Geoscience Reference
In-Depth Information
magnitude-6 earthquakes has been subjected to intensive research. This resulted
in the prediction of a magnitude-6 event between 1988 and 1992. Even though no
earthquake occurred, one is still expected. A very successful prediction was made
and evacuation warning given for the 1975 magnitude-7.3 Haicheng earthquake
in China, but this success was not repeatable: the 1976 Tangshan earthquake was
devastating. Despite concerted research effort in many fields over many years,
the goal of earthquake prediction still remains distant.
4.2.4 Nuclear explosions
Nuclear explosions present particular problems: seismologists are required to
monitor compliance with the ban on nuclear testing. To be certain that an event
is identified correctly as an earthquake or an explosion, it is necessary to con-
sider parameters such as location, depth and shape of the waveforms. Nuclear
explosions have all been shallower than 2 km. Thus an event with a focal depth
of 200 km is exceedingly unlikely to be a nuclear explosion, even assuming the
maximum errors in depth calculation, whereas an event with a very shallow focal
depth of 15 km would be a candidate for an explosion. The first P-wave from an
explosion to arrive at any seismic station, irrespective of azimuth, should be com-
pressional, whereas the first arrival from an earthquake will be either dilatational
or compressional, depending upon the azimuth (Section 4.2.8).
As mentioned in Section 4.1.3, nuclear explosions are not as good as earth-
quakes at generating surface waves or S-waves. An earthquake will generate
more long-period energy than will an explosion because it ruptures over a fault-
plane that is large in comparison with the cavity produced by an explosion. In
general, earthquakes have
m
b
values 1.0-1.5 units smaller than those for nuclear
explosions with the same
M
S
values. The straight line in Fig. 4.11(a),
M
S
=
1
.
68
m
b
−
4
.
82
(4.22)
or
m
b
=
2
.
87
+
0
.
60
M
S
(4.23)
separates most of the explosions and earthquakes.
There is, however, a difficulty with small events. Calculations of
m
b
and
M
S
are
generally performed with data recorded at distances greater than 2000 km from
the epicentre but surface waves from small events (
m
b
<
4.5) cannot be reliably
measured at such distances. A solution is to use
M
L
and
M
0
rather than
m
b
and
M
S
.
M
L
, the local magnitude proposed by Richter, is based on measurements made
close to the event. In Fig. 4.11(b) the earthquakes and explosions are separated
over many orders of magnitude. The problems with this approach, however, are
that an
M
L
scale is, as its name indicates, local and so has to be determined for
each region; and local seismograph networks are required.
