Probable late Messinian tsunamiites near Monte Dei Corvi, Italy, and the Nijar Basin, Spain: expected architecture of offshore tsunami deposits
Three distinct, 30- to 80-cm-thick, graded, multilayered, coarse-grained sandstone layers, intercalated in the late Messinian mudstones of the Colombacci formation in Lago Mare facies of the Trave section are interpreted as tsunamiites (Ts1–Ts3). Each of these layers is sheet-like and could be followed along strike over several tens of meters. The lower two layers (Ts1–Ts2) occur in the lower part of the Colombacci formation and the third (Ts3) just below a conspicuous white “colombacci” limestone near the top of the formation. The three sandstone layers represent unique sedimentary events within the 120-m-thick San Donato-Colombacci mudstones, which contain many thin, fine-grained, possibly storm-related turbidites. Each of the three clastic layers is overall graded and strongly cross-bedded. A single layer consists of a stack of several graded sublayers that are eroded into the underlying mudstones and into each other. Absence of hummocky cross-stratification (HCS) indicates that the layers are not produced during a large, catastrophic storm event. Current ripples such as dm-sized trough cross-beds suggest strong, prolonged, unidirectional currents, capable of carrying coarse conglomeratic sands. Climbing ripples in middle-fine sand units indicate a high suspension load settling under waning current strength. Each of the Ts1–Ts3 beds satisfies a combination of criteria, described in this paper, that allow interpretation as a tsunamiite in an offshore environment. Tsunamiite Ts2 is underlain by a 15-cm-thick meshwork of synsedimentary fissures, filled with coarse sand. Ground movements induced by strong earthquakes probably caused these crevasses. The uniqueness of each layer, the erosion of the base of each of the sublayers into underlying mudstones and previously deposited sublayer and the consistent stacking of graded sandstone beds within each of the three layers, underlain by earthquake-produced fissures, strongly point to deposition by traction currents produced by the surges of a large tsunami event, triggered by a large vertical fault movements. Vertical fault displacements most likely occurred along the thrust faults like the Sibilline thrust at the SW of the Laga foreland basin, which were active at late Messinian times. A series of cyclic graded turbidites, underlain by seismically induced sand-filled fissures in the Late Messinian Feos formation in SE Spain, are interpreted as tsunamiites produced by a tsunami or seiche.
KeywordsPaleo-tsunami deposits Tsunamiite Lago Mare Messinian
The recognition and detection of tsunami deposits in the geological record could be an important tool for the evaluation of the frequency and magnitude of past tsunami activity. Tsunamis are invariably triggered by an energetic (convulsive) event. Often-cited ‘triggers’ include subduction-related earthquakes (Sumatra 2004, Japan 2011), landslides (Dawson and Steward 2007), submarine landslides (Storegga) (Dawson et al. 1988), volcanic eruptions of Santorini, (Cita 2008), Krakatoa (Van den Berg et al. 2003) and large extraterrestrial body impacts in and near ocean basins (Bourgeois et al. 1988; Smit et al. 1996). Sedimentary deposits directly or indirectly caused by tsunamis are known as tsunamites or tsunamiites (Shiki and Yamazaki 2008). We prefer here to use the latter term.
Tsunami deposits are best known from modern onshore, intra-tidal to supra-tidal areas. Most tsunami research is centered in these environments, and these are relevant for the evaluation of tsunami activity in recent and historical context. However, such areas are prone to erosion (Dawson and Steward 2007; Bourgeois et al. 1988). Specifically, those regions near subduction-related earthquakes, where most supra-tidal tsunami deposits can be expected, are on uplifting rather than subsiding crust, facilitating erosion. Such supra-tidal areas are poorly suited for the research of tsunami activity in the more distant geological record.
