6th European Conference on Severe Storms

Short Abstracts

Paper 1:


Medha Deshpande, HCL Infosystems Ltd. working for IITM Pune

Purva Salvekar, IITM Pune


The physics of tropical cyclone formation (genesis) is not well understood, and more is known about mature cyclone than the formation mechanism that produces it. The numerical experiments are conducted to simulate the genesis and further intensification of the Severe Cyclonic Storm (2001), formed over the North Indian Ocean (NIO) using the Fifth Generation of Mesoscale Model (MM5). The model is integrated for 168 hrs starting from 00 UTC of 19 May 2001 to simulate the genesis of the vortex on 12 UTC of 21st May. The impact of domain size (Small 110 x 110 grid points and Large 165 x 200 grid points), resolution, cumulus parameterization schemes and nesting level is investigated. The simulated results obtained with large domain which covers the large scale monsoon flow and high resolution agree reasonably well with observations provided by India Meteorology Department (IMD) in terms of time variation of location (track) and intensity. Among the three cumulus parameterization schemes namely Grell (Gr), Betts-Miller (BM) and updated Kain-Fritsch (KF2) scheme the former one gave least track error. The schemes of BM and KF2 both show movement in the northeast direction which is far away from the IMD estimated data. Regarding intensity variation in Gr, after formation upto 72 hrs of integration intensity remained same then started deepening similar to IMD estimation but with underestimation of intensity. It is important to note that the integration is started on 00:19-05-2001 well before the formation of vortex (on 12:21-05-2001). The best combination experiment gives the formation of the vortex at 13.90N/68.80E with CSLP of 1000 hPa on 12 UTC of 21nd (after 60 hrs of integration) and at that time as per India Meteorology Department (IMD) the observed CSLP was 1000 hPa at 13.50N/69.00E. After genesis the simulated system initially moves in northeastward direction, then in the northward and thereafter northwestward, well matching with the IMD estimated track. The maximum intensity is 950 hPa as against IMD estimated values of 932 hPa at 12 UTC of 24th May 2001. This clearly shows that MM5 model has capability to simulate the genesis and further intensification of TC in the Arabian Sea.

Paper 2:

A Study of Quasi-millennial Extratropical Cyclone Activity Using Tracking and Clustering Methods

Lan Xia, Centre for Coastal Research, Helmholtz-Zentrum Geesthacht

Frauke Feser, Centre for Coastal Research, Helmholtz-Zentrum Geesthacht

Hans Von Storch, Centre for Coastal Research, Helmholtz-Zentrum Geesthacht

Matthias Zahn, University of Reading


There is growing interest in knowing the impact of climate change on extratropical northern hemisphere cyclone activity. In a warming global climate, properties of extratropical storms such as intensity, frequency, and distribution of genesis regions are expected to differ from nowadays climate. To analyze long-term change of these properties, mean sea level pressure fields (MSLP) of a quasi-millennial (1000-1990) global climate simulation by ECHO-G are applied to track storm events using a previously developed tracking algorithm (Hodges 1996). For validation reasons tracks gained from ECHO-G simulations are compared to those from NCEP/NCAR reanalysis data. The numbers of tracks from the ECHO-G simulation data are on a similar level as those derived from the NCEP/NCAR reanalysis data. Densities for genesis regions representing the origins of storm tracks are also compared. By using a clustering analysis that sorts the tracks into different categories, temporal changes in different regional storms activities are examined. Track types are identified and frequency, density, and lifespan of these types are individually studied for temporal changes from simulation year1000 to 1990.

Paper 3:

Modeling Severe Hail Risk in the Contiguous France from Re/Insurance Perspectives

Jianming Yin, Tokio Marine Technologies LLC


Hail can be a significant hazard in Europe especially in the summer months. Though infrequent in nature, catastrophic losses to buildings and automobiles due to extreme hailstorms in Europe underscores the needs for the insurance industry to adequately quantify the hail risk in Europe. In this endeavor, a severe hail risk model is developed for the contiguous France. After assessing the availability and quality of the hail reports from various sources, European Severe Windstorm Database (ESWD) severe hail reports (hail diameter 20 mm or greater) are selected to model the severe hail risk in France. An empirical climatology of severe hail is derived for the contiguous France by applying a bivariate Gaussian Kernel smoother with the optimal bandwidth to the ESWD hail reports. The optimal bandwidth or optimal standard deviation for the Gaussian Kernel smoother is achieved using the out-of-sample Jackknife cross validation technique. This empirical climatology of severe hail reveals the relative frequency of the severe hailstorm occurrences across the contiguous France and is found to be reasonably comparable with other independent studies. It is hence used for the hail risk simulation. The hail risk model developed in this endeavor is not only intended to price individual site with hail coverage but also to allow companies to assess their reinsurance needs at portfolio level. Therefore, a severe hail event concept is introduced and defined as a congregation of individual hailstorms spawned by the same convective precipitation system moving through the contiguous France in a 72-hour time frame. Historical severe hail events are extracted from ESWD hail data. The number of severe hail events in a year is modeled using a Poisson distribution while the number of severe hailstorms in a hail event is modeled using an exponential distribution. Monte Carlo simulation techniques are employed to simulate up to 1,000,000-year worth of future possible severe hail events.

Paper 4:

Wind Field Distribution within Hurricane Force Extratropical Cyclones over North Pacific and Atlantic using QuikSCAT scatterometer Measurements



Joseph Sienkiewicz, NOAA/NWS/NCEP/OPC


Extratropical cyclones (ETCs) that reach hurricane force intensity are a significant threat to the safety of life at sea and a risk to cargo and vessels. ETCs vary on scale from less than 100 km in diameter up to 4,000 km in diameter and have an average life cycle of five days from genesis to death. Associated wind conditions can range from light (10-20 kts) to hurricane force (HF) (>63 kts). Knowledge of the wind structure, the frequency of occurrence and distribution of HF winds in ETCs has been greatly enhanced by data from QuikSCAT scatterometer. QuikSCAT wind measurements have shown that HF winds in ETCs in the Northern Hemisphere are: much more frequent than thought; occur most frequently in the late fall through winter months; are short-lived (on average 24 hours or less); tend to occur in particular locations of the cyclone; and can cover tens of thousands of square miles. ETCs are frequent over the Northern Hemisphere oceans and at any given time during the fall through early spring as many as five to eleven observed cyclones, at varying stages of intensity can be present over North Pacific or Atlantic. In this paper we studied averaged wind speed distribution over ETCs that reached (HF) wind strength in the North Pacific over a period of 7 cold seasons from 2001-2008 using QuikSCAT data. During this period a total of 225 cyclones with HF winds were identified and tracked in the North Pacific. December proved to be most active month with 56 separate storms reaching HF strength over the 7 year period. The average storm motion was found to be ~24 knots. Most HF events last between 6-24 hours. 50% of the 12 hour events occurred during December and 75% of the 30 hour events occurred during November and December.

Paper 5:

Tornadic Storms of 2008 over Bangladesh: Observed by Radar and Simulated by using WRF-ARW Model

Mohan Kumar Das, SAARC Meteorological Research Centre (SMRC)

Sujit Kumar Debsarma, SAARC Meteorological Research Centre (SMRC)

Someshwar Das, National Centre For Medium Range Weather Forecasting (NCMRWF)

Md. Abdul Mannan Chowdhury, Jahangirnagar University


In the present paper the tornadic storms are studied based on field survey, ground and radar observations. Low level moisture influx by southerly flow from the Bay of Bengal coupled with upper level westerly jet stream causing intense instability and shear in the wind fields triggered a series of storms for two weeks during 30 August to 14 September in 2008 . The exact time and locations of the storms are investigated by using the hourly precipitation data retrieved from a S-band radar of BMD located at Dhaka. Subsequently, the storms are simulated by using the WRF-ARW model at 2 km horizontal resolution based on 6 hourly NCEP-FNL data. The maximum intensity of surface wind speed simulated by the model was only 9 m sec-1 (32.4 km hr-1 for the 3rd September 2008 case) among the 3 cases, placing it as F0 on the Fujita scale. This seems to be highly underestimated. The highest vertical velocity (updraft) simulated by the model was 320 m sec-1 (1152 km hr-1) around 700-850 hpa, which is overestimated. The updraft reached up to 250 hPa (~ 10 km above the ground). It seems that the funnel vortex did not reach the ground, and might have passed a few hundred meters above the surface. The maximum values of SREH simulated by the model was 1088 m2 sec-2. The maximum and minimum values of the BRNSHR simulated by the model were 224 and 64 m2 sec-2, which were generally large enough to produce rotating storms according to the prescribed range. Results shows that while there are difference of 2-7 hours between the observed and simulated time of the storms, the distances between observed and simulated locations of the storms are several tens of kilometers. GrADS and RIP software were used to diagnose various parameters.

Paper 6:

Floods in eastern Romania during June and July, 2010

Meda Daniela Andrei, National Meteorological Administration

Simona Andrei, National Meteorological Administration


During June and July, 2010, Romania has experienced a period of strong atmospheric instability. The number of days with heavy rain was high and especially in the eastern regions produced historic growth of the rivers debit and levels. The floods have led to significant economic losses and casualties. During 18th June and 9th July the water amount exceeded the multiannual monthly average (in the eastern regions locally exceeded 200 mm and isolated 300 mm). This situation was caused by the persistent action of two atmospheric blocking structures - one of them acting above the Western Europe and the other over the eastern part of the continent. In these circumstances the south and the south-eastern part of the continent were under the influence of a cyclonic field which also influenced Romanian territory. The paper’s aim is to analyze the synoptic context and the convective storms that caused these disastrous floods in eastern Romania. For this study were used NCEP/NCAR re-analysis, the analysis of the limited area model ALADIN, the EUMETSAT satellite images, radar data and observation from meteorological and hydrological stations of the National Meteorological Administration and National Institute of Hydrology and Water Management. The study concludes that the atmospheric double blocking over the European continent, on both sides of the longitudinal band which Romania is situated, has favored disastrous weather conditions. During this period the most sever episodes were those in which the Eastern European ridge determined a backward trajectory of the Mediterranean cyclones arrived in South-Eastern Europe and their reactivation over the Black Sea basin. The moist air transport from the Mediterranean and Black Sea played an important role in the atmospheric instability amplification. In this period were affected Ukraine and Republic of Moldova, too. Heavy rains in these regions influenced also increasing levels of some rivers in Romania.

Paper 7:

Testing of waterspout forcasting indices over the Adriatic Sea using Aladin model

Tanja Renko, Croatian Hydrological and Meteorogical Service

Tomislav Kozaric, Croatian Hydrological and Meteorogical Service

Martina Tudor, Croatian Hydrological and Meteorogical Service


Waterspouts are severe mesoscale phenomena that can cause great damage to lives and property, especially if they enter mainland. Many waterspout events have been recorded in Croatia over the past several years, especially during the summer months. This motivated us to test and develop some forecasting tools that will help forecast waterspout events. Two forecasting indices are most often cited in the literature: the Szilagyi Waterspout Index (SWI) and KHS – Index to calculate risk of (water)spout development. For the SWI a combination of three parameters correlates strongly with waterspout events: difference between water temperature – 850 mb air temperature, convective cloud depth and 850 mb wind speed. The KHS index is based on four parameters: vertical wind shear 0-3 km, 0-500 m lapse-rate, average humidity in the first kilometer and 10 m wind speed. In this study we test these two indices on several case studies. The SWI and KHS are calculated from the Aladin model used operationally in the Meteorological and Hydrological Service of Croatia. This is a hydrostatic mesoscale model with an 8 km resolution. During the 2010 19 days with waterspouts observed along the Adriatic coast were identified. For all 19 events short description about time, location, synoptic situation (weather type and thermodynamic environment) is given as well as yes/no for the value of both indices. Counting only "yes"-cases, the KHS successfully forecasts the occurrence of waterspouts in 13 cases (hit rate of 68,4%) and the SWN in 14 cases (73,7%). Results of a detail analysis of 4 waterspout events are presented. The research shows that most of the waterspout events were thunderstorm related; however, for those that are not (fair-weather) more than just common thermodynamic instability indices like Totals-Totals Index, KI Index etc are needed. Our study shows that SWI and KHS indices improve the waterspout forecast.

Paper 8:


Aydın Erturk, Turkish State Meteorological Service

Haralambos Feidas, Aristotle University of Thessaloniki


Mediterranean Sea is one of the sources of development of small scale storms. Thunderstorms and tornados originated from Mediterranean Sea affect Turkey and Greece several times a year. They do not only destroy structures but also cause loss of life and property. The severe weather conditions also affect navigation and maritime business in this region. Geostationary meteorological satellites act like primary source for monitoring storms especially for the area which is not well covered by any ground observation system or meteorological radars. This study presents the potential of the high spectral resolution offered by the METEOSAT Second Generation (MSG) SEVIRI (Spinning Enhanced Visible and Infrared Imager) imagery in detecting storms as well as in storm detection and monitoring. In some cases, even a single SEVIRI image, either VIS or IR, are helpful in detection of the storms. In addition, multi-spectral RGB applications provide beneficial help in recognition of the thunderstorms. Looping 15 minute SEVIRI cycles also enables monitoring of the storms. In this study, we have analyzed Visible, Near Infrared, High Resolution Visible and Infrared 3.9, 10.8 and 12.0 microns channels images, RGB applications; Day Microphysics, Air Masses and Storms RGBs to detect the storms as recommended in MSG Interpretation Guide. Five different storms cases are analyzed for the time period of 2003-2010 originated from Mediterranean Sea. We found out that SEVIRI IR channels, HRVIS channel and RGB applications can be used to recognize and monitor small-scale tornados.

Paper 9:

Statistical verification of COSMO-I7 post-processing procedures and soundings for the forecast of thunderstorms over Piemonte region.

Paolo Bertolotto, ARPA Piemonte


Since 2005 in the operational centre of ARPA Piemonte some automatic procedures, based on COSMO-I7 fields (www.cosmo-model.org), have been used in forecasting thunderstorms phenomena over the Piemonte alert areas. Such procedures constitute an aid for the meteorological forecasters in suggesting thunderstorms alerts. This work is separated into two parts. In the first one we show the reliability of these procedures using standard skill scores; moreover, we underline the differences in performance among the suggestions of those automatic procedures against the actual alerts emitted by forecasters. In the second part, COSMO-I7 soundings over highly vulnerable locations are analysed and compared with the corresponding observed soundings in order to look for some discriminant parameters in forecasting thunderstorms. The aim is to achieve a further simple automatic procedure based on the forecasted soundings. This new procedure will become operational and will be verified during summer 2011. It is shown here that the best procedure for the forecast of thunderstorms reaches a POD around 60%, but false alarms are kept around 80%; instead, the operational forecast of ARPA Piemonte weather forecasters has the tendency to decrease false alarms reducing BIAS to 1, but it has as collateral effect the POD decrease to 40% around. Regarding the soundings, it appears that the most reliable parameters in discriminating storms occurrence and, at the same time, are well forecasted by the model are the K index, the Total-Totals Index, the Surface Lifted Index, the SWEAT index, the Surface Relative Elicity and the Precipitable Water.

Paper 10:

Operational forecast of thunderstorms over Piemonte region: a simulation of past cases, using different weather forecaster's output.

Paolo Bertolotto, ARPA Piemonte


The aim of this work is to analyse in detail the difference of performance and the added value of ARPA Piemonte operational tools in issuing the alerts for severe thunderstorms. In particular, we tested the direct model output of ECMWF-IFS (www.ecmwf.int), COSMO-I7 (www.cosmo-model.org) and some post-processing procedures based on COSMO-I7 outputs. First of all, ten past cases have been selected. Concerning the forecast of summer severe thunderstorms, each of them has a particular feature (harms and damages, false alarms, missed alarms); secondly, an intranet web page has been created in which the operational meteorologists of ARPA Piemonte could have a look at the principal meteorological fields related to the ten cases, fields from which the date of the event has been omitted on purpose; eventually the meteorologists simulate an operational framework, releasing the maximum QPF (quantitative precipitation forecast) in a 6 hours interval during the event and the related heavy thunderstorm warnings over one or more Piemonte areas. These forecasts have finally been collected and analysed by seeking the difference in performance using different tools (only ECMWF-IFS, ECMWF-IFS and COSMO-I7, both the models plus the post-processing procedures) and the real operational forecasts at the time in which the event has taken place. The exercise allow to realize the added value in forecasting severe thunderstorms of the local model (COSMO-I7), of the post-processing procedures and weather forecasters.

Paper 11:

IMGW project on SMS presence alert system and IMGW Monitor SH

Lukasz Legutko, Institute of Meteorology and Water Management

Aleksander Plygawko, Institute of Meteorology and Water Management


IMGW SMS presence alert system IMGW would like to present severe storm alarm system based on SMS messages. This is first in the EU alarm system for prevention before impact of storm on society and goods based on "Mobile Presence System". Presence system store localisation information of every mobile phone users in Poland. If people are moving then presence system track their mobile phone and store localisation in data base. This information is used to send alarm to users who are at storm impact area. System based on meteorological information derived from numerical models, satellite and radar information and that from telemetry network. IMGW generate alerts before storm impact and since that moment a sophisticated technology is starting to inform people in risk area about coming storm danger. The agenda of presentation: 1. General description of the system. 2. Technical issues. 3. Presence server and mobile users tracking 4. Use cases. IMGW Monitor SH IMGW would like to present IMGW SH Monitor. This is dedicated system with presentation layer based on data derived from Hydrology System. This is a product designed for crisis centers, local authorities, fire services, etc. The system provides online flood situation (data from 1200 telemetry stations) and presents hydro forecasts, based on hydro numerical models. This system is available from through Internet and can be used in field crisis management as an extension of information received from observers. The agenda of presentation: 1. General description of the system. 2. Use cases.

Paper 12:

Low CAPE, high shear tornadoes

Fulvio Stel, Arpa FVG - CRMA

Dario B. Giaiotti, Arpa FVG - CRMA


Italy is characterized by a relatively high number of tornado events even if CAPE values are not particularly high, at least if compared with other Countries. This work is devoted to analyse a tornado outbreak (two separate tornadic events almost at the same time 11 UTC, 12 LT, distant at least 50 km) that occurred during 12th December 2006 in Friuli Venezia Giulia (Italy) into an environment characterized by low CAPE (of the order of 30 J/kg) but high shear (15 m/s in 1000 m). These events (pictures have been taken) did not show any significant rotation aloft (no mesocyclones detected) neither significant electrical activity (no CG lightning detected by national lightning network and no thunders heard). The common element between these two events is the small distance between tornado occurrence and mountains (roughly 10 km). The dynamical role played by orography is analysed by way of analytical computations and numerical simulations. It is shown that forced convergence exerted by orography might supply the needed upward movement to sustain the event as well as the vorticity stretching to produce a tornado. High shear, moreover, helps to avoid hydrometeor loading, that might suppress the release of CAPE.

Paper 13:

Comparison of different COSMO Model outputs during a test case over north-western Italy.

Massimo Milelli, ARPA Piemonte

Daniele Gandini, ARPA Piemonte

Elena Oberto, ARPA Piemonte

Christian Ronchi, ARPA Piemonte

Paolo Bertolotto, ARPA Piemonte


The aim of this work is to study a severe rainfall event that happened during the 3rd and 4th of October 2010 over north-western Italy. In fact a detailed diagnostics of the event, both from the observation and the forecast point of view, is important to understand the problems of the model and it is helpful for the forecasters in the operational centre. In particular, the Italian operational limited area model (COSMO Model) failed in the description of precipitation in the pre-frontal part of the event, during the morning of the 4th. The comparison of different resolution models outputs (COSMO-I7, COSMO-I2) using a very-high resolution network of rain-gauges and also the radar rainfall estimation has been made in a subjective (eye-ball verification) and objective (statistical verification) way. Also a subjective comparison of the main atmospheric forecasted fields has been carried out against the corresponding ECMWF-IFS analysis fields. The main findings could be summarized as follows: model underestimation of sea surface temperature (SST) over the Gulf of Genoa; overestimation of mean sea level pressure (MSLP) over the Po valley; wrong wind convergence in the lower part of the atmosphere along the coastline of Liguria. These features contributed to the underestimation of the convective part of the event which is almost missing in the operational run of the model.

Paper 15:

Rainfall associated with Mesoscale Convective Systems which affected Rio Grande do Sul State in July, August and September of 2006

Luiz Carlos Bacelar, Universidade Federal de Pelotas

Cláudia Campos, Universidade Federal de Pelotas

Gustavo Rasera, Universidade Federal de Pelotas

Cristiano Eichholz, Universidade Federal de Pelotas


In Rio Grande do Sul (RS) State one of the phenomena that cause damage to society are the Mesoscale Convective Systems (MCS), which are associated with intense convective activity that causes intense rainfall, windstorms and hail. Currently MCS are studied using ForTrACC method (Forecasting and Tracking Active Convective Cells), which allows tracking MCS and analyzing the evolution of their features. Thus, the aim of this study was to identify the MCS that reached RS from July to September 2006 (JAS-2006), using ForTraCC, and evaluate the precipitation associated with these events (PMCS). GOES 12 satellite images from JAS-2004-2008, the ForTrACC method and 12 RS surface meteorological stations rainfall daily data from JAS-1977-2006 were used. It was considered only the MCS whose genesis happened southward of 20ºS, with spontaneous initiation and normal dissipation, with life cycle longer than 6 hours and that reached the RS State (latitudes from 27° to 34° S and longitudes from 49º to 58º W). The results showed that in JAS-2006: occurred dry spell in the RS; the west and north-coast State regions registered lower PMCS occurrence; the major PMCS contribution was observed in the South-eastern State region; approximately 51% of the rainfall recorded was associated with MCS and in this quarter formed less MCS, with bigger mean duration and bigger size. The factor that may have contributed to these results in JAS-2006, was the mean-intensity El Niño Modoki occurrence, that caused the quick passage of the Frontal Systems over the RS, and consequently, did not favor the organization of convection.

Paper 16:

Correlating overshooting tops and severe weather

Petra Mikus, EUMeTrain project, Meteorological and Hydrological Service, Croatia

Natasa Strelec Mahovic, Meteorological and Hydrological Service, Croatia


Overshooting convective cloud top (OT) is a dome-like protrusion above a cumulonimbus anvil, often penetrating into the lower stratosphere. It represents a very strong updraft. A single OT exists for less than 30 minutes and has a maximum diameter of ~15 km. According to some investigations, deep convective storms with OTs often produce hazardous weather conditions such as heavy rainfall, damaging winds, large hail, cloud-to-ground lightning and tornadoes. The OTs also generate gravity waves which can produce significant turbulence. These events can cause considerable property damages, influence everyday activities and even endanger the human lives. Relationship between the occurrence of the OTs and severe weather conditions over Central Europe is established. The OTs are detected from Meteosat 8 and 9 data, using a combination of brightness temperature difference of the water vapor and the infrared channel (6.2-10.8 μm), and of the ozone and the infrared channel (9.7-10.8 μm). This method includes the infrared brightness temperature and brightness temperature difference criteria. Locations and times of appearance of the OTs are compared with the occurrence of the strong wind and wind gusts measured by the automatic stations. Additionally, hailpad and raingauge measurements are used to determine the relationship between the occurrence of the OTs and severe weather conditions. Favorable synoptic conditions, i.e. dominant large-scale weather types and associated wind regimes are revealed for OT related severe weather episodes.

Paper 17:

Infrasound generation by strong convective storms

Mariam Akhalkatsi, Institute of Theoretical Physics, Ilia State University

Grigol Gogoberidze, Centre for Fusion, Space and Astrophysics, Physics Department, University of Warwick


Strong convective storms, such as supercells, are known to be the powerful sources of infrasound with peak frequencies between 0.5 and 1 Hz. Observations revealed that this radiation is strongly correlated with formation of a tornado, but it is generated substantially before (0.5 -1 hrs) tornado appearance. Detection of infrasound has significant potential for improving tornado forecasting, because the acoustic power radiated could be approximately 〖10〗^7 W and infrasound below 1 Hz can travel for great distances without significant absorption. The physical mechanism responsible for infrasound emission remained unexplained. The broad and smooth spectra of the observed infrasound radiation indicate that turbulence is one of the promising sources of the radiation. We study acoustic radiation from turbulent convection taking into account the effects of stratification, temperature fluctuations and moisture in the air. It is shown that in saturated moist air turbulence in addition to the Lighthill’s quadrupole and dipole sources of sound (related to stratification and temperature fluctuations), there exist monopole sources related to heat and mass production during the condensation of moisture. We determine the acoustic power of these monopole sources and show that radiation of a monopole source related to the nonstationary heat production is dominant for typical parameters of supercells. The results are in good qualitative agreement with the main observed infrasound characteristics e.g. total acoustic power and characteristic frequency. We give quantitative explanation of the high correlation between intensity of supercells infrasound and later tornado formation. It is shown that low lifting condensation level and high values of convective available potential energy, which are known to favor significant tornadoes, also lead to a strong enhancement of supercells infrasound radiation. This analysis indicate the potential for infrasonic detection systems to determine potentially tornadic storms and improve tornado forecast.

Paper 18:

Remote sensing obervations of severe cyclonic storm stimulated chlorophyll-a augmentation in the Bay of Bengal

Kailasam Muni Krishna, Andhra University


Tropical cyclones (also known as hurricanes or typhoons) when passing over land may have devastating effects on human lives, but over the ocean they can strongly boost another form of life - ocean primary (phytoplankton) production. Ocean primary production plays significant roles in the earth environmental system. Bay of Bengal is more vulnerable to natural hazards like tropical cyclones compared with Arabian Sea. In recent years the number of intense (category 4 or 5) tropical cyclone are increased in this region. Previous studies revealed that surface chlorophyll-a concentration increased in the stir of cyclones. This increasing biological productivity would be an expected consequence on global climate change. This study explores the upper ocean biological response to tropical cyclones in the Bay of Bengal from 2003–2010. Multi satellite and in-situ (Argo) observations are used to find out mechanism on the biological changes during the study period. I observed that strong cyclones (category 3 to 5) enrich chlorophyll-a concentration upto 15 mg/m3 due to upwelling induced vertical mixing. Vertical profiles from Argo observations also reveals the same phenomena in this region during cyclone period.

Paper 19:

Possibility of the using of “Cold Ring” indicator on satellite images in operational nowcasting/forecasting of severe weather

Inna Sobchenyuk, Ukrainian Hydrometeorological Center


Satellite images are extremely important for synoptic analysis and detection and monitoring of severe convective clouds, especially, when other measurements of atmospheric or cloud state are not available. After the launch of MSG satellite, which has significant advantages compare to other one in spectral bands, spatial resolution and time sampling, forecasters got a reliable tool for a short-range forecasting and nowcasting of severe weather. Our goal is to identify and characterize specific parameters under which one or another case of storm activity is really dangerous. A very important feature on MSG IR10.8 image is so called cold-ring and cold U/V shape storm. These features studies from early 1980’s showed a very close correlation with severe weather or supercells. However, this feature alone does not automatically classify a storm as a supercell. If observed, it indicates a possible severity of the storm, but sometimes it does not prove the severe weather. Nevertheless, due to the high correlation with severe weather, this feature has big significance in nowcasting, and should be well known to forecasters. To get more information on the cold-ring characteristics over Ukraine, some cases were selected in the years 2008 and studied more to details. All selected cases had in common that severe weather was reported. Satellite images we received with a frequency of 15 minutes. We conducted a series of observations (1 month) and try to determine under what weather patterns, or parameters, we can trace the patterns of spread of severe storms activity. Our research is very actual and timely and it has a strong practical side because it gives the opportunity to follow the laws of appearance thunderstorm clouds over the territory of Ukraine in the summer period and predict it with the greatest advance. Warnings about these phenomena will help to minimize their impact and reduce damage for such industries and sectors of the economy such as aviation, agriculture, etc.

Paper 20:

Severe Hail Climatology of Turkey

Abdullah Kahraman, Istanbul Technical University

Seyda Tilev Tanriover, Istanbul Technical University

Mikdat Kadioglu, Istanbul Technical University


A climatology of severe hail for Turkey is being constructed using newspaper records, meteorological observations, government agencies and identical sources. Since hail is a small scale event in both spatial and temporal meanings, it is usually underreported especially over less-populated areas and during night time. Non-severe hail which is not associated with important damage is also subject to underreporting. The preliminary results of the study includes over 600 records between 1950 and 2010, many of which does not have hail diameter, severity or time of the day information. Some of the records have photographs or videos enabling confirmation of the hail size, when some are depended upon eyewitnesses which are sometimes exaggerated. The database is built up using the most reliable records. More than half of severe hail cases is expressed as walnut size. The largest reliably reported size of hail is 65 mm (90 grams) which is observed in Ankara in 06.05.1953, although 300-400 grams of hailstones have been observed in southeast according to some eyewitness records published at a major national newspaper. The collected data shows that hail in Turkey usually occurs in spring and summer months. Approximately 2/3 of the hailstorms are observed during afternoon and evening hours. However, morning hours also have significant number of records. The geographical distribution is more or less homogeneous, but the Mediterranean coast, Marmara region, northeast part of the country as well as central Anatolia have in particular higher records.

Paper 21:

Enhancing tornado climatologies with Doppler radar

Mallie Toth, Purdue University

Jeff Trapp, Purdue University

Josh Wurman, Center for Severe Weather Research

Karen Kosiba, Center for Severe Weather Research


Accepted methods of estimating tornado intensity and recording event occurrences have introduced subjectivity to the existing tornado climatology. Population biases are intrinsic to the dataset. For example, areas without expansive reporting systems and well-defined procedures for post-event damage surveys may be underreporting events and intensity,thereby skewing the database towards certain regions. We are developing a means to supplement subjective methods of evaluating tornado intensity using operational Doppler radar networks. In the United States, the Weather Surveillance Radar – 1988 Doppler (WSR-88D) network provides storm-scale Doppler velocity coverage of almost the entire continental United States for ~20 years. In our approach, we relate WSR-88D data to higher-resolution, tornado-scale velocity measurements from mobile Doppler on Wheels (DOW) research radars rather than to damage. Specifically, we are creating a statistical model that will allow us to objectively estimate ground-level intensity of tornadoes using only the WSR-88D network. For each event, we determine a maximum differential velocity as measured by the DOW (ΔVDOW) and then determine the associated maximum of the WSR-88D differential velocity (ΔV88D) within a 15-minute time window. Our current regression model is based on 10 tornado cases and exhibits the following linear relationship: ΔV88D ≈ 0.4(ΔVDOW) + 19. The linear correlation coefficient for this relationship is 0.84. Work is underway to add more cases, including null (non-tornadic) events. We acknowledge that there are limitations to our simple model, which presumes the velocity data necessarily are associated with a confirmed tornado; we have begun to address the range degradation issues described by Wood and Brown (1997) as well as uncertainty related to the sampling of multiple vortex tornadoes by modeling tornadoes of various sizes and range to a radar and sampling them with a simulated WSR-88D. We anticipate that our resulting model can be applied, in principal, to various convective events worldwide.

Paper 22:

Severe Rainstorms in Brazil, cultural differences expressed in a Social Media Service may signal ways to more effectively apportion prevention and mitigation funding.

Ana Lucia Macedo, Climatempo Consultoria Ltda

Alfredo Behrens, FIA, Business School Sao Paulo

Angela Ruiz, Climatempo Consultoria Ltda


The purpose of this work is to further the understanding that culture may play in thwarting the deployment of effective prevention or mitigation of severe storm damages. Weather prediction has made an important progress in many developing countries. What mostly fails there are prevention measures and timely public announcement of severe storms; besides the deployment of measures to mitigate the damage when storms happen. Attitudes towards the capacity to face natural events and their consequences stem from a peoples´ culture. Culture is slow to change but once its role is identified the adequate amount of resources to counteract negative attitudes can be ascertained. Cultures that favor a perception that the control of people´s destiny is in their hands pay more attention to weather forecasts, build better refuges and resort to them in timely fashion. Other cultures, the fatalist prominent among them, favor perceptions according to which people´s lives are at the mercy of events beyond their control. Consequently they do not feel as guilty about the consequences of not taking prevention measures; and do not pressure authorities for them. Brazil, a country with a strong national identity is large enough to display a variety of sub-cultures, partly as a consequence of the diversity of the immigration currents that contributed to its current 190 million population stock. This work characterizes two population cultures through meta-analysis of literature and compares the different population´s attitudes to recurrent yearly rainfalls, as well as the differences in casualties that stem from similar severe storms. This work suggests that rather than apportioning prevention funds proportionate only to the probability of natural hazards, the federal government should offer greater support to areas where the people´s culture will discourage them from prevention and thus thwart the building of the readiness necessary to limit the storm´s damages.

Paper 23:

A case study of heavy convective precipitation events over the Canary Islands associated with a polar-subtropical cyclone

Olinda Carretero, AEMET

Felisa Aguado, AEMET

Francisco Martin, AEMET


Extreme rainfall intensities and torrential events from convective storms were observed at the Canary Islands on 1st-2nd February 2010. Many places, towns, and resorts areas suffered flash floods and important property damages. So 620 mm of rain was reported in San Mateo (Gran Canary Island) during the episode and 138 mm/h of rainfall intensity was measured at Las Mercedes (Tenerife Island). No injuries were reported. The organized high effective convection cells and the local complex orographic effects played an important role to characterize this hazardous situation. Everywhere evolution was driven by an intense upper level circulation, warm SST anomaly and anomalous high precipitable water environment (NAO/AO indices were negative and a moderate - high El Niño event took place, simultaneously). These elements characterized the 2009-2010 wintertime. At the beginning of February 2010, the synoptic and mesoscale frameworks were favourable for convective developments. The cloud structures evolved and were embedded in complex cyclone evolution from an extratropical cyclone up to a hybrid one, in subtropical latitudes. Initially, a cold air cyclogenesis moved from polar latitudes to subtropical ones and a cold low was formed close to Canary Islands. In the next hours, the frontal zones disappeared or weakened over subtropical warm ocean waters. Subtropical cyclone developed convective clouds around low level depression meanwhile the system slowed down. A low level warm temperature anomaly was formed and symmetry improved in the convective cloud elements. Hybrid cyclone and organized convection moved over the Canary Islands. The main goals of this case study are two fold: firstly, and looking from synoptic and mesoscale points of view, to show a complex case of the development and evolution of a cyclone from extratropical origin to hybrid ending, and secondly, to analyse and to compare the NWP models behaviour in these complex events.

Paper 24:

Remote sensing of convective storms: A-Train observations of storm tops

Martin Setvák, Czech Hydrometeorological Institute, Prague

Alois Sokol, Faculty of Mathematics, Physics and Informatics, Comenius University, Bratislava

Daniel Lindsey, NOAA/NESDIS/STAR/RAMMB, CIRA/CSU, Fort Collins, Colorado

Kristopher Bedka, Science Systems & Applications, Inc. at the NASA Langley Research Center, Hampton, Virginia

Pao Wang, Department of Atmospheric and Oceanic Sciences, University of Wisconsin-Madison

Jindřich Šťástka, Czech Hydrometeorological Institute, Prague

Zdeněk Charvát, Czech Hydrometeorological Institute, Prague


Among the remote sensing platforms used for operational monitoring of convective storms, weather satellite data play an important role in the observation of storm tops. Despite the fact that tops of convective storms do not provide direct information about storm intensity, possible severity and internal structure, much on this can be inferred from various storm-top characteristics observed in the satellite data. Though significant progress has been achieved in this area over the past two or three decades, utilizing multispectral data from various low-Earth orbit and geostationary satellites, as well as numerical modeling efforts, many observed storm-top features are still awaiting a better understanding. From this aspect, the A-Train satellite constellation offers a unique possibility to study details of storm tops, utilizing data from several satellites and their instruments. These include the Aqua satellite with its MODIS radiometer, CloudSat with its cloud profiling radar, and the CALIPSO satellite with its cloud and aerosol lidar. Given the very small time gap between these individual data sets, the storms are scanned by these different satellites at very close stages of their evolution. In the present work, we document some of the most interesting cases, addressing features such as overshooting tops, the infrared brightness temperature field with some of its unique features (cold-U/enhanced-V and cold-ring cloud-top features, coupled with embedded warm spots and areas), the jumping cirrus and above-anvil ice plumes, lower-stratospheric moisture above anvil tops (inferred from brightness temperature differences of specific bands), cloud-top microphysics, etc. The findings of this study should contribute to the improvement of conceptual models of storm tops, and may also enhance the quality of some of the operational meteorological products derived from satellite data.

Paper 25:

Evaluating the Vortex Detection And Characterization (VDAC) technique using real multiple-Doppler observations of supercell thunderstorms

Corey Potvin, National Severe Storms Laboratory

Alan Shapiro, University of Oklahoma

Michael Biggerstaff, University of Oklahoma

Joshua Wurman, Center for Severe Weather Research


The Vortex Detection and Characterization (VDAC) technique is designed to identify tornadoes, mesocyclones and other convective vortices in multiple-Doppler radar data and retrieve useful characteristics such as size, strength and translational velocity. Radial wind data from two or more radars are fit to a simple analytical model of a vortex and its near-environment. The model combines a uniform flow, linear shear flow, linear divergence flow, and modified combined Rankine vortex. The vortex and its environmental flow are allowed to translate. A cost function accounting for the discrepancy between the model and observed radial winds is evaluated over space and time so that observations can be used at the actual times and locations they were acquired. The model parameters are determined by minimizing this cost function. Tests of the technique using analytically-generated, numerically-simulated and one observed tornadic wind field were presented in Potvin et al. (2009). In the present study, an improved version of the technique is applied to additional real radar observations of tornadoes and other sub-storm scale vortices. The technique exhibits skill in detecting these vortices and characterizing their size and strength. Single-Doppler experiments suggest that, in the absence of overlapping radar coverage, the technique may still reliably detect and characterize larger (> 1-km diameter) vortices.

Paper 26:

Improving dual-Doppler retrieval of the vertical wind using a vertical vorticity constraint

Corey Potvin, National Severe Storms Laboratory

Alan Shapiro, University of Oklahoma

Ming Xue, Center for Analysis and Prediction of Storms


Dual-Doppler vertical wind retrieval is subject to significant errors in cases where near-surface regions of strong convergence or divergence are unobserved by the radars. This is particularly true in the common case where a vertical velocity boundary condition can only be safely applied at the ground. This study examines the impact of a vertical vorticity equation constraint on vertical velocity retrievals when near-surface radar observations are lacking. The analysis proceeds in a three-dimensional variational (3D-VAR) framework with the anelastic form of the vertical vorticity equation imposed along with traditional data, mass conservation and smoothness constraints. The technique is tested using emulated radial wind observations of a supercell storm simulated by the Advanced Regional Prediction System (ARPS), as well as with real dual-Doppler observations of a supercell storm that occurred in Oklahoma on 8 May 2003. Special attention is given to procedures to evaluate the vorticity tendency term, including spatially-variable advection correction and estimation of the intrinsic evolution. Volume scan times ranging from 5 min, typical of operational radar networks, down to 30 s, achievable by rapid-scan mobile radars, are considered. The vorticity constraint substantially improves the vertical velocity retrievals in our experiments, particularly for volume scan times < 2 min.

Paper 27:

Severe bow echoes in Germany

Christoph Gatzen, Meteogroup Deutschland


Climatologic findings of severe bow echoes in Germany are presented. Severe wind gusts that exceed 25 m/s are searched in a 20-year data set. 6-hourly (former years) to hourly peak wind gust measurements are available since 1989. Analogous to Burke and Schultz (2004) who investigated cold-season bow echoes in the continental United States over a period of 4 years, radar images archived every 15 minutes at a 2x2 km horizontal resolution are analysed to detect bow echoes for a 15-year period (1996-2011). From 1989 to 2011, the wind gust data combined with WMO thunderstorm observations and historical reports are used to detect severe long-lived bow echoes and derechoes in Germany. The data set is analysed to find out climatologic aspects of severe bow echoes, severe long-lived bow echoes, and derechoes in Germany: The starting time, lifetime, duration, and path of the events are analysed. The seasonal occurrence is presented as well as the preferred synoptic flow patterns. An analysis of proximity soundings is used to analyse the vertical wind shear, instability, and moisture associated with severe bow echoes in Germany. The results are then compared to the findings for cold-season bow echoes over the United States (Burke and Schultz, 2004).

Paper 28:

Identification of Convective Modes and Tracking of Storm Systems Based on Radar Data

Oliver Schlenczek, Institute for Atmospheric Physics - University of Mainz, Mainz; Institute for Meteorology and Climate Research - Troposphere Research, Karlsruhe Institute of Technology, Karlsruhe

Mark G. Lawrence, Max Planck Institute for Chemistry - Department of Atmospheric Chemistry, Mainz


There are many different types of severe convection encountered in the atmosphere, ranging from single cells and multicell storms on the milder side, to supercells and mesoscale convective systems on the more violent side. The hazardous nature of severe convection results in a strong interest from the public for scientific studies, with the goal of being able to issue warnings as far in advance as possible. Radar data are a particularly valuable tool for this, as they can be obtained with high temporal and spatial resolution. However, with radar data it is difficult to distinguish automatically between different convective modes. To improve this situation, a sample of C-band Doppler radar data from IMK, KIT-Karlsruhe, is analysed and categorised by the following criteria: Convective storms may either be isolated from each other or have an overlapping part of their reflectivity area; the latter class contains multicells and mesoscale convective systems (MCS's). Therefore, the non-isolated storms are subdivided into linear and nonlinear MCS's, where the linear MCS's have a greater length-to-width ratio than nonlinear MCS's, and multicells, which are too small to match the size criterion of an MCS. Each of the four classes is divided into two further subcategories of storms: those which contain a mesocyclone, and those without a mesocyclone. This work has started recently and the first set of prototype results for a selected set of storm observations is being analysed and will be presented. The presentation will also give an outlook to the final goal of the project, which is to develop an algorithm which allows an automatic categorisation of convective modes in order to examine large sets of radar data with a reasonable effort, allowing the community to gain insights into the spatial and temporal distribution of different convective modes, and maybe highlighting some "hot-spots", e.g., of supercells.

Paper 29:

On the Development of Large Surface Vorticity in High-Resolution Supercell Simulations

Johannes Dahl, North Carolina State University

Matthew Parker, North Carolina State University

Louis Wicker, National Severe Storms Laboratory


Tornadogenesis has been described as three-stage process, involving i) the genesis of the mid-level mesocyclone, ii) development of rotation at the surface, and iii) amplification of this vorticity to tornadic strength by convergence. This study is concerned with the second step, i.e., the mesocyclogenesis at the surface. To elucidate the processes that lead to intense vertical vorticity at the surface, high-resolution cloud-scale simulations have been carried out using the Bryan cloud model (CM1). It is found that multiple surges of horizontal momentum (which are associated with vertical vorticity at their flanks) precede the development of compact low-level circulation centers at the lowest model level. These circulations are fed by the vorticity associated with the momentum surges. Some of these circulation centers become vertically connected to the mesocyclone aloft. Although the rear-flank downdraft (RFD) may temporally coincide with some of these surges, the RFD does not cause them. The surges (and the vertical vorticity) emanate from the main downdraft northwest of the updraft. Using the ability of CM1 to calculate forward trajectories with high accuracy, we find that the parcels in the low-level circulation centers originate east of the storm, are passing through its forward flank, and gain horizontal vorticity as they traverse the eastern edge of the main downdraft. Their vorticity vectors are reoriented while descending in the downdraft, which results in vertical vorticity at the surface. Subsequently, the parcels are swept southwards and their vertical vorticity is amplified by horizontal convergence. In contrast to previous studies, all of the parcels in the low-level circulation are processed by the main downdraft and none of them originates directly from the inflow sector. Details of how the vortex lines are reorientated in the main downdraft are investigated using idealized downdraft simulations in different wind-shear regimes.

Paper 30:

Derechoes in Europe

Christoph Gatzen, Meteogroup Deutschland


Derechoes in Europe In the United States, widespread convectively induced windstorms are called derechoes (Johns and Hirt, 1987). In Europe, this term was not used to classify these events, although they did likely occur. Recently, five derechoes have been classified and were published in the scientific literature: Three in Germany, one in Finland, and one in Spain. Three of them occurred in the warm season, two in the cold season. Four were associated with strongly-forced synoptic situations and one was described as a “hybrid” case. Two derechoes affected a single country, while the others swept across more than one country. In this study, the author presents an overview of European derechoes that includes a couple of unknown derechoes and possible derechoes especially from Germany. These events were found using a 20-year data set of wind gust measurements in Germany. Additionally, historical records of possible derecho cases are included in this summary. The presented work improves the knowledge of the derecho climatology in Europe in order to achieve a better risk assessment for European convective wind storms.

Paper 31:

15 August 2010 – synoptic and aerological analysis

Tomáš Púčik, Masaryk University, Faculty of Science, Brno

Michaela Valachová, Charles University, Faculty of Mathematics and Physics, Prague

Martin Setvák, Czech Hydrometeorological Institute, Prague


On 15th August 2010, a widespread severe convective weather occurred over parts of Central and Eastern Europe. Many of these storms exhibited in satellite imagery various cloud-top features typically being associated with severe weather. This day was selected by the Convection Working Group of EUMETSAT and ESSL as a case of high importance, which will be studied from various aspects, as well as used as a test case for some of the derived (satellite-based) meteorological products. In the first part, we analyze the basic synoptic-scale setup, upper-tropospheric conditions, tropopause heights, vertical distributions of instability and wind shear for the different locations of the region where these storms formed. Besides the general description of the atmospheric conditions in the area, we seek the reasons for the different anvil-top appearance of these storms from the perspective of satellite observations. In the second part we focus at a storm which formed in the late afternoon south of Prague, hitting parts of the town early evening as a severe hailstorm, causing a total damage of about 180 millions of Euro and several injuries. Despite the storm severity and high radar echotop heights, it exhibited an unusually warm cloud top in satellite imagery, especially when compared to large severe thunderstorms elsewhere in the area. Nevertheless the storm still exhibited a cold-U feature, a well-known indicator of the possible storm severity. Thus, in this part we discuss the initiation and development of the Prague hailstorm and describe the mesoscale conditions that have contributed to its severity, especially in relation to the „atmospheric ingredients“, that were shown to play an important role in the large hail formation. Finally, we focus at the atmospheric conditions which might have influenced the specific satellite anvil top appearance of the Prague hailstorm, contrasting them with those of other storms in the area.

Paper 32:


Dhruba Banerjee, Department of Physics,Swami Vivekananda Institute of Science and Technology,Dakshin Gobindapur, Kolkata-148,West Bengal

Ramaprosad Bondyopadhaya, Mathematics Department,Jadavpur University, Kolkata-32, West Bengal


A prominent example of extreme weather event in India is Cyclonic Storm. In this paper the characteristics of distribution of tropical Cyclonic Storm (CS), Severe Cyclonic Storm(SCS), Very Severe Cyclonic Storm(VSCS) and Super Cyclonic Storm(SuCS) over Bay Of Bengal(BOB) and Arabian Sea(ARS) have been studied during the last 20 years (1990-2009).The analysis revels that the total number of cyclone has increase with high gradient (+1.67 per year) but moderate Correlation Coefficient(C.C.).Of course the number of events of different categories appears to decrease with increase of intensity (T. No.). Attempt has also been made to find out the influence of solar activity on these extreme weather events. Keeping in mind that the Sun Spot Number (SSN) is an indicator of the strength of solar effects on it. It has been found that in most of the times the high value of SSN is associated with small value of total number of cyclone. Specifically when the years of high SSN> 90 (say),are considered then Correlation Coefficient(C.C.) between SSN and number of cyclones comes out to be -0.78 while Correlation Coefficient(C.C.) of cyclones with time is +0.53.Thus it appears that sun’s spot’s influence is predominating only when it’s number (SSN) exceeds certain critical value (90).This is consistent with other analyses of these authors.

Paper 33:

Observation of tropical pre-monsoon thunderstorm by UHF Wind Profiler

Ruchith Ramakrishnan Devaki, Indian Institute Of Tropical Meteorology

Sachin Deshpande, Indian Institute Of Tropical Meteorology

Pulidundi Ernest Raj, Indian Institute Of Tropical Meteorology

Rashmi Joshi, Indian Institute Of Tropical Meteorology


Thunderstorms are interesting and important meteorological phenomena in many aspects. Although they are short lived, they can be intense with strong updrafts and downdrafts and produce heavy precipitation locally. Over the tropical Indian region, the thunderstorm is an important weather phenomenon particularly during pre-monsoon (March – May) and post-monsoon (October – November) seasons. This paper illustrates an investigation of a pre-monsoon thunderstorm that occurred over Pune (18○32’N, 73○51’E), India from observations made using a 404 MHz wind profiler (UHF radar). High resolution profiler data collected has been used to investigate the time-height variations in vertical velocities, Doppler width, reflectivity, Refractive index structure parameter Cn2, horizontal winds and wind shears in the height range 1.05 km to 4.35 km on 03 June 2008 during the period 1650 - 1744 hrs local time. Rapidly varying vertical velocities (updrafts and downdrafts), large values of vertical Doppler width and large Cn2 variations from 10-16 to 10-13 m-2/3 under clear air to heavy precipitation conditions have been observed. Also, enhanced reflectivity associated with the radar bright band/melting band is noticed. It is seen that during precipitation the UHF radar signal is dominated by the precipitation echoes as high as 20 m s-1. Kinetic energy per unit mass also showed a significant variation during the thunderstorm event. Temporal variations in surface meteorological parameters recorded at the radar site showed changes typical of the various stages of a tropical pre-monsoon thunderstorm. UHF radars/wind profilers thus provide information at high temporal and spatial resolution which is useful to study events of intense convection in detail.

