Tip:
Highlight text to annotate it
X
Yes, thank you everybody
And thank you first of all for coming here
Many thanks, special thanks to the BOB committee,
the Best of Bristol lectures committee, for organising this thing
and giving us opportunity, and especially for arranging this
in these halls of residence. It's quite nice to combine the place where you live
and put some lectures in here - little bit of research and a little bit
of interesting topics. And finally, a big thanks from both me and Mike for
voting us to take part in this interesting event
Alright, basically me and Mike didn't know each other before this event
so we were brought together by the BOB committee in order to give a talk
on earthquakes. So we decided okay, this looks like a debate between two totally
different disciplines, because Michael is coming from the School of Earth Sciences
he's a seismologist, and I'm coming from the Faculty of Engineering,
from Civil Engineering and I am working on Earthquake Geotechnics
But in the very end we found that as we started discussing about these things,
that we're not so different. We have a lot of things in common
We're both working on earthquakes, seismology and earthquakes,
and we're both working on the ground, alright?
I'm probably looking into the shallower
layers of Earth, Michael is looking into the deepest interior of Earth
But yeah, in either case it's the same Earth we are talking about
So I think that I would probably change the title, not to a debate,
but probably to a synergy between the two disciplines
And Mike, would you explain how these two are brought together?
Okay I'll take over. Thanks Dimitris. Yeah this is my simplified cartoon
of the basics of what we're going to try and talk about today
So I'm interested in the actual earthquake, phenomena of what causes earthquakes
and what it means about earth structure, and how seismic waves propagate through
the Earth and how we can image what the Earth looks like
And as Dimitris said, he's a bit shallower. So he's interested in how these
seismic waves destroy the built environment.
And actually, to really understand that,
you need to understand my part and conversely
we struggle with interpreting these
waveforms because we don't really understand the ground response
That's something we've realised by talking together, actually, that there is
really indeed a synergy between the two fields
So I'm going to start with some of the real basics, so apologies -
some of the students here have taken my 4th year seismology,
and this will be old hat to them. But what do we mean by an earthquake?
Well, it's often called an event, or tremor,
and it's a sudden release of energy
and so there's this very classic model now of elastic rebound
Essentially as you apply stresses, which are a force per unit area, to a fault
or a boundary, the stresses build up and then they suddenly release
and we get permanent strain, permanent deformation
This is an example of the surface rupture after the Denali Earthquake in Alaska
in the 1990s, which if I remember rightly was about a magnitude 8.2
So this was a huge earthquake and this rupture went for many many kilometres
And between the two of us, we're going to show you some pretty spectacular examples
of some of the damage and effects that can be caused by earthquakes
I put this slide in because, you know, obviously humans have been dealing
with earthquakes for a long time and there's a lot of folklore concerning
why they occur, what causes them. And so in old Japanese folklore,
the Namazu was a giant catfish that lived under the Japanese islands
And if I remember rightly, there's one catfish per island
And this god Kashima stood over the top of the catfish to make sure he didn't move
and cause earthquakes, but occasionally,
he'd lose concentration and then the catfish
would cause earthquakes. But in a very Japanese way, the earthquake was
the bringer of destruction, but also in this woodcut, you can see
he's got all the tools to help rebuild in Japan
And in fact in this one, you can see him here helping rebuild
So it's a bit of the flavour of our talk in a way, like earthquakes are really good
and really exciting in someways, but for other people, they're obviously
not very exciting and not very good
and it's that dichotomy that makes them really an interesting thing to study
I think that the Japanese, in a very zen-like way, sort of captured this
a long time ago
Historically, earthquakes have always been put down to sort of, you know,
forces greater than us, and there's often theology evoked to explain
and really the first earthquake that triggered a proper scientific debate
was the Lisbon Earthquake in 1755. This was offshore Portugal, near the Azures,
It was probably a magnitude 9, triggered 3 huge tsunamis which,
the ground shaking and the tsunamis
destroyed the city of Lisbon, but it was felt
even in North America
This had really generated a lot of interest in the science community
about why we have earthquakes and it was really one of the first times
that people talked about waves propagating from the earthquake to places
But also it had a really profound impact on philosophy and theology at the time
So Voltaire's Candide was written in response to the Lisbon Earthquake
You know, what he really struggled with is how could
a divine God cause such destruction
So this was in really some ways the beginning of modern seismology
So one of the most obvious things is, well, where did the earthquake occur?
