The research team, led by Laura Lipfert of the Institute of Geography and the Oeschger Centre for Climate Change Research at the University of Bern, together with Ralf Hand, Angela-Maria Burgdorf, Christian Pfister, Heinz Wanner and Stefan Brönnimann, set out to quantify a relationship that climate scientists have long suspected but rarely been able to test over long timescales. Most observational studies of the spring precipitation to summer heat link have been confined to the instrumental era, generally extending no further back than the mid-twentieth century. By combining documentary archives, a 600-year paleoclimate reanalysis and thousands of years of model simulations, the Bern group has now pushed the record back to 1421, and with documentary evidence alone, to the early 1300s.

The physical mechanism at the heart of the study is the land-atmosphere feedback. When spring rainfall is scarce, soils enter the summer with depleted moisture reserves. Evapotranspiration, the process by which plants and soil release water vapor that carries away latent heat, becomes restricted. Energy that would otherwise be consumed evaporating water instead directly heats the air, amplifying temperature anomalies. This feedback is most powerful in regions where evaporation is moisture-limited, meaning that water, not energy, is the constraining resource. Southern and southeastern Europe sit squarely in this regime, which is precisely where the new analysis finds the signal to be strongest.

The study’s centerpiece case study is the summer of 1473, arguably the most extreme heat event of the past six centuries in terms of duration and geographical extent. Documentary sources describe an almost unimaginable year: rivers from France to western Russia shrank to trickles that could be forded on horseback or on foot, wells dried up across the Po Valley, forests burned uncontrollably in Poland and Belarus, and trees shed their leaves as if it were midwinter. In Basel, not a drop of rain fell from late June to mid-September. The drought persisted into the following winter, and in 1474 and 1475 swarms of migratory locusts invaded Central Europe. The year belongs to what researchers have described as a European megadrought period spanning roughly 1400 to 1480.

The reanalysis maps of 1473 reveal the atmospheric machinery behind the disaster. The preceding spring showed a pronounced precipitation deficit over Central and Eastern Europe, accompanied by a distinct positive East Atlantic pattern in sea-level pressure, a signature found in 19 of 20 ensemble members. The summer itself featured a large anomaly in 500-hectopascal geopotential height centered over Germany, along with a jet stream over the Atlantic-European sector that was stronger, more zonal and shifted poleward. Crucially, when the team examined the full statistical record, these same circulation fingerprints, a positive East Atlantic pattern and increased atmospheric blocking over north-central Europe in spring, emerged as recurring precursors of hot summers, not quirks of one exceptional year.

To conduct the statistical analysis, the researchers used the ModE-RA paleo-reanalysis, a 20-member global gridded monthly reconstruction spanning 1421 to 2008 at a resolution of roughly 1.8 degrees. ModE-RA blends an ensemble of transient simulations from the ECHAM6 atmospheric general circulation model with natural proxies, documentary records and instrumental measurements through an offline data assimilation scheme. The team supplemented this with ModE-RAclim, a variant in which the model prior is time-invariant so that any variability stems purely from the assimilated observations, and with the raw model ensemble ModE-Sim, which supplied a staggering 11,760 simulated years for the analysis. Anomalies were computed against a moving 31-year climatology using LOESS locally weighted regression, allowing fair comparison across centuries that include both the Little Ice Age and the modern warming era.

The results were consistent across all three datasets. Grid-point correlations between January-to-May precipitation and June-to-August temperature were significantly negative in a band between roughly 36 and 48 degrees north, with the strongest signal over southeastern Europe and near-zero values in northern Europe, where abundant moisture means evaporation is not water-limited. Composite analysis, drawing on the full ensemble to isolate one-in-ten-year events, showed that hot summers are typically preceded by dry springs over southeastern Europe with deficits reaching 1.4 millimeters per day, and that dry springs are followed by summer warming approaching one degree Celsius centered over the same region. Notably, for the Mediterranean region, 23 percent of the hottest summers were also dry-spring years, the highest overlap of any region examined.

Perhaps the most important result concerns causation. A skeptic could argue that atmospheric circulation simply drives both the dry spring and the hot summer, making precipitation a bystander rather than an agent. To test this, the team ran a regression analysis for the northern Mediterranean land area, predicting summer temperature from monthly precipitation, six circulation indices including the North Atlantic Oscillation, the East Atlantic and Scandinavian patterns, and three jet stream indices, plus Atlantic sea-surface temperatures, the NINO3.4 index and volcanic aerosol loading. Even after accounting for circulation in winter, spring and summer, April and particularly May precipitation remained significant predictors. The moisture deficit itself, not merely the circulation that produced it, exerts a direct influence on summer heat. Dry springs were also followed by more frequent and intense summer heatwaves in the model simulations, defined as spells of at least three consecutive days exceeding the 90th percentile.

The documentary record adds rich texture to the statistics. Of 22 historically documented hot summers, 16 were preceded by dry springs, including 1304, when the Rhine could reportedly be crossed on horseback between Strasbourg and Basel and ripe grapes appeared in Alsace by early July, and 1540, the legendary year when Lake Constance fell so low that its lake floor became visible and annual precipitation on the Swiss Plateau reached only about a quarter of the twentieth-century average. Yet the record also contains instructive exceptions: 1590 and 2024 saw hot summers after wet or variable springs, and several dry springs, such as those of 1420, 1559 and 1686, were not followed by drought summers. The link is real but probabilistic, not deterministic.

The authors caution that their setup was not designed to measure seasonal forecast skill, and the weak grid-point correlations, around minus 0.2, underline that a dry spring raises the odds of a hot summer without guaranteeing one. Even so, the implications are sobering. Under EURO-CORDEX projections, summer precipitation is expected to decline across Europe south of about 50 degrees north, while southeastern Europe, the study’s identified hotspot of land-atmosphere coupling, faces its own shifting rainfall regime. As 2022 and 2026 have demonstrated, when a spring precipitation deficit covers the continent, summer heat arrives with a depleted water balance already in place, producing compound heat and drought events that stress agriculture, rivers and power grids simultaneously. Six hundred years of history suggest that the atmosphere remembers a dry April and May, and that the price is often paid in July and August.

Subject of Research: The historical relationship between spring precipitation deficits and summer heat extremes in Europe over the past 600 years

Article Title: Historical climate extremes in Europe and the connection between spring precipitation and summer heat

Article References: Lipfert, L., Hand, R., Burgdorf, A.-M., Pfister, C., Wanner, H., & Brönnimann, S. (2026). Historical climate extremes in Europe and the connection between spring precipitation and summer heat. Climate of the Past, 22(10), 1863-1879. https://doi.org/10.5194/cp-22-1863-2026

Image Credits: AI Generated

DOI: 10.5194/cp-22-1863-2026

Keywords: paleoclimatology, heatwaves, drought, soil moisture, land-atmosphere feedback, Europe, ModE-RA, East Atlantic pattern, atmospheric blocking, 1473 summer, Climate of the Past, compound extremes