Content uploaded by Jon J. Major
Author content
All content in this area was uploaded by Jon J. Major on Jun 08, 2016
Content may be subject to copyright.
Geological Society of America
3300 Penrose Place
P.O. Box 9140
Boulder, CO 80301
(303) 447-2020 • fax 303-357-1073
www.geosociety.org
This PDF file is subject to the following conditions and restrictions:
Copyright © 2004, The Geological Society of America, Inc. (GSA). All rights reserved.
Copyright not claimed on content prepared wholly by U.S. government employees within scope
of their employment. Individual scientists are hereby granted permission, without fees or further
requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in other
subsequent works and to make unlimited copies for noncommercial use in classrooms to further
education and science. For any other use, contact Copyright Permissions, GSA, P.O. Box 9140,
Boulder, CO 80301-9140, USA, fax 303-357-1073, editing@geosociety.org. GSA provides this
and other forums for the presentation of diverse opinions and positions by scientists worldwide,
regardless of their race, citizenship, gender, religion, or political viewpoint. Opinions presented
in this publication do not reflect official positions of the Society.
69
Jibson, R.W., Crone, A.J., Harp, E.L., Baum, R.L., Major, J.J., Pullinger, C.R., Escobar, C.D., Martínez, M., and Smith, M.E., 2004, Landslides triggered by the 13 Janu-
ary and 13 February 2001 earthquakes in El Salvador, in Rose, W.I., Bommer, J.J., López, D.L., Carr, M.J., and Major, J.J., eds., Natural hazards in El Salvador: Boulder,
Colorado, Geological Society of America Special Paper 375, p. 69–88. For permission to copy, contact editing@geosociety.org. © 2004 Geological Society of America
Geological Society of America
Special Paper 375
2004
Landslides triggered by the 13 January and
13 February 2001 earthquakes in El Salvador
Randall W. Jibson*
Anthony J. Crone
Edwin L. Harp
Rex L. Baum
U.S. Geological Survey, Box 25046, MS 966, Denver Federal Center, Denver, Colorado 80225, USA
Jon J. Major
U.S. Geological Survey, 1300 SE Cardinal Court, Suite 100, Vancouver, Washington 98683, USA
Carlos R. Pullinger
C. Demetrio Escobar
Mauricio Martínez
Servicio Nacional de Estudios Territoriales, Km 5 1/2 Carretera a Santa Tecla,
Avenida Las Mercedes, Edifi cio ISTA, San Salvador, El Salvador
Mark E. Smith
U.S. Geological Survey, Box 25046, MS 415, Denver Federal Center, Denver, Colorado 80225, USA
ABSTRACT
During a one-month period in early 2001, El Salvador experienced two devastating
earthquakes. On 13 January, a M-7.7 earthquake centered ~40 km off the southern coast in
the Pacifi c Ocean caused widespread damage and fatalities throughout much of the coun-
try. The earthquake triggered thousands of landslides that were broadly scattered across
the southern half of the country. The most damaging landslide, a rapidly moving mass of
~130,000 m3, occurred in the Las Colinas neighborhood of Santa Tecla, where ~585 people
were killed. Another large landslide (~750,000 m3) near the city of San Vicente blocked
the Pan-American Highway for several weeks. One month later, on 13 February, a M-6.6
earthquake occurred ~40 km east-southeast of San Salvador and triggered additional thou-
sands of landslides in the area east of Lake Ilopango. The landslides were concentrated in a
2500 km2 area and were particularly abundant in areas underlain by thick deposits of poor-
ly consolidated, late Pleistocene and Holocene Tierra Blanca rhyolitic tephras erupted from
Ilopango caldera. Most of the triggered landslides were relatively small, shallow failures,
but two large landslides occurred that blocked the El Desagüe River and the Jiboa River.
The two earthquakes triggered similar types of landslides, but the distribution of trig-
gered landslides differed because of different earthquake source parameters. The large-
magnitude, deep, offshore earthquake triggered broadly scattered landslides over a large
region, whereas the shallow, moderate-magnitude earthquake centered within the country
triggered a much smaller, denser concentration of landslides. These results are signifi cant
in the context of seismic-hazard mitigation for various earthquake scenarios.
Keywords: seismic hazards, landslides, El Salvador, landslide dams.
*jibson@usgs.gov
70 R.W. Jibson et al.
INTRODUCTION
During a one-month period in January and February 2001,
El Salvador experienced two major earthquakes that caused
widespread damage and fatalities throughout much of the coun-
try. On 13 January 2001, a moment-magnitude (M)-7.7 earth-
quake occurred ~40 km off the southern coast of El Salvador at
a depth of ~60 km beneath the Pacifi c Ocean (Fig. 1). Exactly
one month later, on 13 February 2001, a M-6.6 earthquake
occurred ~40 km east-southeast of San Salvador at a depth of
15 km. Both earthquakes triggered thousands of landslides over
a large part of the country. These landslides, most of which
were fairly shallow (<5 m), caused most of the earthquake
fatalities, destroyed hundreds of structures, blocked roads, and
dammed or clogged streams and rivers. Remarkably, this wide-
spread landslide damage was triggered during El Salvador’s
dry season following a below-average rainfall year (Evans and
Bent, this volume, Chapter 3); had the earthquakes occurred
during the wet season, landslide damage probably would have
been considerably worse.
Landslides are a common occurrence in the geologic setting
of El Salvador, a country situated in a very active tectonic envi-
ronment where the Cocos plate is being subducted beneath the
Caribbean plate (Bommer et al., 1996). Rising magma created by
the subduction process has produced a chain of active volcanoes
throughout El Salvador, which have erupted episodically and
deposited widespread, poorly consolidated tephra in many parts
of the country. The combination of active tectonism, steep topog-
raphy, relatively young, weak volcanic rocks, a warm, humid,
subtropical climate with heavy rains, and relatively frequent
moderate and large earthquakes all contribute to an environment
prone to landsliding.
Some tephra deposits that are widespread in El Salvador
(described subsequently) are particularly susceptible to land-
sliding during earthquakes. In fact, Bommer et al. (2002) found
that, for a given earthquake magnitude, both the numbers of
seismically triggered landslides and the areas affected by those
landslides tend to be greater in the geologic environment of El
Salvador than in different geologic, geomorphic, and climatic
environments. Historical records show that El Salvador has suf-
fered damage from earthquake-triggered landslides in at least ten
earthquakes since 1857 (Rymer and White, 1989).
Soon after the 2001 earthquakes, the U.S. Geological Sur-
vey (USGS) responded to requests for technical assistance by
the government of El Salvador and dispatched teams of scien-
tists to El Salvador to work with Salvadoran colleagues in mak-
ing rapid hazard assessments and mitigation recommendations.
Bommer et al. (2002) likewise investigated and documented
post-earthquake effects, including landslides. Preliminary
reconnaissance reports and hazard assessments by the USGS
teams were published soon after the earthquakes (Baum et al.,
2001; Harp and Vallance, 2001; Jibson and Crone, 2001). This
paper draws on these early reports to briefl y describe the source
parameters of the two earthquakes, document major landslides
triggered by the earthquakes, and qualitatively compare and
contrast the landslide types and distributions triggered by the
two earthquakes. Because we conducted rapid reconnaissance
and provided on-the-spot hazard assessments, our investiga-
tions were brief and did not permit a comprehensive assess-
ment of landslide effects or detailed investigations of individual
landslides. However, the observations and interpretations in this
paper provide an important documentation of the types of land-
slides triggered, the environments in which they occurred, and
some of their devastating effects.
Terminology in this paper follows Varnes’s (1978) clas-
sifi cation system, which classifi es slope movements by type of
movement and type of material (Table 1). Also following Varnes,
we use the term landslide generically to describe all varieties of
slope movements; thus, landslide can refer to slope movements
such as rock falls, debris fl ows, slumps, etc.
THE 13 JANUARY 2001 EARTHQUAKE
The earthquake of 13 January 2001 occurred at 11:33 a.m.
local time and had a M of 7.7. The earthquake was located ~40 km
off the coast of El Salvador beneath the Pacifi c Ocean (13.049° N,
88.660° W) at an estimated focal depth of 60 km (Fig. 1). The
focal mechanism indicates normal faulting in the overriding
Caribbean plate (U.S. Geological Survey, NEIC, 2001a).
Strong earthquake shaking affected a large part of El Salva-
dor. Most of the southern two-thirds of the country had ground
accelerations of at least 0.1 g, and several strong-motion stations
in south-central El Salvador recorded peak accelerations of 0.3 g
or greater. The maximum recorded shaking occurred in La Lib-
ertad along the southern coast (Fig. 2), where an acceleration of
1.11 g was recorded (Bommer et al., 2002).
The earthquake affected more than 1.3 million people and
caused 844 deaths and 4723 injuries; more than 277,000 houses
were damaged or destroyed (U.S. Agency for International
Development, 2001). Landslides accounted for most of the earth-
quake damage and fatalities.
