Past megadroughts in central Europe were longer, more severe and less warm than modern droughts

The study addresses whether recent central European droughts (e.g., 2003, 2010, 2013, 2015, 2018) are unprecedented in a millennial context. Droughts, especially when combined with high evapotranspiration, impact the entire hydrological cycle and impose major socio-economic costs. While recent events have been linked to anthropogenic warming, instrumental and satellite records are too short to assess long-term precedence. The authors therefore aim to place modern droughts in a Common Era context by leveraging paleoclimate reconstructions, hydrological observations, and large-scale climate drivers, focusing on central Europe where recent drought impacts were severe and proxy coverage is rich.

The paper situates recent European droughts within prior findings: recent summer warming in central Europe is unprecedented over the last 2500 years, and several European regions have experienced prolonged severe droughts since the early 2000s, often exacerbated by elevated temperatures. Previous work attributes such extremes in part to anthropogenic climate change and warming of land and ocean surfaces. Studies have also documented connections between European summer drought variability and North Atlantic SSTs, blocking, and large-scale circulation patterns. Reconstructions like the Old World Drought Atlas and documentary-based indices provide long records for evaluating hydroclimatic extremes. This background motivates a millennial-scale assessment of drought magnitude, duration, and drivers.

- Region and indices: Central Europe defined as 3°E–20°E, 45°N–56°N. Drought assessed via Palmer Drought Severity Index (PDSI/scPDSI). - Observational data: CRU TS 4.03 monthly precipitation, temperature, and scPDSI at 0.5° resolution; streamflow from the German Hydrological Institute; HadISST SST (1°); 20th Century Reanalysis Z500 and winds at 500 mb; reconstructed Z500 back to 1500. - Drought reconstruction: Extracted OWDA PDSI grid points over the region for 1000–2012 and averaged to a regional index. Scaled OWDA series to match the mean and standard deviation of observational scPDSI (CRU TS4) for valid comparison. Uncertainty estimated as RMSE from residuals between instrumental and original paleoclimate series. Similarity quantified by correlations: r=0.71 (p<0.001, 1901–2012) and r=0.79 (p<0.001, 1901–1978). - Historical documentary data: Seasonal decennial temperature and precipitation indices derived from German and neighboring regions using chronicles, diaries, administrative and harvest records, newspapers, and proxies (e.g., tree rings), calibrated with instrumental data. Indices represent deviations beyond ±0.75 standard deviations relative to 1951–1980. - Additional reconstructions: Seasonal precipitation/temperature reconstructions for Europe (e.g., Alps) and Germany used for cross-validation of drought periods over the last 500 years. - Last millennium reanalyses: PHYDA (proxy assimilation with CESM ensemble; annually resolved fields including PDSI and temperature at ~2.5°×1.9°) and LMR v2.0 (ensemble data assimilation including PAGES2k; annually resolved fields such as SST, PDSI, geopotential height) to examine links between drought, ocean states, and circulation over the Common Era. - Composite analysis: Constructed composites for years with PDSI < −0.75 std. dev. and TSI < −0.75 std. dev. to evaluate associated Z500 anomalies and blocking patterns (DJF and seasonal persistence). Significance assessed via t-tests at 95% confidence. Compared multi-decadal intervals (e.g., 1400–1480, 1770–1840, 1901–2012) using fitted Gaussian distributions of PDSI values and highlighted notable modern drought years (2003, 2015, 2018).

