Storm surge and tsunami deposits along the Moroccan coasts: state of the art and future perspectives

The Moroccan coasts are occasionally inundated by storm surges and tsunamis. Local historical archives recorded some of these events, such as the storm surge of 1913 CE and the tsunami of 1755 CE. The latter remains the most destructive event the country has ever faced, with major human and economic losses recorded mainly between the two cities of Tangier and Safi. The privileged way to prevent any hazard related to these events is to study their past occurrences and impacts. However, historical evidences about these natural hazards are often very scarce to determine their return periods and evaluate their intensities. The scientific community increasingly uses sedimentary archives from coastal environments, since they offer a viable complement to historical archives. Several studies using this approach have been conducted on the Moroccan coast in recent years; however, until now, there has been no review dealing with these studies, which is the main objective of this paper. Twenty sites with traces of coastal inundation deposits have been inventoried during this work, and most of them are located along the Atlantic coast. The Mediterranean side remains poorly studied despite the presence of tsunami and storm surge risks. The review draws attention also to the absence of chronological data for most of the coastal inundation deposits recognized up to now along the Moroccan coasts, which is a major issue that prevents the determination of the return period of these events.


3
2008; Viles and Spencer 2014). Coastal inundation events, caused by storm surges and tsunamis, represent an additional pressure to these settings. These extreme events are today among the costliest natural hazards in terms of economic and human damage. In 2004, the Sumatra tsunami killed 165,708 people in Indonesia and caused economic losses estimated at US$ 4.46 billion (EM-DAT Public). In 2005, the United States suffered an estimated US$ 125 billion in economic losses from Hurricane Katrina, as well as the deaths of 1,833 people (EM-DAT Public). The projected fast increase in coastal populations might expose 75% of the world's population to ocean-related hazards by the year 2025 (World Ocean Network in Finkl and Makowski 2019). Projections estimate that global (annual) losses from coastal flooding will increase from US$ 6 billion in 2005 to US$ 52 billion in 2050 (Hallegatte et al. 2013).
The impact of storm surges and tsunamis on the Moroccan coasts may increase in the future. Firstly, because of the expected sea-level rise due to climate change (Li et al. 2018;Nagai et al. 2020). The latest IPCC report predicts a 2.5 m rise in mean sea level by 2100 for the RCP 8.5 scenario (Pachauri et al. 2014;Pörtner et al. 2019). With this, storm surge and tsunami events that are considered minor today will probably become more numerous and destructive in the future. Secondly, there is a strong demographic and economic growth observed along the Moroccan coastline. The urbanisation process has accelerated since Morocco's independence in 1956, with the development of large coastal urban areas, such as Saïdia, Tanger, Rabat, Casablanca, Agadir and Dakhla. The city of Casablanca, which in 1951 had about 683,000 inhabitants, was home to 3.36 million inhabitants in 2014 (Haut Commissariat au Plan 2014). The city of Rabat, which had no more than 50,000 inhabitants at the beginning of the XX century, saw its population triple in 1951, with 203,000 inhabitants, to reach 577,827 inhabitants in 2014 (Haut Commissariat au Plan 2014). The Moroccan coastline is also a structuring pole of the Moroccan economy. It brings together most of the tourist and industrial activities. The country's main port infrastructures are Tangier-Med, Casablanca, Mohammedia, Jorf-Lasfar and Agadir. There are also specific strategic infrastructures, such as the two large phosphate complexes of Jorf-Lasfar and Safi, as well as energy production plants, located in Jorf-Lasfar (2,016 Mw), Safi (1,386 MW) and Tahaddart (400 MW).
The current state of knowledge on Moroccan coastal inundation events is based mainly on rare textual and instrumental records, which are insufficient to decipher the behaviour of these low-frequency, but high-impact events (Raji et al. 2015). Since 2008, several coastal sites were investigated in search for sedimentary traces left by storm surges and tsunamis, in the hope of dating them and filling the knowledge gap regarding these natural hazards. Multiple sites with sedimentary evidences of these events were identified along the Moroccan coasts, in the form of boulders (Medina et al. 2011) and fine-grained sediments (typically sand-sized) (Mhammdi et al. 2015;Raji et al. 2015;El Talibi et al. 2016;Khalfaoui et al. 2020). Despite this effort, the number of studies remain low compared to the nearby Iberian coasts, which are confronted with the same natural hazards.
This work synthesizes and reviews the main methods and outcomes of all studies conducted on inundation deposits along the Moroccan coasts, in terms of geographic distribution, types of sedimentary records, and proxies used to identify them. The study also proposes some recommendations for future research projects on this topic along the Moroccan coasts. General informations about proxies applied to locate onshore coastal inundation deposits are beyond the scope of this work and can be found in Engel et al. (2020).

