Jökull - 01.12.1990, Side 154

greatest chemical diversity (Sigvaldason et al., 1974).
Seismic evidence supports the hypothesis that this area
is underlain by the central part of a mantle plume of
low velocity material (Tryggvason et al., 1983; Ge-
brande et al., 1980). This section of the neovolcanic
zone is largely covered by the ice cap Vatnajökull
(Fig. 1), and therefore its geological structure is only
poorly known. The area has been volcanically active
in historical times and early seismic work revealed it
as one of the most seismically active areas in Iceland
(Tryggvason et al., 1958; Tryggvason, 1973). Erup-
tions have led to tephra fall and floods of meltwater,
often damaging vegetation, farms and roads in coastal
areas. Because of their remoteness and glacial cover,
the location of many eruptions is uncertain. Many
have been ascribed to the well known Grímsvötn vol-
cano, but the wide distribution of epicentres in the
area and recent attempts to relocate eruption sites from
bearings taken to the eruption plumes indicate that the
active area is considerably larger (Þórarinsson, 1974).
New insight into the geological structure was
gained when satellite images of Vatnajökull became
available (Þórarinsson et ai, 1973). In some areas,
the subglacial bedrock topography is reflected in the
surface forms of the ice. A caldera was seen to ex-
ist beneath the ice cover of Bárðarbunga,which was
thus identified as an active volcano. Caldera struc-
tures were also seen in other volcanoes such as Kverk-
fjöll, Grímsvötn and Tungnafellsjökull (Fig. 2). It be-
came evident that the geological structure of the area is
dominated by several large central volcanoes. Fissure
swarms outside the ice cap have been traced to the
glacier edge (Sæmundsson, 1978, 1980; Jakobsson,
1979, 1980), but so far it has been unclear how these
are connected with central volcanoes beneath the ice
to form extensive volcanic systems. The term ”vol-
canic system“ was originally introduced by Jakobsson
(1979) to cover both the term central volcano and
its associated fissure swarm. One may define a vol-
canic system as a spatial grouping of eruption sites
with particular tectonic, petrographic and geochemi-
cal characteristics. In some volcanic systems the fis-
sure swarm is the most prominent structure, in others
the central volcano is well developed, sometimes with
a caldera, extensive acidic volcanism and geothermal
activity. A volcanic system may contain more than
one central volcano (e.g. Hofsjökull- Kerlingarfjöll)
and fissure swarms sometimes branch into two or more
subswarms. One may therefore easily get into prob-
lems with the definition of individual systems.
During the past ten years, the bedrock topography
beneath the ice has been mapped extensively using
radio echo sounding (Björnsson, 1988; Bjömsson,
in prep.). At the same time increased seismograph
coverage has led to more precise epicentral locations
(Einarsson and Bjömsson, 1987; Einarsson, 1991)-
This information has added significantly to the under-
standing of the geology of the area beneath Vatnajök-
ull. Seismic activity in conjunction with topographic
highs reveals active volcanoes. Ridges, presumably
built up by subglacial fissure eruptions, show the lo-
cation of fissure swarms. The purpose of the present
paper is to integrate this recently acquired knowledge
of the subglacial volcanic regions in Vatnajökull. The
volcanic systems in the area are delineated and the
paths of meltwater produced in eruptions are traced to
the various rivers draining the glacier. Information on
historical eruptive activity is reconsidered in the light
of this evidence.
MAPPING OF BEDROCK TOPOGRAPHY
AND SEISMIC ACTIVITY
BEDROCK TOPOGRAPHY
The topographical maps of Vatnajökull (Figs. 3
and 4) are based on continuous ice-thickness profil-
ing by radio echo-sounding and precision barometric
altimetry (see Bjömsson, 1988). The maps were con-
structed by interpolation between sounding lines using
a digital matrix with a grid spacing of 200 x 200 m.
Due to the large spacing between the sounding lines
(typically 1000 m), the topographic maps do not usu-
ally represent features smaller than some kilometres
across. The accuracy of the absolute ice-thickness
measurements along the sounding lines is of the order
of ± 15 m, and the accuracy of the measurements of
absolute ice-surface elevation is of the order of ± 10
m. Relative resolution of the data with respect to to-
pographical features is considerably better than this,
however. Volcanic and tectonic structures of the order
150 JÖKULL, No. 40, 1990