Journal of Volcanology and Geothermal Research 389 (2020) 106744

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

journal homepage: www.elsevier.com/locate/jvolgeores

Statistical separation of tectonic and inflation-driven components of

deformation on silicic reservoirs, Laguna del Maule volcanic field, Chile
N. Garibaldi a,⁎, B. Tikoff a, D. Peterson b, J.R. Davis c, K. Keranen b
a
Department of Geoscience, University of Wisconsin-Madison, Madison, WI, USA
b
Cornell University, Ithaca, NY, USA
c
Carleton College, Northfield, MN, USA

a r t i c l e i n f o a b s t r a c t

Article history: The reconstruction of the structural history of inflating silicic systems is challenging because their faulting record
Received 20 June 2019 encompasses tectonic and inflation-driven deformation, and separating the two can rarely be accomplished.
Received in revised form 15 November 2019 Here, we present and utilize a statistical methodology to differentiate tectonic from inflation-driven deformation
Accepted 20 November 2019
in the Laguna del Maule (LdM), a post-glacial rhyolitic volcanic field (Southern Volcanic Zone, Chile, latitude 36°
Available online 23 November 2019
S). LdM is cut by the Troncoso fault, a major normal fault that strikes NE and dips to the NW. The Troncoso hang-
ing wall contains abundant, young (b 23 ka) NE-striking normal faults, whereas the footwall is largely unfaulted,
with few NNW-striking faults. Activity within the shallow (b8 km) LdM reservoir has caused two inflation events,
both centered in different areas of the footwall: 1) a geodetic-based (InSAR) inflation, acting since 2007; and 2) a
shoreline-based inflation, evidenced by warping of a 9.4 ka high-stand shoreline. To separate tectonic and
inflation-driven faults in the Troncoso hanging wall, we perform a statistical analysis of fault orientation to com-
pare them to the elongation direction predicted by tectonics. The orientation of normal faults is consistent with
NW-SE tectonic elongation recorded outside the inflating area. To evaluate if either inflation event has
reactivated these tectonic faults, we perform a statistical analysis of orientation-and-heave (horizontal offset
on a fault): we calculate the maximum elongation direction (the accumulated horizontal offset) and compare
it to elongation predicted by tectonics, shoreline-based inflation and geodetic-based inflation. The maximum
elongation direction is sub-parallel to elongation predicted by shoreline-based inflation and is statistically differ-
ent from tectonic elongation. We interpret a first stage of tectonic extension (N19–9.4 ka) where normal faults are
developed on the Troncoso hanging wall. During shoreline-based inflation (b9.4 ka), faults on the hanging wall
are reactivated, suppressing uplift; the largely unfaulted footwall uplifts instead. The differential uplift is accom-
modated by slip on the Troncoso fault. The current, geodetic based uplift, nucleates few faults with large offsets,
and potentially reactivates the Troncoso fault. From the structural history, we suggest that the structural architec-
ture is associated with surface volcanism: deformation of the Troncoso hanging wall appears to facilitate re-
peated, small volume, post-glacial rhyodacitic and andesitic eruptions, whereas the largely unfaulted Troncoso
footwall seems to favor rhyolite accumulation and eruption during the late Holocene.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction ongoing eruptions (Costa et al., 2011). Additionally, pre-existing

crustal-scale tectonic faults channel migration of rhyolite, localizing
Silicic volcanic systems are preferentially located in areas of tectonic effusive loci at the surface (e.g., Aguirre-Díaz and Labarthe-Hernández,
extension, where tectonic deformation regulates several magmatic pro- 2003; Goto and McPhie, 2018). Linking tectonic deformation to the
cesses within the system, potentially affecting the composition of timing, location and composition of eruptive products is fundamental
erupted magmas (Hughes and Mahood, 2008, 2011; Farris et al., to understanding tectonic controls on silicic systems.
2017). Tectonic extension can trigger violent eruptions via nucleation The structural characterization of tectonic deformation in resurgent
or reactivation of surface-to-reservoir faults that facilitate magma as- systems is complicated by the presence of inflation-associated faults.
cent (Petrinovic et al., 2010; Allan et al., 2012, 2017). Tectonic extension Rapid inflation rates of mm/yr to cm/yr above silicic reservoirs result
also modulates the eruptive frequency (Cabaniss et al., 2018) and flux of in concentric uplift patterns that mask tectonic extension in geodetic
measurements (e.g., Le Mével et al., 2016; Magee et al., 2017). These in-
⁎ Corresponding author. flation rates are in stark contrast to what are considered typical strain
E-mail address: garibaldi@wisc.edu (N. Garibaldi). rates (10−13–10−15 s−1; Pfiffner and Ramsay, 1982), which are

https://doi.org/10.1016/j.jvolgeores.2019.106744
0377-0273/© 2019 Elsevier B.V. All rights reserved.
2 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744

equivalent to tens of mms of horizontal deformation distributed over 40 km3 of post-glacial (b26 ka) eruptive products, comprising volumi-
lengths of a few hundred kms. nous rhyolites and small volume rhyodacites and andesites that encircle
For this reason, the record of tectonic deformation in resurgent sys- a lake basin (Hildreth et al., 2010; Singer et al., 2014, 2018; Andersen
tems is typically restricted to surface structures in and around the resur- et al., 2017, 2018). The major structural features in the basin include
gent area. Repeated, high-flux magmatic injections keep resurgence the NE-striking, NW-dipping, Troncoso normal fault that projects di-
cycles active for 105 to 106 years (Coleman et al., 2004; de Saint rectly into the center of the lake from the southwest, and a series of
Blanquat et al., 2011; Schöpa and Annen, 2013; Menand et al., 2015), NE-striking and NW-striking collapse structures on the north edge of
which over decadal to centennial timescales accumulate considerable the lake (Fig. 2). These caldera collapse structures record up to 500 m
strain on the roof-rocks (Puskas et al., 2007; Robertson and Kilburn, of offset, defining the south edge of a caldera produced by the eruption
2016; Kilburn et al., 2017; Singer et al., 2018). Throughout these cycles of the Cajones de Bobadilla ignimbrite ~990 ka (Hildreth et al., 2010;
of magmatic uplift, faults in and around the resurgent area are nucleated Andersen et al., 2017).
and reactivated to accommodate the inflation-driven deformation Within the Troncoso fault zone, normal-sense kinematics are indi-
(e.g., Acocella et al., 2004; Levy et al., 2018). cated by ~20 m vertical offset of a 9.4 kyr old high-stand shoreline; nor-
Studies aiming to reveal the structural evolution of volcanic systems mal motions are also suggested by a ~200 meter vertical offset of a
gain insight from eroded sections and numerical models Miocene-Pliocene unconformity (Singer et al., 2018; this study). The
(e.g., Gudmundsson, 2006). Importantly, these studies also rely on projection of the fault intersects the center of ongoing uplift, the
surface observations of structure orientation and distribution modeled source of deformation (5 km depth, Feigl et al., 2014), and
(e.g., Rowland and Sibson, 2001), vent distribution and morphology the low-density gravity anomaly (2 km depth, Miller et al., 2017) that
(e.g., Corazzato and Tibaldi, 2006), a combination of these two methods suggest the presence of melt-rich pockets. In contrast, a low shear
(e.g., Tadini et al., 2014; Bolós et al., 2015; Isaia et al., 2015), and fault ki- wave velocity body (2–8 km depth, Wespestad et al., 2019) and large
nematic indicators to infer paleostress/strain tensors (e.g., García- magnetotelluric conductors (2–12 km depth, Cordell et al., 2018)
Palomo et al., 2004; Pérez-Flores et al., 2017; Barcelona et al., 2019). interpreted as a crystal-rich mush are located along the downward ex-
These techniques are effective at revealing the structural architecture tent of the Troncoso fault, below the mapped hanging wall, in the north-
and major fault kinematics within volcanic systems, but can rarely dis- west portion of the lake (Fig. 2B).
tinguish between structures accommodating tectonic deformation and On the surface, the Troncoso fault separates distinct styles of defor-
those accommodating deformation associated with magmatism mation and post-glacial (b 22.5 ka) volcanism. The Troncoso hanging
(e.g., Minor, 1995). This difficulty makes these techniques unable to wall (Fig. 2C) localizes NE-striking normal faults on the lake bottom
characterize short-term (104 years) changes in the regime of roof-rock (Peterson et al., 2018) and numerous, small volume rhyodacitic and an-
deformation and link these changes to the magmatic history of the sys- desitic eruptions (Andersen et al., 2017). In contrast, the footwall of the
tem. The ability to differentiate between tectonic and inflation-driven Troncoso fault (the southeast half of the basin) has few NS -striking, ver-
deformation is key for the detailed structural reconstruction of tical faults in the lake bottom (Peterson et al., 2018) and has localized
deforming rhyolitic systems because it allows one to identify tectonic numerous rhyolitic eruptions, particularly during the late Holocene (b
and magmatic triggers of eruptive activity. 8 ka) (Singer et al., 2014; Andersen et al., 2017; Singer et al., 2018).
Here, we develop and employ a methodology to differentiate the
tectonic and magmatic origins of deformation using the observable 2.2. Post-glacial volcanic history of Laguna del Maule
structural record within the Laguna del Maule (LdM) volcanic field.
We adopt a fully empirical approach based on the statistical analysis The most voluminous Quaternary event within LdM is recorded by
of measured faults, which allows to focus on observable deformation the pre-glacial, 990 ka, rhyodacitic ignimbrite of Cajones de Bobadilla
(strain). Our method simultaneously takes into account fault orienta- (Hildreth et al., 2010; Andersen et al., 2017). This caldera-producing
tion and the offset of each fault, which allows to determine confidence eruption is evidenced on the north half of the basin by 500 m thick ignim-
regions for the geographic direction of maximum strain. We perform brites that have been glacially eroded to the south (Hildreth et al., 2010).
orientation-only and combined orientation-and-heave (the horizontal Following the glacial retreat ~23–19 ka (Singer et al., 2000; Andersen
component of the total offset on a fault) analyses of normal faults at et al., 2017), explosive and effusive activity have dominated LdM in two
the Laguna del Maule (LdM) volcanic field. The orientation-only analy- distinctive pulses of volcanism: an early post glacial (22.5–19.0 ka) and
sis (which assumes dip-slip) tests if the elongation direction at the mo- a middle to late Holocene (b 14.5 ka). During each of the pulses, andes-
ment of fault nucleation is consistent with tectonic elongation or ites, dacites and rhyolites were erupted (Hildreth et al., 2010; Singer
inflation-driven elongation. The orientation-and-heave analysis tests if et al., 2014; Andersen et al., 2017). Since the initial eruption of post-
the horizontal component of slip that has accumulated over thousands glacial rhyolites N19 ka, the locus of voluminous rhyolitic activity has
of years on the faults, i.e., the finite elongation, is consistent tectonic migrated within the basin, whereas small volume rhyodacites and an-
elongation or inflation-driven elongation. We find that normal faults desites are almost exclusively restricted to the Troncoso hanging wall,
have a tectonic origin, yet the maximum elongation direction has an in- on the northwestern edge of the basin (e.g., Singer et al., 2000;
termediate trend between tectonic and inflation elongations. We inter- Hildreth et al., 2010; Andersen et al., 2017).
pret that the Troncoso fault (the main normal structure within LdM) The early post-glacial pulse, spanning 22.5–19.0 ka, was initiated by
and other tectonically induced normal faults in the hanging wall have the N19 ka Laguna del Maule eruption, which extruded 20 km3 of rhyo-
been normally reactivated to accommodate inflation. From our findings, litic material (Andersen et al., 2018; Fierstein, 2018). The exact position
we can relate the structural architecture and detailed deformation his- of the vent is unknown, but is assumed to lie beneath the lake. Following
tory to the composition and localization of eruptive products of the this event, three smaller volume rhyolites (0.5–3 km3 each) were
system. erupted northwest, south, and southeast of the lake (Andersen et al.,
2017; Singer et al., 2018). During this early post-glacial stage, six
2. Geologic setting rhyodacites and eight andesites with cumulative volume b 5 km3
were erupted on the northwestern edge of the lake, where pre-glacial
2.1. Tectonics and volcanism (N23 ka) rhyodacites and andesites had previously been erupted
(Singer et al., 2000; Singer et al., 2014; Andersen et al., 2017).
The LdM volcanic field is located 30 km behind the active Southern During the middle to late Holocene pulse, rhyolitic volcanism was
Volcanic Zone, at latitude 36° S, 30 km east of the Andean front concentrated on the Troncoso footwall. Between 14.5 and 8 ka, rhyolites
(Fig. 1). The post-glacial LdM covers ~300 km2 and has generated were exclusively erupted on the Barrancas complex, on the southeast
 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744 3

