Jökull - 01.12.1977, Síða 20

DISCUSSION
The time-depth proíile, and the tephras them-
selves, are of considerable volcanological interest
in that they add to our knowledge of volcanic
activity in the regions surrounding Bárdar-
bunga. In the following, some of the tephra
layers will be described and discussed, and in-
ferences drawn.
A. Volcanic history
Grímsvötn
For clarity the designations of the Grímsvötn
tephras (“G” in the diagram, Fig. 3) appear on
the left of the curve. Of the 14 known or
suspected eruptions in Grímsvötn since 1650
(Thorarinsson 1974), probably 11 appear in the
core. Of these, nine are distinguished by the
“right” chemical composition, and one (1706)
has “wrong” chemical composition. Tephras of
four eruptions are missing, one alleged Gríms-
vötn eruption (1706) is deleted from the list on
the basis of chemical composition, whereas one
tephra layer, corresponding to 1854, may derive
from an unknown eruption in Grímsvötn.
The nine tephra layers which exhibit complete
correspondence between volcanic history (Thor-
arinsson 1974), position in the core, and chemi-
cal composition, are G 1934, G 1922, G 1892,
G 1883, G 1838, 1784, G 1774, G 1716, and
G 1659. Of these, four in particular merit
further discussion becouse of ambiguities re-
garding their existence or origin.
G 1892: A jökulhlaup in Skeidará took place in
March 1892, whereas for the period from Nov-
ember 1891 to January 1892 dispersed references
can be found to an eruption in Vatnajökull
(Thorarinsson 1974, pp. 97—105). The tephra
layer, of correct chemical composition and rea-
sonable grain size characteristics, confirms the
suspicion that an eruption was associated with
the jökulhlaup.
G 1838: In 1838 earthquakes were felt in North
Iceland (June 12), Skeidará flooded, and some
ashfall was observed. Björn Gunnlaugsson the
surveyor, standing atop a mountain, described
smoke billowing up through Vatnajökull (Thor-
arinsson 1974, pp. 70—71). The presence of a
1 8 JÖKULL 27. ÁR
tephra layer in tíie core corresponds to this year,
and its Grímsvötn chemical composition in-
dicates that an eruption in Grímsvötn did,
indeed, take place in 1838.
G 1784: In 1784 Skeidará flooded, and a vol-
canic cloud was observed over Vatnajökull.
Thorarinsson (1974, p. 62) states: “It may be
concluded tliat in the spring and summer of
1784 fires were abroad in Vatnajökull, but not
in Grímsvötn; that a jökulhlaup took place in
early April 1784, perhaps accompanied by an
eruption in Grímsvötn. . . . According to Sveinn
Pálsson, the Lakagígar Fires from 1783 burned
into the winter of 1785.”
This latter statement by Sveinn Pálsson is of
considerable interest, for the Bárdarbunga
sample attributed to 1784 contains Pelé’s hair,
which is characteristic for powerful Hawaiian-
type fountain eruptions. In fact, Pele’s hair has
been found in tephra from Lakagígar (Jón
Steingrímsson 1783, S. Thorarinsson, pers.
comm.), but it must be considered highly un-
likely in Grímsvötn tephra. However, according
to the depth-time profile, the sample must be
allotted to 1784 or 1785 (not 1783), and could
therefore, in the light of Pálsson’s statement,
be derived from a late phase of the Lakagígar
eruption.
G 1659: The written sources make only oblique
references to an eruption in Grímsvötn in 1659
(Thorarinsson 1974, pp. 51—52). Allotting the
Bb 524 tephra to G 1659 engenders an unknown
eruption of unknown chemistry about 1661 (the
sample Bb 522 was lost), wliich might conceivably
be identified with the Katla eruption of 1660-61.
The tephra in that eruption went mainly to
the SE (G. Larsen, pers. comm.). The Gríms-
vötn-like chemical composition (tholeiite) of Bl)
524 renders tliis reconstruction preferable to an-
other possible one, in which the missing layer
Bb 522 is allotted to G 1659, and the lower one
to an unknown subglacial eruption of tholeiite
composition not far from Bárdarbunga in 1657.
Because of its distinctive chemistry, a Katla-type
composition in Bb 522 would have completely
“nailed down” the profile.
A Grímsvötn eruption in 1854? The 1854-tephra,
which has a Grímsvötn-like cliemical composi-