New Zealands volcanoes: Kermadec Islands

By John H. Latter, Edwards F. Lloyd , Ian E.M. Smith , and Simon Nathan

This booklet is concerned with volcanic hazards which may arise in the event of eruptions of volcanoes in the Kermadec Islands, at a submarine volcano midway between the Kermadecs and Tonga, and at several submarine volcanoes between the Kermadecs and the North Island of New Zealand. Although not part of New Zealand landmass, these volcanoes are near neighbours and their future eruptions may well have an impact on our environment.

All these volcanoes lie along the zone of collision of two of the major structural plates of which the Earth’s crust is composed, the Pacific plate to the east, and the Australian plate to the west (see Inside Front Cover). The line of collision is marked by the succession of deep ocean trenches which characterise this part of the south-west pacific, the Tonga Trench in the north, the Kermadec Trench in the centre, and the Hikurangi Trench in the south, off the coast of the North Island of New Zealand (Figure 1). Along this line, the Pacific plate is being forced beneath the overriding Australian plate as a result of the relentless pressure of convergence of these two huge regions of the Earth’s surface. The rate at which this process is taking place is relatively high, about 7cm per year on average (although it is not yet known whether the process is continuous, or whether it proceeds as a series of sudden jumps, perhaps associated with large earthquakes). Earthquakes occur on the upper surface of the Pacific plate down to depths of more than 600 km, as it is forced down at a steep angle below the overriding Australian plate, which is itself buckled upwards by the pressure of collision forming structural highs, two of which are the Tonga and Kermadec Ridges. The line of volcanoes formed when rock melted by the heat of collision, over a range of depths at which the increasing temperature outweighed the effect of increased pressure (which tends to prevent melting), then rose under its own buoyancy through denser overlying rock to intrude the upper crust. The primary melting process, shown by earthquakes taking place in the roots of the volcanoes, occurs at depths of about 100 km below the surface.

Image above: Raoul Island – looking north-west across Raoul Caldera from Moumoukai Peak, with Blue Lake on the right, and Green Lake on the left, and the Weather Station on Fleetwood Bluff in the right background. Photograph by B.D. Scott

Front Inside Cover (above): Raoul Island – 1964 Crater No 1, looking north-west, with part of Green Lake in the foreground. Photograph by E.F. Lloyd.

Front inside cover (above): Diagram showing the Pacific Plate underthrusting the Australian plate beneath Raoul Island. Cross-section along the line AB marked on Figure 1, at right angles to the line of the Kermadec Trench.


The Kermadec Islands, 750 to 1000 km north-north-east of New Zealand, are mainly of volcanic origin. They are uninhabited, except for a weather station manned by a handful of people on Raoul Island (previously known as Sunday Island), the largest and most northerly island in the group (Front Cover). All the islands are scientific reserves for the protection of fauna and flora. They have a moist subtropical climate, and as a result vegetation normally recovers quickly (in tens of years) after being damaged by volcanic activity. However, a very large eruption could well destroy the seed source, and regrowth of vegetation would then be very slow. The islands are administratively part of New Zealand, and any risks that result from eruptions there are accordingly the Government’s responsibility to assess and reduce, if possible.

 

Figure 1: Map showing location of volcanoes and principal features of sea floor topography in the region from North Island, New Zealand, to southern Tonga (from General Bathymetric Chart of the Oceans, and New Zealand Oceanographic Institute data). Bathymetry in metres. Filled triangles are recently active or potentially active volcanoes, and unfilled triangles possible recently active volcanoes.

1. Falcon Island (Tonga)

2. Submarine volcano south-east of Honga Hapai (Tonga)

3. Submarine volcano north-west of Tongatapu (Tonga)

Situated at such a distance from permanently inhabited parts of New Zealand, it may seem unnecessary to consider risks from eruptions in the Kermadecs at all, but the islands lie close to air routes linking New Zealand with Tonga, Niue and Western Samoa, as well as the western United States, and shipping lanes used by hundreds of vessels every year. Besides, at least one Kermadec volcano is capable of producing great eruptions which could generate tsunami (wrongly called tidal waves, since they have nothing to do with the tide) which are liable to cause devastation on northern New Zealand coastlines as well as in other parts of the Pacific.

The Monowai Submarine Volcano (Figure 1), about halfway between Raoul Island and the southernmost island of the Tonga Group (Ata Island), lies just beyond both New Zealand’s and Tonga’s Exclusive Economic Zones. It is a submarine peak that rises to within 120 metres of the surface. Should it ever emerge about the surface, it would be an interesting problem in international relations as to which country could legitimately claim it! When it erupts, it is likely to prove dangerous to ships in the immediate area, and, in the case of a large eruption, also to overflying aircraft: there is, in addition, some risk of tsunami originating at this volcano, which is discussed below.

South of the Kermadecs, some 230 to 330 km north-east of the Bay of Plenty coast of the North Island of New Zealand (Figures 1, 2), is a group of submarine volcanoes known as the Rumbles, or more prosaically as the South Kermadec Ridge Seamounts.

Figure 2. Bathymetry (in metres) between the south-west end of the Kermadec Ridge and North Island, New Zealand. Filled and unfilled triangles as in Figure 1. RI-IV = Rumble I-IV.

These were first identified as volcanoes in 1963, by staff of the Naval Research Laboratory at Devonport, in Auckland. Since 1958, the laboratory had operated an outstation on the east coast of Great Barrier Island, studying, among other things, ambient sea noise. Unusual signals kept being recorded, and were informally christened BUN (for Barrier Underwater Noise) by the laboratory staff. Eventually the source of BUN was traced to noises, presumably of volcanic origin, at one of the Rumble seamounts (Rumble III). Since then, an eruption in late 1986 at Rumble III was observed both from a boat at close range and from the air. The Rumble volcanoes pose dangers mainly to shipping in the near vicinity, but, in the event of large eruptions, also to aircraft; they might also give rise to tsunami, which could affect nearby New Zealand coastlines. Three other large volcanic seamounts have recently been discovered by the New Zealand Oceanographic Institute east, south and south-west of Rumble IV (Figure 2).

Closer to the New Zealand coast, oceanographic studies have led to the discovery of many seamounts, most of which have been identified as submarine volcanoes of various ages. The principal ones are Whakatane Seamount, about 80 km north-north-east of White Island, and a group of small submarine cones at the south-west end of Colville Ridge (Figures 1, 2), about 30 to 50 km north-east of Colville Knoll. A single dated rock at Whakatane Seamount gives an age of about 0.7 million years. The other cones have not yet been dated. Basaltic lavas of very youthful appearance, with little or no sediment cover, have recently been dredged from the eastern slope of the Ngatoro Basin (Figures 1, 2). This may be a rift structure, and is certainly very different to the tectonic environment in which the principal seamounts to the east occur. It is likely that eruptions have taken place quite recently in this area.

