Chapter 3
3. INFORMATION FROM VARIOUS SOURCES
Most of this section describes, in considerable detail, the materials compiled to plot the non-survey information shown in Fig. 5. For the eighth anniversary of the tornado, The Edmonton Journal published their rendition of an early version of Fig. 5, along with an article which described this research (Barret, 1995).
a. Flooding
The reports of sewer flooding made by 94 respondents are represented by circles in Fig. 4. Unfortunately, the Rain Report (Fig. 1b) did not explicitly request "flooded basement", though many survey participants checked 'other' and stated that their basements flooded. The regions where sewer flooding was common (Fig. 4) were, to a considerable degree, in the same areas where basement flooding was common, as reported by Edmonton Water and Sanitation (EWS) and outlined in Fig. 5. Minor flooding at both the Charles Camsell (5.8 E, 11.6 N) and the General (7.1 E, 9.8 N) hospitals proved to be modest inconveniences (Scanlon and Hiscott, 1994).
EWS tended to 271 flooded basements reported from 1500 MDT July 31 through August 4 (Bowen, 1994). Crews also cleaned 64 street locations flooded because of plugged catch basins, and they refit 39 lifted manhole covers (Bowen, 1994). The 4 areas enclosed by bold dotted lines in Fig. 5 encircle 153 of the 271 flooded basements; the fine dotted line encloses all but 18 of the 271 reports. The number of homeowners who drained their basements without assistance could not be determined. EWS divides the city into 25 service areas. The numbers of lifted manhole covers and plugged catch basins, which caused streets to flood, are plotted in Fig. 5 near the developed centres of each service area.
b. Rain in Greater Edmonton
The patterns of flooding in Fig. 4 and Fig. 5 suggest that the heaviest rains fell in south-central and north Edmonton. EWS provided records from their 16 tipping-bucket rain gauges (Ward, 1994). The operation of each gauge was monitored by comparing its measurements with weekly amounts collected in cylindrical rain gauges placed at the same site. Furthermore, the tipping-bucket gauges were calibrated at least once per year. The tipping-bucket gauge at the Municipal Airport (7.2 E, 12.0 N) was shared with Environment Canada. The Canadian Armed Forces and EWS both had a tipping-bucket gauge at CFB Edmonton, known locally as the Namao Airport (9.2 E, 18.7 N).
For each EWS tipping-bucket rain gauge, the information was stored on-site using battery powered equipment, and later transferred to EWS via telephone lines. The transmission of the data compiled during the storm was delayed, but none of the information was lost. The rainfall amounts were available for 5 minute intervals. Figure 5 shows the 24 hour and maximum 60 minute rainfall totals from 16 sites for July 31. The 24 hour rainfall amount plotted in northeast Edmonton (13.5 E, 17.0 N) was from a cylindrical rain gauge located at a climatological station sponsored by Environment Canada.
Figure 5 indicates that most areas in south- central and north Edmonton received more than 40 mm of rain on July 31, and near the regions of frequent basement flooding, approximately 30 to 40 mm of rain fell in 1 hour. The low values at 5.8 E, 10.4 N, that is, the station with the hourly and daily totals of 2 mm, were obviously caused by a malfunction. All other EWS gauges recorded their maximum 60 minute totals as the tornadic storm passed over the city, and they seemed to operate correctly. Heavy showers also accompanied a windy evening storm that struck Edmonton at 1800 MDT after doing extensive hail damage to crops southwest of Edmonton (Charlton et al., 1995), but the 60 minute precipitation amounts did not match those from the afternoon storm. One-hour rainfall amounts of 30 to 40 mm were apparently capable of causing widespread flooding. For any location in Edmonton, the average time between one-hour rainfalls of 30 mm or more is approximately 5 years, and for 40 mm or more, it is approximately 15 years (Bruce, 1968).