Paper 34:

Tornado of 27 July 2011 in Lithuania – a case study

Inga Stankunaite, Lithuanian Hydrometeorological Service


Lithuania‘s climate is relatively mild and ranges between maritime and Continental. However, environmental conditions within Lithuanian territory vary from Severe thunderstorm with heavy rainfall and hail in the summer to heavy snowstorms in the winter. Thus, recently Lithuanian Hydrometeorological Service upgraded the weather surveillance system by installing Vaisala’s WRM200 dual polarization C-band Doppler radar in Western Lithuania near Laukuva town and lightning detection system. A second radar system is planned to be installed in Eastern Lithuania. Weather radar is most important tool in issuing warning especially during the summers like Lithuania and surrounding regions had in 2011. This summer at least 3 tornadoes were spotted over country’s territory causing property damages. Tornado that developed on 27 July 2011 in Northern Lithuania was one of the strongest tornadoes since 29 May 1981. Twister of 27 July caused not only significant damage for human property but some people were also injured during this event. Judging the fact that lorry was overthrown tornado might be at least F2 in damage intensity level. Luckily, there are no reports of people being killed. 27 July 2011 conditions for severe weather were great – Lithuania was in tropical air mass with a lot of moisture and high air temperatures. Cloud tops in Northern Lithuania during tornado event were reaching 14-15 km with cloud top temperature of -54…-58 oC while temperature near the ground was around +25...+28 oC. Convergence at low levels and divergence at higher levels were very favorable for severe storms to form. The aim of this study was to identify the synoptic and mesoscale processes leading to the formation of the tornado, thus leading to the improvement of severe weather forecasting.

Paper 35:

EOF analysis of the hailpad data for the polygon in Croatia

Damir Pocakal, Meteorological and Hydrological Service of Croatia

Zeljko Vecenaj, Department of Geophysics, University of Zagreb


Continental part of Croatia is exposed to the thunderstorms with severe rain and hail, especially during the summer months. This area is very important in terms of agriculture. Hence, in order to document hail events in details, a hailpad polygon 30×20 km in size was installed in the north-western part of this area (Zagorje) before the hailfall season in 2002. A dense network with 150 hail pads with linear spacing of 2 km (1 hailpad per 4 km2) was obtained. Hail parameters that are established from hailpads can be used in climatologically and dynamical studies of hail In this study, we use the data gathered in the hailpad polygon within the period 2002 – 2010. Three hail parameters are analyzed: the intensity (kinetic energy, KE) of hail, the maximum hailstone diameter (d) and the hailfall duration (t). The goal is to investigate what are the additional factors that have influence on spatial and temporal distributions of KE, d and t in this area (polygon). There are some indications that orography could have significant influence (Počakal et al. 2009). Our approach is to implement the empirical orthogonal functions (EOF) analysis to the data from the hailpad polygon in the above mentioned 9-years period. We will investigate how many and to what extend are the modes (components) of EOF analysis significant for particular hail parameter distribution and try to relate this primarily to the orography

Paper 36:

Flash-flood in Pula in the North Adriatic in the night of 24 to 25 September 2010

Dunja Plačko-Vršnak, Meteorological and Hydrological Service

Dunja Mazzocco Drvar, Meteorological and Hydrological Service

Martina Tudor, Meteorological and Hydrological Service


A case with the extreme precipitation in western part of Croatia during the night from 24th to 25th September 2010 is analyzed by means of measurements and NWP model data. The measurements include satellite, radar, lightning and rain-gauge data, whereas for the forecast part ECMWF and ALADIN models were used. Synoptic analysis showed a deep upper-level trough stretching from western Scandinavia to the Gulf of Genoa with the secondary trough located over Sicily. The low-level synoptic situation was characterized by the south-easterly flow at the front side of a cyclone moving eastwards through the Tyrrhenian Sea transporting warm and humid air along the eastern Adriatic coast. In the afternoon of the 24th scattered showers in the area started. As the warm and moist flow persisted, showers intensified and spread to the whole North Adriatic basin. According to the radar data, the only measurements available over the sea, the hourly precipitation amounts exceeded 50 millimeters. Around midnight of the 25th, second cyclone in the Gulf of Genoa caused the strong convergence along the west coast of Istrian Peninsula. Pronounced vertical wind-sheer induced severe convective development in the vicinity of the horn of Istria moving over the land and producing almost 200 mm of rain in less than 6 hours. In south-western Istria roads and settlements were flooded and the electricity supply was interrupted. The rain-gauge amount of 176 mm is the highest daily value ever measured in city of Pula. Operational 8km resolution hydrostatic ALADIN model forecast the existence of twin cyclones quite satisfyingly giving the correct precipitation maximum over the sea. Nevertheless, there was no indication for the secondary maximum over the Istrian land. The processes that lead to the development of the secondary maximum in the rainfall were further explored using the 2km resolution non-hydrostatic ALADIN model.

Paper 37:

Power law behaviour of atmospheric vortices

Lisa Schielicke, FU Berlin

Peter Nevir, FU Berlin


Atmospheric vortices of different size like tornadoes, hurricanes or cyclones are rotating pressure anomalies that are able to produce high wind speeds causing serious damages. Vortices are characterized by their intensity as well as by their geometric properties: Intensity is related to the size of the pressure anomaly while the geometric properties are mass-related. An atmospheric moment which is a combination of both properties has been defined in an earlier work analogue to the seismic moment which describes the size of earthquakes. In the presented work the atmospheric moment is used to study the probability density distributions of cyclones as well as the behaviour of their properties. Cyclones are tracked and analysed using the ERA-interim data set which covers a period of more than 20 years (1989-2010). Probability density distributions of the cyclones are compared to these of tornadoes and earthquakes (Gutenberg-Richter law). The behaviour of the most intense cyclones is studied in detail.

Paper 38:

Investigating historical severe storms in Austria (1604, 1807) and England (1638)

Katrin Pfeifer, University of Salzburg

Niki Pfeifer, LMU Munich


In this talk we survey selected rediscovered historical severe storms from a cultural historical perspective. Specically, we investigate a severe storm that destroyed the Parish Church at Widecombe-in-the-Moor in Dartmoor (UK) in 1638. Moreover, we report on two severe storms that uproofed houses and uprooted trees in Vienna (Austria) in 1604 and about two hundred years later in 1807. We analyze various primary sources ranging from written sources like tracts, contemporary newspaper reports, and pamphlets to pictorial sources (woodcuts and copperplate etchings). We discuss the impact of these severe storms on the society and the cultural responses.

Paper 39:

Progress and Challenges with Warn-on-Forecast

David Stensrud, NOAA/NSSL

Nusrat Yussouf, CIMMS and NOAA/NSSL

Dustan Wheatley, CIMMS and NOAA/NSSL

Jidong Gao, NOAA/NSSL


Warn-on-Forecast is a ten-year National Oceanic and Atmospheric Administration project to extend severe weather warning lead times by incorporating forecasts from a convection-permitting ensemble modeling system into the warning decision process. Initial results with convection-permitting ensembles that assimilate radar observations have shown that while reasonable thunderstorm analyses are produced, obtaining accurate very short-range thunderstorm forecasts is more challenging. Results from several warn-on-forecast data assimilation and prediction experiments for severe weather cases will be shown to highlight where progress has been made. Model error is often blamed as the dominant source in the loss of forecast skill, especially inaccuracies in the model microphysical parameterizations, although the lack of storm-scale observations also contributes to this loss of forecast skill. Observational studies indicate that the densities and intercept parameters of hydrometeor distributions can vary widely among storms and even within a single storm. Therefore, assuming a fixed set of microphysical parameters in a convection-permitting numerical model can lead to significant errors in not only the forecasts but also the analyses of severe storm events. To explore the impact of variations in parameters within the same microphysics scheme, several observing system simulation experiments (OSSEs) are conducted using a range of different realizations of the hydrometeor intercept and density parameters. Results highlight the potential for using a variety of realistic microphysical parameters across the ensemble members in improving the analyses and very short-range forecasts of severe weather events. Other challenges to ensemble design for convection-permitting resolutions will also be discussed.

Paper 40:

Severe thunderstorms and climate change

Harold Brooks, NOAA/NSSL


During the past century, the world's temperature increased unequivocally and is projected to increase more during the next century. It seems reasonable to ask the question of what impact global temperatures have had and will have on the nature of severe thunderstorms. Given the small scale of severe thunderstorms, the connection between global temperatures and storms is not direct. In this presentation, I will address issues with using reports of severe thunderstorms directly, along with some recent work on hail observations. It is also possible to consider environmental conditions in which severe thunderstorms occur and look at how they have and will change. The challenges of such an approach will be discussed, as well as the limits on what can be said. Possible avenues for future research will be laid out.

Paper 41:

A real-time radar wind analysis system for nowcast application

Qin Xu, NOAA/National Severe Storms Laboratory

Kang Nai, CIMMS, University of Oklahoma

Li Wei, CIMMS, University of Oklahoma


A real-time radar wind analysis system has been developed for monitoring low-level wind conditions at high (up to 2 km) spatial and (5-10 min) temporal resolution. The system has capabilities to integrate multi-sensor observations to produce real-time vector wind field. In additional to real-time level II data from 6 radars, Oklahoma Mesonet data and NPN (NOAA Profiler Network) data are also ingested into the system to improve the accuracy of retrieved wind field and derived products. This system can display the retrieved horizontal vector wind field at each selected vertical level or on each conical surface of radar scans synchronized and overlapped with real-time reflectivity and/or Doppler velocity fields from the radar. The real-time wind products have been displayed on NSSL/WDSS-II ORPG and made available to NWS Norman Forecast Office. The early version of the system was delivered to the Pacific Northwest National Laboratory to provide the very needed real-time radar wind retrieval capability to derive high-resolution emergency response dispersion models for homeland security applications. The system’s data quality control package was also tested and adopted at NRL Monterey for radar data assimilation and nowcast applications. The key technical elements developed in the system and the recent upgrades for extended applications will be presented at the conference.

Paper 42:

The 14 July 2010 severe MCS event over parts of central Europe

Helge Tuschy, Deutscher Wetterdienst

Marcus Beyer, Deutscher Wetterdienst


On 14 July 2010 a significant mesoscale convective system (MCS) developed over parts of central Europe, crossing all of northeastern France, Benelux and Germany before dawn. Apart from widespread damaging wind gusts, isolated large hail and at least three tornadoes were reported, causing numerous fatalities. In this poster we first highlight the different phases of the MCS: 1. A well orga- nized squall line exhibiting a sharp reflectivity gradient and a significant rear inflow jet during the mature phase over Benelux and northwest Germany. 2. The outflow dominant phase with decreasing severity over central/east Germany. Among other things we investigate how external changes (environmental synoptic set-up and daytime driven boundary layer modification) but also internal ones (strength of the rear inflow jet, temporal development of the cold pool and depth of the effective inflow layer) assisted in the transformation of the MCS. Thereafter features that are important in the operational nowcasting process will be discussed. Among others: wake low/high, the evolution of mid-altitude radial convergence (MARC) signatures, cold U-shapes, central warm spots along the anvil-layer and the mature phase of embedded bow echoes (with attendant nicks like the rear inflow jet). Subsequently we evaluate how nowcast tools were helpful in this event or how they could assist in the preparation of better forecasts. Finally the thermodynamic and kinematic environment of the line will be evaluated, leading to some interesting subitems with this event: The limited amount of large hail reports, the late occurrence of tornadic thunderstorms and the concentrated swaths of damaging winds. Also differences between the northern and southern part of the line will be addressed. The "Corfidi vector approach" will be used to explain the variable motions of the MCS. This research will be carried out with an extensive archive of surface, remote-sensing and model data (COSMO-EU), visualized by NinJo.

Paper 44:

Heat Transfer in the System Ocean-Cyclone-Atmosphere and Maximum Potential Intensity of Tropical Cyclone

Irakli Shekriladze, Georgian Technical University


A model of equilibrium translation (MET) links development of tropical cyclone (TC) to conformity of thermal and dynamical fields in the system ocean-cyclone-atmosphere. The model discloses non-dimensional alignment number claiming to be the basic criterion of TC development: N =(U.Q)/(q.R) (where U is TC translation speed; Q is average initial hurricane heat potential (HHP) of given sea area before TC entering; q is average integral heat flux (sensitive and latent) from sea surface to TC; R is TC radius). The MET considers TC as an open dissipative system internally geared to maximum intensification. Equilibrium translation takes place when this internal tendency is in tune with large-scale environmental wind. At that TC rapidly intensifies at roughly constant N (alignment effect). Another important peculiarity is reverse dependence between HHP and equilibrium translation speed. The paper presents the results of analysis of the effect of aforementioned reverse dependence on sea surface temperature (SST) and maximum potential intensity (MPI) of TC. Negative feedback between HHP and equilibrium translation speed causes important peculiarity of equilibrium translation: the higher HHP, the slower equilibrium translation, other things being the same. Correspondingly, TC results less intensive cooling of given sea area at low HHPs and, vice-versa, the cooling is much intensive at high HHPs (total cooling effect directly depends on duration of TC translation above given sea area). This circumstance allows to assuming that average SST under TC is roughly uniform during equilibrium translation, irrespective of initial HHP and SST fields. On the other hand, just equilibrium translation is precondition for establishment of MPI. As follows from further analysis, at sufficiently high Q, constancy of average SST results constancy of MPI irrespective of initial HHP and SST fields. This main outcome of the paper is confirmed by wide field data at Q > 50 kJ.cm-2.

Paper 45:

Multi-model simulation of the F4 Hautmont tornado

Valerian Jewtoukoff, Laboratoire de Meteorologie Dynamique ENS


A F4 tornado occurred on the day of 2008 August 3rd in Northern France, and produced severe damages to the small town of Hautmont and its surroundings. As previously mentioned in the litterature, the Météo France operational models suite was not able to predict this tornado. Tornadoes are relatively rare in France compared to other natural hazards such as flash floods, and no tornado watch was issued. The synoptic environment of the storm was characterized by a relative low pressure on the North of Scotland and an undulating front extending from the Atlantic ocean to the Netherlands across Northern France. The region of interest was located right under the diffluence zone preceding the 500 mb jet streak. This zone was associated to strong upper level divergence favorable to synoptic scale vertical forcing. On the mesoscale, one of the cells from the storm system started to evolve rapidly into a severe thunderstorm. Rainfall radar images show evidences of supercellular structures. A remarkable feature is the fact that this tornado occurred in a high-shear low-CAPE (~500 J.kg-1) environment. A previous paper focused on the analysis of convective composite parameters and these proved to be good tools to identify the tornadic region during that case, but is it possible to explicitely simulate this case? WRF-ARW simulations initialized with the NCEP FNL analysis have been performed to investigate if the model is able to discriminate the supercellular thunderstorms and capture the position and timing associated with the event. ARPS idealized high-resolution simulations were also initialized with a sounding extracted from the WRF analysis in order to study the increase in low-level vorticity. Results of those simulations are presented and compared to the satellite and radar-derived mesoscale situation.

Paper 46:

The “Consorcio de Compensación de Seguros” and the wind storm insurance coverage in Spain

Alejandro Izuzquiza, Consorcio de Compensación de Seguros


The Consorcio de Compensación de Seguros (CCS) is a public business institution under the authority of the Ministry of Economy and Finance. One of its main duties entails compensating damage to insured persons and goods as consequence of the “extraordinary risks”, which include natural catastrophes and some violent human actions (terrorism, etc.). Among the natural catastrophes covered by CCS we can find wind storms with gusts above 135 km/h and tornadoes. The storm Klaus (January 2009) caused more than 275,000 claims to be assumed by CCS, which had to pay indemnifications in an amount close to € 500 million. As consequence of the storms Floora and Xynthia (January and February 2010) the CCS faced more than 80,000 claims, with a total cost of €100 million. In Spain, this extraordinary risk coverage must be compulsorily attached to every ordinary policy issued by the private insurance companies covering damages in property lines, or accidents and death in the case of personal insurance lines. The premium receipt of this ordinary policy must include a surcharge to be deposited by companies in the CCS. The permanent nature of the CCS, the wide scope of its natural catastrophe covers and its experience and capacity to manage them –on its own or through ways of collaboration with the insurance industry- entail that the CCS is considered by policyholders and insurance companies as a necessary and efficient complement which gives stability to the Spanish insurance system.

Paper 47:

Convective cell identification and nowcasting using multi-source data

Anna Jurczyk, Institute of Meteorology and Water Management

Jan Szturc, Institute of Meteorology and Water Management

Katarzyna Osrodka, Institute of Meteorology and Water Management


The presented method is being developed for precipitation nowcasting using multi-source data with particular emphasis on severe weather convection-related phenomena. The basic assumption of the proposed technique is to consider each convective cell as a separate object. Before the nowcasting the cells are identified by means of the following algorithms: (i) separation between convective and stratiform areas, (ii) determination of the cell centres, (iii) clustering of convective areas into particular cells. The differentiation between convective and stratiform precipitation is based on multi-source data from: weather radar, satellite, lightning detection, and numerical weather prediction model by means of employing fuzzy logic approach. For identified convective area cell centres are determined through analysis of radar reflectivity field, especially its pattern around local maxima. Clustering of convective radar pixels into cells is performed applying geometrical approach which takes account of distance to the centres and pattern of reflectivity field along the related paths. At this stage of research the overall assumption for the nowcasting is to predict advection of precipitation without considering evolution of rain intensity. Extrapolation approach is employed for both stratiform and convective areas, but in different way. For stratiform precipitation object-oriented extrapolation is performed like in TREC algorithm, whereas for convection area displacement vector is determined for each convective cell separately. It is performed by coupling of cells in subsequent time-steps and then calculation of their movement vectors from the cell centroid positions. Merging and splitting of the cells is considered as well. For selected convective events examples of nowcasts generated employing proposed algorithm will be presented in comparison with both forecasts: persistent and derived from the extrapolation algorithm without consideration of convection.

Paper 49:

Possible Impacts of Climate Change on Snowfall and Snow Depth over Northern Canada under Downscaled Future Climate Conditions

Chad Shouquan Cheng, Atmospheric Science and Applications Unit, Meteorological Service of Canada Branch, Environment Canada

Baoling Wang, Adaptation & Impacts Research Section, Atmospheric Science & Technology Directorate, Environment Canada


The overarching purpose of this study was to project changes in the occurrence frequency and intensity of future daily snowfall events and daily snow depth under downscaled future climate conditions over Northern Canada. The 55-year historical data, including hourly/daily meteorological observations, were used to develop daily snowfall and snow depth simulation models. In addition, nine GCM outputs with two IPCC AR4 scenarios (A2 and B1) were used in the study for future two-time windows (2046–65, 2081–2100). The GCM historical runs (1961–2000) were also used to correct the GCM biases. Different regression methods were used to construct downscaling transfer functions for different meteorological variables. Using downscaled GCM meteorological variables, automated synoptic weather typing integrated with cumulative logit and non-linear regression analyses was applied to project future daily snowfall amounts and daily snow depths. Downscaling transfer functions and snowfall/snow depth simulation models were validated using a cross-validation scheme and comparing data distributions and extremes derived from downscaled GCM historical runs and observations over a comparative time period 1961–2000. The results showed that the models for all variables used in the study performed well. This talk will introduce the research project and outline the modeling exercise and verification process. The major findings on future snowfall/snow depth projections from the study will be summarized in the presentation as well. One of the major conclusions from the study is that the methods used in the study are useful for climate change impact analysis on future snowfall and snow depth.

Paper 49:

Possible Impacts of Climate Change on Snowfall and Snow Depth over Northern Canada under Downscaled Future Climate Conditions

Chad Shouquan Cheng, Atmospheric Science and Applications Unit, Meteorological Service of Canada Branch, Environment Canada

Baoling Wang, Adaptation & Impacts Research Section, Atmospheric Science & Technology Directorate, Environment Canada


The overarching purpose of this study was to project changes in the occurrence frequency and intensity of future daily snowfall events and daily snow depth under downscaled future climate conditions over Northern Canada. The 55-year historical data, including hourly/daily meteorological observations, were used to develop daily snowfall and snow depth simulation models. In addition, nine GCM outputs with two IPCC AR4 scenarios (A2 and B1) were used in the study for future two-time windows (2046–65, 2081–2100). The GCM historical runs (1961–2000) were also used to correct the GCM biases. Different regression methods were used to construct downscaling transfer functions for different meteorological variables. Using downscaled GCM meteorological variables, automated synoptic weather typing integrated with cumulative logit and non-linear regression analyses was applied to project future daily snowfall amounts and daily snow depths. Downscaling transfer functions and snowfall/snow depth simulation models were validated using a cross-validation scheme and comparing data distributions and extremes derived from downscaled GCM historical runs and observations over a comparative time period 1961–2000. The results showed that the models for all variables used in the study performed well. This talk will introduce the research project and outline the modeling exercise and verification process. The major findings on future snowfall/snow depth projections from the study will be summarized in the presentation as well. One of the major conclusions from the study is that the methods used in the study are useful for climate change impact analysis on future snowfall and snow depth.

Paper 50:

Evaluation of lightning data from different detection networks over the Czech Republic

Petr Novák, Czech Hydrometeorlogical institute


The Czech Hydrometeorological Institute (CHMI) utilizes lightning data from the Central European Lightning Detection Network (CELDN) since May 1999 in free experimental mode and since 2002 fully operationally on commercial basis. The CELDN is operated by Siemens AG. It was created by combining LF sensors from German BLIDS network, Austrian ALDIS network and several new sensors installed in Czech Republic, Poland, Slovakia and Hungary. CELDN is primarily targeted on detection of cloud-to-ground (CG) lightning but intra-cloud (IC) lightning detection is also available. In the CHMI, lightning detection is primarily used for the nowcasting of severe storms mainly in combination with other remote sensing data. Climatology of lightning based on the 2002-2008 period showed good quality of CG lightning that fulfilled requirements of the CHMI forecasters and employed nowcasting systems. Quality of IC lightning was worse and spatially inhomogeneous but the first results from 2009 showed improvement IC lightning, most probably as a result of lightning sensor upgrades. During several last years, German company Nowcast GmbH extended their lightning detection network LINET also over the Czech Republic territory. LINET sensors works in similar frequency band as CELDN sensors and their data are physically comparable. Density of detection sensors over the Czech Republic territory is also similar for both networks. The CHMI received full access to the LINET data from 2010 convective season for evaluation purposes. The paper presents updated climatological characteristics of lightning over the Czech Republic territory based on the 2002-2010 CELDN dataset. Changes in lightning data quality after CELDN sensors upgrades are discussed. Comparison of CELDN and LINET data for 2010 is also shown.

Paper 51:

Mesoanticyclones and anticyclonic tornadoes in supercells

Howard Bluestein, University of Oklahoma

Jeffrey Snyder, University of Oklahoma

Michael French, University of Oklahoma


Anticyclonic tornadoes are relatively rare occurrences in supercells, while mesoanticyclones are ubiquitous. When anticyclonic tornadoes have been documented in right-moving supercells, many form a vortex couplet with cyclonic tornadoes (or what remains of their circulation) along the rear-flank gust front. Mesoanticyclones have been documented in midlevels in several ways: (1) as part of the couplet produced by the tilting of horizontal vorticity associated with deep-layer shear by the main updraft; (2) as part of a couplet along the rear-flank gust front, or above it; and (3) on the left-front flank as “wake vortices;” Some of the second appear in association with the “Owl horn” echo (Kramar et al. 2005). In this presentation we will present more evidence from mobile, X-band, Doppler-radar data of mesoanticyclones and anticyclonic tornadoes in supercells, including many from VORTEX2, using data from the U. Mass. X-Pol radar and the Naval Postgraduate School, rapid-scanning, MWR-05XP (a mobile, X-band, phased-array radar). The short-term evolution of intense anticyclonic vortices near the ground, based on data from the MWR-05XP, will be described. The anticyclonic vortices developed at low elevation first, decayed first at low elevation, and were confined to the lowest 3- 4 km. It is suggested that anticyclonic tornadoes in right-moving supercells are like landspouts, in that the vorticity of the mesoanticyclone is stretched underneath updrafts in growing convective towers along the rear-flank gust front.

Paper 52:

Tornado dynamics (invited oral presenation)

Howard Bluestein, University of Oklahoma


In the past decade there have been numerous studies of tornadogenesis and tornado structure, especially in supercells. They have included the analyses of Doppler radar data on both the storm scale and tornado scale, measurements with in situ instruments near the ground in and near tornadoes, numerical simulations of tornado-like vortices in supercells, and LES simulations of intense columnar vortices in contact with the ground. In this presentation, some of the fundamental issues associated with tornadogenesis and tornado structure, with an emphasis on what we know and what we don’t know, will be summarized. This talk will conclude with some thoughts on where research on tornado dynamics will likely be headed.

Paper 53:

Comparison and validation of satellite-based overshooting top detection methods

Petra Mikus, EUMeTrain project, Meteorological and Hydrological Service, Croatia

Kristopher Bedka, Science Systems and Applications, Inc. at the NASA Langley Research Center

Natasa Strelec Mahovic, Meteorological and Hydrological Service, Croatia


Overshooting convective cloud tops (OT), dome-like protrusions above a cumulonimbus anvil, indicate the location of vigorous updrafts where severe weather including heavy rainfall, hail, strong winds and tornadoes most often occurs. OTs are apparent in satellite data with a lumpy textured appearance in visible channel imagery. Since visible channel imagery is only available during daytime, meteorologists must also use the infrared (IR) channels for observing and objective detection of OTs. The aim of this investigation is to compare the detection characteristics and relative accuracy of several different satellite–based OT detection methods. Research by the authors has shown that OTs can be detected from the satellite data using the brightness temperature difference (BTD) of the water vapor and the IR channel (6.2–10.8 μm), BTD of the ozone and IR channel (9.7–10.8 μm) or BTD of carbon dioxide and IR channel (13.4–10.8 μm). A combination of the 6.2-10.8 μm and 9.7-10.8 μm BTDs can also be used. All these methods include thresholds for both IR brightness temperature and the BTD. A more complex method, called IRW-texture, includes a combination of infrared channel brightness temperature and spatial gradient criteria with a numerical weather prediction model tropopause temperature forecast to detect OT signatures at their characteristic spatial scale. The theoretical background of all mentioned methods will be explained and the results of detection will be compared with the High Resolution Visible (HRV) satellite images during day-time in order to validate each method.

Paper 54:

Frequency of supercells over the western part of Serbia in period from 2005. to 2009.

Miodrag Stojanovic, Sector for hail suppression of the Hydrometeorological Service


Western part of Serbia is hilly-mountainous. Because of its geographic position it is exposed to supercells that come from Bosnia and it is a good ground for developing a new ones. Supercells are very important because they usually make flashfloods and large scale destructions from hail. In this study frequency of supercells will be estimated for the 5 year period (2005. - 2009.). Data from weather radar Mitsubishi RC-34A, along with the ground observations and upper air soundings were used. Because Mitsubishi RC-34A is not Doppler radar other criteria for the supercell estimations were set (such as spatial dimension, maximal reflectivity, height of the maximal reflectivity, height of the accumulation zone, height of the cloud top and duration of the cloud life).

Paper 55:

Characteristics of misocyclones observed on Tosa Bay in Japan

Koji Sassa, Kochi University

Ippei Hamada, Kochi University

Yoshiki Hamaguchi, Kochi University

Taiichi Hayashi, Disaster Prevention Research Institute, Kyoto University


Tosa Bay located at south side of Shikoku island in Japan is known to be area that tornadoes frequently outbreak in the world. Its frequency is 56 in the area of 10000 square km per year, which is 19 times larger than the value in Oklahoma in United States. Our observational research aims to clarify the characteristics of tornadic and nontornadic misocyclones. We observed misocyclones by using two X-band Doppler radars and a C-band Doppler radar from 2008 to 2010 and detected 25 misocylones for three years. Most of them were generated offshore and got on shore. Four misocyclones developed to be tornadoes and caused damages from F0 to F1 in Fujita scale. The diameters and vorticities of the misocyclones were measured from Doppler Radar data. They were nearly distributed on a hyperbolic line, which means their circulations were in almost the same value. But the plots of four tornadoes located at the upper side of the hyperbolic line.

Paper 56:

A case stady of severe hailstorm over central part of Serbia on July 08nd 2009

Ruzica Radovanovic-Bulic, Republic Hidrometeorological Service of Serbia

Svetlana Biljic, Republic Hidrometeorological Service of Serbia

Ivan Bulic, Republic Hidrometeorological Service of Serbia


The area of the western part of the Balkan Peninsula is the area with frequent convective cloudiness in the warm half of the year. The Alps mountain ranges have a multiple influence within the meaning of a partial blocking of cold air followed by its downward movement further towards the east. This work deals with an extreme natural disaster that came down on a part of Serbia on July 8, 2009. The storm was accompanied by significant precipitation, hail of all sizes, 1 to 5 cm in diameter, and strong winds reaching even 23.1 m/s. The 12 UTC sounding showed three temperature inversions in the 900 - 460 mb layer which is indicative of frontal zones in this area. The passage of the cold front originating in the west, over the territory of the Western Balkans gave development of deep convections on the Serbian territory. In the frontal line and before the front, five very deep convective cells emerged that were monitored through the hail suppression system. One of these cells was a supercell. The supercell travelled round 200 km moving in a WNW-ESE orientation at the speed of approximately 50 km/h, accompanied all the way by heavy rain and hail, with the size of hailstones ranging from a hazelnut to hens’ egg, and strong winds (19.4 m/s). The strength of the maximum radar reflectivity reached 73 dBz at the altitude of 10.3 km, Cloud top hight was reaching 16 km. The remaining four cells travelled round 150 km, moving in parallel with the supercell as accompanied by rain and strong winds reaching the speed of 23.13 m/s. The work’s objective is to show synoptic situation, the route of these five cells, their radar features, as well as their vertical cuts, satellite images and their manifestation on the ground.

Paper 57:

Simultaneous BTD (WV6.2-IR10.8) anomaly and above-anvil ice-plume observed above the storm of 06 July 2010, North Italy

Mária Putsay, Hungarian Meteorological Service

Martin Setvák, Czech Hydrometeorological Institute

André Simon, Hungarian Meteorological Service

Jochen Kerkmann, EUMETSAT


The study focuses at a convective storm, which occurred late afternoon on 06 July 2010 above Adriatic coast of the North Italy. The storm was interesting by its cloud-top appearance, exhibiting a well-defined cold-U shape in the IR bands, and a distinct above-anvil ice-plume; both well documented in the MSG/SEVIRI data. However, the most interesting feature of this case was the brightness temperature difference, BTD(WV6.2 IR10.8) anomaly, documented in the MSG rapid scan 5-minute data. This way defined, the BTD is usually positive above cold storm anvils, which has been documented in several previous studies. This effect can be attributed either to a warm water vapor (lower-stratospheric moisture, LSM) above tops of convective storms, or to scattering effects atop storm anvils. The first effect (LSM) seems to play a more significant role, and most of the positive BTD observations are thus attributed to the warmer LSM. The positive BTD can be either closely correlated to the IR brightness temperature, or exhibit various local “BTD anomalies”. The first case relates to a semi-homogenous LSM layer being advected from elsewhere, while the second case represents a manifestation of possible local source of LSM – the storm itself. While other recent studies focus at statistical characteristics of the positive BTD, we focus at a distinct BTD anomaly which developed at the top of the above mentioned storm. While a similar BTD anomaly has been already described earlier, this case is unique by a simultaneous occurrence of the BTD anomaly and a nearby ice-plume. In this case the BTD anomaly for a short period of time resembles a distinct plume-like feature, oriented similarly as the ice-plume, while for most of its existence it has rather irregular oval outline. We discuss the significance of the co-existence of both features, as well as details of the BTD evolution.

Paper 58:

Verification of the forecasts of heavy convective rainfalls produced by COSMO model in a high temporal resolution

Vojtech Bliznak, Institute of Atmospheric Physics AS CR

Petr Zacharov, Institute of Atmospheric Physics AS CR


The contribution will be focused on verification of few convective events occurred between 23 June and 5 July 2009 on the territory of the Czech Republic and produced heavy rainfalls. Radar reflectivities measured by Czech radar network CZRAD (radar Brdy and Skalky) every 5 minutes were converted into rainfall intensities using standard Z-R relationship and integrated in time every 15 minutes. The adjustment coefficient that was applied during every time step was obtained by comparing integrated 15min rainfall estimates within 1 hour with hourly adjusted radar-derived rainfall estimates. The 15min precipitation totals were predicted by non-hydrostatic NWP model COSMO, which is integrated with a horizontal resolution of 2.8 km. The radar reflectivity was assimilated into the model run. The extrapolated radar reflectivity was assimilated into the model as well and the extrapolation method is based on the COTREC algorithm. The verification was performed for the first three hours of forecasting and the ends of the assimilation window were in 12-16 UTC. On the whole, 45 forecasts from 9 days of convective precipitation were evaluated for each hour. The evaluation included subjective verification and the objective verification scores Fractions Skill Score and SAL. The aim of our study is to: (1) prepare the gauge-adjusted 15min radar precipitation totals; (2) assess how accurately the model is able to simulate the 15min totals.

Paper 59:

Very short range severe weather warning based on satellite and model data in South Africa

Estelle De Coning, South African Weather Service


To issue warnings of pending severe weather in the time frame two to twelve hours is just one of the many and challenging tasks of operational forecasters. Combining all the instability indices, available moisture, terrain features and wind shear parameters cognitively, is not always easy. Beyond the scope of Doppler radar – which is not commonly available in southern Africa - the only tools remaining to do this are satellite and model data. In South Africa a satellite based Combined Instability Index was developed which combines instability, moisture as well a height above sea level into a probabilistic map for convective activity with at least three hours lead time. This product is available operationally and provides good guidance to forecasters. The next step was to add wind shear into the equation to say something about severe weather probabilities. Such an index was developed and tested for several cases of severe convection over South Africa. Evaluation of the new index was done through observations of large hail and/or wind damage, lightning from the South African Lightning Detection network, radar data as well as severe storm signatures from Meteosat Second Generation imagery. Results of preliminary test will be shared to show the potential value of such an index especially in the southern African region where radar systems are rare.

Paper 60:

Probabilistic Forecast of Severe Weather Events: The long way to the mesoscale

Jose A. Garcia-Moya, Agencia Estatal de Meteorologia - AEMET


Atmospheric predictability depends on forecast length and it is flow dependent as well. Extreme daily temperatures can be forecast up to several days in advance. Convective precipitation events cannot. Probabilistic forecast may be useful to increase the capability to forecast severe weather events even in the short-range. It is widely accepted that forecast risk is the goal for both Medium- and, also, Short-Range Prediction. Additionally, information about the uncertainty of the forecast is also needed. On the other hand numerical weather prediction models increase both horizontal and vertical resolutions to become true mesoscale models. They try to improve the representation of mesoscale weather events (flash foods, wind gales, etc), mainly because forecasting extreme events has an increased interest of the general public. Ensemble Prediction Systems (EPS) are the best tool to give information about uncertainty and probabilistic forecast but they are very expensive in terms of computer resources. Multi-model ensemble prediction systems are showing to be a very good approach to account for model errors but we also need appropriate techniques to perturb initial and boundary conditions. The number of members of the EPS is also a way to improve results and to increase the resolution of the Probability Density Function (PDF) but adequate computer resources are required as well. In the presentation results from the Multimodel Multi-boundaries SREPS (Short-Range Ensemble Prediction System) developed at AEMET are shown. Special attention is paid to surface parameters and, among them, precipitation. SREPS is compared with ECMWF EPS which is focused on the Medium-Range forecast. The evolution of SREPS is GLAMEPS (Grand Limited Area Model Ensemble Prediction System) developed by the Hirlam-ALADIN Consortia. It will become operational at the end of 2011. Results from GLAMEPS (10 Km resolution) are compared with SREPS (25 Km resolution) and ECMWF EPS (32 Km resolution) showing that resolution matters. Finally, characteristics of the new AEMET Convective-permitting scale EPS will be mention. Research project is starting this year and the system will become operational in 2015 once a new computer will be installed at AEMET.

Paper 61:

Planet Boundary layer dynamics involved in northern Italy misocyclonic and mesocyclonic waterspouts as well as in northern and central Adriatic Sea

Angelo Bertozzi, SMS METEOCENTER

Pierluigi Randi, SMS METEOCENTER


The occurrence in central and northern Adriatic of misocyclonic waterspouts is usually associated with mainly 2 different types of mesoscale circulation patterns: - Prefrontal or frontal Apenninic-generated north-easterly moving drylines. In this case misocyclons tend to trigger on convergence boundary between pre-existing flows from SE or ESE and orographic currents from SW sometimes caused by middle troposhere critical level sudden formation; this kind of convective trigger represent a very small amount of total occurred events. - Cold front, dry-line or boundary outflow of shouthward moving convective systems on the Triveneto - northern Adriatic area. whith this conditions misocyclons get triggered mainly along the boundary that separates preexisting currents from SW or SE and northern flows from below which often follow a orographic component such as Bora winds. In this case, the trigger of misocyclons and waterspouts is much more frequent due to the higher availability of low level turbolent kinetic energy, PBL discontinuity and elevated rotational component of motion in north-east flows that most likely induce PBL ondulations and amplify low-mid level SREH and vorticity. This paper aims to analyze the dynamics of the above, especially those related to the role played by "Dinaric Alps" in conjunction with north-east flows. The analysis was conducted by reforecasting events of "severe weather" through high resolution operational models (Parent domain Europe at 7.5 km - Central Mediterranean nest at 2.5 km) forced with boundary conditions obtained from reanalysis datasets (ERA interim CFS reanalysis, MERRA ) The severe weather events were detailed with internal telescopic nests at very high final resolution (200 meters around) and 3D rendered by specific visualization software (Vapor)

Paper 62:

Summertime northern Italy outbreaks, description of "Adriatic slot" circulation dynamics and their role in convective trigger and initiation of miso and mesocyclons. 3D rendering of "Adriatic corkscrew" circulation.

Angelo Bertozzi, SMS METEOCENTER

Pierluigi Randi, MeteoCenter


The present work aims to make a detailed 3 dimensional description of summertime circulation patterns prevailing on north and central Adriatic context in case of Eastward moving and southward moving outbreaks. By 3D renderings obtained through high resolution re-forecasting simulations it may be possible to depict the role of "bora pools" induced PBL discontinuities and west Adriatic nearshore hydraulic jumps in misocyclon initiation and development during southward moving cold front. A similar depiction is made to highlight the low level shear and deep layer shear dynamics role in convective overturing in Noth-Estern PO-Valley during Eastward moving outbreaks. It will also be described with 3D rendering a sort of "corkscrew-like" adriatic circulation pattern, a kind of reproducible pattern retrievable in the central-northern Adriatic during summer outbreaks consisting in a widely extended clockwise rotor shear vortex occupying large part of northern-cental Adriatic slot area This circulation pattern in its early structuring stage develop very favourable mesofrontogenesis conditions.

Paper 63:

Study of some stationary convection cases in the Valencian Community

Juan Simarro, AEMET

José Angel Núñez Mora, AEMET


Valencian Community, as other coastal areas of the Mediterranean Sea, suffers heavy rain episodes more frequent in the autumn season. In this poster we expose a study of flash flood episodes in the Valencian Community from the point of view of the short range forecast and nowcasting. We have selected a small number of not common cases of stationary convection which produced floods in relatively small areas. Stationary convection is potentially dangerous even in non extreme weather environments because as the convective systems do not move or move slowly great amounts of precipitation can be accumulated. The episodes are studied using standard conceptual models and tools. A review of the meteorological literature on this issue is provided. A high resolution numerical model is used to simulate the rain processes and results are compared to observations taken from rain gauges, meteorological radar data and satellite images. Some questions about the difficulty of the forecast of this type of situations are drawn.

Paper 64:

Slantwise circulations and convection in pre-frontal environment over central Italy: a numerical study

Maurizio Fantini, ISAC-CNR

Piero Malguzzi, ISAC-CNR

Andrea Buzzi, ISAC-CNR


The synoptic situation over Central Italy on 30 October 2008 showed a strong south-westerly wind associated with a synoptic-scale trough and cold front passage. Bands of precipitation aligned with the wind developed well ahead of the front, combined with individual convective cells. Severe surface winds and pronounced wave trains in the lee of the Apennines were also observed. The above features were successfully simulated with the non-hydrostatic model MOLOCH at 0.9 km resolution. Idealized experiments without orography were then performed to isolate and diagnose the role of instabilities in the observed development, and the dynamical properties associated with the co-existence of conditional and symmetric instability, and a ‘Delta-M-adjustment’ event.

Paper 65:

Lightning behaviour during the lifetime of severe thunderstorms

Tsvetelina Dimitrova, Agency Hail Suppression

Rumjana Mitzeva, Faculty of Physics, University of Sofia

Hans D. Betz, University of Munich, Department of Physics;Nowcast GmbH

Hristo Zhelev, Agency Hail Suppression; Faculty of Physics, University of Sofia

Sebastian Diebel, Nowcast GmbH


The present work is directed to study the evolution of lightning characteristics in different severe thunderstorms producing hail. The thunderstorms are frequent events from April to September over Bulgaria. In more than 60% of the days there are thunderstorms and half of them produce hail. The number of severe thunderstorms judged by hail size greater than 2 cm varies between 1-5 cases per year. Lightning and radar data concerning the three severe thunderstorms developed over Bulgaria during the summer of 2010 are analysed. Lightning data are taken from the LINET network. Radar information is obtained by S-band Doppler radar MRL5-IRIS observations from Hail Suppression Agency in Bulgaria. The flash rate (FR) is calculated per 4 minutes in accordance with the period of radar volume scan. The two of the thunderstorms are multicellular during their lifetime while the third one is transformed from a multicell into a supercell storm. There is significant difference in the duration of large hail falling on the ground from the tree thunderstorms. The combined analysis of lightning characteristics (FR, multiplicity, polarity and peak current) and radar parameters is carried out. The results reveal that along with the similarities in the behavior of lightning during the lifetime of severe storms there are differences. In the three thunderstorms there is a flash and multiplicity jump before the time of the detection of damaging hail on the ground. However the lag-time is different - between 10 and 30 minutes. The jump is best pronounced in the thunderstorm which transformed into a supercell storm. The significant difference in the ratio of positive to negative flashes is established in this thunderstorm during multicell and supercell periods. In the three thunderstorms positive flashes are detected however their percentage is different. The temporal evolution of lightning and VIL (vertical integrated water) is also examined.

Paper 66:


Jindřich Šťástka, Czech Hydrometeorological Institute, Prague

Michaela Radová, Czech Hydrometeorological Institute, Prague

Martin Setvák, Czech Hydrometeorological Institute, Prague


Positive values of brightness temperature difference (BTD) between water vapor and infrared (IR) window bands observed above cloud tops of convective storms were described and discussed in several previous studies. Presence of water vapor in the warmer lower stratosphere is assumed to be the most likely explanation of this feature. Some of the studies suggest that the storms themselves inject the moisture into the lower stratosphere, and that this happens not only in vigorous tropical convection but also in mid-latitude storms penetrating the tropopause. Observed positive BTDs above parts of a storm's anvil could also be explained by transparency and scattering effects. In general, the processes occurring at and above cloud tops of longer-lived or severe storms are still far from being well-understood, which makes the BTD interpretation quite difficult. The present study focuses on the topic of positive BTD values using mainly the MSG Rapid Scan Service 5-minute data over Europe from 2008 - 2010. The emphasis of this study is on the time evolution of BTDs (in relation to the storm's life cycle) and the relationship between BTD and brightness temperature in the IR window band. Other data, such as soundings, A-Train satellite data (when available), etc., are used to obtain the best possible information about each case. Our results should contribute to a better general understanding of the processes at the upper levels of storms and an improvement of conceptual models of storms. Also, more precise interpretation of the BTD may improve the utilization of this parameter, which is commonly used by various cloud classification and precipitation detection algorithms.

Paper 67:

Climate model simulations of the large-scale conditions associated with the genesis of Medicanes

Javier Martin, UIB

Romualdo Romero, UIB

Maria Tous, UIB


Medicanes (Mediterranean tropical-like storms) are violent windstorms that, once developed over the sea, have the potential to affect islands and coastal regions. This phenomenon operates on the thermodynamical disequilibrium between the sea and the atmosphere like tropical cyclones. This analogy is confirmed through their visual appearance in satellite images: axisymmetric cloud structures with a convective wall around a relatively cloud-free central eye. A genesis probability index “GENpdf” has been formulated in the literature which has been succesfully tested for the observed genesis of tropical storms. It depends on the thermodynamic contrast between the sea surface and overlying air, the low-trophospheric vorticity, mid-tropospheric relative humidity and the deep-layer wind shear. Large values of this index are revealed as a necessary -although not sufficient- diagnostic indicator of medicane producing synoptic environments after analysing twelve different cases. The present study attempts to analyse the changes in frequency and intensity of these medicane potential environments imposed by global warming. Specifically, the ERA-40 reanalysis for the period 1981-2000 are first analysed to obtain the q99 and q99,9 percentiles of GENpdf for the whole Mediterranean basin. On the other hand, climate simulations for 1981-2000 (control) and 2081-2100 (future) under A2 and A1B scenarios provided by five GCMs are considered. Monthly and subregional exceedance of q99 and q99,9 are calculated from these simulations. These exceedances are first compared against the ERA-40 time-spatial patterns to assess the goodness of each GCM for the control period, and then the changes between control and future time slices are evaluated. Furthermore, a first guess of a probabilistic risk map of medicane genesis is produced for the present and future climates. The present map is shown to be compatible with the true occurrence of medicanes. The future map indicates, for most of the regions and GCMs, a lower occurrence of medicane-favorable conditions at synoptic scale.

Paper 68:

The bow echo event of July 14 th 2010 in the Netherlands

Robertus Groenland, KNMI

Jacob Kuiper, KNMI

Sander Tijm, KNMI


In this study, the structure of a bow-echo on a squall line is investigated. This paper will assess the synoptic environment in which this squall line evolved, as well as the mesoscale features of the bow echo, including radar-reflectivity and radial velocity. Furthermore a model study on this case has been performed. With a non hydrostatic model (HARMONIE, resolution 2.5 km) the structure of the bow-segment can be visualized, including the RIJ. Examples from our 3D-visualisation technique will be shown. In addition a damage-survey was performed by meteorologists from the KNMI. This lead to a lot of information about the type and patterns in the damage. On July 14 th 2010 a squall line with several bow segments crossed the eastern part of the Netherlands from south to north. This event damaged many houses and trees and one camping site was hit severely fatally wounding two persons. A mile north of the camping site five electric poles went down. It was the first time in Dutch history that high winds were the cause of this. With a non-hydrostatic model (harmonie, 2.5 by 2.5 km resolution) we did some experiments in simulating this line of convection. With an extension of the domain, to include the entire development of the convection, some improvements and interesting results can be shown. Along the leading edge of the simulated squall line a bow segment developed. With a special 3D technique/projection a clear signal of an RIJ is visible. Still today this type of model projection is used for forecaster training purposes rather than for operational use.

Paper 69:

Regional trends in severe convective weather: a dynamical downscaling approach

Eric Robinson, Purdue University

Robert Trapp, Purdue University

Michael Baldwin, Purdue University

Alex Gluhovsky, Purdue University

Noah Diffenbaugh, Stanford University


The existence of historical trends in hazardous convective weather such as tornadoes and damaging winds has been difficult to assess with confidence due to the fact that the hazard observations are derived largely from eyewitness reports and/or damage. Such reports are convolved with the growth of population, changes in report procedure, development of education and awareness programs, etc. Unbiased time series of severe convective activity must be compiled if we are to understand if such activity has changed in the past, and also how it may change in the future, due to anthropogenic climate change. Our previous work has focused more on classifying the large-scale environment without considering actual convection initiation- a major caveat. Thus, we now explore the use of high-resolution dynamical downscaling for trend correction or reconstruction. Our basic approach involves a multi-decadal sequence of daily integrations of the Weather Research and Forecasting model in a convective-storm permitting setup, with global reanalysis data as initial and boundary conditions. Artificial neural networks are developed to identify and classify when severe convection is occurring in the model and to identify regions of the United States (and in principle, elsewhere in the world) with similar large-scale storm forcings. The relative skill of these networks in representing the spatial and temporal distribution of severe-thunderstorm occurrences over these different geographical areas is examined and differentiation of different severe convective modes is also explored. The connection of these trends to changes in regional dynamics is considered. Lastly, the implications of this study on identifying severe-thunderstorm trends in an anthropogenically enhanced climate are discussed.

