th 5 European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

ECSS 2009 Abstracts by session

ECSS 2009 - 5th European Conference on Severe Storms 12-16 October 2009 - Landshut – GERMANY List of the abstract accepted for presentation at the conference: O – Oral presentation P – Poster presentation

Session 04: Climate change impacts on sever storms, development of adaptation concepts

Page Type Abstract Title Author(s)

Significant Increases in Frequencies and Intensities of 93 O Weather Related Catastrophes – what is the Role of P. Höppe Climate Change? Extreme Precipitation: Current Forecast Ability and A. Champion, K. Hodges, L. 95 O Climate Change Bengtsson High-resolution modeling of the effects of anthropogenic J. Trapp, E. D. Robinson, M. E. O climate change on severe convective storms Baldwin, N. S. Diffenbaugh K.Riemann-Campe, R. Blender, 97 O Future global distributions of CAPE and CIN N. Dotzek, K. Fraedrich, F. Lunkeit Severe hail frequency over Ontario, Canada: recent trend O Z. Cao and variability 99 P Hailpad data analysis for continental part of D. Pocakal RegioExAKT - Regional Risk of Convective Extreme N. Dotzek, the RegioExAKT 101 P Weather Events: User-oriented Concepts for Trend consortium Assessment and Adaptation Climate change impacts on severe convective storms over 103 P J. Sander, N. Dotzek Europe Wind loads and climate change – significance of gust fronts P M. Kasperski, E. Agu, N. Aylanc in the structural design Extreme weather events in southern Germany – A. Matthies, G. C. Leckebusch, T. 105 P Climatological risk and development of a large-scale Schartner, J. Sander, P. Névir, U. identification procedure Ulbrich Study of weather change due to loss of Sunderban Delta P. Chatterjee, U. K. De, D. P Region Pradhan Influence of sounding derived parameters on the strength of 107 P S. Grünwald, H. E. Brooks tornadoes in Europe and the USA from reanalysis data

91

92 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

SIGNIFICANT INCREASES IN FREQUENCIES AND INTENSITIES OF WEATHER RELATED CATASTROPHES – WHAT IS THE ROLE OF CLIMATE CHANGE? Peter Hoeppe

Munich Re, 80791 Munich, GERMANY, [email protected]

As extreme weather events affect the core business exposure levels and the fact that modern technologies are of insurance this industry has quite early addressed potential more vulnerable to losses. The state of Florida in the USA, effects of global warming on natural catastrophe hazards. which has always had a high hurricane exposure, is a good Today climate change is regarded as one of the largest risks illustration of the way that socioeconomic factors can act as for insurance industry. Munich Re's experts have been natural catastrophe loss drivers. The population there has researching loss events caused by natural hazards around the grown from three million in 1950 to the current 19 million. globe for over 35 years. These losses are documented in the NatCatSERVICE database currently documenting more than As the rise in the number of natural catastrophes is 26,000 single events. largely attributable to weather-related events like windstorms and floods (figure 1), with no similarly strong In recent years we have seen many natural increase in geophysical events such as earthquakes, catastrophes with records in intensities and losses caused by tsunamis, and volcanic eruptions, there is some justification them such as: in assuming that anthropogenic changes in the atmosphere, and climate change in particular, play a decisive role. There  The hundred-year flood in the Elbe region in has been more and more evidence to support this hypothesis Germany in the summer of 2002 in recent years. The fourth status report of the  The 450-year event of the hot summer of 2003, Intergovernmental Panel on Climate Change (IPCC 2007) which caused more than 70,000 heat fatalities in regards the link between global warming and the greater Europe frequency and intensity of extreme weather events as  The largest ever recorded number of tropical significant. The report finds, with more than 66% (28) and hurricanes (15) in a single North probability, e.g. that climate change already produces more Atlantic season in 2005, with the strongest (Wilma – heat waves, heavy precipitation, drought and intense tropical core pressure: storms and that such effects will be growing in the future. 882 hPa), fourth strongest (Rita), and sixth strongest (Katrina) hurricanes on record.  Hurricane Katrina, the costliest single event of all times, with economic losses of over US$ 125bn and insured losses of approximately US$ 60bn;  In October 2005, Hurricane Vince formed close to Madeira, subsequently reaching the northernmost and easternmost point of any tropical .  Winter storm Kyrill (January 2007) has caused the second largest losses in Europe caused by a winter storm  Largest losses ever caused by flooding in the UK in June/July, 2007.

The analyses of the NatCatSERVICE data clearly show a dramatic increase in the number of natural FIG. 1: Annual numbers and trend lines of loss relevant natural catastrophes around the globe, with ever growing losses. The events broken down to the different perils (Source: NatCatSERVICE, Munich Re) trend curve indicating the number of great natural catastrophes worldwide (involving thousands of fatalities, billion-dollar losses) reveals an increase from about three The rise in global average temperatures significantly per year at the beginning of the 1950s to around eight at the increases the probability of record temperatures. Higher present time. temperatures also enable air to hold more water vapour, thus increasing the precipitation potential. Combined with more Economic and insured losses resulting from great pronounced convection processes, in which warm air rises to weather disasters have risen sharply. In 2005, a record year, form clouds, this results in more frequent and more extreme economic losses were as high as nearly US$ 180bn and intense precipitation events. Already today such events are insured losses around US$ 90bn. responsible for a large proportion of flood losses.

The main reasons for the sharp increase in losses from weather-related catastrophes are population growth, the settlement and industrialisation of regions with high

93 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

Munich Re’s Geo Risks Research has undertaken hurricane frequency analyses for the events of the past decades, which also take into account natural climate cycles (the Atlantic Multidecadal Oscillation, AMO). The results indicate, that the higher frequency and intensity of Atlantic tropical cyclones in recent years could be due both to the natural cycle (the current warm phase, which started in 1995) and global warming.

Now that a number of changes have already happened and some of the predictions for the coming decades have already been seen, the key issue is no longer if and when there will be conclusive proof of anthropogenic climate change. The crux of the matter is whether the existing climate data and climate models can provide sufficient pointers for us to estimate future changes with reasonable accuracy and formulate adaptation and prevention strategies in good time.

The insurance industry’s natural catastrophe risk models have already been adjusted in the light of the latest findings. For instance, they now incorporate the increased hurricane hazard due to higher sea surface temperatures that will remain above the long-term average due to the ongoing cyclical warm phase in the North Atlantic and the continuous warming caused by anthropogenic climate change.

