Session 11 | Collection of storm data, historical events and damage assessments

Session 11

Collection of storm data, historical events and damage assessments
| Tue, 09 May, 09:00–10:45 (EEST)|Main Conference Room
| Attendance Thu, 11 May, 14:30–16:00 (EEST) | Display Wed, 10 May, 09:00–Thu, 11 May, 18:30|Exhibition area
Orals |
Tue, 09:00
Thu, 14:30

Orals: Tue, 9 May | Main Conference Room

Session 11
keynote presentation
David Sills and Gregory Kopp

The Northern Tornadoes Project (NTP) was founded in 2017 at Canada's Western University, supported by social impact fund ImpactWX, with the aim of better detecting tornado occurrence, improving severe and extreme weather prediction, mitigating against harm to people and property, and investigating the future implications of climate change.

The project was initially limited in scope – attempting to find at least a few undocumented tornadoes in the northern forests of Ontario and Quebec. After demonstrating that the NTP could more than double the annual tornado count in those provinces, the NTP set out to detect, assess and document every tornado that occurs across the country. We also forged other partnerships with Western Libraries, University of Manitoba, York University, Pelmorex’s The Weather Network, Instant Weather, and CatIQ. Research collaborations were undertaken with a number of academic institutions inside and outside of Canada.

As the project evolved, the NTP required new techniques and technologies to allow us to meet our ambitious scientific goals. Cutting-edge remote sensing capacity, including ultra high-resolution satellite imagery and piloted and remotely piloted aircraft systems, needed to be utilized and the latest processing techniques adopted. New means of assessing wind damage were employed. Even a new set of definitions related to tornadoes and related phenomenon had to be developed. The societal impacts of significant events are also increasingly being investigated and becoming part of the event record.

Interestingly, as our assessment tools reach ever-higher resolution, new problems emerge. A lower-resolution image showing what might have been considered a wide tornado path through the forest in the past now often shows evidence of a mix of tornado and downburst damage, and thorough tree-by-tree analysis is needed to untangle the two. Thus, methods for rating and documenting tornado and downburst damage must also evolve as resolution increases.

In addition, the NTP is building new ways of using social media reports of severe weather as crowdsourced data, including creating a community of ‘super-contributors’ that provide high-quality evidence. Sophisticated methods for ‘scraping’ social media for key reports are also under development.

As a result of these detection, assessment and documentation processes, we are generating novel research-quality data sets that are being employed by a number of different types of users. All NTP data are open source in order to foster further innovation.

The presentation will provide a number of detailed examples showing how the NTP has shifted the post-event paradigm in Canada.

How to cite: Sills, D. and Kopp, G.: The Northern Tornadoes Project – Shifting the Post-Event Paradigm in Canada, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-32,, 2023.

Session 11
Filip Skop

Dust storms are an unusual and understudied type of severe weather phenomena in Europe, causing low horizontal visibility, hazardous to human health particulate matter concentrations and economic losses. Despite occurring mostly in arid and semiarid climates, dust storms are also being reported occasionally in Poland during severe drought periods. Regions most prone to dust storms in Poland include Greater Poland, Masovian, Kuyavian-Pomeranian and West Pomeranian Voivodeships. A significant horizontal gradient of sea level pressure, as well as convective phenomena, are considered the main causees of dust storms in Poland. Apart from strong wind, low near-surface level relative humidity, low soil humidity and negative Standarised Precipitation-Evaporation Index (SPEI) values also support dust storms' development. Based on a comparison between particulate matter concentration data, obtained from air quality measurement stations and meteorological data obtained from synoptic stations, dozen of convective dust storm days were identified in the Polish warm period (April-September) between 2003 and 2020. Recorded dust events formed most commonly as a cause of thunderstorm's outflow, connected to cold fronts and low tropospheric convergence zones. ERA5 reanalysis combined with atmospheric soundings data were used  in order to determine the environment supportive for convective systems likely to produce a dust storm.  High Lifted Condensation Level values and low humidity in a lower troposphere strongly supported this type of events. 

How to cite: Skop, F.: Climatology of convective dust storms in Poland, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-29,, 2023.

