A tsunami record in the sedimentary archive of the central Algarve coast,Portugal: Characterizing sediment,reconstructing sources and inundation paths

Abstract

This study describes sedimentation associated with the tsunami generated by the Lisbon earthquake of ad 1755. It is argued here that the tsunami deposited a sand sheet across the Lagoa dos Salgados (central Algarve, Portugal), that is intercalated with late-Holocene estuarine/lagoonal sediments. A wide range of proxies (sedimentological, exoscopic and palaeontological) are used to establish the provenance of the sandy material as well as to constrain the age of the deposit. Stratigraphic criteria are used to distinguish the uniqueness of the event layer. Exoscopic and textural analysis suggest that the source of the event deposit is mainly the dune, beach and underlying layer. Micropalaentological analysis (Foraminifera) indicates a conspicuous increase in diversity and dominance of marine species within the event sediment sheet. The spatial characteristics of the tsunami layer suggests that the barrier prevented widespread overtopping by the incoming tsunami allowing inference of c. 10 m as maximum height at the coast; they also indicate the inlet as the preferential route for both water and sediment transported inland. Dating results (¹⁴C, ²¹⁰Pb and ¹³⁷Cs) allow extrapolation of an age of deposition compatible with the ad 1755 Lisbon earthquake and tsunami, the most devastating event that affected this coastal area in historical times. Correlations with similar deposits detected in nearby lowlands strengthen the argument that the tsunami sediment layer represents a marker horizon in the coastal stratigraphy along the Portuguese Algarve coast. The stratigraphic uniqueness of this event might have implications in the establishment of millennial scale recurrence intervals for this type of high-energy marine inundation.

Keywords

exoscopy Foraminifera Portugal texture tsunami sedimentation

Introduction

The study of tsunamis in the geological record is fundamental to the assessment of tsunami hazard and mitigation of their well-known devastating effects. Although much has been learned in recent years regarding the identification of tsunamis in the geological record many questions remain unanswered (e.g. erosional and depositional patterns, the identification of sediment sources, recurrence intervals). From a geological point of view, tsunamis are notable as short-lived but extremely powerful agents of coastal change associated with complex patterns of erosion and deposition (Bondevik et al., 1997). In contrast with modern tsunamis, for which eyewitness accounts and field measurements of both erosional and depositional effects are known (e.g. Gelfenbaum and Jaffe, 2003; Moore and Moore, 1988; Paris et al., 2009, 2010; Szczucinski, 2011), the recognition of palaeotsunami deposits and reconstruction of the generating events depends solely on the interpretation of geological data. The sedimentary recognition of tsunami deposits is based upon a group of textural, lithostratigraphic, palaeontological and geochemical data that has been used by many authors over recent years (Albertão and Martins, 1996; Andrade et al., 2003; Atwater, 1986; Chagué-Goff, 2010; Clague et al., 2000; Dawson, 1994; Dawson et al., 1988, 1995, 1996; Goff et al., 1998; Minoura et al., 1994; Moore et al., 2006; Nanayama et al., 2000; Switzer and Jones, 2008).

The aim of this study is to investigate the sedimentary characteristics of a distinctive lithostratigraphic unit preserved within the coastal sediments of the central Portuguese Algarve coastline, an area known to have been affected by the destructive Lisbon tsunami of ad 1755, and to discuss its association with this particular inundation using a range of proxies to reconstruct flow patterns, inundation paths and tsunami wave characteristics at the coast.

Study site

Although rarely subject to extreme storms or tsunamis, the Algarve coast is the most vulnerable area in terms of tsunami hazard in Portugal (Baptista and Miranda, 2009). Several historical accounts (e.g. Costa et al., 2005; Lopes, 1841; Sousa, 1919) describe in detail how the tsunami waves affected the coastal area of the Algarve that was devastated by the Lisbon tsunami of ad 1755. At two lowland locations (Martinhal and Boca do Rio) sandy deposits attributed to this tsunami have already been described (Figure 1; Andrade et al., 1997; Hindson and Andrade, 1999; Hindson et al., 1996; Kortekaas and Dawson, 2007). Historical descriptions of older tsunamis believed to have affected the Algarve coast have also been reported and include the Tavira tsunami of ad 1722 and much larger events in 63 bc and ad 382 that may have affected the whole Algarve coast (Baptista and Miranda, 2009).

Figure 1.

Locations where tsunami deposits have been described along the Gulf of Cadiz, Iberian and Algarve coastlines (sandy deposits numbered in red; boulder deposits numbered in black). 1: Corrochano et al. (2000); 2: Andrade et al. (2003); 3: Scheffers and Kelletat (2005); 4: Abrantes et al. (2005, 2008); 5: Morales et al. (2008); 6: Ruiz et al. (2005); 7: Luque et al. (2001); 8: Whelan and Kelletat (2005); 9: Reicherter et al. (2010); 10: Andrade et al. (1997), Kortekaas and Dawson (2007); 11: Costa et al. (2011); 12: Dawson et al. (1995), Hindson and Andrade (1999), Hindson et al. (1996); 13: Costa et al. (2009); 14: Schneider et al. (2009); 15: Andrade (1992).

The highest astronomical tides in this area are c. 1.8 m above mean sea level (a.m.s.l.) and during extreme spring tides exceed 2 m a.m.s.l. (Hindson and Andrade, 1999). The mean tidal range is approximately 2.1 m but reaches c. 3 m during spring tides. The maximum reported elevations reached during the highest astronomical tides coincident with storm surges can reach as high as c. 2.15 m a.m.s.l. (Esaguy, 1984). Taborda and Dias (1992) based on the study of two major storms (14 February to 3 March 1978 and 25 to 31 December, 1981) concluded that the component of storm surge was associated with an average increase in sea surface elevation of 0.42 m in Lagos (see Figure 1d for location).

The wave regime along the Algarve coastline is typically low energy, with mean annual significant wave height (Hs) < 1 m. During storms, the prevailing wave direction is associated with westerly winds with most large waves approaching from the WSW to SW.

The Salgados lowland is located in a bay between Armação de Pêra and Galé (Figures 2 and 3). This coast features a 6 km long intermediate-reflective sand beach backed by a 3 to 17 m high vegetated multiple dune ridge (Figure 2), the latter covering cemented Pleistocene–Holocene aeolianites and beachrock (Moura et al., 2007; Pereira and Soares, 1994; Pinto and Teixeira, 2002). The continuity of the beach-dune system is interrupted by the ephemeral inlets of two infilled lagoons (Alcantarilha and Salgados, Figure 3a), that collect water and sediment delivered by the Alcantarilha stream and the Salgados and Espiche streams, that respectively drain 204 and 41 km² catchments, mostly developed in early-Miocene limestone and late-Miocene siltstone and sandstone, with the sandy Plio-Pleistocene formation of Areias de Faro- Quarteira underlying the Holocene stratigraphic sequence infill of both lowlands. The Salgados lowland covers an area of 1.5 km² with the eastern part having been reclaimed and landfilled for a golf course (Figure 3). The remaining area is a flat-floored lagoon the bottom developing between 1.1 and 1.7 m a.m.s.l., usually covered by a 1 m water column.

Figure 2.

Oblique aerial view of Salgados lowland in March 2011.

Figure 3.

The Salgados lagoon (Lagoa dos Salgados) and sampling scheme. (a) Location in the Armação Pêra-Galé coastal ribbon and nearshore samples – depths in metres below m.s.l.; (b) close up of the lowland showing the area reclaimed for the golf course and location of beach, dune, lagoon and alluvial plain samples; (c) coring locations at the lowland and nearby alluvial plain.

This bay was particularly affected by the ad 1755 tsunami. The consequences registered in historical documents were summarized by Oliveira (2009). Sousa (1919) commented that (author’s translation, comments in italic) ‘[…] in Armação [de Pera], the fortress was destroyed by the sea, and due to its impetus, took the Church of Santo António leaving only a few stones, as also sixty-two persons, that the sea took and afterwards threw them dead’. Lopes (1841) mentioned that ‘the sea left one house standing in the earthquake; rushed more than ½ league [1 league = c. 6 km] inland, flooding everything, leaving salt water lakes in the lowlands, creating islands and drowning 84 people’. Another historical description in Costa et al. (2005) refers to the importance of the backwash: ‘Armação de Pera, counted fifty-five houses, none of which fell with the earthquake, but were instead swallowed by the sea destroying more than half of the houses, the fortress […], leaving no trace of its location; taking sixty-seven people in the backwash’.

Methods

In total, 158 sediment cores (up to 5 m) were obtained from the Salgados lowland (Figure 3) using a hand-operated gouge, Edelman augers, van der Horst, van der Staay and Livingstone corers. Moreover, approximately 50 surface samples from present-day sedimentary environments, including the nearshore between 4 and 19 m below m.s.l., beach face, beach berm, dune, dune crest and alluvial plain were collected by hand or directly collected from the sea bottom by scuba diving.

