Jökull - 01.12.1964, Blaðsíða 6

profile Grímsfjall towards Kverkfjöll, and a
short profile transverse to it, are shown in Figs.
3 and 4.
For computation of the thickness of the ice
cap it is necessary to know that part of the
Bouguer anomaly, which is due to the mass
distribution below sea level. Values as low as
—35 mgal were found by Einarsson in the
centre of the gravity bowl of Iceland, at the
northwestern edge of Vatnajökull. The exact
form of the central part of this bowl is not
known, as only a limited number of measure-
ments is available from the central part of Ice-
land. At Jökulheimar the Bouguer anomaly is
about —7 mgal according to the present mea-
surements. There are thus considerable varia-
tions in the regional Bouguer anomalies at the
edge of Vatnajökull and they cannot be extra-
polated to the Grímsvötn area.
Several seismic ice-thickness determinations
are available on Vatnajökull from the work of
Joset and Holtzscherer (1954). One of their
stations (No. 28), which was considered to give
reliable results, is located close to the profile
line Grímsfjall—Kverkfjöll. Apparently the best
procedure available to estimate the ice-thickness
along the profile, is to use the known ice-
thickness at this station to calculate the re-
gional Bouguer anomaly due to the underlying
rock formations, and then assume this anomaly
to be constant along the profile line. At this
station the thickness of the ice was found to
be 760 meters, and the Bouguer anomaly —90
mgal. Assuming an infinite ice-sheet of uni-
form thickness, the contribution of this sheet
to the Bouguer anomaly would be —54 mgal.
The remaining —36 mgal would then be due
to deeper crustal effects. The ice-thickness on
the remaining points of the Grímsvötn—Kverk-
fjöll profile and the one transverse to it were
then computed from the observed Bouguer ano-
malies by subtracting an assumed regional ano-
maly of —36 mgal and using a conversion factor
of 14.0 m/mgal, which is valid for an infinite
sheet of density difference 1.7 g/cm3.
Several sources of error are possible in the
JÖKULL 1964
64