The 28th February 1969 earthquake and tsunami in the Atlantic Iberian margin

doi:10.21203/rs.3.rs-4959553/v1

Download PDF

Research Article

The 28th February 1969 earthquake and tsunami in the Atlantic Iberian margin

https://doi.org/10.21203/rs.3.rs-4959553/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

On the 28th February1969, a massive earthquake stroke SW Iberia and NE Morocco triggering a tsunami recorded in more than 20 tide stations. The event occurred in the SW Iberian margin, the same seismogenic area of the 1st November 1755 mega event. Several studies were developed in the last 55 years to address its earthquake mechanism and the corresponding tsunami source. In some cases, the study of the 1969 event was also the base for inferences regarding the 1755 earthquake and indirectly to give some light on tsunamigenic processes related with the SW Iberian margin. In this study, we present a comprehensive review of the tsunami data, taking advantage from the great improvement that occurred on the quality of the bathymetric data, particularly on the shallow areas close to the tide stations. We used a larger set of tide-records than previous studies. All records were digitized from the original mareograms and processed them according to modern standards. We address the possible landslide triggered at the NW coast of Morocco as the explanation of the tsunami observation at Casablanca. The new dataset combining both the earthquake and the landslide sources allows a better relocation of the tsunami source, enabling a quantitative comparison of the different source scenarios that have been developed for seismological research. The simulations presented here suggest that a thrust fault of 85 km x 20 km verging to the southeast is the best candidate to be responsible for the 1969 earthquake. The trace of this deep fault follows the one of the “Horseshoe Fault”, a northwest verging structure interpreted from the multichannel seismic data. Moreover, this deep structure may be accountable for both the 1969 event and the later 12th February 2007 M6 earthquake. Even more, the “Deep Horseshoe Fault” is a strong candidate to be the source of the 1st November 1755 event up to now elusive to multiple geological and geophysical studies.

Atlantic

1969 earthquake

tsunami

submarine landslide

The 28 February 1969 earthquake occurred at 2:41 am in the North-East Atlantic Ocean within the Horseshoe Abyssal Plain, southwest of Iberia. The Portuguese Seismological Authority issued the following statement using local time (UTC + 1): "an earthquake was recorded in the seismographic stations of Coimbra [central Portugal] and Lisbon, starting at 3h 41m 41.5s [and] 3h 41m 20.2s, respectively, and with the epicenter about 230 km to SW Lisbon. The magnitude of the earthquake is 7.3 on the Richter scale. The earthquake was felt with intensity VI-VII of the international scale in Lisbon and other locations on the mainland. In Lisbon, another earthquake was felt starting at 5:28 a.m. with intensity III of the international scale. The earthquake was recorded at the Serra do Pilar [northern Portugal] Seismographic Station at 3:41 am and 52".

The earthquake hit the south of the country and the Lisbon region, being the last major earthquake to occur in mainland Portugal, and the most important of the twentieth century. The epicenter of the earthquake was later determined as (36.01º N, 10.57º W) and was assigned the magnitude Ms = 7.9 and magnitude Mw = 8.0 after integration of data from the international network. The event caused alarm and panic among the population, disruption in telecommunications and electricity supply. There were 13 fatalities in mainland Portugal, 2 as a direct consequence of the earthquake, and 11 indirect (Quintino, 1969). Some victims are also reported in Morocco. The highest intensity (VIII) was felt in Algarve, being attributed intensity VI to Lisbon. It was felt up to 1300 km from the epicenter, particularly in Bordeaux (France), and the Canary Islands. Several aftershocks were recorded by the WWSSN station at Serra do Pilar registering 47 events between February 28 and March 24.

The generation area of this event is nowadays interpreted as the major tsunamigenic area in the North-East Atlantic Baptista (2020), being characterized by a complex puzzle of lithospheric blocks deformed by the compression between the Nubia and the Eurasian plates. Other significant events in the same area are the massive earthquake and transatlantic tsunami on the 1st November 1755 (Solares & Lopez-Arroyo 1979; 2004; Levret 1991; Jonhston et al. 1996; Baptista et al. 1998), the 31 March 1761 (Baptista et al. 2006; Wronna et al. 2019) and the instrumental earthquake of 12th February 2007 earthquake of magnitude 6.0 (Stich et al. 2007; 2010).

López-Arroyo & Udias (1972) interpreted the source as a southwest dipping fault, with an estimated depth between 22 km and 30 km. McKenzie (1972) reviewed the fault plane solutions for more than 100 earthquakes along the Africa-Eurasia plate boundary concluding that P waves alone couldn’t be used to select the fault planes. This was achieved by Fukao (1973) based on the analysis of long period surface waves. He computed an epicentre located at 36.01º N, 10.57º W, interpreting the focal mechanism as a thrust dipping to the northwest, with a small left-lateral strike slip component, and a depth 22 ± 6 km (see details in Table 1).

Grimison & Chen (1988) suggested that the earthquake included two subevents, being the second one a strike-slip. The macroseismcity associated with the 1969 earthquake was compiled and published for Portugal and Spain by López-Arroyo & Udías (1972), Mendes (1974), Martínez-Solares et al. (1979), Mezcua (1982), Paula & Oliveira (1996), Solares & López-Arroyo (2004) and Pro et al. (2020). Grandin et al. (2007) studied a set of source scenarios depicted in Table 1, to show that it is possible to reproduce the intensity values considered that the earthquake had a focal mechanism with strike = 233°; dip = 49.5°, rake = 63.5°, and a rupture zone having the length of L = 82.5 km, the width W = 35 km and the average slip of 4 m, which is equivalent to a seismic moment of M₀ = 6 × 10²⁰ Nm. Buforn et al. (2020) reanalyzed the seismic intensities using an improved velocity model and re-located the epicentre of the main shock as 36.04º N, 10.59º W, at a depth of 37 km. This location is only 4 km to NNW of Fukao’s determination.

The tsunami that followed the 28th February 1969 earthquake impacted the coast during the night, in low tide conditions. The tsunami was barely noticed by the populations, but was recorded in the tide gauges of Portugal, Spain, Morocco and United Kingdom. In Portugal, only some fishermen remark unusual waves at the shore (see Baptista & Miranda 2009). The Moroccan press refers to a small "Raz de marée" in Rabat and Casablanca. At Rabat, the sea rose 80 cm. In Casablanca, the same journal describes the tide record saying that at 23:53 (27 February) the tide was at 2.70 m, and 2.40 m at 02:45 (28 February) but at 03:15 the tide gauge suffered a sudden push of one meter. The report finishes by stating that small Spanish boats that felt the shaking at sea sought shelter in the harbour of Casablanca (Le Petit Marocain, 1969).

