By Joshua MacFadyen and Nolan Kressin*, University of Prince Edward Island
For the full lesson on the historical GIS methods and tools described in this post, see the new tutorial in the Geospatial Historian Methods of Visualizing Temporal Data, by Nolan Kressin and Joshua MacFadyen.
The movement of commodities has been an important study within Canadian scholarship since Harold Innis wrote The Fur Trade in Canada (1930), but many forget that in his earlier thesis on the Canadian Pacific Railway (CPR) Innis also focused on the goods that these lines carried to market each year. By examining the records of the CPR, Innis painstakingly summarized the freight capacity and the market conditions that shaped this chapter in Canadian environmental history. Innis was well-known as a “dirt” researcher, digging into archival collections and often cutting-and-pasting his notes across manuscripts (literally, with scissors) to organize the enormous amounts of information he collected. As historians turn to more focused studies of individual commodities, we can parse large historical datasets with tools beyond scissors and glue. In this piece we discuss new ways to take some of the same historical railway data and focus on resources like firewood and lumber in their natural environments. In theory, we could even use the railway locations to focus on forest types (softwood) or even species. Perhaps if he had these digital tools, Innis’s first book after the CPR study might have been called the “Fir Trade in Canada.”
Scholars like Innis might well have imagined a time when computerized systems could assist in the collection and visualization of historical records, but I’m not sure they could have foreseen tools like Geographic Information Systems (GIS) and the ability to create animated maps of railways and the goods they carried between rural and urban environments. In this article, we explain the process we followed to create just such an animated map of a firewood as it was hauled around by not just the CPR, but all railways in Canada. We needed a map that would help shed light on both the geographic distribution of biomass energy production and consumption and the temporal trends during a period of intensive economic growth (and hence energy expansion). We know that seems like a rather specific research question, but we argue that others could follow a similar process to map the movements of any commodity or passenger traffic on historical railways or any other routes (eg road or water transport) over time.
The method we cover in the tutorial is a relatively new feature in Geographic Information Systems known as time mapping, or temporal GIS. GIS was designed to visualize and analyse spatial data, first and foremost. Only in recent years have developers focused on the software’s capacity to represent the temporal attributes that accompany many forms of geographic information. Now there are many striking examples of visualizations that can map everything from the patterns of wind to the shipment of goods, and the movement of migrants. The humanities are close behind with time maps of US population density (1790-2010), trans-Atlantic slavery, and nuclear detonations since 1945. Military historians have created time maps of conflicts from stylized animations of US Civil War battlefields to detailed sequences of Canadian movements at the Battle of Vimy Ridge. For a large collection of projects containing some element of animation see Mapping History from the University of Oregon.
As ESRI’s guide to “Working with Temporal Data in ArcGIS” explains, there are at least four types of temporal GIS data. Moving features are those that move over space, like a tornado or hurricane where the centre is often in a different location every hour. Discrete events can contain data about features that usually happen in a single place and time, like a car accident. Stationary features contain data about a single place over time, such as a sensor, and change/growth features represent a feature that changes in size, such as a forest fire, or in the case of this exercise, railways.
Like many aspects of GIS, its temporal capabilities often exceed what historians need. Any of those data types above can use calendar or non-calendar time scales up to fractions of a second apart. Historians often have much coarser temporal details for our data (months or years), or we want to map features in decades or even centuries. Moreover, we are often missing or uncertain of key dates. GIS lacks the ability to incorporate most of these uncertainties, and cartographers often have to get creative when mapping historical features. In this example of historical railways, the data were reported inconsistently in both time and space. We were missing key dates in the 1890s because officials reported each company’s freight data in an aggregated form. We were also missing some spaces because certain railways existed but did not report their freight data accurately, or at all, in some years. Finally, historians routinely work with a sample of the total available data, particularly when extensive digitization, data entry, and development are necessary. Depending on their objectives, scholars can uncover and demonstrate geographic patterns without entering every known historical observation. In the case of Canadian railways, we decided to work with sample years in the period (1876-1921), selecting dates that we knew were relatively complete and that illustrated the extent of wood shipments in key periods.