Fig. 1. Location map of the Laguna del Maule (LdM) volcanic field in the Southern Volcanic Zone. LOFZ: Liquiñe-Ofqui fault zone. ACFZ: Antiñir-Copahue fault zone.

edge of the basin. After 8 ka, rhyolitic activity continued on the Barran- decrease on elevation that results in a 60 m difference between the
cas complex and expanded to the eastern and southern edges of the high (southeast) and low (northwest) portions of the shoreline. The
lake, with five or more eruptive events distributed throughout the foot- shoreline record of crustal inflation is consistent with the addition of
wall. Coevally with the b8 ka rhyolites, three rhyodacites and a single 13 km3 of magma, leading to the subsequent eruption of the 9 km3 of
andesite were erupted on the Troncoso hanging wall, in the area that lo- rhyolite in the Barrancas Complex (Singer et al., 2018).
calized earlier pre and post-glacial andesitic and rhyodacitic eruptions The second recognized record of inflation started as recently as 2007
(Hildreth et al., 2010; Andersen et al., 2017). and is currently ongoing (Fournier et al., 2010; Feigl et al., 2014). This in-
flation is also centered on the Troncoso footwall, but is adjacent to the
2.3. Magmatically driven inflation trace of the fault, north of the Nieblas rhyolite. Uplift has averaged
rates of ~20 cm/yr that slightly decreased following 2014 (Le Mével
At LdM, magma replenishment in a shallow mush that hosts ephem- et al., 2015). We refer to this record of inflation as geodetic-based infla-
eral rhyolite reservoirs has resulted in two separate inflation records tion. The deformation measured by InSAR is consistent with a shallowly
(Andersen et al., 2017, 2018). The first recorded inflation occurred dipping sill at 5 km depth, growing by ~0.005 km3/yr (Feigl et al., 2014).
b9.4 ka, and was identified by the warping of a paleoshoreline high- In map view, the sill intersects the surface trace of the Troncoso fault
stand (Fig. 2) that defines a doming pattern centered on the southeast and is co-located with a Bouguer gravity anomaly interpreted as
end of the lake, on the Troncoso footwall (Hildreth et al., 2010; Singer crystal-poor rhyolite at 2 km depth within a body of highly crystalline
et al., 2018). We refer to this earlier record as shoreline-based inflation. magma (Miller et al., 2017).
Eruption of the 19.0 ka Espejos rhyolite blocked the northwest outlet of
the lake, rising its level and incising a wave-cut shoreline 200 m above 3. Faulting in Troncoso and LDM
the current lake elevation (Hildreth et al., 2010). Cosmogenic nuclide
(36Cl) dating shows that the high-stand level was abruptly abandoned 3.1. Normal motions and faulting on the Troncoso fault
9.4 ka ago, upon collapse of the Espejos flow (Hildreth et al., 2010;
Andersen et al., 2018; Singer et al., 2018). The originally horizontal The Troncoso canyon is a NE-trending, 5 km long, glacial valley par-
paleoshoreline is now warped, with a progressive northwestward allel to the fault and exposes ~350 m of lapilli and tuffs of the Miocene
4 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744

A) N
n = 63
5 km
n=1 n=4
36.0°S

n = 40
n = 43
Espejos
rhyolite

n = 24 LdM-1
(?)

LdM-2 Laguna
del
LdM-3 Maule

n = 52
Peninsula
andesite
n = 14

LdM-4
geodetic-based
36.1°S

(current)
center of uplift n=6

n = 15

Tr-1 n = 54 Nieblas
rhyolite shoreline-based
(Holocene)
so

center of uplift
ul o
fa ronc

n = 19
t
T

Tr-2 n = 56 Cerro Barrancas

B) C) rhyolitic complex
5 km
2.5-
8 km A
hanging
wall
1.5-7 km footwall
2-6 km
lt

lt
fau

fau
so

so
co

co
n

70.6°S 70.5°W 70.4°W

Tro

A’
Tro

post-glacial eruptions (< 22.5 ka) Fault Faults. See caption Exposed lake
rhyolite andesite Shoreline GPS for color explanation. sediments
rhyodacite measurement Geophysics anomaly/model (InSAR, gravity,
seismic tomography, magnetotellurics). See
pre-glacial ignimbrites Center of uplift
caption for color explanation.
Cajones de Bobadilla (~990 ka) potential vent of
(?) Laguna del Maule normal, reverse, strike-slip and undetermined
rhyolite faults, respectively

Fig. 2. Geologic setting of the LdM volcanic field. A) Faults measured in the LdM basin and Troncoso shown as poles to fault planes in lower equal area projections. Sites of structural analysis
(LdM 1–4, Tr 1–2) denoted with black, square outlines. Paleoshoreline measurements from Singer et al. (2018). In the lake basin, faults are concentrated on the Troncoso hanging wall, an
area of low volume, pre- and post-glacial andesitic and rhyodacitic eruptions. Lake bottom faults mapped by seismic reflection (Peterson et al., 2018.). Bobadilla caldera collapse structures
within the Cajones de Bobadilla ignimbrite are shown as thick black lines (Hildreth et al., 2010). B) Relationship between the Troncoso fault and geophysical anomalies and models. Black
outline: Mogi-source modeled sill at 5 km depth (Feigl et al., 2014); Solid blue outline: low gravity anomaly potentially indicating presence of melt, 2–6 km depth (Miller et al., 2017);
Dashed blue outline: low gravity anomaly interpreted as a crystalline mush, 2–6 km depth (Miller et al., 2017); Pink outline: seismic tomography low shear-wave velocity body,
2–8 km depth (Wespestad et al., 2019). Red outline: magnetotelluric conductors at 2–6 km depth (Cordell et al., 2018). Magnetotelluric and seismic anomalies are interpreted as
crystal-rich mush. C) Schematic representation of the structural style of the Troncoso hanging wall (NW) and footwall (SE). The Troncoso hanging wall is cut by abundant, NE-striking
normal faults. The Troncoso footwall has NS-striking faults (Peterson et al., 2018). A-A′ shows the profile location of the schematic cross-section on Fig. 9.

