Jet stream controls on European climate and agriculture since 1300 CE

The study investigates how variability in the latitude of the North Atlantic–European summer jet stream (EU JSL) influences European climate extremes and societal impacts over the last seven centuries. Under anthropogenic climate change, models project a poleward shift and increased waviness of the Northern Hemisphere jet stream, promoting more persistent anomalies and intensified mid-latitude extremes. However, observational records are short and models exhibit biases, leaving uncertainty about historical EU JSL variability and its impacts prior to significant anthropogenic warming. Europe’s rich documentary archives and availability of temperature-sensitive tree-ring proxies provide a unique opportunity to reconstruct past EU JSL behaviour and assess its connections to regional dipole patterns in temperature, precipitation, and drought, and to agricultural, biophysical, and demographic outcomes.

Prior research links jet stream configuration to European weather regimes, including blocking, and to compound extremes such as heatwaves, droughts, floods and wildfires. Studies suggest enhanced jet waviness and faster upper-level winds under warming, as well as quasi-resonant circulation that can synchronize hemispheric extremes. Agricultural literature documents sensitivity of major crops to heat and water stress and the risk of synchronized low yields across breadbasket regions. Proxy-based reconstructions have previously characterized storm tracks and jet variability over recent centuries, and European historical documents have detailed climate impacts on harvests, grain prices, and epidemics. Nonetheless, the long-term regional expression of EU JSL variability and its societal impacts, particularly before the industrial era, has remained insufficiently quantified.

• Target definition: Summer (July–August) EU jet stream latitude (EU JSL) variability in the North Atlantic–European domain, focusing on the mode most closely associated with European late-summer climate extremes.
• Proxy network: Three temperature-sensitive tree-ring maximum latewood density (MXD) chronologies located at opposite ends of the EU JSL summer dipole and in the Alps: BRIT (British Isles), ALP (Alps), and a newly developed MXD record for the northeastern Mediterranean (NEMED). Each chronology extends to at least 1300 CE and shows robust positive correlations with regional July–August temperatures (r = 0.49–0.61; P ≤ 0.01), consistent across frequency domains.
• Calibration and reconstruction: Multiple linear regression of the three MXD series against instrumental EU JSL (1948–2004 CE). The reconstruction explains 38.5% of the variance (r = 0.65; P < 0.001) with no trends in residuals, preserving interannual to multi-decadal variability. Split-period calibration/verification yields 34–41% variance explained with positive RE and CE.
• Frequency characteristics: Combining chronologies from opposite ends of the dipole optimizes explained variance and replication but constrains low-frequency (centennial-scale) variability preservation.
• Modern composites: Composite analyses of years with extreme positive (northward) and negative (southward) EU JSL positions (D80/D20 or D90/D10) using reanalysis and gridded datasets: NCEP/NCAR 500 hPa geopotential height (Z500), CRU TS4.05 temperature and precipitation, scPDSI. Significance assessed with P ≤ 0.1 and false discovery rate control (qFDR ≤ 0.1).
• Historical field correlations: Correlation maps between reconstructed EU JSL and long-term gridded reconstructions of Z500 (1659–2000 CE), temperature (1766–2000 CE), precipitation (1766–2000 CE), and Old World Drought Atlas scPDSI (1300–2004 CE).
• Agricultural impacts: Analysis of gridded maize and wheat yields (1981–2016 CE; Global Dataset of Historical Yields), detrended with a 20-year smoothing spline. Composites computed for BRIT and NEMED under EU JSL extremes.
• Documentary datasets: Independent historical records of climate extremes (hot/cold; dry/wet; floods), grain prices, grape harvest timing and quality, wildfires (1450–1940 CE), epidemics and human mortality. EU JSL deviations during extreme years assessed using Wilcoxon signed-rank tests; significance thresholds P < 0.05 (filled) and P ≤ 0.1 (translucent).
• Event chronology and extremes: Identification of multi-decadal anomalies (e.g., 1810–1839 CE) and comparison to known volcanic eruptions (e.g., Mayon 1814, Tambora 1815).

