Chapter 5
5. THE GREATEST HAILSTORM
Charlton et al. (1995) argued that the hail which accompanied the Edmonton tornado was unprecedented in Canada, and that the swath of giant hail they described was better documented than hail swaths from any other urban storm. Giant hailstones of various shapes and opacities, collected and photographed by M. Madsen in south-central Mill Woods (10.3 E, 3.0 N), are shown in Fig. 11.
a. Areas by hail-size category and comparisons with the 1991 Calgary storm
Figure 12 is an updated map showing the size category of the largest hailstone reported by each participant who completed the Hail Report in the survey. The categories, plotted as numbers 1 through 7, were denoted by the names of common objects. The minimum accepted dimensions for the categories were established by Charlton et al. (1995): 1 (Shot), 0.2 cm; 2 (Pea), 0.5 cm; 3 (Grape), 1.2 cm; 4 (Walnut), 2.1 cm; 5 (Golfball), 3.3 cm; 6 (Tennis ball), 5.2 cm; 7 (Larger), 7.8 cm. The diameter of a standard golfball is 4.4 cm, for a tennis ball, 6.4 cm. Of the 755 respondents within the area of the maps, 638 provided a maximum hail-size category. Shown in Fig. 12, the boundaries of walnut-, golfball-, and tennis ball-sized hail enclose local areas in which at least one half of all hailstone reports were for those categories or larger categories; for example, at least 50% of the digits plotted within all sectors of the tennis ball boundary are '6' or '7'.
The areas of the regions within the boundaries and the areas of residential housing (see Fig. 2) in these regions were determined: for walnut, 92 km2 including 36 km2 of residential area; for golfball, 53 and 18 km2; for tennis ball, 125 and 57 km2. The total area of large hail, that is, walnut-sized or larger, was 270 km2. Of this, residential zones occupied 111 km2. In 1987, the residential area of Greater Edmonton (Edmonton, St. Albert, and Sherwood Park) was 169 km2. Postal delivery information for 1987 (Canada Post, 1987) indicated that there were 154000 houses in Edmonton, and 174000 in Greater Edmonton. (The population of Greater Edmonton was approximately 3% smaller than that of Metro Edmonton, as defined by Statistics Canada.) In this study, the definition of a house includes duplexes and row houses. Alan Wood of the Insurance Bureau of Canada estimated that there were 32000 successful insurance claims made by holders of homeowner policies; and at least five sixths of these claims (27000) were for hail damage. Assuming that the size of a residential area was proportional to the number of houses in it, 66% of all houses in Greater Edmonton lay within the area of large hail, and 44% lay within the area of giant hail, that is, golfball or larger. The area of tennis ball or larger encompassed 34% of all houses. Yet, at the most, just 18% of all Greater Edmonton homeowners (32000 of 174000) made a successful insurance claim.
Policy holders made successful claims for damage caused by the September 7, 1991 hailstorm to 60000 of the 217000 (28%) houses in Calgary (1991 population 714000, 51o 04' N, 114o 04'W, elevation 1080 m). To augment the preliminary claims information provided by the Insurance Bureau of Canada, Charlton et al. (1995) surveyed Calgary residents requesting that participants report their maximum hail sizes and how many of the 4 houses nearest to their houses needed reshinglings. Forty-five of the 60 participants resided within the 130 km2 damage area, which was later delineated by 2 insurance adjustors who spent a full year assessing claims for damage caused by the storm. Just 3 of these 45 reported tennis ball-sized hailstones, and only 1 reported larger than tennis ball. Of the other 41 in the damage area, 22 reported golfball, 16 reported walnut, and 3 reported pea or grape. The discrepancy between insurance claim rates for the 1991 Calgary storm, where tennis ball- sized hail was a rarity, and the 1987 Edmonton storm, where it was common, will be discussed in Section 7.
The times when hail commenced in Greater Edmonton, as recorded by respondents, were plotted on a map (not included). Hail from the tornadic storm fell on virtually all of east and central Edmonton sometime between 1500 and 1600 MDT. Some regions of west Edmonton were also hit by hail during this period. When a second thunderstorm traversed Greater Edmonton between 1730 and 1800 MDT, St. Albert, west Edmonton, and central Edmonton were buffeted by hail. The funnel cloud sightings associated with this second thunderstorm were discussed in Section 4, and the wind damage attributed to this storm will be revealed in Section 6.
b. Measurements of hailstones
The survey form requested that participants measure their largest hailstones. The largest measurement given by each of the 236 respondents who answered the request are displayed in Fig. 13 as a circle of proportional diameter. Many participants from Mill Woods claimed to measure hailstones with diameters greater than 10 cm.
Hail samples, most consisting of several hailstones, were collected from 63 participants; nearly all of the samples were collected within 3 weeks of the storms. Two exceptional samples were collected after a belated public appeal by Charlton was published by the Edmonton Sun on March 25, 1988. One sample included the Tews hailstone (264 g) which set the Alberta record for hailstone mass, and the other included the Smegal hailstone, which was only a fraction of a gram lighter (a record hailstone is usually named after the person who originally collected it). These 2 hailstones had fallen about 10 km apart and were in good condition when collected from their owners on the same day in the spring of 1988. Tews' experiences while collecting the record hailstone were documented in the Edmonton Journal (Retson, 1994). The previous Alberta record, the Wilson stone (249 g), and the present Canadian record, the Gawel stone (290 g), found near Cedoux, Saskatchewan in 1973, were discussed by Wojtiw and Lozowski (1975).
