Defining functional urban regions in Bahia, Brazil, using roadway coverage and population density variables

• A spatial analysis-based method for the definition of functional urban regions.
• The ESDA technique performs better than SKATER for defining urban agglomerations.
• Both techniques indicate the existence of two large urbanized clusters.
• Findings do not totally match the single official state metropolitan region.
• The approach and related findings are important for adequate planning practices.

The concept of Functional Urban Regions (FURs), also called Metropolitan Regions (MRs), is not simple. It is clear, though, that they are not simply a combination of adjacent municipalities or areas. Different methods can be used for their definition. However, especially in developing countries, the application of some methods is not possible, due to the unavailability of detailed data. Alternative approaches have been developed based on spatial analysis methods and using variables extracted from available data. The objective of this study is to compare the results of two spatial analysis methods exploring two variables: population density and an indicator of transport infrastructure supply. The first method regards Exploratory Spatial Data Analyses tools, which define uniform regions based on specific variables. The second method used the same variables and the spatial analysis technique available in the computer program SKATER – Spatial ‘K’luster Analysis by Tree Edge Removal. Assuming that those classifications of regions with similar characteristics can be used for identifying potential FURs, the results of all analyses were compared with one another and with the ‘official’ MR. A combined approach was also considered for comparison, but none of the results match the existing MR boundaries, what challenges the official definitions.