94 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

EXTREME PRECIPITATION: CURRENT FORECAST ABILITY AND CLIMATE CHANGE Adrian J. Champion, Kevin I. Hodges, Lennart O. Bengtsson

NCEO National Centre for Earth Observation, ESSC, The University of Reading, Harry Pitt Building, 3 Earley Gate, Whiteknights, PO Box 328, Reading, RG6 6AL, UK, [email protected] (Dated: 15 September 2009)

I. INTRODUCTION Figure 2 shows the radar imagery at the same time, Extreme precipitation events are a major cause of which has a resolution of 1km, from the Met Office flooding events, such as the 2007 summer floods in the UK NIMROD radar network. Figure 1 shows that the LAM and other flooding events in Western Europe. Such events picks up the large scale seen in the radar imagery, but does are difficult to forecast due to the convective scale processes not pick up the small scale, or the intensity seen in the radar embedded within the synoptic scale. imagery. The effect of climate change of such events also needs to be addressed. In a warmer climate, the frequency and intensity of extreme events are projected to increase. If similar synoptic conditions as seen in 2007 over the UK are experienced in a future climate, how extreme will the precipitation be, and will such conditions become more frequent? This research uses a Limited Area Model (LAM) of the UK Met Office’s Unified Model at a 12km resolution, to investigate the current forecast ability of extreme precipitation. The ECHAM5 Global Climate Model (GCM) at high resolution (T213/T319) is also used to investigate the change in Atlantic storm tracks in a warmer climate.

II. CURRENT FORECAST ABILITIES

The UK floods experienced during the summer of 2007 were caused by extreme rainfall from local convective storms. These storms were persistent over the UK due to the presence of a stationary cut-off low that remained over the UK for several days, bringing a constant moisture supply to the convective storms (Blackburn, 2008). The small scale of the convective systems, embedded FIG. 2: UK Met Office C-band rainfall radar data, for use with the within the synoptic scale system, meant it was very difficult NIMROD automated weather analysis and nowcasting system, for for the forecast model, which at the time had a resolution of the 20th July 2007 at 1400. 12km, to pick up the extreme rainfall. Figure 1 shows the precipitation output for a LAM re-run for the summer 2007 The 12km resolution of the LAM does not have a UK floods using ECMWF boundary conditions. high enough resolution to model the convective systems. This would explain why during the 2007 floods, the model under-predicted the amount of rainfall by up to 45% of the observed rainfall (Stuart-Menteeth, 2007). However, as noted by Nieto et al., 2005, the precipitation distribution associated with cut-off lows is difficult to predict as they bring moderate to heavy rainfall over large areas. Therefore an increase in resolution does not necessarily mean an improved forecast for such conditions.

III. GLOBAL CLIMATE MODELS

The ECHAM5 T319 GCM is used to investigate whether the intensity and frequency of Atlantic storms change in a warmer climate. This has been investigated by Bengtsson et al. (2009) at the T213 resolution; however a

higher resolution model is expected to predict more extreme FIG. 1: Total Precipitation Rate output from a 12km v6.1 LAM of events. the UK Met Office’s UM at 1400 hours for the 20th July with Storms are tracked using the TRACK software ECMWF boundary conditions. developed by Hodges (1995) which tracks storms based on

95 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY their 850hPa relative vorticity. The tracking was performed larger scale storm. for two periods: a 20th Century (20C) climate, 1980-2000, It is also clear that the number of extreme and a 21st Century (21C) climate, 2080-2100, using the precipitation events is expected to increase in a warmer IPCC A1B scenario for the 21C. climate, and also for the intensity of these events to increase. Figure 3 shows the maximum precipitation along the It is planned to track cut-off lows in the ECHAM5 data, tracks determined as the area averaged precipitation within 5 using the TRACK software. By selecting a case study degree of the storm centre, for DJF and JJA, for both the identified in this data, and downscaling it to use as boundary T319 and T213 resolutions. The insets show the tails of the conditions for a LAM, the effect of the increase in intensity distributions scaled to 90 months. can be investigated.

V. REFERENCES

Bengtsson, L., Hodges, K., Keenlyside, N., 2009: Will Extratropical Storms Intensify in a Warmer Climate? J. Climate, 22 2276-2301. Blackburn, M., Methven, J., Roberts, N., 2008: Large-scale context for the UK floods in summer 2007. Weather, 63 (9) 280-288. Hodges, K., 1995: Feature Tracking on the Unit Sphere. Mon. Weather Rev., V123 3458-3465. Nieto, R., Gimeno, L., Torre, L., Ribera, P., Gallego, D., 2005: Climatological Features of Cutoff Low Systems in the Northern Hemisphere. J. Climate, 18 3085-3103. Stuart-Menteeth, A., 2007: U.K. Summer 2007 Floods, Technical Report, Risk Management Solutions.

FIG. 3: Comparison of T213 (dashed) and T319 (solid) intensity distributions for area (5 degrees) averaged total precipitation rate (mm hr-1), bin width is 0.5 mm hr-1, for 20C (black) and 21C (red) for the Northern Hemisphere for DJF (top) and JJA (bottom).

Figure 3 shows that in the T319 resolution, in comparison to the T213 resolution, there is an increase in the number of extreme events. This increase is a lot smaller for DJF than JJA. Figure 3 also shows an increase in the number of extreme events in 21C, in comparison to 20C, for both DJF and JJA. This appears larger in T319.

IV. CONCLUSIONS

The resolution of weather forecast models has increased in recent years, from 12km to 1.5km (UK Met Office). The benefit of this increase in resolution on cut-off low events, such as the summer 2007 UK floods, needs to be investigated, to determine whether the higher resolutions are able to forecast convective storms when embedded within a

96 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

FUTURE GLOBAL DISTRIBUTIONS OF CAPE AND CIN K. Riemann-Campe1,2, R. Blender2, N. Dotzek3, K. Fraedrich2, F. Lunkeit2

1International Max Planck Research School on Earth System Modelling (IMPRS-ESM), Bundesstraße 53, 21046 Hamburg, Germany, [email protected] 2Meteorologisches Institut der Universität Hamburg, Grindelberg 5, 20144 Hamburg, Germany 3Deutsches Zentrum für Luft und Raumfahrt (DLR) - Institut Physik der Atmosphäre, Oberpfaffenhofen, Münchner Straße 20, 82234 Wessling, Germany (15 September 2009)

I. INTRODUCTION Convective available potential energy (CAPE) is used to categorise and forecast convective storms. Convection might be inhibited by positive values of convective inhibition (CIN) which defines the energy needed to reach the CAPE layer. Therefore, CIN indicates the probability of convection occurring, while CAPE determines the intensity of convection (Colby, 1984). During the last decades CAPE and CIN reveal trends in ERA-40 re-analysis data (Riemann-Campe et al., 2009) which are reproduced by simulations with the coupled atmosphere-ocean general circulation model ECHAM5/MPIOM. Here, the impact of increasing greenhouse gases during the upcoming decades is assessed by the analysis of global distributions of CAPE (100 hPa FIG. 1: Difference of JJA mean of 95th-percentile of CAPE in J/kg mixed layer, pseudo-adiabatically) and CIN computed in the between scenario A1B (2071-2100) and scenario 20C (1901-2000). scenarios A1B and B1 for the years 2001 until 2100. Positive numbers indicate an increase of CAPE with time.

