the tion even at a glance. Secondly, the existence of
tional dimension is added, allowing the user to abanico plot (named after the Spanish term for a many zircon analyses of similar age but varying
visualize the amount of relative uncertainty on each folding fan), includes a PDP on the right side of the uncertainty works to create vectors in the radial
data point on the x-axis, as well as the distribution graph (Vermeesch, 2012), combining properties of plots that are easily differentiated from one another,
of ages on the y-axis (Figs. 3B and 3C). Individual Figs. 3A and 3B. and the slope of these vectors anchored at the ori-
age peaks are calculated, as with KDEs (Fig. 3B), with The advantages of the radial plot being applied gin indicates the age (marked along the radial part
the bonus that xenocrystic populations that may be to magmatic zircon populations are, firstly, that the of the plot). Lastly, the addition of another data
difficult to spot in a histogram appear alongside the consideration of analytical uncertainty immediately visualization method increases the confidence in
rest of the data at a lower point density (see Miller et informs the user about the relative value of each the user’s interpretation, particularly because the

32 Ma

A KDE B radial plot 30 Ma C abanico 28

Central age: 16.5 Ma

Pannonian zircon 24
MSWD = 110 25
(n=1000)
40

Peak 1: 14.8 Ma (31.6%)

Peak 2: 17.2 Ma (66.7%)
Frequency

Peak 3: 27.5 Ma (1.7%) 20

+2
20

+2 16
-2
-2
10

15
13

σ/t 14 2 0.9%
0

15 20 25 30 0 50 100 150 200 250 0 22 44 66 88 110 132

Age [Ma] t/σ t/σ
Figure 3. Comparison of plotting methods for the Pannonian zircon data set (Lukács et al., 2018): (A) histogram and kernel density estimate (KDE, red curve) of the data set,
demonstrating a bimodal distribution; (B) radial plot after Galbraith (1990), where x values are plotted proportionally to 1/σ (uncertainty) and y values are determined from
the deviation of each datum from a central value (e.g., a weighted average). All points on a line through the origin have the same age value that is denoted on the arc of a
circle drawn at a convenient x value. The strong scatter demonstrates variability in data point uncertainties as a result of different measurement uncertainties (e.g., resulting
from variably radiogenic Pb in young zircon); (C) abanico plot, with added probability density plot (PDP, red line on right), combining properties of a histogram and PDP/KDE
curves as in part A with the data representation in part B.

GEOSPHERE | Volume 19 | Number X

interpretation of age spectra is only a semiquan- spiky, provided there is sufficient temporal reso- Pullen et al., 2014; Vermeesch, 2004). For magmatic
titative method that requires a careful qualitative lution (i.e., there could be as many age peaks as studies, where the age range is much smaller, a
assessment as opposed to strict quantitation. zircon crystals or crystal subdomains analyzed). One data set of 100 zircons may encompass the natural
The attractiveness of the radial/abanico plot is must therefore make a rough judgment about the variability of the population. A simple subsampling
that it easily and clearly captures the entire age number of crystals required to accurately capture experiment demonstrates this concept (Fig. 4):
spectrum, making it applicable to volcanic and plu- the variability of their data set. For provenance anal- Using subsamples of 20, 50, and 100 zircons from
tonic zircon suites that may have a large xenocrystic ysis in sedimentary records (where age ranges may a data set of Miocene volcanic zircon crystals from
component. This type of plot additionally excels at span billions of years), it is necessary to look at data Pannonia (Lukács et al., 2018), one can see that the
finding peaks of high sample density, which is par- sets numbering 100–1000+ zircons to resolve low- overall variability of the entire data set is adequately
ticularly applicable to LA-ICP-MS/SIMS data sets. abundance age populations (Dodson et al., 1988; captured by subsampling 50–100 zircons, which is
Application to TIMS data sets, on the other hand,
is complicated by their relatively small size (typi-
A B

5
cally <20 zircons) and high precision, which leads
to the accentuation of individual zircon ages in a

4
radial plot as opposed to a representative age of

1.5
abundant zircon crystallization. (n = 20) (n = 50)

3
1
2.5 How Much is Enough? The Limits of

2
Undersampling

Frequency (relative probability)

0.5
The central supposition behind zircon age spec-

1
tra analysis is that practical sampling of the zircon
population (which is typically limited to tens to

0
hundreds of zircons) still allows generation of an
accurate representation of the true distribution. 14 16 18 20 22 24 26 14 16 18 20 22 24 26
This is, however, far from trivial, because selection
bias due to the method of zircon extraction (which

70
may hinder, e.g., crystals included in phenocrysts C D

10
from being liberated) or picking (where typically

60
large, euhedral crystals are preferred) can hardly
be avoided. With this caveat, we address the more

50
8
(n = 100) (n = 1329/1339)
tangible statistical problem of how many of the

40
crystals that are available should be analyzed after
a rock is subjected to a certain zircon identification 6

30
and extraction procedure (Table 1). The larger the
4

data set, the more accurate the representation of

20
the true spectrum, but in practice, there are upper
2

limits on the number that can be analyzed given

10
time and cost considerations. The optimal size of
n depends on (1) how many modes there are in a
0

0
spectrum, as well as (2) the separation of the peaks 14 16 18 20 22 24 26 14 16 18 20 22 24 26
and the analytical uncertainty, and (3) the model Age [Ma]
used to generate the spectral representation.
Figure 4. Histograms and corresponding kernel density estimate (KDE) plots demonstrating the effects of undersam-
The number of modes in a zircon age distribu-
pling illustrated by subsampling of (A–C) 20, 50, and 100 zircons, respectively, from a data set of Miocene volcanic zircon
tion is effectively unconstrained, and the shape of a crystals from Pannonia, compared to (D) the full population of 1339 zircons with 10 analyses plotting outside the range
natural population will theoretically always appear (Lukács et al., 2018). 206Pb/238U ages are plotted, with typical uncertainty of ca. 0.4 m.y., or ~3%.

GEOSPHERE | Volume 19 | Number X

in line with estimates from Caricchi et al. (2016). suggests that the minimal sample size for resolv- Zürich using SIMS and LA-ICP-MS methods, respec-
Here, the larger subsample also begins to resolve ing bimodality in the Pannonian zircon example tively, and include: (1) The Belfond lava dome from
input from xenocrystic zircons at ca. 25 Ma as well is ~50 zircon analyses, but reliable estimates of the late Pleistocene Soufrière Volcanic Center (Saint
as a shoulder in the data at ca. 18–19 Ma, which the percentage of each mode may require a larger Lucia, Lesser Antilles Arc); (2) the Miocene plutonic
may reflect another population too close in age to number of analyses (e.g., 100; Fig. 4D). In the data complex of La Gloria (Central Andes); (3) the Oli-
be resolved by LA-ICP-MS. Importantly, small data set used here, the range between the modes is gocene volcanic province of the Southern Rocky
sets (10–20 zircons) demonstrate no bimodality, sufficiently large, and analytical uncertainties are Mountains volcanic field (also previously called
which suggests that the KDE adequately protects sufficiently low. Under other circumstances, these the San Juan volcanic field), western USA; and
against over-assigning individual peaks. For such criteria might not be met, and it is difficult to define (4) the Ordovician plutonic complex of the Sierra
small data sets, sampling only suggests an age of a set of rules that applies universally, apart from de Famatina–Valle Fértil in NW Argentina.
roughly 15–17 Ma (Figs. 4A and 4B). In larger subsa- two basic principles: (1) Δt/σ must be at least 3–5, The Soufrière Volcanic Center data (Fig. 5) com-
mples, however, one can more confidently argue for and (2) a consistent algorithm that is commonly prise zircon from samples of the dacitic Belfond lava
the existence of two age peaks and perhaps identify available must be used (e.g., Vermeesch, 2012). dome and cogenetic granodioritic enclaves that
different proportions of xenocrysts and the exis- Slightly different considerations hold for U-Th erupted at ca. 13.6 ka (Barboni et al., 2016; Schmitt
tence of subsidiary age peaks (Figs. 4C and 4D). zircon ages, which cover an age range between ca. et al., 2010). The data set includes zircon rim anal-
However, these latter two interpretations should be 0 ka and 300–350 ka. Because of the exponential yses of pristine growth surfaces, as well as interior
considered strictly qualitative, and they require cor- relationship between isochron slope m (which for analyses of sectioned crystals, but we do not differ-
roborating evidence. A statistically robust approach single zircon ages is usually defined by two points: entiate between those in this comparison (cf. Schmitt
for identifying distinct populations in such mixtures the measured zircon composition and a constant et al., 2010). Lateral overlap of the ~30-µm-diameter
was developed by Sambridge and Compston (1994) model melt composition that is typically derived ion beam used in SIMS analysis for crystal interior
that is based on deconvolving multiple Gaussian from U-Th whole-rock or glass analyses) and time, analyses could lead to mixed ages, whereas this is
peaks from a polymodal PDP. the absolute age uncertainty increases as secular largely mitigated for the rim analyses in which the
The distance between age peaks (Δt) and analyt- equilibrium is approached. Kent and Cooper (2018) SIMS sputtering is parallel to the growth direction,
ical uncertainties plays a large role in determining concluded that >100 zircon analyses are required integrating to a depth of <5 µm. Most of the Bel-
whether age peaks can be resolved. Using the con- to characterize U-Th zircon age spectra and to cor- fond zircon crystallization ages significantly predate
vention of a 95% confidence interval (abbreviated rectly interpret them in the context of a magma the eruption age (dashed lines in Fig. 5) which was
as 2σ, where σ refers to the standard error of an system’s thermal history, but as ages approach independently constrained by (U-Th)/He zircon dat-
individual analysis; Table 1), it follows that for two secular equilibrium, the absolute age uncertainty, ing. Interestingly, the dominant zircon age mode at
age peaks to overlap at less than 5%, their separa- even for the same analytical precision, becomes too ca. 28–29 ka is similar in the zircon populations of
tion must be at least 4σ (2σ for each peak). This is large to meaningfully distinguish different age pop- both the host lava and enclosed plutonic enclaves,
consistent with the Bayesian modeling of Keller et ulations. In favorable cases (e.g., high-radiogenic which suggests recycling of zircon from crystal-rich
al. (2018), who determined that the ideal separation Pb yields), U-Pb geochronology can seamlessly domains of an intrusive reservoir during a younger
of age peaks should fall within the 3–5 σ range (i.e., complement U-Th ages. magmatic episode (Fig. 5). These age modes pre-
Δt/σ = 3–5). In the previous example, with large age date the dome eruption by ~14 k.y. and the last
peaks at 14.8 Ma and 17.2 Ma and an analytical major explosive eruption in the region at ca. 20 ka
uncertainty amounting to ca. 0.7 Ma at the 4σ level 2.6 True Zircon Age Spectra by ~8 k.y. Several minor peaks in decreasing abun-
(the mean of the 3–5σ range identified by Keller et dance are present between ca. 40 ka and 114 ka, as
al., 2018), the corresponding Δt of 2.4 Ma trans- Obtaining large zircon data sets is not always well as some older crystals, in particular in the plu-
lates into a Δt/σ of ~14. The apparent shoulder at practical, especially for ID-TIMS geochronology, and tonic enclaves. The central age of the population
ca. 18.5 Ma presents a conundrum, however. While we therefore showcase well-documented examples (ca. 62 ka and 49 ka for lava and plutonic enclaves,
on the one hand, the Δt between this “peak” and where zircon age spectra have been obtained at respectively) has no geologic significance, but the
the one at ca. 17.2 Ma is 1.3 Ma (well outside the 4σ high-n sampling. To illustrate, we have summarized significance of the older zircon crystals in terms of
criterion), the distribution appears to be continuous in Figs 5–8 PDPs, KDEs, and radial plots from four continuous versus punctuated crystallization is more
toward older ages, which suggests that this period data sets of U-Pb and U-Th zircon data covering a difficult to distinguish, especially when the ion beam
experienced prolonged zircon crystallization and wide age range. Data have been collected over the might have overlapped different growth domains
may in fact be composed of many smaller, and ulti- past few years at the University of California at Los (Fig. 5). The predominance of ages between ca.
mately unresolvable, peaks. The test applied here Angeles/Heidelberg University (UCLA-HD) and ETH 40 ka and 114 ka, however, corresponds well with

GEOSPHERE | Volume 19 | Number X

major eruption pulses in the Soufrière Volcanic Cen- Belfond (St. Lucia) volcanic Belfond (St. Lucia) plutonic
ter and a possible hiatus between ca. 250 ka and 0 20 40 60 80 100 200 ka 0 20 40 60 80 100 200 ka
110 ka, which again supports crystal recycling in a 1.0 1.0
long-lived magma system (Schmitt et al., 2010). A B

Cumulative probability
Relative probability
La Gloria U-Pb zircon LA-ICP-MS data show a 0.8 0.8
nearly Gaussian distribution, with a single peak

eruption

eruption
at ca. 10.4 Ma (Fig. 5), which agrees within error 0.6 0.6
with a published U-Pb zircon LA-ICP-MS age
of 10.3 ± 0.2 Ma (Deckart et al., 2010). However, 0.4 0.4
higher-precision ID-TIMS data on the same sam-
ples establish that the pluton crystallized zircons 0.2 0.2
n = 218
for almost 2 m.y., between 10.2 Ma and older than
n = 139
12.0 Ma (Gutiérrez et al., 2018). Hence, it would be 0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
erroneous to interpret the relatively simple “uni-
zircon-melt isochron slope zircon-melt isochron slope
modal” population seen by the LA-ICP-MS data as
all zircon crystallizing during a single event corre- 1 1
sponding to the weighted mean age of ca. 10.4 Ma. C central age = 61.6 ka (n=139)
200 ka D central age = 48.9 ka (n=218) 200 ka

That such an interpretation is inadequate is also MSWD = 66 0.8 MSWD = 160 0.8
underscored by the elevated MSWD for the entire
0.6 100 100
data set. 0.6
At even larger spatial and temporal scales than 80 80
+2 +2
for individual volcanoes or plutons, zircon age spec- 0.4 60 60
0.4
tra are also found for entire volcanic fields or arc -2 -2
segments. Geological background and additional 40 40
age information is presented in the references pro- 0.2 0.2
erup Peak 1: 27.8 ka (35.2%) erup
vided, and here, we only summarize zircon ages Peak 1: 28.8 ka (25.3%) tion Peak 2: 40.4 ka (19.3%) tion
Peak 2: 46.1 ka (22.7%) 20 20
from these cited studies. Large zircon age data sets Peak 3: 59.9 ka (22.8%)
Peak 3: 84.3 ka (28.8%)
Peak 4: 114 ka (15.2%)
for the Southern Rocky Mountains volcanic field (Sli- Peak 4: 192 ka (23.2%) 0 0
Peak 5: 448 ka (7.5%)
winski et al., 2022) and the Famatina Province show
0 20 40 60 80 100 120 140 0 50 100 150 200
distinct peaks (Figs. 7 and 8). Again, both data sets
t/σ t/σ
display elevated MSWD values and central ages that
lack any significance for dating a specific geological Figure 5. (A, B) Probability density plots (PDPs, with relative and cumulative probabilities in blue and red curves, respec-
tively), and (C, D) radial plots (individual data points are in green). (A, C) Plots for U-Th zircon ages for volcanic and (B, D)
event. For the Southern Rocky Mountains volcanic
plutonic samples of the Belfond lava dome, Soufrière Volcanic Center, St. Lucia (data were compiled from Schmitt et al.,
field, two main peaks stand out at ca. 33–34 Ma and 2010, and Barboni et al., 2016). Bottom line in parts A and B and inner circle in parts C and D indicate the slope of the
ca. 27–29 Ma that correspond to the peaks of activity zircon-melt isochron using glass or whole-rock compositions as an anchor for the melt from which zircon crystallized
in the northern and southern parts of the volcanic (see Schmitt et al., 2010, and Barboni et al., 2016, for details). Dashed line marks the eruption age of the Belfond dome at
13.6 +0.45/–0.39 ka as independently determined by (U-Th)/He thermochronology (Schmitt et al., 2010). MSWD—mean
province, respectively. There is also a younger peak
square of weighted deviates.
at ca. 23 Ma, which corresponds to the beginning of
the Rio Grande Rift in the region, and the emplace-
ment of some large silicic units (in particular, the and the Famatina regions are relatively old, the Curry et al., 2021). Moreover, zircon ages systemat-
Sunshine Peak Tuff that erupted at 22.93 ± 0.02 Ma; LA-ICP-MS uncertainties likely mask more com- ically correlate with geochemical indicators such as
Lubbers et al., 2020; Mehnert et al., 1973). plexity within the zircon age population. Based Yb/Dy, which reflects accessory mineral saturation
For the Famatina region, there is a broad main on high-precision ID-TIMS data, we know that zir- and hence the temperature evolution of the magma
peak, centered on an age of ca. 463 Ma, and a con from single ignimbrite units of the Southern (Bachmann et al., 2007; Wotzlaw et al., 2013). Sim-
younger tail (<350 Ma) that may represent an epi- Rocky Mountains volcanic field has crystallization ilarly, preliminary unpublished ID-TIMS data on
sode of Carboniferous magmatism. However, as intervals on the order of 0.5–1 m.y., with potential some of the Famatinian units suggest peaks of zir-
both the Southern Rocky Mountains volcanic field zircon populations hidden within those peaks (e.g., con growth at ca. 475–480 Ma and ca. 465–470 Ma,

GEOSPHERE | Volume 19 | Number X

La Gloria 12.5 Ma

300
A B 12

Frequency (relative probability)

250
11.5

200
11

standardised estimate
150
+2 10.5
−2

100
10

50
n=564 central age = 10.4 Ma (n=564)
MSWD = 12 9.5

0
||| |||||||||
|||||
||||||
|||||
|||
|||||||||
||||||||||
|||||
|||||||
||||||||||
||||
||
||||||
|||
||||||
|||||
||||
||||||
||||||
||
||||
|||||||
||||
|||||
||||
||||
||||||
|||||
||||||||
||
||||
|||
|||||
||||
|||
|||||
||||||||||
||||||
||||||||||||||
|||||
|||
|||||||||||||||
||||||||||||||
|||
||||||
|||||
|||||
||||||||
||||
|||
||||||||
|||||
|||
||||||||||||||
|||||||||
|||||
|||||||||||||
|||||
|||||||
||||||
|||||
|||||||
|||||||
||
|||||||||||
|||||
| |||||
|||||||||||||||||||
| ||| | | |

9 10 11 12 13 0 50 100 150 200 250

Age [Ma] t/σ
Figure 6. (A) Histogram and kernel density estimate (red curve) and (B) radial plot for U-Pb ages from La Gloria Pluton, Central Andes. The same data are
plotted in parts A and B. MSWD—mean square of weighted deviates.

