Bull Volcanol (2016) 78:24 DOI 10.1007/s00445-016-1018-3

RESEARCH ARTICLE

Early volcanic history of the Rabaul area Chris O. McKee 1 & Robert A. Duncan 2

Received: 10 October 2015 / Accepted: 21 February 2016 # Springer-Verlag Berlin Heidelberg 2016

Abstract We conducted an extensive program of 40Ar–39Ar age determinations on a suite of 27 volcanic rock samples from key stratigraphic units at Rabaul, Papua New Guinea in order to improve understanding of the early eruption history of the multiple volcanic systems present in the area. Analyses of whole rock, plagioclase and groundmass separates yielded statistically significant ages for 24 samples. Replicate analyses (groundmass, plagioclase) for 17 of the samples provided concordant ages. The oldest systems in the Rabaul area (>1 Ma to ≈300 ka) are in the south, associated with the caldera-like Varzin Depression, and in the north, at the stratovolcanoes Watom and Tovanumbatir. The earliest known activity of the Rabaul system occurred between about 330 and 200 ka and involved emplacement of lava flows and scoria deposits. Major explosive activity at the Rabaul system commenced at about 200 ka and produced a sequence of dacitic ignimbrites that culminated with the emplacement of the large-volume Malaguna Pyroclastics at about 160 ka. Calderas may have been formed as a consequence of the large volumes of tephra produced during some of these eruptions. Products of the early activity are found in the northern and northeastern walls of Rabaul Caldera and on the northeastern flank of Tovanumbatir. This leads to the conclusion that the source of the early activity at Rabaul

Editorial responsibility: J. Fierstein * Chris O. McKee [email protected]

1

Port Moresby Geophysical Observatory, Port Moresby, National Capital District, Papua New Guinea

2

College of Earth, Ocean and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331, USA

probably was located in the northern part of the present caldera complex. A shift in the focus of activity at the Rabaul system took place between about 160 and 125 ka. All of the younger (<125 ka) major pyroclastic formations, including the Karavia Welded Tuff, the Barge Tunnel Ignimbrite and the Latlat Pyroclastics, which make up the bulk of the exposure in the southern and western walls of Rabaul Caldera, were erupted from a source or sources in the south-central part of the complex. The stratovolcanoes Palangiangia and Kabiu, which flank the northeastern part of the complex, had commenced activity by about 100 ka while their neighbour to the southeast, Turagunan, was active by about 70 ka. There is little stratigraphic and chronological information about the Tavui system, immediately north of Rabaul. At present, the products of only two eruptions in the Rabaul area can be attributed to Tavui: the ≈79 ka Tokudukudu Ignimbrite and the ≈7 ka Raluan Ignimbrite. However, this evidence of activity and its timing is sufficient to show that the Tavui and Rabaul systems are coeval. The average interval between major eruptions at the Rabaul system over the last ≈18 ky is in the range 2.6 to 6 ky. Eruption frequency likely varied over the life of the system; however, the return period during the early phase of ignimbrite eruptions appears to have been similar to that of the last ≈18 ky. Keywords Rabaul Caldera . Papua New Guinea . 40Ar-39Ar ages

Introduction The volcanoes of the Rabaul area constitute the northern end of the Gazelle Volcanic Zone (GVZ, Johnson et al. 2010), the belt of volcanoes that extends for 65 km across the central to northern part of the Gazelle Peninsula, New Britain Island, Papua New Guinea(Fig.1).TheoldestmembersoftheGVZ areatitssouthern

24

Bull Volcanol (2016) 78:24

Page 2 of 28

W-T Zone

TAVUI

Watom

N N 4º10'

Palangiangia

Tovanumbatir

Kabiu Turagunan

RABAUL Vulcan

Tavurvur

VUNAKANAU 4º20'

Ke a rev . tR

Fig. 1 Northeastern Gazelle Peninsula showing the belt of Late Cainozoic volcanoes, the Gazelle Volcanic Zone (GVZ), that includes the caldera centres of the Rabaul area. Bold arcuate lines denote interpreted caldera margins. Parallel northwesttrending lines mark the WatomTuragunan Zone (WTZ) of dominantly mafic stratovolcanoes. Solid triangles represent volcanic centres in the southeastern part of the WTZ. Solid stars represent the recently active intra-caldera centres of the Rabaul system. Solid circles represent the locations of the stratigraphic columns shown in Fig. 2. Crossed geological picks represent gold mining operations

VARZIN Mt Varzin Wairiki Wara

KEREVAT

ngoi R .

4º30'

SIKUT

NENGMUTKA 0 4º40'

10

5

15

km 152º00'

end. Geological mapping indicates that Nengmutka was active in the late Tertiary, Miocene to Pliocene (Lindley 1988, 2006), and Sikut has been K–Ar dated at 2.90 ± 0.04 Ma, i.e. Late Pliocene (Corbett et al. 1991). The caldera systems of the Rabaul area, Rabaul and Tavui, and the stratovolcanoes on the northeastern to easternflankofRabaulCalderaaretheonlyactiveorpotentially active centres in the GVZ.The mostsignificantrecentactivity was at Rabaul in 1937 when more than 500 people were killed by the VEI 4 Vulcan eruption (Fisher 1939; Johnson and Threlfall 1985) andin 1994whenabout70 %of thetownofRabaulwasdestroyed by the simultaneous VEI 4 Tavurvur and Vulcan eruptions (Blong and McKee 1995; Johnson et al. 1996). Geological studies at Rabaul since the late 1960s (Heming 1974; Walker et al. 1981; Nairn et al. 1989, 1995) have demonstrated that the area has a long and complex eruption history (Table 1), involving multiple eruption sources. In addition to the evidence of numerous eruptions from intra-caldera sources at Rabaul, there are indications of many other eruption sources in the Rabaul area, as mapped by Nairn et al. (1989, 1995) and as

152º10'

152º20'

shown in Fig. 1. The extra-caldera sources closest to Rabaul are the large stratovolcanoes Watom, Tovanumbatir, Palangiangia, Kabiu and Turagunan, which comprise the Watom-Turagunan Zone (WTZ; Johnson et al. 2010), a broad volcanic corridor trending at 310° across the northern to northeastern flank of the Rabaul Caldera Complex. Neighbouring eruption sources to the south of Rabaul Caldera include the broad caldera-like feature named Varzin Depression and its associated stratocones Mount Varzin and Wairiki, and the elevated shield between Varzin Depression and Rabaul Caldera that is surmounted by a relatively small, shallow caldera-like depression named Vunakanau Basin (Nairn et al. 1989). The most recently recognized eruption source in the Rabaul area is the largely submarine Tavui caldera system, immediately north of Rabaul (Tiffin et al. 1986; Taylor et al. 1991; McKee 2015). While the geological studies at Rabaul have resulted in the development of a reasonably detailed stratigraphic sequence (Table 1), a detailed chronology of the sequence has been lacking. Understanding the eruptive history of the Rabaul area is critical in

Bull Volcanol (2016) 78:24 Table 1 Rabaul eruption history prior to this study (from Nairn et al. 1995, modified and updated)

Page 3 of 28 24

Age

Event

Magma type

Volume (km3)

Tavurvur pyroclastic and lava eruptions

A/D

0.3?

AD

1994–2011 1994

Vulcan pyroclastic and cone-building eruptions

D/A

0.3?

1937–1943

Tavurvur pyroclastic eruptions

A/D

0.05?

1937

Vulcan pyroclastic and cone-building eruptions

D/A

0.3

1878

Vulcan (submarine) and Tavurvur pyroclastic eruptions

D/A

0.3?

1850??

Sulphur Creek pyroclastic eruptions, crater forming

A

0.1?

1791

Tavurvur pyroclastic and lava eruptions, cone-building

A/D

0.1?

1767

Rabalanakaia? pyroclastic and lava eruptions, cone-building (<250 year B.P. by 14C) 14 Years B.P. ( C) <750

Rabalanakaia pyroclastic and lava eruptions, cone-building

A

0.2?

>750

Dawapia pyroclastic eruptions, cone-building, pumice rafts

D

0.3?

?

Sulphur Creek lavas and pyroclastics, cone-building

A

0.1?

ca. 1400

Rabaul Pyroclastics eruption and caldera collapse

D

>10

4200–1400

Talili Subgroup, multiple (>7) pyroclastic eruptions producing fine vitric ashes and pumice deposits, roughly coeval with late cone-building Palangiangia Lavas and scorias,

D

5

B

0.3?

and the youngest Kabiu scorias

B

0.2? 0.1?

?

and Turagunan scorias

B

6900

Raluan Ignimbrite and caldera collapse (Tavui)

R

>5

6900

Raluan Scoria

B

0.5?

?–?

Talwat Subgroup, multiple (>4) pyroclastic eruptions producing fine ashes, roughly coeval with older cone-building Turagunan scorias

D?

1?

B

0.5?

?

and Kabiu cone-building scorias/lavas

B

2?

and Palangiangia scorias/lavas early cone-building

B

1?

ca. 15,000

Vunabugbug Pyroclastics eruption and caldera collapse?

D

5?

?

Tatoko Subgroup multiple (>4?), pyroclastic eruptions

D

1?

c.16,000

Namale Pyroclastics eruption and caldera collapse?

D

5?


Four unnamed and uncorrelated pyroclastic eruptions as seen at Kerevat Kulau Ignimbrite, major ignimbrite eruption, caldera collapse?

D

?

D

>10?

B?,R,D

1?

?

Kabakada Subgroup, multiple (>6) pyroclastic eruptions with a least 3 (small?) ignimbrites Vunairoto Lapilli, pyroclastic fall

A/D

0.1?

37,000

Latlat Pyroclastics, caldera collapse

D

5

?–?

Talakua Subgroup, multiple (>7?) pyroclastic eruptions

D

1?

?20 ka

Ma 0.04?

Barge Tunnel Ignimbrite, caldera collapse?

D

5?

?

Karavia Welded Scorias

D

1?

0.08

Tokudukudu Ignimbrite (Tavui)

R

?

?–?

Tavui Subgroup, multiple (>5) pyroclastic eruptions with a least one ignimbrite Malaguna Pyroclastics eruption, and caldera collapse?

D/B

1?

D/A

10?

0.1 0.1

Boroi Ignimbrites, and caldera collapse?

D

10?

?

East Wall Lavas

D

?

0.19

Rabaul Quarry Lavas

D

?

?

Seismograph Lavas

B

?

0.5

Tovanumbatir Lavas and scoria

B

2?

Most volumes are rough estimates A andesite, B basalt, D dacite, R rhyolite

24

Bull Volcanol (2016) 78:24

Page 4 of 28

Fig. 2 Stratigraphic columns showing exposure in the north-„ northwestern wall of Rabaul Caldera and in the southwestern wall of Tavui Caldera. The only common units are the Malaguna Pyroclastics (160 ka), the Raluan Ignimbrite (6.9 ka) and the Rabaul Pyroclastics (1.4 ka). Significant gaps in these columns are due to inaccessibility and lack of exposure

developing a capacity to anticipate and plan for future volcanic events. Radiocarbon dating has provided a measure of chronological control on major eruptions in the Rabaul area in the Late Pleistocene and Holocene. Within the age range of this dating technique (approximately the last 50 ky), ages have been determined and refined for a number of the major eruptions and for one intermediate-scale eruption, as shown in Table 2. While these results may represent a relatively short portion of the complete eruption history of the Rabaul area, they do give an indication of the frequency of the larger eruptions in the Late Pleistocene to Holocene. The paucity of good chronological data for volcanoes of the Rabaul area and indications of recharge of the Rabaul system, e.g. resurgence of the floor of the caldera (McKee 1993) since the latest major eruption, at 1.4 ka BP (Heming 1974; Nairn et al. 1989, 1995), have given some urgency to quantifying the eruptive history. Accordingly, we designed and implemented a program using the 40Ar–39Ar incremental heating dating technique (McDougall and Harrison 1988) that was expected to provide a more complete record of the timing of the larger, devastating eruptions from the Rabaul and Tavui systems and to determine the chronological relationships between these systems and the neighbouring centres. This paper presents the results of the 40Ar–39Ar experiments. These results provide significant new information on the early volcanic history of the Rabaul area, particularly eruption frequency, and thus are of value in assessing the volcanic hazards and risk in the northeastern part of the Gazelle Peninsula in general, and in the Rabaul area in particular.

