International Council for the Exploration of the Sea Reports/Expert Group Report/aco… · Denmark . Telephone (+45) 33 38 67 00 . Telefax (+45) 33 93 42 15 . . [email protected] . Recommended

ICES IBP PLAICE REPORT 2013 ICES ADVISORY COMMITTEE

ICES CM 2013/ACOM:63

REF. ACOM, WGNSSK

Report of the Inter-Benchmark Protocol for Plaice in Subarea IV (IBP Plaice)

April 2013

By correspondence

International Council for the Exploration of the Sea Conseil International pour l’Exploration de la Mer

H. C. Andersens Boulevard 44–46 DK-1553 Copenhagen V Denmark Telephone (+45) 33 38 67 00 Telefax (+45) 33 93 42 15 www.ices.dk [email protected]

Recommended format for purposes of citation:

ICES. 2013. Report of the Inter-Benchmark Protocol for Plaice in Subarea IV (IBP Plaice), April 2013, By correspondence. ICES CM 2013/ACOM:63. 78 pp.

For permission to reproduce material from this publication, please apply to the Gen-eral Secretary.

The document is a report of an Expert Group under the auspices of the International Council for the Exploration of the Sea and does not necessarily represent the views of the Council.

© 2013 International Council for the Exploration of the Sea

ICES IBP Plaice REPORT 2013 | i

Contents

Executive summary ................................................................................................................ 3

1 Introduction .................................................................................................................... 4

2 Material and methods ................................................................................................... 6

3 Exploratory analysis of four index options ............................................................. 11

3.1 Trimming ages (sep_ages) ................................................................................... 11

3.2 Keeping or removing the historic part of the Isis index (combCor_histvs.combCor) ................................................................................... 12

3.3 Comparison of XSAs with separate (sep) and combined BTS indices (combCor_hist) ...................................................................................................... 12 3.3.1 Index weighting ..................................................................................... 12 3.3.2 Standard errors of index mean log-catchability at age ..................... 12 3.3.3 XSA internal and external standard errors ......................................... 14 3.3.4 Residuals ............................................................................................... 16 3.3.5 Numbers and mortality-at-age............................................................. 16 3.3.6 Retrospective patterns and Mohn Rho values ................................... 16

4 Discussion ..................................................................................................................... 24

5 Conclusion .................................................................................................................... 26

6 References ..................................................................................................................... 27

Annex 1: Splitting of SNS and BTS-Tridens tuning indices ......................... 28

Annex 2: Index values for the BTS-Tridens, BTS-Isis and the Combined BTS ............................................................................................................. 30

Annex 3: Residual for the BTS survey indices in in the sep option and the sep_ages option for exploring the effect of trimming the age ranges included ............................................................................................................ 33

Annex 4: Relative numbers-at-age compared between the two options in which the BTS-indices are combined; with and without keeping the early Isis part of the time-series.......................................................... 38

Annex 5: North Sea plaice XSA assessment outputs and diagnostics for index option sep (using separate BTS-Isis and BTS-Tridens indices) .................................................................................................................. 39

Annex 6: North Sea plaice XSA assessment outputs and diagnostics for index option combCor_hist (using the combined BTS index) ...................... 50

Annex 7: Participants list ...................................................................................... 60

Annex 8: Comments by reviewers ...................................................................... 61

ii | ICES IBP Plaice REPORT 2013

Annex 8.1: Modelling the log ratio of the indices to compare modelled relative catchabilities and empirical estimates .................................... 63

Annex 9: Stock Annex Plaice in Area IV ........................................................... 64

ICES IBP Plaice REPORT 2013 | 3

Executive summary

The stock assessment of North Sea plaice uses three different survey indices as tuning fleets: (1) the Beam Trawl Survey RV Isis (BTS-Isis), (2) the Beam Trawl Survey RV Tridens (BTS-Tridens) and (3) the Sole Net Survey (SNS). In 2008 and 2009, WGNSSK and WKFLAT (the ICES Benchmark and Data Compilation Workshop for Flatfish) respectively addressed concerns about differing signals in the three scientific tuning indices. Based on an observed offshore movement of the younger ages relative to the areas covered by each of the surveys, it was concluded that the trends in the indices of these surveys may no longer independently reflect the trend in abundances of age classes in the whole of the North Sea. As it is considered generally to be better to have indices spanning as much of the distribution as possible rather than having separate indices, which are prone to showing local trends, it was suggested to combine the two separate BTS-indices into one.

For comparison of the effect of combining the indices, the plaice assessment was car-ried out using the XSA model (Darby and Flatman, 1994) written in the R language (v2.13.1; R Development Core Team 2008) using the FLR package (FLCore v2.4; Kell et al., 2007) incorporating data up to 2011. This was done for four different option of incorporating different indices. Model diagnostics and results were examined to ascertain the impact of the choice of indices.

The assessment diagnostics show that the XSA model fit improves using a combined BTS index rather than BTS-ISIS and BTS-Tridens as separate indices. The impact on model results is limited, allowing for a smooth transition to using the new index in the assessment that forms the basis of management for the North Sea plaice stock.

4 | ICES IBP Plaice REPORT 2013

1 Introduction

The current ICES stock assessment of North Sea plaice uses three different survey indices as tuning fleets: (1) the Beam Trawl Survey RV Isis (BTS-Isis), (2) the Beam Trawl Survey RV Tridens (BTS-Tridens) and (3) the Sole Net Survey (SNS) in Sep-tember-October. The BTS-Isis was initiated in 1985 and was set up to obtain indices of the younger age groups of plaice and sole, covering the southeastern part of the North Sea. Since 1996 the BTS-Tridens covers the central part of the North Sea, ex-tending the surveyed area. Owing to the spatial distribution of both BTS surveys, considerable numbers of older plaice and sole are caught. Previously age groups 1 to 4 were used for tuning the North Sea plaice assessment, but the age range has been extended to 1 to 9 in the revision done by ACFM in October 2001.

In 2009 the ICES Benchmark and Data Compilation Workshop for Flatfish (WKFLAT) addressed concerns about differing signals in the three scientific tuning indices (ICES 2009a). Traditionally, the BTS‐Isis survey captured trends in the numbers of plaice of younger ages best, because their spatial sampling effort overlapped most with the distribution of fish of these ages. The BTS‐Tridens survey was believed to capture trends in numbers of fish of older ages best. However, changes in the spatial distribu-tion of plaice of younger ages have resulted in a gradual increase in numbers of younger plaice being caught further offshore by BTS‐Tridens (Kraak et al., 2008). Be-cause of this change in the spatial distribution of the younger ages relative to the are-as covered by each of the surveys, it was concluded that the trends in the indices of these surveys may no longer independently reflect the trend in abundances of age classes in the whole of the North Sea.

The use of a preliminary combined BTS index, compiled by WKFLAT, in XSA did not lead to a strong residual pattern for this index (ICES, 2009b). WKFLAT concluded that this was caused by the fact that the combined index for younger ages follows the trend of BTS-Isis in the past, but deviates from this towards the BTS-Tridens in recent years, and that this was “in accordance with the information and knowledge of the changes in the distribution of fish of ages 2 and 3.” They furthermore pointed out that combining the 2 indices depends on (the quality of) the relative gear efficiency esti-mates. WKFLAT recommended that the adoption of a combined index from BTS-Tridens and BTS-Isis should be checked by WGBEAM.

Based on a visual examination of catch rates by age group and year WGBEAM con-firmed that there was evidence of an offshore shift of the younger age groups (2009b). This finding was supported by Grift et al., 2004. However, their results suggested that the offshore shift mainly occurred in the coastal rectangles, which would indicate that the BTS-Tridens and the BTS-Isis indices might not be severely biased by a shift in distribution between these survey areas. Nonetheless, WGBEAM presented three options for the use of survey indices in the assessment model and recommended further investigation by means of runs of the XSA model:

a ) Using separate index series for Tridens and Isis; b ) Using separate series for Tridens ages 2–9 and Isis ages 1–3; c ) Using the combined index series, to be conducted using a correction for

differences in gear efficiencies between the two research vessels in accord-ance with estimations from earlier WGBEAM work (ICES 2005).


In 2010 WGNSSK conducted exploratory analyses examining the use of the combined BTS survey (option (c) above; ICES 2010b). It was concluded that it had a minimal effect on the perception of the stock in comparison to the default assessment (option (a) above; referred to as SPALY - same procedure as last year - in the WGNSSK re-port) in the early 1990s, while resulting in a slightly higher estimate of SSB and a very slight lower estimate of F in the recent period. Because of these negligible effects on the results, the combined index was not included in the assessment. The effects on recruitment estimates however were not specifically addressed, nor was a thorough investigation of the diagnostics performed.

WGNSSK in 2012 again addressed the issue of differing signals from the two surveys and retrospective patterns in the XSA results (ICES 2012b). Since the XSA model as-sumes a constant catchability over all years, it is unable to handle survey indices with trends or changes in catchability over time. As an alternative solution, exploratory analyses were conducted in which the surveys were split into two separate time peri-ods (up to 1999 and from 2000 to present). Splitting survey time-series allows estima-tion of different catchability coefficients for the two time periods of the survey and effectively breaks the link between the early and recent survey data in the assess-ment. While this slightly reduced the magnitude of retrospective patterns it also re-sulted in using relatively short time-series (especially for the BTS-Tridens for which the time period 1996–1999 would be too short for inclusion in the XSA). See also ap-pendix 1. It was concluded not to further pursue this possible makeshift solution. Instead, it was decided to further investigate the use of the combined in advance of WGNSSK 2013.

WGBEAM in June 2012 produced separate and combined indices and made a number of observations on the observed trends. The Isis survey indicated that recruitment has been well below average in most years since the strong 2001 year class became appar-ent as one year olds in 2002. Only in 2009 and 2011, the observed number of one year olds was higher than the long-term mean; with the 2011 recruitment-at-age one being the fifth highest of the time-series. The Tridens survey confirmed the strong 2001 year class, and documented above average incoming year classes from 2007 onwards, with the observed number of one year olds in 2011 being the highest of the time-series. This pattern is visible at all ages in this survey. The combined Isis-Tridens index showed above average numbers at all ages, with an increasing trend since the begin-ning of the 21st century, building up to the biggest 4+ group in the time-series. It is not clear where the higher numbers of four year olds in 2007–2009 come from in the Tridens and combined indices.

The current report describes results of exploratory analyses using different index combinations in the current ICES XSA assessment of the stock. On this basis a rec-ommendation is made on the usage of separate or combined BTS survey indices. This report is presented for review to IBPPlaice 2013.


2 Material and methods

Calculation of the combined indices was based on the same approach as the separate indices, i.e. catch rates (N/hour) were averaged by ICES rectangle and then averaged for all ICES rectangles within the index-area. The index-area consisted of the com-bined Isis and Tridens index areas (see Figure 1) plus ICES rectangle 34F2 (as it is fished in almost all years either by Isis or by Tridens).

Figure 1. Left: Survey area of the BTS survey indicating the index areas of Tridens (blue + green shading) and Isis (yellow + green shading) (from ICES 2009b). Right: Length–frequency distribu-tions of plaice and sole for each of the Isis (BTS) and the Tridens (BTS+) given as the mean annu-al relative abundance (+ s.e.), based on numbers per hour (from ICES 2005).

Both survey vessels use an 8 m beam trawl with 40 mm stretched mesh codend, but the Tridens is rigged with a modified net, using a flip-up rope. It is difficult to obtain reliable gear efficiency estimates for the comparison of the two vessels and determine a correction factor for combining the indices. Data based on simultaneous hauls from the same vessel are scarce (although available for 81 hauls from Rijnsdorp and Groeneveld 1990; see left hand side of Table 1). WGBEAM 2009 however decided to calculate combined BTS indices using gear correction factors as estimated by WGBEAM in 2005 based on the comparison of several thousands of Isis and Tridens hauls in relation to the distance between these hauls (ICES 2005; see right hand side of Table 1). The final indices (separate and combined) are presented in Appendix 2.

Table 1. Estimates of correction factors to account for differences in gear efficiency of the BTS-Isis compared to the BTS Tridens.


Indices

Standardized index values for the separate and combined BTS indices are shown in Figure 2. In general the indices exhibit the same trends over time. Strong cohorts can be identified in all indices; though there are some year effects observed in the older ages in the BTS-Isis index (e.g. age 5 in 2011 and age 8 in 2004). The catchbility of the older ages in the BTS-Isis survey is very low so occasional large hauls can lead to larger than usual cpues.

year

stan

dard

ized

inde

x -10

12

3

2000 2005 2010

1

-10

12

3

2000 2005 2010

2

-10

12

3

2000 2005 2010

3

-10

12

2000 2005 2010

4

-10

12

3

2000 2005 2010

5

-10

12

2000 2005 2010

6

01

23

2000 2005 2010

7

01

2

2000 2005 2010

8

01

23

2000 2005 2010

9

BTS-Comb BTS-Tridens BTS-ISIS

Figure 2. Standardized index values for ages 1–9 of the separate (BTS-Isis and BTS-Tridens) and combined (BTS-comb) indices.

The combined index follows the BTS-Isis very closely for age 1 and BTS-Tridens very closely for ages 5 and older. For ages 2 and 3, and to a lesser degree age 4, the com-bined index averages out the signals from the two separate BTS indices. This is in agreement with how the XSA weights the different ages when fitting the separate indices, with the BTS-ISIS having higher weightings for the younger ages (particular-ly age 1) and the BTS-Tridens having greater weighting for the older ages (see Section 6). When the WGNSSK first tested the combined index in 2010 (ICES 2010) they not-ed that the apparent minor effect on assessment results was likely due to the weight-ings at age used in combining the two indices being very similar to the relative weightings at age assigned to each index when fit separately in the XSA.


In addition to the BTS indices, all assessments conducted used the SNS index for ages 1–3 (see ICES 2009a and ICES 2012b). The internal consistency of the four indices examined is plotted in Figure 3. The BTS-Tridens shows strong relationships between numbers-at-age in consecutive years, with the exception of age 1 vs. age 2. The BTS-Isis generally shows weaker internal consistency, though it is strongest for age 1 vs. age 2. The combined BTS index has better internal consistency than the BTS-ISIS for all ages except age 1 vs. age 2 and has similarly good consistency from ages 4 and up compare to the BTS-Tridens. For age 2 vs. age 3 and age 3 vs. age 4 the combined index is more consistent than the BTS-Isis, but less so than the BTS-Tridens. Overall, the level of internal consistency for the combined index is good, making it suitable for use in an assessment model.

Figure 3. Internal consistency between ages of the four indices examined. Lower right panels show the coefficients of determination (R2) of the correlations. White/light yellow indicates poor correlations, red indicates strong correlations.


Stock assessments

For comparison of the effect of combining the indices, the plaice assessment was car-ried out using the XSA model (Darby and Flatman, 1994) written in the R language (v2.13.1; R Development Core Team 2008) using the FLR package (FLCore v2.4 and FLXSA v2.0; Kell et al., 2007) incorporating data up to 2011. This was done for four different options of incorporating different indices (Section 3).

The XSA uses the reconstructed discard set described in the North Sea plaice stock annex. The data used and control settings for the assessment follow those agreed at the last benchmark of the stock (ICES 2009), and are given in Table 2. No further tun-ing of the assessment was carried out to account for the change in indices. The as-sumptions made regarding the treatment of the indices remain valid for the BTS_comb index. All other data were kept the same so that the impact of the change in indices can be easily examined by comparing the alternate assessment results.

Table 2. Inputs and settings used to fit XSA models to the alternative index options.

SETTING/DATA VALUES/SOURCE

Catch-at-age Landings (since 1957, ages 1–10) + (reconstructed) discards based on NL, DK + UK + GE fleets. Discards reconstruction between 1957–1999), observations since 2000

Tuning indices Index option sep Index option combCor_hist

BTS-Isis 1985–2011 1–8 BTS-Tridens 1996–2011 1–9 SNS 1982–2007 1–3

BTS-Isis 1985–2011 1–8 BTS-Tridens 1996–2011 1–9 SNS 1982–2007 1–3

Plus group 10

First tuning year 1982

Time-series weights No taper

Catchability dependent on stock size for age <

1

Catchability independent of ages for ages >=

6

Survivor estimates shrunk towards the mean F

5 years / 5 years

s.e. of the mean for shrinkage 2.0

Minimum standard error for population estimates

0.3

Prior weighting Not applied

Mohn’s Rho

The rho statistic of Mohn (Mohn, 1999) has been commonly used to measure the ret-rospective pattern. It is defined as the sum of annual relative differences between an estimated quantity from an assessment with a reduced time-series and the same quantity estimated from the full time-series:


where X denotes some variable from the stock assessment such as F or SSB, y denotes year, tip denotes the terminal estimate from an assessment with a reduced time-series, and ref denotes the assessment using the full time-series. We use this statistic in the results section to assess the retrospective error.


