Annex 4 analytical methods

TRUST is a transport network model for the assignment of Origin-Destination matrices at the NUTS3 level of detail for passenger and freight demand on the multimodal transport network of Europe. Road rail, inland waterways and maritime transport modes are covered in separate modules, each with its own matrices, that are then assigned simultaneously on the multimodal transport network. The current version of the TRUST model does not deal with modal split and its main output is the load on road network links in terms of vehicles per day and on non-road links in terms of either passengers or tonnes per day.

TRUST is built in PTV-VISUM software environment. The assignment algorithm used is Equilibrium Assignment which distributes demand for each origin/destination pair among available alternative routes, according to Wardrop first principle. This principle assumes that each traveller is identical, non-cooperative and rational in selecting the shortest route, and knows the exact travel time he/she will encounter. If all travellers select routes according to this principle the road network will be at equilibrium, such that no one can reduce their travel times by unilaterally choosing another route of the same OD pair. This principle has been extended to consider generalised travel cost instead of travel time, where generalised travel cost can include the monetary cost of in-vehicle travel time, tolls, parking charges and fuel consumption costs. The impedance function is defined in terms of generalised time from an origin O to a destination D. Travel costs are defined separately by link types using combinations of fixed, time-dependent and distance-dependent parameters. Travel time is estimated endogenously by the model as result of the assignment. Speed-flow functions are used to model the impact of traffic on free-flow speeds, given links capacity. The model iterates until a pre-defined convergence criterion for equilibrium is reached.

TRUST road transport module

The TRUST road module deals with the assignment of road transport O-D matrices for both passenger (cars) and freight (trucks>3.5t). The road network includes all relevant links between the NUTS3 regions, i.e. motorways, primary roads as well as roads of regional and sub-regional interest. Also ferry connections (Ro-Ro services) between European regions and between European regions and North Africa are explicitly modelled with their travel time and fare.

Road transport demand is modelled in TRUST by means of origin/destination matrices between NUTS3 zones. Intra-NUTS3 demand is not part of the matrices as it is not assigned to the network, but implicitly considered as pre-load on network links. For some non EU countries (e.g. Russia or Ukraine) domestic demand is not part of the matrices.

The passenger matrix includes car trips (coach trips are not modelled) and is segmented into three groups:

• Short distance (< 100 km) commuting
• Short distance (< 100 km) non-commuting
• Long distance (> 100 km)

The freight matrix includes vehicles above 3.5 tonnes between NUTS3 zones and is segmented into the following demand groups:

• Domestic Short distance (<=50 km)
• Domestic average distance (50 –150 km)
• Domestic Long distance (>= 150 km)
• International.

This segmentation allows us to apply dedicated parameters (e.g. considering that short distance domestic transport usually is made of lighter trucks than long distance international transport) and to measure the contribution of the typical vehicles of each segment to link loads. In addition, each demand group is further divided by considering the origin country (the are 242 flows in total) as this allows for the differentiation of fuel costs for the vehicles. Base year (2017) matrices are derived from those estimated in the ETISplus project with further revisions to match Eurostat statistics on road traffic. For forecasting purposes, future matrices are estimated exogenously by applying demand growth rates taken from available sources (e.g. EU Energy and transport trend, ASTRA model, etc.).

Speed-flow functions in TRUST are used to simulate congestion on road links. Since the model assigns daily matrices the speed-flow curves implemented as attributes of the road links are adjusted to take into account that congestion is more hardly recognisable if demand and supply are compared on a 24 hour basis. Speed-flow functions depends on link type, speed and flow/capacity ratio.

Fuel consumption and emissions factors for road modes build on COPERT IV functions but with a relevant modification. Basically, the convex form of the COPERT function has been modified to consider that in real traffic conditions average speeds (the assignment model provides average speeds) are most likely the result of repeated stop-and-go. An average speed of e.g. 70 km/h on motorways means that there is more traffic than when average speed is 110 km/h so one should expect more fuel consumption rather than less fuel consumption as implied by original COPERT functions.

