Mesoscale Discussions are here!

Last 22-23 February the ESTOFEX team congregated in Wiener Neustad in Austria. We decided to stop with Forecast Updates in favour of a new product, the Mesoscale Discussion.
The first public Mesoscale Discussion was issued for 1300-1700Z today by our forecaster Georg Pistotnik:

The severe weather reports from ESSL's European Severe Weather Database are also plotted as they came in.

A Mesoscale Discussion (MD) is intended to be a short text containing a diagnosis of ingredients for severe convective storms during the current day, with a validity of 2-6 hours. It may describe one or more areas of interest indicated by green contours. Like the daily Forecast it can also contain information about the expected storm mode and associated severe weather phenomena, but is not translated into threat levels and thunderstorm probabilities. A Mesoscale Discussion can be issued whenever a forecaster wishes to describe the current situation in an area in detail.

Note that Forecast Updates will no longer be issued. The Update was a product which changed the status of our threat levels. However, forecaster availability and the time required to decide collectively on a Forecast Update made it troublesome to issue it on time when the situation changed.

Soon we will also post our daily forecast map to our Facebook page.

ESTOFEX group photo at ECSS 2009

We profited from a rare sunny spell Tuesday afternoon at ECSS 2009 in Landshut, Germany and had our group photos taken! Click the images for a larger version (1 MB each):

Top row, left to right:
1. Paul Knightley, 2. Chuck Doswell, 3. Oliver Schlenczek, 4. Christoph Gatzen, 5. Tomas Pucik, 6. Marko Korosec, 7. Pieter Groenemeijer
Bottom row: 1. Johannes Dahl, 2. Helge Tuschy, 3. Harold Brooks, 4. Oscar van der Velde

All the forecasters:

Left to right:
Christoph Gatzen, Tomas Pucik, Marko Korosec, Oscar van der Velde, Johannes Dahl, Pieter Groenemeijer; kneeling: Oliver Schlenczek and Helge Tuschy

An impression of the conference venue:

ESTOFEX is attending ECSS 2009

This week (12-16 October) most of ESTOFEX is participating at the European Conference on Severe Storms (ECSS) in Landshut, Germany. During this time, forecasts may not appear. Thanks for your patience!

ESTOFEX starts probabilistic thunder forecasting

Starting September 8th, the ESTOFEX forecast maps indicate the probability of thunderstorms in more detail. By using two lines, areas with three different levels of probability are indicated.

Low probability: 0 - 15%
Medium probability: 15 - 50%
High probability: 50 - 100%

The probabilities refer to the chance of thunder within a distance of 40 km of a point. The areas are bounded by lines, that represent the 15% and 50% contours repectively. The 15% contour will be a thin yellow line and the 50% contour a thick yellow line. Verification results will be published as they become available. We are grateful to EUCLID for the gridded data they kindly provide us for this purpose.

First major severe thunderstorm outbreak of the 2009 summer season.

A severe thunderstorm outbreak occurred over parts of western and central Europe on Monday, May 25th, and Tuesday, May 26th. The synoptic maps on both days featured favorable conditions for severe thunderstorms over an extensive area. An upper-level trough moved in from the Bay of Biscay on the 25th, lifted northwards on the subsequent day and was placed over the North Sea on the morning hours of the 27th. An warm and unusal humid airmass covered NE-France, Belgium, Luxembourg and the Netherlands on the 25th with surface dewpoints in the twenties over NE-France. Steep lapse rates, generated by an elevated mixed layer and a moist boundary layer below, were responsible for extreme amounts of convective available potential energy (CAPE) over these areas with modelled values in excess of 3000 J/kg. Such values are extremely rare in Europe, and occur only once in a few years at a given point.

Additionally, moderate to high amounts of wind shear were in place, prompting ESTOFEX to issue a broad level-2 in its convective forecast. Emphasis of the convective forecast were the possibilities of giant hail (given the extreme CAPE values) and a serious wind gust threat.

After a long-lived cluster of thunderstorms evolved over the Bay of Biscay during the evening hours of the 24th and entered the western English Channel at the 25th around midnight, little convective storms developed. Then, new storms intiated over northern and western parts of France and across Belgium in the early afternoon. A small cluster, including a massive supercell crossed Belgium moving into Germany. Further initiation took place over central France at roughly 5 p.m. CET and this activity organized rapidly into a northeastward moving MCS with numerous, discrete supercells ahead of it. Hail around 10 cm in diameter was reported with these supercells, that naturally caused widespread hail damage. Overnight, a strong squall-line developed that produced gusts in excess of 35 m/s across Northern France, Belgium and the Netherlands. Spectacular photography of those storms can be found in for example, a report by our French colleagues of Keraunos. A radar loop of the beleux countries can be found here.