Overview of Landslides Triggered by the Earthquake
The 13 January earthquake triggered widespread damaging
landslides in the southern half of El Salvador. Landslides also
were reported as far away as west-central Guatemala at epi-
central distances of ~350 km (Bommer et al., 2002). Although
we were unable to accurately locate a landslide limit across the
entire region, we estimated a limit (Fig. 1) enclosing an area of
~25,000 km2 (including parts of Guatemala not shown) based
on aerial and ground reconnaissance and landslides reported by
Bommer et al. (2002). Within this landslide limit is an area of
more concentrated landslides encompassing ~5000 km2 (Fig. 1).
The most signifi cant concentrations of landslides occurred on
slopes around the periphery of San Salvador, in the Cordillera
Bálsamo region west and south of San Salvador, in areas around
Lake Ilopango and Lake Coatepeque, and on steep fl anks of
Landslides triggered by the 2001 earthquakes in El Salvador 71
10 October 2002
San Salvador
San Salvador
San Salvador
Santa Ana
Santa Ana
Santa Ana
San Miguel
San Miguel
San Miguel
Usulután
Usulután
ta
ta
San
San
Tecla
Tecla
Santa
Tecla Lake Ilopango
Lake Ilopango
Lake Ilopango
Lake Coatepeque
Lake Coatepeque
Lake Coatepeque
HONDURAS
HONDURAS
HONDURAS
GUATEMALA
GUATEMALA
GUATEMALA
EL SALVADOR
EL SALVADOR
EL SALVADOR
PACIFIC OCEAN
Usulut
Usulut
‡n
Volcano
Volcano
San Salvador
San Salvador
Volcano
Volcano
San Salvador
Volcano
Usulut‡n
Volcano
San Vicente
San Vicente
Volcano
Volcano
San Vicente
Volcano
km
02040
13 Jan 2001
13 Feb 2001
13 Feb 2001
13 Feb 2001
90˚00´
89˚00´
88˚00´
13˚00´
14˚00´
Fig. 2
Fig. 15
Figure 1. Map of El Salvador showing epicenters (stars) of 13 January (M-7.7) and 13 February (M-6.6) earthquakes; short-dashed line is approximate limit of landslides triggered by
13 January earthquake; long-dashed line outlines area of concentrated landslides from the 13 January earthquake; black lines are roads; gray lines are rivers and streams; large rectangle
outlines area of Figure 2; small rectangle outlines area of Figure 15.
72 R.W. Jibson et al.
some volcanoes in the southern part of the country, particularly
Usulután volcano (Figs. 1 and 2).
Most of the triggered landslides were relatively small (tens
to hundreds of cubic meters), shallow (<5 m) falls and slides in
surfi cial rock and debris. Landslide concentrations were greatest
where two types of Pleistocene and Holocene volcanic rocks crop
out: (1) relatively soft, weak pyroclastic deposits and (2) solid,
indurated rocks that originated as lava fl ows. The largest num-
ber of landslides occurred in pyroclastic deposits; these slides
tended to be highly disrupted masses of rock and earth that fell
and slid into jumbled piles of landslide debris. Earthquake-trig-
gered landslides in this type of material have been documented
in many previous earthquakes in El Salvador (Rymer and White,
1989). Landslides originating in the harder lava fl ows were volu-
metrically small but very hazardous; they consisted primarily of
boulders as large as several tens of cubic meters that were shaken
loose from steep outcrops and then bounced and rolled down
steep slopes. These boulders caused great damage when they hit
buildings, vehicles, or people.
The January earthquake also caused two large, deep (tens of
meters) landslides. The most damaging was the Las Colinas land-
slide in Santa Tecla, a western suburb of the capital city of San
Salvador (Fig. 2). The Las Colinas landslide had an estimated
volume of 130,000 m3 (Evans and Bent, this volume, Chapter 3)
and caused ~585 fatalities when it slid off the north slope of
Bálsamo Ridge; the landslide had an abnormally long runout dis-
tance and destroyed everything in its path in this densely popu-
lated neighborhood. A second large landslide near San Vicente
(Fig. 2), which we estimate had a volume of ~750,000 m3, buried
a few hundred meters of the Pan-American Highway under tens
of meters of debris.
The January earthquake also triggered liquefaction and
associated lateral spreading along coastal areas from La Libertad
eastward to the mouth of the Lempa River and along riverbanks
east of Usulután (Fig. 2). Because our mission focused on assess-
ing landslide hazards, we did not investigate liquefaction areas in
detail but merely noted some locations where liquefaction effects
were observed.
Signifi cant Landslides and Areas of Landslide Activity
Cordillera Bálsamo
The January earthquake triggered widespread landsliding
in the Cordillera Bálsamo, a broad, deeply dissected upland area
southwest of San Salvador. Bálsamo Ridge defi nes the northern
boundary of the Cordillera Bálsamo, and the ridge separates the
cordillera to the south from a broad fl at valley the north, which is
occupied by Santa Tecla (Fig. 2). Cordillera Bálsamo is underlain
by the Bálsamo Formation (Weber et al., 1978), which consists of
volcanic breccias, lavas, and other well-indurated volcanic rocks.
In the area near San Salvador, the top of the Bálsamo Formation
is marked by weathered soil that contains suffi cient clay and
fi ne-grained material to act as a groundwater barrier that locally
perches water in porous, overlying, young volcanic deposits. Such
perched groundwater conditions probably contribute to seismic
instability because high pore-water pressures can develop in satu-
rated layers during earthquake shaking (Harp et al., 1984).
In central El Salvador, the Bálsamo Formation is overlain
by a sequence of latest Pleistocene and Holocene rhyolitic and
andesitic tephras erupted from San Salvador Volcano and Ilo-
pango caldera (Rose et al., 1999). Four named deposits were
erupted from the Ilopango caldera, an 8 × 11 km depression
that is now fi lled by Lake Ilopango (Fig. 1). The four deposits,
named Tierra Blanca Joven (TBJ) and TB2–TB4, erupted from
the caldera since 56.9 ka (Rose et al., 1999). The youngest unit,
TBJ, was deposited about A.D. 430 (Dull et al., 2001). The oldest
unit (TB4) buried the preexisting landscape locally to depths of
2 m or more; TB4 is overlain by younger, interstratifi ed tephra
from Ilopango caldera and San Salvador Volcano along Bálsamo
Ridge. The thickness of the TB4 and younger deposits can vary
considerably over short distances because they were deposited on
and fl owed over preexisting topography; the deposits are com-
monly thicker in ancient valleys compared to adjacent ridges.
The generalized stratigraphy of post–Bálsamo Formation depos-
its in the Bálsamo Ridge area is shown in Table 2.
The Ilopango Tierra Blanca tephras are particularly prone
to landsliding (Rymer and White, 1989; Bommer and Rodríguez,
TABLE 1. LANDSLIDE CLASSIFICATION SYSTEM
Type of material
Bedrock Engineering soils
Type of movement
Primarily coarse Primarily fine
Falls Rock fall Debris fall Earth fall
Topples Rock topple Debris topple Earth topple
Rotational Rock slump Debris slump Earth slump
Slides Translational Rock slide Debris slide Earth slide
Lateral spreads Rock spread Debris spread Earth spread
Flows Rock flow Debris flow Earth flow
Complex Combination of two or more principal types of movement
Note: Modified from Varnes (1978).
Landslides triggered by the 2001 earthquakes in El Salvador 73
P
a
n
A
m
e
r
i
c
a
n
H
i
g
h
w
a
y
Santa
Santa
Tecla
Tecla
Santa
Tecla
San Salvador
San Salvador
San Salvador
Lake Ilopango
PA CIFIC OCEAN
Usulután
La Libertad
San Vicente
Zaragoza
Zaragoza
Zaragoza
San Pedro
Masahuat
Berlin
Alegria
Santiago
Santiago
de Maria
de Maria
Santiago
de Maria
C
o
r
d
i
l
l
e
r
a
B
a
l
s
a
m
o
Comasagua
Comasagua
J
i
b
o
a
R
i
v
e
r
L
e
m
p
a
R
i
v
e
r
CA
4
CA
1
Figs. 3, 4
Fig. 5 Fig. 7
Fig. 6
Fig.8
Fig.14
Fig. 9
Fig.10
Fig. 11
Fig. 12
Fig. 13
020
km
Comasagua
Figure 2. Map showing area of greatest landslide concentration from 13 January earthquake. Location shown in Figure 1. Black lines are roads; gray lines are rivers and streams; airplane
symbol shows location of international airport. Locations of photographs shown in Figures 3–14 shown by arrows.
74 R.W. Jibson et al.
2002; Konagai et al., this volume, Chapter 4; Rolo et al., this
volume, Chapter 5). Weak cementation and generally nega-
tive pore-water pressures provide strength for these deposits
to be stable under most conditions; however, when saturated
by heavy rainfall or shaken during strong earthquakes, they
can lose strength and collapse (Bommer and Rodríguez, 2002;
Bommer et al., 2002; Rolo et al., this volume). Our reconnais-
sance indicated that the distribution of abundant earthquake-
induced landslides in the Cordillera Bálsamo generally coin-
cided with the presence of relatively thick deposits of TB4 and
overlying pyroclastic sediment.