The deeper (below base level) offshore marine environment offers a much better potential for the preservation of tsunamiites in the geological record. Deep-water (up to 4,000 m) tsunamiites are thus far known to result from the Chicxulub impact in the western Atlantic basin (Norris et al. 2000) and the catastrophic Holocene eruption of the Santorini (Thera) volcano (Shiki and Cita 2008; Cita 2008). The latter deposits are known as ‘homogenites’. Cita (2008) distinguished a type ‘A’ homogenite in small basins on the Mediterranean ridge and a type ‘B’ homogenite that is basically a megaturbidite. Homogenite ‘A’ is a pelagic turbidite consisting of oozes remobilized by the shuttle movements of the ‘shallow-water tsunami wave’ on the seafloor (Sugawara et al. 2008) and tend to accumulate as pelagic sediment, muddy gravity flows into flat-floored depressions close to the remobilized areas (Cita 2008; Shiki and Cita 2008). Such ‘shallow-water waves’ are characterized by very long wavelengths—about >20 times longer—compared to the depth of the ocean and a very small (<1–2 m) amplitude. The horizontal water movements of such long waves are the same throughout the water column (Sugawara et al. 2008) and are normally not strong enough to remobilize sea floor oozes. However, the ‘Bernoulli effect’ over the crest of the Mediterranean ridge may have caused the necessary increase in current traction on the seafloor to remobilize the oozes in this specific case. The type ‘B’ homogenite is probably triggered by the backflow(s) of the tsunami (Cita 2008) from shallow waters in the Sirte Gulf in view of the abundance of shallow-water fossils. The sediment-laden high-density backflow developed downslope into something akin to a megaturbidite deposit (Mulder et al. 2009). Yet, characteristic sedimentary features that unequivocally (without evidence tying the deposits to the Santorini eruption) define it as a tsunamiite are not expected from such gravity flow deposits, beyond a few possible grain size ‘jumps’.
Most of the known suspected offshore tsunamite deposits in the geological record are related to large impact events. Coarse, graded, sandstone deposit are associated with the Chicxulub crater at the K/T boundary around the Gulf of Mexico (Bourgeois et al. 1988; Smit et al. 1996), the Manson crater in the late Cretaceous, the Lockne crater in the mid-Ordovician (Izett et al. 1993; Sturkell et al. 2000), and the early Proterozoic tsunamiites with impact spherules in the Hamersley Basin in Australia (Hassler et al. 2000).
Where to look for possible and recognizable tsunamiites? The frequency of large earthquake-triggered tsunamis is unquestionably high at the continental shelf near subduction zones. Passive margins are usually farther away from the source. The effective depth, i.e., the depth where tsunami deposits can be recognized as such, of tsunami-influenced sediment will diminish with distance and may be well in reach of the wave base of large storms and be often obliterated by the reworking effects of the storm waves, changing its texture into those of tempestites. If a typical tsunamiite, standing out from a tempestite, is discovered on distal continental margins, it is expected that these in majority may be triggered by large-body impact. It is questionable whether tsunamiites triggered by large subduction-related earthquakes will be found—or may be extremely rare—in passive margin settings, because the passive margin is often far away from the source, except in smaller basins like the Mediterranean. Subduction-related tsunami waves may have lost so much energy that its effects can be difficult to distinguish from those of tidal currents.
2 Geological setting
The upper part of the Schlier fm, ‘Formazione di letto’, or euxinic Shale unit (Montanari et al. 1997), consists of ~15-m-thick, fully marine pelagic sediments displaying rhythmic alternations of dm-thick sapropels and bioturbated marly limestones, interpreted as orbitally forced, mainly precession, cycles (Hilgen et al. 2007a; Husing et al. 2007). The euxinic shales are the pre-evaporite prelude to the MSC, and the increase in thickness of the black, organic C-rich sapropel interbeds probably indicates the increasing closure of the Mediterranean Basin.
The overlying interval is the ‘Lower Evaporites’ or Gessoso-Solfifera fm. The sapropels of the euxinic shales increase in thickness and organic carbon content toward the ~30-m-thick evaporitic Gessoso-solfifera fm, which consists of 16/17, 100- to 180-cm-thick couplets of laminated euxinic shales and laminated primary gypsum layers. Both layers, but mostly the gypsum beds, contain syn- to post-sedimentary, ingrown, tree-like, selenite gypsum crystal aggregates, up to 80 cm in size. The interval of interbedded laminated black shales and gypsum layers is almost identical to those of the lower evaporite of the Yesares fm of the Sorbas Basin (Krijgsman et al. 2002). Also, these cycles are probably orbitally forced [precession (Hilgen et al. 2007b)]. The contact between the formations is unfortunately not exposed in the Trave section. Twelve of the (precession) gypsum cycles of the presumably 16 (Manzi et al. 2007) could be counted in the adjacent sections 1.3 km toward the SE on the Mezzavalle beach. Both the laminated euxinic shale and the laminated gypsum layers contain many whole specimens of the small, thick-boned, extreme environment–tolerant (high salinity) fish Aphanius crassicaudus, (Reichenbacher and Kowalke 2009). The next interval consists of the san Donato and Colombacci formations, described below. The highest unit of the Trave section is the Trave fm, a 15-m-thick, medium-bedded (20–25 cm), shore-face, and bioturbated medium-to-fine-grained sandstone. The Trave fm marks the final shallowing of the Messinian succession, although (Roveri et al. 2005) suggest a lowermost Pliocene age for the Trave sandstone. The Trave is resistant to erosion and forms a hard ledge—a ‘reef-like’ feature, the “Scoglio del Trave” in the sea (Fig. 4). The Trave contains in situ double-valved brackish water Limnocardium and sedimentary structures including hummocky cross-stratification (HCS) in less-bioturbated levels, in particular near the top. The Trave displays regular rhythmic, cyclic bedding, marked by thin, slightly softer, finer-grained interbeds. The cycles indicate regular changes in storm and/or bioturbation intensity, probably related to cyclic climatic changes. The ~75 rhythmic beds are deposited in less than 80 Ky (Roveri et al. 2005), which suggest cyclicity on a millennial scale (J.S. unpublished data). These are in the range of the ‘Bond cycles’ (Bond et al. 1997). A bored hardground at the top forms the transgressive surface of the fully marine lower Pliocene.
