Recommendation: Level 1 (Direct Delivery, Provisional)

Atlantic mackerel were last assessed in 2021 via a Management Track assessment; the most recent benchmark was in 2017 at S A W 64 . 2021 results indicated the stock was overfished based on S S B in 2019 (42,862 m t ) being 24% of the S S B[M S Y] proxy (S S B [40 percent] =181,090 m t ), and overfishing was occurring based on F 2019 (0.46) being 208% of the F[M S Y] proxy (F[40% S P R ] =0.22). The assessment included three indices: the N M F S Spring bottom trawl survey Albatross years from 1968; to ; 2008; the N M F S Spring bottom trawl survey Bigelow years from 2009; to ; 2019; and a range wide S S B index for 1977; to ; 2019 developed from theCanada D F O dedicatedegg surveyand the N E F S C MAR M A P and ECO MON surveys. Theassessment assumed constant natural mortality (M =0.2) and included one fishery fleet with time-invariant, flat- topped selectivity.

Kiersten Curti presented the proposed assessment plan for Atlantic mackerel in 2023, which will use the current A S A P model configuration with no changes and updated fishery and survey data through 2022. C A M S estimates of commercial landings and discards will be used for 2020; to ; 2022. Survey updates will include the 2021 and 2022 N M F S Spring bottom trawl survey (2020 survey was not conducted) and the S S B index for 2021 and 2022, if available. Reference points will be updated using the S A W 64 projection approach with M S Y level proxies of F[40% S P R ] and S S B [40 percent] . Rebuilding projections for 2023; to ; 2024 will be based on an assumed bridge year catch in 2023, two-stanza recruitment, and F[Rebuild] =0.12 as defined in the Atlantic Mackerel Rebuilding Amendment 2.0. The proposed backup assessment approach is the I smooth method using the S S B index developed from egg surveys.

The A O P raised questions about D F O data to support the S S B index and availability of data to support the assessment. Dr.Curti explained that the 2020 S S B index will be treated as missing, the 2021 S S B index is available, and the 2022 samples to support the S S B index are currently in transit. She expects that the 2022 S S B index will be available to support the assessment but noted that delays are possible. The A O P asked about model sensitivity to terminal year estimates and suggested that sensitivity analysis to examine the impacts of missing the terminal year S S B index may be warranted if the index is not available. The A O P also noted that this is the first iteration of the Atlantic mackerel assessment using C A M S data and recommended comparisons of C A M S landings and discards to outputs from previous methods to assess any substantial differences.

The A O P raised questions about application of the two-stanza recruitment assumptions for refer- ence points and projections. Dr.Curti highlighted previous deliberations by the 2021 Management Track assessment process and the S S C . She noted that there is no clear evidence of environmental conditions impacting recruitment. Despite high adult condition since the mid-2010s, recruitment has been low, but Dr.Curti indicated there is little evidence of a shift in environmental conditions. Research in Canada has indicated that S S B and temporal/spatial overlap of larvae with preferred prey are significant drivers of strong year classes. Without clear evidence that recruitment is environmentally driven, the 2021 Man- agement Track assessment did not change the S A W 64 assumptions for reference points and there are no proposed changes for the 2023 Management Track assessment.

The A O P supported continued development of the Ecosystem and Socioeconomic Profile (E S P ) for Atlantic mackerel, which describes ongoing examinations of natural mortality and stock productivity. The E S P will be provided as supporting information in 2023 and results to date do not indicate that changes to the assessment model are warranted.

The A O P recommended a provisional Level 1 review for Atlantic mackerel. The A O P supported a direct delivery of the assessment to the S S C based on the proposal to maintain the model configuration and update three years of fishery and survey data. The S S C recommended that a Level 2 review may be warranted if the 2022 S S B index is not available for the assessment update or if large differences in C A M S data are detected. The N E F S C will consider all available data in the coming months and determine if the review needs to be elevated to a Level 2.

