A BRIEF SURVEY OF TORNADOES IN SWEDEN

Despite its northem latitude, Sweden experiences tornadoes; some of these have been notable.

This survey begins with mention of early concepts and experiments concerning tornadoes.

Outstanding events from the 18th and 19th centuries are described. Twentieth century events are summarized. A compilation of all the events is included.

Sweden has a long history of tornado occurrence. Although its high latitude would seem to inhibit tornado generation, the long days of summer can counteract this factor. (A similar combination of effects has led to the occurrence of tornadoes to the east of the Baltic Sea (peterson, 1992)). While the frequency of events cannot match that in the United States and other tornado-prone regions, notable torna-does have been reported in the scientific literature.

The existence of tornadoes was sufficient to inspire early conceptual and physical explanations.

In his remarkable book on the life and environment ofthe northem peoples of Europe Olaus Magnus (1555) commented on the origin of atmospheric vortices and their incredible effects.

For the most part he relied on classical authors (e.g., Seneca) for his explanations. The famous physicist Wilcke in Stockholm published results of several experiments simulating

tornadoes in the 1700s in the Proceedings of the Royal Academy of Science. There he described a simple experiment on the general structure of this phenomenon. He took a large cylindrical container of water with a light material on its bottom, e.g., powdered chalk. At a small distance from the center there was a thick steel wire placed vertically a few inches deep from the top surface. This wire was then rapidly rotated by a in a circle around the center of the container. After a while the upper layer of the water came into rotation; the centrifugal force led to a pressure decrease i the middle, and the surface became concave. In response, the deep-lying layer began to rise and rotate. Thus a spiral motion developed up the center of the container. The chalk initially moved toward the center but eventually rose.

It was presumed that in the atmosphere the rising air carries water vapor which eventually condenses and fills the vortex. In the laboratory if the base of the container was rotated, once again the fluid was induced to rotate. A light, colored fluid such as wine or oil put on the upper surface assumed a motion opposite that of the chalk in the previous experiment. Wilcke thought that the ultimate origin of tornadoes lay in electrical effects, although apparently he did not conduct relevant experiments. (The idea that lightning plays a role in tornado formation goes back to Aristotle.)

One of the earliest tornadic events recorded in Sweden occured at Möklinta, Västmanland, on 27 September 1725 (Kalsenius, 1725). In an influential book on tornadoes, the French physicist Peltier (1840) included an extensive extract from the Kalsenius eyewitness report. Not only did Kalsenus witness the tornado itself and its damages, but afterwards he conducted a survey of the track and interviewed other eyewitnesses. One aspect of the destruction was a swath through a forest tract.

The tornado was seen to form as a dense pillar above an expanse of land an eighth of a mile west of a lake. As it crossed the lake, a huge volume of water was raised about the vortex. Thereafter as it passed back over land, the swath experienced a deluge along with the damaging wind. At this stage the tornado was near the Kalsenius residence. Just previously he had been watching the gathering clouds, sensed an ominous rumble and had taken cover as rotating masses of water had begun to engulf him. Viewing the succeeding events from indoors, he saw the granary roof lifted as large timbers were removed like feathers. A massive gateway and roof were thrown down including large pine support posts. A chimney was destroyed with bricks cast all about. A servant woman later reported being caught up and carried 30-40 meters in the air.

Several controversial issues regarding tornadogenesis arose in the discussions of Swedish tornadoes of the nineteenth century. Superseding the lightning generation mechanism, the cause of storm formation was sought in the presence of contrary air currents (horizontal wind shear).

On the other hand some investigators emphasized vertical instability. Once the vortex had formed there was disagreement regarding the dominance of rotational versus vertical motions in the vortex, and whether the vertical motion is upward or downward. In the absence of photography and with the uncertainty of eyewitness reports, the primary data came from damage swath investigations.

Damage swaths through forests have been cited frequently in Europe and North America, even giving evidence of tornadoes in ages past. In a very influential monograph written during World War 1 the subsequently famous German meteorologist Alfred Wegener (1917) gave great attention to the tornado damage pattern, which he termed the "Åsgardsweg". Harkening back to Norse mythology, "Åsgardsweg" combines "åsgards" with the German word "weg". According to Grimm (1883)" souls that have not done so much good as to win heaven, nor yet harm enough to merit hell, drunkards, scoffers, tricksters, are doomed to ride about until the end of the world.. The horses are coal-black, having glowing eyes, and are govemed with fiery rods and iron reins, the noise of the troop is heard from afar..If you hear them come, you must get out of the way, or throw yourself flat on the ground and feign sleep, for there have been cases of living men being dragged along with the moving mass.. In some parts, this ghostly array is called aaskereria, aaskerej, aaskereida, in other hoselreia: the former corrupted from asgardreida, - reid, the Åsgard march. (Åsgard was where the gods lived)... Sometimes you do not see the procession, but only hear it rush through the air".

In Scandinavia - especially Sweden - the tall forests on thin glacial soil provide excellent conditions for the preservation of wind blowdowns and, in particular, tornado swaths. Wegener was aware of surprisingly numerous tornado swaths, notably during the 1800s, in Sweden. The distinctive term "Åsgardsweg" was widely used in the extensive pre-World War 2 German literature on tornadoes. (The term fell out of use after the War.)

