The research, led by Aytaç Paçal of the German Aerospace Center (DLR) together with colleagues at the University of Bremen, Universidad Carlos III de Madrid, the Universitat de València, and DLR, centers on a Variational Autoencoder, or VAE. This type of neural network, first developed as a generative model, works by compressing enormous, high-dimensional inputs into a much smaller latent space and then reconstructing the original data from that compressed code. In this case, the inputs were not photographs but spatiotemporal snapshots of the atmosphere: nine climate variables, including temperature, humidity, wind components, mean sea level pressure, geopotential height at 500 hPa, cloud cover, stream function, and surface solar radiation, tracked over eleven-day windows across the North Atlantic and Europe.

To build those inputs, the team started with the ERA5 reanalysis dataset, a continuous reconstruction of global weather stretching back to 1940. Heatwave onsets were identified over Western European land areas by comparing daily maximum 2-meter temperatures against local 90th percentile thresholds, calculated for each calendar day using a fifteen-day moving window from the 1941 to 1980 baseline. A density-based clustering algorithm, GDBSCAN, then stitched spatially and temporally contiguous hot grid cells into discrete events, yielding 2,565 unique heatwave onset dates between 1941 and 2022. Each onset was wrapped in a window spanning five days before and five days after, so the model could learn not just what a heatwave looks like at its peak, but how the atmosphere builds toward it and breaks down afterward.

The VAE was deliberately trained on the historical period from 1941 to 1990, then evaluated on recent events from 2001 to 2022. This historical-to-recent split creates what the authors call an out-of-distribution setting: a test in which the model must represent climate conditions that differ from anything it learned. The architecture relied on three-dimensional convolutions, which process space and time simultaneously, proving more stable than hybrid designs combining two-dimensional convolutions with LSTM layers. The model achieved reconstruction skill, measured by R-squared scores of roughly 0.72 to 0.76 across training, validation, and test sets, with the strongest performance for the large-scale circulation fields such as geopotential height and sea level pressure, and lower skill for the noisier near-surface winds.

When the researchers encoded recent heatwaves into the latent space and visualized the results, a striking pattern emerged. Events from 2001 to 2022 occupied a ring-like structure with fewer samples at the center and a pronounced accumulation toward one side of the space, regions largely untouched by the historical training events. Statistical tests quantified the drift: using Mahalanobis distance to measure how far the recent distribution had moved, summer samples shifted by D equal to 2.00, with about 83 percent of that signal attributable to thermodynamic warming. When the linear warming trend was removed from the data, statistically significant shifts persisted across all four seasons, and winter emerged as the season with the largest remaining dynamical displacement, at D equal to 1.40. In other words, beyond simple mean warming, the multivariate atmospheric structure of heatwaves itself appears to be evolving.

Clustering the latent space with a Gaussian Mixture Model revealed that four components best describe the test-period heatwaves, and each cluster mapped onto a distinct physical regime. One cluster captured winter warm spells, driven by persistent anticyclonic blocking centered over the British Isles and Western Europe, with suppressed cloud cover, warm advection from the Atlantic, and a structure resembling the omega-block configurations known from the literature. A second, dominant summer cluster showed the classic signature of intense European heat: strong positive temperature, humidity, and geopotential height anomalies, a deep North Atlantic trough feeding a ridge over Europe, and near-stagnant winds under cloud-free skies. Two further clusters both captured transition-season events in spring and autumn, yet displayed nearly opposite anomaly patterns, one ruled by Greenland blocking and enhanced solar radiation, the other by stronger pressure anomalies and warm-air advection.

A crucial validation step came from testing the model against some of the most infamous heatwaves on record. The team located the 2003 Western European event, the 2010 Russian heatwave, and the 2018 Scandinavian heatwave within the learned latent space and constructed composite anomaly maps from each event’s ten nearest neighbors. The resulting 500 hPa geopotential height anomalies reproduced the well-documented circulation signatures of these benchmark events: a ridge over central Europe in 2003, strong positive height anomalies over western Russia in 2010, and Scandinavian height anomalies in 2018. The fact that an unsupervised model, trained without any labels or human guidance, independently organized these events in a physically consistent way underscores the promise of the method.

