2003 Seattle Annual Meeting (November 2–5, 2003)

Paper No. 17
Presentation Time: 1:30 PM-5:30 PM

SOARS (3): MONITORING LANDCOVER CHANGES FROM LANDSAT DATA IN THE CUYAHOGA RIVER WATERSHED OVER A 27 YEAR PERIOD


LARSEN, Karen, Geography, Kent State Univ, Kent, OH 44242, DEGENHARDT, Lindsey, Geology, Bowling Green State Univ, Bowling Green, OH 43402, DUNHAM, Susan, Department of Geography, Southwest Texas State Univ, San Marcos, 78666 and FEDDERS, Mark David, Geography, Univ of Toledo, Toledo, OH 43606, klarsen@kent.edu

The Landsat Earth Observing Satellites have provided researchers with the longest continuous space-based remote sensing dataset in the world. We have used this resource to develop landcover maps to serve as an educational platform and research tool for understanding man-land interactions that continue to influence the environmental health of the Cuyahoga River Watershed in northeast Ohio. The results presented in this paper are part of the larger SOARS (Scientific Outreach and Applications of Remote Sensing) project sponsored by OhioView, NASA, and the Cuyahoga River Remedial Action Plan. All landcover maps and statistics will aid local officials in making informative decisions on land management in the watershed.

Landcover maps were created at approximately 5 year intervals starting in 1976. For each time slice a single category of "urban" was extracted and digital change detection was performed between each dataset to determine the effect of sprawl on the watershed. By county, while the population of Cuyahoga County, the home of Cleveland, decreased by 7.5% since 1980, the population increased dramatically in the adjacent counties that are in the watershed (Portage, 11.9%; Summit, 3.5%; Geauga, 22%). In contrast the landcover that is associated with urban areas in the Cuyahoga watershed has increased by almost 50% in the lower watershed, demonstrating that more land is being utilized per individual than 20-30 years ago.

Starting with the 1984 image, unsupervised classification was performed using ERDAS Imagine. Sixty spectral clusters extracted from each dataset were aggregated into 3 different map types: (1) general landcover (8 categories); (2) percentage of agriculture and grass; and (3)percentage imperviousness. These landcover maps document the spawling of Cleveland. Importantly, these landcover changes, especially increased imperviousness, and general urban sprawl, have been correlated with increased runoff into the Cuyahoga River and its tributaries and increased NPS pollutant concentration.