Saturday, June 04, 2011

Italian seismologists on trial as killers because they have no crystal ball

I have added some estimates of the frequency of quakes of different magnitudes and cost-and-benefits analysis of evacuation at the end

Another example of the insane witch hunts that are gradually returning our world back to Salem 1692.

Vitalik sent me a link to a stunning article at the Fox News server,
Italian Seismologists Charged With Manslaughter for Not Predicting 2009 Quake; Google News
Six earthquake scientists and an official have been charged - it's not just being talked about, it has happened - for manslaughter, a softer version of (mass) murder. They may spend up to twelve years in prison. Why? Because they didn't predict a 6.3 magnitude earthquake in L'Aquila that killed 308 people on April 6th 2009, including 2 Czechs. When did it happen? Now, in the 21st century. Where did it happen? In Italy, a country that is amazingly still a member of the European Union rather than a freshly separated province of Zimbabwe.

Where does the immense stupidity come from? Most things of this sort can't be predicted. At least, none of the major earthquakes has ever been successfully predicted as of 2011. If this were possible, lots of human lives would have been saved in the past. Why would someone believe otherwise? Where does the idea that similar rare events can be predicted come from?

Well, there's an answer to the last question. It comes from the media where other lunatics - mostly global warming alarmists these days - claim to be able to predict all kinds of wonderful things 100 years in advance.

Inside the Earth, things are constantly moving, in systematic as well as chaotic directions. Sometimes a piece of rock crosses a threshold, a tipping point, and in some of those cases, the resulting changes may make a bigger piece of rock cross another tipping point. And a big earthquake may quickly result in a very small percentage of such cases. In Italy itself, shaking similar to the L'Aquila aftershocks appears 100 times every day.

But to predict whether this sequence of events will lead to very large earthquake, you would need to know the detailed position of every grain of sand - and it could still fail to be enough. It's just like the question whether a 1-sigma bump in a scientific graph is going to grow to a 5-sigma bump. No one can know by looking at the bump itself. These bumps are everywhere and they look nearly identical. One may only interpret what happened a posteriori.

I personally think that the predictions of earthquakes that are larger than those that have just occurred at nearby places, days in advance, will always be impossible. There have been various preprints arguing that one could have seen signals of the Japanese earthquake for days but the papers don't actually show what the litmus tests are and were doing when there are no earthquakes. I don't think it's a good science. But even if this methodology could be developed to predict the quakes, we're not there yet.

Note that in a similar business, last year, people were predicting an imminent eruption of Katla volcano in Iceland. I was saying it was unlikely and the specific larger events are almost never "ignited" by smaller ones in such a straightforward way. As always, I was right, they were wrong. Even if Katla erupted now, it couldn't really be said to be the result of the E-ugly-name volcano eruption last year. There's another volcano erupting which would be more likely a "starter" of Katla if Katla decided to act now.

Quakes are a chaotic process. Even quantum randomness may influence the timing of big earthquakes. The only thing you can predict is that after a large earthquake takes place, it's likely that there will be various smaller "echoes" across the region whose locations may be predicted with a limited accuracy, using a probabilistic language. But getting bigger events out of the smaller ones always depends on chance.

There were dozens of aftershocks but only one foreshock, a 4.1 magnitude earthquake 1 week earlier. Earthquakes of the 4.1 magnitude are not even recorded on the USGS maps. Only a supertiny portion of 4.1 earthquakes may be "predictors" of a coming magnitude 6.3+ earthquake. It was just sensible for the seismologists to stop the panic because if every 4.1 earthquake led to an evacuation of a town for weeks, several percent of the mankind would be constant refugees.

If someone suggests that the seismologists should have recommended the evacuation of the town with 80,000 people for weeks because of the 4.1 foreshock (plus some much smaller trembling), just consider how crazy it would be. Exactly a week ago, there was a 4.8 magnitude earthquake 30 miles from Chania, Crete, which also has 80,000 people or so, and similar cities. (You may check hundreds of other examples of earthquakes predicting a doom just in the recent month.) Would you really say that because of that - a possible sign of a coming big earthquake - the local seismologists in Crete, when asked, should have recommended the evacuation of those cities for weeks? Hundreds of thousands of refugees? Have you lost your mind? 99.99... percent of such smaller earthquakes are not foreshocks of much bigger ones.