Paper 70:

Quantitative Precipitation Forecast for convective flash flood rainfalls – evaluation of two models and verification techniques

Daniela Rezacova, Institute of Atmospheric Physics, ASCR

Petr Zacharov, Institute of Atmospheric Physics, ASCR


The contribution will present results of spatial verification for the forecasts of heavy local convective rainfalls. The rainfalls caused flash floods in the Czech Republic in 2009. Nine events occurring between 23 June and 5 July 2009 were analysed and simulated by the COSMO and ALADIN NWP models. The non-hydrostatic model COSMO runs without a cumulus parameterization with horizontal resolution 2.8km and the hydrostatic model ALADIN runs with 9km resolution. The quantitative precipitation forecast will be verified by gauge-adjusted radar data from the Czech radar network CZRAD. The verification procedure applies two spatial methods SAL and FSS. The study aims at answering the questions: (1) What is the difference in heavy convective rainfalls between models with different physics and horizontal resolution? (2) What is the difference between the verification results obtained by SAL and FSS techniques?

Paper 71:

Lightning knowledge and safety behavior in Austria and Bavaria, Germany

Alexander Keul, Salzburg University

Christine Duernfeld, Salzburg University

Julia Holzleitner, Salzburg University

Andreas Schoerghofer, Salzburg University

Lena Weichenmeier, Salzburg University


An Austrian survey on lightning knowledge and safety behavior with 133 respondents (age 20-84, mean 43) was organized in 2008 (Keul et al, 2009). As part of Interreg IVA project REBLAUS, a sample of 108 persons (age 19-81, mean 44) from Bavaria filled in the same questionnaire in 2010. The two surveys are analyzed for possible risk assessment differences. Given a list of natural risks, lightning was rated as medium risk in Austria and Bavaria. 66% of the Austrians felt well-informed about thunderstorms, in Bavaria 74%. Self-reported lightning fear was mostly low. In a distance calculation task („thunder 3 seconds after lightning“), the most frequent (and wrong) answer was 3 km (54-55%). 20% of the Austrians and 29% of the Bavarian answered correctly. Two of three medical knowledge items (killed instantly, cardiac arrest, resuscitation) were answered correctly by 70 to 90%. 26% in Austria knew about cardiac arrest (Bavaria: 67%). General physics questions (e.g. mountain CG maximum) were answered correctly by 50 to 95% (by Bavarians better than Austrians). Relevant safety items in the open - „crouched position“ (correct: A 32%, B 62%) and „3 meters distance to objects“ (correct: A 40%, B 69%) - indicate a lack of information, even for some academics. Historic folk beliefs („churchbells against lightning“) are largely rejected, but some survived („always highest point“ 50% correct, „different trees different danger“ only 24% A/40% B). Behavioral choice tasks (where to hide in a thunderstorm and where not) showed a high level of basic knowledge (buildings, car safe >95% A, >85% B correct) with few exceptions (metal fence, field Austria around 70% correct, Bavaria 95-99%). Although a number of life-saving questions is answered correctly by most respondents, there are still weak points in knowledge and behavioral choice in need of competent information.

Paper 72:

Quality of radio/television weather warnings in the eyes of the Austrian public

Alexander Keul, Salzburg University

Alois Holzer, Austrian Broadcasting Corporation ORF

Thomas Wostal, Austrian Broadcasting Corporation ORF


Media weather reports, the main channel of meteorological and climatological information to the general public, attract evaluations. Former research criticized that for a high-interest topic, recalled information by nonexperts is rather low. Is it realistic to assume instant learning? A 2008 pilot study (Keul et al., 2009) by the Salzburg University in cooperation with the Austrian Broadcasting Corporation ORF used historic radio messages about fair-weather or a storm. Of a sample of 64 adults, 10% did not recall anything, 43% general information, 47% weather details. Males had less recall for fair-weather, but better recall for the warning. With longer weather messages, more (false) details were recalled. In a second field experiment in the spring of 2010 in Salzburg City, Austria, a quota sample of 102 adults was interviewed on fair-weather or a warning situation. The TV or radio reports used gave the latest weather forecasts of the given day. Subjects were asked what they found important and could remember for their next-day use. Of the n=60 TV sample, 93% said they understood the fair-weather report and 100% the warning. 75% said the report speed was OK. 60% recalled their local prognosis (older people 70%). Of the n=42 radio sample, 77% said they understood the fair-weather report and 84% the warning. 81% found the report speed OK. 64% recalled their local prognosis. About 50% of the TV/radio sample liked to have more behavior advices. Practical conclusions were: Austrian TV weather is well-understood (visual + speech channel), radio weather (speech only) is a more complex listening task. Standard reports run too fast for 20-25% of the lay users. About 40% are not able to decode their local prognosis efficiently.

Paper 73:

Austrian media weather communication - strengths and weaknesses

Alexander Keul, Salzburg University

Alois Holzer, Austrian Broadcasting Corporation ORF

Thomas Wostal, Austrian Broadcasting Corporation ORF

Gunda Schuller, Austrian Broadcasting Corporation ORF

Johannes Eberl, Salzburg University

Johannes Mayer, Salzburg University

Louisa Hacking, Salzburg University


Media weather forecasters know that their reports reach an audience with different perception habits and that presentation modes make the difference. User research is needed as quality control. In the spring of 2010, a cooperation of the Austrian Broadcasting Corporation ORF with the University of Salzburg used a 56-item questionnaire to test weather knowledge of lay people and their understanding of weather forecasting and warning features with 237 Austrian adults (72 natives & 52 immigrants Vienna, 113 adults Salzburg). Only 15-20% expressed disinterest in weather. Weather information is mostly used for leisure time activities, clothing and transport mode selection. Television, radio and internet are the most popular information media. Radio weather has the top rank in eastern Austria, TV weather in western Austria. The lay knowledge level is high: 60% can name clouds (mostly cumulus, cirrus), over 90% know the adiabatic temperature gradient, 80% freezing droplets on a cold road surface, over 90% know a car is lightning-safe, but the majority computes a wrong lightning distance. Usual met report cloud cover phrases were correctly ranked (but overcast and heavy clouds mixed up). In 2008 the same was obtained for typical sunshine duration expressions. 50-80% marked the right interpretations for „shower“, 50% for „sleet shower“, 65-90% for „violent storm“. Usual wind velocity phrases were put in a correct rank order, also expressions for rain duration/frequency. Snowfall levels were ascribed to correct Austrian locations by 50-80%. Most important media weather elements were max-temperature tomorrow and next days, weather map, warnings and (rain/snow) prognosis for the next day. 75-80% said the weather elements should always appear in the same order. The most important prognosis range was tomorrow and a 2-3 days trend.

Paper 74:

Mesoscale numerical simulations of medicanes: Comparison against satellite-derived trajectories and isolation of key physical processes

Maria Tous, Meteorology Group, Dept. Physics, Universitat de les Illes Balears

Romu Romero, Meteorology Group, Dept. Physics, Universitat de les Illes Balears

Climent Ramis, Meteorology Group, Dept. Physics, Universitat de les Illes Balears


A few tropical-like cyclones have been developed over the Mediterranean Sea during the last decades according to the inventory of images provided by Meteosat satellite. These extreme small-scale warm-core storms, also called "medicanes", operate on the thermodynamical disequilibrium between the sea and the atmosphere, and sometimes attain hurricane intensity and threat the islands and coastal regions. Despite their small size, mesoscale model runs at moderate horizontal resolutions (7.5 km) made with MM5 are able to simulate the cyclone formation and general trajectory for most of the cases and emphasize the warm-core, axi-symmetrical structure of the storms. Simulation of the timing and precise details of the storm trajectories are shown to be more problematic when compared against the satellite images available for the events. It is hypothesized that the small size of the systems and the crucial role of moist microphysics, deep convection and boundary layer parameterizations are the main factors beyond these errors. On the other hand, a sensitivity analysis examining the role of the large-scale dynamic forcing and sea surface fluxes is conducted. Specifically, the potential-vorticity (PV) inversion technique is used to reduce the amplitude of the upper-level precursor trough in the model initial conditions, and the latent and sensible heat fluxes from the Mediterranean are switched off during the course of the simulations. Using a factor-separation technique, the individual effects of the aforementioned factors, as well as their interaction are determined. Results show that just before their mature phases, the enthalpy fluxes from the sea grow up, helping to develop these cyclones and inducing further cyclone deepening by diabatic heating. The joint action of the upper-level PV anomalies and these fluxes becomes the basic main factor for the genesis and evolution of the medicanes.

Paper 75:

Two examples of last winter convective structures in Spain

Benito Elvira, AEMET

Felisa Aguado, AEMET

Francisco Martín, AEMET


During the winter 2010-11 some deep convection episodes took place in Spain. In this paper two case studies are analyzed: an squall line over the Ebro valley on 28th January 2011 and a wide and lasting convective structure where some supercells and squall lines can be identified in the vicinities of Alboran sea on 4th March 2011. In the first case, the squall line produced 13 cm of accumulated hail (the highway AP-2 had to be closed for more than 1 hour) and moderate wind gusts (50 km/h). The second episode, although produced the most adverse weather on the sea, reached also the coast with about 15 mm/hour rainfall intensity and 55 km/h wind gusts. In this study the synoptic and mesoscale pre-convective environment, using NWP analysis and mainly satellite images, has been analyzed. Both cases show similar features: light instability in terms of CAPE, convergence area between the humid and warm air from the Mediterranean sea and the Atlantic westerly flow where were developed the initial convective cells and a moderate/strong vertical shear environment which indicates an important degree of potential organization. Once the convective lines have been set up, Doppler radar data show how the outflow wind from the downdraft converges, making up a line, with the easterly flow feeding and keeping alive the convection for a few hours. PPI reflectivity reveals narrow precipitation band just behind the convergence line. Radar wind is also used to guess cyclonic (anticyclonic) turns in the northern (southern) part of the squall lines and inside a supercell. Echotop and VIL data show light/moderate values, according to the winter; however, maximum reflectivity is close to 60 dBZ, quite similar to the spring/summer convection. The goal of this paper is to enhance the role of a well trained forecaster who has to deal with these convective situations, quite often associated to severe weather. NWP models are able to predict the synoptic and mesoscale environment, that is the pre-convective conditions, but usually not the convective structures and, what is more important, its surface effects. Forecasters need to know the physical process and the radar conceptual models linked to these structures in order to get a fast and correct identification of them and so a more precise and updated warning broadcast.

Paper 76:

Increasing severity of thunderstorms and cyclones: Environmental and socio-economic impacts in India

Shadananan Nair, Nansen Environmental Research Centre – India


Millions of poor with low adaptive capacity living in climate sensitive regions and an economy closely tied to climate sensitive natural resource base make India highly vulnerable to the impacts of climate extremes. Climate related disasters kill hundreds every year. Associated with rising temperature, significant increasing impacts are observed in case of cyclones, thunderstorms, lightning and floods. There is an increasing tendency of thunderstorm development in the Western Ghats region where the squalls cause widespread damage to agriculture, houses and forest. Lightning, flash floods and landslides cause casualties frequently. Heavy rain from the convective clouds erodes topsoil, threatening biodiversity. The thickly populated coastal zones, especially cities suffer from the increasing intensity and frequency of tropical storms. In addition to structural damages, surges during high tides paralyse coastal urban life, as the floodwater obstructs drainages. Surges contaminate water resources far inland. Strong winds affect the coastal circulation, upwelling and SST. Falling fish catch in the southern coast is linked to this. Impact of storms is noted even in coral environments. Impact on national economy also is large, as fisheries contribute significantly to overseas trade. Climate extremes are likely to retard the present economic growth, because of the massive investment required for adaptation, mitigation, post-hazard recovery and resettlement. Social issues like migration to safe location and competition for resources worsen. However, implementation of the policies and strategies including the coastal zone regulation act and of the measures for adaptation to climate change often fail because of various ecological, socio-economic, technical and political issues. A comprehensive assessment of the socio-economic and environmental impacts of increasing severity of thunderstorms and tropical storms in India, and of the current strategies and polices to face such challenges is made in this study. Suggestions for the improvement of the climate policy and adaptation strategy have been provided.

Paper 77:

Severe Convective Weather Forecasting in Europe

Chuck Doswell, Doswell Scientific Consulting


Convective storms generally develop, mature, and dissipate on time scales of a few hours, and their impacts generally are felt in regions of mesoscale spatial dimensions. Anticipation of the possible severe convective weather inherent in a synoptic-scale situation is critical for being able to response once storms begin. Numerical models, including those that permit explicit convection to develop within the model, are not likely to be adequate for providing accurate information about location and timing of severe convective weather any time in the near future. Therefore, forecasting severe convective weather can be challenging and experience is an important factor in being a success as a severe storms forecaster. The basic principles used for a scientific approach to severe storms forecasting are summarized and discussed. Beyond the meteorological challenge, however, is another important factor: the matter of how such forecasts can provide valuable information for the forecast users. Thus, important issues for forecasting severe storms in Europe include both meteorological and societal topics. A path by which an effective severe convective storm forecasting system can be created for Europe is presented that addresses some of the societal aspects of the challenge that need to be solved.

Paper 78:

Why does the atmosphere produce deep columnar convective vortices?

Chuck Doswell, Doswell Scientific Consulting


Although considerable research over the past several decades has led to extensive knowledge of environments in which supercells and tornadoes occur -- the primary examples of deep columnar convective vortices -- there has been little progress in understanding what role these rotating systems play in the atmosphere. Any comprehensive understanding of deep columnar convective vortices should include: knowledge of the energy source that powers the process, what instability mechanism is involved, and how the resulting system alters the atmosphere to mitigate the need for the process to continue. Two conceptual models have been developed that satisfy these requirements: (1) The energy source for supercells includes the storm-relative helicity of the air ascending in the updraft, in combination with the buoyancy associated with the deep convection. (2) The energy source for tornadoes is dominated by storm-relative helicity within the low-level inflow to the tornado. The instability mechanism in both is the basic exponential growth process associated with the so-called stretching term in the vertical vorticity equation -- the energy tapped from the source of helicity is used to drive the amplification of the kinetic energy of the vortex, which ultimately is dissipated by viscosity. It is suggested that the tornado model is comprehensive; it applies to all tornadoes, including mesocyclonic and nonmesocyclonic tornadoes. Although the models cannot be evaluated conclusively at this time, a plan for testing the validity of the proposed conceptual models is offered.

Paper 79:


Tommy Tolson, Tolson Engineering Services


The Texas Gulf Coast has been the recipient of numerous strong hurricanes in the past six years, most dramatically Hurricanes Rita (2005) and Ike (2008). These storms caused extensive damage to residential and commercial structures inducing a very large claim volume on property insurance carriers; extensive litigation also ensued. No consistent guide exists for data collection to evaluate structural damage to residential and light commercial buildings exposed to these storms. As a result, the breadth and depth of field engineering evaluations and damage assessments performed post-hurricane on these structures is quite variable. This paper proposes some data and information guidelines for properly evaluating hurricane-induced damage to residential and light-commercial structures.

Paper 80:

Modelling flash floods in small forested mountain watersheds using geographically referenced data

Mihai - Daniel Niţă, Transilvania University of Brasov

Ioan Clinciu, Transilvania University of Brasov


Severe storms which appear in mountainous areas in Romania produce highly unpredictable flash floods. Many mathematical models, rather expeditious than analytic, are used in predicting this kind of phenomena. This study used a highly analytic method for modeling flash floods produced in a experimental small predominantly forested watershed, located in mountainous area of Brasov county. The method is based on mathematical quantification of soil potential accumulation of water and uses geographically referenced data on: soil texture, soil porosity, soil humidity, antecedent rains, soil cover and watershed morphometry. Many modeling methods, despite of soil importance in producing the surface and subsurface flow, do not use too many data for characterizing this influence. Therefore the prognosis quality is enhanced in the presented study with the geospatial data on soil and land cover. The method itself predicts the flash floods phenomena by creating a synthetic hydrograph. The flood wave simulation is based on following cell raster characteristics: slope, soil cover, rainfall intensity, overland flow distance, and the discharge is calculated based on rainfall intensity and derived runoff coefficients. After applying the method on several hydrological events, the differences between measured and simulated were minor, especially on important events. Thereby the synthetic hydrograph of the highest peak discharge registered in 2010 (14th of July) had a 10 minutes temporal shift and a 6% overestimated value from the real hydrograph measured in cross section with a Thompson sharp crested weir.

Paper 81:

Observations of Low Level Mesocyclonegenesis Using a New Trajectory Mapping Technique

Daniel Betten, Cooperative Institute for Mesoscale Meteorological Studies at the University of Oklahoma

Michael Biggerstaff, University of Oklahoma

Gordon Carrie, University of Oklahoma


A dual-Doppler dataset was collected on the 29 May 2004 Geary, OK tornadic supercell by a pair of C-Band SMART radars with analyses every 3 minutes and a horizontal resolution of 750 meters. The scale of the low level mesocyclone (~7km in diameter) and the slow storm motion (~12 m/s) and evolution allow for the calculation of Lagrangian trajectories over the mature and dissipation stage of this tornadic mesocyclone. A new visualization technique was developed to map Lagrangian information from the trajectories to a single analysis time. The mapping of trajectories allows for the examination of hundreds of trajectories at a single time, displaying height, wind, or vorticity information at a point in the past or integrated over time. The examination of the evolution of the mapped trajectories from the 29 May storm reveals periodic surges of inflow and forward flank air around the low level mesocyclone that coincide to changes in the mesocyclone motion and strengthening of intensity. There are also surges of rear-flank downdraft air south of the mesocyclone that coincide with the concentration and downscaling of vertical vorticity in the mesocyclone. Examinations of vorticity budgets along specific trajectories compare very favorably to previous numerical modeling studies that show multiple source regions of air in the low level mesocyclone. Integration of the stretching term in the vertical vorticity equation and maximum tilting along the trajectory confirm modeling studies that suggest that enhancement of horizontal vorticity is more important for mesocyclone air originating in the forward flank than from the inflow regions east of the circulation. The contribution of precipitation loading in the baroclinic generation of horizontal vorticity is also estimated.

Paper 82:

Physics and Dynamics around the Overshooting Top of Severe Thunderstorms

Pao Wang, University of Wisconsin-Madison

Shih-Hao Su, National Taiwan University

Martin Setvák, Czech Hydrometeorological Institute, Prague

Zdenek Charvat, Czech Hydrometeorological Institute, Prague


Understanding the physics and dynamics of thunderstorm process greatly facilitates the forecast and nowcast of storms using observational data. This paper will present a study of the physics and dynamics of the overshooting top of some severe thunderstorms utilizing a 3-D cloud model equipped with explicit microphysics to perform simulations of some storms occurred in the past. The model results will be compared with observations (especially satellite visible and IR images) to verify the simulation. Then the model physics will be used to explain the physics of the observed phenomena around the overshooting top. Observations from both geostationary and polar orbiting (such as the NASA A-train) satellites will serve as the data source for comparison. Applications of the new understandings will be discussed.

Paper 83:

Decadal variation of heavy rainfall frequency in Kyushu, Japan and associated synoptic weather patterns

Koji Nishiyama, Kyushu University

Kenji Wakimizu, Kyushu University


The aim of this study is to investigate decadal variation (30 years : 1979-2008) of heavy rainfall frequency in Kyushu area (target area) located in the west of Japan and associated synoptic field patterns by using the Self-Organizing Map (SOM), which is one of unsupervised neural networks (pattern recognition) and can convert complex nonlinear features into simple two-dimensional relationships. Heavy rainfall frequency in the target area is higher for recent 10 years (term3: 1999-2008) than for the other periods (term1: 1979-1980, term2: 1989-1998). Therefore, to recognize what kinds of synoptic fields affected the variation in heavy rainfall frequency, 1) synoptic fields (NCAR-NCEP reanalysis data: grid spacing of 2.5 degrees) for 30 years were divided into 900 patterns (25 groups) using the SOM and 2) decadal variation of heavy rainfall frequency per each group was calculated. The synoptic field patterns are defined by the spatial distribution of wind components at the 850 hPa level and Precipitable Water (PW), which can express the Low-Level Jet (LLJ) and high PW area formed by convective activity, relating the formation of heavy rainfall. Here, heavy rainfall frequency was calculated from the AMeDAS, Automated Meteorological Data Acquisition System, which shows dense observational network with fine resolution of about 17km in Japan. From these analyses, it could be confirmed that heavy rainfall were caused by synoptic field patterns characterized by the Low-Level Jet (LLJ), high PW area, steep PW gradient (frontal activity), and cyclonic activity including Typhoon. Moreover, increasing trend of heavy rainfall frequency could be recognized for recent period (term3) in two groups relating to dominant LLJ with high PW area, and one group relating to dominant frontal activity. On the other hand, decreasing trend of heavy rainfall frequency could be recognized in one group relating to cyclonic activity (excepting Typhoon) with high PW area.

Paper 84:

Nowcast Warnings within AutoWARN - A System for the Automatic Support of the Weather Warning Service at Deutscher Wetterdienst

Bernhard K. Reichert, Deutscher Wetterdienst

Paul James, Deutscher Wetterdienst

Sebastian Trepte, Deutscher Wetterdienst


AutoWARN integrates various meteorological products in a semi-automated warning process with manual monitoring and decision possibilities for the forecaster. It exploits and combines observations, radar products, nowcasting products, statistical forecast products, and NWP (Numerical Weather Prediction) model forecasts. It generates and permanently updates forecast-time dependant automatic warning status proposals. The forecaster manually controls and - if necessary - modifies the automatic proposals. They are then exported to a system generating textual and graphical warning products for customers. Within project AutoWARN the new nowcasting system CellMOS has been introduced. It is a cell identification and Model Output Statistics (MOS)-based tracking system delivering probability information on potential cell tracks for the next two hours. In order to provide a generic optimal solution for nowcast warnings in AutoWARN, all available nowcast input data are processed together in a single grid-based system: the NowCastMIX. Using a Fuzzy-Logic-approach, NowCastMIX combines available nowcast-information from CellMOS, KONRAD (radar-based cell identification and tracking system), lightning observations, VIL (3D radar-based vertically integrated liquid water content), RADVOR-OP (precipitation prediction system), Synop-Observations, and various background fields from the NWP model COSMO-DE. This provides a single optimal set of gridded warning fields every 5 minutes. The goal of NowCastMIX is thus to provide and optimize an ongoing real-time synthesis of the various nowcasting and forecast model system inputs to provide a single, consolidated set of most-probable short-term forecasts. The presentation will focus on illuminating the concepts of NowCastMIX and AutoWARN and demonstrating the currently running pre-operational versions of the systems at the National Warning Centre (NWC) of the Deutscher Wetterdienst.

Paper 85:

Simulation of the storm from 15 August 2010 with a high resolution COSMO NWP model

Zbyněk Sokol, Institute of Atmospheric Physics, ASCR

Petr Zacharov, Institute of Atmospheric Physics, ASCR


On 15 August, 2010, weather in central Europe was influenced by a slowly moving cold front. Strong convective activity accompanied by heavy precipitation, supercell occurrence and in some cases also by hails was observed in various locations in Bohemia. Large hails, reaching up to 5 cm in diameter, were registered in Prague, where caused an estimated 180 million EUR damage to property in less than one hour and hurt several people. Simulations of observed convective storms with the COSMO NWP model will be presented in our contribution. The model will be integrated with a horizontal resolution of 1 km to 2.8 km and with various numbers of model vertical levels. The model will be run with the assimilation of observed radar reflectivity and with one-moment and two-moment microphysics. The aim of our study is to assess how accurately the model is able to simulate the studied event and answer the following questions: (i) How the forecasts depend on the horizontal and vertical resolution of the model; (ii) How the precipitation forecasts depend on one-moment and two-moment parameterization of cloud microphysics; (iii) Is the COSMO model with two-moment parameterization of cloud microphysics able to forecast hails?

Paper 86:

A Diagnostic Study of Explosive Cyclogenesis over Northern European Russia

Andrey L. Vetrov, Perm State University

Nikolay A. Kalinin, Perm State University


Explosive cyclones storms can generate hurricane-force winds and convective precipitation. When track across land, the results can be devastating. Usually strong extratropical cyclone tracks over the north part of Europe. In this work we focus on those cases in which the European Russia are is affected by extreme wind and precipitation episodes during past years. Convection is an important contributor to the total precipitation and wind involved in a midlatitude cyclone. In an occluded front, part of the precipitation is caused by convective instability, while the post-frontal precipitation is exclusively triggered by this mechanism. In this study, heavy precipitation in an intense mid-latitude cyclone is analysed by generation APE (Available Potential Energy) due latent heat. We are used a numerical WRF model (Weather Research and Forecast) for simulations.The data set used is given by the Global Forecast System (GFS) is a global numerical weather prediction computer model of the National Oceanic and Atmospheric Administration (NOAA) with 6-hour resolution. The model was run with a horizontal grid spacing of 10 km and 37 vertical levels. The model top is at 25 hPa. Split explicit time differencing is used with a time step of 120 s. Wind, temperature, relative humidity and geopotential height interpolated to 17 pressure levels are used for calculations APE and generation of APE due latent heat release. Parameterization scheme are used to estimate diabatic heating due latent heat release. Diagnostic calculations reveal that generation available potential energy due to sensible heating influence on development heavy precipitation. High speed and large deepening of the low is related with liberation of latent heat in lower levels.

Paper 87:

Implementation and first evaluation of the nowcasting system INCA in Belgium (INCA-BE)

Maarten Reyniers, Royal Meteorological Institute of Belgium


In 2010, the Royal Meteorological Institute of Belgium (RMI) started the implementation of the INCA system (Integrated Nowcasting through Comprehensive Analysis) in Belgium: INCA-BE. INCA is a nowcasting system for the analysis and nowcasts of several meteorological fields, like temperature, humidity, wind, cloudiness, precipitation and precipitation type, and some derived fields like icing potential, wind chill, wind gusts and visibility. It operates at a horizontal resolution of 1km, and on an hourly basis (10 min for the precipitation fields). It has been developed at the Austrian national weather service (ZAMG), and it is currently used in several European countries as the operational nowcasting system. In this contribution, we will first give an overview of the main characteristics of the system, the adaptation to our local domain, its data sources and the analysis and nowcasting methods used. The precipitation module of the system will be discussed in more detail. Since INCA-BE is running at the RMI for some months now, we are able to present a first qualitative evaluation of the system performance. Especially, we will focus on the results of the precipitation nowcasting in convective situations.

Paper 88:

A provisional climatology of tornadic and non-tornadic linear convective systems in the UK

Matthew Clark, Met Office


A seven-year climatology of cool-season (September to February) quasi-linear convective systems in the UK has been constructed using archived radar data. 103 lines meeting basic length (≥100 km) and duration (≥2 h) criteria were identified over the period September 2003 to February 2010. Of these, 90 occurred in association with well-defined fronts (mostly cold fronts). The remaining 13 occurred within polar-maritime air masses to the rear of frontal systems. The lines occurred most frequently over inland areas of southern and central England, with frequencies decreasing markedly to the north. The environment of each line was investigated by analysis of surface and radiosonde observations, from which representative values of various environmental parameters were obtained. TORRO (Tornado and Storm Research Organisation) data were used to categorise events into ‘tornadic’ and ‘non-tornadic’ classes. The distributions of parameter values within the tornadic and non-tornadic classes were compared. Tornadic lines were found to be associated with significantly larger pre-line CAPE, line-normal forward motion, cross-line horizontal wind veer, and cross-line temperature difference. However, of the various vertical shear parameters analysed (including 0 – 1 km bulk shear, 0 – 3 km bulk shear, line-normal and line-parallel 0 – 1 km shear), no significant differences could be found between tornadic and non-tornadic classes. Some of the implications of these results for the forecasting of tornado risk in cool-season environments will be discussed.

Paper 89:

Thunderstorm activity and its relationships with southern cyclons in Estonia, 1950-2004

Kaupo Mändla, University of Tartu, Department of Geography

Sven-Erik Enno, University of Tartu, Department of Geography


This study examines the activity of thunderstorms over Estonia during 1950-2004. Possible relationships with southern cyclons that transport warm and humid air to the northern Europe are also studied. Thunderstorms activity is studied on the basis of visual data from one maritime, two coastal and two inland stations. Average annual duration of all recorded thunderstorms varied from 26 to 41 hours in the studied stations during 1950-2004. The annual average thunderstorm duration gave statistically significant downward trends at three stations and insignificant upward trends at two stations. Obvious fluctuations in thunderstorm activity appeared. Periodicity of 12 to 13 years was distinguishing from Fourier analysis. It can be also noticed that the overall duration of thunderstorms was slightly higher during 1950-s to 1970-s. Southern cyclones form over the Mediterranean and Black Sea, move northwards and transport warm and humid tropical air to the higher latitudes. During the warm season they have a high potential to cause wide and severe thunderstorms in the northern part of Europe. About 10-13% of all cyclons that affect Estonia are southern cyclons. Our study revealed that 99% of all southern cyclones that caused thunder appeared from April to October. From April to October, 342 southern cyclones were detected in 1000 km radius circle that centre point located in Estonia between 1950 and 2004. About 40% of them caused thunderstorms to Estonia. The highest thunder appearance percentages (about 60-80%) in case of southern cyclones found in summer months. In August even 82.5% from all southern cyclones induced thunder. During the study period, thunder performance with southern cyclones decreased insignificantly.

Paper 90:

Statistical advection model SAM for precipitation nowcasting

Zbyněk Sokol, Institute of Atmospheric Physics ASCR

Petr Pešice, Institute of Atmospheric Physics ASCR


We will present a statistical advection model (SAM) for nowcasting of precipitation for the area of the Czech Republic. The SAM is an advective-statistical algorithm, which utilises radar, satellite and lightning observed data and prognostic values of an operational numerical weather prediction. The advection part of the SAM is based on the COTREC technique and the statistical part of the model is a regression model, which is complemented by the distribution correction of forecasted values. This correction modifies the distribution of regression model outputs to be similar to the observed distribution of precipitation. The model forecasts precipitation totals for boxes of 3 km by 3 km covering Czech catchments. The lead times of the quantitative precipitation forecasts (QPF) are 1, 2 and 3 hours. The evaluation of accuracy of SAM’s forecasts will be presented and compared with forecasts by an extrapolation using the COTREC technique. The accuracy of the QPF will be evaluated by the root-mean-square-error, absolute error and bias. The forecasts will be also assessed by “fuzzy” measures SAL and Fractions Skill Score. The precipitation forecasts will be used as inputs of a hydrological model for selected events and the hydrological forecasts will be evaluated.

Paper 91:

The hailstorm in Valjevo city (west part of Serbia) on 19th June 2010

Vesna Ivkov, Republic Hydrometeorological Service of Serbia


The aim of this work is to present a analysis of a severe thunderstorm with hail, heavy rainfall and strong wind which covered the territory of town Valjevo (west part of Serbia) through the use of radar data obtained by Doppler S-band radar on Fruska Gora Mountain and non-Doppler S-band radar in Valjevo. A thunderstorm developed in district of river Drina. The storm grown and moved from southwest direction and very quickly, it enters the territory of town Valjevo and by moving from southwest transferred across west parts of Serbia. By synoptic station in Valjevo there was registered 33.7 l / m2 precipitation amount in 3 hours period. This thunderstorm also produced hailstones with grain sizes of the corn seed to the hazelnut and caused serious crop damages. Using Rainbow, HASIS and HASIS 3D software we analyzed these thunderstorms during their lifetime. In the radar imagery can be clearly seen an active convective cell with large vertical extension affecting the city of Valjevo, with reflectivity values that surpass the 65 dBz.

Paper 92:

Present status and preliminary results of the Belgian lightning detection network

Dieter Poelman, Royal Meteorological Institute of Belgium


The Royal Meteorological Institute of Belgium (RMI) has been operating a SAFIR lightning detection network (Système d'Alerte Foudre par Interférométrie Radioélectrique) since 1992. The SAFIR network consists out of five stations and is able to detect the total lightning (TL) electrical activity, i.e., cloud-to-ground (CG) as well as cloud-to-cloud (CC) discharges. The location of lightning discharges are determined by triangulation of the signals in the very-high frequency (VHF) band. A capacitive electrical antenna allows then to discriminate between CC and CG. In addition, an alternative processor (TLP) for treating the data runs in parallel, combining VHF interferometry with the time-of-arrival (TOA) technique for the localization and characterization of the TL activity. Furthermore, data from our network is recently combined in real-time with data from a demo-network around Paris consisting out of three sensors, and with four sensors of KNMI (the Netherlands). In this contribution, we focus on the changed configuration of the network, the use of the TOA method and the expected influence on the lightning detection. Data from the present and previous lightning season are analyzed and presented, as well as some preliminary intercomparison studies with other lightning networks.

Paper 93:

Comparisons of kinematical retrievals within a simulated supercell: Dual-Doppler analysis vs. EnKF data assimilation

Corey Potvin, National Severe Storms Laboratory

Louis Wicker, National Severe Storms Laboratory

Alan Shapiro, University of Oklahoma


Mobile Doppler radar datasets permit the retrieval of convective wind fields on spatiotemporal scales unresolved by conventional stationary radars. These retrievals and the dynamical analyses derived from them, such as parcel trajectory and vorticity budget calculations, can provide valuable insights into storm dynamics. Unfortunately, the validity of conclusions drawn from such analyses can be threatened when significant retrieval errors occur. Thorough understanding of the error characteristics associated with different kinematical retrieval techniques is therefore critical to advancing understanding of convective storms. In this study, we compare the performance of two sophisticated analysis techniques – variational dual-Doppler analysis (DDA) and ensemble Kalman Filter data assimilation (EnKF-DA) – in retrieving the 3-D wind field within a simulated supercell thunderstorm. Since EnKF-DA combines information from both observations and model dynamics/physics, it should ideally produce wind analyses superior to those obtained from DDA, given the same observational data. However, errors in the forecast model, observational operators and the EnKF settings can introduce substantial errors into the analyzed wind field. It is therefore conceivable that wind retrievals from EnKF-DA may sometimes be inferior to those obtained using DDA. We perform our truth simulation and data assimilation experiments using the National Severe Storms Laboratory Collaborative Model for Multiscale Atmospheric Simulation (NCOMMAS) and its associated EnKF-DA system. The DDAs are performed using a variational technique that weakly satisfies observational, mass conservation and smoothness constraints. Perfect-model EnKF-DA experiments are performed to place an upper limit on the expected accuracy of storm-scale EnKF-DA wind analyses when data are available from either one or two radars. Imperfect-model experiments are then performed to provide a more realistic assessment of the errors that can be expected in practice. Of particular interest is whether single-radar EnKF-DA can produce storm-scale wind analyses of similar quality to those obtained using DDA. Implications for mobile radar deployment strategies are discussed.

Paper 94:

Time Series of Severe Storms in Austria, 1951 – 2010

Otto Svabik, Central Institute for Meteorology an Geodynamics, ZAMG


As an addition to the severe weather chronicle of ZAMG (mainly based on given press coverage) the annual variability of the number of storms in Austria, gusts and hail, within the last sixty years is investigated. The aim of the investigation is to develop times series of severe weather events able to indicate climate change. Is there an increase of the severe weather events causing heavy damages, like insurance data are telling (as an argument for climate change)? The mentioned severe weather time series are not homogeneous, unfortunately. They include breaks, which are not due to natural climate variability. One type of such breaks is caused by changes of the keepers of the chronicle, specifying the weather impacts individually. To explain the increasing amounts of losses during the last 60 years, we have to consider the increasing density of population and values in towns, the higher standard of living, the changing of regional infrastructure and environmental conditions finally.

Paper 95:

A climatology of supercells for south-eastern Romania

Sorin Burcea, National Meteorological Administration

Bogdan Antonescu, The University of Manchester

Daniel Carbunaru, National Meteorological Administration

Monica Sasu, National Meteorological Administration

Aurora Bell, National Meteorological Administration


In order to understand the spatial and temporal distribution and frequency of prolific severe weather reports in the south-eastern Romania, a radar-based climatology of supercells is presented. Supercell forecasting, their initiation, severity, and associated weather represent a difficult task for Romanian forecasters. Conceptual models for convective storm initiation, based on radar, satellite and synoptic observations have been developed during recent years with the aim to provide guidance for forecasters. Nevertheless, many challenges like recognizing the environmental mesoscale factors that favor the onset of supercellular thunderstorms still remain. This applies mainly in the south-eastern Romania, where the environment tends to be more conductive to severe convection than in other parts of the country. A five years supercells climatology was developed for south-eastern Romania. The current climatology includes the initiation point, the storm track, the termination point, and the duration of the supercells. The events were selected from the radar dataset covering the convective season (May–September) between 2003 and 2007, and provided by the Medgidia WSR–98D weather radar. Using the geographic information systems (GIS), the radar-based supercell climatology was organized into a geospatial database that enable a powerful spatial analysis and future use of the dataset by various GIS applications. A total number of 282 supercells have been identified, with aproximately 60% being observed during July–August. The diurnal distribution of the supercells shows a main maximum at 1400 UTC, and a secondary maximum at 1800 UTC. The spatial distribuiton of the initiation points indicates that the supercells are developing in the vecinity of the Black Sea, and then moving toward north-east (25% of the cases). The mean monthly initiation point of the supercells is moving north-west toward inland during May–July, and then toward Black Sea during August–September.

Paper 96:

Impact of Alps-Apennines orography in distribution of severe weather and tornadic events through northern Italy

Angelo Bertozzi, SMS METEOCENTER

Pierluigi Randi, MeteoCenter

Niccolò Ubalducci, Thunderstorm Team

Andrea Griffa, Thunderstorm Team

Andrea Chini, Thunderstorm Team


The data frequency and territorial distribution during april-october period of tornadic events (of miso and mesocyclonic kind) in the north of Italy (particularly referred at Po Valley) show that these phenomenas tend to be concentrated in restricted geographical areas, with triggering dynamics that seems to be strongly influenced by lower troposphere interactions between Alpine and Apennine mountains and the Adriatic Sea surface flux as well. These forcing factors likely could change, sometimes dramatically, thermodynamic parameters and mesoscale windshear, favoring the trigger of severe weather events in conditions of relatively weak or moderate instability. This work aims to analyze through high resolution reforecasting of about 50 tornadic events occurred in the Po Valley (between 1980 and 2010) the mesoscale orographically induced forcings that assumes deep meaning in determining in the same areas severe weather phenomena, as regards both episodes whether it be mescyclonic or not miscyclonic. Re-forecasting runs was performed with WRF model framework (ARW core solver with variational DAS) running up on high resolution (1.4 km) quite large domains (1400 x 900 - 60 vertical levels) central Mediterranean centered. All runs were forced with first guess and boundary condition obtained from reanalysis (CFS reanalysis, MERR, NCEP, ERA interim) whith re-run forced with GFS operational run archive, in order to assess the model skills in localization and evaluation of mesocyclonic/severe weather events performance in forecasting and reforecasting runs.

Paper 97:

Usefulness of MPEF Divergence product in diagnosing the environment of deep convection

Christo Georgiev, National Institute of Meteorology and Hydrology


EUMETSAT generates the Upper-troposphere Divergence (DIV) product and disseminates it operationally through EUMETCast since 2008. The DIV product is hourly calculated at Meteosat Product Extraction Facilities (MPEF) from Meteosat Second Generation (MSG) channel 6.2 Atmospheric Motion Vectors. In this study, the quality of the MPEF DIV product has been assessed by comparison with NWP model analyses of wind field as well as wind observation data at upper-air. Considering the effect of smoothing out information horizontally by the interpolation procedure applied in DIV product generation, a specific colour palette is used that allows inferring the information content of the product for the purposes of upper-level diagnosis. In order to assess the potential of MPEF Divergence product regarding possible nowcasting applications, 1166 convective cells with cloud top temperature lower than –50 °C that developed over Europe and Northern Africa from 15 June to 22 July 2010 were considered. The results show that 76 % of all convective cells initiated at areas of divergence seen by MPEF DIV product. In such synoptic conditions the MPEF DIV imagery is useful to monitor the evolution of the divergent upper-level flow pattern before the onset of deep convection. The composite images of MSG DIV field can be considered in order to anticipate where the upper-level dynamic environment favours the development of strong convection through forcing for ascent. In addition to other relevant parameters from NWP/satellite products (as upper-level PV anomalies, jet streams, instability) the MPEF DIV product may be used in support to deep convection forecasting. The usefulness of this approach for the purposes of intense convection forecasting during convective events in 2011 is being studied and results will be reported in the paper.

Paper 98:

Thunderstorm nowcasting at the Institute of Meteorology and Water Management

Katarzyna Bednarek, Institute of Meteorology and Water Management

Rafal Kielar, Institute of Meteorology and Water Management


Last year extreme weather events in Poland showed that there is a great need to develop new forecasting methods and nowcasting tools. To meet the needs of emergency services, IMWM decided to create a workstation which should concentrate on forecasting occurrence and movement of thunderstorms and other phenomena such as hail, wind gusts etc. This presentation is planned to describe the organization of the workstation and an example of the decision process based on one extreme case study. Since this project starts at the beginning of June, there are still some problems unsolved.

Paper 99:

GIS analysis of Austrian-Bavarian CG lightning data

Thomas Prinz, iSPACE Research Studios Austria

Wolfgang Spitzer, iSPACE Research Studios Austria

Christian Neuwirth, iSPACE Research Studios Austria

Wolfgang Schulz, OVE ALDIS Austrian Lightning Detection and Information System

Gerhard Diendorfer, OVE ALDIS Austrian Lightning Detection and Information System

Alexander Keul, Salzburg University


The EC-funded Interreg IVA project REBLAUS „Cross-border nature risk management by GIS-based analysis of CG lightning activity“ is a cooperation of several institutes and scientists (electrical engineering, geoinformatics, meteorology, insurers). REBLAUS covers an area of 40,000 square kilometers encompassing the Austrian Federal province of Salzburg and southeastern Bavaria, Germany. From 1998 to 2009, the lightning detection network EUCLID registered about 1 million CG lightning flashes in the REBLAUS area populated by roughly 1.3 millions inhabitants. To analyze CG lightning activity fine-grain with respect to topography, building forms and density, population density, bodies of water and technical infrastructure, geographic information systems (GIS) are used and CG flashes are matched with reported damages. The REBLAUS area, annually analyzed, has a high CG density variance due to varying thunderstorm activity/tracks. An aggregation map of ten years local CG activity showed lightning densities per square kilometer from under 1 to 9. Regional high density (>3 per sq.km) in Germany was observed NE Rosenheim, Bavaria. In Austria, high density spreads across the northern edge of the Alps from Lofer to Schladming. Hot spots are also located S Murau (>5 p.sq.km) and on two mountaintop towers near Salzburg City and Kitzbuehel (upward initiated lightning, <26 p.sq.km). CG density maps were computed using local flash density (LFD), focal flash density (FFD) or probabilistic flash density (PFD). Using SRT NASA data, topographic effects were analyzed with variations of concave vs. convex shapes vs. planes. Mean sea level had a strong effect. No CG clusters due to water bodies were detected. Using 7 years of insurance-reported damages, CG flashes per damage ranged from 1.87 for Salzburg City to 17.27 for Salzburg-Tennengau. Damages usually originated during a few thunderstorms per year. Therefore, a meteorological evaluation of the five major-damage thunderstorms is underway.

Paper 100:

Comparative verification of different nowcasting systems to support optimization of severe weather warnings

Kathrin Wapler, Deutscher Wetterdienst (DWD)

Sebastian Trepte, Deutscher Wetterdienst (DWD)

Martin Goeber, Deutscher Wetterdienst (DWD)

Paul James, Deutscher Wetterdienst (DWD)


Severe weather associated with deep convection pose a significant threat to life, property and economy. Fatalities, injuries and damages might be caused by lightning, gusts, hail, heavy precipitation or tornadoes. Therefore the provision of accurate and timely nowcast information, i.e. warnings provided by the national meteorological services, is essential for the general public as well as special users like emergency services and aviation. Several algorithms exist which detect and nowcasting deep convection. Most of them are based on either radar reflectivity measurements, like KONRAD (CONvektion evolution in RADar products, Lang et al. 2003), CellMOS (Cell Model Output Statistics, Hoffmann 2008) and Rad-TRAM (Radar TRAcking and Monitoring, Kober and Tafferner 2009) or on satellite measurements like RDT (rapid developing thunderstorm, Morel et al. 2002) and Cb-TRAM (Cumulonimbus TRAcking and Monitoring, Zinner et al., 2008). Due to their small horizontal extent, severe weather phenomena resulting from deep convection are rarely entirely and uniquely captured by current observation systems which hamper verification efforts. However, verification is needed to assess the quality of the algorithms, to determine their strengths and weaknesses and consequently to lead to improvements. The use of consistent verification methods is crucial to compare the different systems. Of particular interest is the question how these algorithms can optimally be used to issue warnings of thunderstorms as well as accompanying specific phenomena like gusts or hail. It is shown that the combination of different algorithms improve the quality of the nowcast. The thorough comparative verification enables the optimal use of the nowcasting systems, helps to improve the individual algorithms and supports the development of the most successful combination of the existing algorithms which is a basis for the provision of accurate warning information.

Paper 101:

SVAT modelling approach in support to flood risk assessment in Bulgaria

Julia Stoyanova, National Institute of Meteorology and Hydrology

Christo Georgiev, National Institute of Meteorology and Hydrology


Besides the space–time structure of the rainfall, the initial soil water content is of importance for the accuracy of the simulated flash-flood events and their forecast. A way to account for initial soil conditions at a site scale is running a soil– vegetation–atmosphere transfer (SVAT) model. The approach adopted in our study is evaluating land surface moistening for analyses and forecasts of over-moistening and its exploring in flood risk assessment. Numerical (1D vertical) SVAT model has been developed and operationally used at NIMH of Bulgaria since 2010, to serve as a source of data for land surface analyses, accounting for regional site-specific physical climate. The energy and water transfer processes are simulated and vegetation root zone moisture conditions are daily quantified (at 4 soil depths in the 100 cm layer) for field conditions at a wide range of climatic/vegetation environment (land cover type, stage of growth, temperature, roughness and evaporative demands, soil water content at specific soil type). A 6-level threshold scheme of root zone moisture availability is developed and used operationally as a diagnostic tool in support to issuing warnings of weather and climate over-moistening (or drought) extremes. Colour-coded maps covering the main administrative regions of Bulgaria visualize the moistening conditions. NWP short-range rain forecasts are considered in the view of the probability for soil over-moistening and used as additional quantitative information for the National MeteoAlarm System. Using case studies of strong convection and rapid cyclogenesis as examples, the paper presents the usefulness of this approach to help diagnosis and prediction of flood risk caused by extreme precipitation events. The utility of the results comes also from the recognition of soil moisture as a climatic forcing factor, which may affect both severity and frequency of extreme weather events.

Paper 102:


Montserrat Aran, Meteorological Service of Catalonia

Juan Carlos Peña, Meteorological Service of Catalonia

Nuria Pérez-Zanón, Meteorological Service of Catalonia


Using a database of daily torrential precipitation from 3 sites of the Catalan coastal area (Catalonia, Spain), the main atmospheric circulation patterns producing torrential rainfall in the area are derived. The database has been constructed with the information extracted from the scanned and digitalized pluviograph record from the following meteorological stations: Estartit (northern coast, Girona) for the period 1986-2006; Barcelona-Fabra Observatory (central coast, Barcelona) from 1927 to 2006; and finally, Ebro Observatory (southern coast, Tortosa, Tarragona) from 1905 to 1950. An episode of torrential rainfall is defined as those days that precipitation exceeded 1 mm/min. This threshold is used to obtain the curves of Intensity-Duration-Frequency which are essential for the knowledge of the spatial and temporal evolution of rainfall intensity for hydrological purposes. The total number of selected episodes is 123 for Estartit, 705 for Barcelona-Fabra and 248 for Ebro. The principal sequence pattern analysis has been applied to obtain the main synoptic patterns for torrential episodes. In order to take into account the environment prior to the torrential event (day D), we have used three analyses in the eighteen previous hours (D-6, D-12 and D-18). The applied technique consists of applying the principal component analysis to reduce the dimension of each variables and the cluster analysis on the most relevant S-mode principal component to obtain the atmospheric patterns. In the S-mode data matrix, the variables are the sequential grid point of the 4 temporal analyses linked to the event and days are observations. The variables used are mean sea level pressure, 850-500 hPa thickness and relative humidity at 700 hPa from the NCAR-NCEP reanalysis. The results show between 6 and 8 atmospheric patterns depending on the zone with a predominance of western and southern flows at mid levels associated to the pass of a trough or a low centred over Catalonia.

