ASCE will be sunsetting the Civil Engineering Database (CEDB) in the coming months. We encourage you to start using the ASCE Library to perform content searches. In the ASCE Library, you can save your search results as well as receive email content alerts when new content relevant to your area of interest is published. Try it out today at https://ascelibrary.org.

 

Evaluation of Satellite-Based Impervious Surface Data for the Mid-Atlantic Region Using an Independently Collected, High-Resolution Validation Database

by John W. Jones,



Document Type: Proceeding Paper

Part of: World Environmental and Water Resources Congress 2008: Ahupua'A

Abstract:

Satellite-based estimates of impervious surfaces are being produced for the entire United States as part of the U.S. Geological Survey's (USGS) Landsat-based National Land Cover Database (NLCD). In the Chesapeake Bay region, these impervious surface data support water-quality modeling, restoration, and planning. USGS research developed a protocol for collecting high-resolution impervious surface data from aerial photographs and other imagery sources, for the purpose of assessing the accuracy of satellite-based maps of impervious surfaces. The validation data collected using this protocol, were applied to the assessment of NLCD impervious surface data accuracy across a gradient of urban development in the ChesapeakeBay/Mid-Atlantic region. The protocol specifies a minimum mapping unit of any impervious surface that is ten square meters or larger in a test area. For the study region 240 such test areas, each measuring 500 m by 500 m on a side, were randomly selected. Initial comparison of NLCD estimated impervious surface area by chip showed that between 33 and 46 chips were collected for each of five development categories (none, rural, exurban, suburban, dense suburban, and urban). Since the analysis was conducted within a GIS environment and the protocol requires extensive attribution of data sources, additional ancillary information available for each validation data point allowed us to provide potential explanations for observed errors. For example, some cases where relative error was greater than 100% could be explained by temporal discontinuities between the date of capture of the satellite image and validation source data in rapidly changing areas. Other errors were created by vegetation cover obscuring impervious surfaces and by other factors that bias the satellite processing algorithms. Through such analysis we evaluated the errors, strengths, and weaknesses of the NLCD product and can suggest appropriate uses for regional, satellite-based data in rapidly developing areas of environmental significance.

Subject Headings: Urban and regional development | Satellites | Data analysis | Validation | Data collection | Errors (statistics) | Databases | Mid Atlantic | United States

Services: Buy this book/Buy this article

 

Return to search