University of Southern California Dissertations and Theses

Landslide inventory associated with the 2008 Wenchuan Earthquake and implications for seismic mountain building

(USC Thesis Other)

Landslide inventory associated with the 2008 Wenchuan Earthquake and implications for seismic mountain building

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

LANDSLIDE INVENTORY ASSOCIATED WITH THE 2008
WENCHUAN EARTHQUAKE AND IMPLICATIONS FOR
SEISMIC MOUNTAIN BUILDING

by

Gen Li

_______________________________________

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fullfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOLOGICAL SCIENCES)

December 2014

1

Acknowledgments

In the first place, I would like to express my gratitude to Dr. Josh West for his
supervision, advice and guidance. Josh‟s inspiration, intuition and passion have made
this work impressing and joyful learning experience. It is a privilege to work with and
learn from Josh. Sincere appreciations are given to Bob Hilton and Alex Densmore,
for generously sharing their ideas and knowledge and warmly hosting me during my
staying in Durham, which was indeed a happy and fruitful time. Great thanks are also
given to our collaborators, Zhangdong Jin, Fei Zhang and Jin Wang, from the Chinese
Academy of Sciences, for assistance and discussion. This research work benefited
from motivating discussion with Niels Hovius and the valuable field dataset
generously provided by Siobhan Whadcoat. I would also like to thank my committee
members, Doug Hammond and Will Berelson, for their advice and encouragement.
Thanks to Cindy Waite for helping me patiently on matters related to the graduate
program. Thank you all my friends here at USC, who have made my life colorful and
unforgettable. Most importantly, I would like to convey my deep gratitude to my
family, for all their love and encouragement.

2

Preamble

This work aimed at constructing a refined co-seismic landslide inventory associated
with the 2008 Wenchuan earthquake, and understanding the volumetric budget
between seismically-induced rock uplift and landslide erosion for earthquakes with
varying magnitudes.

I was the main author, mapped the co-seismic landslides associated with the 2008
Wenchuan earthquake, carried out the earthquake moment magnitude-based scaling
calculations and wrote the manuscript. I also developed algorithms for propagating
landslide volume uncertainties via a Monte Carlo approach and for modeling
landslide reverse-gamma probability distribution. Josh West provided the main
supervision. Specifically, he advised on the scaling calculations and relevant modeling
work, aided in the outline of the manuscript, collected remote imagery and was the
main editor. Alex Densmore helped with the framework of the manuscript and
landslide mapping techniques, and provided a significant portion of remote imagery.
Zhangdong Jin assisted in the field work. Bob Hilton provided general guidance.
Robert Parker assisted in landslide mapping and provided his published Wenchuan
landslide inventory which was done by automated mapping. All authors contributed to
paper revision.

This work was supported by the USC Dornsife College of Letters, Arts and Sciences,
the U.S. National Science Foundation, the Chinese Academy of Sciences, and the U.K.
Royal Society, and benefited from imagery provided by the Polar Geospatial Data
Center and DigitalGlobe. Niels Hovius is thanked for helpful discussion. Siobhan
Whadcoat is thanked for providing the field dataset. We thank two anonymous
referees for their insightful comments which significantly improved the manuscript.

This work was published as:
Li, G., A. J. West, A. L. Densmore, Z. Jin, R. N. Parker, and R. G. Hilton (2014),
Seismic mountain building: Landslides associated with the 2008 Wenchuan
earthquake in the context of a generalized model for earthquake volume balance,
Geochem. Geophys. Geosyst., 15, 833–844, doi:10.1002/2013GC005067.

3

Table of Contents

Acknowledgments 1
Preamble 2
List of Figures 4
List of Tables 5

1. Abstract 6
2. Introduction 8
3. Setting 11
4. Method 12
5. Results 13
6. Discussion 15
7. Conclusion 27
Bibliography 29
Appendix 35
Figures and Tables 50

4

List of Figures

Figure 1 Topographic map of the study area 50
Figure 2 Comparisons of the two landslide inventories 51
Figure 3 Illustration cartoon of a 3-D geometric model 52
Figure 4 Individual earthquake volume balance 53
Figure 5 Cumulative earthquake volume balance 54
Figure 6 Contributions from various earthquakes to the net volume change 55
Figure S1 Illustration cartoon of the mapping techniques 56
Figure S2 Inverse gamma distribution of the landslide inventory 57
Figure S3 Schematic diagram of the landslide clustering effect 58
Figure S4 Quantification of the landslide clustering effect 59
Figure S5 Coordination system for uplift volume calculation 60
Figure S6 Results of landslide volume‟s parameters‟ sensitivity test 61
Figure S7 Results of coseismic volume budget‟s parameters‟ sensitivity test 62

5

List of Tables

Table 1 Reported landslide volumes 63
Table 2 Model parameters 64
Table S1 Used satellite images 66
Table S2 Parameters used in the Appendix 77
Table S3 Results of landslide volumes calculated via multiple methods 79
Table S4 Results of parameters‟ sensitivity test 81

6

1. Abstract
Here we assess earthquake volume balance and the growth of mountains in the context of a
new landslide inventory for the M
w
7.9 Wenchuan earthquake in central China. Coseismic
landslides were mapped from high-resolution remote imagery using an automated algorithm
and manual delineation, which allows us to distinguish clustered landslides that can bias
landslide volume calculations. Employing a power-law landslide area-volume relation, we
find that the volume of landslide-associated mass wasting (~2.8+0.9/-0.7 km
3
) is lower than
previously estimated (~5.7-15.2 km
3
) and comparable to the volume of rock uplift (~2.6± 1.2
km
3
) during the Wenchuan earthquake. If fluvial evacuation removes landslide debris within
the earthquake cycle, then the volume addition from coseismic uplift will be effectively offset
by landslide erosion. If all earthquakes in the region followed this volume budget pattern, the
efficient counteraction of coseismic rock uplift raises a fundamental question about how
earthquakes build mountainous topography. To provide a framework for addressing this
question, we explore a group of scaling relations to assess earthquake volume balance. We
predict coseismic uplift volumes for thrust-fault earthquakes based on geophysical models for
coseismic surface deformation and relations between fault rupture parameters and moment
magnitude, M
w
. By coupling this scaling relation with landslide volume-M
w
scaling, we obtain
an earthquake volume balance relation in terms of moment magnitude M
w
, which is consistent
with the revised Wenchuan landslide volumes and observations from the 1999 Chi-Chi
earthquake in Taiwan. Incorporating the Gutenburg-Richter frequency-M
w
relation, we use
this volume balance to derive an analytical expression for crustal thickening from coseismic
deformation based on an index of seismic intensity over a defined area. This model yields
7

reasonable rates of crustal thickening from coseismic deformation (e.g.~0.1-0.5 km Ma
-1
in
tectonically active convergent settings), and implies that moderate magnitude earthquakes
(M
w
≈6-7) are likely responsible for most of the coseismic contribution to rock uplift, because
of their smaller landslide-associated volume reduction. Our first-order model does not
consider a range of factors (e.g., lithology, climate conditions, epicentral depth and tectonic
setting), nor does it account for viscoelastic effects or isostatic responses to erosion, and there
are important, large uncertainties on the scaling relationships used to quantify coseismic
deformation. Nevertheless, our study provides a conceptual framework and invites more
rigorous modeling of seismic mountain building.

8

2. Introduction
Earthquakes are thought to contribute substantially to tectonic uplift and orogenic growth
(Avouac, 2007). However, earthquakes can also result in widespread landsliding in mountain
belts, leading to enhanced erosion and thus working to reduce surface topography (Keefer.,
1994; Dadson et al., 2004; Guzzetti et al., 2009; Korup et al., 2010; Hovius et al., 2011;
Egholm et al., 2013). Notably, it has been observed that large, shallow earthquakes trigger
mass wasting that can effectively offset or even outweigh the coseismic addition of rock mass
or volume to an orogen (Hovius et al., 2011; Parker et al., 2011). Quantifying this earthquake
volume balance, or the net result of coseismic mass wasting and coseismic crustal growth, is
critical for understanding crustal mass budgets, landscape building, and the role of
earthquakes in mountain belt evolution.

Previous studies have estimated the volume balance for individual earthquakes through
mapping landslides from remote imagery, gauging riverine sediment load, and measuring
surface displacements within the zones of concentrated slip and moment release (e.g., Hovius
et al., 2011; Parker et al., 2011), but less attention has been paid to developing a general
volume balance relation for earthquakes. The general volume balance equation for a single
earthquake depends on the deficit term from landsliding and fluvial evacuation, and the
surplus term from coseismic fault slip and rock deformation. Additional volume loss over
repeated earthquake cycles can result from fluvial and diffusive hillslope erosion and
landslides not associated with earthquake triggering, and additional volume gain can result
from aseismic and interseismic slip, and from viscoelastic, isostatic, or dynamic effects.
9

However, here we focus specifically on the coseismic volume balance, which has heretofore
been difficult to isolate from the interseismic processes. An empirical correlation between the
total landslide volume triggered by an earthquake and the earthquake‟s moment magnitude
M
w
has been reported (Keefer, 1994), and this provides a generalizable constraint on the
deficit of the volume balance. Increasing landslide area and volume with earthquake
magnitude is related to the triggering of landslides by ground motion, and is modulated by
topographic effects and seismic wave attenuation, which together have been shown to control
the rate and distribution of coseismic landslides (Meunier et al., 2007).

For the volume surplus term, two-dimensional dip-slip dislocation models simulate coseismic
crustal deformation, and these are well validated by field observations (Okada, 1985; 1992;
Cohen, 1996). Together with statistical correlations between fault rupture parameters and
earthquake magnitudes (Wells and Coppersmith, 1994), these models make it possible to
relate earthquake magnitude to the volume of material added to the upper crust. With the
deficit and surplus terms thus constrained, an analytical volume balance equation for
earthquakes with specified magnitudes can be derived. By introducing the Gutenburg-Richter
frequency-magnitude relation and a regional seismic-intensity factor, we can further derive an
analytical expression for coseismic crustal thickening rates in terms of the frequency of
occurrence of earthquakes. These thickening rates represent the cumulative effects from
coseismic tectonic volume addition and landslide erosion for all earthquakes in a given
region.

10

In this study, we combine new landslide data for the Wenchuan earthquake and the derivation
of this general volume balance relation to investigate the problem of earthquake volume
budgets. The M
w
7.9 Wenchuan earthquake occurred in an area of steep mountainous
topography and caused widespread coseismic landsliding (Dai et al., 2011; Parker et al.,
2011). It thus provides an ideal case study for evaluating earthquake volume balance. We first
report a new coseismic and immediate postseismic landslide dataset for the Wenchuan
earthquake, developed through mapping using high-resolution remote imagery covering the
rupture zones. We calculate the total landslide volume using a power-law landslide
area-volume relation (Guzzetti et al., 2009; Larsen et al., 2010), and compare the landslide
volume to the surface uplift measured from Synthetic Aperture Radar (SAR) (de Michele et
al., 2010; Parker et al., 2011). Our result from Wenchuan is then used, together with data
from the 1999 Chi-Chi earthquake in Taiwan, to test our general model for describing
earthquake volume balance. Although the very large uncertainties in the parameterizations
used in our analysis make it difficult to confidently discern positive versus negative volume
balance, we view the conceptual framework presented here as providing a potentially valuable
foundation for future work that may reduce these uncertainties.

11

3. Setting
The M
w
7.9 Wenchuan earthquake occurred on May 12, 2008 in the Longmen Shan mountain
range, Sichuan Province, central China. The regional lithology is characterized by mixed
assemblages of Proterozoic basement rocks, a Paleozoic passive margin sequence, a Mesozoic
foreland basin succession, and limited exposures of Cenozoic sediment (Burchfiel et al.,
1995). The faults within the region are mainly dextral-thrust oblique-slip faults, which
initiated in the Late Triassic and have been active through the Cenozoic (Densmore et al.,
2007). Based on modern geodetic observations and paleoseismology, the recurrence time of
large earthquakes within the Longmen Shan range is estimated to be ~2000-4000 years (Ran
et al., 2010; Shen et al., 2009). Geophysical observations show that fault displacement varied
greatly along the surface rupture, with two areas, Yingxiu and Beichuan, suffering the largest
slip and moment release (Figure 1) (Liu-Zeng et al., 2009; Shen et al., 2009).

12

4. Methods
We first used unsupervised classification based on spectral intensities (e.g., Borghuis et al.,
2007; Parker et al., 2011; West et al., 2011) to identify and extract landslide areas from
satellite imagery. We used high-resolution satellite imagery (SPOT, Digital Globe WorldView
and QuickBird images) taken within one month after the earthquake. Our landslide mapping
was conducted over 38,270 km
2
in the Longmen Shan, covering over 90% of the surface
rupture area and the zones of most concentrated landslide density. Using manual screening,
we then removed non-landslide objects including roads, buildings and terraces, based on
visual contrast and spatial characteristics of landslide locations. Large clusters of
amalgamated landslides were segmented into their constituent parts and each individual
landslide was delineated manually (see Appendix). The mapped landslides were then
compared to images from before the earthquake to eliminate pre-existing landslides.

13

5. Results
Within the Longmen Shan region, we mapped a total of 57,150 landslides, with a total area of
~396 km
2
, which is much smaller than the previous estimate (~566 km
2
) by Parker et al.
(2011). The probability density of all landslides is well described by a three-parameter
inverse-gamma distribution (see Appendix), as observed for many other landslide inventories
(Malamud et al., 2004).