Digital humanities scholars use a variety of tools to help visualize historical change in a GIS. In the early days of PowerPoint, historians might have created a series of maps with an identical scale and extent and then shown them in sequence (or even as an animation) in order to help audiences quickly recognize spatial patterns that changed over time. Increasingly, scholars now use video and gif photo formats to create a similar effect in a file format that is easier to share online or outside of a lecture setting. GIS companies have made it easier to share series of static map images online or map sequences in a web interface so that they appear to change at the user’s control. Sean Kheraj used ESRI Story Maps and YouTube video animations to show the expansion of pump stations, mainlines, and capacity in Canadian oil pipelines between 1950-1979. These visualizations were part of his research project “Silent Rivers of Oil: A History of Oil Pipelines in Canada since 1947,” available in several talks and publications on this NiCHE project page.
Two of the main GIS developers (ESRI and QGIS) now offer tools that will recognize calendar or non-calendar time attributes in your data and then display them in a sequence using a time slider control feature. These tools have both descriptive and analytical applications, and we would argue that railway maps are a great candidate for both. Similar to Kheraj’s pipeline map, ESRI Canada has created a web map (and accompanying lesson) on ArcGIS Online. The result shows some of the strengths of a time-enabled map of historical transportation systems like the Canadian railways, and readers who simply want to see a basic map of railway expansion in Canada, should certainly check it out. However, for those who want to map and visualize changes in historical data (such as firewood or other freight), keep reading and consider following through to our full Geospatial Historian lesson on temporal GIS.
Temporal GIS also allows historians to conduct analytical exercises based on the temporal attributes of their historical data. In many cases, it would be nearly impossible to answer some of these questions without a GIS. For example, consider a query like “how often did icebergs come within 10 Km of Twillingate Lighthouse in Newfoundland in May of each year, and did those increase or decrease with changes in global and regional climate?” To answer this question you would need something like the iceberg sightings database (c1800-1959) with multiple years of iceberg data including both their precise location at time points that spanned the longer iceberg sighting season. The GIS could then identify the location of each point, measure their distances to — and select only those points within 10 Km of — Twillingate Lighthouse, and then return only those with “May” in their temporal attribute data. Temporal mapping allows historians to consider a range of geospatial questions of their sources, but it’s also great for demonstrating patterns through basic data visualization.
In this exercise we performed a simpler form of temporal analysis (freight hauled per distance on each line). The results allowed us to focus our historical research on railways and regions that “stood out” on the map. Some of the leading firewood lines, in both real and relative amounts, were narrow gauge railways like the Toronto, Grey & Bruce. These were designed specifically to haul firewood to the growing energy markets in Ontario’s cities, so it’s no surprise that they appeared as bright red veins connecting Toronto to its hinterland. By comparing firewood to other commodities hauled on each line we see that the narrow gauge companies dedicated up to 50 percent of their freight to firewood. Others like the Whitby and Port Perry were not known as fuel trains, but they still stand out on the relative map and they actually dedicated increasing amounts to firewood in the early years (rising from 8 to 12 percent). As firewood supplies decreased and coal imports increased in Southwestern Ontario, the concentrations around Toronto disappeared from the temporal map. However, the data visualization revealed some surprising concentrations on short-haul lines connecting Eastern Ontario and Quebec cities with regions in the Canadian Shield. These will be explored in a forthcoming article.
The research required three general categories of sources, historical railway/freight statistics, historical railway GIS data, and general railway histories. We found most of the historical statistics in the Sessional Papers. The annual reports of the Canadian Minister of Railways and Canals usually contained a “Summary Statement of Description of Freight Carried” by railways in the previous year. These tables were also reproduced in the Canada Year Book, and in the later years (1900s), these were the only historical sources available to us. The historical railway GIS data, a collection of line segments representing railways built between 1836-1922, were supplied by Byron Moldofsky of the Canadian Historical GIS Partnership. However, even these require data development, technical knowledge, and collaboration to use them to their full potential. Finally, general railway histories were required to provide historical context and to fill some of the gaps that appeared when trying to match the historical statistics to the historical GIS data. As we mentioned, the historical railway/freight statistics data had many limits. This requires extensive research in railway histories and other secondary sources such as Christopher Andreae’s Lines of Country (1997). In the cases where no match could be found, we had to decide whether the railways existed but failed to report their data, or whether the historical GIS files were incorrect. This is explained in the difference between “grey and white” lines in the tutorial.
