Canadian Science Advisory Secretariat (CSAS) Research Document 2025/048 Newfoundland and Labrador Region September 2025 Assessment of the Northern Cod (Gadus morhua) Stock in NAFO Divisions 2J3KL in 2024 Paul M. Regular, Katherine Skanes, Rajeev Kumar, Rick M. Rideout, Emilie Novaczek, Fatemeh Hatefi, Robert S. Gregory, Mariano Koen-Alonso, and Karen S. Dwyer Fisheries and Oceans Canada Northwest Atlantic Fisheries Center St. John’s, NL A1C 5X1 Foreword This series documents the scientific basis for the evaluation of aquatic resources and ecosystems in Canada. As such, it addresses the issues of the day in the time frames required and the documents it contains are not intended as definitive statements on the subjects addressed but rather as progress reports on ongoing investigations. Published by: Fisheries and Oceans Canada Canadian Science Advisory Secretariat 200 Kent Street Ottawa ON K1A 0E6 http://www.dfo-mpo.gc.ca/csas-sccs/ DFO.CSAS-SCAS.MPO@dfo-mpo.gc.ca © His Majesty the King in Right of Canada, as represented by the Minister of the Department of Fisheries and Oceans, 2025 This report is published under the Open Government Licence - Canada ISSN 1919-5044 ISBN 978-0-660-78262-1 Cat. No. Fs70-5/2025-048E-PDF Correct citation for this publication: Regular, P.M., Skanes, K., Kumar, R., Rideout, R.M., Novaczek, E., Hatefi, F., Gregory, R.S., Koen-Alonso, M., and Dwyer, K.S. 2025. Assessment of the Northern Cod (Gadus morhua) Stock in NAFO Divisions 2J3KL in 2024. DFO Can. Sci. Advis. Sec. Res. Doc. 2025/048. v + 126 p. Aussi disponible en français: Regular, P.M., Skanes, K., Kumar, R., Rideout, R.M., Novaczek, E., Hatefi, F., Gregory, R.S., Koen-Alonso, M. et Dwyer, K.S. 2025. Évaluation du stock de morue du Nord (Gadus morhua) dans les divisions 2J3KL de NAFO en 2024. Secr. can. des avis sci. du MPO. Doc. de rech. 2025/048. v + 130 p. http://www.dfo-mpo.gc.ca/csas-sccs/ mailto:DFO.CSAS-SCAS.MPO@dfo-mpo.gc.ca https://open.canada.ca/en/open-government-licence-canada#:%7E:text=Open%20Government%20Licence%20-%20Canada%201%20Using%20Information,Governing%20Law%20...%208%20Definitions%20...%20More%20items iii TABLE OF CONTENTS ABSTRACT ................................................................................................................................... v INTRODUCTION .......................................................................................................................... 1 REPORTED LANDINGS OF COD ................................................................................................ 2 COMMERCIAL FISHERIES ..................................................................................................... 2 RECREATIONAL FISHERIES ................................................................................................. 3 CATCH AT AGE ....................................................................................................................... 3 Historic Pattern in Catch at Age ............................................................................................ 3 Catch at Age from the Last Five Years ................................................................................. 4 TAGGING ..................................................................................................................................... 4 TAGGING BASED ESTIMATES OF RECREATIONAL CATCH .............................................. 4 Results .................................................................................................................................. 5 DFO RESEARCH VESSEL BOTTOM-TRAWL SURVEYS OF 2J3KL ......................................... 6 SURVEY DESIGN .................................................................................................................... 6 FALL SURVEYS....................................................................................................................... 7 Fall Abundance and Biomass Indices ................................................................................... 7 Fall Survey Mean Catch at Age ............................................................................................ 7 Fall Survey Catch Distribution ............................................................................................... 8 DIVISION 2H ............................................................................................................................ 8 DIVISION 3L SPRING .............................................................................................................. 8 MATURITY ............................................................................................................................... 8 GROWTH ................................................................................................................................. 9 TRENDS IN MEAN RELATIVE CONDITION BY DIVISION AND SURVEY ............................ 9 INDEX OF STARVATION-INDUCED MORTALITY ............................................................... 10 EXTENDED NORTHERN COD ASSESSMENT MODEL (XTENCAM) ...................................... 10 DATA INPUTS FOR XTENCAM ............................................................................................ 11 Fall DFO Research Vessel (RV) survey .............................................................................. 11 Inshore Acoustic Biomass Estimates and DFO RV Survey Catchability Parameter (q) ..... 12 Sentinel Surveys ................................................................................................................. 12 Juvenile Surveys ................................................................................................................. 13 Catch and Catch Bounds .................................................................................................... 13 Tagging ............................................................................................................................... 14 Capelin as a Predictor of Natural Mortality .......................................................................... 15 RESULTS FROM XTENCAM ................................................................................................. 15 Effect of Capelin Availability ................................................................................................ 16 Stock-recruitment Relationship ........................................................................................... 16 Reference Points ................................................................................................................. 16 Stock Size and Mortality Rates ........................................................................................... 16 Retrospective Analysis ........................................................................................................ 17 iv Projections .......................................................................................................................... 17 Conclusions ......................................................................................................................... 18 ACKNOWLEDGEMENTS ........................................................................................................... 18 REFERENCES CITED ................................................................................................................ 18 TABLES ...................................................................................................................................... 23 FIGURES .................................................................................................................................... 89 APPENDIX A: ASSESSMENT MODEL DIAGNOSTICS .......................................................... 113 v ABSTRACT The Northern Cod (Gadus morhua) stock inhabits waters off southern Labrador and eastern Newfoundland eastward to the edge of the continental shelf in the Northwest Atlantic Fisheries Organization (NAFO) Divisions (Divs.) 2J3KL. This stock was assessed through a Regional Peer Review Process (RPR) conducted March 18–21, 2024, in St. John’s, Newfoundland and Labrador (NL) to address terms of reference provided by Fisheries Management. The general objective of this meeting was to provide advice on stock status and trends, to consider the effects of environmental conditions, and provide short- to medium-term projections of the stock. This is the first application of the updated assessment model and new Limit Reference Point (LRP) defined at the October 2023 Framework process. The assessment model was extended, through the inclusion of additional data, back to 1954 (previously starting at 1983). The longer term perspective provided by this model informed a revision of the LRP. While estimates of stock size since 1983 were similar, the previous LRP was roughly 40% higher. Following this revision, the stock was determined to be out of the Critical Zone since 2016. Under the current assessment, the 2024 Spawning Stock Biomass (SSB) is estimated to be 1.2 times the LRP, with a 95% confidence interval (CI) of 0.7 to 2.1. There is an estimated 22% probability that the stock is in the Critical Zone. An upper stock reference has not been defined, however given the proximity to the LRP the stock is considered to be in the Cautious Zone. Following a period of growth from 2010 to 2016, SSB has remained unchanged. SSB in 2024 was estimated at 342 Kt (95% CI = 246–475 Kt). Estimated numbers of recruits (age 0) have remained unchanged since the mid-2010s, corresponding to about 80% of pre-collapse (1954– 90) levels. Natural mortality (M, ages 5+) has varied between 0.29 and 0.90 since 1995 (mean = 0.47) and in 2023 was above average [0.59 (95% CI = 0.32–1.12)]. Fishing mortality (F, ages 5+) has been below 0.05 since 2004 and in 2023 was 0.02 (95% CI = 0.01–0.03). Mortality rates of Northern Cod increase when there are insufficient prey in the system to support the population. Capelin are a key prey species for Northern Cod, and their collapse, coupled with sustained low productivity, has impeded the recovery and growth of the cod stock. Capelin are expected to remain around 10% of their pre-collapse levels, which limits prospects for cod stock growth. Projections of SSB to 2027 show that if total removals for 2023 are set to zero, the probability of the stock declining into the Critical Zone is 42%. However, if removals are doubled to 27,034 t (two times the model-estimated total of 13,517 t), this probability increases to 52%. With total removal levels examined here (0 to 27,033 t) the risk of stock decline from 2024 to 2027 is moderately high to high, ranging from 62% to 76%. There is no level of removals that gives a neutral to high (≥50%) probability of stock growth. Under current ecosystem conditions and total removals, the stock has not grown since 2016 and short-term prospects for stock growth are limited even under zero removals. Although primarily driven by ecosystem factors and natural mortality, increases in removals further add to the risk of stock decline into the Critical Zone. 1 INTRODUCTION This document gives an account of the 2024 assessment of the Northern cod (Gadus morhua) stock that inhabits waters off southern Labrador and eastern Newfoundland eastward to the edge of the continental shelf in Northwest Atlantic Fisheries Organization (NAFO) Divisions (Divs.) 2J3KL (Figure 1). The current evaluation was conducted through a Regional Peer Review Process (RPR) on March 18–21, 2024, in St. John’s, Newfoundland and Labrador (NL) using available data up to the end of 2023. A Science Advisory Report (SAR) for the 2J3KL stock from this meeting will be produced. Details of previous assessments and stock updates for Northern Cod to 2016 are reported elsewhere (Lilly et al. 2006; Brattey et al. 2008a, 2009, 2010, 2018). The last stock assessment was in 2021 (DFO 2022). Since 2016, an integrated state-space, catch-censored population dynamics model developed specifically for Northern Cod (NCAM) has been used to assess the stock (Cadigan 2015, 2016b). In 2021, this model indicated that the