Because of their distance from the New Zealand mainland, future eruptions from Monowai or the Kermadec volcanoes, or from the Rumble and other submarine volcanoes closer to New Zealand, will only constitute a potential danger to life and property in New Zealand if they are exceptionally large. In this case, hazards are likely to be caused by the distribution of ash over a wide area and the development of tsunami associated with powerful explosions. However, there are likely to be severe risks to the few residents of Raoul Island, and to aircraft and shipping near the volcanoes, as a result of much smaller eruptions.

The most likely types of volcanic eruptions, and the dangers associated with them, are described below.

Lava flows are the hot streams of molten rock that flow down the flanks of a volcano. The distance a lava flow travels depends mainly on the viscosity of the lava (ease with which it flows), but is also affected by the volume of lava erupted, and by the exact shape of the landscape over which it is extruded.

All the volcanoes considered in this booklet have erupted lava during past eruptions and are capable of producing lava during future activity. On the islands, lava flows are likely to be moderately thick and slowly moving. They should pose little risk to movable property or to human life. However, where they enter the sea, the very hot molten rock coming into sudden contact with cold seawater may cause steam explosions, which will be very dangerous at close range.

Eruptions of lava in deep water are probably moderately common along the Kermadec Ridge. Historical submarine eruptions at Monowai Volcano, at a submarine vent north-east of Raoul Island, and at Rumble III Volcano (see below) were probably all the result of small lava flows or domes being erupted quietly onto the sea floor, with release of gas at the surface. They show clearly that risk due to small submarine eruptions is negligible. However, in the case of a sudden large eruption, explosions could endanger aircraft and shipping, and could give rise to tsunami. Now eruptions of lava in shallow water are known to have occurred recently in the region.

Explosive eruptions from volcanoes around the rim of the Pacific are typically accompanied by flows of rapidly-expanding gas and solid particles which travel as a fluidised mass. Such flows are hot (up to 1000^oC) and move at speeds of several hundred kilometres per hour. They are called pyroclastic flows, and are also sometimes described as glowing clouds or burning clouds. Naturally they are extremely dangerous. They can travel uphill or across areas of irregular relief, but usually tend to flow along river courses and into depressions. In some cases, pyroclastic flows may flow along the surface of a lake or the ocean. A major source of hazard is heated air moving in front of a pyroclastic flow which can sear vegetation and kill rapidly by heat suffocation.

Large pyroclastic flows form deposits called ignimbrites. When the rate of emission of the erupting magma and its temperature are both sufficiently high, the deposit accumulates in a semi-molten, plastic condition, and individual particles and layers fuse together, forming a rock known as welded ignimbrite, which looks like concrete. At lower rates of emission, or lower temperatures, ignimbrites do not weld, but accumulate as masses of unsorted blocks and pumice.

A lateral blast is a more or less horizontally-directed cloud of gases and very hot rock particles that explodes outwards at very high speeds from the site of an eruption. Velocities of more than 300 km/h have been estimated for the lateral blast that occurred at the climax of the 1980 Mt St Helens eruption in the United States. Lateral blasts are extremely unpredictable and can develop without warning during any major explosive eruption.

There are likely to be pyroclastic flows and lateral blasts in future eruptions of Raoul, Macauley and Curtis Islands. Because the islands are small, their effects will be localised, but offshore areas near the islands are also at risk. The principal risks are due to their extreme heat and rapid velocity. Pyroclastic flows have accompanied large eruptions from Raoul Island during the past 4000 years, and they represent the principal danger to people and property at the Weather Station. Pyroclastic flows also happened during a very large eruption at Macauley Island during the last few thousand years.

The term tephra encompasses all the products of a volcanic eruption that are ejected through the air during explosive eruptions (except the products of pyroclastic flows and lateral blasts which stay close to the ground surface). Tephra is subdivided into three main particle sizes, ash (less than 2 mm diameter), lapilli (2-64 mm diameter), and blocks or bombs (greater than 64 mm diameter). Consolidated ash is also known as tuff. Scoria is dark-coloured tephra, which has been erupted hot but not in a molted condition. It is permeated by holes from which gas has escaped, and in colour and texture resembles cinders or iron slag from a furnace. Pumice is a light-coloured equivalent. Tephra is usually dispersed widely during an eruption. The main factors that control how far it extends are the height of the eruption column (which depends on the violence of the eruption), and the strength and direction of the prevailing wind at various altitudes.

The dangers presented by tephra may be considered as (1) the problem created by its physical presence, and (2) the fact that potentially harmful substances, such as fluorine or chlorine, which may poison the environment, can adhere to the tephra particles. The effects are very dependant on the size of the particles and the way in which the tephra is distributed.

The effects of tephra on human beings, animals, transportation systems, machinery and the infrastructure of cities and industry are well known. In general these have little relevance to the volcanoes of the Kermadec Ridge. However, there are two specific cases of severe hazard that do apply in this region. The first concerns shipping and the second aircraft.

Shipping lanes between New Zealand and Australia and many of the Pacific Island nations, as well as the Americas, cross the Kermadec Ridge. A major explosive eruption, either from one of the island volcanoes or from the submarine volcanoes on the Ridge, could create a tephra cloud and probably a floating raft of pumice, which would obscure visibility, damage shipboard machinery, and create a health hazard for crews. A prolonged eruption would greatly disrupt shipping in the region north and north-east of New Zealand.

An even greater risk is that posed by a widely-dispersed ash cloud to aircraft. Important air routes to the north and north-east of New Zealand lie along or across the Kermadec Ridge. A number of recent incidents in which aircraft have encountered eruption clouds have shown that jet engines are extremely vulnerable to fine ash particles abrading moving parts, and in extreme cases all power has been lost (although fortunately it has so far always been possible to start the engines up again).

Historical eruptions from the submarine volcanoes Monowai and Rumble III have not produced much air-borne tephra, although both Monowai and the unnamed vent off the north-east coast of Raoul Island, mentioned earlier, have given rise to extensive rafts of pumice which could have caused a problem for shipping. During the short sequence of explosions from Raoul Island in 1964 (see below), a steam and ash column rose to an altitude of about 1.2 km, but it dispersed quickly and would not have been a significant risk to either shipping or aircraft. A greater volume of tephra was erupted at Raoul Island in 1814 and 1870, when cinder cones built up in shallower water in Denham Bay: however, the hazardous area was limited to one or two kilometres from the vent. All recent eruptions on land in the Kermadecs have produced tephra. Future hazards will depend on the volume erupted and on the wind direction. For Raoul Island, this is discussed in more detail below.

The possibility of grave danger will arise if any of the Kermadec Ridge volcanoes enters into a major eruption. For most of the submarine volcanoes, this would probably require some initial activity to build the vent area up to sea level, but the summits of both Monowai and Rumble III Volcanoes are within 200 metres of the surface at the present time. Dangers due to tephra are much greater for large eruptions in very shallow water or on land than for those in moderately deep water.

Tsunami are sea waves caused by disturbances of the ocean floor. A large displacement of water is needed to produce such a wave. In the open ocean the height of a tsunami is generally less than one metre, with a period (the time between the crest of successive waves) of between five minutes and one hour. The speed of tsunami waves depends on the depth of the water, and in deep water is as much as 800 km/h.

