Download Depth segmentation of the seismogenic

GEOPHYSICAL RESEARCH LETTERS, VOL. 38, L06305, doi:10.1029/2011GL046897, 2011
Depth segmentation of the seismogenic continental crust:
The 2008 and 2009 Qaidam earthquakes
J. R. Elliott,1 B. Parsons,1 J. A. Jackson,2 X. Shan,3 R. A. Sloan,2 and R. T. Walker1
Received 26 January 2011; accepted 2 February 2011; published 24 March 2011.
[1] The seismic hazard in the immediate vicinity of an
earthquake is usually assumed to be reduced after rupture
of a continental fault, with along‐strike portions being
brought closer to failure and aftershocks being significantly
smaller. This period of reduced hazard will persist as strain
re‐accumulates over decades or centuries. However, this is
only realised if the entire seismogenic layer ruptured in
the event. Here we use satellite radar measurements to show
the ruptures of two Mw 6.3 earthquakes, occurring in almost
the same epicentral location ten months apart in the Qaidam
region, China, were nearly coplanar. The 2008 earthquake
ruptured the lower half of the seismogenic layer, the 2009
event the upper half. Fault segmentation with depth allows
a significant seismic hazard to remain even after a moderate
and potentially devastating earthquake. This depth segmentation possibly exists in the case of the 2003 Bam earthquake where satellite radar and aftershock measurements
showed that it ruptured only the upper half of the 15–20 km
deep seismogenic region [Jackson et al., 2006], and that the
lower, unruptured part may remain as a continuing seismic
hazard. Citation: Elliott, J. R., B. Parsons, J. A. Jackson, X. Shan,
remain close to failure. InSAR measurements for the Qaidam
earthquakes studied here provide the first geodetic constraints on the vertical partitioning of fault rupture within the
continental seismogenic layer for two similarly large, but
temporally separated events.
2. InSAR Observations and Modelling
1
COMET+, Department of Earth Sciences, University of Oxford,
Oxford, UK.
2
COMET+, Bullard Laboratories, Department of Earth Sciences,
University of Cambridge, Cambridge, UK.
3
State Key Laboratory of Earthquake Dynamics, Institute of
Geology, China Earthquake Administration, Beijing, China.
[4] We measure the coseismic displacement field due to
each earthquake separately using a combination of SAR
acquisitions spanning each event from the Envisat satellite
(Table S1 in Text S1 in the auxiliary material).1 For each
earthquake, we constructed radar interferograms for two
look directions on ascending and descending satellite passes.
The methods used are by now standard; an outline is given
in the auxiliary material. We derived fault and slip models
using both look directions simultaneously, but only present
the observed and model interferograms for the descending
pass here (Figure 2; images of the observed and model
interferograms, and residuals for both ascending and descending directions are included in Figures S2 and S3 in
Text S1). The interferograms show peak line‐of‐sight displacements (motion towards the satellite) of 8 and 30 cm for
the two events. The different characteristics of the interferograms and the relationship between them provide strong
pointers to the final model solution. The interferogram for
the 2008 event is much smoother, with smaller amplitudes,
suggesting it has a deeper source, and lies to the south of the
interferometric fringes for the 2009 earthquake. The interferogram for the latter event shows a larger, more complex
pattern of surface displacements, with shorter wavelength
features suggesting a shallower source given the similar
moment magnitudes of both events (Figure 2).
[5] To determine the sub‐surface fault orientation and
extent in each case, we invert the InSAR displacements for
the fault geometry assuming a rectangular uniform‐slip
distribution [Okada, 1985; Wright et al., 1999] (see methods
in auxiliary material). The best fitting model interferograms
for the two events are shown in Figure 2 (and Figures S2
and S3 in Text S1), with the fault parameters given in
Table 1. The 2008 event was deeply buried with a centroid
depth of 16–18 km. Fault dislocation models from inverting
this data‐set are consistent with a thrust dipping 58° SSW. It
is possible to model the InSAR data with a shallow northward dipping thrust fault (22°) by fixing the strike to the
conjugate fault plane (248°) given by the seismological
solutions, but the RMS in fit to the InSAR data increases
from 4.0 mm to 4.5 mm, and the centroid depth shallows to
15 km, less than body wave solution of 18 km. (Figure S4
in Text S1 shows the model and residuals for this case).
Copyright 2011 by the American Geophysical Union.
0094‐8276/11/2011GL046897
1
Auxiliary materials are available in the HTML. doi:10.1029/
2011GL046897.
