Authors: Dr. Jean-Paul Rodrigue and Dr. Cesar Ducruet
Transportation networks are a framework of routes linking locations. The structure of any region corresponds to networks of economic and social interactions.
1. Transport Networks
Transportation systems are commonly represented using networks as an analogy for their structure and flows. Transport networks belong to the wider category of spatial networks because their design and evolution are physically constrained instead of non-spatial networks such as social interactions, corporate organization, and biological systems, which are usually constrained by other factors and where space plays a lesser role.
The term network refers to the framework of routes within a system of locations, identified as nodes. A route is a single link between two nodes that are part of a larger network that can refer to tangible routes such as roads and rails, or less tangible routes such as air and sea corridors.
Transportation networks are the outcome of a trade-off between the goal to connect as many locations as possible and cost and infrastructure development constraints. The territorial structure of any region is expressed as a network that includes all its economic interactions, but more realistically as sub-networks expressing one dimension. The implementation of networks is rarely premeditated, but the consequence of continuous improvements as opportunities arise, investments are made, and conditions change. The setting of networks is the outcome of various strategies, such as providing access and mobility to a region, reinforcing a specific trade corridor or technological developments, making a particular mode, and its network more advantageous over others.
A transport network denotes either a permanent track (e.g. roads, rail, and canals) or a scheduled service (e.g. airline, public transit, train). It can be extended to cover various types of links between points along which mobility can take place. The relevance of a network is related to its connectivity. Metcalfe’s law states that the value of a network is proportional to the square of connected nodes so that complex networks are exponentially more valuable than simple networks since they offer a large number of options in connecting locations. Thus, economic development is commonly associated with network complexity.
In transport geography, it is common to identify several types of transport structures that are linked with transportation networks with key elements such as nodes, links, flows, hubs, or corridors. Network structure ranges from centripetal to centrifugal in terms of the accessibility they provide to locations. A centripetal network favors a limited number of locations, while a centrifugal network does not convey specific locational advantages. Network structures can also be direct or indirect in their connectivity. The most directly connected networks are point-to-point networks where a service originates and ends in a single location. A more complex form involves a route network where intermediary locations are serviced along a linear sequence.
The recent decades have seen the emergence of transport hubs, a centripetal form, as a common network structure for many types of transport services, notably for air transportation. Although hub-and-spoke networks often improve network efficiency, they have drawbacks linked with their vulnerability to disruptions and delays at hubs, an outcome of the lack of direct connections. Evidence underlines that the emergence of hub-and-spoke networks is a transitional form of network development rationalizing limited volumes through a limited number of routes. When traffic becomes sufficient, direct point-to-point services tend to be established as they better reflect the preference of users. Therefore, the more traffic a network supports, the higher its propensity towards direct connections.
Transport networks are better understood by the usage level (e.g. the number of passengers, tons, vehicles, capacity) than by their sole topology based on a binary state (i.e. presence or absence of links). Inequalities between locations can often be measured by the number of links between nodes and the related economic opportunities associated with connectivity and the level of traffic. Many locations within a network have higher accessibility, which is often related to better opportunities. However, economic integration processes tend to change inequalities between regions, mainly through a reorientation of the structure and flows within transportation networks at the transnational level. Economic and commercial changes are associated with changes in network configurations and connectivity.
The efficiency of a network represents its ability to support flows while operating conditions meet performance criteria such as speed, capacity, and safety. It can be measured through graph theory and network analysis. These methods rest on the principle that the efficiency of a network depends partially on its topology, which is the layout of nodes and links. Some network structures have a higher efficiency level than others, but careful consideration must be given to the basic relationship between the revenue and costs of specific transport networks. Rates thus tend to be influenced by the structure of transportation networks since the hub-and-spoke structure, particularly, had a notable impact on transport costs, namely through economies of scale.
The efficiency of transportation networks is also related to their resilience, which is the ability to support disruptions while maintaining a level of service and connectivity. A resilient network remains connected after facing disruptions such as severed nodes or links. A network could be efficient but not very resilient, or the other way around. For instance, a hub-and-spoke network enables a high level of efficiency for air transportation through the consolidation of flows and better usage of airplane assets. Still, such a network is not very resilient, particularly if a hub is disrupted. Thus, depending on the location of the same disruption in a transportation network, its impact could differ widely if it concerns a hub or another node.
