Sea Grant Project R/ET-5

"Development of High Resolution Satellite and GIS Techniques for Monitoring and
Assessment of Coastal Ecosystems"



 
 

Principal Investigator Klemas, Victor V.
Co-principal Investigator Field, Richard
Associate Investigator 1 Hall, William
Associate Investigator 2 Yan, Xiao-Hai


OBJECTIVES:

(1) To develop techniques for monitoring landscape level environmental indicators using new high-resolution satellite imagery for providing quantitative estimates of wetland and estuarine habitat conditions and trends. (2) To determine what ancillary data and image analysis techniques are required to reliably detect critical wetland health indicators, such as biomass change, fragmentation, riparian buffers, invasive species, etc. if medium resolution satellite imagery (Landsat TM) is combined with high resolution imagery (IKONOS) (3) To work with the Marine Advisory Service to prepare websites for high school students, teachers and the general public illustrating the history of Delaware's tidal wetlands and how the new remote sensing techniques help coastal managers to restore, enhance and protect them. METHODOLOGY: We will analyze high-resolution imagery of several well-documented sites, for which we have much ancillary data (e.g. Blackbird Creek and St. Jones River National Estuarine Research Reserve Sites, Milford Neck Wildlife area). About half the required satellite/aircraft imagery is already on hand; the rest will be obtained from NRL and NASA. First we will analyze the spatial variability of our test sites to determine if the spatial resolution of the imagery is suitable for detecting the relevant coastal features. Next we will perform multispectral analysis using the high-resolution images. In addition to wetland loss, changes in critical environmental indicators, such as biomass, will be studied. We will evaluate the accuracy of the multispectral analysis results individually and when combined with ancillary data in a GIS. Ancillary data layers will be used for interpreting the multispectral analysis maps and comparing their accuracy to existing maps and medium resolution satellite results.
During the first year we have performed spectral and spatial analyses of existing satellite/aircraft imagery to determine the spectral characteristics of the wetland classes and features at our test site. We have acquired leaf-on high resolution satellite/aircraft data from overflights and contemporaneous field observations on the ground. During the second year, these first year multispectral observations will be analyzed to map wetland health attributes. A preliminary analysis of map accuracy will be performed. The second year of leaf-on high resolution satellite/aircraft data will be acquired in association with intensive ground observations. During the third year we will perform the final accuracy assessment and will have wetland managers and researchers evaluate the results. An educational computer module in software such as ArcExplorer and/or PowerPoint, of the wetland remote sensing techniques will be completed and tested for use in high schools and regional institutions. Reports and papers will be prepared each year.
RATIONALE: Many coastal features are small and have narrow, elongated shapes, such as mud flats, riparian/shoreline buffers, monospecific wetland stands or wetlands existing along edges of bays and rivers. Therefore they are difficult to detect with 20-30 meter resolution satellite imagery. Furthermore, wetland and water pollutants have complicated spectral reflectance signatures and are difficult to identify with only the available 3 to 6 spectral bands. Recently new satellites are being launched which are providing imagery with spatial resolutions of 1 to 4 meters (e.g. IKONOS). There is an urgent need to develop mapping techniques, which use the spectral and spatial content of these high-resolution images, to enable researchers and managers to monitor wetland habitat cost-effectively. There is also a need to get students educated about the history of Delaware?s wetlands and these new technologies and to demonstrate to the public how these techniques are helping us to restore, enhance and protect coastal ecosystems. PROJECT RESULTS: In our preliminary research we have shown that high-resolution satellite or aircraft imagery can be used to classify wetlands and detect habitat features (e.g. fragmentation, riparian buffers, mudflats, invasive species, biomass) important to wildlife, including migratory birds. Specifically, at 1 meter to 4 meter resolutions we have been able to differentiate wetland types by their spatial, as well as spectral characteristics. It seems that combining spatial and spectral analyses of high-resolution satellite imagery will produce an effective approach for monitoring wetland type, extent and health. We found that the Modified Soil Adjusted Vegetation Index give best results for wetland biomass change determination. Our improved method for mapping wetland biomass changes requires fewer field samples and corrects for seasonal, tidal and soil reflectance variations without the need for local correction factors. ACCOMPLISHMENTS: High resolution multispectral (4m) and panchromatic (1m) IKONOS satellite imagery has been received from NASA for all three test sites. The imagery has been processed and multispectral analysis initiated. The test sites have been segmented between tidal and non-tidal areas. The non-tidal areas are further segregated into upland, forested wetland, and emergent wetland areas with the aid of such ancillary data as the NWI and Delaware?s SWAMP data sets. Historical thematic mapper scenes from 1984 to 1997 are being subset for analysis of the study areas using the Modified Soil Adjusted Vegetation Index to identify areas of wetland change. In addition, a historical examination of the tidal wetlands of the Delaware River and Estuary, which shows that all of the tidal wetlands have been impacted by human use, has been completed and is being prepared for presentation to the public on the World Wide Web. BENEFITS: The new remote sensing and GIS techniques being developed provide reliable information on wetland habitat status and trends to local and regional agencies and researchers. The high-resolution imagery enables us to study wetlands fragmentation, riparian buffers, impervious areas, invasive species, plant biomass and other landscape-level environmental indicators which heretofore were difficult to detect from satellite altitudes. The quality of the high-resolution wetland and related data products will be evaluated by coastal planners, managers and researchers to see how well they can be integrated with other data in a GIS to more effectively monitor coastal ecosystems. If the results of this evaluation are positive the new techniques will form the basis for designing a system for long-term monitoring of Delaware Bay habitats.
The educational computer presentation will be such that its format could be adopted by teachers of marine sciences, earth sciences and physical sciences. SG Marine Advisory Service personnel will help to implement this outreach component.
 
MARINE ADVISORY SERVICE INTERACTION: The Marine Advisory Service (MAS) will guide us in the application of our results to coastal resource  management. The research project will be carried out in conjunction with the Delaware Chapter of The  Nature Conservancy, Delaware Wildlands, Inc., the Delaware Department of Natural Resources and environmental Control (see letter of support) and the Delaware National Estuarine Research Reserve Program. All four groups will make field sites available to our team and help in the field data collection.
Other regional projects which have been collaborating with us and will benefit from this research include:
1) The FWS GAP Analysis Project (FWS and DE, MD, NJ)
2) The Delaware Coastal Management Program (NOAA and State of Delaware)
3) The Delaware Estuary Program (EPA and DE, NJ, PA)
4) The Delaware Inland Bays Estuary Program (EPA and State of Delaware)
The Marine Advisory Service (MAS) will also help us develop an educational computer presentation module for high school students, teachers and the public illustrating how the new remote sensing techniques will help protect, enhance and restore coastal ecosystems. We will explore commercial software packages to determine the most appropriate media for the presentation. Options include using an off-line web browser, Microsoft's PowerPoint, or Adobe Acrobat. The presentation material will be tested by the MAS at local schools and institutions. A final presentation version will be made available on the Internet.




back to the Sea Grant Research

Last modified: October 17, 2001
Xuebin Zhang  --- xbzhang@newark.cms.udel.edu