Where Were Sodom and Gomorrah?

Summary

Two ruined sites near the Dead Sea have been announced as Sodom in the last fifty years.  Neither identification holds up.  What follows is the case for each, the case against each, and what a skeptic should and should not conclude from the failure.

• The two leading candidates are about 65 km apart and both are defended by scholars who believe Genesis is history. Bab edh-Dhra and Numeira sit at the southeast corner of the Dead Sea and burned in the Early Bronze Age. Tall el-Hammam sits northeast of the Dead Sea and burned in the Middle Bronze Age. The fight is not believer against skeptic. It is Bryant Wood and Associates for Biblical Research against Steven Collins and Trinity Southwest University, with mainstream archaeology declining to endorse either.
• The southern sites burned about three centuries too early, and the two of them did not burn together. The excavators date Bab edh-Dhra's destruction to about 2350 BC.¹ A conservative reading of biblical chronology puts the destruction of Sodom at 2067 BC.¹ Worse for the theory, the dates published from the cemeteries by Chesson and Schaub in 2007 put Numeira's end around 2600 BC — roughly 250 years before Bab edh-Dhra, not simultaneous with it.² The claim that both cities show a shared destruction, still repeated in popular apologetics, was true of the 1970s interpretation and is not true now. Steven Collins, who has every reason to press the point, presses it: "nearby Numeira went out of operation 250 years before Bab edh-Dhra."³
• The northern site burned about three centuries too late, and it was rebuilt. Collins dates the terminal destruction of Tall el-Hammam to 1750–1650 BC.⁴ To make that Sodom he has to compress biblical chronology hard: a 215-year sojourn in Egypt instead of the 430 of Exodus 12:40, and patriarchal lifespans read as honorific rather than literal.⁴ And Tall el-Hammam was occupied again in Iron Age II, around 1000 BC, which collides with the prophets who use Sodom as the type of a place never inhabited again — "no Arab will pitch his tent there" (Isaiah 13:20).⁴
• The famous 2021 airburst paper was retracted on 24 April 2025. Scientific Reports pulled "A Tunguska sized airburst destroyed Tall el-Hammam" with a plain statement: "the claims that an airburst event destroyed the Middle Bronze Age city of Tall el-Hammam appear to not be sufficiently supported by the data in the Article. Given these concerns, the Editors no longer have confidence that the conclusions presented are reliable."⁵ Eleven of the twenty-one authors formally disagreed with the retraction.⁵
• The retraction was not a fashion change. Two specific technical failures killed it. Jaret and Harris showed the "shocked quartz" lacked the features that define shocked quartz — the lamellae "pinch and swell in thickness, and have irregular spacing," where real planar deformation features are planar, evenly spaced, and constant in thickness.⁶ They also showed the platinum measurements were made by a method known to overestimate platinum "by up to a factor of 4," at an accelerating voltage too low to separate the peaks at all.⁶ Boslough and Bruno then showed the Tunguska comparison itself rested on numbers inflated through a citation chain that "included non-peer-reviewed or unscientific sources."⁷
• The paper had already claimed Genesis as corroboration. Its authors wrote that the Genesis account "is consistent with having been an eyewitness account of a cosmic airburst."⁸ That sentence, published in a Nature-family journal, is what carried the story worldwide. It is now attached to a retracted paper. The team has republished an expanded version in Airbursts and Cratering Impacts, a journal run by the Comet Research Group, whose director is the retracted paper's corresponding author.⁹
• The Ebla tablets do not mention Sodom. Giovanni Pettinato announced in October 1976 that the Ebla archive named Sodom, Gomorrah and Zoar. Alfonso Archi, his successor as Ebla epigrapher, showed the readings did not hold, and Pettinato himself withdrew the Zoar reading. The settled verdict, in a standard survey: "The present consensus is that Ebla has no bearing on the Minor Prophets, the historical accuracy of the biblical Patriarchs, Yahweh worship, or Sodom and Gomorrah."¹⁰ This claim is still circulating in apologetics fifty years later. It should be dropped.
• The text's own geography pulls south, and the one verse that pulls north is the weakest verse in the file. Genesis 14:3 puts the battle "in the Valley of Siddim (that is, the Salt Sea)."¹¹ Genesis 14:10 puts tar pits there, and bitumen is a southern Dead Sea product. Ezekiel 16:46 puts Sodom south of Jerusalem, in a verse that puts Samaria north of it.¹² Josephus in the first century says the lake "is extended as far as Zoar in Arabia. The country of Sodom borders upon it," and the sixth-century Madaba map puts Zoar at the southern end.¹³¹ Against all of that, Collins has Genesis 13:10–12, where Lot looks over "the whole region of the Jordan" and travels east. Even Walter Rast, who dug Bab edh-Dhra, made the point without a dog in the northern fight: "the directions and locations in Genesis 13 are the most general and obscure of all the texts about Sodom."¹⁴
• But the northern case is not silly, and its best argument is one of scale. Tall el-Hammam is a very large Bronze Age city — on Collins's figures, over 100 acres of occupational footprint, with more than 60 acres inside its defenses.⁴ Bab edh-Dhra covers roughly 10 acres. Genesis 10:19 and 13 treat the Cities of the Plain as a recognizable regional entity, and Genesis 14 has their kings mustering armies. Collins's question is fair: if the biblical writers knew the lower Jordan valley at all, why would they describe a cluster of ten-acre villages and never mention the dominant city of the region?³
• The excavator's own position, now read directly, is more interesting than either side reports. Walter Rast rejected a flatly historical reading — his essay "rejects an understanding which overlooks their sagalike representation" — and then wrote that "encapsuled in the formula of two cities experiencing memorable destruction was a fragment of local recollection of great importance."¹⁵¹⁵ Noth, von Rad, Westermann, Wallis and de Vaux all independently derive the material from an old pre-Israelite Dead Sea tradition.¹⁵ The mainstream position is not that Genesis 19 was invented late; it is that a very old local memory of a destroyed city was absorbed into Israel's ancestral narrative.
• The "did Bab edh-Dhra burn?" question is settled, in between the two camps. Rast: "At Numeira the evidence is extensive that the town was burned during its last major phase, while the destruction of the EB III walled town at Bab edh-Dhra is attested by ruined and abandoned structures and by the demolished upper part of the town wall, the superstructure of which … had suffered exposure to fire."¹⁵ Neither "no significant fire" nor "an enormous conflagration" is accurate. And Rast filed the objection to cosmic-catastrophe readings of Genesis 19 in 1987, endorsing a comparison of them to "attempts to explain biblical events by means of extraterrestrial visits."¹⁵
• Honest weaknesses on our own side. The 2067 BC date that rules out Tall el-Hammam is itself derived, not measured; it depends on a 1446 BC Exodus and a 430-year sojourn, and this project's own article on dating the Exodus concludes that the Exodus cannot presently be dated. The absolute date for the end of Early Bronze III has no Egyptian synchronism at all and is, in Wood's own words, "strictly an educated guess."¹ Kris Udd's Andrews University dissertation on exactly this question concludes only that the southern identification is "a possibility" that merits "further investigation" — which is the honest ceiling.¹⁶ And the strongest apologetic detail in the older literature, the paired destruction of two neighboring cities about twenty years apart matching Genesis 14 and 19, has been undercut by the site's own later radiocarbon work.

