Astronomer Predicts Major Earthquake for Japan

[zoobyshoe]

Originally posted by Ivan Seeking
...actually as soon as I started getting the Zoobitude

"Disposition:Generally retiring,though occasionally
aggressive. Credited with
tearing apart bothersome
dogs."

 

[zoobyshoe]

Originally posted by Tsunami
Any where you want! Just make it out TO ME!

Here I answer your off the wall
quetion with what turns out to
be pretty much what a trained
geologist would say and YOU want
to get paid?? YOU want to get
paid?? YOU owe ME!

Minimun geological consultation
fee is 10,000.00.

 

[Ivan Seeking]

Going back however:

Forecasting quakes is generally considered to be impossible with current technology, and Kushida's method of using anomalies in the VHF range of radio waves to predict the timing and intensity of tremors has not gained many believers in the scientific community.

So his predictive process could still be within 25% of the real value. I assume by intensity they mean Richter value and not energy. To me, this would be an unexpected relationship at best.

 

[zoobyshoe]

Nothing happens in a vaccum, and
the search for an earthquake
"prodrome" (term borrowed from
migraine terminology) is, no
doubt, a logical and scientific
undertaking.

I don't know enough about any of
the specific sciences he is call-
ing to his aid in making his predictions, to form an opinion
as to whether he's way off track
or not, but as far as I can see
from my two inch elevation, there
is nothing absurd looking in his
method.

It could well be the coming typh-
oon threw off his measurements
somehow, such that it inflated
his prediction of the magnitude.
Or it could be that he needs to
make adjustments on how he calculates magnitude.

I think he was close enough that
he shouldn't be discouraged. There
is the ominous possibility this
was just a forehock. A couple more
days to tell...

 

[Ivan Seeking]

If possible, I will start visiting his website. Eventually it is just a matter of hits and misses.

 

[Ivan Seeking]

Mice can 'foretell earthquakes'[?]: BBC

Possibly a related story:

https://www.physicsforums.com/showthread.php?s=&threadid=6112

 

[Ivan Seeking]

For the record then:
It seems that in terms of energy, the error was about 17,000%

 

[zoobyshoe]

What's 17,000% among friends?

 

[Ivan Seeking]

Actually the prediction was for 17000% more energy than we got. But technically:

The energy was about 99.5% less than expected.

The Richter value was in error by 21%

I would think that the odds of a 5.5 quake on any day is no worse than 1:3500. So, if we were off by three days, we might allow for odds like 1:1000 of getting this close by chance. Of course this is just for perspective and not meant as hard numbers.

I didn’t find a good number for the frequency of Japanese quakes, but I did find a couple of interesting, related sites.

http://neic.usgs.gov/neis/current/japan.html

http://www.owlnet.rice.edu/~geol108/eq4/site_directory.htm#Earthquakes

 

[zoobyshoe]

They just had some more, Ivan. Wrong location but definitely the right magnitude:

Big earthquakes shake Japan island
Address:http://www.msnbc.com/news/971921.asp?pne=msntv

-zoob

 

[Zantra]

Hmmm...

Ok I'm not laughing anymore.. anyone else?

 

[Ivan Seeking]

Originally posted by Zantra
Hmmm...

Ok I'm not laughing anymore.. anyone else?

I have no idea how to gauge the odds as a function of distance. Still, it does seem a bit striking. I would imagine that hitting a 7.0 within 1 week and in the vicinity [geologically] could be significant. I think this guy deserves a good look.

 

[zoobyshoe]

Well, I think it's still wide open for debate since he got the location of the first one right and the timing right, but was way off on the magnitude.

These recent ones on the Island of Hokkaido are, really quite far away from the location of his prediction, but a magnitude 8 is huge, and fullfills what the prognosticator said in predicting greater than magnitude 7.

 

[Zantra]

Pretty accurate as far as earthquake predictions go. Of course I'd like to see him do it again, but it's an interesting theory

 

[zoobyshoe]

This is the trouble, as I said in my original post in this thread, that since Japan is one of the most earthquake prone places on earth, predicting an earthquake of appreciable magnitude there anytime in the next ten days is about as risky as flipping a coin.

The original quake was striking because he got the time and location right, but he flopped on the magnitude. The magnitude was his prime concern: he wanted to warn people to save lives.

Here, a week later the magnitude he predicted (very unusual) happens, but he's about a thousand miles off on the location.

You can't call his prediction a hit but it's too close to be a miss either. As Ivan said, he bears watching.

Now another quake predictor I read about has a theory based on a simple pattern of quakes in the pacific rim: major quake on west coast of South America is followed by major quake in Japan is followed by major quake on west coast of North America. In his system quakes around richter 6 count, as they follow the pattern.
I haven't been paying attention and don't know if there were any 6 or better quakes in South America recently, but maybe the three of us should buckle our seat belts just in case.

 

[Ivan Seeking]

Originally posted by zoobyshoe
This is the trouble, as I said in my original post in this thread, that since Japan is one of the most earthquake prone places on earth, predicting an earthquake of appreciable magnitude there anytime in the next ten days is about as risky as flipping a coin.

