Subject: Re: Assessment of predictions (was Earthquake advisory cancelled 11/24/96)
From: stgprao@sugarland.unocal.COM (Richard Ottolini)
Date: 9 Dec 1996 14:06:38 GMT
In article ,
Dennis Gentry wrote:
>This gets back to the reliable argument. If a method can only
>predict a subset of quakes, and that method (at least for the
>posted predictions) displays a better than chance hit rate, would
>that method still be considered reliable?
There is another statistical measure called "significance" that
answers your argument. That measure evaluates the results of
more than one experiment. Successfully predicting a few low probability
events or lots of high probability events can be significant.
I haven't seen that calculated in this newsgroup, although I believe
it has been used to evaluate the Greek electric prediction methods
discussed in the May 26, 2996 issue of Geophy Res Letts(?).
Significance is used in medical epidemology all the time.
Rare events might include leukemia near a power line or common events
such as coffees's link with breast cancer. As with earthquakes,
statistical results are debatable until the precise physical causal
mechanisms are discovered.
Subject: Sulfates bigger than earthquakes!
From: Harold Asmis
Date: Tue, 10 Dec 1996 09:53:09 -0500
From the Newsedge wire service, talking about sulfate concrete rot in
California.
****
One major concrete subcontractor, who asked not to be identified, said
he believes damages "could cost Southern California [builders']
insurance companies as much as the Northridge earthquake."
****
I'm a geotechnical engineer, and this is great stuff! Apparently, those
silly Californians :) don't like to build good Canadian-style
tile-drained basements, but instead lay a concrete slab on grade.
According to this article, the builders (and mix contractors) say they
don't have to follow code, and the customer never wants the expensive
stuff (code specified sulfate-resistant concrete). So now, all these
undrained slabs are sucking up sulfates and rotting. The only solution
is to jack up the whole house and repour the foundations. Can't wait
until the next earthquake hits these houses! That $10 billion of the
CEA is going to be drained instantly. :)
--
Harold W. Asmis harold.w.asmis@hydro.on.ca
tel 416.592.7379 fax 416.592.5322
Standard Disclaimers Apply
Subject: Re: Assessment of predictions (was Earthquake advisory cancelled 11/24/96)
From: russ@seismo.demon.co.uk (Russ Evans)
Date: Wed, 11 Dec 1996 07:48:25 +0100
Dennis Gentry wrote:
> Thanks Russ. Actually, as far as serious statistics, I didn't see to
> much more than has been discussed on this ng. The abstracts were
> very interesting though and enjoyed reading them. At least the ones
> that were available.
The heavier statistics isn't there yet, in part because of problems in
setting mathematics in HTML, and in part because I haven't yet got round
to sorting stuff out. It will certainly be there in the proceedings
volume.
> One gentleman mentioned that "new mathematical and theoretical tools"
> will be needed for earthquake prediction. I wasn't quite clear on what
> the author was trying to get across here. Was it in reference to tools
> for evaluating earthquake predictions or tools for earthquake predictions?
I forget who made that reference. There are some problems with the
available statistical methods, and the principal technical outcome of
the meeting was that the statisticians present felt that they were
converging on ideas of what was needed. (Remember that abstracts tend
to state what the authors feel they already know, rather than their
questions, which are what get discussed at a 'Discussion Meeting').
> Also, I missed seeing anything about Alan Jones method for tracking
> earthquake predictions (a few abstracts did discuss the null hypothesis
> to which I thought Alans' process handled pretty well).
Ummm. There was a key point in Phil Stark's talk that, because
earthquake occurrences are not distributed in time as a Poisson process,
any method which depends on messing about with that series (either by
declustering to generate a quasi-Poissonian distribution or by some form
of randomisation to generate alternate datasets) is liable to fall into
the trap of falsely rejecting the null hypothesis. Last time I looked
at Alan's method, he was employing declustering. (I'm open to
correction or update on that, of course, Alan!) In plain language,
this means that such tests are too weak -- it is possible for schemes
which do not work to score more highly than they deserve. Alan's tests
have the great benefit of simplicity, so, as long as people understand
that they are the first of a series of hurdles which their proposals
have to clear, I think they are right for this forum.
> Another thing that was interesting was that several methods at earthquake
> prediction were discussed but are being discarded as unreliable due to
> one of the big three factors being missing such as location. For those
> that were tagged as being unreliable due to location, I didn't see
> anything about attempts at triagulation. Didn't see triangulation
> discussed at all. Maybe in language that went over my head?
I tried very hard to head people off from talking too much about
specific schemes (and, having attended quite a few such meetings, I do
think this was the main reason that this one made progress). If we mean
the same thing by 'triangulation' (using multiple observations to refine
an estimate for a prediction), I would classify this (at least for the
purposes of the meeting) as belonging at the level of operational
detail.
> Looks like it was a great meeting though.
As I said, I feel that we really moved the issue along a bit, which is
always a very positive feeling.
Russ
p.s. I have part-prepared responses to a number of other points made in
this thread, but completion and posting will have to wait a few days, so
'more later'.
Subject: Re: Alaskan 1964 e/quake
From: jewett@netcom.com (Bob Jewett)
Date: Wed, 11 Dec 1996 19:26:57 GMT
Jon Kirk (Jon@jonkirk.demon.co.uk) wrote:
: Can anybody please tell me where on the Internet I can find specific info
: about the Alaskan earthquake of 1964?
If you go to the Alta Vista search engine (http://altavista.digital.com/)
and type
+Alaska +earthquake +1964
(The plus signs require rather than suggest each of the key words.)
you will get about 400 references including:
1964 Alaska Tsunami Publications
1964 Alaska Tsunami Publication List. This is an all-inclusive
list of publications, organized alphabetically by first
author. Abstracts in HTML often....
1964 Prince William Sound Tsunami Images
Alaska Earthquake 1964
Good Friday, 1964. It was a bright, sunny day in Alaska. The
streets were crowded with people going to Good Friday services,
coming home from work ...
The last is an ad for a 22-minute video "using rare footage, much of which
was shot during the quake ..."
Bob
Subject: Equals the g-force all over the world
From: Knut Ove Hauge
Date: Thu, 12 Dec 1996 11:45:50 -0800
The question is. Will measurement of the g-force equals all over the
world. Beacause the g-force is obtained from the planets,the Moon and the
Sun it could be belived that the g-force would be equal all over the
world,except from a little difference as in the gravty on the poles.
The occurence of large earthquakes also prove that the formula used in
EQF is a common formula all over the world.
But what about the temperature measured in Stavanger,Norway compared to
the g-force at that place. It seems that the temperture is an inverse
funtion of the g-force, and the temperature does not equals all over the
world. Since I started to publish the measuring results 19.th of June
1996, I have 3 visitors every day that have read the measuring
results.After half a year of publishing they could answer the question.
If they are not reading this, or are not willing to comment this, only
further measurements done all over the world could answer the qustion.
Look at http://home.sol.no/ovehauge/
