A simple assessment of the expansion of American urban areas from 1970s to 1990s using the North American Landscape characterization data set

The triplicates of the North American Landscape Characterization (NALC) data set are used to assess the expansion of 10 large urban areas in the United States in two time periods since 1970s. For each of the ten cities, an image window covering the city is selected from the corresponding triplicate. For each site, a decision tree classifier is applied to classify the urban, suburban developed/residential areas from the other land use (the undeveloped/non-residential areas) for the three dates, and to detect the areas changed from vegetated/non-residential areas to developed/residential land surface types in the two time periods (from date 1 to date 2 and from date 2 to date 3). The classification maps are then examined and adjusted for accuracy using visual inspection of the images and the available ancillary maps. The adjusted maps are then used to calculate the percentage increases of the urban and suburban areas. The calculated percentages indicate that cities in western United States were experiencing faster expansion than cities in the eastern US. The increase of residential areas in the time period from 1973 to 1986 ranges from 93% for Las Vegas, NV to only 5% for Washington, DC. In the time period from 1986 to 1991, the increase spreads from more than 37% for Las Vegas, NV to only 3% for Washington, DC again. When percentage increases of the developed/residential areas are plotted against population of the city, significant linear relationships are found for most of the 10 cities.