Long Island (Pono, Arop, Ahrup) is one of a string of volcanic islands off the NE coast of Papua New Guinea (PNG) stretching all the way SE to New Britian.  It is separated from the PNG mainland by the 52 km wide Vitiaz Strait.  The central caldera is 360 m deep. 

The island is dominated by a pair of steep-sided stratovolcanoes, Mount Reaumur and Cerisy Peak on either end of a 10 x 12.5 km wide caldera and crater lake (Lake Wisdom).  There is a small resurgent volcanic island in the middle of the lake, Motmot Island, that has been the location of several eruptions over the last couple centuries.  The island is roughly hexagonal, 27 x 14 km, 328 km2 in area, with the caldera and 86 km2 Lake Wisdom crater lake in the middle.  The base of the edifice is 60 – 70 km in diameter on the sea floor. 

The most recent of at least three caldera-forming eruptions took place around 1665 AD.  This eruption scoured and buried vegetation on most of the island.  In most tropical settings, plants and other wildlife quickly recolonizes volcanic islands following massive eruptions.  This process is going slowly on Long Island, likely due to a relatively dry environment, near constant trade winds, and unusually porous soils.  Prolonged droughts and dry spells are common.  The island gets 280 cm/yr of rain.  Even so, the island does host a variety of plants and animals, though the plant life has not yet grown into thriving rainforests covering the entire island.

Lake Wisdom was first discovered by Europeans in 1928.  Its water level is some 190 m above surrounding sea level.  There does not appear to be much outflow from the lake into the ocean.  Lake temperatures are relatively constant at 28° C, getting a bit cooler at depth.  There are relatively high oxygen levels in the lake and no apparent hydrothermal activity outside of Motmot Island and a fissure just N of it. 

The island supports a population of 2 – 4,000.  There are nearly 79,000 within 100 km, mostly on the mainland.  The mainland was downwind from Long Island during the 1665 eruption, prompting legends of a Time of Darkness associated with the eruption and ash fall.

The island was first inhabited some 1,000 years ago.  Island residents have a cargo cult creation myth.  They are also quite respectful of the crater lake, seldom approaching it in the belief there is another race of strong humans living around the lake.  Local disappearances are often blamed on these people. 

During WWII, Long Island was used as a barge staging area for Japan.  Allies retook the island starting Dec 1943, eventually setting up a radar station in Apr 1944. 

The island is not monitored in any appreciable form nor are there dedicated webcams.  The closest volcano observatory is at Rabaul, some 580 km E, at the far E end of New Britian. 

Region

PNG is a busy volcanic province.  We had a pair of posts of volcanoes on New Guinea proper, though they are a long way from Long Island.  Over the years, we published multiple posts on volcanoes in and around New Britain including Umboi, 2025, Langila, 2019, Dakataua, 2023, Pago and the Witori Caldera, 2022, Uluwun and Bamus, 2017, Lolobau, 2021, Rabaul, 2019 and Ritter Island, 2015.  Any of these would be great places to start looking at regional volcanic activity. 

Long Island is part of a line of volcanic islands offshore, NE of New Guinea, and W of New Britain.  We will limit the review of regional volcanoes to those within 100 km or so from Long Island.  All distances listed below are drawn from the center of the volcanic crater on each island to the neighboring island center.  We are borrowing liberally from the regional descriptions in our Oct 2025  Umboi post. 

Bagabag Island

Bagabag Island is located 114 WNW from Long Island.  Its closest neighbor is Karkar, another 18 km farther NW.  The island is circular, about 7 km in diameter with an area of 37 km2.  The central volcano is topped with a 3 x 5 km caldera breached to the SE by a flank collapse that created Christmas Bay.  The rest of the island is ringed by narrow coral reefs.  The volcano tops out at 692 m.  The island supports a population of perhaps 3,000.