2.1 Upper Messinian San Donato and Colombacci formation
The 92-m-thick San Donato formation (also named Tetto Fm) [P-ev1 of Roveri et al. (2005)] dominantly consist of monotonous silty mudstones with interbedded thin turbidites, in particular in the basal 50 m. A 10-m-thick package in the middle of the formation is slumped. The San Donato fm represents the post-evaporitic Lago Mare phase and has been regarded as the clastic lateral equivalent of the ‘Upper evaporites’ (Roveri and Manzi 2006). Most authors assume the great Messinian dessication of the Mediterranean to occur between the lower and upper evaporites, at the base of the San Donato fm. The San Donato fm abruptly overlies the last gypsum cycle, possibly marking the short 90-ky hiatus caused by the sea level drop (Krijgsman et al. 2002). This contact is not exposed in the Trave section but is well visible on Mezzavalle beach (Fig. 2).
The 76-m-thick Colombacci formation [P-ev2 of Roveri et al. (2005)] is a continuation of the silty mudstone lithology of the san Donato fm., but it contains three 10- to 95-cm-thick complex, multiple-graded, gravel-sand layers, (Ts1–Ts3) at 157, 158, and 217 m, (Fig. 5) described here. Roveri et al. (2005) place the boundary between (P-ev1/P-ev2) at the base of the lowest of the three sandstone layers. The Colombacci fm. in addition contains two about 1-m-thick, conspicuously white limestones (C1 and C2), the enigmatic “colombacci” layers, that contain marine nannofossils (Roveri et al. 2005; Popescu et al. 2007), possibly indicating initial flooding events, a short prelude to the opening of the Mediterranean to the Atlantic after the salinity crisis and before the major flooding at the base of the Pliocene (Popescu et al. 2007). The almost complete absence of microfossils in the san Donato/Colombacci mudstones precludes any reliable depth estimate of the Lago Mare Sea. However, the laminated, unsorted fine silty mudstones with intercalated thin turbidites, and the fine-grained pelagic nannofossil colombacci limestones indicate an offshore environment, below storm wave base.
2.2 Messinian Lago Mare Feos formation, Nijar Basin, Spain
The Feos formation is in the center of the basin >100 m thick and consists of coarse continental clastic layers with paleosols, and white laminated Lago Mare lagoonal to brackish fine-grained mudstones with nannofossil and fine gypsum debris, comparable to the “colombacci” layers in the Trave section. The Feos fm. represents a series of cyclic alternations of continental clastic and lagoonal-lacustrine Lago Mare facies, interpreted as repeated periods of desiccation and reflooding of the Nijar embayment on orbital precession timescales (Krijgsman et al. 2002). Regarding the small size and isolation of the northern Nijar Basin, the embayment may have had a shallow base level. Near the hamlet of los Feos (Fortuin and Dabrio 2008) describe an outcrop in the locally developed Lago Mare facies (labeled LLM) of characteristic micritic, stark white, finely laminated, lacustrine-lagoonal chalks, interbedded with frequent thin-to-medium-bedded, well-sorted, graded sand beds, interpreted as turbidites. The varve-like laminations in LLM are preserved due to the total absence of any bioturbation. Many ‘turbidites’ occur in sets of two to three regularly spaced, thinning upward graded sandstone layers and are here reinterpreted as tsunami event beds.
The Feos fm. is sharply overlain by the shallow marine calcarenites of the Pliocene Cuevas fm.
