Jökull - 01.12.1992, Qupperneq 18

Table 1. Crustal models at 4 OBS stations on the Axarfjarðardjúp profile. The depth is measured from the
seafloor. — Jarðlagasnið (P-bylgjuhraði sem fall af dýpi) í báðar áttir frá neðansjávarskjálftamælunum á
bylgjubrotssniðinu.
OBS 18towardsSE OBS 17 towardsNW OBS 17 towardsSE OBS 13 towardsNW OBS 13 towardsSE OBS 12 towards NW
P-wave velocity km/s depth km P-wave velocity km/s depth km P-wave velocity km/s depth km P-wave velocity km/s depth km P-wave velocity km/s depth km P-wave velocity km/s depth km
2.30 0.00 2.30 0.00 2.30 0.00 2.30 0.00 2.30 0.00 2.30 0.00
2.31 0.50 2.31 0.90 3.00 1.00 2.50 0.05 2.50 0.50 3.20 0.05
2.50 0.60 3.30 0.95 4.30 1.00 2.80 0.70 3.30 1.10 3.95 1.35
3.00 0.90 3.95 2.10 5.90 4.00 3.10 0.71 3.50 1.50 2.60 1.40
3.80 2.30 4.30 2.11 6.80 5.00 3.30 1.00 3.60 1.80 2.80 2.03
4.30 2.31 5.50 4.00 7.00 11.00 2.80 1.01 3.80 2.20 4.55 2.40
4.75 3.00 6.00 5.00 3.20 2.10 4.30 3.10 6.80 6.50
5.90 4.00 6.80 6.00 4.40 2.20 5.50 3.30 7.00 11.00
6.80 5.70 7.00 11.00 5.50 4.30 6.80 5.40
7.00 11.00 6.80 5.00 7.00 11.00
7.00 11.00
depth of 100 m (E36 on Figure 3) and 40 m (E37 and
E38 on Figure 3) on a line perpendicularto the profile.
The shots outside the profile were too widely spaced
to get a good resolution of the crustal structure.
The stations and shot sites were located with
a GPS-pathfinder system with a radial accuracy of
±30 m. All recording instruments were time synchro-
nized with the Rugby time signal at the beginning and
end of each deployment.
DATA ANALYSIS
The OBS records are analogue three-component
seismic recordings (one vertical and two horizontal)
and one time-signal channel. The vertical component
was recorded at three amplification levels. The ana-
logue tapes were digitized with a sample rate of 58 Hz
and stored on a tar-formatted, UNIX tape. The data
from each channel occupies two bytes in the wave data
file which is divided into blocks of 8192 bytes.
The travel-time generating program TTGEN was
used in order to model the regional crustal structure
by the trial and error method. TTGEN is a subrou-
tine from the hypocenter location program HYPOIN-
VERSE(Klein 1978). It generates crustal models with
up to fifteen layers and calculates the travel-time for a
laterally layered model, with linear velocity gradients
within each layer. One embedded low velocity zone
is permitted. Homogeneous layers with zero velocity
gradients are permitted but velocity discontinuities are
not.
The OBS records contain substantial background
noise, in particular OBS station 12. Figure 3 shows
the reduced record section of the vertical component
of the four OBS stations on the refraction line along
with calculated travel-times for the íirst arrivals. No
secondary arrivals are seen. The signals are clipped
later in the coda which does not help identifying later
arrivals. We trace the first arrival out to a distance
of approximately 20 km, beyond which it fades into
the noise. Due to the wide span and the small size
of the explosions, the seismic data were interpreted
as independent unreversed profiles. An attempt was
made to use an inversion model for the data, but the
overlap of the profiles was too small for such a model.
The upper crust of Iceland is best modeled by an
algorithm that works with velocity gradients, rather
than many small step segments. The use of linear
gradients smooths out the discontinuities in the travel
time derivatives which result from homogeneous-layer
modeling. Linear gradients also give a more realis-
tic spread in the emergence angles of the downgoing
rays than occurs when modeling the rays as refracted
from discontinuities. However, both gentle gradients
and steep steps were used in modeling the seismic
traces along the profile. Table 1 lists the best-fitting
16 JÖKULL, No. 42, 1992