2. The Topology and Typology of Networks
Transportation networks, like many networks, are generally embodied as a set of locations and a set of links representing connections between those locations. The arrangement and connectivity of a network are known as its topology, with each transport network having a specific topology. The most fundamental elements of such a structure are the network geometry and the level of connectivity. Transport networks can be classified in specific categories depending on a set of topological attributes that describe them. It is thus possible to establish a basic typology of transport networks that relates to their geographical setting as well as their modal and structural characteristics.
The physical grounding of a network varies in relevance depending on the transport mode considered. Roads and railways are composed of track infrastructure, while maritime and air transports remain vaguely defined due to their higher spatial flexibility. The exception is terminals, whereas maritime networks remain more constrained than airline networks due to port sites being less abundant than airport sites. River networks typically form basins and can be classified as trees or dendrograms. Therefore, there are three types of physical spaces on which the typology of transport networks are set and where each represents a specific mode of territorial occupation:
- Clearly defined and delimited. The space occupied by the transport network is strictly reserved for its exclusive usage and can be identified on a map. Ownership can also be clearly established with defined rights of way, either privately or publicly owned. The most rrelevant examples include road, canal, and railway networks.
- Vaguely defined and delimited. The space of these networks may be shared with other modes and is not the object of any ownership, only of rights of way. Examples include air and maritime transportation networks.
- Without definition. With these networks, space has no tangible meaning, except for the distance it imposes with nodes being the core structure. Little control and ownership are possible, but agreements must be reached for common usage. Examples are radio, television, WiFi, and cellular networks, which rely on the use of specific frequencies granted by regulatory agencies.
Networks provide a level of transport service related to their costs, implying that levels of economic development are related to network density. An optimal network would be a network servicing all possible locations, but such a service would have high capital and operational costs. Transport infrastructures are established over discontinuous networks since many were not built simultaneously, by the same entity, or using the same technology. A subway system could be built over a period of several decades with new segments using a different technology. A rail network could be built by several different railway companies and subject to consolidation in later phases. An air transportation system could be composed of the network of several carriers, each having its hubs and subject to mergers and new entrants. Therefore, operational networks rarely service all parts of the territory directly and homogeneously. Some compromise must often be found among a set of alternatives considering a variety of route combinations, levels of service, and competitiveness. Networks are also labeled depending on their overall properties:
- Regular network. A network where all nodes have the same number of edges. In the same vein, a random network is a network that is formed by random processes. While regular networks tend to be linked with high levels of spatial organization (e.g. a city grid), random networks tend to be linked with development opportunities such as accessing a resource.
- Small-world network. A network with dense connections among close neighbors and few but crucial connections among distant neighbors. Such networks are particularly vulnerable to catastrophic failures around large hubs.
- Scale-free network. A network having a strong hierarchical dimension, with few vertices having many connections and many vertices having few connections. Such networks evolve through the dynamic of preferential attachment by which new nodes added to the network will primarily connect larger nodes instead of being connected randomly.
Investigating the interdependencies among different transport networks, notably when those are of different nature and structure, is challenging. Some crucial aspects and problems related to inter-network relations may be as follows:
- Coevolution. Different transport networks might follow similar or different paths based on spatial proximity and path-dependence of economic development, with a wider variety of networks in core regions than in peripheral regions.
- Complementarity. Some locations may be central in one network but peripheral in another, depending on their specialization and function and the scale of analysis (terminal, city, region, country); the complementarity between networks can be measured based on the number of common nodes and links.
- Interoperability. Typically, cargo flows from a maritime network to a road network shift from a scale-free structure to a regular structure, thus following different topologies that are not easily combined; air and sea terminals remain few in the world due to the difficulty combining and integrating technically air and sea networks physically at the same locations.
- Vulnerability. How do changes in one network affect the other network, on a global level (entire network) or local level (single node or region)? This is particularly important for two networks sharing common nodes, such as global cities, logistics platforms, and multilayered hubs in the case of abrupt conjunctures (e.g. natural disasters, targeted attacks, labor disputes, security, and geopolitical tensions), thus posing the problem of rerouting flows through alternative routes and locations.
