| |
 |
|
|
Science Forum Index » Geology - Earthquakes Forum » The probability of earthquakes is now foreseeable
Page 2 of 2 Goto page Previous 1, 2
|
| Author |
Message |
| Mike Williams |
 Posted: Fri Apr 18, 2008 11:00 am |
|
|
|
Guest
|
"Physikus" <physikus1@gmx.net> wrote in message
news:4f12c8cd-8f65-473a-8526-fa2a7d56fd97@y21g2000hsf.googlegroups.com...
On 16 Apr., 18:03, "Mike Williams" <miklw...@pacbell.net> wrote:
Quote: "Physikus" <physik...@gmx.net> wrote in message
...
What an incredibly archetypical netkook, pseudoscientific response!!
"Magnitudes the system doesn't predict (nobody can know in what way
earth's crusts are prestressed)!"
The three-legged stool of earthquake prediction quite simply, and
rationally, does not provide for the elimination of any of the individual
legs.
Mike Williams
Arroyo Grande, CA U.S.
"pseudoscientific" ?
Quote: Look:
1.
Found at 03:02 pm at http://eqf.homedns.org/ (Germany:UTC+2,
UTC=13:02):
15:02:14 19.15 51.350 8 1 1057 Polen (predicted)
Very close to the time when this happened:
2008-04-18 13:01:12.4 50.09 N 18.42 E 2 ML 2.3 POLAND
(Source: http://www.emsc-csem.org)
2.
Found at 12:35 pm at http://eqf.homedns.org/ (Germany:UTC+2,
UTC=10:35):
12:35:22 36.92 36.61 1 1 2 177 Türkei & (predicted)
12:35:48 36.92 36.37 1 1 3 177 Aleppo/Syrien (predicted)
Very close to the time when this happened:
2008-04-18 10:28:17.2 36.76 N 36.72 E 6 MD 3.3 TURKEY-SYRIA BORDER
REGION
(Source: http://www.emsc-csem.org)
Matching times & matching coordinates!
Physikus
You have GOT to be kidding!
1) - you must realize that anybody can score 100 "hits" per day easily by
simply specifying approximate locations and times (and leaving the degree of
approximation out of the prediction). And leaving magnitude out entirely!
2) - what remarkably small magnitudes seem to fulfill your "predictions"!
Especially for such seismically active regions as the "Turkey-Syria border".
Not so sure about Poland.
3) Can you please provide a complete list of ALL your predictions for, say,
the one-month prior to your two claimed hits, along with the hits that
fulfilled them? Said list to include all predictions that were not
fulfilled, also.
Mike Williams
Arroyo Grande, CA U.S. |
|
|
| Back to top |
|
| Thomas A. Russ |
| Posted: Fri Apr 18, 2008 1:08 pm |
|
|
|
Guest
|
Physikus <physikus1@gmx.net> writes:
Quote: On 16 Apr., 18:03, "Mike Williams" <miklw...@pacbell.net> wrote:
"Physikus" <physik...@gmx.net> wrote in message
...
What an incredibly archetypical netkook, pseudoscientific response!!
"Magnitudes the system doesn't predict (nobody can know in what way
earth's crusts are prestressed)!"
"pseudoscientific" ?
Look:
1.
Found at 03:02 pm at http://eqf.homedns.org/ (Germany:UTC+2,
UTC=13:02):
15:02:14 19.15 51.350 8 1 1057 Polen (predicted)
Very close to the time when this happened:
2008-04-18 13:01:12.4 50.09 N 18.42 E 2 ML 2.3 POLAND
(Source: http://www.emsc-csem.org)
2.
Found at 12:35 pm at http://eqf.homedns.org/ (Germany:UTC+2,
UTC=10:35):
12:35:22 36.92 36.61 1 1 2 177 Türkei & (predicted)
12:35:48 36.92 36.37 1 1 3 177 Aleppo/Syrien (predicted)
Very close to the time when this happened:
2008-04-18 10:28:17.2 36.76 N 36.72 E 6 MD 3.3 TURKEY-SYRIA BORDER
REGION
(Source: http://www.emsc-csem.org)
Matching times & matching coordinates!
Actually, the times and coordinates don't match. In fact, the
coordinates are off by 150km, 24km and 46km. In the first case, I doubt
that one could even feel a 2.3 magnitude quake from 150km away. In the
latter case, it seems you are trying to claim two hits from one event.
So, what precisely are the bounds in time and space for these
predictions?
And if the predictions include events that one can't even detect
(without instrumentation) at the predicted location, of what practical
use are these predictions?