The original quake was striking because he got the time and location right, but he flopped on the magnitude. The magnitude was his prime concern: he wanted to warn people to save lives.

Here, a week later the magnitude he predicted (very unusual) happens, but he's about a thousand miles off on the location.

You can't call his prediction a hit but it's too close to be a miss either. As Ivan said, he bears watching.

Now another quake predictor I read about has a theory based on a simple pattern of quakes in the pacific rim: major quake on west coast of South America is followed by major quake in Japan is followed by major quake on west coast of North America. In his system quakes around richter 6 count, as they follow the pattern.
I haven't been paying attention and don't know if there were any 6 or better quakes in South America recently, but maybe the three of us should buckle our seat belts just in case.

Really, predicting a quake around 7.0 happening just somewhere in the world and getting within one week probably yields a random chance of success in the range of 1:2 - 1:4. 7.0s are much more common than 8.0s.

To predict a quake > 7.0 for anywhere in Japan within a week appears to yield a random chance of 1:50 to 1:200 for success. I am still trying to find a reliable number.


Number of Earthquakes per Year, Magnitude 7.0 or Greater
1900 - 1999 [entire world]



1900 13 1930 13 1960 22 1990 12
1901 14 1931 26 1961 18 1991 11
1902 8 1932 13 1962 15 1992N 23
1903 10 1933 14 1963 20 1993M 16
1904 16 1934 22 1964 15 1994 15
1905 26 1935 24 1965 22 1995E 25
1906 32 1936 21 1966 19 1996 22
1907 27 1937 22 1967 16 1997 20
1908 18 1938 26 1968 30 1998 16
1909 32 1939 21 1969 27 1999 23
1910 36 1940 23 1970 29
1911 24 1941 24 1971 23
1912 22 1942 27 1972 20
1913 23 1943* 41 1973 16
1914 22 1944 31 1974 21
1915 18 1945 27 1975 21
1916 25 1946 35 1976$ 25
1917 21 1947 26 1977 16
1918 21 1948 28 1978 18
1919 14 1949 36 1979 15
1920 8 1950 39 1980 18
1921 11 1951 21 1981 14
1922 14 1952 17 1982 10
1923 23 1953 22 1983 15
1924 18 1954 17 1984 8
1925 17 1955 19 1985 15
1926 19 1956 15 1986# 6
1927 20 1957 34 1987 11
1928 22 1958 10 1988 8
1929 19 1959 15 1989 7

Total 1900-1997 = 1960 events = 20 per year
http://neic.usgs.gov/neis/eqlists/7up.html

 

[Ivan Seeking]

Here is some good info.

Magnitudes of Significant Earthquakes

The USGS Earthquake Magnitude Working Group compiled a list of significant, past earthquakes for which reliable magnitude estimates are available. This list is not static and will be reviewed annually by the National Earthquake Information Center and revised as new research and results become available. (May take some time to load.

http://earthquake.usgs.gov/docs/sign_eqs.htm

I count 9 Japanese quakes greater than 7.0 since 1891. This yields odds of random success - of predicting any quake >=7.0, anywhere in Japan within one week - of around 1:600.

My only caution regarding this number is this sentence in the USGS link:

past earthquakes for which reliable magnitude estimates are available

I don't know how many Japanese quakes this may exclude.

 

[zoobyshoe]

If you go to the little map in the story I linked to you can see that Hokkaido is a separate Island from Japan, and it is very far away from Tokyo, where he predicted the >7 quake. The epicenter is about as far from Tokyo as Vladivostok, Russia, and also as far away from Tokyo as parts of Korea. It is a totally different fault.
Think of it this way: if someone predicted a >7 quake for San Fransisco within a week, but it only got hit with a 5.5, and then a week later Portland got three, one of which was an 8, could we properly say the prediction was a hit? In my mind he was somewhere in a grey area between hit and miss which means, hopefully, that there is something to his methods that could be sharpened up to give more accurate results. He was quite certain Tokyo was in for it.
Knowing for certain that a specific city will be hit by a specific magnitude is the whole goal of trying to predict quakes.

 

[Ivan Seeking]

YIKES!

January 23, 1556

Shensi, China
830,000 fatalities
~8 Mag

From the link above.

 

[hypnagogue]

Originally posted by zoobyshoe
If you go to the little map in the story I linked to you can see that Hokkaido is a separate Island from Japan, and it is very far away from Tokyo, where he predicted the >7 quake. The epicenter is about as far from Tokyo as Vladivostok, Russia, and also as far away from Tokyo as parts of Korea. It is a totally different fault.
Think of it this way: if someone predicted a >7 quake for San Fransisco within a week, but it only got hit with a 5.5, and then a week later Portland got three, one of which was an 8, could we properly say the prediction was a hit? In my mind he was somewhere in a grey area between hit and miss which means, hopefully, that there is something to his methods that could be sharpened up to give more accurate results. He was quite certain Tokyo was in for it.
Knowing for certain that a specific city will be hit by a specific magnitude is the whole goal of trying to predict quakes.