The findings carry weight because heatwaves are intensifying faster in Western Europe than climate models have generally projected, with recent studies showing that unprecedented extreme heat has grown roughly tenfold and that economic losses from events like those of 2003, 2010, 2015, and 2018 reached 0.3 to 0.5 percent of European gross domestic product. By capturing the interplay between circulation, humidity, radiation, and temperature, the VAE approach offers a window into why: enhanced land-atmosphere feedbacks, moisture limitations, and shifting circulation regimes can amplify and restructure heat extremes beyond what a thermodynamic trend alone would predict. The authors caution, however, that linear detrending cannot fully separate forced changes from natural variability, and that the identified clusters depend on the chosen variables, thresholds, and model hyperparameters.

The team also probed an open question with direct forecasting implications: do the pre-onset circulation patterns identified by the model appear during quiet periods that never produce a heatwave? A preliminary comparison against ordinary summer days from 2015 to 2021 found no clear predictive signal, suggesting these large-scale configurations can occur without always triggering extreme heat. Future work, the researchers suggest, could add soil moisture, vegetation indices, or apply the same framework to CMIP6 climate simulations to see how well models capture historical heatwave patterns and where their biases lie.

What makes this study resonate far beyond the technical community is its central message: heatwaves are not simply hotter versions of the past, but events whose underlying atmospheric machinery is shifting. By letting a machine discover those shifts without labels, thresholds, or preconceived categories, the researchers have produced a data-driven atlas of heatwave regimes that aligns with established meteorology yet reveals changes invisible to traditional single-variable methods. As Europe braces for hotter summers, warmer winters, and increasingly erratic shoulder-season extremes, tools like this VAE may become essential for detecting when the atmosphere’s playbook has changed, and for building the climate risk assessments on which adaptation strategies depend.

Subject of Research: Multivariate atmospheric drivers of Western European heatwaves analyzed with unsupervised machine learning

Article Title: A multivariate analysis of atmospheric drivers for Western European heatwaves

Article References: Paçal, A., Hassler, B., Weigel, K., Fernández-Torres, M.-Á., Camps-Valls, G., & Eyring, V. (2026). A multivariate analysis of atmospheric drivers for Western European heatwaves. Earth System Dynamics, 17(4), 955-986. https://doi.org/10.5194/esd-17-955-2026

Image Credits: AI Generated

DOI: 10.5194/esd-17-955-2026

Keywords: heatwaves, Western Europe, Variational Autoencoder, machine learning, ERA5, atmospheric blocking, climate extremes, geopotential height, latent space, Earth System Dynamics, unsupervised learning, climate change

The work, carried out by Christian M. Grams of the Federal Office of Meteorology and Climatology MeteoSwiss, updates and thoroughly documents a regime definition first introduced in 2017. Where traditional European regime classifications were built separately for individual seasons, most famously the four canonical winter regimes identified in the 1990s, the year-round framework handles the entire calendar in one consistent system. The key trick is a normalization of anomalies in 500-hectopascal geopotential height, a standard measure of the large-scale flow at mid-altitudes of the atmosphere. Because winter anomalies are naturally much larger in amplitude than summer ones, an unnormalized analysis across all seasons would be dominated by winter. By dividing the low-pass filtered height anomalies by a seasonally varying measure of their climatological variability, the method removes the seasonal difference in amplitude while preserving the spatial shape of the flow, allowing summer and winter patterns to be compared on equal footing.