Visual qualitative descriptions of sedimentary materials and units in the field included evaluation of colour, lithology, texture, lithostratigraphy and macrofossil content. Sedimentary logs were constructed and preliminary correlations between cores were undertaken.

Bulk samples weighing more than 30 g were sieved using a standard set of sieves at 0.5 φ intervals. In cases where samples weighed less than 30 g the grain-size measurements were conducted on the laser granulometer Malvern Mastersizer Hydro 2000 MU after deflocculating with 30% sodium hexametaphosphate. Results derived from both methods were assumed to be comparable. Statistical grain size parameters (e.g. mean grain size, D₅, D₁₀, D₉₀, median, mode, standard deviation, kurtosis and skewness) of the samples were calculated using the graphic method (Folk and Ward, 1957) in the GranGraf software (Carvalho, 1998). A group of textural features were map-plotted to study their spatial distribution. This was done using the Inverse Distance Weighting (IDW) method of interpolation between sampling points in the ArcGis 9.3 environment. IDW is a method of interpolation that estimates cell values by weight-averaging the values of sample data points in the neighbourhood of each processing cell according to a non-linear function of distance (Davis, 1986). Carbonate content, CaCO₃ (associated with bioclasts), was determined by measuring the volume of carbon dioxide evolved during reaction with hydrochloric acid using an Eijkelkamp calcimeter (Eijkelkamp, 2008). Organic matter content (OM) was determined by loss-on-ignition at 550°C following Bengtsson and Enell (1986). For benthic foraminiferal studies, 60 samples from present-day and cored sediment were wet sieved over a 4 φ (63 µm) mesh and the retained (coarse) fraction oven-dried (Fatela et al., 2007). Samples were then split down to an amount containing approximately 100–200 Foraminifera. These were hand-picked and sorted on slides for identification and counting using a standard binocular microscope.

Scanning Electron Microscope (SEM) exoscopic analysis was undertaken on c. 1000 quartz grains collected from 45 samples obtained in present-day coastal environments (nearshore, beach and dune) and cored materials of lithostratigraphic unit E (see below). In each sample a number of 12 to 30 quartz grains from of the 1 to 3 φ (125–500 µm) size fraction were randomly selected under the binocular microscope and prepared for SEM analyses. In the literature the number of grains reported as necessary for sedimentological interpretation varies within a wide range (cf. Mahaney, 2002 for a summary and examples). In this study, an empirical assessment was conducted to evaluate the minimum size of the sample with statistical meaning, focusing on the number of grains required to hold the median value characterizing each microtextural attribute reasonably constant within the sample (cf. Appendix 1 for details). Our results suggest that the median value is about 20 grains.

Each grain was characterized using a number of descriptive microtextural features revealed in microphotographs (e.g. fresh surfaces, percussion marks, dissolution features, angularity and adhering particles); microtextures in each grain were ranked according to the area of the surface of the grain they occupied, ranging from 0 (absent) to 5 (occupying more than 75%). Additionally, a broader classification was used, dividing the grains in three essential types: A: dissolution features dominating the surface of the grain; B: stronger presence of fresh surfaces; and C: percussion marks dominating the surface of the grain.

Alpha spectrometry of ²¹⁰Po was used for ²¹⁰Pb determinations on a sediment core retrieved for this purpose, assuming secular equilibrium between the two isotopes. ¹³⁷Cs was measured by non-destructive gamma spectrometry (Jouanneau et al., 2008). Both ²¹⁰Pb and ¹³⁷Cs were determined at Université de Bordeaux, France and the vertical concentration profile of ²¹⁰Pb excess used to infer a uniform sedimentation rate using the constant rate supply method (Appleby and Oldfield, 1983). Eleven 1 cm thick organic sediment samples cored from Lagoa dos Salgados were dated at the Beta Analytic Radiocarbon Dating Laboratory, Miami, USA, the dating having considered the carbon-rich organic particulate matter dispersed within the sediment. Calibration was done using the CALIB 6.0 Radiocarbon program coupled with the IntCal09 curve (Stuiver and Reimer, 1993).

Results

Radiocarbon dating, ²¹⁰Pb and ¹³⁷Cs determination and sedimentation rates

Eleven sediment samples were collected from organic-rich layers or laminae in different cores where the major lithostratigraphical units were clearly identified and allowing for lateral correlations (see below; Table 1). Differences in thickness of the same unit across the lowland and lateral intra-unit lithofacies changes complicate in some cases the depth–age relationships. In addition, the large sand/mud ratio found in all sediment packages and all cores, precluded the construction of a global age–depth model. Samples were radiocarbon dated with results shown in Table 1 and Figure 4. The δ¹³C values suggest a strong contribution of terrestrial carbon to the organic material dated (Lowe and Walker, 1997) and the oldest sediment dated roughly corresponds to the mid Holocene.

Table 1.

Radiocarbon dates obtained for Lagoa dos Salgados top Holocene stratigraphic sequence.

Core	Lithostratigraphic unit	Depth below surface (m)	Height (above m.s.l.)	Material	δ¹³C ‰	Conventional radiocarbon age	2 sigma calibrated result	Intercept calibration curve
SG_LV12	D	0.77	0.65	organic sediment	–25.1‰	980±40 yrs BP	yrs cal. BP 790-960	cal. ad 1030
SG92	D	0.46	0.82	organic sediment	–24.9‰	1300±40 yrs BP	yrs cal. BP 1170-1300	cal. ad 680
SG37	D	0.58	0.74	organic sediment	–25.3‰	1380±40 yrs BP	yrs cal. BP 1270-1340	cal. ad 650
SG_LV7	D	0.59	0.68	organic sediment	–25.2‰	1420±40 yrs BP	yrs cal. BP 1280-1380	cal. ad 640
SG39	D	0.62	0.69	organic sediment	–25.4‰	1610±40 yrs BP	yrs cal. BP 1400-1570	cal. ad 420
SG1	B	2.43	–1.04	organic sediment	–22.5‰	2820±40 yrs BP	yrs cal. BP 2850-3030	cal. bc 980
SG6	B	2.92	–1.61	organic sediment	–25.6‰	4710±40 yrs BP	yrs cal. BP 5320-5580	cal. bc 3510
SG46	B	2.71	–1.39	organic sediment	–25.1‰	4760±40 yrs BP	yrs cal. BP 5330-5590	cal. bc 3630 or 3580 or 3530
SG5	B	2.89	–1.52	organic sediment	–25.6‰	4850±40 yrs BP	yrs cal. BP 5480-5650	cal. bc 3640
SG47	B	3.11	–1.37	organic sediment	–25.1‰	4940±40 yrs BP	yrs cal. BP 5600-5740	cal. bc 3700
SG1	A	4.32	–2.93	organic sediment	–19.1‰	5920±50 yrs BP	yrs cal. BP 6650-6880	cal. bc 4790

Figure 4.

Schematic log of Lagoa dos Salgados showing main Holocene lithostratigraphic units resting upon Plio-Pleistocene dissected basement (see text for details); the relative location of cores from which 11 radiocarbon dates and one 210Pb and 137Cs concentration profile were obtained are also shown in Figure 3.

Sedimentation rates were established based upon determination of ²¹⁰Pb and ¹³⁷Cs for the top 0.35 m of the stratigraphic column of core SG48 (Figure 3). Results indicate a time-integrated accumulation rate derived from the vertical profile of ²¹⁰Pb excess of c. 2.6 mm/yr. Moreover, taking ad 1963 as the maximum nuclear testing in the Northern Hemisphere (Ritchie and McHenry, 1990) the accumulation rate obtained from the ¹³⁷Cs peak at 0.10 m (Figure 4) below surface is 2.3 mm/yr, which is in rough agreement with the value inferred from ²¹⁰Pb.

Lithostratigraphy

A schematic summary log (Figure 4) based on cores collected in Lagoa dos Salgados shows six lithostratigraphic units deposited on top of the Plio-Pleistocene (Areias de Faro-Quarteira) basal sand.

Unit A

The lowermost unit, with a maximum thickness of 2.48 m (core SG1), rests unconformably on the reddish fluviatile Plio-Pleistocene materials and is characterized in the seaward region of the basin by alternations of yellowish to grey sand with marine shell fragments and less abundant and thinner clay laminae. One radiocarbon date obtained from an organic mud lamina of the upper section of this unit, at c. 4.32 m below surface (i.e. c. 3 m below m.s.l.) in SG1, yielded a radiocarbon age of c. 4790 yr cal. bc (Figure 4 and Table 1). This indicates that most of the coarser sediment of the basal unit accumulated in an essentially high-energy and shallow environment open to marine influence, preceding the pronounced drop in sea-level rise rate at 6000 to 7000 cal. BP that characterizes the Holocene transgression in the Algarve and elsewhere along the western Portuguese coast (e.g. Freitas et al., 2003; Teixeira et al., 2005; Vis, 2009). Two samples (obtained from cores SG1 and SG46) were used for sedimentological characterization of this unit. The values obtained for each core were conditioned by their position in terms of proximity to the coast confirming field observations: core SG1 presented 77.2% of coarse material (>63 µm), 8.9% for CaCO₃ content and 1.6% of OM. On the other hand core SG46 (farther inland) presented 18.5% of coarse material, 1.1% for CaCO₃ content and 3.8% in terms of OM.

