Presentation

The context

The data used for assessment purpose in fisheries science is usually collected and raised at a National level and then combined to be used in the assessment models. The data collection Regulation (1639/2001) entering to force in 2002 aimed to guarantee the quality of the data and improve co-operation between European countries. At the same time, more pressure was put on the assessment working groups to support their recommendations and some questions were raised about the quality of the model input data. The concept of precision in the sampling was introduced in the very first years of the DCR. The purpose of precision calculation is here to find at a stock level, the optimal intensity and sampling share between countries and/or between metiers. In 2006, it became urgent to investigate the broad discard sampling in term of results and precision obtained in the purpose of improving the assessment. The guidelines for collecting data on maturity, growth, sex-ratio or fecundity are still awaited to enable the adequate collection and use of the data.

The ICES PGCCDBS (ICES 2005b) was created in 2002 with the aim of promoting international co-ordination between Institutes collecting data for assessment purposes. As International co-operation was not promoted directly by the DCR, partners involved in European projects like Study Contract 94/013, FIEFA, EMAS and SAMFISH wanted to keep the co-ordination already implemented and proposed the creation of PGCCDBS. From this Planning Group, age reading, discard and precision workshops were initiated to address specific issues across all European countries.

The discards and first precision workshop WKSCMFD (ICES 2004) aimed to provide guidelines to implement appropriate sampling program and precision calculation, whilst the second precision workshop WKSDFD (ICES 2005a) began to investigate on the data.

One of the principal outcomes of WKSDFD was to show how important and informative the exploratory data analysis was and how results in terms of precision were dependent on the method used. Specific tools have been developed in a few countries, some allow the exploratory analysis and/or calculate the precision of length/age structure, some calculate the precision of maturity and growth data, others can compare age-length keys, none of them covers the full range of the DCR with reliable, comprehensive and agreed methods.

The objectives

The objective of the study is to develop a common "open source" tool (COST) for assessing the accuracy of the biological data and parameters estimates collected for stock assessment purposes within the framework of the Data Collection Regulation.

As a response to the lot N°2 of the call for tenders FISH/2006/15, the common "open source" tool-box will consist of different packages that will develop validated methods to investigate and estimate parameters for (i) discard volume, (ii) length and age structure of catches and landings, and (iii) biological parameters such as growth, maturity and sex-ratio. Where appropriate, the estimates will be calculated according to one out of a fixed number of agreed raising procedures, based on the methods already developed by some Institutes. Linkage with ICES end-users will be consolidated in close cooperation with ICES in order to facilitate their use by the stock assessment working groups. The definitions and methods will be in line with those described and summarised in the report of the ICES Workshop on Sampling Design for Fisheries Data (ICES 2005a).

According to the tender document, the packages should include:

  • Data administration;
  • Exploratory data analysis;
  • Parameter estimation and associated precision;
  • Simulations.

The main tasks should be as follows:

  1. Propose a common format of datasets comprising all the variables needed to raise the data to the population level and estimate statistical properties (existing formats such as Fishframe will be considered)
  2. Based on the common format, propose exploratory analysis of the most disaggregated data to enable the search for outliers, misallocated data and allocation of samples per strata
  3. Based on the common format, develop algorithms and implement software programs to estimate the statistical properties at a strata level and at the population level
  4. Based on the common format, develop algorithms and implement software programs to account for missing data and account for external errors
  5. Based on the common format, develop algorithms and implement software programs to enable the investigation of the number of samples and the number of individuals to sample to achieve a target precision

The development of the common "open source" tool-box should take into account the recommendations from the 2006 ICES Planning Group on Commercial Catches Discards and Biological Sampling (ICES 2006) and from the 2005 ICES Workshop on Sampling Design for Fisheries Data (ICES 2005).

The outcomes of the project should include:

  • Report summarising the data
  • Graphs of the Exploratory analysis results
  • Raised estimates (volume of discards raised by trips, by total landings and/or by an auxiliary variable, length and age structure of catches, biological parameters) by agreed strata associated with their precision estimates
  • Report summarising precision estimates and quality indicators
  • Simulation analysis to investigate the optimal sampling intensity to achieve a target precision
  • Manuals (COST reference manual, COST Tutorial and COST User Manual)
Timetable and Final Report

The duration of the study should be eighteen months from the signature of the contract. An interim report should be made available within six months from the signature of the contract and a draft final report should be made available within seventeen months from the signature of the contract.

The methodology

 Constitution of a core team 

To answer such a technical call for tenders involving very precise expertise, there were two alternatives, namely (i) chose a very compact core team and developers or (ii) open the door to a variety of expertise from different geographical regions. The first alternative would have been an easy and efficient way to carry out the work but it is the second alternative that has been chosen to guarantee (i) that all the country and/or regional specificities would be considered, (ii) ensure that the methods developed correspond to the needs of those countries and (iii) ensure the widest dissemination of knowledge. The counterpart of this choice is that the management package including the work of the core team and the beta-testing of the methods takes a substantial part of the overall budget.

 Area coverage 

The COST methods will develop validated methods to investigate and estimate sampling indicators for (i) discards, (ii) length and age structure of catches and landings, and (iii) biological parameters such as growth, maturity and sex-ratio from all the geographical regions covered by the DCR. In COST, there will be experts from the Baltic, the North Sea, the Atlantic and the Mediterranean, in order to ensure the complete coverage of the European continental waters.