Chapter 2
2. THE NEWSPAPER SURVEY
While assisting employees of the Alberta Weather Centre (ALWC) with their field investigation of the tornado's damage track (Wallace, 1987), Charlton and Wojtiw decided that a public survey, distributed in the daily newspapers, was warranted. On August 3, financing for the survey was obtained from the Central Research Fund of the University of Alberta and the City of Edmonton. Financial arrangements for publishing the full page advertisement in both newspapers were managed by the city. The survey form, published Saturday, August 8 by the Edmonton Journal and Sunday, August 9 by the Edmonton Sun, is shown in Fig. 1a and Fig. 1b. Most of the questions in the survey were taken from a form used by the Volunteer Weather Watchers (VWW), a group that is organized by the weather offices of Environment Canada to report severe-weather events every summer. The form published by the newspapers, however, also had to elicit information from the public about the efficacy of Environment Canada's weather warning system. It also included questions about the availability of frozen hail samples, photographs, and video tapes. Interpretation of some returned survey forms proved to be challenging; these problems are discussed in later sections, and a modified survey form, one that should avoid such difficulties, is available from Charlton.
Figure 2 is a map showing selected Edmonton roads, residential areas, and the tornado path. The units of the coordinate grid are miles because most main roads are 1 mile apart. The legal land location of coordinates 0 east and 0 north is the southwest corner of S 18, T 51, R 25, W4. There are no "correction lines" within the area of the map; therefore, the grid is representative of the legal survey lines. This coordinate system is shown on all the maps presented in this study. The positions of important sites which are not denoted on the maps will often be described by their coordinates.
Of the 815 survey participants, 755 of them were present within the boundaries of Fig. 2 during the afternoon or evening of July 31. The phenomena experienced by each of the 755 respondents were prioritized such that encountering the tornado was given highest priority, and not experiencing severe weather of any kind was given lowest priority. The location of each respondent at the time of his highest priority experience was converted into coordinates. The Archive Report (Charlton et al., 1989) contained tabulated data organized by the respondents' positions within the grid. The tabulated data, available in hard copy and computer diskette (Apple and PC formats), consisted of all parameters derived from the newspaper survey. Analysis of this data, particularly the mapping of it, is the principal purpose of this study. Copies of both the Archive Report and the survey responses were given to the archives at the University of Alberta and the Provincial Museum of Alberta. Respondent's names and addresses have been removed from the forms housed at the Museum. Forms held at the University archives are available only to bona fide researchers.
The path of the tornado, shown in the maps, was based on the damage survey conducted by local meteorologists (Wallace, 1987), including Charlton and Wojtiw. Arrows represent sections of the path where "considerable or greater" damage occurred, and dashed lines indicate regions of only "general" damage. The tornado path is included in Fig. 2 to show its proximity to roads, neighbourhoods, and 4 important industries in east Edmonton: the tornado travelled between 2 operating oil refineries, and it came dangerously close to Edmonton's largest chemical plant, an Edmonton Power thermal-electric generating station, and the Maple Ridge Mobile Home park! Clearly, numerous near- tragedies were barely avoided, a fact that few authors of previous studies have noted. Figure 3 is 1 of a series of the only known set of photographs taken from north of the river in which the tornado was clearly depicted. It shows the tornado passing to the east of Imperial Oil's Strathcona refinery. At that position, the tornado was approximately 2 km east-northeast of the ALWC. An abridged tabulation of facilities struck by the tornado will be included in Section 3.
The dots and circles in Fig. 4 show the locations of the 755 survey respondents within Greater Edmonton. These clearly indicate that reports were received from virtually all residential areas (shown by shading), and that workers in the industrial area, where the tornadic winds apparently reached their greatest speed, were especially generous in reporting their experiences. Instances of sewer flooding, denoted by the circles, will be discussed in Section 3. Figure 4 also shows that the tornado was outside the city's boundary as it moved from 8.2 N to 12.0 N, an area of heavy industries in Strathcona County. As stated earlier in this section, the dots and circles plotted on the map are placed where the respondents encountered their most-severe-weather experiences.
The original survey numbers were plotted by hand at the participants' locations on a 104 cm by 122 cm (1:30000) base map. To prevent overlapping, some of the respondents' positions were shifted by a small amount; consequently, some of the coordinates listed by Charlton et al. (1989) were not perfect representations of the participants' locations. Altering the coordinates, however, made the maps less prone to overlapping data. The errors introduced by the shifting of coordinates are easily seen in 2 areas of Fig. 4: a number of respondents appear to be located in the farm fields to the east of Mill Woods or directly on the river.