Movatterモバイル変換

[0]ホーム
URL:

Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
Thehttps:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

NIH NLM Logo
Log inShow account info
Access keysNCBI HomepageMyNCBI HomepageMain ContentMain Navigation
pubmed logo
Advanced Clipboard
User Guide

Full text links

Nature Publishing Group full text link Nature Publishing Group Free PMC article
Full text links

Actions

Share

.2023 Feb;614(7949):719-724.
doi: 10.1038/s41586-022-05693-y. Epub 2023 Feb 8.

Severe multi-year drought coincident with Hittite collapse around 1198-1196 BC

Affiliations

Severe multi-year drought coincident with Hittite collapse around 1198-1196 BC

Sturt W Manning et al. Nature.2023 Feb.

Abstract

The potential of climate change to substantially alter human history is a pressing concern, but the specific effects of different types of climate change remain unknown. This question can be addressed using palaeoclimatic and archaeological data. For instance, a 300-year, low-frequency shift to drier, cooler climate conditions around 1200 BC is frequently associated with the collapse of several ancient civilizations in the Eastern Mediterranean and Near East1-4. However, the precise details of synchronized climate and human-history-scale associations are lacking. The archaeological-historical record contains multiple instances of human societies successfully adapting to low-frequency climate change5-7. It is likely that consecutive multi-year occurrences of rare, unexpected extreme climatic events may push a population beyond adaptation and centuries-old resilience practices5,7-10. Here we examine the collapse of the Hittite Empire around 1200 BC. The Hittites were one of the great powers in the ancient world across five centuries11-14, with an empire centred in a semi-arid region in Anatolia with political and socioeconomic interconnections throughout the ancient Near East and Eastern Mediterranean, which for a long time proved resilient despite facing regular and intersecting sociopolitical, economic and environmental challenges. Examination of ring width and stable isotope records obtained from contemporary juniper trees in central Anatolia provides a high-resolution dryness record. This analysis identifies an unusually severe continuous dry period from around 1198 to 1196 (±3) BC, potentially indicating a tipping point, and signals the type of episode that can overwhelm contemporary risk-buffering practices.

© 2023. The Author(s).

