Summary

To validate data poor methods, estimates from data rich stock assessments were used to siumulate scenarios related to different level of information and knowledge.

• Methods tested were i) Catch Only Methods (COMs) based on biomass dynamics; ii) Length Based Indicators (LBIs); and iii) life history characteristucs to estimate Productivity (\(r\)) as used in Productivity and Susceptibility Analysis (PSA).
• The data rich datasets used were i) https://www.ramlegacy.org; ii) ICCAT Jabba assessments; and iii) ICCAT Stock Synthesis assessments.
  
• Although the RAM database provides a large sample of stocks with different characteristics the results can be difficult to interpret as the details vary, e.g. the assumptions used when fitting for values of difficult to estinate parameters and how data conflicts are addressed.
  
• Therefore the COM analysis was repeated for the ICCAT assessments using a self-test and a cross-test.
• Evaluation depends on how the estimates could be used in management advice, i.e. to estimate current stock status relative to reference points, or trends.
• COMs performed poorly, both when depletion was know or priors based on life histories were provided. This was particularly true if only a short time series of ctach was available

Self Test

Operating Model was based on JABBA assessments, i.e. where the assumptoins used in fitting corresponded to the method tested

Cross Test

Operating Model was based on an integrated age/length based assessment (SS3)

Summary

ICES have proposed a variety of Length Based Indicators (LBIs) in data poor situtations

• Time series are constructed from length frequencies distributions, i.e. means, modes and upper and lower percentiles; and then
• Compared to a variety of reference levels based on life history characteristics
• Used as proxy \(MSY\) reference points or indicators of overfishing
• Used the length compositions from SS3 to generate time series and
• Life history parameters from FishBase to estimate reference levels
• Performance was variable both across fleets and between indicator type, therefore in data poor cases it will be inporartant to devlop screening methods for indicators and to indentify the Value of improved sampling.
• The results also indicate that the length compositions used in the ICCAT assessments are infomation poor and often in conflict. More effort should go into developing appropriate diagnostics (e.g. Carvalho et al.)

Fishing Mortality relative to \(F_{MSY}\)

Length Base Indicators: Mean Length and \(L_{opt}\)

Proxy for \(MSY\)

Length Based Indicators \(25^{th}\): Percentile and \(L_{mat}\)

Proxy for growth overfishng

Length Based Indicators: Blue Marlin

Summary

Take \(F_{MSY}\) (or \(H_{MSY}\) based on harvest rate) from data rich stocks as the “best” estimatre of productivity and then compare to life history parameters and estimates of \(r\) based upon them.

• Although there are correlations between \(H_{MSY}\) and life history parameters and the estimates of $r, it is poor
• Ranking species on the basis of their productivity is robust for identifying stocks with low productivity
• As productivity increases the relationship between \(H_{MSY}\) and life history parameters becomes less clear
• Therefore although it appears that life history parameters can be used to identify the stocks with low productivity, it is more difficult to derive priors for data rich assessments.

Ranking

Identifcation of the least productive stocks is robust

Summary

• In age-structured models density dependence is mainly accounted for by the stock recruitment relationship. Cury et al. (2014), however, showed that in most cases the stock-recruitment relationship used to estimate productivity and determine reference points, has poor estimation/predictive power and the environmental has a larger effect on productivity
• In ICCAT assessments growth, maturation and natural mortality are assumed not to have varied despite the large changes in the environment and stock biomass seen, indicators for such demographic processes should be developed.
• The presence of clockwise cycling in surplus production in the tuna assessments is due to recruitment anomalies, this implies that future catches are driven by incoming year classes possibly due to environmental drivers rather than a production function.
• This has consequences for management based on target and limit reference points, since it follows that future biomass trends can not be predicted from current biomass based on setting total allowable catches.
• Management Strategy Evaluations for data rich stocks should also take into account such processes (see Sharma et al.,) and MSE could be used to develop indicators for data poor stocks.

Tropical Tunas

Surplus Production

\(B_t+1\) = \(B_t\) - \(C_t\) + \(SP\)

Process Error

\(B_t+1\) = \(B_t\) - \(C_t\) + \(SP\)

Tasks

• Indicators for assessed and unassessed stocks
• Link to habitat

• Extend to Endangered, Threatened and Protected Species

• How to improve our ability to integrate observations made by different platforms covering different spatial extents at different grain sizes?
• EAFM requires documenting and identifying factors that determine the distribution, abundance, movement, demographics, physical or genetic characteristics, behaviour of ecosystem components.

• A need to develop and enhance observational and model-based techniques that will allow the use cross-scale information to advance our understanding of the Atlantic ecosystem and how it changes over time.

Remote Sensing

Electronic Tags

Fishing

Eff Dis

```

AIS data