3 Exploratory analysis of four index options

The plaice assessment was run using the XSA model for four different index options. Table 3 shows these options which incorporate (1) the separate BTS indices and (2) the combined indices corrected for gear efficiency in accordance with estimates by WGBEAM in 2005. Since the BTS-Isis survey started in 1985 and the BTS-Tridens in 1996, the combined index is calculated for the period 1996 until present, while the Isis index is kept for the period 1985–1996 as a separate index. An additional run (3) was done to investigate the effect of removing the early part of the BTS-Isis index. A final option (4) was included because WGBEAM suggested this option to be investigated in 2009. In all cases the model converged within 50 iterations, the maximum limit defined in the stock annex. The XSAs using the combined index converged in fewer iterations. WGMG (ICES 2009c) has advised that XSA outcomes can be impacted if the number of iterations needed to converge it is too high, but in the case of the North Sea plaice XSA convergence is not a significant issue.

Table 3. Four index options for which the assessment was run.

INDEX

OPTION DESCRIPTION NUMBER OF

INDICES CONVERGENCE (NUMBER

ITERATIONS)

1 sep Current assessment: SNS, BTS-ISIS 1985–2011 and BTS-Tridens 1996–2011

3 41

2 combCor_hist

SNS, Combined BTS index 1996–2011, BTS-ISIS index 1985–1995

3 36

3 combCor SNS, Combined BTS index 1996–2011 2 35

4 sep_ages SNS, BTS-ISIS ages 1–3, BTS-Tridens ages 2–9 3 42

3.1 Trimming ages (sep_ages)

First of all we investigate the option suggested by WGBEAM in 2009 which keeps the two BTS-indices included separately in the assessment, but reduces the age ranges that are used in the assessment. Currently, ages 1–9 are used from the Tridens index and ages 1–8 from the Isis index (because there have been a few years with zero catches of nine year old fish in the Isis). In this option only ages 2–-9 are included from the Tridens and only ages 1–3 are included from the Isis, in order to reduce the overlap between the two indices which might resolve part of the conflict between the signals. Despite the fact that we feel that it is generally preferable to use as much of the data as possible we investigated this option since it is suggested by WGBEAM. Selected results are presented in Appendix 3.

It is clear that by trimming the age ranges as suggested, indeed those ages which show the strongest residual patterns are removed from the assessment. This in turn reduces the maximum and minimum residual values in absolute terms. This effect is especially clear for the Tridens age 1 where large residuals are found. Since the age 1 from the Tridens is weighted by the assessment with less than 5% however, it is to be expected that this has very little effect. The patterns in residuals in the ages that are kept are not resolved.

Investigation of the retrospective plots shows that the retrospectives in SSB or F are not affected. The retrospective in recruitment gets worse however. For this reason, as


well as for the fact that we generally would like to keep as much of the available data in the assessment, we decide to not further pursue this option.

3.2 Keeping or removing the historic part of the Isis index (comb-Cor_histvs.combCor) Exploratory diagnostics show that removing the early part of the Isis time-series does not have any effect on the recent retrospective patterns for SSB, F or recruitment (re-sults available on request). However, it does affect the model’s estimates of numbers-at-age around the ‘breakpoint’ between the early part of the Isis index and start of the combined index relative to the current assessment in which the indices are kept sepa-rate. This effect on estimates of numbers-at-age is also present when the early Isis index is kept in, but it is a lot smaller (see Appendix 4). For this reason, we decide to not further pursue this option and thus exclusively focus on further describing the characteristics, diagnostics and results of running the plaice assessment for index options 1 (sep) and 2 (combCor_hist).

3.3 Comparison of XSAs with separate (sep) and combined BTS indices (combCor_hist) The full assessment diagnostics for options sep and combCor_hist are presented in Appendix 5 and Appendix 6, respectively. Selected key diagnostics are presented in plots in this section.

3.3.1 Index weighting

In order to investigate the relative contributions of the indices and F-shrinkage to the estimates of the terminal numbers-at-age in the two options, we investigate the weights attributed to the indices (Figure 4). The most substantial effect can be ob-served in ages 1–3, where the combined BTS index is weighted more than the weight that the two separate BTS indices received together in the separate index option. This is at the expense of the weight put on the SNS survey index.

3.3.2 Standard errors of index mean log-catchability at age

The standard error of the log-catchability at age, a fractional coefficient of variation of the fleet's catchability for that age, is an indicator of the quality of the data (Rivard et al., 2003). Values exceeding 0.5 can indicate problems fitting the assessment to given ages of the index. The log-catchability standard errors for the indices used in the sep and combCor_hist index options are shown in Table 4. When the BTS indices are used separately, the BTS-ISIS has high catchability SEs for the older ages (particularly age 8) and the BTS-Tridens has very high SEs for ages 1 and 2. Using only the old period of the BTS-ISIS (1985–1995; BTS-ISIS_old) leads to smaller SEs at the older ages, but still very high for age 8. The combined BTS index has catchability SEs lower than 0.5 for all ages. The catchbility SEs for the SNS survey are high in both cases, for all ages.


Table 4. Index standard errors of log-catchability for ages with catchability independent of year-class strength and constant over time. Values exceeding 0.5 are indicated in bold.

INDEX OPTION

Sep combCor_hist

age BTS-ISIS BTS-Tridens SNS BTS-comb BTS-ISIS_old SNS

1 0.45 1.60 0.51 0.28 0.61 0.53

2 0.50 0.73 0.83 0.32 0.46 0.86

3 0.47 0.34 1.01 0.25 0.39 1.02

4 0.31 0.32 0.24 0.39

5 0.46 0.30 0.25 0.42

6 0.51 0.30 0.29 0.49

7 0.59 0.30 0.33 0.39

8 0.84 0.31 0.24 0.76

9 0.35 0.37

Figure 4. Relative weights attributed to the different indices by the model.


3.3.3 XSA internal and external standard errors

The internal and external standard error of the logtransformed terminal population can be considered as a weighted mean of the variance of the log-catchability coefficients estimated for the individual set of abundance indices (see Darby and Flatman, 1994; Vinther 2001). The internal standard error corresponds to the within fleet variance whereas the external standard error corresponds to the among fleets variance. The internal and external standard errors for the XSAs fit using the two index options are shown in Figure 5. Both the internal and external standard errors follow the same pattern at age for both index options: higher for the young ages, de-creasing to age 4 and then stabilizing at relatively low levels. Using the combined BTS index decreases the internal standard error on the younger ages and very slightly increases the SEs at older ages. Differences in external standard errors between the two options, though large on a relative scale for the older ages, are minor.


Figure 5. The internal (top) and external (middle) standard errors of the XSA assessments fit to the two index options: sep and combCor_hist. The relative percentage changes are also shown (bot-tom).


3.3.4 Residuals

Figures 6 and 7 show the residual plots for the separate and combined index options. It is clear from Figure 6 that the model fits better to the combined BTS index, since overall the residuals are substantially reduced in size and most importantly, the strong pattern observed in the youngest ages in the BTS-Tridens index is completely removed in the combined BTS index. Figures 7 shows the effect of combining the BTS indices (which is only done for the period of 1996–2011) on the model fit to the early BTS-Isis index which is kept in the model as a separate index for the period 1985–1995. There appears to be no substantial effect on the residuals. Likewise the residual plots for the SNS survey index show that the effect of combining the BTS-indices ap-pears negligible.

3.3.5 Numbers and mortality-at-age

Figure 8 shows the effect of combining the indices on the model’s estimates of num-bers and mortality-at-age for all ages separately. For the youngest ages (1–2) the model estimates approximately 20–30% higher numbers respectively. Consequently, the F on those ages is estimated to have been 20–30 percent lower in most recent years. This is to be expected, considering that the Tridens estimated larger numbers at these ages, but received very little weight by the model for those ages. Numbers for the older ages (3–10) are estimated approximately 10–20% higher and F 10–20% low-er.

Figure 9 shows the assessment outputs for SSB, mean F and recruitment, comparing the sep and combCor_hist options. Recruitment (age 1) as mentioned is estimated ap-proximately 20% higher with the combined indices. SSB is estimated slightly higher in recent years (up to 10% in 2011). Consequently, mean F is estimated approximately 10% lower in 2011.

3.3.6 Retrospective patterns and Mohn Rho values

Figures 10, 11 and 12 show the retrospective plots for both index options as well as bar plots for the comparison of the annual deviations of the terminal retrospective values from the final assessment run. The Mohn’s Rho value is the sum of these an-nual deviations. Table 5 below presents the Mohn Rho values and shows that for SSB, mean F and recruitment for the last five retrospective ‘peels’. In all cases the retro-spective error is slightly reduced.

Table 5. Mohn Rho values as indicator for the 5-year retrospective error for both index options.

SSB MEANF REC

sep 0.67 -1.01 1.51

combCor_hist 0.65 -0.86 1.46


2000 2005 2010

BTS-ISIS residuals: sep

Year

Age

13

57

9

Mod>ObsObs>Mod

-0.590.52-0.840.46-0.830.8-0.510.5

-1.240.83-0.99 1.07

-1.93 0.98-1.66 1.31

max = 1.31; min = -1.93

2000 2005 2010

BTS-Tridens residuals: sep

Year

Age

13

57

9

-3.97 1.83-1.37 1.24

-0.54 0.39-0.49 0.6

-0.5 0.46-0.49 0.56

-0.81 0.38-0.460.56

-0.530.82

max = 1.83; min = -3.97

2000 2005 2010

BTS-comb residuals: combCor_hist

Year

Age

13

57

9

-0.610.38-0.77 0.46

-0.50.37-0.33 0.42

-0.460.4-0.490.46

-0.8 0.35-0.410.4

-0.780.71

max = 0.71; min = -0.8

Figure 6. Survey residuals for the BTS-Isis (top), BTS-Tridens (middle) and combined BTS-index (bottom).


1986 1988 1990 1992 1994


Year

Age

13

57

9

Mod>ObsObs>Mod

-1.22 0.58-0.27 1.23

-0.28 0.95-0.54 0.58

-0.660.68-0.7 0.82

-0.74 0.8-1.15 1.8

max = 1.87; min = -1.22

1986 1988 1990 1992 1994

BTS-ISIS residuals: combCor

Year

Age

13

57

9

-1.15 0.65-0.59 0.9

-0.44 0.74-0.6 0.57

-0.60.74-0.74 0.82

-0.74 0.68-1.18 1.7

max = 1.7; min = -1.18

1985 1990 1995 2000 2005 2010

SNS residuals: sep

Year

Age

12

3

Mod>ObsObs>Mod

-0.731.03

-1.1.2

-1.661.99

max = 1.99; min = -1.66

1985 1990 1995 2000 2005 2010

SNS residuals: combCor_hist

Year

Age

12

3

-0.821.05

-1.1.22

-1.682

max = 2; min = -1.85

Figure 7. Survey residuals for the early part of the BTS-Isis time-series (left) and the SNS-survey (right) for the options sep (top) and combCor_hist (bottom).


Figure 8. Numbers-at-age and mortality-at-age for the combCor_hist option relative to the sep option.

1960 1970 1980 1990 2000 2010

0.7

0.8

0.9

1.0

1.1

1.2

1.

Year

Rel

ativ

e F

At A

ge

Age 2Age 3Age 4Age 5Age 6

1960 1970 1980 1990 2000 2010

0.6

0.8

1.0

1.2

1.4

Year

Rel

ativ

e F

At A

ge

Age 7Age 8Age 9Age 10

1960 1970 1980 1990 2000 2010

0.7

0.8

0.9

1.0

1.1

1.2

1.3

combCor_hist / sep

Year

Rel

ativ

e F

At A

ge

Age 1Age 2Age 3

1960 1970 1980 1990 2000 2010

0.7

0.8

0.9

1.0

1.1

1.2

1.3

combCor_hist / sep

Year

Rel

ativ

e N

umbe

rs A

t Age

Age 1Age 2Age 3

1960 1970 1980 1990 2000 2010

0.7

0.8

0.9

1.0

1.1

1.2

1.3

Year

Rel

ativ

e N

umbe

rs A

t Age

Age 2Age 3Age 4Age 5Age 6

1960 1970 1980 1990 2000 2010

0.7

0.8

0.9

1.0

1.1

1.2

1.3

Year

Rel

ativ

e N

umbe

rs A

t Age

Age 7Age 8Age 9Age 10


Figure 9. Assessment results for recruitment, SSB and mean F for the sep and combCor_hist op-tions.

1960 1970 1980 1990 2000 2010

0.90

0.95

1.00

1.05

1.10

Year

Rel

ativ

e S

SB

1960 1970 1980 1990 2000 2010

0.90

0.95

1.00

1.05

1.10

Year

Rel

ativ

e M

ean

F

1960 1970 1980 1990 2000 2010

0.8

0.9

1.0

1.1

1.2

combCor_hist / sep

Year

Rel

ativ

e R

ecru

itmen

t (A

ge 1

)

1960 1970 1980 1990 2000 2010

0.0

0.2

0.4

0.6

0.8

Year

Mea

n F

(age

s 2-

6)

sepcombCor_hist

1960 1970 1980 1990 2000 2010

0e+0

01e

+06

2e+0

63e

+06

4e+0

65e

+06

Year

Rec

ruitm

ent (

'000

s)

sepcombCor_hist

1960 1970 1980 1990 2000 2010

0e+0

01e

+05

2e+0

53e

+05

4e+0

55e

+05

Year

Spa

wne

r Sto

ck B

iom

ass

(kt)

sepcombCor_hist


1990 1995 2000 2005 2010

050

0000

1500

000

sep

Year

Rec

ruitm

ent (

Age1

)

1990 1995 2000 2005 2010

050

0000

1500

000

combCor_hist

Year

Rec

ruitm

ent (

Age1

)

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Annu

al M

ohn'

s rh

o

-0.6

-0.2

0.2

0.6

sepcombCor_hist

Underestimate

Overestimate

Figure 10. Retrospective plots for recruitment for option sep (top) and combCor_hist (middle). The bottom plot shows the annual deviations of the terminal estimate in the retrospective run vs. the final assessment estimate. (the sum of these represents the Mohn Rho value).


1990 1995 2000 2005 2010

0e+0

02e

+05

4e+0

5

sep

Year

Spaw

ner S

tock

Bio

mas

s

1990 1995 2000 2005 2010

0e+0

02e

+05

4e+0

5

combCor_hist

Year

Spaw

ner S

tock

Bio

mas

s

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Annu

al M

ohn'

s rh

o

-0.4

-0.2

0.0

0.2

0.4

sepcombCor_hist

Underestimate

Overestimate

Figure 11. Retrospective plots for SSB for option sep (top) and combCor_hist (middle). The bot-tom plot shows the annual deviations of the terminal estimate in the retrospective run vs. the final assessment estimate. (the sum of these represents the Mohn Rho value).


1990 1995 2000 2005 2010

0.0

0.2

0.4

0.6

0.8

1.0

sep

Year

Mea

n F

(age

s 2-

6)

1990 1995 2000 2005 2010

0.0

0.2

0.4

0.6

0.8

1.0

combCor_hist

Year

Mea

n F

(age

s 2-

6)

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Annu

al M

ohn'

s rh

o

-0.4

-0.2

0.0

0.2

0.4

sepcombCor_hist

Underestimate

Overestimate

Figure 12. Retrospective plots for the mean F for option sep (top) and combCor_hist (middle). The bottom plot shows the annual deviations of the terminal estimate in the retrospective run vs. the final assessment estimate. (the sum of these represents the Mohn Rho value).


4 Discussion

There have been concerns about differing signals in the three scientific tuning indices used in the North Sea plaice stock assessment since the stock was last benchmarked by ICES in 2009. WGBEAM thoroughly examined the potential for combining the BTS indices and showed that combining the BTS index is practically possible. The surveys occur at the same time of year, with minimal overlap and use similar gears. The use of a flip-up rope in the Tridens survey gear does result in some differences in catcha-bility at age. Quantifying these differences is challenging, but two recent studies were used to test different options. WGBEAM concluded that the use of gear effi-ciency correction factors was more advisable (Loes Bolle, IMARES, pers. comm.).

The resulting combined index follows the BTS-ISIS very closely for age 1 and the BTS-Tridens for ages 5+. For ages 2–4 the combined index lies between the two separate indices. The combined index is internally consistent and covers a large area of the distribution of the stock.

WGNSSK looked at the use of the combined index in 2010. At the time it was decid-ed against changing to the use of the combined index on the basis that incorporating it had minimal impact on assessment results (SSB and F). However, using a single consistent index covering a broader range is still theoretically better than using two separate indices from different areas of the stock distribution. The work presented here represents a more thorough analysis of XSA diagnostics and results to determine if the use of the combined index provides a better basis for management, regardless of the short-term impacts on stock status outputs form the assessment model.

An alternative method for dealing with retrospective patterns is to split indices so that catchabilities can be estimated for particular time periods. If it can be easily demonstrated that a step change has occurred in survey catchability, then splitting the surveys in the appropriate year could remove the model misspecification. This should in turn reduce retrospective patterns and produce more accurate stock and F estimates. In 2012 WGNSSK explored various options for splitting the indices (ICES 2012). However, given the gradual nature in the change of catchability and the loss of data from some indices that needed to be trimmed for the splitting it was decided that this would not adequately address the conflicting signals in the indices. There-fore it was decided once more to further explore the option of combining the BTS indices.

