Jökull - 01.12.2003, Blaðsíða 21

The 1783–1785 Laki-Grímsvötn eruptions
east of Mt. Kaldbakur as either Landnorðursgjá (the
northeast vents) or Austurgjá (the east vents).
The fires at the third eldgjá were seen at least
10 times from July 1783 to May 1785 (Steingríms-
son, 1788). These fires were situated east-northeast
of Kirkjubæjarklaustur according to observers and re-
ferred to as Austur-Landnorðursgjá (the east north-
east vents). The actual location of this site has been
debated in the past, because no historical eruption
sites are known due east-northeast of Prestbakki or
Kirkjubæjarklaustur (Thoroddsen, 1925; Thorarins-
son, 1974; Thordarson and Self, 1988). However, the
following quotations taken from Steingrímsson’s de-
scriptions give a clear indication on the location of this
eldgjá:
26 September 1783: “Another additional fire and steam
column, which had been seen before that summer, far east-
northeast of the Síða district, due north of Lómagnúpur
mountain, but north-northwest of Öræfi district, as it was
oriented by me and 8 others.”
January 1784: “Fires were still seen at the East-
northeast f issure. From Klaustur [i.e., Kirkjubæjarklaust-
ur], these fires were seen behind the northern end of Þverár-
fjall mountain. The inhabitants of Öræfi district had a better
view of these fires due to their location.”
Firstly, these descriptions show that this eldgjá
(i.e., eruption site) was separated from the other two
by a considerable distance to the east. Secondly, de-
spite the fact that Steingrímsson says the fires were
east-northeast of Kirkjubæjarklaustur, his directional
information indicates otherwise. The fires were seen
behind the northern end of Þverárfjall, showing that
they were actually located to the north-northeast (20–
30◦) of Kirkjubæjarklaustur (Figure 4). The only
way these fires could be observed from the vicinity
of Lómagnúpur Mountain was through the valley of
the River Núpsvötn, which is located along the east-
ern side of the mountain. This valley trends 3◦ east
of north. The third orientation given by Steingríms-
son is from the Öræfi district east of Skeiðarársan-
dur, where the fires were seen to the north-northwest.
When these three cardinal references are plotted, it
is clear that the Austur-Landnorðursgjá eruption site
was at the Grímsvötn volcano (Figure 4).
COURSE OF EVENTS AS REVEALED
BY THE CONTEMPORARY ACCOUNTS
THE PRECURSORS OF THE ERUPTION
A farmer, Jón Eiríksson at Ljótarstaðir in the Skaft-
ártunga district observed the first sign of activity in
the area when he felt weak tremors in mid-May 1783.
These tremors largely went unnoticed until 29 May
when they increased to such an extent that the inhab-
itants in Skaftártunga district evacuated their homes
and slept outside in tents for their own safety. On 1
June strong earthquakes were felt over an area extend-
ing from Mýrdalur in the west to Öræfi in the east, to
a distance of about 75km from the Laki fissures. The
intensity and duration of these earthquakes steadily
increased through the first week of June or until the
onset of the eruption on 8 June (A1–3).
Signs indicating volcanic activity were seen in the
mountains or in the glaciers above the Fire districts in
the spring 1783 before the onset of the Laki eruption.
The crew of a brig claimed that they saw fires some-
where above the Fire districts, as well as southwest
of the Reykjanes peninsula, as they were passing the
southern coast of Iceland on their way to Hafnarfjörð-
ur in SW-Iceland (Figure 1, inset). It is not known
whether this brig was the Boesand, which arrived at
the beginning of May or the Torsken, which arrived
around 20 May. It was also reported that the inhab-
itants of the Fire districts often saw a bluish smoke
above the ground, which they related to volcanic ac-
tivity (B1). The location of the above-mentioned
activity is unknown, but may have been within the
Vatnajökull ice cap because the inhabitants of the Fire
districts did not specify the location of the site.
In summary: The Laki eruption was preceded by 3–4
weeks of seismic activity. It began with weak earth-
quakes in mid-May 1783 that increased in intensity
and frequency until the beginning of the eruption on 8
June. If the observation by the brig’s crew is correct,
this precursory phase may have resulted in a short-
lived eruption sometime in May 1783.
FIRST 45 DAYS OF THE ERUPTION
In clear and calm weather on Whitsunday, 8 June
1783, at about 9 a.m., people in the Fire districts
JÖKULL No. 53, 2003 19