Since COPERT functions are different by vehicle type, an average fleet composition is considered to derive the parameters used in TRUST. When the model is run for forecasting purposes for future years, the emission factors are updated considering projections regarding the evolution of fleet in the selected year.

TRUST rail transport module

TRUST rail module does not consider capacity restrictions and follows an AON (All or Nothing) assignment of origin/destination matrices on the minimum path available on the network. This means that the transport volume on the rail links are computed irrespective of the availability of rail services and of transport chains.

The rail network includes different link types according to technical elements (number of tracks, electrification, maximum speed allowed, etc.) as drawn from the ETISplus database. Links dedicated to some type of traffic (e.g. high-speed service or freight trains) are distinguished as well as links where some types of train are not allowed. The rail network is linked to the road network as intermodal transport is modelled. Rail supply includes intermodal terminals where loads are transferred between road and rail. There are 917 intermodal terminals across the EU countries. In case of passenger transport the interchange links between local/intercity services and high-speed services and transfer between car feeder and local/intercity services are modelled as well.

Rail demand is segmented according to types of traffic which correspond to different train types in terms of occupancy of rail capacity. For passenger demand, three segments based on train type are used:

• Regional Trains
• Intercity Trains
• High Speed Trains (or similar, like the German ICE trains)

Two different types of freight trains are considered:

• intermodal trains,
• conventional trains (conventional block trains or single wagon load trains), which is further split into three groups:
  • conventional trains 700 tonnes
  • conventional train 1200 tonnes
  • conventional train 2900 tonnes.

TRUST maritime transport module

The maritime network includes several ports throughout Europe. Fictitious maritime links provide sea routes to link ports and allows the model to compute travel distances of maritime connections.

Maritime ports are classified into three categories: bulk ports, container ports and general cargo ports. Most of the ports belong to more than one category but some ports have only one or two specialisations; ports can host only demand for those freight segments (e.g. if one port is classified as a bulk port only, maritime routes for general cargo and container demand cannot go through that port). For zones without ports there is no direct access to ship mode, which in turn can be accessed through feeder modes (truck, rail or inland waterway according to existing infrastructures). As a consequence, rail, road and inland waterway networks are also used in the TRUST maritime model because trains, trucks and barges are used as feeder modes to connect inland zones with ports and allow a full path between the origin and the final destination of freight shipment.

Maritime demand consists of origin/destination matrices segmented according to the three categories of bulk, container and general cargo. Matrices are in terms of tonnes per year and each segment of demand has its matrix that is assigned independently to the network.

TRUST inland waterway transport module

TRUST inland waterways (IWW) network includes all the relevant canals among all the NUTS3 regions covered by the spatial area of the model. The network includes 70 main inland ports across Europe selected on the basis of the quantities of goods handled or on their strategic role along the international routes. Each IWW network link has specific features in term of free-flow speed. Specific flags are used to identify links belonging to the Core TEN-T Network, to each TEN-T Corridor and to the comprehensive network. Therefore, results can be provided for these subsets of the network. Demand Origin-Destination matrices are segmented according to two main freight categories: container and non-container. Matrices are based on ETISplus project and further refined on Eurostat statistics.

Further information on TRUST is available on http://www.trt.it/en/tools/trust/

TRUST road transport module input

• OD Matrices at NUTS3 level in terms of vehicles
• Speed-flow functions
• Transport costs by mode
• Travel time value
• Average fuel consumption
• Average emission factors

TRUST rail transport module input

• OD Matrices at NUTS3 level in terms of trips or tonnes in an average day (24 hours)
• Transport costs
• Occupancy / Load factors
• Rail link attributes

TRUST maritime transport module input

• OD Matrices at NUTS3 level in terms of tonnes (bulk, container and general cargo)
• Transport costs
• Occupancy / Load factors
• Maritime link attributes

