Chapter 6
6. DAMAGE REPORTS
Figure 15 includes a detailed sketch of the path of tornado damage, the hail boundaries from Fig. 12, and coded reports of damage described by survey respondents. This rendition of the path of tornado damage, along with the times when the tornado was at certain locations, was adapted from the field study organized by Wallace (1987). The appropriate Fujita Tornado Intensity Scales (Fujita, 1973), FTIS, determined by ALWC meteorologists (Wallace, 1987), are plotted beside each region of "considerable or greater damage", depicted by dark shading. Fujita (1971) associated each "F number" with a single word describing damage and with a suggested minimum wind speed: F0 (light, 64 km/h), F1 (moderate, 117 km/h), F2 (considerable, 181 km/h), F3 (severe, 253 km/h), F4 (devastating, 331 km/h), F5 (incredible, 418 km/h), F6 (inconceivable, 513 km/h).
a. Damage reported in the field survey
As meteorologists conducted the field study, they used the specific types of damage associated with each F number to classify the damage along the tornado path. For regions with shingle damage or some snapped or uprooted trees, the damage was labelled F1; if roofs were pulled from frame houses but walls remained, mobile homes (trailers) were destroyed, or empty railroad boxcars were pushed on their sides, the damage was categorized as F2; for areas where roofs and some walls were torn from frame houses, trains were overturned, sheathing was torn from steel framed structures, or automobiles were lifted from the ground and rolled; damage was labelled F3; in areas where frame houses were levelled, the steel frames of structures were badly damaged, trees were debarked, or cars or trains were tossed or rolled for sizable distances damage was ranked F4. No instances were found where frame houses were pulled from their foundations, steel-reinforced pre-stressed concrete buildings were demolished, or automobiles were tossed vast distances like a missile; therefore, none of the damage observed during the field study was ranked F5. Fujita (1971) contained detailed information about the development of the Fujita Tornado Intensity Scale (FTIS): for example, the minimum wind speed needed to produce F1 damage, and thus an F1 tornado, is the same as that needed to upgrade a tropical storm to a hurricane.
Figure 15 includes an inset depicting the 4.8 km damage path south of the city limits, but it does not show the location of the first touchdown, which occurred at 1455 MDT. This time represents 0 minutes in Fig. 10. Wallace (1987) reported that no damage to the fields and trees was found in the region where the first touchdown occurred. Using the average speed of the funnel cloud between the second touchdown and the city limit (0.0 N), the first touchdown was estimated to have occurred 12 km south of the city limit.
b. Damage reported by participants
The responses from the 755 participants within Greater Edmonton were scrutinized for information about damage. Many respondents recorded damage to more than 1 object. Reports about the damaged objects were organized into 8 groups which, if prioritized, corresponded to the table of damage codes shown in Fig. 15. Opinions about the durability of various structures defined the order of the groups. Thus, all of the damage recorded in each survey form could be represented by a maximum of 8 damage groups, each of which was subdivided by the cause of damage, hail or wind; the group of highest priority for each respondent was the one plotted in Fig. 15: lower case letters were used as symbols to denote hail damage, and capital letters were used for wind damage. For example, the damage report of a respondent who saw branches pulled from trees by wind, an automobile dented by hail, and a house without all or part of its roof would be categorized as 'T', 'a', 'H', but his response would be plotted in Fig. 15 as 'H'. Generally, the more money needed to repair the damage, the higher the priority of the damage. Each damage code in Fig. 15 was plotted at the location where a participant observed his damage of highest priority. The symbols were plotted by hand and, when necessary, shifted from their exact positions to avoid excessive overlapping.
Some of the groups included one type of damaged object, others included many types: structural damage to industrial buildings was classified as "Industrial"; "Home structure" required structural damage to wood-frame houses; "Heavy equipment" denoted the tossing or toppling of automobiles, railway boxcars, or massive industrial equipment; "Roof repair" denoted the removal or denting of shingles or chimneys; "Auto" required that automobiles be damaged but not thrown; "Building" represented non-structural damage to buildings, including broken windows and dented siding but excluding roof damage; "Garden" included damage to gardens, lawn furniture, lawns, fences, and garbage cans; "Trees" required the loss of tree limbs or branches or the fracturing of trunks, but did not include the loss of twigs or leaves.