Yomba (Hankow Reef)

Yomba (Hankow Reef) is located 63 km NW from Long Island.  The history of this island is interesting, as oral tradition from many villages along the Madang coast on NE PNG report residents fled from Yomba 8-10 generations ago as the island was destroyed by a massive volcanic eruption and tsunami.  The actual location of the island is uncertain but suspected to be near the location of present-day Hankow Reef.  The seafloor was surveyed in 2002 locating two distinct seamounts and a caldera.  There was no evidence for the existence of the legendary Yomba volcano. 

A Dec 2014 paper by L Wilson suggests that the Yomba was not a typical oceanic stratovolcano. Rather, it was a low island, built of several tuff cones and pyroclastics, probably less than 200 m above sea level.  A violent phreatic eruption may have destroyed major sections of the island leaving small platform(s) of pyroclastics which were rapidly eroded by wave action.  An island built by Surtseyan eruptions was destroyed by the same type of eruption.  This eruption is thought to have taken place a few decades before the 1665 eruption at Long Island. 

Crown Island

Crown Island is located some 30 km NW from Long Island.  It is a deeply eroded, 3 km wide stratovolcano covered by rain forest.  There is no known historic eruptions or current hydrothermal activity.  There are young tephras on the island, likely from neighboring Long Island.  Possible submarine debris avalanche deposits are N and S of the island. 

Tolokiwa

Tolokiwa is located 53 km E from Long Island.  It is an 8 km diameter circular island with a summit crater breached to the NE.  That crater is largely filled by a more recent cone.  The volcano was built sometime in the last 11 ka.  The largest documented submarine debris avalanche deposit mapped in the Bismark Arc is N of Tolokiwa.   

Umboi

Umboi Island is located 99 km ESE from Long Island.  We previously described this volcanic Island in our Umboi post Oct 2025. 

Sakar

Sakar is located 108 km E from Long Island.  It is a heavily eroded 992 m stratovolcano with a summit crater lake.  Erosion cut deep valleys on the flanks of the volcano.  The island itself is partly surrounded by coral reefs.  The 8 x 10 km island is constructed by two volcanoes.  The older one that forms much of the island is mainly basaltic.  The younger andesitic cone was constructed within the older crater, whose rims are exposed on the N and E sides.  There are no historic eruptions known from Sakar.  The island has an active hydrothermal system with warm springs found along the SW coast. 

There is a pyroclastic cone on the S flank of the island that may have been active within the last 11 ka.  There is a large submarine debris avalanche deposit N of Sakar and another pair W and NW of the island.  Sakar and Umboi are usually lumped together in discussions of flank collapse tsunamis from Ritter Island in 1888. 

Smithsonian GVP carries a Sept 2009 Weekly Report that has the Darwin VAAC reporting a possible diffuse 2 km plume from Sakar on 18 Sept.  The plume drifted 165 km NW and was gone within four hours.  The plume may have been from a fire or hydrothermal activity.  Darwin VAAC later confirmed that an eruption did not occur. 

Ritter Island

Ritter Island is 112 km ESE from Long Island.  Granyia posted on the Ritter Island eruption in 2015.  It was mentioned in a few comments on our previous home in 2012. 

Long Island

The andesitic to basaltic Long Island is dominated by two steep-sided stratovolcanoes on either side of the caldera, 1,280 m Mount Reaumur, the tallest point on the island to the N and 1,112 m Cerisy Peak to the S.  Reamur is the highest point on the island.  Collapse of the volcanic complex created a 10 x 12.5 km caldera now filled by Lake Wisdom.  There is no obvious remnant of the original central volcano.  Post caldera eruptions constructed a small cone, Motmot Island in the S central part of the lake.  Moderate explosive eruptions took place in the 20^th Century from vents at and near Motmot Island.  The area is tectonically active with a M 6.7 earthquake in 2007 took place 54 km S of the volcano.  Another M 5.1 took place Feb 2010 16 km SE.

