Table of Contents

1. Introduction

On 4th October 2007 at 1530 UTC, a large MCS (Mesoscale Convective System), passed over the Balearic Island of Mallorca. Associated to the MCS an F2 tornado was reported. The MCS was accompanied by severe wind gusts that spread the damage in numerous areas of the island.

Fig 1.1 - Huge damage in Mallorca streets and highways as a result of the F2 Tornado and severe winds gusts from the MCS itself

The longevity of the MCS was quite remarkable (approx. 12 hours!) , and in some stages of its life cycle, it showed a fast movement, taking a very short time, less than 1 hour, to cross Mallorca island. To show this please have a look at the below illustration of the path of the MCS. Between 07:45 and 10:45 the MCS stay almost stationary before it takes a rapid NE course leaving the Forecasters with almost no time to warn.

Fig 1.2 MCS approximate track and relevant times (UTC). Verify the irregularity of the velocity of the MCS. For example, in the first phase (probably steered by a supercell embedded in the southern flank) the MCS takes more than three times the time employed to cross Mallorca Island, for travelling a similar distance. Also, in this first stage, the direction of movement is not following the steering winds.

Surface wind gusts in excess of 60 knots where registered in a couple of Mallorca weather stations, and also precipitation rates higher than 20 millimetres in 20 minutes in Palma de Mallorca and other spots of the Island.

The convective system initiated at 0700 UTC, in front of Palos Cape, as a Supercell storm, in south-eastern Spain (Murcia region), producing heavy rain along most of its way. It maintained its supercell character until 11:20 - 11:40 UTC approximately; then progressed to NE, crossed Mallorca between 15:30 and 16:15 and reached its dissipating stage by 19:00 UTC, off the northern Catalonian coast.

The social and economic impact of the event was enormous, and can be assessed from many sources, including the internet, newspapers, etc..., but this is not the purpose of this case study. This work tries to analyse the meteorological situation from observations and NWP products, and to emphasize the key factors that, according to the authors, could have played a major role when trying to understand the severity of the event. Another aim of the study is to make available to European forecasters as much material as possible, facilitating further research.

If you think you can contribute to this Case Study or you have suggestions or questions please feel free to contact us.

1.1 Surface Weather Observations

In this section we will try to summarize the impact of MCS arrival to Mallorca Island in terms of available surface weather reports, including synops, metars, and automatic weather stations observations. Finally, we include a subjective analysis of MCS gust front isochrones while passing over the island.

Synop observations at 15 UTC

Surface Observations over the Balearic Islands show a clearly disturbed mesoscale wind field, with 20 kt westerly winds over Mallorca, in place where easterly winds were forecasted for this time. Only the observation over Menorca (the most eastern Island) is consistent with the 12 UTC 3 hours forecasts. By this time, a tornadic bow echo (see radar chapter) still over the sea was about reaching Palma de Mallorca city, (it did by 15:30 UTC). This lets us think that from the location of the two wind plots over the island, to the MCS gust front, there should be a marked wind speed gradient, not resolved by observation network, but confirmed by the more frequent metar and speci reports at Palma de Mallorca airport, and also by authomatic weather stations. Check that the rest of the environment in the image is not windy at all, stressing the fact that the commented wind field is purely a mesoscale feature promoted by MCS activity, and not forecasted by operational available models.

Metar and Speci reports at Palma de Mallorca airport issued from 15 to 16 UTC

Going to a deeper detail, by zooming into Palma de Mallorca airport, we can easily follow surface weather impact of MCS gust front while passing over it. Notice the sudden increase in wind speed and associated gustiness at 15:30 , and also the marked wind direction shift, from easterly to westerly. The opposite change take place at 16:00 as a result of mesoscale feature leaving the airport. Notice also the sharpness of gust front main core of values above 50 kt, that takes around 10 minutes, from 15:30 to 15:39, to cross the airport.

Subjective isochrones meso-analysis

Note: The times on this analysis are local time (e.g. UTC+2)

This analysis is provided by Agustí Jansà (AEMET) who is delegate at the Balearic Islands and gives a quick review of the impact of MCS passage over Mallorca (green and red spots represent damaged electric power towers). Despite we have only observations over land, is not hard to imagine that MCS activity in terms of wind, should have been similar during its whole path, from the beginning off Murcia coast to the end of its life cycle.