In the end this 15-second animation answered some questions and raised many others. Most interesting to us were the ways we could quickly confirm the importance of firewood lines like Toronto Grey and Bruce. However, we were surprised to see those diminish so quickly in the 1880s. We were also intrigued by how the maps identified other Toronto-area wood haulers like the Whitby and Port Perry Railway. Finally, as the next chapters of this research will explore, these maps helped reveal how new lines like J.R. Booth’s “Canada Atlantic Railway” to Ottawa, New York State, and New England created entirely new firewood supply chains in areas that would otherwise have experienced a more rapid transition from wood to coal.
*Nolan Kressin is a research assistant with the UPEI GeoREACH Lab and an undergraduate student in the Applied Climate Change and Adaptation program at the University of Prince Edward Island.
Feature Image: Cords of Firewood Hauled on Canadian Railways, 1878. Source: Sessional Papers of Canada. Map by the Authors.
Andreae, Christopher. Lines of Country: An Atlas of Railway and Waterway History in Canada. Erin, Ontario: Boston Mills Press, 1997.
Churcher, Colin. “Fuelled by Wood.” Branchline (May 2007), pp 6-9. See also Churcher, Colin. “Fuelled by Wood” https://churcher.crcml.org/Articles/Article2007_04.html Accessed 7 October 2020.
Innis, Harold. A History of the Canadian Pacific Railway. London: P.S. King, 1923. E-book on Gutenberg https://gutenberg.ca/ebooks/innis-historyofthecpr/innis-historyofthecpr-00-h.html Accessed 7 October 2020.
Innis, Harold. The Fur Trade in Canada: An Introduction to Canadian Economic History. New Haven: Yale University Press, 1930.
Institute for Ocean Technology, “Ice Data Project: Iceberg Sighting Database,” http://www.icedata.ca/iceberg-sightings/ Accessed 7 October 2020.
Toronto Railway Historical Association, “Port Perry Station.” http://www.trha.ca/trha/history/stations/port-perry-station/ Accessed 7 October 2020.
Latest posts by Josh MacFadyen (see all)
- The Grass Roots of a PEI Potato Farm - May 9, 2022
- The Stubborn Commuter - November 3, 2021
- Post-Doctoral Fellowship – Canada Research Chair GeoREACH Lab – UPEI - April 21, 2021
- The Fir Trade in Canada: Mapping Commodity Flows on Railways - October 8, 2020
- Other Plans: Development and Agriculture in Prince Edward Island - June 27, 2019
- Go Big or Go Spruce - April 2, 2018
- Will it Play in Peoria, Alberta? - January 22, 2018
- Weather Markets: A Business Case for Environmental History - May 17, 2017
- Enseigner les SIG historiques et restaurer les communautés perdues en classe - May 1, 2017
- Teaching Historical GIS and Restoring Lost Communities in the Classroom - November 1, 2016
Josh and Nolan:
This is a fantastic article! Great work on this project. Obviously, I’m in agreement on the utility of temporal GIS tools for yielding historical insights. ESRI’s web-based GIS tools have made this work much easier in recent years. As you note, you can add time-enabled layers to your GIS datasets. And ArcGIS Online now supports this in the web. You can then visualize the time-enabled layers in ESRI’s “Time Aware” web app. Here’s a quick article on this technique:
I used this technique for my project on the Great Epizootic of 1872-73 and created this Time Aware visualization:
I made use of ESRI’s “Canadian Historic Railways” time-enabled map service, as you note:
It is amazing how easy it is to access these tools and shared data and services. Re-mixing this historical data in GIS visualizations is going to open up a number of new ways to think about historical spatial data.