VOGRIPA carries three massive eruptions from the island in its database.  The largest of these was 19 ka, a VEI 7.4 ejecting 100 km3 of the Kiau ignimbrite.  A VEI 6.3 4.4 ka ejected the 10 km3 creating the Biliau beds along with a caldera collapse.  The most recent eruption here was in 1661, a VEI 6.5 that ejected 30 km3, the Tibito tephra.  The most recent of these eruptions deposited andesitic tephras across New Guinea highlands, prompting mainland legends of a “Time of Darkness.”

Reaumur and Clerisy Peak are ancient basaltic stratovolcanoes.  There are prominent satellite cones on the SW flank of Reaumur and E and NW of Clerisy Peak.  Mount Reamur and Cerisy Peak have gentle lower slopes and steep upper slopes.  Both are heavily eroded with little of the original cones remaining. 

There are deep, steep walled V-shaped valleys down their slopes.  Outside the steep valleys, the cones are covered with shallow gullies down the length of the slopes.  Neither volcano was damaged by caldera formation associated with the 1665 eruption, though Reaumur appears to share a ridge with the caldera rim.  They both appear to predate that eruption.  Their actual age is unknown. 

Most of the coastline is soft pyroclastics cut into cliffs and undergoing active erosion due to wave action.  Large trees hang over the cliff edge.  There are locations with older, more consolidated pyroclastics and lavas that better resist wave action. 

The most extensive landform on the island are gentle slopes falling away from the caldera rim.  These are constructed by pyroclastics erupted during the three known caldera-forming eruptions.  There has been some erosion, though there is no constantly flowing water or streams.  These gentle slopes continue to the coastline cliffs.

The recently active vent and Motmot Island are in the lake, 3 – 4 km from the S shore.  Motmot rises 300 m from the caldera floor.  Perhaps 50 m of it is above the lake surface.  Motmot is a basaltic cone.  As of 1999 its area was 0.25 km3.  The surface is lava, cinder and ash.  The island is either barren or sparsely vegetated.  Motmot Island has been rapidly eroded by wind and waves as it is constructed of soft pyroclastics.  Lava flows on the island have reduced the process.  The most recent eruption in 1993 took place through the lake from a fissure about a kilometer N of Motmot. 

Caldera and Lake Wisdom

Lake Wisdom has a surface area of just over 90 km2.  Its surface is 196 m above sea level.  More than 60% of the lake floor is below sea level.  The deepest point is 173 m, E of Motmot Island.  There were 7-8 eruptions from Motmot and one from a neighboring linear vent system in the lake.  Lake volume is estimated at 19.7 km3. 

The highest point on the caldera rim is 630 m at the N end, likely a ridge of Mount Reaumur.  Most of the caldera rim is 375 – 500 m.  Caldera volume is estimated 41 – 51 km3.  30 – 48% of the caldera volume is filled with water.  Surface area of the caldera was probably smaller following the 1665 eruption.  Eroding caldera walls have filled the original volume with sediment.  There is no significant outlet to the sea.  Motmot Island and other resurgent activity also contributed to refilling the caldera. 

The caldera is thought to have grown in stages over the course of the three caldera-forming eruptions rather than being the product of the most recent one.   The embayment on the W side of Lake Wisdom probably subsided before the rest of the caldera.  Some of the terraces suggest a higher lake level which may have been destroyed at other parts of the lake by more recent collapses.  Caldera walls in the embayment, particularly in the S end are much more eroded with gentler slopes, suggesting subsidence took place when the water level was lower.  The bottom of the W embayment is a short shelf, showing two stages of subsidence in its formation. 

There are three separate landforms around Lake Wisdom.  The first of these is a cliff surrounding most of the lake.  Most of the cliff is a steep face with occasional talus slopes at the bottom entering the lake.  There are 5 terrace remnants at the W embayment 10 – 15 m above the lake surface.  The E side of the lake has a broad sloping area and amphitheater sloping S toward the lake thought to be a partial collapse of the caldera wall.

