Greinar (Vísindafélag Íslendinga) - 01.01.1976, Page 50

48
earthquakes evidently occur on more than one fault. This view was
expressed by Tryggvason (9).
The epicenters shown in Figure 1 are not located accurately
enough to reveal any detailed tectonic pattem such as faults. In
this paper an attempt is made to remove some of the major sources
of errors in epicentral locations by locating the epicenters relative
to one another. Epicenters of events that occurred in 1968 and 1969
were relocated in this way. These epicenters appear to define a
WNW striking fault near the island Grimsey. The focal mechan-
ism solution for one earthquake indicates strike-slip motion in a
right-lateral sense. This fault parallels the Húsavík faults and is
situated approximately 40 km to the north of them.
Some evidence is found for a parallel strike-slip fault to the south
of the Húsavík faults. The transform motion between the axial rift
zone of Iceland and the Kolbeinsey Ridge thus appears to be taken
up by possibly three or more parallel strike-slip faults.
RELATIVE LOCATION OF EPICENTERS
Errors in the location of an earthquake using teleseismic data
can arise from a number of sources:
1. Errors in the assumed earth structure under the earthquake
can cause an error in the location, especially if the station distribu-
tion is non-uniform.
2. Unknown station residuals and lateral inhomogeneities along
the ray paths can cause errors in the locations that appear random
if different station sets are used to locate different seismic events.
3. Mis-picking of the first P-wave arrival is an error source. The
first arrival may be picked on a large event but if this arrival is
relatively small, it may be missed on a small event.
Several steps can be taken to avoid these errors. By locating all
earthquakes with respect to one master event, it is possible to elimi-
nate errors caused by uncertainty in the earth structure near the
source as well as errors caused by uncertainty in the origin time.
Travel time differences between stations are observed and compared
with the travel time differences for the master event. By observing
only travel time differences the origin time is eliminated as a vari-