We can feel it, we knew there was ground shaking, but locating where it is,
is actually a challenge, and the fourth year students that took my course
have done a location exercise. And it's a bit of a chicken and egg problem,
It's very simply velocity equals distance over time - you need to know
the velocity structure that the waves propagate through
so the way that we image the velocity structure is by using the earthquakes
but to do that, we need to know where the earthquakes are
So it's a bit of an iterative process. So the
challenges are where is the earthquake
what was its origin time, and what is the velocity structure
that it propagates through? How fast do the seismic waves go through it?
And you can see here, it's quite intuitive, you can triangulate back to the location
- closer stations receive the seismic signals first and the more distant stations
receive the seismic signals later
And what was realised really in the
1960s/1970s was when people started plotting
a map of all the earthquakes in the world - this is every earthquake graded in
magnitude 5 from 1977 to 1994 - and what you can see is there is a very clear
pattern to earthquake locations, and these delineate tectonic plates
Some of them are huge, like the Pacific Plate, some of them are quite small,
like the Cocos Plate or the Caribbean plate
And really it was through these better ideas
of the velocity structure of the Earth
and better estimates of locations of
earthquakes, that these patterns became clear
And this was one of the linchpin arguments to the whole plate tectonic paradigm
As a seismologist, another big question you always get asked is, well, how big was the
earthquake? Or how big can they be?
And really the amount of energy that is stored
in earthquakes is really quite alarming
So the biggest one on record was in Chile, it was magnitude 9.5
and it's hard to say what it is equivalent to, as it is pretty well is the biggest
thing we've had. I mean this is energy released in kilograms of explosives that
doesn't mean anything. But the explosion of Mount St Helen’s, you know the volcano,
Krakatoa, the world's largest nuclear test was up around magnitude 8 earthquake
So for each increase in the magnitude, there is an increase in the magnitude
of energy that is released
The other interesting that is shown on this plot, though, is the number of earthquakes
of these magnitudes. And basically, for each
increase in magnitude of the earthquake
there's an order of magnitude less earthquakes of that magnitude
So for example, magnitude 2 earthquakes which we can't really feel, we can detect
but we can't feel, there's over a million of these a year. In fact there's over
a thousand of these magnitude earthquakes in the UK alone every year
But when you get up to sort of magnitude 9s, you know maybe one every ten years
This last decade has been a bit unusual. We've had 2 magnitude 9s
And so what I've done, or taken - I always like this image - is taken all
of the locations of earthquakes from 1898 and weighting the intensity
of the earthquake by the magnitude. So you can see where they occur
and how big they are
So for example, this is the East Pacific Rise, this is where new tectonic plates
are being formed. Lots of earthquakes, but they're all quite small in magnitude
Same with on the Mid-Atlantic Ridge. The
really big ones are here, like in New Zealand
across Sumatra, Japan, The Kuril Isles, the Aleutians
These are areas where plates collide, and one plate subducts underneath the other
These are where the most damaging earthquakes are. Another notable area is
the Alpine-Himalayan belt. This is more of a diffuse plate boundary
where two bits of continent are colliding into each other
So you can see it's not only where the
earthquakes occur, but also how big they are
that is significant
So I've got a couple of tutorial slides to illustrate breaking a rock, and I was going
to have a demonstration but I forgot to bring my piece of wood
But essentially you take a piece of rock and you have a fault in it or a weakness
a pre-existing weakness - this could be the San Andreas fault - this could be a
tiny crack in an oil reservoir, and you can think of the stresses applied
to that block of material. There's a normal stress which is acting to stop
motion along that fault or that failure plane. And there's a shearing stress that's
trying to instil motion. And it's the balance between that, that decides whether
or not you'll get an elastic release of energy
Other factors - so there's this shearing stress, and this equation describes when
we get failure. When the normal stress and the cohesion for if you think
how sticky the rock is, when this is exceeded, when we get this shearing stress,
we get failure. And this plots as an equation of a straight line
and what you can see is, the higher the
normal stress, the higher the shearing stress
that's required to get the failure
So obviously it's a bit more complicated than this, but these sort of principals
used both in seismology and in earth sciences and in geotechnical engineering
In reality this isn't always a straight line, it curves
and there's a fair bit more complexity
But the interesting effect, and we'll come back to this at least three times
in the lecture actually, is the effect of water. So if you get water into that crack
or that discontinuity, it imparts a pore pressure and that counteracts normal stress
So if you like, it props open the fracture. And you can see this, the demonstration
that I was going to show you, take a block of wood and you try and slide it
across a table - there's a lot of resistance.
