Modelling urban dynamics as driving forces of land use change in port cities

The urban environment is a complex system where general urban morphology patterns emerge from local interactions of many human and natural agents. The nature of these individual interactions is difficult to examine from empirical data due to their auto-correlation and to path-dependence in the evolution of urban systems. For the same reasons, calibrations of urban land-use and transport interaction models based on cellular automata (CA) often rely on manual methods. However, this paper shows that a gradient-descent algorithm can be used to calibrate an agent-based CA model of urban areas, wherein cells are characterised by the levels of each activity type they host rather than by categorical states representing only their dominant land-use. This automated process allows efficient calibration of the model for multiple urban settlements, enabling cross-sectional analysis of whether an observed relationship is unique to a settlement due to its path-dependence, or has similarities with results from other settlements. This paper uses the model to investigate the nature of urban dynamics within port-cities by examining the interactions between geographic features, land-use, and transport in 46 port-based settlements and 10 non-port settlements across Great Britain. It shows that port activities had most pronounced effects on smaller settlements, with urban dynamics in larger settlements affected little by the presence of port. There were though variations in urban dynamics between larger settlements with respect to the effects of their retail and manufacturing activities on other land-use. The automated calibration of agent-based urban CA models in this paper enabled the quantification of these interactions in multiple settlements, providing the ability to look beyond the uniqueness of individual case studies to find general patterns of interaction. This study therefore provides important insights into the driving forces of land-use changes in port-cities which could help facilitate future sustainable transport and land use development.