(Illex Illecebrosus) and Preliminary Annual Landings-Per-Unit-Of-Effort for the Southern (USA) Stock Component

Working Paper Not to be cited without author’s permission Characterization of Body Weight Data from the Landings of Northern Shortfin Squid (Illex illecebrosus) and Preliminary Annual Landings-per-Unit-of-Effort for the Southern (USA) Stock Component

by Lisa Hendrickson and Alicia Miller

Northeast Fisheries Science Center April 20, 2020

Background

Similar to other ommastrephid squid species, the Northern shortfin squid (Illex illecebrosus) stock, hereafter referred to as Illex, is a transboundary resource. Illex fisheries occur within the Exclusive Economic Zones (EEZs) of the USA and Canada, as well as in international waters that are managed by the Northwest Atlantic Fisheries Organization (NAFO). Thus, the northern stock component (NAFO Subareas 3+4) is managed by Canada and NAFO and the southern stock component is managed in the USA (NAFO Subareas 5+6) by the Mid-Atlantic Fishery Management Council (MAFMC; Hendrickson and Showell, 2019). The resource constitutes a unit stock throughout its range in the Northwest Atlantic Ocean (Hendrickson 2004). The only documented spawning grounds for the species is located in US waters on the continental shelf and upper slope in the Mid-Atlantic Bight (Hendrickson, 2004). The small-mesh bottom trawl fishery for Illex occurs on the spawning grounds (Hendrickson and Hart 2006). The fishery occurs when squid are available in commercial quantities, primarily during June-October, near the edge of the US shelf (Hendrickson and Holmes 2004; Hendrickson 2004).

The application of most conventional stock assessment methods are inappropriate for I. illecebrosus and other cephalopod species given their unique life histories and population dynamics (Hendrickson 2004; Arkhipkin et al. 2015; Arkhipkin et al. In Press). Like other ommastrephids, I. illecebrosus is semelparous and spawns throughout the year with several peaks that result in the presence of multiple, overlapping sub-annual cohorts. The species has a lifespan of less than one year (Dawe and Beck 1997; Hendrickson 2004).

Illex stock assessments are data-poor because the existing Northeast Regional fishery-dependent datasets are not collected at the high temporal and spatial resolution (i.e., daily, tow-based catch, effort, fishing location and biological data) required to apply the in-season depletion-type models that are used for other ommastrephid squid stocks (NEFSC 1999; Arkhipkin et al. 2015; Arkhipkin et al. In Press). The existing fishery data requires merging of the Dealer and Vessel Trip Report Databases to obtain trips that are a 1:1 match. Trips that are not matched do not contain the data needed for catch-per-unit-of-effort (CPUE) analyses. In addition, the spatial (large Statistical Areas for reporting fishing locations primarily resulting in a single location per Illex trip) and temporal (average tow duration and number of tows by subtrip) resolution of these data is inadequate for accurate in-season stock assessment. Another requirement for in-season assessment

Working Paper Not to be cited without author’s permission is a pre-fishery Illex survey (NEFSC 1999) conducted with commercial fishing gear similar to the survey conducted in 2000 (Hendrickson 2004). The spring and fall NEFSC bottom trawl surveys are conducted during periods of the species’ annual migration on (spring survey) and off (fall survey) the continental shelf. CVs for the spring survey indices are higher than they are for the fall survey indices. Consequently, estimates of Illex relative abundance and biomass on the US shelf are derived using the fall rather than the spring survey data (NEFSC 1999).

History of Illex illecebrosus in-season stock assessment

The 2017-2019 Illex fishery closures led to the MAFMC’s Science and Statistical Committee (SSC) to revisit the potential for implementation of in-season assessment of the US stock component. Given this objective, it is important to understand what has already been accomplished on this topic to avoid repetitive research.

Since 1996 (Hendrickson et al. 1996), NEFSC assessments of I. illecebrosus have recommended conducting real-time or in-season stock assessments that can be used for adaptive, in-season management. Following a 1998 fishery closure, the Illex fishing industry requested implementation of an in-season management program in order to avoid foregone yield during years of high on-shelf abundance. Conversely, there must also be the recognition of potential fishery closures during periods of low Illex abundance, which emphasizes the need for a pre-fishery survey, conducted by an Illex fishing vessel, the data from which can be used to derive an initial abundance estimate. In order to evaluate the feasibility of conducting in-season Illex assessments, beginning in 1999 I obtained multiple research grants that allowed me to conduct several pilot studies on this topic. Prior to 1999, I also began collecting body weight data from Illex processors. The history of collaborative research conducted to collect the high-resolution CPUE and biological data required for in-season stock assessment such as a weekly, depletion model are summarized in Table 1. Although these projects demonstrated that in-season data collection and assessments were possible, previous fishery managers were not in favor of adopting it because it would require a different type of infrastructure than currently exists. Furthermore, the required data for such assessments was (and still is) lacking and would be more costly to collect. For example, an e-VTR tow-based data collection framework with daily reporting and in-season assessment analysis would be necessary. The current management framework is also not set-up for adaptive fishery management. Real-time or in-season assessment of stock size was recommended in previous assessments to ensure adequate spawner escapement for each sub-annual cohort that is fished in order to increase the probability that subsequent recruitment is adequate to sustain each cohort. In- season assessment has been implemented in other ommastrephid stocks to reduce the likelihood of foregone yield during high abundance years and recruitment overfishing during low abundance years (Basson et al. 1996; Arkhipkin et al. In Press).

