The Gibraltar Arc seismogenic zone (part 2): Constraints on a
shallow east dipping fault plane source for the 1755 Lisbon
earthquake provided by tsunami modeling and seismic intensity
M.-A. Gutschera,⁎, M.A. Baptistab, J.M. Mirandab
aIUEM, Univ. Brest, UMR 6538, Plouzané, France
bInst. of Geophysics, Univ. Lisbon, Portugal
Accepted 7 February 2006
Available online 7 July 2006
The Great Lisbon earthquake has the largest documented felt area of any shallow earthquake and an estimated magnitude of
8.5–9.0. The associated tsunami ravaged the coast of SW Portugal and the Gulf of Cadiz, with run-up heights reported to have
reached 5–15 m. While several source regions offshore SW Portugal have been proposed (e.g.— Gorringe Bank, Marquis de
Pombal fault), no single source appears to be able to account for the great seismic moment as well as all the historical tsunami
amplitude and travel time observations. A shallow east dipping fault plane beneath the Gulf of Cadiz associated with active
subduction beneath Gibraltar, represents a candidate source for the Lisbon earthquake of 1755.
Here we consider the fault parameters implied by this hypothesis, with respect to total slip, seismic moment, and recurrence
interval to test the viability of this source. The geometry of the seismogenic zone is obtained from deep crustal studies and can be
represented by an east dipping fault plane with mean dimensions of 180 km (N–S)×210 km (E–W). For 10 m of co-seismic slip an
Mw 8.64 event results and for 20 m of slip an Mw 8.8 earthquake is generated. Thus, for convergence rates of about 1 cm/yr, an
event of this magnitude could occur every 1000–2000 years. Available kinematic and sedimentological data are in general
agreement with such a recurrence interval. Tsunami wave form modeling indicates a subduction source in the Gulf of Cadiz can
partly satisfy the historical observations with respect to wave amplitudes and arrival times, though discrepancies remain for some
stations. A macroseismic analysis is performed using site effect functions calculated from isoseismals observed during instrumentally
recorded strong earthquakes in the region (M7.9 1969 and M6.8 1964). The resulting synthetic isoseismals for the 1755 event suggest
a subduction source, possibly in combination with an additional source at the NW corner of the Gulf of Cadiz can satisfactorily
explain the historically observed seismic intensities. Further studies are needed to sample the turbidites in the adjacent abyssal plains
to better document the source region and more precisely calibrate the chronology of great earthquakes in this region.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Great Lisbon earthquake; Iberia; Morocco; Subduction; Tsunami
The Great Lisbon earthquake of 1755 with an
estimated magnitude of 8.5–9.0 was felt as far away
as Hamburg, the Azores and Cape Verde Islands and has
Tectonophysics 426 (2006) 153–166
⁎Corresponding author. Tel.: +33 2 98 49 87 27; fax: +33 2 98 49 87
E-mail address: firstname.lastname@example.org (M.-A. Gutscher).
0040-1951/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
the largest documented felt area of any shallow
earthquake (Martinez-Solares et al., 1979; Johnston,
1996) (Fig. 1). The earthquake which struck on 1
November 1755 (All Saints Day) caused up to 100,000
deaths (Chester, 2001) through destruction by ground
shaking, from the ensuing fires and by tsunami waves of
5–15 m which devastated the coasts of Southwest Iberia
and Northwest Morocco (Baptista et al., 1998a). How-
ever, to this day the source of this great earthquake re-
mains unknown (Gutscher, 2004).
While it is difficult to accurately determine the mag-
nitude of great historical earthquakes like the 1755 event,
comparison to recent strong earthquakes can be useful.
The 1969 Cape St. Vincent earthquake with an instru-
mentally determined magnitude of Mw=7.9 provided
such an opportunity (Fukao, 1973; Buforn et al., 1988).
depth) and generated a modest tsunami. The observed
tsunami amplitudes at Cascais (near Lisbon), Lagos (near
Cape St. Vincent) and Cadiz in 1969 were 50 cm, 50 cm
and 10 cm respectively (Gjevik et al., 1997). For com-
these same locations in 1755 were 6 m, >10 m and 15 m,
respectively. This demonstrates the tremendous energy
liberated by the 1755 event and implies that it was
significantly stronger than the 1969 earthquake (Baptista
et al., 1998a).