R. A. Sloan, and R. T. Walker (2011), Depth segmentation of the
seismogenic continental crust: The 2008 and 2009 Qaidam earthquakes, Geo phys. Res. Lett., 38, L06305, doi:10 .1029/
2011GL046897.
1. Introduction
[2] The Mw 6.3 10th November 2008 and Mw 6.3 28th
August 2009 thrust earthquakes occurred in the North
Qaidam thrust system, south of the Qilian Shan/Nan Shan
thrust belt and on the northern margin of the Qaidam basin, NE
Tibet (Figure 1). This fold‐and‐thrust belt is the result of the
ongoing northward convergence of India with Eurasia [Chen
et al., 1999], with the rate of NE–SW convergence across it of
approximately 10 mm/yr (Figure 1) [Gan et al., 2007].
[3] Earthquakes in the continental crust occur by slip on
faults in the seismogenic layer above depths of ∼15–20 km
[e.g., Scholz, 2002]. Elastic stresses arise due to the accumulation of strain produced by the relative motion of crustal
blocks on either side of the fault. If, however, when an
earthquake occurs on the fault, it does not rupture right
through the seismogenic layer, then part of the fault may
L06305
1 of 6
L06305
ELLIOTT ET AL.: CONTINENTAL CRUST DEPTH SEGMENTATION
L06305
Figure 1. (a) Topographic map of the Qaidam Basin region, north–eastern Tibet. The white rectangle delineates the area of
faulting shown in Figure 1b. The 2008 and 2009 earthquakes occurred on the north–eastern edge of the Qaidam Basin south
of the Qilian Shan/Nan Shan mountains. Earthquake focal mechanisms are shown from body wave modelling in this paper
(red) and for Mw 5.5+ events in the period 1976–2009 from the GCMT catalogue (grey). GPS vectors are from Gan et al.
[2007] and are relative to a fixed stable Eurasia. Major faults are from Taylor and Yin [2009]. (b) ASTER satellite mosaic
image (RGB bands 321) of the Qinghai region. Mapped active thrust faults are denoted by solid triangles in the hanging
wall. The axes (dashed lines) of anticlines (open diamonds) and synclines (open crosses) are inferred from the structure. A
south‐facing scarp in alluvial fans, which we interpret as secondary faulting above the buried south‐dipping 2009 earthquake fault, is marked by a white line with the uplifted side denoted by +. (The image without annotations is included as
Figure S1 in Text S1).
2 of 6
L06305
L06305
ELLIOTT ET AL.: CONTINENTAL CRUST DEPTH SEGMENTATION
Figure 2. Observed and model interferograms for the (a and b) 2008 and (c and d) 2009 earthquakes. The observed interferograms were derived using Envisat ASAR data on descending track 319 for the dates given in Table S1 in Text S1. Each
colour cycle from blue through yellow to red equals 4 cm of ground displacement away from the satellite. The surface
projections of the fault planes at depth are outlined by the black rectangles and the up‐dip projection of the faults to the
surface by the thick black lines.
Table 1. The 2008 and 2009 Fault Parameters as Derived From the InSAR Inversion and Seismological Body Wave Modellinga
Strike
(deg)
Dip
(deg)
Rake
(deg)
Slip
(m)
Lon
(deg)
99
±1.5
103
58
±2.0
60
95
±4.0
109
0.6
±0.26
0.6
95.859
±0.5 km
67
±1.6
53
±0.6
45
±1.1
72
58
±1.5
106
±1.3
97
±1.9
89
0.7b
fixed
1.0b
fixed
0.6b
fixed
0.6
95.745c
fixed
95.811c
fixed
95.941c
fixed
BW
122d
fixed
100
±0.3
104
±0.8
111
BW
112
56
90
0.3
Model
InSAR
BW
InSARa
InSARb
InSARc
Lat
(deg)
Length
(km)
Top
(km)
Bottom
(km)
Centroid
(km)
Moment
(1018 Nm)
10th November 2008
37.657
15.6
12.1
±0.4 km
±1.4
±3.3
11.5
11.5
11.3
± 1.1
13
21.6
±1.7
23
16.4
±0.5
18
3.6
±0.2
2.3
6.3
28th August 2009
37.690c
7.0d
fixed
fixed
37.653c
10.2
fixed
±0.1
37.645c
6.3
fixed
±0.2
12
9.7
±0.4
5.9
±0.1
6.7
±0.2
12
3.0
±0.04
2.4
±0.02
1.9
±0.04
0
11.8
±0.4
7.0
±0.1
6.6
±0.3
11
7.4
±0.2
4.7
±0.06
4.3
±0.14
5
1.5
±0.06
1.9
±0.04
0.8
±0.03
2.6
6.1
6.4
1.3
6.6
4
0.4
5.7
31st August 2009
6.4
a
Width
(km)
Mw
6.1
6.1
5.9
6.2
For the seismological body‐wave solutions, the slip, length, top and bottom depth are estimated by calculating a source dimension from the centroid
depth and moment. This assumes a square fault with a fixed slip‐to‐length ratio of 5 × 10−5. Therefore, the length scales as the cube root of the
moment, M0 = 5 × 10−5 mL3, where m is the rigidity equal to 3.23 × 1010 Pa.
