A Methodology to Characterize
Landscape-scale Urban Growth from Satellite Imagery
Emily
Hoffhine Wilson, Research Assistant, James D. Hurd, Research Assistant, Center
for Land use Education And Research (CLEAR), NAUTILUS RESAC, University of Connecticut;
Daniel L. Civco, Associate Professor, University of Connecticut
Suburban and urban growth radically is changing New England’s landscape. However,
little research has been conducted to systematically map, describe and quantify
these changes. While the results of growth are evident as new houses, new shopping
centers, more cars, etc., consistently classified geospatial data on where and
to what extent growth is occurring are lacking. Researchers at the Center for
Land use Education and Research (CLEAR) as part of the Northeast RESAC (Regional
Earth Science Applications Center) have developed a model in ERDAS Imagine that
quantifies and characterizes urban growth. Satellite images are used to create
at least two dates of land cover data that minimally separate developed land,
undeveloped land, and water. The land cover data are inputs to the urban growth
model, which, after several steps, creates an objective and repeatable urban
growth image consisting of five types of growth as well as no change classes.
The five types of urban growth are: infill, expansion, isolated, linear branching,
and clustered branching. The output raster image can be converted to a grid
and used as a GIS dataset to show both the type and location of new developments
across a given area. It is suggested that this approach, using readily available
historic satellite imagery, effectively can be used to help understand the dynamics
of land use change over large geographic areas.
Close Window