Craig Miller

Monitoring seismic precursors to an eruption from the Auckland Volcanic Field, New Zealand

Abstract The Auckland Volcanic Field (AVF) in New Zealand is monitored by a network of five telemetered, vertical‐component, short‐period seismographs. Between 1995 and 2005, 24 earthquakes were located in the Auckland region. Ten of these were located reasonably reliably (position and depth uncertainty ≤10 km) and all of these were <15 km deep. Only one of these earthquakes occurred within the AVF. Magnitudes ranged from ML 1.6 to 3.3, and five earthquakes of ML ≥ 2.4 were felt. There were few reliably located earthquakes because most were not recorded by the whole network owing to their relatively low magnitude and a high level of background noise. The Auckland earthquakes are believed to represent normal background seismicity and are not thought to be eruption precursors. All earthquakes were of high‐frequency, tectonic type; no low‐frequency, volcanic earthquakes were recorded. Based on seismic precursors to eruptions from historically active volcanic fields, we estimate that precursory earthquakes could occur as little as 2 weeks before an Auckland eruption and they could be as large as ML 4.5–5.5. Based on the depth of the background seismicity in Auckland, and previous estimates of the ascent rate and source depth of AVF magmas, we calculate a precursory period as short as a few days. Our best estimate of the length of preeruption seismicity is therefore a few days to a few weeks. The largest precursory earthquakes could be large enough to be felt by most of the population who live in Auckland City. During a magmatic intrusion, deep long‐period earthquakes might occur at c. 30 km as magma ascends into the crust. Earthquakes would probably have to be a lot shallower, perhaps only 5 km, before their epicentres might be useful for estimating the location of any eruption. Geodetic monitoring methods (GPS and InSAR) might perform as well as seismic monitoring for identifying unrest, but they have significant limitations. To better monitor and interpret precursory seismicity from the AVF, an increase in the number of seismographs and an improvement in our understanding of the local crustal structure are needed.

Hazards from hydrothermally sealed volcanic conduits

The 17 March 2006 eruption from Raoul Island (Kermadec arc, north of New Zealand) is interpreted as a magmatic-hydrothermal event triggered by shaking associated with a swarm of local earthquakes. The eruption, which tragically claimed the life of New Zealand Department of Conservation Ranger Mark Kearney, occurred without significant volcanic seismicity or any of the precursory responses the volcanic hydrothermal system exhibited prior to a similarly sized eruption in 1964. Preliminary evidence suggests that the absence of precursory behavior is probably the consequence of hydrothermal sealing of the volcanic conduit since the 1964 eruption, and points to potential hazards associated with quiescent oceanic island volcanoes.

A model for developing best practice volcano monitoring: a combined threat assessment, consultation and network effectiveness approach

This paper presents a combined approach to achieving best practice volcano monitoring through a review of New Zealand’s volcano-monitoring capability as established under the GeoNet project. A series of benchmark, consultation and network performance studies were undertaken to provide a comprehensive review of volcano monitoring in New Zealand and to establish plans for future improvements in capability. The United States Geological Survey National Volcano Early Warning System method was applied to benchmark the established monitoring networks against recommendations for instrumentation based on a volcano’s threat level. Next, a consultative study of New Zealand’s volcanology research community was undertaken to canvass opinions on what future directions GeoNet volcano monitoring should take. Once the seismic network infrastructure had been built, a noise floor analysis was conducted to identify stations with poor site noise characteristics. Noise remediation for poor sites has been implemented by either re-locating the site or placing sensors in boreholes. Quality control of Global Navigation Satellite System networks is undertaken through the use of multipath parameters derived from routine processing. Finally, the performance of the monitoring networks is assessed against two recent eruptions at Mount Tongariro and White Island. This combined approach can be used as a model to assess the need for future monitoring levels on any volcano.

Super-volcanic investigations

Multi-disciplinary analyses of Earth’s most destructive volcanic systems show that continuous monitoring and an understanding of each volcano’s quirks, rather than a single unified model, are key to generating accurate hazard assessments.