Tracking Atlantic Bluefin Tuna from Foraging Grounds Off the West Coast

Home , Mackerel

ICES Journal of Marine Science (2020), doi:10.1093/icesjms/fsaa090 Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020

Tracking Atlantic blueﬁn tuna from foraging grounds off the west coast of Ireland

Thomas W. Horton 1,2,3, Barbara A. Block4, Alan Drumm5, Lucy A. Hawkes2, Macdara O’Cuaig5, Niall O´ Maoileídigh5,RossO’Neill5, Robert J. Schallert4, Michael J. W. Stokesbury6,and Matthew J. Witt1,2* 1Environment and Sustainability Institute, University of Exeter, Penryn Campus, Penryn, Cornwall TR10 9FE, UK 2Hatherly Laboratories, College of Life and Environmental Sciences, University of Exeter, Prince of Wales Road, Exeter EX4 4PS, UK 3Centre for Environment, Fisheries and Aquaculture Science, Pakefield Road, Lowestoft NR33 0HT, UK 4Tuna Research and Conservation Center, Stanford University, Hopkins Marine Station, Pacific Grove, CA, USA 5Marine Institute, Newport Co. Mayo F28 PF65, Ireland 6Department of Biology, Acadia University, 33 University Avenue, Wolfville, Nova Scotia B4P 2R6, Canada *Corresponding author: tel: (þ44) 01392 722268; e-mail: [email protected]. Horton, T. W., Block, B. A., Drumm, A., Hawkes, L. A., O’Cuaig, M., O´ Maoileídigh, N., O’Neill, R., Schallert, R. J., Stokesbury, M. J. W., and Witt, M. J. 2020. Tracking Atlantic bluefin tuna from foraging grounds off the west coast of Ireland. – ICES Journal of Marine Science, doi:10.1093/ icesjms/fsaa090. Received 1 October 2019; revised 7 April 2020; accepted 18 April 2020.

Pop-up archival tags (n ¼ 16) were deployed on Atlantic bluefin tuna (ABT) off the west coast of Ireland in October and November 2016 (199–246 cm curved fork length), yielding 2799 d of location data and 990 and 989 d of depth and temperature time-series data, respectively. Most daily locations (96%, n ¼ 2651) occurred east of 45W, the current stock management boundary for ABT. Key habitats occupied were the Bay of Biscay and the Central North Atlantic, with two migratory patterns evident: an east-west group and an eastern resident group. Five out of six tags that remained attached until July 2017 returned to the northeast Atlantic after having migrated as far as the Canary Islands, the Mediterranean Sea (MEDI) and the Central North Atlantic. Tracked bluefin tuna exhibited a diel depth-use pattern occupying shallower depths at night and deeper depths during the day. Four bluefin tuna visited known spawning grounds in the central and western MEDI, and one may have spawned, based on the recovered data showing oscillatory dives transecting the thermocline on 15 nights. These findings demonstrate the complexity of the aggregation of ABT off Ireland and, more broadly in the northeast Atlantic, highlighting the need for dedicated future research to conserve this important aggregation. Keywords: Atlantic bluefin tuna, diving behaviour, electronic tagging, habitat use, Ireland, migration, Northeast Atlantic

Introduction et al., 2019) with catch indices from Japanese longline fleets work- Atlantic bluefin tuna (Thunnus thynnus, hereafter ABT) are highly ing in the northeast Atlantic also indicating a positive change in migratory, endothermic predators that range widely throughout catchability (Kimoto and Itoh, 2017). However, the underlying the North Atlantic Ocean (Block et al., 2005). They were once fre- mechanisms behind these fluctuations in spatial distribution are quently encountered along the western coasts of Ireland until complex and remain unclear (Ravier and Fromentin, 2004; 2005 (Cosgrove et al., 2008), before becoming regionally scarce. Fromentin, 2009). In recent years, ABT have reappeared in coastal and offshore wa- The ABT population is comprised two or more genetically dis- ters off Ireland (O´ Maoileídigh et al., 2018), the United Kingdom, tinct spawning stocks (Rooker et al., 2008; Rodrı´guez-Ezpeleta Sweden, Denmark (MacKenzie et al., 2020) and Norway (Ferter et al., 2019): the “eastern stock” spawns in the Mediterranean Sea