WGBEAM recommended four index options for WGNSSK to consider for use in the XSA. All options were explored, but the focus was on the current assessment option (BTS-ISIS and BTS-Tridens separate) and compCor_hist (BTS-combined, BTS_ISIS 1985–1995 retained). Theoretically the combined option is considered the most suita-ble alternative to the current assessment. Examination of results for other index op-tions suggested that the XSA model fit did not improve (all results available on request). In addition, reducing the age ranges of the BTS-ISIS and BTS-Tridens (index option sep_ages) results in the loss of a lot of potentially meaningful data on stock abundance. Retaining the older period of the BTS-Isis allows for stability in the esti-mates for the historic time period while having a limited effect on terminal years. During the retained time period (1985–1995) the index is thought to follow the as-sumption of constant catchability.

The results obtained here also highlighted some problems associated with the use of the SNS in the North Sea plaice assessment. The shift of younger ages off shore is more problematic for the SNS survey than for the BTS-Isis. A very clear and strong


residual pattern is observed and the inclusion of this survey is in part to blame for the persistent retrospective pattern. Fortunately, using the combined BTS index, results in a reduced weight in the model for the SNS. Combining the SNS with other sur-veys is not an option since the gear used differs from the BTS survey and correcting for differences in gear efficiency would be a major difficulty. The use of this index in future assessments could be examined further e.g. using tapered time-series weighting, splitting the index or removing it from the assessment entirely.

WGMG 2009 showed through simulation tests that in cases where convergence takes many iterations there is a tendency for further iterations to move the assessment away from the underlying true population state. They concluded that it is essential to determine the convergence characteristics of any XSA assessments, and alternative methods need to be explored in cases where convergence is slow and leads to large changes in perceived stock dynamics. For all the index options examined the XSAs converged in less than the maximum (50) specified by the stock annex, converging in fewer iterations when using the combined BTS index.

The results presented comparing the sep and combCor_hist index options looked at both assessment diagnostics and outputs. Any decision on whether to use the com-bined or separated indices should be based rather on the quality of the index, diag-nostics of the model fit and on its self-consistency over time (retrospective) rather than how the resulting assessment outputs differ. The combined index is self-consistent with strong positive inter-age correlations. The XSA fit to the combined index had reduced standard errors across all ages with no notable high uncertainties for young ages (like the BTS Tridens) or the old ages (like BTS ISIS). This also leads to smaller internal SEs in the XSA (equivalent to narrower confidence bounds), espe-cially for the young ages. There is no significant residual patterns in the combined index and residuals are smaller in size overall. Some year effects remain in the resid-ual (e.g. 2001 -, 2008 +), but the patterns and large size of residuals seen in the young ages of the Tridens are removed. The effect on assessment results is minor (SSB up, F down) and there is a very slight improvement in all five year retrospectives, though the persistent directional retrospective patterns persist.


5 Conclusion

The assessment diagnostics show that the XSA model fit improves using a combined BTS index rather than BTS-ISIS and BTS-Tridens as separate indices. The impact on model results is limited, allowing for a smooth transition to using the new index in the assessment that forms the basis of management for the North Sea plaice stock. It is recommended that in future the assessment for North Sea plaice should use the combined index. It is further recommended to explore the use of the SNS survey in the assessment (time tapered weighting or splitting).


6 References

Darby C.D. and Flatman S. 1994. Virtual Population Analysis: Version 3.1 (Windows/DOS) User Guide. Information Technology Series, No. 1(MAFF, Directorate of Fisheries Re-search, Lowestoft) 85 pp.

ICES. 2005. Report of the Working Group on Beam Trawl Surveys (WGBEAM), 7–10 June 2005, Lowestoft, UK. ICES CM 2005/G:12, 83pp.

ICES. 2009a. Report of the Benchmark and Data Compilation Workshop for Flatfish (WKFLAT 2009). ICES CM 2009/ACOM:31.

ICES. 2009b. Report of the Working Group on Beam Trawl Surveys (WGBEAM). ICES CM 2009/LRC:04.

ICES. 2009c. Report of the Working Group on Methods of Fish Stock Assessment (WGMG). ICES CM 2009/RMC:12.

ICES. 2010a. Report of the Working Group on Beam Trawl Surveys (WGBEAM). ICES CM 2010/SSGESST:17.

ICES. 2010b. Report of the Working Group on the Assessment of Demersal Stocks in the North Sea and Skagerrak (WGNSSK). ICES CM 2010/ACOM:13.

ICES. 2012a. Report of the Working Group on Beam Trawl Surveys (WGBEAM). ICES CM 2012/SSGESST:11.

ICES. 2012b. Report of the Working Group on the Assessment of Demersal Stocks in the North Sea and Skagerrak (WGNSSK). ICES CM 2012/ACOM:13.

Grift, R.E., I. Tulp, L. Clarke, U. Damm, A. McLay, S. Reeves, J. Vigneau and W. Weber, 2004. Assessment of the ecological effects of the Plaice Box. Report of the European Commission Expert Working Group to evaluate the Shetland and Plaice boxes.

Kell, L. T., Mosqueira, I., Grosjean, P., Fromentin, J. M., Garcia, D., Hillary, R., Jardim, E., et al. 2007. FLR: An open-source framework for the evaluation and development of manage-ment strategies. ICES Journal of Marine Science, 64: 640–646.

Kraak, S.B.M., Buisman, F.C., Dickey-Collas, M., Poos, J.J., Pastoors, M.A., Smit, J.G.P., Van Oostenbrugge, J.A.E., and Daan, N. 2008. The effect of management choices on the sus-tainability and economic performance of a mixed fishery: A simulation study. ICES Jour-nal of Marine Science 65 (4), pp. 697–712.

Mohn, R. 1999. The Retrospective Problem in Sequential Population Analysis: An Investigation Using Cod Fishery and Simulated Data, ICES Journal of Marine Science, 56, 473–488.

R Development Core Team. 2008. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.

Rijnsdorp and Groeneveld. 1990. The effect of a flip-up rope on the catch efficiency of an 8-m beam trawl. C.M. 1990/B: 16.

Vinther, M. 2001. Ad hoc multispecies VPA tuning applied for the Baltic and North Sea fish stocks. ICES Journal of Marine Science, 58: 311–320.


Annex 1: Splitting of SNS and BTS-Tridens tuning indices

(From the WGNSSK 2012 report)

In recent years, the XSA catchability residuals exhibit pronounced trends for ages 1–3: they are consistently negative for the SNS and consistently positive for BTS-Tridens. This is likely to be explained by a movement of young plaice out of the area of the SNS into the area of the BTS (Beare et al., 2010). Juvenile plaice have been distributed more offshore in recent years. Surveys in the Wadden Sea have shown that 1-group plaice are almost absent from the area where they were very abundant in earlier years. This could be linked to environmental changes in the productivity or changes in the temperature of the southern North Sea, but these links have not been shown conclusively. The distribution of the SNS overlaps largely with the Wadden Sea, and the SNS receives high weightings in XSA in the tuning of trends of plaice of age groups 1–3 due to its historically stronger correlation with the VPA. The expected net effect of these changes in catchability would be an underestimation of recruitment strength. This is also seen in the retrospective pattern of recruitment in recent assess-ments of the stock.

Following initial tests at the previous working group, further analyses investigating the sensitivity of the assessment output to this were conducted. Various combinations of division (splitting) of the SNS and BTS-Tridens tuning indices were examined (see text table below). In all cases indices were split at year 2000 (<2000 and >=2000) as opposed to year 2004 as done previously. Previous splitting indices were based on the pattern of residuals for the indices, but further examination of available data and the plaice box report (Beare et al.) suggest 2000 to be a more appropriate year to sepa-rate present from past distribution of plaice juveniles.

RUN NAME DESCRIPTION

Original All three indices in full, following stock annex

SplitSNS Only the SNS index split*

SplitBoth Both SNS and BTS-Tridens split*

SplitNew Both SNS and BTS-Tridens split*, only >=2000 BTS-Tridens index retained

SplitOld SNS index split*, only ages 4–9 of BTS-Tridens used (no need to split)

SplitOldrecYng SNS split*, BTS-Tridens divided into two indices: full time-series ages 4–9 and >=2000 ages 1–3

*All splits divide indices into <2000 and >=2000.

Assessment runs have been done with these split tuning indices (Figure 8.3.1). Split-ting the indices raises SSB slightly in all cases except SplitOld. IN this case removing the young ages in the BTS-Tridens index lowers the estimated recruits significantly in the recent period and the general lower level of year-class strength leads to lower SSB. In general splitting the indices has a very limited impact on F, though in most cases this leads to an estimation of higher recruitment in the last two years.

It was decided that while splitting the indices is not the ultimate solution to this prob-lem, it remains clear that recruitment is probably underestimated by the model. This will be taken into account when determining the level of recruitment to use in the short-term forecast.


Figure Split.1. North Sea plaice. Sensitivity of the assessment with respect to assumptions on catch-ability of indices over time (by splitting the SNS and/or BTS Tridens indices at the year 2000; see text for details). XSA results with respect to recruitment (top), F (bottom left) and SSB (bottom right) estimates. Note: some lines may be hidden due to near identical outputs.


Annex 2: Index values for the BTS-Tridens, BTS-Isis and the Combined BTS

BTS-Tridens

YEAR 0 1 2 3 4 5 6 7 8 9

1996 0.000 1.643 6.021 4.451 2.903 2.039 1.566 0.721 0.415 0.190

1997 0.000 0.221 7.119 9.127 3.252 2.105 1.523 0.401 0.819 0.354

1998 0.000 0.228 32.249 9.572 4.874 2.202 1.274 0.929 0.762 0.304

1999 0.054 2.692 7.711 35.228 5.558 2.498 1.928 0.633 0.761 0.309

2000 0.043 4.795 13.445 12.910 16.957 2.882 1.716 0.933 0.805 0.218

2001 0.178 2.154 8.612 9.901 6.681 7.360 1.055 0.592 0.418 0.505

2002 0.000 18.553 12.912 9.541 6.411 4.181 4.420 0.743 0.741 0.394

2003 0.338 3.975 41.692 13.378 9.059 5.077 2.806 3.920 0.703 0.740

2004 0.014 5.985 15.784 31.488 9.430 4.316 2.439 1.242 2.500 0.409

2005 0.043 6.876 23.366 12.234 17.672 2.824 6.871 1.565 0.567 3.574

2006 0.236 6.725 32.192 25.727 11.367 10.918 1.985 3.897 0.864 0.723

2007 0.000 26.571 23.735 19.551 23.175 4.900 10.147 1.974 3.786 0.323

2008 0.000 17.467 50.462 25.585 18.392 18.974 6.243 12.747 2.657 6.749

2009 0.116 12.110 41.685 43.331 19.126 12.052 11.768 3.081 10.119 1.567

2010 0.644 26.180 35.716 34.561 30.093 13.412 5.695 12.234 2.744 6.362

2011 0.174 41.881 71.478 41.593 28.462 31.670 14.284 5.501 11.881 1.172


BTS-Isis

YEAR 0 1 2 3 4 5 6 7 8 9

1985 595.271 136.759 173.893 36.059 10.997 1.273 0.973 0.336 0.155 0.091

1986 9.303 667.441 131.704 50.173 9.208 3.780 0.400 0.418 0.147 0.070

1987 44.126 225.822 764.186 33.841 4.880 1.842 0.607 0.252 0.134 0.078

1988 29.623 680.173 146.993 182.312 9.991 2.810 0.814 0.458 0.036 0.112

1989 31.862 467.877 319.272 38.660 47.305 5.850 0.833 0.311 0.661 0.132

1990 27.000 185.344 146.071 79.339 26.351 5.469 0.758 0.189 0.383 0.239

1991 152.176 291.378 159.424 33.955 13.569 4.313 5.659 0.239 0.204 0.092

1992 26.814 360.890 174.526 29.253 5.961 3.748 2.871 1.186 0.346 0.050

1993 74.272 188.988 283.400 62.783 8.272 1.128 1.130 0.584 0.464 0.155

1994 284.479 193.260 77.139 34.458 10.586 2.667 0.600 0.800 0.895 0.373

1995 108.101 265.634 40.618 13.218 7.527 1.110 0.806 0.330 1.051 0.202

1996 222.510 310.287 206.883 21.469 4.470 3.134 0.838 0.044 0.161 0.122

1997 65.515 1046.845 59.241 17.180 2.670 0.257 0.358 0.157 0.111 0.000

1998 255.654 347.575 402.657 44.960 8.294 1.224 0.339 0.149 0.213 0.072

1999 257.559 293.253 121.551 171.254 3.391 1.956 0.127 0.130 0.027 0.030

2000 209.293 267.473 69.252 29.349 22.359 0.570 0.162 0.502 0.027 0.012

2001 807.932 206.531 72.236 17.840 9.174 8.716 0.270 0.131 0.038 0.040

2002 248.356 519.224 44.475 14.901 4.991 2.539 1.321 0.085 0.128 0.000

2003 225.619 132.754 159.120 10.057 5.550 1.426 1.133 0.638 0.111 0.096

2004 197.940 233.707 39.623 61.912 6.152 2.464 1.492 0.952 2.842 0.000

2005 270.775 163.046 66.176 6.759 12.790 1.084 1.164 0.290 0.152 0.492

2006 250.800 128.615 36.385 18.115 2.982 5.890 0.867 0.757 0.040 0.269

2007 298.086 311.997 67.169 19.707 14.416 2.942 6.085 0.684 0.831 0.156

2008 387.592 221.567 120.728 30.108 9.075 7.205 0.618 1.715 0.292 0.229

2009 555.472 408.995 105.222 45.975 13.013 4.029 3.474 0.574 2.128 0.278

2010 814.363 261.097 84.254 34.244 20.178 4.662 2.162 3.464 0.207 2.547

2011 323.428 486.157 148.217 55.305 20.065 12.904 3.945 2.243 2.263 0.232


Combined BTS index

YEAR 0 1 2 3 4 5 6 7 8 9

1996 102.136 143.896 99.623 13.280 4.266 3.035 1.653 0.676 0.442 0.214

1997 24.190 386.840 28.679 14.886 4.010 2.042 1.538 0.428 0.797 0.327

1998 96.333 131.191 177.631 25.463 7.266 2.500 1.355 0.955 0.808 0.323

1999 100.264 116.989 53.597 96.348 6.493 3.005 1.926 0.659 0.756 0.314

2000 81.459 108.393 38.887 22.880 23.680 3.017 1.725 1.113 0.797 0.219

2001 297.375 80.296 39.788 15.695 8.754 9.300 1.079 0.624 0.420 0.511

2002 87.786 217.276 26.709 14.029 7.616 4.794 4.643 0.754 0.765 0.385

2003 87.985 53.579 94.429 15.858 10.305 5.361 3.081 4.007 0.732 0.760

2004 80.357 101.411 30.306 51.218 11.212 4.961 2.885 1.538 3.402 0.391

2005 106.916 70.845 45.646 13.806 20.392 3.035 6.942 1.568 0.571 3.570

2006 97.992 54.855 42.922 29.187 11.748 12.052 2.106 3.938 0.844 0.767

2007 115.922 139.391 44.429 24.594 26.579 5.681 11.685 2.091 3.947 0.364

2008 143.963 98.909 89.736 33.838 20.735 20.605 6.330 13.054 2.727 6.718

2009 219.268 170.840 76.528 54.059 21.482 12.834 12.192 3.139 10.254 1.585

2010 326.437 144.792 69.544 47.943 40.349 17.914 6.845 15.841 3.179 8.306

2011 120.520 226.465 125.987 58.138 32.752 33.174 15.090 5.808 11.940 1.124


Annex 3: Residual for the BTS survey indices in in the sep option and the sep_ages option for exploring the effect of trimming the age ranges included

2000 2005 2010

BTS-Tridens residuals: sep

Year

Age

13

57

9

Mod>ObsObs>Mod

-3.97 1.83-1.37 1.24

-0.54 0.39-0.49 0.6

-0.5 0.46-0.49 0.56

-0.81 0.38-0.460.56

-0.530.82

max = 1.83; min = -3.97

2000 2005 2010

BTS-Tridens residuals: sep_ages

Year

Age

13

57

9

-1.36 1.29-0.52 0.39

-0.48 0.52-0.480.47

-0.420.51-0.79 0.32

-0.330.41-0.70.75

max = 1.29; min = -1.36


1985 1990 1995 2000 2005 2010


Year

Age

13

57

Mod>ObsObs>Mod

-1.22 0.58-0.841.23-0.830.95

-0.54 0.58-1.240.83

-0.99 1.07-1.93 0.98

-1.661.87

max = 1.87; min = -1.93

1985 1990 1995 2000 2005 2010

BTS-ISIS residuals: sep_ages

Year

Age

13

-1.22 0.58-0.851.23

-0.880.96

max = 1.23; min = -1.22


1990 1995 2000 2005 2010

0.0

0.2

0.4

0.6

0.8

1.0

sep

Year

Mea

n F

(age

s 2-

6)

1990 1995 2000 2005 2010

0.0

0.2

0.4

0.6

0.8

1.0

sep_ages

Year

Mea

n F

(age

s 2-

6)