40 Ma
Southern Rocky Mountains Volcanic Field
2000

A B central age = 29.6 Ma (n=7023)

MSWD = 12
35
Frequency (relative probability)

standardised estimate
n=7023
1500

+2 30
−2
1000

25
500

Peak 1: 23.1 Ma (3.8%)

Peak 2: 26.7 Ma (33.1%)
Peak 3: 28.5 Ma (24.1%)
Peak 4: 32.9 Ma (25.5%)
Peak 5: 35.7 Ma (13.5%)
21
0

||||||||||||||||
||||||
|||
||||||
|||||
|||||
||
|||
||||
|||||
|||
||||||||||
||
||||
||||||
|||||
|||||
||
||||
||||||
||||
|||||
||||||
||||
|||||||
|||
|||
||||||||||
|||||||||||
||||||||
|||||
|||||||||
||||||||||||
||||||||
|||||
|||||
|||||||||
||||||||||||||||||
|||||||||||||||||||||||||||||||||||||||||||||||||
|||||||||||||
|||
|||||
||
||||
||||
||||||
||||
|||||
||||
|||||
|||
|||||||
||||||||||
||||
||||||
|||||
||
|||
||||
||||||
|||||
||||||
|||||
|||||
|||
||||||
||||||
||||
||||||||
|||
||||
|||||||
|||||||||
|||||||||
||||||||||||||||||||||||||||||||||||||||
|||||||||
|||
||||||||||
||||||||||||
|||||||
|||||||||
||||||
||||||||||
||||||
||||||||||
||||||||
|||
||||
|||||||||
|||||||
||||||||
||||||
||||||||
|||||||
||
||||||||
||||
|||||||
|||||||
|||
||||||||||||||||||||
|||
||||| |||

20 25 30 35 40 0 50 100 150 200 250

Age [Ma] t/σ

Figure 7. (A) Histogram and kernel density estimate (red curve) and (B) radial plot for U-Pb zircon ages from Southern Rocky Mountains Volcanic Field. The
same data are plotted in parts A and B; data source is Sliwinski et al. (2022). MSWD—mean square of weighted deviates.

GEOSPHERE | Volume 19 | Number X

550 Ma

Famatinian arc
central age = 448 Ma (n=873) 500
MSWD = 33
A B

standardised estimate
Frequency (relative probability)
500
+2 450
−2

400
300
400

200
350

100
n=873 Peak 1: 344 Ma (10.3%)
Peak 2: 463 Ma (89.7%)
300

0
||||||||||
||||||||||||||||||||||||
||||||||||||||||||||
|||||
|||||||||
||||||||||||||||||
||||||||||||
||||||||
|||||||||||||||||||||||
|||||||||||||||||
|||||||||
||||||||||||
|||||
|||||||||
||||||||
||||||
|||
||||||||
|||
||||
|||
|||||||
||
||||
|||||
||
||||||
||||
||||||
||||
||||||||
|||||||
||
||||
||||||
||||||||||||
||||||
||
||||
|||||||
||||||||
|||
|||
||||
|||||||||
||||||
|||||||
|||
||||
||
|||||
|||||
|||
||||
||||||||||
||
||||
||||
|||
||
||||||||||
|||||||||
||||||||
||||||
|||||||||||||||||||||
||||||||
||| || |

250 300 350 400 450 500 550 600 0 20 40 60 80 100

Age [Ma] t/σ

Figure 8. Histogram and kernel density estimate (red curve) and (B) radial plot for U-Pb zircon ages from the Famatinian arc. The same data are plotted in
parts A and B. MSWD—mean square of weighted deviates.

hidden in the broad Ordovician peak of the com- Unrecognized inheritance, especially when xenocrys- analyzing matching secondary zircon references
paratively low-precision LA-ICP-MS data. Hence, as tic contaminants are only marginally older than the to identify potential under- or over-estimation of
pointed out in section 2.5 above, distinguishable magmatic episode, can also shift ID-TIMS dates analytical error. Thus, we urge caution in invoking
events must be separated by >3–5σ to be isolated. to older ages (e.g., Widmann et al., 2019). Similar “protracted zircon crystallization” as a carte blanche
With typical errors of ~1–2% (at 2σ) for U-Pb zircon considerations hold for data from spatially selec- for explaining away excess analytical scatter, particu-
LA-ICP-MS data, this means that igneous pulses for tive methods, where CA pretreatment also shows larly for geologically old igneous rocks, because the
San Juan and Famatina need to be separated by promise but may not always be the silver bullet to magmatic lifetimes typically observed for young vol-
more than 1–1.5 m.y. and 10–15 m.y., respectively, eliminate zircon affected by Pb-loss (e.g., Kryza et al., canic systems are but a minor fraction of the overall
to be statistically distinguishable. 2012; Sliwinski et al., 2017; von Quadt et al., 2014). analytical uncertainty. Analytical reliability, e.g., from
Bias can also affect spatially selective methods: analyzing homogeneous and well-characterized zir-
LA-ICP-MS dates are sensitive to fractionation correc- con references, is a prerequisite for identifying true
2.7 False Zircon Age Spectra tions that depend on the depth of the ablation crater, zircon age spectra.
and because ablation characteristics vary depending
However, not every zircon age data set where on the structural state of zircon, this may result in
excess scatter beyond analytical uncertainties exists apparent age variability (Marillo-Sialer et al., 2016; ■■ 3 UNDERSTANDING AND MODELING
(commonly inferred from an elevated MSWD value) Sliwinski et al., 2017). SIMS analysis of high-U zircon ZIRCON AGE SPECTRA TO EXTRACT
should be considered a true zircon age spectrum. is known to yield apparent older ages when the rela- INFORMATION ABOUT MAGMATIC
Sampling bias and analytical problems can produce tive sensitivity factor (RSF) is calibrated on normal-U SYSTEMS
excess scatter, and common issues are often specific zircon references (e.g., Schaltegger et al., 2015;
to the analytical method applied. For example, even White and Ireland, 2012). Lastly, assigning errors 3.1 Fundamentals and Modeling Rationale
with chemical abrasion pretreatment, zircon domains is model-dependent in all analytical methods, and
that experienced Pb-loss may still contribute to not all sources of error are always identifiable. Thus, In the second part of this review, we focus on
an ID-TIMS analysis (e.g., Widmann et al., 2019). it is essential to monitor analytical performance by existing and upcoming forward models to generate

GEOSPHERE | Volume 19 | Number X

distributions of zircon crystallization ages with the used for dating can introduce bias (i.e., high spa- spherical magmatic intrusion cooling at depth
goal of quantifying rates and durations of magmatic tial resolution on sectioned crystals in SIMS and (Fig. 9) before addressing more complicated but
input into volcanic plumbing systems (e.g., Caric- LA-ICP-MS versus whole zircon by TIMS; Table 1). realistic scenarios in which a magma system under-
chi et al., 2014, 2016; Friedrichs et al., 2021; Liu et We first introduce these general concepts behind goes recharge. In the simplest case, magma is
al., 2021; Lukács et al., 2021; Tierney et al., 2016; the modeling and then discuss the robustness of the injected instantaneously at its liquidus temperature,
Ratschbacher et al., 2018; Weber et al., 2020; Melnik critical assumptions in the following section (3.2). (T0), and the intrusion cools starting from its outer
et al., 2021). The main idea behind this modeling To explain the fundamentals of the modeling portions. Zircon crystallizes between Tzr down to Ts ,
approach is to identify for which set of input param- approach, we can consider the simple case of a and the rate of zircon crystallization either decreases
eters a computed zircon distribution—including age
and possibly crystal age zonation as well as chemis-
try—is the closest to measured zircon from volcanic Linear 12 9 6 3 T0 magma
or plutonic rocks. These models are based on five A measurable ages
Non linear 12 7 4 1
tenets: (1) the thermal history of a magmatic intru-
Tzr
sion can be adequately predicted, with reasonable
assumptions regarding the shape of an igneous

Temperature
body (e.g., Annen et al., 2015); (2) zircon ages
record magmatic crystallization; (3) zircon crystal-
spherical
lizes between the zircon saturation temperature (Tzr) intrusion
and the solidus temperature (Ts), and the zircon frac-
tion per unit mass of magma decreases from Tzr to TS
Ts according to an empirical zircon saturation model
(e.g., Baker et al., 2002; Boehnke et al., 2013; Borisov
and Aranovich, 2019; Gervasoni et al., 2016; Shao et
Time
al., 2020; Watson and Harrison, 1983); (4) the number B
of zircon crystals crystallizing in a magmatic plumb- 1 3 1 3 1 3 1 3 5 9 5 9 5 9 5 9 1218 1218 1521 2430 3642 2424
ing system at each time interval is proportional to
the change in the temperature distribution between

n. measurable ages
Tzr and Ts; and (5) the fraction of measured zircon of
different ages is representative of the proportions
of zircon that crystallized within each age interval.
This last assumption deserves further com-
ment, as the age distribution ultimately measured
depends on the type of sample investigated (i.e.,
plutonic or volcanic) and the analytical method
used (Table 1). Firstly, volcanic and plutonic zir-
cons will sample different volumes and evolutionary
stages of a magmatic system as plutonic zircons
can grow in areas of very high crystallinity (and
possibly hydrothermally at subsolidus conditions), Zircon ages
whereas rheological thresholds for magma being
Figure 9. Schematic representation of the temporal evolution of temperature and measurable zircon crystallization ages
able to erupt require that volcanic zircon is largely within a spherical intrusion where each temperature interval has a randomly prescribed number of measurable zircon
derived from the liquid-dominated part of a system crystals (cyan and red numbers in part A). (A) Curves of different colors show the evolution of temperature within the
at the time of eruption. As a telling example, basalt reservoir portion with the corresponding color (inset). Gray shadings indicate different temperature intervals within
is typically devoid of zircons, but many gabbros which zircons crystallize at different rates (as shown in the legend for two models of zircon crystallization, linear and
nonlinear). For nonlinear, we presume that with decreasing temperature there is an exponential decrease in the number
contain zircon that crystallized interstitially from
of zircons crystallizing per unit time, as suggested by zircon saturation models. For the linear model, zircon crystalliza-
residual melt at near-solidus conditions (e.g., Hay- tion rates simply decrease linearly with temperature. (B) Schematic distributions of measurable zircon ages resulting
man et al., 2019). Secondly, the analytical technique from cooling of the spherical reservoir and considering a linear (cyan) and nonlinear (red) zircon crystallization model.

GEOSPHERE | Volume 19 | Number X

linearly or (more realistically) exponentially with with respect to zircon collected from the solidified the repeated intrusion of dike- or sill-shaped mag-
temperature (Fig. 9A). The exponentially decaying magma reservoir (Fig. 10B versus Fig. 10C). An matic batches. In these models, dikes or sills are
mass fraction of zircon is consistent with zircon exception could result from the remelting of a kinematically inserted into a model space while
saturation calibrations, where ~40% of the total completely solidified portion of the reservoir, thus moving host rocks to accommodate magmatic
amount of zircon that can form in a batch of magma releasing its zircon cargo to the melt. The zircon input, and the subsequent thermal evolution of the
crystallizes within 25 °C below the onset of zircon age distributions in plutonic and volcanic rocks are crust is monitored while taking partial melting and
saturation, and only <10% crystallizes over the different because an eruption must occur before effects such as latent heat into account (e.g., Dufek
last 25 °C before reaching the solidus temperature the full solidification of a magmatic reservoir, and and Bergantz, 2005; Annen et al., 2006; Caricchi et
(Harrison et al., 2007). Several observations can be therefore the youngest measurable age in volcanic al., 2014; Tierney et al., 2016; Ratschbacher et al.,
made from this simple scenario. The distributions of zircons must be older than the youngest measur- 2018; Weber et al., 2020; Melnik et al., 2021). All
ages calculated assuming linear or nonlinear (expo- able age in plutonic zircon. Additionally, some of models show that the first injections cool rapidly
nential) relationships between T and the amount of the injected magma continuously cools below the and a certain time is required to build up a partially
zircon crystallized differ, with linear models slightly solidus during assembly of a magmatic reservoir molten system. Yet, the details of the different mod-
overrepresenting younger zircon crystals compared (e.g., Annen, 2009; Annen et al., 2006, 2015; Karakas els vary, and therefore results of three numerical
to the nonlinear (exponential) model (Fig. 9B). In the et al., 2017), and zircon locked within the solidified codes for two end-member cases are compared.
absence of dynamic processes such as convection, portion of the magmatic system is less likely to be One model that has been successfully applied
the age distributions in the different portions of a sampled by an eruption (e.g., Shane et al., 2012). for the forward modeling of zircon age spectra is
solidified intrusion would be different, and the age Finally, zircon ages determined by bulk analysis the finite elements (FE) approach of Caricchi et al.
of the youngest measurable zircon would decrease techniques such as TIMS are average ages that (2014, 2016), in which thermal diffusion was solved
from the outer to the inner portions of the intrusion record integrated down-temperature zircon crys- using an FE approach with periodic sill insertion
(Fig. 9B). If, hypothetically, the different shells were tallization from a zircon-saturated melt that may and passive tracers recording the temperature
dynamically mixed prior to eruption (e.g., due to have lasted longer than analytical uncertainties evolution over time for different portions of the
recharge and remobilization; Bachmann and Ber- indicate (Table 1). High spatial resolution ages from magmatic system. A subsequent version of this
gantz, 2006; Huber et al., 2009), a volcanic rock could LA-ICP-MS and SIMS analyses, by contrast, permit model used the finite difference (FD) method to
contain a mix of all ages and different compositions the resolution of different stages of the magma’s discretize the diffusion equation, whereby the tem-
(e.g., Pamukçu et al., 2022). thermal history at 103–104 year intervals (Table 1). perature evolution is recorded in each numerical
Considering a geologically more plausible sys- The example introduced earlier in this section of cell that contains magma above the solidus (Weber
tem assembled by the injection of discrete pulses a single batch of magma cooling in a thermally et al., 2020). These Geneva models track the ther-
(e.g., Coleman et al., 2004), a zircon age distribution uniform crustal environment is of course unrealis- mal evolution of a magmatic system assembled by
that is noticeably different from the previous sce- tically simplified, as long-lived magmatic systems the injection of sills (Fig. 11A), which was previously
nario must be expected (Fig. 10). Importantly, the are sustained by frequent magmatic recharge. shown to be chiefly controlled by the thermal state
calculated distribution of measurable zircon ages Thermal models for such systems that are open of the crust and the vertical rate of magma accu-
would also look different if instead of considering for recharge exist, but different approaches are mulation (Annen et al., 2015; Caricchi et al., 2014).
the time-temperature paths of different portions of rarely vetted against one another. Two of these Variations of the mode of magmatic injection and
the intrusion, we were to only track the evolution models that are in use at the University of Geneva geometry of the magmatic system (either randomly
of the average temperature. This implies that zir- and UCLA-HD are therefore described here in detail in a prescribed volume, or by under-accretion, in
con age distributions calculated using an “average” to assess their similarities and differences regard- which sills are stacked in a downward growth direc-
time-temperature path cannot be used to retrieve ing the inferred magmatic thermal history and to tion) can be modified. These parameters only play
reliable information about the thermal evolution of compare their predictions for zircon age spectra, a minor role in the long-term thermal evolution
magmatic systems assembled by episodic magma primarily in volcanic settings. of magmatic systems and therefore in the calcu-
injection (Figs. 10A and 10B). lated zircon age distributions, but they can produce
The type of sample under investigation is also strongly divergent thermal histories during the
relevant, as zircon age distributions from plutonic 3.2 Thermal Models for Magmatic Systems early stages of a system (Annen et al., 2006, 2015;
or volcanic rocks will not have the same character- Undergoing Repeated Recharge Carrigan, 1988; Karakas et al., 2017).
istics. For instance, if an eruption occurs during the In the Geneva family of models (Fig. 11A), an
assembly of a crustal magma reservoir, zircon sam- There has been a long tradition of modeling the axisymmetric geometry is assumed, magma is
pled by the eruption will record a shorter age range thermal evolution of magmatic systems following injected at a set temperature (e.g., 1000 °C), and

GEOSPHERE | Volume 19 | Number X

Tracers (magma pulse)

A measurable ages 6 5 4 3 2 1
T0 magma
Non linear all tracers 12 7 4 1
Non linear Tav 12 7 4 1
Tzr

Temperature
Tav

Ter
TS

B Time Figure 10. Schematic representation of temporal

evolution of temperature and measurable zircon
1 1 1 1 1 1 1 2 1 5 1 6 4 12 4 16 7 24 7 23 7 23 7 16 7 13 12 12 crystallization ages (same prescribed number as
in Fig. 9) within a spherical magmatic reservoir
growing through input of successive discrete
Plutonic zircon

n. measurable ages
magma pulses. (A) Gray curves show evolution
of temperature for magma within each succes-
sive injected magma pulse. Black line shows
evolution of average temperature within the
growing magmatic reservoir. In this case, we
consider only the realistic case of exponentially
decaying zircon crystallization with decreasing
temperature. (B) Schematic distributions of
measurable zircon ages resulting from thermal
evolution of the reservoir. Note the difference
in the distribution if the temperature evolution
of each pulse (red) is considered rather than
only using the average temperature of the in-
trusion (black). (C) Schematic distributions of
Zircon ages measurable zircon ages for magmatic reservoir

Eruption
portion with temperature sufficiently elevated
C for the magma to be eruptible. This would be the
11 19 12 distribution of zircon age observed in volcanic
products at the eruption.