Spawning Stock Biomass (SSB) was at 52% of the Limit Reference Point (LRP or Blim), in the Critical Zone of Fisheries and Ocean’s Canada’s (DFO’s) Precautionary Approach (PA) Framework (DFO 2009, 2022). This LRP was defined as the average SSB from the 1980s (DFO 2010; DFO 2019a). Projections under six catch scenarios, with removals ranging from zero to 1.3 times (15,360 t) the model estimated catch for 2020 (11,816 t), showed that the probability that SSB will reach the LRP by 2022 is less than 0.02. Partial coverage of the stock area by the standard offshore Research Vessel (RV) survey program prevented a typical update of the assessment in 2022 and 2023 since the leading indicator of stock size could not be calculated. Mechanical issues with ageing research vessels led to partial coverage of the stock area in 2021, and targeted sampling for comparative fishing took precedence over standard survey sampling in 2022 (DFO 2024a). Comparative fishing was prioritized, as data from that program were necessary to quantify differences in catchability between the outgoing vessels Canadian Coast Guard Ship (CCGS) Alfred Needler and CCGS Teleost with the new CCGS John Cabot and CCGS Capt. Jacques Cartier. Data from this program were essential for ensuring the continuity of the RV survey time-series. The absence of assessments, however, provided time to organize and prepare an assessment framework meeting, which was held October 16–20, 2023, in St. John’s, NL (DFO 2024b). The general objective of this meeting was to advance the assessment by addressing several research recommendations from previous assessments of Northern Cod. A series of extensions to NCAM were reviewed and accepted at the framework to produce a revised model that is hereafter referred to as xteNCAM. Nearly 30 years of historic data were added to the revised model, extending the series back to 1954. Historic landings (Schijns et al. 2021) and tagging (Taggart et al. 1995) data were particularly useful for broadening the scope of the assessment model. Additionally, juvenile survey data from two inshore monitoring programs (Fleming survey, Lewis et al. 2022; Newman Sound survey, Gregory et al. 2019) were integrated into the model, which enabled the implementation of a stock-recruitment relationship. An index of Capelin, a key prey species for cod and driver of cod productivity (Koen-Alonso et al. 2021; Regular et al. 2022), was also integrated into the model to predict changes in rates of natural mortality of cod. These revisions provided an opportunity to revisit the LRP for Northern Cod. In alignment with PA guidelines (DFO 2009, 2023a), the new LRP was established at 40% BMSY (SSB that produces the maximum sustainable yield [MSY]). Under this new framework, SSB was estimated at 1.16 times the LRP in 2021, with a 29% probability of being below the LRP (DFO 2024b). This contrasts with the 2021 assessment, which estimated the stock to be 52% of the previous LRP (DFO 2022). While this changed the 2 2021 stock status from Critical to Cautious, it is important to note that both NCAM and xteNCAM provide comparable estimates of SSB in 2021 (approximately 400 kt). The difference in status stems from the change in the LRP, as the estimate of 40% BMSY is roughly 40% lower than the average SSB from the 1980s. In the current assessment, xteNCAM was used to assess the status of the stock with updated input data. In addition to the assessment model results, several other sources of information were reviewed at the assessment (e.g., physical and biological oceanography, information on Capelin, predators, prey, inshore pre-recruit surveys, tagging, industry surveys, etc.). REPORTED LANDINGS OF COD Reported landings from this stock from the 1950s until 2015 are described in detail in previous documents (Lilly et al. 2006; Brattey et al. 2018). Reported landings for 2016–23 are added to the time-series (Table 1; Figures 2–3). Fixed gear landings (from 1975) are also updated to 2023 (Table 2; Figure 4) and show that most of the catch throughout 2006–23 was taken by gillnets. Finally, catch estimates by fishery type (Recreational, Sentinel, and Commercial) are shown in Figure 5. Under the Privacy Act, the Department can no longer provide landings and catch information for a specific fishery when that fishery has fewer than five fishing enterprises, five fishing vessels, and/or five buyers participating in a fishery. As such, aggregate catch statistics are presented to protect the privacy and economic interests of participants in the fishery. COMMERCIAL FISHERIES The directed inshore stewardship fishery and a recreational fishery for cod were re-opened in 2006. Note that the management year extends from April 1 to March 31 of the following year (since 2000), but catch statistics are reported in calendar year as there have been no significant landings from this stock from January to March. There has been no formal Total Allowable Catch (TAC) set since 2006; instead, commercial fishers are allocated a fixed annual allowance per licence holder. Beginning in 2016, the 2J3KL Northern Cod stewardship fishery has been managed using variable weekly catch limits, gear restrictions, seasons, and in 2018 a maximum authorized harvest amount (MAH; 9,500 t) was introduced. The MAH for the stewardship fishery has been 12,999 t since 2021. Provisional total reported landings in 2023 were 12,998 t. Recreational landings are regulated by number of days and number of fish per person (five per person, for a maximum of 15 fish per boat). Currently there is no requirement to report recreational landings. Some effort has gone in to obtaining estimates of recreational landings (see Brattey et al. 2018); currently tagging is used to determine a general approximation. There are known to be discards of cod in the shrimp fishery that operates offshore in Divs. 2J3K. This has been investigated in the past and shows that although the weight of removals is small, the number of small fish removed can be large. Monthly reported landings of cod from NAFO Divs. 2J3KL from 1983–2023 were also compiled from available catch statistics databases for Canada-NL, Canada-Maritimes, and non-Canadian catches (Table 3). Since 2006, most of the landings in the commercial fishery are taken from August to November; catches in the remaining months came mostly from the sentinel fishery and bycatch. These monthly values were used in the assessment model (see below) in conjunction with tag recapture information. These values were used to estimate the fraction of harvesting that took place in any year before tagged fish were released in that year. 3 There remains no mechanism to estimate high-grading (i.e., discarding of smaller fish) or discard rates in the historical catch statistics. Further, recreational fishery landings are not directly measured (see below). The amount of discarding in recreational fisheries is unknown. RECREATIONAL FISHERIES The recreational fishery operates inshore in NL from June to October. Over the past 17 years (2006–23), estimates of recreational fishery landings have only been available for 2006, 2008, 2011, and 2012. These estimates were obtained using a combination of methods, including phone surveys and surveillance. Analyses of tag return information from commercial versus recreational fishers suggests that removals from recreational fisheries have been large and, in some years, higher than reported by Fisheries Officers (see Brattey et al. 2010), averaging about 30% of reported fishery removals during 1997–2018 (see tagging section). In the past three years, this number has been slightly lower (25%) due to increases in the commercial harvest. Estimates from this method are annually quite variable (Brattey et al. 2009, 2010) and are not used in formal compilation of landings statistics. Nevertheless, results suggest that recreational landings are large and that total removals are higher than reported landings by an unknown and varying annual amount. Some sampling of the recreational catch takes place at sea and dockside by Conservation and Protection officers. In the past, there have been signs of discarding or high-grading, but it is difficult to determine whether this is currently ongoing. CATCH AT AGE The age composition and mean length-at-age of the cod landings were initially calculated by gear, unit area, and quarter as described in Gavaris and Gavaris (1983). Sampling of cod lengths and ages (otoliths) from the various sectors of the commercial fishery has been extensive from 2011–23. General sampling objectives for the fishery were to obtain a representative length frequency from each commercial area fished (i.e., NAFO statistical unit area such as 2Jm, 3Kd, 3La, etc.) for each gear type and fishing period. Fishing was generally carried out from July to October and three gear types (gillnet, hand-line, and longline) account for most (greater than 95%) of the landings (Table 2). Sampling aims to measure a minimum of 2,000 fish from each unit area, gear type, and season. A minimum sample of 200 fish per catch is measured, or the entire catch if less than 200 fish are available. In addition, the objective is to collect five otoliths per centimeter (cm) length group, per unit area, gear type, and fishing period. Additional sampling of sentinel fishery catches is conducted by sentinel fishers following a separate protocol described elsewhere (Maddock Parsons 2014; Mello and Simpson 2023). Historic Pattern in Catch at Age The time-series of catch-at-age from the fishery for Northern Cod (inshore and offshore combined) extends from 1962–2023 (Table 4). For the post-moratorium period, almost all of the catch has come from the stewardship fishery which is dominated by gillnets. Descriptions of historical trends in catch-at-age can be found in previous assessment reports (Brattey et al., 2018). Generally, the age composition of catches from the beginning of the last five years show a shift in central tendency from ages of approximately 3–8 in 1980 and 1990 to ages of approximately 5–10 in 2010 and 2020 (Figure 6). 4 Catch at Age from the Last Five Years In the past five years, most of the catch consists of ages 6–11 (Table 4; Figure 6) which is typical of a fishery dominated by gillnets. In assembling the catch-at-age for the recent period (2011–23), there were no major discrepancies in the sum of the products (estimated catch numbers-at-age times weight-at-age) relative to reported landings. The ratio of the sum of products to reported landings was close to 1 in each year. TAGGING For Northern Cod, an extensive time-series of mark-recapture information is available. The tagging data are in two parts: 1. The earlier (i.e., pre-1997) tagging data are summarized in Taggart et al. (1995) and were analyzed by Myers et al. (1996) and Myers et al. (1997). Disc tags were the main tag type used in the tagging program between 1962–66 and they continued to be used extensively between 1978–91, after which they were phased out and T-bar (floy) tags became the main type (Table 5). 2. The current tagging program began in 1997 and are reported in Brattey and Healey (2007) and Brattey et al. (2018). The use of high-reward and low-reward floy tags in the current program allows for the estimation of reporting rates (Konrad et al. 2016). In 2021, 4,143 tags were released, with a further 3,165 deployed in 2022 and 3,972 deployed in 2023 (Table 5). Twenty percent of these tags were “high value” ($100 reward), with the remaining eighty percent as “low value” ($10 reward between 1997–2022; $25 reward since 2023). Tagging occurred in Divs. 2J3KL in 2021–22, and in 3KL in 2023 (Figure 8). Recaptured tags received in 2021–23 were caught in NAFO Divs. 2J, 3K, 3L, 3P, and 4R, with the vast majority (88%) caught in 3KL, coinciding with the location of inshore fishing effort for cod (Figure 8). TAGGING BASED ESTIMATES OF RECREATIONAL CATCH Tag returns from the commercial and recreational fishery provide an annual estimate of recreational catch. Since the commercial landings are reported annually, and the ratio of tags reported by commercial fishers and recreational fishers is also known, a simple ratio estimator can be used to calculate recreational catch using the following equation: 𝐶𝐶rec,𝑦𝑦 𝐶𝐶comm,𝑦𝑦 = 𝑅𝑅rec,𝑦𝑦 𝑅𝑅comm,𝑦𝑦 , where 𝐶𝐶𝑟𝑟𝑟𝑟𝑟𝑟,𝑦𝑦 is the catch/total landings from recreational (rec) and commercial (comm) fisheries in year 𝑦𝑦, and 𝑅𝑅 is the number of tags returned by recreational and commercial fishers in year 𝑦𝑦. 