What is TMfS?

TMfS is a multi modal strategic Transport and Land Use modelling suite that contains an extensive database of road and public transport supply information, travel demand data and land-use information. It has been used extensively to provide input to a number of major Scottish transport scheme and policy applications throughout the previous commission for Transport Scotland and a number of Local Authorities and Regional Transport Partnerships. In addition, it has also been used to provide supporting information for regional and local area models within the TMfS modelled area.

How does TMfS work?

TMfS works as a four stage demand modelling procedure programmed using TRIP/CUBE software. It models Mode Choice, Destination Choice, Time of Day Choice and Route Choice. The route choice provides costs for the demand model and also data for operational analysis. There is also an associated Land Use Model called TELMoS, which provides planning data inputs to the Travel Demand Model.

How does the model converge?

Highway Model - The highway model convergence has two parts. Firstly Global convergence is achieved when the cost of the network changes by less than 0.01 on 3 successive iterations of the model. Once Global Convergence of the network is achieved then Link Based Convergence is examined. This is done by checking whether 95% of the links in the network change in flow by less than 5% on 3 successive iterations. Once both types of convergence have been achieved then the model is said to have converged.

Demand Model - The demand model reaches convergence using the method of Trip Damping. This means that for every loop of the demand model the solution is made up of the current solution and those from previous loops. For example on loop 2, the solution is 50% of the solution from loop 1 and 50% of the solution from loop 2. On loop 3 it is 33% of each loops solution and so on. From testing 5 loops of the demand model have been deemed to be enough to reach convergence.

What software platforms are used in TMfS?

TMfS currently uses the CUBE software platform, while TELMoS (the land-use model) makes use of Delta. As part of the enhancement programme, we are comparing Citilabs (the suppliers of CUBE) against an alternative software package offered by PTV, but this will be a largely internal decision designed to meet Transport Scotland objectives. It is anticipated that we will have to continue to prepare bespoke software where necessary.

How is TMfS Validated?

The Road Assignment Model is validated against a whole range of counts that were not used in the calibration process while additional validation takes place against journey time surveys. Further validation is undertaken against Census travel-to-work data. It is hoped that the new commission will be able to utilise a greater amount of travel time data, Trafficmaster data for example, and forge closer links with the SRTDb. Validation tests on the Public Transport and Demand Model are also undertaken by testing a number of scenarios. However, the validation process for both current and future year tests is dependent on both the range and quality of data available.

How will TMfS model integrated ticketing and ‘soft measures’?

The current status of the proposed enhancements for soft measures means that it is likely to be a desk based study analysing what soft measures are currently in place elsewhere in the UK\Europe or further afield and discuss how these could be integrated into the modelling framework. However, even following such a study, it is unlikely that we will be able to fully and robustly incorporate soft measures into the model until a range of empirical evidence exists and validation to an elasticity of change based on measures introduced can be ascertained. In terms of ticketing, TMfS currently uses a distance based fares structure. There are particular difficulties in attaining season ticket and zonal data that would provide input to the modelling of integrated ticketing, but it is hoped that both new and available data may be able to go some way to support this. The inclusion of National Rail Survey data and information from the Concessionary Travel database will also assist in providing a more robust dataset for PT demand

Does TMfS have the potential to model high speed rail links between Scotland and London?

The main thrust of public transport enhancements in TMfS (including the inclusion of air and ferries) is to model internal trips. TMfS is, in essence, a Scottish model and although it has a representation of cross border movements, it is not specifically interested in ultimate origins or destinations beyond the border. However, there is a possibility that TMfS could be used to feed into other models such as a model designed to evaluate high speed rail between Scotland and London.

Is it possible to model journey time reliability?

Journey time reliability is an important issue but it is almost impossible to systematically model.

Will planning data be enhanced for future models?

we will consult with the Local Authorities to recollect planning data. The information from the SESTRAN area was not validated by the planning authorities, although this may be rectified. To some extent, developer hotspots may be covered by David Simmonds Consultancy’s proposal to monitor investment/development at the zonal level. However, we cannot allow for the situation where developers influence policy changes as this can be a very difficult variable to plan.

What are the procedures for the calibration and validation of TELMoS?

The approach to calibrating and validating a land-use model is different to that of a transport model such as TMfS. Firstly, land-use modelling involves change over time, which means that surveys are difficult to conduct and data is particularly sparse. However, we have drawn upon work undertaken by urban geographers and economists to inform the process. The idea is to test that TELMoS provides the same forecasts, although there is an element of judgement involved in the process. We are currently developing new surveys to try and improve this process and hope to integrate these into the Enhancement Report. However, it will nonetheless remain difficult to find data that is ‘good enough’ and released at sufficiently frequent and regular intervals to encourage improvements. Population estimates are worth considering but are subject to variation and are not disaggregated to Local Authority level. One method of possibly improving validation is to conduct post-study reviews.

Can a robust audit of TELMoS be undertaken with intellectual property rights in place?

Yes, TELMoS does have its weaknesses but we plan to identify these through the risk assessment process. It must be remembered that TELMoS is all we have and we will provide guidance on how the results should be used and where the weaknesses in the model exist.

What are the costs of maintaining an increasingly complex TMfS?

Transport Scotland is fully committed to taking TMfS forward and is willing to meet any realistic resource demands for the model.

Will there be intermediate releases of TMfS following its release for the STPR?

The model is continually evolving and, as a result, there is a need to continually update TMfS. Issues in the model application will be addressed and, most importantly, it must be remembered that more applications of TMfS will lead to a more robust and better TMfS. It would generally be anticipated that the most up to date version of TMfS at the time is appropriate for a new application and that each application and use of TMfS requires application and so can be discussed at that time. Although TMfS itself is envisaged to remain in its current form over the next year, it is likely that data as opposed to the entire model will be updated and can be provided to users as required.