Jökull - 01.12.1983, Qupperneq 15

A natural Calorimeter at Grímsvötn;
an Indicator of geothermal and volcanic Activity
HELGI BJÖRNSSON
Science Institute, University of Iceland,
Dunhagi 3, 107 Reykjavík, Iceland
ABSTRACT
The heat output of the subglacial geothermal area at
Grímsv'ótn is estimated from the total volume of water in
j'ókulhlaups which originate from the caldera. The volume
melted by the subglacial heat source is estimated by sub-
tracting the glacier surface ablation from the total water
volume. Variations in heat output are presentedfor the last
125 years. An average or báse heat Jlux of 4000—5000
MW, derived by this method, is in agreement with a former
estimate of the average heat flux, which was obtained by a
mass balance study on the glacier. This heat flux has been
explained by the heat extraction of water penetrating into
hot rock boundaries of magma. A decline has been observed
in this base flux over the last 80years, which may be due to
lack of recharge to the magma body. In the same period a
comparable increase has been observed in the heat output of
another thermal area 10 km northwest of Grímsv'ótn. Four
peaks in the heatflux at Grímsv'ótn rise by a factor of2to3
above the base flux. We suggest that at least two of them
are caused by intrusion of magma at the base of the glacier.
The total inflow of magma into the Grímsv'ótn area is
estimated to be about 50 X l(f nfyV1 of which 3% has
been extruded on the ice cap surface, 10% transported to the
glacier bed and 87% have been solidified in the upper crust.
This inflow of magma equals the average annual rate at
which magma has been erupted to the surface of lceland.
The heat energy removed by volcanic eruptions is an order of
magnitude less than the heat removed by hydrothermal
convection.
INTRODUCTION
At Grímsvötn in Vatnajökull nature offers a uni-
que calorimeter for estimating the heat flow from a
geothermal area. (Fig. 1). The ice cap covers the
thermal area and almost no heat escapes into the
atmosphere. Hence, the volume of ice melted at the
glacier bed is a measure of the heat released by
thermal activity within thearea. A long-term mass
balance model for the drainage basin has given an
estimate for the average heat flux at Grímsvötn of
5000 MVV (Bjömsson 1974). The heat flow can be
estimated by an alternativeapproach. First we esti-
mate the volume of meltwater which drains down to
Skeidarársandur in jökulhlaups from Grímsvötn
(Fig. 1), second the volume melted at the thermal
area is separated from the total volume of the jökul-
hlaups by subtracting the surface ablation. This
approach does not only givean average value for the
heat flux, but also hints at fluctuations in the heat
output.
OBSERVATIONS
Since 1934 all jökulhlaups from Grímsvötn have
been observed by scientists who have estimated
their volumes. All the way back to 1850 estimates
of volumes exist. They have been inferred from
detailed descriptions of the floods by inhabitants in
the flood area, and based on comparisons with
measured floods after 1954 (Thorarinsson 1974, Rist
1955, 1976, Stefánsson 1983). A conservative esti-
mate of the water volume is given in Fig. 2a. For
calculating the volume melted by the subglacial
heat source, glacier surface ablation of0.15 km3yr_l
was subtracted from the total volume of the jökul-
hlaups. This value for the ablation is based on
observations on Vatnajökull for the last 30 years
(Bj'ömsson 1974) and may be a rather high value for
the whole period considered. Fig. 2b shows the
computed rate of melting by the subglacial heat
source and the heat flux required to melt ice at that
rate. In this estimate we have assumed that each
jökulhlaup drains out the volume of water melted
since the last jökulhlaup. This implies that the lake
is assumed to drop down to the same level in each
jökulhlaup. Exceptions from this rule may intro-
duce false fluctuations. If the lake level fell less
than normal in one jökulhlaup, the rate of melting
since the last flood would be underestimated, but
JÖKULL 33. ÁR 13