County line for a distance of 16 km (10 mi), 

 the shorelands are composed of dunes and 

 barrier beaches. At this point the shore type 

 changes abruptly to rock outcrop at the 

 water's edge. This rock shore extends north 

 to the St. Lawrence River interrupted only 

 by a few pockets of beaches and marshes at 

 the inner end of the deep bays. 



RECOMMENDATIONS 



A list of options for landward and seaward 

 boundaries of Level I, II, and III divisions, along 

 with the pros and cons of adopting each option, 

 was presented in the Results section. The ideal 

 landward and seaward boundaries of divisions 

 would be those which delimit the major coastal 

 processes which occur in each division. This would 

 most accurately reflect functioning of real-world 

 ecosystems. Unfortunately, these are extremely 

 difficult to delimit. In actual practice the land- 

 ward and seaward boundaries described by the 

 Cowardin et al. (1977) classification, as described 

 in Results, are probably as close to these ideal 

 boundaries as can be drawn. The real advantages 

 to adopting the boundaries used by the National 

 Wetlands Inventory are that they are being mapped 

 presently and that a large amount of data are being 

 stored in this format. All other options listed are 

 unacceptable due to the problems inherent in 

 each, as previously described. 



Concerning lateral (perpendicular to the shore- 

 line) boundaries of Level I and II divisions, those 

 which end at the political boundaries of the United 

 States are obviously artificial. They were delin- 

 eated in that manner due to the scope of the study. 

 It is obvious, however, that the boundaries of 

 coastal ecosystems logically should not resemble 

 political boundaries. Thus Table 2 lists more 

 rational boundaries for Level I and II divisions 

 which abut the political boundaries of the United 

 States and overlap into other countries. 



In some instances it may be necessary or useful 

 to lump or further subdivide Level II divisions for 

 the purpose of producing Characterizations or 

 Profiles. For example, one might lump the North 

 and South Bristol Bay divisions into a Bristol Bay 

 Characterization. In the case of lumping, it is ad- 

 visable to lump Level II divisions which are within 

 a Level I division, rather than those from two dif- 

 ferent Level I divisions. Level II divisions within a 

 Level I division are by definition more similar 

 and, thus, may have predictions made about them 



which are more reliable than predictions made 

 about Level II divisions drawn from different 

 Level I divisions. Thus, lumping should occur 

 only within Level I divisions. 



Criteria for separating Level III divisions are 

 suggested in the Methods section. Because of the 

 detailed information which would be needed to 

 accurately delineate the Level III divisions, it is 

 recommended that such divisions, if they are 

 needed, be products of either a characterization 

 or some special study on a specific Level II division. 



SUMMARY 



The objective of this project is to formulate a 

 hierarchical regional classification scheme for 

 coastal ecosystems of the United States and its 

 territories based on the physical characteristics of 

 those areas. The classification is designed to ad- 

 dress the following: "How can the coastline of 

 the United States be partitioned to best separate 

 ecosystems, when the purpose of defining these 

 ecosystems is to make predictions about how spe- 

 cific types of perturbations in specific geographi- 

 cal areas will affect the ecosystems hydrologically, 

 structurally, functionally, and biologically?" 

 Two primary users of this classification are the 

 National Coastal Ecosystems Team and Ecological 

 Services, both within the FWS, who will use the 

 classification for determining locations and 

 boundaries of subject areas for their Characteri- 

 zation Studies, and Profiles (see Glossary). 



Existing coastal classification schemes were 

 examined to determine if any were suitable for 

 fulfilling the above stated objective. Coastal clas- 

 sifications were found to fall into essentially three 

 types— structural, functional, and regional. Struc- 

 tural and functional classifications do not address 

 geographical problems and are thus not appro- 

 priate; only regional classifications address the 

 question being asked. 



There are two types of regionalizations— one 

 based on biogeography and one based on physical 

 (chemical, geological, etc.) parameters. Biogeo- 

 graphical regionalizations are based on the actual 

 distribution of one or a few groups of organisms 

 and do not address distribution of coastal ecosys- 

 tems per se; regionalizations based on physical 

 parameters do address ecosystems. The only re- 

 gionalization found which is based on physical 

 parameters (Dolan et al. 1972) was rejected be- 

 cause of the size of its Elemental Units. Thus it 

 was appropriate to develop a classification scheme 



12 



