Keywords

1 Introduction

The western part of the Eurasia–Nubia plate boundary extends from the Mid-Atlantic Ridge in the west to Tunisia in the east, and includes the Iberian Peninsula and the northern part of the Maghreb region. The interaction between Iberia and Nubia results in a complex region located in the western part of the Eurasian–African plate boundary that has various tectonic features (Buforn et al. 2004; Borges et al. 2007, 2008; Bezzeghoud et al. 2014). This region corresponds to the transition from an oceanic boundary (between the Azores islands and the Gorringe Bank) to a continental boundary where Iberia and Nubia collide.

According to several authors, (Buforn et al. 2004; Borges et al. 2007; Bezzeghoud et al. 2014), in the Ibero-Maghrebian region between the Gulf of Cadiz and Algeria, the Eurasian–Nubian plate boundary corresponds to a well-defined narrow band of seismicity, where large earthquakes occur in association with N–S to NNW–SSE horizontal compression due to the convergence of Eurasia and Nubia (Fig. 1). In this region, earthquakes are concentrated between 2°W and 4°E, and M > 6.0 earthquakes have occurred on 9 September 1954 (Ms 6.5, El Asnam, formerly Orléansville, Algeria); 10 October 1980 (Ms 7.3, El Asnam, formerly Chlef, Algeria); 27 October 1985 (Mw 5.9, Constantine, Algeria); 29 October 1989 (Mw 6.0 Tipasa–Chenoua, Algeria); 18 August 1994 (Mw 5.7, Mascara, Algeria); 22 December 1999 (Mw 5.7, Ain Temouchent, Algeria); 10 November 2000 (Ms 5.5, Beni Ouartilane, Algeria); 21 May 2003 (Mw 6.8, Zemmouri—Boumerdes, Algeria) (Ayadi and Bezzeghoud 2015); 24 February 2004 (Mw 6.4, Al Hoceima, Morocco) (Buforn et al. 2004); 20 March 2006, (Mw 5.2, Laalam, Algeria) (Bouhadad et al. 2010) and 2016 south Alboran earthquake (Mw 6.4) (Buforn et al. 2017).

Fig. 1
figure 1

Seismicity (M ≥ 4.5) of the Ibero-Maghrebian region for 1973–2015. Seismic data were taken from the NEIC database. White squares are the major historical earthquakes (A: 02/01/1365, B: 09/10/1790, C: –/03/1819, D: 02/03/1825, E: 02/01/1867, F: 15/01/1891). White circles are the main earthquakes discussed in this study (1: 1980 El Asnam EQ, 2: 1985 Constantine EQ, 3: 1989 Tipaza Chenoua EQ). The topography and bathymetry data have been extracted from the GEBCO (General Bathymetric Chart of the Oceans: http://www.gebco.net). Figures 1, 2, 3, 4, 5 and 6 are plotted with GMT software (Wessel and Smith 1991)

Full size image

Intermediate-depth seismicity (60–150 km) in the Alboran Sea region extends, in the N–S direction, in a very narrow vertical band of 50 km wide. This intermediate-depth seismicity is distributed in E–W direction and limited by a narrow band less than 20 km wide that broadens as we move to the Strait of Gibraltar. This seismicity could also be associated with the convergence process of the Eurasia–Africa plates (Buforn et al. 2004). This intermediate-depth seismic activity on the eastern side of the Strait of Gibraltar may be explained by the existence of a seismogenic block in the upper mantle that has approximate dimensions of 200 km long, 150 km deep and 50 km wide (Buforn et al. 2004). In this region, the material is relatively rigid, and the stresses are released by larger earthquakes. The presence of very deep earthquakes (650 km) under southern Spain is a further sign of the complexity of this area. In the Alboran Sea region, the material is more fragmented with a large number of small faults, and the stresses are released by frequent small-to-moderate earthquakes. Consequently, the plate boundary is more diffuse and corresponds to a wider area that includes the Betics, the Alboran Sea and the Rif Cordilleras (Buforn et al. 2004), where deformation is manifested by the continuous occurrence of small earthquakes and only occasionally by moderate to strong earthquakes with magnitudes greater than six as shown by the historical seismicity (Martínez Solares and Mezcua 2002; El Mrabet 2005; Bahrouni et al. 2013; Ayadi and Bezzeghoud 2015). This intermediate and deep seismicity recorded beneath the Gibraltar Arc, the westernmost Alboran Sea and southern Spain remains a subject of debate concerning whether it may be attributed to active subduction beneath the Gibraltar Arc or to another origin (Gutscher et al. 2012).