3. Networks and Space
Transportation networks underline the territorial organization of economic activities and the efforts incurred to overcome distance. These efforts can be measured in absolute (distance) or relative terms (time) and are proportional to the efficiency and the structure of the networks they represent. Due to the operational and technical characteristics of their modes and terminals, transportation networks have distinct spatial configurations. By its structure and inherent properties, a transport network by itself reveals information about the spatial structure it supports in terms of the importance of locations, how they are connected, what is of importance, and what is of lesser importance.
The relationships transportation networks establish with space and the information they reveal are related to their continuity, their topographic space, and the spatial cohesion they form. The territory is a topological space having two or three dimensions depending on the transport mode considered. Roads are roughly set over a two-dimensional space while air transport is set over a three-dimensional space. However, flows and infrastructures are linear, having one dimension since they conceptually link two points. Thus, establishing a network is a logical outcome for a one-dimensional feature to service a territory by forming a lattice of nodes and links. Long-distance links tend to connect nodes that are of high importance, while short-distance links tend to connect nodes of lower importance or low importance nodes with a hub higher in the hierarchy.
In order to have spatial continuity in a transport network, three conditions are necessary:
- Ubiquity. The possibility to reach any location from any other location on the network, thus providing general access. Access can be a simple matter of vehicle ownership or bidding on the market to purchase a thoroughfare from one location to another. Some networks are continuous, implying that they can be accessed at any location they service. Roads are the most salient example of a continuous network. Other networks are discrete, implying that they can only be accessed at specific locations, commonly at a terminal. Rail, maritime, and rail networks are considered discrete networks since they can only be accessed through their terminals.
- Fractionalization. The possibility for a traveler or a unit of freight to be transported without depending on a group. It becomes a balance between the price advantages of economies of scale and the convenience of dedicated service.
- Instantaneity. The possibility to undertake transportation at the desired or most convenient moment. There is a direct relationship between fractionalization and instantaneity since the more fractionalized a transport system is, the more likely time convenience can be accommodated. Many networks such as air and maritime shipping networks are subject to scheduling, implying that a connection between two node could only be serviced by, for instance, a daily flight or a weekly port call.
These three conditions are never entirely met as some transport modes fulfill them better than others. For instance, the automobile is the most flexible and ubiquitous mode for passenger transportation. However, it has important constraints, such as low capacity and high levels of space and energy consumption. In comparison, public transit is more limited in the spatial coverage of its service, implies batch movements (busloads, trainloads, etc.), and follows specific schedules (limited instantaneity), but is more cost and energy-efficient if its volume is high enough.
Freight transportation also varies in its spatial continuity, ranging from massive loads of raw materials (oil and ores) that can be handled only in a limited number of ports to highly flexible parcels movements carried by vans. Containerization has been a remarkable attempt to address the issue of ubiquity (the system permits intermodal movements), fractionalization (each container is a load unit), and instantaneity (units can be loaded by trucks at any time of the day, and containerships make frequent port calls).
An important cause of discontinuity is linked to the spatial distribution of economic activities, notably industrial and urban, which tend to agglomerate. Congestion may also alter these conditions. Road congestion in a metropolitan area may impair ubiquity as some locations may be challenging to reach since their accessibility is reduced. Fractionalization may also be reduced under such circumstances as people would consider public transit and carpooling and would thus move as batches. Further, as commuters cope with increasing congestion, several trips may be delayed or canceled altogether, reducing instantaneity.
Transportation networks have always been a tool for spatial cohesion and occupation. The Roman and Chinese empires relied on transportation networks to control their respective territories, mainly collecting taxes and moving goods and military forces. During the colonial era, maritime networks became an important tool of trade and political control, which was later on expanded by the development of modern transportation networks within colonies. In the 19th century, transportation networks also became a tool for nation-building and political control. For instance, the extension of railways in the American hinterland had the purpose of organizing the territory, extending settlements, and distributing resources to new markets. In the 20th century, road and highways systems, such as the Interstate system in the United States and the autobahn in Germany, were built to reinforce this purpose. In the later part of the 20th century, air transportation networks played a significant role in weaving the connectivity of the global economy. For the early 21st century, telecommunication networks have become a means of spatial cohesion and interactions fulfilling the requirements of global supply chains.
