quality criteria into economic and societal terms for use in cost-benefit and 

 risk/damage assessment analyses must be developed. The priority for this is 

 high since established standards for the marine environment do not exist. 



Finally, to state that a model is no better than the data available to 

 construct and quantify it is a much cited truism, but there is more to enhancing 

 confidence in ecosystem models than just improving the quantity and quality of 

 the data. Ecosystem modeling has evolved in an unstructured fashion, and model 

 development is not guided by formal criteria. As a result, ecosystem modeling 

 as a scientific discipline is not firmly grounded in strong theory. Quite 

 often mathematical functions and time-space aggregations of variables are chosen 

 more by tradition than by rigorous scientific principles. Modelers should give 

 high priority to the development of sound theoretical principles for modeling 

 and the establishment of formal criteria for model formulation and evaluation. 

 There was a general feeling that an initial step in this much needed direction 

 would be to devise and implement testing procedures for comparing, validating, 

 and clarifying the properties of several existing models and modeling method- 

 ologies. The knowledge gained from these activities would be of tremendous 

 value to the future development and application of ecosystem modeling as a 

 marine environmental assessment and resource management tool. 



xm 