Tsunami originating at volcanoes may result from collapse of a submarine cone, from a powerful submarine explosion, or from the violent impact of pyroclastic flows or debris avalanches into water (e.g. at Krakatoa in 1883). Such waves spread out in all directions from their source. As they approach land, the height of the waves increases rapidly.

The damage caused by tsunami is due to flooding and the initial force of the wave. Tsunami rank highly as a cause of death and destruction in historical eruptions, and must be viewed as a serious potential threat to northern coastal areas of New Zealand in the event of a major eruption from one of the Kermadec Ridge volcanoes. A serious problem is the likely short warning time (1-2 hours) between initiation of a tsunami and its arrival on New Zealand coasts.

Each potentially active volcano in the region is described briefly below, in order from north-east to south-west, with a summary of its history of eruptions insofar as this is know.

Also known as Orion Submarine Volcano, Monowai Submarine Volcano lies just west of the Tonga Ridge, where the ridge falls at its southern end to below 1000 metres from the surface. The volcano is closely aligned with the Tongan volcanoes to the north, the southernmost of which are shown on Figure 1. It lies off the trend of the Kermadec volcanoes, and is structurally part of the Tonga Ridge rather than of the Kermadec Ridge. A shoal was first reported in this area in 1944, but it was not until 1977 that it was recognised as a volcano. On October 17 of that year the crew of an RNZAF Orion aircraft saw a discoloured patch of water some 5 km long, by 100 to 200 metres wide, with gas bubbles rising to the surface over an area of about 100 metres in diameter (Figure 3). Sonobuoys, dropped close to the bubbling area, picked up pulsating rumbling noises, thought to be due to the violent discharge of the gas bubbles. Similar activity was seen three days later from another RNZAF Orion aircraft, but had died down by the next visit on October 27. The position of the upwelling was 25^o55’S, 177^o14’W.

Figure 3. (a, b). Submarine eruption at Monowai Volcano in October 1977 (aerial photographs by RNZAF aircrews). The diameter of the main area of upwelling, seen in (b), is about 100 metres (see paper by F.J. Davey in Supplementary Reading).

During 1978 and 1979 both HMNZS Tui and HMNZS Monowai carried out several surveys of water depth and noise over the volcano, the most detailed of which was by Monowai in September 1978 (Figure 4). The volcano was found to be conical, rising from a depth of about 1500 metres to a peak of about 120 metres below surface, at 25^o53.2’S, 177^o11.3’W. At a depth of 1000 metres, the cone is about 10 km across, in a north-north-west direction through the summit, by about 7 km at right angles to this direction. No dredge samples have yet been recovered from the cone, and it is therefore not known what types of volcanic rock are erupted at Monowai Volcano.

Minor eruptions at Monowai Volcano probably take the form of quiet outflowing of lava on the sea floor, and some turbulent gas release at the surface, and are of no risk to anyone. In the case of a sudden large eruption, however, explosions could endanger aircraft or ships close to the vent. A more widespread hazard would result from the collapse of a large part of the underwater cone. This would probably give rise to moderately large tsunami, which could affect many parts of the south-west Pacific. Collapse of this kind could readily take place due to overloading of the cone by lava eruptions at the summit or high on the flanks, which might cause the steeper parts of the structure to give way. If the 1944 report of a shoal was correct, collapse of the summit must indeed have taken place since the peak is now about 120 metres below surface. However, from the shape of the cone in Figure 4, it does not look as if a large collapse has occurred recently. Hence it is likely that the shoal reported in 1944 was, in fact, either a pumice raft, or a disturbance of the water due to gas discharge.

The Kermadec Islands are the summits of large volcanoes that have been built up on the crest of the Kermadec Ridge, and have emerged above sea level. The ridge stretches for 600 to 700 km north-north-east from about 33½^oS 180^o to about 28^oS 178^oW (Figure 1), and has been upraised, as already mentioned, by the ongoing collision between the Pacific and Australian plates.

The northernmost known volcanic centre in the Kermadecs is Raoul Island, which is the summit of a large submerged massif about 35 km by 20 km, with its long axis aligned north-east, slightly inclined to the overall trend of the Kermadec Ridge. It is possible, however, and indeed likely, that there are submarine volcanoes, as yet undiscovered, further north along the ridge, which rises in several places to within 500 metres of the surface. Raoul Island itself, the largest of the Kermadec Islands, is an anvil-shaped island, about 30 km² in area, with a maximum length of about 10 km in an east-west direction along the north coast, and about 6 km from north to south (Figure 5). The island has undergone many changes as a result of eruptions during the past few thousand years, and as a result of earlier eruptions. It now contains a large central depression, a little over 3 km east-west by about 2 km north-south, formed largely by subsidence immediately after large eruptions. This type of structure, resembling a very large volcanic crater, is known as a caldera. Just west of Raoul Caldera, and almost touching it at one point, is Denham Bay, a second caldera (also a little over 3 km long, but in a north-south direction, by rather less than 3 km wide), which has been flooded by the sea.

Figure 5. Raoul Island vent areas for eruptions during the last 4000 years.

The geology of Raoul Island has been well studied, especially over the northern half of the island, which was mapped in detail by two of the present authors (E.F.L. and S.N.: see Supplementary Reading). In the southern part, access is difficult and more work needs to be done, especially on the distribution of rocks erupted during the last few thousand years, before the detailed history of eruptions at Raoul can be fully understood. The earliest rocks exposed on Raoul Island, and on some of the small Herald Islands to the north-east, were formed as a result of submarine eruptions. They have been roughly dated at between half a million and about one and a half million years old. The rocks are basaltic andesites, a type common in island arc situations worldwide. Eruptions of basaltic andesite are generally not violently explosive; lava flows are often produced, together with rhythmic ejections of hot lava fragments and "spatter", in a style of eruption known as Strombolian, after the Italian island of Stromboli.

At some stage during the last half million years, the volcano emerged above sea level and a large stratovolcano (a composite cone, made of alternating lava flows, and tephra beds) was built up. A very powerful eruption led to the collapse of this structure, and a single large caldera formed, which probably included the sites of the present Raoul and Denham Bay Calderas. Between 5000 and 10,000 years ago, activity in the eastern part of this caldera built up a large cone of basalt and basaltic andesite in a series of Strombolian eruptions. This filled and extended beyond the area now occupied by Raoul Caldera: it has been named Moumoukai Volcano.