You pour some water on the table,
it slides very easily. And that's that pore pressure acting against
the normal stress. So that has the effect of lowering the critical shear stress
that's required to generate an earthquake, and that's exactly what we think happens
in certain situations. So for example in hydraulic fracture stimulation,
where water is being pushed into the rock, that's why we start
to get small earthquakes. Same things happens when ice sheets start to break up
You get warmer water injected into the ice sheet, and that triggers small ice quakes
and eventual failure
And the other thing we can do by looking at the seismic data is infer something
about the actual geometry of the fault. So this is quite interesting as we're trying
to understand things like the San Andreas fault - which actually isn't one fault
it's a whole series of faults - or the large
fault system that the Tohoku earthquake
occurred on in Japan, so by analysing seismic data, we can infer the strike
of that fault, the dip of it and the sense of motion on that fault
That tells us a lot about the tectonic processes that are at play
So this is a map view, looking down on a mid-ocean ridge
So here at the ridge, new tectonic plates are being formed, and we always get
this style of faulting that is known as "normal faulting"
And there's segments of the ridge that we can see, but between that,
we have what are called transform faults. So
it's a plate boundary, but the plates are
sliding past each other and we get a very different style of earthquake
So by analysing this seismic data, we can remotely probe the tectonic behaviour
of very distant parts of the planet, and actually very deep parts of the planet too
So I'm going to show you just three spectacular earthquakes
They're devastating for different reasons, actually, so this is the third
largest earthquake that has been recorded in modern history, as in
the last 150 or so years. It was a magnitude
9.3, it was depth 30km and it occurred
offshore of Banda Aceh in Sumatra, and it's a mega-thruster earthquake
There, one plate is being thrust underneath the other and eventually the stresses
build up to a point where we get this sudden release of energy
And I'm going to show you a description or a cartoon of the behaviour of
this earthquake that was done by using an array of seismometers in Japan
So a long long way away, but a very dense array to look at this
This event triggered a tsunami that was felt in three oceans, which is probably
the first that we've been able to document that
The sense that the length of the fault that ruptured was 1300km long
and the maximum displacement was about 15m, so it's a bit like taking
houses on either side of the street from here to the Orkney Islands
and all of a sudden, one side of the street is 15m higher than the other side
But that didn't happen all in one go
And what's interesting about this earthquake is it started here
and it propagated northwards. So it took about ten minutes for this part of this
plate boundary to literally unzip all the way northwards towards
the Andaman and Nicobar Islands. And it's only through, ironically,
the dense instrumentation of seismometers in Japan that we were able
to actually see this. In fact this was researchers at Harvard that were able
to see this
Another earthquake coincidentally occurred on Boxing Day but the year earlier
and it was only a magnitude 6.6 event, so a lot smaller, like two orders
of magnitude smaller. 15km deep, a different type of earthquake,
a strike slip earthquake, but over 26,000 deaths and nearly
30,000 people injured. This is what the city of Bam in Iran looked like
which was near the epicentre, and this is what it looked like after the earthquake
And the big problem here is not so much the magnitude, you know, we know
there's earthquakes all through this region,
the Arabian Plate and the Eurasian Plate
are colliding in this area here - the earthquake was there - but it's the
building construction, so the mud brick buildings. And this is exactly what
Dimitris researches, touches on and he'll go into the considerations
of these sorts of effects
And then the final one is a sort of intermediate, a magnitude 7.5,
it was generally pretty deep actually, it was in 1970. But it had the effect
of triggering a massive avalanche high up in the mountains
a glacial dam broke that sent down ice, rock, water and mud, and it completely
buried the town of Yungay and killed 20,000 - in fact, only 92 people
survived in the town
And this thing moved, this avalanche, this landslide,
moved at speeds between 450 and 800km an hour
So again, a very different effect, a side effect, associated with earthquakes
so often it's not so much the ground shaking
of the earthquake, it's the knock-on
consequential effects that it has
So that's a bit about the actual earthquake, how we look at them themselves
so now we are going to move into the part where we talk about how we actually
image the what the earth looks like inside
And it's really only our best means of
imaging the Earth's deep interior is through
seismology, it's how we know that the Earth has a solid inner core and it has
a liquid outer core, and we actually know that that core, both inner and outer
are compositionally very distinct from the mantle. The mantle is rocky,
it's more silicate in composition, the core is essentially primarily iron
- it's metallic. And by looking at how seismic
waves refract and reflect and diffract
through the Earth, we can actually image this structure
And what's shown in background here the mantle is this very rich, three dimensional
structure where blue are higher than average seismic velocities
red are slower than average seismic velocities.