Previous Illex assessments incorporated data collected from a research project designed to demonstrate the feasibility of real-time at-sea collection of tow-based fishery and body weight data. During 1998-2003, tow-based fishery data were reported daily by Illex fishermen to the NEFSC via Boatracs macros and a captain-specific password-protected e-VTR website. The at- sea data were automatically uploaded to the NEFSC e-VTR website and combined with the remainder of the required VTR data fields (Hendrickson et al. 2003). Another grant-funded research project involved a pre-fishery, pilot Illex bottom trawl survey using two chartered Illex

Working Paper Not to be cited without author’s permission fishing vessels (Hendrickson 2004; NEFSC 1999; Hendrickson and Hart 2006). Data from these research projects were then incorporated in the 1999, 2003 and 2006 stock assessments to demonstrate their utility for in-season weekly estimates of stock size using depletion models.

This Working Paper focuses solely on the southern component of the Illex stock with the objective of extending the time-series of landings-per-unit-of-effort (LPUE) and body weight data that have been evaluated in previous Illex stock assessments through 2004. An Illex stock assessment has not been conducted since 2005 due to the lack of adequate data at the resolution required to conduct in-season depletion modeling and to estimate %MSP Biological Reference Points (NEFSC 2006). Data from 15 additional fishing seasons, from 2005-2019, are evaluated here in addition to the 1997-2004 data that were previously evaluated. Characterization of data from these additional fishing seasons will allow for the comparison of body weight and LPUE trends during a broader range of low, medium and high periods of Illex abundance on the U.S. shelf for the purpose of evaluating the potential for these data to be used to conduct in-season stock assessments.

Methods

Mean Body Weight

Mean body weight has been used as a measure of productivity for the Illex stock in previous stock assessments (NEFSC 1999; NEFSC 2003; NEFSC 2006; Hendrickson et al. 2001). This report characterizes both annual and weekly Illex body weight data collected from the landings during 1997-2019. The body weight data for 1997-2003 was collected as part of a cooperative research study that involved real-time, fishery-dependent data collection. I continued collecting these data in order to generate a time-series of high-frequency body weight data with which to evaluate changes in stock productivity. Body weight data for 2004-2006 and 2009-2018 were collected from landings of the directed fishery by QA/QC staffs from the two primary Illex processors. Data were generously provided by Lunds Fisheries in Excel spreadsheets which required some reformatting, for 2016-2019. Data were also provided by Seafreeze Ltd. on their QA/QC sampling forms and required extensive keypunching by staffs from the Population Dynamics Branch. Seafreeze Ltd. provided us with additional information to identify each trip and this allowed me to add “date landed” from the Dealer Database to their dataset in order to assign week of the year to the samples from each trip. Mean body weight for the samples collected by port samplers was computed by dividing the sample weight by the number of lengths in the sample. Samples collected by port samplers included 100 squid per market category. These samples were obtained opportunistically with the objective of collecting the numbers of monthly samples that I requested by market category, fishery region and year. Such samples do not include all market categories from each trip, unlike the Lunds Fisheries and Seafreeze Ltd. samples. As a result of the different sampling protocols, the two mean body weight datasets were summarized separately. For the purpose of comparing trends, stratified mean body weights for the NEFSC fall bottom trawl surveys (Hendrickson and Showell 2019) are also presented.

Catch-per-Unit-of-Effort

Both annual and weekly time-series of landings-per-unit-of-effort (LPUE) have been derived in previous Illex assessments (NEFSC 1999; NEFSC 2003; NEFSC 2006). Discards will be estimated

Working Paper Not to be cited without author’s permission for the upcoming For the subject analysis, I assumed that LPUE was representative of CPUE because the most recent assessment indicated that Illex discards generally comprised a small portion (0.5-6.0%) of the annual catches (NEFSC 2006). I have used the LPUE acronym in this report instead of CPUE to be clear that the subject analysis does not include discard data. Landings, fishing effort and location data were retrieved from the NEFSC commercial fisheries database that includes merged trips from the Dealer Database and the Vessel Trip Report Database. The methodology used to create the merged database is described in (Wigley et al. 2008). Only trips with 1:1 matches between the Dealer and VTR Databases contain the fishing effort and Statistical Area data necessary to compute LPUE. From this subset of trips available for LPUE estimation the dataset was further subset to include only directed trips. Directed trips were defined as trips with Illex landings > 10,000 lbs and > 50% of the total trip weight. This combination of landings thresholds was used to exclude longfin squid trips with Illex bycatch from the LPUE dataset. Data from 1997-2018 were included in the LPUE analysis because although Dealer and VTR reporting became mandatory on May 2, 1996, 1997 was the first complete year of reporting (MAFMC, 1994). The 2019 merged Dealer-VTR dataset was not available in time for inclusion in this report.