Unit B

This unit, with a maximum thickness of 1.32 m (core SG44), is mainly composed of grey to brown mud with occasional intercalated layers of coarser sediment which wedge out laterally, indicating a significant decrease in the energy level of the environment with increased input of fine-grained terrigenous materials. The radiocarbon dates (obtained in different cores at c. 2.4 to 3.1 m below surface, i.e. c. 1.0 to 1.6 m below m.s.l.) constrain the deposition of this unit within the 3700 to 980 yr cal. bc time span (Figure 4 and Table 1), a period of quasi-stability of mean sea level that should have favoured the build-up of the barrier and beach-dune system at the exposed façade of the Armação de Pêra embayment and increased the frequency of sediment choking of the inlet, thus favouring the retention of terrigenous sediment within the lowland. Twelve samples from eight different cores allowed characterization of this unit in terms of percentage of coarse material, CaCO₃ and OM. The average and standard deviation values were 22.9±14%, 10.7±9.3% and 5.4±1.6% respectively, indicating persistence of marine signature but reduction of the energy level across the whole basin.

Unit C

This unit rests conformably with the under- and overlying units. It essentially consists of yellowish to greyish medium to fine sand containing marine shell (bivalves) fragments, and exhibits normal (fining upward) grading. Sand is intercalated with numerous muddy laminae. It presents a maximum thickness of c. 3.25 m (core SG2). An age range (approximately 1500 to 3000 cal. BP) for this unit was estimated extrapolating the radiocarbon ages obtained for units B and D. Thirty-three samples from 13 different cores allowed characterization of this unit in terms of percentage of coarse material, CaCO₃ and OM; the average and standard deviation values for the mud dominated intercalations were 25.3±15.1%, 11.6±9.4% and 5.87±3.0% respectively; and for the sand dominated material 90.1±9.9% and 20.8±9.9%, respectively, the OM being virtually nil.

Unit D

This unit is composed of brownish silt and clay material that becomes darker to the top. It presents a maximum thickness of approximately 0.75 m (core SG46) and exhibits an abrupt contact with the overlying unit E. Five radiocarbon dates were obtained from organic sediment (at c. 0.45 to 0.75 m below surface; i.e. c. 0.65 to 0.85 m a.m.s.l.) of this unit (Table 1) constraining its age of deposition between yr cal. ad 420 and 1030. Thirty-five samples from 32 different cores allowed characterization of this unit in terms of percentage of coarse material, CaCO₃ and organic matter content; the average and standard deviation values were 12.2±16.7%, 9.2±7.0% and 4.1±1.3%, respectively.

Unit E

This unit is essentially composed of medium to fine sand, which contrasts markedly with the underlying sediment of unit D and is separated from it by an unconformity. Unit E is widely distributed in the lowland and present in several cores although its thickness decreases (from 0.79 m (core SG LV 10A) to a few millimetres further inland (e.g. core SG 72)) up to its landward limit c. 800 m from the present-day coastline. The altitude of the basal contact of unit E rises gradually inland from c. −0.05 m up to 1.28 m a.m.s.l. This unit is laterally variable in terms of its thickness and sedimentary characteristics. It is observed in its seaward section as a massive sandy deposit with no identifiable laminations and without sedimentary structures. Several centimetre-sized mud rip-up clasts from the underlying unit were identified throughout this unit although more noticeable at its base. However, further inland this unit is characterized by finer sediments; this spatial contrast is expressed by higher φ₁₀, φ₅₀ and φ₉₀ (implying shift towards finer size interval of the whole >63 µm grain size spectrum) with increasing distance to the inlet, and simultaneous enrichment in mud-sized particles. Forty-five samples from 42 different cores allowed characterization of this unit in terms of percentage of coarse material, CaCO₃ and MO. The average and standard deviation values for the massive sand were 87.6±4.9%, 28.7±9.0%, respectively, with no OM. Further inland the corresponding values were 54.24±15.5%, 23.0±7.7% and 2.32±0.54%.

Unit F

The topmost unit is characterized as a dark brown silty/clayish layer presenting a maximum thickness of 2.20 m (core SG47) and corresponds to the present-day sedimentary environment. In the field, units F and D are indistinguishable in places were unit E is absent. The upper (0.2–0.10 m) centimetres of this unit are, in places, black. Field observations revealed some coarser intercalations within the mud restricted to an area closer to the inlet and extending northwest along the back barrier. ²¹⁰Pb and ¹³⁷Cs sedimentation rates were established for the top 0.35 m of the stratigraphic column (Figure 4) and the results suggest that the low-energy deposits resting on top of the coarser unit E in core SG48 are younger than ad 1865. Forty-four samples from 35 different cores allowed characterization of unit F in terms of the percentage of coarse material, CaCO₃ and MO. The bulk of unit F yielded average and standard deviation values of 18.8±14.2%, 9.2±4.2% and 4.5±1.32%, respectively. In contrast, the coarser intercalations presented values of 66.6±9.6%, 20.3±8.2%, respectively, the OM being virtually absent.

Summary

A repeating lithostratigraphic sequence occurs in the flat-floored depression of Lagoa dos Salgados (Figure 5). Within this stratigraphical sequence a peculiar feature was observed in the geometrical distribution of unit E. This unit is thicker in the coastal edge of the lowland space and it decreases to millimetric thickness in the northern part of the Lagoa dos Salgados, just before wedging out. In Figure 5a, it is observed that unit E decreases its thickness with increasing distance inland, presenting c. 0.50 m in the core section nearest to the coast and decreasing to millimetric thickness at about 400 m inland. On the other hand, profiles B–B’ and C–C’ exhibit unit E consisting of a very thin (centimetric) sand lamina quickly disappearing inland. Thus, one can conclude that unit E is a distinct wedge-shaped and ramping up lens of coarse sediment extending across the lowland for several hundred metres inland, being thickest at the coast, closest to the present-day inlet.

Figure 5.

Cross-sections based in the stratigraphy of cores analysed across the Salgados lowland. A profile N–S; B profile NE–SW; C profile NW–SE.

Grain-size characteristics of unit E

The analysis of the geographical variation of a group of geometrical and textural features (e.g. thickness, percentage of sand, percentage of carbonates, φ₅₀, φ₁₀, φ₉₀, mean, mode, standard deviation, kurtosis and skewness) of sediment from unit E (Table 2) was also conducted. The percentage of sand (Figure 7a) decreases inland with maximum values (up to 100%) in the southeastern areas of the basin and presents lowest values (up to 20%) in the northern and northwestern sectors of the lowland. The lower values found closer to the landward edge of the deposit mirror the decreasing transport competence and capacity of a landward transport mechanism. However, it may, in cases, be also influenced by the very small thicknesses of the unit E in the more distal locations, where field subsampling could not avoid contamination of the sand by the over- and underlying mud. The space distribution of mean grain size (Figure 6a) exhibits a pattern congruent with that of the proportion of sand. The largest values (1.06 ϕ) occur in the southeastern area and the lowest values in the north and northwest sectors. Similar patterns were also observed for median, mode, φ₉₀ and φ₁₀ parameters (Table 2).

Table 2.

Grain size parameters of tsunami deposit (unit E) from Lagoa dos Salgados.

