Use of Quality Control Methods to Monitor the Status of Fish Stocks

Use of Quality Control Methods to Monitor the Status of Fish Stocks

James Scandol

Use of Quality Control Methods to Monitor the Status of Fish StocksThis is part of Fisheries Assessment and Management in Data-Limited Situations
PDF    
To download the free PDF [353 KB], please enter:
-or-

Description

Many fisheries that are data-limited are also of low economic value. Therefore, not only are the fisheries data-limited, but there are limited human resources available for undertaking stock assessment. Qualitative methods such as "eyeballing" the data are then often used to assess such systems. Quantitative methods need to be developed that are objective, but less demanding than dynamic stock assessment models. In particular, simple methods that can signal trends in empirical stock-status indicators need to be explored. One such approach is the use of quality control methods such as Shewhart, moving-average, and CUSUM (cumulative sum) control charts. Originally designed for industrial quality control, these methods can be parameterized to detect transient or persistent causes with specific false-positive and false-negative error rates. These signals can be interpreted within a managerial context as trigger reference points.

Results of a simulated study of yellowfin bream (Acanthopagrus australis) stocks from New South Wales (Australia) are presented. Empirical stock-status indicators including catch, catch per unit effort, mean age, mean length, recruitment fraction, total mortality, and fishery-independent surveys were processed using quality control methods. Performance of these indicators and algorithms were measured with receiver-operator-characteristic curves, which captured both false-positive and false-negative error rates. Biomass surveys performed best, followed by mean age and length, and recruitment fractions. Commercial catch rates and catch had the worst performance but were still acceptable. Age-based total mortality performed poorly unless very large numbers of samples were taken. Potential applications of these methods include a rapid diagnostic tool in data-limited situations, development of empirical reference points, and empirical rule-based management systems. These methods are easily applied even when there is a short time-series of low-contrast data but a range of caveats must always be considered.

Item details