Jökull - 01.12.1984, Qupperneq 60

Fig. 3. Acid fallout from the Lakagígar 1783
A.D. eruption as seen in 3 Greenland ice cores:
Camp Century (A), Créte (B) and Dye 3 (C).
The dates have been obtained by stratigraphical
dating methods. The sampling corresponds to 12
samples over the calculated annual layer thick-
ness at the depth in question. More or less than
12 samples per stratigraphically dated year indi-
cate respectively higher and lower precipitation
than the average annual precipitation at the loca-
tion.
Mynd 3. Súr úrkoma frá Skaftáreldum 1783 í
þremur ískjörnum á Grœnlandi, Camp Century
(A), Créte (B) og Dye 3 (C).
In Fig. 3 the annual layers have been separated
by means of the oxygen isotope seasonal record;
the annual layers are therefore not separated into
exact calendar years, but into oxygen isotope
years, from minimum to minimum of the ð(lsO
seasonal variation — close to a calendar year. For
the moment no method exists by which the true
yearly shift can be established, but using the
minima of the seasonal oxygen profile is a close
enough approximation. (Dansgaard 1964, p. 445,
Fig 4)
Even though the oxygen profiles are smoothed
with time and depth in the core due to diffusion,
they can in many cases be reestablished by a
back-diffusion technique (Johnsen 1977). The
dating can also be supported by other strati-
graphical dating methods (Hammer et al. 1978),
which are less prone to diffusion.
In the case of high accumulation areas i.e.
more than approx. 20 cm precipitation per year
(in water-equivalent) the difference between the
true shift of the year and the ð(lsO) definition of
the shift will in general only change the “true”
yearly average acidities little; unless a high volca-
nic acid signal is confined to a few winter months.
The latter is not very likely, because strong volca-
nic signals tend to last more than a few months.
The Lakagígar signal is of short duration i.e.
approx. 1 year.
The immediate appearance of the Lakagígar
signal in the early summer snow of 1783 is interes-
ting (interpretation from the ð(lsO) profile).
because S02 injected into the stratosphere by a
volcano takes some months to be transformed
into H2S04 which has a residence time of approx.
1 year in the polar stratosphere. The above fin-
dings agree with our present knowledge on the
eruption: The Lakagígar eruption was only a
moderately explosive eruption and the large
amount of liberated S02 was to a large degree
confined to the troposphere, which speeds up its
transformation to sulfate (Junge 1963, p. 68), the
acid deposition on the ice sheet declined quickly,
when the eruptive activity ceased.
Thorarinsson 's minimum estimates on the acid
production (changing his S02 values to H2S04
values) gives 21.1 mill. tons of H2S04 + HCl (1.2
mill. tons), while my own estimates based on the
Créte core, gives 100 mill. tons. In a correspond-
ence with Thorarinsson 1977 he wrote — —
Using ca. 1% water percent in basalt magma, as
some do, and the same fractional composition of
the gas as in Surtsey, we get a max. of ca. 40 mill.
tons or the same order of magnitude as your
estimate.” (translated to English). My estimate
may be a little on the high side, because the
Icelandic eruptions seem to give a somwhat
higher acid deposition on Greenland, than one
would expect from the fission product data.
Selected data from the latter profiles, combined
with the global fallout pattern, was used to relate
ice core acid data to the acid production by an
eruption (see e.g. Hammer et al. (1981)).
Presently the Lakagígar event has been
detected as highly elevated acidity in the follow-
ing ice cores: Dye 3, Dye 2 (South Greenland),
Milcent, Créte (Mid Greenland), Camp Century,
North Central, Hans Tavsen (North Greenland),
Devon Island, Ellesmere Island (North-East
Canada), (D.A. Fisher personal communication,
1982).
58 JÖKULL 34. ÁR