2. Basic and Derived Parameters

This chapter will present you a range of satellite images from MODIS and Meteosat 9 and a combination of these images with a set of basic and derived ECMWF parameters. Starting of which please have a look at the following MODIS image of 4th October 1143UTC. A mesoscale convective system can be seen in Figure 2.1 (Click to enlarge). The radar imagery (see radar chapter) suggest the presence of a Supercell thunderstorm in the southern flank of the MCS during its initial phase. A close look of Fig 2.1 even reveals some eminent signs of the severity of this cell as overshooting tops and some resulting gravity waves are visible. Towards the southern edge of the cell we can identify a fierce V-shaped structure.

Fig 2.1 - Modis image of 4th October 2007 1143UTC. The convective cell is evident at this stage with associated overshooting tops
and due to the inflicting perturbation at the upper layers of troposphere also some gravity waves are seen.

Although we look at a static image it is possible to recognise different air masses. Each having its own characteristic helping us to solve the puzzle together to explain why the developments were so explosive.

  • The image reveals there is a low level flow coming from the east "feeding" the MCS. This is best seen by the gravity waves observed east of the system.
  • On mid-level flow a thickness ridge cloud from the Sahara is observed. It is observed a thick patch of clouds on the far right side of the image. Flanked by a dust cloud from the Sahara it is a sign of the southerly flow.
  • The dust cloud also is the cause for a capped inversion. Continental African mid-level air over Mediterranean colder air, is supposed to produce a strong capped air mass, to the east of the MCS.

In the next three subchapter we will start to look at the different animations of satellite images and combine them with NWP.

2.1 MSG Satellite images

Meteosat 9 IR10.8 - Time sequence
In this series of images the convective developments over the Balearic Islands are described using plain Meteosat infrared 10.8 μm. The channel is appropriate as it pictures the ice particles from high clouds quite clearly. In a sequence of 15 minutes the relevant satellite image are presented.

Meteosat 9 Enhanced IR10.8 - Time sequence
In this chapter again Meteosat 9 infrared 10.8 μm is shown, but the images have been artificially color enhanced. This will improve the discrimination of where most ice particles form during the several convective stages. In a sequence of 15 minutes the satellite images are presented and described.

Meteosat 9 WV6.2 - Time sequence
The chapter presents the several convective developments over Mallorca using Meteosat 9 WV6.2 channel. The channel is suitable as it gives an idea of the upper tropospheric humidity (UTH) and gives a view of the upper air dynamics. 15 minutes sequence of images are presented.

Meteosat 9 HRVIS - Time sequence
This chapter will show the HRVIS images of 4th October. Especially with the smaller scale cells the high resolution channel allows us a very good monitoring of the convective development. In a sequence of 15 minutes the satellite images are presented and described.

Meteosat 9 Severe Convection RGB - Time sequence
The Severe Convection RGB makes use of the NIR1.6 and IR3.9 μm channels. Both of these channels provide information on the microphysics within the clouds. E.g. with this RGB we are able to discriminate between smaller and larger ice particles. This is useful for detecting convection. Cells that are bound to severe updraft will likely be characterised by smaller ice particles, as there is less time available for the ice particles to collide and form bigger particles. When a cell has reached its mature stage we will find larger ice particles. With this in mind click the above link and study the set of Severe Convection RGB images to identify at what stage the updrafts within the MCS is the highest and at what time step the cell reaches its mature stage.

Meteosat 9 Dust RGB - Time sequence
As the names says the Dust RGB is useful for detecting dust. Already in Fig 2.1 (Modis image) we could recognise the presence of dust. However using this RGB is also very helpful when looking at convection. You will learn that with this RGB you can easier discriminate between the thick ice part of a convective cell and the anvil which is a thin cirrus shield.

2.2 Synoptic Analysis (ECMWF)

Meteosat 9 with ECMWF Overlay (Synoptic scale)
This subchapter presents the set of ECMWF parameters combined with the satellite image. The interactive overlays gives you the freedom to combine and study them. A short description on the most interesting features is given to provide you with some extra guideline.

