Japan Earthquake Science

[CraigS]

So the debate simmers to the boil again … how big is it .. how long does it last … when will the next one happen.. how big will it be ?':

Quake could alter Tokyo risk: experts (http://www.physorg.com/news/2011-03-quake-tokyo-experts.html)



So out come all of the prediction models …

The last time I looked into this, the Chaos Modelling conclusion was that:



The definition of an 8.9 magnitude earthquake, is imposed by people on nature .. not the other way around.

Comments welcome.

Cheers

[renormalised]

Thew whole idea of how big they are and how long they last is not unimportant...it has to do with energy released, which is what the Richter Scale is. Each major division of the scale is 32 times more intense than the previous, so a mag 8 quake releases 32 times the energy of a mag 7 quake. How long a quake lasts is independent of how much energy they release, but as a whole, the bigger the quake the longer it lasts.

All Chaos Modeling is doing, is confusing the issue....it's largely irrelevant. All it's doing is just taking a different perspective to how you view quake physics. It makes no difference to the energy released and the length of time they last.

[CraigS]

I'm not sure I could ever see how a predictive model could ever make a difference to "the energy released and the length of time they last", (as per my last statement, in my previous post).

However, clearly the emphasis is still on collecting data, and crunching it in forecast models .. theoretical processing techniques will always be demanded from scientists, in order to make sense of the data ...



Cheers

[renormalised]

Trying to predict earthquakes is like taking a craps shoot at a carnival...in reality I don't think they'll ever be able to predict to within reasonable certainty when a quake will happen. You would have to know every single variable for every single possible occurrence/outcome to be able to predict a quake with 90-100% certainty. In any case, the system would ultimately rest on quantum physics and that immediately throws your predictive abilities right out the window.

In any case, as I mentioned, the Richter Scale is not about prediction, it's about quantifying the amount of energy released in an earthquake. The same with the Mercali Scale, except in that case it's the amount of damage done by an earthquake.

The term aftershock can be very problematical. Unless the series of quakes is related, i.e. belonging to the same system of stresses and geological provenance, then defining what an aftershock is can be difficult. What I mean by this, is you could call every earthquake that ever happened an aftershock of the previous one, no matter where/when it is. Why??....because it's conceivable that an earthquake in California could setup up a stress all the way across the Pacific Plate that might trigger an earthquake in Japan 10 years later.

[CraigS]

… hmmm … a bit like predicting life on exo-planets, eh ?
:P :)
(.. don't take me too seriously … 'twas only in jest .. I'm in another 'mood' today … :) )



You mean like Chaos Theory's 'Butterfly Effect' ?

:)

Cheers

[renormalised]

Like the Butterfly effect....yes and no. Yes, in that one event in a particular spot sets up an event in another, remote spot. No, in that there is an observable connection between the two remote spots and that it is possible to define a cause and effect in that system which can be measured...i.e. earthquake happening in CA, sets up a stress in the same plate on the opposite side of the planet, in Japan, that creates a stress which may take 10 years to release.

In Chaos Theory, there is no observable connection between the two events, even if they are part of the one system (i.e. the atmosphere). The Butterfly Effect is just a random chain of events which leads from the flapping wings to the thunderstorm.

[CraigS]

Hmm … interesting .. I think the knowledge which establishes a demonstrated cause and effect within the closed system, more or less, eliminates the need for predictions of the outcomes within that same system ? :question:

Chaos theory excludes randomness. Its about non-linear systems, which may have the appearance of randomness … but non-linear relationships can be defined from macro-level observation. This is then used to establish the form of the pattern. The non-linearity implies that the causes are not random .. but are defined by physical laws obscured (to the observer), by complexity.

Cheers

[renormalised]

No, despite there being a connection and a cause-effect mechanism present, the fact is that the effect and its timing is not predictable. I said that you can define a cause and effect mechanism, not predict what will happen and when. Any cause will setup a series of events, but whether in the case it's a reduction or increase of the stress in the opposite sides of the plate is a matter of an indefinite set of variables. The CA quake could have easily released the stress in Japan and mitigated a future quake just as easily as it made one inevitable. Prediction, therefore, is still useful in measuring outcomes in such as system.