Furthremore these are apparently also post-dictions rather than pre-dictions.
From looking at your web site, it appears -- at least from the demo --
that you provide "predictions" for 10 minutes into the future, based on
the request time.
So it looks like at 15:02 you "predict" and event that occured about 1
minute earlier. For the Middle East, it appears that the prediction
from 12:35 is "predicting" an event that occurred 7 minutes earlier.
It's pretty easy to get hits if you wait until after the event in order
to predict it.
--
Thomas A. Russ, USC/Information Sciences Institute |
|
|
| Back to top |
|
| Thomas A. Russ |
| Posted: Fri Apr 18, 2008 1:19 pm |
|
|
|
Guest
|
Physikus <physikus1@gmx.net> writes:
Quote: On 16 Apr., 04:54, Skywise <i...@oblivion.nothing.com> wrote:
Physikus <physik...@gmx.net> wrote in news:81a21bd0-6b9b-4aa1-8716-
c88757e90...@m3g2000hsc.googlegroups.com:
But anyway, what we need are defined locations, defined time
frames, and defined magnitudes. Anything less is pretty much
useless.
The service provides both, as you can see, defined locations and
defined time frames.
No it doesn't. It provides a single instant of time, percise to one
second, and a single point, precise to 0.01 degrees (roughly 1km) with
no indication of any intervals in either time or space.
Quote: But nobody is able to predict the strongnes of
predicted EQs, because the parameters of the inner crust is unknown,
at all. Nevertheless the service shows, whereever and whenever a
danger occurs.
Well, if there are no lower bounds on the size of the earthquake, I
think that the word "danger" is rather overstated. Most earthquakes
that occur are so small that they are not felt by anyone on the earth's
surface. So that hardly constitutes a danger.
--
Thomas A. Russ, USC/Information Sciences Institute |
|
|
| Back to top |
|
| Thomas A. Russ |
| Posted: Fri Apr 18, 2008 1:42 pm |
|
|
|
Guest
|
Translation follows.
No absolute guarantee for accuracy, but I am fairly confident about the
accuracy.
There is no reflection of any thoughts I might have on the subject.
Physikus <physikus1@gmx.net> writes:
Quote: On 17 Apr., 20:21, t...@sevak.isi.edu (Thomas A. Russ) wrote:
Physikus <physik...@gmx.net> writes:
...
--
Thomas A. Russ, USC/Information Sciences Institute
Please let me answer in German:
Also denn, was ist eigentlich ein Erdbebenwahrscheinlichkeitspotenial?
Was heisst es, wenn dieses Potenial erhoeht sein soll? Was sollte man
darunter verstehen?
Unter einem Erdbebenwahrscheinlichkeitspotential (EWP) verstehe ich
ein relatives Risiko bezŽüglich des Eintreffens eines Bebens.
An earthquake probability potential (EWP) is the relative risk of the
occurrence of a quake.
Quote: Und was soll man unter "real earthquake" verstehen? Soweit ich geshen
habe, ist da keinerlei anerkannten Groessenangabe (z.B. Richterskala),
kein mindestwert von Erdbebenstaerke. Insofern ist es schwer to
erkennen, was man aus einer Vorhersage machen soll, wenn man nicht weiss
ob das Erdbeben ueberhaupt spuerbar sei.
Die Voraussage bezieht sich auf das Aufkommen und Zusammenwirken
unterschiedlicher (KrŽäfte-) Faktoren zum selben Zeitpunkt, am selben
Ort. Hieraus wird das EWP (s.o.) abgeleitet.
Unter reale Erdbeben sind Beben gemeint, die tatsŽächlich - nach der
Vorhersage - stattfinden.
The prediction is based on the emergence and confluence of different
(force-) factors at the same timepoint and at the same location. From
this, the EWP is derived. The term "real earthquakes" means quakes that
actually occur -- after the forecast.
Quote: Magnituden werden dabei nicht vorhergesagt, da das System nicht wissen
kann, inwieweit die regionale Erdkruste vorgespannt ist; die Prognose
bezieht sich daher lediglich auf das regionale Vorhandensein Erdbeben-
auslŽösender Faktoren, zum vorausgesagten Zeitpunkt.
Magnitudes are not forecast because the system cannot know, to what
extent the regional earth's crust is pre-stressed. The forecast
therefore is based only on the presence of earthquake-initiating
factors, at the predicted timepoint.
Quote: Ist die angegebene Zeit exakt? Oder gibt es ein Interval um die
angegebened Zeit? Oder ist die Zeit der spaetesten Zeitpunkt von so
einem Interval, der mit der Verhersagezeit anfaengt? Und was ist mit
den geographischen Grenzen? Muss das Erdbeben genau an dem angegebenen
Ort stattfinden, genau bis zum 0,01 Grad?
