Evaluation of Don's Mt. St. Helen's forecast

On Sept. 20, Don made the following forecast (#9545 below)
for the Mt. St. Helen's region:

"Mount St. Helens, Washington Earthquake Forecast
Posted by Don In Hollister on September 20, 2001 at 00:48:38:

Magnitude: 2.0-3.6
From: 09/21/2001PDT - To: 09/26/2001PDT
Location: Mount St. Helens, Washington
Lat: 46.2N - Long: 122.2W - Range: 40Km"

Since the time window has now passed, this forecast can be evaluated.
PNSN identified the following events within the 40 km radius that
Don had declared in the past past week:

ML 2.1 2001/09/28 23:44:58 46.4N 122.2W 17 km S of Morton, WA
ML 1.5 2001/09/25 11:03:25 46.2N 122.2W 0 km NE of Mount St. Helens
ML 1.3 2001/09/24 17:52:56 46.2N 122.2W 0 km NE of Mount St. Helens
ML 1.8 2001/09/24 00:12:11 46.2N 122.2W 1 km NE of Mount St. Helens

Source:

http://www.geophys.washington.edu/recenteqs/Quakes/quakes0.html

A swarm of smaller events has also occurred in the Mount St. Helens
vicinity including nearly 100 events of Ml<1.0. The forecast, however
was clearly not for a small event, but for something in the Ml 2-3.6
range. While the event on 24 Sept. matched all parameters of
Don's prediction, it was not in the stated magnitude range. In
my opinion, in region of high microseismicity, it is important
for a forecast event to lie in the given magnitude range as well
as within an acceptable time of the forecast window. The first
3 events and the accompanying swarm did not fall in the magnitude
window, and while the swarm was uncommon for Mt. St. Helen's at this
time, I do not believe it can be considered a match for Don's
prediction.
The earthquake last night S. of Morton, however is another matter.
While 2 days outside the time window, it was in the magnitude and
location range (being about 22 km from the predicted epicenter).

In the case of such a small event, it is particularly necessary,
however to check the probability of random occurrence within
the time frame. In Don's area (40 km of 46.2N 122.6W) there have
been 10 earthquake with Ml between 2.0 and 3.6 in the past year -
about 1 every 36 days. The past several months, however, have
been quiet in the region with the ony event of Ml>=2 in the region
occurring on August 6 (Ml 2.4). Prior to that, there had been
none since May 8, 2001. Earthquakes larger than last night's
occurred on Aug 6 (Ml 2.4); Apr 24 (Ml 2.3) Mar 17 (Ml 2.5) and
Oct 10, 2000 (Ml 2.2). The event of March 17 was the largest
in the region in the past year (Ml 2.5).
From this background, the daily chance over the past year of the
occurrence of an earthquake of Ml 2.0-3.6 in the 40 km radius area was
10/365 = 0.027. In the stated 6-day period, the chance was 0.16
(about odds of 1 in 7), in the observed 8-day period from the
initial day of the forecast to the event, the odds of success
were about 1 in 5. By random chance Don should not have had
a success on this one in more than one in five tries. Not a bad
forecast given the Ml 1.8 which occurred in the middle of the
time window, and the Ml 2.1 which occurred 2 days outside it.
I'd give this a B+, mostly because the probability of random
success is fairly high compared with other predictions Don has
made.

• Earthquakes in the Pacific Northwest