Mislocated Earthquakes From Auckland Islands Earthquake

Minor Earthquake Strikes Auckland Islands: Mislocated Earthquakes on South Island - GeoNet

Mw 5.6, ~415 kilometres south of Invercargill, New Zealand, 13:04 UTC, 01/12/2014

The Mw 5.6 Auckland Islands Earthquake picked up on the New Zealand National Seismograph Network (bottom right of picture). The individual P-waves and S-waves can be seen here on the seismograms of stations such as Wellington or Denniston North.

At 13:04:23 UTC a moment magnitude 5.6 earthquake struck north-west of the Auckland Islands, about 400 kilometres south-south-east of New Zealand, at a depth of 32 kilometres. This earthquake had an oblique reverse faulting mechanism, something consistent with seismic events in this part of the world. The quake was caused by the relatively weak subduction of the Australian Plate beneath the Pacific Plate at a feature known as the Puysegur Trench. The earthquake was picked up by GeoNet, the national natural hazards programme run by the New Zealand government, and mislocated as two small local events on the South Island of New Zealand. In this article the processes behind these mislocations are explained and compared to similar instances in GeoNet’s monitoring.

At 13:05:21 UTC the GeoNet automatic detection system for seismic events computed a local magnitude 4.1 earthquake 10 kilometres north-east of Alexandra, South Island. This earthquake would be a very rare event, the McKenzie Country of western Otago Province generally proving aseismic, even with numerous faults being present in the region. The earthquake was given a likely depth of 15 kilometres, something compatible with previous earthquakes in the area.

About one minute later (13:06:26 UTC) a local magnitude 3.9 earthquake was computed by the automatic detection system about 25km north-east of Murchison, South Island. This earthquake would be a relatively unusual event, not owing to its location, where both strike-slip events, associated with the Marlborough Fault System, and reverse subduction events, associated with the Kaikoura Ranges (the only onshore imprint of the Hikurangi Trench subduction zone) occur. However this earthquake was given a focal depth of 142 kilometres, something more commonly associated with subduction events in the western sections of the Cook Straits or the Taranaki Region of the North Island.

The Auckland Islands Earthquake as seen at the nearest seismograph on the New Zealand National Seismograph Network to the focus - The Paps, on Stewart Island, south of Invercargill. This shows a very clear spike at the bottom left; note the S-waves cannot easily be picked out from the P-waves.

These two automatic detection earthquakes are what have come to be known as “ghost quakes”. When a distant earthquake is detected by seismograms in New Zealand, GeoNet’s automatic detection system attempts to compute a local event as the cause of the arriving waves. Unfortunately the arriving seismic waves are often more distinct than those arriving from a local event. Therefore two separate waves are detected by GeoNet’s seismograph stations – the initial P-waves, followed by the S-waves. The difference between these two sets of waves means that the further away from the true focus of the earthquake a second “S-wave” earthquake is located – the S-wave is far easier determined at these more distant seismograph stations than at stations closer to the source (an example is seen below).

The seismograph at Top House (upper South Island, New Zealand) shows the Auckland Islands Earthquake (left of picture, near 4-hour line); the waveform recorded shows clear P-waves followed by weaker S-waves. The other clear spike (bottom centre, near 2-hour line) is a local magnitude 4.5 aftershock to the Mw 6.7 Hikurangi Earthquake of the 17th November 2014.

In the case today this involved a distant moment magnitude 5.6 earthquake located north-west of the Auckland Islands within the Puysegur Trench system, about 415 kilometres south-south-west of Invercargill. This meant, as the seismic waves were detected further and further to the north-north-east from the focus the waveform recorded became more and more deformed and elongated.

At The Paps Seismograph (north-west of Oban, Stewart Island) the waveform recorded was stretched, but the P-wave and S-wave was not easily distinguishable. This is indicative of a sizable earthquake but at a reasonable distance from the seismograph station – in this case about 400 kilometres away.

At Top House Seismograph (north-east of St Arnaud, South Island) the waveform recorded was incredibly stretched, the P-wave and S-wave easily distinguishable from one another. This sort of waveform is representative of the seismograph station being a great distance from the epicentre – in this case about 1070 kilometres away.

The easily distinguishable S-wave at Top House & nearby seismograph stations in the north part of the South Island led to the automatic detection system to compute an earthquake in this region. Due to the highly deformed waveform it was also computed that the earthquake must have been quite deep. The inability to distinguish the P-waves and S-waves in the lower South Island meant that a shallower earthquake was computed in this region.

Generally speaking this sort of issue is not associated with Puysegur Trench earthquakes but rather Kermadec Trench earthquakes. Numerous examples have been noted in the past by myself, including a moment magnitude 5.6 earthquake which struck the Kermadec Arc on the 21^st September 2013. More distant events of significant magnitudes (such as the moment magnitude 6.9 event on the 1^st November 2014) are also often given “ghost quakes” despite this origin being over 1200 kilometres to the north of New Zealand. The bias here is due to the tectonic setting of New Zealand and the simple geography of the south and north of New Zealand.

The two "ghost" quakes computed by GeoNet, as seen on the quake list on the GeoNet website at about 13:26 UTC. These earthquakes will be subsequently deleted come morning on the 2nd December NZDT.

The Puysegur Trench is an incipient subduction zone which acts as a transfer zone on the Australian-Pacific Plate Boundary. Unlike the Kermadec-Tonga Trench system to the north, the Puysegur Trench does not appear to cause earthquake with focal depths of more than 120 kilometres – these events are generally restricted to the Fiordland Region of New Zealand’s South Island. The Puysegur Trench connects the 600 kilometres long Alpine Fault in the north to the 800 kilometres long Macquarie Fault Zone. Large events have occurred along this part of the greater plate boundary, but they are infrequent. This reflects general seismic activity levels.

The Kermadec Trench takes the majority of the subduction element of the collision of the Pacific Plate and Australian Plate to the north of New Zealand. Here earthquake are incredibly common, perhaps not helped by the parallel Havre back-arc basin which accounts for the extension within the overlying Australian Plate (as it bends and flexes due to the stresses exerted by the subducting Pacific Plate). Earthquakes here can reach depths in excess of 500 kilometres, whereas further to the north nearer Tonga and Fiji earthquake focal depths reach the greatest found globally at around 700 kilometres (only events under the Sea of Okhotsk, part of the Kuril-Kamchatka Trench, can match these depths).

Thus I have recorded such an example of a rare Puysegur Trench earthquake, distant from New Zealand, recorded by GeoNet, and how the issues with the automatic detection system are very similar in their nature to those faced by the system after distant Kermadec Trench earthquakes.

Sources

13:04 Auckland Islands Earthquake: http://geofon.gfz-potsdam.de/data/alerts/2014/gfz2014xnfx/mt.txt

13:05 Alexandra “earthquake”: geonet.org.nz/quakes/region/newzealand/2014p904176

13:06 Murchison “earthquake”: geonet.org.nz/quakes/region/newzealand/2014p904178

Link to previous article on 21^st September 2013 Mw 5.6 Kermadec Earthquake: http://sincalquakes.wordpress.com/2013/09/22/interpreting-a-significant-kermadec-quake/