The ITF Modelling Framework
The ITF framework first estimates the demand for transport, based on a set of socio-economic drivers (population, Gross Domestic Product, trade, etc.) before analysing the way this demand may be satisfied. This second step includes a detailed modelling of mode choice. Finally, the models compute the activity and the generated CO2 emissions linked to transport. Additionally, depending on the sector, other transport-related variables and indicators.
The ITF framework can assess the effect of a large range of policies and exogenous impacts. In all models, policies which may impact transport demand or the related CO2 emissions become input parameters. The models are constantly being worked on, improved and updated. None of the models relies on any commercial (transport) modelling software and are developed entirely in-house. The most recent model documentation was developed under the Decarbonising Transport in Europe project in early 2020 and is accessible via the following links:
- Global urban passenger transport model – see here
- Global urban freight transport model – see here
- Global non-urban passenger transport model – see here
- Global non-urban freight transport model – see here
The information below provides a brief overview of some of the main characteristics of each model.
The urban passenger transport model
The ITF global urban passenger transport model is a strategic tool to test the impacts of policies and technology trends on urban travel demand, related CO2 emissions and accessibility indicators. Outputs for various scenarios can be obtained to 2050. The model represents passenger mobility at the scale of Functional Urban Areas (FUAs).
The model is designed as a systems dynamic model (stock and flow model) to evaluate the development of urban mobility in all cities over 50 000 inhabitants around the world. It combines data from various sources that form one of the most extensive databases on global city mobility to account for fifteen transport modes. These range from the conventional private car and public transport to new alternative modes such as shared mobility.
The urban passenger model represents travel behaviour by modelling aggregate travel behaviour by traveller segment. A traveller segment is defined by socio-economic characteristics of travellers (e.g. their gender, income level and age). While the model is built at the FUA level, the final analysis is carried out for nine world regions.
The urban freight transport model
The ITF Urban Freight Model estimates the impact of policy measures related to decarbonizing urban freight transport under different scenarios. It applies innovative ways to overcome the general lack of data describing urban freight movements.
The first version of the model was developed for a single urban area, Groot-Rijnmond in the Netherlands. The model was then expanded to all European functional urban areas (FUAs) and eventually to those of the rest of the world. The FUAs are divided into a grid of smaller zones of two by two kilometres for obtaining more detailed results with higher spatial resolution.
The model applies a classic four-step modelling approach. It accounts for different commodity types that may expose different shipment characteristics. In the freight generation step, freight production and consumption of a zone, per commodity type, are estimated with a generalised linear model (GLM), based on spatial characteristics of the urban area (such as information on employment and retail space). In a next step, freight flows for six distance ranges are estimated, again for each commodity type. Here, a distance bin split model assesses the shares of shipments falling into each distance bin depending on city characteristics. In a third step, the obtained flows for each distance bin and commodity type are converted into trips using specific vehicle types. For this, an iterative procedure is applied that uses information on the likelihood of use of specific vehicle types for the different commodities. Finally, the emissions of each trip are calculated based on the vehicle type and ton-kilometres of each trip.
The non-urban passenger transport model
The ITF non-urban passenger model is a strategic tool that tests the impacts of multiple policies and trends on the non-urban passenger sector. The model provides scenario forecasts for non-urban transport activity and its related CO2 emissions up to 2050. The model estimates activity between urban areas (intercity travel) and passenger activity happening locally in non-urban areas (intra-regional travel). The latter includes travel in peri-urban and rural areas. The model is developed to assess the impact of transport, economic and environmental policy measures (air liberalisation, carbon pricing, etc.), as well as the impact of technological developments and breakthroughs (electric aviation, autonomous vehicles, etc.).
The model builds on two older ITF models, the international passenger aviation model and the domestic non-urban passenger model. The new non-urban model combines and enhances these two models, now accounting for all multimodal passenger activity that occurs outside of urban areas. The model structure comprises eleven sub-models (or modules) that estimate the non-urban passenger transport activity and its effects on the environment.
The non-urban freight transport model
The ITF non-urban freight transport model assesses and provides scenario forecasts for freight flows around the globe. It is a network model that assigns freight flows of all major transport modes to specific routes, modes, and network links. Centroids, connected by network links, represent zones (countries or their administrative units) where goods are consumed or produced.
The most recent version of the ITF freight model integrates the (previously distinct) surface and international freight models. International and domestic freight flows are calibrated on data on national freight transport activity (in tonnes-kilometres, tkm) as reported by ITF member countries. Reported data is also used to validate the route assignment of freight flows. Trade projections in value terms stem from the OECD trade model and converted into cargo weight (tonnes). These weight movements are then assigned to an intermodal freight network that develops over time in line with scenario settings. These define infrastructure availability, available services and related costs.
The current version of the model estimates freight transport activity for 19 commodities for all major transport modes including sea, road, rail, air and inland waterways. The underlying network contains more than 8 000 centroids, where consumption and production of goods takes place. Each of the more than 150 000 links of the network is described by several attributes. These include length, capacity, travel time (incl. border crossing times), and travel costs (per tkm).