Jökull - 01.12.1983, Blaðsíða 77

thermal system with a replenishable magma re-
servoir at depth (Bjömsson et al. 1982).
CHEMICAL GEOTHERMOMETRY
AND THE GRÍMSVÖTN SYSTEM
In this paper an independent approach is tried
for deriving information about the Grímsvötn
system from the chemistry of the hlaup water. The
method of approach will first be conceptually
described in this chapter and then enlarged upon in
subsequent sections.
The well-researched Krafla area in N-Iceland
(Bjömsson et al. 1979) may be used as a model for the
Grímsvötn volcanic and geothermal system, the
chief difference between the two being that one is
submerged and the other one not. The Kraíla
caldera is traversed by a dike swarm extending N-S.
Below the caldera a magma chamber has been
delineated between 3 and 7 km depth, supplying
heat to the overlying hydrothermal system. In
Grímsvötn the caldera lake and the hydrothermal
system may be viewed as two convection cells,
transporting the magmatic heat from depth to the
bottom of the íloating glacier. The lower system
receives cold water down fissures from the overlying
caldera lake, and probably also from the general
ground water stream flowing toward SW from the
Bárdarbunga — Kverkfjöll high (Bjömsson 1974).
The cold water is heated up and convected toward
the surface. In the upper convective system, that of
the caldera lake, the hot water ascends in a plume
above the geothermal areas to melt the sole of the
floating ice. Chemical analyses of the flood water of
1972 show that the bulk of the caldera lake is of
extremeiy uniform comp>osition, indicating effective
mLxing in the more than 3 km! body of water.
The average temperature in the lake must be 4°C.
Melting of the floating ice will be the most effective
directly above thermal vents and, since near its
melting point ice has very little mechanical strength
the surface would stand lowest where the ice is
thinnest. In Grímsvötn there is such an indication
of thinner ice cover along the south-eastem margin
of the caldera, where the surface of the ice slopes
towards the south and SE, and is lowest at a point
near the SE-comer (Thorarinsson 1974, pp. 227-233).
This slope is some 20 m in height, indicating a 200 m
thick ice cover in Grímsvötn (Thorarinsson 1974).
Apparently the major area of thermal upflow is
along the southern part of the marginal fault in
Grímsvötn, with its most effective part at the SE-
end, near the subglacial outlet of the caldera lake.
Cold water probably percolates down the cooler
parts of the fissure system towards the north to be
convected up again along the southern margin.
Chemical geothermometers have been developed
in recent years and used successfully to evaluate
temperature conditions in geothermal systems (e.g.
Amórsson 1980; Amórsson et al. 1982). The basic
assumption is that water in the hydrothermal
system reaches a temperature-dependent equil-
ibrium with various minerals in the surrounding
rock - in particular that the concentration of dis-
solved silica reaches equilibrium with quartz or
chalcedony at a given temperature at depth and
that the ascending water does not appreciably lose
silica upon cooling. Likewise, the ratio Na/K is
governed by a temperature-dependent exchange
reaction between the minerals albite and micro-
cline.
Unlike simple concentrations like [H4 Si04] the
ratio [Na+]/[K+] is not disturbed by dilution with
pure water. It will be shown in the following that by
using certain correction procedures thc Na/K -
ratio of the Grímsvötn reservoir can be reconstruct-
ed, and the temperature ofthe hydrothermal comp-
onent derived. Then the concentration of silica in
equilibrium with quartz or chalcedony at that temp-
erature is obtained, and by comparison with that
analysed in the hlaup-water the degree of dilution is
calculated. Since the temperature ofthe Grímsvötn
reservoir is maintained at 4°C the heat carried by
the thermal water is used to melt ice, according to
the equation
Ns = NL * Cp (w)/AH[“ * AT
where Ns is the number of moles of solid (ice) that
NL moles of liquid (water) at (4 + AT) °C will melt,
Cp (w) being the heat capacity af water, and AH,1"/
the latent heat of melting for ice.
The difference between the dilution due to melt-
ing and that obtained through the analysis of the
hlaup water must be accounted for by other melt
water, percolating from the surface through the
glacier, or from the continuous melting of the glacier
outside the geothemal system but within the water
basin of Grímsvötn.
THERMAL VVATER IN THE GRÍMSVÖTN
CALDERA
At the end of a jökulhlaup the ice cover of Gríms-
vötn stays flat and smooth, showing that water is left
JÖKULL 33. ÁR 75