Another severe thunderstorm outbreak occurred over parts of Switzerland and S/SE-Germany as a long tracked bow echo raced northeastwards on Tuesday, May 26th. For that day, ESTOFEX again issued a level-2. Thunderstorms developed over Switzerland between 11 and 12 UTC and a continuous line evolved, running from the Black Forest area all the way down to Ticino. The most organized part of the MCS was situated over N-Switzerland with a rapid motion towards the northeast. During the passage of the Lake of Constance, radar scans revealed further organization into a well structured bow echo with a sharp reflectivity gradient along the leading edge and a pronounced rear inflow jet. During the passage of Landsberg, doppler velocity radar data measured 35-40m/s winds at the lowest 500-2000m, which overlaps nicely with a 39m/s wind gusts from Landsberg. Intense rain amounts with flash flooding and 3-4cm hail were observed in this line, which produced a swath from NE-Sitzerland to SE Bavaria with wind gusts in excess of 25m/s. See www.sturmwetter.de. The bow echo finally decayed over SE Bavaria during the evening hours with new initiation next to Salzburg.

Plans for 2009

ESTOFEX has had a meeting in Prague on March 27-28th 2009. There, ESTOFEX members have discussed several subjects, including a number of ideas that already existed for a few years. The most important issues were:

1) Excessive rainfall forecasting
2) Probabilistic forecasting

Excessive rainfall forecasting

Of all threats associated with thunderstorms, flash flooding causes the most casualties and can produce a lot of damage. We have decided to start including convective rainfall threat in our "threat level" system and will be focusing on the meteorological aspects only, since we cannot practically take into account local hydrological conditions. We will only consider convective events to be part of our forecast job. Like severe wind gusts, which can also occur in non-convective situations, we will use the presence of lightning near the place and time of an event to determine whether the event is convective.

The criterion to identify whether convective rainfall was severe has been set at 60mm in 3 hours (and, as said, lightning must be detected). The difficulty, of course, is that there are many different ways precipitation is reported, such as "flash flooding occurred", "90mm collected over 24 hours", or "20mm in just 10 minutes". Our criterion is chosen so that we can exclude the short duration events with small sums and focus on the significant events. We will not use an "extremely severe" criterion for precipitation as we do with the other types of severe convective weather.

With the addition of excessive rain to threat levels, they are likely going to be issued more often. Because there can be several reasons for issuing a threat level, we will, starting May 1st. 2009, add phrases above the forecast text indicating the primary severe weather threat for each threat level area.

Probabilistic forecasting

Thunderstorm areas

Our current forecasts for thunderstorms consists of one line (yellow), which, in the ideal case, separate the areas where thunderstorms will occur and where they will not. In practice, of course, thunderstorms do not always occur everywhere where the regional scale conditions are thought to be favourable. So, if you find yourself in a thunderstorm area, the chance that you will experience a thunderstorm is much lower than 100%. Similarly, outside the thunder areas, the chance is higher than 0%. Forecast verification has shown that on average, in central Europe in the summer season, the true chance to get lightning within a radius of 40 km of your location when you are included inside the area is about 50%. In northwestern Europe this chance is less than 20%.

We have decided to start drawing two lines for the chance of thunderstorms, in the near future. This will allow us to better characterize the true chance of storms. a thin yellow line will indicate where we forecast a low spatial coverage of thunder, or when we have lower confidence in the development of storms, while a thicker yellow line will indicate where confidence of higher storm coverage is large. This change will probably be implemented later in 2009.

Severe weather threat areas

The severe weather forecasts will, starting May 1st, also have a well-defined probabilistic interpretation. The lines will namely represent a chance percentage of severe weather occurring within 40 km of a point on the map. Using the data from the ESSL European Severe Weather Database, Pieter Groenemeijer has conducted a verification of severe weather for our levels. As a part of this procedure, the number of times severe weather occurred inside the different level areas was determined. With this information the threat level lines will now be labelled with a percentages indicating the probability of severe, or extremely severe weather (details in the FAQ soon).