differences change with season. The strongest positive II. DATA & METHODOLOGY differences in CAPE occur over the Central USA, the centre Six-hourly values of the coupled atmosphere-ocean model of Africa, and in Asia south of the Tibetan Plateau during ECHAM5/MPIOM (Roeckner et al., 2003) are used in the June, July and August (JJA). The strongest negative spectral truncation T63 (horizontal resolution ~ 1.875) to difference in CAPE occurs over the southern hemispheric compute CAPE and CIN for 1900-2100. The global eastern tropical Pacific during December, January and distributions of these variables in the future climate February (DJF). scenarios B1 and A1B (2001-2100) are compared with those In comparison with the differences in CAPE are of the present-day climate (simulation 20C, 1900-2000). those in CIN also present over all continents and over all Trends are estimated by subtracting the seasonal mean of the ocean basins, with the exception of Greenland and the time range 2071-2100 from the seasonal mean of the time Antarctic. However, the spatial extend of the differences are range 1901-1930. The differences are applied to mean not as pronounced in the higher latitudes. In addition, the values of CAPE and CIN as well as to their 95th, 80th, 60th, negative differences occur in very few and small regions 40th, 20th and 5th percentiles. only. In general, regions with a positive difference in CAPE show also a positive difference in CIN. This positive III. RESULTS & CONCLUSIONS The general patterns of the differences between present-day and future climates show a general increase of CAPE and CIN over all continents and ocean basins with the exception of Greenland and Antarctica. The differences are stronger pronounced in the A1B scenario than in the B1 scenario due to the stronger pronounced increase in greenhouse gasses. The analyses of the changes in the different percentiles reveal for CAPE and CIN that the higher the percentiles the stronger the changes over time. The magnitude of the difference as well as the spatial extent is strongest in the 95th percentile (FIG. 1 and 2). The differences in CAPE are in general positive and occur over the continents. Negative differences are visible mostly over the ocean basins of the southern hemisphere and over the eastern North Atlantic between 50° and 60° North. FIG. 2: Difference of JJA mean of 95th-percentile of CIN in J/kg The magnitude and the spatial extent of the between scenario A1B (2071-2100) and scenario 20C (1901-2000). Positive numbers indicate an increase of CAPE with time.

97 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

difference leads not only to an increase in potential convection but also in boundary layer stability, which prevents the development of convection. As the differences in CIN are mostly positive, regions with a decrease in CAPE reveal not only a decrease in potential convection but also in the likeliness of convection development. However, the regions of strongest differences in CIN are usually not concordant with the regions of strongest differences in CAPE. The strongest positive difference in CIN occurs in the Mediterranean during JJA, while the strongest negative difference in CIN is visible over the southern hemispheric eastern tropical Pacific.

IV. AKNOWLEDGMENTS Thanks to Frank Sielmann for his support in CAPE calculations, and to DKRZ, DWD, and ECMWF for the data. KRC acknowledges the support by IMPRS-ESM.

V. REFERENCES Colby JR FP., 1984: Convective inhibition as a predictor of convection during AVE-SESAME II. Mon. Wea. Rev., 112 2239-2252. Riemann-Campe K., Fraedrich K., Lunkeit F., 2009: Global climatology of convective available potential energy (CAPE) and convective inhibition (CIN) in ERA-40 reanalysis. Atmos. Res., 93. 534-545. Roeckner E., Bäuml G., Bonaventura L., Brokopf R., Esch M., Giorgetta M., Hagemann S., Kirchner I., Kornblueh L., Manzini E., Rhodin A., Schlese U., Schulzweida U., Tompkins A., 2003: The atmospheric general circulation model ECHAM5 part 1 model description. Max-Planck- Institut für Meteorologie, Report No. 349. Hamburg.

98 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

HAILPAD DATA ANALYSIS FOR CONTINENTAL PART OF CROATIA Damir Pocakal

1Meteorological and Hydrological Service of Croatia, 10000 Zagreb,Gric 3, Croatia, [email protected]

(Dated: 11 September 2009)

I. INTRODUCTION

Mainly in the summer months, Croatia is exposed to This distribution shows that 65.3 % of all recorded hail the thunderstorms with severe rain and hail, especially in its cases in continental part of Croatia have kinetic energy less continental part. In the 1960s, aiming to protect and reduce than 20 Joule /m2. Only three hail cases with intensity scale heavy damage in agriculture and other mobile and immobile H5 and maximum hail stone size between 31.3-35.4 mm, are properties, a hail suppression system was introduced to this recorded in time period 2002-2008. Spatial analysis show area. Several authors (Fraile et all. 2003; Dessens et all. that the location of these three cases was in eastern flat part 2009; Siutas et all. 2009) accent importance of hail of continental part of Croatia. Relative frequency measuring with hailpads for climatology and evaluation distribution of kinetic energy (point hail fall) for Croatia and studies in different countries. In order to receive precise and polygon are shown in Fig. 1. objective hailstone parameters, hailpads were installed during the season 2001 on each main meteorological and hail suppression station in continental part of Croatia. The Average / 2002 2003 2004 2005 2006 2007 2008 total number of hailpads today is 730 (stations and polygon). Year number of 1197 1089 1220 1363 900 990 1172 stones/ m2 II. PRESENTATION OF RESEARCH max. diam. 11.8 14.1 12.3 12.4 12.7 13.6 14.4 (mm) mass During the hail season (01.May-30.September) in 555.4 470.3 472.6 483.1 317.7 459.7 643.5 (g/m2) time period from 1891 to 2008, around 12500 reports of hail are collected. These reports contain exact information about K.E. 30.3 35.9 37.6 33.1 22.1 32.2 53.9 (J/m2) the location, date and time of hail fall. Unfortunately there are not physical parameters of hail in these reports. They contain only a short description or size comparison with TABLE I: Average values for hail parameters in continental part of Croatia. other objects (corn, pea, walnut, golf ball, etc.). Spatial analysis of mean number of days with hail, based on this reports (Pocakal et all. 2009) shows that the area with maximum hail days (Zagorje) is in the west continental part 80 of Croatia near the Slovenian border. 70 The hailpad polygon (POL), 30 x 20 km in size was established in this area before the hailfall season in 2002. 60 Hailpads were installed between the existing hailpads on hail 50 suppression stations. This way, a dense network with 150 40 2 hailpads with linear spacing of 2 km (1 hailpad /4km ) was 30 obtained.

