Tornadoes on the Canadian Prairies: 1826-1939 — Part 1, Project Overview

-By Patrick McCarthy and Jay Anderson-


Our project was to develop the most comprehensive and accurate database of Canadian prairie tornadoes. We have completed the period from 1826 to 1939. How this was accomplished is presented here.

Previous efforts

The systematic collection of Canadian prairie tornado events began with A.B. Lowe and G.A. McKay (1960). Brad Shannon (1976) extended that database to 1973. During the 1970s, the Prairie Weather Centre (PrWC) in Winnipeg established the Meteorological Service of Canada’s (MSC) first summer severe weather program. It was supported by the Regina Weather Office and was expanded to the Alberta Weather Centre. For verification and case study purposes, severe weather events, including tornadoes, were collected and archived for the three Prairie Provinces. These reports were often very incomplete, relying on informal contacts and phone calls to weather observers. At the time it was thought that sightings of only about 1/3rd of the tornadoes reached the weather office. In collaboration with the Ontario Weather Centre, damage surveys became a part of the PrWC records. Weather spotters were added to improve event detection and description. While the collection of annual events was ongoing, the archiving of historic occurrences remained very limited.

In the early 1980s, Dr. Keith Hage, at the University of Alberta, began an extensive study to identify historic Alberta and Saskatchewan tornado and windstorm occurrences. Michael Newark (1984) developed the first Canada-wide tornado database. His work would eventually include events as far back as the 18th century. Thomas Grazulis (1993) published an extensive American tornado dataset that included a few events that straddled the U.S.—Canada border. Hage (1994) published a second database for historic Alberta tornadoes, windstorms, and lightning fatalities. Dr. Alexander Paul (1995), at the University of Regina, produced a Saskatchewan tornado chronology from 1906 to 1991 (690 tornadoes). Grazulis (2000) published a compilation of Canadian “killer” tornadoes. Hage (2001) published his compilation of Saskatchewan tornadoes (720), windstorms, and lightning fatalities from 1880 to 1984.

While employed at the MSC office in Winnipeg, we began consolidating and correcting these sources, beginning in the early 1990s and continuing to the present. This effort formed the basis of the current MSC Prairie Severe Weather archive, and which is updated annually. Eventually, a formal version was made available to the public and researchers (McCarthy, 2010). The database contained roughly 19,000 severe convective events, including 3100 reported tornadoes dating back to 1826.

Figure 1. Part of the tornado damage associated with the June 30, 1912, Regina F4 tornado. (Public domain: image provided by the City of Regina Archives – CORA-RPL-A-0905)


We noted many problems in the 2010 database. These included vague reports, data gaps, unverified assumptions, missing and inaccurate dates and times, possible duplications, and conflicting death and injury totals. We undertook a project to review and update all the tornado reports in the archive and to add previously undiscovered events to the record.

We worked independently to take advantage of our unique but overlapping research approaches. In 2019, our two datasets were merged and differences in our records were resolved. This 1826-1939 portion of the review took about five years to complete. A small amount of new information has since been added; something that will likely occur on an ongoing basis.

The data-collection phase of the project had many challenges. Prior to 1940, there were no damage surveys, no video evidence, damage and tornado photographs were rare, eyewitness accounts were limited, and construction practices were varied. We are both experienced storm-damage surveyors and we used our experience to assess damage reports, leading to a subjective damage rating when sufficient evidence was available.

A multitude of sources, mostly unofficial and difficult to unearth, were used in the research. These included digital databases, particularly newspapers, books, historic school sources, community/rural municipality/county histories, historical societies, and obituaries. Most events appeared in newspapers, though there was typically little follow-up after the initial reports. Downed telephone and telegraph lines and poor roads often delayed the reports, leading to inaccurate dates in the record. Delayed information seldom made it into the newspapers, allowing initially bad information to linger in articles for weeks. To account for these issues, we searched for descriptions of events for up to a month after their occurrence, and in a few instances, much later.

Figure 2. This is photo of a tornado near Vulcan, Alberta on July 8, 1927; one of a number of tornadoes that day. [Public domain, photographer: McDermid Photo Laboratories, Calgary]
Eyewitnesses have an expansive view across the open prairie. Tornadoes could be seen from long distances and reports often resulted in misplaced locations. The “highways” of the prairie were the railroads. Railway station names were often referenced, further misplacing the tornado’s position. Many stations and some communities have disappeared over time. We controlled many of these difficulties by triangulating available reports to narrow the location error. On days with multiple events, the information was plotted using eyewitness accounts, storm information, and damage characteristics, to identify storm damage swaths and potential tornado tracks. The numerical NOAA-CIRES 20th Century Reanalysis (V2c) (, displayed via the ( website, was also used to help assess storm types, potential storm motion, and storm potential. To validate a tornado occurrence, the authors used a decision-tree approach, similar to Sills, et al (2003).