&

Geologic context

& &

Stratigraphic framework and sample collection

&

removal of large parts of the deposited pyroclastic sequence which is expressed as erosional unconformities and the similarities in appearance and in chemical composition of many of the pyroclastic deposits (Nairn et al. 1989, 1995). Wind patterns at Rabaul (McKee et al. 1985) have resulted in significant differences in the distribution of pyroclastic fall deposits in different sectors around eruption sources: pyroclastics deposited in one sector of the edifice may be absent or unrecognizably different in thickness and texture in another sector. Difficult-to-access and discontinuous exposure is available in the walls of the Rabaul and Tavui calderas. Sections for these exposures (Fig. 2) illustrate the challenges of stratigraphic correlation in the Rabaul area: Apart from the youngest deposits, the 1.4 ka Rabaul Pyroclastics and the 6.9 ka Raluan Ignimbrite, the only other formation identified in both exposures is the 160 ka Malaguna Pyroclastics. Thus, the stratigraphy of the Rabaul area has been pieced together from numerous isolated exposures of mostly small portions of the stratigraphic record. A complex stratigraphy has resulted in which there are considerable uncertainties and likely gaps. The eruptive history of the Rabaul area is a record of the interplay of activity from several volcanic sources (Fig. 1): Early centres and systems adjacent to Rabaul Caldera Complex, including early elements of the WTZ Rabaul Caldera Complex Younger stratovolcanoes adjacent to Rabaul Caldera Complex—later elements of the WTZ Tavui Volcano

There are significant difficulties correlating stratigraphy in the Rabaul area. This is a result of poor exposure, erosional Table 2 Revised radiocarbon chronology of major eruptions in the Rabaul area Volcano

Eruption

Date (ka BP)

Volume (km3)

Rabaul Rabaul Tavui

Rabaul Pyroclastics Talili Pyroclastics Raluan Ignimbrite

1.4 4.2 6.9

11 2? 5?

Rabaul Rabaul Rabaul Rabaul

Vunabugbug Pyroclastics Namale Pyroclastics Kulau Ignimbrite Latlat Pyroclastics

10.5 15.1 ≈18 36.7

5? 5? >10? 5?

Information sources: Nairn et al. (1995), Rabaul Volcano Observatory unpublished data.

Early centres and systems adjacent to Rabaul Caldera Complex The 15 × 10 km caldera-like Varzin Depression (Nairn et al. 1989), whose centre lies about 20 km south-southwest of the centre of Rabaul Caldera, appears to be the oldest volcanic system in the Rabaul area. Varzin Depression is mostly buried by pyroclastic deposits assumed to have originated from younger centres to the north (including the Rabaul Caldera Complex), but two apparently late-stage yet eroded features are exposed at the northeastern and eastern edges of the depression. These are the small stratocones Mount Varzin and Wairiki respectively. The analysed products of Mount Varzin cone are basalts having about 51 wt% SiO2 (Nairn et al. 1989).

Bull Volcanol (2016) 78:24

Page 5 of 28 24

RABAUL NNW CALDERA WALL

Rabaul Pyroclastics

1m

TAVUI SW CALDERA WALL

1m Rabaul Pyroclastics

Raluan Ignimbrite Vunabugbug Pyroclastics

Raluan Ignimbrite

Namale Pyroclastics Grey ignimbrite

Brown ignimbrite Brown ignimbrite

Brown ignimbrite Brown ignimbrite

Grey ignimbrite 1m

unconformity Coarse lithic-rich ignimbrite Brown-grey ignimbrite Vunairoto Lapilli Plinian grey pumice Tokudukudu Ignimbrite

Grey ignimbrite Stratigraphy obscured

unconformity 1m

Tavui Scoria - grey-black scoria

Stratigraphy obscured - black welded tuff

Malaguna Pyroclastics - pumice lapilli/coarse ash - black welded scoria - sintered scoria

Malaguna Pyroclastics - pumice lapilli/coarse ash

Stratigraphy obscured

Grey ignimbrite Grey ignimbrite Grey ignimbrite Grey ignimbrite Grey ignimbrite Cliff Ignimbrite Brennan Ignimbrite

Boroi Ignimbrite

24

Page 6 of 28

Immediately north of Varzin Depression is the Vunakanau centre. Vunakanau Basin, a 3 × 4 km shallow depression at the summit of this centre, may represent a buried caldera (Nairn et al. 1989). The asymmetrical flanks of the Vunakanau centre slope gently and may extend as much as 10 km to the northwest and eastsoutheast. There are no exposures of the material that built this structure and only its general form is suggested through the blanket of the overlying pyroclastic sequence from the Rabaul system. Watom Island, which lies about 20 km northwest of the centre of Rabaul Caldera, is the subaerial part of a large stratovolcano having basal diameter of about 10 km. Watom has a broad summit crater about 2 km across and is deeply dissected. A collar of raised coral reef fringes the subaerial part of the volcano. Analysed products of Watom are basaltic andesites having about 56 wt% SiO2 (Nairn et al. 1989). Tovanumbatir appears to have been similar in size to Watom originally, but its lower southern and northeastern flanks were engulfed by caldera formation at the adjacent Rabaul and Tavui centres, respectively. Deep dissection of Tovanumbatir reveals that this centre is a stratovolcano. Much of what remains of the lower flanks of Tovanumbatir is draped by younger pyroclastic deposits, mostly from Rabaul but some also from Tavui. Analysed rock samples from Tovanumbatir range from basalt to high-silica dacite, having 51–68 wt% SiO2 (Nairn et al. 1989; P. Wallace, pers. comm., 2002; Rabaul Volcano Observatory (RVO), unpublished data). The Watom and Tovanumbatir centres are the oldest known elements of the WTZ, the essentially mafic volcanic corridor that lies adjacent to the Rabaul Caldera Complex. Rabaul Caldera Complex The largely sea-filled elliptical caldera at Rabaul measures 14 × 9 km. The margins of at least 3 calderas can be recognized in the nested caldera complex (Fig. 1). The caldera complex probably evolved from an earlier system that developed as a southeastwards migration of activity from the Watom and Tovanumbatir centres. From stratigraphic relationships, the oldest units known from the Rabaul system are lava flows exposed at the base of the northeastern and north-northeastern walls of the caldera. The basaltic Seismograph Lavas and the dacitic East Wall Lavas are present in a section of the northeastern caldera wall adjacent to Tavurvur, and lava flows or domes of dacitic composition, the Rabaul Quarry Lavas, are present in the northnortheastern caldera wall at two sites. Rabaul Quarry Lavas and East Wall Lavas are chemically and petrographically almost identical (Nairn et al. 1989), and similar radial and arcuate joint patterns are seen in their

Bull Volcanol (2016) 78:24

caldera wall exposures. These early lavas occupy sites within the WTZ and possibly represent activity from a WTZ precursor to the later-developed Rabaul Caldera Complex. Exposures in the caldera walls indicate an overwhelming predominance of pyroclastic rocks. Stacked dacitic pyroclastic deposits in the north-northwestern caldera wall (Fig. 2) include Brennan Ignimbrite and Cliff Ignimbrite at base, Boroi Ignimbrite and Malaguna Pyroclastics near the middle and Raluan Ignimbrite and Rabaul Pyroclastics at the top of the sequence (Fig. 2, and see additional details in ‘Early sequence of frequency ignimbrite eruptions’). The basal to middle part of the exposure in the northern wall of the caldera constitutes a part of the Older Formations group (see below). A younger sequence of pyroclastic deposits, which is exposed in the southern to western basal parts of the caldera wall, includes Karavia Welded Tuff, Barge Tunnel Ignimbrite and Latlat Pyroclastics: These formations are part of the Younger Formations group (see below). These formations are primarily dacitic, although some of the deposits include components of andesitic and rhyodacitic compositions. Distinctive andesitic scoriaceous units in the pyroclastic canopy on the lower northeastern flank of Tovanumbatir are the newly identified pyroclastic deposits, Tokalat Tuff and Korere Scoria, together with the previously known Tavui Scoria. These formations are part of the Older Formations group (see below). A deep roadcut at Kabakada, about 12 km northwest of the centre of Rabaul Caldera, exposes a 21-m-thick pyroclastic sequence that represents a large part of the Younger Formations group. At the base is a distinctive flat-lying fall deposit, Vunairoto Lapilli, of andesitic composition. Vunairoto Lapilli is overlain by a flatlying sequence of pyroclastic deposits, the Kabakada Subgroup, which includes the deposits of more than six pyroclastic eruptions at least three of which are ignimbrites. An angular unconformity, which is believed to be associated with Late Pleistocene low sea levels around 18 ka BP (Nairn et al. 1989, 1995), separates the Kabakada Subgroup from the overlying sequence of radiocarbon-dated younger pyroclastics from the Rabaul Caldera Complex. Directly overlying the unconformity is the large-volume Kulau Ignimbrite and other formations that comprise the youngest part of the sequence, capped by the 1.4-ka Rabaul Pyroclastics deposits. The age of the Rabaul Pyroclastics eruption has been revised to AD 667–699 using wiggle-match 14C dating (McKee et al. 2015). Much of the caldera floor topography is very youthful, comprising products from activity following the latest major eruption. The most recent activity occurred at the Karavia Bay

Bull Volcanol (2016) 78:24

vents, Rabalanakaia, Sulphur Creek, Vulcan and Tavurvur (McKee, Kuduon and Patia, unpublished data, 2011).

Page 7 of 28 24

Turagunan may have been active post-1.4 ka as indicated by a lava flow and pyroclastic deposits from this centre which appear to be draped over the 1.4 ka caldera wall (Heming 1974).

Younger stratovolcanoes Tavui volcano The stratovolcanoes Palangiangia, Kabiu and Turagunan dominate the peninsula that forms the northeastern to eastern flank of Rabaul Caldera (Fig. 3) and constitute the younger part of the WTZ. These volcanoes are little dissected and are apparently younger than Watom and Tovanumbatir. Lava flows and pyroclastic deposits from Palangiangia and Kabiu make up most of the exposure in the northeastern caldera wall. Stratigraphic relationships expressed in caldera wall exposures suggest that Palangiangia predates its larger neighbour Kabiu. Turagunan, which lies immediately southeast of Palangiangia and Kabiu, is apparently younger than these two centres. All three centres are believed to be potentially active. There are indications that all three were active during the period of eruptive activity responsible for the deposition of the Talili Pyroclastics Subgroup, starting at about 4.2 ka and continuing until some time in the interval 1.8–1.4 ka (McKee, unpublished data, 2015). Part of Palangiangia was engulfed during the calderamodifying eruption at 1.4 ka BP, but later activity built up a new edifice, Rabalanakaia, within the remnant of Palangiangia. Fig. 3 The young stratovolcanoes Turagunan, Kabiu and Palangiangia dominate the peninsula that forms the eastern to northeastern flank of the Rabaul Caldera Complex. Together with the older centres Tovanumbatir and Watom, these stratovolcanoes form the broad northwest-trending W-T Zone that separates the caldera systems of Rabaul and Tavui. The postcaldera vent Tavurvur (left middle ground) was erupting at the time this photograph was taken, about 2–3 January, 1996. This photograph was supplied by Gazelle Restoration Authority, Kokopo, East New Britain Province

Tavui Volcano contains a steep-walled, mostly submarine caldera, 10 × 9 km across and 1100 m deep, that hosts two intracaldera cones, one of which has a well-preserved crater (Tufar 1990; Tiffin et al. 1990; McKee 2015). A large cone forms the northern flank of Tavui: Its summit is the highest point, 18 m below sea level, on the submerged section of Tavui’s caldera rim. An extensive field of terraced pyroclastic deposits is present on the northern and western flanks of Tavui. The southwestern wall of Tavui Caldera is seen on-shore as the Tavui Fault (Nairn et al. 1995) on the northeastern flank of Tovanumbatir (Fig. 1). Exposures in the Tavui Fault scarp include Rabaul-sourced rocks, two units believed to be from Tavui and other deposits which could be from Tavui but which currently are of unknown origin. The two units attributed to Tavui are the 6.9 ka Raluan Ignimbrite and the 78.7 ka Tokudukudu Ignimbrite (McKee 2015). The chemical characteristics of these deposits, low K2O and high SiO2, distinguish them from rocks sourced at both Tovanumbatir and Rabaul

24

Bull Volcanol (2016) 78:24

Page 8 of 28

(Nairn et al. 1995; Wood et al. 1995). Rocks dredged from Tavui’s caldera characteristically have low contents of K and other incompatible elements compared with Rabaul-sourced rocks (Tufar and Naser 1992; Wallace and Tufar 1998).

details of the samples and the geochemical context of the Rabaul area volcanoes are presented in the following section. Petrology Geochemical data sources

Sample collection In total, 27 rock samples were selected from the volcanoes of the Rabaul area for radiometric dating by 40Ar–39Ar methods. The rock samples were chosen so as to best define the eruptive history of these volcanoes and were evaluated by thin section examination for state of alteration and texture. Details of the samples and their locations are given in Table 3. Chemical

Table 3 Details of rock samples from the Rabaul area submitted for 40Ar–39Ar dating