If something is non linear, but has a definable pattern it cannot be chaotic, by definition. Non linear systems, despite many appearing random, cannot by chaotic because they follow an underlying order which maybe obscured by complexity, but is still there. Chaos, in the strictest sense of the word, means total randomness with no underlying pattern or order driving the mechanism or system. Therefore Chaos Theory, by definition of what you have written, is not chaotic. Which is a contradiction in terms.

It would be better called Non Linear Systems Theory.

[CraigS]

A few definitions may be in order here, (from various sources)…

A non-linear system is one whose output is not directly proportional to its input.

A deterministic system is a system in which no randomness is involved in the development of future states of the system.

Small differences in initial conditions of a system, (such as those due to rounding errors in numerical computation), yield widely diverging outcomes for chaotic systems, rendering long-term prediction impossible in general.

This happens even though these systems are deterministic, meaning that their future behaviour is fully determined by their initial conditions, with no random elements involved. In other words, the deterministic nature of these systems does not make them predictable. This behaviour is known as deterministic chaos, or simply chaos.

More later .. gotta go.

Cheers

[renormalised]

I know the definitions, Craig, but that still doesn't change what I said. In their strictest sense, they're not truly chaotic. It's a case of bad use of words to describe what's happening. They may not be predictable systems, in so far as their outcomes are concerned, but there is an underlying order driving these systems to whatever outcomes they are heading towards. Once the initial conditions are set, the outcomes are what they'll be, but you have no way of predicting what they'll be. Basically, to be able to predict outcomes in systems such as these, you would have to know of every possible existing variable of the system along with every possible combination/cause for those variables. An indefinite set of variable and cause/effect outcomes. Impossible to model, unless it obeyed something similar to the quantum principle of superposition of state and Heisenberg's Principle.

[Robh]

Hi Craig and Carl,

This is an interesting read.
From an article by Balakrishnan Ramasamy and T S K V Iyer
Chaos Modelling with Computers-
Unpredictable Behaviour of Deterministic Systems ...

Read from ...
Chaos and Dynamical Systems page 29
to ...
This property of sensitive dependence on initial conditions is one of the characteristics of chaos. Page 30.

Full text here ...
http://www.ias.ac.in/resonance/May1996/pdf/May1996p29-39.pdf

Regards, Rob

[CraigS]

I'm back .. Ok, Carl .. yep .. the definitions I posted were for the record and weren't intended to counter-argue anything you've said. I was going to say this, but I ran out of time. (Apologies for any misunderstandings).

As far as long term earthquake prediction is concerned, I agree that at the moment, it still seems we still don't have the ability to make long enough predictions, and it may always be that way.

My purpose for introducing the concepts which Chaos Theory introduces, is simply a way to help us step outside of the usual things we go looking for, and ponder them for a while, given that traditional ways of looking to create predictive models of earthquakes, may never get us closer to the 'traditional' end goals, which target providing answers for questions like:
- how big will it be .. how long do clusterings last … when will the next one happen.. etc, etc.

Rob: Thanks for the paper … its interesting .. I don't know about others, but it seems that no matter how many times I read about Chaos related areas, I find it difficult to retain the concepts in memory (over the medium to longer term) …. another reason I posted the basic system concepts/definitions.

Perhaps Chaos is another one for the "Counterintuitive" thread, eh ?
(Or is it just me ?). :)

I'm always interested in reading about the practical outcomes of Chaos Theory and Fractal geometry. Fractal geometry applications are easier to come by, but I'm sure that the concepts of Chaos Theory are well and truly embedded in all sorts of mathematical models thesedays.

Cheers

[CraigS]

Here's another interesting series of Questions and Answers (http://www.physorg.com/news/2011-03-scientists-discoveries-japan-temblor.html) … A quick summary of the questions (of relevance to this thread .. ie: forecasting/modelling) follow:

Q: What caused it?
Q: Was it a surprise?
Q: You mean there were no hints at all?
Q: There was a lot of seismic activity off Japan's coast last week, including a magnitude 7.2 quake on Wednesday. Should that have been a warning sign?
Q: Should we expect more earthquakes?
Q: Will this change the way scientists look at earthquakes around the world?
Q: Does this change our understanding of earthquakes in Southern California and elsewhere on the West Coast?
Q: This sounds a lot like the 9.3 earthquake that struck Sumatra in 2004, generating a tsunami that killed more than 230,000 people in 14 countries.
Q: The 1995 Kobe quake in Japan killed more than 6,000 people. Was it almost as big as this one?