About 2200 B.C., a major change took place in the type of magma (molten rock) erupted at Raoul Island. Previously most of the material erupted was basalt and basaltic andesite. From this time onwards, however, almost all the magma erupted was dacite, which is a more viscous rock than andesite (that is to say, it flows less easily). Consequently gas, which is the driving force of all eruptions, is much less able to escape from dacite magma than from andesite, and as a result high gas pressures build up within the magma, which are released in very powerful explosions. Hence dacite is usually erupted as pumice, which is rock from which the trapped gases have escaped violently through the myriad small holes with which the rock is perforated. Fortunately rhyolite, which is still more viscous than dacite (and hence gives rise to eruptions, such as those of Taupo in New Zealand, which are even more explosive an dangerous than dacite eruptions), is not found, except in small quantity, in the Kermadec Islands. Although dacite pumice has been the normal product of magmatic eruptions at Raoul Island since about 2200 B.C., in one eruption about 1700 years ago the magma was exclusively basaltic. Many eruptions have also occurred which have not ejected any fresh magma. These so-called phreatic (or water-driven) explosions throw out only old (previously-cooled) rock, as a result of sudden violent interaction between hot rock or gas and groundwater.

The volcanic history of Raoul during the last 4000 years has been extensively studied. In addition to three historical eruptions, which were seen taking place in 1814, 1870 and 1964, geological studies have established the approximate dates, volumes and vent areas for a further 12 eruptions, as listed in Table 1. These vent areas are also marked on Figure 5. Approximate sizes (volumes) are given for the eruptions. These "erupted volume magnitudes" (EVM) form a logarithmic scale, like Richter earthquake magnitudes. Each successive number on the scale represents ten times the volume of the previous number, and each number is the most significant (first digit) of the logarithm of the erupted volume in cubic metres. Thus magnitude 6 represents 1 followed by six zeroes, or 1,000,000 cubic metres (10⁶m³), and is ten times the volume of magnitude 5, which represents 100,000 cubic metres (10⁵m³).

Table 1. Volcanic eruptions at Raoul Island during the last 4000 years (in order from youngest to oldest).

Note: Rock formations named tephra or pumice were produced by eruption of fresh magma, and those marked breccia by phreatic eruptions (see text) of old (previously-cooled) rock.

* Erupted volume magnitude: see text.

** In addition, fresh magma was erupted from Denham Bay Caldera during these eruptions, but has not been found on land.

*** Basalt: all other eruptions of tephra and pumice were of dacite.

The present Raoul Caldera began to form with the eruption of Matatirohia Tephra, which initiated collapse of Moumoukai Volcano. Denham Bay Caldera was formed as a result of the Fleetwood Tephra eruption (EVM 10), the largest eruption to have taken place at Raoul during the last 4000 years.

For comparison, the equivalent EVM magnitudes for some well-known eruptions are as follows:

Vulcan, Rabaul, Papua New Guinea June 1937 8
Mount St Helens, United States May 1980 9
Tarawera, New Zealand June 1886 9
Krakatoa, Indonesia August 1883 10
Taupo, New Zealand About 1800 years ago 11
Lake Toba, Indonesia About 75,000 years ago 12

Especially in the case of larger eruptions at Raoul Island, with magnitudes of EVM 8 and above, a great deal of the erupted rock was emplaced as horizontally-directed ground-hugging pyroclastic flows, although much was also emplaced as airfall tephra. Pyroclastic flows are of two kinds, one in which most of the gas comes directly from the magma (these are pyroclastic flows in the strict sense, and when the volume is very large the resultant rocks, as mentioned above, are often called ignimbrites), and the other in which the gas comes mainly from water (these are called pyroclastic surges, or base surges). There is usually a component of base surges in all eruptions. Base surge deposits, for example, have been clearly identified in the 21 November 1964 eruption deposits at Raoul Island.

Pyroclastic flows and surges represent the principal danger to people and property at the Weather Station on Raoul Island. It is practically impossible to predict specific areas which could be overwhelmed by pyroclastic flows and surges, as the exact position and inclination of the vent, in relation to surrounding topography, is critical. In both the Oneraki Tephra and Fleetwood Tephra eruptions, high-energy pyroclastic surges overtopped hills several hundred metres high.

Although pyroclastic flow deposits cannot be accurately predicted, thickness of airfall tephra deposits can be estimated fairly well if wind speeds and directions are known for the various ranges of attitude which eruption clouds of various magnitudes are able to reach.

Estimates of the average fallout pattern of tephra are given in Figure 6 for the commonest wind speeds and directions at Raoul Island, for short-duration eruptions of EVM magnitude 5 to 9. Note that eruptions of magnitude 5 are hypothetical at Raoul, since such small eruptions have neither been observed historically nor detected in the geological record. If they occur at all, it may well be that they take place as part of an overall eruption episode of greater volume. Eruptions are shown arbitrarily centred in the north of Green Lake, in Raoul Caldera, at a point which most nearly describes a common centre of eruption during the last 4000 years; except in the case of the hypothetical magnitude 5 eruption, which for clarity is shown arbitrarily centred in Denham Bay Caldera. Notice that the commonest prevailing wind directions are from the west, except in the case of magnitude 5 eruptions which would not be strong enough to rise higher than about 1 km above the surface. At this, and lower levels, the prevailing wind is from the east. Above about 3 km, which is the level usually reached by magnitude 6 eruptions, prevailing winds are from the west. No diagram has been drawn for magnitude 10 eruptions, in which the whole island is likely to be covered by at least 2 to 3 metres of tephra.

More than 20 cm of tephra is regarded as a serious risk for people, and for buildings; 10 to 20 cm as a moderately high risk; and 1 to 10 cm as a low risk. On this basis, it can be seen that expected risks from tephra at the Weather Station, on the north coast, are low for eruptions of magnitude 8 or less in both Raoul Caldera and Denham Bay Caldera (a magnitude 8 eruption there should still only give about 2 cm thickness at the Station). The risks are serious, however, for magnitude 9 or larger eruptions in either caldera.

In addition to pyroclastic flows and tephra falls, there are a number of other potentially serious hazards which are likely to affect Raoul Island. These include mudflows (lahars) formed through interaction of lake water or rainfall (there are no permanent streams) on pyroclastic flow or tephra deposits, gas discharges, earthquakes caused by gas pressure of magma seeking to force a way to the surface, and tsunami, which could affect a wide area of the south-west Pacific, as well as the coasts of Raoul Island itself. The Weather Station on the north coast stands not far from the sea, on Fleetwood Bluff, about 40 metres above high water mark. It is unlikely that the station is within reach of the great majority of tsunami of local origin. However, an exceptionally large one could conceivably endanger it. Such a tsunami could be generated by the impact of a very large pyroclastic flow into the sea, by collapse of part of the island into the sea, by a large earthquake offshore (which might or might not be related to volcanic activity), or by very powerful submarine explosions.

In addition to pyroclastic flows and tephra falls, there are a number of other potentially serious hazards which are likely to affect Raoul Island. These include mudflows (lahars) formed through interaction of lake water or rainfall (there are no permanent streams) on pyroclastic flow or tephra deposits, gas discharges, earthquakes caused by gas pressure of magma seeking to force a way to the surface, and tsunami, which could affect a wide area of the south-west Pacific, as well as the coasts of Raoul Island itself. The Weather Station on the north coast stands not far from the sea, on Fleetwood Bluff, about 40 metres above high water mark. It is unlikely that the station is within reach of the great majority of tsunami of local origin. However, an exceptionally large one could conceivably endanger it. Such a tsunami could be generated by the impact of a very large pyroclastic flow into the sea, by collapse of part of the island into the sea, by a large earthquake offshore (which might or might not be related to volcanic activity), or by very powerful submarine explosions.

