Jökull - 01.12.1994, Qupperneq 11

others were omitted because they were thin or
crumbly. In compound flows such as DO 18 and DF
26 only one or two flow units were sampled. The total
number of units in Table 1 is 303. Four sites which
were later thought to represent duplicate sampling of
a flow already cored (DB 0A, DB 0B, DK 13A, DL
9A) have been left out. Two lavas, DC10 and DK 21,
were also discarded from the calculation of average
properties of the collection due to their large within-
flow scatter of directions (alpha95 > 23.5°). The qual-
ity of directional results in the remaining 301 lavas, as
measured by the within-site alpha95, is excellent: the
rms value of this parameter is less than 7°.
Mean fields and dispersions
Statistical parameters for the mean fields and virtu-
al poles from the Isafjarðardjúp area were computed
using FisherT statistics. After inversion of reverse
magnetic directions, the mean field of all lava flows
(except the six duplicate or discordant flows men-
tioned above) has D = 357°, I = 74.3°, vector sum R =
264.3 with N = 301, yielding a circular standard devi-
Fig. 9. A simplified transection of the lava succession south
of ísafjarðardjúp showing the zeolite zoning of the lava
flows. Also shown is the series of compound lava flows in
sections DF and DG. A columnar basalt in the northem part
of Skötufjörður can be traced over a wide area; however, it is
thin or absent in our profile DB.
9. mynd. Einfaldað þversnið gegnum firðina sunnan
Isafjarðardjúps, með ummyndunarbeltum og staðsetningu
dyngjusyrpu í sýnatökusniðum DF og DG. Einnig er merkt
þykkt stuðlað hraunlag sem sést vel í klettabeltinu norðantil í
Skötufirði og víðar.
ation (theta63) of 28.6°, uncorrected for within-flow
scatter. The 95% confidence circle has a radius of
about 3°, so that the mean direction is not significantly
different from a geocentric axial dipole field (D = 0°, I
= 77.3°).
Carrying out the same statistical procedures on the
virtual geomagnetic poles, we obtain a slightly far-
sided mean pole position at 87.5° N, 165°E, with R =
242.6, theta63 = 36.3°. Similar results are obtained if
only the 225 flows in the composite section are used,
or if profiles SZ and SF of McDougall et al. (1984) are
included (see Kristjánsson and Jóhannesson 1989).
The circular standard deviation of the directions or
poles in this area is unusually large for Icelandic lavas.
The present authors have observed (Kristjánsson and
Jóhannesson 1989; Kristjánsson 1995) that a long-
term decrease in this statistical property has taken
place within the past 14 Ma. It is suggested that the de-
crease is largely caused by an increasing stability with
time of the geomagnetic field during reversal transi-
tions and major excursions during this period. There
are at least three episodes in the composite section in
ísafjarðardjúp where the virtual poles are changing er-
ratically through a sequence of several lava flows. One
is at the R to N transition in flows DB 37-38 and DC 6
-10. Another is a prolonged excursion or series of
short polarity reversal events at DD 21-23, DE and up
to DF 12. The third episode is at DK 20-28, perhaps
including also DL 1-2. Similar but possibly shorter
episodes in other paleomagnetic collections in Iceland
are mentioned by Kristjánsson and Jóhannesson
(1989, p. 130). This type of behavior is less pro-
nounced in studies on younger lava flows in Iceland
which also yield lower between-lava circular standard
deviation values. Various world-wide paleomagnetic
studies on Quatemary sediments in the literature dis-
play erratically varying pole positions which may also
be due to geomagnetic instabilities. An instability
episode at 16.2 Ma is recorded by the Steens Mountain
lava sequence in the westem U.S.A. (Coe et al. 1995).
MAIN RESULTS ON STRATIGRAPHY
The polarity column of ísafjarðardjúp and to the west
After elimination of overlapping parts of adjacent
JÖKULL, No. 44
9