A problem ESTOFEX faces is, however, that the reporting of severe weather to ESWD is still very low in certain areas of Europe, so that we basically had to rely on verification statistics from the part of the forecast area of which most data are available. Put briefly, we need your reports! So, to support ESTOFEX, please, submit severe reports to the ESWD. Note that an ESWD report has only five required fields: DATE, HOUR, LOCATION, DESCRIPTION, and SOURCE.

Central-European severe weather outbreak August 15th

Many countries were affected in a belt extending from northern Italy to the Baltic Countries. Sadly enough, 8 people lost their lives due to the severe weather, most of them due to falling objects in strong wind gusts, 3 in Poland, 2 in Slovakia, 2 in Italy and 1 victim in Austria. ESTOFEX noticed the seriousness of situation a few days beforehand and an extended forecast was issued on August 13th. The next day it became obvious that the highest threat level, 3, would be necessary for some regions and a convective forecast was posted.

The first storms started quite early, before noon, and in very strong wind fields. They quickly organised into supercells. One of the supercells hit the city of Katowice, accompanied by damaging hailstorm. A video of the falling hail can be found here. But Poland was not the only region, where explosive thunderstorm formation occurred. Northern Italy and Slovenia followed very soon. In very a unstable environment, the storms produced swaths of large hail, some of the hailstones had a diameter of more than 5 cm. Later in the day, the situation became even more serious as a line of supercells developed over southern Poland and moved to the NE. A forecast update was issued approximately soon afterwards. These storms produced at least 8 tornadoes. Here are a few photo's:


Source: AP


Source: Agencja Gazeta

One of them has been rated F3, i.e. a strong tornado, and caused considerable damage to houses and vegetation. Tornadoes were extensively documented by local storm chasers and some videos can be found here. A very dramatic video comes from the inside of a bus, that was overturned by severe tornadic winds. Meanwhile, a line of HP supercells organized into a powerful bow echo over Slovenia and headed to the north. Extensive wind and hail damage occurred with this system across Slovenia, W. Hungary, E.Austria and SW. Slovakia. In Austria and Slovakia, hailstones of 5 cm diameter were found after the storm and roofs of some houses were completely destroyed. What is more, mountainous regions of Middle Europe experienced flash flooding as storms repeatedly moved over the same places.

Violent tornado in northern France

A violent tornado has struck the communities of Hautmont and Maubeuge in northern France on August 3rd at around 20:30 UTC. Severe weather experts of the Belgian Meteorological Institute Karim Hamid and Karel Holvoet have conducted a damage survey and rate the damage F4 on the Fujita scale, with wind speeds likely exceeding 300 km/h or 90 m/s. Their report (in Dutch) can be found here. Pierre Mahieu and Emmanuel Wesolek of Keraunos have published this case study in French. The severity of the damage is also clear from images that have appeared in the press, for example the following...

photo: AFP / P. Frutie

...and from these photos. The tornado sadly caused 3 fatalities and left 18 people injured. Moreover, more than 700 homes were damaged, among which a few dozens have become inhabitable. A television crew of LC1 has flown over the affected areas resulting in the following TV report.
ESTOFEX had initially issued a level 1 for this area. It appeared that strong low-level wind shear and high storm-relative helicity would enable rotating updrafts to form with a chance of tornadoes. However, low amounts of instability forecast by the numerical model precluded the issuance of a higher risk category until the mid-evening. At 20:12, shortly before the deadly tornado struck, a forecast update was issued because it appeared that strong convective storms were ongoing in an environment much more unstable than expected earlier in the day. This is one of the strongest, if not the strongest tornado that occurred since the start of the European Storm Forecast Experiment.

Severe thunderstorm outbreak south of the Alps

Severe thunderstorms have caused extensive damage on Sunday 6th of July and Monday 7th of July. The episode started with the development of intense supercells across the southern flanks of the Alps in northern Italy Sunday afternoon. The updated ESTOFEX forecast for this day can be found here: link. Large hailstones up to 6 cm in diameter and tornadoes have been reported to the ESWD database of the European Severe Storms Laboratory, the organization that keeps track of severe weather in Europe. The reports of both days can be retrieved here. Among the people affected were tourists who were camping close to Lake Garda. A video at Youtube shows the large hailstones: link. Later in the day, the cells progressed southeastward through the level 2 area while continuing to produce large hail. A timelapse video of the rotating storm by Cristiano Corte can be found here: link. This storm continued to produce several tornadoes near Padova. A video of an impressive tornado over water can be found here: link. Several images and video's of the storms and the damage they have caused are presented on the website of Andrea Colombo link.

