7th European Conference on Severe Storms
ECSS 2013


Paper 2:


Jianming Yin, Tokio Marine Technologies LLC, United States

Bo Yu, Tokio Marine Technologies LLC, United States

Junwa Shimada, Tokio Marine & Nichido Risk Consulting Co., Ltd, Japan


Spain is the European country with the greatest agricultural losses caused by severe hailstorms. While there are a few studies on regional severe hail climatology in Spain based on regional hailpad measuring networks and a number of studies on the meteorological conditions and climatological interpretation of individual severe hailstorms in some regions of Spain, there has been a lack of severe hail reports for studies and investigations on the severe hail climatology in Spain at the national level. Severe hailstorm risk is a significant issue for the insurance industry. Therefore, Re/insurance companies need reliable models to analyze and quantify severe hailstorm risk in Spain.
After reviewing the quality of available Spain hail data required for simulation of severe hailstorm risks, historical hailstorm reports from 2005 through 2010 from the European Severe Storm Laboratory (ESSL) European Severe Weather Database (ESWD) are used to model the severe hailstorm risk in Spain. Monte Carlo technique is conducted to simulate the stochastic severe hailstorms using the compiled ESWD historical hail data.
To allow companies to assess their reinsurance needs at a portfolio level, the concept of severe hail event is introduced and defined as a congregation of individual hailstorms occurring within a 72-hour timeframe across Spain. The simulation process starts with simulating the probable number of severe hail events in a simulation year by sampling from the Poisson distribution fitted by the historical annual number of hailstorm events. The number of hailstorms in a hail event is modeled by an exponential distribution. The starting date of a stochastic hail event is modeled by a Weibull distribution derived based on the temporal variations of ESWD historical hail events in Spain. A two-dimensional (latitude and longitude) Gaussian kernel density smoother with the bandwidth is employed to achieve the maximum likelihood of the ESWD historical hail storm genesis locations in Spain. The optimal bandwidth for Gaussian kernel density estimation is determined using Jackknife out-of-sample likelihood maximization procedure. Physical parameters of a stochastic hailstorm, including storm path length, width, orientation, and intensity, are then simulated using different statistical distributions fitted by historical hailstorm data. The same simulation process repeats for all stochastic hailstorms to simulate up to 1,000,000-year worth of future possible severe hailstorm events in Spain.
Comparisons show that the simulated climatology of severe hailstorm risk matches reasonably well with the ESWD historical data. The severe hailstorm model developed in this study provides a reasonable quantification of the severe hailstorm risk in Spain from Re/insurance perspectives. Coupled with the severe hailstorm damage modeling and insurance portfolio exposures, the model can help re/insurance companies to quantify the severe hailstorm loss potentials in Spain at both individual locations and at portfolio levels.

Paper 3:

Integrated In Situ, DOW, and damage observations in tornadoes

Karen Kosiba, Center for Severe Weather Research, United States

Josh Wurman, Center for Severe Weather Research, United States


Analyses of the low-level winds in tornadoes will be presented from two unique data sets, 05 June 2009 (Goshen County, WY) and 25 May 2012 (Russell, KS), which reveal periodic temporal variations in velocity, little variation in wind speed with height, and one- and two-celled low-level vortex structures. Also EF ratings, damage incurred and wind speed measurements suggest inconsistencies between how metrics are applied and actual wind measurements.

During the Goshen County, WY tornado, direct observations of the winds inside the tornado were obtained with an instrumented armored vehicle, the Tornado Intercept Vehicle (TIV) and integrated with Ground Based Velocity Track display (GBVTD) analyses constructed from fine-scale mobile Doppler radar (Doppler On Wheels) data to reveal inward and upward spiraling near-surface flow, upward motion near the surface, and an axial downdraft aloft. FFT analysis of the Rapid-Scan DOW observations every 7 s revealed a periodicity in tornado intensity, with amplitude of ~4 m s-1 and peak energy at periods of 66 s and 108 s. This is consistent with long wavelength, upstream-propagating Rossby-type waves slowly revolving about the tornado at 6-10 m s-1, not short wavelength multiple-vortex type phenomena which would complete orbits in ~20 s. Additionally, simultaneous video documentation of damage occurring during the tornado is related to the direct wind observations, which suggest that wind direction and complex failure modes are important in assigning a damage rating.

During the Russell, KS tornado, data taken by the Rapid Scan DOW, DOW7 and an in situ mobile mesonet provide an instantaneous vertical cross-section (7 levels) of winds from 3.5 m AGL to 30 m AGL. Preliminary analysis indicates little dependency of the wind speeds on height. Additionally, two-dimensional dual-Doppler analyses from DOW7 and the Rapid-Scan DOW with an unprecedented 1.6 km baseline and 7-s GBVTD analyses will be presented illustrating the rapid evolution of the low-level wind structure.

Paper 4:

Synoptic conditions of the occurrence of hail in central Europe

Katarzyna Suwala, Adam Mickiewicz University, Poland


Hail is one of the most complex products of water vapour condensation. It occurs most often during thunderstorms (called hailstorms), usually when the air temperature above ground surface is higher than 0˚C. Not only is it very local, ephemeral and rapid phenomenon, but also it causes a great threat to human life and property. For this reason many studies were carried out, attempting to indicate as many factors contributing to hail occurrence as possible, in order to improve the process of hail forecasting. The main objective of this research is to identify hail regions and their synoptic characteristics which are conductive to hail development in central Europe.
Taking into consideration the spatial variability in the occurrence of hail in central Europe, five hail regions were distinguished using Ward’s group hierarchy method. Furthermore, circulation patterns responsible for hail precipitation in each region were analysed. Composite maps of the sea-level pressure (SLP), 500 hPa geopotential height and 850 hPa-level temperature means were constructed for the days with hail. Additionally, anomalies of the values of each parameter were presented, in order to identify the differences between the days with hail and average conditions.
When it comes to the general characteristic of hail incidents in this part of Europe, it was found that more than 65% of hail incidents were recorded from April to June, with May as the hail-peak month. Furthermore, hail occurred most often in the southern part of Germany and in north-eastern sector of Poland.
In all five regions hail precipitation was associated with negative anomalies of SLP and 500 hPa heights, most often along with low-pressure systems developing within cold air mass. In some cases the spatial distribution of temperature anomalies in lower troposphere implied the presence of cold front passing over Europe.

Paper 5:

Wintertime lightning in Finland in 2003-2007

Joanna Rinne, Foreca, Finland


The purpose of my research was to gather information about the properties and climatology of wintertime lightning events in Finland. In this research a wintertime lightning event was defined as 'one or more detected lightning occurences in a weather situation, where either a temperature of ≤ 0 °C and/or snowing or sleeting was observed in the SYNOP-opservations closest to the lightning location.'
In 2003-2007 there were altoghether 19 events meeting these requirements. The events were spread out quite evenly between the end of october - the beginning of november and april, with small maxima in november and february. Time of day did not seem to affect the probability of wintertime lightning occurrence. The events were mainly located in the southwestern part of Finland, but there were observations of wintertime lightning even in northwestern Lapland. In 16 of the 19 observed lightning events the wintertime lightning was observed in areas with an occluded front (frontal analyzes by FMI and/or UKMO). In one of the cases the front was analyzed as cold and in two the analysis was unclear. The center of a low pressure area was usually situated on the west side of the observed lightning; this was caused by the clockwise turning of occluded fronts around the low pressure centre. The surface temperature was close to zero in all wintertime lightning events and both rain and snow/sleet was observed in the SYNOP observations close to the lightning location. CAPE was zero or very close to it in the soundings closest to the time and place of observed lightning. Warm advection was observed in almost all soundings, also an unfrozen sea was found to be a factor in the formation of wintertime lightning in Finland. In each event, the number of cloud-to-ground-lightning observed was usually 1 or 2 and the amount of strikes in a lightning 1 or 2. 46% of the lightning observed were positively charged, 54% negatively charged. Case studies will be presented if time permits.

Paper 6:

Simulation of a Severe Thunderstorm Event over Bangladesh Using WRF-ARW Model, During SAARC STORM Pilot Field Experiment 2011

Md. Nazmul Ahasan, SAARC Meteorological Research Centre (SMRC), Bangladesh

S. K Debsarma, Bangladesh Meteorolocal Department (BMD), Bangladesh


Thunderstorms are the local severe storms affect the West Bengal and Northeastern parts of India, Bangladesh, Bhutan and Nepal during the pre-monsoon season (MAM) caused huge loss of lives and damage to properties worth millions of dollars annually. ARW dynamic core of WRF model is utilized to simulate the thunderstorm event of 11 May 2011, which occurred over Bangladesh and nearby territory of India. Required data were collected under the SAARC STORM Pilot Field Experiment 2011. NCEP FNL data at six hourly intervals is utilized as initial and lateral boundary condition (LBCs).The thunderstorm event is studied based on field survey, ground and RADAR observations. Doppler Weather RADAR (DWR), Kolkata has recorded the vertical extend of the system was about 16-18 km and the RADAR reflectivity 54-57 dBz. A north-south oriented well established squall line of 400 km length was noticed to the centre of Bangladesh at 0530 UTC. Convection persisted over Bangladesh from 0300 UTC to 1300 UTC. Kalpona-1 Satellite recorded the cloud top temperature (CTT) of -50°C at 0600 UTC. The WRF model products at 9km and 3km resolution of various initial conditions (00, 06, 12 hrs earlier run) showed a cyclonic circulation over Gangetic West Bengal of India at 850 hPa and over Assam and adjoining areas of India at 500 mb. Strong upward motion was noticed over Bangladesh. Significant moisture incursion was observed over centre and south-east parts of Bangladesh at 925 and 850 hPa with southerly/southwesterly flow of 9 to 11 ms-1 (17-21 knots). Result showed that the WRF-ARW model with high resolution could capture thunderstorm event of 11 May 2001 in reasonably well though there are some spatial and temporal biases. The model with 06 hrs earlier initial conditions was appropriate to predict the system.

Paper 7:

Near Term Convective Precipitation Forecasting using Real-Time Lightning Data, Observations, and Radar

Daniel Lennartson, Telvent, United States


Near Term Convective Precipitation Forecasting using Real-Time Lightning Data

Areas of convective precipitation are often associated with a number of significant hazards such as tornadoes, hail, lightning, wind gusts, and flash floods. A scalable technique is featured to enhance near term precipitation forecasts using real-time lightning data which can assist in timeliness and coverage of convective events and provide reasonable quantitative precipitation forecasts, employing a relationship between lightning flash density and radar with the numerical model forecast parameters.
Often we see numerical models have common shortfalls when it comes to convective forecasting such as under forecasting amounts, miscalculating the speed of convective lines, and misplacing the general areas of convective precipitation. Primarily, this is due to a lack of resolution and/or limited initial conditions. When operational forecasters utilize an ensemble of models to manually to make up for these deficiencies, the task can be daunting and there always is a risk an event is missed or not caught until the event has already made a significant impact. The other effect is over forecasting to ensure that missed events are avoided; however, over forecasting can also be detrimental to end users of the forecast data with false alerts to events that sometimes clearly will not happen.
The technique displayed within this paper will demonstrate how lightning flash density, observations, and radar alters a near term quantitative precipitation forecast initially populated with a blend of numerical models. The result is a far superior first guess field that will include all high impact convective events and if needed, amplify the precipitation amounts that are underforecasted while reducing areas of precipitation not covered by radar and/or have low forecast probability of precipitation (POP).

Paper 8:


Skyler Jayden Dembe, Global Initiative Uganda (GIU), Uganda

Cissy Namujju, Millennium Environmental Health Alliance (MEHA), Uganda

Emmanuel Mbabazi, Global Initiative Uganda (GIU), Uganda


It is suffice to constantly assess community vulnerability and capacities with regard to climatic change and build their resilience through adaptation efforts, complementing mitigation efforts aimed at reducing the rate and magnitude of climate change. This framework has shifted from Disaster management to a sustainable approach of Disaster risk Reduction. Disasters are associated with extreme weather events. Climate change directly interacts with the exposure to climatic extremes. The challenge in the context of adaptation is to move from the understanding that climate change is occurring to concrete measures that reduce existing vulnerabilities of human and ecological systems. The focus in this study is the effects and responses of flood risk imposed by storm water among the urban poor living in the highly vulnerable shanty neighborhoods on the outskirts of Kampala city centre. It explores the underlying vulnerabilities of the two areas and the challenging problem of how to effectively shape human institutional responses to the risk of natural disasters with a special focus on floods. The social risk management and asset-based approaches on which the study is based provide a conceptual framework for understanding the sequential links between risks; human exposure and sensitivity; the impacts of risky events; and risk management strategies. The outcome of the study shows marked differences in the vulnerability factors and the management of flood related disasters in the two study areas. Furthermore, it was revealed that the indigenous coping mechanisms employed by the poor may become less effective as increasingly flimsy livelihood systems struggle to withstand disaster shocks. Strategies to reduce vulnerability should be entrenched in vulnerability analysis and greater understanding of both household-level and universal-response options that are available to decrease the vulnerable exposure to climate risk.

Paper 10:

Comparison of Simulations using WRF_ARW and WRF_NMM Model for Local Severe Storms over Bangladesh

Md. Majajul Alam Sarker, SAARC Meteorological Research Centre (SMRC), Bangladesh

Sujit Kumar Debsarma, SAARC Meteorological Research Centre (SMRC), Bangladesh

Nazlee Ferdousi, SAARC Meteorological Research Centre (SMRC), Bangladesh

Mohan Kumar Das, SAARC Meteorological Research Centre (SMRC), Bangladesh


Simulation of severe thunderstorm events which occurred over Bangladesh and adjacent area are performed using ARW and NMM dynamic cores of fine grid version of the Weather Research and Forecasting (WRF) modeling system. The thunderstorms are typical mesoscale systems dominated by intense convection, is one of the most incredible weather phenomena in the atmosphere. They frequently occur over Bangladesh and adjoining regions during hot and humid weather conditions of pre-monsoon season (March to May). These storms are also known as Nor’wester (locally called “Kalbaishakhi”), as they usually travel from northwesterly direction. Mesoscale models are essential for the accurate prediction of such high-impact weather events. In the current research, an attempt has been made to compare the simulated results of two thunderstorm events on 15th April and 3rd May 2012 using ARW and NMM model core of WRFV3.3 with 4 km horizontal resolution and validated the model results with observations. Both models performed well in capturing stability indices which are indicators of severe convective activity. Comparison of model simulated radar reflectivity imageries with observations revealed that NMM model has simulated well the propagation of the squall line than ARW model. It is also observed that model simulated spatial plots of cloud top temperature, NMM model captured the genesis, intensification, and propagation of thunder squall little better than ARW model. The outcomes of these studies show that the high resolution WRF_NMM model has better capability for simulating the thunderstorms in comparison with WRF_ARW model.

Paper 11:

Disaster risk reduction efforts and factors affecting flood disaster management: A case study of Katakwi district-Olupe and Ngariam Camps.

Maxi Mbidde Ssenyondo, Green World Uganda (GWU), Uganda

William Kayemba, Makerere University, Kampala-Uganda, Uganda

Betty Nabatanzi, Crusade for Environmental Awareness Agency (CEAA., Uganda


Objectives: to establish factors that affected flood disaster management in Katakwi district.

Methodology: the study focused on 238 randomly selected internally displaced camp residents. Key informant interviews were conducted and data qualitatively analyzed to assess issues that impeded successful mitigation of flood disasters.

Results: Continued degradation of wetlands for crop cultivation and cattle over grazing were partly responsible for increased flooding during rainy seasons and drought during dry seasons. The absence of accessible micro-finance credit schemes to support recovery efforts of the communities’ drastically undermined measures to reduce the impact of flood disasters. The district was reported to have lacked contingency plans to show the risks and likelihood of related disasters occurring with potential effects at the community level hence impeding disaster management and preparedness. Both the government of Uganda and the local government of Katakwi district had not practically earmarked emergency funds for disaster response. The affected communities had no storage facilities for emergency relief items like medicine and food. The idea of having community level food stores and granaries died out and this amplified the flood disaster with famine making disaster management difficult. The poor nature of the community’s temporary mud bricks and wattle roofed huts exacerbated the impact of the floods since many huts were just washed down prompting more relief items like tents straining the relief efforts.

Conclusion: Gross lack of awareness at the community level and alternative means of livelihood that do not constrain non-renewable resources have persistently provoked natural disasters in the district.

Paper 12:

Exploring the Constraints and Strategies for Climate Change Adaptation

Simon Peter Ssebudde, Rural-Urban Environmental Agency (RUEA), Uganda

Faridah Nakanwagi, Rural-Urban Environmental Agency (RUEA), Uganda

Elisha Nambi, Makerere University, Kampala-Uganda., Uganda


This paper discusses the possibilities and constraints for adaptation to climate change in urban areas in low- and middle-income nations. The paper outlines the potentials for adaptation and the constraints, the scale of urban change, direct and indirect impacts of climate change on urban areas. This highlights how prosperous, well-governed cities can generally adapt, at least for the next few decades – assuming global efforts at mitigation successfully halt and then reverse global emissions of greenhouse gases. But most of the world’s urban population lives in cities or smaller urban centers ill-equipped for adaptation – with weak and ineffective local governments and with very inadequate provision for the infrastructure and services needed to reduce climate-change-related risks and vulnerabilities.

Innovations by urban governments and community organizations and in financial systems that address such problems are explored, including the relevance of recent innovations in disaster-risk reduction for adaptation. Local innovation in adaptation can be encouraged and supported at national scale. Mechanisms for financing this and the larger ethical challenges that achieving adaptation raises – especially the fact that most climate-change-related urban (and rural) risks are in low-income nations with the least adaptive capacity.

Paper 13:

Analysis of lightning activity during thunderstorms with the overshooting tops

Petra Mikus, Meteorological and Hydrological Service, Croatia, Croatia

Natasa Strelec Mahovic, Meteorological and Hydrological Service, Croatia


Thunderstorms’ updraft area, frequently manifested by the appearance of the overshooting tops (OT), is linked to the electric activity of the storm, in a way that updraft surges coincide with an increase in flash rate. Analysis of lightning activity in the thunderstorms with OTs was done for the warm part of the year (May - September) during 2009 and 2010, over the region from approximately 41°N 8°E to 49.5°N 24°E. Deep convective clouds with OTs were detected from the Meteosat 9 SEVIRI data, using method based on the infrared window (IRW, 10.8 μm) channel and the absorption channels of water vapor (WV, 6.2 μm) and ozone (O3, 9.7 μm) in form of brightness temperature differences. Locations and times of the detected OTs were compared to the distribution and types of the lightning strokes, provided by LINET Lightning Location System.

The results show that spatial distribution of lightning activity coincides well with the spatial distribution of the detected OTs. The largest numbers of lightning strokes, as well as OTs were detected in the western Hungary, southeastern Germany, northern Adriatic and southern Adriatic coastal region at the slopes of the Dinaric Alps. In general, the largest number of OTs occur between 14 and 21 UTC, while from 06 to 10 UTC OT detections are rather rare. Lightning activity shows similar temporal distribution. Sharp increase in lightning activity and larger values of the electric current are evident at the time of the OT detections. At that times, lightnings occur well above the tropopause, being clearly related to the OT parts of the Cumulonimbus cloud.

Paper 14:

Synoptic and mesoscale analysis of waterspouts in the Adriatic (2001 – 2011 preliminary climatology)

Tanja Renko, DHMZ, Croatia

Josipa Kuzmic, FER, Croatia

Natasa Strelec Mahovic, DHMZ, Croatia


A comprehensive analysis of waterspout events observed and recorded in the Eastern Adriatic Basin in 2010 was carried out. The synoptic and mesoscale environment was discussed and the corresponding Szilagyi Waterspout Nomogram (SWN) and Szilagyi Waterspout Index (SWI) values were calculated with the help of ALADIN model. Since official observations are rare and occasional information from different media unreliable, an effort was made to establish continuous data collection, in order to ensure better understanding of all possible factors and environmental conditions for waterspout formation, and to provide a preliminary climatology of waterspouts in this region. For that purpose, in spring 2011 a survey: “You saw a waterspout/tornado? Report to us!” was launched on the official web site of Croatian Meteorological and Hydrological Service (meteo.hr). This survey enabled a more complete database encompassing 220 waterspout events from 2001 to 2011 which is presented and discussed in this work.

Spatial coverage of the events shows equal distribution along the eastern part of the Adriatic coast although there is an evident connection between the number of recorded events and locations with largest concentration of population, like the big towns or tourist destinations. Following the previous works that included Adriatic and/or Mediterranean waterspouts we present here an analysis of temporal distributions per year, season, month and time of the day. A separation between ‘fair weather’ and ‘thunderstorm related’ waterspouts are also made and absolute frequencies of five synoptic types relevant to waterspout development are shown. Thermodynamic environment was analyzed with the help of radiosounding data.

Among all the results several of them should be pointed out: waterspouts are more frequent during the summer months, more that half of the events are thunderstorm related. The synoptic environment with dominating south-westerly flow proved to be the most favorable for the development of waterspout phenomena.

Paper 15:

Numerical simulation of squall line by using detailed microphysics

Istvan Geresdi, University of Pecs, Hungary

Gregory Thompson, National Center for Atmospheric Research, United States

Noemi Sarkadi, University of Pecs, Hungary


There is a strong interaction between the dynamics and microphysics occurs in the squall lines. The evaporation of the water drops near to the surface affects the intensity of the cold pool and the low-level wind shear. Numerical experiments prove that appropriate simulation of the melting process is crucial for the correct simulation of cooling due to the evaporation of the water drops. Most of the numerical models suppose that the melted water immediately sheds off the surface of the melting snow flakes and graupel particles. There are numbers of evidences based on laboratory observations, which show that the shedding of melted water from the surface of solid precipitation elements occurs only in the case of graupel particles larger than about 1 cm. Our studies show that the artificial increase of the number concentration of small water drops due to the continuous shedding results in overestimation of cooling rate due to the evaporation.
Two-dimensional versions of WRF with both detailed and bulk scheme are used to investigate the sensitivity of cold pool formation on the applied microphysical description.

Paper 16:

Geostationary Lightning Observations in Support of NWC and Severe Weather Monitoring

Jochen Grandell, EUMETSAT, Germany

Marcel Dobber, EUMETSAT, Germany

Rolf Stuhlmann, EUMETSAT, Germany


The Lightning Imaging Sensor (LIS) onboard the Tropical Rainfall Measurement Mission (TRMM) platform has provided a continuous source of lightning observations in the +/- 35 deg latitude region since 1998. LIS, together with its predecessor Optical Transient Detector (OTD) have established an unprecedented database of optical observations of lightning from a low-earth orbit, allowing a more consistent and uniform view of lightning that has been available from any ground-based system so far. The main disadvantage of LIS is that, since it operates on a low-earth orbit with a low inclination, only a small part of the globe is viewed at a time and only for a duration of ~2 minutes, and for a rapidly changing phenomenon like convection and the related lightning this is far from optimal. This temporal sampling deficiency can, however, be overcome with observations from a geostationary orbit. One such mission in preparation is the Lightning Imager on-board the Meteosat Third Generation (MTG) satellite, which will provide service continuation to the Meteosat Second Generation (MSG) system from 2018 onwards. The MTG system will continue MSG as the primary European source of geostationary observations over Europe and Africa.

One of the new missions on MTG is the Lightning Imager (LI) mission, detecting continuously over almost the full disc the lightning discharges taking place in clouds or between cloud and ground with a resolution of roughly 10 km. The LI mission is intended to provide a real time lightning detection (cloud-to-cloud and cloud-to-ground) and location capability in support to NWC and VSRF of severe storm hazards and lightning strike warning. As lightning is strongly correlated with storm related phenomena like precipitation, hail and gust, a further objective of the LI mission is to serve as proxy for intensive convection related to ice flux, updraft strength and convective rainfall.

The MTG LI mission concept and status will be presented and discussed. An emphasis will be put on describing the processing concepts from raw data to end-user products, and the various related challenges in processing, such as false lightning event filtering.

Paper 17:

Severe Thunderstorm Observation and Regional Modeling Pilot Field Experiment 2012 and Data Assimilation Impacts

Mohan Kumar Das, SAARC Meteorological Research Centre (SMRC), Sher-E-Bangla Nagar, Dhaka-1207, Bangladesh

Sujit Kumar Debsarma, SAARC Meteorological Research Centre (SMRC), Sher-E-Bangla Nagar, Dhaka-1207, Bangladesh

Bishawjit Chowdhury, Main Meteorological Office, Shah Amanat International Airport, Patenga, Chittagong, Bangladesh

Md. Majajul Alam Sarker, SAARC Meteorological Research Centre (SMRC), Sher-E-Bangla Nagar, Dhaka-1207, Bangladesh

Md. Mizanur Rahman, SAARC Meteorological Research Centre (SMRC), Sher-E-Bangla Nagar, Dhaka-1207, Bangladesh

Nazlee Ferdousi, SAARC Meteorological Research Centre (SMRC), Sher-E-Bangla Nagar, Dhaka-1207, Bangladesh

Uma Charan Mohanty, Centre for Atmospheric Sciences, Indian Institute of Technology, Delhi, Hauz Khas, New Delhi-110016, India


SAARC (South Asian Association for Regional Cooperation) Meteorological Research Centre (SMRC) has taken up a long term programme namely SAARC Severe Thunderstorm Observation and Regional Modeling (STORM) Programme since 2009 and is being continued among Bangladesh, Bhutan, east and northeast India and Nepal. The main objective of the programme is to formulate ideas on modeling of mesoscale convective system (MCS) and validate models with the data collected during Field Experiments over the SAARC region. The Experiment focused on Nor'westers (locally known as “Kalbaishakhi”) that occurred in Bangladesh, Bhutan, eastern and northeastern parts of India and Nepal. Afghanistan, northwest India and Pakistan have started the second phase in 2012 and will continue till 2014 to study the genesis of dry convective storms, dust storms and deep convections. Upon successful completion of the field campaign data collected during the Experiment are being prepared for assimilation in WRF-ARW Model. West Bengal of India and Sylhet, Bogra, Madaripur, Faridpur and Chittagong of Bangladesh experienced moderate/severe thunderstorm (60-100 km/hr) on 17 May 2012. Peshawar, Khyber Pakhtunkhwa, Lahore and Punjab of Pakistan were affected by a sand storm of 05 June 2012. These two cases of SAARC STORM phase I and II have been studied in this paper. Six hourly Final Reanalysis (FNL) data of National Centers for Environmental Prediction (NCEP) have been used as input to WRF-ARW Model for the simulations.

The WRF dynamical core ARW modeling system along with 3DVAR Data Assimilation technique has been used to improve the forecast scenarios of severe MCS during SAARC STORM Pilot Field Experiment 2012. The conventional and non-conventional data are used in the assimilation system. The results indicate that, compared to the control (CTRL) run, 3DVAR Data Assimilation technique substantially improves the overall forecast products over the SAARC STORM region.

Paper 18:


Luca Nisi, MeteoSwiss, Switzerland

Paolo Ambrosetti, MeteoSwiss, Switzerland

Lorenzo Clementi, MeteoSwiss, Switzerland


The purpose of the operational-oriented system COALITION (Context and Scale Oriented Thunderstorm Satellite Predictors Development) is to detect early thunderstorms that will later develop into severe ones and consequently help weather forecasters to increase the lead time in issuing severe weather warnings. This innovative object-oriented system integrates data provided by different sources (Meteosat Second Generation – Rapid Scan Service, Weather Radar, Numerical Weather Prediction and Climatology). One of the system’s purposes is to use operationally available information about convective processes and to integrate it into a heuristic model. Furthermore the orographic forcing, which is often neglected in heuristic nowcasting models, is taken into account and included in the system as an additional convection triggering mechanism. This is particularly important for areas characterized by complex orography like the Alpine region. The COALITION algorithm merges evolving thunderstorm properties with selected predictors. The forecasted evolution of the storm is the result of the interaction between convective signatures (referred as “objects”) and their surrounding environment. Eight different “object-environment” interactions are analyzedin eight modules, providing ensemble nowcast of thunderstorm attributes (satellite- and radar-based) for the following 60 minutes. All ensemble nowcasts are then combined through a weighting and thresholding scheme and the results are summarized into a single graphical map to facilitate user’s interpretation. COALITION has an update frequency of 5 minutes; every new graphical output highlights the cells having a high probability of severe convection development in the next 60 min. A statistical verification shows that COALITION is able to nowcast the intensity of developing convective cells with sufficiently good skill scores up to a typical lead time of 20 minutes.

Paper 19:

The influence of mesoscale mid-level vortices on deep convection and implications for tropical cyclogenesis.

Melville Nicholls, University of Colorado, United States


The formation of a strong mid-level vortex in a mesoscale precipitating system influences the subsequent development of convection in a number of ways. It results in warming aloft, cooling at low levels, and large vertical wind shear at the radius of maximum winds. Moreover as the mid-level circulation intensifies the conversion of latent heat energy released in convective towers to kinetic energy of the balanced rotational flow becomes more efficient. Cloud-resolving numerical model simulations that examine the transformation of a weak incipient mid-level vortex into a tropical cyclone illustrate the importance of these factors. A suite of experiments with the Colorado State University Regional Atmospheric Modeling System (RAMS), show two distinct pathways to tropical cyclogenesis. One pathway involves a steady increase of the surface winds to tropical cyclone strength as the radius of maximum winds gradually decreases. A notable feature of this evolution is the creation of small-scale lower tropospheric cyclonic vorticity anomalies by deep convective towers and subsequent merger and convergence by the low-level secondary circulation. The second pathway also begins with a strengthening low-level circulation, but eventually a significantly stronger mid-level circulation develops. Cyclogenesis occurs subsequently when a small-scale surface concentrated vortex forms abruptly near the center of the larger-scale circulation. The small-scale vortex is warm core throughout the troposphere and results in a local surface pressure fall of a few millibars. It usually develops rapidly, undergoing a modest growth to form a small tropical cyclone. Conditions that favor development along the second pathway and the factors that lead to the formation of a very small surface concentrated vortex are investigated. The relevance of these findings to the formation of Mediterranean tropical cyclones is discussed.

Paper 20:

Changes in hailstone size distributions in relation with a rise in the freezing level

Jean Dessens, ANELFA, France

Claude Berthet, ANELFA, France

José Luis Sanchez, Universidad de Leon, Spain


The altitude of the freezing level in the atmosphere, h0°C, has mixed effects on hailfalls at the ground, because it relates both to convective and melting processes. However, mean hailstone size distribution measurements with hailpads for 125 hail days in the Bordeaux area during the 1989-2011 period show some statistically significant correlations between point hailfall characteristics and h0°C when different hailstone diameter classes are considered. There is a significant negative correlation between the mean hailstone number and h0°C in the small diameter class (5-7 mm), no correlation in the medium class (7-9 mm), and positive correlations in all the other larger classes (> 9 mm). The melting effect reduces the number of small hailstones at the ground, while the convective processes prevail for the number of large hailstones. The absence of correlation in the middle class is noteworthy, because it may help study the effects of parameters other than h0°C.

With these correlations, it is possible to explain the hail intensity increase already observed in Southwestern France. In this region, the surface temperature has increased by 1.8°C from 1989 to 2011 during the months of April to June. For the same period and months, the mean hail day h0°C value has increased from 2.3 to 2.8 km. The trend lines of the correlations reported in the above paragraph show that the number of 5-7 mm diameter hailstones has decreased by 21%, and that the number of 11-13 mm hailstones has increased by 28%. These results are in agreement with numerical simulations of hail climatology in Australia and North America, and they constitute the first physically measured indication of the possible effect of global warming on hail frequency and intensity in various regions

Paper 21:

Diurnal variation of thunderstorms and meteorological conditions observed during the longest thunderstorms in Northern and Central Europe

Katarzyna Grabowska, University of Warsaw, Faculty of Geography and Regional Studies, Poland

Joanna Popławska, University of Warsaw, Faculty of Geography and Regional Studies, Poland


Northern and Central Europe is characterised by a lesser thunderstorm activity than regions located more to the south of Europe. However, these regions can also be affected by dangerous weather conditions, including thunderstorms, posing threat to human life and activities.
The purpose of this paper is to demonstrate yearly and diurnal variation of thunderstorms in the selected cities of Central and Northern Europe: Brest, Warsaw, Orenburg, as well as Trondheim, Tampere and Syktyvkar in years 2005-2012.
In the period under study the longest thunderstorm (lasting several hours) for each city was chosen and then examined to see what weather conditions generated the thunderstorm. The changes in temperature, air humidity, wind direction and atmospheric pressure before, during and after the thunderstorm were analysed.
Data on thunderstorms and meteorological conditions comes from the METAR airport reports. METAR reports are coded messages sent every 30 minutes. The frequency of reporting determines thunderstorm duration which cannot be shorter than 30 minutes.
The analysis of weather conditions during the longest thunderstorms in the selected cities was carried out based on synoptic maps, visible and infrared satellite images as well as vertical aerological sounding.
In all cities under study short thunderstorms, lasting less than 30 minutes, occur most often. Long thunderstorms are rare. The longest thunderstorm lasted 9 hours and occurred in Warsaw. In Northern Europe the longest thunderstorms last for a shorter time, about 4 hours. Thunderstorms occur during the entire diurnal cycle in all cities under study. In the zone marked by marine climate the majority of thunderstorms were observed at 1.30 p.m., in the transitional climate zone – at 5.00 p.m., and in the continental climate zone – at 6.00 p.m.

Paper 22:

New insights about tornadogenesis in supercells obtained from idealized three-dimensional numerical simulations involving a heat source and heat sink in a vertically sheared environment

Paul Markowski, Department of Meteorology, Pennsylvania State University, United States

Yvette Richardson, Department of Meteorology, Pennsylvania State University, United States


Idealized simulations are used to investigate the roles of environmental vertical wind shear and baroclinic vorticity generation in the development of near-surface vortices in supercell-like "pseudostorms." A cyclonically rotating updraft is produced by a stationary, cylindrical heat source imposed within a horizontally homogeneous environment containing streamwise vorticity. Once a steady state is achieved, a heat sink is imposed on the northeastern flank of the updraft at low levels. Cool outflow emanating from the heat sink spreads beneath the updraft and leads to the development of near-surface vertical vorticity. The near-surface cyclonic vorticity maxima comprise parcels that have passed through the heat sink and descended from above (500--2500 m altitude, depending on the negative buoyancy acquired). Baroclinity encountered along the trajectories causes the vorticity vector to become tipped upward relative to the trajectories during the descent, ultimately leading to cyclonic vorticity at the nadirs of the trajectories.

A strong cyclonic vortex forms if the parcels acquiring cyclonic vorticity via the aforementioned process subsequently experience intense stretching, which occurs in the case of an intermediate heat sink and strong environmental shear at low levels. A strong cyclonic vortex fails to form in simulations in which the heat sink is excessively strong or weak, or if the low-level environmental shear is weak.

Paper 23:

Determination of the area’s most exposed to the hail damage in the continental part of Croatia

Damir Pocakal, Meteorological and Hydrological Service of Croatia, Croatia

Zeljko Vecenaj, Department of Geophysics, University of Zagreb, Croatia


Settled in the mid latitudes of the Northern Hemisphere, Croatia is exposed to the frequent occurrence of severe thunderstorms, especially in the continental part. Therefore, occurrence of hail with a high possibility of doing heavy damage in agricultural production is very frequent. In the 1960s, aiming to protect agricultural production from hail, a hail suppression system was introduced in that area. In this study, we use observed hail and damage data collected in the 1981 – 2012 period along with the hailpad data from 2002 - 2012 period collected on hail suppression stations in the warm parts of the year (from the beginning of May to the end of September).

For the spatial analysis of the observed hail and damage data, protected area is divided into 9 km x 9 km quadrants in a way that in every quadrant there is at least one hail suppression station. For each quadrant, a “hail threatening index” (HTI) as an empirical function of average number of days with hail, number of cases with hail damage (scale from 5 to 100 %) and cases with severe damage (scale from 50 to 100 %) in agriculture is calculated. The purpose of this index is to indicate locations or areas that are most frequently exposed to the crops damage caused by hail. In addition, EOF (empirical orthogonal function) analysis is performed on the kinetic energy data measured by the network of hailpads, in order to determine areas /quadrants with greater average kinetic energy in the period 2002-2012, and hopefully confirm findings revealed by HTI. Results of this study, therefore, reveal the areas that would require additional measures of hail protection (passive or active) and/or use of the agricultural plants more resistant to hail.

Paper 24:

Linking meteorological conditions to flood and flash flood occurrence - why is it so difficult?

Thea Turkington, University of Twente, Netherlands

J. Ettema, University of Twente, Netherlands

C. J. van Westen, University of Twente, Netherlands


Projected increases in heavy precipitation for Central Europe indicate potential increases in frequency and magnitude of floods and flash floods. However, before future changes in these hazards can be determined, the current relationship between the hazard and meteorological trigger should be understood. It is known that intense short duration precipitation, long-lasting rainfall, snow-melt are all important for mountainous areas in Europe, yet determining the precise triggers and their temporal probability faces many challenges. Triggering conditions can be complex, and limited data availability and quality can make it difficult to determine statistically significant links. There is also often a spatial mismatch between climate observations and what is actually triggering the flood or landslide. This work will explore the current challenges faced when determining the temporal probability of meteorological triggers for floods and flash floods as well as potential solutions. Examples will be drawn from the Ubaye Valley, France and Fella River Basin, Italy, focusing on how data availability and quality, conceptualization of the problem and different statistically based approaches all affect link between meteorological conditions and hazard occurrence. Once a better understanding of the link between meteorological conditions and hazard is understood, this will hopefully provide the basis for more reliable future projections in hazard occurrence.

This work is part of the EU 7th framework Marie Curie ITN project ‘CHANGES’ under Grant Agreement No. 263953.

Paper 26:

The application of selected methods for detection of tornadoes in Poland (case studies)

Joanna Poplawska, University of Warsaw, Poland

Katarzyna Grabowska, University of Warsaw, Faculty of Geography and Regional Studies, Poland


Every year over the area of Poland we record several dozens of events which the media and the public tend to call tornadoes. However, the majority of such occurrences are connected with strong winds produced by e.g. downburst, and only very few events are actually the consequences of a real tornado. Unfortunately, due to the lack of basic emergency warning system, even those few events happen to be highly destructive and pose threat to human life or well-being. This is mainly the outcome of selective monitoring of tornadoes performed in our country and no practical application of radars to detect them. The fact that no multi-faceted research on this topic has been carried out makes it difficult to specify potential conditions conducive to occurrence of tornadoes in this region of Europe. As a result, implementation operations aimed at safety improvement have been abandoned.
The purpose of this paper is then to specify potential meteorological conditions characteristic for these days when dangerous whirlwinds occur in Poland, as well as to evaluate the possibility to detect them by means of available methods and devices. Based on ESWD reports, a set of tornado-type occurrences was selected. Several radar products were analysed in time intervals of 10 minutes. The following parameters were considered: maximum values of reflectivity, vertical wind profile, vertical wind shear, as well as turbulences. Based on satellite images the altitude of storm cloud tops was defined. Due to the limited radar data available, only some of the cases from the last decade were analysed.
The data concerning radar products comes from the Institute of Meteorology and Water Management. The satellite images were acquired from the website. Additionally, vertical aerological soundings and synoptic maps were used. In order to analyse thermodynamic instability, convective indices were used.

Paper 27:

Forecasting severe convection with a high-resolution local model ensemble

Marcus Beyer, Deutscher Wetterdienst, Germany

Marcus Paulat, Deutscher Wetterdienst, Germany


Due to its spatial and temporal resolution forecasting severe convection is a difficult task for models. Improving the model and increasing its resolution creates improved forecasts but cannot fully solve the problem. In consequence of the chaotic nature of small scale weather phenomena one deterministic model run will always differ from another. To overcome this problem the forecast business increasingly changes from deterministic to probabilistic approaches.

In this poster we introduce the ensemble forecast system of the German local model COSMO-DE (2.8 km), which runs in operational mode since May 2012. To illustrate the potential of such a high resolution ensemble technique two case studies are presented.
The first example is the 23rd of May 2012. This event was characterized by a lack of a strong dynamic forcing. Convection in the western parts of Germany was triggered by a weak short wave. Only low winds were present in the different atmospheric levels. Therefore, the wind shear was also very low. This resulted in slowly moving and pulsating thunderstorms that were mostly steered by their internal dynamics. In consequence this finally led to heavy local precipitation events.
A completely different case is the 30th of June 2012 when severe convection was substantially triggered by dynamics. A strong and positively tilted long wave trough was present over Western Europe. Several short waves circled the upper trough. Together with a strong deep layer wind shear this resulted in widespread initiation of severe thunderstorms in south-western Germany that organized into a big MCS. Besides excessive precipitation, severe wind gusts and very large hail were the outcome.

In this poster we illustrate how the use of ensemble technique has improved the forecast of both events. The advantages in comparison to a deterministic model output are worked out by making use of verification results from the summer season 2012.

Paper 28:

Case study: derecho event in Estonia 08 August 2010 - Modelling derecho dynamics with NWP model HARMONIE

Velle Toll, University of Tartu, Estonia

Aarne Männik, University of Tartu, Estonia

Andres Luhamaa, University of Tartu, Estonia


On 8-th of August 2010 derecho-type thunderstorm moved over Estonia, causing widespread wind damage. Operational model in Estonian Meteorological and Hydrological Institute was unable to predict it. Thunderstorm progress over Estonia is recorded on radar reflectivity images, where approximate speed 25 m/s of northward movement of the bow echo is detected. The strongest wind gusts measured in automatic stations were up to 35 m/s and more than 15 m/s in larger area.

Case study of this situation is carried out using non-hydrostatic convection-permitting HARMONIE model (default spatial resolution 2.5 km), which explicitly resolves deep convection. HIRLAM model analyses are used as initial and boundary conditions in the study. Derecho dynamics and predictability is investigated.

Preliminary results of the simulations show capability of HARMONIE model to predict severe convective storm and reliably represent the dynamics of the storm. Model resolution increase has an impact on representation of storm intensity and dynamic details but 2.5 km horizontal resolution appears sufficient for reliable forecast. Modelling results depend strongly on initial data. With appropriate initial data realistic derecho evolution is forecasted and dynamics of the storm can be analysed.

Operational weather forecast in Estonian Meteorological and Hydrological Institute could have benefited from high resolution (2.5 km) numerical weather prediction environment. The wind damage could have been foreseen using HARMONIE model and appropriate warning about the storm in Estonian Meteorological and Hydrological Institute could have been issued.

Paper 29:

Perception and use of severe weather warnings by emergency management professionals in Germany

Thomas Kox, Freie Universität Berlin, Germany


Warnings of severe weather events are, as weather forecasts in general, always uncertain. Since both the chaotic nature of the atmosphere, as well as inaccuracies in the weather observation and computer models lead to ambiguity and incomplete knowledge (NRC 2006). Thus, the communication of weather warnings is a difficult task, even to experienced users such as emergency managers. Probabilistic information is a way to indicate the level of uncertainty. It can be communicated via numeric information (‘80%’) as well as via a verbal statement (‘very likely’). Nevertheless, it is unclear which form is the most appropriate way to communicate weather warnings (Handmer & Proudley 2007), and whether the information is properly understood and interpreted by the receiver (Gigerenzer et al. 2005).
The paper shows the results of an experts survey, which was conducted in autumn 2012 as part of the WEXICOM project at the Hans-Ertel-Centre for Weather Research of the German Meteorological Service (DWD). The study was devoted to the question how weather warnings are communicated to expert users and how they can be evaluated in terms of their content and their effectiveness. In total 161 emergency managers (fire fighters, police officers and civil servants) across Germany answered an online questionnaire. Inter alia questions dealt with the tools professional forecast users rely on, their confidence in forecasts, user specific lead times or the information format (e.g. level of detail, verbal vs. numerical information, etc.) and hence their understanding of probabilistic information.
A large number of open questions were selected to identify new topics of interest, unknown problems and to identify research gaps in the field of communicating weather information in Germany.

Paper 30:

An assessment of weather warnings communication products

Thorsten Ulbrich, Institute for Meteorology FU Berlin, Hans Ertel Centre for Weather Research, Germany

Uwe Horst Ulbrich, Institute for Meteorology FU Berlin, Hans Ertel Centre for Weather Research, Germany

Martin Goeber, Deutscher Wetterdienst, Hans Ertel Centre for Weather Research, Germany


The research project WEXICOM of Hans Ertel Centre for Weather Research (HErz) of DWD applies both meteorology and social sciences to optimize and adapt weather warnings to stakeholders' needs. Focusing on Berlin we bring together weather services and users of weather warnings to improve warning products.
To gain an overview of existing warning products for actors of society in storm loss mitigation we assemble a catalogue. This catalogue compares products by content, media employed, time of provision and target group.
To evaluate the usability of warning products for professional and non-professional actors we develop a heuristic. This heuristic is based on frequently applied guidelines for web usability (Krug 2005, Nielsen& Loranger 2008) and web credibility (Fogg 2002). Utilising the heuristic we assess warnings provided publicly via websites or Smartphone applications and the national weather services' warning product available exclusively for relief forces (FEWIS).
To detect problems that may come up for non- professional users when they search for information on websites and apps, we conduct a scenario based user test. This helps to understand how the products are applied, what information is enquired, how weather warnings affect decision making, how receivers deal with information on uncertainty and it may uncover origins of problems to understand and find information.
We examine weather warnings broadcasted on TV and Radio, asses their effect on the audience and quantify the reach of broadcasted warnings for specific events. Therefore we analyze whether the content of the broadcast is understood and how it affects behaviour and planned activities.
Fogg, B.J (2002): Stanford guidelines for web credibility. A research summary from the Stanford persuasive technology lab. http://www.webcredibility.org/guidelines, Status: 06.2012.
Frick, J.; Hegg, C. (2011): Can end-users' flood management decision making be improved by information about forecast uncertainty? in: Atmospheric Research 100 (2011) pp. 296-303.
Krug, S. (2006): Don't make me think! A common sense approach to Web usability. Berkeley.
Nielsen, J.; Loranger, H. (2006): Prioritizing Web usability. Berkeley.

