University of Twente Student Theses


Towards Integrated Land Use and Transport Modelling: Evaluating Accuracy of the Four Step Transport Model- the Case of Istanbul, Turkey

Gachanja, James Njiraini (2010) Towards Integrated Land Use and Transport Modelling: Evaluating Accuracy of the Four Step Transport Model- the Case of Istanbul, Turkey.

[img] PDF
Abstract:Urban land use and transportation are known to interact and influence each other yet it is common practice for planners in both domains to prepare plans without due consideration for this interaction. Integrated land use and transport models are therefore important because they enable policy makers to foresee and evaluate the effects of transport and urban plans jointly thus providing solutions to common planning problems. The four step model (FSM) is identified as a weak point in integrating land use and transport modelling because it does not deal with land use explicitly. Another concern is that the sequential approach of the FSM creates a situation where no feedback is provided from one step to the other leading to inaccurate outputs. This research focuses on evaluating the (FSM) in order to assess the factors that influence its accuracy. The City of Istanbul in Turkey is used as the case study. The forecast period for this research is between 1996 and 2006. The methodology involves comparing the FSM daily traffic against the actual observed daily traffic derived from RTMS records. Using the 1997 Istanbul transport model developed by ITU and IMM as the basis for analysis, forecasts for population, students and employment as well as the planned road network are validated against actual data for 2006. A total of six FSM runs are then undertaken using different inputs, namely: projected data, actual data, disaggregated TAZs, actual highway network, new model parameters and finally a feedback mechanism. The results from these six model runs are assessed for accuracy by using the %RMSE statistic. The effect size of using each different input is then measured by means of the F-test and omega squared (ώ2). Results show that the population and employment forecasts are overestimated while those for students are underestimated. The planned road network significantly differs from the actual road network. The %RMSE accuracy assessment results for the FSM under different inputs is as follows; projected data (204.43%); actual data (195.02%); disaggregated TAZs (181.81%); actual highway (138.78%); new parameters (100.92%); and feedback (170.75%). In terms of effect size, the use of new parameters has the most significant effect (ώ2=.549) followed by application of a feedback mechanism (ώ2=.501). The use of disaggregation, actual highway and projected data has insignificant effect on the FSM. The land use and socio economic projections account for the greatest inaccuracy in the FSM while the use of new parameters and feedback improves the accuracy of the FSM significantly. A change in model inputs leads to a change in model outputs above and beyond the effect of extraneous factors, hence the variation in model outputs due to model inputs used is significant. The study recommends improving the quality of input data, application of feedback and estimation of the trip generation parameters at different time steps in the forecasting horizon based on anticipated changes in travel behaviour. Keywords: land use and transportation modelling; four step model; accuracy assessment.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page