Mapping and analyzing change of impervious surface for two decades using multi-temporal Landsat imagery in Missouri

Human population growth and associated sprawl has rapidly converted open lands to developed use and affected their distinctive ecological characteristics. Missouri reflects a full range of sprawl characteristics that include large metropolitan centers, which led growth in 1980s, and smaller metropolitan and rural areas, which led growth in 1990s. In order to study the historical patterns of sprawl, there is a need to quantitatively and geographically depict the extent and density of impervious surface for three time periods of 1980, 1990, and 2000 for the entire state of Missouri. We mapped impervious surface using Sub-pixel Classifier (TM), an add-on module of Erdas Imagine for the three time periods, where impervious surface growth was derived as the subtraction of impervious surface mapped from the different time periods. Accuracy assessment was performed by comparing satellite derived impervious surface images with ground-truth acquired from high resolution air photos. Results show that during 1980-2000, 129,853 ha of land were converted to impervious surface. Sprawl was prominent on urban fringe (within the urban boundaries) during 1980s with 23,674 ha of land converted to impervious surface compared to 22,918 ha in 1990s. There was a temporal shift in the rural landscapes (outside the urban boundaries) in the 1990s with 48,079 ha of land converted to impervious surface compared to 35,180 ha in 1980s. Major findings based on analysis of the impervious surface growth include: (i) new growth of impervious surfaces are concentrated on areas with 0.5-1.0% road cover; (ii) most new growths are either inside or close to urban watersheds; and (iii) most new growths are either inside or close to counties with metropolitan cities. This research goes beyond the usual hot spots of metropolitan areas to include rural landscapes where negative impact was exerted to the ecosystem due to the low density development and larger affected areas. (C) 2012 Elsevier B.V. All rights reserved.