Volcanic zircon

n. measurable ages
Zircon ages

GEOSPHERE | Volume 19 | Number X

Sill underaccretion model Recharge rate Recharge into standing melt

Figure 11. Example of thermal structure of magma
9.1×10 km /yr/km (e.g., Spera et al., 2001)
-6 3 2
(e.g., Annen et al., 2005) emplaced into shallow (depth of intrusion is 7 km)
1000
Geneva MTK UCLA-HD MTK crust in (A) sill under-accretion (“Geneva” model) and
0 875 (B) recharge into standing melt in the center of an
2 A B 750
outward-growing ellipsoidal igneous body with horizon-
tal and vertical axes scaling 4:1 (University of California
4
Los Angeles-Heidelberg University [UCLA-HD] model).

Temperature (°C)
6 625 The magmatic recharge flux in both cases was set to 9.1
Depth (km)

8 × 10−6 km3/yr/km2, with magmatic recharge temperature

Tzrn Sat 500
(1000 °C) and initial geotherm (30 °C/km) identical in
10 1100‒0 ka
375 both models. Axisymmetric geometry was assumed
12 1100‒0 ka 3145 km3
Tsolidus with fixed bottom-temperature conditions in the Geneva
3145 km3
14 250 model, whereas the UCLA-HD model kept the bottom
Tsolidus
16 thermal flux constant. The intrusion accumulated over
125 1.5 m.y. in part A, but during an initial (here ~400 k.y.) in-
18 1500‒1100 ka cubation time (Annen et al., 2015), crustal temperatures
1144 km3 0
20 were too low to maintain a steady presence of magma,
0 10 20 30 0 10 20 30 40 and hence the contribution of zircon that crystallized
Width (km) during incubation is negligible. Thus, the model in part
B started later, with the incubation duration in part A
0 being subtracted from the total run duration, so that
the resulting zircon age spectra can be directly com-
2 C D pared, and with an initial magmatic volume of ~10%
4 of the final intrusive volume being prescribed. Thermal
6 structure at the end of the simulation (corresponding
to a zero-age in Fig. 12) is shown, along with the extent
8
Depth (km)

of total intrusive volume (red dashed line), and parts of

10 the intrusion above the solidus (black solid line) and the
12 zircon saturation temperature (black dashed line). The
14 left sides of parts A and B show the results of the Ge-
MTK center MTK center
neva UCLA-HD software, whereas the right sides show
16 Geneva center UCLA-HD center
the results of the MagmaThermoKinematics.jl (MTK)
MTK side MTK side
18 software (see Appendix). (C and D) Temperature profiles
Geneva side UCLA-HD side
20 are shown at the sides and within the centers of the
0 200 400 600 800 0 200 400 600 800 1000 models based on published studies and newly devel-
oped MTK algorithms.
Temperature (°C)

the magma crystallinity (xf) increases with decreas- (t-T) trajectories of the system are then derived and zrm is the maximum fraction of zircon crystal-
ing temperature (T in °C), as: from a large number (e.g., 1000) of passive trac- lizing between zircon saturation (Tzr) and the solidus
ers sampling temperature at constant time intervals temperature (Tsol). At the time when the zircon pop-
1
xf = 1 , (1) between the onset and the end of magmatic injec- ulation is sampled (e.g., in a volcanic eruption), only
1+ e
tions. Each tracer can produce an arbitrary number tracers with a temperature higher than a minimum
where of measurable ages (e.g., 100), and the rate of temperature (Tmin) are considered. Tmin can represent
zircon crystallization decreases with decreasing either the minimum temperature at which magma
=
( 800 T)
. (2) temperature following the derivative of Equation (and zircon therein) can erupt, or the temperature of
23
3 for T (Tierney et al., 2016): the interstitial melt extracted (together with zircon)
Alternatively, the relation between T and xf was 104 to feed a volcanic eruption. The tracer that spends
estimated from a fourth and fifth order polynomial Fzr = 1.62 1.8 10 4 e T
zrm, (3) the longest period within the temperature interval
fit to the experimental melt fraction tempera- of zircon saturation provides the maximum possible
ture curves of Blatter and Carmichael (2001) and where the cumulative fraction of zircon (Fzr) for duration of continuous zircon crystallization. Once
Marxer and Ulmer (2019). The time-temperature each unit of magma increases with decreasing T, this is determined, all measurable ages between the

GEOSPHERE | Volume 19 | Number X

oldest age and the time of eruption are summed as fraction of zircon growth is then calculated using Overall, both model geometries define plausi-
schematically shown in Figure 10. Equation 3. This results in a time-resolved zircon ble end-members of multi-cyclic intrusions, where
The second type of model (UCLA-HD; Friedrichs growth curve for each cell, and each curve is then the Geneva sill geometry leads to initially more
et al., 2021; Tierney et al., 2016) starts with a two- summed at the time of eruption. Only cells that rapid heat dissipation (especially when the depth of
dimensional grid of 0.1 × 0.1 km cells over a 20 × are above 700 °C (super-solidus) are considered sill emplacement is random, rather than vertically
60 km (width × depth) rectangle that represents a and weighted according to their relative volumes stacked) compared to the UCLA-HD model, where
mid–upper crustal block. Parameters for modeling when translating the 2-D grid into 3-D. These zircon the ellipsoid has a smaller surface-to-volume ratio
the thermal and compositional evolution of magma growth curves distinguish between core (e.g., at than a thin cylindrical sill, and recharge always
and its surroundings within this grid are based on 20% of each curve) or rim ages (e.g., at 80%; Frie- occurs in the hottest part of the system. Crustal
equations for recharge-assimilation–fractional drichs et al., 2021). An average total zircon age can thermal conditions prior to intrusion are defined in
crystallization (RAFC; Spera and Bohrson, 2001). also be obtained from the integration of the entire both models as constant mantle and surface heat
Geometrically, the UCLA-HD model differs from zircon growth curve. flow. Country-rock partial melting and assimilation
the Geneva models in that instead of vertically A limitation of existing numerical modeling stud- (Spera and Bohrson, 2001), as well as radioactive
stacking sills from a prescribed depth and typically ies that simulate the thermal evolution of magmatic heating, are irrelevant for comparatively brief
displacing the crustal rocks downward (e.g., Weber systems is that the source code is typically not read- model-run durations and the shallow depth of sub-
et al., 2020), magmatic injection always occurs in ily accessible in publications, and the assumptions volcanic igneous bodies and their low country-rock
the center of an ellipsoidal volume of magma, i.e., behind the models often vary according to a par- temperatures.
the hottest and therefore likely most long-lived ticular geological setting, which makes comparison To understand how differences in the choice
standing volume of magma in the intrusive body of the different studies difficult. Moreover, numer- of material parameters and boundary conditions
(Fig. 11B). Upon growth due to periodic magmatic ical implementation details, such as how to deal affect the thermal structure, we systematically var-
recharge, the body maintains its originally defined with nonlinearities, are beyond the limits of typi- ied many of the parameters for the two scenarios.
aspect ratio. The thermal evolution of individual cal publications, which limits reproducing results. Results show that the central intrusion setup is very
cells is monitored throughout the simulation as Therefore, we developed a new Julia package to robust with respect to variations in the material
they migrate away from the center of magmatic simulate the thermal evolution of magmatic sys- parameters, and that the sill under-accretion case
recharge and are translated into a pseudo-3–D dis- tems (MagmaThermoKinematics.jl, or briefly MTK). has larger variations (Fig. S1; see footnote 1). This
tribution assuming rotational symmetry around the The new package is open source and works in 2-D, difference is mostly likely caused by the sill under-
vertical central axis of the magmatic body in the 2-D axisymmetric, and 3-D geometries. It utilizes accretion case not having reached steady-state yet,
center of the grid. modern hardware such as graphics processing units even over 1.5 m.y. run durations, whereas this is
The crystal fraction, xf, is calculated some- (GPUs) and multicore processors and provides a the case for the central intrusion case.
what differently from that of the Geneva model simple yet extendable way to change mate-
(Caricchi et al., 2014, 2016) by using the following rial parameters (see the Appendix for a detailed
relationship: description). Several dike injection algorithms are 3.3 Describing Calculated and Measured Age
implemented, including one that employs an ana- Distributions
( ) ( )
2
x f = Tliq T / Tliq Tsol , (4) lytical solution to insert coin-shaped cracks into an
elastic half-space, as well as central injection and sill In principle, knowing the thermal conditions
where Tliq and Tsol are the liquidus and solidus tem- under-accretion. Tracers are injected within newly in an intrusion at any time allows calculation of
peratures of the recharge magma, respectively. added dikes and record the subsequent thermal the number of possible zircon crystallization ages
The temperature of magmatic recharge is selected evolution from which zircon age distribution can potentially available for measurement. The cal-
based on reasonable geological constraints (e.g., be calculated. We have benchmarked this third culated number of measurable ages can then be
based on thermometry of co-erupted mafic rocks; modeling approach versus the two end-member converted into a PDP (normal or cumulative) for
Friedrichs et al., 2021). At each time-step during scenarios shown in Fig. 11, which demonstrate that comparison with the distribution of measured ages
the model run, the temperature change of each differences in the thermal structure are generally (Fig. 12). Alternatively, a simple parameter, such as
cell is monitored: For each cell undergoing cooling <5 °C at the end of the simulation (see Fig. S1 in the the age difference (Δt) between stated percentiles
within the saturation range (e.g., 800–700 °C), the Supplemental Material1). of the modeled zircon age distribution, can be used

Supplemental Material. Figure S1. Results of simulations in which we systematically changed a single model parameter, starting from the reference simulations shown in Figure 11. Please visit
1

https://doi.org/10.1130/GEOS.S.22898453 to access the supplemental material, and contact [email protected] with any questions.

GEOSPHERE | Volume 19 | Number X

7.5×10 km /yr/km 9.1×10 km /yr/km 1.1×10 km /yr/km

-6 3 2 -6 3 2 -5 3 2
Figure 12. Comparison of cumulative

Zircon age cumulative probability %

100 probability density plots (PDPs) of (A–C)

84
A B C 84% zircon ages and (D–F) average magmatic
temperature (all domains in the igne-
ous body with temperatures >700 °C)
modeled for a magmatic body gener-
ated by frequent magmatic recharge in
sill geometry (blue lines; Geneva model),
and injection into the center of an ellip-
50 soid (red lines; University of California
Los Angeles-Heidelberg University
[UCLA-HD] model) with recharge fluxes
of 7.5 × 10–6, 9.1 × 10–6, and 1.1 × 10–5 km3/
Δt = 110 ka Δt = 170 ka Δt = 205 ka yr/km2. Runtime for the Geneva models
Δt = 353 ka Δt = 491 ka Δt = 559 ka was always 1.5 m.y.; corresponding du-
16 rations for the UCLA-HD models were
16% shortened to (A–C) 0.7 m.y., 1.1 m.y.,
0 and 1.3 m.y., respectively, to exclude
D E F the incubation time required to estab-
sill central lish continuous melt presence in the
900 underaccretion injection Geneva models (see Fig. 11). Initial mag-
Taverage magma (°C)

matic volumes in the UCLA-HD models

Geneva UCLA-HD
were 10% of the final volumes, which
MTK MTK
are 1641 km3, 3145 km3, and 6100 km3
in parts A, B, and C, respectively. Re-
800
sults from MagmaThermoKinematics.jl
(MTK) models using both geometries
and methods to compute zircon ages
700 yielded equivalent results, which under-
scores the robustness of the different
<1% <1% <1% numerical approaches. Due to differ-
magma magma magma ent geometries, Δt and final average
600 temperatures differ, but the shapes of
0 500 1000 0 500 1000 0 500 1000 1500 model zircon age distributions are gen-
erally similar.
Age (ka)

for comparison (e.g., using the 16–84 percentile to compare the modeled and natural distribution of (Table 1). To demonstrate the potential differences
difference, corresponding to ± 1 s.d., although the zircon ages measured with ID-TIMS or high spatial between these types of data in an idealized case, we
resulting distributions in both types of models are resolution techniques. simulate a pool of 200 zircon crystals (assumed to
strongly skewed toward younger ages; Fig. 12). In be spherical for simplicity) spanning an age range
the model-scenario examples (Fig. 12), Δt system- of 3 m.y., with radii between 50 µm and 100 µm, and
atically increases with the duration of continuous 3.4 Measurement Bias in Age Distributions with the thickness of the different zones selected
melt presence and recharge rate. The input param- randomly but arranged to decrease from the crys-
eters of the model that best match the measured Before discussing applications of these models, tal center toward the rim (a selection is shown in
distribution of zircon ages can then provide a we first focus on the analysis of zircon collected in Fig. 13A). The U content of the different zones and
quantitative estimate of processes operating in a plutons and discuss the modifications to apply to the variation of the zircon zone thickness from core
long-lived crustal intrusive body, such as the rate the modeling when dealing with zircon from vol- to rim affect the distributions of ages obtained from
of magma recharge. However, this would be true canic products. The properties of a natural zircon whole zircon versus high spatial resolution tech-
only if dating could be performed in each zone for age distribution are captured by common analytical niques (Curry et al., 2021). In this simulation, each
a large number of zircon crystals. Instead, actual techniques such as ID-TIMS, LA-ICP-MS, and SIMS, zircon comprises 2–10 zones of decreasing age from
data sets are necessarily limited in number and which afford age resolution to different degrees core to rim and only zircons with 10 zones have a
resolution, and it is thus important to consider how depending on their required sampling volumes core with the oldest age (i.e., 3 Ma). During the

GEOSPHERE | Volume 19 | Number X

ID-TIMS would be obtained using ID-TIMS is significantly

0.25
B Density ID-TIMS ages shorter than the “real” duration of the magmatic

Volume fraction zones of different ages

all (n.=200) event during which zircon was crystallizing (Fig. 13;
bootstrapped (n.=40)

0.20
Glazner and Sadler, 2016). However, it is interest-
bw=0.05 Ma
ing that the median and mode for the calculated
ID-TIMS analysis are relatively close to the real val-

0.15
ues, whereas the standard deviation is generally
underestimated by the subsamples (Fig. 13D). Thus,
for ID-TIMS age determinations, mode and median

0.10
provide relatively reliable metrics for comparison of
"real" distribution modeling results, whereas the true standard devi-

0.05
ation of ages is underpredicted. This test implies
that using ID-TIMS ages in combination with the
Spot size (high spatial resolution) = 30 μm model proposed by Caricchi et al. (2014, 2016) is

0.00
100 μm appropriate and produces relatively good estimates
0 0.5 1 1.5 2 2.5 3 for the rate of magmatic input into the system but
Apparent age [Ma]
underestimates the volume of the system.
LA-ICP-MS or SIMS Using the same simulated population of zircon

C D crystals, we calculated the distribution of ages

1.1
Mode [Ma]
that would be measured by placing 1100 analy-
sis spots of a high spatial resolution technique

0.8
(LA-ICP-MS or SIMS) with a beam diameter of

0.5
30 µm (Fig. 13A). We assume that crystals are ana-
0.6 0.7 0.8 0.9 1 lyzed in cross-sectioned mid planes, disregarding
Median [Ma]
Density

mixing of crystal domains at different depths for

simplicity. High spatial resolution analyses better

0.6
St. dev. [Ma]
capture the total duration of zircon crystallization,

0.3
but because of lateral mixing of growth domains,
they provide biased estimates for the mode and
median of the actual distribution of zircon ages

