DEVELOPMENT OF SALT MARSH CHANGE DETECTION PROTOCOL

USING REMOTE SENSING AND GIS

(Project Progress Slides)

ABSTRACT

The National Park Service has identified that the loss of salt marsh habitat in Jamaica Bay , Gateway National Recreation Area, warrants the development of a cost-effective, long-term salt marsh change detection protocol that can be applied in a broader coastal region of the northeast United States . This proposal addresses the creation of such a working protocol.  This project will evaluate the dynamic boundaries of salt marsh areas in the Jamaica Bay delineated by the NPS from historical aerial photographs. Based on the information obtained from the historical aerial photographs, we will conduct both computer-assisted classification and manual delineation of the salt marshes in the Bay area using high spatial resolution QuickBird-2 satellite image. The QuickBird2 satellite image possesses 0.65-meter spatial resolution on its panchromatic band and 2.5-meter spatial resolution on its multispectral bands. The spatial resolution and spectral coverage of the BuickBird-2 data can satisfy the salt marsh mapping requirements. The capability of repeated satellite data acquisition will meet the goal of salt marsh change detection and monitoring. We will first compare and evaluate the salt marsh mapping results between manual interpretation and digital classification of the QuickBird-2 data. We will also compare and evaluate the salt marsh mapping results obtained from manual interpretation of 2003 color infrared aerial photograph and from the QuickBird-2 data. With the comparisons and evaluations, we will be able to develop the change detection protocol using appropriate data. To support the manual and digital classification of salt marshes, we will conduct field works to establish GPS based Virtual Field Reference Database (VFRDB) for the Jamaica Bay area. The VFRDB contains typical GPS transect data that recorded the salt marsh variables such as the vegetation, hydrology, soils and sediments, as well as other observable indicators such as nekton and birds. The VFRDB also contains GPS-photos that recorded landscape specifics of the photo site in at least 4-campass directions (e.g., North, South, East, and West). The VERDB is part of the protocol that will provide benchmark data for long term monitoring and change detection analysis.

First fieldwork was conducted October 31, 2003. 

Final Project Report, July 2006. 

Please forward questions to yqwang@uri.edu.