And I'll show you some images of that
in a second
The other one, you get a really big earthquake
- the other thing that people don't
often realise - is that the Earth will resonate for many weeks, if not months
depending, so after the Tohoku Earthquake, the Earth resonated for at least 6 months
And there's different modes, they're like overtones in an organ pipe. There's the
so-called breathing mode where the Earth just moves in and out
Obviously this is a bit exaggerated - the displacements are milimetres
There's a so-called football or rugby ball mode, and then you get these
very complicated modes. And each one of these modes is sensitive
to different Earth structure. So these higher
modes, for example, are very sensitive
to the Earth's inner core, these lower modes are very good at getting an estimate
of the density of the mantle, for example
Because they're controlled by gravity, they provide information about Earth's density
So, what you're going to realise is, or hear from Dimitris is that what he'd
sort of needs from me is this -
He wants to see the anatomy of a seismogram because that is going to inform
his geotechnical planning and modeling
And there's many different types of waves that comprise a seismogram
So we have P-Waves - if you think of a Slinky - it's the compressions
and rarefaction if you push a Slinky
S-Waves which are side to side motion, if you oscillate your Slinky
side to side
And then finally, Surface Waves - these are often the very destructive ones
They decay exponentially in strength with depth
But they propagate with large amplitudes on the surface
In fact from bigger earthquakes, these surface waves
can image down to as much as 600-700km into the Earth
So a big failure like something on something like the San Andreas fault
generates this nice rich wave terrain - this is the type of thing I get up early
about, to have a look at
And it's this that's providing our insight into Earth structure
So I thought I'd just show you a little bit about how we actually record
these sort of earthquakes. This is a seismometer deployed in Ethiopia
This is an area we've been working in for over 15 years now
This is a three component broadband seismometer
- we can leave these in the ground
now for many years. They have huge hard disks
What's really incredible is how fast this technology has changed
really in recent years
We get power from solar panels, we get our timing from GPS
If you think back 20 years ago we didn't even have the GPS let alone
a big enough hard disk. And we trickle-charge the battery which powers
the data logger and the seismometer and we usually store that somewhere
separate to this
And this has really changed modern seismology
So here's a snapshot of the type of
things we've been doing in the Horn of Africa - so Ethiopia, Eritrea, Yemen -
Before we started this work there were two seismic stations in the area,
and now there's been through a collaboration
of American, French, UK and Ethiopian
scientists. We've deployed over 200
And to give you an idea of what our picture of the Earth structure
has changed, it's a bit like before we were looking at it without our glasses on
we just knew that the mantle had a depth of 75km, was enormously slow
in velocity. Now we can see all this rich complexity and I often joke that
its the most expensive experiment to tell you where the East Africa Rift is
- this big red blob here
But there's other interesting images, or aspects of imaging
Here's a subducting slab underneath Japan
This is a very famous volcano that borders China and North Korea
It's actually where the legend is that the leaders of North Korea are born
or emerge from this volcano. They call it Paektu, the Chinese call it Chenbai Shen
And tomography, or seismic imaging has imaged the slab that comes down to
about 600-700km. And then it flattens out at this point
This is kind of unusual
But the other interesting thing in this region is that there's lots of earthquakes
that go all the way down to the bottom, like the deepest earthquakes we get
are about 690km deep
And actually given what I told you about the normal, the confining stress and
the shear stress, it's not readily apparent why you'd get an earthquake at depths
of greater than 400-500km
And some of these are big - there was one in 1994 in Bolivia
It was a magnitude 8.2, and I was living in Toronto at the time and it was
actually felt in Toronto. I mean that's many thousands of kilometres away
And so the big question is why do they occur? One idea is it's dehydration, so
the minerals that are being carried down
from the surface have a lot of water in them
and then they suddenly dehydrate and that gives you that water that therefore
lubricates the fault and you can get failure
It's a contentious idea but it's the same mechanism
And there's some recent evidence of minerals of this type extending very deeply
into the bottom of the slabs
So I'm just to finish going to talk a little bit about more recent research
that's going on in Bristol and looking at induced seismicity
So this is not earthquakes that are occurring naturally, these are earthquakes
that we cause. And there's a number of different ways that I don't think
people realise that, actually, that we've been causing earthquakes to some extent
