Bull Volcanol (2017) 79:76 DOI 10.1007/s00445-017-1162-4

RESEARCH ARTICLE

Diverse dynamics of Holocene mafic-intermediate Plinian eruptions at Mt. Taranaki (Egmont), New Zealand Rafael Torres-Orozco 1

&

Shane J. Cronin 1,2 & Magret Damaschke 1 & Natalia Pardo 3

Received: 10 February 2017 / Accepted: 1 October 2017 # Springer-Verlag GmbH Germany 2017

Abstract Over the last 5000 years, at least 53 eruptive episodes have occurred at Mt. Taranaki (western North Island, New Zealand), from either its summit crater (~ 2500 m) or a satellite vent on Fanthams Peak (~ 1900 m). The magmas erupted have a wide range of compositions from basaltic to trachy-andesitic (~ 48–60 wt% SiO2). Five large-magnitude episodes from this sequence were studied so as to characterize a typical range of explosive eruption styles at andesitic stratovolcanoes, including three eruptions from the summit crater and two from Fanthams Peak. Sustained eruption columns characterized the climactic phase of all five eruptions, but these were interspersed with pulsating, collapsing, or oscillating conditions. Eruption columns reached between 14 to 29 km in height and ejected minimum volumes of 0.1– 1.1 km3 at mass discharge rates of 1 × 107–2 × 108 kg/s, indicating magnitudes of 4.1 to 5.1. The simplest eruptions occurred from Fanthams Peak with basaltic magmas producing high-climactic eruption columns rapidly after vent open-

ing, followed by gentle waning phases or a passage into a lava-fountaining phase. Eruptions of higher-silica magmas at the summit vent, by contrast, showed longer pre-climactic eruptive phases with either dome growth or complex phases of vent clearance and blockage producing unsteady eruption columns. The latter eruption types produced block-and-ash flows, lateral-blast surges, and column-collapse pumice-andash flows, with run-out distances of 3–19 km, covering 5– 70 km2 with volumes of up to 0.022 km3. Our results demonstrate that very different eruption scenarios may occur at different vent locations, or with subtly different compositions erupted, on the same stratovolcano so that emergency management planning must take such a range of possibilities into account.

Editorial responsibility: C. Bonadonna

Introduction

Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00445-017-1162-4) contains supplementary material, which is available to authorized users. * Rafael Torres-Orozco [email protected]

1

Volcanic Risk Solutions, Massey University, Private Bag 11 222, Palmerston North, New Zealand

2

School of Environment, University of Auckland, Private Bag 92 019, Auckland, New Zealand

3

Department of Geoscience, University of Los Andes, Cr 1 #18A-12, Bogotá, Colombia

Keywords Andesitic-basaltic volcanism . Eruptive dynamics . Plinian and sub-Plinian eruptions . Volcanic hazards

Andesitic stratovolcanoes often produce long, complex eruptive episodes, with several variations in style and magnitude (e.g., Cioni et al. 2003; Houghton et al. 2004; Carazzo et al. 2012). This unpredictability in eruption style could lead to larger societal repercussions. Plinian eruptions are rarely alike, even if climactic steady phases are of similar intensity and magnitude (e.g., Coltelli et al. 1998; Saucedo et al. 2010; Cronin et al. 2013; Eychenne et al. 2013; Avellan et al. 2014). For instance, rapid changes in eruption processes and style are commonly the most difficult to understand and anticipate at explosive mafic-intermediate composition stratovolcanoes (e.g., Bourdier et al. 1997; Coltelli et al. 1998;

76

Bull Volcanol (2017) 79:76

Page 2 of 27

Macias et al. 1997; Capra et al. 2006). Reconstructing the intricate dynamics of such explosive episodes and understanding the physical processes behind abrupt volcanic styletransitions are thus critical for defining the potential hazards of future large eruptions (e.g., Cioni et al. 2000; Platz et al. 2007; Arce et al. 2012; Shea et al. 2012). Plinian eruptions consist of highly energetic, sustained, steady to quasi-steady eruptive phases lasting hours (Wilson 1976; Cashman et al. 2000; Cioni et al. 2000, 2008). They involve high mass discharge rates (107–108 kg/s) of large volumes of pyroclasts (0.1–10 km3), leading to high eruption columns (20–35 km; Walker 1973; Cioni et al. 2000, 2008). These steady phases commonly alternate with phases of quasisteady to unsteady pulses producing pyroclastic density currents (PDCs) of different types (e.g., Shea et al. 2011, 2012; Carazzo et al., 2012; Cronin et al. 2013; Sulpizio et al. 2014). Sub-Plinian eruptions have a similar style but are of lower intensity (~ 106–107 kg/s), volume (0.01–0.1 km3), and column height (< 20 km; Cioni et al. 2000, 2003, 2008). All of these events cause damage of varying, although usually catastrophic, degrees (e.g., Bourdier et al. 1997; Macias et al. 1997; Andronico and Cioni 2002; Cioni et al. 2008; Cronin et al. 2013). During intermediate-silicic Plinian eruptions, the inherent high viscosity of magmas inhibits bubble nucleation and exacerbates pressure build-up before fragmentation (e.g., Klug and Cashman 1994, 1996; Gardner et al. 1996; Mader 1998; Cashman et al. 2000). In basaltic melts, high viscosities are attained by degassing-induced crystallization leading to second boiling, by magma-water interaction (cf. Alidibirov and Dingwell 1996; Mader 1998), and/or by sticking of material to the conduit walls (cf. Houghton et al. 2004). Mafic explosions can also be formed by decompression-induced microvesiculation of large volumes (> 1 km3) of rapidly ascending magma (Coltelli et al. 1998). Transitions between steady and unsteady explosive eruption phases also occur with shifts in magma decompression rates, fragmentation surface depth, magma composition, gas content and rheology, conduit instability, phreatomagmatism, and change in vent position (cf. Klug and Cashman 1994, 1996; Gardner et al. 1996; Mader 1998; Cashman et al. 2000; Shea et al. 2011, 2012; Bonadonna et al. 2016). These processes are reflected to a large extent in pyroclast properties, including vesicle textures of juveniles (cf. Klug et al. 2002; Houghton et al. 2004; Sable et al. 2006; Shea et al. 2011, 2012; Pardo et al. 2014). Understanding these factors controlling eruption dynamics is important in order to forecast more realistic eruption scenarios (e.g., Andronico and Cioni 2002; Cioni et al. 2008). Mt. Taranaki (Egmont, ~ 2500 m asl), New Zealand, is a stratovolcano that has produced over 200 explosive eruptions in the last ~ 30 ka, including at least 60 with sub-Plinian to Plinian characteristics (Alloway et al. 1995; Turner et al.

2008a, 2009, 2011a; Damaschke et al. 2017a, b). In this work, deposits produced by 5 of the 53 known late-Holocene (< 5 ka) eruptive episodes of Mt. Taranaki (Torres-Orozco et al. 2017) were used to explore the transitions in behavior and mechanisms of explosive eruptions in mafic-intermediate volcanic systems, to quantify the uppermost eruptive limits (in terms of intensity and magnitude) at this volcano, and to understand differences between hazards at separate vents on a single stratovolcano. The deposits of these five eruptive episodes were selected as they comprise the best preserved, and some of the thickest and coarsest, pyroclastic successions at Mt. Taranaki. They also represent contrasting sequences of eruption units (from fallout to PDC), suggesting diverse eruption styles (TorresOrozco et al. 2017), and have mafic-intermediate compositions: three erupted andesite magma and two erupted basaltic-basaltic andesite magma. Most of these five eruptions were produced from the summit vent (current summit 2518 m), and at least one (i.e., Manganui-D, Torres-Orozco et al. 2017) was generated from the satellite cone Fanthams Peak (current summit 1966 m, Fig. 1).

Geological setting Mt. Taranaki is the youngest and southernmost stratovolcano of a chain of the four < 1.75 Ma and NNW-SSE trending volcanoes that comprise the Taranaki Volcanic Lineament (TVL, Neall et al. 1986; Alloway et al. 1995; Zernack et al. 2011; Fig. 1). The TVL constitutes New Zealand’s westernmost center of volcanism and is associated with subduction of the southern Pacific plate beneath the Australian plate along the Hikurangi Trough (Henrys et al. 2003; Fig. 1). The present-day, ~ 12 km3 (Neall 2003), volume of Mt. Taranaki above ~ 1400 m is < 10 ka (Neall 1979; Turner et al. 2011a, b) and comprises a small fraction of the 150 km3 volcaniclastic debris that form the surrounding ring plain (Zernack et al. 2011). The youngest feature is the eastern half of the ~ AD 1800 Sisters Dome, which makes up the present summit (Platz et al. 2012). The western side of the summit collapsed some time during a series of eruptions over the last 800 years (Procter et al. 2010). The conical edifice symmetry is broken in the southeast by the 1966 m in height, Fanthams Peak basaltic satellite cone (Fig. 1). Only eruptions of < 3000 cal BP until ~ 1600 cal BP from this vent have been confirmed (Turner et al. 2008b; TorresOrozco et al. 2017). The last ~30 ka of explosive activity at Mt. Taranaki are recorded by tephras within soil profiles to the northeast and southeast of the volcano (Neall 1972; Whitehead 1976; Alloway et al. 1995), as well as within lake and peat sediments (Turner et al. 2008b, 2009, 2011a; Damaschke et al. 2017a). Recent proximal studies by Torres-Orozco et al. (2017) recognized at least 53 bed-sets which were interpreted as deposits

Page 3 of 27 76

Hik

SI Pacific Plate

60

N

39°16'S

Pouakai

Taranaki Fault

Midhirst

50

Fanthams Peak

D E

Taranaki

F

SH3

40

S

H43

c

C

B

Stratford

Natio nal Park

174°7'E

A

Fanthams 174°4'E

Kaponga

G

39°19'S

FF

Mt. Taranaki

40°S

ngi

ial

100 km

Ax

IGF Inglewood

Kaitake

Tro u

es ng Ra

TF

b

Oakura

R

ura

TVL

gh

Sh

CE FZ

Paritutu

70

ea

New Plymouth

Eg m on t

38°S

NI

Australian Plate

rB

Onaero

178°E

t

Waitara

174°E

el

5680 000m.S

Tasman Sea

a

TV Z

38°58'S

Bull Volcanol (2017) 79:76

Eltham Study sites

MF

30

10 km

Type sections This work and Torres-Orozco et al. (2017)

39°36'S 20

SH45

Lake and peat sediment records (Damaschke et al. 2017a, b) Other revised and/or re-excavated sections

Manaia Hawera

Fault line Roads

b

80 173°58'E

SH State Highways

Residential area

90

1700 000m.E

10

20

30

40

174°43'E 50

Fig. 1 (a) Tectonic setting of the North Island (NI) of New Zealand (modified from King and Thrasher 1996; Henrys et al. 2003; Sherburn and White 2006; Stagpoole and Nicol 2008). CEFZ Cape Egmont Fault Zone, R Mount Ruapehu, SI South Island, TVL Taranaki Volcanic Lineament (yellow line), TVZ Taupo Volcanic Zone. (b) Zoomed area of the Taranaki Peninsula and the TVL. The latter comprises four

< 1.75 Ma and NNW-SSE migrating andesitic volcanoes or their eroded volcanic edifice-remnants (Neall 1979): Paritutu, Kaitake, Pouakai and Mt. Taranaki (and the satellite cone of Fanthams Peak). IGF Inglewood Fault, MF Manaia Fault, NFF Norfolk Fault. (c) Zoomed area of the proximal eastern flanks of Mt. Taranaki, and type sections in this study. Coordinate system: NZGD 2000 New Zealand Transverse Mercator

from 53 explosive eruptive episodes generated from the summit of Mt. Taranaki and Fanthams Peak over the last 5000 years, and consisting of complex successions of intercalated fallout and PDC units. At least 16 of these bed-sets comprise widespread lapilli fall deposits that could represent

sub-Plinian to Plinian eruptions (Torres-Orozco et al. 2017). Here, we revisit the deposits of the five thickest and coarsest eruptive episodes of this pyroclastic succession so as to quantify parameters associated with, and decode the dynamics of the maximum likely events at this andesitic stratovolcano.

76

Page 4 of 27

Methods Bed-sets of the Kokowai, Upper Inglewood, Manganui-D, Kaupokonui, and Burrell eruptive episodes were studied at 112 exposures between 1 and 42 km from the summit crater on the eastern flanks of Mt. Taranaki (Fig. 1). Following Torres-Orozco et al. (2017), in proximal areas (1–3 km from the crater), each bed-set was integrated by intercalated pyroclastic or epi-volcaniclastic layers, had clear lower and upper boundaries characterized by sharp contacts with paleosols or weathered ash deposits, and was produced during a single eruptive episode (e.g., Turner et al. 2008b; Manville et al. 2009). Thereafter, the pyroclastic layers of a bed-set correspond to individual fall or PDC units (cf. Fisher and Schmincke 1984) that could have been deposited during phases of varying magnitude (e.g., waxing, climactic and waning phases) within a single eruptive episode. Most distal (> 10 km) exposures were also studied previously by Neall (1972), Whitehead (1976), Franks (1984), Alloway et al. (1995) and May (2003). These data were ingested and reinterpreted here (Fig. 1). Data from lake and peat sediments (Damaschke et al. 2017a) were also used to extend maps of distal distribution of individual fallouts. The criteria employed to correlate proximal and distal fall deposits produced from Mt. Taranaki can be found in Damaschke et al. (2017b). Maximum deposit thickness and the three axes of the largest five pumice and lithic clasts of the fall deposits were meas u r e d i n th i s w o r k a t a l l l o c a t i o n s l i s t ed ab o v e (Online Resource 1) so as to construct isopach and isopleth maps. Thickness measurements on the ring plain are likely minima, due to upper-deposit reworking. Fall deposits of the Manganui-D bed-set (layers MD1 to MD3) were identified in 107 locations; however, the individual fall layers could not be clearly distinguished at medial-distal locations, and isopachs were constructed for the total deposit thickness. Fall deposits of the Upper Inglewood bed-set (layer Uig7) were recognized in 92 sections. Fall deposits of the Kokowai bed-set (layers Kw4 and Kw7) were identified in only 20 sections; therefore, isopachs and isopleths of the Kokowai are least minimum dispersal models calculated by using the interpolation and extrapolation parameters described below. Isopachs of the Kaupokonui bed-set were modified from Whitehead (1976), and isopleths were drawn using data from 77 sections collected in this study. Fall deposit isopachs and isopleths of the Burrell bed-set (corresponding to 57 sections) were modified from Platz et al. (2007). In addition, maps of the distribution of block-and-ash flow and other PDC deposits from the Kokowai (layers Kw1 to Kw3, Kw6, and Kw8) and the Upper Inglewood (layers Uig1 to Uig6) were constructed using data collected at 14 and 37 sections, correspondingly (Online Resource 1). Isopach and isopleth contours were digitally drawn in ArcMap 10.1 using a natural neighbor interpolation for