The maximum dimension of the largest hailstone in each of 62 of the 63 samples is plotted in Fig. 14 as a circle of proportional diameter. One measurement was lost and could not be retaken because the freezer used to store the samples failed and some melting occurred. Time and travel constraints, as well as the modest size of hailstones reported by some respondents, restricted the collection of samples to, primarily, south Edmonton. If all 240 available samples had been retrieved, this collection would still be smaller than the more than 300 samples retrieved from Edmonton by the Alberta Hail Project after the August 4, 1969 hailstorm (Rogers and Summers, 1971). The largest hailstone collected at that time was only 104 g (Rogers, 1970). The world's most costly hailstorm occurred in Munich on July 12, 1984, but scientists retrieved hail samples from only 10 locations (Binder, 1985). The literature documenting these urban hailstorms was reviewed in Charlton et al. (1995).
The table in Fig. 14 shows the masses and the 3 orthogonal dimensions of the 5 heaviest hailstones and those of a large but desiccated hailstone. The participant who collected the desiccated hailstone, labelled '?' in Fig. 14, reported its maximum diameter to be 17.8 cm; but its maximum dimension, when collected by a researcher more than a year later, was 12.7 cm. This hailstone was originally stored in a self-defrost freezer; it subsequently crumbled when it was examined in the laboratory. Its mass, measured at the laboratory, was less than 200 g. Hailstones stored in most deep freezes did not seem to experience similar decreases in mass and structural integrity. For example, 2 well preserved samples of hailstones that fell on July 31, 1987 were collected in the summer of 1994; these 2 samples are not documented in this study. Examination of the 5 heaviest hailstones, numbered 1 to 5 in Fig. 13, indicated that all were made of solid ice when they fell, that is, none of the 5 were "soft" or "spongy".
The calculated masses of perfect, solid-ice ellipsoids with orthogonal dimensions identical to those of the 5 largest hailstones are shown in the table in Fig. 14. The deficit between a measured mass and its ellipsoidal equivalent suggests the degree to which the hailstone deviates in shape from a simple ellipse. If the largest dimension of a hailstone included a spike or a pronounced lobe, as some of the hailstones shown in Fig. 11 do, the mass of the hailstone would be considerably less than the ellipsoidal mass. None of the 5 heaviest hailstones had identical measured and ellipsoidal masses, but only the mass of hailstone '4' differed radically from its ellipsoidal mass (Fig. 14).
The growth of a hailstone is highly dependent upon its fall speed which is, for a given mass, dependent upon its shape (Charlton and List, 1972; Knight and Knight, 1970). The largest hailstones from the collected samples were, in general, quite smooth. This observation may be useful to researchers testing their numerical cloud models by comparing their results with the hailstones which fell on Edmonton. Such a test has been proposed by H. D. Orville of the Institute of Atmospheric Sciences, South Dakota School of Mines and Technology, Rapid City.
c. Comparison of respondent measurements with laboratory measurements
Fifty of the 63 respondents who provided samples reported measurements of their largest hailstones. They measured their hailstones without assistance from the researchers who subsequently collected their samples. The researchers assumed that the largest hailstones in the samples were the stones measured by the participants and that the participants measured the maximum dimensions. Charlton et al. (1995) included a scatter diagram which compared their measurements with those taken by the participants. One half of the measurements of maximum dimensions by participants were more than 20% larger than those taken in the laboratory. Apparently, many rounded their measurements up to the nearest 0.5 inch (1.27 cm). When that amount was subtracted, agreement with the laboratory measurements was much better.
These 50 respondents, however, were more accurate in categorizing their largest hailstones. Nineteen reported larger-than- tennis-ball hailstones, but 10 had collections which contained no hailstones large enough for that category (7.8 cm); only 1, however, failed to submit a hailstone with a maximum dimension larger than the diameter of an actual tennis ball (6.4 cm). Thus, all but 1 of these 19 were correct. Twenty-three reported tennis ball-sized hailstones; just 3 supplied hailstones smaller than the minimum for that category (5.2 cm). Golfball-sized hailstones were recorded by 6 participants; all 6 were larger than 3.3 cm, the minimum for the category, but one was larger than 7.8 cm, the minimum for larger-than-tennis-ball. None of these 50 participants claimed that their largest hailstone was walnut sized, the minimum size for large hail. Two of these 50 respondents failed to categorize their largest hailstones, although they did provide a measurement. Thus, 43 of the 48 respondents (90%) who categorized their hailstones did so correctly. Such accuracy, however, may not be achieved when citizens categorize small hail (shot, pea, or grape). Many meteorologists claim that the public usually overestimates the size of hailstones, but no other study comparing measurements or size categorizations made by the public with measurements taken in a laboratory was found. The Archive Report (Charlton et al., 1989) included all measurements taken by the respondents and the researchers.