for as long as we've been manipulating and doing things to the ground
So one is water impoundment. If you put in a bit hydroelectric dam, you change
the hydrostatic pressure and that can trigger an earthquake
And one of the worst examples was in India in Koyna
A magnitude 6.5 was triggered in 1967 and that killed 200 people
And there's many other examples - the Aswan Dam for example,
in Egypt also triggered seismicity
Coal mining is something that we're very used to - Britain gets hundreds
of earthquakes every year up to about magnitude 3 associated with coal mining
There's an area in Northamptonshire which is
described as Britain's earthquake capital
gets magnitudes up to 1.7 which you wouldn't feel, so I wouldn't really
call it an earthquake capital, but there has been a lot of them
In fact one of our final year students is looking at these earthquakes
Interestingly if you look at social media, one person was convinced that
it was due to fracking, even though there's no shale gas in the area at all
It was coal mining related
And then one of the more controversial in induced seismicity is associated
with oil and gas, and so in one part of northeastern British Colombia
hydraulic fracture stimulation actually stimulated magnitude 4 earthquakes
But these are very very unusual. And the small town that's near this area
called Fox Creek, it was interesting to see how social media replied
You know, there was no way that could be caused by fracking as far as the expert
that runs the local pizza parlour was concerned
So everybody has got an opinion when it gets to these sort of things
This is an animation and it shows earthquakes in Oklahoma, and this is a
really interesting and complicated area right now
You'll notice some dots going off here from about 2006 up to about 2011
And that they're fairly sparse. And then you'll notice a sharp increase
This is the accumulative number of earthquakes in all of Central US
Up until about 2009, there was a fairly steady rise, and then all of a sudden
there's this sharp huge increase primarily in Oklahoma and to some extent in Kansas
And we now know that this is occurring from conventional oil production
that produces waste water, and that waste water is being injected into
deep saline aquifers
What's interesting about it, you can see all the seismicity up here, that all
of a sudden just shot off. Again it's the same effect - the pore pressure
is reducing that effect of normal stress but the interesting question is
why is it occurring here? This sort of operation occurs all over the place,
so what is unique about Oklahoma? Why is it occurring there?
That's an interesting question.
And so going back to our Namazu, the catfish, nowadays we have
a slightly different Namazu, it's ourselves that are actually causing
some of these earthquakes
So just to finish my bit, so I've tried to describe why and how we get earthquakes
but there's still some issues, you know we don't understand all the details of
induced seismicity, these very ultra deep events. And I haven't mentioned this
but we also get very slow earthquakes
Where? Well for the most part, we know they can occur on plate boundaries
but we also get them in inter-cratonic areas
So Britain has had a magnitude 6.2 earthquake in an area called Dogger Bank
which you might know from the shipping news
When? Well that's the holy grail and that's something that we never really
be able to understand
What we'll be able to say is what the risk and hazards
sort of where, and how big it might be, and what Dimitris will now talk about
are the side effects and the engineering. And in some ways this is a much more
rewarding area of research than trying to decide when
What I'm interested in learning from him, is to understand induced seismicity I
need to go into much finer detail, to understand the processes I'm looking at here
So I will hand over to Dimitris now who will talk you through the next phase
Thank you
Alright, so, thank you Michael for not only the talk but also for this
because basically that's what I was needing in order to design our structure
on top of that
So what we get from as engineers from seismologists is that input motion
which we apply to the base of our buildings
We can then estimate all the internal
forces developing in all these members
And basically design these ones
There's a list of different technologies so you can have shear walls, thicker shear
walls for reinforced concrete buildings, diagonal members for steel structures
and for buildings of high importance, such as hospitals and museums etc.
you can even get different technologies of base isolation at the foundation
so that the building doesn't really feel the earthquake
Or newer technologies now with dampers and all that kind of
nice interesting fancy stuff
Basically, myself, I am a geotechnical engineer so I'm going to focus my talk
on something else
In fact what Mike is giving us, is the input motion at the bedrock
But in most cases it's not the bedrock on which we build the building
Generally between the bedrock and our building there is a layer
which can extend to several tens of metres of soft soils
and the question that I'm going to try to answer today is
"Do these soils actually affect the building's performance?"