A hybrid version of the Fujita ( and Enhanced-Fujita Scales ( was used to rate tornado damage. The EFScale is too modern for most early events, which lack the necessary details in their descriptions. The FScale rating, which allows for a broader interpretation, was primarily used for the database. For events that crossed the U.S. – Canada border, the rating is for damage on the Canadian side, only.

The 1826-1939 timeframe was a period of immigration, spreading from east to west across the Northern Plains. There was a slow growth in population, communities, roads, railways, newspapers, etc., which affected the availability of data. For example, most cemeteries did not appear until after 1910 in Saskatchewan and Alberta.

One of our major goals was to provide a more complete account of the number of deaths and injuries. To make accurate casualty totals, a rigorous effort was made to identify the names of victims. This included victims who may have succumbed to the injuries well after the event. We used cemetery records, obituaries, ancestry/genealogy records, church records, provincial/state censuses, and death records to track down these victims.

Our effort has culminated in the most complete and descriptive archive of Prairie tornadoes and a record for researchers to build on. Some of our findings from this early compilation are presented in Part 2 of this series.

Figure 3. Tornado near Oak Lake and Kenton, Manitoba, July 25, 1909. [Public Domain: photo by A. L. Cameron].


Grazulis, T.P., 1993. Significant Tornadoes, 1880-1991. Volume II: A Chronology of Events. Environmental Films, St. Johnsbury, VT. cad=0#v=onepage&q&f=false

Grazulis, T.P., 2000. Canadian Killer Tornadoes. The Tornado Project, St. Johnsbury, VT.

Hage, K.D., 1994. Alberta Tornadoes, Other Destructive Windstorms and Lightning Fatalities: 1879-1984. Self-published, Spruce Grove, Alberta, Canada.

Kendrew, Wilfrid George, and Balfour W. Currie, 1955. The climate of central Canada– Manitoba, Saskatchewan, Alberta and the Districts of Mackenzie and Keewatin. E. Cloutier, Queen’s Printer, Ottawa, 152.

McCarthy, P. (ed), 2011. Prairie and Northern Region 1826-2010 Severe Weather Event Database. Prairie and Arctic Storm Prediction Centre, Meteorological Service of Canada, Environment Canada, CD-ROM CD_095.

McKay, G.A., A.B. Lowe, 1960. The tornadoes of western Canada. Bull. Amer. Meteor. Soc., 41, No. 1, pp. 1-8.

Newark, M. J., 1984: Canadian Tornadoes, 1950- 1979. Atmos.-Ocean, 22, 343-353.

Paul, A.H., 1995. The Saskatchewan Tornado Project. University of Regina, Department of Geography Internal Report.

Raddatz, R.L., R.R. Tortorelli, M.J. Newark, 1983. Manitoba and Saskatchewan Tornado days 1960 to 1982. Environment Canada, Atmospheric Environment Service, Canada Climate Centre, Downsview, ON, CLI-683, 57 pp.

Raddatz, R.L., J.M. Hanesiak, 1991. Climatology of Tornado Days 1960-1989 for Manitoba and Saskatchewan. Climatological Bulletin. 1991, 25, Num 1, pp. 47-59.

Shannon, T.B., 1976: Manitoba Tornadoes 1960-1973. Atmospheric Environment Service, Meteorological Applications Branch,

Environment Canada, Project Report No. 29.

Sills, D.M.L., Scriver, S.J., and King, P.W.S. 2004. The tornadoes in Ontario project (TOP). Preprints, 22nd AMS Conference on Severe Local Storms, Hyannis, Mass., American Meteorological Society, CD-ROM 7B.5.

Patrick McCarthy is a retired Meteorological Service of Canada meteorologist and former Head of the Prairie and Arctic Storm Prediction Centre. He is an active weather history buff, storm chaser, and the Chair of the CMOS Winnipeg Centre.

Jay Anderson is a meteorologist, formerly with Environment Canada, where he worked primarily in Winnipeg and Vancouver over a 34-year career. Since his retirement, he has been working casually as a consultant, primarily in the travel industry, and teaching storm chasing at the University of Manitoba

Canadian Prairies, Jay anderson, Patrick McCarthy, tornado

For previous issues go to the Archives page on the main CMOS site.

© 2017 Canadian Meteorological and Oceanographic Society

Website Designed & Maintained by Aliado Marketing Group.