Jökull - 01.01.2009, Page 91
Slope creep in East Iceland observed by InSAR
teract with and reflect from objects on the ground that
have a roughness of a similar dimension. The ERS-1
and ERS-2 radar wavelengths are identical and data
from these two satellites can be combined to form in-
terferograms. The Envisat radar, on the other hand,
operates on a slightly different wavelength so Envisat
data cannot easily be combined with ERS data for in-
terferometry.
Radar imaging is fundamentally different from
conventional passive remote sensing techniques that
typically acquire near-nadir photographs in the visible
or near-visible bands. Radar imaging is an active re-
mote sensing technique where radar pulses are trans-
mitted and the ground reflections of these pulses are
collected again by the same antenna (see e.g. Hanssen
(2001) for a review). The primary advantage of this
technique over other remote-sensing techniques is that
radar imaging is not limited to daylight acquisitions
nor to cloud-free conditions, and in addition it can
be used to detect ground displacement. The ERS-
1/2 satellites cover in each pass a swath that is about
100 km wide and the incidence angle varies from
about 19◦ to about 26◦ across the swath. The Envisat
satellite can be operated in several different modes
with different incidence angles. However, the data
requested for East Iceland were acquired in the IS-2
mode, which is similar to the ERS mode.
Shadow
Layover
23
o
Line-of-sight (LOS) direction
to the satellite
wav
efro
nts
Unconsolidated
Deposits
Figure 2. A schematic figure showing layovers and
shadows in radar imaging. Ground displacements can
only be measured along the line-of-sight (LOS) direc-
tion. – Mynd sem sýnir „álögur“ og skugga í radar-
mælingum í hæðóttu landslagi.
The average incidence angle of 23◦ has limitations
and it means that slopes tilting away from the radar
and are steeper than 67◦ are in a ’shadow’ and can-
not be imaged by the radar satellite (Figure 2). How-
ever, not many slopes are so steep. More important is
the imaging limitation of slopes that incline towards
the radar look direction. When the tilt of these slopes
exceeds 23◦, radar returns from the top of the moun-
tain will arrive at the same time or before radar re-
flections from further down the slope, which makes it
impossible to distinguish between these signals (Fig-
ure 2). This phenomena is called a ’layover’ and is
much more limiting than shadowing, as it excludes
virtually all significant slopes facing the radar look di-
rection. Fortunately, radar satellites can acquire radar
data from approximately opposite directions, i.e. dur-
ing ascending and descending passes, so most slopes
can be imaged using one of these two viewing direc-
tions.
The topography of the Eastern Fjords in Iceland
is relatively rough with steep-sided valleys and fjords
and with mountains exceeding 1000 m in elevation.
A large part of the area consists of slopes exceed-
ing 23◦ and some slopes even exceed 67◦. Therefore,
slopes facing ESE result in a layover during descend-
ing passes and WSW slopes cannot be imaged during
ascending passes.
Thousands of ERS-1 and ERS-2 radar images
have been acquired over Iceland since 1991 with mul-
tiple acquisitions for any given location from several
different tracks and from both the ascending and de-
scending directions. The repeat time of these satellites
is 35 days, which means that they orbit along the same
track every 35 days. The revisit time is shorter, as any
given point on ground can be imaged more frequently,
i.e. from overlapping tracks and from ascending and
descending tracks. Although many acquisitions of
Iceland exist, most of these data have been acquired
during descending passes. Other limitations include
a gap in ERS-1 data from 12/1993 to 4/1995, when
the satellite was operated in different orbits. After the
launch of ERS-2 in 1995 the satellite trailed ERS-1 by
only a day, providing an opportunity to form one-day
(tandem) interferograms that have been extensively
used to generate digital elevation models and to study
JÖKULL No. 59 91