• Modern EU JSL dipole: Northern EU JSL excursions produce cool, wet summers over the British Isles (BRIT) and hot, dry conditions over the northeastern Mediterranean (NEMED); southern excursions reverse this dipole. These patterns are reflected in temperature, precipitation, scPDSI, Z500, vegetation productivity (up to ~30% anomalies in modelled GPP) and tree growth (up to ~50% anomalies in beech radial growth).
• Crop yields: In NEMED (1981–2016 CE), maize and wheat yields are anomalously low during northern EU JSL (hot/dry) summers; southern EU JSL (cool/wet) summers are associated with above-average yields. Wheat shows weaker, less homogeneous anomalies than maize. In BRIT, yield responses are weaker but generally of the same sign as in NEMED.
• Reconstruction skill and structure: The tree-ring-based reconstruction explains 38.5% of observed EU JSL variance (1948–2004 CE; r = 0.65; P < 0.001) and is skillful back to 1300 CE (RE, CE > 0; 34–41% variance explained). Variability is dominated by interannual to multi-decadal scales.
• Long-term variability: A pronounced multi-decadal southern EU JSL anomaly occurred during 1810–1839 CE, overlapping the cold early 19th century and coinciding with major tropical eruptions (1813–1816). Recent EU JSL values generally fall within historical ranges, though 2010 exhibits the most northerly instrumental EU JSL, beyond the reconstruction period.
• Increased extremes: The 19th and 20th centuries show ~40% more northern and southern EU JSL extremes than previous centuries.
• Robust dipole through time: Correlations between reconstructed EU JSL and historical fields (Z500, temperature, precipitation, scPDSI) consistently show the BRIT–NEMED climate dipole across centuries, weakened in the 19th century due to volcanic forcing.
• Historical climate and hazards: Composites of reconstructed EU JSL during independently documented historical extremes show that hot/dry NEMED summers align with northern EU JSL anomalies, while cold/wet NEMED summers align with southern anomalies; BRIT exhibits the opposite pattern. NEMED floods occur predominantly during southern EU JSL summers; historical wildfire years (1450–1940 CE) occur mostly during northern EU JSL summers.
• Societal impacts: Historical NEMED grape harvests show delays, low yields, and poor wine quality during southern EU JSL (cool/wet) summers, and early harvests with better quality during northern EU JSL (hot/dry) summers. Grain prices in BRIT and NEMED respond consistently with modern yield patterns (lower prices with southern EU JSL in NEMED; BRIT prices tend to be higher after cold, wet northern EU JSL summers). Plague years in NEMED are associated with southern EU JSL (cool/wet) summers, whereas BRIT epidemics and mortality increase after northern EU JSL (cold/wet) summers.

The reconstruction demonstrates that EU JSL variability has persistently organized a summertime climate dipole between BRIT and NEMED for over seven centuries, with clear imprints on ecosystems, agriculture, and human societies. This long-term perspective contextualizes modern extremes and supports evidence that anthropogenic warming may be associated with more frequent or intense northern EU JSL configurations. Northern EU JSL extremes elevate wildfire risk in NEMED and can reduce crop yields at both ends of the dipole, with potential implications for European food security. The study underscores cascading climate–society interactions: crop failures can elevate grain prices, aggravate food insecurity, and contribute to heightened mortality and epidemics. By linking atmospheric dynamics to historical biophysical and demographic stressors, the work informs risk assessments of compound events. The extended record also aids model development and attribution by illuminating the dynamic (circulation-driven) contributions to extremes, complementing thermodynamic warming influences. Projections of a weak northward shift and enhanced waviness of the jet stream highlight the importance of considering EU JSL configurations in future risk planning.

This study provides a 705-year reconstruction of summer EU jet stream latitude variability using temperature-sensitive tree-ring density records and links EU JSL extremes to a persistent BRIT–NEMED climate dipole and to agricultural, biophysical, economic, and demographic impacts. The reconstruction reveals an increase in EU JSL extremes in the 19th–20th centuries, a notable southern anomaly in the early 1800s coincident with major volcanic eruptions, and robust dipole patterns across centuries. The findings emphasize the need to integrate EU JSL dynamics into assessments of compound climate risks, wildfire management, and food security planning under anthropogenic warming. Future research should improve representation of jet dynamics in models, investigate mechanisms behind low-frequency variability and external forcings (e.g., volcanic), expand proxy networks to better preserve centennial scales, and enhance integration of historical documentary evidence with modern observations and process-based impact models.

• Frequency preservation: Combining MXD chronologies from opposite ends of the dipole optimizes explained variance but reduces retention of low-frequency (centennial-scale) variability; Little Ice Age centennial signals are muted.
• Calibration period and target uncertainty: The instrumental EU JSL record (1948–2004 CE) is relatively short, and reanalysis-based metrics may carry uncertainties and biases.
• Volcanic forcing: The 19th century exhibits weakened dipole patterns due to strong volcanic influences, complicating attribution of EU JSL impacts.
• Proxy–impact linkages: Grain prices are an indirect proxy for agricultural productivity and are influenced by storage, trade, and market dynamics; historical datasets may have spatial and temporal biases and heterogeneous sensitivities.
• Regional heterogeneity: BRIT agricultural and temperature signals show weaker and less homogeneous responses than NEMED; historical maize data are lacking, and crop–climate relationships can evolve with changing varieties and practices.
• Temporal coverage: The reconstruction ends in 2004 and therefore does not capture the exceptionally northern EU JSL of 2010.