NOAA/NERRS REMOTE SENSING APPLICATION PROJECT:USING BIOMASS TO MONITOR CHANGES IN WETLAND HABITAT

摘要

Population growth in coastal regions suggests continuing losses of wetland and adjacenthabitats, as waste loads and competition for limited space and resources increase. Coastalwetlands and submersed habitats are being destroyed by erosion, dredge and fill, impoundments,toxic pollutants, eutrophication, excessive turbidity and sedimentation. Yet salt marsh grasses,mangroves, macroalgae and submersed grasses are essential as nourishment and animal habitat.Many marine finfish, shellfish and migratory birds depend on these coastal habitats for theirsurvival. Continued loss of these wetlands may lead to the collapse of some coastal ecosystemsand associated fisheries; therefore, documentation of the loss or degradation of coastal wetlands isneeded for their conservation and effective management.Remote sensors have been used successfully for several decades to map physical andbiological properties of the open ocean and large upland areas. However, coastal wetlands andestuaries require much finer spatial and temporal resolution and present a challenge to sensorscurrently deployed on satellites and aircraft. For example, many coastal ecosystem are patchyand have narrow, elongated shapes, such as mud flats, riparian/shoreline buffers, nonspecificwetland stands or wetlands existing along edges of bays and rivers. Therefore, they are difficultto detect with the 20-30 meter resolutions of moderate resolution satellites such as Landsat orSPOT. High-resolution satellites, like IKONOS, can detect small wetland sites but are lackingspectral bands, especially in the mid-infrared, needed to discriminate wetland types and forestsfrom wooded wetlands. Also the limited coverage of high-resolution satellites necessitates theuse of many images at high cost, if large coastal areas are to be surveyed. Data from airbornehyperspectral scanners are expensive and require highly specialized analysts.To overcome these problems new sensors and analysis methods are being developed.Spectral analysis techniques will benefit from new hyperspectral mappers being deployed onsatellites. New image analysis approaches include multi-resolution segmentation, form andtexture classification, contextual classification, (proximity, direction, adjacency, inclusion), fuzzylogic classification, neural network analysis, biomass change detection, etc. Also, smallremotely-controlled helicopters are becoming available for field data collection.To help coastal managers and scientists choose the most appropriate imagery and analysistechnique, the NOAA National Estuarine Research Reserve System (NERRS) Program hasselected several NERRS sites for evaluation of airborne and satellite sensors for monitoring andmapping emergent wetlands and submerged aquatic vegetation. As part of this Remote SensingApplication Assessment Project (RESAAP), our team is investigating the cost-effectiveness ofseveral remote sensing techniques, including airborne hyperspectral, airborne digital multimulti-spectral, high-resolution (e.g. IKONOS) and moderate resolution (e.g. Landsat TM) satelliteimagery for observing health-related properties of wetlands and estuaries. (Table 1)