Aftershocks, triggered earthquakes, and Christchurch’s seismic future

A post by Chris RowanAs more scientific information becomes available regarding last week’s magnitude 6.3 earthquake in Christchurch, we can look a bit more closely at the nature of this earthquake, how it fits into the overall tectonic picture in New Zealand, and future seismic risks in the region.

Aftershock, or triggered earthquake?

There seems to have been a bit of ambiguity in discussions about the nature of this week’s earthquake: was it an aftershock of September’s earthquake? Was it a separate earthquake that was possibly triggered by September’s earthquake? What’s the difference, anyway?

When any fault ruptures in an earthquake, movement along the fault plane itself also stresses and deforms the the surrounding crust. These stress changes can often induce further smaller earthquakes – aftershocks – that take the form of a cloud of tremors that encompasses the initial rupture. In other words, aftershocks are mainly caused by the stress added to the crust by the initial earthquake. However, very few faults are found in glorious isolation within the crust. More often that not, other faults, that may have themselves accumulated a large amount of tectonic stress over the previous decades and centuries, will be found nearby. In such cases, the stress change due to the first rupture may be large enough to push these nearby faults over the edge, resulting in a triggered earthquake. The difference between this and an aftershock is that the stress added by the first earthquake is only a small component of the stress released when the second fault fails – most of it was already present, and the fault would have ruptured at some point in the future anyway.

The difference between aftershocks and triggered earthquakes

The difference between aftershocks and triggered earthquakes.

Where things get complicated is in regions where major faults are quite close to each other – perhaps segments of a large fault boundary fault, or different strands of the same fault system – so that the cloud of aftershocks from a large earthquake on one fault overlaps with the trace of another fault that hasn’t ruptured yet. If that second fault then does rupture, what do we call that? It’s in the same region as all the other aftershocks; but most of the stress being released was not a result of the first earthquake – it was already there on the fault, and the additional loading was just the straw that broke the camel’s back.

When aftershocks and triggered ruptures overlap

When major faults are closely spaced, things can get complicated.

This weeks’ events in New Zealand seem to fall into this ‘it’s complicated’ category. Last September’s Darfield earthquake was centred 40 km to the west of Christchurch, but the epicentres of the aftershocks in the months that followed gradually migrated east towards the city. The map below (courtesy of Geonet) shows that these smaller aftershocks – the green circles – do encompass the area containing the fault which ruptured on Tuesday. So it could – technically – be described as an aftershock. However, for reasons that I discuss below, it seems probable (to me, anyway) that this fault is releasing tectonic stress built up over geological timescales – centuries or more – which would also make this a triggered earthquake. If that is the case, simply referring to it as an aftershock obscures its wider tectonic significance.

Christchurch Earthquake Sequence

All earthquakes in the Christchurch region since September 2010. Earthquakes since 22 Feb 2011 in red. Click to enlarge. Source: Geonet.

The big tectonic picture

And what is that wider tectonic significance? In my discussion of last September’s magnitude 7 earthquake, I wrote the following:

…the occurrence of such earthquakes in this particular region of the South Island is probably also linked to ongoing changes in the nature of the plate boundary at the junction between the subduction zone [off the East Coast of the North Island and the continental transform [The Alpine and Marlborough Faults]. If you look at the displacement history of the individual faults in the Marlborough Fault zone, the northern faults are older, were more active in the geological past, and have quite small recent (in the geological sense of ‘the last few 100,000 years’) displacements; the southern faults are younger, and have much larger recent displacements. The most obvious explanation for these changes is that the most northern of the Marlborough faults was originally directly linked with the end of the subduction zone, but that these two structures moved out of alignment as the subduction zone moved south, causing new strands of the Marlborough Fault system to grow in order to more efficiently accommodate plate motions.

This tectonic evolution is ongoing, and since the end of the subduction zone is now actually to the south of the southernmost and youngest of the Marlborough faults. Some of the plate boundary deformation is probably therefore being shunted into the region around Christchurch, where it needs to be accommodated by dextral strike-slip faulting. Eventually, over geological time, this deformation will lead to the formation of a new, more southerly strand of the Marlborough Fault system.