The conversion from area to volume for each individual landslide is implemented via a
power-law scaling relationship (Hovius et al., 2011; Larsen et al., 2010):



n
i
i L L
A V
1

 (1)
where V
L
is the total volume of landslide material, A
Li
is the area of the ith landslide, n is the
number of mapped landslides, and  and  are empirical scaling parameters. Based on this
relationship, the total landslide volume for the Wenchuan landslides is calculated as
~2.8+0.9/-0.7 km
3
by using published scaling parameters (Guzzetti et al., 2009; Larsen et al.,
2010) including those obtained from field measurements of 41 coseismic landslides in the
Longmen Shan (Parker et al., 2011) (Table 1). The value and uncertainty of the total landslide
volume are determined by Monte Carlo simulation taking into account combinations of the
two scaling parameters  and . For each group of parameters, volume calculations (i.e., Eq. 1)
on the Wenchuan landslide inventory were repeated 50,000 times with random sampling of
normally-distributed scaling parameters  and , and the total landslide volume value is
reported based on the median of the Monte Carlo distribution with lower and upper bounds
defined by the 16th and 84th percentiles of the distribution, respectively (Table 1 and Table
14

S3). To account for variations among different combinations of parameters, a combined total
landslide volume value and relevant uncertainties (2.8+0.9/-0.7 km
3
) are then calculated by
applying this sampling algorithm to all combinations of scaling parameters (Table 1, Table S3
and details in Appendix). Sensitivity analysis indicates that the most significant source of
uncertainty in the final calculation of total landslide volume is from the uncertainty in the
parameter  (see Figure S6 and Table S4 in Appendix). The estimated landslide volume range
(2.1-3.7 km
3
) is consistent with the volume range (1.5-3.6 km
3
) reported in a recent study
(Ren et al., 2013), which determined well-constrained volumes within smaller spatial
windows in the Longmenshan and extrapolated these to the total area of coseismic landslides
assuming a lognormal distribution.

The global correlation between the total volume of landslides triggered by an earthquake V
L

and the earthquake moment magnitude M
w
(Keefer et al., 1994; Malamud et al., 2004)
provides context for interpreting the estimated volumes from the Wenchuan earthquake:
) 52 . 0 ( 26 . 11 42 . 1 log   
w L
M V (2)
For M
w
=7.9, this global scaling relationship gives a total landslide volume V
L
of 0.9+2.1/-0.6
km
3
, or a range of 0.3-3.0 km
3
. This compares to our estimation from mapping of V
L
=
~2.8+0.9/-0.7 (2.1-3.7) km
3
. Although the mean volume derived from our mapping is higher
than the mean inferred from the global scaling relationship, the ranges clearly overlap
considering the uncertainties. This type of comparison could be improved by further efforts to
reduce uncertainties both in the global relationships and in the area-volume parameters used
to determine landslide volume from individual earthquakes such as Wenchaun.
15

6. Discussion
6.1. The Wenchuan earthquake volume balance
The coseismic volume addition to the Longmen Shan region resulting from slip during the
Wenchuan earthquake is ~2.6± 1.2 km
3
based on Synthetic Aperture Radar (SAR) data (de
Michele et al., 2010; Parker et al., 2011). This range is close to our estimated total landslide
volume (2.8+0.9/-0.7 km
3
). Although the large uncertainties on both values limit our ability to
confidently distinguish between positive and negative net volume balances, the first-order
similarity between the volume growth and potential reduction due to landslides implies that,
for the Wenchuan earthquake at least, seismically-triggered landslide erosion can significantly
offset coseismic tectonic rock uplift if all of the landsliding material can be evacuated by
rivers before the next comparable earthquake. Incomplete fluvial evacuation of landslide
material is unlikely to affect the long-term volume budget (e.g. over the timescale of repeated
earthquake cycles), because it would require long-term accumulation of very significant
amounts of landslide debris, at odds with the thin alluvial cover on hillslopes and lack of thick
pre-2008 sediment stores in the Longmen Shan (Ouimet et al., 2009; Parker et al., 2011).
Thus, from a volume balance perspective, the contribution from coseismic deformation during
the Wenchuan earthquake to the long-term, regionally averaged topographic evolution of the
Longmen Shan range is considerably reduced by coseismic landslides, and may be close to
insignificant.

Our calculated landslide volume (2.8+0.9/-0.7 km
3
) is lower than the previously reported
volume range of 5-15 km
3
(Parker et al., 2011). This previous work used only
16

an automated algorithm to extract landslides. We added rigorous manual screening after
noticing that the automated routine did not separate amalgamated clusters of landslides into
their component parts. Segregation of amalgamated clusters has a large potential effect on
estimated landslide volumes because of the non-linear relationship between landslide area and
volume (Guzzetti et al., 2009; Larsen et al., 2010). The ratio of the volumes from our study
compared to Parker et al. (2011) falls on a predetermined curve controlled by the splitting of
clumped landslides (see Appendix); any difference between the studies in screening for
non-landslide areas has minimal effect. The significant differences in calculated volumes
demonstrate that large landslide areas not divided into their constituent parts (i.e. the long tail
in Figure 2A) can strongly bias estimates of landslide volumes (consider the contributions of
different-sized landslides to the total landslide volume, as shown Figure 2B).

While the differences between the revised landslide volume estimate presented here and the
previously reported volume from Parker et al. (2011) highlight the importance of
differentiating individual landslides from clustered landslide during automated mapping at
large scales, it is worth noting that both results lead to the same conclusion that
earthquake-triggered mass wasting may effectively offset coseismic volume addition.
Although accurate net coseismic volume difference cannot be confidently determined within
the uncertainties in the data and the methodology used here, the similarity of the volume
estimates is a key observation that requires further consideration.

17

If other earthquakes follow the same pattern, the volume budget for Wenchuan poses
important questions about coseismic mountain building (Parker et al., 2011). One explanation
for the efficient counteraction of coseismic volume addition in the Wenchuan case may be
that erosion and uplift are indeed balanced in the present-day Longmen Shan (Godard et
al.,2009). Another is that isostatic compensation for removed landslide material counteracts
mass wasting and facilitates rock uplift (Molnar, 2012). Simple calculation of the
flexural-isostatic response of the Longmen Shan range, however, indicates that
erosionally-induced rock uplift could only replace ~30% of the mass lost from landslides
(Densmore et al. 2012). Here we explore another possibility: that orogenic growth is
controlled by the imbalance between volume accumulation in small earthquakes which trigger
low volumes of landslides, and volume destruction from large earthquakes which trigger large
landslide volumes (Parker et al., 2011). In that scenario, the coseismic orogenic volume
balance should depend significantly on earthquake magnitudes.

6.2. Volume balance for earthquake events with specified moment magnitudes
We examine the relationship between earthquake volume balance and earthquake magnitude
using a first-order quantitative model informed by empirical scaling relationships. We relate
the volume budget of an individual earthquake event to the earthquake‟s moment magnitude
by considering (1) the total landslide volume-magnitude relation for a given event (Keefer,
1994), (2) the analytical deformation field for an earthquake event from a two-dimensional
dip-slip dislocation model, and (3) scaling relations between fault rupture parameters (e.g.,
rupture area, surface displacement, and rupture length) and earthquake magnitude.
18

Based on the analytical deformation field for a two-dimensional dip-slip dislocation model in
a homogeneous, elastic half-space (Okada, 1985; 1992), we calculate the uplift volume
caused by the surface deformation of a thrust-fault earthquake by integrating the vertical
displacement over the uplifted range (see Appendix). The integrated result can be expressed
geometrically as an extruded volume (the yellow colored region in Fig. 3). The uplifted
volume V
U

is expressed as a function of fault rupture area A, surface displacement D, and
dip angle θ (see Appendix):
  cos sin
2
1
AD V
U
 (3)
To minimize the effects from dip angle θ and to average the uplifted volume over the dip
angle, we define a dip-angle averaging term Θ as:


 
m ax
m in
m ax
m in
cos sin
2
1





  
d
d
(4)
where Θ is used to normalize the uplift volume as a function of dip angle θ over a range of dip
angles. Integration over the dip angle range (θ
min
≤ θ ≤ θ
max
) gives the dip angle-averaged
uplifted volume:
  AD V
U
(5)
The well-constrained fault geometry of the Wenchuan earthquake (Xu et al., 2009) can be
used to examine this uplift volume model in comparison with the results from SAR-based
geodetic observations (de Michele et al., 2010). With the fault geometric parameters (focal
depth ~14-18 km, subsurface-surface dip angle ~40° -~90° , and rupture length ~240 km) from
Xu et al. (2009) and our integration method (Eq. 3, 4 and 5), we estimate the 2008 Wenchuan
earthquake rock uplift volume as 3.5± 0.9 km
3
, which overlaps within uncertainty with the
19

SAR-based rock uplift volume of 2.6± 1.2 km
3
. This suggests that our simplified,
two-dimensional dip-slip dislocation model provides a reasonable first-order constraint on
uplift volumes. To generalize this uplift volume model, we adopted empirical relationships
between fault rupture area, surface displacement and moment magnitude M
w
from Wells and
Coppersmith (1994), which were reported as:
w A A
M b a A   log (6)
w D D
M b a D   log (7)
where a
A
, b
A
, a
D
, and b
D
are empirical constants. By combining these with Eq. 5, we obtain an
expression relating the dip angle-averaged uplifted volume and M
w
:

      log ) ( ) ( log
w D A D A U
M b b a a V

(8)
where
U
V is the dip-angle averaged uplift volume and Θ is the dip-angle average term,
normalizing the uplift volume over a range of dip angles. Values, standard errors and sources
for all the parameters are reported in Table 2.

We substitute the empirically-derived scaling factors for relationships between rupture area
and surface displacement and earthquake magnitudes for reverse fault earthquakes (Wells and
Coppersmith, 1994) into Eq. 8. We consider values for the dip angle  in the range of 20° to
40° , a geologically reasonable range for the majority of orogenic thrust faults (e.g., Abers,
2009; Shen et al., 2009), and obtain the dip-angle averaged uplift volume in terms of
magnitude:
) 44 . 1 ( 40 . 8 ) 22 . 0 ( 06 . 1 log    
w U
M V (9)

20

The relation describing the “destructive” coseismic landslide volume as a function of M
w
(Eq.
2) has a different slope from this relationship, which describes the “constructive” uplift
volume as a function of M
w
. At low M
w
, uplift volume is greater than landslide volume for a
given M
w
, but at the highest M
w
, landslide volume is greater. The relations cross at a value of
M
w
that defines a threshold, beyond which the volume of seismically-induced landslides
outweighs volume addition associated with coseismic deformation on the fault. Taking the
mean values from the parameterization of each relation, this threshold between earthquakes
that have a net positive versus net destructive volume balance would be M
w
≈ 8.0 (Figure 4A).
However, it is important to emphasize that there are very large uncertainties in this analysis
given the poor constraints on key parameters, with the largest uncertainty introduced in the
fault geometry and scaling parameters used in Eq. 8, as demonstrated by sensitivity analysis
(see Table S4 and Figure S7 in Appendix for details).

The coseismic volume budget can then be determined as the difference of the “constructive”
(Eq. 9) and “destructive” (Eq. 2) terms, as:
L U
V V V    (10)
L U
V V V log log log    (11)
Note that Eq. 11 can be rewritten as the ratio of the two volumes:
L
U
V
V
V log log   (12)
Uncertainties on these values (as shown by the dashed line-bounded ranges in Figure 4B and
Figure 4C) are propagated via integrated non-linear error propagation, taking into account
uncertainties in the parameter values used (see Table 2). Our result for the Wenchuan
21

earthquake falls within the range predicted by Eq. 12 (Figure 4B and 4C).The only other
documented assessment of earthquake volume balance is from Hovius et al. (2011), who
estimated the net effect on surface topography for the M
w
7.6 Chi-Chi earthquake in Taiwan
using hydrologic and geodetic data. Their study used the sediment load from the epicentral
Choshui catchment to calculate landslide material export and indicated that over 30% of the
added mass by the earthquake has been removed. Assuming equal density for uplifted rock
and eroded sediment, this value for the Chi-Chi earthquake also falls within the range of
volume change predicted by our model (Figure 4B and 4C).

We emphasize that this scaling relation-based, first-order volume budget has many large
uncertainties, and that the propagation of these uncertainties means that, at present, it is
difficult to confidently define the net volume balance of earthquakes. Many of the scaling
parameters, particularly those associated with coseismic uplift, require further quantification
before this problem can be fully understood. Moreover, the geometry of real fault systems is
considerably more complex than our simplified model, which considers only displacement on
a single reverse fault. In addition, in our analysis we consider deformation in an ideal, elastic
half-space (Okada., 1992), which does not include the effects of viscoelastic response that can
contribute importantly to mountain building and may also scale with earthquake magnitude
(e.g King et al., 1988). Future work might consider a viscous half-space framework. We also
assume no spatial variation in landslide susceptibility, which is controlled by a number of
factors (e.g., lithology, topography, climatic conditions, and epicentral depth). These
parameters are beyond the scope of our study, but will require further consideration to
22

develop a complete picture of the earthquake volume balance problem. Nonetheless, given the
reasonable accordance with data from the two presently-available earthquake events, we
suggest that the concept developed here provides a meaningful general framework for
considering the coseismic contribution to the volume balance of earthquakes, which with
further refinement can be used to probe the role of seismicity in the production of surface
topography and associated crustal thickening.