Previous studies of the 1755 earthquake have
proposed prominent basement highs off the SW Iberia
margin, such as the Gorringe Bank (Johnston, 1996) or
the Marquis de Pombal (Zitellini et al., 2001; Gracia et
al., 2003) to be the likely source (Fig. 1). Yet the
relatively modest surface areas of these source regions
Fig. 1. Isoseismal map of the 1755 earthquake (after Martinez-Solares et al., 1979; Levret, 1991). Inset shows entire felt zone in NWAfrica, western
Europe and the NE Atlantic (after Johnston, 1996). Bathymetry shown as 1000 m contours (Sandwell and Smith, 1997). The 1969 Cape St. Vincent
earthquake occurredon theN60Etrending Horseshoe Fault.Historicallyreportedtsunamiarrivaltimes andamplitudes aregivenfor cities inthe study
area (Baptista et al., 1998a).
154 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
make it difficult to explain the seismic moment, for a
reasonable set of fault parameters (e.g. — co-seismic
displacement, rigidity, recurrence), consistent with the
slow, NW–SE relative convergence between Africa and
Iberia at about 4 mm/a (Argus et al., 1989; Fernandes et
al., 2003). Seismic images of the crustal structure in the
Fig. 2. Cross-section through the subduction fault plane (A–A′ in Fig. 1), showing the geometry of the sources tested (with and without fault splay)
for different amounts of fault slip (10 and 20 m) and the initial seafloor displacement associated with each.
155 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
accretionary wedge, overlying an eastward dipping
basement and connected to a steep, east dipping slab of
cold, oceanic lithosphere beneath Gibraltar (Gutscher et
In this paper, we use the geometry of the shallow east
dipping fault plane of the Gibraltar subduction as deter-
mined in a parallel study (Thiebot and Gutscher, 2006-
this volume) (Fig. 1) and calculate the ensuing seismic
moment, as well as model synthetic tsunami waves
generated by rupture along such a fault plane. Further-
more, we perform a macroseismic analysis, using a
regional site effect function (derived here) and calculate
synthetic isoseismals for a subduction source in addition
maps based on historical observations. Finally, we
consider the recurrence time for great earthquakes
along this subduction plane in terms of sedimentological
observations and regional kinematic data.
2. Subduction plane fault parameters
Deep seismic reflection and refraction profiles image
an east dipping decollement and east dipping basement
linked to the subduction zone beneath Gibraltar.
Numerical modeling of the temperature distribution
along the subduction interface suggests a seismogenic
zone witha downdipwidthofabout200km(Thiebotand
represent the principal fault plane using a series of rect-
angular sub-planes extending from 6.5 km depth east-
dip increases progressively from 2.5°, to 5° to 7.5°. We
the seafloor to 6.5 km depth, where it joins the principal
fault plane (Fig. 2). Fault splays within an accretionary
wedge are considered to be potentially active during co-
seismic slip and may contribute to the generation of
tsunamis (Park et al., 2002; Satake et al., 2003). Moment
magnitudes are calculated here for two different uniform
co-seismic slips of 20 and 10 m. The mean dimensions of
the fault plane are:
N ? S length ¼ 180 km;
totalE ? W width ¼ 210 km;
Seismic moment is calculated according to the equation:
Mo ¼ ASD; A ¼ 3 ? 1010Pa
rupture area S ¼ 37800 km2;
for slip of D ¼ 20 m
Mo ¼ 2:27 ? 1022Nm
Mw ¼ 2=3log Mo−6:03; yields Mw ¼ 8:80
for a slip D ¼ 10 m
Mo ¼ 1:13 ? 1022Nm
yields Mw ¼ 8:64
For comparison, the Great Alaska earthquake of 1964
had a rupture area of
750 km ? 200 km and a mean slip D ¼ 10 m:
Thus; Mo¼4:5?1022Nm and Mw¼9:1:
For the 1964 Alaska earthquake, however, two major
asperities ruptured and were calculated (on the basis of
geodetic and tsunami inversions) to have slips reaching
10–15 m and 20–25 m (Holdahl and Sauber, 1994;
Johnson et al., 1996). The rupture of 1755, may likewise
have had a mean slip of 10 m, but with local peak
displacements up to 20 m.
3. Tsunami modeling
The above simplified, fault plane was used to perform
tsunami wave form modeling of a shallow east dipping
subduction source. The initial displacement of the sea-
floor, that we consider to be similar to the initial dis-
placement of the water surface, is calculated using the
elastic half-space approach (Okada, 1985). The vertical
displacement (for a pure thrust mechanism) is shown
Source parameters of rectangular fault plane segments (given in UTM29 coordinates in m)
SegmentDipSouth Nord Degreesm
X segY segX segY seg StrikeRake SlipWidthLength
Note the dip of the fault plane increases towards the east, with increasing depth (see Fig. 2).