Sample	Depth (m)	% Sand	1% ϕ	10% ϕ	50% ϕ	90% ϕ	Mean (MZ) ϕ	Graphic std deviation (σ1)	Graphic kurtosis (K_G)	Simple skewness measure (SKM)	Mode ϕ	Median ϕ
LV10	0.46-0.47	100.00	–0.19	0.43	1.04	1.74	1.06	0.51	1.05	0.15	1.25	1.04
LV10	0.91-0.92	100.00	–0.44	0.44	1.06	1.85	1.10	0.56	1.18	0.32	1.25	1.06
LV10	0.56-0.57	99.56	–0.02	0.49	1.11	1.84	1.13	0.53	1.08	0.22	1.25	1.11
LV10	0.64-0.65	100.00	–0.22	0.46	1.21	1.94	1.20	0.58	1.06	0.02	1.25	1.21
LV10	0.80-0.81	99.84	0.13	0.62	1.21	1.85	1.22	0.48	1.02	0.08	1.25	1.21
9	0.43-0.77	93.33	–0.05	0.55	1.32	2.43	1.38	0.91	1.62	1.73	1.25	1.32
13	0.36-0.83	91.44	0.15	0.71	1.39	2.74	1.47	0.86	1.74	1.81	1.25	1.39
15	0.45-0.73	95.58	0.06	0.68	1.52	2.65	1.57	0.82	1.29	0.86	1.25	1.52
11	0.445-0.70	92.71	0.16	0.77	1.53	2.94	1.61	0.89	1.49	1.57	1.25	1.53
39	0.42-0.59	90.26	0.05	0.75	1.59	3.45	1.64	0.90	1.61	1.39	1.75	1.59
LV7	0.45-0.46	86.16	0.52	1.07	1.77	4.04	1.95	0.90	1.40	1.45	1.75	1.77
35	0.47-0.55	68.34	–0.19	0.17	1.49	4.12	1.99	1.56	0.52	1.17	4.25	1.49
20	0.43-0.53	84.63	0.70	1.17	1.80	4.05	2.12	0.97	1.43	1.52	1.75	1.80
17	0.39-0.665	83.89	0.31	0.78	1.50	4.05	2.14	1.30	1.28	1.73	1.25	1.50
99	0.44-0.48	79.61	0.23	0.67	1.63	4.08	2.16	1.35	1.04	1.37	1.75	1.63
95	0.43-0.44	72.79	0.00	0.60	1.68	4.11	2.18	1.39	0.51	1.20	4.25	1.68
101	0.45-0.47	66.21	–0.04	0.59	1.77	4.13	2.22	1.41	0.52	0.99	4.25	1.77
103	0.40-0.43	68.90	0.17	0.70	1.84	4.12	2.27	1.36	0.52	0.98	4.25	1.84
106	0.45-0.465	80.87	0.43	0.89	1.88	4.07	2.32	1.26	0.98	1.08	1.75	1.88
93	0.415-0.445	61.29	0.19	0.78	2.03	4.15	2.37	1.34	0.54	0.70	4.25	2.03
7	0.46-0.48	67.86	0.45	0.93	2.10	4.13	2.44	1.25	0.54	0.74	4.25	2.10
37	0.40-0.56	83.13	0.68	1.19	1.97	4.06	2.44	1.14	1.10	1.19	1.75	1.97
LV7	0.40-0.41	82.66	0.62	1.15	2.05	4.06	2.46	1.16	1.01	0.98	1.75	2.05
91	0.39-0.44	64.86	0.38	0.97	2.26	4.14	2.52	1.24	0.56	0.44	4.25	2.26
41	0.55-0.575	85.99	–0.26	1.28	2.41	4.04	2.53	1.01	1.18	0.23	2.25	2.41
137	0.42-0.45	59.39	0.34	0.95	2.41	4.15	2.58	1.25	0.56	0.12	4.25	2.41
89	0.43-0.45	63.89	0.68	1.16	2.34	4.14	2.61	1.17	0.56	0.45	4.25	2.34
52	0.24-0.57	91.04	0.85	1.68	2.66	3.91	2.68	0.79	1.12	0.14	2.75	2.66
141	0.46-0.465	53.96	0.59	1.18	2.74	4.16	2.77	1.17	0.57	–0.35	4.25	2.74
115	0.37-0.40	55.96	0.88	1.32	2.71	4.16	2.79	1.13	0.57	–0.17	4.25	2.71
139	0.41-0.44	51.19	0.37	0.99	2.96	4.17	2.80	1.23	0.58	–0.94	4.25	2.96
59	0.36-0.42	81.23	–0.97	1.39	2.81	4.07	2.89	1.04	1.09	–0.63	2.75	2.81
24	0.50-0.51	57.32	1.16	1.75	3.02	4.16	3.02	0.95	0.58	–0.31	4.25	3.02
31	0.49-0.52	44.89	0.50	1.17	4.02	4.18	3.22	1.17	0.61	–2.92	4.25	4.02
64	0.44-0.455	41.10	0.35	1.18	4.03	4.19	3.24	1.17	0.64	–2.98	4.25	4.03
66	0.44-0.46	45.06	0.59	1.30	4.02	4.18	3.25	1.13	0.62	–2.84	4.25	4.02
22	0.57-0.58	34.58	1.13	1.80	4.05	4.20	3.42	0.94	0.68	–2.33	4.25	4.05
33	0.49-0.53	21.24	0.49	1.68	4.09	4.23	3.52	0.96	8.61	–2.78	4.25	4.09
26	0.495-0.50	39.59	1.56	2.20	4.03	4.19	3.53	0.79	0.67	–1.87	4.25	4.03

Figure 6.

(a) Interpolated values of mean grain size (D50). (b) Interpolated values of D10.

The analysis of textural parameters indicates that unit E is coarser and thicker across the southern part of Salgados basin whereas the majority of the finer-grained sediment occurs in the northern to northwestern part (Figures 6 and 7). These features point to a seaward source of unit E and dispersion path from seaward (coarser and thicker) to landward (finer and thinner). According to Friedman (1961) positive skewness in medium to fine sand could be the result of unidirectional flow or the result of rapid deposition preventing the washing out of finer particles. In our study (Figure 7b) a skewness pattern compatible with the above-mentioned interpretation can be observed. The combination of textural results and geometrical data also suggest a north–south trending area (possibly a palaeochannel) in the middle of the basin as a preferred route through which sediment was transported inland. It is also worth noting that regardless of the distance to this channel, the plot of textural attributes in Figures 6 and 7 trends obliquely to the elongation of the barrier.

Figure 7.

(a) Percentage of sand of unit E sediments. (b) Skewness of unit E sediments.

Although the mean grain size of unit E sediment is c. 2.3 φ, it varies considerably with location within the lowland. When comparison is established between these sediments and their possible sources, represented by the studied present-day analogues, the widest variation in mean sand size is detected in sand from unit E, ranging from 0.65 to 2 φ (Figure 8a). The distal nearshore samples (collected between 4 and 19 m below m.s.l.) and materials from unit D are finer grained, presenting an average mean diameter of c. 2.50 and 2.20 φ, respectively. By contrast, the beach and dune sedimentary environments exhibit lower values (i.e. coarser sediments) of 0.75 and 1 φ, respectively. The dimensional ranges of the mean grain size obtained from the different environments indicate that only nearshore samples do not overlap the mean grain size distribution of unit E sediments (Figure 8a).

Figure 8.

(a) Mean sand size of the Lagoa dos Salgados samples. n indicates the number of samples used for each possible source (unit E only considered cores thicker than 0.02 m). (b) Percentage of sand and carbonate within unit E on core LV10. (c) Percentage of sand and carbonate within unit E on core LV7.

In the few cores that allowed the study of the vertical variation of textural attributes within unit E, the mean grain size, % sand and % carbonate remain virtually invariant or slightly decrease up unit (Figure 8b and c). However, a sharp decrease in the percentages of sand and carbonates, not related with the presence of rip-up clasts, was detected just above the basal contact of unit E. This suggests a change or ephemeral stoppage in the sedimentary process responsible for the deposition of unit E.

Micropalaeontological analysis

The results from the foraminiferal study of cored sediments indicate a marine to brackish signature throughout the sediment infill of the Salgados lowland (Table 3 and Appendix 2).

Table 3.

Summary of Foraminifera data and palaeoecology of Lagoa dos Salgados.