Paper 31:

Understanding the Special features of cyclone Aila (2009) after landfall

Radhika Kanase, IITM, India

P. S. Salvekar, IITM, India


The objective of the study is to simulate the special features of the pre-monsoon Bay of Bengal cyclone Aila (23-26May 2009) after landfall. Weather Research & Forecasting (WRF) model with 3 nested domains (60km, 20km and 6.6km) and two-way interaction, is used. The initial and boundary conditions are supplied from FNL dataset with RTG-SST. The single combination of physical parameterization schemes i.e. BMJ as cumulus, WSM as microphysics,YSU as Planetary Boundary Layer (WSM-YSU-BMJ) is considered. The special feature of Aila is its northward movement throughout its life period and its rapid intensification just after the landfall. It maintained its cyclone intensity upto 15 hours after landfall. The results from the model are examined at every 6hr interval and 1hr interval.
Northward movement throughout its life period and the intensification after the landfall are very well captured in the numerical experiment. The model predicted landfall is 3-4 hrs late than the observed landfall and the landfall error is found to be 83 km. Model simulated cyclone maintained its intensity for 12hrs after model landfall. The detail features could be understood clearly when the computed results are examined at every 1hr interval. Thus the observed special features of the cyclone Aila are well captured by the WRF model though the intensity is over predicted. The vertical structure of Aila is also compared with the observed vertical structure before and after the landfall.

Paper 32:

Long term changes in frequency of thunderstorms in the Baltic countries, 1950-2004

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

Agrita Briede, Department of Geography, Faculty of Geography and Earth Sciences, University of Latvia, Latvia

Inga Stankunaite, Lithuanian Hydrometeorology Service Weather Forecasting Division, Lithuania


Long-term changes in thunderstorm climate in the Baltic countries and their reasons are analyzed during the period of 1950-2004. The study area is located in northeastern Europe and encompasses Estonia, Latvia and Lithuania. Visual thunderstorm observations at 42 meteorological stations that were carefully checked for artificial inhomogeneities were used as the main data source. Annual numbers of thunderstorm days (TDs) showed abrupt inter annual changes, fluctuations with an average periodicity of 13 years and statistically significant downward trend with a decreasing rate of 0.9 TDs per decade. During the study period, average annual number of TDs has decreased about 24% in the Baltic countries. TD frequency was highest around 1960 and lowest around 1990. Weaker TD maxima were detected at the beginning of 1970s, in mid 1980s and at the end of 1990s. Weaker minima appeared at the beginning of 1950s and at the end of 1960s and 1970s. Thunderstorms were most frequent in the Baltic countries in case of the circulation type E in the Vangengeim-Girs classification. This indicates the relationship between the presence of warm southerly and easterly airflows and TD occurrence in the Baltic countries. Changes in TD frequency seemed to be largely related to changes in atmospheric circulation conditions. TD frequency in May and June seems to be related to local air temperatures in Eastern Europe. In July and August, more general atmospheric circulation conditions over the Northern and Eastern Europe seem to affect TD frequency in the Baltic countries. It is also possible that the strongest TD maximum around 1960 was in addition affected by atmospheric electricity anomalies caused by frequent atmospheric nuclear tests.

Paper 33:

Impact of Data Assimilation on Simulation of Thunderstorm Events over Bangladesh using WRF 3DVARDA Technique

Nazlee Ferdousi, SAARC Meteorological Research Centre (SMRC), Bangladesh

Sujit Kumar Debsarma, SAARC Meteorological Research Centre (SMRC), Bangladesh

Md. Majajul Alam Sarker, SAARC Meteorological Research Centre (SMRC), Bangladesh


The purpose of this paper is to demonstrate the effect of data assimilation by using Weather Research and Forecasting (WRF) Model in the simulation of thunderstorm events over Bangladesh and its neighborhood. The thunderstorms or Nor’westers which are locally known as “Kalbaishakhis” are mainly the well known short lived severe weather phenomena that cause a lot of damage to properties and loss of human lives in and around Bangladesh almost every year during the pre-monsoon season (March – May). Two moderate thunderstorm events along with squalls lashed over Ishurdi, Dhaka and Chittagong regions of Bangladesh and Shillong of India on 26 April 2010 and Satkhira Rajshahi, Khepupara, Patuakhali, Dhaka and Chittagong regions of Bangladesh and Alipore, Dumdum, Haldia, Bankura, Gaya and Jamshedpur regions of India on 26 May 2010. WRF Model with horizontal resolution of 9 km and 27 vertical eta levels is run for a period of 24 hours using NCEP-FNL data starting from 0000 UTC of 26 April 2010 and 0000 UTC of 26 May 2010 as initial and boundary conditions for those two cases. YSU (Yonsei University) Planetary Boundary Layer (PBL) parameterization scheme for boundary layer option, Rapid Radiative Transfer Model (RRTM) for long wave and Dudhia for short wave radiation scheme have been utilized to run the model. Noah Land Surface Model (LSM) and Kain-Fritsch (KF) cloud scheme are also considered in the Model. An effort is made to simulate those cases by assimilating SYNOP and Upper Air data of Bangladesh and neighbourhood using WRF 3DVAR Data Assimilation (3DVARDA) technique in WRF Model. The simulated rainfalls (convective and non-convective) are compared with TRMM 3B42RT product and rainfall data of Bangladesh Meteorological Department (BMD) for validation process. The model outputs are also compared with satellite and RADAR imageries for validation and utilized in explaining the development mechanism of the thunderstorms. The WRF Model products like dBZ, rainfall intensity, wind flow pattern, vertical wind shear and vertical wind velocity show improvement after DA over CTRL run in both the cases which are quite encouraging.

Paper 34:

Lightning Applications in Weather and Climate Research

Colin Price, Tel Aviv University, Israel


Thunderstorms, and lightning in particular, are a major natural hazard to the public, aviation, power companies, and wildfire managers. Lightning causes great damage and death every year, but also tells us about the inner working of storms. Since lightning can be monitored from great distances from the storms themselves, lightning may allow us to provide early warnings for severe weather phenomena such as hail storms, flash floods, tornadoes and even hurricanes. Lightning itself may impact the climate of the Earth by producing nitrogen oxides (NOx), a precursor of tropospheric ozone, which is a powerful greenhouse gas. Thunderstorms themselves influence the climate system by the redistribution of heat, moisture and momentum in the atmosphere. What about future changes in lightning and thunderstorm activity? Many studies show that higher surface temperatures produce more lightning, but future changes will depend on what happens to the vertical temperature profile in the troposphere, as well as changes in water balance, and even aerosol loading of the atmosphere. Finally, lightning itself may provide a useful tool for tracking climate change in the future, due to the non-linear link between lightning, temperature, upper tropospheric water vapour, and cloud cover.

Paper 35:

Storm water and climatic change in peri – Urban cities in Sub – Saharan Africa

Mutawe Eddy, Millennium Climate Change Research Alliance (MCCRA), Uganda

Mutebi Emmanuel, Makarere University, kampala, Uganda., Uganda

Nambi Elisha, Research International Consult Uganda (RICU), Uganda


Peri – Urban cities in Sub – Saharan Africa encounter stern storm water. Increased storm frequency and intensity related to climate change are intensified by parameters such as the growing occupation of flood plains, increased runoff from hard surfaces, inadequate waste management and silted up drainage. One can distinguish four types of storm water in urban areas: localized storm water due to inadequate drainage; storm water from small streams within the built up area; storm water from major rivers; and coastal storm water. This participatory action research study focused on vulnerability analysis in cities, to investigate people’s perceptions of why storm water occur, how they adapt, actors responsible for reducing the storm water risks and mitigation measures communities undertake. While local people adapt to storm water, recognition of states and development partners’ responsibility to act to mitigate storm water and its causes, especially the aftermaths of climate change, is required.
KEY WORDS: Storm water, climatic change, peri – Urban cities, adaptation and Mitigation

Paper 36:

The Comparison of GLD360 and EUCLID Lightning Location Systems in Europe

Heikki Pohjola, Vaisala Oyj, Finland

Antti Mäkelä, Finnish Meteorological Institute, Finland


The performance of GLD360 (Global Lightning Dataset 360) long range lightning location system has been studied over Europe. The analysis has been done by comparing the lightning location data of GLD360 from year 2011 (May-September) to a reference system, which here is EUCLID (European Cooperation for Lightning Detection); EUCLID is a good reference system because its performance has been analysed around Europe in many previous studies, especially in Austria. We show results about the relative performance of GLD360 from Austria, Scandinavia, North Sea and Spain, as well as the general relative performance over the whole Europe. Our results indicate that (i) the relative detection efficiency (RDE) of GLD360 varies from about 36% to 170% in the EUCLID coverage area, (ii) GLD360 daily RDE in Austria in July varies from 4.4% to 80.8%, (iii) the median relative location accuracy of GLD360 in Austria in July is 2.8 km, (iv) the diurnal variation in Austria in July shows similar feature for both systems, i.e., largest percentage at 16-17 UTC, (v) the median peak current of negative (positive) strokes for GLD360 is 14 kA (11 kA) and for EUCLID for the same area and period 9 kA (7 kA), (vi) the correlation between EUCLID and GLD360 peak current is r = 0.72, and (vii) the boundary of the efficient detection area of EUCLID can be clearly seen when the GLD360 RDE increases above 100%; this coverage boundary is extremely sharp. Our results show that a comparison of this kind reveals many interesting performance issues regarding both the comparative and the reference network, especially the areas where the reference network performs poorly.

Paper 37:

Cold season thunderstorms in Finland and their effect on aviation safety

Antti Mäkelä, Finnish Meteorological Institute, Finland

Elena Saltikoff, Finnish Meteorological Institute, Finland

Jukka Julkunen, Finnish Meteorological Institute, Finland

Ilkka Juga, Finnish Meteorological Institute, Finland

Erik Gregow, Finnish Meterological Institute, Finland

Sami Niemelä, Finnish Meteorological Institute, Finland


A total of 13 commercial airplanes were struck by lightning in October (10 in one day) and December (3 on 3 separate days) 2011 in the main Finnish Helsinki-Vantaa airport. The number of lightning-struck airplanes is extremely large, considering the time of year and the small number of flashes by the storms. In this paper we show characteristics of these cases regarding the synoptic situation as well as their forecasting. There were remarkable differences in the operational models; the high-resolution non- hydrostatic model was superior in predicting the convective nature of the event compared to the coarser resolution hydrostatic model. The interview of the pilots of the struck airplanes shows that the pilots did not receive detailed information to avoid the situation; also, the lightning strike affected the pilots even causing temporary loss of sight and hearing. Luckily, no fatalities or severe damages to the airplanes occurred. The most interesting case is October 19, 2011; during this single day, a total of 10 airplanes were struck. The analysis suggests that a major cause for the large number of struck airplanes is that the planes took off directly into the convective core of the storm and the planes initialized the flashes themselves. However, the time of the year, the near position of the storm area relative to the takeoff path, and the necessity to use only a certain takeoff path because of the direction of wind makes the convective scenario difficult to predict and avoid. The pilots have expressed interest to receive training for these cold season thunderstorms.

Paper 38:

Extreme precipitation and related weather types over Croatia in the period 2001-2011

Dunja Plačko-Vršnak, Hydrological and Meteorological Service, Croatia


In this study a classification of weather types over Croatia for cases with extreme precipitation during the period 2001-2011 is done. Due to the irregular shape and climatic diversity, Croatia is divided in six sub-regions. The criterion for a daily extreme precipitation is different for every sub-region and is based on climatological return period of 5 or 10 years. In the 2001-2011 period, 41 extreme precipitation events were observed.The highest daily amount (188,2 mm) was measured at the climatological station Lokve Brana, in the Mountainos part of Croatia, in winter. The extremes occur from May till December, and are most frequent in September (30 %) and August (22%). From January till April, and also in July extremes were not recorded. The East Croatia is a region with highest number of extremes according to climatological threshold. The clasification of weather types (developed by Poje, 1965) for cases with extreme precipitaton is done according to surface and upper-level pressure fields. The method is subjective, based on experience of forecaster. This approach is highly sensitive to minor differences in the route or location of synoptic systems in conjuction with the complex topography and sea-land exchange. Out of 29 weather types, 8 were detected during days with extreme precipitation. Even 80 % of detected weather types were connected with the influence of the depression. The most frequent (44%) weather type is C1 – eastern (front) sector of cyclone where the center is positioned westward of Croatia, most often over Gulf of Genoa and the North Adriatic. Upper-level pressure charts were analyzed to determine the prevailing large-scale flow that existed in the days with extreme precipitation. The most common flow regime is the SW, observed with a frequency of approximately 41 %. The SE flow also occurs with a relatively high frequency (28 %).

Paper 39:

Significant hail producing storms in Finland: Storm morphology and environment

Jari Petteri Tuovinen, Finnish Meteorological Institute, Finland

Jenni Rauhala, Finnish Meteorological Institute, Finland


Significant hail (diameter 5 cm or larger) has been documented in Finland during the 40-yr period of 1972−2011 during 22 significant-hail days. The storm environments are studied using observed soundings. Soundings are modified near surface layer based on the nearest surface observation prior to the first significant hail report. Environments of significant-hail days are characterized by relatively high MUCAPE values, but only one third (8, 36%) of the days have 0-6 km bulk shear values higher than 15 m/s. Generally larger CAPE and wind shear seems to support larger diameter hail. The convection is elevated in 5 out of 22 significant-hail cases (23%).

The radar-based convective modes and storm characteristics are determined for storms leading to 18 documented significant hail cases in Finland during 1999−2011. Observed parent storm types were right moving cluster supercells (8 cases), right moving discrete supercells (5), left moving discrete supercell (1), cluster cells (2) and discrete cells (2). The results show that most (14, 78%) of the significant hail causing storms in Finland are supercell storms and have a lifetime of more than 3 hours, 30% more than 5 hours. Majority (67%) of significant hail producing storms move towards north-east quadrant.

Significant hail development in a storm is faster with increasing MUCAPE values. Also significant hail forms faster in cluster supercells than in discrete supercells. Discrete significant-hail producing supercells have longer lifetime than cluster supercells and they form in larger bulk shear environments.

Paper 40:

Convective activity over Ukraine

Inna Semenova, Odessa State Environmental University, Ukraine


Such phenomena as hail, strong thundershowers, squalls (and tornados) are observed in Ukraine annually, that is why an actual task is an improvement of quality of their prognoses with use of synoptic and numerical methods.
The frequency of strong weather conditions in Ukraine accompanied by convective systems is analyzed by the data of European Severe Weather Database (ESWD) during May to September 2000-2010. In this period was observed 264 dangerous convective phenomena, namely: 223 cases with a large hail, 49 cases with strong thundershowers, 21 case with tornados and 21 case with strong gusting. Amount of all phenomena is increased from year to year and can be related with the climatic changes.
The intensity of convective processes has been estimated using synoptic data and numerical radiosonde parameters and specified the convective criteria for territory of Ukraine. It is showed that most acceptable parameters for definition of convective phenomena were the index of SHOW, K-index and Precipitable water, which reaches dangerous criteria in 58-83% cases.

Paper 41:

Anomalies in frequency / intensity correlations in hail climatology

Claude Berthet, ANELFA, France

Jean Dessens, ANELFA, France


There are two main parameters used to represent hail in a climatic region: the frequency of point hailfalls with hailstones larger than 0.5 cm, F, and the mean intensity of these hailfalls, in general characterized by the hailstone total kinetic energy, Eh. These parameters can be computed at different periods or in different time slots. In this paper, the large data sample available in France after 24 years of hailpad measurement is used to present the F/Eh diagrams relative to years, months, days of the week, and hours of the day. This presentation is given for two areas of Southwestern France with different climates, the Atlantic region near Bordeaux, AT, and the Midi-Pyrenean region around Toulouse, MP.

The F/Eh diagrams reveal anomalies in the correlations between the two parameters. The points relative to a time period are generally grouped in packages or along trend lines, but some of them are largely outside the mean correlation, as for example: extremely high value for Eh in 1992 in MP, high April and low May values for Eh both in AT and MP, larger Eh values for Saturdays and Sundays in MP, and a time-lag in the daily evolution of F and Eh, this parameter being the first to decrease in the evening. Some of these anomalies can be explained by meteorological conditions and cloud physics, others may still be due to the sample dimensions, but all need to be surveyed in the future. This study also confirms that F and Eh are two relatively independent parameters which both need to be considered for the hail climatology in a region.

Paper 42:

Increased variability of tornado occurrence in the United States in recent years

Harold Brooks, NOAA/NSSL, United States


Since 2000, a large number of tornado records have been set in the United States, both for maximum and minimum monthly occurrences in many calendar months (F1 and stronger). In addition, the timing of the early part of the tornado season has become dramatically more variable. The extremes come against a background of very small changes in mean occurrence and appear to continue a trend that began over 30 years ago. The number of days with at least one F1 tornado in the US has decreased since the mid-1970s, but the number of tornadoes on the biggest days has increased. It is difficult to explain those two aspects as effects of reporting changes. As a result, it is possible that it is a physically real occurrence and may represent a response to climate change. A simple statistical model explains many of the results. Its relationship to global climate will be discussed.

Paper 43:


Maria Tous, Universitat de les Illes Balears, Spain

Romualdo Romero, Universitat de les Illes Balears, Spain

Climent Ramis, Universitat de les Illes Balears, Spain

Kerry A. Emanuel, Massachusetts Institute of Technology, United States


Medicanes or “Mediterranean hurricanes” are extreme cyclonic windstorms morphologically and physically similar to tropical cyclones. Although medicanes are rare phenomena and their size and intensity are clearly lower than for hurricanes, expected changes in frequency and intensity as a consequence of global warming are a cause for regional concern. Consequently, this study aims to investigate the long-term risk of medicane-induced winds.

First attempts to assess the medicane risk have been undertaken by our group based on two different perspectives. The first approach consists on detecting and tracking symmetric warm-core cyclonic disturbances generated in nested climatic simulations. As an alternative, the second method takes advantage of the statistical-deterministic approach developed by Emanuel and his team in the context of hurricane risk assessment in the tropical regions. This approach generates thousands of synthetic storms with a low computational cost, thus enabling a statistically robust assessment of the spatial-temporal risk function.

Here we present unprecedented medicane risk maps based on both techniques, using ERA40 reanalysis and GCM outputs as input data. Results derived from ERA-40 are compared, on the one hand, against the satellite-based climatology and, on the other hand, against the GCM-derived results for the baseline period. These comparisons allow us to calibrate the methods and to better assess the future effects of climate change on medicane frequency and intensity.

Both methods generally agree with regard to the medicane-prone geographical areas and seasonal incidence, pointing out the central-western Mediterranean basins and the cold season as the most favorable. Regarding the effects of climate change, fewer storms but an increased probability of violent cases at the end of the century compared to present are projected.

Paper 44:

Estimating Severe Thunderstorm Risk in North America

Scott Stransky, AIR Worldwide Corporation, United States

Tomas Girnius, AIR Worldwide Corporation, United States

Eric Robinson, AIR Worldwide Corporation, United States


Severe thunderstorms producing tornados, hail, and straight-line wind can cause significant property losses in North America. This is evidenced by the nearly USD 30 billion of insured property losses inflicted by severe thunderstorms in the United States in 2011. Stochastic modeling of catastrophic events is an important risk assessment tool used by insurers, reinsurers, government agencies, and other stakeholders. We describe the development of a severe thunderstorm hazard model for the United States and Canada, which can be used to estimate the average annual and extreme low-probability financial losses that could occur in this part of the world.
The methodology is based on Monte Carlo simulation which stochastically generates a large number of potential annual scenarios of severe thunderstorm activity along with the associated losses. The event generation module is based on historical severe thunderstorm data available from NOAA’s Storm Prediction Center (SPC), supplemented with information from the Climate Forecast System Reanalysis of the daily large-scale convective environment. The historical storm reports are heuristically de-duplicated, appropriately smoothed and de-trended, and combined with meteorological parameters indicative of favorable conditions for severe storm development, such as CAPE, STP, and SHIP. Where storm reports are available, these parameters help constrain the smoothing and sampling. In addition, the use of these parameters allows for the seamless integration of the U.S. and Canada, where less historical data is available. Each individual simulated occurrence contains an associated geocode, intensity, length, width, and direction that help determine the hazard footprint.
After vulnerability relationships are developed, the simulated events are combined with exposure information to estimate the financial losses.

Paper 46:

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

Shadananan Nair, Nansen Environmental Research Centre (India), 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 of cyclones, thunderstorms, lightning and floods are observed. There is an increasing tendency of thunderstorm development in the Western Ghats region where the squalls cause widespread damage to settlements, agriculture 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 47:


Luca Nisi, 1) Oeschger Centre, University of Bern 2) MeteoSwiss, Switzerland

Olivia Martius, Oeschger Centre, University of Bern, Switzerland

Alessandro Hering, MeteoSwiss, Switzerland

Urs Germann, MeteoSwiss, Switzerland


During the warm season of the year, intense thunderstorms regularly affect the Alpine area. High impact convective phenomena like severe hailstorms can cause substantial hail, water, and storm damage to agriculture, forest, buildings, and cars. In Switzerland severe summer storms are among the costliest high-impact weather events and it is therefore important to understand the dynamics and physics of these phenomena and their interaction with the complex Swiss topography in more detail. Recently a new project focusing on hail storms over the Swiss Alpine region started thanks the collaboration of the University of Bern and MeteoSwiss. An important first step is to compile a short but comprehensive climatology of spatial and temporal characteristics of hail storms in Switzerland. For this purpose two radar-based products referred as Probability Of Hail (POH) and Maximal Expected Size of Hail (MESH) have been reprocessed for the period 2002-2012. These products have been analysed over the whole period in order to investigate spatial and temporal distribution of hail storms. In this paper first results of this new climatology will be discussed and presented. Furthermore, an overview on future works will also be provided.

Paper 48:

Nowcasting of severe weather at DWD using remote sensing and nowcast product data

Tim Böhme, Deutscher Wetterdienst, Germany


At Deutscher Wetterdienst (German National Weather Service, DWD), severe weather warnings in nowcasting base mostly on remote sensing and other observational data analysis and their extrapolation. The presentation will give an overview about the available remote sensing data and nowcast products at DWD.

In autumn 2012, the radar scan strategy was changed by shorten the operational volume scan interval from 15 min to 5 min. Thus, all radar product data are now available in high temporal resolution. This will allow to develop modern 3D and updated nowcast products. Several case studies which show the benefit of these temporally highly-resolved data for nowcast tasks will
be presented.

Beside observational data, already operational nowcast product data for identification and tracking of warning objects and areas (e.g., Konrad, CellMOS and RadVor-OP), new product data resulting from 3D remote sensing data (e.g., convective initiation and meso-cyclone detection) and modern merging techniques (e.g., NowcastMIX and ITWS) which are currently in evaluation stage will be presented. Actual case studies will demonstrate the capability of these products for the nowcasting of severe weather.

All meteorological data are illustrated by visualisation from the system NinJo. NinJo allows to survey the weather situation comprehensively but also in real-time in operational use as all important weather information is included into this system regularly immediately after the data transfer or production. With the case studies the design of the data of remote sensing and nowcast products in NinJo will be presented.

Paper 49:

Improving Forecasters Skill by introducing Convective Initiation at DWD

Pierre Fritzsche, Deutscher Wetterdienst, Germany


DWD role by law is to inform and warn the public about significant and potential dangerous weather. Detecting possible severe Thunderstorms well in advance, would increase the lead time in warning and therefore save lives and losses arising from one event. Increased understanding in the field of convective and pre-convective mechanism and the use of satellite data provides the opportunity to detect areas with future convective clouds (growing over -35 dbZ in radar data) up to 1h in advance. By transferring the SatCast method at EUMETSAT to Europe, the code is analysed and optimized for operational purposes by DWD. The result was a first implementation at DWD, as well as on EUMETSAT.
In a second step a training and evaluation campaign was started. This included a multi-year climatology in order to improve our understanding in the dynamics of the physical based interest fields in different configurations of used AMV’s, clustering and evaluation technics. The presentation includes a description of how the development at DWD is connected with active forecaster to maximise the usability and therefore the impact of the product in operational use.

Paper 50:

Hail Risk Areas in Austria, on the basis of reports 1971-2011 and Weather Radar Images 2002-2011

Otto Svabik, ZAMG, Austria

Vera Meyer, ZAMG, Austria

Lukas Tüchler, ZAMG, Austria

Gernot Zenkl, ZAMG, Austria


ZAMG performed a hail risk map of Austria, as an additional parameter inside eHORA (Natural Hazard Overview & Risk Management Austria), in charge of the federal ministry of environment demonstrating risk areas for storms, floods and earthquake.
Basic data were on the one hand the chronicle of our institute showing severe weather events including hailstorms and resulting damages. Finally 435 cases were analysed, all through the years 1971 -2011. On the other hand weather radar composit images, stored since 2002 (time resolution 10 minutes, spatial resolution of 2x2 km during the year 2002, afterwards 5 minutes, resp. 1x1km resolution). Up to 162 hail cases could be combined with cells, detected within the radar images for the last ten years, and at last the tracks of the hailing cells.
All 435 hail events were classified, following the lines of the TORRO hail intensity scale. The scale extends from H0 to H10 with its increments of intensity or damage potential related to hail size.
The complex topography of the Alps has an influence on frequency and intensity of the hail events, expressed by the detected tracks of the hailing cells and the spatial distribution of the different hail intensities.
The tracks north of the main ridge of the Alps run mainly from southwest to northeast. Within the areas south and southeast of the main ridge from west to east, respectively from northwest to southeast.
The strongest hailstorms (intensity TORRO 7, effecting total losses on fields, severe roof damages and risk of serious injuries) are located along the foothills of the Alps, along the hilly regions close to the lowlands. For instance, the Bohemian Massif close to the Danube valley.
In the presented map of regional hail hazard of Austria, all regions up to 1500 meters above sea level were investigated.

Paper 51:

Using self-organizing maps to classify supercell proximity soundings from the rapid update cycle.

Christopher Nowotarski, Pennsylvania State University, United States

Anders Jensen, Pennsylvania State University, United States


A dataset of 1185 rapid update cycle (RUC) supercell proximity soundings is clustered using self-organizing maps (SOMs) in an effort to identify distinguishing traits of supercell and tornado environments. A SOM is a topologically-ordered array of a pre-specified number of nodes, developed through a learning algorithm which identifies the salient features of the input data. Vertical profiles of relevant kinematic and thermodynamic variables derived from the proximity soundings are clustered according to the node to which they are most similar. The advantage of this technique compared to traditional parameter-based discrimination techniques (e.g., supercell composite parameter, significant tornado parameter) is that information over the entire depth of the vertical profile, including its shape, is considered rather than a point value computed from bulk parameters at fixed, somewhat arbitrary levels. We assess the ability of SOMs of individual variables in differentiating between environments supportive of different observed storm types (nonsupercell, nontornadic supercell, weakly tornadic supercell, significantly tornadic supercell) by comparing the distribution of profiles associated with each storm type in each cluster. In this way, we identify effective variables for future development of the SOM technique as means of issuing conditional probability forecasts of storm type. We also examine patterns in composite hodographs, location, seasonality, and diurnal variability associated with each SOM cluster as a means of identifying regimes that are favorable (or unfavorable) for particular storm types. The sensitivity of the results to the number of SOM nodes and the depth of the vertical profiles used in the SOM are also considered. Finally, we expand the use of SOMs to vertical profiles of multiple variables such as temperature and dewpoint temperature or both zonal and meridional wind components (i.e. a hodograph) and compare these results with SOMs of vertical profiles of single variables.

Paper 52:

Mountain waves forcing deep convection at the East of the highest Andes tops

Rodrigo Hierro, CONICET - Universidad Austral, Buenos Aires - Argentina, Argentina

Horacio Pessano, Facultad Regional San Rafael - Universidad Tecnológica Nacional, Mendoza - Argentina, Argentina

Pablo Llamedo, CONICET - Universidad Austral, Buenos Aires - Argentina, Argentina

Alejandro de La Torre, CONICET - Universidad Austral, Buenos Aires - Argentina, Argentina

Andrés Odiard, Dirección de Agricultura y Contingencias Climáticas - Gobierno de Mendoza - Argentina, Argentina

Peter Alexander, Departamento de Física - FCEN - Universidad de Buenos Aires, Buenos Aires - Argentina, Argentina


In this work we analyze a sample of 400 storm cases, between 2004 and 2011, that produced a considerable damage to cultivated areas at the North of Mendoza, Argentina. This orographic semiarid region at midlatitudes (roughly between 32S and 36S) at the East of the Andes constitutes a natural laboratory where usually diverse sources of gravity waves take place. From WRF model simulations, radar network and radiosounding data, the lifting mechanism required to raise air parcels to levels of free convection, partially supplied by mountain waves (MWs), is analyzed. Based on a previous work, a criterium for the significant presence of MWs is established. Near 30% of the analyzed cases exhibited MW structures during the storm events. Stability conditions in the presence of MWs are evaluated from convective available potential energy (CAPE) and convective inhibition (CIN) indices. The energy provided by MWs is compared with the necessary to trigger convection. A Morlet Continuous Wavelet Transform along zonal and meridional direction is applied to vertical air velocity to isolate parameters belonging to dominant stationary MW modes. Specific features between radar data and dynamic processes found for this sample, with and without significant MWs, are discussed.

Paper 53:

Towards a two-moment bulk microphysical parameterization of hail with explicit liquid water fraction

Vivek Sant, Max Planck Institut for Meteorology, Germany

Axel Seifert, Max Planck Institut for Meteorology, Germany


Most bulk microphysical parameterizations of hail make strong simplifications when it comes to wet growth and melting of hailstones. This can lead to significant problems and biases in simulations of hail growth and the amount of hail reaching the ground. Especially for very high-resolution simulations of deep convection, a microphysics scheme which allows for internally mixed particles with explicit liquid water fraction would be appropriate or even necessary. We will present a framework for such a parameterization and the methodology for the formulation of the closure equations. First results of a one-dimensional model of melting hailstones will be presented. This project is part of the German research program 'High-definition clouds and precipitation for advancing climate prediction', which aims at the large-eddy simulation of clouds and deep convection on the meso-alpha-scale.

Paper 54:

The role of 3D vortex-vortex interaction and superhelicity in tornado maintenance and development

Marcus Büker, Western Illinois University, United States

Gregory Tripoli, University of Wisconsin - Madison, United States


The interaction of quasi-horizontal vortex features with a vertically-aligned vortex (e.g. a mid-level meoscyclone or a tornado) is shown to cause highly localized intensification of the vertical vorticity, given a proper relative geometry. Horizontal vortex tubes are ubiquitous in regions of severe convection; additionally, many examples of video evidence and numerical simulations routinely show these structures wrapped around a parent tornado vortex. The scalar product of vorticity and its curl, termed superhelicity, will be present whenever horizontal vortex structures are curved around an axis parallel to a given background vorticity. Superhelicity has been noted many times in the literature in relation to helicity (more specifically, the viscous destruction thereof), but to our knowledge, has not been related to localized vortex intensification. We will illustrate this relationship through idealized numerical simulations. Additionally, we will postulate situations where favorable relative geometries of updraft and downdraft pulses, mesocyclones and tornadoes would likely result in vortex intensification near superhelical regions.

Paper 55:

Calculation characteristics of catastrophic floods on the mountain rivers of the Crimean peninsula

Valeriya Ovcharuk, Odessa State Environmental University, Ukraine

Elena Todorova, Odessa State Environmental University, Ukraine

Ekaterina Myrza, Odessa State Environmental University, Ukraine


The Crimea rivers belong to flood type. The discharges of water on it during floods sharply increase and can in 200 - 400 times exceed average annual values.
At the rivers, taking began on the north slopes of the Main ridge (Al'ma, Kacha, Bel'bek, Chornaya and other), snow prevails, and for the rivers of the South bank is a rain feeding. Protecting from floods is traditionally conducted on two directions. First is diking, strengthening, straight and river-beds deepening. So on many areas of river-bed of the rivers straight or fixed concrete flags. During a flood in August 1997 on the river of Salgir, at the village of Dvurech'e water on the diked areas did not go out on the flood plain. However much the straight and deepening of river-beds of the rivers often results in sorrowful results. So the straight of river-beds of the rivers is instrumental in formation of landslides, swamping and submergence of flood plain, busy agricultural area. Deepening of the rivers is violated by an ecological equilibrium, violates the systems of cracks and reaches, and on the whole - water gidrobiocenos. The second way is adjusting of floods storage pools, that detention and subsequent gradual up cast of water, not resulting in floods. Here also there are certain problems. For example, an emergency up cast, counted on the discharge of 270 m3/s, has the Simferopol'skoe storage pool, and the river-bed of Salgir in a city line hardly skips 30 m3/s. A therefore control action of storage pool is its far fewer possibilities.
Taking into account aforesaid, development of reliable method for the calculation of maximal flow of floods on the rivers of Crimea is an urgency scientific task for this region.

Paper 57:

Severe weather in North America

Peter Hoeppe, Munich Re, Germany


North America has been particularly hard hit by weather catastrophes such as hurricanes, tornadoes, floods, wildfires, searing heat and drought in recent years. Analyses of the 3,900 natural catastrophe loss records for North America since 1980 stored in Munich Re's NatCatSERVICE database yield an almost fivefold increase in the number of weather disasters, indicating that in this region, weather risks are changing faster than anywhere else in the world.
The main contribution to the upward trend comes from socio-economic factors such as ongoing urbanization and increasing values, but climate variability and climate change play a role, too. Oscillations such as ENSO and AMO have a distinguishable influence on the occurrence and intensity of certain weather extremes. There is also some evidence that anthropogenic influence on the climate has at least lead to more severe thunderstorm-related losses in the U.S.
Annually aggregated thunderstorm losses have reached levels comparable to those from hurricanes, even though the latter can cause larger single-event losses; coastal and inland floods have billion-dollar damage potentials; the 2012 heatwave and drought revealed how far-reaching and expensive temperature extremes can be, setting the stage for wildfires, power breakdowns and huge losses in agriculture; ice-storms in the Northeast, winterstorm-triggered landslides in the Western States, and countrywide subsidence-heave-problems complement the range of costly weather-related hazards.
While hazards are inherently natural, the resulting risk, defined as the integrated product of intensities and probabilities of hazard events and the corresponding consequences, are determined by humans. They decide where values are placed and how they are protected. Hence risk mitigation must consider all three contributing components: hazard, exposure, and vulnerability. Risk reduction can only be effectively achieved given widespread awareness of what is at stake and a willingness to take the necessary mitigation, preparation and defense measures.

Paper 58:

A Subtropical Squall line in Southern China: Kinematic Structure Retrieved from Dual-Doppler Data

Haiguang Zhou, Chinese Academy of Meteorological Science, China


A squall line moved from Guangxi province to Guangdong province in South China during 23 and 24 April 2007 that causes gale, thunderstorm, and heavy precipitation. The 3D wind field of the squall line is retrieved by Guangzhou and Shenzhen dual-Doppler radar volume data in the mature period. In the horizontal, there is storm-relative front-to-rear flow at the low level of the squall line frontal. On the other hand, the storm-relative cold rear-to-front flow is shown at the stratiform region of the squall line on the low level. The rear inflow enhances the convergence in the frontal of the squall line. On the mid-level, the storm-relative front-to-rear flow prevails. The downdrafts are located at the rear of the squall line. Near the density current front, there are updrafts at the low and middle level. The updraft at the midlevel is stronger than the one’s at the low level. There are some convergence centres at the density current front on the low level. The absolute value of the strong convergence is more than 3×10-3s-1. The storm-relative flow in the vertical cross-section perpendicular to the quasi-two-dimensional squall line is also revealed. In the low and middle level of the density current front, there is deep front-to-rear flow. Part of the front-to-rear flow blow backward, and some flow blow upward and then forward. There is a descending rear-to-front cold flow at the low level of the squall rear. The 3D structure conceptual model of the subtropical squall line in South China is proposed.

Paper 59:

Climatic Characteristics of Thunderstorms in Latvia

Zanita Avotniece, Latvian Environment, Geology and Meteorology Centre, Latvia

Māris Kļaviņš, The University of Latvia, Latvia

Agrita Briede, The University of Latvia, Latvia

Lita Lizuma, Latvian Environment, Geology and Meteorology Centre, Latvia


Convective atmospheric phenomena such as thunderstorms, which are commonly accompanied with heavy showers, hail and squalls, have been recognised as impetuous and hazardous meteorological events associated with major threats to both natural ecosystems and human health and safety. Therefore evaluation of the distribution and long-term changes in both the intensity and frequency of thunderstorms is crucial for climate change risk assessment.
This study focuses on the long-term characteristics of thunderstorm climatology, describing the spatial distribution and trends in the frequency and intensity of thunderstorms in Latvia over the period 1950-2010. The study covers 14 major meteorological observation stations in Latvia, and in addition to ground observations, satellite observation data provided by CM SAF (The Satellite Application Facility for Climate Monitoring) were used to estimate areas subjected to the most frequent thunderstorm development. Daily data on atmospheric circulation conditions following the GWT classification recommended by the COST Action 733 were used for the analysis of circulation conditions favourable for thunderstorm development.
The analysis of ground observation data revealed significant differences in the spatial distribution of thunderstorms in Latvia, with a high connection to the orography and proximity to the Baltic Sea. In addition the analysis of satellite data highlighted areas subjected to an increased convective activity that was not covered by the weather stations. The most favourable conditions for the development of thunderstorms have been detected during days with south-west and west cyclonic flows or the centre of a low pressure area located over the Baltic Sea area. The long-term trends in the thunderstorm frequency showed a decreasing tendency, however the increased frequency of hail and thundershower intensity could be an indicator of an increase in thunderstorm intensity. Besides the highest number of thunderstorms on the record has been observed by the end of the study period – in the year 2010.

Paper 60:

INCA system severe weather nowcasting fields - a new tool in the forecasting toolbox.

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


Proper understanding, interpretation and application of various severe weather information sources is crucial for human economy. Weather events inflicting water management, road maintenance and civil protection fields resulting in life and financial losses are more and more common in Poland as we are facing progressing climate changes. Correct recognition and instant nowcasting information transmission is in focus of attention for both meteorologists and hydrologists.

INCA nowcasting system (Integrated Nowcasting through Comprehensive Analysis) was originally designed in ZAMG (Vienna). Now it is being developed under consortium INCA-CE gathering 16 institutional partners from 8 Central European countries to provide ultra-short range meteorological forecast for CE area. Local implementations of INCA system are also working in China and Israel. INCA system gathers all real time meteorological information available (radar, automatic rain gauges, temperature, pressure, humidity and wind among others) and processes it to final products: analysis and prognosis fields of numerous meteorological parameters. Very dense 1km grid of the system output fields makes it very important tool for local use in small scale regions as well as in mountainous and other land surface diverse areas.

In the presentation the objectives of INCA-CE Project will be discussed – developing of nowcasting system specifically adopted to its day-to-day use in operational hydrology, civil protection and road safety, provision of common internet forecast platform for CE area, improvement of decision making process for end-users – just to name few of them. In addition a case study comparison of chosen INCA field will be conducted to present its effectiveness and mutual relations with data fields used in severe storm forecasting routine up to now. Since all ‘ground truth’ data fields have their characteristic systematic errors INCA fields were considered as an alternative.

Paper 61:


Mária Putsay, Hungarian Meteorological Service, Hungary

Zsófia Kocsis, Hungarian Meteorological Service, Hungary

Marianne Koenig, EUMETSAT, Germany

André Simon, Hungarian Meteorological Service, Hungary

Ildikó Szenyán, Hungarian Meteorological Service, Hungary

Márta Diószeghy, Hungarian Meteorological Service, Hungary


The aim of a Convective Initiation (CI) product is to identify those cumulus clouds which are likely to develop into mature cumulonimbus. A cloud tracking CI algorithm for 15-minute Meteosat Spinning Enhanced Visible and InfraRed Imager (SEVIRI) data was developed at EUMETSAT (Kocsis et al., 2012). This algorithm is based on the Mecikalski and Bedka (2006) method, and uses the NWCSAF Cloud Type product to select the low- and mid-level cloudy pixels and HRW product to track these clouds.

In the frame of a EUMETSAT Scientific Study, Hungarian Meteorological Service (OMSZ) adapted this algorithm and studied CI and satellite derived environmental (atmospheric water vapour content and instability) parameters together in severe and non-severe storm cases. The aim is to investigate the possibility to combine the CI information with the environmental instability parameters to further improve the reliability of the product, and to study whether one can find any trace of possible later severity already at the developing stage of the thunderstorm. The CI interest fields (brightness temperature values, differences and their trends characterizing updraft strength, glaciations and cloud depth) are studied together with the environmental parameters and with other information on the severity level of the convective storm (system), like maximums of the surface measured wind guest, radar reflectivity, radar derived hail probability and vertical integrated liquid (VIL) values (characterizing the size of the hail particles), severe storm reports, etc. A case study and statistical results will be demonstrated in the poster.

Paper 62:

Object-based fuzzy logic fusion of multiple data sources for nowcasting of CI and storm lifecycles

Dennis Stich, DLR, Germany

Caroline Forster, DLR, Germany

Arnold Tafferner, DLR, Germany


Information gained by storm tracking and nowcasting tools, based on remote sensing data from radar or satellite, can be improved by using additional data sources. A methodology to include multiple additional data to improve the satellite-based CI (convection initiation) detection of DLRs Cb-TRAM (Cumulonimbus Tracking and Monitoring) algorithm by postprocessing was successfully developed and is presented here. Data which indicates the availability of the three main ingredients of a storm development – moisture, instability, and lift – is evaluated per each object which was detected as CI suspect from the satellite view. After testing many additional data, the results for equivalent potential temperature at the surface (moisture indicator), the KO-Index (instability), and vertical motion in 500 hPa (lift) were chosen to be the best candidates for a combination with fuzzy logic. The combination results in a CI forcing value for each CI detection. With the help of verification statistics for a whole convective season over Central Europe, the CI forcing value was translated in a “CI probability” which describes the probability that a cell detected as CI will further develop. The CI probability is an additional information and, depending on user-requirements, can be used for the filtering of 5-65% of all false CI alarms in Cb-TRAM resulting in a considerable improvement of the CI detection.
In recent and ongoing work, this concept is adapted for studies on storm lifecycles. The mature storm objects used here are Rad-TRAM (Radar Tracking and Monitoring) storm detections which shall be further analyzed with indicators for the trend of the storm cell - is it growing, unaltered, or decaying? The additional data for the fuzzy logic combination here are, for example, cloud top temperatures, radar reflectivity, radar cell size, lightning density, size of the lightning cell within the radar object, etc.

Paper 63:

Relationship between lightning characteristics and radar estimated parameters during pre-severe and severe stages of hail producing thunderstorms developed over Bulgaria

Tsvetelina Dimitrova, Agency Hail Suppression, Bulgaria

Rumjana Mitzeva, Faculty of Physics, University of Sofia, Bulgaria

Hans D. Betz, University of Munich, Department of Physics, Germany

Yana Pisarova, Agency Hail Suppression, Bulgaria

Elisaveta Peneva, Faculty of Physics, University of Sofia, Bulgaria


Several studies have shown that there is a relationship between lightning activity and radar characteristics of thunderstorms. Some studies search for correlation between radar echo top, maximum radar reflectivity and flash rate, FR. Others are directed to establish if there is a relationship between vertically integrated liquid, VIL and FR. However, the obtained results vary depending on geographical location, seasons, etc.
Various radar estimated parameters (e.g. radar cloud top, height of radar reflectivity factors, VIL, VIL density) and lightning characteristics (flash rate, multiplicity, polarity and peak current) are analyzed for hail producing thunderstorms, developed over Bulgaria in the period 2010-2012 years. Since the non-inductive mechanism plays the major role in the electrification of thunderstorms and it depends on the supercooled liquid water content, a special attention is given to the analysis of VIL and thickness of cloud cells in the sub-freezing cloud region. The studies are performed separately for non-severe and severe stages of hail producing thunderstorms.
Lightning data are taken from the LINET network. Radar information is obtained by radar MRL5-IRIS (wave length 10 cm) observations from Hail Suppression Agency in Bulgaria. The flash rate is calculated per 4 minutes in accordance with the period of radar volume scan.
Preliminary results indicate that in the studied thunderstorms FR increases to particular value of some radar parameters and after that there is a decrease of FR.

Paper 64:

Electrification of thunderstorms developed over Bulgaria - combined analysis of radar data and numerical simulations by MésoNH

Rumjana Mitzeva, Faculty of Physics, University of Sofia, Bulgaria

Boryana Tsenova, National Institute of Meteorology and Hydrology, Bulgaria

Tsvetelina Dimitrova, Agency Hail Suppression, Bulgaria


Several studies show that there is an apparent relationship between lightning activity and microphysical and dynamical characteristics of thunderstorms. However, the conclusions based on the investigations conducted in different geographical regions are often contradictory. Results vary also depending on the type of thunderstorms (severe and non-severe). The present study is focused on the relationships between thundercloud microphysics and its electrification. Cases of severe hail producing thunderstorms that developed over Bulgaria are studied. Numerical simulations with the three-dimensional non-hydrostatic mesoscale model MésoNH are performed. Radar data (taken from MRL5-IRIS with wave length 10 cm) for the same cases is considered. A combined profound analysis of the numerical simulations and radar data will be presented.

Paper 65:

NowCastMIX – automatic integrated warnings from continuously monitored nowcasting systems based on a fuzzy-logic approach with optimized estimates of storm cell vectors

Paul James, Deutscher Wetterdienst, Germany

Sebastian Trepte, Deutscher Wetterdienst, Germany

Bernhard K. Reichert, Deutscher Wetterdienst, Germany

Dirk Heizenreder, Deutscher Wetterdienst, Germany


The German Weather Service’s AutoWARN system integrates various meteorological data and products in a warning decision support process, generating real-time warning proposals for assessment and possible modification by duty forecasters before dissemination to customers. On nowcasting timescales, several systems are continuously monitored. These include the radar-based storm-cell identification and tracking methods, KONRAD and CellMOS; 3D radar volume scans; precise lightning strike locations; the precipitation prediction system, RadVOR-OP, as well as synoptic reports and high resolution numerical forecast models. These provide a huge body of valuable data on rapidly developing mesoscale events. However, without additional pre-processing, the forecasters could become overwhelmed with information, especially during major convective outbreaks. NowCastMIX thus processes all available data together in an integrated grid-based analysis, providing an optimal warning solution with a 5-minute update cycle, combining inputs using a fuzzy logic approach for estimating likely storm attributes and deriving optimized estimates for the storm cell motion vectors. The resulting gridded warning fields provide regionally-optimized warning levels for differing thunderstorm severities which can be managed adequately by the duty forecasters. NowCastMIX thus delivers an ongoing real-time synthesis of the various input data to provide consolidated sets of most-probable short-term forecasts. The system has now run over two complete convective seasons, yielding a comprehensive, high resolution dataset of thunderstorm analyses and corresponding warning forecasts for the whole of Germany. This provides a valuable research resource for developing methods to improve quality. A verification of NowCastMIX forecasts against its own analyses shows that the predicted cell speeds for generating downstream warnings are already nearly optimal, but that newly developing cells tend to form somewhat rightwards of existing trajectories, in association with typical discrepancies between the vectors of individual cells and those of the larger-scale troughs in which they are embedded.