Verdict: Sodom has not been found.  No excavated site carries an inscription naming it, and every proposed identification requires either the archaeology or the biblical chronology to move by about three centuries.  That is not evidence that Genesis 19 is fiction — the region contains real Bronze Age cities that really burned, sitting on a real fault line above real petroleum and sulfur deposits, and the biblical tradition about them is geographically specific and old.  But an honest reader should notice that the two most confident claims made in the last twenty years, the Ebla tablets and the Tall el-Hammam airburst, have both been withdrawn by the scientific process rather than by their critics.  Anyone still citing either one has not checked.

What the text actually requires

Before weighing sites, it is worth writing down what a candidate has to satisfy.  Bible quotations here are from the NET Bible.

A cluster of cities, not one city.  Genesis 14:2 names five kings: Sodom, Gomorrah, Admah, Zeboyim, and "the king of Bela (that is, Zoar)."¹⁷  So the site must sit in a group.

Near the Salt Sea, in a valley with bitumen.  "These last five kings joined forces in the Valley of Siddim (that is, the Salt Sea)."¹¹  "Now the Valley of Siddim was full of tar pits."¹⁸

Well watered and farmed.  "all of it was well watered … like the garden of the Lord, like the land of Egypt, all the way to Zoar."¹⁹

Fortified, with a gate.  "Lot was sitting in the city's gateway."²⁰

Destroyed by fire from above, along with its farmland.  "Then the Lord rained down sulfur and fire on Sodom and Gomorrah … So he overthrew those cities and all that region, including all the inhabitants of the cities and the vegetation that grew from the ground."²¹

Visible as smoke from Abraham's vantage point.  "He looked out toward Sodom and Gomorrah and all the land of that region.  As he did so, he saw the smoke rising up from the land like smoke from a furnace."²²  Abraham was at Mamre, near Hebron.

Zoar survives.  Lot flees there and it is spared.  Zoar then stays on the map: Isaiah 15:5, Jeremiah 48:34, Josephus, Eusebius, and the Madaba mosaic.¹  This is the single most useful constraint in the file, because Zoar is the one city of the five that later sources can still point to.

And Zoar sits at the southern limit of the view from Mount Nebo.  Deuteronomy 34:1–3 has Moses shown the land from Nebo — "the Negev, and the plain of the Valley of Jericho, the city of date palm trees, as far as Zoar."²³  Zoar marks the far end of the panorama.  Ghor es-Safi, at the southern end of the Dead Sea, is roughly 50 km south of Nebo and fits;  Tall el-Hammam lies north-west of Nebo and does not.  This article had not previously stated this constraint, and the northern proposal has to absorb it.