(MEDI) (Abascal et al., 2016) and the “western stock” spawns in from the study animal if they remained at a constant depth the Gulf of Mexico (Wilson et al., 2015). ABT larvae and mature (62.5 m) for a period of 4 d, which may indicate death or prema- adult fish (with fully developed gonads) have also been found in ture detachment. the Slope Seas between the Gulf Stream and the northeast US Animal locations were reconstructed using the Global Position continental shelf seas, indicating that ABT may use other, lesser- Estimator 3 (GPE3, Wildlife Computers), which uses the tag known spawning grounds in the North Atlantic Ocean records of light, temperature, depth, and reference data on sea (Richardson et al., 2016). Throughout their range, ABT stocks surface temperature (NOAA OI SST, psl.noaa.gov/data) and ba- have been subjected to over-exploitation (Taylor et al., 2011). In thymetry (ETOPO1-Bedrock, Amante and Eakins, 2009) with a 1 2007, after considerable stock depletion, a multi-annual stock re- user-defined movement speed (set at 2 m s ) to determine the Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020 building programme was implemented by the International most likely location. For each tag, up to seven estimates of loca- Commission for the Conservation of Atlantic Tunas (ICCAT). tion were available per day, which were averaged (geodesic mean) The most recent ICCAT stock assessment suggested that the east- to create “daily locations”. All reported locations are most likely ern stock had grown “substantially” (ICCAT, 2017), leading to a locations and are subject to positional error (grand mean for all revision of the recovery plan in 2018 and staged increases in total tags of 1.5 6 0.1 longitude and 1.2 6 0.1 latitude). allowable catches up to 36 000 tonnes in 2020 (ICCAT, 2018). The status of the western Atlantic stock is more uncertain, and re- Behaviour classification cent research has indicated that recovery of the eastern stock may ABT that were tracked for longer than 60 d (n ¼ 14) were classified have resulted in an increase in mixing rates between sub- into two types based on whether they crossed the Mid-Atlantic populations (Hanke et al., 2017). Ridge (median longitude 28W, “east-west”) or not (eastern resi- To avoid their repeated over-exploitation, understanding the dent). Periods where distances between successive relocations indi- complex spatio-temporal life-history of ABT is key. ABT move- cated faster movements were classified as “fast migration” using ments appear to vary ontogenetically, with larger ABT ranging fur- the adehabitatLT package in R (Calenge, 2006; further details of ther and occupying more northerly regions (Block et al., 2005; classification methods are supplied in the Supplemental Materials) Walli et al., 2009). There may also be differences in the movement and vertical movements analysed separately. ABT with tags that of ABT from different stocks (Aranda et al., 2013; Fromentin et al., remained attached after 1 July 2017 (the year following tag attach- 2014a; Wilson et al., 2015). Otolith microchemistry and tracking ment and the time of year schools of ABT begin to be observed off studies have linked ABT present in northeast Atlantic aggregations west Ireland, A. Molloy, pers. comm.) were classified as return to both eastern (east of the 45W meridian) and western (west of migrants if either tag pop-up or daily locations were present in the the 45W meridian) stock management areas, and to spawning area east of 30W and north of 50N. grounds in the MEDI (Block et al., 2005; Stokesbury et al., 2007; Rooker et al.,2019). Whilst research into the underlying drivers be- Horizontal movements hind the changing abundance of ABT in the northeast Atlantic Areas of relative importance for ABT were determined by divid- continues (e.g. Fromentin et al.,2014b; Faillettaz et al., 2019), ing the number of summed daily locations by counts of unique aspects of the genetic provenance, migration patterns, and putative tags in 100 km diameter hexagonal bins (11 555 km2 per hexagon; spawning behaviour of ABT that seasonally reside in the northeast maximum count 16). The study area was partitioned into ecore- Atlantic remain unclear. In the present study, we build on the work gions following Longhurst (2010, Fig. S1). detailed in Stokesbury et al. (2007) to further investigate the movements, habitat preferences, and area-specific behaviours of poten- Diving behaviour from time series data tially sexually mature ABT captured on their seasonal foraging After detachment, tags transmitted each 8-h time-series segment grounds off the west coast of Ireland. across three data messages, which were often only partially recovered via satellite. Mean depth values derived from incomplete Methods transmitted time series (one or two messages received) were Electronic tagging found to differ from values derived from downloaded tag Between October and November 2016, ABT (n ¼ 16, mean size archives, where tags were physically retrieved (see Supplementary 220 6 13 cm, 1 Standard Deviation, curved fork length, CFL) Table S1 and Supplementary Figure S3 for details), and hence were captured off the west coast of Ireland by recreational “rod were removed from future time-series analysis. In addition to us- and line” fishermen trolling lures. Tagging was conducted on- ing transmitted time series (three data messages received), down- deck, during which a saltwater hose was used to irrigate the gills loaded datasets from three recovered tags were down sampled to and a cloth soaked in fish-slime replacement (PolyAqua) was match transmitted time-series frequency and included in diving placed over the eyes to reduce stress. Electronic tags (Wildlife analyses. Time-series data for each tag were subset into day-time Computers MiniPAT 247A and 348F, tagware v2.4n, hereafter and night-time summary periods using tag-derived sunrise and “tags”) were attached via percutaneous darts as detailed in sunset times. Diving metrics (mean depth, vertical movement Wilson et al. (2015) and programmed to detach from ABT after rate, mean temperature, and rate of ambient temperature change) 316–365 d. Tags recorded light, pressure (depth) and temperature were then calculated for each day-time or night-time period for every 15 s for model 247A tags (n ¼ 8) and every 5 s for model each tag. Vertical movement rate and rate of ambient temperature 348F tags (n ¼ 8). The entire procedure (removal from the water change were calculated by summing the absolute depth and tem- to release) took 3–5 min. After detachment, tags were pro- perature change, respectively, and dividing by the time elapsed in grammed to transmit 8-h-long segments of depth or temperature hours for a given summary period (either day time or night time-series data at a 10-min resolution (hereafter “transmitted time). Generalized linear mixed models (GLMMs, gamma family) time series”, n ¼ 9 tags). All tags were programmed to release were fit to log-normalized depth and temperature dive metrics, Tracking Atlantic bluefin tuna from Ireland 3 with fixed terms for ecoregion and day or night and tag as a ran- travelled by an ABT (234 cm CFL) in a single day was 276 km (1.4 dom effect using the package “lme4” (Bates et al., 2014). The body lengths s1). ABT exhibited a fast migration behaviour for most appropriate model was selected by removing individual between 2 and 50 d per tag (mean 15 6 12 d tag1), which was fixed effects and comparing with the null model using a likeli- observed in five out of eight ecoregions (CNRY, MEDI, NADR, hood ratio test. t-Tests using Satterwaite’s method were used to NASE, and NECS), with the highest proportion occurring in the test the differences between fixed-level effects. The final model MEDI and CNRY ecoregions (30 and 13% of daily movements in was validated by visually inspecting standardized residuals. All each region, respectively, Table 2). errors are reported as one standard deviation. In addition to investigating general behaviours over longer periods (hours), puta- Vertical movements Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020 tive spawning behaviour was investigated using the downloaded The dataset of complete time-series dive data comprised 990 and archive of a recovered tag. For this analysis, rates are reported at 989 d of depth and temperature time-series data, respectively the base sampling rate of 5 s. (36% of days with geolocation data). Dive data were collected in Results all ecoregions except the NASW, and only one ABT occupied the Atlantic Subarctic (SARC) for a period of 11 d. Due to the paucity Fieldwork and tag performance of data in these two regions, they were excluded from modelling. The mean tag retention time was 224 6 99 d (n ¼ 15, Supplementary Figure S4), with three tags remaining attached for Depth preferences from time-series data the entire programmed attachment period (307–365 d). One tag Tracked ABT occupied the shallowest depths in the MEDI ecore- detached following a putative mortality event, with the tag gion (19 6 19 m, t ¼ 0.1, p ¼ 0.32), which were similar to depths remaining at a constant depth for 4 d (14P0251, Supplementary occupied in the CNRY (21 6 3m, Table 2) and Gulf Stream; N. Figure S6), and one tag only transmitted for 7 h post- Atlantic Subtropical Gyre (West) (GFST) ecoregions (27 6 19 m, detachment. Three tags were physically recovered, and raw time t ¼ 0.42, p ¼ 0.68). ABT occupied significantly greater depths in series data were downloaded. The resulting dataset comprised the NADR (49 6 22 m, t ¼ 7.33, p 0.001), NASE (45 6 11 m, 2779 d (n ¼ 14 tags) of geolocation data. t ¼ 5.24, p 0.001) and NECS (33 6 7m, t ¼ 5.6, p 0.001) ecoregions. ABT occupied significantly shallower depths during Horizontal movements the night whilst in the MEDI (day 24 6 24 m, night 13 6 3m,t ¼ Tracked ABT dispersed up to 4628 km from the tagging site (cu- 2.3, p ¼ 0.02), NADR (day 64 6 41 m, night 32 6 11 m, t ¼ mulative along-track straight-line distance, mean 2780 6 721 km, 2.5, p ¼ 0.01), and NECS (day 39 6 11 m, night ¼ 24 6 10 m, t Figure 1a), but most remained in the eastern Atlantic, with 96% ¼4.7, p ¼ <0.001) ecoregions. ABT vertical movement rate of daily locations occurring east of the 45 W meridian. No ABT was positively correlated with mean occupied depth (Spearman’s moved north immediately after tagging, and 98% of all daily loca- rank, S ¼ 258214, q ¼ 0.78, p ¼ <0.001). Low ABT vertical move- tions were south of the tagging site. ABT moved west into sover- ment rates (<15 m h1) were observed in all ecoregions and 90% eign waters of the United States and Canada, as far south as the of vertical movement rates were <186 m h1 (Supplementary Canary Islands, as far east as the coast of Libya, and as far north Figure S8). as the Faroe Islands, as well as visiting known spawning grounds in the eastern and central MEDI (Figure 1a). Eight ABT travelled Temperature preferences from time-series data to the Bay of Biscay, where they either remained for 26–107 d (range, mean 48 6 30 d) or migrated west to the Central North Mean ambient temperature experienced by ABT differed signifi- Atlantic (Figure 1b). ABT occupied eight different ecoregions in cantly between all ecoregions except the CNRY and GFST 6 6 the North Atlantic and MEDI (Figure 2). Areas of high relative (17.3 1.7 and 17.7 0.7 C respectively, t ¼1.6, p ¼ 0.64, Table 2) and the GFST and NASE ecoregions (17.7 6 0.7 and mean residency for ABT (Figure 1c) were in the Celtic Sea and Goban Spur area [maximum 19 d per tag per grid cell, NE 17.3 6 1.1 C, respectively, t ¼0.42, p ¼ 0.99). Coolest ambient Atlantic Shelves (NECS) ecoregion], the Bay of Biscay [maximum temperatures were experienced by ABT in the SARC (day 12.0 C, 18 d per tag per grid cell, N. Atlantic Drift (NADR), NECS and night 12.5 C) and NECS ecoregions (day 12.9 6 1.1 C, night Canary Coastal (CNRY) ecoregions] and central North Atlantic/ 13.2 6 1.2 C) and the warmest in the MEDI ecoregion (day Flemish Cap region [maximum 12 d per tag per grid cell, NADR 21.0 6 3 C, night 21.8 6 1.7 C). In addition, mean temperatures and N. Atlantic Subtropical Gyre (East) (NASE) ecoregions]. Ten occupied by ABT were significantly cooler between day and night ABT (77%) exhibited an east-west migratory pattern, crossing the periods for all ecoregions (GLMM, cooler by 0.2 C, df ¼ 1, Mid-Atlantic Ridge into the Newfoundland Basin, and three ABT F ¼ 11.6, p 0.001) and were positively correlated with the rate 4 (23%) exhibited an eastern resident migratory pattern, remaining of ambient temperature change (Spearman’s rank, S ¼ 53 521 , in waters east of 20W and making latitudinal migrations q ¼ 0.35, p ¼ <0.001). Low rates of ambient temperature change 1 (Figure 3). Six tags remained attached after 1 July 2017, of which (<0.5 Ch ) were observed in ABT occupying every ecoregion 1 five (83%) showed a return migration to the northeast Atlantic and 90% of ambient temperature change rates were <3.1 Ch (Table 1), with the sixth migrating to the Scotian Shelf where the (Supplementary Figure S8). tag detached on 5 July 2017. The grand mean speed of travel for all tracked ABT was Spawning ground visitation and behaviour 46 6 9 km day1. Ninety-three percent of pooled daily move- During the present study, four ABT entered the MEDI between ments for ABT were <125 km day1 (n ¼ 2570 d) with 67% May and July—the known period for spawning (Aranda et al., <50 km day1 (n ¼ 1850 d, Supplementary Figure S2). The maxi- 2013). ABT entered through the Straits of Gibraltar between 16 mum distance (along track straight line, range <1–276 km) May 2017 and 23 May 2017 (mean 19 May 2017), but only one 4 T. W. Horton et al. Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020

