In the late afternoon our expert guest Darrel Kingfield from NOAA/NSSL presented an application-based presentation on elementary radar analysis and integrating polarimetric radar variables (differential reflectivity, correlation coefficient, specific differential phase) into the storm interrogation process. He also shown few examples of tornadic supercells on the radar data.

In the evening we had few thunderstorms nearby Wiener-Neustadt and few of us went chasing to take some pictures. Unfortunately we did not catch anything, but it was great fun!

Thursday’s thunderstorm activity was forecast to be less intense than Wednesday’s and would cover the south-eastern and south-central part of Europe. Most of the severe weather events took place in south-eastern Austria where large hail and heavy precipitation were reported. After the verification of the mid-european day1 and day2 forecasts for the 12th of June (shown below), participants had an interesting discussion on the nowcastMIX and VIL products. The group found that high values of VIL can be used to detect large hail. However, since this parameter is calculated from a vertical column, updraft tilting (in sheared environment) and the speed of radar scanning affect the VIL values, and may limit its usability. In nowcastMIX product, participants suggested that instead of predicting all threats in one tool, it would be useful to use seperate categories for hail, wind gusts and heavy precipitation. This would make the comparison with occurred severe weather reports easier.

VILtrack product

nowcastMIX product

Today our participants had hands full of work. In the environment of high CAPE and low shear numerous intense thunderstorms have formed in central-european region.
Even though situation was diffucult, with the use of available Testbed tools participants managed to make a quite good predictions. Let’s take a look at yesterday’s and today’s day1-forecast and nowcasts:

11.06.2014 12UTC Meteosat Sandwich, surface observations and lightning data

Today, after expert lecture held by Paul James on NowcastMIX product, participants began to monitor thunderstorm activity across Europe. Around 12UTC the most spectacular feature was MCS over NE Germany. Thanks to Szymon Walczakiewicz from Polish Stormchasing Society we were able to see shelf-cloud associated with this feature over Szczecin city:

In the late afternoon, isolated convective cells near German, Czech Republic and Polish border that grew in the environment of large CAPE (> 3000 J/kg) and low wind shear (< 10m/s) produced hail up to 5cm in diameter. Similiar as in situations from previous days, we found VIL track parameter as usefull tool in defining places where large hail is likely to occur. On the image below it can be seen that even though MCS from NE Germany was active, it did not show any potential to produce large hail.

11.06.2014 Meteosat E-view, VIL track and ESWD large hail reports

Due to the public holiday on Monday, the second Testbed’s session starts on Tuesday this week. New participants from Poland, Finland, Germany, Austria, UK and USA will try to deal with the difficulty of forecasting severe weather that is going to occur this week in Europe.

This week participants will be supported by Georg Pistotnik and Tomas Pucik from ESTOFEX association. Very interesting expert lecture “An overview of NOAA/NSSL decision assistance tools, current & future capabilities” was presented by Darrel Kingfield from NOAA/NSSL. Darrel approached available NSSL nowcasting tools that support forecasters in making decision and also underlined the importance of adaping warnings visualisation to the needs of society.

A lot of severe thunderstorms occcurred on sunday and monday. During weather briefing our forecasters discussed supercells and bow echo features that occurred in western Europe. Conditions were very supportive for severe weather, very large hail of ~12cm was reported in France while Dusseldorf experienced wind gust of 40m/s. The situation from 1945UTC is presented below:

09.06.2014 1945UTC Meteosat IR BT image with surface observations and PCAPPI.

With the use of radar products, we’ve also analyzed the case of supercell that occurred around 12UTC in western Germany. The classic “hook-echo” shaped structure was detected on the radar in PCAPPI reflectivity. By using VIL product (vertically integrated liquid) it was also possiblie to conclude that the cell was right-moving and high values indicated possiblie hailstorm. As it turned out, numerous reports of large hail reaching up to 4cm were reported in ESWD. We also checked mid-level rotation product and mesocyclone detection algorithm. Although according to the algorithm rotation was not strong, it was detected.

09.06.2014 12UTC VIL track – vertically integrated liquid algorithm (left), PCAPPI reflectivity (center), mid-level rotation track with ESWD reports (right)

Currently ESSL Testbed participants work on day1 & day2 experimental forecasts and nowcasting in central Europe.

This week we are expecting a lot of severe thunderstorms, so participants are going to have a lot of interesting work!

The first week of ESSL Testbed 2014 comes to an end. Today our participants issued their last nowcasting, long and short term forecasts. The most interesting forecast for Sunday evening we provide below:

“Forecaster: ESSL TESTBED

A trough over Western Europe with embedded short wave troughs moving northwards is still nearly stationary located with its eastern border just offshore the Western coasts/ British Isles. Diurnal heating and warm air advection from South West as well as large instability and strong deep layer shear lead to thunderstorms. In the area large hail, heavy precipitation and strong wind gusts can be expected.

Severe weather is expected in Eastern Europe (Serbia, Romania, Moldavia, Ukraine, W Russia): Former air mass boundary between hot dry air to the East and moist air to the West is moving SE as a cold front. Due to high CAPE values and existing moisture thunderstorms with large hail and heavy precipitation are expected. In the SW part of the area significant low level shear could support local tornadic storms.

Subjective weighting of guidance by these products when making the forecast, on a scale of 0 to 4 (0 not used at all, 4 strongest weight):
GFS: 2
ECMWF: 3
ECMWF-EFI: 0
ALARO: 0
COSMO-EU: 0″

_____________________________

Indeed, as we take a look at GFS and ECMW predictions of CAPE, DLS and SRH, they seems to overlay quite well. However, since current forecast indicate that QG-lift is not favorable there is uncertainty whether thunderstorms will occur. Situation is interesting and worth to monitor in following hours.