This demonstrates the 'traditional' type of questions which put pressure on forecasting models ..

The answers are interesting (I have provided a couple below) but don't really lead to any conclusions (as expected given the nature of the phenomenon, I guess).



.. so the field of possible scenarios broadens, rather than converges …

By the way, I'd just like to express my deep concerns for the human tragedy situation over there as well … (its just awful, and I wish I could help).

I think at the end of the day, people will eventually turn to Science for better predictive tools and maybe Engineering knowledge, for protection.

Rgds.

[snas]

Craig
I hope that your hope that people will turn more to science proves correct. However, my experience in medicine is that many people have become unable to grasp modern science based medicine's concepts, have been scared off from use of modern drugs (normally because of the rantings of people who seem to have some strong bent against modern science based medicine). As a result, I am sadly not making more room to house the sudden influx of "modern science converts".
Already I have seen a site saying that the quake was god's retribution against the Japanese for killing whales and for killing Americans in WW2. What the!? You can't deal with crazies!

Stuart

[CraigS]

Hi Stuart;
My motivation is not so much in the hope that people will turn to Science (I don't really care what people turn to .. they have a free choice).
But when something like this happens, interrogative style questions invariably get directed at the scientific agencies who receive funding.

What can the agencies do in a situation which has no definitive predictable outcomes (over the long term)?

Its really a conundrum, I think.

But clearly, the focus at the moment is developing ways to warn of impending quakes, which give adequate preparation time for those at risk.
Which all makes good sense. It'll probably never result in sufficient warning to save as many lives as the general public might expect, however.

Defensive building codes, tsunami warning technologies and tsunami protection barriers are really the best we've come up with so far.

I think it would be great if people understood the true nature of the phenomenon, from a more theoretical forecasting perspective, however. This resets expectations, which on the whole, I think, aren't generally particularly realistic.

Cheers

[renormalised]

I sometimes wonder how the human species even got as far as it did, considering the general level of intelligence that seems to be exhibited by 90% of them. Or, more to the point, common sense. If it wasn't for a few clever people and a lot of graft by people nearly as smart, most of them would still be running around scratching their bare backsides out in the middle of the bush. What the real conundrum is that the more technologically advanced we become, the more imbecilic the general populace becomes. The more fundamentalist many turn out to be and the less advanced their outlook and apparent education becomes.

That's why you get ridiculous sites like the one you mentioned.

Trying to explain the science behind earthquakes and their prediction would be a waste of time with many...it would go straight over their heads. High school was hard enough for them, let alone having to try and understand uni level geology/geophysics.

Even many, so called, "well educated" people aren't as well educated as they might like to think they are. Science for them was something only done by nerds in lab coats. And that's the general attitude of society towards science.

[avandonk]

A good model for plate movement is a brick on a rubber band pulled up an inclined plane of carborundum paper. No matter how many times you run the experiment every outcome is different. It is called a complex system.

Bert

[Barrykgerdes]

The theories will soon come fast and furious and probably none of them will get near the truth. We have had a relatively calm period in the "ring of fire" for many years but now it appears we may be moving into a more active period. The "big one" in California is yet to come but I don't think they are prepared for it. Yesterday I watched a program on TV that stated that "California has more nuts than any other state".

I loved science at school. We were taught many things in theory and then had the practical work to prove them. I think today's "science" is based on statistics and "political correctness" without any real testing of its theories.

It's surprising, and then maybe not so surprising the number of people who base their life around the stars and the predictions of the astrologers that don't understand anything about the planet movements.

One thing you can be sure of. Things will not change!

Barry

PS Here is the cause of the latest earthquake. China is putting so much carbon into the air on the west side of the Pacific and Julia so much hot air on the Eastern side that the Pacific plate had to tilt to compensate!