Growth of new plate boundary faults on the South Island of New Zealand in response to southward propagation of the subduction zone

Looking at the traces of the Darfield Fault, and the fault that ruptured this week, mapped out the Geonet aftershock plot above, it looks to me like two strands of the same, mainly strike-slip, fault system, which is exactly what you would expect for the early stages of a new branch of the Marlborough Fault. Over geological time, then, you would expect to see these faults developing into a much more prominent part of the plate boundary zone. But that’s over the next couple of million years. What about the next few decades?

The future seismic risks for Christchurch

People in Christchurch are unsurprisingly feeling a little overwhelmed by the recent seismic chaos, and fearful for the future. Now that New Zealand geologists have been made aware of the active faults running across the Canterbury Plains, they will be racing to study them and their past behaviour, in order to assess how much risk they pose to Christchurch in the decades and centuries ahead. But the very fact that these faults were unknown actually provides us with some information about them – namely, that although they are active, they have not been particularly active in the recent past. Earthquake-generating faults are identified from the historical record – quite short in the case of New Zealand – and looking for features of the landscape that indicate fault motion, such as scarps and uplifted terraces. The fact that these faults don’t seem to have generated such features, and have instead managed to be totally buried beneath the river-borne debris coming off the uplifted Southern Alps to the west, tells us that these faults do not rupture particularly frequently; if they were, the Canterbury Plains would not be so flat.

The Canterbury Plains

Not a particularly tectonically shaped landscape (mountains not included, obviously)

This makes sense: we know how much motion occurs across the plate boundary that bisects New Zealand, and we know that the motion on the main boundary faults of the South Island – the Alpine and Marlborough Faults – accounts for around 80% of that motion. The remainder is probably distributed across many faults on the South Island, not just the ones near Christchurch, which means that they need much longer to build up enough stress to rupture in an earthquake – probably the high hundreds, or low thousands of years. It would be unwise to relax before some detailed geological work is done, of course, but I suspect that these particular faults have done all the damage that they are going to do to Christchurch for the foreseeable future.

Categories: earthquakes, geohazards, society, tectonics
Tags: , , , , , ,

Comments (10)

  1. Ron Schott says:

    I’m inclined to agree that these quakes have probably relieved the strain buildup across this particular set of faults for the foreseeable future, but I’m a bit less sanguine that that recurrence interval is in the “high hundreds, or low thousands of years” based on the topographic evidence from the Canterbury Plains. For one thing, the most recent quake appears to be on a blind fault that didn’t rupture the surface (I only hesitate to call it a blind thrust because of its highly oblique slip direction). Also, the Darfield quake last September was nearly pure strike slip, so even thought it did leave a scarp, little or no vertical displacement of the Plain would be expected from previous quakes on this fault, presuming similar paleo-focal mechanisms. Furthermore, its orientation was generally parallel with drainages in the region – and certainly not at high angles to them (which would be expected to lead to telltale dogleg stream channels). Finally, I suspect (and this is more speculative) that these relatively youthful braided river systems that drain the Alps and supply sediment to the Plain probably do a fairly effective job of erasing (by erosion and/or deposition) what little topographic evidence might have been left by the previous Darfield-type quake. If these observations stand up to scientific examination, then it’s possible earthquakes on the Darfield-Christchruch fault system do not leave much of a noticable, lasting surface expression and thus may occur more frequently than the lack of such evidence elsewhere might suggest.

  2. Kurt Scholz says:

    I was also thinking similar thoughts about the plains river system and the type of slip the region seems to be experiencing. I would have thought that the liquafaction experienced would tend to suggest that the quakes may leave limited evidence over time, of having occured.

  3. I am a theoretical physicist at U Canterbury, and live in Cashmere in the Port Hills near the middle of the February 22 buried fault trace as shown the GNS map you reproduce, in an area with serious structural damage to buildings, surface cracks and collapsed retaining walls. (The 5.7 magnitude aftershock 15 minutes after the 6.3 one, with an epicentre under our hill, is the one that did most damage locally.) The worst damage is at the end of the street (my end unfortunately) near the bottom of the hill, just above the plain.