156M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
along a profile perpendicular to the fault plane for the 4
cases tested,10m and 20m slip,without andwitha splay
10 m slip and no splay fault, with a maximum of + 3.5 m
and a minimum of −2.4 m. The greatest seafloor dis-
placement is obtained for the case with 20 m slip, in-
cluding the splay fault, and ranges from +10 m at the
western end of the Gulf of Cadiz to −5 m at the eastern
end.This initial sea-surface displacementthenpropagates
throughoutthe Atlantic inthe Iberia— NWAfricaregion
(Fig. 3). For a tsunami wave in deep water, the pro-
pagation velocity is approximately proportional to the
square root of the water depth v=(gh)1/2. Across regions
at 4000 m water depth for instance, the tsunami wave
drops to 360 km/h and so on. The velocity of wave
shelf regions are encountered. Thus, a good regional
bathymetric model is required for accurately modeling
tsunami arrival times. Here we use the global 2 min
combining satellite altimetry with available depth sound-
ings along ship tracks. Finite difference software SWAN
Code (Mader, 1988) was applied to calculate wave pro-
pagation, using a shallow water non-linear wave model
and a cell size of 0.025°. The regional propagation of the
stations in the SW Iberia, NW African region (Fig. 4).
These stations were selected for the most part on the
non-linear propagation effects in shallow water estuaries,
which may induce significant modeling errors. For in-
stance, we did not model the travel time and wave heights
for Lisbon, where propagation is deeply affected by the
shallow Tagus Estuary, which cannot be adequately re-
presented by our bathymetric model. Instead we consid-
ered Oeiras, a small town just west of Lisbon harbor,
when modeling travel times and wave heights in Huelva
(Spain), we also considered the neighboring Portuguese
city of Vila Real de San Antonio (which was severely
damaged by the 1755 earthquake and tsunami and com-
pletely rebuilt similar to downtown Lisbon, but is not
influenced by propagation in an estuary).
Synthetic mareograms provide information on wave
phase as well as the period (duration) of a tsunami wave.
There are contradictions between the different compila-
tions of historical records on the polarity of the first
wave at the SW Spanish coast. Martinez-Solares et al.
(1979) report a withdrawal, while Baptista et al. (1998a)
cite 1756 Spanish documents describing a flux from the
sea. Our synthetic mareograms for Huelva and Cadiz
show, respectively, a temporary drop in sea-level of 1.5,
and 5 m, respectively, before the positive tsunami wave
arrived (Fig. 4A). The mean period of the modeled
Fig. 3. Three-part time sequence showing propagation of tsunami
wave in map view. At time zero, the initial displacement of the
seafloor, corresponding to the initial displacement of the surface of the
sea, is shown. Bathymetry as in Fig. 1.
157M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
tsunami waves for the Cadiz station is about 20–40 min,
which is in good agreement with historical reports
(Baptista et al., 1998a).
Calculated travel times are compared to historical ob-
servations in Table 2. Modeled tsunami travel times show
and Madeira stations (5-15 min difference). The modeled
arrival time forCadiz is about 40mintooearly(36minvs.
Fig. 4. Synthetic mareograms A) Portuguesewest coast stations (south): Cape St. Vincent, Oeiras, B) Portuguesewest coast stations (north): Figueira,
Porto, C) Gulf of Cadiz stations: Huelva, Villa Real, Cadiz, D) Tangiers and Gibraltar, E) Madeira Island stations: Porto Santo, Madeira, F) Safi
158 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
basis of thermal modelling of forearc thermal structure
(Thiebot and Gutscher, 2006-this volume). A shorter
and provide a better travel time fit, while slightly reducing
the total seismic moment. Arrival times for the Portuguese
west coast and Safi, are about 25–40 min later than
historical reports. This may be due to a combination of
several possible causes: substantial delays due to poorly
constrained shallow shelf bathymetry, inaccuracy of
historical reports or multiple tsunami sources.
Calculated wave amplitudes are compared to histor-
ical observations in Table 3 for the four cases tested.
Modeled tsunami amplitudes are greatest in the Gulf of
Cadiz — Cape St. Vincent region, in agreement with
historical observations. However, most modeled ampli-
tudes are significantly less than historical reports (30–
50% of the reported amplitudes for the 20 m fault splay
source). An even greater discrepancy exists for Safi
(Morocco) where only a 1.2 m high wave is modeled but
a 6 m wave was reported. For many stations some of this
discrepancy may be due to run-up effects, which have
been estimated to locally increase tsunami amplitudes
by 30% (Baptista et al., 1998b). But for Safi another
explanation must be sought.
