Log In | Users | Register
Edit | Attach | New | Raw | Diff | Print | | Tools

myCOSMO-NExT Case: 2005 07 18

Short description of case; if available reference to published literature and other (possibly attached) documents.

On the afternoon of July 18th 2005, a particularly intense supercell thunderstorm struck the Lake Geneva region. The storm initiated just southwest of Lyon, France and tracked over 300 km towards the northeast before losing its supercell characteristics in the Swiss Alpine foothills around the town of Interlaken. During its 3-hour lifespan, the storm’s forward translation averaged 60-80 km h-1. At the height of its severity, this supercell was responsible for hail the size of golf balls, a microburst with measured wind gusts up to 160 km h-1 and two confirmed tornadoes. Miraculously, nobody was killed nor seriously injured. However, the material losses were considerable including ravaged vineyards, damage to buildings and vehicles and sections of forests completely destroyed. Post event analysis was undertaken utilizing radar imagery/algorithms, satellite images, lightning data as well as surface data, eyewitness reports and a damage survey. This case provides a unique look at a supercell evolving within an Alpine environment and helps confirm prior research concerning certain storm features and signatures that have been observed in North American Great Plains supercells and elsewhere. Concerning tornadogenesis, this case helps confirm via observations and simulations the hypotheses of several previous papers that low-level wind flow modified through channeling by mountains can periodically provide a locally favorable wind shear environment for tornadogenesis. For this particular case, inflow winds channeled around mountain features appear to have been instrumental in the formation of the second tornado, since important topographical obstacles prevented any significant low-level gradient wind shear from operating on the eastern end of Lake Geneva where this tornado occurred.

ppt presentation @ ECSS 2007 conference : J:\gve\pel\ECSS2007\Peyraud_ECSS2007.ppt

While it is unrealistic to expect the COSMO-1 model to be able to simulate all of the micro-scale details of this supercell event (all storm-scale flows and interactions), it will nevertheless be interesting to note whether the model is able to better simulate certain mesoscale flows around and emanating from the supercellular convection than the coarser COSMO models. Particularly interesting will be whether COSMO-1 is able to detect any type of low-level flow regime in the Rhône Valley ahead of the supercell's arrival in Le Bouveret that may have helped increase the low-level storm-relative helicity and hence the tornadic potential of this specific supercell.

Version 1

COSMO-1 forecast: 18.07.2005, 06 UTC run

* valid @ 12 UTC (H+6hr): pre-convective environment

  • Instability fields : @ 12 UTC,

  • Low-level wind fields : @ 12 UTC,

  • LCL heights : @ 12 UTC,

MLCAPE Cosmo1 12utc 06utcrun.jpg

MLCAPE Cosmo7 12utc 00utcrun.jpg

MUCAPE Cosmo1 12utc 06utcrun.jpg

MUCAPE Cosmo7 12utc 00utcrun.jpg

10mwind Cosmo1 14utc 06utcrun.jpg

10mwind Cosmo2 14utc 00utcrun.jpg

Version 1

COSMO-1 forecast: 18.07.2005, 12 UTC run

QPE 6hr Cosmo7 18utc.jpg

QPE 6hr Cosmo1 18utc 12utcrun.jpg

10mwind Cosmo1 14utc 12utcrun.jpg

Summary of this simulated convective event

Overall, the above COSMO-1 simulations (6 UTC and 12 UTC runs) seem to have done an equivalent job in simulating the synoptic and mesoscale environments than the COSMO-2 re-analysis that had been undertaken of this event. However, contrary to the COSMO-2 re-analysis of the event which had been conducted with radar data assimilation (LHN), these COSMO-1 simulations did not benefit from this additional initial condition data. As a result, while the COSMO-2 re-analysis with LHN was able to reasonably simulate part of the supercell's storm-scale wind field (i.e. gustfront structure), the COSMO-1 COSMONExT? simulations were not able to do so since the lack of radar data assimilation did not allow the correct placement of the convective cells. In the COSMO-1 simulations, the convective activity was placed a bit further south along with the associated storm-scale flow. In conclusion then, this supercell event over the Lake Geneva region and its tornadic potential was very dependent on storm-scale flow interactions between the storm and the terrain surrounding it. As a result, the only way to adequately simulate such a storm (and perhaps catch its tornadic phase) would be to use a storm-scale model with a much finer resolution (below 100 m) and include radar data assimilation (and perhaps add satellite and lightning data) in order to simulate the storm in the correct location. It is not sure whether such a model is really necessary or wished for at this point, since other more basic and important simulating challenges remain to be solved within the very heterogeneous environment of the Swiss Alpine region. (For a more detailed look at the COSMO-2 re-analysis of this event with LHN, please consult the article in Weather and Forecasting, December 2013 issue).