Figure 1. Blueﬁn tuna horizontal movements in the North Atlantic. (a) Estimated daily locations obtained from 14 electronic tags attached to ABT in 2016 (n ¼ 2779 tracking days). (b) Hundred-kilometre hexagon grid showing the number of unique tags in each grid cell for the 14 tags that yielded data and (c) 100-km hexagon grid showing mean residency of tagged ABT (days per tag). Black broken line at the 45W meridian denotes the ICCAT stock delimitation line, and blue broken line denotes the Mid-Atlantic Ridge. White broken line (b and c) denotes 200 m depth contour. Tracking Atlantic blueﬁn tuna from Ireland 5 Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020

Figure 2. Seasonal occupancy of North Atlantic ecoregions by Atlantic blueﬁn tuna. (a) Map of the North Atlantic displaying estimated daily locations obtained from 14 electronic tags attached to ABT in 2016. Black broken line at the 45W meridian denotes the ICCAT stock delimitation line, and blue broken line denotes the Mid-Atlantic Ridge. (b) A Gantt chart displaying the temporal pattern of ecoregion usage for all electronically tagged ABT. Numbers above the plot denote the total number of active tags, and ﬁlled circles indicate the number of active tags in each ecoregion, both at a weekly resolution. Fast migration—periods where distances between successive relocations indicate faster movements (thick black lines).

(16P1265) was tracked returning to the North Atlantic after 47 d and returned on 21 July 2017 after having spent 47 d in the MEDI residency (exit on 6 July 2017, see below). The third ABT over the spawning season. Archival data (5-s resolution) reveal (16P1170) experienced a similar temperature profile to 16P1265 that the fish undertook high-frequency shallow (10 6 10 m) div- indicating entry to the MEDI but lacked light data to reconstruct ing around the thermocline on 15 occasions over two periods the track beyond 3 June 2017 (Supplementary Figure S7). Two (the 3 May 2017–5 May 2017 and 14 May 2017–24 June 2017), eastern resident ABT migrated to the central MEDI and two east- between 00:00 and 04:00 (UTC), whilst off the Balearic Islands west ABT migrated to the western MEDI (Figure 3). Two of these (Figure 5). Rates of vertical ascent/descent peaked at 9 m 5 s1 four tags were physically recovered. Depth data from the first tag (mean 1.1 6 1.1 m 5s1) and the fish experienced temperature (14P0031) suggest that the fish was caught near the Strait of fluctuations of up to 2.9C5s1 (mean 0.2 6 0.3C5s1). Messina on 4 June 2017 and potentially towed to a fish farm off Outside of these periods, whilst still in the MEDI, the fish occu- Malta, where the tag was recovered 76 d later. The second tag pied depths of 13 6 30 m and experienced mean ambient temper- (16P1265) detached from the fish after collecting a dataset over a atures of 21.5 6 2.8C with mean rates of vertical ascent/descent full migratory cycle. The fish (234 cm CFL) moved an estimated of 0.8 6 0.9 m 5 s1 and rates of temperature change of minimum straight-line distance of 17173 km over 307 d 0.1 6 0.2C5s1. Tags attached to the other two ABT that en- (Figure 4). The ABT departed the NECS on 7 November 2016 tered the MEDI detached 130 km off the Libyan coast on 25 June 6 T. W. Horton et al. Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020