It is also important to note that the apparent volume balance of a given earthquake as
developed here also does not predict the total topographic effects of a single event. For
example, in the Wenchuan case, the zone of surface uplift (<15 km from the surface rupture
assuming this zone is defined by focal depth× tan(θ), and using focal depth d ~ 14-18 km and
dip angle θ ~ 40-90° ) is much smaller that the region of observed coseismic landslides (which
range up to ~100 km from the surface rupture). Moreover, small spatial scale topographic
features such as observable fault scarps may remain evident despite regional-scale volume
loss via landslides. The difference in the spatial scale of rock uplift versus landslide removal
will play a role in mass redistribution within mountain ranges, and may have important
consequent effects on tectonic processes (Egholm et al., 2013; Parker et al., 2011; King et al.,
1988), but the overall volume balance of earthquakes in terms of topographic evolution only
makes sense when integrated over large areas and over long periods of time (e.g., the
timescale of seismic cycles on major faults).

23

6.3. Integrated volume balance for multiple earthquake events
The seismic contribution to mountain building and crustal thickening is not controlled by one
single earthquake, but is the accumulated result of multiple seismic cycles (Avouac, 2007).
This means that understanding the seismic role in mountain building is further complicated by
the varying recurrence rates of earthquakes with dissimilar magnitudes in different tectonic
settings. To account for these variable seismic parameters, we incorporate the
Gutenberg-Richter frequency-magnitude relation and a seismic-intensity factor to estimate the
cumulative effect of earthquakes on coseismic crustal volume balance and the coseismic
crustal thickening that contributes to building mountainous topography.

The Gutenberg-Richter frequency-M
w
relation (Gutenburg and Richter, 1954) describes the
frequency of earthquakes in a specified region as a function of earthquake moment magnitude
M
w
, and thus allows the integration of volume budget effects from earthquakes of varying
magnitudes. Using the Gutenberg-Richter frequency-M
w
relation, we can calculate the total
coseismic uplift and landslide volumes from all earthquake events (Figure 5A; see details in
Appendix). The rate of net coseismic volume addition is then derived as the difference of the
surplus term and the deficit term:
m ax m ax
) ( ) ( ) (
10 10
w N L L w N D A D A
M b b a
N L
N N M b b b a a
N D A
N N
b b
b a
b b b
b a
V
     


 

 
 

(13)
where a
A
, b
A
, a
D
, and b
D
are parameters in the fault rupture geometry parameters-M
w
relations,
a
N
and b
N
are parameters in the Gutenberg-Richter relation, a
L
and b
L
are parameters in the
landslide volume-M
w
relation, Θ is the dip-angle averaging term, and M
wmax
is the maximum
observed magnitude for all earthquakes within the study area.
24

Introducing a regional seismic-intensity factor, I
M
, makes it possible to consider the spatial
variation of the frequency of mountain-building earthquakes. This factor expresses the
regional seismic intensity as the number of earthquakes with magnitudes greater than or equal
to M
w
in specified region (defined as a 1° long. by 1° lat. area) per year (Malamud et al.,
2004). Although a complete dataset recording all earthquakes is not available, Kossobokov et
al. (2000) has compiled a global map of I
4
, showing the spatial variation of occurrence rates
for earthquakes with magnitude greater than or equal to M
w
= 4. The global dataset of the
seismic-intensity factor I
4
thus allows calculation of earthquake frequency in different settings.
Note that magnitude 4 is also the approximate cutoff magnitude of earthquakes observed to
generate landslides (Keefer et al., 2002). Combining the seismic-intensity factor, the global I
4

dataset, and the Gutenberg-Richter relation-based cumulative earthquake volume budget
model (Eq. 13), we convert volume change to a rate of crustal thickening by assuming that the
additional material is spread uniformly over the 1° × 1° area of defined seismic intensity from
the Kossobokov et al. (2000) compilation. We thus obtain a generalized analytical expression
for estimating the rate of seismically-induced crustal thickening in terms of the observed
maximum earthquake moment magnitude M
wmax
in the study area, and the regional
seismic-intensity factor I
4
(see Appendix for derivation):
N w N L L w N D A D A
b M b b a
N L
N M b b b a a
N D A
N
E
I
b b
b
b b b
b
A
h
4
4
) ( ) ( ) (
10 10 10
1
m ax m ax










 


     

(14)
where h

is the crustal thickening rate (km Ma
-1
) during coseismic deformation, and A
E
is the
equivalent normalized 1° × 1° area at the equator (~111× 111 km
2
).

25

The numerical results can be derived by substituting all parameters (Table 2) into Eq. 14,
giving:
4
6 52 . 0 3 16 . 0
) 10 07 . 3 10 70 . 6 (
max max
I h
w w
M M  
   

(15)
The dependence of crustal thickening rate on M
wmax
and I
4
is shown in Figure 5B. Referring to
the compiled global spatial distribution of I
4
and M
wmax
in Kossobokov et al. (2000), we may
derive first-order coseismic crustal thickening rates for specific regions using Eq. 15. Then
this generalized coseismic mountain building and crustal thickening model can be compared
to observed orogenic uplift and exhumation rates. The Himalaya, a typical
continental-continental collision zone characterized by thrust-fault earthquakes, provides an
example. In this case, I
4
≈ 1-5 earthquakes yr
-1
and M
wmax
≈ 9 on the global I
4
and M
wmax
maps
(Kossobokov et al., 2000), and the resulting modeled coseismic crustal thickening rates are
around 0.1-0.5 km Ma
-1
. Though this is potentially geologically reasonable, rigorous
validation and refinement of this generalized model will require further careful assessment. It
is also important to note that our model is based on empirical parameterizations that are
specific to thrust fault settings. Nonetheless, considering the large uncertainties on
geodynamic parameters at orogenic scales, our model provides a first-order estimation of
coseismic cumulative crustal volume change, and most importantly presents a new conceptual
approach for considering this problem quantitatively.

26

6.4. Relative overall contributions to coseismic mountain building from earthquakes of
varying magnitude
We can estimate the contributions to the total volume of uplifted rock as a result of coseismic
deformation from earthquakes with specified magnitudes, and define a volume contributing
fraction function f(M
w
):
total
w w w
w
V
M V M M V
M f

 

   

) ( ) (
) (


(16)
where
total
V

 is the total uplift volume across the range of all magnitudes (determined as the
difference between the maximum net uplift volume and the net uplift volume at M
w
= 4), and
δM is the size of each magnitude bin (set here to be δM
w
= 0.1). For 4 ≤ M
w
≤ 9, the
contributions to the total rock uplift volume from different earthquakes are shown in Figure 6.
The distribution of added fractions suggests that moderate to large earthquakes (6 ≤ M
w
≤ 7)
add most to the total orogenic volume through coseismic deformation, while the very largest
earthquakes may be net destructive, because of the much more significant landslide volume
reduction effects at larger magnitudes.

27

7. Conclusion
We mapped the Wenchuan earthquake-triggered landslides by combining automated
extraction and manual segregation of landslide clusters. The total calculated landslide volume
is ~2.8+0.9/-0.7 km
3
. This revises the initial estimate of Parker et al. (2011), but confirms that
coseismic volume removal may significantly counteract seismically-induced orogenic growth.
The uncertainty in our estimated volume of Wenchuan-triggered landslides remains
significant, largely because of weak constraints on the parameters in the area-volume scaling
relationships used in the calculation. Better understanding of the sources of parametric
uncertainties in this scaling relation, or other direct approaches to determine volumes, will be
required to reduce such uncertainties. To consider whether the observations from Wenchuan
are generalizable, we develop a model of thrust-fault associated uplift that allows us to
evaluate the volume balance for earthquakes with specified magnitudes. We find that there
may be a threshold earthquake magnitude above which volume wasting from coseismic
landsliding exceeds earthquake-triggered volume growth. The very large uncertainties in the
parameterizations used in our analysis hamper clear definition of the magnitude of this
threshold, with the greatest uncertainty coming from the parameterization of coseismic uplift
but uncertainty in global landslide volume-M
w
relation also significant. Future work building
on this conceptual framework might reduce these uncertainties. By incorporating the
Gutenburg-Richter relation and the seismic-intensity factor I
4
, we estimate the cumulative
effect for all earthquakes in different regions. Assuming efficient erosion and fluvial export,
and using the mean values of fault geometry and coseismic landslide volume from our
parameterization, the net crustal thickening rates associated with coseismic deformation in
28

typical continental-continental collision zones (e.g., the Himalaya) would be on the order of
~0.1-0.5 km Ma
-1
.
Based on our analysis and the global earthquake inventory, the earthquakes
that contribute most to crustal thickening and mountain building are probably medium to
large events with magnitudes of 6 to 7, although future work to reduce uncertainties is clearly
warranted. Such work might make it possible to better quantify the coseismic contribution to
deformation, and to then be able to relate this to other sources of information such as geodetic
observations.

29

Bibliography
Abers, G. A. (2009), Slip on shallow-dipping normal faults, Geology, 37(8), 767-768,
doi:10.1130/focus082009.1.
Avouac, J. P. (2007), 6.09 - Dynamic Processes in Extensional and Compressional Settings –
Mountain Building: From Earthquakes to Geological Deformation, in Treatise on
Geophysics, edited by S. Gerald, pp. 377-439, Elsevier, Amsterdam, doi:
10.1016/B978-044452748-6.00112-7.
Borghuis, A. M., K. Chang, and H. Y . Lee (2007), Comparison between automated and
manual mapping of typhoon-triggered landslides from SPOT-5 imagery, Int J Remote
Sens, 28(7-8), 1843-1856, doi:10.1080/01431160600935638.
Burchfiel, B. C., Z. L. Chen, Y . Liu, and L. H. Royden (1995), Tectonics of the Longmen
Shan and adjacent regions, central China, International Geology Review, 37(8), 661-735,
doi: 10.1080/00206819509465424.
Cohen, S. C. (1996), Convenient formulas for determining dip-slip fault parameters from
geophysical observables, Bulletin of the Seismological Society of America, 86(5),
1642-1644.
Dai, F. C., C. Xu, X. Yao, L. Xu, X. B. Tu, and Q. M. Gong (2011), Spatial distribution of
landslides triggered by the 2008 Ms 8.0 Wenchuan earthquake, China, Journal of Asian
Earth Sciences, 40(4), 883-895, doi: 10.1016/j.jseaes.2010.04.010
Dadson, S. J., N. Hovius, H. Chen, W. B. Dade, J. C. Lin, M. L. Hsu, C. W. Lin, M. J. Horng,
T. C. Chen, J. Milliman and C. P. Stark (2004), Earthquake-triggered increase in
sediment delivery from an active mountain belt, Geology, 32(8), 733-736, doi:
30

10.1130/G20639.1.
de Michele, M., D. Raucoules, J. de Sigoyer, M. Pubellier, and N. Chamot-Rooke (2010),
Three-dimensional surface displacement of the 2008 May 12 Sichuan earthquake (China)
derived from Synthetic Aperture Radar: evidence for rupture on a blind thrust, Geophys J
Int, 183(3), 1097-1103, doi: 10.1111/j.1365-246X.2010.04807.x.
Densmore, A. L., M. A. Ellis, Y . Li, R. J. Zhou, G. S. Hancock, and N. Richardson (2007),
Active tectonics of the Beichuan and Pengguan faults at the eastern margin of the
Tibetan Plateau, Tectonics, 26(4), doi: 10.1029/2006TC001987.
Densmore, A. L., R. N. Parker, N. J. Rosser, M. de Michele, Y . Li, R. Q. Huang, S. Whadcoat,
and D. N. Petley (2012), Reply to „Isostasy can't be ignored‟, Nat Geosci, 5(2), 83-84,
doi:10.1038/ngeo1385.
Egholm, D. L., M. F. Knudsen, and M. Sandiford (2013), Lifespan of mountain ranges scaled
by feedbacks between landsliding and erosion by rivers, Nature, 498(7455), 475-478, doi:
10.1038/nature12218.
Fu, Z., C. B. Hu, H. M. Zhang, Y . G. Cai, and Y . J. Zhou (2011), The possibility of inferring
rupture depths of fault earthquakes from zero-strain points of coseismic surface
deformation, Seismol Res Lett, 82(1), 89-96, doi: 10.1785/gssrl.82.1.89.
Godard, V ., R. Pik, J. Lave, R. Cattin, B. Tibari, J. de Sigoyer, M. Pubellier, and J. Zhu (2009),
Late Cenozoic evolution of the central Longmen Shan, eastern Tibet: Insight from
(U-Th)/He thermochronometry, Tectonics, 28, TC5009, doi: 10.1029/2008TC002407.
Gutenberg, B., and C. F. Richter (1954), Seismicity of The Earth and Associated Phenomenon
(2nd ed), Princeton University Press, Princeton, NJ.
31