Units	Foraminifera assemblages					S	α-Fisher Index	Density			Paleoecology
	Dominant species	Average (%)	Max. (%)	Min. (%)	Acessory species (average)			Forams/g dry sediment
	Dominant species	Average (%)	Max. (%)	Min. (%)	Acessory species (average)			Min.	Max.	Average
F	J. macrescens	87.0	87.0	87.0	H. germanica (3%)	15	4	5.5	173	57	High marsh
	M. fusca	10.0	10.0	10.0	H. germanica (3%)						High marsh
	H. germanica	29.0	45.2	10.5	C. lobatulus (5%)						Brackish low intertidal to subtidal lagoon
	H. depressula	21.0	39.0	5.7	E. discoidalis (3%)
	Haynesina spp. (juveniles)	17.4	37.7	7.4	A. mamilla (3%)
	Cibicides spp.	10.0	28.2	0.8	Marine and brackish species (%<3%)
E	Cibicides spp.	21.0	40.3	6.9	A. mamilla (5%)	29	9	7.2	420	105	Subtidal, allocthonous marine forms
	C. lobatulus	9.0	14.6	1.6	Elphidium spp. (4%)
	C. cf pseudoungerianus	6.0	12.4	0.8	M. concentrica (4%)
	E. discoidalis	7.0	16.9	2.9	Haynesina spp. Juvenils (3%)
	E. macellum	6.0	10.6	2.2	E. crispum (3%)
	Quinqueloculina spp.	6.0	10.6	2.2	C. gerthi (3%)
	H. depressula	7.0	22.1	2.2	C. refulgens (3%)
	H. germanica	6.0	14.5	0.6	Marsh, lagoon and marine species (%<3%)
D	H. germanica	31.0	54.0	8.1	Cibicides spp. (4%)	11	3	0.4	513	105	(Bulk of unit) Brackish low intertidal to subtidal lagoon
	H. depressula	20.1	52.9	7.5	A. tepida (4 %)
	Haynesina spp. (juveniles)	29.8	56.0	6.5	Elphidium spp. (3%)
	C. lobatulus	10.0	28.2	0.8	Marine and brackish species (%<3%)
					E. crispum (4%)						(Base of unit) Subtidal lagoon with allocthonous marine forms and reworked tests in beach/dune
	Cibicides spp.	44.0	53.5	35.7	A. mamilla (4%)
	Elphidium spp.	13.0	25.8	6.7	C. lobatulus (5%)
	E. macellum	8.0	15.7	4.3	Glabratella spp. (4%)
	E. discoidalis	8.0	11.3	2.8	M. concentrica (3%)
					Marine species (%<3%)
C					A. tepida (5 %)	20	6	0.7	438	95	Subtidal lagoon with more abundant allocthonous marine forms and reworked tests in beach/dune
					A. mamilla (5%)
	Cibicides spp.	31.0	57.0	6.0	Glabratella spp. (4%)
	C. lobatulus	9.6	19.5	2.0	C. gerthi (3%)
	C. refulgens	9.8	9.8	9.8	E. crispum (3%)
	E. macellum	6.0	13.7	0.7	E. pulvereum (3%)
	E. discoidalis	5.3	11.5	2.2	M. concentrica (3%)
					P. mediterranensis (3%)
					H. germanica (3%)
					Marine and brackish species (%<3%)
B	H. germanica	22.0	48.0	5.2	A. mamilla (5%)	19	5	0.3	131	31	Brackish subtidal, intertidal lagoon with allocthonous marine forms
	Cibicides spp.	19.2	40.3	6.0	E. crispum (4%)
	A. tepida	13.8	25.0	2.6	E. macellum (4%)
	Elphidium spp.	9.1	20.9	3.6	C. gerthi (3%)
	H. depressula	6.2	18.9	1.3	E. complanatum (3%)
	Haynesina spp. (juveniles)	6.0	6.0	6.0	E. excavatum (3%)
					Discorbis spp. (3%)
					Marine and brackish species (%<3%)
A	H. germanica	33.5	51.0	10.4	A. mamilla (5%)	22	7	2	68	26	Brackish subtidal to low intertidal lagoon with allocthonous marine forms
	A. tepida	19.9	28.7	12.4	C. gerthi (4%)
	Discorbis spp.	13.5	13.5	13.5	Elphidium spp. (4%)
	Cibicides spp.	10.7	13.5	8.0	E. excavatum (4%)
	Quinqueloculina spp.	7.3	7.3	7.3	E. discoidalis (4%)
	E. macellum	5.3	5.3	5.3	H. depressula (3%)
					Bolivina spp. (4%)
					Marine and brackish species (%<3%)

Units A to C are characterized by brackish to marine species and suggest subtidal to low-intertidal environments, the neatness of the marine signature increasing towards the top of unit C, as shown by the diversity index α Fisher and species richness (S) in Table 3 and data in Appendices SM1 and SM2. The textural contrast between unit C and the overlying units D and F is also evidenced in the foraminiferal contents: the low-energy sediments yielded 61% to 98% brackish intertidal forms, namely Haynesina germanica, Haynesina depressula and Haynesina spp. (juveniles) with less abundant Ammonia tepida, suggesting a pronounced decrease in salinity. Agglutinated high salt marsh species such as Jadammina macrescens (87.4%) and Miliammina fusca (9.6%) were only detected in one sample at the top of unit F.

Unit E again comes out as peculiar in the foraminiferal content (Table 3 and Figure 9), showing the widest diversity throughout the sedimentary column (total number of species in each core varying from 19 to 37 (mean=29) and number of Foraminifera per gram of dry sediment between 7 and 420; Table 3, Appendices SM1 and SM2, available online). This sedimentary unit is characterized predominantly by marine assemblages (Cibicides spp., Cibicides lobatulus, Cibicides pseudoungerianus, Elphidium macellum, Elphidium discoidalis, Elphidium crispum, Quinqueloculina spp., Mississippina concentrica and Asterigerinata mamilla) and by a considerable proportion of fully subtidal marine species in relation to estuarine species, 70% to 92%, the variation being explained by location within the lowland (i.e. increasing with proximity to the inlet). Some brackish forms were also detected (e.g. H. germanica and H. depressula) although in smaller proportions (5% to 28%) indicating reworking of the bottom. The dominance of marine Foraminifera throughout unit E clearly demonstrates that this horizon was deposited either as a result of a marine transgressive episode or a short-lived marine flood event.

Figure 9.

Percentage of open marine versus brackish benthic foraminifera in unit E and over- and underlying sediment in short cores from Lagoa dos Salgados. See Figure 3 for core location.

Exoscopic studies

The 181 quartz grains sampled from unit E are distinctive from the ones derived from other depositional environments (Figure 10 and Table 4). The statistical representativeness of grains was tested and results are presented in Appendix 1. Beach samples, with highest percentage of grains of the B type (Figure 10b), can be characterized by higher values for angularity and fresh surfaces, contrasting with lower values of adhering particles and dissolution. By contrast, nearshore samples present higher values of adhering particles and dissolution and almost no fresh surfaces. The grain type that is more frequent in nearshore samples is type A (Figure 10a). Dune samples present the widest range of features and a more balanced distribution in terms of the grain type. It is noticeable that the similarity in values of angularity and percussion marks found in grains from both dune and unit E sediment. The latter show almost no evidence of dissolution or adhering particles (Table 4) and exhibit high values of angularity, fresh surfaces and the higher concentrations of micro-percussion marks; moreover, the highest value of percussion marks was detected in core SG 14, close to the lagoonal inlet and where the thickness of the unit E is considerable. Furthermore, the grain types B and C (that associate with more dynamic processes) represent more than two-thirds of the grains of unit E, thus indicating that strong impacts between grains occurred and therefore demonstrating the high energy involved in their transport before deposition inland.

Figure 10.

Exoscopic images of quartz grains from the Lagoa dos Salgados. (a) Dune; (b) beach; (c) proximal nearshore; (d) distal nearshore; (e) unit E; (f) detail of percussion marks (approximately 12 μm in length) imprinted on the surface of a grain from unit E. Images (a), (b) and (e) are typical of grain types A, B and C, respectively.

Table 4.

Statistical comparison of microtextural features of grains from Lagoa dos Salgados (i.e. nearshore, beach, dune and unit E).

Sed env	Angularity	Fresh surfaces	Dissolution	Adhering particles	Percussion marks	% Type A	% Type B	% Type C
Beach	2.81	2.65	1.45	1.68	1.51	20.70	62.50	16.80
Dune	2.40	1.68	2.39	2.92	1.86	47.70	30.50	21.80
Nearshore	1.96	0.78	3.09	3.42	1.08	72.60	14.90	12.50
Unit E	2.45	2.31	1.04	1.31	2.04	30.50	43.60	25.90

Discussion

The Lagoa dos Salgados is actively accreting and this coastal depositional environment exhibits significant potential as a sedimentary archive of paleoenvironmental changes that occurred throughout the Holocene. In this case, the data described above suggest that the lagoonal infill accumulated essentially after 7000 cal. BP, in progressively lower-energy conditions, favoured by the shelter offered by the presence of a coastal barrier. Sediments of units A to C exhibit a neater marine signature when compared with the topmost deposits of units D and F, which correspond to terrestrial-sourced, mud-dominated sedimentation that persisted throughout the last c. 1500 years across the whole lowland. This later sedimentation pattern was disrupted by a unique event responsible for the deposition of unit E, a single laterally continuous and thin lamina of coarser material carrying abundant marine ecological proxies, which suggests a return to higher-energy conditions and marine-sourced sedimentation, affecting most of the lowland. Although the Lagoa dos Salgados is subject to occasional breaching of the barrier promoted by marine overwash, providing water and sediment exchange with the ocean, the textural signature of these events is, at present, as it was in the recent past, confined to the inlet and flood shoals, where marine sand reaches the top of the stratigraphic column. Further inland and at a short lateral distance from the channel, the top sedimentary unit is represented by mud of unit F.

Within the upper mud-dominated units D and F (global thickness of 0.70–1.90 m) the conspicuous medium sand lamina (unit E) occurs at c. 0.40 m below the surface and consists essentially of quartz sand with marine bivalve shell fragments and mud intraclasts (rip-up clasts). The thickness of this layer decreases from c. 0.80 m close to the barrier to a few millimetres c. 850 m at its landward limit, well beyond normal overwash, storm and tidal influence.