What the seismologists recommended not only fails to be a crime: it was the right thing for them to do given the information they had and the current state of their discipline. A Czech seismologist, Mr Jan Zedník, nicely says that reliable warnings before big earthquakes are only possible in Hollywood movies in which a blonde female seismologist calls the mayor sufficiently many times so that they finally evacuate Los Angeles. ;-) This is spot on. Billions of people fail to distinguish the movies (and equally fictitious predictions of climate-change-related catastrophes in the media) from the reality. And billions of people began to worship their governments, being certain that the governments should protect them even against the laws of physics.

Now, just try to evaluate how many complete idiots have to exist in the whole hierarchy of the "dignified luminaries" in Italy. The general prosecutor of that country probably has to be a complete imbecile because it had to go through him and the same has to be true for a whole pyramid of his subordinates, all of their secretaries, and many others who had to have some tools to stop this insanity.

Meanwhile, if you live in a country where morons such as these Italian lawyers or global warming alarmists prevail, the only relatively safe way how not to spend your life in prison or at stake is to say that all conceivable kinds of catastrophes, all earthquakes, volcano eruptions, tornadoes, hurricanes, typhoons, floods, collisions with other planets etc. will occur at every place of the Earth at every second in the future. Almost no one will harass you for talking rubbish and many people will applaud you. That's the only way how you may "become" innocent. Otherwise, together with the fellow seismologists and industrialists who use fossil fuels, you will be charged of manslaughter for causing all these events and killing the people!

You're a witch! Burn you!

Sorry, I got carried away by the mass hysteria myself. It's everywhere. ;-) I am sure that the Italian judges will be inspired by the video above and they will weigh the seismologists. If they weigh as much as a duck, they must be made of wood, and therefore they are witches.

Frequency of quakes and foreshocks

About 40 percent of L'Aquila-like, larger earthquakes have foreshocks. But can you evacuate the nearby towns every time you experience trembling? For weeks or months or years?

The Gutenberg-Richter law says that the frequency gets reduced ten-fold every time you increase the magnitude of the quake by 1 degree. So the frequency of magnitude 6+ earthquakes is about 100-times lower than that of the magnitude 4+ earthquakes.

Moreover, they're not "quite" correlated. At most 40% of the larger magnitude 6+ earthquakes have magnitude 4+ foreshocks - and probably many fewer because among the 40%, the foreshocks included may be much weaker, too. Also, foreshocks at various delays - time scales - are included as well, making the predictive usefulness of foreshocks even lower. Moreover, the later quakes don't necessarily have to occur on the same place, so our predicted odds that the quake could occur at the bad place will be overestimates.

It follows that the probability is less than 0.5% (and probably much, much less) that an ongoing magnitude 4+ earthquake is a sign of a coming magnitude 6+ earthquake beneath the same town. And probably much smaller. Now, think sensibly: would you evacuate the town of 80,000 which would be needed to systematically try to save almost all lives?

Note that some greatest earthquakes such as the 1950 Medog earthquake display no foreshock activity at all. The foreshocks are simply a very poor litmus test to predict larger earthquakes.

Take a sensible quantification. First of all, you must quantify the human lives whether it looks bad or not. Take one million euros per life which is a lot. So the earthquake ultimately killed 300 people which is 300 million euros in damages. However, the probability that this would occur was only 0.5%. So the expectation value of the damages from the dead people is just 1.5 million euros.

On the other hand, you have the expenses needed to evacuate the town. In order for the evacuation to be a good investment, you would need the evacuation of 1 person - among 80,000 - to cost just 1,500,000 euros divided by 80,000 = 19 euros. I guarantee to you that the actual evacuation is going to be much more expensive than 19 euros. If you want to store the people for a month - before the threat goes away - in a different place, they will have to pay hundreds of dollars for some "rent" or its equivalent, or count the corresponding reduction of the living standards. You should really subtract their salaries as well because they will probably not work. That's already thousands of euros per person.

So at least by an order of magnitude, and probably two orders of magnitude or more, it would be a lossy investment to evacuate the town. You would have to say that a human life is 10 million euros. And it's just crazy. Almost no life insurance can get that high. People ultimately die - and they may also die from the shortage of money (that's the case of millions in the third world) which is why it is simply wrong to say that the missing money can never match the loss of human lives. Putting the price of a human life to an infinite value is incompatible with a rational policymaking.