Bull Volcanol (2017) 79:76

reference and tested against contours generated from a singlesegment regression combined with an ordinary kriging interpolation in a similar way to Yang and Bursik (2016). For consistency, distal isopachs of < 5 cm in thickness were partially, or fully, extrapolated using constant minimum isopach areas1/2 (A1/2) of d0.1 cm (of 60 to 80 km). These constants were applied over interpolated isopach A1/2 of d5 cm (of 20 to 40 km) (cf. Klawonn et al. 2014). The extrapolated distal isopachs were assumed to have elliptical and sub circular shape factors of less than one to one, aspect ratios of one to more than one, and eccentricities (0 < ellipse < ~ 1.05) equal to the mean eccentricity of the proximal and medial isopachs (e.g., Cioni et al. 2000; Sulpizio 2005). Isopleths of < 0.8-cm in diameter were partially extrapolated using constant minimum isopleth areas1/2 of d0.1 cm (of 15 to 20 km, or from 40 to 50 km). These constants were applied over interpolated isopleth A1/2 of d1.6 cm (of less than 10 km, or ranging from 15 to 25 km, correspondingly). The constant minimum A1/2 values employed were averaged from exponential regressions (cf. Pyle 1989, 1995; Klawonn et al. 2014) of all the isopachs or isopleths interpolated from the field data collected in this work. Samples of 0.5 to 5 kg from each layer of every bed-set, and in some cases from different stratigraphic levels (base, middle, and top) of a single layer, were collected in proximal areas. After cleaning in distilled water, samples were dry sieved at 0.5-ϕ intervals (ϕ) between – 5 and 4.5 ϕ. The granulometric median diameter (Mdϕ) and graphical standard deviation (σ1) of Inman (1952; Fig. 2) and the sorting categories of Folk and Ward (1957) were determined for each sample. Following the criteria of Avellan et al. (2014), componentry analyses were completed on 350 particles from the common coarse mode (− 1 to − 3.5 ϕ) and from the finer secondary mode (commonly 0.5 ϕ) using a binocular stereoscope. Lithology classes included vesicular juvenile clasts, dense juvenile clasts, accessory (altered and non-juvenile volcanic fragments) and accidental lithics (non-volcanic clasts), and free crystals. Three sets of 350 counts were made for each fraction, giving uncertainties of 0.6 to 0.2 vol%. The vesicular juvenile clasts of each studied bed-set were qualitatively classified following the scheme of Houghton and Wilson (1989) and Cas et al. (2008) into the following: (1) finely (i.e., ~ 0.5 to 3 mm vesicles) to coarsely (i.e., ~ 3 to 6 mm vesicles) vesicular yellow-white, pink, gray, or brown pumice; (2) dense, microvesicular (i.e., often deformed vesicles only visible under microscope) gray, dark-gray or brown pumice; (3) dense, finely vesicular, and microvesicular (i.e., firm, heavy clasts of thin to thick-walled vesicles), dark-gray and dark-brown juvenile clasts (scoriaceous) only present in Manganui-D deposits; and (4) crystalline (i.e., crystal rich, ~ 60 vol% crystals) yellow-white or gray-violet pumice. For density analyses, a total of 1542 cylinders (of 10– 20 mm-diameter) were drilled from vesicular juvenile lapilli clasts (− 4 to – 3 ϕ), and 32 cylinders were drilled from dense

Bull Volcanol (2017) 79:76

a

Page 5 of 27 76

b

5

σ1ϕ

3

PDCs 2

Kw4

Kaupokonui (Kp) Manganui-D (MD)

Kp

0.1

MD

Upper Inglewood, layer Uig7 Upper inglewood, layers Uig1-6

0.01

Bu

7

Fallouts

1

Burrell (Bu)

1

Kw

Log Thickness (m)

4

10

Ui

Kokowai, layer Kw7 Kokowai, layers Kw5-6 and Kw8

g7

Kokowai, layer Kw4 Kokowai, layers Kw1-3

0.001

0 -5

-4

-3

-2

-1

0

1

2

3

0

Mdϕ

10

20

30

40

50 1/2

Isopach Area

60

70

80

(km)

Fig. 2 a Diagram of median diameter (Mdϕ) vs. Inman sorting coefficient (σ1ϕ), modified from Walker (1971), and corresponding to grain-size of samples from different < 5000 to 300 cal BP fall (Fallouts)

and pyroclastic density current (PDCs) deposits of Mt. Taranaki. b Bulkdeposit isopach data plotted on square root (Area) vs. log (Thickness) of each fall deposit studied

andesitic juvenile clasts for reference. Bulk densities were obtained with a Micrometrics GeoPyc 1360 envelope density analyzer (± 1.1% reproducibility). Skeletal densities (accounting for gas-accessible porosity) and solid densities (excluding both connected and isolated porosities) were acquired with a Quantachrome Ultrapycnometer using pure nitrogen as the flowing gas (± 0.2% reproducibility). Five measurements per sample were performed and averaged. All cylinders were then crushed in a porcelain mortar (down to 2 to 3 ϕ) in order to perform analyses of solid density. Bulk porosities with uncertainties of 0.1 to 0.5% were calculated following Houghton and Wilson (1989). Porosities accessible (i.e., connected porosity) or inaccessible (i.e., isolated porosity) to the nitrogen gas were calculated following Klug and Cashman (1996) (Online Resource 1). Samples of vesicular, dense juvenile clasts and bulkdeposits were hand-picked from several horizons within every bed-set and ground to powder in a tungsten carbide ring mill. The powders were fused into disks following the Norrish fusion method (Harvey et al. 1973) with a dilution factor of 2:6 (ignited sample at 12:22 lithium tetraborate:lithium metaborate flux). The disks were analyzed in triplicate using the PANalytical Axios 1 kW wavelength dispersive X-ray fluorescence (XRF) spectrometer at the University of Auckland (New Zealand). For calibration, 28 international standards analyzed in triplicate were used. U.S. Geological Survey glass standard BCR-2G was analyzed to provide an independent assessment of accuracy and precision (Wilson 1997). The results for BCR-2G are consistent with the reference values (within < 1%) for elements considered in this study, and replicate analyses indicate an analytical precision (σ) of < 1% (mostly ~ 0.7%). Minimum fall deposit volumes were calculated using a range of methods that were applied to both interpolatedextrapolated isopachs and interpolated isopachs (Table 1).

Minimum PDC volumes were calculated by mapping a minimum deposition area and assuming an average deposit thickness (cf. Arce et al. 2005; Platz et al. 2007). Dense-rockequivalent (DRE) volumes were calculated using the weighted average bulk density and the weighted average solid density estimates for each fall or PDC unit based on the volume percentage of pumice and lithics. The total DRE volume of an eruptive episode was then calculated by addition of the individual fall and PDC DRE volumes. Minimum, maximum, and total average column heights were calculated following the methods of Carey and Sparks (1986), Sparks (1986), Pyle (1989), Sulpizio (2005), and Bonadonna and Costa (2013). The results were used to estimate Bbc^ (i.e., the half distance of the maximum clasts extrapolated from the isopleth data; Pyle 1989) and neutral buoyancy column heights (HB), by substituting the total average column heights (HT) in equations –bc = 0.41HB/ (HB 1/2 − 7.3) and HT = HB/0.7, postulated by Sparks (1986) and Pyle (1989). Mass-and-volume eruption rates (MER and Q, respectively) were calculated by substituting the total average column height of each fallout in equations HT = 1.67(Q0.259) of Sparks (1986) and Carey and Bursik (2000) and HT = 0.236 × MER1/4 of Wilson and Walker (1987). These were also determined by interpolating the total average column height vs. the volumetric and mass eruption rates in the diagram of Carey and Bursik (2000). Following Wilson (1976), the total erupted mass (mT) of each fall or PDC unit was calculated as a product of the corresponding solid density and DRE volume. Thereafter, the eruption magnitude (M) was estimated by applying equation M = Log10(mT) − 7 of Pyle (2000). The total magnitude of an eruptive episode was estimated using the total weighted average solid density of a bed-set, obtained from the individual solid densities of both fall and PDC deposits and the total DRE volume.

76

Bull Volcanol (2017) 79:76

Page 6 of 27

Table 1 Eruptive volumes obtained by different methods using bulk-isopach data of fall deposits, and area and bulk-average thickness of pyroclastic density current deposits Method

Bu int

Kp int

MD1–3 int-ext

Fall deposits Cole and Stephenson (1972), Pyle (1989) 0.07 0.07 0.30 1 segment km3 Legros (2000); selected isopach (d..) in cm (d10) 0.06 (d5) 0.08 (d20) 0.46 km3 Fierstein and Nathenson (1992), Bonadonna and Houghton (2005) 3 0.07 0.09 0.39 2 segment km Bonadonna and Houghton (2005) 0.09a 0.47 3 segment km3 Sulpizio (2005); method Aip1/2 (kt or k0) plus ext distal data 26.86 31.06 35.60 Aip1/2(kt) km 17.05 22.05 15.36 Aip1/2(k0) km Aip1/2(kt) km3 0.05 0.05 0.21 0.03 0.03 0.08 Aip1/2(k0) km3 Sulpizio (2005); method k1 vs. Aip1/2 Vp/Vt < 0.3 km3 0.09 0.09 0.39 0.3 < Vp/Vt < 0.7 km3 0.18 0.17 0.71 3 0.53 0.50 2.11 Vp/Vt > 0.7 km Bonadonna and Costa (2012, 2013) 0.07 0.13 0.24 km3 Average total volume 0.08 0.10 0.35 km3 σ1 0.04 0.04 0.09 Pyroclastic density current deposits E.g., Arce et al. (2005), Platz et al. (2007) Uig4–Uig6 Uig3 Uig2 Minimum thickness (m) 0.08 0.04 0.02 Maximum thickness (m) 0.53 0.38 0.46 Average thickness (m) 0.31 0.21 0.24 7.2 2.9 5.8 Minimum area (107 m2) 22.0 6.0 14.0 Volume (106 m3) Volume km3 0.0221 0.0060 0.0140

MD1–3 int

Uig7 int-ext

Uig7 int

Kw7 int

Kw4 int-ext

Kw4 int

0.34

0.16

0.23

0.06

0.29

0.37

(d20) 0.46

(d20) 0.25

(d20) 0.25

(d20) 0.07

(d30) 0.36

(d30) 0.36

0.06

0.31

0.47a

a

0.47

0.23

0.27

0.47a

0.25

0.28a

39.64 15.36 0.25 0.07

27.80 10.44 0.11 0.04

38.20 10.44 0.16 0.04

17.33 29.34 0.04 0.07

25.37 10.89 0.20 0.08

30.46 10.89 0.26 0.08

0.45 0.81 2.41

0.20 0.38 1.13

0.30 0.54 1.61

0.07 0.15 0.44

0.37 0.73 2.12

0.48 0.91 2.67

0.44

0.18

0.21

0.06

0.30

1.10b

0.41 0.08

0.22 0.10

0.24 0.04

0.06 0.01

0.31 0.05

0.51 0.28

Uig1 0.16 1.46 0.81 0.8 6.8 0.0068

Kw8 0.08 0.23 0.15 1.2 1.8 0.0018

Kw6 0.04 0.15 0.10 1.2 1.2 0.0012

Kw3 0.06 0.08 0.07 0.5 0.36 0.0004

Kw2 0.07 0.28 0.17 2.7 4.6 0.0046

Kw1 0.14 0.78 0.46 1.0 4.5 0.0045

0.36

Bed-sets: Burrell (Bu), Kaupokonui (Kp), Manganui-D (MD), Upper Inglewood (Uig) and Kokowai (Kw). Fall deposit volumes were calculated from solely field-interpolated (int) isopachs, or combined with extrapolated (ext) distal isopachs Italic numbers correspond to volumes estimated by using the method of Sulpizio (2005) and preferred to calculate the average total volume Aip1/2 distance from vent expressed as the square root of the isopach area, calculated with either kt or k0, kt thinning rate when the total data fit a single segment, k0 thinning rate of the proximal segment when data are divided in multiple segments, k1 thinning rate calculated from Aip1/2 (kt), Vt total volume, Vp proximal volume a

Volumes calculated by using AshCalc 1.1 of Daggitt et al. (2014)

b

The Kw4 int. volume calculated by following the method of Bonadonna and Costa (2012, 2013) was preferred for representing the best fit of this work’s data to the Weibull distribution

Late-Holocene large explosive eruptions at Mt. Taranaki The 4700–4600 cal BP Kokowai eruptive episode The Kokowai eruptive episode (Torres-Orozco et al. 2017) produced the thickest and coarsest fall deposits of the lateHolocene history of Mt. Taranaki, along with a series of PDC deposits (Fig. 3 and Online Resource 1). The most complete exposures of the Kokowai bed-set include eight layers (Kw1 to Kw8) at proximal locations (e.g., section A, Fig. 1), which are poorly preserved in the ring plain. However, such layers are distinctive in lake sediments recovered 24–25 km

east of Mt. Taranaki (Damaschke et al. 2017b) and within sands forming coastal cliff exposures 42 km northeast of the crater, as deposited below the ~ 4–3.9 ka BStent^ tephra marker of Alloway et al. (1995). Layers Kw4 and Kw7 are clast-supported, coarse angular lapilli fall deposits, and they are the most distinctive layers of the Kokowai bed-set (Fig. 3). They are preceded by and separated by PDC deposits (i.e., Kw1 to Kw3, and Kw5 to Kw6; Torres-Orozco et al. 2017). In proximal sections, while Kw4 is up to 180 cm thick and mostly massive, Kw7 is 60–65 cm thick and stratified. The lowermost 10 cm of Kw4 is reversely graded, but the uppermost 30 cm is either pumice-rich and normally graded or lithic-rich and reversely graded. The

Bull Volcanol (2017) 79:76

Kw7 has a 20-cm-thick, normal to reversely graded base, capped by ~ 40-cm-thick, massive or normally graded lapilli (Fig. 3). In medial sites, Kw4 is a 30–40-cm-thick, massive to weakly bedded medium-grained lapilli, and layer Kw7 is a 20cm-thick, weakly to normally graded deposit of fine lapilli and coarse ash. In distal locations, Kw4 is a 4–5-cm-thick layer of medium-sized lapilli, and Kw7 is a 1–3-cm-thick layer of coarse ash (Damaschke et al. 2017b). Only ~ 2 cm of the Kw4 lapilli is preserved at the most distal site studied, which is in the Onaero coastal cliffs (Fig. 1). At proximal sites, Kw4 and Kw7 are poorly sorted and have overall unimodal grain-size distributions (Fig. 3). Kw4 has a mode of – 3 ϕ at the base, − 3.5 ϕ across middle levels, and − 2.5 ϕ in upper levels. From its base to its top, Kw7 has main modes of − 2.5, − 3.5, and – 3 ϕ and a secondary mode at − 4.5 ϕ at its base. This reflects the presence of ballistic clasts at this level. Most layers representing PDC deposits (Fig. 2) are very poorly sorted and have bimodal distributions (Fig. 3). Lithic-rich layers Kw1, Kw2, and Kw5 were associated with block-and-ash flows (Torres-Orozco et al. 2017) and have the most extended grain-size distributions of all deposits sampled, with modes ranging from − 5 to 4ϕ. In contrast, pumice-rich layers Kw3, Kw6, and Kw8 were associated with column-collapse PDCs (Torres-Orozco et al. 2017) and have uniform distributions with modes at − 3.5 and 3 ϕ (Fig. 3). The Kokowai succession comprises yellow, gray, and minor pink pumice, as well as gray and dark-gray dense andesite juvenile clasts, free crystal fragments (plagioclase, hornblende, pyroxene and Fe-Ti oxides), and other accessory and accidental lithics (Fig. 3). Pumice contents of layers Kw4 and Kw7 are up to 70–85 vol% in their middle stratigraphic levels and in both cases decrease towards the upper levels to 20 and 62 vol%, respectively. The top part of Kw4 is rich in both dense juvenile clasts (56 vol%) and accidental lithics (24 vol%; Fig. 3). While layers Kw3, Kw6, and Kw8 contain 70–80 vol% pumice, layers Kw1, Kw2, and Kw5 are composed of up to 70–80 vol% dense andesite juvenile clasts. Pumice lapilli from layers Kw4 and Kw7 is predominantly finely vesicular to microvesicular. Yellow and gray pumice clasts of layer Kw4 become denser upwards, increasing from a bulk density of 0.8–0.9 g/cm3 at the base (70–75 vol% bulk porosity), through 1.0 g/cm3 in the middle (70 vol% bulk porosity), to bimodal 0.9 and 1.2 g/cm3 at the top (75 and 60 vol.% bulk porosities; Fig. 3). There is a strong contrast between yellow and pink pumice (0.7–0.8 g/cm3, 80 vol% bulk porosity) vs. dense gray pumice (1.2 g/cm3, 60 vol% bulk porosity) across the basal level of layer Kw7. In contrast, the middle part of Kw7 contains only yellow pumice (0.7 g/cm3, 80 vol% bulk porosity; Fig. 3). The connected and bulk porosities of both Kw4 and Kw7 layers (Table 2) correlate graphically (Fig. 3). In addition, both layers have also low isolated porosities of 0.1 and 0.6 vol%, respectively (Online Resource 1), indicating high vesicle interconnectivity.

