Session 5

On June 30^th, 2021, a low-pressure formed in the center of Europe, causing the advection of warm air and the infiltration of cold air from the rear sector of the low-pressure field. This resulted in the formation of a strong convergence line in the Moldavia region, which generated a multicellular system. We analyzed the forecast atmospheric sounding for this date at 12 UTC in various locations in Moldavia, using the European Centre for Medium-Range Weather Forecasts (ECMWF) to identify areas with high convective potential. The planetary boundary layer (PBL) was determined through the analysis of selected atmospheric sounding, and a well-mixed PBL was observed. As the air particle rises, it cools adiabatically, carrying moisture with it. However, slight variations in the height of PBL were noted at different locations, with the northern-central part of Moldavia experiencing more convection than the southern part, which was less affected by severe weather. The region of the atmosphere closest to the ground surface that is directly influenced by it in terms of turbulence, wind direction and speed, moisture, heat, and convective activity is known as PBL. Most of the water vapor and aerosols in the atmosphere are concentrated in this layer, where turbulent mixing occurs and atmospheric fronts and their associated phenomena are manifested. It is important to analyze the height of PBL and how it varies, as it does not have the same height everywhere and changes over time and space. In our work we determined the height of  PBL through the use of atmospheric sounding.

Keywords: forecast, convection, PBL, atmospheric sounding

How to cite: Cosmina, A., Ingrid-Mihaela, M., Diana-Corina, B., Adrian, T., and Marius-Mihai, C.: The extreme phenomena from the end of June 2021 and the variations of the planetary boundary layer height, in Moldavia region, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-23, https://doi.org/10.5194/ecss2023-23, 2023.

The precise forecast of convective cells is essential for meteorological services as they can be accompanied by life-threatening severe hail, wind gusts, or heavy rain. However, state-of-the-art NWP models usually possess update frequencies of several hours so that forecasters must use predictions that are outdated when new thunderstorm cells develop. NWP models do often accurately simulate the intensity of convective cells, but with shifts in space and time. Object-based nowcasting algorithms with higher update frequencies became necessary to deliver information on the evolution of convective storms for the first two hours since observation. Furthermore, the combination of nowcasting and model data enables the relocation of simulated cells towards observed cells.

Many deterministic object-based nowcasting tools as DWD’s KONRAD3D algorithm assume that detected cells will have persistent intensity. Within the SINFONY (Seamless INtegrated FOrecastiNg sYstem) project at DWD, we aim at modelling the life-cycles of storm cells in a truthful way and capturing the uncertainties of object-based nowcasts. Hence, we extended our nowcasting algorithm towards an ensemble prediction system called KONRAD3D-EPS. Each ensemble member is initialized by drawing from parameterized distributions of storm lifetime and maximum severity. Inspired by previous studies, e.g. Wapler (2021), KONRAD3D-EPS uses a set of horizontally flipped parabolas to model the life-cycle of convective cells in terms of their severity. In case of redetection of a convective cell, the algorithm corrects the previously estimated lifetime and severity maxima. Thus, the parabolas can be adapted individually for any convective storm in any weather condition.

Besides life-cycle predictions, KONRAD3D-EPS delivers information on the probability of thunderstorm occurrence for the next 2 hours depending on detected cells and their severity. In order to condense the ensemble data, we also provide the representative member for each convective cell. This is done by applying the pseudomember algorithm by Johnson et al. (2020) to the ensemble data.

We will give an overview of our probabilistic object-based nowcasting algorithm KONRAD3D-EPS and present its predictions for prominent example cases. Moreover, we will show first verification results.

How to cite: Josipovic, L., Pante, G., Schnoor, I., Brechtel, A., and Blahak, U.: Object-based Ensemble Prediction System KONRAD3D-EPS, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-49, https://doi.org/10.5194/ecss2023-49, 2023.

How to cite: Rempel, M., Schultze, M., and Blahak, U.: A New Ensemble System at DWD for Seamless Short-Term Forecasts of Convective Precipitation, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-73, https://doi.org/10.5194/ecss2023-73, 2023.

Terminal aerodrome forecast (TAF) is a standard ICAO product used for flight planning purposes worldwide. It is created by forecasters and states forecasted weather conditions significant for aviation and their changes for the airport in the next 24 hours. TAF Verification in Croatia control follows the approach proposed by Mahringer (2008).

Forecasting of convection is challenging and nowadays mostly leans on ingredients-based methodology. Ingredients for deep moist convection are never ideally known–especially mesoscale lift–so uncertainty in local initiation is always present. On the other hand, the standard reference area for thunderstorm (TS) forecast in TAF is a 16 km diameter circle around the airport. The commonly used procedure of verifying TS in TAFs is against observed TS stated in standard reports for the airport (METAR) where a thunderstorm is observed in the same area. But is this enough and fair considering the inherent spatial uncertainty of convection initiation?

For example, the forecaster predicts conditions for TS at the airport for an afternoon with a probability of 40%, and TS was observed 40km away. There is no doubt that it was a  miss event for the (small) airport area but still, TAF had some value. On the other hand, when TS is forecast with high probability and it was not observed within 200km of the airport it is apparently a false alarm. 