Wegener' s book provides a detailed analysis of tornadoes reported in Europe. The dozen Swedish events he included provide numerous drawings of damage swaths, and an extended

extract of a tornado on 9 June 1882 in Nöttja (Hildebrandsson, 1884). A previous report by Hildebrandsson (1875) describing a tornado at Hallsberg was included in an earlier book on tornadoes and cyclones (Zurcher and Margolle, 1876).

The tornado which struck a farm three kilometers northwest of Hallsberg on 18 August 1875 was studied in detail by Hildebrandsson. After a morning of occasional rain, a very heavy shower was followed by the sudden formation of a tornado. Eyewitnesses described it as an inverted cone initially destroying a forested area before laying waste to several buildings and taking off the second story of the main building. The damage swath revealed trees uprooted lying with their crowns inward toward the path direction. Debris extended 2 km along the path; the width was up to 150 m.

Hildebrandsson attributed the formation to rapid destabilization due to strong surface heating, resulting in ascending motion. The convergent orientation of the trees was taken as verification of strong low-level converging winds.

The prominent French meteorologist Faye, on the other hand favored sinking motion and explained the tree patterns as due to the "screwing" effect of the twisting descent.

Wigert (1891) reported on damage due to a tornado on 4 July 1890 near the village of Vimmerby in Småland. The tornado was witnessed to form about 1600 hr moving to the north-northeast. The swath through forested areas was mapped in detail, revealing the gyratory motion of the tornado. Wegener laid great importance on these sketches.

One of the most detailed studies of tornadoes in Sweden focussed on events in 1939 and 1942 (Bath, 1945). In addition to detailed eyewitness descriptions, Bath presents weather maps and attempts a calculation of the energy available to tornadoes based on the vertical structure of the atmosphere. On the afternoon of 13 August 1939 a tornado occurred in Södermanland and Uppland (from Södra Björkfjärden to Göksby). The effects were experienced over a 3 hr period for a distance of 116.5 km, with a translation towards the north-northwest of 10.8 m/s

Eyewitnesses did not report a classical conical funnel, and there were interruptions in the swath.

Nevertheless the displacement of objects and the damaged tree patterns give evidence of the presence of a cyclonic tornado.

About three years later (11 August 1942) a tornado swept through Småland and Västergötland (from Örreryd to Broholm). It began in the late afternoon and moved north-northeast at about 11 m/s for a distance of 60 km. Once again the circulation appeared cyclonic, with most of the

damage to trees. An interesting aspect was the formation well along in the lifetime of a second vortex 200 m to the east of the original tornado; their paths were parallel.

On the same 11 August 1942 another tornado occurred in Närke. It formed about 1600 hr near Tångeråsa and lasted to Stenkulla, being on the ground on 1.5 km moving north-northeast. In this case a farmer described the formation and appearance of a classical funnel cloud.

Tornadoes in Sweden during the last forty years have not received wide attention. Numerous events have been noted in the Swedish press however. Many tornadoes in Sweden develop near a cold front. Vedin (1986) reported on such an event on the afternoon of 2 October 1986. The

strong tornado occurred near where Jämtland, Ångermanland and Medel-pad meet. It moved east to east-southeast cutting a 300 m swath in the forest. The track extended 9 km. (This was an area struck previously in 1942, with that damage swath still apparent!)

In the early evening of 16 August 1990 a tornado struck near Årjäng. Vedin (1990) visited the scene to gather reports. The tornado moved north about 21 km with a breadth of 100 m. At one point eyewitnesses saw three whirls crossing a lake, while over land there was evidence of two damage swaths.

Vedin (1996) noted a tornado 4 September 1996 near Simrishamn, as well as the simultaneous occurrence of six funnels outside Gotska Sandön, 4 August 1994; he also include a sketch of a tornado 10 August 1975 in Hofors.

Two tornadoes struck 6 November 1996 (Vedin, 1996; Vedin and Eggertson Karlström, 1996) one near Moheda, one at Åsle. In midafternoon the tornado formed west of Moheda, passing through the community towards the east-northeast for about 1.5 km. There was substantial damage on the grounds of the church and to businesses (roof tiles and windows). The Åsle tornado formed in early evening and more northeast for a distance of 5 km. Over 70 injured animals had to be put to death.

Table 1 gives a chronological listing of published tornado events. There is a great unevenness in the list; very few tornadoes are found after the 1880s until the last decade or so. This could reflect changes in the weather pattems, however more likely it results from changing policies in reporting, editorial interests and personal attention of meteorologists. (In the United States, all of these factors can be demonstrated to have affected the completeness of the record.) Perhaps there are sources of tornado observations that may yet be tapped. (For Estonia, for example, Tarand (1995) discovered many tornado reports in the annals of the Tallinn Botanical Gardens.)

Despite the probably incomplete record, it is of interest to consider a few of the characteristics.

Alexandersson (1986) reported a survey of tornadoes amounting to 150 events; Table 2 includes the monthly distribution he found.

The history of tornadoes in Sweden is yet to be fully described; many reports undoubtedly can be assembled to allow a more detailed characterization of events. The present evidence supports the following; tornadoes in Sweden

1. occur in association with active fronts,

2. are summer phenomena, with a peak frequency in August,

3. may exhibit quite long tracks,

4. often are not accompanied by a distinctive conical funnel, and

5. can cause considerable damage, particularly in forested areas.

At this point many aspects of tornado activity are not apparent from published reports; e.g., the appearance of the parent storm (visually, on radar or from satellite), the location beneath the storm (near the updraft along the gust front, etc.), nor the relation to hail swaths. With the most modern meteorological technology available, we can anticipate future clarification of these issues.