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Proxies of drier to drought climate from three different detrending methods applied to the Gordion tree-ring dataset.
a, Tree-ring record from 1497–797 bc (Methods). The driest 25% of years is shown in orange (other 75% of years in green shading); division indicated by black horizontal line; 28-point Savitzky–Golnay (SG) filter is shown. Instances of 2 or 3 consecutive dry to very dry years at various levels are indicated. There are 3 instances of the driest 6.25% of years occurring consecutively (1494–1492 bc, 1198–1196 bc and 871–869 bc). The grey bar indicates the 12-year period from 1198 to 1187 bc with 3 consecutive years from 1198–1196 bc in the lowest 6.25% of all years and with 6 or 7 years (50–58%) in the lowest 20% of values. GOR, Gordion.b, Close-up of the period 1275–1125 bc, showing annual ARSTAN (ARS) index values (Methods) highlighting those in the lowest 20% and the lowest 6.25% of values.
Fig. 2
Fig. 2. δ13C record from Gordion tree rings compared with δ13C and δ18O records from Sofular and Kocain Caves, Turkey, and the period referred to in texts mentioning famine or grain shortage in Hittite lands.
δ13C and δ18O records from Sofular (bottom) and Kocain (top) Caves compared with the Gordion overall combined averagez score δ13C time-series, δ13CCorZ, chronology (middle). The Gordion chronology is also represented smoothed with a 28-point Savitzky–Golnay filter (middle).Z-transformed values greater than zero (drier) are highlighted with orange shading above the line and contrasted with green shading for wetter conditions below the line. Dryness increases towards the top of the graph. Bottom, E indicates the End Hattusha–Hittite Kingdom period during the reign of Suppiluliuma II and before Ramesses III year 8. Bottom, out of 9 contemporary texts that refer to famine or grain shortage in Hittite lands,, 4 are from the 13th centurybc, 3 are undated but also probably originate from the 13th centurybc (orange shading), and 2 are from the end of the 13th centurybc (red shaded area). The dashed arrows highlight indications of a drying trend in each record.
Fig. 3
Fig. 3. Reconstructed summer dryness levels across Europe and the Mediterranean region around the consecutive major drought yearsad 1607–1608 andad 1927–1928.
Climate proxy data from the Old World Drought Atlas (http://drought.memphis.edu/OWDA/) derived from tree-ring measurements, showing self-calibrating summer (June–July–August (JJA)) Palmer Drought Severity Index (scPDSI) values (negative values indicate drier and positive values indicate wetter) across the Old World region during and around the consecutive major drought yearsad 1607–1608 andad 1927–1928 (2 consecutive years). Our study suggests that there were 3 consecutive drought years from 1198 to 1196 bc in Anatolia.
Extended Data Fig. 1
Extended Data Fig. 1. Map of East Mediterranean-Near East indicating the major ancient empires (Hatti = Hittites, Egypt, Mitanni, Assyria, Babylonia), several other main political entities sometimes within, sometimes on the margins of, the Hittite Empire (Kaska, Wilusa, Seha River Land, Arzawa-Mira, Lukka, Pitassa, Tarhuntassa, Alashiya, Kizzuwatna),, the main sites/locations mentioned in this paper (Hattusa next to modern Boğazkale, Gordion, Polatlı, Sofular Cave, Kocain Cave, Kirikkale, Sivas, Kayseri, Karkemish, Kadesh), and the approximate Continental Central Anatolia (CCAN) precipitation zone, or the combined two main simplified composite ‘inland’ precipitation regimes identified for Anatolia in ref. .
Note: Wiyanawanda, Masa, and Ikkuna, mentioned in the main text, are all thought to lie in, or close to, the Lukka territory in southwest Anatolia. Map created from free and open (public domain) Natural Earth data (https://www.naturalearthdata.com/), using the Natural Earth II dataset.
Extended Data Fig. 2
Extended Data Fig. 2. Gordion tree-rings series and crossdating grid.
a, Raw (non-standardized) measured tree-ring series for the Gordion samples (all same scaling on Y axis) indicating the overall growth trends in each case.b, Crossdating grid for the Gordion tree-ring samples listing the Baillie-Pilcher t-scores (upper value) and length of overlap in rings/years (lower value) for each sample series versus the others from COFECHA. Potentially satisfactory values where there is strong overlap (e.g. > 50-100 years) are tBP ≥3.5 (unsatisfactory values are indicated by grey shading)—although in reality higher tBP-values are necessary for a secure match (see Methods). Higher values and especially tBP-scores >5-6 usually indicate a good crossdate. Here all samples exhibit at least several strong to very strong crossdates against a number of other samples in the chronology with long overlaps, demonstrating an overall robust chronology.
Extended Data Fig. 3
Extended Data Fig. 3. Ring-width measurements for the Gordion chronology and processing with ARSTAN, and detrending to Standard Chronologies for the six different approaches used (see Fig. 1, Extended Data Figs. 5, 8a), each offering good correspondence with the data.
Sample depth is shown under each figure element.
Extended Data Fig. 4
Extended Data Fig. 4. Residual and ARSTAN (ARS) chronologies derived from the six Standard Chronologies in Extended Data Fig. 3 using ARSTAN,.
Sample depth is shown under each figure element.
Extended Data Fig. 5
Extended Data Fig. 5. Sample number per year and Expressed Population Signal for a set of six ARSTAN (ARS) chronologies derived from the Gordion samples (see Extended Data Figs. 3, 4) using the ARSTAN software, employing different standardization curves, each aimed at highlighting the likely underlying climate signal (see Methods).
The shaded region indicates where the EPS values are ≥0.85. Note for the Hugershoff detrending one sample (GOR-3) receives the ‘itmax exceeded in amoeba’ warning—we used the age dependent spline for this sample.
Extended Data Fig. 6
Extended Data Fig. 6. Modern Instrument-derived self-calibrating Palmer Drought Series Index (scPDSI) values and correlationsad 1901–2012 for January-August for central Anatolia from the Old World Drought Atlas (OWDA).