Our aerial and ground reconnaissance showed that the earth-
quake triggered thousands of landslides throughout much of the
Cordillera Bálsamo, most of which were shallow (<5 m thick),
disrupted earth or debris slides and rock falls. Keefer (1984) clas-
sifi ed these types of earthquake-triggered landslides as “disrupted
slides and falls” because they tend to involve materials that are
unconsolidated, very weakly cemented, or highly fractured, and
the triggered landslides disaggregate and form deposits of highly
disrupted material. Much of the Cordillera Bálsamo is blanketed
with fi ne, weakly cemented tephra that collapsed and disaggre-
gated during the strong earthquake shaking. Collapse and disag-
gregation of tephra deposits left thick accumulations of loose,
fi ne debris on the ground surface. Many steep-walled ridges in
this region were thoroughly shattered by amplifi ed shaking of the
weak surfi cial deposits. Fissuring and incipient landsliding along
the edges of such ridges were widespread.
The road from Santa Tecla to Comasagua traverses the
Cordillera Bálsamo (Fig. 2). Along most of its length, this road
follows the crest of a narrow, steep-sided ridge where ground
shaking was amplifi ed; the ridge is thoroughly shattered along
most of its length (Fig. 3). The shattering, in turn, led to wide-
spread slope failures that produced large volumes of landslide
material. The ground surface along much of the ridge was fi s-
sured, road cuts had failed extensively, and nearly all man-made
structures were damaged or destroyed. The shaking left the road
in very poor condition; disaggregated tephra covered the road
with as much as 0.5 m of fi ne, powdery dust, and loose sediment
continued to accumulate from the failed road cuts (Fig. 4). At the
western edge of Santa Tecla, the road crossed a large, deep debris
slide similar in size and scale to the Las Colinas landslide, which
is described below.
Las Colinas Landslide
By far the most devastating landslide triggered by the January
earthquake was the Las Colinas landslide (GPS: 13° 39.662′ N,
89° 17.188′ W) that cascaded off the steep northern fl ank of
Bálsamo Ridge. The landslide destroyed hundreds of houses and
killed ~585 people, the largest loss of life in one location caused
by this earthquake. Detailed descriptions and analyses of the Las
Colinas landslide are the subject of two other papers in this vol-
ume (Evans and Bent, Chapter 3; Konagai et al., Chapter 4), and
so we provide only a brief overview.
The Las Colinas landslide, a rapid earth fl ow having an
estimated volume of 130,000 m3, originated at an altitude of
1075 m on Bálsamo Ridge and traveled 735 m northward
through the Las Colinas neighborhood of Santa Tecla (Fig. 5);
the landslide dropped ~165 m from source to terminus (Evans
and Bent, this volume). Evans and Bent (this volume) estimate
TABLE 2. GENERALIZED STRATIGRAPHY OF LASTEST
QUATERNARY VOLCANIC DEPOSITS OF BALSAMO RIDGE
Depth
(m)
Deposit
0–2 Stratified, loose, basaltic tephra from San Salvador
Volcano.
2–4 Massive, tan to orange-brown, weathered ash.
4–6 Interstratified, dark, loose basaltic tephra, brown
weathered massive ash, and loose volcanic lapilli.
6+ TB4 volcanic ash, white, soft to loose, well sorted.
Figure 3. Shattered ridge near Comasagua (see Fig. 2 for location).
Figure 4. Road-cut failures near Comasagua (see Fig. 2 for location).
Landslides triggered by the 2001 earthquakes in El Salvador 75
that the overall landslide velocity was ~15 m/s and that it may
have reached velocities as great as 30 m/s in some places. The
near-vertical scarp at the landslide headwall exposed the upper
25–30 m of material involved in the landslide, which appeared
to be only somewhat moist at the time of the landslide (Evans
and Bent, this volume). Deposits of the Cuscatlán Formation
and the Tierra Blanca units of the San Salvador Formation,
typical of deposits covering Bálsamo Ridge and the Cordillera
Bálsamo (Weber et al., 1978), were exposed in the main scarp
(Evans and Bent, this volume).
A strong-motion seismometer located on the valley bot-
tom in Santa Tecla, 1.4 km from the Las Colinas landslide
source, recorded a peak ground acceleration of 0.5 g (Bommer
et al., 2002), a very high level of shaking considering the site
is ~100 km from the earthquake source. Evidence along the
crest of Bálsamo Ridge, however, indicated that shaking there
probably was even greater. We saw trees snapped off, boulders
thrown vertically and then laterally from their sockets, and deep
fi ssures along the edge of the ridge, all of which suggest shak-
ing at or above 1 g. Such amplifi cation of shaking is common
along the edges of steep ridges (Ashford et al., 1997). This
very strong shaking, along with the thick, loose to poorly con-
solidated, volcaniclastic deposits, the steep topography on the
northern fl ank of Bálsamo Ridge, and possibly the presence of a
relatively impermeable soil at the top of the Bálsamo Formation
were contributing factors to the Las Colinas landslide.
Zaragoza
In and around Zaragoza (Fig. 2), rock falls occurred where
fracturing in lava fl ows formed large boulders that were suscepti-
ble to being shaken loose during earthquakes. A road cut on high-
way CA-4 north of Zaragoza (GPS: 13° 36.176′ N, 89° 17.161′
W) produced one such damaging rock fall that had a volume of
~2500 m3 (Fig. 6). One house was undermined in the initial rock
fall, and several additional houses were perched precariously
along the newly formed scarp. The rock below these houses
was extensively fractured and dilated by earthquake shaking and
posed signifi cant risk of failing in aftershocks, heavy rainfall, or
future large earthquakes.
Santa Elena
Rock falls involving large boulders occurred in the Santa
Elena neighborhood in the southwestern part of San Salvador
(Fig. 2). A steep ridge that has several outcrops of massive, hard
lava fl ows extends along the southern edge of Santa Elena (GPS:
13° 39.639′ N, 89° 15.711′ W), and these rock outcrops produced
rock falls during the earthquake. Fractures in the outcrops create
rock blocks 1–20 m3 in volume that can be shaken loose during
earthquakes; the blocks then roll and bounce down the 30°–40°
slope at high speed. Several such boulders are strewn through the
forest all the way to the base of the slope, indicating that this pro-
cess has occurred previously. During the 13 January earthquake,
a 5 m3 boulder broke from the rock outcrop and rolled downslope
rapidly enough to knock down a 0.5 m diameter tree and travel to
within 30 m of the houses at the base of the slope (Fig. 7). Had
this boulder continued downslope, it would have severely dam-
aged any structure it hit.
The hazard from rock falls in Santa Elena varies with posi-
tion along the ridge. The hazard is greatest on the western end
of the ridge and lessens somewhat eastward because the rock-
fall sources are higher on the slope on the eastern end (boulders
released higher on the slope are more likely to be slowed and
stopped by the thick forest covering the slope). Although heavy
rainfall can also trigger rock falls in this area, the greatest
Figure 5. Las Colinas landslide (view to
southwest) (see Fig. 2 for location).
76 R.W. Jibson et al.
hazard is from rocks shaken loose in moderate or large earth-
quakes. Houses in this area are built directly against the base
of the slope, and houses in the uppermost part of this neighbor-
hood are at signifi cant risk from rock-fall hazards.
La Cima
The earthquake cracked the ground surface and overlying
structures in La Cima, a large housing development built on cut-
and-fi ll pads on the southern outskirts of San Salvador (Fig. 2;
GPS: 13° 39.817′ N, 89° 12.965′ W). During the earthquake,
cracks as wide as 10 cm formed in some parts of La Cima. The
cracks extended through road pavement and through several
houses, causing serious damage (Fig. 8). The cracks appeared
to have formed at or near the boundary of the cut slope and the
fi ll slope, a correlation that has been observed in earthquakes
elsewhere. Cracking along this boundary results from either
differential settlement or an impedance contrast created when
seismic waves encounter the boundary between stiffer intact
material and looser fi ll (Stewart et al., 1996). Renewed crack-
ing or failure of the steep edge of the fi ll in future earthquakes
or rainy periods is possible.
Lake Ilopango and Lake Coatepeque
Steep slopes cut by streams in the area around Lake Ilo-
pango (Figs. 1, 2) produced hundreds of debris falls (Fig. 9)
having volumes as great as several thousand cubic meters (Harp
and Vallance, 2001). Slopes in this area are blanketed with thick
deposits of weak, rhyolitic Tierra Blanca tephra.
The earthquake also triggered dense concentrations of
debris falls in rhyolitic pumice deposits in the area around Lake
Coatepeque (Fig. 1). Tephra layers in this area, however, are not
as thick as those around Lake Ilopango, and so the number and
volumes of landslide were smaller on average.
Las Leonas Landslide
The January earthquake triggered a large debris slide
(Fig. 10) from a high, south-facing road cut on the Pan-American
Highway at Las Leonas (GPS: 13° 39.963′ N, 88° 48.738′ W),
~4 km northwest of San Vicente (Fig. 2). The landslide buried
several cars and killed 12 people. The roadway at the landslide
location was a divided four-lane highway that traversed the
Figure 7. Boulder shaken loose by 13 January earthquake that came
to rest near houses in Santa Elena (see Fig. 2 for location). Measuring
stick is 2 m long. Note scars on tree at right where the boulder tore
bark off the trunk.
Figure 6. Rock fall from road cut north of Zaragoza (see Fig. 2 for
location).
