Jökull - 01.12.1983, Qupperneq 85

DISCUSSION AND CONCLUSIONS
The pivotal argument of this paper states that the
Na/K ratio of the hlaup water in Skeidará can, with a
suitable correction, yield the temperature of the
geothermal water entering Grímsvötn. Another
crucial argument is that the dissolved silica gives a
measure of the dilution of that geothermal water in
the caldera lake assuming, in accordance with
empirical geothermistry, that low-temperature
water is in equilibrium with chalcedony and high-
temperature water in equilibrium with quartz. If
quartz equilibria were to be used for the 1972-water
the component calculation of Table 5 would give a
negative number for VP, the non-hydrothermal
component. Assuming equilibrium with chalced-
ony for the high-temperature water of 1982, which
in it self would be geochemically incorrect, would
yield somewhat but insignificantly higher silica in
the geothermal water.
As seen in Table 3 the corrective measure of
subtracting the solute concentrations of normal
Skeidará water from those of the hlaup water has
insignificant effect on the Na/K ratios, and hence
the calculated temperatures. A 10% error in Na and
K, in opposite directions, would result in 10% error
in the calculated temperature, which would not
aífect the main conclusions of this paper. We see no
way in which those results, i.e. the drastic change in
geothermal temperature in Grímsvötn between the
periods prior to 1972 and 1982, can be wrong.
Granting the correctness of our geochemical
argument a number ofconclusions follows:
1) I’he temperature of the geothermal water
increased between 1972 and 1982 by some 90°C.
Taken in conjunction with the evidence leading
Tómasson et al. (1974) to suggest a minor eruption in
Grímsvötn during the 1972-jökulhlaup, and the
tenuous indication shown in Table 1 of enhanced
rate of accumulation in Grímsvötn after 1972, the
idea of a volcanic event in 1972 is supported by the
geochemical results.
2) The „meteorological component“ (VP) in the
hlaup water has diminished by a half from 1972 to
1982, suggesting cooling climate and less surface
melting. This is in accordance with the observation
(H. Bjömsson 1982, pers. comm.) that the caldera
lake surface stood exceptionally high before the
1982-hlaup.
3) The geothermal component is water, not
steam, as shown by its high solute concentration. In
order to bring hot water from depth to the caldera
lake a hydrostatic head outside the Grímsvötn area
itself is required. The ice surface of the caldera lake
oscillates between 1350 and 1450 m elevation
(Thorarinsson 1974) whereas the glacier surface in
the Grímsvötn basin to the north (Bjömsson 1974)
reaches above 1700 m. Since the glacier is temper-
ate the eífective water table should stand at about
9/10’th of the glacier thickness below its surface, for
the wet and plastic ice should act as a continuation
of the groundwater below.
JÖKULHLAUPS IN RIVER SKAFTÁ —
CHEMICAL EVIDENCE FOR
A SUBGLACIAL FUMAROLE AREA
The jökulhlaups in river Skaftá (Bjómsson 1977) in
1971, 1972 and 1982 were sampled and analyzed for
the common ions. The overall chemistry of the hlaup
water is diíferent from that of Grímsvötn with
regard to Na, K, and silica relations.
Analyses No. 10-13 in Table 3 show that some
increase in silica and the alkalis occurred in the
1972 hlaup (No. 12) as compared to the normal
Skaftá water (No. 10). In the 1971 and 1982 hlaups
(No. 11 and 13) the silica increase is significant but
the alkalis are unaltered or even diluted as compar-
ed to the normal Skaftá water. Of particular note is
the significant increase in sulphate and carbonate.
The sulphate in the analyses is in fact total sulphur
calculated as sulphate, but about 1/3 of the amount
is sulphide (H2S). The Skaftá river delivers no
homogeneous discharge composition, but a well de-
fined maximum of the geothermal component is
observed during the floods. The chemical data at
hand will not be treated thermochemically in this
paper, but the remark may be made in conclusion
that geothermal water is not a significant part of the
flood water, and that the gases characteristic for
fumarole activity are its dominating dissolved
species. A tentative conclusion is therefore that the
Skaftá reservoir derives its energy mainly from geo-
thcmal steam issuing from a subglacial fumarole
ground. This might suggest different topographic
and tectonic conditions from those of Grímsvötn,
with a localized water pocket overlying the thermal
area (Bjömsson 1977) and no well developed fissure
system for cold water to percolate down.
MONITORING OF SOLUTE CHEMISTRY
For the purpose of predicting jökulhlaups by
monitoring glacial rivers the following points need
JÖKULL 33. ÁR 83