Jökull - 01.12.1983, Síða 84

higher degree of boiling at depth as inferred from its
lower C02/H2S ratio.
5) The overall chemistry of the restored Gríms-
vötn fluid, particularly that of the 1982 hlaup,
matches in many respects the composition of geo-
thermal water from Lýsuhóll in Snaefellsnes, west-
ern Iceland (No. 9, Table 3). Ofparticular note are
high levels ofcarbonate, and similarities in sulphate
and alkaline earth metals.
GEOCHEMICAL MODELS OF
THE GRÍMSVÖTN SYSTEM
As already indicated in the Introduction the
analysis of the hlaup water from Grímsvötn permits
several independent inferences to be made about
the geothermal system:
1) The discharge peak is homogeneous, indi-
cating an efíicient convective cell in the Grímsvötn
reservoir.
2) Geothermal water is added to the caldera
lake, as reflected by the high amount of dissolved
metallic ions.
3) The temperature of the geothermal water ent-
ering the caldera lake is estimated from the meas-
ured Na/K ratio, and hence the relative proport-
ions of the three water components can be calculat-
ed, i.e. geothermal (G), melted ice due to the geo-
thermal activity (M), and other (P). This latter
component is constituted of melt water from the
surface and within the ice, and from melting at the
bottom of the glacier due to heat flux outside the
geothermal area.
Two models can be distinguished for the origin of
the geothermal water, i.e. (1) that it is externally
derived ground water, in which case it adds to the
volume of the caldera lake, and (2) that it is re-
circulated focal water and as such only transports
energy from depth to the system.
(I) The first model may be expressed in the follow-
ing equations: VT = VG + VM + VP (the volume of
water in the caldera lake is at any time the sum of
the geothermal component, that due to melting at
the bottom of the floating ice, and that melted else-
where). D = VT/VG (dilution factor). VM = kVG T
(volume of melted water is proportional to the
temperature and volume of the geothermal water).
(II) In the second model VG is included in VP and
VM, and the volume of the lake is VT = VM + VP.
From these equations the proportions of the three
constituents can be calculated for the two models,
as shown in Table 5 — the total volume VT merely
scales the numbers up and down.
Of these models the second one is more realistic
in view of the fact that the water basin ofGrímsvötn
is about the highest-lying area in Iceland, and the
ground water must be derived from within it. The
energy output calculated here is similar to that
obtained previously by totally different methods
(Bjömsson 1974 and Bjömsson et al. 1982) if similar
volumes are assumed for the Grímsvötn reservoir
prior to the hlaups of 1972 and 1982. It is noteworthy
that in both models the part (relative and absolute)
ofVP is ntuch smaller in the 1982 hlaup which might
be in accord with cooling climate and diminishing
melt water at the surface of the glacier. Prior to the
1982-hlaup the water in the Grímsvötn iake stood
higher than hitherto recorded (H. Bjömsson, pers.
comm. 1982), which would be in accord with a
thickening ice-barrier at the outlet of Grímsvötn.
Table 5. Relative proportions ofwater-constituents in the Grimsvötn caldera lake, according to Models I
and II.
Tafla 5. Hlutfall jaröhitavatns (G), brœðsluvatns af völdum jarðhita (M) og annars bræðsluvatns (P) skv. tveimur
líkönum.
Model VT (%) VG (%) VM (%) VP (%) V (km3) VP (km3) Power (MW)
I. 1972 100 35.7 45 19.3 3.2 0.62 2675
1982 100 26.6 63 10.4 3.5 0.36 4095
II. 1972 100 - 70 30 3.2 0.96 4160
1982 100 — 86 14 3.5 0.49 5590
The 1972-models are calculated for equilibrium with chalcedony, the 1982-ones for equilibrium with
quartz. The volume of the 1972 hlaup is as measured (Table 1) whereas that of 1982 was estimated
assuming the monthly rate of accumulation of the period 1972—1976, and the observed period between
1976 and 1982, 65 months (Table 1).
82 JÖKULL 33. ÁR