Mapping of this lamina suggests a wide, thin and fan-shaped feature, wedging out within the lagoonal basin, implying an exceptionally large inundation of marine origin. The erosive base together with the presence of rip-up clasts and grain size contrast with the framing sediment favours an emplacement mechanism governed by an abrupt single high-energy marine inundation, rather than a positive increment in the sea level rise rate. The spatial distribution of characteristics of sediment from unit E and the inferred vectorial properties of flow and associated sediment dispersion path are compatible with a massive inflow of marine water and sediment. This is in clear contrast with the outflow that must have followed the inundation and apparently failed to leave a distinct sedimentary record. Moreover, the sedimentary signature of the inundation demonstrates that the inflow path of water and sediment was essentially conveyed through the inlet rather than extensive overtopping of the barrier, thus providing an objective limit to the maximum elevation of the water level at the time of inundation.

The linear extrapolation of sedimentation rates derived from ²¹⁰Pb and ¹³⁷Cs concentration profiles to the contact separating units E and F indicates that the resumption of muddy sedimentation following the emplacement of unit E is more recent than c. ad 1865.

Sediment of unit E presents essentially a unimodal grain size distribution, which suggests predominance of one sedimentary source. Bahlburg and Weiss (2007) analysed the grain-size distributions of tsunami materials deposited by the 2004 tsunami in the Indian and Kenyan coasts. Those deposits were mostly characterized by symmetrical or slightly positively skewed size distributions and this was interpreted as indicating that the tsunami sediment was entrained from well sorted pre-tsunami materials available in the nearshore and beach environments. In the case of Lagoa dos Salgados, a similar pattern was detected in the size distribution of sediments from both unit E (Figures 6, 7 and 8) and dune/beach material from present-day analogues. The comparison of the median grain size of unit E sediments and of possible sources (Figure 8), supported by SEM studies (Table 4) of the same sediments, suggest that the coarser fraction of unit E was mainly derived from beach and dune sediments, whereas the underlying unit D may have contributed to the finer fractions. The beach comes out as a very likely source in terms of angularity and fresh surfaces of the grains, whereas dunes are suggested as a major source by the presence of percussion marks. Most probably both types of environments have contributed as primary and essential sources for materials washed inland and deposited as unit E. Sato et al. (1995) found that two Japanese tsunami deposits (1983 and 1993) were mainly sourced from the beach sediments using limited textural data. Moreover, Paris et al. (2007) used similar methods to suggest that in Lhok Nga (Banda Aceh, Indonesia) the 2004 Indian tsunami deposit may have nourished from the beach, although the shallow shelf, the coral reef, the lagoons and the dunes also provided sediment. In addition to the textural and exoscopic results presented here, inedited results obtained by the authors from distal and proximal nearshore, beach, dune and storm and tsunami deposits in a wide range of geographical areas in Portugal and elsewhere provide grounds to suggest that microtextures can be applied to relate tsunami deposit and possible source materials and also to distinguish storm and tsunami deposits. However, further exploration of this methodology and results clearly falls beyond the scope of this study.

As suggested by Goff et al. (1998, 2001) the shell richness is one of the attributes characterizing tsunami sediments, and a good marker of the high-energy involved in the sediment transport. In Lagoa dos Salgados, unit E is characterized by a sharp increase in the carbonate content associated with shells, mostly represented by fragments of thick-shelled marine bivalves.

Changes (increases) in marine microfauna (including Foraminifera) have been previously used as an indicator of marine invasions in tsunami studies (e.g. Clague et al., 1999, Goff et al., 1998, 2001; Hindson et al., 1998). Atwater and Moore (1992), Dominey-Howes (1996), Dominey-Howes et al. (1998) and Dominey-Howes et al. (2000) demonstrated that tsunami deposits contained a wide variety of benthic Foraminifera species. In the case of Lagoa dos Salgados, the benthic Foraminifera associations of unit E present an abrupt increase of the species richness and diversity index (α Fisher > 5, clearly indicating a marine environment) as well as a strong percentage increase of the relative proportions of open marine/brackish forms (e.g. Cibicides spp., Elphidium macellum, Elphidium discoidalis, Elphidium crispum, Quinqueloculina spp., Mississippina concentrica and Asterigerinata mamilla). This corroborates a marine source and is compatible with a short-lived but massive inundation of the lowland, as previously outlined elsewhere by Mamo et al. (2009).

The sediment grain size in unit E decreases inland and upwards (typically within one φ interval). A pronounced contrast in grain size is detectable at the very base of unit E, when compared with the underlying materials, and further upward the sediment is virtually homogeneous in size or normally graded. In numerous places the incorporation of rip-up clasts within unit E is in agreement with the abrupt and high energy character of an inundation with a strong erosive potential, which is also in line with the unconformable/erosive character of the basal contact. This erosion may have been considerable since radiocarbon dates of the top of unit D are at least 800 cal. years older than the age inferred for the base of unit F. These observations are in agreement with the findings of Minoura and Nakaya (1991), Bondevik et al. (1997), Nanayama et al. (2000) and Dawson and Stewart (2007) that correlate the presence of intraclasts ripped from soft underlying units with the erosive capacity associated with a tsunami inundation.

In Martinhal and Boca do Rio (Figure 1), sedimentary deposits associated with the ad 1755 tsunami have been previously described and discussed (e.g. Andrade et al., 1997; Dawson et al., 1995; Hindson and Andrade, 1999; Hindson et al., 1996; Kortekaas and Dawson, 2007). According to Kortekaas and Dawson (2007), in Martinhal, the tsunami breached the barrier and laid down an extensive sheet of sand, differentiated by its larger extension inland from other sand units present in the stratigraphic column, representing either the permanent sedimentation regime or storm washovers. The tsunami origin for this deposit was supported by a varied suite of attributes, such as presence of rip-up clasts from the underlying layer, basal erosive contact and contrast between increased content of marine Foraminifera in the tsunamigenic unit and the essentially brackish assemblages yielded by the under- and overlying layers. In the case of Boca do Rio, according to Hindson and Andrade (1999), the same inundation event deposited a sedimentary horizon marking a distinct sedimentological and micropalaeontological break from the enclosing deposits. This tsunami layer consists of several distinct subunits that varied greatly in both vertical and horizontal sections. The highly variable nature of the horizon appears to be due to rapid variation in the hydrodynamic characteristics of the depositional event and diversity in source materials. The presence of upward fining sequences and of landward thinning and grading, as well as the erosional contact with the underlying lagoonal/estuarine mud, strongly suggested a tsunamigenic origin. Furthermore, paleontological and dating data also supported the reasoning for a deposition of this horizon associated with the ad 1755 tsunami. Although the documentary record indicates that the Algarve coast has been affected by other tsunamis in the last two millennia (Baptista and Miranda, 2009), no clear sedimentary record of their activity has been found, until present. If the time window is extended further back, a single reference to a peculiar layer, inferred as tsunamigenic and dated from c. 6th century bc, was described by Schneider et al. (2009) in a central Algarve lowland.

The lack of correspondence between historical and sedimentary record of tsunami events in the late Holocene can be attributed to: (a) eventual predominance of erosional patterns, as discussed by Szczucinski (2011) in the case of the 2004 Indian Ocean tsunami, thus inhibiting the preservation and deposition of tsunamigenic signature; (b) lack of textural contrast hindering distinction between tsunami and framing sediments; (c) tsunami waves incapable of overtopping dune fields, thus limiting their sedimentary signature further inland of the coastline.

This study shows that the sedimentary signature recognized in the western Algarve as produced by the ad 1755 tsunami extended further east. In fact, unit E from Lagoa dos Salgados shares many stratigraphic and chronological characteristics with the nearby deposits detected in Boca do Rio and Martinhal, thus suggesting a regional and synchronous event as responsible for its deposition. This conclusion is in agreement with the documentary record of this particular inundation in the Algarve coast. Considering all these lines of evidence we can infer that unit E was deposited by an exceptional abrupt marine invasion, most probably the ad 1755 tsunami. This hypothesis is supported by the sedimentological, palaeontological, exoscopic and chronological data. The spatial characteristics of the tsunami layer at Lagoa dos Salgados suggests that the barrier was effective in preventing extensive overtopping by the waves and this provides grounds to infer the maximum elevation of the sea surface in this coastal stretch of c. 10 m at the time of inundation.

The uniqueness of this record within the top low-energy sedimentary sequence lasting for about 1500 years might be used, with caution, to extrapolate a recurrence interval for high magnitude tsunami-borne inundations of this coast, suggesting that such interval may be of millennial scale.

Conclusions

A peculiar deposit (unit E) detected in the late-Holocene sedimentary record of the Lagoa dos Salgados is characterized by a group of sedimentological features that are in agreement with many of the known published criteria to identify palaeotsunamis. It contrasts with the under- and overlying layers in being represented by a horizon of coarser sediment sandwiched between finer-grained sediments. It rests with unconformity on the underlying layer with an erosive contact and ablation of the top of the lower sediments. The Foraminifera content indicates a dominance of marine species throughout the sedimentary unit E, including open marine forms, and exhibits high diversity index values (α Fisher) characteristic of marine environment, in contrast with assemblages from framing sediments. The exoscopic data point to unusual percussion marks scattered across individual grain surfaces and considered as indicative of multiple grain collisions during tsunami wave transport. Both microtextural and grain-size data suggest that most of the sediment in the tsunami deposit was sourced from adjacent beach and dune materials with smaller contribution from eroded underlying sediments. The inferred age of the sediment unit is consistent with the late 18th century and all the above mentioned features indicate the ad 1 November 1755 tsunami as the most suitable candidate for the deposition of unit E within the top stratigraphic sequence of Lagoa dos Salgados.