To do a fairer TAF verification we used lighting data network (LINET) data and adapted it for several diameters around airports: 16, 30, 50, 100, 150, and 200 km. Hourly series with observed thunderstorms were derived from raw data and verified with forecasted ones from TAFs. We also tried to investigate differences in using this TAF verification method on the homogeneous mainland of Croatia, at the coast, or on complex topography near mountains.

Our results reflect the inherent spatial uncertainty of convection and could be used in analyzing and diagnosing where more improvements could be made in local thunderstorm forecasting.

How to cite: Jurkovic, J. and Soljan, V.: Addressing spatial uncertainty of convection in thunderstorm forecast verification, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-121, https://doi.org/10.5194/ecss2023-121, 2023.

Severe weather events, such as intense precipitation, hail and thunderstorms, are common summer phenomena in Bulgaria and are associated with large economic losses. An active hail suppression is taking place in North West and South Central Bulgaria over the dense agricultural regions from May to September. A joint venture between Sofia University "St. Kliment Ohridski'' and the Bulgarian Hail Suppression Agency contributed to the project “Balkan-Mediterranean real-time severe weather service” (BeRTISS, 2017-2020). As part of the BeRTISS project, a pilot operational service was established by exploiting GNSS tropospheric products in support of safety, quality of life and environmental protection in the region. To facilitate the service, in 2018 a GNSS network with 12 reference stations was installed and since February 2020, the Sofia University GNSS Analysis Center provides operational near-real time products. First results of combining GNSS derived Integrated Water Vapour (IWV) and Instability Indices (InI) are reported for Sofia plain for the period May-September (2010 - 2015). Based on statistical regression analysis, classification functions are obtained that contribute to the thunderstorm forecasting skill. The majority of the classification functions combining IWV and InI are reported with the best performance in May, followed by June and September. In this work, IWV and InI classification functions are presented for South Central Bulgaria. The first results for the period May-September 2020-2021 indicate the probability of detection 0.71 for IWV, 0.85 for InI, and 0.89 for IWV and InI combined. While false alarms decreased from 0.42 for IWV to 0.39 for InI and 0.28 for IWV and InI combined. These results will be implemented in the GNSS Strom Demonstrator web portal.  

How to cite: Slavchev, M., Guerova, G., and Dimitrova, T.: Combining GNSS meteorology and Instability Indices for derivation of classification functions in South Central Bulgaria., 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-125, https://doi.org/10.5194/ecss2023-125, 2023.

Storms deeply impact human safety and economic activities by generating severe hazards such as large hail, strong wind gusts or floods. The more critical phenomena often imply mesoscale convective systems (MCS) that originate from the aggregation of intense cells. The induced circulation reinforces the system and maintains it for a long time.

Convective-scale numerical weather prediction (NWP) models are now able to simulate realistically such storms. Nevertheless, model and displacement errors often alter the forecast accuracy and ultimately the weather alerts efficiency.

In order to study the predictability of MCS, present work attempts to automatically detect, track and characterize such system in the outputs of the operational high resolution nowcasting model used at Météo-France, Arome-NWC. Such system provides 6h forecast every hour by updating the last AROME-France forecast towards the latest conventional and radar observations, thanks to a 3D variational (3DVar) data assimilation system.

Three methods are developed and compared. The first one applies a segmentation algorithm on images of maximum simulated reflectivity (Zmax). The second one applies a watershed transformation algorithm on Zmax, using as seed the 10,8 µm brightness temperature (TB10.8) from the SEVIRI imager onboard geostationary satellite MSG, that is diagnosed by applying the RTTOV radiative transfer algorithm to simulated variables. Finally, the last one trains a convolutional neural network (CNN) from MCS that are hand-labelled from simulated Zmax and TB10.8 images. Subjective analyses and objective evaluation based on objects-oriented scores depict the third approach as the more reliable.

The operational benefits of this object-oriented approach are investigated on several convective cases using synthesis plots. Superimposing the MCS detected in the different available hourly runs of Arome-NWC, as well as comparing tracked parameters within the MCS throughout the different forecasts, help in assessing the behavior of the successive runs. That could facilitate the analysis of the forecaster in tricky situations, leading to a positive impact on the weather alert efficiency.

How to cite: Arnould, G., Montmerle, T., Moisselin, J.-M., and Rottner, L.: A Deep Learning approach for MCS detection: application for storm tracking and nowcasting, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-38, https://doi.org/10.5194/ecss2023-38, 2023.

Several previous studies have shown that near-surface thermal or outflow boundaries might have a significant impact on supercell development – especially low-level mesocyclone intensification. Supercells that cross surface boundaries at an acute angle usually produce rapid and intense tornadoes on the cool side, even local tornado outbreaks, and in some cases, large hail and damaging wind gusts can occur as well. Similar interactions can be present in the Carpathian Basin as well, however, the investigation of these processes in this region is more difficult due to the complex terrain. This complexity can lead to processes on a much more localized scale. Based on observations, these local thermal baroclinic zones are typically associated with shallow surface lows over Hungary. The surface confluent flow and the equivalent potential temperature gradient zone created by preceding precipitation bears substantial horizontal vorticity near the ground that might be favorable for the low-level intensification of mesocyclones in supercells.

In this research, we investigate these local baroclinic zones and their influence on supercell intensification using non-hydrostatic WRF (Weather Research and Forecasting) real and idealized simulations. We inspect how the presence of the surface baroclinic zone affected the dynamics of the simulated thunderstorm. The gained results may add significant contributions to our understanding of these processes, and support the early recognition of their patterns in the operational routine.