Given the systematically late arrivals and low am-
plitudes for the Portuguese west coast stations, it ap-
pears that the subduction source proposed here, cannot
alone reproduce the observed tsunami for the entire
region and that a contribution from an additional source
further to the NW may be necessary. However, the sub-
duction source is better able to account for the reported
tsunami arrival times for the stations Huelva, Cape St.
Vincent, Safi and Gibraltar than earlier modelsbased ona
single source on the Marquis de Pombal (Baptista et al.,
1998b) (Table 2) and amplitudes are somewhat closer to
4. Macroseismic analysis
Macroseismic analysis can be helpful in reconstruct-
ing sources for historical earthquakes where instrumen-
tal recordings are absent. The Iberia — NW African
region, has been struck by strong earthquakes in the
recent past, which can be used to establish a regional
attenuation function (Levret, 1991; Johnston, 1996), as
well as to determine local site effects. The two strongest
recent events were the M7.9 St. Vincent earthquake of
1969 (Fukao, 1973) and an M6.5 event in the northern
Gulf of Cadiz in 1964 (Udias and Arroyo, 1970) (Fig. 5).
Once the site effects have been determined, synthetic
as demonstrated in similar studies (Mendes Victor et al.,
1999; Baptista et al., 2003). Attenuation functions (of
epicentral distance vs. intensity) have been proposed for
the 1755 and 1969 earthquakes (Levret, 1991). Here, we
subtract the radial attenuation functions (Fig. 6A,B) from
published maps of seismic intensity and determine the
local site effect for both the 1969 and the 1964 events
influence the felt intensity, certain general patterns are
reflected in both site effect maps. Next, we calculate a
mean site effect map taking an average of the 1969 and
1964 site effect maps (Fig. 7). (N.B. — This averaging
Comparison of historically observed tsunami arrival times with
calculated arrival times from tsunami modeling of a shallow east
dipping source and compared to calculated arrival times from a source
at the Marquis de Pombal “source B“ (N160) (Baptista et al., 1998b)
StationHistorical TT Modelled TT Marq. Pomb. TT
Cape St. Vincent
Safi (NW Morocco) 30 min
(not modelled, shallow water)
28 min25 min
Comparison of historically observed tsunami wave amplitudes with
calculated amplitudes from tsunami modeling of a shallow east
dipping source and compared to calculated amplitudesfrom a source at
the Marquis de Pombal (N160 source) (Baptista et al., 1998b)
10 m 20 mModelled
Porto Santo 3 m
>10 m 2.2 m 4.4 m 2.5 m5.0 m 6.0 m
(17 m ??)
1.1 m 2.3 m 1.3 m
2.4 m 4.8 m 2.7 m
0.7 m 1.3 m 1.4 m
1.4 m 2.8 m 1.4 m
1.1 m 1.7 m 2.2 m
1.4 m 1.5 m 2.7 m
0.5 m 0.6 m 1.1 m
1.1 m 1.1 m 2.2 m
(not modelled, shallow water)
0.2 m 0.4 m 0.2 m
0.3 m 0.6 m 0.3 m
2.2 m 10.0 m
159M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
process results in 0.5 MSK intensity contours. While we
recognize that such precision is not possible with respect
to felt seismic intensity, one can consider the probability
of the intensity at this site being the next highest full I
value, or the next lowest full I value, to be 50%).
The averaged site effect map shows strong attenua-
tion from the southern Portuguese coast (Algarve) tow-
ards the NE to the Variscan Massifs of Central Iberia.
Strong attenuation is observed as well in the internal
zones of the Betic–Rif orogen and the Western Alboran
Sea. Conversely, there are moderate amplification ef-
fects in the Guadalquivir Basin (Betic foreland basin),
and the Rharb Basin (Rif foreland basin). Finally, strong
amplification is found in the Agadir Basin (SW
Morocco) and in the Porto region of NW Portugal. In
almost all cases the amplification is related to Cenozoic
sedimentary basins, a correlation pointed out in a pre-
vious study (Levret, 1991). The regions of greatest
attenuation commonly correlate with crystalline massifs,
except for the thickly sedimented and thinned crustal
domain of the West Alboran Sea. Here attenuation may
be due to the low p-wave velocity and low Q anomaly in
the upper mantle of the West Alboran domain (Calvert et
al., 2000), caused by vigorous convection in the back-
arc. Next, by taking a published regional attenuation
function for the 1755 event (Levret, 1991) and adding
our mean site function, we calculated synthetic iso-
seismals for the 1755 earthquake.