Note that the argument above doesn't change if you imagine that the seismologists would only advise "voluntary" evacuation to a subset of the people. For these people, the expectation value of the expenses would still exceed the expectation value of the risk of their life converted to the money.

If someone wants to put his life to 100 million euros and use a very different cost-and-benefits analysis than what is above, he has to do it privately because the officials are working with the population statistically, as an average. If he's waiting for someone else to escape even though he's scared of the slightest risk of a quake, well, then I would say that it shows that his intelligence is probably low enough for the claim that his life is worth 100 million euros, anyway. The price could be closer to 5 euros for the beer his parents drank before they made him.

If you really hate converting human lives to the money, I can actually offer you an argument that avoids it altogether and the conclusion is still the same. Evacuation is not safe, either. When New Orleans was evacuated, dozens of people died in the Dome and other centers for the refugees. That's more than 1 percent of Katrina's casualties.

So the evacuation would statistically save just 0.5% (chance of larger quake) times 300 = 1.5 people, but it would kill 1% (percentage of casualties due to evacuation) times 300 = 3 people in average because of the problems with the evacuation. So regardless of the expenses, the evacuation would actually kill more people than the large quake if both things are computed as proper expectation values.

The actual merits of the evacuation are even worse. In reality, the same trembling was observed in that region at least since December. So you would need to evacuate the people for half a year if not permanently. That could be equivalent to building 50,000 euro worth of new housing for each average inhabitant of L'Aquila. Compare this with the expected loss due to the possible larger earthquake which was 19 euros.

All this reasoning may look too cruel and cold-blooded but if you use pure emotions, you simply can't decide whether people such as the Italian seismologist gave a good advice according to their best knowledge at that time. They did.


  1. My understanding is that the situation
    was this: there were tremors for days.
    Many people had started to leave the town fearing a large earthquake.
    These guys went on television and
    told people not to leave town because
    they were certain there would be no

  2. Fine. Even though it's obvious that small earthquake make larger ones more likely than normally, they determined that it was very unlikely that a big earthquake would occur, and by saying that they were "certain", they wanted to give it some weight to avoid a mass hysteria that was inappropriate in that case.

    Mother Nature proved their guess wrong.

    But it's still important and obvious that one can never be certain that there won't be a big earthquake in an earthquake area. Never, ever. If someone is able to make better predictions than they did, he should have left despite their words. But such people will also be doing hysterical things in thousands of other contexts where they will be proved wrong.

    It's impossible for those people who try to predict similar things to be always right because they depend on chance. So as I mentioned, the only way how to "satisfy" insane people who expect the seismological "shamans" to always warn them is to say that there is always a coming tragedy.

    I strongly protest against this pressure. Scientists must be erring equally often in both directions. Scientists saying that there will always be a tragedy are not scientists. I prefer a few hundred Italians to be randomly killed by an unpredictable disaster over sacrificing the pillars of the whole modern civilization.

  3. Yes I agree with you.
    It is insane to prosecute
    these guys. But is is also
    bad for scientists to go on TV
    and not acknowledge uncertainty.
    We (scientists) should be leading
    by example.

    So I am saying they are guilty
    of bad science but should not be
    legally culpable.

  4. The matter is not about prediction, which is impossible, I agree. The matter is prevention. The matter is people education, without making alarmism. The matter is to check if the basic antiseismic rules was respected for buildings (which are not, someone used too much sand) in that high risk zone.

    I suggest this article from SISSA University to Italian readers.

    (sorry for my bad english)

  5. There are lots of buildings and other facilities, and whole nations, across the world located at needlessly dangerous places which are severely suboptimal. That's how the civilization has developed. Civilization was not designed by perfectionist 25th century engineer. It was produced by ordinary people in the past centuries.

    One can't blame seismologists for the fact that Italian cities are located on wrong spots. It's really not their business.

    It's excellent to acknowledge uncertainty but TV viewers these days just don't understand the concept quantitatively. If the seismologists think that the chances of a big quake are 0.01% or less, and they apparently did, they will not talk about uncertainty because that would lead to mass hysteria that could, by the way, lead to the same number of deaths.

    They were wrong but there are many other cases in which they're right. But because no one dies. no one discusses it. The discipline of seismology is complex and noisy so that the guesses of the seismologists may not be too much better than the nearly random guessing based on overall frequencies and/or some simple laws how the frequencies are affected by foreshocks.