Paper 67:

The 2.5-minute Meteosat 10 rapid scan experiment and storm-top observations

Martin Setvák, Czech Hydrometeorological Institute, Prague, Czech Republic

Johannes Müller, EUMETSAT, Germany


During the commissioning of the Meteosat 10 (MSG-3) satellite, EUMETSAT has carried out an experimental 2.5-minute rapid scan test. According to the original commissioning plans for the Meteosat 10 satellite, a 3-minute rapid scan was initially scheduled, however after suggestion by the EUMETSAT/ESSL Convection Working Group (CWG) it was decided to modify the test to shorter, 2.5-minute intervals. This interval is exactly half of the present 5-minute RSS (Rapid Scan Service), which is being operationally performed by the Meteosat 8 satellite; also it will be the native rapid scan interval of the future Meteosat Third Generation – Imager (MTG-I) satellites. The 2.5-minute rapid scan test started on 11 September 2012, and lasted exactly 24 hours.

Despite the relatively late time of the year for convective storms, several storms still occurred over various parts of Europe, and were captured by the Meteosat 10 satellite during this test. In the presentation some of these storms will be shown in detail, with emphasis on the big variability of the storm tops – namely their overshooting tops.

The shorter periodicity data such as the 2.5-minute rapid scan by Meteosat 10, or even shorter-period ones such as those from the U.S. GOES-14 satellite (http://cimss.ssec.wisc.edu/goes/srsor/GOES-14_SRSOR.html) can help improving our understanding of the overshooting tops, which are frequently used in nowcasting as one of the satellite-based indicators of possible storm severity.

Paper 69:

SCENE: convective cell identification and nowcasting using multi-source data

Anna Jurczyk, IMGW-PIB, Poland

Jan Szturc, IMGW-PIB, Poland

Katarzyna Ośródka, IMGW-PIB, Poland


The presented model SCENE (Storm Cell Evolution and Nowcasting) 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 system, and numerical weather prediction model employing fuzzy logic approach. For identified convective area cells’ centres are determined through analysis of radar precipitation 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 precipitation 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 cells’ centroid positions.

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 70:

Sounding-derived parameters and their ability to forecast individual severe weather threats for the region of central Europe.

Tomas Pucik, Masaryk University, Slovakia

Miroslav Kolar, Masaryk University, Czech Republic

David Ryva, Amateur meteorological society (Skywarn Czechoslovakia), Czech Republic


Study is devoted to the analysis of more than 35 environmental parameters (indices) and their ability to discriminate between the intensity categories of individual severe weather types (hail, wind gusts, excessive precipitation). Intensity categories were delineated as non-severe, severe and extremely severe while severity for each of the events was assigned using the ESWD data. Environmental parameters (indices) have been processed from 1962 proximity soundings acquired for 4 years of thunderstorm activity (2008-2011) in the region of central Europe. Ability of parameters to discriminate between the intensity categories has been determined via non-parametric statistical tests (using p-values and test score values).

One of the primary findings is that the discriminatory ability of a certain parameter is often very different for individual severe weather types, another being that even parameters with the best scores exhibit at least some overlap between the values of individual intensity categories. Statistical results showed that the best discriminatory skills were found for the hail events, worst for the excessive precipitation. Moreover, we found that it is easier to discriminate non-severe and severe events than severe and extremely severe ones.

Severe hail environments were well predicted using the buoyancy-related parameters, while severe wind gust environments were more sensitive to the vertical wind shear. Simple combinations of buoyancy and wind shear proved as the best predictors for both severe weather types. In case of the excessive precipitation, results were different with a tendency for severe events to occur in the environments favoring high precipitation rates / efficiency. Also, tornado environments were examined with the finding that even in the favorable range of parameters, non-tornadic cases dominate.

Generally speaking, primary goal is an assessment of the severe weather predictability using simple, sounding derived parameters and the outline of main limitations that stem from using them, especially regarding the parameter “threshold” values.

Paper 71:

Wide-spread severe convective storm events in Bulgaria (1991-2010)

Liliya Bocheva, National Institute of Meteorology and Hydrology, BAS, Bulgaria

Ilian Gospodinov, National Institute of Meteorology and Hydrology, BAS, Bulgaria

Petio Simeonov, National Institute of Meteorology and Hydrology, Bulgaria

Tania Marinova, National Institute of Meteorology and Hydrology - BAS, Bulgaria


The objective of this work is to give an overview of the spatial and temporal distribution of the occurrence of severe convective storms (SCS) simultaneously in a large part of the country. The covered period is 1991-2010. Only days in which there is thunderstorm activity combined with 24-hour precipitation amount above 30 mm in at least 4 out of 27 administrative regions of the country are selected and analyzed. The choice of 30 mm/24hours as a bottom limit is motivated by the fact that for most of the stations in Bulgaria it presents at least 30% of the mean monthly precipitation amount. These massive SCS events have become more frequent during the last 10 years. Most of them have also been documented as being the cause of floods and economic loss in the affected regions. They are more often found in Central and East Bulgaria than in the western part of the country. About 80% of them occur in the second half of summer and the first half of autumn. They are associated with stationary frontal systems and Mediterranean cyclones. Analysis of the available radiosounding data is given in support.

Paper 72:

Radar-based hail climatology for the Czech territory

Katerina Skripnikova, Institute of Atmospheric Physics AS CR, Czech Republic

Daniela Rezacova, Institute of Atmospheric Physics AS CR, Czech Republic


Severe hailstorms are rare, but can cause considerable damages in Czechia. Because of the time and space variability, the exact detection of hail from ground measurement is almost impossible. It is problematic to rely on the ground station data when creating hail climatology. An alternate approach uses information from weather radars, which provide information with high temporal and spatial resolution. In Czechia, two C-band Doppler radars are operated by the Czech Hydrometeorological Institute and cover with measurements the whole territory. Several algorithms for hail detection using single-polarization radar data were tested for use on the Czech territory. New “combi” criterion was proposed based on a combination of previously tested techniques. Because of the data used for criteria testing, principally large hail is detected. The “combi” criterion is applied to the radar data from summer months (May-August) of the period 2007-2012. Hail risk distribution and hail climatology of Czechia based on these data is formed. Severe hail distribution over the Czech territory and the dependence on orography is discussed.

Paper 73:

Regionalization of cloud-to-ground lightning in Iberia

João A. Santos, UTAD, Portugal

Margarida Reis, UTAD, Portugal

Fernando De Pablo, University of Salamanca, Spain

Luís Rivas-Soriano, University of Salamanca, Spain

Solange Leite, UTAD, Portugal


The present study deals with temporal and spatial variability of cloud-to-ground discharges covering the Iberian Peninsula, the Balearic Islands and nearby seas (36-44ºN; 10ºW-5ºE) during the period 2003-2009, based on a 17 sensors network. Hourly, monthly and seasonal variability of cloud-to-ground discharges, peak current intensity, polarity and multiplicity are first analyzed. Secondly, daily mean, maximum and minimum 2 m air temperature, precipitation and altitude are selected as forcing factors of cloud-to-ground discharges in an Iberian and regional spatial scale. High monthly, seasonal and regional correlations between these forcing factors and discharge patterns were identified. CGD occurrences are mainly located over high-altitude areas during summer, while they tend to occur over sea in autumn. CGD patterns are controlled by temperature fields and maximum (minimum) temperatures tend to lead the correlations in the warmer (cooler) months. A clear correspondence between the precipitation patterns and CGD is also found when precipitation is mostly convective (summer half of the year).

Paper 74:

Modeling of cloud-to-ground lightning activity in Portugal

João Sousa, IPMA, Portugal

João A. Santos, UTAD, Portugal

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


Dynamically coherent lightning regimes in Portugal are identified and a statistical-dynamical logistic modeling of the daily discharges over the country is implemented. A dataset of cloud-to-ground discharges over Portugal, collected by the Portuguese lightning detection network in the period of 2003-2009, is used. The high-resolution MERRA reanalysis is also used for the analysis of the atmospheric dynamics during each regime. Three lightning regimes are identified, two over a wintertime period (October-April), WREG and WREM, and one over a summertime period (May-September), SREG. The WREG is characterized by a low pressure system westwards of the Iberian Peninsula, with southerly and westerly winds over Portugal, and shows stronger lightning activity over the southern half of the country. The WREM presents a strong trough over Portugal, accompanied by southwesterly flow over the country and reveals a relatively spread pattern of lightning activity. The SREG is related to an inverted low pressure trough over Portugal, which generates air parcel rising and hints at high lightning activity over more inland areas. For the statistical-dynamical approach, based on logistic regressions, two classes of daily lightning activity are used in each regime, i.e. values below the median (weak activity) and above this threshold (strong activity). Some meteorological variables are used as predictors: equivalent potential temperature, specific humidity, relative humidity, potential vorticity, lifted index and moist static stability index at 850, 700, 500 and 300 hPa. Although the best predictor for the three regimes is the difference of the equivalent potential temperature in the 700-500 hPa layer, the best 4-layer lifted index is also significant for the three regimes. These models present an efficiency of 65-70% and can be of great value for weather forecasting in Portugal.

Paper 75:

Re-analysis of a high-end severe weather event on 18 August 1986.

Tomas Pucik, Masaryk University, Slovakia

Petr Zacharov, Institute of Atmospheric physics, Czech Republic

Martin Setvak, Czech Hydrometeorological Institute, Czech Republic

Zbynek Sokol, Institute of Atmospheric physics, Czech Republic

Zbynek Cernoch, Amateur Meteorological Society, Czech Republic


On this date, an extraordinary severe weather event was observed over the Czech Republic and some other countries (e.g. Germany or Northern Italy) featuring intense hailstorms. In case of the Czech Republic, a single supercellular thunderstorm managed to produce a 230 km long hail swath with many reports of very large hail (up to 12 cm in diameter) and severe wind gusts. This resulted in an extreme damage with some settlements losing every single roof, and also several casualties.

Even though a relatively thorough analysis has been done shortly after the event (Pavlik - Kakos - Strachota, 1988), little has been known in those times about the supercell or large hail prediction. Thus, we would like to shed more light on the environmental conditions (using the ingredient-based methodology), which allowed for perhaps the most intense hailstorm in the modern history of the Czech Republic. Large scale conditions will be assessed using the ECMWF reanalysis data with the horizontal resolution of 0.5°. More regional and detailed analysis will be based on the run of COSMO model featuring 50 vertical levels and a horizontal resolution of 2.8 km.

This case became well-known also from the satellite perspective, because hailstorm exhibited very significant cold-U feature with a "plume" of cirrus extending over its "warm" interior part (Setvak - Doswell, 1991). In the light of a new research that has been done on the storm-top characteristics, we return to this case and discuss its significance.

Eventually, this study should mostly serve as a re-analysis of the event from the forecaster point of view, contrasting the low and high-resolution model output and their ability to reproduce severe storm environment. We also point to the nowcasting possibilities (spatial-temporal resolution of the data) that were available at that time versus the current situation.

Paper 76:

Simulation of the hail storms with the COSMO NWP model

Zbynek Sokol, Institute of Atmospheric Physics ASCR, Czech Republic

Vojtech Bliznak, Institute of Atmospheric Physics ASCR, Czech Republic

Petr Zacharov, Institute of Atmospheric Physics ASCR, Czech Republic


Simulations of several observed convective storms accompanied by hail with the COSMO NWP model will be presented. The model will be integrated with a horizontal resolution from 1 km to 2.8 km covering the domain of the Czech Republic and with the assimilation of observed radar reflectivity and satellite data. The model will use one-moment and two-moment microphysics. The aim of our study is to assess on which conditions the model is able to simulate hail occurrence 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 applied microphysics (one-moment and two-moment).

Paper 77:

Analysis of the 4th October 2007 squall line predictability by means of an Ensemble Kalman Filter

Victor Homar Santaner, Universitat de les Illes Balears, Spain

Lorena Garcies Artigues, Universitat de les Illes Balears, Spain

David Stensrud, National Severe Storms Laboratory, United States

Louis Wicker, National Severe Storms Laboratory, United States


The western Mediterranean is a region climatologically prone for the development of high impact weather events. On 4th October 2007, a severe squall line formed southwest of the Western Mediterranean basin and swept across the Balearic Islands, producing extensive damage and one dead. The prediction of such events is a great challenge for current operational offices due to the fundamental imbalance between degrees of freedom in the forecasting system (i.e. high spatial resolution) and the operationally available atmospheric information in the area (i.e. very low density of in-situ observations over the sea). We assess the value of an Ensemble Kalman Filter for the initialization of a mesoscale ensemble prediction system by comparing different configurations, including the basic downscaling from the operational ECMWF EPS predictions on a case of a squall line ocurred over the Western Mediterranean on 4th October 2007. The results from these experiments will be transferred to the field activities within the HyMeX program context. Results suggest that the use of EnKF is significantly beneffiting the accuracy of the probabilistic predictions as the filter transfers information from over land towards maritime areas. The particular effects of this proces are discussed.

Paper 78:

Statistical model of convection

Nazim Huseynov, National Academy of Aviation, Azerbaijan

Bahruz Malikov, National Academy of Avation, Azerbaijan


In this article the instability indices and some numerical prognostic methods that used for thuderstorm foercast at the airdrome Heydar Aliyev (Baku, Azerbaijan) were assesed and verified. For analyses the numerical methods of Sosin, Beil, Reshetov, and instability indeces Li, Si, TT, CAPE etc. were considered.
In this connection statistical analyses was made on thermodynamic parameters of the atmosphere during instability at the airdrome Heydar Aliyev. The statistical investigation was made with the purpose to develop convection model.
Model was developed on the basis of regularity of spatio-temporal correlation of meteorological element and parameters in instability weather conditions at the Heydar Aliyev airdrome and on flight route in the territory of Azerbaijan Republic. Obviously the verification of the model is higher if it’s including more statistical predictors. The model was developed on the basis of empirical probability of convective parameters vertical distribution.
For the development of the model of the cloud formation processes climatological conditions was investigated.
The physical process development was objectively assessed with the evaluation of potential energy of instability [31, 135, 148, 153]. The following parameters were considered as the criteria of instability: CAPE, vertical distribution of RH, moister content of air on isobaric levels (qm), depth of instability layer (∆h).
In the calculations radiosounding data for the period of 2004-2012 of the aerological station Mashtaga (WMO code 37860) was used.
The methodology of the development of the model included next steps:
 systematization of raw data;
 calculation of parametric values of diagnostic parameters of convection;
 correlation analyses of main values of diagnostic parameters, and evaluation of predictors;
 definition of the calculation method according to the type of the task;
 definition of the informative parameters for the model.
Statistical calculation was realized with the help of the “STATSOFT” companies software “STATISTICA”.
1) Statistic parameters of the instability conditions in atmosphere has been defined;
2) Parametric values of meteorological elements was calculated;
3) Statistic model for forecasting of storm clouds at the airdrome Heydar Aliyev was developed on the basis of discriminant analyses.

1. Бочарников Н.В. Начальные условия для оперативной модели конвективного облака. – В. Сб.: Радиолокационная метеорология. – Л.: Гидрометеоиздат, 1984, с. 50-53.
2. Брукс К., Краузерс Н. Применение статистических методов в метеорологии. – Л: Гидрометеоиздат, 1963, 416 с.
3. Гусейнов Н.Ш. Вертикальное распределение метеорологических элементов в пограничном слое атмосферы // Труды РГГМУ, 2006, №2, с. 79-85.
4. Гусейнов Н.Ш. Применение вероятностной логики и теории статистических решений для альтернативного прогноза гроз // Научные труды Национальной Академии Авиации, Том 5, № 3, 2003, том 5, № 3, с. 20-23.
5. Шметер С.М. Термодинамика и физика конвективных облаков. — Л.: Гидрометеоиздат, 1987, 288 с.
6. Шметер С.М. Физика конвективных облаков.— Л.: Гидрометеоиздат, 1972, 232 с.
7. Doswell C., Schutz D., 2006: On the use of indices and parameters in forecasting severe storms. Electronic J. Severe Storms Meteor., 1(3), 1-22.
8. Huseynov N.Sh., Malikov B.M. Synoptic situations generating the convective processes over the airdromes of Azerbaijan Republic / 10th Plinius Conference on Mediterranean Storms European Geosciences Union Tropical Conference Series, Nicosia, Cyprus, 22-24 September 2008, pp.53-54.
9. Huseynov N.Sh., Malikov B.M. Statistical analysis of the storms event at the airdrome Baku. Modeling of the accuracy isolines of the storms registration systems. // Atmospheric Research vol. 93, pp. 93-99.
10. Huseynov N.Sh., Malikov B.M. Regularity of distribution of precipitations on the airports of Azerbaijan Republic // Scientific Journal Advances in Geosciences (An Open Access Journal for Refereed Proceedings and Special Publications), Vol. 20, 2009, p. 9-12.

Paper 79:

Understanding the effects of horizontal convective rolls on the organization of low-level vorticity in simulated supercell thunderstorms.

Christopher Nowotarski, Pennsylvania State University, United States

Paul Markowski, Pennsylvania State University, United States

Yvette Richardson, Pennsylvania State University, United States

George Bryan, National Center for Atmospheric Research, United States


In previous work using idealized numerical simulations of supercell thunderstorms, the authors and others have found changes in the arrangement of low-level vorticity in these storms when their environment features a convective boundary layer. Most notably, heterogeneity associated with boundary layer convection may disrupt the organization of the low-level mesocyclone while accelerating the development of misovortices along the rear-flank gust front. In an effort to better quantify these effects and understand the physical processes by which they occur, we present analysis of trends in circulation about the low-level mesocyclone, vorticity budgets along parcel trajectories, and a simplified density current simulation. Sensitivity of the results to the orientation of the HCRs relative to storm motion is also discussed.

Paper 80:

The derecho of 8 August 2010 in Estonia

Aarne Männik, University of Tartu, Estonia

Helve Meitern, Estonian Meteorological and Hydrological Institute, Estonia


Summer thunderstorms are common in Estonia, but those which reach the levels of severe thunderstorm are quite rare. In the afternoon of August 8, 2010, violent straight-line thunderstorm was moving with a speed of 75 km/h across Latvia and Estonia and reached Finland. It was later identified as a derecho. This presentation provides an overview of the weather conditions before derecho formation; thunderstorm development and progress in Estonia. The synoptic conditions appear to be sympthomatic for formation of the severe storm in the region. Weather stations on the way of the derecho recorded wind gusts of about 23,4...36,5 m/s and the derecho reached its maximum destructive force moving over Eastern Estonia. All storm damages together were estimated to reach near 2 Million Euros in Estonia alone. An interesting feature of the situation was a coincidence of severe thunderstorm formation together with the presence of smoke from heavy forest fires in neighbouring Russia. It may be speculated that the aerosol from forest fires may have influenced the formation of the storm and is a good opportunity to study aerosol and severe convection interactions.
The presentation documents information from observations including remote sensing and analyses the synoptic situation for the purpose of following modelling studies. Overview of occurred damages is presented.

Paper 81:

Sounding-derived parameters associated with tornado occurrence in Poland and universal tornadic index

Mateusz Taszarek, Adam Mickiewicz University, Poland


Study is mainly devoted to operational meteorology, to improve tornado forecasts in Poland and create a universal tornadic index formula. A study is focusing on climatology of sounding-derived parameters associated with tornadoes in Poland and their potential value for tornado forecasting. The data was collected from soundings made in 10 stations in and around Poland which were closely in time and space connected with tornado occurrence. The main aim of the study was to analyze the thermodynamic and kinematic parameters and formulate an index. The information about tornado incidents was taken from media reports and the European Severe Weather Database for the years 1977-2012. Total of 97 tornado cases were divided according to their strength for significant (F2/ F3), weak (F0/F1) and unrated cases, and also according to their environmental surface temperature, for warm (>18oC) and cold (<18oC) tornadoes. As it turned out, depending on the temperature, tornadoes tended to present different environmental conditions for tornadogenesis. In warm cases, the most important factor was instability while for cold cases it was wind shear. It was also proven that significant tornadoes in Poland occur in conditions accompanied by high moisture content, moderate instability and dynamic wind field. The results of this study were used to create a universal tornadic index designed to forecast activity in warm and cold, and weak and strong tornadic environments. The quality of this index was tested for the period with increased tornado activity in Poland from 2008 to 2010.

Paper 82:


Diego Christian Araneo, IANIGLA, Argentina

Silvia Carmen Simonelli, PRM, IANIGLA-CONICET, Argentina

Federico Augusto Norte, PRM, IANIGLA-CONICET, Argentina

Jorge Ruben Santos, ICB-UNCuyo, Argentina

Carlos Bustos, DACC-Mendoza, Argentina


The objective of this work is to determine the vertical profile patterns of temperature (T) and dew point (Td) for the Northern of Mendoza Province and to obtain a statistical forecast model of convection using these profiles as predictors.
Rawinsondes at 12 UTC from Mendoza-Aero weather station (National Weather Service) were used in order to obtain the profiles and to adjust the forecast model. Several 1987-2005 October-March periods were used for the analysis and the 2006-2010 November-March periods were used to validate the predictions. We defined a Convection Occurrence Index (C) according to the reported observations during 24 hours, starting at 9 UTC before the rawinsonde observation.
Principal Component Analysis (PCA) was applied in order to determine the vertical profiles patterns of T and Td. In order to build the forecast model, the resulting component loadings were used as predictors in a logistic multiple-regression fit with the C index as response variable. The helicity and 850-400 hPa shear were also incorporated as predictors in the logistic multiple-regression fit.
PCA resulted useful to obtain the vertical profile patterns of T and Td. The resulting patterns in direct and reverse modes represent real cases. The analysis of T and Td sheds 6 significant components explaining 92% of the system variability.
In the studied cases, over than 70% of the forecast model effectiveness is obtained using only the T and Td profiles (similar to that found in other studies by traditional indices). The effectiveness could be improved by changing the definition of C and the incorporation of shear and helicity as predictors.
Profile patterns of T and Td reveals that the probability of convection increases (decreases) with strong (weak) T lapse-rate between low and middle levels of the troposphere and high (low) moisture content in the lower layers.

Paper 83:

Impact assessment of thunderstorms in Portugal in the 2003-2011 period: overall characterization and analysis of a noteworthy case study in Lisbon area


This study is focused on the assessment of material and human damage caused by thunderstorms in Portugal. A systematic search over the 2003-2011 period on the web archived material by 19 mass media national sources, complemented by the available online published news from 47 regional newspapers, allowed to gather information related with the impacts of thunderstorms in Portugal, including accidents affecting living beings (people or animal deaths, injured) and occurrences of material damages in several structures (e.g. buildings, vehicles, social equipments) or forested (wildfires) and agricultural areas. All selected occurrences in this (Access) database are related with lightning and hail damage. The study period starts in 2003 because is the first year with available data on lightning activity, collected by the Portuguese lightning network, providing instrumental observations of cloud-to-ground discharges to check the origin of the accidents. Temporal and spatial variability of the occurrences are analysed applying GIS (geographical information system) tools. A total of 234 accidents were registered in the database, occurring in 111 days with thunderstorm damage (DTD), which means an average frequency of 12 DTD/year. Hail related occurrences accounts for 30% of the total of the accidents, being 70% caused by lightning. The case study of the severe hailstorm occurred in the Lisbon area in the 29th April 2011 was selected to carry out a detailed survey of the produced damages during this extreme event. This case illustrates the complex diversity of impacts triggered by a severe hailstorm in this urban area resulting in a very harmful and costly natural disaster, responsible for inundations in several locations and disruptions on public and private services.

Paper 84:

Overshooting top: Physics and Dynamics

Pao Wang, University of Wisconsin-Madison, United States


Overshooting top is an important feature of thunderstorms and a subject of increasing studies recently, especially on the aspect of satellite observations. To correctly interpret the observations, however, requires a sound understanding of the physics and dynamics of the overshooting top, what it is and what it is not. This presentation will utilize a physics-based cloud model to simulate the formation of overshooting top of typical thunderstorms, and use the model results to illustrate the physical and dynamical processes of the formation of the overshooting top. Thermodynamic structure (temperature, theta and theta-e) will be examined in great detail. Cloud microphysical structure will be presented and compared with CloudSat observations. Implications of the sighting of overshooting tops on the stratosphere/troposphere exchange will be discussed.

Paper 85:

Vapor volume reduction - an additional factor that contributes to the central low pressure in a cyclone

Dhananjay Mardhekar, No organization - Independent researcher, India


Tropical cyclones are central low pressure systems with torrential rains in the eye-wall. The discovered phenomenon explains an additional factor that contributes to this central low pressure. According to Avogadro’s law 18 grams of water occupies 22.414 liters in vapor state at standard temperature and pressure (STP). i.e., 1.0 ml water occupies 1245 ml vapor. Conversely, 1245 ml vapor at STP when condenses forms 1.0 ml water. Latent heat is released. Due to the phase change, from vapor to liquid, this is instantaneous; enormous and instantaneous reduction in volume takes place. 1245 ml vapor (STP value) instantaneously diminishes to meagre 1.0 ml water. In cyclone continuous torrential rains implies continuous condensation, therefore continuous and instantaneous vapor volume reduction leading to low pressure zone formation, continuous formation of sudden pressure gradient forces, continuous additional suction of vapor-rich air, the vapor part again condenses. Thus, continuously forming sudden local low pressure zones contribute continuously to the central low pressures. If the vapor volume reduction phenomenon was not to occur, then the central low pressure formed would not be enough to maintain the observed wind speeds. Vapor continuously enters the cyclonic system with a very large volume and comes out of the system on condensation with a very small volume obviously forming low pressures. Rains go out of the system but the instantaneous empty spaces formed remain in the system. Thus, in a cyclone condensation of vapor provides two things simultaneously: - 1) it provides sudden low pressures; 2) it provides latent heat. Since cyclonic engine is well organized, we can say that in a cyclone these two different processes work together in a well coordinated manner. Continuous condensation and the concomitant vapor volume reduction phenomena can be applied to tornado dynamics explaining the initiation and maintenance of the air flow.

Paper 86:

Derecho-like event in Bulgaria on 20 July 2011

Ilian Gospodinov, BAS, National Institute of Meteorology and Hydrology, Bulgaria

Tsvetelina Dimitrova, Agency Hail Suppression, Bulgaria

Lilia Bocheva, BAS, National Institute of Meteorology and Hydrology, Bulgaria

Petio Simeonov, BAS, National Institute of Meteorology and Hydrology, Bulgaria

Rumen Dimitrov, Agency Hail Suppression, Bulgaria


Derecho events are rare in Bulgaria. Only one such event has been registered there within the last 5 years. The Bulgarian hail suppression agency (BHSA) has been equipped with state-of-the-art meteorological Doppler radars since 2008. The available data allows more thorough analysis of the evolution of big-impact convective clouds occurring in the vicinity of the radar location. In this work we present a severe convective storm with derecho event that occurred on 20 July 2011 in the northwestern region of Bulgaria. The heaviest damage has been found in the municipality of Knezha. We present an overview of the damage in the region. The weather station there reported strong thunderstorm with hail and 41 mm of rainfall for about an hour. We present analysis of the evolution of the convective cloud based on the radar data provided by BHSA. The Doppler radar data revealed wind speed above 30 m/s, cloud top height of 17 km, and maximum radar reflectivity factor of 63 dBZ. Analysis of the available radiosounding data is also given in support. The storm is associated with a rapid cold front passing through in midday. The field investigation carried out by the local stuff of the BHSA revealed derecho-like pattern in the debris.

Paper 87:

Lay severe weather competence - A pilot study on Brazil, India, and Germany

Alexander G. Keul, Salzburg University, Austria

Luci Hidalgo Nunes, UNICAMP Campinas, Brazil

Maria Luiza De Andrade Benini, UNICAMP Campinas, Brazil

Sanjay Sharma, Kohima Science College, India

Devajyoti Dutta, Kohima Science College, India

Melanie Korff, Salzburg University, Austria


Weather knowledge, interest, and risk assessment of laypeople is a key element of severe weather warning quality management. As a pilot study for an international data base, the situation was assessed in Brazil, India, and Germany. Questionnaires covered weather (report) interest/sources/legibility, basic weather knowledge, subjective risk assessment, preparedness, self-reported behavior, physical damage by weather events, and sociodemographic data. Brazilian data (104) came from the relatively flat area of Campinas northeast of the Sao Paulo metropolis. The Indian sample (100) was collected in northeast, mountainous Nagaland. A German survey (80) chose the Rosenheim, Bavaria, foothills north of the Alps. Mean ages were 35.8 (Ind), 37.2 (Brz) and 38.8 (Ger); the samples were gender-balanced, but had an education bias.
Weather interest was above 50% in all three countries (Ger 80%). Interest in weather reports varied (Ind 31%, Brz & Ger 53%). The main sources were TV/newspapers in India, TV/internet in Brazil, and internet/radio/TV in Germany. Report legibility was rated medium by all samples. Subjective risk assessment identified landslides as Indian top-feared risk, floods in Brazil, tornadoes in Germany. Meteorological lay knowledge was low in India, medium in Brazil, high in Germany. Severe weather information (good: 25% Ind, 14% Brz, 50% Ger) and preparation (good: 31% Ind, 7% Brz, 24% Ger) show room for improvement. Physical damage events were sparse in India (11% storm, 9% flood), medium in Brazil (30% storm, 28% lightning) and in Germany (41% storm, 29% flood, 23% lightning).
Meteorological interest or behavior did not correlate with education in Brazil and Germany. In India, higher educated people were more weather-interested. In all three countries, weather interest and weather report interest were statistically related. In flood-risky Brazil, estimated flood risk correlated with local weather information and preparedness. Physical weather-related damage did not produce attention or attribution effects.

Paper 88:

Can lightning really hit you? Risk, damage and preparedness surveys in India, Brazil, and Austria

Alexander G. Keul, Salzburg University, Austria

Sanjay Sharma, Kohima Science College, India

Luci Hidalgo Nunes, UNICAMP Campinas, Brazil


Cloud-ground lightning shows geographical hotspots in the Americas, Africa, India and Indonesia (Christian et al., 2003). Does the objective risk level shape lightning interest, knowledge and preparedness of the local population? Will education and physical damage play a role? Three areas of India, Brazil and Austria are compared. The Brazilian and Indian regions show OTD lightning density 10-20 flashes/km/year, Austria 2-4. In questionnaire surveys, 104 (54% high education) responded at Campinas, Sao Paulo province, Brazil, 100 (66% high education) at Nagaland province, India, and 133 (10% high education) from the Salzburg and Upper Austria provinces of Austria.
In a correlation analysis, formal education levels showed no significant outcome in Brazil. In India, more educated residents had higher weather interest and assumed a higher lightning risk. In Austria, higher education even lowered the search for information.
Actual lightning strikes at respondent’s homes in Brazil (reported by 28%) went with higher lightning information scores (i.e. stimulated information seeking), whereas in India (4% strikes) and Austria (12% strikes), no correlations were present. In Brazil, raised weather interest corresponded with higher awareness that lightning could hit the house. Also, reported preparedness in Brazil correlated with the assumed lightning risk and the lightning information score, and in India with weather interest, weather report attention and the lightning information score. In Austria, higher lightning information scores paradoxically meant a lower expected lightning risk, whereas the subjective lightning risk rose with weather report attention and wish for more information.
It is concluded that a) a high CG flash rate influences population risk parameters when it results in physical damage, b) formal education is no predictive factor as lightning is no school topic, and c) emotions stimulating attention and interest should be activated when providing public lightning information.

Paper 89:

First experiments with the „Rapid Development Thunderstorm“ algorithm

Vojtech Bliznak, Institute of Atmospheric Physics AS CR, Czech Republic

Zbynek Sokol, Institute of Atmospheric Physics ASCR, Czech Republic


The contribution will present the first experiments with the “Rapid Development Thunderstorm” (RDT) algorithm, which is applied over the territory of the Czech Republic (CR). The main goal of the RDT algorithm is to identify, monitor and track intense convective clouds using data derived from the satellite observations of Meteosat Second Generation (MSG) every 15 minutes. The RDT algorithm and corresponding software has been developed by French meteorological service Météo-France in the framework of the EUMETSAT Satellite Application Facilities (SAF) in support to nowcasting.
We will present a modified algorithm, which will be adapted to the local data. The verification of the algorithm will be performed subjectively as well as objectively including comparison with lightning data used as the ground truth. The main motivation of this work is to prepare suitable information for the assimilation of observed convective cells including their extrapolated trajectories into the numerical weather prediction model. This approach, which has already been tested with radar data, should improve the accuracy of precipitation forecasts.

Paper 90:

Relationships between the cyclones of southern origin and thunderstorms in Estonia

Kaupo Mändla, University of Tartu, Estonia

Sven-Erik Enno, University of Tartu, Estonia


This study presents relationships between the frequency and duration of southern cyclones and thunderstorms over Estonia during the period 1950-2010. Cyclones of southern origin were filtered out from the database of cyclones in the Northern Hemisphere. In this study, “southern cyclones” are defined as lows that have formed south of 47°N, east of the 0° meridian and west of 60°E, and have entered into the circle of 1000-km radius around Estonia. 557 southern cyclones were detected during the 61 years under observation. Based on the data of five meteorological stations, 1039 days with thunderstorms were detected in Estonia during this time. Only 10.4% of all observed thunder days were associated with southern cyclones. At the same time, 28.5% of all southern cyclones were accompanied by thunderstorms in Estonia. Thereby, 40.6% of all southern cyclones induced a thunderstorm during the thunder season from April to October. The highest activity of thunderstorms was detected in summer months when 60-80% of southern cyclones caused thunder. 99% of all thunderstorms induced by southern cyclones appeared during the thunder season. The
largest number of thunder days was detected when southern cyclones passed a station at a closer distance than 500 km. At two stations, the annual average number of thunderstorms induced by southern cyclones has increased in the cases when the cyclones passed Estonia from the east within the distance up to 500 km. In the cases when the cyclones passed Estonia from the west at a longer distance than 500 km, the annual length of a thunderstorm increased by 1.5 hours. The number of cloud-to-ground lightning strikes, induced by southern cyclones, was larger than that of any other thunder events. Therefore, the intensity of southern cyclones-related thunderstorms is higher than that of other thunderstorms.

Paper 91:

High-resolution climatology of lightning in Central Europe

Kathrin Wapler, Deutscher Wetterdienst, Germany


Lightning pose a significant threat to life, property and economy. Hence, the detailed knowledge of the occurrence of lightning is important. A high-resolution climatology allows assessing the local risk of lightning. A 5-year analysis (including data of >30 million strokes measured by the LIghtning detection NETwork LINET) of the spatial and temporal occurrence of lightning in Germany and neighbouring areas is presented. The analysis on a high-resolution grid with spatial resolution of 1 km enables identifying local features, e.g. this resolution is high enough to identify TV towers which trigger lightning. The data set allows studying local effects, e.g. the influence of orography on the occurrence of thunderstorms. The analysis reveals spatial and temporal patterns. The highest numbers of lightning strokes occur in the pre-alpine region of southern Germany; further local maxima exist in low mountain ranges. The lowest number of lightning is present in areas of the North Sea and Baltic Sea.
Despite a high year-to-year variability of lightning rates, on average a clear annual cycle (maximum June to August) and diurnal cycle (maximum in the afternoon) is present. Additionally to this well-known annual and diurnal pattern, the data show that those are intertwined: the diurnal cycle has an annual cycle, visible in the time of daily maximum which occurs later in the afternoon in summer compared to spring and autumn. Furthermore an annual cycle of mean IC height, i.e. rising IC height during the year with a maximum in late summer, is shown. Additionally, the statistics are shown in dependence of synoptical situations. This reveals environments which favour the development of thunderstorm in different areas.

Paper 92:

Analysis of lightning flash characteristics in Central Europe

Kathrin Wapler, Deutscher Wetterdienst, Germany

Christopher Frank, University of Bonn, Germany


Lightning pose a significant threat to life, property and economy. Hence, the detailed knowledge of the occurrence of lightning and its characteristics is important. A better knowledge of lightning characteristics helps understanding the effects responsible for cloud electrification. An improved understanding will lead to better detection and forecasting of lightning.
The characteristics of lightning in Germany and neighbouring areas measured by the LIghtning detection NETwork LINET are presented. Especially, the following stroke attributes are studied: type (cloud-to-ground (CG) and intra-cloud (IC)), height, polarity, amplitude. Several statistics of these stroke attributes are presented. Only a few are mentioned here: the amplitude show higher average values of CG (cloud-to-ground) versus IC (intra-cloud) strokes; the IC height has a Gaussian distribution with an average around 8 km.
Furthermore, the multiplicity of lightning flashes (the number of strokes within a flash) is analysed along with the characteristics of the strokes within a flash. These analyses reveal interesting flash characteristics, e.g. the average amplitude of all strokes within a flash of high multiplicity is larger than the average stroke amplitude of flashes with low multiplicity.

Paper 93:

Lagrangian diagnostics of convective cells using combined satellite, lightning and radar observations

Akos Horvath, Leibniz Institute for Tropospheric Research, Germany

Kathrin Wapler, Deutscher Wetterdienst, Germany

Fabian Senf, Leibniz Institute for Tropospheric Research, Germany

Hartwig Deneke, Leibniz Institute for Tropospheric Research, Germany


The nowcasting of severe convective events remains a challenging endeavour that suffers from relatively low skill and high false alarm rates. In order to further our understanding of severe weather systems, we have performed a Lagrangian analysis of precipitation cells over Germany combining geostationary satellite, ground-based lightning and radar observations. The radar signature of tracked precipitation cells has been combined with cloud macro- and microphysical retrievals from Meteosat-9 rapid scan data, provided by the EUMETSAT Climate Monitoring and Nowcasting/Very Short-Range Forecasting Satellite Application Facilities and lightning measurements from the Lightning detection NETwork LINET. The life cycle of radar reflectivity, rain rate, cell size, cloud-top height, cloud optical thickness, drop effective radius, cloud water, lightning rate, and lightning rate tendency has been analyzed for a large number of tracks. The mean time evolution of convective systems has then been characterized by synchronizing individual events relative to various measures of peak intensity (maximum reflectivity, maximum lightning rate tendency, maximum effective radius, etc.). Interesting characteristics are seen in the analysis: e.g. decreasing effective radius along with increasing cloud optical thickness is a precursor of further electrification and increasing flash rate; strong trends of lightning rates are more often followed by high reflectivity and thus an increased chance of hail.

Paper 94:

A modeling study to evaluate the silver iodide dispersion from ground generators seeding flares

Enrique Puliafito, Grupo de Estudios de la Atmósfera y el Ambiente, Argentina

David Allende, Grupo de Estudios de la Atmósfera y el Ambienta UTN-FRM, Argentina

Pablo Cremades, Grupo de Estudios de la Atmósfera y el Ambienta UTN-FRM, Argentina

Celeste Mulena, Grupo de Estudios de la Atmósfera y el Ambienta UTN-FRM, Argentina


Annually an average of 10% of the agricultural production in the Mendoza Province in western Argentina is lost as a result of severe hailstorms, and approximately 20% of the agricultural area receives hail precipitation. Operational hail suppression programs have been based in Mendoza since 1970’, designed to lessen the economic impact in the vineyards. Lately cloud seeding program have been done with planes and ground generators. This modeling study is designed to assess the evidence of silver iodide dispersion in order to evaluate if potential clouds are reached and secondarily to optimize the location of such generators.
In that sense, 3-D simulations with the version 3.4 of the Weather Research and Forecasting model with Chemistry (WRF/Chem) are conducted for several typical hailstorms, which peak activity commonly occurs near sunset (approximately 0000 UTC, or 2100 local time), and storms lasting past local midnight. Three nested modeling domains are set to downscale the physical properties of the atmosphere in the predominant storm generation zones. The model physical and chemical parameterizations were previously tested and found to produce the more accurate estimation of temperature, humidity and wind fields at surface and upper-air levels. The emission sources are properly characterized with emission rates, temperatures and velocities from direct measurements on the generators.
Finally, silver iodide mass concentrations at different heights and dry deposition rates are estimated with the modeling system for the entire modeling domain.
Results of the model are compared to TITAN radar echo reflectivity’s.

Paper 95:

Characteristics of lightning behavior in hailstorms and hail properties in storms with overshooting tops

Natasa Strelec Mahovic, Meteorological and Hydrological Service, Croatia

Petra Mikus, Meteorological and Hydrological Service, Croatia

Damir Pocakal, Meteorological and Hydrological Service, Croatia


Previous research showed that in most cases intensification of the storms’ updraft is accompanied by an increase in total flash rate. Rapidly growing flash rate increases the potential for severe weather, but doesn’t imply the development of severe weather phenomena itself. During severe thunderstorms CG lightning production seems to be decreasing, while significant increase in the number of IC flashes is registered. In the hailstorms, regions with a reduced number of lightning strokes can be detected. These regions are usually short lived and related to the wet hail growth which is not conducive to hydrometeor charging.

In this work lightning characteristics are studied for ~ 40 cases of hail-producing thunderstorms occurring over Croatia in summer months (May to August), from 2008 to 2012. Lightning distribution, registered by LINET Lightning Detection Network, was compared to hail occurrences, based on data collected from a hail-pad polygon located in NW Croatia. The polygon, consisting of 150 hailpads, covers the area of app. 600 km2 with approximate distance between hailpads of up to 2 km. The results will show spatial and temporal characteristics of lightning distribution and polarity as well as portion of cloud to ground (CG) and intra-cloud (IC) lightning before, during and after the time of hail occurrence.

Additionally, all studied hailstorms were checked for the appearance of the overshooting tops (OT) using a method based on SEVIRI brightness temperature differences and verified in high-resolution visible (HRV) images, where available. For cases with apparent OT, detailed analysis of lightning properties and physical properties of hailstones on the ground (number, diameter and kinetic energy) is made. In order to find possible typical differences, these parameters are compared with the same parameters in cases where OT was not apparent.

Paper 96:

Properties of embedded warm areas and surrounding cold rings and cold-Us

Michaela Radová, Czech Hydrometeorological Institute, Prague, Czech Republic

Jindřich Šťástka, Czech Hydrometeorological Institute, Prague, Czech Republic

Martin Setvák, Czech Hydrometeorological Institute, Prague, Czech Republic


Distinct long-lived embedded warm areas together with surrounding cold rings or cold-Us, which are observed atop some of deep convective storms in satellite imagery in IR window bands, belong among features that are not sufficiently explained yet. Importance of understanding of these features is given, among others, by their connection with occurrence of severe weather. Knowledge of their origin and understanding of their evolution are hence important not only for better interpretation of satellite data but also for nowcasting of the severe weather.

This study aims at mid-latitude convective storms which exhibit the above mentioned features. It mainly focuses on spatial properties of the embedded warm areas, cold rings and cold-Us and their time evolution during existence of the features, including also the transformation processes between the cold ring and cold-U. Further, we explore relationship of the spatial properties of these features with other meteorological variables and with the storm evolution. The study is based on IR window band data (especially the MSG SEVIRI Rapid Scan data), complemented by satellite data from other spectral bands, radar data and soundings.

Paper 97:

Downscaling past severe storms in Finland

Pauli Jokinen, Finnish Meteorological Institute, Finland


The study of past severe storms (deep moist convection) relies heavily on the availability of weather radar information. One can derive many important pieces of information from radar data, e.g. the storm path and affected areas, the timing, convective mode, potential hazards (hail, gusts, tornadoes, flooding and lightning). In Finland the digital radar archive is only available from late 1990s. Therefore it is challenging to study older storms due to the lack of radar data.

The purpose of this study is to investigate the feasibility of simulating past Finnish storms with a convection permitting numerical model. This is done by dynamically downscaling two reanalysis data sets (CFSR, ERA-Interim) with the WRF ARW model at a 4 km spatial resolution. This setup results in the output of simulated radar reflectivities and other storm related parameters. The usability of this data as a radar proxy will be analyzed by comparing the results to actual weather observations, lightning data, newspaper and scientific reports. In addition the differences between the two reanalysis data sets will be studied. Two interesting severe storm cases in Finland from the 1980s were selected: one extratropical storm with remnants of a hurricane and one long-lived MCS.

Paper 98:

Effects of hail damage mitigation on grapevines crop in Mendoza, Argentina, by atmospheric weather modification with cloud seeding.

Martín Alejandro Cavagnaro, Dirección de Agricultura y Contingencias Climáticas, Gobierno de Mendoza, Argentina

Eduardo Martín, Ministerio de Agroindustria y Tecnolología - Gobierno de Mendoza, Argentina

Diego Araneo, Departamento de Ciencias Ambientales, Instituto Argentino de Nivología y Glaciología –IANIGLA-, CONICET Centro Científico Tecnológico Mendoza, Argentina

Leonardo Insegna, Dirección de Agricultura y Contingencias Climáticas, Gobierno de Mendoza, Argentina

Jorge Carbonari, Dirección de Agricultura y Contingencias Climáticas, Gobierno de Mendoza, Argentina


Mendoza is the province with the most severe storm hail events in Argentina every summer, giving rise to several prevention systems tested over the last sixty years. The assessment of hail damage is here statistically compared between years during which any system was applied, versus control seasons with active fighting. We use available information from radar, hail pad network, weather station network and field survey data, from the Dirección de Agricultura y Contingencias Climáticas of Mendoza, Argentina, including an historical grape production area data base obtained from the Instituto Nacional de Vitivinicultura. We discuss the impact of hail damage mitigation in vinegrape crop using active defense by aircraft seeding during the period 1998 - 2012, evaluating the effects of application of atmospheric weather modification by silver iodide and comparing the results with the period 1994-1998, this last without active defense.