A 4 m x 8 m oval-shaped, storm-water tunnel buried in the Kennendale Ravine of south Clareview ('B' in Fig. 5 and Table 1) suffered damage when the tornado passed its outlet to the river causing a local pressure decrease. The bottom of 330 metres of the pipe was pulled towards the top of the pipe! Replacement of this section of the tunnel cost more than $2 million. Christopher Ward and Andy Bowen, employees of EWS, stated that the blockage of the storm-water tunnel probably did not contribute significantly to the street, sewer, or basement flooding in the catchment area to the west. Immediately after the tunnel folded upward, it probably burst, allowing storm water to flow beneath the collapsed section. Bowen does not know whether the storm sewer was plugged by heavy rain when the tornado passed its outlet at approximately 1545 MDT. At that time, the 2 tipping-bucket gauges closest to the storm sewer, Kennendale Yard and Norwood, had 5 minute rainfall rates which were equivalent to 70 mm per hour and 80 mm per hour, respectively. The tornado also drew air through a 300 m long building at Stelco Inc., though the building, unlike the Kennendale storm sewer, was not sucked flat. The damage to Stelco Inc. is discussed in sub-section g.
c. Weather across Alberta
Charlton et al. (1995) noted that rainfall continued intermittently in Edmonton from 1500 MDT July 31 through August 2, and the total for these 2 days was roughly 100 mm. They also reported that an intensifying low pressure system, which remained in northern and central Alberta for 30 hours, brought an average of 33 mm of rain on July 31 to an enormous area northwest of Edmonton. This area, 277000 km2, is about 80 percent the size of Germany! Using the data from the 0600 MDT, July 31 radiosonde weather balloon released near Edmonton, Charlton et al. (1995) estimated precipitable water to be only 32 mm. That is, if all the water vapour in the atmosphere that day was condensed, it should yield only 32 mm of rain. Clearly the weather system was very efficient at producing widespread rain. Paruk (1988) reported that widespread flooding occurred in a region northwest of Edmonton. In this region, rainfall amounts for the period July 31 to August 2 exceeded 300 mm at 1 location and 200 mm at 2 other locations. Even if the tornado and swath of enormous hail had not occurred, the rainfall in Alberta from July 29, when the tropical air arrived, to August 4, when the low pressure system dissipated, would be regarded as exceptionally severe.
Charlton et al. (1995), utilizing lightning flash records, showed that the thunderstorm which eventually spawned the tornado moved northeastward from the Rocky Mountains, then gradually turned eastward when it was between 100 km and 200 km southwest of Edmonton, and it finally proceeded northward just as the tornado formed. Other storms traversing central Alberta that day had relatively straight paths moving from the southwest to the northeast. Equipment failure caused by lightning on July 30 prevented detailed weather-radar data from being recorded. Thus, the exceptional path of the tornadic storm could not be examined using radar imagery.
d. An introduction to the examination of hail damage using reshinglings
The billing records of the A Clark Shingle Company were examined to determine the distribution of reshinglings that this company performed after the tornado. Reshinglings completed in 1990, a year when no severe hailstorms struck Edmonton, were also inspected. The large area enclosed by the fine long-dashed line in Fig. 5 contains 84 percent of the reshinglings done between August 1, 1987 and July 31, 1988. The 2 small areas delineated by bold long-dashed lines contain 53 percent of the reshinglings completed in those 12 months. These 2 areas of south Edmonton were relatively new communities; the area west of 8 E is known as Kaskitayo, and the area to the east of 9 E is the western part of Mill Woods.
Figure 5 indicates that the 4 regions where flooded basements were common and the 2 areas where shingles replacements were frequent were clustered well away from each other and from the path of the tornado.
e. The distribution of disaster assistance
Frank Nesbitt, an employee of Alberta Public Safety Services, an agency of the Government of Alberta, provided a list of 259 businesses which suffered damage between July 25 and August 3 and applied to the provincial government for disaster assistance. Loans and grants totalling $20.8 million were provided to 214 successful business claimants. The agency dispensed federal and provincial funds to other groups as well: $4.6 million to 472 individuals, $2.6 million to 335 farmers, $4.7 million to 8 municipalities, and $3.6 million to 12 government departments (Alberta Public Safety Services, 1990). The Edmonton Journal (1996) reported that a lawsuit which delayed the completion of the assistance program was settled in 1996.