𝐶𝐶comm,𝑦𝑦 is reported annually with no error reported. The total numbers of tags associated with each fishery is a function of the total number of tags reported, which is the number of high reward tags reported (𝑁𝑁(𝐻𝐻)), and the number of low reward tags reported (𝑁𝑁(𝑆𝑆)) adjusted by the fishery-specific (type) reporting rate (𝜆𝜆) for low reward tags. That is, 𝐸𝐸�𝑅𝑅type,𝑦𝑦� = 𝑁𝑁(𝐻𝐻)type,𝑦𝑦 + 𝑁𝑁(𝑆𝑆)type,𝑦𝑦 𝜆𝜆type,𝑦𝑦 . Reporting rates (𝜆𝜆) are calculated with a random-walk model developed by Konrad et al. (2016), which are region (2J3KL versus 3Ps) and year-specific for the commercial fishery and constant over the years for the recreational fishery. Assuming that the proportion of tags removed from 5 the two fisheries is estimated from the number of tags (corrected) removed in each fishery, the number of tags (corrected) returned by each fishery can be treated as a random variable where, 𝑅𝑅rec,𝑦𝑦 ∼ Bin�𝐸𝐸�𝑅𝑅rec,𝑦𝑦�, 𝐸𝐸�𝑅𝑅rec,𝑦𝑦� 𝐸𝐸�𝑅𝑅rec,𝑦𝑦� + 𝐸𝐸�𝑅𝑅comm,𝑦𝑦� � , and 𝑅𝑅comm,𝑦𝑦 = �𝐸𝐸�𝑅𝑅rec,𝑦𝑦� + 𝐸𝐸�𝑅𝑅comm,𝑦𝑦�� − 𝑅𝑅rec,𝑦𝑦. To incorporate the random variables, random normal variates were drawn for annual reporting rates and random binomial variates were drawn for 𝑅𝑅rec,𝑦𝑦 to calculate the recreational catch (𝐶𝐶rec,𝑦𝑦). This process was repeated 1,000 times to obtain a sample of 𝐶𝐶rec,𝑦𝑦, and a mean and standard deviation were extracted from those 1,000 samples. The approach of using tag return data from commercial and recreational fisheries to estimate total recreational landings makes a few critical assumptions. The first, and likely most important, is that the general behavior of the two fisheries is similar. In other words, if the commercial fishery requires 1,000 tonnes of catch to return 10 tags, then 1,000 tonnes of recreational catch would be needed to return the same number of tags. Reporting rate differences in the two fisheries are taken into account directly, so any differences in reporting rate by different fisher types can be addressed explicitly. It is differences in the fish themselves and whether the population of fish exposed to the commercial fishery and the recreational fishery is more or less the same, and most importantly, have a similar distribution of tags available to the two fisheries that are most important. For the 2J3KL fishery, the recreational and commercial catch are both summer to early fall and inshore fisheries, so are most likely pursuing similar fish. The gears used are different; 5” gillnets are the main gear used in the commercial fishery, while the recreational fishery is conducted with handlines. Gillnets are size-selective, with fish ranging in lengths between 60–80 cm being most vulnerable. In theory, handlines are less size-selective; however, hook size will influence the size of fish captured, and fisher behavior will lead to some size selectivity (recreational fishers are expected to move to other grounds if they catch a few small fish). The tagging program only tags fish greater than 44 cm fork length (FL), so only fish of this size or greater are sampled by tag returns. If the proportion of fish less than 45 cm FL in the commercial and recreation catch differs substantially and makes up a non-trivial component of total landings, the estimate of recreational catch based on tag returns alone may be biased. Another assumption is that the tagged fish are equally available to each fishery. Tagging efforts are distributed throughout 2J3KL but are concentrated in certain areas, and the spatial distribution of tags deployed in each year is not entirely consistent. If many tags are deployed in areas of intense commercial fishing but modest recreational fishing, estimates of recreational catch will be biased low. Timing of tagging may also lead to biases; tagging takes place from July to October in 2J3KL, and the recreational fishery opens before the Stewardship Fishery in 2J3KL. If a large number of tags are deployed in August, for example, a time when the recreational fishery is well underway (and tending to taper off) but the stewardship fishery has yet to open, then the commercial fishery may have more tags available for capture, again leading to a recreational catch estimate that is biased low. Results Annual tag returns are presented in Table 5. The number of tags returned in recent years is low for the time series, likely a function of lower tagging efforts in 2017–20 and declining reporting rates (Figure 7). Estimated reporting rate from the commercial fishery has declined consistently from a high of 0.85 in 2001 to an average of 0.46 in recent years (2019–23). The reporting rate from the 6 recreational fishery is estimated as a constant for the time series, and also sits at 0.46 (Figure 9). These return numbers may increase slightly as fishers continue to send in the tags they captured in the past year. To get enough returns for this method, tagging numbers must be kept relatively high and reporting should be encouraged among participants of both fisheries. In addition to increased tagging efforts since 2020, the value of low-reward tags was increased from $10 to $25, with the first deployment of the new tags in 2023. The ratio of tags returned by recreational fishers over commercial fishers in 2J3KL has varied considerably over the time series. Some years are very low, including years with no recreational fishery (2004–05), while in other years (e.g., 2014) almost as many tags were returned by recreational fishers as commercial fishers. The ratio of recreational:commercial tag returns was generally low and stable in the early years (1997–2006) with a mean of 0.14 (range = 0–0.32), high and variable between 2007–2015 (mean = 0.54, range = 0.25–0.85), and has returned to a low and stable period since 2016 (mean = 0.20, range = 0.07–0.26; Figure 10). Estimated recreational catch in 2J3KL has ranged in the hundreds of tonnes to a few thousand tonnes in years with a recreational fishery (Figure 11). Recreational catch estimates range from 0.14–3.9 kt, with no recreational catch in 1997, 2004, or 2005. Generally, recreational catch varied between 0.0–1.5 kt (mean = 0.63 kt) in the early period (1997–2006), and have increased since 2007, varying between 0.7–3.9 kt (mean = 2.3 kt). In 2023, and the estimate of recreational catch was 2.2 ± 0.63 kt. In the absence of reported recreational catch, this method provides an independent estimate of the general level of removals by the recreational fishery. In some years, total fishing removals can be much higher than the reported commercial landings due to substantial recreational catch. This estimate is not directly integrated into the assessment model, however, to date the catch bounds applied within xtenCAM are sufficient to account for the levels of recreational catch estimated from tag returns. DFO RESEARCH VESSEL BOTTOM-TRAWL SURVEYS OF 2J3KL Research bottom trawl surveys have been conducted by Canada during the fall in NAFO Divs. 2J, 3K, and 3L since 1977, 1978, and 1981, respectively, and the information from these surveys was updated to 2023. Spring surveys have been conducted by Canada in Div. 3L during the years 1971–82 and 1985–23. In 2020, due to Covid-19, there was no spring survey. Both the spring and fall surveys were interrupted in 2021 and 2022 by necessary comparative fishing sampling efforts aimed at quantifying differences in catchability between the outgoing and new CCGS vessels (DFO 2024a); however, there was a fall survey of Divs. 2J and 3K in 2021. SURVEY DESIGN Details of the stratified random trawl survey design as well as changes in gear and vessels are described in previous documents (Lilly et al. 2005, 2006; Brattey et al. 2008a; Wheeland and Rideout 2023). Additional information on surveys conducted by DFO since the introduction of the Campelen trawl in 1995 is provided by Rideout and Ings (2021) and references therein. Details of survey performance statistics, timing, and spatial coverage are summarized elsewhere (see Power et al. 2016 and references therein). Note that all the survey catch rate information presented below for the 1983–94 period are in Campelen equivalent units, while values for 1995 onwards are based on actual Campelen catches. 7 The fall 2023 survey of Divs. 2J3KL took place earlier than average, particularly in Div. 2J (Figure 12). It is uncertain if the earlier timing resulted in a different underlying cod distribution than during the fall survey in previous years. FALL SURVEYS Fall Abundance and Biomass Indices Indices of cod abundance and biomass are based on the stratum-area weighted arithmetic mean catch (numbers and weights) per standard tow using an R based version of STRAP (Stratified Analysis Programs; Smith and Somerton 1981), called Rstrap, with inconsistently (or non) fished strata being omitted from the calculations. There have been some changes over time in the depths covered during the survey; consequently, trends in the indices of abundance and biomass of cod have been monitored for those strata that have been fished most consistently since the start of the surveys. These “offshore index strata” are those in the depth range 100–500 m in Div. 2J and 3K and 55–366 m (30–200 fathoms) in Div. 3L. The inshore strata fished intermittently from 1996 onwards are not included in this index, nor are deep-water strata (greater than 200 fathoms in Div. 3L or greater than 500 m in Div. 2J and 3K). Separate estimates of abundance and biomass by stratum have been calculated for the inshore and deep-water strata (see Brattey et al. 2008a), but coverage in these areas has been poor for several years, and few cod have been observed in the deep-water strata, so these are not updated here. Lilly et al. (2006) provides details on the interpretation of the fall survey data with respect to depth and timing of the survey. The full-time series of fall DFO research vessel survey index values for Divs. 2J, 3K, and 3L begins in 1983 and shows that the abundance and biomass indices have been low since the start of the moratorium in 1992 (Table 6; Figure 13). Both abundance and biomass indices increased since 2011, however there has been no clear trend in either direction. Most of the abundance and biomass (greater than 75%) is found in the northern portion of the stock area (Divs. 2J and 3K; Figure 14). The average abundance and biomass indices from the last three years are approximately 30% of the average during the 1980s. The 2023 survey abundance and biomass index values were 407 million fish and 340 kt, respectively. The decrease in numbers of older fish from the 2017 survey could be due to changes in the timing of cod being inshore later than usual and not fully in the survey area at the time of the RV survey or that some of the older cod were in poor condition and did not make it through the summer to be caught in the RV survey. It is possible that 2017 was a survey effect. Fall Survey Mean Catch at Age Rstrap was used to calculate index values through the entire time series. Survey numbers-at- age were obtained by applying age-length keys (ALKs) to the numbers of fish at length in the samples from indexed strata. A time series of survey catch rates at-age, expressed as mean numbers per tow at-age for Divs. 2J3KL combined, is given in Table 7 and Figure 15. Mean catch per tow was generally high (mostly 50–200 fish per tow) in the 1980s, but declined rapidly to generally less than 10 fish per tow during 1990–93. The age structure also contracted during the collapse period, with few old cod (age 6+) in the survey catches by the early 1990s. The catch rates at-age remained low for more than a decade, but catch rates have been increasing since about 2010. Generally, age structure had been expanding since 2012, with cod spawned in the early 2000s onwards surviving through to older ages in recent surveys (Table 7). 