Go to top

The Dead Fish and Soldiers Quake

~142 BCE

by Jefferson Williams









Introduction & Summary

In ~142 BCE, the Seleucid Empire experienced a power struggle between competing monarchs. Two generals, Sarpedon and Diodotus Tryphon , led their troops into battle. Tryphon’s army won. As Tryphon’s army marched up the Lebanese coast after the battle, a tsunami is reported to have struck and drowned parts of the army; leaving a scene of Dead Fish and Soldiers in its wake.

This tsunami appears to be reported from an earlier lost source - Posidonius (~135 BC – ~ 51 BC) - repeated by Strabo in Geographicum and Athenaeus of Naucratis in The Deipnosophistae. The accounts are more or less identical. Athenaeus and Strabo do not explicitly mention that the tsunami was caused by or accompanied by an earthquake however, after describing the tsunami, Strabo speculates that the earth moved up or down to create the displacement of water, perhaps hinting that this tsunami was accompanied by some type of earthquake report by Posidonius or others. Evidence that the tsunami account may describe an earthquake comes from the archaeological site of Tel Ateret in the Galilee and paleoseismic evidence in Syria, the Galilee, the Dead Sea, and the Arabah. At Tel Ateret, there is archaeoseismic evidence for an earthquake on or near to ~142 BCE. At Bet Zayda in the Galilee and the Tekieh trenches in Syria, there is paleoseismic evidence for an earthquake around this time. Although an earthquake on the Lebanese coast was too far from the Dead Sea to have generated seismites, seismites from around this time (mid second century BCE) have been observed in En Feshka, En Gedi, and Nahal Ze 'elim. The seismites thicken towards the south and are thick and brecciated in Nahal Ze 'elim - indicating high levels of local Intensity. There is additional paleoseismic evidence from the Taybeh trench in the Arabah. This suggests an earthquake couplet may be at play where a southern quake triggered a northern quake or vice-versa.

See the Notes section for brief discussion on whether the Dead Fish and Soldiers Quake was conflated in the historical sources with the Malalas Confusion Quake and the Seventeenth of Adar Quake

Textual Evidence

Section
Geographicum by Strabo
The Deipnosophistae by Athenaeus of Naucratis


Geographicum by Strabo

Strabo (~ 64 BCE – ~ 24 CE), in his book Geographicum, possibly using Posidonius (~135 BCE – ~ 51 BCE) as his source (Kidd,1988:40), reports sea wave flooding between Tyre and Acre (aka Ptolemais). Although he does not specifically cite an earthquake as the cause of the sea wave, he speculates that ground movement may have caused the sea wave comparing this event to another possible earthquake and tsunami reported in ~20 BC near Mount Casius (aka Cassium) in Egypt. If there was an earthquake, there is some uncertainty about its date. According to Ambraseys (2009), it could have occurred between 138 BCE and 125 BCE. According to Karcz (2004), it likely struck between 145/144 BCE and in 138/137 BCE. Karcz (2004) added that if the Dead Fish and Soldiers Quake and the Seventeenth of Adar Quake are the same event, this event would likely have occurred in 143/142 BCE - a date which coincides with archaeoseismic evidence at Tel Ateret. The description in Geographicum (Book XVI Chapter 2 Paragraph 26) follows:
A marvellous occurrence of a very rare kind is reported as having taken place on this shore between Tyre and Ptolemaïs: at the time when the Ptolemaeans, after joining battle the Sarpedon the general, were left in this place, after a brilliant rout had taken place, a wave from the sea, like a flood-tide, submerged the fugitives; and some were carried off into the sea and destroyed, whereas others were left dead in the hollow places; and then, succeeding this wave, the ebb uncovered the shore again and disclosed the bodies of men lying promiscuously among dead fish. Like occurrences take place in the neighborhood of the Mt. Casius situated near Aegypt, where the land undergoes a single quick convulsion, and makes a sudden change to a higher or lower level, the result being that, whereas the elevated part repels the sea and the sunken part receives it, yet, the land makes a reverse change and the site resumes its old position again, a complete interchange of levels sometimes having taken place and sometimes not. Perhaps such disturbances are subject to periodic principles unknown to us, as is also should be the case of the overflows of the Nile, which prove to be variant but follow some unknown order.
Ambraseys (2009) notes that an earthquake was not specifically mentioned adding that assuming that such a large event in fact occurred, it should have caused havoc in the coastal area of southern Lebanon and Palestine, for which there is not a hint in the sources. However, In quoting Strabo, Ambraseys (2009) neglected to include Strabo’s ground movement and tsunami-like speculations. Ben-Menahem (1991) assigned a Local Magnitude of 7.0 and a Maximum Local Intensity of X to this supposed earthquake also stating that there was partial subsidence of Sur Island and that earthquake shaking was strong in Cyprus. Although Ben-Menahem (1991) cites Strabo as his source, Strabo does not mention strong earthquake shaking in Cyprus and merely speculates about the possibility of uplift followed by subsidence with no permanent change in elevation in the vicinity of Sur Island. These observations are interpretations by Ben-Menahem (1991).

The Deipnosophistae by Athenaeus of Naucratis

Athenaeus of Naucratis writing in his book The Deipnosophistae in the early 3rd-century AD also records this event (Book VIII 332 B 24) while explicitly stating that Posidonius [c. 135 BCE – c. 51 BCE ] was his source.
I know also that Poseidonius the Stoic speaks of a great quantity of fishes in these words: When Tryphon of Apameia, who had seized the kingdom of Syria, was attacked near the city of Ptolemais by Sarpedon, Demetrius's general, the latter was defeated and forced to retreat into the interior with his troops. Tryphon's army were marching along the coast after their victory in the battle, when suddenly a wave from the ocean lifted itself to an extraordinary height and dashed upon the shore, engulfing all the men and drowning them beneath the waters. And when the wave receded it left behind a huge pile of fishes among the dead bodies. The followers of Sarpedon, hearing of this disaster, came up and gloated over the bodies of their enemies, while they also carried away an abundance of fish and offered sacrifice to Poseidon, god of the rout, near the suburbs of the city.

