The Filkins-Broadwell Tornado

With all of the damage locations plotted on the overall map, it is possible to see a clear track running from WNW to ESE. The following map and aerial photograph cover exactly the same area. The photograph has been plotted with all of the damage points collected (red for trees down, yellow for building damage). The track of the tornado has been plotted assuming that all of these data points are within the tornado's footprint. The results are quite pleasing, in terms of verifying that this was indeed a tornado.

Aerial photograph of the Filkins/Broadwell/Alveston region, showing tornado path.
Image from www.getmapping.com

Overhead map of the Filkins/Broadwell/Alveston region, showing tornado path.
Image from www.multimap.co.uk

Based on these overview plots alone, the evidence is fairly conclusive that this path of damage could not have been caused by any other wind phenomena other than a tornado. The damage is almost non-existent outside of the main affected track, giving a path which is far too coherent to have been caused by the likes of straight-line winds and violent downdrafts.

At the widest point, the damage track as estimated on the above plots is around 170m wide (~190 yards). However, most of the damage we witnessed showed toppling to the east, implying that all damage we saw was most likely caused by westerly winds. In a cyclonic vortex, the westerly component is found on the southern edge of the tornado. In this instance, there was enough evidence to suggest that the tornado was indeed cyclonic.

The following table is taken from TORRO, and is used to classify tornadoes in the UK. The scale classifies tornadoes up to T10, but in the interest of this particular vortex, we've only included from T0 to T4.

TORRO Intensity	Description Of Tornado & Windspeeds	Description Of Damage (for guidance only)
T0	Light Tornado 17 - 24 m s-1 (39 - 54 mi h-1)	Loose light litter raised from ground-level in spirals. Tents, marquees seriously disturbed; most exposed tiles, slates on roofs dislodged. Twigs snapped; trail visible through crops.
T1	Mild Tornado 25 - 32 m s-1 (55 - 72 mi h-1)	Deckchairs, small plants, heavy litter becomes airborne; minor damage to sheds. More serious dislodging of tiles, slates, chimney pots. Wooden fences flattened. Slight damage to hedges and trees.
T2	Moderate Tornado 33 - 41 m s-1 (73 - 92 mi h-1)	Heavy mobile homes displaced, light caravans blown over, garden sheds destroyed, garage roofs torn away, much damage to tiled roofs and chimney stacks. General damage to trees, some big branches twisted or snapped off, small trees uprooted.
T3	Strong Tornado 42 - 51 m s-1 (93 - 114 mi h-1)	Mobile homes overturned / badly damaged; light caravans destroyed; garages and weak outbuildings destroyed; house roof timbers considerably exposed. Some of the bigger trees snapped or uprooted.
T4	Severe Tornado 52 - 61 m s-1 (115 - 136 mi h-1)	Motor cars levitated. Mobile homes airborne / destroyed; sheds airborne for considerable distances; entire roofs removed from some houses; roof timbers of stronger brick or stone houses completely exposed; gable ends torn away. Numerous trees uprooted or snapped.

The TORRO tornado intensity scale (from www.torro.org.uk)

For this investigation, we systematically looked at the damage descriptions for each tornado class, and compared them to what we had observed in the Filkins area. Firstly, we observed dislodged slates and snapped twigs, thus we can be sure that this tornado exhibited at least a T0 rating. Most of the characteristics for the T1 category were also more than satisfied, so we could therefore upgrade the Filkins event to T1.

However, we saw a good deal of damage in the category rated T2, and many of the images in this report back that up. There were no mobile homes or caravans in the region, but numerous garden sheds were indeed destroyed. Tiled roofs that were in the path of the tornado were badly damaged, and there was at least one instance of a toppled chimney stack. However, the damage to the roofs, and also a large number of the trees, was still indicative of stronger windspeeds than those rated T2.

T3 is the next most powerful tornado rating, and surprisingly we observed many of the suggested characteristics indicative of this intensity. Again, no caravans were observed, but a large garage structure was destroyed in Broadwell. In many of the badly damage properties, the roof timbers were exposed as all tiles, leading and roof felt were totally removed. Large trees were uprooted or significantly damaged, as many of the images prove. We can therefore suggest that this tornado, at least using the guidance scale, is a solid T3 classification.