Figure 3. Migration patterns and spawning ground visitation of Atlantic blueﬁn tuna. (a) East-west ABT that exhibited movements crossing the Mid-Atlantic Ridge, (b) eastern resident ABT that remained east of the Mid-Atlantic Ridge, and (c) ABT from both categories that visited previously described spawning areas in the Mediterranean Sea (unique colours for tags are the same for all plots). Black broken line at the 45W meridian denotes the ICCAT stock delimitation line, and blue broken line denotes the Mid-Atlantic Ridge. Tracking Atlantic blueﬁn tuna from Ireland 7

2017 (14P0330), and 300 km southeast of Iceland on 1 September 2017 (16P1264). The coarseness of transmitted depth and temperature time-series data received from the tags prevents an in-

Xmit days (light days) vestigation of putative spawning behaviour for these ﬁsh. Discussion

Return The distribution of blueﬁn tuna in the Atlantic has ﬂuctuated migrant? markedly over time (Fromentin et al., 2014b), and positive abun-

dance trends indicate a recent resurgence in northeast Atlantic Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020 (Kimoto and Itoh, 2017; Faillettaz et al., 2020). Consequently, understanding the movements and habitat selections (e.g. Walli et al., 2009; Galuardi et al., 2010) that constitute the overall distribution of ABT is a key step in conserving them throughout their range. Here, we show that ABT tagged off Ireland appear to spend Days at large Trajectory most time in eastern management regions, were comprised of at least two migratory groups, and visited known spawning grounds in the MEDI, where, in at least one case, diving behaviour consis- reason Release time series with respective sampling frequencies shown in pa- tent with spawning was observed. ¼

Irish ABT migrations 9.17 Pin broke 262 E-W N 16 (200) 9.57 Complete 307 E-W Y 16 (111) 9.39 Pin broke 311 – Y 0 (7) 15.79 Premature 202 E-W – 18 (147) 42.04 Pin broke 133 E-W – 20 (121) 16.18 Complete 308 E-W Y 15 (156) 41.52 Premature 75 E-W – 3 (34) 26.63 Pin broke 135 E-W – 19 (128) 13.94 Premature 56 – – 17 (27) 50.91 Pin broke 260 E-W – 16 (235) 41.48 Premature 106 E-W – 18 (96) 10.62 Complete 365 E-W Y 12 (79) To meet the requirements of a metabolically demanding lifestyle whilst storing sufﬁcient energy reserves to reproduce, ABT must prioritize prey capture by exploiting a patchwork of seasonally

histogram and “TS” productive feeding areas (Walli et al., 2009; Wilson and Block, ¼ 2009). Records indicate that ABT have been present off Ireland since at least the 1970s and they have been caught by commercial pelagic trawlers and recreational fishers since 1999 (Cosgrove et al., 2008), although no commercial fishery exists. A pilot recreational catch, tag, and release programme was recently sanctioned to aid data collection for management (ICCAT, 2018). ABT arrive in Irish waters from foraging grounds in the Atlantic and spawning grounds in the Mediterranean from July onwards to feed on a diverse array of forage fish including sprat (Sprattus sprattus), Atlantic saury (Scomberesox saurus), and Atlantic mackerel 8.82 19 May 2017 27.56 8.81 10 March 2017 39.78 8.86 1 September 2017 61.10 8.82 1 July 2017 47.04 8.78 25 June 20178.87 34.01 8 12.74 January 2017 45.07 Pin broke 256 East – 18 (184) 8.82 5 August 2017 35.858.63 14.64 6 March 2017 Pin broke8.81 44.09 298 1 September 2017 53.92 East – 16 (283) 8.84 19 August 2017 57.60 8.59 28 December 2016 40.11 8.79 26 June 2017 42.38 8.80 23 January 2017 39.82 8.81 12 October 2017 54.50 (Scomber scomber, Cosgrove et al. 2008). The continental shelf edge is closest to Ireland and the United Kingdom in this area (ca. 100 km) and is a region of seasonally high productivity (Van De Poll et al., 2013) and an established migratory pathway for pelagic forage fish (Jansen et al., 2012). ABT leave these foraging

Deployment Pop-up grounds for the Bay of Biscay and/or the Central North Atlantic in the late autumn when stratiﬁcation of the water column breaks