Die Prognosen werden derzeit etwa alle 5-10 Minuten erstellt. Das
betreffende Event sollte sich innerhalb von +/- 0.00-3.00Ž° (LŽänge &
Breite) und zeitlich innerhalb von +/-10 Minuten ereignen; genauer ist
das vorliegendende Kartenmaterial leider nicht (fast immer liegen die
Treffer aber innerhalb eines Intervalls von +/- 0.00-1.00Ž° und +/- 4
Minuten).
The forecasts are currently created every 5-10 minutes. The target
event should occur within +/- 0.00-3.00 degrees (latitude and
longitude), and temporally within +/- 10 minutes. The available map
information is unfortunately not any more exact than that. (The hits
are almost always within +/- 0.00-1.00 degree and +/- 4 minutes.)
Quote: Infolge der Kartenungenauigkeiten sind Grenzregionen etwas
problematisch: So kann z.B. eine Grenzkoordinate der TŽürkei-Syrien-
Grenze noch innerhalb der TŽürkei oder noch innerhalb Syriens
ausgegeben werden; an dieser Problematik wird jedoch intensiv
gearbeitet.
Because of the inexactness of maps, the border regions are somewhat
problematic. For example, a border coordinate of the Turkish-Syrian
border can be output inside Turkey or else inside Syria. This problem
is being worked on intensively.
Quote: Gradangaben werden dennoch auf 2 Stellen hinter dem Komma ausgegeben,
da sie intern mit dieser Genauigkeit berechnet werden.
Degree values are listed to 2 decimal places because they are interally
calculated with that precision.
Quote: Diese angaben sind natuerlich sehr schwer auszuwerten. Und keiner von
ihnen sind exakt, d.H. die Vorhersage spricht nur von erhoehter
Wahrscheinlichkeit. Wie ist es denn moeglich zu behaupten, dass diese
Methode 100% richtig ist, da die Vorhersagen selber oft wehniger als 50%
Wahrscheinlichkeit beinhalten.
Der EWP-Wert gibt die vom System errechnete zu erwartende Gesamtkraft
an, durch welche ein Beben regional getriggert werden kann.
GrundsŽätzlich werden nur dann DatensŽätze ausgegeben, wenn das System
ein erhŽöhtes Potential (>0) feststellt. Ist also der EWP-Wert >0, so
ist ein Beben grundsŽätzlich indiziert, unabhŽängig davon, wie hoch der
EWP-Wert ist; denn - wie schon erwŽähnt: man kann nicht wissen, ob die
Kruste auf die vorherrschenden KrŽäfte reagiert. BenŽötigt die regionale
Kruste als Triggerung nur wenig Kraft, genŽügt bereits ein kleiner EWP-
Wert - und umgekehrt.
The EWP-value presents the system-calculated expected total force
through which a quake could be triggered in a region. Fundamentally, a
data item will only be output if the system detects an elevated
potential (> 0). If the EWP-value is greater than 0, then an earthquake
is fundamentally indicated, independent from the size of the EWP-value.
This is because -- as already mentioned -- one cannot know if the crust
will react to the prevailing forces. If the regional crust only needs a
small amount of force to trigger, then a small EWP-value suffices -- and
vice versa.
Quote: Ich hoffe, alle Fragen zufriedenstellend beantwortet zu haben.
I hope I have adequately answered all the questions.
Quote: Da meine Englischkenntnisse fŽür eine so ausfŽührliche Antwort leider
nicht ausreichen, wŽäre es sehr nett, wenn jemand, der die deutsche
Sprache ausreichend beherrscht, wie offensichtlich Herr Russ (vielen
Dank fŽür die zweisprachige AusfŽührung Ihrer Fragen!), meine Antwort
fŽür alle Žübersetzen kŽönnte. HierfŽür vielen Dank im Voraus.
Since my English ability is not sufficient for such an extensive answer,
it would be nice if someone who understands German sufficiently well,
like apparently Mr. Russ, (Thanks for the bilingual presentation of the
questions) could translate the answers for everyone. Many thanks in
advance.
--
Thomas A. Russ, USC/Information Sciences Institute |
|
|
| Back to top |
|
| Physikus |
| Posted: Sat Apr 19, 2008 12:03 am |
|
|
|
Guest
|
On 18 Apr., 20:08, t...@sevak.isi.edu (Thomas A. Russ) wrote:
Quote: ...