Page 7 of 27 76

The 3300 cal BP Upper Inglewood eruptive episode The Upper Inglewood eruptive episode (Torres-Orozco et al. 2017) produced the thickest PDC succession of the last 5 ka at Mt. Taranaki. Most of the seven layers that comprise the Upper Inglewood succession (Uig1–Uig7; Fig. 4 and Online Resource 1) are exposed across the eastern flanks. PDC deposits can be traced for up to 15 km east of the summit crater, while fallouts are widely preserved on the medial ring plain, as well as within lake sediments 24–34 km southeast, northeast, and north of the summit crater (Damaschke et al. 2017b). Lithic-rich layers Uig1 to Uig3 are restricted to few locations. They comprise block-and-ash, or lapilli-and-ash, massive deposits, which correlate distally with strongly stratified ash deposits (Online Resource 1). In contrast, the most distinctive proximal layers of the Upper Inglewood succession are the Uig4 to Uig6 pumice-rich PDC deposits, and the Uig7 fallout (Torres-Orozco et al. 2017; Fig. 4). Layers Uig4 to Uig6 contain pinch-and-swell structures in matrix-supported ash-and-lapilli beds that are up to 45 cm in thickness at proximal sites. These thin into 5–10-cm-thick massive beds of coarse ash with rare fine lapilli at medial locations. Layer Uig7 comprises a 70-cm-thick, mostly massive deposit of clast-supported coarse lapilli. The basal 10 cm of Uig7 are reversely graded and the uppermost 10 cm are normally graded (Fig. 4). At medial locations, layer Uig7 thins into a 10–20cm-thick, massive deposit of fine to rare medium lapilli. At distal sites, only a 1–4-cm-thick deposit of coarse ash and very fine lapilli is preserved (Damaschke et al. 2017b). Whereas the deposits of layers Uig1 to Uig3 are very poorly to extremely poorly sorted, and often bimodal (− 5 to 3.5 ϕ; Fig. 4), the deposits of layers Uig4 to Uig6 are very poorly sorted with uniform distributions between − 3 and 3 ϕ (Fig. 4). The base of layer Uig7 has very poor sorting with a main mode at – 4 ϕ, but the middle and top levels comprise moderately well-sorted lapilli with unimodal distributions and main modes at – 5 ϕ (Fig. 4). Layers Uig1 and Uig2, which have been associated with block-and-ash flows (Torres-Orozco et al. 2017), are rich in dense andesitic clasts (66–72 vol%) and free crystals (18– 21 vol%), with rare pumice (2–5 vol%; Fig. 4). Deposits of layer Uig3 contain gray-violet pumice (20 vol%) and fresh andesitic clasts (61 vol%), and deposits of layers Uig4 to Uig6, produced by PDCs (Torres-Orozco et al. 2017; Fig. 2), have 52–57 vol% yellow and gray pumice, with a dense andesitic clast content that progressively decreases upwards from 35 vol% at the base of the deposit to 19 vol% (Fig. 4). The free crystal (plagioclase, hornblende, Fe-Ti oxides, pyroxene, and rare biotite) content of layers Uig4 to Uig6 increases from 7 to 17 vol%. Finally, the Uig7 layer is vertically homogeneous (Fig. 4) and contains pumice (67 vol%), dense andesitic clasts (13–15 vol%), accessory lithics (8 vol%), and free crystals (12 vol%).

50 cm

40

Kw7

Kw6

b-t

b-m t

FA FL B CA CL

60

Kw7

m

r

S

Kw4

40

m

R

Bulk-p (vol.%)

80

t

#

>

r

#

m

m

#

> >

#

m

=

m n-r

n-r

t

b-t

60

b-m

80

Deposit level t: top m: middle b: base

Kw1

Kw2

Kw3

Kw4 b

Kw4 m

Kw4 t

Kw5

Kw6

Kw7 b

Kw7 m

Kw7 t

r

m

100 Kw8

=

>

50

vol.% m

m r,p # fx

#

#

> >

=

=

>

0

Grain-size FA: fine ash CA: coarse ash FL: fine lapilli CL: coarse lapilli B: blocks/bombs

Vesicular juveniles Dense andesitic clasts

-5

Kw5

Kw6

0

Kw1

Kw2

Kw3

Kw4 base

Kw4 middle

Kw4 top

Kw7 base

Kw7 middle

Kw7 top

Kw8

σ1= 3.9

σ1= 3.6

σ1= 3.1

σ1= 1.3

σ1= 1.6

σ1= 1.4

σ1= 3.2

σ1= 3.2

σ1= 1.6

σ1= 1.8

σ1= 1.9

σ1= 3.1

30

30

30

30

30

30

1 1.1 1.2 1.3

Kw4 base

Kw4 middle

1 1.1 1.2 1.3

Bulk-d (g/cm3)

0.6 0.7 0.8 0.9

Kw4 top

0.6 0.7 0.8 0.9

Kw7 base

Yellow pumice Dense grey pumice

Pink pumice

Kw7 middle

Kw7 top

60

60

70

70

80

80

Bulk-p (vol.%)

50

50

0.5

0.9

0.5 1.3

0.9

0.5 1.3

0.9

1.3

0.5

0.9

0.5 1.3

0.9

0.5 1.3

0.9

1.3

50

50

Kw4 top

80

70

80

90

90

Kw4 base

Kw4 middle

70

Kw7 base

Kw7 middle

Kw7 top

Bulk-p (vol.%)

60

60

S: Sorting (left to right bars): extremely-poor, very-poor, poor, moderate, moderately-good (Folk and Ward 1957) R: Particle rounding (left to right bars): angular, subangular, subrounded, rounded, well rounded

Pink pumice Yellow pumice Dense grey pumice

20

20

20

20

20

20

20

20

20

20

20

20

wt.%

Dense lithics Free crystals

Fig. 3 Lithostratigraphic characteristics, modified from Torres-Orozco et al. (2017), componentry, granulometry, and pumice density and porosity data of the 4700–4600 cal BP Kokowai bed-set (layers Kw1 to Kw8). Grain-size modified from White and Houghton (2006). Notice the slight bimodality in some fall deposit histograms (e.g., basal and middle deposit levels of layer Kw7) produced at 4 ϕ mode by ash deposits from associated pyroclastic density currents. Connected and bulk porosities (p) and bulk densities (d) were calculated from texturally different pumice clasts, hand-picked from different vertical fall deposit levels (30 clasts per texture per level)

40

50

60

70

80

90

Kw1

Kw2

Kw3

Kw4

Kw5

Kw8

Kokowai

n: normal grading r: reverse grading fn/fr: faint normal/reverse grading m: massive p: parallel-stratified x: cross-stratified fp/fx: faint parallel-/cross-stratified

Topography mantling Pinching Pinch-and-swell Clast-support # Matrix-support # Clast-rich matrix-(~10%) support

# clasts

= >

Page 8 of 27

Connected-p (vol.%)

Bulk-d (g/cm3) Bulk-d (g/cm3)

76 Bull Volcanol (2017) 79:76

Bull Volcanol (2017) 79:76

Page 9 of 27 76

Table 2 Bulk and solid densities, bulk porosities, dense rock equivalent volumes, and eruptive parameters calculated from fall and pyroclastic density current deposits produced by different eruptions at Mt. Taranaki

Fall deposits Burrell avg. pumice avg. lithic wt. avg. Kaupokonui avg. pumice avg. lithic wt. avg.

Content vol.%−1

bulk-d g/cm3

so-d g/cm3

bulk-p vol.%

avg.vol km3

DREa km3

mTb 1011 kg

Qc 104 m3/s

MERd 108 kg/s

Te h

avg.HT km

Mf

0.76 0.24

0.90 2.80 1.36

2.42 2.90 2.54

62.8 2.6 48.3

0.08 0.02 0.10

0.03 0.02 0.05

0.72 0.56 1.4

0.39

0.13

2.87

14.2

4.1

1.20 2.50 1.80

2.56 2.60 2.58

53.0 3.0 30.0

0.05 0.05 0.10

0.02 0.05 0.07

0.60 1.3 1.8

0.37

0.12

4.04

14

4.3

3.01 2.90 2.98

37.4 5.0 27.7

0.30 0.10 0.40

0.19 0.1 0.29

5.7 2.8 8.6

3.0

1.10

2.24

24

4.9

2.70 2.80 2.73

65.0 8.0 46.2

0.20 0.10 0.30

0.07 0.09 0.16

1.9 2.6 4.5

2.1

0.76

1.66

22

4.7

2.91 2.90 2.90

70.3 4.5 48.6

0.07 0.03 0.10

0.02 0.03 0.05

0.60 0.84 1.5

1.6

0.58

0.72

20.6

4.2

2.86 2.90 2.87

67.6 7.0 56.7

0.33 0.07 0.40

0.11 0.07 0.17

3.0 1.9 5.0

5.0

1.80

0.75

27.5

4.7

2.86

67.6

0.90

0.29

8.3

2.90 2.87

7.0 56.7

0.20 1.10

0.19 0.48

5.4 14.0

5.0

1.80

2.06

27.5

5.1

3.27 2.93 3.11

0.012 0.010 0.022

0.004 0.008 0.012

0.13 0.25 0.39

–

–

–

–

3.6

2.84 2.78 2.79

0.001 0.005 0.006

0.001 0.004 0.005

0.03 0.11 0.15

–

–

–

–

3.2

2.74 2.73 2.73

0.001 0.020 0.021

0.003 0.017 0.020

0.08 0.46 0.55

–

–

–

–

3.7

2.88 2.90

0.002 0.001

0.001 0.001

0.03 0.03

2.89

0.003

0.002

0.06

–

–

–

–

2.7

0.54 0.46

Manganui-D, layers MD1–MD3 avg. vsc juv 0.70 1.89 avg. lithic 0.30 2.80 wt. avg. 2.16 Upper Inglewood, layer Uig7 avg. pumice 0.67 0.96 avg. lithic 0.33 2.60 wt. avg. 1.50 Kokowai, layer Kw7 avg. pumice 0.67 0.85 avg. lithic 0.33 2.80 wt. avg. 1.50 Kokowai, layer Kw4 avg. pumice 0.82 0.92 avg. lithic 0.18 2.70 wt. avg. 1.24 g Kokowai, layer Kw4 avg. pumice

0.82

0.92

avg. lithic 0.18 2.70 wt. avg. 1.24 Pyroclastic density current deposits Upper Inglewood, layers Uig4–Uig6 avg. pumice 0.54 1.13 avg. lithic 0.46 2.47 wt. avg. 1.74 Upper Inglewood, layer Uig3 avg. pumice 0.20 2.15 avg. lithic 0.80 2.49 wt. avg. 2.42 Upper Inglewood, layers Uig1 and Uig2 avg. pumice 0.04 1.96 avg. lithic 0.96 2.26 wt. avg. 2.25 Kokowai, layers Kw3, Kw6, and Kw8 avg. pumice 0.75 0.89 avg. lithic 0.25 2.75 wt. avg. 1.39 Kokowai, layers Kw1 and Kw2

76

Bull Volcanol (2017) 79:76

Page 10 of 27

Table 2 (continued) Content vol.%−1

bulk-d g/cm3

avg. pumice 0.01 0.89 avg. lithic 0.99 2.75 wt. avg. 2.86 Whole bed-set Upper Inglewood, layers Uig1–Uig7 wt. avg. 1.98 Kokowai, layers Kw1–Kw8 wt. avg. a 1.75 wt. avg. b

1.75

avg.vol km3

DREa km3

mTb 1011 kg

Qc 104 m3/s

MERd 108 kg/s

Te h

avg.HT km

Mf

2.88 2.90 2.90

0.001 0.009 0.010

0.001 0.009 0.010

0.03 0.26 0.30

–

–

–

–

3.4

2.84

0.35

0.24

6.9

2.1

0.76

2.55

22

4.8

2.89

0.51

0.31

8.9

5.0

1.80

1.34

27.5

5.0

2.89

1.21

0.73

21.0

5.0

1.80

3.19

27.5

5.3

so-d g/cm3

bulk-p vol.%

Content: average vol.%−1 of pumice or vesicular juveniles (vsc juv) and dense components (lithic) per layer(s). The prefix Bavg.^ stands for average and Bwt. avg.^ for weighted average. Notice that specific pumice and lithic wt. avg. of bulk-d and so-d, i.e. ((vol.%−1 ) (bulk-d or so-d)), are not included bulk-d bulk density, so-d solid density, bulk-p bulk porosity, avg.vol average volume, DRE dense-rock-equivalent, mT total mass, Q volume eruption rate, MER mass eruption rate, T duration of the eruption in hours (h), avg.HT average column height, M magnitude a

DRE = ((wt. avg. bulk-d)(avg.vol))/wt. avg. so-d

b

Wilson (1976): mT = so-d(DRE)

c

Sparks (1986) and Carey and Bursik (2000): HT = 1.67((Q)(0.259))

d

Estimated from Q and Wilson and Walker (1987): HT = 0.236(MER1/4 )

e

Wilson (1976): T = mT/MER

f

Pyle (2000): M = Log10(mT)-7

g

Average volume calculated with the method of Bonadonna and Costa (2012, 2013)

Typically, the total average bulk density of the pumice clasts contained in the Upper Inglewood succession decreases upwards (Fig. 4). The crystalline, dense, and violet Uig1 to Uig3 pumice clasts have unimodal bulk density distributions of 2.0 to 2.3 g/cm3 (25–30 vol% bulk porosities). Uig4 pumice clasts are predominantly yellow-white (0.9 g/cm3, 80 vol% bulk porosity) and gray (1.3 g/cm3, 55 vol% bulk porosity). Uig5 pumice clasts are mostly dense gray (1.6 g/cm3, 45 vol% bulk porosity), yellow/ white (0.8 g/cm3, 75–80 vol% bulk porosity), and gray (1.2 g/cm3, 60 vol% bulk porosity) at the base and across middle levels. However, the bulk density at the top level of Uig5 increases because of the abundance of crystalline violet pumice of 2.0 g/cm3 (40 vol% bulk porosity). Uig6 pumice clasts are mainly yellow/white and gray. Uig7 pumice clasts are mostly yellow/white (0.7–0.9 g/cm3, 80 vol% bulk porosity) and gray (1.1 g/cm3, 65 vol% bulk porosity), alongside dense gray (1.4 g/cm3, 50 vol% bulk porosity) and rare violet (2.1 g/cm3, 25–30 vol% bulk porosity; Fig. 4). Overall, the total connected porosities and bulk porosities of the Upper Inglewood succession correlate graphically and the isolated porosities are generally low (Fig. 4), suggesting a high vesicle interconnectivity. Both, the total average bulk density (1.0 g/cm3) and the total average solid density (2.7 g/cm3) of pumice from layer Uig7 are equivalent to a total average bulk porosity of 65 vol% (Table 2). In addition, the

total average solid density of the Uig7 pumice is consistent with the solid density of 2.8 g/cm3 of the accompanying dense andesitic clasts (Table 2). The ~ 2600 cal BP Manganui-D eruptive episode The Manganui-D eruptive episode (Torres-Orozco et al. 2017) generated the thickest and most widespread fallout of the basaltic Manganui Formation. Torres-Orozco et al. (2017) defined the Manganui Formation as comprising seven members (Manganui A–G), all of which were erupted from Fanthams Peak. Three layers of the Manganui-D succession (MD1 to MD3; Fig. 5 and Online Resource 1) are very well preserved at eastern proximal sites. In southern locations (e.g., section G, Fig. 1), lava flows from Fanthams Peak have eroded into the uppermost Manganui-D layer (MD3). At most ring plain locations (Fig. 1), partially reworked Manganui-D fallout comprises the top part of the exposed profiles. Distally, ManganuiD deposits were identified by Damaschke et al. (2017b) in lake and peat sediments recovered at 24–34 km southeast, northeast, and north of the summit crater. In proximal sections, layer MD1, which is 40 cm in thickness, and layer MD2, which is 80 cm in thickness, are friable, reverse to massive and normally graded, clast-supported deposits of fine-medium lapilli and coarse ash (Fig. 5). Layer MD3 consists of a 50-cm-thick set of four reversely graded, clast-supported sub-layers. In medial locations, MD1 and