Paper 103:

The major hail days of 1988-2010 in Southwestern France

Claude Berthet, ANELFA

Jean Dessens, ANELFA

Jose Luis Sanchez, University of León


Before the hailpad networks were developed in France by the ANELFA, a major hail day in a region was just a day with strong hailfalls on precious crops, or on a large city. With the hailpad networks installed from 1988 in Southwestern France, and then, a few years later, in central and Mediterranean areas, severity indices of hail days could be computed on a physical basis. In this paper, a daily classification is established for the 1988-2010 period in the Atlantic and Pyrenean areas by successively considering the hailfall frequency (percent of stations impacted per day), the hail intensity (mean kinetic energy per hailfall per day) and the hail amount (mean kinetic energy per installed station per day). In the Atlantic region, May 11, 2009 is number one both for the frequency and for the hail amount. In the Midi-Pyrénées region, July 16, 2009 is the primary hail day of the period (number one for the amount, number two 2 for the frequency), followed by April 16, 2007. For these major hail days, and for the few other following in the series, we examine the hail parameters (hailstone size distributions) and the meteorological conditions (origin of air mass, freezing level height, wind velocity, CAPE, etc.) in order to look for the processes at the origin of very severe hail and to improve the forecast and prevention of these extreme events.

Paper 104:


Michalis Sioutas, ELGA-Meteorological Applications Centre

Wade Szilagyi, Meteorological Service of Canada

Alexander Keul, Vienna University of Technology


One of the most impressive waterspout occurrences are outbreaks, which are referred to as multiple waterspouts in a given period of time over a region and generally produced by the same weather system. Waterspout outbreaks are classified as either small SO (2-9 waterspouts), moderate-MO (10-19), or large-LO (20 or more) on a single day. The areas of research are those of Eastern Mediterranean, the Baltic Sea and the Great Lakes of North America. On September 5, 2002, Keul documented a series of 13 waterspouts in the Eastern Mediterranean off the Cretan north coast. He also documented two outbreak days (August 11 & 13, 2006) over the Baltic Sea in which 10 waterspouts occurred on each day. During the fall of 2003, Szilagyi documented a record breaking outbreak of 66 waterspouts over a seven day period over the Great Lakes of North America. During the period 2000-2010, 163 individual waterspouts were reported in 41 outbreak days over the Aegean and Ionian Sea, with 39 days corresponding to SO category, 1 day to MO and 1 day to LO. During the same period, 176 waterspouts were reported in 27 outbreak days over the Baltic Sea, and were classified as SO (25 days), MO (2 days) and LO (0 days). During the period 1994-2010, 731 waterspouts were reported in 158 outbreak days over the Great Lakes, and were classified as SO (149 days), MO (7 days) and LO (2 days). Meteorological aspects in terms of synoptic, mesoscale and thermodynamic environments are examined in order to identify weather patterns most conducive to waterspout outbreaks. The Szilagyi waterspout nomogram forecasting tool was also applied, incorporating a larger single waterspout data sample. Inter-comparisons between waterspout outbreak characteristics for the different geographical areas are also examined.

Paper 105:

Synoptic factors conditioning extraordinary instances of thunderstorms in Central Europe in the years 1951-2010

Zuzanna Bielec-Bakowska, Department of Climatology, Faculty of Earth Sciences, University of Silesia

Leszek Kolendowicz, Department of Climatology, Adam Mickiewicz University in Poznań

Grazyna Knozova, Czech Hydrometeorological Institute, Brno Regional Office


Contemporary climate changes are used as the basis for numerous research projects concerning changes in the occurrence of climatic events that are considered dangerous or strongly impacting the economic activities of man. Such events also include the occurrence of thunderstorms and accompanying phenomena. The objective of research was to determine types of synoptic situations that are accompanied by the occurrence of extraordinary instances of thunderstorms over Central Europe. Data concerning thunderstorms were taken from 51 synoptic stations located in Germany, Poland, the Czech Republic and Slovakia for the years 1951-2010. Atmospheric circulation on thunderstorm days has been illustrated by the distribution of the pressure fields over Europe, both at sea level and at a baric level of 500hPa. In addition, use has been made of information concerning the instability of the atmosphere and of synoptic maps. The first stage of research consisted in separating thunderstorm regions on the basis of the long-term and annual variability of the occurrence of thunderstorms. Next, a synoptic analysis was performed for days in which thunderstorms occurred over the majority of the analysed area or were observed at all of the stations in the identified storm regions. Particular attention was also given to days with thunderstorms that accompanied vast high pressure systems as well as to winter thunderstorms. The aforementioned situations were usually connected with dynamically shifting low pressure systems and the accompanying cold atmospheric front, the appearance of which was preceded by the advection of tropical air masses. As regards winter thunderstorms, we also observed the occurrence of a very strong advection of air from the west or north-west. In turn, thunderstorms connected with high pressure systems occurred most frequently when a low pressure trough appeared over the area impacted by high pressure.

Paper 106:

Impact of storms on society, impact mitigation.

Bransby Bubele Bulo, South African Weather Service


Introduction: In South Africa during summer, normal to above-normal rainfall occurrs mostly over the eastern half, including the central parts but the south-western half receives below-normal rainfall. In January 2009, heavy falls of rain were experienced in the northeastern and eastern parts where damages to houses, business etc were estimated to run into millions of rands. People lost lives in thunderstorms that occurred in areas in the east. Impacts: (a) Lightning- A rescue operation saved 167 miners after they were trapped in the Blyvooruitzicht gold mine near Carletonville on the 17th after a lightning struck an Eskom power substation. The lightning caused an electrical surge to the mining substation, destroying cables and cutting the electrical supply to the shaft. (b)Storms/strong winds- on 02/01/2009 forty-five families were left homeless when a storm accompanied by strong winds destroyed areas of Taung in the North-West. Disater management had to provide the affected families with blankets, food parcels, tents etc. The damage was estimated at more than R2 million.

Paper 107:

Kalman Filter based nowcasting for severe weather elements

Zuohao Cao, Environment Canada


Meteorological hazards frequently occur near surface over a short time period. However, mesoscale numerical weather prediction (NWP) model based short range forecasts/nowcasts have difficulty accurately predicting near surface meteorological parameters mainly due to small scale high impact weather systems and the infancy of nowcasting methodologies. Because a majority of human activities, such as high-level sports competition and airport control, take place in the atmospheric boundary layer, there is a growing concern on inaccuracies in mesoscale model predicted surface fields. This study aims at improving Canadian severe weather nowcasting through correcting mesoscale model forecasts of meteorological parameters by effectively using high resolution in-situ observations at the sites of interest because the site-specific nowcasting (e.g., Toronto international airport) of meteorological parameters has a greater economic value. Since a Kalman Filter (KF) approach can take into account both NWP model errors and observational errors simultaneously, and it can immediately update and produce optimal nowcasting when new observations become available, the KF technique is employed to provide nowcasts of severe weather parameters. This study also aims at establishing the climatology at the sites of interest for better estimation of the observational error covariance and the error covariance in mesoscale model short-range forecasting, and for performance evaluation of the mesoscale models, and the KF system. The detailed results will be presented at the conference.

Paper 108:

Damage survey in The Netherlands after the squall line passage of July 14th 2010

Jacob Kuiper, KNMI


In the late afternoon of July 14th 2010 a squall line crossed the south-eastern part of the Netherlands. At several locations huge damage has been reported due to the accompanied severe weather. Many thousands of trees were uprooted, large greenhouse firms were completely destroyed, houses sometimes severely damaged. In one place a camping site was hit directly, 2 persons fatally wounded, many others sometimes severely injured. Close to that location a part of a high-voltage power line failed, knocking down five iron masts, including the 150 kV power line along a 1 kilometre long path. Never before in Dutch history this type of power lines have been destroyed by ‘wind only’. A KNMI damage survey team travelled through the most severe hit areas very soon after the event. Many eyewitnesses were heard and a large photo-archive (including high resolution aerial imagery, obtained by the National Traffic Police (aviation department)) of this case has been produced. Especially the areas near the small towns of Neerkant and Vethuizen were visited intensively. Also a large amount of images and small video’s of this event were shot by eyewitnesses during the passage of the squall line. Combining all information sources has led to a detailed image of the development of special phenomena along the damaged swath. The work of the damage survey team is shown, including the search for detailed imagery and eyewitnesses. The findings were published in KNMI publication nr. 228, in December 2010.

Paper 109:

Using the Spatial Coverage of Favorable Severe Weather Parameters in Forecasts of Severe Convection

Chad M. Shafer, University of South Alabama

Charles A. Doswell, Cooperative Institue for Mesoscale Meteorological Studies

Michael B. Richman, School of Meteorology, University of Oklahoma

Lance M. Leslie, School of Meteorology, University of Oklahoma

Andrew E. Mercer, Northern Gulf Institute, Mississippi State University


Severe storm forecasters have recognized that most major tornado outbreaks occur on days with collocated favorable vertical wind shear and buoyant instability spanning a large region. To provide an objective test of this notion, we have identified a large number of severe weather outbreak events (6000+) in the contiguous United States using kernel density estimation (KDE), which evaluates the number, density, and clustering of the severe reports received in a given time period. The cases then are ranked based on their perceived relative severity using the scores of a multivariate index, and are categorized as characteristic types (e.g., major tornado, hail-dominant, wind-dominant). Additionally, the region associated with the event of interest is identified using a threshold value of the approximated probability density function from the KDE. This permits evaluation of analyzed or simulated fields of severe weather parameters in the region in which an outbreak occurs. The average or total magnitude of the severe weather parameter in an outbreak region is used as a diagnosis of the outbreak’s severity, and is compared to the relevant ranking index score. The results of these diagnoses and forecasts suggest comparable accuracy and skill with corresponding operational forecasts of the events. Suggestions for implementing these techniques globally are discussed.

Paper 110:

Automatic Recognition of Cold-Ring Pattern on IR Satellite Images

Mateja Irsic Zibert, Environmental Agency of the Republic of Slovenia

Benedikt Strajnar, Environmental Agency of the Republic of Slovenia

Janez Žibert, University of Primorska, Faculty of Mathematics, Natural Sciences and Information Technologies


Many extreme events in Slovenia were characterized with the presence of a cold-ring pattern on the Meteosat SEVIRI IR10.8 image. In this article a system for automatic detection of cold-ring patterns on IR images was designed and tested. The main goal of doing automatic detection of cold-rings was to perform a statistical analysis of how often cold-rings are present over Slovenia and in which cases this is connected with occurrence of severe weather. Secondly, we wanted to evaluate, on a many year data archive, the hypothesis of long lasting cold-rings being correlated with large hail (>3 cm) phenomenon. Automatisation based on pattern recognition techniques is important since it is independent of colorscale, which is very relevant and needs to be adjusted in the case of human recognition of cold-rings. Automatic detection of cold-rings also gives additional information to operational forecasting for issuing warrings and can be included in expert systems of different meteorological sources (radar data, lightning data etc…). Automatisation is also needed in case of large amount of data for processing, which is the case for Meteosat Rapid Scanning with 5-minute update frequency and in the future with the Meteosat Third Generation data which will be available every 2.5 minutes. The system for automatic detection was built based on selected cases in years 2008 and 2009. We defined proper characteristics of cold-ring patterns and tested several different pattern-recognition techniques to adequately model cold-ring and non-cold-ring data. The characteristics of cold-rings were derived from IR images, where the representations of data were made from different feature sets including the typical size of cold-ring, the difference between minimum temperature in cold-ring and maximum temperature in its Central Warm Spot, temperature difference relative to tropopause temperature and other similar properties. In addition also information from HRV during daytime was used.

Paper 111:

Contribution to an European adaptation of the Enhanced Fujita Scale : Analysis of Damage caused by Tornadoes in France

Emmanuel Wesolek, KERAUNOS

Pierre Mahieu, KERAUNOS


The Enhanced Fujita Scale has opened new perspectives in the domain of tornado intensity rating. The approach proposed by the Enhanced Fujita Scale is based on damage indicators and degrees of damage (DOD). All these indicators and DODs have been initially calibrated for the United States territory, and they can not be automatically transposed to the European territory. Indeed, if some damage indicators may be valid in the United States as in Europe (damage to vegetation, to some modern buildings ...), some others are specific to the American continent and thus need to be adapted to European building characteristics. In this perspective, the study analyzes the entire 600 cases of tornadoes recorded to date in France and list all the associated damage. This approach allows to establish a set of typical French tornadic damage indicators, which can be added to the damage indicators already provided by the Enhanced Fujita Scale. Proposals for the classification of each of these indicators in degrees of damage are also detailed.

Paper 112:

Consequences of Tornadic Storms in Urban Areas: Case Study of the Paris Tornado (France) in September 10, 1896

Emmanuel Wesolek, KERAUNOS

Pierre Mahieu, KERAUNOS


Urban tornadoes present a major challenge in terms of safety and economic impact. Even if urban tornadoes are not uncommon in France, they have not been analysed in contemporary studies yet. The tornado that struck the city of Paris in September 10, 1896 is particularly interesting because it concerned the most dense urban landscape in France. This tornado caused extensive damage to buildings, vegetation and street furniture. Five people were killed and over 70 were injured and hospitalised. In order to propose a detailed analysis of this case, the study first deals with the meteorological situation prevailing in the Paris region during this tornadic event, by using surface meteorological observations of that day and reanalysis data from the 20th Century Reanalysis Project (PSD-CIRES). Then, the study determines precisely the path followed by the tornado in the heart of Paris city. The nature of the damage caused to infrastructure is analysed to determine the tornado behaviour in such urban and rugged areas. The characteristics of injuries caused by the tornado are also analysed in order to assess the nature of risk associated to urban tornadoes on the local population.

Paper 113:

The benefits of GLD360 lightning location data in operational weather forecasting

Heikki Pohjola, Vaisala Oyj

Antti Mäkelä, Finnish Meteorological Institute

Nick Demetriades, Vaisala Inc

Nikki Hembury, Vaisala Inc

Ron Holle, Vaisala Inc


Typically, weather radars in operational weather radar networks have measurement ranges of 250 km. This is specifically the case in cold climates, where precipitation occurs in a shallow layer limiting the detection range. The situation is different in the tropics, where the height of the tropopause can be more than 15 km and deep convection with thunderstorms can reach extreme altitudes. Traditionally the combination of radar data and national lightning data enlarges our measurement domain and provides more information on the severity of the weather pattern we are measuring. However, national lightning detection networks also have limited coverage. Vaisala’s Global Lightning Dataset GLD360 can be combined with weather radar data, like national lightning network data, to improve the limited range of weather radar without the limited range of national lightning network data. In this paper we introduce a new way of enlarging severe storm detection that can applied all over the world based on GLD360 data and weather radar. The case studies show that combined GLD360 lightning and radar data could give up to an hour or more of lead time beyond the range of available weather radar measurements. An addition benefit of the combination of lightning and radar data is that it provides more information about the type, severity and life cycle of the phenomena compared to measurements taken with a single instrument. In the second part of the paper we compare GLD360 measurements to the NORDLIS lightning detection network located in Scandinavia. This comparison shows GLD360 performance in northern Europe, which is at the outer edge of the efficient GLD360 coverage area. The results indicate that although the GLD360 detection efficiency is lower than the NORDLIS network (which also detects cloud flashes), GLD360 still provides usable lightning data even at high latitudes.

Paper 114:

Characteristics of cloud-to-ground lightning activity over Portugal in relation to air pollutants

Solange Leite, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

João Santos, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

Margarida Marques, CENSE – Center for Environmental and Sustainability Research - Environmental assessment, monitoring and remediation

Margarida Reis, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

João Sousa, Portuguese Meteorological Institute

Sandra Correia, Centro de Estudos Geográficos, IGOT, Universidade de Lisboa

Marcelo Fragoso, Centro de Estudos Geográficos, IGOT, Universidade de Lisboa


Lightning plays a key role in determining the regional air chemistry. For instance, lightning directly releases NOx throughout the entire troposphere. One main objective of this study is to assess how lightning, a natural source, influences lower tropospheric air pollution in Portugal. Air quality data is recorded on a daily basis by the Portuguese Environmental Agency at eight air quality monitoring stations. Cloud-to-ground (CG) lightning flash data, collected by the lightning detection network installed by Portuguese Meteorological Institute, is used to analyze the mutual effect on lightning activity and air pollutant concentrations over Portugal for the period of 2003-2009. Negative and positive flash density distributions, as well as the number, density, intensity (peak current) and multiplicity of CG flashes are also studied. CG lightning activity is considered in relation to annual, seasonal and monthly averages of concentration of ozone (O3), sulphur dioxide(SO2), nitrogen oxides (NOx) and particulate matter with an aerodynamic diameter smaller than 10 m (PM10). Additional attention is also devoted to the extreme episodes, particularly to two long-lasting lightning episodes: from the 12th to the 18th of June 2006 and from the 9th to the 22nd of September 2007. Flash numbers in these episodes are respectively 10592 and 12053. Both episodes enclose the absolute maximum in 2006 (4317) and in 2007 (4952).

Paper 115:

A seven-year study about cloud-to-ground lightning characteristics in Portugal

João Santos, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

Solange Leite, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

Margarida Reis, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

João Sousa, Portuguese Institute of Meteorology

Sandra Correia, Centro de Estudos Geográficos, IGOT, Universidade de Lisboa

Marcelo Fragoso, Centro de Estudos Geográficos, IGOT, Universidade de Lisboa


A seven-year study of cloud-to-ground (CG) lightning characteristics in Portugal is presented. The study is based on approximately 180 thousand flashes recorded by a lightning detection network from January 2003 to December 2009. The longest available dataset over Portugal is considered. It provides a unique opportunity to study the long-term annual, seasonal, monthly, and daily time distributions of the number, density, intensity (peak current), polarity and multiplicity of CG flashes in Portugal, as well as the spatial variability of these parameters. One of the main results is that the pattern of mean CG flashes clearly divides Portugal into its east-west halves; the eastern half (leeward side) is more continental, with high lightning activity, while the western half (windward side) is exposed to strong maritime influence and generally low lightning activity. In the eastern half, the absolute maximum activity is identified in southern Portugal (over Alentejo), which might be related to the presence of mesoscale convective systems, often located over the region during the summer semester and fed by thermal effects. The maps related to negative polarity are more informative than those related to positive polarity. The maximum over Alentejo is maintained, as well as the reliability parameter and the multiplicity. When analyzing the CG flashes by season, it is found that during the summer semester the maxima occur over northeastern Portugal, a mountainous region with high summertime temperatures. Flash maximum duration time is also found in this region. During winter, northeastern Portugal has almost no electric activity, mostly due to the relatively low wintertime temperatures and to snow cover, whereas the central coastal region experiences a maximum. This study also examines fundamental statistics, such as mean, median, standard deviation, 10th and 90th percentiles of the number, density, intensity, polarity and multiplicity of CG flashes.

Paper 116:

Dynamical characterization of thunderstorm days in Portugal

João Sousa, Portuguese Institute of Meteorology

João Santos, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD

Marcelo Fragoso, Centro de Estudos Geográficos, IGOT, Universidade de Lisboa

Solange Leite, Centre for the Research and Technology of Agro-Environmental and Biological Sciences - CITAB - UTAD


Thunderstorms in the atmosphere are produced in large convection cells (Cumulonim-bus) under conditions of strong atmospheric instability. Their electrical discharges (flashes) may represent a serious threat to the electric power networks, may trigger forest fires and cause injuries and losses of human lives. In Portugal, since 2002, the Me-teorological Institute maintains four detection sensors in order to effectively cover the national territory. Although this database is still very recent, the RAIDEN project (Lightning activity in Portugal: variability patterns and socioeconomic Impacts) intends to systematize all the available data so as to characterize its spatial and temporal varia-bility and to assess the main dynamical mechanisms underlying the occurrence and de-velopment of thunderstorms in Portugal. In this study we have identified two weather regimes that are commonly associated with a significant number of discharges on a daily basis. Considering the 50th percentile (25 discharges per day over Portugal) as the threshold for the selection of thunderstorm days in Portugal, 416 days are keyed to the Frontal (116) and to the Convective (300) regimes. Their synoptic characteristics are then analyzed in greater detail using a number of atmospheric fields provided by the NCEP-NCAR reanalysis project. This analysis clearly highlights their distinct dynamical features and suggests some dynamical predictors for the occurrence of thunderstorms in Portugal. These outcomes are expected to have great applicability not only to weather forecasting, but also to the development of downscaling strategies applied to climate model data.

Paper 117:

Cloud-to-ground lightning activity in Portugal: overall characterization, spatial and temporal patterns of associated thunderstorms over the 2003-2009 period

Marcelo Fragoso, Centro de Estudos Geográficos, IGOT-Universidade de Lisboa

Sandra Correia, Centro de Estudos Geográficos, IGOT-Universidade de Lisboa

Solange Leite, UTAD, CITAB, Physics, Vila Real,

João Santos, UTAD, CITAB, Physics, Vila Real,

João Sousa, Instituto de Meteorologia, Portugal


This study aims to achieve two main objectives. First an overall characterization of the geographic incidence and temporal rhythms of lightning activity in mainland Portugal territory over the 2003-2009 period will be presented and discussed. This summary characterization will focus on the cloud-to-ground (CG) lightning activity, analysing parameters like the total number of flashes, the frequency of negative and positive flashes and peak current intensity (kA). Temporal rhythms and spatial patterns of CG lightning at annual, seasonal and monthly scales will be briefly characterized. The second objective is devoted to the identification of the spatial and temporal patterns of thunderstorm events, based on the associated clusters of CG flashes. Taking into account the large number of days with CG lightning in Portugal over the studied period (754 days, with maximum daily count of 4951 CG flashes), only CG lightning days whose strikes number exceeded the 50 percentile (376 days) were selected for this study. The production of hourly maps of CG lightning allowed the identification of thunderstorm events by the interhourly clustering of CG flashes patterns in space. Considering the mature stage of each thunderstorm event, a classification was made according to the associated CG flashes pattern, distinguishing several categories (four main categories of events: local, multilocal, regional and widespread thunderstorms). Results from this classification of thunderstorm events will be discussed analysing several features (e.g. annual distribution and regional frequency) of each type of thunderstorm event. Finally, case studies of each type of thunderstorm event will be shown as also daily composites of selected atmospheric variables (e.g. sea-level pressure, 500 hPa geopotential heights) that might be relevant to understand the identified CG pattern.

Paper 118:

A close look at a severe mesoscale convective system surrounding the Air France 447 crash

Humberto Barbosa, Laboratório de Análise e Processamento de Imagens de Satélites – LAPIS

Michel Mesquita, Bjerkness Centre for Climate Research

Ivon Da Silva Júnior, Laboratório de Análise e Processamento de Imagens de Satélites – LAPIS


On 1 June 2009, the Air France (AF) flight number 447 crashed in the tropical Atlantic Ocean between 02:00 and 02:15 UTC. An analysis of the severe Mesoscale Convective System (MCS) surrounding the AF 447 flight could provide some insights. No study, to our knowledge, has looked at the degree of severity of the cloud top features associated with this specific MCS. Although we do not engage in any discussion concerning the causes of the AF 447 crash, we will present results that highlight the severeness of the MCS when the accident occurred. In order to investigate the degree of severity of the MCS cloud–top features and its weather characteristics, we have used the multispectral data of Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard Meteosat-9 based on detection of overshooting tops. Two SEVIRI water vapor (WV) band images (6.2 μm and 7.3 μm) and the SEVIRI thermal infrared (IR) band images (10.8 μm) provided information about the structure and the microphysics of the MCS cloudiness. While the Meteosat-9 (Meteosat Second Generation, MSG) measurements do not directly provide information on vertical motions, we focus on the combination of water vapor and infrared window channels to depict overshooting tops at these MCS’s as proxies for convective intensity. Analysis of the difference between the water vapor and the infrared brightness temperature reveals unique details on the severity of the MCS formed in a line west-east across the flight path. WV–IR differences larger than +3.0 °C are associated with deep convective clouds (overshooting clouds) that have a large amount of ice and strong updrafts. Several overshooting tops cooling below -80 ºC were found in our analysis. These were associated with a severe MCS cloud cluster on the route of the flight. A striking feature is that this cloud cluster resulted from the merging of four smaller clusters with cloud top temperature that reached -81°C – this cloud structure could potentially be responsible for the extremeness of the MCS.

Paper 119:

A Tool for Severe Convective Storm Evaluation Using Satellite Infrared and Lightning Network Observations

Humberto Barbosa, Laboratório de Análise e Processamento de Imagens de Satélites – LAPIS

Ivon Da Silva Junior, Laboratório de Análise e Processamento de Imagens de Satélites – LAPIS

Robert Holzworth, University of Washington


This paper presents a tool developed at LAPIS (Laboratory for Analyzing and Processing Satellite Images) for severe convective storm evaluation that combines satellite infrared with lightning network observations. Taking full advantage of the high temporal sampling of Meteosat-9 and its wide gamma of spectral bands, a number of cases of severe weather in Brazil was performed. Locations of lightning discharges are retrieved from the World Wide Lightning Location Network (WWLLN). Both the EUMETCast system – a new C-band satellite reception facility to collect data from Meteosat-9 – and the WWLLN sensor are presently installed at Universidade Federal de Alagoas (UFAL). Moreover this tool, based on open source codes for geographical information systems, written for overlaying locations of lightning discharges over cloud-top infrared radiances, the validity of the evolving convective storm has demonstrated a great flexibility and ease of use. Results show that difference values larger than -20K for brightness temperature differences between water vapor (WV) and infrared (IR) bands are found to correspond well with cloud-ground lightning activity. The study opens up an avenue for successive validation of the analyses together with their improved implementation for operational nowcasting applications.

Paper 120:

Observations from VORTEX2: The genesis of the Goshen County, Wyoming tornado (05 June 2009)

Karen Kosiba, Center for Severe Weather Research

Joshua Wurman, Center for Severe Weather Research

Paul Markowski, Penn State University

Yvette Richardson, Penn State University

Paul Robinson, Center for Severe Weather Research


A dual-Doppler and thermodynamic analysis of the genesis of an EF-2 tornado observed on 5 June 2009 in Goshen County, WY during the VORTEX2 field campaign will be presented. Prior to tornadogenesis dual-Doppler analyses revealed the beginning of a wind surge in the rear-flank downdraft. This surge was followed by the development of a secondary gust front located behind the primary rear-flank gust front and adjacent to the developing tornadic circulation. Through trajectory analysis and the integration of mobile mesonet data, the importance of this change in rear-flank downdraft structure and its link to tornadogenesis will be discussed in the context of vorticity generation and/or transport. Changes in reflectivity structure, including the development of a low reflectivity ribbon, during this time interval also will be examined.

Paper 121:

Rapid-Scan DOW radar observations of tornadoes during VORTEX2

Josh Wurman, Center for Severe Weather Research

Karen Kosiba, Center for Severe Weather Research

Paul Robinson, Center for Severe Weather Research


The rapid-scan Doppler on Wheels (DOW), a 6-beam, multi-frequency passive phased array radar, is capable of simultaneously collecting 6 levels of data in less than 10 seconds. A 2.5 m square slotted waveguide antenna produces a 0.8° beam. Range resolution as fine as 25 m is possible with short pulses and radar gating. This capability makes the rapid-scan DOW ideally suited to study rapidly evolving phenomena, such as tornado vortices. During the VORTEX2 2009 season, the rapid-scan DOW collected data during the 5 June 2009 Goshen County, WY tornado. The tornado passed within 400 m of the DOW. Over 30 minutes of continuous data were collected. Volumetric data were collected every 7 seconds, yielding over 250 volumes of tornado data with simultaneously fine-scale spatial and temporal resolution. The early stages immediately after genesis, intensification and demise of the tornado were documented from 2202 UTC to 2234 UTC. The rapid evolution of the structure and dynamics of the tornado are analyzed. Data from the rapid-scan DOW are interpreted in the context of traditionally produced dual-Doppler analyses from DOW6 and DOW7. Dual-Doppler analysis using the 7-second updates of the Rapid-Scan DOW and the 1 minute low level updates of the DOW7 will be presented. During the 2010 VORTEX2 season, data in several tornadoes were collected with the Rapid-Scan DOW, some in configurations which permit dual-Doppler analysis. Preliminary results from these analyses will be presented.

Paper 122:

In Situ and radar observations of low-level winds in tornadoes

Karen Kosiba, Center for Severe Weather Research

Josh Wurman, Center for Severe Weather Research

Paul Robinson, Center for Severe Weather Research


A DOW radar, several tornado pods, and a manned tornado intercept vehicle (TIV) obtained data in two tornadoes during the spring of 2008. The core flow and near core flow regions of two tornadoes crossing in and near the towns of Stuttgart, AR and Glen Elder, KS were observed. During the VORTEX2 2009 field phase, 11 tornado pods, the TIV and 3 DOW radars collected data in the Goshen County, WY tornado on 5 June 2009. The core flow passed directly over the TIV and within 400 m of the rapid-scan DOW. Additionally, video documentation of real-time damage in tandem with wind measurements allowed for assessment of the wind/damage relationship. A comparison of low-level DOW winds and 1 and 3.5 m in situ winds obtained in the aforementioned tornadoes suggest that strong asymmetries may exist in the lowest levels of the tornado boundary layer with strong inflows in some zones and nearly tangential flow in others.

Paper 123:

Fine-scale Dual-Doppler analysis of the boundary layer in Hurricane Frances (2004)

Josh Wurman, Center for Severe Weather Research

Karen Kosiba, Center for Severe Weather Research

Paul Robinson, Center for Severe Weather Research


Dual-Doppler data were collected by the DOW radars in the northern eyewall and eye of Hurricane Frances as it made landfall near Stuart, FL. A 5.7 km baseline was established south and east of Fort Pierce, FL. The range resolution varied from 13 to 38 m, with the crossbeam oversampled resolution of approximately 30 m in the dual-Doppler lobes. This is one of the smallest dual-Doppler baselines ever established in a hurricane, providing an extremely fine scale three-dimensional analysis of the hurricane boundary layer. The DOWs deployed at approximately 1400 UTC on 04 September 2004 and undeployed at approximately 1100 UTC on 05 September 2004. Landfall of Hurricane Frances occurred at approximately 0600 UTC on 5 September 2004. The DOW data indicate the presences of boundary layer structures of different length scales, most of which are substantially < 1 km. The dual-Doppler domain was chosen to optimize coverage and spatial and temporal resolution. As such, the horizontal domain comprised a 5.5 km x 5.5 km region, with a grid spacing of 20 m, thereby allowing for the resolution of horizontal structures with a characteristic size of O ~ 80m. Furthermore, the domain captured the boundary layer flow on both the windward and leeward sides the barrier island, which allowed for the assessment of the impact of the barrier island on the onshore flow. Three-dimensional analysis revealed updraft/downdraft couplets associated with the boundary layer rolls. Characteristic wavelengths of the boundary layer structure were determined through a 1-dimensional Fourier decomposition. Structures were examined both windward and leeward of the barrier islands, as well as over the entire domain. Horizontal cross-sections were taken through the features at various locations and a characteristic wavelength was determined. The results of these analyses and implications on boundary layer momentum fluxes will be presented.

Paper 124:

Severe storms over eastern region of India, Bangladesh plain and northern Bay of Bengal as observed from the TRMM sensors.

Sanjay Sharma, Kohima Science College

Devajyoti Dutta, Kohima Science College

Rakesh Gairola, Space Application Center


The combined region of eastern India, Bangladesh plain and the northern Bay of Bengal is one of the most favourable zone for the occurrence of severe storms. The complex topography of the region coupled with suitable environmental and synoptic conditions favours the formation of severe storms . The main objectives of the present work are to study spatial and temporal variability of convective activities with the help of Tropical Rainfall Measuring Mission (TRMM) sensors and to further analyse the thermodynamics of convective systems. The study region is stretch between 180 to 280 N and 850 to 950 E . For the present study, different proxies for the severity of the convective systems such as (i). echo top height at 20 dBZ (ETH 20 dBZ) from TRMM - Precipitation Radar (PR), (ii). minimum Polarization Corrected Temperature (PCT) at 85 and 37 GHz from TRMM Microwave Imager (TMI) and (iii). lightening flash rates from Lightening Imaging Sensors (LIS) are used . The utilized data products of TRMM –PR, TMI and LIS are 2A25, IB11 and LIS daily granule respectively. The associated thermodynamics of the storms are analysed with the help of radiosonde sounding data. The study is carried out with the help of data sets during 1998 -2008. The seasonal and diurnal variation of the selected proxies of severe storms are analyzed separately over different terrains of the study region . Distinct characteristics of convective storms are observed over each terrain. The relation between ETH 20 dBZ and PCT85/PCT37 is established over each types of terrain. Though, severe convective systems are found to be associated with high values of CAPE in the range of 4000-5000 J/kg, but one to one relation could not be established. Further work is in progress to identify better environmental indices to be associated with severe storms.

Paper 125:

Observations from VORTEX2: The pretornadic phase of the Goshen County, Wyoming, supercell (5 June 2009)

Paul Markowski, Penn State University

Yvette Richardson, Penn State University

Josh Wurman, CSWR

Karen Kosiba, CSWR

Paul Robinson, CSWR

Jim Marquis, Penn State University


We present an analysis of the pretornadic phase (2110-2148 UTC; tornadogenesis occurred at 2152 UTC) of the Goshen County, Wyoming, supercell of 5 June 2009 intercepted by the Second Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX2). The storm resembles tornadic supercells that have been observed in the past. For example, it develops a couplet of counterrotating vortices that straddle the rear-flank downdraft and are joined by arching vortex lines, with the cyclonic vortex becoming increasingly dominant in the time leading up to tornadogenesis. The outflow in the hook echo region, where sampled, has relatively small negative buoyancy during this stage of evolution. In addition to discussing the evolution of the kinematic fields derived from dual-Doppler mobile radar observations, we will present a Lagrangian analysis of the 3D vorticity vector along trajectories that approach the mesocyclone region. Changes in the orientation and magnitude of the vorticity will be related to barotropic and baroclinic processes along a trajectory, with the latter being much more difficult to diagnose given the challenges of assessing horizontal buoyancy gradients.

Paper 126:

The effects of low-level shear on the intensity and longevity of low-level mesocyclones in simulated supercell thunderstorms

George Bryan, National Center for Atmospheric Research

Leigh Orf, Central Michigan University


Some recent observational studies have found that tornadoes are more likely with strong low-level wind shear (from roughly the surface to 1 km AGL). We have been conducting three-dimensional numerical simulations of supercell thunderstorms to test whether this relationship is reproducible in numerical models, and to evaluate the dynamical processes that might explain the observational studies. In our series of idealized simulations, we use a wind profile that consists of two straight-line segments when plotted on a hodograph. The length of the (relatively) upper-level segment is held fixed for every simulation. The segment of the hodograph at low-levels is varied in several ways (e.g., length, depth, etc). In one especially interesting series of simulations, we vary the angle between the low-level segment and the upper-level segment. Consistent with observational studies, our modeling results show that low-level shear is especially favorable for strong, persistent mesocyclones, and particularly when the low-level shear vector is approximately at right angles to the upper-level shear vector. In other words, hodographs that are “L-shaped” (or “sickle-shaped”) are especially favorable for strong low-level mesocyclones. At the conference, we will discuss two dynamical reasons why this hodograph shape might be favorable: 1) the interaction between cold pools and shear; and 2) the nonhydrostatic pressure perturbations associated with updrafts in shear.

Paper 127:

An assessment of Australian severe thunderstorm environments 1989-2010

John Allen, The University Of Melbourne

David Karoly, The University of Melbourne


Severe thunderstorms present a frequent and significant threat to property and life in Australia during the spring and summer. However, these relatively infrequent events are scattered over large areas, making it difficult to evaluate the meteorological environments conducive to their occurrence using the sparse upper air sounding network in Australia. Adding to this difficulty, the varied and complex problem of initiation means that simply identifying thunderstorms based on the presence of a conducive environment does not necessarily guarantee that severe thunderstorms will occur. A proximity climatology of known severe thunderstorm environments in Australia during 2003-2010 has been used to derive covariate discriminants to identify the potential of an environment to produce severe thunderstorms. Convective variables have been determined from a dataset of 1.5 degree gridded ERA-Interim reanalysis profiles for this purpose. The covariates use variables describing the potential for organization (deep-layer wind shear) and the potential for instability over the depth of the atmosphere (CAPE). This allows separation of environments into several categories stratifying between those conducive to non-severe and severe thunderstorms, and between severe and significant severe thunderstorms. The discriminants were then applied to obtain a climatology representing the frequency and spatial distribution of environments conducive to severe and significant severe thunderstorms for Australia over the period 1989-2010. The inter-annual variability, frequency and spatial distribution of the convective environments over this period are considered and examined in relation to phenomena such as the El Niño-Southern Oscillation. The results of this climatology are also placed into a global context by comparison with the severe thunderstorm environments found in both Europe and the United States over a comparable period.

Paper 128:

Verification of Model-Derived Proximity Soundings for Convective Environments in Australia

John Allen, The University Of Melbourne

David Karoly, The University of Melbourne

Graham Mills, Formerly CAWCR, Bureau of Meteorology, Australia


Model or reanalysis-derived proximity soundings are an effective method of providing atmospheric profiles with regular spatial and temporal intervals. However, their application as an accurate representation of the atmosphere at any one point requires verification against observed datasets. The verification process used involved two steps; the first approach was to compare Australian radiosonde station profiles to a high resolution NWP model, and the second to a reanalysis dataset. As the intention was to assess viability for application as proximity soundings to severe thunderstorm reports, the comparison used variables commonly derived to analyze convective environments. Variables have been determined for 00UTC profiles produced from the Australian Radiosonde dataset of 16 sounding stations, 12.5km gridded MesoLAPS Numerical Weather Prediction analysis and forecast data, and a dataset of 1.5 degree gridded ERA-Interim reanalysis data over the period 2003-2010. The variables include quantities that represent atmospheric instability (Convective Available Potential Energy, Lifted Index, Convective Inhibition), vertical wind shear (magnitude of shear in the lowest kilometre, six kilometres and between one and six kilometres) and the height of atmospheric reference levels, such as the lifted condensation level or equilibrium level. A regression and direct comparison of the model and reanalysis data are presented for the entire database of 4988 soundings, as well as the performance at selected stations to analyze more localized differences. The results from both verifications, and the analysis of differences between the datasets are discussed in the context of factors that are important for the application of these datasets to long-term climatological studies of convective environments.

Paper 129:

Radar Derived Storm Properties Over Central Greece

Dimitrios Bampzelis, Aristotle University of Thessaloniki

Theodore Karacostas, Aristotle University of Thessaloniki


An attempt is made to obtain isolated-storm properties from a C-band weather radar, located close to the area of interest, by using the cell tracker TITAN (Thunderstorm Identification, Tracking, Analysis, and Nowcasting - NCAR). The study aims in building a storm climatology over central Greece, as a tool for investigating possible rain enhancement feasibility potential. The objective on this study is two folded: firstly, to analyze and describe storm properties -and secondly- to identify storm properties differences among the seven distinct synoptic situation patterns, which had been established for the examined area of central Greece. The TITAN algorithm is used to obtain storm properties during the months April to September, for the five-year period of 2006 to 2010. The storm days were distributed into the seven distinct synoptic situation patterns, in an attempt to identify differences in storm properties, structure and behavior among them. The radar-based storm properties were compiled by using the 35dBz radar reflectivity and the minimum storm lifetime of 15 minutes as the TITAN storm-tracking thresholds. Following this approach, the total storm days and the isolated storms, that affected the area, were identified. The storm days were classified into the seven distinct synoptic situation patterns, using the 500 hPa and the surface charts. Isolated storm properties include: storm initiation time, duration, direction, speed, mass, volume, area, precipitation area, total area covered, cloud base, cloud top, reflectivity and precipitation flux. Frequency analyses of the aforementioned findings bring out evidences of the storm characteristics, their extends and limitations, providing thus an integrated view of the experienced isolated storms over the examined area. Moreover, the classification of the storm days reveals certain differences in several storm properties indicating how isolated storms over central Greece vary under different synoptic situation patterns.

Paper 130:

Study of Two Severe Storm Events That Affected the Major Area of Thessaloniki, Greece

Dimitrios Bampzelis, Aristotle University of Thessaloniki

Nicolas Tzonichakis, 3D S.A.

Theodore Karacostas, Aristotle University of Thessaloniki


Two different storm events that occurred over the major area of Thessaloniki are analyzed, studied and compared. The one took place during the warm season (August 2009), and the other during the transitional period (October 2010). The main objective on this paper is to integrate the storm characteristics, derived: from radar reflectivity measurements, synoptic charts and thermodynamic diagrams. The cell tracker TITAN (Thunderstorm Identification, Tracking, Analysis, and Nowcasting - NCAR) approach s used to analyze the radar reflectivity measurements and derive storm properties from both storm events. Data used include successive weather reflectivity measurements obtained at 3,5 min intervals from the C-band weather radar located 10 km northeast of the target area. The radar-based storm properties are derived from radar reflectivity measurements and they include: storm duration, direction, speed, mass, volume, area, precipitation area, cloud top, reflectivity and precipitation flux. Upper air (500 hPa) and surface weather charts, developed from ECMWF analyses, and radiosonde information from Thessaloniki synoptic station, are used for the synoptic, dynamic and thermodynamic analyses and studies of the examined storm events. Both events exhibited prolonged duration and caused at least twice amount of precipitation from the monthly mean. The events vary in several characteristics, mainly based upon the developing time-period (August and October) and the triggering mechanism. Synoptic and dynamic analyses reveal an upper air short wave trough to trigger the August event, while the October well organized and deep barometric low was associated with the surface frontal activity. The thermodynamic instability indices appear to be more intense during the August event. Spatial and temporal distribution of the several storm properties is also performed. This study can act as a useful tool for possible mitigation measures and policies regarding severe storm events over the major area of Thessaloniki, Greece.

Paper 131:

First operational results of RELASE (Rainfall Estimation from Lightning And SEviri data) software at CNMCA

Davide Melfi, Italian Air Force Weather Service - CNMCA

Francesco Zauli, Italian Air Force Weather Service - CNMCA

Daniele Biron, Italian Air Force Weather Service - CNMCA

Massimiliano Sist, Selex Galileo


First results of precipitation retrieval from RELASE (Rainfall Estimation from Lightning And SEviri data) software are here presented. The RELASE software was born during a collaboration that aimed for studying a possible use of lightning data in hydrological field between CNMCA (Centro Nazionale di Meteorologia e Climatologia Aeronautica - Italy) and SELEX-GALILEO (a Finmeccanica company). The software’s rainfall retrieval technique that use geostationary satellite Infrared (IR) observations and lightning information retrieved from LAMPINET (lightning network of the Italian Air Force Meteorological Service) is here explained and a comparison with HSAF (EUMETSAT Satellite Application Facility on Support to Operational Hydrology and Water Management) products is also presented. In addition, considering that for the next generation of geostationary meteorological satellites (Meteosat Third Generation - MTG) an optical Lightning Imaging (LI) mission is planned, the use of the RELASE software on simulated data is evaluated.

Paper 132:


José María Sánchez-Laulhé, AEMET


The objective of this presentation is to show synoptic and mesoscalar environment aspects associated with a local event of heavy rainfall in the south of the Iberian Peninsula on 16 August 2010, exceeding 200 mm in 5 hours and originating a significant flash flood, by using observational data and European Centre for Medium-Range Weather Forecasts (ECMWF) operational analyses and forecasts. The event was associated with the poleward transport of deep tropical moisture along the weatern flank of the upper and middle level summer anticyclone over North Africa. The interaction between a positive potential vorticity anomaly in high levels associated to a middle latitude trough, and a negative potential vorticity anomaly associated with the tropical moisture band originated a jet streak and exported the tropical moisture to the Iberian Peninsula. On the synoptic scale the heavy precipitation occurred where this tropical moisture intersected a region of forced ascent situated beneath the equatorward jet streak-entrance region. On the mesoscale the event was focused in a zone of orographic wind convergence. The synoptic environment was similar to that of the climatology one associated to predecessor rain events (PREs) that occur over the United States east of the Rockies approximately 1000 km poleward of recurving tropical cyclones.

Paper 133:

Impact of Radar Data Assimilation on Numerical Simulation of a Severe Storm in Croatia

Antonio Stanesic, DHMZ

Keith Brewster, Center for Analysis and Prediction of Storms


A severe thunderstorm hit the north-western part of Croatia in the late afternoon and evening of 24 June 2008. Strong wind gusts and hail were observed and there were even reports of a small tornado. Hail was the size of a hazelnut (~13 mm) or in some locations the size of an egg (~50 mm). The storm caused significant damage to crops, buildings and cars. The storm was initiated in Austria and traveled southeast through Slovenia and Croatia reaching as far as Bosnia and Herzegovina. This severe event was used as a test case for the Advanced Regional Prediction System (ARPS) high resolution numerical prediction model and as testbed for exploring the impact of assimilating conventional and radar data. Nested model grids were used with horizontal grid spacing of 24 km (ARPS24), 8 km (ARPS8), and 2.5 km (ARPS2.5). Results show that without data assimilation, the models were unable to represent the development of the storm nor the proper environment for it. Assimilation of SYNOP data in ARPS8 sets up proper environment for storm development. However, because the resolution is too coarse ARPS8 does not properly resolve the storm development. Using ARPS8 with assimilation of SYNOP data as a driving model enables ARPS2.5 model to produce more realistic (but still not satisfactory) storm development. To achieve better results high resolution data must be assimilated. This is done via assimilating radar data from the Bilogora radar, located in northern Croatia. Radar radial velocity was assimilated using three-dimensional variational analysis (3DVAR). Radar reflectivity data were used through a cloud analysis procedure where hydrometeors and cloud fields are defined, and adjustments to the in-cloud temperature and moisture fields are made. The best results were obtained by assimilating radar data in two sequential steps 15 minutes apart.

Paper 134:

Automatic Hazard Level Approximation of Convective Storms Using Real Time Emergency Data

Pekka J. Rossi, Finnish Meteorological Institute

Vesa Hasu, Aalto University, Department of Automation and Systems Technology

Antti Mäkelä, Finnish Meteorological Institute

Jarmo Koistinen, Finnish Meteorological Institute

Heikki Pohjola, Vaisala Oyj


Issuing warnings on severe convective weather is nowadays a part of operational weather services. Thanks to the development of real time accurate observations and computer-based models, operative forecasters have access to numerous automatic severe weather nowcasting and warning tools that facilitate their work. However, these methods do not include reported damages that have already taken place due to severe weather. In this paper, we study automatic real time hazard level determination of convective storms using a new information source: real time emergency reports. During severe convective storms, emergency call centers log a large number of reports, for example, due to flash floods, lightning damages and uprooted trees. Since 2006, Finnish Meteorological Institute has received these reports from the emergency call centers in real time. In addition to the location of the emergency, the reports contain a coarse classification of the emergency type and a short verbal description of the incident for on-line use. We propose attaching this information automatically to weather radar detected storms to characterize their hazardous properties. The proposed method uses a weather radar based convective storm tracking algorithm in the background. Detected tracks of individual storms and incoming emergency reports are analyzed to determine the relationship between each report and a convective storm. Then, the method estimates the hazard level for each storm based on number of associated emergency reports. Finally, based on the hazard levels, we can highlight potentially dangerous convective storms in nowcasting products. As the flow of emergency reports is more intense in densely populated areas, the estimated hazard level is scaled using the population density of incoming emergency report locations. Moreover, with the history information from the tracking algorithm, storms preserve the estimated hazard level over several time frames. The potential of the method is demonstrated through several case studies in Finland.

Paper 135:

Long-term variability and distribution of thunderstorms in non-mountainous regions of Bulgaria (1961 – 2010)

Liliya Bocheva, National Institute of Meteorology and Hydrology - BAS

Tania Marinova, National Institute of Meteorology and Hydrology - BAS

Petio Simeonov, National Institute of Meteorology and Hydrology - BAS


The aim of the presented study is to investigate thunderstorm activity and distribution of stormy days in non-mountainous regions in Bulgaria during the period 1961 - 2010. The frequencies of thunderstorm days with different categories of precipitation versus the total thunderstorm days are also analyzed. Daily observational data for thunderstorm and precipitation from 70 meteorological stations in Bulgaria is used. These stations are representative for all geographical regions of the country with altitude below 1000 m. Five daily precipitation categories for the thunderstorm days have been defined as follows: Light precipitation (class A) 0.0-4.9 mm; Light-Moderate (B) 5.0-14.9 mm; Moderate-Heavy (C1) 15.0-29.9 mm; Heavy (C2) 30.0-59.9 mm, and Torrential (D) ≥ 60.0 mm. During the investigated period, no statistically significant tendency is found in the monthly, annual and decadal distribution of thunderstorm days almost all over the country. The values do not differ much from the typical ones for different geographical and climatic regions in Bulgaria. The annual frequencies of the defined precipitation categories are similar in almost all regions in Bulgaria. Analysis of the time variability of these categories shows a significant increase (with about 60%) of the cases from the heaviest category D during the last 10 years. Recently, thunderstorm days with torrential precipitation more frequently occurred in the end of the summer (in August and September).

