2005 BES Annual Meeting Presentation and Poster Abstracts
Remotely Sensing Analysis of Urban Land Cover Components
Abstract: Urban landscapes are composed of diverse materials (concrete, asphalt, metal, water, vegetation and soil) arranged by humans in complex ways to accommodate a variety of uses. The relative percentage of these materials has important implications for the larger environmental system processes controlling the cycling and processing of water, nutrients, carbon, and energy, as well as issues related to human quality of life and conserving biodiversity. We examined the utility of the several different image analysis techniques with remotely sensed imagery at multiple scales to estimate three key land cover components (namely, impervious surface, managed grass and tree cover) in an urban/suburban setting. Results for digital aerial photography (1m Ground Resolution Cell - GRC), IKONOS (4m GRC) and Landsat Thematic Mapper (30m GRC) were compared for a coastal watershed in the state of New Jersey, USA. Sub-pixel analysis of the coarser scale Landsat image data was possible with a root mean square error of ±10 to 15% for impervious surface and ±12 to 22% for grass and tree cover. This model was then successfully applied to provide statewide estimates of the 3 urban land cover components. Using this sub-pixel analysis approach we were able to separately quantify the grass vs. tree components in urban vegetation, which is of major importance to the study of the urban/suburban ecosystems as well as watershed assessment.