Micro-simulation urban land use change modelling :the case of Ladprao, Bankok, Thailand

Abstract

This thesis focuses on modelling the spatial pattern of urban growth of Ladprao, a district of Bangkok, Thailand. The first part of the thesis reviews the urban growth and land use change problems in Bangkok as well as the current role of urban planning and its limitations, in order to provide the context of this study. A GIS-based cellular automata (CA) model has been developed, where the multinomial logistic regression (MNL) and multicriteria decision analysis (MCDA) methods have been integrated to identify the potential cells for development. Customized tools have been developed using a VBA macro within the ARCGIS environment to facilitate the implementation of urban simulation. The developed model has been applied to replicate the spatial pattern at the detail of the district level, focusing on the change of land from vacant to residential, commercial, and industrial during the period 1993 - 2001. Validation of the model has been undertaken through the comparison between the 2001 simulated and actual land use maps. The simulation was unsuccessful in reproducing the actual growth. In terms of the spatial agreement, the overall accuracy was about 30% (31.59% and 32.01% with MNL and MCDA respectively). In terms of urban morphology, the results showed the emergence of urban development in a space-filling pattern. Urban growth over discrete time-steps acted as a process of building accretion, appearing as a growing cluster around the existing development. In the actual pattern, the emergence of development was dispersed over the study area. The unexpected, but interesting, results of this observation have led to the conclusion of the three possible reasons; the inappropriateness of the CA approach to simulate the pattern of urban district level growth, the inability to include all significant development factors of the study site, and finally the distinctive characteristics of Ladprao and Bangkok area itself. Though the results are unpromising, the developed model can be considered as the first in the Bangkok area that attempts to be used as a spatial micro simulation tool operated at the district level. Future research work, if data permits, also suggests adding more development factors, adapting the agent-based modelling to the application, and extending the simulation to the growth of other areas of Bangkok both in the district and city level in order to help improve the understanding of Bangkok's growth.