Trapa-Trapa formation (Fig. 3). The Trapa-Trapa volcanics are uncon- motion on the fault (Fig. 3A). The southeast flank of the canyon exposes
formably covered by basaltic flows of the Plio-Pleistocene lavas of the triangular facets morphologically similar to those produced by active
Cola de Zorro formation. Across the canyon, the ~200 m vertical offset normal faults (Fig. 3B) (e.g., Strak et al., 2011). Similarly, the 9.4 ka
of the Miocene-Pliocene unconformity suggests northwest-side down paleoshoreline shows a northwest-side down offset of ~20 m across
 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744 5

A) looking SW
PPlcz

? PPlcz

200 m
Mtt

B) looking SE
facets

PPlcz

Mtt

Faults Tr-2

C) 2430 X X’ Laguna
Troncoso trace

del
maule
2410
Paleoshoreline

X
elevation (m)

2390

o
X’

lt os
fauronc
Nieblas
2370 rhyolite
Easting (x 105 m)

T
3.57 3.59 3.61 3.63 3.65

Fig. 3. Panoramic and flank views of the Troncoso canyon. A) View from the head of the canyon to the southwest. Dashed yellow line shows the unconformity between the Pleistocene Cola
de Zorro (PPcl) and Miocene Trapa-Trapa (Mtt) formations. The unconformity shows a difference of ~200 m in elevation between the southeast (higher) and northwest (lower) flanks.
B) Southeast flank of the canyon showing faceted surfaces and transverse valleys where fault data was collected. C) Elevation profile of the paleoshoreline on the south of the basin,
modified after Singer et al. (2018). The shoreline abruptly drops 20 m in elevation from east to west, as it crosses the trace of the Troncoso fault. White points in the map inset are
paleoshoreline locations not included in the profile.

the projection of the Troncoso canyon (Fig. 3C; Singer et al., 2018). Ad- The yellow-tan, fine diatomaceous clay beds (1–500 mm thick) are ex-
ditional evidence of northwest-side down kinematics is provided by six posed as lake-level surfaces or vertical walls up to 2 meter high
mesoscopic fault zones at the south end of the canyon. The fault zones (Hildreth et al., 2010), with interbedded lapilli and ash fall tuffs centi-
strike NNE, dipping moderately to the NW and SE, and have fault meters thick. Within the sediments, thicker strata are often lens shaped
cores that are 3–25 cm wide (Fig. 4). Northwest-side down offsets of and interstratified with fine to coarse sands.
~2–8 m cause the staircase-like topography and abrupt downstream The faults within the Troncoso hanging wall are NE-striking, normal
widening observed in the valleys between the triangular facets. discrete slip surfaces organized in arrays (Fig. 4D). Fault spacing is
In the canyon, we documented 152 additional faults. The faults are denser closer to the trace of the Troncoso fault. On the surface, rare
discrete slip planes that are best exposed at the base and walls in trans- fault zones are co-located with and parallel to major faults mapped by
verse valleys on the southeast flank (Fig. 4A). There are multiple gener- seismic reflection in the lake basin (Peterson et al., 2018). The faults in
ations of faulting, as inferred from cross-cutting relations and the the arrays have millimeters to centimeters of offset, and few faults
presence/absence of gypsum and iron oxide mineralization. The fault with tens of centimeters of offset are located at the base of the Nieblas
surfaces often preserve slickenlines, yet the absence of offset markers and Espejos rhyolites. The faults rarely preserve striations or shear
make it difficult to constrain the sense of motion or offset magnitude: sense indicators, so we relied on offset stratigraphic markers to inter-
60% of the faults have unconstrained sense of motion, 17% are normal, pret fault motion. The identification of strike-slip faults was particularly
15% are strike-slip, and 8% are reverse. A first set of normal faults strike challenging given the sub-horizontal attitude of bedding. As a general
NE with moderate to steep dips to the SE, whereas a second set strikes rule, we interpreted faults with steep (N80°) dips and small (b0.5 cm)
NS with moderate dips to the E (Fig. 5A). Strike-slip faults strike NW vertical offset of subhorizontal beds as recording strike-slip movement,
and dip steeply. Reverse faults have a spread of strikes (N-NW-W) but we could rarely determine the sense of slip (dextral vs. sinistral).
with moderate to steep dips to the SW and NE (see Supplement 1 for We documented a total of 229 faults in the sediments, of which 49%
more detailed fault descriptions). are normal, 22% are strike-slip, and 13% are reverse. The remainder
16% have an unknown sense of motion. Normal faults have two main
3.2. Post-glacial faulting in the LdM basin orientations: one set strikes NE-SW with moderate to steep dips to
the NW and SE, and a second, minor set strikes NW and dips steeply
In the LdM basin, young faults are concentrated on the hanging wall to the NE (Fig. 5B). Strike-slip faults have two main orientations:
of the Troncoso fault, and are hosted by subhorizontal, post-glacial 1) EW-striking with steep dips to the N and S; and 2) NS-striking,
(b23 ka, Singer et al., 2000) lake sediments (Fig. 2). In contrast, there with steep dips to the E and W. We infer that these strike-slip faults
are few faults in the sediments exposed in the footwall of the fault. are conjugate sets: simultaneous dextral motions on NS-striking,
6 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744

A) B)

U
D

n=6

C) D)
lin
ea
tio
n

Fig. 4. Faults in the lake basin and Troncoso canyon. A) Fault zones on the southeast flank of the Troncoso canyon with interpreted northwest-side down motion, forming steps. Lower
equal area net shows the orientation of fault zones in the southeast flank of the Troncoso canyon. Picture facing southeast B) Inset in A. Fault core on fault zones causing the step-like
topography. C) Discrete fault-slip planes in the Troncoso canyon. D) Fault arrays in LdM lake sediments are discrete faults with millimeter to centimeter displacements.

vertical faults and sinistral motions on vertical, EW-striking faults methodology, the assumptions underlying the analyses, the determina-
would be consistent with NW-SE elongation recorded by NE-striking tion of representative strain/stress tensors for fault populations, and the
normal faults. Reverse faults have three main orientations: 1) NNE- reliability of the results.
striking with dips to the E; 2) NE-striking with dips to the NW; and Fault inversion results indicate regional NW-SE elongation (Fig. 6).
3) NW-striking with dips to the SW (see Supplement 1). At LdM, and sites Tr-1 and Tr-2, inversions have similarly oriented min-
imum compressive stress axes (σ3), gently plunging to the SE (Table S2
4. Tectonic elongation direction in Supplement). The maximum compressive stress (σ1) at LdM and Tr-1
sites gently plunge to the NE and SW respectively. At the Tr-2 site, the
To determine the direction of tectonic elongation, we performed maximum compressive stress plunges steeply to the NE. Results from
paleostress (e.g., Angelier, 1984) and paleostrain (e.g., Marrett and the inversion of the kinematic axes are generally consistent. The maxi-
Allmendinger, 1992) inversions on faults in the Troncoso canyon (out- mum shortening axes (P) plunge shallowly to the NE at LdM and Tr-1,
side the actively doming area) and within the LdM basin. A single and plunge moderately to the NE at Tr-2. The maximum extension
fault-slip datum includes fault orientation, sense of motion, and stria- axes (T) plunge shallowly to the SE at LdM, moderately to the SE at
tion orientation. Ideally, fault inversions are performed on multiple Tr-1, and shallowly to the NW at Tr-2.
(25+), variedly oriented fault slip data recorded on a single outcrop
(e.g., Sperner and Zweigel, 2010). In the Troncoso canyon, we inverted 5. Orientation analysis
faults from sites Tr-1 and Tr-2 (Fig. 2). Tr-1 is a series of exposures in
close proximity to hydrothermally altered tuff; Tr-2 is a continuous Fault inversions indicate a regional NW-SE extensional component
~200 m exposure of tuff (Tr-2). In the LdM basin analysis, we used faults of deformation. We perform a fault orientation analysis to test if normal
from the entire basin for a single inversion because faults with faults in LdM were developed under a homogeneous tectonic elonga-
slickenlines are rare. We partially alleviated the problems associated tion direction. Under this assumption, the mean orientation of faults at
with data from discontinuous outcrops by incorporating in the inver- different sites should be indistinguishable, and consistent with faults
sion faults hosted in the same lithology (clay-rich sediments) and that predicted by tectonic elongation. We used the R package
showed small offsets and clear kinematic indicators. We avoided faults GeologyGeometry (e.g., Davis and Titus, 2017), available at http://
with any signs of reactivation, or that belonged to older generations www.joshuadavis.us/software/, for all our statistical tests. We first test
(i.e., cross-cut by younger faults or presence of hydrothermal minerali- for distinguishable fault orientations within the basin by bootstrapping
zation). This approach reduced our data to nine faults at LdM, six faults (105 iterations) normal fault orientations of the entire dataset and from
at Tr-1, and nine faults at Tr-2. The strain inversion was conducted by four sites (LdM-1, 2, 3, 4) to generate 95% confidence regions of the
estimating P- (shortening) and T- (extension) axes of incremental mean (see Supplement 3, section 1 for details). In this test, non-
strain using Faultkin (Marrett and Allmendinger, 1990; Allmendinger overlapping confidence regions indicate different mean orientations.
et al., 2012). The paleostress inversions were made with the FSA soft- We perform a second test (via calculation of p-values) to check the
ware (Célérier, 2010). See Supplement 2 for an account of the consistency of fault orientation with elongation directions predicted
 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744 7

A) Tronocoso B) Lake Sediments Lake Sediments Tr-1 Tr-2

faults
normal

n=9 n=7 n=9

n = 25 n = 113

paleostress
σ1 σ2
σ2
strike-slip

σ1
σ3 σ1 σ3
σ2 σ3

paleostrain
n = 24 n = 56 T
P B P
B P
reverse

T
T B

σ1 - P axis σ2 - B axis σ3 - T axis

n = 14 n = 30
Fig. 6. Paleostress (dynamic analysis) and paleostrain (kinematic analysis) results on
lower hemisphere, equal area projections. Each column is a specific site (Lake Sediments
Kamb contours (%) 3 6 9 12 at LdM, Tr-1, Tr-2. The top row shows fault planes with sense of motion of the hanging
wall. Center row shows the orientation of the 10 paleostress tensors with the least mean
Fig. 5. Poles to fault planes measured in LdM and the Troncoso canyon. Lower hemisphere, misfit angle. Selected tensor is represented in red symbols (circle: σ3, triangle: σ2,
equal area stereographic projections with Kamb contour intervals of 3%. A) Faults in the rhombus: σ1). Bottom row shows P (shortening) and T (elongation) axes and the
Troncoso canyon. B) Faults in LdM lake sediments. corresponding Bingham means of the paleostrain analysis.

