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

The study investigates whether a specific, consecutive multi-year drought coincided with and potentially contributed to the collapse of the Hittite Empire around 1200 BC. The Hittites, a major power centered in semi-arid central Anatolia with extensive regional linkages, endured numerous sociopolitical, economic and environmental challenges from ca. 1650–1200 BC. While prior work often links a broad, low-frequency shift to drier, cooler conditions in the Eastern Mediterranean and Near East to Late Bronze Age disruptions, precise, annually resolved associations between climate and the Hittite collapse have been lacking. The authors posit that while societies often adapt to gradual climatic shifts, rare sequences of consecutive extreme drought years can overwhelm resilience strategies. The paper aims to provide annually dated, high-resolution climate proxies from the Hittite administrative core to assess whether such an extreme episode occurred contemporaneously with the Empire’s collapse and to evaluate its potential historical significance.

Recent scholarship has increasingly explored climatic drivers for Late Bronze Age collapses, shifting emphasis from invasions, earthquakes, or political-economic causes to environmental forcing, citing multiple proxies indicating generally drier, cooler conditions from the 13th to 10th centuries BC across the region. However, these datasets are frequently low-frequency and loosely dated, limiting tests of synchronization and causation with historical events. Texts from the 13th century BC mention grain shortages in Hittite lands, though their context remains debated. Studies of risk, resilience, and agrarian strategies in semi-arid Mediterranean settings suggest that one bad year is manageable, but multiple consecutive poor years can trigger famine and crisis. Prior dendroclimatic and speleothem work in Anatolia and adjacent regions demonstrates spatially coherent drought variability, yet few records offer annual resolution from the Hittite core. This study addresses that gap using tree-ring width and carbon isotope data from central Anatolia and compares them with regional cave isotope records and modern precipitation analogs.

• Materials and study area: Archaeological juniper (Juniperus excelsa and J. foetidissima) timbers from Gordion in central Anatolia (ca. 230 km west of Hattusa), near the western frontier of Hittite lands, were used to construct an annually resolved tree-ring chronology spanning approximately 1775–748 (±3) BC. Twenty-three samples representing ~18 different trees include the years around 1200 BC.
• Dendrochronological processing: Individual ring-width series were detrended to maximize the common climate signal using multiple standard approaches (age-dependent splines, negative exponential curves, and Hugershoff functions) and ARSTAN indexing. The resulting chronologies were examined for periods of reduced growth (narrow rings) as proxies for dryness, given juniper sensitivity to late spring–summer precipitation at lower elevations in semi-arid central Anatolia.
• Dryness/drought thresholds: Years were categorized by percentile thresholds of growth indices (for example, lowest 25% = dry; lowest 20% = substantially reduced harvest risk; lowest 6.25% = severe drought likely leading to serious harvest reduction/failure), based on analogs to modern precipitation distributions.
• Modern climate analog: Records from the nearby Polatlı meteorological station (AD 1929–2009, ~17 km from Gordion) were used to contextualize thresholds (≈20% of years <300 mm annual precipitation; ≈6.25% of years <250 mm), the latter approximating the minimum for viable cereal cultivation in the region. Regional coherence of extreme aridity across central Anatolia was considered using published datasets.
• Temporal smoothing: A 28-point Savitzky–Golay filter (approximating a generational timescale) was applied to evaluate multi-year to decadal drought contexts.
• Stable isotope analyses: δ13C was measured on α-cellulose from annual rings of four Gordion trees; values were z-transformed and interpreted with higher δ13C as indicating drier conditions. The δ13C series (ca. 1400–1050 BC) was compared with δ13C and δ18O speleothem records from Sofular Cave (NW Anatolia) and Kocain Cave (SW Anatolia) to assess regional moisture trends and corroborate tree-ring inferences.
• Regional analogs for multi-year drought: Old World Drought Atlas (OWDA) reconstructions of summer scPDSI for well-attested back-to-back droughts (e.g., AD 1607–1608; AD 1927–1928) were reviewed to illustrate the typical spatial extent of severe Anatolian droughts relative to the Hittite realm.
• Chronological precision: The Gordion chronology is nearly absolutely dated with ±3-year uncertainty around the target interval. Consecutive drought sequences were identified by runs within the defined percentile thresholds.

