The variability of the northern westerlies has been considered as one of the key elements for modern and past climate evolution. Their multiscale behavior and underlying control mechanisms, however, are incompletely understood, owing to the complex dynamics of Atlantic sea-level pressures. Here, we present a multi-annually resolved record of the westerly drift over the past 6,500 years from northern Italy. In combination with more than 20 other westerly-sensitive records, our results depict the non-stationary westerly-affected regions over mainland Europe on multi-decadal to multi-centennial time scales, showing that the direction of the westerlies has changed with respect to the migrations of the North Atlantic centers of action since the middle Holocene. Our findings suggest the crucial role of the migrations of the North Atlantic dipole in modulating the westerly-affected domain over Europe, possibly modulated by Atlantic Ocean variability.
The Mediterranean Basin has been a cradle of civilizations since the middle Holocene (8.2–4.2 thousand years before 1950 C.E., kyr BP). Today some 400 million people live in this region. As a “climate hot spot,” the Mediterranean features large hydroclimate variability in response to global climate change1 and has been experiencing exceptionally low rainfall over the past two decades2, resulting in significant impacts on ecosystems, human society, and the economy3. The drought has been attributed to an enhanced sea-level pressure (SLP) contrast between the Azores High and the Icelandic Low (the SLP dipole, hereafter; Supplementary Fig. 1), known as the positive phase of the North Atlantic Oscillation (NAO)4, which led to a northerly migration of the westerlies, transporting moisture away from the Mediterranean region (Supplementary Fig. 2a). The location of the SLP dipole is not stable through time. Its spatial displacement determines the angle of the westerly tracks and the phase of the East Atlantic (EA) pattern5,6,7,8,9,10,11, defined by a pressure anomaly centered over the eastern North Atlantic (52°30’N, 27°30’W; Supplementary Fig. 2b)12. The variabilities of the SLP dipole and varying westerly tracks can therefore result in unstable NAO-correlated rainfall regimes (red and blue areas in Supplementary Movie 1) in Europe on decadal to multidecadal timescales, termed “non-stationary NAO” behavior13.
Climate proxy records are essential to develop an in-depth understanding of the westerlies on a range of timescales. For example, terrestrial and marine sediments in Iberia14, and Moroccan tree-ring data combined with Scottish speleothem records15 suggest a prolonged positive NAO phase during the Medieval Climate Anomaly (1050–850 yr BP). However, Northern Hemisphere proxy assemblages and model simulations argued that proxy-based paleo-NAO reconstructions are sometimes contentious16,17,18, largely because of the complex interactions among Atlantic SLP. Only few proxy records reliably reflect Atlantic SLP patterns, rendering tracking the westerlies’ position across mainland Europe challenging19. Since NAO-correlated rainfall regimes (Supplementary Movie 1) reflect the position of the westerlies5,6,7,8,9,10,11, past positions of the westerlies can be constrained by comparing a series of westerly-sensitive records in the North Atlantic region.
Here, we present a multi-annually resolved precipitation record from such a sensitive region (northern Italy). In conjunction with other regional westerly-sensitive archives, these data depict large-scale hydroclimate changes and the variability of the SLP dipole and the westerlies over the past 6500 years.
Results and discussion
Bàsura cave (44°08′N, 8°12′E; Supplementary Figs. 1 and 3 and Supplementary Text 1) is located in Toirano, Liguria, northern Italy, an area characterized by a typical Mediterranean climate. More than 70% of the annual precipitation, 1276 (± 310) mm (1σ, 1833–2008 C.E.), falls during the rainy seasons from September to February (Genoa meteorological station; 44°24’N, 8°05’E; 56 m above sea level; Supplementary Fig. 3a). The 1-km-long cave has a mean annual cave air temperature of 15.6 °C (2013–2014 C.E.) and 97–100% relative humidity beyond 100 m from the entrance. Two stalagmites, BA14-1 and BA18-4, were collected at sites with 98–100% relative humidity (Supplementary Figs. 3b and 4). Their U-Th-based age model (Supplementary Fig. 5) shows that they cover the time interval from 6437 ± 12 to 5 ± 18 yr BP. About 1000 subsamples were extracted along the growth axis of the stalagmites for δ18O analysis (Supplementary Figs. 4 and 6) to establish a composite δ18O record (∆18O; Supplementary Fig. 6; Methods). Hendy tests20 (Supplementary Figs. 4 and 7) of 13 growth layers suggest that calcite precipitated at near isotopic equilibrium (Supplementary Text 2). In combination with the coeval Sr/Ca data (Supplementary Text 3 and Supplementary Fig. 8; Methods), the ∆18O is interpreted as a westerly-sensitive record, with negative/positive values corresponding to strong/weak westerlies over northern Italy and high/low rainfall amount (Supplementary Text 2).
