Measuring urbanization in United States, from the first trading post to the Metropolitan Areas (1790-2000)

We describe here an original methodology for building an harmonized data base relative to urban populations in United States, from 1790 to 2000 (one point every 10 years, i.e. 22 dates). This method is based on two theoretical principals. The first one consists in the using of a spacetime referential in order to build agglomerations, by aggregating cities or towns from 1870 (inside a one-hour limit at each period of time). The second consists in the using of a functional approach of urban entities as soon as possible, i.e. from 1940. We have two different types of functional areas in our data base, metropolitan and micropolitan. The metropolitan areas definition has changed several times (Standard Metropolitain Areas in 1949, Standard Metropolitan Statistical Areas in 1958 and Statistical Metropolitan Areas in 1983). We elaborate different criteria to correct some perimeters, in order to obtain a temporal coherent trajectory of populations and surfaces. As these metropolitan areas are very few (less than 400 in the 2000 census), we use the definition of micropolitan areas (given in 2000 by the Census Board) to build such areas in prior decades, up to 1940. The different maps that are built from our harmonized data base (urban frontier, urban patterns and surfaces) give a global and coherent vision of the U.S. urbanization through the last two centuries. Lastly, we compare urban growth and inequality size indicators computed with our data base to other computed with the Census definition or with other data bases built by researchers. The results are sometimes very different, which confirm the major impact of the delineation methodology in urban growth measures.