BG

[ballaratdragons]

That is exactly why I am excited that my 15 y.o. son asked the Principal at his school if he can do Advanced Physics next year instead of the normal Physics class. :thumbsup:
Physics excites him!

The Principal said yes :)
There will be two in his class :lol:
Sums up your post perfectly.

[CraigS]

So now, GPS measurements have confirmed a 2.4 metre (8 feet) shift of one of the GPS stations in Japan (not sure which one at this stage) …



They've also calculated that the Earth's rotation speed has sped up by 1.6 microseconds .. slightly more than that caused by last year's Chile earthquake (8.8 Richter). Sumatra's 2004 quake (9.1 Richter), caused a 6.8 microsecond shortening of the day.

It'll be interesting to see whether measurements confirm the 1.6 microsecond increase in speed. The relationship between large quakes and the accelerated rotation speed is interesting, given that they also just recently discovered, (from seismic measurements), that the core is actually moving much slower than previously believed – approximately 1 degree every million years.

As a comparison, over the course of a year, the length of a day varies by about 1 millisecond, getting longer in the North American winter and shorter in the summer. This is thought to be mainly due to the average exchanges of energy between the solid Earth and fluid motions of Earth's atmosphere and its oceans.

Longer fluctuations of about 4 millisecs occur also, (over a 65 to 80 year cycle), and are believed to be due to flow of liquid iron in the outer core, and are measured in the magnetic field variation patterns. These fluctuations also correlate strongly with the average global surface temperature patterns.

Seems like it would take a lot of very large earthquakes, to make any appreciable impacts, when compared with the fluid movement variations.

Cheers

[CraigS]

Here's some more perplexing 'chatter' (from this morning's Herald):



Hmm Mr McCue must have superior modelling skills .. then we get this …



What is it that they are actually trying to say ?
.. absolutely nothing !??! … (and very well said too, I might add).
:)
Cheers

[Outbackmanyep]

Hi guys,

FYI

I subscribe to USGS alerts, the first alert i have is from last Wednesday

Region: NEAR EAST COAST OF HONSHU, JAPAN
Geographic coordinates: 38.510N, 142.792E
Magnitude: 7.2 M
Depth: 14 km
Universal Time (UTC): 9 Mar 2011 02:45:18
Time near the Epicenter: 9 Mar 2011 11:45:18
Local standard time in your area: 9 Mar 2011 13:15:18
Location with respect to nearby cities:
169 km (105 miles) E (80 degrees) of Sendai, Honshu, Japan
196 km (122 miles) SE (133 degrees) of Morioka, Honshu, Japan
221 km (137 miles) ENE (66 degrees) of Fukushima, Honshu, Japan
416 km (258 miles) NE (40 degrees) of TOKYO, Japan

Since then, including the big earthquake, i have had 97 more alerts.

There were 3 alerts on Wednesday, 4 of Thursday, in the range of 5.7 to 7.2 (being the first alert), so if anything these initial earthquakes should have sounded warning signs although minor.

Looking at the rest of the earthquake alerts, the average depths of the occurences were around 30kms deep.

Amazing!

[renormalised]

They are saying something, Craig. The first statement is saying that because of the size of the aftershocks, there's a good possibility of there being another large quake, in his estimation.

The second is saying that the 80's and 90's were quite and that the last two years or so has been active, more so than usual. He thinks it may not be random and they don't know why.

Probably not all that important in the overall scheme of things, but they're not just throw away lines:)

[CraigS]

I see what they're saying .. and it sounds logical, but I still question the value of what they're saying. Because of the known nature of the beast, even logical statements add little-to-no value.

If I was living over there, I doubt that I'd be needing someone to tell me that there's a good possibility of there being another large quake, in someone else's estimation. I think I could estimate that as well as anybody (probably to the same degree of accuracy, too) .. but that still doesn't give me anything I can rely on.

In my opinion, the second statement is almost self-redundant.

Cheers

[renormalised]

Well, they have to sound like they know what they're talking about...they are scientist, you know:):P

[CraigS]

This gets interesting. Found some interesting words:



So, it appears most serious attempts at earthquake prediction, take more recognition of earthquakes as complex systems, rather than chaotic.