Date Latitude Longitude Date Latitude Longitude down and the water cools (Van Aken, 2001). The Bay of Biscay hosts small (55–110 cm straight fork length) ABT throughout the summer and autumn (Rodrı´guez-Marı´n et al., 2003; Arregui et al., 2018). Here, we show that large ABT tagged off Ireland (199–224 cm CFL) do not occupy the area until winter and spring (October–April), highlighting differential, age-structured use of the area between conspeciﬁcs. The Central North Atlantic is a region of high productivity (Daniault et al., 2016) that has been shown to attract sharks (Queiroz et al., 2016), birds (Dias et al., 2012), whales (Silva et al., 2013) and multiple size cohorts of ABT (Block et al., 2005; Stokesbury et al., 2007; Arregui et al., 2018) from both eastern and western stocks (Rodrı´guez- Ezpeleta et al., 2019), including ABT from foraging aggregations off Ireland. 230220 MiniPAT 247A MiniPAT 247A TS (5 min) TS (5 min) 11 October 2016 11206 October 54.54 2016 54.53 MiniPAT 348F –8.78 HS (24 h)234 14 October 22 2016 October 2016 MiniPAT 54.54 348F 54.53 TS (5 min) –8.79 29 October 2016 Mortality 54.76 0 – – 6 (2) CFL (cm) Tag type Programming Understanding and characterizing the movements of large

Deployment and pop-up satellite tag metadata. ABT is a key step in conserving the spawning stock. Wilson et al. a a,b a a (2015) tracked ABT that had visited the Gulf of Mexico spawning ground and showed that they exhibited high fidelity to the forag- An ABT that is thought to have been caught by a fishing vessel on approximately 4 June 2017. Tags that were physically recovered. rentheses. “Xmit days” denotes the length of time the tag transmitted for post-release with the number of days with light data given in parentheses. 16P1265 16P1249 240 MiniPAT 348F TS (5 min) 29 October 2016 54.74 16P1267 220 MiniPAT 348F TS (5 min) 28 October 2016 54.78 16P1264 224 MiniPAT 348F TS (5 min) 28 October 2016 54.70 14P0062 215 MiniPAT 247A TS (5 min) 12 October 2016 54.55 14P0330 206 MiniPAT 247A HS (24 h)16P1253 207 12 October 2016 MiniPAT 348F 54.54 TS (5 min) 25 October 2016 54.71 14P0031 Table 1. Tag serial Summary statistics for 16 electronic tags attached to ABTa off the northwest coast ofb Ireland during 2016. For programming: “HS” 14P0441 212 MiniPAT 247A HS (24 h) 12 October 2016 54.54 16P1263 246 MiniPAT 348F HS (24 h) 1 November 2016 54.59 14P035914P0251 224 MiniPAT 247A HS (24 h) 11 October 2016 54.54 –8.74 31 July 2017 63.69 –4.17 Pin broke 293 East Y 18 (237) 14P0307 216 MiniPAT 247A HS (24 h) 9 October 2016 54.53 14P0337 216 MiniPAT 247A HS (24 h) 9 October 2016 54.53 16P1268 16P1170 199 MiniPAT 348F TS (5 min) 12 October 2016 54.53 ing ground they were first captured on whilst also remaining west 8 T. W. Horton et al.

Table 2. Vertical habitat use of tagged Atlantic blueﬁn tuna. Fast Tags migration T-S tags SST Mean VMR Region (days) (%) (nDjnT) (C) depth (m) (m h1) Mean temperature (C) ATC (Ch1) CNRY 8 (145) 13 5 (81j83) 16.6 6 2 (14.1) 22 6 5 (775) 77 6 19 (212) 17.2 6 1.5 (10.7)a 2.5 6 2.6 (11.8) 19 6 2 (698) 56 6 8 (167) 17.4 6 1.9 (10.7)a 1.2 6 0.9 (4.4) GFST 2 (117) 0 2 (19j17) 17.2 6 1.6 (10.8) 25 6 16 (609) 87 6 47 (233) 17.7 6 0.7 (0)a,b 1 6 0.5 (3.6) 29 6 21 (531) 108 6 62 (324) 17.8 6 0.6 (11.1)a,b 1.5 6 0.9 (2.3)