Version 2

COSMO-1 forecast: 18.07.2005, 06 UTC run

* Version2_Cosmo1_MUCAPE_h6hr_12utc.png:
Version2_Cosmo1_MUCAPE_h6hr_12utc.png

* Version2_Cosmo1_ppn1hr_h8hr_14utc.png:
Version2_Cosmo1_ppn1hr_h8hr_14utc.png

  • Version2_Cosmo1_ppn1hr_h9hr_15utc.png:
    Version2_Cosmo1_ppn1hr_h9hr_15utc.png

  • Version2_Cosmo1_10mwind_h8hr_14utc.png:
    Version2_Cosmo1_10mwind_h8hr_14utc.png

  • Version2_Cosmo1_10mwind_h9hr_15utc.png:
    Version2_Cosmo1_10mwind_h9hr_15utc.png

* Summary of this simulation : The COSMO-1 modeled convective precip signal from this version 2 of the 6 UTC run is not much different than the corresponding version 1 COSMO-1 6 UTC run. Nevertheless, here are slight differences. The version 2 run simulated a slightly less unstable airmass over the Swiss Plateau in the pre-convective environment around 12 UTC (200-400 J/Kg less CAPE). As a result, the modeled convection is a little weaker in version 2 compared to version 1 but the convective cell trajectories are pretty similar to each other (both slightly to the south of the supercell's actual trajectory). Concerning the simulated low-level 10 m wind fields during the 14 and 15 UTC timewindow when the modeled convection was in proximity to Lake Geneva, while both versions adequately resolve the drier post-frontal northwest flow along the Jura, the prefrontal windflow ahead and in the immediate proximity to the modeled convection is very chaotic and large differences in wind speed and direction are apparent between the 2 versions. This is of course a result of the differences in modeled convective t-storm outflow regimes between the 2 versions that don't place the convective cells in the exact same locations. As a result, it is impossible to draw any conclusions based on these simulations regarding the exact mechanisms that ultimately favored tornadogenesis on the east end of the lake. That being said, it is unrealistic to expect a 1km resolution model without radar data assimilation to adequately resolve a thunderstorm's storm-scale flows and local interactions (one would need a 100 m horizontal resolution model and very good assimilation of obs data (synop,metar,lightning) and passive instrument data (sat+radar)).

Version 2

COSMO-1 forecast: 18.07.2005, 12 UTC run

* Version2_12utcrun_Cosmo1_MUCAPE_h0hr_12utc.png:
Version2_12utcrun_Cosmo1_MUCAPE_h0hr_12utc.png

  • Version2_12utcrun_Cosmo1_ppn1hr_h1hr_13utc.png:
    Version2_12utcrun_Cosmo1_ppn1hr_h1hr_13utc.png

  • Version2_12utcrun_Cosmo1_ppn1hr_h2hr_14utc.png:
    Version2_12utcrun_Cosmo1_ppn1hr_h2hr_14utc.png

  • Version2_12utcrun_Cosmo1_ppn1hr_h3hr_15utc.png:
    Version2_12utcrun_Cosmo1_ppn1hr_h3hr_15utc.png

  • Version2_12utcrun_Cosmo1_10mwind_h1hr_13utc.png:
    Version2_12utcrun_Cosmo1_10mwind_h1hr_13utc.png

  • Version2_12utcrun_Cosmo1_10mwind_h2hr_14utc.png:
    Version2_12utcrun_Cosmo1_10mwind_h2hr_14utc.png

  • Version2_12utcrun_Cosmo1_10mwind_h3hr_15utc.png:
    Version2_12utcrun_Cosmo1_10mwind_h3hr_15utc.png

* Summary of this simulation : The COSMO-1 modeled convective precip signal from this version 2 of the 12 UTC run is not much different than the corresponding version 1 COSMO-1 12 UTC run. Paradoxically, the modeled convective ppn signal is slightly weaker in version 2 compared to version 1 despite the fact that the modeled CAPE is higher around Lake Geneva in version 2 compared to version 1. However, the modeled convective cell trajectory is almost identical in the 2 versions, both slightly to the south compared to the actual supercell trajectory. Concerning the modeled low-level wind field, once again no conclusions can be drawn concerning tornadogenesis since the modeled convection is in the wrong location and hence topographical windflow interactions are unrealistic in the model compared to reality.

For a condensed summary of the assessment for this case as well as all other myCOSMO-NExT cases refer to the myCOSMO-NExT Overview page.

Note: This page is both world-readable and world-writeable.

Attach
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:12
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:14
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:13
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:25
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:25
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:26
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:26
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 16:00
jpg
version 1 uploaded by LionelPeyraud on 06 Jan 2014 - 15:59
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:34
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:35
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:35
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:33
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:34
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:34
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:34
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:09
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:10
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:05
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:08
png
version 1 uploaded by LionelPeyraud on 23 Jul 2015 - 14:08
spacer
This site is managed by the Center for Climate Systems Modeling (C2SM).
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors. Ideas, requests, problems? Send feedback!