8 Fall Survey Catch Distribution An extensive time series of age-aggregated and disaggregated distribution plots of fall survey catch numbers and catch weights per tow are available but, for brevity, only a subset of these are shown here to illustrate key changes in cod distribution in the recent period. A detailed history and description of changes in the distribution of cod at the time of the surveys to 2009 is given elsewhere (Shelton et al. 1996; Lilly et al. 2006; Brattey et al. 2010). Patterns of distribution in the most recent four years indicate that both number and weight per tow are generally spread widely throughout Div. 2J, 3K, and northern 3L (Figure 16). In some years, there are large tows of fish on the edge of the continental shelf. In 2020 and 2021, there are some large tows of cod (greater than 500 fish per tow) in Div. 2J, however there were fewer large tows of cod in 2023. Additionally, the distribution in 2020 and 2023 shows more tows with cod in southern 3L than in the recent past (Figure 16). Distribution of cod aged 2–4 indicate that cod at ages 2–4 were generally found further inshore in 2021 (Divs. 2J3K only) than in 2020, however they are distributed fairly evenly over the survey area in 2023 (Figure 17). There were few large (less than 500 cod per tow) catches of cod aged 2-4 in 2023. DIVISION 2H Trends in the abundance and biomass indices and length frequencies were examined from bordering Div. 2GH to the north. Div. 2G has not been surveyed since 1999, and Div. 2H has been covered intermittently from 1999–10. After 2010, Division 2H has been covered as part of regular multispecies surveys. Cod in Div. 2H showed a large increase in both abundance and biomass indices in 2015, which was likely spillover from Div. 2J (Figure 18). Distribution plots show that these fish were found close to the northern part of Div. 2J (not shown). DIVISION 3L SPRING The spring survey covers only part of the stock area, but this time of year may be important in showing certain trends, such as condition related to starvation (Regular et al. 2022). As expected, there is more variation in the estimates of abundance and biomass in this area during spring (fish may be moving into and out of the survey area) but generally there is a trend of increased numbers in fish since 2013 (Figure 19). MATURITY Annual estimates of age at 50% maturity (A50) and proportions mature at-age for females from the 2J3KL cod stock, collected during annual fall DFO research bottom trawl surveys, were calculated as described by Morgan and Hoenig (1997). A cohort-specific binomial logistic regression model was used to estimate the proportion mature as a function of age and these estimated proportions (Table 8) are used in the calculation of spawning stock biomass. The estimated age at 50% maturity (A50) is used as a metric for monitoring changes in age at maturity. A50 was generally between 6.0 and 7.0 among cohorts produced in the late 1950s and around 6.0 among those produced during the late 1960s to the early 1980s but declined thereafter (Figure 20). Age at maturity has remained low but variable (4.8–5.7) for the 1990– 2016 cohorts, with no clear trend. Estimates for the last cohort (2018) are often more uncertain because only younger ages (1-5 years) are available to estimate A50. The estimate of A50 for the 2010 cohort (5.8) is unusual and is the highest observed in the recent period; the confidence intervals for the fit to this recent cohort are not large and close inspection of the raw data indicate consistently lower observed proportions mature at each age. Estimates of A50 for the 1990 cohort onwards from the previous stock assessment in 2021 are overlaid on the 2024 9 assessment results (Figure 20). This comparison shows that the addition of one more year of data resulted in very minor retrospective changes in A50 among 2013–16 cohorts. Males show a similar trend in A50 over time (data not shown) but tend to mature about one year earlier than females. GROWTH The lengths-at-age and weights-at-age of cod sampled during the fall surveys were calculated following equations in Morgan and Hoenig (1997) and are shown in Figures 21 and 22. There was a strong decline in length-at-age in Div. 2J and Div. 3K from the late 1970s to the early 1990s followed by an increase in length-at-age, while there was little or no decline in Divs. 3L over that period. Mean length-at-age was mainly above average in 2011–14 in all Divs. However, since then, mean length at-age was below average for almost every age in all divisions. Weight-at-age also showed a steep decline in Divs. 2J and 3K during the same period that lengths were declining and as with length-at-age there was less of a trend in Divs. 3L. Mean weight-at-age was above average in 2011–14 in all divisions and near average in 2015. Since then, mean weight-at-age was below average for most ages in all Divs. Annual variation in mean weight-at-age for Div. 2J3KL combined was examined over ages 3–7 by analyzing deviation from the average as a proportion over the time series for each age. The average mean weight-at-age from 1981–2023 was calculated for each age (Figure 23). Deviation was calculated for each age in each year by subtracting the mean for the age for the time series from the annual observation for that age and then dividing this by the mean for that age. Mean weight-at-age decreased from the beginning of the time series to the early 1990s. It increased to well above average by 1997. From 1997 to 2015, mean weight-at-age fluctuated but remained at or above average. Mean weight-at-age from 2011–13 was among the highest in the time series, but this was followed by a steady decline to below average since then. Beginning-of-year and mid-year weights-at-age were also generated for the assessment, where beginning-of-year weights were used in the calculation of SSB and mid-year-weights were used in the calculation of predicted landings. These weights were generated using a cohort-based Von Bertalanffy growth model developed for Northern Cod (Cadigan 2016a). This model is fit to the aforementioned mean weights-at-age from the RV survey, providing estimates that smooth through the variation in the means, while also filling gaps in the time-series and allowing adjustment of timing (i.e., expected weight at the beginning and middle of the year). Estimates of beginning-of-year and mid-year weights-at-age are presented in Table 9 and Table 10, respectively. TRENDS IN MEAN RELATIVE CONDITION BY DIVISION AND SURVEY Relative condition was calculated by first fitting a log-log regression of gutted weight as a function of length for cod sampled by the spring and fall RV surveys and the Sentinel survey. Individual weights were then divided by weights predicted by the length-weight regression to compute relative condition values. These values were averaged by division and survey to obtain trends in mean relative condition from the spring RV survey (primarily May–June in Div. 3L), summer Sentinel survey (primarily July–September in Divs. 3K3L), and fall RV survey (primarily October–December in 2J3KL) (Figure 24). Cod caught in the spring survey of 3L tend to be in the lowest condition. Condition is relatively stable in the fall across all divisions, varying with little trend around a level slightly above 1. In the fall, cod in Divs. 2J and 3K showed lower relative condition in the early 1990s and all divisions showed a peak in condition around 2011. There are no clear trends in condition in the summer in 2J; however, these averages are based on 10 relatively few samples (<100). Mean relative condition showed local maxima around 1996, 2005, and 2013 and minima around 2010 and 2016 in the 3K and 3L Sentinel series and the 3L spring RV series. The most pronounced decline in mean relative condition was observed during spring in 3L in the early 1990s. Focusing on the most recent years, condition appears to have improved slightly relative to 2016, especially in the fall; however, spring and summer condition remains at relatively low levels. INDEX OF STARVATION-INDUCED MORTALITY An index of starvation-induced mortality was calculated following methods outlined in Regular et al. (2022) using data from the spring RV survey, the summer Sentinel survey, and the fall RV survey. In short, the index is derived from a seasonal model of gutted condition where cod with residual levels below -0.18 are assumed to succumb to starvation (sensu Dutil and Lambert 2000; Casini et al. 2016). Since the actual critical condition level below which cod die from starvation is unknown, this metric is considered an index that is thought to be proportional to rates of mortality from poor condition. Patterns across ages 2–4, 5–7, and 8+ were calculated and are shown in Figure 25. The index exhibited a relatively low rate for both the 5 to 7 and 8+ age groups from 1977 to 1989. However, it showed a rapid increase in the early 1990s, increasing by approximately five times 1989 levels for the 5 to 7 and 8+ age groups. Since the early 1990s, the index has shown variability for the 5 to 7 and 8+ age groups, indicating peaks in starvation-induced mortality around 1998, 2003, 2009, and 2016. The index has declined across all ages since the last spike in 2016. EXTENDED NORTHERN COD ASSESSMENT MODEL (XTENCAM) Since 2016, Northern Cod has been assessed using an integrated state-space catch-censored population dynamics model developed specifically for this stock (Northern Cod assessment model (NCAM); Cadigan 2015, 2016b; DFO 2016a). This model integrates much of the existing information about the productivity of the stock and, as such, is capable of estimating annual rates of natural (M) and fishing (F) mortality. The model also estimates the catch, rather than assuming that reported landings are an exact measure of catch. An interval identifying a likely range of catch (lower and upper bounds) is specified, and this interval was determined during discussions with stakeholders present at the assessment meeting. At a recent framework meeting (DFO 2024b), a series of stepwise changes were implemented to augment the base case NCAM (formulation accepted at the last framework; DFO 2016a), which served as the foundation of the extended assessment model, xteNCAM. Details of the extensions are outlined in Regular et al. 2025a, Regular et al. 2025b. Broadly, the objective of each step was to: 1. improve the fit of the model to catch composition data, 2. minimize conflicts between survey indices, 3. extend the time series back to 1954 (base case NCAM started in 1983), 4. integrate data from juvenile cod coastal monitoring programs, 5. implement a stock-recruitment relationship and calculate per-recruit reference points, 6. estimate baseline levels of natural mortality (M), and 7. quantify the effects of Capelin and seals on the cod stock. Most of these objectives were listed as research recommendations from preceding Canadian Science Advisory Secretariat (CSAS) processes for Northern Cod (e.g., DFO 2019b, 2023b). 