How to cite: Komjáti, K., Csirmaz, K., Breuer, H., Kurcsics, M., and Horváth, Á.: Supercell interactions with surface baroclinic zones in the Carpathian Basin, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-39, https://doi.org/10.5194/ecss2023-39, 2023.

Warm seasons of 2021 was quite exceptional if we consider the hail occurrence in the continental part of Croatia which attracted a lot of public attention due to the announced changes in the Law on the Hail Suppression System. One of the events that sparked a great debate in the public occurred on 25 June near the town of Požega due to hail size and damage, although this event was very well forecasted, and orange warning for thunderstorm was issued one day in advance.

This work aims to explore the atmospheric conditions present during this hailstorm but also a forecasting potential of such events using the convection-permitting km-scale numerical models.

On that day ahead of an advancing trough that stretched from Scandinavia to the western Mediterranean all necessary ingredients for severe deep moist convection were recognized. Convective initiation started first in Bosnia and Herzegovina ahead of the surface cold front. For damage in the vicinity of town Požega, most important were convective cells that were advected into the area of Slavonian Posavina. They experienced "explosive" development, and then moved in the southwestern and southern flow across western Slavonia. The estimated convective available potential energy of the air parcel raised from the ground to the level of free convection (SBCAPE) in the Zagreb area, based on radiosounding at 12 UTC was about 2700 J/kg and the deep layer wind shear was estimated at 20-25 m/s, which favored the formation of a supercell thunderstorm.

Here, convection-permitting km-scale WRF model is utilized to inspect the ability of the WRF model to reproduce the atmospheric conditions leading to formation and evolution of an extremely damaging hailstorm. Additionally, the ability of HAILCAST and Lightning Potential Index (LPI) diagnostics to reproduce the main characteristics of observed hail and lightning is explored. HAILCAST is a one-dimensional hail growth model that forecasts the maximum hail diameter at the ground. Similarly, LPI highlights the areas with the potential for developing lightning activity. Moreover, the sensitivity experiments are performed to investigate the impact that convection parameterization has on the simulated timing and characteristics of convection.

The results reveal that the WRF model, when run at the km-scale resolution (1.5 km) can reproduce the synoptic and mesoscale conditions present during an extremely damaging hailstorm occurring over Croatia. Moreover, HAILCAST and LPI diagnostics reproduce hail and lightning characteristics comparable to those observed.

How to cite: Renko, T., Malečić, B., Mikuš Jurković, P., Kozarić, T., and Horvath, K.: Hailstorm in Eastern Croatia - analysis and forecasting potential by convection-permitting numerical model, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-41, https://doi.org/10.5194/ecss2023-41, 2023.

Severe thunderstorm with gale-force and most likely hurricane-like wind gusts, heavy rain showers and locally large hail affected area of Bjelovarsko–bilogorska county in the central part of Croatia in the afternoon hours on 15^th September 2022. Some locals also mentioned the possibility of a tornado. The thunderstorm caused significant damage to property, crops and forests estimated at approximately 8 million EUR. A natural disaster was declared in two cities and five county municipalities.

Due to very high impact and even the possible occurrence of a tornado, detail synoptic and mesoscale analysis of severe thunderstorm was performed. Moreover, for the first time in recent history, an official on-site meteorological inspection of affected areas was done. That resulted in abundant photo and video materials which were used for the analysis of damage patterns and determination of the phenomena that caused the damage.

On 15th September 2022, the ridge over Croatia weakened and the southwesterly flow strengthened bringing a significant amount of moisture. The dew point temperature was between 16 and 18 °C. A cold front was approaching from the northwest, and in the upper levels a colder air was advected above the very heated surface, due to which the instability of the atmosphere increased. According to radio sounding measurement in Zagreb–Maksimir at 12 UTC, the most unstable CAPE (MU CAPE) was 888 J/kg, and 0-6 km bulk shear was 24.5 m/s which are characteristic values of the environment where supercells can form. The wind shear was mainly concentrated in the lower layers of the troposphere, from 0 to 3 km (23.6 m/s), and it was also significant in the layer from 0 to 1 km (12.1 m/s), which favors the formation of tornadoes.

Radar reflectivity analysis confirmed the presence of a supercell, and a bow-shape was detected what is characteristic for storms with damaging winds. In the satellite imagery during the mature phase of the thunderstorm, the overshooting tops were observed indicating very strong convective updraft. By analyzing the forest damage areas from the ground and using the aerial drone footage, no evidence of wind spinning was found. All trees were knocked down in the approximately same straight-line direction, only slightly divergent damage pattern was observed. Such a damage pattern is usually associated with downbursts.

How to cite: Mikus Jurkovic, P., Kozaric, T., Kalin, L., Renko, T., and Pirak, T.: Severe thunderstorm in the central part of Croatia on 15th September 2022, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-45, https://doi.org/10.5194/ecss2023-45, 2023.