Geochemical data from the Rabaul area volcanoes considered here are derived from Heming and Carmichael (1973), Heming (1974), Nairn et al. (1989) and unpublished data of Dr R. F. Heming. The analytical methods used were classical wet chemical for major element data reported by Heming and Carmichael (1973) and Heming (1974) and standard XRF for most other data. Four new analyses supplied by Dr S. Eggins

System

Location/Formation

Sample

Lithology

Grid Ref.a

1. Early centres adjacent to Rabaul

Varzin Watom Watom Tovanumbatir Tovanumbatir Vunairoto Lapilli Latlat Pyroclastics Barge Tunnel Ignimbrite Karavia Welded Tuff Tavui Scoria Tavui Scoria Malaguna Pyroclastics Malaguna Pyroclastics

RP98021 RP98016 RP98025 RP98003 RP98026 RP98006 RP98015 RP98027

066 143 972 448 982 454 085 362 076 377 995 351 109 239 060 263

Boroi Ignimbrite Boroi Ignimbrite Brennan Ignimbrite

RP98009 RP98029 RP98028

Cliff Ignimbrite

RP98030

Rabaul Quarry Lavas East Wall Lavas Seismograph Lava

RP98004 RP98018 RP98017

Korere Scoria Tokalat Tuff Turagunan Kabiu Palangiangia 410 Ignimbrite

RP98012 RP98011 RP98005 RP98002 RP98001 RP98410

Lava flow, north centres Lava flow, south crater rim Lava flow/dome in crater Lava flow, south base Lava flow, upper west flank Lapilli, Kabakada road cutting Pumice at base, Blue Lagoon Pumice, base of cliff at Barge Tunnel Plinian pumice, Rabarua Scoria, Tavui road cutting Scoria, Tavui road cutting Plinian pumice, Tavui road cutting Plinian pumice, Tunnel Hill road cutting Pumice, Tavui road cutting Pumice, Tunnel Hill road cutting Pumice, base northwest caldera wall Pumice, base northwest caldera wall East base of quarry face Lava flow, east caldera wall Lava flow, Tavurvur instrument tunnel Scoria, Tokalat Creek Welded tuff, Tokalat Creek Lava flow, west base Lava flow, south base Lava flow, south base Pumice, northeast Rabaul Caldera

Tokudukudu Ignimbrite

RP98414

Pumice, Tavui road cutting

087 407

2. Rabaul Caldera Complex

3. Younger stratovolcanoes 4.Tavui volcano

a

RP98013 RP98010 RP09000 RP98007 RP09008

063 230 087 407 087 407 089 404 064 360 089 404 064 360 071 363 071 363 098 362 125 323 127 321 091 391 091 393 129 309 118 328 111 334 102 341

Grid references are with respect to sheet 9389 (Edition 1) Series T601, Rabaul, Papua New Guinea 1: 100,000 Topographic Survey

Bull Volcanol (2016) 78:24

(pers. comm.) were obtained by ICPMS. The largest aggregate data set is that of Nairn et al. (1989) which comprises 95 analyses, 56 of which are analyses of samples collected in 1985–1986, and 39 are from the earlier work of Heming and Carmichael (1973) and Heming (1974, and from unpublished data). These data were used to construct the K2O vs SiO2 covariation plot for volcanic rocks of the Rabaul area (Wood et al. 1995) shown here in modified form in Fig. 4. Major and trace element data for the stratigraphic units at Rabaul for which 40Ar–39Ar ages have been determined or attempted are presented in Table 4. Location data and other details of the chemically analysed samples from these units are given in Table 5. K2O and SiO2 values for these samples are identified in Fig. 4.

Geochemistry and mineralogy The geochemical data set assembled by Nairn et al. (1989) was evaluated by Wood et al. (1995). Most of the analyses of rocks from the Rabaul area volcanoes form a Bmain series^ of compositions (Wood et al. 1995) that is dominated by highK dacite and high-K, high-SiO2 andesite but which also includes medium-K basaltic andesite and medium-K basalt (Fig. 4). Scatter of the geochemical data for the Bmain series^ of compositions is greater at the high-SiO2 end of this trend. This variation may be a function of changes in the source region and zone of magma collection with time (Wood et al. 1995). A simple test of possible changes with time is shown in Fig. 4, where K2O and SiO2 are differentiated according to age relative to the major unconformity below the Kulau Ignimbrite (Nairn et al. 1989). The most K-rich andesites, dacites and Fig. 4 K2O vs SiO2 for volcanic rocks from the Rabaul area, modified after Wood et al. (1995). Analyses of units for which 40 Ar–39Ar dates were determined or attempted are identified. Analyses are broadly differentiated by age: black squares are for units <20 ka and open circles are for older units. Fields and nomenclature are after Gill (1981)

Page 9 of 28 24

rhyodacites are from the earlier part of the stratigraphic sequence, prior to 18 ka BP. Rhyodacites are not known from the younger part of the sequence (<18 ka), as shown in Fig. 4. The youngest rhyodacites are within the Talakua Pyroclastic Subgroup (75–72 ka) and the Barge Tunnel Ignimbrite (75 ka). The only other rhyodacite is the plinian phase of the Malaguna Pyroclastics (160 ka). As pointed out by Wood et al. (1995), the absence of rhyodacite in the younger part of the sequence (and the overall paucity of rhyodacite) may be surprising given that only 15 % more fractionation is required to produce rhyodacite from dacite. This is discussed further in ‘Geochemical trends—Rabaul Caldera Complex, Tovanumbatir’. Rhyolitic rocks are rare in the stratigraphic sequence of the Rabaul area. Only two subaerially outcropping rhyolitic rock units are known: the Raluan Ignimbrite and the Tokudukudu Ignimbrite (Nairn at al. 1995; Wood et al. 1995; McKee 2015). Rock samples dredged from Tavui Caldera in 1990 included rhyolitic pumices (Tufar and Naeser 1992: Wallace and Tufar 1998). Raluan Ignimbrite and Tokudukudu Ignimbrite and the Tavui rhyolitic dredge samples have lower K2O and other incompatible element contents than those expected for Rabaul main series rhyolite of the same SiO2 contents (McKee 2015 and Fig. 4), which has led to the conclusion that the Rabaul main series of compositions and the analysed rhyolitic rocks of the Rabaul area appear to be genetically unrelated (Wood et al. 1995; McKee 2015). Tavui Volcano is suggested as the most likely source of rhyolitic rocks in the Rabaul area (McKee 2015). Minerals present in the basalt-to-dacite main series are plagioclase, augite, hypersthene, Fe–Ti oxides and apatite (Nairn

48.85 0.72 20.27 9.29 0.14 5.44 13.25 2.22 0.45 0.12 −0.53 100.22 48.49 0.45 316 25

24 64 17 5 543 13 29 <5 68 2 11 <10 4

47.48 0.78 19.25 10.68 0.18 5.35 11.78 2.16 0.38 0.17 1.32 99.53 48.35 0.39 – –

– – – – – – – – – – – – –

SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 LOI Total SiO2* K2O* V Cr

Ni Zn Ga Rb Sr Y Zr Nb Ba La Ce Nd Pb

2

17 103 19 10 594 20 47 <1 211 8 20 13 3

50.70 1.01 18.10 10.73 0.21 5.08 10.68 2.68 0.83 0.19 0.36 100.21 50.59 0.83 378 10

4

174 77 19 28 650 13 56 7 235 5 37 20 –

50.88 0.99 18.01 10.58 0.25 4.83 10.09 3.11 0.95 0.26 1.21 101.16 50.91 0.95 310 15

5

18 94 18 14 476 22 64 <5 146 2 <10 13 7

53.37 0.97 16.63 10.94 0.17 4.77 9.61 2.81 1.09 0.18 −0.56 99.98 53.08 1.09 386 15

6

28 70 18 18 554 24 59 <1 225 6 12 12 3

53.35 0.78 17.62 8.30 0.15 5.38 9.89 2.95 1.22 0.18 0.39 99.82 53.44 1.22 285 60

7

<1 89 17 48 399 33 134 3 381 7 22 18 10

56.32 1.14 15.98 9.84 0.21 3.39 7.40 3.70 1.76 0.39 100.18 56.22 1.76 240 4

8

<5 111 16 49 335 36 160 <5 398 14 34 20 14

58.55 1.11 15.34 8.21 0.18 2.07 4.65 3.81 2.83 0.39 1.40 99.54 60.68 2.88 146 <5

9

<1 89 17 48 399 33 134 3 381 7 22 18 10

60.71 0.93 15.63 6.84 0.18 2.14 4.80 4.02 2.71 0.25 98.30 61.76 2.76 144 2

10

<5 96 17 44 410 33 130 <5 373 12 30 19 12

60.74 0.91 15.56 6.62 0.19 2.18 4.80 3.83 2.75 0.29 1.41 99.28 62.06 2.81 119 <5

11

<1 105 17 47 394 32 131 2 415 13 31 20 9

62.35 0.94 15.86 6.14 0.18 2.21 4.96 4.16 2.83 0.28 0.48 99.91 62.41 2.83 145 2

12

7 86 17 39 340 36 147 <5 386 14 31 21 11

61.94 0.87 15.45 6.57 0.16 1.98 4.68 4.62 2.59 0.34 0.60 99.80 62.44 2.61 108 7

13

5 64 17 56 348 37 179 <5 440 15 35 27 11

64.71 0.81 15.93 5.29 0.13 1.44 3.83 4.60 3.14 0.31 0.01 100.20 64.59 3.13 60 <5

14

<5 78 16 43 305 39 171 <5 444 13 39 26 13

64.89 0.84 15.78 4.98 0.17 1.24 3.27 4.60 2.85 0.25 1.21 100.08 65.63 2.88 70 <5

15

4 88 16 48 310 46 161 8 464 14 37 24 11

64.43 0.79 15.05 4.68 0.15 1.29 2.97 4.69 3.34 0.27 2.14 99.80 65.97 3.42 45 1

16

<2 104 15 67 325 38 196 4 465 16 35 20 12

63.37 0.83 15.31 4.58 0.17 1.14 2.84 3.91 3.60 0.22 95.97 66.03 3.75 26 2

17

3 90 15 39 302 46 157 8 436 16 38 21 10

64.73 0.80 15.03 5.04 0.16 1.26 3.34 4.38 2.83 0.27 1.75 99.59 66.16 2.89 72 3

18

<2 86 13 62 295 37 185 4 430 14 35 15 16

62.29 0.74 14.44 4.19 0.15 1.29 2.80 3.77 3.60 0.19 93.46 66.65 3.85 34 <2

19

7 95 14 64 294 37 192 <5 481 14 38 22 17

64.82 0.77 14.78 4.11 0.15 0.99 2.69 4.20 3.77 0.22 3.40 99.90 67.17 3.91 10 <5

20

<5 74 15 50 212 38 180 <5 458 16 31 26 12

65.28 0.71 14.39 3.96 0.12 1.04 2.50 4.35 3.29 0.18 3.98 99.80 68.13 3.43 44 <5

21

2 84 17 61 283 42 168 7 501 18 36 22 10

66.15 0.68 15.04 2.92 0.13 0.87 2.04 3.70 3.81 0.15 3.93 99.42 69.27 3.99 20 1

22

– – – – – – – – – – – – –

70.13 0.32 13.22 2.33 0.08 0.72 2.53 3.39 2.19 0.04 5.05 100.00 73.86 2.31 – –

23

<5 42 12 21 174 32 139 <5 335 12 22 18 9

73.28 0.31 13.03 2.06 0.08 0.47 1.96 4.44 1.66 0.05 2.91 100.25 75.28 1.71 13 <5

24

Page 10 of 28

16 92 18 8 561 17 41 <1 160 7 8 8 4

50.26 0.98 18.50 10.41 0.20 5.29 11.60 2.54 0.74 0.17 0.16 100.71 49.91 0.73 379 9

3

Major and trace element data for the stratigraphic units at Rabaul for which 40Ar–39Ar ages have been determined or attempted

Analysis 1

Table 4

24 Bull Volcanol (2016) 78:24

Bull Volcanol (2016) 78:24 Table 5

Page 11 of 28 24

Chemically analysed sample details

Analysis

Sourcea

Stratigraphic correlation

Field no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

1 2 3 3 4 2 3 5 2 5 2 3 2 2 2 2 5

Turagunan Lavas Palangiangia Lavas Seismograph Lavas Varzin Lavas Tovanumbatir Lavas Kabiu Lavas Watom Lavas Tavui Scoria Vunairoto Lapilli Korere Scoria Malaguna Pyroclastics Tokalat Tuff Latlat Pyroclastics Rabaul Quarry Lavas East Wall Lavas Karavia Welded Tuff Brennan Ignimbrite

18b RAB 53 RP98017 RP98021 8082 RAB 400 RP98025 RP98010 RAB 11a RP98012 RAB 179c RP98011 RAB 184d RAB 70 RAB 403 RAB 198c RP98028