Paper 99:

Road Weather Nowcasting

Tatiana Bazlova, IRAM, Russian Federation

Nikolay Bocharnikov, IRAM, Russian Federation

Mark Vinogradov, IRAM, Russian Federation

Alexander Solonin, IRAM, Russian Federation

Nadezhda Iakimainen, IRAM, Russian Federation


Control of snowfalls and ice buildup on roadways during severe winter storms is challenging for road maintenance entities. Some of the critical challenges include short-range forecasts of the onset and cessation of precipitation, its location, rates and types to make effective decisions for treatment types and their timing. The MeteoTrassa system has been developed and applied for tactical use (0-4 hours) in Saint-Petersburg region since 2000. Nowadays the Doppler radar data is used for nowcasting of local severe weather and particularly for quantitative precipitation estimations. The objective of this paper is to provide an overview of the capabilities of the system as they relate to the monitoring and forecasting of weather that may impact the roadway/runway maintenance operations.

Paper 100:

Severe storm reports of the 17th Century: Examples from the UK and France

Katrin Pfeifer, University of Salzburg, Austria

Niki Pfeifer, Munich Center for Mathematical Philosophy, Germany


In this talk we survey reports on selected severe storms from the 17th Century. Specifically, we investigate a severe storm which was accompanied by a ball lightning phenomenon in Cornwall (UK) in 1640. The "fiery Ball", which was reported to be accompanied by a "ter[r]ible sound", entered the church, broke stones and smashed windows. It made holes in stone walls and injured about 14 people. Furthermore, we report on a storm that tore down houses, blew down a stone wall and uprooted trees in Bedford (UK) in 1672 and we report on two severe thunderstorms that tore off roofs and uprooted trees in Oxfordshire (UK) and Blois (F) in 1680. In Oxfordshire, hail stones killed farm animals, and later lightning caused a fire, which damaged houses and burned down barns. In Blois, houses were torn down by the wind, eight perishes were ruined by the hail (hailstones were the size of a "man's fist"). Furthermore, houses were damaged and glass windows were shattered. Based on various primary sources, we discuss the impact of these severe storms on society. Moreover, we briefly discuss how people perceived and explained atmospheric phenomena like storms, tornadoes, and hail. Finally, we discuss selected key issues of investigating historical severe storms, like how the contemporaries’ perceptions influence their eyewitness accounts, and how modern tornado- and wind-scales could be adapted to investigate historical severe storms.

Paper 101:

August 1st 2012 convective episode over eastern France and western Switzerland : observations, forecasts and model deficiencies

Lionel Peyraud, MeteoSwiss, Switzerland


On August 1st 2012 an MCS affected regions of eastern France and western Switzerland. The convection initiated as isolated cells along a prefrontal trough/ convergence line in the Loire valley during mid-afternoon. Over the next couple hours, the convection rapidly grew upscale towards the east through outflow boundary merging and interaction with the nearby topography. The resulting MCS took on an assymetric leading-line / trailing stratiform configuration as it advected further east over Switzerland during the overnight hours. Observations also seemed to show the development of an apparent weak Mesoscale Convective Vortex (MCV) circulation within the MCS after several hours. While this MCS was only of moderate intensity, several numerical models including the COSMO 7km and COSMO 2km models had significant trouble in adequately simulating this convective event. Several reasons as to why this may have been the case will be highlighted in this convective case study. Results from the COSMO-2 Limited Ensemble Prediction System for this case will also be presented.

Paper 102:

Radar-based Precipitation Nowcast at Deutscher Wetterdienst (DWD)

Tanja Winterrath, Deutscher Wetterdienst, Germany

Wolfgang Rosenow, Deutscher Wetterdienst, Germany


Quantitative precipitation forecasts with high temporal and spatial resolution are essential for the meteorological warn management as well as hydrological applications in the context of flood risk management. Here, we present the DWD radar-based nowcast suite focusing on high-resolution forecasts of precipitation quantity, aggregate state, and type.

Radar composites covering the area of Germany serve as input for a tracking algorithm providing nowcasts for up to two hours that is based on the advection of precipitation elements using the displacement vector field that is derived from the mapping of similar precipitation structures in successive radar images. The forecast amounts of precipitation are quantified making use of the most recent gauge-adjustment procedure assuming persistence of the precipitation frequency distribution resulting in the quantitative precipitation forecasts for the next two hours with an update frequency of 15 minutes.

For hydrological and meteorological applications not only the quantity but also the type of the precipitation, covering the distinction between solid and liquid precipitation as well as the forecast of hail events, is of paramount importance. Combining the radar-based precipitation nowcasts with data from complementary observations and numerical modelling, e.g. synoptic weather observations, MSG satellite data, numerical model temperature and humidity profiles, and surface temperature, under consideration of the underlying orography, allows an estimation of the precipitation phase, while the concurrency of high reflectivity values in the radar signal and the occurrence of strong lightning intensity points to an enhanced probability of hail within a thunderstorm event.

Products of the presented nowcast suite are a key component of the automatic warning support system currently under development at DWD.

Paper 103:

Sensitivity to horizontal resolution and turbulence parametrization on high resolution idealized simulations of thunderstorms

Antoine Verrelle, CNRM-GAME Meteo-France, France

Didier Ricard, CNRM-GAME Meteo-France, France

Christine Lac, CNRM-GAME Meteo-France, France


Idealized simulations of deep moist convection are performed with the Meso-NH model at kilometric and sub-kilometric scales.

To trigger convection a thermal perturbation is inserted in the low levels, superimposed on a homogeneous initial state derived from Weisman and Klemp (1982, 1984) studies.
In an idealized framework (without land surface, radiation scheme, orography or Coriolis force), two different sensitivities are analysed:
- the impact of horizontal resolution with increasing grid spacing (4km, 2km, 1km and 500m);
- the impact of one dimensional turbulent (T1D) versus three dimensional turbulent (T3D) parameterization.

The simulations generate one cell which splits into two convective systems: one right-moving supercell and one left-moving multicellular system. Typical features are clearly identified for the supercell (hook echo, FFD, RFD, mesocyclone…)

Results show that the surface precipitation and the size of convective systems increase with increasing resolution. The main convective updraft of the supercell becomes larger at high resolution, but the mean of vertical velocity inside the updraft core (>3 m/s) is weaker.
The percentage of resolved TKE decreases with increasing resolution. This is not consistent and shows the limit of the turbulence parameterization in convective clouds. At high resolution, the source of subgrid TKE, mainly located inside the upper part of the clouds, is mainly due to dynamical processes. Parameterization of dynamical production is function of resolved horizontal and vertical wind shear. At low resolution, these gradients are poorly resolved, this leads to a deficit of subgrid TKE.
Furthermore, inside the clouds, TKE is higher for T3D than for T1D. The cool pools are less intense for T3D. Moreover, kinetic energy spectra show a deficit in subgrid mixing at the smallest wavelengths with T1D at 500m resolution.

To further investigate the structures of turbulence inside the clouds, LES simulations of convective clouds will be carried out.

Paper 104:

A 10-year Radar-based Precipitation Reanalysis for Germany – First Steps and Future Directions

Tanja Winterrath, Deutscher Wetterdienst, Germany

Elmar Weigl, Deutscher Wetterdienst, Germany


The area-covering German radar network comprising 16 operational Doppler C-band systems provides a database of paramount importance for temporal and spatial high-resolution precipitation analyses for real-time and - as the archive encompasses ten years of reflectivity data by now - for climatological applications. Based on the combination of radar based precipitation estimates and comprehensive surface precipitation observations, Deutscher Wetterdienst (DWD) operationally provides high-resolution quantitative precipitation estimation (QPE) products for real-time hydrological applications in the context of flood risk management since 2005. Meanwhile, a valuable database of radar reflectivity and QPE data has accumulated, demanding its reanalysis and evaluation for climatological applications in various subject areas e.g. water engineering, climate monitoring, and climate modelling.

DWD has defined two approaches for the quality-controlled high-resolution precipitation reanalysis of the archived data:

• the post-correction of six years of operational QPE products and
• the reanalysis of ten years of radar reflectivity data.

The first short-term approach is based on the archived operational QPE composite products with a temporal resolution of one hour and a spatial grid size of one kilometre squared covering Germany. To allow for statistical analyses with a focus on extreme precipitation events a post-correction suite is applied to the data to eliminate residual false echoes and further enhance data quality. A first reanalysis has been performed focussing on the metropolitan region of Cologne.

The second long-term approach will comprise a complete reanalysis of the reflectivity data applying a consistent analysis suite. It is based on the local five-minute precipitation scans of the 16 operational weather radars of the DWD network since 2001. Furthermore, additional algorithms will be developed in order to improve the high data quality and comply with climatological standards.

At the conference, we present the first results of the post-correction approach as well as first steps toward the 10-year radar-based precipitation reanalysis.

Paper 105:

Changes of Thunderstorm and Hail Potential in Climate Change

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

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


Severe thunderstorms and associated extreme events such as hail represent a substantial hazard potential for buildings, crops, and critical infrastructure. Due to their local-scale extent and a lack of appropriate monitoring systems, hailstorms are not captured reliably and comprehensively for a long period of time, which hampers statistical analyses including estimation of trends. To overcome these constraints, the examinations will be based on proxy data such as hail-relevant convective parameters. The aim of the study is to quantify how and to what extent the convective potential related to hailstorms has changed in the past and what changes will be expected for the future, particularly in light of global warming.

Within the frame of the project ‘Haris-CC’ (HAil RIsk and Changing Climate), changes of convective conditions are analyzed using both observational data and regional climate models (RCM). According to statistical analyses of various European soundings, the convective potential of the atmosphere has increased significantly in the last 20 – 30 years in most regions in Central Europe. Even though also high-resolution RCMs are not able to simulate single hail storms, they can sufficiently reproduce the convective potential in the atmosphere. Analyses from reanalysis data confirm the general findings of the observed trends. An ensemble of seven high resolution RCM simulations does not show significant variations of the convective potential between a future (2021 – 2050) and a past period (1971 – 2000).

To improve the prediction quality of hail events, a logistic hail model was developed by means of multivariate statistics (logistic regression). This mathematic model is based on a combination of appropriate hail-relevant meteorological parameters (e.g., convective parameters, moisture content, weather conditions) and enables the calculation of a combined index for the analysis of the hail potential, which is termed as Potential Hail Index (PHI). Applied to an ensemble of different RCMs, it shows that the PHI will slightly increase in the future. This implies that the potential of severe hail will (slightly) increase in future decades. This finding is in line with other studies that investigate convective stability for Central Europe.

Paper 106:

The three-dimensional microphysical and dynamical structure of convective storms

Thorwald Stein, University of Reading, United Kingdom

Robin Hogan, University of Reading, United Kingdom

Emilie Carter, Met Office, United Kingdom

Carol Halliwell, Met Office, United Kingdom

Kirsty Hanley, University of Reading, United Kingdom

Humphrey Lean, Met Office, United Kingdom

John Nicol, University of Reading, United Kingdom

Robert Plant, University of Reading, United Kingdom


Forecasting centres routinely run simulations at convection-permitting resolutions, but there is an urgent need for novel radar-observation techniques to evaluate the storm structures produced by these models. A data set of high-resolution radar observations for forty days with convective storms is used to evaluate such storms in the UK Met Office forecast model for the DYMECS project (Dynamical and Microphysical Evolution of Convective Storms). The 3 GHz Chilbolton radar was set up to automatically track convective storms in real-time through a scan-scheduling algorithm linked to a database of storms identified in the Met Office rainfall radar network. Many configurations of the Met Office model have been tested against the Chilbolton observations for their representation of convective storms. In terms of the detailed three-dimensional microphysical structure, modelled storms are shown to have a wider horizontal structure for different reflectivity thresholds compared to radar observations, whilst the storms are not as deep as observed. For instance, the model does not produce reflectivities above 40 dBZ above the melting layer, which is frequently observed in intense convective storms, but this is improved by the inclusion of prognostic graupel in the microphysics scheme. The dynamical storm structures are analysed in terms of vertical velocity and the size of convective cores, derived from vertical profiling radar scans. Two existing retrieval methods for vertical velocities are presented and compared statistically. Both methods are then combined with radar reflectivity observations to define convective cores and to evaluate the size and intensity of such cores in the model. The results presented here will help improve the microphysics and sub-grid mixing schemes, as well as determine whether even higher resolution (down to 100m) models provide a notably better representation of the three-dimensional evolution of convective storms.

Paper 107:

Evolution of Modeled Convective Storms with and without Lightning Data Assimilation

Antti Pessi, Vaisala, United States


Radar data provide valuable information about storm development, track, and intensity. However, radar data are not available or are very limited over most parts of the Earth, including oceans, mountains, and regions with sparse or no radar networks. Accurate, timely lightning data provide supplemental information about convective activity and can be used as a proxy for radar reflectivity. Moreover, data from global lightning detection networks can be used for rainfall estimation and data assimilation into NWP models over data-sparse regions.

To investigate the lightning-reflectivity relationship, lightning and radar data over the continental United States and adjacent ocean areas were collected and analyzed. The lightning data were obtained from Vaisala’s National Lightning Detection Network (NLDN) and Global Lightning Dataset (GLD360) and the radar data from the National Weather Service NEXRAD. The results show a strong log-normal correlation between the lightning rates and reflectivity values.

A new kind of lightning data assimilation (LDA) technique was developed that uses GLD360 and NLDN data and utilizes the aforementioned relationship between the lightning rate and radar reflectivity. The LDA method introduces significant differences in storm structure and dynamics compared to the cases without LDA. These differences are discussed together with model verification results.

Paper 108:

High-resolution model evaluation of convective storm evolution using precipitation radar data

Thorwald Stein, University of Reading, United Kingdom

Robin Hogan, University of Reading, United Kingdom

Emilie Carter, Met Office, United Kingdom

Carol Halliwell, Met Office, United Kingdom

Kirsty Hanley, University of Reading, United Kingdom

Humphrey Lean, Met Office, United Kingdom

John Nicol, University of Reading, United Kingdom

Robert Plant, University of Reading, United Kingdom


Forecast models are routinely run at convection-permitting resolution, specifically in the United Kingdom, where the Met Office UK 1.5km model (referred to as the UKV) is run four times a day to provide a 36-hour forecast to the public. At this resolution, individual convective storms are expected to be resolved, but comparisons with surface rain-rate observations from the Met Office radar network still reveal substantial errors in the timing and characteristics of these storms. In this study, we evaluate the Met Office model at horizontal resolutions of 1.5km, 500m, 200m, and 100m, against rainfall-radar observations and we analyse the model sensitivity to the mixing length in the Smagorinski sub-grid mixing scheme, microphysical parameterisation of graupel and ice, and vertical resolution. The evaluation is first performed on a storm-by-storm basis. For instance, domain-averaged rain rate improves with higher horizontal resolution, which is partly due to an increase in number of small storms (smaller than 40 square kilometres), and an intensification of intermediate-sized storms (of order 100 square kilometres). However, similar results are obtained for a given horizontal resolution by decreasing the mixing length in the sub-grid mixing scheme. Using a storm-identification and tracking algorithm, the models are evaluated in a Lagrangian framework. The life cycle of convective storms in the UKV and rainfall-radar data is presented in terms of storm-average rain rate and storm area, which indicates a typical life cycle in the models with little variation compared to observations. Analysis of the fate of newly initiated storms indicates existence of rapid intensification in observations that does not appear in the models, whilst modeled storms go through fewer break-up and merger events and have a longer life time. Finally, the findings from the Lagrangian framework are used to interpret model evaluation results of rainfall on a catchment-area basis, or an Eulerian framework.

Paper 109:

On the identification of synoptic-scale controls associated with the presence or absence of tornado outbreaks

Chad Shafer, University of South Alabama, United States

Charles Doswell, Cooperative Institute for Mesoscale Meteorological Studies, United States

Lance Leslie, University of Oklahoma, United States

Michael Richman, University of Oklahoma, United States

Andrew Mercer, Mississippi State University, United States

Mason Rowell, University of Oklahoma, United States

Stacey Hitchcock, University of Oklahoma, United States


Recently, studies have investigated the accuracy and skill with which analyzed and simulated fields of meteorological covariates distinguish tornado outbreaks from other less significant severe weather events. These studies have indicated that the magnitudes of the covariates and the spatial extent to which the magnitudes are favorable for significant severe weather both are associated strongly with the severity of the outbreaks. Current findings suggest model-simulated fields of meteorological covariates, using initialized gridded fields consistent with the synoptic scale, exhibit considerable skill in the identification of tornado outbreaks up to at least three days in advance of their occurrence. However, diagnoses of tornado outbreaks occur with a considerable number of false alarms, suggesting a continuing lack of understanding of the synoptic-scale processes associated with the occurrence of tornado outbreaks and a need for objectively derived probabilistic diagnoses of tornado outbreaks. These recent studies examining the identification and discrimination of tornado outbreaks are reviewed, emphasizing (1) the outstanding science questions remaining regarding the skillful prediction of these events, (2) the proposed methods for developing new meteorological covariates that could help distinguish tornado outbreaks from less severe outbreaks, and (3) the development of probabilistic methods in the diagnosis of significant severe weather events.

Paper 111:

A Comparison of Buoyancy and Baroclinity within Tornadic and Non-tornadic VORTEX2 Storms

Christopher Weiss, Texas Tech University, United States

David Dowell, NOAA/ESRL, United States

Paul Markowski, Penn State University, United States

Yvette Richardson, Penn State University, United States


Over the course of the two-year Verification of the Origin of Rotation in Tornadoes Experiment 2 (VORTEX2), approximately 650 transportable “StickNet” probes were deployed near target thunderstorms. Most of the deployment strategies utilized in VORTEX2 favored the simultaneous sampling of two scales. A coarse array of probes (~30 km wide on average, ~3 km spacing) was deployed to sample heterogeneity on the storm scale, primarily tasked with an assessment of the thermodynamic deficits experienced within downdraft regions. Nested within this coarse array was a fine array (~5 km wide, ~1 km spacing) intended for the evaluation of thermodynamic and kinematic gradients very near low-level mesocyclones. As such gradients have been shown to be potentially critical in the evolution of pulses and overall buoyancy within the updraft, and the generation of baroclinic horizontal vorticity for parcels approaching the low-level mesocyclone, the composite depiction of the atmospheric state for tornadic and non-tornadic storms will be beneficial in this regard.

At this conference, we will present the results from a continuing StickNet and mobile mesonet analysis of thermodynamic deficits across a sample of VORTEX2 storms [10 May 2010 (Seminole, OK), 18 May 2010 (Dumas, TX) and 7 June 2009 (Oregon, MO)]. Initial results indicate a fairly disparate presentation of thermodynamics across the 10 May 2010 and 18 May 2010 cases, generally confirming the notion from some previous studies (e.g., Markowski et al. 2002) that negatively buoyant air, though required for the generation of horizontal baroclinic vorticity, can, in excess, preclude the generation of strong low-level vertical vorticity. Further, the shape of the thermodynamic gradients across these cases does support the notion of separate baroclinic zones within the forward flank of these storms, in line with the diagnosis of vorticity budgets for low-level mesocyclone-bound trajectories in recent storm-scale numerical simulations.

Paper 112:

Forecaster Tornado Warning Decision Processes in PARISE 2012

Pam Heinselman, NOAA National Severe Storms Laboratory, United States

Daphne Ladue, OU CAPS, United States

Darrel Kingfield, OU CIMMS, United States

Robert Hoffman, Institute for Human and Machine Cognition, United States


How do forecasters decide to issue or not to issue a tornado warning? What are their radar-based conceptual models? Does rapid-scan radar data enhance the application of those conceptual models and aid forecaster warning decisions? These are the questions we sought to answer via the Phased Array Radar Innovative Sensing Experiment (PARISE). Twelve National Weather Service forecasters participated in PARISE, which ran for six weeks during June – August 2012 at the NOAA Hazardous Weather Testbed in Norman, Oklahoma, USA. Two forecasters participated each week. The experiment's overarching goal was to test whether rapid, adaptive sampling with the phased array radar increased NWS forecasters' ability to effectively cope with tough tornado warning cases.

During the experiment each forecaster worked four cases, two tornadic and two non-tornadic, in displaced real time. For each case, the forecasters' goal was to decide whether a tornado warning was warranted. RecordMyDesktop software recorded forecaster interaction with the radar data. After each case, the forecasters reviewed the video replay of their desktop activity and retrospected aloud about their reasoning and observations. The researcher prompted the forecaster to describe his or her actions and thought processes, and typed out a timeline of these. After completing the timeline, forecasters were asked to identify key judgments during the case and the information used to make them. These data are being analyzed to discern forecasters’ tornado warning processes during the four events. Of particular interest are the decision processes that best helped forecasters assess correctly the tornado threat. Quantitative results indicate strong warning performance by participants. The average tornado lead time was 21 min, the probability of detection values were high (0.75 to 1.0), and the probability of false alarm values were fairly low (0 to 0.5).

Paper 113:

Climate Extremes and Water Balance in the City of Sao Paulo - SP, Brazil: Subsidy for Public Policies

Fernanda Zanon, University of State Julio de Mesquita Filho - UNESP, Brazil

Magda Adelaide Lombardo, University of State Julio de Mesquita Filho - UNESP, Brazil

Bruna Jesus, University of State Julio de Mesquita Filho - UNESP, Brazil


In the last fifty years it is been increasingly evident the anthropogenic interventions in the climate change. Besides the temperature increase, several studies suggest that climate change is also responsible for interference in precipitation, where we observe a systematic increase in the frequency of intense rainfall. In Brazil, the most frequently disasters are associated with atmospheric instabilities that are responsible for triggering extreme weather events such as floods, windstorms, tornadoes, hail and landslides.
The general objective was to analyze, based on the use of a set of indicators, the inter-relationship between precipitation and water behavior that prevails in the city of São Paulo in the last decade, with the support of remote sensing techniques and Geographic Information Systems in analysis of the use and occupation of land.
The analysis resulted in the creation of thematic maps and charts, thus contributing to the discussions that has been proposed to different cities in the world, as the adaptation and serve as a subsidy for public policy.

Paper 114:

Extreme events in the Metropolitan Region of São Paulo (MRSP) - SP, Brazil: public subsidy policies.

Magda Lombardo, University of Sao Paulo, Brazil

Amanda Lombardo, Federal University of Sao Carlos, Brazil


The intense urbanization causes many environmental changes in the cities, including the occurrence of heat islands, thermal anomalies and changes in precipitation. In the case of precipitation, in recent decades, São Paulo has intensified the occurrence of heavy rains more often cause flooding and floods that threaten the quality of life for residents. The extreme weather events are occurring more frequently in the MRSP (Metropolitan Region of Sao Paulo), and the rains has increased in the last 50 years.
The overall objective was to analyze the occurrence of extreme weather events in the MRSP, with emphasis on rainfall during 2010 and the environmental risk, in other words, the identification of points of water logging and flooding and the combination of factors that contribute to urban disasters and that affect the population.
The methodology used for the analysis was to evaluate the dynamic rhythmic with emphasis on climate extremes of temperature and precipitation, the use of Remote Sensing in the analysis of the use and occupation of land and the use of Landsat thermal infrared radiometric temperature for extraction and mapping of regions risk. Extreme weather events, with changing patterns of precipitation, heat waves and prolonged drought cause serious problems to the population, especially those most vulnerable and needy.
With the results of the analysis were generated thematic maps and graphs that can contribute to propose strategies to support public policies.

Paper 115:


Ovcharuk Valeriya, Odessa State Environmental University, Ukraine

Eugene Gopchenko, Odessa State Environmental University, Ukraine


World experience in the sphere of calculation of the maximum runoff characteristics allows classifying of the methods, being used in practice, into few groups:
1.So called formula of maximum intensity or “rational formulas”
2.Volume formulas
3.Empirical, mostly reduction formulas
4.Methodology of “peak-peak”
5.Formulas, based on theory of bed isochrones
IN the first case the empiric dependence between maximum intensity of precipitation for some calculation period and maximum modules of drainage is established.
Volume and reduction formulas in their base depend on single-mode floods, at that the common decision is only volume structure, since variant of reduction formula represents only private case, when duration of water influx from the slopes to the channel net through the territory is changed a little. It is very difficult to realize volume structure for normalization of characteristics of maximum drainage, since duration of flood depends not only on size of basins, but also on factors of hillside regulating of drainage and river-bed – floodplain regulation of floods. To the contrary, too simple methods of generalization of data in structure of reduction formulas and methodology of “peak-peak” for long years displaced development of modern theoretical base for normalization of maximum drainage characteristics. The idea of formation of formulas of maximum intensity (rational method) is not well-founded in theoretical respect, as in the model “precipitations-channel flow” the operator of precipitation transformation in hillside drainage is missed.
It is obvious that two operators must describe the process of formation of channel runoff: “precipitations – slope influx” and “slope influx – channel runoff”. The authors proposed operating structure, having put in a basis model of bad isochrons. In this case calculation formula of the maximal ranoff bases on model slope influx and two transformation functions caused by time of bed running to and of channel and flood plain storage.

Paper 116:


Jefferson Polizel, University of Sao Paulo, Brazil

Magda Lombardo, University of Sao Paulo, Brazil

Demostenes Silva Filho, University of Sao Paulo, Brazil


This research aims to use tools to map the geo-spatial distribution of
thermal field in Piracicaba, SP, and compare the different types of
urban surfaces. Samples were performed in temperature in 3 different
areas, at certain times, in the summer season and winter. These
collection points were set up temperature circular polygons and
polygons through these circulars were derived from images of high
resolution multispectral aerial videography, using the technique of
supervised classification were separated from the percentages of
different types of urban surfaces. Comparisons were made with thermal
imaging and correlated with pixels taken from NDVI (Normalized
Difference Vegetation Index) the multispectral images of multispectral
aerial videography, the results of R2 = 0,68. Other comparisons were
made with the temperatures collected and scenes of the band 6 of
Landsat 5 (TM). For processing the scenes, we used the algorithm
processing software IDRISI 3.2. It was possible to obtain thematic
maps with radiant values temperature of the surface of the town of
Piracicaba. The results obtained by comparing the classes of coverage,
and canopy temperature were adequate to yield an R2 of 0,56 for
circular polygons of 50 meters, other results such as lake/pond R2 was
0,72 and for shade 0,24. With the development of resources of
geotechnology, remote sensing, geographic information system, more
detailed information will be obtained from the urban fabric.

Paper 117:

Numerical Simulation of Supercell Tornadogenesis: The 2012 Tsukuba F3 Tornado Event

Wataru Mashiko, Meteorological Research Institute, Japan


On 6 May 2012, F3 supercell tornado, which is one of the most destructive tornadoes ever in Japan, hit Tsukuba City in eastern Japan and caused severe damage. The environment around the supercell storm was characterized by strong low-level veering sheer and a high CAPE of about 2000 J/kg. Backward trajectories and vorticity budget analysis revealed that the secondary rear-flank downdraft (RFD) surge played a key role in tornadogenesis by transporting the large streamwise vorticity associated with baroclinicity along the forward-flank gust front. Moreover, the secondary RFD outflow surge enhanced the horizontal convergence, contributing to the rapid amplification of the vertically tilted streamwise vorticity. The evolution of vortex lines within low-level mesocyclones is also examined. The vortex lines emanating from the low-level mesocyclone form the arches straddling the hook-shaped hydrometers distribution. This result indicates that baroclinic vorticity generation within RFD region is effective for a low-level mesocyclone formation as in the recent observational studies of suprecells in midwestern United States.

Paper 123:

Hail storms genesis and evolution in the Andes Region (Mendoza, Argentina) derived from radar data

Horacio Pessano, Facultad Regional San Rafael - Universidad Tecnológica Nacional, Mendoza - Argentina, Argentina

Rodrigo Hierro, CONICET - Universidad Austral, Buenos Aires - Argentina, Argentina

Pablo Llamedo, CONICET - Universidad Austral, Buenos Aires - Argentina, Argentina

Alejandro de La Torre, CONICET - Universidad Austral, Buenos Aires - Argentina, Argentina

Andrés Odiard, Dirección de Agricultura y Contingencias Climáticas - Gobierno de Mendoza - Argentina, Argentina

Peter Alexander, Departamento de Física - FCEN - Universidad de Buenos Aires, Buenos Aires - Argentina, Argentina


An environmental lapse rate conditionally unstable, moist air and some mechanism to lift up the parcel up to its level of free convection (LFC) are necessary or the development of deep convection. The rising motions associated with synoptic scale processes are too weak to lift a potentially buoyant parcel to its LFC. Mendoza, Argentina, is a semiarid region situated at midlatitudes (roughly between 32S and 36S) at the east of the highest Andes tops. There, frequent cold fronts arrive from SW and provide the required lift mechanism to trigger severe deep convection events. An appropriate instrument to study hail storms is the Weather radar, which provides a quantitative estimation of hail precipitation with high spacial and tine resolution. In this work, from radar data, we explore the possible mesoscale mechanisms able to trigger deep convection in the study region. We analyze 80 severe storms occurred during the austral summer, between 2005 and 2011.Precipitation flux, storm velocity, cell displacement, storm life time, among other radar variables, are evaluated along complex terrain in two different regions during the selected events.

Paper 125:

Mitigating the impact of environmental degradation on climatic change and global warming

Rehema Namuddu, Rakai Community Development Trust (RACDET), Uganda

Ryan Evans Ntambi, True East African Environmental Alliance (TREAEA), Uganda

Samuel John Ssemwanga, Faculty of Social Sciences, Makerere University, Kampala-Uganda, Uganda

Cissy Irine Namujju, Environmental Awareness Agency (EAA)., Uganda


1. Introduction – The program aim at contributing to improved effectiveness and efficiency of climatic management systems by strengthening the capacity to intervene in an appropriate, effective and timely fashion in the districts of Kampala, Tororo and Jinja.
2. Results and Discussion - The districts of implementation were selected because they hold the most eminent industrial centres and trading hubs of the country that are sources of environmental degradation and industrial emissions. The program seeks to reduce the vulnerability of Ugandans and improving their coping mechanisms to effects of climate change. Environmental aspects in Uganda that have an effect on climate change include; industrial emissions and discharges, vehicle fumes, poor garbage disposal, encroachment on swamps and forest reserves. In return climate change affects the environment as well agricultural production. The unpredictable rains have led to loss of crops because rain has been either too much or too little and too late during planting or harvesting seasons. Climate change is making it difficult for people to feed themselves as crops become susceptible to diseases and unpredictable weather patterns.
3. Conclusions - In order to achieve long term positive effects, the capacity of environment users should be built such that they provide eco-system services through adoption of environmentally friendly practices that mitigate carbon activities, protect and restore water catchments system like massive tree planting, installation of waste treatment mechanisms, advocacy and enactment of legislations as well as installing locally designed incinerators at institutional, community public points, municipal garbage centers and households

Paper 126:

Modelling of the hail hazard in Germany

Marc Puskeiler, Karlsruhe Institute of Technology, Germany

Michael Kunz, Karlsruhe Institute of Technology, Germany

Manuel Schmidberger, Karlsruhe Institute of Technology, Germany


Every summer half-year, extreme hailstorms cause high amounts of losses due to damage to buildings, crops, cars and critical infrastructure in Germany. In the federal state of Baden-Württemberg, for example, most of the damage to buildings is caused by large hailstorms (1986-2011). According to our analyses of past events, track lengths of these systems can be in excess of several hundred kilometers. The time scale from the initiation of the hailstorm to its maximum damage potential may be less than 30 minutes.
Due to the local-scale impacts of a few hundred meters to some kilometers only, hailstorms and their intensities are not captured accurately and uniquely by a single observation system. Therefore, we tried to reproduce tracks of past severe hail events from a combination of different meteorological datasets, such as 2D- and 3D-Radar data, lightning data, Reanalysis and witness observations (e.g., from ESWD), with damage data of insurance companies. Furthermore, a tracking algorithm was applied to the Radar data to analyze also whole tracks of the events.
This approach allows to estimate the number and intensity of hail events in Germany in a very high spatial resolution of 1x1 km². The results show a high spatial variability of the occurrence probability including several hotspots. We found that in particular flow direction, convection potential in terms of convective available energy and orographically-induced flow modifications are most relevant for the spatial variation.

Paper 127:

Detecting and forecasting severe storms using lightning data

Mark Vella, Meteo Malta, Malta


Presentation of a new low-cost instrument for detecting, mapping, and localisation of lightning within 200 km with accuracy of 100 metres, and prediction of lightning events in real time, with archive of recorded historical lightning data. Lightning energy calculation, differentiate between Cloud to Ground and Inter Cloud lightning and thunderstorm tracking will be also the key features of this instrument.

Paper 128:

C-band dual-polarization weather radar hail signatures observed in South Finland

Ljubov Nevvonen, Finnish Meteorological Institute, Finland

Dmitri Moisseev, University of Helsinki, Finland

Venkatachalam Chandrasekar, Colorado State University, United States

Heikki Pohjola, Vaisala Oyj, Finland


Radar hail signatures at C-band are ambiguous. In several studies (Vivekanandan et al., 1990), it was reported that differential reflectivity values of rain-hail mixtures could be higher than 5 dB. This result is contrasting with S-band radar observations where differential reflectivity values close to 0 dB are expected. Due to a high spatial and temporal variability of hailstorms, it is very difficult to provide sufficient ground-truth observations to validate these measurements. Furthermore, for the same reason, performance of current hydrometeor classification algorithms is very rarely quantified in terms of such metrics as for example probability of hail detection. In this study we utilize a unique dataset collected at Helsinki Testbed to address the above-mentioned problems. The Helsinki Testbed was established in 2005 jointly by the Finnish Meteorological Institute and Vaisala. As a part of the Testbed instrumental setup more than 60 Vaisala WXT 510 weather transmitters were installed in the Helsinki Metropolitan area. The weather transmitters are capable of precipitation intensity measurements, as well as of discrimination between rain and hail. Since the establishment of the Testbed, the transmitters have reported more than 100 hail hits. This is a unique dataset since it is not only providing records of hail occurrences, but also provides exact location and times of those events. Our preliminary comparison of radar observations and the WXT hail reports, show a good agreement between those observations. It was interesting to observe that in majority cases hail measurements have shown high Zdr (Differential Reflectivity) values, but in some cases the traditional Zdr 'hole' was observed. Also were compared of radar observations and the Probability-Of-Hail (POH). Those observations were typically taken during late spring several cases are from October and December.

Paper 129:

Thunderstorm forecasting by a fuzzy logic combination of model data

Martin Köhler, DLR, Germany

Arnold Tafferner, DLR, Germany


Within the framework of a PhD study, a new fuzzy logic approach was developed with the aim of combining different data fields of the COSMO-DE weather model in order to forecast thunderstorms for up to six hours. The chosen parameters are omega (500 hPa), CAPE, synthetic radar and satellite data. These are transformed within the approach into linguistic variables with the help of so-called fuzzy input sets under meteorological aspects, and are then combined with an “if .. then” rulebook. The output of the fuzzy logic method is finally a thunderstorm likelihood for each grid point of the COSMO-DE weather model. Currently, an object-based validation of its forecasting performance through a statistical analysis for the summer period of 2012 is performed, including the calculation of different verification scores like probability of detection (POD) or false alarm ratio (FAR). The results of a comparison between the fuzzy logic method and the internal COSMO-DE thunderstorm probability (“neighborhood method”) will help to identify possible improvements of the new method.
The presentation illustrates the concept of the new fuzzy logic approach to forecast thunderstorms up to six hours by a combination of COSMO-DE model data. Additionally the application on some example cases and a comparison with the internal COSMO-DE thunderstorm probability will be shown along with a full statistical analysis for the summer period of 2012. An additional part of the PhD study will focus on certain improvements like tuning of the fuzzy logic system (e.g. by different weighting of the parameters) and a “best member selection” system for the COSMO-DE EPS model. The final goal is to include the new forecasting method into DLR's WxFUSION system, an user-oriented integrated forecasting system for air traffic.

Paper 130:

Radar-based Hail Statistics Over Belgium

Maryna Lukach, RMI, Belgium

Laurent Delobbe, RMI, Belgium


The Royal Meteorological Institute (RMI) of Belgium has been operating a C-band Doppler weather radar for more than ten years (2001-2012). Archived volume radar data are available since 2002 and constitute an invaluable source of information for various statistical studies. Next to many other applications the high resolution volumetric reflectivity measurements of the radar are used for operational hail detection. The hail detection algorithm is based on the criteria proposed by Waldvogel and calculates the Probability Of Hail (POH) as a function of the radar echotop and freezing level heights.
In this study, the Severe Hail Index (SHI) algorithm, which also estimates the Probability of Severe Hail (POSH) and the Maximum Expected Size of Hail (MESH), will be implemented next to the operational algorithm. The temperature profile data, required by the SHI algorithm, will be extracted from a NWP model. Based on the volume radar data and temperature information, the two detection algorithms will be run for the ten-years period. This will allow us to derive statistics on the occurrence of hail events and on their severity. These statistics can be used later for verification of regional climate models.

Paper 131:

Vaisala's new Airport Lightning Information System (ALIS): Using Vaisala's GLD360 to improve cloud-to-ground lightning warnings, present weather reporting, and low level windshear situational awareness at airports anywhere in the world

Nick Demetriades, Vaisala OYJ, Finland


Thunderstorms pose various hazards to airplanes during terminal approach as well as to ground operations personnel. The hazards include (1) dangerous cloud-to-ground lightning, (2) low level windshear, (3) poor visibility, (4) heavy rainfall, (5) hail, (6) severe turbulence, and (7) tornados. Detection and tracking of lightning activity in thunderstorms is relevant to all of these thunderstorm hazards.

The Global Lightning Dataset (GLD360) is produced by a lightning detection network consisting of sensors installed around the world by Vaisala. Breakthroughs in lightning waveform recognition at extreme distances as well as in sensor sensitivity allow the GLD360 network to produce unmatched, near-uniform high performance lightning detection on a global-scale. GLD360 provides ≥70% cloud-to-ground flash detection efficiency and ≥5% cloud flash detection efficiency with a median cloud-to-ground stroke location accuracy of 2-5 km. To ensure the GLD360 network meets these high performance specifications, continuous stringent validation studies are being performed in North America, South America, and Europe.

Vaisala has recently launched a new airport weather offering called the Airport Lightning Information System (ALIS). ALIS combines a web display showing Vaisala's GLD360 data with on-screen, text, and email alerts. This information can be used by air traffic controllers, meteorological observers, ground operations managers, and airlines to improve cloud-to-ground lightning warnings, present weather reporting, and low level windshear situational awareness at airports anywhere in the world.

Paper 132:

Characteristics and properties of overshooting tops

Michaela Valachova, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic

Martin Setvak, Czech Hydrometeorological Institute, Czech Republic


Overshooting tops directly manifest strength of the updrafts which form them and give us essential information about the convective storm from the satellite perspective. This contribution will focus on high-resolution details of distinct overshooting tops and their relationship with other cloud-top phenomena.

In high-resolution visible imagery, the overshooting tops can be revealed by their typical “bubble-like” structure and shadows they cast, however during day-time only. In “microphysical bands” (at 1.6, 2.2 and 3.5-4.0 µm) they may differ from their surroundings by their reflectivity, depending on the particle size and shape within them. In the colour enhanced IR-window brightness temperature field the overshooting tops usually are collocated with the coldest pixels atop of convective storms, however in some cases the overshooting tops do not differ significantly by their brightness temperature from the surrounding cloud top, which makes their IR-based detection problematic.

In our study we will focus on all of these characteristics in detail, to enhance our understanding of storm top processes, using namely the high-resolution data from polar orbiting satellites - such as AVHRR, MODIS and VIIRS, combining these with MSG rapid scan data and other available data sources.

Paper 133:

Comparison of two mesoscale LAM-EPS generation methods for the prediction of heavy rains over the Western Mediterranean: the HyMeX IOP8 event.

Mar Vich, Meteorology Group, Dpt. Física, Universitat de les Illes Balears, Palma de Mallorca, Spain, Spain

Octavio Jaume, Meteorology Group, Dpt. Física, Universitat de les Illes Balears, Palma de Mallorca, Spain, Spain

Víctor Homar, Meteorology Group, Dpt. Física, Universitat de les Illes Balears, Palma de Mallorca, Spain, Spain

Romualdo Romero, Meteorology Group, Dpt. Física, Universitat de les Illes Balears, Palma de Mallorca, Spain, Spain


This study compares a limited area model (LAM) ensemble prediction system (EPS), hereafter PV-gradient, whose initial and boundary conditions perturbations are statistically drawn from a climatology of Potential Vorticity (PV) errors with a reference system consisting of an ensemble of dynamically downscaled members from the ECMWF global EPS. The PV-gradient system has shown better skill than a multiphysics EPS for western Mediterranean cyclonic situations associated with high-impact weather phenomena such as heavy rain. The current study compares the performance of these two generation methods for a heavy precipitation case that affected the eastern and southeastern Iberian peninsula.

The ECMWF global EPS, which initial perturbations are based on the dominant singular vectors, has proven its value and skill for the medium range large-scale forecast. The use of this 50+1 members global EPS is fairly popular for the generation of mesoscale LAM EPSs and has become a benchmark for experimental LAM EPS generation methods. The dynamical downscaling is performed with a 4-km ARW-WRF setup, rendering a considerably higher resolution version of the global predictions for the western Mediterranean area.

In this context, we compare the performance of each strategy for the precipitation field on a typical situation of heavy precipitation event in the Spanish Mediterranean, the HyMeX IOP8 event that occurred on 28-29 September 2012. During the event, heavy precipitations occurred in Andalusia, Murcia, Valencia and Catalonia with more than 200 mm of precipitation values recorded in 24 hours. The most affected regions were Andalusia and Murcia with 11 casualties and material losses estimated at 120 M€. The evaluation procedure compares the results of both EPSs against the observed precipitation values provided by raingauges covering these regions.

Paper 134:

The OASE project: Object-based Analysis and Seamless prediction

Kathrin Wapler, Deutscher Wetterdienst, Germany

Silke Troemel, University of Bonn, Germany

Theresa Bick, University of Bonn, Germany

Hartwig Deneke, Institute for Tropospheric Research, Germany

Malte Diederich, University of Bonn, Germany

Akos Horvath, Institute for Tropospheric Research, Germany

Fabian Senf, Institute for Tropospheric Research, Germany

Clemens Simmer, University of Bonn, Germany

Juergen Simon, University of Bonn, Germany


The research group on Object-based Analysis and SEamless prediction (OASE) is part of the Hans Ertel Centre for Weather Research (HErZ). The group consists of scientists at the Meteorological Institute, University of Bonn, the Leibniz-Institute for Tropospheric Research in Leipzig and the German Weather Service.
One of the central goals of the project is a near real-time radar and satellite remote sensing-driven 3D observation-microphysics composite covering Germany. It contains gridded observations and estimated microphysical quantities. Observations and microphysics are intertwined via forward operators and estimated inverse relations, which also provide uncertainties for model ensemble initialisations. The lifetime evolution of dynamics and microphysics in (severe) convective storms is analysed based on 3D scale-space tracking, which will eventually evolve into a novel nowcasting framework. An object-based analysis condenses the information contained in the dynamic 3D distributions of observables and related microphysics into descriptors, which will allow identifying governing processes leading to the formation and evolution of severe weather events. The object-based approach efficiently characterises and quantifies the process structure and life cycles of severe weather events, and facilitates nowcasting and the generation and initialisation of model prediction ensembles. The object-based perspective also offers a promising strategy for the validation of atmospheric models with observations. The poster gives an overview of the OASE project.

Paper 135:

Detection of mesocyclone signatures in the 3D Doppler data of the DWD radar network

Thomas Hengstebeck, Deutscher Wetterdienst, Germany

Dirk Heizenreder, Deutscher Wetterdienst, Germany

Paul Joe, Environment Canada, Canada

Peter Lang, Deutscher Wetterdienst, Germany


The radar network of the Deutscher Wetterdienst (DWD) offers the possibility of detecting and tracking rapidly developing, small scale severe weather phenomena in an operational environment. Recently, this approach has been further accomplished by an update of the scan strategy towards a 5-minute-repetitive volume scan pattern. The volume scan consists of 10 sweeps with elevation angles ranging from 0.5° to 25° and provides 3D Doppler radar data. Furthermore, the radar network is currently being modernized by replacing the old systems with state-of-the-art dual-polarization radars. In order to adequately exploit the possibilities of the polarimetric data, a software framework called POLARA has been developed (POLArimetric Radar Algorithms), which also hosts "classical", non-polarimetric algorithms. POLARA contains various modules like quality control algorithms, a hydrometeor classification and an improved quantitative precipitation estimation, so that there is the possibility of synergetic exchanges. As part of the quality control the correction of dual-PRF unfolding errors has been improved to eliminate spurious high shear regions and restore corrupted real mesocyclone vortices. Within POLARA a mesocyclone detection algorithm is running in real time identifying mesocyclonic shear features in the measured radial wind field and creating 3D mesocyclone objects from vertically aligned features. Despite of using sets of radar station data rather than a composite, there are no multiple detections of the same mesocyclone in overlapping radar regions, since all features with similar geographical coordinates are merged into a single mesocyclone object. The mesocyclone objects are ranked according to a severity metric taking into account properties like shear, momentum and VIL and made visible to the forecaster at DWD by means of the NinJo meteorological workstation system. The mesocyclone detection in its current implementation will be introduced and the performance will be discussed by means of selected weather cases.

Paper 136:

Analyzing targeting guidance for a DTS-MEDEX-2009 case study: misleading sensitivity products.

Lorena Garcies, University of Balearic Islands, Spain

Víctor Homar, University of Balearic Islands, Spain


The DTS-MEDEX-2009 targeting campaign was carried out during the autumn of 2009 and it was focused on improving the forecast skill of HIW events linked to Mediterranean cyclones. To guide the decision about where additional observations would most benefit the quality of the forecast of each potential adverse event, five different sensitivity analysis techniques were computed: Singular Vectors (SV) from the ECMWF; Ensemble Transform Kalman Filter (ETKF) and Kalman Filter Sensitivity (KFS) from Météo France; and ensemble and adjoint sensitivities from the University of Balearic Islands. Despite not all sensitivity computations were available to the forecasters/scientific teams in real-time, all these sensitivity computations were devised to identify the best location for additional observations. Therefore one immediate question arises: which sensitivity method best advise decision makers on where to deploy an extra observation? This talk attempts to shed light on this question and other such observational and sensitivity concerns by analyzing the guidance provided by these five sensitivity products for one case study of the DTS-MEDEX-2009 campaign.

