Contextualizing the Geographic Influence on Infantry Manoeuvrability in a Historical Battlefield Using GIS: A Case Study of the Canadian Corps in the Second Battle of Passchendaele, First World War

Abstract

The Second Battle of Passchendaele (October 26 - November 10, 1917) remains one of the most enduring symbols of the Canadian experience in the First World World, yet its historical prominence contrasts with a limited comprehension of the geographic realities faced by the soldiers. This gap stems largely from the absence of geographic contextualization in teachings of military history, compounded by a historical under-examination of former battlefields in Flanders. This study addresses this gap via a GIS-based quantitative examination of the influence of geographic conditions on infantry manoeuvrability during the Second Battle of Passchendaele.
High-resolution digital elevation models and historical British Army maps were used to reconstruct key battlefield features — slope, flooding, roads and railways, and enemy fields of fire (viewsheds) — as "manoeuvrability indicators". Infantry battlefield traversal was simulated using least-cost path calculations across multiple individual and 3 aggregated indicators (terrain passability, hazard exposure, manoeuvrability [all]) for 18 battlefield subdivisions (corridors). Statistical analyses were then conducted to evaluate (1) the influence of each geographic variable on manoeuvrability using the Mantel test for correlation and Welch's t-test between equally and independently weighted regression models; (2) the spatial variability of these relationships via pairwise Games-Howell post hoc tests per regression model weighting type; and (3) the alignment between predicted manoeuvrability and historical accounts of battlefield conditions using Cohen's weighted kappa coefficient under ordinal and binary classification schemes.
The results showed least-cost paths to be strongly influenced by spatial concentrations of high- and low-cost features, but also balancing the avoidance of high-cost features with minimization of total costs. All geographic variables exhibited very significant influence on infantry manoeuvrability, but still exhibited substantial spatial variability. Independently weighted regression models generally improved manoeuvrability predictability over equally weighted models, but the complexity of geographic interactions showed some areas with no significant difference between model types. The Games-Howell tests revealed significant variability across corridors, indicating considerable divergence between models for different geographic contexts. Cohen's weighted kappa showed moderate alignment between predicted manoeuvrability and historical accounts, with near-exclusive significance between classification schemes suggesting a trade-off between classification granularity and indicator complexity. Ordinal classifications better captured localized variations, whereas binary more effectively captured broader terrain patterns.
The findings consistently emphasized the complex and spatially variable interplay of geographic variables in influencing infantry manoeuvrability. Rather than challenging established historical interpretations, this study complements them by developing a replicable GIS-based framework to analyze the influence of geographic conditions on infantry manoeuvrability, offering a promising foundation for incorporating geographic context into historical military analysis.

Popular Abstract

The word "Passchendaele" is one that conjures images of the worst aspects of the First World War, from the senseless slaughter to the mud and misery of the battlefield conditions faced by soldiers. It was during the Second Battle of Passchendaele, from October 26 - November 10, 1917, where the Canadian Corps took over the front line and found themselves having to assault a German army which had been in occupation of the high ground of Passchendaele Ridge since 1914. However, the preceding months of fierce fighting combined with incessant autumn rains in the already poorly drained region of West Flanders had turned the rather short two kilometres separating the Canadians from Passchendaele into a shell-pocked landscape which is best described as lunar.
The Second Battle of Passchendaele remains one of the most enduring symbols of the Canadian experience in the First World World, in large part due to the traditions and symbols of remembrance such as John McCrae's 1915 In Flanders Fields poem, which popularized the poppy as a symbol of remembrance. Yet the battle's historical prominence contrasts with a limited comprehension of the geographic realities faced by the soldiers, compounded by a historical under-examination of former Flanders battlefields due to the rapid return to agricultural activities following the First World War which effectively erased much of the evidence of past conflicts.
Advances in geographical information systems (GIS) and the increasing availability of high resolution datasets covering large areas make it possible to digitally reconstruct historical landscapes with increasing detail. In this study, several geographic variables influencing the movement of infantry throughout the battlefield were reconstructed from a combination of historical and modern data sources, including: slope, flooding, roads and railways, and German fields of fire (rifle and machine gun viewsheds). Using these variables, least-cost paths were calculated and analyzed to quantify the influence of each variable on infantry movement throughout the battlefield under different geographical conditions.
The results showed that each geographic variable had a major, but uneven, influence on infantry movement depending on local conditions. In many areas, the influence of geographic variables was clear and predictable, such as steep slopes or extensive flooding forcing large detours in infantry movement. However, no single variable was able to sufficiently explain all variability in movement on its own as most variables exhibited complex interactions with one another. In general, the results exhibited greater complexity at lower elevations where flooding was more extreme and exposure to enemy fire was greatest, but there was a clear and consistent reduction in complexity as conditions improved uphill towards the village of Passchendaele itself.
A comparison of different modelling approaches showed that assigning unique weights to each geographic variable generally provided a more realistic picture of how battlefield conditions constrained infantry movement. Finally, a comparison of the analysis results with historical accounts showed moderate agreement in the classification of battlefield conditions, underscoring the value of a systematic approach even when evaluated against highly variable and often personal narratives of historical events and conditions.
Overall, this study highlights how complex and uneven the influence of geographic variables were throughout the battlefield and provides insight into the experience of the Canadian Corps at Passchendaele. Rather than challenging established historical interpretations, this study complements them by developing a replicable GIS-based framework to analyze the influence of geographic conditions on infantry manoeuvrability, offering a promising foundation for incorporating geographic context into historical military analysis.

• author: Stewart, Jesse Michael
• supervisor: David Tenenbaum; Jonathan Seaquist
• organization: Dept of Physical Geography and Ecosystem Science
• course: GISM01 20252
• year: 2025
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: Geography, GIS, Geographical Information Systems, Passchendaele, First World War, Manoeuvrability, Military Terrain Analysis, Least-Cost Paths
• publication/series: Master Thesis in Geographical Information Science
• report number: 199
• language: English
• id: 9215702
• date added to LUP: 2025-11-24 17:07:03
• date last changed: 2025-11-24 17:07:03