The stress regime obtained from the focal mechanisms of shallow events is compatible with the horizontal N–S to NW–SE convergence of Eurasia and Africa. However, in the Betics–Alboran area, there is also horizontal extension in an approximately E–W direction (De Vicente et al. 2008; Van der Woerd et al. 2014). According to several authors (e.g. Houseman 1996; Buforn et al. 2004) different tectonic models have been proposed for this region such as subduction process, extensional collapse of thickened continental lithosphere, continental lithospheric delamination, back-arc extension caused by subduction rollback, convective thinning, or subduction and breaking of a vertical slab of material.

Thus, for the Central Maghreb, the stress regime obtained from the inversion of focal mechanisms reflects a ‘pure’ convergence dynamic in the NW–SE direction and for both sides (the Rif in northern Morocco and northern Tunisia) reflects a strike–slip regime. Both sides are affected by the geodynamics of the adjacent areas that are influenced by the dynamics of the Alboran Sea in northwestern Morocco (Ousadou et al. 2014) and by lateral slab migration/segmentation and deep dynamics such as lithosphere–mantle interactions in the Tunisian Atlas (Soumaya et al. 2015).

The convergence rate determined by seismic strain analysis decreases from the Azores Plateau to the Ibero-Maghrebian region (Bezzeghoud et al. 2008, 2014; Buforn 2008). The corresponding values of slip velocity given by Bezzeghoud et al. (2014) range between 1.4 and 6.7 mm/year from the triple junction to the northwestern part of Algeria and the Tell Mountains, where the velocity is approximately 3.7 mm/year. These values may underestimate the geological deformation and do not include the energy released by aseismic processes (e.g. folding, thickening, plastic deformation and slow aseismic slip). Additionally, the slip velocity may be considered to be instantaneous and independent from that derived from the geodetic data because the earthquake cycle is much shorter than the history of available earthquakes (Borges et al. 2007).

2 Seismicity of Algeria in the Ibero-Maghrebian Context

2.1 Seismicity of the Ibero-Maghrebian Region

The seismicity of the Maghreb region is located along the African–Eurasian plate boundary starting from the Azores triple junction and continuing to Tunisia, crossing the Strait of Gibraltar, Morocco and northern Algeria. It is related to the closure of Quaternary basins under an oblique NW–SE convergent stress regime. Most of the activity is linked to fault-related folds striking NE–SW and is distributed along the main tectonic features. Because the plate boundary is not defined by a single discontinuity, as found in the western region of Gibraltar, but is more diffuse, the earthquakes are distributed over a wide zone. Most of the seismicity is shallow, located between depths of 5 and 20 km (Fig. 1). Nevertheless, intermediate and deep seismic events have been observed beneath the Gibraltar Arc, the westernmost Alboran Sea and southern Spain. The seismicity distribution in the Maghreb displays a clear arrangement for the activity since 1973, with an E–W trend along the African–Eurasian plate boundary (Fig. 1). Morocco and Algeria appear as more active areas than Tunisia, and the seismicity is mainly concentrated in the Rif, the High Atlas of Morocco, the Algerian Tell and Middle Atlas and the Tunisian Atlas, where the seismicity spreads along a band extending south to the Saharan Atlas (Fig. 1). The easternmost Maghrebian region, the Tunisian Atlas area, has not experienced strong earthquakes; most of the seismicity is of low magnitude located in the Northwestern Atlas, Southern Atlas and Eastern Pelagic platform (Bahrouni et al. 2013, 2014). However, Tunisia is still a key region because it is located at the junction between the eastern and western Mediterranean domains and straddles the orogenic province and the stable platform (Bouaziz et al. 2002).