Landslides triggered by the 2001 earthquakes in El Salvador 77
slopes on the northern fl ank of the Istepeque River valley. Our
fi eld observations indicated that the slide was ~250 m wide at
its base, 120 m high, and 25 m thick, yielding an estimated vol-
ume of 750,000 m3.
The main scarp of the landslide formed along the fault plane
of a south-dipping normal fault that bounds a graben that forms
the valley. This fault created a zone of weakened, sheared teph-
ritic material, which facilitated slope failure during earthquake
shaking. We inspected the slope above the crown of the landslide
and did not see open cracks or other evidence of impending fail-
ure of material above the scarp, which suggests that the failure
location was controlled strongly by the preexisting fault-plane
surface in the slope.
The formation of a large landslide at this location can be
attributed primarily to the zone of weakness along the fault plane
that defi ned a large, deep mass of weak volcanic rock that failed
during strong shaking. Also, the toe of the slope was oversteep-
ened by a road cut, which, in turn, was destabilized by people
excavating rocks from the cut to sell as building aggregate.
Usulután Region
The region around Usulután (Figs. 1, 2) experienced wide-
spread but sparsely scattered landslides, primarily rock falls from
road cuts and rock and debris slides from gully walls on the
fl anks of volcanoes.
At a geothermal plant near Berlín (Fig. 2), a rock fall of sev-
eral hundred cubic meters occurred on the steep slope behind one
of the wells (GPS: 13° 30.947′ N, 88° 30.740′ W). The slope that
failed is the scarp of an older, larger landslide. Several boulders
from the rock fall stopped within a few meters of the well equip-
ment; the biggest single boulder was ~5 m3 (Fig. 11). Another
landslide headed just below the road to the well, and the road had
cracks parallel to the edge of the slope.
Several road cuts produced rock falls and rock slides along
roads linking Berlín, Alegría, and Santiago de Maria (Fig. 2). The
larger rock falls had volumes of several thousand cubic meters
(Fig. 12). At least one slope in this area also produced rock falls
during Hurricane Mitch in 1998, which shows that identical
types of landslides in this area can be triggered by heavy rainfall
or earthquake shaking. Strong shaking dilated fractures 10 cm
or more in some of the rock-fall scars, and many precariously
perched boulders were ready to fall.
Near the town of Alegría, a small lake basin in an old
volcanic crater has steep walls tens of meters high. The earth-
quake triggered several rock falls from the crater walls, which
had volumes as large as a few thousand cubic meters (GPS:
13° 29.563′ N, 88° 29.523′ W), and some rock falls temporarily
closed the road around the lake.
Volcanoes in the southern part of the country shed numerous
rock and debris slides from the walls of preexisting gullies. These
failures had volumes of several cubic meters to several thousand
Figure 8. Cracks at the cut-fi ll boundary in La Cima (see Fig. 2 for
location).
Figure 9. Landslides in Tierra Blanca ash deposits near Lake Ilopango
(see Fig. 2 for location).
78 R.W. Jibson et al.
cubic meters. Usulután Volcano had perhaps the highest concen-
tration of landslides (Fig. 13) in this region.
San Pedro Masahuat
Many steep walls of deeply incised gullies around San Pedro
Masahuat (Fig. 2) failed during the earthquake. San Pedro Masa-
huat is built on a thick (tens of meters) deposit of white, powdery
volcanic ash, and deep gorges are eroded into the ash on the edge
of town. These gorges have near-vertical walls 70–80 m high and
are 50 m or more wide. The earthquake triggered massive fail-
ures along walls of the gorges; blocks and slabs of ash hundreds
to thousands of cubic meters in volume collapsed and cascaded
to the bases of the steep slopes (Fig. 14). Numerous cracks and
fi ssures parallel to the edge of the gorge opened as far as 30 m
from the edge. These cracks typically had several centimeters
of displacement downward and toward the bluff edge and thus
defi ned incipient landslides that likely will fail in future rainy
seasons or earthquakes.
THE 13 FEBRUARY 2001 EARTHQUAKE
Exactly one month after the devastating 13 January earth-
quake, a second major earthquake struck El Salvador. The 13
February 2001 earthquake occurred at 8:22 a.m. local time and
had a moment magnitude (M) of 6.6. The earthquake was located
~40 km east-southeast of San Salvador in the area east of Lake
Ilopango (13.671° N, 88.938° W; Figs. 1, 15) at an estimated
focal depth of 15 km. The focal mechanism indicated strike-slip
crustal faulting in the overriding Caribbean plate (U.S. Geologi-
cal Survey, NEIC, 2001b).
Owing to the smaller magnitude and shallower depth of the
February earthquake, the effects and damage were more local-
ized than those in the January event. Nevertheless, the earth-
quake still affected more than 1.5 million people and caused 315
deaths and 3399 injuries. About 57,000 houses were damaged or
destroyed by strong shaking or ground failure (U.S. Agency for
International Development, 2001).
Overview of Landslides Triggered by the Earthquake
The February earthquake triggered thousands of landslides
in an area of perhaps 2500 km2. We could not establish an accu-
rate landslide limit, because to distinguish landslides triggered
Figure 10. Las Leonas landslide block-
ing the Pan-American Highway near
San Vicente (see Figure 2 for location).
Figure 11. Boulders from rock fall above geothermal well near Berlín
(see Fig. 2 for location).
Landslides triggered by the 2001 earthquakes in El Salvador 79
in the February earthquake from those triggered in January was
generally impossible outside the immediate epicentral area. The
most concentrated landsliding occurred in an area of ~1000 km2
(Fig. 15). Thick deposits of late Pleistocene and Holocene
rhyolitic tephras that are particularly susceptible to failure during
seismic shaking (Rymer and White, 1989; Rolo et al., this vol-
ume, Chapter 5) underlie the epicentral area (Weber et al., 1978).
These deposits erode easily and formed steep-sided valleys that
produced thousands of relatively small (<1000 m3), shallow
(<5 m) falls and slides. In places, these landslides were so highly
concentrated that they coalesced into nearly continuous failures
along canyon walls (Fig. 16).
Figure 12. Landslide on road from Ber-
lín to Alegría (see Fig. 2 for location).
Figure 13. Rock and debris slides in gullies on the slopes of Usulután
Volcano (see Fig. 2 for location).
Figure 14. Failures of near-vertical slopes in ash near San Pedro Masa-
huat (see Fig. 2 for location).
80 R.W. Jibson et al.
The earthquake also triggered large landslides in river val-
leys east of Lake Ilopango and at San Vicente Volcano (Fig. 1).
Particularly large landslides occurred in the drainages of the El
Desagüe and Jiboa Rivers near Lake Ilopango (Fig. 15); these
slides dammed the rivers and impounded lakes, which posed a
potentially signifi cant downstream threat. Several large land-
slides also occurred on the northern slopes of San Vicente Vol-
cano. Prominent failures included a large rock slide in the upper
part of the Quebrada Del Muerto and numerous failures along the
steep valley walls in the upper part of the Quebrada El Blanco
and adjacent drainages (Fig. 15).
Earthquake-induced liquefaction and lateral spreading
caused localized damage around the northeastern shore of Lake
Ilopango. The greatest damage from liquefaction and lateral
spreading occurred in San Agustín, ~10 km northwest of the
earthquake epicenter (Fig. 15), where fi ssures and slump scarps
formed on a populated alluvial fan. Buildings that spanned the
fi ssures and scarps suffered considerable damage.
Signifi cant Landslides and Areas of Landslide Activity
Jiboa River Landslide
About 2 km upstream from the confl uence of the El Desagüe
and Jiboa Rivers (Fig. 15), a landslide having an estimated vol-
ume of 12 million m3 dammed the Jiboa River (Fig. 17). The
Jiboa River landslide consists primarily of rhyolitic tephra and
pyroclastic debris of the Cuscatlán Formation (Weber et al.,
1978), and most of the material is poorly compacted and very
porous. The landslide dam is ~700 m long (parallel to the river
drainage), 250 m wide (across the valley), and 60–70 m thick.
The dam impounded a lake that was ~25 m deep, 400 m long,
and fi lling at a rate of less than 0.1 m3/s (<8600 m3/d) by mid-
March 2001. The lake-surface altitude at that time was ~475 m. If
fi lled completely, the lake would have had an estimated maximum
depth of 60 m, a length of ~2 km, and a volume of ~7 million m3.
The surface of the landslide deposit was hummocky and
irregular. Because of the surface irregularity, two possible breach
Lake
Ilopango
Jiboa
E
l
D
e
s
a
g
ü
e
R
i
v
e
r
San Vicente
Volcano
River
+
P
a
n
A
m
e
r
i
c
a
n
H
i
g
h
w
a
y
C
A
1
Guadelupe
San
Vicente
Ver apaz Tepetitán
Cojutepeque
Santo
Tomás
San
Sebastián
San
Bartolo
Tonacatepeque
San
Martín
Guadalupe
13 Feb. 2001
13 Feb. 2001
4
3
1
2
5
89˚00'
13˚35'
13˚40'
13˚45'
88˚50'
510150
Kilometers
EXPLANATION
Major highway
Primary road
Landslide site
Figure 15. Map showing area of great-
est concentration of landslides in the 13
February earthquake. Figure 1 shows
location. Numbered symbols corre-
spond to signifi cant landslide locations:
1—Jiboa River landslide (photo in Fig-
ure 19); 2—El Desagüe River landslide
(photos in Figs. 17 and 18); 3—Que-
brada del Muerto rock slide (photo
in Fig. 20); 4—Quebrada El Blanco
debris slides (photo in Fig. 21); 5—San
Agustín–Quebrada El Chaguite area.