    But no one can do better.

    These folks just concluded that the probability of such a thing was nearly zero. No one else could have known that it was wrong. Their argument was utterly sensible - the energy was being gradually released which looked innocent. Well, it wasn't. But they didn't know it in advance. No one did. So they can't be blamed for it.

  6. As it is not possible to say wether an earthquake is to occurr, is equally impossible to say that it is not occurring. But when you have an unusual activity the only thing you can do is to warn people that a bigger event may happen. Warned people can decide to wait the end of the activity period in their house, or to move. Obviously everyone can detect a little earthquake, it's not a matter of sofisticated instruments, so everyone was aware that something unusual was going on. What those seismologists did wrong is to "certify" that nothing was to happen. Thay have not be condemned for having foreseen the big event, but for having assured that it was not going to happen. So they have been condemned precisely for having predicted something

  7. "As it is not possible to say wether an earthquake is to occurr, is equally impossible to say that it is not occurring."

    Dear Alessandro, except that it is vastly more likely that a big quake will *not* occur.

    The people couldn't have been "warned" because there was no good scientific reason to expect that exactly at that moment, there would be a big quake.

    "What those seismologists did wrong is to 'certify' that nothing was to happen."

    Because that was the best result they could get from their science.

    "So they have been condemned precisely for having predicted something."

    Right, exactly. They were condemned for being scientists and for doing predictions which is an inseparable part of science, moreover in a country where the judges and much of the public has no idea what science is and in a moment that just happened to see an earthquake which is a coincidence that have absolutely nothing to do with.

  8. I do not completely agree with you because one thing is to say "it's very unlikely that an earthquake will occurr. Despite of the signals we're having chances are really low" and the other is to say that the event will not occurr. They deliberately discouraged people from taking action accordingly with the possibility of a bigger avent, as they people was already doing. Moreover there are other scientists ( that believe that the bigger event was to take at least into account because of the unusual activity. Anyway they have not been condemned, but charged and now the trial is starting. They are officially culprits.
    I do agree with you when you say that noone has idea of what is and what is not scientific.

  9. I never wrote that no one has any idea what is scientific. I wrote that you had no idea what is scientific.

    Scientists may have other guesses what will happen with the seismic activity at a given place and time but one earthquake is not enough to evaluate who actually has better methods and knowledge of the underlying science.

    You have no evidence that the "other scientists" you are mentioning actually understand this discipline more than the people whom you openly want to harass in the court.

  10. you said that "judges and much of the public has no idea what science is", not that I have no idea of what science is :)
    anyway the guy that i've mentioned, as others and evidently the judges, never thought that anyone has an understandment of the underlying science. The fact here is that someone gave wrong indication about the possibility of a major event. I think that we are looking at this from the same point of view, but we see different things. Since noone has clues about the next big event, why those guys were so sure about the fact that it wouldn't happen? so it's them to have thought to know the underlying science. I believe it's quite scientific a trial to determine if they could have done better if they hadn't say anything.
    Exactly they are members of the "national high risks committee" and they had the responsibility of deciding if the situation in L'Aquila was "zero danger" or "danger" or "evacuate". They decided, after 4 months of daily quakes, that it was "zero danger". From my point of view it is enough for letting those people answer for the responsibility they are paid for.

  11. You're just making it up, medieval bigot. Show me the place in their job contract that gives them responsibility for earthquakes in the sense that if a large enough earthquake occurs and kills people in Italy, they will be arrested for 12 years.

    It is completely preposterous. People employed under such a job contract wouldn't be scientists, anyway. They would be playing Russia roulette.

    Every year, between hundreds and hundreds of thousands of people are killed by earthquakes globally. Chances are that every decade, to say the least, it includes Italian people. It is very likely that someone in Italy would die because of that - and that a seismologist signing to the insane job contract you mention would go to prison.

    But it is still extremely unlikely that such a large earthquake will occur under a particular town in a particular month, which is why it was correct to translate the probability to plain Italian as zero.

    They did the right thing and idiots want to blame them for the earthquakes because they're idiots, and malicious ones. Scum like you was preventing the progress and the birth of science on the Earth for long centuries, and if it will be necessary to nuke people like you to avoid the return to the Dark Ages, I won't hesitate in offering my full endorsement for the operation.

  12. I laughed so hard when I read this. Thanks for articulating some of my thoughts on this subject. :)