Jökull - 01.06.2000, Page 7

Ground deformation at Katla
The station Kötlukriki is definitely of lower
quality than the station Jökulkvísl as the computed
standard error of tilt is higher and increases with time
between the measurements used to compute the tilt.
Tilt computed from observations made one year apart
usually has standard error of about 0.5 /rrad while
tilt computed from measurements made more than 10
years apart has standard error of slightly more than 1.0
/trad. This suggests that the long time stability of indi-
vidual markers in this tilt station is not very good, or
that gradual irregular ground deformation takes place
within the area of the tilt station. This tilt station
lies on lava which has flowed from the Mýrdalsjök-
ull glacier. Beneth the lava is probably a thick layer of
sand, similar to that found outside the lava. This sand
layer may be subject to continuing compaction.
The station Höfðabrekkuheiði is of far inferior
quality relative to the two tilt stations east of Mýr-
dalsjökull. Two or more markers of the tilt station
were unstable during the first years of observations,
probably because frost action moves sizeable slabs
of the tuff in which the markers are cemented. Later,
one of the markers became loose because of erosion.
These problems at the station Höfðabrekkuheiði make
it of limited value in interpreting the 1967 to 1995 tilt
observations near Mýrdalsjökull.
INTERPRETATION
The computed tilt exceeds in a number of cases the
computed standard error by a factor of more than
three, suggesting more than 99 per cent probability
that the observed tilt is real. If tilt between two
successive observations is considered, this occurred
in 5 out of 12 cases at the station Kötlukriki and
in 7 out of 13 cases at the station Jökulkvísl. There
is quite good correlation between tilt at these two
stations. If tilt at one station is up in an easterly or
westerly direction, the other station is also tilted in
the same general direction (Figure 3). This suggests
that the observed tilt at these two stations is caused by
a common source.
The observed tilt varies significantly from one per-
iod to another, although tilt azimuth either toward east
or west appear most common. This points towards the
glacier Mýrdalsjökull to the west of these stations as
a possible source, either as a source of uplift which
would cause tilt up in westerly direction, or, at other
times, a source of subsidence, which would cause tilt
down towards a westerly direction.
Prior to an eruption of Katla, an increase of
pressure is expected below the volcano as magma
moves toward a magma reservoir which supplies
material for the eruptions. This process will cause
more or less continuous uplift of the volcano and tilt
up towards a westerly direction at the tilt stations east
of Mýrdalsjökull. This has not been observed.
The observations indicate periods of uplift and
periods of subsidence of the crust beneth the Mýrdals-
jökull glacier. This alternating uplift and subsidence
has been interpreted as caused by variable ice mass of
the glacier (Tryggvason, 1973). It is well known that
a great amount of snow accumulates on the glacier
each winter and a similar amount melts each sum-
mer. Tilt observations were made in the summer of
1967 to 1973, sometimes as early as mid June, someti-
mes as late as mid August. Estimates of glacier load
(Figure 4) at times of tilt observations correlate fairly
well with the observed east component of ground tilt
at the two tilt stations east of the glacier. After mea-
surements were resumed in 1986, most observations
have been made in September or October.
The two tilt stations east of Mýrdalsjökull were
observed in May 1994 in addition to observations in
October 1993 and October 1994. Both stations show
tilt down toward a westerly direction between October
1993 and May 1994, a period of increasing glacier
load, and both stations show tilt up towards west
from May to October 1994 (Figure 3), a period of
decreasing glacier load.
A model study of the Mýrdalsjökull ice field
(Tryggvason 1973) suggested that annual variation in
the glacier load is equivalent to about 2 m of water
over the entire area of the ice field, about 700 km2.
Observed tilt near the edge of the ice field indicate
that the amplitude of the annual tilt oscillation is about
6 //rad. A model of elastic lithosphere of uniform
thickness floating on viscous íluid, using the approx-
imation of Hertz (1884) suggest that the amplitude of
the vertical ground oscillation at the center of the ice
JÖKULL No. 48 5