Three potential source regions were considered. In
each case the source region (with constant intensity) has
a radius of 100 km and beyond this distance intensity
Levret (1991). The first source centered at 8° W, 35.5° N
corresponds to the subduction fault plane (Fig 8A). The
resulting synthetic isoseismal map (Fig. 8B) predicts
very high intensities in SW Spain and NW Morocco
(I=8–9). While such intensities were reported for Cadiz
and Sevilla, these values are 1–2 intensities higher than
historical reports from Morocco. And the predicted
intensity for Lisbon at 7.5 is about 1.5 too low. The
second source centered at 8.5° W, 36° N (Fig. 8C) is at
the NW edge of the subduction fault plane. This could
represent either a large slip asperity along the fault plane
in this vicinity, or alternatively, the combined moment
release from the subduction fault plane and from a N75E
oriented fault, representing the Africa–Eurasia plate
boundary in the northern Gulf of Cadiz (Jimenez-Munt
et al., 2001; Negredo et al., 2002) and situated at the
northern edge of the accretionary wedge. Slip along this
fault has been interpreted to be the source of the 1964
M6.5 earthquake in the Northern Gulf of Cadiz (Udias
and Arroyo, 1970). The resulting synthetic isoseismal
map (Fig. 8D), predicts I=10 along the Algarve coast,
I=8–9 in SW Spain and I=8–9 in NW Morocco.
Lisbon with a predicted intensity of 8is also closer to the
historically reported I=9. Thus, the overall fit is
satisfactory. Finally, the third source is centered at 9.5°
W, 36.5° N (Fig. 8E) about 100 km SW of Cape St.
Vincent (SW tip of Portugal). This could represent the
Marquis de Pombal together with an east trending
segment of the Africa–Iberia plate boundary. The
resulting synthetic isoseismal map (Fig. 8F) predicts
I=10 at the SW corner of Portugal,I=8–9in Lisbon and
I=6–7 throughout NW Morocco. The overall fit to the
historically reported isoseismals is very good in Iberia
and satisfactory in Morocco. However, some uncer-
Fig. 5. Isoseismal maps a) M7.9 Cape St. Vincent earthquake of 1969, b) M6.8 Gulf of Cadiz earthquake of 1964. Bathymetry as in Fig. 1.
160 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
Fig. 6. Radial attenuation function for; A) M7.9 Cape St. Vincent earthquake of 1969, B) M6.8 Gulf of Cadiz earthquake of 1964. Calculated site
effect maps for C) M7.9 Cape St. Vincent earthquake of 1969, D) M6.8 Gulf of Cadiz earthquake of 1964.
161 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
tainties remain regarding reports of I=8 in Meknes, Fez
and Marrakech and I=7 in Agadir. These historically
than predicted by such a source.
Three other candidate sources have been proposed for
the 1755 earthquake; Gorringe Bank, Marquis de Pombal
and the Horseshoe fault (Fig. 1). Following the M7.9 St.
Vincent earthquake in 1969, which occurred on the
Horseshoe Fault in the Horseshoe abyssal plain, some
authors proposed this to be the source of the 1755 earth-
solution indicates conjugate N60° striking, 45° dipping
fault planes with a reverse sense of motion. Other authors
proposed the nearby Gorringe Bank, an uplifted flake of
oceanic crust (rising to a water depth of 25 m), to be the
source (Johnston, 1996). Seismic profiles reveal steep
deformation here (Sartori et al., 1994; Hayward et al.,
1999). The potential dimensions of a lithospheric scale
thrust fault for Gorringe Bank are about 200 km×80 km
(for a 50 km thick elastic lithosphere, and 40° fault dip).
entirely in oceanic lithosphere), a magnitude of 8.7 was
a more typical rigidity value of about 3×1010Pa, a uni-
form slip of 24 m would be required to produce an M8.7
More recently, authors have focussed on the Marquis
(Zitellini et al., 2001; Terrinha et al., 2002; Gracia et al.,
2003).Tsunami modelingsuggestsa sourcelocated atthe
Marquis de Pombal structure can provide a reasonable fit
for arrival time for most stations, though discrepancies
exist for the Gulf of Cadiz and NW Moroccan coast
(Baptista et al., 1998b). The primary objection to the
Marquis de Pombal source, however, is that the dimen-
sions of this structure (about 100 km×70 km) are too
etal.,2001).Anevent ofthismagnitude wouldnot befelt
in Hamburg or the Cape Verde Islands. Indeed, an em-
pirical relation has been established between fault rupture
length and moment magnitude (Wells and Coppersmith,
region of an M8.7 earthquake should be at least 200–
300 km in length (Fig. 9). Of the known tectonic struc-
tures off SW Iberia, only Gorringe and the subduction
fault plane have the necessary dimensions (200 km). The
Marquis de Pombal is 70–100 km in size and is alone
incapable of generating the appropriate seismic moment.
tectonic shortening is considered to be the result of
ary. Plate kinematic models estimate this motion to be
about 4 mm/a (Argus etal.,1989; Fernandesetal.,2003).
period of 3, 4 or 6 thousand years, respectively, would be
search for a candidate source, is what is the recurrence
time of great earthquakes like the 1755 event? Two inde-
pendent sets of sedimentological data provide clues as to
this recurrence time, turbidites in the abyssal plains, and
tsunami related overwash deposits in lagoons.
