Forecasting Sting-Jet Storm Kristin over mainland Portugal: A Multi-Resolution and Multi-Model Perspective

On 28 January 2026, a rapidly intensifying cyclone named “Kristin” struck mainland Portugal, causing widespread damage, particularly in the central region of the country. Extreme wind gusts associated with the storm’s sting jet destroyed houses and industrial facilities, caused at least 13 fatalities, damaged key electrical and communication infrastructures, and severely impacted the national economy. This event ranks among the most extreme ever recorded in mainland Portugal in terms of wind gusts, setting new records that exceeded 200 km/h at weather observation stations.

A sting jet is a narrow, transient airstream that descends from the cloud head, accelerating towards the surface and generating damaging winds. These events are often challenging to forecast due to their short duration, small spatial scale, and rapid intensification. In this context, we present an overview of different numerical weather prediction (NWP) models to forecast this storm: IFS at 9km resolution (ECMWF’s HRES), IFS at 4.4km resolution (the DestinE initiative’s global Digital Twin), AROME at 2.5km (IPMA’s operational limited area model), and AROME at 500m resolution (the DestinE initiative’s regional component, produced by project DE_330, the so-called Destination Earth on Demand Extremes, or DEODE).  

All models capture the rapid deepening of the cyclone, with central pressure decreasing by more than ~10 hPa in approximately 12 hours. However, marked differences are detected in both the timing and magnitude of the minimum pressure. The DEODE simulation predicts the deepest cyclone, whereas AROME estimates a weaker cyclone, not necessarily in terms of deepening rates or wind speeds, but rather in minimum MSLP values at the center of the cyclone. 

There is a general agreement among the forecasts regarding wind gust values and the accumulated precipitation pattern associated with both the cloud head and the cold front. However, AROME better captures the location of the observed maximum gusts, whereas both IFS and DEODE predict the extreme winds further north than observed at the weather stations. In contrast, AROME tends to overestimate the accumulated precipitation, while both IFS configurations better represent precipitation amounts over the affected regions.

The intercomparison between these NWP configurations and observations provides valuable insights into the challenges of forecasting extreme events, highlighting the importance of high-resolution modelling to better represent storms such as “Kristin”.