On the N side of the lake, near vertical cliffs rise 200 m above the water.  The caldera wall around the W embayment has a well-developed stream network.  The floor of this bay is much shallower than the rest of the caldera.  There is a terrace 10 – 15 m above present lake level.  This bay is likely older than the rest of the caldera, meaning there were multiple phases of caldera collapse.

There is a 3 km long section of the E margin of the caldera with a 400 m wide breach.  This breach allows some drainage from the subsided section to the sea.  Other parts of the drainage drain into the caldera.  The interior caldera wall is eroded with a mix of steep V-shaped valleys and gullies sloping into the caldera from the rim.  No part of the original surface remains.

Pyroclastics

There are four main pyroclastic sequences found on the island.  The lowest exposed parts of the volcano are found in the caldera walls.  Most of the middle and upper sections are inaccessible due to the steepness of the walls.  While lava flows are present in some locations, they are generally rare.  The initial series of eruptions including some quiet periods deposited the Sauro Beds.  This was followed by three catastrophic caldera-forming eruptions.  The events were separated by a period of relative quiet with small, intermittent tephra falls.  The three Plinian eruptions all took place after activity at Reaumur and Clerisy stopped. 

The lower sequence, the Sauro Beds, are generally airfall tephras.  The upper three sequences produced by the caldera-forming eruptions are similar.  Each has a base airfall covered by ignimbrites which are finally covered by another airfall layer.  The pyroclastic layers are juvenile magma.  Degree of welding has yet to be determined, though not much has been observed. 

The Sauro Beds are a group of at least 12 bedded red airfall tephras that cover lavas near Mount Reaumur.  They are covered by either other lavas or tephras at the bottom of the Kiau Beds.  Individual Sauro beds are up to 30 cm thick, though some are thicker.  They are heavily weathered and only found near Mount Reaumur on the N part of the island.

There are three identified major eruptions from the caldera.  The first of these created the Kiau Beds, a thick sequence of ash and lapilli covered by a firm ignimbrite.  Carbon dating of burned wood at the base gives 16.0 ka as the eruption date.  The Kiau Beds are chemically similar to the widespread Ep Tephra on the mainland highlands.  Two dates of this tephra near Kuk (Mt Hagen) are between 15 – 12 ka. 

The Kiau Beds are covered by the Biliau Beds.  Bilau has a similar sequence of an early airfall covered by a 1 – 8 m later layer of pyroclastics.  Once again, charred wood at the base of the deposit was used to date the eruption at 4.0 ka.  The Bilau Beds on Long Island are believed to correlate with the widespread Kim Tephra on the PNG mainland highlands.  This tephra dates 3.8 – 3.5 ka. 

The most recent caldera eruption on Long Island around 1650 AD deposited the Matapun Beds.  This covers a palosol on top of the Biliau Beds.  The lower layer is up to 2 m of coarse ash and lapilli airfall.  The middle layer is unwelded, poorly bedded pyroclastic flows of lapilli and blocks in a sandy matrix.  The pyroclastics are covered by up to 1.5 m of airfall ash and lapilli.  On the mainland, this eruption deposited the Tibito Tephra, mapped across 94,000 km2 with a volume over 10 km3.  Bulk tephra volume was estimated at 21 km3 with a 28 km plume by Blong and Magill in 2018.  Most of the island’s coastline and more than half of the interior are covered by the Matapun Beds.  These beds extended the island’s area for a while before it was eroded by wave action. 

The relative size of the three eruptions is not well known, but the Ep Tephra appears to be the thickest of the three tephra layers on the mainland.  The Kim Tephra may be the thinnest and least widespread.  All three eruptions were larger than VEI 6.3.  It is likely that all three eruptions helped created the current caldera volume of at least 40 – 50 km3.  Subsequent resurgent activity between major eruptions and following the 1650 AD event built new cones and refilled some portion of the caldera.  It is not possible to tie caldera size to the volume ejected with each major eruption.