This paper present analysis of all hailpad data (2049 frequency relative (%) 20 cases) collected in period from 2002 to 2008, for the whole 10 protected area, different sub regions and especially for 0 hailpad polygon. H0 H1 H2 H3 H4 H5 0-20 20-100 100-300 300-500 500-800 > 800 CRO 65,3 25 8,2 1,1 0,3 0,1

POL 67,5 23,9 8,3 0,3 0 0 III. RESULTS AND CONCLUSIONS

FIG. 1: Hail kinetic energy distribution for Croatia and polygon On the bases of all 2049 hail cases recorded on based on the Torro scale (2002-2008). hailpads in the continental part of Croatia average values per square meter, for max. diameter, number of stones, mass and

kinetic energy are calculated: (dmax.= 13.2 mm; n = 1197; Relative frequency distribution of hailstone size, show that m = 492.0 g and K.E.= 36.5 J). Annual average values for the first four size classes (interval 2.5 mm) from 5.0 mm till continental part of Croatia are shown in Tab.1. For easier 15.0 mm, contain more than 95 % of all stones. Annually comparison with similar analysis in other countries, Torro distribution for these four size classes show decreasing trend scale (Webb et all. 2001; Sioutas et all. 2009) is used for of smaller stones (5.0 – 7.5 mm) and increasing trend for kinetic energy distribution.

99 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

greater stones (10.1 – 15.0 mm). This trend is larger for the polygon, then for the whole protected area in Croatia (Fig. 2.).

Results show that the western parts have the highest frequency of hail fall with intensity between H0 – H4 of Torro scale. Increase of damaged hailpads per hail day on polygon from 2.1 in 2002 to 7.8 in 2008, together with the greater number of larger stones indicate a possible increase of hailstorm intensity in continental part of Croatia.

5.0-7.5 7.6-10.0 10.1-12.5 12.6-15.0 80

70

60

50

40

30 relative frequency (%) frequency relative

20

10

0 2002 2003 2004 2005 2006 2007 2008 POL

FIG. 2: Distribution and trend of different hail stone sizes (mm) on polygon (2002-2008).

IV. REFERENCES

Dessens J. Berthet C. and Sanchez J.L., 2009: Seeding optimization for hail prevention with ground generators. Jour.of Wea. Mod., 41 119-126. Fraile R., Berthet C., Dessens J. and Sanchez J.L., 2003: Return periods of severe hailfals computed from hailpad data. Atmos. Res., 67-68 189-202. Pocakal D., Vecenaj Z., and Stalec J., 2009: Hail Characteristics of Different Regions in Continental Part of Croatia. Atmos. Res., 93 516-525. Sioutas M., Meaden T. and Webb J.D.C., 2009: Hail Frequency, Distribution and Intensity in Northern Greece. Atmos. Res., 93 526-533. Webb J.D.C., Elsom D.M., and Reynolds D.J., 2001:Climatology of severe hailstorms in Great Britain.. Atmos. Res., 56 293-310.

100 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

REGIOEXAKT – REGIONAL RISK OF CONVECTIVE EXTREME WEATHER EVENTS: USER-ORIENTED CONCEPTS FOR TREND ASSESSMENT AND ADAPTATION Nikolai Dotzek1 and the RegioExAKT consortium2

1 Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, 82234 Wessling, Germany. e-Mail: [email protected] 2 DLR, FUB, RUB, RWTH, MR, FMG, TÜV, BM, IKT, NC, Sky, and ESSL. www.pa.op.dlr.de/RegioExAKT/info/index_en.html (Dated: 15 September 2009)

I. INTRODUCTION (using polarimetric radar and total lightning data) and Extreme weather events from severe thunderstorms drainage is developed based on the current situation and (damaging winds, hail, heavy precipitation and tornadoes) climate change scenarios. On behalf of the whole pose a threat to life and safety of European citizens and lead RegioExAKT consortium, this paper highlights selected to significant property damage. For Germany, the Munich results after two thirds of the project time. Reinsurance Group estimates a total damage of € 1 to 2 billion per year. For Europe as a whole, thunderstorms are II. RESULTS likely to cause € 5 to 8 billion annual total damage. A field Meteorological reanalysis and model data (ERA-40, regional of particular concern, and also at the cutting edge of science, and global climate, weather forecasting), wind- and water is the estimation of regionalised severe convective storm risk engineering, socio-economic approaches, advanced remote- in a changing climate with time horizon 2030 and beyond. sensing and in-situ observational tools are available to There is a strong demand for regionalised hazard address and satisfy user demands for adaptation guidance. assessments and adaptation strategies by weather-sensitive The assessment of the economic and climatologic hazard of economic sectors like the insurance industry, airports, water severe thunderstorms is strengthened by the recent management, and also national weather services like the availability of a pan-European severe weather database, DWD in Germany in its efforts for optimisation of forecasts ESWD (Fig. 2, www.essl.org/ESWD/, Dotzek et al., 2009). and warnings of such events. The adaptation of existing building codes with respect to wind loads and precipitation maxima to climatic trends in extreme weather events is also economically relevant. From these target groups, Munich international airport (Fig. 1) and the Munich Re Group were chosen as exemplary users and project partners.

FIG. 2: Incidence (reports per year and per 10,000 km2) of large hail (left) and heavy precipitation (right) in Germany between 1999 and 2008 on a 0.5° x 1° latitude-longitude grid based on the ESWD.

Hazard assessment for various severe thunderstorms phenomena based on ESWD reports is enhanced by evaluation of thunderstorm parameters in ERA-40 reanalysis data. The physical parameters evaluated are surface-based FIG. 1: Munich international airport, one of the targeted users. and mixed-layer convective available potential energy (CAPE), convective inhibition energy (CIN), deep-layer (0-6 1/2 The BMBF-funded klimazwei-project links the users km AGL) wind shear and the product of (2 CAPE) x deep- with an interdisciplinary research group (see project layer shear. The latter is the main metric for severe website). The three-year project which started in January thunderstorm potential. Note that CAPE alone is not a useful 2007 develops hydro-meteorological and insurance-related metric of thunderstorm potential. The ERA-40 analysis does scenarios of extreme weather events following from not show significant trends in severe thunderstorm potential regionalised climate and vulnerability projections compared over Germany in the period 1979-2001. to an assessment of the present state. Together with new ERA-40 data are also being applied to evaluate wind zone maps for Germany, this helps to enable timely precipitation extremes against rain gauge observations in the adaptation of insurance business strategies or building codes. recent decades and to identify the main weather patterns For Munich airport, an optimised thunderstorms nowcasting responsible for high precipitation events in (southern)

101 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

Germany. In addition, a long-term ECHAM5 climate model The same total lightning data sampled by the LINET run pointed to no significant changes in annual maxima of system is also used in the nowcasting algorithm ec-TRAM daily precipitation until about 2030, but an increased which combines radar data (cell identification and tracking) variability of these extremes afterwards. and LINET (also grouped in “cells” and tracked). Fig. 5 gives an example of a major thunderstorm day in southern Germany. This algorithm can also exploit polarimetric radar data as available with DLR’s POLDIRAD radar and the next generation of DWD’s operational radar network. This will improve the identification of the thunderstorms life cycles. Aside from the early detection and warning of hazardous cells approaching Munich airport, it is also possible to nowcast the time when the hazardous situation at the airport will be over – an important, cost-saving point in air traffic management (ATM) and air traffic control (ATC).