by tectonics and magmatic inflation. We assume that the tectonic elon-

gation direction is parallel to the σ3 and T axes from fault inversions in the normal faults in the basin. Overall, normal faults in the basin have
Troncoso, and that horizontal elongation caused by inflation is perpen- consistent and indistinguishable strikes.
dicular to the contours defining the surface doming and estimate it by The NW-SE-trending poles of normal faults in the hanging wall are
visual inspection. For shoreline-based inflation (b9.4 ka), we used inter- significatively different (p = 0.00) to the elongation direction predicted
polated elevations of the paleoshoreline (Singer et al., 2018). For the by shoreline-based inflation, whereas it is indistinguishable (p = 0.62)
current geodetic-based inflation, we used InSAR interferograms span- from tectonic elongation (red arc in Fig. 8; Table 1).
ning the longest period that showed a concentric doming pattern
(Feigl et al., 2014; Le Mével et al., 2016).
We calculate p-values from bootstrapped means (105 iterations) of 6. Combined orientation-and-heave analysis
trends of fault poles and use the horizontal elongation directions pre-
dicted by tectonics and inflation as null hypotheses. The p-value is the To assess if magmatic inflation has affected slip on normal fault pop-
probability to observe the data, or data more extreme (i.e., more distant ulations in LdM, we perform a combined orientation-and-heave (hori-
from the null hypothesis), under the assumption that the null hypothe- zontal offset) magnitude analysis to estimate the maximum heave
sis is true (e.g., Greenland et al., 2016). A p-value smaller than the cho- direction assuming dip-slip motion on each normal fault (for details
sen significance level α (typically α ≤ 0.05) results in the rejection of the on the methods followed here, see Supplement 3, sections 2–4). We cal-
null-hypothesis, i.e., the population has a mean different to the null hy- culate a p-value using 95% confidence regions of the maximum heave
pothesis. A p-value larger than α is an inconclusive test and does not direction, which we calculate by weighted bootstrapping (105 itera-
confirm the null hypothesis (see Supplement 3, section 4). We adopt a tions) the trends of poles to normal faults. The weight of each fault is
significance level of α = 0.05. equal to the heave on it, so that the contribution of a fault to the mean
Normal faults at each of the four LdM sites (1, 2, 3, and 4) follow the is proportional to the horizontal offset. This approach is in contrast to
dominant orientation in the basin: they strike NE with moderate to unweighted (equal probability) bootstrapping, where infrequent struc-
steep dips to the NW and SE (Fig. 7). The bootstrapped confidence re- tures with larger offset (e.g., Marrett and Allmendinger, 1992) have an
gions calculated from the entire normal fault dataset and LdM-1, 3, overall smaller contribution to the mean, despite their larger offset. In
and 4 strike NE, and dip steeply to the SE. The confidence regions of sites where the confidence region was wide, and visibly distinct fault
these sites overlap, resulting in the rejection of the null hypothesis populations had pole trends parallel to any of the predicted elongations,
that the populations have different means. The 95% confidence region we subdivided faults according to the predictions. To show that the 95%
calculated at LdM-2 strikes NE and dips to the NW, and does not overlap confidence regions are a reasonable estimate of the maximum elonga-
with the confidence region at LdM-4, which indicates that they are sta- tion direction, we calculate the maximum horizontal elongation direc-
tistically different. The significant difference between the two sites is tion from faults. We performed this procedure for the entire fault
caused by opposite dip directions: The mean strike of the faults at the population and for sites LdM-1, 2, 3, and 4. We subdivided faults in
two sites follows the dominant NE strike observed on the majority of sites LdM-3 and 4 into subsets (3a, 3b, 4a and 4b). Confidence regions,
8 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744

All normal faults LdM - 1 LdM - 2

inflation record and are simultaneously parallel to that predicted by tec-
tonics. Further evidence for activity on the Troncoso fault is the smaller
fault spacing and larger NW-SE directed heaves adjacent to its trace
(LdM-4a; Fig. 2).

7.2. Shoreline-based inflation b9.4 ka

n = 113 n = 26 n = 17 Geological, numerical, and geophysical studies show that during in-
flation, pre-existing tectonic faults are reactivated and localize the
LdM - 3 LdM - 4 Bootstrapped means
uplifting blocks (e.g., Tibaldi and Vezzoli, 1998; Acocella and
Funiciello, 1999; Lindsay et al., 2001; Acocella et al., 2004; Levy et al.,
2018). We argue that inflation-induced fault reactivation is observed
on the Troncoso fault. Larger offsets are observed in hanging wall struc-
tures adjacent to the Troncoso fault (LdM-4a). Moreover, the ~20 m ver-
tical offset of the 9.4 ka shoreline across the Troncoso fault is likely too
large to result solely from tectonic movement ~350 km away from the
n = 22 n = 27 plate boundary. The magnitude of the offset requires that inflation
3 6 9 12 Kamb contours (%)
caused normal slip on the fault, resulting in uplift of the footwall relative
to the hanging wall.
Fig. 7. Orientations of normal faults measured in the lake sediments. Lower hemisphere,
The localization of inflation-driven deformation on pre-existing tec-
equal area projections of poles to normal faults measured in the entire LdM basin and tonic faults likely occurred on the hanging wall of the Troncoso fault.
sites LdM 1–4. The last projection shows the results of the orientation analysis: While the fault orientation analysis shows that faults have a tectonic or-
bootstrapped averages with 95% confidence regions of fault poles. Colors in the igin, the orientation-and-heave analysis (maximum heave direction)
bootstrap correspond to fill colors at each site.
shows that faults striking perpendicular to shoreline-based doming
elongation accommodate more deformation. The maximum heave di-
rection estimated with all the normal faults is parallel to the average
centers, p-values, and null hypotheses for each site are reported in elongation direction expected from shoreline-based inflation b9.4 ka.
Table 1. However, the orientation of the faults is consistent with the tectonic
Confidence regions of the maximum heave directions estimated elongation direction. At three (LdM-1, 2, 4a) of the four sites, the maxi-
from all faults and at LdM-1, 2, 3a and 4a trend NW-SE (Fig. 8). The mum heave direction trends between elongations predicted by tecton-
trend of the maximum heave direction of the full fault dataset is within ics and shoreline-based inflation.
the range of predicted elongation from shoreline-based inflation and is We interpret this pattern as partial overprint of the original tectonic
statistically different from elongation predicted by tectonics (ptectonic = elongation by deformation caused by shoreline-based inflation. This
0.00). At LdM-1, the maximum heave direction trends between the pre- overprint indicates that a subset of the normal faults on the Troncoso
dicted tectonic (ptectonic = 0.26) and shoreline-based inflation hanging wall have slipped as a response to inflation caused by magmatic
(pshoreline = 0.00) elongation directions. This intermediate trend of the injections. In our study area, however, we cannot unambiguously distin-
maximum heave direction between elongation predicted by tectonic guish between faults that were reactivated or formed during doming,
and shoreline-based inflation is also observed at LdM-2 (ptectonic = because the elongation direction from shoreline-based inflation is at a
0.00, pshoreline = 0.07), and at LdM-4a (ptectonic = 0.01, pshoreline = low angle to the elongation direction caused by tectonic extension.
0.00). Maximum heave direction at LdM-3a has a wide confidence re- We propose that in the hanging wall of the Troncoso fault, elongation
gion caused by large, SW-trending heaves, which are possible outliers. caused by shoreline-based inflation b9.4 ka is accommodated, at least
At this site, tectonic elongation is indistinguishable from maximum partially, by reactivation of tectonically formed, NE-striking normal
heave direction (ptectonic = 0.08). faults.
At LdM-3b and 4b, the maximum heave directions trend EW to ENE-
WSW, subparallel to, but statistically different from, the elongation pre- 7.3. Current geodetic-based inflation
dicted by geodetic-based inflation (p = 0.01 and p = 0.00, respectively).
These sites also yield p ≤ 0.01 for elongation predicted by tectonic and The current inflation, starting in 2007, is causing doming at LdM
shoreline-based inflation. (Feigl et al., 2014). The maximum heave direction of faults with large
offsets at LdM-3b and 4b are simultaneously subparallel and statistically
7. Discussion different from elongation predicted by geodetic-based inflation. These
sites are located on the Troncoso hanging wall, adjacent to the main
7.1. Tectonic deformation fault plane and immediately west of the point of maximum uplift. If
the current, high rates of inflation (N20 cm/yr) are accommodated by
In the LdM volcanic field, the normal Troncoso fault is the main faulting, then the faults are expected to have large offsets. We interpret
structure accommodating NW-SE tectonic extension. The subhorizontal faults with large offsets at LdM-3b and 4b as newly formed and accom-
minimum (σ3) paleostress and principal extension (T) axes calculated modating deformation associated to the current inflation. Field studies
at LdM, Tr-1 and Tr-2 show that the LdM basin and the Troncoso canyon and analog models of inflation show that normal faults form in the re-
have experienced NW-SE directed elongation. Because these sites are surgent area to accommodate differential uplift and extension associ-
both inside and outside the volcanic field, we attribute this similarity ated with doming, filling space created by asymmetric block uplift
to regional tectonic deformation. Normal motion across the Troncoso (e.g., Self et al., 1986; Orsi et al., 1991; Acocella et al., 2000).
fault, indicated by northwest-side down offset of the 9.4 ka
paleoshoreline (~20 m) and the Mio-Pleistocene unconformity 7.4. Structural evolution of LdM
(~200 m), accommodates NW-SE elongation. Within the Troncoso
hanging wall (the northwest portion of LdM), NW-SE tectonic elonga- From our orientation-only and orientation-and-heave analyses, we
tion is recorded by the abundant NE-striking normal faults. These faults propose the following model for the volcano-tectonic evolution of
have orientations statistically different to that predicted by either LdM since ~19 ka. Prior to the shoreline-based b9.4 ka inflation record,
 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744 9