Since radiosonde and commercial aircraft data (AMDAR) profiles were the only observational means available during the DTS-MEDEX targeting campaign, this study tests the ability of each sensitivity product in identifying the region where a plausible sounding leads to a greater impact on the forecast of a potential high impact cyclone over Southern Italy on December 5th 2009. All targetable radio-soundings sites are also tested and a severe weather meteorologist is used as a confronting reference. The verification testbed comprehends single sounding experiments and multiple sounding strategies by using the WRF Data Assimilation system. Single sounding tests reveal that sensitivity products fail to recognize the best location for a primary observation since most of the soundings added over operational radio-sounding stations have a larger influence on the intense cyclone forecast than the areas indicated by the objective sensitivity calculation methods. Additionally, it is shown that human-based decisions, after evaluating available sensitivity information, are not optimal neither in single nor in multiple sounding strategies. These results evidence the need for a more robust objective guidance to operations centers during targeting campaigns.

Paper 137:

Validation of MPEF GII product against sounding and lightning data - implications on convection forecast

Ivan Smiljanic, DHMZ, Croatia

Zrinko Bahoric, DHMZ, Croatia

Natasa Strelec Mahovic, DHMZ, Croatia


Due to increase in spatial and temporal resolution, satellite data have become an unavoidable part of convection nowcasting systems and a basis for convection-related research. Besides detecting the initiation of convection and following its development, satellite data also enable the assessment of pre-convective conditions. This can be achieved through Global Instability Index (GII), combining SEVIRI radiance data and NWP model fields into stability indices similar to these calculated from upper-air sounding data. The biggest advantage of stability indices derived from MSG satellite data is the capability of the continuous monitoring, insured by the 15 min repeat cycle and up to 3km pixel resolution, providing forecasters with useful information much more frequently than the soundings available only twice daily at a very limited number of stations. GII product is derived within the Meteorological Product Extraction Facility (MPEF) of EUMETSAT in 15 or 5 min (Rapid Scan) cycle and disseminated via EUMETCast.

In this work indices contained in MPEF GII product, i.e. K-index, KO-index, Lifted Index and Total Precipitable Water, will be validated against the same indices derived from upper-air sounding data over southern part of Central Europe and Adriatic region. The 3-hourly average GII index values will also be statistically evaluated against the occurrence of lightning (which is considered an indication of convective activity) over the region, for each day during 3 summer seasons (May-September), 2009-2011. This should give the idea about the threshold values for each stability index and the probability of lightning connected to each index value. The results will be compared to similar results over South-Africa. Moreover, the forecast application of a combination of indices with other factors (orography, land/sea variability) will be tested in order to derive a parameter that could be operationally used for the forecast of convection.

Paper 138:

Severe storm in the western coast of Lithuania on 6 August 2012

Ieva Vitkutė, marine forecast division, Lithuania


Over the past few years it has been observed that the recurrence of severe storms cases in the seaport of Klaipėda and along the Baltic coast of Lithuania is increasing. The poster presentation will analyze a severe storm case, which occurred last summer – on the 6th of August 2012.
The models (HIRLAM, DWD, GFS) were predicting that in the second half of the day there would be an average intensity rainfall, but no strong squall had been forecasted. Meteorologists of the LHMS Marine Forecasting Division were expecting squalls in Klaipėda, on Curonian Lagoon and Southeastern Baltic in the first half of the night on August 6th. Unfortunately, strongly evolved powerful stormy Cb clouds, with tops at 15 km altitude, reached Klaipėda much earlier – around 16:30 UTC (19:30 local time). Sudden squall (32 m/s) accompanied by lightning resulted in 46 mm precipitation measured during 1 hour. The poster will demonstrate in greater detail a mechanism of this phenomenon including its thermal, dynamic and physical conditions and their development. The Author will show the main reasons why these storms are continuing to recur at Lithuanian seashore and how the marine forecasters are working on prediction of severe weather cases using information from satellite images, momentary data from weather radar and sometimes inaccurate numerical weather predictions.
Damage inflicted by very strong wind and heavy rain will be shown as well.

Paper 139:

A new stochastic event catalogue for hail in Europe

Heinz Jürgen Punge, KIT, Germany

Kristopher Bedka, Science Systems and Applications, Inc., United States

David B. Stephenson, University of Exeter, United Kingdom

Michael Kunz, KIT, Germany

Marc Puskeiler, KIT, Germany

Angelika Werner, Willis Re, Germany


Damage by hail fall constitutes one of the major atmospheric risks to mobile, buildings and agricultural values. The quantification of this risk is of particular interest for insurance companies. However, there is little knowledge beyond local historical damage, due to the rarety of events and the lack of uniform detection methods.
Here, we present the hazard component of a stochastic risk model for hail in Europe. The stochastic event catalogue is based on hail observations from the ESWD network and satellite observations of overshooting cloud tops (OT), indicating very strong convection and thereby favorable conditions for hail formation.

Historic hail events are defined based on OT occurrences detected from infrared brightness temperatures of the MSG SEVIRI satellite between 2004-2011. The stochastic catalogue uses this historic event properties compromised by hailstone observations from the ESWD network to derive more than 1 million individual events with an event footprint resolution of 1km.
The presented catalogue is the first hail event catalogue based on a single homogeneous observation source over Europe. Key hail risk areas in central and southern Europe are represented, and high frequencies occur in regions neighbouring the Alps and the Pyrenees. A further maximum occurs in central Eastern Europe. However, major hail events can occur everywhere in Europe.

Paper 140:

The Fitzroy Stormglass - a europeanwide old instrument of weather forecasting

Claus Schott, Caritas Hospital of Bad Mergentheim, Germany


It is said that the liquid within the Fitzroy stormglass is a mixture of several ingredients, most commonly distilled water, ethanol, potassium nitrate, ammonium chloride, and camphor. This specific mixture is said to be developed by Admiral Robert FitzRoy and used on his voyage with Charles Darwin on the HMS Beagle in 1835.
During the historic voyage, FitzRoy carefully documented how the storm glass would predict the weather.

In this historical examination it could be found out that the stormglass of Admiral Fitzroy was not only an instrument used by Charles Darwin and Admiral Fitzroy. It was an old well known instrument not only used in seafaring, but also mainly used as a agriculture prognostic instrument spread all over europe. Starting up with the first experiments about the cristalisation process in 1748 under Romieu in France (without mentioning the cristalisation processes as a weather predictor), it was first mentioned as a stormglass 1780 in France and 1782 in Germany. Afterwards it was found also in Denmark, Netherlands and later on in the USA up to the 19 th century under different names. All europeanwide spreaded stormglasses are containing the same ingredients as the ancient british stormglass in different concentrations filled in a closed glas bottle. It is also said that this instrument can predict weather changes because of changes of the cristalisation formations 6-9-12-16-24-36 up to 48 h before a storm.

Paper 141:


Khalid Muwembe, Uganda Department of Meteorology, Uganda

Charlie Williams, University of Reading, United Kingdom

Thorwald Stein, University of Reading, United Kingdom


Several studies suggest that weather patterns over Lake Victoria are highly variable, with wind gusts in the vicinity of thunderstorms suddenly thrashing up high waves capable of capsizing small fishing boats. In order to improve safety on the lake, the Mobile Weather Alert (MWA) service is implemented by utilising mobile phone technology to provide daily weather forecasts that reduces vulnerability of fishermen to weather hazards. This service was developed as a pilot scheme under the WMO Severe Weather Forecast Demonstration Project (SWFDP). An evaluation of MWA forecasts over Lake Victoria has been carried out using binary forecast verification methods based on a standard contingency table.

In MWA forecast verification, different observation thresholds are used to assess the accuracy of the MWA forecasting system in a binary (yes-no) situation: These include 0.1mm (Observation), 2mm, 5mm, 10mm and 20mm thresholds. Furthermore, three observation data sets are used in the study: averaged rainfall from three coastline rain-gauges, TAMSAT RFE, and ATD lightning data. Forecast performance was assessed using different scores; Frequency Bias Index, Proportion correct; Critical Success Index; Equitable Threat Score (ETS) and True Skill Statistics (TSS). The MWA forecasts were found to have high success rate of around 70% when verified against averaged rainfall and RFE, however 60% success rate was found when forecasts are verified against ATD data. These success rates are based on the frequency of hits and correct rejections forecasted during the period February – April 2012. The study illustrated that TSS and ETS are the best scores for verifying the MWA forecasts since both measures demonstrated that the forecasting process has skill in predicting severe storms as well as calm days over the Lake Victoria region.

On assessment of forecasting tools, interviews and questionnaires completed by forecasters at Uganda Department of Meteorology (UDoM) indicate a wide range of forecast tools are available to operational forecasters. The main limitation is lack of in situ observations over the lake which makes day-to-day forecast verifications a little difficult. Various suggestions are made on improving the accessible forecasting tools, among others, the need to develop hazard diagnostics from the available high- resolution model products and tuning the science configuration in the 4 km model for the tropical regions which is expected to improve model forecasts.

Paper 142:

Observational Investigation of a Tornadic HP Supercell Storm in China

Xiaoding Yu, Training Centre/China Meteorological Administration, China


A detailed analysis on a tornadic HP supercell’s environmental conditions, structure and evolution is made based on the Doppler weather radar data, routine upper-air and surface observation, and intense automatic weather station observations. The main results are as following:
1) This HP supercell occurred in the moderate CAPE and significant vertical wind shear environment, with high value of low level vertical wind shear and low LCL at same time. The moderate CAPE and significant vertical wind shear favors the generation of supercell, while the high value of low level vertical wind shear and low LCL favors the occurrence of strong tornadoes.
2)This HP supercell began to develop when the pre-exist quasi-linear convective system (QLCS) becoming wider and shorter, the mesocyclone first appeared within one cell locatde in the middle of the QLCS, beginning at 4km height and then developing upward and downward. Afterwards, the mesocyclone also developped in the southern cell of the QLCS, and gradually the southern cell merged with the middle cell, forming a strong HP supercell, with a 12km diameter and 9 km height mesocyclone embedded. The vertical vorticity associated with the mesocyclone is 1.5×10-2 s-1. The radar echo of this supercell successively displayed kedney, spiral, “S” shape, and finally evolved into bow echo, lasting more than 2 hours.
3) The F3 tornado occurred during the “S” shape period. Before tornado touching down, a TVS appeared in the central part of the large mesocyclone, corresponding to a vertical vorticity value of 6.0×10-2 s-1. When the tornodo was underway, strong divergence present at the storm top just above the position of the ground tornado, with a divergent value of 0.8×10-2 s-1。The mesocyclone that lead to the tornado lasted 2 hours and 13 minutes.
The mechanisms for mesocyclone generation and HP echo revolution are discussed in details.

Paper 143:

Tornadogenic versus non tornadogenic SuperCell discrimination using radar in Portugal

Paulo Pinto, IPMA-PortugueseMetService, Portugal


At the Portuguese Met Service (IPMA) current strategy towards severe weather Nowcasting depends on several tools. One of such is the automatic radar based warning products catalogue. General purpose and aeronautical forecasters are trained, on an ongoing basis, in order to optimize the use of these, in the scope of an operational environment. Diagnose and short term forecast of severe weather phenomena or, more properly, High Impact Weather phenomena is the target of such catalogue. One of these products is the rotation circulation detection product (RCD) developed in order to diagnose supercell (SC) and mesovortices (MV) like circulations. The configuration of such automatic products makes use of the IRIS (Interactive Radar Information Systems) ability in capturing, in a combined way, several 2-D radar product fields, thresholds and areas. Areas prone to the onset of a particular phenomena are, thus, identified. In the case of the aboved stated product it will be a SC or MV. The tune of the configurations depends on the synoptic climatology and the corresponding observational series. Fortunately, Portugal has a quite low tornado rate, strongly dependent on the prevailing winter circulations. This fortune has a draw back on the tunning side, for obvious reasons, mainly due to very scarce ground truth reliable records. Nevertheless, on the last November the 16th, a tornado outbreak has taken place in southern Portugal. An instability line developed under a very warm and moist airmass, creating favourable conditions for these phenomena to occur, in a low system circulation where wind shear played its role. We could confirm 4 tornados, 3 of which producing damage, the strongest classified has an F3/T6 (F, Fujita; T, Torro). Such events affected, mainly, quite densely populated areas and this allowed an unusually reliable ground truth to be established. These episodes and a few more events that occured between October 2009 and the present time, aloowed us to test and verify configurations with a new philosophy. This task has accomplished two different goals: a) to enhance CSI of the forecast of SC detection as such, compared to the current RCD product CSI; b) to allow a promissing discrimination between regular SC and more prone to tornadic events SC.

Paper 144:

On the extreme summer precipitation in Ukraine over the last decades

Vladyslav Tymofeyev, Ukrainian Hydrometeorological Institute, Ukraine

Alexander Scheglov, T. Shevchenka Kiev national University, Ukraine


Recent decades’ climatology of precipitation in Ukraine including extreme events is studied as well as case studies in individual years. Precipitation is analyzed via data of individual stations as well as in space, in the sight of the recent climate change. New seasonal features of near-surface warming were identified: Both maximum and minimum Summer SAT showed further growth in the recent decade, greater than winter season. Trends in total precipitation amount, days with precipitation and single daily intensity were obtained; Significant difference among individual months is detected, with an increase in total precipitation amount and days with precipitation in June and July, and decrease – in August. Detail specification is done by means of threshold of significant precipitation. Greatest growth in the daily intensity occurred between 1990s and 2000s in central-southeast Ukraine; this parameter can be used for regionalization of changes in space.
The main large-scale synoptic patterns, causing the growth in the daily precipitation in June or July Is deep European trough with cut-off low. On the other hand blocking high is dominated in the East Europe in late July-August, providing great weather contrast between West and East Europe. This cause smaller frequency of droughts in east Europe in the early summer, and greater frequency in August.
Types of mesoscale fields responsible for extreme events are recognized within the larger-scale synoptic systems, such as convergence lines or mesocyclones. Transition between atmospheric scales can be used to assessing general threat of extreme precipitation.

Paper 145:

Assessment of the skill of nowcasting to predict high-impact Heavy Precipitation Events

Joan Bech, Universitat de Barcelona, Spain

Marc Berenguer, Centre de Recerca Aplicada en Hidrometeorologia (CRAHI). Universitat Politècnica de Catalunya, Spain


Recent advances in very short-term precipitation forecasting, or nowcasting, systems have contributed to improve weather surveillance and allow issuing early warnings for heavy rainfall events. These warnings are particularly necessary to mitigate flash flood impacts when the event takes place over a small catchment or in a highly populated area, typical features of Western Mediterranean coastal regions. In this study, using a nowcasting system based on Lagrangian extrapolation of radar-derived precipitation fields, we examine the predictability of several high-impact events that took place in Catalonia (NE Spain). Weather radar quality controlled data are used for generating up to 3-h forecasts which are verified against both weather radar precipitation estimates and raingauge observations. Precipitation events include the 7 September 2005 Llobregat Delta river tornado outbreak and the 2 November 2008 tornadic thunderstorms, with total precipitation amounts exceeding 100 mm in 24 h and intense rainfall rates that caused local flooding, with maximum 30 minute raingauge accumulations of 38.2 and 12.3 mm respectively – in total, 3456 h of precipitation forecasts were verified. To assess the quality of precipitation forecasts a number of verification indices were calculated, including the critical success index (CSI), the probability of detection (POD), the false alarm rate (FAR) or the linear correlation for several rainfall rate thresholds and lead times. Results comparing nowcasts with raingauge 30’ amounts indicate that forecasts provided useful guidance for moderate rainfall intensities (up to 1 mm in 30’) for lead times of 60’ (POD above 0.60 and FAR below 0.15), while comparable scores were obtained for rainfall occurrence (0.1 mm in 30’) up to 2.5h (POD above 0.5 and FAR below 0.2). Similarly, linear correlations of radar estimates and forecasts exceeded Eulerian persistence of precipitation estimates for lead times of 1.5 h for moderate intensities (up to 0.8 mm/h) while predictability of higher rates decreased quickly. The analysis concludes with a discussion of event-to-event variability and the practical limits to nowcasting derived automated precipitation warnings. This study is partly funded by the ProFEWS project (CGL2010-15892).

Paper 146:


Vincent Sakwa, Kenya Meteorological Department, Kenya


Numerical Weather Prediction (NWP) systems have become increasingly relevant and indeed essential to the severe weather forecasting process, with a growing number and variety of sophisticated outputs, currently available from NWP producing centres, which are beneficial to severe weather forecasting to many National Meteorological and Hydrological Services (NMHS). The Severe Weather Forecasting Demonstration Project (SWFDP) is a series of regional subprojects whose scope is to explore and test the usefulness of the products currently available from NWP systems of global and regional meteorological centres, with the goal to improving severe weather forecasting services in countries where sophisticated model outputs are not currently used.
The goals of the SWFDP are the following:
• to improve the ability of NMCs to forecast severe weather events;
• to improve the lead time of alerting of these events;
• to improve interaction of NMCs with Disaster Management and Civil Protection Authorities (DMCPA) before and during events;
• to identify gaps and areas for improvements to improve the skill of products from Global Data-Processing and Forecasting System (GDPFS) Centres through feedback from NMCs.
The regional subproject for Eastern Africa was implemented in September 2011 and it focuses on the following severe weather events;
• Heavy rain/flooding and deficit of precipitation/dry spells
• Strong winds in relation to thunderstorms and any other phenomena over the Indian Ocean and major Lakes.
• Hazardous Indian Ocean and major lake waves. .
The demonstration project uses a cascading approach to provide greater lead-time for severe weather and at the same time contribute to capacity building and improving links with Disaster Management and Civil Protection Authorities (DMCPA). The cascading process involves global centres (National Centers for Environmental Prediction-NCEP, European Centre for Medium Weather Forecasting-ECMWF, Germany Weather Service-DWD, UK Met. Office), two regional centres (Kenya and Tanzania met. Services) and participating NMHSs located within the footprint of the subproject. NMHSs involved in the subproject are Rwanda, Burundi, Uganda, Kenya, Ethiopia and Tanzania. The three levels applied in cascading approach for forecasting severe weather are as follows:
• global NWP centres which provide available NWP products, including in the form of probabilities;
• regional centres which interpret information received from global NWP centres, run limited-area
models (Consortium for Small scale Modelling-COSMO and Weather Research and Forecasting-WRF) to refine products, liaise with the participating NMCs;
• NMCs who issue alerts, advisories, severe weather warnings to the public, decision makers and DMCPA;

Since the inception of the subproject the following achievements have been realized in the region:
• improved the ability of NMCs to forecast severe weather events.
• Improved on weather alerts related to severe weather events hence reduced weather related disasters such as floods, landslides and destructive winds.
• improved the lead time of alerting of these events.
• Improved the interaction of NMCs with DMCPA before and during events.

However, storms continue to destroy life and property over Lake Victoria due to poor communication in the Lake. There is poor station network of cell phone provider in the Lake and hence a weak network of the cell phone communication.

There is concerted effort to improve station observation network in the lake region. Two data buoys have been installed on the Kenyan side of the lake. A number of member countries are planning to install weather radars in the region. Already Radio and internet (Ranet) stations have been installed in some areas and are being used to communicate weather information to the fishermen.

Paper 147:

A study of the 21 March 2012 tornadic quasi linear convective system in Catalonia

Joan Bech, Universitat de Barcelona, Spain

Joan Arús, Agencia Estatal de Meteorología (AEMET). DT Catalunya, Spain

Salvador Castán, Agencia Pericial, Spain

Nicolau Pineda, Servei Meteorològic de Catalunya, Spain

Oscar van der Velde, Lightning Research Group. Universitat Politècnica de Catalunya, Spain

Joan Montanyà, Lightning Research Group. Universitat Politècnica de Catalunya, Spain


This study presents a description of a quasi linear convective system that took place in Catalonia (NE Spain) on 21 March 2012 producing heavy rainfall, intense lightning and a weak tornado in the village of Ivars d’Urgell around 19 UTC. A post-event survey indicated F0 and F1 damage in houses of the village –roofs and ceilings, broken windows, fences and walls and trees knocked down– along a track approximately 2 km long and about 200 m wide. Doppler radar observations show that the parent thunderstorm that spawned the tornado was one of a series that developed along a convective line that moved from S to N, initiating convective activity in terms of precipitation and lightning in the Mediterranean Sea and moving inland in S Catalonia (Tarragona and Salou coastal area, producing local flash floods). Convective activity remained several hours with series of thunderstorms developing along the same paths. The synoptic situation was dominated by a high pressure ridge extending from northern Africa to central Europe, with a closed maximum sea level pressure area around 1036 hPa over northern France, southern Germany and Austria. On the other hand a relative low pressure area seen on 850 hPa and upper levels was present over the Iberian Peninsula, favouring a southern maritime flow from the Mediterranean between the forward part of the low pressure area and the high pressure system which blocked the advance of the low to the east. In the study we examine both the synoptic environment and storm scale observations with Doppler radar and total lightning data (cloud to ground and intracloud flashes) that lead to this cool-season severe convective event which is remarkable given the fact that, unlike in this case, most reported tornadoes in the region occur during the warm season (with peaks in August and September) and during daylight hours (6 to 18UTC).

Paper 148:

COSMO-CZ-EPS – the first evaluation

Petr Zacharov, Institute of Atmospheric Physics ASCR, Czech Republic

Daniela Rezacova, Institute of Atmospheric Physics ASCR, Czech Republic

Zbyněk Sokol, Institute of Atmospheric Physics Academy of Sciences of the Czech Republic, Czech Republic


The study presents a first experiences with a new ensemble COSMO-CZ-EPS. This short-range ensemble prediction system is computed on the Institute of Atmospheric Physics ASCR on the initial, lateral and boundary conditions from COSMO-SREPS and COSMO-LEPS ensembles. The COSMO-SREPS is a multi-model 3-member ensemble, driven by three deterministic global models: GME, GFS and IFS global model. The COSMO-LEPS ensemble is driven by 16 selected members from ECMWF ensemble. The COSMO-CZ-EPS is integrated in 2.8km horizontal resolution on the domain covering the Czech Republic and near neighbourhood. The integration started at 0600UTC and finished at 2400UTC of the same day.
This study shows the results of QPF from July 2012 with several events with heavy convective precipitation or other severe convective phenomena. The precipitation fields are verified by gauge adjusted radar measurement. The verification used traditional scores and modern SAL technique.
The ensemble COSMO-CZ-EPS will serve as the base for ensemble spread-skill relationship assesment in the future.
The work is supported by the grant GACR 209/12/P701.

Paper 149:

The impact of multiple rain classes on surface DSDs in idealized simulations of supercell storms

Matthew Kumjian, National Center for Atmospheric Research, United States

Hugh Morrison, National Center for Atmospheric Research, United States


In deep convective storms, rain is generated by a number of mechanisms. These include warm-rain processes, shedding from hailstones, and melting of smaller ice particles. Each of these processes can produce different drop size distributions (DSDs) that, when combined, produce DSDs not well modeled by the inverse exponential or gamma functions typically employed in bulk microphysics parameterization schemes. Indeed, polarimetric radar and disdrometer measurements often reveal large inhomogeneities in precipitation characteristics and exotic DSDs in convective storms at small spatial scales, particularly in supercells.

Following a concept that has been used in the Straka Atmospheric Model, the Morrison microphysics scheme implemented in WRF is modified to contain multiple classes of rain, partitioned based on the physics of its origin (warm-rain processes, shed rain, and melted ice). Idealized WRF simulations of a supercell storm are performed with the standard and newly modified (partitioned rain) microphysics. DSDs at points of interest (e.g., hook echo, forward flank) simulated by the two versions of the scheme are analyzed and compared to past polarimetric radar and disdrometer observations. The impacts of differences in simulated DSDs on rates of evaporative cooling are discussed.

Paper 150:

Electrical and microphysical characteristics of a winter storm over Colorado

Matthew Kumjian, National Center for Atmospheric Research, United States

Wiebke Deierling, National Center for Atmospheric Research, United States


On 10-11 November 2012, a winter storm affected parts of north-central Colorado, producing > 12 cm of snow and graupel accumulations across the region. In addition, the storm was unusually electrically active, producing numerous in-cloud flashes and several cloud-to-ground strikes. The storm was well sampled by the Colorado Lightning Mapping Array (COLMA), as well as the polarimetric WSR-88D radar near Denver (KFTG). Additionally, lightning data were collected by other lightning detection networks that detect cloud-to-ground (and a portion of in-cloud) lightning over the continental United States.

Polarimetric radar data from times leading up to, during, and after lightning discharges are investigated to explore changes in the microphysical characteristics during the storm's life cycle. For the time period where the storm was electrically active polarimetric radar data will also be compared to the three dimensional lightning information provided by COLMA. Additionally, time series of graupel and ice crystal mass retrieved from the radar data are presented.

Paper 151:

Estimation of Stability Indices from ATOVS/NOAA and ATOVS/METOP measurements over Ukraine

Oleksandr Bobryshev, Ukrainian Hydrometeorological Research Institute, Ukraine

Oleksii Kryvobok, Ukrainian Hydrometeorological Research Institute, Ukraine


The scheme of the estimation of stability indices over Ukraine using ATOVS/NOAA and ATOVS/METOP measurements is developed. Stability Indices are defined as the stability of the atmosphere in hydrostatic equilibrium with respect to vertical displacements and is a critical factor in severe weather development. Atmosphere profiles, required for estimation of Stability Index are retrieved use the IAPP package which reads ATOVS level 1D data and produces temperature and water vapour profiles. K index, KO index, lifted index, and maximum buoyancy are obtained.
The retrieved temperature and water vapour profiles based on satellite-date are in a good agreement with observed ground radio sounding data. In order to test the usefulness of this estimation of stability indices for short-term forecast several cases of rapidly developing convective storms observed over Ukraine were analysed. The comparison between stability indices estimation from satellite derived atmospheric profiles and NWP forecast shows that the first one is able to predict instability better over pre-convective areas. Thus, it is expected that the nowcasting and short-term forecast can be improved by the operational Stability Index products derived from satellite measurements of atmospheric profiles.

Paper 152:

The impact of future thunderstorm activity on the air traffic system

Otto Hyvärinen, Finnish Meteorological Institute, Finland

Andrea Vajda, Finnish Meteorological Institute, Finland

Stephan Kral, Finnish Meteorological Institute, Finland


Convective events (thunderstorms) affect transport system in various ways causing sudden disruptions in traffic. By exploring the uncertainties in the future changes in local severe thunderstorm activity, major options can be identified to reduce the vulnerability of aviation system and facilitate the development of methodologies of quick reaction in emergency situations. This study provides an estimate in the possible short and long-term changes in local severe thunderstorm activity in the vicinity of selected European hubs. The ERA-Interim reanalysis dataset provided by ECMWF and Europe-wide surface data (originating from ESWD) in combination with remote sensing data are utilized for the present day study of severe thunderstorms. In order to assess the possible future changes in the convective events the high-resolution regional climate model runs, available from the ENSEMBLES project, will be used. Variations in convective available potential energy (CAPE) as well as other robust indices are defined and ranking of convection intensity is provided. The defined indices are applied to the RCM data to define the changes in the probability of severe thunderstorms under the projected future climate conditions.

This study is associated with the EU FP7 project MOWE-IT (Management of weather events in transport system). The major goal of the project is to identify existing best practices and to develop methodologies to assist transport operators, authorities and transport system users to mitigate the impact of natural disasters and extreme weather phenomena on transport system performance.

Paper 153:

Synthetic NWC-SAF products for regional NWP forecasts of summer convection over Central Europe

Fabian Senf, Leibniz institute for tropospheric research, Germany

Hartwig Deneke, Leibniz institute for tropospheric research, Germany

Kathrin Wapler, German weather service, Germany


The EUMETSAT satellite facility on support to Nowcasting (NWC-SAF) provides a variety of cloud products to forecasters for assessing current weather conditions and possible hazardous changes, especially valuable for fastly developing severe summer convection. The set of NWC-SAF products ranges from cloud properties, like type and top altitude, estimates of precipitation and convective development to dynamical and environmental analysis, e.g. atmospheric motion vectors and stability indices.

We present a study in which the NWC-SAF software for product generation is applied to synthetic satellite data which are calculated from the regional weather prediction model COSMO-DE via the radiative transfer model RTTOV. We give an overview of the applications of synthetic NWC-SAF products. For the high-resolution wind product HRW, retrieval accuracy is estimated within a closure study. Cloud motions are commonly used to estimate wind components of the large-scale geostrophic flow. For severe summer convection, however, the impact of mesoscale circulations on the HRW product is demonstrated. In addition, a generic framework for forecast verification is established based on synthetic satellite products. We demonstrate the use of the RDT (Rapid Development Thunderstorms) product as object-based tool, and derive a relationship between derived cloud top expansion rates and simulated updraft strengths.

Paper 154:

The Effect of Climate Change Adaptation on Rural Community Livelihoods

Robertson Evans Mbidde, Research Environmental Agency (REA), Uganda

Rose Mary Nalubega, Global Change Agency (GCA), Uganda

Ronald Lutwama, Faculty of Science and Technology, Makerere University, Kampala, Uganda., Uganda


Background: Dependence on rain fed crop production means the impact of stresses and shocks are felt keenly by rural poor people, who depend directly on food system outcomes for their survival, with profound implications for the security of their livelihoods and welfare. However, such stresses and shocks will not necessarily lead to negative impacts, as risks and uncertainties, often associated with seasonality, are embedded in the practice of agriculture and there is considerable experience of coping and risk management strategies among people working in this sector. This study explores the opportunities for linking social protection, CCA and DRR in the context of rain fed crop production, establishing whether these three approaches would help enhance resilience to shocks and stresses in rain fed crop production areas.

Methods: The study reviewed the conceptual and policy-related similarities and differences between the three disciplines. Case studies where climate change and resilient social protection approaches were collected and an adaptive social protection framework that highlighted opportunities was developed.

Results: With climate change, the magnitude and frequency of stresses and shocks is changing and approaches such as social protection, disaster risk reduction (DRR) and climate change adaptation (CCA) will be needed to bolster local resilience and supplement people’s experience. Social protection and DRR measures designed to limit damages from shocks and stresses may not be sufficient in the longer term. For social protection to be resilient to climate change impacts, it will need to consider how reducing dependence on climate sensitive livelihood activities can be part of adaptive strategies.

Conclusion: Climate Change Adaptation and Disaster Risk Reduction cannot effectively address the root causes of poverty and vulnerability without taking a differentiated view of poverty, integration with social protection can solve the problem.

Paper 155:

Trends of Natural Disasters and Morbidities in Uganda

Robertson Evans Mbidde, Research Environmental Agency (REA), Uganda

Rose Mary Nalubega, Global Change Agency (GCA), Uganda

Ronald Lutwama, Faculty of Science and Technology, Makerere University, Kampala, Uganda., Uganda


Background: The concepts of natural disasters and complex emergencies are quite distinct, with different strategies for mitigation and response. However, natural disasters and epidemics occur concurrently in the same geographic location.

Methods: Natural disasters, complex emergencies and epidemics that met the inclusion criteria were included. Results: The study revealed that 63% of the complex emergencies had ≥1 epidemic compared with 23% of the natural disasters. 27% of the largest natural disasters occurred in areas with ≥1 ongoing complex emergencies while 87% of the largest complex emergencies had ≥1 natural disaster.

Conclusion: Epidemics commonly occur during complex emergencies. The data presented in this article do not support the often-repeated assertion that epidemics, especially large-scale epidemics, commonly occur following large-scale natural disasters. This observation has important policy and programmatic implications when preparing and responding to epidemics. There is an important and previously unrecognized overlap between natural disasters and complex emergencies. Training and tools are needed to help bridge the gap between the different type of organizations and professionals who respond to natural disasters and complex emergencies to ensure an integrated and coordinated response.

Paper 156:

Intense precipitation patterns by means of Concentration Index – examples of two Brazilian sites

Lucí Hidalgo Nunes, UNICAMP-Universidade Estadual de Campinas, Brazil

Javier Martín-Vide, UB-Universitat de Barcelona, Spain

Guilherme Henrique Gabriel, UNICAMP-Universidade Estadual de Campinas, Brazil


The study evaluated patterns of intense rainfall at daily basis by means of the Concentration Index (CI) proposed by Martín-Vide (2004) that considers the contribution of the days of greatest rainfall to the total amount, and thus permit to identify the rainfall distribution. The index was used in two regions of Brazil: Santos and Campinas. The former (419,400 inhabitants) is the oldest settlement of the country, with plains extending from the Atlantic Ocean to the Serra do Mar Range and rich ecosystems; however accelerated changes in this fragile region have speed up natural processes such as mass movements, mainly triggered by intense rainfall. Campinas (1,080,999 inhabitants) is a vigorous economic centre - its airport, for instance, had recently the second highest growth in the world - but rapid and unplanned changes have contributed to the increase in the severity of hazards triggered by intense precipitation. Thus, an understanding of the rainfall distribution is essential for identifying the susceptibility to high-intensity events in both sites in view of reducing loss of life, damages and the number of affected people.
Daily rainfall of homogeneous precipitation records with good registration accuracy spanning from 1958 to 2003 was distributed in classes of 5.0 mm (0-4.9 mm; 5.0-9.9 mm and so on). For Santos region values of CI were 0.64 (2 stations) and 0.72 (2 stations), while for Campinas values were 0.69, 0.70 and 0.71 (2 stations). Although the mechanisms that generate rainfall in the regions are distinct the indices for both are similar and can be considered high, enhancing that Santos and Campinas present important variability, fact that poses difficulties for forecasting and planning any activity in the sites, which are amongst the most dynamic and with greater potential for economic growth of Brazil, attracting investments and population which, notwithstanding, are at permanent risk.

Paper 158:

Dense weather radar network for observation of severe storm in Helsinki metropolitan area

Roberto Cremonini, University of Helsinki, Dept. of Physics, Finland, Finland

Dmitri Moisseev, University of Helsinki, Dept. of Physics, Finland, Finland

V. Chandrasekar, Colorado State University, United States

Pekka J Rossi, Finnish Meteorological Institute, Earth Observation, Finland, Finland

Susanna Lautaportti, University of Helsinki, Dept. of Physics, Finland, Finland

Laura Rojas, University of Helsinki, Dept. of Physics, Finland, Finland

Annakaisa von Lerber, University of Helsinki, Dept. of Physics, Finland, Finland


The United Nations expects that today’s urban population of 3.2 billion will increase to nearly 5 billion by 2030, resulting in three out of five people living in cities worldwide. The growth of large cities has been rapid over the last half of the 20th century, and this trend seems set to continue to the extent that megacities – defined by the UN as cities with ten million or more inhabitants – have been called the ‘urban phenomenon of the 21st century’. Rapid, uncontrolled spatial growth and densification create settlements in inappropriate areas most likely to be exposed to natural hazards. Moreover, according to the United Nations, among the 450 urban areas with a population of at least one million in 2011, almost 60% are exposed to the risk of a natural disaster.
Stormy weather, heavy rainfall and flash floods are the main sources of risk in urban areas. However, urban areas influence the weather causing local amplifications and damping of atmospheric phenomena. In order to be useful for pluvial flood forecast, up to 100-meters resolution quantitative precipitation estimation (QPE) is needed by hydraulic and hydrological models. High density meteorological network are needed in urban areas to improve the understanding in local atmospheric and environmental processes and to reduce the risk exposition. Helsinki urban area is one of the best monitored areas in the world, with tree polarimetric C-band radars with 16-km average distance, and surface weather stations, that are part of Helsinki testbed, and FMI weather stations. The weather radar network, combined with ground weather stations, allows high resolution observation of rainfall and wind field over the metropolitan area of Helsinki.
Operation of the network is demonstrated by high resolution observations of a severe storm and supercell thunderstorms, that hit Helsinki metropolitan area on 8th August 2010.

Paper 159:

Past and projected future changes of North Atlantic polar low frequency

Matthias Zahn, University of Reading, United Kingdom


Polar lows are vigorous and fast developing storms threatening human offshore activities in polar waters. Due to their small spatial scale polar lows are under-represented in most global reanalysis and climate model data. By means of dynamical down-scaling it has become possible to simulate polar lows with a Regional Climate Model (RCM) and track them in the output fields of long-term RCM simulations. To investigate their long-term frequency changes over the North Atlantic a RCM was driven with NCEP reanalysis data for the past as well as with global data of IPCC future projections and polar lows were counted in the output data.

While there could no systematic change of polar low frequency be found in the past, a significant decrease in the number of annual cases was discovered associated with projected anthropogenic warming of the atmosphere. This decrease is linked to an increase in atmospheric stability with warming, which is a result of faster rising air temperatures compared to sea surface temperatures. The influence of changing large scale ocean currents on the atmospheric stability over the North Atlantic will also be discussed.

Paper 160:

Statistics of extreme areal rainfall depths based on radar observations

Edouard Goudenhoofdt, Royal Meteorological Institute, Belgium

Laurent Delobbe, Royal Meteorological Institute, Belgium


Weather radars can provide quantitative precipitation estimates (QPE) over large areas at high spatial and temporal resolution.
In this study, a long-term analysis of QPE is performed.
Raw volume data from two C-band weather radars have been archived since 2002 and 2004.
This allows a careful processing of the reflectivity data to derive surface rainfall (e.g. ground echoes removal, application of a vertical profile of reflectivity).
To further improve QPE, radar-based estimates will be combined with hourly raingauge measurement using the most efficient methods.
From this 10-year dataset, areal rainfall depth statistics will be computed with a focus on extreme values.
In particular the effects of the area size and duration will be investigated.
As an example, the exceedance probability of rainfall depth for several catchments will be shown.
The goal of this study is to produce useful information for the verification of regional NWP and climate models and for flood risk assessment.

Paper 161:

Operational programme for the exchange of lightning location data: feasible or not?

Antti Mäkelä, Finnish Meteorological Institute, Finland

Jussi Haapalainen, Finnish Meteorological Institute, Finland

Dieter Poelman, Royal Meteorological Institute, Belgium

Francoise Honore, Meteo France, France

Daniele Biron, Servizio Meteorologico Aeronautica Militare, Italy

Ladislav Meri, Slovak Hydrometeorological Institute, Slovakia

Jerzy Konarski, Institute of Meteorology and Water Management, Poland

Bojan Lipovscak, Meteorological and Hydrological Service, Croatia

Asko Huuskonen, Finnish Meteorological Institute, Finland

Balint Varga, Hungarian Meteorological Service, Hungary


The exchange of national weather radar data on European scale has been carried out successfully within the framework of the OPERA network. In a similar manner, lightning location data could also be shared and combined and this has actually been done over smaller regions in Europe, e.g., between some neighbouring countries. Larger areas like Europe are also covered by the European Cooperation on Lightning Detection (EUCLID) community. However, because of the commercial nature of EUCLID, the accurate data is not available free of charge in real time.

The main reason why lightning location data is more difficult to share and combine compared to weather radar data is that most of the European national meteorological services are not the owners of a lightning location systems, but are customers. This means that an OPERA-style cooperation is much more difficult to establish since it is easily hampered by commercial issues. In addition, the performance and coverage of a national lightning location system is subject to variations, making a composite more challenging.

In this study we explore the methods and requirements on how lightning location data could be shared and combined in Europe between national meteorological services. We also address the question whether such “European lightning composite” is even feasible in practice.

Paper 162:

Synoptic-scale characteristics in European derecho events

Christoph Gatzen, Meteogroup Deutschland, Germany


For this work, severe wind events from 1989 to 2012 are analyzed to classify central European derecho events. For each derecho, a systematic analysis of the synoptic-scale situation is performed. This analysis is based on an ingredients-based methodology for organized deep moist convection. Accordingly, the synoptic-scale flow is analyzed with respect to the boundary-layer moisture, lapse rates, vertical wind shear, and synoptic-scale forcing mechanisms. Additionally, the position of synoptic-scale features like low- and mid-level jet streaks, or low-level moisture fields are analyzed with respect to the path of derecho events. Based on this analysis, the events are grouped to clusters that represent the spectrum of synoptic situations observed. Furthermore, this analysis is done for the intensification, mature, and dissipation stage of derechos separately. A comparison to American derecho events is done.

Paper 163:

Sustainable Natural Hazzard Mitigation.

Ssebudde Simon Peter, Rural-Urban Environmental Agency (RUEA), Uganda

Faridah Nakanwagi, Rural-Urban Environmental Agency (RUEA), Uganda

Elisha Nambi, Makerere University, Kampala-Uganda., Uganda


Employment of sustainable development principles to natural disaster mitigation is explored in this text. Three main domains are examined namely; land-use planning and policies; shelter design, building materials and construction methods; and institutional organization at local, regional, national and international levels. These three domains are illustrated on the basis of experiences of human settlements and housing of the rural poor in general. Taking into consideration the scale of the problem and the variety of conditions, the most pressing issues are identified along with the different remedies and the major areas for policy intervention. However, transferring these ideas into implementation strategies, in which creative combinations of solutions, priorities, timeframes and resources are to be identified will depend on a particular disaster situation and obviously cannot be carried out without detailed examination of the circumstances. Adjustments and changes are proposed to the ways in which human settlements are shaped, grown and managed in order to ensure harmonious interactions between natural and human systems, so that vulnerability to natural disasters is minimized.

Paper 164:

Validation of satellite-based CI detection of convective storms via backward trajectories

Felix Dietzsch, Leibniz institute for tropospheric research, Germany

Fabian Senf, Leibniz institute for tropospheric research, Germany

Hartwig Deneke, Leibniz institute for tropospheric research, Germany


Within this study, the rapid development and evolution of several severe convective events is investigated based on geostationary satellite images, and is related to previous findings on suitable detection thresholds for convective initiation. Nine severe events have been selected that occurred over Central Europe in summer 2012, and have been classified into the categories supercell, mesoscale convective system, frontal system and orographic convection. The cases are traced backward starting from the fully developed convective systems to its very beginning initial state using ECMWF data with 0.5 degree spatial resolution and 3h temporal resolution. For every case the storm life cycle was quantified through the storm's infrared (IR) brightness temperatures obtained from Meteosat Second Generation SEVIRI with 5 min temporal resolution and 4.5 km spatial resolution. In addition, cloud products including cloud optical thickness, cloud phase and effective droplet radius have been taken into account. A semi-automatic adjustment of the tracks within a search box was necessary to improve the tracking accuracy and thus the quality of the derived life-cycles. The combination of IR brightness temperatures, IR temperature time trends and satellite-based cloud products revealed different stages of storm development such as updraft intensification and glaciation well in most casesconfirming previously developed CI criteria from other studies. The vertical temperature gradient between 850 and 500 hPa, the Total-Totals-Index and the storm-relative helicity have been derived from ECMWF data and were used to characterize the storm synoptic environment. The results suggest that the storm-relative helicity also influences the life time of convective storms over Central Europe confirming previous studies. Tracking accuracy has shown to be a crucial issue in our study and a fully automated approach is required to enlarge the number of cases for significant statistics.

Paper 165:

Nowcasting system at Danish Meteorological Institute

Claus Petersen, Danish Meteorological Institute, Denmark

Ulrik Korsholm, Danish Meteorological Institute, Denmark


There is an increasing use of weather data to optimize different systems such as management of wind and solar power. Air traffic and winter maintenance of roads, management of waste and rain water and ordinary forecasts to warn the public of dangerous weather are other types of use of weather data. The focus in nowcasting is both on manual monitoring and on frequent and very accurate short range forecasts as a tool for forecasters and for use in automatic systems.
For these reasons a dedicated nowcasting model setup have been developed at Danish Meteorological Institute (DMI) . The HIgh Resolution Limited Area Model (HIRLAM) is the core in the system. Every hour a traditional surface analysis and 3D variational analysis are done.
The keystone in the system is a dense network of surface stations which together with radar and satellite derived cloud products are received with a frequency of 10 minutes. These frequent and high resolution data are the primary justification for running data-assimilation and a Rapid Update Cycle weather model every hour. The near future plan is to run the model 2-4 times every hours or with the same frequency as new observations are received. The model provides 12 hours forecast of relevant weather parameters with output every 10 minutes so the forecast model can mimic the radar and satellite input.
The system has been tested on individual cases and is now running in real time for testing. Results from test cases and verification of the model will be presented.

Paper 166:

Satellite Based Climatology of (Sub-) Tropical Cylcones in Europe

Alois M. Holzer, ESSL, Austria

Pieter Groenemeijer, European Severe Storms Laboratory, Germany


In the waters surrounding Europe, a fully developed tropical cyclone has rarely been documented. However, in the Mediterranean Sea, and Black Sea, and in the Atlantic Ocean southwest of the Iberian Peninsula, storms that reveal a high level of subtropical or tropical cyclone type organization can be observed on satellite imagery. These storms reach storm strength and pose a threat to shipping, tourist industry and the coastal communities. Moreover, their frequency and intensity may change within a changing global climate.
Before this study, a sound objective baseline climatology of such systems, or methods to obtain such a climatology, were lacking in Europe - even though this would have direct applications to risk management, e.g. on behalf of shipping companies. To obtain such climatology, the manual Dvorak-Method was applied to Meteosat 1 imagery, back to the onset of the operational meteorological satellite era in Europe.
This work presents the outcomes, researching (sub-)tropical cyclones in the waters surrounding Europe. For these systems we assessed their
i) climatological frequency,
ii) geographical distribution and
iii) typical characteristics in various channels, using METEOSAT satellite imagery.
According to the Dvorak method suspect systems were classified as (sub-) tropical depressions, storms or cyclones via assigning T-numbers (T for tropical) related to the satellite appearance of the system. The goal was to identify every (sub-) tropical storm while digging through the archive. In a second step the single cases were studied in depths.
Climatology was finally set up, containing the general number of systems in each of the three basins, their seasonal occurrence, mean peak intensity, peak intensity distribution, and lifetime.
We will not only present the data, but also examples of imagery and the most prominent storms.

Paper 167:

Tornadogenesis in a high-resolution simulation of the 8 May 2003 Oklahoma City tornadic supercell.