III. CONCLUSIONS FIG. 3: Adaptation options of the Munich airport drainage and After about three quarters of the project, we can identify this water management system. set of conclusions and recommendations:

The analysis of the water management and drainage  Severe thunderstorm parameters can be evaluated from facilities at Munich airport helped to identify quite a number large-scale fields, like those produced in the ERA-40 reanalysis or CLM regional climate model simulations; of options to adapt to potentially higher precipitation extremes in the future and to optimise the current  For recent decades and the present state, there is procedures. These measures are summarised in Fig. 3 and apparently no significant trend in severe thunderstorm encompass both the climate change control strategy and environments; environmental protection or the reduction of periods in  Initial results for the future climate rather point to an which the runways and taxiways are not fully serviceable enhanced variability of extremes after about 2030, than due to heavy rain or post-rain water layers. to a trend in the average level of severe convective storm activity;  The popular quantity CAPE should not be used on a stand-alone basis to draw conclusions on future severe thunderstorm occurrence. This holds in particular for surface-based CAPE being dominated by the predicted surface values of temperature and water vapour. Instead, mixed-layer CAPE should be used in concert with parameters like CIN and deep-layer wind shear;  For drainage and water management in urban infrastructures, effective adaptation measures can be developed. Similar arguments hold for current and

FIG. 4: Left: 20-h COSMO-DE forecast of DSI for 14 May 2007, future wind loads, for which RegioExAKT provides 2000 UTC, right: verification by LINET-measured number of new wind zone maps and novel high-resolution flashes between 2045 and 2115 UTC. measurements of gustiness for various wind events (winter storms, squall lines, downbursts); The development of short-term forecasting and  The application of DSI in short-term thunderstorm nowcasting algorithms is illustrated in Fig. 4 which shows a forecasting and of radar and total lightning data in 20-hour forecast field of the dynamic state index (DSI, nowcasting algorithms shows promising results and a Névir, 2004), verified by total lightning observations large potential for user-friendly display interfaces. (ground and cloud flashes). The remaining project time will focus on the climate change impact analysis and refinement and synthesis of results.

IV. ACKNOWLEDGMENTS This work was partly funded by the German Ministry for Education and Research (BMBF) under contract 01LS05125 in the project RegioExAKT (Regional Risk of Convective Extreme Weather Events: User-oriented Concepts for Climatic Trend Assessment and Adaptation, www.regioexakt.de).

V. REFERENCES Dotzek, N., P. Groenemeijer, B. Feuerstein, and A. M. Holzer, 2009: Overview of ESSL's severe convective storms research using the European Severe Weather Database ESWD. Atmos. Res., 93(1-3), 575-586.. Névir, P., 2004: Ertel´s vorticity theorems, the particle FIG. 5: ec-TRAM nowcast 21 July 2007, 2000 UTC. relabelling symmetry and the energy-vorticity theory of fluid mechanics. Meteorol. Z., 13, 485-498.

102 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

CLIMATE CHANGE IMPACTS ON SEVERE CONVECTIVE STORMS OVER EUROPE 1 1 Julia Sander and Nikolai Dotzek

1 Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, 82234 Wessling, Germany, [email protected] (Dated: 09 October 2009)

I. INTRODUCTION II. RESULTS

Forecasting thunderstorms is one of the most difficult CAPE provides a measure of the maximum updraft of tasks in weather prediction, due to their rather small spatial a statically unstable air parcel, the thermodynamic speed 1/2 and temporal extension and to the insufficient horizontal and limit: wmax = (2 CAPE) . Yet for the development of severe vertical resolution of operational numerical models. A thunderstorms not only thermo-dynamic conditions have to workshop on extreme weather and climate change, initiated be favourable but also dynamic conditions like deep-layer by the Intergovernmental Panel on Climate Change (IPCC, vertical wind shear between 0 and 6 km AGL. Vertical shear 2002) noted that observations of severe thunderstorms are has a particular effect on the evolution of convective storms not collected uniformly and that long, consistent records since it enables the internal organisation of convection by exist only for few locations. An emphasis on analysis of inducing local pressure gradient forces. An increase in environmental conditions for initiation of severe convection CAPE would mean a destabilisation of the atmosphere was recommended, and a variety of thermodynamic and which could be caused by an increase in moisture near the dynamic parameters have been derived (Brooks et al., 2007). surface or by cooling of the middle and upper troposphere. This is the basis for possible changes in distribution of the But high CAPE values will not necessarily correspond to environmental conditions associated with possible global actual convective activity since the convective initiation climate change (Brooks et al., 2007). might be inhibited by a capping inversion. The situation in Europe is very different from the Yet, Brooks et al. (2003) concluded that the higher situation in the U.S. because lapse rates in Europe are lower, both CAPE and deep-layer shear (DLS), the greater the reflecting the absence of a source of high lapse rate air probability that environmental conditions would support comparable to the Rocky Mountains. In Europe, high lapse severe convective weather, given the initiation of storms. rate values tend to be associated with low values of mixed- There are many other parameters that are also important for layer moisture qML. So on average, peak values of severe thunderstorms, e.g. lifted condensation level (LCL), Convective Available Potential Energy (CAPE) are lapse rate between the 800 and 500 hPa levels and mixing significantly lower there than over the Central Plains of ratio. The mean mixing ratio in the lowest 100 hPa, qML, North America (cf. Brooks et al., 2003). provides a direct measure of the lower-tropospheric A first analysis of European trends in thunderstorm moisture. Lapse rates between 800 hPa and 500 hPa can environments from ERA-40 data was presented by Sander provide information on the potential instability. Because the et al. (2008). The present study focuses on regional climate lapse rate calculation is tied to standard pressure levels here, model runs over Europe for the present, and one scenario of it cannot be a complete representation of the potential the future climate. The 20th century control climate is instability. Inversion layers just below 500 hPa, for example, evaluated for the “present-day” period 1979-2000 by might lead to underestimating the lapse rate and in turn the analysing convection indices calculated from meteorological potential instability. parameters of one simulation of the regional climate model CLM (www.clm-community.eu, Rockel et al., 2008). Future trends are derived from an A1B scenario model run in the timeframe 2079-2100. The A1B scenario describes a possible future world of very rapid economic growth, global population peaking in mid-century, and rapid introduction of new and more efficient technologies. A balance between energy sources is assumed.