0
0.6 0.7 0.8 0.9 1
bw = 0.15 Ma (Fig. 13D). Therefore, the comparison between
Median [Ma]
the calculated distribution of zircon ages and
“real” distribution ID-TIMS (n. 200) LA-ICP-MS or SIMS
0 0.5 1 1.5 2 2.5 3 ID-TIMS (n. 40) (n. zircons=200) measurements performed using high spatial res-
Apparent age (Ma) (n. spots=1100)
olution techniques should focus on the total spread
of zircon ages (e.g., as described by the standard
Figure 13. (A) Simulated, randomly assigned zircon age zonations, and resulting age distributions for methods sampling
different volumes of zircon (B: isotope dilution–thermal ionization mass spectrometry [ID-TIMS], C: laser ablation– deviation) rather than on the mean or median. The
inductively coupled plasma–mass spectrometry [LA-ICP-MS] or secondary ionization mass spectrometry [SIMS]; bw = exceptions are analyses of unsectioned crystals
sampling bandwidth). Part D plots median age versus mode and standard deviation for the “real” distribution (red open acquired in depth-profiling mode (e.g., Schmitt
circle) and different samplings using multiple spots on various types of crystal. and Vazquez, 2017), in which the ion beam ideally
samples individual growth domains with minimal
prescribed 3 m.y. of magmatic history that zircon section 3.3 (Fig. 13B). We then use this simulated mixing as the depth resolution of the ion beam
could potentially record, the mass fraction of zir- zircon production curve to calculate the distribution (<1 µm) is small relative to the dimensions of
con crystallized increases as the amount of magma of zircon ages that would be obtained by perform- the growth zones. Surface analysis of individual
capable of crystallizing zircon grows, but then drops ing ID-TIMS analyses by summing the product of and unsectioned zircon crystals also allows age
toward the end of the interval to mimic the wan- the mass fraction of each sector and the respective determination of the last growth episodes largely
ing stages of the magma system, consistent with age of the sector (Fig. 13B). Not surprisingly, this without mixing different growth zones, and thus
the expected zircon age distributions introduced in calculation shows that the range of zircon ages that not only the total spread can be modeled, but also

GEOSPHERE | Volume 19 | Number X

the shape of the probability density function (i.e., the average magmatic temperatures are always (Tierney et al., 2016). These exemplary estimates
the relative proportions of newly crystallized zircon lower for the Geneva model than for the UCLA- significantly exceed the commonly inferred intru-
relative to antecrystic grains). HD model, with temperatures gradually increasing sive to extrusive ratio of 10:1 (Crisp, 1984). The
and decreasing with the run duration, respectively intrusive to extrusive ratio also changes as a func-
(Fig. 12). This is expectedly matched by the longer tion of time and tends to decrease in thermally
3.5 Comparison of Different Thermochemical duration of zircon crystallization (measured as the mature volcanic systems (Giordano and Caricchi,
Models for Zircon Crystallization in Magmatic difference between the 16th and 84th percentile in the 2022, and references therein).
Systems Evolving through Recharge cumulative zircon age distribution) in the overall Zircon growth models also quantitatively pre-
cooler Geneva models relative to the hotter UCLA- dict volumes of eruptible magma in the subsurface
Although direct comparability between the cur- HD models at the modeled magmatic recharge that can be directly compared with those estimated
rently applied models as summarized in section rates of 7.5 × 10−6 to 1.1 × 10−5 km3/yr/km2 (Fig. 12). from other petrological constraints (e.g., Weber et
3.2 is complicated by the fundamental differences Despite these differences, the overall shape of the al., 2020; Friedrichs et al., 2021) or geophysical
in how they are conceptualized and coded, we cumulative probability distribution of expected zir- observables (e.g., Lukács et al., 2021). To illustrate
can demonstrate general similarities between the con ages is similar, which implies that under the how zircon age spectra and their quantitative mod-
predicted thermal histories for magmatic systems selected conditions, sampling of younger zircon is eling can reveal fundamental differences between
open for recharge over typical lifetimes of volca- more likely than sampling of older crystals due to the underlying magma systems, we compare data
nic systems (e.g., 1–1.5 m.y.; Weber et al., 2020). progressive growth of the zircon-saturated ther- from four volcanoes recently studied (Fig. 14). Here,
To mitigate model-dependent bias, we calculated mal layer. this can only be accomplished in an abridged form,
zircon crystallization ages for both models using and we refer to the cited publications for more
identical boundary conditions. For this, conditions detail (Weber et al., 2020; Friedrichs et al., 2021;
typical of a long-lived arc volcano were selected, 3.6 Implications of Zircon Age Spectra Lukács et al., 2021).
which include a normal initial geothermal gradi- Modeling One key outcome of these studies is that poten-
ent of 30 °C and a magma recharge temperature tially eruptible volumes of magma can be significant
of 1000 °C (Weber et al., 2020). Furthermore, both Despite conceptual and computational differ- in systems that are currently dormant, or even con-
types of models were set up to produce the same ences in the setup of thermochemical models for sidered extinct by conventional standards, but also
total magmatic volumes with an approximately zircon crystallization in evolved magmatic sys- that volumes of potentially eruptible magma vary
identical depth range of the accumulated intrusion tems fed and maintained by a repetitive influx significantly in different systems. This is import-
after the end of each run (Fig. 12). A corollary of the of more primitive magma, they provide broadly ant in assessing mid- to long-term hazards from
cylindrical sill geometry with downward displace- similar results. This can yield unique quantitative volcanoes (e.g., Lukács et al., 2021; Weber et al.,
ment is that the magma always comes into contact constraints on long-term magmatic fluxes and 2020). The estimates for the total amount of erupt-
with comparatively cold crust, and the magma will magmatic crustal accretion in a specific tectonic ible magma (at <50% crystal content) derived from
thus be more affected by rapid initial cooling than environment. In contrast to geophysical observa- modeling zircon age spectra are comparatively
the outward growing ellipsoid, as earlier intruded tions that monitor present-day conditions, modeling small (albeit non-negligible) for the Ciomadul dome
and hot magma/rock shields the interior of the the thermal evolution of magmatic chambers as field of Romania at ~5–10 km3 (Lukács et al., 2021),
ellipsoid from heat loss to cold country rocks. To tracked by their zircon inventory can reveal long- whereas for Nevado de Toluca, the total amount of
compensate for a protracted initial incubation term variations in magmatic input in long-lived eruptible magma is about one order of magnitude
time in the Geneva models during which zircon systems, and it can also be applied to fossil mag- larger (Figs. 14A and 14B; Weber et al., 2020). This
productivity is negligible, the run duration of the matic systems in the form of plutons. agrees with the overall differences in cumulative
UCLA-HD models was shortened to only match the One important insight gained from the model- eruptive volumes (8–10 km3 for Ciomadul versus
zircon-producing duration of the Geneva models ing of zircon age spectra is that substantial intrusive ~60 km3 for Nevado de Toluca), and eruptive fre-
(as defined by the presence of ~1% of final mag- volumes accumulate under arc volcanoes. Hence, quency (e.g., last major eruption at ca. 30 ka for
matic volume). The volume of the initial intrusion intrusive to extrusive ratios determined by this Ciomadul versus ca. 14 ka for Nevado de Toluca).
that becomes inflated by continuous recharge (in method for volcanic arcs range from 25:1 to 50:1 Zircon age spectra can also indicate variations
pulses every 5000 years) was prescribed as ~10% for Nevado de Toluca, a stratocone in the Trans- in magmatic recharge fluxes over time (Friedrichs
of the final intrusive volume in the UCLA-HD model, Mexican Volcanic Belt (Weber et al., 2020), to et al., 2021; Angeles-De La Torre et al., 2023). This
and the aspect ratio of the ellipsoid was held con- 75:1 for lava domes in the Altiplano-Puna region, is the case for two adjacent stratovolcanoes in
stant (horizontal:vertical axes = 4:1). As a result, which erupted after an ignimbrite flare-up episode Central Anatolia (Mt. Hasan and Mt. Erciyes) that

GEOSPHERE | Volume 19 | Number X

Model setup: Geneva UCLA-HD

10-5 10-5 Figure 14. Zircon age spectra as ranked order plots plotted
with model zircon age curves as cumulative probability
10-6 10-6
km /ka/km
3 2
km /ka/km
3 2 density plots based on different intrusion geometries
10-7 10-7 in the (A, B) Geneva and (C, D) University of California
Los Angeles-Heidelberg University (UCLA-HD) models.
30 A 80
C (A) Ciomadul dome complex representing a small, dor-
mant system, for which U-Pb zircon interior ages from
25 the youngest eruptions (Phase 5 is between ca. 50 ka and
60 30 ka; Lukács et al., 2021) are plotted; (B) Nevado de Toluca,
20 a large stratocone that is currently dormant but has experi-
enced multiple large-volume explosive eruptions in the late
15
40 Pleistocene, with U-Th and U-Pb interior analyses shown
for the 12.5 ka Upper Toluca Pumice (UTP) eruption (Weber

Relative probability
10
et al., 2020); (C) Mt. Hasan; and (D) Mt. Erciyes, both strato-
Ranked order

20 volcanoes in central Anatolia (Friedrichs et al., 2021); data

5 Ciomadul Phase 5 Mt. Erciyes ca. 9 ka sets comprise U-Th and U-Pb interior ages with xenocrystic
zircons from Mt. Erciyes (with the youngest ages down
0 0
10-5 to ca. 2 Ma) omitted. Mt. Hasan has an eruptive history
10-5
of steady andesitic lava flow effusion with a ca. 5–15 k.y.
10-6 10-6
km /ka/km
3 2 km3/ka/km2 recurrence interval in the late Pleistocene, whereas Mt. Erci-
10-7 10-7 yes experienced its penultimate eruptions at ca. 105–85 ka
followed by an eruptive hiatus before an early Holocene
80
50 B D reawakening, when a series of peripheral rhyolitic domes
erupted (Friedrichs et al., 2021, and references therein); only
data from the most recent eruptions of both volcanoes are
40 60
plotted. Curves above the ranked order plots show (A, B)
constant or (C, D) variable recharge rates used in different
30 models. Models in A and B were run between 1.5 Ma and
40 2 Ma and 0 ka (eruption age), with an incubation time
20 during which no zircon was produced (see text); eruption
ages were subtracted from measured zircon ages. Model
20 curves in parts C and D were calculated from 800 ka (Mt.
10
Nevado de Toluca UTP Mt. Hasan ≤20 ka Erciyes) and 550 ka (Mt. Hasan) to the respective eruption
ages; measured ages are shown. For ease of comparison,
0 0 all areal rates were calculated here for a magmatic system
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700 circumscribed with a 10 km radius; for specific areal extent
of volcanic edifices, see original studies.
Zircon age (ka)

show different styles of continuous versus pulsed intermediate (andesitic) composition for Mt. Hasan As these examples illustrate, interpretation
recharge throughout the late Pleistocene (Figs. 14C and initially highly explosive eruption bursts of of zircon age spectra should ideally integrate as
and 14D; Friedrichs et al., 2021). One important evolved (rhyolitic) magmas from multiple vents much geological information as possible, including
aspect of the study by Friedrichs et al. (2021) is for Mt. Erciyes (Friedrichs et al., 2021). Zircon age the eruptive pre-history of a volcanic system. The
that zircon crystals were separated and analyzed spectra modeling suggests that the active mag- onset of magmatic input into the system is espe-
following identical protocols for samples from matic system underneath Mt. Hasan is larger and cially important for refining estimates of the rates
both volcanoes so that the observed differences more frequently rejuvenated by mafic magmatic of magmatic input at depth. This is because the
in the age spectra are robust. Moreover, different input, whereas that of Mt. Erciyes has shrunk to dominant factor controlling the thermal evolution
patterns in long-term recharge rates, steady and contain mostly residual evolved melts in what of the magmatic system is the total amount of sen-
frequent for Mt. Hasan versus fluctuating and infre- may represent the waning stages of its magmatic sible heat added into the crust, which is the product
quent for Mt. Erciyes, correlate with the eruptive system, despite the overall larger volcanic edifice of the rate of magmatic input and time. The onset
frequency and type for each volcano, which are volume compared to that of Mt. Hasan (Friedrichs of magmatic input can either be obtained from
characterized by frequently recurring lava flows of et al., 2021). the oldest (non-xenocrystic) zircon or by dating

GEOSPHERE | Volume 19 | Number X

the first eruption of a volcanic system, e.g., with saturation models for a single batch of melt under- δ18O and εHf) that can be used alongside and in sup-
40
Ar/39Ar or (U-Th)/He chronometers. On one hand, going monotonic cooling (e.g., Harrison et al., 2007). port of the age data. Lastly, the understanding of
both approaches will likely underestimate the dura- The models described in section 3.2 suggest zircon the significance of zircon ages is limited by the fact
tion of magmatic input into the plumbing system, distributions for long-lived silicic magma systems that, in most studies, textural relationships of zir-
as zircon only starts crystallizing once sufficiently with younger zircon predominating over older con to other mineral phases are unknown because
evolved melts are generated, and products of older zircon, the opposite of the distribution expected zircon is extracted from its matrix by physical and
eruptions may be covered or destroyed. On the for cooling of a single-magma batch (Figs. 9 and chemical methods. Therefore, it remains ambigu-
other hand, it is sometimes difficult to distinguish 10). This implies that it is not straightforward to ous whether an individual zircon crystal stopped
between antecrysts and xenocrysts (Miller et al., prescribe an a priori age distribution to zircon age crystallizing because it became entombed in major
2007), which may lead to an overestimation of the spectra to extract eruption ages from zircon data. minerals and therefore was shielded from the melt,
onset of magmatism. Thus, extrapolating from zircon ages to a precise or because of subsolidus cooling and complete
The thermal models presented in section 3.2 eruption age (e.g., Keller et al., 2018) requires firm solidification of the magma.
illustrate how for similar zircon age distributions, knowledge of the pre-eruptive zircon crystallization Although zircon saturation is well constrained,
the average temperature can change significantly history of a particular magmatic system. Because zircon can crystallize before whole-sale satura-
when different geometries of magmatic input into zircon ages track the pre-eruptive accumulation tion is reached when melt composition and Zr
the plumbing system are considered (Figs. 12D– of magma, they are an important complement to abundance become locally conducive for zircon
12F). Such differences should be reflected in eruption ages for unraveling potential leads and crystallization, e.g., when zircon crystallizes next
different abundances and temporal trends in the lags regarding climatic variability (e.g., sea-level to fast-crystallizing oxides (Bacon, 1989). A related
chemistry of volcanic products, as well as the trace- changes and glaciation–deglaciation) and associ- critical aspect is that evolved (per-)alkaline sys-
element characteristics of zircon. Hence, knowing ated lithospheric stresses that may influence the tem compositions currently defy modeling zircon
the temporal and chemical evolution of volcanic productivity of magma, recharge fluxes, and erup- crystallization because experimental constraints
products and zircon from eruptions spanning the tive frequency (e.g., Satow et al., 2021). for zircon saturation in such melts are scarce (cf.
widest possible time interval provides additional Gervasoni et al., 2016). Empirically, zircon crystal-
constraints for identifying the most appropriate lizes late in such systems due to the high solubility
thermal model. This, in turn, allows the retrieval 3.7 Future Developments for a Better of Zr and Hf in alkaline melts (Linnen and Keppler,
of more accurate and precise estimates of the Interpretation of Zircon Age Spectra 2002), but then crystallization from supersaturated,
average rates of magmatic input into the volcanic volatile-rich, and low-viscosity melts can produce
plumbing system. The same approach could be To further advance zircon as a magmatic ther- large amounts of zircon, often much larger in grain
applied to exposed crustal sections by focusing mochronometer, improvements in analytical size than those that crystallized from metaluminous
on plutons spanning the widest possible range of methodology and more realistic models for zir- or peraluminous melts. The processes of zircon
emplacement ages. Studies that complemented con crystallization, residence, and mobilization in saturation and growth in such melts, along with the
zircon age spectra with thermochemical indicators, magmas are required. Because zircon age spectra formation of hydrothermal zircon (e.g., Troch et al.,
such as trace element data, have demonstrated this based on a few analyses may not reliably repre- 2018), should receive more attention in the future.
potential (e.g., Ti-in-zircon; Samperton et al., 2017; sent the underlying distribution, large-n data sets
Kent and Cooper, 2018; Weber et al., 2020; Fried- (n > 100) will potentially provide additional infor-
richs et al., 2021). mation, in line with the approach advocated for 3.8 Unifying Reactive Two-Phase Thermo-
Models such as those shown in Figures 11 and state-of-the-art detrital zircon studies (Pullen et al., Mechanical Models of Magma Transfer
12 indicate that protracted zircon crystallization in 2014). Moreover, there are analytical limitations
a long-lived system open for recharge results in a in resolving the true duration of zircon crystalli- The thermal models presented in section 3.2
distribution of zircon ages in volcanic rocks that zation (Table 1). To further utilize the potential to are severely restricted in realistically describing
differs significantly from that expected from zircon reconstruct thermal histories from individual zircon mixtures between crystals and melt where relative
saturation in a magmatic volume that (unrealisti- crystals, improved spatial resolution and precision proportions of solids and liquids are expected to
cally) cools without any internal thermal gradients. are essential. Zircon trace-element geochemistry fluctuate spatially and temporally over the lifetime
Zircon age spectra, successfully modeled in terms also furnishes complementary information such of a magma system. Re-mobilization of crystal-
of open-system magmatic recharge, thus may as temperature (Ti-in-zircon) and melt evolution rich marginal domains or convective overturn in a
have little in common with the simple exponential (e.g., indices for differentiation such as Zr/Hf and magma prior to volcanic eruptions are disregarded
decrease in zircon production predicted from zircon Eu/Eu*, as well as for crustal assimilation such as in these models, although both processes are likely