The decadal to multi-centennial westerly fluctuations since 6.5 kyr BP documented by the Bàsura ∆18O record are consistent with regional reconstructions, largely reflecting hydroclimate changes (Supplementary Text 2). This ∆18O series is in good agreement with a lacustrine record21 and a stalagmite-based wet/dry index from Italy22,23 as well as hydroclimate reconstructions from Spain24, Portugal25, and Algeria26 (Supplementary Fig. 9 and Supplementary Table 2). For example, the correlation coefficients of Bàsura ∆18O with lacustrine flood records from northern Italy (Supplementary Fig. 9a)21 and cave records from central Italy (Supplementary Fig. 9b, c)22,23 are –0.42 (n = 31, p < 0.05) and 0.67 (n = 31, p < 0.05), respectively (Supplementary Table 2). The drought between 5 and 4 kyr BP was concurrent with dry periods recorded by speleothems from Turkey27, Lebanon28, Israel29, and Morocco30,31 (Supplementary Fig. 10a–d). The following wet period during 4–3 kyr BP is in good agreement with records from circum-Mediterranean countries (Supplementary Figs. 9d–f, 10a, b, and 11c, d)24,25,26,27,28,32,33. These events likely had a profound impact on ancient Mediterranean human societies (Supplementary Fig. 12). For example, the 5.2 kyr BP dry period has been suggested to have resulted in the demise of the Uruk period29, while the dry period at 4.8–4.1 kyr BP (4.2-kyr event) possibly induced the collapse of the Old Kingdom34 and the Akkadian Empire35.
Hydroclimate changes in northern Italy are strongly affected by North Atlantic SLP variability (Supplementary Fig. 2a, b) and westerly dynamics. Instrumental precipitation data from Genoa (Genoa PP) are positively/negatively correlated with the 850-mb zonal wind around Gibraltar/Iceland during the rainy season (Supplementary Fig. 2c), suggesting that changes in the position of the westerlies affect Genoa PP. Genoa PP show a strong negative correlation with SLP variations centered over northwestern Europe (Supplementary Fig. 2d), resembling the EA pattern12, implying that spatial changes of the SLP dipole dominate regional rainfall patterns. The SLP in northwestern Europe (averaged over [45–55°N, 20°W–0]) is also significantly anti-correlated with the 850-mb zonal wind around Toirano (averaged over [42–47°N, 5–15°E]), with a correlation coefficient of −0.6 ± 0.04 (n = 97; p < 0.01; 1920–2016 C.E.) during winter (December–February, DJF) based on NCAR/NCEP Reanalysis v3. This anticorrelation emphasizes that the westerlies dominate the hydroclimate in Toirano. The leading empirical orthogonal function (EOF) 1 of European-Atlantic (60°W–40°E, 20–80°N) precipitation represents an EA teleconnection (Supplementary Fig. 13a) that is positively correlated with Genoa PP (r = 0.43 ± 0.04, n = 173, p < 0.01). Combined with the NAO pattern observed in EOF2 (Supplementary Fig. 13b), our results show that the variability of the SLP dipole influences Toirano precipitation (Supplementary Fig. 2a, b). The precipitation patterns documented by the Toirano records hence reflect movements of the westerlies under the two climate modes, consistent with previous studies36,37,38.
On multi-decadal to multi-centennial time scales, Bàsura stalagmite-inferred paleo-precipitation is positively correlated (r = 0.69, n = 292, 99%, Methods) with a NAO reconstruction16 of the past millennium (1049–1969 C.E.; Supplementary Fig. 14). In contrast to the negative correlation of modern Genoa PP with the NAO index39 in winter (DJF; Supplementary Fig. 2a), the comparison suggests that Toirano was in a positive NAO-correlated region over the past millennium, where the westerlies were strong and precipitation was high during the positive NAO phase. When compared with lacustrine sediment records from western Greenland40 (Fig. 1d) that reflect NAO-driven local temperature changes and westerly dynamics, Bàsura ∆18O inferred precipitation patterns also show a positive correlation in the interval 2.2–1.2 kyr BP (r = 0.37, n = 64, 95%) and 0.8–0.3 kyr BP (r = 0.69, n = 41, 95%), but a negative correlation (r = –0.48, n = 84, 99%) between 4.2 and 2.2 kyr BP (Fig. 1d). In other words, the relationship of Toirano precipitation with the reconstructed NAO index16 remained positive after 2.2 kyr BP, but was negative in the interval 4.2–2.2 kyr BP, revealing that the NAO-correlated regions over Europe have changed since the middle Holocene.
Non-stationary North Atlantic Oscillation
Over the past 150 years, correlations of Toirano precipitation with western Greenland temperature (Fig. 1c; black line) and with the NAO index39 (Fig. 1c; blue line) have been unstable, reflecting the non-stationary NAO behavior (Supplementary Movie 1). Correlations between NAO index and Toirano precipitation were, for example, more significant for 1870–1900 and 1950–1980 C.E. than for 1905–1935 C.E. (Fig. 1c), suggesting that Toirano was located in negative NAO-correlated regions in the late 19th and late 20th centuries (Fig. 1a), but close to the boundary of (or even temporally in positive) NAO-correlated regions in the early 20th century (Fig. 1b). Such changes in NAO-correlated regions have been suggested to be modulated by the migrations of the SLP dipole (i.e., EA phases)5,6,7,8,9,10,11. Northward or counterclockwise rotated displacements of the SLP dipole could lead the westerly tracks towards a southwest-northeast tilt and shift the positive NAO-correlated regions towards higher latitudes5,9 (e.g., Fig. 1a).