To document the effects of the storms which raked central Alberta on July 31, Charlton et al. (1995) mapped the distribution of disaster-assistance applicants from rural Alberta and of farmers who received crop insurance payments. The map showed a good relationship between the paths of lightning activity and damage in rural areas.
The list (Nesbitt, 1994) had the names and mailing addresses of the 259 business applicants. It did not differentiate between successful and unsuccessful claimants, nor did it note the amount of assistance that applicants were seeking. Some of the mailing addresses were likely not the locations at which damage occurred; more suitable addresses were sought in a 1987 telephone book or, if necessary, by telephoning the businesses, provided that they still existed in 1995. At least 152 of the 259 business applicants had assets within the fine short-dashed line in Fig. 5. This line forms a crude replica of the tornado path and encompasses an area which is nearly 3 km wide in places. Wallace (1987) concluded that the maximum width of "general" damage was 1.3 km. Of the 259 business claimants, 135 resided within the 5 small areas outlined by bold short dashes in Fig. 5. Eighty-five applicants with addresses in Greater Edmonton were located outside of the fine short-dashed line; presumably, their losses were not caused by the tornado. Claims were made by several transportation, holding, and property companies, and the locations of their damaged assets could not be readily ascertained. Firms with mailing addresses outside of Greater Edmonton were ignored, except if the locations within Greater Edmonton where these firms had assets were known.
Although some of Canada's most prominent industries were included on the list of disaster assistance applicants, the $20.8 million dispersed to all businesses by governments was small compared to $150 million lost to businesses by the insurance industry.
f. Air surveys
A set of 1:1500 colour air-survey photographic prints was obtained from Alberta Public Safety Services (Nesbitt, 1995). The photography was conducted on August 4, 1987 by Western Remote Sensing (now defunct). The firm did not give Alberta Public Safety Services the negatives from the survey, and the negatives have not been located. These excellent photographs were used to locate and identify nearly all of the industries damaged by the tornado, though the names of all of these industries are not documented in this study. The colour prints were donated to the Department of Earth and Atmospheric Sciences at the University of Alberta. The Edmonton Space and Science Centre made duplicates of some of the prints for a severe weather display shown in 1996.
Two sets of black and white stereo photographs from air surveys conducted on August 3 and August 6, 1987 by Global Remote Sensing were also examined. The negatives from the first flight remain with the firm, and those from the second flight, along with 1:2500 prints with overlays for use in public displays, are owned by the Public Works department of the City of Edmonton. The set of prints from the later flight included photographs of east Mill Woods and the undeveloped area southeast of Mill Woods, areas which are not depicted in the colour photographs. Several instances of significant house, farm, or tree damage, shown in Fig. 5, were found when these photographs were examined.
A set of colour slides was kept at the ALWC archives (Vickers, 1995). This set consists of 13 images of damage sites taken from an aircraft by Brian Smith for research conducted at the University of Chicago. Smith, now an employee of the National Weather Service in Valley, Nebraska, provided information about the photographs. In 1995, T.T. Fujita stated that all of Smith's photographs had been donated by the University of Chicago to the United States archives after Fujita retired. The limited-edition memoirs of Fujita (1992) included 4 images of tornadic damage in Edmonton selected from B. Smith's set; each photograph was described briefly.
Charlton's collection of video tapes of television newscasts also proved useful for developing Fig. 5, especially those segments taken from aircraft a few hours after the tornado dissipated. Unfortunately, aviators with the Canadian Forces, stationed at the Namao Airport, were not asked to conduct an aerial survey of the tornado path; presumably, the military could have flown on the evening of July 31 or on August 1, a day of inclement weather, and carefully photographed the damaged areas before much of the debris was moved.