Uig2

Uig3

Uig1

Uig4

Uig5

Uig6

Uig7

20

40

60

80

#

#

>

>

>

r

m

n

p m

#

>

R

40

60

Bulk-p (vol.%)

20

S

80

Uig7 Uig6 Uig5 top Uig5 base Uig4 Uig3 Uig1-2

Uig1

Uig2

Uig3

m n

Uig4

Uig5

Uig6

Uig7 b

Uig7 t

100

# p,x

Upper Inglewood

FA FL B CA CL

50

# m # p

# p,x

fr

m

# fp/x

>

=

fn

0

vol.%

20

20

20

20

20

20

20

20

20

60

-5

wt.%

Uig5 top

Uig6 base

0

Uig1

Uig2

Uig3

Uig4

Uig5 base

Uig6 top

Uig7 base

Uig7 top

σ1= 3.5

σ1= 2.1

σ1= 4.1

σ1= 2

σ1= 2.5

σ1= 2.3

σ1= 2.2

σ1= 1.5

σ1= 2.5

σ1= 0.6

30

30

30

50

30

30

0.6

0.8

1

1.4

1.6

1.8

Uig1-2

Uig4

Bulk-d (g/cm3)

1.2

Grey pumice Dense grey pumice Yellow-white pumice Crystalline Y-W pumice Crystalline violet pumice

Uig3

Uig5 top

Uig7

2

2.2

Uig5 base

Uig6

30

30

30

30

30

30

30

Uig4

Uig1-2

Uig3

Bulk-p (vol.%)

20 30 40 50 60 70 80

Uig5 base

Uig5 top

Uig6

Uig7

Bulk-d (g/cm3)

# clasts

0.5

1.5

0.5 2.5

1.5

0.5 2.5

1.5

0.5 2.5

1.5

0.5 2.5

1.5

0.5 2.5

1.5

0.5 2.5

1.5

2.5

20

60

80

Uig1-2

Uig3

Uig4

Uig5 top

Bulk-p (vol.%)

40

Uig5 base

Uig6

Uig7

Fig. 4 Lithostratigraphic characteristics, modified from Torres-Orozco et al. (2017), componentry, granulometry, and pumice density and porosity data of the 3300 cal BP Upper Inglewood bed-set (layers Uig1 to Uig7). Notice the slight bimodality in some fall deposit histograms (e.g., basal level of layer Uig7) produced at ~ 3 ϕ mode by ash deposits from associated pyroclastic density currents. Connected and bulk porosities (p) and bulk densities (d) were calculated from texturally different pumice clasts, hand-picked from each pyroclastic layer (30 clasts per texture per layer). See Fig. 3 for complete symbology and definitions

50 cm

Upper Inglewood

Connected-p (vol.%)

30

Bull Volcanol (2017) 79:76 Page 11 of 27 76

MD1

MD1

MD2

MD3

20

30

40

50

20

#

#

#

m m r

r-n

r-n n m r

m r r r-n r-n

30

R

b

b-m

b-m

40

m-t

t

S

Bulk-p (vol.%)

t

t

Manganui-D

FA FL B CA CL

= > >

=

=

>

0

50

L-DF MD1

MD1

MD2

MD3

U-DF

100

Deposit level t: top m: middle b: base

MD3 MD2 MD1

50

vol.%

-5

20

20

20

20

20

20

30

20

20

20

20

30

30

20

wt.%

0

L-DF

MD1 base

MD1 middle

MD1 top

MD2 base

MD2 middle1

MD2 middle2

MD2 top

MD3 base

MD3 (p1)

MD3 (p2)

MD3 (p3)

MD3 (p4)

U-DF

σ1= 2.9

σ1= 1.5

σ1= 1.2

σ1= 1.6

σ1= 1.3

σ1= 1.1

σ1= 1

σ1= 1

σ1= 1.2

σ1= 1.3

σ1= 1.3

σ1= 1

σ1= 1

σ1= 3.9

30

30

30

1.5

1.5

1.5

1.7

1.7

1.9

2.3

2.1

2.3

MD 2 top base-middle

(g/cm3)

2.1

2.1

2.3

Bulk-d (g/cm3)

1.9

2.5

2.5

2.5

MD 1 middle-top base

Bulk-d (g/cm3)

1.9

Bulk-d

1.7

MD 3 top middle base

10

MD 1

10

MD 2

10

MD 3

30

40

50

30

40

50

30

40

50

Bulk-p (vol.%)

20

Bulk-p (vol.%)

20

Bulk-p (vol.%)

20

60

60

60

1

2

3

1

2

3

1

2

3

20

20

20

40

40

40

Bulk-p (vol.%)

30

middle-top base

MD 1

Bulk-p (vol.%)

30

top base-middle

MD 2

Bulk-p (vol.%)

30

top middle base

MD 3

50

50

50

Page 12 of 27

Fig. 5 Lithostratigraphic characteristics, modified from Torres-Orozco et al. (2017), componentry, granulometry, and density and porosity data of vesicular juvenile clasts of deposits of the 2600 cal BP Manganui-D bed-set (layers MD1 to MD3). Deposits of lower and upper debris flows (L-DF and U-DF) are also represented. Connected and bulk porosities (p) and bulk densities (d) were calculated from similar vesicular juvenile clasts, hand-picked from different vertical levels within each pyroclastic layer (30 clasts per level per layer). See Fig. 3 for complete symbology and definitions

50 cm

Manganui-D

Connected-p (vol.%)

# clasts # clasts # clasts

Bulk-d (g/cm3) Bulk-d (g/cm3) Bulk-d (g/cm3)

76 Bull Volcanol (2017) 79:76

50 cm

50 cm

r

m

n

40

50

60

70

80

Kp

40

b

m

m

60

t

Grey pumice Brown pumice

Burrell

FA FL B CA CL

t

=

=

b

FA FL B CA CL

Kaupokonui

Bu

Burrell

t

m

80

30

40

50

60

S

S

0

50

vol.%

50

vol.%

30

top base

50

Kaupokonui

R

R

Bulk-p (vol.%)

r

r

m

n

0

70

Kp t Kp b

100

Bu b

20

20

20

30

-5

-5

Deposit level t: top m: middle b: base

Bu m

Bu t

100

30

wt.%

0

>4

Kp top

Kp base

0

Bu base

Bu middle

Bu top

σ1= 2.3

σ1= 1.5

Md = -3.1 σ1= 1.7

Md = -4.5 σ1= 0.7

Md = -4.1 σ1= 0.9

20

20

20

1

Bulk-d

(g/cm3)

1.1 1.2 1.3 1.4 1.5 1.6

Kp top base

1 1.1 1.2 1.3

Bu base

Bulk-d (g/cm3)

0.6 0.7 0.8 0.9

Grey pumice Dense brown or dark grey pumice

Bu middle

30

60

70

80

50

60

70

Bulk-p (vol.%)

40

Bulk-p (vol.%)

50

Bulk-d (g/cm3) 0.5

0.9

0.5 1.3

0.9

0.5 1.3

0.9

1.3

0.9

1.3

1.7

40

60

70

30

50

60

80

Bulk-p (vol.%)

40

Bulk-p (vol.%)

50

Bu base

Bu middle

Bu top

70

Fig. 6 Lithostratigraphic characteristics, modified from Torres-Orozco et al. (2017), componentry, granulometry, and pumice density and porosity data of the 1200 cal BP Kaupokonui bed-set (Kp), and the AD 1655 Burrell fall deposit (Bu). Connected and bulk porosities (p) and bulk densities (d) were calculated from texturally different pumice clasts, hand-picked from different vertical fall deposit levels (30 clasts per texture per level). See Fig. 3 for complete symbology and definitions

Connected-p (vol.%)

# clasts # clasts

Bu top

Bulk-d (g/cm3)

20

Bull Volcanol (2017) 79:76 Page 13 of 27 76

76

Page 14 of 27

MD2 merge into a 14–17-cm-thick, massive, although faintly stratified deposit of coarse ash. Instead, MD3 becomes a 1- to 6-cm-thick, massive deposit of coarse ash and fine lapilli. Beyond 20 km from the crater, the three layers merge into a single 8–12-cm-thick and massive deposit. At distal locations, only a 0.2–8-cm-thick deposit of coarse ash and extremely fine lapilli is preserved (Damaschke et al. 2017b). Layers MD1 to MD3 are poorly to moderately sorted, and all have unimodal grain-size distributions whose main modes vary with depth, reflecting grading (Fig. 5). The four sub-layers within MD3 have their primary modes at between − 2.5 and – 4 ϕ. In order of abundance (Fig. 5), layers MD1 to MD3 comprise dense dark-gray and dark-brown vesicular and microvesicular juvenile clasts (scoriaceous), dense dark-gray/ black non-vesicular basaltic juvenile clasts, other accessory and accidental lithics, and free crystals (hornblende, plagioclase, Fe-Ti oxides, pyroxene, and rare olivine). In MD1, while vesicular juvenile clasts increase from 74 vol% at the base to 77 vol% at the top, with dense juvenile clasts increasing from 8 to 12 vol%, accessory lithic contents decrease from 12 to 4 vol%. In MD2, all components are vertically homogeneous (76 vol% vesicular juvenile clasts, 14 vol% non-vesicular juvenile clasts, 8 vol% free crystals, and 2 vol% accessory lithics). In MD3, the volume percentage of vesicular juvenile clasts decreases from 76 vol% at the base of the layer, through 40 vol% across middle levels, before increasing again to 76 vol% at the top. Accessory and accidental (gabbroic) lithics together comprise 1 vol% at the lowermost levels of MD3, 27 vol% in the middle of the layer, and back to 4 vol% at the top. Non-vesicular juvenile clasts (11–14 vol%) and free crystals (7–9 vol%) are generally homogeneous throughout MD3 (Fig. 5). MD1 to MD3 have homogeneous vesicular and microvesicular juvenile clast textures, which vary only in color (dark-gray to dark-brown). Within each layer, vesicular juvenile clasts become denser from base to top (Fig. 5). Density ranges from 1.9 to 2.0 g/cm3 in MD1 (35–40 vol% bulk porosity), 1.7 to 2.5 g/cm3 in MD2 (45 to 25 vol% bulk porosity), and 1.6 to 2.2 g/cm3 in MD3 (50 to 35 vol% bulk porosity). The low isolated porosities of the vesicular juvenile clasts of the Manganui-D succession (0.3–0.7 vol%; Online Resource 1) indicate high vesicle interconnectivity. Moreover, the total average solid density of the Manganui-D vesicular juvenile clasts (3.0 g/cm3) is consistent with the solid density of the corresponding dense juvenile clasts (2.9 g/cm 3; Table 2). The ~ 1200 cal BP Kaupokonui eruptive episode The Kaupokonui eruptive episode, as defined Torres-Orozco et al. (2017), produced a single fallout SE of the volcano (Whitehead 1976; Fig. 2). In proximal sections, the Kaupokonui fallout formed an 18–30-cm-thick layer of

Bull Volcanol (2017) 79:76

reversely graded, fine to coarse-grained subangular lapilli (Fig. 6 and Online Resource 1). In ring plain sections, the Kaupokonui layer is a 5–15-cm-thick, reversely graded deposit of coarse ash to fine lapilli. Deposits of the Kaupokonui layer are very poorly to poorly sorted and have unimodal grain-size distributions with a mode of − 2.5 ϕ at the base and − 4.5 ϕ at the top (Fig. 6). On the ring plain, Whitehead’s (1976) data show an overall grain-size with modes between 0 and 1 ϕ. Kaupokonui deposits consist of finely to coarsely vesicular and crystal-rich brown pumice (54 vol%), gray and dark-gray andesitic clasts (35 vol%), dark-red or dark-violet accessory lithics (8 vol%), and free crystals (4 vol% of plagioclase, hornblende, Fe-Ti oxides, and pyroxene). Kaupokonui pumice clasts become slightly less dense with height in the layer, decreasing from 1.3 to 1.2 g/cm3 (55 to 60% bulk porosities; Fig. 6). The isolated porosities of pumice clasts, sampled from different vertical levels (base, middle, top) within the Kaupokonui layer, range from 2 to 6 vol% (Online Resource 1), indicating high vesicle interconnectivity (i.e., poor preservation of isolated vesicles). In addition, the total average solid densities of both pumice and dense andesitic juvenile clasts are equivalent (2.6 g/cm3; Table 2). The AD 1655 Burrell eruptive episode The Burrell eruptive episode produced a fall deposit (Fig. 2), which represents the most recent large explosive event of Mt. Taranaki (Druce 1966; Topping 1971; Platz et al. 2007). The Burrell fallout was generated from a 14-km-high eruption column, following a lava-dome collapse at the summit crater (Platz et al. 2007), and it is represented by a single layer, best exposed to the east (Torres-Orozco et al. 2017). In proximal sections, the Burrell fallout formed a 50-cm-thick, variably graded, fine to coarse, angular to subrounded lapilli layer (Fig. 6and Online Resource 1), which thins on the ring plain into a 5–10-cm-thick and normally graded deposit of finemedium lapilli and coarse ash. At proximal sites, the Burrell fall deposit ranges from very poorly to moderately well sorted and have unimodal grainsize distributions with a mode of – 3 ϕ at the base, and − 4.5ϕ at the top (Fig. 6). In ring plain sections, the overall grain-size mode is between − 1 and 1 ϕ (Whitehead 1976). The lithology comprises gray and brownish dark-gray pumice, bluish gray and dark-gray dense andesitic juvenile clasts, accessory lithics (red iron-stained and white-bleached clasts), and free crystals (plagioclase, hornblende, Fe-Ti oxides, pyroxene, and biotite). In middle stratigraphic levels, the pumice content decreases (base 79 vol%, middle 69 vol%, and top 81 vol%), and the dense andesitic clast content increases (24 vol%). The content of accessory lithics slightly increases upwards from 3 to 6 vol%, whilst the content of free crystals remains uniform (2–3 vol%; Fig. 6).