Session 11
Tomas Pucik, Pieter Groenemeijer, Miroslav Singer, David Ryva, Miloslav Stanek, Georg Pistotnik, Rainer Kaltenberger, and Alois Holzer

In the afternoon of 24 June 2021, severe hailstorms affected Austria, Czechia, and Poland and an F4 tornado occurred in southeastern Czechia. Along the 27.1 km long path, it damaged 1200 buildings and caused 6 fatalities and more than 280 injuries. The width of the damage path was extreme for European standards, up to 2500 m across. The zone with significant damage of F2 or stronger was up to 520 m wide. Isolated instances of F4 damage were noted in 3 villages with the destruction of well-constructed brick walls and significant debarking of trees. We discuss the challenges associated with surveying the tornado from an organizational point of view to the strategy onsite. Improvements are proposed to make surveys of such large-scale events more effective.

The event was not well forecast even by expert forecasters present at the ESSL Testbed 2021. Although the environment was clearly conducive for supercells capable of very large hail with high values of CAPE (> 3000 J/kg) and strong vertical wind shear (0-6 km bulk shear > 20 m/s), lower tropospheric shear was forecast to remain fairly weak by most of the NWP models with 0-1 km bulk shear < 10 m/s and 0-1 km SRH < 100 m2/s2. The fact that only one tornado (and of such a high intensity) occurred in the area despite numerous supercells present points to the importance of mesoscale modifications to the environment. We address the storm-scale evolution starting from the merger of two storms through updraft intensification with giant hail production, and subsequently, low-level mesocyclone strengthening and tornado production. We also discuss the importance of local mesoscale boundaries and modification to the environment shortly before the tornado.

The event illustrates a number of difficulties with tornado forecasting in Europe. The first is the lack of sufficient data exchange among countries. The tornado passed within 10 km of the borders of Austria and Slovakia and the tornado-producing supercell formed over Austria. No exchange of automatic station surface observations and volumetric radar data between those countries takes place and this likely limited the situational awareness of forecasters. While the tornado occurred over Czechia, the storm was best detected from a Slovakian radar. Another difficulty was the aggressive filtering of doppler velocity data that masked the core of the low-level mesocyclone preventing forecasters to appreciate the intensity of the event as it unfolded. 

How to cite: Pucik, T., Groenemeijer, P., Singer, M., Ryva, D., Stanek, M., Pistotnik, G., Kaltenberger, R., and Holzer, A.: Damage survey, environment and storm-scale evolution of the giant hail and F4 tornado producing supercell on 24 June 2021, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-143,, 2023.

Session 11
Maarten Reyniers, Laurent Delobbe, and Sylvain Watelet

The detection and prediction of very local weather effects, like wind gusts or hail storms, remains a challenging issue in meteorology. Especially in the case of severe storms, there is often a lack of detailed ground truth data to assess the precise conditions of the event. One promising approach to supplement the information obtained with the classical instruments like radar and satellite, is the collection of citizen weather reports through smartphones. In August 2019, the Royal Meteorological Institute of Belgium (RMIB) added such a reporting feature in its smartphone app. Currently the database of citizen reports holds more than two and a half million records. Since May 2022, the app users are also able to add a photo to their observation.

In this presentation, we describe the general concept of the app feature and how the quality control of the reports and the filtering of the photos are organised. Then, we give some general statistics of the aggregated dataset. We will show that this unconventional observation system brings uncommon sources of uncertainties and biases related to the human nature of the observations. The collected data have already been exploited in several use cases at RMIB, such as the verification of the official weather warnings and the forecasts per commune as sent out by the weather office. We will further focus on the verification of the operational radar-based hail detection algorithm used at RMIB. Among all reported observations, 23 % concern precipitation (rain, snow or hail) and 0.7 % hail. We will examine the extent to which such a dataset allows evaluating the performance of the algorithm in terms of probability of correct and false detections. These applications show that the citizen observations are an extremely valuable new source of very localised information for many applications in research and operations at RMIB.

How to cite: Reyniers, M., Delobbe, L., and Watelet, S.: Citizen observations via smartphone in Belgium: data collection and applications, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-50,, 2023.