If this is the case, then even predictions citing ‘higher probabilities’ of impending large ‘aftershocks’, (ie: capable of generating tsunamis), would be a subset of the overall complexity and would thus, still be complete conjecture. There is just as much chance of there not being large aftershocks.

Interesting .. this all depends on the modelling approach taken from the outset which again, is dependent on basic observational data taken over a given timeframe.

Cheers

[CraigS]

Here's another geophysicist dude speaking authoritatively (http://media.smh.com.au/more-megaquakes-likely-soon-2233569.html?from=newsbox) … (he's from the "Richard H Hagemeyer Pacific Tsunami Warning Center") … take a look at the 1:00 mark where he ends up saying that he thinks we have 'a few more (big ones) to worry about' (referring to the perceived clustering).

So, here's a recent paper (http://arxiv.org/pdf/1010.5589v2) (dated 28 Oct 2010), which uses the latest (sophisticated) earthquake modelling techniques, which concludes:



I'm overwhelmed by the lack of commonality on the issue of unpredictability (as is evidenced time and time again from the theory), and the message these guys deliver. The two are completely at odds with eachother.

Am I misunderstanding some thing here (I'm prepared to admit that this is more than possible :question: ).

Cheers

[CraigS]

Ok .. the scientific plot thickens …

He's a video of 'John Rundle' (http://www.physorg.com/news/2011-03-tokyo-massive-aftershock-expert-video.html), UC Davis seismologist predicting the numbers and magnitudes of the aftershocks. A gloomy outlook for Tokyo, it seems.

He mentions these predictions are in accordance with the 'well known' Laws of Seismology … he mentions 'Bass' Law of Seismology (at around the 25 sec mark).

Anyone know anything about Bass' Law ?

(I'm very willing to learn more about this … )
:)
Cheers

[renormalised]

It's to do with the statistical analysis of earthquake prediction, namely the scaling of earthquake seismicity and modeling between simulated and actual earthquakes. They try to predict aftershocks using the models generated. Rundle helped develop the BASS model for earthquake prediction:)

Here's the article:)

[CraigS]

Thanks Carl;

Got a reference so I can read up on it ?

Cheers

[renormalised]

Just added it to my previous post:)

[CraigS]

Cool .. thanks.

[Brian W]

Not exactly pleasant science but does anyone here have any idea of how quickly the 'faceless fifty' receive their lethal dose of radiation. As I understand it the Japanese government has raised the legal radiation daily dosage to 5x the U.S. standard?

Brian

[CraigS]

Here ya go … the aftermath and a reminder that earthquakes cannot be predicted (because their patterns are part of a "Complex System"):

Japan's seismologists blinded to March 11 quake: journal (http://www.physorg.com/news/2011-04-japan-seismologists-quake-journal.html)



More evidence that we should all be more focused on Chaos/Complexity Theory, as our survival implicitly depends on it !

The arrogance here, is to ignore what we know, which is a mathematical certainty.

Nature at the macro scales, does not allow predictability !
… This is in spite of the predictability of the laws of classical physics, which drive it all !

Cheers, Captain Chaos.

[yusufcam]

my feeling is that these things are unpredictable only because of the paradigms being used (or more precicisely the assumptions upon which those paradigms are based).

Not that they are unfathomable, beyond comprehension, or in the final analysis, inherently (in a sense) unpredictable.

but this is belief

[CraigS]

Hi Colin;
Nature does its own thing. The models we create, are our feeble way of trying to convince ourselves that we can master (via predictions), phenomena which may ultimately, be not predictable, by their very nature.

The unpredictability of earthquakes is demonstrable. Our predictions fail all tests, so far applied.

This gives absolute weight to the assertion that they are unpredictable, and thus also goes a long way towards helping us to subdue our intuitive reliance on feelings and opinions, when it comes to physical phenomena.

Sounds like these Japanese scientists mentioned in this article, may not have yet tweaked to the unpredictability of it all, (as is supported by empirical, observational evidence).

