According to the to the USGS/NEIC QUICK EPICENTER DETERMINATIONS (Golden, CO — NO. 1-014 JAN 14, 2001) the following earthquakes of magnitude 7.5 or more happened in July-December 2000.
Date and Time
Lat Lon Depth
Magnitudes (mb, MS, other)
2000/11/16 04:54
-3.958  152.268  33
5.8  8.1  8.00MWHRV  8.00MEGS
2000/11/16 07:42
-5.179  153.054  33
6.2  7.8  7.60MWHRV  7.30MEGS
2000/11/17 21:01
-5.453  151.685  33
6.2  8.0  7.60MWGS  6.90MEGS
2000/12/06 17:11
39.625  54.772  30
6.6  7.5  7.00MWGS  6.90MEGS
The last one in TURKMENISTAN (where the background seismic rate is insufficient to run the M8 algorithm in the original version) is outside our test area. The other three are a cluster of earthquakes from NEW IRELAND and NEW BRITAIN REGIONS, P.N.G. None of them are predicted. According to QED all the three are «complex events» of which the largest two are clear outliers of the activated aftershock zone. I was surprised to find that their station magnitudes scale with the epicentral distance: The larger is the distance to the station — the larger is its MS! Here are the two MS averages estimated over the stations from epicentral and antipodal hemispheres (data from the USGS/NEIC mchedrqed.dat file as on JAN 14, 2001 was used) for each of the earthquakes:
Date and Time
Epicentral hemisphere
Antipodal hemisphere
Average MS
Number
Average MS
Number
2000/11/16 04:54
7.6
16
8.1
37
2000/11/16 07:42
7.5
27
7.8
60
2000/11/17 21:01
7.4
17
8.0
53
2000/12/06 17:11
7.2
28
7.5
58
The enlarged Figure shows the M8-MSc alarms issued for the second half of 2000 for M8.0+ and M7.5+ along with the epicenters of Nov 16 and 17 earthquakes and their aftershocks. The M7.5+ prediction misses the earthquakes by 200 km (which is less than the distance between centers of neighbor circles of investigation).