Jökull - 01.12.1989, Qupperneq 45

Fig. 3. The Bouguer anomaly map calculated with a
Bouguer density of 2500 kg/m3. Contour lines are 5
milligals apart.
Mynd 3. Bouguer þyngdarkort, reiknað með Bougu-
er eðlisþyngdinni 2500 kg/m3. Milli jafngildislína
eru 5 milligal.
The source of this anomaly could lie both in the
crust and the mantle. Volcanic products erupted dur-
ing the Pleistocene are likely to contain a high pro-
portion of hyaloclastites with lower density than the
surrounding basalts erupted in ice-free environment.
The surface formations then subside as they are
buried by younger volcanics and are carried side-
ways out of the volcanic zone by plate movements.
The crust formed in this way during the Pleistocene
is of lower average density than the crust formed by
dense lava flows or intrusions, which gives rise to
negative gravity anomalies. Another possible source
is the layer of partially molten material thought to
underlie most of Iceland (Gebrande and others,
1980; Beblo and Bjömsson, 1978). This layer is at
shallower depth beneath the volcanic zone than
under the older Tertiary and Quarternary areas. A
thinner crust and a higher degree of melt at the
TABLEI. ANOVA for polynomials fitted to the
gravity data.
TAFLAl. ANOVA reikningar fyrir bakgrunnssvið
þyngdarmœlinganna.
order of polynom. goodness of fit degrees of freedom F value F test, a = 0.95
1 82.2% 2, 288 667.23 3.00
2 89.5% 5,285 65.23 2.21
3 89.5% 9,281 0.08 1.88
Goodness offit is measured by the coefficient of determi-
nation and shows how much the variance in the data is
reduced by the polynomial. Degrees offreedom are shown
for the polynomial and the residual data. The F value
indicates the improvement in goodness of fit over the
increase in variables in the polynomial. We state, with
a=95% confidence, that the added parameters
significantly improve the fit if the F value exceeds the F
test statistic listed in the last column.
crust-mantle interface beneath the volcanic zone
may produce a gravity low (Hersir and others, 1984;
Björnsson, 1985). We note that a zone of no
reflections was identified by Zverev and others
(1980) on a seismic reflection profile that crosses the
volcanic zone in SW-Iceland. This zone, interpreted
to be a volume of high degree of partial melt, was
located at about 8 km depth directly beneath Skjald-
breiður.
At Langjökull in the north, the gravity low
appears to tum to the east, but no data are available
from the glacier. In the south, near Hengill, it turns
WSW but does not seem to continue out to the tip of
the peninsula as might be expected from the earth-
quake distribution (Klein and others, 1977;
Halldórsson and others, 1984).
There are three main local minima in the gravity
low. Counting from the south, the first one is in the
tuffaceous mountain complex on the peninsula
southwest of Hengill. This low may at least partly be
a spurious creation of the Bouguer correction. The
second occurs just east of the lake Þingvallavatn,
which is in a topographic low in the rifting zone.
The third is at Skjaldbreiður, discussed above.
Smaller minima appear at the westem and eastem
JÖKULL, No. 39, 1989 43