Bull Volcanol (2017) 79:76

Page 15 of 27 76 Studied in this work

Trachy-andesite

Burrell * Basaltic trachy-andesite

Na2O + K2O wt.%

7

Kaupokonui * Manganui-D Upper Inglewood Kokowai, Kw7 Kokowai, Kw4

Trachy-basalt 5

Manganui-D (May 2003) 4.6 ka (Kokowai) pumice deposits (Turner 2008)

Other Mt. Taranaki-sourced pyroclastic deposits

Andesite

Maero Formation * Manganui-A and Manganui-C Manganui Formation *

Basaltic andesite Basalt

3.6-3.3 ka (Lower and Upper Inglewood) deposits *

3 52

47

57

9-1.5 ka pyroclastic deposits *

62

SiO2 wt.% 13

Fe2O3 (t) wt.%

MgO wt.%

7

CaO wt.%

13

11 11

5 9 9

3

7

7

5 22

5

1 6

Al2O3 wt.%

Na2O wt.%

K2O wt.% 3

20

5

18

4 2

16

3

14

2

1.3

1

TiO2 wt.%

MnO wt.%

P2O5 wt.%

0.4

0.2

1.1

0.3 0.9

0.2 0.7

0.5 47

57

52

SiO2 wt.%

62

0.1 47

57

52

SiO2 wt.%

62

0.1 47

57

52

62

SiO2 wt.%

Fig. 7 Whole-rock analyses of volcanic rocks from Mt. Taranaki, normalized to anhydrous basis, plotted on the total alkalis vs. silica diagram modified from Le bas et al. (1986), and on binary diagrams of some major elements. Dotted line discriminates between alkaline (above

the line) and sub-alkaline (below the line) series (Irvine and Baragar 1971). Data taken from previous work (*) correspond to Franks (1984), May (2003), Turner (2008), Turner et al. (2011b) and Platz et al. (2007, 2012)

Two types of pumice were identified in the Burrell fall deposit: finely vesicular gray pumice and densemicrovesicular brownish to dark-gray pumice. The basal stratigraphic level of the deposit is only composed of gray pumice with a bimodal bulk density distribution (modes at 1.0 and 1.1 g/cm3, 65 vol% bulk porosity; Fig. 6). The middle and top levels of the deposit contain gray pumice with a bulk

density of 0.7–0.8 g/cm3 (70–75 vol% bulk porosity) and brownish to dark-gray pumice with a bimodal bulk density distribution (1.0 and 1.3 g/cm3, 55 to 65 vol% bulk porosities; Fig. 6). The isolated porosities of pumice clasts from the Burrell fall deposit range from 3 to 12 vol%, indicating good vesicle interconnectivity (Online Resource 1). In addition, both

Bull Volcanol (2017) 79:76

Tasman Sea

a

Manganui-D

Waitara

90

90

Page 16 of 27

b

Burrell

Tasman Sea

(modified from Platz et al. 2007)

Waitara

New Plymouth New Plymouth

0.2 3

1

Stratford

20

1 Eltham

2

5620000m.S

Hawera

20 km

90

1700000m.E

50

Upper Inglewood (Uig7)

Tasman Sea Waitara

c

SH3

SH3

SH45

Stratford

10

8 6

Eltham

5620000m.S

6

7 11

40 27 22 35 9 9 62 20 41 11 34 23 14

1

43

SH Inglewood

20

40

8

5

24 23 24 25

10

20 20 33 20

10

14

43

SH

Inglewood

90

9

SH45

Hawera

20 km

1700000m.E

50

d

Kaupokonui

Tasman Sea

(modified from Whitehead 1976)

Waitara New Plymouth

New Plymouth

0.3 1 7 11

4 1

5

Eltham

20 km

1700000m.E

50

Tasman Sea

Kokowai (Kw4)

e

5620000m.S

2

Hawera

Waitara

Stratford

6 19 12 18 10 10

17 10

SH3

SH45

5 15 9 2 4

14

3 4

5 SH45

20 km

Hawera

1700000m.E

50

Tasman Sea

2 New Plymouth

4

4

20

22 25

83

25

Isopach Area (cm-thick) 43

SH

176 83

Stratford

20

40

90

22 Stratford

Inglewood

SH3

43

SH

23 170

5

10

5

2

5 Inglewood

40

40 46

20 km 1

New Plymouth

4

SH3 Hawera

1700000m.E

20 km 50

5620000m.S

Eltham

SH45

f

Kokowai (Kw7) Waitara

20

5620000m.S

SH Inglewood

10

20

20 24 21 19 12 20 24 24 15 27 28 18 16 50 10 19 18 21 10 Stratford 20 10 10 10 Eltham

90

43

SH

7

SH3

7 7

43

Inglewood

21 17 21

90

10

5620000m.S

76

SH45

1700000m.E

Study sites Summit crater Fanthams Peak vent Extrapolated isopach Interpolated isopach Roads Residential area

1

>150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 <10 50

Bull Volcanol (2017) 79:76

ƒFig. 8

Isopach maps of fall deposits corresponding to bed-sets: a Manganui-D, layers MD1 to MD3, sourced at the satellite vent of Fanthams Peak; b Burrell, modified from Platz et al. (2007); c Upper Inglewood, layer Uig7; d Kaupokonui, modified and corrected from Whitehead (1976); e Kokowai, layer Kw4; and f Kokowai, layer Kw7. Gray numbers represent thickness data measured in centimeters. Some contours are labeled with black numbers inside white squares (in centimeters). Black contours were interpolated from field data (Online Resource 1) using methods described in the text. Red contours were partially or fully extrapolated using parameters described in the methodology. Coordinate system: NZGD 2000 New Zealand Transverse Mercator

pumice and dense juvenile clasts have the same total average solid density of 2.9 g/cm3 (Table 2). Our analyses of the Burrell fall deposit are consistent with earlier results obtained by Platz et al. (2007).

Whole-rock chemistry A total of 29 whole-rock analyses were carried out (Online Resource 2) on pumice, vesicular/microvesicular juvenile clasts (scoriaceous), and dense juvenile clasts of different stratigraphic levels, sampled from the Kokowai, Upper Inglewood, and Manganui-D bed-sets. Vesicular juvenile clasts of the Manganui-A and Manganui-C members (Torres-Orozco et al. 2017) of the Manganui Formation were also analyzed for comparison. Whole-rock data of pumice and dense juvenile clasts of the Burrell and Kaupokonui fall deposits were taken from Franks (1984), May (2003), Platz et al. (2007), and Turner et al. (2011a, b). Samples from Mt. Taranaki summit are sub-alkaline basaltic trachy-andesites and trachy-andesites (52.8 to 60.0 wt% SiO2, 5.3 to 7.7 wt% NaO + K2O; Fig. 7), with internally homogeneous pumice clasts (Online Resource 2). Kokowai pumice data cluster at 53.5–54.3 wt% SiO 2 , Upper Inglewood at 58.5–58.8 wt% SiO2, and Burrell at 56.0– 57.0 wt% SiO2. Deposits produced by the most explosive late-Holocene eruptions from Fanthams Peak (e.g., Manganui-A, Manganui-C, and Manganui-D) vary in composition from alkaline and sub-alkaline basalts to rare trachybasalts (47.9 to 517 wt.% SiO2, 3.6 to 5.3 wt% NaO + K2O; Fig. 7). In particular, dense juvenile clasts of the Manganui-D bed-set are generally more silicic than their respective vesicular juvenile clast compositions (50.6 vs. 48.3 wt% SiO2; Online Resource 2). The Manganui compositions are consistent with the silica contents of deposits formed by less explosive and effusive eruptions produced at the same vent (48.0– 53.0 wt% SiO2; Stewart et al. 1996; May 2003; Turner et al. 2011b). Deposits of the Kaupokonui bed-set are alkaline and subalkaline basaltic trachy-andesites (52.5–55.5 wt% SiO2), being intermediate between the summit and the Fanthams Peak series in the binary diagrams of SiO2 vs. MgO, Al2O3,

Page 17 of 27 76

and TiO2 (Fig. 7). In addition, the dense juvenile clasts of the Kaupokonui are more silicic than the corresponding pumice compositions (57.3 vs. 54.0 wt% SiO 2 , Online Resource 2).

Eruptive metrics Dispersal The isopach and isopleth maps constructed in this work (Figs. 8 and 9) are characterized by elliptical and some subcircular shapes. While isopachs have shape factors of 0.4 to 1.0, aspect ratios of 1.0 to 5.4, and eccentricities of 0.7 to 1.0 (Online Resource 1), isopleths have shape factors of 0.7 to 1.0, aspect ratios of 1.0 to 2.7, and eccentricities of 0.7 to 1.0 (Online Resource 1). The most widespread fallouts correspond to deposits of fall layers MD1–MD3, Uig7, and Kw4, which have the largest isopachs (A1/2 of d5 cm = 39, 32, and 29 km, correspondingly; Fig. 2) and isopleths (A1/2 of d1.6 cm = 13–14, 9–14, and 22–25 km, respectively; Online Resource 1) produced in this work. In contrast, the Kaupokonui and Burrell fall deposits have smaller isopachs (A1/2 of d5 cm = 20 and 27 km, respectively; Fig. 2) and isopleths (A1/2 of d1.6 cm = 6–7 km). Fall layer Kw7 has the smallest isopachs (A1/2 of d5 cm = 18 km; Fig. 2) generated in this work, but has isopleths (A1/2 of d1.6 cm = 14–16 km) comparable to the largest fallouts, suggesting that the Kw7 deposit thickness is underestimated. In every case, the fallout thickness decreases exponentially with distance from the vent (Fig. 2), and the decay can be modeled by two or three linear segments (i.e., in Log 10 thickness vs. A 1/2 diagrams; Online Resource 3; cf. Bonadonna and Houghton 2005). The isopachs of > 30 cm and isopleths of > 1.6 cm of fall deposits of layers Uig7 and MD1–MD3 indicate an initial proximal dispersal to the northeast, before the main axis of deposition was set eastwards from the volcano (Figs. 8 and 9). A similar pattern, but towards southeast, was recognized for the Burrell fallout, as also noted by Platz et al. (2007). This is consistent with Plinian eruption columns driven by their own convective forces before being affected by prevailing winds at stratospheric neutral buoyancy levels (Carey and Sparks 1986). In contrast, dispersals of the Kaupokonui and the two Kokowai fall layers (Kw4 and Kw7) were constant throughout. The Kaupokonui fall deposit was distributed to the southeast, whereas fall layers Kw4 and Kw7 were dispersed to the northeast and north-northeast, respectively (Figs. 8 and 9). Maps of the distribution of PDC deposits indicate that block-and-ash flow (Uig1 to Uig3; Fig. 2) and pumice-rich PDC deposits (Uig4 to Uig6; Fig. 2) were formed directly to the east of Mt. Taranaki during the Upper Inglewood eruptive episode (Fig. 10). In contrast, similar block-and-ash flow and pumice-rich PDC deposits produced during the Kokowai

Bull Volcanol (2017) 79:76

Page 18 of 27

2.8 0.8 1 0.8 1.5

Stratford

0.2

2 0.5 2.5 1.1 1 0.8

0.5

0.7

0.4

0.5

Eltham 1690000m.E

New Plymouth

25

Inglewood 1 1 0.3 1.4 1.2 2 2 0.9 1.4 10 km 1.2 4 3 3 1.3 0.9 6 1.5 1 0.8 3.5 Stratford 0.8 5 0.7 0.9 0.6 0.6 0.3 0.4 1.4

0.1

0.3

e 25

3.2

Stratford

Kw4 pumice

i

1685000m.E 25

New Plymouth

Stratford

Kw7 pumice Eltham

25

New Plymouth

1.5

1.6

1685000m.E

1.2

1.7 1.2

0.9

1.2 0.8

0.8

Eltham

2.4

2 3.7

4.3

4 2.4

5.2

Stratford 0.2

h

5

5

1685000m.E

Fig. 9 Isopleth maps of the diameters of pumice clasts, vesicular juvenile clasts (scoriaceous) of the Manganui-D bed-set, and dense andesitic clasts in centimeters, corresponding to fall deposits of bed-sets: a Burrell, modified from Platz et al. (2007); b and c Kaupokonui; d and f Manganui-D, layers MD1 to MD3, sourced at the satellite vent of Fanthams Peak; g and h Upper Inglewood, layer Uig7; i and j Kokowai, layer Kw4; and k and l

Stratford

2.5

Kw4 dense clast Eltham

4

j

1685000m.E

0.2

3.2

10 km

2.4

1.4 1.2 2 1.5 2.6 1.3

1.6

1 0.9

5635000m.S

5635000m.S

3

1 1.3

2

3.2

1.2 1.7

1

0.8 2.2

0.3 Inglewood

10 km

Inglewood 2.8

5635000m.S

0.9

0.9

0.2 1.4

10 km

0.1

1

0.1 0.9 Inglewood 0.8 0.9 0.9

0.8

0.8

k

1685000m.E

75

Eltham

75

0.1

5.3 4.2

5635000m.S

5635000m.S

5 3.1

1.6

25

4

6.3 5.7 5.4

75

Uig7 dense clast

New Plymouth

g

0.5

0.5

2.5 3

2

0.8

1

5635000m.S

Eltham 0.1

5.7

5.3 4.6

10 km

0.3

Inglewood

1.6

4.2

0.2

10 km

3.2

1.3 1.5 1.6 1.2 1.9 1.8 1.8 1.3 1 3 3.5 2.2 1.9 1.5 5 3 1.9 1.7 4 1.2 Stratford 1.9 1.8 1.3 1.2 1.9 1.3 1.2 0.5

1685000m.E

25

New Plymouth

0.8 0.2

4

1 1.2

Uig7 pumice

3.2 Inglewood

10 km

1 Inglewood 1 1.2 1.2

0.3

f

1.5

0.8 1.2

Eltham

75

75

New Plymouth

0.4

1685000m.E

75

Eltham

1685000m.E

5635000m.S

0.3

1 1 Inglewood 1 1.2 1 1 1.5 1.2 1.4 0.9 1.3 2 1.4 10 km 1.5 4 3.5 1.4 0.9 5 6 1.5 1.6 Stratford 2 1 2 1.4 0.9 1.5 1.5 0.9 0.4 0.5

0.8

0.8

5635000m.S

0.1

d

0.3

Oakura

0.1

1

0.8

6

1.

5635000m.S

0.2 0.3

0.1 1

MD3 vesicular clast 25

1

1

25

New Plymouth

1690000m.E

0.2 Oakura

1685000m.E

Stratford

Eltham

New Plymouth

75

MD2 vesicular clast

3

1.8 0.5 1.6

75

3.4

1 0.4 0.2 1 0.5 0.2 2.5 2 0.5 0.3 0.9 1 0.5 0.7 0.6 0.3 0.5 0.3 0.3 3.5

0.2

0.7

2.4 1.6

MD1 vesicular clast 25

0.3 1

75

1690000m.E

0.5 4.5

5635000m.S

a

0.2 0.3 0.8 0.9 Inglewood 1 0.9 0.5 1 1 0.9 1.5 1 1.2 1.2 0.8 1 1.3 2 10 km 1.6 0.8 3 2.2 0.8 3 5 0.8 3 Stratford 0.8 1.9 0.6 0.6 0.5 0.4 0.2 Eltham 0.3

60 10 km 0.3

0.5

5635000m.S

3

2

Eltham

15

Inglewood

Kaupokonui dense clast

0.1

Oakura

c

0.8

4

15

Inglewood

10 km

Isopleth Area (cm-diameter)

10 km

Kaupokonui pumice

0.8

dense clast (modified from Platz et al. 2007)

>5 3.2-5 ~2-3.2 1-2 0.8 <0.8

New Plymouth

b

0.8 0.1

5635000m.S

15

Burrell

60

Inglewood

Study sites Summit crater Fanthams Peak vent Extrapolated isopleth Interpolated isopleth National Park limit Roads Residential area