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Annu

al M

ohn'

s rh

o

-0.4

-0.2

0.0

0.2

0.4

sepsep_ages

Underestimate

Overestimate


1990 1995 2000 2005 2010

0e+0

02e

+05

4e+0

5

sep

Year

Spaw

ner S

tock

Bio

mas

s

1990 1995 2000 2005 2010

0e+0

02e

+05

4e+0

5

sep_ages

Year

Spaw

ner S

tock

Bio

mas

s

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Annu

al M

ohn'

s rh

o

-0.4

-0.2

0.0

0.2

0.4

sepsep_ages

Underestimate

Overestimate


1990 1995 2000 2005 2010

050

0000

1500

000

sep

Year

Rec

ruitm

ent (

Age1

)

1990 1995 2000 2005 2010

050

0000

1500

000

sep_ages

Year

Rec

ruitm

ent (

Age1

)

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Annu

al M

ohn'

s rh

o

-0.6

-0.2

0.2

0.6

sepsep_ages

Underestimate

Overestimate


Annex 4: Relative numbers-at-age compared between the two options in which the BTS-indices are combined; with and with-out keeping the early Isis part of the time-series

1960 1970 1980 1990 2000 2010

0.0

0.5

1.0

1.5

combCor / sep

Year

Rel

ativ

e N

umbe

rs A

t Age

Age 1Age 2Age 3Age 4Age 5Age 6Age 7Age 8Age 9Age 10

1960 1970 1980 1990 2000 2010

0.0

0.5

1.0

1.5

combCor_hist / sep

YearR

elat

ive

Num

bers

At A

ge

Age 1Age 2Age 3Age 4Age 5Age 6Age 7Age 8Age 9Age 10


Annex 5: North Sea plaice XSA assessment outputs and diagnostics for index option sep (using separate BTS-Isis and BTS-Tridens indices)

Assessment summary

Table APPsep1. Assessment outputs from the current ICES XSA assessment of the North Sea plaice stock (index option ‘sep’).

recruits ssb catch landings discards fbar2-6 fbar hc2-6 fbar dis2-3 Y/ssb 1957 460521 285168 78443 70563 7880 0.28 0.23 0.12 0.25 1958 700348 293812 88191 73354 14837 0.33 0.25 0.19 0.25 1959 864884 299088 109164 79300 29864 0.37 0.24 0.24 0.27 1960 760715 308720 117334 87541 29793 0.37 0.27 0.23 0.28 1961 866068 321356 118474 85984 32490 0.35 0.24 0.27 0.27 1962 593500 371851 125375 87472 37903 0.38 0.24 0.29 0.24 1963 694668 373663 148376 107118 41258 0.42 0.27 0.35 0.29 1964 2254803 366000 147571 110540 37031 0.46 0.30 0.32 0.30 1965 701913 349390 140223 97143 43080 0.38 0.28 0.24 0.28 1966 594040 366356 166552 101834 64718 0.40 0.24 0.33 0.28 1967 407185 421538 163365 108819 54546 0.42 0.25 0.31 0.26 1968 438880 410962 139521 111534 27987 0.33 0.21 0.21 0.27 1969 658778 385429 142820 121651 21169 0.35 0.26 0.17 0.32 1970 664188 336818 159982 130342 29640 0.46 0.34 0.28 0.39 1971 420274 325084 136939 113944 22995 0.38 0.29 0.21 0.35 1972 374203 327465 142475 122843 19632 0.41 0.32 0.18 0.38 1973 1319873 276800 143783 130429 13354 0.47 0.41 0.13 0.47 1974 1135201 286019 157485 112540 44945 0.49 0.41 0.20 0.39 1975 863731 297149 195235 108536 86699 0.58 0.39 0.43 0.37 1976 690980 308599 166917 113670 53247 0.42 0.30 0.27 0.37 1977 988683 312634 176689 119188 57501 0.51 0.34 0.31 0.38 1978 917356 298716 159639 113984 45655 0.46 0.35 0.22 0.38 1979 898137 296875 213282 145347 67935 0.67 0.49 0.36 0.49 1980 1133823 272614 171485 140405 31080 0.56 0.50 0.15 0.52 1981 864375 262727 173596 140565 33031 0.55 0.48 0.16 0.54 1982 2019231 261090 204508 155381 49127 0.61 0.52 0.22 0.60 1983 1302743 310503 219386 144903 74483 0.60 0.48 0.27 0.47 1984 1254969 318155 227848 157032 70816 0.59 0.43 0.28 0.49 1985 1845858 340658 221419 160870 60549 0.53 0.43 0.23 0.47 1986 4753000 368095 296472 166519 129953 0.66 0.50 0.34 0.45 1987 1958536 445340 345628 155104 190524 0.70 0.49 0.51 0.35 1988 1763230 386904 312684 156261 156423 0.67 0.40 0.51 0.40 1989 1179868 410985 279112 171319 107793 0.61 0.38 0.46 0.42 1990 1030954 374680 229016 157791 71225 0.57 0.39 0.40 0.42 1991 910217 344625 230278 149343 80935 0.64 0.41 0.48 0.43 1992 775019 282064 183326 126277 57049 0.63 0.42 0.41 0.45 1993 528748 245464 153043 118027 35016 0.64 0.50 0.28 0.48 1994 442733 224637 135227 111442 23785 0.61 0.51 0.24 0.50 1995 1162490 218074 121063 99235 21828 0.65 0.56 0.21 0.46 1996 1291971 180057 134647 82598 52049 0.67 0.52 0.35 0.46 1997 2144115 205817 184297 84152 100145 0.80 0.52 0.69 0.41 1998 773058 225637 176282 72531 103751 0.74 0.39 0.60 0.32 1999 840965 200712 152696 81720 70976 0.67 0.38 0.39 0.41 2000 982008 226943 126783 82472 44311 0.47 0.32 0.26 0.36 2001 541134 268314 183182 82873 100309 0.78 0.32 0.72 0.31 2002 1722612 195812 125777 71387 54390 0.58 0.38 0.42 0.36 2003 527319 223856 144964 67172 77792 0.61 0.38 0.45 0.30 2004 1266619 204624 116536 62070 54466 0.48 0.29 0.44 0.30 2005 768477 241169 110133 56257 53876 0.41 0.20 0.38 0.23 2006 934392 250554 120299 58453 61846 0.38 0.20 0.39 0.23 2007 1162662 256931 89783 50348 39435 0.32 0.16 0.35 0.20 2008 1008347 355614 95309 49434 45875 0.24 0.14 0.22 0.14 2009 1034612 396133 100671 55446 45225 0.22 0.12 0.21 0.14 2010 924410 493431 106980 61163 45817 0.21 0.11 0.20 0.12 2011 1258796 468861 108523 67963 40560 0.23 0.11 0.20 0.14


FLR XSA Diagnostics 2013-03-15 00:14:27

cpue data from xsa.indices Catch data for 55 years. 1957 to 2011. Ages 1 to 10. fleet first age last age first year last year alpha beta 1 BTS-Isis 1 8 1985 2011 0.66 0.75 2 BTS-Tridens 1 9 1996 2011 0.66 0.75 3 SNS 1 3 1982 2011 0.66 0.75 Time-series weights : Tapered time weighting not applied Catchability analysis : Catchability independent of size for all ages Catchability independent of age for ages >= 6 Terminal population estimation : Survivor estimates shrunk towards the mean F of the final 5 years or the 5 oldest ages. S.E. of the mean to which the estimates are shrunk = 2 Minimum standard error for population estimates derived from each fleet = 0.3 prior weighting not applied Regression weights year age 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 all 1 1 1 1 1 1 1 1 1 1 Fishing mortalities year age 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 1 0.210 0.147 0.215 0.142 0.286 0.074 0.153 0.166 0.212 0.103 2 0.594 0.627 0.664 0.491 0.552 0.460 0.332 0.319 0.301 0.287 3 0.530 0.621 0.472 0.479 0.450 0.426 0.256 0.220 0.224 0.224 4 0.632 0.491 0.502 0.367 0.420 0.222 0.277 0.204 0.204 0.298 5 0.681 0.686 0.252 0.449 0.252 0.300 0.166 0.220 0.148 0.201 6 0.464 0.618 0.526 0.257 0.234 0.182 0.161 0.125 0.166 0.121 7 0.528 0.489 0.331 0.447 0.130 0.118 0.155 0.110 0.097 0.100 8 0.264 0.302 0.151 0.266 0.308 0.097 0.098 0.091 0.084 0.092 9 0.171 0.131 0.107 0.091 0.182 0.099 0.022 0.038 0.035 0.058 10 0.171 0.131 0.107 0.091 0.182 0.099 0.022 0.038 0.035 0.058 XSA population number (Thousand) age year 1 2 3 4 5 6 7 8 9 10 2002 1722612 456526 343259 134427 80304 107691 12084 8014 4916 10528 2003 527319 1262822 228057 182873 64629 36782 61289 6449 5568 7334 2004 1266619 412076 610177 110892 101278 29447 17933 33998 4315 5651 2005 768477 923997 191894 344488 60722 71201 15752 11649 26441 7387 2006 934392 603590 511493 107506 215949 35058 49820 9119 8082 8560 2007 1162662 635383 314601 295252 63895 151946 25116 39569 6064 14157 2008 1008347 977340 362918 185831 213949 42823 114577 20206 32482 43235 2009 1034612 783293 634500 254099 127517 164055 32981 88799 16578 22938 2010 924410 793288 515163 460634 187445 92618 130943 26723 73347 37716 2011 1258796 676700 531241 372680 339757 146316 70983 107575 22230 50599


Estimated population abundance at 1st Jan 2012 age year 1 2 3 4 5 6 7 8 9 10 2012 0 1027230 459473 384149 250378 251361 117300 58138 88772 18991

Fleet: BTS-Isis Log-catchability residuals. year age 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 1 -1.222 -0.568 -0.813 0.407 0.420 -0.407 0.225 0.582 0.322 0.481 -0.195 -0.164 0.524 0.504 0.264 -0.021 0.281 0.144 -0.081 -0.343 -0.255 -0.587 -0.068 -0.213 0.384 0.080 0.317 2 0.336 -0.271 0.586 -0.265 0.613 0.136 0.401 0.657 1.228 0.314 -0.224 0.457 -0.750 0.420 0.317 -0.388 -0.298 -0.334 -0.053 -0.297 -0.714 -0.844 -0.346 -0.281 -0.206 -0.454 0.260 3 -0.052 0.401 -0.249 0.530 -0.278 0.522 0.004 0.069 0.946 0.411 -0.129 0.478 -0.492 0.633 0.800 -0.007 -0.216 -0.567 -0.487 0.241 -0.811 -0.827 -0.273 -0.112 -0.273 -0.356 0.093 4 -0.291 -0.141 -0.541 -0.107 0.495 0.579 0.098 -0.381 0.138 0.525 0.294 0.179 -0.188 0.505 -0.142 0.014 0.244 -0.069 -0.370 0.241 -0.256 -0.510 -0.084 -0.045 -0.049 -0.205 0.067 5 -0.562 0.018 -0.353 0.291 0.675 -0.342 -0.001 0.238 -0.658 0.304 -0.331 0.833 -1.237 0.367 0.507 -0.517 0.485 0.367 0.011 -0.196 -0.367 -0.083 0.475 0.067 0.042 -0.248 0.213 6 0.298 -0.632 -0.701 -0.020 0.169 -0.319 0.816 0.546 0.210 -0.175 0.161 0.515 -0.165 0.048 -0.910 -0.990 -0.347 -0.390 0.640 1.072 -0.248 0.149 0.595 -0.441 -0.082 0.044 0.156 7 0.065 0.102 -0.224 -0.232 -0.261 -0.667 -0.736 -0.063 -0.584 0.797 -0.019 -1.930 -0.440 -0.341 -0.435 0.920 -0.654 -0.901 -0.536 0.982 0.004 -0.411 0.164 -0.408 -0.289 0.120 0.300 8 -0.112 -0.052 -0.412 -1.149 0.861 0.541 0.105 0.402 -0.490 0.186 1.870 0.012 -0.271 0.454 -1.467 -1.660 -1.336 -0.266 -0.165 1.309 -0.468 -1.528 -0.110 -0.484 0.017 -1.117 -0.112 Mean log-catchability and standard error of ages with catchability independent of year-class strength and constant w.r.t. time 1 2 3 4 5 6 7 8 Mean_Logq -8.0324 -8.4137 -9.0342 -9.6163 -10.1790 -10.5215 -10.5215 -10.5215 S.E_Logq 0.4519 0.4988 0.4705 0.3107 0.4626 0.5104 0.5907 0.8373 Fleet: BTS-Tridens Log-catchability residuals. year age 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 1 -1.440 -3.974 -2.860 -0.461 -0.077 -0.317 0.777 0.376 -0.043 0.543 0.427 1.433 1.212 0.829 1.745 1.830 2 -1.373 -1.163 -0.399 -0.734 -0.321 -0.718 0.136 0.314 0.489 -0.048 0.740 0.320 0.553 0.574 0.394 1.237 3 -0.513 -0.542 -0.332 -0.199 -0.246 -0.222 -0.430 0.381 0.148 0.364 0.106 0.302 0.308 0.251 0.235 0.390 4 -0.478 -0.217 -0.253 0.127 -0.489 -0.299 -0.044 -0.106 0.443 -0.158 0.603 0.165 0.435 0.111 -0.031 0.191 5 -0.416 0.046 0.135 -0.068 0.284 -0.503 0.047 0.462 -0.455 -0.229 -0.285 0.166 0.216 0.318 -0.011 0.292 6 -0.172 -0.029 0.060 0.498 0.058 -0.296 -0.494 0.235 0.251 0.215 -0.335 -0.206 0.560 -0.174 -0.300 0.131 7 -0.446 -0.814 0.177 -0.164 0.228 -0.458 -0.045 -0.033 -0.064 0.378 -0.084 -0.089 0.285 0.079 0.070 -0.115 8 -0.353 0.415 0.416 0.559 0.423 -0.251 0.177 0.368 -0.131 -0.463 0.233 0.094 0.412 0.264 0.155 0.234 9 -0.277 0.245 0.073 0.068 -0.241 0.223 -0.032 0.446 0.091 0.435 0.086 -0.491 0.816 0.040 -0.048 -0.530 Mean log-catchability and standard error of ages with catchability independent of year-class strength and constant w.r.t. time 1 2 3 4 5 6 7 8 9 Mean_Logq -11.9973 -10.1199 -9.6165 -9.3906 -9.3597 -9.2093 -9.2093 -9.2093 -9.2093 S.E_Logq 1.5957 0.7250 0.3434 0.3212 0.2999 0.3041 0.3030 0.3053 0.3465


Fleet: SNS Log-catchability residuals. year age 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 1 0.434 0.146 0.509 -0.375 -0.102 -0.294 -0.085 0.236 0.022 1.027 0.969 0.606 0.602 -0.138 -0.335 0.564 0.600 0.508 -0.050 -0.176 -0.250 NA -0.466 -0.605 -0.515 -0.730 -0.495 -0.501 -0.574 2 0.690 0.384 0.549 0.879 -0.053 0.517 0.487 0.800 0.240 0.823 1.197 0.925 0.647 0.173 0.528 -0.260 0.904 0.920 -0.638 -0.440 -0.957 NA -0.603 -1.234 -0.912 -0.722 -1.007 -1.086 -1.220 3 0.366 -1.098 0.424 0.386 0.182 -0.011 1.456 1.098 0.854 0.451 1.067 0.368 0.271 -0.120 1.137 -0.479 1.994 1.836 0.053 -0.536 -1.041 NA -0.251 -1.119 -1.123 -1.665 -0.729 -1.255 -0.971 year age 2011 1 -0.530 2 -1.531 3 -1.543 Mean log-catchability and standard error of ages with catchability independent of year-class strength and constant w.r.t. time 1 2 3 Mean_Logq -3.5546 -4.5796 -5.6816 S.E_Logq 0.5052 0.8261 1.0061

year

age

2

4

6

8

1985 1990 1995 2000 2005 2010

BTS-Isis BTS-Tridens

2

4

6

8

SNS

Scale

4

3

2

1

0

-1

-2

-3

-4

Figure APPsep1. Log-catchability residuals for the three indices used in the current ICES XSA assessment of the North Sea plaice stock (index option ‘sep’): BTS-Tridens, BTS-Isis and SNS.