Sources of storm surges and tsunamis along the Moroccan coasts
Morocco is in the northwesternmost part of Africa, with around 3500 kms of coastlines (500 km on the Mediterranean Sea and 3000 km on the Atlantic Ocean). The waves reaching the Atlantic coast are created by westerly winds travelling across the North Atlantic Ocean (Charrouf 1991). They are not uniform, but they show a general direction between WNW and N, with heights (Hs) between 0.5 and 3 m and periods (Ts) around 9-13 s (Charrouf 1991). The tides on the Atlantic coast are mesotidal and semidiurnal.
On the Mediterranean coast, around 90% of the significant wave heights (Hs) are under 1.5 m and only 5% above 3 m (El Mrini et al. 2012). The dominant wave periods range between 5 and 6 s (El Mrini et al. 2012). The most frequent directions are E and NE (El Mrini et al. 2012). The tides on the Mediterranean coast are microtidal and semidiurnal.

Storm surges
During the winter season, North Atlantic extratropical cyclones generate strong westerly winds that can produce extreme swells to the Moroccan Atlantic coast. According to the standards of the Moroccan National Weather Service, a wave is considered dangerous when its height (Hs) exceeds 3 m on the Mediterranean sea and the Strait of Gibraltar, and 4 m on the Atlantic side (http:// vigil ance. maroc meteo. ma/). These waves can be dangerous, especially during spring tides. More rarely, tropical cyclonic systems can reach the Moroccan coast. Examples include Vince (8-11 October 2005), Delta (22-30 November 2005) and Leslie (11-14 October 2018) (Fig. 1). These cyclonic systems often reach the end of their life cycle on the Moroccan coasts (Mhammdi et al. 2020). Mhammdi et al. (2020) synthesised the storm surge events that affected the Moroccan Atlantic coast between 1905 and 2018. This record brings together several tidal databases, such as Simonet andTanguy (1956) andEl Messaoudi et al. (2016). The work of Simonet and Tanguy (1956) contains wave data collected by the Earth Physics Department of the Cherifian Scientific Institute from several ports during the 1928-1952 period. El Messaoudi et al. (2016) present a list of some major storm events recorded by the National Weather Service between 1966 and 2014. This catalogue contains other events compiled by Minoubi et al. (2013) for the Safi region during the winters of 1948-1949, 1965-1966, 1973-1974 and 1985-1986. To our knowledge, there is no storm catalogue for the Mediterranean coast of Morocco. Raji (2014) mentioned some historical events in 1889 and 1941 CE that affected the Nador lagoon and breached its coastal barrier. Niazi (2007) also cited some historical storms that impacted the coast of Tetouan in 1963Tetouan in , 1989Tetouan in and 1990. The event of 1963 completely destructed the main jetty of the port of M'diq and was the result of strong eastern winds, known locally as "Chergui". Other local technical reports mention storms affecting the coast of Fnideq andM'diq in 1956 and1957.

Mediterranean coast
Tsunami events recorded in the western part of the Mediterranean Sea are less destructive compared to the Atlantic ones. Two areas were defined by Papadopoulos and Fokaefs (2005) as part of an intermediate tsunamigenic zone: (i) the Alboran Sea environment, characterized by the dominance of strike-slip fault systems and, (ii) the Tell Atlas belt, Northern Algeria, characterized by a compressive type of deformation forming thrusts and folds ( Fig. 2a and c). In the Alboran Sea, Álvarez-Gómez et al. (2011) identified twelve probable tsunamigenic seismic sources. The most dangerous source would be the Alboran Ridge fault system, which is capable of generating tsunami wave heights up to 1.5 m at the receiving shores (for water depths < 80 m) in Spain and Morocco (Álvarez-Gómez et al. 2011). The seismic activity of the North Algerian Fold and Thrust Belt also represents a moderate tsunami source for the Iberian and North African coasts (Álvarez-Gómez et al. 2011). For the 1980 El Asnam earthquake, the maximal water heights were between 0.3-0.4 m, recorded at tide gauges between Cartagena and Alicante (South-Eastern Spain) (Roger et al. 2011). However, historical archives indicate the occurrence of multiple major tsunami events associated with Mw > 6 earthquakes, such as 1522, 1680, 1790 and 1804 CE (Kaabouben et al. 2009;Maramai et al. 2014).
Recent studies reported that tsunami events in the Alboran basin could also be triggered by sedimentary instability processes affecting the seafloor surface. The development of exploration techniques has revealed the presence of numerous potentially tsunamigenic landslides on the seafloor surface, at the southern Alboran Ridge (Al-Borani landslide) (Macías et al. 2015), along the northern Alboran continental slope (Baraza Slide) (Casas et al. 2011), along the northern Xauen-Tofino bank (Rodriguez et al. 2017) and northeastern seamounts (Alonso et al. 2014). Macías et al. (2015) simulated a possible tsunami generated by the Al-Borani submarine landslide and found out that it can produce tsunami waves with a maximal amplitude of 1 m at the Moroccan coastline (single landslide; water depths < 80 m). Following the same approach, Rodriguez et al. (2017) showed that a submarine landslide similar to the one located Along the northern Xauen-Tofino banks, can produce a tsunami with an elevation of the order of 1 m at the Moroccan coast (similar to the Al-Borani landslide). The most dangerous characteristic of this tsunami is the short propagation time between the source and the Moroccan coastline, on the order of 13 min, which represents a major threat to cities like Al-Hoceima. Reicherter and Becker-Heidmann (2009) indicate that the 1522 Almeria earthquake may have generated a submarine   in the Gulf of Cadiz. c main faults and historical earthquakes  in the Alboran sea. Historical earthquakes are from the NOAA database. Major quaternary faults of Iberia are from QAFI database (García-Mayordomo et al. 2012) landslide along its offshore segment, which probably produced a tsunami (with run-up between 1 and 3 m) well recorded in coastal sediments.

Atlantic coast
Most of the historical tsunamis that affected the Atlantic coast of Morocco were generated by submarine earthquakes, distributed along the western segment of the Eurasia-Nubia plate boundary, between the Azores archipelago and the Strait of Gibraltar ( Fig. 2a and  b). This boundary is controlled by different tectonic structures with different regimes from the West to the East: (i) an extensional regime near the Azores archipelago; (ii) a lateral strike-slip regime along the Gloria Fault; and (iii) a compressional regime in the South West Iberian Margin (Serpelloni et al. 2007). Seismic tsunamis in the area are produced mainly by the Gloria fault and the South West Iberian Margin (SWIM) (Luque et al. 2002;Baptista and Miranda 2009). The latter is characterized by the presence of several tectonic active structures, such as the Marquês de Pombal, the Gorringe bank, the Horseshoe, and the Cadiz faults (Johnston 1996;Zitellini et al. 1999;Gràcia et al. 2003;Matias et al. 2005;Gutscher 2006). The major historical events in the area include the 218-209 BC Lacus Lingustinus tsunami (Campos 1991), the 1 November 1755 Lisbon earthquake and tsunami (Mw ≥ 8.5) (Baptista and Miranda 2009), and the recent 28 February 1969 Horseshoe earthquake and tsunami (Mw 7.9-8.0). The 1755 CE Lisbon tsunami caused massive damage to the Iberian and Moroccan Atlantic coasts and remains the most destructive in the history of the European region (Baptista et al. 1998). The waves propagated to the eastern Lesser Antilles, Brazil and Newfoundland, Canada (Kozak et al. 2005;Roger et al. 2010;Biguenet et al. 2021).
The second source of tsunamis is the volcanic activity of some archipelagos located near Morocco (Canary, Azores and Cape Verde Islands). For example, studies reported that the explosion of the La Palma volcano in the Canary Islands could produce a slip of one of its flanks to the sea, which in turn will cause large tsunami waves (Abadie et al. 2012(Abadie et al. , 2020. One of the most remarkable tsunami events is that of the volcanic island of Fogo in Cape Verde. The eruption of this volcano, dated 70,000 years ago, caused its eastern flank to collapse into the sea, which is believed to have formed a gigantic sea wave of almost 300 m in height (Ramalho et al. 2015). Boulders weighing 700 tonnes were found on Santiago Island, 650 m from the coast, between 150 and 200 m above sea level (Ramalho et al. 2015).
Far-field tsunami sources can also threaten the Moroccan Atlantic coast. On the other side of the Atlantic, these events can be caused by the Northwestern Atlantic Ocean Submarine Landslides, the Puerto Rico Trench, and the northern Cuba fold-and-thrust belt (ten Brink et al. 2014;Costa et al. 2021). The 1929 CE Grand Banks earthquake in Canada triggered a tsunami that was instrumentally measured across the Atlantic Ocean to the coasts of Portugal and the Azores islands (Fine et al. 2005).

Geological evidence of past coastal inundation events
Despite the important work conducted so far on Moroccan coastal Quaternary deposits, few studies have been devoted to storm surge and tsunami deposits. Gigout (1957) reported the presence of isolated blocks along the coastal region of Rabat and connected their movement and overturning to storm surge events. The first work on marine highenergy deposits was the one conducted by Mhammdi et al. (2008), to re-examine the blocks mentioned by Gigout (1957) and to assess their potential relationship to the 1755 CE Lisbon tsunami. Table 1 synthesises the work done so far on coastal inundation deposits along the Atlantic and Mediterranean coasts of Morocco between 2008 and 2022. The locations of these studies are presented in Fig. 3.

Marchica Lagoon (Nador)
The only work on coastal inundation events along the Mediterranean coastline was conducted by Raji et al. (2015) in the Marchica lagoon (Nador) (1 in Fig. 3c). In this study, sediment cores were collected inside the lagoon following long-shore and cross-shore transects. Geochemical and sedimentological analysis of the MC45 core showed the presence of three sandy levels of coastal origin, trapped within the lagoon's fine stratigraphy. Chronological data ( 14 C, 210 Pb ex and 137 Cs) placed these three events over the last 500 years. Based on local and regional historical records, Raji et al. (2015) associated these deposits with the storm of 1889 CE and the two tsunamis of 1522 CE and 1790 CE.
In terms of best practices, the work of Raji et al. (2015) is an excellent example to follow in the study of coastal inundation deposits along the Moroccan coast. The multiproxy approach, based on grain-size analysis, XRF geochemistry and geophysics, increases the chances of identifying these types of deposits in low-energy coastal environments. The use of different radioisotopes ( 137 Cs, 210 Pb ex and 14 C) refines the age of the storm surge or tsunami deposits. On the other hand, the combination of results from sedimentary archives and seismic is suitable to evaluate the maximum extension of the most intense coastal inundation events over the whole lagoon (Raji et al. 2018).

Tanger bay
A small coastal marsh in Tanger bay was one of the areas surveyed by Genet (2011) (2 in Fig. 3c). It is a small coastal alluvial plain about 1 km long and 300 m wide, crossed by a small river called Oued Mlaleh. Five cores, with a length between 1 and 3 m, were collected within the plain along a cross-shore transect. The records revealed the presence of a few levels of coastal sand trapped inside the fine estuarine stratigraphy (Fig. 4a). Genet (2011) believes that these deposits were emplaced by extreme marine events.   1 3

Cap Spartel-Tahaddart coast
The coastal section between Cape Spartel and the Tahaddart thermal power plant was prospected by Genet (2011) andEl Talibi et al. (2020) (3-6 in Fig. 3c). This coastal section contains a few ponds and marshes, located behind a low dune ridge and seems to be favourable for the sedimentary record of coastal inundation events. Genet (2011) conducted manual coring and trenching in the Boukhalef lagoon and the periphery of the Houara marsh. Trenches cleared in a construction site near Houara marsh revealed sandy levels interspersed in the local stratigraphy and resting discontinuously on a paleosol. The presence of rip-up clast and some mud-drapes, associated with bioclasts of marine origin, were all indicators of a marine high-energy deposit (Genet 2011 Table 1 Fig . 4 Field photos of fine-grained deposits along the Moroccan coastlines. a Tsunami deposit recovered with an auger from a small marsh in the bay of Tanger (Genet 2011). b Tsunami-related sediments in the Carla-11 core at 6 km from the mouth of Loukkos river. 15 cm level showing more or less well-preserved sands and shells intercalated in wetland mud sediments (Mhammdi et al. 2015). c Washover of Oualidia lagoon (Google Earth). (d) Washover of Sidi Moussa lagoon (Google Earth) the mouth of the Boukadou River, located north of Tahaddart. The study focused on three trenches, 20 to 30 cm long, oriented along a cross-shore transect. A sandy level was identified in all three trenches and rests discontinuously on a paleosol. The marine origin of this level was justified by its richness in marine bioclasts (foraminifera and molluscs), and by its grain size, which is similar to present-day beach sediments. The investigated deposit becomes thinner from the coast, reflecting a progressive decrease in the energy of the waves (El Talibi et al. 2020). The authors agreed on the 1755 CE event as a potential source of these high-energy deposits. At the Tahaddart estuary, El Talibi et al. (2016) identified morphological and sedimentological evidences of an ancient coastal inundation event between coastal dunes of Pleistocene age (3-6 in Fig. 3c). The interdune spaces were filled mainly with sediments eroded from the Pleistocene dunes during the inundation. The authors attempted to reconstruct the wave direction associated with the event using grain size and anisotropy of magnetic susceptibility measurements. The results showed two main directions: N91°-171° and N280°-325°, which corresponded respectively to the uprush and backwash currents. Taking into consideration the different possible sources of tsunamis, the obtained wave directions were tentatively interpreted as a result of the Lisbon tsunami in 1755 CE. Wave heights were estimated to be between 6 and 8 m for this event.
Not far from the site investigated by El Talibi et al. (2016), several sediment cores were collected in 2017 from the salt-marsh area (behind the coastal dunes) by Khalfaoui et al. (2020Khalfaoui et al. ( , 2023 (3-6 in Fig. 3c). The objective was to reconstruct the paleoenvironmental evolution of the Tahaddart lower estuary during the mid-to late Holocene and to identify possible records of ancient coastal inundation events. These cores were examined using a multi-proxy approach, combining sedimentological (visual description; laser grain-size), geochemical (LOI, CaCO 3 and pXRF) and micropaleontological (benthic foraminifera) analyses, supported by 210 Pb ex , 137 Cs and 14 C time series data. This work revealed the first well-dated deposits of the 1755 CE tsunami on the Moroccan Atlantic coast (named SM1). Three older events, with a mean age of 625, 2700 and 3800 cal BP (named SM2, SM3 and SM4 respectively), were also identified in the framework of this work, and correspond probably to some tsunami events, recorded in the Iberian Atlantic coast. Mhammdi et al. (2015) analysed a sediment core collected from the Loukkos estuary (CARLA-11) using multiple proxies, including grain size, magnetic susceptibility and carbon content (7-9 in Fig. 3c; Fig. 4b). The results revealed the presence of a 15 cm sandy level rich in marine shells trapped in fine estuarine sediments. Given the distance of the core from the shoreline (6 km), the authors interpreted this sandy layer as a tsunami event. An estimated age, between 5000 and 3000 BP, was given to this deposit based on chronological data provided by Carmona and Ruiz (2009) from fluvial terraces surrounding the CARLA-11 archive. Medina et al. (2011) mentioned the presence of boulder deposits in the same area (7-9 in Fig. 3c; Fig. 5a and b). They are small compared to the ones found by the same authors on the Rabat-Bouznika coast, with a respective volume and maximum weight of ~ 10.7 m 3 and ~ 23.5 tons. These boulders are mostly leaning against the cliff in a nested position.

Larache
In the rocky shore platform of Laghdira (south of Larache city), Sedrati et al. (2022) used unmanned aerial vehicle (UAV)-based digital photogrammetry and SfM differential models to investigate the potential boulder transport and detachment triggered 1 3 by a strong storm in 2019, and to identify possible boulder transport modes (7-9 in Fig. 3c). The authors reported the detachment and mobilization of three boulders and the emplacement of a new one, as a result of the storm. The types of movements documented are sliding, saltation and boulders overturn with contrasting sizes and volumes.

Rabat-Bouznika coast
On the Rabat-Bouznika coastline, two types of high-energy deposits have been identified: boulders and fine dune deposits (10-17 in Fig. 3c). Mhammdi et al. (2008) reinvestigated the boulders recognised by Gigout (1959). A total of four sites were studied, namely Harhoura, Temara, Val d'Or and south of Skhirat. According to these authors, the boulders were detached from the active cliff during a high-energy marine event(s) and then transported over variable distances, from a few meters to 300 m. The shape of these boulders is generally flat, with a maximum length (axis A) of 9.8 m at Val d'Or. The weight of these structures is estimated between 4 and 100 tons. They can be found as single blocks, as a train of interlocking boulders or as a chaotic agglomerate. The study conducted by Medina et al. (2011) provides more quantitative results. It focused on the sites of Val d'Or, Harhoura and Cité Yacoub El Mansour. The direction of inclination and imbrication of the boulders was variable (N, NW and W). Their displacement distance reaches 150 m. According to hydrodynamic modelling, it is necessary to have waves with an amplitude of 5 to 11 m to move these boulders, which can be encountered in tsunami events (Medina et al. 2011). Brill et al. (2021) tried to date some of the Rabat-Bouznika boulders using optically stimulated luminescence rock surface exposure dating measurements (OSL-RSED) (10-17 in Fig. 3c; Fig. 5c). The results revealed large variability in the age of the boulders and were considered not conclusive. According to these authors, the variability was related to the low amounts of quartz and potassium feldspar in the source rock of the boulders, and to post-depositional erosion affecting some blocks. The major finding of this work was that the tsunami of 1755 CE was not the only event responsible for the deposition of these boulders, and that storm surge events have also played an important role in this process. According to these authors, storm surge events moved or reversed boulders with a weight greater than 24 tons. These results were in agreement with those of Medina et al. (2018) who noticed also the displacement of a large boulder (33 tons) in Dohemy beach (near Bouznika) during the winters of 2011-2012.
Another work carried out by Chahid et al. (2016) in the Rabat-Bouznika area revealed the presence of other high-energy deposits (10-17 in Fig. 3c). These authors studied two stratigraphic sections cleared in two Holocene dune formations located in Harhoura and Skhirat. The Harhoura section is positioned 20 m from the shoreline and is about 2 m thick. It consists of a succession of massive sandy deposits, composed of poorly classified sands and shells, similar to the sand deposited on the present beach. These levels, interpreted by the authors as high energy levels, are separated by horizons of young soils rich in gastropods and marine shells. This Harhoura section is also characterized by the presence of several clusters of boulders, which reinforces the eventdriven aspect of the deposits that form this Section. 14 C dating, performed on this section, places these high-energy deposits between 9900 and 2200 cal BP (Chahid et al. 2016). The second section is in Skhirat, north of the mouth of the Cherrat river. The authors describe a dune belt formed by a succession of poorly consolidated calcarenites, emplaced probably by high-energy marine events. The sedimentary structure and texture of these sandy levels (bedding, load figures, shell beds and pebbles) are related to wave breaking and washover deposition (Chahid et al. 2016). The construction of this second section was estimated to be between 8000 and 4000 cal BP (Chahid et al. 2016). The presence of cemented calcarenite boulders on their surfaces is another indicator of high-energy event driven deposition (Chahid et al. 2016). Benamri et al. (2023) also investigated the Harhoura coastal deposits to determine this time the source of the sediment forming the storm deposits described by Chahid et al. (2016). Using a multiproxy approach (sand exoscopy and petrography, microfaunal analysis (benthic foraminifera) and elemental geochemistry), the authors showed that the storm deposits were sourced from the shallow platform with a palaeobathymetry ranging between 0 and 300 m. Before the coastal inundation event, these platform sediments were in the form of ancient shelf deltaic fans, formed as a result of the erosion of the Caledonian and Hercynian Sehoul basement and its sedimentary cover. Leorri et al. (2010) and Mellas (2012) studied two washover deposits, as well as boulders located in the lagoon of Oualidia and Sidi Moussa (18 and 19 in Fig. 3b). The washover deposit in the Oualidia lagoon is located about 900 m north of the main channel (Fig. 4c). It is about 200 m wide by 120-130 m long. Two cores and a trench were performed by Mellas (2012) in the washover to characterize it in depth. The extracted sequences are constructed from coarse sands rich in shell debris, as well as a few marine benthic foraminiferas, like Cibicides lobatulus, Ammonia beccarii, and Elphidium crispum. This deposit lays discontinuously on autochthonous lagoonal fine sediments. Chronologically, no dating has been performed on this washover deposit. The Sidi Moussa washover is located about 1 km north of the main pass. It is 150 m long by 130 m wide (Mellas 2012) (Fig. 4d). This deposit was also studied at depth through coring. The sedimentological and micropaleontological results show the presence of a succession of shell-like coarse sandy levels of variable thickness (10 to 110 cm). These deposits are underlain by muddy lagoonal sediments. According to Mellas (2012), these deposits are probably the result of several marine high-energy waves (tsunamis or storm surges). Several isolated or clustered boulders (about 20) were also reported in this lagoon near the washover deposit. They were distributed parallel to the coastline and characterized by their elongated shapes with a major axis of 0.65-3.5 m (A-axis). No chronological data was provided for the washover and the boulders.

Safi coast
About 12 km south of Safi, Theilen-Willige et al. (2013) reported the presence of a few boulders, located on the intertidal platform, either arranged in the form of an arc curved towards the ocean or leaning directly on the cliffs present in this area (20 in Figs. 3b; Fig. 5d and e). Other blocks have been observed in overturned or upright positions with streak marks on the surface as evidence of their displacement by the waves. According to these authors, the weight of some blocks exceeds 100 tons.

Chronological correlation of identified events with regional coastal inundation deposits
To recognize possible regional events, the Moroccan coastal inundation deposits have been chronologically correlated with other deposits along the Iberian coasts. The only deposits dated along the Moroccan Atlantic coast are those of Khalfaoui et al. (2020Khalfaoui et al. ( , 2023 (Fig. 6). The 1755 CE tsunami is the only confirmed event through both historical and geological approaches (Kaabouben et al. 2009;Khalfaoui et al. 2020). It is a well-documented regional tsunami with multiple deposits already identified and well-dated along the Spanish and Portuguese coasts, notably in the Salgado Lagoon (Costa et al. 2012), Boca Do Rio (Hindson and Andrade 1999), Martinhal (Kortekaas and Dawson 2007) and in the Gulf of Cadiz (Cuven et al. 2013). The three additional deposits dated by Khalfaoui et al. (2023) around 3800, 2700 and 625 cal BP are considered by the authors as possible events. Further studies are needed to confirm their occurrence. However, they can also be correlated with some Spanish coastal inundation deposits. For example, Rodríguez-Ramírez et al. (2015) recognised three coastal inundation deposits from sedimentary records collected in the Guadalquivir estuary. The first one (event A) was dated around 4000 cal BP and was considered the most intense and destructive of the three events. According to the authors, it significantly transformed the geomorphology of the Guadalquivir estuary and impacted a human settlement in the region established during the Neolithic and Copper Age periods. Event B, dated at ~ 3550 cal BP, was considered of lesser magnitude compared to A. It was correlated with a 3600 cal BP earthquake, located through turbidite deposits identified by Vizcaino et al. (2006) on marine cores from the southwestern margin of Portugal. The third deposit (C) covered an extensive geographic area in the Doñana estuary and was dated at ~ 3150 cal BP. Taking into account the uncertainties of 14 C dating, it is possible to correlate between the 3800 and 2700 cal BP events in the Tahaddart estuary, respectively with the deposits A and C of Rodríguez-Ramírez et al. (2015). Other works on the Iberian Atlantic coasts agree on the existence of a submergence event between 3000 and 2500 cal BP, whose sedimentary traces are present in Guardiana, Rio Piedras, Tinto-Odiel and Guadalquivir (see the review of Costa et al. 2022). Coastal inundation deposit similar in term of age to the 625 cal BP deposit was recognized by Morales et al. (2008) in the Tinto-Odeil estuary, in the form of a marine shell-rich tsunamigenic level (HEL2). Another suspected  (Khalfaoui et al. 2023). A, B and C correspond to the events identified by Rodríguez-Ramírez et al. (2015). The grey and green vertical bars respectively correspond to possible and confirmed events tsunami deposit was identified by Becker-Heidmann et al. (2007) near Cape Camarinal (Bolonia), dated around 500 cal BP. The three events detected by Raji et al. (2015) in the Nador lagoon are the only known and dated coastal inundation events along the Moroccan Mediterranean coast (the two tsunamis of 1522 and 1790 CE, and the storm surge of 1889 CE). The 1522 CE event has been recorded in multiple sites along the Spanish coasts. Becker-Heidmann et al. (2007) managed to date this event in Algeciras, using radiocarbon techniques (Fig. 7). Sedimentary evidences of the same event were discovered by Reicherter and Becker-Heidmann (2009) in shallow drilling in the lagoon of the Cabo de Gata area. According to historical records, the 1522 Almeria earthquake affected large areas in the western Mediterranean and caused more than 1000 casualties (Reicherter and Becker-Heidmann 2009).

Gaps and needs
The work done so far on the Moroccan coasts to document coastal inundation deposits detected several coastal sites containing sedimentological evidence of these events. Despite this effort, the number of studies remains low compared to neighbouring countries (Spain and Portugal). Also, most of the detected sites are located along the Atlantic coast, which leaves the Mediterranean side poorly studied. On the one hand, it is important to revisit all the investigated sites, to conduct more complete studies, especially with sediment cores that can cover a large period of time. On the other hand, several coastal sites have not yet been investigated and deserve to be prospected. On the Mediterranean coast, the lagoon of Smir seems to be a good spot to search for coastal inundation deposits. The background mud-dominated sedimentation of the lagoon and the low elevation of the dune system (2-5 m) may favour the record of tsunami or storm surge events. On the Atlantic coasts, we can site the lagoons of Moulay Bousselham and Khenifiss, as well as the coastal lac of Sidi Boughaba.
The offshore part of the Moroccan coasts can be a valuable source of information on coastal inundation events. Shallow sea and deep-sea canyon environments are also capable of retaining marine high-energy deposits, especially tsunamis. A comprehensive compilation of studies concerning backwash deposits in marine shallow waters is given by Riou (2019). Beyond the storm-wave base, the high sedimentation rates can give rapid coverage for tsunami backwash deposits (Costa and Andrade 2020). Using sediment cores collected from the Pago Pago Bay (American Samoa), Riou et al. (2020) identified backwash deposits belonging to the 2009 south Pacific tsunami and the 1960 great Chilean earthquake tsunamis. The tsunami layers exhibited a sharp basal contact and were non-or poorly graded, compared to the background sedimentation. They also showed high values of Ti/Ca ratio, as a result of the land-sea transport of sediment (Riou et al. 2020). In the same sequences, the authors also recognized fine-grained terrestrial flashflood deposits belonging to tropical cyclones Ofa (1990 CE) or Val (1991 CE) and potential coarse-grained terrigenous run-off deposits emplaced during Cyclone Rene just four months after the 2009 SPT, or by Cyclone Wilma in early 2011. Quintela et al. (2016) detected deposits belonging to the 1755 Lisbon tsunami on offshore records collected south of Portugal. The tsunami deposits were in the form of coarse layers, rich in coastal foraminifera species, compared to the background sedimentation (Quintela et al. 2016). It is worth mentioning that such an approach is challenging in offshore environments. The seafloor investigations and access to sediments are logistically difficult, expensive, and time-consuming (Schwarzer 2020). To find the right spot for sampling, a good knowledge of seafloor architecture, geomorphology, and sediment distribution patterns is required.
The majority of coastal inundation deposits identified along the Moroccan coasts lack chronological data, making it difficult to place them in a chronological framework. Consequently, the chronological correlation between coastal inundation deposits located on the same site, or in different locations (e.g., regional scale), also becomes challenging. Without chronological data, there is no recurrence interval for these extreme events. For the few available 14 C dates, the absence of a local DeltaR value poses another problem that needs to be dealt with. This value is usually used to calibrate 14 C dates obtained on marine organisms. It represents the regional deviation from the pre-industrial marine reservoir age, estimated to be ~ 400 14 C yr in subtropical oceans (Heaton et al. 2020). This local deviation from the marine reservoir age is caused by several factors such as upwelling currents, which are very abundant along the Atlantic coast of Morocco. Usually, the DeltaR value can be estimated by comparing 14 C dates obtained on pairs of contemporary marine and continental organisms, collected for example at the same depth in the same sediment core (or trench), or found at the same coastal archaeological site. In the Tahaddart estuary (NW of Morocco), Khalfaoui et al. (2020) tried to estimate this value using a pair of contemporary wood and marine shell, collected from the same depth in one of the studied cores. The authors found a value of -75 ± 20 years, which is relatively close to the one obtained by Martins and Soares (2013) for the Andalusian coast in Spain (-108 ± 31 years). However, the value found by Khalfaoui et al. (2020) must be taken with caution, since the oceanographic conditions are different along the Moroccan Atlantic coast. On the Mediterranean side, Raji et al. (2015) found a DeltaR value of about − 218 years using sediment cores collected from the Nador lagoon. Additional studies are certainly needed to determine the geographical and temporal variation of DeltaR along the Moroccan coasts, which will help to precise the chronology of coastal inundation deposits.
There is a great need to work in partnership with local historians and archaeologists to better understand the impact of these extreme events on the local population, which has occupied the Moroccan coasts since the Neolithic phase. During the Roman period, trade activities, based on fishing, salt and garum production were probably affected by these disasters. For example, a sudden interruption of garum production in the third century CE has been reported in several coastal archaeological sites in Spain, Portugal and Morocco. The origin may be political (fall of the Roman Empire) or natural, related to one or more extreme marine events (Alonso et al. 2015;Trakadas 2015;González-Regalado et al. 2018).

Conclusion
coastal inundation events represent a major threat to the Moroccan coastlines, especially with the current socio-economic development and the future rise of the mean sea level caused by global warming. Research projects on coastal inundation deposits can help extend further back in time the Moroccan tsunami and storm surge records, to determine their return periods and maximum intensities. The present work synthesis the processes responsible for storm surges and tsunamis on the Moroccan coasts, as well as the sedimentary records produced by these events on these coasts. Storm surges on the Atlantic side are caused by extratropical cyclones, which circulate between the Azores High and the Icelandic Low mostly during the winter season. The Mediterranean coast is exposed to extratropical cyclones and sometimes to Medicans (Mediterranean cyclones). Tsunamis on the Atlantic coast are caused mainly by submarine earthquakes, generated by the Azores-Gibraltar seismic zone. On the Mediterranean side, tsunamis are less intense and are related mostly to the seismic activity of the strike-slip fault systems of the Alboran Sea. Twenty sites with traces of coastal inundation deposits have been inventoried during this work. They are for the most part located on the Atlantic coast, especially along the Cap Spartel-Tahaddart and Rabat-Bouznika coasts.
Amid the gaps raised during this review is the still low number of studies on this topic in Morocco, compared to neighbouring countries (Spain and Portugal), and the lack of chronological data for most of the coastal inundation deposits identified so far. Among the recommendations, it is important to revisit the already known sites and to conduct more complete studies using multiproxy approaches on sediment cores, which can cover a larger time scale. In addition to the Moroccan continental shelf, some coastal sites remain uninvestigated and deserve to be prospected, such as the Smir lagoon and the coastal lake of Sidi Boughaba.