Actually, the times and coordinates don't match. In fact, the
coordinates are off by 150km, 24km and 46km. In the first case, I doubt
that one could even feel a 2.3 magnitude quake from 150km away. In the
latter case, it seems you are trying to claim two hits from one event.
The System puts a matrix with 1.200.000 points (locations) around the
world. From these points it finds up to 2% locations with higher EQ-
potentials. This is not a gag or a game! Behind this methode stands a
theory, developed since 1990.
Because the System cannot use a matrix with a higher density (to hold
the database in a functionally great), the matrix-points are
approximately 110 km and also approximately 8 minutes in distance to
each other. At this time the forecast cannot be more accurate.
The country-borders are not very exactly yet, within the system, so a
point might be given out as a part of two countries.
Quote: And if the predictions include events that one can't even detect
(without instrumentation) at the predicted location, of what practical
use are these predictions?
The forecast of locations and times with heigher potentials, delivered
by the system, can safe lives.
If a weather-forecast predicts a thunderstorm at your location, you
will usually stay at home. If you know about a danger of high EQ-
potential within a region, you will not enter the location (or leave),
for example.
Quote: Furthremore these are apparently also post-dictions rather than pre-dictions.
No. the times are within an interval of ca. 10 minutes, depending on
the scale (distance of the points) of the matrix.
Quote: From looking at your web site, it appears -- at least from the demo --
that you provide "predictions" for 10 minutes into the future, based on
the request time.
No, not at all. The site delivers only up to 20 minutes at a test-
request. But the system can give a forecast for up to several weeks in
the future.
Quote: So it looks like at 15:02 you "predict" and event that occured about 1
minute earlier. For the Middle East, it appears that the prediction
from 12:35 is "predicting" an event that occurred 7 minutes earlier.
It's pretty easy to get hits if you wait until after the event in order
to predict it.
No. I already explained this point above. If you get a forecast of 10
to 20 minutes, you might find a real earthquake later on, matching
the list.
Physikus
P.S. I tried to give this answer with my (kn) own english words. So,
you should find several grammatical mistakes, i think. |
|
|
| Back to top |
|
| Physikus |
| Posted: Sat Apr 19, 2008 12:06 am |
|
|
|
Guest
|
On 18 Apr., 20:42, t...@sevak.isi.edu (Thomas A. Russ) wrote:
Quote: Translation follows.
No absolute guarantee for accuracy, but I am fairly confident about the
accuracy.
...
Thomas A. Russ, USC/Information Sciences Institute
It is very kind of you, to spend your time in such a professionel
translation!
Thank you very much!
Physikus |
|
|
| Back to top |
|
| Thomas A. Russ |
| Posted: Mon Apr 21, 2008 12:44 pm |
|
|
|
Guest
|
Physikus <physikus1@gmx.net> writes:
Quote: And if the predictions include events that one can't even detect
(without instrumentation) at the predicted location, of what practical
use are these predictions?
The forecast of locations and times with heigher potentials, delivered
by the system, can safe lives.
No it can't. There is not enough specificity.
Quote: If a weather-forecast predicts a thunderstorm at your location, you
will usually stay at home. If you know about a danger of high EQ-
potential within a region, you will not enter the location (or leave),
for example.
If a weather-forecast system predicts a thunderstorm somewhere within
500km of my location, I'll likely not do anything different at all.
Perhaps a much better analogy would be to say that the forecasts you are
predicting are for an increased chance of a windy day. But weather this
wind will be high enough to be noticeable is not part of the
prediction. So, all we get is a forecast that there is a "higher"
chance of wind occurring, it is somewhere within 500km and the wind
could be anything from a light breeze that you would not even feel
against your skin to a hurricane -- but we can't tell you which one it
is.
Given that lack of specific characteristics, one would have to be insane
to leave a region, just because there is some unquantified elevated
chance of something happening, where it is much more likely to be not
felt and far away. Because of the power law distribution of
earthquakes, low-magnitude earthquakes that are generally detectable
only through instrumentation will occur much, much more frequently than
large earthquakes.
Quote: Furthremore these are apparently also post-dictions rather than pre-dictions.
No. the times are within an interval of ca. 10 minutes, depending on
the scale (distance of the points) of the matrix.
From looking at your web site, it appears -- at least from the demo --
that you provide "predictions" for 10 minutes into the future, based on
the request time.
No, not at all. The site delivers only up to 20 minutes at a test-
request. But the system can give a forecast for up to several weeks in
the future.
So it looks like at 15:02 you "predict" and event that occured about 1
minute earlier. For the Middle East, it appears that the prediction
from 12:35 is "predicting" an event that occurred 7 minutes earlier.