In the Ibero-Maghrebian region, the NW–SE oblique convergence between Africa and Eurasia plates is consistent with the earthquake focal mechanisms that show mostly reverse faulting on NE–SW trending often combined with a strike–slip component. Most of the focal mechanisms in this area exhibit strike–slip faulting in northern Morocco, reverse faulting in northwestern Algeria and strike–slip mechanisms in northeastern Algeria and Tunisia (Fig. 2) (Ayadi et al. 2002; Bezzeghoud and Buforn 1999; Harbi et al. 1999; Buforn et al. 2004; Bahrouni et al. 2014). This agrees with earlier studies that give a NW–SE convergence direction for Africa and Eurasia with translation of the African plate to the east.

Fig. 2
figure 2

Focal mechanisms for M ≥ 5.0 shallow earthquakes (depth ≤ 40 km) in the Ibero-Maghrebian region for the period 1950–2015. Size is proportional to magnitude. Solution parameters for focal mechanisms are listed in Table 2

Full size image

2.2 Seismicity of Algeria

The seismicity of the Tell Atlas, which extends from the Algerian margin to the South Atlasic fault system, is related to the dynamics of Quaternary basins under an oblique NW–SE convergent stress regime, including the basins of Mleta and L’Habra in the west, Cheliff and Mitidja in the centre, and Soummam, Hodna and Guelma in the east. Most of the activity is linked to fault-related folds striking NE–SW and is distributed along the main tectonic features. This seismicity is characterized by moderate to low magnitude with strong events occurring generally once a decade. Over the last six decades, several strong events occurred and were associated with extensive and severe damage. The seismic catalogues show several moderate, strong and major events offshore or inland, such as those listed in the introduction in addition to numerous large historical events such as those that occurred in Algiers (1365 and 1716, Io = X), Oran (1790, Io = X), Mascara (1819, Io = X), Djidjelli (1856, Io = VIII) and M’sila (1885, Io = IX) (Benouar 1994; Mokrane et al. 1994; Harbi et al. 2010; 2015; Ayadi and Bezzeghoud 2015).

The focal mechanisms in Algeria are dominated by reverse faulting for earthquakes located in the western part of the country from 2°W to 5°E, which are associated with the strong and moderate earthquakes that occurred in the area (Fig. 2). Meanwhile, the focal mechanisms in the eastern part are dominated by strike–slip faulting for earthquakes located from 5°E to 8°E (Fig. 2). Most of these are associated with identified active Quaternary faults and folds (Bouhadad et al. 2003; Ayadi et al. 2002, 2008; Meghraoui 1988; Bounif et al. 1987; Maouche et al. 2011; Abbes et al. 2016).

3 Significant Earthquakes in Algeria Since 1980

Among all the earthquakes that have occurred in Algeria, we present only those with magnitudes greater than or equal to 5.7 and caused serious damage (Table 1) for the period 1980–2016. These earthquakes include the following: (i) the El Asnam earthquake of 10 October 1980, with Mw 7.1; (ii) the Constantine earthquake of 27 October 1985, with Mw 5.9; (iii) the Tipasa–Chenoua earthquake of 29 October 1989, with Mw 6.0; (iv) the Mascara earthquake of 22 August 1994, with Mw 5.7; (v) the Ain Temouchent earthquake of 22 December 1999, with Mw 5.7 and (vi) the Zemmouri—Boumerdes earthquake of 21 May 2003, with Mw 6.8.

Table 1 Damage caused by the most significant earthquakes occurred in Algeria since 1980 discussed in the text. The damage is given by Mr. Chergui Abdelkader and Mrs. Bradai Kheira from the ‘Direction Générale de la Protection Civile’ (Algiers, Algeria)
Full size table
Table 2 Source parameters of focal mechanisms plotted in Fig. 2
Full size table