Paper 136:

Analysis and forecasting waterspouts on Lake Constance

Marco Stoll, MeteoSwiss

Gabriela Kaufmann, MeteoSwiss


Waterspouts are observed globally along coastal maritime waters as well as on inland lakes. In Lake Constance area, rather than being a threatening phenomenon, they primarily cause significant attention and reporting in media. At the same time, physical-meteorological processes that cause waterspout formation in this region are not well understood and attempts to forecast their occurrence have not been made on a regular basis, therefore. In this study observational and numerical model data is analysed for a small number of recent cases with the aim to shed some light on the meteorological conditions that favour waterspout formation. Special emphasis is taken on the role of mesoscale flow fields which are strongly influenced by the Alpine orography. An attempt is made to set up basic forecasting recipes and compare them with techniques applied in other waterspout-prone areas.

Paper 137:

Trusted Spotter network Austria

Thomas Krennert, ZAMG, Central Institute for Meteorology and Geodynamics, Vienna, Austria

Alois Holzer, ESSL, European Severe Storms Laboratory e. V., Wessling, Germany

Michael Staudinger, METEOPICS, ZAMG

Mathias Stampfl, SKYWARN Austria

Christian Ortner, ZAMG


The TRUSTED SPOTTER NETWORK AUSTRIA TSN is a cooperation between the Austrian meteorological service ZAMG (www.zamg.ac.at), SKYWARN AUSTRIA (www.skywarn.at) and the ESSL with its EUROPEAN SEVERE WEATHER DATABASE ESWD (http://www.eswd.eu). METEOPICS (www.meteopics.eu) provides a public forum for images of severe weather and damage surveys. A “trusted spotter” is a member of SKYWARN AUSTRIA, providing reports about significant or severe weather and consecutive damages to the Austrian national weather service ZAMG. Since these reports were formerly delivered by fax or email, the usability for immediate response by the forecaster was limited so far. Also, the reliability of the information suffered from the anonymity of these reports. In order to avoid such difficulties the TSN was established to build up a reliable network between spotters (chasers) and operational forecasters. For this purpose ZAMG offers an individual training program for spotters, regular workshops and scientific support. Further, the activity of a “trusted spotter” is facilitated by real time weather information from ZAMG, easily accessible via the internet. An interface also allows the display of the reports within the ESWD data base and an additional web based real time platform for operational forecasting at ZAMG. The reports are strictly following the ESWD data format and threshold guidelines. The platform is provided by ESSL / ESWD, web based and hence independent from visualization tools and operating systems. Further, reports from “trusted spotters” are accepted by ESWD with QC1 clearance. Since ESWD does not operate a forum for severe weather images and photos of succeeding damages, METEOPICS generously offered a sub domain of its homepage specifically designed for contributions from TSN and ESWD as an official image data base. The operational forecaster is now able to use reliable weather information from TSN in real time to evaluate and verify warnings during severe weather periods and to conduct adjustments to warnings if necessary, joint case studies can be swiftly released to the public.

Paper 138:

Assessment of the hail hazard from a combination of different meteorological datasets and insurance data

Marc Puskeiler, Karlsruhe Institute of Technology (KIT)

Michael Kunz, Karlsruhe Institute of Technology (KIT)


Severe thunderstorms and extreme events such as hail often lead to significant damage to buildings, crops and critical infrastructure. In the federal state of Baden-Württemberg, for example, most of the damage to buildings is caused by large hailstorms (1986-2008). Due to the high damage potential, quantifications of the hail hazard and risk as accurate as possible are essential for the economy, especially for the insurance industry. Within the frame of the project “HARIS-SV” (Hail Risk SparkassenVersicherung), it is aimed at quantifying the hail hazard for different parts of Germany in a high spatial resolution. Due to the local-scale impacts of a few hundred meters to some kilometres only, hailstorms and their intensities are not captured accurately and uniquely by a single observation system. Therefore, several appropriate meteorological datasets (3D / 2D radar, lightning, satellite and radiosounding data) are used and combined to identify tracks of single hailstorms in the past. To identify events with loss-relevant hail, loss data from insurance companies are included in the analysis. The hailstorm tracks are projected on an equidistant grid (e.g., 10 x 10 km2) to quantify their local density and their intensity. Application of extreme value theory (peaks-over-threshold method and generalized Pareto distribution function) then allows for estimating the intensity as a function of the probability or return period. The results reveal a high spatial variability of the intensity and probability of hail tracks that can be (partly) explained by orographic flow modifications. In the future, a hail loss model will be created to convert measured and modelled intensities (e.g., radar reflectivity) into monetary parameters like mean loss or maximum loss. From that, it is will be possible to quantify the local-scale hail risk for certain return periods.

Paper 139:

Remote sensing information on convective precipitation in Poland – radar and ATS datasets as a validation sources for H-SAF satellite precipitation products – convective case study analysis.

Rafal Iwanski, Satellite Remote Sensing Centre Institute of Meteorology and Water Management National Research Institute


Proper understanding, interpretation and application of various precipitation information sources is crucial for human economy. Precipitation events inflicting water management and resulting in flood danger are more and more common in Poland as we are facing progressing climate changes. Correct recognition and validation of satellite precipitation products is in focus of attention for both meteorologists and hydrologists. Network of over a thousand Automatic Telemetric Stations (ATS) collects precipitation information from all over the country in the near-real-time mode. Meteorological radar network consists of eight devices covering whole Poland and providing unified precipitation field using different precipitation products from which Surface Rainfall Intensity (SRI) is used on operational basis. Both systems provide quality spatial and temporal distribution of meteorological information for purposes of scientific studies as well as operational meteorology. The main goal of EUMETSAT Satellite Application Facility in Support to Operational Hydrology and Water Management (H-SAF) is to provide satellite products in near real time mode to be useful for operational hydrology. Among them, the pre-operational precipitation products based on both passive microwave sensors (conical and cross track scanning) and IR sensors calibrated by MW have been available since 2009 for cooperating teams for detailed validation before release of operational products. One of the products, PR-OBS-3, is to be validated with use of both radar and ATS datasets. In the presentation, the ability of the H-SAF PR-OBS-3 product to reproduce the convective precipitation patterns will be discussed on the base of quantitative and temporal relations with radar and ATS precipitation information datasets (as ground based sensors) within chosen case analyses. Radar precipitation information used for the very first time to validate H-SAF satellite product in Poland will be introduced. Finally, the quality of the satellite products in estimation of convective precipitation accuracy will be presented.

Paper 140:

Trend analysis of meteorological parameter relevant to hail from soundings and reanalysis data

Susanna Mohr, Institute for Meteorology and Climate Research (IMK-TRO), Karlsruhe Institute of Technology (KIT)

Michael Kunz, Institute for Meteorology and Climate Research (IMK-TRO), Karlsruhe Institute of Technology (KIT)


Severe thunderstorms and associated extreme events such as hail represent a substantial hazard potential for buildings, crops, and critical infrastructure. In the last decades, damage caused by severe hailstorms has significantly increased in Central Europe. In the light of global warming, the questions arise whether the thunderstorm potential of the atmosphere already has changed and what changes will be expected in the future. Within the frame of in the project ‘Haris-CC’ (HAil RIsk and Changing Climate), changes of convective conditions are analyzed for past and future periods using both observational data and regional climate models (RCM). Because hailstorms are not captured well by current observation systems, the examination will be based on larger-scale proxy data. According to statistical analyses of various European soundings, the convective potential of the atmosphere has changed significantly in the last decades. Most of the convective indices that were found to be relevant for hail events derived from insurance loss data show a trend towards more intense convection within the last two decades. In Germany, the change in the trend direction occurred later in the northern parts (mid-1990s) compared to the south (beginning 1990s). By varying the time series with respect to their start and end date, the stability of the linear trends and their significance was tested. Furthermore, past hailstorms are described within a multi-dimensional parameter space of appropriate meteorological indicators (e.g., convective parameters, moisture flux convergence, wind shear, large-scale weather patterns), obtained from soundings and reanalysis data. By comparing combinations of appropriate proxies with hail losses provided by a building and agriculture insurance company, it is aimed at describing hail events at best. The methods will be applied to an ensemble of high resolution RCM simulations. Future changes of the hail hazard will then be projected from the current hazard and the climate change signal.

Paper 141:

Severe convective precipitation at Pula on 25th September 2010 - NWP model simulations

Martina Tudor, DHMZ

Antonio Stanesic, DHMZ

Stjepan Ivatek-Sahdan, DHMZ

Kristian Horvath, DHMZ


A case with the severe precipitation in the night and early morning on 25th September 2010 in Istra, Croatia, is used as a testbed for the NWP model ALADIN. The operational ALADIN forecast at Meteorological and Hydrological Service of Croatia is run with 8 km horizontal resolution, it uses ARPEGE initial and boundary conditions and digital filter initialization. The parallel suite has similar characteristics but it uses initial conditions obtained from data assimilation cycle. Forecasts obtained from operational and parallel suites show that the precipitation event west of Istria peninsula was forecasted by both suites even several days in advance, but high precipitation maximum at southern part of the peninsula was not captured. Possible reason for that is that observed severe precipitation was caused by convective activity supported by synoptic conditions and/or local conditions that were not represented correctly in the initial conditions or the model was not able to represent its development. In order to investigate this, different sets of experiments have been performed. These tests include the sensitivity to different initial and lateral boundary conditions (driving models: ARPEGE or IFS LBC and data assimilation), model resolution and convective parametrization scheme. Several experiments using high-resolution (2 km) non-hydrostatic ALADIN model runs have been performed. The experiment where the prognostic parametrization of convection has been used have indicated the existence of the secondary maximum over the Istrian land. Most of the precipitation was given by the convection scheme. This result suggests the importance of using the convective parametrization even in the resolutions in which it is assumed that the convection is resolved. Further experiments, with a diagnostic convection scheme and without any deep convection scheme will be performed in order to investigate this problem further.

Paper 142:

Convection Initiation – Nowcasting by data fusion and its Verification

Dennis Stich, DLR

Caroline Forster, DLR

Tobias Zinner, LMU München

Arnold Tafferner, DLR


This study aims at combining different data sources appropriately in order to detect convection initiation (CI) and to extend the nowcasting horizon of thunderstorms to the CI stage. Results for classical verification measures like probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI) are presented with regard to CI for DLR’s satellite-based thunderstorm detection and nowcasting algorithm Cb-TRAM. Based on the question if pixel based scores are appropriate for the problem of CI nowcasting, object based and fuzzy verification methods are also taken into account. For the object based method, a hit is counted, if an analysed Cb-TRAM polygon overlaps a nowcast polygon. The fuzzy method makes use of a search radius around the polygon, growing with lead-time of the nowcast to allow little nowcast positioning errors. The object based method is found to be most reasonable for the task of CI verification. The measures are calculated for nowcast lead-times, and additionally for the timeframe of up to one hour lead-time as accumulated hits, misses etc. It is shown that Cb-TRAM is able to detect CI in some cases, but fails in other cases. The object based accumulated measure results in a POD of 0.409 and a FAR of 0.545. The question now is whether the combination of Cb-TRAM with additional data sources containing information on CI improves Cb-TRAM’s ability to identify CI. Therefore, in future work, data sources like surface observations analyzed with VERA and model simulations by COSMO-EU and COSMO-DE are examined and will be combined with Cb-TRAM, if useful signals are found. The presentation outlines the concept of this approach. Verification measures for the combinations will be evaluated in comparison to those for Cb-TRAM alone in order to quantify the possible improvements of the scores resulting from the fusion of the data sources.

Paper 143:

Climate change impacts on severe thunderstorms in Finland

Andrea Vajda, Finnish Meteorological Institute

Jenni Rauhala, Finnish Meteorological Institute


This study provides an estimate of the possible changes in frequency and intensity of severe thunderstorms in Finland under one of the projected future climate scenario. Our tool for this study is the high-resolution (25 km) regional atmospheric model (RCA3) developed by the Rossby Centre. The model runs follow the SRES A1B emission scenario. Two 30-year time periods were analyzed: a 30-year control run describing the present climate conditions (1971–2000) and a scenario run corresponding to 2071–2100. The tornado and large hail observations are compared to the outputs of the control run to define severe weather probabilities within specific storm environments. We use the convective available potential energy (CAPE) and vertical wind shear as the quantitative measures characterizing the storm environment. The environmental thresholds for delineating severe thunderstorm occurrence are derived from the values of CAPE and wind shear in the present climate. The probability of any severe weather, large hail or tornado events in specific storm environments in the current climate as well as the changes in their frequency and intensity under the projected future climate conditions are analyzed.

Paper 144:

An extratropical precipitation cell: numerical simulation of heat and energy conversions

Julia Palamarchuk, Black and Azov Seas Hydrological and Meteorological Centre

Sergiy V. Ivanov, Odessa Environmental University

Galina P. Ivus, Odessa Environmental University

Alexander S. Matygin, Black and Azov Seas Hydrological and Meteorological Centre


The reliable forecast of heavy precipitation is substantial component in preventing socioeconomic losses and risk assessment. Numerical simulation is the appropriate way for evaluation and prediction of those weather events and their outcomes, which is broadly used in forecasting services all over the world in their daily releases. The focus of this work is to demonstrate the ability of a high resolution model (MM5) in reproducing mesoscale special and temporal variations of heat and energy fluxes within an extratropical precipitation cell developing behind a cold front. The fine vertical resolution is used in order to detect atmospheric layers where generation, intensification and decaying of ascending motions occur. A mask approach allows to clearly outline both the structure and size of the cell. Quantitative estimations of heat and mass transport associated with the convective circulation as well as exchange between the low and middle troposphere are obtained. The results demonstrate the sensitivity of simulated humidity and precipitation fields to a choice of the planetary boundary layer and microphysical processes parameterization schemes on both the coarse and fine resolution domains. Thus, the numerical modeling provides better understanding of thermodynamics and energy conversion occurred in extratropical mesoscale convective cells.

Paper 145:

Convective instability indexes as thunderstorm predictors for Catalonia

Montserrat Aran, Servei meteorològic de Catalunya

Nicolau Pineda, Servei Meteorològic de Catalunya


Convective instability indexes (CII) are a useful tool to evaluate the potential instability of the atmosphere and can be used as thunderstorm predictors. However, its effectiveness depends on, among other things, the chosen threshold and the region. In order to optimise the skills of the CII to diagnose and predict thunderstorms, the Servei Meteorològic de Catalunya (SMC) has started a study to obtain optimal thresholds for the region of Catalonia (NE Spain). The CII database (2004-2010) has been built using the Barcelona radiosonde data at 00 and 12 UTC. Thunderstorm activity was monitored by the total lightning location system of SMC. The “thunderstorm day” used as a predictand was defined as a day with at least 10 cloud-to-ground flashes. Following the work done by Haklander and van Delden (2003) for the Netherlands, different verification parameters such as the True Skill Statistic, the Heidke Skill Score or the Rank Sum Score have been applied to assess the forecast skill of the CII. Different tests have been carried out to achieve our goal. Firstly, about 70 indexes and parameters (obtained from the RAOB program) were analyzed, resulting in a selection of 50. Secondly, a deeper analysis was carried out for the selected indexes, looking to its temporal evolution. Following the seasonal variations found, different seasonal thresholds were defined. Furthermore, the thunderstorm activity in the region concentrates in the warm season (May to October), so the following studies have focused on this half-year period. Finally, the skill of the indexes was evaluated choosing different forecast periods (3, 6 and 12 hours).

Paper 146:

Implementation of the INCA system in the Czech Hydrometeorological Institute

Hana Kyznarova, Czech Hydrometeorological Institute

Petr Novak, Czech Hydrometeorological Institute

Lucie Brezkova, Czech Hydrometeorological Institute


Nowcasting and very short range forecasting of meteorological phenomena have been studied for several decades but they are still far from being solved. Today there exist many forecasting systems of different complexity. The INCA system was developed originally by Austrian meteorological service (ZAMG), which made it available for other meteorological services. Implementation of the INCA system in the Czech Hydrometeorological Institute (CHMI) began last year within INCA-CE project of the CENTRAL EUROPE Programme co-financed by the European Regional Development Fund. The INCA is a complex forecasting system, which uses surface station measurements, remote-sensing data, NWP forecasts and detailed topographic data. It includes time-dependent blending of extrapolated quantities with the forecasts of NWP model. This paper describes current status of the INCA system implementation in the CHMI. It includes use of surface station measurements, radar data, NWP ALADIN model forecasts and hi-res topographic information. Precipitation analysis and forecasts are computed every 10 minutes and analysis and forecasts of other meteorological quantities (e.g. temperature, humidity, wind) are updated hourly. All analyses and forecasts are done on 1x1km grid. Visualization of the INCA analyses and forecasts in the CHMI will be done via web-based application which is currently under development; preliminary version is already available. Preliminary results are positive, and further testing is planned in operative use in the CHMI. The paper also shows the first results of quantitative evaluating of INCA precipitation forecast as an input into a hydrological model for a flash flood case. INCA precipitation forecasts are compared with other precipitation nowcasting systems developed and used in the CHMI.

Paper 147:

Observations of a supercell storm structure, its visual appearance and relationship with severe weather

Jenni Rauhala, Finnish Meteorological Institute


During the late afternoon and evening of 28 June 2009, a supercell thunderstorm moved over central eastern Finland. The storm moved from north to south causing vast damage during its over 5 hour life time. The hail was up to 8 cm in diameter and downbursts caused F1 wind damage with two injuries. A unique observational material of the event was collected, including 33 hail reports, 18 wind damage reports from Emergency Response Center and the public as well as several photos of both hail and damage. Additionally, aerial photographs of the wind damage taken from low-flying aircraft were available, and photographs of the visual appearance of the thunderstorm at different stages of storm development were obtained from storm spotters. The material includes also a unique video footage of starting downburst and gustnadoes over a lake. The storm passed close to the Kuopio Doppler-radar, which allowed detailed analysis of storm development and its features during and before severe weather. The radar analysis showed supercell features such as mesocyclone, bounded weak echo region and a persistent hook echo. Collapsing reflectivity cores were observed prior to the major downbursts. The storm went through a mesocyclone occlusion during its lifetime and its structure resembled a hybrid storm with characteristics of both supercellular and multicellular storms. This presentation offers an analysis of storm development and a comparison of the reflectivity structures with both the visual appearance and the observed severe weather. Furthermore, the usefulness of a new large hail nowcasting tool and hail algorithm is tested.

Paper 148:

Detection of Rapidly Developing Cumulus Areas from MTSAT-1R Short-Time-Interval Images

Akira Sobajima, Meteorological Satellite Center/Japan Meteorological Agency

Izumi Okabe, Japan Meteorological Agency

Takahito Imai, Meteorological Satellite Center/Japan Meteorological Agency

Yasushi Izumikawa, Meteorological Satellite Center/Japan Meteorological Agency


The Meteorological Satellite Center (MSC) of the Japan Meteorological Agency (JMA) is developing a nowcasting satellite product called ‘rapidly developing cumulus areas (RDCA)’. The objectivity of this product is to provide aviation user with the information of severe weather expected to be evolved into thunderstorms earlier than the similar information generated from rain radar observations. RDCA is produced from short-time-interval images over the same observing area captured by an infrared and visible imager on board a geostationary satellite. The RDCA algorithm is developed in MSC consulting the EUMETSAT’s Convective Initiation product. First, cumulus clouds are selected by not only examining the values of brightness temperatures and reflectivities but also employing an index called ‘slope index’ indicating updraft strength. Second, the clouds in the developing stage of their lifecycle are extracted. For the extraction, the temporal variations of brightness temperatures and reflectivities at the clouds in the satellite images are examined by taking into account for the movements of the clouds. Third, clouds at the peak of the lifecycle are also extracted. Parameters for RDCA are tuned by using MTSAT-1R five-minute-interval images captured in August 2010, during which many severe thunder storms were occurred over Japan. Rain radar observations and the lightning observations of the lightning detection network system (LIDEN) are also used as references. The skills of RDCA are about 0.6 for the probability of detection (POD) and about 0.3 for the threat score (TS) measured by using the same images used in the parameter computation. MSC/JMA plans to develop the algorithm, recalculate the parameters by populating more satellite observations and perform independent validation. In the conference, the latest results will be presented.

Paper 149:

Synoptic conditions of heavy snowstorms in Europe

Ewa Bednorz, Adam Mickiewicz University


Synoptic conditions of heavy snowstorms were analyzed for five European stations: Oslo, Smolensk, Poznań, Bremen and Budapest. Circulation patterns responsible for snowstorms in each location were shown on composite maps of the SLP and 500 hPa geopotential heights means and anomalies. Maps were constructed for the days with snowstorms resulting in daily increase in snow cover depth by at least 10 cm, separately for each of five stations. Similar contour maps of precipitable water content and temperature at the geopotential level of 850 hPa were constructed. Furthermore, back trajectories of air masses for the chosen days with the most effective snowstorms, were presented. Snowstorms in Europe are associated with negative anomalies of sea level pressure and 500 hPa heights, which means low pressure systems over the analyzed region. The lows differ in the intensity and localization. In the south (Budapest) heavy snowstorms appear as a result of fronts in the colder parts of Mediterranean cyclones, the centers of which are often situated over the Italian Peninsula. Another typical location of cyclones or troughs of low pressure is the Baltic Sea region, where meteorological fronts appear quite frequently and may cause snowstorms in western part of East Europe (Smolensk). The cyclonic activity over the continent, which brings very snowy weather in Poznan and Bremen, is simultaneous to the weakening of Icelandic Low and significant positive pressure anomalies over the north Atlantic, which indicates the negative phase of North Atlantic Oscillation. Positive NAO phase expressed by negative pressure anomalies over the North Atlantic, with a low pressure centre over the Norwegian Sea may cause snowstorms in Scandinavia. Very snowy conditions require negative temperature anomalies, which are strongest when snowstorms appear in the south, central or western Europe.

Paper 150:

Evaluation of the convection-allowing model HARMONIE for heavy precipitation events over the Western Mediterranean

Ana Genoves, AEMET

Javier Calvo, AEMET


An accurate forecast of convective weather systems, and specially those related to high impact weather, continues being a major challenge of weather forecasting. Nowadays, the increasing resolution of Numerical Weather Prediction (hereafter, NWP) models, the improvement of observing systems and the use of more adjusted physics have allowed the improvement of the skill of predictions. Large scale weather systems forecasts are remarkably better some days in advance, when compared to the past of weather forecasting. Nevertheless, despite some advances have been made in recent years by means of the use of SREPSs and probabilistic forecasts, the characteristic temporal and spatial scales of convection and related weather (wind gusts, heavy rain or tornadic winds) make them unpredictable from these NWP models. Another approach for producing reliable and accurate forecast can be the use of convection-allowing models. This is a topic of very active research. Moreover, some National Weather Services are already using convection-allowing models for operational forecasts. Those models can be the base for producing real and accurate nowcasting systems. Since some time ago, AEMET is running, in daily non-operational mode, the HARMONIE system, a non-hydrostatic model, of 2.5 km grid spacing. The behaviour of the model and the first results of verification against observations for a Western Mediterranean case of heavy rain are showed in this study.

Paper 151:

Detection of cyclone attributes by Hofmöller diagrams

Miloslav Müller, Institute of Atmospheric Physics

Marek Kaspar, Institute of Atmospheric Physics


The climate in midlatitudes is substantially influenced by frequency and attributes of extratropical cyclones passing across the region. Usually, tracking of cyclones is performed by detection of SLP/geopotential minima or vorticity maxima. Nevertheless, there are other anomalies of meteorological variables which also deserve our attention: divergence and vertical velocities, moisture fluxes, temperature gradients etc. Our previous studies confirmed that significant anomalies in specific variables are a common feature of cyclones producing heavy rains in Central Europe. Since the Hofmöller diagram depicts changes in a characteristic along a line during a period, it seems to be a useful tool for description of the development of cyclones existing several days. Two approaches are presented: (i) Eulerian analysis of cyclones passing across a fixed area; (ii) Lagrangean analysis of individual cyclones. The combination of both approaches enables us to recognize e.g. the magnitude of the studied anomalies, their duration, velocity of the cyclone movement, and direction of the cyclone track. Among others, Mediterranean cyclones moving to the Central Europe along the well-known Vb-track can be easily detected and analyzed by Hofmöller diagrams. Our presentation is based on reanalysis data. We present attributes of several cyclones which produced severe weather in Europe, with the main emphasis on heavy rains and subsequent floods. We also demonstrate typical patterns which can be recognized in the Hofmöller diagrams in case of such a cyclone. Finally, we discuss possible differences in these patterns related to heavy rains and floods within various regions.

Paper 152:

Forecasting tornado – one case lesson

Adam Jaczewski, Institute of Meteorology and Water Management

Piotr Baranski, Institute of Geophysics, Polish Academy of Sciences

Jan Parfiniewicz, Institute of Meteorology and Water Management


This work focuses on tornado predictability on the basis of multi-scale synoptic analysis of the 8-minute tornado case in Poland. The analysis is carried out with use of 3D Virtual Laboratory concept – Vis5d and a number of tools like the one to produce reflectivity composites and Doppler retrievals. It was recognized that the dominant driving process for the abrupt convection over Europe was an inflow in the upper troposphere of a cold arctic air over a warm and humid tropical air. Deep massive drop of cold Arctic air started moving over the Atlantic from Greenland towards the British Isles, eventually modifying the formation of the Atlantic branch of the polar jet stream. After 8 days of the cold air movement over the North Atlantic, the Atlantic branch of the jet stream extended north-easterly over Europe parallel to elongated tongue of hot and moist tropical air masses on its east side. 3D inspection of the potential temperature topography form the NWP model outputs shows quite significant slope of 315 K isentropic surface – the upper cold front. And this cross-isobaric sliding over hot subtropical air beneath must have been responsible for the abrupt convection. The W-E vertical cross-section of potential temperature, with overlapping radars reflectivity at 16 UTC demonstrates that a process of this type generates the so-called convective alias potential instability and releases a strong convection. The predictability analysis concerned the potential of particular components of the system, i.e., NWP model output, radar and lightning data. The first macro stage, comprising several days of the pre-convective period and the second one-day forecast stage developing over Europe, which provide necessary growth of instability are hoped to be well predicted by our operational models. However the 3rd stage lasting for e few hours and precede supercell formation has been recognized as difficult to predict. Taking into account the steps and results of our reanalysis listed below: 1.restoring the typical wind jump and related vortex tube in the examined supercell event by the COSMO model application, 2.providing a realistic approximation of the tornado funnel oscillation by the assimilation of Doppler radar data, i. e., the rotational wind component, 3.observing the time derivate of the total lightning rate and the occurrence of considerable abrupt jump of the time derivate of total lightning rate preceding by approximately 5 minutes the moment of visible tornado appearance being a good indicator of increasing storm severity, it is reasonable to expect that similar synoptic tools and procedures and continuous monitoring of the total lightning activity in Poland by the SAFIR/PERUN system should be included into routine short-term predictions and warnings of extremely severe thunderstorm hazards such as tornado or downburst or heavy hail or rain gush incidents connected with supercells or mesoscale convective systems (MCS).

Paper 153:

The use of stability indices for severe storms prediction in Poland

Adam Jaczewski, Institute of Meteorology and Water Management

Jan Parfiniewicz, Institute of Meteorology and Water Management

Bogumil Kois, Institute of Meteorology and Water Management

Barbara Brzoska, Institute of Meteorology and Water Management


Since 90s, the rise in the number of dangerous phenomena, including severe storms is observed in Poland. An important role in their development plays deep convection phenomenon, which can be described by the stability indices like CAPE, Showalter and others. The use of as most as possible number of indices is desired because there is no one index which can be perfectly connected with storm occurrence. For the CAPE, results of calculations by few methods will be compared. The scope of the work is to develop the prediction system for storm weather. As a main tool the operationally working COSMO NWP model is used and 48 hours forecasts with 1h steps are postprocessed to obtain atmospheric profiles (geopotential height, temperature, pressure, humidity, wind direction and speed) and surface data connected to severe phenomenon (precipitation type and amount, maximum wind speed, pressure tendency). As the first step, chosen stability indices are calculated from COSMO's atmospheric fields for 2011 spring and summer season. Then, the indices' values are verified with upper-air data from Polish and surrounding stations. For the verification, the data from the system for lightning detection and localization (PERUN) and information on storm magnitude deduced from surface synoptic observations will be also used. For that, for the COSMO model every grid point, the most representative station will be selected on the basis of optimal interpolation method and the information enclosed in SYNOP report will be converted to the value from originally elaborated 8-numbers storm scale. The verification results for the prediction system will be presented for individual indices and advantages and disadvantages of using them will be reported.

Paper 154:

Mesoscale features of severe hailstorm days classified in clusters

Eduardo García-Ortega, University of León

Andrés Merino, University of León

Laura López, University of León

José Luis Sánchez, University of León


The forecast of precipitation using NWP models is an open problem. Various parameterizations used by cloud resolving models provide different results, with higher or lower scores. Hailstorms in particular are associated to convective phenomena that are highly localized in time and space. The triggering and subsequent development of convective storms is linked to mesoscale as well as to synoptic factors. The Ebro Valley is one of the regions in Europe with the highest frequency of hailstorms. The topography and the climate are crucial in the formation of convective storms and the short-term forecast of these phenomena is still a challenge. The GAP database (Group for Atmospheric Physics, University of León) on hailstorms between 2001 and 2008 was used by García-Ortega et al. (2011) to establish a classification of the synoptic patterns related to hail events in the study area. The results were five clusters with marked differences. The MEV (Middle Ebro Valley) is a small area and the intra-cluster synoptic variability leads to different spatial precipitation patterns which are highly influenced by the local topography. The present study will focus on the days with more intense and extensive precipitation events. The main aim is to establish a relationship synoptic cluster classification-mesoscale features in those days. Using the MM5 we have studied the factors that enhance the development of convection from a mesoscale perspective, taking into account atmospheric fields such as convective instability, water vapor convergence or the low-level flux. The results show well-defined patterns with clear differences between the various clusters. The weight and the spatial pattern of the three factors are different in the days belonging to the different clusters and leading to different spatial patterns precipitation intra and inter cluster. García-Ortega, E., Lopez, L., Sánchez, J. L. doi: 10.1016/j.atmosres.2010.08.023

Paper 155:

A GPM simulator to improve the NWP of severe events

Eduardo García-Ortega, University of León

Francisco Javier Tapiador, University of Castilla-La Mancha

Laura López, University of León

José Luis Sánchez, University of León


Precipitation forecasting is a challenging task. On the one hand, physics formulation and parameterization schemes in NWP models are still not precise enough to obtain a correct spatial and temporal description of precipitation events, especially when covection is present. In such cases, increasing the spatial resolution of the model does not a guarantee of improved results by itself. On the other hand, ground-based precipitation measurements suffer for well-known uncertainties. Better precipitation estimates are instrumental to improve the forecast of severe storms using NWP models, as it has been proven that assimilation of precipitation into NWP models improves the forecast. The forthcoming Global Precipitation Measuring (GPM) Mission will provide, for the first time, adequate sampling of precipitation in high latitudes using an orbital radar. The mission will provide global retrievals of precipitation, with 1-2 hour sampling over land thanks to the mixed non-sun-synchronous/sun-synchronous orbits of the constellation and the inclusion of MW sounders. This configuration will expand the TRMM capability of direct sensing of precipitation in higher latitudes. The spatial resolution of the estimates will range from 5 km for the DPR to 7 to 32 km (depending on operating frequency) for the GMIs. Precipitation products from GPM will be more accurate and suitable for assimilation into NWP models. However, parametric algorithms for the GPM mission rely at its turn on better modelling at high spatial resolution. Here, we present a case study of intensive and extensive precipitation patterns in Northeastern Spain simulated with WRF. Three-nested domains were defined, the inner domain covering the entire Iberian Peninsula with a horizontal resolution of 3 km. Test simulations have been performed with a radiative transference code, and some examples are shown.

Paper 156:

The use of NOAA/ATOVS satellite data for convective storms nowcasting and early warnings

Bozena Lapeta, Satellite Remote Sensing Centre, Institute of Meteorology and Water Management, National Research Institute, Krakow Branch


Extreme convective precipitation events resulting in casualties and economical losses are more and more common in Central Europe. Therefore its forecasting and nowcasting are significant elements of early warning system allowing limitation of the natural disasters effects. However, despite the fast and strong developments in numerical weather prediction techniques, forecasting of convection processes is still very difficult. The application of atmospheric instability indices may allow for improving convection prediction skills. The main source of information about the atmospheric instability are temperature and water vapour profiles. As the ground sounding stations network is rather sparse and does not allow for analyzing the atmospheric stability over the regions, the satellite data are commonly used for this purpose. The Advanced TOVS (ATOVS), onboard of NOAA and Metop satellite, includes the High Resolution Infrared Radiation Sounder Version 3(HIRS/3), the Advanced Microwave Sounding Unit-A (AMSU-A) and the Advanced Microwave Sounding Unit-B (AMSU-B) and measures radiation in the infrared and microwave spectral channels, what allows for temperature and water vapour profiles retrievals with 1K accuracy. In the paper, the application of NOAA/ATOVS data for atmospheric instability indices calculations and storms nowcasting is presented. K Index (KI) and TT Index (TTI)as well as Total Precipitable Water derived from ATOVS data were used with collocated data from Polish lightning detection network PERUN (Vaisala SAFIR 3000 system) to define the values range that may be used in storms nowcasting operational system. The obtained results were verified for selected cases of heavy storms occurrence.

Paper 157:

Variants of meteorological conditions during large-scale rain floods

Marek Kaspar, Institute of Atmospheric Physics ASCR

Miloslav Müller, Institute of Atmospheric Physics ASCR


While convective torrential rain can cause severe flash floods in small streams, intense large-scale precipitation produces significant rain floods in major rivers. An array of studies has confirmed the close relationship between the large-scale rainfalls and circulation conditions in synoptic scale in midlatitudes. The aim of the presented work is to distinguish the main variants of the synoptic-dynamic conditions of the rain floods in various parts of Central Europe using a uniform methodology which makes a quantitative comparison of the flood events possible. The individual sets of the most significant flood events, which occurred in the Czech Republic and in selected major rivers in Germany and Austria, are compiled from the period 1961-2010. The selection criterion takes into account the mean daily discharge or both the aggregate area of affected catchments and the return period of peak flows where available. Synoptic-dynamic conditions are described in terms of the strength of anomalies of meso-alpha scale which regularly occurred over Europe and the Northern Atlantic for the duration of causal precipitation. In order to detect meso-alpha anomalies in various dynamic and thermodynamic variables, reanalysis data with the horizontal resolution of 2.5 deg. are employed. An anomaly is defined as a contiguous area of statistically outlying values compared to their area related climatology. The strength of the anomaly is assessed by its size and by the values of distribution function. Cluster analysis is applied on each set of the events using the strength of the anomalies as a similarity criterion. The circulation characteristics of the most important clusters and meso-alpha anomalies typical of them are introduced and discussed. The presented approach opens the door, for instance, to the detection of significant flood events from the past using only reanalysis data which can be beneficial when other data are not accessible.

Paper 158:

Mechanisms producing heavy precipitation event in the southern Adriatic area

Branka Ivančan-Picek, Meteorological and Hydrological Service

Kristian Horvath, Meteorological and Hydrological Service

Natasa Strelec Mahovic, Meteorological and Hydrological Service


Extreme precipitation event occurred in the morning hours on 22 November 2010 over Dubrovnik, Croatia and the surrounding area of the southern Dinaric Alps mountain range. The rain gauge measurements in Dubrovnik exceeded 160 mm / 24 h, with a peak intensity of 71.5 mm/h. The event caused severe flash floods, landslides, interruption of traffic and electricity as well as other infrastructural damage. The synoptic environment was characterized by a deep cyclone moving from northern Europe over the Alps and farther east. The associated frontal system covered most of the central Europe and the mid-Mediterranean. The mechanisms responsible for the formation of convection have been analyzed though synoptic measurements, satellite data and numerical experiments performed with the WRF model, which was set up at the convection-resolving resolution in the innermost domain. The numerical simulations highlighted the essential role of southerly low-level jet stream (LLJS) in the transport of warm and moist air towards the affected area. In addition, numerical sensitivity experiments shaded light on the role of the complex orography of the southern Dinaric Alps on the initialization and development of convection over the target area.

Paper 159:

Analysis of hail storm characteristics obtained by radar, shot-and long-term impact

Svetlana Stevanovic, Republic Hydrometeorological Service of Serbia, Kneza Viseslava 66, Belgrade,Serbia

Mira Stepanovic-Nikolic, Republic Hydrometeorological Service of Serbia, Kneza Viseslava 66, Belgrade,Serbia

Julijana Nadj, Republic Hydrometeorological Service of Serbia, Kneza Viseslava 66, Belgrade,Serbia


Some hail storm characteristics (observed by radar); certainly indicate in shorter period of time that will be damage in area of moving storm. These characteristics can be traced/analyzed using the reflectivity products, differential reflectivity products and radial velocity products. Analyzing certain characteristics could contribute to mitigation the impact of the storms. Storm case happen on 07 august 2005 with damage in rural area. This paper considering the characteristics in the field of reflectivity, differential reflectivity and radial velocity and damage caused in a short period of time during the life time of this storm.

Paper 160:

Data collection in system of hail suppression and analysis of thunderstorms and damage provided by them

Svetlana Stevanovic, Republic Hydrometeorological Service of Serbia, Kneza Viseslava 66, Belgrade,Serbia

Mira Stepanovic-Nikolic, Republic Hydrometeorological Service of Serbia, Kneza Viseslava 66, Belgrade,Serbia

Julijana Nadj, Republic Hydrometeorological Service of Serbia, Kneza Viseslava 66, Belgrade,Serbia


Observation of cloudiness this service is permanent during the year, but activity provided on hail suppression are conducted during the period from 15 April to 15 October. This period coincides with most common occurrences of thunderstorms development in Serbia, which implies with collection of data about weather condition, radar parameters of observed cloudiness, atmospheric phenomena and caused damage. Hail suppression process is aimed at reducing damage from hail storm, but the completion of any action and any eventual damage is analyzed in particular report. As an example how system operate with this kind of data, has been taken one year period from 15 April to 15 October in smaller part of territory of Serbia, which represents the area of the intensive agricultural production.

Paper 161:

Windstorm Xhyntia: local effects and operational aspects in Basque Country case.

Santiago Gaztelumendi, Tecnalia - Meteo Unit / Basque Meteorology Agency

Joseba Egaña, Tecnalia - Meteo Unit / Basque Meteorology Agency

Ivan R. Gelpi, Tecnalia - Meteo Unit / Basque Meteorology Agency

Kepa Otxoa De Alda, Tecnalia - Meteo Unit / Basque Meteorology Agency

Roberto Hernandez, Tecnalia - Meteo Unit / Basque Meteorology Agency

David Pierna, Tecnalia - Meteo Unit / Basque Meteorology Agency


At the end of February 2010, a strong extratropical cyclone named Xynthia brought hurricane-force winds and high waves to Western Europe. 58 deaths were reported in France, Spain, Portugal, Belgium, Germany and England. Most of the deaths occurred in France as a consequence of rapid rise of water and inundations in west coast. In this paper we present, focusing on the Basque Country Area, an analysis of Windstorm Xynthia event. A study and local characterization of this severe weather episode, considering Basque Meteorology Agency (Euskalmet) operational aspects was made. Data collected in Basque Country Automatic Weather Stations Network (BCAWSN) and other data available in the area including numerical models are presented. We also present some Euskalmet operational aspects related with forecast and severe weather warning issues and actions during this severe weather episode. In the Basque country, hurricane wind gusts were recorded in various locations across the region. A 226 kilometers per hour gust was recorded in a mountainous area in the interior. These are some of the strongest winds observed since records are available from BCAWSN owned by the Basque Government.

Paper 162:

A case study of convective severe storms episode in Basque Country.

Santiago Gaztelumendi, Tecnalia - Meteo Unit / Basque Meteorology Agency

Joseba Egaña, Tecnalia - Meteo Unit / Basque Meteorology Agency

David Pierna, Tecnalia - Meteo Unit / Basque Meteorology Agency

Ivan R Gelpi, Tecnalia - Meteo Unit / Basque Meteorology Agency

Roberto Hernandez, Tecnalia - Meteo Unit / Basque Meteorology Agency

Javier Lopez, Tecnalia - Meteo Unit / Basque Meteorology Agency

Kepa Otxoa De Alda, Tecnalia - Meteo Unit / Basque Meteorology Agency


In this paper a study of a convective summer event in Basque Country is made. During this day heavy storms were produced all over the area, some single cells affect many points in Basque Country, in particular Vitoria-Gasteiz city was affected by a extraordinary big size hail episode as a consequence of a rapid evolving deep convection system that crosses the city. We analyzed different aspects of the environment characterization during this day, focusing on synoptic and mesoscale aspects and the analysis of data sets coming from the Automatic weather station network and the different imagery products from a Dual Doppler radar available in the area and MSG. During the afternoon of July 1st 2009, many shower storms are formed all over the area, due to the heat accumulation all day long. The environment is characterized by a relatively high degree of thermal instability, large content of water vapor in surface and other factors that favor the development of deep convective cells, with a slowly eastwards displacement. As a consequence, at some places all over the Country, heavy precipitations are registered. In the well-populated area of Vitoria-Gasteiz city hail size over 5 cm diameter causes many incidents and promotes important material losses. During this day in the radar imagery it can be seen a single active convective cell with large vertical extension and reflectivities over 60 dBz.

Paper 163:

Analysis of a heavy rainfall episode in the Basque Country focusing on total lighting detection.

Javier Lopez, Tecnalia - Meteo Unit / Basque Meteorology Agency

Santiago Gaztelumendi, Tecnalia - Meteo Unit / Basque Meteorology Agency

Joseba Egaña, Tecnalia - Meteo Unit / Basque Meteorology Agency

Kepa Otxoa De Alda, Tecnalia - Meteo Unit / Basque Meteorology Agency

Roberto Hernandez, Tecnalia - Meteo Unit / Basque Meteorology Agency

Ivan R Gelpi, Tecnalia - Meteo Unit / Basque Meteorology Agency


In this work we present a study of a heavy rainfall episode that took place over the Biscay coast (Basque Country) in September 2009, with special focus on the analysis of data obtained from total lightning detection systems available in the area. We include synoptic and mesoscale analysis and some information from other data sources as weather radar, MSG and automatic weather stations network. In September 16th 2009, a cut-off low pressure system crossed over the Basque Country following the NW direction, it turned back on next day and placed over the Basque Country, moving towards the south of France later on. As a consequence of this dynamic instability situation and humid north wind in low layers, some locations in the Biscay coastline were affected by heavy precipitations and flash floods. Particularly heavy rainfall in September 18th affected mainly the area of the village of Bermeo, in the Basque Country seaside, close to Cape Matxitxako. The rainfall lasted for hours and left intense and very heavy rain, over 50 mm rain in one hour and over 220 mm in the whole episode. During this event, total lightning detection was performed using the two lightning detection networks available in the area. These networks are a LF/VHF network and the LF/VLF sensors of Linet network, each of them operating with four sensors in the territory of the Basque Country. The LF/VHF network reported 269 cloud-to-ground lightning discharges in a 10 km radius area centred in Bermeo. These locations were performed using the combination of TOA and MDF technologies in this network. 263 of these events were time-correlated to the detections performed by the LF/VLF network sensors, which represents a high correlation level.

Paper 164:

Windstorm Xhyntia: overview and comparison with other explosive ciclogenesis cases affecting Basque Country area.

Joseba Egaña, Tecnalia - Meteo / Basque Meteorology Agency

Santiago Gaztelumendi, Tecnalia - Meteo / Basque Meteorology Agency

David Pierna, Tecnalia - Meteo Unit / Basque Meteorology Agency

Ivan R Gelpi, Tecnalia - Meteo / Basque Meteorology Agency

Roberto Hernandez, Tecnalia - Meteo / Basque Meteorology Agency

Kepa Otxoa De Alda, Tecnalia - Meteo Unit / Basque Meteorology Agency


In this work we present a study of Xynthia event, focusing on main characteristics of its formation and evolution including some local aspects and a comparison with other severe windstorms cases that had affected the Basque Country area in the past. On February 27th 2010 the zonal circulation is undulated creating favorable conditions for the cyclogenesis at the southwest area of the Iberian Peninsula. The generated cyclone deeps quickly, with pressure values compatible with the definition of explosive cyclogenesis, reaching pressure minimums of 967 mb. The system moves north-eastward into the Cantabric Sea on the next days. Xhyntia quickly travels over Bay of Biscay, affecting Basque Country area on late 27th early 28th January, surpassing the French coast at 06 UTC of 28th. We present some aspects related with this severe weather episode, including synoptic and mesoescale features. We focus on comparisons with other similar events that have affected Basque Country. Particularly we compare with Klaus and Martin events, two relatively recent and similar episodes that in the Xynthia case. The Xynthia trajectory differs in a significant way from Martin and Klaus trajectories. Klaus and Martin moves from west to east, while Xynthia have a trajectory southwest-northeast, moving over the sea at the west of Iberian Peninsula, when the minimum pressure values of 967 mb is reached the cyclone is located in the vicinity of Galicia region, moving finally towards the north-east part of France over Cantabric Sea. This displacement is similar to February 1941 deep cyclone, with hurricane-force south wind, causing the Santander city big fire. In Xynthia case minor fires are produced too.

Paper 165:

Meteorological analysis of the extraordinary hailstreak on 26 May 2009

Michael Kunz, Karlsruhe Institute of Technology, Institute for Meteorology and Climate Research

Jan Handwerker, Karlsruhe Institute of Technology, Institute for Meteorology and Climate Research

Susanna Mohr, Institute for Meteorology and Climate Research (IMK-TRO), Karlsruhe Institute of Technology (KIT)

Marc Puskeiler, Karlsruhe Institute of Technology, Institute for Meteorology and Climate Research

Bernhard Muehr, Karlsruhe Institute of Technology, Institute for Meteorology and Climate Research

Manuel Schmidberger, Karlsruhe Institute of Technology, Institute for Meteorology and Climate Research

Rainer Langner, Vereinigte Hagelversicherung, Gießen


On 26 May 2009, between 1200 and 2100 UTC, parts of Switzerland and southern Germany were hit by a severe hailstorm of exceptional severity, associated with significant damage to buildings, crops, and infrastructure in the order of several hundred Million Euro. The large-scale convective system was triggered near the Lake Geneva and moved over south Germany to the Czech Republic, yielding a hail swath approximately 600 km in length. The convective system is thoroughly analyzed by combining different observation data (3D radar reflectivity, surface stations, soundings, lightning data, satellite data, loss data from insurance companies, and model analyses), complemented by additional model studies with the non-hydrostatic limited-area model COSMO. According to the MAX-CAPPI composite derived from 6 different radars, the fast moving storm (70-80 km h-1) featured several forward and left-flank developments of new cells, triggered by a significant gust front. Dynamically, it resembled the famous Munich hailstorm on 1984 with a complex development, evolving as a multicell, through a supercell, and developing into a mesoscale convective system and complex (MCS/MCC). The mechanisms and meteorological conditions most decisive for maintaining the convective activity of this storm are analyzed and discussed. Furthermore, it is found that radar reflectivity considering the height of the 55 dBZ signal shows a good agreement to the damage data, both in the location and intensity.

Paper 166:

A study of instability indexes in summer severe storms case in the Basque Country area.

Joseba Egaña, Tecnalia - Meteo Unit / Basque Meteorology Agency

Santiago Gaztelumendi, Tecnalia - Meteo Unit / Basque Meteorology Agency

David Pierna, Tecnalia - Meteo Unit / Basque Meteorology Agency

Kepa Otxoa De Alda, Tecnalia - Meteo Unit / Basque Meteorology Agency

Roberto Hernandez, Tecnalia - Meteo Unit / Basque Meteorology Agency

Ivan R Gelpi, Tecnalia - Meteo / Basque Meteorology Agency


In this work is presented a study about the correlation between several instability indexes and the severe summer storm situations, focus on the usefulness supporting the actions of operational weather forecast for the Basque Country area in this type of episodes. In the Basque Country severe storms take place usually from May to September (summer period). In this work the cases are selected basing on a precipitation rate threshold criteria and lightning activity presence. As representative for severe storms cases summertime events with precipitations that exceed ten millimetres in any rain gauge from the Basque Country AWSN are studied in the last years. To characterize severe storm events, an analysis of the most significant parameters, that have influence in the development or activity of convective cells, will be made. The most used instability indexes in the Basque meteorology agency operational tasks for the forecast of these events will be taken into account. The most useful instability indexes or those which warn better the risk of these extreme events in the Basque Country case are shown. Among others TTI (Total Total Index), LI (Lifted Index) and CAPE (Convective Available potential energy) have been analyzed. In the same way is studied some index combination in order to improve the representation of probability of severe storm occurrence in a deep convection environment. We must consider that sometimes, severe storms are generated outside from the study area, and later moves crossing the Basque Country, mainly from southwest to northeast.