b
The slip for each fault segment is fixed in the InSAR inversion to allow it to converge to a sensible solution. The choice of slip amount is scaled to the
fault length Scholz, 2002.
c
For the multi‐segmented 2009 rupture, the surface projections of each fault were fixed in the inversion, based upon the pattern of surface‐deformation
observed in the interferometric fringes.
d
The fault strike and length for the most westerly segment was also fixed due to the lack of data in this region.
3 of 6
L06305
ELLIOTT ET AL.: CONTINENTAL CRUST DEPTH SEGMENTATION
Furthermore, this latter solution has a fault plane striking
WSW at a high angle to the mapped thrusts in the region
which generally strike SE–NW to E–W (Figure 1). Conversely, the steeper SSW dipping modelled fault is consistent with the strike of faulting in the region and the overall
NE–SW convergence shown by the GPS, so it is our preferred interpretation. A dip angle of 50–60° is at the upper
end of dips for thrust mechanisms [Sibson and Xie, 1998],
but the unambiguously steep dip for the shallower 2009
event confirms that the conditions for high‐angle thrusts
exist in this area.
[6] The 2009 event has three peaks in coseismic surface
deformation which are larger and much more localised then
for the 2008 event. Given that each earthquake was of
approximately the same magnitude, this indicates a much
shallower event. The surface displacements require three
separate fault segments also dipping to the SSW to model
this pattern of surface deformation (Figure 3). A single fault
could not match the curvature of the deformation pattern for
the western portion, nor the northwards 2–3 km offset in
deformation seen in the east. Due to the many unknowns in
the fault parameter inversion from the use of three segments,
it was necessary to fix their location a priori using the
InSAR fringe pattern. Furthermore, the slip was scaled to the
fault length [Scholz, 2002] to further reduce the number of
unknowns and permit the inversion to converge on a stable
solution. The depth interval of slip found in the fault models
is confirmed to be much shallower, with slip occurring from
2 km down to 7–11 km. The model gives a good fit to the
InSAR data (Figure S3 in Text S1) with an RMS value of
1.1 cm. The near‐field, fault related residuals can be reduced
slightly to 0.9 cm RMS using a finite fault model (Figure S9
in Text S1).
[7] Both results are in agreement with the fault parameters
found from the body‐wave seismological solution (Table 1),
except for the seismological estimate of dip for the 2009
event which is too high an angle for a pure thrusting mechanism. The centroid depths are given as 18 km and 5 km for
the 2008 and 2009 events, respectively. Given that some of
the fault parameters are fixed for the multi‐segmented 2009
event, the errors in Table 1 may be an underestimate due to the
assumptions required to place the fault location and constrain
the slip to be proportional to the fault length, as this reduces
the trade‐offs between parameters.
[8] We performed a Coulomb stress change calculation
[e.g., Lin and Stein, 2004] due to the slip modelled in 2008
event (Figure 3). For faults in the orientation modelled as
that for the 2009 central segment, the Coulomb stress at the
base of this segment (8 km) due to the 2008 event is shown
to increase by over 0.5 MPa (taking the low‐angle conjugate
thrust plane instead would cause a much smaller stress
increase < 0.1 MPa (Figure S8 in Text S1)).
[9] The largest aftershock, three days after the 2009
earthquake, was a Mw 5.7 event; due to its size relative to the
mainshock (Mw 6.3) we neglected it in modelling the surface
displacements, although tentatively it could be attributed to
the easternmost fault segment based upon the fault depth
and orientation.
