Borehole seismometer monitoring of changing properties on Mt Ruapehu Volcano

Abstract

We installed borehole seismometers on Mt. Ruapehu volcano and examined seismic waveforms of earthquakes and explosions recorded on these and other GeoNet and portable seismometers. We developed and applied techniques to compare changes in seismic wave propagation on volcanoes to magma movements and to other techniques that measure stress changes on volcanoes. We determined baselines of seismic velocity values against which changes can be measured. We also found that on Ruapehu and other volcanoes, changes in anisotropy correlate to changes in stress indicators measured from other techniques such as GPS baseline lengths, focal mechanisms and the relative proportion of small versus large earthquakes (b-values).

---

Technical Abstract

On Mt. Ruapehu Volcano, previous studies suggested that changes in seismic anisotropy and attenuation occurred due to stress changes associated with magma emplacement. It was not clear, however, whether the changes would be large enough and close enough in time to predict future eruptions. A key to recognising subtle changes in geophysical properties is to determine background measurements during times of little change in the volcanic processes and to then use repeating seismic sources and a borehole seismometer.

We addressed three specific objectives:

1. To determine suitable repeating sources to use to measure changes in seismic properties over time, we installed a borehole seismometer, which was supplemented later by a second borehole seismometer funded by a Marsden grant. We also installed a 16-station portable network to complement the permanent GeoNet network around Mt. Ruapehu during the year 2008. We left a 5-station portable network running above a cluster of earthquakes in the Waiouru area so that we could use the highly correlated events among them as sources of repeating energy. Finally, we carried out a series of explosions in Lake Moawhango, which is near to Mt. Ruapehu and also to the Waiouru earthquake swarm, to provide a benchmark against which future changes in seismic properties can be measured.
2. We developed a set of automatic shear wave splitting computer programs to quickly measure anisotropy on Mt. Ruapehu and other volcanoes. The shear wave splitting parameters of fast polarisation give the orientation of the anisotropic medium, and the delay time is a product of the strength of the anisotropy and the length of the path in the anisotropic medium. Fast polarisations are expected to be parallel to the maximum principal stress direction if the anisotropy is caused by fluid-filled, stress-aligned cracks. Other possibilities are that faulting will cause minerals to align with the fault planes, or that polarisations will align with anisotropic minerals in rocks such as schists. We developed a code to perform a two-dimensional tomographic inversion for shear wave splitting delay time and spatial averages of fast directions. The codes were applied to the results of the portable deployment to serve as a benchmark against with other changes could be measured. We found that Mt. Ruapehu is more complicated than previously thought, with some regions being controlled by nearby faults and others by stress.
3. We compared the shear wave splitting measurements to areas where other types of stress indicators had been suggested to change, and we used the Coulomb 3.0 stress modelling code to try to determine if stress changes could explain all the phenomena. At Ruapehu changes in b-values for earthquakes in the Erua region occurred at the same time as changes in anisotropy for paths these earthquakes and permanent station FWVZ. At Mt. Asama volcano in Japan anisotropy changes correlated with changes in strain as measured by GPS baselines and at Souffriere Hills volcano in Montserrat, anisotropy changes correlated with changes in focal mechanisms. At both these latter volcanoes, calculated stress from inferred dykes could explain the general orientation of the anisotropy. At Okmok volcano in Alaska, changes in splitting were considered to be caused by changes in earthquake location; repeating multiplets yielded near constant splitting measurements, and stopped repeating after the eruption began.