In addition to the obvious risks of volcanic eruptions (especially pyroclastic flows) at the Weather Station on Raoul, eruptions of magnitude 8 or greater are likely to endanger overflying aircraft and shipping in the vicinity. Large eruption clouds will spread downwind for a considerable distance, so that aircraft over a wide area may be at risk. If an eruption continues for long enough for the wind direction or speed to change significantly, the actual pattern of tephra fallout, and hence also the distribution of ash at high levels in the atmosphere, will be much more complicated than the patters shown in Figure 6, which assume eruptions of short duration (during which wind speed and direction remain constant).

Figure 6 (a – d). Estimated tephra thickness in eruptions at Raoul Island of EVM magnitude 5 to 9, assuming the most prevalent wind conditions at altitudes reached by the eruption columns, and assuming a constant wind speed and direction during the eruptions. Note similar patters are to be expected from eruptions of other Kermadec Ridge volcanoes.

1. Magnitude 5 eruption in Denham Bay caldera (with 13 knot wind from east, at an altitude of about 1 km); and magnitude 6 eruption in Raoul Caldera (with 22 knot wind from west, at an altitude of about 3 km).
2. Magnitude 7 eruption in Raoul Caldera (with 27 know wind from west, at an altitude of about 6 km).
3. Magnitude 8 eruption in Raoul Caldera (with 47 knot wind from west, at an altitude of 9 to 10 km).
4. Magnitude 9 eruption in Raoul Caldera (with 27 knot wind from west, at an altitude of 16 to 17 km).

Historically observed eruptions at Raoul Island give a limited view of the possibilities for future activity, since the eruptions of the last 100 years have been much smaller (magnitudes 7 or less) than those of the last 4000 years.

During the 1814 eruption, a new island formed in Denham Bay. In addition, an unobserved phreatic eruption (Smith Breccia, see notes to Table 1 above), know from the presence of its deposits, occurred in Raoul Caldera at about the same time. From tree ring counts, a date of 1815±4 years has been obtain for this eruption. Since this is compatible with 1814, it suggests that it probably occurred at approximately the same time as the observed eruption in Denham Bay.

In 1870, from some date in the second half of June, a settler family named Covat, living at Denham Bay, and an injured sailor left by visiting ship, the Crowninshield, experience very frequent, almost incessant earthquakes. These were accompanied by suffocating sulphur fumes, and continuous submarine explosions in Denham Bay, where two new islands formed as a result of the activity. Steam and ash clouds rose to heights of 2000 or 3000 feet (about 1 km), and on 24 June the inhabitants fled into the interior of the island. When the ship returned in early October to pick up the injured sailor, Raoul Island was seen to be "burning fearful from four different parts", and the Covat family and the injured sailor were found "in the mountain very much frightened". Geological investigations have shown that the depression now occupied by Green Lake formed during the eruptions which were seen between mid June and early October 1870 in Raoul Caldera, at the same time as the Denham Bay eruption. A crater about 600 metres in diameter was excavated by a phreatic (water-driven) explosion which covered an area of about 2½ km² in debris.

The 1964 eruption, which began just before 6.00 a.m. on 21 November, was a short but dramatic sequence of explosions clearly seen from the Weather Station. The climax of the eruption, a steam and ash column up to an altitude of about 1.2 km, was reached within seconds of the initial outbreak (figure 7), and the bulk of the activity was over within about half an hour. Nevertheless, the island was evacuated shortly afterwards as a precautionary measure because of uncertainty about the possibility of continuing activity, recognition at this time that there had been eruption of extreme violence in the past, and the inevitable delays in help arriving from New Zealand, should this become necessary. A large crater, about 100 metres in diameter and 80 metres deep, was blasted out about 250 metres north-west of Green Lake, together with 11 smaller craters, as shown in Figure 5. In addition, fragments of fresh pumice were found floating in an area of discoloured sea water in Denham Bay, and it appears that a small eruption took place there also (Figure 5). It is notable that this eruption was considerably smaller, at magnitude 6, than any other recognised in the geological record. Also notable is the fact that in the last three eruptions at Raoul Island activity ha taken place in both calderas at approximately the same time.

 
Figure 7. (a-c). Eruption at Raoul Island on 21 November 1964, seen from the Weather Station. Photographs by J A Peart.
a) Just after the beginning of the eruption shortly before 6.00 a.m. The central eruption cloud, behind the radio mast, is erupting from the main crater about 250 metres north-west of Green Lake. The light-coloured expanded cloud on the right is the initial eruption cloud from the main crater which has drifted to the south-east (towards the right) owing to the wind. The convoluting cloud on the left, rising above white steam, marks a base surge (see text) travelling to the left.

 
b) Climax of the eruption at the main crater, photographed a few seconds after (a) above. A dense eruption cloud has reached an estimated height of 1.2 km. Rocks falling in spearhead thrusts, on ballistic trajectories, can be seen on the left of the column, just under the high mass of light-coloured steam. White steam rising at the base of the column marks the passage of base surges.

(c) Fall-out of debris from the convoluting eruption cloud some seconds after (b) above. Light coloured steam from the base surges is rising beneath, and to the left of the column.

Apart from the 1814, 1870 and 1964 eruptions, which were all witnessed and recorded, as outlined above, there is archaeological evidence that Polynesian people probably experienced some of the earlier eruptions. Stone adzes found on Raoul have been dated to the 14^th or 15^th centuries A.D., and it is thought possible that people were living on the island before the Rangitahua eruption, 300 or 400 years ago, and may have left as a result of it. It is not known whether there were people living on the island during the later Tui Breccia and Sentinel Tephra eruptions.

In addition to volcanic activity on and very near Raoul Island, as described above, there has been a single historical case of a submarine eruption at 29^o11’S 177^o52’W, a position about 8½ km north-east of the Weather Station. This point appears to lie on the northern slope of a small seamount, about 240 metres high, which reaches to about 560 metres below the surface. Pumice was seen rising from this spot by Thomas Bell, on board the whaler Othello in March 1886, and spreading out to form a raft estimated as 3 miles (about 5 km) long. No steam accompanied the upwelling.

With regard to future activity at Raoul Island, the prognosis is for more of the same. Existing data on the timing and scale of eruptions cannot define recurrence intervals between eruptions, or the timing of the next eruption, except in vague terms. The best estimate that can be made for average recurrenc intervals is as follows:

The "likelihood index" given above is a very rough measure of the likelihood of an eruption of a given size. The index is obtained by dividing the time since the last eruption by the average interval between eruptions of the same size. An eruption that occurs at about the average recurrence interval after the previous event of the same magnitude is less unexpected than one that takes place either much earlier or much later than this average interval. On this basis, either a moderately small eruption (magnitude 7), or a very large eruption (magnitude 9), appear quite probable, because the times that have elapsed since the previous eruptions are longer than average. The elapsed times since magnitude 6 and 8 eruptions are approaching the average, whereas a magnitude 10 eruption has a long way to go before the average recurrence interval is reached. It must be stressed that actual intervals depart widely from the average recurrence intervals given above, and eruptions of any size (within the range actually observed or inferred from existing deposits) are possible at any time.