Recommendation: Level 2 (Expedited Review)

Bluefish was last assessed in the Management Track in 2021 with data updated through 2019. That assessment utilized an A S A P statistical catch at age model to conclude that the stock was overfished, but overfishing was not occurring. Bluefish completed a Research Track assessment that was peer reviewed in December 2022. The newly accepted assessment developed a W H A M state space statistical catch- at-age model with deviations on the numbers at age estimates. Natural mortality, which was previously assumed constant at age-2, is now assumed to vary by age. The model employs two fishery fleets (recre- ational landings & discards and commercial landings), and 5 fleet selectivity blocks (2 commercial and 3 recreational). Three new indices were added to the model: M R I P C P U E Guild Approach index (1985; to ; 2021), S E A M A P Age-1 (1989; to ; 2021) and CHES M A P Trawl survey (1985; to ; 2018). The 2022 Research Track assessment (data through 2021) concluded that the stock was not overfished and overfishing was not occurring.

The 2023 Management Track assessment will update the current Research Track assessment with one year of additional data (2022). To address concerns of Research Track peer reviewers, the assessment will shift from full multi nomial age length keys to only using multi nomial approaches to fill in holes in age length keys (consistent with the approach used by Stock Eff). This may allow for exploration of alternate likelihoods for age compositions.

The assessment update will conduct short term projections in W H A M , which allows for incorpora- tion of model uncertainty, auto-regressive processes and uncertainty in recruitment and numbers-at-age. Removals in 2023 will be assumed to be equal to the 2023 A B C (13,890 m t ) and projections will be carried forward for years 2024; to ; 2026.

The A O P was concerned that the spawning biomass target has declined by 50% and is likely caused by changes in M using the Lorenzen curve resulting in a reduction in the recruits to fishable sizes. How- ever, the previous target had never been achieved in the fishery and was likely overinflated. This approach may represent a more reasonable level of reference points. It was noted that the S S C was concerned that the average weight of discards has disparities between the M R I P and angler surveys, likely due to higher average weights of large discarded fish. It was noted that the N E F S C and G A R F O have agreed to use the same values in setting specifications. Previously, the approach was overestimating discards, particularlyin the south (southern fish are generally smaller). The use of regionally stratified estimates is considered to be a more realistic and appropriate approach. The A O P recommended a Level 2 review for bluefish due to the significant reduction in the biomass target and proposed changes to the age-length key approach.

Recommendation: Level 2 (Expedited Review)

Deep-sea red crab is a data poor species that has not been considered in previous Management Tracks. A specifications update was completed in 2019 to set specifications for fishing years 2020; to ; 2023. This update included a time series from 2002; to ; 2019 including landings data from the limited access fleet, inci- dental landings, L P U E estimates for the limited access fleet and biological information from port samplers and observed trips. There is no assessment model, no biological reference points for this stock and none will be developed during this Management Track cycle.

This data update will add 4 years of data (2019; to ; 2022) including landings; L P U E estimates; port sampled carapace lengths for landed males; observer sampled carapace lengths for males; females and discarded males; and observer data on egg-bearing females and discards. There are no issues with missing 2020 survey data since the update relies on reported catches and some observed trips. The sexes segregate by depth and the fishery targets areas with higher densities of males. During the C A M S review, there were issues with the discards for some gear types. However, C A M S data are not used in the data update. A tagging project had low return rates indicating the potential for high mortality of tagged individuals, or a super abundant population. A Level 2 review of the available data and to highlight the limitations of analyses that have been attempted for this species is recommended to suggest potential approaches and generate useful research recommendations.

Recommendation: Level 2 (Expedited Review)

Longfin squid was last assessed in 2020 at a Management Track assessment. The overfishing status was unknown, but the stock was not overfished. The “not overfished" status was based on a comparison of the average of the 2018 and 2019 annualized, q -adjusted swept area biomass estimates (i.e., averages of the N E F S C spring and fall survey biomass for each year), 63,349 m t , to the threshold B[M S Y] proxy (B[Threshold] ) based on a long-term average (1976; to ; 2019). The threshold B[M S Y] proxy is 50% of the B[M S Y] target (i.e., 0.5×42,405 m t =21,203 m t ). It was also noted that the N E A M A P fall survey biomass estimates are added to those from the N E F S C fall survey.

