Modelling Land Use in the Generic Urban Model



Modelling Land Use in the Generic Urban Model

Authors

Michiel de Bok, Barry Zondag, Eric Petersen, RAND Europe, NL

Description

This research presents the development of a land use component of the Generic Urban Models (GUM) Phase 2 project, sponsored by the UK Department for Transport.

Abstract

Land-use and transport interaction models are appropriate instruments to forecast land-use responses to transport change. Such models are dynamic in structure, and iterate between transport and land-use components to model how the system evolves over time. This research presents the development of a land use component of the Generic Urban Models (GUM) Phase 2 project, sponsored by the UK Department for Transport. The land use component is part of a forecasting model for the Leicestershire region in the UK. Besides the application of the model for the Leicestershire County Council, the generalised approach can be used by the UK Department for Transport to test impacts of transport policies and exogenous developments on larger urban areas throughout Great Britain.

The land use component is based on the TIGRIS XL model, an integrated land use and transport model that has been developed for the Transport Research Centre in the Netherlands. The land-use model uses discrete time steps of one year to generate population segmentation and employment figures by zone. A dynamic integration is made with the Central Leicestershire Transport Model every 5 years.

The land use model incorporates four specific land-use components: a demography module, a land-use and real estate market module, a housing market module and a labour market module.

- The demographic module:
addresses the transition processes of the population and households. It simulates aging of the population, birth and deaths, and in and out migration.

- The land-use and real estate market module:
accounts for changes in land use and buildings, office space and houses, and addresses both brown field and green field developments. The level of market regulations when modelling land use changes can be varied between a regulated land use planning system to a free market.

- The housing market module:
simulates the annual moves (if any) of households. It simulates two choices: the choice to move or stay and residential location choice, conditional on a move. A wide set of explanatory variables are used in the choice models, such as household characteristics, neighbourhood amenities, prices and accessibility. Accessibility is taken from the transport model in the form of logsums, providing a direct linkage between the two models. These measures are household type specific and combine mode and destination effects in a consistent way.

- The labour market module:
models the changes in number of jobs by sector and changes in the workforce at the zonal level. Separate models have been estimated on time series data (1986-2000) for seven economic sectors to account for differences in location behaviour between sectors.

In the application of the model to the Leicestershire case study, data from various sources has been used. The base year data are derived from the UK Census, Inter Departmental Business Register (IDBR) and National Travel Survey (NTS). Exogenous scenario inputs are taken from the Office for National Statistics (ONS) Population Statistics and the TEMPRO database. The household and employment location coefficients are taken from estimations on extensive Dutch datasets. These coefficients will be verified with the available UK data. Further developments could include the estimation of new coefficients for the UK.

The presented modelling approach provides consistent forecasts of the spatial distribution of residents and jobs, tailored to the needs of the transport model. In practice, bringing together the land use and transport models will contribute to the ex ante analysis of policy packages targeting at sustainable development. As the EC fifth framework project PROPOLIS has indicated that policy packages combining infrastructure and zoning policies are much more efficient then individual measures. The GUM model is well suited to evaluate such integrated policy packages, particularly because green field and brown field developments are explicitly accounted for within the land use model component.

Publisher

Association for European Transport