- Validation: The OWDA-reconstructed regional PDSI explains 52% of the variance in observed scPDSI over 1901–2012 (r=0.71, p<0.001); correlation is r=0.79 (p<0.001) over 1901–1978. - Driest years: Over the last millennium, central Europe’s driest years include 1102, 1419, 1503, 1504, 1858, 1865, and 1921. In observations, 1949, 1976, and 1990 are among the driest summers; within 1901–2012, 1921 and 1976 are notably dry. The 1921 drought was strongest in northwestern Germany, while 1976 affected the entire region; these align with Rhine, Elbe, and Weser low flows. - Two megadroughts: Two prolonged, severe drought periods are identified: ~1400–1480 (mid-15th century) and ~1770–1840 (late 18th to early 19th century). The first aligns with the Spörer Minimum (low solar activity with large eruptions) and featured dry summers (driest decade 1471–1480) and colder-than-average winters and springs. The second aligns with the Dalton Minimum and exhibited decades of negative precipitation anomalies across winter, spring, and summer, with colder winters/springs and harsh Baltic winters. - Modern droughts not unprecedented: Statistical distributions of PDSI show megadrought periods had markedly different distributions from the 20th–21st centuries; recent notable events (2003, 2015, 2018) fall within the range of natural variability over the last millennium in terms of drought magnitude and/or duration, though modern events are unusually warm. - Mechanisms: Low total solar irradiance (TSI) phases are followed by weak AMOC states after several decades; TSI minima are quasi-synchronous with weak AMOC. Negative AMO/weak AMOC states induce persistent high-pressure (blocking) over the British Isles/northern–central Europe, suppressing ascent and precipitation, promoting drought. Proxy and paleo-reanalysis composites show positive Z500 anomalies over western/central Europe during low TSI and low PDSI years. Cold North Atlantic conditions and enhanced blocking also characterize observed dry spells (e.g., 1971–1976), coinciding with an abrupt AMOC weakening. - Contrast with Maunder Minimum: Despite low solar activity, the Maunder Minimum (~1655–1715) featured a relatively warm North Atlantic (positive AMO) and precipitation excess in the study region, underscoring interplay between external forcing and internal variability. - Future context: While anthropogenic warming will exacerbate European drought risk, natural variability (e.g., potential TSI decreases, AMOC/AMO phases, blocking) will modulate future drought frequency and severity.

By combining observational, proxy, historical, and reanalysis evidence, the study demonstrates that central Europe has experienced megadroughts in the past that exceeded modern events in duration and severity, addressing the question of modern drought uniqueness. The findings clarify that recent droughts, though exceptional in warmth due to contemporary warming, are not unprecedented in hydroclimatic severity over the last millennium. Mechanistically, the study links past megadroughts to cold North Atlantic conditions and weak AMOC, likely triggered or modulated by reduced solar irradiance and volcanic activity, which favored persistent high-pressure (blocking) over western and central Europe. This highlights the importance of coupled ocean–atmosphere processes and external forcing in shaping multi-decadal drought regimes. The contrast during the Maunder Minimum, when the North Atlantic was relatively warm and the region was wetter, underscores that solar minima alone are insufficient; the combined state of ocean variability and atmospheric circulation is crucial. These insights refine attribution of recent droughts and inform future risk assessments by emphasizing the joint roles of anthropogenic warming and natural variability.

Central Europe experienced two major megadroughts (~1400–1480 and ~1770–1840) that were longer and more severe than 21st-century droughts. Recent notable droughts (2003, 2015, 2018) lie within natural variability when viewed over the last millennium, albeit with unusually high temperatures. The study identifies a mechanistic chain linking low solar irradiance to weak AMOC and cold North Atlantic conditions, fostering persistent blocking over western/central Europe and prolonged drought. Looking ahead, climate projections indicate substantial European drying even under lower-emission scenarios (SSP1-2.6, SSP2-4.5). Future drought risk will reflect both anthropogenic warming and natural variability; potential TSI decreases could further increase central European drought frequency, compounding anthropogenic drying. Future research should quantify joint anthropogenic-natural interactions (e.g., AMOC variability, AMO phase, blocking dynamics) and refine multi-proxy reconstructions and reanalyses to better constrain drought magnitude, persistence, and triggers.

- Reconstruction uncertainties: Differences among proxy-based reconstructions (OWDA, PHYDA, LMR, regional precipitation/temperature reconstructions) arise from varying proxy types, spatial coverage, and methods, leading to amplitude differences and uncertainties, especially near ~1820–1840. - Instrumental record length: Sparse and relatively short instrumental records (and short satellite era) limit direct comparison and calibration precision for extremes. - Scaling and regional averaging: Scaling OWDA to match observational mean/variance and using regional averages may mask sub-regional heterogeneity and introduce assumptions. - Reanalysis constraints: Paleo-reanalyses (PHYDA, LMR) have annual resolution and assimilation/model dependencies, limiting seasonal detail and potentially underestimating drought amplitude. - Forcing attribution: While associations with low TSI, volcanism, weak AMOC, and blocking are strong, causality and the partitioning between internal variability and external forcing remain partly constrained by data and model uncertainties.