Variable correlations between Toirano precipitation and hydroclimate in the North Atlantic (Fig. 1c; black and blue lines) match the trend of the SLP time series over southwestern England (Fig. 1c; brown line) and the reconstructed EA index (Fig. 1c; red line)41. This confirms that EA phase changes could affect the locations of the SLP dipole and hence the NAO-correlation regions over Europe, in accordance with previous studies9. For longer timescales, isotope-enabled general circulation models indicate that the multi-decadal correlation patterns of NAO-precipitation δ18O in Europe can be affected by the EA phases42. The Bàsura ∆18O also shows similarity with the 700-mb height pressure variation in the eastern Atlantic on centennial to millennial scales obtained from a transient simulation (TraCE-21ka) with all forcings using the Community Climate System Model version 3 (Fig. 2e). Accordingly, correlation changes between the NAO index and Bàsura ∆18O (Fig. 1d) over the past thousands of years can be attributed to the variability in the position of the SLP dipole, which modulates the NAO-correlation regions.
Our conclusions are supported by the comparisons with westerly-sensitive records across Europe. For example, during the positive NAO phase 5.4–3.5 kyr BP40,43 (Fig. 2a), proxy records reveal a cool/dry climate in the regions below ~45°N21,22,23,24,25,26,27,28,29,30,31,44 (Fig. 2d, e, g and Supplementary Figs. 9 and 10a–d) and a wet/warm climate in northern Europe45,46,47,48,49 (Fig. 2b, c, g). However, during another positive NAO phase 2.2–1.2 kyr BP40,43 (Fig. 2a), southern21,24 (Fig. 2e), central44 (Fig. 2d and Supplementary Fig. 9b, d), and northern Europe45,46,47,48,49,50,51,52 (Fig. 2b, c) were characterized by multi-centennial wet and warm conditions, while this period featured ambiguous hydroclimate fluctuations in the eastern Mediterranean27,28,29 (Supplementary Fig. 10a–c) and northern Africa31,53 (Supplementary Fig. 10d, e). In addition, intense storminess in northern Europe49 coincides with droughts in southern Europe during 5.8–5.5, 4.5–3.95, and 3.3–2.4 kyr BP (Fig. 2f) but with wetness in southern Europe during 1.9–1.05 and 0.6–0.25 kyr BP. These observations suggest that the NAO-correlated regions over Europe and the North Atlantic westerly tracks changed. A running correlation analysis (Supplementary Fig. 15) between Bàsura ∆18O and proxy records from Norway46, Germany45 and Austria44 reveals that Norway and Germany were in a different precipitation domain than Bàsura cave before versus after ~3 kyr BP. Austria, on the other hand, remained in the same precipitation domain as Bàsura cave over the entire period. The shifts of the SLP dipole might be time-transgressive and an exact time boundary is difficult to define, but in general, the SLP dipole migrated southward after 2.2 kyr BP. This is supported by European pollen-based wet/dry reconstructions54 and is consistent with a southward migration of the oceanic Azores front between 6.5 kyr BP and the late Holocene55.
The observation of the changing NAO-correlated regions demonstrates a centennial to millennial mixing effect of NAO and EA modes on European hydroclimate changes. The NAO mode in principle dominated the “dipole” precipitation pattern over mainland Europe (Fig. 1a, b) over the past 6500 years. The EA mode also played an important role in modulating the centennial to millennial positions of the NAO dipole. This phenomenon confirms the combined effect of NAO and EA on positions of the NAO dipole over the instrumental period8,9. For example, a concurrent positive NAO phase and negative EA phase during the period 5.4–3.5 kyr BP (Fig. 2g) forced the Azores High northeastward. Coinciding positive NAO and EA phases from 2.2–1.2 kyr BP, on the other hand, resulted in a southwestward shift of the Azores High.
Migration of the westerlies since the middle Holocene
SLP dipole movements could alter the route of the westerlies over the East Atlantic5,6,7,8,9,10,11. A northward displacement of the Azores High, such as prior to 2.2 kyr BP, favors a SW-NE direction of the westerlies over Europe. In this case, the pivot point of the SLP dipole (the NAO-zero correlation line) can also rotate into a SW-NE direction, restricting the positive NAO-correlated regions to northwestern Europe (Fig. 2g). After 2.2 kyr BP, the SLP dipole migrated southwards and the westerlies tilted less over Europe. The positive NAO-correlated regions shifted southward and northern Italy was thus positively correlated with the reconstructed NAO index (Fig. 2h).