3.1 The El Asnam Earthquake of 10 October 1980

The strongest event that occurred in northern Africa is that of El Asnam on 10 October 1980, at 12 h 25 min (TU) with Ms 7.3 (Table 1). This event has been studied by numerous researchers (Deschamps et al. 1982; Ouyed et al. 1981, 1983; Philip and Meghraoui 1983; Yielding et al. 1981, 1989; Ousadou et al. 2014). The event was relocated by the International Seismological Centre (ISC) at 36.16°N–1.40°E with a fixed depth of 10 km based on 514 station records. The Harvard CMT seismic moment is 5.07 × 1019 N m (Mw 7.1). The mainshock mechanism constructed using P-wave polarities at teleseismic distances shows reverse faulting along a plane striking N45°E and dipping N54° with an 83° slip angle (Ouyed et al. 1981; Deschamps et al. 1982), which is similar to the Harvard CMT solution (azimuth 50°N, dip 61°, slip 81°). The P and SH wave modelling show that the source consisted of two sub-events separated by 4 s (Yielding et al. 1981) (Fig. 3). This event is associated with 40-km-long surface ruptures that have been clearly mapped (King and Vita-Finzi 1981; Philip and Meghraoui 1983). Three segments are identified on the rupture zone as shown by the aftershock distribution and surface fault traces. The aftershock zone extends for a length of approximately 40 km in the NE–SW direction (Fig. 3), ranging between 5 and 14 km in depth, as given by the relocation using HypoDD by Ousadou et al. (2014). The distribution exhibits three segments, and the NE edge is leading in approximately the N–S direction.

Fig. 3
figure 3

The focal mechanism (Deschamps et al. 1982) and the corresponding aftershock sequence (relocated by Ousadou et al. 2014) following the mainshock of the El Asnam earthquake of 10 October 1980. Red star: mainshock after ISC relocation (2005). The regional stress tensor and principal stress (Ousadou et al. 2014) are shown in the lower right corner (Shmax in red and Shmin in blue colours). The scale bar units are in km

Full size image

The focal mechanisms of the 1980 El Asnam aftershocks are well constructed using the P arrivals (Ouyed et al. 1983; Yielding et al. 1989). Most of the focal solutions for the aftershocks show reverse faulting along the El Asnam rupture zone, with an important strike–slip component both northeast and southwest of the fault. The aftershock focal mechanisms are inverted to calculate the stress field for the Cheliff basin yielding a reverse tensor with maximum and minimum horizontal stress Shmax and Shmin striking in the N325° and N235° directions, respectively (Ousadou et al. 2014). This is in good agreement with the directions of the P and T axes of the mainshock (Fig. 3).

The Cheliff basin experienced the 9 September 1954 Orléansville earthquake, which had a magnitude of Ms 6.7, maximum intensity of Io X (Mercalli scale) and macroseismic depth of approximately 9 km (Rothé 1955; Rothé et al. 1977). According to Heezen and Ewing (1955), breaks in submarine cables were caused by the motion of a mass of sediments detached from the continental slope by the mainshock of 1954 and transformed into a strong turbidity current. All of the focal mechanisms established from the P arrivals show reverse solutions with a fault plane dipping to the NW (Shirokova 1967; McKenzie 1972; Espinosa and Lopez 1984), which is not much different from the El Asnam one. Referring to the work of Dewey (1990), the 1954 earthquake was relocated by the Joint Epicenter Determination method, taking the 1980 earthquake near the central segment of the El Asnam fault as the reference. As a result, Dewey proposes a model of the 1954 fault, associated with the 1980 thrust fault. This model agrees with the dislocation model proposed by Bezzeghoud et al. (1995) deduced from geodetic measurements.

3.2 The Constantine Earthquake of 27 October 1985

The eastern part of Algeria experienced a moderate-sized earthquake (Mw 5.9) on 27 October 1985 at 18 h 34 min (TU) (Table 1). It was the strongest earthquake recorded in the Tell Atlas in the 5 years following the 1980 El Asnam earthquake. The mainshock has been relocated and showed a remarkable jump of the epicentre from outside the cloud of aftershocks (Bounif et al. 1987) to inside it (Ousadou et al. 2013). The focal solution using P and SH waveform inversions shows a left-lateral strike–slip mechanism on a fault striking N217° and dipping 84° with a rake of 19°, a depth fixed at 9 km and a seismic moment of 5.2 × 1017 N m (Deschamps et al. 1991).

The aftershocks have been located (Bounif and Dorbath 1998) and, once more, relocated in a 3D model using TomoDD (Ousadou et al. 2013). The aftershock relocation by TomoDD is better defined and less scattered than that of the 1980 El Asnam earthquake (Fig. 4). It extends for a length of approximately 30 km, but only the 14-km-long central part shows an alignment in the N210°/215° direction, which corresponds to one of the two nodal planes given by Harvard CMT (2005) and Deschamps et al. (1991). This segment extending from 5 to 15 km down has been subsequently interpreted as the source of the mainshock rupture, and the 4.5-km-long surface breaks, organized in an en échelon system, are secondary ruptures due to ground shaking around the epicentral area (Ousadou et al. 2013).

