Modeling Land-Use Change in Complex Urban Environments

This article introduces how advances in the science of complex urban systems can potentially improve current land-use change (LUC) modeling, with discoveries of nonlinearity in urban development, spatial dependencies among neighboring places, and complicated social network interactions. We introduce recent development in statistical, computational, and spatial econometrics fields, adopting their general ideas and adapting some technical details for potential incorporation of those progresses into urban LUC modeling approach, which includes (1) Geographic Automata Systems for flexible modeling framework; (2) semiparametric spatial models to for spatial dependencies calibration; (3) high-performance statistical inference for enhanced forecasting capability; (4) parallelism for complex and high-resolution urban models; and (5) spatio-temporal analysis for LUC models to be more adaptive to different data types and functionalities. We then point out three critical challenges and development that urban LUC community need to address, which include (1) model integration with other relating models (such as transportation and economic models); (2) user-aware modeling interface in planning applications; and (3) a set of communal acknowledged modeling criteria. Finally, we present a case of University of Illinois Land-Use Evolution and Environmental Impact Assessment Model's application in Chicago metropolitan area, with emphasis on paralleled model construction and cloud-based interface.