Geophysical Research Letters

Comment on “Holocene tsunamis from Mount Etna and the fate of Israeli Neolithic communities” by Maria Teresa Pareschi, Enzo Boschi, and Massimiliano Favalli


[1] Pareschi et al. [2007] (hereafter referred to as PBF07) suggested that the tsunami generated by the collapse of Mount Etna ca. 8,300 yr B.P., destroyed the Neolithic village of Atlit-Yam on the Israeli coast. The main issues raised by PBF07 are examined here as they relate to finds from the site, as well as from other inland Neolithic sites from the Levant.

[2] Pareschi et al. [2006] first suggested that the tsunami occurred ca. 8,000–7,500 yr B.P. and later PBF07 shifted the date to ∼8,300 yr B.P, but no explanation for this shift was offered. If the tsunami occurred at 8,000–7,500 yr B.P., then sediments from Water Well 11 which date to 8,370–8,210 yr B.P. (Figure 1) (but which are said to be tsunami related according to PBF07) would have actually pre-dated the assumed tsunami. (All dates are calibrated years B.P., groups of dates from the same structure were averaged with ±1sigma. Dating by E. Boaretto, Radio-carbon Dating Laboratory, Weizmann Institute, Israel.) Similar sediments from Water Well 66 dated to 8,640–8,540 yr B.P. [Galili et al., 2002] and the concentrations of fish bones and wheat from Locus 10/A dated to 8,425–8,360 yr B.P. also pre-date the tsunami (Figure 1). Alternately, if the tsunami destroyed the village in 8,300 yr B.P. (PBF07), how can the settlement exhibit an uninterrupted sequence of occupation from ca. 9,400 to 8,000 yr B.P.? [Galili et al., 2002] (Figure 1). The same argument applies to the more recent proposed Etna cone collapse of 7,590 ± 130 B.P. [Calvari and Groppelli, 1996]. As noted above, this event also post-dates many of the so-called ‘tsunami features’ identified by PBF07. The end of the occupation of Atlit-Yam ca. 8,000 BP, clearly relates to a well-documented Mediterranean sea level rise following the end of the last glaciation [Bard et al., 1996; Galili et al., 2005a].

Figure 1.

Comparative time scales of the Etna tsunami and the assumed tsunami features, as proposed by Pareschi et al. [2006, 2007], and the occupation of Atlit-Yam.

[3] The human skeletal pathologies identified at Atlit-Yam are mainly associated with infectious diseases resulting from chronic health problems and dental diseases, and are not associated with natural disaster. “Fresh” injuries relating to trauma, which are expected to be found on victims of such a violent event, were not detected. Loss of teeth and partially burnt bones, both features specified by PBF07 as tsunami related, are in fact common in Neolithic human osteological assemblages in the region [Hershkovitz and Galili, 1990].

[4] Burial practices at Atlit-Yam were similar to those identified in other Pre-Pottery Neolithic (PPN) sites in the Levant. Most of the human skeletal material was recovered from formally prepared graves with the deceased interred in a flexed position [Galili et al., 2005b]. Isolated bones were found throughout the site but represent primary graves disturbed in antiquity by human activities such as building, as well as more recent marine agents. Such isolated human bones are common in other submerged Pottery Neolithic (PN) sites [Galili et al., 1998], as well as terrestrial PPN sites.

[5] In all features the faunal assemblage resembles that reported for neighboring submerged PN sites [Horwitz et al., 2006] and other Levantine PPN sites. It clearly differs from assemblages that have undergone sudden and violent catastrophes, where animal bones are usually found in anatomical articulation [Lyman, 1994]. About half of the animal bones bear cut marks attesting to intentional butchery, which in turn indicates that they represent food refuse and not natural mortalities. Moreover, the species and age distribution of animals does not resemble that expected for a catastrophic population, but points to selection and management of a limited range of taxa for meat production [Horwitz and Tchernov, 1987; Galili et al., 1993].

[6] The concentrations of fish bones and cereal grains, cited by PBF07 as evidence for the sudden abandonment of the site, differ in nature and each appears to have accumulated at a different time [Galili et al., 2004]. Such food concentrations have been reported from terrestrial PPN villages in the southern Levant and represent food stores that had not been consumed.