We can act on the acceptance of the unpredictability. In this case, assuming that earthquakes/tsunamis are 'more likely' to strike along the south coast of Japan, demonstrates a reliance on a (flawed) prediction. And yet, the evidence demonstrates that the phenomenon is not predictable.

If it was me, I would've acted assuming that we have no idea as to where and when, the next earthquake would strike. This would then seemingly lead to a program of building reactors inland, instead of just in the northern part of the country and on the coast ! (For example).

We can make sound, very practical decisions based upon the assurance of unpredictability.

Cheers

[yusufcam]

its odd because i agree with a lot of what you are saying that i would still argue in part agianst it (its like yes, but :)).

i am not sure this is considered an established fact in the realm of science but isn't there the idea that animals can sense impending natural disasters?

its a widely accepted fact that migratory animals (birds, whales) have barely understood navigational skills which enable them to navigate huge distances across the globe.

they are in the natural world

what do you think?

[CraigS]

It is .. with mathematical assurance.


Perhaps, but as you say, this is just an idea.
When it comes to mathematical assurance vs ideas, I'll go where with a statement of assurance, with demonstrable evidence (testability), and pure logic backing it up.



Well, as you say, this is barely understood.

The unpredictability of earthquakes, on the other hand, is well understood, and assured.
:)
Cheers

[Outbackmanyep]

Earthquake alert period should be 3-4 days approximately prior and after new of full moon, where i can't tell you!
May be a little science behind it, i am not an expert, but from my observations of really big ones they are mostly 2-5 days after new or full moon.
The more active regions that have been making headlines recently are a good place to start!

If you subscribe to USGS Quake alerts you'll find that there are HEAPS of quakes every day!

[CraigS]

G'day Yep !

A little science, (maybe) :question:

Mathematical assurance behind unpredictability, though.

Hmmm :question:

I'll go with the flo' bro !
(Just playing around .. and having some fun here .. with a serious side to it all as well). :)

Cheers & good to see you back ! :)

[yusufcam]

the animals sensing natural disasters appears to be an idea that members of the scientific community are taking an interest in rather than dismissing.

the mathematical certainty of unpredictability, one could say as logically, is that they have yet to figured it out.

i suppose i lean towards the idea that its an unknown rather than unpredictable.

i'll stop here ...

[Brian W]

If it was me, I would've acted assuming that we have no idea as to where and when, the next earthquake would strike. This would then seemingly lead to a program of building reactors inland, instead of just in the northern part of the country and on the coast ! (For example).

We can make sound, very practical decisions based upon the assurance of unpredictability.


Let me see if I understand you correctly;

Because you have absolutely no idea where the next earthquake will strike you are going to build your reactor inland.

Basing this decision soundly upon the assurance of unpredictability!

Seems to me your process for making a decision is equivalent to drawing straws. The locality that draws the short straw gets the reactor cause there is no way to decide which place is safer.

As someone once said to me... garbage.

There is a little concept called probability. You build on a flood plain probably you will get flooded. Build on a major fault line and there is more of a probability of an earthquake levelling your structure than if you don't. Build on the coast and there is a higher probability that a really big wave or surge will hit you than if you build in the interior.

Certainly at the present it is not possible to predict when and where Mother Earth will turn violent but it is certainly possible now to state statistically which areas are more likely to be affected.

Brian

[CraigS]

But there again Brian, this is about mathematics !
:)

Cheers

[CraigS]

With the mathematical unpredictability … one could say logically that they may never figure out how to predict it !

… and that's regardless of how much we learn about it, from the present day observations. If we can never know precisely, the initial conditions of the process, then we can never predict the outcome ! If the outcomes we see are fractal in nature, then all of what I have said, applies.



That's cool Colin .. its an interesting area, and very much counterintuitive for me.
I have no agendas other that to project the perspective, for consideration.

I also lean towards it being 'unknown'.
However, I'll also rise to wrap my mind around the certainty that we will never know, and make plans on that basis (as well as a bit of intuition/curiosity, etc ;) ).

Counterintuitive overall however, to say the least.
:)
Cheers & Rgds

[yusufcam]

we are in agreement then......

life can be interesting