GEOSPHERE | Volume 19 | Number X

common in nature. Rounded and partially melted Although significant breakthroughs have been composition cannot predict the chemical evolution
plutonic enclaves are frequent ingredients of vol- achieved during the last decade, to date there are no of the magmatic chamber. Alternative methods,
canic deposits, which suggests that the reheating elasto-plastic thermo-mechanical models that are such as a pre-computed petrological database cov-
of largely solidified magmatic crusts can liberate able to account for volume change and phase reac- ering the expected compositional space (Rummel et
zircon from otherwise non-eruptible domains of a tions simultaneously. Instead, the second category al., 2020), or direct coupling with a Gibbs free energy
magmatic system (e.g., Schmitt et al., 2010). Fur- has focused on coupling two-phase flow models minimizer (Riel et al., 2019, 2022) is a step forward.
thermore, massive recharge events are expected to with phase reaction while disregarding thermo- The coupling of a physical model of melt extraction
cause remobilization, stirring, and homogenization mechanical deformation (e.g., Jackson et al., 2003, and crystal accumulation in a magma chamber to
over large distances in a long-lived crystal mush 2005; Annen et al., 2006; Bouilhol et al., 2011; Solano a petrological database (Rummel et al., 2020) or a
(e.g., Bachmann and Bergantz, 2006; Huber et al., et al., 2012; Jackson et al., 2018; Riel et al., 2019). Gibbs energy minimizer (e.g., Riel et al., 2019), while
2009). This seems to be a prerequisite for indistin- Here, several 1-D models of reactive two-phase challenging on its own, can allow the mechanical
guishable zircon age spectra at the hand-specimen flow have been successfully conducted, but only and chemical evolution of a magmatic reservoir to
scale for the same eruption despite these crystals a limited number of 2-D applications have investi- be modeled. With adequate zircon saturation mod-
originating in different parts of a magmatic reser- gated the formation of a reactive melt-filled network els or a thermodynamic database that includes Zr
voir and thus recording different thermal histories (e.g., Bouilhol et al., 2011), and all ignore the role (Kelsey and Powell, 2011), the growth and resorption
(e.g., Pamukçu et al., 2022). The frequency and of elasto-plastic thermo-mechanical deformation of zircon can be readily predicted. However, the slug-
intensity of these remobilization and overturning of the host-rock, including diking. Moreover, the gish dissolution kinetics of zircon in silicate melts
events in long-lived magmatic systems are empir- coupling with phase reactions is mainly achieved (e.g., Watson, 1996; Bindeman and Melnik, 2016) also
ically poorly constrained. using simplified parameterizations (e.g., Jackson play an essential role in controlling the available Zr
Creating a petrologically grounded thermo- et al., 2003, 2005, 2018; Solano et al., 2012), which pool for zircon crystallization. Therefore, the cou-
mechanical model applicable to magmatic systems provides useful insights into the overall model pling of thermo-mechanical models for magmatic
is one of the biggest challenges of geodynamic behavior but makes it difficult to directly compare chamber dynamics and petrological reactions also
modeling. Indeed, such a model has to integrate: the simulations with geochemical data. The third needs to account for non-equilibrium thermody-
(1) complex elasto-viscoplastic rheologies for the category aims to better understand the relationship namics. A possible solution could be to use stable
deforming host rock; (2) several modes of magmatic between magmatic chamber recharge and erup- phase equilibrium prediction as a directional step
transport/segregation, including mass convection, tion. Using a 0-D modeling approach, these models (e.g., Lasaga, 1986), while the length of that step,
pervasive flow within an interconnected solid net- consider up to three phases (solid, liquid, and gas), which corresponds to the reaction kinetics, could
work (low melt/rock ratio), settling of crystals within visco-elasto-plastic rheology of the surrounding be constrained by experimental melting studies
a magmatic body (high melt/rock ratio), and diking crust, parameterized thermodynamically consis- (e.g., Hou et al., 2020).
through the host-rock (high melt/rock ratio); (3) end- tent phase changes within the magma chamber,
member cases where either fully solidified rock or as well as pressure variations related to recharge
fully liquid magma is present; (4) extremely large and eruptions events (e.g., Degruyter and Huber, ■■ 4 CONCLUDING REMARKS
viscosity difference between magma and solid host; 2014; Townsend et al., 2019; Townsend and Huber,
(5) thermodynamically consistent phase stabilities 2020; Townsend, 2022; Liao et al., 2021). Although Zircon, although a tiny and seemingly neg-
and transitions, involving reaction kinetics, volume the latter category allows scaling relationships to be ligible component in rocks, is nonetheless an
change, and thermal stress; and (6) the coupling drawn based on eruption frequency and size (e.g., essential probe into magmatic systems concealed
of tectono-magmatic processes occurring over a Degruyter and Huber, 2014; Townsend et al., 2019; at depth. It is not only a time capsule, but also a
wide range of temporal and spatial scales in a self- Townsend and Huber, 2020; Townsend, 2022), it falls faithful recorder of the thermal and chemical con-
consistent and numerically stable way. short of describing the internal dynamics of a mag- ditions in magmas. For these reasons, zircon has
State-of-the-art approaches are currently divided matic chamber and has yet to be applied to higher become indispensable for studying differentiation,
into three main categories that tackle these prob- spatial dimensions. longevity, and pre-eruptive priming of magmatic
lems from different angles. The first category aims Melt segregation within magmatic chambers systems, where it adds a retrospective capabil-
to better couple thermo-mechanical deformation and cycling between magmatic recharge and ity that geophysical monitoring restricted to brief
with two-phase flow while disregarding or simpli- extraction in and out of a magmatic reservoir imply human timescales lacks. Although zircon age dis-
fying phase reactions (e.g., Keller et al., 2013, 2017; that compositions of crystals and melts are con- persion is seemingly a burden from the traditional
Keller and Katz, 2016; Turner et al., 2017; Rummel et stantly evolving. Hence, the use of phase equilibrium perspective of an event-oriented geochronology,
al., 2020; Keller and Suckale, 2019; Katz et al., 2022). parameterization or look-up-tables of fixed bulk-rock the increasing power to unravel the complexity of

GEOSPHERE | Volume 19 | Number X

accessory mineral crystallization in time and space To amend this situation, we have developed a new, open if using Equation (A2b) directly. It requires knowledge of the
source, software package—MagmaThermoKinematics.jl (https:// derivative of the melt fraction versus temperature, ∂ϕ ∕ ∂T, which
also bears tremendous potential. In this review, we github.com/boriskaus/MagmaThermoKinematics.jl) (MTK)— can be computed analytically for most melt parameterizations
emphasized the need for improved statistical tools to using Julia (Bezanson et al., 2017; https://julialang.org), which and is obviously zero below the solidus and above the liquidus.
adequately portray and interpret geochronological can be employed to simulate the thermal evolution of mag- Yet, an important aspect for stable numerical implementations
matic systems. Julia is a high-level programming language that is that this derivative must be continuous throughout the full
data and outlined potential strategies to advance temperature range. Some melting parameterizations result in a
has some similarities to MATLAB or Python (and allows the
conceptual models for dynamic magma systems. creation of short, easy-to-understand code), but is compiled significant jump in ∂ϕ ∕ ∂T around the solidus (or liquidus), which
Progress in this regard requires the integration of and therefore significantly faster than the above-mentioned puts severe time-step restrictions on the numerical simulation.
interpreted languages. Moreover, it has outstanding support One solution is to apply a small amount of smoothening to
empirical studies of volcanic and plutonic rocks that
for multi-threaded parallel central processing units and GPUs, existing parameterizations around the solidus and liquidus to
combine accessory mineral geochronology with an excellent testing framework, and a fairly rich ecosystem of enforce continuity. This is an option implemented in GeoParams.
other geological and petrological information such readily available packages that are written in a highly compos- jl and allows larger time-steps to be taken while giving essen-
able manner, which makes it easy to combine different packages. tially the same results.
as (chrono-)stratigraphy, as well as correlative trace-
The governing advection-diffusion equation relevant for the In MTK, the time-derivative in Equation (A3) is discretized
element and isotope data, and lastly refinement of thermal evolution of magmatic systems with intruding magma is: in an explicit manner:
numerical simulations of the thermomechanical evo-
T T T T n+1 T n T
lution of multiphase magmatic systems in dynamic cp +vj = k + Hr + QL s
. (A1) c p + QL = k + Hr
t xj xi xi t T t xi xi
settings. Further study of the thermal and compo- Here, T denotes temperature [K], v host-rock velocity
t T
sitional evolution of magmatic systems has direct (induced by injecting a new sill), t time, ϕs = (1−ϕ) the volume T n+1 = T n + k + Hr . (A4)
xi xi
fraction of solid rock, ϕ the melt fraction (both dimensionless c p + QL
implications for assessing volcanic hazards, con- T
parameters between [0–1]), ρ density [kg m–3], cp heat capacity
straining the heat sources of magmatic geothermal T
[J kg–1 K–1], k conductivity [W m–1 K–1], QL latent heat [J kg–1], The spatial derivative, k , is implemented using a
systems, and understanding the formation of ore and Hr radioactive heat sources [W m–3]. Many of the material xi xi
staggered grid finite difference scheme, which defines thermal
deposits associated with magmatic intrusions. properties can also be a function of temperature, which makes
this a nonlinear equation. conductivity in-between temperature grid points (e.g., Gerya,
In general, we solve Equation (A1) into two steps. 2019). In 1-D, this would be:
An advection step is applied only when a new sill is i+
1
T i +1 T i i
1
Ti Ti 1

ACKNOWLEDGMENTS intruded, which happens every few 1000 years in most cases k 2
k 2
T ∆x ∆x
discussed here. k , (A5)
This contribution is the outcome of a workshop, “ZASSy” (Zircon xi xi ∆x
T T
Age Spectra Systematics), which was generously supported by +vj = 0. (A2a)
t xj where Δx is the (constant) spacing between temperature grid-
the Terrestrial Magmatic Systems (TeMaS) Research Platform
We employ a Semi-Lagrangian advection scheme for this nodes that are defined at points indicated with a superscript i.
of Heidelberg and Mainz Universities, as well as by a MAGMA
(e.g., Staniforth and Côté, 1991; Becker and Kaus, 2016), which The thermal conductivity is physically located at a point that
ERC grant to Boris Kaus. Luca Caricchi received funding from the
takes into account that the host rock moves to create space for is exactly in between two temperature grid points. In case of
European Research Council (ERC) under the European Union’s
the intruding magma. The host-rock velocity is either set to be temperature-dependent conductivity, we first interpolate tem-
Horizon 2020 research and innovation program (grant agreement
no. 677493 – FEVER) and the Swiss National Science Foundation a constant below a horizontal sill (in the Geneva setup), can be T i +1 +T i i+
1
perature to that same grid point, i.e., T . 2
=
(grant no. 200021_184632). Olivier Bachmann acknowledges elliptical in mass-conserving manner (in the UCLA-HD setup), or 2
The advantage of employing the explicit approach is that it
funding from the Swiss National Science Foundation (grant no. can be set from an elastic dike intrusion algorithm.
results in an algorithm that can be quite efficiently implemented
200021_178928). We thank journal reviewers Adam Kent and A diffusion step is applied more frequently:
on the GPU, for example by using the ParallelStencil.jl Julia
Blair Schoene, as well as associate editor Jonathan Miller for
T T package (e.g., Räss et al., 2022). A disadvantage is that there
insightful comments. Marco Pfennig of Gau-Bickelheim, Ger- cp = k + Hr + QL s
. (A2b)
t xi xi t are time-step restrictions due to the Courant-Friedrichs-Lewy
many, is thanked for a stimulating mid-workshop excursion.
criteria, which state that:
The last term in the equation involves the time derivative
of the solid fraction. It turns out that this is numerically rather max ( ) ∆t
< 0.5, (A6)
unstable and requires a very small time-step. Therefore, we min ( ∆ x 2 , ∆y 2 , ∆ z 2 )
APPENDIX. MagmaThermoKinematics.jl, AN OPEN- apply the chain rule to this term and write this as:
SOURCE SOFWARE PACKAGE TO SIMULATE THE THERMAL k
where = is the thermal diffusivity and Δx, Δy, Δz
EVOLUTION OF MAGMATIC SYSTEMS T T T c p + QL
cp = k + Hr + QL s
t xi xi T t T
Modeling the thermal evolution of magmatic systems has a
the grid spacing in three directions (for 3-D). In case material
long tradition, and a range of commercial and non-commercial
T T T parameters are spatially variable or depend on temperature,
codes have been employed that intrude dikes and/or sills, and cp + QL = k + Hr
t T t xi xi one can use an upper bound to estimate κ.
simulate the resulting thermal evolution of the intrusion and
Equation A4 works for the case that parameters are inde-
surrounding rocks. As the modeling assumptions, boundary T T
c p + QL = k + Hr . (A3) pendent of temperature, which is generally not the case (ϕ and
conditions, and material parameters employed often differ T t xi xi ∂ϕ ∕ ∂T obviously depends on T ). Therefore, one typically has to
between publications, it is not always straightforward to com-
employ nonlinear iterations. In the general case that ρ, cp, k, and
pare results from different studies. In addition, none of the Here, we made use of the fact that ∂ϕs /∂T = −∂ϕ/∂T. This
∂ϕ ∕ ∂T depend on T, this gives:
software currently in use is available as an open source software, formulation shows that the latent heat effect can be imple-
which makes it difficult to reproduce results or understand the mented by locally changing the heat capacity (e.g., Simpson, T n+1,it +1 = T n +
t
k T ( ) Tn
( T n ) +Hr . (A7)
finer implementation details. 2017; Gerya, 2019), which is numerically much more stable than ( ) ( )
T c p T + QL
T
( )
T
xi xi

