During autumn 2024, different parts of Europe experienced severe rainfall, such as Spain, France, and Italy. While there are many important aspects to evaluate for all these events, we will here focus on the rainfall related to Storm Boris that affected central Europe.
Between 13 and 16 September 2024, the low-pressure system Boris brought record-breaking rainfall that led to severe flooding in parts of Austria, Czechia, Slovakia, Poland, Romania and Hungary. The flooding caused widespread damage and disruption with costs estimated in the billions of euros. Sadly, 27 lives were reported lost as a result.
Based on ECMWF forecasts and other sources of information, national meteorological and hydrological centres and other civil authorities within Member and Co-operating States issue weather, flood and other emergency warnings at the national and regional level.
About Storm Boris
During 11 and 12 September cold air was driven south across western Europe into the Mediterranean to meet warm, moist air to the east, creating a strong temperature contrast across central Europe. A low-pressure system formed along the interface over northern Italy on 11 September and was named Boris. It moved east and then north, and an upper-level cut-off low developed. This followed the classical 'Vb' ('five-b') pattern, which typically leads to large amounts of precipitation in the eastern Alpine region. High-pressure ridges to the northeast and the northwest helped to anchor the system and its very active rainfall zone for several days over central Europe. The largest three-day rainfall totals (13 September 00 UTC to 16 September 00 UTC) occurred over eastern Austria and along the border between Czechia and Poland (see the top panel of the first figure). A three-day total of 442 mm in the mountains in northern Czechia was the highest value found in observed station data available at ECMWF.
On top of the precipitation observations included in SYNOP weather station reports, ECMWF regularly receives additional observations from Member and Co-operating States. Triggered by this case, observations from Czechia started to be collected as well, and Slovakia also provided additional observations for evaluation purposes at ECMWF. The benefit of these additional observations for the evaluation of this case is clearly seen in the top panel of the first figure.
Prediction of the rainfall from Storm Boris
The bottom panel of the first figure shows the mean precipitation in the box outlined in the observation map, covering large parts of Czechia, northeastern Austria, and parts of Poland and Slovakia for 13 to 16 September. Approaching the event, the signal in the ensemble (ENS) became gradually stronger. The ensemble mean for the Integrated Forecasting System (IFS) ensemble passed the 99th percentile of the model climate around 6 September, a week before the onset of the event. After a slight dip in the signal on 10 September 00 UTC, the ensemble mean was above the model climate maximum (based on 1,800 re-forecasts valid at the same time of the year as the event). One can note that the Artificial Intelligence Forecasting System (AIFS) ensemble (currently running with a 1° resolution) did not suffer from the drop in the signal on 10 September, giving a more consistent evolution of the signal compared to the IFS.
Role of sea-surface temperature
The very warm summer of 2024 over southeastern Europe (see the previous Newsletter: https://www.ecmwf.int/en/newsletter/181/news/extremely-warm-summer-southeast-europe) resulted in strong anomalies of sea-surface temperature (SST) in the Mediterranean and the Black Sea. It raises the question of whether these anomalies increased evaporation and contributed to the extreme rainfall during Storm Boris. One way to test this is to run forecast experiments where, to constrain the dynamics, the wind is nudged to ECMWF's ERA5 reanalysis globally, while the moisture-related variables are freely evolving. (Note that temperature and surface pressure will be affected by the nudging due to geostrophic adjustments.) In a first test, we ran three lower-resolution 11-member ensembles initialised on 1 September, where one did not use any nudging (No Nudge), one used nudging together with the observed SST (Nudge realSST), and one used nudging but with the climatological SST in the Mediterranean and the Black Sea (Nudge climSST). The resulting ensemble distributions are included in the forecast evolution plot, placed on 1 September. While the 'No Nudge' experiment resembles the longest (15-day) ensemble forecasts, both nudged experiments reproduce well the extreme seen in the short-range forecasts for the event. But by replacing the SST with the climatology, the ensemble mean precipitation in the box is reduced by around 10 percent. It indicates that the warm SST contributed to the extreme precipitation. However, one aspect not tested by this setup is the feedback on the atmospheric flow of the warm SST and increased latent heat release, which could have an additional effect.
Flood forecasts through EFAS
The European Flood Awareness System (EFAS) is managed by the European Commission's Joint Research Centre and is part of the EU's Copernicus Emergency Management Service (CEMS). As the computational centre, ECMWF generates flood forecasts using an open-source hydrological model, LISFLOOD, based on meteorological forecasts from ECMWF, the German National Meteorological Service (DWD) and the COSMO Limited-Area Ensemble Prediction System consortium, and meteorological and hydrological observations.
The first strong signal of a severe flood associated with Storm Boris was suggested by EFAS on 9 September with the 12 UTC forecast run. Large parts of the Oder River basin (in western Poland and the border with Czechia and Germany) were forecast to be at risk of river discharge exceeding a 20-year return period. The flood signal extended to the upper Danube (in Austria) with the 10 September 12 UTC forecast run. The hydrographs shown in the second figure, based on forecasts initialised on 12 September 00 UTC, show the risk of exceptional flood levels – with some ensemble members well beyond the maximum level within the model climatology (shown by the dashed line in the figure).
In this article we showed a few examples of evaluating forecasts of Storm Boris, mainly focusing on large-scale aspects of the precipitation and on flood forecasts. However, there are many more details, especially about local impacts, to be explored.