    I first started thinking about the questions discussed in your post about a month after the September 4 event, while driving from home to Lincoln on the plains southwest and observing the pattern of surface cracks and other damage around Cashmere Road and the area between Halswell and Tai Tapu. Being a physicist, I try to think in terms of simple mechanical models, and as my thoughts based on the observed damage pattern last year seem confirmed by recent events, I would like to run them past those less geologically naive than me.

    My basic post-September 4 question is how does Banks Peninsula interact with the main movement on the Alpine fault? Banks Peninsula is a former volcanic island that was here before the plains existed in their present form. The Alpine fault is dragging the Canterbury plains southwest at a few centimetres a year. My intuitive guess is that the mass of the basalt of Banks Peninsula resists this, so we get a build up of strain on any fault systems parallel to where the general southwesterly movement of our bit of the Pacific plate hits the change in density from the sediments of the plains to the basalt of the Port Hills. There is probably also some twisting of Banks Peninsula in response to the general southwest movement of the plains. Actually, I recall seeing a GPS determined map of the velocity field of annual surface movement many years ago, which I think showed some rotation of Banks Peninsula – but I don’t know where to find that map now.

    My hypothesis then is that the “main” Greendale fault is then just the middle segment of a larger fault system extending to the Alps, the eastern part of which we are sitting on in the Christchurch hill suburbs. On September 4 it gave way first where it was weakest: in plains sediment at a location halfway between the rock mass of the Southern Alps on the one hand, and the rock mass of Banks Peninsula on the other. These two rock masses define the boundaries of the system.

    If you play “Christchurch earthquake map” http://www.christchurchquakemap.co.nz/ with “sticky dots” for 1 month from 4 September then the aftershock pattern is consistent with the presence of this larger fault system that extends east of the Greendale fault from Rolleston, hits the Port Hills between Halswell and Tai Tapu and then traces the northern edge of Banks Peninsula out to sea. What is striking from the 1-month “sticky dot map” is that the number of epicentres in the first month is very thick around the main fault and its extension as long as it is on the plains. As soon as the aftershock line hits the Port Hills it is still very distinctive, but also much thinner than on the plains. As a physicist, my intuitive guess would be that the mass of the basalt of the Port Hills damped what would have otherwise been the natural aftershock pattern from the main September 4 event in the region where the underlying fault meets the Port Hills. In other words, the “sticky dot map” of 1st month aftershock epicentres would have been equally thick in the Port Hills segment as on the plains if that area was composed of alluvial sediments rather than basalt. This is consistent with the amount of aftershock energy initially released being a bit on the low side as compared to some statistical expectations. Instead of an initial release of energy then we have had a further building up of the strain on the under-Port Hills segment of the fault system, until it gave.

    The GNS diagram http://all-geo.org/highlyallochthonous/wp-content/uploads/2011/02/Christchurcheqs.jpg of the two fault traces you have reproduced also lends some support to this idea. If you follow the yellow dotted line of the February 22 fault trace you will see that it extends into a green line of pre 22-February aftershocks which is very thick across the plain. Rather than there being a single east-west fault that joins the Greendale fault to the Port Hills fault it appears that the Port Hills fault runs a little more northeast-east – southwest-west, but that it extends out in the plains to roughly the area around Springston and Ellesmere, and then there is a northward jump to the western edge of the Greendale fault near Rolleston. From the density of green pre-22 February aftershock, it actually looks as if this jump is a small fault which runs north-northeast – south-southwest, i.e. parallel to the two other sub-surface ruptures which branch off the Greendale fault in the GNS map. I.e., rather than talking about one fault, I would guess that we are talking about an inter-connected fault system with different elements – the small branch segments parallel to the main Alpine fault, and the onger Greendale and Port Hills faults angled (not exactly orthogonal) to them. I think that the pattern of this fault system is still consistent with a basic physical mechanism that the resistance of the dense basalt of Banks Peninsula to the southwesterly movement on the Alpine fault is what is leading to the build up of stresses on the Port Hills and Greendale faults.

    In regard to the future seismic history of Christchurch, then in the immediate future we should perhaps watch out for a very substantial aftershock on the part of the Port Hills fault that extends onto the plain which has not ruptured according to the GNS map – i.e., I would watch out for one with an epicentre near Lincoln, or on the spur between Ellesmere and Rolleston. However, the density of pre-February 22 aftershock in these regions is the largest anywhere on the map. So it could be most likely that the stresses in these areas have already sorted themselves out, and the worst is over.