The PROC GENMOD SAS procedure was used to derive a standardized LPUE time-series. LPUE data (mt per day fished), for 1997-2018, were fit to Type 3 GLMs with normal, gamma and negative binomial error structures. Goodness-of-fit for the three error type models was determined based on model deviance divided by the degrees of freedom. Main effects included in each initial set of models included all possible combinations of year, week, permit and vessel type (i.e., freezer trawler (FT), refrigerated seawater (RSW) or ice (ICE) boat), without interactions. Permit and vessel type were included in the models because LPUE is known to vary by vessel type (NEFSC 1999) and because permit incorporates both vessel type and vessel-specific properties that can affect LPUE (e.g., specific captain and crew size). Vessel types identified in previous stock assessments were used in the analysis along with vessel types which were confirmed by industry members for vessels which entered the fishery after 2004. Some subsequent model configurations also included Statistical Area (area). LPUE data included in the GLM models were also subset for the Illex fishing season, which was identified based on an examination of weekly nominal LPUE data during 1997-2018.

Trends in the NEFSC fall survey biomass indices (stratified mean kg per tow) and the standardized LPUE indices were compared using Spearman’s Rank-Order correlation analysis.

Results and Discussion

Mean body weight

Research survey trends in annual mean body weight are associated with annual trends in Illex relative abundance, such that stratified mean body weight is generally lower during year of low relative abundance, and vice versa, on the US Shelf (Hendrickson et al. 2004) and Scotian Shelf (Hendrickson and Showell 2019). Annual mean body weights of individuals caught during NEFSC fall bottom trawl surveys decreased substantially in 1982 following the collapse of the northern stock component. Mean body weights then declined and remained below the 1982-2018 average during most years since 1995 (Hendrickson and Showell 2019).

Working Paper Not to be cited without author’s permission Changes in Illex mean body weight represent the combined effects of growth, mortality (both fishing and natural mortality), and emigration and immigration. As is typical for squid species, age rather than body size data must be used for cohort identification because same-sized individuals can be from two different overlapping seasonal cohorts which have different growth rates (Dawe and Beck 1997; Arkhipkin et al. In Press). As documented in previous assessments, mean body weight gradually increases during the fishing season, and for 1997-2018 combined, the peak occurred in week 34 (Figure 1). Thereafter, mean body weight decreased through the end of the fishing season as squid migrated back offshore and south. The ongoing Illex aging study funded by the MAFMC will be used to identify whether these changes in size involve one or more cohorts.

Body weight data obtained from the squid processors consisted of much larger annual sample sizes than the length-based data obtained by NMFS port samplers (Figure 2). Regardless, smooths of both datasets showed a similar W-shaped trend with the presence of larger squid at the beginning, middle and end of the time-series (Figure 3). Annual mean body sizes of squid sampled by the Illex processors ranged between 100 and 200 g, but the mean body sizes of squid collected by the port samplers were larger and ranged between 180 and 480 g. This disparity requires further investigation. However, as described in the Methodology section, other than for trends, the two datasets are not directly comparable due to the differences in sampling protocols. As a result of larger sample sizes and sampling of unculled samples from all RSW and ice boat trips and all market categories from FTs, the processor dataset should be more accurate than the port agent samples. However, further investigation is needed to confirm this conclusion.

When trends between the fishery mean body weight time-series and the stratified mean body weight time-series from the NEFSC fall surveys are compared, the fishery time-series does not show the gradual decrease exhibited by the survey data (Figure 3). Body weight data collected from the directed fishery landings represents cluster sampling due to the inherent nature of fishing behavior. When fishing, clusters of tows are conducted in close proximity to one another in areas of high squid abundance and can result in biased fishery-dependent data. For example, Illex mean body size increases with latitude (Hendrickson 2004) so if Illex body weight samples are predominately obtained from northern fishery areas body size will appear to be larger for that time period. In contrast to the fishery data, the survey’s stratified random sampling protocol ensures that these body weight samples are representative of the population that is present on the shelf during the fall. Although the fall survey can be viewed as a post-fishery index because it occurs at or near the end of the fishing season, the gradual decreasing trend in body weight is expected to be present during the fishing season as well. As part of the next steps, a spatial analysis of the fishery body weight data will be conducted to determine its temporal and spatial representativeness.

