Hi Roger,

This is a relatively easy statistics type of project that might be of interest to you. If it does not sound interesting or you are pressed for time then don’t bother.

Now that the format of the Chart A earthquake forecasting picture file on the Data.html Web page is largely settled I am starting to look at certain assumptions behind the forecasting program such as the one being discussed here.

QUESTION

What is the percentage of deadly earthquakes that have occurred since the beginning of 1990 that had another deadly earthquake occur previously within plus or minus one longitude degree?

Earthquake latitudes should not be considered, just their longitudes.

Say that a deadly earthquake occurred on January 1, 1990 at 100 E longitude. It would be one for the “no previous deadly earthquake category.” Then another deadly one occurred at 101 E longitude on February 3, 1991. That would then be one for the “previous deadly earthquake category.”

Once that is done the next step would be to do the calculations for deadly earthquakes that had one occur in the past within plus or minus 2 longitude degrees, then 3, 4, and 5, again going back to the beginning of 1990.

As you are aware, the USGS has a Web page that contains earthquake information plus the number of fatalities for the earthquake.

THE GOAL HERE

My forecasting program matches EM Signals with past earthquakes. It then draws a map that shows the locations of the 50 or so best matching earthquakes that were not deadly plus the 50 or so best that did cause fatalities.

The assumption that I have been working with is that if a destructive earthquake is going to occur somewhere then the probability is fairly high that it will have been preceded by a deadly one that occurred around that longitude since the beginning of 1990. So having the program plot those 50 deadly ones in addition to the ones that were not deadly is valid even though many of the others that were not deadly would have better matches with the EM Signals.

However, if the study showed for example, that only 5% of our deadly earthquakes followed another one that occurred in the past around the same longitude then that basic assumption would have to be reassessed.

While the program is running, the number of earthquakes that are plotted, deadly or not deadly can be changed and a new plot generated. That takes about 20 seconds.