back to home, WP4, CS2

Case study 2: Approach 2: Evaluating North Sea (and VIa) cod and haddock recovery plans

Introduction

This approach has aimed to evaluate the biological effects of North Sea cod recovery plans addressing management issue one by evaluating effects of certain management measures in relation to cod recovery plans.

Another scientific work made under the FRS Marine Laboratory national work uses FLR (FL-Core, FL-Fisboat and FL-Ogive) to evaluate management plans under the umbrella of current cod recovery plans. Focus here was the development and application of an evaluation framework implemented under FLR. The stock considered was VIa cod rather than North Sea cod but both stocks are currently subject to a recovery plan utilising the same harvest control rules.

A third study under ICES WGNSSK and FRS Marine Laboratory used FLR to evaluate North Sea haddock recovery plans. Also this work has been put in context of the above approach. Focus is here also development and application of an evaluation framework implemented under FLR.

Description of the fishery, stocks and management system

See general description under the Case Study 2 Main Page.

Description of the base case and scenario evaluations

See general description under the Case Study 2 Main Page.

Data and parameters

Stock data

VIa cod case study: Data by age and year on weights (catch, landings, discards, stock), maturity, natural mortality, and one research-vessel survey index were taken from the 2006 report of the ICES Working Group on the Assessment of Northern Shelf Demersal Stocks (WGNSDS), (ICES 2006a). These data covered ages 0-7+, and years 1978-2005.

North Sea haddock case study: Data by age and year on weights (catch, landings, discards, stock), maturity, natural mortality, and three research-vessel survey indices were taken from the 2007 report of the ICES Working Group on the Assessment of Demersal Fish Stocks in the North Sea and Skagerrak (WGNSSK). These data covered ages 0-15+, and years 1963-2006.

Fleet data

North Sea haddock case study: Catch numbers-at-age and yield (tonnes) for total catch, landings and discards were also taken from the 2007 ICES WGNSSK report.

Modelling

FLR (Operating Model)

VIa cod case study: The operating model was implemented using the following packages: R (2.2.1), FLCore (2.1), FLFisboat, FLOgive. Each iteration consisted of a 26-year forward projection (2005 to 2030). There were 100 iterations for each variant of the model.

North Sea haddock case study: The operating model was implemented using the following packages: R (2.4.1), FLCore (1.4.3), FLXSA (1.4.2), FLAssess (1.4.1). Each iteration consisted of a 25-year forward projection with assessment-driven management feedback: there were 50 iterations for each target F, and 5 target Fs (0.1, 0.2, 0.3, 0.4, 0.5). Simulations were summarised using the number of years in each iteration for which biomass B < B(pa), which is one way of describing the risk that a particular target F will reduce stock sustainability.

Conditioning of Operating Model

VIa cod case study: Initial numbers at age were taken from a TSA assessment performed at the 2004 meeting of WGNSDS (ICES, 2004) and 2005 was taken as the first forecast year. This allowed TACs to be set by the candidate HCRs from the first year they could be applied under the cod recovery plan. Resulting SSB values could be compared to results gained from applying the actual TACs set for the stock in 2005-2007. Also SSB values for 2005-2007 could be compared to results from the latest stock assessment. The performance of HCRs were tested under two sets of conditions. The first, referred to as the base case scenario and representing ideal conditions, allowed the HCR to be applied to the output of the biological model (perfect knowledge of stock condition) and assumed perfect compliance, i.e. the total quantity of fish removed from the stock exactly matched the TAC specified by the HCR. The second characterized uncertainties and problems known to exist in fisheries management assessment and rule compliance.

Biological model

Weight at age was incorporated in two ways. The first made use of the weights at age supplied to WGNSDS, (ICES, 2006a). The mean and standard deviation of weights at age was calculated from data spanning the years 1978 to 2004. The second used the length weight relationship of Coull et al. (1989) after modeling fish length using the von Bertalanffy growth equation. Length at age data came from the FRS first quarter West Coast groundfish surveys 1995-2005. Cohorts were identified by birth year. Since the von Bertlanffy model is nonlinear in its parameters, parameter values were estimated by nonlinear least squares using the R function “nls”. Exploratory analysis revealed that allowing variation in Lmax,and μ was not necessary for West Coast cod and a simple von Bertlanffy curve sufficed (Clarke, 2007).