And the most characteristic example to go through that is an earthquake in 1985
in Mexico. This wouldn't have been really really interesting actually -
if you look to the damage that happened in the neighbouring areas, it was only
mild to moderate, but more than 350km away, there was more than 400 buildings
that collapsed, more than 3000 ones that were seriously damaged
In this particular case here, in this complex, one building
actually fell on top of the other
So what caused such devastating effect in such a long distance?
Basically it was the soil conditions, so if you go into this particular area
Mexico essentially is located on a lake and the lake is made of some very
stiff hard basaltic rock, but inside the lake, there is these very soft
and highly plastic silt and volcanic clay sediments
which extend to depths of 20, 30, 40 and even more metres in depth
And if you go to the edge, to the outer parts of the city, and you look at
outer recordings over there, you see a very very weak motion
however in the city centre, this was much much more intense
So what caused that?
So if we looked schematically, there was the waves coming from the epicentre
they went to the seismic bedrock, and then we got waves traveling up and down
into this soft soil deposit
And if we want to go into more detail, we will need to understand
a little bit of theory about waves
So waves have a direction of propagation, they do have an amplitude
- self explanatory - and they have a wavelength
So this is measured from maximum to maximum
And this wavelength now depends on two things
Firstly on the soil medium itself. So you know that sound travels faster
in water than in air, right? So when we have stiff soils, the wave velocity
is faster, so the wavelength is longer
And the second thing - I don't know how many of you are familiar
with gymnastics - but you all probably can understand when you have the ribbon
when you move it fast, at the high frequency you get small wavelengths
and at the lower frequency you get larger wavelengths
So all these waves travel up and down in our soil deposit
and because they reflect at the surface, they become stationary
Stationary meaning that they have the maximum amplitude here
no movement here, no movement here
What's happening here?
Well it depends on how many wavelengths you can fit in there
So the thing is that because here you get the maximum
and this thing depends on how many waves you can fit in there
you actually get an amplification of this motion from the bedrock to the top
And what we tend to use quite often is this amplification ratio
which is the ratio of the motion at the top to the motion at the bottom
and obviously this depends on the wavelength
and on the thickness of our soil layer
So if we do a little bit of math - I'm not going to go through the equations today
you are lucky - if we do a little bit of math, we're going to see that
this amplification ratio is larger than 1, meaning that you get a more intense motion
at the top, and actually in some cases, you get infinite amplification ratios
This looks like this is quite similar to resonance - it's not exactly resonance
but it is a quite similar thing
It is when you can fit exactly one quarter, three quarters, five quarters etc.
of a wavelength into our soil layer
In these cases, you would get zero movement here, and maximum movement here
Well, practically this never happens, okay? Why?
Because the soil is not a perfect elastic material. The soil has damping
we lose energy through soil, so actually you never reach infinity, you do get
significant amplifications, but definitely not infinite ones
And for different damping, you would get different kinds of amplifications
and by the way, damping depends on the motion itself
What is damping in the first place?
It's the energy being lost as you get dislocation of soil grains
So the more intense the motion and the formation in the soil
the more intense the energy lost, the more intense the damping
And by the way, for large values and damping you can also get deamplification
values and not only amplification because of the soil
So essentially, the soil works like a nice and beautiful equaliser
You've got the seismic motion at the bedrock coming from the epicentre
It consists of several different motions with different frequencies
and as an equaliser, the soil decides which one to amplify
and which ones to deamplify so that what you get on the surface
is totally, totally different
Good news about this is that we can now understand the behaviour
of soil in there. We have tests in the laboratory where we can simulate
what happens in them, then we can develop numerical models
And using these models, we can now get these maps which use the geotechnical
information for each particular location and they can give us a detailed result
about how intense the earthquake motion is going to be in each particular location
This procedure is called Seismic Microzonation, something quite useful
for engineering today
The question is is that enough? Is that all that soil is doing because of earthquakes?
