PREP Publications

Abstract

Estimates of impervious surface acreage for 2010 were generated for the 59-town region covered by the Piscataqua Region Estuaries Partnership (PREP). The project extended previous work done in the region for the years 1990, 2000, and 2005 and relied on the same satellite-based data sources and image processing methodologies. As a result, standardized impervious surface estimates are now available for a 20-year time period in the PREP region. The current project mapped impervious surfaces (buildings, pavement, etc.) based on a Landsat 5 Thematic Mapper (TM) image acquired on April 24, 2010. Processing used both traditional and sub-pixel image classification techniques, as described in previous efforts (see Justice and Rubin, 2006 and Justice and Rubin 2003 for a complete processing description). The current study utilized comparable satellite imagery, and applied consistent techniques to map the PREP area for 2010. It should be noted that since the completion of the 2005 project, the U.S. Geological Survey (USGS) has modified its image offerings and now only serves TM data resampled using cubic convolution techniques. (Resampling algorithms are applied as part of the image registration processing, and describe the way output values are assigned to individual image pixels.) This is an important consideration because our previous sub-pixel classifications were conducted on data sets processed using nearest neighbor resampling (which is the recommended approach). While the classification process can be completed using data resampled using alternative techniques, the results may not be as reliable. Since the three previous data sets (1990, 2000, and 2005) were developed using images processed using the preferred nearest neighbor method, and because it is unlikely that features would revert from impervious to non-impervious status, we opted to use the 2005 impervious surface raster as the starting point for the current iteration. Accordingly, the 2005 impervious surface data was used to mask the 2010 TM image with any remaining, unmasked pixels classified using the techniques described in the documents referenced above. A second departure from previous processing streams is that the TM data is now delivered as a georeferenced, terrain-corrected file. This eliminated the requirement for GRANIT to perform these steps locally. The data set has been archived in the GRANIT GIS clearinghouse, thereby making it available to the coastal resource community as well as the general public. The data are appropriate for watershed and subwatershed level characterizations. Users are discouraged from accessing them to support larger scale mapping and applications.

Publication Date

10-2011

Publisher

New Hampshire Estuaries Project

Document Type

Report

Share

COinS