The lithology of the Trave section was logged in detail in the field (Fig. 5), and 250 samples were taken to perform a laser grain size analysis and thermogravimetric analysis (TGA) for carbonate content. The samples chosen for grain size analysis were mainly taken from the fine mudstones. Turbidites, except the coarse intervals, were avoided as much as possible. The interval between Ts2 and Ts3 could not be sampled because of landslides covering the outcrop. The mean grain size of the samples is coarse silt (phi scale 6) at the base of the section and constant fine silt of the upper part of the formation. The modal grain size gradually decreases in top of the san Donato, up till the first thick Ts1 Tsunamiite sandstone at the base of the Colombacci fm. The coarse grain size at the basal 30 m reflects the influx of thin turbidites. The slight decrease in modal grain size above the slumps coincides with the decrease in number of turbidites.
An aliquot of the crushed samples was further finely ground for TGA to measure carbonate contents, organic C content and water.
The TGA method used includes a stepwise loss on ignition (LOI) method, developed at the VU university (Konert 2010 in preparation). At specific temperature intervals, from low to high temperatures, successively water (<60–80°C), organic carbon (90–105°C), CaCO3a (155–190°) (aragonite), and CaCO3b (230–300°) (calcite) are lost (vaporized) and quantified.
4.1 Ts1 (tsunamiite 1)
4.2 Ts2 (tsunamiite 2)
The 5–10 cm top of the mudstones directly underlying Ts2 contains a meshwork of conspicuous dark, 5- to 15-cm-long, fissure-like features that are filled with coarse sand (Fig. 10). The fissures often bifurcate and are in cross section elongated, and sometimes ellipsoid. The fissures are often a few cm offset in cracks, demonstrating soft-sediment deformation just after their origin. The origin of the injected sand is probably related to the deposition of the basal coarse sand bed. Although the coarse layers are cross-bedded, reliable current directions could not be measured due to its lateral channel-fill characteristics. On perpendicular cross-cuts, trough infill migration is toward the SW.
4.3 Ts3 (tsunamiite 3)
4.4 Tsunamiites in the Feos formation, Nijar Basin, SE Spain
In the outcrop near Los Feos, the sand layers capping the conjugate sand-filled fissures are not a single graded layer but come in pairs and triplets (Fig. 13). The first sand layer is followed at regular distances by a second (and often by a third or fourth) thinner and finer-grained sand layer. The regular spacing and thinning/fining upward suggests that these series of sand layers are genetically linked and are indirectly induced by the same tectonic event.
5.1 Occurrence and expected sedimentary architecture of offshore tsunamiites
This contrast has profound implications for the resulting sedimentary deposits at shallow lower shore-face to offshore (10–100 m) water depths. Storm waves basically produce short-term, large oscillatory currents on the seafloor at this depth. The passage of a storm event will lead to a tempestite deposit, a single graded bed, characterized by erosion of the seafloor, subsequently filled with a lag deposit, followed—at waning wave energy—by a graded sandy deposit showing hummocky cross-stratification (HCS) indicative of oscillating currents on the seafloor (Fig. 15).
The arrival of a train of shallowing tsunami waves, which transfer the waves into short-duration pseudo-tidal surges, will ultimately result in a single stack of multilayered, overall graded deposit at shallow-to-deep offshore depths. Each of the surges produces unidirectional currents, either in onshore or in offshore direction, which differ fundamentally from the oscillation currents produced by storm waves. The first tsunami wave (surge) erodes into underlying soft offshore mudstones and should leave sole marks (flute-casts and tool marks) that can be very similar to those produced at the base of turbidites. Successive wave (surge) arrivals will erode into the top of the previously deposited surges. Due to waning strength of the tsunami waves, each following surge will have less energy and will be less capable of downcutting into the former surge layers and be finer-grained (Fig. 15).
As with homogenite ‘B’ (Cita 2008), the up-rush and in particular the backflow of the tsunami surge will erode a large amount of sediment and bring it into suspension. Because of its high density and supported by backflow currents, the sediment-laden suspension will move downslope as gravity flow. The following up-rush surge will slow down the gravity flow, leading to rapid sedimentation of the suspended sand. This phase is often represented by climbing-ripple sets. The next backflow surge of the next tsunami wave will initially be strong enough to erode down into the previously deposited layer of sediment and, when slowed down by the next up-rush surge, lead the next cycle of deposition of suspended sediments.
A complete offshore tsunami deposit should thus reflect the passage of a series of tsunami waves, in general waning in strength. Each of these waves—or surges—should reflect a sequence of erosion, transport, and deposition (Dawson and Shi 2000).
The expression (architecture) of the resulting tsunami deposits may be variable with water depth, wave height, or size, the confinement of the body of water where the tsunami occurred [a narrow lake or seaway, a confined sea (Mediterranean, Gulf of Mexico), bottom topography (Shiki and Cita 2008)], and availability of unconsolidated sediment. In general, tsunami surges carry a high sediment load, when substantial amounts of sediments were eroded from either the seafloor or coastal areas, as observed near Aceh Banda in Indonesia after the 2004 tsunami (Fig. 1), or the type 2 homogenites (turbidites) of the Santorini tsunami (Cita 2008; Shiki and Cita 2008).