Graphs of body size by week of the year (Figures 4-6) show that Illex body weight trends do not always follow the characteristic rise-and-fall pattern during the fishing season. Although some of the variability in the weekly trends may be attributable to low sample sizes, body size trends three months prior to the fall survey (June-August) would also be expected to show a decreasing trend if they were representative of the population. For ease of identifying trends, smooths of weekly body size data are shown and they indicate that squid weighed more during years of high landings (1998, 2004 and 2017-2019) and less during years of low landings (Figure 7).

Working Paper Not to be cited without author’s permission Landings-per-Unit-of-Effort

Based on weekly nominal LPUE values, the Illex fishing season occurred during weeks 17-45 and averaged 18 weeks in duration during 1997-2018. The first week of the fishing season ranged between weeks 17 and 25, with an average start week at week 22. Discounting years with fishery closures (i.e., 1998, 2004 and 2017-2018), the duration of the fishing season ranged between 14 and 25 weeks and averaged 20 weeks.

The proportion of total Illex landings represented by trips with 1:1 matches between the Dealer and VTR Databases (i.e., landings from trips available for LPUE analysis) gradually improved during 1997-2018, but proportions were much lower during the early part of the time-series (dashed line in Figure 8). Nearly all of the landings from the trips available for LPUE analysis were used to derive the LPUE indices (blue line versus red line in Figure 8). Therefore, the landings proportions indicated by the dashed line in Figure 8 are also representative of the landings proportions included in the LPUE dataset. During 1997-2001, the proportions were lowest and ranged between 0.51 in 1997 and 0.67 in 2001. During most years between 2002 and 2018, the landings proportion were near or above the mean of 0.78. The proportions were highest (average = 0.95) during 2011-2018. In summary, the 2011-2018 LPUE data comprise the highest proportion of total Illex landings, followed by the 2002-2010 and then the 1997-2001 estimates.

Fleet size during 1997-2018 ranged from 4 vessels in 2015 to 27 vessels in 1998 (Figure 9). With respect to fishing effort, the number of trips conducted and fleet size showed trends that were similar to the number of nominal days fished during most years. The exceptions were 2017 and 2018, when days fished did not increase as rapidly as the numbers of trips and vessels. Days fished reached a peak in 2011 and the numbers of trips and vessels peaked in 2018 and 1998, respectively.

When characterized by vessel type, landings (mt) and effort (days fished) showed similar trends, although more they were more variable for FTs, which was likely due to the lengthier trips of FTs in comparison to RSW and ice boats (Figure 9). Landings and effort trends varied slightly between vessel types whereby peaks in both variables occurred during 2011 for FTs, but landings peaks for ice and RSW boats occurred in 2019 when the ice and RSW boats harvested most of the landings (Figure 9). Landings were high for all three vessel types during 1998 and 2004.

A major change in the fleet composition occurred during 1997-2018. Prior to 2008, the fleet was dominated by FTs which harvested most of the landings (75% on average; Figure 9). After 2008, FTs still harvested most (63%) of the landings, but landings by RSW boats increased in conjunction with the increase in numbers of RSW boats. The numbers of RSW and ice boats increased rapidly after 2017 and peaked in 2019 at 14 and 12, respectively (Figure 9). The number of FTs peaked at 11 in 2004 and decreased to only 7 vessels by 2019. The 2019 data for number of vessels by vessel type were obtained from the VTR data to illustrate the change in fleet composition in recent years. Based on discussions with Illex fishermen and processors, the increase in RSW boats coincided with a reduction in FTs because some have been converted to RSW boats and multiple FTs have either sunk or sold their Illex permits. Although FTs converted to RSW boats may have similar per-trip harvest capacities, their annual harvesting capacities can surpass those of FTs because RSW boats make shorter, more frequent trips than FTs and several of the RSW boats are large capacity vessels. RSW and ice boat trip durations are shorter because they

Working Paper Not to be cited without author’s permission are limited by the rapid degradation of Illex catches and averaged four days during 1997-2018. FTs make longer trips (eight on average) than RSW or ice boats because they sort and freeze their landings at sea (Figure 10). Crew sizes for FTs and RSW and ice boats averaged 9 and 4, respectively. The number of days fished reached a peak in 2011 for both FTs and RSW, but FT days fished was much higher than for RSW boats. This trend reversed during 2018 and 2019. The average days fished during 1997-2018 was more variable for FTs than ice and RSW boats and comprised 24%, 18% and 20% of the average trip duration, respectively. Steam time comprises a large portion of the trip duration because the fishery occurs offshore near the shelf edge (Figure 11). The remainder of the trip duration is comprised of time spent searching for productive fishing locations (Powell et al. 2003), sorting and freezing the landings for FTs and longer steams between fishing locations, which tend to occur at night because Illex fishing occurs during the daytime.