Archeoseismic Evidence

Location Status Intensity Notes
Tel Ateret probable ≥9
  • ~2.5 m of slip. MW ≈ 7.1-7.4
  • Dated 3rd century BCE - ~142 BCE but likely struck around ~142 BCE
Tell Anafa possible a massive leveling and terracing operation may have obliterated archaeoseismic evidence

Tel Ateret

Satellite Photo of Tel Ateret Fig. 8.5a

A satellite photo (GoogleEarth) of Tel Ateret – Benot Ya’aqov bridge. Vadum Iacob castle straddles the fault trace, that runs through an aqueduct system south of the castle.

JW: White box labeled B encompasses the archaeological site

Agnon in Garfunkel et al (2014)


Names

Transliterated Name Source Name
Tel Ateret
Metzad ‘Ateret Hebrew מצד אטרט
Vadum Jacob Crusader
Vadum Iacob Crusader
Chastellet Crusader
Le Chastelez 13th century CE
Bayt al-Ahzan 12th century CE Arabic بايت الأهزان
Qasr al-'Ata Modern Arabic قاسر الء'اتا
Qasr al-'Atara Modern Arabic قاسر الء'اتارا
Introduction

Tel Ateret is situated atop a structural high that oversees a crossing of the Jordan River. It's military/strategic value has led to multiple occupations which from bottom up include an Iron Age II fortification, a Hellenistic Complex, a medieval Crusader Castle known as Vadum Jacob, and the last structure - a Mumluk and Ottoman pilgrimage site with a mosque ( Ellenblum, et al, 2015). The Hellenistic settlements dating from the 3rd to the 1st centuries BCE may have been Pharanx Antiochus captured by Hasmonean King Alexander Jannaeus in 81 BCE (Ellenblum et. al., 2015 citing Ma'oz, 2013). In some traditions (e.g. 12th century Muslim), the site was associated with the dwelling place of the Biblical Patriarch Jacob when he learned of the disappearance of his son Joseph ( Ellenblum, 2003). Structures built on the site straddle the active Jordan Gorge Fault thus providing a unique location to resolve slip from earthquake events.

Note: Some papers refer to the fault intersecting the structures as the Dead Sea Fault (DSF). This refers to the Jordan Gorge Fault (JGF) segment which is one of several large active faults comprising the Dead Sea Transform. Chronology

Ellenblum et al (2015) found 8 m of accumulated slip at Tel Ateret since the construction of an Iron Age IIA fortification early in the 1st millenium BCE. They were able to resolve this slip into a sequence of time periods summarized below. In a few cases, they were able to estimate the slip of individual earthquake events allowing for an estimate of Moment Magnitude (MW) using the scaling laws of Wells and Coppersmith (1994)
Date Slip1
(m)
Slip2
(m)
Moment3
Magnitude
MW
Intensity4
I
Intensity5
I
Slip2
Velocity
(m/s)
Comments
980 BCE - 142 BCE >2 n/a unresolved unresolved n/a n/a At least 2 m of slip displaced Iron Age IIa walls in an unknown number of events
probably ~142 BCE ~2.5 n/a 7.1 - 7.4 ≥ 9 n/a n/a Excavated coins suggest this event occurred around and no earlier then 142 BCE
~50 BCE - 1178/9 CE ~1.5 n/a unresolved unresolved n/a n/a
  • Undated earthquakes after mid Hellenistic time (~50 BCE) and before construction of Vadum Jacob in 1178/9 CE
  • The site appears to have been abandoned in the mid 1st century CE (no earlier then 64/65 BCE) which may have been due to an earthquake
probably 20 May 1202 CE ~1.6 1.25 7.0 - 7.2 ≥ 9 9 3 Crusader Fortress Vadum Jacob Damaged
Ottoman period ~0.5 0.5 6.6 - 6.8 ≥ 8 7 1 Ottoman Mosque was damaged
Footnotes

1 from Ellenblum et al (2015)
2 from Schweppe, et al (2021) - Schweppe et al (2021) used detailed laser scans of the site and discrete element models to estimate slip and slip velocity for the last two events.
3 computed from Ellenblum et al (2015)'s slip using Wells and Coppersmith (1994)
4 Estimated by Jefferson Williams primarily using Earthquake Archeological Effects chart of Rodríguez-Pascua et al (2013: 221-224)
5 Schweppe, et al (2021) converted slip velocity to Intensity using using Wald et al (1999)

Note - Schweppe, et al (2021) also produced Magnitude estimates which are in fairly good agreement with the table above but because their Magnitude estimates required an input of presumed fault rupture length, they are not repeated in the interest of avoiding circular reasoning.
Iron Age IIA and later Earthquake(s) - 980 BCE - 142 BCE

Ellenblum et. al. (2015:6) uncovered Iron Age IIA remains (ca. 980-830 BCE) to the south and partially beneath the Hellenistic ruins in the southern part of the site. Preliminary dating was based on architectural style and pottery typologies. Ellenblum et. al. (2015:6) estimated that the Iron Age IIA wall was displaced 8 m across the fault with 6 m of displacement taking place after the early Hellenistic period. This left 2 m of displacement in an unknown number of events during the first millennium BCE prior to the 142 BCE earthquake.

Early Hellenistic Earthquake - probably ~142 BCE

  • Hellenistic walls in the southern part of Tel Ateret from Ellenblum et al (2015).
Ellenblum et. al. (2015:5) estimated a displacement of ~2.5 m from this event which, though dated from the 3rd century BCE - ~142 BCE, probably struck around ~142 BCE. Ellenblum et. al. (2015:3) described two Hellenistic building phases - one early and one late - with the later phase built atop the earlier one. Although the Northern Hellenistic Complex was too heavily damaged to allow tracing of walls across the fault zone, the southern complex in Area E allowed for such analysis which is described as follows:
Two building phases are discernible in this excavated Hellenistic compound with crosscutting relations that determine their temporal relations. The walls of both phases are deformed and truncated at the fault immediately south of the faulted Crusader wall.

...

