摘要：An impervious surface is any surface that prevents water from infiltrating the ground. As impervious surface area increases within watersheds, stream networks and water quality are negatively impacted. The Multi-Resolution Land Characteristic Consortium developed a percent impervious dataset using Landsat imagery as part of the 2006 National Land Cover Database. This percent impervious dataset estimates imperviousness for each 30-meter cell in the land cover database. The percent impervious dataset permits study of impervious surfaces, can be used to identify impacted or critical areas, and allows for development of impact mitigation plans; however, the accuracy of this dataset is unknown. To determine the accuracy of the 2006 percent impervious dataset, reference data were digitized from one-foot digital aerial imagery for three study areas in Arkansas, USA. Digitized reference data were compared to percent impervious dataset estimates of imperviousness at multiple 900m2 , 8,100m2 , and 22,500m2 sample grids to determine if accuracy varied by ground area. Analyses showed percent impervious estimates and digitized reference data differ modestly; however, as ground area increases, percent impervious estimates and reference data match more closely. These findings suggest that the percent impervious dataset is useful for planning purposes for ground areas of at least 2.25ha.