Fig. 4
figure 4

The focal mechanism (Deschamps et al. 1991) and the corresponding aftershock sequence (relocated by Ousadou et al. 2013) following the mainshock of the Constantine earthquake of 27 October 1985. Red star: main shock after relocation (Ousadou et al. 2014). The regional stress tensor and principal stress (Ousadou et al. 2014) are shown in the lower right corner (Shmax in red and Shmin in green colours). The scale bar units are in km

Full size image

The aftershock focal mechanisms have been manually constructed, and the majority of them exhibit strike–slip faulting with some reverse or normal components, located between 10- and 15-km depths and spread all along the seismic zone (Ousadou et al. 2013). The inversion of aftershock focal mechanisms yields a strike–slip tectonic regime with quasi-horizontal Shmax and Shmin striking in the N347° and N257° directions, respectively. This is in accordance with the tectonic observations made in the Constantine basin, which is defined by the left-lateral strike–slip character of the NE–SW trending Ain Smara fault (Fig. 4). The Shmax and Shmin orientations correspond to the P and T axes of the mainshock mechanism (Ousadou et al. 2014).

3.3 The Tipasa–Chenoua Earthquake of 29 October 1989

The moderate earthquake that occurred on 29 October 1989 with magnitude Ms 5.9 struck the northwestern part of the Mitidja basin in the locality of Tipasa–Chenoua at 19 h 09 min (TU) (Table 1). The mainshock was located by NEIC and relocated by Bounif et al. (2003) (Fig. 5). The P and SH body wave modelling reveals almost pure reverse faulting along a plane striking 246°N and dipping 56° with an 86° slip angle and a seismic moment of 8.2 × 1017 N m, Mw 6.0 (Bounif et al. 2003) (Fig. 5).

Fig. 5
figure 5

The focal mechanism (Bounif et al. 2003) and the corresponding aftershock sequence (relocated by Ousadou et al. 2014) following the mainshock of the Tipasa–Chenoua earthquake of 29 October 1989. Red star: mainshock after relocation (Bounif et al. 2003). The regional stress tensor and principal stress (Ousadou et al. 2014) are shown in the lower right corner (Shmax in red and Shmin in blue colours)

Full size image

The epicentral distribution of the aftershocks depicts an elongated area with a general SW–NE direction approximately 15 km long and 7 km wide. Most of these aftershocks lie offshore, and there is no clear segmentation of the aftershock zone (Fig. 5).

The aftershock focal solutions reconstructed manually (Ousadou et al. 2014) show predominantly pure reverse faulting with a weak left-lateral strike–slip component on a plane dipping to the NW and striking NE–SW. This is in agreement with the geometry of the aftershock area and the focal mechanism of the mainshock (Fig. 5).

The set of aftershock focal mechanisms has been inverted to calculate a compressive stress tensor with well-defined horizontal Shmax and Shmin trending N326° and N239°, respectively (Ousadou et al. 2014). These two directions are close to those of the P and T axes of the mainshock focal solution and very close to that obtained in Cheliff Basin, located approximately 100 km southwestward.

3.4 The Mascara Earthquake of 18 August 1994

The Oranie region (northwestern Algeria) has experienced several significant earthquakes in the last centuries; the most important one is that of Oran city on February 9, 1790, Io = XI, which destroyed the town completely and caused the loss of many lives. Since 1790, no other event has been as disastrous, except that of 18 August 1994, with a magnitude of Mw 5.7 (Ayadi et al. 2002), which struck Mascara Province (Algeria) at 01 h 13 min GMT (Table 1). The Mascara earthquake is located in a zone with low seismic activity but with a maximum intensity Io of IX from past earthquakes (Ayadi and Bezzeghoud 2015). However, since the beginning of this century, the region has been dominated by seismic quiescence, and no event with magnitude larger than 5.5 has occurred in this area.

