LEFT2 ? a Strategic Freight Transport Model for Great Britain

LEFT2 ? a Strategic Freight Transport Model for Great Britain


A S Fowkes, A E Whiting, D Johnson, ITS, University of Leeds, UK


This paper reports a new strategic freight transport model, calibrated to GB data, with policy test results. The major innovation is that it incorporates a feedback from composite costs to market size in addition to mode split effects.


This paper reports some innovative model building that yield a useful addition to models of freight in Great Britain. Currently, the Great Britain Freight Model (GBFM) is part of the Department for Transport?s national model suite. That model is over-detailed for some purposes and can only model mode choice effects. Freight modelling requirements in Great Britain were reviewed in work for the department for Transport, the Strategic Rail Authority, the Highways Agency and Transport for London. Arising from our work on that project, and our involvement with GBFM, we identified a need for a strategic freight model, near instantaneous to run, that could handle both mode choice effects and market size effects, the latter driven by changes in composite cost. Such a model has been constructed within funds available as part of the Department for Transport funded ITeLS project. We have called it the LEFT2 model, where LEFT stands for the LEeds Freight Transport. Documentation is available of the ITeLS website, www.ITeLS.org.uk.

LEFT2 is a spreadsheet based model, split in three dimensions:
1. Whether or not the traffic (rail and road only considered) is suitable for through train operation (ie. a block train siding to siding);
2. The commodity group (from 7 exhaustive groups);
3. The road OD distance band (from 9 exhaustive bands).

Cost models are provided that calculate costs in each cell for Road and Rail. Base traffic matrices are provided for 2000 and 2010. Policy changes are allowed to affect the costs of each mode in each of the 126 cells (=2x7x9). A simple logit model, calibrated to exactly reproduce the two base matrices, gives the cell traffic percentage split by mode. The same logit model parameters are used to form a composite cost for each cell that is allowed to affect the total (road plus rail) traffic in that cell. Interesting data outputs for tonnes and tonne-kms are presented on an output sheet. Example outputs of the effect of the EU Working Time Directive, SRA Company Neutral revenue Support, DfT Road User Charging, and increased road vehicle efficiency will be presented.

The purpose of the paper is to present our work for discussion within a timescale that will allow feedback into a later version of the model which we currently have funding to develop. The paper will therefore discuss a range of the short-cuts and simplifications that we have employed in LEFT2. In some cases we have probably solved some difficulties that will face others, for instance our method to reallocate traffic over distance bands to achieve a change in the average length of haul. Other difficulties have been parked for later work, which is now underway. By the time of the conference we hope to have refined the modelling and added outputs for vehicle kilometres and emissions.


Association for European Transport