As well as post a few facts to back up our statements.

Firstly, I would agree that a prediction covering hundreds of miles is relatively useless in the sense that it wouldn't provide disaster authorities with a specific location to hone in on. But on the other side of the coin, all of the disaster response teams in and around an area of hundreds of miles would be ready to respond and provide immediate assistance to the specific area that did take the hit. Predicting over a large area would also preclude any one area from getting into a potential panic situation.

In addition you say, "In addition i cannnot accept a spatial window of "hundreds" of kilometers with a given magnitude below 5 as SCIENTIFICALLY acceptable,, because that is like predicting that the sun will rise
tommorow !". I would take that to mean a 100% probability of occurring anyway. Well, lets get to the facts.

In my past prediction I had predicted an M3.5 to an M5.0 with a 2 day window covering a circular area of 300 miles in any one direction. To the east this would take me to about 50 miles short of Prescott, Arizona. To the SE about 25 miles short of San Felipe down in Baja California. To the North, to about the south shore of Walker Lake and to the NW about Fremont, California.

In Bolt's "Earthquakes", he has a probability chart on page 207 with a coverage area of northern and central California or a 280,000 Square kilometer area.

In addition, a FAQ had been posted on sci.geo.earthquakes, that listed 3 different probability tables covering different areas of California. It was compiled by a Dr. Robert Uhrhammer at UC Berkeley. You can look for this FAQ on deja news by looking for the following subject in newsgroup sci.geo.earthquakes:

[lm 7/27/95] Frequently asked earthquake references, part III

Deja News is at: www.dejanews.com

From the statistics presented, an M3.5 has the following percent probility in:

Bolt's "Earthquakes
1 day 6.3 1 week 36.0
FAQ
CENTRAL COAST RANGES, CALIFORNIA
1 day 3.0807 1 week 19.7308
NORTHERN AND CENTRAL CALIFORNIA
1 day 7.2716 1 week 41.1554
CALIFORNIA
1 day 21.4772 1 week 81.6993

There is a difference in the probabilities shown for northern/central California between Bolts chart and the FAQ. Reason is that Bolt chart was based on the seisimicity recorded from 1949 - 1983 and the FAQ from 1949 - 1988 that gave different formulaes in the calculation.

Bolts log N = 4.23 - 0.815 ML
FAQ log N = 4.44 - 0.857*M

My coverage area is about a third larger then all of California (from my calculations, if I did that right, is about 282,744 square miles where the total square miles for California is 158,706). But then again, about 40% of the coverage area is out in the ocean where quakes aren't known to occur in addition to other land based areas. So in that sense, I'm needlessly adding to my coverage area. But to keep it simple, I had chosen to just go with the radius.

So based on the California statistics above (1 day have a 21% probability), you can see that predicting an M3.5 in a 300 mile radius over a 2 day window is somewhere less then a 50% probability and that predicting anything less then an M5.0 over hundreds of miles is nothing near predicting the sun will rise tomorrow.

I thought I would also say that people are expecting too much in terms of earthquake prediction and science in general. Nothing that I know of was perfect the first time it was introduced without needing any additional refinements.

Dennis