Two forms of natural mortality were chosen. The first applied a natural mortality rate of 0.2 to all ages consistent with the practise used in WGNSDS for Division VIa cod. The second made use of the natural mortality values derived for North Sea cod using the multi species VPA package (MSVPA) (Pope, 1991).

Two sources of maturity at age were considered. Values were those as used at the WGNSDS and those taken from West Coast survey results. In the latter case an average was taken over all fish sampled for age and maturity at each age over the years 1995-2005.

Three models of recruitment were considered; Ricker, hockey stick and Saila Lorda. All models were conditioned using the SSB and recruitment time series from the 2004 VIa cod assessment. The data and resultant fits are shown below.

North Sea haddock case study: In order to condition the model, an initial assessment was run on the available stock and fleet data using FLXSA. Conditioning was required for the following operating model components:

  1. Recruitment time-series. Every future 20-year projection period was assumed to contain exactly two large year-classes, the size of which were taken from a Normal distribution for which the mean was the estimated size of the 1999 year-class at age 0, and the variance was 10% of the mean. The remaining year-class values for each projection were given by a second Normal distribution for which the mean was the geometric mean of the estimates for the 1995-2005 year-classes (without 1999), and the variance was the CV of the same estimates. Finally, the two large year-classes were not allowed to occur consecutively.
  2. Catchability q for each survey. This was obtained by minimising the sum-of-squares between the observed survey indices and the fitted indices q * N^p, where N is estimated abundance and p is a power term.
  3. Selection ζ. This is a measure of how fishing mortality F varies with age a, and is given for each age by the mean of the last three historical F estimates. This is then rescaled so that the sum of selection across all ages is equal to 1.0. In simulations modelling effort-based management, selection ζ must be assumed to be known throughout the simulation period: if this were not the case, there would be no unique solution to the estimation of F that results in the required catch. In this case study ζ is assumed to be fixed throughout the simulation period, although more complicated models of time-varying selection would be possible.
Fisheries & Fleets

VIa cod case study: Under the base case scenario discards and unallocated landings were zero. Under the alternative scenario unallocated landings were modeled according to a curve analogous to the von Bertalanffy growth curve, i.e. landings were assumed to equal

(1)

where Lrep is reported landings, SL is a fitted slope term and Lvb is given by

(2)

where Lumax represents the level of actual landings beyond which the relationship to reported landings is linear and μL is analogous to the von Bertalanffy growth curve parameter. Parameters for equations (1) and (2) were derived from fits to the reported landings and final stock assessment estimates of landings from the 2006 Division VIa cod assessment, using data over the years 1978-2005, (ICES, 2006a). The data and model fit are shown below.

Discards were incorporated into the Fleet model using gear selectivity data coupled with sorting ogive and length distribution at age data following Casey (1996). See case study 2, approach 3 for a fuller explanation of this approach to modelling discards.

North Sea haddock case study: No modelling of fisheries or fleets is included in this case study. Fleets are assumed to adhere precisely to management measures, and there is also assumed to be no changes in targetting behaviour or other fishery behaviour during the simulation period. The analyses are limited to the available scale, which is that used by the relevant ICES working group (i.e. the whole North Sea and Skagerrak).

Economics

VIa cod case study: No economic or socio-economic modelling is included in this case study.

North Sea haddock case study: No economic or socio-economic modelling is included in this case study.

Reference Points

VIa cod case study: Reference points are those agreed by ICES.

North Sea haddock case study: Reference points are those agreed by ICES.

Observation Error Model

VIa cod case study: Under the base case scenario SSB, recruitment and mean F were all considered to be assessed without error. The alternative scenario took the SSB produced from the operating model and applied a random, normal error with C.V. of 30%, following the report on the Evaluation of Research Surveys in Relation to Management Advice (EC, 2003).

North Sea haddock case study: Normally-distributed noise (with a CV of 0.1) was added to observed values of total catch, landings and discards to replicate observation error. Similar noise (with a CV of 0.2) was added to research-vessel survey indices.