g. Selected damage sites
The letters A through T, plotted in Fig. 5, represent an interesting selection of damage sites; these are described in Table 1. A variety of sources were used to develop the concise descriptions contained in Table 1: Charlton's extensive collection of newspaper articles; notes made during conversations with dozens of people who lived or worked in the vicinity of the tornado path; conversations with Scott Alexander and David Ungstad, structural engineers associated with the Department of Civil Engineering, University of Alberta, who examined damaged buildings in the industrial area; the 4 sets of aerial photographs discussed above; the responses from survey participants; and the reports from various agencies contained in Tornado: A Report (Alberta Public Safety Services, 1990). Information gained from a meeting with Greg Smith, an employee of the Occupational Health and Safety branch of the Department of Labour, Government of Alberta, was also helpful for developing the descriptions. He had experience with search and rescue procedures for mine disasters and, thus, was given the task of coordinating the search and rescue teams which rummaged through the collapsed buildings in the industrial area. He has an impressive set of photographs showing damaged buildings and debris-laden fields, as well as an excellent recollection of the damage he witnessed in the industrial area. The descriptions in Table 1 were developed prior to discovery of the study by Lux (1990). Lux's descriptions of sites E to R were in good agreement with those given in the table.
The descriptions for the sites where the Emergency Response Team or dangerous goods inspectors responded to reports of hazardous waste spills (sites D, F, G, J, K, N, and T) were, to a considerable degree, derived from a chapter in Tornado: A Report, but the locations of Tiger Chemicals and Pounder Emulsions shown in that report were incorrect; they have been plotted at their appropriate locations in Fig. 5. At site D, several Canadian National Railways (CNR) boxcars and tank cars derailed; at site S, 1 boxcar was toppled. Two employees of CNR provided detailed information about these sites: J. Albert, a yard supervisor, and W. Logozar, a locomotive engineer. Logozar was in the cab of the locomotive at site S. He submitted a completed survey in 1987 (Charlton, 1989).
Some of the events at the Archaeological Centre (site C) were documented in a letter which accompanied a returned survey (Charlton, 1989). To supplement this information, 2 employees responsible for the County of Strathcona's science park in 1994, the warden, Cliff Lacey, and a ranger, Ed Whitelock, were contacted.
The building enclosing the steel-rolling mill at Stelco Inc. ('I' in Fig. 5 and Table 1), a 300 m long structure, suffered substantial wind damage to its interior and exterior as the tornado passed its east end. The building had numerous ventilation louvers, and its east and west doors were open; consequently, the damage was not likely induced by static pressure loss. Apparently, Stelco suffered the greatest financial loss of any business: $12 million damage to equipment and a number of buildings (Webster, 1987).
A modest number of the badly damaged buildings in the industrial area are described in Table 1. (Lux (1990) described the damage to several other buildings in this area.) Annotated black and white photocopies of the colour aerial photographs of the industrial area were made. The name and address of nearly every damaged business shown was written on the copies. These copies were invaluable for quickly identifying buildings of interest and were borrowed by the Edmonton Space and Science Centre for developing their display about the Edmonton tornado.
h. Damage to houses and farm buildings and forest blowdowns
There are 4 locations where symbols for structural damage to wood-frame houses are plotted in Fig. 5: Clareview, where 16 houses lost roofs and some walls, and another house retained only a small section of 1 wall; Mill Woods, where approximately 10 houses were structurally damaged but did not need to be rebuilt; one rural area of Edmonton (12.8 E, 1.8 N), where 3 houses lost much of their roofs, and another (12.6 E, 0.8 N), where 1 house was structurally damaged. Many of the badly damaged houses were modern, two-storey buildings (Lux, 1990). No media reports about the structural damage at the 2 rural locations were found. On August 1, 1987, Charlton observed the damage to the 3 houses at 12.8 E, 1.8 N while he was exploring the region for evidence of tornado damage. The damage to the house at 12.6 E, 0.8 N was found while examining the air photographs kept by Public Works, City of Edmonton.
Aerial photographs were used to search for destroyed farm buildings; the destruction of a few of these buildings was documented in the newspapers and by Wallace (1987), but most sites shown in Fig. 5 were undocumented.
