NASA/University of Maryland
Coastal Marsh Project

photograph of salt marsh with large open water bodies

Methodology

The methodology of the Coastal Marsh Project involves the processing of remotely sensed TM image data and thematic GIS data layers through a model conceived by Dr. Michael Kearney.   The result is a Marsh Surface Condition Index GIS data set where coastal marshes are classified into a range from healthy to completely deteriorated wetlands.


Contents

Research Rationale

Data Base Components

Classification Process


Research Rationale

Geomorphology

Coastal salt marshes deteriorate due to many causes, but essentially through one process:  the surface of the salt marsh drops below the level of the local tides, and the salt marsh can no longer maintain its delicate balance with the salt water that brings nutrients and sediment to build its footing. As a result, the marsh deteriorates into open water.

The causes differ from place to place:  the extraction of fluids from beneath the substrate, as in Louisiana, may cause the marsh to drop in elevation relative to the local sea level;  a decline in sediment because of damming of a contributary stream may deprives the marsh of material to build elevation after the tides wash it away;  the deposition of sediment from urbanization of an area can change the hydrology of the marsh too quickly for the plants to adjust;  even the rebound of a formerly-glaciated area can change that balance between elevation and water levels.

Some researchers believe that wetlands, like barrier islands, are naturally dynamic entities, and that they can adjust continuously to changes in their elevation by "migrating" upland: as the water level rises, the plants begin to root at higher elevations. Some believe that wetlands erode from the edges and slowly retreat to a smaller acreage.

Dr. Michael Kearney of the University of Maryland has worked for many years on the question of salt marsh deterioration, and has proposed a model in which the marsh surface erodes from within, as the balance of elevation and substrate is disturbed. That is,  he believes that although all salt marshes are necessarily very wet, there is a correlation between the amount of water in a given area and the integrity of the surface of the marsh.  Marshes become waterlogged and the open water areas begin to expand, causing a cascade effect as the marsh deteriorates.

Policy issues

Since the 1970s, when wetland conversion was at an all-time high, and the Environmental Protection Act was passed,  many pieces of legislation have proposed to protect coastal marshes from development and other anthropogenic depredation. The decline in acreage since then has slowed abruptly.

Wetlands protection policy is often set in a research vacuum, however, with local coastal managers not really knowing what is happening to their natural wetlands.  The behavior of declining wetlands determines the policy that a given coastal manager will take,  from banning construction near them to dumping sewage sludge on the surface to raise the elevation.  Dr. Kearney's model can demonstrate a predictable rate of change in the wetland as the water level increases, which means that intervention at the right time can prevent further deterioration.

Research needs

In order to make informed decisions, managers need a data base that shows the condition of their wetlands.  Many small data bases exist, funded locally and undertaken in problem areas;  the procedures to delineate a wetland, not to mention its condition, are very labor-intensive, so their production is limited and expensive. What is clearly needed is a data base that covers a large area, that can be readily and cheaply updated, and that can be incorporated into the geographic information systems of the many researchers working on this problem.

Therefore, the Coastal Marsh Project team has developed a method for constructing a data base that can be added to indefinitely and used as a rough model for coastal managers who need to know the condition of their local marshes.  If an area is found to be in peril, local managers can then seek out more precise analyses and determine the most cost-effective intervention.


Return to the Top of the page.


Data Base Components

Thematic Mapper images

Thematic Mapper (TM) images from Landsat satellites provide spectral reflectance of surface properties. TM scenes of the eastern United States were used to determine the amount of water in the salt marsh areas.

National Wetlands Inventory data sets

National Wetlands Inventory (NWI) maps provide large-scale (1:24000) delineations of all wetlands.  Many of the NWI maps for the eastern United States have been digitized into ARC/INFO vector GIS format and are available to download from the USGS GeoData web site.  The NWI polygons were used to distinguish palustrine and estuarine wetlands from each other and from non-wetland areas.

U.S. Geological Survey topographic maps

USGS topographic quadrangle maps provide accurate large-scale representations of general surface features and their relative positions. Topographic maps for the study area were used for general reference and orientation purposes.  In some cases the paper maps were used on digitizers to register ground control points in ARC/INFO  GIS  software.

ArcUSA thematic layers

ESRI's ArcUSA 1:2M thematic data layers were employed within ARC/INFO and ArcView GIS software for orientation and cartographic purposes.


Classification Process

The Coastal Marsh Project employed the following methodology to produce the Marsh Surface Condition Index data layer:

1. National Wetlands Inventory data (NWI) were acquired wherever available within the study area. An example of this data set for the Wingate, Maryland quandrangle demonstrates the NWI wetlands designations within the vector GIS polygons.  The boundary of this data set follows the USGS 1:24000 quadrangle standard.
        Wingate example NWI data

2. Landsat Thematic Mapper scenes (TM) were acquired for the study area. The Wingate example is displayed in false-color infrared.  The black areas are open water (note how easy it is to tell the edge of the land at the water interface); the blue-grey areas are wetter and the reddish areas are drier land. This illustrates how the spectral signal can be used to determine the percentage of water in a given area.
        Wingate example TM data

3. Using the ARC/INFO GIS software, the NWI data were reclassified and similar polygons were combined to form new, larger polygons. In this manner, the many NWI wetland classifications were simplified into estuarine, palustrine and non-wetlands areas within the quadrangle.  The new, individual, quadrangle-based vector GIS coverages were then joined to create a single, state-wide vector GIS.
        Wingate example NWI reclassification data

4. Using the PCI Image Processing software, the TM scenes were registered to ground control points and corrected for geometric distortions. The data was then processed through an atmospheric correction algorithm developed at the University of Maryland. A mixture model program separated out the different signals from the satellite image to determine the percent of water across the scene.  The classified TM data was then converted to raster GIS format, where each pixel value represents the percent of water.
        Wingate example TM mixture data

5. The NWI and TM output GIS data sets were combined to produce a Marsh Surface Condition Index.  The Coastal Marsh Project methodology assumes that a coastal marsh deteriorates as it becomes increasingly waterlogged. The Marsh Surface Condition Index provides a classification which demonstrates the vulnerability of a given area to disintegration.   A coastal manager could use this data in a GIS to show areas of concern.  The healthiest marsh surface is green, ranging from yellow to red to dark blue, with increasing deterioration.  Areas of 'complete deterioration' indicate new areas of open water. This final dataset is in vector GIS format and has been combined so that each square is now about 200 meters on a side.
        Wingate example output data.



Return to the Coastal Marsh Project main page.