Jökull - 01.01.2009, Page 95

Slope creep in East Iceland observed by InSAR
e) 20040531−20040809
Sey
ðis
fjör
ður
La
yo
ve
r
~1 km
a) 19950718−19950823
Sey
ðis
fjör
ður
La
yo
ve
r
−10
−5
0
5
10
f) 20050516-20050620
Sey
ðis
fjör
ður
La
yo
ve
r
b) 19970618−19971001
Sey
ðis
fjör
ður
La
yo
ve
r
c) 19980708−19981125
Sey
ðis
fjör
ður
La
yo
ve
r
d) 19990519−19991110
Sey
ðis
fjör
ður
La
yo
ve
r
Atmospheric
artifact
Figure 4. Several interferograms of Þófi in radar coordinates spanning 1–6 month intervals in 1995–1999 and
2004–2005. The black arc marks the upper edge of the Þófi site. – Nokkrar bylgjuvíxlmyndir af Þófanum frá
1995–1999 og 2004–2005. Hver mynd spannar 1–6 mánuði. Efri mörk Þófans eru sýnd með svartri línu.
is significant here is the abrupt and consistent change
in interferometric phase values that can be seen in the
Þófi area, which probably represents ground move-
ment of an area that is only about 2–300 m broad. A
large area located along the coast to the northeast also
shows consistent interferometric phase value shift of
similar magnitude (Figure 4c). However, the smooth
appearance of this signal and its single appearance in
this interferogram strongly suggest that it is of atmo-
spheric origin, but not due to ground displacement.
The location of the small moving area coincides
well with the area in the southwestern-half of the
Þófi site that had been found moving from the GPS
measurements (Jensen and Jóhannesson, 2002), ac-
cording to a geocoded version of this interferogram
(Figure 3a). The geocoding involves a projection
of the interferogram from its radar geometry coordi-
nates to geographical coordinates. The projection re-
quires interpolation onto a very different grid and does
therefore modify the observed data significantly. The
geocoded version of the interferogram has the same
25 m posting as the DEM used in this process and is
therefore somewhat coarser than the original interfer-
ogram. The resulting image shows the location of the
moving area to be on the southwestern side of the Þófi
area and mostly located just above 100 m elevation
(Figure 3a). Neither the DEM nor the geocoding pro-
cedure are very accurate, so the location of the mov-
ing area in the interferogram should be regarded as
only an approximation. However, this location corre-
JÖKULL No. 59 95