The landings (mt) and effort (days fished) data used to estimate the LPUE indices were high for all three vessel types during 1998 and 2004, but the peaks varied between vessel types. FTs exhibited an effort and landings peak in 2011 and RSW and ice boats exhibited landings peaks in 2018 (Figure 12). During 2017 and 2018, landings by RSW boats reached their highest levels of the RSW time-series; a period when the total LPUE landings were dominated by landings from RSW boats (Figure 9). These landings and effort trends translated into the highest nominal LPUE (mt/df) indices of the RSW and FT time-series during 2017 and 2018 (Figure 9).

LPUE Indices

The results of the GLM model runs for annual LPUE standardization are summarized in Table 2. All model runs converged and all main effects were significant at the 5% alpha level except for vessel type in the normal error structure model. The model with a negative binomial error structure showed the best fit to the LPUE data according to the deviance/degrees of freedom values (Table 2). Based on the AIC values for this model run, the three-factor model that included year, week and permit provided the best fit for the initial model runs. AIC values were identical for the negative binomial and gamma models, which can occur due to the high number of degrees of freedom and given the high flexibility of the gamma distribution as implemented here with SAS GENMOD, which can allow it to mimic the negative binomial. LPUE estimates for the best fit model for all three error types showed similar trends with the exception of 2018, which showed a decrease for the negative binomial model and an increase for the normal error structure model (Figure 13).

Additional model runs that included Statistical Area showed that this factor was not statistically significant at the 5% alpha level. However, the data associated with the largest Statistical Area coefficient were investigated further by examining the associated VTR images. Misreporting of a Statistical Area for a single vessel during most of an entire season was found and corrected in the LPUE dataset prior to development of a four-factor model that included Statistical Area. The results indicated that Statistical Area was statistically significant for all three error type models and improved the fit of the three-factor negative binomial model. The negative binomial model that included year, week, permit and Statistical Area provided diagnostics indicating the best fit (Table 2). LPUE estimates were fairly precise for most of the time-series, but were lowest during 1997-1998 and 2006-2009 (Figure 13). The standardized LPUE indices and the NEFSC fall survey biomass indices (stratified mean kg per tow) showed some similarities in trends (Figure 14) and

Working Paper Not to be cited without author’s permission were moderately correlated (r = 0.577, p = 0.006) during 1997-2018, but were highly correlated during 2008-2018 (r = 0.845, p = 0.002).

Trip limits were imposed when plant processing capacity was reached (Wayne Reichle pers. comm.). Nominal LPUE and the standardized LPUE estimates were impacted by these trip limits. The unaccounted for landings resulted in an underestimation of Illex LPUE indices. Unfortunately, these impacts could not be quantified because the trip limit quantities and dates of implementation, by vessel, were unknown. Any future implementation of trip limits in the Illex fishery should consider these impacts on Illex biomass estimates derived using LPUE.

Next steps

With respect to the fishery body size data, a spatial analysis will be conducted to determine their temporal and spatial representativeness. This will require obtaining the Statistical Areas for each subtrip and trip by matching the date landed fields in the body weight dataset and the VTR database.

With respect to LPUE analyses, the next steps will include estimation of annual discards for the time-series and estimation of the 2019 LPUE index when the data are available. In addition, standardized LPUE indices will be derived for week of the year, for weeks 17-45, during 1997- 2019. Variability of LPUE in relation to changes in fishing location will also be examined and further analysis of the 2017-2019 VMS data in relation to the VTR effort and LPUE data will be conducted.

References

Arkhipkin, A. I., L. C. Hendrickson, I. Paya´, G. J. Pierce, R. H. Roa-Ureta, J.-P. Robin and A. Winter. In Press. Stock assessment and management of cephalopods: advances and challenges for short-lived fishery resources. ICES Journal of Marine Science,doi:10.1093/icesjms/20fsaa038.

Arkhipkin A. I., Paul G. K. Rodhouse, Graham J. Pierce, Warwick Sauer, Mitsuo Sakai, Louise Allcock, Juan Arguelles, John R. Bower, Gladis Castillo, Luca Ceriola, Chih-Shin Chen, Xinjun Chen, Mariana Diaz-Santana, Nicola Downey, Angel F. Gonz_alez, Jasmin Granados-Amores, Corey P. Green, Angel Guerra, Lisa C. Hendrickson, Christian Ib_a~nez, Kingo Ito, Patrizia Jereb, Yoshiki Kato, Oleg N. Katugin, Mitsuhisa Kawano, Hideaki Kidokoro, Vladimir V. Kulik, Vladimir V. Laptikhovsky, Marek R. Lipinski, Bilin Liu, Luis Mari_ategui, Wilbert Marin, Ana Medina, Katsuhiro Miki, Kazutaka Miyahara, Natalie Moltschaniwskyj, Hassan Moustahfid, Jaruwat Nabhitabhata, Nobuaki Nanjo, Chingis M. Nigmatullin, Tetsuya Ohtani, Gretta Pecl, J. Angel A. Perez, Uwe Piatkowski, Pirochana Saikliang, Cesar A. Salinas-Zavala, Michael Steer, Yongjun Tian, Yukio Ueta, Dharmamony Vijai, Toshie Wakabayashi, Tadanori Yamaguchi, Carmen Yamashiro, Norio Yamashita, and Louis D. Zeidberg. 2015. World Squid Fisheries. Reviews in Fisheries Science and Aquaculture, 23:2, 92-252, doi: 10.1080/23308249.2015.1026226.