From this normalized data, an empirical orthogonal function analysis combined with fuzzy clustering identified seven distinct regimes, and the number seven was not chosen arbitrarily. The author tested configurations ranging from two to eleven clusters and used an objective similarity index, the maximum anomaly correlation coefficient between cluster patterns, to judge when clusters stopped being genuinely distinct. With four clusters the familiar canonical regimes reappear, but pushing to seven reveals finer structure while keeping the patterns well separated. Beyond seven, an eighth weak pattern emerges that closely resembles existing regimes, and the similarity between clusters jumps sharply. The seven regimes split into two families: three cyclonic regimes, Atlantic Trough, Zonal, and Scandinavian Trough, dominated by negative height anomalies and enhanced storm activity, and four blocked regimes, Atlantic Ridge, European Blocking, Scandinavian Blocking, and Greenland Blocking, dominated by positive anomalies and high pressure.

One of the study’s central contributions is an objective life cycle definition. Rather than simply labeling each day with a regime category, the method computes a continuous index for each regime, a standardized projection of the current atmospheric state onto the regime pattern. When this index exceeds a threshold of 1.0 for at least five days, a life cycle is identified, complete with onset, maximum, and decay stages. Crucially, life cycles can overlap, so one regime can be building while another is still dissolving, which matches how the atmosphere actually behaves. Roughly thirty percent of all time steps belong to no regime at all, corresponding to flow situations close to climatology. This continuous, dynamical view transforms regimes from a static classification into something closer to living entities with births, mature phases, and deaths.

The practical payoff of this framework is considerable. Weather regimes modulate surface conditions across entire continents for days to weeks, exactly the window where conventional weather forecasts lose skill but before seasonal averages take over. They shape wind speeds critical to Europe’s renewable power supply, including the dreaded Dunkelflaute episodes when low winds and weak sunshine coincide, and they modulate heat waves, cold spells, storm series, and atmospheric river landfalls. Previous work has linked prolonged heat-induced mortality to European blocking and winter stress on the UK health system to Greenland blocking. The new study shows that the most extreme regime life cycles, those lasting more than a month, coincide with Europe’s most extreme seasons. A Zonal regime life cycle in early 1990 lasted an extraordinary 73 days amid severe winter storms, while the longest Scandinavian Blocking life cycle began in July 2003 and lasted 32 days during Europe’s record-breaking heat. A 54-day Greenland Blocking episode starting in November 2010 brought a brutal cold wave.

The regimes also reveal a striking seasonal rhythm. Cyclonic regimes together account for up to forty percent of winter days but only about twenty-two percent in July, while blocked regimes dominate summer, exceeding fifty percent of days in June through August. European Blocking and Scandinavian Blocking alternate in preference: the former peaks in relative frequency in winter and spring, the latter in summer and autumn. Inter-annual variability is enormous. The early 1990s saw a run of strongly zonal years, 2010 brought a dominant negative North Atlantic Oscillation winter, and the hot years 2021 and 2022 featured frequent blocking. No single year matches the climatological average, underscoring how much year-to-year chaos overlays the seasonal cycle.

The most provocative question the study tackles is whether climate change is shifting regime frequencies. Extending the identification back to 1950 using the backward extension of the ERA5 reanalysis, the author finds that inter-annual variability dominates and significant trends are scarce, with one exception: Scandinavian Blocking shows a significant positive trend, concentrated in summer and autumn. But here the story takes a subtle turn. The 500-hectopascal geopotential height across the regime domain has been rising at about 5.4 geopotential meters per decade since 1979, consistent with the thermal expansion of the troposphere as the planet warms. When the regime identification is repeated with this area-averaged trend removed, the significant Scandinavian Blocking trend vanishes entirely. The apparent increase in summer blocking over Scandinavia, the analysis suggests, is likely primarily a thermodynamic signature of global warming rather than a genuine change in atmospheric circulation.

This finding raises what the author describes as an almost philosophical question: in a warmer climate, should a blocked regime be defined as a circulation anomaly deviating from the climatological background, or as a region of very high pressure with strong impacts on the surface? From an impact perspective, the thermodynamic effect of warming alone can already partly explain the recent increase in high-pressure episodes reported in some studies, and higher background temperatures likely amplify the local consequences of blocking, particularly for summer heat waves. The study also cautions that simple linear trend analyses are highly sensitive to the period chosen, and proposes that non-stationary methods, such as the LOESS filters increasingly used by European meteorological services, will be needed to cleanly separate natural variability from climate change signals in future regime research.

The paper is as much an infrastructure project as a scientific analysis. It provides a comprehensive technical appendix documenting every step of the method, from the Lanczos low-pass filtering to the iterative life cycle identification criteria, and the complete regime dataset from 1950 to the present has been openly released on Zenodo along with auxiliary scripts. The author shows that switching from the older ERA-Interim reanalysis to ERA5 changes the attribution of only a tiny fraction of time steps, confirming the robustness of the framework. With ECMWF and MeteoSwiss now working toward operational use of the year-round regimes in sub-seasonal forecasting and climate monitoring, these seven atmospheric archetypes are poised to move from the research literature into the daily toolkit of forecasters, energy traders, and health planners across Europe.

Subject of Research: Year-round weather regimes and their life cycles, variability, and trends in the North Atlantic European region

Article Title: A life cycle definition of year-round weather regimes in the North Atlantic European region

Article References: Grams, C. M. (2026). A life cycle definition of year-round weather regimes in the North Atlantic European region. Weather and Climate Dynamics, 7(3), 1641-1680. https://doi.org/10.5194/wcd-7-1641-2026

Image Credits: AI Generated

DOI: 10.5194/wcd-7-1641-2026

Keywords: weather regimes, atmospheric blocking, North Atlantic, Europe, ERA5 reanalysis, sub-seasonal forecasting, weather regime life cycles, Scandinavian Blocking, climate trends, geopotential height, heat waves, renewable energy

The research, led by scientists from China University of Geosciences in Wuhan and the National University of Defense Technology, tackles a stubborn problem that has plagued climate assessments of the region. The Western Pacific Subtropical High, or WPSH, is a cornerstone of the East Asian summer monsoon. Its intensity, coverage, and position strongly modulate where summer rainfall falls and how hot the season becomes. When the high shifts westward or strengthens, rainbands can stall over the Yangtze River valley or abandon northern China entirely, while subsiding air beneath the system suppresses clouds and bakes the land below. In recent years, a growing number of extreme precipitation events and heatwaves across East Asia have been traced to anomalous behavior of this single circulation feature, making its long-term trajectory a matter of acute societal concern.

Yet measuring that trajectory has proven surprisingly contentious. In operational practice and much of the research literature, the WPSH has traditionally been defined by a fixed geopotential-height contour, essentially drawing the boundary of the high wherever the atmospheric pressure surface reaches a chosen value. The approach is simple and convenient, but it carries a hidden flaw in a warming world. As greenhouse gases thicken the tropospheric column, geopotential heights rise nearly everywhere, regardless of any genuine change in the circulation itself. A fixed contour therefore encloses a larger area over time even if the high has not dynamically strengthened, producing what many researchers suspected was a spurious trend of intensification baked into the metric rather than the atmosphere.

To escape this trap, earlier studies proposed a patchwork of alternative indices, some built on thermodynamic criteria and others on dynamical ones. The trouble was that these alternatives frequently disagreed with one another, yielding inconsistent trends and divergent spatial footprints for the WPSH under climate change. Without a consensus metric, the community could not say with confidence whether the subtropical high was strengthening, weakening, or merely drifting. That uncertainty is precisely what the new study set out to resolve, and its solution is elegantly physical: stop tracking a contour and start tracking the dynamics that make the subtropical high matter in the first place.

The team’s innovation is a definition based on what they call thermodynamic-dynamic synergy. Rather than fixing the boundary at an arbitrary height value, the method examines the joint probability distribution frequency, or JPDF, of geopotential height and vertical velocity, two fundamental atmospheric variables. By analyzing how these variables co-vary, the technique identifies a dynamically constrained height threshold that corresponds to the characteristic sinking motion, or subsidence, at the heart of the subtropical high. This linkage to descent is not an arbitrary choice. Subsidence is the engine of the WPSH’s weather influence: it suppresses cloud formation, clears the skies, and warms the surface, while the circulation around the high steers moisture-laden monsoon flow and positions the rainbands that deliver East Asia’s summer rainfall.

In essence, the JPDF metric does not simply delineate a region of high pressure or an anticyclonic swirl. It isolates the WPSH’s dynamically active main body, the portion of the system characterized by organized subsidence that is physically tied to East Asian summer temperature and rainfall anomalies. The warming-induced mean rise in geopotential height, which contaminates fixed-contour definitions, is effectively filtered out, while the physical connection to weather is retained. The result is a measure that behaves consistently across a changing climate, a property the older definitions could not guarantee.

The payoff shows up in the validation. Indices derived from the JPDF framework, covering the high’s area, intensity, western ridge point, northwestward extension, and coverage over land, all correlate more strongly with East Asian precipitation and temperature anomalies than conventional measures. The metric successfully captures the upward motion associated with monsoon rainbands on the high’s flank as well as the subsidence linked to high-temperature anomalies beneath and within the system. In other words, the new definition tracks the part of the subtropical high that people actually experience in their weather, which is exactly what any climate index should do.

Armed with this more trustworthy yardstick, the researchers turned to the latest generation of climate models, drawing on projection experiments from the Coupled Model Intercomparison Project Phase 6, the international standard for future climate scenarios. The projections reveal robust increases in both the area and the intensity of the WPSH under future warming, together with a northwestward extension of its boundary and greater coverage over East Asian land areas. Crucially, these changes are not artifacts of a few outlier models. They are supported by strong inter-model consensus, particularly under medium- and high-emission scenarios, giving the findings a level of confidence that earlier, metric-dependent studies could not achieve.

The study goes beyond gradual, century-scale trends to examine the behavior of individual extreme events on synoptic timescales, the episodes when the subtropical high surges to record strength and lingers over a region. Under continued warming, the analysis projects that such record-breaking WPSH events will become more frequent, more intense, and longer-lasting. Strong cases may push farther inland over East Asia, and they become more likely to combine with large-scale high-pressure systems over Eurasia, forming continental-scale heat domes. That combination raises the risk of persistent, record-shattering heatwaves and, at the same time, intense monsoon rainfall extremes, since a stronger, displaced high can funnel moisture into concentrated rainbands while its core desiccates and heats the land it covers.

The implications reach well beyond atmospheric science. The changes imply increasing exposure to monsoon rainfall extremes and heatwaves that lie beyond the range of historical experience, a challenge for infrastructure, agriculture, public health, and disaster preparedness across densely populated East Asia. By supplying a physically consistent framework for diagnosing WPSH-related weather anomalies and assessing the system’s long-term evolution under anthropogenic warming, the study provides a foundation for future extreme-event attribution, for evaluating how well climate models represent this pivotal circulation, and for quantifying climate risk in one of the world’s most monsoon-dependent regions. In effect, the research converts a long-standing measurement controversy into a clear, actionable warning: the atmospheric engine behind East Asia’s summer weather is shifting into a more powerful and more hazardous regime, and the new JPDF-based lens allows scientists to see that shift unambiguously for the first time.

Subject of Research: Response of the Western Pacific Subtropical High to anthropogenic greenhouse warming

Article Title: Climate change causes the Western Pacific Subtropical High stronger and more extreme

Article References: Climate change causes the Western Pacific Subtropical High stronger and more extreme. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: Western Pacific Subtropical High, climate change, East Asian monsoon, heatwaves, extreme rainfall, geopotential height, subsidence, CMIP6, National Science Review, monsoon rainfall extremes, climate models, East Asia