Status:Finalised Material from: Tim, Linus
Discussed in the following Daily reports:
Picture
1. Impact
A blizzard hit the region Annapurna in Nepal around the 14 October, killing at least 39 hikers.
The snapshot from Goggle Map below shows the location.
2. Description of the event
The heavy precipitation was mainly due to the aftermaths of the tropical cyclone Hudhud, which made landfall on the Indian east-coast on the 12 October (see below for track forecast after the landfall). The cyclone formed on the 8 October in the Bay of Bengal.
The plots below shows 12-hour forecasts of MSLP and precipitation for the days before the blizzard.
MSLP and precipitation from 12-hour forecasts
The animation below shows the change in the snow cover in the forecast from 11 October 00z.
For some reason the precipitation observations are missing from Nepal for the days of interest.
3. Predictability
3.1 Data assimilation
3.2 HRES
The plots below shows the 24-hour precipitation for the 14 October. The hourglass symbol marks the Annapurna district.
24-hour precipitation (14 Oct)
Show cover on the 15 Oct 00z
The plots above shows the snow cover in HRES at 15 October 00z from different initial times. Please note that points with very high values are glaciers. As reference, the plot below shows the snow cover on the 12 October.
3.3 ENS
The plots below show the EFI and SOT for snowfall valid on the 14 October. Already 7 days before the event the SOT was above 5 for the affected area.
EFI and SOT for snowfall
The EFI and SOT for snowfall are shown below for extended ranges beyond day 7 (non-operational ranges) and they still capture the signal of extreme snowfall. Such a strong signal for the time ranges displayed is very rare.
EFI and SOT for snowfall - extended ranges
The next plot shows the CDF for snowfall for the point 28.8N, 84E, which is in the worst affected area.All forecasts from 7 days before event and later, had 50% risk for exceeding the maximum of the model climate!
3.4 Monthly forecasts
Weekly precipitation anomalies 13-20 October
The plots above show weekly means of precipitation anomalies. Already the forecast from 6 October had a strong anomaly for northern Nepal.
3.5 Comparison with other centres
CDF for precipitation 28.8N, 84E
The plots above show CDF of precipitation for 28.8N, 84E for different centres in the TIGGE data base. For all lead times, the ECMWF ensemble is much more extreme than the other centres. This could be due to model and/or archived resolution or different meteorological conditions.
4. Experience from general performance/other cases
5. Good and bad aspects of the forecasts for the event
- A very good prediction of the extreme event. Probably due the the predictability of the tropical cyclone, which formed 6 days before the snowfall event.
6. Additional material
- Newsletter article in Winter 2014/2015