MEDI 4 (107) 30 4 (76j72) 22.4 6 2 (16.8) 24 6 24 (482)U 92 6 59 (391) 21 6 3 (13.3) 6.3 6 3.1 (17.3) Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020 13 6 3 (454)U 53 6 13 (426) 21.8 6 1.7 (13.3) 4.4 6 2.3 (14.7) NADR 14 (1 217) 8 10 (353j358) 14.5 6 1.9 (9.5) 64 6 41 (756)U,* 113 6 33 (359) 14.7 6 1.4 (3.7) 1 6 0.6 (5.9) 32 6 11 (871)U,* 103 6 28 (332) 14.7 6 1 (7.1) 0.7 6 0.3 (5.4) NASE 9 (623) 5 7 (277j265) 17.2 6 1.5 (12.9) 43 6 14 (958)* 105 6 45 (515) 17.2 6 0.9 (7.2)b 1.4 6 1.5 (7.2) 45 6 17 (931)* 136 6 63 (568) 17.3 6 1.2 (7.9)b 1.2 6 1.5 (9.6) NASW 1 (51) 0 0 – – – – – –– – – NECS 14 (299) 8 7 (165j166) 12.6 6 1.7 (9.7) 39 6 11 (195)U,* 114 6 52 (228) 12.9 6 1.1 (9.7) 0.9 6 1.1 (8.4) 24 6 10 (196)U,* 92 6 42 (207) 13.2 6 1.2 (9.9) 0.5 6 0.4 (4) SARC 1 (11) 0 1 (3j5) 12.5 6 0.3 (11.6) 16 (199) 56 (68) 12 (4.6) 2.5 (3.2) 13 (190) 46 (88) 12.5 (5.1) 1.2 (2.6) Fast migration 14 (209) N/A 7 (16j23) 15 6 1.9 (12.1) 30 6 30 (365) 77 6 59 (214) 14.6 6 1.3 (10.2) 0.7 6 0.5 (2.8) 41 6 31 (710) 146 6 119 (698) 15.1 6 1.8 (9.9) 1 6 0.5 (3.3) Totals 14 (2 779) 8 14 (990j989) Day 40 6 29 (958) 98 6 45 (515) 15.7 6 2.6 (0) 1.6 6 2 (17.3) Night 30 6 19 (931) 101 6 66 (698) 16.1 6 2.8 (5.1) 1.4 6 1.4 (14.7) All 35 6 25 (958) 100 6 56 (698) 15.9 6 2.7 (0) 1.5 6 1.7 (17.3) Maximum depth, maximum VMR, and maximum rate of ATC are shown in parentheses for “mean depth (m)”, “VMR (m h1)”, and “ATC (Ch1)”, respectively. Minimum temperature is shown in parentheses for “mean temperature (C)”. “T-S tags” denotes the number of tags that transmitted usable time-series data, with the number of days given in parentheses for each of depth (nD) and temperature (nT). White boxes denote day time, and grey-shaded boxes denote night-time periods. For “mean depth”, U denotes ecoregions where mean occupied depths were signiﬁcantly deeper than the GFST ecoregion and * denotes ecoregions where mean occupied depths were signiﬁcantly different between day and night summary periods (at the 5% level). For “mean temperature”, letters denote similarity between mean occupied temperatures and ecoregions without letters are statistically unique (at the 5% level). Fast migration, periods where distances between successive relocations indicate faster movements; VMR, vertical movement rate; ATC, ambient temperature change.

of the 45W meridian. Stokesbury et al. (2007), albeit with a small In recent years, ABT have been observed more frequently in sample size (n ¼ 3 ABT), demonstrated that ABT tagged off waters off Denmark, Sweden, and Norway after having been rare Ireland visited the MEDI and western Atlantic regions. Here, we since the 1960s (Ferter et al., 2019; MacKenzie et al., 2020). Here, build on this work, demonstrating that ABT tagged off Ireland we provide no evidence of connectivity between foraging grounds constitute two movement types (east-west and eastern resident) off Ireland and these Nordic regions. However, it is likely that the and visit known spawning grounds in the central and western migration of ABT into waters north of the study site, such as MEDI. Furthermore, we show a high degree of spatio-temporal these, is under-represented in this dataset due to premature tag variability in individual fish movements, with multiple ecoregions shedding. Given the similar sizes of ABT in the Nordic (Denmark inhabited simultaneously (although we note not all ecoregions and Sweden, mean 232 6 16 cm CFL, MacKenzie et al., 2020) and are similar in area) by fish of a size considered to be sexually ma- Irish aggregations, it could be that ABT tracked from Irish forag- ture. Reasons underlying this are likely varied and reflect the chal- ing grounds (i) visited Nordic regions but the temporal range of lenge of locating sufficient forage fish in a heterogeneous geolocation data in this study did not capture this part of the an- environment. Some ABT tracked in the present study did not visit nual migration, or (ii) show foraging ground fidelity (as for known spawning grounds, which is consistent with other ABT spawning grounds, i.e. Rooker et al., 2008) and Ireland and electronic tagging studies (Block et al., 2005; Walli et al., 2009; Denmark represent unique cohorts of the ABT population with Galuardi et al., 2010; Wilson et al., 2015) and may be because differing spatial habits. they (i) use alternate spawning locations, for instance the Slope Sea (Richardson et al., 2016; e.g. Supplementary Figure S5h)or waters off the Canary Islands (Druon et al., 2016; e.g. Patterns in diving behaviour Supplementary Figure S5f), (ii) could be immature fish from the Many marine fishes dive extensively for reasons including forag- proportion of the western stock that spawn in the Gulf of Mexico ing (Wilson and Block, 2009; Thorrold et al., 2014; Whitlock (these fish show first spawning ground visitation at larger sizes et al., 2015), thermoregulation (Teo et al., 2007), reproduction than the eastern stock, Block et al., 2005), or (iii), as with other (Aranda et al., 2013; Cermeno~ et al., 2015), and navigation iteroparous fishes, may choose to defer breeding due to a defi- (Brunnschweiler et al., 2009) and employ different diving strate- cient diet or poor nutritional condition (Rideout and gies depending on their movement mode (i.e. transiting—Walli Tomkiewicz, 2011). et al., 2009) and the time of day (Gilly et al., 2006; Queiroz et al., Tracking Atlantic bluefin tuna from Ireland 9 Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020