Quite a long time ago in 1964, in Niigata in Japan there was an earthquake
which was a landmark because from this earthquake we took this very famous picture
I don't know how many of you have seen this picture before
Well, it was not only these buildings, it was tens of buildings that settled
and tilted with maximum settlements reaching 3.8m into the ground
Okay, spectacular settlement
Why? Again it's the soil that did it
Niigata is a city built close to the river, and the river used to bring
all these soft sediments. These are moving deposits all over this region here
which had depths of up to 15m, and basically after the end of the earthquake
you could see them right on the surface
Well, what happened?
What happens in granular soil when you shake it?
You will probably not know but if you're not on a diet
you definitely know what happens to sugar when you shake it
So when you buy sugar and you open the tin and you try to fit it in there
you pour it in, and then
oh my God, you've got a little bit more left outside - what do I do?
Tap the tin, it settles, alright, now I can pour the rest in!
So yes the same thing happens to sand. When you shake them, they compact
This though happens to dry sand only. What happens to saturated sands?
Well what happens when water is there?
Now in this case, these ones want to settle, but they can't actually
because if they want to settle, water needs to go away
And if we do this very very fast, as would happen for example in an earthquake,
then there's not enough time for the water to go away
So essentially the different grains start floating inside the water
and now it's the water which is now carrying the weight of the soil
And you can see that because you get an increase of pore pressure
of pore water pressure
And in this situation, in this case right here, you've got grains that are
not touching each other. This means that the soil essentially behaves like
a very viscous fluid, like a soup
This phenomenon is called Liquefaction
Of course in the end you've got the soil settling and the ground settlement
at the end of the earthquake, and the water gradually moves towards the top
and as it moves towards the top it gets all these things with it
This photo was from Christchurch in New Zealand
You can see these sand boils that started forming after the end of the earthquake
Well, this is quite zoomed in - if you zoomed out, you'd see quite
a devastating situation. It was sand boils all over the city
And is that what liquefaction is about?
Sand boils - actually just doing a cleaning up at the end would be enough
No, that's not it and in order to demonstrate
what the effects of liquefaction are,
I could start an experiment but most of you know how clumsy I am
And I wouldn't want you to get told off by the people from the Clifton Hall
by spilling sand and water all over the place
So I'm going to give you some instructions on how to do this in the summer
in a beach - I would like to say in Greece but alright, any beach would do
So what you need is a container, sand, some water, one rock and one table tennis ball
which is optional
And what you have to do is place the container on a table, fill it up with water
- not to the top - pour the sand inside until you reach the "water table"
or the water the surface, and while pouring this, just remember to place
the table tennis ball in there and in the end place the rock at the top
And all you have to do then is shake it
Any ideas what is going to happen to the rock?
Yep, it's going to sink just like these buildings in recent earthquakes on top
of liquefiable soil. You can see significant settlements over there
but very importantly, you don't see any
structure damage on the buildings themselves
Why? Because actually the liquefied soil became something like a fluid
so shear waves could not propagate through the liquefied soil
and they did not really cause any structural damage to the superstructures
And of course what is going to happen to the tennis ball?
Buoyancy - it's going to come up, okay, just as any light underground structure
Well you might say that this is not very important as compared to the buildings
but imagine we are in an emergency situation and we really need these roads
not blocked immediately after the earthquake
so this really is an important problem
Many other effects from liquefaction, one of them being lateral spreading
And in this case, you've got mildly sloping ground. And because the soil liquefies
and loses its strength completely, even at very mild slopes you get movements
downstream that can reach several metres actually
And in this case it would be nice, I guess, because you could go to the swimming pool
and swim directly into the sea, in this hotel in Turkey
which moved when we had lateral spreading by about 3m
I'm not really sure it would be quite that nice in that particular case in Kobe
in Japan where this was a small lateral spreading but right behind that
this pier moved and the whole deck of a bridge fell over
Yeah, liquefaction - quite detrimental, quite devastating. Now what?
In fact, to be honest, we do not really understand liquefaction and all these
phenomena so well. So basically we are starting to understand them, but
we're still quite scared of them
So what we are trying to do in current practices, is just avoid it
Just find the remedial measures for example, compact the soil
so that it does not liquefy. How can we do that?