The focal solutions computed by Harvard, Bezzeghoud and Buforn (1999) using waveform modelling and Thio et al. (1999) using single station inversion (hybrid method), show a reverse mechanism on a single rupture striking in the NE–SW direction at a shallow depth (4.5 km) and a scalar seismic moment of 3.3 × 1017 N m. These focal mechanisms seem to be in good agreement with the tectonic observations (Thomas 1985) and with the macroseismic map (Ayadi et al. 2002).

3.5 The Ain Temouchent Earthquake of 22 December 1999

The Ain Temouchent region situated in the northwestern part of Algeria is characterized by low-to-moderate seismic activity in comparison with the Oran and Mascara regions. Nevertheless, the region was shaken on 22 December 1999, by an earthquake of magnitude Mw 5.7 (Table 1). The focal mechanism of the mainshock has been established using broadband data at regional and teleseismic distances (Yelles-Chaouche et al. 2004), and the scalar seismic moment estimated from waveform modelling is 4.7 × 1017 N m. Unfortunately, there is no aftershock study for this earthquake.

Indeed, throughout history, no important earthquakes have been mentioned in the seismic catalogues for this region (Fig. 1) (Roussel 1973; Benhallou 1985; Mokrane et al. 1994; Benouar 1994; Ayadi and Bezzeghoud 2015). These catalogues all indicate a low level of seismic activity with the occurrence of earthquakes having magnitudes less than 5.5. The Ain Temouchent earthquake is the largest seismic event ever recorded in the region with a maximum observed intensity of VII (MSK scale).

3.6 The Zemmouri—Boumerdes Earthquake of 21 May 2003

The easternmost edge of the Mitidja Basin experienced a strong earthquake of Mw 6.8 on 21 May 2003, at 18 h 44 min (TU), located to the east of the city of Algiers; this was the largest seismic event felt since that of 3 February 1716, which had Io = X (Rothé́ 1950; Harbi et al. 2015) (Table 1). The Zemmouri—Boumerdes earthquake occurred along the complex thrust and fold system of the Tell Atlas in the northeastern continuation of the Blida Mountain front and the related Mitidja Quaternary Basin.

Bounif et al. (2004) relocated the mainshock hypocenter, using three major aftershocks, with results that show a shift of this hypocenter from offshore to the coastline at 8–10 km depth. This coastal epicentre suggests a rupture along a previously unidentified offshore fault. According to Ayadi et al. (2003), this fault is an offshore continuation of the south Mitidja fault system. This is also supported by field observations of coastal uplift marked by a continuous white band of some sort of deposit attributed to a period of uplift, which is visible in exposed cliff faces (Meghraoui et al. 2004). Geodetic measurements of the 50-km-long coast show an average uplift of approximately 0.55 m along the shoreline with a maximum of 0.75 m east of Boumerdes and a minimum close to 0 near Cap Djenet (Meghraoui et al. 2004).

According to the study on teleseismic waveform inversion performed by Delouis et al. (2004), the Zemmouri—Boumerdes event was associated with an SE-dipping thrust fault mechanism with a seismic moment of 2.86 × 1019 N m (Mw 6.9), implying a 50-km-long fault rupture that should appear on the sea bottom at 6–12 km offshore from the coast (Fig. 6). Santos et al. (2015) investigate the rupture process to shed light on the location and geometry of the seismic fault and slip distribution using methodology based on teleseismic data, uplift measurements and synthetic aperture radar data. The methodology used by Santos et al. (2015) processes the available seismic and geodetic data to evaluate the two complementary planes of the focal mechanism. The location and geometry of the model proposed by Santos et al. (2015) are in agreement with the aftershock relocation proposed by Ayadi et al. (2008), where one aftershock cloud corresponded to a fault that dipped to the SE.

Fig. 6
figure 6

The focal mechanism (Braunmiller 2005) and the corresponding aftershock sequence (relocated by Ayadi et al. 2008) following the mainshock of the Zemmouri—Boumerdes earthquake of 21 May 2003. Red star: mainshock after relocation (Bounif et al. 2004). The regional stress tensor and principal stress (Ousadou et al. 2014) are shown in the lower right corner (Shmax in red and Shmin in blue colours). The scale bar units are in km

Full size image

The event was followed by a large sequence of aftershocks. More than 900 aftershocks that were relocated with TomoDD extend NE–SW along an approximately 60-km-long zone crossing the coastline and exhibit at least three seismic clouds and a well-defined SE-dipping main fault geometry (Ayadi et al. 2008). The distribution of seismic events presents a clear contrast between a dense SW zone and a NE zone with scattered aftershocks (Fig. 6).

