Model Comparisons with Site Assessments 



There are few opportunities for validation of our regional model of coastal response to sea 

 level rise. Knowing that the model has an accuracy definable at a particular level would be of 

 great help in interpreting the findings of the study. One approach to validation, although an 

 imperfect one, is to compare model results with detailed studies of local sites. Two such studies 

 are available— a study of the impact of sea level rise on wetlands in Tuckerton, New Jersey, by 

 Kana et al. (1988 and Chapter 3, this report), and in Charleston, South Carolina, by Kana, Baca, 

 and Williams (1986 and Chapter 2, this report). However, it must be recognized that true 

 validation cannot be obtained because of the radically different approaches being compared. 

 Thus, our simulations for Charleston suggest that a greater capacity for marsh migration exists 

 than fine-scale analysis suggests. As stated above, fine-scale disturbances and landscape 

 complexity, which limit marsh migration, could not be simulated using a square kilometer grid. 

 The New Jersey site, however, with greater landscape homogeneity on a coarser scale, provides a 

 quasivalidation of the SLAMM model. 



Several of the major differences in methodology of the regional model and site-specific 

 approaches should be understood before making comparisons. First of all, the model approach 

 operates at a much larger geographic scale and consequent loss of local scale accuracy, in 

 keeping with the major objectives of the study. Thus, for example, high and low salt marsh are 

 not distinguished in the model as they are in the site studies. The 1 km 2 cell which forms the 

 spatial unit of the model is defined only by the predominant land category type present. There- 

 fore, in areas where salt marsh may be an important but secondary land category, it will be under- 

 represented in the regional analysis. Similarly, where salt marsh predominates, it could be over- 

 represented as the only category present, and if conditions for migration are favorable, an over- 

 estimate of migration results. In the comparisons to follow, this latter situation is believed to be 

 more significant than the former. 



The data limitations in the modeling approach are defined by the accuracy and timeliness of 

 the USGS 7 and 1/2 minute (and occasionally 15 minute) quadrangle topographic map series. 

 Necessarily, then, a set of generalized properties results. This is most apparent with elevation 

 because the quadrangle series frequently presents elevational contours at five- or ten-foot inter- 

 vals, which are quite coarse for subdividing coastal land categories. Consequently, subtle 

 differences which show up in a detailed study as a loss or gain of one category or another are not 

 recognized in the regional analysis. 



Freshwater and saltwater marshes are not distinguishable based on the USGS maps. There- 

 fore, the raw data recognizes only "marsh," and our model used an algorithm based on elevation 

 with respect to spring high tide to differentiate the two types. 



Other aspects affect both regional and local interpretations. These include limited data on 

 subsidence rates and accretion rates as well as on actual marsh migration rates, and lack of any 

 empirical knowledge of coastal land responses to sea level rise at a rate as rapid as that projected 

 for the next century. 



The major response at the New Jersey wetland site to the low scenario through 2075 is the 

 replacement of high salt marsh with low salt marsh (Kana et al. 1988 and Chapter 3, this report). 

 Also projected is the loss of over half the transition marsh in the Tuckerton area, but an increase 

 of the same area in the Great Bay Boulevard area. However, at both locations no change in 

 overall wetland area is projected under the low scenario. The conversion of high to low salt 

 marsh noted by Kana et al. would not be detected in our model; furthermore, because the 

 distinction between saltwater and freshwater marsh cannot be made in the input data but is 

 based on imprecise elevation determinations, we prefer to consider total wetland changes. 

 Adjustments to transitional marsh in the New Jersey and South Carolina studies would occur 

 within the framework of our general freshwater marsh category. We project a 9 percent decline in 

 total wetland area by 2075, growing rapidly to a 75 percent decline by the year 2100. For the year 



123 