Studies on the numerical simulation of the tsunami include Heinrich et al. (1994), Guesmia et al. (1996), Guesmia et al. (1998), Miranda et al. (1996) and Gjevick et al. (1997). All these studies were based on the fault mechanism suggested by Fukao (1973), a thrust with a small strike slip component, striking N55ºE, parallel to the Gorringe Bank (see Fig. 1); its fault dimensions are 80 km x 50 km, the dip is 52º, and the average slip is 3.85 m, corresponding to a seismic moment of 6.0x10²⁰ Nm and a magnitude 7.9, considering a rigidity of 5x10¹⁰ Pa. All these simulations fairly reproduce the observations in what concerns the first arrival and the amplitude of the first peak. All studies discard a part of the mareograms available. Casablanca mareogram is not reproduced by any author, even in terms of the polarity of the first wave, using only the earthquake as the single source.

Heinrich et al. (1994) shallow water model does not reproduce well enough the first wave in Cascais; the discrepancies are attributed to poor bathymetry or to the instruments’ response. Guesmia et al. (1996) finite element Boussinesq model tests only six tide stations and fails to reproduce the observation in Cadiz (the amplitude of the first peak) and the arrival time at Terreiro do Paço, close to Lisbon downtown. Gjevik et al. (1997) show that tsunami travel times are coherent with a source that encompasses the earthquake epicenter deduced by Fukao (1973). Their finite difference Boussinesq model is not able to reproduce the observations at Las Palmas de Gran Canária, Faro (named herein Culatra), Cadiz and Lagos. Renou et al. (2011) and Allgeyer et al. (2013) both use Fukao’s source parameters to propagate the tsunami along the North coast of Morocco and the French Atlantic coast, respectively; they are not able to reproduce the tsunami polarity at Casablanca.

Pires & Miranda (2001) developed a method for the inversion of tsunami waveforms and use it to optimize the source parameters of the 1969 horseshoe tsunami. They suggest a change of the initial Fukao (1973) source parameters, taken as “first guess” for the inversion scheme, to an optimal finite source, a pure thrust, 113.5 km long, 29.3 km wide, strike 53º, depth 23.5 km, dip 68º, rake 90º and slip 2.97 m. This optimal source provides improved fit with respect to the observations when compared to Fukao solution. Casablanca mareogram is only possible to reproduce considering the hypothesis raised by Gjevik et al. (1997) that the earthquake could have triggered a landslide close to the Moroccan coasts. The addition of the landslide effect slightly degrades the global misfit obtained for the optimal solution (Pires & Miranda 2001). Once again, the differences between observed and synthetic travel times are interpreted as the result of instrumental errors and poor resolution of the bathymetry.

Table 1

Source scenarios following Grandin et al. (2007).

Source A1 corresponds to Buforn (1988) source modified by Grandin et al. (2007), sources A2 and C2 correspond to Buforn et al. (1988) and Fukao (1973), respectively. In our simulations we use 10.395º W, 35.845º N, the minimum obtained by BRT, as the mean location of the upper edge of the finite fault that represents the rupture area. In the table below, depth refers to “depth to the top” of the fault.
Source Nº	Depth (km)	L (km)	W (km)	Strike	Dip	Rake	Slip
A1	8	85.0	20	221.0	47.0	44.0	5.29
A2	8	85.0	20	231.0	47.0	54.0	5.29
A3	8	85.0	20	241.0	47.0	64.0	5.29
B1	8	82.5	35	223.0	49.5	53.5	4.16
B2	8	82.5	35	233.0	49.5	63.5	4.16
B3	8	82.5	35	243.0	49.5	73.5	4.16
C1	8	80.0	50	225.0	52.0	63.0	3.85
C2	8	80.0	50	235.0	52.0	73.0	3.85
C3	8	80.0	50	245.0	52.0	83.0	3.85

The tsunami was recorded by tide stations in Portugal mainland, Morocco, Azores Islands, Spain mainland and Canary Islands and United Kingdom (see Fig. 1 for locations). A total of 21 waveforms were obtained in Portugal, Spain, France, Morocco and the United Kingdom (see Table 2). The tide records of Saint Mary (Scilly Islands) and Gibraltar in the United Kingdom, Ceuta and Sanlucar de Barrameda in Spain, and Aveiro in Portugal, are used for the first time. Due to their location, the stations of Saint Mary, Gibraltar and Ceuta impose additional constraints to the north and the east in the location of the source.

Table 2

Tsunami records used in this study. Locations were obtained from the description of the tide stations. Depths were deduced from the bathymetric grid with horizontal resolution of 0.01 degrees. Observed travel-times were deduced from the de-tided waveforms. Depths refer to the closest bathymetric grid node. Arrival times and first peak amplitudes for the landslide correspond to the source depicted in Fig. 5.
Station	Acronym	East Longitude	North Latitude	Depth (m)	Observed TTT (min)	Landslide Modelling
Station	Acronym	East Longitude	North Latitude	Depth (m)	Observed TTT (min)	TTT (min)	1st peak (m)
Terreiro do Paço	TPA	-9.13	38.71	4.9	67	95	-0.00
Aveiro Sacor	AVS	-8.70	40.66	4.2	109	-	-
Cascais	CSC	-9.41	38.69	11.0	39	71	0.03
Lagos	LGS	-8.66	37.10	9.0	38	58	0.05
Casablanca	CSB	-7.60	33.62	15.4	48	48	0.37
St Cruz de la Palma	STC	-17.75	28.67	11.6	91	104	0.01
Puerto de la Luz	PDL	-15.42	28.13	4.9	94	103	0.01
Angra do Heroismo	ADH	-27.21	38.65	18.0	139	-	-
Horta	HRT	-28.62	38.53	7.7	151	-	-
Cadiz	CDZ	-6.28	36.54	3.4	77	80	0.02
Coruna	CRN	-8.39	43.36	-	141	-	-
Ceuta	CTA	-5.32	35.89	3.8	76	72	0.00
Culatra	CLT	-7.87	36.98	2.6	42	54	0.02
Sanlucar de Barrameda	SB	-6.39	36.78	4.6	101	96	0.02
Chipiona	CHP	-6.44	36.75	7.4	75	83	0.03
Arrecife	ARR	-13.52	28.96	18.5	-	-	-
Pedrouços	PDR	-9.23	38.69	11.4	52	86	0.01
Leixoes	LXS	-8.71	41.17	9.3	85	116	0.01
Cabo Ruivo	CBR	-9.09	38.75	-	87	110	0.00
Gibraltar	GBR	-5.36	36.13	0.4	72	70	0.00
St Mary	STM	-6.32	49.92	3.0	235	-	-

The tide stations closest to the epicenter lie along the west and south coast of Portugal, the Gulf of Cadiz and Morocco. The tsunami was recorded as far north as the Scilly Islands, and to the South in the Canary Islands; the westernmost records are those in the Azores (Horta and Angra) and the easternmost are Ceuta and Gibraltar. Tide-records from Belfast (UK) and La Rochelle (France) were obtained and analyzed, but they do not show any sizable tsunami signal. This confirms the conclusions already presented by Allgeyer et al. (2013) and so these records are not presented here.

