Storms of the 22nd April 2008 above Bulgaria, Romania and Ukraine

• Martin Setvak
• Aurora Stan-Sion
• HansPeter Roesli
• Gergana Kozinarova
• Jarno Schipper

On the afternoon of 22nd April 2008, a series of significant and severe convective storms has developed in the region of Bulgaria, Romania and south-west Ukraine. Two of these storms have developed several significant cloud-top features, observed by weather satellites, commonly ranked as indicators of possible storm severity. These were: high 3.7/3.9 μm cloud top reflectivity, well-developed plume above one of these storms, and namely a cold-U shape feature atop one of these storms, and a cold-ring shape on top another storm. It is namely the successive occurrence of these two later features which makes this case extremely important for conceptual models of the two features, and of convective storms in general.

Meteosat 9 Enhanced IR10.8 - 22 April 2008 1345UTC

Meteosat 9 Enhanced IR10.8 - 22 April 2008 1645UTC

The first of the two major storms (labeled “A”) began to form above north Bulgaria at about 1215 UTC, forming a distinct cold-U shape shortly after, at 1300 UTC. From then on, the storm began to move rightwards, south of BG/RO borders. The storm lost its rightward propagation at about 1415-1430 UTC, crossing the borders to Romania. The cold-U shape disappeared between 1515-1530 UTC, and the storm has dissipated shortly after. The second storm (labeled “B”) formed at about 1600-1615 UTC approximately in the same area in Romania, where the storm A has weakened, just behind it. From 1615 UTC the storm formed at its top a well-defined cold-ring feature, which persisted atop of this storm till about 1800 UTC, when the storm weakened and dissipated above south-west Ukraine.

The importance of this case is in the successive occurrence of the two features (cold-U and cold-ring shapes) in such a close proximity. It is believed that the main difference between the two types of storms is linked to the environmental wind shear. If so, the occurrence of the cold-ring shaped storm B in the track of the first one, the cold-U shaped storm A, can be explained by several ways: (1) modification of the environment by the storm A; (2) formation of the cold-U shape not only as a result of the environment, but also some internal processes inside the storm A (at the same time forming the distinct plume above the storm) and/or by the “plume-masking mechanism”; (3) different cloud-top heights reached by the two storms, thus interacting with slightly different environmental winds. The possible mechanisms above will be further studied, together with other satellite and radar characteristics of these two storms.

In this case the relevant satellite imagery from the SEVIRI instrument as well as some imagery of NOAA polar satellite will be presented and described. The severe convection will be further pictured using RADAR. The case was associated to reports of thunderstorms, hail and tornados!

This footage of a tornado was taken by Tihomir Velikov (Hydrologist at NIMH-BAS) in Razgrad (Bulgaria) with his cellular phone (click here to start the movie)

KNMI Surface Chart - 22 April 2008 1200UTC

Meteosat 9 Airmass RGB - 22 April 2008 1800UTC

In the above images a frontal system can be recognised over Romania and Bulgaria. Within the cold front the Meteorologists of KNMI analysed a further wave development that with the corresponding PVA maxima in the upper layers may help to understand the weatheractivity of this case on a synoptic scale. The Airmass RGB shows a large MCS that is embedded to the frontal depression at 1800UTC

From Estofex the following forecast for April 22 was issued for Romania and Bulgaria:

SYNOPSIS

A dipolar pressure configuration is situated over Europe, with Scandinavia under a blocking high while a low pressure system affects mostly Italy and the Balkan, as well as more central parts of Europe. Unstable airmass can lead to isolated to scattered thunderstorms with the focus appearing to be Romania, northern Bulgaria and the nearly stationary cold front stretching from northern Austria into Slovakia. The vertical shear is weak, except for the eastern parts of the Balkan, where a strong upper jetstream lies… with the left exit region of a 60 m/s jetstreak (300 hPa) situated near NW Bulgaria and southern Romania (in GFS model 12Z) during the afternoon. Low level convergence however runs ahead of this zone, being over eastern Romania at 15Z.

DISCUSSION

…Romania and northern Bulgaria….

GFS model simulates weakly capped instability with 2000m cloud base height at 12Z. Several hundred J/kg MLCAPE should be available from a deep source layer. However, Monday 12Z soundings in the area are quite dry and barely have any CAPE, large scale lifting is needed to create more substantial CAPE.

The picture shown by various bulk shear magnitudes is supportive of supercellular convection, but the 0-3 km shear and storm-relative helicity are offset to the east of the most likely triggering zone indicated by low-level convergence. If anything, this environment may favour splitting cells.

The larger scale forcing imposed by the jetstreak could yield a MCS going over S to E Romania and Moldova, while N-Bulgaria may see more of a preference for isolated cells that also have more SREH available for their rotation, and slightly enhanced (8 m/s) 0-1 km shear marginally conducive to tornadoes but relatively high LCL (1500m) is a limiting factor.

Large hail can be produced by supercells and large multicells, as well as isolated severe gusts. The relatively dry inflow air will likely limit hail size but evaporative cooling could enhance gusts. The anticipated MCS could become the main producer of strong to severe gusts. An upgrade to level 2 may be considered according to convective mode.

