Storm-resolving climate models are gaining attention for their improved simulation of mesoscale processes. Yet, how finer resolution benefits synoptic-scale phenomena remains unclear. We assess atmospheric blocking in the Next Generation Earth Modelling Systems, European Eddy-Rich Earth System Models, and Destination Earth projects, identifying key bias drivers and their response under a high-emissions climate change scenario.

Forecast biases of winter extratropical cyclones are quantified by varying diabatic heating intensity. A southwest shift and underestimated intensity are found in the strong heating group. The weaker heating group mainly shows an intensity bias. Specific biases are identified for wind, moisture, temperature, and upper-level circulation fields. Findings highlight that representing moist processes and their interaction with atmospheric dynamics is a key area for future model developments.

The Quasi-biennial Oscillation (QBO) dominates variability in the tropical stratosphere, & it impacts surface climate in several parts of the world. However, climate models have been shown to systematically under-estimate the influence of the QBO. Here, we re-evaluate this finding using much larger ensemble sizes than have been previously available based on four separate models. We find that the models are comparatively more successful in capturing QBO influences than reported by previous work.

Human-caused warming intensified the late October 2024 Valencia extreme precipitation event. Using storyline simulations, we compared today’s climate with a cooler past climate while keeping the synoptic weather pattern the same. Warmer air and sea temperatures increased moisture, instability, and rainfall, showing that climate change amplified an already extreme storm.

We present an unsupervised clustering methodology and post-clustering analysis framework that identifies recurring monsoon rainfall patterns and quantifies their transitions. Analysis reveals breaks are prolonged while monsoon depression are transient. The decomposition of frequency and intensity changes quantifies their contributions in long-term rainfall changes.

We improve the detection of long corridors of intense moisture transport, known as atmospheric rivers, that affect Antarctica. Our new method can track these rivers even when they curve, cross the South Pole, or break into smaller parts. We show that they cause even more heavy snowfall and warm weather on the continent than previously thought, clarifying their role in shaping Antarctic climate and ice loss.

Textbooks describe the atmosphere’s north–south motion as closed circulation cells. This study shows that the water cycle also drives a subtle one-way flow of air, moving it from the humidified subtropics toward regions dried by rain and condensation. This hidden transport helps explain gradients of inert gases and suggests that large-scale atmospheric circulation may be more open than commonly assumed.

Predicting the start of the summer monsoon over the South China Sea is important because it affects rainfall, farming, and flood risk across East Asia, but it is difficult to forecast far in advance. We tested a new way to define monsoon onset based on a lasting shift in large-scale weather patterns rather than a single wind threshold. Using seasonal forecast data, we found that this approach improves prediction skill and gives a more physically consistent picture of monsoon onset.

Sting jets are air flows that can form in cloudy and unstable regions of intense windstorms. They can cause damaging wind gusts as they descend and accelerate while releasing their instability. We use a numerical simulation of Storm Ciarán to identify the individual cloud processes, such as condensation and sublimation, driving the onset of this instability. We reveal their complex interplay and highlight the unique properties of the three-dimensional environment in which a sting jet develops.

Mass coral bleaching on the Great Barrier Reef is driven by unusually warm ocean temperatures, but cloud cover also plays a critical role by regulating how much sunlight reaches the ocean. This study examines daily patterns of clouds and rainfall over north-east Queensland during bleaching seasons using satellite, radar, and regional weather data. Clouds and rainfall patterns differ between land, ocean and wind conditions, helping improve understanding of processes influencing reef heat stress.

Tropical cyclones that move into the midlatitudes become extratropical cyclones, so-called extratropical transition. In this study, we detected transition events in the Southern Hemisphere based on the model data. We found that weather patterns during the transition are different from case to case. In some cases, strong tropical cyclones interact with the midlatitude flow and lead to great changes in weather nearby.

By analysing data from 1940 to 2024, we found that upper atmosphere conditions in early winter directly influence the number of North Sea storms occurring weeks or months later. We used a climate model and machine learning to improve these forecasts. Results show that December patterns can predict January storminess with high accuracy, with effects persisting for up to 60 d. Stratospheric data and machine learning together improve winter storm prediction for this region.

We use the Large Ensemble Single Forcing simulations to assess the role of individual forcings to the Mediterranean drying and to clarify the dynamical origin of the model’s prediction variability. A more pronounced North Atlantic warming hole, a stronger stratospheric polar vortex, and a larger poleward shift of the subtropical jet correlate with a stronger drying trend. Aerosols had a detectable influence on Mediterranean climate. Hence, their removal may have an impact in future decades.

This study aims to seasonally predict springtime tornado activity using a weather-regime-based hybrid model and to identify the physical sources of predictability to explain the results. Tornado outbreaks, days with several tornadoes, exhibit model skill and should be a primary focus of future work given their societal impacts. Low-frequency climate modes are important sources of predictability for weather regimes, providing forecasts of opportunity for springtime tornado outbreaks.

This study investigates yearly heat extremes simulated with the global climate model CESM2. A trajectory-based method is used, which allows quantifying the contributions to temperature anomalies from advection, subsidence, and diabatic heating. The results show that the magnitude of CESM2 heat extremes agrees fairly well with ERA5 reanalyses, but it is often for the partly wrong physical reasons.

A key pattern of tropical weather, the Madden-Julian Oscillation, has become significantly harder to predict since the late 1990s. We discovered this by comparing forecasts from major models across two time periods. The decrease in forecast skill is linked to changes in large-scale climate patterns, not just model errors. This means to improve long-range weather forecasts, models must better simulate how these large-scale patterns interact with tropical weather.