Management Procedure

VIa cod case study: As pointed out at the 2006 ICES methods working group (ICES, 2006b) there are points of ambiguity to the cod recovery plan. In particular it is not stated what rule to apply when/if SSB rises above Bpa. As part of the bilateral agreements between the EU and Norway a long term management goal of mean F at or below 0.4 has been set for cod in the Skagerrak, North Sea and eastern Channel. This long term objective was endorsed by an STECF review of the cod recovery plan (STECF, 2007) and so was incorporated into the simulations. Cod recovery is defined as two successive assessments of SSB above Bpa. In the period before recovery five alternatives were considered to illustrate the possible impact of differences in interpretation of the HCR at the point where SSB < Blim. Runs were performed where a) TAC was allowed to be reduced to whatever level necessary to return predicted SSB to above Blim, b) reductions in TAC were always restricted to 15% irrespective of predictions of SSB or mean F and c) reductions in TAC were always restricted to 50%. The forth model applied TACs corresponding to a fishing mortality of 0.4 at all times i.e. regardless of SSB estimates. The final alternative adopted the ‘F sliding scale’ approach (CEC, 2004b) as used for Northeast Arctic cod (ICES, 2007).

(3)

where Flt is the long term fishing mortality used to set TACs (0.4 in this study), Fd is a fishing mortality used to set TACs when the stock is at or below Blim, i.e. depleted and the target F value changes linearly for SSBs assessed between these two values. Fd = 0 in these simulations.

North Sea haddock case study: In each forward projection year in each iteration of the simulation, a series of processes were implemented.

  1. The previously-defined TAC was assumed to be taken in full, unless there were insufficient fish to do so. The pattern of selection at age was assumed to be constant throughout the simulation. The proportion of fish discarded at each age was also assumed to be constant.
  2. Data collation from commercial vessels and research-vessel surveys was simulated.
  3. The data produced were then passed to a stock assessment module (FLXSA) to generate estimated stock abundance and mortality values.
  4. The stock summaries were subsequently used as the basis of a harvest control rule which encapsulated the EU-Norway management plan for North Sea haddock. The TAC for the following year was set according to this plan, on the basis of data taken from the stock assessment.

The process is summarised in the following flowchart:

Results

Work in relation to Evaluation of North Sea Cod Recovery Plans

The first addressed work has been done in cooperation with ICES and CEFAS. Results of the evaluation of the biological effects of North Sea cod recovery plans in relation to various management measures using the modelling in B-Adapt. The B-Adapt Model has been formulated into R/FLR by CEFAS.

The second addressed work has mainly been made under FRS Marine Laboratory national work as an internal research project (ROAME 0352). Results of the work have been reported in the ROAME final report and through a series of FRS internal reports. A figure summarising the results is reproduced below. Earlier work showing results of using FLR to evaluate management plans under the umbrella of current North Sea cod recovery plans are presented in ICES WGNSSK (2006), section 16 and Working Document 18: ICES WGNSSK 2006, section 16 and Working Document 18.

 A comparison of average SSB, average TAC, and recovery probability - P(recovery) for 4 different  HCRs: cod recovery plan (black), cod recovery plan without TAC constraints (red), F sliding scale (green) and a constant F=0.4 (blue). Left panel: under idealised conditions; right panel: under additional mortality which is not properly taken into account by the stock assessment and HCRs.

Work in relation to Evaluation of North Sea Haddock Management Plans

This work has been done in cooperation with ICES and FRS. Results of the evaluation of North Sea haddock management plan in relation to various management measures using modelling in F/RLR is presented in ICES WGNSSK Report (2006), section 16: ICES WGNSSK 2006, section 16.

Dissemination and Links to Other Work

CEFAS, 2006. Results of the evaluation of the biological effects of North Sea cod recovery plans in relation to various management measures using the modelling in B-Adapt is so far given in a national report produced by CEFAS (UK). The B-Adapt Model has later been formulated into R/FLR. CEFAS Report, CEFAS, UK. 2006.

Clarke, E. D. 2007. Evaluating the West of Scotland cod recovery plan using computer simulations, FISHupdate, June 2007.

Clarke E.D. 2007. ROAME MF0352: The evaluation of fisheries management strategies using simulation. Case study: the West of Scotland cod recovery plan. Part 2: Model parameter estimation. Fisheries research Services internal Report, No ##. 10pp.

Clarke E. D. and Holmes, S. J. 2007. The evaluation of fisheries management strategies using simulation. Case study: the West of Scotland cod recovery plan. Part 4: Evaluation of the west of Scotland cod recovery plan. Fisheries research Services internal Report, No ##. 29pp.

Holmes, S.J. 2007. ROAME MF0352: The evaluation of fisheries management strategies using simulation. Case study: the West of Scotland cod recovery plan. Part 3: Discards model parameter estimation and validation. Fisheries research Services internal Report, No ##. 20pp.