Sediment cores from the Horseshoe Abyssal Plain
have sampled a series of turbidite deposits covering the
entire abyssal plain with thicknesses ranging from
20 cm to over 2 m (Lebreiro et al., 1997). The volumes
of several of these turbidite flows, exceeds 1 km3. A
maximum of 21 individual turbidites were cored span-
ning roughly the period since 35 ka (Lebreiro et al.,
Fig. 7. Mean site-effect map using 1969 and 1964 earthquakes. Alg =
Algarve, VarMas = Variscan Massifs, GuB = Guadalquivir Basin,
IntBet = Internal Betic Zone, IntRif = Internal Rif Zone, WAS = West
Alboran Sea, RhB = Rharb Basin, AgB = Agadir Basin.
162 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
Fig. 8. Radial attenuation function and synthetic isoseimal maps, respectively, calculated for; A,B) a subduction fault plane source centered at 8° W,
35.5° N, C,D) a source at 8.5° W, 36° N, E,F) a source on the Marquis de Pombal, situated 100 km SWof Cape St. Vincent at 9.5° W, 36.5° N. In all
cases there is zero attenuation (constant intensity I) within the source region.
163 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
1997). The upper most turbidite in both the Horseshoe
and Tagus abyssal plains has been dated as being con-
temporaneous with the 1755 earthquake (Thomson and
Weaver, 1994). The older turbidites are likely to have
been triggered by great earthquakes in the past. The
chronology of these turbidite flows suggests a period-
icity of 1000–2000 years.
Coarse-grained deposits in the lagoon behind a sand-
spit near Cadiz were identified as overwash deposits
generatedbya tsunami,because the 6m highsandbarrier
is well above the maximum storm surge height. Dating
allowed this tsunami deposit to be correlated to the 1755
earthquake. An older tsunami deposit beneath the first
time on the order of 2000 years is indicated.
Thus, these two different types of sedimentological
databothsuggest a recurrencetimeonthe order of1000–
2000 years. This recurrence time is generally consistent
with the convergence rates indicated by GPS measure-
ments in the Morocco–Iberia region, suggesting west-
ward motion of stations in the Gibraltar block at
velocities of 5–10 mm/a (Reilinger et al., 2001; Mourabit
et al., 2002). For such velocities, 10–20 m of slip could
accumulate during the inter-seismic phase of the seismic
cycle along a locked fault zone.
The dimensions of the Gibraltar subduction fault
plane (roughly 200×200 km) are sufficient to generate
an earthquake of moment magnitude M=8.80 for a co-
Fig. 8 (continued).
Fig. 9. Empirical relation between fault rupture length and earthquake
moment magnitude (Wells and Coppersmith, 1994), with known
values for 1964, M9.2 Alaska earthquake and 1960, M9.5 Chile
earthquake and probable values for the 1755 Lisbon earthquake added.
164 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
seismic slip of 20 m or of M=8.64 for a co-seismic slip
of 10 m. Tsunami modeling indicates general agreement
with historical travel time and amplitude observations
for about half the cities in the study area. A subduction
source alone offers a better fit to the historical tsunami
reports for stations in the Gulf of Cadiz–Morocco area
than the Marquis de Pombal or other source further west
(Baptista et al., 1998b). However, late arrivals and low
amplitudes remain for the modeled tsunami along the
Portuguese west coast and for Safi (Morocco). Macro-
seismic analysis suggests a source at the NWof the Gulf
of Cadiz is best able to produce the historically observed
seismic intensities and that a subduction fault plane
alone may not be sufficient. But the subduction source
appears likely to have contributed to the 1755 earth-
quake in generating the tremendous seismic moment,
which no other tectonic structure in the region is large
enough to produce alone. Also, the recurrence interval
of great earthquakes in the region (1000–2000 years) is
better explained by a slip along the subduction fault
plane than other candidate sources discussed thus far. A
combination of the subduction fault plane and an
additional source further to the NW may offer the best
explanation for the historically observed tsunami and
seismic intensity patterns in the region. Historical
reports of multiple (2–3) ground shaking events over
the span of 10 min (Vilanova et al., 2003) support the
notion of a complex multiple rupture. One possible
scenario is that rupture of a ENE trending segment of the
Africa–Iberia plate boundary in the Gorringe—North-
ern Gulf of Cadiz area, may have triggered rupture along
the subduction fault plane.