18 19 20 21 22 23 24

2 5 2 2 2 2 2

Rabaul Pyroclastics Cliff Ignimbrite Boroi Ignimbrite Barge Tunnel Ignimbrite Malaguna Pyroclastics Tokudukudu Ignimbrite Raluan Ignimbrite

RAB 125 RP98030 RAB 167 RAB 95 RAB 168a RAB 182 g RAB 182a

Grid ref.b

Chemical classification

Sample lithology, locality

138 339 982 454 087 407 995 351 091 391 065 360 091 393 109 239 098 362 127 321 061 257 071 363

High Alumina Basalt High Alumina Basalt High Alumina Basalt High Alumina Basalt High Alumina Basalt Basaltic Andesite Basaltic Andesite High K Andesite High K Andesite High K Andesite High K Andesite High K Andesite High K Andesite High K Dacite High K Dacite High K Dacite High K Dacite

Lava flow, S base Turagunan Lava flow, NW base Palangiangia Lava flow, NE wall Rabaul Caldera Lava flow, N flank Varzin Lava flow, S flank Tovanumbatir Scoria, quarry E base Kabiu Block, post-caldera dome Scoria, Tavui Fault scarp Lapilli, Kabakada road cut Scoria, NE flank Tovanumbatir Welded tuff, Malaguna road cut Welded tuff, NE flank Tovanumbatir Scoria, S wall Rabaul Caldera Lava, NE wall Rabaul Caldera Lava, E wall Rabaul Caldera Welded tuff, W wall Rabaul Caldera Pumice, NW wall Rabaul Caldera

011 280 071 363 071 364 060 263 089 365 087 407 087 407

High K Dacite High K Dacite High K Dacite High K Dacite High K Dacite Rhyolite Rhyolite

Pumice, Ramale road cut Pumice, NW wall Rabaul Caldera Pumice, NW wall Rabaul Caldera Pumice, W wall Rabaul Caldera Pumice, N wall Rabaul Caldera Pumice, Tavui Fault scarp Pumice, Tavui Fault scarp

098 345 127 321 066 143

a

Sources of analyses: 1, J.B. Gill, University of California, unpublished data; 2, Nairn et al. (1989); 3, R.A. Duncan, Oregon State University, unpublished data; 4, Heming and Carmichael (1973), Heming (1974); 5, S.E. Eggins, Australian National University, unpublished data

b

Grid references are with respect to Sheet 9389 (Edition 1) Series T601, Rabaul, Papua New Guinea 1:100,000 Topographic Survey

et al. 1989; Wood et al. 1995). The phenocryst assemblage of the rhyolites includes quartz and hornblende, in addition to all of the minerals present in the basalt-to-dacite main series (Wood et al. 1995). Magmatic systems and processes The magmatic systems of the Rabaul area comprise the following: Varzin, Vunakanau, WTZ, Rabaul Caldera Complex and Tavui. As little is known about the Varzin and Vunakanau systems commentary here is restricted to WTZ, Rabaul Caldera Complex and Tavui. The WTZ appears to be a major volcanic corridor, possibly representing a deep fracture system (Johnson et al. 2010). Basalt is the dominant product but andesite and dacite are also represented in the eruptive products of some centres in the WTZ. The Rabaul Caldera Complex may have evolved from the WTZ. An active, shallow, large-volume sub-caldera magma reservoir has been interpreted from seismic tomographic studies (Finlayson et al. 2003; Bai and Greenhalgh 2005; Itikarai

2008). The eruptive output is dominated by dacite of medium to high K contents (Wood et al. 1995; Johnson et al. 2010). There is abundant evidence of basaltic material entering the system, preserved as crystals and fragments of mafic composition (Heming 1974; Patia 2003; Johnson et al. 2010). Sources of basaltic injections could include the WTZ (Johnson et al. 2010). Tavui appears to be a youthful caldera system (McKee 2015) but a shallow, active magma system has not been detected within the 12 km depth limit of the seismic tomographic studies to date (Finlayson et al. 2003; Bai and Greenhalgh 2005; Itikarai 2008). Tavui’s eruptive output appears to consist of a low to medium K rock series comprising basaltic andesite, andesite, dacite and rhyolite (Tufar 1990; Wallace and Tufar 1998). Fractional crystallization is a principal magmatic process in the Rabaul area. It accounts well for the basalt-andesite-dacite Bmain series^ of rock compositions (Wood et al. 1995; Johnson et al. 2010) and implies the importance of basalt as the starting point for the more evolved compositions at Rabaul. It should be noted that the main series includes

24

Bull Volcanol (2016) 78:24

Page 12 of 28

material from both WTZ and Rabaul Caldera Complex. Fractional crystallization also may account for the Tavui rock series, but the necessary mass balance computations have not been carried out to confirm this. Magma mixing appears to be an important process at Rabaul Caldera Complex. This was first emphasized by the work of Walker et al. (1981) in terms of the likely triggering of major eruptions by the interaction between mafic and felsic melts. Wood et al. (1995) identified examples of co-genetic mixing and hybrid mixing and Johnson et al. (2010) noted the common occurrence of contrasting co-eruption magmas. Partial melting of crustal material also may play a role in the generation of magma in the Rabaul area. Smith and Johnson (1981) and Wood et al. (1995) concluded that crustal partial melting was the most likely process responsible for the generation of Raluan rhyolite.

Geochronological analytical methods We employed 40Ar–39Ar total fusion and incremental heating methods to determine ages for samples of whole rock, plagioclase and groundmass separates from rocks of the Rabaul area volcanic systems. This analytical work was carried out at the Argon Geochronology Laboratory in the College of Earth, Ocean, and Atmospheric Sciences, Oregon State University, USA. For additional details of analytical methods, see Duncan and Keller (2004). Samples were first crushed and sieved to retain the 200– 300 μm grain size. Plagioclase feldspar was separated from phyric rocks using a Frantz iso-dynamic magnetic instrument. Whole rock (aphyric) chips, groundmass free of phenocrysts, or separated plagioclase crystals were ultrasonically cleaned in distilled water to remove dust and dried at low temperature (60 °C), then packaged in Cu-foil, loaded into evacuated quartz vials and irradiated with fast neutrons at the Oregon State University TRIGA reactor for 6 h at 1 MW power. After decay of short-lived radionuclides (~3 weeks), the samples were loaded on a manifold attached to a double-vacuum, low blank resistance furnace. Incremental heating was monitored by thermocouple, usually in 5–6 discrete temperature steps from 600 to 1400 °C. Active gases were removed with a series of Zr–Al getters. Ar-isotopic compositions (40Ar, 39 Ar, 3 8 A r, 3 7 Ar and 3 6 Ar) were measured mass spectrometrically using a MAP 215/50 instrument, equipped with an electron multiplier. We calibrated our age determinations against standard monitor FCT-3 biotite (28.03 Ma, Renne et al. 1998), which was irradiated with the samples at 7 positions. We made two measurements of the fluence parameter, J, at each position, then regressed the values vs position with a second order polynomial fit. Analytical uncertainties in J (included in the reported 2σ errors) are 0.2–0.3 %.

Ages were calculated in two ways. Individual heating step ages were calculated assuming the initial composition of Ar in the rock or mineral was atmospheric (40Ar/36Ar = 295.5). Evidence that this assumption is true is found in the concordance of step ages (especially middle and high temperature steps), producing a Bplateau^ age, which is the weighted mean (by inverse variance) of the step ages. The statistical significance of the plateau age is assessed by an F-statistic, the mean square of weighted deviations (MSWD) which compares step age uncertainties with the variability among the ages. All plateau ages are statistically significant, with MSWD values of about 2 or less. A second calculation uses the Ar-isotopic compositions of the steps (40Ar/36Ar vs 39Ar/36Ar) to construct an isochron, whose slope gives the age of the sample and whose 40Ar/36Ar intercept gives the initial Ar composition in the rock or mineral. All such intercepts are within uncertainty of the atmospheric value, indicating that no samples have been affected by undegassed (Bexcess^) 40Ar, further corroborating the plateau age calculations. In all cases but one, sample 98018 (discussed below), plateau ages for whole rock or groundmass and plagioclase from the same sample are concordant (within analytical uncertainty) confirming the simple interpretation of plateau ages. A small number of whole rock samples were analysed by total fusion experiments, in which the samples were heated to fusion in a single step, followed by gas cleanup and isotopic measurements. These are equivalent to conventional K–Ar ages.

40

Ar–39Ar geochronology of volcanoes of the Rabaul area 40

Ar–39Ar total fusion and incremental heating ages for whole rock, plagioclase, groundmass, glass and ash from volcanic centres in the Rabaul area are presented in Table 6. The listed analytical uncertainties are 2σ. We measured ages from two fractions (and for sample 98025 three fractions) for 18 of the 27 samples: In 17 cases, the repeated age determinations agreed within 2σ error. All incremental heating ages meet the accepted criteria for reliability (e.g. Duncan and Keller 2004), and we have calculated mean ages for the multiple analyses as the best estimate of sample crystallization ages. Early centres and systems adjacent to Rabaul Caldera Complex Watom is the oldest dated centre in the Rabaul area. There is agreement among the plateau ages on whole rock, groundmass and plagioclase (sample 98025: 1292 ± 43, 1220 ± 100 and 1645 ± 519 ka) for a late-stage lava body within Watom’s summit crater. The mean of these results is 1283 ± 39 ka. No reliable age data were obtained from a second sample (98016) from Watom.

Bull Volcanol (2016) 78:24 Table 6

Page 13 of 28 24

40

Ar–39Ar incremental heating ages for samples from the Rabaul area volcanoes

Location/formation

Sample

Material

Plateau age ±2σ (ka)

Tavui Volcano 410 Ignimbrite 98410 whole rock 246 ± 125 Tokudukudu Ignimbrite 98414 whole rock 372 ± 215 Rabaul Caldera Complex—Younger Formations and Younger Stratovolcanoes Turagunan lavas 98005 whole rock 70 ± 158 Latlat Pyroclastics 98015 plagioclase 61 ± 213 98015 whole rock 74 ± 82 Barge Tunnel Ignimbrite 98027 plagioclase 70 ± 57 98027 whole rock 78 ± 40 Karavia Welded Tuff 98013 plagioclase 51 ± 86 98013 whole rock 66 ± 116 Vunairoto Lapilli 98006 plagioclase 134 ± 129 98006 whole rock 79 ± 94 Palangiangia lavas 98001 whole rock 102 ± 149 Kabiu lavas 98002 whole rock 134 ± 151 98002 groundmass 105 ± 90 Rabaul Caldera Complex—Older Formations Tavui Scoria 98010 plagioclase 303 ± 62 98010 whole rock 425 ± 61 Malaguna Pyroclastics 09000 groundmass 169 ± 51 98008 plagioclase 174 ± 72 98008 groundmass 164 ± 21 Malaguna Pyroclastics 98007 plagioclase 159 ± 57 98007 groundmass 156 ± 19 Boroi Ignimbrite 98029 whole rock 200 ± 77 Boroi Ignimbrite 98009 plagioclase 156 ± 59 98009 groundmass 255 ± 238 Cliff Ignimbrite 98030 plagioclase 241 ± 113 98030 groundmass 168 ± 98 Brennan Ignimbrite 98028 plagioclase 201 ± 75 Rabaul Quarry Lavas 98004 plagioclase 289 ± 141 98004 whole rock 187 ± 53 East Wall Lavas 98018 plagioclase 196 ± 118 98018 groundmass 43 ± 11 Korere Scoria 98012 plagioclase 206 ± 40 98012 whole rock 199 ± 27 Tokalat Tuff 98011 plagioclase 332 ± 123 98011 glass 288 ± 28 Seismograph Lavas 98017 plagioclase 330 ± 56 Early Adjacent Centres Tovanumbatir Lavas 98026 plagioclase 438 ± 125 98026 groundmass 297 ± 31 Tovanumbatir Lavas 98003 plagioclase 382 ± 105 98003 groundmass 315 ± 39 Varzin Lavas 98021 whole rock 524 ± 95 98021 groundmass 509 ± 52 Watom Lavas 98025 plagioclase 1645 ± 519 Watom Lavas 98025 whole rock 1292 ± 43 98025 groundmass 1220 ± 100

Mean age ±2σ (ka)

72 ± 77 75 ± 33 56 ± 69 98 ± 76

112 ± 77

365 ± 43

165 ± 20 156 ± 18

162 ± 57 199 ± 74

200 ± 50 44 ± 11 201 ± 22 290 ± 27

305 ± 30 323 ± 37 512 ± 46

1283 ± 39

Number

Isochron age ±2σ (ka)

Ar/36Ar intercept ± 2σ

4/5 4/6

232 ± 164 198 ± 347

297 ± 14 302 ± 11

4/6 4/5 3/5 3/5 3/5 4/5 4/5 4/5 4/6 5/5 4/5 1

64 ± 344 70 ± 393 74 ± 173 84 ± 115 36 ± 101 32 ± 165 40 ± 130 69 ± 177 84 ± 155 48 ± 613 100 ± 343

296 ± 4 295 ± 39 296 ± 4 293 ± 21 298 ± 6 297 ± 18 299 ± 3 301 ± 12 295 ± 3 298 ± 8 296 ± 5

5/6 5/5 7/10 5/5 1 4/5 1 3/5 4/5 4/5 4/5 3/5 4/5 4/5 4/5 5/5 3/5 4/5 3/5 3/6 5/5 5/5

310 ± 134 516 ± 256 108 ± 48 113 ± 163

294 ± 29 292 ± 9 300 ± 4 302 ± 19

91 ± 119

303 ± 28

335 ± 259 228 ± 138 15 ± 40 265 ± 239 106 ± 152 100 ± 139 222 ± 127 81 ± 96 88 ± 211 21 ± 40 190 ± 92 192 ± 55 290 ± 262 260 ± 55 282 ± 1132

288 ± 14 291 ± 8 298 ± 14 293 ± 22 297 ± 4 302 ± 9 331 ± 40 301 ± 5 318 ± 48 306 ± 18 309 ± 68 296 ± 4 300 ± 23 299 ± 6 296 ± 13

458 ± 242

295 ± 9

334 ± 221

298 ± 11

556 ± 188

294 ± 9

1984 ± 800 1261 ± 46

272 ± 44 306 ± 10

4/5 1 4/5 1 5/6 1 3/5 5/5 1

40

Ages are reported relative to biotite standard FCT-3 (28.03 ± 0.16 Ma, Renne et al. 1998). Age (n = number of heating steps used/total), weighted by the inverse of variance. n = 1 signifies total fusion analysis. Plateau age is the best estimate of crystallization age, due to greater precision compared with concordant isochron age.