The aftershock focal mechanisms were determined by inverted seismic broadband (Braunmiller and Bernardi 2005) and constructed using P-wave onsets (Ayadi et al. 2008; Ousadou et al. 2014). The majority of the solutions are reverse mechanisms, similar to the mainshock and two of the three largest aftershocks. However, at the eastern extremity of the seismic cloud, some events show reverse faulting on planes striking approximately N–S. Some strike–slip and normal faulting solutions are found, especially at the westernmost end of the aftershock cloud (Ousadou et al. 2014).

The aftershock focal mechanism database was inverted to calculate the stress tensor for the eastern part of the Mitidja basin, which is a compressive stress triaxial tensor with well-defined horizontal Shmax and Shmin in the N340° and N251° directions, respectively (Ousadou et al. 2014).

4 The Impact of Earthquakes in Algeria

Earthquakes are the expression of the dispersion of energy through seismic waves. The quantification of this seismic energy led seismologists to define different magnitudes. Several energy–magnitude empirical relations have been determined, the most suitable ones being those proposed by Gutenberg and Richter (1956), which express, in this study, the energy in joule from the magnitudes mb, and Ms given by the formulas log E = 2.4mb − 1.2 and log E = 1.5Ms + 4.8, respectively, as well as the relationship established by Kanamori (1977) giving the energy in joules from the moment magnitude Mw by the same formula that for Ms. However, to calculate the energy released from the historical earthquakes (1365–1900) we use Mokrane et al. (1994) empirical relation which is the magnitude versus intensity: Mm = 0.97Io − 2.24 for IX < Io ≤ XI.

Based on the catalogue of Ayadi and Bezzeghoud (2015) and using the formulas above, we estimated the seismic energy released by the 23 earthquakes with M > 5.5 that occurred in Algeria during the period between 1900 and 2015 (Fig. 1). The results show that 69% of the total energy (Fig. 7) comes from the El Asnam earthquake of 10 October 1980 (indicated by a red star in Fig. 4), if we consider that its magnitude was Ms 7.3. The El Asnam earthquake is the seismic event that dominates all the seismicity of the region for this period (1900–2015). A more recent earthquake that occurred in the Zemmouri—Boumerdes region on 21 May 2003 with a magnitude Mw 6.8 (Fig. 6) represents only 12% of the total energy (Fig. 7) released by the earthquakes occurring in the same period (1900–2015). Meanwhile, the other five recent earthquakes discussed in the previous sections released very low amounts of energy (1%, Fig. 7) when compared with the total seismic energy released by the El Asnam earthquake of 10 October 1980.

Fig. 7
figure 7

Diagram representing the percentage of seismic energy released by the earthquakes occurred in Algeria for the period 1900–2015 based on the catalogue of Ayadi and Bezzeghoud (2015). Note that the El Asnam earthquake of 10 October 1980 represents 69% of the total seismic energy for this period. See text for details

Full size image

However, for the total period (1365–2015) listed in the catalogue of Ayadi and Bezzeghoud (2015), the seismic energy released by the El Asnam earthquake of 10 October 1980 and the Zemmouri—Boumerdes earthquake of 21 May 2003, at 8 and 1%, respectively, is much less significant when compared to that released (88%) (Fig. 8) by the six major historical earthquakes (2 January 1365; 9 October 1790; March 1819; 2 March 1825; 2 January 1867 and 1 January 1891) (Fig. 1). The other seven recent earthquakes discussed in the previous sections released an insignificant amount of energy (<1%, Fig. 8) when compared with the total seismic energy released by all the seismic events recorded in the period 1365–2015.

