Remote Sensing of Wetlands

This paper demonstrates three examples of applying remote-sensing technology to the delineation of wetlands in the United States. First, the results of a technique to identify roseau cane (Phragmites communis) using LANDSAT and aircraft MSS data are presented. Approximately 9700 ha. of Phragmites communis were identified in the Mississippi River delta using LANDSAT data. Second, the results of a technique to identify wetland communities, principally mangroves, in Florida using LANDSAT and aircraft MSS are presented. A cost analysis of the remote versus conventional ground-survey technique is provided. The LANDSAT MSS average classification accuracy for mangroves only versus all classes were 87 percent and 74 percent, respectively. The third example presents the results of a technique to delineate surface cover types in the southeast tidal/riverine system of the Savannah River. Both forested and nonforested wetlands were delineated using multi-temporal LANDSAT MSS data. The average classification accuracy for wetland classes only was about 82 percent and for all cover types was about 78 percent.