Hm... He predicted >= 7 for Tokyo, which only turned out to be a 5.5, but was followed up by an 8 in Hokkaido. Could it be that his model predicted the amount of energy churning about in the Earth's crust around Tokyo more accurately than the data bears out, but for some reason not all of that energy was dissipated in the Tokyo earthquake, and went on to contribute to the large magnitude of the event at Hokkaido?

 

[Ivan Seeking]

Originally posted by zoobyshoe
If you go to the little map in the story I linked to you can see that Hokkaido is a separate Island from Japan, and it is very far away from Tokyo, where he predicted the >7 quake. The epicenter is about as far from Tokyo as Vladivostok, Russia, and also as far away from Tokyo as parts of Korea. It is a totally different fault.
Think of it this way: if someone predicted a >7 quake for San Fransisco within a week, but it only got hit with a 5.5, and then a week later Portland got three, one of which was an 8, could we properly say the prediction was a hit? In my mind he was somewhere in a grey area between hit and miss which means, hopefully, that there is something to his methods that could be sharpened up to give more accurate results. He was quite certain Tokyo was in for it.
Knowing for certain that a specific city will be hit by a specific magnitude is the whole goal of trying to predict quakes.

It was about 600 miles from Tokyo. Vladivostok, Russia, is about 900 miles. However, we are using data from the entire Japanese area. Some of the quakes listed were centered 300 or 400 miles out at sea. One was near the location of the present quake. Also, the fact that Japan proper is so active, to me further supports the use of this data for comparison.

Finally, considering the magnitude, this may have been a subduction zone quake. In this situation, local faults may be of little significance. I don't know if a subduction zone quake is possible in this location. I will look.

Edit. It seems that one internet map used was not very accurate. The distances indicated are adjusted.

 

[zoobyshoe]

Originally posted by hypnagogue Hm... He predicted >= 7 for Tokyo, which only turned out to be a 5.5, but was followed up by an 8 in Hokkaido. Could it be that his model predicted the amount of energy churning about in the Earth's crust around Tokyo more accurately than the data bears out, but for some reason not all of that energy was dissipated in the Tokyo earthquake, and went on to contribute to the large magnitude of the event at Hokkaido?

I don't see any reason why this couldn't have been the case. If it were shown to be the case then it just points out the need for a lot of refinements in his methods, because it doesn't do much good for him to be off by a thousand miles for the people who get hit. If he is unable to get the location right then the fact he can predict within a thousand mile radius is a mere curiosity.

 

[Ivan Seeking]

Originally posted by hypnagogue
Hm... He predicted >= 7 for Tokyo, which only turned out to be a 5.5, but was followed up by an 8 in Hokkaido. Could it be that his model predicted the amount of energy churning about in the Earth's crust around Tokyo more accurately than the data bears out, but for some reason not all of that energy was dissipated in the Tokyo earthquake, and went on to contribute to the large magnitude of the event at Hokkaido?

This is similar to what I'm thinking. Also, we have no idea what his method is. Geographical range, timing, and magnitude are all variables that surely would result in different degrees of precision and accuracy in any foreseeable predictive model. To me, the crude but simple chance of success indicated is striking. I am at least 100% more interested than before. This means my left eyebrow is raised at least 0.250 inches now.

 

[hypnagogue]

zooby, I agree that his method, if it's really valid, could use some refining (if applicable-- maybe his method is just an inherently crude measure). But assuming it wasn't a fluke, even a crude estimate of a high magnitude earthquake like this would be valuable, certainly better than nothing. So in the future, to be more accurate he might have to warn not just Tokyo but all of Japan-- there's obvious practical problems with this, eg Japan residents wouldn't know of a safe place within Japan to hide out for a while, and most areas would suffer from anxiety and the cost of preparations for nothing. But preparations made in those areas that actually are struck could still ultimately save lives-- there could still be actual practical benefits.

 

[Ivan Seeking]

Some related info showing the location of the Japanese subduction zones:
http://atlas.geo.cornell.edu/education/instructor/topography/japan.html

http://geology.about.com/library/bl/maps/bljapanseismap.htm

It seems that this could be a major subduction zone event; we will have to wait and see. If so, it would seem possible that certain predictive indicators might have to be viewed in a broader geographical context than was indicated.

 

[kleinma]

perhaps the first quake he predicted was accurate, and the following quake was a result of the initial one...

regardless, they should look into his methods and see about possibly improving on them, because his ways for detecting quakes obviously shows he is doing something right.

I think the fact that his predictions were not 100% accurate is minor compared to the fact that he obviously has something going on for himself

 

[Ivan Seeking]

Originally posted by hypnagogue

By the way hypnagogue, thanks much for your kind words and your vote.

 

[hypnagogue]

No problem my man, you earned it. A board like this absolutely needs a solid and composed voice for open-mindedness, otherwise it just stagnates into arguments of dogma.

 

[hypnagogue]

Plus, the articles you post are thought provoking and sometimes provide a member with a pretty new avatar.

 

[Ivan Seeking]

The man who predicted last week's quake has gained sudden credibility, writes Shane Green in Tokyo.

http://www.theage.com.au/articles/2003/09/26/1064083188094.html

Well, maybe...