Fig. 8
figure 8

Diagram representing the percentage of seismic energy released by the earthquakes occurring in Algeria for the period 1365–2015 based on the catalogue of Ayadi and Bezzeghoud (2015). Note that the six major historical earthquakes (02/01/1365; 09/10/1790; 03/1819; 02/03/1825; 02/01/1867 and 15/01/1891) (see Fig. 1) represent 84% of the total seismic energy for this period. See text for details

Full size image

5 Discussion and Conclusion

Generally, northern Africa has experienced moderate to strong earthquakes. However, between 1920 and the present, three earthquakes with magnitudes of approximately 8.0 occurred in the oceanic region (Mw 8.2, 25 November 1941; Ms 8.0, 28 February 1969; Mw 7.9, 26 May 1975), and four earthquakes with magnitudes of approximately 7.0 (Mw 7.1, 8 May 1939, Santa María Island; Mw 7.1, 1 January 1980, Azores; Ms 7.1, 20 May 1931, Azores–Gibraltar fracture zone; Mw 7.3, 10 October 1980, El Asnam, Algeria) occurred along the western part of the Eurasian–Nubian plate boundary (Bezzeghoud et al. 2014). In general, the major earthquakes (M ≥ 7.0) have occurred within the oceanic region, except for the El Asnam (Algeria) earthquake. The deadliest earthquakes that occurred in the study area since 1920, measured in numbers of victims, were those recorded in El Asnam, north Algeria (1954 and 1980), Azores, Portugal (1980 and 1998), Al Hoceima, north Morocco (1994 and 2004) and Zemmouri—Boumerdes, north Algeria (2003) (Benouar 1994; Bezzeghoud et al. 1996, 2014; Bezzeghoud and Buforn 1999; Ayadi et al. 2003).

The Ibero-Maghrebian region is vulnerable due to the shallow character of its seismicity generated by the convergence of the Eurasian and Nubian tectonic plates, to the specific mechanical properties of the crust and to local site effects (Bezzeghoud and Buforn 1999; Mourabit et al. 2014). To lead to a better assessment of the hazards and risks in and around urban areas, we need to know the detailed behaviour of the fault rupture processes that cause the earthquakes. Next, to produce reliable seismic hazard maps and mitigate the destructive impact of the earthquakes, the regional seismic hazards must be assessed using different approaches (e.g. deterministic and probabilistic) based on available historical and instrumental seismicity data (Ayadi and Bezzeghoud 2015; Harbi et al. 2015), earthquake sources and seismotectonic zoning studies and structural models associated with attenuation laws (Mourabit et al. 2014).

In summary, Algeria presents a seismotectonically complex area in which earthquakes responsible for high social and economic impacts have occurred and will continue to occur in the future. An Earthquake Early Warning System (EEWS) for this zone could be very useful in mitigating the damage caused by large earthquakes (Buforn et al. 2015). The complexity of the Ibero-Maghrebian region results in the need to address in depth the problem of the design of the early warning network to develop an appropriate EEWS. One of the issues that is particularly important is whether a single alert system could effectively cover the entire region or whether more than one system would be required instead, which takes into consideration the different seismic behaviours found in the areas under study. A preliminary feasibility study for an EEWS in the Gulf of Cadiz was performed as part of the ALERT-ES project coordinated by the Complutense University of Madrid with the participation of the Royal Institute and Naval Observatory of San Fernando, Cádiz, the Geological Institute of Catalonia in Barcelona and the University of Évora. The extension of this system to the entire Ibero-Maghrebian region is now in progress with the ALERTE-RIM project, which began in 2015 with the same participants.

Data and resources

Data are available from the literature listed in the references section and from (1) the GCMT, Global Centroid Moment Tensors: http://www.globalcmt.org; (2) Harvard CMT (2005) Centroid Moment Tensor Project: http://www.globalcmt.org/CMTsearch.html; (3) IAG, Instituto Andaluz de Geofisica: http://www.iagpds.ugr.es; (4) IGN (2004), Instituto Geográfico Nacional, Spanish seismic network: http://www.ign.es/ign/en/IGN/home.jsp; (5) ISC (2008), International Seismological Centre: http://www.isc.ac.uk, Internat. Seis. Cent., Thatcham, United Kingdom; (6) NEIC, National Earthquake Information Center: http://neic.usgs.gov; (7) RCMT, Regional Centroid Moment Tensors: http://www.bo.ingv.it/RCMT; and (8) ZUR_RMT, Swiss Seismological Service (SED): http://www.seismo.ethz.ch/en/. Topography and bathymetry data have been extracted from the GEBCO (General Bathymetric Chart of the Oceans: http://www.gebco.net).