How current and future urban patterns respond to urban planning? An integrated cellular automata modeling approach

While many publications predict future urban scenarios, few have deliberated the impact of issued urban planning on scenario prediction. We propose a planning-constrained model (named PCGA-CA) that integrates cellular automata (CA) and genetic algorithm (GA) to simulate current and future urban patterns under the spatial constraints of urban planning. The planning regulations include three types: fully allowed area (FAA), partially allowed area (PAA), and strictly prohibited area (SPA), where we propose a planning implementation parameter (PIP) to represent the stringency in PAA. Under different PIPs, we apply the PCGA-CA model to simulate the 2015 urban patterns and predict the 2030 and 2045 scenarios for Ningbo city, China. The results show that the regulations substantially affect the simulation accuracy and urban pattern. As the planning regulations become less stringent, the accuracy decreases from 90.3% to 89.4% and the urban pattern becomes less compact. In particular, the urban pattern is the most compact when the regulations are not imposed. The PCGA-CA predicts the quantity and location of illegal urban development, and identifies spatially varying urban growth across planning regulations. For the same year, the urban patterns with different PIPs illustrate substantial differences in landscape metrics. The simulations of the current urban pattern should help urban planners and local authorities assess past implementations of urban planning, while the scenario predictions can offer a view of the future by evaluating the consequences of different planning regulations