Jökull - 01.12.1984, Qupperneq 42

ppm (see Ellis and Mahon 1977) and in Icelandic
geothermal waters (NEA data and Table 3) only
saline or C02- waters have magnesium concentra-
tions exceeding 0.1 ppm. We consider the com-
parison with low-temperature C02-water from
Lýsuhóll (40-60 °C, 20 ppm Mg) as done by
Steinthórsson and Óskarsson (1983) to be mis-
leading. Even carbonate-water from high temper-
ature areas would have considerably lower Mg
content as the solubility of Mg minerals increases
with lower temperatures. Therefore, we do not
believe that mixing of meltwater and geothermal
water can explain the observed magnesium con-
centration in jökulhlaups. We suggest instead the
interaction between water and highly reactive
volcanic glass to be a more likely explanation of
the observed Mg concentrations in the water
from Skeidará during jökulhlaups. The magne-
sium was leached from the volcanic glass at
temperatures not exceeding 40 °C, and the water
has not been exposed to higher temperatures in
the lake for any length of time.
The Na/K ratio has been successfully used as a
thermometer of geothermal water by many
geochemists (Ellis 1979, Fournier 1981, Arnórs-
son et al. 1983). However, the use of the Na/K
geothermometer for highly diluted geothermal
water is generally considered questionable (see
Ellis 1979, Fournier 1981, and Benjamin et al.
1983). For waters originally at temperatures
below 100 °C it is applicable only in special cases
(Kristmannsdóttir and Johnsen 1982, Krist-
mannsdóttir 1983). Steinthórsson and Óskarsson
(1983) suggested that the Na/K ratio of water
from jökulhlaups could be used to calculate the
reservoir temperature in the Grímsvötn geoth-
ermal system. The mean value of the Na/K ratio
in samples from the jökulhlaup in February 1982
was 24±2 (standard deviation), in December
1983 24±6 and in 1976 the ratio was 19±4. Sam-
ples of normal water in Skeidará in 1982 and 1983
show a mean value of 24 ±6 for the Na/K ratio.
The variations of this ratio are greater from one
month to another in 1982 and 1983 than between
the jökulhlaups in 1976 and 1983. According to
Steinthórsson and Óskarsson (1983) this ratio was
20 in samples from the jökulhlaup in 1982, 50 in
samples from the jökulhlaup in 1972 but it was 40
in their „normal” water (see their table 3, p. 78).
For comparison, the average ratio of Na/K in the
rivers in SW-Iceland ranged from 17 to 40 in the
years 1972/1973 (Ármannsson et al. 1972, Rist
1974) and the variation in samples from a single
river was in the same range. In the river Súla the
Na/K ratio in normal water varies from 14 to 65
and from 17 to 45 in the river Djúpá. In light of
the described circumstances the use of a Na/K
geothermometer on water in the jökulhlaups
from Grímsvötn (Steinthórsson and Óskarsson
1983) appears questionable.
Substances whose concentrations
are not changed by reactions
in the lake.
Some substances are not expected to have
leached out of basaltic glass in the Grímsvötn
lake. First because the glass does not contain the
substances in appreciable quantities. This applies
for carbonate, fluoride and chloride. Secondly,
the substances are poorly soluble in cold water.
That applies for silica.
Carbonate is not a major component of basal-
tic glass and thus would not be leached in the
lake. The concentration of total carbonate in the
jökulhlaups is the result of plain mixing of the
geothermal component and the meltwater. The
concentration of carbonate (as C02) in normal
river water is 10-30 mg/kg as shown in Fig. 8. The
concentration for the meltwater component in
Grímsvötn is probably lower and an estimate of
20 mg/kg (as C02) seems to be an upper limit. It
is difficult to estimate the carbonate concentra-
tion of the geothermal component. This con-
centration is variable from one high-temperature
geothermal area to another as shown in Table 3.
The concentration of C1 and F cannot be signi-
ficantly affected by water/rock interaction in the
lake but may be slightly changed by increased
volcanic activity as HCl and HF are components
in volcanic gases.
Silica is poorly soluble in cold water. Hence,
no significant leaching is expected in the lake.
Therefore, the observed silica concentration of
the jökulhlaups reflects the original concentra-
tion of the fluid that enters the lake. The discus-
sion of silica will be continued separately in the
next section.
GEOTHERMAL MASS FRACTION
AND THE BALANCES OF MASS
AND ENERGY.
Energy and mass balance equations have so far
been used to describe the heat and mass flow in
the Grímsvötn area (Björnsson 1974, 1983). Now
40 JÖKULL 34. ÁR