    • John Delano says:

      “You listened to Albert Einstein.”
      His comment was if you want to learn something about an object ,”look at it.”

      Difference between Christchurch and Manhattan Island- Where you have tall buildings you hve “bedrock.”
      Where you have gravel and swampland you r foundation is not so good.
      Christchurch has a bad foundation.

      Your comment about the strange shape and reason for lyttelton harbor is my point that you looked at the thing. The “blownout Caldera wall” is going on in Iceland as we speak. One of its largest volcanoes [under 1,000's of feet of ice ] is blowing out its wall with a huge volume of escaping melt water. The same for Lyttelton harbor.

      http//www.Christchurchquakemap.co nz/ with “stickey dots” is ahuge discovery, a fire under the study of the geology “pot of learning.”
      A “Pot ” because so many stange ununderstandable geology terms ,names maker the science a “pot” of memorizing stuff.
      You in your unrelated field are a genius, with a fire burning. I only hope your sucess will be greater than Albert Einsteins. In 1905 he made his discovery, he died in 1955 at age 75. Most of his photographs are of “The old Guy”, and he was not recognized as a genius for many years.
      Yes the gravels of the canterbury plains are at there “angle of repose” and no surface faults show. What is hidden due to the liquefaction and the gravel is the “rebound” -The vertical movement of the land.
      Review my other comments, while I tell you “we stand on the sholders of giants.”
      Google Earth, there “tunneling tool”, There Latitude and Longitude placements of earthquakes on the Google Earth map.
      Listen to this- Using the Tunneling tool confirm this “neat example.”
      Take a flashlite and shine it through a cleal plastic globe, and look at the shadow on a darkened surface. The “Boot”of Italy is where you chine the lite through. On the ceiling -are two boots “side by side” as if the hunter took them off and placed them in the corner.
      New Zealand is the other boot!
      Why.
      The countries 2-9 million years ago were being pushed up from the ocean bottom as they sat near the earths equator by massive ice cap/and ice shell flows at the same time.
      E-mail me and you will receive my book -free “The origin of Mountains.”
      sunnyday1@optonline.net

  4. Georg Winkler says:

    Hi Ron,

    I can’t help but observe that the geomorphology of the Lyttelton Harbour, in particular its orientation, is strikingly similar to the orientation of the earthquakes associated with the Feb 22 event. If the harbour is simply an erosion feature with no other influences then I would have expected that it would have formed a more radial orientation from the centre of the cone, much like almost all of the other valleys that disect the Banks Peninsula. Instead it cuts right across these at quite an odd angle suggesting that something else has controlled its formation. I am not a seismologist, but is there strong evidence for a south dipping fault beneath the Port Hills as reported by GNS, or is there a possibility that it could actually be north dipping (like the Greendale Fault), with a surface fault trace in the Lyttelton Harbour? This would also result in the reported uplift of the Port Hills, and may also explain why a fault trace has not been found (or has it?).

    Interested to hear your comments.

  5. Georg,

    Lyttleton Harbour is not simply an “erosion feature”, and there is no “centre of the cone” on Banks Pensinsula. Rather both Lyttleton Harbour to the north and Akaroa Harbour to the south are the separate eroded dominant craters of the Miocene volcanoes which formed Banks Peninsula, and which were active at different times. Cutting and pasting from Wikipedia:

    Geologically, Banks Peninsula comprises the eroded remnants of two large composite shield volcanoes (Lyttelton formed first, then Akaroa). These formed due to intraplate volcanism between approximately eleven and eight million years ago (Miocene) on a continental crust. The peninsula formed as offshore islands, with the volcanoes reaching to about 1,500 m above sea level.

    The craters eventally eroded so that the sea flowed in from the eastern end in the case of Lyttleton harbour, and from the southern end in the case of Akaroa harbour. Lyttleton harbour is a little bit wider than Akaroa harbour, as the original crater has had somewhat longer to erode.

