Jökull - 01.12.1983, Blaðsíða 76

two settlements lying on either side of the ílood
plain and a fairly busy road crossing it.
Before the Askja eruption of 1961 a great increase
in hydrothermal activity was observed at the site
where the eruption was to occur a month or so later
(iSigvaldason 1964). This led to the idea that the
advent of a subglacial eruption could be predicted
by monitoring the chemistry of the rivers draining
the area with regard to dissolved ions that are of
hydrothermal origin (Sigvaldason 1963, 1965; Stein-
thorsson 1972). For this reason a program was carried
out of periodically collecting water samples from
various rivers in South Iceland to test the feasibility
of this method of prediction empirically. The pres-
ent paper is based on data obtained through that
effort.
FLOOD MECHANISM
The mechanism of jökulhlaups has been under
dispute for a long time, but the currently accepted
hypothesis of jökulhlaups in Skeidará and Skaftá is
that proposed by Bjömsson (1974, 1977): „In a
period of five or six years the water level in the
ice-covered caldera lake (Grímsvötn) rises about
100 m. The glacier is lifted oífa subglacial ridge east
of the lake and water is forced subglacially through
a 50 km long route beneath Skeidarárjökull, causing
vast floods. When the water level has fallen about
100 m the waterway is sealed by rapid plastic de-
formation of the ice at the eastern edge of the Gríms-
vötn depression. The level is not lowered to the
subglacial rim.
The primary cause of jökulhlaups from Gríms-
vötn is the melting of ice in the geothermal area.
The slope of the glacier surface is altered and ice
and water flow into the Grímsvötn depression.“
(Björnsson 1974, p.l).
In the context of the present paper it is important
to note that the ratio ofwater to ice in the caldera at
the time of a jökulhlaup is of no consequence for the
mechanism and frequency of bursts, because the
water squeezing underneath the barrier ”does not
know“ if the overlying column is water or ice. The
water, be it liquid or frozen, is meteoric in origin and
derives from three sources: precipitation in the
caldera itself, glaciers flowing into the caldera from
around, and melt water seeping along the sole of the
glacier into the caldera (all Icelandic glaciers are
temperate, and their temperature is maintained at
the water/ice univariant). Given a constant climate,
i.e. constant precipitation and glacier thickness,
variation in the geothermal activity should result in
variation in the volume of the Grímsvötn floods
rather than their frequency. This is illustrated in
Fig. 1.
Fig. 1. The sketches are meant to illustrate the
efFect of diffcrem activity in the geothermal system
beneath Grímsvötn. A glacier flows into a caldera
lake from the right; the caldera wall to the left. The
floating ice is l/10th above water, and the critical
water level for triggering a jökulhlaup is indicated in
both figures.
A. The geothermal activity causes little melting
at the bottom of the glacier which remains thick.
When critical water level is reached the volume of
the lake water is small.
B. Much thermal activity melts the ice and the
volume of the lake is large when the critical water
level is reached, resulting in a large jökulhlaup.
Mynd 1. Myndimar eiga að sýna áhrif mismunandi jarð-
hitavirkni á vatnsmagn í Grímsvötnum þegar þau hlauþa.
Jökull skríður inn í lónið frá hœgri ogýlýlur á vatninu, en
jarðhitinn brteðir íshelluna neðan frá. A efri myndinni (A)
er tiltölulega lítil jarðhitavirkni og ísinn er þykkur þegar
vatnsborðið hefur náð þeirri hteð að Vótnin hlauþi. A neðri
myndinni (B) er íshellan þynnri þegar sömu vatnsborðshreð
er náð og rúmmál vatnsins meira. Að óbreyttri afkomu
jökulsins retti það því að vera rúmmál jökulhlauþa en ekki
tíðni þeirra sem breytist með virkni jarðhitasvteðisins.
Björnsson (1974) and Bjömsson et al. (1982)
estimated the power of the hydrothermal system of
Grímsvötn to be about 5000 MW (thermal)
comparing the volume of the jökulhlaups and the
mass balance in the area yielding melt water to the
system. This output ofenergy can be accounted for
by an effective thermal exchange of the hydro-
74 JÖKULL 33. ÁR