Eruptions

While the Smithsonian GVP carries reported eruptions in 1933, 1938 and 1943, the earliest detailed report of an eruption was 1953 – 1954, some 4 km from the S shore of the lake.  These marked the emergence of Motmot Island above the lake surface, two cones connected by a small ridge.  After a decade or so, eruptions resumed Mar 1968 on the shore of the new island.  The eruptions built an ash cone.  By the end of the eruption sequence Nov 1969, the crater was completely formed along with a new crater lake in it.  Note that there is an unconfirmed report of a small eruption in 1961. 

Smithsonian GVP carries 20 years of Bulletin Reports 1973 – 1993.  The first of these May 1973 reports a small eruption from Motmot Island within the crater lake 18 Apr.  There was increased fumarole activity before the eruption.  It ejected ash and steam and produced a small lava flow.  Considerable new fumarole activity at Motmot was reported Jan 1979.  There was no new ejecta observed.  Follow-up airborne inspection in Feb confirmed increased vapor emissions only.  Another airborne inspection Sept 1990 observed vegetation spreading in some areas in the crater of the 1974 cone.

The most recent eruption took place Nov 1993.  First indication of new activity was a change of color of Wisdom Lake.  Source of the color change did not appear to be Motmot, but a new vent a couple kilometers farther N into the lake.  Permanent seismograph station at Karkar Island, 160 km NW recorded unusually strong, continuous tremor from Long Island starting 3 Nov.  The eruption ejected steam and ash from the eruption location under the lake. 

Direct visual observation of water and ash ejection from the lake were made on 5 Nov.  Ground-based inspections 11 Nov heard muffled thudding noises and felt earthquakes.  A seismograph set up at the time recorded volcanic tremor.  Most of the activity took place 11 – 16 Nov and started tailing off by 17 Nov.  The eruption that began under the caldera lake, N of Motmot Island ended by Dec.  Plumes of discolored water continued after ash ejection stopped.  

Dating the most recent caldera-forming eruption

Accurate dating of the most recent eruption has been something of an adventure, narrowing to 1665 – 1668 AD from initial estimates ranging between late 16^th Century to early 20^th Century.  There was even a line in a mid-19^th Century Victorian novel that influenced a proposed date of 1665 – 1668.  Dating this eruption relies on inspection of tephras on the mainland, revegetation, and ice core sampling. 

The story begins with early 1970s archeological investigations at the Kuk Swamp on the mainland.  This work documented a number of thin tephras. The youngest of these was the 10 cm Tibito Tephra, dated in 1971 at around 1810 AD.  This widespread tephra was eventually traced back to Long Island, about 300 km from the Kuk site.  More refined estimates on the eruption based on oral tradition of a time of darkness reported at more than 40 sites and at least 26 language groups.  The island was first visited in 1700 providing more observations.  Blong in 1982 proposed 30 years for revegetation of the island based on observations in 1700 and concluded the eruption took place 1630 – 1670 AD.  This is the current estimate of the eruption used by Smithsonian GVP.

The GVP estimate is not without controversy.  There is significant variation between calendar years and radiocarbon years, as use of these curves was relatively new in the early 1980s.  Worse, some oral traditions reported the eruption in the early 20^th Century, which was not helpful.  Various publications citing tree ring and ice core records listed the eruption as one of several unknown eruptions, usually dated 1640 – 1690 AD. 

The location of Long Island, the large size and ejected volume of the eruption, and estimated sulfur content of the magma confirmed that the eruption should have been recorded in Greenland and Antarctica ice cores 1650 – 1675 AD.  Final analysis has narrowed the date to 1665 – 1668 AD.