All records were re-digitized from the original paper records. Digitized waveforms were linearly interpolated into a one-minute time step and redrawn over the original records to check their accuracy. De-tiding was made using a least squares algorithm to fit the interpolated data into polynomials whose degrees guaranteed an adjusted R-square parameter above 0.995. To further ensure the goodness of our de-tiding a band-pass filter was applied. The parameters of the filter are: end of the first stop band at 91 min, beginning of first pass-band at 89 min, end of the pass-band at 5 min and beginning of the second stop band at 2 min. Details are given in Baptista et al. (2016). All used de-tided waveforms are represented in Fig. 2, with the same horizontal scale (between 0 and 20000 sec after the earthquake occurrence) and variable vertical scale to account for the varying amplitudes. De-tided tsunami waveforms show a tsunami signal above the noise level and in most cases sharp tsunami arrival times can be directly deduced from the de-tided records.

To further support the assessment of arrival times, and detect frequency changes associated with first arrivals, wavelet spectral analysis was made with Morlet’s continuous wavelet approach:

$$\:{W}_{{\Psi\:}\left|f\right.}\left(a,b\right)={\int\:}_{-\infty\:}^{+\infty\:}\frac{1}{\sqrt{a}}{{\Psi\:}}^{*}\left(\frac{t-b}{a}\right)f\left(t\right)dt\:\:\:\:a,b\in\:\mathcal{R},a>0$$

Where the mother function is given by a plane wave modulated by a Gaussian:

$$\:{{\Psi\:}}_{0}\left(\eta\:\right)={\pi\:}^{-1/4}{e}^{i{\omega\:}_{0}\eta\:}{e}^{-{\eta\:}^{2}/2}\:\:$$

ω₀ is the non-dimensional frequency, here taken to be 6, to satisfy the admissibility condition (see details in Baptista et al. 2016). The combination of the direct analysis of the tsunami records with their wavelet spectra (Fig. 3) allowed the assessment of arrival times for all tide stations, presented in Table 2.

3.1 Source location using tsunami travel time data

The simplest way to compute the approximate location of the tsunami source area is by Backward Ray Tracing (Baptista et al. 1998). This technique is based on the reversibility of the linear tsunami propagation: each tide station is considered as a point source generating a forward travel-time map. After the subtraction of the observed tsunami travel-time observed at the station, the zero contour corresponds to the set of all possible source locations. Using the same procedure for all stations, the minimum of the root mean square error is the most probable location for the tsunami source. The minimum is usually not zero due to instrumental errors, the finite extension of the source and the non-linearity of the tsunami propagation.

The tsunami travel time dataset used in the Backward Ray Tracing simulation uses 17 out of the 21 stations shown in Table 2: the record of Cabo Ruivo (CBR) does not permit the identification of the first arrival due to its low signal to noise ratio; the record of Arrecife (ARR) has no data between 0 a.m. and 3 a.m. on the 28th February and so cannot be used to infer the tsunami arrival time; La Coruña (CRN) tide station is located in a long narrow inlet (ria) that is not correctly represented in our bathymetric model; finally, Casablanca (CSB) was excluded because both the arrival time and the first wave polarity are incoherent with all other tide stations, as described above. The locations of the different stations were carefully checked, using the new bathymetric compilation and their descriptions mostly available on the web.

The bathymetric grid used in the Backward Ray Tracing simulation has a resolution of 0.005⁰ (approximately 0.5 km in latitude). Results show that it is possible to locate a source area encompassing the earthquake epicenter with an average travel-time error of 5.6 minutes. The minimum is located at 10.415º W, 35.835º N (see Fig. 4) using a dataset of 17 stations, while previous studies needed to discard part of the observations (Heinrich et al. 1994; Guesmia et al. 1996; Gjevik et al. 1997; Pires & Miranda 2001). It is worth mentioning that the set of stations used for this calculation guarantees a good azimuthal coverage. The minimum of the misfit error of the backward raytracing is located 24 km SE of Fukao (1973) epicenter (see Fig. 4).

3.2 Casablanca Landslide

The first-wave tsunami polarity at Casablanca is incoherent with what is observed in nearby stations. It was early interpreted by Gjevik et al. (1997) as the consequence of a local secondary landslide triggered by the earthquake. To explore this hypothesis a series of simulations were made to find a landslide source able to reproduce the 55 cm retreat observed at Casablanca tide gauge 54 minutes after the earthquake origin time. Numerical simulation of the landslide dynamics and the ensuing tsunami generation and propagation was made using an in-house developed multi-layer visco-plastic shallow water numerical model, solved in a finite volume scheme (details in Omira & Ramalho 2020). In this numerical code, the landslide dynamics is governed by the visco-plastic Bingham model (Imran et al. 2001), and the wave generation and propagation are approximated by the non-linear shallow water equations. For the Casablanca landslide the yield (or Bingham) stress ($\:{\tau\:}_{\text{y}}$) is considered as $\:{\tau\:}_{\text{y}}=10\:kPa$.

The morphology of the platform of Morocco, and the sharp leading depression observed at Casablanca favours a landslide located west of this harbour, moving downslope in an east-west direction. We performed numerical tests considering different locations, volumes and onset times for the landslide, starting with the basic hypothesis that the landslide was triggered by the earthquake shock at 02:41:32, as there are no aftershocks recorded by the seismic network before 04:25:34.1 (López-Arroyo & Udias 1972).

We were unable to find any solution enough far away from Casablanca, to be triggered at 02:41:32, which would correspond to the tsunami travel time depicted in Table 2, and able to generate a leading wave arriving at the tide gauge shortly prior to the “seismic” tsunami. The observed waveform can only be fairly reproduced with a source closer to Casablanca and triggered several minutes after the earthquake origin time. Our preferred solution corresponds to a submarine landslide with a volume of 3 km³, located 50 km off Casablanca, occurring ~ 15 min after the earthquake. Additional parameters of the landslide model selected in this study are listed in Table 3. The comparison between the waveform generated at Casablanca and the actual mareogram is shown in Fig. 5. We also show, using as example one of the sources discussed in the next section (source A1), that the tsunami generated by the co-seismic deformation arrives to Casablanca much later, and with opposite polarity.