• The Gordion tree-ring chronology (1497–797 BC segment emphasized) exhibits regular dry years, but consecutive extreme droughts are rare. Across 701 years, only 80–85 years (11.4%–12.1%) belong to sequences of two or more consecutive years in the lowest 20% of growth values.
• Only 13–16 years (1.9%–2.3%) in total are part of sequences of two or more consecutive years within the lowest 6.25% (severe drought) across the 701-year series.
• Between 1270 and 1135 BC (best replication: 10–18 trees/year), there are only six sequences of two or more consecutive years in the lowest 20%, and only one interval with two or more consecutive years in the lowest 6.25%: 1198–1196 BC, comprising three consecutive severe drought years.
• The 12-year span 1198–1187 BC contains 6–8 years (50%–67%) in the lowest 20% of values, indicating a prolonged dry phase. With 28-point Savitzky–Golay smoothing, the interval around 1198–1187 BC is the driest or second-driest multi-year period between 1400 and 1000 BC.
• δ13C from Gordion tree rings shows a drying gradient in the later 13th century BC with sustained dry to very dry conditions from 1232–1192 BC, including drier spikes at 1222–1221 BC and around 1195 BC. These features align closely with ring-width minima and coincide with textual references to grain shortages in Hittite lands.
• The timing of the severe consecutive drought (1198–1196 ±3 BC) aligns with the historically inferred window of Hittite imperial collapse and administrative abandonment at Hattusa.
• Regional context and analogs suggest that severe Anatolian droughts typically affect much of the Hittite realm, though with spatial variation in intensity, potentially creating strong differentials relative to neighboring areas.

The annually resolved dendroclimatic and isotopic proxies from central Anatolia indicate a rare, consecutive three-year severe drought (1198–1196 BC), embedded in a longer dry interval (ca. 1198–1187 BC; isotopic dry phase ca. 1232–1192 BC). In semi-arid agrarian systems such as the Hittite core, typical resilience strategies (diversification, storage, water management, and social networks) are calibrated to withstand single-year shocks; however, multi-year consecutive harvest failures are historically far more destabilizing and can precipitate famine, disease, social unrest, and political fragmentation. The evidence suggests that the Hittite heartland, largely land-locked and dependent on regional grain and pastoral systems, would have faced acute logistical constraints in importing bulk staples during sustained drought. Contemporary texts hint at constrained external grain support, further exacerbating stress. While climate alone does not cause historical outcomes, the concurrence of this rare extreme episode with known sociopolitical fault lines and external pressures provides a plausible contextual trigger that could have overwhelmed established adaptations, catalyzing administrative collapse and reorganization. The broader Eastern Mediterranean shows heterogeneous trajectories around 1200 BC, consistent with spatially variable drought impacts and diverse local responses, supporting a model where severe Anatolian drought was a contributing—though not sole—factor in the Hittite collapse.

This study provides an annually resolved, nearly absolutely dated moisture reconstruction from the Hittite administrative core, identifying an exceptionally severe, consecutive three-year drought (1198–1196 ±3 BC) during a prolonged dry interval that coincides with the collapse of the Hittite Empire. The integration of ring-width analyses, carbon isotopes from Gordion timbers, and regional speleothem records affords unprecedented temporal specificity, moving beyond low-frequency correlations to a plausible, historically salient episode that could have overwhelmed resilience strategies. The results suggest climate extremes contributed to, but did not solely cause, the Hittite collapse. Future research should: (1) expand multi-proxy records across Anatolia and adjacent regions to refine spatial patterns and synchronicity; (2) improve chronological precision and replication to test robustness of the 1198–1196 BC signal; (3) integrate archaeological-economic datasets (storage, subsistence, trade, and mobility) to model system sensitivity to consecutive droughts; and (4) develop statistical frameworks to evaluate causation pathways between extreme climate sequences and sociopolitical outcomes.

• Causation cannot be directly established; the study documents temporal concurrence and plausible mechanisms but acknowledges other human factors.
• Geographic representativeness: Primary data derive from Gordion (western edge of Hittite core); extrapolation to the entire Hittite realm relies on modern/regional analogs and comparative proxies.
• Chronological uncertainty of ±3 years around the key interval limits exact alignment with historical events.
• Sample size and species/site-specific responses may introduce biases despite detrending and cross-validation.
• Dryness thresholds are inferred via modern precipitation analogs and percentile-based ring-width metrics, which, while reasonable, are indirect proxies for harvest outcomes.
• Speleothem comparisons and isotope interpretations entail assumptions about moisture sources and seasonality, adding uncertainty to regional coherence assessments.