management/assessment issue, (3) the resources available to develop, implement, 

 and operate the model, and (4) the capability of the various modeling method- 

 ologies to provide the required information within known and acceptable 

 co nf idence limits . 



The workshop perceived a definite need for modelers and managers to 

 collaborate and work closely together through all phases of the modeling 

 project as well as the overall program to assure that modeling efforts are 

 responsive to the program objectives and pertinent to the information needs of 

 the decisionmakers. Modeling proceeds through an iterative process which 

 continually redefines the conceptual and functional structure of the model. 

 Similarly, program objectives may be redefined and refocused in light of new 

 data. Modelers and managers must be continuously aware of each other's sphere 

 of limitations and options. Modeling activities which take place in isolation 

 from the other aspects of the overall management/assessment program will lead 

 to misunderstandings and misconceptions about what information is required of 

 the model and what information can be provided by the model. As a result, the 

 model's utility will be less-than-satisf actory. Thus, it is imperative that 

 modelers and managers interact every step of the way including model operation 

 and interpretation of the results. This interaction should involve others 

 associated with the program (e.g., field scientists, economists, demographers) 

 and proceed through well-defined stages (see Panel B's report for a detailed 

 treatment) . 



Lack of critical data and knowledge of basic ecological processes and 

 interactions including pollutant impacts and synergisms are major impediments 

 to the overall effectiveness of ecosystem modeling as a management and decision- 

 making tool. The participants concurred that the quality of the data base on 

 which a model is built is a major factor influencing the confidence limits of 

 the modeling results. It was agreed that the value of ecosystem modeling as a 

 marine resource management and impact assessment tool would be greatly enhanced 

 if crucial data gaps were filled. For example, pre- and post-recruitment 

 multispecies ecosystem models have virtually no current application to fisheries 

 management concerns because production, mortality, and dispersal processes 

 (both biotic and abiotic) which affect egg, larval, and juvenile survival are 

 poorly known. A quantitative understanding of these processes is mandatory 

 before fisheries models as resource management tools can progress beyond the 

 more traditional single species, population dynamics type of models currently 

 in use. 



With regard to environmental quality ecosystem modeling, key research 

 priorities which were identified are (1) fluxes and processes of material 

 exchange between estuaries and coastal waters, sediments and water column, and 

 sediments and organisms, and (2) pollutant effects on organismal physiology, 

 reproduction, and survival for different life stages. Participants expressed 

 the viewpoint that ecosystem models used as research tools to test and formulate 

 hypotheses of causal relationships and to identify data needs are of as much 

 value to long-term management and assessment objectives as are predictive 

 models. Modelers, managers, and administrators should not lose sight of this 

 important aspect of ecosystem modeling. 



In addition to ecological data needs, there was general recognition that 

 algorithms which will translate ecosystem modeling results and environmental 



xii 