60

76

3

Stratford

Kw7 dense clast Eltham

l

Kokowai, layer Kw7. Gray numbers indicate average diameters measured in the field. Black contours were interpolated from field data (Online Resource 1) using methods described in the text. Red contours were partially extrapolated using parameters described in the methodology. Coordinate system: NZGD 2000 New Zealand Transverse Mercator

a rk lP i ona

55

Egmo nt Nat

Summit crater Study sites Roads Residential area

50

on

tN Egmon

ati

ar k 55

Layer Uig4-6 Uig3 Uig2 Uig1

50

Upper Inglewood

al P

a

Page 19 of 27 76 5660000m.S

Bull Volcanol (2017) 79:76

45

Stratford 5 km

1690000m.E

1700

10

5645000m.S

Layer Kw3,6,8 Kw2 Kw1

b

Kokowai 1690000m.E

5 km 95

1700

Fig. 10 Maps of the area of distribution of pyroclastic density current deposits from bed-set Upper Inglewood (layers Uig1 to Uig6) and Kokowai (layers Kw1 to Kw3, Kw6, and Kw8). Coordinate system: NZGD 2000 New Zealand Transverse Mercator

eruptive episode (Fig. 2) were distributed directly to north and south of the volcano (Fig. 10). Eruptive volumes Total volumes (averaged across multiple methods; Table 1) of each fall deposit range from 0.1 to 1.1 km3 with uncertainties of ± 0.01–0.3 km3. Volumes calculated from one segment by the methods of Cole and Stephenson (1972) and Pyle (1989), using V = 13.08(To × bt2) (where Bbt^ corresponds to the half distance of the maximum thickness extrapolated), are consistent within uncertainty of the overall average. Volumes calculated from two and three segments following Fierstein and Nathenson (1992) and Bonadonna and Houghton (2005) are consistent with the results generated by using other methods (e.g., Table 1). Following the method of Sulpizio (2005), Kw7 fall layer was the only case in which the proximal volume divided by the total volume was larger than 0.7 (Vp/ Vt > 0.7; Table 1), yet the resulting volume was one order of magnitude higher than the results produced by using other methods (Table 1). Differences between volumes calculated from combined interpolated and extrapolated isopachs and those from solely field-interpolated isopachs generally range from ± 0.01 to ± 0.2 km3, but are significantly higher in some cases (e.g., volumes of layer Kw4 calculated using Bonadonna and Costa 2012, 2013). Nevertheless, in the case of fall layer Kw4, we preferred the volumes calculated by the method of Bonadonna and Costa (2012, 2013) (Table 1) because they represent the best fit of this work’s data to the Weibull distribution. For evaluation of the results obtained by most empirical methods, all volumes were re-calculated b y u s in g A s h C a l c 1 . 1 o f D a g g i t t e t a l . ( 2 0 1 4 ) . Corresponding dense-rock-equivalent (DRE) fallout volumes

are ~ 0.1 km3 for Burrell, Kaupokonui, and layer Kw7, and 0.2–0.5 km3 for Manganui-D, layer Uig7, and layer Kw4 (Table 2). Minimum volumes calculated from PDC deposits produced by the Upper Inglewood and Kokowai eruptive episodes (Fig. 2) range from 0.0004 to 0.05 km3 (Table 1). Block-and-ash flow and laterally directed PDCs represented by the deposits of Uig2 and Uig4 to Uig6 (Fig. 10) have the largest volumes, being 0.014 and 0.022 km3, respectively, compared with a total PDC volume of 0.05 km3 generated during the Upper Inglewood eruption (Table 1). In comparison, the largest block-and-ash flow (layers Kw1 to Kw2) and the largest column-collapse PDC deposits (layer Kw3) produced during the Kokowai eruption (Figs. 2 and 10) represent 0.005 km3 of the total PDC volume of 0.013 km3 (Table 1). Small 0.001–0.003 km3 block-and-ash flow and column-collapse PDC deposits (0.0007–0.002 km3 DRE) were produced by the Burrell eruption (Platz et al. 2007). Corresponding total PDC DRE volumes produced during the analyzed eruptions range from 0.002 to 0.02 km3 (Table 2). Column heights, wind speed, and classification Minimum, maximum, and total average column heights were determined for the studied fallouts. For simplicity, the height estimates obtained by applying different methods (Online Resource 1) were represented using a normal distribution (Fig. 11a). Heights obtained by the method of Carey and Sparks (1986) (Fig. 11b, c), based on isopleth maps, were in nearly every case consistent with those calculated by the methods of Sulpizio (2005) and Bonadonna and Costa (2013). Heights calculated using the half distance of the

76

Bull Volcanol (2017) 79:76

Page 20 of 27

Dispersal index (km2)

5

10

15

20

25

30

35 4 0.1

avg. HT (km)

nia n

n

Pli

coarse

3

0

Total population

27

MD

bc = 15 km

21

2

km

0.05

24

24.7

bc = 8

19

16.8

-Pl inia

Uig7

0.1

sub

22

16.2

km

Kw4

Bu

1

bc = 3

28.5

km

26.4

22

19

14.2

11.7

Surtseyan

bc = 1

Kw7

Kp

11.8

50000

27.5

14

0.15

500

d

Strombolian

HT Distribution

0

20.6

0.2

5

fine

a

Half-distance ratio (bc/bt)

0.25

cone-like

sheet-like 1

10

100

Thickness half-distance bt (km )

20

m

10

0

0.8-2 cm (1-2.5 x103 kg/m3) Wind speeds 15-28 m/s

m

/s

/s

/s

20

m

30

m

ultr

a-P

Plin

ian

35.6 km

sub

1

-Pl

λML / λth

Crosswind range (km)

/s

Diameters 30

28.3 km 10

e

lini

an

inia

sm

41

n

all-

Mo

der

ate

b

6.8 km

0

0

10

20

30

40

50

10

0.1

10

14

km

-HT

60

Maximum downwind range (km)

-HT

24

21.0 km 13.8 km

km

km

-HT

km

-HT

100

λth

30

Mass eruption rate (kg/s) 107 108

30

m

f 30

35.6 km

21.0 km 13.8 km 6.8 km

0 0

10

20

c 30

40

20

~1-2E8 kg/s ~5-8E4 m3/s

HT 27.5

Spa

28.3 km

HT

~5-8E7 kg/s HT 22

HT (

avg. H (km)

25 10

109

35

986)

/s m

m

106

/s

20

m

10

/s

rks 1

20

/s

1.6-4 cm (1-2.5 x103 kg/m3) Wind speeds 15-33 m/s

0

~2-3.4E4 m3/s

HB

15

~8E6 -1.5E7 kg/s HT 14

50

~3.5-6E3 m3/s

Q

10

ER

M

Maximum downwind range (km)

5 0

Burrell (Bu)

2

4

sub-Plinian

Plinian

104

105

103

6

bc (km)

Kaupokonui (Kp)

106

Volume eruption rate (m3/s)

Manganui-D, layers MD1-3 Whole deposit Upper inglewood (Uig1-7) Upper Inglewood, layer Uig7

30

Kokowai, layer Kw7 Kokowai, layer Kw4 AD 1913 Colima (Saucedo et al. 2010) AD 1300 Mt. Pelee (P1, Carazzo et al. 2012) AD 79 Vesuvius (Carey and Sparks 1986) 122 BC Etna (Coltelli et al. 1998) ~10 ka Tongariro (U-Pk, Pardo et al. 2012) 12.1 ka Nevado de Toluca (MTP, Arce et al. 2005) 12.4 ka Apoyeque (UAq, Avellan et al. 2014) 16 ka Vesuvius (L0-L3, L5, Cioni et al. 2003)

Original sources in Bonadonna and Costa (2013) AD 1982 Chichon A AD 1912 Novaraptu C AD 1886 Tarawera AD 1563 Fogo

Etna LTP Colima

25

h

i

SP MTP

20 SP

AD 186 Hatepe 0.8 ka Cotopaxi L3 2 ka Montaña Blanca

<21.8 ka Ruapehu (Mgt, Pardo et al. 2012)

4.1 ka Agnano M Spina (D1) 4.1 ka Agnano M Spina (B1) 4.6 ka Fogo A

23.5 ka Tacana (SP, Arce et al. 2012)

>60 ka Fontana Lapilli E

21.7 ka Nevado de Toluca (LTP, Capra et al. 2006)

g

UAq

Whole deposit Kokowai (Kw1-8)

HT (km)

Crosswind range (km)

Diameters

15 0

1

2

Min volume DRE (km3)

3 4

4.5

5

Log10 mT[Kg]-7

0

1

2

3

Min duration (hr)

4

Bull Volcanol (2017) 79:76

ƒFig. 11

Eruptive parameters and classification of the studied deposits. a Normal distribution of total column heights (HT) in kilometers, calculated by using different methods (Table 1 and Online Resource 1). Numbers indicate minimum, maximum, and average HT corresponding to individual fall deposits. b and c Isopleth data plotted in the diagram of Carey and Sparks (1986) to determine HT and wind speeds, based on diameters of dense andesitic clasts and pumice clasts (or vesicular juvenile clasts of the Manganui-D bed-set), ranging from 0.8 to 2 and 1.6 to 4 cm, respectively, and on the corresponding clast’s bulk densities (kg/ m3). d Diagram of Pyle (1989) of classification of the eruptions based on parameters (bc, bt) calculated from isopach and average isopleth (pumice and dense andesitic clast) data, and on the dispersal index of Walker (1973) (Online Resource 1). e Diagram of Bonadonna and Costa (2013) of classification of eruptions based on HT, and on parameters (λth, λML) calculated from isopach and isopleth data (Online Resource 1). f Diagram modified from Sparks (1986) and Carey and Bursik (2000) to determine volume and mass eruption rates (Q and MER, respectively), considering the average HT calculated from each individual fall deposit. The curve of HT calculated by using the model of Sparks (1986) is indicated. Neutral buoyancy column heights (HB) were estimated by using the average HT of each fall deposit and Bbc^ of Pyle (1989) (Online Resource 1). Fields corresponding to Plinian and sub-Plinian eruptions were modified from Bonadonna and Costa (2013). g Plot of average HT vs. minimum denserock-equivalent (DRE) eruptive volumes. The latter were calculated from average total volumes, and bulk and solid densities (Table 2). h Plot of average HT vs. magnitude (M = Log10(mT)-7) calculated using the method of Pyle (2000). Total mass in kilograms (mT = (solid density) (DRE volume)) calculated using the method of Wilson (1976). i Plot of average HT vs. minimum duration of the eruption, in hours (T = mT/MER; Table 2), estimated using the method of Wilson (1976)

maximum clasts extrapolated from the isopleth data (bc; Pyle 1989), following the methods of Sparks (1986) and Pyle (1989), were unrealistically high and out of range from the other results (Online Resource 1) and thus were not considered in the final results. However, the total average heights calculated using Carey and Sparks (1986), Sulpizio (2005), and Bonadonna and Costa (2013) were substituted into the equations of Sparks (1986) and Pyle (1989) so as to calculate minimum bc and the maximum neutral buoyancy column heights. The maximum estimated average column height of 27.5 km was for layer Kw4 (Fig. 11a). The next greatest average height was 24.7 km for layers MD1–MD3. The lowest average column heights of 16–17 km were estimated for Burrell and Kaupokonui fallouts (Fig. 11a). By applying the method of graphical interpolation and extrapolation of Carey and Sparks (1986), wind speeds of 15 to 33 m/s were estimated as affecting all the eruptive columns studied in this work, with the strongest winds occurring during the Kaupokonui and Kokowai eruptive episodes (Fig. 11b, c). The bt and bc parameters (Pyle 1989), defined above, were calculated from the isopach and isopleth maps of each fall deposit (Online Resource 1). Layers MD1 to MD3, Uig7, and Kw4 have bt values consistent with ranges expected for Plinian eruptions and have dispersal areas inside the 0.01Tmax isopach, as defined by Walker (1973), of 2000 to

Page 21 of 27 76

4000 km2 (Fig. 11d). In contrast, Burrell, Kaupokonui, and layer Kw7 have bt values that correspond to sub-Plinian eruptions, and have dispersal areas inside the 0.01Tmax isopach of 600 to 2200 km2 (Fig. 11d). These results are consistent with the classification of eruptions into Plinian and sub-Plinian categories, as defined by Bonadonna and Costa (2013) (Fig. 11e). Eruption rates, magnitudes and durations Fall layer Kw4 has the highest volumetric and mass eruption rates of 5–8 × 104 m3/s and 1–2 × 108 kg/s, respectively (Table 2), consistent with Plinian eruptions, following the classification of Bonadonna and Costa (2013) (Fig. 11f). Likewise, deposits of layers Uig7 and MD1–MD3 have also Plinian volumetric and mass eruption rates of 104 m3/s and 107–108 kg/s, correspondingly (Fig. 11f). Layer Kw7 plots at the uppermost limit of sub-Plinian eruptions, whereas Burrell and Kaupokonui sit at the lowermost sub-Plinian limit, having volumetric and mass eruption rates of 103 m3/s and 106– 107 kg/s, respectively (Fig. 11f). Eruption magnitudes range from 4.7 to 5.1 (Table 2) for the Plinian columns that produced layers MD1–MD3, Uig7, and Kw4 (Fig. 11h) and range from 4.1 to 4.3 for the sub-Plinian columns that formed the rest of layers studied. In contrast, PDC deposits were associated with magnitudes of 2.7–3.7 (Table 2). Based on relationships defined by Wilson (1976), the minimum duration of each eruption was roughly three to four hours for the Burrell and Kaupokonui eruptive episodes, one hour for layer Kw7, two hours for layers Uig7 and MD1– MD3, and one to two hours for layer Kw4 (Table 2 and Fig. 11i). By roughly integrating PDC deposits, minimum durations of two and a half to three hours result for the Upper Inglewood and the Kokowai eruptive episodes (Fig. 11i).

Discussion Constraints on the eruptive volume estimates Due to the poor preservation in medial and distal areas of the different fall deposits studied in this work, most volumes were expected to be underestimated as all empirical strategies applied would not be capable of capturing the actual tephra deposit thinning. Thereafter, the average total volumes are minimum estimates in every case (Table 1). In fact, the identification of the break-in-slope between segments (Online Resource 3), when applying the exponential strategies of Fierstein and Nathenson (1992) and Bonadonna and Houghton (2005), is often arbitrary due to the lack of data, and this can significantly affect the final result (e.g., Bonadonna and Costa 2012, 2013). Furthermore, the Weibull strategy is more sensitive to

76

Page 22 of 27

the distal points than the exponential strategy (Bonadonna and Costa 2012, 2013), and consequently, the resulting volumes will be undervalued if the distal thickness is underestimated (as may be the case here for extrapolated distal fall deposits of layers MD1–MD3, Uig7, and Kw4). For this reason, volumes calculated from solely field-interpolated data were preferred when applying the Weibull method of Bonadonna and Costa (2012, 2013), yet volumes calculated using exponential strategies with two and three segments were considered to be likely the most reliable by being less sensitive to the missing distal data. To test such reliability, the identification of segments and all volumes estimated manually were compared to segments and volumes estimated automatically, applying the same exponential and Weibull strategies in AshCalc 1.1 of Daggitt et al. (2014). Most results generated by AshCalc were consistent with volumes calculated manually, and improved the 2-segment interpolated-extrapolated and interpolated volumes of layers Uig7 and Kw4, respectively (Table 1). However, the re-segmentation had no significant impact on the 3-segment volumes (e.g., the fall deposit of Kaupokonui, and layers MD1–MD3 and Uig7; Table 1) owing to the underrated distal data, which produces inconsistencies (e.g., 3-segment < 2-segment volumes). Mt. Taranaki summit vent: eruptive successions and processes Of the < 5000-year-old units examined here, the Plinian phases represented by layers Kw4 and Uig7 record the highest energy and largest volume eruptions produced at the summit vent of Mt. Taranaki. Smaller sub-Plinian phases are represented by layer Kw7 and Burrell fall deposits. Overall, erupted fallout volumes ranged from 0.1 to 1.1 km3 (0.05 to 0.5 km3 DRE) and column heights from 12 to 29 km, with volumetric eruption rates of 103–104 m3/s, mass eruption rates of 107– 108 kg/s, and durations lasting roughly one to four hours (Fig. 11). The largest eruption, represented by layer Kw4, had a magnitude of 5.1. Its peak column height and fall dispersal were comparable to Plinian eruptions such as Hatepe (AD 186), Agnano (4.1 ka—B1), and Tarawera (AD 1886; original sources in Bonadonna and Costa 2013; Fig. 11). Tephra dispersal from this episode was also comparable to the largest known eruptions at the larger North Island stratovolcano Mt. Ruapehu (cf. < 21,800 cal BP—Mgt; Pardo et al. 2012). The next largest eruption, represented by layer Uig7, had an overall magnitude of 4.8 and a minimum DRE volume of 0.16 km3. This Plinian phase had comparable dynamics, column height, dispersal, and magnitude to the AD 1913 eruption of Colima (Saucedo et al. 2010) and the AD 1300 eruption of Mt. Pelée (Carazzo et al. 2012; Fig. 11). Sub-Plinian eruptions represented by layer Kw7 and the Burrell fall deposit were comparable in dispersal and column height to sub-Plinian

