Jökull - 01.12.1984, Page 45

ture, and its concentration is used as a geoth-
ermometer (see Ellis 1979, Fournier 1981,
Arnórsson et al. 1983). When water is close to
magma, self-sealing due to precipitation of silica
puts an upper limit of 330 to 350 °C to the
temperature of the fluid (Fournier 1983). With
increasing temperature, quartz has a solubility
maximum at constant pressure. When this max-
imum is reached (at about 340 °C) precipitation
of quartz deep in hydrothermal systems may
decrease the permeability to such an extent that
convecting meteoric water no longer can attain
temperatures higher than that given by the quartz
solubility maximum. Known reservoir tempera-
tures in Icelandic high- temperature geothermal
areas range from 240 to 350 °C. Table 3 shows
chemical concentrations of well discharges for
five liquid dominated geothermal areas in Ice-
land. The first three areas have boiling reservoirs
with dilute fluid of meteoric origin. The table
shows concentrations for both total discharge and
deep water. The two other areas have saline
reservoir water. The table shows the deep water
concentrations. Further, calculated concentra-
tions are given for water boiled at 235 °C for all
the areas. Grímsvötn is a high-temperature
geothermal system and the reservoir temperature
is presumably above 300 °C. The fluid is dilute
and probably liquid dominated. Boiling would
occur at 235-250 °C on the lake floor, depending
on the height of the lake level.
The concentration of silica in the deep reser-
voir water may be Cgw = 700 mg/kg. If fluid of
that concentration were discharged into Gríms-
vötn, we would estimate the geothermal mass
fraction k=0.13 from equation (8). According to
the calculations, illustrated in Fig. 9, the mass
and energy balances would require the steam
mass fraction to be x= 0.45 when the fluid enters
the lake. During upflow, however, deep water as
well as condensed steam would equilibrate with
the formation rocks at or above 235 °C (given a
few hours or days, see Rimstidt and Barnes 1980).
Hence, we estimate the silica concentration Cgw=
400-600 mg/kg in the water entering the Gríms-
vötn lake (see Table 3 for comparison). Further,
we estimate the geothermal mass fraction k=
0.14-0.16 and the energy balance requires steam
mass fraction x=0.20-0.35 for the fluid dis-
charged to the lake; the mass flow of geothermal
water Mgw= 0.60-0.83T011 kg/yr and Mgv=
0.24-0.34T011 kg/yr of steam. The mass of ice
melted in the lake is estimated to be M;=
4.0-4.2T011 kg/yr. Furthermore, we expect the
total thermal power of the Grímsvötn system to
be 4700-4900 MW, of which 2100-3000 MW are
transported by steam and 1900-2600 MW by
water (see Fig. 9).
Calculations similar to those for silica are diffi-
cult for carbonate. A plausible estimate is not
available for the carbonate concentration of the
geothermal component as it varies from one high-
temperature area (>300 °C) to another (see
Table 3). But if we assume k=0.15 and if the
meltwater component contains 20 mg/kg carbon-
ate (as C02), (and Cr=Ca=Q), we can calculate
the concentration for the geothermal component
that would be consistent with the measured con-
centrations in the jökulhlaups. The calculations
show variations from C=2000 to 4500 mg/kg (as
C02) for the geothermal component. This is high
but not unlikely in an active volcanic area. Direct
interaction with magma has been observed in the
geothermal systems in Krafla and Námafjall
(Björnsson et al. 1979). The concentration of C02
in geothermal fluids in the Krafla area increased
considerably during the recent volcanic events
(Armannsson et al. 1982).
The concentrations of fluoride and chloride
may well be consistent with a geothermal mass
fraction k=0.15 (see Table 3).
VOLCANIC ACTIVITY DEDUCED
FROM WATER CHEMISTRY
The high concentrations of sulphate and iron
(as well as carbonate) during the jökulhlaup in
December 1983 suggest direct contact between
magma and geothermal fluid.
Sulphate (S04) in the Grímsvötn lake origin-
ates from oxidation of H2S as well as from the
S04 in the geothermal discharge. The contribu-
tion from the meltwater is small as is evidenced
by the glacier rivers when not influenced by
jökulhlaups (Fig. 8). The concentration of sul-
phate will be influenced by volcanic activity. We
may even expect a sharper increase in sulphate
than carbonate shortly after volcanic activity
because H2S (and S02) is more soluble in water
than C02. This may explain the very high con-
centration of sulphate in the jökulhlaup of
December 1983 as compared to those of 1972,
1976 and 1982. The reported concentration of
sulfate in the jökulhlaup in 1965 was also very
high (Sigvaldason 1965).
JÖKULL 34. ÁR 43