The walls of the earlier phase are invariably straight along 20 m except for within the fault zone, where they are crooked left laterally. Heaps of cobbles at the bottom of the walls, which have fallen from the upper parts of the early phase walls, buried indicative artifacts including candles, vessels, cooking pots, decorated fishplates, and relief bowls, as well as imported Hellenistic wares and a hoard of coins. All these findings unambiguously belong to the Hellenistic period. The most special finding is a hoard of 45 small bronze coins buried within the debris of the older walls. A numismatic analysis of the 32 well-preserved datable coins limits the range of the hoard to 150s-140s BCE. The latest dated coin was minted in 143/142 BCE. About 60% of the coins cluster close to this date.
Ellenblum et. al. (2015:4) interpreted the coin hoard as consistent with a scenario of sudden collapse of the wall, possibly triggered by an earthquake, adding:
The types and the arrangement of the walls indicate that the two successive construction periods were separated by a destruction event within the Hellenistic period, which left a considerable amount of debris along the fault. Based on the stratigraphy and lateral displacement, we attribute the termination of the older phase to an earthquake that tore apart the earlier phase during the second century BCE. The latest dated coin in the hoard, minted in 143/142 BCE, provides a lower bound for the date of this earthquake [i.e. a terminus post quem], after which the late Hellenistic walls were built. In another two-phase Hellenistic settlement some 20 km north of Ateret - Tell Anafa, an abrupt termination of a well-developed settlement with elaborate construction [Sharon Herbert in Stern et al (1993:58-61, v. 1)], may be re-interpreted as a result of an earthquake destruction.
Ellenblum et. al. (2015:5) estimated slip from earthquake events by measuring displacement of walls across the fault:
The Hellenistic walls are bent immediately south of the faulted Crusader wall. Reconstruction of the early wall to its original straight disposition requires about 6 m (Figures 3a & 3b ). The later wall (highlighted yellow), dated to the late 2nd - 1st century BCE, is exposed along ~8 m. It is also curved leftward at the center by about 20°. The builders of the Crusader wall destroyed the eastern segment of the late Hellenistic wall, making direct measurement of the displacement impossible.
The 6 m of displacement represents slip from the Late Hellenistic Earthquake and all subsequent earthquakes. Ellenblum et. al. (2015:5) estimated a displacement of ~2.5 m for just the Late Hellenistic Earthquake. Using the scaling laws of Wells and Coppersmith (1994), this corresponds to a magnitude of 7.1 - 7.4 (see Calculator).

Late Hellenistic Earthquake ? - Mid 1st century BCE

Ellenblum et. al. (2015) estimate ~1.5 meters of fault slip occurred on the site between its abandonment probably in the middle of the first century BC and when a Crusader fortress was built at the end of the 12th century CE. Due to the sites abandonment and lack of identified new constructions during this time, it is difficult to resolve the ~1.5 meters of slip into individual earthquake events. However, abandonment of the site may have been precipitated by an earthquake. The latest Hellenistic coin excavated from the site dates to 65/64 BCE indicating desertion of the site occurred afterwards.

post Hellenistic Earthquakes - ~50 BCE - 1178/9 CE

Ellenblum et. al. (2015) estimate ~1.5 meters of fault slip occurred on the site between its abandonment probably in the middle of the first century BC and when a Crusader fortress was built at the end of the 12th century CE. Due to the sites abandonment and lack of identified new constructions during this time, it is difficult to resolve the ~1.5 meters of slip into individual earthquake events.

Vadum Jacob Earthquake - probably 1202 CE

  • Plan of Vadum Jacob fortress with offsets from Ellenblum et al (1998).
Ellenblum et al (1998:304) report, based on historical sources (e.g. William of Tyre, Abu-Shama, Ibn-al-Athir, and 'Imad al-Din al-Isfhani), that the foundation stone of the castle [of Vadum Jacob] was laid in October 1178 CE. The castle, only partially constructed at the time, was besieged and destroyed 11 months later, on 30 August 1179 CE providing a terminus post quem of 1179 CE for its seismic destruction. Up to ~2.1 m of lateral slip was observed in the southern and northern defense walls along with up to 10 cm. of vertical slip. 0.5 m of the lateral slip was attributed to a later seismic event which damaged an Ottoman Mosque which was later built on the site . This left ~1.6 m of slip between the military destruction of the castle in 1179 CE and the seismic event that damaged the Ottoman mosque. In order to produce a terminus ante quem for initial seismic damage to Vadum Jacob, a trench was dug parallel to the southern face of the castle in which 4 units were identified. Units 1 and 2 were recorded as having been deposited on or prior to the castle's military destruction on 30 August 1179 CE. A fallen ashlar block on top of Unit 2 was presumed to have fallen immediately after the Muslim conquest as a historical source document (Abu-Shama) details partial dismantling of the castle soon after it was conquered. Colluvial Unit 3 was dated from 1179 CE to present and was presumed to have accumulated in the centuries after the Muslim conquest. Unit 4 is a modern bioturbated soil horizon. Faults within the trench were associated with seismic displacement of the Crusader wall and a later seismic event. Ellenblum et al (1998:305) described the faults as follows:
The faults extend to two different stratigraphic levels: One group of faults displaces the alluvium of unit 1 and the limy level of unit 2, but extends only a few centimeters into post-1179 unit 3; the second group of faults breaks much higher into the colluvial wedge, up to the base of the modern soil horizon, and possibly to the surface. These observations suggest that at least two earthquakes produced the 2.1 m offset of the southern wall that is now observed. One event occurred soon after the outer ashlar wall was removed, i.e., very soon after 1179. The second post-1179 earthquake also produced rupture at Vadum Jacob, but well after removal of the wall and the accumulation of the colluvium, probably much closer to the present.
Although a strict terminus ante quem was not established, the trench suggests that an earthquake struck soon after military destruction of the castle leaving the 1202 CE earthquake as the most likely candidate.
  • Trench Log from Vadum Jacob fortress from Ellenblum et al (1998) .