Alexander Schenkman, Center for Analysis and Prediction of Storms, United States

Ming Xue, University of Oklahoma, United States

Ming Hu, Earth System Research Laboratory, United States


On 8 May 2003, a supercell thunderstorm produced a strong, long-track tornado that caused F-4 damage on the south side of Oklahoma City. A 50-m grid spacing simulation of this tornadic supercell was conducted using the Advanced Regional Prediction System (ARPS) model in order to determine the important processes associated with the tornado and tornadogenesis. This simulation is nested within a lower-resolution model simulation that was initialized via 3DVAR data assimilation. The high-resolution simulation produces two tornadoes that track within 10 km of the location of the observed long-track tornado. The stronger of these two tornadoes reaches high-end F-3 intensity.
The development of both simulated tornadoes is analyzed in great detail in order to determine the processes responsible for tornadogenesis. This analysis reveals that tilting of low-level vorticity generated by surface drag plays an important role in the origin of vertical vorticity near the ground for both tornadoes. A simulation is run without surface drag and is found to be considerably different from the simulation with drag included. A tornado still develops in the no-drag simulation, but it was shorter-lived and took a substantially different track than the tornadoes in the drag simulation. Tilting of environmental vorticity in an outflow surge was determined to be the most likely cause of the tornado in the no-drag simulation. Baroclinic vorticity generation was found to be unimportant in the development of the tornadoes in both the drag and the no-drag simulation. Implications and most recent analysis of tornado and tornadogenesis dynamics in the simulation will be presented.

Paper 168:

Development of a multilayer perceptron based artificial neural network model to asses the intensity of mesoscale convective systems from reanalyzed data set

Sanjay Sharma, Kohima Science College, India

Devajyoti Dutta, Kohima Science College, India


An attempt has been made to develop a Multilayer Perceptron (MLP) based Artificial Neural Network (ANN) model to asses the intensity of mesoscale convective systems (MCSs) by using NCEP/NCAR reanalyzed data i.e. (i). Convective Available Potential Energy, (ii). Convective Inhibition, (iii). Vertical wind shear at lower heights (1-5 km), (iv). Vertical wind shear at higher heights ( >5 – 9 km) and (v). precipitable water at lower height. Rain echo top height at 20 dBZ (ETH20dBZ) from Tropical Rainfall Measuring Mission -Precipitation Radar is considered as a proxy for intensity of MCSs. MCS with ETH20dBZ ≥ 15 km is considered as severe. The model is developed for a data set over north eastern part of India. For this purpose, MLP network is trained with the help of atmospheric parameters as an input and ETH20dBZ as a target. Training is carried out by using back propagation algorithm. A network of 5:10:20:1 (5 nodes in input layer, 10 nodes in 1st hidden layer, 20 nodes in 2nd hidden layer and one node in output layer) is found to be optimum. With the help of developed weight matrices, ETH20dBZ is computed for a given input data set. Estimated values of ETH20dBZ are compared with the observed values. Error statistics are studied in terms of bias (0.51 km), root mean square error (3.44 km) and correlation coefficient (0.59 ). By using the developed model, Probability of Detection (POD) of severe MCSs is found to be 85%. Further work is in progress to take into account ETH40dBZ as an additional convective proxies to detect hail features in MCSs.

Paper 169:


Eric Defer, CNRS-LERMA, France

Sylvain Coquillat, LA, France

Jean-Pierre Pinty, LA, France

Serge Soula, LA, France

Jean-Michel Martin, LA, France

Serge Prieur, LA, France

Evelyne Richard, LA, France

William Rison, NMT, United States

Paul Krehbiel, NMT, United States

Ronald Thomas, NMT, United States

Daniel Rodeheffer, NMT, United States

Veronique Ducrocq, MeteoFrance, France

Olivier Bousquet, MeteoFrance, France

Odile Roussot, MeteoFrance, France

Laurent Labatut, MeteoFrance, France

Thomas Farges, CEA, DAM, DIF, France

Christian Vergeiner, TU Graz, Austria

Wolfgang Schulz, OVE, Austria

Graeme Anderson, UK Met Office, United Kingdom

Stephane Pedeboy, Meteorage, France

Hans-Dieter Betz, nowcast, Germany

Kostas Lagouvardos, NOA-IERSD, Greece

Pascal Ortega, UPF, France

Gilles Molinie, LTHE, France

Patrice Blanchet, ONERA, France


The PEACH (Projet en Electricité Atmosphérique pour la Campagne HyMeX) project is the Atmospheric Electricity component of the HyMeX (Hydrology cycle in the Mediterranean Experiment) experiment. PEACH aims at measuring and analyzing the lightning activity and electrical state of thunderstorms over the Mediterranean Sea. During the SOP1 (Special Observation Period; September-October 2012), records of four European operational lightning detection networks (ATDNET, UKMO; EUCLID; LINET, nowcast; ZEUS, NOA) and the NMT Lightning Mapping Array (LMA) were used to document the total lightning activity over South-Eastern France. Other research instruments such as induction rings (LA), electric field measurements (OVE; LA; NMT), video camera (OVE; ONERA), microbarometer and microphone arrays (CEA) were deployed to document the properties of the lightning flashes as well as the electrical state of parent thunderclouds. These observations are used to characterize the evolution of the electrical activity during the life cycle of SOP storms in conjunction with microphysics and kinematics description of the parent storms as derived from ground-based radar, ground-based and airborne in situ observations. Interpretation of the results are supported by the use of cloud models (MM5; MESO-NH with electrification and lightning schemes).
We will discuss some of the recorded cases at both flash and storm scales, including the 14 October 2012 tornado case that occurred North of Marseille. We will also introduce some of the products that will be made available to the HyMeX Community.

Paper 170:

Meteorological Causes of Flashflood in Pila village (Slovakia) on 07/06/2011.

Martin Benko, Slovak Hydrometeorological Institute, Slovakia

Norbert Polčák, Slovak Hydrometeorological Institute, Slovakia

Martina Sadloňová, Slovak Hydrometeorological Institute, Slovakia

Paulína Valová, Slovak Hydrometeorological Institute, Slovakia


Several villages on the Eastern foothill of Male Karpaty mountains were affected by flashflood, especially Píla village, in the afternoon on 7th June 2011. The first intensive rainfall was observed in Modra - Piesky station over 12-12:30 UTC, immediate intensity was of 100 to 140 mm/h and during next 3 hours 65 mm of total amount of precipitation was observed. The above mentioned flashflood was caused by a few thunderstorm cells initiated by instability in warm wet air mass and supported by upward motions on the slopes of Male Karpaty mountains. Meteorological conditions in the atmosphere just before the flashflood occurrence are analyzed in the short case study. For instance following parameters have been analyzed: pressure field, air temperature and relative humidity in different levels, wind field in lower levels of the troposphere and vertical wind shear, and last but not the least stability indexes as well.

Paper 171:

TRUSTED SPOTTER NETWORK AUSTRIA – New Developments and Applications at ESWD

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

Barbara Chimani, ZAMG, Central Institute for Meteorology and Geodynamics, Vienna, Austria, Austria

Konrad Türk, ZAMG, Central Institute for Meteorology and Geodynamics, Vienna, Austria, Austria


The TRUSTED SPOTTER NETWORK AUSTRIA (TSN) constitutes the collaboration 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).

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. With TSN the reliability of the information for operational forecasters had been improved significantly.
For this purpose ZAMG offers an individual training program, regular workshops and scientific support. This program for collaborating TSN spotters and storm-chasers seems to be unique among European national weather services. It also consists of job-shadowing, which is obligatory for each individual TSN member and is following a comprehensible and standardized procedure provided at every regional centre in Austria.

Further, the activity of a “trusted spotter” is facilitated by real time weather information from ZAMG, easily accessible via the internet.

Additionally, a web based, real time platform allows displaying ESWD data base reports for the operational forecaster at ZAMG. These reports are strictly following the ESWD data format and threshold guidelines. A trusted spotter is only allowed to give information from personal experience, secondary or forwarded information should be excluded. Therefore, reports from “trusted spotters” are accepted by ESWD with the higher QC1 clearance.

In order to use available potential for further collaboration between ZAMG and Skywarn Austria a joint web page for extreme weather reports was constructed. In this way, ESWD criteria are now also applied for reports from the general public.

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.

In a climatological study about extreme precipitation events (>40mm/6hrs, since 2000) the frequency of reports in ESWD was analyzed in relation to the growth and development of the Trusted Spotter Network. First results of this analysis will be shown.

Paper 172:

Structure and Formation Mechanisms of Two Adjacent Shear Lines Accompanied by Wind Gusts in the Japan Sea Coastal Region during a Cold-Air Outbreak

Wataru Mashiko, Meteorological Research Institute/JMA, Japan

Hanako Y. Inoue, Meteorological Research Institute/JMA, Japan

Syugo Hayashi, Meteorological Research Institute/JMA, Japan

Kenichi Kusunoki, Meteorological Research Institute/JMA, Japan

Syunsuke Hoshino, Aerological Observatory/JMA, Japan

Kenichiro Arai, East Japan Railway Company, Japan

Kenichi Shimose, National Institute of Advanced Industrial Science, Japan

Masako Kusume, Alpha-denshi Co., Ltd./MRI, Japan

Masahide Nishihashi, Alpha-denshi Co., Ltd./MRI, Japan

Hiroshi Yamauchi, Meteorological Research Institute/JMA, Japan

Osamu Suzuki, Japan Meteorological Agency, Japan

Hiroyuki Morishima, East Japan Railway Company, Japan


Two adjacent shear lines accompanied by wind gusts exceeding 20 m s−1 passed over the Shonai region on the Japan Sea coast during a moderate cold-air outbreak on 12 December 2010. The structure of the shear lines were examined using simulation results and observational data including high-frequency upper-air soundings. Simulation results revealed that cold-airmass outflow from the Eurasian Continent had a split-flow pattern due to the blocking effect of the Changbai Mountains. One is from the south side of the mountains (SCF), and the other is from the north side (NCF). NCF also has two streams: a shallow southward-meandering flow on the south and a slightly deeper straight flow on the north. A shallow shear line was formed between SCF and the shallow southward-meandering flow of NCF. The deeper straight flow of NCF sharply formed another shear line on the north at around 1-km height with SCF. This result implies that the shear lines accompanied by wind gusts on the Japan Sea coast during a coldair outbreak have various structures and that the horizontal wind shear a little above the surface could cause surface wind gusts.

Paper 175:

Applying Mean-Shift – Clustering for 3D object detection in remote sensing data

Jürgen-Lorenz Simon, Meteorological Institute Bonn, Germany

Malte Diederich, Meteorological Institute Bonn, Germany


The prediction skill of numerical weather prediction (NWP) in is extremely limited in the range from 0-3 hours for a number of reasons, which gave rise to a number of other approaches to close the gap to NWP now often subsumed under the name NOWCASTING. New NOWCASTING prediction techniques are under development in the Hans-Ertel – Zentrum. The approach of Themenbereich 1 (HErZ-TB1) employes an object-detection scheme for identifying convective cells at an early stage in a 2D/3D satellite/radar composite. The object identification (clustering) is performed using a technique borrowed from the rule book of computer vision, called mean-shift clustering. It’s application on remote-sensing data, exemplified by our composite, will be presented.

Paper 176:

Reanalysis of the fourth-deadliest Tornado in European history

Alois M. Holzer, European Severe Storms Laboratory, Austria

Mathias Stampfl, ESSL, Austria

Thomas Schreiner, ESSL, Austria

Pieter Groenemeijer, European Severe Storms Laboratory, Germany


On the 10th of July 1916 a large tornado passed over the northern neighbourhoods and industrial zones of the town of Wiener Neustadt, Lower Austria. This resulted in 32 fatalities and 328 wounded people among massive structural damage to residences and industrial buildings. According to the European Severe Weather Database (ESWD) this event is the fourth-deadliest tornado in Europe’s currently known history.
We performed a careful reassessment of this historic event within a project TORNeustadt, on order of the City of Wiener Neustadt, with the aim to integrate the results into the town’s civil emergency plan. More precisely, we collected and reanalysed historical material, including more than 100 damage pictures and performed a damage assessment based on the available damage images and descriptions. This was done by assigning an F-/T-Scale rating to every damage indicator based on the respective degree of damage. The individual ratings were used to compose a detailed map of the damage path across the city and allowed us to determine the tornado’s maximum intensity
While an earlier estimate of the maximum F-Scale rating of the Wiener Neustadt tornado of 1916 was F3, the reanalyses resulted in an upgrade to F4, based on several damage indicators in the area of a locomotive factory, where brick walls with a thickness close to one metre collapsed. In addition, we have estimated the tornado risk for the entire Vienna basin based upon climatological ESWD data, including the reanalyzed Wiener Neustadt case of 1916.
In our presentation we focus on our methodology and present the results on maps and in statistics.

Paper 177:

A derecho in northeastern Europe on 8 August 2010

Anniina Törmä, Finnish Meteorological Institute, Finland

Jenni Rauhala, Finnish Meteorological Institute, Finland

Andris Viksna, Latvian Environment, Geology and Meteorology Center, Latvia


On 8 August 2010, a mesoscale convective system moved over northeastern Europe causing 950 km long path of wind damage. The storms started over Belarus and Lithuania and over Latvia formed a large scale bow echo that moved northward over Estonia to Finland. Damage was mainly fallen trees and damaged buildings and mostly of F1 intensity. The strongest measured wind gust was 36.5 m/s observed in Estonia, but also both in Latvia and Finland above 26 m/s wind gusts were reported in several locations.
During the event northern Europe was in a vast surface low pressure area of which center was over southern Baltic Sea allowing flow of warm and very moist continental low-level air into Baltic countries and Finland from southeast. A weak northward moving 500-hPa short-wave trough influenced the storm environment by both increasing the vertical wind shear and by increasing the instability. The low-level wind shear was increased by the northwestward oriented low-level jet. The mesoscale convective system developed close to a frontal boundary and then moved in a direction nearly parallel to the boundary. It had features commonly described with progressive derechoes. This study summarizes both the weather conditions leading to this extreme event and the observed damage path over the affected countries.

Paper 179:

Experimental forecasting of severe storms in Europe: a summary of the first ESSL Testbed

Pieter Groenemeijer, European Severe Storms Laboratory, Germany

Alois Holzer, European Severe Storms Laboratory, Austria

Georg Pistotnik, European Severe Storms Laboratory, Austria

Kathrin Riemann-Campe, European Severe Storms Laboratory, Germany


From 4 June to 6 July 2012, the first ESSL Testbed took place at the Research and Training Centre of the European Severe Storms Laboratory in Wiener Neustadt, Austria. During this time, researchers and forecasters worked closely together putting new forecast supporting products to the test.

The Testbed's main activity is to prepare experimental forecasts for severe weather for day 1, 2, 3, 4 and 5 as well as “nowcasts” for the following 2 hours using the available Testbed tools and standard meteorological data. Subsequently, a verification of these forecasts is performed using the European Severe Weather Database, followed by an evaluation of forecasting tools and techniques.

Inspired by the annual Spring Program at NOAA's Hazardous Weather Testbed (HWT), the ESSL Testbed has a stronger focus on forecaster training than the HWT. Given the various backgrounds of the participants, an important Testbed goal is to acquaint its participants with severe weather forecasting methods and techniques that work universally. Among the tools that were evaluated were visualizations of high-resolution ensemble NWP, the satellite-based cloud top cooling and overshooting top detections, lightning detection, ECMWF's Extreme Forecast Index, and cell-tracking algorithms.

In daily “Expert Lectures” that were broadcast online to remote participants, researchers provided background information on their products and internationally renowned experts in forecasting presented their viewpoints on storm forecasting and its scientific roots.

Organized by ESSL in close cooperation with the Austrian Central Institute for Meteorology and Geodynamics (ZAMG), the Testbed was supported - among others - by the German Weather Service (DWD), EUMETSAT, WMO, ECMWF, VAISALA, and the GOES-R programme, providing products for evaluation and real-time meteorological data or financial support to participants. The first edition of the European Severe Storms Laboratory's Testbed has brought together 67 forecasters and developers from 21 countries across the world, a majority of whom were associated with the European National (Hydro-) Meteorological Services.

We will report on the first edition of the ESSL Testbed, present the forecasting, verification and evaluation activities using an example case, discuss its outcomes, and present plans for the 2013 Testbed.

Paper 180:

LEWP along squall line

Robertus Groenland, KNMI, Netherlands


In this study, the structure of a series of LEWP-segments on a wintertime squall line is investigated. One of the highest convective windgusts in recent history was reported with this line (96 knots). 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.

Paper 181:

The analysis of heavy precipitation events over small mountain catchments

Daniela Rezacova, Institute of Atmospheric Physics ASCR, Czech Republic

Petr Zacharov, Institute of Atmospheric Physics, Czech Republic

Sarka Blazkova, TG Masaryk Water Research Institute, Czech Republic


Results of the analysis of heavy convective precipitation events are presented for 7 small mountain catchments in the Jizera Mountains (border region of the Czech Republic).
Event oriented precipitation analysis considers several definitions of precipitation event which take into account threshold rainfall values and a possible interruption in precipitation duration. Meteorological analysis comprises synoptic pattern, ground temperature, wind direction, stability conditions and the precipitation antecedent the event.
Input precipitation data consist of hourly rainfalls determined from radar-based and ground measurements. Radar-based rainfalls were provided by product MERGE (operational product of the Czech hydro-meteorological service) which merges ground precipitation reports from Czech synoptic stations and rough radar precipitation. The MERGE data cover the Czech territory by rainfall values in 1km×1km grids and they are available for summer periods 2002-2012. In addition, data from 21 independent gauges were used. The gauges measured rainfall data in the catchments of interest in 2010 and several of them also in 2005-2006.
The study aims at (i) improving the gauge adjustment of radar rainfalls by catchment oriented adjustment procedure, and (ii) testing the parameters of stochastic precipitation generator used in semi-distributed model of catchment hydrology (TOPMODEL) in order to simulate asymptotic flood frequency statistics.
In the presentation we will discuss the results of the event oriented analysis for duration, mean and maximum rainfall intensity, and inter-storm period.

Paper 182:

The use of forecasted radar reflectivity with hail detection algorithm

Katerina Skripnikova, Institute of Atmospheric Physics ASCR, Czech Republic

Petr Zacharov, Institute of Atmospheric Physics ASCR, Czech Republic

Zbynek Sokol, Institute of Atmospheric Physics ASCR, Czech Republic


Severe hailstorms are rare, but can cause considerable damages. Hail prediction is a hard task of meteorological forecasting. We are testing forecasted radar reflectivity with radar-based hail detection algorithm. The NWP model COSMO is able to simulate a three-dimensional radar reflectivity field in a comparable scale with actual measurements. We will compare the results from hail detection algorithm computed on forecasted radarreflectivity and measured radar reflectivity and also with the deterministic hail forecast of COSMO model. A severe hail event, which occurred on 15 August 2010 on the Czech territory, was selected for the comparison.The COSMO model run in 1.1km horizontal resolution and with the assimilation of observed radar reflectivity. The measured radar reflectivity came from the two C-band Doppler radars operated by the Czech Hydrometeorological Institute which cover the whole Czech territory.

Paper 183:

Assessment of the European Severe Convective Storm Climatology using Reanalysis Data

Georg Pistotnik, European Severe Storms Laboratory, Germany

Pieter Groenemeijer, European Severe Storms Laboratory, Germany

Thilo Kühne, European Severe Storms Laboratory, Germany


Thunderstorms and their accompanying phenomena like large hail, severe wind gusts, tornadoes and excessive precipitation are increasingly recognized as an important hazard to life and property in Europe. Within the project STEPCLIM (“Severe Thunderstorm Evaluation and Predictability in Climate Models”), we link historic severe thunderstorm events in Europe to atmospheric conditions resolved by reanalysis data. The aim is to find a relation between quantities, which can be represented by relatively coarse climate models, and the occurrence of short-lived and local severe weather phenomena associated with convection. At a later stage, this relation will be applied to climate forecast data, so that the effects of climate change on the frequency and intensity of severe convective storms can be investigated.

A set of parameters is defined for the characterization of the local state of the atmosphere at any given point in place and time. These parameters represent the “ingredients” for severe thunderstorms, namely instability, vertical wind shear, and a measure of support for convective initiation. They are calculated from 6-hourly ERA-Interim reanalysis fields (1979-2011). The such derived fields are compared with the occurrence or non-occurrence of severe weather phenomena according to the quality-controlled severe storm reports collected in the European Severe Weather Database (ESWD).

A logistic regression is then fitted in order to find the best relation between a given combination of parameters and the associated severe weather probability. This relation is postulated to be invariant against any changes in climate, an assumption that is justified by the physics-based nature of the chosen parameters. With this restriction, the derived relation can be used to estimate future severe weather frequencies based on modeled climatic changes of the underlying parameter distributions. In this presentation, emphasis is put on the used methodology and on a way to deal with inhomogeneities of the observational data.

Paper 184:

Tornado and waterspout climatological risk for Greece

Michalis Sioutas, ELGA-Meteorological Applications Centre, Greece


Tornado and waterspouts are extreme weather events, which though rare, however can be hazardous in many aspects including economic losses for damage to infrastructure and property and a serious threat for injures and fatalities. Estimating probabilities of occurrence and risk levels is crucial for research and forecasting, damage prevention and life protection. Background tornado and waterspout data for Greece are based on an updated 14-year database. The climatological probability of tornado and waterspout days for the various Greek regions and regional unities (prefectures) is estimated and spatial and temporal threat patterns are identified. Western Greece islands and coasts exhibit the greatest tornado risk, with maxima in Elias area, northwest Peloponnesus and Corfu Island, in the Ionian Sea. Waterspout peak threat is located over north offshore of Crete Island, in the southern Aegean Sea. The annual risk cycle is different for the various parts of Greece. Northern Greece faces maximum tornado threat during summer, while southern Greece during winter. Western Greece exhibits a longer tornado seasonality, from autumn up to spring. Waterspouts develop in almost all seas with seasonal maxima in autumn.

A principal component analysis (PCA) was applied on tornado and waterspout day frequency distributions for the various Greek regional unities (prefectures). The results indicated the first component (PC1) grouping regions of western Greece, the Ionian Islands and the southern Aegean, and the second component (PC2) regions of north-central Greece. Waterspout frequency PCA analysis indicated the first component (PC1) including the north Ionian Sea and north offshore of Crete Island and the second component (PC2) areas of the north-central Aegean Sea. The total risk analysis permitted an identification of the most tornado and waterspout prone areas in Greece. Determining of tornado alleys and waterspout high risk spatial patterns indicated a large seasonal influence and differentiation.

Paper 185:

Floods in Southeastern Brazil- Observations, Simulation, Projection and Uncertainties

Iracema Cavalcanti, National Institute for Space Research, Brazil

Sin Chou, National Institute for Space Research, Brazil

Jorge Gomes, National Institute for Space Research, Brazil


Cases of floods in Southeastern Brazil are analyzed, discussing the associated synoptic features and the convection in the systems shown in satellite images. Cases of heavy precipitation in this region are identified in a climate simulation of present time and in a climate projection under global warming using the Regional Eta Model. The synoptic and mesoscale conditions, associated with heavy precipitation, simulated by the model are analyzed. The frequencies and intensities of extreme precipitation cases are obtained from four members of the simulations and projections, and an uncertainty analysis is performed with the results. One of the extreme precipitation cases is illustrated: the case that happened in Southeastern Brazil (Rio de Janeiro) from 31st December 2009 to 1st January 2010. There were several landslides in the region, one of which destroyed a beach hotel killing many people. The satellite image showed the influence of several synoptic systems in the region, as the Bolivian High displaced from its normal position, the upper level cyclonic vortex close to Northeastern Brazil coast and a frontal system extending from the region towards the ocean. The synoptic condition, with low level wind confluence over the region, high relative humidity above 80% and strong divergence at high levels favored by the displacement of the Bolivian High southeastward contributed to the heavy precipitation associated with the oceanic squall line embedded in the frontal system. A simulation of this case with high resolution of Eta Model shows the ability of the model in representing the extreme precipitation.

Paper 186:

Climate-driven increase in the variability and multi-year mean level of severe thunderstorm-related losses and thunderstorm forcing environments in the U.S. since 1970

Julia Sander, German Aerospace Centre, Institute of Atmospheric Physics, 82234 Oberpfaffenhofen-Wessling, Germany, Germany

Eberhard Faust, Munich Reinsurance Company, 80802 Munich, Germany, Germany

Jan Eichner, Munich Reinsurance Company, 80802 Munich, Germany, Germany

Markus Steuer, Munich Reinsurance Company, 80802 Munich, Germany, Germany


A substantial increase in the variability of normalised direct economic and insured severe thunderstorm-related losses in the U.S. east of the Rocky Mountains over the period 1970-2009 (March – September) has been detected. Besides variability, also the multi-year mean level of losses has strongly increased. Analysed are sizeable severe thunderstorm events causing at least US$ 250 million in normalized economic losses. The high threshold guarantees homogeneity over time, because such events regularly covered several states and thus are very unlikely to have been missed at any time due to reporting variability.

To shed light on the question whether the strong increase was driven by an external climate driver, the time series of normalized losses (annual counts and annual loss aggregate) was correlated with the time series of thunderstorm forcing environments. The latter were inferred from NCEP/NCAR reanalysis data and comprise 6-hourly CAPE and vertical wind shear data combined to form the variable Thunderstorm Severity Potential (TSP). From the notable correlation found between the time series of normalized losses and meteorological thunderstorm forcing environments (TSP) it could be inferred that climate was the dominant driver for the increase in variability and average level of thunderstorm-related losses in the period 1970-2009. An important factor in the rise of TSP over time could be identified in CAPE, as we found a substantial rise in the annual number of exceedances of a high CAPE threshold in the reanalysis data.

Recent studies imply that the changes observed in our study, particularly regarding an increase in high-level CAPE environments, are consistent with the modelled effects of anthropogenic climate change. The physical chain of climate change-driven increasing levels of specific humidity leading to rising levels of CAPE as one of the preconditions of more severe thunderstorm forcing environments has already been established by measurements and climate model experiments.

Paper 187:

CAPS Storm-scale ensemble forecasting for the NOAA HWT 2013 Spring Experiment

Fanyou Kong, University of Oklahoma, United States


The Center for Analysis and Prediction of Storms at the University of Oklahoma continues to produce realtime storm-scale ensemble forecast (SSEF) for the NOAA 2013 Hazardous Weather Testbed (HWT) Spring Experiment. The CAPS SSEF will consist of 25 members from three NWP model systems (WRF-ARW, COAMPS, and ARPS), covering the full continental United States with convection-allowing resolution at 4-km horizontal grid pacing. SSEF members are configured with a combination of initial perturbations extracted from the coarser grid Short-Range Ensemble Forecast (SREF) ensemble members and various physics options in microphysics, PBL and land-surface model, and radiation. Up to 140 Doppler weather radar data, both radial wind and reflectivity, are analyzed into the SSEF in realtime using the ARPS 3DVAR and Complex Cloud Analysis system. A wide range of ensemble products will be 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, as well as simulated synthetic GOES satellite cloud top brightness temperature will also be produced from the SSEF starting 2013 Spring. Objective evaluation of the CAPS SSEF dataset will be performed. The impact from lateral boundary perturbations on SSEF spread and skill will be investigated using two different strategies, one direct driven with perturbed member forecasts that are corresponding to IC perturbations and another with perturbations extracted. The behavior of stochastic kinetic energy backscatter perturbation in the convection-allowing high-resolution NWP will also be discussed.

Paper 188:

Flash Flood in Madeira Island in autumn 2012

Flavio Tiago Couto, Évora Geophysics Centre, University of Évora, Portugal

Rui Salgado, Évora Geophysics Centre, University of Évora, Portugal

Maria João Costa, Évora Geophysics Centre, University of Évora, Portugal


Madeira (32º75'N and 17º 00'W) is a Portuguese mountainous island with the highest peaks above 1800 m, where after the catastrophe of 20 February 2010, which has shown to be vulnerable to the flood occurrences from significant socio-economical losses. Using ground observations and the MESO-NH model, this study investigates the main atmospheric aspects related to the flash flood and landslides occurrences on 05 November 2012 in the island. The period between 25 October and 06 November 2012 have been considered in the analysis. The numerical simulations are performed using different configurations, with a single domain of 25 km horizontal resolution and using grid nesting technique with 9, 3 and 1 km resolution. During this period, two main events were chosen based on rain gauge data analysis, to be simulated at high resolution (1km): 30 October and 05 November, when the highest amounts of precipitation in Madeira's highlands and landslides were observed, respectively. Previous studies show that the heavy rainfall events in Madeira depend on both synoptic and mesoscale features, justifying the simulations in different resolutions. The accumulated precipitation simulation pointed out the prominent role of orography in the intensification of precipitation over the island. The possibility of severe weather development was confirmed from some instability indices obtained from radiosonde observations, despite the CAPE index, which indicated a weak instability environment. On the other hand, large-scale environment was characterized by the development of extratropical cyclones near the island, where the frontal structure acting over Madeira's archipelago is evident. A remarkable feature is the fact that on 30 October, coupled to a frontal system, the main moisture source in low-levels was related to an atmospheric river structure passing over the island.

Paper 189:

Lightning channels emerging from the top of thunderstorm clouds

Oscar van der Velde, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia, Spain

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

Serge Soula, Université de Toulouse III -Paul Sabatier, Observatoire Midi-Pyrénées, Laboratoire d'Aérologie, Toulouse, France, France

Nicolau Pineda, Servei Meteorològic de Catalunya, Spain


In recent years, research of transient luminous events is shifting from the rather common elves and sprites high above thunderclouds to the much less frequently observed phenomena issued by the storm cloud itself: gigantic jets (GJ) connecting to the ionosphere, and high-energy terrestrial gamma-ray flashes (TGFs) recorded at spacecraft. These phenomena are observed more often at tropical latitudes, and a link may or may not exist between the two. Both likely share the requirement of high-altitude leaders of negative polarity, which in the case of a GJ escapes from the cloud top and transforms into a long streamer discharge. While this should be easier at lower air densities (higher altitude), studies showed GJs need not be produced by storms with the highest tops. TGFs have still unclear origins, but may be related to production in negative leaders or other regions with strong vertically directed electric fields by runaway electron mechnisms.
In December 2009, a GJ was observed in the Mediterranean Sea region. During the same night, a nearby storm produced repeatedly multiple leaders piercing through the cloud top, without any sign of streamers reaching higher altitudes (unlike jets or starters). Similar observations of upward cloud-to-air lightning have been obtained recently by low-light cameras over storms near the Catalonian coast in different seasons. The production conditions are currently being investigated, with a focus on optically determined altitudes of lightning and evolution of storm tops (and their temperature level). The initial impression is that cloud flashes escape into the air above during stages when the growing convective cloud top is very close to the main charge production region. Upward cloud-to-air lightning has also been mapped by the Ebro Lightning Mapping Array, exhibiting inverse bolt-from-the blue characteristics, and as a by-product of a bolt-from-the-blue lightning strike to ground, recorded by a high-speed camera.

Paper 190:

Severe Convective Storms in the European Societal Context

Charles Doswell Iii, Doswell Scientific Consulting, United States


Perceptions of the severe convective storm threat and the capacity to deal with that threat are strongly influenced by the social setting in which the storms occur. The social context for severe convection in Europe is distinctly different from that in the United States of America (USA), so a comparison between the two reveals certain factors that govern the responses to severe convective storm events. Although perceptions of the threat are different in Europe than in the USA, the objective hazards are more similar than they are widely understood to be. The development of infrastructure for dealing with the hazards necessarily depends on changing the existing perceptions and breaking down certain barriers to the creation of appropriate infrastructure.

Paper 191:

Verification of multi-categorical thunderstorm forecasts at the European Storm Forecast Experiment over the period 2009-2013

Oscar van der Velde, Lightning Research Group, Electrical Engineering Department, Technical University of Catalonia, Spain

Pieter Groenemeijer, European Severe Storms Laboratory, Germany

Helge Tuschy, Deutscher Wetter Dienst, Germany

Christoph Gatzen, MC-Wetter, Germany

Johannes Dahl, Convective Storms Group, North Carolina State University, United States

Oliver Schlenczek, Cloud Microphysics Group, Johannes Gutenberg Uni Mainz, Germany

Tomas Pucik, Czech Hydrometeorological Institute, Czech Republic

Marko Korosec, DARS, Slovenia

Georg Pistotnik, European Severe Storm Laboratory, Austria

Thilo Kühne, European Severe Storm Laboratory, Germany


The European Storm Forecast Experiment (ESTOFEX) has been making forecasts of thunderstorms and accompanying severe weather for most of Europe for more than 10 years. The forecasts consisted of three severe weather threat lines as well as one thunderstorm line. The latter allowed only a “yes/no” forecast of thunderstorms in an area. The verification of these dichotomic lightning forecasts was presented at ECSS 2007. In mid 2009 ESTOFEX switched to the use of two thunderstorm probability lines, tentatively marked “15% probability” and “50% probability” of thunderstorms within 40 km from each location. Treating the areas between the lines as three categories separating areas with very low, medium and high probability, the present verification aims to: 1) quantify the observed mean frequency of occurrence in each class over the entire region, by counting for all forecasts all grid points with and without lightning in each of the three probability categories. 2) quantify the observed frequency of lightning per grid cell over time when it was included in each class, which allows mapping of biases. 3) study the overlap in probability classes by histograms of the number of grid cells per interval of percentage of lightning occurred over the times it was included in a probability class. 4) study seasonal performance and 5) the consistency among forecasters. Similar procedures can be used for the verification of threat levels 0, 1, 2 and 3 for convective severe weather phenomena (damaging wind gusts, large hail, tornadoes and excessive precipitation) using the ESSL European Severe Weather Database.

Paper 192:

Hail occurrence in Poland (1966-2006)

Zuzanna Bielec-Bakowska, Silesian University,Department of Climatology, Poland


The main aim of this study was to present results of research into the occurrence of days with hail in Poland and then on the relationship between hail fall occurrence and synoptic situations. The analysis was based on meteorological observations from 23 stations in Poland in the years 1966-2006 (from April to September). In the second part of the study the types of synoptic situations elaborated by T. Niedzwiedz and synoptic maps were also used.
As a result, it was determined that there is strong regional variation, both annually and in the long-term variability of hail occurrence. The annual pattern of hails occurrence seems to be its most significant characteristic. A springtime maximum dominates at stations in the north and in the western half of the country, while a far more even distribution with a maximum shifted towards the summer season is found in the south of Poland. The study identified no clear trend in the number of days with hails during the analysed period. The long-term record of days with a hails was very even at most of the stations. It is also difficult to find periods of distinctively increased or decreased number of days with hail at most of the stations simultaneously.
The results obtained in the study reveal that hail fall events most often accompanied cyclonic situations and the most favourable situations for the hail fall were occurrences of a centre of low pressure over Poland and northern cyclonic situations. The relationships were different at various stations. Stations located in south part of Poland stood out the most, because of strong influence of orography. The investigation of relations between hail fall events and synoptic situations also shows that about 60% cases of hail fall were connected with atmospheric fronts crossing over Poland. Hail events of inter air mass type frequently occurred on the edges of high-pressure areas in an unstable air masses: after a cold atmospheric front passing or accompanied convergence zones.

Paper 193:

Forecast of Severe Weather in Ukrainian Hydrometeorological Centre based on different kind of meteorological information

Inna Sobchenyuk, Ukrainian Hydrometeorological Center, Ukraine


Monitoring of severe weather in the Ukrainian Hydrometeorological Centre (UHC) is based on a
combination of different kind of meteorological information – radar, satellite and ground data. In
our operational work, we try to provide the most important and typical details, such as synoptic
situations and other precursors that could be useful to us to predict developing convective
clouds. Since the convective activity has a high frequency, one of the most important issues is to
compare the current weather patterns with satellite data. This approach would help forecasters
in different cases to distinguish between the stage, the intensity of convective clouds and the
possibility of the actual appearance of convective storms.
Satellite data began to play an extremely important role for the detection and monitoring of
severe convective clouds, when EUMETCast station was installed in UHC in 2007. A big
variety of RGB images from MSG are generated operationally, which provide much better
meteorological interpretation of satellite images. One of the critical point in monitoring of severe
weather is a time slot of image generation, which on MSG is 5 and 15 min that can provide
detection even rapid developing thunderstorms. Special feature of convective storms - is a “cold
ring”. Monitoring of convective storms based on a “cold ring” includes many different aspects
- the area of distribution, the temperature difference, the speed of development, brightness
temperature and so on.
Analysis of severe weather during the time period from May to October 2012 is discussed. Some
cases of extreme convective activity, with people death are discussed in details.

Paper 194:

Evaluation of Stability Indices for Thunderstorm Prediction in Mendoza Argentina

Carlos Bustos, Dirección de Agricultura y Contingencias Climáticas, Gobierno de Mendoza, Mendoza, Argentina, Argentina

Hugo Videla, Dirección de Agricultura y Contingencias Climáticas, Gobierno de Mendoza, Mendoza, Argentina, Argentina

Jorge Santos, Instituto de Ciencias Basicas. Universidad Nacional de Cuyo. Mendoza Argentina, Argentina

Diego Araneo, Programa Regional de Meteorología IANIGLA-CCT-Mendoza, Argentina, Argentina

Silvia Simonelli, Programa Regional de Meteorología IANIGLA-CCT-Mendoza, Argentina, Argentina


The Province of Mendoza, Argentina, has been described as a hot spot in severe convection. Many thermodynamic parameters and indices are currently being used as thunderstorm predictors because of their high correlations with the onset and development of convection. Some of these indices have been developed for one specific area and their forecasting accuracy has generally been assessed in that zone and not in others. It is a highly intriguing question whether there are parameters or indices that may function adequately as thunderstorm predictors, as far as the Probability of Detection (POD) is concerned, irrespective of the latitude of the study zone.
The effectiveness of a number of forecasting indices for non-frontal thunderstorm activity has been investigated for Mendoza. The indices include the Total Total index (TT), the Showalter index (SHI), K index (KI), Lifted index (LI), S index (SI) and CAPE.
The best parameters were evaluated by using two methods: skill scores and discriminant analysis. For every thermodynamic parameter, six skill scores were derived, but only the values of the TSS (True Skill Statistics) were considered to verify the success of the thunderstorm prediction
In order to assess the occurrence or non-occurrence of storms in the study zone additional data were used from an S-band radar network in the province of Mendoza.
For the decision whether a thunderstorm day was expected or not, the best results were obtained with the original SHI which showed the highest TSS skill score (0.506), POD= 0.85. The guidance value for SHI was determined to be 4.3. In this case thunderstorms were predicted as long as this value was not exceeded. The next highest TSS skill score for the same prediction was reached by using the TT with the guidance value 44.7, TSS(0.49) and POD (0.73). The CAPE was not found to be a successful predictor of air mass thunderstorms TSS (0.418), POD (0.67).

Paper 195:

Characteristics of Drop Size Distribution observed by dense disdrometer network and Ku band radar in Tokyo metroporitan area during TOMACS.

Sadao Saito, Meteorological Research Institute, Japan

Cyusei Fujiwara, Meteorological Research Institute, Japan

Kenichi Kusunoki, Meteorological Research Institute, Japan

Eiichi Sato, Meteorological Research Institute, Japan


During warm season, strong isolated thunderstorms are often occured and it causes floods or some rainfall-related hazards over Kanto Plain. Tokyo Metroporitan Area Convection Study (TOMACS) has been designed for better understanding for mechanisms of thunderstorms.
For this purpose, we developed dense network of surface observation stations(Surface-network) and Ku-band fast scanning Doppler radar(Ku-radar).
Surface- network has 12 observation stations with these spatial interval is about 3 km and each station contains WXT520(vaisala) and parsivel(OTT). Ku-radar has its observation range 20 km and it makes volume scan in 1 minutes. So, from both instruments spatial and temporal high structure of thunderstorm can be observed.
In this study, I will show characteristics of DSD of several isolated thunderstorm cases with Surface-network and Ku-radar data.

Paper 196:

Heavy Rain Forecast Improvement with Radar Data Assimilation

Seyda Tilev Tanriover, Istanbul Technical University, Turkey

Abdullah Kahraman, Turkish State Meteorological Service, Turkey


Number and impact of natural disasters are increasing worldwide, within the last 50 years. The International Database reports indicate that 30 natural disasters occured per year in 1950s and more than 400 in 2000s in average, besides, number of effected people has risen from about 25 millions per year in the 1960s to 300 millions since 2000. The main cause of this trend is the increasing number of climate related disasters. Floods are responsible for 40% of all natural disasters and half of disaster related loss of lives all around the world. Furthermore, most flood causalities are due to flash floods. As a concequence quantitative precipitation forecasting is an ongoing issue, especially for early warnings of flash floods caused by severe convective storms.
Although current nonhydrostatic mesoscale models have fine resolutions with the order of 1 km, operational data which are used to initialize these models have resolutions with the order of 10 km. Since numerical weather prediction is an initial value problem, to improve the forecasts of severity, location and timing of these mesoscale events, all available data must be used to initialize the models. Doppler radars provide reflectivity and radial wind data with very high spatial and temporal resolutions.
In this study, the impact of radar data assimilation on short range precipitation forecasts is enquired for an excessive precipitation event, resulting flash floods with 31 causalities on 8-9th of September 2009 in Istanbul. A control run without assimilation, and a run with assimilation of Istanbul Doppler radar data operated by Turkish State Meteorological Service have been conducted. Advanced Research WRF is used as atmospheric model and three dimentional variational technique as the assimilation method. Results indicate that assimilation of radar data improves the forecast, verified against precipitation observations and radar reflectivity.

Paper 197:

Application of radar observation data to predict a landslide due to localized heavy rain

Seok-Hwan Hwang, Korea Institute of Construction Technology, Korea, Republic of

Sanghun Lim, Korea Institute of Construction Technology, Korea, Republic of

Dong-Ryul Lee, Korea Institute of Construction Technology, Korea, Republic of

Dae Heon Ham, Korea Institute of Construction Technology, Korea, Republic of

Kyotaek Hwang, Korea Institute of Construction Technology, Korea, Republic of


During the heavily rainy season in Korea, typically between June and September, landslides due to localized torrential downpours have been often occurred. On July 28 2011, a severe landslide at Mountain Umyeon located in Seoul came about, which was directly caused by the tremendous amount of concentrated precipitation for an hour. Debris flow generated by the landslide resulted in losses of both life (18 people) and property. At that time, the ground was already weakened by heavy rain for 15 hours with total amounts of 230.0 to 266.5 mm. The localized heavy rainfall was assessed as a main cause of the landslide since maximum precipitation occurred concurrently with the peak runoff in the area where time of concentration was less than 5 minutes.
In Korea national standards of landslide watch and warning provided by the Korea Forest Service, are based on hourly, daily, and total cumulative rainfall obtained from ground rain gauge. However, this standards, which were established 25 years ago, are not sufficient to efficiently predict a landslide especially caused by the recent concentrated precipitation. Moreover, a ground rain gauge has a limitation to forecast localized rainfall because of low network density and difficulties to predict movement of rain front. Considering climatic, topographic, and anthropogenic conditions in Korea, shorter than 1-hr prediction of flash floods in urban and mountainous areas are crucial to reduce loss of lives and properties. Therefore, the very short forecasting using radar observations have lately become as an important issue.
In this study, application of radar observation data to predict a landslide due to localized heavy rain was examined for Mountain Umyeon in July 2011 case. We also compared occurrence time of the landslide using ground gauge measurements and radar rainfall estimations to investigate utility of the radar data for warning lead time.

Paper 198:

Experiences in using simulator for convection and nowcasting training

Vesa Nietosvaara, EUMETSAT, Germany

Kristina Petraityte, EUMETSAT, Germany


EUMETSAT supports various courses intended for operational forecasters. One of such courses is the NOMEK course, developed jointly by the Nordic meteorological services.

In recent NOMEK courses, we have used EUMETCAL Weather Simulator. The participants have been challenged to work on an intense convective storm case, using the available archived data. While the clock is running, the forecasters work on analyzing and monitoring the situation, preparing timely nowcasts and issuing warnings.

To create such a simulation we need to retrieve and edit the necessary data from the case and arrange it in the simulator. In the actual exercise the forecasters will get the new data almost in the same way as they get it during the forecast shift.

Our experience have shown that an archived case displayed in a simulator format engages the learners better than a traditional case study presentation. Working against the clock and producing real forecasts makes the exercise feel more authentic. We believe that even a limited amount of data presented in a simulator format is enough to create a real forecasting atmosphere.

The presentation will describe how the simulator exercises can be set up. Special emphasis will be put on demonstrating how the simulator has been used for the NOMEK course convection training.

Paper 200:

Automatic CB detection for METAR messages in Finnish Meteorological Institute

Harri Hohti, Finnish Meteorological Institute, Finland

Otto Hyvärinen, Finnish Meteorological Institute, Finland

Pirkko Pylkkö, Finnish Meteorological Institute, Finland


The information of the existence of Cumulonimbus clouds (Cb) is very important for aviation. The standard way of dissemination is through METAR messages. Traditionally these messages are constructed manually from human observations of the clouds in the vicinity of the aerodrome. However, because of the cost issues, automatization of this process is a growing trend.

The operational automatic detection of Cb for METAR in Finnish Meteorological Institute, based solely on radar data, is presented. Polar volume radar data of 500 m radial and one degree azimuthal resolution is used for the Cb detection. Doppler filtered data of each elevation is first postprocessed using dedicated anomaly
removal filter. Possible Cb range gates are recognized by using two distance dependent reflectivity thresholds. The amount and vertical consistency of passed range gates are then checked before acceptance as detected Cb. The process is repeated every five minutes, and detections in radar network are combined to create final cartesian Cb detection grid of 1 km resolution.

Radar data itself is, arguable, the most suitable data source for the Cb detection, but other data sources could be of complementary nature. We explore the possibility to substitute radar with lighting data and satellite products of EUMETSAT's NWCSAF program in cases of temporary breaks or other errors of data transmission in the radar network.

Paper 201:

Linear and non-linear methods to evaluate the usefulness of meteorological station data for thunderstorm nowcasting

Arturo Pucillo, OSMER - ARPA FVG, Italy

Agostino Manzato, OSMER - ARPA FVG, Italy


Thunderstorm nowcasting is very important in Po Valley (Italy) because storms can be particularly severe during summer season. The complex thunderstorm forecasting problem is based very much on the dynamical approach (LAM) even if the statistical approach (MOS, statistical adaptation, observation extrapolation, etc) is still used for nowcasting purposes (e.g. INCA software). In a previous work, the authors have explored two statistical techniques, based on multiregression and Linear Discriminant Analysis classification, applied to a dataset of meteorological surface observation used as predictors of 5-minute maximum radar VMI in the plain of the Friuli Venezia Giulia region (at the eastern boundary of the Po Valley, NE Italy). The results obtained indicated that the relevance of a linear model using observed data for nowcasting purposes is more promising in the classification approach; that suggests to try further steps in that direction. Moreover, also the multiregression approach can be improved if more refined techniques are implemented (e.g. exhaustive search for the best predictor subsets). So, in this development of the work, the authors test on one side the improvement of the dataset adding more candidate predictors (such as the pressure, the Sea Surface Temperature, etc.), restricting the analysis to a well defined period of the year according to the frequency of most severe storm occurrence in Friuli Venezia Giulia region, and by studying only the cases that present maximum VMI values above a significantly high threshold. On the other side, a non-linear statistical approach is tested in order to evaluate the improvement achieved by implementing a more complex model, taking in account the limits shown by linear approach. An objective verification is provided in order to compare the results of linear and non-linear approaches.

Paper 203:

A severe convective episode triggered by accumulated precipitation in the coast of Parana State, Brazil

Jefferson De Lima Picanço, Universidade Estadual de Campinas, UNICAMP, Brazil

Lucí Hidalgo Nunes, Universidade Estadual de Campinas, UNICAMP, Brazil


In March 2011 significant accumulated rain severely hit 2 coastal towns in Paraná State, Southern Brazil. The affected area comprises a flat narrow region in the western side of the elevations of Serra do Mar Escarpment. The catastrophe caused floods, flash floods and landslides and led 4 people to death besides 8 injured, left around 33,000 displaced and caused major damages, forcing the government to declare the state of calamity to facilitate the use of funds to aid affected families and expedite rehabilitation efforts.
One of the rain gauges located in the town of Morretes recorded 265mm of rainfall in 24 hours, which corresponds to the average annual precipitation (268mm) and 536 mm between the 10th and 12th March. Notwithstanding, the accumulated rainfall of the precedent days must be contributed to the severity of the event, as a sequence of rainy days cause soil saturation and thus, can trigger major disasters.
The largest impacts occurred on 11th March, connected to a strong convective activity, which is quite common during summer season in the area. An eastward instability was observed and remained in the coast, associated with a trough with high gradient and considerable instability at different levels which brought in turn strong winds. Transport of humidity from the ocean associated with the subtropical high circulation over the Atlantic also contributed for the heavy precipitation. Clouds were not too developed as temperatures at ground and values of CAPE were not very high, but temperatures were close to dew point due to the high instability and associated humidity.
Residents declared that it was the biggest disaster caused by precipitation ever and the registers showed that rainfall totals were quite high and concentrated in a relatively short period. However, the consequences also enhance the ever-growing vulnerability of population to weather episodes.

Paper 204:


Jose Luis Sanchez, GFA. IMA. University of Leon, Spain

Estibaliz Gascón, GFA. University of Leon, Spain

Sergio Fernandez, GFA. University of Leon, Spain

Laura Lopez, GFA. University of Leon, Spain

Pablo Melcón, GFA. University of Leon, Spain

Eduardo García-Ortega, GFA. University of Leon, Spain

Andres Merino, GFA. University of Leon, Spain

Jose Luis Marcos, GFA. University of Leon, Spain


Traditionally, the static stability has been obtained from indices such as KI, HI, TT, Showalter or CAPE, calculated twice a day from the vertical profiles obtained from radiosondes. There are a number of references that have established the likelihood of occurrence of a hailfall or thunderstorms from calibrated either a single index, or a combination thereof.

However, the analysis of the static stability returns to again be a topic open to scientifically consider take continuous measurements of the vertical profiles of temperature and humidity. It seems necessary to compare the results with those obtained when calculated from data from a sounding.

So, a multichannel microwave radiometer (MMWR) can be used to calculate, in real time and continuously, the temporal evolution of different stability indices. In our case a MP-3000A Microwave Profiler and database of severe weather situations was used to study the “traditional indices” and the “continuous stability indices”.

Finally, a model to assess the probability of occurrence of storms in the Middle Ebro Valley was built for making use of continuous indices calculated from the data of MP-3000A.

Acknowledgments: This study was supported by the Plan Nacional de I+D of Spain, through the grants CGL2010-15930, Micrometeo IPT-310000-2010-022 and the Junta de Castilla y León through the grant LE220A11-2.

Paper 205:

Environment of tornadogenesis in outer rainband of tropical cyclone

Masahiko Ebara, Weather Information & Communications Service LTD., Japan

Koji Sassa, Kochi University, Japan


Tropical cyclone is one of major sources of tornado outbreak in Japan.More than 70 percents of tornadoes occur in the outer rainband of tropical cyclone. However, the environmental parameters, e.g., SReH, EHI and so on, cannot represent that outer rainband is the region suitable for tornado outbreak. The present study aims to clarify what environmental factors of outer rainband are important to generate tornadoes and/or mesocyclones.

We simulated some cases of tornado outbreaks in outer rainband, using the Weather Research and Forecasting (WRF) model version 3.4.1. As their initial data, the NCEP reanalysis data and the GPV data obtained from the mesoscale non-hydro static model of Japan Meteorological Agency. We also used using NEAR GOOS Merged satellite and in situ data Global Daily Sea Surface Temperatures (MGDSST) as sea surface temperature data. The spatial resolutions of three domains were 9, 3 and 1 km, respectively.

Our numerical simulation will show the mechanism of generating supercell in outer rainband. We expect that such understanding will help to find the relevant environmental parameter to predict tornado outbreak in outer rainband.

Paper 206:

Short range forecasting and nowcasting of severe weather operational products developed in UHMI

Oleksii Kryvobok, Ukrainian Hydrometeorological Research Institute, Ukraine


Short range forecasting operational products based on NWP model are developed in the Ukrainian Hydrometeorological Research Institute. The outputs of regional WRF model were used for estimation of pre-convective conditions in summer based on KI, PW, LI and TT indices. The verification of these parameters has been done in the early morning against the occurrence of lightning later in the day. The special “look‐up” table where a value of the index corresponds to a percentage chance that lightning occurs was generated. Combined Instability Index has been identified as a combination of Total Totals, K Index, Lifted Index and Precipitable Water weighted with their ability to predict lightning with a lead time of at least 3 hours. A probability map of convection for Ukraine is generated operational.
Severe weather nowcasting operational products based on satellite data include:
- different MSG RGB images;
- combined enhanced MSG IR10.8 image and AMV;
- Nowcasting SAF Rapid Developing Thunderstorm product.
Using remote sensing data is crucial in the first 12 hours of a forecast. MSG data is very important because of a spectral variety of channels, which offers a generation of different kind of RGB images and provides much better meteorological interpretation of satellite images. Using satellite and model data, as well as hybrids of these, we have a very important information on convection and severe convection.

Paper 207:

The relationship between tropical storm and precipitation patterns of heavy rain in Kochi, Japan

Hitomi Makigusa, Kochi University, Japan

Koji Sassa, Kochi University, Japan


Kochi was known as the area where heavy rain frequently occurs in Japan. Tropical storms are major sources causing heavy rain in Kochi. The present study aims to categorize mesoscale convective systems yielding heavy rain and to investigate relationship between the locations of tropical storms and the categorized convective systems.
We checked radar echo patterns when any rain gauges at 29 Japan Meteorological Agency’s (JMA)’s observation points in Kochi prefecture recorded more than 50 mm/h and categorized them to five types; spiral rainband, fixed echo, streak on slope, quasi-linear rainband and multi-cell types. The spiral rainband type is the convection system of tropical storm itself. The fixed echo type is quasi-stationary convective system locked at certain area due to orographic effect. The streak on slope is also a kind of orographic rainfall. The other two types are major convective systems that appear with extra tropical storms and Meiyu front. We extracted 236 echo patterns for 25 years from 1986 to 2010. The tracks of tropical storm’s eye when the convective system appeared were also checked from the typhoon best track data collected by JMA.
The heavy rain was caused mainly by orographic systems.The sum of the fixed echo and the streak on slope types was more than 70 percents of total systems. This fact shows that heavy rain is mainly caused by the interaction between the terrain of Kochi and the wet swirling airflow of tropical storms. We found that the major convective system yielding heavy rain was different depending on the location of tropical storm. The probability of heavy rain was also found to depend on the location of tropical storm.

Paper 209:

Comprehensive comparison of tornado and earthquake statistics

Lisa Schielicke, FU Berlin, Germany

Peter Nevir, FU Berlin, Germany


Tornadoes and earthquakes are characterized by a high variability in their properties, e.g. in their geometrical extent, their magnitude and their temporal occurrence. Additionally, both are observed to cluster in space and time. Whereas earthquake statistics are known for their power-law behavior (e.g. Gutenberg-Richter law, Omori law), especially the temporal tornado behavior has not been studied in a comprehensive manner so far. In this work, tornado data was analyzed with the help of seismological methods and compared to earthquake data. In general, tornado statistics show power-law behavior partly coextensive with characteristic scales when the temporal resolution is high (10 to 60 minutes). These characteristic timescales match with the observed diurnal behavior of tornadoes and peak around multiples of one day. Additionally, we discuss a possible similar underlying structure of earthquakes and tornadoes composed of heterogeneous, coupled, interactive threshold oscillators that possibly explains the observed behavior.

Paper 210:

Quantitative expression of uncertainty in nowcasting heavy convective precipitation in Central Europe by extrapolation methods

Jan Mejsnar, Department of Meteorology and Environment Protection MFF UK, Czech Republic

Zbyněk Sokol, Institute of Atmospheric Physics AS, Czech Republic

Petr Pešice, Institute of Atmospheric Physics AS, Czech Republic


Methods extrapolating observed radar reflectivity along Lagrangian trajectories are frequently used for nowcasting of precipitation. The forecast errors of these methods have two basic components. The first one stems from inaccurate estimates of the motion fields used for calculating trajectories. The second one follows from the assumtion that is frequently applied and that consists in an assumption that observed radar reflectivity does not change in another way that by the advection. Some research of these aspects has already been done but it concentrated to another reagions than to Central Europe. Because scale characteristics of convective storms differ for various parts of the world we think that such study can bring original results.
The aim of this study is to quantify these two errors and then quantify predictability and uncertainty in forecasting convective precipitation. We will analyze heavy convective storms which occurred in the Czech Republic in 2009 and 2010 by means of several extrapolation methods. The main attention will be devoted to analyses of forecasted and observed accumulated precipitation, in contrast to usually analyzed forecasted reflectivity, over 60 and 30 minutes. We will concentrate on error analyses of precipitation forecasts for „typical“ Czech small catchments, the size of which is from 100 to 200 km2.

Paper 211:

Tornado and Severe Hail Environments of Turkey

Abdullah Kahraman, Turkish State Meteorological Service, Turkey


Proximity soundings derived from radiosonde data or global reanalyses are one of the main tools for investigating synoptic scale environments associated with severe convective storms and related hazardous phenomena. Results from distinct geographies suggest that although some of the variables representing particular ingredients of severe convection are accepted to be suitable for identifying a specific type of storm/event in most of the globe, it is well known that there are considerable regional differences in choice and threshold of parameters due to contrasting climatological and topographical characteristics.
This study aims understanding synoptic scale environments of severe convective storms in Turkey using ERA-Interim data from 1979 to 2012 with six-hour intervals. Tornado, waterspout and severe hail database recently built from various sources is used as severe weather observations. In addition to these categories, an “other thunderstorms” class is defined, of which data are obtained from Turkish State Meteorological Service’s climatological stations. Derived variables prior to tornado, waterspout and severe hail occurences are studied for determination of discriminatory parameters. Results are compared to similar studies performed for USA and European countries.

Paper 212:


Zarmandukht Petrosyan, Armstatehydromet, Armenia

Hamlet Melkomyan, Armstatehydromet, Armenia

Rita Abraamyan, Armstatehydromet, Armenia

Hrach Astsatryan, Institute for Informatics and Automation Problems, Armenia

Vladimir Sahakyan, Institute for Informatics and Automation Problems, Armenia

Yuri Shoukourian, Institute for Informatics and Automation Problems, Armenia

Levon Vardanyan, State Hydrometeorology and Monitoring Service of Armenia, Armenia

Anna Shahnazaryan, Institute for Informatics and Automation Problems, Armenia


The intensity and frequency of hazardous hydro-meteorological phenomena, especially strong winds and heavy rainfall, are expected to increase in Armenia further harming e.g. agriculture and infrastructure (Second Report, 2010). In the mean time the accuracy of severe weather prediction, e.g. thunderstorms, stormy winds and flash floods is difficult to predict for the territory of Armenia due to its complex geography and diverse terrain. At present Armenian State Hydrometeorological and Monitoring Service , responsible for the operational forecasting for the territory of Armenia, operates 48 meteorological stations, and only 3 meteorological station synoptic data are involved in global exchange through World Meteorological Organization Global Telecommunication System as well as historical data and monthly updates to the Global Climate Observing System Surface Network. The implementation of a numerical weather prediction regional model allows to resolve mesoscale phenomena with involving additional local specific data. This is crucial for Armenia, because the number of stations involved in the global data exchange is limited. The aim of this article is to present the results of the data assimilation impact on heavy rainfall within the Weather Research and Forecasting (WRF) model (Michalakes, 1998) based on the heavy rainfall event for the period 15-18 April 2013. The will be important step towards to the operational usage of the model, which will allow to investigate the potential impact of station`s data on short-range forecasts.

Paper 213:

A significant late-winter severe weather outbreak in subtropical South America in the synoptic and climatological contexts.

Ernani L. Nascimento, Universidade Federal de Santa Maria, Brazil


On 7 September 2009 a significant severe weather outbreak was observed over a large area encompassing southern Brazil and extreme northeastern Argentina. Different convective modes, including a bow echo and supercells, were observed at distinct stages of the outbreak that initiated in the morning and lasted until the late night hours. The outbreak occurred under strong synoptic forcing within the warm sector of a developing extratropical cyclone, with concurrent strong conditional instability, vertical wind shear and a low-level jet. The role played by the Northern Argentinean Low (inverted trough; NAL) and its interaction with the developing extratropical cyclone in setting up favorable conditions for severe convective storms is assessed. A qualitative comparison also is performed between this South American synoptic pattern and the one typically observed over the Central Plains of North America on situations of severe weather with strong synoptic-forcing. It is highlighted that the presence of the NAL represents an important distinction between the surface synoptic patterns observed on the two continents. Finally, as an effort to quantitavely assess how the 7 September 2009 episode ranked in a climatological context, the magnitude of the severe weather parameters observed during this event is compared to a 10-year sounding-based climatology of some traditional convective indices over subtropical South America. It is found that the magnitude of the weather parameters during this event outstands in the climatological context.

Paper 214:

Probabilistic rainfall warning system with an interactive user interface

Jarmo Koistinen, Finnish Meteorological Institute, Finland

Harri Hohti, Finnish Meteorological Institute, Finland

Janne Kauhanen, Finnish Meteorological Institute, Finland

Juha Kilpinen, Finnish Meteorological Institute, Finland

Vesa Kurki, Finnish Meteorological Institute, Finland

Tuomo Lauri, Finnish Meteorological Institute, Finland

Antti Mäkelä, Finnish Meteorological Institute, Finland

Pertti Nurmi, Finnish Meteorological Institute, Finland

Pekka Rossi, Finnish Meteorological Institute, Finland

Dmitri Moisseev, University of Helsinki, Finland


A real time 24/7 automatic alert system is in operational use at the Finnish Meteorological Institute (FMI). It consists of gridded forecasts of the exceedance probabilities of rainfall class thresholds in the continuous lead time range of 1 hour to 5 days. Nowcasting up to six hours applies ensemble member extrapolations of weather radar measurements. Nowcasting exploits also lightning density and satellite based pseudo rainfall estimates. The extrapolation technique applies atmospheric motion vectors (AMV) originally developed for upper wind estimation with satellite images. Exceedance probabilities of four rainfall accumulation categories are computed for the future 1 h and 6 h periods and they are updated every 15 minutes. For longer forecasts exceedance probabilities are calculated for future 6 and 24 h periods during the next 4 days. From approximately 1 hour to 2 days Poor man's Ensemble Prediction System (PEPS) is used applying e.g. the high resolution short range Numerical Weather Prediction models HIRLAM and AROME. The longest forecasts apply EPS data from the European Centre for Medium Range Weather Forecasts (ECMWF). The blending of the ensemble sets from the various forecast sources is performed applying mixing of accumulations with equal exceedance probabilities. Ensemble data from other sources (satellite, lightning, NWP) are converted to the same geometry as the radar data and blended as was explained above. A verification system utilizing telemetering rain gauges has been established. Alert dissemination e.g. for citizens and professional end users applies SMS messages and, in near future, smartphone maps. The interactive user interface facilitates free selection of alert sites and two warning thresholds (any rain, heavy rain) at any location in Finland. The pilot service was tested by 1000-3000 users during summers 2010 and 2012.

Paper 215:

Variability of strong convective storms in Ukraine

Vira Balabukh, Uhmi, Ukraine

Stepan Yagodinets, Uhmi, Ukraine

Ludmila Malytska, Uhmi, Ukraine


The paper considers the conditions of formation of hazardous convective weather phenomena (strong showers, hail, squalls, tornadoes) in Ukraine and their spatial and temporal variability during 1981 – 2010.
Research of convection processes was based on daily radiosonde data for the warm season (May-September 1981 - 2010s), reanalysis ERA-Interim ECMWF data for 1989 - 2010 years , daily observations at 187 meteorological stations in Ukraine, as well as observations of the natural phenomena in other regions (different from the meteorological stations). Indices of atmospheric instability, the magnitude of the Convective Available Potential Energy (CAPE), the moisture, the height of the condensation and equilibrium level was used to quantify the intensity of convection.
The criteria for the intensity of convection for Ukrainian territory were refined on the basis of these data. Features of the development of convection for various hazardous convective weather events were investigated and identified the necessary conditions for the occurrence of showers, hail, tornadoes and squall in Ukraine.
Spatio-temporal variability of convection intensity in Ukraine, its regional characteristics and dynamics for the past 30 year was analyzed. Significant tendency to an increase the average temperature and moisture of the troposphere is observed during 90s of the twentieth century in Ukraine in the warm season that led to the growth of CAPE of the atmosphere, the speed of updrafts, raising the level of condensation and convection, and have increased the instability of the atmosphere. The number and intensity of strong showers, hail, squalls, tornadoes and the number of days with thunderstorms have increased due to such changes in Ukraine.

Paper 216:

Cold rings and cold U/V shapes as seen atop of deep convective clouds in infrared satellite imagery in years 2006-2012 over Slovenia

Mateja Irsic Zibert, Slovenian Environment Agency, Slovenia


This paper presents an analysis of cold rings and cold U/V shapes (Setvak et al., 2012) as observed in enhanced infrared (IR) window satellite imagery atop of deep convective storms. The data from SEVIRI/METEOSAT measurements are used. The cold rings and cold U/V shapes are automatically detected during daytime and night-time using SATSEVERE algorithm (Iršič Žibert et al., 2012). The distribution and the typical characteristics of cold rings and cold U/V shapes in the period 2006-2012 over the southern part of Central Europe are presented and discussed.
Additionally, a recently established concisely verified database of severe weather events (strong wind, strong precipitation and/or large hail) based on multiple sources of ground reports (NMS ground network, ESWD database, other national reports, …) for the period 2006-2012 is presented. Such database is indispensable when studying the aforementioned features of satellite imagery and their connection to severe weather occurrence.

Iršič Žibert, M., J. Žibert. Monitoring and automatic detection of the cold-ring patterns atop deep convective clouds using Meteosat data, Atmos. Research, 2012.
Martin Setvák, Kristopher Bedka, Daniel T. Lindsey, Alois Sokol, Zdeněk Charvát, Jindřich Šťástka, Pao K. Wang . A-Train observations of deep convective storm tops, Atmos. Research, 2012. http://dx.doi.org/10.1016/j.atmosres.2012.06.020

Paper 217:

Information system „The Dangerous Meteorological Phenomena» in Ukraine

Vira Balabukh, UHMI, Ukraine

Stepan Yagodinets, UHMI, Ukraine

Elena Lavrinenko, UHMI, Ukraine

Tamara Sotnik, UHMI, Ukraine

Nataliya Talerko, UHMI, Ukraine


Information system „The Dangerous Meteorological Phenomena» is intended for the storage and processing of data for Dangerous weather phenomena in Ukraine and about atmospheric processes that cause them. System consists from information and handbooks parts, specialized software, database query and the user interface.
The database archival and operational information on dangerous phenomena weather and condition of atmosphere are components of the information system. They include information about the intensity, duration, location and the formations conditions of meteorological phenomenas: strong winds, squalls, tornadoes, intense showers, hail, very strong and prolonged rains, snowfalls, severe snowstorms and dust storms, sleet, complex deposition, cases of extreme heat and frost.
Databases systems include more than 120,000 cases of dangerous weather phenomena (1968- 2012s), daily observation data network aerological sounding Ukraine (more than 160 000 radiosonde from 1973 to 2012s) and data on the state of the atmosphere obtained based reanaliz Era-Interim ECMWF (1981 - 2012s).
The handbooks part of the system include: directories "Dangerous phenomenon in Ukraine" and "Weather and Health"; monthly and annual weather reviews and atmospheric circulation in Ukraine, Europe and Northern Hemisphere, cloud atlas, glossary of meteorological terms.
Information system "The Dangerous Meteorological Phenomena" enables to do integrated research spatial and temporal distribution of dangerous weather phenomena and conditions of their formation, study atmospheric processes and their regional features, create a problem-oriented databases.
The technologies and databases of information system allowed the first in Ukraine to establish a system of monitoring changes conditions of formation, frequency, intensity and location of hazardous weather phenomena. They make it possible also to identify the most dangerous and favorable for living and investment regions of Ukraine.

Paper 218:

Estimating road accident responsiveness to bad weather and weather warnings

Adriaan Perrels, Finnish Meteorological Institute FMI, Finland

Väinö Nurmi, Finnish Meteorological Institute FMI, Finland


Even though other factors, such as alcohol use, cover a larger share of the causes of road traffic accidents in many countries, adverse weather conditions definitely contribute to accident rates as well. The paper discusses the estimation of an accident occurrence function for road traffic accidents in Finland during the winter months (roughly: mid October - mid April). The data set comprises of the reported number of accidents per day by province in combination with daily weather observation and warning data per region as well as road infrastructure quality information. The total number of observations is about 80000. The model is estimated for regional road accident incidence in relation to weather conditions, while correcting for other influences, such as type of day. The output of the model indicates the expected rise in the number of accidents in a region as compared to a normal day. The model accounts for the effect that traffic flows diminish in response to bad weather.
As the Finnish road transport system is well adapted to winter weather weather conditions have to be really quite bad (winter storms; blizzards) before accident incidence rates start to rise clearly. The paper also shows the difference between well pre-warned and less well pre-warned adverse weather conditions. Last but not least it provides some cost estimates of changes in accident rates.

Paper 219:

Remote sensing analysis of a severe convective storm in Basque Country.

Santiago Gaztelumendi, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Joseba Egaña, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Javier Lopez, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Roberto Hernandez, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Mercedes Maruri, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

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


In this paper an analysis of a severe convection event in Basque Country is made focusing on operational remote sensing information, including Radar, Meteosat and lighting detection system, available for the study area in Basque Meteorology Agency. We also present different aspects of the environment characterization affecting the storm evolution, at synoptic, mesoscale and local scale, including numerical modeling and registered data from Automatic Weather Stations (AWS) network.

This heavy hailstorm episode, affects the central part of Basque Country on 30th May 2011. During this event convective cells with large vertical development reaching or exceeding 10 km are present, as a consequence radar reflectivity values observed are exceptional, mainly due to rain rate and droplet size rather than to hail size.

Some heavy (>15mm/hour) and very heavy (>30mm/hour) rain episodes, accompanied with hail, were observed all over the region. Showers are especially intense in Zuia area, where some AWS register more than 30 mm in one hour with ten-minutes rain-rates over 15 mm. In Sarria AWS this hailstorm left 115 mm precipitation in one hour, a hourly accumulation rainfall data never registered before for this location. In the surroundings of Gorbea Mountain, damages due to hailstone and water accumulation where produced in some places including Murgia Village.

Paper 220:

Analysis of simulated precipitation in a 2D non-hydrostatic model

Maurizio Fantini, ISAC-CNR, Italy


The importance of low-level shear in the development of organized convective structures is well established in observations and numerical modeling. More generally, the formulation of quantitative relationships between the multiplicity of environmental parameters and measures of convective activity, by which to validate conceptual models, is a complex field of research.

With the intent to establish a baseline to interpret the variability of organized convective storms in response to changes in environmental conditions, we perform numerical experiments with a simplified non-hydrostatic model, with reference wind and thermodynamic profiles.

For these idealized simulations, the non-hydrostatic model MOLOCH is reduced to a two-dimensional domain (x-z) with flat bottom and no soil physics, while diffusion and microphysics are the same of the operational model, and a diurnal cycle is imposed, uniform over the domain.

Preliminary results are given in terms of number and intensity of precipitation episodes for a range of numerical and physical parameters.

Paper 221:

A study of a flood episode in Basque Country

Joseba Egaña, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Santiago Gaztelumendi, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Miriam Ruiz, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Roberto Hernandez, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

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

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


In this study we present an analysis of a flood episode affecting San Sebastian area on 6 November 2011. A heavy and persistent rainfall episode affects Basque Country along 4-7 November 2011, as a consequence some floods are produced mainly in the east part during 6 November.
During this episode a Cut-Off Low moves from south of Pyrenees towards Mediterranean Sea over Balearic Islands area. In the surface, a low is generated over the northeast of Iberian Peninsula. As this low gets stronger, northerly flow intensifies over the Bay of Biscay and different fronts fully affect the Basque Country during several days. The characteristics of the Mediterranean air mass, its posterior passage over Cantabrian Sea, and the front persistence over the Basque Country, cause high efficiency in precipitation leaving extraordinary amounts of precipitation in some parts of northeastern Basque Country, during day 5 and 6.
During the event a mix of stratiform and low vertical developed convective cells are observed. Although heavy showers are produced, most serious damages are originated by rainfall persistence.
In some areas in the northeast of the Basque Country, accumulated precipitation during this episode represents over twice the November average rainfall. The worst affected area was the northeast of the Basque Country. Near San Sebastian total amount of precipitation is over 300 mm.
During the event, a person died and a thousand people have been affected by floods. On November 6, the east of the Basque Country was seriously affected: a hundred people were evacuated, a power cut affected 5400 people, a lot of properties were affected with significant damages and several roads were closed due to flooding and landslides.

Paper 222:

Four Year Lightning Climatology in the Basque Country

Javier Lopez, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Santiago Gaztelumendi, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Joseba Egaña, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Mercedes Maruri, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Roberto Hernandez, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain

Ivan R Gelpi, Tecnalia - Meteo Area / Basque Meteorology Agency, Spain


Since the end of 2008, the Basque Country Lightning Detection Network has been fully operative. This paper presents a four year climatology of the cloud-to-ground lightning strokes detected by the network in the region of the Basque Country (NE of Spain) since 2009.
In these four years of operation, nearly 31000 strokes have been detected and located in the Basque Country. 2009 and 2011 were very active years, with more that ten thousand detections each. On the contrary, 2010 and 2012 were less active years, counting a number of detections not higher than 5350 each year.
According to the latitudes in which the Basque Country is located, the peak of the lightning activity is associated to the warm months, from May to September. The moisture accumulated during the cold season and the beginning of warmer temperatures make May and June typical months for high lightning activity that usually concentrates on the south of the region. From July on, the warmth accumulated by the sea contributes to more seaside thunderstorms and a higher electrical activity close to the coast.
Regarding the polarity of the strokes, the percent of positive strokes in each season ranges from 5% (2011) to a maximum of 10.7% (2009). In these four years, between 50% to 60% of the positive strokes detected in the Basque Country presented a peak current less that or equal to 20 kA, whereas nearly 80% of the negative strokes in all the four cases studied presented a peak current below this maximum.
According to the data retrieved by the network, the most intense positive CG stroke registered in this period had a peak current of 243 kA, and it occurred in the month of December of 2010. The most intense negative record corresponds to a stroke of -239 kA in November 2011.

Paper 223:

Recent observations of meteotsunamis on the Finnish coast

Hilkka Pellikka, Finnish Meteorological Institute, Finland

Hilppa Gregow, Finnish Meteorological Institute, Finland

Jenni Rauhala, Finnish Meteorological Institute, Finland

Juha Aalto, Finnish Meteorological Institute, Finland

Pauli Jokinen, Finnish Meteorological Institute, Finland

Kimmo Kahma, Finnish Meteorological Institute, Finland

Pentti Pirinen, Finnish Meteorological Institute, Finland


Several occasions of exceptionally strong and rapid sea level variations were observed on the Finnish coast in the summer months of 2010 and 2011. We interpret these variations as small meteotsunamis: waves having tsunami-like properties, but caused by meteorological processes, namely small-scale air pressure disturbances moving over the sea at a resonant speed. According to the eyewitness reports, the height of the sea level variation exceeded 1 m in some locations and the oscillations induced strong currents in shallow straits. Smaller but clear fluctuations were observed at the Finnish tide gauges. At least one of the events (8 August 2010) was related to a mesoscale convective system, which caused strong wind gusts up to 32 ms-1.

We try to explore whether the onset and development of these sea level oscillations are traceable from the radar imagery, or from shoreline observations of temperature, wind speed, humidity and surface pressure. Additionally, we will investigate how a gridded data set with high resolution over land can help in following the propagation of the meteorological phenomenon that caused the meteotsunamis. Our hypothesis is that by combining tide gauge data to meteorological data it is possible to detect historical meteotsunami events. To what degree we can be certain without radar data and, on the other hand, how much more we can learn about meteotsunamis by using the radar data is also the scope of our research. The findings will be of great importance for the safety of coastal citizens, as well as the safety and planning of the nuclear power plants along the Finnish coast.

Paper 224:

Analysing spatial distribution of cloud-to-ground lightning in Portugal: patterns identification and relationship with geographical factors.

Eusébio Reis, Institute of Geography and Spatial Planning (University of Lisbon), Portugal

António Lopes, Institute of Geography and Spatial Planning (University of Lisbon), Portugal

Marcelo Fragoso, Institute of Geography and Spatial Planning (University of Lisbon), Portugal


This work aims to characterize the geographic incidence of cloud-to-ground lightning (CGL) activity in mainland Portugal. For this purpose, a seven-year period, from 2003 to 2009, was considered. The annual and monthly spatial distribution of cloud-to-ground (CG) flashes were analysed and compared with several geographical factors: altitude, aspect, geology, soil types, land occupation and distance to ocean. Regardless of the strong interannual variability among this period, some annual spatial patterns can be detected. Basically, there is not a general annual or monthly pattern for all the territory: a clear opposition between North and South, as well as between interior (more continental, with stronger lightning activity) and littoral (exposed to strong maritime influence, with lower lightning activity) lands was found, that reflects a combined influence of altitude and distance to ocean. In the northern sector there is a clear evidence of the influence of altitude, since there is a correspondence between the mountains and a higher flash density; in the South, where low altitudes prevail, other factors as distance to ocean are more important.
On the other hand, these factors act differently when distinct periods of the year are considered. It is clearly observed the variation of lightning density along the year, with higher densities during the warmer period, specially May, June and September, and very low densities in December and January; additionally, and not less relevant, a changing spatial patterns along these months can be detected: apart absolute density differences, from October on the concentration of CG flashes becomes lower in the interior of Portugal, especially in the northeast sector, characteristics which extent South along the Winter months; in this period, the CGL activity is very scarce or absent in the North Interior of the country, and there is not a relevant difference associated to distance to ocean in the South; from March on, this pattern changes in to another direction, since the concentration of CG flashes becomes higher in the interior of Portugal, starting from South, and the presence of these occurrences near the ocean are almost irrelevant; at this point, from June to August the presence of CGL activity in the North Interior, especially in northeast sector, becomes relatively dominant.

Paper 225:

Analysing spatial distribution of damaging floods and mass movements in Portugal from 1865 to 2010 (DISASTER database): geographical factors, weather types and human impacts.

Eusébio Reis, Institute of Geography and Spatial Planning (University of Lisbon), Portugal

José Luís Zêzere, Institute of Geography and Spatial Planning (University of Lisbon), Portugal

Marcelo Fragoso, Institute of Geography and Spatial Planning (University of Lisbon), Portugal


The GIS DISASTER Portuguese database includes all the damaging floods and mass movements events detected in mainland Portugal since 1865. This database results from three years of investigation; during this period, 16 newspapers were inquired and 145344 copies of newspapers analyzed. In these 146 years, 1903 occurrences were detected (85% are floods), which caused 1310 human deaths, 14191 evacuated and 41844 displaced, mostly as a consequence of flash floods generated from extreme rainfall events.
Although it is not possible detect any clear trend over time, the spatial distribution of both phenomena reflects the influence of natural factors and the very irregular geographical distribution of the urban settlements and human activities. These human variables helps to explain why such a different phenomena, that are conditioning for different factors, have a relatively similar distribution (r2=0.73, for relation between density of floods and mass movements by municipality). The analysis of the floods shows five main areas of high density occurrences: Lisbon metropolitan area; Oporto and surrounding municipalities; Douro valley; Tagus valley; area of Coimbra city (centre of Portugal). In the remainder territory the damaging floods are also present, but their occurrence are much more scattered.
On the other hand, the damaging mass movements, much less frequent, have a more concentrated distribution, that reflects the important influence of local factors, in special the slope and lithology: Lisbon city and very near surrounding areas; Oporto city; slopes along Portuguese middle sector of Douro valley; coastal sectors, mainly in cliffs of Algarve.
With respect to floods, some spatial patterns can be defined, related with different weather types; three of them can be highlighted: very concentrated events, along small areas or linear trajectories, related with very convective depressions, that generate flash floods; a regional pattern distribution, normally separated in several sub-regions, whose influence can be detected only in a part of the territory (north, centre, northwest, etc.); a linear distribution along the main rivers, related with very long rainfall periods caused by successive surface frontal zones, that generate progressive floods.

Paper 226:

The Effects of Extreme Weather on the European Transport System: an Analysis Based on Media Reports

Riitta Molarius, C, Finland

Pekka Leviäkangas, Technical University of Oulu, Finland

Jaana Keränen, VTT Technical Research Centre of Finland, Tampere, Finland

Ilkka Juga, Finnish Meteorological Institute, Finland

Andrea Vajda, Finnish Meteorological Institute, Finland


Europe-wide storms cause serious and long lasting damage to society and especially to transportation networks, in regard to the rebuilding costs of roads, railways, bridges and tunnels, lost income in aviation, and delays in logistics systems of European industry. In order to determine which extreme weather events have had the most significant impacts on different transportation systems in the various parts of Europe, a media analysis approach was performed. The focus was on determining what kind of weather is regarded as extreme or adverse in different countries. As the databanks from Meteorological Offices typically don’t include any information about the consequences of the events, it was decided to generate a database of media news. The use of media reports allowed a different insight into the disaster and its consequences from the injured society point of view, which was not possible by using only meteorological statistics.
The results of the study indicate that the effects, conditions and consequences of extreme weather events differ according to not only location, geography, population distribution, and infrastructure, but also with regards to the preparedness of the society for different extreme weather conditions. There are clear differences between the northern and southern parts of Europe, but there are also significant variations between eastern and western areas. The study was done in European 7th Framework Programme project EWENT (Extreme weather impacts on European networks of transport).

Paper 227:

Photogrammetric analysis of Tsukuba tornado

Koji Sassa, Kochi University, Japan

Hiromori Miyagi, Miyazaki University, Japan


The Tsukuba tornado occurred about noon at 6 May 2012. It was one of the most violent tornadoes in Japan and its Fujita scale was F3. It destroyed 234 houses and resulted in one fatality and 37 injured. Many observers filmed the Tsukuba tornado with high-resolution cameras and smartphones from various locations.
In this study, we made photogrammetric analysis of the Tsukuba tornado in order to clarify its intensity and dimensions.The tangential velocity of the Tsukuba tornado was measured by using a particle image velocimetry method. Three movies filmed from different locations were synchronized through the image of spark from electric cables cut by the tornado.
The tornado track with accurate time of passage was accurately determined from these films. The tornado showed multiple-vortex structure with two suction vortices in the middle of the track. The diameter of the tornado near ground changed from 60 to 35 m. The tangential velocity of debris near ground was found to be 56 m/sec when the tornado attacked urban area and the moving velocity of the tornado was 16 m/s. Then, the maximum gust velocity was 72 m/sec. Such evaluation of the tornado velocity is as accurate as the measurement by a Doppler radar.

Paper 228:

Analysis of spatial and temporal variations of cloud-to-ground lightning in Romania

Daniel Carbunaru, National Meteorological Administration, Romania

Sorin Burcea, National Meteorological Administration, Romania


Severe weather events generating a significantly large number of cloud-to-ground lightning (CG) flashes have been analyzed. The CG flashes were recorded by the SAFIR type Romanian Lightning Detection Network, during 2003–2008. The research was focused on the overall dynamics of the CG associated with different severe convective storms which have developed in the same time, rather than focusing on the individual study of CG flashes associated with a given convective cell. For each severe event, the total number of CG flashes was grouped into a primary series which, afterwards, was used to analyze the spatial and temporal evolutions of CG flashes by building subsequent spatial (the distance between two consecutive CG flashes from primary series) and temporal (time interval between two consecutive CG flashes from primary series) series. Analysis revealed two classes of CG flashes: the ones that occurred in the same time and the ones occurring at different time moments. Time intervals between two CG flashes occurring at two time points are grouped around an average value. For each of the two classes, spatial intervals have been analyzed, the main resulted feature being that the majority of CG flashes occurring at the same moment in time are located at less than 10 km one from another. Frequency of detection associated to spatial intervals with various lengths, for every 1000 CG flashes from the primary series, have been analyzed. Results have shown that the majority of CG activity at a given time is grouped around a convective system, after which it is “transferred” to another convective system. For each severe event, distributions of the maximum currents at each 1000 CG flashes from the primary series were also obtained. The resulted distribution is invariant.

Paper 229:

New radar forward operator and first experiences of LETKF radar volume data assimilation within the convection-allowing NWP model COSMO-DE

Ulrich Blahak, German Weather Service (DWD), Germany

Yuefei Zeng, Karlsruhe Institute of Technology (KIT), Germany

Dorit Epperlein, Karlsruhe Institute of Technology (KIT), Germany

Hendrik Reich, German Weather Service (DWD), Germany

Andreas Rhodin, German Weather Service (DWD), Germany

Klaus Stephan, German Weather Service (DWD), Germany

Christoph Schraff, German Weather Service (DWD), Germany


The new weather radar network of the German Weather Service (DWD) will, after its complete update in 2014,
comprise 17 C-Band dual polarisation
Doppler radar systems evenly distributed throughout Germany for complete coverage.
They provide unique information about cloud structure and precipitation in
three dimensions and high resolution.
Up to now these data are not used in the operational COSMO-model of DWD (a non-hydrostatic limited
area numerical weather prediction model), except within the
framework of the latent heat nudging and for a simple nudging method of the radial wind.
Future applications are however planned to make better use of radar data within an
upcoming new LETKF data
assimilation system (Localized Ensemble Transform Kalman Filter),
which will be based on the operational convection-allowing
high resoluton ensemble forecasting system COSMO-DE-EPS (grid spacing
of 2.8 km, rapid update cycle, Central Europe domain).
Here, the use of weather radar data, which provide measurements of dynamical and
microphysical characteristics of precipitating clouds at high temporal and
spatial resolutions, is a promising means for improvements of
short-term precipitation forecasts, especially in convective situations.

However, the observations (reflectivity, radial velocity,
polarisation parameters) are not directly comparable to the prognostic variables of the model.
In order to, on one hand, enable radar data
assimilation in the framework of the above-mentioned LETKF-assimilation system
and, on the other hand, to
facilitate comparisons of numerical simulations with radar observations in
the context of cloud microphysics verification, a
comprehensive modular radar forward operator has been development.
This operator calculates the radar
observables reflectivity, radial wind and (currently under development) polarisation parameters from the prognostic
model output, which are then used in the LETKF system instead of, e.g., radar-derived hydrometeor contents.
To meet operational demands, this operator is
applicable on supercomputer architectures, and efficiency is a major design
criterion, which requires for good parallelization and vectorization
properties of the code.

The operator consists of several modules, each of which handles a
special physical process (scattering, extinction, microwave propagation,
etc.). Each of these modules offers different formulations to choose from,
which enables variable possibilities to configure the forward operator with
respect to the often conflicting constraints of physical accuracy and
computational costs. Here, a main goal is to find in some sense “optimal”
configurations for both applications, data assimilation and model
verification. For example:
1) the radar beam can be considered to propagate as a
simple ray or with the actual volume averaging
characteristics (beam function).
2) The beam bending can be either derived from a 4/3
earth radius concept or from the actual simulated vertical gradient of the refractive
index of air.
3) Radar reflectivity may be calculated from the full Mie-theory or
from various (more efficient) approximations, and attenuation effects may be
taken into account or not.

After having developed the operator and integrated the processing of radar data into the DWD LETKF software package, we are in the process
of conducting the first data assimilation experiments. For this, we run
a COSMO-DE ensemble for June 2011 (mainly convective precipitation), driven by a test ensemble of our global model GME.
The poster will summarize the method and our first experiences.

Paper 230:


Jorge Santos, Instituto de Ciencias Básicas (ICB), Universidad Nacional de Cuyo, Mendoza, Argentina, Argentina

Federico Norte, Programa Regional de Meteorología IANIGLA-CCT-Mendoza, Argentina, Argentina

Badrinath Nagarajan, National Center for Atmospheric Research (NCAR), Boulder, Colorado, USA, United States

Diego Araneo, Programa Regional de Meteorología IANIGLA-CCT-Mendoza, Argentina, Argentina

Silvia Simonelli, Programa Regional de Meteorología IANIGLA-CCT-Mendoza, Argentina, Argentina

Carlos Bustos, Dirección de Agricultura y Contingencias Climáticas, Gobierno de Mendoza, Mendoza, Argentina, Argentina


Mendoza in Argentina, experiences severe weather events such as, Zonda winds (Chinook-type winds) and hailstorms. The Zonda and hail storms produce damaging winds and hail leading to significant economical loss in the form of crop and property damage, besides loss of human lives. A satisfactory forecast of these events will not only mitigate damage but also help optimize the efficient allocation of resources of the emergency response agencies. As a first step, we successfully deployed in real time, a global state-of-the-art numerical weather prediction model, the Global Environmental Multiscale Model (GEM). The 48-h GEM forecasts has successfully forecasted several Zonda episodes during 2011 and 2012 with wind speeds, temperatures and dew points that show good agreement with the observed. As a second step, Mendoza hailstorms are forecast by the Advanced Research Weather and Research Forecasting (WRF) model. Verification of the hailstorm forecasts against radar observations indicate good (poor) skill of the WRF model in predicting the timing (location) of deep convection. These preliminary results will be further investigated and a series of sensitivity experiments will carried out by testing different parameterization schemes (convection and planetary boundary layer)and vertical and horizontal computational grid-spacing.

Paper 231:

Explicit simulations of mixed-phase electrified clouds of the HyMeX-SOP1 experiment with the mesoscale model MesoNH

Jean-Pierre Pinty, Laboratoire d'Aerologie, France

Michel Chong, Laboratoire d'Aerologie, France

Eric Defer, LERMA-Observatoire de Paris, France

Christelle Barthe, Laboratoire de l'Atmosphère et des Cyclones, France

Evelyne Richard, Laboratoire d'Aerologie, France

Paul Krehbiel, New Mexico Tech, United States

William Rison, New Mexico Tech, United States

Ronald Thomas, New Mexico Tech, United States


We present recent numerical results obtained with the cloud electrical module CELLS inserted in the cloud-resolving mesoscale model MesoNH (Barthe et al., 2012; Pinty et al., 2012). The case studies are taken from the first Special Observing Period of HyMeX (Hydrology Mediterranean eXperiment). The simulations are carried out at 1 km resolution and cover a large domain of several hundred km. The model is using Arome analyses of Meteo-France. The novelty of the study is brought by the comparison of the simulated flash characteristics to lightning network observations. In particular data from a LMA (Lightning Mapping Arrray) are used together with operational Lightning Detection Networks (EUCLID and others) to validate the outputs of MesoNH.

Paper 232:

Retrospective analysis of synoptic favourable conditions for deep convective events in the Mediterranean Sea

Massimiliano Pasqui, CNR-IBIMET, Italy

Samantha Melani, CNR-IBIMET, LaMMA, Italy

Francesco Pasi, CNR-IBIMET, LaMMA, Italy

Bernardo Gozzini, CNR-IBIMET, LaMMA, Italy

Marco Gaetani, JRC, CNR-IBIMET, Italy

Edmondo Di Giuseppe, CRA-CMA, Italy

Vincenzo Levizzani, CNR-ISAC, Italy


Deep convective systems generating over the Mediterranean Sea represent a substantial hazard for coastal communities and structures.
The dynamic behaviour of intense convective systems is the effect of a strong interaction between large scale atmospheric features, small scale dynamics, and energy fluxes. The span and duration of these precipitating systems are thus the result of such interaction, which, in turn, is strictly linked with their impact on coastal areas.
In the present study selected atmospheric circulation features at the synoptic scale were related to a four-year (2007-2010) convective events database over the Mediterranean basin. System detection was performed using geostationary satellite data according to Melani et al. (2012), while their “severity” characteristics were assigned by means of the European Severe Weather Database (ESWD) reports.
An objective daily circulation classification, developed in the framework of the COST-733 Action, is used to identify prominent synoptic conditions and atmospheric regimes during and before the convective initiation. Circulations are derived from a spatial variability principal component analysis of geopotential height data at 700hPa over the Mediterranean basin. The data used in this respect come from the Modern Era Retrospective-analysis for Research and Applications (MERRA) dataset over the time period 1979 – 2010.
Such a daily synoptic characterization is firstly used to find some recurrence characteristics and intra-annual behaviours of the detected events as well as to analyse Mediterranean large scale atmospheric variability (regimes and features) and its major forcing mechanisms over the entire MERRA period.
The analysis shows several positive significant trends in occurrence frequencies of two most prominent circulation classes characterized by pure cyclonic circulation and high zonal index in late summer and fall season. This is also confirmed in recent years along with an increasing role of the longitudinal sea surface temperature gradient of the Mediterranean basin.

Paper 233:

Flash floods evolution in Catalonia: from precipitation to societal aspects

Mª Carmen Llasat, Dept. Astronomy and Meteorology, Faculty of Physics, University of Barcelona, Spain

Raül Marcos, Dept. Astronomy and Meteorology, Faculty of Physics, University of Barcelona, Spain

Montse Llasat-Botija, Dept. Astronomy and Meteorology, Faculty of Physics, University of Barcelona, Spain

Joan Gilabert, Dept. Astronomy and Meteorology, Faculty of Physics, University of Barcelona, Spain

Marco Turco, Euro-Mediterranean Centre for Climate Change (CMCC), Italy

Pere Quintana-Seguí, Observatori de l'Ebre, URL-CSIC, Spain


The analysis and characterization of the flash flood events occurring in Catalonia (NE of Spain) for the period 1982-2007 has showed an increase in summer and early autumn, that probably has been mainly due to inter-annual and intra-annual changes in population density (Llasat et al, 2010). In fact, a recent work that analyses trends of the ETCCDI (Expert Team on Climate Change Detection and Indices) precipitation indices in this region from 1951 to 2003 has showed that are no general trends at a regional scale for those indices related with high precipitations (Turco and Llasat, 2011) neither for maximum daily precipitation. Another experiment dealing with the analysis of the evolution of rainfall rate for the last 15 years has been conducted. Its results have showed that there is a significative but slightly positive trend in the greatest part of the region with the exception of the North-Eastern part of Catalonia, where the yearly precipitation is also diminishing, mainly as a consequence of the decrease of precipitation in summer.

These results points towards the need of an in-depth analysis of the evolution of flash floods and precipitation in Catalonia. With this aim the period 1981-2010 has been selected. The study has been focused in two regions, one characterized by an increase of floods in spite of the decrease of precipitation (the Empordà region) and the other one characterized by the maximum frequency of floods (the Maresme region). Furthermore a comparative with a major period starting in 1900 has been done. 213 flood events have been identified for the period of 30 years (365 starting in 1900) 77 of them were ordinary, 114 extraordinary and 22 of them catastrophic, being the autumn the season with the maxima value. Almost 20% of the events in this period (1981-2010) caused 110 casualties. Factors like the evolution of population density, changes in uses of soil, and other socio-economical aspects have been considered. Finally, future evolution of extreme rainfall (by means of a MOS method) and its potential impacts in the variation of floods have been discussed.

Paper 234:

The angular propulsion engine – a missing link in tornadogenesis?

Anthony Straatman, Western University, Canada

Jeliazko Polihronov, Western University, Canada


Research on tornadoes outlines a history of empirical and theoretical studies aimed at the causal relationship between processes in a tornadic storm [1-2]. A critical gap in the current understanding of tornadoes is the mechanism that suddenly amplifies rotating winds to tornadic severity and then sustains the tornado for a life cycle that typically spans several minutes to an hour. Researchers agree that a necessary condition for power-up of a supercell tornado is the presence of a supercell storm, which serves as the cold reservoir and suction sink, and a warm lower atmosphere, which serves as the warm, high-pressure reservoir. However, beyond this point, despite the observations of numerous researchers [1, 3-4], considerable uncertainty exists as to how the mature tornado forms. Here we propose that amplification and maintenance occur as a result of formation of the rotating vortex into an angular propulsion engine that extracts energy from the warm source air to provide itself with continuous rotary thrust. A recent study by Polihronov and Straatman [5] presents a thermodynamic model of angular propulsion that unravels the long-standing question of how temperature separation occurs in a vortex tube. The key element is that conditions must exist that enable the fluid to give up energy as propulsion to power the rotating system. The same key element provides a possible answer as to how a rotating conduit of wind can suddenly be amplified and maintained at tornadic intensity. The concept of a thermodynamic engine has been used to explain the function of hurricanes [6], but as the scale and the mechanisms responsible for the genesis of tornadoes are vastly different than for hurricanes, similar modeling has presented significant scientific challenges. Characterization of this crucial link in tornadogenesis and maintenance gives fresh insight into the operation, prediction and possible interruption of tornadoes.

1. Markowski, P. M., Richardson, Y. P., “Tornadogenesis: our current understanding, forecasting considerations, and questions to guide future research,” Atmospheric Research, 93, 3-10 (2009).
2. Bluestein, H. B., “History of tornado research,” American Meteorological Society 25th Conference on Severe Local Storms, Oct 11, Article 6.1 (2010).
3. Wurman, J., Kosiba, K., Markowski, P., Richardson, Y., Dowell, D., Robinson, P., “Finescale single- and dual-doppler analysis of tornado intensification, maintenance, and dissipation in the Orleans, Nebraska supercell,” Mon. Weather Rev., 138, 4439-4455 (2010).
4. Marquis, J., Richardson, Y., Markowski, P., Dowell, D., Wurman, J., “Tornado maintenance investigated with high-resolution dual-doppler and EnKF analysis,” Mon. Weather Rev., 140, 3-27 (2012).
5. Polihronov, J. G., Straatman, A. G., “Thermodynamics of angular propulsion in fluids,” Phys. Review Letters, 109, 054504 (2012).
6. Emanuel, K., “Hurricanes: Tempests in a greenhouse,” Phys. Today, 59, 74-75 (2006).

Paper 235:

Study of the 5th July 2012 severe hailstorm in Pla d'Urgell (NE Spain)

Carme Farnell, Department of Geography, University of Barcelona, Spain

Montse Aran, Meteorological Service of Catalonia, Spain

Muntsa Busto, Meteorological Service of Catalonia, Spain

Jordi Mateo, Meteorological Service of Catalonia, Spain

Nicolau Pineda, Meteorological Service of Catalonia, Spain

Tomeu Rigo, Meteorological Service of Catalonia, Spain

Maite Torà, Agrupació de Defensa Vegetal de les Terres de Ponent, Spain


On the 5th July 2012 a severe hailstorm hit western Catalonia (NE Spain), in particular the Pla d'Urgell county, with hailstones up to 7 cm in diameter. A first evaluation made by the Department of Agriculture estimates that 25.000 Ha of crops were affected, with loses up to 20M euros. Huge damages were reported in orchards (apples, pears and peaches) and corn fields.
The present study examines the episode from different point of views: from the synoptic analysis to an extensive surface fieldwork. Different observation systems where available: METEOSAT imagery, C-band radar network data, total lightning detection records, and automatic weather stations registers. Moreover, the region is covered with hailpads, which will allow the study of the hailstones.
The synoptic situation was characterised by a low located at the north-west of France coast. Its trough axis was orientated from north to south, and was over the centre of the Iberian Peninsula at 12 UTC. Although the strongest synoptic forcing was in the Bay of Biscay, Catalonia was in the warm sector where there was a remarkable divergence zone. At surface level, a relative low developed the previous day in the southeast of the Iberian Peninsula impinges easterly winds on Catalonia. Consequently, the air mass in Catalonia was quite humid and presents convective instability. The radiosounding data of Barcelona showed a CAPE of more than 1936 J/kg at 12 UTC. Another factor that played an important role was the high values of wind shear between 0-6 km and the veering profile detected.
The thunderstorm, which affected the region for only two hours, presented a well-defined configuration of severe convection, and in fact, the observations at surface (wind, hail) are in agreement with the remote sensing patterns.

Paper 236:

Composite mean and anomaly of synoptic conditions for waterspout days over South Aegean Sea (S. Greece)

Ioannis T. Matsangouras, University of Athens, Greece

Panagiotis T. Nastos, University of Athens, Greece


Recent studies and several reports submitted to the European Severe Weather Database (ESWD) and the Greek tornado-report-database at University of Athens, have given evidence that south Aegean Sea is a vulnerable area for waterspouts and funnel clouds to occur. Occasionally waterspouts have been appeared on shore over the north coasts of Crete island, causing significant impacts to the local society. In this study, daily composites (averages) of the mean and anomalies (with respect to 30 years climatology 1981-2010) of synoptic conditions for waterspout days over the south Aegean Sea were quantified, based on NCEP/NCAR Reanalysis datasets for the period September, 1948 to December, 2012. Additionally, an updated climatology of the phenomena is presented and analyzed. The analysis was based on seasonal variability of phenomena and was carried out for the specific barometric pressure level of 500 hPa, the sea level pressure (SFC) and the dynamic Lift index (LI). The interpretation of the results indicate specific seasonal weather patterns both in the lower and the middle atmosphere associated with the incidence of waterspouts and funnel clouds

Paper 237:

Can a physics-based algorithm correctly analyze severe convective wind gusts?

Georg Pistotnik, European Severe Storms Laboratory (ESSL), Germany


Three processes contribute to the kinetic energy of convective wind gusts at the earth's surface: (1) vertical transport of momentum, (2) negative buoyancy due to evaporation of falling precipitation, and (3) “forced” downward accelerations due to hydrometeor load aloft. While these contributions are well known in theory, their representation in a fully physics-based analysis or forecast algorithm still remains challenging. Reasons include the high variability and elusive nature of gust-supporting situations, their accordingly poor representation in numerical models and in observational data, and an insufficient knowledge of those processes which occur below resolved scales.

An algorithm capable of computing the “kinetic gust potential” from 3-dimensional fields of temperature, humidity and wind has been developed. A universal applicability to various modes (analysis, nowcast and forecast) and resolutions was envisaged, as well as a close tie to the underlying physics in order to enable constant advancements. A set of adaptable “steering parameters” was introduced in order to handle sub-scale effects like friction, turbulent entrainment of environmental air into a downdraft, and storm-scale perturbation pressure gradients. This approach allows both to statistically tune a “deterministic” diagnosis of expected maximum gust speeds and to follow an ensemble approach with user-defined probabilistic output.

In a first effort to examine the algorithm’s functionality, it was applied to ERA-Interim reanalyses (1979-2011) and historic severe wind cases according to the European Severe Weather Database (ESWD). The restriction to "positive" cases of actually measured or otherwise documented wind gusts above 25 m/s allows to evade the complicated question of where, when and how an existing gust potential is converted into an actual convective gust event. This presentation focuses on the methodology, presents a statistical evaluation, addresses questions of predictability, and outlines the further path how to make the algorithm fit for a transformation from coarse reanalyses to high-resolution nowcasts.

Paper 238:

Severe squall lines in Hungary. Is it viable to use a simple convective parameter check-list based on hydrostatic NWP in operational forecasting?

Mark Kohlmann, Eötvös Loránd University of Science (ELTE), Hungary

Kornél Kolláth, Hungarian Meteorological Service (OMSZ), Hungary

Kálmán Csirmaz, Hungarian Meteorological Service (OMSZ), Hungary

Tamás Mona, Eötvös Loránd University of Science (ELTE), Hungary


Linear mesoscale convective systems and their associated severe events can cause serious damage to human lives, property and agriculture. To assess the possible scenarios for such events regarding severe storm environments characterized by sounding derived indices, a great number of studies have already been conducted, such as Kuchera and Parker (2005) or Doswell and Evans (2001).

In this study, a series of cases involving severe linear MCS-s over Hungary are examined, considering buoyancy, vertical wind shear, lower and upper-air moisture and other, more recently developed parameters like damaging winds (DMGWIND). Severe cases and false alarm cases are also inspected. Examinations are performed through model analysis derived convective indices opposed to the standard method of relying solely on upper-air measurements of any sort. Convective parameters and indices are calculated from 0.25°×0.25° 3 hourly ECMFW fields and additional radar and observational data is used, all available from 2005 to 2011.

The main goal of this research is to determine whether it is possible to create a simple but efficient decision process for potentially suitable environments for severe sqall lines, based on hydrostatic NWP data

Paper 239:

Recent Developments in Our Understanding of Tornadic Storms

Yvette Richardson, Penn State University, United States


Results in the past decade from field campaigns, numerical models, and theoretical studies have increased our understanding of certain aspects of storm initiation, tornado genesis (and its failure), and tornado maintenance. In particular, we discuss: the influence of small-scale vortices on the water vapor field and storm initiation, the presence of surges within supercell rear-flank downdrafts and their influence on tornado genesis and maintenance, the presence of particular vortex line configurations and their implications regarding possible vorticity generation mechanisms, and the interactions between ongoing tornadoes and nearby vortices.

Paper 240:

A radar-based climatology of tropopause folds and deep convection

Bogdan Antonescu, University of Manchester, United Kingdom

Geraint Vaughan, University of Manchester, United Kingdom

David M. Schultz, University of Manchester, United Kingdom


Tropopause folds have been shown to enhance or inhibit convective storms, but how does a forecaster know which will happen in any given situation? How does the structure of tropopause folds affect the occurrence, location and morphology of convection? As a start to answering these questions, we constructed a 5-yr (2006--2010) radar-based climatology of tropopause folds and convective storms for Wales, United Kingdom. Based on the continous, high-resolution data from a VHF wind-profiling radar located at Capel Dewi, 183 tropopause folds were identified. A maximum of tropopause-fold cases occurred in January with a secondary maximum in July. A clear correspondence was found between the occurrence of folds and the North Atlantic Oscillation index, with a positive phase of the oscillation associated with an increased number of folds in winter 2006--2009, and springs 2007 and 2009. Based on the data from the weather radar network operated by the UK Met Office, a radar-based climatology of convective storms was developed, resulting in 685 cases. Unlike much of the US where a clear annual cycle peaking in the warm season occurs, convective storms over Wales display a relatively consistent occurrence throughout the year, except for a strong minimum in late winter and early spring. Multicellular lines were the most common (43%) with a monthly distribution that peaks in October, followed by isolated cells (33%) most common in May--September and nonlinear clusters (24%) with a maximum in November--January. Convective storms are associated with 57% of the tropopause folds identified in this study. The monthly distribution of these cases reveal a significant seasonal cycle with a maximum in December. The tropopause folds were then classified according to their location with respect to an upper-level trough into four categories: east of the trough, underneath the tropopause fold on eastern side of the trough, underneath the trough and underneath the fold on the western side of the trough. From the 55 tropopause folds observed underneath the tropopause fold on the eastern side the trough, 67.3% were associated with convective storms. The convective mode in this region was dominated by multicellular lines (41.8%). From the 128 the tropopause folds observed underneath the fold on the western side of the trough, 32.8% were associated with a convective storm. In this region the convective mode was dominated by isolated cells (23.4%). These results show that the stronger the synoptic forcing, the more likely storms are to be more highly organized. Future work aims to better understand the reasons for this result.

Paper 241:

Thunderstorm interception with a mobile automated surface weather station in Brazil.

Diogo M. Custodio, UFSM, Brazil

Ernani L. Nascimento, UFSM, Brazil

Mauricio I. Oliveira, UFSM, Brazil

Otávio C. Acevedo, UFSM, Brazil


A prototype portable surface automated weather station (AWS) capable of high-frequency sampling of surface atmospheric features produced by deep convective storms has been deployed in research mode for the interception of thunderstorms in southern Brazil since January 2012. This mobile AWS, which is named "Mesomóvel", follows a configuration similar to the north-american Texas Tech University Stick-Net system, in which the instrumentation is mounted on a tripod. One Mesomóvel prototype was built with the goal of developing local know-how on mobile instrumentation and on storm interception with scientific purposes in Brazil. The focus is on providing research-quality documentation of local and short lived surface convective phenomena that are poorly resolved by the Brazilian synoptic-scale automated observing system and that can occasionally lead to severe weather conditions, such as gust fronts, cold pools, mesohighs, mesolows and heavy rain. Variables measured by the Mesomóvel station (up to 1/3 Hz sampling frequency) include atmospheric pressure, air temperature and relative humidity, wind speed and direction, and rain rate and accumulation.

Results indicate that the Mesomóvel AWS succesfully detects short lived (and sometimes intense) variations on the sampled variables that are typical of features that accompany convective activity and that are usually missed by the operational AWS from the Brazilian surface observing system. Examples are discussed of successful storm interceptions that documented the passage of gust fronts and cold pools, as well as the strategies used for the short-range forecasting and nowcasting in support to the interception activity. Challenges regarding the local road network and data access in real time during the interceptions are also discussed.

Paper 242:

An investigation of the tornadic stage of the Goshen County, Wyoming, supercell of 5 June 2009 using EnKF assimilation of mobile radar data collected during VORTEX2

James Marquis, Pennsylvania State University, United States

Yvette Richardson, Pennsylvania State University, United States

Paul Markowski, Pennsylvania State University, United States

David Dowell, NOAA, Earth System Research Laboratory, United States

Joshua Wurman, Center for Severe Weather Research, United States

Karen Kosiba, Center for Severe Weather Research, United States

Paul Robinson, Center for Severe Weather Research, United States


Mobile radar velocity observations collected on 5 June 2009 during VORTEX2 are assimilated into an idealized supercell simulation using the Advanced Weather Research and Forecasting model and the ensemble Kalman filter (EnKF) technique. The resulting EnKF analyses fill gaps in the dual-Doppler and in situ observation dataset so that a thorough analysis of mesocyclone-scale processes relating to tornadogenesis, maintenance, and decay can be performed. Observations are assimilated at 2-minute intervals into a 50-member ensemble with horizontal model grid spacing of 500 m.

Trends of the single-Doppler-estimated tornado intensity are compared to the EnKF representation of the parent supercell. Emphasis is placed on the buoyancy of the surrounding low-level outflow air, the strength of the low-level mesocyclone-scale circulation, and the relative placement of the tornado to the mid-level updraft and mesocyclone. The trajectories of parcels located in the low-level mesocyclone are traced backward and forward from times representative of the formation, mature, and dissipation stages of the tornado to determine if origins of low-level vertical vorticity differ throughout its lifecycle and to assess the evolution of connections between updrafts near the surface and aloft.

Paper 243:

Toward Developing a Storm-scale Prediction System for Hazardous Weather: An Update

Louis Wicker, NOAA National Severe Storms Lab, United States

David Stensrud, NOAA National Severe Storms Lab, United States


Beginning in the spring of 2010, the U.S. National Oceanic and Atmospheric Administration began funding a new research program at the National Severe Storms Laboratory called “Warn on Forecast” (“WoF”). WoF’s primary goal is extending the lead-time for tornado warnings issued by the U.S. Weather Service offices from their current average lead-time of 15 minutes toward 60 minutes. This requires transforming the current operational process from detection of tornadic precursors within thunderstorms from Doppler radar signatures into one where numerical weather prediction is used augments the entire process. WoF is therefore a program to extend numerical weather prediction down to convective scales (~ 1 km) or even sub-convective scales (~100’s of meters) using high-resolution radar, satellite, and other dense in situ meteorological instrumentation. An important aspect of WoF technology will be the use of probability information, via ensemble forecasting, to help mitigate the uncertainties associated with convective-scale weather prediction.

WoF is a multi-group effort, with NSSL coordinating the research and development of these technologies with several partners, including the Center for the Prediction of Storms (CAPS) at the University of Oklahoma, the Environmental Systems Research Laboratory (ESRL) in Boulder CO, the Cooperative Institute of Mesoscale Meteorological Studies (CIMMS), and two weather service operational groups collocated with NSSL; the Norman OK Weather Service Forecast Office and the U.S. Storm Prediction Center. Social science groups are also incorporated into the program as an important aspect of our research and development is to determine how best to present this new and complex data source to operational forecasters and other potential users.

The focus of this talk will be on the current status of the technical WoF research. The talk will discuss several data assimilation methods being tested for convective scale analysis and forecasting and examples will be shown from these various approaches. Assimilation of radar data at these scales is a very different problem from large-scale data assimilation because the state vector being retrieved is indirectly measured, under-determined, and there are few inherent dynamic balances at these spatial and temporal scales that can be be exploited for the analysis. Nevertheless, some promising results have been already generated. These results as well as our current challenges will be discussed.

Paper 244:

Radar detection of local severe storms in Austria and their impact to aviation.

Rudolf Kaltenboeck, Austrocontrol, Austria


Austria is frequently affected by severe thunderstorms. Interesting storms have been investigated by using the operational polarized Doppler weather radar data. The weather radar network covers entire Austria - the Alpine mountainous areas and the lower parts in the surrounding of the Alps, where mesocale influences are predominant.
Deep convections of several years are classified in different types and systems as well as deviation of storm motion related to low level shear is presented. Radar fine line examples of convergence lines highlight the importance of watching low level radar signatures for dry and wet cases like thunderstorm outflow or triggering of new convection. Doppler and dual polarized weather radar data reveal subscale, mesocale features relevant for aviation like hail detection and quantification, location of up- and downdraft, mesocyclones and strong shear/turbulence areas, and lightning forecasts.
A case study will be shown, where severe convective weather had strong influence to aviation and varied impact to air traffic flow over Austria.

Paper 245:

Comparison of Three Summer Bow Echoes in Serbia

Maja Rabrenovic, Ministry of Interior, Serbia


The objective of this paper is to examine meteorological factors that affect bow echo development. For this purpose, we shall present and compare three summer bow echoes that appeared on the Serbian territory in the period 2003-2008: the first, which was developed from single supercell and observed on 31 August 2003; the second that occurred from stagnant weather pattern with weak synoptic-scale features on 19 June 2007; and the third that emerged from low-pressure frontal system on 8 August 2008. Although all three bow echoes contained elements of severe weather with supercells, heavy rain, large hail and strong straight-line winds, they evolved from different synoptic-scale situations. These events shall be examined and presented using satellite and radar images in the area of Northern, Central and Southern Serbia.

Paper 246:

The Role of Ambient Horizontal Vorticity in Near-ground Rotation of Supercells

Johannes Dahl, North Carolina State University, United States

Matthew Parker, North Carolina State University, United States

Louis Wicker, National Severe Storms Laboratory, United States


It has long been known
that a downdraft is required for supercell tornadogenesis to occur, and that
baroclinicity at the periphery of the downdraft is an important source
of the tornado's vorticity in full-physics simulations and observed supercells.
However, it has also been proposed that mere barotropic
rearrangement of pre-existing vortex lines could explain tornadic near-ground
In this study the authors seek to quantify both contributions.
To this end, high-resolution supercell
simulations were performed, and a new Lagrangian technique, allowing for
a separation of baroclinic and barotropic vorticity, is used to determine
the role of ambient vorticity in vortex genesis.

Preliminary results suggest
that the ambient vorticity is reoriented into the vertical in the main downdraft,
but that the sign of the resulting vertical vorticity may be positive or negative.
This suggests that the ambient vorticity may either reinforce or offset the
baroclinically generated cyclonic vorticity.

Paper 247:

Toy model simulations of baroclinic and barotropic processes in downdrafts

Matthew Parker, North Carolina State University, United States

Johannes Dahl, North Carolina State University, United States


In full simulations of supercells, surface vertical vorticity is rather unsteady and tied to surges in the main downdraft. The configuration of these vorticity maxima, including their orientation, intensity, and storm-relative motion, results from some combination of baroclinic and barotropic processes in the downdraft. Operationally, parameters based upon the low-level wind profile appear to be somewhat skillful in identifying significant tornado days; the physical explanation for this skill is not yet totally clear.
The difficulty in studying these processes within full supercell simulations is that, when the wind profile is varied, many storm-scale features also change.

Here we take the approach of idealized simulations, using the same basic model configuration as Dahl et al (2013, this conference) except that the model is run dry. There is no initial vertical vorticity, and the wind profile is varied to test the role of the orientation of the shear (horizontal vorticity) vectors. There is no parent storm updraft, and downdrafts are artificially triggered by several means. Under a wide range of conditions, surface vorticity in excess of 0.01 s-1 develops in these idealized simulations. Here we assess the causes and sensitivities of these resultant vortices; several key points emerge.

The surface vortices in this idealized study are solely downdraft/outflow phenomena. Even so, it appears that the wind profile (particularly the vertical wind shear) may directly influence downdraft production of vertical vorticity. This may occur due to modifications in either the baroclinic or barotropic production of vorticity, as both effects co-exist in these simple simulations. In many cases, the surface vorticity is often small beneath the zone of barotropic production and the baroclinically-produced surface vorticity tends to move away rapidly from the downdraft. However, in special cases (to be discussed), a steady maximum in vorticity persists next to the downdraft. This situation would seem to be more favorable for tornado genesis and maintenance.

Paper 248:

Obstacles and barriers in research work on historical tornadoes in Central Europe

Thilo Kühne, European Severe Storm Laboratory, Germany

Georg Pistotnik, European Severe Storm Laboratory, Austria

Emmanuel Wesolek, KERAUNOS, France

Pierre Mahieu, KERAUNOS, France

Artur Surowiecki, Skywarn Polska, Poland


With the foundation of the European Severe Storm Laboratory (ESSL, 2002) and the establishment of the European Severe Weather Database (ESWD, 2004), severe storm events could be listed from all over Europe. Notwithstanding, there is a great lack of information regarding severe storm events from the period before the year 2000. This lack of information concerns also tornado events which occurred over parts of Central Europe. While finding previously unknown tornado events in newspaper archives and historical scientific publications, it was found that there are communicative obstacles and barriers in the field of local historical and present coverage. Therefore it was made to focus to analyze reasons and causes of these communicative difficulties.

Referring to the ESWD data base before starting the archive research work, a lack of tornado data could be detected over areas which had changes in political and administrative affiliation.
This mainly concerns German and Polish territories before and after 1945, including former national and occupied area. Focusing on Germany in three time-spatial sections (Pre-WWI (before 1919), after WWI to the end of WWII (from 1920 to 1945), and after WWII), it shows that communicative transmission of severe weather events was under the influence of various changing factors, e.g. administrative official languages, expulsion of local minorities, renaming of towns and areas, razing villages and devastating acts of war.

The data amount of 104 investigated historic tornado reports from the time-period 1363 to 1940 in the local territories of Alsace (France), Sudety (Czech Republic), Dolnośląskie, Lubuskie, Zachodniopomorskie, Pomorskie, Warminsko-Mazurskie (Poland) and the Oblast Kaliningrad (Russia) shows very different local barriers in the fields of language, research possibilities and archiving methodology. According to Poland and Russia, these obstacles prevented scientists to construct a climatology on tornado events for some parts of their national area.

Paper 249:

Dection of cloud convection from COMS

Eunha Sohn, KMA, Korea, Republic of

Minbum Choi, KMA, Korea, Republic of

Sungrae Chung, KMA, Korea, Republic of


NMSC/KMA has been detecting convective cloud from COMS (Korean satellite) observation using module (PGE11, NWC SAF RDT) developed by NWC SAF/EUMETSAT and providing it mainly for aviaton and forecasters.

COMS can normally observe over korean area with interval times of 2 and 13 mins. It means that it is not easy to identify rapidly developing convective cloud within about 10 mins.

We are trying to optimize NWC SAF RDT algorithm by inspecting parameters which make an important role in detecting convective over korean area and looking for the feasibility whether convective cloud develops or decays.

Paper 250:

Statistical Analysis of Extreme Rainfall Events and their Impact on the City of Bogota

Eddy Herrera, Pontificia Universidad Javeriana, Colombia

Canal Clima, Canal Clima, Colombia


Extreme precipitation events are able to trigger significant geomorphic processes and significant magnitude consequences in economic and human losses and generate problems within urban area.
This study identified and analyzed the extreme precipitation events presented between the period of 2008 to 2012, in the city Bogota Colombia, in order to find and determine their effects within the urban area studied.
Statistical analysis applied is intended to determine the degree of association between extreme precipitation event and the kind of impact that this produced in the area.

Paper 251:

Characteristics of Overshooting Tops derived from MeteosatSeviri data.

Monika Pajek, Institute of Meteorology and Water Management - National Research Institute, Poland


Overshooting tops (OST) are the cloud structures extending above Cumulonimbus anvil, which locate strong updraft in convective cell connected with severe weather events, such as:storm, hail, heavy rain, lightnings, tornado, strong wind. One of the techniques that could be used to indicate OST using satellite data is difference of brightness temperature(BTD) between two spectral channels of Meteosat/SEVIRI: WV( water vapour) 6,2 um and IR (infra red) 10,8 um. Positive difference is interpreted, in empirical way, as an overshooting tops occurence. although, physical interpretaton is still examinated.

Wider and wider applicability of overshooting tops satellite products create demand for detailed studies concerning temporal and spatial variability of this phenomena. For this reason comprehensive studies for the whole year in mid-latitude European conditions was presented. Every fifteen minutes data from Meteosat Saviri channels were prepared and processed. Distributions of overshoting tops in different time, spatial and meteorological conditions were investigated. Preliminary verification and comparison with data from Synop ground stations, lightning detection system, radiosoundings and other satellites were analysed together with their correlations.

Detailed statistical study is necessary for better understanding and interpretation of severe convection and overshooting tops phenomena and allow to implement OST products to automatic system. The analysis performed has methodological, operational and cognitive aspect.

Paper 252:

Using the European Severe Weather Database for climatological analyses

Pieter Groenemeijer, European Severe Storms Laboratory, Germany

Georg Pistotnik, European Severe Storms Laboratory, Germany


With the establishment of the European Severe Weather Database (ESWD), a single source for European extreme weather data has become available to the research community. Now that the data set is steadily growing at a rate of over 5000 events per year and consists of more than 50000 events, time is due to assess how well the ESWD lends itself for climatological analyses.

It will be shown that daily and seasonable frequency distributions of events can be derived in a relatively straightforward manner. Also, the peak season of severe weather across Europe can be identified for most areas. However, absolute quantitative results are challenged by spatial and temporal inhomogeneities of the ESWD reports.

It is therefore useful to be able to identify "reliable" subsets of data across which such inhomogeneities do not prevail. In order to do so, reanalysis data (ERA-Interim) were used to compute macroscopic atmospheric variables like CAPE and deep-layer shear and related to the probability of severe weather via a logistic regression.

The results indicate that the reporting rate of severe weather events is reliably high across parts of Central Europe and that the estimated number of unreported cases in early times is large. In many eastern European countries, this situation still continues up to the present day.

Paper 253:

What is the uncertainty of weather warnings? Can we predict it? How can this uncertainty estimate be used?

Martin Goeber, DWD, Hans-Ertel Centre for Weather Research, Germany


Weather warnings are still mostly deterministic forecasts. On the face of it we pretend to predict severe weather to happen somewhere in an area with 100 % certainty. Yet verification of warnings reveals that the conditional probability for something severe to happen, given we forecasted it, varies between 90 % and 5%, depending on: the event to be warned, the lead time of the warning, the observational coverage in the area warned and the strictness of the usage of rigid warning criteria. Thus the forecasted probability (100%) and the observed probability of occurence, given we warned (5-90%) , are very far apart. This is also true for the other side of the warning problem, i.e. the non-warning. Here the probability of something happening, given we did not warn (0 % forecast), is far from 0 %, e.g. between 5-70 %.

Yet often we have knowledge about the uncertainty of severe weather forecasts from forecasters experience as well as dynamical and statistical models. Verification results for all 3 approaches will be presented for gusts and precipitation. They reveal positive skill in the estimation of uncertainty of severe weather forecasts. The usage of this kind of uncertainty information is currently tested with emergency managers in the city of Berlin. First experiences and results from this experiment will also be presented.

Paper 256:

Lightning-related Fatalities and Injuries in Turkey

Seyda Tilev Tanriover, Istanbul Technical University, Turkey

Abdullah Kahraman, Turkish State Meteorological Service, Turkey


A research on lightning fatalities and injuries in Turkey will be presented. In Turkey, a big portion of loss of lives related with convective storms occurs due to lightnings similiar to many parts of the globe. Together with station reports from Turkish State Meteorological Service, newspaper and news agency archives, the online media database were also searched to collect lightning fatalities and injuries between 1939 and 2012. Annual average fatality rate exceeds 20 people during the last decade (~0.3 per million). Most of the victims are young and male. There are a lot of records of dead or injured shepherds, in rural areas, particularly under trees. Geographical, annual and seasonal distributions of the events will be discussed.

Paper 257:

Flood And Flash Flood Resulting From Convective Storms In The SWIO Islands And Their Impacts On The Local Community: Case Studies From Mauritius

Kumar Ram Dhurmea, Mauritius Meteorological Services, Mauritius

Prem Goolaup, Mauritius Meteorological Services, Mauritius


During southern hemisphere summer, the South West Indian Ocean (SWIO) Islands experience a weak cross equatorial flow from the northern tropics and the monsoon trough gradually establishes itself close to 10 °S, reaching an extreme position over central Madagascar around February. Summer is also the rainy and tropical cyclone season.
Occasionally convective storms develop, the oceanic environment of the tropical islands being very buoyant along the inter-tropical convergence zone and also over the land areas due to excessive thermal heating. It has been observed that convective storms are likely to develop under the following meteorological situations: active sea-breeze, easterly wave, dissipating tropical storm, feeder band and rain-band associated with tropical cyclone, and active convection in the monsoon trough. Usually these convective storms develop within hours and cause squally and thundery weather with heavy down pours causing flash floods and disruption in the socio-economic activities within the local community and even causing loss of life and property. The only monitoring tool available is the METEOSAT satellite imageries and therefore early warning to the Island population is quite a challenge. Each situation is briefly discussed and illustrated with examples from the Seychelles (January 2013), Comoros (April 2012), Reunion (January 2013) and Mauritius (September 2008, March 2009).
The second part of this paper compares two flash flood events that occurred in Mauritius, namely: rain-band associated with a tropical storm that crossed the island on 22-24 March 2005 and severe storm triggered by the remnant of a dissipating storm on 26 March 2008.
GIS analysis gave a comprehensive understanding of the degree to which extreme rainfall duration and intensity associated with severe storms is likely to affect the development of flash flood. The study also provides background Information on how extreme rainfall intensity and frequency of event durations ranging from hours to multiple days can have a bearing on the flood impacts, which information can be used by the Disaster Management and Civil Protection Authorities to put in place appropriate mitigation measures.
Results show different flood duration in these two episodes which lasted for almost 48 hours in the former (2005 event) while lasting for only 6 hours in the latter (2008 event). Marked differences in the rainfall intensities are observed in these two events with extreme values limited to 25 mm/half hour in March 2005 while peaking up to 42 mm/half hour in March 2008. The results also show that rainfall need not be continuous over several days before soil saturation that may lead to flood, but rather a flash flood can occur in a matter of few hours exacerbated by poor or inadequate drainage.
This study also highlights the impact of the aerial extent of storm covering an area. In March 2008 the whole of Mauritius was under very active clouds leading to unprecedented runoff once the rain touched the ground while in March 2005 the island was swept by cloud bands which involved partial coverage and thus discontinuities in extreme rainfall intensity.
The necessity of appropriate tools, such as radar is direly felt in such emergency situation. Radar data will contribute to assess the severity of a thunderstorm prior to its impact and provide adequate knowledge and insight of the spatial and temporal nature of the impending heavy rainfall.

Paper 258:

Severe wind gust thresholds for Meteoalarm derived from uniform return periods in ECA&D

Andrew Stepek, KNMI, Netherlands

Ine Wijnant, KNMI, Netherlands

Gerard Van der Schrier, KNMI, Netherlands

Else Van Den Besselaar, KNMI, Netherlands

Albert Klein Tank, KNMI, Netherlands

Rob Groenland, KNMI, Netherlands


An alternative wind gust warning guideline for Meteoalarm, the severe weather warning website for Europe, is presented. There are unrealistically large differences in levels and issuing frequencies of all warning levels currently in use
between neighbouring Meteoalarm countries. This study provides a guide for the Meteoalarm community to review their wind gust warning thresholds. A more uniform warning system is achieved by using one pan-European return period per warning level. The associated return values will be different throughout Europe because they depend on local climate conditions, but they will not change abruptly at country borders as is currently the case for the thresholds. As return values are a measure of the possible danger of an event and its impact on society, they form an ideal basis for a warning system. Validated wind gust measurements from the European Climate Assessment and Dataset (ECA&D, http://www.ecad.eu) were used to
calculate return values of the annual maximum wind gust. The current thresholds are compared with return values for 3 different return periods: 10 times a year return periods for yellow warnings, 2 year periods for orange and 5 year periods for red warnings. So far 10 countries provide wind gust data to ECA&D. Due to the ECA&D completeness requirements and the fact that some countries provided too few stations to be representative for that country, medians of the return values of annual maximum wind gust could be calculated for 6 of the 10 countries. Alternative guideline thresholds are presented for Norway, Ireland, The Netherlands, Germany, the Czech Republic and Spain and the need to distinguish between coastal, inland and mountainous regions is demonstrated. The new thresholds based on uniform return periods differ significantly from the current ones, particularly for coastal and mountainous areas.

Paper 259:

Convective storm hazards from a reinsurance perspective

Angelika Werner, Willis, Germany


The US Tornado season 2011 brought Severe Convective Storm (SCS) risk back to insurers and reinsurers attention. Aggregated economic losses reached US$ 20-23bn with the Joplin tornado on May alone causing losses of US$ 7bn. However, SCS risk is existent in many regions across the globe, with major events like the Munich Hailstorm 1984 (~US$ 1.1bn), the Sydney Hailstorm 1999 (~US$ 1.5bn) or Hailstorm Wolfgang 2009 (~US$ 1.8bn; all original economic losses).
This presentation will review tornado and hail risk from an insurance perspective and provide a summary on current model developments for SCS, as well as the science behind it."

Paper 260:

Investigation of high shear, low CAPE severe convection in the Southeastern and Mid-Atlantic United States

Jason M. Davis, Dept. of Marine, Earth, and Atmospheric Sciences, North Carolina State University, United States

Keith D. Sherburn, Dept. of Marine, Earth, and Atmospheric Sciences, North Carolina State University, United States

Matthew D. Parker, Dept. of Marine, Earth, and Atmospheric Sciences, North Carolina State University, United States


High shear, low CAPE (HSLC) severe weather presents a significant forecasting challenge in the United States, particularly in the Southeastern and Mid-Atlantic states. These environments are rather similar to the typical severe weather environments in Europe, and are less well understood than the classical U.S. Great Plains severe weather environments. HSLC severe weather, including tornadoes and severe wind gusts, is common during the cool season and overnight, coinciding with the climatological annual and diurnal peaks in false alarm warnings and missed events. Ongoing collaborative research between the US National Weather Service and North Carolina State University is focused on improving the understanding of, forecasting of, and warning operations during HSLC events through an environmental, parameter-based climatology in addition to a radar-based climatology of tornadic and non-tornadic vortices.

Findings from the environmental climatology indicate that while conventional severe weather guidance tools such as the Significant Tornado Parameter are useful in discriminating between significantly severe and non-severe HSLC events, the values at which they are skillful fall well below traditionally accepted thresholds. Furthermore, a combination of lapse rates and low-level shear is statistically more skillful than any existing composite parameter in detecting HSLC environments favorable for significantly severe convection. The radar climatology portion of this study finds that the convective mode distribution for HSLC tornadoes in this region indicates a greater relative frequency of tornadoes associated with quasi-linear convective systems (QLCSs) compared to environments with higher CAPE. This includes tornadoes associated with embedded supercells in QLCSs, whose structure can be less recognizable compared to the discrete classic supercells commonly found in the Great Plains, and tornadoes associated with embedded mesovortices rather than mesocyclones, whose dynamics are still not well-understood. Documentation and assessment of the most useful environmental and radar-derived parameters will be presented.

Paper 261:

Damage surveys at the Royal Meteorological Institute of Belgium

Karim Hamid, RMIB, Belgium


Since 2008, damage surveys are carried out systematically after cases of significant damage in Belgium, produced by severe convection. These surveys are only one part of a procedure to investigate the damage and is part of the Fujita-project at the RMIB. After the survey is done, a detailed investigation is needed of the RADAR data, for both the reflectivity- and velocity scans. By this, specific structures can be found which tell us something about the morphology of the convective storms. In particular, we are searching for indications of strong winds, like Rear Inflow Jets, several types of vortices and supercell structures. All types of strong wind storms are possible in Belgium and even supercells do not seem to be very rare. Especially the low topped supercell is an important source of tornadoes in Belgium. Several cases of tornadoes, associated with this type of severe convection, have been recorded during the last few years. Especially for wind damage, associated with bow echoes or squall lines with embedded bowing structures, a damage survey on the field seems to be very useful. By means of these surveys, we could determine the type of wind damage. It's clear that the investigation of convective wind damage is very complex and multidisciplinairy. The investigator needs to be familiar with the morphology of all types of convection, capable to interpretate the, often complex, RADAR imagery and he must be very alert for details in the damage pattern on the field. Ideally, he should also have some knowledge about structural engineering. Finally, the most challenging part is to put all the information together and make some conclusions about the origin of the wind damage and the occurred wind speeds, using for example the Enhanced Fujita Scale.

Paper 262:

An analysis of convective parameters in the Northern Hemisphere from the ERA Interim re-analysis and CMIP5 projections

Andrew Russell, Institute for the Environment, Brunel University, United Kingdom


Convective parameters (i.e. convective available potential energy (CAPE), convective inhibition (CIN) and vertical shear) are calculated from the European Centre for Medium-Range Weather Forecasts Interim re-analysis (ERAI) for 1979-2012 and from a range of models and experiments from the Coupled Model Intercomparison Project Phase 5 (CMIP5) for 1950-2100. Of particular interest, the analysis shows approximate 20% and 50% increases in mean CMIP5 multi-model European summer 1200UTC CAPE for the 4.5 W/m^2 and 8.5 W/m^2 Representative Concentration Pathways (RCPs), respectively, relative to the Historical period of the CMIP5 analysis. However, in the model spread in this analysis is very high and shows some relationship with model resolution. Further key results, plus the spatial and temporal variability of the convective parameters from the different datasets, will be presented and interpreted in the presentation.

Paper 263:

Review of Supercell and Tornado Dynamics

Robert Davies-Jones, National Severe Storms Laboratory (Emeritus), United Kingdom


Thunderstorms that form in strong vertical wind shear quickly evolve into supercell storms. Supercells are well-organized, monolithic units of vigorous convection. A supercell consists of a rotating updraft (mid-level mesocyclone) and a downdraft that coexists symbiotically with the updraft in an almost steady state.

Doppler-radar and visual observations along with computer simulations reveal that tornadic supercells evolve through three stages. Firstly, the updraft starts rotating and a mesocyclone forms aloft, secondly a narrower vortex, the tornado cyclone, develops near the ground (thereby completing a rotating column that extends from the ground to a great height), and lastly a tornado forms within the tornado cyclone.

The updraft tilts environmental horizontal vorticity upwards. The updraft rotates cyclonically as a whole if this vorticity is streamwise in the updrafts’ reference frame (i.e., in the direction of the storm-relative wind). Updraft rotation and motion are linked so a complete theory of mid-altitude mesocyclones requires understanding of how storms propagate. There are two principle propagation mechanisms, one linear and the other nonlinear.

The process whereby rotation develops in rising air cannot explain how rotation starts near the ground where updrafts are normally weak. Near-ground rotation does not commence without a downdraft. In computer simulations, low-altitude air that goes into rotation has previously subsided and travelled along isotherms (with cooler temperatures to the right of this flow). Horizontal vorticity is generated here and subsequently tipped upwards. The air then enters the updraft where it is stretched vertically and spun up into a tornado cyclone.

This cyclone collapses into a tornado only if the model includes surface friction, which paradoxically causes the extreme upward and rotary winds. With friction, inflowing air parcels penetrate much closer to the rotation axis and revolve much faster (like a spinning ice skater) despite some loss of angular momentum to the ground.

Paper 264:

The NWS Storm Prediction Center: An overview, and a look at new techniques

John Hart, NWS Storm Prediction Center, United States


The Storm Prediction Center (SPC) in Norman Oklahoma is tasked with forecasting severe thunderstorms and tornadoes across the contiguous United States. The center has been in operations for over 60 years, and has developed numerous forecasting methods that remain in use today. This presentation will provide an overview of the philosophy and operations at the SPC including an example of products issued for the April 27, 2011 tornado outbreak. Several new tools and methods that have only recently been put into use at SPC will also be discussed.

Paper 265:

Collaborative adaptive weather radar network for major metropolitan regions: The Helsinki implementation

V. Chandrasekar, University of Helsinki and Colorado State University, Finland

Dmitri Moisseev, University of Helsinki, Finland

Roberto Cremonini, University of Helsinki, Finland

Markku Kulmala, University of Helsinki, Finland

Jarmo Koistinen, Finnish Meteorological Institute, Finland

Yrjo Viisanen, Finnish Meteorological Institute, Finland

Ari-Matti Harri, Finnish Meteorological Institute, Finland

Heikki Pohjola, Vaisala Oyj, Finland

Heikki Turtiainen, Vaisala Oyj, Finland


Public safety, transportation and water managers require accurate and timely quantitative, weather and precipitation information (QWPI) in order to make decisions regarding infrastructure operations and resource allocations. In addition to public safety, existing infrastructure dedicated to transportation, water supply, and waste water can be negatively impacted by both long- and short-duration weather events. During the recent times, the migration of population from rural to urban regions has already tipped the scale, with majority of the world living in urban regions, while at the same time increasing the vulnerability of metropolitan population, to natural hazards. Consequently high resolution (in space and time), quantitative weather and precipitation information systems are becoming integral part of smart, resilient and sustainable cities.
The current hazardous weather monitoring and forecasting system is inadequate for many coastal cities around the world. Advanced quantitative, weather and precipitation information system can also aid water managers in securing supplies while mitigating flood risk and minimizing water quality impacts to the coastal regions from storm runoff and combined sewer overflows. The Center for collaborative Adaptive Sensing of the Atmosphere (CASA) has demonstrated the advantage of the collaborative radar networks, that are being deployed in several metropolitan regions around the world ( Dallas Fort Worth, Tokyo, Osaka). These radar networks are based on small inexpensive X band radars. This paper presents possibly the first deployment of a metropolitan adaptive weather radar network using C –band radars. This paper will examine the scale process of using C band versus X band radars, for Collaborative Adaptive Radar Networks (CARN). While the radar network forms the anchor, integration of other sensors within this network will also be explored. While the Helsinki implementation may not encounter all the hazards, it will serve as an international demonstration test-bed to be implemented in other regions of the world.