Terminal year survivor and F summaries

Age = 1 . Catchability constant w.r.t. time and dependent on age Year class = 2010 Fleet = BTS-Isis 1 Survivors 1409781.000 Raw weights 4.258 Fleet = BTS-Tridens 1 Survivors 6401923.000 Raw weights 0.333 Fleet = fshk 1 Survivors 573176.00 Raw weights 0.25 Fleet = SNS 1 Survivors 604384.000 Raw weights 3.416 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "1409781" "0.46" "Inf" "Inf" "1" "0.516" "0.076" [2,] "BTS-Tridens" "6401923" "1.645" "Inf" "Inf" "1" "0.04" "0.017" [3,] "fshk" "573176" "1.899" "Inf" "Inf" "1" "0.03" "0.178" [4,] "SNS" "604384" "0.514" "Inf" "Inf" "1" "0.414" "0.17" Weighted prediction: Survivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "1027230" "" "" "" "0.103" Age = 2 . Catchability constant w.r.t. time and dependent on age Year class = 2009 Fleet = BTS-Isis 2 1 Survivors 596082.000 497852.000 Raw weights 2.909 2.866 Fleet = BTS-Tridens 2 1 Survivors 1583298.000 2631399.000 Raw weights 1.344 0.224 Fleet = fshk 2 Survivors 316596.00 Raw weights 0.25 Fleet = SNS 2 1 Survivors 99383.000 258752.0 Raw weights 1.063 2.3


Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "545118" "0.343" "0.09" "0.262" "2" "0.527" "0.247" [2,] "BTS-Tridens" "1702695" "0.682" "0.178" "0.261" "2" "0.143" "0.086" [3,] "fshk" "316596" "1.732" "Inf" "Inf" "1" "0.023" "0.394" [4,] "SNS" "191217" "0.44" "0.445" "1.011" "2" "0.307" "0.587" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "459473" "" "" "" "0.287" Age = 3 . Catchability constant w.r.t. time and dependent on age Year class = 2008 Fleet = BTS-Isis 3 2 1 Survivors 421428.000 244014.000 563837.000 Raw weights 3.482 2.292 2.366 Fleet = BTS-Tridens 3 2 1 Survivors 567380.000 569730.000 880031.000 Raw weights 6.379 1.059 0.185 Fleet = fshk 3 Survivors 259723.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 82126.000 113459.000 232691.000 Raw weights 0.763 0.838 1.898 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "393225" "0.283" "0.228" "0.806" "3" "0.417" "0.22" [2,] "BTS-Tridens" "573794" "0.316" "0.048" "0.152" "3" "0.391" "0.156" [3,] "fshk" "259723" "1.788" "Inf" "Inf" "1" "0.013" "0.316" [4,] "SNS" "156120" "0.41" "0.317" "0.773" "3" "0.179" "0.481" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "384149" "" "" "" "0.224" Age = 4 . Catchability constant w.r.t. time and dependent on age Year class = 2007 Fleet = BTS-Isis 4 3 2 1 Survivors 267766.000 175322.000 203731.000 202401.000 Raw weights 7.416 2.586 1.672 1.749 Fleet = BTS-Tridens 4 3 2 1 Survivors 303082.000 316763.000 444541.000 841092.000 Raw weights 6.773 4.738 0.772 0.137


Fleet = fshk 4 Survivors 284851.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 94772.000 84517.000 152590.000 Raw weights 0.567 0.611 1.403 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "229984" "0.216" "0.102" "0.472" "4" "0.468" "0.32" [2,] "BTS-Tridens" "319232" "0.231" "0.079" "0.342" "4" "0.433" "0.241" [3,] "fshk" "284851" "1.723" "Inf" "Inf" "1" "0.009" "0.266" [4,] "SNS" "119494" "0.41" "0.191" "0.466" "3" "0.09" "0.546" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "250378" "" "" "" "0.298" Age = 5 . Catchability constant w.r.t. time and dependent on age Year class = 2006 Fleet = BTS-Isis 5 4 3 2 1 Survivors 311015.000 204775.000 191385.00 189805.000 234722.000 Raw weights 3.684 6.657 2.33 1.487 1.683 Fleet = BTS-Tridens 5 4 3 2 1 Survivors 336420.000 243693.00 322909.000 436965.000 1053731.000 Raw weights 8.556 6.08 4.268 0.687 0.132 Fleet = fshk 5 Survivors 230956.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 71668.000 91774.000 121130.00 Raw weights 0.511 0.543 1.35 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "225099" "0.198" "0.093" "0.47" "5" "0.414" "0.222" [2,] "BTS-Tridens" "306993" "0.187" "0.094" "0.503" "5" "0.516" "0.168" [3,] "fshk" "230956" "1.808" "Inf" "Inf" "1" "0.007" "0.217" [4,] "SNS" "101764" "0.408" "0.151" "0.37" "3" "0.063" "0.439" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "251361" "" "" "" "0.201" Age = 6 . Catchability constant w.r.t. time and dependent on age Year class = 2005


Fleet = BTS-Isis 6 5 4 3 2 1 Survivors 137133.000 91513.000 111706.000 104898.000 82958.00 65240.00 Raw weights 3.279 3.444 6.224 2.101 1.18 1.08 Fleet = BTS-Tridens 6 5 4 3 2 1 Survivors 133673.000 116032.000 131010.000 159576.000 161459.000 179818.000 Raw weights 9.017 7.998 5.684 3.848 0.545 0.085 Fleet = fshk 6 Survivors 79472.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 56569.00 56982.000 70085.000 Raw weights 0.46 0.431 0.866 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "104960" "0.192" "0.086" "0.448" "6" "0.372" "0.134" [2,] "BTS-Tridens" "131542" "0.163" "0.047" "0.288" "6" "0.585" "0.109" [3,] "fshk" "79472" "1.882" "Inf" "Inf" "1" "0.005" "0.174" [4,] "SNS" "62981" "0.422" "0.075" "0.178" "3" "0.038" "0.215" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "117300" "" "" "" "0.121" Age = 7 . Catchability constant w.r.t. time and dependent on age Year class = 2004 Fleet = BTS-Isis 7 6 5 4 3 2 1 Survivors 78463.000 60737.000 60621.000 55556.000 44253.000 25007.00 45026.000 Raw weights 2.211 2.837 2.773 4.661 1.327 0.68 0.719 Fleet = BTS-Tridens 7 6 5 4 3 2 1 Survivors 51806.000 43071.000 79921.00 89844.000 78593.000 121865.000 100093.000 Raw weights 8.801 7.802 6.44 4.257 2.432 0.314 0.056 Fleet = fshk 7 Survivors 46937.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 11001.000 23344.000 31737.000 Raw weights 0.291 0.249 0.577


Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "56521" "0.194" "0.096" "0.495" "7" "0.326" "0.102" [2,] "BTS-Tridens" "61196" "0.151" "0.12" "0.795" "7" "0.645" "0.095" [3,] "fshk" "46937" "1.903" "Inf" "Inf" "1" "0.005" "0.122" [4,] "SNS" "22489" "0.42" "0.312" "0.743" "3" "0.024" "0.24" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "58138" "" "" "" "0.1" Age = 8 . Catchability constant w.r.t. time and dependent on age Year class = 2003 Fleet = BTS-Isis 8 7 6 5 4 3 2 1 Survivors 79338.000 100078.000 81755.000 94957.000 81618.000 38828.000 43482.00 62985.000 Raw weights 1.183 2.022 2.703 2.789 4.951 1.378 0.75 0.736 Fleet = BTS-Tridens 8 7 6 5 4 3 2 1 Survivors 112137.000 95153.000 74557.000 110212.000 104699.000 98716.000 84562.000 85025.000 Raw weights 7.129 8.051 7.433 6.477 4.522 2.524 0.346 0.058 Fleet = fshk 8 Survivors 58902.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 28876.000 25840.000 55720.000 Raw weights 0.302 0.274 0.591 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "77378" "0.188" "0.102" "0.543" "8" "0.303" "0.105" [2,] "BTS-Tridens" "97225" "0.139" "0.056" "0.403" "8" "0.671" "0.085" [3,] "fshk" "58902" "1.91" "Inf" "Inf" "1" "0.005" "0.136" [4,] "SNS" "39242" "0.42" "0.253" "0.602" "3" "0.021" "0.198" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "88772" "" "" "" "0.092" Age = 9 . Catchability constant w.r.t. time and dependent on age Year class = 2002 Fleet = BTS-Isis 8 7 6 5 4 3 2 1 Survivors 6214.000 14223.000 12223.000 30540.000 11406.000 8437.000 14107.000 17514.000 Raw weights 1.126 1.898 2.448 2.208 3.215 0.868 0.398 0.418 Fleet = BTS-Tridens


9 8 7 6 5 4 3 2 1 Survivors 11175.000 22176.000 20553.000 33242.000 22418.000 34695.000 27338.000 30951.000 27644.000 Raw weights 7.194 6.785 7.557 6.733 5.128 2.936 1.591 0.184 0.033 Fleet = fshk 9 Survivors 6369.00 Raw weights 0.25 Fleet = SNS 3 2 Survivors 6202.00 10391.000 Raw weights 0.19 0.145 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-Isis" "13458" "0.202" "0.168" "0.832" "8" "0.245" "0.08" [2,] "BTS-Tridens" "21598" "0.136" "0.13" "0.956" "9" "0.743" "0.051" [3,] "fshk" "6369" "1.943" "Inf" "Inf" "1" "0.005" "0.163" [4,] "SNS" "7755" "0.685" "0.256" "0.374" "2" "0.007" "0.136" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "18991" "" "" "" "0.058"


Figure APPsep2. Catchability at age of the three indices used in the current ICES XSA assessment of the North Sea plaice stock (index option ‘sep’): BTS-Tridens, BTS-Isis and SNS.


Annex 6: North Sea plaice XSA assessment outputs and diagnostics for index option combCor_hist (using the combined BTS index)

Assessment summary

Table APPsep1. Assessment outputs from the XSA assessment of the North Sea plaice using the combined BTS index (index option ‘combCor_hist’).

recruits ssb catch landings discards fbar2-6 fbar hc2-6 fbar dis2-3 Y/ssb 1957 460521 285168 78443 70563 7880 0.28 0.23 0.12 0.25 1958 700348 293812 88191 73354 14837 0.33 0.25 0.19 0.25 1959 864885 299089 109164 79300 29864 0.37 0.24 0.24 0.27 1960 760715 308720 117334 87541 29793 0.37 0.27 0.23 0.28 1961 866068 321356 118474 85984 32490 0.35 0.24 0.27 0.27 1962 593501 371851 125375 87472 37903 0.38 0.24 0.29 0.24 1963 694669 373664 148376 107118 41258 0.42 0.27 0.35 0.29 1964 2254808 366001 147571 110540 37031 0.46 0.30 0.32 0.30 1965 701915 349390 140223 97143 43080 0.38 0.28 0.24 0.28 1966 594043 366357 166552 101834 64718 0.40 0.24 0.33 0.28 1967 407188 421540 163365 108819 54546 0.42 0.25 0.31 0.26 1968 438884 410964 139521 111534 27987 0.33 0.21 0.21 0.27 1969 658787 385433 142820 121651 21169 0.35 0.26 0.17 0.32 1970 664197 336823 159982 130342 29640 0.46 0.34 0.28 0.39 1971 420286 325091 136939 113944 22995 0.38 0.29 0.21 0.35 1972 374222 327476 142475 122843 19632 0.41 0.32 0.18 0.38 1973 1319975 276814 143783 130429 13354 0.47 0.41 0.13 0.47 1974 1135327 286043 157485 112540 44945 0.49 0.41 0.20 0.39 1975 863855 297190 195235 108536 86699 0.58 0.39 0.43 0.37 1976 691176 308669 166917 113670 53247 0.42 0.30 0.27 0.37 1977 988966 312745 176689 119188 57501 0.51 0.34 0.31 0.38 1978 918590 298883 159639 113984 45655 0.46 0.35 0.22 0.38 1979 898255 297187 213282 145347 67935 0.67 0.49 0.36 0.49 1980 1134708 273061 171485 140405 31080 0.56 0.49 0.15 0.51 1981 868049 263500 173596 140565 33031 0.54 0.48 0.16 0.53 1982 2019588 262189 204508 155381 49127 0.61 0.52 0.22 0.59 1983 1301830 311944 219386 144903 74483 0.59 0.48 0.27 0.46 1984 1257651 320397 227848 157032 70816 0.58 0.43 0.28 0.49 1985 1847164 343047 221419 160870 60549 0.52 0.43 0.23 0.47 1986 4722689 371935 296472 166519 129953 0.65 0.49 0.34 0.45 1987 1913904 446271 345628 155104 190524 0.70 0.48 0.51 0.35 1988 1762584 385898 312684 156261 156423 0.68 0.40 0.53 0.40 1989 1179736 407935 279112 171319 107793 0.62 0.38 0.47 0.42 1990 1032297 360169 229016 157791 71225 0.60 0.41 0.40 0.44 1991 911030 327272 230278 149343 80935 0.71 0.47 0.48 0.46 1992 776145 266189 183326 126277 57049 0.69 0.47 0.40 0.47 1993 531093 226026 153043 118027 35016 0.64 0.50 0.28 0.52 1994 443369 193486 135227 111442 23785 0.61 0.51 0.24 0.58 1995 1164758 179431 121063 99235 21828 0.64 0.55 0.21 0.55 1996 1293443 181088 134647 82598 52049 0.66 0.51 0.34 0.46 1997 2151531 207765 184297 84152 100145 0.78 0.51 0.68 0.41 1998 772423 229503 176282 72531 103751 0.72 0.38 0.60 0.32 1999 842259 204292 152696 81720 70976 0.66 0.37 0.39 0.40 2000 991931 231463 126783 82472 44311 0.46 0.31 0.26 0.36 2001 544146 275820 183182 82873 100309 0.76 0.31 0.71 0.30 2002 1742334 200583 125777 71387 54390 0.57 0.37 0.41 0.36 2003 536428 231367 144964 67172 77792 0.60 0.37 0.44 0.29 2004 1283387 213227 116536 62070 54466 0.47 0.28 0.43 0.29 2005 776702 252693 110133 56257 53876 0.39 0.19 0.37 0.22 2006 949881 262257 120299 58453 61846 0.36 0.19 0.38 0.22 2007 1193449 271938 89783 50348 39435 0.30 0.15 0.34 0.19 2008 1021125 379649 95309 49434 45875 0.23 0.13 0.21 0.13 2009 1094944 419027 100671 55446 45225 0.21 0.11 0.21 0.13 2010 1170126 525715 106980 61163 45817 0.20 0.10 0.19 0.12 2011 1476620 511240 108523 67963 40560 0.20 0.10 0.16 0.13


FLR XSA Diagnostics 2013-03-15 00:17:02

cpue data from xsa.indices Catch data for 55 years. 1957 to 2011. Ages 1 to 10. fleet first age last age first year last year alpha beta 1 BTS-Isis_old 1 8 1985 1995 0.66 0.75 2 BTS-comb 1 9 1996 2011 0.66 0.75 3 SNS 1 3 1982 2011 0.66 0.75 Time-series weights : Tapered time weighting not applied Catchability analysis : Catchability independent of size for all ages Catchability independent of age for ages >= 6 Terminal population estimation : Survivor estimates shrunk towards the mean F of the final 5 years or the 5 oldest ages. S.E. of the mean to which the estimates are shrunk = 2 Minimum standard error for population estimates derived from each fleet = 0.3 prior weighting not applied Regression weights year age 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 all 1 1 1 1 1 1 1 1 1 1 Fishing mortalities year age 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 1 0.208 0.143 0.212 0.140 0.280 0.072 0.150 0.156 0.163 0.087 2 0.589 0.615 0.641 0.481 0.543 0.447 0.321 0.313 0.278 0.208 3 0.517 0.612 0.456 0.451 0.434 0.415 0.246 0.211 0.219 0.203 4 0.622 0.472 0.489 0.349 0.383 0.212 0.266 0.194 0.193 0.289 5 0.687 0.666 0.239 0.431 0.235 0.263 0.156 0.210 0.139 0.188 6 0.444 0.630 0.498 0.239 0.220 0.168 0.137 0.117 0.157 0.113 7 0.482 0.457 0.341 0.409 0.120 0.110 0.141 0.092 0.090 0.093 8 0.232 0.264 0.138 0.276 0.272 0.088 0.091 0.082 0.069 0.085 9 0.158 0.112 0.091 0.082 0.191 0.085 0.020 0.035 0.032 0.047 10 0.158 0.112 0.091 0.082 0.191 0.085 0.020 0.035 0.032 0.047 XSA population number (Thousand) age year 1 2 3 4 5 6 7 8 9 10 2002 1742109 459342 349527 135959 79770 111396 12955 8988 5279 11309 2003 538766 1280464 230605 188544 66016 36299 64641 7237 6449 8496 2004 1284072 422434 626140 113198 106409 30701 17496 37031 5028 6585 2005 776772 939789 201266 358932 62809 75844 16887 11253 29186 8155 2006 950574 611095 525782 115987 229018 36946 54021 10146 7724 8180 2007 1194174 650025 321392 308181 71569 163772 26824 43370 6993 16328 2008 1021603 1005853 376167 191976 225648 49766 125277 21752 35922 47815 2009 1095321 795287 660300 266087 133077 174641 39263 98481 17977 24873