A high degree of similarity between our Bàsura ∆18O record and a stalagmite δ18O series from Kaite cave, Spain (Supplementary Fig. 9d)24 strengthens our argument of the relationship between westerly routes and the variability of the SLP dipoles since the middle Holocene. Variability in the Kaite cave δ18O record reflects the zonal migration of the Icelandic Low and the positive δ18O values are related to westward shifts of the Icelandic Low. Positive Bàsura ∆18O values, on the other hand, are associated with an eastward migration of the Azores High. The similar patterns of Bàsura ∆18O and Kaite δ18O (Supplementary Fig. 9d), therefore, suggest the co-occurrence of westward shifts of the Icelandic Low with eastward shifts of the Azores High. The induced counterclockwise rotation of the SLP dipole in turn can lead to a change in the NAO-correlated regions over Europe8,9. Combined with studies that suggest more frequent positive NAO phases from the middle to the late Holocene40,56,57, our results reveal the complex interaction between the NAO state and its dipole locations on centennial to millennial scales, with associated changes in the orientation of westerly routes across Europe.
Possible effect of latitudinal oceanic thermal gradient on the sea-level pressure dipole
The movements of the SLP dipole are likely modulated by the latitudinal gradient of sea-surface temperatures (SSTs) in the North Atlantic. On decadal to centennial scales, observational reanalysis and model simulations have shown that the multidecadal SLP response to Atlantic Multidecadal Variability (AMV) projects on migrations of the SLP dipole (i.e., changes in EA phases)58,59. Warm AMV phases that feature a warm North Atlantic Ocean correspond to a southward shift of the SLP dipole (positive EA). Over the past 2.2 kyr, SSTs have increased in northern Norway60 (K23258, Fig. 3b) and eastern Greenland61,62 (MD99-2269, Fig. 3d; MD99-2322, Fig. 3e), in line with elevated sea-surface salinity (SSS) levels around the Subpolar Gyre63,64 (Fig. 3a; RAPiD-12-1K, Fig. 3f; MD99-2227, Fig. 3g) and an increased proportion of Atlantic water-sensitive coccoliths in the Nordic Sea65 (MD95-2011, Fig. 3c). These observations reveal that, over the past 2.2 kyr, more warm, salty, low-latitude Atlantic water has been delivered towards high latitudes compared to before, possibly reducing the latitudinal SST gradient and hence leading to a southward migration of the SLP dipole. The linkage of AMV and EA is also in accord with results of the preindustrial controlled Community Earth System Model 2 (CESM2)66 (Methods), which shows a strong coherence between SLP anomalies over the East Atlantic and SST changes along the North Atlantic Current (Fig. 3a), suggesting that Atlantic Ocean dynamics play an important role in the migration of the SLP.
Our ∆18O series in combination with other westerly-sensitive records show multiple patterns of European westerly drift on decadal to millennial scales since the middle Holocene in response to the non-stationary behavior of the NAO that was possibly modulated by Atlantic sea-surface conditions. Our results underscore the impact of changing Atlantic Ocean circulations on the position of the westerlies over Europe that in the near future might be associated with variability of the meridional overturning circulation, increasing greenhouse gases, and/or varied aerosol concentrations59.
Stalagmite samples, U-Th dating, and age model
Two stalagmites, 190-mm-long BA14-1 and 90 mm-long BA18-4, were collected in two cave chambers of narrow side passages of Bàsura cave, 500 and 800 m from the entrance (Supplementary Fig. 3a), in January 2014 and June 2018, respectively. The stalagmites were cut into halves and polished (Supplementary Fig. 4). High resolution scanning electron microscope (SEM) analysis shows that BA14-1 formed as needle-shaped aragonite and BA18-4 as rhombic calcite crystals (Supplementary Fig. 4).
A total of 78 (BA14-1) and 18 (BA18-4) subsamples, 10–100 mg each, were drilled for U-Th chemistry67 and dating67,68 (Supplementary Data 1) at the High-Precision Mass Spectrometry and Environment Change Laboratory (HISPEC), Department of Geosciences, National Taiwan University. All U-Th isotopic measurements were conducted on a multi-collector inductively coupled plasma mass spectrometer, Thermo-Finnigan Neptune68. A gravimetrically calibrated69 triple-spike, 229Th-233U-236U, and the isotope dilution method were employed to correct for mass bias and to determine U-Th isotopic compositions and contents. Half-lives of U-Th nuclides used for U-Th age calculation are given in ref. 69. Uncertainties in isotopic data and dates, relative to 1950 C.E., are given at the two-sigma (2σ) level or two standard deviations of the mean (2σm) unless otherwise noted. U-Th contents, isotopic compositions and ages are given in Supplementary Data 1. Age corrections for the initial 230Th are 0–2 years, smaller than dating errors of ± 2–30 years. A Monte-Carlo-derived age-depth model was constructed using StalAge70 techniques (Supplementary Fig. 5).
Stable oxygen isotopes
A total of 732 (BA14-1) and 265 (BA18-4) powdered subsamples, 10–50 μg each, were micro-milled at 0.05–0.10 mm intervals along the growth axis (Supplementary Data 2 and Supplementary Figs. 4 and 6). Four to seven coeval subsamples of 13 layers were drilled for Hendy tests20. The oxygen isotopic composition was analyzed on Thermo-Finnigan MAT 253 mass spectrometers at the College of Geography Science, Nanjing Normal University, the Department of Geography Science, Fujian Normal University and the Department of Natural History Sciences, Faculty of Science, Hokkaido University. Subsamples for Hendy tests were analyzed on a Micromass IsoPrime mass spectrometer at the Department of Earth Sciences, National Taiwan Normal University. All δ18O values are reported in per mil (‰), relative to the Vienna PeeDee Belemnite (VPDB) and standardization was accomplished using NBS-19. Reproducibility of δ18O measurements was ± 0.08–0.12‰ at the 1-sigma level.
Average δ18O values of aragonitic BA14-1 are 1.2‰ higher than those of calcitic BA18-4 during the overlapping period from 702 to 752 yr BP (Supplementary Fig. 6). This offset falls within the range of 0.6–1.4‰ suggested by previous speleothem studies71. To combine the two series, the BA14-1 δ18O data was subtracted by 1.2‰ and attached to the BA18-4 δ18O series. The overlapping δ18O (702–752 yr BP) was averaged and resampled at 4-year intervals to match the average resolution of the overlapping δ18O. We also introduced an addition error of ± 0.2‰ derived from the one standard deviation of δ18O differences of BA14-1 and BA18-4 during the overlapping period. The maximum error in combination with the instrumental error (± 0.12‰) is ± 0.23‰ for the new time series. The composite Δ18O data (Supplementary Data 2; Supplementary Fig. 6) were then normalized to the mean δ18O value (–5.8‰) of the new series over the entire period.
Powdered subsamples, 10–50 μg each, were micro-milled at 0.1–0.5 mm intervals along the growth axis for Sr/Ca determination on BA14-1 and BA18-4, with external matrix-matched standards for every 4–5 samples on an inductively coupled plasma sector-field mass spectrometer, Finnigan Element II, at the HISPEC, National Taiwan University72. The 2-sigma reproducibility is ± 0.5%.
BA14-1 Sr/Ca, with a mean of 1.2 (± 0.2, 1σ) mmol/mol, vary from 0.45 to 1.9 mmol/mol. BA18-4 Sr/Ca range from 0.035 to 0.25 mmol/mol and the mean is 0.048 (± 0.017, 1σ) mmol/mol. The offset between the two Sr/Ca datasets is caused by different partition coefficients, 0.1–0.2 for calcite73 and 0.8–2.0 for aragonite74. To combine the two Sr/Ca series, BA14-1 and BA18-4 Sr/Ca data were first converted to z-standard records using the mean and standard deviation of each dataset. The converted z-standard records (Supplementary Data 3) were zigzagged in the overlapping interval, 728–875 yr BP, based on the assumption that the precipitation changes influenced Sr/Ca of aragonite (BA14-1) and calcite (BA18-4) in the same direction (Supplementary Text 3).
Error propagation of correlation analysis
For correlations between data without age uncertainties, the upper/lower limits for the correlation coefficient were added, derived from the probable errors75 (P.E.) in correlations with P.E. = 0.6745 × (1−r2)/N0.5, where r is the correlation coefficient and N is the number of observation pairs.
For the correlation of two time series with age uncertainties, the method of Fohlmeister et al. (2012)76 was applied. This method embraces a Monte Carlo approach to simulate the best correlation between two series. The maximum correlation coefficient between two real series is presented in the text followed by the significance level. Specifically, more than two thousand artificial random time series that have the same characteristics as the original time series (e.g., variance, auto-correlation coefficients, data resolution and absolute ages with age uncertainties) were generated. By tuning these artificially generated time series, the distribution of the best correlations and the significant level can be obtained. For example, if 5% of tuned artificial series yield a correlation coefficient above 0.5, the best correlation of real series with a coefficient of > 0.5 is significant at the 95% significance level. The maximum correlation coefficient between two series is presented in the text followed by the significance level.
For Supplementary Table 2, the correlation analysis was conducted on centennial to multi-centennial scales. We first resampled each of the 20 proxy records, using 15-year time windows and averaged these values for consecutive 150-year periods, resulting in time series of X values (degrees of freedom = X–1), each representing a 150-year average. We then calculated correlation coefficients between each of these time series and the transformed Bàsura record.
Community Earth System Model Version 2
A Community Earth System Model Version 2 (CESM2) preindustrial simulation66 was performed at approximately 2° × 2° resolution to represent natural climate variability. To verify that, the proposed processing holds on multidecadal to centennial time scales as in our records, a 5-yr running mean on this long preindustrial simulation was performed in a 300-year window. The sea-surface temperature data were correlated with the 700-mb height pressure variation in the eastern Atlantic (52˚30’N, 27˚30’W).
U-Th report, δ18O and Sr/Ca of BA14-1 and BA18-4 are available in the Source Data file.
Giorgi, F. Climate change hot-spots. Geophys. Res. Lett. 33, 1–4 (2006).
Hoerling, M. et al. On the increased frequency of Mediterranean drought. J. Clim. 25, 2146–2161 (2012).
Nelson, G. C. et al. Climate change: Impact on agriculture and costs of adaptation (International Food Policy Research Institute, Washington, D.C., 2009).
Hurrell, J. W. Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science 269, 676–679 (1995).
Woollings, T., Hannachi, A. & Hoskins, B. Variability of the North Atlantic eddy-driven jet stream. Q. J. R. Meteorol. Soc. 136, 856–868 (2010).
Wang, Y.-H., Magnusdottir, G., Stern, H., Tian, X. & Yu, Y. Decadal variability of the NAO: Introducing an augmented NAO index. Geophys. Res. Lett. 39, L21702 (2012).
Woollings, T. & Blackburn, M. The North Atlantic jet stream under climate change and its relation to the NAO and EA patterns. J. Clim. 25, 886–902 (2012).
Moore, G. W. K., Renfrew, I. A. & Pickart, R. S. Multidecadal mobility of the North Atlantic Oscillation. J. Clim. 26, 2453–2466 (2013).
Comas-Bru, L. & Mcdermott, F. Impacts of the EA and SCA patterns on the European twentieth century NAO-winter climate relationship. Q. J. R. Meteorol. Soc. 140, 354–363 (2014).
Mikhailova, N. V. & Yurovsky, A. V. The East Atlantic Oscillation: Mechanism and impact on the European climate in winter. Phys. Oceanogr. 4, 25–33 (2016).
Mellado-Cano, J., Barriopedro, D., García-Herrera, R., Trigo, R. M. & Hernández, A. Examining the North Atlantic Oscillation, East Atlantic Pattern, and jet variability since 1685. J. Clim. 32, 6285–6298 (2019).
Barnston, A. G. & Livezey, R. E. Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Weather Rev. 115, 1083–1126 (1987).
Vicente-Serrano, S. M. & López-Moreno, J. I. Differences in the non-stationary influence of the North Atlantic Oscillation on European precipitation under different scenarios of greenhouse gas concentrations. Geophys. Res. Lett. 35, 1–6 (2008).
Moreno, A. et al. The Medieval Climate Anomaly in the Iberian Peninsula reconstructed from marine and lake records. Quat. Sci. Rev. 43, 16–32 (2012).
Trouet, V. et al. Persistent positive North Atlantic Oscillation mode dominated the Medieval Climate Anomaly. Science 324, 78–80 (2009).
Ortega, P. et al. A model-tested North Atlantic Oscillation reconstruction for the past millennium. Nature 523, 71–74 (2015).
Hernández, A. et al. A 2,000-year Bayesian NAO reconstruction from the Iberian Peninsula. Sci. Rep. 10, 14961 (2020).
Cresswell-Clay, N. et al. Twentieth-century Azores High expansion unprecedented in the past 1,200 years. Nat. Geosci. 15, 548–553 (2022).
Trouet, V., Babst, F. & Meko, M. Recent enhanced high-summer North Atlantic Jet variability emerges from three-century context. Nat. Commun. 9, 1–9 (2018).
Hendy, C. The isotopic geochemistry of speleothems—I. The calculation of the effects of different modes of formation on the isotopic composition of speleothems and their applicability as palaeoclimatic indicators. Geochim. Cosmochim. Acta 35, 801–824 (1971).
Sabatier, P. et al. 6-kyr record of flood frequency and intensity in the western Mediterranean Alps – Interplay of solar and temperature forcing. Quat. Sci. Rev. 170, 121–135 (2017).
Regattieri, E. et al. Late glacial to Holocene trace element record (Ba, Mg, Sr) from Corchia Cave (Apuan Alps, central Italy): paleoenvironmental implications. J. Quat. Sci. 29, 381–392 (2014).
Isola, I. et al. The 4.2 ka event in the central Mediterranean: New data from a Corchia speleothem (Apuan Alps, central Italy). Clim. Past 15, 135–151 (2019).
Domínguez-Villar, D., Wang, X., Krklec, K., Cheng, H. & Edwards, R. L. The control of the tropical North Atlantic on Holocene millennial climate oscillations. Geology 45, 303–306 (2017).
Thatcher, D. L. et al. Hydroclimate variability from western Iberia (Portugal) during the Holocene: Insights from a composite stalagmite isotope record. Holocene 30, 966–981 (2020).
Ruan, J. et al. Evidence of a prolonged drought ca. 4200 yr BP correlated with prehistoric settlement abandonment from the Gueldaman GLD1 Cave, Northern Algeria. Clim. Past 12, 1–14 (2016).
Fleitmann, D. et al. Timing and climatic impact of Greenland interstadials recorded in stalagmites from northern Turkey. Geophys. Res. Lett. 36, L19707 (2009).
Cheng, H. et al. The climate variability in northern Levant over the past 20,000 years. Geophys. Res. Lett. 42, 8641–8650 (2015).
Bar-Matthews, M. & Ayalon, A. Mid-Holocene climate variations revealed by high-resolution speleothem records from Soreq Cave, Israel and their correlation with cultural changes. Holocene 21, 163–171 (2011).
Wassenburg, J. A. et al. Reorganization of the North Atlantic Oscillation during early Holocene deglaciation. Nat. Geosci. 9, 6–11 (2016).
Ait Brahim, Y. et al. North Atlantic ice-rafting, ocean and atmospheric circulation during the Holocene: Insights from Western Mediterranean speleothems. Geophys. Res. Lett. 46, 7614–7623 (2019).
Demény, A. et al. Middle Bronze Age humidity and temperature variations, and societal changes in East-Central Europe. Quat. Int. 504, 80–95 (2019).
Lončar, N., Bar-Matthews, M., Ayalon, A., Faivre, S. & Surić, M. Holocene climatic conditions in the Eastern Adriatic recorded in stalagmites from Strašna peć Cave (Croatia). Quat. Int. 508, 98–106 (2018).
Drysdale, R. et al. Late Holocene drought responsible for the collapse of Old World civilizations is recorded in an Italian cave flowstone. Geology 34, 101–104 (2006).
Cullen, H. M. et al. Climate change and the collapse of the Akkadian empire: Evidence from the deep sea. Geology 28, 379–382 (2000).
Serrano, A., García, J. A., Mateos, V. L., Cancillo, M. L. & Garrido, J. Monthly modes of variation of precipitation over the Iberian Peninsula. J. Clim. 12, 2894–2919 (1999).
Lorenzo, N., Taboada, J. & Gimeno, L. Links between circulation weather types and teleconnection patterns and their influence on precipitation patterns in Galicia (NW Spain). Int. J. Climatol. 28, 1493–1505 (2008).
Josey, S. A., Somot, S. & Tsimplis, M. Impacts of atmospheric modes of variability on Mediterranean Sea surface heat exchange. J. Geophys. Res. Ocean. 116, 1–15 (2011).
Jones, P. D., Jonsson, T. & Wheeler, D. Extension to the North Atlantic Oscillation using early instrumental pressure observations from Gibraltar and south-west Iceland. Int. J. Climatol. 17, 1433–1450 (1997).
Olsen, J., Anderson, N. J. & Knudsen, M. F. Variability of the North Atlantic Oscillation over the past 5,200 years. Nat. Geosci. 5, 808–812 (2012).
Comas-Bru, L. & Hernández, A. Reconciling North Atlantic climate modes: Revised monthly indices for the East Atlantic and the Scandinavian patterns beyond the 20th century. Earth Syst. Sci. Data 10, 2329–2344 (2018).
Comas-Bru, L., McDermott, F. & Werner, M. The effect of the East Atlantic pattern on the precipitation δ18O-NAO relationship in Europe. Clim. Dyn. 47, 2059–2069 (2016).
D’Andrea, W. J., Huang, Y., Fritz, S. C. & Anderson, N. J. Abrupt Holocene climate change as an important factor for human migration in West Greenland. Proc. Natl Acad. Sci. 108, 9765–9769 (2011).
Fohlmeister, J., Vollweiler, N., Spötl, C. & Mangini, A. COMNISPA II: Update of a mid-European isotope climate record, 11 ka to present. Holocene 23, 749–754 (2013).
Fohlmeister, J. et al. Bunker Cave stalagmites: An archive for central European Holocene climate variability. Clim. Past. 8, 1751–1764 (2012).
Vasskog, K., Paasche, Ø., Nesje, A., Boyle, J. F. & Birks, H. J. B. A new approach for reconstructing glacier variability based on lake sediments recording input from more than one glacier. Quat. Res. 77, 192–204 (2012).
Gilbertson, D. D., Schwenninger, J.-L., Kemp, R. A. & Rhodes, E. J. Sand-drift and soil formation along an exposed North Atlantic coastline: 14,000 years of diverse geomorphological, climatic and human impacts. J. Archaeol. Sci. 26, 439–469 (1999).
de Jong, R., Björck, S., Björkman, L. & Clemmensen, L. B. Storminess variation during the last 6500 years as reconstructed from an ombrotrophic peat bog in Holland, southwest Sweden. J. Quat. Sci. 21, 905–919 (2006).
Sorrel, P. et al. Persistent non-solar forcing of Holocene storm dynamics in coastal sedimentary archives. Nat. Geosci. 5, 892–896 (2012).
Sundqvist, H. S., Holmgren, K., Moberg, A., Spötl, C. & Mangini, A. Stable isotopes in a stalagmite from NW Sweden document environmental changes over the past 4000 years. Boreas 39, 77–86 (2010).
Spencer, C. D., Plater, A. J. & Long, A. J. Rapid coastal change during the mid- to late Holocene: The record of barrier estuary sedimentation in the Romney Marsh region, southeast England. Holocene 8, 143–163 (1998).
Baker, A. et al. A composite annual-resolution stalagmite record of North Atlantic climate over the last three millennia. Sci. Rep. 5, 10307 (2015).
Sha, L. et al. How far north did the African Monsoon fringe expand during the African Humid Period? Insights from southwest Moroccan speleothems. Geophys. Res. Lett. 46, 14093–14102 (2019).
Mauri, A., Davis, B. A. S., Collins, P. M. & Kaplan, J. O. The climate of Europe during the Holocene: A gridded pollen-based reconstruction and its multi-proxy evaluation. Quat. Sci. Rev. 112, 109–127 (2015).
Repschläger, J., Garbe-Schönberg, D., Weinelt, M. & Schneider, R. Holocene evolution of the North Atlantic subsurface transport. Clim. Past. 13, 333–344 (2017).
Orme, L. C. et al. Past changes in the North Atlantic storm track driven by insolation and sea-ice forcing. Geology 45, 335–338 (2017).
