Imperviousness, NLCD 2001

Imperviousness for the Baltimore MSA from the NLCD 2001 impervious dataset. Impervious surfaces refers to impenetrable surfaces such as rooftops, roads or parking lots. Quantification of imperviousness can offer a relatively objective measure of urban density and provide a forum for its classification. For NLCD 2001, imperviousness was chosen as the surrogate for the urban intensity classification in an effort to improve the precision of urban characterization used in the original NLCD 1992. Modeling empirical relationships between imperviousness and Landsat data is accomplished using regression tree techniques. Several one-meter digital orthophoto quadrangles are used for each Landsat scene to derive reference impervious data needed for calibrating the relationships between percent imperviousness and Landsat spectral data, which are then modeled using a commercial regression tree algorithm called Cubist. The models are then applied to all pixels in a mapping zone to produce a per-pixel estimate of imperviousness in urban areas (Yang et al., 2002). This procedure quantifies the spatial distribution of impervious surfaces as a continuous variable for urban areas from 1 to 100%, and offers a consistent and repeatable method to characterize urban areas across the Nation. This data layer is then masked to ensure only urban pixels are included and thresholded (Table 1) into NLCD 2001 urban classes and inserted into the land cover. Imperviousness information will be available as an independent product of NLCD 2001. The National Land Cover Database 2001 for mapping zone 60 was produced through a cooperative project conducted by the Multi-Resolution Land Characteristics (MRLC) Consortium. The MRLC Consortium is a partnership of federal agencies (www.mrlc.gov), consisting of the U.S. Geological Survey (USGS), the National Oceanic and Atmospheric Administration (NOAA), the U.S. Environmental Protection Agency (EPA), the U.S. Department of Agriculture (USDA) Forest Service (USFS), the National Park Service (NPS), the U.S. Fish and Wildlife Service (FWS), the Bureau of Land Management (BLM) and the USDA Natural Resources Conservation Service (NRCS). One of the primary goals of the project is to generate a current, consistent, seamless, and accurate National Land cover Database (NLCD) circa 2001 for the United States at medium spatial resolution. For a detailed definition and discussion on MRLC and the NLCD 2001 products, refer to Homer et al. (2003) and http://www.mrlc.gov/mrlc2k.asp. The NLCD 2001 was created by partitioning the U.S. into mapping zones. A total of 66 mapping zones were delineated within the conterminous U.S. based on ecoregion and geographical characteristics, edgematching features and the size requirement of Landsat mosaics. Mapping zone 60 encompasses whole or portions of several states in the mid-Atlantic region, including the states of New Jersey, Delaware, Maryland, Pennsylvania, Virginia, and the District of Columbia. Questions about the NLCD mapping zone 60 can be directed to the NLCD 2001 land cover mapping team at the USGS EROS Data Center (EDC), Sioux Falls, SD [phone] or [e-mail]. This is part of a collection of 221 Baltimore Ecosystem Study metadata records that point to a geodatabase. The geodatabase is available online and is considerably large. Upon request, and under certain arrangements, it can be shipped on media, such as a usb hard drive. The geodatabase is roughly 51.4 Gb in size, consisting of 4,914 files in 160 folders. Although this metadata record and the others like it are not rich with attributes, it is nonetheless made available because the data that it represents could be indeed useful. This is part of a collection of 221 Baltimore Ecosystem Study metadata records that point to a geodatabase. The geodatabase is available online and is considerably large. Upon request, and under certain arrangements, it can be shipped on media, such as a usb hard drive. The geodatabase is roughly 51.4 Gb in size, consisting of 4,914 files in 160 folders. Although this metadata record and the others like it are not rich with attributes, it is nonetheless made available because the data that it represents could be indeed useful.