Ottoman Mosque Earthquake

Ellenblum et al (1998:305) described archaeoseismic evidence from Mamluk and Ottoman mosques built on the site as follows:

In the northern part of the castle, we also unearthed a Muslim mosque whose northern wall is displaced sinistrally by 0.5 m. A mikhrab (the Muslim praying apse) is well preserved in the southern wall. According to the study of the pottery, the mosque was built, destroyed, and rebuilt at least twice: the initial structure was built in the Muslim period (12th century) and later rebuilt once or twice during the Turkish Ottoman period (1517-1917). The 0.5 m displacement is observed in the northern wall of the latest building phase . The repetitive building of this site might be due to earthquakes.
The latest rebuilding phase was not dated. Ellenblum et al (2015) suggested that the first of the Baalbek Quakes on 30 October 1759 CE was responsible while Ellenblum et al (1998:305) and Marco et al (1997) entertained the possibility that the Galilee Earthquake of 1837 CE is also a possible candidate

Seismic Effects
Early Hellenistic Earthquake - probably ~142 BCE

Ellenblum et al (2015) report the following seismic effects:

  • The walls of both phases are deformed and truncated at the fault immediately south of the faulted Crusader wall.
  • Heaps of cobbles at the bottom of the walls
  • a destruction event within the Hellenistic period, which left a considerable amount of debris along the fault.
  • The Hellenistic walls are bent immediately south of the faulted Crusader wall
  • ~2.5 m of displacement

Vadum Jacob Earthquake - probably 1202 CE

  • Plan of Vadum Jacob fortress with offsets Fig. 2a&b from Ellenblum et al (1998).
Ellenblum et al (1998:304) described seismic effects as follows:
We discovered offset archaeological remains at four excavated locations within the castle (Fig. 2). Three of these locations are Crusader structures and the fourth is a late medieval/early modern Muslim structure. The offset, fully expressed in the southern and northern defense walls, reaches 2.1 m in sinistral displacement with less than 5 cm of vertical slip (Marco et al., 1997). The 50 m long walls crossing the hill are ideal baselines, having meticulously laid masonry and a fault-perpendicular (east-west) orientation. The displacement measured on the walls records the cumulative slip on the fault since 1179, the year the castle was conquered by Saladin. Displacement is distributed over about a 10 m wide zone, and the deformation is accommodated primarily by small offsets and rotations of the carved limestone blocks. All the displacements on the southern wall are purely horizontal (all the blocks retain their original level), and all the rotations are about vertical axes. A minor vertical component of slip, up to 10 cm, is observed in the northern wall. North of the southern main gate, the fault trace bends westward and a Crusader floor is torn, forming a 2 m wide graben (Fig. 2A). The geometry of the graben is compatible with a small left bend in the trace of the fault.

Ottoman Mosque Earthquake

Marco et al (1997:14) described the archaeoseismic evidence as follows:

In the northern part of the fortress, additional walls from the Arabic period show lesser left-lateral displacements of the order of 0.2-0.3 m at points which coincide with the N—S lineament. A Muslim-style room is interpreted as a mosque due to the presence of a south-facing round niche identified as a "makhreb" or a praying apse. The mosque wall is bent and offset sinistrally, and a layer of building stones, that lies imbricated directly on the floor, suggests collapse.
Ellenblum et al (2015) updated the displacement estimate to 0.5 m.
  • Photo of faulted wall of the Ottoman mosque from Marco (2009).

Slip History of Tel Ateret

Tel Ateret Slip History Figure 4

Schematic illustration of the stages of slip accrual (values are rounded) in the Ateret structures, timeline from bottom to top.

Ellenblum et. al. (2015)


Tel Ateret Slip History Figure 3b

A schematic illustration of the archaeological strata at Tell Ateret offset by the Dead Sea Fault. The older the strata the larger the offset

Ellenblum et. al. (2015)

Intensity Estimates
Early Hellenistic Earthquake - probably ~142 BCE

Effect Description Intensity
Collapsed Walls VIII+
Displaced Walls VII+
The archeoseismic evidence requires a minimum Intensity of VIII (8) when using the Earthquake Archeological Effects chart of Rodríguez-Pascua et al (2013: 221-224) .

Vadum Jacob Earthquake - probably 1202 CE

Effect Description Intensity
Displaced Walls VII+
Although the reported archeoseismic evidence requires a minimum Intensity of VII (7) when using the Earthquake Archeological Effects chart of Rodríguez-Pascua et al (2013: 221-224) , the reported ~2.1 m of coseismic slip suggests a much higher intensity - IX (9), X (10), or higher.

Ottoman Mosque Earthquake

Effect Description Intensity
Collapsed Walls VIII+
Displaced Walls VII+
The archeoseismic evidence requires a minimum Intensity of VIII (8) when using the Earthquake Archeological Effects chart of Rodríguez-Pascua et al (2013: 221-224) .

Calculator
Moment Magnitude from Strike-Slip Fault Displacement

Source - Wells and Coppersmith (1994)

Variable Input Units Notes
cm. Strike-Slip displacement
cm. Strike-Slip displacement
Variable Output - not considering a Site Effect Units Notes
unitless Moment Magnitude for Avg. Displacement
unitless Moment Magnitude for Max. Displacement
  

Notes and Further Reading
References

Ellenblum, R., et al. (2015). "Archaeological record of earthquake ruptures in Tell Ateret, the Dead Sea Fault." Tectonics 34(10): 2105-2117.

Ellenblum, R., et al. (1998). "Crusader castle torn apart by earthquake at dawn, 20 May 1202." Geology 26(4): 303-306.

Marco, S. et. al., (1997). "817-year-old walls offset sinistrally 2.1 m by the Dead Sea Transform, Israel." J. Geodyn. 24: 11.

Schweppe, G., et al. (2021). "Reconstructing the slip velocities of the 1202 and 1759 CE earthquakes based on faulted archaeological structures at Tell Ateret, Dead Sea Fault." Journal of Seismology.

Marco, S. (2009). The history of the Frankish Castle of Vadum Iacob. Dead Sea Workshop, pp. 35-41

Agnon, A. (2014). Pre-Instrumental Earthquakes Along the Dead Sea Rift. Dead Sea Transform Fault System: Reviews. Z. Garfunkel, Z. Ben-Avraham and E. Kagan. Dordrecht, Springer Netherlands: 207-261.

Maʿoz, Z. U. (2013). "A Note on Pharanx Antiochus." Israel Exploration Journal 63(1): 78-82.