Because of concern for the safety of staff at the Weather Station on Raoul, as well as for purely scientific reasons, a seismograph is kept operating at the Station and staff are instructed in how to recognise signs that might precede eruptions. Radio-telephone communications are adequate, if not particularly good, and people and the Station are encouraged to get in touch with DSIR scientists at Rotorua and Wellington if unusual activity is seen on the island, or recorded on the seismograph. Among other measures taken, a duplicate set of photographs of typical seismic recordings is held on Raoul and at DSIR Geology and Geophysics in Wellington, which aids in identifying events on the records. This method of "tele-analysis" has proved useful on several occasions.

There are problems, however, with rapid evacuation of people from Raoul. The island is so far away that it takes an RNZN frigate two days sailing from Auckland to reach it, and the presence of shipping nearby cannot be relied upon. The situation has recently improved with the construction of an emergency landing strip on the north coast of Raoul (Figure 5), but in the event of a large eruption this might be unusable. Until recently, the island was beyond the range of helicopters, except those carried on board ship. However, in May 1992, a helicopter based at Taupo equipped with extra fuel tanks, successfully flew from New Zealand to Raoul Island and back, with the aid of a fuel dump on L’Esperance Rock. An "all weather" boat is kept for emergencies, but unfortunately it cannot be launched in all weathers. Furthermore, road access between the Weather Station and the usual place to launch the boat, about a kilometre east-north-east of Blue Lake, runs along the narrow northern rim of Blue Lake Crater (see Figure 5), and is likely to be cut in the event of a moderate eruption in Raoul Caldera.

2. MACAULEY VOLCANO

Macauley Island, the second largest in the Kermadecs, lies about 110 km south-south-west of Raoul Island. Its highest point is Mount Haszard (238 metres above sea level) in the north-west part of the island, which is roughly circular in outline with a diameter of about 3 km. Its total area is only about 300 hectares (see Figure 8 and 9 for plan and photograph of the island). Haszard Islet lies off the south-east coast, separated from the island by a channel about 300 metres wide.

 
Figure 8. Bathymetry (in metres) of Macauley Caldera (from paper by E.F. Lloyd and S. Nathan, in preparation; see Supplementary Reading). Survey lines are shown dotted. Bars marked 72-75 are sites of dredge samples.

No historical eruption has been observed on or very near Macauley Island, but two submarine eruptions have been reported within 50 km distance. On 6 September 1825, a small ring-shaped island, christened Brimstone Island, was seen at a reported position about 45 km west of Macauley. The island has not been seen since, and the position given seems very doubtful, as it lies in water more than 2000 metres deep. The second submarine eruption was seen on 1 December 1887 at a reported position about 22 km north-north-west of Macauley, which again appears to be in deep water, in this case a little more than 1500 metres deep.

The geology of Macauley Island was first described in detail by R.N. Brothers and K.R. Martin of Auckland University in 1970. It has since been examined more fully by two of the present authors (E.F.L. and S.N.: see Supplementary Reading). The later studies, together with oceanographic work carried out from the Russian research ship Vulkanolog, have firmed that Macauley Island is a small fragment of land on the rim of a large submarine caldera, called Macauley Caldera, which is centred about 8 km north-west of Macauley Island (Figure 8). The caldera is about 13 km in an east-west direction, by about 11 km north-south, and more than 1000 metres deep. A steep slope separates it from Macauley Island, and it is likely that part of the island collapsed into the caldera during a later eruption than the one responsible for its formation.

On the island, which is composed mainly of basaltic lava flows erupted above sea level, there is a prominent thick deposit of dacitic pumice (Figure 9), now recognised as an unwelded ignimbrite (Sandy Bay Ignimbrite). The likely reason for it being unwelded is that it was emplaced relatively cool, having lost heat by impacting into the sea. The Sandy Bay Ignimbrite is similar to the Taupo Ignimbrite erupted about 1800 years ago in New Zealand, and its eruption was the cause of the formation of Macauley Caldera. The volume of the eruption has been calculated as about 100 km³ (EVM magnitude 11). On the assumption that half of this was emplaced as airfall tephra from a high eruption column, and assuming typical high-level westerly winds and average wind speeds, the resulting deposit probably covered approximately the area shown in Figure 10. Although the 1 cm contour reaches nearly as far as Pitcairn and Easter Island (Morotiri) in the southern Tubuai (Austral Islands) group, should both have received 1 to 2 cm thickness of fine ash. No studies have yet been carried out on these islands to determine whether this was indeed the case. Nor have Sandy Bay eruption deposits yet been found on Curtis or Raoul Islands. At present it has not been possible to date the Sandy Bay Ignimbrite, although it is clear that the eruption is at most a few thousand years old and my even be only a few hundred years old.

Figure 9. (a-b) Macauley Island. Photographs by B.J. Scott.

(a)View looking north-west showing light-coloured Sandy Bay Ignimbrite overlying dark lava flows produced by earlier eruptions. Haszard Islet can be seen on the right.

(b)Lava Cascade, a lava flow of the Haszard scoria and lava eruption filling a gully cut in Sandy Bay Ignimbrite. The dark shore platform is composed of earlier basalt lavas.

Figure 10. Estimated tephra thickness in the Sandy Bay Ignimbrite eruption (EVM magnitude 11) from Macauley Volcano, assuming the most prevalent wind conditions (27 knot wind from west, at an altitude of 16 to 17 km), and assuming that wind conditions remained constant throughout the eruption. Note that a similar pattern would result from any magnitude 11 eruption on the Kermadec Ridge.

Possibly a hundred years or so after the ignimbrite was erupted, there was a much smaller basaltic eruption which originated on what was then land, one or two kilometres north of the present Macauley Island. Lava flows and scoria deposits were first erupted onshore (Figure 9b), then part of the island collapsed, the sea flooded the vent, and the remainder of the eruption was submarine. Two phases of this eruption, the Haszard Scoria and associated lavas, and the overlying Parakeet Tuff, have left deposits on Macauley Island large enough for the EVM magnitudes of the tephra eruptions to be estimated. Haszard Scoria, erupted while the vent was above sea level, has a magnitude of 8, and Parakeet Tuff, the product of submarine explosions, a magnitude 7.

The Sandy Bay eruption of magnitude 11 from Macauley Caldera is the largest eruption so far identified in the south-west Pacific, in the entire region between New Guinea and New Zealand, with the exception of certain eruptions on the New Zealand mainland. The fact that it occurred so recently, within at most a few thousand years ago, highlights the risk of future eruptions from Macauley Island. It must be expected that magnitude 8, 9 and 10 eruptions, as well as smaller events, are fairly frequent occurrences from this volcanic centre.

