Shaking in Christchurch boosted by seismic lensing?

A post by Chris RowanEven taking into account how close the rupture point of Tuesday’s earthquake was to Christchurch, the intensity of the shaking – and the amount of damage that the city suffered as a consequence – seems to be very high for a magnitude 6.3 earthquake. The fact that the city is built on soft sediments that amplify shaking is an obvious factor here, but an article in the New Zealand Herald raises the possibility that geological structures in the region may have acted as a ‘seismic lens’, focussing the seismic energy released in the earthquake towards Christchurch.

If we plot the earthquake’s epicentre on a geological map of the central South Island of New Zealand (courtesy of this handy interactive map from GNS), we can see it is close to the edge of the Banks Peninsula – the eroded remnants of two basaltic shield volcanoes that were active around 10 million years ago.

Location of Feb 22 earthquake plotted on a geological map of the Christchurch region, showing the Miocene volcanics of the Banks Peninsula (pink) rising above the young clays, silts and gravels of the Canterbury Plains (tan). Source: GNS

The hard volcanic rocks that make up the Banks Peninsula are very different from the soft sediments that underlie Christchurch. The huge contrast in physical properties across the boundary between these two units means that any seismic waves hitting the boundary would mostly be reflected back the way they came. So, instead of spreading out equally in all directions, a large proportion of the seismic energy released in a nearby earthquake would end up being sent away from the boundary with the basalts – which, unfortunately, is towards Christchurch.

How seismic lensing might concentrate seismic energy and boost shaking intensity.

As it stands, this idea is rather speculative – more work will be required to pin down the location of the fault before it can be proven one way or the other. But if seismic lensing was a factor in the very intense shaking Christchurch endured, the location of this week’s earthquake – in an ideal position to produce this effect – was doubly unfortunate.

Update Callan requested a shakemap that includes the Banks Peninsula. This is from GNS’ Geonet site, and included felt earthquake reports (circles) and instrumental peak ground accelerations (squares). As you can see, there is not much useful coverage on the Banks Peninsula, and although what is there could be interpreted as showing less shaking at an equivalent distance from the epicentre, this does not account for the effects of being on hard rock (basalt) rather than soft sediment, which will probably amplify the shaking even in the absence of other factors.

Shakemap for the Feb 22 earthquake (star)

Categories: earthquakes, geohazards, geophysics
Tags: , , ,

Comments (17)

  1. Mike says:

    I read your first post on the earthquake and did some thinking along similar lines.
    I’m a hydrographic surveyor so thought about it in terms of multibeam acoustics.

    /cast copypaste
    Reading in more detail and looking at peak acceleration plotted on the map (plus a bit of knowledge about beam forming) the hard volcanic layer under the Port Hills appears to have channeled seismic waves from the fault rupture horizontally. These horizontally focused seismic waves generated the strong back and forth motion and the east-west nature of the hard volcanic layer under the hills generated lobes of great intensity as reflected seismic waves interacted. From a rough look at the map and a few assumptions on the nature of the hard layer (based on overlying Port Hills topography) a lobe would focused northward up towards Horseshoe lake area (which had a lot of liquefaction) and secondary lobes through the CBD and under the estuary and a tertiary lobe out towards Hornby. This is supported by the layout and intensities of measured peak acceleration. Not that I’m a seismologist or geologist but the idea checks out when run through my geek/bullshit filters…

  2. Lab Lemming says:

    But wasn’t shaking also bad in the hills?

  3. Do we have a shake map that includes the Banks Peninsula?

  4. Matt Hall says:

    Fascinating story, nicely written. I don’t know much about earthquake seismology, but it seems like it would be fairly easy to model, given the subsurface geology. The reflection off the hard surface would be the same polarity as the incident wave I suppose. You say ‘most’ of the energy would have been reflected; I don’t know the details of the rocks involved but the reflection coefficient (proportion of incident energy reflected) is probably lower than 0.5 is it not?

    As the reflecting waves interact with non-reflected ones, could there even have been an interference pattern in shaking intensity, so that badly shaken neighbourhoods would be interspersed with much less badly damaged ones?

    Lastly, I wonder if ‘lensing’ is the right description or metaphor? A lens works by refracting rays to focus energy, whereas the phenomenon you seem to be writing about is reflection. You can focus energy with reflections too, of course, but only with a concave geometry. Just a thought.


    • Chris Rowan says:

      I agree that there are some problems with the ‘lensing’ description, but it’s the one that seems to be out there in the media so I thought I should use it. ‘Focussing’ might be more accurate.

