Jökull - 01.01.2016, Side 53

The marginal zone of Tungnaárjökull, Iceland
set against the dynamics of Tungnaárjökull was un-
dertaken by the authors during several research expe-
ditions to the forefield of the glacier between 1995
and 2005 (e.g. Molewski, 2005). The main objective
of the expeditions was to create a geomorphological
map of the entire forefield of Tungnaárjökull. Due to
limited time and accessibility, individual landforms
and sediments were identified and studied to different
degrees. The geomorphological study resulted in the
first identification of the morphogenesis of the anal-
ysed area. Previous research on the glacial geomor-
phology of the forefield of Tungnaárjökull includes
work by Andrzejewski and Molewski (1999, 2000) in
the central and southern parts of the marginal zone of
the glacier which focused upon the tectonics of glacial
and glacifluvial deposits resulting from glacial surges.
Later work by Andrzejewski (2002) identified five
morphogenetically different areas within the forefield
of Tungnaárjökull which he related to the glacier’s
varied dynamics, which are mainly conditioned by the
topography of the underlying volcanic bedrock. A ge-
omorphological map of the forefield of Tungnaárjök-
ull was published by Molewski (2005) and Evans et
al. (2009). Evans et al. (2009) also described the diag-
nostic qualities of landforms and deposits associated
with glacier surges.
The objective of this study is to reconstruct
palaeogeomorphological events in the northern part
of the marginal zone of Tungnaárjökull since the end
of the LIA, based on geomorphological mapping and
sedimentological analysis, as well as published his-
torical maps and aerial photographs of the region
(Freysteinsson, 1972; Thorarinsson and Sigvaldason,
1972). The research presented here focuses upon two
key areas: one located in the northern part of the
forefield; and the other, a smaller area including a
part of the Jökulgrindur bedrock ridge and adjacent
glacigenic sediments (Figure 1B).
METHODS
Geomorphological field mapping of the Tungnaárjök-
ull forefield was carried out at a scale of 1:30,000
using aerial photographs from 1986 and 1995. The ge-
ographical position of the sites examined in the field
were determined using a standard GPS receiver and
calibrated aerial photographs. The geometric levelling
of landforms was conducted using a clinometer.
Sedimentological and lithofacies analysis of the
glacilacustrine and glacifluvial sediments was carried
out upon exposed natural sections through these de-
posits. Particular emphasis was placed on recording
sediment type and the range of sedimentary struc-
tures present, as well as the nature of bedding and
lithological contacts, enabling a series of lithofacies
to be established. The lithofacies code used in gen-
eral follows that proposed by Miall (1978). The ori-
entation of ripple cross-lamination, low-angle cross-
stratification and inclined stratification surfaces was
recorded using a compass clinometer to provide a
record of palaeocurrent directions active during the
deposition of the glacigenic sequence.
In order to establish the most common transition
between pairs of adjacent lithofacies and to detect
the deterministic and random components of that suc-
cession (i.e. the transition probability), Markov chain
analysis (e.g. Miall, 1973; Gradziński et al., 1986)
were conducted on the longest lithofacies sequence,
i.e. from the kame terrace. Markov chain analysis is a
technique which can be used to determine the proba-
bility of one state transitioning into another in con-
texts where it is assumed that states are ’memory-
less’, i.e. that the probability of the subsequent state
depends only on the present state and not on any pre-
ceding state or sequence of states. Because the statis-
tical reliability of any resultant transition probabilities
depends on the abundance of any given facies in the
profile, z-tests were used to evaluate the statistical sig-
nificance of the probability of individual transitions.
The critical z-statistic value for 0.05 reliability is 1.64.
It is assumed that transitions for which Z exceeds this
value are preferred by the mechanisms of sedimen-
tation processes; that is to say that sedimentation pro-
cesses result in transitions between layers which are to
some extent predictable or regular. These transitions
are used in the construction of so-called "transition
diagrams" (Gradziński et al., 1986).
Average grain sizes of various deposits were
largely recorded in the field using a comparison grain-
size visual estimation chart. However, 10 bulk sam-
ples collected from the deposits within the kame
JÖKULL No. 66, 2016 53