TRUST inland waterways transport module input

• OD Matrices at NUTS3 level in terms of tonnes (container, non-container)
• Transport costs
• Occupancy / Load factors
• Iww link attributes

TRUST road module outputs

• Average daily loads on road links split by demand segment and by country of origin
• Road traffic activity (passenger-km, tonnes-km, vehicle-km) per year by country (based on territoriality principle).
• Road traffic activity (passenger-km, tonnes-km, vehicle-km) per year on TEN-T core network and on TEN-T corridors.
• Origin-destination journey time.
• Origin-destination journey (perceived) cost.
• Road accessibility measures by NUTS-III region.
• Origin-Destination Paths.
• Energy consumption by link. This can be aggregated to get results by country (territorial principle), on TEN-T core network and on TEN-T corridors.
• Emissions by link for NOx, PM, VOC, CO and CO2.This can be aggregated to get results by country (territorial principle), on TEN-T core network and on TEN-T corridors.

TRUST rail module outputs

• Average daily loads on rail links split by demand segment.
• Rail traffic activity (passenger-km, tonnes-km) per year by country (based on territoriality principle).
• Rail accessibility measures by NUTS-III region.

TRUST maritime module outputs

• Seaport throughput (tonnes) per year by port and cargo type (container, bulk, other)
• Share of feeder modes transporting freight to/from seaports
• Maritime accessibility measures by NUTS-III region

TRUST inland waterways module outputs

• Average daily loads on iww links split by demand segment
• Iww traffic activity (tonnes-km) per year by country (based on territoriality principle). 

TRUST is particularly suitable for modelling road charging schemes for cars and heavy goods vehicles, and policies in the field of infrastructure (e.g. completion of the core and comprehensive Trans-European Transport (TEN-T) network).

More specifically the policy measures that can be simulated with TRUST are:

Road sector

• Road charging (e.g. Eurovignette): Charges can be coded directly if they are based on demand segments of the model, otherwise average charges based on e.g. fleet composition should be estimated exogenously
• Energy taxation: average change of operating cost can be coded according to fleet composition by country
• Road infrastructure changes: Changes can consist of new links and improved links. Given the scale of the model, simulation is meaningful for major modifications (e.g. one corridor) rather than for single links.
• Speed limits       
• Technology – transport information system, management & service: As far as technology is supposed to modify elements like travel speed or link capacity. The entity of the modification should be estimated exogenously
• Truck driver regulations: Indirect simulation based on exogenous assumption on expected impact of regulation on driving cost.

Rail sector

• Infrastructure charging: Charges can be coded directly if they are based on demand segments of the model otherwise average charges should be estimated exogenously
• Rail infrastructure changes: Changes can consist of new links and improved links. Given the scale of the model, simulation is meaningful for major modifications (e.g. one corridor) rather than for single links.
• Technology – transport information system, management & service: As far as technology is supposed to modify elements like travel speed or operational costs. The entity of the modification should be estimated exogenously

Maritime sector

• Infrastructure charging: As far as ports can be charged
• Technology – transport information system, management & service: As far as technology is supposed to modify costs or times at ports. Modification should be estimated exogenously
• Port regulations: As far as regulation is supposed to modify costs or times at ports. Modification should be estimated exogenously

Inland waterways sector

• IWW infrastructure changes: Changes can consist of new links and improved links. Given the scale of the model, simulation is meaningful for major modifications.
• Port regulations : As far as regulation is supposed to modify costs or times at IWW ports. Modification should be estimated exogenously
• Technology – transport information system, management & service: As far as technology is supposed to modify elements like travel speed or reduce operation costs. The entity of the modification should be estimated exogenously.

Impact types that can be assessed with the models include:

Transport

• Transport impact, Environmental impact, Economic impact
  • Transport volumes
  • Modal split
  • Network impacts
  • Emissions
  • Noise
  • Transport costs

Can be assessed through: Modelling of specific scenarios in combination with ASTRA

• Transport