The land stays ruined.  "The whole land will be covered with brimstone, salt, and burning debris;  it will not be planted nor will it sprout or produce grass."²⁴  Later prophets use Sodom as the type of permanent desolation.

A date, if the biblical numbers are taken serially.  On a 1446 BC Exodus and a 430-year sojourn, Abraham lives 2166–1991 BC and the cities fall in 2067 BC.¹  That figure is a derivation, not a datum, and it is the pivot on which the whole archaeological argument turns.

Part 1 — The southern candidate: Bab edh-Dhra and Numeira

What is there

William Albright surveyed the southeast Dead Sea plain in 1924 looking for the Cities of the Plain.²⁵  Bab edh-Dhra was excavated by Paul Lapp in the 1960s and then, from 1973, by Walter Rast and R. Thomas Schaub, whose survey found four more Early Bronze sites strung south of it — Numeira, Safi, Feifa and Khanazir.  Five sites, matching five cities, in exactly the region the Valley of Siddim texts point to.  That correspondence is what made the identification attractive, and it is still the strongest single argument the southern theory has.

Bab edh-Dhra is a walled town with a gate, a temple, and a very large cemetery whose latest phase buried the dead in above-ground mudbrick charnel houses.  Wood reports that five excavated charnel houses were burned, and that in the largest, A22, the excavation showed the fire started in the roof and spread inward when the roof fell — which he reads as fire falling from above.¹

The site sits on the eastern fault of the Jordan rift.  Bitumen seeps and sulfur are present in the southern Dead Sea basin, and Dead Sea bitumen has been chemically matched to asphalt used in Egyptian mummification, showing a working Early Bronze trade in the stuff.¹  Paleobotany at Bab edh-Dhra and Numeira found barley, wheat, grape, figs, lentils, flax, and less commonly chickpeas, peas, broad beans, dates and olives — several of which require irrigation, which matches "well watered."¹

The southern basin of the Dead Sea, where the Valley of Siddim would have been, was dry in the Early Bronze Age and flooded later — which is what Genesis 14:3 says, describing a valley that "is the Salt Sea."¹

What is wrong with it

The date.  The excavators put the end of Bab edh-Dhra at 2350 BC (Rast) or 2300 BC (Schaub).¹  The biblical figure, on the conservative reckoning, is 2067 BC.  That is a gap of roughly 250 to 300 years.

Wood's answer is that the absolute date for the end of Early Bronze III is not actually known.  Bronze Age dates in the Levant depend on synchronisms with Egypt, and for EB III there are none: "no synchronisms have yet been found for the EB III period."  The preceding EB II period is tied loosely to Egypt's Archaic Period, itself known only to within 200 years, and how the intervening 800 years should be split between EB III and EB IV is "strictly an educated guess."¹  Published end-dates for EB III in three standard reference works range from 2350 to 2200 BC.¹

This is a real point and it is fair.  It is also a weak kind of argument, because it works by widening the error bars rather than by producing a date.  Kris Udd's 2011 Andrews University dissertation is the most careful attempt to run it to ground, examining both the Mesopotamian and Egyptian chronologies and the radiocarbon record.  His finding was mixed: chronologies "have been steadily lowered over the past century," which helps, but "Radiocarbon was found to generally favor higher dates," which hurts.  His conclusion is deliberately modest — the combination "indicates the possibility that Bab edh-Dhra' and Numeira could be two of the biblical cities of the plain.  Further investigation into this possible identification is merited."¹⁶  That is a long way from a discovery.

The two cities did not burn together, and this used to be the best argument.  Genesis 14 has the cities sacked by four kings, and Genesis 19 has them destroyed some years later.  The older southern literature made much of two matching destructions about twenty years apart at Bab edh-Dhra and Numeira, and of similar pottery in the top layers at both sites putting their ends close together.  The later cemetery dates published by Chesson and Schaub in 2007 separate the two destructions by about 250 years, with Numeira ending around 2600 BC and Bab edh-Dhra around 2350.²  Collins, arguing from the other side, states the gap flatly.³  Any presentation of the southern case that still claims a shared destruction horizon is out of date.

The size.  Bab edh-Dhra covers roughly 10 acres.²⁶  Genesis 10:19 lists Sodom in a passage that also lists Babylon, Akkad and Uruk;  Genesis 14 has its king leading an army.  Collins's objection lands: "How could a small, marginally-existing EB city like Bab edh-Dhra be Sodom which is listed along with the great Mesopotamian urban centers such as Babylon, Akkad and Uruk (Erek) in Genesis 10?"³

Whether Bab edh-Dhra even burned — now answered.  Wood publishes photographs of ash layers and describes burned charnel houses;  the standard reference summary says Bab edh-Dhra, unlike Numeira, does not appear to have been destroyed by a significant fire.  This article previously flagged the disagreement without adjudicating it.  Rast's own 1987 essay settles it, and the answer is in between:

At Numeira the evidence is extensive that the town was burned during its last major phase, while the destruction of the EB III walled town at Bab edh-Dhra is attested by ruined and abandoned structures and by the demolished upper part of the town wall, the superstructure of which, made of mudbrick in some cases, had suffered exposure to fire.¹⁵

Read that carefully, because both sides have overstated it.  Numeira burned;  that is not in doubt.  Bab edh-Dhra was destroyed — ruined structures, a demolished wall — and part of the wall superstructure shows fire exposure.  That is a real destruction with a fire component.  It is not "no significant fire," and it is not an "enormous conflagration" consuming the town either.  Anyone leaning on "a fiery destruction at Bab edh-Dhra" should quote Rast's sentence rather than paraphrase it.