The Chicxulub impact–related tsunamiites (Smit et al. 1996) in the Gulf of Mexico and the Proterozoic tsunamiites in the Hamersley Basin (Hassler et al. 2000) characteristically display rather thick climbing-ripple intervals often followed by a graded but visually homogenous silt–clay tail (Smit et al. 1996), attesting to this high sediment load.
Stands out from the sedimentary sequence above and below and should not be the culmination of a clearly observable sedimentary trend.
Should reflect the repeated high-energy input of recurring powerful currents.
Consists of a series of graded beds, each one reflecting passage of one of the tsunami surges. As the initial waves are the largest, the lowest beds will usually be the coarsest, although as the Indian Ocean tsunami learned (Fig. 15) that the second or third one is often the largest.
Each of the higher beds should in general indicate less current energy and show a thinning of beds and an overall decrease in grain size, but see 3.
Reflects the high sediment load in suspension in the water from eroded areas, although the end product is dependent on the characteristics of the eroded sediment (climbing ripples, high-density flows).
Should not display typical hummocky cross-stratification (HCS), although in the waning stages, short-term oscillatory movements in the water may occur.
May contain evidence of the triggering event, such as earthquake-induced soft-sediment deformation below, but not above the tsunamiite; volcanic ashes; impact ejecta.
Be synchronous in different environments when found over a large region.
May contain debris backwashed from back-barrier supra-tidal areas flooded by the tsunami. This may include abundant plant debris.
Could display a burrow fabric at the top, consisting initially of back-burrowing animals like crabs and shrimps, followed by a normal multitiered burrow fabric that is normal for the local seafloor. Back-burrowing animals may also penetrate into the top of the lower layers deposited by preceding tsunami surges because there is sufficient time for these [up to 1 h, see Fig. 3)], but not enough time to develop a normal, tiered burrow fabric.
It is unlikely that in any tsunamiite, all of the above criteria will be met, but the combination of a number of these may identify a tsunamiite. Note however that the above criteria in isolation do not provide proof of a tsunami deposit (Shanmugan 2011). Some of the above criteria are by itself also valid for gravity flows or turbidity currents or tempestites.
Tsunami wave deposits (tsunamiites) should be different from tempestites (Fig. 17) in many aspects. Because they can reach greater depths (depending on the distance from and magnitude of the triggering event), they stand out from storm events. The frequency of unusually large storm events [taken the data from hurricane deposits around the Gulf of Mexico (Donnelly and Woodruff 2007)] will be one catastrophic event every 10–250 years at a single location. In offshore sedimentary archives, prone to such storm events, tempestite deposits should therefore frequently occur, as long as the depositional area remains at a suitable depth. Such depositional intervals are known, e.g., from transgressive base of the Cretaceous Del Rio fm of the Western Interior Seaway in the United States (Lock et al. 2009), where the tempestites are stacked upon one another, and display characteristic large HCS.
The complex sandstone beds in numerous outcrops at the KPg boundary around the Gulf of Mexico are clear examples of tsunamiites and share many—and often all—of the above-indicated characteristics (Smit 1999; Smit et al. 1996; Smit 2010 in preparation). Striking features in these tsunamiites are thick (10–40 cm) backflow direction climbing-ripple intervals, displaying decreasing current strength upwards, just before the up-rush of the next incoming tsunami surge.
Shanmugan (2011) states that single sedimentologic criteria alone cannot distinguish between storm (hurricane) and tsunami event deposits. He summarized the criteria of reported tsunami-produced sediment beds in a wide variety of environments, but none of the tsunami-related sedimentary characteristics are diagnostic on its own, and he concluded that up till now, it is not possible to make a distinction between storm- and tsunami event–produced deposits. The wide variety of observed cross-beds [low-angle, HCS, climbing ripples, (Morton et al. 2007)], planar beds, and normal and inverse grading reported from tsunami beds are very similar to those of turbidites, mass flows or debris flows and tempestites. Erosional bases are also characteristic for tempestites and debris flows/turbidites. Anomalous coarse sand layers (event beds) are also reported from storm deposits. Yet, the vast majority of the described characteristics are from coastal, shore-face to coastal plain environments. Very little is reported from the (shallow) offshore environment.
