Jökull - 01.01.2020, Side 61

Einarsson and Jakobsson
describing the effects of historic earthquakes, and 3)
data from analog seismographs from the early years
of seismography.
One of the rules of experimental sciences is that
a scientific experiment must be reproducible, i.e., that
a repeated experiment shows the same result. This
rule is hard to follow in many branches of the natural
sciences. We may anticipate future experiments and
set up our observational networks to obtain the neces-
sary data. But nature itself determines when the actual
experiment takes place and may not agree with our
anticipated course of events. The physical and chem-
ical systems at work are complicated and frequently
we must design our experiments in retrospect and rely
on data that may be available, sometimes by chance.
This is particularly evident when dealing with self-
destructive systems, like the source area of a large
earthquake or an active volcanic system. The next
event is almost never a repeat of the last. This sit-
uation gets worse the larger the previous event was.
As a result the scientific value of a specific data set
may not be recognized until long after it was made. In
the meantime, the data are in constant danger of be-
ing destroyed or lost. The seismological community
is responding to this reality of lost data by launching
extensive projects to copy analog seismograms into
digital formats that can be preserved and used in fu-
ture research (Richards and Hellweg, 2020).
The first analog seismograph was installed in Ice-
land in 1909 and the last one was taken out of opera-
tion in 2010. The number of instruments varied a lot
in the time between these years, reaching a maximum
of about 50 stations in the 1980’s. These were mostly
short-period instruments, suitable for the recording
of local earthquakes and for determination of arrival
times and magnitudes. The seismograms are primary
sources of data on a century of seismic and volcanic
activity in the Icelandic crust. The paper seismograms
were stored in various insecure locations. A project to
preserve the seismograms by digitization was initiated
in 2017. The main objectives were twofold:
1. To save the data from destruction, accidental or oth-
erwise.
2. To make the data easily accessible to the scientific
community for future studies.
In this paper we give an account of the background
of the project and provide guidance to the use of the
data.
A HISTORY OF ANALOG
SEISMOGRAPHS IN ICELAND
Early decades
The first seismograph in Iceland was set up by Ger-
man scientists in 1909 in Reykjavík. It was located in
the old building of the Nautical School (Stýrimanna-
skólinn) at Öldugata 23, that still stands (Figure 1).
This location was presumably chosen because of the
precise time keeping. A visible time mark was given
every day from this location for the benefit of naviga-
tion of ships in the nearby Reykjavík harbor.
The seismograph was a Mainka seismograph that
measured the horizontal N-S component of the ground
movement. Another seismograph of the same type
was added in 1913. It measured the E-W compo-
nent. The operation of these instruments did not go
very smoothly. It was frequently interrupted and fi-
nally stopped in 1914 because of the First World War
(Tryggvason, 1951; Garðarsson, 1999). These seis-
mographs were mechanical, i.e., no electronics. The
amplification was achieved by connecting rods be-
tween the horizontal pendulum and the recording pen.
The recording was by a fine needle on smoked paper
taped on a rotating drum. The soot on the paper was
then fixed in a bath of shellac. The amplification was
low, only 100.
The operation of the Mainka seismographs was re-
sumed in Reykjavík in 1925 and 1927, soon after the
Icelandic Meteorological Office was founded, at the
initiative of Þorkell Þorkelsson, its first director. The
instruments were moved to the new Nautical School
in 1946 and the operation was continuous until 1952,
when improved instruments were installed.
A set of three short-period Sprengnether seismo-
graphs was set up in Reykjavík in 1951–1952. These
had a much higher amplification than the old instru-
ments but were seriously limited by the high level
of microseisms, i.e., continuous tremor originating in
the Atlantic Ocean, with a high spectral peak near
6 sec. period. These instruments had an electromag-
58 JÖKULL No. 70, 2020