We thank the French–Portuguese cooperation project
for funding the travel between France and Portugal and
the Portuguese MATESPRO project for organizing the
Lisbon meeting which helped initiate this collaborative
research project. We also thank the reviewers for their
constructive suggestions. And we particularly thank L.
Matias for the assistance in calculating the attenuation
functions and determining the site effects for the study
area. This article is a IUEM contribution 993.
Argus, D.F., Gordon, R.G., Demets, C., Stein, S., 1989. Closure of the
Africa–Eurasia–North America plate motion circuit and tectonics
of the Gloria fault. J. Geophys. Res. 94, 5585–5602.
Baptista, M.A., Heitor, S., Miranda, J.M., Miranda, P.M.A., Mendes
Victor, L., 1998a. The 1755 Lisbon; evaluation of the tsunami
parameters. J. Geodyn. 25, 143–157.
Baptista, M.A., Miranda, P.M.A., Miranda, J.M., Mendes Victor, L.,
1998b. Constraints on the source of the 1755 Lisbon tsunami
inferred from numerical modelling of historical data on the source
of the 1755 Lisbon tsunami. J. Geodyn. 25, 159–174.
Baptista, M.A., Miranda, P.M.A., Chierici, F., Zitellini, N., 2003. New
study of the 1755 earthquake source based on multi-channel
seismic survey data and tsunami modeling. Nat. Hazards Earth Sci.
Syst. 3, 333–340.
Buforn, E., Udias, A., Colombas, A., 1988. Seismicity, source mechan-
isms and tectonics of the Azores–Gibraltar plate boundary. Tectono-
physics 356, 89–118.
Calvert, A., Sandvol, E., Seber, D., Barazangi, M., Roecker, S.,
Mourabit, T., Vidal, F., Alguacil, G., Jabour, N., 2000. Geody-
namic evolution of the lithosphere and upper mantle beneath the
Alboran region of the western Mediterranean: constraints from
travel time tomography. J. Geophys. Res. 105, 10871–10898.
Chester, D.K., 2001. The 1755 Lisbon earthquake. Prog. Phys. Geogr.
Africa and Eurasia as derived from ITRF2000 and GPS data.
Geophys. Res. Lett. 30 (N.16), 1828, doi:10.1029/2003GL017089.
Fukao, Y., 1973. Thrust faulting at a lithospheric plate boundary: the
Portugal earthquake of 1969. Earth Planet. Sci. Lett. 18, 205–216.
Gjevik, B., Pedersen, G., Dybesland, E., Harbitz, C.B., Miranda, P.M.
A., Baptista, M.A., Mendes Victor, L., Heinrich, P., Roche, R.,
Guesmia, M., 1997. Modeling tsunamis from earthquake sources
near Gorringe Bank southwest of Portugal. J. Geophys. Res. 102
Gracia, E., Danobeitia, J.J., Verges, J., PARSIFAL Team, 2003.
Mapping active faults offshore Portugal (36°N–38°N): implica-
tions for seismic hazard assessment along the southwest Iberian
margin. Geology 31, 83–86.
Gutscher, M.-A., Malod, J., Rehault, J.-P., Contrucci, I., Klingelhoefer,
F., Mendes-Victor, L., Spakman, W., 2002. Evidence for active
subduction beneath Gibraltar. Geology 30, 1071–1074.
Gutscher, M.-A., 2004. What caused the Great Lisbon earthquake?
Science 305, 1247–1248.
Hayward, N., Watts, A.B., Westbrook, G.K., Collier, J.S., 1999. A
seismic reflection and GLORIA study of compressional deforma-
tion in the Gorringe Bank region, eastern North Atlantic. Geophys.
J. Int. 138, 831–850.
Holdahl, S.R., Sauber, J., 1994. Coseismic slip in the 1964 Prince
William Sound earthquake: a new geodetic inversion. Pure Appl.
Geophys. 142, 55–82.
Jimenez-Munt, I., Fernandez, M., Torne, M., Bird, P., 2001. The transition
from linear to diffuse plate boundary in the Azores–Gibraltar region:
results from a thin sheet model. Earth Planet. Sci. Lett. 192, 175–189.