Short-term warnings for severe thunderstorms are produced at the German Weather Service (DWD) with the support of NowCastMIX, which automatically creates warning areas for the next 60 minutes. In short, NowCastMIX processes meteorological fields from various sources such as NWP, radar, surface station reports and lightning detections. Based on the available data, the potentials for heavy rain, hail and severe gusts are calculated every 5 minutes by a hierarchy of fuzzy logic sets. From these potentials, categorical thunderstorm warnings are issued for detected cells. By condensing the information into clusters, regions are then identified which require warnings.
One system that NowCastMIX currently uses as input is KONRAD. KONRAD is a method for automatic detection, tracking, and prediction of thunderstorm cells based on two-dimensional weather radar data. In recent years, a new scheme, KONRAD3D, has been developed, which provides three-dimensional objects of detected cells with the help of 3D radar volume scans. KONRAD3D also provides new state-of-the-art approaches to capture features such as the storm track more smoothly and consistently over time.
In order to let NowCastMIX benefit from this new development, KONRAD will be replaced by KONRAD3D. Subsequently, reanalyses with the new KONRAD3D module will be calculated for 194 convective days over a period of three convective seasons. A comparison of both data sets will show in which areas NowCastMIX benefits from the three-dimensional objects provided by KONRAD3D. The fuzzy logic of NowCastMIX will be tuned based on the reanalyses in order to optimize the probability of detection and false alarm ratio for the administrative districts of Germany.

How to cite: Debertshäuser, M., James, P., and Werner, M.: Integrating KONRAD3D into the nowcasting guidance system NowCastMIX at DWD, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-48, https://doi.org/10.5194/ecss2023-48, 2023.

Since July 2020 the DWD WarnWetter-App comprises the Crowdsourcing module “User Reports”. This module provides users the functionality to report observations about current weather conditions and severe weather to DWD and other users.

The user reports represent the current meteorological conditions at a certain place at a certain point of time. The Crowdsourcing module provides 10 different meteorological categories (lightning, wind, hail, rain, wet icy conditions, snowfall, snow cover, cloudiness, fog, tornado), each of which contains specific characteristic levels and optionally additional attributes. In addition, the user has the option of setting the location and time of the event manually.

The benefit of the data is that meteorological information at ground level is collected at places where no weather station is located in the immediate vicinity. The dataset is able to complement the existing synoptic station network. Forecasters from DWD already benefit from user-based observations that are available in near real-time.

In recent years, a new nowcasting algorithm has been developed at DWD, called KONRAD3D. The algorithm aims to automatically detect, track, and nowcast convective cells in order to support DWD’s warning management.

KONRAD3D uses three-dimensional radar reflectivity data as main input. In addition, also lightning data and information about hydrometeor types based on polarimetric radar data is regarded. In particular, in the latest version KONRAD3D features the new hail flag - a parameter that assesses a cell’s threat of hail. The new parameter rests upon the hydrometeor data and should roughly estimate the expectable near-ground hail size.

This is where the crowdsourcing data comes into play. Hail reports from app users are able to confirm expected hail sizes on the ground. Our analyses will show, whether the hail threat estimates were reasonable and at which point the user reports could complement the real-time operation of KONRAD3D.

How to cite: Spitzer, A., Kempf, H., Jerg, M., Werner, M., and Blahak, U.: DWD-Crowdsourcing: Are User Reports beneficial for Object-based Nowcasting?, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-54, https://doi.org/10.5194/ecss2023-54, 2023.

Summer thunderstorms can cause strong socio-economic impacts over Germany. The project SINFONY (Seamless integrated forecasting system) at German Weather Service has the goal to improve short-range predictions of these storms. Nowcasting (NWC) currently is superior to numerical weather prediction (NWP) on the very-short range up to about two hours in predicting convective cells while NWP performs better afterwards. Within SINFONY products are developed that integrate both approaches for a seamless prediction. High-resolution reflectivities from the German radar network are used as observational data base and for NWC initialization. The respective reflectivities from NWP models are derived by employing the radar forward operator EMVORADO. Convective cells are identified from these reflectivities using the KONRAD3D cell detection tool.

We present forecasts of convective cells from standalone NWC and NWP predictions and from a product that combines both systems. The recently developed NWC ensemble system KONRAD3D-EPS comprises 20 members with stochastic differences in the positions and life cycles of NWC objects. NWP objects come from ICON-D2-EPS (20+1 members) simulations employing a two-moment microphysics scheme with a forecast horizon of 8 hours. To combine these 41 members the NWP objects are clustered spatially, compared with each observation and the cluster closest to an observation is selected. Several properties of objects within a selected cluster are compared with the matching observed object using the Total Interest. Model objects that are similar enough to the observed object are selected and spatially shifted to make their centroid position equal to the observed cell. The trajectories of shifted NWP objects are then used, together with the NWC objects, as forecasts of convective cells. NWP objects that develop later in the forecast are considered as well but without the assignment to an observed object. Thus, with increasing lead time and the decease of cells that existed during the initialization of the combination, the product smoothly transitions into a purely model-based forecast.

Having an ensemble of predicted objects necessitates some kind of information reduction. Here the pseudomember method is employed that selects the locally most representative objects from the ensemble. For the object-based verification we use the Median of Maximum Interest. It reveals for JJA 2022 that the methods described above deliver a seamless object-based forecast product for convective cells which unifies the strengths of NWC and NWP. For lead times between 30 and 120 minutes the combined product performs even better than each prediction type standalone.

How to cite: Schnoor, I., Brechtel, A., Josipovic, L., Pante, G., Posada, R., Feige, K., Keller, J., and Blahak, U.: Predictions of convective cells from nowcasting, numerical forecasting and a combination of both, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-62, https://doi.org/10.5194/ecss2023-62, 2023.