3. Discussion
[10] The location, dips and depths of faulting in each
Qaidam earthquake are well constrained (Table 1). The
L06305
faults lie beneath the Olunbuluk Shan [Fang et al., 2007],
though neither earthquake broke the surface. We have
mapped active faults bounding the mountains to both the
south and north on ASTER satellite imagery (Figure 1) that
clearly would produce upward relative vertical movement of
the mountains. In cross‐section it is possible to see that two
of the fault segments of the later 2009 rupture lie almost
exactly up‐dip of the deeper 2008 event (Figure 3) and that
the calculated dips are in close agreement. These fault
segments are along strike from the active thrust fault
bounding the northern side of the Olunbuluk Shan further to
the east. There is also an active thrust bounding the mountains to the south. The termination of slip in the 2008 event
at 10 km and the subsequent slip in the top part of the
crust in 2009 may have been a stochastic process, but we
hypothesise that the break in rupture between the two
earthquakes can be explained if this NNE dipping thrust
fault intercepts and offsets the two earthquake rupture planes
at 8–12 km (precisely where depends on the assumed dip of
the fault). The two fault segments can be modelled as precisely coplanar by fixing the location of the deeper event in
the modelling, but the total RMS in the InSAR increases
from 4.0 mm to 4.4 mm, indicating a small offset in the
faults of 1 km is possible, despite this geometry not being
stable in the long term.
[11] This fault geometry has parallels with that for the
2003 Mw 6.6 Bam earthquake in SE Iran. InSAR observations of the Bam earthquake indicated that slip occurred
down to 10 km, whilst most seismically recorded aftershocks were in the deeper range of10–17 km [Jackson et al.,
2006]. In the Bam area, there is a fault with a thrust component visible at the surface that may well join, at a depth
that marks the top limit of the intense aftershock zone, with
the base of the buried strike‐slip fault on which most of the
rupture occurred [Funning et al., 2005; Jackson et al.,
2006].
[12] Elsewhere in Iran, the rupture of the Qeshm Island,
Iran, Mw 6.0 earthquake in 2005 was constrained by InSAR
to the upper 10 km of sedimentary cover [Nissen et al.,
2010] and a series of Mw 4.9–5.7 earthquakes in 2006 at
Fin in the Zagros ruptured the lower sedimentary cover to a
depth of 5–10 km [Roustaei et al., 2010]. In both the latter
cases, aftershocks occurred in the depth range of 10 to
20 km, in the upper part of basement, and it is likely that
faulting is segmented into two competent layers at different
depths separated by weaker lithologies of salt and marls.
[13] If we assume that the aftershocks occur where seismogenic crust is brought closer to failure [King et al., 1994],
the obvious question is whether or not the unruptured part
of the seismogenic crust will fail subsequently in a large
earthquake. An example of delayed segmented rupture of
the seismogenic crust is thought to have occurred on the
Gowk Fault, Iran. The Mw 7.1 1981 Sirch earthquake
[Berberian et al., 1984] produced little surface offset on the
ground trace of the Gowk Fault, and occurred on a deeper
part of the fault system (with a centroid depth of 18 km).
Seventeen years later, the Mw 6.6 1998 Fandoqa earthquake
occurred at a much shallower depth (centroid depth 5 km)
on the same section of fault, and was associated with a much
larger surface rupture [Berberian et al., 2001]. Until now,
however, there has been no example of successive earthquakes in the same locality where the capability of satellite
radar interferometry to provide precise measurements of
4 of 6
L06305
ELLIOTT ET AL.: CONTINENTAL CRUST DEPTH SEGMENTATION
L06305
Figure 3. (a) Landsat (RGB 742) map view of the fault locations for the 2008 (white) and 2009 (black) earthquakes from
the InSAR analysis. The focal mechanisms are the body wave seismology solutions for each event. The barbed lines
represent the surface projections of the faults. Red lines denote thrust faults mapped from the ASTER satellite imagery.
(b) Cross section of InSAR determined fault surfaces and projected seismological focal mechanisms for the two earthquakes.
The central and eastern portions of the 2009 rupture lie up‐dip of the region of faulting modelled for the 2008 earthquake.
The white and black lines indicate the 2008 and 2009 central segment faults respectively. The red dashed lines indicate (for
a range of dip angles shown) the down‐dip locations from the mapped surface trace of the main north‐dipping thrust. Over
the dip range 45–60°, the projected fault trace intersects the break in fault rupture between the 2008 and 2009 events. This
geometry is not stable as offset along the fault cannot continue for very long where the faults intersect and must indicate an
evolving fault system. The sharp range front suggests the NNE dipping fault dominates and a small reverse offset of the
SSW dipping fault segments is consistent with this. (c) Coulomb stress change at 8 km depth due to the 2008 event (for
faults orientated as the 2009 central segment). The 2009 rupture lies within the region of positive Coulomb Stress change.
