The research team built their simulation by dynamically downscaling a Community Earth System Model run spanning 1821 to 2100 with the Weather Research and Forecasting model, known as WRF. The global model supplied the large-scale circulation, while WRF resolved the regional detail at 20-kilometre horizontal resolution with hourly output, 49 vertical levels and a domain covering the whole Euro-Atlantic area. This approach matters because global climate models, typically running at resolutions of 100 kilometres or more, struggle to capture the small, short-lived and topographically trapped cyclones of the Mediterranean, which are smaller and shallower than their Atlantic and Pacific cousins but capable of devastating impacts nonetheless.

To track the storms, the researchers applied a cyclone detection algorithm to the 850-hectopascal geopotential height field, regridded to one-degree resolution to filter out weak and unphysical lows over the region’s complex terrain. Cyclones had to show a minimum pressure gradient, persist for at least twelve hours and avoid grid cells above 1000 metres of orography. Around a thousand cyclones were identified per region per period. The team then defined extreme cyclones by the 95th percentile of six-hourly accumulated precipitation and 850-hectopascal wind speed within a 500-kilometre radius of each storm centre, ranking every system by how many standard deviations it departed from the mean.

The headline finding concerns frequency. Comparing the winter half-years of 1821 to 1880, a proxy for the pre-industrial climate, with those of 2039 to 2099 under the high-emission RCP8.5 scenario, the simulation shows a decline in mean cyclone frequency across nearly the entire Mediterranean. The drop is largest over the long-recognised hotspot stretching from Italy towards the Levant and over the Anatolian Plateau, where frequencies fall by up to 0.02 cyclones per day, a relative decrease of roughly one-third. Earlier studies with global models attributed this decline to reduced baroclinicity, increased static stability and a positive shift in the North Atlantic Oscillation, and the new regional results confirm that picture at much finer scale.

Frequency, however, is only half the story. The study reveals that precipitation extremes within cyclones peak about six hours before the storm reaches its minimum core pressure, while wind extremes coincide with the pressure minimum itself. In the western Mediterranean, future extreme precipitation cyclones intensify significantly during their most intense phase, with six-hourly rainfall increasing by roughly five millimetres north of the storm centre twelve hours after peak precipitation, a doubling compared with pre-industrial storms. In the eastern Mediterranean, by contrast, cyclone-related precipitation remains roughly unchanged, and the storms themselves shift southward over the warmer sea, a statistically significant migration that concentrates future hazards over open water.

The physical mechanism behind the western intensification lies in the vertical structure of the storms. Vertical cross-sections through the cyclone cores reveal so-called potential vorticity towers, columns of high potential vorticity air that extend through the entire troposphere, fed from above by stratospheric intrusions and from below by diabatically produced potential vorticity generated where condensation releases latent heat. In future western Mediterranean storms, equivalent potential temperature in the warm sector rises by up to nine degrees Celsius, reflecting a warmer and moister atmosphere, and potential vorticity values in the lower part of the tower increase by up to one unit, roughly fifty percent. This enhanced low-level potential vorticity strengthens the cyclonic circulation, allowing the storms to deepen more rapidly and remain more intense through their mature phase.

Wind extremes tell an even more consistent story. Unlike precipitation, wind speed extremes intensify in both the western and eastern Mediterranean. The explanation lies not in the low levels but aloft, in the behaviour of the jet stream. Composite analysis of the 200-hectopascal wind field shows that extreme wind cyclones typically form on the northern edge of the jet and, as they mature, migrate into the left exit region of a jet streak, the zone where upper-air divergence promotes rising motion and favours cyclone development. Under future conditions the subtropical jet strengthens by up to ten metres per second, so mature wind cyclones find themselves positioned beneath a substantially more powerful jet streak, a configuration that fuels their intensification.

The study also examined compound events, cyclones that simultaneously deliver extreme precipitation and extreme wind, whose combined damage can exceed the sum of the individual hazards. These compound storms cluster over the warm waters of the Tyrrhenian and Adriatic Seas and over northern Italy in the west, and over the Ionian and Aegean Seas in the east. Their future behaviour largely mirrors the individual extreme categories: precipitation in compound events follows the west-intensifies, east-unchanged pattern, while their wind response resembles that of pure wind extremes, with one notable exception. In the eastern Mediterranean, the wind speed of compound extreme cyclones actually decreases in the future, a divergence the authors cannot yet fully explain but which highlights how sensitive compound hazards are to regional dynamics.

Seasonality, remarkably, remains stable. Precipitation extremes peak in autumn when the sea is still warm enough to feed moisture into passing storms, wind extremes occur more evenly across the colder months, and compound events favour the early winter half-year. None of these patterns shifts significantly between the pre-industrial and future periods, meaning that societies will face the same seasonal rhythm of risk, only amplified. The authors validated their simulation against the ERA5 reanalysis for 1981 to 2010, finding that WRF captures the main cyclone hotspots and storm structures well, though it overestimates precipitation and wind intensity in the eastern Mediterranean, where it also produces more cyclones than observed.

The implications are sobering for a region already identified as a climate change hotspot. A drier Mediterranean with fewer storms might sound like relief, but the study shows that mean and extreme precipitation are decoupling: while average winter rainfall declines, the rain that does fall increasingly arrives in concentrated, violent bursts delivered by fewer but fiercer cyclones. Stronger winds across the whole basin compound the threat of coastal flooding, which these storms are known to drive. The authors caution that their analysis rests on a single simulation member, and they call for ensemble regional downscaling and kilometre-scale global models to firm up the projections. Even so, the message is clear: for the Mediterranean, the future of storm risk lies not in how often the cyclones come, but in how hard they hit when they do.

Subject of Research: Projected changes in extreme Mediterranean cyclones and their wind, precipitation and compound hazards under climate change

Article Title: Mediterranean cyclones from pre-industrial to future climate: changes in extreme wind, precipitation and compound precipitation–wind events

Article References: Doensen, O., Messmer, M., Dolores-Tesillos, E., & Raible, C. C. (2026). Mediterranean cyclones from pre-industrial to future climate: changes in extreme wind, precipitation and compound precipitation–wind events. Weather and Climate Dynamics, 7(3), 1899-1917. https://doi.org/10.5194/wcd-7-1899-2026

Image Credits: AI Generated

DOI: 10.5194/wcd-7-1899-2026

Keywords: Mediterranean cyclones, climate change, extreme precipitation, extreme wind, compound events, RCP8.5, regional climate modelling, WRF, potential vorticity, jet stream, cyclogenesis, Weather and Climate Dynamics