Working Paper Not to be cited without author’s permission Basson, M., J. R. Beddington, J. A. Crombie, S. J. Holden, L. V. Purchase and G. A. Tingley. 1996. Assessment and management techniques for migratory annual squid stocks: the Illex argentinus fishery in the Southwest Atlantic as an example. Fisheries Research, 28: 3–27.

Brodziak, J.K.T. and L.C. Hendrickson. 1999. An analysis of environmental effects on survey catches of squids Loligo pealei and Illex illecebrosus in the northwest Atlantic. Fish. Bull. (U.S.) 97: 9-24. Dawe, E.G., and P.C. Beck. 1997. Population structure, growth and sexual maturation of short- finned squid (Illex illecebrosus) at Newfoundland. Can. J. Fish. Aquat. Sci., 54: 137-146.

Hendrickson, L.C. 2004. Population biology of northern shortfin squid (Illex illecebrosus) in the northwest Atlantic Ocean and initial documentation of a spawning site in the Mid-Atlantic Bight (USA). ICES J. Mar. Sci. 61: 252-266.

Hendrickson L. C. and D. R. Hart. 2006. An age-based cohort model for estimating the spawning mortality of semelparous cephalopods with an application to per-recruit calculations for the northern shortfin squid, Illex illecebrosus. Fish. Res., 78: 4–13.

Hendrickson, L. C. and E. M. Holmes. 2004. Essential fish habitat source document: northern shortfin squid, Illex illecebrosus, life history and habitat characteristics, Second Edition. NOAA Tech. Memo. NMFS-NE-191. 36 p.

Hendrickson L. C. and M. A. Showell. 2019. 2019 Assessment of Northern Shortfin Squid (Illex illecebrosus) in Subareas 3+4. NAFO SCR Doc. 19/042, Serial No. N6973, 38 p.

Hendrickson L. C., E. G. Dawe and M. A. Showell. 2001. Assessment of Subarea 3+4 Northern shortfin squid (Illex illecebrosus) for 2000. NAFO SCR Doc. 01/61, Serial No. N4439, 13 p.

Hendrickson, L.C., J. Brodziak, M. Basson, and P. Rago. 1996. Stock assessment of Northern shortfin squid, Illex illecebrosus, in the Northwest Atlantic during 1993. Northeast Fish. Sci. Cent. Ref. Doc. 96-05g; 63 p.

Hendrickson, L. C., D. A. Hiltz, H. M. McBride, B. M. North and J. E. Palmer. 2003. Implementation of electronic logbook reporting in a squid bottom trawl study fleet during 2002. Northeast Fish. Sci. Cent. Ref. Doc. 03-07. 30 p.

NEFSC [Northeast Fisheries Science Center]. 1999. Report of the 29th Northeast Regional Stock Assessment Workshop (29th SAW): Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. Northeast Fisheries Science Center Ref. Doc. 99-14; 347 p.

NEFSC [Northeast Fisheries Science Center]. 2003. Report of the 37th Northeast Regional Stock Assessment Workshop, Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. NEFSC Ref. Doc. 03-16. 597 p.

Working Paper Not to be cited without author’s permission NEFSC [Northeast Fisheries Science Center]. 2006. 42nd Northeast Regional Stock Assessment Workshop (42nd SAW) stock assessment report, part A: silver hake, Atlantic mackerel, and northern shortfin squid (CRD 06-09a). Northeast Fish. Sci. Cent. Ref. Doc. 06-09a; 284p.

Powell, E. N., A. J. Bonner, B. Muller and E. A. Bochenek. 2003. Vessel time allocation in the US Illex illecebrosus fishery, Fisheries Research, 61: 35–55.

Working Paper Not to be cited without author’s permission

Figure 1. Box and whiskers plot of Illex illecebrosus body weight (g), by week of the year during 1997-2019, based on samples from Illex fishery landings. Body weight data for 1999-2003 were obtained from a combination of Illex squid processors and a collaborative fishery data collection study, whereas data for other years were obtained from one or both of the predominant Illex processors depending on year. During 1997-2019, the fishery occurred between weeks 17 and 45. The red dots represent values greater than the largest value; up to 1.5 times the interquartile range.

