Modelling Uncertainties in Long-Term Predictions of Urban Growth: A Coupled Cellular Automata and Agent-Based Approach

Modelling the growth of urban settlements is of considerable interest for different applications, amongst which integrated flood management. This study aims at modelling urban growth for a long time horizon up to 2100 and to integrate the model outcomes with a hydrological model for the same time horizon. Forecasting land-use change over such time frames entails very significant uncertainties. In this regard, the main focus of this paper is attributed to the handling of uncertainty in an urban growth model. To this end, we examine a Monte Carlo Simulation method, which is integrated in the proposed urban growth model. Transition probabilities for each non-urban cell are estimated by a coupled Cellular Automata-Agent-Based ap-proach. The results help to handle uncertainty over long time horizons and to assess the increment in degree of uncertainty at every time-step.