Lisa Hendrickson’s presentation highlighted the complexity of longfin squid life history and the sea- sonal nature of the fishery which has both inshore and offshore components. Unlike Illex squid, longfin squid are neritic (i.e. residents of shelf waters). Hence, both the spring and fall N E F S C bottom trawl surveys likely sample most of the stock inhabiting United States waters. The S A R C 51 (2010) assessment approach considered the seasonal dynamics of the fisheries by calculating exploitation rates (catch/survey biomass) between the seasonal surveys. S A R C 51 concluded that annualized survey biomass estimates were more appropriate. Dr.Hendrickson noted cohort-based estimates of biomass and exploitation rates have alwaysbeen computed for squid caught in the spring versus fall surveys because the two cohorts have different growth rates and productivity levels. Although an approach based on analyzing each intra-annual cohort independently would be more realistic since it would capture the reliance of summer and fall fisheries on the recruits produced from the spring stock estimates, this approach has been deemed not permissible un- der the Management Track and will be suggested for exploration under the next Research Track. Similarly, the winter and spring fisheries depend on recruits produced from the fall survey stock estimates. Such a model would also allow inclusion of seasonal differences in growth rates.

Dr.Hendrickson recommended a change in the time series used to compute the B[M S Y] average from 1976; to ; 2008 to 1997; to ; 2022. The rationale was based on consideration of rapid warming and other changes in environmental conditions, and possibly productivity in recent years. Changes in fleet characteristics, data quality (i.e. mandatory fishery data reporting as of 1997), and in-season management as of 2000 were also considered relevant by the assessment lead to this proposed change. Questions from the A O P addressed the basis for the proposed change in years to compute the B[M S Y] average and whether there was any evidence of trends in the surveys. No trends have been observed but further analyses are needed. The selection of appropriate stanzas of years for projections or measures of productivity are always controver- sial, so justifying any changes should be data driven and well supported.

Additional questions from the A O P and other meeting attendees included the methods used to es- timate catchability in the trawl surveys, comparisons with assessments of species similarly impacted by environmental changes (e.g., Atlantic mackerel), and whether any preliminary changes had been detected. To account for diel vertical migrations, abundance and biomass estimates are based on daytime tows where “daytime" is defined by solar zenith angle because the species is most available to bottom trawls during the daytime. These values vary with location and date. The exclusion of tows outside the solar zenith angle ranges for the N E F S C spring and fall surveys reduces the frequency of low and zero tows, and generally improves precision but also reduces sample sizes within strata. The N E A M A P fall trawl survey swept area estimates will be updated because they are added to those of the N E F S C fall surveys. The N E A M A P spring survey’s intermittent encounters of longfin squid are attributed to varying availability of squid to the survey area; the stock is generally farther offshore in the spring.

Collectively, these considerations led the A O P to recommend a Level 2 review and a continuation of the current assessment methodology. The selection of an alternative basis for the B[M S Y] average should be fully explored and compared to the existing span of years. Results of both approaches should be presented to the M T A reviewers. The inclusion of newly developed C A M S estimates of landings and particularly discards, should be fully explored. Finally, M T A review can lay the groundwork for the Research Track assessment now scheduled for 2026. The groundwork could include any pending or required research on basic biology, alternative modeling approaches, and required data streams from the commercial fleets. The Terms of Reference for the assessment have not been set; the newly chartered Research Track Steering Committee of the N R C C will likely be involved in this process.

Recommendation: Level 3 (Enhanced Review)