Table 3

Parameters of the landslide source
Total volume (km³)	Maximum thickness (m)	Depth of initial trough (m)	Depth of initial bulge (m)	Average sliding slope (º)	Runout distance (km)	Yield stress (kPa)
3	20	200	1000	2.2	25	10

The tsunami waveforms observed by the tide gauges must be a combination of effects from both the landslide and the earthquake sources. In Table 2 we compare observed arrival times with ones that would be generated by the landslide. We also indicate the first peak synthetic amplitude. We can expect significant overlap in the case of CDZ, SB, CEU and GBR, but only in the first two cases the amplitude is larger than 1 cm, and even in CDZ and SB it is three to four times smaller than the amplitude of the observed leading wave. To the north, the effect of the landslide on the tsunami waveforms either does not affect the first waves or has an amplitude much smaller than the one generated by the co-seismic deformation.

3.3 Selecting the “best Okada-type” tsunami source

Tsunami modelling is made with NSWING code based on the discretization and explicit leap-frog finite difference scheme to solve the shallow water equations in spherical and Cartesian coordinates developed by Liu et al. (1998). Details and previous studies developed with this code can be found in Miranda et al. (2014), Wronna et al. (2015), Omira et al. (2015) and Baptista et al. (2016).

The bathymetry data grid used in the simulations has a 0.01º cell size; near the harbours where the tide stations are located, swath bathymetric maps or harbour charts were used to refine the seafloor morphology and ensure a fair representation of local conditions within the limits of the horizontal resolution.

Sea surface deformation at time t = 0 s, used to initiate the simulation, assumes that the initial sea surface elevation mimics the co-seismic deformation of the ocean bottom. Mansinha & Smiley (1971), Okada (1985) established the analytical equations to compute the elastic deformation of a rectangular fault with a set of fault parameters: Length, Width, strike, rake, dip and depth of the fault. All candidate source solutions presented by Grandin et al. (2007), presented in Table 1, were studied and the synthetic tsunamis waveforms were compared with the observations, to select the “Best Okada-type” source.

The visual comparison between synthetic and observed waveforms is simple: it takes into account the amplitude, the frequency content and the arrival time, in a more complex way than what is done by a simple misfit computation between the synthetic and the observed waveforms. One of the most important problems regards the existence of a “time lag” between the two waveforms that can be the consequence of grid resolution or clock errors in tide gauges. It is worth to mention that small clock errors, which are irrelevant in terms of normal oceanographic harbour operation, are critical for numerical tsunami modelling, and cannot be accurately estimated neither from the identification of the earthquake signal (not represented in most of the tide gauge records) nor from the comparison between the observed and the synthetic tide wave. In this study the largest “time-lags” correspond to Cadiz and Sanlucar de Barrameda both shallow areas poorly represented by the morphological grid.

To take into account the occurrence of such “time lags” we computed the cross-correlation between each observed signal and the synthetic one produced by all candidate sources, assuming that the maximum of the cross-correlation corresponds to the best computation of the “time lag”. We used a 20minute window centered in the assumed arrival time for each station and all results were visually checked. We were unable to compute robust estimations for AVS, PDL, ARR and LXS because the leading wave isn’t clear enough. Table 4 shows the results for the remaining 14 stations and all candidate sources.

Time lags are fairly constant and almost independent of the input source (see Table 4). Thus, we can compute misfit errors, correcting the synthetic signal from the “time lag” and using the differences corresponding to the first 30 data points after the tsunami arrival time. Therefore, the later arrivals that are mainly a consequence of multiple propagation accidents close to the shore are not used to assess the quality of the different candidate sources, and the focus is made on the waves more informative of the source.

Results shown in Table 4 and Fig. 6, indicate that A-sources are better than B-sources, and B-sources better than C-sources. The lowest average fit error corresponds to source A1 (see Table 4 and Fig. 6) with a slightly small average misfit than A2, both close to the seismological analysis made by Buforn et al. (1998).

Table 4

Computation of time lags and misfit errors for 16 stations of the network and the nine source models proposed by Grandin et al. (2007). For each model the time lag corresponding to the least misfit is indicated in minutes. The misfit is computed as the mean square error between observed and modelled data, over 20 points, starting from the assumed tsunami travel time.
Station	TT	A1		A2		A3		B1		B2		B3		C1		C2		C3
Terreiro do Paço	67	0	0.039	0	0.028	1	0.029	1	0.035	1	0.041	1	0.063	1	0.067	2	0.096	2	0.118
Cascais	39	0	0.114	1	0.114	1	0.139	1	0.120	1	0.126	1	0.177	1	0.138	1	0.160	2	0.206
Lagos	38	2	0,094	3	0,111	3	0,111	2	0,128	2	0,140	3	0,147	2	0,173	3	0,184	4	0,195
Angra Heroismo	139	4	0.044	4	0.050	4	0.058	5	0.081	5	0.087	5	0.092	5	0.109	5	0.109	5	0.112
Horta	151	3	0.045	5	0.047	5	0.047	7	0.057	7	0.058	6	0.059	8	0.065	7	0.067	7	0.070
Cadiz	77	12	0.024	13	0.025	13	0.023	12	0.037	13	0.036	13	0.034	13	0.049	13	0.047	14	0.045
Ceuta	76	-4	0.016	-4	0.016	-5	0.015	-5	0.015	-4	0.015	-4	0.015	-4	0.016	-4	0.016	-4	0.015
Culatra	42	3	0.085	3	0.087	3	0.095	3	0.116	3	0.123	4	0.129	3	0.150	7	0.151	7	0.150
Sanlucar de Barrameda	101	-10	0.069	-10	0.073	-10	0.074	-10	0.100	-10	0.100	-9	0.098	-10	0.123	-10	0.120	-15	0.117
Chipiona	75	3	0.089	3	0.090	4	0.092	3	0.109	4	0.112	4	0.110	4	0.140	-8	0.138	5	0.130
Pedrouços	52	3	0.078	4	0.067	4	0.050	4	0.054	5	0.042	5	0.049	5	0.076	5	0.080	5	0.090
Gibraltar	72	2	0.005	2	0.005	3	0.004	2	0.004	3	0.004	3	0.004	2	0.007	2	0.007	3	0.008
St Mary	235	-8	0.013	-7	0.014	-7	0.017	-7	0.030	-6	0.031	-6	0.032	-6	0.049	-5	0.051	-6	0.053
Average			0.055		0.056		0.058		0.068		0.070		0.077		0.089		0.094		0.100

The 28th February 1969 event was the object of many studies focused not only on the analysis of this particular event, but also to infer the seismogenic process related with the 1755 “Lisbon” earthquake.

