Jökull - 01.12.1984, Page 63

Fig. 6. Acid fallout in 3 Greenland ice cores;
probably from an eruption in Katla 1179 A. D.
Sampling as in Fig. 3.
Mynd 6. Súr úrkoma í þremur ískjörnum frá
Grœnlandi; líklega frá Kötlugosinu 1179.
good correspondence between all the cores for
this „event“. A similar correspondence between
the nonvolcanic background of the 3 cores can
not be expected, as that would require unrealistic
assumptions with respect to the meteorological
conditions over the ice sheet. The deposition of
acids from a large volcanic eruption is of course
also influenced by meteorological conditions.
Flowever, the additional volcanic acids in the
atmosphere increase the acidity of the deposited
snow, as compared to the situation, where no
volcanic acids are present in the atmosphere.
To what extent acid profiles along ice cores
from different locations on the ice sheet correlate
over periods of no or little volcanic activity has
not yet been solved. Clearly a high correlation
can only be expected between locations, where
summer temperatures above 0°C are rare and
where the yearly precipitations are not too small.
Even in this case meteorological conditions, the
transport of the non-volcanic acid from the
source regions and the strength of the sources
would influence the degree of correlation.
The Hekla I eruption 1104 A.D.
and dating accuracy.
There is clear evidence, that the Hekla I erup-
tion was an explosive eruption (Thorarinsson
(1956 p. 21), but how much acid gas it released is
less known.
If measured in freshly fallen tephra the erup-
tion produced approximately 2 km3 of material
(Larsen and Thorarinsson 1977). It should there-
fore be possible to detect the acid traces in the
Greenland Ice Sheet. The explosive character of
the eruption is revealed in the isopachs of the
Icelandic tephra deposition, and the plume must
have reached far into the stratosphere. It is there-
fore to be expected, that the deposition of the
volcanic acids on the ice sheet took some years.
Again, the Créte acidity record is the most
trustworthy, Fig. 7, while, as mentioned earlier,
the Dye 3 record is rather noisy around the time
in question. As this eruption is the oldest histori-
cally well dated eruption, which left the Green-
land Ice Sheet with an undoubtful acid volcanic
signal, it was used to check our dating accuracy
for the Créte core, (Hammer et al. 1980). The
maximum acid deposition from the eruption is in
1105: Not a too bad dating, considering that all
Icelandic annals give 1104 as the eruption year
(Thorarinsson personal communication, 1979).
Such an accurate dating is only possible for
Central Greenland ice cores and only if the yearly
precipitation exceeds some 20 g/cm2 year of water
equivalent. The Dye 3 dates are not too bad, but
melting and a less “isolated” position with respect
to the cyclone paths of the North Atlantic Ocean
give rise to a larger number of irregular seasonal
variations in both ó(lsO), dust, acidity etc. Hence
a larger number of “doubtful years” appears in
the ice record. If e.g. 9 such years appear over
1000 annual layers, and there is an equal chance
of missing a real “year” or adding a wrong
“year”, the dating accuracy will be ± 3 years over
1000 years.
JÖKULL 34. ÁR 61