Weather-Related Risk Assessment of Canadian Transcontinental Freight Railway: Case Study of Saskatchewan ^†

Abstract

Railway transport is the backbone of Canada’s supply chain, valued for its reliability and safety despite its inherent complexity, which brings a variety of associated risks. Considering Canada’s climate, analyzing weather-related risks in the railway network is valuable for decision-makers in transportation, especially in preparation for cold seasons and railway network development. In this study, a comprehensive spatial analysis using ArcMap was applied to identifying Saskatchewan’s hotspots and vulnerable geographical areas for the railway network in terms of minimum temperature, rain, and snow, which are the major weather-related risk factors in this province. Risk maps were generated for each factor mentioned, which demonstrate hotspots and hazardous areas that require more attention and planning to maintain the continuity of the supply chain. These results can be used to enhance safety, reduce service disruptions, and ensure the smooth operation of the railway network.

1. Introduction

The advantages of railway transportation, such as its reliability, capacity, and environmental friendliness [1], have solidified its role as a critical component of North America’s transportation and supply chain. This mode of transportation is associated with various risks. Studies indicate that weather-related issues are common causes of railway transport failures [2]. This significance is emphasized by the Canadian government’s commitment to investing over $100 million in railway safety and security in 2021, driven by concerns related to climate change and extreme weather events, which present significant risks. Given the pivotal role of railway transportation in Canada’s supply chain and the increasing impact of weather-related hazards linked to climate change [3], this study focuses on illustrating the impact of weather-related hazards on the CN, one of the two Canadian Transcontinental Railway networks and the sole transcontinental freight railway, in Saskatchewan. In this study, to conduct a spatial analysis and identify hotspots and vulnerable areas for the freight railway in Saskatchewan, we first identify the most significant weather-related factors through a comprehensive literature review. Subsequently, we utilize GIS and ArcMap tools to pinpoint and map these vulnerable areas.

2. Literature Review

Taking into consideration their intensity and frequency, it is evident that snow and rain are among the most impactful weather-related hazards [4]. Additionally, extreme weather-related conditions, particularly extreme temperatures, have a substantial effect on transportation infrastructure [5]. Following this, there are some major papers that have studied the effect of the weather-related risk factors mentioned on the railway network.

Maurer et al. [6] conducted a study that examined six different cases within the European transport system affected by extreme weather conditions. Half of these cases were railway accidents attributed to severe weather events. One of these cases involved a railway accident between Vienna and Prague, which was damaged by flooding resulting from heavy rainfall. Additionally, another case analyzed in their study was a railway failure due to heavy snow in Sweden. Another study focused on the European railway system was carried out by Misnevs et al. [7]. They used European accident history and a likelihood heuristic method, which relied on expert opinion, to identify the weather-related factors affecting railway safety. In this study, they identified ten factors that contributed to accidents and risks in railway transportation, with extreme temperatures, rainfall, and snowfall being among them.

Ma et al. [8] conducted an experimental research study on the effect of low temperatures on the railway network. In this study, the scholars examined the damage that different ranges of low temperatures can cause to rail structures and the damage behavior in extreme cold conditions. The study discussed how rail damage primarily occurred during the cold season, especially from November to March.

Dindar et al. [9] applied the popular railway safety method of Bayesian networks. Using this model, they analyzed railway derailments linked to weather-related causes, including high/low temperature, floods, extreme wind, snowfall, rainfall, and fog. For this research, they utilized USA derailment accident data from 2005 to 2015. They categorized different types of derailments caused by weather-related factors. Then, using Bayesian networks, they studied the root causes and conditions caused by each root, and finally, they applied a fuzzy-based approach to determining the probabilities of derailment being caused by each weather-related factor.

One important study, which conducted spatial analysis on the railway network and included weather-related factors such as snowfall and rainfall (among others), was carried out by Zhao et al. [10]. To perform this analysis, the authors first constructed a database for China’s railway accidents and then, using the relationship between disaster occurrences and natural hazard intensity, they created susceptibility maps. This study revealed that the most significant factor contributing to railway accidents in China is rainfall, while snowfall has the lowest impact. Furthermore, this research demonstrated that 16% of China’s railway network is in a high-risk susceptibility range for disasters.

S. Sun et al. [4] conducted another significant spatial analysis to assess weather-related hazards in China’s railway system. In this study, they utilized a hazard level matrix model to generate risk maps, visualizing the high-risk areas of China referred to as ‘hotspots’. The weather-related factors studied in this paper included gales, rainfall, and snowfall. The hazard level matrix in this research was constructed by considering both risk intensity, calculated using the maximum measure of the factor, and risk frequency, based on annual days of snow and gale, as well as annual heavy rain hours. Then, using this hazard matrix and susceptibility, a risk matrix model was introduced, categorizing risks as very low, low, mild, moderate, severe, or critical. The findings of this research revealed that 29.3% of China’s high-speed railway is exposed to rainfall risk, and 20% is exposed to snowfall risk.