Future eruptions at Macauley Volcano are likely to be submarine. Resulting hazards include large eruption columns and downwind tephra plumes which will prove dangerous to aircraft (within a large distance, in the case of the larger eruptions). Impact of substantial pyroclastic flows into the sea, possibly with future collapses of Macauley Island into the caldera, is also likely. These may endanger shipping in the vicinity. Both the pyroclastic flows and the collapses will give rise to large, potentially destructive tsunami, which may affect coastlines at a great distance. Environmental damage, through the discharge of large volumes of gas and fine tephra particles, may also occur. It is fortunate, however, that prevailing winds are generally from the west, and will carry the bulk of gas and ejecta into the South Pacific, where population is sparse or non-existent.

2. CURTIS VOLCANO

Curtis Island (Figure 11), about 35 km south-south-west of Macauley Island, is a small island, about 800 metres in a north-west by south-east direction, and 500 metres at right angles. Cheeseman Island, about 500 metres by 300 metres, and similarly aligned, lies to the west, across Stella Passage, which is about 600 metres wide. The islands are the eroded peaks of a volcano at the summit of a broad topographic rise, measuring about 25 km by 15 km, on the crest of the Kermadec Ridge. Some 20 km south-east of Curtis Island, on the same topographic rise, is a recently discovered submarine seamount, which has been tentatively identified as volcanic, on the basis of a plume of gas and a chemical anomaly in the water. Curtis Island, unlike Macauley and Raoul Islands, has active fumaroles (gas vents) in its crater, a structure 300 to 400 metres long by about 200 metres wide, with its floor only about 10 metres above sea level.

Pyroclastic deposits, part of which are welded, make up the bulk of Curtis Island. The rocks are massive, poorly-bedded ash deposits with larger blocks. Fragments in the ash are of pumice and andesite, and the fact that some of the rocks are welded suggests that they may be ignimbrites. The vent from which the rocks were erupted is unknown and almost certainly lies offshore. The rocks have not been dated, and the volcanic history of Curtis and Cheeseman Islands is not yet known. A study in 1979 by A.C.Doyle, R.J.Singleton and J.C.Yaldwyn (see Supplementary Reading) demonstrated that remarkable uplift has taken place in the islands, totalling 18 metres in about the last 200 years, and 7 metres between 1929 and 1964. This, and the presence of active fumaroles, suggests that the volcano is potentially active, and the presence of pumice fragments means that the eruptions are certainly explosive, and probably large. There are no authentic reports of historically observed eruptions at Curtis Island, although there have been several occasions on which a greater discharge of acrid fumes than usual has been seen.

Figure 11. Curtis Island. View of the crater in 1977, looking south-east. Aerial photograph by S.Nathan.

Hazards to be expected are similar to those for Macauley Island. Dangers are principally to aircraft and to shipping in the vicinity. Because of the likely explosive nature of eruptions at Curtis Island, it is to be expected that tsunami will be generated.

3. L’ESPERANCE ROCK

About 100 km south-south-west of Curtis Island is L’Esperance Rock (Figure 12), previously known as French Rock, an islet 70 metres high, measuring about 250 metres in diameter. It is composed of andesite lava flows and associated intrusions. Some 8 km to the north-west is Havre Rock, which scarcely projects above the water at low tide. Both these features, the most southerly points above sea level in the Kermadec Islands, are erosional summits of a large topographic high, about 30 km north-south by 25 km east-west, on the crest of the Kermadec Ridge. Although L’Esperance Rock is certainly a volcanic remnant, it shows no sign of either volcanic land form or volcanic activity, and no dates have been obtained from the andesite of which the Rock is composed.

4. STAR OF BENGAL BANK

Approximately 100km south along the Kermadec Ridge from L’Esperance Rock is a large broad topographic rise known as Star of Bengal Bank (Figure 1). It has not yet been closely studied. In all likelihood, active or potentially active submarine volcanoes exist in this area.

3. SOUTH KERMADEC RIDGE SEAMOUNTS

The South Kermadec Ridge Seamounts (Figure 2), at the southern end of the Kermadec Ridge, include the Rumble Seamounts (I to IV) and two seamounts described by A.C. Kibblewhite as Silent I and II (because there was apparently no possibility of the noises recorded on Great Barrier Island – BUN, as described above – originating there ). There are also three recently discovered cones (see below) and a number of other features about which little is known.

Figure 12: L’Esperance Rock. View from the west. Aerial photograph by I.E.M. Smith.

Fresh glassy andesite lava has been dredged from the summit of Rumble III, and basaltic lava from the flanks, and it is now believed that all recorded instances of BUN originated at this seamount. Dramatic confirmation of the active nature of Rumble III was obtained on 13-14 July 1986, when the crew of a Japanese fishing boat reported light-coloured steam rising from the ocean at 35 º 44’S 178º29’E, and a sulphur slick covering an area of 500 m². Figure 13 shows discoloured water due to a rising gas plume at Rumble III, photographed by the crew of an RNZAF aircraft about three weeks later. The area was subsequently surveyed by HMNZS Tui , and later by the Russian ship Vulkanolog, and it was found that the summit of Rumble III lies about 200 metres below the surface. In the course of the same investigations, a plume of rising gas was also detected above the summit.

Rumbles I and II have also been dredged, yielding somewhat less fresh andesite, together with limestones and other sediments, which suggest that these volcanoes may now be extinct. Fresh glassy lava comparable to that found on Rumble III has, however, been dredged from Rumble IV, and this and the fact that a column of gas has been detected in the water above the cone suggests that it too may be an active volcano. Three other large stratovolcanoes, from which andesite and basaltic-andesite rocks have been dredged, have recently been discovered east, south and south-south-west of Rumble IV, at distances of 17, 28 and 46 km, respectively. All are as yet unnamed: they are shown in Figure 2. Each stands more than 1000 metres above the sea floor and reaches within 1500 metres of the surface. Rising gas has been detected above the most easterly of these three cones, which, like Rumble III and Rumble IV, is therefore regarded as active.

Only about 125 km north-east of the Bay of Plenty coast of the North Island , New Zealand, just beyond the 2000 metres depth contour, lies Whakatane Seamount (Figure 1, 2). Fresh andesite lava has been dredged by DSIR’S Oceanographic Institute from this locality.

Further inshore towards New Zealand, there are numerous seamounts, some of which show signs of recent activity. Of particular interest is Mahina Knoll (Figure 2), about 20 km north-west of White Island, which shows evidence of a recent partial collapse, and has yielded dredge samples of dacite and andesite.

Figure 13: (a,b). Water discoloured by rising gas at Rumble III Submarine Volcano on 5 August 1986 (aerial photographs by RNZAF aircrew)

Basaltic lavas, which appear to be very young, have recently been dredged from the eastern slope of the Ngatoro Basin (Figure 2). Little or no sediment covers them, and it is clear that they must have been erupted very recently. The Ngatoro Basin is a tensional feature, possibly a rift structure, and an entirely different tectonic environment to that of the stratovolcanoes to the east.

