Figure 1: Research Blocks (red: catch advice not required ; blue: catch advice required) and Reference Areas (orange) in the Convention Area. The fishable depth range (600-1800m) is highlighted in green.

Figure 1: Research Blocks (red: catch advice not required ; blue: catch advice required) and Reference Areas (orange) in the Convention Area. The fishable depth range (600-1800m) is highlighted in green.

Background

For data-limited toothfish fisheries in the Southern Ocean there are generally no fishery-independent data on the status of the stock. Therefore, the collection of such data is included in the CCAMLR management process that sets the requirements for vessels to participate in those fisheries.

Determining the appropriate catch limits in data-limited toothfish fisheries that allow sufficient data collection for stock assessments, but that do not place stocks at risk in the interim of having an assessment, is a widespread topic in fisheries generally and one that has been a longstanding issue for CCAMLR (SC-CAMLR-XXIX, paragraphs 3.128-3.129).

In 2016, CCAMLR agreed to an approach for estimating biomass in data-limited areas that uses two methods, the CPUE-by-seabed area analogy and the Chapman mark-recapture estimation (WG-SAM-16 paragraph 2.28).

In theory the two methods for biomass estimation should provide similar results, however, differences in estimates between methods may arise due to limitations in these methods (e.g., these do not account for varying tag survival or fish migrations).

In 2017, CCAMLR (WG-FSA-17 paragraphs to 4.28 to 4.38) agreed that, where time-series of such biomass estimates are available, the trends in these estimates, as well as the most recent estimates of biomass, should be used to provide information on setting catch limits.

CPUE-by-seabed area analogy

A central element that underpins any assessment is the estimation of the biomass of the target species in the proposed fishing area. While CCAMLR has a number of data-limited toothfish fisheries it also has data-rich fisheries that are assessed using an integrated assessment software (CASAL), and, information from these assessed fisheries is used to guide the assessment of data-limited fisheries.

For data-limited fisheries, an initial estimate of biomass can be obtained using the biomass estimated in a data-rich area and assuming the same density of fish in the data-limited area. As fishing data becomes available this simple ‘seabed area’ approach scales biomass by the ratio of the Catch Per Unit Effort (CPUE; kg of fish caught per km of fishing line) in the data-rich area to that in the data-limited area (i.e., the CPUE-by-seabed area analogy method; Agnew et al., 2009) such that the biomass B can be estimated as:

\[ B_x = \frac{C_x \times A_x \times B_r}{C_r \times A_r} \]

Where the subscripts x and r denote parameters from the research block and reference/assessed area respectively. C is the median of the haul by haul CPUE where the total catch (t) on a line, including fish that are tagged and released, is divided by the length of line (km). A is the seabed area (km2) in the depth range 600-1800m and \(B_r\) is the current biomass estimate (t) from the most recent assessment in the reference area.

Chapman mark-recapture estimation

A requirement for participating in all CCAMLR data-limited toothfish fisheries is to tag and release fish at a rate of 3 to 5 fish per tonne caught such that mark-recapture data can be used to estimate biomass (e.g., via the Peterson/Chapman estimation methods).

The Chapman mark-recapture-based biomass estimation relies on the assumption that the ratio of tagged and un-tagged fish in the population is represented by that ratio in the catch. The biomass B in fishing season s can be estimated as:

\[ B_s = \frac{C_s \times ( n_{s-1} + 1)}{m_s + 1} \]

where \(n_{s-1}\) is the number of tagged fish available for recapture at the end of the previous fishing season, \(C_s\) is the catch in season s (as with CPUE the catch includes fish that are tagged and released) and \(m_s\) is the number of tagged fish recaptured in season s (excluding within-season recaptures).

Applying the Chapman method, only fish that have been tagged and recaptured within a research block (after adding a 5km buffer, WG-SAM-16 para 2.47), and have been at liberty for 1 to 3 years, are included. Time-at-liberty is further constrained to a single year in 486_2, 486_3 and 882H, due to evidence of extensive migration in these areas that results in overestimates of biomass (WG-FSA-17, paras 4.37(iv), 4.78, 4.80). Finally, the trend in Chapman biomass estimates is only considered if a threshold of adequate recaptures is reached (at least 3 recaptures per year in at least 2 of the last 3 years WG-FSA-17, para 4.34).

Trend Analysis decision rules

A set of trend analysis decision rules were developed by the Working Group on Fish Stock Assessments (WG-FSA-17) to determine a catch limit for each research block in a data-limited fishery. These decision rules have evolved over time to adapt management to fisheries dynamics.

In recent years, due to the lack of consensus over fishing in some areas (e.g., CCAMLR-40 paragraph 6.43), WG-SAM-21 (paragraph 3.32 (iii)) agreed that in order to establish catch limits in research blocks where fishing had not taken place in recent fishing seasons, if data were not available from the most recent fishing season, the previous catch limit should be carried forward. Such an approach should be limited to five years, after which time the catch limit would need to be re-evaluated outside the current trend analysis framework.

The time-series of biomass estimates (and their uncertainty; see WG-SAM-17/12 for details on the calculation of uncertainty) from both methods are used to evaluate overall trends in biomass. An inverse variance weighted least-squares regression is used to incorporate the confidence of each biomass estimate in the determination of the trend in the biomass time-series. To compare between trends across research blocks the standardized regression (beta) coefficient of the slope is estimated.

Each trend is then evaluated using a threshold of beta=0.1 so that the overall trend is determined to be:

Knowing the overall trend, whether there was an adequate number of recaptures and if the CPUE trend is declining enables a decision to be made based on the trend analysis decision rules. As a result, the recommended catch limit is computed in one of five ways (see Fig. 5):

The trends in biomass estimates from the CPUE-by-seabed area and Chapman mark-recapture methods are shown in Figures 2-4. The outcomes of the trend analysis decision rules are shown in Figure 5, and the resulting biomass estimates and catch limits for the next fishing season arising from these rules are in Table 1.