The spatial distribution of textural and geometric attributes of unit E suggests that the tsunamigenic waves and inundation entered the lagoonal space through the inlet instead of overtopping the dune field, thus constraining to about 10 m the maximum tsunami height at this coastal stretch. The sedimentological data preserved a distinct signature of the inflow but failed to register the subsequent outflow.

Furthermore, the singular character of this record, in Lagoa dos Salgados and in nearby Martinhal and Boca do Rio lowlands, suggests that the recurrence interval of similar high-magnitude events in the Algarve coast maybe of millennial timescale.

Footnotes

Appendix 1

Appendix 2 Acknowledgements

The authors are grateful to SB Teixeira (ARH-Algarve) for providing the nearshore samples and oblique aerial view in , to S Leroy (Brunel University, UK) for providing the Livingstone corer, and to all GeoFCUL researchers, postgraduate and undergraduate students who helped during the demanding coring campaigns. O Voldoire and J-M Hénot provided invaluable assistance in the SEM facilities in Clermont-Ferrand. The reviews by R Paris and D Dominey-Howes provided relevant contributions to improve this paper.

Funding

This work was supported by projects FCT – GETS (PTDC/CTEGEX/65948/2006) and EU – NEAREST (EU-037110-GOCE-2006). P Costa and A Oliveira benefited from PhD grants from FCT.

References

Abrantes

Alt-Epping

Lebreiro

. (2008) Sedimentological record of tsunamis on shallow-shelf areas: The case of the 1969 AD and 1755 AD tsunamis on the Portuguese Shelf off Lisbon. Marine Geology 249: 283–293.

Abrantes

Lebreiro

Rodrigues

. (2005) Shallow-marine sediment cores record climate variability and earthquake activity off Lisbon (Portugal) for the last 2000 years. Quaternary Science Reviews 24: 2477–2494.

Albertão

Martins

(1996) A possible tsunami deposit at the Cretaceous–Tertiary boundary in Pernambuco, northeastern Brazil. Sedimentary Geology 104: 189–201.

Andrade

(1992) Tsunami generated forms in the Algarve Barrier Islands (South Portugal). Science of Tsunami Hazards 10(1): 21–34.

Andrade

Kortekaas

. (1997) Sedimentological traces of tsunamigenic overwash of the Martinhal lowland (Western Algarve – Portugal). Proceedings Seminário da Zona Costeira do Algarve. Faro, 10–12 July 1997, Eurocoast-Portugal, pp. 11–18.

Andrade

Freitas

Miranda

. (2003) Recognizing possible tsunami sediments in the ultradissipative environment of the Tagus estuary (Portugal). Coastal Sediments’03 – The 5th International Symposium on Coastal Engineering and Science of Coastal Sediment Processes. 18–23 May, Clearwater Beach FL, CD-ROM, 14 pp.

Appleby

Oldfield

(1983) The assessment of ²¹⁰Pb data from sites with varying sediment accumulation rates. Hydrobiologia 103: 29–35.

Atwater

(1986) Evidence for Great Holocene earthquakes along the outer coast of Washington State. Science of Tsunami Hazards 236: 942–944.

Atwater

Moore

(1992) Tsunami about 1000 years ago in Puget Sound, Washington. Science 258: 1614–1617.

10.

Bahlburg

Weiss

(2007) Sedimentology of the December 26, 2004, Sumatra Tsunami deposits in intermediate and large distances, Eastern India (Tamil Nadu) and Kenya. International Journal of Earth Sciences 96: 1195–1209.

11.

Baptista

Miranda

(2009) Revision of the Portuguese catalogue of tsunamis. Natural Hazards Earth System Science 9: 25–42.

12.

Bengtsson

Enell

(1986) Chemical analysis. In: Berglund

(ed.) Handbook of Holocene Palaeoecology and Palaeohydrology. Chichester: John Wiley & Sons Ltd, pp. 423–451.

13.

Bondevik

Svendsen

Mangerud

(1997) Tsunami sedimentary facies deposited by the Storegga tsunami in shallow marine basins and coastal lakes, western Norway. Sedimentology 44: 1115–1131.

14.

Carvalho

(1998) GranGraf v2B. Grain-size software developed at Science Faculty, University of Lisbon, Portugal.

15.

Chagué-Goff

(2010) Chemical signatures of palaeotsunamis: A forgotten proxy? Marine Geology 271(1–2): 67–71.

16.

Clague

Bobrowsky

Hutchinson

(2000) A review of geological records of large tsunamis at Vancouver Island, British Columbia, and implications for hazard. Quaternary Science Reviews 19: 849–863.

17.

Clague

Hutchinson

Mathewes

. (1999) Evidence for late Holocene tsunamis at Catala Lake, British Columbia. Journal of Coastal Research 151: 45–60.

18.

Corrochano

Merino

Barba

. (2000) Los sedimentos del canal de Mira (Aveiro, Portugal): Propriedades texturales, procedencia y modelo paleogeografico. Studia Geologica Salmanticensia 36: 143–160.

19.

Costa

Andrade

Seabra

. (2005) 1755 – Terramoto no Algarve. Faro: Centro Ciência Viva do Algarve, 237 pp.

20.

Costa

PJM

Andrade

Freitas

. (2009) Preliminary results of exoscopic analysis of quartz grains deposited by a palaeotsunami in Salgados lowland (Algarve, Portugal). Journal of Coastal Research 56: 39–43.

21.

Costa

PJM

Andrade

Freitas

. (2011) Boulder deposition by extreme marine events. Earth Surface Processes and Landforms DOI: 10.1002/esp.222810.1002/esp.2228.

22.

Davis

(1986) Statistics and Data Analysis in Geology. New York: Wiley, 646 pp.

23.

Dawson

(1994) Geomorphological effects of tsunami run-up and backwash. Geomorphology 10: 83–94.

24.

Dawson

Stewart

(2007) Tsunami deposits in the geological record. Sedimentary Geology 200: 166–183.

25.

Dawson

Hindson

Andrade

. (1995) Tsunami sedimentation associated with the Lisbon earthquake of 1 November AD 1755: Boca do Rio, Algarve, Portugal. The Holocene 5(2): 209–215.

26.

Dawson

Long

Smith

(1988) The Storegga Slides: Evidence from eastern Scotland for a possible tsunami. Marine Geology 82: 271–276.

27.

Dawson

Smith

Ruffman

. (1996) The diatom biostratigraphy of tsunami sediments: Examples from recent and middle Holocene events. Physics and Chemistry of the Earth 21: 87–92.

28.

Dominey-Howes

DTM

(1996) The geomorphology and sedimentology of five tsunamis in the Aegean Sea region, Greece. Ph.D. Thesis, Coventry University.

29.

Dominey-Howes

Cundy

Croudace

(2000) High energy marine flood deposits on Astypalaea Island, Greece: Possible evidence for the AD1956 southern Aegean tsunami. Marine Geology 163: 352–361.

30.

Dominey-Howes

Dawson

Smith

(1998) Late Holocene coastal tectonics at Falasarna, western Crete (Greece): A sedimentary contribution. In: Stewart

Vita-Finzi

(eds) Coastal Tectonics. Geological Society London, Special Publication 146, pp. 341–350.

31.

Eijkelkamp (2008) Operating Instructions: 08.53 Calcimeter. Eijkelkamp, 6 pp.

32.

Esaguy

(1984) Ria de Faro – Barra da Armona – Evolução 1873–1983. Lisboa: Direcção Geral de Portos, 5 pp.

33.

Fatela

Moreno

Antunes

(2007) Salinity influence on Foraminiferal tidal marsh assemblages of NW Portugal: An anthropogenic constraint? Thalassas 23(1): 51–63.

34.

Folk

Ward

(1957) Brazos river bar: A study of significant of grain size parameters. Journal of Sedimentary Petrology 27: 3–26.

35.

Freitas

Andrade

Rocha

. (2003) Lateglacial and Holocene environmental changes in Portuguese coastal lagoons: 1. The sedimentological and geochemical records of the Santo André coastal area (SW Portugal). The Holocene 13: 433–446.

36.

Friedman

(1961) Distinction between dune, beach and river sands from their textural characteristics. Journal of Sedimentary Petrology 31(2): 514–529.

37.

Gelfenbaum

Jaffe

(2003) Erosion and sedimentation from the 17 July, 1998 Papua New Guinea tsunami. Pure and Applied Geophysics 160: 1969–1999.