a, Comparison of a west-east transect across the core Hittite region in central Anatolia from Polatlı (39.58N, 32.14E) to Kirikkale (39.85N, 33.5E) to Boğazkale (40.02N, 34.61E) (location of Hattusa) to Sivas (39.75N, 37.02E) and for a major center to the southeast (Kayseri) (38.72N, 35.48E) (see Extended Data Fig. 1). The scPDSI series from Polatlı has Pearson correlation coefficients (r) = 0.78, 0.7, 0.48, and 0.6 respectively; thus positive and suggesting a general regional pattern, except for Sivas in the east.b, Critically, the one extreme drought event in the dataset, centered 1928, occurs as a < scPDSI -4 event across all five loci, andc., is reconstructed across the entirety of central Anatolia, suggesting that extreme arid events recorded at Polatlı are likely regionally effective (see also Fig. 3 and Extended Data Fig. 7).
Extended Data Fig. 7
Extended Data Fig. 7. Tree-ring derived reconstructed self-calibrating Palmer Drought Series Index (scPDSI) values and correlationsad 1500–2000 (when there is reasonable data representation for Anatolia) for June-July-August for central Anatolia from the Old World Drought Atlas (OWDA).
a, Comparison of values from a west-east transect across the core Hittite region in central Anatolia as in Extended Data Fig. 6. Pearson correlation coefficients (r) with Polatlı (in vertical order shown) = 0.86, 0.85, 0.89 and 0.89 indicating a strong regionally coherent pattern. Instances where Polatlı and either 3 (of 4) or all 4 other locations have scPDSI less than −2.7 (value indicating lowest 6.25% of values at Polatlı) are indicated, as are some historically attested or reconstructed droughts startingad 1567 (ref. at Table 3), the reconstructed 2yr dry episodes for western Anatoliaad 1786-1930 (W dry 2yr) (ref. at Table 6) and instances of noted occurrences of 2yrs or 5yrs of drought in southern Anatolia (S dry 2/5yr) since 1500. All the five OWDA PDSI series record the same major arid episodes withb, showing that there is a strong wider regional expression across central Anatolia for the extreme arid events such as the five driest years in the Polatlı series in the OWDA dataset. In contrast, some of the other historically recorded droughts or those reconstructed for western or southern Anatolia do not correlate with especially dry conditions in one or more of the Polatlı, Kirikkale, Boğazkale, Sivas, and Kayseri series—events where the dashed grey bars do not correspond with the magenta lines ina—and so may not reflect serious drought expressed broadly across central Anatolia.
Extended Data Fig. 8
Extended Data Fig. 8. Additional ARSTAN (ARS) climate (moisture availability) proxies and comparison of ARS index values for a SW Anatolia juniper chronology versus observed low precipitation years.
a, Three additional alternative versions of the Fig. 1 analyses. Drier calculated conditions (low index values) toward the bottom (compare with Fig. 1). See Fig. 1 for description of analysis. Grey bar indicates 12-year period ~1198–1187bc with 3 consecutive years ~1198–1196 in lowest 6.25% of all years and with 7 or 8 years (58–67%) altogether in lowest 20% of values.b, ARS rt-gt values for three juniper tree-ring chronologies from SW Anatolia,ad 1931–2000, versus the lowest observed precipitation totals for May-June, the period critical to annual precipitation response, from proximate meteorological stations. The lowest (minimal) 37.1% of ARS values include the lowest 20% of observed precipitation values. The lowest 20% of ARS values include 50% of the lowest 20% of observed values and 7 of 9 (78%) in the periodad 1931–1961. The lowest 10% and 6.25% of ARS values are also indicated; with 5 of 7 (71%) of the former and 3 of 4 (75%) of the latter corresponding with years in the lowest 14.3% of observed years.
Extended Data Fig. 9
Extended Data Fig. 9. Gordion δ13C data, raw and residual values, and trends.
a, Upper: plots of the δ13C data (with linear extrapolation between data points). The linear fit lines indicate small gradients towards less negative values with increasing tree age. The arrows indicate a marked trend towards less negative values from ~GOR RY1230 to 1320. Lower: residual plots (from linear fits) removing different baselines and the age-trends. Gradient ~GOR RY 1230–1320 is evident.b, δ13C values from the two series on tree GOR-3. Similar baselines, and, despite noise, similar trend GOR RY1220–1320 (indicated by arrow).c, Average residual values (froma, lower) with 28 and 56 pt Savitzky-Golnay (SG) smoothing lines (one and two human generations) for period mainly with ≥2 data per year. Trend to less negative values GOR RY1220–1320 is indicated (see also inset). The plot at the bottom indicates the number of data values per year.
Extended Data Fig. 10
Extended Data Fig. 10. Indicative δ13C data and analysis from four Gordion trees (see SI, Table 2).
Z scores for the five δ13C series comparing these via 28 pt Savitzky-Golnay (SG) smoothings (or 28 pt adjacent average smoothing in the case where the data are not all evenly spaced) versus the 28 pt SG smoothed curve of the average z values for the combined δ13C series (δ13CCorZ) (see Methods). The correlation coefficients between compared smoothed curves versus the smoothed average curve are GOR-3E (0.62), GOR-3S (0.9), GOR-77 (0.73), GOR-82 (0.80) and GOR-87 (0.78).
See this image and copyright information in PMC

Comment in

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Kaniewski D, et al. Cold and dry outbreaks in the eastern Mediterranean 3200 years ago. Geology. 2019;47:933–937. doi: 10.1130/G46491.1. - DOI
    1. Drake BL. The influence of climatic change on the Late Bronze Age collapse and the Greek Dark Ages. J. Archaeolog. Sci. 2012;39:1862–1870. doi: 10.1016/j.jas.2012.01.029. - DOI
    1. Langgut D, et al. Climate and the Late Bronze collapse: new evidence from the southern Levant. Tel Aviv. 2013;40:149–175. doi: 10.1179/033443513X13753505864205. - DOI
    1. Cline, E. H. 1177 B.C.: The Year Civilization Collapsed (Princeton Univ. Press, 2021).
    1. Halstead, P. & O’Shea, J. eds. Bad Year Economics: Cultural Responses to Risk and Uncertainty (Cambridge Univ. Press, 1989). - PubMed

Publication types

MeSH terms

Related information

LinkOut - more resources

Full text links
Nature Publishing Group full text link Nature Publishing Group Free PMC article
Cite
Send To

NCBI Literature Resources

MeSHPMCBookshelfDisclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

[8]ページ先頭
©2009-2025 Movatter.jp