The geologic complexity of the SW Iberian margin, marked by deep basins and oceanic ridges, challenged geologists and geophysicist in search for the mechanisms beyond the triggering of such mega-events (Zitellini et al. 1999; 2001; Gracia et al. 2003). Many bathymetric and multi-channel seismic surveys were performed, leading to the identification of NE-SW trending thrusts thought to be able to generate large earthquake events: the Horseshoe Fault, the Marques de Pombal Fault and the Portimão–Guadalquivir Bank Fault, dipping southwest and Coral Patch faults (Zitellini et al. 2001; 2004; 2009; Gracia et al. 2003), and even tectonic activity related with the Cadis Wedge (Gutscher et al. 2002). Detailed studies focused on some of these structures, looking for markers of neo-tectonic activity which could be related with seismic mega-events, namely the 1st November 1755 earthquake (see Fig. 7 for locations).

The large amount of tsunami descriptions available from the Atlantic Basin (Baptista et al. 1998) increased the significance of tsunami research as a tool to address the study of the earthquake and tsunami sources. Data from Zitellini et al. (1999) multi-channel seismic survey up to 12–15 km depth, showing that Marques de Pombal Fault is an active structure dipping to the east, were used to design a possible complex source for the 1755 tsunami (Baptista et al. 2003). The Horseshoe Fault was also the target of tsunami modelling studies (Omira et al. 2009; Baptista et al. 2011) and its detailed analysis by Martinez-Loriente (2018) based on high-resolution seismic data up to a depth of 10 km, showed its neo-tectonic activity, being interpreted as a NW trending thrust-fault dipping to southeast. Other authors used different combinations (or extrapolations) of geological structures but none was able to fairly cope with all existing historical descriptions and a few field surveys targeting the tsunami deposits in southern Portugal.

The 1969 event is the largest instrumental earthquake and tsunami generated in this tectonic environment. Nevertheless, despite the large amount of geophysical data recovered since the nineties its geological source has been elusive to map. As described above most of the candidate sources are small variations to the original Fukao (1973) solution, bearing no clear relationship with the main tectonic units mapped by seismic stratigraphy. We arrive to a paradox situation where the SW Iberian margin is the site of mega-earthquakes and transatlantic tsunamis, and where large tectonic faults have been mapped by geophysical methods, but with no direct connection between them.

This study of the 28th February 1969 tsunami tries to challenge this apparent paradox, refining what can be obtained from tsunami data and checking it against up-to-date tectonic mapping and seismological research developed in the last decades.

Tsunami analysis using modern digitization tools and wavelet analysis made possible the compilation of a dataset of 18 well constrained tsunami travel times, distributed along the NE Atlantic coast. The combination of this new set of tsunami waveforms with high-resolution bathymetry made possible the location of the tsunami source area without discarding stations located in complex and shallow bathymetric environments, as done by previous studies. Results presented in Fig. 2 show that it is possible to find a Backward Ray Tracing solution encompassing the earthquake epicenter without discarding the stations of Las Palmas de Gran Canaria, Faro (named herein Culatra) Cadiz and Lagos as concluded by Gjevik et al. (1997) and Pires & Miranda (2001). The minimum of the misfit error is 5.6 minutes at 10.415º W, 35.835º N, on the Horseshoe fault trace, approximately 24 km southeast of the epicenter given by Fukao (1973).

An important point concerns the existence of a landslide close to the coast of Morocco as the source of the tsunami waveform record at Casablanca. Our results suggest that the earthquake shaking destabilized the slope sediments creating a weak layer favouring the occurrence of a landslide. Few minutes later (~ 15 min), the landslide failure is initiated and its downslope movement (~ 5 min duration) generated a tsunami wave up to 1 m in height (Fig. 5) that propagated towards Casablanca coast. No aftershocks were recorded by the seismological network 10–15 min after the main-shock but the extensive modelling made for this study is strong enough to support the hypothesis that a 3 km³ landslide, located 50 km off Casablanca is the responsible for the waveform recorded at Casablanca tide gauge. In the absence of a post-event bathymetric survey to confirm the landslide, nearby stations' seismograms were inspected. However, this search turned unsuccessful as all nearby stations saturated for more than one hour following the earthquake (Buforn, Batlló, pers. communication).

Earthquake-triggered submarine landslides, occurring within distinct continental margins and involving failure volumes ranging from few cubic kilometres (i.e., 1998 Papua New Guinea, ~ 4 km³ volume (Synolakis et al. 2002)) to hundreds of cubic kilometers (i.e., 1929 Grand Banks, ~ 200 km³ (Piper et al. 1999)), account for a common source of tsunami. The 1998 Papua New Guinea event is a classic example of landslide-tsunami that caused a heavy local impact, ~ 15 m of runup, despite the relatively small volume of the landslide volume (~ 4 km³) (Synolakis et al. 2002). Although of comparable volumes, the landslide off Casablanca resulted in small tsunami compared to the 1998 Papua New Guinea event. The reason for this difference mainly lies in considering a thin landslide (20 m of max. thickness). Other factors, such as the landslide location and its dynamics when moving downslope, have also a significant control on the generated tsunami.

The shallow water simulations using an Okada type source computed from the parameters provided by seismology succeed in reproducing the observed tsunami in all stations (see Fig. 4) except in Casablanca. Among all tested sources presented in Table 1, Source A1–85 km x 20 km, strike 221º, rake 44º and dipping 47º to northwest is the one that better reproduce tsunami observations. This conclusion is coherent with the source dimensions proposed by López-Arroyo & Udias (1972) from the aftershock sequence, and the results of the inversion procedure developed by Pires & Miranda (2001) who suggested that observations could be better reproduced if Fukao (1973) finite fault parameters were replaced by a longer and narrower finite fault.

This conclusion is different from the one deduced by Grandin et al. (2007) from the numerical simulation of macroseismic intensities. In their study the best agreement is found for source B2, even if it is also concluded that source A2 reproduces well the observations. In the analysis presented here we show that, from the tsunami point of view, sources “A” are systematically better than sources “B” (see misfits for all stations in Table 4). Both studies agree that sources “C” which correspond to Fukao (1973) solution, and some small variations of parameters, underperform for both macroseismic and tsunami modelling.

What is the geological structure associated with the 1969 event? Despite the fact that the minimum misfit of the BRT solution matches the Horseshoe Fault trace, we address two different structures. The Horseshoe Fault described by Martinez-Loriente et al. (2018) is interpreted as NW-verging active thrust affecting the whole sedimentary sequence up to the seafloor, where it generates a scarp higher than 1 km (Martinez-Loriente et al. 2018). The top of the fault used for the tsunami simulations is located at 8 km depth, suggesting that the seismic rupture area of the 1969 earthquake should be located at a depth greater than 10 km, on a northwest dipping fault plane (see Fig. 7).