2010 1171087 848220 526016 483978 198293 97648 140522 32407 82107 42222 2011 1477812 899903 580945 382500 360880 156131 75535 116242 27373 62313 Estimated population abundance at 1st Jan 2012 age year 1 2 3 4 5 6 7 8 9 10 2012 0 1225416 661446 429122 259264 270479 126182 62258 96616 23649 Fleet: BTS-Isis_old Log-catchability residuals. year age 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1 -1.151 -0.495 -0.733 0.476 0.490 -0.338 0.294 0.649 0.387 0.549 -0.127 2 0.019 -0.589 0.271 -0.566 0.291 -0.183 0.078 0.336 0.904 -0.014 -0.547 3 -0.247 0.206 -0.444 0.339 -0.445 0.319 -0.195 -0.137 0.743 0.205 -0.343 4 -0.353 -0.200 -0.599 -0.165 0.445 0.570 0.025 -0.446 0.060 0.450 0.213 5 -0.509 0.080 -0.283 0.360 0.744 -0.258 0.155 0.276 -0.602 0.334 -0.296 6 0.250 -0.692 -0.743 -0.044 0.140 -0.350 0.820 0.663 0.121 -0.228 0.062 7 0.022 0.069 -0.277 -0.251 -0.249 -0.670 -0.735 -0.004 -0.378 0.678 -0.060 8 -0.143 -0.075 -0.424 -1.180 0.883 0.603 0.159 0.472 -0.364 0.478 1.702 Mean log-catchability and standard error of ages with catchability independent of year-class strength and constant w.r.t. time 1 2 3 4 5 6 7 8 Mean_Logq -8.1022 -8.0940 -8.8342 -9.5486 -10.2285 -10.4561 -10.4561 -10.4561 S.E_Logq 0.6052 0.4577 0.3852 0.3905 0.4191 0.4918 0.3859 0.7644


Fleet: BTS-comb Log-catchability residuals. year age 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 1 -0.088 0.372 0.377 0.188 -0.087 0.176 0.106 -0.166 -0.348 -0.255 -0.614 -0.057 -0.188 0.292 0.066 0.227 2 0.431 -0.771 0.304 0.208 -0.261 -0.202 -0.146 0.110 0.101 -0.402 0.010 -0.084 0.093 0.164 -0.021 0.464 3 0.149 -0.485 0.214 0.374 -0.097 -0.197 -0.497 0.107 0.172 -0.008 -0.232 0.076 0.119 -0.001 0.112 0.195 4 -0.288 -0.185 -0.038 0.096 -0.333 -0.188 -0.055 -0.186 0.421 -0.233 0.368 0.087 0.351 0.008 0.040 0.134 5 -0.128 -0.127 0.159 -0.004 0.210 -0.378 0.115 0.401 -0.455 -0.284 -0.336 0.094 0.159 0.251 0.136 0.188 6 -0.141 -0.075 0.021 0.462 0.000 -0.331 -0.488 0.354 0.363 0.154 -0.333 -0.145 0.411 -0.202 -0.171 0.120 7 -0.576 -0.795 0.104 -0.267 0.354 -0.504 -0.127 -0.082 0.186 0.289 -0.158 -0.098 0.215 -0.085 0.257 -0.123 8 -0.238 0.267 0.399 0.395 0.233 -0.309 0.077 0.272 0.087 -0.409 0.082 0.042 0.364 0.172 0.103 0.161 9 -0.367 0.265 -0.070 -0.036 -0.454 0.025 -0.130 0.318 -0.113 0.334 0.202 -0.519 0.714 -0.027 0.108 -0.783 Mean log-catchability and standard error of ages with catchability independent of year-class strength and constant w.r.t. time 1 2 3 4 5 6 7 8 9 Mean_Logq -8.8781 -9.1212 -9.1907 -9.2256 -9.2795 -9.2141 -9.2141 -9.2141 -9.2141 S.E_Logq 0.2761 0.3151 0.2455 0.2350 0.2543 0.2891 0.3278 0.2414 0.3731 Fleet: SNS Log-catchability residuals. year age 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 1 0.457 0.170 0.532 -0.352 -0.078 -0.262 -0.065 0.258 0.042 1.048 0.988 0.622 0.621 -0.118 -0.315 0.583 0.622 0.527 -0.037 -0.160 -0.242 NA -0.460 -0.595 -0.514 -0.737 -0.488 -0.544 -0.823 2 0.712 0.407 0.574 0.903 -0.029 0.543 0.527 0.820 0.262 0.841 1.216 0.943 0.660 0.192 0.546 -0.241 0.921 0.943 -0.621 -0.434 -0.945 NA -0.622 -1.237 -0.910 -0.733 -1.023 -1.084 -1.281 3 0.383 -1.083 0.441 0.406 0.201 0.008 1.479 1.146 0.865 0.467 1.076 0.379 0.279 -0.120 1.145 -0.472 2.002 1.843 0.070 -0.531 -1.054 NA -0.273 -1.172 -1.147 -1.680 -0.758 -1.287 -0.982 year age 2011 1 -0.680 2 -1.851 3 -1.633 Mean log-catchability and standard error of ages with catchability independent of year-class strength and constant w.r.t. time 1 2 3 Mean_Logq -3.5762 -4.6009 -5.696 S.E_Logq 0.5309 0.8621 1.024


year

age

2

4

6

8

1985 1990 1995 2000 2005 2010

BTS-comb

1985 1990 1995 2000 2005 2010

SNS

Scale

3.00

2.25

1.50

0.75

0.00

-0.75

-1.50

-2.25

-3.00

Figure APPcombCor1. Log-catchability residuals for the three indices used in the combined BTS XSA assessment of the North Sea plaice stock (index option ‘combCor_hist): BTS-comb, BTS-Isis_old and SNS.

Terminal year survivor and F summaries

Age = 1 . Catchability constant w.r.t. time and dependent on age Year class = 2010 Fleet = BTS-comb 1 Survivors 1537079.000 Raw weights 10.182 Fleet = fshk 1 Survivors 625238.00 Raw weights 0.25 Fleet = SNS 1 Survivors 620515.000 Raw weights 3.143 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F


[1,] "BTS-comb" "1537079" "0.3" "Inf" "Inf" "1" "0.75" "0.07" [2,] "fshk" "625238" "1.915" "Inf" "Inf" "1" "0.018" "0.165" [3,] "SNS" "620515" "0.54" "Inf" "Inf" "1" "0.232" "0.166" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "1225416" "" "" "" "0.087" Age = 2 . Catchability constant w.r.t. time and dependent on age Year class = 2009 Fleet = BTS-comb 2 1 Survivors 1052301.000 706263.000 Raw weights 7.701 7.665 Fleet = fshk 2 Survivors 328972.00 Raw weights 0.25 Fleet = SNS 2 1 Survivors 103932.000 290352.000 Raw weights 1.056 2.366 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "862495" "0.221" "0.199" "0.9" "2" "0.807" "0.163" [2,] "fshk" "328972" "1.803" "Inf" "Inf" "1" "0.013" "0.382" [3,] "SNS" "211450" "0.461" "0.475" "1.03" "2" "0.18" "0.544" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "661446" "" "" "" "0.208" Age = 3 . Catchability constant w.r.t. time and dependent on age Year class = 2008 Fleet = BTS-comb 3 2 1 Survivors 521329.00 420193.000 574877.000 Raw weights 9.07 5.858 5.876 Fleet = fshk 3 Survivors 269969.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 83843.000 119189.000 249143.000 Raw weights 0.753 0.804 1.814 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "504349" "0.181" "0.086" "0.475" "3" "0.852" "0.175"


[2,] "fshk" "269969" "1.807" "Inf" "Inf" "1" "0.01" "0.306" [3,] "SNS" "163861" "0.427" "0.331" "0.775" "3" "0.138" "0.463" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "429122" "" "" "" "0.203" Age = 4 . Catchability constant w.r.t. time and dependent on age Year class = 2007 Fleet = BTS-comb 4 3 2 1 Survivors 296480.000 290038.000 305370.000 214799.000 Raw weights 8.323 6.689 4.172 4.206 Fleet = fshk 4 Survivors 305432.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 97151.000 87706.000 159124.000 Raw weights 0.555 0.572 1.299 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "279502" "0.158" "0.072" "0.456" "4" "0.897" "0.271" [2,] "fshk" "305432" "1.731" "Inf" "Inf" "1" "0.01" "0.25" [3,] "SNS" "123504" "0.428" "0.194" "0.453" "3" "0.093" "0.533" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "259264" "" "" "" "0.289" Age = 5 . Catchability constant w.r.t. time and dependent on age Year class = 2006 Fleet = BTS-comb 5 4 3 2 1 Survivors 326477.000 281557.000 270276.000 296927.000 255568.000 Raw weights 9.203 7.584 6.143 3.803 4.149 Fleet = fshk 5 Survivors 251789.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 74674.00 97260.000 129491.000 Raw weights 0.51 0.522 1.281 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "289990" "0.142" "0.044" "0.31" "5" "0.923" "0.177" [2,] "fshk" "251789" "1.82" "Inf" "Inf" "1" "0.007" "0.201"


[3,] "SNS" "107521" "0.426" "0.159" "0.373" "3" "0.069" "0.42" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "270479" "" "" "" "0.188" Age = 6 . Catchability constant w.r.t. time and dependent on age Year class = 2005 Fleet = BTS-comb 6 5 4 3 2 1 Survivors 142233.000 144560.000 127248.000 142052.00 115984.000 68270.000 Raw weights 9.923 8.635 7.112 5.56 3.034 2.687 Fleet = fshk 6 Survivors 86970.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 59132.000 60648.000 75428.00 Raw weights 0.461 0.416 0.83 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "130257" "0.134" "0.086" "0.642" "6" "0.95" "0.11" [2,] "fshk" "86970" "1.89" "Inf" "Inf" "1" "0.006" "0.16" [3,] "SNS" "66969" "0.44" "0.082" "0.186" "3" "0.044" "0.203" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "126182" "" "" "" "0.113" Age = 7 . Catchability constant w.r.t. time and dependent on age Year class = 2004 Fleet = BTS-comb 7 6 5 4 3 2 1 Survivors 55062.000 52490.000 80018.000 88387.000 67166.000 62904.00 48228.000 Raw weights 7.407 8.652 7.016 5.375 3.548 1.76 1.793 Fleet = fshk 7 Survivors 52212.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 11607.000 25065.000 34319.000 Raw weights 0.294 0.241 0.554 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "64192" "0.133" "0.086" "0.647" "7" "0.964" "0.091" [2,] "fshk" "52212" "1.909" "Inf" "Inf" "1" "0.007" "0.11"


[3,] "SNS" "23881" "0.438" "0.322" "0.735" "3" "0.03" "0.227" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "62258" "" "" "" "0.093" Age = 8 . Catchability constant w.r.t. time and dependent on age Year class = 2003 Fleet = BTS-comb 8 7 6 5 4 3 2 1 Survivors 113477.000 124947.000 78902.000 113241.000 105382.000 76626.00 64627.000 68194.00 Raw weights 10.205 6.829 8.297 7.096 5.742 3.72 1.962 1.86 Fleet = fshk 8 Survivors 66937.00 Raw weights 0.25 Fleet = SNS 3 2 1 Survivors 30684.000 28037.000 60985.000 Raw weights 0.309 0.269 0.574 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F [1,] "BTS-comb" "98846" "0.121" "0.079" "0.653" "8" "0.97" "0.083" [2,] "fshk" "66937" "1.917" "Inf" "Inf" "1" "0.005" "0.121" [3,] "SNS" "42311" "0.438" "0.259" "0.591" "3" "0.024" "0.185" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "96616" "" "" "" "0.085" Age = 9 . Catchability constant w.r.t. time and dependent on age Year class = 2002 Fleet = BTS-comb 9 8 7 6 5 4 3 2 1 Survivors 10805.0 26221.000 21721.00 35674.000 25981.000 34172.000 23452.000 26162.000 20020.000 Raw weights 6.4 9.901 6.61 7.875 6.052 4.127 2.628 1.181 1.199 Fleet = fshk 9 Survivors 6759.00 Raw weights 0.25 Fleet = SNS 3 2 Survivors 7323.000 12687.000 Raw weights 0.218 0.162 Fleet Est.Suvivors Int. s.e. Ext. s.e. Var Ratio N Scaled Wgts Estimated F


[1,] "BTS-comb" "23996" "0.122" "0.128" "1.049" "9" "0.986" "0.046" [2,] "fshk" "6759" "1.954" "Inf" "Inf" "1" "0.005" "0.154" [3,] "SNS" "9256" "0.705" "0.272" "0.386" "2" "0.008" "0.115" Weighted prediction: Suvivors Int.s.e. Ext.s.e. Var.Ratio F [1,] "23649" "" "" "" "0.047"

Figure APPcombCor2. Catchability at age of the three used in the combined BTS XSA assessment of the North Sea plaice stock (index option ‘combCor_hist): BTS-comb, BTS-Isis_old and SNS.


Annex 7: Participants list

NAME ADDRESS PHONE/FAX E-MAIL Aukje Coers Wageningen IMARES

PO Box 68 1970 AB Ĳmuiden Netherlands

Phone +317 487 136

[email protected]

Colin Millar Invited Expert

Joint Research Centre Institute for Protection and Security of the Citizen Via E. Fermi 2749 21027 Ispra (VA) Italy

Phone +39 0332 785208 Fax +39 0332 789658

[email protected]

David Miller Chair

Wageningen IMARES PO Box 68 1970 AB Ĳmuiden Netherlands

Phone +31 3174 853 69 Fax +31

[email protected]

Barbara Schoute International Council for the Exploration of the Sea H. C. Andersens Blvd. 44–46 1553 Copenhagen V Denmark

Phone +45 33 38 67 56 Fax +45

[email protected]

David Stokes Invited Expert

Marine Institute Rinville Oranmore Co. Galway Ireland

Phone +353 (0)91 387200 Fax +353 (0)91 387201

[email protected]

Clara Ulrich DTU Aqua - National Institute of Aquatic Resources

Jægersborg Allé 1

2920 Charlottenlund

Denmark

Phone +45 21157486

Fax +45 3588 3333

[email protected]


Annex 8: Comments by reviewers

Reviewers: Colin Millar ([email protected]) and David Stokes ([email protected]). These comments were received by e-mail.

Colin Millar

Firstly I agree with the combination of the survey indices to reduce the effect off po-tential offshore movement of juvenile plaice. What follows is an independent look at the plaice data.

There are two approaches that come to my mind. 1) combine the indices as has been done, or 2) model the change in distribution. Both approaches, if there is sufficient information, are valid; but there are benefits and drawbacks in both approaches.

1 ) Combining the indices produces a new series of indices, and is the most at-tractive option, not least because it utilizes \emph{raw} data by combining abundance observations by haul. This crux of this approach is to be able to estimate sufficiently well the relative catchabilities between the surveys be-ing combined. A second issue is that this relative catchability should be constant in time; but this is commonly assumed for surveys in the North Sea.

2 ) The second approach is to model the catchabilities at age. This approach maintains the different survey indices as independent time-series, and combines them at a higher level. This has the benefit of modelling the changes in relative catchability rather than fixing them, but this requires the stock assessment model to allow changes in catchability.

If XSA is the chosen assessment method, then it is really the only option.

We can investigate differences between the survey indices quite simply by plotting the log ratio of the Tridens and Isis survey indices, which are plotted in the supple-mentary document. From the document it is clear that the changes in availability occur for ages 1 to 4, with the effects being greater the younger the age classes. For ages 5 to 8 there is no evidence of a change in availability.

This document also brings out the problem that availability and relative catchability are confounded. For example, in Figure 1 is the difference in intercept between age 5 and age 6 due to differences in relative catchability or availability? Unfortunately it is not possible to assess relative catchability directly from these data. However, if this model could be incorporated into a stock assessment model then estimates of relative catchability could be estimated as part of the fitting process. The residuals from such a fit are presented in the document (Figure 4).

To conclude, there is a clear problem with the use of survey indices covering different areas of the plaice stock, when the proportion of the stock with in each survey region is changing with time. This is appears to be the case here, and the pattern in time seems to be fairly smooth in time. It appears that not only does the Tridens and Isis suffer from this issue, but also the SNS survey; in fact the problem seems to be greater in the SNS survey than the Isis survey. This raises the question of whether combining only the Tridens and Isis will actually solve the problem.

I recommend the combination of the surveys, but I would prefer to have more evi-dence of the values of relative catchability used to combine the indices.

mailto:[email protected]

mailto:[email protected]


I recommend that a future benchmark should consider using a statistical catch-at-age model. Not only will this allow for observation errors in the catch data, but would also allow the uncertainty in the relative catchability between the surveys to be in-cluded and for the assumptions about relative catchability to be better investigated.

I recommend that a future benchmark considers a way to incorporate the SNS survey index into the combined index, or considers an approach that models changes in catchability.

David Stokes

Firstly, there is a “Yes” for the combined index. The feeling of the WGISDAA group and my own is broadly that there is some improvement in precision as one might expect when combining data.

Secondly, the accuracy of a combined index is acknowledged in the documents pro-vided to rely heavily on the conversion factors. There is disagreement in the conver-sion factors published and there has been some discussion at WGISDDA about that. Neither calibration document however presents selectivity by length or even cohort. In the conversion used for IB Plaice to combine data, WGBEAM 2005 catches are ag-gregated into two groups across hauls, large and small. It’s hard to get a sense of any length or cohort effects in terms of either catchability or availability.

Beam trawl selectivity may not be as complex as otter trawl work, but I think the approach taken by Fryer et al., 2003 (see attached) which also fed into the IPROSTS calibration project as part of IBTS does outline some important assumptions around gear selectivity in general. I’d also suggest looking at the inter-calibration section in the ICES Survey Analysis and Design workshop WKSAD 2004 for discussion on as-sumptions around using conversion based on catches vs. indices vs. lengths.

