Discussed in the following Daily reports:
Material from: Linus. Esti, ...
Discussed in the following Daily reports: |
On 22–23 August 2026, an unusually intense, rapidly developing cyclone moved northeast from the Alpine/Polish region towards the Baltic states, producing extreme precipitation and damaging winds. Forecast predictability was unusually short: deterministic IFS displaced the heaviest precipitation roughly 300 km eastwards by moving the cyclone too quickly, whereas AIFS represented its position and rainfall substantially better. Observations showed exceptionally heavy rainfall, with some Latvian stations receiving around half their average monthly rainfall in one day, alongside winds exceeding 107 km/h in Riga.
We will focus on 24-hour precipitation on 22 August in a 1x1 degree box centred on Western Latvia (57.5N, 22.5E).
The plots below show analyses of MSLP and 6 hour rainfall from 20 August 00UTC to 24 August 00UTC, every 12th hour.
The plots below show z500 and T850 from 18 to 24 August 00UTC, every 24 hour.
Observations and analysis for the event
Control forecast (IFS 9-km resolution)
The IFS control forecasts from 20 and 21 August placed the precipitation too far east. Earlier forecasts have only a weak signal for the precipitation in the Baltic states.
DestinE (IFS 4.4km resolution)
AIFS-single (AIFSv1.0 ~0.25 resolution)
AIFS-single had a much better signal in the forecasts from 20 and 21 August compared to IFS control. But earlier forecasts had the system far to the east as in IFS control.
AIFS-ENS control (~0.25 degree resolution)
Hybrid control
The Hybrid control looks like the IFS control for this case.
EFI (based on IFS 9-km ensemble)
The plots below shows the EFI and SOT for 24-hour precipitation on 22 August. Also there is it clear that the signal for the Baltic states appeared on 21 August, only one day before the event.
Forecast Evolution plot