Finally, north-west of the Ngatoro Basin, at the south-west end of the Colville Ridge (Figure 2) is a group of submarine volcanoes, believed recently active. This group is aligned with Mayor Island, close to the New Zealand coast, which is a dormant volcano that characteristically erupts an unusual variety of rhyolite.

Risks resulting from eruptions of Rumble III and IV, and the seamounts mentioned above , cannot be reliably estimated. It is probable, however, that all are capable of generating moderate-sized tsunami. There is certainly danger to shipping in the vicinity, and, in the event of a large eruption, also to aircraft. There is no information at present on recurrence intervals of eruptions at these volcanoes.

An awareness of possible risks associated with eruptions in the Kermadecs or at the Rumble seamounts is not the only reason for drawing attention to these volcanoes. As an integral part of our environment, knowledge about them for its own sake is important. The existence of these fascinating submarine mountains, whose peaks, even the Kermadecs, barely reach above the surface of the ocean, is a fact of which all ought to be aware.

Thanks are due to L.H. Hall, J.A. Peart, B.D. Scott, B.J. Scott and I.C. Wright for supplying unpublished information, maps or photographs; to V.E. Neall, H.W. Wellman and I.C. Wright for helpful criticism of the text; to H.J. Anderson, D.O.R. Bright, F.J. Davey, B.W. Davey, P.J. Dickson, G.P. Glasby, R.D. Maunder, C.C. Piesse, M.E. Reyners and staff of the Weather Station on Raoul Island, who contributed vital information; to C. Hume who did the drawings, to L. Homer and W. St. George who reproduced some of the photographs, and to W. Esam, P Bratton and S. Nepe who typed the manuscript. The Lottery Grants Board is gratefully acknowledged for a grant towards publication of this booklet.

SUPPLEMENTARY READING

Blong, R.J. 1984: Volcanic Hazards: A sourcebook on the effects of eruptions. Academic Press Sydney, 424 pages. This is the best general account of volcanic hazards available.

Carter, L. 1976 Cuvier Bathymetry. New Zealand Oceanographic Institute, Chart, Coastal Series, 1:200,000, D.S.I.R. Wellington.

Davey, F.J. 1980 : The Monowai Seamount : An active submarine volcanic centre on the Tonga-Kermadec ridge. New Zealand Journal of Geology and Geophysics, Volume 23 (4), pages 533-536.

Doyle, A.C., Singleton, R.J. and Yaldwyn, J.C. 1979. Volcanic activity and recent uplift on Curtis and Cheeseman Islands, Kermadec Group, Southwest Pacific. Journal of the Royal Society of New Zealand, Volume 9 (1), pages 123-140.

Frohlich, C. and Barazangi, M. 1980: A regional study of mantle velocity variations beneath eastern Australia and the southwestern Pacific using short-period recordings of P, S, PcP, ScP and ScS waves produced by Tongan deep earthquakes. Physics of the Earth and Planetary Interiors, Volume 21, pages 1-14. In spite of its alarming title, this is a good recent account of the seismicity and shape of the colliding plates in the Kermadec-Tonga region.

Healy, J., Lloyd, E.F., Banwell, C.J. and Adams, R.D. 1965: Volcanic eruption on Raoul Island, November 1964. Nature, London, Volume 205 (4973), pages 743-745.

Hydrographer of the Navy. 1963: Tongatapu Island to L’Esperance Rock. Admiralty Chart, No. 2283, RN Hydrographic Office, Taunton, England.

Hydrographer of the Navy. 1980: South Pacific Ocean, New Zealand, Kermadec Island to East Cape. Chart New Zealand 22, RNZN Hydrographic Office, Auckland.

Hydrographer of the Navy. 1987: New Zealand Pilot. 14th Edition. RN Hydrographic Office, Taunton, England. This and the two Charts above give useful local information about the islands and surrounding waters.

Kibblewhite, A.C. 1966: The acoustic detection and location of an underwater volcano. New Zealand Journal of Science, Volume 9 (1), pages 178-199. Describes BUN, and reports discovery of the Rumble Volcanoes.

Kibblewhite, A.C. 1967: Note on another active seamount in the South Kermadec Ridge Group. New Zealand Journal of Science, Volume 10 (1), pages 68-69.

Lloyd, E.F. and Nathan, S. 1981: Geology and tephrochronology of Raoul Island, Kermadec Group, New Zealand. D.S.I.R. New Zealand Geological Survey Bulletin 95, 105 pages. The most comprehensive report available on the geology of Raoul Island: has a full bibliography of papers on the Kermadec Islands.

Lloyd, E.F. and Nathan, S. In prep.: Volcanic history of Macauley Island, Kermadec Ridge, New Zealand. New Zealand Journal of Geology and Geophysics.

Oliver, W.R.B. 1910. The geology of the Kermadec Islands. Transactions and Proceedings of the New Zealand Institute, Volume 43, pages 524-535. An interesting early account of the geology, especially of Raoul Island.

Pantin, H.M., Herzer, R.H. and Glasby, G.P. 1973. Bay of Plenty Bathymetry. New Zealand Oceanographic Institute, Chart, Coastal Series, 1:200,000, D.S.I.R. Wellington.

Smith, S.P. 1895: Volcanic activity in Sunday Island in 1814. Transactions and Proceedings of the New Zealand Institute, Volume 28, pages 47-49. Includes eyewitness account of the 1814 eruption of Raoul Island.

Wright, I.C. 1990: Bay of Plenty-Southern Havre Trough Physiography. New Zealand Oceanographic Institute, Chart, Miscellaneous Series No. 68, 1:400,000, D.S.I.R., Wellington.

Wright, I.C. 1992: Shallow structure and active tectonism of an offshore continental backarc spreading system: the Taupo Volcanic Zone, New Zealand. Marine Geology, Volume 103, pages 287-309. The most definitive account of bathymetry and structure for the region between Whakatane Seamount and the New Zealand coast.

Opposite Back Inside Cover: Fresh basaltic lava flows in 1745 metres of water, on south flank of an unnamed seamount east of Rumble IV. The underwater camera is triggered by the line and weight shown. Photograph by I.C. Wright.

This text is taken from one of a series booklets which cover volcanic hazards at each active or potentially active volcanic centre in New Zealand.

The series was produced by the Volcanic Hazards Working Group of the Civil Defence Scientific Advisory Committee, which includes scientists from the Institute of Geological and Nuclear Sciences and the Universities.

• Booklets published in the series so far are:
• Number One ‘Egmont Volcano’.
• Number Two ‘Okataina Volcanic Centre’.
• Number Three ‘White Island’.
• Number Four ‘Kermadec Islands’
• Number Five ‘Auckland Volcanic Field’
• Number Six ‘Mayor Island’
• Number Seven "Taupo Volcanic Centre’