11 The current assessment is based on the xteNCAM formulation that estimates a constant baseline M and an age-invariant Capelin effect, as this was the formulation that was selected as the new assessment model (DFO 2024b). DATA INPUTS FOR XTENCAM The xteNCAM model uses the following data sources: • age-disaggregated information from the fall DFO offshore bottom-trawl survey (ages 2–14, 1983–2003, 2005–20, 2023), • inshore combined 2J3KL Sentinel 5½” mesh gillnet index (ages 3–10, 1995–2023), • inshore acoustic estimates of biomass in Smith Sound (1995, 1997–2004, 2006–09), • inshore juvenile indices from the Fleming (1959–64, 1992–97, 2001, 2020–21) and Newman Sound surveys (1996, 1998–2023), • fishery catch age-composition information (ages 2–14, 1962–2023), • partial fishery landings information (1954–2023), • tagging information (release years: 1954–55, 1962–66, 1978–93, 1995–2023), and • indices to calculate Capelin/cod biomass ratio [acoustic estimates of Capelin biomass in 3L (1985–92, 1996, 1999–2005, 2007–15, 2017–19, 2022–23) and total biomass estimates of cod from the fall DFO offshore bottom-trawl survey (1983–2003, 2005–20, 2023)]. Model process errors and observation equations are included for all data sources, and xteNCAM provides estimates of population parameters and variance parameters as well as modeled estimates for each of the data sources described above. Further details, including model likelihood equations for the various data sources, and likelihoods for random effects such as process errors, are also described in detail in Cadigan (2015); Cadigan (2016b) and Regular et al. 2025a and Regular et al. 2025b. Fall DFO Research Vessel (RV) survey The main source of information on trends in the status of Northern Cod is the fall DFO RV survey. This survey is based on a stratified random design, has a long time-series (1983–2023), captures a broad range of ages, and covers a large portion of the stock area. This survey is not intended to provide a direct estimate of stock size but can provide an age-disaggregated index of stock size, which can be used to scale the entire population provided the catchability coefficient (q) can be estimated for each age class in the population. Some cod can be outside the surveyed area, such as in shallow coastal water (e.g., small juveniles), or in deep water off the shelf edge, but provided there is no trend in the proportion of the population outside the area surveyed, the survey can still be used to infer stock trends. The DFO RV survey index time series included the results of the latest (2023) survey and the age range 2–14 years. The survey catchability (q) was constrained to be equal for ages 6 to 14 and the 2004 and 2021 DFO RV survey indices were not used because of problems with survey coverage, and the 2022 survey was not conducted. Age-disaggregated stratified mean number per tow indices for ages 2–14, from 1983–2023, were used in xteNCAM (Table 7). 12 Inshore Acoustic Biomass Estimates and DFO RV Survey Catchability Parameter (q) Following the collapse of the stock in the early 1990s, there is evidence from several sources that q changed for Northern Cod, through a distributional shift of the remnant stock to inshore regions that are not part of the fall DFO RV survey area. A relatively large aggregation of cod was observed in Smith Sound, Trinity Bay (Div. 3L) in the spring of 1995 (Rose et al. 2011). Subsequent surveys in winter and spring gave acoustic biomass estimates ranging from 10,000 to 26,000 t during 1995-2007, but the biomass estimates declined rapidly during 2007-09 to less than 1,000 t. Conventional tagging experiments and acoustic telemetry of cod in Smith Sound and neighboring areas indicated overwintering cod in Smith Sound dispersed around the coast in late spring and summer (Brattey et al. 2008b; Brattey 2013) and returned to Smith Sound in fall. Furthermore, these studies showed that the decline in biomass during 2007–09 was not due to F or M, but was more likely a redistribution of these cod to other areas, including the offshore. The fall DFO RV survey indicated that the biomass of cod in the offshore was increasing, particularly for older cod, when the Smith Sound biomass was declining and further acoustic telemetry and conventional tagging of offshore cod released in 2008 confirmed that the seasonal shoreward migration of offshore cod was taking place, similar to the pattern observed in the pre-moratorium period. Cadigan (2015) concluded that this change in migration was highly plausible and should be included in the stock assessment model through the estimation of a multiplicative q at-age adjustment for the DFO RV survey during 1995–2009. The size of the Smith Sound component of the stock was estimated using published Rose et al. (2011) acoustic biomass estimates (Table 11) with samples (trawl and hand-line) taken during the acoustic survey providing age composition information (Table 12). Sentinel Surveys Sentinel surveys for cod in the inshore were conducted by fishing enterprises operating from many communities in Divs. 2J, 3K, and 3L at various times during summer and fall from 1995 onwards. Lilly et al. (2006) summarized sentinel data up to 2005, and the most recent detailed account of the sentinel program is provided by Kumar et al. (in prep1), who extended the time series to 2023. The primary goal of these surveys was to obtain information on trends in the relative density of cod on traditional inshore fishing grounds during the moratorium. The sentinel surveys were also intended to provide samples that would yield information on various aspects of cod biology in the inshore, including age compositions, size-at-age, condition, maturity, and feeding. An age-disaggregated index of standardized catch rates for cod in the inshore of 2J3KL was calculated from data gathered from sentinel fishing with gillnets (3¼” and 5½” mesh) and line-trawls (Stansbury et al. 2000). The methodology for calculating the index and full details of the sentinel fishery are described elsewhere (Kumar et al. in prep1). The standardized age-disaggregated catch rate index (fish/net, ages 3–10) from 5½” mesh gillnets (fixed) from Div 2J3KL combined was updated, and results from 1995–2023 were used as an index in xteNCAM (Table 13). Sentinel survey q was assumed to follow a 2D age-year autoregressive process as the portion of the stock that makes itself available to the inshore sentinel survey may change through time (Regular et al. 2025a). 1 Kumar, R., et.al. In prep. Sentinel Surveys 1995–2023 – Catch Rates and Biological Information on Atlantic Cod (Gadus morhua) in NAFO Divisions 2J3KL. DFO Can. Sci. Advis. Sec. Res. Doc. 13 Juvenile Surveys There are two long-term surveys designed to monitor the abundance of juvenile cod: 1. the Fleming survey (see Lewis et al. 2022 for details), and 2. the Newman Sound survey (see Gregory et al. 2019 for details). The Fleming survey was conducted along the Northeast Coast of Newfoundland from 1959 to 1964. The survey ceased in 1964, but was reinstated from 1992 to 1997, and again in 2020–21. The survey methods remained consistent over time, sampling the same season (October- September) and sites using a beach seine. Though the time series is fragmented, consistent methods permit direct comparisons of catch rates of cod aged 0–1. The Newman Sound survey uses similar methods as the Fleming survey, except it is isolated to Newman Sound, Bonavista Bay. The area is, however, sampled more intensely, sampling is conducted from July to November, and an index has been produced for nearly every year since 1995. Both indices are used to inform trends in age 0 and 1 cod in the model. The Fleming survey offers an historic perspective on the relative abundance of recruits while the Newman Sound survey provides more recent trends. These data also proved to be useful for informing a stock-recruitment relationship (Regular et al. 2025a). The Newman Sound index has been updated to 2023 and these data are used in xteNCAM along with past indices from both surveys (Table 14). Independent q parameters are estimated for age 0 and 1 cod; however, q is assumed to be time-invariant and equal across both surveys. Catch and Catch Bounds A time-series of catch-at-age information is used to account for fishery removals and therefore provide key information about F in many assessment models; however, in xteNCAM, the catch data was treated somewhat differently, with separate likelihood components for the total removals (with bounds, see below) and the catch proportions-at-age based on the sampling of commercial (and recreational, where possible) landings. The total landings from 1954–2023 (second last column of Table 1), and catch-at-age from 1962–2023 (Table 4) converted to proportions with zeros replaced (see Cadigan 2015), were used as inputs to xteNCAM. Model-predicted landings were derived from the sum-product across age of predicted catch numbers-at-age and mid-year stock weights-at-age (Table 10). In many fisheries, catch is not measured exactly, but it is an important quantity for scaling the total estimated population size. Issues that have typically generated concern about catch accuracy for Northern Cod are unreported or misreported landings (either domestic or non-Canadian), discarding due to quality concerns, and dumping / discarding based on size, particularly when size-based price-differentials are in place. In the post-moratorium period, the determination of the recreational catches has also been a concern, as recreational catches are difficult to estimate, and for Northern Cod, no standard procedures are in place to do so. The magnitude of these potential biases in overall catch reporting and how they change over time have not been quantified for Northern Cod. For the current assessment model, an interval was required for the likely range of catch (C). According to Cadigan (2015), this approach is better than simply assuming an exact catch or ad-hoc adjustment of reported landings. Note that xteNCAM estimates catch almost freely within these bounds, unless there is a strong indication otherwise based on all other input data and the entire model structure. In the 2016 assessment (DFO 2016b), and later at the 2023 framework meeting (DFO 2024b), participants at the meeting agreed that potential catch inaccuracies for Northern Cod likely 14 varied over time and decided to consider four time periods based on different fishery dynamics through the pre-moratorium and post-moratorium period, as well as different states of knowledge about potential catch inaccuracies: 1. A pre-200 nautical mile limit period (1954–77), where catches were high and may have been relatively unconstrained; 2. A pre-moratorium period (1978–91), when catches were high; 3. An early post-moratorium period (1992–2005) when there were small inshore commercial fisheries and recreational fisheries; and 4. A recent period (2006–23) when a directed inshore fishery took place for a few weeks during summer (stewardship fishery) and a dock-side catch monitoring program (DMP) which had been in place since the 1990s was redesigned to incorporate a tier-based approach. This period also included short seasonal recreational fisheries. Although the stewardship and recreational fisheries were expanded in 2016 and 2017, the catch bound estimates for these two years were status quo. There are no direct estimates of recreational landings in the total reported catch for some years, but information from tagging suggested that recreational landings were a substantial fraction of commercial removals throughout the recent period (DFO 2016b). For all four time periods, the lower catch bound (CL) was considered by the meeting to be 10% above reported landings. This value was arbitrary but given the potential for discards and misreported landings, 10% seemed reasonable. The catch bounds were as follows: 1954–77 [CL, CU] = [1.1, 2] where a wide bound accounts for a broad range of possible catches through this period, 1978–91 [CL, CU] = [1.1, 1.5] with the upper bound to account mainly for discards and non-Canadian misreporting, 1992–2005 [CL, CU] = [1.1, 2] with the higher upper bound to account for the period of no DMP and more variable catch. During 2006–23, CL was again fixed at 1.1, but annual estimates of CU were made through differing adjustment up from reported commercial landings. Computation of CU for 2006–23 was based on a simple formula: CU = 1.1 × 1.3 × 1.1 × (Commercial/Total) The rationale for the first scaling factor (1.1) was a minimum adjustment to account for discarding and misreporting, and bias in visual estimation of catch weight evident from the confirmation of estimated weights on DMP authorization numbers. The second factor (1.3) was an adjustment to account for recreational fisher landings, where the analysis of tag returns indicated that recreational landings were on average about 30% of reported commercial landings. A third upward adjustment of 1.1 was included to account for other issues such as tagging estimates being based on numbers rather than weight and a general consideration that it was better to have the interval too wide rather than unreasonably narrow. The final ratio term (commercial/total) ensured that scaling factors were applied