Guzzetti, F., F. Ardizzone, M. Cardinali, M. Rossi, and D. Valigi (2009), Landslide volumes
and landslide mobilization rates in Umbria, central Italy, Earth and Planetary Science
Letters, 279(3-4), 222-229, doi: 10.1016/j.epsl.2009.01.005.
Hanks, T. C., and H. Kanamori (1979), A moment magnitude scale, Journal of Geophysical
Research: Solid Earth, 84(B5), 2348-2350.
Hovius, N., P. Meunier, C. W. Lin, H. Chen, Y . G. Chen, S. Dadson, M. J. Horng, and M.
Lines (2011), Prolonged seismically induced erosion and the volume balance of a large
earthquake, Earth and Planetary Science Letters, 304(3-4), 347-355, doi:
10.1016/j.epsl.2011.02.005.
Keefer, D. K. (1994), The importance of earthquake-induced landslides to long-term slope
erosion and slope-failure hazards in seismically active regions, Geomorphology, 10(1-4),
265-284, doi: 10.1016/0169-555X(94)90021-3.
Keefer, D. K. (2002), Investigating landslides caused by earthquakes - a historical review,
Surv Geophys, 23(6), 473-510, doi:10.1023/A:1021274710840.
King, G. C. P., R. S. Stein, and J. B. Rundle (1988), The growth of geological structures by
repeated earthquakes 1. Conceptual framework, Journal of Geophysical Research: Solid
Earth, 93(B11), 13307-13318, doi: 10.1029/JB093iB11p13307.
Korup, O., A. L. Densmore, and F. Schlunegger (2010), The role of landslides in mountain
range evolution, Geomorphology, 120(1-2), 77-90, doi:
10.1016/j.geomorph.2009.09.017.
Kossobokov, V . G., V . I. Keilis-Borok, D. L. Turcotte, and B. D. Malamud (2000),
Implications of a statistical physics approach for earthquake hazard assessment and
32

forecasting, Pure Appl Geophys, 157(11-12), 2323-2349, doi: 10.1007/PL00001086.
Larsen, I. J., D. R. Montgomery, and O. Korup (2010), Landslide erosion controlled by
hillslope material, Nat Geosci, 3(4), 247-251, doi: 10.1038/ngeo776.
Liu-Zeng, J., Z. Zhang, L. Wen, P. Tapponnier, J. Sun, X. Xing, G. Hu, Q. Xu, L. Zeng, L.
Ding, C. Ji, K. W. Hudnut, and J. van der Woerd (2009), Co-seismic ruptures of the 12
May 2008, Ms 8.0 Wenchuan earthquake, Sichuan: East-west crustal shortening on
oblique, parallel thrusts along the eastern edge of Tibet, Earth and Planetary Science
Letters, 286(3-4), 355-370, doi: 10.1016/j.epsl.2009.07.017.
Malamud, B. D., D. L. Turcotte, F. Guzzetti, and P. Reichenbach (2004), Landslides,
earthquakes, and erosion, Earth and Planetary Science Letters, 229(1–2), 45-59, doi:
10.1016/j.epsl.2004.10.018.
Molnar, P. (2012), Isostasy can't be ignored, Nat Geosci, 5(2), 83-83, doi: 10.1038/ngeo1383.
Meunier, P., N. Hovius, and A. J. Haines (2007), Regional patterns of earthquake-triggered
landslides and their relation to ground motion, Geophys. Res. Lett., 34(20), doi:
10.1029/2007GL031337.
Okada, Y . (1985), Surface deformation due to shear and tensile faults in a half-space, Bulletin
of the Seismological Society of America, 75(4), 1135-1154.
Okada, Y . (1992), Internal deformation due to shear and tensile faults in a half-space, Bulletin
of the Seismological Society of America, 82(2), 1018-1040.
Ouimet, W. B., K. X. Whipple, and D. E. Granger (2009), Beyond threshold hillslopes:
Channel adjustment to base-level fall in tectonically active mountain ranges, Geology,
37(7), 579-582, doi: 10.1130/G30013A.1.
33

Parker, R. N., A. L. Densmore, N. J. Rosser, M. de Michele, Y . Li, R. Q. Huang, S. Whadcoat,
and D. N. Petley (2011), Mass wasting triggered by the 2008 Wenchuan earthquake is
greater than orogenic growth, Nat Geosci, 4(7), 449-452, doi: 10.1038/ngeo1154.
Ran, Y . K., L. C. Chen, J. Chen, H. Wang, G. H. Chen, J. H. Yin, X. A. Shi, C. X. Li, and X.
W. Xu (2010), Paleoseismic evidence and repeat time of large earthquakes at three sites
along the Longmenshan fault zone, Tectonophysics, 491(1-4), 141-153, doi:
10.1016/j.tecto.2010.01.009.
Ren, Z. K., Z. Q. Zhang, J. H. Yin, F. C. Dai and H. P. Zhang (2013), Morphogenic
uncertainties of the 2008 Wenchuan earthquake: generating or reducing? Journal of
Earth Science, accepted.
Shen, Z. K., J. B. Sun, P. Z. Zhang, Y . G. Wan, M. Wang, R. Burgmann, Y . H. Zeng, W. J. Gan,
H. Liao, and Q. L. Wang (2009), Slip maxima at fault junctions and rupturing of barriers
during the 2008 Wenchuan earthquake, Nat Geosci, 2(10), 718-724,
doi:10.1038/ngeo636.
Singh, S. J., and S. Rani (1993), Crustal deformation associated with two-dimensional thrust
faulting, Journal of Physics of the Earth, 41(2), 87-101, doi: 10.4294/jpe1952.41.87.
Singh, S. J., and M. Singh (1998), A note on the relationship between dip-slip fault parameters
and plate parameters in the two-dimensional modelling of preseismic deformation,
Proceedings of the Indian Academy of Sciences-Earth and Planetary Sciences, 107(2),
155-160, doi: 10.1007/BF02840466.
Wells, D. L., and K. J. Coppersmith (1994), New empirical relationships among magnitude,
rupture length, rupture width, rupture area, and surface displacement, Bulletin of the
34

Seismological Society of America, 84(4), 974-1002.
West, A. J., C. W. Lin, T. C. Lin, R. G. Hilton, S. H. Liu, C. T. Chang, K. C. Lin, A. Galy, R.
B. Sparkes, and N. Hovius (2011), Mobilization and transport of coarse woody debris to
the oceans triggered by an extreme tropical storm, Limnol Oceanogr, 56(1), 77-85, doi:
10.4319/lo.2011.56.1.0077.
Xu, X., X. Wen, G. Yu, G. Chen, Y . Klinger, J. Hubbard, and J. Shaw (2009), Coseismic
reverse- and oblique-slip surface faulting generated by the 2008 Mw 7.9 Wenchuan
earthquake, China, Geology, 37(6), 515-518, doi: 10.1130/g25462a.1.

35

Appendix
A1. Imagery
The satellite images (IKONOS, QuickBird, SPOT and WorldView) used in this work are
listed in Table S1 and the attached spreadsheet file. For mapping coseismic and immediate
postseismic landslides, we used images mostly collected within 30 days of the earthquake,
with some images up to around one year.

A2. Landslide mapping methods
To map landslides over the large rupture zone, we first extracted large landslide clusters using
an automated algorithm, following Parker et al. (2011), and then carried out manual
refinement, including segmenting clustered landslides into constituent parts, removing
non-landslide objects, and eliminating pre-existing landslides, based on visual contrast and
spatial characteristics of landslide locations. The Wenchuan earthquake-associated landslides
were mapped over an area of 38,270 km
2
within the Longmen Shan mountain range (Figure 1
of the main text), covering over 90% of the surface rupture area and the zones of most
concentrated landslide density.

The automated algorithm principally used maximum likelihood-based unsupervised
classification for separating landslide and non-landslide areas based on the distinct spectral
intensities of landslides on images (Borghuis et al., 2007; Parker et al., 2011; West et al.,
2011). We applied this method to all multispectral images and selected landslide and
non-landslide classes based on the best visual results. For panchromatic images, similarly, a
36

binary classification of landslide and non-landslide areas was performed by defining a pixel
intensity threshold, based on the best visual effects. The automatically selected landslide
classes included landslide clusters consisting of several small landslides. Since the total
landslide volume is determined from the total number and individual areas of all landslides
(Larsen et al., 2010), for correct landslide volume estimation, we manually split large
landslide clusters into their constituent parts. Non-landslide objects (e.g., roads, buildings,
terraces) were removed through naked-eye screening. These two procedures are shown in
Figure S1. Pre-existing landslides were eliminated through comparison of images acquired
prior to the earthquake.

A3. Distribution of the landslide inventory
We calculated the probability densities for our landslide inventory. The probability density
function p(A) is defined by Malamud et al. (2004), as:

L
L
A
N
N
A p

 1
) (  (S1)

where A
L
is the area of an individual landslide, N is the number of landslides in the inventory,
and δN is the number of landslides with areas in the range A
L
to A
L
+δA
L
.

The probability distribution was found to be in good agreement with a three-parameter
inverse gamma distribution, as shown in Figure S2. As proposed by Malamud et al. (2004),
the three-parameter inverse gamma distribution provides a general description of the
37

probability distribution for landslides that fits well with three complete landslide inventories
from Northridge, Umbria, and Guatemala:















 


s A
a
s A
a
a
s a A p
L L
L
exp
) (
1
) , , ; (
1 



(S2)
where Γ(ρ) is the gamma function of ρ.

When fitting a three-parameter inverse gamma distribution to our Wenchuan landslide
inventory, the last two points at extremely large areas (Figure A2) were excluded because they
each represent only one event and thus have an insignificant effect on the probability
distribution statistics. However, it is important to note that the three-parameter inverse gamma
distribution is a purely empirical fit that appears to describe multiple landslide inventories,
and no analytical solutions are provided for the parameters. The three-parameter inverse
gamma distribution can be introduced as a general context to evaluate statistical distributions
of landslide inventories, but not a rigorous criteria for calibration.

A4. Comparing results from this study with previous work: quantifying the influence on
total landslide volume from landslide cluster segmentation versus non-landslide object
removal
The landslide volumes calculated in this study differ from these reported by Parker et al.
(2011), and here we explore reasons for this difference in detail. Specifically, we want to
distinguish the effect of (1) landslide segmentation, versus (2) manual screening for
non-landslide objects.
38

Because of the large size (over 50,000 landslides) and great variation in landslide area
(varying from ~10
2
to 10
7
m
2
) of the Wenchuan landslide inventories (Parker et al. (2011) and
this study), it is difficult to directly determine the relative importance of landslide cluster
segmentation and non-landslide object removal in causing the observed volume difference. To
resolve this problem, we generalize these two manual procedures using two simple models, as
shown in Figure S3.

To model the theoretical effect of splitting clustered landslides, we assume that a landslide
cluster of area A
1
is divided into several small landslides with the same area A
2
. This case
represents an end-member of area-based segmentation (Figure S3A). Before area
segmentation, the volume of the landslide of area A
1
is calculated based on the landslide
area-volume power-law relationship (Larsen et al., 2010):



1 1
A V  (S3)
After dividing this landslide of area A
1
into a group of N landslides, each with the same area
A
2
, the total landslide volume becomes:



2 2
A N V   (S4)

N, the number of the divided landslides, can be obtained from area conservation:

1 2
A A N   (S5)
39

Combining Eq. (S1), (S2) and (S3), the ratio of the two volumes V
2
/V
1
, can be written as a
function of the ratio of the two areas A
2
/A
1
:

1
1
2
1
2











A
A
V
V
(S6)

To estimate the theoretical effect of removing non-landslide objects, we treat this procedure as
only reducing landslide area without changing the total number of landslides (Figure S3-B).
This is a reasonable simplification because of the very large number of landslides and the
large variations of landslide features within the landslide inventory. In this scenario, the ratio
of the two volumes V
2
/V
1
can also be related to the ratio of the two areas A
2
/A
1
, as:










1
2
1
2
A
A
V
V

(S7)

The two relations (Eq. S6 for landslide cluster segmentation and Eq. S7 for landslide area
reduction) can be displayed as curves in V
2
/V
1
-A
2
/A
1
space (Figure S4). These simple
calculations and the predicted curves suggest that the total landslide volume is highly
sensitive to the changes both in the number of landslides and in the area of each individual
landslide. To associate the two Wenchuan landslide studies with the end-member curves
representing the two ideal cases, we take the ratio of the two estimated volumes, the ratio of
the two areas averaged over landslides contributing most to the total volumes (Figure 2B in
the main text), and the scaling factor γ for Longmen Shan landslides (Parker et al., 2011), and
40

plot the resulting point on the V
2
/V
1
-A
2
/A
1
diagram. As shown in Figure S4, the point falls
along the curve determined by area segmentation. This same pattern holds independent of
what scaling parameters (e.g., combinations of α, γ in Eq. 1) are used. The coincidence
indicates that the segmentation of clustered landslides is likely to be the major cause of the
volume difference between the two studies.

This conclusion is not surprising, for two reasons. First, Parker et al. (2011) also removed
false (i.e. non-landslide) objects, in their case using object-oriented filtering. This minimizes
the influence of removing additional non-landslide areas by the manual approach used in this
study. Secondly, the conversion from landslide area to volume depends on the non-linear
power-law relationship, which is very sensitive to the changes in the total number of
landslides (N) and individual landslide area (A
i
), because the total landslide volume changes
non-linearly with the landslide area. Very large landslide areas, i.e. large clusters that have not
been divided into their constituent landslide parts, can thus strongly bias calculated areas.
This highlights the importance of differentiating individual landslides from clustered
landslides, and calls for careful consideration of potential biases in determining total landslide
volumes by automated mapping.

These two simplified models do not cover all conditions, and provide only theoretical
end-members representing the manual refinements that were conducted. In practice these two
procedures are not so clearly distinct, and in some conditions they are hard to differentiate.
For example, some large landslide clusters can also contain non-landslide parts, and manual
41

refinement of these clusters must include both procedures of landslide cluster segmentation
and removal of non-landslide areas. While it is difficult to completely separate the effects of
these two procedures, our models provide a quasi-quantitative way to evaluate these two steps
under conservative and reasonable assumptions.