Figure 16. Coalescing failures in ash deposits east of Lake Ilopango
(see Fig. 1 for location of Lake Ilopango).
Landslides triggered by the 2001 earthquakes in El Salvador 81
points existed on the landslide dam: one along the northern
margin of the dam at an altitude of ~510 m and another along
the southern margin of the dam at an altitude of ~515 m. (All
altitudes were estimated from fi eld measurements made with
barometric altimeters and hand-held GPS receivers, which have
considerable inaccuracy.) On the basis of our estimates of the
potential maximum lake volume and stream-gage records of
average fl ow rates for the Jiboa River, we projected that, without
engineering intervention, the lake could fi ll and overtop the dam
within several months.
Initial hazard assessment of landslide dam. The Jiboa
River landslide and the lake it impounded presented a signifi cant
threat to people, property, and infrastructure downstream. If the
lake fi lled to capacity, it would have overtopped the dam and
generated a potentially devastating fl ood. On the basis of a case
study of a similar landslide dam on the Pisque River in Ecuador
(Asanza et al., 1992), we speculated that a catastrophic fl ood
could develop quickly and that the lake could drain in several
hours if it overtopped the dam. To quantify the threat by overtop-
ping of the landslide dam, we estimated a range of values for
peak discharges at the point of breach and the time it would take
to reach peak discharge. We used a model that predicts fl ood
discharges after failures of natural dams (Walder and O’Connor,
1997). The model is based on the kinematics of breach formation
and the hydraulics of fl ow through the breach. Parameters that
must be estimated for the model include drop in lake level (d),
water volume released (V), and breach erosion rate (k). Walder
and O’Connor (1997) show that d is typically 50%–100% of the
dam height and that k generally ranges from 10 to 100 m/h. From
the projected hypsometry of the maximum lake impoundment,
we determined that ~95% of the lake volume (~6.7 million m3)
could empty if a breach eroded the dam by 40 m (~66% of the
dam height). Therefore, by setting d = 40 m, V = 6.7 million m3,
and using bounding values of k, we estimated that the maximum
discharge at the breach could range from ~1000 to 10,000 m3/s
and that peak discharge could occur in less than one hour. Dis-
charges of the Jiboa River were recorded between 1961 and
1996 at the Montecristo gaging station, 20 km downstream from
the dam (División de Meteorológico y Hidrológico, 2001). The
greatest peak discharge recorded during this period was 642 m3/s.
Therefore, the range of estimated peak discharges caused by over-
topping of the landslide dam is ~1.5 to 15 times greater than the
maximum discharge recorded along the Jiboa River valley.
A catastrophic fl ood on the Jiboa River caused by failure of
the landslide dam would follow a complex fl ow path and possi-
bly increase in magnitude downstream. Below the dam, the fl ood
would fl ow violently down a narrow valley to the confl uence
with the El Desagüe River, 2 km downstream. At the confl u-
ence, some fl oodwater would fl ow westward up the El Desagüe
River toward the outlet of Lake Ilopango; the remainder would
turn southward and fl ow down the Jiboa River. Floodwater that
surged up the El Desagüe River would likely overtop and erode
part of the El Desagüe landslide dam (described subsequently).
Erosion of that landslide dam would release some of the water
it impounds, but the maximum volume of water contributed
from that impoundment (~500,000 m3) is less than 10% of
the potential maximum volume released from the Jiboa River
impoundment (~7 million m3). These estimates assume a worst
case in which all of the impounded waters are released, which is
probably unlikely.
Discharge records from gaging stations along the upper and
middle reaches of the Jiboa River can be used to roughly estimate
the travel times of fl ood peaks through the parts of the valley
downstream from the landslide dam (División de Meteorológico
y Hidrológico, 2001). The upstream gage, Jiboa River at San
Ramon, was situated at the site of the Jiboa River landslide, and
the most downstream gage, Jiboa River at Montecristo, was situ-
ated ~20 km downstream from the landslide. Three fl ood peaks
(all less than 200 m3/s) that occurred during the common period
of operation for these gages (1972–80) traveled the 20 km from
the site of the landslide dam to the Montecristo gage in 2–4 h.
Furthermore, the fl ood hydrographs showed little attenuation
between the San Ramon and Montecristo gages because this sec-
tion of the Jiboa River is incised into a bedrock canyon.
Directly downstream from the Montecristo gage site, the
Jiboa River valley widens abruptly, and the river fl ows onto
the broad coastal plain. In the confi ned, narrow bedrock valley
a fl ood would retain much of its velocity, energy, and erosive
power, but in the valley’s broad lower reaches a fl ood would
spread across the coastal plain dissipating much of its energy and
attenuating its peak while inundating large areas of fl ood plain. A
catastrophic fl ood of 1000 to 10,000 m3/s would inundate much
of the river’s fl ood plain and could travel downstream faster than
the smaller, seasonal fl oods.
The impact of a catastrophic fl ood on the lower reaches of
the Jiboa River is diffi cult to assess accurately because of the
complex effects of water fl owing up, and then back down, the
El Desagüe River valley, the subtle relief of the coastal plain,
Figure 17. Landslide dam on the Jiboa River (view southwest). The
impounded lake behind the landslide is at the lower left (see Fig. 15
for location).
82 R.W. Jibson et al.
a lack of accurate topographic data along the fl ood reach, and
the uncertainty in the timing and magnitude of the fl ood peak.
Nevertheless, such a fl ood in the lower Jiboa River valley would
likely affect settlements on the lower fl ood plain, the coastal
highway bridge that spans the river, and perhaps partly inundate
the international airport, which lies adjacent to the Jiboa River
fl ood plain (Fig. 2).
Mitigation of landslide-fl ood hazard. To avoid a poten-
tially catastrophic fl ood along the Jiboa River caused by uncon-
trolled overtopping of the landslide dam, the government of El
Salvador excavated a spillway across the landslide to reduce the
maximum volume of the impounded lake and to provide some
control on the timing of water release. The north breach point
was selected as the most advantageous site for spillway excava-
tion for several reasons: (1) The north breach point coincided
with an existing roadway, providing relatively easy access for
heavy equipment to construct a spillway. (2) A spillway on the
northern side of the landslide dam required less excavation than
a spillway along the southern margin: ~50–60 m long versus
~700 m long. (3) The northern margin of the landslide dam
provided a safer work site. The southern margin of the dam
lies directly beneath the steep (~70°), 200 m high headscarp
of the landslide, where rock falls that occurred nearly continu-
ously would pose a constant hazard to workers. (4) Two breach
points existed along the northern margin of the landslide dam:
an upstream (eastern) point adjacent to the roadway, and a
downstream point several hundred meters to the west. If the
lake overtopped the upstream breach point before completion
of the spillway, the fl oodwater would fl ow into and inundate a
large tributary valley (Quebrada Seca) of the Jiboa River, but
the downstream breach point would prevent this water from
immediately fl owing down the main Jiboa River channel.
The spillway that was eventually excavated across the Jiboa
River landslide dam was 20 m deep, ~100 m long, and ~15 m
wide at the bottom of the cut (Fig. 18). Lake water began fl ow-
ing over the spillway on 1 July 2001. Soon after water began
fl owing over the dam, a 2–3 m deep gully eroded through the
spillway, but the channel rapidly stabilized, and no signifi cant
additional erosion has occurred. Reasons for this stabilization
are unclear and may involve various factors: (1) Precipitation
and surface runoff in El Salvador subsequent to emplacement
of the landslide have been below normal, which possibly con-
tributed to the low initial outfl ow over the spillway and a conse-
quent lack of rapid erosion and catastrophic fl ooding. (2) Even
without signifi cant infl ow, once the lake began to spill over
the landslide, it could have emptied catastrophically if rapid
erosion persisted. The fact that the erosion stabilized suggests
that a resistant lens of material may exist that impeded erosion.
(3) The dam may be porous enough that seepage reduces the
head available for fl ow over the spillway.
As of January 2003, a lake averaging ~15 m deep having an
estimated volume of 2.5 million m3 remained behind the land-
slide dam, which could be stable for a number of years. However,
the erodible nature of the material composing the landslide dam
makes it questionable whether the lake will remain indefi nitely or
whether a rapid breach of the spillway may yet occur in average
or above-average water years when outfl ow over the spillway
will be greater.
Figure 18. Spillway excavated through
the toe of Jiboa River landslide dam
(see Fig. 15 for location).
Landslides triggered by the 2001 earthquakes in El Salvador 83
The spillway was excavated to reduce the magnitude of a
fl ood caused by overtopping of the dam and to provide some
control over the timing of water release. A more diffi cult prob-
lem is the possibility of a dam failure induced by piping, in
which lake water percolates through the dam and erodes con-
duits through which the lake water can drain. As impoundment
depth increased, the potential for piping increased because the
impounded water exerted increasing pressure on the dam. During
and after spillway construction the dam was monitored visually
for seepage indicative of piping and potentially decreasing dam
stability. No seepage had been observed as of January 2003.
