Here's a question to tickle your brain cells - and somthing to be considered
in making the Far-field aftershock maps, too.

An interesting question is being raised by the distribution of
the great Kunlun Mt. far-field aftershocks (FFA). In general,
it is expected that the strongest triggering around great earthquakes
will occur at the p- and s-wave shadow zone boundaries. When the
greatest moment (energy) release is coincident with the hypocenter,
there is no problem identifying the source of the energy, but
when the epicenter (the start of the rupture) is substantially
different from the location of greatest energy release as determined
by the moment tensor solution, the question arises regarding which location
to use in determining probability of subsequent FFA.
Such a case has arisen in the Kunlun Mountains great quake. The
location of rupture initiation appears to be substantially different
from the location on the fault where the most energy was released.
The epicentral data from NEIS gives the hypocentral parameters as:

O: 14NOV2001 09:26:10 35.9N 90.5E MS=8.0 NEIS QINGHAI, CHINA.

Whereas the moment tensor solution and distribution of aftershocks
from the Harvard determination gives the maximum moment at:

Source:
http://www.seismology.harvard.edu/CMTsearch.html

O: 14NOV2001 09:26:10 35.5N 92.8E MS=8.0 NEIS QINGHAI, CHINA.

Since there is more than 200 km separating these two locations,
the difference becomes important. In my opinion, the second (the
CMT solution) is the better to use in far-field studies since
it is really the distribution of energy from the source of energy
which is important here. The energy actually is smeared out over
the entire length of the rupture - in this case from 90E to 95E.
The difference became clear when viewing the largest or most
unusual earthquakes since the Kunlun event, many of which have
occurred near the shadow zone boundaries. Where EPID is the
distance from the NEIS epicenter and CMTD is the distance from
the CMT moment tensor solution, these include:
EPID CMTD

Central Mid-Atlantic Ridge (Nov 15; Mw 6.3) 104.1 105.9
Sevier Utah (Nov 19; Ml 3.8) 103.7 102.7
Macquarie Islands (Nov. 18, Mb 5.6) 108.4 107.5
Nelson N.Z. (Nov 17; Ml 4.8) 107.4 106.3
Ross N.Z. (Nov, 18, Ml 4.4) 107.0 105.9

This becomes important in making assessments of hazard related
to FFA. For example, using the NEIS epicenter, most of the New
Zealand earthquake zone falls just within the shadow zone, and
outside the region where strong FFA is expected, however, using
the CMT solution, the regions of New Zealand within the region
where FFA are expected. A warning would be made for New Zealand
using the CMT solution, but not using the NEIS solution. Since
the purpose of using FFA triggering to give a general warning
to areas where strong subsequent seismicity is expected, the difference
in using the CMT vs the PDE solutions is extremely important.

Any ideas from the board on approaches to this problem?