Overall the combined index looks fine; we’ll have to assume the conversion is precise enough to not add significant bias at this point. I would strongly recommend further work in the coming surveys re. a more detailed analysis of the calibration data. I un-derstand both survey trawls have been towed simultaneously in the past by Tridens at least and would therefore provide an important data dataset for analysing relative catchability at the length/cohort level. This type of approach should also target the catchability issue directly, by eliminating availability from the equation which is an important confounding issue here.

I couldn’t see the table (8.2.12) outlining additional recruitment indices in the version of the stock annex I have. I assume the IBTS Q1 and Q3 indices would be in there, but aren’t used. It might be worth asking WGNSSK to request IBTS to have a look at the NSIBTS indices in case some additional data could be provided for the problem co-horts in particular that seem to be on the move in this area. This combined index sur-vey covers the entire North Sea in both quarters 1 and 3 and should give an additional spatial and temporal picture to what is going on.


Annex 8.1: Modelling the log ratio of the indices to compare mod-elled relative catchabilities and empirical estimates

Let

I_ay1 = Q_a1 A_ay1 N_ay, be the survey index for age a, year y for the Tridens, say (Index = catchability(Q) x availability (A) x Abundance (N))

and similarly for the other survey

I_ay2 = Q_a2 (1 - A_ay1) N_ay

if the surveys together cover the whole (or a constant proportion of the) stock then the availability should sum to 1 (or a constant) for each age and year (I think this is implicit in your calculations also). The catchabilities are constant in time, but the availability is changing with time. And so, when you combine the indices you only need to know the relative catchabilities of the gears. - as you are doing.

then the log ratio of the two indices is

log(I_ay1 / I_ay2) = log(Q_a1/Q_a2) + logit(A_ay1)

i.e.

r_ay = q_a + p_ay

(let’s say) - r for ratio, q for log relative catchability, and p for logit availability

i.e. the log of the relative catchbility at age plus the logit of the availability which has been changing over time.... also taking logs will homogenize the variance

so if you fit a linear model (gaussian errors) to the log ratio of the survey indices, and include year as a smooth term then the intercept should give you the log relative catchability (with confidence intervals), which you can compare to the value you use in you calculations

You can either assume constant CV across ages and fit something like

r ~ factor(age) + s(year, by = age)

or do separate regressions for each age


Annex 9: Stock Annex Plaice in Area IV

Stock North Sea plaice

Working Group WGNSSK

Date April 2013

By Aukje Coers and David Miller

A. General

A.1 Stock definition

The North Sea plaice is defined to be a single stock in ICES Area IV. However, data from data storage tag experiments reveal that about one third of plaice released in the Southern Bight of the North Sea visit the eastern English Channel in December and January. In contrast, analysis of the movements of mark-recapture experiments with plaice of a similar size and released at similar times indicates that only 13% of plaice released in the Southern Bight visit the eastern English Channel at this time (Hunter et al., 2004). This difference between DST and mark–recapture experiments is not observed in the central North Sea and German Bight, where the movements of plaice derived from the two approaches are relatively similar (Bolle et al., 2005). The differ-ences may possibly be due to the fact that these fish migrate to their spawning grounds by selective tidal stream transport. Studies (Kell et al., 2004) have shown that the migration between North Sea and the adjacent areas is more problematic for the smaller adjacent areas than it is for management in IV.

Genetic analysis of plaice population structure in northern Europe using microsatel-lites and mitochondrial DNA data (Hoarau et al., 2004) reveals relatively strong dif-ferentiation between “shelf” plaice and those from Iceland and Faroe, suggesting that deep water may serve as a barrier to movement between these populations. Howev-er, within the area of the European continental shelf, only weak differentiation could be detected between North Sea–Irish Sea and other areas (Norway, the Baltic and the Bay of Biscay, Hoarau et al., 2004). Although the spatial location of sampling within the North Sea was not sufficient to reveal any substructure. The lack of any genetic differentiation between Irish Sea and North Sea plaice populations (Hoarau et al., 2004) despite the evidence from mark–recapture studies that indicate extremely low transfer of individuals between these sea areas (0.36% over 17 years, calculated from (Dunn and Pawson, 2002)) shows how differently genetic and tagging studies pro-vide an understanding fish population structure. Nonetheless, it seems unlikely that Irish Sea and North Sea plaice are a single “stock”, at least in a fisheries management sense.

A.2 Fishery

North Sea plaice is taken mainly in a mixed flatfish fishery by beam trawlers in the southern and southeastern North Sea. Directed fisheries are also carried out with seines, gillnets, and twin trawls, and by beam trawlers in the central North Sea. Due to the minimum mesh size enforced (80 mm in the mixed beam trawl fishery), large numbers of (undersized) plaice are discarded. Fleets exploiting North Sea plaice have generally decreased in number of vessels in the last ten years. However, in some in-stances, reflagging vessels to other countries has partly compensated these reduc-tions. For example, approximately 85% of plaice landings from the UK(England and


Scotland) are landed into the Netherlands by Dutch vessels fishing on the UK regis-ter. Vessels fishing under foreign registry are referred to as flag vessels, which may have different fishing patterns from the rest of the fleet. Besides having reduced in number of vessels, the fleets have also shifted towards two categories of vessels: 2000 HP (the maximum engine power allowed) and 300 HP (the maximum engine power for vessels that are allowed to fish within the 12 mile coastal zone and the plaice box). Also, the decrease in fleet size may partially have been compensated by slight increases in the technical efficiency of vessels. In the Dutch beam trawl fleet indications of an increase of technical efficiency of around 1.65% by year was found over the period 1990–2004 (Rijnsdorp et al., 2006). Because the commercial tuning-series are not currently used in the assessment, these estimates do not affect the cur-rent assessment.

The Dutch beam trawl fleet, one of the major operators in the mixed flatfish fishery in the North Sea, has seen a shift towards more inshore fishing grounds, changing the catchability of the fleet. This shift may be caused by a number of factors, such as the implementation of fishing effort restrictions, the increase in fuel prices and changes in the TAC for the target species (Quirijns, 2008). However, the contribution of each of these factors is yet unknown. Other factors affecting the catchability of the fleet in-clude the changes in the fishing speed of the vessels, and discarding marketable fish in certain seasons and areas, as a result of the TAC management (Rijnsdorp, 1991).

Conservation schemes and technical conservation measures

Fishing effort has been restricted for demersal fleets in a number of EC regulations (EC Council Regulation No. 2056/2001; EC Council Regulation No 51/2006; e.g. N°40/2008, Annex IIa). For example, for 2007, Council Regulation (EC) No 41/2007 allocated different days at sea depending on gear, mesh size, and catch composition: Beam Trawls could fish between 123 and 143 days per year. Trawls or Danish seines could fish between 103 and 280 days per year. Gillnets could allowed to fish between 140 and 162 days per year. Trammelnets could fish between 140 and 205 days per year.

Several technical measures are applicable to the plaice fishery in the North Sea: mesh size regulations, minimum landing size, gear restrictions and a closed area (the plaice box).

Mesh size regulations for towed trawl gears require that vessels fishing North of 55°N (or 56°N east of 5°E, since January 2000) should have a minimum mesh size of 100 mm, while to the south of this limit, where the majority the plaice fishery takes place, an 80 mm mesh is allowed. In the fishery with fixed gears a minimum mesh size of 100 mm is required. In addition to this, since 2002 a small part of North Sea plaice fishery is affected by the additional cod recovery plan (EU regulation 2056/2001) that prohibits trawl fisheries with a mesh size <120 mm in the area to the north of 56°N.

The minimum landing size of North Sea plaice is 27 cm. The maximum aggregated beam length of beam trawlers is 24 m. In the 12 nautical mile zone and in the plaice box the maximum aggregated beam length is 9 m. A closed area has been in opera-tion since 1989 (the plaice box). Since 1995 this area was closed in all quarters. The closed area applies to vessels using towed gears, but vessels smaller than 300 HP are exempted from the regulation. An evaluation of the plaice box has indicated that: from trends observed it was inferred that the Plaice Box has likely had a positive ef-fect on the recruitment of plaice but that its overall effect has decreased since it was


established. There are two reasons to assume that the Plaice Box has a positive effect on the recruitment of Plaice: 1) at present, the Plaice Box still protects the majority of undersized plaice. Approximately 70% of the undersized plaice are found in the Plaice Box and Wadden Sea, and despite the changed distribution, densities of juve-nile plaice inside the box are still higher than outside; 2) In the 80 mm fishery, discard percentages in the box are higher than outside. Because more than 90% of the plaice caught in the 80 mm fishery in the box are discarded, any reduction in this fishery would reduce discard mortality. There is, however, no proof of a direct relationship between total discard mortality and recruitment.

Generally, it is assumed that the majority of discarded animals do not survive (Beek et al., 1990; Chopin et al., 1996). Reviews of studies that have tested this assumption acknowledge that discard mortality is determined by a range of biological, technical, or environmental factors or 'stressors' (Broadhurst et al., 2006). Biological factors re-late to e.g. the species, physiology, size, catch weight/volume, composition; technical stressors relate to e.g. gear design, deployment duration, fishing speed; environmen-tal stressors relate to e.g. temperature, hypoxia, depth, wind force, availability of sunlight.

For the beam trawl fishery, discard mortality is influenced by the duration the organ-isms are confined in the codend and concurrent injuries (Beek et al., 1990; Broadhurst et al., 2006). If the fish were brought onboard alive, then the processing of the catch onboard would also matter. However, in fact, processing on board hardly affects the survival of the discards because approximately 70% of the catch is moribund upon landing already (Beek et al., 1990). It is estimated based on experimental studies onboard commercial vessels that less than 10% of the plaice and sole discards in the beam trawl fisheries survive the process of discarding (Bult and Schelvis-Smit, 2007; Beek et al., 1990).

A.3 Ecosystem aspects

Adult North Sea plaice have an annual migration cycle between spawning and feed-ing grounds. The spawning grounds are located in the central and southern North Sea, overlapping with the distribution area of Sole. The feeding grounds are located more northerly than the sole distribution areas. Juvenile stages are concentrated in shallow inshore waters and move gradually off-shore as they become larger. The nursery areas on the eastern side of the North Sea contribute most of the total re-cruitment. Subpopulations have strong homing behaviour to specified spawning grounds and rather low mixing rate with other subpopulations during the feeding season (de Veen, 1978; Rijnsdorp and Pastoors, 1995). Genetically, North Sea and Irish Sea plaice are weakly distinguishable from Norway, Baltic and Bay of Biscay stocks using mitochondrial DNA (Hoarau et al., 2004).

Juvenile plaice were distributed more offshore in recent years. Surveys in the Wad-den Sea have shown that 1-group plaice is almost absent from the area where it was very abundant in earlier years (van Keeken et al., 2007). The Wadden Sea Quality Status Report 2004 (Vorberg et al., 2005) notes that increased temperature, lower lev-els of eutrophication, and decline in turbidity have been suggested as causal factors, but that no conclusive evidence is available; taking into account the temperature tol-erance of the species there is ground for the hypothesis that a temperature rise con-tributes to the shift in distribution.

A shift in the age and size at maturation of plaice has been observed (Grift et al., 2007; Grift et al., 2003): plaice become mature at younger ages and at smaller sizes in recent


years than in the past. This shift is thought to be a genetic fisheries-induced change: Those fish that are genetically programmed to mature late at large sizes are likely to have been removed from the population before they have had a chance to reproduce and pass on their genes. This results in a population that consists ever more of fish that are genetically programmed to mature early at small sizes. Reversal of such a genetic shift may be difficult. This shift in maturation also leads to mature fish being of a smaller size at age, because growth rate is reduced after maturation.

B. Data

B.1 Commercial catch

Discard sampling programmes started in the late 1990s to obtain discard estimates from several fleets fishing for flatfish. These sampling programmes give information on discard rates from 1999 but not for the historical time-series. Observations indicate that the proportions of plaice catches discarded are high (80% in numbers and 50% in weight: (van Keeken et al., 2004)) and have increased since the 1970s (51% in numbers and 27% in weight: (van Beek, 1998)). The discards time-series are derived from Dutch, Danish, German and UK discards observations for 2000–2007. For the period prior to that, a reconstructed discard time-series for 1957–1999 exists, based on a re-constructed population and selection and distribution ogives (ICES, 2005a).

The discard data from the sampling programmes in the individual countries are raised totals, based on samples from onboard observers. These observers generally take length-structured samples of which a subsample is aged through otolith read-ings.

The UK discards estimates have strong interannual variation, caused by the low sample sizes, and sampling different strata in the UK fleet. For example, the UK dis-card samples for 2007 were taken mainly from the UK Nephrops and otter trawl fish-ery. These fisheries represent only a small fraction of the total UK plaice landings, and raising the UK discards using only samples from this fleet would potentially lead to incorrect estimates. Since the UK landings represents 24% of the total nominal landings, obtaining accurate discard estimates is crucial. In order to gain better esti-mates of discards, the proportionality of the English discards to the Dutch discards is calculated in the observations since 2000. The UK estimates are recalculated assuming a constant ratio between the UK and Dutch discard numbers-at-age:

NLya

y

NLya

y

UKya

UKya D

D

DD ,2007

2002,

2007

2002,

,ˆ ×=

∑

∑

=

=

1

where UKyaD , , UK

yaD ,ˆ , and NL

yaD , are the observed and estimated UK, and observed Dutch

discard numbers of year y and age a, respectively.

After raising to the fleet total and estimation of discards-at age using age–length keys from the Dutch BTS surveys, discard observations at age are thus available from the Dutch, Danish, German and the UK discard sampling programmes. The sampling effort in the Dutch and UK programmes is given in the quality of the estimation of total discards numbers-at-age depends on the quality of the available discards data,


which are derived from low sampling level discards observations within the four countries that have provided discard estimates.

Discards at age were raised from the Dutch and UK sampling programmes by effort ratio (based on hp days at sea for the Dutch fleets, and on trips for the UK fleets). Discards at age from the Danish and German sampling programmes were raised by landings. Discards at age for the other fleets for which no estimates were available, were calculated as a weighted average of the Dutch, Danish, German and UK dis-cards at age and raised to the proportion in landings (tonnes). This is the same meth-od as used in the final assessment by WGNSSK 2005 (method B).

A self-sampling programme for discards was started by the Dutch beam trawl fishery in 2004, and is still running. This sampling programme has a large number of sam-ples, taken onboard by the fishermen, estimating the percentage of discards by vol-ume. The program indicates a strong spatial pattern in the discarding of the fleet. The percentage discards estimated in the self-sampling programme is significantly lower than that in the Dutch sampling programme in the same years (Aarts and van Helmond, 2007).

To reconstruct the number of plaice discards at age before 2000 that are required for an XSA assessment, catch numbers-at-age are calculated from fishing mortality-at-age corrected for discard fractions, using a reconstructed population and selection and distribution ogives (ICES, 2005a). Alternatively, the discards previous to 2000 can be estimated using the statistical catch-at-age approach as described in (Aarts and Poos, 2009).

Landings

The landings by country are collected by different countries, segregated by sex for the Netherlands and Belgium (accounting for approximately 50% of the landings). Age structure is available for the Netherlands, France, Germany, Denmark and Belgium (accounting for approximately 75% of the landings). The total age-structured land-ings are estimated using a weighed procedure for the age structure by country, based on the proportionality of the weight of the total landings.

During the benchmark of the eastern channel (VIId) plaice stock (WKFLAT, 2010) it was decided that 50% of Q1 landings taken in the eastern channel are actually plaice from the North Sea stock migrating in and out of the area. The impact was found to be minimal, given that as a percentage of the total catch, these eastern channel land-ings, available back to 1980, account for less than 1% each year. Since 2012, 50% of the Q1 eastern channel (VIId) plaice landings are included in the assessment of the North Sea plaice stock. See the stock annex for plaice in Division VIId for further de-tails.

B.2 Biological

Weight-at-age

The stock weights of age groups 1–4 are calculated using modelled mean lengths from survey and back-calculation data (ICES, 2005a) and converted to mean weight using a fixed length–weight relationship. Stock weights of the older ages are based on the market samples in the first quarter. Stock weight-at-age has varied considerably over time, especially for the older ages. Discard weights-at-age are calculated the same way as the stock weights of age groups 1–4, after which gear selection and dis-carding ogives are applied. Landing weights-at-age are derived from market sam-


pling programmes. Catch weights-at-age are calculated as the weighted average of the discard and landing weights-at-age. There appear to be cohort effects on landings weight-at-age, which are also reflected in the stock weights-at-age. In addition to the cohort effects, there is a long-term decline in weight-at-age for the older ages. The stock weights of the older ages are based on the market samples in the first quarter. In these market samples, the sex ratio for the older ages may be skewed towards one of the sexes. The WG suggests a more in depth study into the causes and consequenc-es of the perceived decreases in stock weights for the next benchmark assessment.

Natural mortality

Natural mortality is assumed to be 0.1 for all age groups and constant over time. The-se values are probably derived from war time estimates (Beverton and Holt, 1957).