El Desagüe River Landslide
The earthquake triggered a second large landslide in the
valley of the El Desagüe River (Fig. 15). The El Desagüe River
drains eastward out of Lake Ilopango and is a tributary to the Jiboa
River. Approximately 6.5 km east of the outlet of Lake Ilopango,
a landslide consisting of coarse, bouldery debris and having an
estimated volume of 1.5 million m3 dammed the El Desagüe River
and impounded a shallow lake (Fig. 19). The landslide originated
from the south wall of the valley and dammed the river ~100 m
upstream from its confl uence with the Jiboa River. About 70% of
the landslide deposit consists of andesitic breccia of the Bálsamo
Formation; the remainder of the deposit consists of rhyolitic pyro-
clastic debris from the Cuscatlán Formation (Weber et al., 1978).
The andesite breccia contains clasts ranging in size from coarse
sand to boulders as large as 3 m in diameter.
Flow in the El Desagüe River consists mainly of outfl ow
from Lake Ilopango. Despite the large size of Lake Ilopango,
the amount of water that drains down the El Desagüe River is
modest. Streamfl ow records from 1961 to 1973 indicate that the
average annual discharge in the El Desagüe River ranges from
0.01 to 0.53 m3/s, and typical maximum fl ow during the rainy
season is ~2–5 m3/s (División de Meteorológico y Hidrológico,
2001). The maximum peak discharge reported during the period
of record was 7.48 m3/s.
The lake impounded by the El Desagüe landslide is ~1.5 km
long and contains an estimated 500,000 m3 of water. The lake’s
maximum depth is estimated at 5 m on the basis of heights of
treetops exposed near the landslide dam (Fig. 19). Soon after the
earthquake, a spillway channel was excavated by hand across
the toe of the landslide to allow the lake to drain. In mid-March
2001, the fl ow in this 3 m wide, 2 m deep channel was ~0.2 m3/
s, and it was not eroding the spillway. Pebble- to cobble-sized
clasts in the landslide debris effectively armored the spillway
and prevented erosion under ambient low-fl ow conditions. Since
the earthquake, drainage across the spillway has maintained the
lake at a nearly constant volume. Higher rainy season discharges
since the earthquake have not signifi cantly modifi ed the chan-
nel armor along the spillway; the armor has minimized spillway
erosion and prevented an erosional breach of the landslide dam.
The degree of channel armoring, the shallowness of the lake,
and the typically modest discharges along the El Desagüe River
suggest that failure of this dam by overtopping is unlikely. Even
discharges as great as the maximum recorded discharge on the El
Desagüe River are unlikely to signifi cantly disrupt this channel
armor and cause a catastrophic breach of the landslide dam.
The most signifi cant event that could affect the lake
impounded behind the El Desagüe landslide would be a cata-
strophic dam-break fl ood resulting from a failure of the Jiboa
River landslide dam. Such a fl ood would fl ow up the El Desagüe
River valley owing to the T-shaped confi guration of the confl u-
ence of the El Desagüe and Jiboa Rivers (Fig. 15). That type
of fl ood could fl ow across and erode the landslide dam, which
could then release the water impounded behind the El Desagüe
dam catastrophically. However, the volume of water impounded
behind the El Desagüe dam is small compared to the volume of
water that could potentially be released by a catastrophic failure
of the Jiboa River landslide dam. Although release of the water
impounded in the El Desagüe River would add to the fl ood vol-
ume downstream, its release under the scenario described above
would likely cause little additional downstream damage.
Las Leonas Landslide
The 13 February earthquake triggered failure of the slope
above the 13 January scarp of the Las Leonas landslide, and again
the Pan-American Highway was blocked by landslide debris
(Fig. 2). Removal of debris from the 13 January landslide that had
blocked the Pan-American Highway had been completed by ~7
February, and the highway had been reopened to traffi c. After this
second major failure, the government of El Salvador rerouted the
highway around the landslide to avoid future instability problems.
San Vicente Volcano
The earthquake triggered landslides in several deeply
incised drainages on the northern fl ank of San Vicente Volcano
(Fig. 15). A large rock slide occurred in the upper part of the Que-
Figure 19. Slump/rock avalanche that dammed the El Desagüe River
near its confl uence with the Jiboa River (see Fig. 15 for location). A
shallow lake upstream from the landslide (left side of photo) is drain-
ing through a spillway that was hand-dug across the toe of the slide.
84 R.W. Jibson et al.
brada Del Muerto drainage on the volcano’s northwest fl ank. On
the northern fl ank, and lower on the volcano’s slopes, extensive
failures occurred along steep walls of three drainages that are
deeply incised into an apron of coarse, unconsolidated debris-
fl ow, pyroclastic, and alluvial deposits. These failures coalesced
into debris slides that fi lled channel bottoms with loose deposits.
Potential remobilization and transport of that loose sediment as
debris fl ows or by sediment-laden fl oods poses a signifi cant haz-
ard to communities downstream, particularly in the Quebrada El
Blanco, because the town of Tepetitán is located on the alluvial
fan of this drainage.
Quebrada Del Muerto rock slide. The Quebrada Del
Muerto rock slide has an estimated volume of 200,000 m3 and
is composed mainly of large blocks of lithifi ed andesite. The
rock slide occurred high on the northwest fl ank of San Vicente
Volcano at the head of the channel of Quebrada Del Muerto
(Fig. 15). The channel passes directly east of the village of Gua-
dalupe, and because this drainage is so close to the village, com-
munity offi cials were concerned that the rock slide could remo-
bilize as debris fl ows and endanger the village. Debris within the
rock slide consists primarily of large boulders, some as much as
several meters in diameter (Fig. 20). The large size and indurated
character of the rock fragments and the minimal drainage area
above the deposit make it unlikely that this landslide debris will
remobilize into signifi cant debris fl ows that could threaten down-
stream settlements or structures, even during heavy rainfall.
Quebrada El Blanco debris slides. Low on the northern
slope of San Vicente Volcano, numerous coalescing debris slides
were triggered by the earthquake along the 40 to 50 m deep
incised drainage of Quebrada El Blanco (Fig. 15). Unconsoli-
dated pyroclastic, epiclastic, and debris-fl ow deposits in the steep
valley walls failed during the earthquake and produced numerous
debris slides that were typically 1–3 m thick; we estimate their
cumulative volume to be ~250,000 m3 (Fig. 21).
Debris from these slides clogged the channel bottom pre-
dominantly with silty sand. This sediment could remobilize as, or
be entrained by, debris fl ows during rainy seasons. Large debris
fl ows triggered in Quebrada El Blanco could extend down to the
broad alluvial fan formed near the base of the channel on which
the small town of Tepetitán is located. About 1 km upslope from
Tepetitán, discharges in Quebrada El Blanco are confi ned to a
channel that is incised several meters into older alluvial deposits.
However, closer to the town the drainage emerges onto a broad
open slope that is cut by shallow channels less than 2 m deep.
This slope is the depositional surface of previous debris fl ows,
and the active channels of the fan have routed fl ows near and
toward Tepetitán in the recent past. The most recent debris fl ow
that reached the town occurred in 1934, caused several fatalities,
and buried parts of the old village under ~60 cm of sediment
(Major et al., this volume, Chapter 7). After the 1934 event most
of the town was relocated to a higher, safer site, but over the years
the original site has been resettled.
The shallow channels and fl at topography near Tepetitán
make it diffi cult to determine the exact paths future debris fl ows
Figure 20. View toward source of the Quebrada del Muerto rock slide
on the northwest fl ank of San Vicente Volcano (see Fig. 15 for loca-
tion). Largest boulders in foreground are about 2 m across.
Figure 21. View to the northwest down the upper reaches of Quebrada El
Blanco channel showing coalescing debris slides from the valley walls
(see Fig. 15 for location). The valley bottom is clogged with landslide
debris that could mobilize into debris fl ows during heavy rainfall.
Landslides triggered by the 2001 earthquakes in El Salvador 85
will take. Shallow channels near Tepetitán could get clogged
with debris and divert fl ows westward, away from the town, or
northeastward toward the town. The channels could also remain
open and guide the fl ows toward the town. Debris-fl ow simula-
tion models (Major et al., this volume, Chapter 7) suggest that
debris fl ows having volumes greater than 100,000 m3 could
potentially reach the town.
San Agustín Delta
Liquefaction and consequent lateral spreading along the
northeastern shore of Lake Ilopango during the 13 February
earthquake damaged or destroyed many homes in the village of
San Agustín (Fig. 15). The village is situated on an active alluvial
fan of Quebrada El Chaguite, which drains ~10 km2 of deeply
dissected terrain between San Agustín and the northern edge of
the Ilopango caldera rim. Lateral spreading on this fan locally
formed fi ssures and small scarps in parts of the village and in
other areas adjacent to the lakeshore (Fig. 22). The water level in
Lake Ilopango is a major control on the local depth of the water
table; therefore, sites close to the shoreline were more prone to
liquefaction and lateral spreading than sites more distant from the
lake or at higher altitudes.