24

Page 14 of 28

The dating results for the stratocone Varzin indicate that its activity was considerably younger than that of Watom. For sample 98021, the plateau ages on whole rock (524 ± 96 ka) and groundmass (509 ± 52 ka) give a mean age of 513 ± 46 ka. Tovanumbatir is slightly younger than Varzin cone. The plateau age on plagioclase (382 ± 105 ka) from a basaltic lava flow, sample 98003, exposed in the engulfed southern lower flank of Tovanumbatir is accepted here. The total fusion age on groundmass for this sample (315 ± 39 ka) is concordant, and we calculate a mean age of 323 ± 37 ka. A K–Ar date of 497 ± 50 ka (2σ uncertainty) was reported by Nairn et al. (1989, 1995) for a sample from the same location. A second sample, 98026, from an andesitic lava flow in the upper western flank of Tovanumbatir provided two similar results. The plateau age on plagioclase (438 ± 125 ka) for this sample is concordant with the total fusion age on groundmass (297 ± 31 ka) providing a mean age of 305 ± 30 ka. 40 Ar–39Ar age spectra for samples from the early systems adjacent to Rabaul are shown in Fig. 5. Rabaul Caldera Complex Dating results for the Rabaul Caldera Complex can be divided into two groups reflecting older and younger formations. 40 Ar–39Ar age spectra for the older and younger formations are shown in Figs. 6 and 8, respectively. The results, shown spatially in Figs. 7 and 9, demonstrate some degree of geographical separation of the dated deposits, with the older formations exposed only in the northern and northeastern walls of the caldera and on the northeastern flank of Tovanumbatir, and the younger formations exposed in the western and southern walls of the caldera and at Kabakada. Older Formations The oldest dated rocks within the Rabaul Caldera Complex are the basaltic Seismograph Lavas from the base of the northeastern wall of the caldera. A plagioclase plateau age of 330 ± 56 ka for this formation is accepted here. A large volume, up to 80 m thick, partly welded andesitic scoria flow deposit, the Tokalat Tuff, has patchy exposure in a broad area on the northeastern and eastern lower flanks of Tovanumbatir. Plateau ages of 288 ± 28 and 332 ± 123 ka were returned for glass and plagioclase ,respectively, from sample 98011. The mean of these results is 290 ± 27 ka and is the best estimate of the age of this formation. Directly overlying the Tokalat Tuff is the andesitic Korere Scoria. For sample 98012, the plateau ages for both whole rock (199 ± 27 ka) and plagioclase (206 ± 40 ka) are concordant and give a mean age of 201 ± 22 ka for this deposit. The Rabaul Quarry Lavas and the chemically and petrographically similar East Wall Lavas are amongst the earliest dated rocks of dacitic composition attributed to the Rabaul

Bull Volcanol (2016) 78:24

Caldera Complex. For sample 98004 of the Rabaul Quarry Lavas, the plateau ages for plagioclase and whole rock are 289 ± 141 and 187 ± 53 ka, respectively, giving a mean age of 200 ± 50 ka. This mean age is also consistent with the K– Ar dating result for a sample from the same part of the lava body, reported by Nairn et al. (1995), which is 188 ± 76 ka (2σ uncertainty). For sample 98018 of the East Wall Lavas, the accepted age is the plateau age result on plagioclase, 196 ± 118 ka. The corresponding groundmass analysis produced a significantly younger age (43 ± 11 ka) but with a high MSWD (>2), suggesting some disturbance of the K–Ar system, such as Ar loss. In view of the stratigraphic position of this formation, we accept the older age. The similarity of the ages for these lava bodies allows the possibility that they may be correlatives. A series of early, small to large volume pyroclastic eruptions emplaced a stack of unwelded dacitic ignimbrites around the northern part of Rabaul Caldera Complex. The Cliff Ignimbrite, which is near the base of the stack, yielded a plateau age of 241 ± 113 ka on plagioclase from sample 98030. The age spectrum for the groundmass fraction is compatible at 168 ± 98 ka. The mean of these results is 199 ± 74 ka, which is accepted here as the age of this deposit. The Brennan Ignimbrite is another unit in the pyroclastic sequence near the base of the northern to northwestern wall of the caldera, but its true stratigraphic position is uncertain. The age accepted here for the Brennan Ignimbrite is the plateau age on plagioclase, from sample 98028, 201 ± 75 ka. Near the middle of the ignimbrite stack around the northern part of Rabaul Caldera are the Boroi Ignimbrite and the overlying Malaguna Pyroclastics. These two deposits are also seen in the pyroclastic sequence that is exposed in the Tavui fault scarp on the lower northeastern flank of Tovanumbatir. The accepted ages for the Boroi Ignimbrite at the two locations are the plateau age on plagioclase for sample 98029 (200 ± 77 ka) and the mean (162 ± 57 ka) of the plateau ages on plagioclase (156 ± 59 ka) and groundmass (255 ± 238 ka) for sample 98009. For the Malaguna Pyroclastics, the accepted ages are the means of the plateau ages on plagioclase and groundmass for two samples, 98007 (156 ± 18 ka) and 98008 (165 ± 20 ka). These ages are somewhat greater than those reported by Nairn et al. (1995) which are: Boroi Ignimbrite 100 ± 20 ka and Malaguna Pyroclastics 110 ± 40 ka (2σ uncertainties). Despite its name and the location of its only known outcrop, in the Tavui Fault scarp on the lower northeastern flank of Tovanumbatir and southwestern wall of Tavui Caldera, the Tavui Scoria probably did not originate from Tavui and is more likely to have come from the Rabaul system, as indicated by its relatively high K content. Plateau ages on plagioclase

Bull Volcanol (2016) 78:24

Page 15 of 28 24

98026 Tovanumbatir Lavas plagioclase

200

437.7 ± 124.5 ka 800

400

0

98003 Tovanumbatir Lavas plagioclase

1600

Age [ ka ]

Age [ ka ]

1600

1200

382.2 ± 104.9 ka

800

400

0

10

20

30

40

50

60

70

80

90

0 0

100

98025 Watom Lavas plagioclase

10

20

30

40

50

60

70

80

90

100

98025 Watom Lavas whole rock

2000

6000

4000

Age [ ka ]

Age [ ka ]

1600 5000

1645.0 ± 519.4 ka

1200

1291.8 ± 43.3 ka

3000 800 2000 400 1000 0

0

10

20

30

40

50

60

70

80

90

100

0

0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [ % ]

98021 Varzin Lavas whole rock

3000

Age [ ka ]

2400

1800

524.3 ± 95.4 ka

1200

600

0

0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [ % ]

Fig. 5 40Ar–39Ar age spectra for samples from the early systems adjacent to Rabaul. Heating step ages are shown with ± 2σ uncertainties, plotted against cumulative 39Ar released. Bold horizontal bar indicates concordant steps used in calculation of plateau age (mean weighted by inverse variance)

and whole rock for sample 98010 (303 ± 62 and 425 ± 61 ka) and on groundmass for sample 09000 (169 ± 51 ka) were obtained here. Stratigraphic relationships, showing that the Tavui Scoria overlies the Malaguna Pyroclastics, favour the groundmass age, 169 ± 51 ka, which is the accepted age.

Younger Formations The flat-lying pyroclastic sequence at the Kabakada road cutting has not been convincingly correlated with the sequences at any other location in and around Rabaul Caldera. The ages obtained in this study for

24

Bull Volcanol (2016) 78:24

Page 16 of 28

a 98017 Seismograph Lavas whole rock

2000

98028 Brennan Ignimbrite plagioclase

1000

Age [ ka ]

Age [ ka ]

1600

1200

330.1 ± 56.2 ka

800

600

800

201.2 ± 74.9 ka 400 400

200

0

0 0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

800

98011 Tokalat Tuff glass

700

98011 Tokalat Tuff plagioclase

1000

Age [ ka ]

Age [ ka ]

600

500

288.3 ± 27.5 ka 400

800

332.3 ± 123.3 ka

600

300 400 200

200 100

0

0 0

10

20

30

40

50

60

70

80

90

100

0

98012 Korere Scoria groundmass

1000

10

20

30

40

50

60

70

80

90

100

98012 Korere Scoria plagioclase

700

600

Age [ ka ]

Age [ ka ]

800

600

199.3 ± 27.1 ka

500

205.6 ± 40.4 ka 400

300

400

200

200 100

0

0 0

10

20

30

40

50

60

70

80

90

100

0

10

Fig. 6

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [%]

Cumulative 39Ar Released [%] 40

Ar–39Ar age spectra for samples from older formations of Rabaul Caldera Complex

sample 98026 from the Vunairoto Lapilli, which lies near the base of the Kabakada sequence, are 134 ± 129 ka (plateau age for plagioclase) and 79 ± 94 ka (plateau age for whole rock). The mean of these results, 98 ± 76 ka, is accepted here as the age for the Vunairoto Lapilli. This date provides an older limit for the overlying Kabakada sequence.

Three large volume formations comprise the bulk of the exclusively pyroclastic material exposed in the southern and western walls of Rabaul Caldera. At the base is the dacitic Karavia Welded Tuff. The age for sample 98013 from this deposit is 56 ± 69 ka, which is the mean of plateau ages on plagioclase (51 ± 86 ka) and whole rock (66

Bull Volcanol (2016) 78:24

Page 17 of 28 24

b 98018 East Wall Lavas groundmass

350

98018 East Wall Lavas plagioclase

4000

Age [ ka ]

Age [ ka ]

300

250

200

2000

196.2 ± 118.3 ka

42.8 ± 10.8 ka

150

3000

1000

100

0 50

0 0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

1000

98030 Cliff Ignimbrite groundmass

800

98030 Cliff Ignimbrite plagioclase

1600

1400

Age [ ka ]

Age [ ka ]

1200

600

168.3 ± 97.7 ka 400

1000

800

241.2 ± 112.9 ka 600

400

200

200

0

0 0

10

30

40

50

60

70

80

90

100

0

98004 Rabaul Quarry Lavas whole rock

1000

800

600

187.1 ± 52.7 ka

10

20

30

40

50

60

70

80

90

100

98004 Rabaul Quarry Lavas plagioclase

1000

Age [ ka ]

Age [ ka ]

20

800

288.5 ± 140.5 ka 600

400

400

200

200

0

0 0

10

20

30

40

50

60

70

80

Cumulative 39Ar Released [%]

0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Relesaed [%]

Fig. 6 continued.