Paper 167:

8-year statistical analysis of 3D radar storm tracks

Edouard Goudenhoofdt, Royal Meteorological Institute

Laurent Delobbe, Royal Meteorological Institute


The knowledge of convective storms occurrence and characteristics over Belgium remains relatively limited. These phenomena, which occur at a spatial scale of a few kilometers, can be well observed by weather radar, which provide 3D reflectivity scans at high spatial and temporal resolution. The complete volume data of 2 C-band radars (from 2003 and 2005 respectively) is archived at the Royal Meteorological Institute of Belgium. An automatic cell identification and tracking system (TITAN) has been used to generate a convective storm database. A sensitivity analysis of TITAN's parameters have been conducted to optimise its performance. After several quality checks on the database, storm tracks are classified into different storm categories based on their characteristics such as shape and intensity. In a first step, different statistical properties are extracted from the convective storm database. The 8 years time-series of convective activity will be analysed through frequency analysis and cycle analysis (seasonal, diurnal). The spatial variability of convective storms initiation and evolution will be analysed and its possible factors tested (sea/land and orographic effect). The distribution of aggregated and instantaneous storm track properties (e.g. area, echo-top, max reflectivity) will be described and fitted to probabilistic distribution models. The temporal and spatial variabilities of these characteristics will also be shown. Possible linear and non-linear relations between pairs of storm track properties will be assessed. In a second step, the dynamic of convective storms will be analysed (i.e the evolution of storm properties from growth to decay and their impact on storm duration). Possible applications of this study include satellite communication attenuation mitigation and regional climate model verification. However, the main motivation of the work is to develop precipitation nowcasting techniques (e.g. probabilistic models of storm evolution) that can be integrated in the new INCA-BE nowcasting system.

Paper 168:

Severe hail size discrimination using dual-polarized weather radar data. A dual- wavelength comparison between C and S band.

Rudolf Kaltenboeck, Austrocontrol, Vienna, Austria; IMGI, University of Innsbruck, Austria and CIMMS, University of Oklahoma, Norman, USA

Alexander Ryzhkov, Cooperative Institute for Mesoscale Meteorological Studies (CIMMS) - University of Oklahoma and NOAA/NSSL – Norman - OK, USA


Severe, large hail inflicts the most damage to property and is a serious risk for aviation. The detection and forecasting of very large hail is still one of the main topics in nowcasting, where weather radar data are used extensively to monitor the three-dimensional storm structure. One of the advantages of dual-polarization weather radars is their capability to discriminate between different precipitation types and nonmeteorological echo. Therefore, many European countries are implementing polarimetry into their operational networks, which offers new possibilities to improve existing algorithms for hail detection and determination of its size. Polarimetric characteristics of hailstones depend on their size, shape, falling behavior, and are strongly affected by the degree of melting and the probing radar wavelength. Strong attenuation and differential attenuation in hail at C band further complicates the issue of hail detection / sizing. In this study, severe hail cases in Oklahoma / USA are investigated by analyzing the data simultaneously collected by two closely located polarimetric weather radars operating at S and C bands. One of them is the OU-PRIME C-band radar operated by ARRC / University of Oklahoma and another one is a polarimetric prototype of the WSR-88D S-band radar (KOUN) belonging to NOAA / NSSL. Polarimetric radar variables measured in the presence of hail at C band are quite different from the ones at S band due to more pronounced effects of resonance scattering and much stronger impact of attenuation. The differences are particularly strong in melting hail below the freezing level, but they can be substantial even at higher altitudes where hail is dry or grows in wet regime. As a consequence, the algorithms for hail detection and determination of its size developed at S band can not be directly applied to C band. In this study, we utilize surface hail reports and S-band polarimetric measurements as a “ground truth” for defining the rules for hail detection and rough estimation of its maximal size at C band assuming that hail detection is done reliably at S band. The differences between vertical profiles of radar reflectivity Z, differential reflectivity ZDR, and cross-correlation coefficient ρhv in hail-bearing parts of the storms are in the focus of our analysis. It is shown that in the presence of hail, ZDR(C) is usually higher than ZDR(S) and ρhv(C) < ρhv(S). The height of radar resolution volume with respect to the freezing level has to be taken into account in polarimetric hail detection / sizing. It is also demonstrated that giant hail is commonly associated with pronounced depression of ρhv in the areas of hail generation above the freezing level and the corresponding drop in ρhv at C band is much stronger than at S band. Vertical profiles of polarimetric variables show different behaviors and developments in hail- bearing cells with respect to different observed hail sizes at ground and to the height of hailstones relative to the melting layer. These offer the possibility to distinguish between areas of high reflectivity with and without hail and to discriminate between smaller-size and giant hail. Secondary polarimetric radar signatures which are caused the three-body scattering, side-lobe contamination, and depolarization commonly associated with severe hail will be also addressed in this presentation.

Paper 169:


Mauricio Antonio, São Paulo State University


This study presents evaluations by one S-band weather radar of heavy rain that resulted in severe floods in an urban watershed (150 km2) located just over 200 km from the radar, and in a small urban basin (4 km2) in Bauru, Brazil, near the radar site (22.3583o S; 49.0272o W). In the first case, the basin of Ribeirão das Anhumas, located in Campinas, had a flood on February 17th, 2003 which resulted in significant property damage and six people dead. The disruption of three rural dams maximized damage from flooding. The rainfall that caused the flooding lasted about three hours with values from 110 to 120 mm in rain gauges of the area. The radar images and accumulations showed the highest rain rates hit the south border of the watershed. In the second case, the area was hit by a convective rain event during less than two hours on November 30th, 2010, which resulted in significant damage (mainly flooded cars) and one person dead by a flash flood. The radar data were collected every 7.5 minutes and the rain quantification was made using two distinct Z-R relationships: for Campinas for intense rain and long distance (>180 km): Z = 5 R1.6, and for Bauru for intense rain: Z = 17 R1.54. The radar accumulations were compared with the measurements of surface rain gauges showing differences between values from 4% to 20%, in both cases. These differences do confirm using a specific Z-R relationship for individual situations and distances. The discrepancies between the results do recommend using radar data for the rain quantification in remote areas. The rain data from radar monitoring used as input in flood forecast models can help the anticipation of flood episodes, making possible previous warnings to save lives and properties, and reduction of damage.

Paper 170:


Mauricio Agostinho Antonio, São Paulo State University

Carlos Alberto Agostinho Antonio, São Paulo State University


A set of seven rain gauge stations were selected to characterize the precipitation regime at the extreme west region of the State of São Paulo, Brazil, in an area named “Pontal do Paranapanema”. After the filling of the dam of Porto Primavera HPP (10 km wide and more than 120 km long) in November-December 1988, it was verified a change in the rainfall regime, with decrease of 9% in the historical annual amount, from 1350 mm to 1230 mm. After 1988, there were also changes in the contribution of each month. It can be detached the months of January and February when both combined had 26% and changed to 33%, however the contribution of the rainy period (October to March) maintaining the value of 70% as a contribution to the annual precipitation. Using the data from one S-band weather radar of the region, it was verified as cause of the precipitation increase during January and February the occurrence of squall lines in the neighborhood of the dam. As samples of squall lines were selected the events of January 23 and 31, 2007. In those cases, the whole of the precipitation areas was over 2500 km2, lasting up to 5 hours, and with accumulated precipitation reaching nuclei exceeding 50 mm. In both occasions the precipitation echoes extended along more than 100 km, standing at a distance of 5 to 10 km from the lake borders, and were moving eastward, up to a distance of 40-50 km, and were confined to the “Pontal” area. The radar reflectivity values found were exceeding 50 dBZ in isolated nuclei, with tops above 13 km. These kinds of squall lines were not identified during the remainning months of the rainy period in the area and are associated to the meteorological circulation.

Paper 171:

Storm-scale ensemble forecasting for the NOAA Hazardous Weather Testbed

Fanyou Kong, Center for Analysis and Prediction of Storms, University of Oklahoma

Ming Xue, Center for Analysis and Prediction of Storms, University of Oklahoma

Keith Brewster, Center for Analysis and Prediction of Storms, University of Oklahoma

Xuguang Wang, Center for Analysis and Prediction of Storms, University of Oklahoma

Kevin Thomas, Center for Analysis and Prediction of Storms, University of Oklahoma

Yunheng Wang, Center for Analysis and Prediction of Storms, University of Oklahoma

Steven Weiss, Storm Prediction Center/NCEP, NOAA

Adam Clark, Natioanl Severe Storm Laboratory, NOAA

Jack Kain, Natioanl Severe Storm Laboratory, NOAA


Accurate prediction of convective-scale hazardous weather events continues to be a major challenge because of the small spatial and temporal scales of the associated weather systems, and the inherent nonlinearity of their dynamics and physics. Uncertainties in numerical weather prediction (NWP) and the high-nonlinearity of the weather systems at the convective-scales render probabilistic forecasts from high-resolution ensemble forecasting systems especially valuable to operational forecasters. Though ensemble forecasting has long become the cornerstone of medium-range global NWP in weather centers around the world and the regional short-range ensemble forecasting (SREF) has been in operation over a decade, the storm-scale ensemble forecasting (SSEF) study is in its very early stage. A multi-year realtime SSEF effort has been conducted since 2007 at the center for Analysis and Prediction of Storms as part of the NOAA Hazardous Weather Testbed Spring Experiment. The CAPS SSEF consists of 20-50 members from three NWP model systems (WRF-ARW, WRF-NMM, and ARPS), covering the full continental United States at 4-km horizontal grid pacing. SSEF members are configured with a combination of initial perturbations extracted from the coarser grid SREF ensemble members and various physics options in microphysics, PBL and land-surface model, and radiation. Over 100 Doppler weather radar data, both radial wind and reflectivity, are analyzed into the SSEF in realtime using the ARPS 3DVAR/Cloud Analysis system. A wide range of ensemble products are made available, including QPF and PQPF (both grid-wise and neighborhood), and the probability matched ensemble mean QPF. Experimental probability products of lightning threat and CI parameters are also produced from the SSEF starting 2011 Spring. Objective evaluation of the CAPS SSEF dataset shows promising potentials in improving QPF and probabilistic QPF skills. For QPF at storm-scale, probability matched means are good replacement to the simple ensemble means, as the former outscore the later and all individual SSEF members, and also outscore the operational 12km NAM by a wide margin. ROC areas from the SSEF are far greater than the operational coarser-resolution PQPF from the NCEP operational SREF. Bias removal and calibration remain a big challenge for the SSEF.

Paper 172:

Dual-polarized weather radar analyses of local severe storms in Austria.

Rudolf Kaltenboeck, Austrocontrol, Vienna, Austria


After the first installation of operational dual-polarized weather radar in Austria at Valluga mountain in 2006 at an altitude of 2800m msl, Austro Control is going to renew the full weather radar network by installing four further simultaneous transmitting dual polarized radars till 2012. A new dual-polarized weather radar was installed close to Vienna airport in December 2010 and is now collecting first radar data in the eastern part of Austria. These lowlands are frequent affected by severe storms. The dual pol data in Austria will be used for nowcasting of local severe weather for aviation purpose e.g. as hail detection/quantification, to localize up/downdraft positions within convective systems, for lighting forecasting and to detect mesoscale features in more detail by analyzing new dual pol moments. These data will additional used to improve the data quality by discrimination between different precipitation types and nonmeteorological echoes and for quantitative precipitation estimations. Dual pol radar analyses of local severe storm cases will be presented which occurred during the summer season 2011 in the eastern part of Austria. Additional, the influence and impact of severe weather to aviation and air traffic flow will be shown.

Paper 173:





Eduardo Garcia-Ortega, UNIVERSITY OF LEÓN


The estimation of kinetic energy of hail precipitation using meteorological radar has caught the interest of some authors. In our case, we used the databases obtained by hailpad networks and the images of C-band and S-Band Radar to build an algorithm to estimate the vertical component of kinetic energy produced by a hail precipitation. In order to compile the database, we have established two categories of kinetic energy: low (< 20 J m-2) and moderate or high (> 20 J m-2). With this information, we have looked to establish a differentiation between hail precipitation that hardly produces damage and, on the contrary, that which does (Dessens et al.,2007) . Once these two categories were established, we applied a logistic function and statistical techniques that establish a correspondence between radar variables and the two categories of kinetic energy. The results show great uncertainty in determining kinetic energy using C-band radar. However, for the S-band radar the results have shown that the probability of detection is 85,7% with a FAR of 14,3% and a variance explained of 61,2%. This result allows us, with corresponding caution, to make a first estimation of the areas in which a hailstorm could produce damage. References Dessens J., C. Berthet, J.L. Sanchez, 2007: A point hailfall classification based on hailpad measurements: the ANELFA scale. Atmos. Res., 83, 132-139.

Paper 174:

Seasonal Analysis of Hailstorm Trends and Their Relationship to Meteorological Factors



Eduardo Garcia-Ortega, UNIVERSITY OF LEÓN




The Ebro Valley is one of the European regions with the highest incidence of hailstorm days. For the past ten years, a series of research studies have been developed to improve algorithms to discriminate between hail cells and rain cells (López and Sánchez, 2009) and to establish the climatology of hailfall in the area. In applying this identification tool, we have been able to analyze weather trends. The results show an increase in the average extension of hailstorms in the area, although, as expected, with a considerable annual variance. Additionally, making use of synoptic scale models that classify meteorological situations which produce hail in the Ebro Valley (Garcia-Ortega et al., 2011), a relationship between these models and the areas affected by hailstorms can be established. This allows us to describe the atmosphere on a synoptic scale in which severe convection processes are favored. Finally, to determine preconvective thermodynamic factors that most influence the extension of hailstorms, correlations between said extension and diverse preconvective parameters obtained using radiosounding have been developed. Of the stability indexes studied, the SI shows a strong correlation to the surface area affected by hail. References García-Ortega, E. , López, L., Sánchez, J.L., 2011. Atmospheric patterns associated with hailstorm days in the Ebro Valley, Spain. Atmos. Res. DOI: 10.1016/j.atmosres.2010.08.023. López, L. and Sánchez, J.L., 2009. Discriminant methods for radar detection of hail. Atmos. Res., 93, 358-368.

Paper 175:

Minimum temperature as an indicator of hail storm severity in a scenario of Global Climate Change




J. Dessens, ANELFA




Since storms that produce hail are one of the natural risks that cause the greatest damage in mid-latitudes, it is very important to be able to know how climate change can affect the frequency and intensity of such storms. Given the difficulty of measuring hail, there are not integral hailfall databases, or large trends of frequency or intensity. In some studies, this lack of data has been replaced by statistics that show loss registered by insurance companies for hailfall (Botzen et al., 2010). Nevertheless, the information given by insurance companies is biased and incomplete, and furthermore, it provides information about hail frequency, but not about its intensity. In this study, data from various hailpad networks has been used from different regions around the world. These hailpad networks, although local, allow us to obtain objective information about hail characteristics. The object of this study is to be able to establish a significant correlation between the variables obtained from the hailpads and to find a meteorological indicator that is simple and easily obtained. With this information, a correlation between simple meteorological parameters—such as temperature—and other variables that show the intensity of hail fall—such as diameter and kinetic energy—can be found. Minimum temperature was found a meteorological indicator to estimate the evolution of the severity of hail storms. References Botzen, W.J.W., Bouwer L.M. and van den Bergh J.C.J.M. 2010. Climate Change and Hailstorm Damage: Empirical Evidence and Implications for Agriculture and Insurance. Resource and Energy Economics, 32, 341-362.

Paper 177:

Realtime storm-scale forecasting in Southern China region

Fanyou Kong, Center for Analysis and Prediction of Storms, University of Oklahoma

Tuanjie Hou, Center for Analysis and Prediction of Storms, University of Oklahoma

Xunlai Chen, Shenzhen Meteorological Bureau

Ming Xue, Center for Analysis and Prediction of Storms, University of Oklahoma

Fei Cai, Shenzhen Meteorological Bureau


An hourly updated WRF-ARW based forecasting system, called Hourly Assimilation and Prediction System (HAPS), has been run in quasi-operational mode since March 2010 in Shenzhen Meteorological Bureau in Southern China. The storm-scale forecasting system originally had a domain of 840km by 600km covering the entire Guangdong province at 4-km grid spacing, and produced 6-h forecasts every hour utilizing all seven S-band Doppler radars in the province by analyzing both reflectivity and radial wind using the ARPS 3DVAR/Cloud Analysis system. In order to assure realtime storm-scale forecasts, two 48-h background forecasts are initiated at 00 and 12 UTC, respectively, at 12-km grid spacing covering a larger region using the NCEP GFS analysis and forecast as IC and LBC. HAPS is recently upgraded to have a larger 4-km domain (1680km by 1440km) that covers the entire five Southern China provinces and to extend the forecast lead time from 6-h to 12-h. In addition to all available Doppler weather radar data, other available surface and upper air observations, including AWS and wind profiler data, are also assimilated. Case studies have been conducted comparing the use of ECMWF 0.25 deg high resolution global analysis and forecast data versus the 0.5 deg NCEP GFS data as IC background and LBCs. QPF products from the HAPS 4-km grid forecasts, including 1-h, 3-h, and 6-h accumulated precipitation and derived reflectivity, along with other forecasted weather variables, are evaluated through the Spring-Summer season. Factors that impact the QPF forecast skill in HAPS are investigated through case studies.

Paper 178:


Luiz Carlos Bacelar, Universidade Federal de Pelotas

Cláudia Rejane Jacondino De Campos, Universidade Federal de Pelotas-Brasil

Gustavo Rasera, Universidade Federal de Pelotas

Cristiano Eichholz, Universidade Federal de Pelotas


Socioeconomic losses associated with weather events are constantly reported in the media. In Brazil, that is the South American country with the largest number of people affected by natural disasters, one of the leading phenomena that bring on damage to society, are the mesoscale convective systems (MCS) that modify weather conditions causing severe weather events, like intense precipitation, fall of hail, and intense electric activity. Currently MCS are studied using ForTrACC method (Forecasting and Tracking Active Convective Cells), which allows tracking these systems and analyzing the evolution of their features. Thus, the aim of this study was to analyze the relation between MCS and the occurrence of severe weather events in Rio Grande do Sul State (RS)-Brazil, in JFM (jan, feb, mar) and JAS (jul, aug, sep) from 2004 to 2008. First, channel 4 Geostationary Operational Environmental Satellite (GOES) images, provided by DSA/CPTEC/INPE (Satellite Division and Environmental Systems/Center for Weather Forecast and Climate Studies/National Institute for Space Research), from JFM and JAS from 2004 to 2008, with 4 km x 4 km spatial resolution and ½ hour temporal resolution were used in FORTRACC to select the MCS whose genesis and maintenance happened southward of 20oS, with life cycle of at least 6 h, with spontaneous initiation and normal dissipation and that reached the region covering the RS. During the study period were detected 179 MCS in JFM and 142 MCS in JAS. After, the data of severe weather events, that was supplied by the RS Defesa Civil (http://www.defesacivil.rs.gov.br), were used to verify if during the occurrence of each MCS there were any severe weather event warning. The results showed that ~23.5% and ~19.7% of MCS that reached RS in JFM and JAS, respectively, caused warning of severe weather event in the cities of RS. It was observed that 85.7% and 66.5% of the cities affected by MCS were associated to the severe weather events in JFM and JAS, respectively. Thus, the majority of severe weather events were associated to the MCS that reached RS in JFM, due to higher convective activity of this period.

Paper 179:

Numerical simulations of intense precipitation in Madeira Island using the MESO-NH Model

Flavio Tiago Couto, University of Évora

Rui Salgado, University of Évora

Maria João Costa, University of Évora


Located in the Atlantic Ocean, the Madeira Island presents favourable conditions for orographic precipitation development, sometimes responsible for high records, floods and several social-economics damages. The use of high-resolution simulations is becoming popular for diagnosis and prognosis of intense precipitation events all over the globe. Therefore, this study aims at analyzing the orographic effects on the atmospheric circulations, verifying the model's ability for the diagnostic of intense precipitation in a region where the orographic effects are predominant. For such, several events of intense rainfall over the Island have been studied using surface station data and numerical simulations with the MESOscale Non-Hydrostatic Model (MESO-NH). Making use of the nesting grid technique, the model is configured for 3 domains (resolutions of 9, 3 and 1 km), with convective parameterization applied in the outer domains (9 and 3 km) and explicit precipitation in the 1 km resolution domain. From the accumulated precipitation simulated with MESO-NH and through point to point comparison against records of various surface stations, it was possible to observed that the temporal evolution of the precipitation is well represented in the model simulation. The orographic effects could also be observed on other variables, namely, Relative Humidity, Temperature, Vertical Velocity, and Convective Available Potential Energy (CAPE). The results confirm the ability of high resolution models to simulate heavy precipitation over isolated mountains, in this case, Madeira Island, suggesting a few mechanisms to precipitation generation as some sort of relation between large scale flow, atmospheric hydro and thermal structure in the presence of local complex and steep orography.

Paper 180:

The operational usability of symmetric instabilities during convective initiation

Thomas Krennert, ZAMG, Central Institute for Meteorology and Geodynamics, Vienna, Austria

Matthias Langer, ZAMG, Central Institute for Meteorology and Geodynamics, Vienna, Austria


In literature, the operational use of moist symmetric instabilities is broadly used and agreed upon in the context of banding of wintertime precipitation events, for the summer season these mechanisms are described more seldom. Further, the coexistence of moist gravitational and moist symmetric instabilities during convective activity is discussed elaborately (Xu [1], Schultz and Knox [2]). Diagnostics from model fields (i.e. Dixon [3]) indicate the possible release of moist symmetric instabilities (MSI) within areas of existing low moist gravitational stability. Examples for similar diagnostics from studies with the Austrian LAM Models ALADIN / ALARO will be shown for the initiation of isolated thunderstorms in the alpine region. Furthermore, the relation between areas of probable MSI and the dynamics of upper tropospheric moisture gradients will be discussed, their application as capable indicators for regions of MSI release within non frontal pre-convective environments will be presented. [1]Xu, Q., 1986a: Conditional symmetric instability and mesoscale rainbands, Q. J. R. Met. Soc., 112, 315pp. [2]Schultz, D. M., Knox, J. A., 2007: Banded Convection Caused by Frontogenesis in a Conditionally, Symmetrically, and Inertially Unstable Environment, Mon. Wea. Rev., 137, 2095pp. [3]Dixon, R. S., 2000: Diagniostic studies of symmetric instability, Ph.D Thesis.

Paper 181:

Usefulness and skill of station-derived predictors in forecasting storm occurrence and intensity

Arturo Pucillo, OSMER ARPA FVG


Some recent occurrences in Friuli Venezia Giulia region (northeastern Italy, hereafter FVG) consisting in severe weather events driven by cold front advection (i.e. the suspected bow echo on 8 August 2008) have shown an apparently common characteristic: the northwest-to-southeast propagation path and strong northwestern winds blowing at 1500 m on the Alps (as measured by a meteorological station on the top of a mountain). Starting from this consideration, this work aims to characterize the skill of one or more predictors extracted from the measurements of the OSMER – ARPA FVG meteorological stations (e.g. wind magnitude, equivalent potential temperature, moisture fluxes, etc.), using the highest radar Vertical Maximum Intensity (VMI) measured on the plain and coast of FVG as predictand of the storm occurrence. The dataset here used consists of 10’ maximum time resolution VMI to define the storm occurrence on plain and coast areas of FVG, and 5’ time resolution observations derived from a mesonet of about 30 stations, considering a time-lag varying from 1 to 3 hours before the storm occurrence. All measures refer to a period 11 years long (2000-2010). The statistical approach has been based on the discriminant analysis and the skill of the forecast has been assessed through categorical analysis, in particular the ROC curve. Moreover a linear regression or multiregression for forecasting the VMI field as storm intensity continuous predictand versus the station-derived predictors has been performed.

Paper 182:

Forecasting thunderstorms on the Po Valley using sounding derived and satellite (Eumetsat IASI) derived indicies with Neural Networks

Agostino Manzato, OSMER - ARPA FVG

Paolo Antonelli, Space Science Engineering Center, University of Wisconsin


The aim of this work is to compare remote-sensing (IASI data from MetOp satellite) derived informations and in-situ (radiosoundings) derived instability indices for short term forecast of thunderstorm occurrence. Thunderstorm occurrence in two areas of the Po Valley (one centered over the Lombardia region and another centered over the Friuli Venezia Giulia region) were defined as number of cloud-to-ground lightnings, fallen between 11 and 17 UTC, exceeding a certain threshold. Instability indices derived from Milano and Udine radiosoundings, launched at 11 UTC, were used to produce a set of about 150 candidate predictors. A subset of about 7 predictors has been chosen by a step-wise forward selection algorithm as optimal input set of a neural-network-based thunderstorm forecasting system, when measuring the performance in terms of minimum cross-entropy error and maximum Peirce skill score on an independent test dataset. Vertical profiles of temperature and water vapor mixing ratio were retrieved from Eumetsat IASI radiances observed in clear sky above the target areas from the MetOp satellite at around 9:15 UTC, during the whole period from April to October 2007-2010, and were used to derive instability indices (not wind-depended). Hence, IASI-derived indices along with some linear combination of IASI radiances (Principal Component Scores) were used as candidate predictors to build another neural-network-based thunderstorm forecasting system. Despite the fact that the systems based on sounding-derived indices perform better than those based on IASI-derived data, encouraging results were obtained for both systems, indicating that space-borne high spectral resolution observations might have an impact on short term forecasting of thunderstorms. To further emphasize this concept, as last experiment, the sounding and satellite predictors were combined together in order optimize the forecasting system of thunderstorm occurrence in the Po Valley.

Paper 183:

Evaluating atmospheric instability from high spectral resolution IR satellite observations

Paolo Antonelli, Space Science Engineering Center, University of Wisconsin

Agostino Manzato, OSMER - ARPA FVG


The presented work aimed to investigate the statistical links between the instability indices derived from in-situ rawinsondes and space-borne IASI products over the areas of Udine Campoformido, and Milano Linate, Italy, with the occurrence of convective activity as detected by cloud-to-ground lightnings. Statistical relationships were evaluated as a first step in understanding and comparing the skills of individual predictors in forecasting the occurrence of convection. The approach used was based on bivariate discriminant analysis (Peirce Skill Scores). The satellite based predictors, retrievals and principal components of radiances, showed good correlation with those derived from rawinsondes, providing encouraging evidence of the usefulness of IASI products in the short term forecasting of instability activity. Results also showed that single predictors were found to have different skills, and indicated the need for an objective methodology capable of selecting and combining all the optimal and most statistically independent predictors to fully exploit their individual forecasting skills. Beyond the described results, significance of the presented material relies on the correlation found between linear combination of IASI radiances (Principal Component Scores) and the occurrence of convection, and on the validation of the IASI level 2 and level 3 products with rawinsondes and lightning observations.

Paper 184:

Tropical Rainfall Characteristics at the Eastern Coast of Northeastern Brazil

Ricardo S. Tenorio, Sistema de Radar Meteorológico de Alagoas - UFAL

Marcia C. Da S. Moraes, Sistema de Radar Meteorológico de Alagoas - UFAL

Byung Hyuk Kwon, Pukyong National University


Usually a single literature-suggested Z-R relationship, where Z the radar reflectivity factor and R the rain rate, is used for weather radar data interpretation. It is desirable to calculate a Z-R relationship by precipitation type to improve the accuracy of quantitative rainfall rate in case of coexistence of different precipitation types, such as, in the area of precipitation produced from a Mesoscale Convective System (MCS). In general, in the MCS trailing anvil, the stratiform precipitation does not fall as drizzle. Rather, the rainfall can assume significant intensity (~10 mm h-1) with showery character. For that reason, in this study, the precipitations were classified into convective and stratiform type, to produce optimum rainfall estimates. Therefore, Z-R relationships were developed for the Eastern Coast of Northeastern Brazil (NEB) using rainfall raindrop size distribution (DSD) data collected with a disdrometer RD-69, aiming their utilization to start the operation of a weather radar system. In this study, due to operation and maintenance facilities, the disdrometer was installed in the Campus of the Universidade Federal de Alagoas (The Federal University of Alagoas) in Maceió in 2001, 12 months before the complete installation of a new radar system. The DSD was stratified by rainfall rate classes. It is found that the DSD are clearly dependent on the parameters of the analytical distribution functions are, and show a marked monthly variability. The parameters of the frequency distributions are dependent on R. The forms of DSDs are similar but the amount of droplets in each one very strongly. This may be possible due to the short period of data collection or to the intraseasonal rainfall variability. The general relationship for the Eastern Coast of NEB was found to be Z = 176.5 R1.29, with correlation coefficient equal to 0.83. This equation is in accordance with the ones for stratiform rain reported in the literature. We found that the convective rain observed is produced by convective cells imbedded into stratiform cloud layers. However, when separating stratiform and convective rainfall we found that the linear coefficient of the Z-R relation is significantly smaller for convective rainfall than for stratiform one (65 and 167 respectively), but the exponential coefficient is higher for convective than for stratiform (1.69 and 1.26 respectively).

Paper 185:

Use of severe weather outlooks for damage prevention and civil protection in Finland

Ari-Juhani Punkka, Finnish Meteorological Institute

Jenni Rauhala, Finnish Meteorological Institute


The Finnish Meteorological Institute has delivered severe weather outlooks for rescue authorities and government organizations since 2005. In March 2010, a questionnaire was sent to all outlook users, which resulted in 172 responses from about 30 organizations. The purpose of this paper is to present results of this questionnaire; especially, how different authorities operate before and during severe weather events. In the ranking of the most significant weather phenomena among the authorities, convective weather is well represented. For fire brigades, thunderstorm wind gusts are the most significant severe weather events, while road maintenance authorities are more interested in phenomena that have a direct effect on the road surface (e.g. hail). A clear majority (75%) of the respondents assessed that the outlooks have helped to cope with difficult weather situations. The main reason for this benefit was the induced longer preparation time, which offered the possibility to rearrange work shifts and allowed better scheduling of the work tasks and internal communication within rescue organizations. Preparedness capabilities naturally depend on the available lead time, and, a 12-hour lead time already enables 95% of the respondents to take at least some preparedness actions. Furthermore, the chosen preparedness actions strongly depend on the lead time. Short lead times allow calling in extra personnel, checking the equipment and forwarding severe weather information. With a 24-hour and longer lead times, it becomes possible to reorganize work shifts and relocate equipment. Thanks to the outlooks, rescue and civil protection authorities gain time and are better prepared for severe weather. Higher level of preparedness may cause them extra costs, but nevertheless, these expenses are easily compensated by the benefits from better response to interruptions caused by severe weather as well as faster recovery of society from harsh weather conditions.

Paper 186:

Verifying sensitivity climatologies of Mediterranean intense cyclones for optimal adaptive observations

Lorena Garcies, Universitat de les Illes Balears

Víctor Homar, Universitat de les Illes Balears


Motivated by the ever present interest in understanding and improving the short-range forecasting of high impact weather events related to Mediterranean cyclones, a variety of sensitivity climatologies of Mediterranean intense cyclones has been recently built. The verification of these climatologies is essential to ensure the reliability of the sensitivity products and ultimately provide robust guidance to policy-makers on plans to redefine routine observational strategies. This work tackle the complex and multifaceted task of verifying a catalog of (an adjoint-based and two ensemble-based) sensitivity climatologies of Mediterranean intense cyclones. We perform Observing System Simulation Experiments with the WRF ARW model for 25 of the most intense Mediterranean cyclones detected in the ERA-40 database to test the ability of each method in identifying areas where perturbations in the initial conditions lead to the greatest impact on the forecast of the intense events. For the sake of a sensible reference, the performance of the available climatologies is tested against the judgement of an experienced severe weather meteorologist. In addition, the background random response to perturbations is also explored. The impact on the numerical forecasts of prescribed perturbations to the initial conditions is evaluated comparing each perturbed experiment to a control simulation. Furthermore, a quantitative analysis of the linearity of the evolution of the perturbations is performed using twin perturbations. Results confirm a statistically significant superior skill of the human and adjoint sensitivity fields against both ensemble sensitivity climatologies. Climatological ensemble sensitivities only show a noticeable improvement upon non-sensitivity experiments when an ad hoc classification of cyclones is used. This reveals one fundamental limitation of the ensemble sensitivity technique in climatological mode when it is applied to rare events insufficiently sampled in the available datasets, a common obstacle in statistical analysis of rare events such as severe weather.

Paper 187:


M. Angeles Picornell, DT de AEMET en Illes Balears

Joan Campins, DT de AEMET en Illes Balears

Agustin Jansa, DT de AEMET en Illes Balears


The Mediterranean region is rarely affected by tropical-like cyclones called medicanes (MEDIterranean hurriCANES), however they produce extremely strong winds. The small size, the tendency to develop over the sea and the low frequency of these cyclones make their detection and prediction difficult. Currently much effort is devoted to identify medicanes. In this sense, five criteria had been established by some authors to define a medicane from the satellite images and a first data base of twelve medicanes had been derived from the IR image data (1982-2005) of Meteosat satellite. The goals of this work are to contribute to a better understanding of these structures and to assess the capability of the numerical models to simulate them. To do that, a method for detection and tracking of the Mediterranean cyclones has been adapted to small-scale intense cyclonic perturbations. First, the algorithm has been modified to properly describe these small cyclones. Next, some parameters have been tuned to discriminate between medicanes and other small cyclones. The procedure has been applied to a numerical simulation output of the tropical-like cyclone that affected the Balearic Sea on September 12th 1996, a medicane selected from satellite-derived data base. This simulation has been done with the ECMWF operational model (T1279L91Cy36r1, grid length ~15km). The predicted cyclone and its evolution has been compared against some available observational data and numerical analyses. The suitability of the presented objective detection of medicanes from numerical model outputs is evaluated.

Paper 189:

On space-time distribution of tornado events in Bulgaria (1961-2010): analyses of two severe tornadic storms

Petio Simeonov, National Institute of Meteorological and Hydrology

Ilian Gospodinov, National Institute of Meteorological and Hydrology

Lilia Bocheva, National Institute of Meteorological and Hydrology


In the first part of this study the available information (date, time, location) about 47 thunderstorm occurrences with tornado events and thunderstorms with the development of waterspouts (5 of which have evolved waterspouts near the Black Sea coast and 3 are formed over dams) in the period 1961-2010 is presented. The 18 of all tornadoes development over land are formed in mountainous and hilly terrain forms and usually moved over the river valleys. Almost all tornado events occurred during the warm half of the year (maximum in June), with the exception of two cases from the last years: on February 15, 2005 in South Bulgaria and March 21, 2007 in northeastern Bulgaria. Commonly synoptic situation associated with the tornado formation is: slow-moving cold atmospheric fronts separating two types of air masses with large temperature gradients; advection of cold air in the lower layers of the troposphere from east or northeast and a warm stream in height from west or southwest; thermal convection. A comprehensive analysis of two severe thunderstorms, connected with development of tornado events is made. Meteorological, satellite, radar data and data from the field observation are used in study. The characteristics of these tornadoes give us the reason to classify them as moderate, i.e. F2 according to Fujita scale. The synoptic analysis suggests there is one typical structure that favours the development of tornados. It is a deep trough or a detached cyclone system to the west of the Balkans so that the frontal jet goes through Bulgaria from southwest to northeast. In this circumstance the generated convective systems are forced to migrate rapidly northeastward and this provides the Coriolis effect to take place and strengthen the cyclonic rotation of the up-moving air below the convective clouds.

Paper 190:

Sensitivity tests to different IC and DA strategy to investigate the impact of radar data assimilation on WRF rainfall.

Ida Maiello, Department of Engineering and Environmental Physics, University of Basilicata & Centre of Excellence CETEMPS -Department of Physics, University of L’Aquila

Rossella Ferretti, Centre of Excellence CETEMPS -Department of Physics, University of L’Aquila

Sabrina Gentile, Centre of Excellence CETEMPS -Department of Physics, University of L’Aquila

Mario Montopoli, University of L’Aquila/CETEMPS Department of Electrical and Information Engineering

Errico Picciotti, HIMET s.r.l.

Frank Silvio Marzano, CETEMPS/Department of Information Engineering, Sapienza University of Rome.


To evaluate the impact of radar data assimilation on a heavy rainfall case, different simulations are performed using WRFV3.2 model and 3DVAR assimilation technique. The case study is the Aniene flood, occurred during May 19-22 2008 in the urban area of Rome causing severe damages, and Monte Midia Radar data (located at the border between the Abruzzo and Lazio regions) have been assimilated to improve high resolution initial conditions. A set of experiments is performed making sensitivity tests both to different set of Initial Conditions (ECMWF analyses and ‘warm start’) and to a different assimilation strategy (3h-DA cycle). In addition, sensitivity tests to outer loops to include the non-linearities in the observation operators are performed for each of the previous experiments. A future step will be tuning of the length scale of the background error covariance and observation error parameters. Finally, to objectively identify the best simulations, statistical indicators are used as FBIAS, RMS and EQTS for the accumulated precipitation.

Paper 191:

Convective storms in a storm-permitting NWP ensemble

Nigel Roberts, Met Office

Giovanni Leoncini, Met Office

Changgui Wang, Met Office

Emilie Carter, Met Office

Humphrey Lean, Met Office


National Weather Centres are now running operational “storm-permitting” NWP forecast models with a focus on the prediction of severe convective storms. These models should provide more realistic and more accurate forecasts of disruptive local weather. Whilst it is becoming increasingly clear that storm-permitting models are more realistic and more skilful (provided an appropriate metric is used), the local accuracy can be extremely variable from forecast to forecast. This uncertainty can be crucially dependent on the error in the positioning and structure of the mesoscale dynamical forcing that is typically inherited from the coarser-resolution driving model (e.g. fronts, Potential Vorticity anomalies) as well as uncertainty in the initial conditions and model error. In order to represent the uncertainty associated with forecasting high-impact weather, including the development of local storms, the Met Office is setting up a storm-permitting model ensemble that is embedded within the regional MOGREPS-R ensemble. This is due for implementation in 2012. The benefit of doing this is not only to improve operational forecasting. Multiple differing realisations of the same events also provide an extremely useful scientific framework for determining the factors that that are responsible for storm development in particular environments and for detecting systematic deficiencies in the model representation of key processes. This presentation will show results from initial case studies in which ensemble forecasts were run at 1.5 km and 2.2 km, with a focus on the representation and uncertainty associated with some disruptive convective storms in the UK. It will include our perspective on the issues raised by a storm-permitting ensemble and how to extract useful information about the threat of severe weather from the output.

Paper 192:

Computational Fluid Dynamics (CFD) analysis of downbursts affecting overhead electricity transmission lines

Simon Tschannett, Weatherpark GmbH

Hildegard Kaufmann, Weatherpark GmbH

Matthias Ratheiser, Weatherpark GmbH

Wolfgang Gepp, Weatherpark GmbH

Herbert Lugschitz, Austrian Power Grid AG

Christoph Karner, Austrian Power Grid AG


On March 1st 2008 the powerful late winter cyclone “Emma” caused widespread damage over Central Europe. In Austria a thunderstorm with a downburst was embedded in “Emma”. This downburst led besides damages to buildings and forests (Pistotnik etal., 5thECSS 2009) to collapses of overhead electricity transmission lines. A total of 12 towers of two transmission lines failed in an area of approximatly 1 km in diameter, whilst another line 600m away remained nearly unaffected. In 1990 the cyclone “Vivian” caused similar but less damage in a neighbouring area but was not brought into connection with a downburst at that time. The question arose if the terrain triggers wind speed-up responsible for such high winds that could leed to the collapse of overhead transmission lines, or if downbursts were causing these failures. To investigate the influences of terrain and/or downbursts on overhead transmission lines, a high resolution Computational Fluid Dynamics (CFD) modell was adapted to this task. Steady state RANS simulations (no "physics" included) were carried out for four different scenarios: storm winds within "Emma" and "Vivian" were simulated in a 2.4x2.4x0.7km domain with and without downburst (as boundary conditions typical downward wind speed values (assumption) used, ambient winds taken from meteorological observations). The highest grid resolution of the simulation is 0.8m, forests were parametrized and included as porous media. Analysis of the 3D wind fields show, that the relatively flat terrain is only responsible for small wind speed-up not capable of damaging transmission lines. Downdraft simulations of "Emma" ("Vivian") result in wind speed-up – with respect to the downdraft wind speed of 29m/s (33m/s) – of up to 2 (2.3). This means, the critical wind speed is well exceed, and the failure of the overhead transmission lines can be contributed to the downbursts.

Paper 193:

Remote sensing observations of a Mediterranean thundersnow and low-altitude heavy snowfall event

Joan Bech, Meteorological Service of Catalonia / Universitat de Barcelona

Montserrat Aran, Meteorological Service of Catalonia

Clara Brucet, Meteorological Service of Catalonia

Nicolau Pineda, Meteorological Service of Catalonia

Tomeu Rigo, Meteorological Service of Catalonia

Eliseu Vilaclara, Meteorological Service of Catalonia


Heavy snowfalls at low altitude and thundersnows, i.e. snowstorms with lightning, are not frequent in the Mediterranean area. A combination of these two relatively rare phenomena occurred on the 8th March 2010 in Catalonia (NE Iberian Peninsula). This episode produced substantial disruption in the communications and transport networks at a regional level and, particularly, in the metropolitan area of Barcelona. It also damaged dramatically large areas of low altitude forests, and a number of power lines as well, due to the combined effect of wind and accumulation of wet snow. Previous documented low-altitude heavy snowfall events in this area also caused important problems, mainly due to its infrequence – though 16 heavy snowfalls in Barcelona were recorded from 1947 to 2009. The work presented here is based on the analysis of remote sensing observations from operational systems (MSG images, total lightning data, and C-band Doppler radar observations), complemented with surface observations, radiosoundings, and NWP products. MSG visible and infrared images show clearly the development of cloud spiral bands associated to a deepening surface low level pressure system, which was the main synoptic feature of this event. Lightning observations indicated not particularly high rates (310 intra-cloud and 128 cloud-to-ground strikes) compared to warm season thunderstorms, though were concurrent to some local intensification of precipitation, which in several observatories exceeded 100 mm in 24 h. Radar observations indicated the presence of embedded convective cores within stratiform areas, which were predominant in the precipitation field. Radar echoes were characterised by moderate vertical development (maximum heights about 8 km). The evolution of individual storm cells are examined in terms of lightning strikes and radar observations (echo intensity, maximum height, and areal extension) similarly as was done in recent thunderstorm case studies in the area.

Paper 194:

Numerical modelling of orographic effects in heavy precipitation episodes

Laura Trapero, Centre d'Estudis de la Neu i la Muntanya d' Andorra, Sant Julia Loria, Andorra (CENMA-IEA)

Joan Bech, Meteorological Service of Catalonia / Universitat de Barcelona

Jeroni Lorente, Universitat de Barcelona


Orographic enhancement mechanisms can play a major role in heavy precipitation events taking place in complex topography regions. A number of conceptual models to describe these mechanisms have been suggested by several authors. For instance, Lin and co-authors employed in 2000 the so-called orographic moisture flux index (OMF), given by the product of the air flow impinging on a mountain ridge, the terrain height, and the water vapour mixing ratio in a given atmospheric layer. In a recent case study of a heavy rainfall event in Catalonia (NE Spain) the OMF index was used with to GFS (global model) data depicting heavy rainfall areas in the highly complex terrain region of the Pyrenees. In that area obtaining not only precipitation forecasts, but also radar quantitative precipitation estimates (QPE), is a rather challenging task, as shown by Trapero and co-authors where the radar QPE quality over Catalonia was examined. The objective of the study presented is examining the use of the OMF index to analyse heavy precipitation events in order to describe quantitatively orographic enhancement mechanisms. The work is based on high resolution non-hydrostatic simulations over the eastern Pyrenees using the WRF model at 3 km horizontal resolution. Initial results for a heavy precipitation event in the Pyrenees are comparable to one alpine event analyzed by Lin and co-authors where OMF values exceeded 4.7 (g/kg m/s). The ultimate goal of this study is contributing to improve our understanding of the orographic enhancement processes that take place in complex topography regions in order to better describe these effects both in diagnostic products such as weather radar QPE and in numerical forecasts.

Paper 195:

ProFEWS: Towards the assessment of the impact of rainfall uncertainty in the framework of Flood Early Warning Systems

Joan Bech, Meteorological Service of Catalonia / Universitat de Barcelona

Marc Berenguer, Centre de Recerca Aplicada en Hidrometeorologia, Universitat Politècnica de Catalunya,


Heavy rainfall events and subsequent flash floods are among the most significant natural hazards in the Mediterranean area. Therefore, successful hydrological management in this region includes prioritizing Flood Early Warning Systems, especially in small- to medium-sized basins. In such a context, the simulation of the basin response is affected by a number of errors of different nature, which limit the performance of the system. The development of a probabilistic approach able to provide additional information on the reliability of flow simulations would certainly be valuable for the operators in charge of issuing warnings and decision-making. The aim of the work presented here is to develop a methodology to couple probabilistic rainfall inputs -mostly based in weather radar quantitative precipitation estimates- with a distributed rainfall-runoff model to generate probabilistic flow forecasts within the framework of Flood Early Warning Systems. This will explain a part of the overall uncertainty affecting flow simulations (that resulting from the propagation of the uncertainty in rainfall inputs), and from an operational perspective, it would inform decision-makers about the expected uncertainties due to rainfall inputs (main forcing of the hydrological model). The developed methodology is being implemented to produce preliminary probabilistic rainfall and runoff forecasts in the Besòs and Llobregat basins (1020 and 5040 km2, respectively) in Catalonia, NE Spain, within the ProFEWS project (CGL2010-15892).

Paper 196:

The Mesocyclone Detection Algorithm of DWD

Thomas Hengstebeck, Deutscher Wetterdienst

Dirk Heizenreder, Deutscher Wetterdienst

Paul Joe, Environment Canada

Peter Lang, Deutscher Wetterdienst


Observations by weather radar networks offer the possibility of scanning the low- and mid-level atmosphere in high spatial and temporal resolution providing the possibility of detecting and tracking rapidly developing, small scale severe weather phenomena in an operational environment. A mesocyclone, defined as a vortex of rotating air within a thunderstorm, is such a dynamic feature that can be captured by Doppler radar networks. Mesocyclones are frequently found as rotating updrafts in supercells and often occur in connection with severe weather events like heavy rain, hail, strong surface winds and tornadoes. Therefore, the detection of mesocyclonic shear regions in Doppler radar data gives valuable information for issuing nowcasts and severe weather warnings. The approach of mesocyclonic shear detection as implemented at Deutscher Wetterdienst (DWD) can be divided into a preprocessing, a processing (the actual algorithm) and a postprocessing part. During preprocessing, dual-PRF unfolding errors are corrected, which otherwise would produce spurious high shear regions and corrupt real mesocyclone vortices. The algorithm uses the pattern vector approach to identify the regions of high shear within the centre of mesocyclonic rotation and provides 3D mesocyclone objects. A planned, unique aspect of the DWD algorithm is the merging of 2D mesocyclone features from multiple network radars in overlapping regions to create a single 3D detection. During postprocessing these mesocyclone objects are validated using secondary features such as VIL. A severity metric is created using mesocyclone object properties (shear, momentum) and the secondary features and the various mesocyclone objects are then ranked and issued to the forecaster as guidance. The details of the Mesocyclone Detection Algorithm of DWD will be described and a preliminary validation by means of selected weather case studies will be provided.

Paper 197:

A four years (2007-2010) classification of convective storms in the Mediterranean basin

Samantha Melani, Consorzio LaMMA/IBIMET-CNR

Francesco Pasi, Consorzio LaMMA/IBIMET-CNR

Bernardo Gozzini, Consorzio LaMMA/IBIMET-CNR

Alberto Ortolani, Consorzio LaMMA/IBIMET-CNR


The Mediterranean basin is an area rich in deep and long-lasting convective systems, due to its position in-between a subtropical and mid-latitude zone. The spatial and seasonal variability of these Mediterranean convective events have been investigated for the 2007-2010 years, using both geostationary (e.g., Meteosat Second Generation) satellite data for their detection, and ECMWF analysis data for finding out the most favourable synoptic conditions for their formation. The performed analysis shows the existence of some preferential areas of genesis, mainly located in the central Mediterranean basin (i.e, Sicilian channel, Ionic and Tyrrhenian seas), where the storms develop and grow preferentially in the autumnal season. The selected synoptic features, as precursors of convective activity genesis, show how the totality of the identified cases occurs in mid-troposphere (500 hPa) troughs or cut-off circulation within southerly flow, with values of mean sea level pressure ranging mainly between 1010-1020 (hPa), and high Thetae (850 hPa) values distributed on average between 50-60 °C. Further, the genesis and dynamics of these events appear to be well correlated with the presence of upper and low level jets, together with a dynamical tropopause anomaly. Among all the detected (269) cases, 94 ones are associated with severe weather reports extracted from the ESWD archive: a statistical analysis was performed on these cases resulting in valuable information about the nature of the selected systems. The long-term objective of this preliminary study is to build up a climatological database of deep convective events occurring in the Mediterranean sea which may critically impact on the Italian peninsula and potentially affect population: an algorithm which can help regional meteorological services in making early decision and a better forecasting of the development and effects of these extreme events has planned to be developed in the very next future.

