|A short review of precautionary reference points and some proposals for their use in data-poor situations (1998)|
This paper ends with a plea for empiricism in setting reference points in conditions of uncertainty and calls for a recognition that LRPs are only components of a management system, and the key criterion that will lead to functionality is acceptance by industry and management. There is a need to test these approaches both in simulation and in practice in order to detect and overcome possible problems of practical implementation of control laws and determine their stability of performance under a range of environmental conditions. The standard approach of seeking to define a management control rule or law obviously is relevant here, but it is not excluded that reliance may also have to be placed on empirical measures. Using a 'basket' of empirical measures, each derived from fundamentally different data sources in a 'traffic-light' mode, may provide a good indicator of the need for management action, just as, in another sphere, the Dow-Jones index of the New York Stock Exchange provides a sensitive indicator to performance of the whole market based on the performance of a small number of key stocks. Once a degree of confidence has been established as to the performance of such multiple criteria, they may be incorporated into the fishery management scheme as one way of achieving a flexible response to changing conditions.
Obviously there are some serious difficulties in reproducing many of the S-R based reference points favoured in the ICES/NAFO area for stocks where long series of recruits and/or age structure are not available. At the same time, fisheries often offer considerable level of detail in catch statistics by subarea (see, for example, Carocci and Majkowski, 1996), which could provide the basis for novel indicators of over-exploitation such as have only been hinted at in this paper. For tunas, these would require further information on migration and could be tested by a combination of simulation and detailed tagging data.