The papers mentioned above are among the most significant studies that have conducted extensive research on the effects of weather-related factors on railway transportation systems. Each of these papers delved into the impact of one or more weather-related factor on railway networks. The selection of weather-related factors in each paper was influenced by various factors, including the geographical area under study and the methodology employed. These studies may vary in terms of the specific factors they investigate, the data they utilize, or their chosen methodologies. However, they all share one common theme: the critical importance and significant role of weather-related risks in the context of railway transportation. As previously mentioned, railway transportation is a crucial component of the global supply chain, particularly in North America.

3. Research Methodology

To generate risk maps for Saskatchewan, the following steps are taken:

Step 1: Initially, a literature review and technical documents, such as the Canadian National Railway (CN) winter plan [11], are used to identify weather-related factors that pose significant risks, especially in this province due to its climate. The final factors for which risk maps are created are as follows:

• Rain: According to CN, rain not only poses a risk to the railway by causing floods but can also be hazardous by itself.
• Snow: According to the Railway Association of Canada [12], snow is a crucial factor in railway operation risk, and many instructions have been introduced to reduce this risk.
• Minimum temperature: Low temperatures can have a severe impact on the railway network, including issues like rail breakage [13].

Government of Saskatchewan reports indicate that the province experiences cold winters with heavy snowfall, justifying = consideration of these factors.

Step 2: Data for the CN and weather station reports across Saskatchewan as reported by the Government of Canada, are collected. These data include minimum temperature reports from each station, as well as the maximum rainfall and snowfall reports from each station. The data cover the period from January 2022 to April 2023.

Step 3: The next step is to select the type of spatial analysis, which, in this case, is Geographic Information Systems (GIS). GIS is a tool used to display and analyze geographical data, making it a practical visualization and spatial analysis tool, particularly useful for handling large datasets [14]. After collecting the data, we utilized ArcMap 10.8.1 software, which is a renowned GIS tool.

Step 4: To determine the values for all the points on the map of Saskatchewan, we employ ArcMap, utilizing the values from each station and various interpolation methods. We experiment with different interpolation methods and select the one with the lowest root mean square (RMS). This selection is guided by prior studies, indicating that the method with the lowest RMS yields improved interpolation results [15].

Table 1. Interpolation methods.

Factor	Method	RMS	Selected
Rain	IDW	24.7	√
	COKRIGING-SIMPLE	25.2
	KRIGING-ORDINARY	24.8
	KRIGING UNIVERSAL	24.8
Snow	IDW	14.4
	COKRIGING-SIMPLE	13.8	√
	KRIGING-ORDINARY	14
	KRIGING UNIVERSAL	14
Minimum Temperature	IDW	2.14
	KRIGING-SIMPLE	2.09
	COKRIGING-ORDINARY	2.07	√
	KRIGING UNIVERSAL	2.1

displays the interpolation methods used for each weather-related factor.

4. Results

Following all of the steps mentioned above and using the GIS tool, risk maps for rain, snow, and minimum temperature are generated.

If we examine the generated risk maps, we can observe the risk ranging from dark blue (indicating the lowest risk) to red (indicating the highest risk). For rain, rainfall ranges from 21.1 mm to 183 mm (based on the maximum rain reported by each station during the study period). For snow, it ranges from 29.4 cm to 85 cm. For minimum temperature, this range is −31.9 to −48.8 degrees Celsius (based on the minimum temperature reported by each station).

Upon reviewing Figure 1b,c, it becomes clear that snow and minimum temperature pose the highest risk in the northern parts of the province, which are less populated than the southern parts. Consequently, there is no railway network passing through these northern areas. However, in the densely populated southern parts, the hazard of minimum temperature is considerable—much greater than that of snow—and there are areas shown in yellow where this hazard is noticeable, and the railway passes through those regions.

Now, looking at Figure 1a, the rain hotspot map, we can see that compared to minimum temperature and snow, the overall rain risk in the whole province appears to be less hazardous. However, in the southern regions, where populated cities and a railway network exist, the rain hazards are more significant than the snow hazards, especially in the south–central regions. The lowest risk in the northern province among the three factors studied is rain, which seems less hazardous in less populated regions.

5. Conclusions

In examining the significance of the railway in Canada’s supply chain, this study employs historical data, related studies, and GIS tools to generate risk maps for rain, snow, and minimum temperature—identified as threats to the Canadian national railway. These maps reveal that the most significant hazards are in the northern areas, where the sparse population makes railway development unjustified. Additionally, in the southern populated regions, the overall risk of minimum temperature surpasses that of snow, necessitating increased attention for planning during cold seasons. While rain poses a lower overall risk, it is more considerable in the southern areas that the railway passes through, requiring focused planning due to its potential for supply chain disruption.

One limitation to this study was the unavailability of data. In Saskatchewan, some of the weather stations did not consistently report weather factors or ceased functioning after a certain period. More data could lead to more precise risk maps. For future studies, the effects of climate change on changes in the risk maps in the coming years can be explored. Additionally, different statistical methods can be tested to compare the risk patterns across various generated risk maps and compare them to historical accidents and data to estimate the best possible method.