A5. Calculating uplift volume from fault slip using plane strain models
Analytical deformation fields of a two-dimensional dip-slip dislocation model in a uniform,
elastic half space were presented in Okada (1985) and Okada (1992). For a rectangular
thrust fault of finite width W and infinite length, the analytical surface uplift of the hanging
wall is given by (Singh and Rani, 1993; Singh and Singh, 1998):






 
 




 
 

1 cos 2
sin
sin
cos
tan
2
sin
2
1



 

 X X
X X b
u (S8)

where X (= x/W) is the ratio of the distance to the surface rupture over the fault width, u is the
surface vertical displacement, b is the imposed slip, and θ is the dip angle (0 < θ < π/2)
(Figure S5).

For a fault with length L, width W, and slip b, the associated surface uplift volume can be
calculated by integration of the vertical displacement over the uplift range:

dxdy
W x W x
W x W x b
V
L x
U
  





 






 
 

1 cos ) / ( 2 ) / (
sin ) / (
sin
cos ) / (
tan
2
sin
2
1



 



(S9)
42

In this integration, y is integrated over the full fault length L, but the integration range for x
needs more consideration. Theoretically, x is integrated from the position of surface rupture (x
= 0) to the cutoff of zero vertical displacement (x = x
u
). No simple analytical solution can be
derived for x
v
, the zero-displacement position, but using a numerical solution we find that the
zero-displacement cutoff (x = x
u
) is very close to the zero-strain position where x = Wcosθ,
and that the resulting difference in the calculated uplift volume between choosing the two
cutoffs (x = x
u
versus x = Wcosθ) is within 1%. For the convenience of deriving analytical
expressions for the questions discussed in the main text, we take the zero-strain cutoff (x =
Wcosθ) for integration:

dxdy
W x W x
W x W x b
V
L
W
U
  





 






 
 





 


cos
0
2
1
1 cos ) / ( 2 ) / (
sin ) / (
sin
cos ) / (
tan
2
sin
(S10)
and obtain:

  cos sin
2
1
WLb V
U


(S11)

With the rupture area A = W× L, and the surface displacement D = b in an ideal condition, we
have:

  cos sin
2
1
AD V
U


(S12)

43

and the obtained volume is the same as the uplift volume calculated from the geometric model
introduced in Figure 3 of the main text.

Note that the surface displacement D is different from the mean displacement on the rupture
plane D’, which is involved in the definition of seismic moment (M
o
= μAD’, M
o
: seismic
moment, A: rupture area (Hanks and Kanamori, 1979)). The mean displacement D’ can be
calculated from the definition of seismic moment, the formula converting seismic moment to
moment magnitude (Hanks and Kanamori, 1979), and the statistical scaling relation between
rupture area and moment magnitude (Wells and Coppersmith, 1994). As shown in Wells and
Coppersmith. (1994), in the case of reverse faults, the measured average surface displacement
is generally less than the calculated mean displacement, and this difference can be attributed
to the incomplete propagation of the fault plane dislocation to the ground surface.

A6. Calculating cumulative mountain building and crustal thickening from multiple
earthquakes
The Gutenberg-Richter frequency-magnitude relation and the seismic-intensity factor can be
used to quantify the cumulative effect of earthquakes on the volume balance of topography,
following Malamud et al. (2004). The Gutenberg-Richter frequency-magnitude relation
(Gutenburg and Richter, 1954) describes the rate of occurrence of earthquakes in a specified
region as:

N w N CE
a M b N 

log log   
(S13)

44

where N
CE
(yr
-1
) is the total number of earthquakes that occur with moment magnitudes
greater than or equal to M within a specified time in a region characterized by an unique b
N

value. The constant
N
a 

(year
-1
) represents the regional seismicity over the specified area.

Combining the occurrence rate
CE
N

of earthquakes (Eq. S13) with the relations for volume
addition (Eq. S14, or Eq. 2 in the main text) and loss (Eq. S15, or Eq. 8 in the main text,
derived from Eq. S12 and Wells and Coppersmith (1994)), yields Eq. S16 and Eq. 17,
respectively:

) 52 . 0 ( 26 . 11 42 . 1 log   
w L
M V (S14)
      log ) ( ) ( log
w D A D A U
M b b a a V (S15)

 
m ax
m in
] [ ] [
w
w
M
M
w CE w L L
M N d M V V
 

(S16)

 
m ax
m in
] [ ] [
w
w
M
M
w CE w U U
M N d M V V
 
(S17)

We obtain the total landslide volume rates
L
V

and the uplift volume rate
up
V

by integrating:

m ax
) (
10
w N L L
M b b a
N L
N N
L
b b
b a
V
 





(S18)
m ax
) ( ) (
10
w N D A D A
M b b b a a
N D A
N N
U
b b b
b a
V
   
 





(S19)

with Eq. S14 rewritten as
w L L L
M b a V   log

(S20)
45

The difference of the two volume rates then gives the net rate of crustal volume addition
V


:

m ax m ax
) ( ) ( ) (
10 10
w N L L w N D A D A
M b b a
N L
N N M b b b a a
N D A
N N
b b
b a
b b b
b a
V
     


 

 
 

(S21)

We can now relate this net crustal volume addition rate to the regional level of seisimicity.
The intensity of seismicity is quantified by the seismic-intensity factor I
M
, the number of
earthquakes with magnitudes greater than or equal to M
w
in the latitude‟s cosine-normalized
1° × 1° areas per year (Malamud et al., 2004). With a global map of I
4
(Kossobokov et al.,
2000), the rate of net volume addition can be coupled to the regional seismic-intensity. To
relate the derived volume addition (Eq. S21) to the seismic-intensity factor I
4
and the
complied global I
4
dataset (Kossobokov et al., 2000), we first set M
w
= 4 in the
Gutenberg-Richter frequency-magnitude relation (Eq. S13), and obtain:
N
b
N
I a
4
4
10 

(S22)

By substituting Eq. S22 into Eq. S21, the net rate of volume addition induced by earthquakes
with M
w
≥ 4 is then derived as a function of seismic-intensity factor I
4
and observed regional
maximum earthquake moment magnitude M
wmax
:

N w N L L w N D A D A
b M b b a
N L
N M b b b a a
N D A
N
I
b b
b
b b b
b
V
4
4
) ( ) ( ) (
10 10 10
m ax m ax










 

 
     

(S23)
46

To convert the crustal volume addition rate
V


to the crustal thickening rate
h

, we define
A
E
as the equivalent normalized 1° × 1° area at the equator (~111 km× 111 km). The net crustal
thickening rate
h

is then obtained from
V


divided by A
E
:
E
A
V
h



 (S24)

By substituting all parameters (auxiliary Table S1) into Eq. S23 and Eq. S24, we obtain the
net crustal thickening rate as function of M
wmax
and I
4
:
4
6 52 . 0 3 16 . 0
) 10 07 . 3 10 70 . 6 (
max max
I h
w w
M M  
   


(S25)
with h

in km Ma
-1
. This equation is given as Eq. 15 in the main text.

A7. Calculating the errors for the total landslide volume
We calculate the standard errors for the total landslide volume using two different methods
and report the results in Table S3.

(1) Linear error propagation method
Following Parker et al. (2011), a linear error propagation method accounting for deviations in
γ is expressed as:

σV(α, γ) = Σv
i
(α±σα, γ±σγ)- Σv
i
(α, γ) (S26)

where σV(α, γ) refers to the standard error of the total landslide volume, v
i
(α± σα, γ± σγ)
indicates the individual landslide volume in terms of parameter α+σα, α-σα, γ+σγ or γ-σγ for
47

each individual landslide, and v
i
(α, γ) is the individual landslide volume as a function of α and
γ. This method returns a first-order estimation of the standard errors but does not account for
the non-linear property of the function V(α, γ), which may introduce bias in the error
estimation.

(2) Monte Carlo (MC) sampling method
As there are no robust analytical solutions for calculating uncertainties of summing variables
following a logarithmic normal distribution, we apply a Monte Carlo sampling method to
simulate the distribution of the final results and estimate relevant uncertainties. We randomly
sample the normally-distributed parameter space multiple times (5000, 10,000 and 50,000
times in this case) and obtain a group of total landslide volume values in each case. For each
group of parameters, we report a median value with lower and upper bounds defined by the
16th and 84th percentiles of the MC distribution (i.e., ranges of ± 1 S. D. in a standard normal
distribution). To account for variations among different parameter combinations, we calculate
a combined total landslide volume value by applying this MC sampling algorithm to all
combinations of parameters α and γ (Table S3): for each individual landslide, all groups of
parameter combinations are randomly sampled and corresponding landslide volumes are
obtained; a mean of these landslide volumes is taken as a combined volume for each
individual landslide, and all means are summed to get a total landslide volume; this process is
repeated 50,000 times and then we get a MC distribution for the final results; we report the
median value with lower and upper bounds defined by the 16th and 84th percentiles of the
MC distribution in the main text. As shown in Table S3, once the sample numbers are large,
48

the difference among different times is negligible. We report the results from 50,000 MC
samples in the main text..

A8. Sensitivity analysis
To assess the importance of error sources in the landslide area-volume calculations and in our
global model, we performed a sensitivity analysis for total landslide volume (Eq. 1) and uplift
volume (Eq. 3 and 8) on related parameters.

For landslide area-volume scaling, we define the sensitivity of each parameter as the
contribution of error from each parameter to the total error:






i
i
i
i
i
i
p
p
V
p
p
V
p s
2
2
) (
) (
) (



(S27)

where p
i
refers to parameters α and γ, s(p
i
) is the sensitivity to parameter p
i
, σp
i
is the standard
error on p
i
, and V is the total landslide volume as a function of parameter p
i
. As shown in
Figure S6 and Table S4, error on γ is the major source of uncertainty for total landslide
volume.

For coseismic volume budget, we define the sensitivity of each parameter (a
L
, a
A
, b
A
, a
D
, b
D

and θ) as the contribution of error from each parameter to the total integrated error.
Integration over M
w
and θ works to normalize the error over a range of M
w
and θ:
49

 

 





m ax
m in
m ax
m in
m ax
m in
m ax
m in
2
2
) (
) (
) (




 
 
d dM p
p
V
d dM p
p
V
p s
w
w
w
w
M
M
w
i
i
i
M
M
w i
i
i

(S28)
where p
i
refers to parameters V
L
, a
A
, b
A
, a
D
, b
D
and θ, s(p
i
) is the sensitivity to parameter p
i
,
σp
i
is the standard error on p
i
, and V is the uplift volume as a function of parameter p
i
. As
shown in Figure S7 and Table S4, errors on V
L
, a
A
, b
A
, a
D
and b
D
(the fault geometry-M
w

scaling parameters) contribute most to the total error.

50

Figures and Tables
Figure 1.

Figure 1. Regional topography of the Longmen Shan region, China, and location (red star) of
the M
w
7.9 Wenchuan earthquake, China, with surface rupture traces (red lines) (Liu-Zeng et
al., 2009; Xu et al., 2009) and mapped coseismic landslides (yellow polygons). The
black-outlined polygon shows the landslide mapping area.

51

Figure 2.

Figure 2. Comparison between the landslide inventories from this study and from Parker et al.
(2011). (A) Landslide area distributions in logarithmic space. The landslide dataset in Parker
et al. (2011) contains substantially more large landslides. This long tail is due to clustered
landslides and explains the larger volume calculated by Parker et al. (2011) compared to this
study (see Appendix). (B) Contributions to total calculated volumes from different landslide
groups. Compared to Parker et al. (2011), landslides of moderate area contribute most to the
total volume in this study.

52

Figure 3.

Figure 3. A 3-D geometrical model showing the uplift volume due to slip on a thrust fault in
terms of rupture area A, surface displacement D, and dip angle θ. The volume added by fault
slip in this geometrical model is effectively equal to the calculated volume from the analytical
deformation field for the two-dimensional dip-slip dislocation model (see Appendix). The
dashed line L
D
represents the position of zero surface strain (Fu et al., 2011).

53

Figure 4.

Figure 4. Dependence of volume balance for individual earthquakes with given magnitudes
and all earthquakes with varying magnitudes on earthquake moment magnitude M
w
. (A)
V olume addition due to coseismic rock uplift (red) and volume reduction due to landsliding
(blue) for individual earthquakes as a function of M
w
. The means of the two terms match at
M
w
≈ 8.0. The blue solid line indicates the least-square best-fit straight line to the reported
earthquake-triggered landslide inventories, adopted from Malamud et al. (2004). The blue
dashed lines represent ±1 standard deviations with respect to the best fit. The red solid line
indicates the theoretical uplift volume estimated from this study, and the red dashed lines
show ±1 standard deviations determined from error propagations of all parameters in Eq. 8.
(B) The difference between volume addition and volume reduction in terms of M
w
. (C) The
dependency of the ratio of volume addition and volume reduction in logarithmic space. The
two events for which volume budgets have been constrained, the Chi-Chi and Wenchuan
earthquakes, both fall within the predicted ranges in (B) and (C). The solid lines in (B) and (C)
represent the derived volume difference from this study, and the dashed lines and grey areas
show ±1 standard deviations from error propagations.

54

Figure 5.

Figure 5. Dependence of coseismic volume balance for all earthquakes with varying
magnitudes on earthquake moment magnitude M
w
. (A) Integrated volume surplus and deficit
for earthquakes with varying magnitudes versus M
w
. (B) Integrated crustal thickening rate
during coseismic deformation as a function of I
4
and M
wmax
.

55

Figure 6.

Figure 6. Contributions to the total coseismic orogenic volume from earthquakes of varying
magnitudes. The total coseismic orogenic volume is calculated as the total uplift volume in
the range of all magnitudes. The contributed volume fraction refers to the contribution to the
total coseismic uplift volume from earthquakes with magnitudes in each bin (set here to be
δM
w
= 0.1), as indicated in Eq. 16.