Session 11
William Gallus, Elizabeth Tirone, Subrata Pal, Somak Dutta, Ranjan Maitra, Jennifer Newman, and Eric Weber

In the United States, the official database of severe thunderstorm wind reports arguably has more serious deficiencies than those of tornadoes and hail. Roughly 90% of the thunderstorm wind reports in the Storm Events database during the period 2007-2021 are estimates without any nearby measurement, and the fact that 40% of the estimates have a value of exactly 50 knots compared to only 13% of the measurements strongly suggests that many may be overestimates since 50 knots is the minimum threshold to be considered a severe wind.   The problems in the database negatively impact development of new forecasting tools and verification.  We have tested six different machine learning approaches, training on roughly 20,000 measured reports during 2007-2017 to create a tool that assigns a probability that any severe thunderstorm wind report is due to winds of 50 knots or greater.  Training is based on date, time, location, and episode and event narrative data from the Storm Events database along with 31 near-storm weather parameters from the Storm Prediction Center mesoanalysis output.  In addition, population density and elevation are used.  Land use and radar reflectivity were also tested but found to not improve the performance. The best-performing algorithm, the Stacked Generalized Linear Model has been found to show large skill with Areas Under ROC curves as high as .90 and Brier Scores around 0.1.  When a supplemental sub-severe database is added for testing, reliability is shown to be good.  Subjective evaluations from testing during three years of NOAA Hazardous Weather Testbed Spring Forecast Experiments have been favorable and will be discussed, along with implications for forecasters.  A recent test found that the average probability for estimated 50 knot wind reports is only 57% whereas it is 81% for measured 50 knot reports, supporting the view of many forecasters that overestimates are a large problem among the estimated reports in the database.

How to cite: Gallus, W., Tirone, E., Pal, S., Dutta, S., Maitra, R., Newman, J., and Weber, E.: A machine learning approach to mitigate problems with estimated winds in severe thunderstorm wind damage reports, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-6,, 2023.

Session 11
Pieter Groenemeijer, Alois M. Holzer, Thilo Kühne, and Tomáš Púčik

Here we report on the development of the International Fujita (IF) scale, developed by ESSL and partners, which provides a globally applicable framework for rating tornado and convective wind damage. 

Around the world, tornado damage is rated using the Enhanced Fujita (EF-)scale, the original Fujita scale, the T-scale, or national adaptations of the Enhanced Fujita scale. ESSL started the development of the IF-scale when noted that the original Fujita scale damage description provided insufficient guidance for rating tornadoes in Europe, mostly because of varying sturdiness of damaged buildings, an issue that Fujita himself addressed in his later work. The development of the IF-scale was catalysed by the introduction of the EF-scale in the USA in 2007, which drastically reduced wind speed estimates for higher wind speeds while simultaneously raising it for the F0/F1 boundary compared to the original F-scale: a change apparently motivated by a need to correct for biases in tornado rating practices.

Instead of referring to building types typical for a specific region, the IF-scale instead defines categories of Damage Indicators that are more universal and can be adapted to other regions, using a damage indicator ‘building’ with an attribute ‘sturdiness’. Any building can be a damage indicator after being assigned a level of sturdiness. The IF-scale also retains the original Fujita-scale wind speed estimates, at least until measured data are available that give better estimates. The wind speed that the scale relates to is instantaneous rather than the average wind speed at 10 m above ground, as evidence is accumulating (through high-quality videos and mobile doppler radar measurements) that the wind speeds responsible for tornado damage often have a much shorter duration than the typical averaging periods for wind gust measurements. The IF-scale heavily borrows from experience in regions that have adapted the EF-scale locally, such as the Japanese and Canadian adaptations, and adds additional damage indicators such as for trees, road and rail vehicles, and many other objects.

Recently, ESSL evaluated a preliminary version of the IF-scale by applying it to a violent tornado case in Czechia on 24 June 2021. This application revealed weaknesses that have since been addressed. We will present the latest version of the IF-scale that has become the standard scale used in ESSL’s record of severe weather in Europe, the European Severe Weather Database.

How to cite: Groenemeijer, P., Holzer, A. M., Kühne, T., and Púčik, T.: The International Fujita Scale and its implementation, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-151,, 2023.

Posters: Thu, 11 May, 14:30–16:00 | Exhibition area

Display time: Wed, 10 May 09:00–Thu, 11 May 18:30
Session 11
Jerome Schyns, Christoph Gatzen, and Lisa Schielicke

In the six-year period of 1965 to 1971, an unusually high number of violent tornadoes of category F4 and F5 was observed in Europe including six F4 tornadoes in Switzerland, France (2), Italy (2) and Germany as well as one F5 tornado in France. We use this tornado series to analyze the conditions present in violent European tornadoes.

