Developing a multi-network urbanization model: A case study of urban growth in Denver, Colorado

Abstract

Urbanization is an important issue concerning diverse scientific and policy communities. Computational models quantifying locations and quantities of urban growth offer numerous environmental and socioeconomic benefits. Traditional urban growth models are based on a single-algorithm fitting procedure and thus restricted on their ability to capture spatial heterogeneity. Accordingly, a GIS-based modeling framework titled multi-network urbanization (MuNU) model is developed that integrates multiple neural networks. The MuNU model enables a filtering approach where input data patterns are automatically reallocated into appropriate neural networks with targeted accuracies. We hypothesize that observations classified by individual neural networks share greater homogeneity, and thus modeling accuracy will increase with the integration of multiple targeted algorithms. Land use and land cover data sets of two time snapshots (1977 and 1997) covering the Denver Metropolitan Area are used for model training and validation. Compared to a single-step algorithm – either a stepwise logistic regression or a single neural network – several improvements are evident in the visual output of the MuNU model. Statistical validations further quantify the superiority of the MuNU model and support our hypothesis of effective incorporation of spatial heterogeneity.