An analogous situation was discussed by Stich et al. (2007) when analyzing the source of the 12th February 2007, Mw 6.0 Horseshoe earthquake. This is the largest event occurring beneath the Horseshoe Abyssal Plain that was recorded by the global and regional modern seismic broadband networks. Stich et al. (2007) conclude that this event took place on an oblique reverse fault 40 km below seafloor, subparallel to the trace of Horseshoe Fault, but dipping in the opposite direction, and attribute it to a secondary fault in the footwall. Other studies on this event include Custódio et al. (2012) and Pro et al. (2013). The finite source properties of the 12th February 2007 earthquake show a larger than average stress drop, which reflects the high rigidity of the source area ($\:\mu\:\sim70\:GPa)$ similar to what was found by Fukao (1973) for the 28th February 1969 event. The moment tensor inversion conducted by Stich et al. (2010) locate the epicenter at 10.31º E, 35.90º N, a depth of 50 km, and two nodal planes defined by strike 122º, dip 55º, rake 147º and strike 232, dip 63º, rake 40º. This source shares similarity with sources A1 and A2, and points to a deep fault plane, below the basement mapped by Martinez-Loriente et al. (2018), questioning the interpretation, from multi-channel seismic sections, that the Horseshoe Fault is rooted at the Moho (considered to be located about 9 km depth below the seafloor), or at the uppermost mantle.

When discussing the source of the large earthquakes that affect the SW Iberian margin it is virtually impossible to not address the source of the 1755 “Lisbon” earthquake and tsunami. Most of the candidate sources studied by different authors (Zitellini et al. 2001; 2004; 2009; Gutscher et al. 2002; Baptista et al. 2003; Gracia et al. 2003; Omira et al. 2009; Baptista et al. 2011) suffered from the difficulty to identify a geological structure large enough to trigger a mega event. One of the outcomes of the re-analysis of the 1969 earthquake and tsunami is that the “Deep Horseshoe Fault” can provide such a solution. It is a deeply rooted active structure, able to trigger very large earthquake events with tsunamigenic potential.

The 28th February 1969 tsunami data is the largest dataset concerning tsunami generation in the Gulf of Cadiz area. Waveforms from a large number of tide gauge stations, with good azimuthal coverage, are used to study the generation and propagation of this event, giving additional information on the earthquake source, the largest instrumental event in the southwest Iberian margin. Despite the amount of research studies conducted in the last 50 years focusing both the earthquake and the tsunami, the re-analysis of this dataset provides the opportunity for a new focus on the source and on the implications concerning the triggering of mega-events in this important source area.

Main conclusions can be summarized as:

The point source which better reproduces tsunami data arrival times is located at 10.415º W, 35.835º N, approximately 20 km south-southeast of Fukao (1973) or Buforn et al. (2020) epicenter, and close to the trace of Horseshoe Fault. This solution is coherent with waveforms from 18 different tide stations, except Casablanca.
Casablanca waveform can be fairly reproduced by a landslide initiated 10–15 min after the main shock and located ~ 50 km to the west of the tide gauge. Its effects on the whole network have amplitudes smaller than 5 cm, and meaningful only for the mareographs located at Cadiz, Sanlucar de Barrameda, Ceuta and Gibraltar, where its arrival overlaps the leading waves of the tsunami generated by the co-seismic deformation associated with the main shock. Nevertheless, only in Cadiz and Sanlucar de Barrameda its amplitude is larger than 1 cm and, in all cases, amplitudes are much smaller than the observed waveforms.
Among the nine solutions studied by Grandin et al. (2007), the “A” sources better reproduce the observations than all others. Source A1, with a length of 85 km, a width of 20 km, strike 221º, dip 47º, rake 44º and slip 5.29 m corresponds to the “Best Okada Source”. This result agrees with the seismological analysis of Buforn et al. (1998).
The two larger instrumental earthquakes that took place in the southwest Iberian margin, the 28th February 1969 and the 12th February 2007, have their epicenters close to the trace of the northwest verging Horseshoe Fault, and both show similar mechanisms: southeast verging thrusts, rooted below the Moho, on the lower crust or the uppermost mantle. This mechanism is different from what is inferred from multichannel seismic studies.
Among all hypotheses put forward for the source of the 1st November 1755 mega-earthquake and tsunami, a major fault which trace follows approximately the Horseshoe Fault, but deeper rooted and verging to the southeast, can provide a “single source” which has the geological and rheological characteristics to be considered as the most probable candidate solution for this megaevent. This “Deep Horseshoe Fault” must be considered for seismic and tsunami risk assessment.

Author Contribution

M.A.Baptista led the work and wrote the first draft of the manuscript with the contributions of J.M.Miranda.M.A. Baptista collected and interpreted the tsunami records. Rachid Omira modelled the tsunami propagation and landslide.M.A.Baptista, J.M.Miranda and Rachid Omira contributed to the interpretation of the results and to the final writing of the manuscript.

Acknowledgement

The authors wish to thank Elisa Buforn and Josep Batlló for making available the seismograms of Rabat, Averroes and Barcelona and Frederic Dias and Daniel Giles for kindly providing several tide records used in this study.The authors wish to thank FCT, the Portuguese Science Foundation for funding received through project project MAGICLAND - MArine Geo-hazards InduCed by underwater LANDslides in the SW Iberian Margin (Ref: PTDC/CTA-GEO/30381/2017), funded by the Fundação para a Ciência e Tecnologia (FCT), Portugal and

