Jökull - 01.01.2016, Síða 84

Leó Kristjánsson
by the rock during its original cooling, and viscous
remanence (secondary VRM) assumed to have been
built up gradually in the current interval of normal po-
larity. In most of those lava flows in Iceland which
have not suffered re-heating above 100◦C in situ with
hydrothermal alteration, the latter component can be
eliminated by appropriate treatment with alternating
magnetic fields (AF demagnetization). Usually, peak
fields of 20 or even 10 milli-Tesla (mT) will suffice
to isolate a TRM component which is consistent from
one sample from another. Such internal consistency
in a unit is indeed a crucial test of its reliability as a
recorder of the geomagnetic field direction during its
emplacement.
In igneous rocks that have been heated to more
than 100◦C on burial or been deformed, the primary
remanence has decayed, with concurrently increas-
ing preponderance of secondary components. Above
250◦C, the original magnetic minerals may be re-
placed. A number of experimental and statistical tech-
niques have been developed for research on poor-
quality rocks, involving assumptions which are some-
times not substantiated. Coe et al. (2014) discuss an
example where a far-reaching incorrect conclusion on
the properties of the geomagnetic field persisted in the
geoscience literature for a quarter-century.
Each polarity transition may take a few thousand
years. Ideas of these transitions, their frequency and
relations to the secular variation have changed much
in recent decades and continue to be debated. This
also applies to so-called excursions of the geomag-
netic field, when VGPs appear to wander for a while
far from their usual habitat in the vicinity of one ge-
ographic pole or the other without a polarity rever-
sal taking place. The evidence for the transitions
and excursions has been derived from many different
kinds of geological materials. The coverage of each
of these records is generally incomplete, and the pri-
mary remanence of formations which originally pre-
served them may be unstable or disturbed by subse-
quent processes.
Some relevant results from abroad on transitions
and excursions
The extensive literature on geomagnetic transitions
and excursions will not be reviewed here. Examples
of two apparent types of behavior in lava sequences
are given below. One type yields a cluster of mid- and
low-latitude VGP positions (often termed intermedi-
ate or transitional, T). In the other type, VGP positions
seem to have moved irregularly all over the globe.
A complex geomagnetic event close in age to the
one described in the present paper is an R-T-N-T-N
transition in the island of Gran Canaria (Leonhardt et
al., 2002). Some 32 lava units were erupted during the
central T-N-T part of this transition, which occurred
at about 14 Ma. It is largely composed of clusters of
VGPs; one at around 40◦S is recorded in 13 flows.
Directions in units of that cluster correlated well be-
tween two sampling profiles overlapping in age, about
half a kilometer apart. In another example of a VGP
cluster (Glen et al., 2003), nine successive flows with
low-latitude poles occur in a partial record of a tran-
sition or excursion of about 10 Ma age, also in the
Canary Islands.
Watkins (e.g. 1969) reported results from a 71-
lava section in Steens Mountain in southeastern Ore-
gon where the geomagnetic field direction appeared to
be changing with unusual irregularity between thick
zones of reverse polarity below and normal polar-
ity above. The lava flows of Steens Mountain have
subsequently been studied in detail (Coe et al., 2014
and references therein), resulting in the definition of
a complex R-T-N-T-N-T-R-T-N transition, including
some clustered groups. The age of this episode has
been determined to be about 16.7 Ma.
Relevant paleomagnetic research in Iceland
The lava pile above sea level in Iceland probably con-
tains millions of flows, of ages from 0 to at least
16 Ma. Research on magnetically stable lava series
in Iceland contributed early on to the development
of some key concepts in this branch of geoscience.
Among these concepts are the conventional statistical
parameters used in the processing, interpretation and
comparison of directional results. Thus, the consis-
tency of N directions having a vector sum of length
R is estimated through a precision parameter k =
N-1/N-R, an angular standard deviation (asd) of the
sample directions around the vector sum, and a 95%
confidence angle (α95) for the sum direction (Fisher,
1953).
84 JÖKULL No. 66, 2016