1 of 1

Managing fisheries by linking indicators with process control techniques�Deepak George Pazhayamadom�A SEA CHANGE PROJECT OF MARINE INSTITUTE FUNDED BY NATIONAL DEVELOPMENT PLAN 2007-2013, IRELAND

Many fish stocks are now declining due to fisheries exploitation and changes in the ocean environment. To avoid depletion, we have to assess fish stocks and develop appropriate management plans or strategies. Traditional models demand large amount of fisheries data and are not suitable to apply in data-limited or data-poor situations. As a result, more than 85% of the global catch are from un-assessed fisheries and lack management advice.

​

Objective:

​

Develop an indicator-based method to monitor, assess and manage data poor fisheries.

​

To test the application of CUSUM control chart in fisheries management.

​

Background

Experiments

Methods

Results

Summary

References

Can we manage a fishery if no historical data are available? - YES

Simulation model:

​

An age structured virtual fish population was developed and a trawl fishery was simulated using codes written in the R language.

​

Data collection:

​

More than 100 stock indicators were computed from the operating model. These indicators are ideally available in data poor situations.

​

The fish stock was simulated for 20 years with a stepwise increase in fishing mortality. Further the indicators were correlated with spawning stock biomass (SSB) of the fish population. Indicators with significant correlations indicate that they can be used to track the trend of mature fish abundance in the fish population (Pazhayamadom et al., 2013).

STEP 1: Find potential indicators

STEP 2: Monitor indicators using CUSUM

The indicators are monitored using a trend detection algorithm known as the self-starting CUmulative SUm (CUSUM) control chart. They are efficient in detecting gradual and persistent changes in the indicator time series. The CUSUM was tested for its detection capabilities that whether it signal an ‘out-of-control’ situation if under- or over-fishing occurs in the fish population. Results were evaluated using Receiver Operator Characteristic (ROC) curves and that indicates the sensitivity and specificity of self-starting CUSUM with a wide range of stock indicators (Pazhayamadom et al., 2013).

Illustration: An example where the mean length indicator is monitored using self-starting CUSUM. The upward shift in indicator time series was detected efficiently by the CUSUM. Note that the method works even when only few observations are available in the indicator time series.

STEP 3: Assess indicator trends using EPC

Once a signal (or trend) is detected by CUSUM, the next step is to estimate the shift (Ŝ) occurred in the indicator time series. Techniques in ‘Engineering Process Control (EPC)’ theory are useful for estimating this indicator shift.

​

Four types of EPC techniques were used to estimate the indicator shift in an overfishing scenario: (1) Taguchi’s method; (2) Montgomery’s method (3) Grubbs harmonic rule and (4) Using CUSUM observations. The results were evaluated by comparing the estimated shifts (Ŝ) with the actual change occurred in the underlying population biomass.

Illustration: An example where Grubbs harmonic rule is used for estimating the upward shift in the indicator time series. Here, the last four observations in CUSUM are out-of-control. Hence the corresponding indicator values in these years are used for the computation of upward shift (Ŝ).

STEP 4: Adjust catch using CUSUM-HCR

The CUSUM-HCR is a ‘feedback-response’ system where the CUSUM signals are integrated with EPC techniques to invoke a management response. This management framework use Harvest Control Rule (HCR) where the Total Allowable Catch (TAC) will sustain, increase or decrease based on the trend in the indicator time series (Pazhayamadom et al., 2015).

If current year = i then,

STEP 1: Find potential indicators

Results from iterated simulations showed that indicators based on old (CN10, CW10, PN10, PW10), mature (On, Ow) and large fishes (Pn, Pw) in the catch correlated significantly with SSB of the fish population.

STEP 2: Monitor indicators using CUSUM

Best

Good

Worst

The closer the apex of the ROC curves towards upper left corner, the better is the detection. The Large Fish Indicators (LFI) were more sensitive and specific when monitored using self-starting CUSUM.

STEP 3: Assess indicator trends using EPC

Results showed that the Grubbs harmonic rule is most efficient EPC technique that can be used to estimate the indicator shift accurately.

STEP 4: Adjust catch using CUSUM-HCR

Results showed that the CUSUM-HCR was successful in sustaining the initial state of the stock (at 50% MSY) despite having no historical data for the stock or fishery.

The CUSUM-HCR is a management framework based on empirical indicators and hence they can be used at data poor situations.

Pazhayamadom, D.G., Kelly, C.J., Rogan, E., and Codling, E.A. 2013. Self-starting CUSUM approach for monitoring data poor fisheries. Fish. Res. 145: 114–127.

​

Pazhayamadom, D.G., Kelly, C.J., Rogan, E., and Codling, E. A. 2015. Decision Interval Cumulative Sum Harvest Control Rules (DI-CUSUM-HCR) for managing fisheries with limited historical information. Fish. Res. doi:10.1016/j.fishres.2014.09.009.