GEOSPHERE | Volume 19 | Number X

We employ dampened Picard iterations, which are computed as: REFERENCES CITED origin?: Geochimica et Cosmochimica Acta, v. 65, no. 21,
Angeles-De La Torre, C.A., Schmitt, A.K., Lovera, O.M., Gassert, p. 4043–4065, https://doi.org/10.1016/S0016-7037(01)0 0708-6.
T = T n+1,it +1
+ (1 )T n+1,it
, (A8) Boehnke, P., Watson, E.B., Trail, D., Harrison, T.M., and Schmitt,
H., Gerdes, A., and Harvey, J.C., 2023, A common magma
source for plutonic and volcanic rocks of the Geysers A.K., 2013, Zircon saturation re-revisited: Chemical Geol-
where ω = [0 − 1] is a dampening parameter and the superscript
geothermal field, California: Volume and intrusive history ogy, v. 351, p. 324–334, https://doi.org/10.1016/j.chemgeo
it
denotes the iteration step. Equations (A7) and (A8) are solved
derived from zircon: Chemical Geology, v. 624, https://doi .2013.05.028.
until the relative difference between subsequent iteration steps,
.org/10.1016/j.chemgeo.2023.121414. Borisov, A., and Aranovich, L., 2019, Zircon solubility in silicate
max(abs(T n+1,it +1 T n+1,it )) Annen, C., 2009, From plutons to magma chambers: Thermal melts: New experiments and probability of zircon crystalliza-
err = , (A9)
(
max abs (T n+1,it +1 ) ) constraints on the accumulation of eruptible silicic magma tion in deeply evolved basic melts: Chemical Geology, v. 510,
in the upper crust: Earth and Planetary Science Letters, v. 284, p. 103–112, https://doi.org/10.1016/j.chemgeo.2019.02.019.
becomes smaller than a predefined tolerance (e.g., 10−5).
no. 3–4, p. 409–416, https://doi.org/10.1016/j.epsl.2009.05.006. Botev, Z.I., Grotowski, J.F., and Kroese, D.P., 2010, Kernel density
In MTK, the governing Equations (A7) and (A8) are imple-
Annen, C., Blundy, J., and Sparks, R., 2006, The genesis of estimation via diffusion: Annals of Statistics, v. 38, no. 5,
mented in 2-D, 2-D axisymmetric, and 3-D geometries. Finite
intermediate and silicic magmas in deep crustal hot zones: p. 2916–2957, https://doi.org/10.1214/10-AOS799.
difference stencils are implemented using ParallelStencil.
Journal of Petrology, v. 47, no. 3, p. 505–539, https://doi.org Bouilhol, P., Connolly, J.A., and Burg, J.-P., 2011, Geological evi-
jl (https://github.com/omlins/ParallelStencil.jl), which is a Julia
/10.1093/petrology/egi084. dence and modeling of melt migration by porosity waves
package for efficient stencil computations that allows the sim-
ulations to be run on either parallel multicore processors or Annen, C., Blundy, J.D., Leuthold, J., and Sparks, R.S.J., 2015, in the sub-arc mantle of Kohistan (Pakistan): Geology, v. 39,
on GPU cards. Construction and evolution of igneous bodies: Towards an no. 12, p. 1091–1094, https://doi.org/10.1130/G32219.1.
Parameterizations for material properties (such as thermal integrated perspective of crustal magmatism: Lithos, v. 230, Caricchi, L., Simpson, G., and Schaltegger, U., 2014, Zircons
conductivity, melt fraction, or heat capacity) are implemented p. 206–221, https://doi.org/10.1016/j.lithos.2015.05.008. reveal magma fluxes in the Earth’s crust: Nature, v. 511,
in a separate package (GeoParams.jl1, https://github.com Bachmann, O., and Bergantz, G.W., 2006, Gas percolation in no. 7510, p. 457–461, https://doi.org/10.1038/nature13532.
/JuliaGeodynamics/ GeoParams.jl), which allows reuse of those upper-crustal silicic crystal mushes as a mechanism for Caricchi, L., Simpson, G., and Schaltegger, U., 2016, Estimates of
in independent software packages (while having to implement upward heat advection and rejuvenation of near-solidus volume and magma input in crustal magmatic systems from
and test them only once). We provide a range of melt-intrusion magma bodies: Journal of Volcanology and Geothermal zircon geochronology: The effect of modeling assumptions
algorithms, which include elliptical or cylindrical geometries, Research, v. 149, no. 1–2, p. 85–102, https://doi.org/10.1016 and system variables: Frontiers of Earth Science, v. 4, p. 48,
or follow the analytical solution for intrusion of a dill/dike in an /j.jvolgeores.2005.06.002. https://doi.org/10.3389/feart.2016.00048.
elastic half-space. Tracers can be injected to track the thermal Bachmann, O., Oberli, F., Dungan, M., Meier, M., Mundil, R., and Carrigan, C.R., 1988, Biot number and thermos bottle effect:
evolution of the intruded magma and host rock, and there is Fischer, H., 2007, 40Ar/39Ar and U–Pb dating of the Fish Can- Implications for magma-chamber convection: Geology, v. 16,
a simple mechanism to change the type of material properties yon magmatic system, San Juan Volcanic field, Colorado: no. 9, p. 771–774, https://d oi. org/ 10. 1130/ 0091- 7613( 1988)016
employed (e.g., using a temperature-dependent conductivity Evidence for an extended crystallization history: Chemical <0771:BNATBE>2.3.CO;2.
instead of a constant value). Both flux and isothermal bottom Geology, v. 236, no. 1–2, p. 134–166, https://doi.org/10.1016 Cherniak, D., and Watson, E., 2001, Pb diffusion in zircon: Chem-
boundary conductions can be implemented. /j.chemgeo.2006.09.005. ical Geology, v. 172, no. 1–2, p. 5–24, https://d oi. org/ 10. 1016
The results of each calculation comprise, besides the ther- Bacon, C.R., 1989, Crystallization of accessory phases in magmas /S0009-2541(00)0 0233-3.
mal field, time-temperature paths on every tracer, which can by local saturation adjacent to phenocrysts: Geochimica et Cherniak, D.J., Hanchar, J.M., and Watson, E.B., 1997, Diffusion
be used in a postprocessing step to compute zircon age data Cosmochimica Acta, v. 53, no. 5, p. 1055–1066, https://doi of tetravalent cations in zircon: Contributions to Mineral-
(using a Julia implementation of R-scripts provided by Weber et .org/10.1016/0016-7037(89)90210-X. ogy and Petrology, v. 127, no. 4, p. 383–390, https://doi.org
al., 2020, which is provided in GeoParams.jl). The file example Baker, D.R., Conte, A., Freda, C., and Ottolini, L., 2002, The /10.1007/s004100050287.
2D_ZASSy.jl reproduces the results shown in Figures 11 and 12, effect of halogens on Zr diffusion and zircon dissolution Claiborne, L.L., Miller, C., Walker, B., Wooden, J., Mazdab, F., and
which include both end-member models. in hydrous metaluminous granitic melts: Contributions to Bea, F., 2006, Tracking magmatic processes through Zr/Hf
Petrological-thermal modeling studies of magmatic systems Mineralogy and Petrology, v. 142, no. 6, p. 666–678, https:// ratios in rocks and Hf and Ti zoning in zircons: An example
often use slightly different assumptions, such as constant versus doi.org/10.1007/s00410-001-0328-3. from the Spirit Mountain batholith, Nevada: Mineralogical
temperature-dependent conductivity, 2-D versus axisymmetric Barboni, M., Boehnke, P., Schmitt, A.K., Harrison, T.M., Shane, Magazine, v. 70, no. 5, p. 517–543, https://doi.org/10.1180
geometries, or employ different melt-curve parameterizations. P., Bouvier, A.-S., and Baumgartner, L., 2016, Warm storage /0026461067050348.
It is often unclear what the impact of those assumptions will be for arc magmas: Proceedings of the National Academy of Claiborne, L.L., Miller, C.F., and Wooden, J.L., 2010, Trace element
on the final temperature distributions of the magmatic system. Sciences of the United States of America, v. 113, no. 49, composition of igneous zircon: A thermal and compositional
For that reason, we performed a sensitivity analysis using the p. 13,959–13,964, https://doi.org/10.1073/pnas.1616129113. record of the accumulation and evolution of a large silicic
two end-member scenarios described in the manuscript (Figs. 11 Becker, T.W., and Kaus, B.J.P., 2016, Numerical modeling of batholith, Spirit Mountain, Nevada: Contributions to Min-
and 12). In each case, we kept all parameters the same as in the Earth systems, an introduction to computational meth- eralogy and Petrology, v. 160, no. 4, p. 511–531, https://doi
reference simulation but changed one parameter. We report the ods with focus on solid Earth applications of continuum .org/10.1007/s00410-010-0491-5.
final temperature profiles in the center and at the side of the mechanics: University of Southern California Lecture Notes, Coleman, D.S., Gray, W., and Glazner, A.F., 2004, Rethinking the
domain. Results are shown in Fig. S1 (see text footnote 1) and http://www-udc.ig.utexas.edu/external/becker/preprints emplacement and evolution of zoned plutons: Geochrono-
demonstrate that the numerical parameters (such as resolution /Geodynamics557.pdf (last accessed April 2023). logic evidence for incremental assembly of the Tuolumne
or time-step) only have a minor effect, but that latent heating Bezanson, J., Edelman, A., Karpinski, S., and Shah, V.B., 2017, Intrusive Suite, California: Geology, v. 32, no. 5, p. 433–436,
and temperature-dependent conductivity can have an impact Julia: A fresh approach to numerical computing: SIAM https://doi.org/10.1130/G20220.1.
that is equal to or greater than reducing the magma flux for Review, v. 59, p. 65–98, https://doi.org/10.1137/141000671. Costa, F., Shea, T., and Ubide, T., 2020, Diffusion chronometry
the sill under-accretion setup, whereas they have little impact Bindeman, I.N., and Melnik, O.E., 2016, Zircon survival, rebirth and the timescales of magmatic processes: Nature Reviews:
on the central intrusion setup. This is likely because the cen- and recycling during crustal melting, magma crystalliza- Earth & Environment, v. 1, no. 4, p. 201–214, https://doi.org
tral intrusion setup has nearly reached a steady state, where tion, and mixing based on numerical modelling: Journal /10.1038/s43017-020-0038-x.
magma intrusion balances thermal diffusion and the partially of Petrology, v. 57, no. 3, p. 437–460, https://doi.org/10.1093 Coutts, D.S., Matthews, W.A., and Hubbard, S.M., 2019, Assess-
molten zone remains more or less constant in size. The sill under- /petrology/egw013. ment of widely used methods to derive depositional ages
accretion case, on the other hand, has a partially molten zone Blatter, D.L., and Carmichael, I.S., 2001, Hydrous phase equilibria from detrital zircon populations: Geoscience Frontiers, v. 10,
that still grows in space after 1.5 m.y. (Fig. S1). of a Mexican high-silica andesite: A candidate for a mantle no. 4, p. 1421–1435, https://d oi. org/ 10. 1016/ j. gsf. 2018. 11. 002.

GEOSPHERE | Volume 19 | Number X

Crisp, J.A., 1984, Rates of magma emplacement and volcanic improved model for aluminous and alkaline melts: Contri- Keller, C.B., Schoene, B., and Samperton, K.M., 2018, A stochas-
output: Journal of Volcanology and Geothermal Research, butions to Mineralogy and Petrology, v. 171, no. 3, https:// tic sampling approach to zircon eruption age interpretation:
v. 20, no. 3–4, p. 177–211, https://doi.org/10.1016/0377-0273 doi.org/10.1007/s00410-016-1227-y. Geochemical Perspectives Letters, v. 8, https://doi.org
(84)90039-8. Gerya, T., 2019, Introduction to Numerical Geodynamic Model- /10.7185/geochemlet.1826.
Crowley, J.L., Schoene, B., and Bowring, S., 2007, U-Pb dating ling: Cambridge, UK, Cambridge University Press, 484 p., Keller, T., and Katz, R.F., 2016, The role of volatiles in reactive
of zircon in the Bishop Tuff at the millennial scale: Geology, https://doi.org/10.1017/9781316534243. melt transport in the asthenosphere: Journal of Petrology,
v. 35, no. 12, p. 1123–1126, https://d oi. org/ 10. 1130/ G24017A. 1. Giordano, G., and Caricchi, L., 2022, Determining the state of v. 57, no. 6, p. 1073–1108, https://doi.org/10.1093/petrology
Curry, A., Caricchi, L., and Lipman, P.W., 2021, Magmatic evo- activity of transcrustal magmatic systems and their volca- /egw030.
lution of zoned and unzoned ignimbrites: Evidence for a noes: Annual Review of Earth and Planetary Sciences, v. 50, Keller, T., and Suckale, J., 2019, A continuum model of multi-
complex crustal architecture feeding four rapid-sequence, p. 231–259, https://doi.org/10.1146/annurev-earth-032320 phase reactive transport in igneous systems: Geophysical
caldera-forming eruptions in the San Juan Mountains, Col- -084733. Journal International, v. 219, no. 1, p. 185–222, https://doi
orado: Journal of Petrology, v. 62, no. 5, https://doi.org/10 Glazner, A.F., and Sadler, P.M., 2016, Estimating the duration of .org/10.1093/gji/ggz287.
.1093/petrology/egab006. geologic intervals from a small number of age determina- Keller, T., May, D.A., and Kaus, B.J., 2013, Numerical modelling
Deckart, K., Godoy, E., Bertens, A., Jerez, D., and Saeed, A., 2010, tions: A challenge common to petrology and paleobiology: of magma dynamics coupled to tectonic deformation of
Barren Miocene granitoids in the Central Andean metallo- Geochemistry, Geophysics, Geosystems, v. 17, no. 12, lithosphere and crust: Geophysical Journal International,
genic belt, Chile: Geochemistry and Nd-Hf and U-Pb isotope p. 4892–4898, https://doi.org/10.1002/2016GC006542. v. 195, no. 3, p. 1406–1442, https://d oi. org/ 10. 1093/ gji/ ggt306.
systematics: Andean Geology, v. 37, no. 1, p. 1–31, https:// Gutiérrez, F., Payacán, I., Szymanowski, D., Guillong, M., Bach- Keller, T., Katz, R.F., and Hirschmann, M.M., 2017, Volatiles beneath
doi.org/10.4067/S0718-71062010000100001. mann, O., and Parada, M., 2018, Lateral magma propagation mid-ocean ridges: Deep melting, channelised transport, focus-
Degruyter, W., and Huber, C., 2014, A model for eruption fre- during the emplacement of La Gloria Pluton, central Chile: ing, and metasomatism: Earth and Planetary Science Letters,
quency of upper crustal silicic magma chambers: Earth and Geology, v. 46, no. 12, p. 1051–1054, https://doi.org/10.1130 v. 464, p. 55–68, https://doi.org/10.1016/j.epsl.2017.02.006.
Planetary Science Letters, v. 403, p. 117–130, https://doi.org /G45361.1. Kelsey, D., and Powell, R., 2011, Progress in linking accessory
/10.1016/j.epsl.2014.06.047. Harrison, T.M., Watson, E.B., and Aikman, A.B., 2007, Tem- mineral growth and breakdown to major mineral evolution
Dodson, M., Compston, W., Williams, I., and Wilson, J., 1988, perature spectra of zircon crystallization in plutonic rocks: in metamorphic rocks: A thermodynamic approach in the
A search for ancient detrital zircons in Zimbabwean sed- Geology, v. 35, no. 7, p. 635–638, https://doi.org/10.1130 Na2O‐CaO‐K2O‐FeO‐MgO‐Al2O3‐SiO2‐H2O‐TiO2‐ZrO2 system:
iments: Journal of the Geological Society, v. 145, no. 6, /G23505A.1. Journal of Metamorphic Geology, v. 29, no. 1, p. 151–166,
p. 977–983, https://doi.org/10.1144/gsjgs.145.6.0977. Hayman, N.W., Rioux, M., Anma, R., Tani, K., Dunkley, D.J., Crow- https://doi.org/10.1111/j.1525-1314.2010.00910.x.
Dufek, J., and Bergantz, G.W., 2005, Lower crustal magma ley, J., and Schmitz, M., 2019, Accretion and oxidation of Kent, A.J., and Cooper, K.M., 2018, How well do zircons record
genesis and preservation: A stochastic framework for the a superfast-spread axial melt lens: TIMS and SIMS zircon the thermal evolution of magmatic systems?: Geology, v. 46,
evaluation of basalt–crust interaction: Journal of Petrology, analyses of the IODP Hole 1256D gabbros: Lithos, v. 348, no. 2, p. 111–114, https://doi.org/10.1130/G39690.1.
v. 46, no. 11, p. 2167–2195, https://doi.org/10.1093/petrology https://doi.org/10.1016/j.lithos.2019.105184. Kohn, M.J., 2016, Metamorphic chronology—a tool for all ages:
/egi049. Hou, T., Botcharnikov, R., Moulas, E., Just, T., Berndt, J., Koepke, Past achievements and future prospects: The American Min-
Fedele, L., Tarzia, M., Belkin, H.E., DeVivo, B., Lima, A., and J., Zhang, Z., Wang, M., Yang, Z., and Holtz, F., 2020, Kinet- eralogist, v. 101, no. 1, p. 25–42, https://doi.org/10.2138/am
Lowenstern, J.B., 2006, Magmatic-hydrothermal fluid inter- ics of Fe–Ti oxide re-equilibration in magmatic systems: -2016-5146.
action and mineralization in alkali-syenite nodules from Implications for thermo-oxybarometry: Journal of Petrology, Kohn, M.J., Corrie, S.L., and Markley, C., 2015, The fall and rise
the Breccia Museo pyroclastic deposit, Naples, Italy, in De v. 61, no. 11–12, https://doi.org/10.1093/petrology/egaa116. of metamorphic zircon: The American Mineralogist, v. 100,
Vivo, B., ed., Volcanism in the Campania Plain-Vesuvius Huber, C., Bachmann, O., and Manga, M., 2009, Homogeniza- no. 4, p. 897–908, https://doi.org/10.2138/am-2015-5064.
Campi Flegrei and Ignimbrites: Elsevier, Developments in tion processes in silicic magma chambers by stirring and Kryza, R., Crowley, Q.G., Larionov, A., Pin, C., Oberc-Dziedzic,
Volcanology, v. 9, p. 125–161. mushification (latent heat buffering): Earth and Planetary T., and Mochnacka, K., 2012, Chemical abrasion applied to
Flaherty, T., Druitt, T.H., Tuffen, H., Higgins, M.D., Costa, F., and Science Letters, v. 283, no. 1–4, p. 38–47, https://doi.org/10 SHRIMP zircon geochronology: An example from the Vari-
Cadoux, A., 2018, Multiple timescale constraints for high- .1016/j.epsl.2009.03.029. scan Karkonosze Granite (Sudetes, SW Poland): Gondwana
flux magma chamber assembly prior to the late Bronze Jackson, M., Gallagher, K., Petford, N., and Cheadle, M., 2005, Research, v. 21, no. 4, p. 757–767, https://doi.org/10.1016/j
Age eruption of Santorini (Greece): Contributions to Min- Towards a coupled physical and chemical model for tonalite– .gr.2011.07.007.
eralogy and Petrology, v. 173, no. 9, https://doi.org/10.1007 trondhjemite–granodiorite magma formation: Lithos, v. 79, Lasaga, A.C., 1986, Metamorphic reaction rate laws and devel-
/s00410-018-1490-1. no. 1–2, p. 43–60, https://doi.org/10.1016/j. lithos. 2004. 05. 004. opment of isograds: Mineralogical Magazine, v. 50, no. 357,
Friedrichs, B., Schmitt, A.K., Lovera, O.M., and Atıcı, G., 2021, Jackson, M., Blundy, J., and Sparks, R., 2018, Chemical differ- p. 359–3 73, https://d oi. org/ 10. 1180/ minmag. 1986. 050. 357. 02.
Zircon as a recorder of contrasting magma recharge and entiation, cold storage and remobilization of magma in the Liao, Y., Soule, S.A., Jones, M., and Le Mével, H., 2021, The
eruptive recurrence patterns: Earth and Planetary Science Earth’s crust: Nature, v. 564, no. 7736, p. 405–409, https:// mechanical response of a magma chamber with porovisco-
Letters, v. 571, https://doi.org/10.1016/j.epsl.2021.117104. doi.org/10.1038/s41586-018-0746-2. elastic crystal mush: Journal of Geophysical Research: Solid
Galbraith, R.F., 1990, The radial plot: Graphical assessment of Jackson, M.D., Cheadle, M.J., and Atherton, M.P., 2003, Earth, v. 126, no. 4, https://doi.org/10.1029/2020JB019395.
spread in ages: International Journal of Radiation Appli- Quantitative modeling of granitic melt generation and Linnen, R.L., and Keppler, H., 2002, Melt composition control
cations and Instrumentation. Part D: Nuclear Tracks and segregation in the continental crust: Journal of Geophys- of Zr/Hf fractionation in magmatic processes: Geochimica
Radiation Measurements, v. 17, no. 3, p. 207–214, https:// ical Research: Solid Earth, v. 108, no. B7, https://doi.org/10 et Cosmochimica Acta, v. 66, no. 18, p. 3293–3301, https://
doi.org/10.1016/1359-0189(90)90036-W. .1029/2001JB001050. doi.org/10.1016/S0016-7037(02)0 0924-9.
Gebauer, S.K., Schmitt, A.K., Pappalardo, L., Stockli, D.F., and Karakas, O., Degruyter, W., Bachmann, O., and Dufek, J., 2017, Liu, P.-P., Caricchi, L., Chung, S.-L., Li, X.-H., Li, Q.-L., Zhou, M.-F.,
Lovera, O.M., 2014, Crystallization and eruption ages of Lifetime and size of shallow magma bodies controlled by Lai, Y.-M., Ghani, A.A., Sihotang, T., and Sheldrake, T.E.,
Breccia Museo (Campi Flegrei caldera, Italy) plutonic clasts crustal-scale magmatism: Nature Geoscience, v. 10, no. 6, 2021, Growth and thermal maturation of the Toba magma
and their relation to the Campanian ignimbrite: Contribu- p. 446–450, https://doi.org/10.1038/ngeo2959. reservoir: Proceedings of the National Academy of Sciences
tions to Mineralogy and Petrology, v. 167, no. 1, https://doi Katz, R.F., Jones, D.W.R., Rudge, J.F., and Keller, T., 2022, Physics of the United States of America, v. 118, no. 45, https://doi
.org/10.1007/s00410-013-0953-7. of melt extraction from the mantle: Speed and style: Annual .org/10.1073/pnas.2101695118.
Gervasoni, F., Klemme, S., Rocha-Júnior, E.R., and Berndt, Review of Earth and Planetary Sciences, v. 50, p. 507–540, Lubbers, J., Deering, C., and Bachmann, O., 2020, Genesis
J., 2016, Zircon saturation in silicate melts: A new and https://doi.org/10.1146/annurev-earth-032320-083704. of rhyolitic melts in the upper crust: Fractionation and

