Katrina Jacobs (supervised by Martha Savage), Victoria University of Wellington (EQC funded project 14/U698)
Slow slip events (SSEs) occur regularly on the Hikurangi subduction zone along the East coast of the North Island. These SSEs sometimes trigger small local earthquakes, usually offshore.
To study the general influence SSEs have on seismicity rates in New Zealand, we did a statistical analysis of earthquake rates comparing times when SSEs were occurring to all other times. We found that although seismicity rates remain relatively low, for most regions there is a significant increase in the number of earthquakes that occur during SSEs. This shows that while the changes associated with SSEs are too small and too slow to be felt, they are large enough to influence tectonic processes and earthquake rates. Future New Zealand hazard models that take into account time variations could incorporate this information.
Given the connection to earthquake rates, we have investigated whether these SSEs affect the timing of volcanic activity at Taupo caldera beneath Lake Taupo. In 2008 a small SSE was accompanied by a large increase in seismicity and gas release at Taupo caldera. Despite other similar and larger SSEs in the area, no other accompanying volcanic unrest has been observed at Taupo in association with these events. By calculating strain from GPS instruments around Lake Taupo we have shown that the 2008 SSE was unique in that it produced dilational strain, which acted to stretch out the caldera. All other SSEs examined appear to have the opposite effect, pushing the caldera together.
We also looked for changes in seismic velocity associated with these SSEs, but have not found any links. The seismic velocity does vary with time and some of these changes may be seasonal. We suggest that the 2008 SSE pulled apart the caldera enough to allow existing gases to rise to the surface and be released. In the future increases in extension at the caldera should trigger increased monitoring and scrutiny of other volcanic parameters.
An increasing number of studies globally are recognising that tectonic processes can have a major influence on volcano behaviour. This has been most striking in studies of triggered seismicity following large regional and global earthquakes. Slow slip events (SSEs) have been recognized more recently to trigger tremor and small local seismicity.
We have investigated the influence that regular SSEs along the Hikurangi margin have on the timing of seismicity in New Zealand, and volcanic unrest at Taupo caldera. This study addresses a time and spatial scale often over looked in other triggering studies. A systematic statistical analysis was carried out to compare seismicity rates during SSEs to times when they are not occurring. We found that six out of nine regions studied in the North Island has significant increases in seismicity during SSEs. This information should inform future time-‐ varying hazard models currently being developed.
In 2008 a small SSE was associated with the largest increase in seismicity in the last decade. Gas release was also recorded during this time. To investigate the effects of SSE induced strain at Taupo caldera we analysed areal strain for the 2008 SSE, a similar SSE in 2006, and a larger, more distant, SSE in 2010. Areal strain was first calculated from cGPS data. Then we compared these strain observations to strains predicted from a finite element model using published slip distributions for the SSEs. The observed and modelled strains match reasonable well directionally, but the modelled strains are an order of magnitude smaller than the observations. The directional match indicates that large dilational strain at Taupo during the 2008 SSE could be tectonic, and does not require an inflating volcanic source. The 2008 SSE is the only one which shows dilational strain, and we suggest that is what allowed gas release and unrest. Other SSEs have not been associated with any unrest, but have had almost purely contractional strain.
Ambient noise analysis was also carried out to look for seismic velocity changes associated with SSEs and strain changes. However, time variations were not found to be useful as background fluctuations in relative velocity were on the order of 0.1% and are too large to confidently identify smaller fluctuations. We propose that GPS baselines be monitored as a proxy for strain, and increases in baseline length across the caldera should initiate increased monitoring protocols.