Search for temporal changes in seismic attenuation on Mt Ruapehu volcano

Abstract

We tested the hypothesis that changes in local stress around Mt Ruapehu volcano caused by the 1995/96 eruptions alter the amount by which seismic wave amplitudes are damped. Previous studies observed changes in shear waveforms (Miller and Savage, 2001; Gerst and Savage, 2004), which were associated with changes in the local stress field caused by repeated filling and depressurising of a magmatic dyke system under Mt Ruapehu. Temporal variations of seismic amplitude damping have been observed at different volcanoes around the world and seem to correlate with volcanic activity.

To test this hypothesis, we measured attenuation around Mt Ruapehu between 1990 and 2005 using regional earthquakes originating from the nearby Waiouru earthquake swarm recorded on 5 permanent seismometer stations as the main data set. We analysed the similarity of events originating from the swarm by waveform comparison in order to ensure a constant source over time. Attenuation measured from direct compressional waves revealed a distinct change in attenuation at two stations that correlates with the beginning of the eruption period in late 1995. Attenuation of later wave arrivals provided stable results over the entire time period, yet we observe an increase in frequency dependence of attenuation during the eruption period.

Our study shows that changes in seismic attenuation can be related to changes in stress caused by volcanic eruptions and thus can be used as an additional method for volcano monitoring at Mt Ruapehu. However, measurements were complicated by insufficient data. Further studies of seismic properties in the vicinity of Mt Ruapehu volcano using improved and large data sets that are now provided by denser station coverage with 15 seismometers are strongly recommended. We found that the nearby Waiouru swarm offers a constant source of remarkably similar events and thus is especially suitable to monitor changes over time.

---

Technical Abstract

We measure seismic attenuation in order to test if temporal changes can be associated with changes in local stress around Mt Ruapehu Volcano caused by the 1995/96 eruptions, as has been observed for changes in seismic anisotropy (Miller and Savage, 2001; Gerst and Savage, 2004). Temporal variations of seismic attenuation have been observed at different volcanoes around the world and seem to correlate with volcanic activity. Attenuation around Mt Ruapehu is measured for the period October 1990 to May 2005 using regional earthquakes originating from the Waiouru earthquake swarm recorded on 5 permanent seismometer stations as main data set. Cross-correlation of P and S waves reveals a remarkable similarity of events from this ongoing seismic source and shows the presence of seismic families, which are especially suitable for the study of temporal changes. The attenuation (Qc-1) of coda waves following the direct P and S waves is calculated for five frequency bands centred at 1.5, 3.0, 9.0, and 12.0 Hz using the single-scattering model. A best-fit equation of Qc = Q0.fa with Q0 = 54 (±7) and a = 1.02 (±0.06) shows an overall high frequency dependence of coda attenuation, which is characteristic for tectonically active regions. a increases at all stations in 1995 and decreases slowly in the years following the 1995/96 eruptions. Small fluctuations of Qc-1 are observed at all stations over the entire time period and do not act as apparent indicator for volcanic activity. Relative change in integrated direct wave attenuation is obtained using a spectral-ratio method in which spectra of individual 1­­second P wave windows are compared to a reference spectrum. Results reveal a distinct change in the low-frequency band (1.5 – 6 Hz) at two stations in 1995. This observation correlates well with variations detected in seismic anisotropy studies, which suggest a model of the change in stress caused by pressurisation of a filling magma dyke under the volcano. A simple assumption of fluid-filled cracks in a region about 5 km broad that respond a varying stress field can explain both the anisotropy and attenuation changes. The absolute attenuation measured from direct waves is poorly constrained due to a lack of data. Nevertheless, estimates from direct P and S waves suggest high attenuation in the vicinity Mt Ruapehu, as expected due to the highly scattering shallow layers of recent volcanic deposits.

We recommend using seismic attenuation as an additional method for volcano monitoring because this study shows that changes in seismic attenuation can be related to changes in stress caused by volcanic eruptions. However, measurements were complicated by insufficient data. We strongly recommend further studies of seismic properties in the vicinity of Mt Ruapehu Volcano using improved and larger data sets that are now provided by denser station coverage with 15 3-component seismometers.