Allgeyer, S., Daubord, C., Hébert, H., Loevenbruck, A., Schindelé, F. & Madariaga, R., 2013. Could a 1755-like tsunami reach the French Atlantic coastline? Constraints from twentieth century observations and numerical modeling, Pure Appl. Geophys., 170(9-10), 1415-1431. https://doi.org/10.1007/s00024-012-0513-5.
Baptista, M. A., 2020. Tsunamis Along the Azores Gibraltar Plate Boundary, Pure Appl. Geophys., 177, 1713–1724. https://doi.org/10.1007/s00024-019-02344-8.
Baptista, M. A., Miranda, P. M. A., Miranda, J. M. & Victor, L. M., 1998. Constrains on the source of the 1755 Lisbon tsunami inferred from numerical modelling of historical data on the source of the 1755 Lisbon tsunami, Journal of Geodynamics, 25(1-2), 159-174.
Baptista, M. A. & Miranda, J. M., 2009. Revision of the Portuguese catalog of tsunamis, Nat. Hazards and Earth Syst. Sci., 9 (1), 25-42.
Baptista, M. A., Miranda, J. M., 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 Syst. Sci., 3, 333–340, https://doi.org/10.5194/nhess-3-333-2003.
Baptista, M. A., Miranda, J. M. & Luis, J. F., 2006. In search of the 31 March 1761 earthquake and tsunami source, Bull. Seismol. Soc. Am., 96(2), 713-721.
Baptista, M. A., Miranda, J. M., Omira, R., & Antunes, C., 2011. Potential inundation of Lisbon downtown by a 1755-like tsunami, Nat. Hazards and Earth Syst. Sci., 11, 3319–3326.
Baptista, M. A., Miranda, J. M., Batlló, J., Lisboa, F., Luis, J. & Maciá, R., 2016. New study on the 1941 Gloria Fault earthquake and tsunami, Nat. Hazards and Earth Syst. Sci., 16(8), 1967-1977. https://doi.org/10.5194/nhess-16-1967-2016, 2016.
Buforn, E., Udias, A. & Colombas, M., 1998. Seismicity, source mechanisms and tectonics of the Azores – Gibraltar plate-boundary, Tectonophysics, 152, 89-118.
Buforn, E., López-Sánchez, C., Lozano, L., Martínez-Solares, J. M., Cesca, S., Oliveira, C. S. & Udías, A., 2020. Re-evaluation of Seismic Intensities and Relocation of 1969 Saint Vincent Cape Seismic Sequence: A Comparison with the 1755 Lisbon Earthquake. Pure Appl. Geophys., 177, 1781–1800. https://doi.org/10.1007/s00024-019-02336-8.
Custodio, S., Cesca, S. & Heimman, S., 2012. Fast kinematic waveform inversion and robustness analysis: Application to the 2007 Mw 5.9 Horseshoe Abyssal Plain Earthquake Offshore Southwest Iberia. Bulletin of the Seismological Society of America, 102, 361–376. https://doi.org/10.1785/0120110125.
Fukao, Y., 1973. Thrust faulting at a lithospheric plate boundary the Portugal earthquake of 1969, Earth and Planetary Science Letters, 18(2), 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., Oceans, 102(C13), 27931-27949.
Gracia, E., Danobeitia, J., Verges, J. & Bartolome, R., 2003. Crustal architecture and tectonic evolution of the Gulf of Cadiz, SW Iberia, at the convergence of the Eurasian and African plates. Tectonics 22 (4), 1033.
Grandin, R., Borges, J. F., Bezzeghoud, M., Caldeira, B. & Carrilho, F., 2007. Simulations of strong ground motion in SW Iberia for the 1969 February 28 (M s= 8.0) and the 1755 November 1 (M∼ 8.5) earthquakes-II. Strong ground motion simulations, Geophysical Journal International, 171(2), 807-822.
Grimison, N.L. & Chen, W.P., 1988. Source mechanisms of four recent earthquakes along the Azores–Gibraltar plate boundary, Geophysical Journal International, 92(3), 391-401.
Guesmia, M., Heinrich, P., & Mariotti, C., 1996. Finite element modelling of the 1969 Portuguese tsunami, Physics and Chemistry of the Earth, 21(1-2), 1-6. https://doi.org/10.1016/S0079-1946(97)00001-3
Guesmia, M., Heinrich, P. H. & Mariotti, C., 1998. Numerical simulation of the 1969 Portuguese tsunami by a finite element method, Natural Hazards, 17(1), 31-46.
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(12), 1071-1074.
Heinrich, P., Baptista, M. A. & Miranda, P., 1994. Numerical simulation of the 1969 tsunami along the Portuguese coasts. Preliminary results, Sci. Tsunami Hazards, 12(1), 3-23.
Imran, J., Harff, P. & Parker, G., 2001. A numerical model of submarine debris flow with graphical user interface, Computers & Geosciences, 27(6), 717-729. https://doi.org/10.1016/S0098-3004(00)00124-2.
Le Petit Marocain 1969. Le Petit Marocain, 55^eme année nº16279, Samedi 1^er mars 1969 (News Paper in french) pages 3-4.
Levret, A., 1991. The effects of the November 1, 1755 “Lisbon” earthquake in Morocco. Tectonophysics, 193(1-3), 83-94.
López-Arroyo, A. & Udias, A., 1972. Aftershock sequence and focal parameters of the February 28, 1969 earthquake of the Azores-Gibraltar fracture zone, Bull. Seismol. Soc. Am., 62(3), 699-719.
Mansinha, L. & Smylie, D. E., 1971. The displacement fields of inclined faults, Bull. Seismol. Soc. Am., 61(5), 1433-1440.
Martínez‐Loriente, S., Gràcia, E., Bartolome, R., Perea, H., Klaeschen, D., Dañobeitia, J. J., Zitellini, N., Wynn, R.B. & Masson, D. G., 2018. Morphostructure, tectono‐sedimentary evolution and seismic potential of the Horseshoe Fault, SW Iberian Margin, Basin Research, 30, 382-400. https://doi.org/10.1111/bre.12225.
McKenzie D., 1972. Active Tectonics of the Mediterranean Region, Geophys. J. R. astr. Soc., 30, 109–185. https://doi.org/10.1111/j.1365-246X.1972.tb02351.x.
Mendes, A., 1969. Notícia acerca do sismo de 28 de Fevereiro de 1969 em Portugal Continental. Finisterra, 4(8) (in Portuguese).
Mezcua, J., 1982. Catalogo general de isosistas de la Península Ibérica. IGN, Publ 202, Madrid.
Miranda, J. M., Miranda, P. M. A., Baptista, M. A. & Victor, L. M., 1996. A comparison of the spectral characteristics of observed and simulated tsunamis, Physics and Chemistry of the Earth, 21(1-2), 71-74.
Miranda, J. M., Luis, J., Reis, C., Omira, R. & Baptista, M. A., 2014. Validation of NSWING, a multi-core finite difference code for tsunami propagation and run-up. In American Geophysical Union (AGU) fall meeting, San Francisco. Paper (No. S21A-4390).
