Jökull - 01.12.1989, Blaðsíða 64

standard normal variable. The third step was to test
the relationship of the form parameters with grain
size within samples. This was done to decide
whether it would be feasible to use sample means
based on all measured grains, or whether comparis-
ons between samples would have to be limited to
certain grain sizes. The fourth step was to use the
SYSTAT statistics module to calculate means, stan-
dard deviations and summary statistics for the
parameters. In the fifth step, cluster analysis was
employed to exclude highly correlated variables,
and then to cluster samples on form and chemical
parameters. Finally, the relationship between vari-
ous form parameters to each other and to chemical
composition was analyzed on plots (step six).
The measurement of the shape of sand sized parti-
cles under the microscope is a tedious business, and
is not likely to be adopted as a routine analysis.
However, computerized image analysis may easily
be applied to a two dimensional analysis, and the
two and three dimensional form parameters were
therefore compared.
CHEMICAL PARAMETERS
The only chemical parameter included in the
present study is the amount of Si02. Data were
obtained from various published and unpublished
sources (Table II). The chemical data set has at least
two drawbacks. Firstly, some of the analyses
represent whole rcck analyses while others represent
microprobe analyses of glass particles. Secondly,
the chemical analyses were not carried out on the
same samples that were measured for form, so that if
a given tephra unit has a range in chemical composi-
tion, it is not known if the Si02 value applies to the
form sample. This is true for all the samples except
the H4 ones. These drawbacks should certainly be
avoided in a detail analysis of individual tephra
units. They are not, however, considered serious
with respect to the objectives of the present study as
the range in silicity between samples is much larger
than the suspected range within most of the tephra
units.
*Surtsey
Fig. 3. The location of sampled volcanoes/tephra
layers in Iceland.
Mynd 3. Gosstöðvar og jarðvegssnið, sem sýnin eru
œttuð úr.
ORIGIN OF THE TEPHRA SAMPLES
The sampling programme was designed to cover
tephra from a wide range of chemical and physical
eruption types from various Icelandic volcanic sys-
tems. A brief description of each eruption and the
volcano in question will be given below.
Öræfajökull is a major eruptive centre at the
southem margin of Vatnajökull (Fig. 3). It has been
built up during the Pleistocene in successive mag-
matic and phreatomagmatic eruptions (Thorarins-
son, 1958). Rhyolitic intrusions occur in the vol-
canic edifice, and one of them forms the highest
peak in Iceland, Hvannadalshnúkur (2119 m). Ör-
æfajökull lies outside the active zone of spreading
and volcanism in Eastem Iceland, but it is the larg-
est active volcano in Iceland. The tephra sample 942
(Rjúpnafell/Hólmsárbrú) is from the AD 1362
eruption which has been investigated by Thorarins-
son (1958), who mapped the rhyolitic tephra layer in
soil sections. According to him, some 10 km3 of
rhyolitic tephra were produced in the 1362 eruption.
This is in excess of any other volcanic eruption in
Iceland in historical time. The top of the Öræfajök-
ull volcano is covered by an ice cap, and the emp-
tion began with an explosive phase. The Si02 value
in Table II is from Sigurdsson (1982).
62 JÖKULL, No. 39, 1989