± 116 ka). The overlying rhyodacitic Barge Tunnel Ignimbrite yielded ages within error of the Karavia Welded Tuff, viz. for sample 98027, 70 ± 57 ka (plateau age for plagioclase) and 78 ± 40 ka (plateau age for whole rock), giving a mean age of 75 ± 33 ka. The Barge Tunnel Ignimbrite was also dated by Nairn et al. (1995) who reported an age of 40 ± 40 ka (2σ uncertainty). The

uppermost formation in the sequence is the dominantly dacitic Latlat Pyroclastics, dated here at 74 ± 82 ka (plateau age for whole rock, sample 980152) and 61 ± 213 ka (plateau age for plagioclase), giving a mean of 72 ± 77 ka. Radiocarbon dates for this formation reported by Nairn et al. (1989, 1995) are 36,690 ± 380 BP and greater than 38,000 BP. A conservative interpretation of the new results

24

Bull Volcanol (2016) 78:24

Page 18 of 28

c 98009 Boroi Ignimbrite groundmass

1000

98009 Boroi Ignimbrite plagioclase

2000

1600

Age [ ka ]

Age [ ka ]

800

254.7 ± 237.6 ka 600

1200

156.0 ± 59.0 ka

800 400

400 200

0 0 0

10

20

30

40

50

60

70

80

90

0

100

10

20

30

40

50

60

70

80

90

100

4000

98029 Boroi Ignimbrite whole rock

2000

09000 Tavui Scoria groundmass

3500

1600

Age [ ka ]

Age [ ka ]

3000

1200

2500

2000

1500 800

199.7 ± 76.6 ka

169.4 ± 50.8 ka 1000

400 500

0

0 0

10

20

30

40

50

60

70

80

90

0

100

1200

10

20

30

40

50

60

70

80

90

100

1200

98010 Tavui Scoria whole rock

1000

98010 Tavui Scoria plagioclase

1000

Age [ ka ]

Age [ ka ]

800 800

424.8 ± 61.1 ka 600

600

303.0 ± 61.6 ka

400

200 400

0 200

0 0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [ % ]

0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [ % ]

Fig. 6 continued.

suggests that the age of the Karavia Welded Tuff is <125 ka and that the ages of the Barge Tunnel and Latlat deposits are younger than 108 ka. The new dating results are consistent with known stratigraphic relationships and suggest that the major formations exposed in the southern and western parts of the caldera wall are younger than about 125 ka.

Younger stratovolcanoes Early products of Palangiangia and Kabiu exposed near the base of the northeastern wall of Rabaul Caldera have yielded similar ages. For Palangiangia, the age accepted here is the plateau age on whole rock from sample 98001, 102 ± 149 ka. For sample

Bull Volcanol (2016) 78:24

Page 19 of 28 24

d 98007 Malaguna Pyroclastics plagioclase

1000

98008 Malaguna Pyroclastics plagioclase

1000

Age [ ka ]

Age [ ka ]

800 800

600

600

174.3 ± 72.4 ka 400

158.5 ± 56.6 ka 400 200

200

0

0 0

10

20

30

40

50

60

70

80

90

0

100

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [%]

Cumulative 39Ar Released [%]

Fig. 6 continued.

98002 from Kabiu, the plateau ages for whole rock (134 ± 151 ka) and groundmass (105 ± 90 ka) give a mean age of 112 ± 77 ka, which is accepted here. Sample 98005 from an early-emplaced lava flow at the western base of Turagunan has returned an age of 70 ± 158 ka (plateau age for whole rock). Fig. 7 Locations and 40Ar–39Ar ages for samples from older formations from Rabaul Caldera Complex

In the context of the stratigraphy and chronology of the major pyroclastic eruptions from the Rabaul Caldera Complex and also the chronology available for Tovanumbatir, these ages seem reasonable, although with large uncertainties. 40Ar–39Ar age spectra for samples from the younger stratovolcanoes are shown

OLDER FORMATIONS

Tavui Sc 169 ka Malaguna P 156 ka

4°10'

Korere Sc 201 ka

Brennan Ig 201 ka

Tokalat Tuff 290 ka

Cliff Ig 199 ka

Rabaul Quarry L 200 ka

Malaguna P 165 ka

East Wall L 196 ka

Boroi Ig 200 ka

4°20'

Boroi Ig 162 ka

Seismograph L 330 ka

152°10'

24

Bull Volcanol (2016) 78:24

Page 20 of 28

in Fig. 8, and the sampling localities and accepted ages are shown in Fig. 9. Tavui volcano The timing of volcanic activity at Tavui Volcano was not resolved by this study. From other investigations, Tavui activity was indicated at 6.90 ± 0.04 ka BP for the emplacement of the Raluan Ignimbrite and at 78.7 ± 12.5 ka BP for the deposition of the Tokudukudu Ignimbrite (McKee 2015). The only products of Tavui for which dating was attempted in this study were the 410 Ignimbrite and the Tokudukudu Ignimbrite, samples 98410 and 98414 respectively. 40Ar–39Ar age spectra for these samples are shown in Fig. 10. For the 410 Ignimbrite, the plateau age (246 ± 125 ka) has a large uncertainty. Similarly for the Tokudukudu Ignimbrite, the plateau age (372 ± 215 ka) and the isochron age (198 ± 347 ka) have very large uncertainties. These large uncertainties reflect large proportions of atmospheric Ar. In addition, these dates appear to be too old considering the respective stratigraphic positions of these deposits. Initially, the 410 Ignimbrite was thought to be a newly discovered formation; however, it became clear later that this deposit is, in fact, the Raluan Ignimbrite which is the product of the penultimate major eruption in the Rabaul area. The Tokudukudu Ignimbrite is younger than Tavui Scoria (169 ± 51 ka) on the basis of stratigraphy, as seen in the Tavui Fault scarp. Thus, the new 40Ar–39Ar ages for the Tokudukudu Ignimbrite are clearly not consistent with its stratigraphic position, nor with the age determined separately, 78.7 ± 12.5 ka BP, using the thermoluminescence method (McKee 2015).

Discussion Synthesis of the early eruption history of volcanoes of the Rabaul area 1. Va r z i n D e p r e s s i o n ≈ 1 . 5 M a . I n t h e M i d d l e Pleistocene (≈1.5 Ma), a large volcanic system started to form at the northern margin of the Gazelle Peninsula, at a location south of the laterformed Rabaul system. One or more episodes of caldera formation led to the development of Varzin Depression. 2. Watom ≈1.4 Ma. Also in the Middle Pleistocene activity commenced at the major volcanic centre Watom at the northern end of the GVZ. The structure that developed is a stratovolcano. A 2-km-wide large crater or small caldera at its summit indicates

3.

4.

5.

6.

that it experienced major explosive activity in the latter part of its life. The last stage of the development of Watom was at about 1.28 Ma when a lava dome or coulée was emplaced in the summit crater/ caldera. Vunakanau ≈1 Ma. The next major development was in the south where a new volcanic structure, the Vunakanau centre, began to develop on the northern flank of the Varzin system. The timing of the development of the Vunakanau centre is unknown, but is estimated to be Middle–Late Pleistocene, post-dating the early development of the Varzin system and pre-dating the Rabaul Caldera Complex. An age of 1 Ma is assigned tentatively. Mount Varzin Cone ≈510 ka. The final stage of activity of the Varzin system was at about 510 ka when the basaltic cone Mount Varzin was formed at the northeastern edge of Varzin Depression. Probably at about the same time, another small cone, Wairiki, was formed at the eastern edge of Varzin Depression. Tovanumbatir ≈500–300 ka. Activity then shifted northwards again and a new centre, Tovanumbatir, began to develop about 14 km southeast of Watom Volcano. Early basaltic lavas from Tovanumbatir, exposed at the southern base of the volcano, were erupted at about 500–320 ka. Andesitic lavas were emplaced on the upper western flank of the volcano at about 300 ka. Explosive activity at Tovanumbatir generated numerous scoria deposits which are interspersed with lava flows. There are indications of further geochemical evolution of the magmas supplying the Tovanumbatir centre as highsilica dacitic rocks have been recovered from parts of the edifice. Rabaul Caldera Complex ≈330 ka–present. The oldest known formation from the Rabaul system is the basaltic Seismograph Lavas (≈330 ka) exposed at the base of the northeastern wall of the caldera. These lavas probably originated from a precursor to the caldera complex, representing a southeastwards migration of activity from the older northwestern part of the WTZ. Slightly younger than the Seismograph Lavas is the large-volume, welded, andesitic scoria flow deposit, the ≈290 ka Tokalat Tuff, which is found only on the eastern and northeastern flanks of Tovanumbatir. Another andesitic scoria flow deposit, the Korere Scoria, which overlies the Tokalat Tuff on the northeastern flank of Tovanumbatir, was emplaced at about 200 ka. The distribution of these scoria flow deposits and their ages raise the question of whether they might be products of Tovanumbatir. Taking into account the relatively large uncertainties for some of the new dating results, it is possible that there could have been a period of overlapping activity at Tovanumbatir and at the precursor of the Rabaul Caldera system. Dacitic volcanism had become established at the

Bull Volcanol (2016) 78:24

a

Page 21 of 28 24

1200

1200

98006 Vunairoto Lapilli whole rock

1100

1000

900

900

800

800

Age [ ka ]

Age [ ka ]

1000

700

600

79.4 ± 93.5 ka

500

98006 Vunairoto Lapilli plagioclase

1100

700

600

134.4 ± 128.8 ka

500

400

400

300

300

200

200

100

100

0

0 0

10

20

30

40

50

60

70

80

90

0

100

2400

10

20

30

40

50

60

70

80

90

100

90

100

1200

98013 Karavia Tuff groundmass

1800

98013 Karavia Tuff plagioclase

1000

Age [ ka ]

Age [ ka ]

800 1200

66.3 ± 115.7 ka 600

600

51.0 ± 86.2 ka

400

0 200

0

0

10

20

30

40

50

60

70

80

90

100

0

1200

10

20

30

40

50

60

70

80

1000

98002 Kabiu Lavas whole rock

98001 Palangiangia Lavas whole rock

1000

800

Age [ ka ]

Age [ ka ]

800

600

101.6 ± 149.3 ka 400

600

133.8 ± 150.5 ka 400

200

200 0

0 0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [%]

0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar Released [ % ]

Fig. 8 40Ar–39Ar age spectra for samples from younger formations of Rabaul Caldera Complex and for samples from the younger stratovolcanoes at Rabaul

Rabaul system by about 200 ka with the emplacement of dacitic lavas: the Rabaul Quarry Lavas and the East Wall Lavas, in the north-northeastern and northeastern parts of the system, respectively. Structural features of these lava bodies suggest that their source or sources may have been located in the

northeastern part of the Rabaul system or within the WTZ. Dacitic ignimbrites began to be produced at about 200 ka. Individual ignimbrite thicknesses near the base of the ignimbrite stack in the northern to northwestern part of the caldera wall are only a few meters suggesting that the

24

Bull Volcanol (2016) 78:24

Page 22 of 28

b

1200

1200

98015 Latlat Pyroclastics whole rock

1000

Age [ ka ]

Age [ ka ]

800

600

73.5 ± 82.4 ka

400

98015 Latlat Pyroclastics plagioclase

1000

800

600

61.4 ± 212.5 ka 400

200

200

0

0 0

10

20

30

40

50

60

70

80

90

100

0

400

10

20

30

40

50

60

70

80

90

100

1200

98027 Barge Tunnel Ignimbrite whole rock

350

98027 Barge Tunnel Ignimbrite plagioclase

1000

300

Age [ ka ]

Age [ ka ]

800

250

78.4 ± 39.5 ka

200

600

70.3 ± 56.6 ka

400 150

200 100

0

50

0 0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar released [ % ] 2000

98005 Turagunan Lavas whole rock

Age [ ka ]

1600

1200

800

70.3 ± 158.1 ka 400

0 0

10

20

30

40

50

60

70

80

90

100

Cumulative 39Ar released [ % ]

Fig. 8 continued.

early ignimbrites may have been of modest volume. This changed at about 160 ka with the eruption of the voluminous Boroi Ignimbrite and Malaguna Pyroclastics. Both of these formations are found beyond Rabaul Caldera, notably on the eastern and northeastern flanks of Tovanumbatir. It is

likely that caldera formation was associated with the emplacement of the Malaguna Pyroclastics and possibly with the Boroi Ignimbrite. A phase of predominantly andesitic volcanism, that produced the scorias of the Tavui Subgroup, marks the end of the early dacitic ignimbrite sequence.