appropriately depending on whether estimates of recreational landings were already included in the total reported catch (i.e., 2007, 2009–10, and 2011–13). Thus, for years with no recreational landings estimate in the total reported catch, CU equals total reported catch times 1.573, but if estimated recreational catch was already included, the adjustment is proportionally less. Table 15 summarizes the catch bounds settings and includes estimates of recreational catch that were used to adjust the bounds in the recent period. Tagging The tagging data comprise an extensive series of tagging experiments, where batches of cod are tagged and released in a specific geographic area and time. Tagged fish are subjected to initial tagging mortality due to the stress of capture and handling in the year of release. In 15 addition, depending on the time of year fish were released and the timing of the fishery only a fraction of F and total mortality (Z) were applied in the year of release; the fraction of fishing that occurred was estimated from a table of monthly landings (Table 3). The population of tagged cod from an experiment diminishes over time due to a combination of initial tagging mortality, tag loss, as well as F and M. For all experiments during 1954–2023, irrespective of capture gear type, short-term tagging survival was assumed to be 97% for tag releases in November-June, and 78% for those from July to October (Brattey and Cadigan 2004). Coarse estimates of tag loss from Barrowman and Myers (1996) was used for disc tags, and loss of floy tags was estimated using double tagging and applied using Kirkwood’s model (Kirkwood 1981) with parameter estimates as described in Brattey and Healey (2007). Harvesters do not return the tags from all they fish that are captured; consequently, reporting rates have to be estimated and this was achieved using a high-reward tagging scheme initiated in 1997. Tag reporting rates for Northern Cod have been extensively studied (Cadigan and Brattey 2006; Konrad et al. 2016) and for the 1997 experiments onwards, reporting rates and uncertainties were estimated within xteNCAM. This was achieved by considering reporting rates in xteNCAM as random effects and adding a likelihood component for these reporting rates (Cadigan 2016b). The likelihood was based on the externally derived estimates and their estimated covariance matrix from the model described by Konrad et al. (2016). Capelin as a Predictor of Natural Mortality Capelin have long been known to be an important prey item for Northern Cod (Templeman 1965) and numerous studies have demonstrated links between Capelin biomass and cod productivity (e.g., Krohn et al. 1997; Rose and O’Driscoll 2002; Buren et al. 2014; Koen-Alonso et al. 2021). Further research demonstrated general correspondence between the availability of Capelin, an index of starvation rates of Northern Cod estimated using body condition data, and rates of M estimated by NCAM (Regular et al. 2022). These results are indicative of a mechanistic link between Capelin availability and rates of M experienced by Northern Cod. To move beyond post-hoc analyses and further test these potential links, Capelin data were incorporated into xteNCAM to use it as a predictor of the rates of M for Northern Cod. Specifically, we used the ratio of Capelin/cod biomass as a predictor of M (sensu Regular et al. 2022) since rates of starvation are expected to be worsened when there are insufficient Capelin to support the cod biomass currently in the system. The Capelin-to-cod ratio requires and index of Capelin and cod biomass. For Capelin, the spring acoustic survey index of Capelin biomass was used as it is the best available data we have for the Capelin stock in Divs. 2J3KL (DFO 2025). The Capelin survey has been conducted for most years since 1985 and the latest estimates up to 2023 are used in xteNCAM. For cod, the biomass index from the fall RV survey is used (final column in Table 6). Since the RV survey is conducted in the fall, it is considered a better indicator of the cod population that will be in the system consuming Capelin through the winter and spring of the following calendar year. The RV survey index of cod was therefore lagged by one year to better indicate the biomass of cod relative to the biomass of Capelin in a given year. RESULTS FROM XTENCAM Several xteNCAM fit diagnostics to the various data sources are presented, and these show that the final model fit the productivity data for Northern Cod well. Several of these plots, showing observed and model-predicted values to each of the data sources along with various residual plots, as well as a 10 year retrospective analysis, are given in Appendix A. Key parameter estimates are shown in Table 16. 16 Effect of Capelin Availability The collapse of Capelin in the early-1990s was followed by the collapse of cod, after which both populations slowly increased (Figure 26). Observed dynamics have resulted in periods of relatively high and low Capelin-to-cod biomass ratios (i.e., Capelin availability). Using this ratio as a predictor of M, variation in Capelin availability explains some of the variation in the realized rates of M estimated by xteNCAM (Figure 26). Further, the inclusion of Capelin in xteNCAM is effectively introducing mechanistic processes to explain changes in M, which previously were only done in an post hoc way through the M-shift formulation. Though the Capelin effect does not impose as large of a spike in M in the early 1990s as the M-shift formulation, the ratio does capture previously unexplained peaks in M around 2001, 2009, and 2017. These peaks also correspond to peaks in the starvation-induced mortality index (Figure 25; Regular et al. 2022). Both Capelin and cod biomass, and the Capelin-to-cod ratio has remained stable since 2017 (Figure 26). Stock-recruitment Relationship The use of data back to 1954, and the addition of juvenile survey data, enabled the estimation of a stock-recruit relationship. Several stock-recruitment relationships were tested for the framework; however, the standard formulation of the Beverton-Holt curve was most parsimonious and, as such, was accepted as a plausible description of the relationship between stock size and recruitment (DFO 2023b). This relationship appears to provide reasonable predictions of recruits for the most recent years (Figure 27). Reference Points The use of a stock-recruitment relationship also enabled the internal estimation of reference points such as SSB at maximum sustainable yield, BMSY (Albertsen and Trijoulet 2020). These calculations underpin the new LRP for Northern Cod, which was set at 40% BMSY in accordance with Canada’s Precautionary Approach guidelines (DFO 2009; DFO 2023b). Refined with data up to 2023, these internal calculations yield the current LRP estimate of 276 Kt (95% CI = 180– 423 Kt). An Upper Stock Reference point (USR) has yet to be defined for Northern Cod. Stock Size and Mortality Rates Trends in stock size and mortality rate estimates from xteNCAM are illustrated in Figure 28, and estimated values summarized in Tables 17-24. The abundance of Northern Cod varied around approximately 4 billion individuals through the 1960s to the 1980s, but rapidly declined in the early-1990s. Abundance remained low through the 1990s, but increased in recent decades from 571 million (95% CI = 357–914 million) in 2005 to 3 billion (95% CI = 2–4 billion) in 2017. Growth in abundance has since stalled and levels in 2024 are similar those of 2017 (Figure 28; top panel). Levels of recruitment (age 0) fell to lowest observed levels around 1995, but recruitment has slowly improved and the average number of age 0s from the last five years correspond to about 80% of the average numbers of age 0s observed prior to 1990. Total biomass shows a similar trend to abundance and increased from 86 Kt (95% CI = 71–105 Kt) in 2005 to 700 Kt (95% CI = 590–830 Kt) in 2017 (Figure 28; middle panel). Likewise, SSB increased from 26 Kt (95% CI = 22–31 Kt) in 2005 to 451 Kt (95% CI = 381–534 Kt) in 2017. Growth in both total biomass and SSB has stalled since 2017. Under the newly adopted LRP of 40% BMSY, the stock has been out of the Critical Zone since 2016. As of 2024, the stock is 1.2 (95% CI = 0.7–2.1) times the LRP and, accounting for statistical uncertainties, there is an estimated 22% probability that the stock is in the Critical Zone. 17 Rates of mortality have been highly variable through the history of the stock. Average (ages 5+) F exceeded M through most of the 1950s to the 1980s; however, M has exceeded F since the collapse (Figure 28). Average F declined when the moratorium was imposed in 1992 and again when an inshore fishery was closed in 2003. While directed inshore fisheries for cod have continued throughout most of the post-moratorium period, average F has remained below 0.05 since 2004. Average F in 2023 is estimated to be 0.02 (95% CI = 0.01–0.03). Average M increased rapidly from levels below 0.4 to a peak of 2.5 around 1992–94, then declining to approximately 0.35 during 1995–99. Additional periods of high M are evident in 2000–03 (M approximately equal to 0.7 to 0.9), 2009–10 (M approximately equal to 0.6), and 2017 (M approximately equal to 0.6). Average M decreased to approximately 0.35 between each of these periods. In 2023, average M again appears to be elevated and is estimated to be 0.59 (95% CI = 0.32–1.12). These results on the relative magnitudes of F and M around the time of the moratorium are different from some published studies which argue that the collapse can largely be attributed to illegal fishing (e.g., Hutchings and Myers 1995; Myers et al. 1996, 1997; Shelton and Lilly 2000; Rose and Walters 2019). In the xteNCAM model, the rate of M is estimated and the model and can assign the sudden disappearance of cod from the DFO RV survey to either F or M. However, the model assigns much of the mortality to M to be consistent with the tagging data that are directly integrated into the model. Changing rates of M are further informed by changes in the relative abundance of Capelin, which indicates that the collapse of Capelin explains some of the spike in M during the early 1990s (Figure 26. This aligns with previous research which suggests that prey limitation contributed to the collapse of cod (Koen-Alonso et al. 2021; Regular et al. 2022). However, it remains possible that a portion of the M spike could be attributed to unreported catch by Canadian and/or non-Canadian fleets, especially if recaptured tags from these fish were not returned. Retrospective Analysis A 10-year retrospective analysis was carried out by removing one year of data at a time and re-running the model (Figure A14). The retrospective analysis indicated that overall, there are no systematic patterns of bias in estimates of stock size or rates, but in some years, the estimate of M can be overestimated, and the subsequent SSB underestimated. This is due, in part, to uncertainty in M in the terminal year because of the incomplete tagging from that year (all tags have yet to be returned). There were some signs of retrospective revisions to SSB relative to Blim. This is due to revisions of estimates of BMSY and, consequently, 40% BMSY, given directional revisions of long-term average rates of M, fisheries selectivity, weights-at-age, maturity-at-age, as well as revisions to the stock-recruitment relationship parameter estimates. Projections Projections were conducted to meet the Terms of Reference, which specified the following objectives: • Three-year projections of SSB, in relation to the LRP (with 95% CIs) assuming total removals are (0, 0.5, 1.0, 1.5, and 2) times the 2023 estimated value. • The level of removals that provide estimates of growth of 10%, 20%, 30%, and 50% from the current SSB with a neutral to high probability (50%, 75%, and 95%) over the short-term (3 years). If possible, identify these levels of removals for the medium- to long-term (5 years). Capelin are expected to remain at current levels in the short term4 and, as such, status quo levels of Capelin were assumed across all scenarios. These projections indicated low to 