093650.8* 43.348N 146.785E 33N 5.00.49 KURIL ISLANDS
094717.5* 43.992N 147.164E33N 4.40.67 KURIL ISLANDS
113131.2? 43.81 N 146.76 E33N 4.21.29 KURIL ISLANDS
114154.6? 20.93 S 174.61 W33N 4.40.35 TONGA ISLANDS
115937.2* 43.503N 147.714E33N 4.60.921 KURIL ISLANDS
120504.543.443N 148.026E33N 5.2 5.1 0.875 EAST OF KURIL ISLANDS
121158.143.771N 148.230E33N 5.1 5.4 0.937 EAST OF KURIL ISLANDS
[skipping 15 more in the Kuril Islands to save space.....]
As you can see, predicting an M5-6 quake after a quake like this one is not very difficult. Not to
be disrespectful, but I have noticed in my experience with many prediction attempts, that a lot
predictions appear immediately after a large event and tend to predict aftershocks. This results in
predictions of no statistical significance because of the very high probability of success by
random chance. (It also suffers from failure to predict all similar aftershocks).
>magnitude 5-6 on Richter scale. This morning at 07.40 LT 9-10-94 or at
>23.40 UST 8-10-94 a cloud was observed whose compass direction and length
>pinpoints the North Island of New Zealand with an estimated magnitude of
The New Zealand prediction seems to have failed. A single failure no more dismisses your
method than a single success would confirm it. But it casts serious doubt on the method in the
absence of any notes of success or sound physical basis for the prediction model.
Note that statistical evaluation of any prediction method would involve establishing the
probability of success by random chance and then consideration of 4 categories of result: positive
(a predicted event occurred), false positive (an event was predicted and did not occur), false
negative (no event was predicted an one did occur) and negative (no event occurred and none
was predicted). Your model has already been indicated to suffer severely from false negatives.
>listmembers who might be able to suggest how such phenomena in the sky can relate to remote
>quakes
I would argue there can be no such mechanism. The physics of cloud formation are entirely
locally determined. The atmospheric properties leading to such formation are so chaotic that
predictions of weather conditions even twelve hours in advance, even with the help of powerful
computers and dense observations, is frought with error, so certainly precise cloud geometry is
completely undetermined twelve hours in advance. As far as energy considerations, local
variations in circulation, humidity, etc., would easily dominate the most powerful energy sources
(earthquakes no exception) even at relatively short ranges, let alone 85 degrees.
>given me a better than 50 % success rate so far, I will be pleased to post these observations.
That would be a start. However, without independant confirmation of your statistics we cannot
necessarily accept them at face value. Your New Zealand prediction is the start of an
independently-verifiable track record of predictions.
Richard Stead
Arlington, VA
stead@seismo.css.gov
|
1 2 3 4 