DISCUSSION
The occurrence of two major earthquakes only one month
apart had devastating consequences for El Salvador. Both earth-
quakes caused signifi cant damage and fatalities, but each had
different source characteristics that resulted in different damage
patterns. These earthquakes typify two fairly common hazard
scenarios that are anticipated in seismically active regions: a
large-magnitude, distant earthquake and a moderate-magnitude,
proximal earthquake. Comparison of landslide patterns from
these two events helps us understand what to expect in such
seismic scenarios.
The 13 January earthquake was a large-magnitude (M-7.7),
moderately deep (60 km) and distant (40 km offshore) event. The
resulting landslide pattern was generally what might be expected
from such an event: landslides were scattered over a very broad
region in the southern half of El Salvador and Guatemala. Land-
slide concentrations appeared to relate more to localized areas
of high susceptibility (steep, weak slopes) than to localization
of strong shaking. One possible exception to this was the Las
Colinas landslide, which occurred in an area that appears to have
experienced an abnormally high intensity of ground shaking
compared to surrounding areas.
The 13 February earthquake was a moderate-magnitude
(M-6.6), shallow (15 km) event located near the center of the
country. Landslides triggered by this earthquake were more
densely concentrated in a much smaller area near the epicenter.
This concentration appeared to relate closely to a limited area of
most intense shaking.
Figure 23 shows the approximate areas affected by landslid-
ing in the two earthquakes on a plot of landslide areas versus
magnitudes from 40 worldwide earthquakes (Keefer, 2002).
Areas from both earthquakes plot above the mean regression
line but are in the main cluster of data. Thus, areas affected by
landslides in these earthquakes are above the worldwide average
but are well within the upper bound of worldwide data. Using the
same data set, Keefer (1984) also plotted maximum epicentral
distance to landslides versus magnitude; his data indicate that the
13 January earthquake (M-7.7) could have triggered landslides
as far as ~330 km from the epicenter. Although their precise loca-
tions are unknown, landslides from the January earthquake were
reported in west-central Guatemala (Bommer et al., 2002) at or
slightly beyond that distance. These observations are consistent
Figure 23. Area affected by landslides from worldwide earthquakes
(modifi ed from Keefer, 2002). Dashed line is upper bound; solid line
is best-fi t regression line. Stars are the 2001 El Salvador earthquakes,
which plot above the regression line but below the upper bound.
Figure 22. Lateral-spread landslide on the northeastern shore of Lake
Ilopango (see Fig. 15 for location).
86 R.W. Jibson et al.
with conclusions of other studies that suggest that earthquakes
in El Salvador tend to trigger landslides over larger than aver-
age areas (Bommer and Rodríguez, 2002; Bommer et al., 2002;
Rodríguez et al., 1999).
Types of landslides triggered by the two earthquakes were
broadly similar and were consistent with observations from
earthquakes worldwide: The large majority of triggered land-
slides were shallow (<5 m), disrupted falls and slides in rock and
debris (Keefer, 1984, 1999; Rodríguez et al., 1999). Each earth-
quake also triggered localized liquefaction and lateral spreading
as well as a few large, deep landslides that caused great damage
and, in the case of Las Colinas, great loss of life.
A major factor in the distribution of landslides associ-
ated with both earthquakes was the presence of thick deposits
of poorly consolidated Holocene and late Pleistocene tephra.
These materials are particularly susceptible to failure during
seismic shaking owing to their low tensile strength and their
tendency to erode readily and form deeply incised, steep-walled
gullies (Bommer and Rodríguez, 2002; Rolo et al., this volume,
Chapter 5). Such materials have been observed in other geologi-
cally similar areas to be particularly susceptible to failure during
earthquakes (Bommer and Rodríguez, 2002; Harp et al., 1981).
Seismic shaking can cause these types of materials to collapse,
disaggregate, and fi ll gully bottoms with large masses of loose
debris that are then remobilized during wet periods and trans-
ported through the alluvial system (Keefer, 1999). This can lead
to dramatically increased sedimentation in subsequent rainy sea-
sons and possibly to the generation of debris fl ows during intense
or prolonged rainfall.
The January earthquake also triggered damaging rock
falls from hard, fractured lava fl ows. When such rock falls are
triggered on steep slopes, boulders can bounce and roll at high
speeds for signifi cant distances and do great damage to anything
they hit. Potential sources of rock falls are generally easy to iden-
tify because the hard lava fl ows tend to crop out prominently.
The earthquakes occurred in the dry season when soil mois-
ture and groundwater levels were at annually low levels. Had
these earthquakes occurred during the wet season, the number,
size, and areal extent of landslides likely would have been far
greater. For example, the 1906 San Francisco, California, earth-
quake (M = 8.2) occurred near the end of the rainy season in an
above-average rainfall year; thousands of landslides were trig-
gered over a very broad region including the Santa Cruz Moun-
tains, which extend along the fault-rupture zone (Lawson, 1908).
By contrast, the 1989 Loma Prieta, California, earthquake (M =
7.1), centered in the Santa Cruz Mountains, occurred at the end
of the dry season after ~5 yr of sustained drought. This earth-
quake produced far fewer landslides, many of which moved a
very limited distance or did not develop completely (Keefer and
Manson, 1998). Schuster et al. (1998) and Keefer and Manson
(1998) concluded that the signifi cant difference in landslide trig-
gering from these two earthquakes is attributable not just to the
magnitude difference but to soil-moisture conditions that relate
to seasonal rainfall.
SUMMARY AND CONCLUSIONS
In early 2001, El Salvador experienced two major earth-
quakes in a period of just one month. The 13 January M-7.7
earthquake occurred at a depth of 60 km and was located ~40
offshore. The 13 February M-6.6 earthquake occurred at a depth
of 15 km and was located near the center of the country. Both
earthquakes triggered thousands of landslides that caused most of
the damage and fatalities. In both earthquakes, the large majority
of landslides were shallow slides and falls in rock and debris, but
both earthquakes also triggered a few large, deep landslides that
were very damaging.
Landslides from the 13 January earthquake were broadly
scattered over the southern half of El Salvador and Guatemala
extending across an area of perhaps 25,000 km2, somewhat above
average for earthquakes in this magnitude range. Areas of great-
est landslide concentration included the Cordillera Bálsamo, an
upland area of deeply dissected terrain composed of young, weak
volcaniclastic deposits; areas around Lake Ilopango; and gullies
on the fl anks of volcanoes in the Usulután region. The most
devastating landslide triggered by the earthquake was the Las
Colinas landslide from the north slope of Bálsamo Ridge. This
130,000 m3 slide detached from a steep ridge above Santa Tecla
and fl owed 735 m into the densely populated Las Colinas neigh-
borhood, killing ~585 people. Another large slide (~750,000 m3)
blocked the Pan-American Highway near San Vicente.
The 13 February 2001 earthquake in central El Salvador
triggered thousands of landslides in the area between Lake
Ilopango and the city of San Vicente. In much of this area,
strong ground shaking caused extensive failures on steep walls
of deeply incised valleys. Two large landslides also were trig-
gered that blocked two major rivers in the area: the El Desagüe
and Jiboa Rivers. The El Desagüe River landslide impounded a
small, shallow lake of ~500,000 m3; lake level quickly stabilized
because the low-volume outfl ow from the lake was channeled
through a small, hand-dug spillway. The Jiboa River landslide
impounded a lake that could have reached a maximum volume
of ~7 million m3, which poses a more signifi cant hazard. A 20 m
deep spillway was excavated through the landslide dam to pre-
vent the lake from reaching its projected capacity; although some
erosion of the spillway has occurred, its channel appears to have
stabilized, at least temporarily.
The distributions of triggered landslides from these earth-
quakes were consistent with what has been observed in other
earthquakes in the region (Bommer et al., 2002; Rymer and
White, 1989) and is intuitively expected. The large, moderately
deep and distant earthquake of 13 January triggered broadly scat-
tered landslides across a large area encompassing the southern
parts of El Salvador and Guatemala. Because moderately high
levels of ground shaking occurred throughout much of this area,
landslide concentrations were related primarily to areas of very
high susceptibility owing to steep slopes in young, weak volcanic
material. The shallow, moderate-magnitude, onshore earthquake
of 13 February triggered landslides over a much smaller area,
Landslides triggered by the 2001 earthquakes in El Salvador 87
about one-tenth the size of that affected in the January earth-
quake. Landslides were densely concentrated in the epicentral
area of highest-intensity ground shaking. In both earthquakes,
the areas most heavily affected by landslides were underlain by
thick deposits of unconsolidated late Pleistocene and Holocene
volcanic tephra.
Because El Salvador is located along a major plate bound-
ary, it will continue to experience earthquakes in the future. The
earthquakes of January–February 2001 demonstrated that El
Salvador must be prepared for a broad spectrum of seismic-haz-
ard scenarios, ranging from large, distant earthquakes to small
and moderate, proximal earthquakes. While these two types of
earthquakes produce different areal distributions of hazards, they
are both capable of causing enormous damage and need to be
factored into seismic-hazard planning throughout El Salvador.