Maturity

A fixed maturity ogive is used for the estimation of SSB from the assessment in North Sea plaice, assuming maturity-at-age 1 is 0, maturity-at-age 1 and 2 is 0.5, and older ages are fully mature. However maturity-at-age is not likely to be constant over time (Grift et al., 2003; Grift et al., 2007). The effects of assuming a constant maturity-at-age on the management advice was discussed in a study by (Kell and Bromley, 2004). However, a study of the effect of the fluctuations of natural mortality on the SSB by the WG in 2004 showed that incorporating the historic fluctuations had little effect on SSB estimates in the period 1999–2003.

B.3 Surveys

The stock assessment of North Sea plaice uses indices from three surveys: (1) the Beam Trawl Survey RV Isis (BTS-Isis), (2) the Beam Trawl Survey RV Tridens (BTS-Tridens) and (3) the Sole Net Survey (SNS). The Beam Trawl Survey RV Isis (BTS-Isis) was initiated in 1985 and was set up to obtain indices of the younger age groups of plaice and sole, covering the southeastern part of the North Sea (RV Isis). Since 1996 the BTS-Tridens covers the central part of the North Sea, extending the survey area of the surveys. Both vessels use an 8 m beam trawl with 40 mm stretched mesh codend, but the Tridens beam trawl is rigged with a modified net. Owing to the spatial distri-bution of both BTS surveys, considerable numbers of older plaice and sole are caught. Previously age groups 1 to 4 were used for tuning the North Sea plaice assessment, but the age range has been extended to 1 to 9 in the revision done by ACFM in Octo-ber 2001.

In 2009, WKFLAT (the ICES Benchmark and Data Compilation Workshop for Flat-fish; ICES, 2009) addressed concerns about differing signals in the three scientific tuning indices. It was concluded that the trends in the indices of these surveys may no longer independently reflect the trend in abundances of age classes in the whole of the North Sea. As it was considered generally to be better to have indices spanning as much of the distribution as possible rather than having separate indices, which are prone to showing local trends, it was decided to combine the two separate BTS-indices into one. Following an Inter-Benchmark Protocol, in March 2013, (Miller and Coers, 2013) the two BTS-indices were combined applying a gear efficiency correction as recommended by WGBEAM (ICES, 2005b). The assessment is conducted using the combined BTS-indices, while maintaining the early part of the BTS-Isis index in the assessment separately (since this survey started eleven years before the BTS-Tridens and so only part of the time-series can be combined). The survey areas of the BTS-Tridens and BTS-Isis are shown in the figure below.


The Sole Net Survey (SNS & SNSQ2) was carried out with RV Tridens until 1995 and then continued with the RV Isis. Until 1990 this survey was carried out in both spring and autumn, but after that only in autumn. The gear used is a 6 m beam trawl with 40 mm stretched mesh codends. The stations fished are on transects along or perpen-dicular to the coast. This survey is directed to juvenile plaice and sole. Ages 1 to 3 are used for tuning the North Sea plaice assessment; the 0-group index is used in the RCT3. In an attempt to solve the problem of not having the survey indices in time for the WG, the SNS was moved to spring in 2003. However, because of the gap in the spring series these data could not be used in the plaice assessment or in RCT3. In 2004, the SNS was moved back to autumn as before, based on the recommendation of the WGNSSK in 2004.

The 1997 survey results for the 1995 and 1996 year classes (at ages 1 and 2) in the BTS and SNS surveys cannot be used in the assessment, owing to age reading problems in that year. Also, the research vessel survey time-series have been revised in May 2006 by WGBEAM, because of small corrections in databases and new solutions for miss-ing lengths in the age–length-keys.

The three different survey indices used are:

• Beam Trawl Survey combined for RV Tridens and ISIS (BTS-combined); (1996–2012);

• Beam Trawl Survey RV Isis (BTS-Isis) for the older part of the time-series; (1985–1995);

• Sole Net Survey (SNS); (1982–2012).

An additional Survey index that can be used for recruitment estimates are:

• Demersal Fish Survey (DFS).

The Demersal Fish Survey (DFS) is the more coastal of the surveys, conducted by several countries. This survey is not used in the assessment, but rather used to esti-mate the recruitment of juvenile fish in the RCT3 analysis. The survey estimates abundances for North Sea plaice age 0 and age 1. However, the age 1 has not been used for recruitment estimation since a number of years, and the time-series for this age was stopped in 2005. Since 2013, a revised time-series from 1990 onwards is used,


because the UK survey was terminated and thus removed from the international combined index (ICES, 2012).

B.4 Commercial lpue

Commercial age-structured lpue series (consisting of an effort-series and land-ings-at-age series) that are available to be used (but are currently not included in the assessment anymore) as tuning fleets are:

• The Dutch beam trawl fleet (since 1989); • The Dutch beam trawl fleet corrected for spatial effort allocation (since

1997); • The UK beam trawl fleet excluding all flag vessels (between 1990 and

2002).

Effort has decreased in the Dutch beam trawl fleet since the early to mid-1990s. Up until 2002, the age classes available in both the Dutch and the UK fleets generally show equal trends in lpue through time.

The WG used both survey data and commercial lpue data for tuning until the mid-1990s. The commercial lpue was calculated as the ratio of the annual landings over the total number of fishing days of the fleet. At that time, however, it was realized that the commercial lpue data of the Dutch beam trawl-fleet, which dominated the fishery, were likely to be biased due to quota restrictions. Vessels were reported to adjust their fishing patterns in accordance to the individual quota available for that year. Fishers reported to leave productive fishing grounds because they lacked the fishing rights and moved to areas with lower catch rates of the restricted species with a bycatch of non-quota, or less restricted species.

A method that corrects for the spatial effort allocation is to calculate lpues at a smaller spatial scale, e.g. ICES rectangles, and then calculate the average of these ICES rec-tangle-specific lpues. Age information is available at this spatial level since 1997, and lpue series could be used for tuning an age-structured assessment method (alterna-tively, age-aggregated tuning-series could be used in other analytical assessment methods than XSA). Only under the assumption that discarding is negligible for the older ages, the lpue represents cpue, and this time-series could be used to tune age-structured assessment methods.

Also, age-aggregated lpue series, corrected for directed fishing under a TAC-constraint (see Quirijns and Poos, 2007), by area and fleet component, can be used as indication of stock development. Available are:

• The Dutch beam trawl fleet (only large cutters with engine powers above 221 kW);

• The UK beam trawl flag vessels landing in the Netherlands (only large cut-ters with engine powers above 221 kW);

• Several Danish fleets (trawl, gillnet and seines) mainly operating in the Northern area;

• Effort of the Dutch beam trawl fleet and of the English beam trawl vessels landing in the Netherlands, by area and fleet component.

B.5 Other relevant data

To be done.


C. Historical stock development

There are currently two methods that could be used to provide an assessment of North Sea plaice, being XSA, and a model developed by (Aarts and Poos, 2009). The XSA uses the reconstructed discard set described in the catch section. The Aarts and Poos method estimates the discards from the mortality signals in the surveys, the landings-at-age and the discards-at-age in the most recent period. WKFLAT 2009 suggest to run both models concurrently, in order to estimate the stability of the Aarts and Poos method.

Model used as a basis for advice

The North Sea plaice is based on the XSA stock assessment. Settings for the final as-sessment are given below:


Catch-at-age Landings (since 1957, ages 1–10) + (reconstructed) discards based on NL, DK + UK + GE fleets. Discards reconstruction between 1957–1999), observations since 2000.

Tuning indices Beam Trawl Survey combined for RV Tridens and ISIS (BTS-combined); (1996–present); 1–9. Beam Trawl Survey RV Isis (BTS-Isis) for the older part of the time-series; (1985–1995); 1–8. Sole Net Survey (SNS); (1982–present); 1–3.

Plus group 10

First tuning year 1982

Time-series weights No taper

Catchability dependent on stock size for age <

1


6

Survivor estimates shrunk towards the mean F

5 years/5 years

s.e. of the mean for shrinkage 2.0

Minimum standard error for population estimates

0.3



The Aarts and Poos model


Catch-at-age Landings (since 1980, ages 1:9) + discards based on observations since 2000 NL, DK + UK + GE fleets (ages 1:8). No reconstruction

Tuning indices Beam Trawl Survey combined for RV Tridens and ISIS (BTS-combined); (1996–present); 1–9. Beam Trawl Survey RV Isis (BTS-Isis) for the older part of the time-series; (1985–1995); 1–8. Sole Net Survey (SNS); (1982–present); 1–3.

Plus group No plus group

First tuning survey year 1980


8 (for catches)

Minimum standard error for likelihood function

0.05


D. Short-term projection

Because the assessment on which the advice is based is currently a fully deterministic XSA, the short-term projection can be done in FLR using FLSTF (1.4.3). Weight-at-age in the stock and weight-at-age in the catch are taken to be the average over the last three years. The exploitation pattern was taken to be the mean value of the last three years, scaled to the intermediate year F. The proportion of landings-at-age was taken to be the mean of the last three years; this proportion was used for the calculation of the discard and human consumption partial fishing mortality.

Population numbers at ages 3 and older are XSA survivor estimates. Numbers at age 2 are based on RCT3 estimates if the estimates from RCT3 show sufficient consisten-cy. Numbers at age 1 and recruitment of the incoming year class are taken from the long-term geometric mean of age 1 assessment estimates, where the most recent years are removed from the time-series, or are based on RCT3 estimates if the estimates from RCT3 show sufficient consistency.

The management options are given for two different assumptions on the F values in the intermediate year;

a ) F is assumed to be equal to the estimate for F in the final year of the as-sessment;

b ) F is set such that the landings in the intermediate year are equal to the TAC of that year.

E. Medium-term projections

Generally, no medium-term projections are done for this stock.

F. Long-term projections

Generally, no medium-term projections are done for this stock.


G. Biological reference points

The current reference points were established by the WGNSSK in 2004, when the discard estimates were included in the assessment for the first time. The stock–recruitment relationship for North Sea plaice did not show a clear breakpoint where recruitment is impaired at lower spawning stocks. Therefore, ICES considered that Blim be set at 160 000 t and that Bpa then be set at 230 000 t using the default multiplier of 1.4. Flim was set at Floss (0.74). Fpa was proposed to be set at 0.6 which is the 5th per-centile of Floss and gave a 50% probability that SSB is around Bpa in the medium term. Equilibrium analysis suggests that F of 0.6 is consistent with an SSB of around 230 000 t. In 2008, a target F was added to the reference points, based on the F stated in the long-term management plan for plaice and sole. This target F is supposedly based on an estimates of FMSY.

TYPE VALUE TECHNICAL BASIS

Management SSBMP 230 000 t Stage one: Article 2.

Plan FMP 0.6 0.3

Stage one: Article 2; Stage two: Article 4.

MSY MSYBtrigger 230 000 t Default to value of Bpa.

Approach FMSY 0.25 Simulation studies and equilibrium analyses taking into account a number of possible stock–recruitment relationships (range of 0.2–0.3).

Blim 160 000 t Bloss = 160 000 t, the lowest observed biomass in 1997 as assessed in 2004.

Precautionary Bpa 230 000 t Approximately 1.4 Blim.

approach Flim 0.74 Floss for ages 2–6.

Fpa 0.60 5th percentile of Floss (0.6) and implies that Beq>Bpa1) and a 50% probability that SSBMT ~ Bpa.

(unchanged since 2011).

The FMSY, FMAX and F0.1 should be estimated given the ten most recent years of the stock assessment.

H. Other issues

None identified.

I. References Aarts, G. and Poos, J. J. 2009. Comprehensive discard and abundance estimation of North Sea

plaice. ICES Journal of Marine Science, 66.

Aarts, G. and Van Helmond, A. T. M. 2007. Discard sampling of Plaice (Pleuronectes platessa) and Cod (Gadus morhua) in the North Sea by the Dutch demersal fleet from 2004 to 2006. ICES Document Report number C120/07. 42 pp.

Beek, F.A. van, Leeuwen, P.I. van, and Rijnsdorp, A.D. 1990. On the survival of plaice and sole discards in the otter-trawl and beam trawl fisheries in the North Sea. Netherlands Journal of Sea Research 26: 151–160.

Beverton, R. J. H. and Holt, S. J. 1957. On the dynamics of exploited fish populations, Her Majesty's Stationery Office, London (UK).


Bolle, L. J., Hunter, E., Rijnsdorp, A. D., Pastoors, M. A., Metcalfe, J. D. and Reynolds, J. D. 2005. Do tagging experiments tell the truth? Using electronic tags to evaluate conventional tagging data. ICES Journal of Marine Science, 62: 236–246.

Broadhurst, M.K., Suuronen, P., and Hulme, A. 2006. Estimating collateral mortality from towed fishing gear. Fish and Fisheries 7: 180–218.

Bult, T.P., and Schelvis-Smit, A. A. M. 2007. Een verkenning van de mogelijkheden van outriggen door vissers uitgevoerd, in het kader van het Advies van de “Task Force Duurzame Noordzeevisserij”. IMARES Report C022.07, 34 pp.

Chopin, F, Inoue, Y and Arimoto, T. 1996. Development of a catch mortality model. Fisheries Research 25: 377–382.

De Veen, J. F. 1978. On selective tidal transport in the migration of North Sea plaice (Pleuronectes platessa L.) and other flatfish species. Netherlands Journal of Sea Research, 12: 115–147.

Dunn, M. R. and Pawson, M. G. 2002. The stock structure and migrations of plaice populations on the west coast of England and Wales. Journal of Fish Biology, 61: 360–393.

Grift, R. E., Heino, M., Rijnsdorp, A. D., Kraak, S., B. M. and Dieckmann, U. 2007. Three-dimensional maturation reaction norms for North Sea plaice. Marine Ecology Progress Series, 334: 213–224.

Grift, R. E., Rijnsdorp, A. D., Barot, S., Heino, M. and Dieckmann, U. 2003. Fisheries-induced trends in reaction norms for maturation in North Sea plaice. Marine Ecology Progress Series, 257: 247–257.

Hoarau, G., Piquet, A. M.-T., van der Veer, H. W., Rijnsdorp, A. D., Stam, W. T. and Olsen, J. L. 2004. Population structure of plaice (Pleuronectes platessa L.) in northern Europe: a comparison of resolving power between microsatellites and mitochondrial DNA data. Journal of Sea Research, 51: 183–190.

ICES. 2005a. Report on the Assessment of Demersal Stocks in the North Sea and Skagerrak. 7–16 September 2004, Bergen, Norway. CM 2005/ACFM:07.

ICES. 2005b. Report of the Working Group on Beam Trawl Surveys (WGBEAM), 7–10 June 2005, Lowestoft, UK. ICES CM 2005/G:12, 83pp.

ICES. 2009. Report of the Benchmark and Data Compilation Workshop for Flatfish (WKFLAT 2009). ICES CM 2009/ACOM:31.

ICES. 2012. Report of the Working Group on Beam Trawl Surveys (WGBEAM). ICES CM 2012/SSGESST:11.

Kell, L. T. and Bromley, P. J. 2004. Implications for current management advice for North Sea plaice (Pleuronectes platessa L.): Part II. Increased biological realism in recruitment, growth, density-dependent sexual maturation and the impact of sexual dimorphism and fishery discards. Journal of Sea Research, 51: 301–312.

Kell, L. T., Scott, R. and Hunter, E. 2004. Implications for current management advice for North Sea plaice: Part I. Migration between the North Sea and English Channel. Journal of Sea Research, 51: 287–299.

Miller, D. C. M. and Coers, A. 2013. An examination of the use of the combined BTS index in the North Sea plaice assessment. CM 2013/ACOM:63. 78 pp.

Rijnsdorp, A. D. 1991. Selection differentials of male and female North Sea plaice Pleuronectes platessa L. and changes in maturation and fecundity. In The exploitation of evolving populations. Ed. by R. Law, T. K. A. Stokes and J. M. McGlade. Springer Verlag.

Rijnsdorp, A. D., Daan, N. and Dekker, W. 2006. Partial fishing mortality per fishing trip: a useful indicator of effective fishing effort in mixed demersal fisheries. ICES Journal of Marine Science, 63: 556–566.


Rijnsdorp, A. D. and Pastoors, M. A. 1995. Modelling the spatial dynamics and fisheries of North Sea plaice (Pleuronectes platessa L.) based on tagging data. ICES Journal of Marine Science, 52: 963–980.

van Beek, F. A. 1998. Discarding in the Dutch beam trawl fishery. ICES CM 1998/ BB:5: 1–15.

van Keeken, O. A., Quirijns, F. J. and Pastoors, M. A. 2004. Analysis of discarding in the Dutch beamtrawl fleet. 96 pp.

van Keeken, O. A., van Hoppe, M., Grift, R. E. and Rijnsdorp, A. D. 2007. Changes in the spatial distribution of North Sea plaice (Pleuronectes platessa) and implications for fisheries management. Journal of Sea Research, 57: 187–197.

Vorberg, R., Bolle, L. J., Jager, Z. and Neudecker, T. 2005. Chapter 8.6 Fish. In Wadden Sea Quality Status Report 2004.

Documents

International Council for the Exploration of the Sea Reports/Expert Group Report/aco… · Denmark . Telephone (+45) 33 38 67 00 . Telefax (+45) 33 93 42 15 . . [email protected] . Recommended