N All faults N LdM − 1 N LdM − 2

103 103
102
102 102
101 101 101

W E W E W E

7.7
17.4 16.6
n = 103 S n = 21 S n = 16 S

7.3
N LdM − 3 N LdM − 3a N LdM − 3b
102 102 102
1 1
10 10 101 12.1
W E W E W E

n = 22 S n = 18 S n= 4
6.8 S

N LdM − 4 49.9 N LdM − 4a N LdM − 4b

103 103 103
102 102 2
80.3
10
101 101 101
W E W E W E

50
n = 23 S n = 17 S n= 6 S

KEY predictions data

maximum heave
tectonic direction
elongation N (95%) N
horizontal
shoreline-based maximum 103 offset
(<9.4 ka) elongation 2 on fault
direction 10
inflation-driven (mm)
elongation 101

W E W E
geodetic-based
(current)
inflation-driven Trend of fault pole.
elongation Distance from center population 1
indicates horizontal population 2
S offset S

Fig. 8. Heave azimuth plots showing measured fault heave (horizontal offset) in their corresponding azimuth. In the diagram, each circle is an azimuthal grid. The distance of each point to
the center is proportional to the log10(heave), where heave is in mm, increasing towards the primitive circles. Blue arrows indicate the maximum elongation direction calculated from fault
offset (see Supplement 3); blue numbers show the average heave per fault in the maximum elongation direction. Blue arcs show 95% confidence regions of bootstrapped maximum heave
direction (orientation-and-heave analysis). Red arc on the ‘all fault’ dataset shows the 95% confidence region of trends of faults poles (orientation-only analysis). Orange, green and blue
lines and symbols indicate the elongation directions predicted (i.e., null hypotheses) by tectonics, shoreline-based (b 9.4 ka) inflation and geodetic-based (current) inflation, respectively.
A predicted elongation direction (null hypothesis) that falls within the arc indicates an inconclusive test (p N 0.05); a predicted elongation direction that falls outside the arc indicates a
significative difference (p b 0.05), rejecting the null hypothesis.

NW-SE directed tectonic extension was accommodated by normal slips oriented NE-striking faults were normally reactivated, and possibly
on the Troncoso fault and the NE-striking, normal faults in the hanging new faults were also developed. As pre-existing faults slipped normally
wall (Fig. 9). During magmatically active periods, deformation associ- as a response to inflation, they caused horizontal elongation of the
ated to inflation can be accommodated in the roof-rocks by uplift, hori- hanging wall, possibly resulting in crustal thinning, which (at least par-
zontal elongation, or a combination of the two. The inflation event tially) counteracted uplift. That is, pre-existing faults allowed strain
b9.4 ka, was accommodated as elongation on the hanging wall and as caused by inflation to be accommodated as horizontal extension rather
uplift on the footwall. In the hanging wall, the pre-existing, favorably than uplift. Contrastingly, the Troncoso footwall is largely unfaulted at
10 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744

Table 1
Null hypotheses (H0), confidence regions (CR), and p-values for each of the null hypotheses calculated from the orientation only and combined orientation-and-heave (i.e., maximum
heave direction) analyses of normal faults. p-Values below the significance level (α b 0.05) are in bold.

Orientation only Orientation ± heave

All faults All faults LdM-1 LdM-2 LdM-3 LdM-3a LdM-3b LdM-4 LdM-4a LdM-4b

H0 tectonic 138 138 138 138 138 138 138 138 138 138
H0 shoreline 120 120 112 116 130 130 130 120 120 120
H0 current – – 167 105 95 95 95 52 52 52
CR center 136 119 133 110 135 154 82 110 130 67
CR Angle (95%) 10 8 10 7 47 18 9 18 6 3
ptectonic 0.62 0.00 0.26 0.00 0.85 0.08 0.00 0.00 0.01 0.00
pshoreline 0.00 0.72 0.00 0.07 0.75 0.02 0.00 0.32 0.00 0.00
pgeodetic 0.00 0.00 0.00 0.00 0.07 0.00 0.01 0.00 0.00 0.00

the surface, and the few subsurface faults identified (Peterson et al., Following a period of quiescence, magmatic activity resumed, caus-
2018) are not oriented favorably to have been normally reactivated by ing the current, geodetic-based inflation. Our data show that few nor-
inflation b9.4 ka. Inflation was accommodated primarily by uplift of mal faults are consistent with the elongation predicted by the current
the footwall. During or following b9.4 ka inflation, the differential uplift inflation event, and have strikingly different attitudes and offsets to nor-
between hanging wall and footwall was accommodated by normal slip mal faults interpreted as tectonic or derived from shoreline-based infla-
on the Troncoso fault, lowering the shoreline on the hanging wall tion. This indicates that the current inflation event has nucleated few
(Fig. 3C). Andesites-rhyodacites and rhyolites were erupted on the faults that exhibit the largest offset near the area of maximum uplift,
hanging wall and footwall, respectively. showing that this is the area of maximum deformation. The area of

Troncoso fault
A A’
TIME Troncoso fault
hanging footwall
wall
hanging
>19 ka wall footwall

tectonic elongation
A A’

<19 - 9.4
ka

uplift
inflation-driven
A elongation A’

< 9.4 ka

uplift

A A’
? uplift

present

Fig. 9. Model of the structural evolution of LdM following rhyolitic starting at 19 ka. Prior eruptions are not shown for simplicity. The rightmost set of diagrams show the evolution of
deformation in the LdM on a schematic cross-section. The central set of models schematically highlights the deformation components and shows the state of the crust prior to
deformation as a dashed rectangle. The leftmost bar indicates timing. Between 19 and 9.4 ka, NW-SE tectonic extension is accommodated by normal faulting on the Troncoso hanging
wall. During inflation b9.4 ka, pre-existing tectonic faults are favorably oriented to accommodate elongation on the hanging wall. On the footwall, the lack of faults results in uplift
instead. The differential uplift between hanging wall and footwall is accommodated by normal slip on the Troncoso fault. Rhyodacites and andesites are erupted on the hanging wall;
rhyolites are erupted on the footwall. During the current inflation episode, the orientation of pre-existing faults in the hanging wall is not favorable to accommodate elongation, which
results in the development of new faults to accommodate uplift. The cross-section location is shown in Fig. 2C. Unit colors are the same as in Fig. 2.
 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744 11