18 moderate probabilities (24–38%) of growth over the next three years (Table 25). Probabilities of exceeding the LRP remains neutral to moderately high under the 0 to 1.0 catch multiplier scenarios (53–69%), but they fall to moderate by 2027 under the 1.5 and 2.0 catch multiplier scenarios (48–51%; Table 26). Under current catch levels (approximately 13,517 t), SSB relative to Blim is projected to be 1.06 (95% CI = 0.30–3.80) by 2027 (Table 27; Figure 29). Given that even without directed fishing (catch multiplier = 0.0) the stock is not projected to be capable of achieving the lower 10% level of growth requested in the Terms of Reference (Table 25), there are no level of removals that can deliver any of the identified levels of growth. Conclusions Although the status of the stock has improved, growth in the stock has stalled since 2017 and, accounting for statistical uncertainties, there is an estimated 22% probability that the stock is in the Critical Zone. Following DFO’s PA Framework, management actions must encourage stock growth in the short term and minimize risk of preventable decline (DFO 2009). Current catch levels increase the risk of pushing the stock into the Critical Zone in the short term. ACKNOWLEDGEMENTS We express our gratitude to the numerous individuals who contributed to the surveys and the processing of data that were used in this assessment, including the ship’s crew, harvesters, and research personnel. These surveys received support from various organizations, including DFO, the Fish, Food and Allied Workers (FFAW) Union, the Natural Sciences and Engineering Research Council of Canada (NSERC) Fisheries Conservation Chair, and the Centre for Fisheries Ecosystems Research (CFER) at the Fisheries and Marine Institute, Memorial University of Newfoundland. We also thank Noel Cadigan for building the foundations of the Northern Cod assessment model and for continued discussions and advice. Finally, we thank the many scientists at DFO who built the foundations of this document and for those who reviewed earlier versions of this document. REFERENCES CITED Albertsen, C.M., and Trijoulet, V. 2020. Model-based estimates of reference points in an age- based state-space stock assessment model. Fisheries Research 230: 105618. Barrowman, N.J., and Myers, R.A. 1996. Estimating tag-shedding rates for experiments with multiple tag types. Biometrics 54: 1410–1416. Brattey, J. 2013. Harvest rates and movements of cod (Gadus morhua) in NAFO Divs. 3KL from tagging and acoustic telemetry. DFO Can. Sci. Advis. Sec. Res. Doc 2013/104: v + 31 p. Brattey, J., and Cadigan, N.G. 2004. Estimation of short-term tagging mortality of adult Atlantic cod (Gadus morhua). Fisheries Research 66: 223--233. Brattey, J., and Healey, B. 2007. Exploitation and movements of Atlantic cod (Gadus morhua) in NAFO division 3KL: Tagging results from the reopened fishery in 2006. DFO Can. Sci. Advis. Sec. Res. Doc 2007/027: iv + 43 p. Brattey, J., Cadigan, N.G., Dwyer, K., Healey, B.P., Morgan, M.J., Murphy, E.F., Parsons, D., and Power, D. 2008a. Assessments of the cod (Gadus morhua) stock in NAFO divisions 2J3KL (April 2007 and April 2008). DFO Can. Sci. Advis. Sec. Res. Doc. 2008/086: x + 150 p. https://doi.org/10.1016/j.fishres.2020.105618 https://doi.org/10.1016/j.fishres.2020.105618 https://doi.org/10.2307/2532854 https://doi.org/10.2307/2532854 https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2013/2013_104-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2013/2013_104-eng.html https://doi.org/10.1016/S0165-7836(03)00203-0 https://doi.org/10.1016/S0165-7836(03)00203-0 https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2007/2007_027-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2007/2007_027-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2008/2008_086-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2008/2008_086-eng.htm 19 Brattey, J., Healey, B., and Porter, D. 2008b. Northern cod (Gadus morhua) 16 years after the moratorium: New information from tagging and acoustic telemetry. DFO Can. Sci. Advis. Sec. Res. Doc 2008/047: iv + 25 p. Brattey, J., Cadigan, N.G., Dwyer, K., Healey, B.P., Morgan, M.J., Murphy, E.F., Parsons, D., and Power, D. 2009. Assessment of the cod (Gadus morhua) stock in NAFO divisions 2J3KL in 2009. DFO Can. Sci. Advis. Sec. Res. Doc 2009/061: viii + 92 p. Brattey, J., Cadigan, N.G., Dwyer, K., Healey, B.P., Morgan, M.J., Murphy, E.F., Maddock Parsons, D., and Power, D. 2010. Assessment of the cod (Gadus morhua) stock in NAFO divisions 2J3KL in 2010. DFO Can. Sci. Advis. Sec. Res. Doc 2010/103: viii + 108 p. Brattey, J., Cadigan, N., Dwyer, K.S., Healey, B.P., Ings, D.W., Lee, E.M., Maddock Parsons, D., Morgan, M.J., Regular, P.M., and Rideout, R.M. 2018. Assessment of the Northern cod (Gadus morhua) stock in NAFO divisions 2J3KL in 2016. DFO Can. Sci. Advis. Sec. Res. Doc. 2018/018: v + 107p. Buren, A.D., Koen-Alonso, M., and Stenson, G.B. 2014. The role of harp seals, fisheries and food availability in driving the dynamics of Northern cod. Marine Ecology Progress Series 511(10): 265–284. Cadigan, N. 2016a. Weight-at-age growth models and forecasts for Northern cod (Gadus morhua. DFO Can. Sci. Advis. Sec. Res. Doc 2016/016: v + 19 p. Cadigan, N.G. 2015. A state-space stock assessment model for Northern cod, including under- reported catches and variable natural mortality rates. Canadian Journal of Fisheries and Aquatic Sciences 73(2): 296–308. Cadigan, N.G. 2016b. Updates to a Northern cod (Gadus morhua) state-space integrated assessment model. DFO Can. Sci. Advis. Sec. Res. Doc. 2016/022: v + 58 p. Cadigan, N.G., and Brattey, J. 2006. Reporting and shedding rate estimates from tag-recovery experiments in Atlantic cod (Gadus morhua) in coastal Newfoundland. Canadian Journal of Fisheries and Aquatic Sciences 63: 1944–1958. Casini, M., Eero, M., Carlshamre, S., and Lövgren, J. 2016. Using alternative biological information in stock assessment: Condition-corrected natural mortality of Eastern Baltic cod. ICES Journal of Marine Science 73(10): 2625–2631. DFO. 2009. A fishery decision-making framework incorporating the precautionary approach. DFO. 2010. Proceedings of the Newfoundland and Labrador Regional Atlantic Cod Framework Meeting: Reference Points and Projection Methods for Newfoundland cod stocks. DFO Can. Sci. Advis. Sec. Proceed. Ser. 2010/053. DFO. 2016a. Proceedings of the Northern cod framework review meeting; November 30 –- December 4, 2015. DFO Can. Sci. Advis. Sec. Proceed. Ser. 2016/031. DFO. 2016b. Stock assessment of Northern cod (NAFO Divs. 2J3KL) in 2016. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep 2016/026. DFO. 2019a. Evaluation of the biomass Limit Reference Point for Northern cod (NAFO Divisions 2J3KL). DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2019/058. DFO. 2019b. Proceedings of the Regional Peer Review of the stock assessment of Northern cod (Divisions 2J3KL), March 19–23, 2018. DFO Can. Sci. Advis. Sec. Proceed. Ser. 2019/007. https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2008/2008_047-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2008/2008_047-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2009/2009_061-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2009/2009_061-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2010/2010_103-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2010/2010_103-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2018/2018_018-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2018/2018_018-eng.html https://doi.org/10.3354/meps10897 https://doi.org/10.3354/meps10897 https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2016/2016_016-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2016/2016_016-eng.html https://doi.org/10.1139/cjfas-2015-0047 https://doi.org/10.1139/cjfas-2015-0047 https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2016/2016_022-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2016/2016_022-eng.html https://doi.org/10.1139/f06-096 https://doi.org/10.1139/f06-096 https://doi.org/10.1093/icesjms/fsw117 https://doi.org/10.1093/icesjms/fsw117 https://www.dfo-mpo.gc.ca/reports-rapports/regs/sff-cpd/precaution-back-fiche-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/Publications/Pro-Cr/2010/2010_053-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/Pro-Cr/2010/2010_053-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/Pro-Cr/2016/2016_031-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/Pro-Cr/2016/2016_031-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2016/2016_026-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2019/2019_058-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2019/2019_058-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/Pro-Cr/2019/2019_007-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/Pro-Cr/2019/2019_007-eng.html 20 DFO. 2022. Stock assessment of Northern cod (NAFO divisions 2J3KL) in 2021. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep 2022/041. DFO. 2023a. Science advice on guidance for Limit Reference Points under the Fish Stocks Provisions. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep 2019/009. DFO. 2023b. Proceedings of the Regional Peer Review for the evaluation of the Limit Reference Point (LRP) for 2J3KL Atlantic cod, January 22–24, 2019. DFO Can. Sci. Advis. Sec. Proceed. Ser. 2023/019. DFO. 2024a. Newfoundland & Labrador Comparative Fishing Analysis – Part 1. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep 2024/002. DFO. 2024b. Northern (2J3KL) Atlantic cod assessment framework. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep 2024/046. DFO. 2025. Assessment of Divisions 2J+3KL Capelin to 2024. DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2025/034. Dutil, J.-D., and Lambert, Y. 2000. Natural mortality from poor condition in Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 57(4): 826–836. Gavaris, S., and Gavaris, C. 1983. Estimation of catch at age and its variance for groundfish stocks in the newfoundland region. Sampling commercial catches of marine fish and invertebrates. Edited by WG Doubleday and D. Rivard. Can. Spec. Publ. Fish. Aquat. Sci 66: 178–182. Gregory, R.S., Dalley, K.L., Newton, B.H., Sargent, P.S., and Cooke, E.L. 2019. Relative strength of three cohorts (2015-17) of Atlantic cod, from nearshore surveys of demersal age 0 and 1 juveniles in Newman Sound, Bonavista Bay. DFO Can. Sci. Advis. Sec. Res. Doc. 2019/012: v + 15p. Hutchings, J.A., and Myers, R.A. 1995. The biological collapse of Atlantic cod off Newfoundland and Labrador: An exploration of historical changes in exploitation, harvesting, technology and management. In The North Atlantic fisheries: Successes, failures and challenges. Edited by R. Arnason and L. Felt. Institute of Island Studies, Charlottetown, PEI, Canada. pp. 38–93. Kirkwood, G.P. 1981. Generalized models for the estimation of rates of tag shedding by southern bluefin tuna (thunnus maccoyii). Journal du Conseil / Conseil Permanent International pour l’Exploration de la Mer 39: 256–260. Koen-Alonso, M., Lindstrøm, U., and Cuff, A. 2021. Comparative modeling of cod-capelin dynamics in the Newfoundland-Labrador shelves and Barents Sea ecosystems. Frontiers in Marine Science 8: 579946. Konrad, C., Brattey, J., and Cadigan, N.G. 2016. Modelling temporal and spatial variability in tag reporting-rates for newfoundland cod (Gadus morhua). Environmental and Ecological Statistics 23: 387. Krohn, M., Reidy, S., and Kerr, S. 1997. Bioenergetic analysis of the effects of temperature and prey availability on growth and condition of Northern cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 54(S1): 113–121. Lewis, R.S., Ings, D.W., and Rogers, B. 2022. Demersal juvenile Atlantic cod (Gadus morhua) survey of the northeast coast of Newfoundland: 2020 Fleming survey. DFO Can. Sci. Advis. Sec. Res. Doc. 2022/056: v + 32p. https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2022/2022_041-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2023/2023_009-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2023/2023_009-eng.html https://publications.gc.ca/site/eng/9.922715/publication.html https://publications.gc.ca/site/eng/9.922715/publication.