Simulation of land use changes by capturing the different impacts of rail transit in both mother city and new towns

Rail transit system plays an important role in connecting the urban center, new towns and suburbs in the metropolitan area. Exploring the methods of integrating rail transit and land use in different locations is essential to improve trip rates and promote transit-oriented development (TOD). Taking Wuhan, China, as a case study, this research uses a cellular automaton-based random forest (CA-RF) model to simulate land use change surrounding rail transit by capturing the impact differences in mother city (MC) and new towns (NT). An impact area identification model is proposed to recognize the most sensitive threshold of land use change in both MC and NT stations, which are 1000m and 1200m, respectively. The characteristics of land use change from 2010 to 2020 are analyzed and compared. The CA-RF model calibration shows that transportation and spatial accessibility present wide variations in driving land use change surrounding the MC and NT stations. The high accuracy of CA-RF model, which is 88.1% for MC and 82.1% for NT, indicating its effectiveness in quantifying the non-linear relationship of land use change and spatial attributes. The forecasted land use maps for proposed rail stations provide a reference for governments and planners in policy intervention and decision-making. This research framework can be applied to other metropolitans to explore the regulations of land use development and promote transit-oriented sustainable development.