Bull Volcanol (2016) 78:24

Page 23 of 28 24

Fig. 9 Locations and 40Ar–39Ar ages for samples from younger formations from Rabaul Caldera Complex and from the young stratovolcanoes at Rabaul

YOUNGER FORMATIONS

4°10'

Palangiangia 102 ka

Vunairoto Lapilli 98 ka

Kabiu 112 ka Turagunan 70 ka

Barge Tunnel Ig 75 ka Karavia Welded Tuff 56 ka 4°20'

Latlat P 72 ka 152°10'

125 ka. The major formations exposed in the southern and western walls of the caldera, the Karavia Welded Tuff, the Barge Tunnel Ignimbrite and the Latlat Pyroclastics (see Fig. 9), are all younger than about 125 ka. The Vunairoto Lapilli, which is younger than 100 ka, and the overlying Kabakada sequence also may be linked to the southern eruptive focus of the Rabaul system. All of the younger

It appears that the early activity at Rabaul Caldera had its eruptive focus in the northern part of the Complex. All of the early products are found in the northeastern to northnorthwestern walls of the caldera and on the eastern and northeastern flanks of Tovanumbatir (see Fig. 7). A shift in eruptive focus to the southern part of the Rabaul Caldera Complex took place in the interval 160–

2400

2400

98410 Ash glass

2000

98414 Ash glass

2000

Age [ ka ]

Age [ ka ]

1600

1200

245.7 ± 125.1 ka

1600

1200

372.2 ± 215.3 ka

800

800 400

400 0

0 0

10

20

30

40

50

60

70

80

Cumulative 39Ar Released [ % ]

Fig. 10

40

Ar–39Ar age spectra for samples from Tavui Volcano

90

100

0

10

20

30

40

50

60

70

80

Cumulative 39Ar Released [ % ]

90

100

24

Bull Volcanol (2016) 78:24

Page 24 of 28

Table 7 Geochemical trend in early formations at Rabaul

Formation

Age (ka)

Rock type

Malaguna Pyroclastics Cliff Ignimbrite Rabaul Quarry Lavas Korere Scoria Tokalat Tuff

165 ± 20, 156 ± 18 199 ± 74 200 ± 50 201 ± 22 290 ± 27

High silica dacite – 69 wt% SiO2 Dacite – 67 wt% SiO2 Dacite – 65 wt% SiO2 High silica andesite – 62 wt% SiO2 High silica andesite – 62 wt% SiO2

Seismograph Lavas

330 ± 56

Basalt – 50 wt% SiO2

(radiocarbon-dated) pyroclastic formations emanated from the southern part of Rabaul Caldera. 7. Younger stratovolcanoes ≈110 ka–present. Early activity at the major stratocone system Palangiangia–Kabiu took place at about 110 ka. This system erupted dominantly mafic material, basalts and basaltic andesites, but more evolved products (dacites) are known from later activity at Kabiu. Turagunan started to form at a later time and was active by about 70 ka. These volcanoes are still potentially active, having erupted within the last few thousand years, demonstrating that these centres have very long lives, as shown earlier by Tovanumbatir. 8. Tavui >80 ka–present. Although establishment of the Rabaul Caldera Complex was the main development in the area over the past ≈300 ky, another major system, Tavui Volcano, started to form within the same interval. The oldest rocks attributed to Tavui have been dated at about 80 ka, but it seems likely that Tavui was active before this time. The record of activity from the Tavui system is sparse; however, it is believed to have been active quite recently, at about 6.9 ka BP, when the Raluan Ignimbrite was produced (McKee 2015).

Geochemical trends—Rabaul Caldera Complex, Tovanumbatir A geochemical trend of progressively increasing silica contents may be evident in the products of the early stages of development of the Rabaul system, and there may be a similar trend over the life of Tovanumbatir. For the Rabaul system, the possible geochemical trend is shown in Table 7. The oldest known eruptives attributed to the Rabaul system, the 330 ka Seismograph Lavas, are amongst its most mafic products, having basaltic composition with about 50 wt% SiO2. A progressive trend of more felsic compositions in later-erupted products is evident until about 200 ka, near the beginning of a series of ignimbrite eruptions. The earliest dated member of this sequence of ignimbrites, the 199 ka Cliff Ignimbrite, has a dacitic composition with about 67 wt% SiO2. Similar compositions are common for many of the products of the later major eruptions of the Rabaul system, although high silica dacite,

about 69 wt% SiO2, was erupted in the opening (plinian) phase of the Malaguna Pyroclastics at about 160 ka. The trend of increasing SiO2 contents of successive eruptive deposits was broken within the Malaguna Pyroclastics. Marked geochemical variation occurred within the Malaguna Pyroclastics from the high-SiO2 dacite (69 wt% SiO2) of the initial plinian pumice phase to high-SiO2 andesite (62 wt% SiO2) of the overlying welded scoria phase, to andesite (≈59 wt% SiO2) in the glassy andesite streaks of banded pumice within the later-erupted ignimbrite (Nairn et al. 1989). The overlying, and perhaps related, Tavui Subgroup (see ‘Shift of eruption focus within Rabaul Caldera’) is dominated by scoria deposits having andesitic composition (≈56 wt% SiO2). The possible geochemical trend for products of Tovanumbatir is based on limited sampling and perhaps should be referred to as a geochemical range rather than a trend. The oldest known products of Tovanumbatir are the basaltic lavas having about 51 wt% SiO2 exposed in the engulfed southern base of the edifice dated at 497 ± 50 ka (K–Ar method, uncertainty 2σ) and 323 ± 37 ka. A lava flow of andesitic composition (58 wt% SiO2) on Tovanumbatir’s upper western flank was dated at 305 ± 30 ka. High-silica dacites have been sampled (P. Wallace, pers. comm. 2002) from other parts of Tovanumbatir, although their stratigraphic positions are unclear. These results suggest that following early activity involving relatively primitive mafic magmas, the erupted products began to show evidence of chemical evolution. At both the Rabaul system and at Tovanumbatir, this process occurred over periods of tens of thousands of years. Rhyolitic products are not known from either the Rabaul system or from Tovanumbatir, suggesting that the fractional crystallization process which accounts well for the basaltic-andesitic-dacitic products of the Rabaul system (Heming 1974; Wood et al. 1995) was never able to reach its full felsic endpoint. The reason for this could be a high rate of recharge of magma chambers with basaltic melts, as has been suggested by Wood et al. (1995). In addition, there is abundant evidence of magma mixing through a large portion of the stratigraphic record at Rabaul (Wood et al. 1995; Johnson et al. 2010), possibly buffering magma compositions towards dacite.

Bull Volcanol (2016) 78:24

Rare lava flows at the Rabaul Caldera Complex Lava flows are rare at the Rabaul Caldera Complex. Only three lava formations are known for the interval from the earliest developments at the complex to the time of the latest caldera forming eruption, a period of more than 300 ky. They are the Seismograph Lavas (≈330 ka), the Rabaul Quarry Lavas (≈200 ka) and the East Wall Lavas (≈200 ka). All of these lava formations are in the northeastern quadrant of the caldera. In the post-1400 BP caldera, the only known lava flows are from vents in the northeastern sector of the caldera: Tavurvur, Rabalanakaia and Sulphur Creek. There is evidence of mafic material in these otherwise felsic lava flows and in the felsic pyroclastic products of these centres (Heming 1974; Patia 2003). The consistent occurrence of lavas in the northeastern part of the caldera and the evidence of magma mixing may suggest a link to the source(s) of mafic magmas that feed the stratovolcanoes that fringe the northeastern flank of the caldera, viz. Palangiangia, Kabiu and Turagunan. The geochronological results presented here show that these dominantly mafic systems have been active for very long periods, more than 100 ky in the case of Palangiangia and Kabiu. We suggest that the mafic magmas that have fed these volcanoes may also have interacted with the more felsic magmas that have fed eruptions within Rabaul Caldera, permitting effusive activity in the northeastern sector of the caldera for at least the last 100 ky. The apparent restriction of the earliest known lava bodies to the northeastern quadrant of the caldera poses the question of whether external mafic magma influences acted during the early development of the Rabaul system. Because of the possibility of overlapping activity from Tovanumbatir and the Rabaul complex, it is possible that interaction between their magma systems also may have taken place. Other possible external mafic magma influences could be one or more assumed sources on the southeastern extension of the early part of the WTZ, which could include a precursor of the Palangiangia-Kabiu centre.

Page 25 of 28 24

and Malaguna Pyroclastics. If this assumption is correct, then about 15 ignimbrites could have been emplaced between the Cliff and Boroi Ignimbrites during a period of about 40 ky. The implied average ignimbrite eruption frequency would be of the order of one every 3 ky. The apparent sequence-ending ignimbrite, the Malaguna Pyroclastics, is particularly interesting because of its chemical and lithological diversity. An early-erupted high-SiO2 dacitic pumiceous phase (≈69 wt% SiO2) is succeeded by variably welded and sintered andesitic scoriaceous tuff (≈62 wt% SiO2), followed by non-welded ignimbrite containing distinctive mixed brown and grey streaky pumice having andesitic bulk composition, ≈61 wt% SiO2 (Nairn et al. 1989). Based on geochemical data and on the new 40Ar–39Ar dating results, it is possible that the andesitic components of the Malaguna Pyroclastics are related to the overlying andesitic scorias of the Tavui Subgroup, the uppermost unit of which, the Tavui Scoria, has SiO2 contents of about 56 wt%. The similarity of the new 40Ar–39Ar ages for these formations—Malaguna Pyroclastics (156 ± 18 and 165 ± 20 ka) and Tavui Scoria (169 ± 51 ka)—allow for the possibility that these deposits could represent a single eruptive sequence. The volume and the variations in composition of the Malaguna Pyroclastics indicate a major magma mixing event. In addition, if the Tavui Subgroup is a continuation of the Malaguna Pyroclastics, the two together would constitute one of the largest and most complex eruptive events at the Rabaul centre. The large volume of the combined Malaguna Pyroclastics and Tavui Subgroup would represent a significant emptying of the magma chamber that fed the sequence of ignimbrite eruptions. Stratigraphically, it appears that the emplacement of the Malaguna Pyroclastics and Tavui Subgroup, whether as separate or related deposits, represents the climax of the early ignimbrite sequence. Additionally, and perhaps more importantly, the eruption of the Malaguna Pyroclastics and Tavui Subgroup may represent the last major activity from the early (northern) eruption focus of the Rabaul system.

Early sequence of frequent ignimbrite eruptions

Shift of eruption focus within Rabaul Caldera

Dacitic ignimbrite-producing eruptions had commenced at the Rabaul system by about 200 ka when the Brennan Ignimbrite and Cliff Ignimbrite were emplaced. In a 20-m-high exposure at the north-northwestern base of the caldera wall, the Brennan Ignimbrite and the Cliff Ignimbrite are at the base of a stack of 7 ignimbrites (Fig. 2). The Boroi Ignimbrite and Malaguna Pyroclastics (erupted at about 160 ka) are not exposed at this locality but are present in neighbouring exposures and lie about 30–40 m above the basal ignimbrite sequence. Assuming upward continuation of the ignimbrite sequence, there is space for another 10 ignimbrites between the basal stack and the stratigraphic positions of the Boroi Ignimbrite

The main rock formations exposed in the southern and western walls of Rabaul Caldera (Karavia Welded Tuff, Barge Tunnel Ignimbrite and Latlat Pyroclastics) are all younger than 125 ka. It is possible that they are present in the northern part of the caldera, overlying the much older pyroclastic sequence, but have not been recognized there, because of vegetation cover and access difficulties. The coarse textural characteristics of the Karavia Welded Tuff at its type locality at the southwestern base of the caldera wall, immediately southwest of the Vulcan headland, have been interpreted as evidence that this location is close to source (Nairn et al. 1989, 1995). Similarly, clast size variation and total thickness parameters

24

Bull Volcanol (2016) 78:24

Page 26 of 28

for the Latlat Pyroclastics indicate a source in the southwestern part of the caldera (Nairn et al. 1995). The new dating results together with distribution, thickness and textural information for the Bsouthern^ pyroclastic formations indicate that the focus of eruptive activity within the Rabaul Caldera Complex was in the more southerly part of the system for much of the last 100 ky. The shift in eruption focus seems to have taken place during the interval between the time of emplacement of the Malaguna Pyroclastics and Tavui Subgroup (≈160 ka) and the time of the eruptions responsible for the Karavia Welded Tuff, Barge Tunnel Ignimbrite and Latlat Pyroclastics (<125 ka). Palangiangia, Kabiu and Turagunan Early accounts of the geology of the Rabaul area proposed that Kabiu is older than Palangiangia (Fisher 1939; Heming 1974), and that Palangiangia represented a migration of activity from Kabiu (Fisher 1939). Part of the reason for this interpretation could be the relative sizes of the two cones, Kabiu being considerably larger and thus conceivably older than Palangiangia. However, exposures in the northeastern wall of the caldera show Kabiu lavas and pyroclastics apparently overlying early-erupted products from Palangiangia. Considering the associated uncertainties, the new dating results for Palangiangia (102 ± 149 ka) and Kabiu (112 ± 77 ka) are consistent with the observed stratigraphic relationship. The dating result for Turagunan (70 ± 158 ka) is broadly similar to the results for Palangiangia and Kabiu. The large uncertainties associated with these results preclude any attempts to correlate the commencement of activity at these centres with developments elsewhere in the Rabaul area.