Working Paper Not to be cited without author’s permission

Figure 2. Samples sizes of Illex illecebrosus body weights (g, top panel) and dorsal mantle lengths (cm) from the landings sampled by Illex processors (top panel) and NMFS port samplers (bottom panel). The data were used to compute mean body weights (g). Mean body weight for the port sampler data were computed by dividing the sample weight by the number of lengths in the sample.

Working Paper Not to be cited without author’s permission

Figure 3. Illex illecebrosus mean body weights (g) for landings data sampled by NFMS port samplers (top panel) and Illex processors (middle panel) in comparison to stratified mean body weights (g) for Illex samples collected during NEFSC fall bottom trawl surveys (bottom panel) during 1997-2019. Mean body weight was not computed for the 2017 fall survey due to a lack of sampling a majority of I. illecebrosus habitat.

Working Paper Not to be cited without author’s permission

Figure 4. Illex illecebrosus body weights (g), by week of the year, for landings sampled by Illex processors during 1997-2004. Fishery closures occurred during 1998 and 2004.

Working Paper Not to be cited without author’s permission

Figure 5. Illex illecebrosus body weights (g), by week of the year, for landings sampled by Illex processors during 2005-2006 and 2009-2012. Data were not available for 2007 and 2008.

Working Paper Not to be cited without author’s permission

Figure 6. Illex illecebrosus body weights (g), by week of the year, for landings sampled by Illex processors during 2013-2019. Fishery closures occurred during 2017-2019.

Working Paper Not to be cited without author’s permission

Figure 7. Generalized additive mode smooths and 95% confidence intervals for Illex illecebrosus body weight (g), by week of the year, for landings sampled by Illex processors during 1997-1999, 2001-2006 and 2009-2019. Fishery closures occurred during 1998, 2004 and 2017-2019.

Working Paper Not to be cited without author’s permission

Figure 8. Proportion of the total annual Illex illecbrosus landings for trips available for computing landings-per-unit-of-effort (LPUE; dashed line) and total landings (mt), landings available for computing LPUE (blue line) and landings used to compute LPUE (top panel) and numbers of trips, vessels and nominal days fished associated with the LPUE dataset (bottom panel) for 1997-2018. The black dashed line represents the mean days fished.

Working Paper Not to be cited without author’s permission

Figure 9. Numbers of vessels and trips (top panels), number of nominal days fished and percentage of annual landings used to estimate LPUE, by vessel type (FT=freezer trawler, RSW=refrigerated seawater and ICE boats; bottom panels), for directed Illex illecebrosus trips during 1997-2018.

Working Paper Not to be cited without author’s permission

Figure 10. Average trip duration (days; top panel), crew size (middle panel) and nominal days fished, by vessel type (FT=freezer trawler, RSW=refrigerated seawater and ICE boats; bottom panel) for directed Illex illecebrosus trips during 1997-2018.

Working Paper Not to be cited without author’s permission

Figure 11. Distribution of Illex illecebrosus landings, by ten-minute square, reported in the Vessel Trip Reports during 2019. Landings are shown as cumulative percentages, by quartile, such that the red squares contain the highest amount of landings (top 25%).

Working Paper Not to be cited without author’s permission

Figure 12. Trends in effort (days fished) and landings (mt) data used to estimate Illex illecebrosus landings-per-unit-of-effort (LPUE mt landed/day fished; top panels) and nominal LPUE (mt/day fished and mt/day absent from port; bottom panels), by vessel type (FT=freezer trawlers, RSW=refrigerated seawater boats and ICE boats), during 1997-2018.

Working Paper Not to be cited without author’s permission

Figure 13. Illex illecebrosus landings (mt), days fished and number of trips for the dataset used to compute landings-per-unit-of-effort (LPUE, mt/day fished) (top), scaled nominal and standardized LPUE indices estimated from Generalized Linear Models with normal, gamma and negative binomial error structures (middle) and scaled standardized LPUE indices and their 95% confidence limits from the negative binomial (best fit) model (bottom) for 1997-2018.

Working Paper Not to be cited without author’s permission

Figure 14. Nominal and standardized LPUE (red line) indices (mt/day fished) for Illex illecebrosus in relation to stratified mean kg per tow I. illecebrosus indices derived from NEFSC fall bottom trawl surveys during 1997-2018. The 2017 fall survey index was not computed due to a lack of sampling a majority of I. illecebrosus habitat. All indices are scaled to their means.

Working Paper Not to be cited without author’s permission Table 1. Timeline of research projects and stock assessment improvements regarding Illex illecebrosus southern stock component. “SAS” represents the Illex stock assessment scientist.