56

Figure S1.

Figure S1. Example illustration of manual procedures for checking landslide inventories
defined from semi-automated mapping, involving (A) segregating of landslide clusters; and
(B) eliminating non-landslide objects.

57

Figure S2.

Figure S2. Three-parameter inverse-gamma distribution fitting of landslide probability density
for the landslide inventory in this study (best fit parameters: ρ=1.78, a=5.93× 10
3
m
2
, and
s=-0.89× 10
3
m
2
, with a coefficient of determination r
2
=0.88). The fitting is done based on the
maximum likelihood fitting algorithm in Matlab. The last two points are excluded in fitting
because they both represent only one event and are not statistically viewed as significant.

58

Figure S3.

Figure S3. Schematic diagram showing the effect of (A) segregation of landslide clusters, and
(B) removal of non-landslide areas, noting that the diagram simplifies the procedures by
assuming the same area for segregated landslides in (A), and conservation of the number of
landslides in (B).

59

Figure S4.

Figure S4. Influence on the total landslide volume from area segregation and area reduction.
V olumes and areas are normalized to the values in Parker et al. (2011). V
2
and V
1
are
estimated landslide volumes from this study and Parker et al. (2011), respectively. A
2
and A
1

are average areas for the landslides dominating total volumes of the two studies. The result of
this study falls on the curve determined by segregating clustered landslides, indicating that
this explains most of the difference between the two results.

60

Figure S5.

Figure S5. The coordinate system used in calculating the thrust-fault surface uplift volume.
The original point is taken at the position of surface rupture. W and L are the width and length
for the thrust fault plane, respectively. The product of W and L is the rupture area A.

61

Figure S6.

Figure S6. Sensitivity test for total landslide volume on parameters α and γ. Parameter values
are in Table S1. For parameter groups G, standard error on α was not reported, and so the
sensitivity of α for G is set as 0. Values are reported in Table S4.

62

Figure S7.

Figure S7. Sensitivity test for coseismic volume budget (Eq. 11 and Eq. 12) on parameters V
L
,
a
A
, b
A
, a
D
, b
D
and θ. Parameters are defined in Table S2. Sensitivity values are reported in
Table S4. Errors on the uplift-magnitude scaling parameters introduce the most uncertainty to
the overall volume balance.

63

Table 1. Landslide area-volume scaling relationships applicable to the Wenchuan region
(Parker et al., 2011) and estimated total landslide volumes. For the Longmen Shan landslide
parameters, P1 refers to parameters obtained from an ordinary regression on the Longmen
Shan landslide field measurement dataset (Parker et al., 2011); P2 refers to parameters from a
robust regression on the same dataset. For each group of scaling parameters, volumes and
errors are obtained via a Monte Carlo sampling strategy. For the reported total landslide
volume in the text, calculations for each individual landslide are repeated 50,000 times using
all the combinations of the cited scaling parameters, and mean values for each landslide are
summed to give the total volume. All values are reported based on the median of the Monte
Carlo distribution with lower and upper bounds defined by the 16th and 84th percentiles of
the distribution (i.e., ranges of ± 1 S. D. in a standard normal distribution), respectively.
Parameter log
10
α γ V olume (km
3
) Reference
L1 (global landslides) -0.84± 0.02 1.33± 0.01 1.80+0.12/-0.1
1
Larsen et al.,
2010
L2 (global bedrock
landslides)
-0.73± 0.06 1.35± 0.01 2.81+0.39/-0.
35
Larsen et al.,
2010
L3 (mixed Himalayan
landslides
-0.59± 0.03 1.36± 0.01 4.34+0.55/-0.
49
Larsen et al.,
2010
G (global landslides) -1.13 1.45± 0.01 3.49+0.39/-0.
35
Guzzetti et al.,
2009
P1 (Longmen Shan
landslides)
-0.97± 0.37 1.30± 0.09 2.50+6.72/-1.
81
Parker et al.,
2011
P2 (Longmen Shan
landslides)
-0.99± 0.37 1.39± 0.09 2.48+6.62/-1.
79
This study
Combined (see Appendix) 2.83+0.86/-0.
65

64

Table 2. Model parameters
Variable Description Equation
Introduced
Reference
A
L
area of landslide 1
V
L
volume of landslide 1
α coefficient in landslide area-volume
relation
1 Larsen et al.,
2010; Guzzetti et
al., 2009; Parker
et al., 2011
γ coefficient in landslide area-volume
relation
1 Larsen et al.,
2010; Guzzetti et
al., 2009; Parker
et al., 2011
M
w
earthquake moment magnitude 2
V
U
coseismic uplift volume 3
A rupture area 3
D rupture surface displacement 3
θ dip angle of thrust faults 3
θ
max
θ
min
maximum and minimum dig angles of
mountain building-associated thrust faults
4
Θ theta function for averaging the dip angle
associated term over a given range
4
dip angle-averaged uplift volume 5
a
A
scaling factor in rupture area-magnitude
relation, -3.99± 0.36 (± 1 S.D.), for reverse
faults
6 Wells and
Coppersmith,
1994
b
A
scaling factor in rupture area-magnitude
relation,0.98± 0.06 (± 1 S.D.), for reverse
faults
6 Wells and
Coppersmith,
1994
a
D
scaling factor in surface
displacement-magnitude relation,
-0.74± 1.40 (± 1 S.D.), for reverse faults
7 Wells and
Coppersmith,
1994
b
D
scaling factor in surface
displacement-magnitude relation,
0.08± 0.21 (± 1 S.D.), for reverse faults
7 Wells and
Coppersmith,
1994
a
L
scaling factor in landslide
volume-magnitude relation, -11.26± 0.52
(± 1 S.D.)
13 Keefer, 1994;
Malamud et al.,
2004
b
L
scaling factor in landslide
volume-magnitude relation,1.42
13 Keefer, 1994;
Malamud et al.,
2004
a
N
coefficient in Gutenberg-Richter relation 13 Gutenburg and
Richter, 1954
U
V
65

b
N
coefficient in Gutenberg-Richter relation,
global average b
N
= 0.9
13 Gutenburg and
Richter, 1954;
Malamud et al.,
2004
M
wmax

M
wmin

upper and lower bounds of earthquake
moment magnitudes in specified area and
time
13

Earthquake recurence rate 13 Gutenburg and
Richter, 1954

net crustal volume addition rate 13
A
E
equivalent normalized 1° lat.× 1° long. area
at the equator
14 Malamud et al.,
2004

net crustal thickening rate 14
I
4
number of earthquakes with M ≥ 4 in
cosine(latitude)-normalized 1° lat.× 1° long.
areas per year
14 Kossobokov et
al.,2000;
Malamud et al.,
2004
f(M
w
) volume contributing fraction function: the
volumes of landslides with magnitudes
between M
w
and M
w
+ δM
w
, divided by the
total uplift volume
16
δM
w
the size of each magnitude bin 16
CE
N

V


h

66

Table S1. Satellite images used for landslide mapping and checking.
Table S1A. Satellite images for landslide mapping:
Sensor* Scene ID Acquired Date Spectral Type Resolution (m)
QuickBird 2 QB0208JUN030415479M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208JUN030415512M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
SPOT 5 52612860810130340531A 10/13/2008 Panchromatic 5
SPOT 5 52592870812150328292J 12/15/2008 Multispectral 10
SPOT 5 52602870810130341052J 10/13/2008 Multispectral 10
SPOT 5 52582880901150333322J 01/15/2009 Multispectral 10
SPOT 5 526028790807060346492A 07/06/2008 Panchromatic 5
SPOT 5 52612860806040402262A 06/04/2008 Panchromatic 5
SPOT 5 526128620810130340531T 10/13/2008 Panchromatic 3
SPOT 5 525928740905200332522T 05/20/2009 Panchromatic 3
SPOT 5 52602870906030403582T 06/03/2009 Panchromatic 3
SPOT 5 42612860903240356491I 03/24/2009 Panchromatic 10
SPOT 5 52612850806040402202J 06/04/2008 Multispectral 5
WorldView 1 WV0108DEC110407452P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407474P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407497P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407522P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150402058P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150402039P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150402019P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150401599P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150401580P1BS1020010005E28800 03/15/2009 Panchromatic 0.5

Table S1B. Satellite images for landslide checking:
Sensor* Scene ID Acquired Date Spectral Type Resolution (m)
IKONOS 1 IK012007012603574880000011618005grn3097N 01/26/2007 Multispectral 3
IKONOS 1 IK012007020303493930000011627409grn3096N 02/03/2007 Multispectral 3
IKONOS 1 IK012008062804024360000011615566grn3044N 06/28/2008 Multispectral 3
IKONOS 1 IK012007091404122610000011611075grn3100N 09/14/2007 Multispectral 3
IKONOS 1 IK012008062804022210000011616349grn3097N 06/28/2008 Multispectral 3
IKONOS 1 IK012008062804023290000011616350grn3099N 06/28/2008 Multispectral 3
IKONOS 1 IK012008052303512050000011605298grn3163N 05/23/2008 Multispectral 3
IKONOS 1 IK012008061704031820000011602831grn3139N 06/17/2008 Multispectral 3
IKONOS 1 IK012008062804023290000011616350grn3101N 06/28/2008 Multispectral 3
IKONOS 1 IK012008052303530530000011605288grn3117N 05/23/2008 Multispectral 3
IKONOS 1 IK012007091404120740000011611074grn3077N 09/14/2007 Multispectral 3
IKONOS 1 IK012009022204082550000011629080grn3110N 02/22/2009 Multispectral 3
IKONOS 1 IK012008052303521830000011605311grn3184N 05/23/2008 Multispectral 3
67

IKONOS 1 IK012008051804095110000011632249grn3104N 05/18/2008 Multispectral 3
IKONOS 1 IK012008070104120820000011619800grn3092N 07/01/2008 Multispectral 3
IKONOS 1 IK012008100504084300000011613066grn3107N 10/05/2008 Multispectral 3
IKONOS 1 IK012008062804020410000011616352grn3108N 06/28/2008 Multispectral 3
IKONOS 1 IK012008060904105010000011625728grn3127N 06/09/2008 Multispectral 3
IKONOS 1 IK012008052604015420000011608984grn3098N 05/26/2008 Multispectral 3
IKONOS 1 IK012008052603594500000011608952grn3177N 05/26/2008 Multispectral 3
IKONOS 1 IK012008061704013470000011602838grn3141N 06/17/2008 Multispectral 3
IKONOS 1 IK012007072904004990000011622288grn3073N 07/29/2007 Multispectral 3
IKONOS 1 IK012008060904110290000011625730grn3105N 06/09/2008 Multispectral 3
IKONOS 1 IK012008062503543110000011612239grn3057N 06/25/2008 Multispectral 3
IKONOS 1 IK012007081703525760000011612702grn3082N 08/17/2007 Multispectral 3
IKONOS 1 IK012008062503545540000011612654grn3159N 06/25/2008 Multispectral 3
IKONOS 1 IK012008100504090020000011613025grn3147N 10/05/2008 Multispectral 3
IKONOS 1 IK012008082204070020000011621983grn3056N 08/22/2008 Multispectral 3
IKONOS 1 IK012007081703534330000011612700grn3072N 08/17/2007 Multispectral 3
IKONOS 1 IK012008060904103510000011625729grn3137N 06/09/2008 Multispectral 3
IKONOS 1 IK012008053103441170000011615032grn3104N 05/31/2008 Multispectral 3
IKONOS 1 IK012008062503531540000011612664grn3106N 06/25/2008 Multispectral 3
IKONOS 1 IK012007081703532440000011612701grn3073N 08/17/2007 Multispectral 3
IKONOS 1 IK012008062804014950000011616348grn3177N 06/28/2008 Multispectral 3
IKONOS 1 IK012008061704015300000011602839grn3106N 06/17/2008 Multispectral 3
IKONOS 1 IK012007072903595900000011622287grn3074N 07/29/2007 Multispectral 3
IKONOS 1 IK012008091803501670000011624217grn3147N 09/18/2008 Multispectral 3
QuickBird 2 QB0204OCT250346391M1BS1010010003591700 10/25/2004 Multispectral 2.5
QuickBird 2 QB0207AUG200412147M1BS1010010007162F00 08/20/2007 Multispectral 2.5
QuickBird 2 QB0208FEB290410199M1BS1010010007BD4500 02/29/2008 Multispectral 2.5
QuickBird 2 QB0208JUN030415482M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208JUN030415578M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208MAY190400054M1BS10100100080E0200 05/19/2008 Multispectral 2.5
QuickBird 2 QB0209MAR130410011M1BS10100100094C6900 03/13/2009 Multispectral 2.5
QuickBird 2 QB0208JUN030415545M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208JUL220410498M1BS1010010008518A00 07/22/2008 Multispectral 2.5
QuickBird 2 QB0208DEC260405092M1BS1010010008F74A00 12/26/2008 Multispectral 2.5
QuickBird 2 QB0205JUL220402270M1BS1010010004635900 07/22/2005 Multispectral 2.5
QuickBird 2 QB0208JUN030415479M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208FEB290410167M1BS1010010007BD4500 02/29/2008 Multispectral 2.5
QuickBird 2 QB0208JUL220410325M1BS1010010008518A00 07/22/2008 Multispectral 2.5
QuickBird 2 QB0208JUL220410463M1BS1010010008518A00 07/22/2008 Multispectral 2.5
QuickBird 2 QB0208JUN030415448M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208JUN030415449M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0208JUL220410326M1BS1010010008518A00 07/22/2008 Multispectral 2.5
QuickBird 2 QB0208JUN030415414M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0205JUN260411209M1BS101001000454A600 06/26/2005 Multispectral 2.5
68

