Status:
...
Finalized Material from: Fernando, Linus, Mohamed, Ivan
Discussed in the following Daily reports:
...
http://intra.ecmwf.int/daily/d/dreport/2014/06/10/sc/
http://intra.ecmwf.int/daily/d/dreport/2014/06/11/sc/
Picture
1. Impact
Excerpt |
---|
On 9 June severe convection affected western Europe. In Germany 6 people were killed, mainly by falling trees |
...
and wind gusts up to 42 m/s |
...
was reported from |
...
Düsseldorf airport. The severe convection followed a period of extremely hot weather. |
The aftermath (in Dusseldorf) - like a hurricane had hit:
Media reports:
http://www.bbc.co.uk/news/world-europe-27776189
2. Description of the event
The strong convection appeared on the north-western edge of a very warm air mass located over central Europe. The satellite imagery below (from late afternoon 9 June) shows convective activity occurring in France, NW Germany and Holland (also NE eastern England). The lightning reports indicate two major areas of strong activity; in France and Germany/Holland.
Satellite sequence from EUMETSAT:
https://www.youtube.com/watch?v=bf0uLDqdoc4
3. Predictability
...
3.1 Data assimilation
3.2 HRES
Precipitation
Gallery | ||||||
---|---|---|---|---|---|---|
|
The figure below shows animation
Wind gusts
Gallery includeLabel wg_ec sort comment title 24-hour maximum wind gusts
CAPE and CIN
Animation of RGB product from EUMETSAT, the max CAPE for +(6-24)h period (top-right) and maximum CAPE blended with CIN (CAPE is removed where CIN >= 200, bottom-left) based on 09@00 run for the HRES. One can see a good agreement between the high values of CAPE and the regions where the convection took place in particular over France and Benelux region. The convection triggered ahead of a cold front crossing the Iberia Peninsula seems to be in contradiction with the relative low values of CAPE forecast. It is possible that a dynamical lifting process helped to trigger the convection in the region.
Satellite sequence from EUMETSAT:
https://www.youtube.com/watch?v=bf0uLDqdoc4
3.
...
Predictability
3.1 Data assimilation
The plot below shows the time series of surface pressure for a set of observations inside a domain in west Europe. The figure includes time series for the first guess and analysis departures, bias and rms. From the 9th of June one can seen the impact of the strong convective activity on the model and the effort of the assimilation to bring closer the first guess to the observations.
3.2 HRES
The figures below show the 24-hour precipitation (06-06UTC). The first plot shows the observations and the following HRES forecasts valid for the same period. In western Germany and Belgium several stations reported more than 30 mm. In the forecasts the band of convection was shifted somewhat to the west. For central France it seems like the convection was overestimated.
Gallery includeLabel precip_ec sort comment title 24-hour precipiation
The figures below show the 24-hour maximum wind gusts (00-00UTC). The first plot is the observed values and the next is the HRES forecast for 9 june 00UTC. In general, no strong wind gusts are present in connection to the convection.
Gallery includeLabel wg_ec sort comment title 24-hour maximum wind gusts
The plot below shows the evolution of the individual convective cells in the forecast (run 09@00) and compared with the satellite images. The top row is the satellite images every 3hrs between 15 UTC (9th) and 00 UTC (10th) and the lower panels are the simulated satellite images from HRES. The stars identify the individual cells. The red star in the northern Germany seen in satellite picture was not predicted by the model while the blue star was forecast by the HRES. The trajectory forecast of this cell is very close to the observed one but slightly displaced to west (consistent with the precipitation pattern discussed above). The model succeeded to predict the size and the life cycle of the blue star. The cells developing in France late in the evening are captured also by the high resolution model (amber star).
The plot below compares the 24h acc. convective precipitation forecast initialised on 09@00 and the reported precipitation. The lines are an attempt to identify the trajectory of the convective cell (model and observed) which had a major weather impact in Germany. From the plot is clear how difficult it is to represent closely each individual cell in terms of its strength and location.
3.3 ENS
The plot above shows the CDF for Duesseldorf valid for
...
70% ENS percentile total PPN in 6h (top-left). Vis image at 1500 UTC and 3hrly PPN from HRES based on the latest run (bottom).
CDF for Duesseldorf valild 9 June. The probabilities for extreme temperatures was well captured 5 days in advance. However, not signal for heavy precipitation and strong wind gusts where present.
Gallery includeLabel efi_cape sort comment title EFI CAPE valid for 9 June
The figures above show the EFI for CAPE (shades) and precipitation (green symbols) valid for 9 June. The hourglass symbol marks the location of Duesseldorf. Already in the forecast from 4 June had high EFI values for the region for CAPE. The extreme rainfall in the forecast is located on the western side of the high CAPE values, probably connected to the CIN values (see figure above).
3.4 Monthly forecasts
3.5 Comparison with other centres
The plots bleow below shows the value of the maximum member for 18-hour precipiation precipitation (06-00UTC) for EC ENS (32km), , UKMO, NCEP, COSMO-LEPS and COSMO-DE-EPS. All forecasts are initialised 9 June 00UTC. Note that the number of members in each ensemble affects this diagnostics (50 for EC, 23 for UKMO, 20 for NCEP, 16 for COSMO-LEPS and 20 for COSMO-DE-EPS). The two last are high-resolution limited area ensembles. Although the maximum values of precipitation is higher in these models over The Netherlands, they also missed the Germany part.
Gallery includeLabel lam_max sort comment title Ensemble max precip
4. Experience from general performance/other cases
...
- Case for ~19 June last year, see DR http://intra.ecmwf.int/daily/d/dreport/2013/06/20/sc/
5. Good and bad aspects of the forecasts for the event
- Extreme CAPE-values already 6 days before for the region
- The HRES forecast (and most of the ensemble members) had the convection somewhat too north, leading to a miss for the worst affected areas in Germany (similar to other centres in TIGGE).
- Too weak convective wind gusts in the model for this case.