38.

Goff

Chague-Goff

Nichol

(2001) Palaeotsunami deposits: A New Zealand perspective. Sedimentary Geology 143: 1–6.

39.

Goff

Crozier

Sutherland

. (1998) Possible tsunami deposits from the 1855 earthquake, North Island, New Zealand. Geological Society Special Publication London 146: 353–376.

40.

Hindson

Andrade

(1999) Sedimentation and hydrodynamic processes associated with the tsunami generated by the 1755 Lisbon earthquake. Quaternary International 56: 27–38.

41.

Hindson

Andrade

Dawson

(1996) Sedimentary processes associated with the tsunami generated by the 1755 Lisbon earthquake on the Algarve coast, Portugal. Physics and Chemistry of The Earth 21: 57–63.

42.

Hindson

Andrade

Parish

(1998) A microfaunal and sedimentary record of environmental change within the late Holocene sediments of Boca do Rio Algarve, Portugal. Geologie en Mijnbouw 77: 311–321.

43.

Jouanneau

J-M

Weber

Champilou

. (2008) Recent sedimentary study of the shelf of the Basque country. Journal of Marine Systems 72: 397–406.

44.

Kortekaas

Dawson

(2007) Distinguishing tsunami and storm deposits: An example from Martinhal, SW Portugal. Sedimentary Geology 200(3–4): 208–221.

45.

Lopes

JBL

(1841) Corografia ou Memória Económica, Estatística e Topográfica do Reino do Algarve. Lisboa, 528 pp.

46.

Lowe

Walker

MJJC

(1997) Reconstructing Quaternary Environments. 2nd edition, London: Longman Group Ltd, 446 pp.

47.

Luque

Lario

Zazo

. (2001) Tsunami deposits as paleoseismic indicators: Examples from Spanish coast. Acta Geologica Hispanica 36: 197–211.

48.

Mahaney

(2002) Atlas of Sand Grain Surface Textures and Applications. Oxford University Press, 237 pp.

49.

Mamo

Strotz

Dominey-Howes

(2009) Tsunami deposits and their foraminiferal assemblages. Earth Science Reviews 96: 263–278.

50.

Minoura

Nakaya

(1991) Traces of tsunami preserved in inter-tidal lacustrine and marsh deposits – Some examples from northeast Japan. Journal of Geology 99: 265–287.

51.

Minoura

Nakaya

Uchida

(1994) Tsunami deposits in a lacustrine sequence of the Sanriku coast, northeast Japan. Sedimentary Geology 89: 25–31.

52.

Moore

Nishimura

Gelfenbaum

. (2006) Sedimentary deposits of the 26 December 2004 tsunami on the northwest coast of Aceh, Indonesia. Earth Planets Space 58: 253–258.

53.

Moore

(1988) Large scale bedforms in bolder gravel produced by giant waves in Hawaii. Special Paper Geological Society of America 229: 101–109.

54.

Morales

Borrego

San Miguel

. (2008) Sedimentary record of recent tsunamis in the Huelva Estuary (southwestern Spain). Quaternary Science Reviews 27: 734–746.

55.

Moura

Veiga-Pires

Albardeiro

. (2007) Holocene sea level fluctuations and coastal evolution in the central Algarve (southern Portugal). Marine Geology 237: 127–142.

56.

Nanayama

Shigeno

Satake

. (2000) Sedimentary differences between the 1993 Hokkaido-nansei-oki tsunami and the 1959 Miyakojima typhoon at Taisei, southwestern Hokkaido, northern Japan. Sedimentary Geology 135: 255–264.

57.

Oliveira

(2009) Influência da Geomorfologia local na preservação de assinaturas sedimentares de eventos de alta energia no Algarve Ocidental. Msc Thesis, Science Faculty, University of Lisbon, 272 pp.

58.

Paris

Fournier

Poizot

. (2010) Boulder and fine sediment transport and deposition by the 2004 tsunami in Lhok Nga (western Banda Aceh, Sumatra, Indonesia): A coupled offshore–onshore model. Marine Geology 268: 43–54.

59.

Paris

Lavigne

Wassmer

. (2007) Coastal sedimentation associated with the December 26, 2004 tsunami in Lhok Nga, west Banda Aceh (Sumatra, Indonesia). Marine Geology 238(1–4): 93–106.

60.

Paris

Wassmer

Sartohadi

. (2009) Tsunamis as geomorphic crises: Lessons from the December 26, 2004 tsunami in Lhok Nga, West Banda Aceh (Sumatra, Indonesia). Geomorphology 104(1–2): 59–72.

61.

Pereira

Soares

(1994) A estabilização holocénica do nível do mar. Vestígios no litoral de Armação de Pêra. Gaia 9: 91–93.

62.

Pinto

Teixeira

(2002) Morphodynamics of the sandy barrier of Salgados coastal lagoon, Armação de Pêra Bay (Algarve – Portugal). In: Eurocoast Portugal (ed.) Eurocoast Littoral 2002, The Changing Coast. Porto: Eurocoast/EUCC, pp. 403–409.

63.

Reicherter

Vonberg

Koster

. (2010) The sedimentary inventory of tsunamis along the southern Gulf of Cádiz (southwestern Spain). Zeitschrift für Geomorphologie, Supplementary Issues 54(3): 147–173(27).

64.

Ritchie

McHenry

(1990) Application of radioactive fallout Cesium-137 for measuring soil erosion and sediment accumulation rates and patterns: A review. Journal of Environmental Quality 19: 215–233.

65.

Ruiz

Rodríguez-Ramírez

Cáceres

. (2005) Evidence of high-energy events in the geological record: Mid-holocene evolution of the southwestern Doñana National Park (SW Spain). Palaeogeography, Palaeoclimatology, Palaeoecology 229: 212–229.

66.

Sato

Shimamoto

Tsutsumi

. (1995) Onshore tsunami deposits caused by the 1993 southwest Hokkaido and 1983 Japan Sea earthquakes. Pure and Applied Geophysics 144: 693–717.

67.

Scheffers

Kelletat

(2005) Tsunami relics on the coastal landscape west of Lisbon, Portugal. Science of Tsunami Hazards 23(1): 3–16.

68.

Schneider

Höfer

Trog

. (2009) Holocene estuary development in the Algarve Region (Southern Portugal) – A reconstruction of sedimentological and ecological evolution. Quaternary International 221(1–2): 141–158.

69.

Sousa

FLP

(1919) O terremoto do 1º de Novembro de 1755 em Portugal e um estudo demográfico: Distritos de Faro, Beja e Évora, Lisboa. Tipografia do Comércio, 277 pp.

70.

Stuiver

Reimer

(1993) Extended ¹⁴C database and revised CALIB radiocarbon program. Radiocarbon 35: 215–230.

71.

Switzer

Jones

(2008) Large-scale washover sedimentation in a freshwater lagoon from the southeast Australian coast: Sea-level change, tsunami or exceptionally large storm? The Holocene 18(5): 787–803.

72.

Szczucinski

(2011) The post-depositional changes of the onshore 2004 tsunami deposits on the Andaman Sea coast of Thailand. Natural Hazards: DOI 10.1007/s11069-011-9956-810.1007/s11069-011-9956-8.

73.

Taborda

Dias

(1992) Análise da sobreelevação de origem meteorológica durante os temporais de Fevereiro/Março de 1978 e Dezembro de 1981. Geonovas Número Especial 1 ‘A Geologia e o Ambiente’, pp. 89–97.

74.

Teixeira

Gaspar

Rosa

(2005) Holocene sea-level index points on the Quarteira coast (Algarve, Portugal). In: Freitas

Drago

(eds) Proceedings of Iberian Coastal Holocene Paleoenvironmental Evolution, Coastal Hope 2005. Lisbon, pp. 125–127.

75.

Vis

G-J

(2009) Fluvial and marine sedimentation at a passive continental margin: The late Quaternary Tagus depositional system. PhD thesis. ISBN 978 90 8659 388 0. Faculty of Earth and Life Sciences, University of Amsterdam.

76.

Whelan

Kelletat

(2005) Boulder deposits on the Southern Spanish Atlantic Coast: Possible evidence for the 1775 AD Lisbon Tsunami. Science of Tsunami Hazards 23: 25–38.

A tsunami record in the sedimentary archive of the central Algarve coast,Portugal: Characterizing sediment,reconstructing sources and inundation paths

Abstract

Keywords

Introduction

Study site

Methods

Results

Radiocarbon dating, 210Pb and 137Cs determination and sedimentation rates

Lithostratigraphy

Unit A

Unit B

Unit C

Unit D

Unit E

Unit F

Summary

Grain-size characteristics of unit E

Micropalaeontological analysis

Exoscopic studies

Discussion

Conclusions

Footnotes

Appendix 1

Appendix 2

Acknowledgements

Funding

References

Radiocarbon dating, ²¹⁰Pb and ¹³⁷Cs determination and sedimentation rates