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2024/2024_002-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2024/2024_046-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/SAR-AS/2025/2025_034-eng.html https://doi.org/10.1139/f00-023 https://doi.org/10.1139/f00-023 https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2019/2019_012-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2019/2019_012-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2019/2019_012-eng.html https://doi.org/10.1093/icesjms/39.3.256 https://doi.org/10.1093/icesjms/39.3.256 https://doi.org/10.3389/fmars.2021.579946 https://doi.org/10.3389/fmars.2021.579946 https://doi.org/10.1007/s10651-016-0344-0 https://doi.org/10.1007/s10651-016-0344-0 https://doi.org/10.1139/f96-159 https://doi.org/10.1139/f96-159 https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2022/2022_056-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2022/2022_056-eng.html 21 Lilly, G.R., Murphy, E.F., Healey, B.P., and Brattey, J. 2005. An assessment of the cod (Gadus morhua) stock in NAFO Divisions 2J3KL in March 2005. DFO Can. Sci. Advis. Sec. Res. Doc 2005/018: iv + 211 p. Lilly, G.R., Murphy, E.F., Healey, B.P., and Brattey, J. 2006. An assessment of the cod (Gadus morhua) stock in NAFO divisions 2J3KL in April 2006. DFO Can. Sci. Advis. Sec. Res. Doc 2006/043: vi + 159 p. Maddock Parsons, D. 2014. Update of Sentinel survey results in NAFO Divisions 2J3KL for 1995–2012. DFO Can. Sci. Advis. Sec. Res. Doc 2014/014: v + 32 p. Mello, L.G.S., and Simpson, M.R. 2023. Sentinel surveys 1995–2020 – catch rates and biological information on Atlantic cod (Gadus morhua) in NAFO Divisions 2J3KL. DFO Can. Sci. Advis. Sec. Res. Doc 2023/041: iv + 33 p. Morgan, M.J., and Hoenig, J.H. 1997. Estimating maturity-at-age from length stratified sampling. Journal of Northwest Atlantic Fishery Science 21: 51–63. Myers, R.A., Barrowman, N.J., Hoenig, J.M., and Qu, Z. 1996. The collapse of cod in eastern Canada: The evidence from tagging data. ICES Journal of Marine Science 53(3): 629–640. Myers, R.A., Barrowman, N.J., and Hutchings, J.A. 1997. Inshore exploitation of newfoundland Atlantic cod (Gadus morhua) since 1948 as estimated from mark-recapture data. Canadian Journal of Fisheries and Aquatic Sciences 54(1): 224–235. Power, D., Ings, D.W., Rideout, R.M., and Healey, B.P. 2016. Performance and description of Canadian multi-species bottom trawl surveys in NAFO Subarea 2 + Divisions 3KLMNO, with emphasis on 2014–2015. NAFO SCR Doc. (16/28). Regular, P.M., Buren, A.D., Dwyer, K.S., Cadigan, N.G., Gregory, R.S., Koen-Alonso, M., Rideout, R.M., Robertson, G.J., Robertson, M.D., Stenson, G.B., and others. 2022. Indexing starvation mortality to assess its role in the population regulation of Northern cod. Fisheries Research 247: 106180. Regular, P.M., Robertson, G.J., Kumar, R., Varkey, D.A., Gregory, R.S., Lewis, R.S., Skanes, K., Gullage, N., Koen-Alonso, M., and Dwyer, K.S. 2025a. Extending the Northern Cod (Gadus morhua) Assessment Model - Part I: Bridging Gaps with Additional Data and Model Variations. DFO Can. Sci. Advis. Sec. Res. Doc. 2025/034. iv + 79 p. Regular, P.M., Kumar, R., Varkey, D.A., Koen-Alonso, M., and Stenson, G.B. 2025b. Extending the Northern Cod (Gadus morhua) Assessment Model - Part II: Quantifying the Impact of Capelin and Seals. DFO Can. Sci. Advis. Sec. Res. Doc. 2025/035. iv + 36 p. Rideout, R.M., and Ings, D.W. 2021. Research vessel bottom trawl survey report (NL region): A stock-by-stock summary of survey information up to and including the 2019 spring and autumn surveys. Canadian Technical Report of Fisheries and Aquatic Sciences 3425: vii + 52. Rose, G.A., and Walters, C.J. 2019. The state of Canada’s iconic Northern cod: A second opinion. Fisheries Research 219: 105314. Rose, G., and O’driscoll, R. 2002. Capelin are good for cod: Can the Northern stock rebuild without them? ICES Journal of Marine Science 59(5): 1018–1026. Rose, G.R., Nelson, R.J., and Mello, L.G.S. 2011. Isolation or metapopulation: Whence and whither the Smith Sound cod? Canadian Journal of Fisheries and Aquatic Sciences 68: 152–169. https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2005/2005_018-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2005/2005_018-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2006/2006_043-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2006/2006_043-eng.htm https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2014/2014_014-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/publications/resdocs-docrech/2014/2014_014-eng.html https://waves-vagues.dfo-mpo.gc.ca/library-bibliotheque/41190166.pdf https://waves-vagues.dfo-mpo.gc.ca/library-bibliotheque/41190166.pdf https://doi.org/10.2960/J.v21.a4 https://doi.org/10.1006/jmsc.1996.0083 https://doi.org/10.1006/jmsc.1996.0083 https://doi.org/10.1139/f96-166 https://doi.org/10.1139/f96-166 https://doi.org/10.1016/j.fishres.2021.106180 https://doi.org/10.1016/j.fishres.2021.106180 https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2025/2025_034-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2025/2025_034-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2025/2025_034-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2025/2025_035-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2025/2025_035-eng.html https://www.dfo-mpo.gc.ca/csas-sccs/Publications/ResDocs-DocRech/2025/2025_035-eng.html https://doi.org/10.1016/j.fishres.2019.105314 https://doi.org/10.1016/j.fishres.2019.105314 https://doi.org/10.1006/jmsc.2002.1252 https://doi.org/10.1006/jmsc.2002.1252 https://doi.org/10.1139/F10-135 https://doi.org/10.1139/F10-135 22 Schijns, R., Froese, R., Hutchings, J.A., and Pauly, D. 2021. Five centuries of cod catches in eastern Canada. ICES Journal of Marine Science 78(8): 2675–2683. Shelton, P.A., and Lilly, G.R. 2000. Interpreting the collapse of the Northern cod stock from survey and catch data. Canadian Journal of Fisheries and Aquatic Sciences 57(11): 2230– 2239. Shelton, P.A., Stansbury, D.E., Murphy, E.F., Lilly, G.R., and Brattey, J. 1996. An assessment of the cod stock in NAFO Divisions 2J3KL. NAFO SCR Doc. 96/62: 56 p. Smith, S.J., and Somerton, G.D. 1981. STRAP: A user oriented computer analysis system for groundfish research trawl survey data. Can. Tech. Rep. Fish. Aquat. Sci 1030: iv + 66 p. Stansbury, D.E., Maddock Parsons, D., and Shelton, P.A. 2000. An age disaggregate index from the sentinel program for cod in 2J3KL. DFO Can. Stock Ass. Sec. Res. Doc: 64 p. Taggart, C., Penney, P., Barrowman, N., and George, C. 1995. The 1954-1993 Newfoundland cod-tagging database: Statistical summaries and spatial-temporal distributions. Canadian Technical Report of Fisheries and Aquatic Sciences 2042: 441 p. Templeman, W. 1965. Some instances of cod and haddock behaviour and concentrations in the Newfoundland and Labrador areas in relation to food. International Commission for the Northwest Atlantic Fisheries (ICNAF) Special Publication 6: 449–461. Wheeland, T., L., and Rideout, R.M. 2023. Coverage of the 2022 Canadian (Newfoundland and Labrador region) multi-species RV bottom trawl survey with notes on comparative fishing. NAFO SCR Doc. (23/042). https://doi.org/10.1093/icesjms/fsab153 https://doi.org/10.1093/icesjms/fsab153 https://doi.org/10.1139/f00-191 https://doi.org/10.1139/f00-191 https://publications.gc.ca/collections/collection_2015/mpo-dfo/Fs70-5-2001-044-eng.pdf https://publications.gc.ca/collections/collection_2015/mpo-dfo/Fs70-5-2001-044-eng.pdf 23 TABLES Table 1. Reported landings (t) of cod from NAFO Divisions 2J3KL from 1954 onward. Proportion of Canadian catch (%) indicated in parentheses. Provisional catches for current year. A dot (.) indicates no data. Year 2J Offshore Mobile gear 2J Fixed gear 2J Total (% Can) 3K Offshore Mobile gear 3K Fixed gear 3K Total (% Can) 3L Offshore Mobile gear 3L Fixed gear 3L Total (% Can) 2J3KL overall Total (% Can) 2J3KL TAC 1954 . . . . . . . . . . . . 315,843a . . 1955 . . . . . . . . . . . . 232,858a . . 1956 . . . . . . . . . . . . 236,210a . . 1957 . . . . . . . . . . . . 254,456a . . 1958 . . . . . . . . . . . . 206,710a . . 1959 46,372 17,533 63,905 (27%) 97,678 56,264 153,942 (37%) 56,030 85,695 141,725 (64%) 359,572 (46%) . 1960 164,124 15,418 179,542 (9%) 75,052 47,676 122,728 (39%) 71,340 94,192 165,532 (61%) 467,802 (35%) . 1961 243,145 17,545 260,690 (7%) 64,023 31,159 95,182 (33%) 78,574 70,659 149,233 (50%) 505,105 (25%) . 1962 226,841 23,424 250,265 (9%) 47,015 42,816 89,831 (48%) 94,659 72,271 166,930 (46%) 507,026 (28%) . 1963 197,869 23,767 221,636 (11%) 79,331 47,486 126,817 (37%) 87,461 73,295 160,756 (48%) 509,209 (29%) . 1964 197,372 14,787 212,159 (7%) 121,423 40,735 162,158 (25%) 152,528 75,806 228,334 (38%) 602,651 (23%) . 1965 246,650 25,117 271,767 (9%) 50,118 26,467 76,585 (35%) 137,740 58,943 196,683 (34%) 545,035 (22%) . 1966 226,283 22,645 248,928 (9%) 58,920 32,208 91,128 (35%) 128,459 55,990 184,449 (35%) 524,505 (23%) . 1967 217,283 27,721 245,004 (11%) 78,801 24,905 103,706 (24%) 213,821 49,233 263,054 (24%) 611,764 (19%) . 1968 359,758 12,937 372,695 (5%) 121,627 40,768 162,395 (26%) 227,592 47,332 274,924 (23%) 810,014 (15%) . 1969 405,261 4,328 409,589 (1%) 81,005 24,923 105,928 (24%) 170,200 67,973 238,173 (36%) 753,690 (15%) . 1970 212,961 1,963 214,924 (1%) 78,366 21,512 99,878 (22%) 152,311 53,113 205,424 (33%) 520,226 (18%) . 1971 154,700 3,313 158,013 (2%) 61,537 21,111 82,648 (26%) 160,742 38,115 198,857 (25%) 439,518 (17%) . 1972 149,435 1,725 151,160 (1%) 133,376 14,054 147,430 (10%) 113,432 46,273 159,705 (32%) 458,295 (14%) . 1973 54,108 3,619 57,727 (8%) 159,761 13,190 172,951 (8%) 98,992 24,839 123,831 (21%) 354,509 (12%) 666,000 24 Year 2J Offshore Mobile gear 2J Fixed gear 2J Total (% Can) 3K Offshore Mobile gear 3K Fixed gear 3K Total (% Can) 3L Offshore Mobile gear 3L Fixed gear 3L Total (% Can) 2J3KL overall Total (% Can) 2J3KL TAC 1974 119,463 1,804 121,267 (1%) 149,208 10,747 159,955 (7%) 68,798 22,630 91,428 (26%) 372,650 (10%) 657,000 1975 78,988 3,000 81,988 (4%) 112,867 15,518 128,385 (12%) 54,440 22,695 77,135 (30%) 287,508 (15%) 554,000 1976 30,785 3,851 34,636 (11%) 80,311 20,879 101,190 (21%) 43,185 35,209 78,394 (48%) 214,220 (29%) 300,000 1977 40,109 3,523 43,632 (9%) 27,827 28,818 56,645 (53%) 32,161 40,282 72,443 (63%) 172,720 (46%) 160,000 1978 22,228 6,638 28,866 (39%) 13,400 29,623 43,023 (85%) 21,476 45,194 66,670 (82%) 138,559 (74%) 135,000 1979 15,731 8,445 24,176 (73%) 38,462 27,025 65,487 (74%) 26,877 50,359 77,236 (84%) 166,899 (78%) 180,000 1980 21,029 17,210 38,239 (81%) 28,750 37,015 65,765 (90%) 29,486 42,298 71,784 (81%) 175,788 (84%) 180,000 1981 26,885 14,251 41,136 (88%) 26,959 23,002 49,961 (92%) 36,824 42,827 79,651 (81%) 170,748 (86%) 200,000 1982 67,307 14,429 81,736 (89%) 12,955 42,141 55,096 (93%) 36,452 56,490 92,942 (90%) 229,774 (90%) 230,000 1983 41,434 10,748 52,182 (92%) 34,436 40,683 75,119 (96%) 50,043 55,001 105,044 (90%) 232,345 (92%) 260,000 1984 12,013 13,150 25,163 (89%) 59,173 35,143 94,316 (88%) 63,641 49,351 112,992 (86%) 232,471 (87%) 266,000 1985 1,544 10,211 11,755 (99%) 81,825 30,368 112,193 (88%) 68,039 39,306 107,345 (71%) 231,293 (81%) 266,000 1986 13,593 12,916 26,509 (70%) 67,867 28,384 96,251 (94%) 111,751 32,202 143,953 (63%) 266,713 (75%) 266,000 1987 43,343 16,022 59,365 (93%) 45,846 27,442 73,288 (92%) 70,528 36,743 107,271 (76%) 239,924 (85%) 256,000 1988 41,477 17,112 58,589 (100%) 40,310 33,820 74,130 (100%) 84,553 51,405 135,958 (80%) 268,677 (90%) 266,000 1989 34,629 23,304 57,933 (98%) 38,529 20,711 59,240 (98%) 77,579 59,238 136,817 (73%) 253,990 (85%) 235,000 1990 18,066 14,505 32,571 (99%) 27,424 27,516 54,940 (99%) 56,675 75,266 131,941 (81%) 219,452 (88%) 199,262 1991 703 2,214 2,917 (97%) 30,423 13,332 43,755 (99%) 79,924 45,416 125,340b (60%) 172,012 (71%) 190,000 1992 0 18 18 (100%) 857 884 1,741 (84%) 28,237 10,960 39,197c (63%) 40,956 (64%) 0 1993 0 13 13 (100%) 0 541 541 (100%) 2,427 8,411 10,838d (78%) 11,392 (79%) 0