Many respondents who noted damaged objects failed to state the locations where they observed damage. Their damage reports were ignored. Some respondents did not mention whether wind or hail caused the damage that they witnessed. For these reports, the damage was assumed to have been caused by hail provided that they reported hailstones of golfball-size or larger; conversely, if a respondent who recorded damage ignored the Hail Report (Fig. 1b), claimed that it was not applicable (N/A), or reported the largest hailstone as walnut-sized or smaller, the damage was assumed to have been caused by wind.
All industrial damage was caused by wind ('I'), and all lay either in the regions of considerable or greater damage, determined by the field study, or within 0.5 km of these regions (Fig. 15). This damage implied winds of F2 to F4 intensity. Respondents reported structural damage due to wind to 4 houses ('H') in Mill Woods (Fig. 15), where the roofs of at least 9 houses were fully (F2) or partially removed (Tower, 1989). This damage could not be confirmed using the black and white air photographs borrowed from the City or the colour air photographs, which showed only a small portion of the northeastern edge of Mill Woods. Sixteen houses in Clareview had their roofs and, in some instances, parts of their walls removed. Another Clairview house had even more damage. It was left with only the floor and part of 1 wall (F3 to F4). The structural damage to these 17 houses (Lux, 1990) was confirmed by the colour air photographs and by 2 homeowners who responded to the survey. Their 2 'H's are plotted in Fig. 15.
Four respondents knew of roofs that were pierced by hail, but only 3 of them noted the locations; each is plotted as an 'h' in Fig. 15. The records of a roofing firm provided additional examples of roofs pierced by hail; this will be discussed in Section 7. Most of the reports of non-structural roof damage and automobile damage due to hail, plotted as 'r' and 'a', respectively, were confined to the golfball and tennis ball areas. Many respondents who recorded garden and tree damage also recorded higher priority damage types; therefore, garden and tree damage, due to hail and wind, were much more common than is suggested by Fig. 15. The number and distribution of reports of wind damage suggested that strong wind gusts buffeted nearly all of Greater Edmonton.
c. The wind and hail of the evening storm
Numerous reports of wind damage to trees, roofs, gardens, and buildings came from west and northwest Edmonton, as well as St. Albert (Figure 15). The times of occurrence accompanying these reports suggested that a storm traversed this region at approximately 1800 MDT. Weather observations taken at the Municipal and Namao Airports, the locations of which are shown in Fig. 2, attest to the ferocity of the evening storm. At 1800 MDT, the weather observer at the Municipal Airport (7.2 E, 12.0 N) experienced the strongest measured winds of the day, 83 km/h, accompanied by small hail. At 1812 MDT, the strongest winds of the day at the Namao Airport (9.2 E, 18.5 N), 97 km/h, were reported as a squall, that is, a prolonged gust. According to the Beaufort Wind Scale, which states that such speeds are seldom experienced over land, these speeds were sufficient to uproot some trees and cause incidents of structural damage. The speed measured at the Municipal Airport, 83 km/h, was in the range for strong gales but close to the minimum for storms. The speed measured at the Namao Airport, 97 km/h, was in the Beaufort range for storms. A heavy hailstorm and an intense thunderstorm were recorded by the observer at the Namao Airport between 1812 and 1821 MDT, but the size of the hailstones was not recorded. Earlier in the day, as the tornado ravaged northeast Edmonton, 2.5 cm hail had fallen at the Namao Airport, and "marble-sized" hail had been observed at the Municipal Airport.
The severe winds from the evening storm proceeded north of the Namao Airport and beyond the northern boundary of the figures. This was confirmed by 2 survey respondents. One noted that part of the gymnasium roof of the Sturgeon Composite High School (8.0 E, 22.0 N), located 5 km north of the Namao Airport, was lifted; he also knew of a site 3 km northeast of the high school where more than 100 trees were blown down. The second participant, located 4 km north of the high school, reported experiencing strong winds at 1825 MDT. In the evening, damaging winds were not restricted to west and northwest Edmonton; a survey participant, who was a trained weather observer, reported a "downburst" which flattened a wooden fence at his home (9.8 E, 8.5 N) southeast of the city centre. The most intensive rainfall passed with the tornado, but these strong, widespread winds hampered the provision of emergency services (Alberta Public Safety Services, 1990).