Year Illex Research and Stock Assessment Improvements Conducted Citation

Hendrickson et 1995 Stock assessment - recommended plan for in-season stock ssessment al. (1996); NEFSC (1996) 1996 Doubled number of fishery lengths per port sample

and changed length sample quotas from quarterly to monthly Increased number of Illex trips sampled by observer program and

included freezer trawlers

SAS visited Illex processors in RI, NJ and VA; SAS and Rutgers student collected Illex fishery data on J. Ruhle's vessel; SAS began creating mean body weight dataset from processor body 1997 weight data from Seafreeze andduring some yrs from Lunds Lunds boats implemented fishery start date of June 21 to allow body weight to reach 100 g

Brodziak and Environmental effects on Illex indices for the NEFSC fall surveys; Hendrickson SAS held squid acoustics workshop with experts from Falklands and (1999) 1998 Canada; Illex fishery closure and industry requests in-season assessment and adaptive quotas

SAS obtained grant and implemented coop. res. study fleet program for collection of real-time, at-sea tow-based catch, effort, loc. and biol. Data for in-season stock assessment Stock assessment - 1999 real-time fishery data incorporated into weekly 1999 depletion model and fishery mean body weight trends and delayed fishery NEFSC (1999) opening effects on YPR assessed

SAS created website to post Illex study fleet results Study fleet workshop held with fishermen to summarize results and plan next steps for 2000 SAS obtained grant and implemented first industry-based survey - Hendrickson 2000 resulted in pre-fishery stock size estimate, discovery of Illex spawning (2004) grounds, first US age and growth study

SAS obtained grant and worked with NEFSC programmers to develop first e-VTR platform on US E. Coast - password-protected website, queriable LPUE maps and ability to print customized reports; Hendrickson et 2002 SAS created Boatracs trip, catch and haul macros and received grant to al. (2003) pay for Boatracs emails sent from sea to NEFSC in real-time and then loaded into Oracle tables; grant also funded gear-mounted depth- temperature loggers for study fleet vessels

Working Paper Not to be cited without author’s permission

Table 1 (cont.)

Stock assessment - included depletion model and new weekly maturation- NEFSC (2003); 2003 natural mortality model that fed into a weekly per-recruit model for Hendrickson semelparous species; end of study fleet project and Hart (2006)

Updated Illex Essential Fish Habitat (EFH) source document; Hendrickson 2004 Second Illex fishery closure and Holmes (2005) Stock assessment – improved upon weekly depletion model and weekly 2006 maturation-natural mortality model that fed into the per-recruit model for NEFSC (2006) semelparous species

Working Paper Not to be cited without author’s permission Table 2. Goodness-of-fit results for the GLM models used to estimate standardized Illex illecebrosus LPUE indices for 1997-2018. The best goodness-of-fit for each of the three model types is shown in bold-faced text and the final model run is shaded gray.

Normal Log- Model Deviance/DF Likelihood AIC Converge Effects Significant (Neg Hess PD) at 5% Year 0.8012 -4038 8121 Y Y Year-Week 0.7890 -4000 8101 Y Y Year-VessT 0.7288 -3890 7830 Y Y Year-Permit 0.5215 -3343 6849 Y Y Year-Week-VessT 0.7091 -3833 7773 Y Y Year-Week-Permit 0.4911 -3236 6691 Y Y Year-VessT-Permit 0.5208 -3341 6846 Y Y Year-Week-Permit-Area 0.4825 -3200 6651 Y Y Year-Week-VessT-Permit 0.4907 -3235 6689 Y N - VessT

Gamma Log- Model Deviance/DF Likelihood AIC Converge Effects Significant (Neg Hess PD) at 5% Year 0.8011 -41158 82362 Y Y Year-Week 0.7877 -41113 82328 Y Y Year-VessT 0.7162 -40965 81979 Y Y Year-Permit 0.5336 -40437 81036 Y Y Year-Week-VessT 0.6993 -40908 81923 Y Y Year-Week-Permit 0.5110 -40349 80917 Y Y Year-VessT-Permit 0.5331 -40435 81034 Y Y Year-Week-Permit-Area 0.5014 -40309 80868 Y Y Year-Week-VessT-Permit 0.5105 -40347 80914 Y Y

Negative Binomial Log- Model Deviance/DF Likelihood AIC Converge Effects Significant (Neg Hess PD) at 5% Year 1.1219 9245219333 82362 Y Y Year-Week 1.1298 9245219378 82328 Y Y Year-VessT 1.1122 9245219526 81979 Y Y Year-Permit 1.1083 9245220054 81036 Y Y Year-Week-VessT 1.1196 9245219582 81923 Y Y Year-Week-Permit 1.1139 9245220141 80917 Y Y Year-VessT-Permit 1.1086 9245220056 81034 Y Y Year-Week-Permit-Area 1.1182 9245220182 80868 Y Y Year-Week-VessT-Permit 1.1141 9245220144 80914 Y Y