Paper 198:

The new system of flash flood forecasting in the Czech Republic

Lucie Březková, Czech Hydrometeorological Institute

Petr Novák, Czech Hydrometeorological Institute

Hana Kyznarova, Czech Hydrometeorological Institute

Milan Šálek, Czech Hydrometeorological Institute

Martin Jonov, Czech Hydrometeorological Institute

Petr Frolík, Czech Hydrometeorological Institute


Czech Hydrometeorological Institute (CHMI) issues daily flow forecast with lead time 48 hours for many catchments of size of hundreds to thousands square kilometers. But forecasting of local flash floods caused by heavy precipitations, which hit very small catchments, still remains an open problem. First test of using precipitation nowcast for flash flood forecasting showed that some flash floods can be predicted several tenths of minutes in advance (Šálek et al., 2006). In Czech Republic the deterministic approach inspired by NOAA system called Flash Flood Guidance was set up in testing operation in 2010. However, the simulation of flash floods provided by hydrological models is associated with great uncertainty of not only quantitative precipitation forecast (QPF), but also of quantitative precipitation estimates (QPE).Thus the effort of stochastic approach is the logical step.. Nowadays, several methods of precipitation nowcasting are run operationally in CHMI. The set of all real-time available QPF makes so called „poor man ensemble“. The related discharge ensemble can give us more realistic approximation of the flood development. The attention is paid to the estimation of the probability of exceedance of the limit discharge in the threatened catchment.. The paper shows the first results of the proposed system which was set up in testing operation on several pilot catchments within the territory of the Czech Republic. Reference: Šálek M., Březková L., Novák P., 2006: The use of radar in hydrological modelling in the Czech Republic - case studies of flash floods. Natural Hazards and Earth System Sciences, 6, 229-236

Paper 199:

Analysis on the Meso-scale Structure of the Snowstorm on 25-29 January 2008 with Dual-Doppler

Haiguang Zhou, Chinese Academy of Meteorological Science


The heavy freezing rain and snowstorm in January and February 2008 brought the severe damage in South China. During 25 and 29 January, there was a snowstorm in Jiangsu province. The snow depth in Nanjing, Zhenjiang and Changzhou city break its record since 1961. The three-dimensional wind field was retreived by the Nanjing and Maanshan dual-Doppler weather radar synchronization volume data in order to investigate the 3D kinematic structure and the maintenance mechanism of the snowstorm. The retrieval wind field shows that the vertical wind shear and the southwest jet stream play the important role on the form and maintenance of this snowstorm. The northeast wind prevails at the low level, on the contrary, the southwest wind prevails above 2km height. There is strong southwest jet stream above 3.5km. In the vertical cross-section, the echo top of 10dBZ is about 7-8km, and the maximum reflectivity is about 35dBZ that is 4km height. The updraft in the strong echo core is more than 4m/s at 3.5km level. The updrafts and downdrafts were distributed alternately. The updraft and the downdraft form the vertical circulation that is conducive to the continuance of the snow. When the vertical wind shear at the low level weakens, the snowstorm fades away. The three-dimensional kinematic conceptual model of this snowstorm is also discussed. Acknowlegements The work was supported by the National Science Foundation of China (grant 40975015), the National High Technology Research and Development Program of China (grant 2007AA061901), and the foundation of state key laboratory of severe weather.

Paper 200:

Objective Satellite-Based Overshooting Top and Enhanced-V/Cold Ring Signature Detection: Validation and Their Relationships With Severe Weather

Kristopher Bedka, Science Systems and Applications Inc. at the NASA Langley Research Center

Jason Brunner, Cooperative Institute for Meteorological Satellite Studies, UW-Madison

Richard Dworak, Cooperative Institute for Meteorological Satellite Studies, UW-Madison

Wayne Feltz, Cooperative Institute for Meteorological Satellite Studies, UW-Madison


Overshooting tops (OTs) are the product of deep convective storm updrafts of sufficient strength to rise above the storms’ general equilibrium level and penetrate into the lower stratosphere. OTs appear as small clusters (< 15 km diameter) of very cold pixels relative to the surrounding cirrus anvil cloud in infrared satellite imagery. An OT with a significant vertical protrusion above the anvil can act as an obstruction to the jet stream wind flow and is partially responsible for producing the cold-ring and enhanced-V signatures in IR satellite imagery, which often indicate the presence of a significant severe storm. Objective satellite-based OT and cold-ring/enhanced-V detection algorithms have been developed over the past 3 years for future operational use within the U.S. GOES-R ABI program. OT detection algorithm output has been used to create long-term databases of OT-producing storms over the U.S., Europe, and North Africa using GOES and MSG SEVIRI data. Quantitative comparisons between ground-based radar data and satellite OT signatures indicate that an OT often corresponds with the timing of the tallest precipitation echo top and relative rainfall maxima. Validation of OT detections relative to ground- and space-based radar profiles show an OT false detection rate ranging from 6-30%, depending on the validation platform, cold vs. warm season, and OT minimum IR temperature. The cold-ring/enhanced-V signature detection algorithm features a false alarm rate of ~20% when compared against 400+ events in MODIS and AVHRR imagery. Large hail, severe wind, and/or tornadoes were found near 25% (76%) of OT (cold-ring/enhanced-V) detections. This presentation will describe these detection algorithms and highlight validation and radar/severe weather comparison results, which show that these algorithms provide a detection accuracy exceeding previous techniques and can be used by forecasters to diagnose hazardous convective storms especially in the absence of reliable Doppler weather radar data.

Paper 201:

Mechanism of the TC/submonthly Wave Pattern Propagation in the Western North Pacific: A Barotropic View

Ken-Chung Ko, National Kaohsiung Normal Univ.

Huang-Hsiung Hsu, National Taiwan Univ.


The mechanisms of propagation of the TC/submonthly (7-30 day) wave pattern in the Western North Pacific during two extreme ISO phases were explored using the vorticity and kinetic energy budget at 850 hPa. The background mean flow advection played a dominant role in the pattern propagation. Eddies tended to grow where the zonal wind gradient was largest and were oriented in a particular direction to efficiently extract kinetic energy from the background flow. The results suggest that the TC/submonthly wave pattern occurring in the confluent zone between the monsoon trough and the anticyclonic ridge, where kinetic energy can be efficiently extracted, and with a circulation pattern in the northeast-southwest orientation, were the configuration to have optimal growth and propagation. It is also found that the amplitudes, shape and propagation of the eddies in East Asia and the Western North Pacific were strongly constrained by the configuration of the monsoon trough and the anticyclonic ridge. Therefore, this is likely the reason why the northwestward propagating and northeast-southwest orienting TC/submonthly wave pattern was frequently observed in the Western North Pacific during boreal summer.

Paper 202:

Assessing the impact of the assimilation of satellite-retrieved precipitation and humidity products into the hydrostatic BOLAM: Two Italian case studies

Stefano Mariani, ISPRA - Institute for Environmental Protection and Research

Oxana Drofa, ISAC-CNR - Institute of Atmospheric Sciences and Climate-Italian National Research Council

Silvio Davolio, ISAC-CNR - Institute of Atmospheric Sciences and Climate-Italian National Research Council

Antonio Speranza, UniCam - University of Camerino

Andrea Buzzi, ISAC-CNR - Institute of Atmospheric Sciences and Climate-Italian National Research Council


Satellite-borne instruments represent a useful data sources additional and, in some case alternative, especially in terms of coverage, to more traditional in-situ instruments. Satellite estimations can have indeed a high impact in operational flood forecasting and early warming systems. This was the goal of a pilot research project funded by the Italian Space Agency, as part of its Earth Observation Program, dealing with the exploitation of satellite-based measurements for the flood risk management (Progetto Pilota "Protezione Civile dalle Alluvioni: il Nowcasting"). With reference to the numerical modelling activity, one of the main objectives of this pilot project was to assimilate satellite-retrieved precipitation and humidity products into the hydrostatic BOLAM model and to evaluate its performance in terms of quantitative precipitation forecasts (QPFs). BOLAM QPFs obtained without any assimilation scheme were also considered as control simulations. The work presents the verification results for two intense events occurred in Italy in 2009. A multi-method verification approach based on categorical skill scores and subjective and object-oriented techniques is applied for the QPF assessment. Rain gauge data available all over Italy are employed for producing observational analyses. In addition, the verification results obtained for BOLAM (with and without assimilation) are compared and contrasted against those related to the non-hydrostatic MOLOCH model.

Paper 203:

The Verification of the Origins of Rotation in Tornadoes 2 (VORTEX2)

Josh Wurman, Center for Severe Weather Research

Louis Wicker, National Severe Storms Laboratory

Yvette Richardson, Penn State University

Erik Rasmussen, Rasmussen Systems

Paul Markowski, Penn State University

David Dowell, National Oceanic and Atmospheric Administration

Don Burgess, The Cooperative Institute for Mesoscale Meteorological Studies

Howie Bluestein, University of Oklahoma


This presentation will describe the second Verification of the Origins of Rotation in Tornadoes Experiment, VORTEX-2, the field phases of which occurred in 2009 and 2010.The VORTEX2 experiment was designed to explore (a) the mechanisms of tornadogenesis, maintenance, and demise, (b) the wind field near the ground in tornadoes, (c) the relationship between tornadoes and their parent thunderstorms and the relationship among tornadoes, tornadic storms and the larger scale environment, and (d) how to improve numerical weather prediction and forecasting of supercell thunderstorms and tornadoes.VORTEX-2 was by far the largest and most ambitious observational and modeling study of tornadoes and tornadic storms ever undertaken.It employed fourteen mobile mesonet instrumented vehicles, ten ground based mobile radars, several of which were dual-polarization and two of which were phased-array rapid-scan, five mobile balloon sounding systems, thirty-eight deployable in situ observational weather stations, unmanned aerial systems, video and photogrammetric teams, damage survey teams, deployable disdrometers, and other experimental instrumentation as well as extensive modeling studies of tornadic storms. Participants were drawn from over 20 universities and laboratories, and at least five nations.Over 60 students participated in field activities. The VORTEX-2 field phases spanned two years in order to increase the probability of intercepting significant tornadoes, which are uncommon.VORTEX-2 made special efforts to operate near a unique, extensive, and diverse network of stationary instrumentation in Oklahoma including a phased array radar, an array of small stationary radars, a prototype dual-polarization operational radar, and a statewide mesonet.The project also made special efforts to operate in regions where unmanned aerial system flights were permitted. Data were collected in dozens of supercell thunderstorms and over a dozen different tornadoes.These included the violent and fatal Oklahoma City tornadoes of 10 May 2010, the Goshen County / Lagrange tornado of 5 June 2009, and several other tornadoes that are obscure now, but will likely become well known to science because of the VORTEX2 data sets collected. These include several more tornadic storms observed with integrated observations during 2010, including the Clinton/Weatherford tornadoes on 12 May 2010, the Dumas tornadoes of 18 May 2010, the Kingfisher tornadoes of 19 May 2010, the Tribune tornadoes of 25 May 2010, and the Booker tornadoes of 13 June 2010.The data in several other non-tornadic or marginally tornadic supercells will permit VORTEX2 researchers to study the differences between storms that make tornadoes and those that do not, thereby allowing for more accurate warnings.

Paper 204:


Jari Petteri Tuovinen, Finnish Meteorological Institute

David M. Schultz, Centre for Atmospheric Science, School of Earth, Atmospheric and Environmental Sciences (The University of Manchester)


Historical hail articles were examined using the National Library’s Internet database (http://digi.lib.helsinki.fi). Altogether 250 different newspapers are stored in the database during 1833 to 1909, covering both Finnish and Swedish languages. The database can be accessed easily via Internet and it operates with word search. We made searches using multiple hail-related keywords on two languages (e.g. raekuuro, hagel and hagelskur), that dated between May and September. Moreover, we focused upon the period between 1 May and 15 October, because there was an up to two-week lag between the event and hail article. The completed search highlighted parts of each newspaper article where the keyword(s) occurred, so there was no need to look through the whole paper (typically 4–8 pages with small font size). During this time span, Finland was occupied by Sweden and Russia, but we collected cases only from the area that belongs to Finland today. Farming was the most important livelihood during the time. Grain, rye, barley and potatoes were farmed from south coast to Southern Lapland of Finland. Crop damage and broken windows were the most typical hail damage with over 50% cases. All sized hail reports were included and the amount and detail of information differed for each report. However, a few descriptions were surprisingly informative and high detailed. In many cases, the hail size was compared to some known object or the hailstone’s weight was measured. Close to 400 hail reports were collected, where about 54% (215) were severe-hail size (2 cm or greater in diameter) and 17% cases had sizes smaller than 2 cm with the remaining 30% having sizes unreported. There are many similarities when we compare these results to a more recent climatology (1930–2006) previously published.

Paper 205:

Heavy and Extreme Precipitation from Supercell Thunderstorms in the United States

Nathan Hitchens, NOAA/National Severe Storms Laboratory

Harold Brooks, NOAA/National Severe Storms Laboratory


The hazards attributed to supercell thunderstorms are primarily considered to be tornadoes, large hail, and damaging winds, but flooding from heavy and extreme rainfall is not as often considered with these storms. As a result there has been little research on the role that supercells play in the production of heavy and extreme rainfall events. Smith et al. (2001) examined four cases of supercells that had extreme rainfall rates at timescales of one hour and less, and suggested that supercell thunderstorms play a significant role in short-duration extreme precipitation, but their climatological role was difficult to assess at that time. In order to assess the contribution of supercells to the climatology of short-duration precipitation extremes the present study uses the Warning Decision Support System – Integrated Information to objectively identify supercell thunderstorms from mosaicked radar data and associate them with high-resolution, accurate multisensor precipitation estimates. Using a feature-based approach, the characteristics of the precipitation related to supercells are examined, with an emphasis on heavy and extreme precipitation. Future work will focus on characterizing model-simulated supercell rainfall and comparing it to estimated rainfall of observed supercells. Smith, J. A., M. L. Baeck, Y. Zhang, and C. A. Doswell, 2001: Extreme rainfall and flooding from supercell thunderstorms. J. Hydrometeor., 2, 469-489.

Paper 206:

Sensitivity study of BRAMS forecast in an organized mesoscale convective event

Matilde Nicolini, Dpto. Ciencias Atmósfera Océanos/UBA/Centro de Inv. Mar y la Atmósfera/CONICET

Yanina García Skabar, Servicio Meteorológico Nacional/Fac. Agronomía/UBA

Luciano Vidal, Dpto. Ciencias Atmósfera Océanos/UBA, Serv. Met. Nacional

Paola Verónica Salio, Dpto. Ciencias Atmósfera Océanos/UBA/Centro de Inv. Mar y la Atmósfera/CONICET


Heavy rainfall and other severe phenomena related to deep convection and their forecast is a major problem in different geographical regions. Subtropical South America (including La Plata basin) is particularly affected by large and intense mesoscale convective systems (MCSs) within which severe events develop specially during the warm season. One of the main objectives is to progress in the forecast of these intense events in order to reduce related damages. During early morning of January 12, 2010 an extended convective line developed in association with a cold front that propagated over the central and northern part of Buenos Aires Province, Argentina. Related storms produced severe winds (reported gusts exceeding 30 m s–1) in different locations around the city of Buenos Aires, causing material damage and even lost of lives. Numerical forecasts of the case study were performed using the Brazilian model Regional Atmospheric Modeling System (BRAMS), reaching a horizontal resolution of around 2km. In the present paper, model forecast performance was evaluated against measurements from radar, satellite estimations of precipitation and surface observations available in the region. The sensitivity to different initial conditions, horizontal and vertical resolution and settings on cloud microphysics scheme are addressed and discussed in order to maximize the BRAMS forecast potential for this particular case.

Paper 207:

Interannual variations in tornado frequency and their association with regional atmospheric conditions: Implications for seasonal predictions

Robert Trapp, Purdue University


The past several years highlight the potential range of interannual variability in tornado occurrence across the United States and elsewhere. In this study, we seek meteorological explanations for such variability in the form of anomalies in regional-scale atmospheric conditions. We first establish the degree to which tornado occurrence can indeed be considered anomalous. Our approach involves local fits of the gamma distribution to the frequency of tornado occurrence. We show, not surprisingly, that considerable geographical variation exists, with significant positive (or negative) anomalies in some region despite a relatively low (or high) tornado occurrence overall for the country. Data from 2006 and 2008 illustrate this well. The tornado-day anomalies are then compared to anomalies constructed using the North American Regional Reanalysis data from the period 1979-present. Insight is provided by some fairly basic sets of variables. We find, for example, that positive anomalies in warm-season (March-April-May) soil moisture, 850-hPa specific humidity, and negative anomalies in mean sea-level pressure (and corresponding cyclonic circulation anomalies) associate with positive anomalies in 2008 warm-season tornado days in the southern Great Plains. A multi-linear regression model that includes these and a few other predictor variables quantifies this association. Some discussion will be devoted to how these findings could lead to seasonal outlooks for tornadoes and other convective weather hazards.

Paper 208:

Using historical information sources to reconstruct historical severe storm cases in Central Europe

Thilo Kühne, European Severe Storms Laboratory (ESSL e.V.)

Heiko Wichmann, Skywarn Deutschland e.V.


With the foundation of the European Severe Storm Laboratory (ESSL e.V.) and the establishment of the European Severe Weather Database (ESWD), severe storm events could be listed from all over Europe. Furthermore, a strongly rising interest in severe weather in Europe and foundations of volunteer observation networks (e.g. Skywarn Deutschland, etc.) contributed to the fact that severe events are better recorded than ever before. Notwithstanding, there is a great lack of information regarding sever storm events from the period before the year 2000. In the period from 2008 to 2011, extensive research work on historical events were carried out with the aim to expand the knowledge about severe storms over the past centuries, using historical information, and to analyse the potentiality whether historical storms can be reconstructed. Historical sources were analyzed in terms of containing information about tornadoes, hailstorms, heavy rainfall, straight line winds and lightning. More than 2,700 previously unknown events could be secured. The majority of reports were related to hail damages from the period of the years 1720 to 1935. To establish a reference period, it was decided to consider a 250-year period (1700-1949) with a focus on Central Europe. The quality of historical information included known or mentioned quantities and very variable qualitative statements regarding details. Furthermore, only about 50% of the reports had a precise statement concerning the exact day and day period. In a few cases there was information about the direction of movements of thunderstorms / hailstorms. Therefore, in these specific cases and with the addition of all available information, it was possible to reconstruct severe storms using schematic visualization.

Paper 209:

Tornado and Water Spout Climatology in Brazil

Luci Hidalgo Nunes, UNICAMP

Laura De Bona, UNICAMP (undergraduation student)

Daniel Henrique Candido, UNICAMP (post graduation student)


The study presents a preliminary climatology of tornadoes and water spouts in Brazil. Altogether 199 events were found between 1960 to 2009 and two prior to this period (1923 and 1957), 85.6% of them in the central-southern region of Brazil, the most developed and populated area of the country, encompassing the states of Rio Grande do Sul, Santa Catarina, Paraná, Mato Grosso do Sul, São Paulo, Rio de Janeiro and Minas Gerais. It is of note that 16 of the 27 Brazilian states registered tornadoes and/or water spouts along the period. Events were not associated with any damage scale like Torro or Fujita because the assignment of any of these scale rating depends on particular structures being present to be damaged, and Brazilian structures (materials and architectonic patterns) do not correspond to the same ones considered in these scales, fact which pose difficulties for comparisons. Most of the registers (73.6%) occurred in the period of 2000 to 2009, but because any significant climate anomaly can be associated with the increase in the number of the events this upward trend is probably due to the facility to document these episodes in recent years and to the growth of urbanization and population in the region. This aspect underlines that prior to the 1960s tornadoes and water spouts were most likely underreported. The monthly distribution for both the entire country and the central-southern region shows that tornadoes and water spouts are more common in November, January, February and March, when more energy is available. May also presents a high frequency of tornadoes due to the conflict between prevailing tropical systems and transient polar masses, more frequent and intense in late autumn. Other aspects which seem to be related to tornadoes are topography, hydrography and the Jet Stream position.

Paper 210:

Severe convection checklist for warning operations at the MeteoSwiss forecasting offices

Lionel Peyraud, MeteoSwiss

Lionel Moret, MeteoSwiss

Christophe Salamin, MeteoSwiss


Severe convective storms regularly affect Switzerland during the warm season. The Alpine region is subject to a wide range of convective modes spanning the entire thunderstorm spectrum from airmass thunderstorms to supercells. In order to better anticipate the convective mode and virulence of upcoming convective events, MeteoSwiss has developed a severe convection checklist. The checklist is ingredient based and is meant to determine whether severe thunderstorm watches are warranted in the next 12 to 36 hours. It is broken up into 4 sub-sections, each representing a key convective ingredient : synoptic configuration, instability, wind shear and humidity. Each sub-section contains several parameters and/or indices spanning a range of representative values and deemed most relevant to measure the type and severity of the forecasted convection. In order to better differentiate non-organized from organized convection and to better estimate the overall convective threat, different weights are attributed to each ingredient as well as to each parameters ranges of values. The checklist will be utilized officially for the first time during the 2011 convective season.

Paper 211:

Conceptual model for tornadic supercell initiation in SE Romania

Aurora Bell, National Meteorological Administration

Daniel Carbunaru, National Meteorological Administration

Sorin Burcea, National Meteorological Administration

Felicia Carbunaru, National Meteorological Administration


Forecasters confess that the most difficult part of the forecasting process of severe weather associated with deep, moist convection is to assess of the moment and place of the storm initiation. The present study tries to contribute to the scientific approach that should be used by forecasters in identification of the place and moment when instability, moisture and lift act together to initiate convection. It provides a conceptual model for severe convection initiation in SE Romania in a specific synoptic circulation. The reason of this work is that the most severe tornadic supercells in SE Romania have been initiated in this synoptic type scale flow, named here convergence of southern flows. We present the contribution of local features, like see breeze or orography, and also the specific structure of the PBL in the initiation of these storms. We this conceptual model we enable the forecaster with an important tool in assessing the synoptic environment prolific to severe weather, mainly tornadoes, and to monitor the local circulations that can enhance or diminish these conditions. Comparison with tornado environment in US is also done. The study is based on Doppler radar data and different radar structures of the storms are presented, with special interest on convergence lines or frontiers that contribute to the storm initiation, the patterns of the relative to the storm flows and the flows before initiation, intensity of the low level jets and mesocylone characteristics. For this study we have also used satellite, observation and NWP data. We present the important steps of the forecasting process of the convection initiation based on this conceptual model. We also analyze and upgrade the data base of the ESSL on tornadoes in Romania for SE part of the country, and add radar information on reported cases, for the 2002-2010 period.

Paper 212:

Radar and in situ observations in a winter bow echo and associated meso-vortices over the Japan Sea area

Kenichi Kusunoki, Meteorological Research Institute

Hanako Inoue, Meteorological Research Institute

Masahisa Nakazato, Meteorological Research Institute

Kotaro Bessho, Meteorological Research Institute

Shunsuke Hoshino, Meteorological Research Institute

Wataru Mashiko, Meteorological Research Institute

Syugo Hayashi, Meteorological Research Institute

Hiroyuki Morishima, East Japan Railway Company

Keiji Adachi, East Japan Railway Company


On Japan railroads, wind conditions affect operating efficiency, infrastructure, and safe passage of people and freight. For instance, strong and gusty winds cause regional delays or shutdowns, and especially hazardous crosswinds may lead to overturn of railcars. The Shonai area railroad weather project will investigate fine-scale structure of wind gust dynamics and kinetics such as tornadoes, downbursts, and gustfronts. Winter tornadic storms over the Japan Sea area have been observed during this project between 2007 and present. In this presentation, we will introduce radar and in-situ measurement in a winter bow echo on January 24 2008. The major features of the winter bow echo are as follows. 1) Radar high temporal observations (at 30-s interval) at low elevation angle reveal clearly the transition from the bow shaped echo to comma-shaped echo. 2) Doppler wind fields confirm that the strong rear-inflow jet (RIJ) existed at the bow apex. Several mesovortices were also observed along the leading edge of the bow. 3) The region of the bow apex was passed over the surface station in the study area and in situ measurement data for the RIJ eternalized. Furthermore, in situ measurement data for a mesovortex embedded in the bow echo were obtained and analyzed simultaneously.

Paper 213:

The Vertical Structures within a winter tornadic storm during landfall over the Japan Sea area

Kenichi Kusunoki, Meteorological Research Institute

Hanako Inoue, Meteorological Research Institute

Kenichi Shimose, Alpha-denshi / MRI

Masahisa Nakazato, Meteorological Research Institute

Kotaro Bessho, Meteorological Research Institute

Shunsuke Hoshino, Meteorological Research Institute

Wataru Mashiko, Meteorological Research Institute

Syugo Hayashi, Meteorological Research Institute

Ken-Ichiro Arai, East Japan Railway Company

Masahide Nishihashi, Alpha-denshi / MRI

Hiroyuki Morishima, East Japan Railway Company

Keiji Adachi, East Japan Railway Company


Winter tornadic storms over the Japan Sea area have been observed during the Shonai Area Railroad Weather Project between 2007 and present. The MRI portable X-band Doppler radar (XPOD: X-band, POrtable Doppler radar) has been installed on the roof of Shonai Airport building in Sakata City during this project. The volume scan radar data were obtained at very close range (<1km) in one tornado on 25 January 2008. In this presentation, the detailed vertical structures of the high-resolution reflectivity and velocity fields of this tornado during landfall is described. Before landfall, the reflectivity field revealed that the center of the tornado had a weak-echo "eye" partially surrounded by an eyewall. After landfall, the eyewall was transformed to spiral structures spiraling outward from the eye. The diameter of the maximum tangential wind contracted from 320m to 160m at the height near 300m. The radar data indicate that the tornado was tilted downstream and the diameter increased with height. The landfalling structure of the other tornado will be also described with some images from a web camera mounted next to the antenna in order to correlate visual features with the radar data.

Paper 214:

Monitoring climatic extremes with the European Climate Assessment & Dataset (ECA&D)

Gerard Van Der Schrier, Royal Netherlands Meteorological Institute (KNMI)

Albert Klein Tank, Royal Netherlands Meteorological Institute (KNMI)

Else Van Den Besselaar, Royal Netherlands Meteorological Institute (KNMI)


Monitoring extremes in climate and the changes in the frequency of occurrence of these extremes heavily depends on the availability of a quality controlled climate dataset. These datasets are usually archived at the National Meteorological Services and not easily accessible by the research community. The need for exchanging high resolution observational data is the motivation to start the European Climate Assessment & Dataset project (ECA&D) in 1998. In this project the participating countries collaborate successfully to form an international observational dataset with daily data for a large area covering Europe, inclusing northern Africa and the Middle East. The database contains daily station data for over 3600 stations, and this number is continuously increasing. The elements available are temperature (daily max., min, and average), daily sums of precipitation, sea-level pressure, wind gust, daily averaged windstrength and five others. Based on the station data for these elements, 60 descriptive indices of extremes and trends therein are derived. These include number of days with extreme precipitation, number of days with extreme temperatures and a storminess index. Also a daily gridded observational dataset (E-OBS), consisting of minimum, mean and maximum temperature, precipitation amount and sea level pressure, is derived. The website offers possibilities to monitor and analyse observed changes in weather and climate extremes by making maps showing trends, anomalies or return values of the indices of extremes. There are also fact sheets of extreme events available as well as more detailed metadata such as station history and pictures of observing sites. In its capacity as WMO-endorsed Regional Climate Centre (RCC) for Europe, ECA&D cooperates with Meteoalarm (which communicates severe weather alarms for Europe) to suggest warning thresholds which give a pan-European uniform severe weather issue frequency. In the presentation, the possibilities and benefits of using ECA&D for the research community are highlighted.

Paper 215:

A 300 years climatology of tornadoes, waterspouts and funnel clouds over Greece

Ioannis Matsangouras, Laboratory of Climatology and Atmospheric Environment, Faculty of Geology and Geoenvironment, University of Athens

Panagiotis Nastos, Laboratory of Climatology and Atmospheric Environment, Faculty of Geology and Geoenvironment, University of Athens

Michalis Sioutas, ELGA-Meteorological Applications Centre


It is widely known that tornadoes and waterspouts could be characterized as the most spectacular whirlwind phenomena, associated in many cases with extreme convective weather, causing extended damage and even loss of life. The complex inland terrain of Greece along with the Ionian Sea at the west and the Aegean Sea at the east, appear to be an area for fury phenomena such as tornadoes, waterspouts and funnel clouds. In this study we present the spatial and temporal variability of tornadoes, waterspouts and funnel clouds from the 18th century to the latest reports up to 2010 over Greece. The extensive dataset was based on academic literature, administrative records, Hellenic National Meteorological Service reports and to our open-ended database of whirlwind events in Greece (http://tornado.geol.uoa.gr), which has been developed and maintained by the Laboratory of Climatology & Atmospheric Environment of the University of Athens. The findings of the performed analysis showed that specific sub-regions favour these extreme phenomena, while the time of formation was examined. In addition to the climatology of tornadoes, waterspouts and funnel clouds over Greece, the daily composite anomalies of meteorological variables (NCEP/NCAR Reanalysis), such as geopotential height, pressure, wind shear etc, at several levels of middle and lower atmosphere, during each recorded event are analyzed and clustered. The goal of this approach is to present significant weather patterns of occurrence for tornadoes and waterspouts over Greece, aiming at the development of an early warning system of such fury phenomena.

Paper 216:

Observation study during the pre-monsoon season over northeast Bangladesh

Masashi Kiguchi, The University of Tokyo

Fumie Murata, Kochi University

Toru Terao, Kagawa University

Azusa Fukushima, Tokyo Metropolitan University

Taiichi Hayashi, Kyoto University

Taikan Oki, The University of Tokyo


The moisture variability during pre-monsoon season over Bangladesh was investigated using the upper balloon observation, OLR, and reanalysis data sets. As synoptic phenomena, the passage of trough on middle troposphere and inflow of warm and moisture air from Bay of Bengal was suggested from these aspects; high potential temperature on lower troposphere during active convection, increase of moisture quantity and relative humidity, prevailing of southwesterly, and decrease of potential temperature on middle troposphere by intensive observation of upper balloon on Dhaka, Bogra, and Sylhet. According to the OLR spatial pattern in each year, the northern India has no rainfall during pre-monsoon, even though in same latitude of Bangladesh. On the other hand, the northern Myanmar, and the southern China as same latitude with Bangladesh have also rainfall activity during pre-monsoon. It is suggested that the active phase of convection during pre-monsoon is the synoptic phenomena. A large lower troposphere integrated moisture flux flows into Bangladesh from the northern India during active phase of convection. In contrast, this moisture flux is weak during break phase of convection. It is interesting that the northern India where has no pre-monsoon rainfall, is located the large moisture flux. The existence of upper trough is confirmed around Bangladesh during active phase of convection from composite analysis of wind fields on 500 hPa and 300 hPa. From the viewpoint of local circulation, it was observed that a strong subsidence in 1–5 km over Dhaka and Bogra and in 4–5 km over Sylhet, whereas topographic uplifting of southwesterly airflow in 1–3 km was observed over Sylhet. Temperature profiles of Sylhet up to 3 km were lower than that of other stations, and easterly wind was observed at the surface.

Paper 217:

Three-dimensional characteristics of winter lightning observed in the Shonai area railroad weather project

Masahide Nishihashi, Alpha-denshi / Meteorological Research Institute

Kenichi Shimose, Alpha-denshi / Meteorological Research Institute

Kenichi Kusunoki, Meteorological Research Institute

Syugo Hayashi, Meteorological Research Institute

Kenichiro Arai, East Japan Railway Company

Hanako Inoue, Meteorological Research Institute

Wataru Mashiko, Meteorological Research Institute

Osamu Suzuki, Meteorological Research Institute

Kotaro Bessho, Japan Meteorological Agency

Shnsuke Hoshino, Meteorological Research Institute

Masahisa Nakazato, Japan Meteorological Agency

Hiroshi Yamauchi, Meteorological Research Institute

Masako Kusume, Alpha-denshi / Meteorological Research Institute

Hiroyuki Morishima, East Japan Railway Company

Keiji Adachi, East Japan Railway Company


In order to develop an automatic strong gust detection system for railroad, the Shonai area railroad weather project has investigated fine-scale structure of wind gust using two X-band Doppler radars and the network of 26 surface weather stations since 2007. In 2009, the project was expanded and started lightning observation to investigate the mechanism of winter lightning and the application to strong gust prediction. Lightning discharge is known to be related to microphysical and dynamical processes within storms. Many scientists have indicated that lightning activity is associated with severe weather. Therefore, integration of continuous three-dimensional (3D) lightning monitoring (intracloud and cloud-to-ground lightning) and comprehensive high-density meteorological observation can provide useful index for predicting strong gust. We developed a lightning observation system. Azimuth and elevation of VHF radiation sources originated from lightning flashes are computed using arrival time difference of three VHF pulses received at three antennas in one site. After operation test at Meteorological Research Institute, we installed this system in the north of Shonai, coastal area of the Japan Sea, in October 2009. Moreover, we constructed three lightning observation sites in the Shonai area in September 2010, in order to visualize lightning discharges in 3D. Our sensors detected lightning discharges at 16:13:32 UTC on 3 December 2010. Using the VHF waveform data, we conduct 3D lightning mapping. The duration of discharge is about 60 ms and divided into two stages. The locations of radiation sources are compared with the X-band radar echo data. As a result, the distribution of VHF radiation sources is consistent with the rim of the strong echo region (around 40 dBZ). The lightning flash was also recorded with the network cameras at each site. In this presentation, we will show the lightning discharge process and the characteristics of the winter thunderclouds in detail.

Paper 218:

Numerical simulation of low-level misocyclones associated with winter convective cells over the Japan Sea area.

Ken-Ichi Shimose, Alpha-denshi / MRI

Shugo Hayashi, Meteorological Research Institute

Wataru Mashiko, Meteorological Research Institute

Kenichi Kusunoki, Meteorological Research Institute

Kotaro Bessho, Japan Meteorological Agency

Shunsuke Hoshino, Meteorological Research Institute

Hanako Inoue, Meteorological Research Institute

Masahisa Nakazato, Japan Meteorological Agency

Hiroshi Yamauchi, Meteorological Research Institute

Keiji Araki, Railway Technical Research Institute

Hiroyuki Morishima, East Japan Railway Company

Keiji Adachi, East Japan Railway Company


This study describes the numerically simulated structure of low-level misocyclones associated with winter convective cells observed at 11 December 2008 during the Shonai area railroad weather project (Shonai area is located in the coast of the Sea of Japan, ~39N). In this case, convective cells were initiated along the cold front over the Japan Sea and misocyclones were embedded in the surface convergence line. Misocyclones were distributed at 2-4 km intervals and some of them landed. The nonhydrostatic numerical model, JMA-NHM (Japan Meteorological Agency – Non-Hydrostatic Model), is used to simulate the misocyclones. The model is initialized by MANAL, which data is an objective analysis made by JMA and horizontal resolution is 10 km, and no special observed data is assimilated in this case. The finest horizontal grid spacing is 250 m. The model succeeds in simulating convective cells and misocyclones along the surface convergence line. The simulated misocyclone is initiated below the height of 500 m, where the horizontal convergence associated with the cold front is significantly strong. Then the misocyclone extends upward and reaches the height of 2 km. The vorticity analysis indicates that the stretching term is dominant for increasing the vertical vorticity of the misocyclone. It is suggested that the low-level misocyclone is first generated by the horizontal shear instability along the cold front and then stretched by the updraft of the convective cell.

Paper 219:

T-re plots generated from MSG data in severe storms forecasting – testing in Central Europe

Michal Pokorny, Charles University in Prague, Faculty of Mathematics and Physics, Department of Meteorology and Environment Protection

Michal Zak, Charles University in Prague, Faculty of Mathematics and Physics, Department of Meteorology and Environment Protection


T-re plots derived from satellite data represent vertical distribution of effective particle size, which is closely related to updraft strength inside convective storms, developing within the target area. Considering the shape of this plots and the particle size, special severe storm forecasting technique was tested in the USA. It was used to forecast and predict dangerous phenomena occurring in severe storms. There are tests done for the Central Europe also. This paper presents preliminary results of testing the T-re plots concept during selected severe storms episodes in the Czech Republic. Discussion of the possible benefits for Czech forecast service is also given.

Paper 220:

Nowcasting and very short range forecasting of supercell thunderstorms in a weakly- or moderately-sheared environment

Kálmán Csirmaz, Hungarian Meteorological Service

Zoltán Polyánszky, Hungarian Meteorological Service

Georg Pistotnik, Central Institute for Meteorology and Geodynamics

André Simon, Hungarian Meteorological Service

Michal Neštiak, Slovak Hydrometeorological Institute

Zsolt Nagykovácsi, Hungarian Association of Stormchasers and Storm Damage Surveyors

Alois Sokol, Comenius University, Faculty of Mathematics, Physics and Informatics


The study investigates several cases of supercell thunderstorms, which occurred in Austria, Slovakia, and Hungary and formed in an environment of relatively low- or moderate wind shear (well below 20 m/s) in the lowest 6km layer of the troposphere. Development of supercells is usually not expected by forecasters in such situations, in contrast to high wind shear cases, where the possibility of thunderstorm rotation is mostly indicated by convective parameters computed from soundings or large-scale atmospheric models. For the selected cases, the properties of the thunderstorm environment were examined by several diagnostic parameters and convective indices (CAPE, BRN, SREH etc.) computed from soundings and analysis of the INCA and MEANDER nowcasting systems. According to these, the conditions preferred the development of multicell thunderstorms rather than supercells. Still, the high-resolution, non-hydrostatic simulation of the cases performed by WRF produced at least one supercell with persistent mesocyclone both at mid- and low-levels (in agreement with observations of Hungarian storm-chasers). Vorticity equation terms and helicity fields were analysed to explain the origin of the rotation and its relation to the environmental wind shear and wind profile. The results indicate that vertical vorticity generated first at nearly 1.5 km height by tilting of horizontal vorticity. At higher levels, the subsequent positive vertical vorticity tendency is largely produced by both tilting term and the vertical advection of the vertical vorticity from below. Unlike to cases with high environmental shear, the horizontal vorticity was rather produced locally, in the close vicinity of the simulated thunderstorm. The study evaluates the accuracy of the numerical simulations in predicting the storm propagation and evolution and compares the results with the outputs of the nowcasting systems. Adjustments of nowcasting techniques and of the current operational diagnostics are proposed to improve the forecasting capabilities of similar events.

Paper 221:

Lightning detection on Milešovka observatory

Petr Pesice, Institute of Atmospheric Physics ASCR


Lightning detection is important for identifying the areas with strong convection activity. Among other things the lightning data can be assimilated to NWP models to refine the other assimilated information such as radar and satellite data. Milešovka hill is well known for frequent thunderstorm occurrence (no wonder that the German name is Donnersberg). The Vaisala Thunderstorm Local Lightning Sensor TSS928 was installed on the observatory in 2005. In 2006 the sensor was destroyed by the stroke of lightning and installed again in 2007. TSS928 can detect lightning in the maximal distance of 56 km. It can distinguish between cloud and cloud-to-ground flashes. For cloud-to-ground strokes the instrument can report the distance in three range intervals (0-9km, 9-19km and 19-56km) and direction in eight sectors. The comparison of lightning sensor data with the data from Central European Lightning Detection Network (CELDN) will be presented. We will focus mainly on timely and space correlations between both data sources. With regard to coarser space resolution of TSS928 the data in individual range intervals and sectors will be compared. Radar data from the CZRAD network will be used to position the lightning more precisely. Thunderstorm cells will be identified in radar images within the instrument range and these cells will be considered as the lightning source in the individual measurement sectors. Google maps based web application for visualization of lightning sensor data will be presented. This application will be available on IAP web pages.

Paper 222:

Variations of Atmospheric Moisture for South America – Global Teleconnections

Humberto Barbosa, Laboratório de Análise e Processamento de Imagens de Satélites – LAPIS

Ivon Da Silva Junior, Laboratório de Análise e Processamento de Imagens de Satélites – LAPIS


Estimates of atmospheric water vapor are of vital importance to meteorological, correct atmospheric imagery satellite and climatological modeling, contributing to the understanding of a wide variety of processes from small-scale weather systems to global climate change. The water vapor was extracted from the daily records of atmospheric sounding compiled by Department of Atmospheric Science, from the University of Wyoming. It is made up of a total of 1,312 stations around the world, of which 95 are in South America. Each sounding is composed of records of pressure, geopotential height, temperature and water vapor in 17 mandatory levels. The variations of the water vapor based on different geographical regions of South America are for the respective test years are compared with the normal (1979 to 2011) to understand the impact of different modes of global teleconnections during the austral summer period over South America. Attempts are also made to emphasize the latitudinal variations in the water vapor as it revealed the gradual decline with the increase of latitude over South America. The paper then explains the relation between all South America water vapor with the ENSO indices such as Sea Surface Temperature of Nino 3 region and the Southern Oscillation Index.

Paper 223:

The biggest damage by hailstorm in Kagoshima Prefecture, Japan on Feb.25, 2009

Kenji Wakimizu, Kyushu University

Koji Nishiyama, Kyushu University


The number of damages by hailstorm shows the tendency to increase recently in Kagoshima Prefecture, Japan. For example, a heavy damage by hailstorm was happened in Ibusuki City (in southern part of Kagoshima Prefecture) on Feb.25, 2009. The details of this damage were vegetables (e.g. broad bean and cabbage), flowers and agriculture facilities (e.g. hard, plastic house and vinyl house). The sum total of this damage was 410million yen. This was the biggest damage by hailstorm for 56 years (1955-2010). The aim of this study is research of the meteorological reason why this biggest damage by hailstorm was happened. We obtained the following main meteorological results. 1)Low pressure advanced from west toward east in the vicinity of the southern part of Kagoshima Prefecture on Feb.24-25, 2009. 2)Ground air temperature trended to descent at 1800JST on Feb.4 in Ibusuki City. However, air temperature rose at about 2C at 0030JST on Feb.25, when hail fell. And air temperature rose about at 4C by the wind in the south and southwest. 3)The wind direction of AMeDAS {Automated Meteorological data Acquisition System operated by JMA(the Japan Meteorological Agency)} observation points in the center of Kagoshima Prefecture changed greatly from 0040JST to 0050JST on Feb.25. The cold front moved from northwest to southeast. 4)In other words, a strap strong rainfall area moves from northwest to southeast in Kagoshima Prefecture. This can be said from the JMA radar echoes, too. It is thought that this hailstorm was happened by many tall cumulonimbi generated along this cold front, and the influence of geographical features of mountains.

Paper 224:


Michalis Sioutas, ELGA


Hail is classified among the major weather threats, resulting in significant economic losses to agriculture and property. Establishing of hail climatologies is of a great importance, offering both to hail research and to estimation of risk and expected level of damage. Climatological data can also be significant inputs in natural catastrophe modelling, in quantifying the hail risk and the financial loss in the insurance and reinsurance industry. In this research, a climatological analysis of hail occurrence in Greece is presented, based on the meteorological stations data of the Hellenic National Meteorological Service (HNMS). Hail records are provided from about 50 meteorological stations span the entire country and in operation during the 30-year "climatological reference" period 1961-1990. The Kriging method was used to grid the irregularly spaced station data and produce hail contour patterns. Influence of latitude and distance from the sea based on GIS data, was examined. Spatial patterns and distribution of hail day frequency was also studied, by using of Principal Component Analysis. Based on the HNMS stations data, hail is occurred with a higher frequency over the western areas of Greece. The highest hail day number is determined for the western parts of Greece with a yearly average “point” frequency ranged from 4 to 5 hail days. Hail days generally increase from the east to west and from coastal to interior and to mountain areas. In the Aegean Sea, hail days increase from west to east, towards the coasts of Turkey. A seasonal hail variability is also revealed, with the continental and north-central parts of Greece affected by hailfall activity mainly during the warm period of the year (April to September), while, coastal western and southern Greece is mostly affected during the cold period (October to March).

Paper 225:

Investigating Hector convective development by microphysical analysis using TRMM satellite data and high resolution model simulations

Sabrina Gentile, Department of Physics/CETEMPS, University of L'Aquila, L'Aquila, Italy

Rossella Ferretti, Department of Physics/CETEMPS, University of L'Aquila, Italy

Frank Silvio Marzano, CETEMPS/Department of Information Engineering, Sapienza University of Rome, Via Eudossiana 18, 00184 ITALY


One event of the convective cell Hector, a storm developing in the tropical northern part of Australia, have been analyzed using high resolution MM5 Mesoscale Model simulations, Berrimah ground-based radar reflectivity and TRMM satellite data. The analysis has been performed focusing on the different convective development phases of the storm. The first stage corresponds the pre-Hector phase characterized by non precipitating organized cells; a brief analysis has been performed to verify their compatibility with Rayleigh-Bénard convection. In the second stage weak precipitation starts and merges into a convergence line in the central part of the islands producing an intense growth of the convection. This phase corresponds to the mature stage; this maximum development of the storm is characterized by a cloud top height reaching the tropopause and updraft speeds of several tens meters per seconds. At this stage a further comparison of the microphysical content has been carried out using the observations from TMI of the TRMM platform evaluating the mean vertical profiles of four hydrometeors: cloud liquid water, precipitating water, cloud ice and precipitating ice. Finally the cell decays and the precipitation stops. The comparison between MM5 and radar reflectivity fields suggests a fairly good model reproduction of the Hector dynamics. Moreover, the comparisons between the model and observation with TRMM Microwave Imager show a good agreement even though a test of reliability for the TMI data is necessary to asses a proper use.

Paper 226:

Lightning nowcasting towards lightning warnings using geostationary satellite data

Prof. John Mecikalski, University of Alabama in Huntsville

Lori Schultz, University of Alabama in Huntsville


Nowcasting 0-1 hour lightning initiation (LI) from geostationary satellite data is a novel, new concept that leverages a mature convective initiation algorithm, the SATellite Convection AnalySis and Tracking (SATCAST) system. Harris et al. (2010) quantified the behavior of ten Geostationary Operational Environmental Satellite (GOES-12) infrared interest fields in the 1-hour in advance of LI, within the SATCAST system. A total of 172 lightning-producing storms that occurred during the 2009 convective season were manually tracked and studied over four regions of the US: Northern Alabama, Central Oklahoma, the Kennedy Space Center and Washington D.C. Four-dimensional cloud-to-ground total lightning mapping array data provide a total cloud lightning picture (in-cloud, cloud-to-cloud, cloud-to-air, cloud-to-ground) and thus precise LI points for each storm in both time and space. Main results indicate that at least a 15 minute advance nowcast of LI can be made with SATCAST, with 35 minutes being the average, and 1 hour being the maximum. The goal now is towards implementing an LI prediction algorithm into realtime satellite-based systems. The hypothesis of this work is that enhanced predictability of lightning via LI nowcasts will lead to more efficient airport terminal operations to positively impact aviation, and can also be used to enhance so-called lightning warnings as produced by National Weather Service offices. In addition, satellite-based LI nowcasts will enhance 0-1 hour convective initiation (CI) nowcasts offshore, beyond the range of NEXRAD radars. This presentation will highlight recent advances to incorporate the SATCAST LI methodology into the Corridor Integrated Weather System (CIWS) as part of the US Federal Aviation Administration's (FAA) 0-8 hour convective weather nowcast system, and also toward advancing geostationary satellite-based LI nowcasting capabilities to meet public sector interests. Details on how the GOES-based algorithm applies to the Meteosat Second Generation SEVIRI sensor will also be provided.

Paper 227:

Evolution of Mesoscale Convective System in Southern Serbia

Maja Rabrenovic, Hydrometeorological Service of Serbia

Tatjana Majstorovic, Hydrometeorological Service of Serbia


This paper analyses the use of WV satellite maps and PV charts for forecasting of initiation of deep convection in the atmosphere. These prediction tools are compared with classical forecasting charts. It is argued that classical forecasting charts are not sufficiently precise to forecast convective development. A case of evolution of mesoscale convective system that occurred in Southern Serbia on 19th of June 2007 is used to illustrate these points. The paper explores the reasons for occurrence of this mecoscale convective system and its evolution into bow echo by using radar and satellite images.