maximum deformation (inflation) is adjacent to the Troncoso fault: (Carlino et al., 2014; Cabaniss et al., 2018). The lack of a transition
then, it is possible that the Troncoso fault has experienced or might ex- zone further promotes the propagation of dikes leading to eruptions,
perience normal reactivations caused by the current inflation event. precluding magma accumulation and the resulting inflation. The re-
duced residence times of magmas decreases the potential to differenti-
7.5. Fault control on eruptive composition within silicic systems? ate them, resulting in less felsic liquids (andesites and rhyodacites) that
are erupted.
In LdM, there seems to be a correlation between faulting and post- Early post-glacial (22.5–19 ka) rhyolitic volcanism spanned the en-
glacial rhyodacitic and andesitic volcanism, as well as rhyolitic volca- tire lake basin, yet the rhyolitic activity has only occurred in the
nism b8 ka. In this section, we speculate that faulting style is a causal Troncoso footwall after 14.5 ka. Findings from geodesy, geophysics
mechanism for the observed spatial patterns of erupted composition. and geomorphology suggest that the footwall seems to have localized
We first discuss how the structural architecture of the Troncoso hanging magmatic injections causing surface inflation potentially leading to
wall may be related to rhyodacitic and andesitic volcanism; we follow the eruption of rhyolites. The inflating subhorizontal sill at ~5 km
this with the discussion of rhyolitic volcanism on the Troncoso footwall. depth, inferred from InSAR deformation, is located in the Troncoso foot-
Crystal-rich mush zones derived from seismic tomography and wall (Feigl et al., 2014; Le Mével et al., 2015). The low Bouguer gravity
magnetotellurics are hosted in the Troncoso hanging wall, beneath anomaly at 2 km depth suggesting a crystal-rich mush with potential
rhyodacitic and andesitic volcanics, and follow the fault at depth. The presence of melt is also in the footwall (Miller et al., 2017). Last, the
450 km3 low shear-wave velocity body is interpreted as a crystal-rich warped, 9.4 ka shoreline is consistent with the addition of 13 km3 of
mush (5% melt) that is adjacent to the fault, strikes parallel to it, and magma ~7 km below the Cerro Barrancas complex, in the Troncoso foot-
dips to the NW, spanning 2–8 km depth (Wespestad et al., 2019). This wall (Singer et al., 2018).
seismic velocity zone partially overlaps with a NW-dipping During the late Holocene (b 8 ka), the limited faulting in the
magnetotelluric conductor at 2–6 km depth (C3, Cordell et al., 2018) Troncoso footwall might favor the accumulation of melt-rich rhyolite
on the northwest edge of the lake. This conductor is continuous with a pockets within the reservoir. In the footwall, the absence of extensional
second, deeper conductive body (C4 in Cordell et al., 2018) spanning structures reduces the vertical permeability of the crust to silicic
6–12 km depth that also dips to the NW. Both conductors are magmas. Magmatic injections beneath the unfaulted footwall remain
interpreted as crystal-rich regions with interstitial melt (Cordell et al., within the mush for longer periods. This scenario promotes the devel-
2018). opment of a thick transition zone that further arrests rising magmas,
We speculate that crustal extension and normal faulting within the causing them to accumulate, resulting in inflation (e.g., Galetto et al.,
Troncoso hanging wall preclude magma accumulation, instead favoring 2017). The increased residence times can produce melt-rich bodies
recurrent, low volume rhyodacitic and andesitic eruptions and leaving where crystal-poor rhyolite is segregated: trace element data suggests
behind a crystal-rich mush. Pre and post-glacial rhyodacitic and andes- that LdM rhyolites are possibly derived from long-lived hot zones
itic eruptions within LdM have low volumes and are localized primarily within a larger, colder mush (Andersen et al., 2019). The rhyolite
on the heavily faulted Troncoso hanging wall. During the early post- pockets in the footwall can reach an eruptive state by overpressure de-
glacial (22.5–19.0 ka) magmatic pulse, five rhyodacites and eight andes- rived from volatile exsolution (Pistone et al., 2013; Andersen et al.,
ites were erupted on the hanging wall, on the west edge of the lake. 2018, 2019), or triggered by surface-to-reservoir fractures that accom-
Contrastingly, only one pre-glacial rhyodacite (20 ka) was erupted on modate crustal block uplift (e.g., Jellinek and DePaolo, 2003; de Silva
the footwall of the fault (Singer et al., 2000; Hildreth et al., 2010; et al., 2006; Gottsmann et al., 2009; Gregg et al., 2012; Kennedy et al.,
Singer et al., 2014; Andersen et al., 2017). During the middle to late Ho- 2012; de Silva and Gregg, 2014; Gregg et al., 2015; Tramontano et al.,
locene, three rhyodacites and a single andesite, all b8 ka, were erupted 2017). We favor a fault-limited footwall block as the explanation for
on the Troncoso hanging wall, and are now adjacent to earlier andesitic the localization of late Holocene (b 8 ka) rhyolites on the southeast
and rhyodacitic eruptions. The rhyodacitic and andesitic units have rel- side of the LdM volcanic field.
atively low volumes of 0.5–3 km3 and are in contrast with the larger ac-
cumulations of rhyolite (N 9 km3) that have preferentially erupted on
the footwall since 14.5 ka (Andersen et al., 2017; Singer et al., 2018). 8. Conclusions
The numerous normal faults on the Troncoso hanging wall are a re-
cord of active horizontal extension, and are spatially associated to nu- In the LdM volcanic field, the Troncoso normal fault controls the
merous, small-volume rhyodacitic and andesitic eruptions. This structural architecture of the basin. We have statistically analyzed the
association leads us to hypothesize that within LdM, normal faults slip- orientation and maximum heave direction of normal faults in the hang-
ping as a result of active horizontal extension promote dike propagation ing wall and compared them to horizontal elongation directions pre-
leading to frequent, less felsic, small-volume eruptions simultaneously dicted by tectonics and magmatic inflation. The orientation analysis
suppressing uplift. Stagnant mushes within the shallow crust progres- shows that normal faults have a tectonic origin, as fault orientations
sively develop a transition zone (a viscosity gradient within a magma within the LdM volcanic field are consistent with tectonic elongation.
reservoir) that has an increasingly lower viscosity contrast with the The orientation-and-heave analysis shows that normal faults in the
newly injected magmas (Galetto et al., 2017). This transition zone ar- hanging wall are affected by inflation. The maximum heave direction
rests dikes rising through the mush, promoting resurgence and surface of the entire normal fault population is consistent with that expected
inflation, leading to increased residence times potentially facilitating from shoreline-based inflation b9.4 ka. At individual sites, the maxi-
magmatic differentiation (e.g., Reid et al., 1997; Andersen et al., 2019). mum elongation direction has an intermediate trend between the elon-
Slipping faults on the Troncoso hanging wall may have aided in the fre- gations predicted from tectonics and shoreline-based inflation. We
quent extraction of magmas from the mush, and the mush fails to de- interpret our results as an indication that some of the pre-existing, tec-
velop a transition zone. The faults may have been used as conduits to tonic normal faults in the hanging wall and the Troncoso fault itself
evacuate overpressurized magmas (e.g., Wicks et al., 2011), sheared were reactivated to accommodate elongation and uplift caused by infla-
the mush causing dilatancies that collect and transport melt (Geshi, tion. A minor population of newly developed faults are consistent with
2001), or produced shear thinning by localizing strain in the crystalliz- the elongation direction predicted by ongoing, geodetic-based inflation,
ing mush thereby facilitating magma transport (Picard et al., 2013). Ad- which is centered in a different part of the Troncoso footwall than the
ditionally, numerical models suggest that normal faults favor eruptions, shoreline-based inflation. From these results, we derive a model of the
as tectonic extension may reduce the times at which the roof-rocks are structural evolution of LdM that separates instances of tectonic and
stable (100 to 1000 years) and promote melt transport to the surface inflation-driven surface deformation, showing that tectonic and
12 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744