A Research Track assessment for spiny dogfish was peer reviewed in December 2022. The Stock Synthesis 3 (S S 3 ) model was used with a time series of 1989; to ; 2019. The stock was not overfished based onReproductive Output in 2019 (239.9 million pups) being 65% of the S S B[M S Y] proxy (Reproductive Output Target =370.8 million pups), and overfishing was occurring based on F 2019 (0.032) being 128% of the F[M S Y] proxy (F[60% S P R ] =0.025). The assessment included the N M F S Spring and Fall bottom trawl survey indices and lengths, two landings fleets and three discard fleets, Lorenzen natural mortality estimates, and two maturity-growth relationship blocks. Dvora Hart presented the proposed assessment plan for spiny dogfish in 2023, which will use the current S S 3 model configuration with explorations and potential modification to the influence of the stock-recruit relationship and updated fishery and survey data through 2022. C A M S and S B R M derived estimates of commercial landings and discards and M R I P estimates of recreational landings will be used for 2020; to ; 2022. Landed and discarded length and sex data by gear type will be updated based on available information. Survey updates include the 2021 and 2022 (2020 survey was not conducted) N M F S Spring and Fall bottom trawl survey indices and lengths. Reference points will be updated using the Research Track approach with M S Y level proxies based on 60 percent of S P R . The projection method will be investigated to consider disproportional landings and discards. The proposed backup assessment approach is the previously used Stochastic Estimator model, which estimates F and S S B using swept area from the N M F S Spring survey with propagation of uncertainties.

The A O P raised questions about the backup assessment plan and potential challenges with applying reference points from the S S 3 model to the outputs from the Stochastic Estimator model. The A O P noted that it is unlikely that the S S 3 model would be rejected during the Management Track Peer Review as it was recently approved during the Research Track assessment. Dr.Hart noted that the new B R P s were approved through the Research Track assessment and would remain in place. The A O P asked about the influences of changes in natural mortality assumptions and age information included in the S S 3 assessment. Dr.Hart commented that the use of the Lorenzen M provides better results from the model and is more biologically realistic. Estimates of M range from 0.3 for newborn pups to 0.08 for large adult females, which influence the per recruit calculations and result in less females reaching the reproductive age. She noted that the only ageing study with a large scope was conducted ≈40 years ago and there is evidence that growth rates have changed. Length at maturity has decreased suggesting that either growth has slowed, or females are maturing at earlier ages. Smaller, slower growing females indicate reduced reproductive output. The 2022 Research Track assessment suggested that reproductive output has rapidly declined in recent years, and the stock may be approaching an overfished status.

The A O P and other meeting attendees inquired about providing estimates of F and biomass for males. Dr.Hart commented that this question has been raised in the past and she could provide these estimates but does not propose deriving reference points for males.

The A O P recommended a Level 3 review for spiny dogfish. The A O P supported the proposed explo- rations and potential changes to the assessment and recommended that ample time be allotted for presenta- tion and review during the Management Track Peer Review. The A O P noted the need for review of C A M S data and the potential for a change in stock status. They also highlighted that the 2023 Management Track is the first iteration of the S S 3 assessment since the Research Track in 2022 and highlighted major changes in estimates of natural mortality and length at maturity.

Recommendation: Level 2 (Expedited Review)

Scup were most recently assessed in 2021 via a Management Track assessment; the most recent benchmark assessment was in 2015 at S A W 60 . Mark Terceiro presented the proposed assessment plan for scup in 2023 that will rely on the model structure (A S A P ) but include updated fishery and survey data through 2022. C A M S estimates of commercial landings and discards will be used for 2020; to ; 2022, but questions remain about the commercial discard estimates. N E F S C trawl survey indices will include “by tow" area swept estimates but the effect of such changes on general trends are negligible. Some minor changes in tuning parameters will be used to improve model diagnostics. These parameters include the Coefficients of Variation (C V ) for some state abundance indices. Population projections will assume a catch in 2023 equal to the approved A B C of 13,458 m t .

The A O P inquired about the potential effects of missing N E F S C survey data in 2020 and the effects of large year classes now moving into the plus group of the population. Such factors can increase the likelihood of retrospective patterns. Dr.Terceiro acknowledged these concerns and noted that noisy indices might cause problems in future years. The model also includes a dome shaped selectivity pattern for the fishery. This creates a large “cryptic" biomass. Consideration of age-based natural mortality rates might be necessary in future benchmarks for this species.