Although one or two of the characteristics in T4 were satisfied in this instance, these may have been circumstantial. One point is worth bearing in mind however. The T3 rating would apply in general to "average" properties, and average tree-state. In this Cotswold region however, many of the properties are very strong indeed, with very thick stone walls, and even use the stone as roof coverings rather than using conventional slates. If some of the properties were less-firmly built, one could hypothesise that the damage may have been more substantial. Due to the time of year, the trees in the region were not in leaf. Trees which have a full canopy can have upwards of 20 times more wind resistance than trees in a winter state. In other words, how does this TORRO scale allow for tree-state? T3 requires that large trees are badly damaged or uprooted. This was indeed the case, but what if these trees, and surrounding trees which were only minorly damaged, had been in full leaf? We would suggest strongly that the damage to trees would have been substantially higher if this tornado had occurred in the spring, and this may have increased the intensity rating of the tornado.

However, based on the facts gathered from this tornado, in this location and at this time of year, we can conclude that from our findings that the Filkins tornado should be rated T3. Now, T3 corresponds roughly with a strong F1 on the Fujita scale, which has windspeeds up to 113mph. From isolated damage we observed, and due to the winter state of the trees, it is possible that the tornado may have intensified briefly to weak F2, with windspeeds exceeding 113mph for brief moments. The map and aerial photograph above displays the region where occasional damage was observed, which could be attributed to weak F2 damage, and this is shown by the green track colour. However, once again this is based on only scattered damage, and using the assumption that trees were in winter-state, and buildings were very strongly constructed.

In the interest of clarity, we will therefore conclude that this tornado was indeed a strong T3 (F1) with windspeeds peaking at 115mph. Now, the speed of the actual storm cell was in the region of 35-45mph. From residents' reports, and based on the strong flow at the time, it can be assumed that this vortex was moving eastwards with considerable speed. Based on how quickly it caught people out, one could believe that the tornado was in fact moving at up to 40mph. If we assumed this to be the case, and also that the strongest windspeeds on the south side of the tornado were 115mph (T3max), then we can roughly conclude that the actual rotating velocity of the tornado was about 75mph.

This therefore leaves the windspeeds on the northern edge of the vortex, at only 35mph (i.e. rotational velocity minus system speed). Clearly, 35mph would be extremely unlikely to cause damage. In tornado research, the largest degree of damage has always been observed where the south side of the vortex encountered obstacles, with damage on the north side more limited. In the case of the Filkins tornado, 115mph on the south side would easily cause considerable damage. Windspeeds of 75mph on both the eastward and westward flanks of the tornado would also cause some damage, although not as extreme. With windspeeds of 35mph, obstacles on the north side of the tornado would remain largely undamaged.

Looking back at the map and aerial photographs however, this leaves a dilemma. The damage path is up to 180 metres wide, with the most severe damage randomly scattered across this width. As we assumed that the strongest damage would occur on the south side of the vortex, this resulting array of scattered damage could point to one of two things. Firstly, every element of damage we encountered was encompassed by the south side of the tornado, and thus the actual vortex diameter was twice the width of the damage path. However, a tornado of over 350 metres wide in the UK would be quite spectacular, and is most unlikely. Another theory may be that the tornado took an oscillatory track from start to finish. In other words, rather than running directly along our hypothesised track, it oscillated about the centre of the track, moving northwards and southwards during its lifetime.

This is all speculation however, and cannot be proved without further data.

However, for the purposes of this report, it can be overwhelmingly proved that this was indeed a tornado with the following attributes:

• Wind speeds of 115mph theoretical maximum
• Estimated minimum track length of 3 miles
• Estimated minimum track width of 150-200 yards
• Estimated duration of grounded tornado of 4.5 minutes

This development and confirmation of this tornado now adds credit to the strength of this storm. Our final conclusions follow.

Page 15