…Adriatic Sea…

A convergence line is indicated by GFS to cross the Adriatic Sea along with some instability, most prominent near Albania at 18Z. Its proximity to the jet and enhanced SREH and low level shear may yield an organized storm or MCS capable of isolated large hail… but no confidence in a threat level for this small area.

Meteosat 9 - Enhanced IR10.8: time sequence

This first loop shows the convective development using the IR10.8 channel. The images are artificially enhanced to ensure a better discrimination of the several convective stages. The change from U-shaped storm into ringshaped storm can be well observed.

Meteosat 9 - BTD WV6.2-IR10.8: time sequence

The BTD stands for brightness temperature (WV6.2-IR10.8). The area of negative BTD (WV6.2 - IR10.8) pixels is shown in greyscale, while the pixels of positive BTD are depicted in color, according to the inserted color scale. So, generally the colder the cloud top, the larger the positive BTD.

Meteosat 9 - IR3.9 Radiance: time sequence

SEVIRI IR3.9 channel displayed as “radiances”. The darker areas in the animation therefor represent the areas with a higher cloud top 3.9 micron refelctivity, resulting from a presence of very small ice particles, most likely generated by strong updrafts.

Meteosat 9 - Severe Convection RGB: time sequence

The convective development over Central Europe is here pictured using the so-called “severe convection RGB”. On red the brightness temperature difference (BTD) of the two water vapour channels 6.2 and 7.3. On green the BTD of the infrared channels 3.9 and 10.8 and on blue the BTD of the two visible channels 1.6 and 0.6, respectively, are pictured. The yellowish colors here correspond to the areas of higher 3.9 micron reflectivity..

Meteosat 9 - HRVIS: time sequence

The HRV (High Resolution Visible) files were tuned manually in Photoshop, using unsharp mask enhancement (under “filters” there). Striking in this loop is that the storm seems to maintain a well-defined above-anvil plume for substantial period of its life

Meteosat 9 - Dust-RGB: time sequence

Allthough more suited for the detection of dust in the atmosphere this RGB can also be of use for monitoring convection. The asscosiated icing makes the cells appear as red. Second remarkable is that with the use of this RGB one is able to make a clear distinction between the thick “active” part of the thundercloud and the thin cirrus shield (Anvil) surrounding the same cloud. As the convective development continues this thin cirrus shield surrounding the MCSs appears as black.

Meteosat 9 - WV6.2: time sequence

In the WV6.2 animation a carefull look reveals an even interesting feature, at the early stages of the storm evolution (about 12-15), something resembling “dark diverging rays” are observed which are artefacts of the SEVIRI instrument. This occurs is along the scan line of SEVIRI. The instrument always scans 3 lines in one go with three different detectors. So, there must have been a slight hiccup with one of the 3 WV detectors or its read-out chain. It is a subtle deviation of the order of 0.1K (see images below) that emerges because of our particular focus.

Meteosat 9 Enhanced WV6.2 - IR10.8 - 22 April 2008 1300UTC. Artefacts indicated by arrows

Meteosat 9 Enhanced IR10.8 - WV6.2 - 22 April 2008 1300UTC. Artefacts indicated by arrows

Meteosat 9 WV6.2 - 22 April 2008 1300UTC. Artefacts indicated by arrows

Meteosat 9 WV6.2 - 22 April 2008 1300UTC. Artefacts indicated by arrows

NOAA15 - AVHRR

The set of NOAA15 images (which passed over the storm at about 14:29 - 14:30 UTC), remapped to a Mercator projection, pixel size 500m are presented in this animation. The band 3B image is displayed as reflectivity, band 4 is enhanced using my standard color enhancement (color bar inserted). The combination of RGB bands 1,2,4 with the color-enhance band 4 is done in Adobe Photoshop, playing a bit with layer blending and transparency.

An intense updraft lifts moisture high into the troposphere and strong upper level divergent winds move the extremely cold moisture (ice crystals) downwind. The moisture fans out as it moves downwind generating a V-shape of the cloudiness on a satellite image (see the cell situated to the South of Romania). The HRVIS image was taken at 1515UTC.

The movement was from south to north so soon Bucharest was affected by the storm. The radar image below shows the intense reflectivity (over the Bulgarian teritory). In this case, the satellite image (!) provided more information about the severity of the system (strong updraft, V-shape).

A few hours later, the same airmass that produced cells with V-shape on satellite image, also produced a High Precipitation tornadic supercell, detected by the WSR-98D S band radar in SE Romania.

Doppler radial velocity image, at the bottom, has an intense mesocyclone and in the reflectivities, we can see the supercell structure, with a very pronounced “hook echo” to the S, and V-notch structure, produced by the flow splitting around a very strong updraft. Very high reflectivites, up to 70 dBZ, due to hail presence. Tornadoes were reported in Bulgaria and Romania

Radar: time sequence

National Romanian Radar Mosaic - Base reflectivity

Radar: time sequence

National Romanian Radar Mosaic - Echo tops

Radar: time sequence

National Romanian Radar Mosaic - Composite reflectivity