In the morning our participants also issued forecast for Paris domain for Saturday, which also seems to be interesting:

“Forecaster: ESSL TESTBED

A level 1 was issued for central and northern France, Benelux, north-western Germany and southern part of British Isles for severe convection.

Central and northern France, Benelux, north-western Germany and southern part of British Isles is expected also for heavy rain and locally hail (especially in northern France and Benelux countries). Central and northern France is expected also for a gusty wind up to 90 km/h”
__________________

But let’s get back to our last day of Testbed’s week 1: Expert lecture about convective windstorms in Europe was presented by Christoph Gatzen from ESTOFEX. Christoph pointed that when forecasting severe wind events, we have to take into account much more factors than only dry air in mid-levels and 850hPa windspeed. In general we need deep moist persistent convection to produce severe wind gusts. During weather briefing there was also a birthday suprise for Georg! 🙂

In the end of the day all participants completed surveys concerning the ESSL Testbed and provided feedback to experimental products.We would like to thanks all participants for interesting forecasts and curious discussions about the weather. You provided a great international atmosphere during the whole week!

See you soon and have a safe trip!

An interesting lecture of the day concerning lightning detection was remotely presented by Ryan Said (USA) from VAISALA. At the Testbed participants use GLD360 lightning data to define thunderstorm position and verify their forecasts. Yes, what about verification of our participants day1&2forecasts? Let’s take a quick look at them:

It is worth to praise Paris domain team for predicting a strong thunderstorm (50mm rain in 1 hour and high lightning activity) that moved S of the Alps during the night. Their forecast was partially based on the COSMO-DE ensemble product of “total water” parameter (using fraction of members exceeding the value of 40kg/m2). This NWP prediction was very distinctive! Bravo COSMO!

It is also worth to check COSMO ensemble 15Z forecast of convective cells in central Poland from previous blog entry. As it can be seen on the 15Z nowcasting situation below, COSMO-DE ensamble forecast again succeded:

A COSMO-day! 😉

Yesterday lecture concerning the value of ensemble forecasting in COSMO model was conducted by Lars Wiegand from Deutscher Wetterdienst. Contrary to deterministic model outputs, ensemble products show various model scenarios (ensemble members) obtained from the different initial conditions (coming from 4 global models) and model physic pertubations. This kind of product finds its practical application in evaluating places where convection is likely to occur.

Participants at the Testbed in short-range forecasting are equipped with experimental 4-panel chart of COSMO-DE ensemble data concerning:
– maximum value of all ensemble members
– areas with particular values of any ensemble member
– fraction of members exceeding an established treshold
– deterministic forecast

On the image below today’s 15Z forecast of radar reflectivity indicate line of convective cells in north-central Poland. The greatest assurance concerns coastal areas while inland areas denote high uncertainty. Are they going to occur? We will see.

It is worth spending a little attention to introduce new experimental forecasting products which participants are able to test during ESSL Testbed 2014. Today we will focus on NWP roaming sounding tool.

How does it work? You can see it in our video below:

click -> Roaming Sounding Tool <- click (composite of METEOSAT Rapid Scan E-View product, OPERA radar reflectivity, GLD360 CG lightning data, overshooting tops algorithm, ECMWF CAPE and ECMWF roaming sounding)

In contrast to previous grid-oriented soundings, this allows you to make a continuous and quick check how the vertical temperature, moisture and wind profile varies spatially.

This tool has been just used to issue an experimental nowcast for a tornado risk in NW Turkey. The lockwise curved hodograph looks especially dangerous. Environmental conditions with SRH ~ 350-400m2/s2, DLS 20-25m/s, LLS 9-12m/s, LCL 600-800m AGL and CAPE ~ 1500 J/kg points out good favorable conditions for mesocyclonic tornadoes. Taking into account that few supercells just occurred in this region, it is worth to monitor situation!

A composite of the METEOSAT Rapid Scan RSS IR product with cloud top temperature, GLD360 CG lightning data, overshooting tops algorithm, ECMWF storm relative helicity and ECMWF roaming sounding)

Today’s verification session revealed again quite a good performance of our participants in day-1 forecasting. Lightning probability areas especially in Alpine and Carpathian region denoted accurately places where lightnings occurred yesterday.

After verification procedure, participants in 2 teams were obliged to conduct weather briefing on their own. Chrisoph Gatzen prepared guidelines to highlight the most important aspects which can be used in presenting day1 and day2 forecast.

Team 1 focused on the trough in Paris domain, while second group prepared analysis for Sofia domain. With the use of experimental forecasting tools group leaders managed to do it quite well. Their presentations were transmitted using Saba software, so that remote participants were also able to be involved.

Today more attention has been devoted to nowcasting products. Thomas Hengstebeck, our expert from DWD, presented remotely how mesocyclones can be detected with radar data. Since this year’s Testbed display offers experimental mesocyclone detection algorithms, this topic was important for our participants. In the afternoon, Pieter Groenemeijer also presented how these products could be used by showing a situation from 23rd May, when large hail up to 4.5cm was reported in Poland.

Nowcasting composite of METEOSAT Rapid Scan E-View product, EuRadCOM radar reflectivity of the lowest level (pseudo-CAPPI), GFS 700hPa wind barbs and geopotential contours and mesocyclone detection algorithm (circles: see operations plan for description).

In the late afternoon, group with Georg Pistotnik prepared day3-5 forecast, while group of Christoph Gatzen was doing nowcasting. Today’s analysis also included good old style map drawing based on measurement data!