Raoul Island is the largest of the Kermadec Islands, over 1,000 kilometres north-east of the North Island. It is the emergent part of a large volcano, almost 20 kilometres in diameter. Although apparently isolated, the island lies on the Kermadec ridge, a chain of submarine volcanoes.
The island’s irregular anvil shape is due to a combination of volcanic activity and marine erosion. Detailed onshore and offshore investigations show that there are two large collapse calderas (Raoul and Denham), both of which have erupted frequently over the last few thousand years.
Eruptions have been observed and recorded in 1814, 1870, 1964 and 2006.
On 9 March 1814 Captain Barnes of the Jefferson, who had landed in Denham Bay several days earlier, reported a thick, dark cloud above Raoul. Returning two months later, he found that there was a volcanic island made of scoria in the middle of Denham Bay. Explosive activity occurred at the same time in Raoul caldera.
The scoria island had disappeared by 1854, when a survey in Denham Bay was made by HMS Herald.
Heading for the hills
William Covat and his family were living in Denham Bay when eruptions started in June 1870. Alarmed by earthquakes, volcanic explosions and suffocating fumes they fled to the hills, and were very pleased to be rescued in early October. They never returned.
Primary records indicate that there were eruptions in both the Denham Bay and Raoul calderas between June and October 1870. Captain Preble of the Ellen Goodspeed described an erupting volcano in Denham Bay on 5 July that blasted a dark column of ash and steam up to 1,000 metres. In early October 1870, there were two separate volcanic islands in Denham Bay.
During the same period there was an eruption in Raoul caldera which blasted a crater 600 metres across, knocking over trees and covering the floor of the crater with mud and boulders.
There was no further volcanic activity for the next 90 years, and Raoul Island has been occupied almost continuously. The area devastated in the centre of the Raoul caldera was soon covered in trees that grow rapidly in the subtropical environment.
In 1954 the British government requested permission to test a hydrogen bomb in the Kermadec Islands. Although there had been eruptions in 1814 and 1870, no one recognised that Raoul was an active volcano. Indeed, the central flat area surrounded by steep cliffs (now recognised as the Raoul caldera) seemed an excellent testing spot. Fortunately the request was declined by the New Zealand government – for political reasons.
A series of strong earthquakes started on 10 November 1964, and by 15 November volcanic tremor was recorded. At about the same time Green Lake started to heat up, and the level rose several metres. An explosive eruption commencing about 6am on 22 November opened a large vent near Green Lake which devastated the immediate area, and there were eleven other small vents within the Raoul caldera. Activity gradually tailed off after early February 1965. Mud ejected from the vents was confined to the central part of the Raoul caldera.
A plume of discoloured bubbling water containing pumice fragments was observed in Denham Bay in November and December.
An earthquake swarm started near Raoul Island on Sunday 12 March 2006, but the tremors rapidly declined over the next few days. Then on 17 March at 8.21 a.m. there was a sudden explosive eruption near Green Lake. Mud and rocks were ejected, together with a plume of steam and gas. Mark Kearney, a Department of Conservation observer, was killed by the explosion. Green Lake was heated, and the level rose several metres during the next few days.
Pattern of eruption
All historic eruptions include explosive activity within the Raoul caldera close to Green Lake, probably caused by disruption of the geothermal system as magma moves upwards. There appears to be a close connection with vents in the Denham caldera, as the 1814, 1870 and 1964 eruptions occurred there at the same time. It is highly probable that there will be future eruptions from the same vents.
Acknowledgements to Ted Lloyd (formerly New Zealand Geological Survey, DSIR), Vince Neall (Massey University), Brad Scott and David Johnston (GNS Science), and David J. Lowe and Richard Smith (University of Waikato).