SUMMARY 



The consensus of the workshop was that mathematical ecosystem modeling has 

 proven to be a useful and practical scientific tool for addressing marine 

 environmental assessment and resource management concerns . Modeling is a 

 valuable mechanism for quantifying and integrating multidisciplinary data into 

 a meaningful information base which would otherwise have been unattainable. 

 Models allow decisionmakers to evaluate a priori the consequences of various 

 actions and management strategies. Viewed from this perspective, virtually 

 all modeling efforts can be of some value to decisionmakers provided they are 

 timely and cost effective. However, the continued role of ecosystem modeling 

 in a management /assessment context will require modelers to develop new 

 methodologies which will make modeling more relevant to the specific information 

 needs of managers, decisionmakers, and other users of modeling results. Two 

 key areas of special importance are (1) restructuring traditional modeling 

 approaches to incorporate methodologies suitable for addressing social and 

 economic consequences of environmental impacts and resource management strategies, 

 and (2) developing appropriate stochastic methodologies which will provide 

 users with probability of occurrence, risk assessment information. 



The workshop cautioned that ecosystem modeling results should never be the 

 sole or even the primary basis for a management/assessment decision. Models, 

 regardless of their apparent complexity and exhaust iveness , are simplified math- 

 ematical representations of natural systems and, as such, they are imprecise 

 mimics and predictors of reality. It is unreasonable to expect that any model 

 will ever provide definitive results free of uncertainty to complex and often 

 multi-objective management /assessment issues. While these points may seem 

 trivial and obvious, they are crucial to the success of any modeling project. 

 Management expectations of modeling efforts must be tempered by recognition 

 and acceptance of the many factors which limit the predictive power of models. 

 Models are tools which provide insight to environmental/ecological relationships; 

 they do not provide answers per se, and they are not ends in themselves. 



The workshop was unanimous in its appraisal that the concept of holistic, 

 site-adaptable and problem-adaptable generic ecosystem models was neither a 

 prudent nor feasible management goal to pursue either now or in the foreseeable 

 future. The necessity to incorporate numerous untested assumptions and 

 theoretical relationships into the overly complex structure of such universal 

 application models would completely erode any potential predictive effectiveness 

 and preclude any usefulness to real world management/assessment needs. Rather, 

 the general feeling was that ecosystem models are best suited to their intended 

 purposes when tailored to meet the site-specific and problem-specific nature 

 of the management /assessment concern. Model complexity should be limited to 

 that which is necessary to provide the required information. 



A broad spectrum of diverse modeling approaches and techniques is available 

 to the modeler. The panels reviewed these in generalized terms and found that 

 each has its advantages and disadvantages and must be assessed in light of the 

 particular situation and application. In any case, the suitability of a given 

 approach/technique to a given problem must be determined not by personal 

 preference but by objective criteria such as (1) the quality and resolution of 

 data available to construct the model, (2) the program objectives of the 



XI 