4. Network Expansion
During the industrial revolution in England, the co-evolution of roads, canals, and ports revealed noticeable interdependencies among the different nodes and networks over time, based on spatial and functional proximity. Initial network developments are often done to support and complement an existing network. Then, the new network competes with the existing network by expanding geographically and topologically in ways unavailable to the prior network. This was particularly the case when rail networks expanded the opportunities created by canal networks. Later, highway networks expanded the opportunities offered by rail networks.
As transport networks expand, existing transport infrastructures are being upgraded to cope with spatial changes. Airports and ports are being transformed, expanded, or relocated. In the air transport sector, the emphasis is being given to integrating airports within fully-fledged multimodal transport systems, networking air with rail and road transport. In maritime transport, networks are also being modified with increasing attention paid to the expansion of the Panama and Suez Canals, increasing traffic on inland waterways, and creating new inland passages between semi-enclosed or enclosed seas.
The global land transport network structure is a function of the density and intensity of economic activities, interconnected systems of cities, and efforts made to access inland resources. Network length tends to be a function of the population and the level of economic development, with geography playing a role in terms of suitable (flat) landscape. Based upon this density, the road and rail networks shift from a grid (high density) to linear corridors (low density). While the global road and rail networks appear to be integrated and interconnected, this is far from being the case. Road networks are designed to service local and regional flows, and only a few corridors are used for long-distance trade. Most rail networks are national in scope with limited international services except for Europe and North America.
The growing competition between the sea and land corridors is not only reducing transport costs and encouraging international trade but prompting many governments to reassess their land-based connections and seek shorter transit routes. Existing land routes are also being extended. Passages through difficult terrain are being investigated with a view to creating fully-fledged land-based continental connections, notably through railways. These land network expansions are driven by economic globalization and inter-regional cooperation and eventually become multimodal transcontinental corridors for rail, road, pipelines, and trunk telecommunications routes. But the impact of increasing world trade on land network expansion, notably over railways, is scale specific. The development of railways has permitted inter and intra-continental connections, namely landbridges in North America and Eurasia.
In recent years, new rail routes in North America, Eurasia, Latin America, and Africa have been developed or are being considered. There is scope for shippers to increase their trade through these new routes, particularly if rising insurance premiums, charter rates, and shipping risks prompt them to opt for a land route instead of the sea route through the Suez or Panama canals. These developments linked to the integration of regional economies to the world market are part of a rationalization and specialization process of rail traffic. However, the success of these rail network expansions depends on the speed of movement and general cargo unitization through containerization. Railways servicing ports tend to consolidate container flows, which allows an increase in capacity and the establishment of door-to-door services through a better distribution of goods among different transport modes. New links are establishing and reshaping new trade flows, underpinning outward cargo movements and the distribution of goods. As some coastal gateways are emerging as critical logistics services centers that rationalize distribution systems to fit new trading patterns, the land network development and cross-border crossings throughout the world have far-reaching geopolitical implications.
- 1.1 – What is Transport Geography?
- 1.2 – Transportation and the Physical Environment
- Graph Theory: Definition and Properties
- Graph Theory: Measures and Indices
- Network Data Models
- Barthelemy, M. (2010) Spatial networks, Physics Reports, No. 499, pp. 1-101.
- Briggs, K. (1972) Introducing Transportation Networks, London: University of London Press.
- Dalton, R., J. Garlick, R. Minshull and A. Robinson (1978) Networks in Geography, London: George Philip & Son Ltd.
- Leinbach, T. (1976) “Networks and Flows”, Progress in Human Geography, Vol. 8, pp. 179-207.
- Newman, M. (2010) Networks: An Introduction. Oxford: Oxford University Press.
- O’Kelly, M. (1998) “A geographer’s analysis of hub and spoke networks”, Journal of Transport Geography, Vol. 6(3), pp. 171-186.
- Scott, D., D.C. Novak, L. Aultman-Hall, and F. Guo (2006) “Network robustness index: A new method for identifying critical links and evaluating the performance of transportation networks”, Journal of Transport Geography, Vol. 14 (3), pp. 215- 227.
- Taaffe, E.J., H.L. Gauthier and M.E. O’Kelly (1996) Geography of Transportation, Second Edition, Upper Saddle River, NJ: Prentice-Hall.