Both CLM datasets have a horizontal resolution of 0.2° (257 x 271 grid points = 4500 x 5000 km), 6 pressure levels vertically and a 3 h temporal resolution. For the actual climate of the 20th century, the global forcing that drives the regional model is based on the same C20 control run, but initiated at three different initialization times. This gives three different realisations of the present climate. Our ongoing study will eventually provide an assessment of the thunderstorm probability for the current FIG. 1: Trend of severe thunderstorm potential (wmax x DLS, top) atmospheric situation and how this state is going to change and CIN (bottom) significant at the 90%-level (shading) for C20 under climate change conditions. Within the research project (1979-2000, left) and A1B scenario (2079-2100, right) in summer. RegioExAKT, the main objective is to estimate, if and how climate change might impact convection over Europe in the First, a summer seasonal mean of convection metrics at 12 future and what adaptation measures can be taken. UTC was derived. Then diurnal, seasonal and annual cycles

103 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY were removed to better estimate trends. Fig. 1 shows that scenario shows for CIN that areas with a trend significant at some areas with a positive trend are significant on the 90%- the 90%-level are shifted from the southern part of Norway level. While for present-day climate (C20), positive trends in towards the southeast to continental Europe although the wmax x DLS (Fig. 1, top left) appear at the Atlantic coastal absolute values essentially stay constant. We argue that the regions, in scenario A1B (Fig. 1, top right) such regions are probability of severe convection might increase in the shifted to the European northern mid-latitudes with a northern part of Central Europe due to an increase of severe maximum in the northern part of Germany and Poland. In convective storm potential wmax x DLS with little or no comparison with the C20 data, areas with negative trends in- change in CIN.

FIG. 2: Trend of convection metrics (top left: TLCL; top right: TEL; FIG. 3: As Fig. 2, but for the A1B scenario run (2079-2100). bottom left: qML; bottom right: lapse rate) for C20 (1979-2000). Ongoing further investigations will show how the crease in the southern parts of Europe. Also, the results of storm parameters evolve and interact and identify the factor the A1B scenario for CIN show that areas with a significant which yields the highest potential for a changing probability trend on the 90%-level are shifted from the southern part of of severe local storms over the 21st century. Norway towards the southeast to continental Europe. But there is hardly any trend, so values remain the same. IV. ACKNOWLEDGMENTS The temperature at the lifted condensation level (TLCL) increases over a wide area in scenario A1B (Fig. 3, top This work was partly funded by the German Ministry for Education right), however, the absolute values of trends are relatively and Research (BMBF) under contract 01LS05125 in the project small. A closer look how TLCL changed between 1979 and RegioExAKT (Regional Risk of Convective Extreme Weather 2079 reveals that TLCL increases on a large scale. In addition, Events: User-oriented Concepts for Climatic Trend Assessment and Adaptation, www.regioexakt.de) within the research programme the temperature of the equilibrium level (TEL) increases above continental Spain in scenario C20 (Fig. 2, lower left), klimazwei. but decreases in scenario A1B (Fig. 3, top right). But trends are more positive in areas of Central Europe e.g. southern V. REFERENCES Germany, , and northern . Absolute values show an increase over Spain and Northern Africa and Brooks, H. E., J. W. Lee, and J. P. Craven, 2003: The spatial a decrease over . The higher T values and distribution of severe thunderstorms and environments from LCL global reanalysis data. Atmos. Res., 67-68, 73-94. decreasing TEL and lapse rates values over France suggest Brooks, H. E., A. R. Anderson, K. Riemann, I. Ebbers, and H. that advection of warm moist air will increase and smaller Flachs, 2007: Climatological aspects of convective parameters from lapse rates result in the formation of taller thunderstorm the NCAR/NCEP reanalysis. Atmos. Res., 83(2-4), 294-305. clouds which may have a higher potential for severe IPCC, 2002: Workshop Report, IPCC Workshop on Changes weather. However, trends in specific humidity (qML) widely in Extreme Weather and Climate Events, Beijing, China, 11–13 June decrease in both CLM runs (Figs. 2 and 3, lower left). 2002, 107 pp. [Available at http://www.ipcc.ch/pdf/supporting- Trends in lapse rate weaken over a wide area, and are shifted material/ipcc-workshop-2002-06.pdf] towards the Iberian Peninsula and northeast of Europe in Rockel, B., A. Will, and A. Hense, 2008: The Regional Climate Model COSMO-CLM (CCLM). Meteorol. Z., 17(4), 347- scenario A1B (Fig. 3, lower right). 348. Sander, J., N. Dotzek, and R. Sausen, 2008: First results of III. CONCLUSIONS climate change impacts on severe convective storms over Europe. Preprints, 24th Conference on Severe Local Storms, Savannah, 27- 31 October 2008, Amer. Meteor. Soc., Boston, 4 pp. [Available at In the C20 data, areas of positive trends in wmax x DLS are apparent at the Atlantic coastal regions, whereas in http://ams.confex.com/ams/24SLS/techprogram/paper_142105.htm] scenario A1B the regions with positive trends are shifted to the European northern mid-latitudes with a maximum in northern Germany and Poland. Furthermore the A1B

104 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

EXTREME WEATHER EVENTS IN SOUTHERN GERMANY – CLIMATOLOGICAL RISK AND DEVELOPMENT OF A LARGE-SCALE IDENTIFICATION PROCEDURE A. Matthies1, G.C. Leckebusch2, T. Schartner3, J. Sander4, P. Névir5, U. Ulbrich6

1Freie Universität Berlin, Berlin, Germany, [email protected] 2 Freie Universität Berlin, Berlin, Germany, [email protected] 3 Freie Universität Berlin, Berlin, Germany, [email protected] 4DLR, Oberpfaffenhofen, Germany, [email protected] 5 Freie Universität Berlin, Berlin, Germany, [email protected] 6 Freie Universität Berlin, Berlin, Germany, [email protected] (Dated: 28. August 2009)

