would be used to bidance flows within the model 

 and to identify primary areas of concern for 

 management and development activities. 



We proposed to illustrate the model with 

 Forrester Diagrams, foUowing the pattern adapted 

 by the IDOE-CITRE group in their proposal 

 (1972). The units of the compartments and flows 

 would change in relationship to the subsystem 

 under study, i.e., gC/m"^ for energy flows, 

 mg/m /yr for nutrient flows, etc. Because the 

 Forrester Diagrams (Forrester 1961) would quickly 

 become unmanageable in an ecosystem as complex 

 as sea islands, each set within each subsystem was 

 to be treated independently. The subsystems 

 would then be abbreviated when combined to form 

 the principal model. It would thus be possible to 

 maintain a manageable matrix for the ecosystem 

 model as a whole and still have high resolution as 

 each major entity is encountered. 



In practice, the major entities (habitats) incor- 

 porated into the sea island ecosystem model would 

 include, but not be limited to, the following: off- 

 shore euhaline, inshore euhaline, ocean beach (in- 

 cluding shifting dunes), stable dunes, maritime 

 forest, pine forest, coastal plain, marsh (including 

 tidal creeks, river beaches, mud flats, freshwater 

 marsh, brackish water marsh, salt marsh, high 

 marsh, low marsh, marsh impoundments), fresh- 

 water, and estuary. Within each subsystem the 

 principal physical, chemical, geological, and bio- 

 logical entities would be compartmentalized. For 

 example, in modeling the chemical processes of an 

 estuary, the important variables would include 

 salinity (as an index of mixing and a habitat 

 determinant), temperature, concentration of dis- 

 solved oxygen, pH, alkalinity, concentrations of 

 organic materials (dissolved and particulate), 

 nutrient levels, concentrations of certain metals, 

 etc. Biological modeling within subsystems such as 

 estuaries would not proceed to the individual 

 species level but would deal with spatial variation 

 as distributed sources and sinks (Nihoul 1975). 

 Biological subsystems would be comprehensively 

 resolved into component biotic subsets (e.g., phyto- 

 plankton, zooplankton, nekton, benthos, etc.) and 

 linked through major variables (nutrients, carbon, 

 etc.) within the system. In addition, external driv- 

 ing forces (temperature, salinity, light, alloch- 

 thonous materials, etc.) for each subset, and export 

 links to other subsets or subsystems, would be 

 identified. The final model was envisioned as a 

 block diagram with blocks representing the major 

 components and lines indicating flows (of carbon. 



energy, etc.) from one component to another and 

 the relationships between subsystems (Patten 1971; 

 Odum and Odum 1972). 



CONCEPTUAL MODELING- 

 INTERIM PROCEDURE 



The above protocol for conceptual modeling 

 was initiated in February 1977. However, we 

 quickly found that these models actually had oiily 

 narrow application to the project. Also, it became 

 apparent that the list of major entities (habitats) to 

 be incorporated into the ecosystem model would 

 have to be revised. The revision was accomplished 

 by using a synthesis of aquatic and terrestrial 

 terminology and the U.S. Fish and Wildlife Service's 

 Interim Classification of Wetlands and Aquatic 

 Habitats of the United States (Cowardin et al. 

 1976). This synthesis resulted in the identification 

 of seven primary systems (marine, estuarine, 

 riverine, palustrine, lacustrine, maritime, and 

 upland to be modeled encompassing a total of 32 

 major subsystems (fig. 1). Various subsystems will 

 also be modeled. The ecosystem models were 

 to be used to identify system components and to 

 structure them into an expanded subject outline 

 for the characterization. 



The value of the conceptual model in relating 

 functional interactions and regulatory processes, as 

 well as identifying system components, prompted 

 us to pursue models which could be integrated 

 with the characterization atlas and narrative. There 

 the models would present a preface summary of 

 each ecosystem and also function as a user tool in 

 understanding the impact of impingments or 

 perturbations on system components. To perform 

 as part of a user package, the complexity of the 

 master" models would often be dissected into sub- 

 system models or submodels. Submodels are 

 generally divided into four formats: 



1. Terrestrial or hydrological submodels (soil 

 types, elevation, wind, wave action, cur- 

 rents, tidal action, dispersed, diffusion, 

 etc.); 



2. Environmental quality submodels (physical 

 states, chemistry, etc.); 



3. Microbiological submodels (viruses, bac- 

 teria, fungi, microscopic algae, and inverte- 

 brates); and 



4. Macrobiological submodels (macroscopic 

 plants and animals, population dynamics, 

 etc.). These submodels are rarely indepen- 



20 