 Recently, severe weather such as heavy rainfall, typhoon, hail, and lightning has become more frequent due to climate change. Especially heavy rains and localized precipitation systems accompanied by lightning strikes has not only human casualties but also social and economic impacts. To minimize the damage caused by lighting, research into the causes of lighting and their predictions are necessary.

 Korea Meteorological Administration(KMA) has been operating the Lightning Forecasting model to predict lightning forecast information on the Korean Peninsula. This model uses the energy density of lightning strikes detected by a total of 21 sensors of the LINET(LIghtning NETwork) systems(Nowcast, Germany). The energy density is calculated as the energy per unit area per unit time(kAkm^-2hr^-1) using CC(cloud-to-cloud) and CG(cloud-to-ground) lightning observation. This provides quantitative information on where strong lightning strikes are concentrated. In addition, the vector fields derived by applying the variational Echo Tracking(VET) algorithm are performed on the energy density and these are advected to the entire density field calculated by the semi-Lagrangian backward advection method. Finally, This very short-term quantitative lightning forecast data is produced every 10 minutes up to 6 hours ahead.

 The evaluation of lightning forecasting is performed from 2019 to 2021 based on each month and season which is a total of 192 cases, and regions for the Korean Peninsula. The verification methods range from simple traditional skill scores to methods for quantitative comparison between the prediction model and lightning frequency observation. Based on the results of the evaluation, the record score of the total critical success index(CSI) is 0.48 within 1 hour, autumn is the highest season(0.51), and winter season(January, CSI=0.17) is the lowest(0.17).

Acknowledgements:

This research was supported by the ”Development of Integrated radar analysis and customized radar technology (KMA2021-03021)” of “Development of integrated application technology for Korea weather radar” project funded by the Weather Radar Center, Korea Meteorological Administration.

How to cite: Son, M., Kim, H.-L., and Nam, K.-Y.: Development and Evaluation of Operational Lightning Forecasting System based on Lighting Detection Network, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-72, https://doi.org/10.5194/ecss2023-72, 2023.

As known, the presence of vertical wind shear leads to the organization of supercell thunderstorms. Changes in the wind profile can affect isolated convection especially in the development stage and in relation to its behavior.

This study takes a closer look at the organization of storms initiating in an urban environment that can have impacts on public life and infrastructure. Therefore an exemplary weather event over east Germany should be discussed in greater detail. While considering our theoretical and conceptual knowledge about supercelluar convection, it is important to ask further questions about the given atmospheric parameters or the specific time of convective initiation, when covering these events. Also it is essential to be able to make statements about the static stability and the sources of moisture. The local effects of a region, that may have influences on the convective scale, can help us to explain observed short-term changes. Especially the DBZ radar reflectivity, as one of the most common nowcast tools, can provide useful insights into atmospheric dynamics in context to the current flow pattern.

By evaluating NWP model data and using different observations, a summary of the synoptic pattern will be given that set the stage for isolated thunderstorm development over Berlin on this day. With using our understanding of dynamics in relation to the storm scale, the accuracy of nowcasting these cells should be clarified. For this purpose, the concept of the dynamic split and the storm motion as predictive indicators will be applied. Could the convective mode have been estimated already in the forecast process?

The case study can be considered as a useful complement for nowcasting silimar events in the future.

How to cite: Feige, H., Horn, C., and Gatzen, C.: A Case study - Splitting Supercells over Berlin on July 10, 2020, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-86, https://doi.org/10.5194/ecss2023-86, 2023.

An Additive Logistic Regression model for large hail (AR_hail) was developed using convective parameters from the ERA5 reanalysis, hail reports from the European Severe Weather Database (ESWD) and lightning observations from the Met Office Arrival Time Difference network (ATDnet). This model was shown to accurately reproduce the climatological distribution and the seasonal cycle of observed hail events in Europe. To explore the value of this approach to medium-range forecasting, a similar four-dimensional model was developed using predictor parameters retrieved from the ECMWF reforecasts: Mixed Layer CAPE, Deep Layer Shear, Mixed Layer Mixing Ratio and the Wet Bulb Zero Height. This model was applied to ECMWF reforecasts to compute probabilistic large hail forecasts for all available 11 ensemble members, from 2008 to 2019 and for lead times up to 228 hours. First, we compared the hail ensemble forecasts for different lead times with observed hail occurrence from the ESWD focusing on a recent hail outbreak. Secondly, we systematically evaluated the model’s predictive skill depending on the forecast lead time using the Area under the ROC Curve (AUC) as a validation score. This analysis showed that AR_hail has a very high predictive skill (AUC > 0.95) for forecasts up to 60 hours lead time. Although the performance scores progressively decrease with increasing lead time, AR_hail retains a high predictive skill even for extended forecasts (AUC = 0.86 at 180 hours lead time) showing that it can provide useful guidance in hail forecasting well in advance. Finally, the performance of the four-dimensional model was compared with that of composite parameters such as the Significant Hail Parameter (SHP) and the product of CAPE and Deep Layer Shear (CAPESHEAR). Results show that AR_hail outperforms CAPESHEAR (at all lead times) and SHP (especially at short lead times). This suggest that the developed Additive Logistic Regression model can improve hail forecasting compared to currently used composite indices in Europe.