inflation-driven deformation can be differentiated from the structural Andersen, N.L., Singer, B.S., Costa, F., Fournelle, J., Herrin, J.S., Fabbro, G.N., 2018.
Petrochronologic perspective on rhyolite volcano unrest at Laguna del. Earth Planet.
record. Sci. Lett. 493, 57–70. https://doi.org/10.1016/j.epsl.2018.03.043.
We speculate that the subsurface magmatic conditions (crystal-rich Andersen, N.L., Singer, B.S., Coble, M.A., 2019. Repeated rhyolite eruption from heteroge-
mush underneath the hanging wall and melt-rich pockets in the foot- neous hot zones embedded within a cool, shallow magma reservoir. J. Geophys. Res.
Solid Earth, 0–3 https://doi.org/10.1029/2018JB016418.
wall) are linked to the structural architecture of the Troncoso fault, Angelier, J., 1984. Tectonic analysis of fault slip data sets. J. Geophys. Res. Solid Earth 89,
and are reflected on eruptive products at the surface. The largely 5835–5848.
unfaulted footwall concentrates b14.5 ka rhyolitic volcanism, whereas Barcelona, H., Yagupsky, D., Vigide, N., Senger, M., 2019. Structural model and slip-dilation
tendency analysis at the Copahue geothermal system: Inferences on the reservoir ge-
the heavily faulted hanging wall has localized post-glacial (b 22.5 ka) ometry. J. Volcanol. Geotherm. Res. https://doi.org/10.1016/j.jvolgeores.2019.03.007.
rhyodacitic and andesitic volcanism. Extension and faulting in the hang- Bolós, X., Martí, J., Becerril, L., Planagumà, L., Grosse, P., Barde-Cabusson, S., 2015. Volcano-
ing wall facilitates transport of rhyodacitic and andesitic magmas to the structural analysis of La Garrotxa Volcanic Field (NE Iberia): Implications for the
plumbing system. Tectonophysics 642, 58–70. https://doi.org/10.1016/j.
surface, whereas the comparatively fewer faults and overall lack of ex-
tecto.2014.12.013.
tension in the Troncoso footwall may promote accumulation of Cabaniss, H.E., Gregg, P.M., Grosfils, E.B., 2018. The role of tectonic stress in triggering
crystal-poor rhyolites during the late Holocene. This model explains si- large silicic caldera eruptions. Geophys. Res. Lett. 45, 3889–3895. https://doi.org/
multaneous magmatic injection, uplift, and compositional diversity of 10.1029/2018GL077393.
Carlino, S., Tramelli, A., Somma, R., 2014. Caldera subsidence in extensional tectonics. Bull.
eruptive products linking them to the structural architecture of the Volcanol. 76, 1–7. https://doi.org/10.1007/s00445-014-0870-2.
system. Célérier, B., 2010. FSA: Fault & Stress Analysis Software. version 33.7. http://www.
Supplementary data to this article can be found online at https://doi. bcelerier.univ-montp2.fr/software/dcmt/fsa/fsa.html.
Coleman, D.S., Gray, W., Glazner, A.F., 2004. Rethinking the emplacement and evolution of
org/10.1016/j.jvolgeores.2019.106744. zoned plutons: geochronologic evidence for incremental assembly of the Tuolumne
Intrusive Suite, California. Geology 32, 433. https://doi.org/10.1130/G20220.1.
Corazzato, C., Tibaldi, A., 2006. Fracture control on type, morphology and distribution of
parasitic volcanic cones: an example from Mt. Etna, Italy. J. Volcanol. Geotherm.
CRediT authorship contribution statement Res. 158, 177–194. https://doi.org/10.1016/j.jvolgeores.2006.04.018.
Cordell, D., Unsworth, M.J., Díaz, D., 2018. Imaging the Laguna del Maule Volcanic Field,
N. Garibaldi: Conceptualization, Methodology, Writing - original central Chile using magnetotellurics: evidence for crustal melt regions laterally-
offset from surface vents and lava flows. Earth Planet. Sci. Lett. 488, 168–180.
draft, Formal analysis. B. Tikoff: Conceptualization, Writing - review & https://doi.org/10.1016/j.epsl.2018.01.007.
editing, Funding acquisition. D. Peterson: Conceptualization, Writing - Costa, A., Gottsmann, J., Melnik, O., Sparks, R.S.J., 2011. A stress-controlled mechanism for
review & editing. J.R. Davis: Methodology, Software, Writing - review the intensity of very large magnitude explosive eruptions. Earth Planet. Sci. Lett. 310,
161–166. https://doi.org/10.1016/j.epsl.2011.07.024.
& editing, Formal analysis. K. Keranen: Writing - review & editing.
Davis, J.R., Titus, S.J., 2017. Modern methods of analysis for three-dimensional orienta-
tional data. J. Struct. Geol. 96, 65–89. https://doi.org/10.1016/j.jsg.2017.01.002.
De Silva, S., Zandt, G., Trumbull, R., Viramonte, J.G., Salas, G., Jiménez, N., 2006. Large ig-
Declaration of competing interest nimbrite eruptions and volcano-tectonic depressions in the Central Andes: a
thermomechanical perspective. Geol. Soc. Lond., Spec. Publ. 269, 47–63. https://doi.
org/10.1144/GSL.SP.2006.269.01.04.
The authors declare that they have no known competing financial Farris, D.W., Cardona, A., Montes, C., Foster, D., Jaramillo, C., 2017. Magmatic evolution of
interests or personal relationships that could have appeared to influ- Panama Canal volcanic rocks: a record of arc processes and tectonic change. PLoS One
https://doi.org/10.1371/journal.pone.0176010.
ence the work reported in this paper. Feigl, K.L., Le Mevel, H., Tabrez Ali, S., Cordova, L., Andersen, N.L., DeMets, C., Singer, B.S., 2014.
Rapid uplift in Laguna del Maule volcanic field of the Andean Southern Volcanic zone
Acknowledgements (Chile) 2007-2012. Geophys. J. Int. 196, 885–901. https://doi.org/10.1093/gji/ggt438.
Fierstein, J., 2018. Postglacial Eruptive History Established By Mapping and Tephra Stratig-
raphy Provides Perspectives on Magmatic System Beneath Laguna del Maule, Chile.
This research was funded by National Science Foundation (NSF) Presented at. AGU Chapman Conference on Merging Geophysical, Petrochronologic,
grant EAR-1411779. We thank B. Singer, C. Cardona, N. Andersen, H. and Modeling Perspectives of Large Silicic Magma Systems, Quinamavida, Chile.
Fournier, T.J., Pritchard, M.E., Riddick, S.N., 2010. Duration, magnitude, and frequency of
Le Mével, C. Miller, D. Cordell, A. Tassara, N. Roberts, R. Williams and
subaerial volcano deformation events: New results from Latin America using InSAR
the UW-Madison Tectonics and Structure group for helpful discussions. and a global synthesis. Geochem. Geophys. Geosyst. 11. https://doi.org/10.1029/
L. Torres, J. Torres, J. Martínez, and D. Sellés are thanked for fieldwork as- 2009GC002558.
sistance. A. Gudmundsson and P. Gregg are thanked for helpful reviews. Galetto, F., Acocella, V., Caricchi, L., 2017. Caldera resurgence driven by magma viscosity
contrasts. Nat. Commun. 8, 1–11. https://doi.org/10.1038/s41467-017-01632-y.
García-Palomo, A., Macías, J.L., Espíndola, J.M., 2004. Strike-slip faults and K-alkaline vol-
References canism at El Chichón volcano, southeastern Mexico. J. Volcanol. Geotherm. Res. 136,
247–268. https://doi.org/10.1016/j.jvolgeores.2004.04.001.
Acocella, V., Funiciello, R., 1999. The interaction between regional and local tectonics dur- Geshi, N., 2001. Melt segregation by localized shear deformation and fracturing during
ing resurgent doming: the case of the island of Ischia, Italy. J. Volcanol. Geotherm. Res. crystallization of magma in shallow intrusions of the Otoge volcanic complex, central
88, 109–123. https://doi.org/10.1016/S0377-0273(98)00109-7. Japan. J. Volcanol. Geotherm. Res. 106, 285–300. https://doi.org/10.1016/S0377-0273
Acocella, V., Cifelli, F., Funiciello, R., 2000. Analogue models of collapse calderas and resur- (00)00271-7.
gent domes. J. Volcanol. Geotherm. Res. 104, 81–96. https://doi.org/10.1016/S0377- Goto, Y., Mcphie, J., 2018. Tectonics, Structure, and Resurgence of the Largest Quaternary
0273(00)00201-8. Caldera in Japan: Kutcharo, Hokkaido. pp. 1–16.
Acocella, V., Funiciello, R., Marotta, E., Orsi, G., de Vita, S., 2004. The role of extensional Gottsmann, J., Lavallée, Y., Martí, J., Aguirre-Díaz, G., 2009. Magma–tectonic interaction
structures on experimental calderas and resurgence. J. Volcanol. Geotherm. Res. and the eruption of silicic batholiths. Earth Planet. Sci. Lett. https://doi.org/10.1016/
129, 199–217. https://doi.org/10.1016/S0377-0273(03)00240-3. j.epsl.2009.05.008.
Aguirre-Díaz, G.J., Labarthe-Hernández, G., 2003. Fissure ignimbrites: fissure-source ori- Greenland, S., Senn, S.J., Rothman, K.J., Carlin, J.B., Poole, C., Goodman, S.N., Altman, D.G.,
gin for voluminous ignimbrites of the Sierra Madre Occidental and its relationship 2016. Statistical tests, P values, confidence intervals, and power: a guide to misinter-
with Basin and Range faulting. Geology 31, 773–776. https://doi.org/10.1130/ pretations. Eur. J. Epidemiol. 31, 337–350. https://doi.org/10.1007/s10654-016-0149-3.
G19665.1. Gregg, P.M., De Silva, S.L., Grosfils, E.B., Parmigiani, J.P., 2012. Catastrophic caldera-forming
Allan, A.S.R., Wilson, C.J.N., Millet, M.A., Wysoczanski, R.J., 2012. The invisible hand: tec- eruptions: thermomechanics and implications for eruption triggering and maximum
tonic triggering and modulation of a rhyolitic supereruption. Geology 40, 563–566. caldera dimensions on Earth. J. Volcanol. Geotherm. Res. 241–242, 1–12. https://doi.
https://doi.org/10.1130/G32969.1. org/10.1016/j.jvolgeores.2012.06.009.
Allan, A.S.R., Barker, S.J., Millet, M.A., Morgan, D.J., Rooyakkers, S.M., Schipper, C.I., Wilson, Gregg, P.M., Grosfils, E.B., de Silva, S.L., 2015. Catastrophic caldera-forming eruptions II:
C.J.N., 2017. A cascade of magmatic events during the assembly and eruption of a the subordinate role of magma buoyancy as an eruption trigger. J. Volcanol.
super-sized magma body. Contrib. Mineral. Petrol. 172, 1–34. https://doi.org/ Geotherm. Res. 305, 100–113. https://doi.org/10.1016/j.jvolgeores.2015.09.022.
10.1007/s00410-017-1367-8. Gudmundsson, A., 2006. How local stresses control magma-chamber ruptures, dyke in-
Allmendinger, R.W., Cardozo, N.C., Fisher, D., 2012. Structural Geology Algorithms: Vec- jections, and eruptions in composite volcanoes. Earth Sci. Rev. 79 (1–2), 1–31.
tors & Tensors: Cambridge. Cambridge University Press, England (289 pp). Hildreth, W., Godoy, E., Fierstein, J., Singer, B., 2010. Laguna del Maule Volcanic Field: Erup-
Andersen, N.L., Singer, B.S., Jicha, B.R., Beard, B.L., Johnson, C.M., Licciardi, J.M., 2017. Pleis- tive History of a Quaternary Basalt-to-Rhyolite Distributed Volcanic Field on the An-
tocene to holocene growth of a large upper crustal rhyolitic magma reservoir beneath dean Range Crest in Central Chile. Servicio Nacional de Geología y Minería-Chile.
the active Laguna del Maule Volcanic Field, Central Chile. J. Petrol. 58, 85–114. https:// Hughes, G.R., Mahood, G.A., 2008. Tectonic controls on the nature of large silicic calderas
doi.org/10.1093/petrology/egx006. in volcanic arcs. Geology 36, 627–630. https://doi.org/10.1130/G24796A.1.
 N. Garibaldi et al. / Journal of Volcanology and Geothermal Research 389 (2020) 106744 13