2.3 Mesoscale Analysis (ECMWF)

Although ECMWF is not a mesoscale model, we found it of interest to zoom in into the area of interest to have a more detailed, full resolution view, of different relevant parameters that helps to understand which mesoscale processes could have played a major role in this convective event.

Meteosat 9 HRV with ECMWF Overlay (Meso scale)
The subchapter on the mesoscale situation shows you for 12 and 15UTC the HRVIS images in combination with the following ECMWF parameters; 1000, 925 and 850 hpa streamlines, Mean Sea Level Pressure, Isotachs at 700 and 300 hPa and the potential temperature at 700 hPa. The interactive overlays gives you the freedom to combine and study them. A short description on the most interesting features is given to provide you with some extra guideline.

3. Radar

Unfortunately when this case happened in October 2007 over Mallorca no radar was (yet) available on the island. To follow and monitor the convective system we will rely on the radar images from three stations of the Iberian mainland, namely Valencia, Murcia and Barcelona.

Fig 3.1 - Radar layout of east coast of Spain with the three radar stations Murcia, Valencia and Barcelona. Since 2008 also a radar is available at Mallorca.

Murcia Radar - Time sequence
From this sequence we can have an idea about the first stages of the convective systems life. Despite the fact that the main radar core velocity field is not seen by Murcia radar, because it is out of range, there are some indirect indicators that tell us the system could have been a supercell during some time of its youth, but we can not be sure of this at all, as no radar derived wind information is available.

Valencia Radar - Time sequence
From 1310 UTC to 1510 UTC the best viewer among the available radars, was the one located at Valencia. During this period a very intense (reflectivity cores above 60 dBZ) bow echo structure was formed that finally reached Mallorca Island at 1530 UTC.

Barcelona Radar - Time sequence
The Mesoscale Convective System arrived in Mallorca Island at around 1530 UTC. The radar situated in Barcelona showed the best reflectivity view for this time. This sequence shows the passage of the system through the island.

Vertical Integrated Liquid Water
Radar derived Vertical Integrated Liquid water, VIL, is one of the key products for severity watching purposes. In Figure 3.2 the VIL values from Murcia radar for 1100 UTC is shown.

Fig 3.2. Vertical Intergated Liquid Water derived from Murcia Radar for 4th October 2007 1100UTC overlayed on the corresponding Meteosat 9 HRVIS image.

It shows which can be the most active part of the convective system, in terms of severe surface weather potential, by the time when it showed more clear supercell signatures.

Verify yourself that the percentage of area covered by the VIL high values is quite small compared to the size of the system from a satellite perspective. The main VIL core in the image, as a first guess, could also be considered the area with the highest precipitation rates at the at the surface or even large hail.

4. Convective Parameters

In chapter 2 we have given you the set of basic and derived parameters from the ECMWF model. This chapter will go a bit more in depth in the topic of convection by presenting you a range of convective parameters from the HIRLAM model. In addition we also try to make the link in this direction by demonstrating you the potential of the various products from the NWCSAF and from MPEF. For latter we will show you the Global Instability Index product that uses a first gues of ECMWF and updates it with MSG information to retrieve a sounding and a stability analysis. Finally we this chapter will give some vertical profiles made over Mallorca and compare them with the forecasted soundings.

HIRLAM: Precipitation forecast

The model run from HIRLAM of 3rd of October 0000UTC is the basis for the calculation of the total precipitation that was forecasted during the events of this case over the Iberian Peninsula and the Balearic Islands. These forecast images are shown in combination with the enhanced IR10.8 satellite image to make a comparison between model
and real time imagery.

HIRLAM: Convective Parameters

In the above two image there are 8 panels plotted that a forecaster in Spain has operational to see from the model where there is potential for convection. In this part of this chapter we will zoom in to this product and look at these different parameters from the HIRLAM model and describe the details we see in relation to this severe weather over the Balearic Islands. We will start with shortly identifying which these parameters are:

  1. PW, or precipitable water, is, in this interpretation of AEMET, the amount of liquid water, in mm, if all the atmospheric water vapour in a column from the surface to 300 hPa. were condensed. High values of PW in clear air often become antecedent conditions prior to the development of heavy precipitation and flash floods. When high PW values areas present a lifting mechanism and warm advection in low levels, heavy precipitation often occurs. These data can provide to forecasters an important tool for very short range forecasting.
  2. The opposite to CAPE is CIN which we also see plotted in the right image in the top right panel. CIN, which stands for convective inhibition is a numerical measure in meteorology that indicates the amount of energy that will prevent an air parcel from rising from a given level to the level of free convection. Sometimes is is also referred to as CAP or CAP inversion. There are many ways of compute CIN, AEMET’s diagnostic tool takes an average parcel represented by first 100 mb of the atmosphere to compute it.
  3. The winx in the bottom left is derived from the WINDEX or Wind Index which is based on observations and numerical models and gives an estimation of max wind gust that can be generated by convective processes. It is considered that when a gust front moves perpendicular and towards high values of WINDEX contours, downburst activity has a chance to take place.
  4. CIZ6 is an indirect measure of vertical wind shear computed from the difference between the lowest 6 km mean wind and the lowest 500m mean wind. It is equivalent to twice the square root of the BRN shear, which is more widely use in forecast offices for the same purposes. The advantage of Ciz6 is that it offers to the forecaster values in m/s directly, (instead of m2/s2 given by BRN shear) which are more intuitive for the assesment task. According to most climatologies performed in the US, this parameter have a very good skill for discriminating organized convection: values between 12 and 24 m/s have to be taken into account for this pourpose. Values higher than 24 m/s tend to be detrimental for convective development, and values below 12 m/s tend to be not enough for well organized convection.
  5. Top left the Lifted Index is plotted. Negative values indicate potential buoyancy for a low level average parcel.
  6. Top right CAPE is plotted. CAPE which stands for (Convective Available Potential Energy) is a measure of the amount of energy available for convection.
  7. To the bottom left we see SRH which stands for Storm Relative Helicity. In this case SRH refers to storm relative helicity in the first 3km. It helps to estimate the capacity of a thunderstorm to produce rotating updrafts in a given environment. The computation is dependant on the direction and speed of the storm, so an approach has to be taken into account before computing it. In the case of AEMET diagnostic tool, the classical approach is adopted, that is to consider that the storm will move 30º to the right of the environmental mean flow and at a speed which is a 75% of that of the mean flow. A forecaster added value to this paramenter is straightforward by just changing the classical approach with the radar observed storm speed and direction, which would give much more realistic SRH values.
  8. Bottom right ACON is a combination of several convective parameters where different thresholds are set:

Deep convection LI<0; CAPE>600J Kg-1; CIN<300 J Kg-1
Structurised deep convection CAPE>700J Kg-1; CIZ>9 m s-1; RH (70 - 50 kPa)<60%
Supercell CAPE>700J Kg-1; CIZ>9 m s-1; RH (70 - 50 kPa)<60%; SRH>150 m2 s-2

If you look closer you also see in some of the figures some lines and windbarbs plotted. This is the direct output from the model that provide guidance about the dynamics in the area of interest.

Nowcasting SAF products

Cloud Top Height
The cloud top temperature and height (CTTH), developed within the SAF NWC context, aims to support nowcasting applications. This product contributes to the analysis and early warning of thunderstorm development. Other applications include the cloud top height assignment for aviation forecast activities. The product may also serve as input to mesoscale models or to other SAF NWC product generation elements.
The CTTH product contains information on the cloud top temperature and height for all pixels identified as cloudy in the satellite scene.

Convective rainfall rate
The objective of the CRR product is to estimate the precipitation rate associated to convective clouds. The final output is a numerical calibrated product (in mm/hr) divided into classes in an image format. This product provides to forecasters complementary information to other SAF NWC products related to rain and convection monitoring as Precipitating clouds and Cloud type.

Precipitating clouds
The objective of the PC product is to support detailed precipitation analysis for nowcasting purposes. The focus is on the delineation of non-precipitating and precipitating clouds for light and heavy precipitation, rather than quantifying the precipitation rate. Particular attention will be given to the identification of areas of light frontal precipitation.
The product provides probability results, i.e. probabilities of precipitation intensities in pre-defined intensity intervals. From the probabilities a categorical estimate of precipitation intensity may be derived. It is not intended to provide information on the type of precipitation.