And the same paragraph shows exactly when the paired-destruction argument was still live.  Rast wrote in 1987 that the two towns "seem to have participated in a common fate, since ceramic evidence from carbon-14 tests have shown that they were terminated at approximately the same time toward the end of EB III."¹⁵  So this was the excavators' own position on the then-available radiocarbon, honestly held.  The cemetery dates published twenty years later separated the two by about 250 years.  The argument was not careless when it was made;  it is out of date now, and the twenty-year lag is the whole point.

The claim as it is still being made

It is worth showing that the superseded version is not a straw man.  Associates for Biblical Research, the organisation that has done the most to promote the southern identification, still carries an article by Bryant Wood which opens:

The ruins of Sodom and Gomorrah have been discovered southeast of the Dead Sea.  The modern names are Bab edh-Dhra, thought to be Sodom, and Numeira, thought to be Gomorrah.  Both places were destroyed at the same time by an enormous conflagration.²⁷

The Chesson and Schaub cemetery dates that separate the two destructions by about 250 years were published the year before that article was posted.²  This is not a criticism of anyone's honesty;  it is an illustration of how long a retired argument stays in circulation, which is the point the method article makes about the whole genre.

One argument in the same piece deserves better treatment than the rest.  The geologist Frederick Clapp proposed that an earthquake could force subterranean bitumen — sulfur-rich — out along a fault, where it could ignite and fall back as burning material.  Wood's point is that Clapp formulated this before the southern sites were located, and the sites turned out to lie on the fault line.²⁸

That has the shape of a real prediction, which is rare in this field, and it deserves to be recorded as one.  Two limits.  Lying on a fault line in the Jordan rift valley is not a demanding prediction — the whole rift is a fault system.  And the mechanism explains the manner of destruction described in Genesis 19 without doing anything to establish that these particular ruins are Sodom, which is the question at issue.

And one sentence in it should not be repeated.  Of the burned charnel houses, Wood writes: "The only conceivable explanation for this unique discovery in the annals of archaeology is that burning debris fell on the buildings from the air."²⁷  Roof fires have other causes — a roof is the most flammable part of a mudbrick building and the first thing an attacker fires.  "Only conceivable" is the kind of claim that this article's own diagnostic-versus-consistent test exists to catch, and it should be caught on our own side as readily as on the airburst paper's.

The excavators' own period summary

The Rast and Schaub cemetery volume states the plain's occupational range plainly, and it is worth having from them rather than from either side's summary:

Material remains from the Early Bronze Age are found in various degrees at Bab edh-Dhra', Numeira, es-Safi, Feifa, and Khanazir … although they do not always represent the same phase.  A few sherds dating to the Middle Bronze Age have been cited … but so far they do not constitute evidence of substantial settlement during this period.  Up to the present no Late Bronze has come to light, but Iron II sherds were found at es-Safi, Feifa, and Khanazir.²⁹

Two things follow.  The southern sites really are Early Bronze and nothing else — the "they do not always represent the same phase" clause is the one that matters, and it is the excavators conceding, in 1989, the point that the later cemetery dating would sharpen.  And there is no Middle Bronze option in the south, which means the northern and southern proposals are not competing readings of the same evidence;  they are looking at different periods entirely.

On Zoar, the one city that was supposed to survive.  The same volume records that Nabataean and late Roman remains at Rujm en-Numeira and in the Ghor es-Safi "have been connected with references in contemporary sources to the Roman fort of Zoora," that sixth-century Hebrew tombstones at Ghor es-Safi attest a Jewish settlement there, and that a Cairo Geniza document refers to Zoar.²⁹  So the historical Zoar at es-Safi is well attested from the Roman period onward — and the Bronze Age levels beneath it, which are what would actually test Genesis 19, remain the unexplored question this article's closing section names.

What the excavators themselves concluded

Rast and Schaub did not claim to have found Sodom, and Rast's actual position is more interesting than either side's summary of it.

He rejects a flatly historical reading.  His essay opens by saying that "the perspective on the texts adopted here rejects an understanding which overlooks their sagalike representation," and that what the Sodom material represents "was often not historical but rather theological or social in its intent."¹⁵

And then he does not stop there.

