Status: Finalized Material from: Linus

Discussed in the following Daily reports:

http://intra.ecmwf.int/daily/d/dreport/2013/12/02/sc/

http://intra.ecmwf.int/daily/d/dreport/2013/12/03/sc/

http://intra.ecmwf.int/daily/d/dreport/2013/12/04/sc/

http://intra.ecmwf.int/daily/d/dreport/2013/12/05/sc/

 

This case is documented in ECMWF Newletter 139 (Hewson et al):   http://www.ecmwf.int/sites/default/files/Newsletter_139_smaller.pdf

1. Impact

On the 5 December the cyclone swept from Scotland, over the North Sea and hit Denmark, Germany and Sweden. Problems were caused both by the wind and the following storm surge. The surged reached 6 metres in Hamburg for example and was the highest along the England east-coast for 60 years. In the aftermaths of the cyclone a blizzard hit Sweden.

2. Description of the event


The figures above show analyses of MSLP and Eady Index from 4 December 00UTC to 6 December 00UTC, every 12th hour. The cyclone developed south of Iceland after 12 UTC on the 4 December in a confluent flow with high baroclinicity.

Satellite images (Meteosat from yr.no) from 5 December 00UTC to 21TUC, every 3 hours.

The figures above shows maximum observed wind gusts (left) and mean wind (right) during 5 December (24-hour period).

3. Predictability

 

3.1 Data assimilation

 

3.2 HRES


The series of figures above shows HRES forecasts of the maximum wind gust during the 5 December and the MSLP valid 12 UTC. The colour scale is the same as for the observations in the previous section. The first forecast is from 5 Dec 00 UTC and the following is 4 Dec 00UTC and so on (with 1 day apart). Already the forecast from 9 days before (second last panel) had a cyclone in the area, but further west than the latest forecast.

These figures snow 10-metre wind speed for Ekofisk in the North Sea(56.3N, 3.1E) and Torsminde in western Denmark (56.4N, 8.1E). The HRES forecast is in red and observations in blue. The forecast is initialised 4 December 00UTC. One interesting feature in the model, for both sites, is the large variations from hour to hour close to the peak of the wind speed.

3.3 ENS

The figures above shows the EFI for wind gusts valid on 5 December from 1 (5 Nov 00UTC), 2, 3, 4, 5 days before the event. The figures also includes the values for the 99th percentile.

Probability of maximum wind gust>33 m/s on the 5 December. The first forecast is from 5 December 00UTC and the following are with one day apart. Up to 3.5-4 days before the event, the probability of hurricane wind gust where close to 100% in western Scotland and over the North Sea.

CDFs of maximum wind gusts for the 5 December for Ekofisk in the North Sea(56.3N, 3.1E) and Torsminde in western Denmark (56.4N, 8.1E). Ekofisk reported between 30 and 40 m/s and Torsminde between 40-50 m/s.

Dalmatian plots of cyclonic features where the colour indicates the maximum wind speed at 850 hPa within 600 km radius. The first forecast is from 5 December 00UTC and the following are with one day apart.


Strike plots of cyclonic features where the maximum wind speed at 850 hPa exceeds 60 kt. The first forecast is from 5 December 00UTC and the following are with one day apart.


3.4 Monthly forecasts

The figure above shows the MSLP anomaly for 2 to 8 December. The top panel shows the anomaly from the analysis and the other panels from the monthly forecasts. In the analysis we see a strong pressure gradient in north-western Europe, in the region where the storm was more intense (although the British Isles are in the high-pressure region in the weekly anomaly). This pattern was captured two weeks in advance.

3.5 Comparison with other centres

-

4. Experience from general performance/other cases

5. Good and bad aspects of the forecasts for the event

  • Good indication of very strong winds 3-4 days in advance.
  • The cyclone was present in the forecasts 8-9 days before the event.
  • Did the cyclone move too slow in the forecast?