ACKNOWLEDGMENTS
Field investigations in El Salvador were facilitated through
the assistance and support of the U.S. Agency for International
Development (USAID) and the El Salvadoran Ministerio de
Medio Ambiente y Recursos Naturales. Assistance in conducting
fi eld investigations during the January earthquake was provided
by Luz Antonina Barrios and Arturo Quezada of Geotérmica
Salvadoreña (GESAL). Willie Rodriguez (U.S. Geological Sur-
vey) assisted with planning and logistics for our fi eld operations.
Margo Johnson composed the maps in Figures 1 and 2. Gerald
Wieczorek and Robert Jarrett reviewed the manuscript.
REFERENCES CITED
Asanza, M., Plaza-Nieto, G., Yepes, H., Schuster, R.L., and Ribadeneira, S.,
1992, Landslide blockage of the Pisque River, northern Ecuador, in
Bell, D.H., ed., Landslides, Glissements de Terrain, Proceedings of the
Sixth International Symposium, Christchurch, New Zealand, 10–14 Feb-
ruary, 1992: Rotterdam, A.A. Balkema, p. 1229–1234.
Ashford, S.A., Sitar, N., Lysmer, J., and Deng, N., 1997, Topographic effects
on the seismic response of steep slopes: Bulletin of the Seismological
Society of America, v. 87, p. 701–709.
Baum, R.L., Crone, A.J., Escobar, D., Harp, E.L., Major, J.J., Martinez, M.,
Pullinger, C., and Smith, M.E., 2001, Assessment of landslide hazards
resulting from the February 13, 2001, El Salvador earthquake: U.S. Geo-
logical Survey Open-File Report 01-119, 20 p.
Bommer, J.J., and Rodríguez, C.E., 2002, Earthquake-induced landslides in
Central America: Engineering Geology, v. 63, p. 189–220.
Bommer, J.J., Hernandez, D.A., Navarrete, J.A., and Salazar, W.M., 1996,
Seismic hazard assessments for El Salvador: Geofísica Internacional,
v. 35, p. 227–244.
Bommer, J.J., Benito, M.B., Ciudad-Real, M., Lemoine, A., López-Menjívar,
M.A., Madariaga, R., Mankelow, J., Méndez de Hasbun, P., Murphy, W.,
Nieto-Lovo, M., Rodríguez-Pineda, C.E., and Rosa, H., 2002, The El
Salvador earthquakes of January and February 2001: Context, character-
istics, and implications for seismic risk: Soil Dynamics and Earthquake
Engineering, v. 22, p. 389–418.
División de Meteorológico y Hidrológico, 2001, Dirección general de recursos
naturales renovables—Anuario Hidrológico 1961–1996: Ministerio de
Agricultura y Ganadería, El Salvador.
Dull, R.A., Southton, J.R., and Sheets, P., 2001, Volcanism, ecology, and cul-
ture: A reassessment of the Volcán Ilopango TBJ eruption in the southern
Mayan realm: Latin American Antiquity, v. 12, p. 25–44.
Evans, S.G., and Bent, A.L., 2004, The Las Colinas landslide, Santa Tecla, El
Salvador: A highly destructive fl owslide triggered by the 13 January 2001
earthquake, in Rose, W.I., et al., eds., Natural hazards in El Salvador:
Boulder, Colorado, Geological Society of America Special Paper 375,
p. 25–37 (this volume).
Harp, E.L., and Vallance, J., 2001, Landslide hazards in El Salvador triggered
by the 13 January 2001 earthquake: U.S. Geological Survey report to U.S.
Agency for International Development, 6 p.
Harp, E.L., Wilson, R.C., and Wieczorek, G.F., 1981, Landslides from the
February 4, 1976, Guatemala earthquake: U.S. Geological Survey Profes-
sional Paper 1204-A, 35 p.
Harp, E.L., Sarmiento, J., and Cranswick, E., 1984, Seismic-induced pore-
water pressure records from the Mammoth Lakes, California, earthquake
sequence of 25 to 27 May 1980: Bulletin of the Seismological Society of
America, v. 74, p. 1381–1393.
Jibson, R.W., and Crone, A.J., 2001, Observations and recommendations regard-
ing landslide hazards related to the January 13, 2001, M-7.6 El Salvador
earthquake: U.S. Geological Survey Open-File Report 01-141, 19 p.
Keefer, D.K., 1984, Landslides caused by earthquakes: Geological Society of
America Bulletin, v. 95, p. 406–421.
Keefer, D.K., 1999, Earthquake-induced landslides and their effects on alluvial
fans: Journal of Sedimentary Research, v. 69, p. 84–104.
Keefer, D.K., 2002, Investigating landslide caused by earthquakes—A histori-
cal review: Surveys in Geophysics, v. 23, p. 473–510.
Keefer, D.K., and Manson, M.W., 1998, Regional distribution and characteris-
tics of landslides generated by the earthquake, in Keefer, D.K., ed., The
Loma Prieta, California, Earthquake of October 17, 1989—Landslides:
U.S. Geological Survey Professional Paper 1551-C, p. C7–C32.
Konagai, K., Johansson, J., Mayorca, P., Uzuoka, R., Yamamoto, T.,
Miyajima, M., Pulido, N., Sassa, K., Fukuoka, H., and Duran, F., 2004,
Las Colinas landslide: Rapid and long-traveling soil fl ow caused by the
January 13, 2001, El Salvador earthquake, in Rose, W.I., et al., eds.,
Natural hazards in El Salvador: Boulder, Colorado, Geological Society of
America Special Paper 375, p. 39–53 (this volume).
Lawson, A.C., 1908, The California Earthquake of April 18, 1906—Report of
the State Earthquake Investigation Committee, Volume 1: Washington,
D.C., Carnegie Institution, 451 p.
Major, J.J., Schilling, S.P., Pullinger, C.R., and Escobar, C.D., 2004, Debris-
fl ow hazards at San Salvador, San Vicente, and San Miguel volcanoes,
El Salvador, in Rose, W.I., et al., eds., Natural hazards in El Salvador:
Boulder, Colorado, Geological Society of America Special Paper 375,
p. 89–108 (this volume).
Rodríguez, C.E., Bommer, J.J., and Chandler, R.J., 1999, Earthquake-induced
landslides: 1980–1997: Soil Dynamics and Earthquake Engineering,
v. 18, p. 325–346.
Rolo, R., Bommer, J.J., Houghton, B.F., Vallance, J.W., Berdousis, P., Mavrom-
mati, C., and Murphy, W., 2004, Geologic and engineering characteriza-
tion of Tierra Blanca pyroclastic ash deposits, in Rose, W.I., et al., eds.,
Natural hazards in El Salvador: Boulder, Colorado, Geological Society of
America Special Paper 375, p. 55–67 (this volume).
Rose, W.I., Conway, F.M., Pullinger, C.R., Deino, A., and McIntosh, W.C.,
1999, An improved age framework for late Quaternary silicic eruptions in
northern Central America: Bulletin of Volcanology, v. 61, p. 106–120.
Rymer, M.J., and White, R.A., 1989, Hazards in El Salvador from earthquake-
induced landslides, in Brabb, E.E., and Harrod, B.L., eds., Landslides:
Extent and economic signifi cance: Rotterdam, A.A. Balkema, p. 105–109.
Schuster, R.L., Wieczorek, G.F., and Hope, D.G., 1998, Landslide dams
in Santa Cruz County, California, resulting from the earthquake, in
Keefer, D.K., ed., The Loma Prieta, California, earthquake of October 17,
1989—Landslides: U.S. Geological Survey Professional Paper 1551-C,
p. C51–C70.
Stewart, J.P., Seed, R.B., and Bray, J.D., 1996, Incidents of ground failure
from the Northridge earthquake: Bulletin of the Seismological Society of
America, v. 86, no. 1B, p. S300–S318.
U.S. Agency for International Development, 2001, El Salvador—Earthquake II
Fact Sheet #8 (FY 2001) 15 March 2001: www.usaid.gov/hum_response/
ofda/elsalveq2_fs8_fy01.html (last visited 15 Jan. 2003).
U.S. Geological Survey, NEIC (National Earthquake Information Center),
2001a, Earthquake Bulletin: http://neic.usgs.gov/neis/bulletin/01_
EVENTS/010113173329/ (last visited 23 Sept. 2003).
U.S. Geological Survey, NEIC (National Earthquake Information Center),
2001b, Earthquake Bulletin: http://neic.usgs.gov/neis/bulletin/01_
EVENTS/010213142205/ (last visited 23 Sept. 2003).
Varnes, D.J., 1978, Slope movement types and processes, in Schuster, R.L.,
and Krizek, R.J., eds., Landslides—Analysis and control: Transportation
88 R.W. Jibson et al.
Research Board, National Academy of Science, Special Report 176,
p. 11–33.
Walder, J.S., and O’Connor, J.E., 1997, Methods for predicting peak discharge
of fl oods caused by failure of natural and constructed earthen dams: Water
Resources Research, v. 33, p. 2337–2348.
Weber, H.S., Wiesemann, G., Lorenz, W., and Schmidt-Thome, M., 1978,
Mapa geólogico de la República de El Salvador/América Central: Bunde-
sanstalt fuer Geowissenschaften und Rohstoffe, Hannover, Germany,
scale 1:100,000, 6 sheets.
MANUSCRIPT ACCEPTED BY THE SOCIETY JUNE 16, 2003
Printed in the USA
