A Decade of SLEUTHing: Lessons Learned from Applications of a Cellular Automaton Land Use Change Model

This paper reviews the numerous and various applications of the SLEUTH urban growth model over the last decade. SLEUTH is a Cellular Automaton-based urban growth model that uses historical geospatial data for calibration of its parameters. Applications have covered the major cities of the United States, including Detroit, Chicago, New York, Washington, San Francisco and Albuquerque. The model has also been applied in the Netherlands, Portugal, South America, Africa and Australia. Applications have examined the likelihood of urban encroachment on waste disposal sites, the generation of planning scenarios for public decision-making, and the expansion of informal settlements in Yaoundé, Cameroon. Regions covered have varied from a single small town to a whole multi-city urban region, and spatial resolutions have gone from tens of meters to kilometers. The paper assembles for the first time the calibration results from the model applications, and attempts an inter-city comparison, expanding on Silva's concept of the “DNA of regions.” This is possible because the successful calibration of the SLEUTH model from historical land use data gives a set of five parameters that capture the nature of the urban growth in a region. The discussion also summarizes the various lessons learned, especially in how to best calibrate the model. Recent results have shown not only that model performance can be significantly improved using high performance computing, but also that adaptations are possible for searching the parameter space that vastly improve how well the model forecasts growth and change into the future.