International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
a random error and, as such, is impossible to quantify at any
given point. To reduce its effects, additive noise/speckle is
depressed by multi-look SAR image processing. The third,
baseline effects on correlation, is reduced by removing the flat
phase. Concerning on topographic effects, although average
altitude of the study area is high, but the relative height is
within 600m, the layover and foreshortening appearances for
existing steep hills on SAR images are relatively few. So the
large, extent of decorrelation area on interferogram may be
attributed to temporal decorrelation, but it is impossible. Since
that the study area is above 4800m altitude with rare vegetation
cover and consists mainly of a series of Gobi and monadnocks,
the backscattering variation between imaging dates should be
less difference.
Figure 2. The changing interferogram of Mani Earthquake,
Tibet on Nov. 8, 1997
According to above analysis of influencing effects, we can
conclude that the fringe disturbance around the Margaichace-
Ruolacuo fault zone exclusively attribute to effects for surface
deformation, which is exceeded the range SAR interferometry
can measure. A step further, we infer from the differential
InSAR interferogram that the slips or block offsets, along
secondary fractures within the main rupture around the
Margaichace-Ruolacuo fault zone, bring in the fringe
disturbance and displacement field decorrelation. In summary,
analyzing the resultant deformation interferogram, we draw a
conclusion that the area around the Margaichace-Ruolacuo fault
is the ultimately possible surface rupture zone and is the most
severe deformation area; epicenter also is located in this zone.
The fringe pattern around the Margaichace-Ruolacuo : main
rupture is dense and elongated parallel to the azimuth of the
rupture. The density of fringes decreases logarithmically across
the rupture from near to faraway as is expected from elastic
“half-space modeling of the lithosphere. Because of asymmetric
fringes distribution along the two sides of the fault, this pattern
is indicative of a left-lateral horizontal shear in the direction of
earthquake rupture (Peltzer, G. and P. Rosen, 1995). If no
vertical deformation and only horizontal shifts parallel to the
azimuth of the rupture are assumed, this evidence, combined
with the viewing geometry mentioned above, indicates that the
ratio of deformation displacements along radar viewing
direction to the presumed direction, is equal to
singsin@ = 0.32 , where 0 « 23? is the radar incidence and
@ = 10° is the angle between the nadir trace of radar and the
presumed deformation offset direction. So we can obtain the
horizontal offsets using the deformation displacement along the
radar viewing direction by dividing the value of 0.32. Under the
presumption, if the relative displacement of horizontal offset
between two neighboring pixel exceeds & = 2.8/0.32 = 8.75 cm,
the displacement along radar viewing direction will exceeds the
range one fringe cycle can measure on differential
interferogram, and fringe disturbance will be emergent. On the
resultant interferogram of this study, the southern side of the
fault from disturbance area outward exists at least 32 fringes,
and the northern side of the fault from disturbance area outward
exists at least 20 fringes. We can infer that the horizontal offset
displacement for the southern side at least reaches 2.8m and
1.75m at least for the northern side of the fault. Apparently, the
deformation displacement in the southern side is more severe
‘ than that in the northern side. But it should be pointed that
horizontal left-lateral shear in the direction of earthquake
rupture is the whole surface deformation form, as to some local
points there maybe exist vertical deformation components, and
also maybe superimpose the later shifts. For example, near the
epicenter area maybe superimpose subsidence after earthquake
because the southern fringes of epicenter show evidently curves
of arc.
Field observations by Xu Xiwei (China National Earthquake
Agency's Staff, 2000) manifest that the surface rupture zone of
Mani earthquake in uninhabited area of northern Tibet lies in
the northern boundary around the Margaichace-Ruolacuo fault
fracture zone. The nearly vertical rupture zone is the seismic
fault of Mani earthquake, trending NE-E direction, extending
120km long and 300-400m in width. The nature of fault
movement is dominant by left-rotary shear with the maximum
displacement 4.5m and the horizontal shift along its east and
west sides is about 2-3m. This agrees with the mapping results
obtained by analyzing the differential interferogram around
Mani earthquake area.
Another research by Xu Lishen et al. (1999) also points that the
seismic fault of Mani earthquake is a left-rotation reverse fault
rupturing from west to east, trending 250°direction, and the tilt
angle is nearly vertical, about 88° (Xu Lisheng and Chen
Yuntai,1997). Their research concludes that Mani earthquake,
on the whole, is characterized by single-side rupturing. All
these conclusions, at some extent, agree with the analysis
results from differential interferogram.
4. CONCLUSIONS
With ERS-1/2 SAR data set, we use differential SAR
interferometry to capture the co-seismic displacement field
produced by the Mani earthquake occurred on Nov. 8, 1997 in
Tibet, China. From the changing interferogram, we concluded
that the zone around Margaichace-Ruolacuo fault is the most
severely deformed and the most possibly rupturing area. Also
we concluded from analyzing the fringe patterns that left-lateral
shear movement is the whole deforming mechanism and
furthermore the offsets are also quantitatively estimated. The
734
Internat
observe
satellite
displace
elastic d
With su
different
co-seisim
on the o
earthqua
topograr
different
along ra
displacei
slant d
deforma
deforma
fringe p
there is
Mani ea
discrepai
displacer
lithosphe
Radar c
interest |
point to
is grandi
studies.
to monit
with GP!
et al. rer
techniqu
interfero
radar ren
Feng Ha
quake O
Develop
Gabriel
Mapping
radar inte
Gens R,
technique
1803-18.
Goldsteit
radar in
Radio Sc
Kimura
Ground
pp.442-4
Li Jianhu
neighbor
Seismolo
Peltzer, (
May 199
SAR inte