Morley, A., Rosenthal, Y. & DeMenocal, P. Ocean-atmosphere climate shift during the mid-to-late Holocene transition. Earth Planet. Sci. Lett. 388, 18–26 (2014).
Kwon, Y. O., Seo, H., Ummenhofer, C. C. & Joyce, T. M. Impact of multidecadal variability in Atlantic SST on winter atmospheric blocking. J. Clim. 33, 867–892 (2020).
Ruggieri, P. et al. Atlantic multidecadal variability and North Atlantic jet: A multimodel view from the decadal climate prediction project. J. Clim. 34, 347–360 (2021).
Sarnthein, M. et al. Centennial-to-millennial-scale periodicities of Holocene climate and sediment injections off the western Barents shelf, 75°N. Boreas 32, 447–461 (2003).
Justwan, A., Koç, N. & Jennings, A. E. Evolution of the Irminger and East Icelandic Current systems through the Holocene, revealed by diatom-based sea surface temperature reconstructions. Quat. Sci. Rev. 27, 1571–1582 (2008).
Jennings, A., Andrews, J. & Wilson, L. Holocene environmental evolution of the SE Greenland shelf north and south of the Denmark Strait: Irminger and East Greenland current interactions. Quat. Sci. Rev. 30, 980–998 (2011).
Thornalley, D. J. R., Elderfield, H. & McCave, I. N. Holocene oscillations in temperature and salinity of the surface subpolar North Atlantic. Nature 457, 711–714 (2009).
de Vernal, A. & Hillaire-Marcel, C. Provincialism in trends and high frequency changes in the northwest North Atlantic during the Holocene. Glob. Planet. Change 54, 263–290 (2006).
Giraudeau, J. et al. Millennial-scale variability in Atlantic water advection to the Nordic Seas derived from Holocene coccolith concentration records. Quat. Sci. Rev. 29, 1276–1287 (2010).
Danabasoglu, G. et al. The community earth system model version 2 (CESM2). J. Adv. Model. Earth Syst. 12, 1–35 (2020).
Shen, C.-C. et al. Measurement of attogram quantities of 231Pa in dissolved and particulate fractions of seawater by isotope dilution thermal ionization mass spectroscopy. Anal. Chem. 75, 1075–1079 (2003).
Shen, C.-C. et al. High-precision and high-resolution carbonate 230Th dating by MC-ICP-MS with SEM protocols. Geochim. Cosmochim. Acta 99, 71–86 (2012).
Cheng, H. et al. Improvements in 230Th dating, 230Th and 234U half-life values, and U-Th isotopic measurements by multi-collector inductively coupled plasma mass spectrometry. Earth Planet. Sci. Lett. 371–372, 82–91 (2013).
Scholz, D. & Hoffmann, D. L. StalAge – An algorithm designed for construction of speleothem age models. Quat. Geochronol. 6, 369–382 (2011).
Lachniet, M. S. Climatic and environmental controls on speleothem oxygen-isotope values. Quat. Sci. Rev. 28, 412–432 (2009).
Shen, C.-C. et al. High precision measurements of Mg/Ca and Sr/Ca ratios in carbonates by cold plasma inductively coupled plasma quadrupole mass spectrometry. Chem. Geol. 236, 339–349 (2007).
Day, C. C. & Henderson, G. M. Controls on trace-element partitioning in cave-analogue calcite. Geochim. Cosmochim. Acta 120, 612–627 (2013).
Wassenburg, J. A. et al. Determination of aragonite trace element distribution coefficients from speleothem calcite–aragonite transitions. Geochim. Cosmochim. Acta 190, 347–367 (2016).
Student. The Probable Error of a Mean. Biometrika 6, 1–25 (1908).
Fohlmeister, J. A statistical approach to construct composite climate records of dated archives. Quat. Geochronol. 14, 48–56 (2012).
The authors thank Henry de Lumley for his contribution to obtain access authorization for sampling in the cave, Meng-Lun Li and Chia-Hao Hsu for technical computer support, and Tsai-Lun Yu for help in the laboratory. TraCE-21ka was made using the DOE INCITE computing program, and supported by NCAR, the NSF P2C2 program, and the DOE Abrupt Change and EaSM programs. We are thankful for the financial support provided by grants from the National Science and Technology Council (NSTC), Taiwan, ROC (110-2123-M-002-009, 111-2116-M-002-022-MY3, 111-2926-I-002-510-G to C.-C.S.), the National Taiwan University (109L8926 to C.-C.S.), the Higher Education Sprout Project of the Ministry of Education, Taiwan, ROC (111L901001 to C.-C.S.), the Graduate Students Study Abroad Program of the NSTC, Taiwan, ROC (109-2917-I-002-003 to H.-M.H.), the US National Science Foundation (AGS-1349942 to VT), National Natural Science Foundation of China (42050410317 to C.P.M.), and CEPAM, PRC-INEE project (France).
The authors declare no competing interests.
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Hu, HM., Trouet, V., Spötl, C. et al. Tracking westerly wind directions over Europe since the middle Holocene. Nat Commun 13, 7866 (2022). https://doi.org/10.1038/s41467-022-34952-9
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.