Ellenblum, R. (2003), Frontier activities: The transformation of a Muslim sacred site into the Frankish Castle of Vadum lacob, Crusades, 2, 83-97.

Ellenblum, R., 1998, Frankish Rural Settlement in the Latin Kingdom of Jerusalem: Cambridge, England, Cambridge University Press, 321 pp.

William-of-Tyre, 1986, Willelmi Tyrensis Archiepis-copi Chronicon (ed. R.B.C. Huygens): Corpus Christianorum Continuatio Mediaevalis, 63-63a, Turnhout, book 21, chapter 25 (26), p. 997.

Abu-Shama, 1872-1906, Le livre des deux jardins. Histoire des deux regnes, celui de Nour ed-Din et celui de Salah ed-Din, in Recueil des historiens des croisades, Historiens orientaux: Acadèmie des Inscriptions et Belles-Lettres, Paris, v. 4, p. 194-208.

Ibn al-Athir, 1872-1906, Extrait de la chronique intitulée Kamel-altevarykh, par Ibn-Alatyr, in Recueil des historiens des croisades, ed. Académie des Inscriptions et Belles-Lettres: Paris, Historiens orientaux, p. 189-744.

’Imad al-Din al-Isfahani, 1971, Bundari, Sana al-barq al-Shami, abridged by al-Bundari, Part 1: Beirut, R. Sesen.

Tell Anafa

Names

Transliterated Name Source Name
Tel Anafa Hebrew תל אנפה‎
Tel el-Hader Arabic تل الأخضر‎
Introduction

Tel Anafa was inhabited from the Early Bronze age through the Early Roman period and then again as attested by some later Arab structures. The most significant remains were found in the Hellenistic period (Herbert in Stern et al, 1993). The name of the town in Hellenistic times is unknown.

Chronology

Herbert in Stern et al (1993) summarized Tel Anafa's Hellenistic and Early Roman history as follows :
The evidence of the coins and stamped amphora handles together with the architectural remains of the Hellenistic levels indicates that there was a minor Ptolemaic settlement on the site in the third century BCE. This was succeeded in the second half of the second century BCE by a prosperous Seleucid settlement that took an active part in the trade of the Late Seleucid empire and flourished through the first quarter of the first century BCE. The chaotic conditions surrounding the disintegration of the Seleucid empire probably led to the site's abandonment some time shortly after 75 BCE. The mound was reoccupied in the early years of the first century CE as part of the reorganization of the Galilee under Herod Philip, whose capital was at nearby Caesarea Philippi (Banias). It was again abandoned toward the end of that century and was not reoccupied in antiquity.
Ellenblum et. al. (2015:4) suggested that an earthquake could be interpreted from the results of excavations in the Hellenistic period at Tel Anafa.
In another two-phase Hellenistic settlement some 20 km north of Ateret - Tell Anafa, an abrupt termination of a well-developed settlement with elaborate construction [Sharon Herbert in Stern et al (1993:58-61, v. 1)], may be re-interpreted as a result of an earthquake destruction.
Although Herbert in Stern et al (1993) did not report any evidence for seismic destruction in Hellenistic times, they did date construction of a Late Hellenistic stuccoed building around ~125 BCE noting that a coin of Alexander Zebina (128-125 BCE) found in the construction fill of the bath's southern room is the latest find under any of the building's original floors. Herbert in Stern et al (1993) also reports that a massive leveling and terracing operation took place with the construction of the Late Hellenistic stuccoed building, obliterating earlier architectural remains. This could explain an absence of archaeoseismic evidence. Coins and [] stamped amphora handles strongly suggested that the site was abandoned in second quarter of the first century BCE; similar to Tel Ateret which was abandoned sometime after 65/64 BCE.

Notes and Further Reading
References

Tsunamogenic Evidence

Paleoseismic Evidence

Paleoseismic Evidence for the Southern and Northern Dead Fish and Soldiers Quake of ~150 BC is summarized below:

Location Status Intensity Notes
Tekieh Trenches possible Event B - ~2 m left lateral displacement
Bet Zayda possible Event CH4-E6 (392 BCE - 91 CE)
Dead Sea n/a n/a
En Feshka possible 2 possible candidates - 1 and 1.5 cm. thick Type 4 seismites
Nahal Darga possible 20 cm. thick seismite - Deformed Unit 8 in Stratigraphic Unit 10 (250 BCE +/- 200)
En Gedi possible 1cm. thick Type 4 seismite
Nahal Ze 'elim probable 15 cm. thick Type 4 seismite at ZA-1 and 15 cm. thick Type 4 seismite at ZA-2
Taybeh Trench probable Event E6 (160-117 BCE)
Qatar Trench unlikely Event E7 (338-213 BCE) - Fortress at Arad Quake is a better fit
Conclusion n/a n/a


Tekieh Trenches

Gomez et. al. (2003) may also have seen evidence for this earthquake in paleoseismic trenches in Syria (Event B – below 170 BC – 20 AD colluvium).

Tekieh Trench Seismic Events Figure 13

Summary of events observed in the trenches and the interpreted palaeoseismic history of the Serghaya fault. Colluvial wedge deposits post-date palaeoseismic events. Stratigraphic ties provide additional constraint on the relative timing of events. Ages represent calendar corrected radiocarbon ages for given features (2σ uncertainties provided).

Gomez et al (2003)


Bet Zayda

Wechsler at al. (2014) may have seen evidence for this earthquake as event CH4-E6 (modeled age 392 BCE – 91 CE) in paleoseismic trenches at Bet Zayda just north of the Sea of Galilee.

Bet Zeyda Earthquakes
Figure 9

Probability density functions for all paleoseismic events, based on the OxCal modeling. Historically known earthquakes are marked by gray lines. The age extent of each channel is marked by rectangles. There is an age uncertainty as to the age of the oldest units in channel 4 (units 490-499) marked by a dashed rectangle. Channel 1 refers to the channel complex studied by Marco et al. (2005).