In the recent eruption history of the Rabaul area (Tables 1 and 2), there were at least five and perhaps as many as nine major (≥5 km3) eruptions in the past 18 ky (Nairn et al. 1995). Most of these eruptions emanated from the Rabaul system, but one probably was sourced at Tavui. These data indicate average return intervals for major eruptions of between 2.6 and 6 ky. This frequency is similar to that estimated for a time interval at a very early stage of the development of the Rabaul system. The implied early ignimbrite eruption frequency determined from exposures in the northwestern wall of Rabaul Caldera is of the order of one every 3 ky. These results suggest that a Bmodal^ frequency of ignimbriteproducing eruptions in the Rabaul area could be in the range 2.6 to 6 ky. The apparent lull in major activity between the ≈160 ka eruption of the Malaguna Pyroclastics (and perhaps Tavui Subgroup) and the sequence of eruptions that produced the flat-lying Kabakada deposits and the major formations exposed in the southern and western walls of the caldera (<125 ka) may be an indication of variability in the eruption frequency at Rabaul over the longer term. The eruption frequency for Tavui Volcano is likely to be much greater than that inferred from the two events in the past ≈80 ky as is indicated by the stratigraphic record and the available radiometric dates. Tavui may always have been a submarine system, and the great depth of water covering its vents would likely prevent material from many eruptions breaking through the sea surface. The implication of this is that most eruptions probably would not leave any deposits in the subaerial pyroclastic sequence. Only the larger eruptions, of the scale of the Raluan Ignimbrite event (at 6.9 ka BP), would be powerful enough to leave a subaerial record.

Concluding remarks Frequency of eruptions in the Rabaul area and volcanic risk Determination of the range of return intervals or frequency of eruptions in the Rabaul area is important in assessing the risk from volcanic hazards. However, the long-term eruption frequency at Rabaul is difficult to determine because of stratigraphic problems: poor exposure, patchy deposition and unconformities render stratigraphic sequences incomplete. In addition, some radiometric dates are imprecise. In reality, eruptions from any volcanic system are likely to be episodic, exhibiting a broad range of return intervals. Thus, while it is not possible to reliably determine a long-term eruption frequency or range of return intervals for the Rabaul area, the new dating results do provide some insight into the frequency of eruptions at an early stage in the development of the Rabaul system. This information makes an interesting comparison to eruption frequency calculations for the Late Pleistocene and Holocene.

This 40Ar–39Ar incremental heating dating program has demonstrated that: 1. The Bolder^ volcanoes of the Rabaul area, systems to the south (Varzin Depression) and northwest (Watom Island), probably commenced activity considerably more than 1 my ago. The youngest of the Bolder^ centres were active until about 500 ka (Mount Varzin) and 300 ka (Tovanumbatir). Watom and Tovanumbatir represent early stages of development of a major northwest-trending volcanic corridor (WTZ) that has been exploited by mafic magmatism. 2. The Rabaul Caldera Complex has a long history of eruptions, spanning >330 ky. Early eruptions, occurring between about 330 and 200 ka, were basaltic and andesitic possibly from a precursor of the caldera system and represent southeastwards migration of activity from the older parts of the WTZ. Dacitic compositions appeared at about

Bull Volcanol (2016) 78:24

3.

4.

5.

6.

7.

200 ka, leading to a sequence of small to large volume explosive eruptions and the generation of many ignimbrites. The larger ignimbrite eruptions were accompanied by formation or modification of calderas. The focus of activity at the Rabaul Caldera Complex was initially in the northern part of the system as products of the early activity are found in the northern and northeastern walls of the caldera and on the northeastern and eastern flanks of Tovanumbatir. A shift in the focus of activity within the caldera to a more-southerly location took place between 160 and 125 ka. All of the younger (<125 ka) major pyroclastic formations, which make up the bulk of the exposure in the southern and western walls of Rabaul Caldera, were erupted from a source or sources in the south-central part of the complex. The stratovolcanoes Palangiangia, Kabiu and Turagunan on the northeastern flank of Rabaul Caldera may have commenced activity in the period 112–70 ka. However, considering the large uncertainties associated with these results, activity at these centres may have begun much earlier, possibly as early as about 190–250 ka. The stratovolcanoes at Rabaul appear to be capable of remaining active for at least 100 ky. Activity at Tovanumbatir appears to have spanned the interval 500– 300 ka. Early eruptions from Kabiu and Palangiagia may have occurred before about 100 ka, and both of these centres are regarded as still being potentially active, having erupted within the last few millennia. The Tavui system has developed in parallel with the Rabaul Caldera Complex as Tavui eruptions are suggested at ≈79 and ≈7 ka. However, the sparsity of Tavui eruptives in the subaerial stratigraphic record hinders efforts to determine the eruptive history of this system. The frequency of major explosive eruptions at Rabaul has been of the order of one per 2.6 to 6 ky during two periods: in the interval 200–160 ka and during the last 18 ky. However, activity rates at volcanic systems can be highly variable, and vigilance must be maintained, especially for any signs of unrest that could constitute preparation for a major eruption.

Acknowledgments AusAID provided funding for the fieldwork for this project and for the subsequent analytical work at Oregon State University, USA. Rabaul Volcanologial Observatory provided staff members Jonathan Kuduon and (late) Herman Patia and a vehicle to assist in the fieldwork. We gratefully acknowledge this support. We thank Professor Hugh Davies of the Earth Sciences Department, University of Papua New Guinea, for thoughtful reviews of earlier versions of the manuscript. Helpful reviews were also provided by Dr R. Wally Johnson, formerly of Geoscience Australia, Dr Andrew Calvert of USGS and an anonymous reviewer. We thank Marisa Sari Egara of Port Moresby Geophysical Observatory for help with word processing of the manuscript and Sonick Taguse of Papua New Guinea’s Mineral Resources Authority

Page 27 of 28 24 for preparing the line diagrams. Assistance in the Argon Geochronology Laboratory was provided by John Huard. COM publishes with the permission of the Secretary, Department of Mineral Policy and Geohazards Management, Papua New Guinea.

References Bai C, Greenhalgh S (2005) 3-D multi-step travel time tomography: imaging the local, deep velocity structure at Rabaul Volcano, Papua New Guinea. Phys Earth Planet Inter 151:259–275 Blong R, McKee CO (1995) The Rabaul eruption, 1994: destruction of a town. Natural Hazards Research Centre, Macquarie University Corbett G, First DM, Hayward SB (1991) The Maragorik Prospect, East New Britain, Papua New Guinea. In: Rogerson R (ed) Proceedings of the PNG Geology, Exploration and Mining Conference 1991, Rabaul. The Australasian Institute of Mining and Metallurgy, Melbourne, pp 112–124 Duncan RA, Keller RA (2004) Radiometric ages for basement rocks from the Emperor Seamounts, ODP Leg 197. Geochemistry Geophysics Geosystems 5:Q08L03. doi:10.1029/2004GC000704 Finlayson DM, Gudmundsson O, Itikarai I, Nishimura Y, Shimamura H (2003) Rabaul Volcano, Papua New Guinea: seismic tomographic imaging of an active caldera. J Volcanol Geotherm Res 124:153– 171 Fisher NH (1939) Geology and Vulcanology of Blanche Bay, and Surrounding Area, New Britain. Territory of New Guinea Geological Bulletin 1 Gill JB (1981) Orogenic andesites and plate tectonics. Springer, Berlin Heming RF (1974) Geology and petrology of Rabaul Caldera, Papua New Guinea. Geol Soc Am Bull 85:1253–1264 Heming RF, Carmichael ISE (1973) High temperature pumice flows from Rabaul Caldera, Papua New Guinea. Contrib Mineral Petrol 38:1– 20 Itikarai I (2008) The 3-D structure and earthquake locations at Rabaul Caldera. Papua New Guinea. Unpublished Master of Philosophy thesis, Australian National University, Canberra Johnson RW, Threlfall NA (1985) Volcano Town – The 1937–43 Rabaul Eruptions. Robert Brown and Associates, Bathurst, Australia Johnson RW, McKee CO, Eggins S, Woodhead J, Arculus RJ, Chappell BW, Sheraton J (1996) The 1994 eruptions at Rabaul Volcano, Papua New Guinea: taking petrologic pathways towards understanding a restless caldera. Eos 76:171 Johnson RW, Itikarai I, Patia H, McKee CO (2010) Volcanic Systems of the Northeastern Gazelle Peninsula. Synopsis, Evaluation, and a Model for Rabaul Volcano. Geoscience Australia, Canberra, Australia, Papua New Guinea Lindley ID (1988) Early Cainozoic stratigraphy and structure of the Gazelle Peninsula, East New Britain: an example of extensional tectonics in the New Britain arc-trench complex. Aust J Earth Sci 35:231–244 Lindley ID (2006) Extensional and vertical tectonics in the New Guinea islands: implications for island arc evolution. Annals of Geophysics, Supplement to Vol. 49, N.1:403–426 McDougall I, Harrison TM (1988) Geochronology and Thermochronology by the 40Ar/39Ar Method. Oxford Monograph on Geology and Geophysics, no. 9, Oxford University Press, New York, 212 pp McKee CO (1993) IAVCEI Canberra 1993 Excursion Guide, Rabaul Caldera—Papua New Guinea. Australian Geological Survey Organization, Record 1993/69, 42pp McKee CO (2015) Tavui Volcano: Neighbour of Rabaul, and likely source of the Middle Holocene penultimate major eruption in the Rabaul area. Bulletin of Volcanology

24

Page 28 of 28

McKee CO, Johnson RW, Lowenstein PL, Riley SJ, Blong RJ, de Saint OP, Talai B (1985) Rabaul Caldera Papua New Guinea: volcanic hazards, surveillance and eruption contingency planning. J Volcanol Geotherm Res 23:195–237 McKee CO, Baillie MG, Reimer PJ (2015) A revised age of AD 667–699 for the latest major eruption at Rabaul. Bull Volcanol 77:65 Nairn IA, Talai B, Wood CP, McKee CO (1989) Rabaul Caldera, Papua New Guinea-1:25,000 Reconnaissance Geological Map and Eruption History. A report prepared for the Ministry of External Relations and Trade, New Zealand Nairn IA, McKee CO, Talai B, Wood CP (1995) Geology and eruptive history of the Rabaul Caldera area, Papua New Guinea. J Volcanol Geotherm Res 69:255–284 Patia H (2003) Petrology and geochemistry of the recent eruption history at Rabaul Caldera, Papua New Guinea: implications for magmatic processes and recurring volcanic activity. Thesis, The Australian Natural University, Canberra, MPhil Renne PR, Swisher CC, Deino AL, Karner DB, Owens TL, DePaolo DJ (1998) Intercalibration of standards, absolute ages and uncertainties in 40Ar/39Ar dating. Chem Geol 145:117–152 Smith IEM, Johnson RW (1981) Contrasting rhyolite suites in the late Cenozoic of Papua New Guinea. J Geophys Res 86:10,257–10,272 Taylor B, Crook KAW, Sinton JM, et al. (1991) Rabaul Caldera, Papua New Guinea, Sea MARC II sidescan sonar imagery, aerial photograph, bathymetry and topography, 1:25,000, Pacific Seafloor Atlas. Hawaii Institute of Geophysics, Honolulu, Sheet 8

Bull Volcanol (2016) 78:24 Tiffin DL, Taylor B, Crook KAW, Sinton J, Frankel E (1986) Surveys in the Solomon Islands and Papua New Guinea using SeaMARC II – Report of the RV MOANA WAVE, 29 November 1985 – 9 January 1986. CCOP-SOPAC Cruise Report 117, South Pacific Applied Geoscience Commission, Suva, 25 pp Tiffin DL, Taylor BD, Tufar W, Itikarai I (1990) A SeaBeam and sampling survey of newly discovered Tavui Caldera near Rabaul, Papua New Guinea. SOPAC Cruise Report 132, South Pacific Applied Geoscience Commission, Suva Tufar W (1990) Sonne – OLGA II Research Cruise, April 29 to June 5, 1990, Preliminary Cruise Report. SOPAC Cruise Report 136, South Pacific Applied Geoscience Commission, Suva Tufar W, Naser G (1992) Lagerstattenkundliche Detailuntersuchungen rezenter Erzvokommen im Bereich einer Platten-Kollisionzone, speziell in der Back-Arc Spreizungszone am Beispiel des ManusBeckens (Papua-Neuguinea). Forschungsfahrt Sonne 68 – OLGA II, Philipps Universitat, Marburg/Lahn, Bundesrepublik Deutschland Walker GPL, Heming RF, Sprod TJ, Walker HR (1981) Latest major eruptions of Rabaul Volcano. In: Johnson RW (ed) Cooke-Ravian Volume of Volcanological Papers, Geological Survey of Papua New Guinea Memoir 10:181–193 Wallace DA, Tufar W (1998) OLGA II Oceanographic Survey of Tavui Caldera, Rabaul. Australian Geological Survey Organisation, Canberra Wood CP, Nairn IA, McKee CO, Talai B (1995) Petrology of Rabaul Caldera, Papua New Guinea. J Volcanol Geotherm Res 69:285–302