The seismo-acoustic dynamics of volcanic unrest and eruptions.
Volcanic systems are inherently complicated and therefore difficult to understand and forecast from monitored data. In an effort to understand volcanic processes, this thesis aims to estimate characteristics of varied seismo-acoustic signals during volcanic unrest and eruption at the southwest Pacif...
| Main Author: | |
|---|---|
| Format: | Other/Unknown Material |
| Language: | English |
| Published: |
University of Canterbury
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10092/102097 https://doi.org/10.26021/11152 |
| Summary: | Volcanic systems are inherently complicated and therefore difficult to understand and forecast from monitored data. In an effort to understand volcanic processes, this thesis aims to estimate characteristics of varied seismo-acoustic signals during volcanic unrest and eruption at the southwest Pacific volcanoes: Ngauruhoe and Whakaari/White Island in New Zealand and Ambae in Vanuatu. These active volcanoes have formed through plate subduction and resultant magmatic activity over centuries and have experienced a variety of eruptive and non-eruptive unrest that has been monitored with variably dense monitoring networks. Using three case studies, this thesis addresses two main issues: 1) seismic waveform classification and analysis, and 2) the relationship between subsurface and surface volcanic observations. Volcano seismology techniques have been applied for many years in New Zealand to monitor the volcanoes. However, systematic studies of long-term trends are ongoing and crucial for developing methodologies to forecast eruptions. Building upon the existing research, I focus on two case studies associated with classification of long-term seismicity at New Zealand volcanoes. I apply different waveform detection/classification approaches depending on signal types. At Ngauruhoe, numbers of low frequency (LF) earthquakes persistently occurred from 2005 to 2010, which marked a new unrest episode since the last eruption in 1975. The LF signals are detected by a short-term averaging/long term averaging algorithm and classified using the analysis of a cross matrix. Micro-earthquakes similar to the detected waveforms are then recognised through application of a master event technique. Detailed analyses of temporal variations suggest a relationship between LF seismicity and meltwater-magma interaction at Ngauruhoe. At White Island, I apply more systematic approaches to very long period (VLP) signals recorded from 2007 to 2019. Volcanic VLPs are detected by a waveform semblance technique and clustered based on Pearson ... |
|---|