The existence of risks to stocks associated with fishing 

 have been demonstrated repeatedly as fishery resources have 

 been overexploited , or reduced through a combination of 

 harvesting and weather stresses. Examples include the great 

 whales, Peruvian anchovies, and California sardines. Exper- 

 ience with these sometimes unanticipated population changes, 

 along with advancing understanding of the ecological relation- 

 ships that keep an ecosystem resilient or able to recover 

 rapidly from disturbances, make at least preliminary evalua- 

 tions of risk possible. Focusing only on a particular stock, 

 without also considering associated species and environment, 

 is too limited a perspective for evaluating the risk of ad- 

 verse impacts on a fishery resource. Interactions between 

 harvesting and other impacts such as pollution, climate change 

 or available alternative habitats must also be considered. 



Determining acceptable levels of risk of short or long 

 term adverse impacts on stocks, habitats, or marine ecosys- 

 tems is a policy question. Environmental legislation dic- 

 tates caution when impacts of an activity cannot be antici- 

 pated, and the minimization of the risk of long term adverse 

 impacts to both stocks and the marine environment. 



In cases where risk levels cannot be assessed, the data 

 required to make those assessments should be indicated in 

 the plans . In such cases , gradual as opposed to abrupt in- 

 creases in fishing would seem in order to allow for time lags 

 and a wide margin of safety with respect to possible adverse 

 impacts on nontarget species. 



Future Developments in Fishery Management Theory 



The assumptions underlying single species management and 

 MSY calculations are limiting and preclude ecosystem level 

 considerations. The development of new theories for eco- 

 system level fisheries management will have to begin with 

 changing those assumptions. Ecosystem level management will 

 be more complex than single species management because the 

 problem is more complex. In particular, the number of vari- 

 ables will increase. Aspects of the environment, particular- 

 ly natural mortality, now assumed to be constant, will have 

 to be treated as variables. Consequently, data requirements 

 will also increase. Some serious consideration of which 

 parameters will be most useful and practical in ecosystem 

 level management should be undertaken by scientists and man- 

 agers familiar with present ecological theory and fisheries 

 management and with environmental legislation. 



46 