I. INTRODUCTION precipitation was found to be a suitable threshold for Extreme weather events like thunderstorms, hail and heavy extreme impact relevant precipitation events (Matthies et al., rain or snowfall can pose a threat to human life and to 2008). considerable tangible assets. For example on July 12th 1984 At step two of this study, parameters are identified capable an intensive hot spell ended by advection of cool westerly to assess large-scale conditions leading to extreme events. air masses entailing extreme weather phenomena around the For correct classification of the large-scale flow conditions frontal zone. One of these events, known as the “Munich in southern Germany an objective scheme for the Hailstorm”, caused overall losses of about 950 mill. US-$ classification of Lamb's circulation weather types (CWT's) including damages on more than 200.000 cars, 70.000 according to Jones et al. (1992) has proved to be most buildings and 180 aircrafts. Several greenhouses have been suitable. Certain CWT's have been turned out to be prone to devastated by hailstones of up to 5 cm in diameter (testified heavy (e.g. northern, western, cyclonic and anticyclonic) and maximum 9.5 cm) accumulating to vast layers of more than extreme precipitation (e.g. western and cyclonic, in winter 10 cm. About 400 people have been injured by hail and also anticyclonic) or on the other side to have a very low risk consequences of heavy rain and wind gusts. Yet there is a of such events (e.g. southern CWT’s). Analysing the lack of knowledge about present day climatological risk, its probability of precipitation exceeding the 90th percentile economic effects and its changes due to rising greenhouse spatial patterns show systematic underestimation in ERA40 gas concentrations. Therefore, parts of economy particularly compared to observations. Characteristics of horizontal sensitive to extreme weather events such as insurance distribution of extreme precipitation probability are companies and airports, require regional risk-analyses from reproduced well and further analyses of processes causing warning time scale to longer term estimations. extreme events are thus reasonable. Other large-scale parameters are tested additionally and in connection with CWT's to analyse the most suitable combination revealing the highest skill to identify extreme precipitation events in II. METHODOLOGY climate model data. The newly developed Dynamic State In this study an attempt to evaluate climatological risk of Index (DSI) already showed good skills tagging hurricanes extreme weather events for southern Germany in close (Weber and Névir, 2008) and frontal precipitation cooperation with stakeholders is made in a three-step- (Claussnitzer et al., 2008). For this study the DSI was used strategy. At first the extreme weather periods in summer and in two ways. At first to identify severe convective situations winter are identified via the connection of meteorological a Thunderstorm Index (TI) was created combining DSI at station data and impact data of project partners Munich Re the 600 hPa level and CAPE (1). and Munich Airport. The representation of extreme 5.0 precipitation events in model data of recent climate is DSITI ∗= CAPE 5.0 validated for ERA40 reanalyses. Diagnosis of large-scale (1) processes causing these events is accomplished by This index was tested in a case study of the “Munich classifying characteristics, intensity and frequency of Hailstorm” in ERA40 (Fig.1). The value reached in the grid- relevant situations. In order to estimate the risk under box containing Munich is the highest in 40 summer half anthropogenic climate change, findings will be transferred to years (1961-2000). Further results indicate that in simulations of AOGCM ECHAM5-OM1 and RCM CLM combination with the anomaly of the amount of precipitable (driven by ECHAM5-OM1) by assessing changes in water in the atmosphere, the TI is capable to identify relevant circulation structures compared to recent climate. extreme convective events in summer. Systematic tests are in progress. Concerning extreme large-scale events such as frontal precipitation the DSI in this study was secondly combined with differential advection of temperature as a III. RESULTS measure of instability. First results for both summer and Before starting to identify extreme weather events thresholds winter are not as good as for the combination identifying have to be defined to distinguish extreme from non-extreme convective summer events. Testing vertical profiles of DSI events with respect to stakeholder’s requirements. instead of single pressure levels is part of ongoing work. Comparing ERA40 and station data with impact records of Munich Re and Munich Airport, the 90th percentile of

105 5h European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

development of a nowcasting procedure. Geophys. Res. Abstr., Vol. 10, EGU2008-A-10029, EGU General Assembly 2008, ISSN 1029-7006 Weber, T. and Névir P. (2008): Storm tracks and cyclone development using the theoretical concept of the Dynamic State Index (DSI). Tellus A, 60(1): 1-10

Munich

(a)

(b)

FIG. 1: (a) TI on July 12th 1984, 18UTC and (b) satellite image of same date and time (German Weather Service)

IV. CONCLUSIONS Parameterisation of precipitation in global climate models suffers from large uncertainties. In this study large-scale parameters are tested in connection with CWT’s to find a combination that has the highest skill to identify extreme precipitation events in gridded data. For convective summer events a combination of CWT’s, the new Thunderstorm Index and precipitable water seems to work well. Large- scale events so far are identified best by a combination of CWT’s, DSI and differential temperature advection, but testing of additional parameters is in progress. In the third step of this study the findings will be transferred from reanalyses to simulations of global and regional climate models. The aim will be to detect if the frequency of exceeding the thresholds of these extreme event combinations changes or if there is a change in intensity of exceedance that can be credited to anthropogenic climate change according to IPCC scenario A1B.

V. REFERENCES

Claussnitzer, A., Névir P., Langer I., Reimer E. and Cubasch U. (2008): Scale-dependent analyses of precipitation forecasts and cloud properties using the Dynamic State Index. Meteorologische Zeitschrift, 17(6): 813-826 Jones, P.D., Hulme M. and Briffa K.R. (1992): A comparison of Lamb circulation types with an objective classification scheme. International Journal of Climatology, 13: 655-663 Matthies A., Schartner T., Leckebusch G.C., Rohlfing G., Névir P. and Ulbrich U. (2008): Extreme weather events in southern Germany – Climatological risk and

106 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

Influence of sounding derived parameters on the strength of tornadoes in Europe and the USA from Reanalysis data Stefanie Grünwald1, Harold Brooks2

1Meteorologisches Institut der Universität Hamburg, Bundesstrasse 55 D-20146 Hamburg, Germany, [email protected] 2NOAA/National Severe Storms Laboratory, 120 David L. Boren BLVD, Norman, OK 73072, USA, [email protected] (Dated: 15 September 2009)

I. INTRODUCTION these parameter combinations have also been generated for the unrated, the F1 and F0 tornadoes. The influence of parameters such as convective Before generating the distributions, a gaussian available potential energy (CAPE), wind shear and the smoother has been applied to the parameter combinations. lifting condensation level (LCL) on the formation of While applying the smoother on each combination, an tornadoes has been examined by several authors and these analysis grid was created where each analysis grid point parameters have been found to discriminate well between contains a value of how likely it is for the combination of tornadic and non-tornadic thunderstorms (e.g. Brooks et al., parameter values that it represents to appear in a tornado 2003; Rasmussen and Blanchard, 1998). However, what is sounding. The density distributions are based on the values still open to question is how they influence the strength of of these analysis grids. tornadoes. Thus, the effect of combinations of the mentioned parameters are analyzed on weak (F0, F1) and significant III. RESULTS AND CONCLUSIONS (F2+) tornadoes in Europe as well as in the US. In Europe, the dataset for the weak tornadoes is a) small. Also, the underreporting of F0 tornadoes is apparent

in Europe (Dotzek et al., 2009). Thus, for Europe, 100 significant tornadoes distributions of the parameter combinations for the F0 and weak tornadoes the F1 tornadoes are compared to the distributions of the 80 unrated tornadoes to see if the unrated tornadoes resemble

the F1 or the F0 tornadoes in order to include them into the 60 dataset for the weak tornadoes for extension of that data set.