Wechsler at al. (2014)


2D and 3D Paleoseismic Study at Bet Zayda

Results are based on a 2D and 3D paleoseismic study conducted over multiple years utilizing multiple trenches. Trenches were dug to examine paleo-channels which intersect the active Jordan Gorge Fault. A few paleo-channels were active long enough to record paleo-earthquakes. Initial work done by Marco et al (2005)) identified fault ruptures with two historical earthquakes which were dated as follows:

Date Displacement (m)
1202 CE ~2.2
1759 CE 0.5
Another channel dating between 3 and 5 ka was displaced up to 15 meters.

Subsequent work at the same location by Wechsler at al. (2014) revealed 8 more surface-rupturing earthquakes in two paleo-channels which were labeled as Channels 3 and 4. Radiocarbon sampling appears to have been sufficiently dense except for Event CH4-E6..

Bet Zayda Plots and Charts

Description Image Source
Age Model Wechsler at al. (2014)
Age Model
Big
Wechsler at al. (2014)
Age Model
really big
Wechsler at al. (2014)
Map of
Trenches
Fault
Channels
Wechsler at al. (2014)

Dead Sea

Although mid second century BCE dates were assigned to seismites in En Feshka, En Gedi, and Nahal Ze e'lim by Kagan et. al. (2011), Migowski et. al. (2004), and Kagan et. al. (2011) respectively, it is unlikely that this northerly earthquake would have created Dead Sea seismites due to the distance involved. At the most southerly Nahal Ze ‘elim site, Kagan et al (2011) records thick brecciated seismites (15 cm. and 8 cm. at the ZA-1 and ZA-2 sites respectively) which correlate well with this event; within the 1σ range. At the more northerly sites of En Gedi and En Feshka, seismite thicknesses for mid second century earthquake(s) are not reported to exceed 1.5 cm. Williams (2004) and Agnon et. al (2006) both revised an original date assignment of Event A at Nahal Ze’elim by Ken-Tor et al. (2001a) from ~64 BCE to approximately mid second century BCE since the mid second century date better matched the radiocarbon profile and estimates of sedimentation rate.
Seismite Types

Seismite Types of Wetzler et al (2010) are used in Intensity Estimates. Seismite Types from Kagan et al (2011) were converted to those of Wetzler et al (2010) to estimate Intensity.

Seismite Types (Wetzler et al, 2010)
Type Description
1 Linear waves
2 Asymmetric Billows
3 Coherent vortices
4 Breccia
Seismite Types (Kagan et al, 2011)
Type
(Kagan)
Type
(Wetzler)
Description
A 4 Intraclast breccia layer
B 4 Microbreccia
C 4 Liquefied sand layer within brecciated clay and aragonite
D 1, 2, or 3 Folded laminae
E 1 Small Fault millimeter -scale throw

En Feshka
Kagan et al (2011) in Table 3 report two seismites from En Feshka at depths of 393 and 402 cm. depth which might fit this earthquake.

En Feshka Plots and Charts

Image Description Source
Age Model Kagan et al (2011)
Age Model - big Kagan et al (2011)
Age Model Kagan et al (2010)
Age Model - big Kagan et al (2010)

En Feshka Core (DSF) Photos

This core was taken in 1997 by GFZ/GSI

Image Description Image Description Image Description Image Description Image Description
Composite Core DSF
Sections B1-B5

0-499 cm.
Section B1

0-93 cm.
Section B2

100-197 cm.
Section B3

200-298 cm.
Section B4

300-396 cm.
Section B5

400-499 cm.

Nahal Darga
In the coarser grained lithology present at Nahal Darga, Enzel et. al. (2000) report a 20 cm. thick seismite in Deformed Unit 8 in Stratigraphic Unit 10 which is dated to 250 BC +/- 200 (2000-2400 BP) (see Table 2).

En Gedi
Migowski et al (2004) assigned a 140 BC date a 1cm. thick seismite at a depth of 302 cm. (3.02 m) in the 1997 GFZ/GSI core DSEn (Table 2).

En Gedi Core (DSEn) Charts and Plots

Image Description Source
Floating Varve Chronology
and Radiocarbon dates
Migowski et al (2004)
Floating Varve Chronology
and Radiocarbon dates -large
Migowski et al (2004)
Migowski's Date shift Migowski (2001)
Recounted Age-depth plot Neugebauer at al (2015)
Recounted Age-depth plot - large Neugebauer at al (2015)
Correlated Age-depth plots
of DSEn and ICDP 5017-1
Neugebauer at al (2015)
Comparison of paleoclimate proxies
from DSEn to other sites
Neugebauer at al (2015)
Core correlation
DSEn to ICDP 5017-1
Neugebauer at al (2015)
Core correlation
DSEn to ICDP 5017-1 -big
Neugebauer at al (2015)
Thin Section of Jerusalem Quake
showing varve counts
shallow section
Williams et. al. (2012)
Thin Section of Jerusalem Quake
showing varve counts
deep section
Williams et. al. (2012)
Thin Section of Jerusalem Quake
showing varve counts
shallow section - big
Williams et. al. (2012)
Thin Section of Jerusalem Quake
showing varve counts
deep section - big
Williams et. al. (2012)

En Gedi Core dating ambiguities

The En Gedi Core (DsEn) suffered from a limited amount of dateable material and the radiocarbon dates for the core are insufficiently sampled in depth to produce an age-depth model that is sufficiently reliable for detailed historical earthquake work in the Dead Sea. Migowski (2001) counted laminae in the core to create a floating varve chronology for depths between 0.78 and 3.02 m which was eventually translated into a year by year chronology from 140 BCE to 1458 CE . The seismites in the "counted interval" were compared to dates in Earthquake Catalogs [Ambraseys et al (1994), Amiran et al (1994), Guidoboni et al (1994), Ben-Menahem (1991), and Russell (1985)]. Relatively minor additional input was also derived from other studies in the region which likely relied on similar catalogs. Some of these catalogs contain errors and a critical examination of where the dates and locations of historical earthquakes reported in these catalogs came from was not undertaken. Migowski (2001) shifted the dates from the under-sampled radiocarbon derived age-depth model to make the floating varve chronology in the "counted interval" match dates from the earthquake catalogs. Without the shift, the dates did not match. This shift was shown in Migowski (2001)'s dissertation and mostly varies from ~200-~300 years. The "counted interval" dates are ~200-~300 years younger than the radiocarbon dates. Some of Migowski's shift was justified. Ken-Tor et al (2001) estimated ~40 years for plant remains to die (and start the radiocarbon clock) and reach final deposition in Nahal Ze'elim. This could be a bit longer in the deep water En Gedi site but 5 to 7.5 times longer (200-300 years) seems excessive. Although uncritical use of Earthquake catalogs by Migowski (2001) and Migowski et al (2004) led to a number of incorrectly dated seismites , the major "anchor" earthquakes (e.g. 31 BC, 1212 CE) seem to be correct.