Forest blowdowns were also located using the aerial photographs. Most undeveloped land in Greater Edmonton has been cleared for farming, but some woodlots have survived. The blowdowns noted in Fig. 5 were levelled stands of trees; the destruction of an isolated pocket of trees was not considered to be a blowdown. The 2 blowdowns in the river valley, near sites A and B, had the feather-like fall pattern typical of a violent downburst of wind, rather than the swirl pattern expected of a suction vortex embedded within the tornado's violent circulation. Patterns of tree falls due to a long-lived tornado were discussed at length by Fujita (1989). He examined aerial photographs depicting the path carved through the Teton National Forest, Wyoming by a tornado which struck on July 21, 1987. Brian Smith explained that the short time between this storm and the Edmonton tornado, 10 days, was, to a large degree, the reason why researchers at the University of Chicago never completed their study of the Edmonton tornado.
i. Power line damage
In Tornado: A Report (Alberta Public Safety Services, 1990), TransAlta Utilities detailed the magnitude of the damage that they suffered. The tornado toppled 49 steel towers suspending 240 kV lines, 17 steel towers carrying 138 kV lines, and more than 200 wooden power poles. Rebuilding the network of power lines cost $8 million. Figure 5 shows the locations of 2 sections of downed lines: 12 km stretching south from a substation near the oil refineries (shown in Fig. 2) and 2 km positioned approximately 100 m east of the row of badly damaged houses in Clareview. On May 31, 1985, Ontario Hydro lost a total of 15 towers at 4 locations along the 117 km path of the Grand Valley-Tottenham tornado in southern Ontario (Gorski, 1985). On 3 occasions in the previous 7 years, tornadoes had destroyed 29, 15, and 7 of Ontario Hydro's towers.
Edmonton Power's main generating station, known as the Clover Bar station and labelled 'P' in Fig. 2, is located in the river valley, approximately 2.5 km south of TransAlta's 2 km section of downed 240 kV transmission towers and 200 m from the path of the tornado. It also lies directly across the river from the outlet of the Kennendale Ravine storm sewer ('B' in Fig. 5). The station is connected to TransAlta's network of 240 kV lines, but the power sent from this site to the residents of Edmonton was unaffected by the loss of these lines. Its power was carried by 72 kV lines which cross the North Saskatchewan River in a tunnel beneath the river. John Mulka, the manager of the station at that time, stated that the plant temporarily lost the use of 1 of its 2 operating generators after debris struck an insulator. Repairs to the plant cost approximately $50000. The document submitted by Edmonton Power to Tornado: A Report does not mention the Clover Bar generating station, but it does describe the company's efforts to restore power to the badly damaged regions of the city, including the devastated Evergreen Mobile Home Park. Few residential areas lost power for a significant period. Thus, most residents with AC- powered radios could receive bulletins as the tornado passed through the region.
j. Other tornadoes in greater Edmonton
The tracks of 2 small but destructive "spin-off" tornadoes are shown in Fig. 5. Both of these tornadoes occurred while the main tornado travelled 1 to 2 km to the west. Their existence was not reported until 1995. The southernmost one was mentioned by Wilf Seutter during a conversation about damage caused by the main tornado. Seutter's family watched this small tornado move northward from their acreage, near the start of its path, after it had flattened some trees in their woodlot. Discussions with Seutter prompted a visit to the nearby farm of Stanley Stannard who had 30 m of his barn roof (14.0 E, 6.7 N) blown from over his head. His wife had been watching the main tornado from their house and telephoned him just before the barn roof was removed. The roof came to rest against their house.
The other spin-off tornado apparently formed at 13.1 E, 7.4 N, immediately west of the Maple Ridge Mobile Home Park. Its approach was observed by Rick Scott of Scott Steel Ltd (13.4 E, 7.7 N). This tornado destroyed the company's 45 m by 15 m truss roof and apparently caused a 200 tonne car-ferry deck under construction to hover; it continued moving northward across an open field and then damaged the roofs of Guardian (then Hyalog) Oilfield Services (13.3 E, 8.0 N) and Capital Industrial Sales (13.3 E, 8.2 N). Shortly thereafter, it lifted and passed over an undeveloped region, turned westward, and finally dissipated near the main tornado. The latter stages were witnessed by employees of Blanchard Transport (13.9 E, 8.3 N) and the Shell-Sherwood oil marketing terminal (13.7 E, 9.1 N). This tornado may have caused the local centre of disaster assistance claims (13.3 E, 8.0 N) which lies east of the main tornado path (Fig. 5). This area is known as Elmjay Business Park.