Bull Volcanol (2017) 79:76

eruptions from Vesuvius (16 ka L0—L3 and L5; Cioni et al. 2003), Nevado de Toluca (12.1 ka MTP; Arce et al. 2005), Fogo (AD 1563; Bonadonna and Costa 2013, Fig. 11), and Tungurahua (AD 2006; Eychenne et al. 2013). All of these eruptions occurred from the summit vent, were complex, and underwent multiple phases with transitions in eruptive style. Dome-collapse block-and-ash flows commonly marked the onset of each eruptive episode, indicating that most of these episodes began with lava-dome effusion, terminated by dome collapse or blasts. Highly heterogeneous pumice porosities and textures in deposits of subsequent pyroclastic phases indicate that the conduit was partly occluded with differential domains of magma density and vesiculation (cf. Shea et al. 2011, 2012). Opening explosive phases gradually unplugged the conduit to produce unsteady and partially collapsing columns and associated PDCs. Eventually, volatilerich magma in the lower conduit was decompressed, triggering a rapid increase in explosivity and the proliferation of pumice in PDC deposits (e.g., from layers Uig1 to Uig3 during the Upper Inglewood eruptive episode). This pattern is indicative of continued magma unroofing and conduit decompression (e.g., Villemant et al. 1996; Carazzo et al. 2012) following multiple explosive lava-dome collapses. During climactic phases, Plinian eruption columns of 22– 29 km in height produced falls with steady pumice: dense clast ratios (64–85 vol% pumice, 8–13 vol% dense clasts). Only minor marginal collapses occurred from the eruption columns, producing thin ash-rich PDC deposits. The areas now occupied by the urban locations of New Plymouth, Waitara, Inglewood, Stratford, and Eltham (Figs. 8 and 9) would have experienced two hours of tephra fall. The sub-Plinian columns that deposited layer Kw7 and the Burrell fallout were 14– 22 km in height, but changes in grainsize, sorting, and pumice/dense clast contents in the corresponding fall deposits indicate, especially for layer Kw7, oscillatory columns. Throughout the post-climactic phases of the studied summit vent eruptions, progressive increases in density and pumice crystallinity were ubiquitous. These suggest progressive degassing leading to waning column heights. Such degassing could have also increased the density of the pyroclastic mixture (e.g., > 2 g/cm3 pumice), leading to intermittent column collapse (cf. Carey and Sigurdsson 1989; Shea et al. 2011, 2012; Carazzo et al. 2012). During the Upper Inglewood eruptive episode, the eruption was followed by rain-triggered lahars that rapidly remobilized primary deposits. Fanthams Peak vent: parasitic cone/flank vent eruptions The isopach and isopleth maps drawn up here (Figs. 8 and 9) indicate that the Kaupokonui fallout could have erupted from either Fanthams Peak or the summit of Mt. Taranaki (cf. Whitehead 1976; Franks 1984). Fanthams Peak was the most likely vent based on the chemical and textural (e.g., density)

Bull Volcanol (2017) 79:76

similarities between pumice and dense pyroclasts of the Kaupokonui fall deposit, and scoria and dense clasts produced during different eruptions of the Manganui Formation (Figs. 4 and 6). Moreover, the Kaupokonui, Manganui-A, ManganuiC, and Manganui-D eruptive episodes all comprised rapid onsets to climactic steady eruptive phases. The Kaupokonui eruptive episode was comparable in dispersal and column height to sub-Plinian eruptions from Vesuvius (16 ka L0–L3 and L5; Cioni et al. 2003), Nevado de Toluca (12.1 ka MTP; Arce et al. 2005), Fogo (AD 1563; Bonadonna and Costa 2013; Fig. 11), and Tungurahua (AD 2006; Eychenne et al. 2013). On the other hand, the Manganui-D eruption had dispersal characteristics comparable to the Plinian phase of the Upper Inglewood eruptive episode, represented by layer Uig7, and the 122 BC eruption of Etna (Coltelli et al. 1998; Fig. 11). The Kaupokonui and Manganui-D eruptions produced conformable and reversely graded fall deposits that represent rapid vent opening and establishment of a climactic eruption column and indicate the absence of early dome formation. In addition, these fall deposits consist of dense (1.2–1.9 g/cm3), crystalline (~ 40–50 vol%), and invariably finely vesicular to microvesicular (mainly < 0.1 mm vesicles) pyroclasts, with total porosities of 35–53 vol% (Figs. 4 and 5), and having high microlite contents. These pyroclast textures were most likely formed by late bubble nucleation (microvesicularity) induced by sudden decompression of a stalled magma (cf. Coltelli et al. 1998). Short-term magma stalling is suggested by high microlite contents, and all at once, microlite crystallization may have generated a second boiling triggering eruption (cf. Alidibirov and Dingwell 1996; Houghton et al. 2004). The Kaupokonui fall deposit was produced from a sustained and steadily rising sub-Plinian column of 12– 16 km in height (Figs. 8 and 9), lasting roughly four hours. In contrast, the Manganui-D eruptive episode produced an oscillating Plinian column, with the climax reached at heights of 24–27 km. Both eruptions produced fall deposits distributed to the east and lasting several hours. The most proximal fall deposits of the Manganui-D eruption have repetitive vertical variations in grainsize and pumice:lithic clast ratios (Fig. 4), which suggest that fluctuations in eruption intensity were possibly associated with rapidly degassing low-viscosity magma causing temporary clogging of the conduit (cf. Cashman et al. 2000; Houghton et al. 2004). The uppermost fall deposits of this succession (layer MD3) have strong stratification in particle sizes and grading and are rich in dense juvenile pyroclasts (2.2 g/cm3), accessory, and accidental lithics (Fig. 4), including gabbroic clasts. These features suggest that the final phases of the episode might have been associated with vent widening/collapse and deep conduit erosion along with unsteady eruption of increasingly gas-depleted magmas. Vertical grainsize and componentry variations in proximalmedial and distal fall deposits have been ascribed in some

Page 23 of 27 76

cases to the effects of wind shifts (cf. Cioni et al. 2000; Bonadonna et al. 2016). Estimated wind speeds varied from 18 to 29 m/s for strongly stratified Manganui-D fall deposits and from 25 to 33 m/s for non-stratified fall deposits corresponding to layers Kw4 and Uig7. The largest wind variations occurred during the Manganui-D eruptive episode and might have favored the formation of shower-bedded fall deposits. However, the isopach and isopleth maps in this study suggest that winds did not strongly displace plumes in proximal areas close to the erupting vent, which is consistent with fall deposition adjacent to the axis of rotation of the eruption column (e.g., Carey and Sparks 1986). The post-climactic phase of the Kaupokonui eruption was short lived with a rapid decrease in energy before cessation. In contrast, during the Manganui-D eruptive episode, lava flowed down the eastern valleys of Fanthams Peak and incorporated fall deposits of layer MD3. The eruption of lava suggests magma degasification and overall gas-exhaustion (cf. Houghton et al. 2004; Carazzo et al. 2012). Implications for Plinian and sub-Plinian eruptions Mt. Taranki’s eruptive episodes studied in this work were produced from two vent areas fed by different compositions. These may have resulted from two co-existing magma feeding systems above a polybaric system (cf. Zernack et al. 2012), where transient and occasionally separate magma reservoirs produced different paths of crystallization and fractionation. It is clear that the compositional variation has had an influence on the eruptive sequences. The climactic Plinian phase of the Kokowai eruptive episode represents the largest late-Holocene eruption of Mt. Taranaki (Fig. 11). This episode exemplifies the most common onset of an eruption at the summit crater, that of dome effusion, which has been recognized here and in previous studies (e.g., Turner et al. 2008c, 2011a; Platz et al. 2007, 2012; Torres-Orozco et al. 2017). Collapse of a growing or pre-existing dome may suddenly unroof a gas-rich magma to cause explosive eruptions (e.g., Villemant et al. 1996; Carazzo et al. 2012; Cronin et al. 2013). The distinguishing features of the Kokowai episode, compared to others studied here, were its larger magma volume and its homogenous composition and pumice texture. Rheologically homogenous magma-expansion, coupled with a stable vent/conduit geometry and a constant and rapid magma rise, produced steady eruption columns (cf. Carey and Sigurdsson 1989) and hindered crystallization-induced changes in viscosity (cf. Alidibirov and Dingwell 1996) and development of foam contrasts in the conduit (cf. Shea et al. 2011, 2012). The Plinian Manganui-D and Upper Inglewood eruptive episodes were nearly identical in intensity, magnitude, column height, fall deposit dispersal, volume, and duration (Fig. 11), but erupted from different vents and had contrasting magma

76

Page 24 of 27

compositions of 50 vs. 60 wt% SiO2 (Fig. 7). Similarly, the sub-Plinian Kaupokonui and Burrell episodes were almost identical in scale (Fig. 11), but were also sourced from different vents and had divergent magma compositions. In general, summit eruptions were associated with highviscosity andesitic and high-Si andesite magmas, which had experienced large degrees of pressure build-up under close-system conditions. Magma trapped within a shallow conduit for long enough would have developed variations in volatile and crystal content and viscosity. This, in-turn, favored eruptions with many unsteady phases and rapid shifts between stable and unstable eruption columns (cf. Mader 1998; Gardner et al. 1996; Cashman et al. 2000). However, even during the most unsteady summit-sourced eruptive episodes, homogeneous magma rheology also characterized the steadiest climactic column phases. In contrast, eruptions produced from the parasitic-cone Fanthams Peak were associated with low-viscosity basaltic and basaltic-andesitic magmas that experienced profuse microlite crystallization. This crystalization possibly generated second boiling and rapid microvesiculation, driving explosive eruptions (cf. Alidibirov and Dingwell 1996; Coltelli et al. 1998; Houghton et al. 2004). These examples show that the intensity and magnitude values alone, commonly employed to quantify and compare eruptions worldwide, do not provide a comprehensive hazard assessment since they are not capable of capturing and describing complex eruptive dynamics, magma composition, and changes in eruptive style. Consequently, in order to evaluate volcanic hazards, the eruptive dynamics and successions must be understood from a deep geological basis. The data also show that given similar volumes and eruptive conditions, magma compositions (often related to vent locations on stratovolcanoes) are very important in influencing how eruption sequences evolve.

Conclusions This study of five of the largest late-Holocene intermediate and mafic eruptive episodes at Mt. Taranaki has provided new insights into the maximum likely eruptive conditions at this volcano and exemplified how Plinian eruptions may vary considerably at andesitic volcanoes. Eruptions of comparable intensities and magnitudes occurred with diverse magma compositions at different vents, which radically affected eruptive styles and especially the progression of eruptive phases. The eruptive style diversity is systematic and depends on the following: (1) the volume of magma being erupted, (2) its composition, and (3) its rheological homogeneity. These hypotheses need to be tested by further investigations, but our observations can be used to define three cases of andesitic Plinian eruptive sequences, useful for emergency management. The

Bull Volcanol (2017) 79:76

first case corresponds to a rapid scaling from brief preclimactic phases and is associated with the largest volumes, typically erupted at andesitic volcanoes (± 1 km3), of rheologically homogeneous andesitic magma. The second case corresponds to smaller in volume (~ 0.1–0.3 km3), sustained or pulsating Plinian eruptions, associated with lower viscosities and mafic magmas. The third case represents the most complex eruptive sequences, which are characterized by Plinian climactic phases occurring only after long unsteady onset phases, and followed by frequent long post-climactic phases. The latter sequence is driven by the most heterogeneous melts, chemically and rheologically, and involves repeated unsteady eruption columns and many consequent PDCs. The studied eruptive episodes were produced from two separate yet associated vents. This fact yields insights into the possible separate feeding systems at volcanoes such as Mt. Taranaki. Earlier work has indicated that Fanthams Peak may produce large eruptions (Neall et al. 1986; Alloway et al. 1995; Torres-Orozco et al. 2017). In this work, we have provided evidence that demonstrates that, even if they are of similar scale, Plinian eruptions at the lower-elevation parasitic vent contrast with dome-forming andesitic Plinian explosions from the summit crater. Andesitic systems have often been hypothesized as being fed by multiple dikes that transport magma into transient polybaric reservoirs. At Mt. Taranaki, they are likely not connected, leading to near-synchronous eruption of the two vents. Tephra fall from the maximum likely eruptions at Mt. Taranaki may affect large areas of the Taranaki Peninsula, including all major urban locations (total population > 100,000; Statistics NZ 2013). From this study, it is also apparent that PDCs from this volcano may surmount high topographic obstacles and reach distances of up to 19 km from the summit. The most violent eruptions at Mt. Taranaki in the past began in two distinct ways at each of its craters. From the summit crater, where andesitic and high-silica andesite was erupted, explosive eruptions began with sudden vent unroofing by dome collapse or explosion. This promoted rapid decompression and fragmentation of viscous andesitic magma stalled in the conduit. At the Fanthams Peak vent, eruption onsets were simple rapid vent opening phases where basaltic and basaltic andesite magma was promptly decompressed, inducing fast microvesiculation under open-system conditions. The largest volumes of rheologically homogeneous magma generated the steadiest Plinian phases at both vents. At Fanthams Peak, rapid degassing of magmas caused cessation of steady explosive phases and changes into effusive conditions. At the summit vent, transient blockage of the system by domes or conduit plugs generated conditions where heterogeneous degassing and crystallization occurred, generating highly variable magma viscosities. This, in-turn, generated highly complex eruptive episodes with unsteady columns and several phases of PDC generation.

Bull Volcanol (2017) 79:76

This work shows that the numerical modeling of future eruptive and hazard scenarios should not over-simplify the complexity of geological processes involved during large eruptions at andesitic volcanoes. It is clear that even subtle variations in magma composition, vent position, and conduit condition may strongly influence the nature and types of hazards during eruptive episodes.

Acknowledgements We are in debt to A.S. Palmer for his support in the field and to K. Arentsen for her assistance with logistics. Constructive comments from C. Bonadonna, A. Harris, and two anonymous reviewers improved this article. We thank the Department of Conservation in Taranaki for their assistance with permits. SJC is supported by the BQuantifying exposure to specific and multiple volcanic hazards^ programme of the NZ Natural Hazards Research Platform. This work forms part of the first author’s doctoral thesis. RTO is supported by a Massey University Doctoral Scholarship, a CONACyT (Mexico) Doctoral Scholarship, and the George Mason Trust of Taranaki.