Neugebauer (2015) and Neugebauer at al (2015) recounted laminae from 2.1 - 4.35 meters in the En Gedi Core (DsEn) while also making a stratigraphic correlation to ICDP Core 5017-1. Nine 14C dates were used from 1.58 - 6.12 m but samples KIA9123 (inside the Late Bronze Beach Ridge) and KIA1160 (the 1st sample below the Late Bronze Beach Ridge) were discarded as outliers. These two samples gave dates approximately 400 years older than what was expected for the Late Bronze Age Beach Ridge - a date which is fairly well constrained from other studies in the Dead Sea. This left 7 samples distributed over ~4.5 m - an average of 1 sample every 0.65 meters - not a lot. Their DSEn varve count, anchored to an age-depth model derived from these 7 samples, produced an average shift of ~300 years compared to Migowski et al (2004)'s chronology (i.e. it is ~300 years older). Although two well dated earthquakes were available to use as time markers (the Josephus Quake of 31 BCE and the Amos Quake(s) of ~750 BCE), they chose not to use earthquakes as chronological anchors (Ina Neugebauer personal communication, 2015). Instead, they used the Late Bronze Age Beach Ridge as evidenced by discarding the two radiocarbon samples. Using the Beach Ridge as a chronological anchor was likely a good decision as the Late Bronze Age Beach ridge is fairly well dated. Their newly counted chronology produced a paleoclimate reconstruction that aligned fairly well with data from other locations . Although paleoclimate proxies are not necessarily synchronous and suffer from greater chronological uncertainty than, for example, well dated earthquakes, the problem with their recount for our purposes does not lie with their relatively good fit to other site's paleoclimate proxies. That is probably approximately correct. The problem is they calibrated their count to the bottom of their counted interval (Late Bronze Age Beach Ridge) but did not have a calibration marker for the top.

In the En Gedi core (DSEn), the Late Bronze Age Beach Ridge (Unit II of Neugebauer et al, 2015) is found from depths 4.35 to 4.55 m. It's top coincides with the bottom of the recounted interval - far away from the overlap (2.1 - 3.02 m) with Migowski's counted interval. Thus, if there were any problems with the recounted dates (e.g. hiatuses or accumulating systemic errors) as one moved to the top of the recounted interval, they would go unnoticed. Varve counts in the overlapped interval were fairly similar - 583 according to Migowski (2001) vs. 518 according to Neugebauer et al (2015). There wasn't a major discrepancy in terms of varve count interpretation. But, the lack of a calibration point near the top of the recounted interval leaves one wondering if the recounted dates in the overlap are accurate and why Migowski's pre-shifted chronology doesn't correlate well with the reliable parts of the earthquake record.

Neugebauer at al (2015:5) counted 1351 varves with an uncertainty of 7.5% (Neugebauer at al, 2015:8). That leads to an uncertainty of ~100 varves by the time one gets to the top of the recounted interval away from the Late Bronze Age Beach Ridge calibration point. The Beach Ridge itself likely has an uncertainty of +/- 75 years. Add the two together and the uncertainty approaches Migowski's shift. In addition, roughly 15% of the recounted interval went through intraclast breccias (seismites) where the varves were uncountable and the varve count was interpolated with a questionable multiplication factor of 1.61 applied to the interpolated varve count (Neugebauer at al, 2015:5). Migowski et al (2004) also interpolated through the intraclast breccias however in her case she used the interpolation to line up with events out of the Earthquake catalogs.

Unfortunately, Neugebauer at al (2015)'s study did not resolve the uncertainties associated with Migowski's varve counts. Both studies lack a sufficiently robust calibration over the entire depth interval. Dead Sea laminae are difficult to count. They are not nearly as "well-behaved" as they are in the older Lisan formation or in Glacial varves. This was illustrated by Lopez-Merino et al (2016). Their study, which used seasonal palynology to ground truth varve counts, showed that between 1 and 5 laminae couplets (ie varves) could be deposited in a year . This study, undertaken in Nahal Ze'elim, represents a worst case scenario. It is essentially impossible to count varves in Nahal Ze 'elim because the site receives too much fluvial deposition which muddies up the varve count (pun intended) compared to the deeper water site of En Gedi. While the conclusions from Lopez-Merino et al (2016) cannot be generalized to the entire Dead Sea, it does point out that Holocene Dead Sea varve counts need to be calibrated to be used in Historical Earthquake studies. The calibration can come through anchor events such as strong earthquakes and/or clearly defined and dated paleoclimate events, seasonal palynology work (determining the season each laminae was deposited in), and/or dense radiocarbon dating - much denser than what is available from the En Gedi core (DESn). There may also be geochemical ways to calibrate varve counts.

In 2018, Jefferson Williams collected ~55 samples of dateable material from an erosional gully in En Gedi (aka the En Gedi Trench) located ~40 m from where the En Gedi Core (DsEn) was taken in 1997 . This erosional gully was not present when the En Gedi core was taken. It developed afterwards due to the steady drop in the level of the Dead Sea which has lowered base levels and creates continually deeper erosional features on the lake margins. Due to cost, these samples have not yet been dated but lab analysis of this material should resolve dating ambiguities in En Gedi. The samples are well distributed in depth (68 - 303 cm. deep) and can be viewed here in the Outcrop Library. Radiocarbon from the En Gedi Core can be viewed here. In the Google sheets presented on the radiocarbon page for the En Gedi Core, Neugebauer's radiocarbon samples and a reconciliation table can be viewed by clicking on the tab labeled Nueg15.