On the other hand, it will be proposed below that encapsuled in the formula of two cities experiencing memorable destruction was a fragment of local recollection of great importance for the history of the occupation in the Dead Sea region.¹⁵

"A fragment of local recollection of great importance" is a long way from "invented." Rast thinks there is real memory in the story.

He is also not alone, and the convergence is worth setting out, because it is a line of critical scholars arriving independently at a pre-Israelite Dead Sea origin for the tradition.  Noth called it "the fundamentally non-Israelite story of Sodom";  von Rad wrote that "that in the case of the story of Sodom we are dealing with an originally independent saga has long been recognized";  Westermann derived it from a "local tradition from the region of the Dead Sea";  Wallis was blunter — "at the heart of the narrative certainly lies a very old pre-Israelite tradition about the destruction of a city called Sodom";  and de Vaux spoke of "traditions originating in Transjordan which have been integrated into the history of Abraham."¹⁵

What a conservative reader should take from this.  The mainstream position is not that Genesis 19 was made up in the Persian period.  It is that a very old local memory of a destroyed city near the Dead Sea was absorbed into Israel's ancestral narrative.  That concedes an ancient core and denies the narrative's historicity as written, and both halves are real.

One more thing in Rast's essay deserves quoting, because it anticipates a paper published thirty-four years later.  On attempts to explain Genesis 19 by geological catastrophe, he cites Keel approvingly for "comparing such a hybridization of the texts and theories of geologic upheaval to recent attempts to explain biblical events by means of extraterrestrial visits," and adds that importing lost-city stories from other cultures as parallels "seems to be based on the misunderstandings of cataclysmic interpretation."¹⁵  His own explanation of why this destruction became memorable while others did not is environmental: the story "helped to explain a degenerating environment around the Dead Sea, which became serious in the latter quarter of the third millennium."¹⁵  A reader who has followed Part 3 will notice that the excavator of the southern sites had filed the objection to cosmic-catastrophe readings of Genesis 19 in 1987.

Part 2 — The northern candidate: Tall el-Hammam

The case

Tall el-Hammam sits in the lower Jordan valley about 12 km northeast of the Dead Sea.  Nelson Glueck surveyed it in 1941;  Kay Prag surveyed it for the British Institute at Amman in 1975–76 and again in 1990.  Steven Collins of Trinity Southwest University has excavated it since 2005.

Collins's argument runs on four criteria — geography, chronology, stratigraphy and architecture.¹

Geography.  Genesis 13 has Abram and Lot separating in "the environs of Bethel/Ai," and Lot looking out over "the whole region of the Jordan" before traveling east.  Collins reads kikkar hayarden as a technical term for the alluvial disk of the Jordan north of the Dead Sea, and argues it cannot extend past the Jordan's mouth.  In his own words:

The descent of the living water ends when the descent itself ends: at the N end of the Dead Sea.  The living hayarden "dies" as soon as it enters the dead waters of the "Arabah Sea."  Thus, the Descent = Jordan cannot pass beyond the physical termination at the lowest spot on the face of the earth, the surface of the Dead Sea.³

Scale.  Over 100 acres of Bronze Age occupation, more than 60 of them fortified — on Collins's account, "the largest continuously-occupied city in the southern Levant" averaged over its 3,000-year history.⁴

Stratigraphy.  A violent terminal destruction — "melted pottery, scorched foundation stones and several feet of ash and destruction debris" — followed by a long abandonment.⁴

Architecture.  Middle Bronze rampart fortifications, satisfying Genesis 19:1's city gate.

The case against

The date, again, but in the other direction.  Collins puts the terminal destruction at 1750–1650 BC.⁴  Standard conservative chronology puts Sodom at 2067 BC.  To close a 300–400 year gap Collins lowers Abraham, and the way he does it is the real problem.  He adopts a 215-year Egyptian sojourn against the 430 years stated in Exodus 12:40, and he treats the patriarchal lifespans as honorific rather than literal.⁴  Simon Turpin's assessment is blunt: "These views are not the result of careful biblical exegesis but rather are the result of eisegesis and special pleading."⁴

A reader who does not share Turpin's inerrantist commitments should still notice the structure of the move.  Collins is not deriving a date and then finding it matches a site.  He has a site, and he adjusts the chronology until it fits.  Wood does something formally similar in the opposite direction — widening the EB III error bars until 2067 becomes reachable.  Both operations are legitimate as hypotheses.  Neither is evidence.

The reoccupation.  Tall el-Hammam was inhabited again in Iron Age II, about 1000 BC.⁴  Isaiah, Jeremiah and Zephaniah all use Sodom as the standard of a place that stays empty: Babylon "will never be inhabited or lived in for all generations;  no Arab will pitch his tent there" (Isaiah 13:19–20).  Collins answers that the simile applies only to the destruction, not to its duration.⁴  That reading is available, but it drains the comparison of the very feature the prophets are using it for.

