Status: Finalised Material from: Linus


 


1. Impact

On  the 7-9 January south-western part of the Alps has got a lot of snow, and several of the ski resorts have been cut off (http://www.bbc.co.uk/news/world-europe-42623293). Reports on internet talks about 2-3 metres in ski resorts like Tignes (Fr), Bonneville-Sur-Arc (Fr), Cervinia (It) and Zermatt (Ch). The ski report for Tignes and Val d'Isere say 110-160 cm of fresh snow in two days (Monday-Tuesday).

2. Description of the event

The plots below show short forecasts of MSLP and 6-hour precipitation spanning from 7 January 00z to 10 January 10z. The precipitation over the south-western Alps developed in the south-easterly low caused by a large-scale cyclone located in the western Mediterranean.


3. Predictability

  

3.1 Data assimilation

 

3.2 HRES

The plots below show 24-hour precipitation accumulated between 8 January 06z and 9 January 06z from observations (left) and short forecasts from HRES (middle) and one member from COSMO-LEPS (right).



3.3 ENS

The plots below shows EFI and SOT for 3-day precipitation valid from Sunday (7 January) to Tuesday (9 January) from different initial dates.

The figure below show the evolution of the2-day (8-9 January) forecast for precipitation in Val d'Isere in France for HRES (red dot) and distributions from ECMWF ENS (blue) and COSMO-LEPS (green). The model climatology distribution is in red. The ENS clearly gained signal on 30 January and again on 3 January.


3.4 Monthly forecasts

The plots below show weekly anomalies for precipitation for 8 to 14 January. Please note that the alignment with calendar weeks are not optimal for this case.

The plots below show weekly anomalies for MSLP for 8 to 14 January.


3.5 Comparison with other centres

See comparison with COSMO-LEPS above.

4. Experience from general performance/other cases


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

  • Difficult to evaluate the short-range forecasts due to lack of observation so far
  • Good performance in medium-range 9-10 days before
  • No signal on monthly scale (longer than 10 days)

6. Additional material