[7] The use of abandoned installations and water wells as garbage pits is a common phenomenon in prehistoric sites throughout the Eastern Mediterranean (e.g., Mylouthkia, Cyprus [Peltenburg, 2003]). Thus, the deposit recovered from the upper parts of the Atlit-Yam water wells, containing a mixture of consumed animal bones and cultural material, is typical of debris associated with human activities, and does not represent debris introduced into the wells by a tsunami.

[8] The three uppermost built stone courses of Water Well 11 survived in situ above the sea floor. Thus the large stones and artifacts found inside Well 11 could not have circumvented this wall unless humans deliberately threw them in. The presence of insect and rodent remains inside the wells may be attributed to natural mortalities or remains of prey introduced by raptors. The fine sediments in the well may have been introduced by wind and sea storms, although seasonal overflow by the neighboring Nahal Oren stream cannot be excluded.

[9] The site is embedded within the upper layer (ca. 10 m. thick) of a hard homogeneous clay of terrestrial origin (Carmel Coast Clay Formation), which is dated to ca. 10,000–12,000 yr B.P. [Galili and Weinstein-Evron, 1985]. Given its date and structure, there is no basis for considering this to be a “tsunami mud layer” (PBF07).

[10] The presence of well-preserved human skeletons still in their original burial position, undisturbed paved stone surfaces, wall foundations and stone-built installations (all of which were constructed without cement), the preserved stone super-structure of Water Well 11 which lies above the sea floor and the 1.5 m high standing stones in Structure 56, could not have survived a destructive tsunami. Similar features are recorded from PN settlements off the Carmel coast and are also typical of terrestrial PPN settlements in the southern Levant.

[11] According to PBF07, the tsunami funneled into the village from the northeast (typo?). Given the morphology of the Atlit Bay and the general setting of the site, a northwest direction of a potential tsunami is more logical.

[12] Historic tsunamis did hit the Levant [Salamon et al., 2007] and the location of the village before abandonment, about 1.5–2 m above sea level, could have exposed it to potential tsunami hazards. In our opinion however, the destruction of the village some 8,300 years ago by a tsunami as proposed by PBF07, finds no support in the archaeological, anthropological, faunal, botanical or sedimentary record from the site (Table 1). Instead, the data indicate that the village was abandoned ca. 8000 years B.P. due to the gradual post-glacial rise in sea level rise, similar to coastal Neolithic villages all over the world. The site was first covered by coastal sand dunes that protected it from abrasion by marine agents and then submerged by the rising sea. Due to the sea level rise, the subsequent PN villages in the region were built farther to the East [Galili et al., 1993].

Table 1. Field Evidence and Finds Observed at Atlit-Yam Versus Expected From a Tsunami Impact
SubjectExpected in Case of a Tsunami ScenarioObserved Field Evidence and Finds
Dating7,600–8,000 yr B.P. [Pareschi et al., 2006]The proposed tsunami post-dates sediments Pareschi et al. [2007] argue are tsunami deposits (TD)
Dating∼8,300 yr B.P. [Pareschi et al., 2007]The site was occupied until 8,000 yr BP, i.e., 300 years after the proposed tsunami.
ArchitectureNo structures or installations are expected to survive the tsunami while in an in situ and upright positionWell-preserved un-cemented stone structures, paved stone surfaces, including in situ 1.5 m high stone megaliths in upright position
Human burialsUnburied skeletons randomly scattered, twisted and in strange positions, formal graves destroyed, various mortuary contextsMajority in formal, prepared graves, individuals in flexed position, isolated bones result from disturbance to primary graves in antiquity.
Human pathologiesEvidence of fresh trauma, mainly to the skullMainly chronic diseases, no fresh injuries or penetrating objects relating to trauma.
FaunaArticulated skeletons with complete bonesNo articulated skeletons, few complete bones, all bones consumed, many cut marks.
Fauna distributionRandom, no patterningMainly in water wells fill with other garbage, or in open spaces, rarely in dwellings.
Fish/cereal concentrationsMany contemporaneous cases of abandoned food, representing a single disastrous eventFish and grains are not contemporaneous; food stores/caches are common in terrestrial Neolithic sites.
Water wellsTotal destruction of the superstructuresThe courses of the upper structure of Well 11 perfectly preserved.
Wells fillContemporary typical tsunami deposits dated to a single event in all pits, shafts and wellsDeposits in wells and pits represent anthropogenic refuse and are not contemporary.