Paper 228:

Sensitivity to initial condition and model resolution of precipitation forecasts modelled by the hydrostatic BOLAM model over the Mediterranean Basin

Marco Casaioli, ISPRA - Institute for Environmental Protection and Research

Piero Malguzzi, ISAC-CNR - Institute of Atmospheric Sciences and Climate-Italian National Research Council

Stefano Mariani, ISPRA - Institute for Environmental Protection and Research


Since 2000 the meteo-marine forecasting system called Sistema Idro-Meteo-Mare (SIMM) is operational at the Institute for Environmental Protection and Research (ISPRA, former APAT). The system, which runs on a SGI-Altix parallel machine, is formed by a cascade of four numerical models, telescoping from the Mediterranean basin to the Venice Lagoon, and is initialized by means of ECMWF analyses and forecasts. A hydrological model, tailored over two Italian river basins, is also integrated in a research configuration into the SIMM forecasting chain in cascade to the meteorological BOLAM model. This latter model, which is at the basis of the forecasting chain and covers the entire Mediterranean Basin, has been recently fully updated by implementing the new version developed by the Institute of Atmospheric Sciences and Climate of the Italian National Research Council (ISAC-CNR). Such new version includes, among others, more efficient parameterization schemes for cumulus convection, radiation, soil and turbulence. In the framework of the forthcoming system development activities, the authors decided to test possible improvements of the quantitative precipitation forecast in terms of input data, horizontal grid step and extension of the BOLAM domain. In order to have a statistically robust assessment of such improvements the MAP D-PHASE observational dataset, which ranges from June to November 2007, is considered. Thus, over the six-month MAP D-PHASE period, two series of simulations are compared against a control series obtained with the recently updated BOLAM model by means of categorical skill scores. The first experiment is performed by using initial and boundary conditions at higher time and space (horizontal and vertical) resolutions; the second one includes the model resolution enhancement and the increase of the domain extension. The two main aims of this sensitivity study are to test the overall performance of a hydrostatic model when grid step is pushed significantly below 10 km and to evaluate how precipitation forecast in the Mediterranean environment is sensitive to the domain extension of the meteorological model.

Paper 229:

A Nowcasting Approach to Detecting South African Hazardous Convective Storms

Caroline H. Hardy, University of Johannesburg

André L. Nel, University of Johannesburg


This work extends the previous study of Hardy and Nel that defined the Hazardous Potential Index (HPI) to detect potentially hazardous convective storms over South Africa. Hazardous storms are destructive and pose a threat to life and property. The HPI was designed to identify storms, in the absence of ancillary meteorological radar data, using only satellite imagery from Meteosat Second Generation (MSG)/SEVIRI. Potentially hazardous storms are identified by evaluating indicators of hazardous storms in infrared data from the Meteosat-8 and Meteosat-9 satellites. The HPI made use of six indicating factors; namely the cumulus cloud, active region, overshooting top, small ice particle, active region growth and cumulus cloud (rainfall estimation) temperature factors. The false alarm rate associated with the HPI may be reduced through an adaptation of the HPI algorithm by removing the cumulus cloud and active region factors and by incorporating an evaluation of the change in the cloud top temperature of a candidate storm. A decrease in the average cloud top brightness temperature is indicative of vertical development within the candidate storm. The modified HPI algorithm monitors the areal growth and brightness temperature changes of candidate regions of cumuliform cloud, establishes the presence of overshooting cloud tops and small ice particles in the cloud by using multispectral band-differencing and estimates the rainfall rate associated with a candidate cloud. The sum of these indicator factors is used to generate a graphical product that may be used in nowcasting applications to identify hazardous storms. The performance of the HPI and the modified HPI are evaluated against the CAPSAT ``Severe Convective Storms'' composite product over 20 historical case studies to demonstrate the effectiveness of the HPI products in detecting hazardous convective storms and to illustrate the improved performance of the modified HPI algorithm.

Paper 230:

Mesoscale Convective Vortices in Nothern East Serbia

Maja Rabrenovic, Hydrometeorological Service of Serbia


The objective of the paper is to analyise emergence of two mesoscale convective vortices, which commenced in Samos (South-Eastern Serbia) and moved into Eastern Romania on the occasion of 30 Juny 2006 and on 19 May 2008. Both MCVs were accompanied with intense convection, which produced a wide range of severe weather throughtout Serbia. The paper shall examine the instability conditions which facilitated occurrence of MCVs in order to be able to predict future occurences of this circulation complex in the region.

Paper 231:

Comparison of May and August 2010 floods in Central Europe

Marek Kaspar, Institute of Atmospheric Physics ASCR

Miloslav Müller, Institute of Atmospheric Physics ASCR

Jozef Pecho, Institute of Atmospheric Physics ASCR


Two significant river floods occurred during the warm season 2010 in Central Europe. In May, heavy regional rains affected the Czech Republic, Slovakia and Poland. Intense precipitation lasted several days and produced large flooding in many rivers (Odra, Wisla, Váh etc.). In August, on the contrary, rains were more concentrated in space and time; they mainly affected the border regions of the Czech Republic, Germany and Poland and produced extreme flooding there. Our presentation compares the events with respect to their hydrological course as well as their meteorological causes. From the hydrological point of view, the impact of precipitation distribution and antecedent saturation on the course of floods is studied. The main emphasis is put on the meteorological analysis and comparison of the events. It is based on NCAR/NCEP reanalysis data. Our previous studies confirmed that circulation conditions before rain floods in Central Europe were characterized by significant large-scale anomalies in selected dynamic and thermodynamic variables. Nevertheless, the events can substantially differ in the size and extremity of these anomalies. Therefore, we evaluate magnitude of the characteristic anomalies before and during the May and August 2010 floods to interpret the differences in their course from the meteorological point of view.

Paper 232:

Assimilation of Water Vapour Airborne Lidar Observations: Impact Study on the COPS Precipitation Forecasts

Evelyne Richard, CNRS

Soline Bielli, CNRS

Matthias Grzeschick, Universty of Tübingen

Cyrille Flamant, CNRS

Christoph Kiemle, DLR


The Convective and Orographically-driven Precipitation Study carried out in summer 2007 over northeastern France and southwestern Germany provided a fairly comprehensive description of the low-troposphere water vapour field thanks, in particular, to the deployment of two airborne differential absorption lidar systems. These lidar observations were assimilated using the 3D VAR assimilation system of the Application of Research to Operations at MEsoscale (AROME) numerical weather prediction mesoscale model. The assimilation was carried out for the period of 4 July -3 August by running a continuous sequence of 3-hour forward intermittent cycles. First, the impact of the lidar observations was assessed by comparing the analyses with a set of more than 200 independent soundings. The lidar observations were found to have a positive impact on the analyses by reducing the moist bias in the first 500 m above ground level and by diminishing the root mean square error by roughly 15\% in the first km. Then, the impact of the lidar observations was assessed by comparing the precipitation forecasts (obtained with and without the lidar observations for the period of 15 July-2 August) with the gridded precipitation observations provided by the Vienna Enhanced Resolution Analysis. In general, the impact was found to be positive but not significant for the 24h precipitation and positive and significant for the 6h precipitation, with an improvement lasting up to 24h. Some selected case studies show that the improvement was obtained through a better depiction of the convection initiation or through a more accurate positioning of the precipitation systems.

Paper 233:

Explicit simulations of electrified mixed-phase clouds in MesoNH: from semi-idealized to first real-case studies

Jean-Pierre Pinty, Laboratoire d'Aerologie

Christelle Barthe, Laboratoire de l'Atmosphère et des Cyclones

Michel Chong, Laboratoire d'Aerologie


We present recent results obtained with the revised version of the explicit cloud electrification module developed in the french mesoscale model MesoNH (Barthe and Pinty, 2007). The module includes three ingredients: 1/ a prognostic budget equation of the electrical charges carried by each type of hydrometeor and by the positive and negative ions, 2/ a computation of the electric field by inverting the Gauss equation and 3/ a statistical algorithm to produce intra-cloud and cloud-to-ground discharges. Electrical charges are generated by non-inductive processes involving ice particles that collide in the presence of supercooled water. The geometry of the lightning flashes obeys a fractal law that characterizes the densely branched flash structures with a marked horizontal extend in the charged regions of the clouds. We first show results obtained for the semi-idealized STERAO test case of severe deep convection (Skamarock et al., 2000 and 2003; Barthe et al., 2007) where thousands of flashes were recorded. For this experiment, the convection is triggered by three warm bubbles in the low levels. At 1 km resolution, MesoNH is able to reproduce the storm structure that evolves from multicellular to supercellular. The electrification processes follow the evolution of the storm with a period of transition that is characterized by a reduction of the flash number as observed. The overall simulation retrieves the huge number of intra-cloud flashes with a good approximation. The next test case is taken from the COPS experiment (Wulfmeyer et al., 2008) that took place in the SW of Germany during summertime in 2007. Here we simulate an isolated storm that forms above the Black Forrest and then produces a few intra-cloud flashes. The simulation is initialized from meteorological analyses and MesoNH is run using the grid nesting facility. The inner model is able to timely capture the short lasting convective event and to produce a dipole structure of charge in the deep cloud. For this case the occurrence of flashes, a few hundreds were detected, is depending on whether a hail category of hydrometeor is included or not in the simulation. The last case is closely related to the Hymex purpose to simulate heavy precipitation over the Cevennes area in the South of France. MesoNH is initialized with "Arome" analyses of Meteo-France for the 7-8 September 2010 events that produced locally 250 to 300 mm of precipitation on the Cevennes ridge. A 48 hour simulation is performed directly at 1 km resolution with MesoNH over a 380X380 km^2 area in order to check the ability of the model to reproduce the very fine scale structure of the rainfall. A good agreement is found between the rain gauges and the MesoNH outputs (rates and locations). The next step is to restart the model now running with the electrical module in order to simulate the flash rates during short periods of time and to compare them with the records of several detection networks.

Paper 234:

On the Difficulties of separating Downburst Damages from Tornadic Damages of fast-moving Storms: Case Study of a high-impact Event in Austria

Georg Pistotnik, Central Institute for Meteorology and Geodynamics (ZAMG)

Rudolf Kaltenböck, Austrocontrol GmbH, Aeronautical Meteorological Service


23 July 2009 was an outstanding severe weather day in Austria and other parts of Central Europe. In the early evening, two supercells formed over Switzerland and Southern Germany in a prefrontal “loaded gun” environment of high instability, a pronounced cap and strong vertical wind shear. While rapidly moving eastward into Austria, these supercells left a path of destruction with hailstones of up to 10 centimetres. Later on, the thunderstorms took on a more line-shaped appearance, and consequently their main threat shifted to widespread severe wind gusts which caused extensive damage and probably met the criteria of a derecho. Several dozens of people were badly injuried. All in all, this day was the most costly day in the 60-year history of the Austrian Hail Insurance company. Site surveys in the surroundings of Vienna, one of the areas particularly affected by the mature squall line, revealed a series of very concentrated forest damages, situated at the Southern flank and often detached from the widespread downburst damages. The convergent pattern of uprooted trees and other characteristics allowed the conclusion that two tornadoes (one of them rated F2) had left their marks, although a closed circulation never developed due to the rapid displacement of the storms. This study aims to elaborate on the difficulties of separating downburst damages from tornado damages in situations when the parent storm exhibits a high translation speed. Furthermore, radar data, objective analysis fields and eye observations are used to address the question if indeed another supercell storm had managed to form at the Southern flank of the squall line, or if the tornadoes were rather gustnadoes caused by the interaction of the downburst with hilly and forested terrain. Implications on the judgment of other storm damages of the same evening are also included.

Paper 235:

Vorticity induction: alignment of RFD vortex ring segments within the mesocyclone

Marcus Büker, Western Illinois University

Gregory Tripoli, University of Wisconsin-Madison


The baroclinic generation of vortex rings and arches in the rear-flanking downdraft (RFD) portion of a supercell has been shown to be one of the likely mechanisms contributing to tornadogenesis. However, the multi-scale interaction of vorticity in this environment is highly nonlinear and complex. We propose a simplified, quasi-barotropic mechanism for the alignment of RFD-vortex segments with the mesocyclone itself. Observations of vortex arch structures show the two counter-rotating vertical vortices to be oriented on an axis perpendicular to the mesocyclone; thus there must exist a vorticity preference leading to alignment of the ‘near’ (embedded) vortex segment with the mesocyclone. Using an established mathematical analogy with electromagnetism (e.g. vorticity vs. magnetic induction), we view the system of the mesocyclone and RFD vortex ring to behave as that of an induction coil and a partially embedded, flexible toroidal magnet. The embedded (near) portion of an RFD vortex ring experiences a torque toward alignment with the large-scale ‘induction field’ (vorticity) of the mesocyclone. This distribution of torque results in the observed perpendicular orientation of vortex arches, as opposed to a parallel orientation that would result from lifting of the ring by the updraft associated with the mesocyclone. This viewpoint, while inspired from analogous behavior in electromagnetism, can be explained through an analysis of differential accelerations through Lamb vector decomposition, as well as from conservation of angular momentum arguments. We will show animations and analysis of our results from numerical simulations.

Paper 236:

Detection of Convective Initiation by Objective Analysis Methods and its Use for Precipitation Nowcasting

Georg Pistotnik, Central Institute for Meteorology and Geodynamics (ZAMG)

Stefan Schneider, Central Institute for Meteorology and Geodynamics (ZAMG)

Christoph Wittmann, Central Institute for Meteorology and Geodynamics (ZAMG)


Convective situations pose the greatest challenge to automatic nowcasting of precipitation. Conventional nowcasting techniques, which use a translation of the latest precipitation analysis into the near future, cannot anticipate any temporal development of the observed precipitation field and are therefore unable to account for the life cycle of convective cells. Recent work at the Central Institute for Meteorology and Geodynamics (ZAMG) has focused on the predictability of the initiation and life cycle of convective cells, based on analysis data of the INCA system (“Integrated Nowcasting through Comprehensive Analysis”, horizontal resolution of 1 kilometre) and forecast data of the numerical weather prediction model AROME (horizontal resolution of 2.5 kilometres). Analysis fields of Convective Available Potential Energy (CAPE), Convective Inhibition (CIN) and near-surface Moisture Flux Convergence have been used to identify areas that are prone to intensification or weakening of existing convective cells, or to the formation of new ones. The results of these “convective nowcasts” of precipitation for 31 days of the summers of 2009 and 2010 are compared against the conventional nowcasts as well as against a previous version of the same convective nowcasts, which used coarser ALADIN forecasts instead of AROME forecasts as a background for the convective analysis fields. While the convective nowcasts show a superior performance for situations of local convection tied to orografic features, the results worsen when the convection gets more widespread and/or supported by stronger dynamic forcing. It is hypothesized that this mainly reflects the differences between primary convection, which is strictly tied to quasi-stationary orographic forcing mechanisms, and secondary convection, which requires stronger transient forcing on local, meso- or synoptic scales. Thoughts on convective initiation in the absence of significant CIN (i.e., primary convection) and in its presence (i.e., secondary convection) are presented, as well as their significance for future work.

Paper 237:

Detailed Diagnostic Analysis of Tornadogenesis within Simulated Supercell Storms

Ming Xue, Center for Analysis and Prediction of Storms, University of Oklahoma

Alexander Schenkman, Center for Analysis and Prediction of Storms, University of Oklahoma

Brett Roberts, Center for Analysis and Prediction of Storms, University of Oklahoma


In this study, detailed diagnostic analyses are performed on several very high-resolution simulations of supercell storms. The simulations used horizontal resolutions up to 12.5 meters, and included different microphysics options. In some simulations, the simulated tornado are very realistic and reaches EF5 intensity. Analyses are performed in both Eulerian and Lagrangian frameworks, including detailed force and vorticity component analyses along parcel trajectories. Detailed time evolutions of the vorticity that lead to the intensification of low-level vortex and tornadogenesis will be presented.

Paper 238:


Ricardo S. Tenorio, Universidade Federal de Alagoas

Adriano A. S. Barros, Universidade Federal de Alagoas

Marco A. M. Lemes, Universidade Federal de Alagoas


This study aims to identify a possible correlation between the radar images (reflectivity factor) generated by the Weather Radar System of Alagoas - SIRMAL, and the flow of the Mundaú River in the Northeastern Brazil. The flows data were obtained from the fluviometric station located at the Boa Fortuna Farm in Alagoas State. The floods in the river Mundaú are recurrent, and have brought great losses in material resources and human lives for decades. The monitoring is very important for better planning and management water resources and for civil defense actions. For this work it was assembled a computer code in object-oriented language, Object Pascal. The program reads the image in jpeg format, generated and provided by the SIRMAL, and spatially integrates all the values of radar reflectivity factor on the watershed determining the accumulated volume in the river Mundaú watershed. This value is associated with the reflective state of the basin, at a given time that indicates the amount of precipitated water on the basin for a given period. The accumulated rainfall is related to the flow of the river. To verify the correlation between them, five months of observation from 2004 to 2006 were chosen; take into consideration abrupt changes in flow and important variations in rainfall rate. The results showed that the correlation coefficients exceeding 70%, and from daily analysis of the radar images and rivers flows it was determined a qualitative 24 hour forecast: Low, Normal or High flow. The method shows potential in forecast river flow by the use only of radar images.

Paper 239:

Global comparison of regional severe tornado maxima

Grayum Vickers, Center for Ocean-Atmospheric Prediction Studies


Differences in tornadogenesis mechanisms around the world are assessed for regionally-preferred environmental characteristics. Regional maxima of significant tornado frequency exist due to respective geographical features, such as proximity to water, orography, etc. Variations between these local “tornado alleys” are poorly understood despite the potential for increasing forecast skill by documenting regional tendency relating to geography. Moreover, many regions across the globe with local maxima of significant (EF3+) tornadogenesis are under-represented in scientific studies due to under-reporting and poor information dissemination discrepancies. Building on previous work by Broyles, et al 2003, and Feuerstein, et al 2005, which address the existence of local “tornado alleys”, this work shows an inter-comparison between several different “tornado alleys” and describes potential causes for their existence. Using tornado reports from around the world, topographic and geographical mechanisms are observed to influence the likelihood of tornado formation. Regional “tornado alleys” are defined by comparing the average F-scale mean distribution of tornadoes versus a regional synoptic average. In addition, the meteorological profile of an archetypal significant mesocyclone tornado is created from composited reanalysis data. Such composites are used to bolster the importance of geography in aiding tornadogenesis, as well as to identify specialized meteorological parameters that contribute to severe tornado formation in different regions.

Paper 240:

Extreme precipitation events and their socio-spatial impacts in a Brazilian urban centre: Campinas-SP

Marina Soria Castellano, State University of Campinas (former student)

Luci Hidalgo Nunes, State University of Campinas


The research analysed the socio-spatial pattern of the impacts of extreme precipitation events between 1958 and 2007 for Campinas, Brazil, an important technological centre located 80km north of São Paulo, with 1,080,999 inhabitants (census of 2010) from which at least 5,000 families living in risk areas. Daily extreme precipitation events were defined by means of percentiles (totals exceeding the 90,00th, 95,00th and 95,75th percentiles). The impacts associated with precipitation were collected in different sources, such as newspapers and the municipal Civil Defense. Analyses showed a considerable increase in both the number of the impacts recorded (from 129 in the first decade to 3,837 in the last) and in the number of the types of impacts (from 16 to 31). During the same period the urban area and the population also presented an impressive increase, the latter of 500%. However, any important change in the frequency of extreme precipitation events was recorded, emphasizing that the consequences of extreme precipitation occurrences depend not only on the magnitude of the record itself, but much upon the vulnerability of the affected group. This aspect becomes clear by the socioeconomic evaluations: extreme precipitation episodes affected more routinely and dramatically low income social groupings, but neighborhoods of all social classes were affected, showing that the entire population of Campinas is exposed to risk, although of different magnitudes.

Paper 241:

Observations from VORTEX2: The intensification of the Goshen County, Wyoming tornado (05 June 2009)

Josh Wurman, Center for Severe Weather Research

Karen Kosiba, Center for Severe Weather Research

Paul Markowski, Penn State University

Yvette Richardson, Penn State University

Paul Robinson, Center for Severe Weather Research


The evolution the Goshen County, Wyoming, 5 June 2009 tornado during its intensification period is studied using integrated dual-Doppler and Rapid-Scan radar data as well as mobile mesonet thermodynamic observations, and particle size measurements from disdrometers. Dual-Doppler analyses reveal that the secondary surge, which was present since near tornadogenesis, is strongest several minutes prior to peak tornadic intensity (as measured by the maximum wind speed), but then decreases in strength and coverage as the tornado reaches peak intensity. Additionally, primary updraft strength and vorticity generation also peak prior to maximum tornado intensity. Integrated dual-Doppler and thermodynamic analysis, and trajectory analysis, are used to diagnose the forcing related to tornado intensification.

Paper 243:


Oleksii Kryvobok, UHMI


Severe weather event associated with deep moist convection such as flush floods large hail, damaging winds and even occasional tornadoes are reported every year in Ukraine. Thus, as a support for disaster mitigation procedures, there is an increase demand for assessment of operational strategies for the short-forecast and/or nowcasting of severe convective weather in this country. The forecast of the exact location for severe convective clouds formation several hours before is not possible or is under research. The only way at this moment is to use remote sensing techniques such as radar or satellite measurements to detect and tracking of convective clouds. Because of installation of some EUMETCast stations in the different parts of Ukraine, which provide real-time geostationary Meteosat Second Generation satellite images, the experience of the use of it for detection and tracking of severe convective clouds is discussed. Based on years of experience in the interpretation of the MSG images in research and operational departments of the Ukrainian Hydrometeorological Service the recommendations for the use of satellite images for short-range forecasting and nowcasting of severe convective clouds were developed. The recommendations based on consecutive analysis of different kind of MSG and MSG-based images. HRV, RGB SEVERE STORM and IR10.8 images are the most useful for detection of convective clouds. From our point of view the RGB SEVERE STORM image provides the most accessible images, because it highlights only convective clouds during the daylight. In the night the most effective for detection of severe convective clouds is the use of IR10.8 images for estimation of cloud top temperature or the cooling rate of cloud tops (analysis of two successive IR10.8 images).

Paper 244:

Analysis of the weather and oceanographic conditions that contribute to the garbage accumulation on the southeastern Adriatic coast in Croatia

Martina Tudor, Croatian Meteorological and hydrological service

Ivica Janekovic, University of Hawai'i at Manoa, Honolulu, Hawaii, USA


On 21st November 2010, garbage accumulation has been observed on the southeast coast of Croatia, from Pelješac Peninsula to Dubrovnik. The labels on the garbage items suggested that it arrived mostly from Albania. It is assumed that strong southeasterly wind has intensified the sea current system that brought the garbage and that it washed into the sea by flash floods. These hypotheses will be investigated using available meteorological, oceanographic and hydrological data. The meteorological conditions are analyzed using available satellite and in-situ measurements from the surrounding area as well as the NWP model data obtained using ALADIN model on 8 and 2 km resolution. Satellite data and meteorological model results suggest two cases with strong convection and rainfall over Albania, 8-10 and 17-19 November 2010. The rainfall rate and southeasterly wind was more intensive in the first case. The strength of the southeasterly wind was underestimated by the operational ALADIN model forecast. The water level measurements on the Bojana river and Skadar lake in Montenegro show significant increase in the water level on 9th November 2010, followed by an increase in the water level of the Skadar lake as well. Finally the sea currents responsible for pollution transport are computed using ROMS ocean model, forced with ALADIN model outputs, major Adriatic river inflows and tides at the open boundary. The results from ocean model are used as input to the forward trajectory computations initiated from the sea surface close to the Albanian coast. Most of them end on the coast of southeast Croatia and Montenegro, although not so many as far northwest as Pelješac Peninsula. This can be attributed to the too weak southeasterly wind in the model that used as forcing to the ocean model.

Paper 245:

Heavy rainfall and flash flood in Dubrovnik on 22nd November 2010

Martina Tudor, Croatian Meteorological and hydrological service

Antonio Stanesic, Croatian Meteorological and hydrological service

Dunja Mazzocco Drvar, Croatian Meteorological and hydrological service

Dunja Placko Vrsnak, Croatian Meteorological and hydrological service

Stjepan Ivatek-Sahdan, Croatian Meteorological and hydrological service


Extreme precipitation case in the morning on 22nd November 2010 in Dubrovnik, Croatia, is used as a testbed for NWP model ALADIN. The synoptic situation was characterized by a huge low moving over the Alps with the frontal system covering most of Central Europe and Mid-Mediterranean. During the passage of the cold front over the southern Italy and Adriatic Sea, a secondary cyclone developed and a meso-scale convective system grew within. The rain gauge measurements exceeded 100 mm/24hr in the area and the one in Dubrovnik measured 161.4 mm/24hr, with a peak intensity of 71.5 mm/h. The traffic and electricity supply were interrupted, water flow invaded and damaged homes and roads and the famous Stradun street was flooded by 40 centimetres of water. The operational forecast at 8km horizontal resolution uses ARPEGE initial and lateral boundary conditions (I+LBC) and digital filter initialization. The parallel suite uses data assimilation coupled to ARPEGE. Test suites uses data assimilation coupled to IFS. All three runs that cover the 24 hour period started at 06 UTC on 22nd November 2010 have forecast precipitation ranging from 20 to 100 mm/24hr for Dubrovnik with the amount exceeding 100 mm/24hr in the neighbouring Montenegro. Results of the high-resolution (2km) non-hydrostatic model runs show strong dependency on the model input data from I+LBC, the position of the precipitation maxima in 2km is very similar to the driving 8km run. In the 2km experiments, significant part of the precipitation was given by the convection scheme, when it was switched on. This result suggests the importance of using the convective parametrization in the ALADIN model even for resolutions in which it is assumed that the convection is resolved. Further experiments with a diagnostic convection scheme and without any deep convection scheme will be performed in order to investigate this problem.

Paper 247:






In June and July of 2010 there has been a significant flooding in north-eastern Romania (Moldova), which comes through both water quantity and the disasters recorded in the string of major storm events in Romania. The present study proposes a short review of the events of the period June-July 2010, with emphasis on the synoptic and mesoscale analysis of the structures responsible for the most virulent episodes of the entire interval. In most episodes analyzed the convective systems responsible for severe weather events in the north-east of our country were established in supercell structures (with very high reflectivity on the radar imagery and the typical satellite structures) that developed one behind the other, approximately in the same direction. The presence of an orographic barrier in the same direction was responsible for additional forcing and for a further intensification of the phenomena on the slopes of the northern Carpathians. The synoptic and mesoscale evaluation of cases of severe weather was performed using ECMWF numerical model analysis and also outputs of a limited aria numerical models ALADIN, ALARO and COSMO, satellite products (Meteosat 9) and radar (Doppler), frontal analysis (SatRep) and data from meteorological stations. Also, there have been used daily and monthly weather bulletins issued under the National Meteorological Administration of Romania.

Paper 248:






The warm season 2010 was characterized by strong contrasts in the entire European territory that alternated periods of extremely high temperatures and excessive rains, causing flooding. In this respect the flooding in Central and Eastern-Central Europe (Czech Republic, Poland, Germany, Hungary, the Baltics, Moldova, etc.) have detached including the ones in Romania in the period June-July 2010. The recorded amounts of water and the long period of time in which have been reported flooding and damage-generating phenomena induced by them are both outstanding. Since practice has shown that operational numerical models often have difficulty in predicting the evolution of mesoscale convective systems, in the achievement of short term forecast of the National Meteorological Administration (NMA), the daily estimates made by the European Storm Forecast Experiment (ESTOFEX) are also taken into consideration. This study is aimed at assessing the quality of the forecasts made during summer 2010 by ESTOFEX for Romania area. The assessment was made with products available in NMA. Mainly, there was used the data provided by meteorological stations, respectively information regarding the amounts of rainfall recorded, wind gusts, hail and lightning. In this study we also followed to detail a severe weather episode with good forecast, and also a less successful example, to identify the difficult predictable structures responsible for severe weather phenomena.

Paper 249:

Results of the 2010-2011 lightning measurement campaigns in Spain

Oscar Van Der Velde, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

Joan Montanyà, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

David Romero, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

Robert Rico, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

Ferran Fabró, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

Victor March, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

Glòria Solà, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia

Nicolau Pineda, Meteorological Service of Catalonia


One of the recent discoveries is the detection of X and gamma rays produced by lightning flashes, sometimes detected from space as terrestrial gamma-ray flashes (TGFs). In northeastern Spain, several facilities and approaches are now used to study the production of these emissions. One of them is the mountain top instrumented tower near La Molina in the Pyrenees. It is capable of measuring lightning currents and in 2011 will be equipped with scintillation detectors for energetic radiation detection, an electric field change sensor, and a high-speed video camera with fisheye lens underneath. The other approach involves driving into a storm with similar equipment in a specially prepared van. During the 2010 season, this approach resulted in the capture of a spectacular positive cloud-to-ground flash with lots of recoil leader activity overhead, though it struck still too far away (a few kilometers) to be able to detect emissions. An analysis of optical and electrical data will be presented. New in 2011 is the installation of a 3D lightning mapping array (LMA) in the Ebro delta and surrounding area, which maps the propagation of lightning leaders in three dimensions. The first results will be presented at the ECSS.

Paper 251:

Hailstorm identification with crop damage and weather radar in a region in the south of Santa Fe and northern Buenos Aires, Argentina.

Romina Nahir Mezher, National Institute of Agricultural Technology

Santiago Banchero, National Institute of Agricultural Technology

Pablo Mercuri, National Institute of Agricultural Technology


Hail is the weather phenomenon that causes the major economic losses in crops. In Argentina, over 85% of agricultural insurance coverages are related to hail damage. This study explored the relationship between damage observed in different field crops (soybeans, wheat and maize) and the variables obtained from a C-band radar located in the northern region of Buenos Aires (INTA Pergamino) and a C-band polarimetric radar in the province of Entre Rios (Parana INTA) during the agricultural seasons, from August to April, 2008-2009 and 2009-2010. Moreover, with the dual polarization system that INTA-Parana radar has, the polarimetric variables Zh, Zdr, Kdp and Phidp were analyzed allowing a higher quality evaluation of hail damage.

Paper 252:

Predictability and predictive ability of severe rainfall processes

Antonio Parodi, Fondazione CIMA


A better understanding of predictability and predictive ability of deep moist convective processes is a key requirement for the prevention and mitigation of risk associated with the occurrence of extreme hydrometeorological events. New generation ground based and spaceborne remote sensing techniques, as well as current petascale and forthcoming exascale computing approaches, are allowing to study these phenomena at unprecedented spatial and temporal scales. In this paper some recent research results achieved by the CIMA Research Foundation in this framework will be presented. The discussion will range from the formulation of a novel theory for deep moist convection, through the investigation of the joint effect of turbulence and microphysics parameterizations on the uncertainty in numerical modeling of deep moist convective processes, to the hydrometeorological and dynamical classification of a rich dataset of observed extreme rainfall scenarios.

Paper 253:

Wall clouds, scud and turbulence

Gregory Tripoli, University of Wisconsin - Madison

Amanda Schulz, University of Wisconsin - Madison


A new theory explaining the existence of a supercell wall cloud is presented. The theory states that the wall cloud is formed by turbulence induced evaporation and melting of precipitation in the mesocyclone. The evaporation occurs at mean humidity’s of near 100%, but with a high variance resulting from turbulence. Pockets of unsaturated air allow evaporation that drives down moist theta-e mixtures that have lower cloud base than the adiabatically rising parcels. This departs from conventional theories that depend on the recirculation of rain-cooled subcloud base air into the updraft to drive the formation of the wall cloud. This process is also hypothesized to be a primary mechanism responsible for the formation of scud, often seen in the vicinity of severe storms. Moreover, it is hypothesized that that process will allow for the formation of a tornado funnel at somewhat higher pressures than could be expected from adiabatic ascension of the updraft. Numerical simulations of this process employing explicitly simulated turbulence intensity and a partly cloud scheme for both precipitation and cloud droplets will be presented demonstrating this process.

Paper 254:

Overshooting Tops derived from Meteosat/SEVIRI data – climatological approach.

Monika Pajek, Institute of Meteorology and Water Management


Overshooting Tops derived from Meteosat/SEVIRI data – climatological approach. Overshooting tops are the cloud structures extending above the Cumulonimbus anvil, which locate strong updraft in convective cell connected with severe weather events, such as: storm, hail, heavy rain, lightning,tornado. To indicate this phenomena difference of brightness temperature (BTD) between two spectral channels of Meteosat satellite - WV (water vapour) 6,2 µm and IR 10,8 µm - could be used. Positive difference is interpreted as an overshooting tops occurrence. Physical interpretation is still examined. However, in empirical way , previous studies of deep convective cases shown usefulness of this satellite product to detection and monitoring of strong convection. Wider and wider applicability of overshooting tops satellite product create demand for detailed studies concerning temporal and spatial variability of this phenomena. For this reason an attempt to prepare comprehensive statistical studies for the whole year in mid-latitude European conditions was made. Every fifteen minutes data from Meteosat/ Seviri channels WV 6,2 and IR 10,8 um were prepared, checked and processed. Distribution of the overshooting tops in different time, spatial and meteorological conditions together with relation between BTD of WV 6,2 and IR 10,8 um and brightness temperature in component channels were investigate. The analysis performed had methodological, operational and cognitive aspect. Detailed statistical information is also necessary for implementing product to automatic expert system.

Paper 255:

Extended analysis of severe convection event with use of satellite products.

Monika Pajek, Institute of Meteorology and Water Management


Satellite data from Meteosat due to very good temporal and spatial resolution become essential source of information for detection and monitoring of rapid convection and storm. In this study wide range of satellite products were used to investigate summer case of deep convection in Poland. These products – from single SEVIRI channel ones, thru RGB compositions to advanced NWCSaf Rapid Developing Thunderstorm - were tested from two points of view. First their usefulness in operational forecasting process were checked. Second their applicability to simple automatic, objective system were analyzed. Additionally products were compared and verified with data from independent sources, such as observations from ground station network and measurements from lightning detection system.

Paper 256:

Complex quality control of Barcelona radiosounding database

Esteban Abellán, Dept. of Astronomy and Meteorology.University of Barcelona

Montserrat Aran, Servei Meteoròlogic de Catalunya

Bernat Codina, Dept. of Astronomy and Meteorology.University of Barcelona

Jordi Cunillera, Servei Meteoròlogic de Catalunya


Since 1998 the Catalan Meteorological Service and the University of Barcelona perform a radiosounding at 00 and 12 UTC. As the quality of radiosonde data is critical in many applications, we have devoted this study to develop appropriate quality-checking methods. Different filters have been developed to detect errors such as climatological outliers, vertical and temporal meteorological inconsistencies and error calibration. The control has been applied to the two files generated after every launch: the ASCII file with a temporal resolution of 10 or 2 seconds and the TEMP report. Most of the filters have been devised following the work by the Integrated Global Radiosonde Archive project (Durre et al. 2008). For the high temporal resolution ASCII files the filters used are: detection of the climatological outliers, detection of the duplicity in pressure or geopotential height, verification of the hydrostatic equation and detection of vertical gradient inconsistencies. With regard to the TEMP reports the filters applied to the mandatory levels are: extreme value checking and detection of temporal inconsistencies. Finally, inconsistent or erroneous data have been manually checked and corrected whenever possible. For that, a comparative analysis of Barcelona radiosonde profiles has been done with nearby radiosonde stations and model reanalysis. The inconsistencies at the initial level have been checked with the surface weather station. After this filtering process we have found that about 4% of our database contains erroneous data, most of them during its earlier years. After an update of the Vaisala equipment, the number of errors became practically negligible.

Paper 257:

The ARPA-EMILIA-ROMAGNA procedures for forecasting and monitoring severe events: two years of experiments

Anna Fornasiero, ARPA-SIMC

Miria Celano, ARPA-SIMC

Pier Paolo Alberoni, ARPA-SIMC

Antonella Morgillo, ARPA-SIMC

Paolo Mezzasalma, ARPA-SIMC

Rosanna Foraci, ARPA-SIMC

Sandro Nanni, ARPA-SIMC


In this work two years of special project focused on the forecasting and managing of severe storms associated to effects on local territory (e.g. roads, small catchments and urban areas hydrological problems) are presented. The goal is to provide an operational procedure for the warning emissions during summer season. In the first year of experiment (2010) there was not occurrence of this type of events in the testbed area (Province of Modena), therefore only a post-event analysis was carried out. The principal result was the separation of two main classes of situations: events were the synoptic forcing is well defined and thus numerical models are able to forecast correctly and events where models fail due to weak synoptic forcing or very localised intense precipitation for which monitoring tools play a crucial role. For the 2011 summer the experiment is extended to the whole Emilia Romagna Region to increase the probability to catch severe events; moreover the forecasting, nowcasting and monitoring activities are defined in a more detailed way. At first, in order to forecast the occurrence of heavy thunderstorms, a chain of high resolution limited area models, spanning from a 7 km to 1 km mesh grid, is used. A set of instability indices is also computed from the observed and forecasted soundings to assess the severity of convection. The result of these actions is summarized in a table evaluating the probability of thunderstorms occurrence and the associated risk, for the current and following day. Then instruments for nowcasting and monitoring, as radar, satellite, high frequency ground stations and lightning detector, are used. On this purpose, some additional operational radar products have been developed to support these activities, as accumulated precipitation in a floating interval and a tool providing the reflectivity above a critical threshold on a high resolution visualization support.

Paper 258:

The representation of convective wind gusts in coarse numerical models

Pieter Groenemeijer, European Severe Storms Laboratory

Heikki Tuomenvirta, Finnish Meteorological Institute

Pauli Jokinen, Finnish Meteorological Institute

Andrea Vajda, Finnish Meteorological Institute


The spatiotemporal resolution of contemporary global climate and weather forecasting models does not permit free convection to be simulated explicitly. Instead, parameterizations must be used to account for convective processes. These parameterizations return the average effect of convection taking place across the area represented by a model gridpoint within the timestep of two calls of the parametrization scheme. As a consequence, these models do not directly provide information of extremes occurring on smaller scales. We discuss the representation of strong wind gusts that occur in concert with deep, moist convection in a frequently-used reanalysis data set, the ECMWF ERA-Interim. We will report on our study on how to use climate models with a similar resolution to predict the future climatology of convective wind gusts.

Paper 259:

Climatological aspects of convective severe weather events in Europe

Pieter Groenemeijer, European Severe Storms Laboratory


Since the establishment of the European Severe Weather Database one single source for extreme weather events has become available to the European research community. Now that the data set is steadily growing at a rate of over 4000 events per year, consists of more than 35000 events, some preliminary climatological analysis is due. The distribution of the frequency of events over the year and the day will be presented. Moreover, the variation of the peak season of severe weather across Europe will be highlighted. Additionally, the spatial and temporal association of severe weather event types with each other is shown. Finally, estimates of high-end convective event occurrence are given.

Paper 260:

Towards an improved wind speed scale vs. damage description adapted for Central Europe

Bernold Feuerstein, European Severe Storms Laboratory e. V.

Pieter Groenemeijer, European Severe Storms Laboratory e. V.

Erik Dirksen, European Severe Storms Laboratory e. V.

Martin Hubrig, European Severe Storms Laboratory e. V.

Alois M. Holzer, European Severe Storms Laboratory e. V.

Thilo Kühne, European Severe Storms Laboratory e. V.

Nikolai Dotzek, European Severe Storms Laboratory e. V.


Determination of tornado or downburst intensity is a difficult task, as it is mostly done using the damage occurring with a storm. Only in very rare cases are there reliable measurements of wind speed. A grading of intensity is done either using the Fujita-scale (F- or EF-scale) or the TORRO-scale, or using both classifications. In order to determine which typical property-, building-, and vegetation damage occurs with the different classes of the F- or T-scale, one needs a verbal description of that damage. A simple adoption of the description valid in the USA (EF-scale) is not very helpful in Europe since the construction of homes differs significantly. A verbal description of the TORRO- and Fujita-scales adapted for Central Europe considering building structure and vegetation characteristics was developed by ESSL, Skywarn Germany and Munich Re members in 2004, but so far was only available in German. In a joint project, ESSL and Skywarn Germany supplemented this description by photos of typical damage. Based in this, we present an updated wind speed scale description adapted for Central Europe considering wind impact to buildings as well as to vegetation. The scale is motivated by the need of a broadly applicable, accurate and consistent tornado or downburst intensity rating system based on a standardised wind speed scale for the purpose of climatological homogeneity. The description comprises building and vegetation damage characteristics, which can be found in Central Europe - but similar in other parts of the world, occurring with the various classes of the Fujita- and T-scales. For practical application, an ensemble-based use of a decision matrix for specific building structures and vegetation types is suggested.

Paper 261:

The Pentecost Monday tornadic supercell in eastern Germany on 24 May 2010

Bernold Feuerstein, European Severe Storms Laboratory e. V.

Thilo Kühne, European Severe Storms Laboratory e. V.


We will give an overview on a tornadic high-precipitation (HP) supercell on 24 May 2010 (Pentecost Monday) which affected the federal states Saxony-Anhalt, Brandenburg and Saxony in eastern Germany and eventually stretching into the northern parts of Czech Republic while developing into a bow echo. It produced at least two strong tornadoes (F2 and F3), a series of heavy downbursts, flash floods and large hail up to 9 cm diameter. With an overall damage swath of more than 200 km length, one fatality, dozens of injuries and a total damage to property and vegetation of more 100 Mio € this supercell was one of the most significant severe events in Europe in 2010. More than 40 entries in the European Severe Weather Database are associated to this storm. The radar images of the supercell showed a persistent pronounced hook echo and typical HP features. The impressive rotating rain and hail curtain was documented by numerous photos and videos but so far no image of the tornadoes emerged since they were likely wrapped by the precipitation. The most affected region was the city of Großenhain (Saxonia) about 30 km NW of Dresden where an F3 tornado caused severe damage to buildings including the collapse of a 90 m high brick chimney. The strong downbursts occurred mainly in the hook and rear flank downdraft region of the supercell close to the presumed tornado tracks resulting in complex damage patterns which are difficult to analyse. Based on local surveys and a comprehensive series of aerial photos we will report on the present status of the damage assessment.

Paper 262:

New Event Types for the European Severe Weather Database (ESWD)

Alois M. Holzer, European Severe Storms Laboratory e. V. (ESSL)

Pieter Groenemeijer, ESSL

Nikolai Dotzek, ESSL

Bernold Feuerstein, ESSL

Thomas Krennert, ZAMG

Thilo Kühne, ESSL

Zhongjian Liang, ESSL

Susanne Haeseler, ESSL


The European Severe Weather Database (ESWD) is a database of severe weather reports for Europe and the Mediterranean Region containing over 35 000 individual reports. The phenomena it contains are primarily associated with severe convective weather: tornadoes, large hail, gustnadoes, heavy precipitation, and lesser whirlwinds (gustnadoes and dust devils). Since its start, inclusion of new types of phenomena has been a topic of discussion between European Severe Storms Laboratory (ESSL), its partners (like NMHSs or spotter organizations) and other ESWD users. Yet, because of a general lack of data on severe convective weather occurrence across Europe, more priority was given to increasing the coverage than to introducing new event types. Within the European Union FP7 Project EWENT (Extreme Weather impacts on European Networks of Transport) an inclusion of new event types has been implemented. The poster presents these new event types for ESWD reports that extend the usability of the ESWD to non-convective severe weather events. In late summer 2011, the new event type “snow” was introduced, covering severe snow or snowstorm events. This allows e. g. the recording of snow accumulations or unusual blowing snow conditions, both for convective and non-convective types. At the same time, two other winter time events were introduced to the ESWD: Icing events (like freezing rain or drizzle, or like hazardous rime accumulations) and avalanches (with definitions based on the input coming from alpine avalanche warning services). Last, one additional event type for convective severe weather was introduced: lightning. This type does not include all strokes, but merely those having caused significant damage or injuries to aircraft, vehicles, ships, people, animals, etc. or strokes having an exceptional peak current, according to lightning detection networks.

Paper 263:

Initial results from convective-scale analysis and prediction of the 14 June 2011 Norman Oklahoma macroburst using conventional and rapid-scan weather Doppler radar data

Louis Wicker, NOAA National Severe Storms Lab

Edward Mansell, NOAA National Severe Storms Lab

Daniel Dawson, Univ of OK Cooperative Institute of Mesoscale Meteorological Studies

David Dowell, NOAA Environmental Systems Research Lab


During the early evening of June 14 2011 Norman Oklahoma (USA) was struck by a wet downburst wind event which produced damage across a ~7 km wide and ~20 km long swath in the northern and eastern portions of the city. Thousands of homes were damaged with an estimated $10 million in insurance claims incurred. The storm formed in an environment characterized by deeply mixed adiabatic boundary layer, weak vertical wind shear, and the storm's interaction with slowly moving weak cold front may have help enhance its intensity. The storm developed approximately 30 km northwest of Norman at 2315 UTC, and was producing damage in Norman by 0015 UTC. A new prototype next-generation operational radar system, NSSL's 10 cm Multifunction Phased Array Radar (MPAR), was located some 25 km southeast of the initial storm cell and eventually was in the storm's damage path. MPAR electronically scans in the both horizontally and vertically within a 90 degree sector with no moving parts, i.e., it can scan very rapidly. For this case, MPAR was able to collect 19 elevation scans in a 90 degree sector every minute during a one hour period before the severe storm went over the radar site. This event presents a unique opportunity to use the MPAR data to determine whether the 5x increase in temporal scanning frequency (relative to convention US operational radars) can help improve analyses within a high-resolution convective storm models. Using the ensemble Kalman Filter (enKF) we will assimilate the various data sets to compare the forecasts of the severe weather parameters using the Twin Lakes WSR-88D data set (located approximately 15 km northeast of Norman) versus the MPAR data set.

Paper 264:

Hailstorms Analysis using Polarimetric Weather Radars and Microwave Sensors in Argentina

Romina Mezher, Climate and Water Institute (INTA) and Department of Atmospheric Sciences (FCEyN/UBA)

Luciano Vidal, National weather Service (SMN) - CONICET and Department of Atmospheric Sciences (FCEyN/UBA)

Paola Salio, Department of Atmospheric Sciences (FCEyN/UBA) and Research Center of the Sea and the Atmosphere (CONICET/UBA)


In the year 2009, during November 29th strong thunderstorms affected East-Central Argentina producing important damages in crops because of big size hail and intense wind gusts. Double polarization weather radars are the best tool to determine hydrometeor types inside the storms and to identify the presence of hail. They are used to broadcast early alerts to the population and in its operational mode to explore the great potential of convection. This study describes the environment related to both storms from the point of view of the mesoscale. It investigates the characteristics of storms producing hail at the surface through an inspection from multiple platforms: observations from TRMM (microwave and radar) and polarimetric variables from dual polarization radar located in the city of Parana (Argentina). The analysis of this information during these extreme events allowed a better understanding of the dynamics of these storms that develop in the southeast of South America and to investigate environmental factors that help their forecast.

Performance Evaluations of the European Lightning Location System EUCLID

Wolfgang Schulz, Austrian Lightning Detection and Information System


In 2001 several countries (Austria, France, Germany, Italy, Norway and Slovenia) started a cooperation called EUCLID (EUropean Cooperation for LIghtning Detection). It is the goal of this cooperation to provide European wide lightning data with nearly homogeneous quality. Subsequently also Spain, Portugal, Finland and Sweden joined EUCLID. The EUCLID cooperation is special in the sense that the individual partners are highly motivated to run their individual networks with state-of-the-art lightning sensors. All the partners employ dedicated technicians to supervise and maintain the network and to react fast in case of sensor or communication problems. As of August 2011 the EUCLID network employs 142 sensors, 4 LPATS III, 13 LPATS IV, 1 SAFIR, 16 IMPACT, 42 IMPACT ES/ESP and 66 LS7000 sensors (oldest to newest), all operating over the same frequency range with individually-calibrated gains and sensitivities. Data from all of these sensors are processed in real-time using a single common central processor, which also produces daily performance analyses for each of the sensors. This assures that the resulting data are as consistent as possible throughout Europe. In fact, the Europe-wide data produced by EUCLID is frequently of higher quality than the data produced by individual country networks, due to the implicit redundancy produced by shared sensor information. Further, this cooperation also acts as a platform for exchange of knowledge related to lightning location technology. Since the beginning of the cooperation the performance of the EUCLID network has been steadily improved, e.g. with improved location algorithms, with newer sensor technology and by adapting sensor positions because of bad sites. Over the next 1-2 years, at least 15 of the older sensors are expected to be upgraded to the newest (LS700x) instruments.
To compare the estimated performance to real values several measurement campaigns were carried out in Europe, e.g. in Slovenia where LLS data was compared to data from GPS synchronized flash counters installed on mobile phone towers [1] – [2], in France where video surveys were used to determine the actual network performance of the French lightning location system [3]. In the presentation detailed performance results from a comparison with direct lightning current measurements on a tower and from video and E-field measurements in Austria will be provided.


[1] Djurica V., Kosmač J.: LLS Accuracy improvements by measurements collected by the RLDN, 19th International Lightning Detection Conference, Tucson, USA, 2006.

[2] Djurica V., G. Milev , J. Kosmač: Lightning location networks performance validation with RLDN, 16th International Symposium on High Voltage Engineering, South Africa, 2009.

[3] Berger, G. and S. Pedeboy: Comparison between real CG flashes and CG flashes detected by a lightning detection network, ICOLSE, Blackpool, 2003.