Raoul Is Review Confirms GNS Science's Actions
16 JUNE 2006
Review Confirms GNS Science’s Actions
A comprehensive review of GNS Science’s actions relating to the volcanic eruption on Raoul Island has concluded that the organisation’s advice and procedures were sound. The conclusions were based on a review of events immediately preceding the eruption, as well as the seismic and volcanic history of Raoul Island.
GNS Science Chairman, Con Anastasiou, said today the review confirmed GNS Science had made the best decisions it could at the most appropriate time. The phreatic (steam) eruption, on 17 March 2006, occurred without warning and was very brief, lasting about three minutes.
Department of Conservation employee, Mark Kearney, went missing at the time of the eruption. Mr Kearney was carrying out routine work involving measuring, monitoring, sampling and observation of the volcano when the eruption occurred.
Mr Anastasiou said that his deepest sympathies and those of GNS Science go out to Mr Kearney’s family, friends and colleagues at this difficult time.
The monitoring work on Raoul Island was being carried out for three reasons:
- To warn of any possible build-up to a large-scale eruption which would require the rapid evacuation of Conservation Department staff from the island.
- To support the international aviation warning system designed to ensure pilots could avoid volcanic ash clouds.
- To provide data to enable GNS Science volcanologists to study trends and search for signs of increasing unrest in one of the world’s most active volcanic regions.
Mr Anastasiou said that safety was of paramount importance for GNS Science.
“The nature of our work often involves scientific investigation in remote and difficult areas which are potentially hazardous. GNS Science’s safety procedures are designed to address these potential hazards with the aim of making sure people are as safe as practicable,” he said.
“ The Department of Conservation staff who carried out the monitoring work on Raoul Island were given specific training in recognizing and dealing with potential hazards associated with volcanic areas by GNS Science, as well as specific training in the science involved.
“However, it was important for us to conduct an in-depth review of the circumstances surrounding the volcanic eruption on Raoul Island and to assess what lessons could be learned both by GNS Science and by others exposed to these kinds of environments."
Mr Anastasiou said the report resulting from the review emphasised the difficulty in predicting volcanic eruptions and the critical importance of monitoring. Both the report, and the international experts who reviewed it, concluded that GNS Science had done all it could do in the circumstances and had made sound judgments based on the events at the time and the known history of Raoul Island.
“ GNS Science remains committed to maintaining and enhancing its safety procedures. This was an unfortunate and unpredictable accident, but we must do our best to learn from events of the past and strive for improvement in the future."
The 2006 Raoul Island Eruption – a Review of GNS Science’s Actions
16 June 2006
The review of GNS Science’s actions relating to the volcanic eruption on Raoul Island, contained in a report entitled “The 2006 Raoul Island Eruption – a Review of GNS Science’s Actions”, was conducted by a panel of three. The panel members were:
- Dr Hugh Cowan, Research Manager, Earthquake Commission
- Professor Jim Cole, Department of Geological Sciences, University of Canterbury
- Dr Terry Webb, General Manager, Natural Hazards, GNS Science.
The terms of reference were as follows:
- Were the scientific judgments made in the lead-up to the Raoul Island eruption soundly based.
- Were GNS Science’s procedures followed.
- What improvements can be made in light of this experience.
The report was peer reviewed by international experts in volcanology. They were:
- Dr David Hill, Scientist-in-Charge, Long Valley Observatory, California
- Dr Margaret Mangan, Associate Scientist-in-Charge, Long Valley Observatory, California
- Randall White, Chief Seismologist for the USGS Volcano Disaster Assistance Program, California.
The review panel found that:
- Each volcano is unique. What happens at one cannot be used to predict accurately what will happen at another.
- Despite scientific advances, it is still difficult to predict both the timing and the size of an eruption.
- The best guide to the way a particular volcano behaves is to study its history.
In relation to events leading up to the eruption on Raoul Island, the review panel noted: - GNS Science identified a high-frequency earthquake swarm on Sunday 12 March 2006. This was intense for about 14 hours and then declined steadily. It was estimated to be about 12-15kms away from the crater lake.
- GNS Science volcanologists did not consider this to be a sign or precursor leading to an eruption. Earthquake sequences had been recorded in 1989, 1990, 1991, 1992, 1993 and 1995, but there had been no eruptions.
- The earthquake swarm in March 2006 was considered to contain mostly highfrequency events, whereas earthquakes preceding eruptions are known to be a mixture of both high and low-frequency earthquakes. The characteristics of the swarm, therefore, were not indicative of an impending eruption.
- In earlier eruptions, such as the last one in 1964, the crater lake temperature had increased prior to the eruption. There was no increase in lake temperature or ground temperature in March 2006.
- Prior to the 1964 eruption, the level of Green Lake had risen by six metres. There was no increase in lake level in March 2006.
- Salinity increased in Green Lake prior to the 1964 eruption. However, chemical analysis of lake water, sampled before the 2006 eruption and analysed after the eruption, showed no increase in salinity or other changes to the lake chemistry in March 2006.
- Ground cracking had occurred in 1964. It did not occur in March 2006.
- Measurements showed no lake tilting in March 2006 – a key tool for detecting the early stages of an eruption.
- As it had done during an earthquake swarm in 1993, GNS Science had increased monitoring from weekly to daily on 13 March 2006. This would identify the onset of any volcanic unrest that might require the evacuation of Conservation Department staff.
- On 14 March 2006, GNS Science noted that the swarm was declining.
- On 16 March 2006, the swarm was continuing to die away.
- The eruption occurred on 17 March 2006.
The panel concluded that the volcanic eruption of Raoul Island was a small eruption that occurred without warning. Such events are very difficult to predict because they have few precursors and our understanding of them is incomplete.
The panel concluded GNS Science’s procedures were current. They were last reviewed seven months prior, in November 2005.
The panel also looked closely at the procedures GNS Science staff had applied to deal with Scientific Alert Levels. These define the current status of each volcano. The panel noted that the Scientific Alert Levels do not predict what is going to happen; they define what is happening.
The panel concluded that adequate attention was paid to the earthquake swarm activity and good teamwork was evident in the way the matter was handled and that procedures were followed. The panel also found that the scientific advice based on the available information was sound.
With regard to future improvements, the panel concluded that:
- The lesson learned concerning the Raoul volcano, that small eruptions may occur there without precursors, should be incorporated into hazard control plans for people on Raoul Island. The Department of Conservation and GNS Science are working closely on this.
- Detailed consideration should be given on how the volcano is monitored in future.
There must be a balance between less monitoring and more monitoring. Less monitoring could increase the risk to Department of Conservation in the long-term through lack of knowledge and warning of a large eruption. More monitoring could increase the risk for the people making the measurements.
- Some remote monitoring may be a possibility, though it is difficult to maintain equipment in remote and corrosive volcanic environments.
- A planned upgrade of the communications link to Raoul Island will improve monitoring of seismic activity on the island.
- Ongoing research at Raoul Island and other New Zealand volcanoes will improve understanding of pre and post-eruptive behaviour. This should lead to an enhanced ability to forecast the potential impacts of such events on New Zealand communities.
The international experts concluded:
- The judgments made by GNS Science were sound and reasonable, given the declining earthquake swarm and the absence of other signs of an imminent eruption.
- The GNS Science team carried out their responsibilities in an exemplary manner that does credit to the scientific and technical competence of the organisation.
- Every new eruption is a learning experience. The risk assessment and recommendations made by the review panel were reasonable.