ESWD (European Severe Weather Database) data, ERA5 reanalysis data and historic weather maps are used to classify the large-scale weather patterns with respect to the strength of synoptic-scale forcing, the position of the jet stream across Europe and relative to the tornadic thunderstorms, location of fronts including drylines, coastal fronts, and pre-frontal convergence zones that could have influenced the thunderstorms development. We use an ingredients-based method to analyze the large-scale flow, focusing on the origin and advection of lapse rates, low-level moisture, zones of lift, and vertical wind shear. For example, we distinguish between situations with a large area affected by an overlap of CAPE and strong vertical wind shear and situations where only a small “sweet spot” could have influenced the development and organization of the thunderstorms.

The vertical profile close to the convective storms are analyzed in ERA5 data that are compared with observation data. Here, we concentrate on the profile of vertical wind shear and the occurrence of low-level streamwise vorticity, expected cloud base, and expected cold pool potential due to evaporative cooling. Additionally, we estimate the motion vectors of supercells.

Proximity soundings serve to initialize supercell simulations using cloud model 1 (CM1, using the pre-configured, idealized supercell set-ups). We discuss the development of convection in the idealized model simulation and characterize the thunderstorms with respect to the lifted condensation level, the height of the mesocyclone, storm motion vector, and col pool potential. The results are compared to findings from the literature.

How to cite: Schyns, J., Gatzen, C., and Schielicke, L.: Revisiting the unusual period of F4/F5 tornadoes in Europe from 1965 to 1971, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-106,, 2023.

Session 11
Martin Lainer, Killian Brennan, Jérôme Kopp, Samuel Monhart, Daniel Wolfensberger, Alessandro Hering, and Zaira Schauwecker

Hail is a major threat connected to severe thunderstorms and an estimation of the hail size is important to issue warnings for the public. Radar real-time products exist that estimate the size of the expected hail. For the verification of such products, ground based observations are necessary. Automatic hail sensors, as available within the Swiss hail network, can provide information about hail diameters observed on the ground. Unfortunately, due to the small size of these sensors (e.g. 0.2 m2) the estimation of the hail size distribution (HSD) can have large uncertainties. To overcome this issue, aerial drone-based 2D orthophotos can be analyzed by using state-of-the-art custom trained AI-object detection models to identify hail stones in the images and to estimate the HSD.

A large right moving supercell with a lifespan of more than 6 hours crossed the midlands of Switzerland from south west in the afternoon of 20th June 2021. The hail swath of this classical supercell was intercepted near Entlebuch and aerial images of the hail on the ground were taken by a DJI Matrice 300RTK drone immediately after the storm has passed. The drone was equipped with a 50 megapixels full frame camera. The average ground sampling distance is 1.5 mm per pixel, which is set by the mounted camera objective with a focal length of 35 mm and a flight altitude of 12 m above ground level.

A 2D orthomosaic model of the survey area (soccer field) is created based on 116 captured images during the first drone mapping flight. The orthomosaic covers an area of about 750 m2 and is then used to detect hail by using a region-based Convolutional Neural Network (Mask R-CNN) model. First, we characterize the hail sizes based on the individual hail segmentation masks resulting from the model detections and investigate the performance with respect to manual hail annotations from experts that are used as validation and test data sets. We present the final obtained HSD from more than 18000 hail stones (Dmax = 39 mm, Dmed = 9 mm) and compare it with nearby automatic hail sensor observations and weather radar based hail products like MESHS (Maximum Expected Severe Hail Size).  Furthermore, we provide first insights into hail melting processes that can be inferred from the information retrieved from a total of 5 subsequent flights performed with the drone within about 20 minutes after the passage of the supercell.

How to cite: Lainer, M., Brennan, K., Kopp, J., Monhart, S., Wolfensberger, D., Hering, A., and Schauwecker, Z.: Drone-based hail observations and the retrieval of the hail size distribution after a supercell passage in summer 2021 in Switzerland, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-5,, 2023.