QuickBird 2 QB0208JUN030415512M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0205JUN260411240M1BS101001000454A600 06/26/2005 Multispectral 2.5
QuickBird 2 QB0205SEP090359461M1BS101001000481E500 09/09/2005 Multispectral 2.5
QuickBird 2 QB0208JUL220410291M1BS1010010008518A00 07/22/2008 Multispectral 2.5
QuickBird 2 QB0208SEP010416060M1BS1010010008822200 09/01/2008 Multispectral 2.5
QuickBird 2 QB0208SEP010416028M1BS1010010008822200 09/01/2008 Multispectral 2.5
QuickBird 2 QB0208SEP010416027M1BS1010010008822200 09/01/2008 Multispectral 2.5
QuickBird 2 QB0208SEP010416026M1BS1010010008822200 09/01/2008 Multispectral 2.5
QuickBird 2 QB0205JUN260411272M1BS101001000454A600 06/26/2005 Multispectral 2.5
QuickBird 2 QB0206MAR310413468M1BS1010010004E4C300 03/31/2006 Multispectral 2.5
QuickBird 2 QB0208JUN030416011M1BS10100100081D5D00 06/03/2008 Multispectral 2.5
QuickBird 2 QB0205JUN260411335M1BS101001000454A600 06/26/2005 Multispectral 2.5
QuickBird 2 QB0208JUL220410258M1BS1010010008518A00 07/22/2008 Multispectral 2.5
QuickBird 2 QB0205SEP090359493M1BS101001000481E500 09/09/2005 Multispectral 2.5
QuickBird 2 QB0205SEP090359527M1BS101001000481E500 09/09/2005 Multispectral 2.5
QuickBird 2 QB0205SEP090359595M1BS101001000481E500 09/09/2005 Multispectral 2.5
QuickBird 2 QB0205JUN260411304M1BS101001000454A600 06/26/2005 Multispectral 2.5
QuickBird 2 QB0205SEP090359561M1BS101001000481E500 09/09/2005 Multispectral 2.5
QuickBird 2 QB0208JUL220410394M1BS1010010008518A00 07/22/2008 Multispectral 2.5
SPOT 5 52612860810130340531A 10/13/2008 Panchromatic 5
SPOT 5 52592870812150328292J 12/15/2008 Multispectral 10
SPOT 5 52602870810130341052J 10/13/2008 Multispectral 10
SPOT 5 52582880901150333322J 01/15/2009 Multispectral 10
SPOT 5 526028790807060346492A 07/06/2008 Panchromatic 5
SPOT 5 52612860806040402262A 06/04/2008 Panchromatic 5
SPOT 5 526028740705060403372T 05/06/2007 Panchromatic 3
SPOT 5 526028740906030403582T 06/03/2009 Panchromatic 3
SPOT 5 52612860611060345561T 11/06/2006 Panchromatic 3
SPOT 5 52602870705060403372T 05/06/2007 Panchromatic 3
SPOT 5 526228610703210348441T 03/21/2007 Panchromatic 3
SPOT 5 525928730511070345341T 11/07/2005 Panchromatic 3
SPOT 5 526128620810130340531T 10/13/2008 Panchromatic 3
SPOT 5 525928740905200332522T 05/20/2009 Panchromatic 3
SPOT 5 52602870906030403582T 06/03/2009 Panchromatic 3
SPOT 5 42612860903240356491I 03/24/2009 Panchromatic 10
SPOT 5 52612850806040402202J 06/04/2008 Multispectral 5
WorldView 1 WV0108DEC030355231P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355233P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355466P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356054P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356076P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356099P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356121P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356144P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
69

WorldView 1 WV0108DEC030356166P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356189P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356211P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356227P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356250P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356272P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356295P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030356318P1BS1020010005DAF600 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407202P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407224P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407474P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407477P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407497P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407499P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407520P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407522P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC120345598P1BS102001000520CB00 12/12/2008 Panchromatic 0.5
WorldView 1 WV0108DEC120346012P1BS102001000520CB00 12/12/2008 Panchromatic 0.5
WorldView 1 WV0108DEC120346017P1BS102001000520CB00 12/12/2008 Panchromatic 0.5
WorldView 1 WV0108FEB150345045P1BS1020010001260300 02/15/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344521P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344544P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344567P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344590P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345013P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345036P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345059P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345071P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345083P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345095P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345106P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070342530P1BS10200100017C9B00 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070342550P1BS10200100017C9B00 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070343352P1BS10200100013BE500 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070343376P1BS10200100013BE500 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070343400P1BS10200100013BE500 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355173P1BS1020010003C14E00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355451P1BS10200100039C1B00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352171P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352233P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352247P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352258P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352271P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352282P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
70

WorldView 1 WV0108MAR030352296P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352307P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352320P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352331P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352391P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352394P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352415P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352418P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352439P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160340130P1BS10200100021EC100 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160340150P1BS10200100021EC100 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160340169P1BS10200100021EC100 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160340448P1BS102001000254F900 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160340495P1BS102001000254F900 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160341169P1BS1020010002CF7500 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160341189P1BS1020010002CF7500 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160341210P1BS1020010002CF7500 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160341231P1BS1020010002CF7500 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347018P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347037P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347056P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347075P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347094P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347113P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347132P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347150P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347169P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347188P1BS10200100023B9500 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347405P1BS10200100023BFC00 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347429P1BS10200100023BFC00 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347454P1BS10200100023BFC00 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347478P1BS10200100023BFC00 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347503P1BS10200100023BFC00 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY200347522P1BS10200100023BFC00 05/20/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070403356P1BS1020010004C0DD00 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070403376P1BS1020010004C0DD00 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070404056P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070404058P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070404059P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108NOV110409288P1BS102001000547D600 11/11/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358370P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358385P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358394P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0109DEC290413278P1BS102001000B2E0600 12/29/2009 Panchromatic 0.5
71

WorldView 1 WV0109DEC290413404P1BS102001000ABCB900 12/29/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406334P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406341P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406355P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406363P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406377P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406384P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406393P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406399P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109FEB130406406P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150402039P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150402058P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108DEC030355195P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355396P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407430P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407432P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100359005P1BS1020010002959900 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358362P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355373P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407452P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407455P1BS1020010005C8D700 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358420P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355220P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108FEB020344283P1BS10200100013A1C00 02/02/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070404083P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407181P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355428P1BS10200100039C1B00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358397P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400503P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108APR110352026P1BS1020010002942C00 04/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407301P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352443P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0109FEB130406320P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0109MAR150401560P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108DEC110407287P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355320P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407266P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407291P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC240402424P1BS1020010005359A00 12/24/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404288P1BS10200100032E6500 06/27/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150401501P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108DEC030355113P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355350P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407245P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
72

WorldView 1 WV0108MAR030352345P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAY240354105P1BS1020010002276900 05/24/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355213P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355111P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355200P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355202P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355474P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352209P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352355P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355218P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070404107P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108APR190405364P1BS10200100024CBD00 04/19/2008 Panchromatic 0.5
WorldView 1 WV0108NOV110409294P1BS102001000547D600 11/11/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355327P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352184P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355129P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355131P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355420P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355479P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070342510P1BS10200100017C9B00 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404018P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404015P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352369P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150401599P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108DEC030355443P1BS1020010005947C00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352367P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108JAN110351183P1BS10200100012F8600 01/11/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355184P1BS1020010003C14E00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404000P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404007P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400480P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108JAN110351185P1BS10200100012F8600 01/11/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150402019P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108JUN270404004P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403593P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359163P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403002P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150401580P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108APR150358442P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345048P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403585P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAY240355053P1BS1020010002B12F00 05/24/2008 Panchromatic 0.5
WorldView 1 WV0108APR280357193P1BS1020010002828500 04/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352341P1BS102001000159B500 02/19/2008 Panchromatic 0.5
73

WorldView 1 WV0108JUN270404029P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355195P1BS1020010003C14E00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403582P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359183P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358309P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359122P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358331P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352317P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404022P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0107DEC250342432P1BS1020010001D43200 12/25/2007 Panchromatic 0.5
WorldView 1 WV0108JUN270404026P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108OCT040358482P1BS1020010004180300 10/04/2008 Panchromatic 0.5
WorldView 1 WV0108OCT210400509P1BS10200100041AE700 10/21/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359204P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108OCT040358481P1BS1020010004180300 10/04/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403034P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352483P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358353P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108MAY240355037P1BS1020010002B12F00 05/24/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355142P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352436P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150401521P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108JUL060355405P1BS10200100039C1B00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0109FEB130406312P1BS1020010006CFCD00 02/13/2009 Panchromatic 0.5
WorldView 1 WV0108DEC110407308P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070403336P1BS1020010004C0DD00 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355146P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0109MAR150401540P1BS1020010005E28800 03/15/2009 Panchromatic 0.5
WorldView 1 WV0108DEC030355149P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC240402442P1BS1020010005359A00 12/24/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358464P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070343328P1BS10200100013BE500 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108APR150358375P1BS10200100019CA700 04/15/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403539P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404036P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403541P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344556P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403546P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352412P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400526P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403491P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355093P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358347P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355184P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
74

WorldView 1 WV0108DEC030355095P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070342471P1BS10200100017C9B00 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344579P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352459P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359142P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355160P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100359267P1BS1020010002623800 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359101P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355182P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355164P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403495P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403571P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352304P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355166P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403528P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070404131P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403550P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108NOV030358038P1BS10200100049EA600 11/03/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355355P1BS10200100039C1B00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403498P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070403317P1BS1020010004C0DD00 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403527P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355358P1BS10200100039C1B00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404252P1BS10200100032E6500 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400350P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070342490P1BS10200100017C9B00 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108JAN070343305P1BS10200100013BE500 01/07/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355178P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352364P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359224P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403051P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0108OCT210401191P1BS1020010003669B00 10/21/2008 Panchromatic 0.5
WorldView 1 WV0108DEC030355177P1BS102001000582BF00 12/03/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403520P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAR200359081P1BS102001000171A800 03/20/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403560P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403563P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403517P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352463P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403476P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358323P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400457P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403564P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100359285P1BS1020010002623800 06/10/2008 Panchromatic 0.5
75

WorldView 1 WV0108JUN270403509P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108OCT040358302P1BS1020010004180300 10/04/2008 Panchromatic 0.5
WorldView 1 WV0108FEB190352388P1BS102001000159B500 02/19/2008 Panchromatic 0.5
WorldView 1 WV0108MAR030352467P1BS10200100016C1D00 03/03/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344522P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108DEC110407330P1BS1020010005C99600 12/11/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100358553P1BS1020010002959900 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270403473P1BS102001000364B000 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345025P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108NOV070403298P1BS1020010004C0DD00 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUL060355381P1BS10200100039C1B00 07/06/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358300P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0107NOV290339224P1BS1020010001827600 11/29/2007 Panchromatic 0.5
WorldView 1 WV0108MAY240354597P1BS1020010002B12F00 05/24/2008 Panchromatic 0.5
WorldView 1 WV0107NOV290339223P1BS1020010001827600 11/29/2007 Panchromatic 0.5
WorldView 1 WV0108NOV070404155P1BS1020010004272400 11/07/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040357360P1BS1020010004CB0600 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280345002P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100358570P1BS1020010002959900 06/10/2008 Panchromatic 0.5
WorldView 1 WV0109JAN100404432P1BS102001000514BD00 01/10/2009 Panchromatic 0.5
WorldView 1 WV0108JUN100358163P1BS1020010002512700 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108FEB280344533P1BS1020010001E42200 02/28/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403068P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0108JUL230359540P1BS102001000372F800 07/23/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401482P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100358172P1BS1020010002512700 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108SEP040358324P1BS102001000338D500 09/04/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401355P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100359248P1BS1020010002623800 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404200P1BS10200100032E6500 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108JUL190352500P1BS1020010003702600 07/19/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401373P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400434P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108JUL230359560P1BS102001000372F800 07/23/2008 Panchromatic 0.5
WorldView 1 WV0108JUL230359482P1BS102001000372F800 07/23/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401501P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100358536P1BS1020010002959900 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100402017P1BS10200100039A7E00 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUL230359521P1BS102001000372F800 07/23/2008 Panchromatic 0.5
WorldView 1 WV0108JUL230359464P1BS102001000372F800 07/23/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100358146P1BS1020010002512700 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108OCT080405032P1BS1020010004580A00 10/08/2008 Panchromatic 0.5
WorldView 1 WV0108JUN270404217P1BS10200100032E6500 06/27/2008 Panchromatic 0.5
WorldView 1 WV0108MAY240354577P1BS1020010002B12F00 05/24/2008 Panchromatic 0.5
76

WorldView 1 WV0108JUL100401519P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108AUG220400411P1BS1020010002D1C600 08/22/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100358154P1BS1020010002512700 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403084P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0108MAY160340472P1BS102001000254F900 05/16/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100402035P1BS10200100039A7E00 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403101P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401409P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUL230359502P1BS102001000372F800 07/23/2008 Panchromatic 0.5
WorldView 1 WV0108JUN070330554P1BS1020010002954A00 06/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401464P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN070330519P1BS1020010002954A00 06/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUN070330536P1BS1020010002954A00 06/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUN070331015P1BS10200100024EBF00 06/07/2008 Panchromatic 0.5
WorldView 1 WV0108JUN100359230P1BS1020010002623800 06/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN070331033P1BS10200100024EBF00 06/07/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403117P1BS1020010002B62000 05/15/2008 Panchromatic 0.5
WorldView 1 WV0108MAY240354557P1BS1020010002B12F00 05/24/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401427P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUL100401446P1BS1020010002D3A400 07/10/2008 Panchromatic 0.5
WorldView 1 WV0108JUN070331050P1BS10200100024EBF00 06/07/2008 Panchromatic 0.5
WorldView 1 WV0108MAY150403134P1BS1020010002B62000 05/15/2008 Panchromatic 0.5

*Sensor Bands: SPOT 5: Panchromatic, 450-800 nm; Blue, 450-510 nm; Green, 510-580 nm;
Red, 655-690 nm; Near Infra Red, 780-920 nm; Shortwave IR, 1,580-1,750 nm; QuickBird 2:
Blue, 450-520 nm; Green, 520-600 nm; Red, 630-690 nm; Near Infra Red, 760-900 nm;
IKONOS 1: Blue, 450-520 nm; Green, 520-600 nm; Red, 625-695 nm; Near Infra Red
760-900 nm; WorldView 1: Panchromatic, 400-900 nm.

