Jökull - 01.01.2010, Side 77

Intraplate earthquake swarms in Central Europe
The source mechanisms of the 2000 swarm differ
substantially from those of the 1997 swarm. Simi-
larly as in the swarm of 1997, both oblique-normal
and oblique-thrust events occurred in 2000, however,
the oblique-normal faulting prevailed; besides, the
oblique-thrust events were limited to the first swarm
phases. The predominant strikes and dips of the 2000-
oblique-normal events matched the geometry of the
main fault plane Nový Kostel (NK) as well as the
strike, dip and rake of the largest ML=4.6 earth-
quake of the swarm of 1985/86 (Antonini, 1988).
This, together with the localization results (Fischer
and Horálek, 2003), suggests that both location and
faulting in 1985/86 and 2000 swarms were similar.
The resultant mechanisms of all events are dominantly
DCs with only insignificant non-DC components usu-
ally not exceeding 10% as a whole. The massive dom-
inance of the DC in all investigated events implies that
the 2000-swarm consisted of a great number of pure
shears along the fault plane.
The resultant source mechanisms of 1997 and
2000 swarms are represented in two ways (Figure 8).
The orientation of the fault planes (DC parts) is de-
picted by the composite focal mechanism plots and
the non-DC parts using the Hudson at al. (1989)
source-type diagrams. Here, the moment tensor is
parameterized by the quantities k and T , both vary-
ing between –1 and 1. In general, a source with
k=1 corresponds to a purely explosive event, while
a source with k=–1 corresponds to a pure implosion.
The parameter T corresponds to the content of CLVD
part. The pure DC sources are uniquely described by
k=T=0 (for more details refer to Hudson at al., 1989).
The source type plots of 1997 and 2000 swarms
significantly differ: while the 1997 sources align
along dipole-type axis, the 2000 sources form a tight
cluster near the DC source type. The common dipole
character of the 1997 sources shows that all events
were of the same type; however, the polarity of the
dipole varied with time. The slightly compressive
dipoles dominating in the first part of the swarm (blue
symbols) were replaced by tensile dipoles in the sec-
ond part of the swarm (red and yellow symbols). The
larger non-DC part of the late events corresponds
to the B-type events of tensile character, while the
smaller non-DC part of early events corresponds to
the A-type events of nearly DC character. However,
the systematic non-DC part of most events (extension
along common line in the source-type plot) suggests
that the non-DC components may not be random.
The plausibility of the non-DC character of the
1997 sources is supported by number of tests per-
formed by Horálek et al. (2002) to rule out the possi-
ble effect of insufficient focal sphere coverage, inac-
curate or noisy input data, inexact localization of the
hypocentre and poor velocity model on the origina-
tion of spurious non-DC components. Another sup-
porting argument is the fact that the A and B events
were located within a compact cluster and the whole
swarm was recorded by the same configuration of
the stations. Thus the waves of the A and B events
propagated along the similar travel paths. Further-
more, we checked the reliability of MTs of the 2000
swarm by means of the jack-knife trials in which the
number of stations used in the MT inversions varied
from a complete set down to 7 and 6 (Horálek et al.,
2010). The reduced data set differed in only two sta-
tions (which did not operate in 2000) from that one
used to retrieve the MT of the 1997 swarm. The tests
proved stability and good constraints of these MT so-
lutions. Besides, they evidenced high susceptibility
to focal sphere coverage by stations: sparse station
distribution resulted in the appearance of false non-
DC components, mainly CLVD. However, for all the
analyzed events of both 1997 and 2000 swarms, the
focal-sphere coverage was sufficient thanks to suit-
ably distributed stations NKC, KRC, KOC and LAC
which were used in all MT inversions performed. The
test also showed that the non-DC parts <15 % can not
be consider as significant even in case of a sufficient
focal-sphere coverage (for more information of the
appearance of false non-DC components due to noise,
mislocation of the hypocentre and velocity structure
mismodelling, see Horálek et al., 2010).
It is evident that the type of faulting depends on
the orientation of the fault plane with respect to the lo-
cal tectonic stress field. Providing that the stress esti-
mations by Havír̆ (2000), Vavryc̆uk (2002) and Plene-
fisch and Klinge (2003) (which are fairly similar) are
consistent with a steady local stress in the NK zone,
JÖKULL No. 60 77