Okada, Y., 1985. Surface deformation due to shear and tensile faults in a half space, Bull. Seismol. Soc. Am., 75, 1135–1154, 1985.
Omira, R., Baptista, M. A. & Matias, L., 2015. Probabilistic tsunami hazard in the Northeast Atlantic from near-and far-field tectonic sources, Pure Appl. Geophys., 172(3-4), 901-920. https://doi.org/10.1007/s00024-014-0949-x.
Omira, R., Baptista, M. A., Matias, L., Miranda, J. M., Catita, C., Carrilho, F. & Toto, E., 2009. Design of a Sea-level Tsunami Detection Network for the Gulf of Cadiz, Nat. Hazards Earth Syst. Sci., 9, 1327–1338. https://doi.org/10.5194/nhess-9-1327-2009.
Omira, R. & Ramalho, I., 2020. Evidence-Calibrated Numerical Model of December 22, 2018, Anak Krakatau Flank Collapse and Tsunami, Pure Appl. Geophys., 177, 3059–3071. https://doi.org/10.1007/s00024-020-02532-x.
Paula, A. & Oliveira, C.S., 1996. Evaluation of 1947–1993 macroseismic information in Portugal using the EMS-92 scale, Ann. Geof., 39(5), 989–1003
Pires, C. & Miranda, P., 2001. Tsunami waveform inversion by adjoint methods, J. Geophys. Res., Oceans, 106(C9), 19773-19796.
Piper, D.J.W., Cochonat, P. & Morrison, M.L., 1999. The sequence of events around the epicenter of the 1929 Grand Banks earthquake: initiation of debris flows and turbidity current inferred from sidescan sonar. Sedimentology, 46, 79 – 97.
Pro, C., Buforn, E., Bezzeghoud, M., & Udías, A., 2013. The earthquakes of 29 July 2003, 12 February 2007, and 17 December 2009 in the region of Cape Saint Vincent (SW Iberia) and their relation with the 1755 Lisbon earthquake. Tectonophysics,
583, 17–27. https://doi.org/10.1016/j.tecto.2012.10.010.
Pro, C., Buforn, E., Udías, A., Borges, J. & Oliveira, C. S., 2020. Study of the PGV, Strong Motion and Intensity Distribution of the February 1969 (Ms 8.0) Offshore Cape St. Vincent (Portugal) Earthquake Using Synthetic Ground Velocities, Pure Appl. Geophys., 177, 1809–1829. https://doi.org/10.1007/s00024-019-02401-2.
Quintino, J. 1969. O Sismo de 28 de Fevereiro de 1969. Algumas informações geofísicas deduzidas do seu estudo. Bol. Mus. Lab. Min. Geol. Fac. Ciênc. Lisboa, 11(2), 265-292.
Renou, C., Lesne, O., Mangin, A., Rouffi, F., Atillah, A., El Hadani, D. & Moudni, H., 2011. Tsunami hazard assessment in the coastal area of Rabat and Salé, Morocco, Nat. Hazards Earth Syst. Sci., 11(8), 2181. https://doi.org/10.5194/nhess-11-2181-2011.
Solares, J. M. & Lopez-Arroyo, A., 2004. The great historical 1755 earthquake. Effects and damage in Spain, Journal of Seismology, 8(2), 275-294.
Stich, D., Mancilla, F., Pondrelli, S. & Morales, J., 2007. Source analysis of the February 12th 2007, Mw 6.0. Horseshoe earthquake: Implications for the 1755 Lisbon earthquake, Geophys. Res. Lett., 34, L11208, doi: 10.1029/2007GL030012.
Stich, D., Martín, R. & Morales, J., 2010. Moment tensor inversion for Iberia–Maghreb earthquakes 2005–2008, Tectonophysics, 483(3-4), 390-398.
Synolakis, C.E., Bardet, J.P., Borrero, J.C., Davies, H.L., Okal, E.A., Silver, E.A., Sweet, S. & Tappin, D.R., 2002. The slump origin of the 1998 Papua New Guinea tsunami. Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 458, 763-789.
Wronna, M., Omira, R. & Baptista, M. A., 2015. Deterministic approach for multiple-source tsunami hazard assessment for Sines, Portugal, Nat. Hazards Earth Syst. Sci., 15, 2557–2568, https://doi.org/10.5194/nhess-15-2557-2015.
Wronna, M., Baptista, M. A. & Miranda, J. M., 2019. Reanalysis of the 1761 transatlantic tsunami, Nat. Hazards Earth Syst. Sci.,19(2), 337-352.
Zitellini, N., Chierici, F., Sartori, R. & Torelli, L., 1999. The tectonic source of the 1755 Lisbon Earthquake, Annali di Geofisica, 42(1), 49–55.
Zitellini, N., Mendes, L. A., Cordoba, D., Danobeitia, J., Nicolich, R., Pellis, G., Ribeiro, A., Sartori, R., Torelli, L., Bartolome, R., Bortoluzzi, G., Calafato, A., Carrilho, F., Casoni, L., Chierici, F., Corela, C., Corregiari, A., Della-Vedova, B., Garcia, E., Jornet, P., Landuzzi, M., Ligi, M., Magagnoli, A., Marozzi, G., Matias, L., Penitenti, D., Rodriguez, R., Rovere, M., Terrinha, P., Vigliotti, L. & Zahinos-Ruiz, A., 2001. Source of 1755 Lisbon earthquake, tsunami investigated, EOS, 82(26): 285–291.
Zitellini, N., Rovere, M., Terrinha, P., Chierici, F., Matias, L. & Bigsets Team, 2004. Neogene through Quaternary Tectonic reactivation of SW Iberian Passive Margin, Pure Appl. Geophys., 161, 565–585.
Zitellini, N., Gràcia, E., Matias, L., Terrinha, P., Abreu, M. A., DeAlteriis, G., Henriet, J.P., Dañobeitia, J.J., Masson, D.G., Mulder, T. & Ramella, R., 2009. The quest for the Africa–Eurasia plate boundary west of the Strait of Gibraltar, Earth and Planetary Science Letters, 280(1-4), 13-50.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
02 Nov, 2024
Reviews received at journal
16 Oct, 2024
Reviews received at journal
18 Sep, 2024
Reviews received at journal
09 Sep, 2024
Reviewers agreed at journal
06 Sep, 2024
Reviewers agreed at journal
02 Sep, 2024
Reviewers agreed at journal
30 Aug, 2024
Reviewers invited by journal
30 Aug, 2024
Editor assigned by journal
24 Aug, 2024
Submission checks completed at journal
23 Aug, 2024
First submitted to journal
22 Aug, 2024

You are reading this latest preprint version

The 28th February 1969 earthquake and tsunami in the Atlantic Iberian margin

Status:

Version 1

Abstract

Figures

1. Introduction

2. Tsunami data analysis

3. Tsunami Simulations

3.1 Source location using tsunami travel time data

3.2 Casablanca Landslide

3.3 Selecting the “best Okada-type” tsunami source

4. Discussion

5. Conclusions

Declarations

Author Contribution

Acknowledgement

References

Additional Declarations

Status:

Version 1