(d) Cross section through X−X′ indicating the Coulomb stress change with depth.
fault location, geometry and slip, could be brought to bear
on the relationship between the two events.
[14] The radar measurements presented here provide
strong geodetic evidence of what had been hinted at by the
observations described above, namely that, where either
fault geometry or lithological properties allow only part of
the seismogenic layer to rupture, the occurrence of a large
earthquake does not necessarily result in a reduction of the
immediate seismic hazard. In the case of the Qaidam
earthquakes, the segmentation of the seismogenic crust with
depth appears to be due to fault geometry; offsets on the
fault plane may have prevented the deeper earthquake rup-
5 of 6
L06305
ELLIOTT ET AL.: CONTINENTAL CRUST DEPTH SEGMENTATION
ture from immediately propagating nearer to the surface.
However, the deeper earthquake will have increased the
Coulomb stress on the shallower segment, taking it closer to
failure [King et al., 1994]. Segmentation with depth like this
allows the possibility that an earthquake of similar magnitude can happen beneath almost the same location within a
short interval of a large event.
[15] Such a possibility has significant implications for the
interpretation of the paleoseismological record and the
assessment of seismic hazard. The deeper earthquakes will
have a relatively small surface effect and a smaller influence
on surface sedimentation, e.g. surface offsets for the deeper
Mw 7.1 Sirch earthquake were negligible compared to those
of the shallower Mw 6.6 Fandoqa earthquake. The mean
repeat time of earthquakes, summed across the depth range
of the seismogenic portion of the crust and therefore
affecting a given area at the surface, is therefore likely to be
overestimated from the paleoseismological record. In addition, it might appear that lack of knowledge of fault
geometry and lithology, particularly in areas where faults
are young and growing, considerably complicates seismic
hazard assessment. Nonetheless, although the time separation between successive up‐dip or down‐dip ruptures is
likely to be highly variable — that for 2008 and 2009
earthquakes in the Qaidam basin was 10 months, that for the
1981 Sirch and 1998 Fandoqa earthquakes on the Gowk
fault in Iran was 17 years — geodetic mapping of the spatial
distribution of strain release in earthquakes as they occur
can provide a means of updating seismic hazard in the short
term. For instance, if it had been possible at the time to
accurately determine the depth distribution of slip for the
1981 Sirch earthquake, it would have been clear that, in the
absence of any historical evidence for an earlier, recent
earthquake in that locality, there was a strong probability of
an earthquake occurring at shallower depths on fault segments visible at the surface.
[16] Acknowledgments. This work is supported by the Natural Environment Research Council through the National Centre for Earth Observation, of which the Centre for the Observation and Modelling of
Earthquakes, Volcanoes and Tectonics (COMET+) is part. We are grateful
to JPL/Caltech for use of the ROI_PAC software. All figures were made
using the public domain Generic Mapping Tools, [Wessel and Smith,
1998].
[17] The Editor thanks Garth Funning and an anonymous reviewer.
References
Berberian, M., J. A. Jackson, M. Ghorashi, and M. H. Kadjar (1984), Field
and teleseismic observations of the 1981 Golbaf‐Sirch earthquakes in SE
Iran, Geophys. J. Int., 77, 809–838, doi:10.1111/j.1365-246X.1984.
tb02223.x.
Berberian, M., J. A. Jackson, E. Fielding, B. E. Parsons, K. Priestley,
M. Qorashi, M. Talebian, R. Walker, T. J. Wright, and C. Baker
(2001), The 1998 March 14 Fandoqa earthquake (Mw 6.6) in Kerman
province, southeast Iran: Re‐rupture of the 1981 Sirch earthquake fault,
L06305
triggering of slip on adjacent thrusts and the active tectonics of the Gowk
fault zone, Geophys. J. Int., 146, 371–398, doi:10.1046/j.1365-246X.
2001.01459.x.
Chen, W., C.‐Y. Chen, and J. Nabelek (1999), Present‐day deformation
of the Qaidam basin with implications for intra‐continental tectonics,
Tectonophysics, 305, 165–181, doi:10.1016/S0040-1951(99)00006-2.