ICES, 2007(a). Results of the evaluation of the biological effects of North Sea cod recovery plans in relation to various management measures using the modelling FLR to evaluate management plans under the umbrella of current North Sea cod recovery plans are presented in ICES WGNSSK (2006), section 16 and Working Document 18: ICES WGNSSK 2006, section 16 and Working Document 18.

ICES, 2007(b). Results of the evaluation of North Sea haddock management plan in relation to various management measures using modelling in F/RLR is presented in ICES WGNSSK Report (2006), section 16: ICES WGNSSK 2006, section 16 and Working Document 18. ICES C.M. 2007/ACFM:35

Needle, C. L. (2006a). Evaluating harvest control rules for North Sea haddock using FLR. Working Paper for the ICES Working Group on Methods of Stock Assessment, Galway, Ireland, 21-26 June 2006.

Needle, C. L. (2006b). Further evaluations of harvest control rules for North Sea haddock using FLR. Discussion document for FRS and SEERAD.

Needle, C. L. (2006c). Revised FLR-based evaluation of candidate harvest control rules for North Sea haddock. Working paper for the ICES Advisory Committee for Fisheries Management, Copenhagen, October 2006.

Needle, C. L. (2007a). Management strategy evaluation for North Sea haddock, Fisheries Research, doi:10.1016/j.fishres.2008.03.004

Needle, C. L. (2007b). Management strategy evaluation for North Sea haddock (poster). Haddock 2007, Sheraton Harborside Hotel, Portsmouth NH, USA. 25 October 2007.

References

Casey, J. 1996. Estimating discards using selectivity data: the effects of including discard data in assessments of the demersal fisheries in the Irish Sea. Journal of Northwest Atlantic Fishery Science, 19: 91-102.

CEC. 2004b. EU-Norway ad hoc scientific working group on multi-annual management plans for stocks shared by EU and Norway, Brussels 14 to 18 June 2004. 224pp.

Coull, K. A., Jermyn, A. S., Newton, A. W., Henderson, G. I. and Hall, W. B. 1989. Length/weight relationships for 88 species of fish encountered in the north east Atlantic. Scottish Fisheries Research Report, No 43. 81 pp.

EC. 2003. EVALUATION OF RESEARCH SURVEYS IN RELATION TO MANAGEMENT ADVICE (EVARES), Final Report to European Commission Director - General Fisheries, February 2003, FISH/2001/02 - Lot 1

ICES. 2006a. Report of the Working Group on the Assessment of Northern Shelf Demersal Stocks (WGNSDS), 9-18 May 2006, ICES Headquarters, Denmark. ACFM:30. 870 pp.

ICES. 2006b. Report of the Working Group on Methods of Fish Stock Assessments (WGMG), 21-26 June 2006, Galway, Ireland. ICES CM 2006/RMC:07. 83 pp.

ICES, 2007. Report of the Working Group on the Assessment of Demersal stocks in the North Sea and Skagerrak (WGNSSK). ICES C.M. 2007/ACFM:35.

Needle, C. L. (2006a). Evaluating harvest control rules for North Sea haddock using FLR. Working Paper for the ICES Working Group on Methods of Stock Assessment, Galway, Ireland, 21-26 June 2006.

Needle, C. L. (2006b). Further evaluations of hervest control rules for North Sea haddock using FLR. Discussion document for FRS and SEERAD.

Needle, C. L. (2006c). Revised FLR-based evaluation of candidate harvest control rules for North Sea haddock. Working paper for the ICES Advisory Committee for Fisheries Management, Copenhagen, October 2006.

Pope, J. G. 1991. The ICES Multispecies Assessment Working Group: evolution, insights and future problems. ICES Marine Science Symposia, 193: 22-33.

STECF. 2007. STECF working group on evaluation of the cod recovery plan. Report of the STECF Sub-Group SGRST-07-01. Hamburg, 26-30 March 2007.

Acknowledgements

EFIMAS Contribution to the work

This work has been done in cooperation between EFIMAS, ICES and the National Fisheries Research Institutes in England (CEFAS) and Scotland (FRS).

Participants

Participants: Coby Needle, Steven Holmes, Liz Clarke (FRS), Laurence Kell, Chris Darby (CEFAS), J. Rasmus Nielsen (DIFRES)

 
efimas1/wp4/cs2/appr2/main.txt · Last modified: 2008/11/16 00:20 by admin
 
Except where otherwise noted, content on this wiki is licensed under the following license:CC Attribution-Noncommercial-Share Alike 3.0 Unported
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki