Status Status:Ongoing analysisFinalized Material from: Linus, MeteoFrance (via Florence)
...
The plot below shows the observed 24-hour percipitation from the EFAS network (6 October 06z - 7 October 06z). It seems like we did not get any observations from the narrow band of very high precipitation rates.
The next plot shows gridded precipitation values interpolated from rain gauges (provided from MeteoFrance).
3. Predictability
3.1 Data assimilation
3.2 HRES
The plots above shows the HRES forecast for 24-hour precipitation (06-06) from 6 Oct 00z (left) and 5 Oct 12z (right). As seen in the report from MeteoFrance, the band of very high precipitation rates are missing.
3.3 ENS
Gallery includeLabel prob20_ec sort comment title Prob >20mm/24h
The plots above shows the probability of more than 20 mm in 24 hours (6 Oct 06z-7 Oct 06z) from different lead times. Also the ensemble lacks the signal in the worst affected area.
Gallery includeLabel prob20_leps sort comment title Prob >20mm/24h
The plots above shows the probabilities from COSMO-LEPS. Here the ensemble from some of the initial times has a signal over the worst affected region but on other hand lacks the signal on the southern tip of France.
3.4 Monthly forecasts
3.5 Comparison with other centres
...
4. Experience from general performance/other cases
- Compare this case with 201409 - Rainfall - S.W France (flash flood in Montpellier)
5. Good and bad aspects of the forecasts for the event
- The ECMWF forecasts missed the heavy rainfall over l’Héraul, while the MeteoFrance models got it.
- Due to model resolution? (Orographic precipitation)