Flash drought is a rapidly developing form of drought characterised by anomalously rapid transitions and accelerated by soil moisture depletion, resulting in pronounced vegetation stress. It typically occurs during the growing seasons at intraseasonal time-scales. This phenomenon has been attracting little scientific research attention in western Europe because of its historically low frequency of occurrence, despite its potentially severe impacts. Such impacts are projected to increase in severity due to increase from atmospheric thirst which has the potential to lead to more frequent and intense flash droughts in the future. This study investigates whether the drought event that occurred over western Europe during spring-summer 2022 can be deemed as a flash drought. Using the standardised evaporative stress ratio derived from the fifth-generation global climate reanalysis data ERA5 from the European Centre for Medium-range Weather Forecasts (ECMWF) over 1980-2024, several clusters emerged as flash drought, mainly over many parts of France, Switzerland, southern England but also in some areas of Spain, Italy, and Germany. These clusters occur between May and July 2022 and all terminated between August and September. Large-scale analysis highlights a mid-tropospheric circulation anomaly pattern characterised by a subsidence over central western Europe which is part of a stationary Rossby wave train persisting from May to July, coinciding with the flash drought onset period. This unusually strong subsidence leads to extreme hot and dry conditions at the surface, favouring flash drought development. This research illustrates the value of applying targeted flash drought frameworks to better understand the nature of rapid drying. In an increasingly water-insecure future, such approaches will be critical to the long-term resilience of drought-prone communities in both Europe and further afield.
The Global Observing System (GOS) includes drifting buoys with barometers whose surface pressure observations are essential for estimating the current state of the atmosphere and ocean for Numerical Weather Prediction. The Global Drifter Program, a scientific project based at the Scripps Institution of Oceanography, provides the majority of these buoys (barometer drifters). However, regions with sparse in situ observation coverage can limit weather forecast skill. We describe the Mediterranean Extreme Events Experiment (M3E) - a project with partners from Europe and the United States - which, via the deployment of barometer drifters into the data-sparse Ionian Sea and central Mediterranean, aims to (1) improve the weather forecasting of extreme events, (2) understand cyclone development across the Mediterranean, and (3) contribute to the Global Atmospheric River Reconnaissance Program (GARRP). The deployments started in September 2025, and to date M3E has released 20 buoys, some of which were available in January 2026 when storm Harry affected the Mediterranean region. Harry's synoptic conditions and extreme weather, including three-day precipitation totals over 400 mm in southern Italy, are discussed. The European Centre for Medium-Range Weather Forecasts Integrated Forecasting System was skilful in resolving Harry's large-scale circulation and the precipitation in southern Italy seven days ahead. There were forecast challenges in capturing the North African precipitation due to mesoscale low-pressure systems. The benefit of the M3E buoys was shown using the Forecast Sensitivity to Observation Impact analysis, and an Observing System Experiment suggested sensitivity of the mesoscale features in storm Harry to the buoy observations in the Ionian Sea. These extra observations positively affected the extreme event forecasts and provide evidence for supporting future M3E deployments. In turn, this would help to sustain the GOS surface network, leading to socioeconomic benefits via improved weather forecasting of extreme weather systems, and contribute to the goals of GARRP.
Subseasonal prediction of extreme weather is an important but challenging task. It depends on both initial and boundary conditions. Using a conditional nonlinear optimal perturbation method, we examine the nonlinear joint effect of Arctic atmosphere initial perturbations and sea ice boundary perturbations on subseasonal prediction of Ural blocking. The influences of sea ice and atmosphere are highly nonlinear and non-additive. The coordinated effects of sea ice and atmosphere can exert stronger influence than the individual perturbations or their linear summation. Our results also reveal the sea ice perturbations in the Barents and Okhotsk Seas, along with atmospheric perturbations over the North Atlantic and Bering Strait, are most critical. Further diagnosis shows that the nonlinear joint effects influence the air-sea heat transfer there, and enhance zonal wind through a temperature gradient, thereby affecting energy transport and weakening the Ural blocking. The above results provide theoretical insights into potential predictability improvement of subseasonal prediction of midlatitude extreme events.
Convection that overshoots the tropical cold‐point tropopause can irreversibly transport water vapour, providing an important pathway for hydration of the tropical lower stratosphere. We formulate an idealised model of this process by treating a convective overshoot as a buoyant plume penetrating a stably stratified layer carrying tracers that represent vapour and ice. We use large‐eddy simulations to study a scaled‐down model flow, which resolves turbulent tracer transport better compared with earlier numerical studies. The vapour and ice tracers evolve according to a minimal moisture model, with equal and opposite rates of condensation and sublimation and a fixed sedimentation velocity for the ice tracer. We explore the behaviour of the model in two experiments where the sedimentation velocity and partitioning of moisture into vapour and ice are varied. To understand the influence of mixing, we extend a formalism that captures the simultaneous modification of buoyancy and tracer concentration to include terms representing the effect of moist processes. We find that our model is an effective representation of the hydration mechanism observed in the tropical tropopause layer when the plume penetrates the stratified layer carrying large amounts of ice, which are mixed with the warmer environment and sublimated to form vapour. Hydration of the stratified layer beyond the direct transport of vapour by the plume requires ice to remain in suspension during mixing in the plume cap, which is controlled by the relative sizes of the sedimentation velocity and turbulent eddy velocity.
Abstract Building on the stochastic particle flow filter (SPFF) introduced in Part 1, this study develops a stochastic particle flow smoother (SPFS) for fully nonlinear data assimilation in high‐dimensional dynamical systems. Whereas the SPFF successfully provides unbiased ensemble spread and handles non‐Gaussian distributions in sequential filtering, many applications require historical state estimates over an extended assimilation window. The SPFS is formulated as a strong‐constraint smoother in which observation information is propagated backward in time to optimize the initial ensemble. In this study, two backward‐propagation strategies are examined. One uses the exact adjoint model, and the other uses an ensemble‐averaged sensitivity matrix approximation. Experiments with the 1000‐dimensional Lorenz 96 system demonstrate that the SPFS with an adjoint model achieves superior trajectory fits compared with the filter, particularly in strongly nonlinear regimes, by allowing particles to follow individual gradients and represent multimodal posteriors. In contrast, the ensemble‐averaged sensitivity matrix approximation, though computationally appealing, degrades in highly nonlinear settings and requires shorter assimilation windows. These results establish the SPFS as an effective solution to the fully nonlinear smoothing problem in high‐dimensional systems.