Session 11
Jérôme Kopp, Agostino Manzato, Olivia Martius, Urs Germann, and Alessandro Hering

We present an investigation of the first observations from the Swiss Hail Network Project, a network of 80 fully automatic hail sensors (Kopp et al., 2022) installed between 2018 and 2020 in the three most hail-prone regions of Switzerland: the Jura, the Napf and Ticino (NCCS 2021). Those sensors provide new hail ground-based information, not only about each hailstone size (estimated by an indirect measures of its kinetic energy), but also about the precise timing of hailstone individual impact (Löffler-Mang et al., 2011), allowing to obtain a time-resolved hail size distribution (HSD).

More specifically, we investigate the point (local) duration of hailfalls, the event hit rate (impacts per second) and time-resolved HSD. We also present and discuss the potential sources of uncertainty specific to the hail sensor, such as the dead time (minimum time between two consecutive observations).

We then compare our observations to measurements from an hailpads network in northeastern Italy (Manzato et al., 2022). While our sample is still limited (around 10’000 hailstone impacts registered during 4 warm seasons) with respect to the hailpads records (29 warm seasons), we found that the HSD obtained with both measurements’ devices are very close to each other.

Finally, we discuss the further combination of the sensor data with radar hail products (Kopp et al., 2022), the exceptionally high density of crowdsourced hail reports collected in Switzerland (Kopp et al., 2022) and the recent drone measurements of hail (Martin Lainer et al., abstract ECSS2023-5), which could pave the way to new and exciting research avenues on hail understanding and forecasting.

Kopp, J., Schröer, K., Schwierz, C., Hering, A., Germann, U. and Martius, O. (2022), The summer 2021 Switzerland hailstorms: weather situation, major impacts and unique observational data. Weather.

Löffler-Mang, Martin, Dominik Schön, and Markus Landry. 2011. « Characteristics of a New Automatic Hail Recorder ». Atmospheric Research 100 (4): 439‑46.

Manzato, Agostino, Andrea Cicogna, Massimo Centore, Paolo Battistutta, and Mauro Trevisan. 2022. « Hailstone Characteristics in NE Italy from 29 Years of Hailpad Data ». Journal of Applied Meteorology and Climatology, août.

NCCS (2021) National Centre for Climate Services : Hail climatology Switzerland., accessed 26 December 2022

How to cite: Kopp, J., Manzato, A., Martius, O., Germann, U., and Hering, A.: Unique observational data from an automatic hail sensors network in Switzerland, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-19,, 2023.

Session 11
Jannick Fischer and Michael Kunz

Information about the hail threat of a thunderstorm is typically limited to rather indirect data from remote sensing or reanalysis. Ground observations of maximum hail diameters provide a more accurate assessment but often suffer from other problems, such as limited or non-uniform coverage. Despite these shortcomings, the above data sources are commonly used for nowcasting of hail producing storms and in hail climatologies. However, only very few studies have actually compared the skill of these different proxies. One reason for this is the lack of ground truth data which could be used to verify whether damaging hail was falling. This is because of the described issues with hail reports, and the fact that insurance datasets, which could provide a more reliable confirmation of hail damage, are usually not made available for research.

To fill this gap, this study uses a 5-year dataset of crop damage claims of a German agricultural insurance. These insurance claims cover large parts of Germany and most accurately reflect hail damage during the growing season of crops from May to August, which is also the time of year with the strongest thunderstorm activity. The damage claims are used to verify and compare common proxies for hail, which include the radar-based (I) column maximum reflectivity and (II) hail tracks using the more refined TRACE3D tracking algorithm, (III) lightning density, (IV) European Severe Weather Database hail reports, (V) observed overshooting tops from geostationary satellites, and (VI) microwave imager hail signatures from polar-orbiting satellites. Their skill in predicting damaging hail is assessed by categorical verification with probability of detection, false-alarm rate, and Heidke Skill Score, including a sensitivity analysis to varying thresholds.

Preliminary results based on 30 events in 2014 indicate that none of the proxies alone can predict hail damage with high accuracy. However, all of them show at least some skill, except for the microwave imager. The radar-based predictors show the largest skill on average. If these findings can be confirmed over more cases, while also including null cases, this would support the use of proxies I-V for hail climatologies but discourage nowcasts of hail for an individual thunderstorm with one of the proxies alone.

How to cite: Fischer, J. and Kunz, M.: A Verification of Hail Proxies with Insurance Data, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-80,, 2023.