Jökull - 01.12.1977, Qupperneq 23

glass, olivine tholeiite and andesite. One layer
is tholeiitic (1919), whereas no analysis exists o£
three (1661, 1739, 1769).
These tephras do not lend themselves to mucli
discussion. The fact that tliey were apparently
not accompanied by jökulhlaups tends to indi-
cate tltat they were producecl in subaereal erup-
tions.
It is tempting to allot the 1919 tephra to an
eruption in Dyngjufjöll (Askja), which was very
active about this tirne. Likewise, the 1661 layer
is tentatively allotted to the Katla eruption of
1660—61 — an analysis would have beautifully
clinched the matter. As to the others, too little
is known about the volcanic history and the
chemical cliaracteristics of the areas surrounding
Vatnajökull to allow but a suggestion as to their
origin. However, as stated earlier the Dyngju-
háls fissure swarm may well have been active
in the 18th century along with the Ivrafla
swarm east of Lake Mývatn, and the Kverkfjöll
swarm (Thorarinsson 1950; for niaps of these
swarnis see Sœmundsson 1974). In his 3-volume
book Ódáðahraun (1945) Ólafur Jónsson has
described tliis inaccessible area most thoroughly.
He states (Vol. II, p. 240) that the crater rows
on Dyngjuháls extend beneath the glacier
Dyngjujökull, and suggests that eruptions can
take place without jökulhlaups being observed
downriver, for the open lavas can absorh tre-
mendous amounts of water that would seep
gradually into the rivers. Tliis, then, miglit
account for the presence of tepliras and the
lack of accompanying jökulhlaups.
Summary
Twenty two of the Bárdarbunga tephras have
been attributed witli various degrees of certainty
to known or suspected eruptions in Grímsvötn
(10), Kverkfjöll (2), other volcanoes beneath
Vatnajökull (5), Katla (2 or 3), and Askja (2 or
3, i.e. 1961, 1875, ancl ? 1919). This leaves six
tephras completely unaccounted for. A new
drillhole strategically placed would probably
acld ntuch to this picture — given a pair of ice-
cores the grain size parameters would come
strongly into the picture in addition to the
íeatures chiefly considered here.
B. Glaciological — Climalological Inferences
Tliickness of Ihe glacier and annual balance
The depth-age data allow an attempt at a
glaciological and climatological interpretation.
Dansgaard (1961) derived an equation to de-
scribe the depth-age relationship in an ice sheet:
t = — ln — (a)
x y w
where H is the total thickness of the ice sheet,
t the age of a given depth horizon at distance
y above the glacier floor, and \ the annual
balance (m yr-1) which amounts to the thick-
ness of the annual layers at the surface.
To derive the equation a number of simpli-
fying assumptions are made, sucli as tliat \
is constant, the glacier is a horizontal sheet
of constant thickness, the horizontal velocity
component o£ the ice does not vary with depth,
and no melting takes place — i.e. the thinning
of the layers is solely due to movement (“squash-
ing out”) of the glacier.
For a given H and \ the relationship between
t and ln H/y is linear, tj = k In (H/H—dj),
where dj is the depth (from the surface) to a
given horizon i. In Fig. 7 this relationship is
plotted for a number of H-values — as seen, a
straight line obtains for H = 518 m, yielding
\ = 2.56 ni/yr. The linear best fit through the
data points was calculated for a number of H-
values, and the correlation coefficicnt determin-
ed. The following Table shows the variation:
H (m) Correlation coefficient
1000 .............991559
600 .............998251
580 .............998702
521 9996060
520 .............9996473
519 .............9996483
518 9996485
517 9996477
500 .............9993650
In Fig. 7 the deviation in depth from the best
fit line for H = 519 m is shown for the 29
data points.
Equation (a) does not allow for the fact that
the density of tlie glacier varies as the snow at
the surface (p = 0.4 g/cm3) recrystallizes to ice
JÖKULL 27. ÁR 21