GEOSPHERE | Volume 19 | Number X

remobilization of an intermediate cumulate at Lake City Räss, L., Utkin, I., Duretz, T., Omlin, S., and Podladchikov, Y.Y., Schmitt, A.K., and Vazquez, J.A., 2017, Secondary ionization
caldera, Colorado, USA: Journal of Volcanology and Geo- 2022, Assessing the robustness and scalability of the mass spectrometry analysis in petrochronology: Reviews
thermal Research, v. 392, https://doi.org/10.1016/j.jvolgeores accelerated pseudo-transient method: Geoscientific Model in Mineralogy and Geochemistry, v. 83, no. 1, p. 199–230,
.2019.106750. Development, v. 15, p. 5757–5786, https://doi.org/10.5194 https://doi.org/10.2138/rmg.2017.83.7.
Lukács, R., Harangi, S., Guillong, M., Bachmann, O., Fodor, L., /gmd-15-5757-2022. Schmitt, A.K., Stockli, D.F., Lindsay, J.M., Robertson, R., Lovera,
Buret, Y., Dunkl, I., Sliwinski, J., von Quadt, A., and Pey- Ratschbacher, B.C., Keller, C.B., Schoene, B., Paterson, S.R., O.M., and Kislitsyn, R., 2010, Episodic growth and homog-
tcheva, I., 2018, Early to mid-Miocene syn-extensional Anderson, J.L., Okaya, D., Putirka, K., and Lippoldt, R., 2018, enization of plutonic roots in arc volcanoes from combined
massive silicic volcanism in the Pannonian Basin (East- A new workflow to assess emplacement duration and melt U–Th and (U–Th)/He zircon dating: Earth and Planetary Sci-
Central Europe): Eruption chronology, correlation potential residence time of compositionally diverse magmas emplaced ence Letters, v. 295, no. 1–2, p. 91–103, https://doi.org/10
and geodynamic implications: Earth-Science Reviews, v. 179, in a sub-volcanic reservoir: Journal of Petrology, v. 59, no. 9, .1016/j.epsl.2010.03.028.
p. 1–19, https://doi.org/10.1016/j.earscirev.2018.02.005. p. 1787–1809, https://doi.org/10.1093/petrology/egy079. Schmitt, A.K., Perfit, M.R., Rubin, K.H., Stockli, D.F., Smith, M.C.,
Lukács, R., Caricchi, L., Schmitt, A.K., Bachmann, O., Karakas, O., Reid, M.R., Coath, C.D., Harrison, T.M., and McKeegan, K.D., 1997, Cotsonika, L.A., Zellmer, G.F., Ridley, W.I., and Lovera, O.M.,
Guillong, M., Molnár, K., Seghedi, I., and Harangi, S., 2021, Prolonged residence times for the youngest rhyolites asso- 2011, Rapid cooling rates at an active mid-ocean ridge from
Zircon geochronology suggests a long-living and active ciated with Long Valley Caldera: 230Th–238U ion microprobe zircon thermochronology: Earth and Planetary Science Let-
magmatic system beneath the Ciomadul volcanic dome dating of young zircons: Earth and Planetary Science Let- ters, v. 302, no. 3–4, p. 349–358, https://d oi. org/ 10. 1016/ j. epsl
field (eastern-central Europe): Earth and Planetary Science ters, v. 150, no. 1–2, p. 27–39, https://doi.org/10.1016/S0012 .2010.12.022.
Letters, v. 565, https://doi.org/10.1016/j.epsl.2021.116965. -821X(97)0 0077-0. Schmitt, A.K., Klitzke, M., Gerdes, A., and Schäfer, C., 2017, Zircon
Marillo-Sialer, E., Woodhead, J., Hanchar, J., Reddy, S., Greig, A., Riel, N., Bouilhol, P., van Hunen, J., Cornet, J., Magni, V., hafnium–oxygen isotope and trace element petrochronol-
Hergt, J., and Kohn, B., 2016, An investigation of the laser- Grigorova, V., and Velic, M., 2019, Interaction between ogy of intraplate volcanic rocks from the Eifel (Germany)
induced zircon ‘matrix effect’: Chemical Geology, v. 438, mantle-derived magma and lower arc crust: Quantitative and implications for mantle versus crustal origins of zircon
p. 11–24, https://doi.org/10.1016/j.chemgeo.2016.05.014. reactive melt flow modelling using Styx, in Ferraro, S. megacrysts: Journal of Petrology, v. 58, no. 9, p. 1841–1870,
Marxer, F., and Ulmer, P., 2019, Crystallisation and zircon satu- Lanari, P., Goncalves, P., and Grosch, E.P., eds., Metamor- https://doi.org/10.1093/petrology/egx075.
ration of calc-alkaline tonalite from the Adamello Batholith phic Geology: Microscale to Mountain Belts: Geological Schoene, B., 2014, 4.10-U–Th–Pb geochronology, in Holland,
at upper crustal conditions: An experimental study: Contri- Society, London, Special Publication 478, no. 1, p. 65–87, H.D., and Turekian, K.K., eds., Treatise on Geochemistry:
butions to Mineralogy and Petrology, v. 174, no. 10, https:// https://doi.org/10.1144/SP478.6. Elsevier, v. 4, p. 341–378, https://doi.org/10.1016/B978-0-08
doi.org/10.1007/s00410-019-1619-x. Riel, N., Kaus, B.J.P., Green, E.C.R, and Berlie, N., 2022, MAGE- -095975-7.00310-7.
Mattinson, J.M., 2005, Zircon U–Pb chemical abrasion (“CA- Min, an efficient Gibbs energy minimizer: Application to Shane, P., Storm, S., Schmitt, A.K., and Lindsay, J.M., 2012, Tim-
TIMS”) method: Combined annealing and multi-step partial igneous systems: Geochemistry, Geophysics, Geosystems, ing and conditions of formation of granitoid clasts erupted
dissolution analysis for improved precision and accuracy v. 23, no. 7, https://doi.org/10.1029/2022GC010427. in recent pyroclastic deposits from Tarawera Volcano (New
of zircon ages: Chemical Geology, v. 220, no. 1–2, p. 47–66, Rubatto, D., 2017, Zircon: The metamorphic mineral: Reviews Zealand): Lithos, v. 140, p. 1–10, https://doi.org/10.1016/j
https://doi.org/10.1016/j.chemgeo.2005.03.011. in Mineralogy and Geochemistry, v. 83, no. 1, p. 261–295, .lithos.2012.01.012.
Mehnert, H., Lipman, P., and Steven, T., 1973, Age of the Lake https://doi.org/10.2138/rmg.2017.83.9. Shao, T., Xia, Y., Ding, X., Cai, Y., and Song, M., 2020, Zircon
City caldera and related Sunshine Peak Tuff, western San Rummel, L., Kaus, B.J., Baumann, T.S., White, R.W., and Riel, N., saturation model in silicate melts: A review and update:
Juan Mountains: Colorado: Isochron-West, v. 6, p. 31–33. 2020, Insights into the compositional evolution of crustal Acta Geochimica, v. 39, no. 3, p. 387–403, https://doi.org/10
Melnik, O.E., Utkin, I.S., and Bindeman, I.N., 2021, Magma cham- magmatic systems from coupled petrological-geodynamical .1007/s11631-019-00384-4.
ber formation by dike accretion and crustal melting: 2-D models: Journal of Petrology, v. 61, no. 2, https://doi.org/10 Simpson, G., 2017, Practical Finite Element Modeling in Earth
thermo‐compositional model with emphasis on eruptions .1093/petrology/egaa029. Science Using Matlab: Chichester, UK, John Wiley & Sons,
and implication for zircon records: Journal of Geophys- Sambridge, M., and Compston, W., 1994, Mixture modeling of Ltd., https://doi.org/10.1002/9781119248644.
ical Research: Solid Earth, v. 126, https://doi.org/10.1029 multi-component data sets with application to ion-probe zir- Sliwinski, J.T., Guillong, M., Liebske, C., Dunkl, I., Von Quadt, A.,
/2021JB023008. con ages: Earth and Planetary Science Letters, v. 128, no. 3–4, and Bachmann, O., 2017, Improved accuracy of LA-ICP-MS
Miller, J.S., Matzel, J.E., Miller, C.F., Burgess, S.D., and Miller, p. 373–390, https://doi.org/10.1016/0012-821X(94)90157-0. U-Pb ages of Cenozoic zircons by alpha dose correction:
R.B., 2007, Zircon growth and recycling during the assembly Samperton, K.M., Bell, E.A., Barboni, M., Keller, C.B., and Chemical Geology, v. 472, p. 8–21, https://doi.org/10.1016
of large, composite arc plutons: Journal of Volcanology and Schoene, B., 2017, Zircon age-temperature-compositional /j.chemgeo.2017.09.014.
Geothermal Research, v. 167, no. 1–4, p. 282–299, https://d oi spectra in plutonic rocks: Geology, v. 45, no. 11, p. 983–986, Sliwinski, J.T., Guillong, M., Lipman, P.W., Zimmerer, M.J.,
.org/10.1016/j.jvolgeores.2007.04.019. https://doi.org/10.1130/G38645.1. Deering, C., and Bachmann, O., 2022, Zircon U-Pb geochro-
Oberli, F., Fischer, H., and Meier, M., 1990, High-resolution Satow, C., Gudmundsson, A., Gertisser, R., Ramsey, C.B., nology and trace element dataset from the Southern Rocky
238
U-206Pb zircon dating of Tertiary bentonites and Fish Bazargan, M., Pyle, D.M., Wulf, S., Miles, A.J., and Hardi- Mountain Volcanic Field, Colorado, USA: Data in Brief, v. 43,
Canyon Tuff; a test for age “concordance” by single-crystal man, M., 2021, Eruptive activity of the Santorini Volcano https://doi.org/10.1016/j.dib.2022.108362.
analysis, in Proceedings, Seventh International Conference controlled by sea-level rise and fall: Nature Geoscience, v. 14, Solano, J., Jackson, M., Sparks, R., Blundy, J.D., and Annen, C.,
on Geochronology, Cosmochronology and Isotope Geology: no. 8, p. 586–592, https://doi.org/10.1038/s41561-021-00783-4. 2012, Melt segregation in deep crustal hot zones: A mech-
Geological Society of Australia Special Publication 27, p. 74. Schaltegger, U., 2007, Hydrothermal zircon: Elements, v. 3, no. 1, anism for chemical differentiation, crustal assimilation and
Pamukçu, A.S., Schoene, B., Deering, C.D., Keller, C.B., and p. 51–79, https://doi.org/10.2113/gselements.3.1.51. the formation of evolved magmas: Journal of Petrology,
Eddy, M.P., 2022, Volcano-pluton connections at the Lake Schaltegger, U., Schmitt, A., and Horstwood, M., 2015, U–Th–Pb v. 53, no. 10, p. 1999–2026, https://d oi. org/ 10. 1093/ petrology
City magmatic center (Colorado, USA): Geosphere, v. 18, zircon geochronology by ID-TIMS, SIMS, and laser abla- /egs041.
no. 5, p. 1435–1452, https://doi.org/10.1130/GES02467.1. tion ICP-MS: Recipes, interpretations, and opportunities: Spera, F.J., and Bohrson, W.A., 2001, Energy-constrained
Pullen, A., Ibáñez-Mejía, M., Gehrels, G.E., Ibáñez-Mejía, J.C., Chemical Geology, v. 402, p. 89–110, https://doi.org/10.1016 open-system magmatic processes I: General model and
and Pecha, M., 2014, What happens when n= 1000? Cre- /j.chemgeo.2015.02.028. energy-constrained assimilation and fractional crystalliza-
ating large-n geochronological datasets with LA-ICP-MS Schmitt, A.K., 2011, Uranium series accessory crystal dating of tion (EC-AFC) formulation: Journal of Petrology, v. 42, no. 5,
for geologic investigations: Journal of Analytical Atomic magmatic processes: Annual Review of Earth and Planetary p. 999–1018, https://doi.org/10.1093/petrology/42.5.999.
Spectrometry, v. 29, no. 6, p. 971–980, https://doi.org/10 Sciences, v. 39, p. 321–349, https://doi.org/10.1146/annurev Staniforth, A., and Côté, J., 1991, Semi-Lagrangian integration
.1039/C4JA00024B. -earth-040610-133330. schemes for atmospheric models—A review: Monthly

GEOSPHERE | Volume 19 | Number X

Weather Review, v. 119, no. 9, p. 2206–2223, https://doi.org Aso: Geochemistry, Geophysics, Geosystems, v. 20, no. 3, and implications for isotopic inheritance: Earth and Envi-
/10.1175/1520-0493(1991)119<2206:SLISFA>2.0.CO;2. p. 1574–1591, https://doi.org/10.1029/2018GC008103. ronmental Science Transactions of the Royal Society of
Storm, S., Shane, P., Schmitt, A.K., and Lindsay, J.M., 2012, Troch, J., Ellis, B.S., Schmitt, A.K., Bouvier, A.-S., and Bachmann, Edinburgh, v. 87, no. 1–2, p. 43–56, https://doi.org/10.1017
Decoupled crystallization and eruption histories of the rhy- O., 2018, The dark side of zircon: Textural, age, oxygen isotopic /S0263593300006465.
olite magmatic system at Tarawera volcano revealed by and trace element evidence of fluid saturation in the subvol- Watson, E.B., and Harrison, T.M., 1983, Zircon saturation revis-
zircon ages and growth rates: Contributions to Mineralogy canic reservoir of the Island Park–Mount Jackson Rhyolite, ited: Temperature and composition effects in a variety of
and Petrology, v. 163, no. 3, p. 505–519, https://doi.org/10 Yellowstone (USA): Contributions to Mineralogy and Petrol- crustal magma types: Earth and Planetary Science Letters,
.1007/s00410-011-0682-8. ogy, v. 173, no. 7, https://doi.org/10.1007/s00410-018-1481-2. v. 64, no. 2, p. 295–304, https://doi.org/10.1016/0012-821X
Szymanowski, D., Forni, F., Wolff, J.A., and Ellis, B.S., 2020, Turner, A.J., Katz, R.F., Behn, M.D., and Keller, T., 2017, Magmatic (83)90211-X.
Modulation of zircon solubility by crystal–melt dynamics: focusing to mid‐ocean ridges: The role of grain‐size variabil- Weber, G., Caricchi, L., Arce, J.L., and Schmitt, A.K., 2020, Deter-
Geology, v. 48, no. 8, p. 798–802, https://doi.org/10.1130 ity and non‐Newtonian viscosity: Geochemistry, Geophysics, mining the current size and state of subvolcanic magma
/G47405.1. Geosystems, v. 18, no. 12, p. 4342–4355, https://doi.org/10 reservoirs: Nature Communications, v. 11, no. 1, https://doi
Tierney, C.R., Schmitt, A.K., Lovera, O.M., and de Silva, S.L., 2016, .1002/2017GC007048. .org/10.1038/s41467-020-19084-2.
Voluminous plutonism during volcanic quiescence revealed Vermeesch, P., 2004, How many grains are needed for a prov- White, L.T., and Ireland, T.R., 2012, High-uranium matrix effect
by thermochemical modeling of zircon: Geology, v. 44, no. 8, enance study?: Earth and Planetary Science Letters, v. 224, in zircon and its implications for SHRIMP U–Pb age deter-
p. 683–686, https://doi.org/10.1130/G37968.1. no. 3–4, p. 441–451, https://doi.org/10.1016/j.epsl.2004.05.037. minations: Chemical Geology, v. 306, p. 78–91, https://doi
Townsend, M., 2022, Linking surface deformation to thermal Vermeesch, P., 2012, On the visualisation of detrital age distri- .org/10.1016/j.chemgeo.2012.02.025.
and mechanical magma chamber processes: Earth and butions: Chemical Geology, v. 312, p. 190–194, https://doi Widmann, P., Davies, J., and Schaltegger, U., 2019, Calibrating
Planetary Science Letters, v. 577, https://doi.org/10.1016/j .org/10.1016/j.chemgeo.2012.04.021. chemical abrasion: Its effects on zircon crystal structure,
.epsl.2021.117272. von Quadt, A., Gallhofer, D., Guillong, M., Peytcheva, I., Waelle, chemical composition and U-Pb age: Chemical Geology,
Townsend, M., and Huber, C., 2020, A critical magma chamber M., and Sakata, S., 2014, U–Pb dating of CA/non-CA treated v. 511, p. 1–10, https://doi.org/10.1016/j.chemgeo.2019
size for volcanic eruptions: Geology, v. 48, no. 5, p. 431–435, zircons obtained by LA-ICP-MS and CA-TIMS techniques: .02.026.
https://doi.org/10.1130/G47045.1. Impact for their geological interpretation: Journal of Analyt- Wotzlaw, J.-F., Schaltegger, U., Frick, D.A., Dungan, M.A.,
Townsend, M., Huber, C., Degruyter, W., and Bachmann, O., ical Atomic Spectrometry, v. 29, no. 9, p. 1618–1629, https:// Gerdes, A., and Günther, D., 2013, Tracking the evolution
2019, Magma chamber growth during intercaldera periods: doi.org/10.1039/C4JA00102H. of large-volume silicic magma reservoirs from assembly
Insights from thermo‐mechanical modeling with applica- Watson, E.B., 1996, Dissolution, growth and survival of zircons to supereruption: Geology, v. 41, no. 8, p. 867–870, https://
tions to Laguna del Maule, Campi Flegrei, Santorini, and during crustal fusion: Kinetic principals, geological models doi.org/10.1130/G34366.1.