    No surface fault trace has been found for the Port Hills fault. (There is a lot of basalt above much of it.) The closest that the rupture plane gets to the surface is at the eastern end below the Avon/Heathcote estuary where it is about 1-2 km deep, as is explained in this video by Jarg Pettinga, Head of Geological Sciences at U Canterbury:

    http://www.youtube.com/watch?v=nTvFDSpmh3A

    You are correct that the fault trace is aligned to the harbour; and that is simply because the fault trace coincides roughly with the northern edge of the basalt, which is aligned with the crater that originally formed it. It is this alignment (which was already evident from the aftershock patterns within just one month of the September 4 event) which has led me to question whether Banks Peninsula plays a significant role in the build-up of stresses in the combined Greendale – Port Hills fault system, as discussed in my post above.

    Cheers, David

  6. I am inclined to suspect that the 6.3 Christchurch earthquake is more related to the Banks peninsula trying to achieve isostasis with its geological environment following millenia of erosion, making the volcano lighter while the surrounding land has been getting heavier due the outwash from both the volcano and the Alps to the West. It just happened that the Sept4 fault line pointed directly at the Northern edge of the volcano and weakened the rock there until it allowed the banks peninsula to lurch upwards.
    http://mtkass.blogspot.com/2011/03/christchurch-earthquake.html

  7. former JAFA says:

    As a geologist I disagree that you would see escarpments and tilted fault blocks as signs of earlier fault movement above the blind faults — *in this particular topographic area*. This classic indicators of faults and previous motion are reliable when you have hard rock, but look at the muddy swamp, with a very slightly congealed top, that is the sediment upon which the people of Christchurch have unwittingly built their homes. Look at this video at about 1:10 minutes in: http://www.stuff.co.nz/national/christchurch-earthquake/5145310/Neighbours-sticking-it-out That man can sink his shovel handle all the way in the ground — with almost no resistance! How are you ever going to see tilted fault blocks in liquefied land that hasn’t solidified into rock? I believe these faults have been active previously but with this type of “jello” sediment poured over them, how would you ever know? With the chance of a 7.0 now at 30% in the next 12 months, I’d be moving the hell away from (at a minimum) the eastern suburbs of Christchurch, govt support or not. How large will the mud volcanoes be if there is a 7.0 on the Port Hills fault? Could the resulting mud volcanoes swallow whole buildings? No one can answer, but again, I’d be using my common sense and getting the hell away from such unstable ground that is not going to solidify again (at least not in human life-span terms) – it’s too shook up & full of water.

    • Chris Rowan says:

      You’re slightly conflating the long-term risks after the current sequence ends (what I was talking about) with the short-term risks as the region adjusts to the stress released in the Darfield earthquake.

      It’s true that landscape is a complex interaction between several factors, including tectonics and lithology. But as you ramp up the deformation rate, the tectonic influence on the landscape will eventually show up – you only have to go up the coast to Kaikaoura to see that . So I still think that the lack of any discernable tectonic features on the Canterbury plains places a maximum limit on the repeat rate of these faults. I am thinking, though, that the widespread liquefaction and associated subsidence that seem to result from earthquakes in this area should also show up in the geological record – I wonder if anyone has looked into that?

  8. john delano says:

    Great presentation-there is no doubt in my mine that the land under Christchurch is “Rebounding” from the 2 and 1/2 mile high glacier that melted 9,600 years ago.

    Christchurch would have been 300 feet below sealevel with the 2and 1/2 mile high on the land. When 9,600 years ago all the glacial ice melted the sea-level rose 320 feet.
    Christchurch is only 23 feet above sea-level now ,but it is still rising -”rebounding” from theabsence of the huge load on the land.

    Thats what is happening- watch the land rise and crack the sediment under Christchurch. I will now check why the GPS is not recording this rising land.
    If you can shed lite of the land under Christchurch rising please leave a comment.

    Links (4)
  1. Pingback: Magnitude 6.3 earthquake rocks Christchurch | Highly Allochthonous

  2. Pingback: Aftershocks in Christchurch « The Trembling Earth

  3. Pingback: Aftershocks of the Sendai earthquake | Highly Allochthonous

  4. Pingback: A fresh temblor in the Canterbury sequence « The Trembling Earth