The southern geography in the rest of the Bible.  Genesis 14:3 identifies the battlefield with the Salt Sea, not the Jordan plain.  Genesis 14:10 puts tar pits there.  Ezekiel 16:46 places Sodom south of Jerusalem in a verse that places Samaria north of it — and Tall el-Hammam lies east-northeast of Jerusalem, not south.¹  Bill Schlegel adds the textual point that Genesis 13:11–12 implies movement over time: Lot chose the kikkar of the Jordan and then "pitched his tent as far as Sodom," which reads as a separation between the two, not an identification.³⁰

Abraham's line of sight.  Genesis 19:28 has Abraham, at Mamre near Hebron, seeing smoke over the land of the plain.  The southern basin is visible from the Hebron highlands.  Tall el-Hammam is not.  This cuts against the north, and it is the mirror image of Collins's own best point that Bab edh-Dhra is not visible from Bethel/Ai.  Each side has a sightline verse the other cannot satisfy.

Kikkar.  Collins takes the word to mean a circular disk.  Wood answers that the Jordan plain north of the Dead Sea is not circular but rectangular in shape;  that the plain there is called the kikkar of the valley of Jericho in Deuteronomy 34:3, not the kikkar of the Jordan;  and that a different kikkar of the Jordan lay between Zarethan and Succoth, far to the north (1 Kings 7:46).  He notes that kikkar is also used of a loaf of bread and of a talent of metal, which suggests the sense transferred to a plain is flatness rather than roundness.¹

Excavation practice.  A 2016 Oxford team recorded significant disruption of the mound from the excavation itself, and archaeologists have raised the concern that publicly branding a site as Sodom stimulates looting and the antiquities trade.³¹  This is a professional criticism rather than an evidential one, but it belongs in an honest account.

How the argument is conducted

It is worth showing the register in which this dispute is carried on, because it is part of the evidence about how much confidence anyone should extend.  Collins, in public comment:

All of this is absolutely incontrovertible.  The biblical text is clear.  The geography and topography are clear.³

and, of the southern proposal:

The Sodom/Bab edh-Dhra scenario is categorically nonsensical, even silly.³

A site identification with no inscription, resting on a contested reading of one Hebrew noun and a chronology the author had to revise to make it work, is not incontrovertible.  Certainty of that grade in the absence of a name on a potsherd is itself a reason to discount.

Part 3 — The airburst paper and its retraction

What was claimed

In September 2021, twenty-one authors published in Scientific Reports the claim that a cosmic airburst larger than the 1908 Tunguska event destroyed Tall el-Hammam at about 1650 BC.  They reported 26 radiocarbon dates from the destruction layer, converging on 1650 BC ± 50, consistent with a pottery seriation date of about 1750–1650 BC.⁸  They described a distinctive "blow-over layer" at the top of the destruction debris, found on the northeast sides of southwest-facing wall remnants — that is, deposited by a blast wave with a direction.⁸  They argued the event raised soil salinity, ended agriculture, and caused a 300–600 year abandonment of about 120 settlements within a 25 km radius.⁸

And they invoked Genesis explicitly:

The description in Genesis of the destruction of an urban center in the Dead Sea area is consistent with having been an eyewitness account of a cosmic airburst, e.g., (i) stones fell from the sky;  (ii) fire came down from the sky;  (iii) thick smoke rose from the fires;  (iv) a major city was devastated;  (v) city inhabitants were killed;  and (vi) area crops were destroyed.⁸

That paragraph, in a journal carrying the Nature imprint, is why this story reached every newspaper in the world.

Why it failed

Two published comments took it apart.

Jaret and Harris (2022) on the mineralogy.  Their central point is a methodological one that generalizes well beyond this case: only certain kinds of evidence can diagnose an impact, because everything else has ordinary explanations.  Their objection to the shocked quartz is specific and checkable:

Specifically, the "lamellae" in the grains are not planar, and instead pinch and swell in thickness, and have irregular spacing.  These are in direct contrast to shocked quartz produced in nature, nuclear explosions, and laboratory experiments which have planar morphologies, uniform spacing, and do not change thickness.⁶

They add that of twelve indexed lamellae, "all but two … have orientations common to frequently documented twin planes in tectonically or thermally stressed quartz."⁶

On the geochemistry, they note that the paper reported only platinum and iridium, where a genuine meteoritic signature requires the full suite of platinum-group elements to match a known meteorite class;  and that the measurement technique used, energy-dispersive spectrometry, "has been shown to be inadequate and overestimates Pt abundances by up to a factor of 4," and at the 10 kV accelerating voltage used "there is not enough energy to separate the PGE peaks in EDS."⁶

Their deepest point is about context.  Ultra-high-temperature melt features are used to identify impact glass because nothing else on a landscape gets that hot — but that reasoning breaks down at an archaeological site, where humans were smelting and firing pottery.  Much of the melt glass in the paper "is closely associated with pottery;  many specimens are pieces of remelted ceramics like one might find in spoils and dumps."⁶

Boslough and Bruno (2025) on the physics.  Mark Boslough is an airburst physicist whose own work was cited in the original paper.  His comment argues that the Tunguska comparison was built on inflated figures inherited through a citation chain:

Bunch et al. make several incorrect statements about the Tunguska event and about airbursts in general.  Collectively, these errors have led to assertions of evidence that do not appear to be supported by the data, and to conclusions that are not factually supported.  Some of these mistakes come from literature in which the primary sources have been misunderstood, exaggerated, and/or misquoted.  In other cases, misinformation was introduced in a chain of citations that included non-peer-reviewed or unscientific sources.  A Tunguska sized airburst cannot possibly generate the claimed temperature or wind speed effects on the ground from an air blast over the distances suggested by Bunch et al. and required to support their conclusions.⁷

Image integrity.  Elisabeth Bik, who investigates figure manipulation for a living, flagged altered images on PubPeer;  a correction in May 2023 republished originals and fixed a figure that Boslough said had been mislabeled and misinterpreted.³²  Boslough's summary of the affair: "Their evidence simply didn't support their conclusions, and they withheld most of the data that they claimed provided the basis for many of their graphs, maps, and age interpretations."³³

The retraction

Scientific Reports posted an editor's note on 15 February 2023 and retracted the paper on 24 April 2025.  The note is short and worth quoting whole, because it says exactly what it found:

Following publication of this Article concerns have been raised by Jaret & Harris about errors in methodology, analyses, and interpretation of the mineralogic and geochemical data in the Article.  In addition, Boslough & Bruno show that comparisons between the Tall el-Hammam site and the Tunguska event are not adequately substantiated, as a result of errors propagated from the original sources that overestimated the temperature, wind speeds, and impact of the air blast at Tunguska.

Therefore, the claims that an airburst event destroyed the Middle Bronze Age city of Tall el-Hammam appear to not be sufficiently supported by the data in the Article.  Given these concerns, the Editors no longer have confidence that the conclusions presented are reliable.⁵

Eleven authors formally disagreed;  nine did not respond to the editors;  the first author had died.⁵

What the retraction does and does not settle

It does not show Tall el-Hammam was never destroyed.  It plainly was: the ash, the melted pottery and the abandonment are in the ground, and no critic disputes the destruction layer.  What the retraction removes is the cause, and with it the only thing that made the site look like Genesis 19 in particular rather than like any of the dozens of Bronze Age cities in the Levant that burned.  Cities in that period were destroyed routinely, by armies and earthquakes.  A destroyed Middle Bronze city near the Dead Sea is not a discovery about Genesis;  it is the ordinary condition of the archaeological record.

It also does not settle the physics for good.  The authors have republished an expanded version, and defenders point to newer mineralogical work at Tunguska itself.  But that republication went into Airbursts and Cratering Impacts, a journal the Comet Research Group launched and whose journal coordinator is Allen West — the retracted paper's corresponding author and the group's director.⁹  A result that could not survive review at Scientific Reports and reappears in a journal run by its own authors' organization has not gained support;  it has changed venue.

Part 4 — Claims that should be retired

"The Ebla tablets mention Sodom and Gomorrah."  Pettinato announced this at the Society of Biblical Literature meeting on 29 October 1976 and repeated it in a 1977 survey article, based on tablets whose facsimiles had not been published for anyone else to check.  Archi contested the readings in 1979 and 1981, arguing among other things that "Ebla place names do not easily correspond to the Bible names or the 'cities of the plain.'"¹⁰  Pettinato withdrew the Zoar reading.  Mark Chavalas's survey states the outcome: "The present consensus is that Ebla has no bearing on the Minor Prophets, the historical accuracy of the biblical Patriarchs, Yahweh worship, or Sodom and Gomorrah."¹⁰  Chavalas adds the part that should sting: "Many of these preliminary analyses came into the popular press as truth, and thus great amounts of misinformation leaked to the public."¹⁰

"Bab edh-Dhra and Numeira were destroyed at the same time, matching Genesis 14 and 19."  Superseded by the site's own later dating, as above.

"A Tunguska-class airburst destroyed Sodom."  Retracted.

Each of these was, in its day, the flagship argument.  Each is still in circulation.  A reader deciding how much to trust apologetic archaeology should weigh not just whether the claims were wrong but how long they took to be withdrawn from popular use.

Part 5 — The mainstream position, stated fairly

Most academic archaeologists and Hebrew Bible scholars hold that Genesis 19 is an etiological narrative: a story that explains a strikingly ruined and lifeless landscape — the salt, the bitumen, the pillars of the Mount Sedom salt dome, the visible tells of dead Bronze Age towns — by attributing it to divine judgment.  Rast, who dug the southern sites, framed his own work in something close to these terms, and his own version of the etiology is specific: the tradition "helped to explain a degenerating environment around the Dead Sea, which became serious in the latter quarter of the third millennium."¹⁵

This position has real strengths.  It explains why the tradition is geographically vague at the crucial point.  It explains why no city name has ever turned up.  It explains why later writers, from the prophets to Josephus, describe the region's condition rather than the event.  Josephus in the first century AD is doing visible etiology himself: he reports that "the traces [or shadows] of the five cities are still to be seen, as well as the ashes growing in their fruits," and concludes that the story "hath these marks of credibility which our very sight affords us."¹³  He is reasoning from a landscape back to a story, which is exactly the mechanism the etiology thesis proposes.