Rapid developing thunderstorm
The Rapidly Developing Thunderstorms product is a tool for monitoring convection from MSG data and is used for early detection of storm clouds.
The basic objectives of the Rapidly Developing Thunderstorms product are twofold: the automatic identification, monitoring and tracking of intense convective systems as well as the detection of rapidly developing convective cells. In other words, this product aims to assist in the automated detection of convection clouds.
The convective systems are presented like "objects" within the satellite images together with their most relevant properties (size, movement, minimum temperature, area and temperature trends, etc.) generated for the expected significant weather. The product highlights the most active convective cells. It may also serve as input for the automated convection detection used in aviation meteorology.
The Rapid developing thunderstorm product is used during the early stages of thunderstorm identification - from the time when convective clouds form to the stage at which the cloud shield has developed at the vertical level of the tropopause.

Total precipitable water
Total Precipitable Water (TPW) is the amount of liquid water, in mm, if all the atmospheric water vapour in the column were condensed. High values of TPW in clear air often become antecedent conditions prior to the development of heavy precipitation and flash floods. When high TPW values areas present a lifting mechanism and warm advection in low levels, heavy precipitation often occurs. These data can provide to forecasters an important tool for very short range forecasting. Within the SAF NWC context, the main goal is to provide TPW data in clear air pixel by pixel in image format for Nowcasting purposes.

Stability analysis imagery
The Stability Analysis Imagery (SAI) was developed by the NWC SAF. The central aim of the SAI is to provide estimations of the atmospheric instability in cloud-free areas. Among all potential indices the Lifted Index (LI) has been implemented and codified and presented in this case for central Europe on the 19th June 2006. The lifted index of SAI is only done for clear sky conditions, therefor for SAI the first step is to compute the Cloud Mask product (CMa). This CMa allows the identification of cloud free and cloud contaminated areas. The SAI product itself uses the corrected normalized IR SEVIRI radiance values of the following channels WV6.2, WV7.3, IR8.7, IR9.7, IR10.8, IR12.0 and IR13.4μm), and provides as output the normalized lifted index.


Global instability index (GII) is an airmass parameter indicating the stability of the clear atmosphere. The GII product should serve as a nowcasting tool to identify the potential of convection and possibly of severe storms in still preconvective conditions. The applied retrieval method makes use of six MSG SEVIRI thermal bands, and together with the a priori information of forecast profiles, the scheme infers an updated atmospheric profile for each MSG pixel, from which instability indices can be computed. Several instability indices are used in this case and presented. The images are presented in 1 hour sequence.

The K-index is a widespreaded method amongst meteorologists to make a stability analysis of the atmosphere. In the above link the K-index as computed by the GII algorithm is presented in 1 hourly interval. To find out more on how the K-index is computed you can look at the following animation.

Lifted Index
A second index that is computed from GII is the Lifted Index. In 1 hour interval the Lifted Index is presented for the 19th June 2006 over Central Europa. If you want to learn more on Lifted Index and how it is derived click "here".

Precipitable Water
One final product to be presented is the precipitable water. For a Meteorologist this product can be of extreme value when doing a nowcast. It represents the total atmospheric water vapor contained in a vertical column of unit cross-sectional area extending between any two specified levels, commonly expressed in terms of the height to which that water substance would stand if completely condensed and collected in a vessel of the same unit cross section.

Vertical Profiles

In the below image the observed sounding of Palma de Mallorca is plotted.

Fig 4.1 - Observed sounding Palma de Mallorca. 4 October 2007 12UTC

Striking and main features are the:

  1. Presence of low level and deep layer shear, with a 70 kt southwesterly jet at around 5km
  2. Large convective inhibition, which will prevent initiation unless a significant forcing available. CIN will decrease with time as a result of low level advections. Also, a mesoscale forcing source will arrive in place at 15 UTC, as we know now, after the event.
  3. A shallow layer of latent instability near the ground, with no so large associated potential bouyancies (most unstable CAPE =MUCAPE=335 J/kg)
  4. Presence of an elevated mixed layer (EML) from 875 to 700 hpa (elevated source of African air, as pointed out in chapter 2 when the MODIS image comments)

The source of the EML can be traced back and confirmed by checking the sounding from Dar El Beida in Algeria. The sounding for 4th october at 00 UTC, confirms that the EML found in Mallorca at 12UTC is advected by a southerly and mid-level flow which is directed towards the Balearic Islands.