How to cite: Battaglioli, F., Groenemeijer, P., Tsonevsky, I., and Púčik, T.: Forecasting Large Hail using Additive Logistic Regression Models and the ECMWF Reforecasts, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-103, https://doi.org/10.5194/ecss2023-103, 2023.

An updated version of JSMeteoView2 web-based application has been developed in the Czech Hydrometeorological Institute (CHMI). It has been used as a primary application for display of weather radar data in CHMI forecasting offices. It is used primarily for displaying of radar data which are available with horizontal and vertical resolution of 0.5 km. It also enables display and combination of other types of data such as data from synoptic stations, data from meteorological satellites, lightning data and requests for intervention of integrated rescue system. Users can use various geographical layers to precisely localize individual meteorological events.

New JSMeteoView2 layer displaying results of object based CELLTRACK algorithm is used specifically in convective situations. CELLTRACK identifies cores of radar reflectivity 44 dBZ and higher and tracks these identified reflectivity cores. It also displays various direct and derived characteristics of the identified cores based primarily on radar data but also on lightning data.

Among other reflectivity core characteristics various hail products were implemented. They are based on Waldvogel algorithm, severe hail algorithm described by Witt and also Vaisala HydroClass algorithm based on dual polarization weather radar data. Long-term evaluation of these hail products was carried out using records from European Severe Weather Database and also on continuous observations from CHMI professional stations, which enabled assessment of probability of detection but also assessment of false alarms occurrence.

The contribution will introduce the updated JSMeteoView2 application with main focus on description of CELLTRACK object layer and overview of implemented hail products. The contribution will also describe methods used for evaluation of hail products and results of this evaluation.

How to cite: Kyznarova, H. and Novak, P.: Overview of Recent Advancements in Convective Storm Nowcasting in the Czech Hydrometeorological Institute, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-119, https://doi.org/10.5194/ecss2023-119, 2023.

The development of additive logistic regression models (AR-CHaMo) for large hail, severe convective wind gusts, and F1 or stronger tornadoes for Europe and parts of North America allowed us to identify how the best predictors vary among different threats and different forecast domains. The best predictors were identified using the variance explained, based on the skill of logistic models for individual parameters as well as on investigating pairs of different parameters and their relation to hazard frequency.

For the models, we have chosen predictors that perform well over both domains and could thus be used to develop a global convective hazard model. In the case of large hail, CAPE was found to be a better predictor across Europe than across North America, where mid-tropospheric lapse rates discriminate better between environments with and without large hail. We found that CAPE below the -10 °C level was a skillful predictor in both domains. For severe convective wind gusts, it was found that they occurred with lower CAPE and lower amounts of absolute moisture in Europe than in North America. Height of the LCL or a parameter that predicts the cold pool strength worked better in Europe than in North America. Strong mean wind in the bottom troposphere was found among the best predictors of severe wind gusts in both domains. Regional differences among the best predictors were also found for F1 and stronger tornadoes, even though the amount of SRH in the lower troposphere is universally a skillful predictor.

We applied models using the best predictors of large hail across North America and Europe to the ERA-5 reanalysis to obtain a global model of large hail hazard. Then, we compare the model to existing hail climatologies worldwide and discuss its limitations and potential improvements.

How to cite: Battaglioli, F., Pucik, T., Groenemeijer, P., and Taszarek, M.: Identifying predictors of large hail, severe convective wind gusts, and tornadoes across Europe and North America: towards the development of global convective hazard models., 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-128, https://doi.org/10.5194/ecss2023-128, 2023.

Exceptionally heavy rainfall are one of the most difficult and challenging problems for forecasters. The synoptic situation at the beginning of September 2022, which led to significant precipitation in the Karlovo valley, was analysed. 24-hour rainfall amount of more than 200 mm were recorded, which is an event exceeding the climatic precipitation characteristics established for the region, including the rare events. The synoptic situation was based on a cyclogenesis in the Gulf of Genoa, and cyclone which passed through the southern regions of the Balkans towards the Black Sea. The precipitation is related to the formed convergent zone over Bulgaria and to the divergence on the 200/300 hPa level. Comparison with the monthly rainfall amount in the area of Karlovo as a climate showed almost 4 times exceeded values for 24-hours. The environmental conditions for development of heavy rainfall over Bulgaria was investigated using proximity sounding obtained by the numerical model GFS. Several instability indices (CAPE, Lifted index, K index, Total-Total index and etc.) and vertical wind shears between various layers are analysed in order to verify their ability to forecast rainfall. Their skill to predict severe storm types is evaluated by calculation of probability of detection (POD) and false alarm ratio (FAR). The data at the surface necessary for the calculations of the instability indices are taken from synoptic stations of National Institute of Meteorology and Hydrology (NIMH) located in centralen Bulgaria.

How to cite: Markova, B., Kleshtanova, V., and Kirilova, A.: Exceptionally heavy rainfall in Karlovo valley, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-129, https://doi.org/10.5194/ecss2023-129, 2023.