It's pretty easy to get hits if you wait until after the event in order
to predict it.
No. I already explained this point above. If you get a forecast of 10
to 20 minutes, you might find a real earthquake later on, matching
the list.
No, the problem is not finding an earthquake later on, it is that you
seem to be claiming hits from earthquakes that occure BEFORE the
forecast was issued. You didn't actually address the apparent problem
with the example you cited where it certainly appears that the
prediction time is later than the event time. That is a post-diction.
You also can't claim "later" events, if they are outside the forecast
window. That is the whole point of having a forecast window, so that it
is clear to everyone what is being predicted.
--
Thomas A. Russ, USC/Information Sciences Institute |
|
|
| Back to top |
|
| Physikus |
| Posted: Fri Apr 25, 2008 3:17 pm |
|
|
|
Guest
|
On 21 Apr., 19:44, t...@sevak.isi.edu (Thomas A. Russ) wrote:
Quote: ...
No, the problem is not finding an earthquake later on, it is that you
seem to be claiming hits from earthquakes that occure BEFORE the
forecast was issued. ...
That is not true, because you can see the forecasts - _before_ an
event occurs (look at http://eqf.homedns.org and type in, for example,
Greece or Turkey and note the given forecasts. After a while, you
should find matching events at a eq-monitoring-site (like http://www.emsc-csem.org)!
All forecasts are whithin a radius of approx. 1.5° (the system can not
work more accurate, because a higher resolution would effect in too
much data).
This is a further example from today:
Forecast from 06:31 pm (=UTC+2h, German Time):
2008-4-25;18:57:03; 24.4; 36.8;Greece;; 1; 6; 2
The matching real event (Time: UTC, source: emsc-csem):
2008-04-25 16:53:40.0 36.69 N 25.16 E ML 3.0 DODECANESE ISLANDS,
GREECE
As you can see, the longitude has a difference of -0.75° W. This
occurs, because the event was predicted 3 minutes before the event
happened (1°=4 Min.).
The forecast for the fairly high earthquake-potential was made at
06:31 p.m. (German Time [MESZ]) for 06:53 p.m - what means: 22 minutes
before a real EQ happened!
So I must say again: The system predicts all higher potentials of
earthquakes, and each real happened earthquake matches a point within
the forecasts.
Physikus
(Sorry about my english - i am still learning the language  . |
|
|
| Back to top |
|
| Weatherlawyer |
| Posted: Fri Apr 25, 2008 6:10 pm |
|
|
|
Guest
|
On Apr 26, 2:17 am, Physikus <physik...@gmx.net> wrote:
Quote: On 21 Apr., 19:44, t...@sevak.isi.edu (Thomas A. Russ) wrote:
...
No, the problem is not finding an earthquake later on, it is that you
seem to be claiming hits from earthquakes that occure BEFORE the
forecast was issued. ...
That is not true, because you can see the forecasts - _before_ an
event occurs (look athttp://eqf.homedns.organd type in, for example,
Greece or Turkey and note the given forecasts. After a while, you
should find matching events at a eq-monitoring-site (likehttp://www.emsc-csem.org)!
All forecasts are whithin a radius of approx. 1.5° (the system can not
work more accurate, because a higher resolution would effect in too
much data).
This is a further example from today:
Forecast from 06:31 pm (=UTC+2h, German Time):
2008-4-25;18:57:03; 24.4; 36.8;Greece;; 1; 6; 2
The matching real event (Time: UTC, source: emsc-csem):
2008-04-25 16:53:40.0 36.69 N 25.16 E ML 3.0 DODECANESE ISLANDS,
GREECE
As you can see, the longitude has a difference of -0.75° W. This
occurs, because the event was predicted 3 minutes before the event
happened (1°=4 Min.).
The forecast for the fairly high earthquake-potential was made at
06:31 p.m. (German Time [MESZ]) for 06:53 p.m - what means: 22 minutes
before a real EQ happened!
So I must say again: The system predicts all higher potentials of
earthquakes, and each real happened earthquake matches a point within
the forecasts.
If you post the actual forecast, editing out the data which you say is
just too much, it will be date and time stamped on all of Usenet. And
available for all to see that's the point.
People don't like to follow links off page. They may fear bugbeds. Or
some other waste of time.
Not a lot will be done in 20 minutes but not a lot is going to be done
anyway. It will open up possibilities though. |
|
|
| Back to top |
|
| |
Page 2 of 2 Goto page Previous 1, 2
All times are GMT - 5 Hours
The time now is Sat Oct 11, 2008 6:38 pm
|
|