Fang, X., W. Zhang, Q. Meng, J. Gao, X. Wang, J. King, C. Song, S. Dai,
and Y. Miao (2007), High‐resolution magnetostratigraphy of the
Neogene Huaitoutala section in the eastern Qaidam Basin on the NE
Tibetan Plateau, Qinghai Province, China and its implication on tectonic
uplift of the NE Tibetan Plateau, Earth Planet. Sci. Lett., 258, 293–306,
doi:10.1016/j.epsl.2007.03.042.
Funning, G. J., B. Parsons, T. J. Wright, J. A. Jackson, and E. J. Fielding
(2005), Surface displacements and source parameters of the 2003 Bam
(Iran) earthquake from Envisat advanced synthetic aperture radar imagery,
J. Geophys. Res., 110, B09406, doi:10.1029/2004JB003338.
Gan, W., P. Zhang, Z.‐K. Shen, Z. Niu, M. Wang, Y. Wan, D. Zhou, and
J. Cheng (2007), Present‐day crustal motion within the Tibetan Plateau
inferred from GPS measurements, J. Geophys. Res., 112, B08416,
doi:10.1029/2005JB004120.
Jackson, J., et al. (2006), Seismotectonic, rupture process, and earthquake‐
hazard aspects of the 2003 December 26 Bam, Iran, earthquake, Geophys.
J. Int., 166, 1270–1292, doi:10.1111/j.1365-246X.2006.03056.x.
King, G. C. P., R. S. Stein, and J. Lin (1994), Static stress changes and the
triggering of earthquakes, Bull. Seismol. Soc. Am., 84(3), 935–953.
Lin, J., and R. S. Stein (2004), Stress triggering in thrust and subduction
earthquakes and stress interaction between the southern San Andreas
and nearby thrust and strike‐slip faults, J. Geophys. Res., 109,
B02303, doi:10.1029/2003JB002607.
Nissen, E., F. Yamini‐Fard, M. Tatar, A. Gholamzadeh, E. Bergman,
J. R. Elliott, J. A. Jackson, and B. Parsons (2010), The vertical separation
of mainshock rupture and microseismicity at Qeshm island in the Zagros
fold‐and‐thrust belt, Iran, Earth Planet. Sci. Lett., 296, 181–194,
doi:10.1016/j.epsl.2010.04.049.
Okada, Y. (1985), Surface deformation due to shear and tensile faults in a
half‐space, Bull. Seismol. Soc. Am., 75(4), 1135–1154.
Roustaei, M., E. Nissen, M. Abbassi, A. Gholamzadeh, M. Ghorashi,
M. Tatar, F. Yamini‐Fard, E. Bergman, J. Jackson, and B. Parsons
(2010), The 2006 March 25 Fin earthquakes (Iran)‐insights into the vertical extents of faulting in the Zagros Simply Folded Belt, Geophys. J.
Int., 181, 1275–1291, doi:10.1111/j.1365-246X.2010.04601.x.
Scholz, C. H. (2002), The Mechanics of Earthquakes and Faulting,
Cambridge Univ. Press, Cambridge, U. K.
Sibson, R. H., and G. Xie (1998), Dip range for intracontinental reverse
fault ruptures: Truth not stranger than friction?, Bull. Seismol. Soc.
Am., 88(4), 1014–1022.
Taylor, M., and A. Yin (2009), Active structures of the Himalayan‐Tibetan
orogen and their relationships to earthquake distribution, contemporary
strain field, and Cenozoic volcanism, Geosphere, 5(3), 199–214,
doi:10.1130/GES00217.1.
Wessel, P., and W. H. F. Smith (1998), New, improved version of generic
mapping tools released, Eos Trans. AGU, 79, 579–579, doi:10.1029/
98EO00426.
Wright, T. J., B. E. Parsons, J. A. Jackson, M. Haynes, E. J. Fielding,
P. C. England, and P. J. Clarke (1999), Source parameters of the 1 October
1995 Dinar (Turkey) earthquake from SAR interferometry and seismic
bodywave modelling, Earth Planet. Sci. Lett., 172, 23–37, doi:10.1016/
S0012-821X(99)00186-7.
J. R. Elliott, B. Parsons, and R. T. Walker, COMET+, Department of
Earth Sciences, University of Oxford, Parks Road, Oxford OX1 3PR,
UK. ([email protected])
J. A. Jackson and R. A. Sloan, COMET+, Bullard Laboratories,
Department of Earth Sciences, University of Cambridge, Madingley
Road, Cambridge CB3 0EZ, UK.
X. Shan, State Key Laboratory of Earthquake Dynamics, Institute of
Geology, China Earthquake Administration, Beijing 100029, China.
6 of 6