The most severe hailstorm in recent decades in Germany was the Munich hailstorm on July 12, 1984 with an insured loss of 1.5 billion DM. The second most expensive event in the last few decades was the Reutlingen hailstorm on 28 July 2013 in Baden-Württemberg, which caused insured losses of more than 600 million Euros at residential buildings in the densely-populated area on the western flank of the Swabian Jura.
Within the frame of the research project "HARIS" (HAil RISk), past and future hail hazard and risk for Germany is being assessed at KIT using a combination of various meteorological data sets and loss data provided by different insurance companies.
As all weather phenomena associated with deep moist convection (thunderstorms with a vertical extent to the troposphere, approx. 12 km), the regions affected by hail, referred to as hail streaks, usually exhibit a very limited spatial extent. According to a study by Changnon (1970), 80 per cent of all recorded loss-related hailstorms in the US have an extent of less than 40 km2. Due to the small affected area and a lack of suitable measurement systems, hail events are not recorded comprehensively over a long period, thus making it impossible to provide meaningful information concerning their spatial distribution and trends. Due to the inadequate recording of hail it has hitherto been difficult – particularly in Germany – to draw any conclusions regarding the hail hazard.
At KIT, hail streaks are reconstructed over a long-term period using a multicriterial approach that includes different data sets, especially radar data. Furthermore, the researchers are using indirect climate data (proxies) for long term analyses of the atmospheric conditions that are necessary for the formation of thunderstorms and hail.
KIT scientists identify hail streaks from radar data
Hailstorms cannot (yet) be reliably and realistically simulated by high-resolution weather forecast or climate models. For this reason, weather radar devices with a very high spatial (up to 0.5 km) and temporal (up to 5 min) resolution are best suited to identify hailstorms. However, the intensity measured by the radar (radar reflectivity) depends on various factors such as the size distribution of the precipitation particles or the distance of the radar beam to the ground. A clear relationship between radar reflectivity and hail intensity on the ground on the basis of these data alone can therefore not be established, but only estimated.
As part of the "HARIS" project at KIT, radar data of the German Weather Service (Deutscher Wetterdienst, DWD) are evaluated with respect to the vertical extent of high reflectivities, which serves as proxy for severe convection in combination with lightning and insurance loss data. Applied over several years, the analysis estimates the radar-derived hail hazard for Germany in a very high spatial resolution of 1 × 1 km² (Puskeiler, 2013). The reconstructed hail streaks and applied extreme value statistics confirm the considerable spatial variability of hail probability and intensity. For Baden-Württemberg, for example, a pronounced maximum was identified in the region south of Stuttgart (Kunz and Puskeiler, 2010).
Similar patterns can also be found in other regions of Germany, as is also confirmed by analyses of the lightning density. On the large-scale, the radar-derived hail probability increases from north to south, as the conditions for the occurrence of thunderstorm events are more favourable in the south due to higher available convective energy. Furthermore, some pronounced maxima are apparent, which are mostly located at the lee side (the side facing away from the wind) of the low mountain ranges – e.g., the lee of the Black Forest or the Brocken. Some of these patterns can be ascribed to the typical flow conditions prevailing on hail days, where the flow goes partly around the mountains rather than over, creating low-level flow convergence in the lee (frequent southwesterly weather conditions; Kunz and Puskeiler, 2010; Kapsch et al., 2012; Puskeiler, 2013).
Pilot project: first automatic measuring network for hail
In a pilot project, ten measuring stations of the Landesanstalt für Umwelt, Messungen und Naturschutz Baden-Württemberg (State Office for the Environment, Measurements and Nature Con¬servation – LUBW) were equipped by KIT with newly developed hail sensors (developed at htwsaar; Löffler-Mang et al., 2011) in 2013. The measuring device is able to register kinetic energy and momentum of hailstones by measuring sound waves with small piezomicrophones in a weather and UV-resistant Makrolon case. From this, hail sizes and the size spectrum can be inferred. The focus of the monitoring lies on areas with a high hail probability, for example on the plains of Filderstadt south of Stuttgart or in Villingen-Schwenningen. The aim is to observe the size distribution of the hailstones during a hailstorm and to derive new knowledge of the damaging effect of hail in combination with other meteorological measurements.
Hail probability derived from regional climate models
Over recent decades, damage caused by severe hailstorms in Central Europe and other regions has consid¬erably increased (Schiesser, 2003; Cao 2008; Kunz et al., 2009). The question arises to what extent this increase in hail damage in the past is, along with changes in vulnerability, determined by the anthropogenic climate change and what changes are to be expected in the future.