Its weakness is that etiology is unfalsifiable in the direction its proponents use it.  Any real event that left ruins would also produce a landscape people explained;  "this is an etiology" is compatible with the story being true, and is usually asserted rather than demonstrated.  The honest statement is that the etiological reading is the best available explanation given that no site has been identified, and that identifying a site would immediately displace it.

Common objections

"Archaeologists have found Sodom, so Genesis is confirmed."  No.  Two sites have been proposed and neither is established.  The evidence for the specific identification of either one is weaker than the confidence with which both have been announced.

"Nothing has been found, so Genesis 19 is a myth."  This overreaches in the other direction.  The setting is right in a way that is hard to fake: an unusual, bitumen-and-sulfur-bearing rift valley with a salt sea, containing real fortified Bronze Age towns that really burned, and a surviving town, Zoar, at the point the text says one survived.  A late invention has no particular reason to place its city cluster in exactly the geologically anomalous corner of Canaan, or to say that the valley "is the Salt Sea," which describes a basin that was dry earlier and flooded later.

"Doesn't the retracted airburst paper prove creationists will believe anything?"  The paper was not creationist.  Its authors were mostly geologists and physicists in the Comet Research Group, publishing in a mainstream journal, and it passed that journal's peer review.  One member of the excavation team was an author.  The failure here was a general one about extraordinary claims and diagnostic evidence, not a religious one.  It should be said clearly, though, that the paper's Genesis paragraph was what made it famous, and that Christian outlets amplified it far past what its data could carry.

"The Bible's chronology is flexible enough to reach either site."  It is flexible enough to reach 1650 BC only by shortening the sojourn to 215 years and reading the patriarchal lifespans as honorific.  It reaches 2350 BC only by widening the Early Bronze error bars beyond what any Egyptian synchronism controls.  Both moves are possible.  Doing both at once — which is what the field as a whole is doing — should tell a reader that the chronology is not currently constraining anything.

"Tall el-Hammam is far bigger, so it must be Sodom."  Size is a real argument for the site being important, and Collins is right that a Genesis writer who knew the region should have known about it.  But size is not an identification, and the site's own excavator's chronology puts its end centuries after any straightforward biblical date.

What would settle it

• An inscription. A seal, tablet or ostracon naming the city. This is how Ekron, Dan and Gath were settled, and it is the only class of evidence that closes an identification. None has been found at any candidate.
• A securely dated Bronze Age Zoar — and this article previously framed the question wrongly. It said the Bronze Age levels under historical Zoar were unexplored. They are not: Konstantinos Politis surveyed and excavated the Ghor as-Safi from 1997 to 2018, published as Ancient Landscapes of Zoara I and II.³⁴ That work established the Roman and Byzantine Zoara — two churches, one dated 571/2 by an inscription naming Bishop Petros, an early Christian bishopric, the Sanctuary of Lot above the town, and independent attestation in the first-century Babatha archive and on the Madaba map.³⁵ The identification of es-Safi as historical Zoar is secure. The Bronze Age question is what remains, and there the datum is Rast and Schaub's: es-Safi is one of the five plain sites with Early Bronze material, and also yielded Iron II sherds.²⁹ So the town the Bible says survived does have Early Bronze occupation on the same plain as the two that burned. What would settle it is a published Early Bronze stratigraphy at es-Safi showing occupation continuing through the horizon at which Bab edh-Dhra and Numeira end. That is a narrower and answerable question than this article had posed — and somebody is now digging it. Titus Kennedy lists "Zoar (Ghawr al-Safi) Renewed Excavations 2025" in the Near East Archaeological Society Bulletin 70 (2025).³⁶ This project has not obtained the article and states nothing about what it reports. A reader following this question should start there, and this article should be revisited when the report is in hand. Attempts to obtain it have failed: the NEASB has no DOIs, is absent from Crossref, OpenAlex and IA Scholar, and the society's own issue page is behind a members-only login. Two cautions to carry when it does surface. Kennedy is the NEASB's editor and the society's president is Scott Stripling, so the review is not independent of the people whose conclusions the report bears on.³⁷ And an excavation run by someone who expects to find continuity at Zoar is exactly the excavation whose stratigraphy most needs checking by someone who does not.
• An Egyptian synchronism for the end of Early Bronze III. This would remove Wood's escape route or confirm it. At present the period floats.
• Directional blast evidence that survives review. The "blow-over layer" at Tall el-Hammam is a genuinely interesting observation about deposition direction, independent of what caused it. It deserves testing by people who are not committed to the airburst.

Until one of those lands, the correct answer to "where was Sodom?" is that nobody knows.

Sources and further reading

The literature reviewed for this article, including sources considered and not used, is in literature-review.md.