2075, Kana et al. project a slight increase in the total marsh area, whereas we project a 9 percent 

 decline. However, as late as 2045 we project a 1.0 percent decline in wetland area, a figure not 

 significantly different from theirs given the limits of both studies. Our simulation through the 

 year 2100, however, suggests that the trend toward migration onto adjacent lowland would soon 

 come to an end and that many of the gains would be lost. 



Agreement is more pronounced under the high scenario. Kana et al. project an 86 percent 

 decline in salt marshes of the Tuckerton area by 2075, compared to a loss of 75 percent by 2075 

 and a loss of 99 percent by 2080 in our study. Consequently, our conclusion with respect to salt 

 marshes in the Tuckerton area is that the two methods, despite being dissimilar in many respects 

 and covering different areas, represent reasonably well an unstable coastal situation which leads 

 to either salt marsh gains or salt marsh losses, depending on rates of sea level rise. 



FUTURE RESEARCH NEEDS 



Although the implementation of the SLAMM model has provided a useful analysis of 

 probable coastal wetland responses to accelerated sea level rise, increased accuracy, reliability, 

 and credibility would follow from additional refinement and study. We recommend that the 

 following steps be implemented: 



(1) Increase the resolution by using a 0.25 km 2 or 0.125 km 2 grid cell for most areas. This 

 would avoid the under- or over-representation of categories such as marshes and would 

 permit the elevation of the dominant category to coincide more closely with the average 

 cell elevation. The reliability of results would be significantly increased through these 

 more realistic estimates of the distributions of the major categories. 



(2) Obtain statistically unbiased samples of sufficient size for quantitative inferences. To do 

 this, a method for stratified random sampling within each region must be developed 

 which takes into account variation in wetland types and coastal topography. With such a 

 method, large-scale changes could be estimated for specific regions, with a level of 

 accuracy sufficient to guide policymaking at the regional level. 



(3) Distinguish among wetland types, including freshwater, transitional, and high and low salt 

 marshes, using the Fish and Wildlife Service habitat classification maps. This would 

 provide a better basis for understanding changing ecological relationships and their 

 implications for future conservation and resource management. 



(4) Analyze the change in the boundary between wetland and open sea. Although wetland 

 loss is recognized as deleterious to fisheries and other marine resources, the relationship 

 is not linear. Recent model analysis using a 1 km 2 cell grid (Browder, Bartley, and Davis 

 1985) shows that as the total "interface" of a coastal marsh (area of marsh surface 

 exposed to tidal water) changes as marsh shoreline disintegrates or becomes increasingly 

 indented, nutrient exchange increases to a point and then declines rapidly, affecting the 

 coastal fishery. An analysis of the changing marsh area exposed to tidal waters could be 

 made from the database and SLAMM model used in the present study; such an analysis 

 would help diagnose the changing resource values of the wetlands. 



(5) Validate the model, using historic data on changes in coastal wetlands, beaches, and 

 lowlands, accompanying anomalously large subsidence in areas such as the Mississippi 

 Delta in Louisiana, Galveston and Houston, Texas, and San Jose, California. 



(6) Use data for remote sensing. This would make it possible to more accurately characterize 

 existing vegetation types. TVansect studies could be used to characterize the relationship 

 between vegetation type, frequency of flooding, and elevation, as described by Kana et al. 



124 