77

Table S2. Variables used in the Appendix.
Paramete
r
Description Equation
Introduced
Reference
p(A
L
) probability density function S1 Malamud et al., 2004
A
L
landslide area S1
N number of landslides in the
inventory
S1
δA landslide area bin S1
δN number of landslides in each bin S1
a inverse gamma function parameter S2 Malamud et al., 2004
s inverse gamma function parameter S2 Malamud et al., 2004
ρ inverse gamma function parameter S2 Malamud et al., 2004
Γ(ρ) gamma function of ρ S2 Malamud et al., 2004
V
i
total landslide volume of inventory i S3
α coefficient in landslide area-volume
relation
S3 Larsen et al., 2010; Guzzetti
et al., 2009; Parker et al.,
2011
γ coefficient in landslide area-volume
relation
S3 Larsen et al., 2010; Guzzetti
et al., 2009; Parker et al.,
2011
X ratio of distance to surface rupture
over fault width
S8
b rupture plane dislocation S8
u rupture surface vertical displacement S8
θ dip angle of thrust faults S8
x distance to surface rupture S9
W fault width S9
L fault length S9
D rupture surface displacement S12
A rupture area S12
M
w
earthquake moment magnitude S13

Earthquake recurence rate S13 Gutenburg and Richter,
1954
a
N
coefficient in Gutenberg-Richter
relation
S13 Gutenburg and Richter,
1954
b
N
coefficient in Gutenberg-Richter
relation, global average b
N
= 0.9
S13 Gutenburg and Richter,
1954
V
L
total landslide volume S14
dip angle-averaged uplift volume S15
Θ theta function, averaged the dip
angle associated term over a given
range
S15
CE
N

U
V
78

coseismic landslide volume
reduction rate
S16

coseismic landslide volume addition
rate
S16
a
A
scaling factor in rupture
area-magnitude relation, -3.99± 0.36
(± 1 S.D.)
S19 Wells and Coppersmith,
1994
b
A
scaling factor in rupture
area-magnitude relation, 0.98± 0.06
(± 1 S.D.)
S19 Wells and Coppersmith,
1994
a
D
scaling factor in surface
displacement-magnitude relation,
-0.74± 1.40 (± 1 S.D.)
S19 Wells and Coppersmith,
1994
b
D
scaling factor in surface
displacement-magnitude relation,
0.08± 0.21 (± 1 S.D.)
S19 Wells and Coppersmith,
1994
a
L
scaling factor in landslide
volume-magnitude relation,
-11.26± 0.52 (± 1 S.D.)
S20 Keefer, 1994; Malamud et
al., 2004
b
L
scaling factor in landslide
volume-magnitude relation, 1.42
S20 Keefer, 1994; Malamud et
al., 2004
I
4
number of earthquakes with M ≥ 4
in cosine(latitude)-normalized
1° lat.× 1° long. areas per year
S22 Kossobokov et al.,2000;
Malamud et al., 2004

net crustal volume addition rate S23
net crustal thickening rate S24
A
E
equivalent normalized 1° lat.× 1° long.
area at the equator
S24 Kossobokov et al.,2000;
Malamud et al., 2004
M
wmax

M
wmin

upper and lower bounds of
earthquake moment magnitudes in
specified area and time
S25
v individual landslide volume S26
pi parameter i S27
s(pi) sensitivity of parameter i S27
D' mean displacement on the rupture
plane

M
0
earthquake seismic moment

h

V


L
V

U
V

79

Table S3. Values and errors of the total landslide volume for the coseismic Wenchuan
landslide inventories (this study and Parker et al. (2011)) using different methods.
a: linear error propagation; b: 5000 MC samples; c: 10,000 MC samples; d: 50,000 MC
samples. P1 refers to the parameters calculated from an ordinary least-square regression on
the Longmen Shan landslide field measurement dataset (Parker et al., 2011); P2 refers to the
parameters from a robust regression on the same dataset.
log
10
α γ V olume
(km
3
)
a

V olume
(km
3
)
b

V olume
(km
3
)
c

V olume
(km
3
)
d

Reference
Landslides from this
study

L1 (global landslides) -0.84± 0.02 1.33± 0.01 1.80+0.1
7/-0.15
1.80+0.1
2/-0.11
1.80+0.12/-
0.11
1.80+0.12/-
0.11
Larsen et al.,
2010
L2 (global bedrock
landslides)
-0.73± 0.06 1.35± 0.01 2.81+0.8
0/-0.62
2.80+0.4
0/-0.34
2.81+0.40/-
0.34
2.81+0.39/-
0.35
Larsen et al.,
2010
L3 (mixed Himalayan
landslides
-0.59± 0.03 1.36± 0.01 4.33+0.8
6/-0.72
4.32+0.5
4/-0.48
4.33+0.56/-
0.47
4.34+0.55/-
0.49
Larsen et al.,
2010
G (global landslides) -1.131 1.45± 0.01 3.49+0.3
9/-0.35
3.48+0.4
0/-0.34
3.49+0.39/-
0.35
3.49+0.39/-
0.35
Guzzetti et
al., 2009
P1 (Longmen Shan
landslides)
-0.97± 0.37 1.39± 0.09 2.45+13.
16/-2.05
2.42+6.2
5/-1.74
2.46+6.64/-
1.79
2.50+6.72/-
1.81
Parker et al.,
2011
P2 (Longmen Shan
landslides)
-0.99± 0.37 1.39± 0.09 2.45+13.
15/-2.05
2.45+6.8
6/-1.78
2.48+6.67/-
1.80
2.48+6.62/-
1.79
This study
80

Landslides from Parker et al. (2011)
L1 (global landslides) -0.84± 0.02 1.33± 0.01 5.73+0.6
2/-0.56
5.73+0.4
6/-0.42
5.72+0.47/-
0.43
5.73+0.46/-
0.42
Larsen et al.,
2010
L2 (global bedrock
landslides)
-0.73± 0.06 1.35± 0.01 9.36+3.0
0/-2.26
9.41+1.5
0/-1.35
9.41+1.53/-
1.35
9.38+1.56/-
1.34
Larsen et al.,
2010
L3 (mixed Himalayan
landslides
-0.59± 0.03 1.36± 0.01 14.86+3.
44/-2.79
14.90+2.
29/-2.00
14.84+2.28/
-1.99
14.86+2.29/
-2.00
Larsen et al.,
2010
G (global landslides) -1.13 1.45± 0.01 15.21+2.
09/-1.83
15.26+2.
09/-1.89
15.21+2.13/
-1.80
15.23+2.07/
-1.83
Guzzetti et
al., 2009
P1 (Longmen Shan
landslides)
-0.97± 0.37 1.388± 0.08
7
9.07+63.
53/-7.90
9.73+32.
43/-7.50
9.21+31.42/
-7.02
8.98+31.10/
-6.85
Parker et al.,
2011
P2 (Longmen Shan
landslides)
-0.99± 0.37 1.39± 0.09 9.14+64.
14/-7.96
9.35+31.
81/-7.23
9.05+31.71/
-6.93
9.19+31.28/
-7.08
This study

81

Table S4. Sensitivity test results. The upper rows are the results for total landslide volume on
parameters α and γ; data sources are listed in Table S3. The lower rows are the results for the
coseismic volume budget on parameters V
L
, a
A
, a
A
, b
A
, a
D
, b
D
and θ.
Parameters G1 G2 G3 L P
s(α) 0.161 0.443 0.166 0.000 0.275
s(γ) 0.839 0.557 0.834 1.000 0.725
Parameters V
L
a
A
b
A
a
D
b
D
θ
s 0.059 0.028 0.035 0.428 0.426 0.024

___________________________________________________________________________

Abstract (if available)

Abstract Here we assess earthquake volume balance and the growth of mountains in the context of a new landslide inventory for the Mw7.9 Wenchuan earthquake in central China. Coseismic landslides were mapped from high-resolution remote imagery using an automated algorithm and manual delineation, which allows us to distinguish clustered landslides that can bias landslide volume calculations. Employing a power-law landslide area-volume relation, we find that the volume of landslide-associated mass wasting (~2.8+0.9/-0.7 km³) is lower than previously estimated (~5.7-15.2 km³) and comparable to the volume of rock uplift (~2.6±1.2 km³) during the Wenchuan earthquake. If fluvial evacuation removes landslide debris within the earthquake cycle, then the volume addition from coseismic uplift will be effectively offset by landslide erosion. If all earthquakes in the region followed this volume budget pattern, the efficient counteraction of coseismic rock uplift raises a fundamental question about how earthquakes build mountainous topography. To provide a framework for addressing this question, we explore a group of scaling relations to assess earthquake volume balance. We predict coseismic uplift volumes for thrust-fault earthquakes based on geophysical models for coseismic surface deformation and relations between fault rupture parameters and moment magnitude, Mw. By coupling this scaling relation with landslide volume-Mw scaling, we obtain an earthquake volume balance relation in terms of moment magnitude Mw, which is consistent with the revised Wenchuan landslide volumes and observations from the 1999 Chi-Chi earthquake in Taiwan. Incorporating the Gutenburg-Richter frequency-Mw relation, we use this volume balance to derive an analytical expression for crustal thickening from coseismic deformation based on an index of seismic intensity over a defined area. This model yields reasonable rates of crustal thickening from coseismic deformation (e.g.~0.1-0.5 km Ma⁻¹ in tectonically active convergent settings), and implies that moderate magnitude earthquakes (Mw≈6-7) are likely responsible for most of the coseismic contribution to rock uplift, because of their smaller landslide-associated volume reduction. Our first-order model does not consider a range of factors (e.g., lithology, climate conditions, epicentral depth and tectonic setting), nor does it account for viscoelastic effects or isostatic responses to erosion, and there are important, large uncertainties on the scaling relationships used to quantify coseismic deformation. Nevertheless, our study provides a conceptual framework and invites more rigorous modeling of seismic mountain building.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Earthquake-driven landsliding, erosion and mountain building: from the eastern Tibetan mountains towards global models

PDF

Heterogeneity of earthquake stress drops, focal mechanisms and active fault zones

PDF

Stress-strain characterization of complex seismic sources

PDF

Tectonic control on landsliding revealed by the 2015 Gorkha earthquake

PDF

Quantifying ground deformation of large magnitude earthquakes using optical imgaging systems

PDF

Multi-scale imaging and monitoring of crustal and fault zone structures in southern California

PDF

Integration and validation of deterministic earthquake simulations in probabilistic seismic hazard analysis

PDF

Extreme events as drivers of landscape evolution in active mountain belts

PDF

Microseismicity, fault structure, & the seismic cycle: insights from laboratory stick-slip experiments

PDF

Detailed properties of seismic waveforms and earthquake sources in Southern California

PDF

Rupture synchronicity in complex fault systems

PDF

The impact of faulting and folding on earthquake cycles in collisional orogens

PDF

Tectonic and kinematic evolution of the Alborán domain during late orogenic extension

PDF

Elements of seismic structures near major faults from the surface to the Moho

PDF

Dynamic rupture processes and seismic radiation in models of earthquake faults separating similar and dissimilar solids

PDF

Analysis of waveform and catalog data of aftershocks for properties of earthquakes and faults

PDF

Volumetric interactions between major ruptures and fault zones illuminated by small earthquake properties

PDF

Multi-scale imaging of the fault zone velocity structure: double-difference tomography, inversion of fault zone headwaves, and fault zone sensitivity kernels

PDF

Critical zone response to perturbation: from mountain building to wildfire

PDF

High-resolution imaging and monitoring of fault zones and shallow structures: case studies in southern California and on Mars

Asset Metadata

Creator Li, Gen (author)

Core Title Landslide inventory associated with the 2008 Wenchuan Earthquake and implications for seismic mountain building

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geological Sciences

Publication Date 10/30/2014

Defense Date 10/30/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag earthquake volume balance,landslide,OAI-PMH Harvest,seismic mountain building,Wenchuan Earthquake

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor West, A. Joshua (committee chair), Berelson, William M. (committee member), Hammond, Douglas E. (committee member)

Creator Email genli@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-510906

Unique identifier UC11297448

Identifier etd-LiGen-3045.pdf (filename),usctheses-c3-510906 (legacy record id)

Legacy Identifier etd-LiGen-3045.pdf

Dmrecord 510906

Document Type Thesis

Format application/pdf (imt)

Rights Li, Gen

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA