Assessment of prospective geological hazards in Torrevieja-La Mata coast (western Mediterranean) based on Pleistocene and Holocene events

The coastal zone in which the lagoons of La Mata and Torrevieja (Eastern Spain) developed can be described as a compilation of geo-hazards typical of the Mediterranean realm. This study has focused mainly on those linked to recent tectonics. Extensive use of the amino acid racemization dating method allowed us to establish the evolution of all the geomorphological units differentiated in the area, the most striking manifestation being at the La Mata Lagoon Bar, where MIS 5 deposits settled on MIS 7 sediments along a marked erosive unconformity, thereby attesting coastal uplift between these two stages. In addition, recent uplift processes were reflected on stepped abrasion platforms and, in some cases, enormous boulders were transported over these platforms by extreme surge waves. Furthermore, we obtained feasible evidence that, during the end of MIS 5, an earthquake with an offshore epicenter linked to Torrevieja Fault, Bajo Segura Fault or the set of faults linked to the former, was responsible for tsunami surge deposits represented in accumulations of randomly arranged and well-preserved Glycymeris and Acanthocardia shells. Recent catastrophic effects linked to the earthquakes were also detected. In this regard, comparison of the paleontological and taphonomic analyses allowed us to discern between wave and tsunami surge deposits. Therefore, evidence of these hazards undoubtedly points to important future (and present) erosive and/or catastrophic processes, which are enhanced by the presence of tourist resorts and salt-mining industry. Thus, these sites are also threatened by future increases in sea level in the context of warmer episodes, attested by raised marine fossil deposits. At the north of Cervera Cape, beaches will be eroded, without any possibility of sediment input from the starved Segura River delta. At the south of this cape, waves (and tsunamis) will erode the soft rocks that built up the cliff, creating deep basal notches.


Introduction
The area of Torrevieja-La Mata is subjected to active tectonic processes, and the epicenters of many recent earthquakes have been in the inland and offshore surroundings of this zone. The largest tectonic event, named the "Torrevieja Earthquake," occurred on 21 March 1829 and registered an intensity of IX-X and estimated magnitude of between 6.3 and 6.9 (Delgado et al. 1993;Muñoz and Udías 1991). This earthquake caused 389 deaths, injured 377 people and destroyed thousands of houses (2965). Indeed, it was the most devastating earthquake in Spanish history (Larramendi 1829). Earthquakes in this area are related to a series of known active faults which have controlled ground uplift or subsidence processes, sometimes coupled with changes in global sea level (Alfaro et al. 1995;Montenat et al. 1990; Martínez Solares and Mezcua Rodríguez 2002;Giménez et al. 2009). As in other Mediterranean regions, the occurrence of extreme events such as tsunamis and severe storms cannot be discarded (Silva et al. 1993;Somoza 1993;Dezileau et al. 2011;Shah-Hosseini et al. 2013Deguara and Gauci 2017).
The area of Torrevieja-La Mata has been affected by the activity of recent tectonics, which produced earthquakes and with them erosive and depositional processes. Because the area is highly populated, as attested by the continuous development of tourist resorts, any tectonic event can have catastrophic effects. In this regard, it would be pertinent to have access to a realistic predictive model of future geological hazards. To this end, a complete geological and geomorphological framework of the area should be drawn up, including the characterization of the different deposits and related morphologies, discriminating between tectonic and/or climate forcings.
Raised marine deposits extend from the surroundings of the locality of Torrevieja to the Segura River mouth. The Cervera Cape acts as a geographical division: to the south, a lineal cliff extends southwards Torrevieja port interrupted by some coastal deposits, and a long sandy bar protects the lagoon of Torrevieja from wave influence. To the north, another sandy bar limits seawards the lagoon of La Mata. Previous studies focused on the raised marine deposits of La Mata Bar (Somoza et al. 1986;Zazo et al. 1990). However, the area of Torrevieja-La Mata has never been examined in detail, despite being highly exposed.
The aim of this study was 1) to establish the origin of some deposits in the area of Torrevieja-La Mata, discerning between two possible sources (surge storm vs. tsunami) based on the sedimentary architecture, fossil content and taphonomy, 2) to determine the age of these deposits and their relationship with other Pleistocene deposits in the area, through amino acid racemization dating, and 3) to obtain feasible information about the geological hazards of this highly populated flat coastal zone. To put these deposits into context, relating them to structural elements (faults), our study addressed the whole area. Also, we compared the observations made in the raised marine deposits with those of an active beach, El Alquián, to the south of the province of Almería, with mollusc shell accumulations produced by wave action (storms) (Torres et al. 2013).

Geographical and geological setting
The study area is in the province of Alicante in Southeastern Spain (Fig. 1). The area has two characteristic lagoons, Torrevieja (14 km 2 ) and La Mata (7 km 2 ), both unconnected to the Mediterranean Sea, although the latter, a protected natural park, is connected to the sea along an artificial ditch (El Acequión). La Mata Lagoon feeds the salt works of Torrevieja Lagoon through an excavated ditch (Canal Salinero).
Both lagoons are slightly below sea level and separated by the Torrevieja hill (46 m.a.s.l.). The Rojales hill (127 m a.s.l), near the village of Rojales, separates La Mata Lagoon from the Segura River.

Geology and tectonics
From a regional point of view, the area of Torrevieja-La Mata is located in the Bajo Segura zone which is the northern termination of the East Betic Shear Zone, a transpressive system that accommodates part of the stress related to the convergence between Africa and Eurasia-Iberia. The Bajo Segura basin underwent NNW-SSE horizontal compression (Montenat et al. 1990;Alfaro et al. 1995) during the Alpine Orogeny and a series of folds control the structure of this area: the Guardamar and Cervera Cape anticlines define the higher elevations while synclines host the La Mata and Torrevieja lagoons (Fig. 1).
The sedimentary record is of Miocene-Quaternary age (Montenat 1973(Montenat , 1977Montenat et al. 1990). According to Martínez et al. (1977) and Almela et al. (1978), the most extensive deposits in the area belong to a diachronic unit named the Sucina Formation (Pliocene-Pleistocene), which comprises red muds and sands of alluvial origin, usually capped by a caliche (Montenat 1973). Towards its north boundary, near Rojales, Pliocene sands and muds of marine origin dominate.
The reverse Bajo Segura Fault (BSF), which is one of the main active faults of the Eastern Betic Shear Zone , is located at the northern boundary of the study area (Fig. 1). Of note, Guardamar and Cervera Cape anticlines are part of the Bajo Segura Fault. Active strike-slip faults of regional development have also been defined (Giménez et al. 2009;García-Mayordomo et al. 2012), namely Torrevieja (TF) and San Miguel (SMF). According to some authors (Alfaro et al. 2002Martínez-Díaz 2006 Perea et al. 2012) and the Quaternary Faults Database of Iberia (QAFI, García Mayordomo et al. 2012), these faults are still active.
The activity of Torrevieja and San Miguel Faults produced a raised block from Torrevieja to the south boundary of Cervera Cape and lowered one block on its northern edge, thus conditioning the present-day landscape (Silva et al. 1993). The activity of these faults (BSF, TF, SMF) is reflected by the remarkable seismicity occurred in the zone during historical times (Martínez Solares and Mézcua 2002). According to Somoza (1993), seismic activity in the area has caused vertical movements between 0.2 and −0.2 mm/year. In 1829, an earthquake almost destroyed the village of Torrevieja, which was one of the most important earthquakes in Spain (IGN 2021) with an intensity of IX-X and estimated magnitude of 6.3-6.9 (Delgado et al. 1993;Muñoz and Udías 1991). This was followed by a rapid drop in sea level and surges that destroyed fishing boats in the port. The earthquake also affected the local landscape, causing the Cervera Cape cliff to rise 20 cm and the appearance of a fracture (Silva et al. 1993). The location of the epicenter of this earthquake has been a matter of discussion. Some authors have proposed that the activity of Bajo Segura Fault was responsible (Somoza 1993;Sanz de Galdeano et al. 1995;Alfaro et al. 2002;García-Mayordomo and Martínez Díaz 2006;Giménez et al. 2009;Silva et al. 2017), while others (López Casado et al. 1992;Delgado et al. 1993) postulated that Torrevieja Fault was the most probable source. Based on offshore geophysical research, Perea et al. (2012) did not discard an offshore epicenter, although the lack of tsunami-related deposits makes this scenario unlikely. In this regard, earthquakes with an estimated magnitude of 6.6 produce vertical displacements of some decimeters if the focus is shallow and would produce short wave tsunamis.

Material and methods
We build a Digital Terrain Model (DTM) at different scales to gain an accurate view of the geomorphology of the area, with special attention to the area of Cervera Cape Cliff. For this purpose, we employed DTMs with a grid resolution of 5 m (DTM05) and 2 m (DTM02) available on the web page of the Instituto Geográfico Nacional. Contour lines were obtained from the DTMs with the tool "contour" of the Software ArcGIS. Six distinct units can be distinguished from the geomorphological and stratigraphical perspective: Torrevieja Glycymeris Bed (TOG), Torrevieja Cliff (TOC), Cervera Cape Cliff (CCC), Cervera

Ancient deposits
In CCG and TOG, we performed a detailed survey of the paleontological content, preservation and arrangement. In this regard, we examined the bed surface, as well as all the exposures of its inner part along the abundant incisions (gullies) linked to current wave action, which allowed observation of the total thickness of these deposits. We recovered mollusc shells and examined their preservation (good preservation: unworn and complete; bad preservation: unworn broken or worn broken). We also considered shell arrangements in order to identify the origin of these deposits.
Given that LMB deposits were studied in detail by Somoza et al. (1986) and Zazo et al. (1990), and in TOC and CCC erosive morphologies occurred, these localities were not sampled to study their palaeontological content.

Recent deposits of El Alquián
To characterize heavy-storm-linked mollusc taphonomy, and for comparison purposes with the ancient deposits of CCG and TOG, we collected modern mollusc shells from 30 m 2 of El Alquián beach, in Almería (SE Spain), just after a period of heavy winter surge (Fig. 4a, b). All shell remains up to a depth of 5 cm were recovered and identified at species level. This relatively remote area was chosen due to the low impact of tourism. The preservation of each Glycymeris shell was examined. For species identification, we used the atlases of D'Angelo and Gargiullo (1987), Gofas et al. (1991), Salas et al. (2011), Giannuzzi Savelli et al. (1997 and Nolf and Swinnen (2013).

Sampling
To obtain a chronological framework of the geological processes that determined the geomorphological evolution of the area, 15 Glycymeris shells from CCG were sampled for amino acid racemization dating. To avoid recent contamination (recent algae and/or decay products of other organisms), shells were picked as far as possible from the sea. Two Tethystrombus latus specimens were also recovered for analysis.
We recovered and analyzed a similar number of Glycymeris shells at LMB sites: 9 samples from LMB-1 and 5 from LMB-2. We also selected 20 Glycymeris shells from TOG and recovered samples (5 Glycymeris shells) from the megadune exposed at CCC, which comprises bioclastic sands and some beds are visibly made of thin wind-glided pelecypod shell fragments. We also took 2 samples of Vermetidae shells from an isolated large angular boulder lying on the intermediate platform of CCC.
Deposits of LML were sampled at the mouth of El Acequión ditch. For this purpose, we used a gas-powered Wacker BH-23 jackhammer to drill two 2-m long boreholes with continuous core recovery, named LML-A and LML-B. We refer to the horizons of the cores by their depth (in cm) from top to bottom (e.g., level LML-A-80 is at 80 cm).
Core samples were examined for a sedimentological study and ostracode and gastropod shell recovery. Samples were washed and sieved at 63 µm, and carapaces and shells were recovered under a binocular microscope.

Amino acid racemization
A hollow diamond drill was used to remove a discoid sample (8 mm in diameter) from an area close to the beak of the Glycymeris shells recovered from TOG, CCC, CCG, LMB-1 and LMB-2. A 25-mm diameter discoid sample was obtained from the ventral side of the last whorl of T. latus shells from CCG using a hollow diamond drill, and a small sample near the aperture of gastropod shells belonging to the Helicidae family was recovered from LML. In all cases, peripheral parts (approximately 20-30%) were removed after chemical etching with 2 N HCl. Afterwards, 10-20 mg of samples were selected for amino acid racemization analysis.
Ostracode valves from LML cores were carefully cleaned by sonication in distilled deionized water and rinsed to remove sediment. Some valves were cleaned with a small brush under a binocular microscope to eliminate fine debris. To remove secondary organic molecules adsorbed to the valves, they were submerged in 3% hydrogen peroxide (H 2 O 2 ) for 2 h following Kaufman (2000) and Hearty et al. (2004) protocols.
Amino acid concentrations and D/L (Dextro-isomer/Levo-isomer) values were quantified using a high-performance liquid chromatograph (HPLC) in the Biomolecular Stratigraphy Laboratory (Madrid School of Mines and Energy), following the sample preparation protocol described by Kaufman and Manley (1998) and Kaufman (2000). This procedure involves hydrolysis, which was performed under an N 2 atmosphere in 7 mL of 6 M HCl for 20 h at 100 °C. The hydrolyzates were evaporated to dryness in vacuo and then rehydrated in 7 mL 0.01 M HCl with 1.5 mM sodium azide and 0.03 mM Lhomo-arginine (internal standard). Samples were injected into an Agilent-1100 HPLC equipped with a fluorescence detector. Excitation and emission wavelengths were programmed at 335 nm and 445, respectively. A Hypersil BDS C18 reverse-phase column (5 mm; 250 mm; 4 µm i.d.) was used for the separation. Derivatization occurred before injection by mixing the sample (2 mL) with the pre-column derivatization reagent (2.2 mL), which comprised 260 mM isobutyryl-L-cysteine (chiral thiol) and 170 mM o-phthaldialdehyde, dissolved in 1.0 M potassium borate buffer-solution at pH 10.4. Eluent A consisted of 23 mM sodium acetate with 1.5 mM sodium azide and 1.3 mM EDTA, adjusted to pH 6.00 with 10 M sodium hydroxide and 10% acetic acid. Eluent B was HPLC-grade methanol, and eluent C consisted of HPLC-grade acetonitrile. A linear gradient was performed at 1.0 mL/min and 25 °C, from 95% eluent A and 5% eluent B upon injection to 76.6% eluent A, 23% eluent B, and 0.4% eluent C at 31 min. This approach allowed the separation of aspartic acid (Asx), glutamic acid (Glx), serine (Ser), glycine (Gly), alanine (Ala), valine (Val), phenylalanine (Phe), isoleucine (Ile), leucine (Leu), threonine (Thr), arginine (Arg) and tyrosine (Tyr). Age was determined using the age calculation algorithms for ostracode shells described by Ortiz et al. (2004a,b). We focused on three amino acids (Ile, Asx and Glx) in Glycymeris shells, as Asx and Glx are the most abundant in this genus (cf. Demarchi et al. 2015), while Ile is the amino acid most commonly used for dating purposes in this genus (Belluomini et al. 1986(Belluomini et al. , 1993Hearty 1986Hearty , 1987Hearty et al. 1986;Torres et al. 2000Torres et al. , 2010Torres et al. , 2013Ortiz et al. 2004b;De Santis et al., 2014, 2020. For Ile, the values obtained by HPLC and IE-LC required conversion in order to be compared with D-aIle/L-Ile values measured by gas chromatography (GC) (cf. Ortiz et al. 2004b). This conversion was supported by the analysis of samples from an inter-laboratory comparison (ILC) exercise (cf. Wehmiller 1984;Torres et al. 1997;Wehmiller et al. 2010). The Ile epimerization (D-aIle/L-Ile) values, ranging between 0.45 and 1.1 by IE-LC and HPLC, were approximately 0.10 lower than those obtained by GC, i.e., D-aIle/L-Ile GC = D-aIle/L-Ile HPLC-IE-LC + 0.10. For lower D-aIle/L-Ile values, the difference was between 0 and 0.05.
Asx and Glx D/L values calculated by HPLC can be compared directly with those obtained by GC because of the similarities found between the racemization ratios of the samples from Wehmiller´s (1984) ILC exercise and several samples analyzed by GC and HPLC in our laboratory (Ortiz et al. 2004b(Ortiz et al. , 2009. Asx and Glx D/L values were selected for the age calculation of ostracodes because they account for over ca 50% of the amino acid content in most of the valves (Kaufman 2000;Bright and Kaufman 2011). The numerical age of each bed was determined by introducing the Asx and Glx D/L values obtained in ostracode valves collected at each level into the age calculation algorithms established by Ortiz et al. (2004a) and Ortiz et al. (2015).
In gastropod samples (Helicidae), the D/L values of Asx and Glx were introduced into the age calculation algorithms of Torres et al. (1997) for central and southern Iberian Peninsula.
The age of a single bed is the average of the numerical dates obtained for each amino acid D/L value measured in gastropods or ostracodes from that level, and the age uncertainty is the standard deviation of the numerical ages calculated at each level.

Geomorphological/stratigraphical units
The geological characteristics of the 6 geomorphological/stratigraphical units are described below.

Torrevieja glycymeris bed (TOG)
It is a tabular body gently dipping seawards, located 20-50 cm above sea level although eastwards is underwater, and it is developed above red alluvial materials (Fig. 3a), formerly named the Sucina Formation (Montenat 1973). It consists of a relic of an originally larger reddish bed (> 35 cm thick) containing well-preserved Glycymeris and Acanthocardia shells randomly oriented in a sandy matrix that shows carbonate cementation. The bed is under wave erosion and many decimeter wide crevasses are excavated on it leaving visible the substratum and its internal structure.

Torrevieja cliff (TOC)
This unit can be followed tens of kilometers south of Cervera Cape while at the north do not appears as a wide sandy beach extends until the Segura River Mouth. This vertical cliff was carved by wave action on the reddish alluvial deposits of the Pliocene-Pleistocene age belonging to the Sucina Formation (Figs. 3a, 4c). A thick massive bed (caliche) at its top protected the land from a fast retreat. At the cliff foot, there are short-lived notches due to wave erosion, which also frequently causes rock falls and large boulders accumulate at foot of the cliff. An active abrasion platform is visible and the narrow beach mainly consists of gravels.

Cervera cape cliff (CCC)
Cervera Cape Cliff constitutes a singular geoform in the Iberian Mediterranean realm mostly because it can be described as a giant indurated eolian complex attached to an older anticline. These sediments, according to Almela et al. (1978), are of Pleistocene age, although Silva et al. (2019) suggested a Pliocene age. The sand cosets present trough crossbedding are 2-3 m high and dip backshore, marking the dune movements. The sand-sized particles are carbonate in nature, quartz is not abundant and, in some cases, cross-beds comprise almost exclusively fragments of shell bioclasts (Fig. 4d). The eolian complex is capped by a thick caliche.
These eolian deposits form a cliff produced by wave erosion, which provides an excellent perspective of the inner organization of the deposits. Indeed, three stepped wave  (Figs. 3b,4e,4f,5) and associated minor cliffs, the recent one underwater. The flattened abrasion platforms do not show sediment accumulation, probably due to the poor cementation of the eolian sandstones, which easily disaggregate. Diverse fractures can be observed (Fig. 5). Indeed, one of them might correspond to the "grieta" (crack) that opened after the 1829 earthquake described by Larramendi (1829).
At the foot of the cliff, there is a chaotic accumulation of fallen sandstone boulders (Fig. 4f). Of note, an angular boulder (> 2m 3 ) on the intermediate platform shows abundant holes drilled by Lithophaga lithophaga, which, in some cases, remain in life position, as well as some tens of Vermetidae shells that attest that it was underwater before being lifted to the platform (Fig. 4g, h).

Cervera cape glycymeris bed (CCG)
This area (Fig. 3b) attracted the attention of researchers (Montenat 1977;Somoza et al. 1986;Zazo et al. 1990) because of the presence of Tyrrhenian fauna representatives. The bed is under wave erosion and many decimeter wide crevasses are excavated on it leaving visible its internal structure and the substratum. Northwards this bed disappeared due to recent wave erosion and southwards Cervera Cape, an active creek that presumably accommodates to La Mata Fault trace produced the erosion of this bed.
CCG appears as an isolated cemented blanket made of a myriad of disarticulated Glycymeris bimaculata shells at 20-50 m above sea level (a.s.l.) (Fig. 6a). The relic outcrop is 350 m long (SE-NW) and 50 m wide (E-W). Northwards, it has been wiped out by recent erosion. The shell bed lies on scoured reddish sediments from the Sucina Formation (Fig. 6b). It is 25-35 cm thick and dips gently seawards. Wave erosion produced gullies, which allow observation of the internal architecture. At its rear, there are loosely cemented dune sands, intensely bioturbated by plant roots.
According to the field observations, CCG is a biorudite with a sandy-silty matrix and carbonate cement. In the inner part of the bed, Glycymeris shells appear randomly oriented, and vertical arrangements are also common, in some cases, adopting a steep stacking pattern, like plates in a dishwasher (Fig. 6b). Shells at the top of CCG show a horizontal arrangement, either concave-up or concave-down. Of note, there are a number of boulders made of carbonate-cemented bioclastic sandstones (coquina of small pelecypods) pulled from an older Pleistocene deposit (Fig. 6c). At the top of CCG, we detected five shells belonging to large gastropods (3 T. latus (Strombus bubonius), 2 Stramonita (Thais) haemastoma). These shells are poorly preserved, being deeply eroded and broken in such a way that only fragments of the last whorl or the columella are observable (Fig. 6d). These shells contrast with the pristine appearance of Glycymeris and Acanthocardia shells.

La Mata Lagoon bar (LMB)
This bar, which consists of fossiliferous sandy beds, extends northwards from CCG and closes off La Mata Lagoon (Fig. 3c). Somoza et al. (1986) and Zazo et al. (1990) gave a detailed description of the sedimentary architecture of LMB, interpreting that these deposits corresponded to a sandy bar, but the paleontological content was only briefly commented on. It was assumed to be of Tyrrhenian age, as reflected by the presence of T. latus. Of note, we observed abundant Glycymeris spp. shells scattered in a bioclastic pale-yellow sand matrix and horizontally arranged.
Many years ago (1907), a ditch (El Acequión) was excavated to feed La Mata Lagoon. Today the ditch is covered by concrete as well as the ancient fossil dune cap to allow the development of a tourist resort. Near the entrance to the ditch, there was an ancient Roman Port constructed on this bar (LMB-1) (Fig. 6e, f). At the south of this port, there were other deposits with abundant shell remains (LMB-2) (Fig. 6g). Indeed, Somoza et al. (1986) identified two bar systems (Fig. 7) caused by the development of a marked erosive surface separating a lower unit of unknown age (which corresponded to LMB-1) from the overlying one with the presence of T. latus (LMB-2).
Towards the lagoon, a small cliff (now covered by concrete) was carved on the materials of the lower unit (LMB-2) and fallen boulders were incorporated in the overlying  Fig. 2. a LML deposits, and b LMB (modified from Somoza et al. 1986). 1: alluvial deposits; 2: lagoon sediments of MIS 7; 3: bioclastic calcitic sandstone with cosets dipping offshore (LMB-2) of MIS 7; 4: bioclastic calcitic sandstone coset dipping foreshore (LMB-1) of MIS 5; 5: Holocene eolian sands; 6: small erosive paleocliff; 7: Holocene backshore/lagoonal sands dipping landwards with abundant and well-preserved mollusc remains formation (LMB-1) sediments. The morphology of the bar is very difficult to determine because in a large part is building covered while northwards vegetated and an active-dune field blurred its outline.

La Mata Lagoon (LML)
A gentle slope connects LMB and the La Mata lagoon body water. Two cores were drilled ( Fig. 3c; 7), showing in all cases calcitic-bioclastic sands and muds with parallel lamination typical of back-barrier lagoonal deposits. Mollusc and ostracode remains are usually scarce, although at the top of the record, there are plenty of very well-preserved gastropod and pelecypoda shells frequently articulated. Near the lagoon water mass, sands are capped by organic-rich muds in which only terrestrial gastropods appeared. At the point where the ditch (El Acequión) entered LML, it is possible to observe nine beds made of alternating sands and muds with abundant articulated shells of pelecypods and also gastropods (Fig. 6h).

Chronology
The mean D/L values of three amino acids (Ile, Asx and Glx) in Glycymeris shells of TOG, CCC, CCG and LMB are given in Table 2. The samples from LMB-1, CCG and TOG showed similar racemization and epimerization values as those from raised marine terraces of MIS 5e age (Aminozone E of Hearty et al. 1986;Hearty 1987). Similar D/L values were also reported by Torres et al. (2000Torres et al. ( , 2010Torres et al. ( , 2016, Ortiz et al. (2017) and De Santis et al. (2018, 2020 for MIS 5 deposits in the Mediterranean realm. In contrast, samples from LMB-2 showed similar D/L values and correlated with MIS 7 (Aminozone F of Hearty et al. 1986;Hearty 1987). It is important to note that Asx D/L values did not perform as well as Ile and Glx in the dating of raised marine deposits in the Iberian Peninsula, as pointed out by Torres et al. (2000). In fact, Asx shows a higher racemization rate than Ile and Glx in land snails (Goodfriend 1991;Goodfriend and Meyer 1991) and marine molluscs (Kvenvolden et al. 1979;Lajoie et al. 1980;Kimber and Griffin 1987).
Indeed, consistent with the findings of other studies (Hearty 1986(Hearty , 1987Hearty et al. 1986;Torres et al. 2000Torres et al. , 2010Torres et al. , 2013, we observed that Ile epimerization analysis showed a greater capacity to discriminate between Pleistocene sites of distinct ages than Glx and Asx, as closer Asx and Glx D/L values were obtained for MIS 5, MIS 7 and MIS 11 localities than D-aIle/L-Ile values.
For T. latus samples from CCG (Table 2), we compared the D/L values with those obtained from sites in Tunisia and Spain (Hearty 1986;Torres et al. 2010Torres et al. , 2013. We found that they may belong to MIS 5. No further age-accuracy than Stage level could be established. The mean D/L values and standard deviations of ostracodes and gastropods (Helicidae) of each horizon from cores drilled in LML are shown in Tables 3 and 4, respectively, with their corresponding age. In general, the numerical datings obtained in LML cores were consistent with depth and revealed that materials belonged to MIS 7, MIS 5 and MIS 1 (Fig. 7) although the cores diameter did not allow the observation of any unconformity. MIS 7 deposits were thicker eastwards, while westwards MIS 5 materials increased. MIS 1 deposits were recorded only in the uppermost ca. 20 cm, The amino acid D/L values obtained in Vermetidae samples from a boulder on the intermediate abrasion platform of CCC were very low (Table 2). Although age calculation algorithms were not developed for this taxon, the low amino acid racemization values indicated a recent age, of only a few centuries or even younger.

Diversity
All the species identified in the deposits of El Alquián beach are listed in Table 1. We recovered a total of 1,604 shells from these beach deposits. Gastropods accounted for ca. 39% and pelecypods ca. 61%. Among gastropods, representatives of the Muricidae and Thaididae families accounted for 63% of the whole sample, together with individuals of the species Cancellaria cancellata (7%) ( Table 1), although its presence is not usually common (D'Angelo and Gargiullo 1987). The remaining gastropod species (30%) formed a short selection of common Mediterranean species. The pelecypod species were dominated by G. nummaria (dominant) and Glycymeris glycymeris, accounting for 74% of bivalves, followed far by Acanthocardia tuberculata-paucicostata.
With some negligible exceptions, all pelecypods were inhabitants of shallow waters with soft bottoms. Something similar occurred with the gastropod group, although Hexaples trunculus was an adaptative opportunistic wandering-in group predator (D'Angelo and Gargiullo 1987).
Thus, the mollusc content of El Alquián beach was characterized by high diversity (36 species), typical of surge wave deposits, with pelecypods predominating over gastropods.
Moreover, the pelecypod shells appeared on the surface of the beach and showed a horizontal arrangement, either concave-up or concave-down, and they were rarely tilted, due to the presence of gravel particles forcing this position.

Shell preservation
Small gastropod shells were generally well preserved on El Alquián beach, whereas the largest ones (mostly H. trunculus) showed traces of abrasion and may have undergone depredation. To determine the preservation state of the shells, we followed Rogalla and Amler (2007). Thus, from a total of 1,100 Glycymeris and Acanthocardia shells, ca. 56% showed strong evidence of erosion (Fig. 8a): -22% showed a deeply eroded umbo region, which progressed to form a hole, making them easy to be turned into beads.
-24% showed broken shell margins and the fracture lines were smoothed by abrasion. Due to their thinness, the ventral borders of the shells were broken and/or abraded, to such an extent that we decided to measure the anteroposterior diameter of the shell.

Diversity
In TOG and CCG, almost 95% of the faunal content corresponded to the pelecypods G. nummaria, with A. tuberculata being the associated species. A single well-preserved Trochidae specimen was the exception. Two highly eroded T. latus shells were found. Furthermore, we observed that the shells were positioned on the top of the beds according to hydrodynamic factors: predominantly convex up, although in the inner part of the bed most shells were randomly oriented, and in stepped positions, they were even stacked vertically, like plates in a dishwasher.

Preservation
Following the criteria of Rogalla and Amler (2007), most of the shells from TOG and CCG were well preserved, as they did not show signs of erosion and were not broken (Fig. 8b), except for some specimens, in particular three T. latus shells, which were deeply eroded.

Tsunami versus storm deposits
LMB deposits offer little doubt about their origin and have been widely described as sandy bars that protected and isolated the lagoonal deposits of LML (Somoza et al. 1986; Zazo  . 1990). Likewise, the sandy deposits of CCC were previously interpreted to be of eolian origin (Almela et al. 1978;Silva et al. 2019). In this regard, we dated the Glycymeris fragments of these deposits at MIS 5 (Table 2). These fragments accumulated to form a giant dune, covering Cervera Cape anticline when the platform was exposed after regression. The origins of CCG and TOG need prior considerations. It should be noted that Glycymeris species burrow to colonize wave-protected areas, with an average density of 1.6 individuals/m 2 , have a life span of about 20 years, and are associated with a small number of other pelecypods (Peharda et al. 2010(Peharda et al. , 2012Crncevic et al. 2013;Royer et al. 2013). Despite being protected from the normal wave regimen, Glycymeris shells can be moved towards the foreshore zone and later to the backshore zone (transfer environment) during storms, finally reaching the beach-berm (sedimentary environment). Under "normal wave" conditions, some shells can be transported towards the shoreline, forming narrow rills at the bottom, which appear as lenticular bodies in the geological record (Muñoz and Udías 1991;Reinhard et al. 2006), and are abraded/broken at the surf zone.
The taphonomy of El Alquián beach, a re-sedimented entity, matched very well with the abovementioned characteristics: diversity of mollusc species with dominance of Glycymeris ssp. (pristine diversity), most shells being eroded and/or broken. The shells were horizontally arranged, with no differences between those positioned convex up or convex down. In this regard, LMB-1 and LMB-2 deposits showed these characteristics and were considered beach deposits.
In contrast, in CCG and TOG, species diversity was very low (95% Glycymeris) and showed little fragmentation, chaotic arrangement of shells with elements oriented in a stepped manner (> 60°) with respect to the stratification, and sharp shell fragments (not smoothly eroded by continuous wave abrasion). These findings are consistent with the criteria of Puga-Bernabeu and Aguirre (2017) for tsunamites.
It is unlikely that the Glycymeris shells of CCG and TOG broke during the rapid short transport over a sandy bottom, which deposited its load in a chaotic arrangement where shells did not present preferential orientation. The presence of boulders pulled from preexisting deposits reinforces this interpretation and the lack of bioerosion and encrustation support this hypothesis. The singularity of these deposits is reinforced by the absence of other marine sediments above and below them.
Therefore, the characteristics of CCG and TOG are similar to the sedimentological and paleobiological features of tsunami deposits rather than surge deposits (Dawson and Stewart 2007;Becker-Heiman et al. 2007;Donato et al. 2008;Engel and Brückner 2011;Smedile et al. 2011;Marriner et al. 2017;Mathes-Schmid et al. 2013;Puga-Bernabéu and Aguirre 2017) based mainly on the biodiversity (surge deposits usually consists of a huge number of taxa), preservation (shells of surge deposits are usually more abraded and fragmented) and pattern (in tsunamites a chaotic arrangement of mollusk shells are common, sometimes with elements oriented in a stepped manner). Thus, we tentatively interpret CCG and TOG deposits to be linked to a tsunami that occurred near the end of MIS 5, although we are also aware that this differentiation is a major challenge (Dawson and Stewart 2007;Engel and Brückner 2011) because of the influence of the local ecological conditions.
In this case, the waves could be generated not far from the coast supported by the large number of faults affecting the platform sediments of Pleistocene age in the offshore area of La Mata-Torrevieja or even the present-day seafloor defined by marine geophysics . Indeed, Muñoz et al. (1991) and García-Mayordomo and Martínez-Díaz (2006) situate the epicenter of the Torrevieja earthquake in a foreshore location. In this regard, Torrevieja Fault, which extends 12 km on land and 12 km offshore (Fig. 1) (García Mayordomo and Martínez-Díaz 2006;Perea et al. 2012), has been associated to the origin of diverse earthquakes (Rodríguez de la Torre 1984; Muñoz and Udías 1991; Albini and Rodríguez de la Torre 2001; Martínez Solares and Mézcua 2002). Similarly, Bajo Segura Fault has an offshore prolongation Silva et al. 2017) and a number of faults affect the platform sediments . In this regard, a coseismic regional uplift related to the Torrevieja earthquake has been described to have produced a 20 cm elevation in CCC (Silva et al. 2014(Silva et al. , 2017(Silva et al. , 2019. Nevertheless, Álvarez-Gómez et al. (2010) modelled that processes generated on the coast of Algeria would produce 30 m high tsunami waves in La Mata area, and Álvarez Gómez et al. (2011) modelled that earthquakes originating in the Alboran Sea would cause 0.5 m of maximum wave elevation in Torrevieja.
We consider that the accumulation of TOG and CCG is probably associated with the activity of Torrevieja or Bajo Segura Faults, although a remote source could not be excluded such as in Northern Algeria (Álvarez-Gómez et al. 2010), the Alboran Sea (Álvarez-Gómez et al. 2011) or the Ibero-Maghrebian region (Muñoz et al. 1991). This tsunami swept the ramp where pelecypod (Glycymeris) shells accumulated and reached a MIS 5e bar, leaving a shell bed inland after eroding the red alluvial muddy deposits (Sucina Formation) that appear on the waterline. The erosion of the MIS 5e bar released some deeply abraded shells that were re-sedimented, which appeared at the top of the strata where the ebb arranged the shells in a normal position. Our interpretation of TOG and CCG as tsunami deposits linked to faults activity was supported by the recent work of Ott et al. (2021): after a modelling of an area with a scenario of plate collision, where reverse and normal faults coexist, being responsible of a historical strong earthquake/tsunami event that affected the Crete island coast at 365 AD.
Thus, our observations support the notion that local sources of seismicity or the collision of the African and Iberian plates cause tsunami surges, which pose a serious threat to lowland tourist resorts and salt mining. Indeed, in the context of a recently approved ambitious resort project (at least 18 high rise buildings of 29 floors) for a wide area to the south of Cervera Cape, the seismic hazard should be taken into consideration.

Landscape evolution
Based on the ages obtained here (Tables 3, 4) and the interpretation of the characteristics of the different deposits, we can identify the following phases in the evolution of the Torrevieja-La Mata landscape ( Fig. 9): 1. LMB was active during MIS 7 (240-180 ka), as the age LMB-2 deposits indicated (Table 2), protecting preexisting brackish pond deposits, as attested by the lagoonal back-barrier deposits in LML cores (Fig. 7), which were also dated at MIS 7 (Tables 3,  4). These deposits consisted of a series of foreshore dipping bioclastic sand cosets. 2. During MIS 5 (130-73 ka), probably MIS 5e, a new bar stacked on the former one, as Somoza et al. (1986) and Zazo et al. (1990) observed, showing a strongly marked erosive unconformity. In this regard, LMB-1 deposits were dated at this Stage (Table 2). This bar consisted of bioclastic sands forming foreshore dipping cosets (Fig. 7). As occurred in MIS 7, lagoonal deposits (LML) also appeared, as datings from cores LML-A and B showed (Tables 3, 4). According to Somoza et al. (1986), an erosive cliff (now covered by concrete) was visible at the back-bar, with fallen boulders present at its foot. 3. The fact that TOG and CCG were dated at MIS 5 (Table 2) and interpreted as tsunami deposits (see Sect. 5.1) indicated that during this Stage (MIS 5, 130-73 ka), an earthquake with epicenter at the foreshore took place (although it is not discardable an epicenter located at a larger distance) and was responsible for a tsunami of unknown energy that hit the unconsolidated MIS 5 bar, deeply reworking wave-transported and previously eroded T. latus shells, as well as angular boulders of bioclastic sandstones. Moreover, this tsunami wiped the seafloor, where a paucispeciphic thanatocoenosis made of disarticulated Glycymeris and Acanthocardia shells occurred, that were quickly transported to the shore, building a tabular bioclastic bed with randomly orientated shells (see Sect. 5.1). 4. Associated with the general lowering of the sea level at the end of MIS 5 (Cawthra et al. 2018), the emerged former shoreface deposits were the sand and bioclast source of the CCC eolian deposits. Indeed, some samples of small Glycymeris shells were dated at MIS 5 (Table 2). 5. Samples from the top of short cores in LML revealed deposition during MIS 1 (Table 3).
Thus, La Mata Lagoon also persisted during the Holocene, probably linked to the eus- tatic rise in the level of the Mediterranean Sea at ca. 7.5-7.0 ka cal BP (Somoza et al. 1998;Zazo 1999Zazo , 2006Zazo et al. 2003Zazo et al. , 2008. 6. The CCC uplift during or after MIS 5 were responsible for stepped abrasion platforms and accumulation of fallen boulders (Fig. 5). Moreover, the low D/L values obtained in Vermetidae samples from a large boulder in CCC indicated a recent age ( Table 2). The abundant holes drilled by Lithophaga lithophaga and the presence of Vermetidae shells in this boulder attest that it was underwater before being lifted to the platform, and point to the occurrence of a large storm/tsunami in recent times. Likewise, TOC showed erosive morphologies linked to wave action.
Thus, the major known effect of fault activity in this area was the destructive Torrevieja earthquake of 1829, although low-intensity earthquakes are continuously recorded. Our interpretation of TOG and CCG as tsunami deposits linked to Torrevieja and Bajo Segura Faults or a set of marine faults implies that this geo-hazard has been present since at least the Upper Pleistocene times.

Conclusions
Here, we combined a large number of amino acid racemization datings with paleontological, taphonomic and sedimentological analyses to obtain an interpretation of geological hazards in a coastal area. Indeed, the paleontological and taphonomic analyses allowed us to identify tsunami (TOG, CCG) and surge (LMB) deposits under local paleoenvironmental conditions.
The Pleistocene evolution of the coast revealed that a series of very important processes took place during MIS 7 and MIS 5. The spatial relationship of these deposits with a bar dated at MIS 5 lying over ancient bar deposits of MIS 7 through erosive unconformity suggested elevation processes affecting the first one linked to recent tectonics. In this regard, based on the mollusc shell taphonomy of some raised deposits, the area studied showed evidence of a tsunami during MIS 5, with its epicenter near the coast, probably associated with Torrevieja and Bajo Segura Faults or the set of faults linked to Torrevieja Fault. At the north of Cervera Cape, this tsunami was responsible for CCG deposits, which spread far into the hinterland. At the south of Cervera Cape, this event hit the cliff carved on the sediments of the Sucina Formation cropped out and deposited TOG.
Torrevieja and Bajo Segura Faults controlled and continue to control the present-day morphology of the coast, driving the uplift of Cervera Cape at the south, while producing subsidence in the north. Indeed, in the area of Cervera Cape, a continuous tectonic-related (earthquakes) uplift has been recorded, even in historical reports, confirming a true geological hazard. The faults in the area caused the Torrevieja earthquake in 1829, one the most destructive in Spain, although low-intensity earthquakes are registered continuously. The evidence of tsunami deposits (CCG, TOG) linked to these faults also during the Pleistocene implies that this hazard has been present over history.
Therefore, given that this area is highly populated and new resorts (high rise buildings of 29 floors) are planned, the seismic hazard should be taken into consideration. Thus, the area between Torrevieja and La Mata, as well as La Mata Lagoon (Natural Park) and the Torrevieja Lagoon saltworks, could be strongly affected by earthquake-induced tsunamis, a rise in sea level, and cliff retreat, among other geological hazards.