Status: Material from: Linus, Ivan, Tim ...
1. Impact
In the end of October severe weather hit different parts of Italy, Austria and France, with casualties. On 29 October a cyclone brought a storm surge to Venice and extreme rainfall and precipitation over the southern Alps. Extreme winds were also registered in several countries. In parts of Italy a complete shutdown of public services was triggered by issue of a red warning related to the storm.
Storm surge generated as a result of the explosive cyclogenesis over the western Mediterranean affected Venice causing massive flooding. While floods of this sort are relatively common in Venice, this time the water levels were 1.54 m above the reference level which ordered this storm as the 4th most intense one in history. The meteorological contribution to this level was more than 1.5 m. In this case the peak intensity of the storm coincided with the astronomical tidal minimum. If this had happened several hours earlier or later, it could have been the record flood. Luigi Cavaleri from Instituto di Scienze Marine, also shared with us that significant wave height (SWH) reached 6 m at their oceanographic tower in the northern Adriatic sea. In a similar storm of 22 Dec 1979 (ranked as 2nd in history) crests reached 9 m above the sea surface. The maximum of SWH was in the late afternoon and evening hours (at about 18 UTC) during the tidal minimum and later on the waves quickly diminished in height.
2. Description of the event
The plots below show analyses of MSLP and short forecasts of 6-hour total precipitation covering 27 October 00UTC to 30 October 12UTC, every 12th hour.
The plots below show analyses of z500 (contour) and T850 (shade) covering 27 October 00UTC to 30 October 12UTC, every 12th hour.
3. Predictability
3.1 Data assimilation
3.2 HRES
The plots below show 24-hour accumulated precipitation 29 October 06UTC to 30 October 06UTC in observations (first plot) and forecasts with different initial times. Already the forecast 5-days before the event (second last plot) had a similar precipitation pattern as the last forecast before the event (second panel). All forecasts underestimated the maximum over the southern Alps by ~half.
The plots below show forecasts of 24-hour maximum wind gusts valid 29 October (we are missing wind gusts in synop observations from Italy).
3.3 ENS
The plots below show EFI and SOT 3-day precipitation (27-29 October).
The plots below show EFI and SOT for 24-hour maximum wind gusts valid for 29 October.
The plot below shows the forecast evolution of 24-hour precipitation (29 October 06UTC to 30 October 06UTC) for a 2x2 degree box centred on Ponte Nelle Alpi (46.2N,12.3E). The blue box and whisker shows the ensemble, red dot HRES and red box-and whisker model climate. The right plot shows the daily precipitation (thin lines) and accumulated precipitation since 1 October (00-00UTC) for observation stations inside the box (black line), 24-hour HRES forecasts (blue) and ensemble mean from +144h and +240h (pink and green). This plot shows the short-range HRES, which correspond to the second last red dot in the left panel) Note that the forecast values in the two panels are not identical as the right panel is for forecasts at observation stations and slightly different accumulation periods. Anyway the right panel indicates that the short-range forecast had a good magnitude of the rainfall 29-30 October.
3.4 Monthly forecasts
The plots below shows weekly anomalies of precipitation valid 29 October to 4 November.
The plots below shows weekly anomalies of MSLP valid 29 October to 4 November.
3.5 Comparison with other centres
3.6 EFAS
The plots below show EFAS forecasts for a point in northern Italy.
4. Experience from general performance/other cases
5. Good and bad aspects of the forecasts for the event
- Extreme rainfall captured from 7 days before (22 October) and with high confidence from 24 October. However external reports provided to ECMWF indicate that HRES shorter range forecast over parts Italy were unsatisfactory - totals were too large.
- Though not discussed directly above, it is clear from separate feedback that details of the strong winds and the very rapid deepening of the low centre were not handled particularly well by the IFS. It seems that with regard to these details this was a low predictability case where very high resolution was critical for capturing the full ferocity of the low.