Relevant to this and to demonstrate the performance of the model we will show you the analysis of ECMWF for 12UTC and the +3 hour forecast for 15UTC for the sounding over Mallorca Bay (39.5°N, 2.6°E)
Fig 4.2 - ECMWF analysis of 12UTC at Mallorca bay (39.5°N, 2.6°E). In red: temperature / green: dew point / blue: wet bulb temperature / black: most unstable parcel lifting path / 17.5 critical saturated adiabat outlined for clarification of latent instability layers

Fig 4.3 - The 3 hour forecast valid for 15 UTC vertical profile at Mallorca bay (39.5°N, 2.6°E). The layer of latent instability is forecasted to increase markedly, mainly as a result of temperature and dew point advection at low levels

Regarding the three hours forecasted evolution of the vertical profile, two main features are clearly seen at low levels:

  1. A marked increase in depth of latent instability layer, as a result of positive advection of temperature and dew point at low levels. This is also reflected in the increase in MUCAPE value from 687 J/kg at 12 UTC to 1002 J/kg at 15UTC. Is important to stress here, that despite CAPE values are not impressive, there is a large number of parcels with moderate positive buoyancies, which is very relevant for the assesment of convection intensity.
  2. An increase in low level shear and helicity, as a result of the intensifying low level easterly flow, and also medium level southerly flow.

Fig 4.4. - Zoom to the lower layers of troposphere. ECMWF analysis (12Z) (top) and 3 hour forecasted (15Z) soundings (bottom)

Verify that most unstable parcel at 12 UTC is at the surface, but at 15 UTC this parcel is at 850 hpa, where advections of temperature and humidity are forecasted to be larger (as a result of a marked increase of wind speed at this levels, from 15 to 30 knots).

NOTE: It should be taken into consideration that these profiles have been obtained from model output in standard pressure levels, so vertical resolution is not so high, and errors in convective parameters are to be expected. Still this model output is enough to show the main features we want to stress here.

At upper levels, we can also confirm the approaching of the jet streak from 12 to 15 UTC. The max winds are located around 9-10 km from MSL (see mesoscale analysis and forecast section). A corresponding increase in deep layer shear is also to be expected. For instance, forecasted BRN shear for 15 UTC is 202 m2 s-2, which is really a large value, implying that for convection to be sustained in this environment, a high degree of organization and intensity is required, otherwise, so much shear would be detrimental for the maintenance of convective activity.

Fig 4.5 - Comparison between the analysis and the three hour forecasted sounding of ECMWF.

Finally, although is not easy to interpret them for this case, we show a comparation of 12 UTC ECMWF run, analysis (left) and 3 hour forecasted (right) hodographs, at the same location as before (Lat, Lon) = (39.5N, 2.6E), and again calculated with standard pressure levels. Together with the hodographs, storm motion vector and storm relative winds are plotted.

For convective parameters computation, a storm motion vector has to be adopted, and here we follow the traditional approach, so the storm is supposed to move with 75% of mean flow speed and 30º to its right, which is not very realistic in this case, because our MCS, by 15 UTC, did not move to the right of the mean flow, even did not moved at the forecasted speed, but markedly faster.

Still, this simplified approach, give some insight on the evolution of kinematic convective parameters, that can be of use in an operational context in future cases and also, from the training perspective, we consider good idea to start using hodographs information as a side tool in the nowcasting proccess.

Please make note of the clockwise curvature of the hodographs, especially at low levels, in agreement with the low level warm advection, that allows for positive SRH values. Check that 3 km SRH is under-calculated by the software, because of the lack of two sampling points between 2 and 3 km height, so, obviously, a value much larger than 212 m 2 s-2 at 15 UTC should be expected. Because of similar reasons, 1 km and 2 km SRH’s are also under-calculated. Regarding the 3 hours evolution of the hodograph, apart from the increase of SRH values, three other key ingredients are present:

1) A marked increase in low levels storm relative flow by 15 UTC (size of the purple vectors)
2) A marked increase in low level shear in layers 1000-925 and 925-850 hPa.
3) A quite large streamwise vorticity at 1000 and 925 hpa (shear vector perpendicular to storm relative wind)

Allthough these three elements are considered necesary for tornadogenesis, obviously, their presence can never be taken as a predictor. Also, the environment attached to the arriving forcing source (the MCS gust front) can be quite different than the forecasted one, but this can not be assessed in this case study, because of the lack of radar derived wind information at low levels.