Additional questions from the A O P inquired about the potential utility of methods to aggregate sev- eral indices to detect relative abundance and trend. Dr.Terceiro noted that various GLM approaches had been explored but previous reviewers expressed concerns about over-smoothing of abundance estimates outside of the assessment model. In theory, modern models are designed to address competing signals in the composite likelihood function, but they do not address the spatial arrangement of the indices or their covariance. Recent recruitment indices have been low, but attempts to estimate a parametric stock recruitment relationship have not been successful. The low values in recent years do not seem sufficient to support a change in the stanza of years used for stock and catch projections. Moreover, in view of Dr.Terceiro’s responsibilities for summer flounder at the June M T A , the analyses to justify such a change are unlikely to be completed.

No R T A are currently planned for scup but likely topics for consideration include the aforementioned topics of age-specific M and aggregation of young of the year indices as well as concerns about discard estimates in the earlier decades of the assessment. The model currently starts in 1963, but estimates of age structure only began in 1984. There appears to be sufficient contrast in recent survey indices such that the earlier years of the time series could be dropped. The tradeoff between contrast in the surveys and uncertainty in the discards and age composition of earlier years may justify truncation at an R T A . The A O P unanimously endorsed a Level 2 review for Scup.

Recommendation: Level 2 (Expedited Review)

The currently approved stock assessment model for summer flounder is a 2021 Management Track assessment (M T A ) with data through 2019, that builds on the 2018 S A W 66 benchmark assessment. This is an A S A P model with four fishery fleets, three selectivity periods, and a variety of federal, state, and academic fishery surveys. Results of the 2021 M T A indicated the stock is not overfished, and overfishing is not occurring.

New sources of information considered for the 2023 M T A include an update of fishery and survey catches for 2020; to ; 2022. C A M S will be queried for commercial landings and discards for these same years; Mark Terceiro noted that preliminary comparisons of S B R M and C A M S for 2018; to ; 2021 were within 10%, likely due to differences in stratification, while differences in landings were trivial. Revision of the N E F S C trawl survey indices for the Bigelow years (2009; to ; 2020) to include “by-tow” swept area calculations are also proposed. Model configuration changes that are proposed as part of this M T A include changes to survey input C V s and adjustments to input catch E S S ; each expected to have minor changes on assessment results, but result in improved model diagnostics. Dr.Tercerio noted that if time allows (depending on exactly when data are available), he will experiment with splitting the terminal fishery selectivity blocks from 2008; to ; 2022 to 2008; to ; 2015 and 2016; to ; 2022 to determine impacts on model performance.

Consistent with past summer flounder assessments, B R P s will be derived from projections that in- clude recruitment estimates that use the entire time series (1982; to ; 2022), while O F L projections will extend the S S C -recommended low-recruitment time series that started in 2011 (2011; to ; 2022). For 2024; to ; 2025 O F L projections, Dr.Terceiro will assume catch in 2023 as the final A B C (15,023 m t ), and follow M A F M C risk policy for A B C (e.g., O F L C V =60%).

Dr.Terceiro is proposing as a backup assessment plan, should one be necessary, of either recent trends in all normalized survey indices (e.g., the S S C data update procedure) or I smooth using the N E F S C Bigelow spring and fall indices.

The A O P further inquired about the source(s) of differences between S B R M and C A M S , to which Dr.Terceiro noted that the differences were not consistent in one direction and that further diagnosis of specific differences will require a line-by-line, stratum-by-stratum examination of discards; he noted there may not be sufficient time to perform that analysis. The A O P noted that highlighting differences to the review panel could be helpful. The A O P supported extending the low recruitment timeseries for O F L projections but inquired as to whether there was a contingency plan if one of the new recruitment estimates (2020; to ; 2022) was anomalously high to which Dr.Terceiro indicated that early signs suggest there is a low risk of this happening, but if it should, he is likely to explore an alternate projection run with the anomalous recruitment(s) removed. The A O P also discussed whether exploration of revisions to historical selectivity blocks would elevate the assessment to Level 3, but the A O P felt comfortable that should time allow for this exploration, given the other modest changes proposed for this assessment, there would be adequate review time under a Level 2 review; moreover, it appears as though past practice has been to maintain Level 2 assessments for this type of proposed change (e.g. scup).

The A O P supported a Level 2 assessment review for summer flounder. Justification for this rec- ommendation included the notion that the time available for a Level 2 review is sufficient to address all proposed changes.