En Gedi Core (DSEn) Photos

Core Depths were measured from surface. The core was taken about a meter above the Dead Sea level which was ~ -411 m in 1997. In 2011, Jefferson Williams measured the elevation of the surface where the En Gedi Core (DSEn) was taken using his GPS. The recorded elevation was -411 m however GPS is less accurate measuring elevation than it is for Lat. and Long. so this depth measurement should be considered approximate.

Image Description Image Description Image Description Image Description
Composite Core
Sections C1, A2, A3, A4

19-397 cm.
Litholog and
Composite Core

47-325 cm.
Litholog
Entire Core

-30 cm.-1022 cm.
Litholog
Legend
Section C1

19-114 cm.
Section A2

114-196 cm.
Section A3

200-296 cm.
Section A4

300-397 cm.
1458 CE Quake

65-80 cm.
1202, 1212, and 1293 CE Quakes

90-115 cm.
1033 CE Quake

131-143 cm.
Thin Section
A3_3_1a

259.7-269.9 cm.
Thin Section
A3_3_2

271.5-273.7 cm.
Thin Section
A3_3_3

273.5-283.5 cm.
Thin Section
A3_4_1

283.3-293.4 cm.
SEM Image
250x Magnification
Sample EG13

Nahal Ze 'elim
ZA1
Kagan et al (2011) in Table 4 reports a 15 cm. thick seismite which they assigned a date of mid second century BC. This is the same seismite which was called Event A by Ken-Tor et al (2001a) and which was redated to ~150 BC by Williams (2004) and Agnon et. al (2006)
ZA-1

Image Description Source
Lithosection with dates Ken-Tor et al. (2001a)
Lithosection with dates - big Ken-Tor et al. (2001a)
Lithosection with dates - really big Ken-Tor et al. (2001a)
Correlated Trench Logs for ZA-1.
Lithosections above are composite lithologs
derived from multiple sites as far as, I think,
300 meters apart. ZA-1 refers to the location
where a large part of the composite
lithosection was derived. More landward gullies
were used to capture the most recent
earthquakes (Events G (1834) and H (1927))
Bookman referred to site ZA-1 as site 2.
Revital Bookman (nee Ken-Tor)
Map of Gullies where Revital Bookman
(nee Ken-Tor) did her work
Revital Bookman (nee Ken-Tor)
Events B (Josephus Quake - 31 BCE)
and C (Jerusalem Quake - 26-36 CE)
at site ZA-1
Jefferson Williams
Thin Section Slide from ZA-1
showing Event C (Jerusalem Quake)
Jefferson Williams

ZA2
Kagan et al (2011) in Table 3 report an 8 cm. thick intraclast breccia at a depth of 516 cm. which could fit this earthquake.

ZA-2

Image Description Source
Age Model Kagan et al (2011)
Age Model - big Kagan et al (2011)
Age Model with annotated dates Kagan (2011)
Age Model with annotated dates - big Kagan (2011)
Annotated Photo of ZA-3
ZA-3 = N wall of gully
ZA-2 = S wall of same gully
Kagan et al (2015)

Arava

On-site fault rupture suggests a minimum moment magnitude MW of 6.5 (Mcalpin, 2009:312).
Taybeh Trench
In paleoseismic trenches near Taybeh, Jordan, LeFevre et al. (2018) dated an earthquake Event (E6) on the Arava Fault to have occurred between 160 BC and 117 BC.

Taybeh Trench Earthquakes
Figure S5

Computed age model from OxCal v4.26 for the seismic events recorded in the trench.

LeFevre et al. (2018)


Taybeh Trench

Image Description Source
Age Model Lefevre et al (2018)
Age Model - big Lefevre et al (2018)
Trench Log Lefevre et al (2018)
Annotated Trench photomosaic Lefevre et al (2018)
Stratigraphic Column Lefevre et al (2018)
Stratigraphic Column - big Lefevre et al (2018)

Qatar Trench
Klinger et. al. (2015) hypothesized that Event E7 in a paleoseismic trench in the southern Arava near Qatar, Jordan may have been caused by an earthquake in ~150 BC or an earlier event (Date Range 338 BC – 213 BC). The Fortress at Arad Quake may be a better fit for this event than the Southern Dead Fish and Soldiers Quake of ~150 BC.

Qatar Trench

Image Description Source
Age Model Klinger et al (2015)
Age Model - big Klinger et al (2015)
Trench Log Klinger et al (2015)
Simplified Trench Log Klinger et al (2015)

Conclusion
Since the earthquake record appears to contain several earthquake couplets (e.g. 750 BC Amos Quakes, 363 AD, 1202/1212 AD) which struck to the north and south of the Dead Sea (Kagan et al, 2011), it is possible there may have been an earthquake couplet with northern and southern epicenters sometime around the middle of the second century BC. This would mean that the paleoseismic data in the Arava and the Dead Sea were likely generated by the southern part of the couplet which may have ruptured the Arava fault. Although there are no known extant textual records for a mid second century BC earthquake on the Arava fault, this should not come as a surprise. Textual sources for earthquakes in the southern Levant are rare before 31 BC – particularly for the sparsely inhabited region south of the Dead Sea.

Notes

Considering Malalas' confused chronology on the Malalas Confusion Quake it is possible that the Dead Fish and Soldiers Quake and the Malalas Confusion Quake refer to the same event. Karcz (2004) speculated on this possibility suggesting Malalas may have misreported the ruling Seleucid King when the alleged Malalas Confusion Quake struck Antioch. Karcz (2004) further suggested that the Seventeenth of Adar Quake may refer to the Dead Fish and Soldiers Quake. Based on the historical sources, Ambraseys (2009) constrained the years of the Dead Fish and Soldiers Quake to between 138 BC and 125 BC while assigning a date of 139 BCE and questioning whether it is a spurious earthquake report. Karcz (2004) used historical sources to constrain the date of the Dead Fish and Soldiers Quake to between 145/144 BCE and in 138/137 BCE and notes that if the Dead Fish and Soldiers Quake and the Seventeenth of Adar Quake are the same event, this event would likely have occurred in 143/142 BCE.

Paleoclimate - Droughts

References