On August 18, 2022, Europe experienced the most severe storm of the year. In the early morning, a derecho formed, causing significant damage along a path of over 1000 km as it travelled at an exceptional speed of up to 40 m/s, with wind gusts reaching a maximum of 62.2 m/s. The storm had a widespread impact, affecting Spain, France, Italy, Slovenia, Austria, and the Czech Republic. The severe weather caused fatalities in several countries, with the highest death tolls reported on Corsica and in the Alps. The focus of this analysis is to understand the factors that led to the derecho's exceptional intensity. We examine the storm's chronological development, and the intensification phases it underwent. We also investigate how external and internal forcing contributed to the rapid upscale growth of the storm.

From the perspective of external factors, we show the importance of a mid-level front and warm air advection at the lower levels and their relation to the distribution of CIN in the vicinity of the event. We show that the development and movement of convective cells were influenced by the mid-level front. The system was also strongly influenced by the internal forcing as it formed in an environment of high CAPE values and strong low-level vertical wind shear, and intense storm-relative inflow. The combination of internal and external forcing factors eventually resulted in the extreme forward speed of the derecho. Due to the interaction with complex topography, the derecho underwent several weakening and strengthening phases, which we associate with the changes to the internal and external forcing. Finally, we compare the environment of the derecho to typical environments of severe convective wind gusts in Europe and compare the event to similar Mediterranean derecho cases from the past.

How to cite: Kreitz, M., Gatzen, C., and Púčik, T.: Analysis of the high-end derecho in Corsica in 2022, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-148, https://doi.org/10.5194/ecss2023-148, 2023.

After several years of development and evaluation, KONRAD3D, the new tool for detection, tracking, and nowcasting of convective cells at Deutscher Wetterdienst (DWD), reaches operational status. The scheme is intended to further improve the performance of DWD’s automated warning decision support system and to equip external customers with improved information on thunderstorm development and severe weather risks.

KONRAD3D essentially relies on 3D radar reflectivity data, but also regards lightning data as well as hydrometeor information from DWD-products based on dual-polarization radar algorithms to assess hail risk. It is planned to integrate further data sources and products, e.g., rotation information derived from radial velocity measurements and mesocyclone detection outputs.

During the past three convective seasons, KONRAD3D has been intensively evaluated by forecasters at DWD as well as in the ESSL (European Severe Storms Laboratory) Testbed with many valuable comments and suggestions on how to modify, improve or better tune the method. Inspired by these findings, various extensions have been implemented, sub-algorithms have been revised, and tuning parameters have been adapted.

This work gives an overview on the status of KONRAD3D’s design and setup at operational kick-off. We present the current status of its visualization in DWD’s meteorological workstation NinJo and illustrate how KONRAD3D is integrated in follow-up applications and end customer interfaces.

How to cite: Werner, M., Feger, R., Josipovic, L., and Böhme, T.: Operational usage of KONRAD3D, DWD’s scheme for detection, tracking, and nowcasting of convective cells, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-150, https://doi.org/10.5194/ecss2023-150, 2023.

The European Storm Forecast Experiment (ESTOFEX) is a well known project of volunteer meteorologists making forecasts of thunderstorms and their severe weather threats for Europe for 20 years (since October 2002). Originally, the forecasts consisted of three severe weather risk lines as well as one thunderstorm line. The verification of the dichotomic lightning forecasting skill was presented at the 4^th ECSS in Trieste, 2007.

In 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. A verification of these lightning probability forecasts over a 4-year period was presented at the 7^th ECSS in Helsinki in 2013 (doi: 10.13140/RG.2.1.1026.1845).

We update the verification of lightning probability areas to include 8 more years, covering 2009-2021 (incl.), and evaluate consistency of each forecaster by histograms and maps of thunderstorm occurrence (observed mean frequency for a location) for each category of forecast. Other aspects of this work will be presented in Part 2 (poster by Mateusz Taszarek). Gridded lightning data from ATDNET is used, as well as severe weather reports from the ESSL European Severe Weather Database. The ESTOFEX forecast category definitions are 0-15%, 15-50%, >50% forecasts for thunder, 0-5% (Level 1) and 5-15% (Level 2) for severe weather, and >15% (Level 3) for extremely severe weather, respectively.

As in the previous work, two key questions are answered:

(1) “Of all the times a location was included in a certain forecast category (e.g. 15-50% area), how often was the phenomenon observed?"

This returns a percentage for each location on the map. These locations can be plotted on the map or grouped into 3 overlapping histograms, one for each forecast category. The overlap and calibration can be judged for each forecaster and by geographic regions. The spatial criterion (e.g. phenomenon within 40 km from a point) can be varied.

(2) “If a spatial thunderstorm density was observed, which forecast probability area was it included in?"

For each category of observed spatial density around a point (or a Practically Perfect Hindcast, see Part 2), a map and histogram can be plotted of the corresponding mean forecast category frequency (or difference), to find areas where forecasters most frequently under- or overestimate activity, for example by season.

How to cite: van der Velde, O. and the ESTOFEX Team: Evaluation of ESTOFEX convective outlooks from 2007 to 2021. Part 1: forecasters and regional performance of lightning predictions, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-158, https://doi.org/10.5194/ecss2023-158, 2023.

Optical flow based nowcasting is a powerful technique to compute forecasts for small lead times up to a couple of hours. Forecasts of radar data are very useful as standalone products and serve as input data for multiple operational products, e.g., for the seamless combination of radar and NWP data and for cell-based analysis and prediction products. However, the optical flow technique has several drawbacks. It assumes stationarity in both the data values as well as the advection information. Further, it is a deterministic technique and the nowcasts have no dynamic properties. Recently, machine learning techniques have shown promising results for producing nowcasts with dynamic properties. However, for radar reflectivity and precipitation data, the predictions often lack high-intensity values and tend to become blurry for larger lead times.

