Jökull - 01.01.2009, Page 93

Slope creep in East Iceland observed by InSAR
ÞÓFI IN SEYÐISFJÖRÐUR
Seyðisfjörður is one of the Eastern Fjords (Figure 1).
To the west and east of the Seyðisfjörður town are
steep mountains, Bjólfur and Strandartindur, which
both are over 1000 m high (Figure 3a). The Strandar-
tindur side has problems related to active slope-creep,
debris flows, water flooding, and rockfall (Jensen and
Sönser, 2002) that threaten buildings in the eastern
part of the town. One of the problem areas is related
to a shelf or a step on the Strandartindur slope that
is called Þófi. This shelf is located at an elevation of
1–200 m above sea level, directly above large fish-
factory buildings that line the coast. The Þófi shelf
has developed due to intersections of WNW-ESE and
NNE-SSW fractures and it is now mostly covered by
till (Sæmundsson and Pétursson, 1999a; Jensen and
Sönser, 2002). The deposits have been interpreted as
being "fossil" rock-glacier deposits (Guðmundsson et
al., 2003).
An indication of recent mass movements in the
Þófi area was detected in the field in 2000 when fresh
and open ground cracks were found (Jensen and Jó-
hannesson, 2002). This discovery was followed by
an installation of 25 GPS benchmarks in Fall of 2001
to monitor the movement and the first full reoccupa-
tion of the network was carried out in Fall of 2002
(Jensen, 2001; Jensen and Jóhannesson, 2002). Dur-
ing the first year the measurements showed systematic
displacement rates of up to 33 cm/year in the south-
western part of Þófi and the moving area was found to
be larger than expected, or at least 400× 250 m2, and
thus the network was expanded in 2002 (Jensen and
Jóhannesson, 2002). The 2002–2003 measurements
show continuing fast movements in Þófi while later
measurements in 2003–2004 and 2004–2005 indicate
little or no displacements (Tómas Jóhannesson, pers.
comm., 2007).
Wangensteen et al. (2006) used cross-correlation
matching of two digital orthophotos to evaluate sur-
face displacements near Seyðisfjörður town. The pho-
tos were taken in 1964 and 1994 and primarily fo-
cussed on the area southwest of Þófi, an area known
as Botnar, while only a small part of Þófi was cov-
ered. The results show an average displacement rate
of 1–2 cm/year, which is near to the estimated accu-
racy level of 1 cm/year (Wangensteen et al., 2006).
However, the derived displacement vectors are coher-
ent and point downhill and thus seem be revealing
real displacements, although the average velocities are
only a fraction of the fast moving Þófi GPS sites in
2001–2003.
The location of Þófi on the steep northwestern
slopes of Strandartindur mountain means that it can
only be imaged by descending radars that look to-
wards the WNW, while ascending imaging results in
a layover. The surface conditions and the rather low
altitude of 1–200 m make observation at this site rel-
atively favorable and the area retains some degree of
coherence for long time periods of up to a couple of
years. The prior knowledge about the extent of the
Þófi creep from the GPS measurements, which indi-
cates that the moving area is only about 400 m broad,
poses a challenge for InSAR as the resolution of the
radar is about 4 m × 20 m on the ground and usually
approximately square pixels of 20 m× 20 m are used.
This means that the expected size of the Þófi creep is
only a couple of tens of pixels, which requires a care-
ful analysis and does not allow much image filtering.
While interferograms in East Iceland retain a signifi-
cant amount of interferometric correlation over a pe-
riod of several years, filtering of the interferograms
is usually needed to obtain useful information from
longer time-span interferograms. Therefore, due to
the small extent of the moving area in Þófi, only in-
terferograms spanning less than about 6–12 months
proved to be useful.
Interferograms from before 1998 indicate no
movement in the Þófi area. Three 1–2 month inter-
ferograms from the summer of 1995 show no signs
of surface displacement during the time these inter-
ferograms span (see an example in Figure 4a). Two
interferograms, one spanning the year 1996–1997 and
another spanning 3.5 months in 1997 do not either hint
at any displacement in Þófi (Figure 4b).
An interferogram from 1998 spanning 4.5 months
from 8 July to 25 November (Figure 4c) is decorre-
lated at elevations above ∼400 m, presumably due to
snow during the November acquisition. However, al-
though the interferogram is noisy, its filtered version
exhibits a rather coherent phase along the coast. What
JÖKULL No. 59 93