5. Conclusions

Main conclusions, key ingredients and relevant considerations related to this severe weather episode are presented below

1 - Causes of main MCS initiation and first stages peak intensity

  • The setting of a low level mesoscale cyclonic circulation, off Murcia coast (see mesoscale overlays), before MCS initial stage, which was quite well forecasted by ECMWF model in time and place.
  • The presence of an upper level vorticity maxima collocated nearly in place, just a bit to the west, above the before mentioned low level circulation, close to the left exit region of an upper level southerly jet (see chapter 2.1 – WV6.2 animation).
  • High values of low and deep layer shear together with high values of storm relative helicity in the area of initiation, that helped the system to develop rotation.
  • CAPE values above 800 J/kg for a quite deep layer in the easterly low level stream entering the cyclonic circulation, together with non buoyancy driven vertical accelerations, helped the convection to be very intense and particularly deep (see chapter 2.1 – Severe Storms RGB) in these first stages.

2 - The onset of supercell phase: impact on model performance

  • It is believed by the authors, that first stages of the MCS (from 08 to 11 UTC) were in supercell form, more specifically, a supercell with a quite slow right mover character, that produced very large amounts of accumulated precipitation during its 3 hours of life (see exercise 4).
  • Although no doppler information was available, other radar and satellite signatures, together with background experience, support further this hypothesis.
  • It is also believed that because the anomalous propagation to the right, operational numerical models lost track of the system from this initial stage onwards, therefore reproducing most of its impact (including precipitation) well to the west from where it really happened, leaving the forecasters with no reliable model guidance, not in terms of precipitation, neither in terms of surface wind.

3 - MCS unexpected track and reasons for longevity

  • After the MCS left its supercell stage, it got catched and steered by the mid level flow and ended up in Mallorca island, by 15:30 UTC, having followed a path clearly shifted to the east with respect to operational forecasts, mainly because the right moving character of initial phase.
  • Also, a not enough accurate short range forecast of upper level jet position (see chapter 2.1 – WV6.2 animation) could have played an important role in reducing the accuracy of the forecasted MCS track.
  • Reasons for MCS longevity can be derived from the presence of a well defined low level thermal boundary (see mesoscale overlays) attached and to the east of MCS track. This thermal boundary, very importantly, was nearly parallel to the mid level flow, allowing the system no to get cut off from the low level ageostrophic moist and warm easterly advection during its whole track. Also, the boundary, acted as a “slider” preventing the MCS from entering the large CIN area to the eastern side of it (see comments below fig. 2.1 in Chapter 2).

4 - Tornadic nature of MCS at its arrival in Mallorca

  • Although we are somehow blind to see the insides of the system when arriving in Mallorca, because of the lack of nearby radar information, the radar structure of the cell by this time, as shown by the distant Barcelona radar, is very similar to a bow echo (meteograms at Santa PonÇa station helped us to support this idea; see exercise 1)
  • The tornado itself, could have been produced or favoured by interaction of the leading edge of MCS gust front, with a deep, sharply defined low level moisture ridge (see exercise 7) that provided the large buoyancy / low NCA environment usually needed for tornadogenesis.
  • Speculations on prior rotation provided by book-end vortices in the bow echo structure or baroclinically generated rotation derived from the sharp thermal and moisture gradients, can be made, but are difficult to demonstrate, because of the lack of nearby radar wind derived information.

5 - Utility of proximity forecast soundings and hodographs information

  • The analysis of vertical profiles shown in section 4.3, and the evolution of them over Palma de Mallorca, from 12 to 15 UTC, has proved to be quite valuable, although not conclusive, for the assessment of tornado potential, showing clearly favourable tendencies of the main convective parameters commonly used for this type of assessment, although, we again want to remark, this can never be taken as conclusive.

  • Also the 3 hours evolution of the hodographs over Palma seems to be quite of help in the assessment of kinematic parameters tendencies, specially in the task of the assessment of source levels with large rotation potential (large streamwise vorticity values), which in this case were coincident with levels with very large potential buoyancies.