Figure 2: Annual biomass estimates (tonnes, with 95\% confidence intervals) and updated trends (dashed lines) for each Research Block in Area 48. Grayed panels indicate those research blocks that require catch advice for the upcoming season. Numbers in parentheses correspond to slope coefficients (|values|>0.1 are deemed significant) of trends while the other numbers indicate sample sizes (red: recaptures, blue: hauls).

Figure 2: Annual biomass estimates (tonnes, with 95% confidence intervals) and updated trends (dashed lines) for each Research Block in Area 48. Grayed panels indicate those research blocks that require catch advice for the upcoming season. Numbers in parentheses correspond to slope coefficients (|values|>0.1 are deemed significant) of trends while the other numbers indicate sample sizes (red: recaptures, blue: hauls).


Figure 3: Annual biomass estimates (tonnes, with 95\% confidence intervals) and updated trends (dashed lines) for each Research Block in Area 58. Grayed panels indicate those research blocks that require catch advice for the upcoming season. Numbers in parentheses correspond to slope coefficients (|values|>0.1 are deemed significant) of trends while the other numbers indicate sample sizes (red: recaptures, blue: hauls).

Figure 3: Annual biomass estimates (tonnes, with 95% confidence intervals) and updated trends (dashed lines) for each Research Block in Area 58. Grayed panels indicate those research blocks that require catch advice for the upcoming season. Numbers in parentheses correspond to slope coefficients (|values|>0.1 are deemed significant) of trends while the other numbers indicate sample sizes (red: recaptures, blue: hauls).


Figure 4: Annual biomass estimates (tonnes, with 95\% confidence intervals) and updated trends (dashed lines) for each Research Block in Area 88. Grayed panels indicate those research blocks that require catch advice for the upcoming season. Numbers in parentheses correspond to slope coefficients (|values|>0.1 are deemed significant) of trends while the other numbers indicate sample sizes (red: recaptures, blue: hauls).

Figure 4: Annual biomass estimates (tonnes, with 95% confidence intervals) and updated trends (dashed lines) for each Research Block in Area 88. Grayed panels indicate those research blocks that require catch advice for the upcoming season. Numbers in parentheses correspond to slope coefficients (|values|>0.1 are deemed significant) of trends while the other numbers indicate sample sizes (red: recaptures, blue: hauls).


Figure 5: The decision tree of the trend analysis, showing where each Research Block (grey ovals) requiring catch advice belongs this year.

Figure 5: The decision tree of the trend analysis, showing where each Research Block (grey ovals) requiring catch advice belongs this year.


Table 1. Research Blocks biomasses (B, tonnes) and catch limits (CL, tonnes) estimated using the trend analysis. Grayed cells indicate research blocks that require catch advice for the upcoming season. PCL: previous catch limit; ISU: increasing, stable or unclear; D: declining; Y: yes; N: no; -: no fishing in the last Season; x: no fishing in the last 5 Seasons; []: insufficent data. Recommended catch limits are subject to approval by the Commission.
Area Subarea or Division Research Block Species PCL Trend decision Adequate recaptures CPUE Trend Decline B Bx0.04 PCLx0.8 PCLx1.2 Recommended CL for 2023
48 48.1 481_1 D. mawsoni 43
43
481_2 D. mawsoni 43
43
481_3 D. mawsoni 0 x x x x x x x x
48.2 482_N D. mawsoni 75
75
482_S D. mawsoni 75
75
48.6 486_2 D. mawsoni 134 ISU Y N 3074 123 107 161 123
486_3 D. mawsoni 36 ISU N N 934 37 29 43 37
486_4 D. mawsoni 196 D Y Y 5366 215 157 235 157
486_5 D. mawsoni 210 D Y Y 40087 1603 168 252 168
58 58.4.1 5841_1 D. mawsoni 138
138
5841_2 D. mawsoni 139
139
5841_3 D. mawsoni 119 x x x x x x x x
5841_4 D. mawsoni 23 x x x x x x x x
5841_5 D. mawsoni 60
60
5841_6 D. mawsoni 104
104
58.4.2 5842_1 D. mawsoni 72 ISU Y Y 9935 397 58 86 86
5842_2 D. mawsoni 55 [] N [] 6450 258 44 66 258
58.4.3 5843a_1 D. eleginoides 19
19
58.4.4 5844b_1 D. eleginoides 18
18
5844b_2 D. eleginoides 14
14
88 88.2 882_1 D. mawsoni 230
230
882_2 D. mawsoni 223 ISU Y Y 9977 399 178 268 268
882_3 D. mawsoni 204 ISU N N 5193 208 163 245 208
882_4 D. mawsoni 154 ISU Y N 5862 234 123 185 185
882H D. mawsoni 102 ISU Y Y 10834 433 82 122 122
88.3 883_1 D. mawsoni 16
16
883_2 D. mawsoni 20
20
883_3 D. mawsoni 60 ISU N Y 6668 267 48 72 48
883_4 D. mawsoni 60 ISU N Y 2788 112 48 72 48
883_5 D. mawsoni 8
8
883_6 D. mawsoni 30 ISU N N 2289 92 24 36 36
883_7 D. mawsoni 30 ISU N N 2500 100 24 36 36
883_8 D. mawsoni 10
10
883_9 D. mawsoni 10 x x x x x x x x
883_10 D. mawsoni 10 x x x x x x x x

Additional Resources