In the current work we explore the potential of deterministic convolutional neural networks (CNN) to improve the operational optical flow nowcasting at DWD. A two-year dataset consisting of radar, NWP and orography data is used for training modified UNet based neural networks. Each network predicts radar reflectivity composites for a specific lead time, in 5-minute steps. Several optimization techniques are combined, both for the input data and the network architecture. The input data contains optical flow based nowcasts of previous radar timesteps that are mapped to the target lead time. The NWP input parameters are chosen for their known importance in convective processes. To understand their impact in this application, an ablation study is performed. The network architecture is optimized. The classic UNet architecture is augmented with additional horizontal computation blocks. This adds more nonlinearity to the finer scales of the network and reduces the validation error. Individual encoders are used for the radar and NWP data and combined with affine linear transformations. Experiments with classical pointwise loss functions as well as losses with spatial context (e.g., FSS) are conducted. The new forecasts are compared with the operational nowcasting at DWD as well as a closed-loop NN approach (RainNet).

How to cite: Friedrich, U.: Analysis and application of CNN to improve deterministic optical flow nowcasting at DWD, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-166, https://doi.org/10.5194/ecss2023-166, 2023.

The European Storm Forecast Experiment (ESTOFEX) is a team of volunteer forecasters that have been providing experimental convective outlooks for Europe since 2002. Probabilistic storm forecasts issued by ESTOFEX address threats posed by severe convective storms, i.e. lightning, large hail, severe wind gusts, tornadoes and excessive precipitation. ESTOFEX also serves as a platform for exchange of knowledge about forecasting severe convective storms with a goal of improving their understanding among both members of ESTOFEX and others. While not official, ESTOFEX products have been widely used by national meteorological services, severe storm communities and the public. ESTOFEX forecasters have regularly contributed to the ESSL Testbeds and are using an ingredients based forecasting methodology to forecast severe storms. Consistently improving severe storm reporting in the European Severe Weather Database (ESWD) and availability of ground-based lightning detection measurements over the last decade enabled the verification of a large number of ESTOFEX forecasts. Thus, in this work we evaluate 4019 convective outlooks issued by ESTOFEX forecasters since 2007. Our goals are to detect spatiotemporal patterns in convective outlooks and test the reliability of issued threat level polygons, i.e. for a low and high probability of lightning, and an increasing probabilities of severe weather: level 1, level 2 and level 3. We performed the verification by applying a number of methods, including contingency table statistics, receiver operating characteristic curves, practically perfect hindcasts and by calculating spatial coverage of detected lightning (ATDnet network) and local storm reports (ESWD) within issued polygons. Results indicate that products issued by ESTOFEX over the last 15 years, when combined together, are consistent with convective climatologies based on reanalyses and lightning detection data. However, we note that forecasters tend to issue outlooks relatively more often for severe weather outbreaks across western and central Europe. We found that while 95% of the issued lightning probability areas fulfilled the required criterion of coverage, this was only true for 40% of the severe weather probability areas. One reason is that while lightning observations are relatively homogeneous across the forecast domain, the same cannot be said about severe weather observations. These are lacking in regions such as southeastern or eastern Europe, while forecasters calibrated themselves to the higher observed coverage in western and central Europe. The reliability of ESTOFEX forecasts increased over the time, but we found underestimation of lightning probabilities over southern Europe and an overestimation of lightning probabilities over British Isles and Scandinavia.

How to cite: Taszarek, M., Groenemeijer, P., Pucik, T., van der Velde, O., and Dafis, S.: Evaluation of ESTOFEX convective outlooks from 2007 to 2021. Part 2: climatology and reliability of threat level polygons, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-167, https://doi.org/10.5194/ecss2023-167, 2023.

The European Severe Storms Laboratory (ESSL) has developed logistic models (AR-CHaMo) for the occurrence of lightning and (very) large hail based on the ERA5 reanalysis, with the primary purpose of investigating long-term chance in severe weather occurrence. These models can, however, also be used in a forecasting context. 

We have set up a routine that uses AR-CHaMo models to post-process the output of three different NWP models, the ECMWF IFS HRES, DWD's ICON-EU, and NCEP's GFS to obtain probabilities of hail and lightning occurrence. The results were evaluated as a part of the ESSL Testbeds in 2022.

Forecast maps showing the calculated probabilities for the current and following day are calculated from the average of the three models and are made available on a website. The website interface allows overlaying reports from the European Severe Weather Database, which give an impression of the quality of the forecasts and help to identify existing weaknesses. Work is underway to include ensemble forecasts, and quantitative real-time verification, and to extend the forecast horizon as part of projects supported by the Austrian Science Fund and ECMWF. We show several AR-CHaMo lightning and large hail forecasts that resulted in no hail, large hail, and very large hail and discuss their performance.  

How to cite: Groenemeijer, P., Battaglioli, F., and Púčik, T.: Stormforecast.eu: real-time automated forecasts for hail and lightning based on post-processed NWP, 11th European Conference on Severe Storms, Bucharest, Romania, 8–12 May 2023, ECSS2023-169, https://doi.org/10.5194/ecss2023-169, 2023.