Hughes, G.R., Mahood, G.A., 2011. Silicic calderas in arc settings: characteristics, distribu- Picard, D., Arbaret, L., Pichavant, M., Champallier, R., Launeau, P., 2013. The rheological
tion, and tectonic controls. Geol. Soc. Am. Bull. 123, 1577–1595. https://doi.org/ transition in plagioclase-bearing magmas. J. Geophys. Res. Solid Earth 118,
10.1130/B30232.1. 1363–1377. https://doi.org/10.1002/jgrb.50091.
Isaia, R., Vitale, S., Di Giuseppe, M.G., Iannuzzi, E., Tramparulo, F.D.A., Troiano, A., 2015. Stra- Pistone, M., Caricchi, L., Ulmer, P., Reusser, E., Ardia, P., 2013. Rheology of volatile-bearing
tigraphy, structure, and volcano-tectonic evolution of Solfatara maar-diatreme (Campi crystal mushes: mobilization vs. viscous death. Chem. Geol. 345, 16–39. https://doi.
Flegrei, Italy). Bull. Geol. Soc. Am. 127, 1485–1504. https://doi.org/10.1130/B31183.1. org/10.1016/j.chemgeo.2013.02.007.
Jellinek, A.M., DePaolo, D.J., 2003. A model for the origin of large silicic magma chambers: Puskas, C.M., Smith, R.B., Meertens, C.M., Chang, W.L., 2007. Crustal deformation of the
precursors of caldera-forming eruptions. Bull. Volcanol. 65, 363–381. https://doi.org/ Yellowstone–Snake River Plain volcano-tectonic system: campaign and continuous
10.1007/s00445-003-0277-y. GPS observations, 1987–2004. J. Geophys. Res. Solid Earth 112.
Kennedy, B., Wilcock, J., Stix, J., 2012. Caldera resurgence during magma replenishment Reid, M., Coath, C., Harrison, T., McKeegan, K.D., 1997. Prolonged residence times for the
and rejuvenation at Valles and Lake City calderas. Bull. Volcanol. 74, 1833–1847. youngest rhyolites associated with Long Valley Caldera: 230Th—238U ion micro-
https://doi.org/10.1007/s00445-012-0641-x. probe dating of young zircons. Earth Planet. Sci. Lett. 150 (1–2), 27–39.
Kilburn, C.R.J., De Natale, G., Carlino, S., 2017. Progressive approach to eruption at Campi Robertson, R.M., Kilburn, C.R.J., 2016. Deformation regime and long-term precursors to
Flegrei caldera in southern Italy. Nat. Commun. 8, 1–8. https://doi.org/10.1038/ eruption at large calderas: Rabaul, Papua New Guinea. Earth Planet. Sci. Lett. 438,
ncomms15312. 86–94. https://doi.org/10.1016/j.epsl.2016.01.003.
Le Mével, H., Feigl, K.L., Córdova, L., DeMets, C., Lundgren, P., 2015. Evolution of unrest at Rowland, J.V., Sibson, R.H., 2001. Extensional fault kinematics within the Taupo Volcanic
Laguna del Maule volcanic field (Chile) from InSAR and GPS measurements, 2003 to Zone, New Zealand: Soft-linked segmentation of a continental rift system. N. Z.
2014. Geophys. Res. Lett. 42, 6590–6598. https://doi.org/10.1002/2015GL064665. J. Geol. Geophys. 44, 271–283. https://doi.org/10.1080/00288306.2001.9514938.
Le Mével, H., Gregg, P.M., Feigl, K.L., 2016. Magma injection into a long-lived reservoir to de Saint Blanquat, M., Horsman, E., Habert, G., Morgan, S., Vanderhaeghe, O., Law, R.,
explain geodetically measured uplift: application to the 2007–2014 unrest episode at Tikoff, B., 2011. Multiscale magmatic cyclicity, duration of pluton construction, and
Laguna del Maule volcanic field, Chile. J. Geophys. Res. Solid Earth 121, 6092–6108. the paradoxical relationship between tectonism and plutonism in continental arcs.
https://doi.org/10.1002/2016JB013066. Tectonophysics 500, 20–33. https://doi.org/10.1016/j.tecto.2009.12.009.
Levy, S., Bohnenstiehl, D.R., Sprinkle, P., Boettcher, M.S., Wilcock, W.S.D., Tolstoy, M., Schöpa, A., Annen, C., 2013. The effects of magma flux variations on the formation and
Waldhauser, F., 2018. Mechanics of fault reactivation before, during, and after the lifetime of large silicic magma chambers. J. Geophys. Res. Solid Earth 118, 926–942.
2015 eruption of Axial Seamount. Geology 46, 447–450. https://doi.org/10.1130/ https://doi.org/10.1002/jgrb.50127.
G39978.1. Self, S., Goff, F., Gardner, J.N., Ight, J.V.W., Kite, W.M., 1986. Explosive rhyolitic volcanism in
Lindsay, J.M., De Silva, S., Trumbull, R., Emmermann, R., Wemmer, K., 2001. La Pacana cal- the Jemez Mountains: vent locations, caldera development and relation to regional
dera, N. Chile: a re-evaluation of the stratigraphy and volcanology of one of the structure. J. Geophys. Res. 91, 1779–1798.
world’s largest resurgent calderas. J. Volcanol. Geotherm. Res. 106, 145–173. de Silva, S.L., Gregg, P.M., 2014. Thermomechanical feedbacks in magmatic systems: Im-
https://doi.org/10.1016/S0377-0273(00)00270-5. plications for growth, longevity, and evolution of large caldera-forming magma reser-
Magee, C., Driscoll, B.O., Chambers, A.D., 2017. Crystallization and textural evolution of a voirs and their supereruptions. J. Volcanol. Geotherm. Res. 282, 77–91. https://doi.
closed-system magma chamber: insights from a crystal size distribution study of org/10.1016/j.jvolgeores.2014.06.001.
the Lilloise layered intrusion, East Greenland. 147, 363–379. https://doi.org/ Singer, B., Hildreth, W., Vincze, Y., 2000. 40 Ar/39 Ar evidence for early deglaciation of the
10.1017/S0016756809990689. central Chilean Andes. Geophys. Res. Lett. 27, 1663–1666. https://doi.org/10.1029/
Marrett, R., Allmendinger, R.W., 1990. Kinematic analysis of fault-slip data. J. Struct. Geol. 1999GL011065.
12, 973–986. https://doi.org/10.1016/0191-8141(90)90093-E. Singer, B.S., Andersen, N.L., Le Mével, H., Feigl, K.L., DeMets, C., Tikoff, B., Thurber, C.H.,
Marrett, R., Allmendinger, R.W., 1992. Amount of extension on “small” faults: an example Jicha, B.R., Cardona, C., Córdova, L., Gil, F., Unsworth, M.J., Williams-Jones, G., Miller,
from the Viking graben. Geology 20, 47–50. https://doi.org/10.1130/0091-7613 C., Fierstein, J., Hildreth, W., Vazquez, J., 2014. Dynamics of a large, restless, rhyolitic
(1992)020b0047:AOEOSFN2.3.CO;2. magma system at Laguna del Maule, southern Andes, Chile. GSA Today, 4–10
Menand, T., Annen, C., de Saint Blanquat, M., 2015. Rates of magma transfer in the crust: https://doi.org/10.1130/GSATG216A.1.
insights into magma reservoir recharge and pluton growth. Geology 43, 199–202. Singer, B.S., Le Mével, H., Licciardi, J.M., Córdova, L., Tikoff, B., Garibaldi, N., Andersen, N.L.,
https://doi.org/10.1130/G36224.1. Diefenbach, A.K., Feigl, K.L., 2018. Geomorphic expression of rapid Holocene silicic
Miller, C.A., Williams-Jones, G., Fournier, D., Witter, J., 2017. 3D gravity inversion and ther- magma reservoir growth beneath Laguna del Maule, Chile. Sci. Adv. 4, 1–11.
modynamic modelling reveal properties of shallow silicic magma reservoir beneath https://doi.org/10.1126/sciadv.aat1513.
Laguna del Maule, Chile. Earth Planet. Sci. Lett. 1, 1–14. https://doi.org/10.1016/j. Sperner, B., Zweigel, P., 2010. A plea for more caution in fault-slip analysis. Tectonophysics
epsl.2016.11.007. 482, 29–41. https://doi.org/10.1016/j.tecto.2009.07.019.
Minor, S.A., 1995. Superposed local and regional paleostresses: fault-slip analysis of Neo- Strak, V., Dominguez, S., Petit, C., Meyer, B., Loget, N., 2011. Interaction between normal
gene extensional faulting near coeval caldera complexes, Yucca Flat, Nevada. fault slip and erosion on relief evolution: insights from experimental modelling.
J. Geophys. Res. Solid Earth 100, 10507–10528. https://doi.org/10.1029/95jb00078. Tectonophysics 513, 1–19. https://doi.org/10.1016/j.tecto.2011.10.005.
Orsi, G., Gallo, G., Zanchi, A., 1991. Simple-shearing block resurgence in caldera depres- Tadini, A., Bonali, F.L., Corazzato, C., Cortés, J.A., Tibaldi, A., Valentine, G.A., 2014. Spatial
sions. A model from Pantelleria and Ischia. J. Volcanol. Geotherm. Res. 47, 1–11. distribution and structural analysis of vents in the lunar crater volcanic field (Nevada,
https://doi.org/10.1016/0377-0273(91)90097-J. USA). Bull. Volcanol. 76, 1–15. https://doi.org/10.1007/s00445-014-0877-8.
Pérez-Flores, P., Veloso, E., Cembrano, J., Sánchez-Alfaro, P., Lizama, M., Arancibia, G., Tibaldi, A., Vezzoli, L., 1998. The space problem of caldera resurgence: an example from
2017. Fracture network, fluid pathways and paleostress at the Tolhuaca geothermal Ischia Island, Italy. Geol. Rundsch. 87, 53–66.
field. J. Struct. Geol. 96, 134–148. https://doi.org/10.1016/j.jsg.2017.01.009. Tramontano, S., Gualda, G.A.R., Ghiorso, M.S., 2017. Internal triggering of volcanic erup-
Peterson, D.E., Keranen, K.M., Garibaldi, N., Lara, L., Tikoff, B., Tassara, A., Thurber, C.H., tions: tracking overpressure regimes for giant magma bodies. Earth Planet. Sci. Lett.
Lanza, F., 2018. Pulsed Magmatic Inflation and Deflation in a Silicic System Facilitated 472, 142–151. https://doi.org/10.1016/j.epsl.2017.05.014.
By Trapdoor Faulting: Laguna del Maule, Chile. AGU Fall Meeting Abstracts. Wespestad, C.E., Thurber, C.H., Andersen, N.L., Singer, B.S., Cardona, C., Zeng, X.,
Petrinovic, I.A., Martí, J., Aguirre-Díaz, G.J., Guzmán, S., Geyer, A., Paz, N.S., 2010. The Cerro Bennington, N.L., Keranen, K., Peterson, D.E., Cordell, D., Unsworth, M., Miller, C.,
Aguas Calientes caldera, NW Argentina: an example of a tectonically controlled, poly- Williams-Jones, G., 2019. Magma reservoir below Laguna del Maule volcanic field,
genetic collapse caldera, and its regional significance. J. Volcanol. Geotherm. Res. 194, Chile imaged with surface-wave tomography. J. Geophys. Res. Solid Earth, 0–3
15–26. https://doi.org/10.1016/j.jvolgeores.2010.04.012. https://doi.org/10.1029/2018JB016485.
Pfiffner, O.A., Ramsay, J.G., 1982. Constraints on geological strain rates: arguments from Wicks, C., de la Llera, J.C., Lara, L.E., Lowenstern, J., 2011. The role of dyking and fault con-
finite strain states of naturally deformed rocks. J. Geophys. Res. Solid Earth 87 (B1), trol in the rapid onset of eruption at Chaitén volcano, Chile. Nature 478, 374–377.
311–321. https://doi.org/10.1038/nature10541.