Chapter 9
9. CONCLUSIONS AND DISCUSSIONS
Studies related to the Edmonton tornado and to the 1985 Ontario tornadoes had numerous similarities. Both events were documented by Environment Canada forecasters. Epidemiological reviews of the human injuries associated with both disasters were completed. The failures of buildings in Edmonton and Ontario were described by engineers. In homes, fatalities occurred in those in which the ground floors were lifted. Substantial segments of the power grids in the 2 regions were destroyed, though the loss of electricity in Ontario appeared to interfere with the dissemination of weather warnings to affected communities. The loss of power in Edmonton apparently had little affect upon the broadcasts of warnings. Two investigations were unique to the Edmonton storms: the newspaper survey and the Hage review of the weather warning system.
The newspaper survey provided an extraordinary opportunity to document the meteorological phenomena observed by the citizens of Greater Edmonton. To ensure that the information would be available to future generations; copies of the returned surveys were donated to the University and Provincial Archives. Information derived from the survey was supplemented with data from unusual sources. A map displaying this non-survey information (Fig. 5) indicated that nearly all of Edmonton was affected by the storms of July 31, 1987: most floods occurred in the northern half of Edmonton; claims for disaster assistance submitted by businesses denoted the tornado path; reshinglings were common in central Edmonton and prodigious in the suburbs of south central Edmonton. Information about particular damage sites, both residential and commercial, illustrated the breadth and severity of the Edmonton tornado.
The existence of 1 of the 2 spin-off tornadoes in east Edmonton seemed to be confirmed by the numerous claims for disaster assistance made by businesses located east of the main tornado's path (Fig. 5). Apparently none of the survey respondents saw a spin-off tornado. Participants, however, saw numerous funnel clouds over west Edmonton during the afternoon and evening storms (Fig. 6), but none of the funnels seemed to touch down. Video tapes and photographs taken from the air and the ground confirmed damage at many obscure sites. The reports of wind damage by respondents in west and northwest Edmonton were supported by official observations of violent winds at the airports.
The shape of the tornado varied considerably as it travelled northwards. Shapes reported by participants at various locations (Fig. 6) were in good agreement with those shown in photographs and video tapes. Many respondents reported seeing multiple funnels, a phenomenon associated with suction vortices and an indicator of a particularly severe storm. Video tapes of the tornado were used to calculate the wind speeds at 2 locations: the estimates were lower than the speeds implied by the damage. The shapes recorded in northeast Edmonton suggested that heavy rainfall obscured the approach of the tornado during its last few minutes.
As the tornado moved northwards, there was an increase in the median warning times for viewers of the tornado (Fig. 10). This trend ended when the tornado became obscured by rainfall. At most locations, those who claimed to have been assisted by radio or television broadcasts had approximately 10 minutes more warning than those without such assistance. The additional warning time provided by broadcasts did not increase significantly as the tornado moved northwards. This suggested that the value of warnings carried by broadcasters was limited but genuine. The warning times reported in other surveys (Table 2) were not consistent with one another; these were not examined for relationships between the positions of the respondents and the locations of the tornadoes. The analysis of warning times recorded by participants in the newspaper survey should interest social scientists and meteorologists studying public responses to hazards.
The maximum hail sizes recorded by 638 participants were plotted and thoroughly analyzed. A record was established for the largest hailstone to fall in Alberta, 264 g. The total insured loss, $250 million, was a record for natural disasters in Canada, and 50000 of the 60000 insurance claims were for damage to houses and automobiles caused by hail.
The maximum dimensions of the largest hailstones in the hail samples collected from 50 respondents were categorized. These categories were then compared with the categories recorded by respondents. For 90% of the 50 hailstones, the 2 categorizations were the same, though one half of the measurements by participants were more than 20% larger than those taken in the laboratory.
The enormous hailstones that buffeted Edmonton were much larger than those which fell in Calgary on September 7, 1991. The insured loss caused by the Calgary storm, however, set the record for the most expensive natural disaster in Canadian history, $400 million.
Figure 15 is a map showing the highest-priority damage types recorded by respondents. It demonstrated, as did the map of non-survey information (Fig. 5), that most of Greater Edmonton was affected by severe wind, giant hailstones, or heavy rain on July 31, 1987.
Three Alberta thunderstorms (the Calgary hailstorm of 1981, the Edmonton tornado and hailstorm of 1987, and the Calgary hailstorm of 1991) set successive Canadian records for an insurance loss caused by a natural disaster, though insurance payments for plant and infrastructure repairs and for production losses caused by the July 1996 floods in the Saguenay region of Quebec may surpass the record. (Alan Wood suggested that a significant portion of the insured loss in Saguenay may not be attributed to surface flooding claims but to lawsuits against the owners of some of the dams.) For industrial plants, damage caused by surface flooding is usually reimbursed by insurance companies. Insurance policies for houses, however, typically do not include financial protection for losses caused by surface flooding.
A summary of the analysis of the relationship between various hail parameters and shingle damage was provided at the end of Section 7. A few points should be emphasized. Claims for roof repairs typically consumed more than one-half of the total insurance settlement for urban hailstorms in Alberta (Charlton and Kachman, 1996). This provided the motivation for the attempt to determine the relationship between shingle damage and hailstone size. For Edmonton, the severity of shingle damage was found to be related to the maximum and the most common hailstone sizes, the duration of the hailfall, and the ages of the shingles. Asphalt shingle manufacturers may be especially interested in the analysis of reshinglings conducted in the year following the storms, as well as the summary of laboratory studies about the resistance of shingles to impacts.
Thunderstorm events in Edmonton and Calgary are comparable in frequency but Edmonton has many more tornado sitings while Calgary has more hailstorms. However, hailstorm information for Calgary dating back to the 1950's suggested that severe events have been unusually frequent in Calgary during the 1990's.
The survey responses provided an exceptional source of information about the Edmonton tornado and hailstorm, perhaps the most violent thunderstorm to strike a major Canadian city this century. If a similar storm should strike another city, any researcher who wishes to study the event would do well to consider surveying the public through the local newspapers.
Finally, this study, with a decade of hindsight woven into it, should be useful to anyone who needs information about the Edmonton tornado and hailstorm.