Apparently no mention of the damage to the Sturgeon Composite High School was made in the daily newspapers or other literature documenting the storms of July 31, 1987. Confirmation was obtained from Alf Sadae, an operations supervisor with the Sturgeon School Division, who stated that one half of the gymnasium roof lifted and folded over the remaining roof. Repairs were completed on August 3, just 4 days after the tornado. Locations where windstorms had damaged schools in Alberta, and the dates of these events, were found in Alberta Tornados, Other Destructive Windstorms and Lightning Fatalities, 1879-1984, by K.D. Hage (1994). In the 106 years before 1985, tornadoes destroyed or badly damaged 11 schools and removed roofs from 6 others; non-tornadic windstorms destroyed or badly damaged 9 schools and removed roofs from 9 others. There were no fatalities at any of these destroyed or damaged schools; nearly all of the incidents occurred in the normal period for summer holidays when the schools would have been unoccupied. The Annual Summer Severe Weather Reports from the Alberta Weather Centre (1984-1996) do not have consistently organized summaries of the types or numbers of buildings damaged by windstorms. No incident of school damage was found in these reports. The reference book by Hage (1994) is a valuable and unique source of historical information about damaging windstorms. It confirmed the rarity of the lifting of school roofs by winds in Alberta.
d. Interpretation of the reported damage
As expected, numerous reports of hail damage to automobiles and roofs ('a' and 'r', respectively, in Fig. 15) came from the 2 communities where reshinglings were common (Fig. 5). The areas of cross-hatching in Fig. 15 are regions where numerous respondents observed large hail (see Fig. 12) but did not report any damage. These regions consist principally of light commercial buildings or parkland. Reports of wind damage from east Edmonton and Strathcona County denoted, to a reasonable degree, the path of the tornado; these reports corresponded to local values of the FTIS (Fig. 15) reported by Wallace (1987). Numerous reports of damage in the industrial area, however, came from locations east of Wallace's damage path. Either the damage path was wider in some places than depicted by Wallace (1987), or the spin-off tornado that passed Maple Ridge (Fig. 2) caused the wreckage east of the damage path.
An informed examination of buildings damaged by a tornado requires some knowledge of both meteorology and structural engineering. Marshall (1993) concluded that variations in the construction of wood-framed buildings might lead to an F scale number with a confidence, at best, of plus or minus 1 F scale. Grazulis (1993) noted that the FTIS neither discriminates between types of house construction, nor does it define descriptive terms like "destroyed". For example, a tornado that removes the roof of a house is rated F2, but the FTIS does not state the minimum percentage of the roof which must be removed to warrant the F2 rating. Furthermore, Grazulis listed 29 types of property damage which should be attributed to F1 winds but are usually classified F2. Using his criteria, he found that 2567 of 8273 F2-F5 ratings contained in the National Severe Storms Forecast Center's records (1950-1989) should have been rated F1. The destruction of steel- framed buildings is classified an F4 event (Fujita, 1971), but the annihilation of such structures in the industrial area of east Edmonton and Strathcona County suggested that the anchorage of a building to its foundation and the connection of its components to one another are vital factors in its resistance to wind (Lux, 1990). David Ungstad (see subsection 3g) noted that sections of numerous concrete-block buildings in the industrial area collapsed because the roofs were not adequately connected to the floors through the walls. Moreover, he stated that the quality of the welds joining components in some steel-framed buildings was poor, and, consequently, the welds were not sufficiently strong to prevent winds from twisting these buildings apart. Ungstad also shared his opinions with Barrett (1995), who interviewed him for the Edmonton Journal. Warehouses without internal walls to assist in bracing the building seemed to be particularly vulnerable to the winds of the Edmonton tornado. Perhaps the demolition of steel-framed buildings should not be assumed to be an F4 event.
Tornadoes are often said to 'skip' when some houses in a block are severely damaged while others are nearly unaffected; this pattern of damage is more likely caused by inconsistencies in the quality of the houses (Rosenfeld, 1994) or suction vortices (Fujita and Smith, 1993). A row of 17 houses in Clareview was severely damaged, including 1 house which was practically levelled. To the east of these houses, approximately 100 m, were some twisted, steel transmission towers, but to the west of these houses, across the street, was a row of houses that, with the exception of a few which lost sections of their roof sheathings, suffered little more than shingle and missile damage (Lux, 1990). Perhaps the difference in wind speeds between the 2 rows of houses was modest, but the wind along the row of badly damaged houses was just strong enough to weaken the houses' structural integrity and cause the loss of roofs and some walls.
The aerial photographs discussed in Section 3 could be used to reevaluate the tornado's path and FTIS values depicted in Fig. 15. This task should employ the recent research compiled in The Tornado: Its Structure, Dynamics, Prediction and Hazards (Church et al., 1993) and the techniques for damage evaluation developed by Bunting and Smith (1993) for use by the National Weather Service in the United States.