 |
Remotely Sensed Indices of Land Cover for Watershed-Level Monitoring
|
Project Title: Remotely Sensed Indices of Land Use Intensity for Watershed-Level Monitoring
Research Locations: Jacques Cousteau, NJ
Last Update: 2-02-04
Rapid development along coasts can lead to increased input of contaminants into coastal waters. In order to properly protect water quality, coastal managers need to understand the pattern of development, both inland and near the shore, and have the tools to predict contaminant inputs. Traditional methods of monitoring development, including aerial photography and visual interpretation, are expensive and time consuming, preventing managers from regularly gauging the effect of rapid growth on coastal resources. This project tested newly available high and medium resolution remote sensing and Geographic Information Systems technologies for their utility and cost-effectiveness in monitoring the intensity of land cover and water quality.
The Approach
Researchers began by developing three reliable indices of human land cover - that is, three variables of
land cover that could be used to track changes in development over time. Those were: the amount of impervious surface, the amount of managed lawn and the amount of tree cover. They then obtained imagery data using three acquisition systems: Landsat ETM, a medium-resolution, public domain data system; IKONOS, an expensive, highresolution imagery system; and traditional aerial photography. Focusing on the impervious surface, managed lawn and tree cover indices, investigators explored several different image analysis techniques for quantifying human development, and compared them for quality and cost-effectiveness.
Results
Results showed that one of the different image analysis methods, the Self-Organizing Map Neural Network (SOM), provided the best estimates of the three land cover components. The Landsat ETMderived measures of impervious surface and managed lawn provided closely comparable estimates of land cover as IKONOS and aerial photography. These findings suggest that medium-resolution remotely sensed imagery such as Landsat ETM can provide cost-effective data for coastal watershed monitoring.
Setting land-use indices
Impervious surface is now widely adopted as a land cover index, and managed lawn is increasingly recognized as an additional component of land use that affects water quality.
A cost-effective alternative
Using medium-resolution imagery such as Landsat ETM can provide useful information for watershed managers at a fraction of the cost of IKONOS and aerial photography. While IKONOS can cost approximately $270 per square km, Landsat ETM provides accurate results for just $.02 per square km. Information can be obtained faster, more frequently and at minimum expense.
Looking Ahead
Researchers successfully applied the SOM technique to Landsat imagery covering the entire state of New Jersey. Future research will focus on improving estimates of the tree cover index.
|
Project Coordinator
|
Dr. R.G. Lathrop Rutgers University (732) 932-1580 lathrop@crssa.rutgers.edu |
|
Start -End Date:
|
09/01/2000 - 09/01/2002
|
|
NERR Reserve(s):
|
Jacques Cousteau, NJ
|
| For more information: |
Dolores Jalbert Leonard Phone: (603) 862-3685 Email: dolores.leonard@unh.edu |