Johnson, J.M., Satake, K., Holdahl, S.R., Sauber, J., 1996. The 1964
Prince William Sound earthquake: joint inversion of tsunami and
geodetic data. J. Geophys. Res. 101, 14,965–14,991.
continental regions — III. New Madrid, 1811–1812, Charleston
1886 and Lisbon 1755. Geophys. J. Int. 126, 314–344.
Lebreiro, S.M., McCave, I.N., Weaver, P., 1997. Late Quaternary
turbidite emplacement on the horseshoe abyssal plain (Iberian
margin). J. Sediment. Res. 67, 856–870.
Levret, A., 1991. The effects of the November 1, 1755 “Lisbon’’
earthquake in Morocco. Tectonophysics 193, 83–94.
Luque, L., Lario, J., Zazo, C., Goy, J.L., Dabrio, C.J., Silva, P.G.,
2001. Tsunami deposits as paleoseismic indicators: examples from
the Spanish coast. Acta Geol. Hisp. 36, 197–211.
165 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166
Mader, C., 1988. Numerical modelling of water waves. Los Alamos Download full-text
Series in Basic and Applied Sciences. 206 pp.
Martinez-Solares, J.M., Lopez, A., Mezcua, J., 1979. Isoseismal map of
the 1755 Lisbon earthquake obtained from Spanish data. Tectono-
physics 53, 301–313.
Mendes Victor, L., Baptista, M.A., Miranda, J.M., Miranda, P.M.A.,
1999. Can hydrodynamic modelling of tsunami contribute to
seismic risk assessment? Phys. Chem. Earth 24, 139–144.
Mourabit, T., Ben Sari, D., Reilinger, R., McClusky, S., Gomez, F.,
Barazangi, M., 2002. Preliminary evidence of active deformation
in Morocco from repeat GPS observations, Proceedings, Wegener
2002 Conference, June 2002, Athens, Greece (abstract).
Negredo, A., Bird, P., Sanz de Galdeano, C., Buforn, E., 2002.
Neotectonic modeling of the Ibero–Maghrebian region. J. Geophys.
Res. 107 (B11), 2292, doi:10.1029/2001JB000743.
Okada, Y., 1985. Surface deformation due to shear and tensile faults in
a half-space. Bull. Seismol. Soc. Am. 75, 1135–1154.
Park, J.-O., Tsuru, T., Kodaira, S., Cummins, P.R., Kaneda, Y., 2002.
Splay faults branching along the Nankai subduction zone. Science
Reilinger, R., McClusky, S., Ben Sari, D., Mourabit, T., Gomez, F.,
Barazangi, M., 2001. Active deformation in Morocco from repeat
GPS observations, EOS Trans. AGU, Fall Meeting 2001 suppl.,
Sartori, R., Torelli, L., Zitellini, N., Peis, D., Lodolo, E., 1994. Eastern
segment of the Azores–Gibraltar line (central-eastern Atlantic): an
oceanic plate boundary with diffuse compressional deformation.
Geology 22, 555–558.
of the 1700 Cascadia earthquake inferred from Japanese tsunami
descriptions. J. Geophys. Res. 108 (B11), 2535, doi:10.1029/
Smith, W.H.F., Sandwell, D.T., 1997. Global seafloor topography from
satellite altimetry and ship depth soundings. Science 277, 1956–1962.
Terrinha, P., Pinheiro, L.M., Henriet, J.-P., Matia, L., Ivanov, A.K.,
Monteiro, J.H., Akhmetzhanov, A., Volkonskaya, Cunha, M.R.,
Shaskin, P., Rovere, M., 2002. Tsunamigenic–seismogenic struc-
tures, neotectonics, sedimentary processes and slope instability on
the Southwest Portugese Margin. Mar. Geol. 195, 55–73.
Thiebot, E., Gutscher, M.-A., 2006. The Gibraltar Arc seismogenic
zone (part 1): constraints on a shallow east dipping fault plane
source for the 1755 Lisbon earthquake provided by seismic data,
gravity and thermal modeling. Tectonophysics 246, 135–152 (this
the emplacement time of recent deep-sea turbidites. Sediment. Geol.
Udias, A., Arroyo, A.L., 1970. Body and surface wave study of source
parameters of the March 15, 1964 Spanish earthquake. Tectono-
physics 9, 323–346.
Wells, D., Coppersmith, K., 1994. New empirical relationships among
magnitude, rupture length, rupture width, rupture area and surface
displacement. Bull. Seismol. Soc. Am. 84, 974–1002.
Zitellini, N., etal., 2001. Source of 1755 Lisbon earthquake and tsunami
investigated. Eos (Transactions, American Geophysical Union) 82,
166 M.-A. Gutscher et al. / Tectonophysics 426 (2006) 153–166