High Speed Rail Demand: Empirical and Modelling Evidences from Italy
E Cascetta, Federico II University of Naples, IT; P Coppola, Tor Vergata University of Rome, IT
We present a demand forecasting study on High Speed Railways in Italy, based on models and empirical observations.
Major investments in High Speed Railways (HSR) have been carried out in Italy resulting in a current network approximately 1300 Km long. The HSR service, started in 2005 between Rome and Naples, now includes several city pairs. The level of service is expected to be further improved with the completion of the new stations in Naples, Florence, Turin and Rome that will allow direct service avoiding central stations in dense urban area, and, on the other hand, with the entrance in the HSR market of a new private operator (i.e. Nuovo Trasporto Viaggiatori- NTV), competing with the national railways one. Such changes creates the conditions in the Italian national transportation market, for a unique case study to investigate the behavior of long-distance passengers.
In this paper we present the results of the second stage of a study to forecast the passenger demand on HSR in Italy, for different Italian macroeconomic, transport supply, and HSR marketing scenarios.
In the first part of the study, presented at the ETC last year, we developed the methodology to forecast the Origin-Destination (OD) passenger volumes by HSR services, based on the following integrated demand models has :
- the national demand growth model projects the base year total OD volumes to future years, according to assumed macroeconomics trends;
- the mode/service choice model estimates the market shares of different inter-urban transportation modes, including alternative rail services, such as Intercity, High-Speed, 1st and 2nd class; in other terms, this model aims at simulating the competition between modes on a given OD pair and the completion among different HSR trains operated by Trenitalia and NTV on the same rail track (i.e. competition within HSR mode);
- the induced demand model which estimates the additional HSR demand due to the improvement of HSR level of services (i.e. new services, travel time reductions, etc.).
Such models were estimated, based on surveys and source data, carried on in 2009 and applied to predict the impacts on national passenger volumes, of the new HSR services and operators as of 2012, under different macroeconomic assumptions and marketing strategies of the main passenger transportation competitors on the long distance (i.e. NTV vs. Trenitalia and HSR operators vs. airlines).
In the second part of the study (here presented), a new campaign of (RP-SP) surveys and traffic counts have been carried on in year 2010, in order, firstly, to validate the model and, secondly, to upgrade the mode-service choice model and the induced demand model.
To validate the model we have compared the model forecasting in the 2010 scenario to the observed traffic volumes on the different modes. The results of such analysis have shown an outstanding increase of HSR demand in year 2010 (+45%), partially due to the increase of the level of service (e.g. in year 2010 the new HSR segment between Bologna and Firenze was completed, leading to significant reductions in travel times on the whole network) and partially due to the recovering of the national economy. This outstanding increase has not resulted to be entirely predicted by our models, leading us to develop new model specifications.
In facts, a run-based mode-service choice model has been developed, aiming at simulating the competition between the two different operators (i.e. Trenitalia and NTV) based not only on the fare structures but also on the actual departure time of the trains (i.e. the HSR timetables. The model is set as a nested-logit models (see Ben-Akiva and Lerman, 1985; and Cascetta, 2009) with a nesting structure to capture higher degrees of substitutions among specific subsets of modal alternatives, particularly the HSR alternatives provided on the same route by different operators, i.e. NTV vs. High-Speed Trenitalia.
The new model specifications, estimated for two trip purposes (i.e. Business and Non-Business) using Maximum Likelihood method, will be presented and discussed in the paper, together with the results of the empirical observations of the national passenger volumes.
Ben-Akiva M. and Lerman S. (1985) Discrete Choice analysis MIT Press.
Ben-Akiva M., Cascetta E., Coppola P., Papola A., Velardi V. (2010) High Speed Railways Demand forecasting in a competitive market: the Italian case study. Proceedings of the ETC 2010, Glasgow, (UK).
Association for European Transport