      Interference between the original and the slightly lagging reflected waves probably would boost the amplitude of shaking in some places and weaken it in others. The original article in the Herald seems to touch on this concept as an explanation for why certain buildings seemed to have suffered much more than other nearby ones.

  5. Chip says:

    Makes sense to me. Looking at the map in light of the article it seems the peninsula inner edge of the basalt looks like a convex mirror and the squares seem placed right where you’d expect standing waves to be produced as the shock wave and the reflections collided. Nice article and I think it may shed more light on other earthquake patterns if they are also examined in this way.

  6. Kim says:

    I seem to remember similar discussions after the Northridge earthquake (1994) – it also had peak ground accelerations higher than gravity (though not as high as 188% – yikes!).

  7. Nick says:

    Thanks for these interesting backgrounders on the quake.

    What do we know about Banks Peninsula volcanics? I read they are in the 8-11 million y.o area. Are the edges of this volcanic ‘dollop’ sitting over older consolidated outwash from the mountains and /or the older sedimentary mix that makes up the NZ Alps? Is the epicentre at c.5km deep under Lyttleton in the sediments/deposits overlain by the volcanics?

  8. Helgi says:

    In your seismic-lensing diagram; shouldn’t the epicentre be located in the basalt not recent sediments? If so, does the ‘lensing’ model still hang together?

  9. EdK says:

    Wasn’t something similar suggested for the Port-au-Prince quake?

    Helgi, the epicenter of an earthquake is the point on the surface above the initial rupture point on a fault. Seismic lensing would not change that.

  10. The Geonet shaking map circles reflect reports from people who felt the quake. Banks Peninsula is sparsely populated, so there will be few circles regardless of intensity. Then there’s then question of who had power and also the inclination to go online (around 20% of then region is still powerless). A lot of the Banks farmers have also come into the city to shovel silt, or bring food – wonderful people.

    It’s a useful map, but not exactly perfect :)

  11. John says:

    Mike, I think there is insufficient data and too many possible factors to be able to sort things out into specific lobes. Firstly the local material and liquifaction effects are going to make a big difference to the intensity and the effects at a particular location, and secondly with an earthquake we are dealing with two different types of waves with different propagation speeds. The latter means they will have different wavelengths, which is going to affect the interference pattern seen. I suspect that in your acoustics, as with the radio antennaes that I have had experience with, you are dealing with reasonably pure sine wave signals of a single frequency, and I would not think that the signals from earthquakes would be quite so pure and easy to deal with.

    That does not mean that the hypothesis is not reasonable. I had a discussion a year or so back with one of the NZ GNS scientists about the possibility of beamforming taking place with Tsunami waves like the event near Samoa. Since the undersea fault movement is generally a linear feature, it seems quite reasonable that it might form a pair of lobes out each side rather like a dipole antennae. Given that such waves take some hours to cross oceans, it might even be possible to use early information about the movement to predict the likely intensity of the waves when they reach the other side of the ocean.

  12. Helgi says:

    Thanks for the epicentre definition. Just wondering, why has GNS mapped the epicentre above Basalt; when the Lensing diagram locates the epicentre at depth, within recent sediments, away from the basalt.

  13. Rod says:

    The Miocene volcanoes of Banks Peninsula sit on a basement high of greywacke. This is the western end of the Chatham Rise. We are assuming that the movement took place on a fault within the basement, not within the Miocene volcanics. So if the basalt lies above the hypocentre (considered to be 3-5 km below the surface), the seismic energy in a lensing model should have been reflected back downwards? I think it is much more complicated than seismic lensing.

  14. Chris says:

    Faszinating story. And it sounds logical for me.
    Have these two volcanoes been islands in front of New Zealand back when they where active? And the land between the mountains and the old volcanoes is just sediment?

    • Rod says:

      As far as we know, the Miocene stratovolcanoes were erupted subaerially. They are likely to have been islands throughout their history. The basement greywacke high appears to terminate abruptly immediately to the west of Banks Peninsula, but at depth, (prob. around 750 m), greywacke basement is still connected to the mountains. Of course, the Canterbury gravels don’t sit directly on the greywacke. There are some Cretaceous to Eocene (140-55 m.y.) rocks in between. Some of these are exposed on Banks Peninsula.

    Links (1)
  1. Pingback: Magnitude 6.3 earthquake rocks Christchurch | Highly Allochthonous