TorDACH Topics

More severe local storms due to global climate change?

Many people and almost all journalists raise this question nowadays. They argue that global warming would lead to more water vapor in the atmosphere and thereby to a higher available potential energy for thunderstorm formation. In principle, this cannot be ruled out, but conditions for severe thunderstorm formation are highly special and largely depend on small-scale influences. There cannot be a simple relation between global warming and thunderstorm activity during European summer. And besides, numerical climate model forecasts for e.g. southern regions of Germany show a trend towards only slightly warmer, but considerably dryer summers. The winters, however, are likely to become much warmer than today - and also significantly moister. That does not exactly support the speculation of more (severe) thunderstorms during summer, but instead indicates more and longer-lasting rain or large amounts of snow in winter.

Naively and superficially inspected, the TorDACH data of recent years indeed show an impressive increase of tornado and downburst reports. Looking at the time series of tornado reports from the 1940s on, this increase looks even more dramatic, cf. the following Fig. 1:

Fig. 1: Time series of tornado reports in Germany since 1800.

But this must not lead to the conclusion that this rise in tornado reports was primarily due to more tornado events, caused by anthropogenic "global change". Much more important here in Europe is the still strongly varying efficiency by which severe storm phenomena are being recognized and reported. Early in the 20th century, Alfred Wegener has collected many tornado reports from Germany and published them in his 1917 book Wind- und Wasserhosen in Europa. Therefore the increase in reports from 1880 on. In the 1930s, Johannes Letzmann did even more intensive research on tornadoes and tornado reports. Result: The 1930s still have the highest number of reported tornadoes in the record. Only the decade 2000-2009 will presumably exceed this number. So the question is: What percentage of actual events is being reported?

To substantiate that, let's make the following simple back-of-the-envelope calculation: In the 1940s, for obvious reasons, hardly anyone in Germany had paid attention to downbursts and tornadoes, and probably only roughly 5 to 10% of all events became known to the public. During the last, say, five years using internet-based weather newsgroups, TorDACH, Skywarn, and storm chaser reports, it is probable that about 80 to 90% of all severe storms happening were documented. This is equal to a dramatic increase by a factor from eight to eighteen! And this solely by more careful documentation of these cases - not because there really have been more severe storms occurring. Even if any heating up of the global climate caused a slight trend towards more thunderstorms, it would completely vanish within the extreme variation of observation efficiency during the last decades. One should keep in mind here that official authorities in the USA even in the 1920s had stated that the whole USA only experienced about 25 tornadoes a year (at the time, this served as evidence that coordinated tornado or general severe storm research was not worthwhile in the USA). Nowadays, the climatological average number of tornadoes per year in the USA varies in the range 1000 to 1200. Again, only looking closer at severe storms (and also looking at weaker cases) has increased the numbers so much, and not a climatic trend.

Yet if the probability that tornadoes are being detected and reported has risen so much in the USA and Europe over the last decades, then there is of course reason to ask if this increase affects all tornado intensities. It is often claimed that there are "more and more significant, i.e. strong and violent tornadoes". Is that true?

To clarify this, it is sensible to look at the USA data first, and then make a comparison using the German data. Table 1 shows the tornado reports in the USA, grouped by intensity on the Fujita scale in the decades from 1950 to 1999. It becomes obvious that a real upward trend in reports can only be diagnosed for the weak tornadoes (F0, F1). The number of reported significant tornadoes (F2 to F5), however, has remained nearly constant or has even slightly decreased.

Table 1: Tornado reports per decade and Fujita scale class in the USA, 1950-1999.

	F-2	F-1	F0	F1	F2	F3	F4	F5	F6
USA, 1990s	0	0	7370	3274	1065	339	81	10	0
USA, 1980s	0	0	3313	3329	1172	313	62	3	0
USA, 1970s	0	0	2396	3653	1910	570	107	16	0
USA, 1960s	0	0	1951	2615	1769	584	103	9	0
USA, 1950s	0	0	1038	1945	1346	466	112	8	0

Looking at the same data in Fig. 2a depicted as normalized intensity distributions, this increase in number of reported weak tornadoes becomes even more apparent. Coupled to this increase is a decrease in percentage of significant tornadoes in the USA during the last decades.

Fig. 2: a) Decadal tornado intensity distributions in the USA from 1920 to 1999, b) German tornado intensity distributions grouped into individual typical epochs of tornado research.

Fig. 2b gives such a picture for Germany, sorted by specific epochs with either intensive or more "coincidental" tornado research. Even based on this much smaller tornado record compared to the USA, the intensity distributions show a substantial increase of weak-tornado reports, especially since initiation of TorDACH in 1997. Simultaneously, a decrease in reports of strong or violent tornadoes has occurred over time. Concerning reports of weak F0 tornadoes, we can expect continued strong growth rates in Germany. However, there is no evidence for an increase of significant tornadoes: The F3 Acht tornado in 2003, for instance, was an event exactly fitting into the existing climatology.

"No evidence for increasing frequency of tornadoes" - This result is also supported by the two reports of the Intergovernmental Panel on Climate Change (IPCC, 2001, 2007). Further information on this subject is contained in the highly readable book by H. Kraus and U. Ebel, 2003: Risiko Wetter: Die Entstehung von Stürmen und anderen atmosphärischen Gefahren. Springer Verlag, Berlin, 250 pp. (in )

But even without climate-induced increase in severe local storms, there is a strong need for research in Europe. A reliable severe weather climatology is just evolving, and the risk assessment of severe local storm threat is still far from sophistication in some aspects. What we need therefore in Europe to work out and quality-control a European severe weather climatology, is an ongoing "looking closer", not only this and next year, but from now on continuously and sustainably. This is only possible with a professionally organized, European severe weather research institute, so to speak a European NSSL - the European Severe Storms Laboratory (ESSL). It should span a similarly broad spectrum of research topics as the NSSL: Thunderstorm-related flash floods, i.e. generally hydrometeorology over complex terrain, hail, downbursts, tornadoes, lightning, winter storms with snow and ice, and local regions of extreme impact in extra-tropical cyclones.