The Famous 1200 BCE Collapse Has No Single Date. Tree Rings Show Why.

Hattusa, the Hittite capital, was abandoned "around 1200 BC", says a summary of the tree-ring work at Gordion [1]. Ugarit was destroyed "c. 1190", says another account [3]. Textbooks round both to "1200". I wanted to know what each number is made of, and whether the two describe one event.

My working thesis was that recalibrated radiocarbon series had moved the collapse by decades. I could not support that. What I found is a weaker headline and a better piece of history. The best clocks do not disagree by decades. They also do not agree on one year. The tidy "1200" hides a spread of about 50 years of destructions.

I did not read every paper in full. The Nature drought study and the Knapp and Manning review were closed to me. For those I use press summaries and search summaries, and I mark them.

The question

Is there one date for the end of the Late Bronze Age? If not, how wide is the spread, and which assumption widens it most?

Data and where it came from

Four kinds of evidence carry the dates.

Tree rings at Gordion. Juniper timbers from Gordion in central Anatolia, about 230 km west of Hattusa, give a ring series from 1775 to 748 BC with an error of 3 years [7][2]. Researchers read ring width and carbon isotope ratios in juniper from the Midas Mound Tumulus to find moisture stress [1]. They place the worst drought at 1198 to 1196 BC, with an error of 3 years (dendrochronology tied to the Gordion series) [1][7]. A university summary adds that six to eight years in the lowest growth class fall between 1210 and 1187 BC, three of them in a row [2].

Egyptian texts and objects. An Egyptian sword with the name of Merneptah (reign about 1213 to 1203 BC) lay in the destruction layers at Ugarit. A tablet shows the destruction came after his death [4]. Ugarit was gone by 1178 BC, the eighth year of Ramesses III, in the same account [4]. That is a historical date, from king lists and regnal years. It is not a measurement.

Radiocarbon. I only have secondary reports here. One search summary says recent radiocarbon work suggests a destruction date of 1192 to 1190 BC, and that a letter from the Egyptian official Bay to King Ammurapi sets an upper limit [9]. I could not open the paper behind it. I treat the 1192 to 1190 figure as a claim I cannot check.

Regional calibration. Manning and colleagues measured calendar-dated tree rings from southern Jordan (AD 1610 to 1940) and found a mean offset of about 19 radiocarbon years from the IntCal13 curve, which varies in time [5][6]. When they re-ran earlier studies, the 95.4% ranges were about 60% different on average [6]. Note the limits. That test used rings from recent centuries, and I read it only in summary. It says the offset is real. It does not say the offset is 19 years at 1200 BCE.

Method

I put each site or event on one line with its method and its stated error. Then I ask which clocks measure the same thing. I did this by hand, without the Lab. Every figure below is a published number or a subtraction of two of them.

The key point: the clocks answer different questions. Tree rings date a climate event. Egyptian synchronisms date a letter or a sword. Radiocarbon dates the death of the plant or seed that was sampled, and the burning of a city only by inference. Hattusa has a further problem: one source says it burned "at an unknown date", though it may have been abandoned instead [4].

Result with numbers and uncertainty

Event	Date (BCE)	Method and range
Worst Gordion drought	1198 to 1196	Tree rings, 3 years [1][7]
Ugarit destruction	about 1190	Texts plus pottery context, no range given [3]
Ugarit, narrower claim	1192 to 1190	Radiocarbon, as reported secondhand [9]
Ugarit, text limit	after 1194 to 1190	Bay letter synchronism, as reported secondhand [9]
Emar destruction	1187 to 1175	Historical and archaeological, stated range [4]
Ugarit gone by	1178	Regnal year of Ramesses III [4]
Cline's shorthand	1177	A convention, not a measurement [8]

Three results follow, each from simple subtraction.

First, the Hittite drought and the Ugarit destruction sit 6 to 8 years apart if I take the midpoints (1197 and 1190 to 1191). With a tree-ring error of 3 years, a gap of 3 to 11 years is possible. That is close. It is not a single year, and it does not show that the drought destroyed Ugarit. The authors themselves call the drought "a tipping point, at least for the Hittites", and they say not all of the Near East suffered a crisis [1].

Second, the 1177 convention sits 13 years after the Ugarit figure of 1190 and 21 years after the drought midpoint of 1197. A search summary says revisions to Egyptian chronology suggest a date about 11 years earlier than 1177 [8]. I could not read the reasoning. If it holds, the gap to Ugarit shrinks to about 2 years. I give this low weight until I read the paper.

Third, the whole sequence of destructions runs from about 1220 to about 1170 BC, a span of 50 years [3]. The collapse is not an event in this record. It is a pile of local events, and the word "collapse" was chosen later.

I also found an internal wrinkle. The account that dates Ramesses III from 1186 also says Ugarit was gone by "the eighth year" in 1178 [4]. If I count the first year as year 1, year 8 is 1179. A one-year slip is not important. It shows how fragile a bare date is when two chronologies are mixed in one paragraph. How do we know that date? Often a writer picked one convention and did not say which.

Sensitivity: which assumption moves the result most

I rank the assumptions by how far each could move the result.

1. Whether the 1177 shorthand is read as a date. It moves the "collapse" by 13 to 21 years against the other anchors. This is the largest swing, and it comes from a convention, not data [8].
2. Whether the sampled event is the event of interest. The Hittite drought is tied to Gordion, 230 km from Hattusa [7]. Treating a ring series as the date of a palace's end adds an unknown gap. I cannot size it from what I read.
3. The regional radiocarbon offset. An offset of about 19 radiocarbon years [5] matters because the radiocarbon curve around 1200 BCE is flat in places. I cannot say from my sources how many calendar years 19 radiocarbon years become there. That calculation needs the IntCal20 curve itself, and I have not run it. The 2018 paper reports a 60% average change in range width [6], so the effect on a 3-year claim like 1192 to 1190 could be large.
4. The tree-ring error. The 3-year error is small, and annual ring counting is the most precise tool here. It moves the result least, and it is the one clock I trust to the year, provided the series is correctly anchored.

This ordering surprised me. I expected calibration to be the big lever. On my evidence the big lever is how people round and name the event. My prior was that calibrated methods shifted accepted dates by decades. For this case, I now hold that more weakly. The shift I can document is 19 radiocarbon years in a regional offset, not decades in a destruction date.

What I think is secure

The tree-ring drought date, 1198 to 1196 BC with a 3-year error, looks secure as a dated climate event, though I read it only through summaries. A spread of destructions across about 1220 to 1170 BC is well supported by several accounts [3][4]. The 1192 to 1190 Ugarit radiocarbon figure is unverified by me. The 1177 date is a label.

I would call the claim "the collapse happened around 1200 BCE, plus or minus a few decades, as a sequence" secure. I would call the claim "it happened in one year" unsupported. I know the other blind spot here. I like material evidence, and I may have under-rated the written sources that give Ugarit its tightest bounds.

What would overturn this view? A published Bayesian model of short-lived samples from the Ugarit destruction layer, calibrated with a regional offset, whose 95.4% range excludes 1190. A Hattusa-specific tree-ring or radiocarbon date for its last burning would also change it. If that fell outside 1198 to 1187, the link between the drought and the Hittite end would weaken.