On Monday, the severe weather continued. The latest forecast update on Monday looked like this: link. During the afternoon strong storms moved eastward over the southern Italian Alps again, before even more intense supercell storms developed over the regions of Veneto and Friuli Venezia Giulia. One particularly intense supercell storm crossed the northern Adriatic Sea and produced serious tornado damage north of the city of Rijeka in Croatia. In the towns of Lisac and Pozega, houses were damaged and the nearby forests partly flattened, see: link. Damage also occurred in the areas of Klenovnik and Lipovnik: link. Slovenia was hard hit as well as damaging winds swept through the country after the supercells had clustered into a linear system. At the same time, more isolated convective storms, also including supercells had developed over parts of Hungary, where widespread large hail occurred.

New: Extended Forecasts

Starting today, ESTOFEX will issue Extended Forecasts, offering a discussion and categorical map of a severe weather situation an extra day in advance, allowing visitors to become more aware of the developing situation early-on.
The Extended Forecast is a product that will be offered depending on availability of forecasters and the apparent urgency of the weather pattern.

European Conference on Severe Storms

Next week the 4th ECSS takes place in Trieste, Italy (10-14 September)
ESTOFEX contributes to this conference with two oral presentations and a poster. Three forecasters additionally present a part of their own thesis work. Detailed results on the verification of our forecasts will be presented, as well as analyses of a few interesting recent large hail and tornado cases.
Very interesting talks and posters are scheduled, also from severe storms researchers overseas.
You can find the program and abstracts at the ECSS website

As all of our forecasters will attend the meeting, it is quite likely that no forecasts will be present for most (or all) of these days, starting Sunday 9th. We apology for the inconvenience and hope to be back to schedule by Sunday 16th of September!

Estofex starts objective forecast verification

In the process of "learning how to forecast thunderstorms" ESTOFEX wants to monitor the quality of its forecasts. This is of course done by comparing the weather that was forecast with that what occurred in reality. Since the spring of 2006, lightning location data and (mostly unverified) severe weather reports from the European Severe Weather Database have been plotted on the maps of old forecasts. This gave the possibility to subjectively see if the threat level and thunderstorm areas had more or less been placed correctly.

Now we take a next step, by analyzing all forecasts issued since April 30th 2006. The results can be used to become aware of systematic forecast errors. This data is made available to our visitors as well and can be found in a new section of the ESTOFEX website.

ESTOFEX meeting of November 2006

At least once a year ESTOFEX comes together to talk and make decisions about forecasting procedures, verification, plans for our future, and outreach. The last meeting was from 3-6 November 2006 in Linkenheim, Germany. Four of the five forecasters were attending and one active member.

A lot has happened over the year. In January 2006 we had a meeting in Toulouse where we decided to use newly defined risk levels and a new colour scheme to go with it. Pieter Groenemeijer has been very active programming a forecast map drawing tool, storing the coordinates, a major step that helped both uniformity of the look of our maps, as well as offering possibilities to do verification. It went into use on April 29th. Pieter has started a first verification based on lightning data from the British arrival-time difference system, the well-known SFLOC reports. This covers a large part of Europe, and is reasonably reliable, enough to use it for our purpose. The results will soon be available under a special section of our site.

We decided to make a change to the forecasting scheme (see FAQ). We simplified it by not making a distinction between tornadoes and other severe weather anymore, and we adjusted the unit area and number of events expected for the threat levels. We will likely make another adjustment to facilitate verification, by using a unit radius around a point within which at least one event should happen to verify the point as true. This is not likely to affect the meaning of our threat levels much, but may influence the style of drawing thunder and threat areas.

Other things we discussed were forecast text uniformity (necessary elements, correct grammar, no abbreviations), and the possibility of adding a severe convective weather FAQ, case studies, longer and shorter term forecasts (rejected: lack of forecasters!), outreach, offering workshops, and organizational aspects.
We also decided on enabling this News blog, so that we have more means of informing you of our changes and to have the possibility of issuing statements and reflection on recent major events.

First test of our ESTOFEX news blog

On this page we will issue news about ESTOFEX and we will have the opportunity for additional discussion on recent major severe weather events.