References Alidibirov M, Dingwell DB (1996) Magma fragmentation by rapid decompression. Nature 380:146–149 Alloway B, Neall VE, Vucetich CG (1995) Late Quaternary (post 28,000 years B.P.) tephrostratigraphy of northeast and central Taranaki, New Zealand. J Royal Soc NZ 25:385–458 Andronico D, Cioni R (2002) Contrasting styles of Mount Vesuvius activity in the period between the Avellino and Pompeii Plinian eruptions, and some implications for assessment of future hazards. Bull Volcanol 64(6):372–391 Arce JL, Cervantes KE, Macías JL, Mora JC (2005) The 12.1 ka Middle Toluca Pumice: A dacitic Plinian–subplinian eruption of Nevado de Toluca in Central Mexico. J Volcanol Geotherm Res 147:125–143 Arce JL, Macías JL, Gardner JE, Rangel E (2012) Reconstruction of the Sibinal Pumice, an andesitic Plinian eruption at Tacaná Volcanic Complex, Mexico–Guatemala. J Volcanol Geotherm Res 217:39–55 Avellan DR, Macías JL, Sosa-Ceballos G, Velasquez G (2014) Stratigraphy, chemistry, and eruptive dynamics of the 12.4 ka plinian eruption of Apoyeque volcano, Managua, Nicaragua. Bull Volcanol 76:792 Bonadonna C, Costa A (2012) Estimating the volume of tephra deposits: a new simple strategy. Geology G32769–1 Bonadonna C, Costa A (2013) Plume height, volume, and classification of explosive volcanic eruptions based on the Weibull function. Bull Volcanol 75(8):1–19 Bonadonna C, Houghton BF (2005) Total grain-size distribution and volume of tephra-fall deposits. Bull Volcanol 67:441–456 Bonadonna C, Cioni R, Costa A et al (2016) MeMoVolc report on classification and dynamics of volcanic explosive eruptions. Bull Volcanol 78:84 Bourdier JL, Pratomo I, Thouret JC et al (1997) Observations, stratigraphy and eruptive processes of the 1990 eruption of Kelut volcano, Indonesia. J Volcanol Geoth Res 79:181–203 Capra L, Carreras LM, Arce JL, Macías JL (2006) The Lower Toluca Pumice: A ca. 21,700 yr BP Plinian eruption of Nevado de Toluca volcano, México. Geol Soc of America Special Papers 402:155–173 Carazzo G, Tait S, Kaminski E, Gardner JE (2012) The recent Plinian explosive activity of Mt. Pelée volcano (Lesser Antilles): The P1 AD 1300 eruption. Bull Volcanol 74:2187–2203

Page 25 of 27 76 Carey S, Bursik M (2000) Volcanic plumes. In: Sigurdsson H, Houghton BF, McNutt SR, Rymer H, Stix J (eds) Encyclopedia of volcanoes. Academic Press, San Diego, pp 527–544 Carey S, Sirgudsson H (1989) The intensity of Plinian eruptions. Bull Volcanol 51:28–40 Carey S, Sparks RSJ (1986) Quantitative models of the fall and dispersal of tephra from volcanic eruption columns. Bull Volcanol 48:109– 125 Cas R, Porritt L, Pittari A, Hayman P (2008) A new approach to kimberlite facies terminology using a revised general approach to the nomenclature of all volcanic rocks and deposits: descriptive to genetic. J Volcanol Geotherm Res 174:226–240 Cashman KV, Sturtevant B, Papale P, Navon O (2000) Magmatic fragmentation. In: Sigurdsson H, Houghton BF, McNutt SR, Rymer H, Stix J (eds) Encyclopedia of volcanoes. Academic Press, San Diego, pp 421–430 Cioni R, Marianelli P, Stantacroce R, Sbrana A (2000) Plinian and subplinian eruptions. In: Sigurdsson H, Houghton BF, McNutt SR, Rymer H, Stix J (eds) Encyclopedia of Volcanoes. Academic Press, San Diego, pp 477–494 Cioni R, Sulpizio R, Garruccio N (2003) Variability of the eruption dynamics during a Subplinian event: Greenish Pumice eruption of Somma-Vesuvius (Italy). J Volcanol Geotherm Res 124:89–114 Cioni R, Bertagnini A, Santacroce R, Andronico D (2008) Explosive activity and eruption scenarios at Somma-Vesuvius (Italy): Towards a new classification scheme. J Volcanol Geotherm Res 178:331–346 Cole JW, Stephenson TM (1972) Calculation of the volume of a tephra deposit. In: Cole JW (Ed) Distribution of high alumina basalts in the Taupo Volcanic Zone. Geology Department. Victoria University of Wellington 1:13–15 Coltelli M, Del Carlo P, Vezzoli L (1998) Discovery of a Plinian basaltic eruption of Roman age at Etna volcano, Italy. Geol 26(12):1095– 1098 Cronin SJ, Lube G, Dayudi DS et al (2013) Insights into the OctoberNovember 2010 Gunung Merapi eruption (Central Java, Indonesia) from the stratigraphy, volume and characteristics of its pyroclastic deposits. J Volcanol Geoth Res 261:244–259 Daggitt ML, Mather TA, Pyle DM, Page S (2014) AshCalc–a new tool for the comparison of the exponential, power-law and Weibull models of tephra deposition. J Appl Volcanol 3(1):7 Damaschke M, Cronin SJ, Holt K, Bebbington M, Hogg A (2017a) A 30, 000-year high-precision eruption history for the andesitic Mt Taranaki, North Island, New Zealand. Quaternary Res, 1–23. https://doi.org/10.1017/qua.2016.11 Damaschke M, Cronin SJ, Torres-Orozco R, Wallace RC (2017b) Unifying tephrostratigraphic approaches to redefine major Holocene marker tephras, Mt. Taranaki, New Zealand. J Volcanol Geoth Res 337:29–43. https://doi.org/10.1016/j.jvolgeores.2017.02. 021 Druce AP (1966) Tree-ring dating of recent volcanic ash and lapilli, Mt Egmont. NZ J Botany 4:3–41 Eychenne J, Le Pennec J-L, Ramón P, Yepes H (2013) Dynamics of explosive paroxysms at open-vent andesitic systems: Highresolution mass distribution analyses of the 2006 Tungurahua fall deposit (Ecuador). Earth Planet Sci Lett 361:343–355 Fierstein J, Nathenson M (1992) Another look at the calculation of fallout tephra volumes. Bull Volcanol 54:156–167 Fisher R, Schmincke HU (1984) Pyroclastic Rocks. Springer-Verlag, Berlin 472 pp Folk RL, Ward WC (1957) Brazos River bar: a study in the significance of grain size parameters. J Sedimentary Res 27(1) Franks AM (1984) Soils of Eltham County and the tephrochronology of central Taranaki. Dissertation, Massey University, Palmerston North, New Zealand

76

Page 26 of 27

Gardner JE, Thomas RME, Jaupart C, Tait S (1996) Fragmentation of magma during Plinian volcanic eruptions. Bull Volcanol 58(2–3): 144–162 Harvey PK, Taylor DM, Hendry RD, Bancroft F (1973) An accurate fusion method for the analysis of rocks and chemically related materials by X-ray fluorescence spectrometry. X-Ray Spectrom 2:33– 34 Henrys S, Reyners M, Bibby H (2003) Exploring the plate boundary structure of North Island, New Zealand. EOS, Trans Am. Geophys Union 84:289–294 Houghton BF, Wilson CJN (1989) Vesicularity index for pyroclastic deposits. Bull Volcanol 51:451–462 Houghton BF, Wilson CJN, Del Carlo P et al (2004) The influence of conduit processes on changes in style of basaltic Plinian eruptions: Tarawera 1886 and Etna 122 BC. J Volcanol Geoth Res 137:1–14 Inman DL (1952) Measures for describing the size distribution of sediments. J Sedimentary Res 22(3) Irvine TNJ, Baragar WRAF (1971) A guide to the chemical classification of the common volcanic rocks. Canadian J of Earth Sciences 8(5): 523–548 King PR, Thrasher GP (1996) Cretaceous-Cenozoic geology and petroleum systems of the Taranaki Basin, New Zealand. Institute of Geological and Nuclear Sciences Monograph 13 Klawonn M, Houghton BF, Swanson DA et al (2014) From field data to volumes: constraining uncertainties in pyroclastic eruption parameters. Bull Volcanol 76(7):1–16 Klug C, Cashman KV (1994) Vesiculation of May 18, 1980. Mount St. Helens magma. Geology 22:468–472 Klug C, Cashman KV (1996) Permeability development in vesiculating magmas: implications for fragmentation. Bull Volcanol 58:87–100 Klug C, Cashman KV, Bacon CR (2002) Structure and physical characteristics of pumice from the climactic eruption of Mt Mazama (Crater Lake), Oregon. Bull Volcanol 64:486–501 Le Bas MJ, Le Maitre RW, Streckeisen A, Zanettin B (1986) A chemical classification of volcanic rocks based on the total alkali-silica diagram. J Petrology 27(3):745–750 Legros F (2000) Minimum volume of a tephra fallout deposit estimated from a single isopach. J Volcanol Geotherm Res 96:25–32 Macias JL, Sheridan MF, Espindola JM (1997) Reappraisal of the 1982 eruptions of El Chichón Volcano, Chiapas, Mexico: new data from proximal deposits. Bull Volcanol 58:459–471 Mader HM (1998) Conduit flow and fragmentation. In: JS Gilbert and RSJ Sparks (Eds) The physics of explosive volcanic eruptions. Geol Soc Lon Spec Publ 145:51–71 Manville V, Németh K, Kano K (2009) From source to sink: a review of three decades of progress in the understanding of volcaniclastic processes, deposits, and hazards. Sediment Geol 220:136–161 May DJ (2003) The correlation of recent tephra with lava flows on Egmont volcano, Taranaki, New Zealand using evidence of mineral chemistry. Dissertation, University of Auckland, Auckland, New Zealand Neall VE (1972) Tephrochronology and tephrostratigraphy of western Taranaki (N108-109), New Zealand. NZ J Geol Geoph 15:507–557 Neall VE (1979) Sheets P19, P20, and P21. New Plymouth, Egmont and Manaia. 1st ed. Geological map of New Zealand 1:50,000. 3 maps and notes. NZ Dept of Scientific and Industrial Res, Wellington Neall VE (2003) The volcanic history of Taranaki. Institute of Natural Resources, Massey University, Soil & Earth Sciences Occasional Publication 2 Neall VE, Stewart RB, Smith IEM (1986) History and petrology of the Taranaki volcanoes. In: Smith IEM (ed) Late Cenozoic volcanism. Royal Society of New Zealand Bulletin 23:251–263 Pardo N, Cronin SJ, Palmer A et al (2012) Andesitic Plinian eruptions at Mt. Ruapehu: quantifying the uppermost limits of eruptive parameters. Bull Volcanol 74:1161–1185

Bull Volcanol (2017) 79:76 Pardo N, Cronin SJ, Wright HMN et al (2014) Pyroclastic textural variation as an indicator of eruption column steadiness in andesitic Plinian eruptions at Mt. Ruapehu. Bull Volcanol 76:822 Platz T, Cronin SJ, Cashman KV et al (2007) Transitions from effusive to explosive phases in andesite eruptions-A case-study from the AD 1655 eruption of Mt. Taranaki, New Zealand. J Volcanol Geoth Res 161:15–34 Platz T, Cronin SJ, Procter JN et al (2012) Non-explosive dome-forming eruptions at Mt. Taranaki, New Zealand. Geomorphology 136:15– 30 Procter JN, Cronin SJ, Platz T et al (2010) Mapping block-and-ash flow hazards based on Titan 2D simulations: a case study from Mt. Taranaki, NZ. Nat Hazards 53:483–501 Pyle DM (1989) The thickness, volume and grainsize of tephra fall deposits. Bull Volcanol 51:1–15 Pyle DM (1995) Assessment of the minimum volume of tephra fall deposits. J Volcanol Geotherm Res 69(3):379–382 Pyle DM (2000) Sizes of volcanic eruptions. In: Sigurdsson H, Houghton BF, McNutt SR, Rymer H, Stix J (eds) Encyclopedia of volcanoes. Academic Press, San Diego, pp 263–269 Sable JE, Houghton BF, Wilson CJN, Carey RJ (2006) Complex proximal sedimentation from Plinian plumes: the example of Tarawera 1886. Bull Volcanol 69:89–103 Saucedo R, Macías JL, Gavilanes JC, Arce JL et al (2010) Eyewitness, stratigraphy, chemistry, and eruptive dynamics of the 1913 Plinian eruption of Volcán de Colima, México. J Volcanol Geoth Res 191: 149–166 Shea T, Gurioli L, Houghton BF et al (2011) Column collapse and generation of pyroclastic density currents during the AD 79 eruption of Vesuvius: the role of pyroclast density. Geology 39(7):695–698 Shea T, Gurioli L, Houghton BF (2012) Transitions between fall phases and pyroclastic density currents during the AD 79 eruption at Vesuvius: building a transient conduit model from the textural and volatile record. Bull Volcanol 74(10):2363–2381 Sherburn S, White RS (2006) Tectonics of the Taranaki region, New Zealand: earthquake focal mechanisms and stress axes. NZ J Geol and Geoph 49:269–279 Sparks RSJ (1986) The dimension and dynamics of volcanic eruption columns. Bull Volcanol 48:13–15 Stagpoole V, Nicol A (2008) Regional structure and kinematic history of a large subduction back thrust: Taranaki Fault, New Zealand. J Geoph Res: Solid Earth 113(B1) Statistics New Zealand (2013) 2013 Census: Taranaki Region. http:// www.stats.govt.nz/. Accessed 1 Jan 2017 Stewart RB, Price RC, Smith IEM (1996) Evolution of high-K arc magma, Egmont volcano, Taranaki, New Zealand: evidence from mineral chemistry. J Volcanol Geoth Res 74:275–295 Sulpizio R (2005) Three empirical methods for the calculation of distal volume of tephra-fall deposits. J Volcanol Geotherm Res 145(3–4): 315–333 Sulpizio R, Dellino P, Doronzo DM, Sarocchi D (2014) Pyroclastic density currents: state of the art and perspectives. J Volcanol Geotherm Res 283:36–65 Topping WW (1971) Burrell Lapilli eruptives, Mount Egmont, New Zealand. NZ J Geol Geoph 15:476–490 Torres-Orozco R, Cronin SJ, Pardo N, Palmer AS (2017) New insights into Holocene eruption episodes from proximal deposit sequences at Mt. Taranaki (Egmont). New Zealand Bull Volcanol 79(3):1–25. https://doi.org/10.1007/s00445-016-1085-5 Turner MB (2008) Eruption cycles and magmatic processes at a reawakening volcano, Taranaki, New Zealand. Dissertation, Massey University, Palmerston North, New Zealand Turner MB, Cronin SJ, Bebbington MS, Platz T (2008a) Developing probabilistic eruption forecasts for dormant volcanoes: a case study from Mt Taranaki, New Zealand. Bull Volcanol 70:507–515

Bull Volcanol (2017) 79:76 Turner MB, Cronin SJ, Smith IEM et al (2008b) Eruption episodes and magma recharge events in andesitic systems, Mt Taranaki, New Zealand. J Volcanol Geoth Res 177:1063–1076 Turner MB, Cronin SJ, Stewart RB, Bebbington M, Smith IEM (2008c) Using titanomagnetite textures to elucidate volcanic eruption histories. Geology 36:31–34 Turner MB, Bebbington MS, Cronin SJ, Stewart RB (2009) Merging eruption datasets: building an integrated Holocene eruptive record for Mt. Taranaki, New Zealand. Bull Volcanol 71:903–918 Turner MB, Cronin SJ, Bebbington MS et al (2011a) Integrating records of explosive and effusive activity from proximal and distal sequences: Mt. Taranaki, New Zealand. Quaternary Intl 246:364–373 Turner MB, Cronin SJ, Bebbington MS et al (2011b) Relating magma composition to eruption variability at andesitic volcanoes: A case study from Mount Taranaki, New Zealand. GSA Bull 123(9/10): 2005–2015 Villemant B, Boudon G, Komorowski JC (1996) U-series disequilibrium in arc magmas induced by water-magma interaction. Earth Planet Sci Lett 140:259–267 Walker GPL (1971) Grain-size characteristics of pyroclastic deposits. J Geol 79:696–714 Walker GPL (1973) Explosive Volcanic Eruptions -A new Classification Scheme. Geologische Rundshau 62:431–446

Page 27 of 27 76 White JDL, Houghton BF (2006) Primary volcaniclastic rocks. Geology 34:677–680 Whitehead SJ (1976) Granulometric studies on selected tephra eruptions, North Island, New Zealand. Dissertation, Massey University, Palmerston North, New Zealand Wilson L (1976) Explosive volcanic eruptions III. Plinian eruption columns. Geophys J Res Astron Soc 45:543–556 Wilson SA (1997) The collection, preparation and testing of USGS reference material BCR-2, Columbia River Basalt. US Geological Survey, Open-File Report 98-00x Wilson L, Walker GPL (1987) Explosive volcanic eruptions VI. Ejecta dispersal in plinian eruptions: the control of eruption conditions and atmospheric properties. Geophys J Res Astron Soc 89:657–679 Yang Q, Bursik M (2016) A new interpolation method to model thickness, isopachs, extent, and volume of tephra fall deposits. Bull Volcanol 78(10):68 Zernack AV, Cronin SJ, Neall VE, Procter JN (2011) A medial to distal volcaniclastic record of an andesite stratovolcano: Detailed stratigraphy of the ring-plain succession of south-west Taranaki, New Zealand. Intl J Earth Scs 100:1936–1966 Zernack AV, Price RC, Smith IEM, Cronin SJ, Stewart RB (2012) Temporal Evolution of a High-K Andesitic Magmatic System: Taranaki Volcano, New Zealand. J Petrol 53(2):325–363