Well, throwing a mass on top. A procedure which is called "Dynamic Compaction"
Or if we want to go deeper, we've got these huge massive vibrators that go
at large steps into the soil and compact it
Or, you remember that this has to do with water, and excess pore pressures
we can use drains in the soil so that when these excess pore pressures develop
we can dissipate through the drains that come up to the surface
and we can avoid the whole problem
Or a combination of all these, which is "Vibro-replacement" where we just construct
stone columns out of gravel in there, and there's a double effect in there
because A - it reinforces the soil and B - because gravel has a very large
permeability, they also work like drains and they let all the excess
pore pressures dissipate
However recently, in the last couple of decades essentially, we've started modeling
these soils and trying to get bigger insight into the behaviour
using different devices in the labs
And using these we can now develop different types of models
mathematical models that simulate the behaviour of soil element
under all these conditions until liquefaction
And we can use now these models, taking advantage of the computational facilities
that we have available today, and we can run simulations for all different
sorts of problems. For example, shallow foundations and how they behave
on liquefiable soil - that's what I did in my PhD -
lateral spreading and how it affects pile foundations, or get some better
understanding of how gravel drains actually work. And this
basically is a project that is currently
in progress. Two students from the third year of civil engineering
are actually working on this, about pipe lines and how much
they uplift you to liquefaction
Of course though, we are talking about very complicated phenomena here
We should not trust our numerical models so easily
We need to verify our results and we need to make experiments
in order to make sure that we get very good simulation of experimental data first
And we need to be also very careful with these experiments, because there's
quite a lot of things which need to be considered - how we apply the shaking,
what are the boundary conditions, where we place the instrumentation, what kind
of instrumentation we use, and all that kind of stuff
But in the end, we can get innovative ideas and new design methodologies
so in a feasibility study for a typical bridge deck,
like for example, which was founded on
liquefiable sand, under conventional methods this could involve
ground improvement of this liquefiable layer and then using piles to transfer
the loads of the loads from the bridge deck to the lower stiffer soils
But now by understanding and being able to simulate the behaviour of liquefied soil
I think that we can also get away with only a partial improvement
at the top of the liquefied layer, leaving all this to liquefy
And A - this is going to save us money and B
- this might also save the superstructure
Because remember what I said about soil amplification and deamplification?
In this case you get significant deamplification because of liquefaction
and the seismic wave at the top is much much weaker than the one at the base
something like a natural mechanism of seismic isolation
Well, this however - one final point and with
this I would like to close my talk today
in order to be able to do all that stuff, in order to be able to understand
how buildings are behaving on liquefiable sand, in order to do the experiments -
by the way this is a photo from the biggest shaking table in the UK, which I suppose
you know is here in Bristol - and also to make numerical simulations
we need to be very careful going into detail and understand how soil behaves
what's the interaction between grains, and all that kind of stuff
and make sure that we use the appropriate models for that
And of course this model will definitely be useless for Mike
because in order to calibrate this, you would need to take magma out of the centre
of the Earth and put it in the triaxial apparatus, which is impossible obviously
but on other hand, the more long model that Mike showed would not be enough
for these cases because obviously it does not predict all these phenomena
with excess pore pressure build up that was liquefaction
And trying to generalise this in conclusion,
when we look into practice, we need to be careful
and go deep inside and understand the theoretical background behind it
And as we go deeper and deeper and deeper and we specialise, we should never forget
the broader picture, the context of what we're working on
And this does not only extend into this field, it goes further beyond
So in some cases, we should go even deeper and understand the micro mechanics
of the grains in order to estimate soil behaviour, but we should never ever
forget the broader picture of all our structures and where we're building them
And understand that all these different components form part of
the same urban environment. Mike?
Yeah, so thank you Dimitris
So, I guess in the interest of symmetry, I just have one last slide
Essentially what we've talked about is the interface between geology, Earth sciences
and engineering. And actually in discussions with Dimitris and others in
civil engineering, I've realised that Bristol has probably one of the best
concentrations of people working in
seismology and earthquake studies in general
in Europe. Right from risk and hazard assessment through to understanding the
earthquake itself to imaging, to geotechnical engineering
And it spans many departments and that's a two way street between the two disciplines
which is what we've tried to say today, but there's all these other disciplines
- oh actually my updated slide didn't come through, we copied the wrong one over -
what I wanted to say here was that essentially,
there's input from applied mathematics
in understanding wave propagation, in physics in understanding physically
how waves propagate, in instrumentation in how we build the next
generation of seismometers or accelerometers,
and then there's implications
in risk in hazards, which have implications for legal frameworks,
environmental regulations, economics, social sciences - so there's actually
quite a broad range of topics that you can tap into by considering, you know,
the consequences of an earthquake, so we'll finish there. Thank you