The second piece of this was revegetation of Long Island following the eruption.  The 1700 visit described the island as covered by mixed savannah and woodland. The question becomes how quickly can a tropical island become green again?  Fortunately, we have an example with Krakatau, which was closely watched following the 1883 eruption.  The island’s location and eruption size are similar to Long Island.  Descriptions of greenery on Krakatau and Long Island should be similar at 39 years after the VEI 6 eruptions, though Long Island revegetation is lagging. 

The mention of unusual sunsets in London is even more odd.  Ainsworth wrote in 1841 about unusual sunsets in 1665 that coincided with an outbreak of the Plague.  Turns out that he picked this up from a line in a Defoe novel written in 1722.  Unusual sky phenomenon was not reported in any publications discussing the plague in London in 1665 or in any historical climate record from the US since 1640. 

Tectonics

Regional tectonics for Long Island are similar to Umboi some 99 km SE.  I will reprise the description of regional tectonics from that post below. 

At its most basic, New Britain is a subduction-driven region.  There are numerous microplates jostling between the ongoing collision between the Indo – Australian and Pacific Plates.  New Britain resides on the South Bismarck Plate.  To the north, is the North Bismarck Plate, which moves generally to the west.  The north boundary of the North Bismarck Plate is generally subduction driven with the Caroline and Pacific Plates subducting under it.

The western portion of the north boundary of the South Bismarck Plate is generally moving in a strike – slip fashion (similar to the motion along the San Andreas Fault system).  The eastern portion is a series of spreading centers.

The southern boundary of the South Bismarck Plate is subduction, with the Solomon Sea Plate / Woodlark Plate subducting under it.  It is this subduction that drives volcanic activity in New Britain.  There is some dispute about the extent of the Woodlake Plate.  Some depictions of it show its northern portion (currently rifting) as the Solomon Sea Plate.

The area is tectonically complex, with recent formation and movement of subduction zones, rifts, spreading centers.  Due to the forces of large plate impact (Indo – Australian and Pacific), activity along the boundary of microplates defining the impact zone have changed significantly in a very short period of geologic time.

As with any tectonically complex volcanic region, this region is also subject to large earthquakes in the M 7 – 8 range.

Conclusions

Long Island is a sparsely inhabited volcanic island with resurgent, post caldera activity.  It has multiple eruptions over the last century both from the resurgent island and the caldera floor.  It should be considered an active, dangerous system, not unlike most of its neighboring volcanic islands off the NE coastline of PNG. 

Additional information

Colonization of an island volcano, Long Island, Papua New Guinea, and an emergent island, Motmot in its caldera lake.  I.  General Introduction. Harrison, et al, Nov 2001

Long Island, Papua New Guinea:  Introduction, Specht, et al, 1982

Colonization of an island volcano, Long Island Papua New Guinea, and an emergent island, Motmot, in its caldera lake.  II. The vascular flora, Harrison, et al, 2001

Long island (Papua New Guinea), Wikiwand

Steps and missteps on the path to a 1665 – 1668 CE date for the VEI 6 eruption of Long Island, Papua New Guinea. Blong & Kubatov, Apr 2020

Colonization of an island volcano, Long Island Papua New Guinea, and an emergent island, Motmot, in its caldera lake.  VII. Overview and discussion, Harrison, et al, 2001

A Lapita presence on Arop / Long Island, Vitiaz Strait, Papua New Guinea?,  Gaffney, et al, terra australis 52

Significance and timing of the mid-17^th-century eruption of Long Island, Papua New Guinea, Blong, et al, Nov 2017

Pyroclastic deposits and eruptive sequences of Long Island, Pain, et al, 1981

Preservation of thin tephra, Blong, et al, Jun 2017

A summary of precursors to volcanic eruptions in Papua New Guinea, Mori, et al, Feb 1986

Long Island, Papua New Guinea:  aspects of landforms and tephrostratigraphy, Blong, et al, Jul 1982

Long Island, Papua New Guinea – aspects of the prehistory, Egloff & Specht, 1982