40 II. PRESENTATION OF RESEARCH 0.5 1.5 1 20 8 6 4

The values of the parameters WMAX (CAPE in 0-6 km Wind Difference (m/s) terms of updraft velocity, here based on a parcel that is 0 2 2 mixed over the lowest 100 hPa), LCL, DLS (deep layer 0 10 20 30 40 shear, 0 to 6 km wind difference) and LLS (low level shear, LCL (m*100) 0 to 1 km wind difference) that are associated with tornadic b) environments have been used to analyse how they affect the strength of tornadoes in Europe and the USA. The updraft 100 significant tornadoes velocity WMAX is based on a parcel theory and is defined weak tornadoes as WMAX=sqrt(2xCAPE) (Holton, 1992). 80 These parameters have been derived by taking the information about where and when a tornado occured from 60 data from the European Severe Weather Database (ESWD) for the years 1958 to 1999 and from data from the Storm 40

Prediction Center (SPC) for the years 1991 to 1999 for the 0.1 0.05 0.2 US. Then, for these times and places the proximity 20 0.150.3 0.1

soundings deduced from the National Center for 0-6 km Wind Difference (m/s) Atmospheric Research (NCAR)/United States National 0 Center for Environmental Prediction (NCEP) reanalysis 0 10 20 30 40 (Kalney et al., 1996) were used and with help of the the LCL (m*100) Skew-t/Hodograph Analysis and Research Program (SHARP) (Hart and Korotky, 1991) the mentioned parameters were derived from the reanalysis soundings. FIG. 1: Density distributions of weak and significant tornadoes for Density distributions for the parameter combinations the parameter combination LCL/deep-layer shear for the US (a) and Europe (b). Values of contour lines, from inside to outside, are 1.5, WMAX/DLS, LCL/DLS and LCL/LLS have been generated 1 and 0.5 for significant and 8, 6, 4 and 2 for weak tornadoes (a) and for weak as well as for significant tornadoes for Europe and 0.3, 0.2 and 0.1 for significant and 0.15, 0.1 and 0.05 for weak the US. In addition, for Europe, density distributions for tornadoes (b)

107 5th European Conference on Severe Storms 12 - 16 October 2009 - Landshut - GERMANY

The density plots for the combination LCL/DLS an influence on the strength of tornadoes that compensate show that in the US (Fig. 1a) most significant tornadoes for the higher LCL heights. Future research should be done occur at higher DLS and lower LCL than most weak to find out about this. tornadoes, with a density maximum at LCL heights between The combination WMAX/DLS (not shown here) is a 650m and 1400 m and DLS values between 10 m/s and 21 better discriminator between the strength of tonadoes in m/s for the weak tornadoes and between 550m and 1100m Europe, than it is in the US, because in the US the WMAX and 16 m/s and 23 m/s for the significant tornadoes. In does almost not vary for weak and significant tornadoes. In Europe on the other hand, most significant tornadoes occur the US, most significant tornadoes occur at higher DLS, but at slightly higher DLS, but at slightly higher LCL than the at about the same WMAX as most weak tornadoes. In weak tornadoes, with a density maximum at LCL heights Europe, most significant tornadoes occur at higher DLS, but between 500 m and 1300 m and DLS values between 10 m/s also higher WMAX, than the weak tornadoes. and 22 m/s for the weak tornadoes and between 600 m and The density plot for the combination LCL/DLS for 1300 m and 15 m/s to 24 m/s for the significant tornadoes. the unrated and the F1 tornadoes (Fig. 2a) shows that most Thus, whereas in Europe the DLS shows similar behaviour unrated tornadoes occur at lower DLS and slightly lower concerning the strength of tornadoes as in the US, the LCL LCL than most F1 tornadoes. Since it was found for Europe shows the opposite behaviour. This is also true for the that the higher the DLS and the higher the LCL, the stronger combination LCL/LLS (not shown here). In the US, most a tornado, the unrated tornadoes should be weaker than the significant tornadoes occur at higher LLS and lower LCL F1 tornadoes and accordingly they must be F0 tornadoes than most weak tornadoes, whereas in Europe most then. Fig. 2b shows that the F0 tornadoes correspond rather significant tornadoes occur at slightly higher LLS, but well to the unrated tornadoes. However, since the density slightly higher LCL. maximum of the F0 tornadoes is located at the lower DLS Since the differences between weak and significant and slightly lower LCL part of the maximum for the unrated tornadoes for the combination LCL/DLS are bigger in the tornadoes the unrated tornadoes appear to be slightly US, than in Europe (Fig. 1), the combination LCL/DLS is a stronger than the F0 tornadoes. This implies that the unrated better discriminator between weak and significant tornadoes tornadoes mostly contain F0 tornadoes, since they show the in the US than in Europe. This is also the case for the best correspondence to the unrated tornadoes, but also some distribution LCL/LLS. The reason for the different stronger tornadoes, which might be F1 tornadoes only. behaviour of the LCL height in Europe compared to the US Comparison between unrated and F1 and F0 tornadoes for might be that in Europe another factor or other factors have the other two parameter combinations (not shown here)

a) showed the same results. Since the unrated tornadoes resemble the F0 tornadoes rather well, it is reasonable to

100 unrated tornadoes include the unrated tornadoes into the dataset for the weak F1 tornadoes tornadoes in future studies. 80

IV. AKNOWLEDGMENTS 60

The authors would like to thank Dr. Nikolai Dotzek for 40 providing ESWD data.

0.05 0.1 20 0.15

0.3 0.2 V. REFERENCES 0-6 km Wind Difference (m/s) 0 0.1 0.1 0 10 20 30 40 Brooks H.E, Lee J.W., Craven J.P., 2003: The spatial LCL (m*100) Distribution of severe thunderstorm and tornado b) environments from global reanalysis data. Atmos. Res., 67-68, 73-94

100 unrated tornadoes Dotzek N., Groenemeijer P., Feuerstein B., Holzer A.M., F0 tornadoes 2009: Overview of ESSL’s severe convective storms 80 research using the European Severe Weather Database ESWD. Atmos. Res., 93 575-586

60 Kalney E., Kanamitsu N., Kistler R., Collins W., Deaven D., Gandin L., Iredell M., Saha S., White G., Woollen J., Zhu

40 Y., Chelliah M., Ebisuzaki W., Higgings W., Janowiak J., Mo K. C., Ropelewski C., Wang J., Leetmaa A., Reynolds

20 0.02 B., Jenne R., Joseph D., 1996: The NCEP/NCAR 40-Year 0.3 0.2

0.025 Reanalysis Project. Bull. Am. Meteorol. Soc., 77 437-472 0-6 km Wind Difference (m/s) 0 0.005 0.010.1 0.015 0.1 0.005 Hart J.A., Korotky W.D., 1991: The SHARP workstation- 0 10 20 30 40 v1.50. A skew-t/hodograph analysis and research program LCL (m*100) fort the IBM and compatible PC. User’s manual. 62 pp. Available from NOAA/NWS Forecast office, Charleston, WV FIG. 2: Density distributions of unrated and F1 tornadoes (a) and of unrated and F0 tornadoes (b) for the parameter combination Holton J.R., 1992: An Introduction to Dynamic LCL/deep-layer shear. Values of contour lines, from inside to Meteorology. Academic Press inc., 3rd edition outside, are 0.3, 0.2 and 0.1 for unrated (a and b), 0.15, 0.1 and 0.05 Rasmussen E. N., Blanchard D. O., 1998: A baseline for F1 (a) and 0.025, 0.02, 0.015, 0.01 and 0.005 for F0 tornadoes Climatology of Sounding-Derived and Tornado (b). Forecast Parameters. Weather Forecast., 13 1148-1164

108