GEOSPHERE | Volume 19 | Number X

Private Health Institutions Law
100% (1)
Private Health Institutions Law
22 pages
Casio AP500
0% (1)
Casio AP500
42 pages
Paper 5
No ratings yet
Paper 5
17 pages
Considerations in Zircon Geochronology by SIMS
No ratings yet
Considerations in Zircon Geochronology by SIMS
28 pages
Zircon SIMS U-Pb Age of The Shaxinan Melagabbro Eastern Tianshan and Constraints On Fe-Ti-V Oxide Mineralization
No ratings yet
Zircon SIMS U-Pb Age of The Shaxinan Melagabbro Eastern Tianshan and Constraints On Fe-Ti-V Oxide Mineralization
2 pages
Cochrane 2013
No ratings yet
Cochrane 2013
209 pages
Revista - Elements Magazine - Zircon
100% (1)
Revista - Elements Magazine - Zircon
6 pages
Most Granitoid Rocks Are Cumulates
No ratings yet
Most Granitoid Rocks Are Cumulates
14 pages
U PB Analytical Methods Zircon
No ratings yet
U PB Analytical Methods Zircon
11 pages
MDMW Zircon04
No ratings yet
MDMW Zircon04
10 pages
Grimesetal 2015
No ratings yet
Grimesetal 2015
27 pages
Belousova Et Al 2006 ZR Morph
No ratings yet
Belousova Et Al 2006 ZR Morph
25 pages
(Gehrels, 2014) PDF
No ratings yet
(Gehrels, 2014) PDF
27 pages
The Elements of Geology; Adapted to the Use of Schools and Colleges
From Everand
The Elements of Geology; Adapted to the Use of Schools and Colleges
Justin R. Loomis
No ratings yet
Chemical Abrasion: The Mechanics of Zircon Dissolution
No ratings yet
Chemical Abrasion: The Mechanics of Zircon Dissolution
25 pages
Fisher Et Al (2017) - Combining ND Isotopes in Monazite and HF Isotopes in Zircon To Understand Complex Open-System Processes in Granitic Magmas
No ratings yet
Fisher Et Al (2017) - Combining ND Isotopes in Monazite and HF Isotopes in Zircon To Understand Complex Open-System Processes in Granitic Magmas
4 pages
Keller 2016 PHD Dissertation
No ratings yet
Keller 2016 PHD Dissertation
184 pages
Frontiers in Geochemistry: Contribution of Geochemistry to the Study of the Earth
From Everand
Frontiers in Geochemistry: Contribution of Geochemistry to the Study of the Earth
Russell Harmon
No ratings yet
Reading Material 1
No ratings yet
Reading Material 1
6 pages
2025 Loucks p511
No ratings yet
2025 Loucks p511
14 pages
Detrital Zircon Record and Tectonic Setting: P.A. Cawood, C.J. Hawkesworth and B. Dhuime
No ratings yet
Detrital Zircon Record and Tectonic Setting: P.A. Cawood, C.J. Hawkesworth and B. Dhuime
4 pages
Introducing Volcanology for tablet devices: A Guide to Hot Rocks
From Everand
Introducing Volcanology for tablet devices: A Guide to Hot Rocks
Dougal Jerram
5/5 (1)
Reviews in Mineralogy & Geochemistry Vol. 58, Pp. 205-238, 2005
No ratings yet
Reviews in Mineralogy & Geochemistry Vol. 58, Pp. 205-238, 2005
33 pages
Zircon (Reviews in Mineralogy and Geochemistry Vol53 - 2003)
100% (1)
Zircon (Reviews in Mineralogy and Geochemistry Vol53 - 2003)
501 pages
Edocsp 24 V 2 CH 10
No ratings yet
Edocsp 24 V 2 CH 10
22 pages
02 Whole
No ratings yet
02 Whole
160 pages
Temperature Spectra of Zircon Crystallization in Plutonic Rocks Harrison2007
No ratings yet
Temperature Spectra of Zircon Crystallization in Plutonic Rocks Harrison2007
5 pages
Seven Papers: Debating the Creation of Our Modern Brain
From Everand
Seven Papers: Debating the Creation of Our Modern Brain
Dan M. Mrejeru
No ratings yet
The Principles of Stratigraphical Geology
From Everand
The Principles of Stratigraphical Geology
J. E. Marr
4/5 (1)
U PB Methods PDF
No ratings yet
U PB Methods PDF
44 pages
U PB IgneousSamples Method
No ratings yet
U PB IgneousSamples Method
3 pages
Chemical Geology
No ratings yet
Chemical Geology
13 pages
Seven Papers 2nd Edition
From Everand
Seven Papers 2nd Edition
Dan Mrejeru
No ratings yet
Spencer - 12 - U-Pb HF Isotope Zircons Garhwal
No ratings yet
Spencer - 12 - U-Pb HF Isotope Zircons Garhwal
10 pages
Herrmann, M. Et Al. (2021)
No ratings yet
Herrmann, M. Et Al. (2021)
11 pages
Dating UHP Rocks (Italy) On WM, Comparaison Ar Ar Et RBSR Analysis
No ratings yet
Dating UHP Rocks (Italy) On WM, Comparaison Ar Ar Et RBSR Analysis
27 pages
Paper 4
No ratings yet
Paper 4
15 pages
Ghosh Et Al 2023 GJ
No ratings yet
Ghosh Et Al 2023 GJ
28 pages
Liu 2007
No ratings yet
Liu 2007
8 pages
Turner 1988
No ratings yet
Turner 1988
1 page
The Temporal Evolution of Chemical and Physical Properties of Magmatic Systems
No ratings yet
The Temporal Evolution of Chemical and Physical Properties of Magmatic Systems
15 pages
Contoh Analisis Sayatan Tipis 2 Paper
No ratings yet
Contoh Analisis Sayatan Tipis 2 Paper
19 pages
Bachmann 2004
No ratings yet
Bachmann 2004
18 pages
Magma Sources and Gold Mineralisation in The Mount Leyshon
No ratings yet
Magma Sources and Gold Mineralisation in The Mount Leyshon
27 pages
Pizarro Et A., 2020. Porphyry Indicator Zircons (PIZs) - Application To Exploration of Porphyry Copper Deposits
No ratings yet
Pizarro Et A., 2020. Porphyry Indicator Zircons (PIZs) - Application To Exploration of Porphyry Copper Deposits
18 pages
Zircon Tiny But Timely
No ratings yet
Zircon Tiny But Timely
7 pages
Paper 1
No ratings yet
Paper 1
25 pages
Sharma Et Al 2022 U PB Zircon Geochronology of A Pyroclastic Rock From The Parsoi Formation - Geological Journal
No ratings yet
Sharma Et Al 2022 U PB Zircon Geochronology of A Pyroclastic Rock From The Parsoi Formation - Geological Journal
17 pages
Geochronology in Myanmar (1964-2017)
No ratings yet
Geochronology in Myanmar (1964-2017)
47 pages
Alasino Et Al 2019 Schlieren Sierra Nevada Folded and Faulted
No ratings yet
Alasino Et Al 2019 Schlieren Sierra Nevada Folded and Faulted
26 pages
Artemieva Et Al., 2002
No ratings yet
Artemieva Et Al., 2002
15 pages
End-Triassic Mass Extinction Started by Intrusive CAMP Activity Pres
No ratings yet
End-Triassic Mass Extinction Started by Intrusive CAMP Activity Pres
12 pages
Evidence For Extremely Rapid Magma Ocean Crystallization and Crust Formation On Mars
No ratings yet
Evidence For Extremely Rapid Magma Ocean Crystallization and Crust Formation On Mars
6 pages
Data Repository Material
No ratings yet
Data Repository Material
31 pages
Grondahl - 2022 - Ore Fertility of Arc Magmas
No ratings yet
Grondahl - 2022 - Ore Fertility of Arc Magmas
11 pages
Temperatura de Cristalização Do Zircão
No ratings yet
Temperatura de Cristalização Do Zircão
4 pages
U-Pb Zircon Dating by Laser ablation-MC-ICP-MS Using A New Multiple Ion Counting Faraday Collector Array
No ratings yet
U-Pb Zircon Dating by Laser ablation-MC-ICP-MS Using A New Multiple Ion Counting Faraday Collector Array
10 pages
Dahlquist Et Al 2013
No ratings yet
Dahlquist Et Al 2013
14 pages
(Penting) Pairwise Sample Comparisons and Multidimensional Scaling of Detrital Zircon Ages With Examples From The North American Paltform, Basin, and Passive Margin Setting
No ratings yet
(Penting) Pairwise Sample Comparisons and Multidimensional Scaling of Detrital Zircon Ages With Examples From The North American Paltform, Basin, and Passive Margin Setting
15 pages
Frontiers in Geofluids
From Everand
Frontiers in Geofluids
Bruce W. D. Yardley
No ratings yet
Seismoelectric Exploration: An Intro to New Methods
From Everand
Seismoelectric Exploration: An Intro to New Methods
Samir Ganaka
No ratings yet
2019 - Whalen & Hildebrand
No ratings yet
2019 - Whalen & Hildebrand
20 pages
Aramaki, S. Et Al. (1984)
No ratings yet
Aramaki, S. Et Al. (1984)
16 pages
Kohn, B. Et Al. (2024)
No ratings yet
Kohn, B. Et Al. (2024)
30 pages
Sakagami, S. (1998)
No ratings yet
Sakagami, S. (1998)
24 pages
Madella, A. Et Al. (2022)
No ratings yet
Madella, A. Et Al. (2022)
14 pages
Reimann, C. Et Al. (2006)
No ratings yet
Reimann, C. Et Al. (2006)
4 pages
Harrison, T. Et Al. (2005)
No ratings yet
Harrison, T. Et Al. (2005)
21 pages
Reiners, P. Et Al. (2005)
No ratings yet
Reiners, P. Et Al. (2005)
18 pages
Neoformation of Exotic Copper Minerals From Gel-Li
No ratings yet
Neoformation of Exotic Copper Minerals From Gel-Li
21 pages
Pohl, W. (2022)
No ratings yet
Pohl, W. (2022)
36 pages
De Oliveira, S. Et Al. (2021)
No ratings yet
De Oliveira, S. Et Al. (2021)
20 pages
Wilson, J. (1963)
No ratings yet
Wilson, J. (1963)
34 pages
Polliand, M. Et Al. (2005)
No ratings yet
Polliand, M. Et Al. (2005)
12 pages
K-Ar Age Quaternary: Determinations of Late Andean Volcanic Rocks, Tertiary and Southern Peru
No ratings yet
K-Ar Age Quaternary: Determinations of Late Andean Volcanic Rocks, Tertiary and Southern Peru
7 pages
DeMartinoetal 2022
No ratings yet
DeMartinoetal 2022
20 pages
Plumbing Tools and Their Uses
No ratings yet
Plumbing Tools and Their Uses
6 pages
Porsche Case Study
No ratings yet
Porsche Case Study
4 pages
Intervention21120-5570393 152823
No ratings yet
Intervention21120-5570393 152823
10 pages
Aluminum and Glass Company in Qatar
No ratings yet
Aluminum and Glass Company in Qatar
5 pages
Financial Kake Da Hotel (N)
No ratings yet
Financial Kake Da Hotel (N)
10 pages
Project Scope Statement1
No ratings yet
Project Scope Statement1
6 pages
Statement Up0510110008421
No ratings yet
Statement Up0510110008421
3 pages
Mining Terms 1
No ratings yet
Mining Terms 1
23 pages
5 Muscle
No ratings yet
5 Muscle
3 pages
Mpa 17 em PDF
No ratings yet
Mpa 17 em PDF
9 pages
Grade 5 Write Expressions A
No ratings yet
Grade 5 Write Expressions A
2 pages
Java4s Com Hibernate
No ratings yet
Java4s Com Hibernate
5 pages
BS en 60974-6-2003 (2005)
No ratings yet
BS en 60974-6-2003 (2005)
24 pages
Trabajo Final de Ingles Técnico
No ratings yet
Trabajo Final de Ingles Técnico
5 pages
2018 Oakland Linuxmalware
No ratings yet
2018 Oakland Linuxmalware
15 pages
Web of Science Core Collection:: Journal Evaluation Process and Selection Criteria
No ratings yet
Web of Science Core Collection:: Journal Evaluation Process and Selection Criteria
35 pages
The Relationship of Endodontic-Periodontic Lesions
No ratings yet
The Relationship of Endodontic-Periodontic Lesions
7 pages
Teaching Behavioral Ethics by Robert A. Prentice
No ratings yet
Teaching Behavioral Ethics by Robert A. Prentice
41 pages
SCBA Pre-Use Inspection
No ratings yet
SCBA Pre-Use Inspection
2 pages
Domino Squares
100% (2)
Domino Squares
1 page
Syllabus
No ratings yet
Syllabus
7 pages
Irr 7920
No ratings yet
Irr 7920
15 pages
FICM Unit 3
No ratings yet
FICM Unit 3
6 pages
Hyaluronic Acid
No ratings yet
Hyaluronic Acid
7 pages
Computerised Assessment of Handwriting
No ratings yet
Computerised Assessment of Handwriting
15 pages
Equlibrium
No ratings yet
Equlibrium
20 pages
WAH5 - Functional Language Worksheets
No ratings yet
WAH5 - Functional Language Worksheets
6 pages
1.1 Survey of The History, Growth and Role of Translation in India
No ratings yet
1.1 Survey of The History, Growth and Role of Translation in India
50 pages

Schmitt, A. Et Al. (2023)

Uploaded by

Schmitt, A. Et Al. (2023)

Uploaded by

Research Paper THEMED ISSUE: Different Personalities of Granites and Rhyolites: Silicic Magmas from the Lower Crust

GEOSPHERE Zircon age spectra to quantify magma evolution

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

pre-eruption post-eruption involve high-​temperature, mafic-​to-​intermediate

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

A KDE B radial plot 30 Ma C abanico 28

Central age: 16.5 Ma

Peak 1: 14.8 Ma (31.6%)

Peak 3: 27.5 Ma (1.7%) 20

15 20 25 30 0 50 100 150 200 250 0 22 44 66 88 110 132

GEOSPHERE | Volume 19 | Number X

Frequency (relative probability)

data set, the more accurate the representation of

limits on the number that can be analyzed given

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

Frequency (relative probability)

9 10 11 12 13 0 50 100 150 200 250

A B central age = 29.6 Ma (n=7023)

Peak 1: 23.1 Ma (3.8%)

20 25 30 35 40 0 50 100 150 200 250

Age [Ma] t/σ

GEOSPHERE | Volume 19 | Number X

250 300 350 400 450 500 550 600 0 20 40 60 80 100

Age [Ma] t/σ

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

Tracers (magma pulse)

B Time Figure 10. Schematic representation of temporal

GEOSPHERE | Volume 19 | Number X

Sill underaccretion model Recharge rate Recharge into standing melt

8 × 10−6 km3/yr/km2, with magmatic recharge temperature

of total intrusive volume (red dashed line), and parts of

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

7.5×10 km /yr/km 9.1×10 km /yr/km 1.1×10 km /yr/km

Zircon age cumulative probability %

matic volumes in the UCLA-​HD models

GEOSPHERE | Volume 19 | Number X

ID-TIMS would be obtained using ID-TIMS is significantly

Volume fraction zones of different ages

C D crystals, we calculated the distribution of ages

mixing of crystal domains at different depths for

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

Model setup: Geneva UCLA-HD

20 volcanoes in central Anatolia (Friedrichs et al., 2021); data

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

GEOSPHERE | Volume 19 | Number X

You might also like

pre-eruption post-eruption involve high-temperature, mafic-to-intermediate

matic volumes in the UCLA-HD models