Within the project HARIS-CC ("Hail Risk and Climate Change") at KIT various proxies (indi¬rect climate data) for hail were examined with respect to their long-term variability. Appropriate proxies for hail are convective parameters and indices, which quantify static stability of the atmosphere. In contrast to direct thunderstorm or hail observations, proxy data are uniformly available for long periods (> 30 years), which enables their statistical analysis.. However, these proxy data only reflect the potential of the atmosphere for deep moist convection. Direct statements con¬cerning the actual occurrence or frequency of thunderstorms or hail cannot be derived.
Long-term statistical analyses of convective parameters derived from radiosoundings show that those considering surface-based tempera¬ture and moisture values exhibit statistically significant positive trends over the last three decades (Mohr and Kunz, 2013a). Accordingly, the probability of thunderstorms and hail over large parts of Europe has increased. To quantify changes in the thunderstorm or hail potential in the past and the near future in Germany, simulations of various high-resolution regional climate models (RCM) were evaluated. Evaluations with observations confirm that RCMs with a spatial resolution of about 10 km are basically capable to reflecting the convective conditions of the atmosphere. RCMs driven by reanalysis (ERA-40) show an increase in the thunderstorm potential between 1971 and 2000, mainly due to an increase in moisture in the lower troposphere. On the other hand, atmospheric stability shows no uniform results.
A similar picture emerges for general weather patterns determined from both reanalyses and various realisations of RCMs. In accordance with the objective weather type clas¬sification of DWD, four of the total of 40 large-scale weather types coincide with an increased probability of damage-relevant hail events. These four weather situations show a statistically sig¬nificant increase in the period from 1971 to 2000. For the future (2010 to 2050) the model data show only a slight increase in these hail-realted weather types. The strengths of the changes, however, are strongly controlled by the period considered (Kapsch et al., 2012).
Development and application of a logistic hail model
To take a sound combination of different meteorological parameters that are relevant for the occurrence of hail into account, a logistic hail model (multivariate analysis method) was developed at KIT. This approach aims at improving the diagnostics for hail events (Mohr and Kunz, 2015). The result of the model is a new index which describes the potential of the atmosphere for the formation of hail. It is therefore referred to as Potential Hail Index (PHI). Applying to a mini-ensemble of seven RCM simulations, it becomes clear that the potential for hail in the future will only increase slightly (statistically sig¬nificantly) in northwestern and southern Germany.
Changnon, S. A., 1970: Hailstreaks. J. Atmos. Sci., 27, 109 – 125.
Kapsch, M. L., M. Kunz, R. Vitolo, und T. Economou, 2012: Long-term variability of hail-related weather types in an ensemble of regional climate models. J. Geophys. Res., 117, D15107. DOI:10.1029/2011JD017185
Kunz, M., J. Sander, und C. Kottmeier, 2009: Recent trends of thunderstorm and hailstorm frequency and their relation to atmospheric characteristics in southwest Germany. Int. J. Climatol., 29, 2283 – 2297.
Kunz, M. und M. Puskeiler, 2010: High-resolution assessment of the hail hazard over complex terrain from radar and insurance data. Meteor. Z., 19, 427 – 439. DOI 10.1127/0941-2948/2010/0452
Löffler-Mang, M., D. Schön, und M. Landry, 2011: Characteristics of a new automatic hail recorder. Atmos. Res., 100, 439–446.
Mohr, S., 2013: Änderung des Gewitter- und Hagelpotentials im Klimawandel. Wiss. Berichte d. Instituts für Meteorologie und Klimaforschung des Karlsruher Instituts für Technologie, Band 58, Karlsruhe, Germany.
Mohr, S. und M. Kunz, 2013: Recent trends and variabilities of convective parameters relevant for hail events in Germany and Europe. Atmos. Res., 123, 211–228. DOI: http://dx.doi.org/10.1016/j.atmosres.2012.05.016
Mohr, S., Kunz, M. und Keuler, K. (2015): Development and Application of a Logistic Model to Estimate the Past and Future Hail Potential in Germany. J. Geophys. Res. submitted.
Mohr, S. und Kunz, M., 2014: Changes in the Hail Potential Over Past and Future Decades: Using a Logistic Hail Model. J. Geophys. Res. (Eingereicht).
Pruppacher, H. R. und J. D. Klett, 1997: Microphysics of clouds and precipitation, Vol. 18. Kluwer Academic Publishers, Dordrecht, Niederlande.
Puskeiler, M., 2009: Analyse der Hagelgefährdung durch Kombination von Radardaten und Schadendaten für Süddwestdeutschland. Diplomarbeit am Institut für Meteorologie und Klimaforschung (IMK), Universität Karlsruhe (TH), Karlsruhe, Deutschland.
Puskeiler, M., 2013: Radarbasierte Analyse der Hagelgefährdung in Deutschland, Dissertation am Institut für Meteorologie und Klimaforschung (IMK-TRO), Karlsruher Institut für Technologie (KIT), Karlsruhe, Germany.
Text and data in cooperation with CEDIM, an interdisciplinary research institution by the Karlsruhe Institute of Technology.