The three graded coarse sandstone beds in the Messinian Trave succession are similar in texture, structure, and the composition of the sand grains, and it is therefore likely that they share a common origin. The three beds were earlier described by Roveri et al. (2005) as (op cit): “thin, coarse grained bioclastic sandstone bodies containing (brackish water) Limnocardium, having an erosional base, displaying normal and inverse grading, medium scale cross-beds at the base, sharply overlain by fine-grained sandstones with HCS”. These sandstone bodies are supposedly rhythmically alternating with thick mudstone bodies. Roveri et al. (2005) interpreted such strata as common in the shallow marine record and as “lobes deposited from high-density gravity flows generated by fluvial floods, or alternatively, by severe storm events”.
However, a rhythmic occurrence of the three coarse sandstone beds is far from obvious in the Trave section; their occurrence is anomalous and random. The angular clasts and admixture with Limnocardium fragments do not argue for a fluvial related origin, but for a local, coastal, source of these components. A high-density gravity flow is an acceptable explanation for some of the individual graded gravelly and poorly sorted sublayers. The grain-supported texture and high-angle, trough cross-bedding of other sublayers argues for deposition by traction currents. However, the atypical, multistoried, repeatedly graded cross-bedded stack of sublayers of all three beds argues for another origin of the layer as a whole. Roveri et al. (2005) support their fluvial origin interpretation by the interpretation of coarse sand-filled “root traces” just below the second sandstone bed Ts2, as an indication for emergence. However, as mentioned earlier, we interpret these “traces” not as tubular root traces, but more planar, seismically induced fractures and cracks, which allowed sand from the sand-laden tsunami-induced current or sand of the overlying sandstone bed deposited by the first tsunami surge to penetrate into those fissures as sedimentary dykes. A similar situation is discussed below for the Feos seismites/tsunamiites.
The alternative interpretation by Roveri et al. (2005) as severe storm deposit remains another possibility. The many thin, fine-grained turbidites interbedded in the lower part of the san Donato mudstones are most likely triggered by either distal storm events or distal earthquakes and indicate deposition in a rather deep offshore environment. However, the three Ts1–Ts3 beds are very different from those. The supporting evidence for storm deposit in Ts1–Ts3 is the interpretation of the cross-bedding in the upper part of the beds as HCS.
HCS is often difficult to establish, and controversial interpretations as storm deposits are based on improper interpretation of cross-beds as HCS (Fig. 9). Shanmugan (2011) warns that the origin of HCS is controversial. However, he allows for a broad interpretation of HCS in general. He cites (Mulder et al. 2009) who interpret anti-dune-like ripples as HCS-like structures in deep-water turbidites. Identical structures are found at the base of Chicxulub impact tsunamiites at la Lajilla, Mexico (Smit et al. 1996), but those are clearly not related to storm activity. At the K/Pg boundary at Brazos River, Texas, Smit et al. 1996 disagree about the interpretation as HCS of cross-bedding in the upper part of the K/Pg sandstone bed (Gale 2006; Yancey 1996). Cross sections perpendicular to the current direction of climbing-ripple sets were here misidentified as HCS. HCS is a sedimentary structure produced by large, but short-term oscillatory currents caused by passage of large waves. Very long tsunami waves produce longer-term oscillations, producing tidal-like, opposing direction currents that will produce different cross-bedding structures such as climbing ripples because of the high sediment suspension load of the currents.
Because an interpretation of sandstone beds as produced either by storm (tempestite) or tsunami event often hinges on proper identification of HCS, we urge for a careful interpretation of such cross-bedding structures.
The grain size analysis in the colombacci-san Donato mudstones demonstrates that the occurrence of the Ts1–Ts3 tsunamiites is not related to a trend in the under- and overlying offshore mudstones. This argues against a connection of Ts1–Ts3 to a gradual shift in facies, such as caused by climate-related shifts in base level as suggested by Roveri et al. (2005).
It is the combination of criteria that makes interpretation of the Ts1–Ts3 beds as tsunamiites likely: (1) the beds stand alone in the mudstone sequence, without preceding facies shifts toward the sandstone beds. (2) They are composed of a stack of graded and cross-bedded sublayers, each eroding into the preceding one and each connected to an individual tsunami wave surge within the tsunami event. (3) They display climbing-ripple sets rather than HCS in the top of each sublayer, in particular, near the end of the backflow stage when suspension-laden tsunami currents lose energy because of the next tsunami wave up-surge. (4) They are linked to earthquake-produced fissured in the seafloor, indicating the triggering event of the tsunami event. The architecture of Ts1–Ts3 satisfies a good number of the above-mentioned criteria and leads us to the conclusion that all three beds are probably offshore tsunamiites.
The graded sand layers in the Messinian Feos formation of southern Spain differ (Figs. 7, 13) in aspect from the Ts1–Ts3 layers in the Trave section. Fissures were observed below graded this sandstone layers (Fortuin and Dabrio 2008) and were interpreted as sedimentary dykes induced by seismic shaking from displacement along the Carboneras fault and filled with sand from the overlying tsunami-induced sandstone beds. This situation is analogous to the dark fissures underneath Ts2 in the Trave section. The sand layers are typical for turbidites, they are graded and contain flute-cast at the base, and they were interpreted by Fortuin and Dabrio (2008) as turbidites directly triggered by the earthquake that caused the sand-filled fissures. However, the turbidites are not single, but come in sets of pairs and triplets, decreasing in grain size and thickness upwards. It seems unlikely that the second and third (fourth) sand layers are directly triggered by (regularly reoccurring) aftershocks, and therefore, we interpret the occurrence these turbidite sets not as directly seismically induced but as a seismically induced tsunami or seiche event. The origin of the tsunami waves is linked to the origin of the sand-filled fissures i.e., large vertical movements along large Cabrera thrust fault.
The origin of the Ts1–Ts3 tsunamiites in the Trave section can be found in large NE vergent thrust faults, like the Sibillini thrust front (Roveri et al. 2005), that were active on the western margin of the large Laga foreland basin (Fig. 3), because of the compression of the Laga foreland basin due to movement of the Apennines toward the Apulian platform in the late Messinian (Govers et al. 2009). The Ts1–Ts3 tsunamites are located on the eastern margin of the Apulian platform, favorably located to receive large tsunamis from the Sibillini thrust front. The components of Ts1–Ts3 are dominated by bioclastic, shallow-water limestones of various ages, readily available on the shallow-water Apulian platform underlying the Adriatic Sea. Abundant plant debris in Ts2 also argues for backwash from an emerged area on the Apulian platform. Similar abundant plant remains were found in of Mimbral, Mexico, in the tsunamiites triggered by the Chicxulub impact. The folded and up-thrust sedimentary sequence west of Ancona is dominated by the deep-marine Umbre-Marche pelagic sequence (Alvarez and Montanari 1988; Montanari et al. 1997) and therefore an unlikely source of the components in Ts1–Ts3. The trough cross-bedding of the lowest bed of Ts2 show a paleocurrent direction toward the SW (Fig. 11). Paleocurrent directions toward SW were also measured on a turbidite in the underlying Tortonian Corve sequence (Montanari et al. 1997). These paleocurrent directions demonstrate a north-easterly source on a NE–SW slope on the Apulian platform margin. Alternatively, the triggering faults can be sought in the many thrust faults off the present Adriatic coast. However, these are mostly active in post-Messinian Plio-Pleisticene times and thus an unlikely source of the tsunami waves.
They occur as unique coarse clastic layers within monotonous mudstones without any indication of a trend in the sediments pre- or post-dating the layers.
They form tabular or channeled layers that can be followed (Ts1) over several hundred meters and are not confined to narrow (fluvial) channels.
They are each built of a stack of 4–6 graded sublayers, each decreasing in grain size and thickness and therefore in energy. Each of the graded beds erodes into the previously deposited beds or in the mudstones below its base.
They are cross-bedded in dm-thick high-angle trough cross-beds in the coarse basal parts, with current direction consistent with tsunami backsurge.
They contain climbing-ripple intervals in fine-to-medium sand and the same grain size as in a Bouma Tc interval, which points to deposition under waning current strength and a high suspension load, consistent with the erosion of large amount of clastic material from near-coastal areas.
They contain bivalves that are eroded from near-shore areas by the tsunami backwash but do not contain double-valved shells as is common in tempestites.
They contain layers of fine plant debris, probably derived from water-saturated plant material buried (as peat) in supra-tidal lagoons, transported to offshore regions by the tsunami backsurge.
They are underlain in one case by a meshwork of sand-filled planar fissures in the underlying mudstone. These are probably not root traces indicating a soil, because in cross section, the fissures are not round or elliptical. These fissures may be related to the large earthquake triggering the tsunami vent.
Regular spaced sets of doublets and triplets of graded thin sand layers, each decreasing in thickness and grain size, intercalated in the Lago Mare mudstones of the Feos formation in the Nijar Basin, Spain, are interpreted as tsunamiites deposited by oscillating seiche/tsunami-driven currents, which cause cyclic water movements decreasing in strength. Vertical movements along the nearby Cabrera fault, active in late Messinian times, trigger these oscillating seiches.
We thank Luuk Kleipool, Lisa Meijer, and Mart Zijp for help in the field. Martin Konert is acknowledged for help with the TGA and grain size analyses. Two anonymous reviewers provided a great help in getting the paper in correct shape.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Alvarez W, Montanari A (1988) The Scaglia limestones (Late Cretaceous-Oligocene) in the northeastern Apennines carbonate sequence: stratigraphic context and geological significance. In: Premoli Silva I, Coccioni R, Montanari A (eds) The Eocene-Oligocene boundary in the Marche-Umbria Basin (Italy). International Subcommision on Paleogene Stratigraphy of the International Union of Geological Sciences, Ancona, pp 13–30Google Scholar
- Cita MB (2008) Deep-sea homogenites: sedimentary expression of a prehistoric megatsunami in the eastern Mediterranean. In: Shiki T, Minoura K, Tsuji Y, Yamazaki T (eds) Tsunamiites. Elsevier, Amsterdam, pp 185–202Google Scholar
- Hilgen F, Kuiper K, Krijgsman W, Snel E, van der Laan E (2007) Astronomical tuning as the basis for high resolution chronostratigraphy: the intricate history of the Messinian Salinity Crisis. Stratigraphy 4:231–238Google Scholar
- Husing SK, Hilgen FJ, Abdul Aziz H, Krijgsman W (2007) Completing the Neogene geological time scale between 8.5 and 12.5 Ma. Earth Planet Sci Lett 253:340–358Google Scholar
- Lock BE, Butler RW, Franklund RT (2009) Tempestite sedimentation: an example from the Del Rio formation of West Texas. Trans Gulf Coast As Geol Soc 59:463–476Google Scholar
- Manzi V, Roveri M, Gennaria R, Bertinib A, Biffi U, Giunta S, Iaccarino SM, Lanci L, Lugli S, Negri A, Riva A, EdoardoRossi M, Taviani M (2007) The deep-water counterpart of the Messinian Lower Evaporites in the Apennine foredeep: the Fanantello section (Northern Apennines, Italy). Palaeogeogr Palaeoclimatol Palaeoecol 251:470–499CrossRefGoogle Scholar
- Montanari A, Beaudoin B, Chan L, Coccioni R, Deino A, DePaolo DJ, Emmanuel L, Fornaciari ES, Kruge M, Lundblad S, Mozzato C, Portier E, Renard M, Rio D, Sandroni P, Stankiewicz A (1997) Integrated stratigraphy of the middle to upper Miocene pelagic sequence of the Cònero Riviera (Ancona, Italy). In: Montanari A, Odin GS, Coccioni R (eds) Miocene stratigraphy: an integrated approach. Developments in paleontology and stratigraphy vol 15, pp 409–450Google Scholar
- Popescu SM, Melinte M, Suc J-P, Clauzon G, Quillevere F, Suto-Szentai M (2007) Earliest Zanclean age for the Colombacci and uppermost Di Tetto formations of the ‘‘latest Messinian’’ northern Apennines: new palaeoenvironmental data from the Maccarone section (Marche Province, Italy). Geobios 40:359–373CrossRefGoogle Scholar
- Roveri M, Gallo AB, Rossi M, Gennari R, Iaccarino SM, Lugli S, Manzi V, Negri A, Rizzini F, Taviani M (2005) The Adriatic foreland record of Messinian events (Central Adriatic Sea, Italy). GeoActa 4:139–158Google Scholar
- Smit J, Roep TB, Alvarez W, Montanari A, Claeys P, Grajales-Nishimura JM, Bermudez J (1996) Coarse grained, clastic sandstone complex at the K/T boundary around the Gulf of Mexico: deposition by tsunami waves induced by the Chicxulub impact? In: Ryder G, Fastovski D, Gartner S (eds) The Cretaceous–Tertiary event and other catastrophes in earth history, vol Sp. Pap. 307. Special Paper. Geol. Soc. of Amer., Boulder, pp 151–182Google Scholar
- Yancey TE (1996) Stratigraphy and depositional environments of the Cretaceous–Tertiary boundary complex and basal Paleocene section, Brazos River, Texas. In: Jones JO, Freed RL (eds) Gulf Coast Association of Geological Societies, forty-sixth annual meeting, American Association of Petroleum Geologists regional meeting and Gulf Coast Section SEPM forty-fifth annual convention. Transactions—Gulf Coast Association of Geological Societies. Gulf Coast Association of Geological Societies, New Orleans 46, pp 433–442Google Scholar
Open AccessThis is an open access article distributed under the terms of the Creative Commons Attribution Noncommercial License (https://creativecommons.org/licenses/by-nc/2.0), which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.