Figure 4. High-resolution diving behaviour of an Atlantic blueﬁn tuna. Physical archive time series (5-s resolution) of depth, temperature, and rate of ATC for a tag recovered from the Outer Hebrides in September 2017 (16P1265). (a) The full time series with a labelled horizontal colour bar denoting which ecoregion the tag was in (colours as per Figure 2), vertical dashed lines represent temporal range of subsequent plots, (b) the deepest recorded dive of any blueﬁn tuna in this study, (c) mesopelagic diving in the NADR region, (d) deep-diving during exit from the Mediterranean Sea, and (e) diel vertical migration in the NECS ecoregion. ATC, ambient temperature change.

2016; Jansen et al., 2019). Here, we demonstrate that ABT dive ABT in this study spent most time in the mixed layer. This behav- extensively and follow a diel diving pattern, which likely reﬂects iour has been observed previously for ABT (Walli et al., 2009; foraging effort as ABT follow the vertical migrations of their prey Lawson et al, 2010; Galuardi and Lutcavage, 2012). An explana- (Darbyson et al., 2003; Gilly et al., 2006; Olson et al., 2016; Jansen tion for the relationship could be physiological, as ABT seek et al., 2019). In addition, low rates of ambient temperature warmer (near-surface) temperatures after visiting deeper (cooler) change across all except the MEDI ecoregion indicate that the depths. ABT likely trade-off between maintaining internal 10 T. W. Horton et al. Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020

Figure 5. Putative spawning behaviour of an Atlantic bluefin tuna in the Mediterranean Sea. (a) Time series of depth and ambient temperature at 5-s resolution. Black line denotes 4-hourly mean depth, and black bar denotes periods when high-frequency shallow diving profiles were observed (see Supplementary Figure S8 for individual profiles and classification details), (b) time series of hourly VMR, (c) time series of the hourly rate of ATC, and (d) daily time series of Moon illumination as a fraction. For all plots, vertical dashed lines represent the date of entry and exit from the Mediterranean Sea. Grey-shaded boxes in (a) and (d) represent full night-time periods and in (b) and (c) the period 00:00–04:00, the putative spawning time. VMR, vertical movement rate; ATC, ambient temperature change.

temperatures at 12–20C above ambient (Lawson et al. 2010), et al. (2007) proposed that heat production through metabolic and visiting colder but more productive waters to replenish de- processes likely increases (internally placed tags demonstrated vis- pleted energy reserves after migration or spawning by foraging on ceral warming) and Reglero et al. (2018) demonstrate that oocytes an energy rich food source (e.g. spawning Atlantic herring, need to be released into warmest surface waters to maximize Pleizier et al., 2012; Wilson et al., 2015). To further this, we pre- growth and development. Consequently, high-frequency shallow sent the first data showing ABT residing in inshore waters of the dives intersecting the thermocline detailed in the present study in Inner Hebrides, Scotland, and the Celtic Deeps off Wales, the the MEDI ecoregion may reflect a thermoregulatory behaviour to coolest regions inhabited by ABT in this study (mean temperature balance the physiological effects of potential heat stress (Teo 12.6C). et al., 2007) whilst also releasing oocytes and sperm above the thermocline at temperatures best for growth and development (Reglero et al., 2018). This specific behaviour, coupled with warm Spawning behaviour of Irish ABT surface waters and the shallow, intense stratification in the MEDI Archival tags have been previously used to identify ABT putative ecoregion likely resulted in observed high rates of ambient tem- spawning behaviours (e.g. Teo et al., 2007; Aranda et al., 2013; perature change over comparatively small changes in depth, Cermeno~ et al., 2015; Hazen et al., 2016). During spawning, Teo reflecting behavioural thermoregulation. Tracking Atlantic bluefin tuna from Ireland 11

Overview during long-distance movement in the western Indian Ocean. Recent years have seen ABT return to waters of the northeast Journal of Fish Biology, 74: 706–714. Atlantic, including the waters off Ireland. Here, we show that the Calenge, C. 2006. The package ‘adehabitat’ for the R software: a tool for the analysis of space and habitat use by animals. Ecological ABT in this aggregation spend most of their time in eastern stock Modelling, 197: 516–519. management units and exhibit high fidelity to foraging grounds Cermeno,~ P., Quı´lez-Badia, G., Ospina-Alvarez, A., Sainz-Tra´paga, S., of the northeast Atlantic. We link ABT present in aggregations off Boustany, A. M., Seitz, A. C., Tudela, S. et al 2015. Electronic tag- northwest Ireland to established high-use areas of the Central ging of Atlantic bluefin tuna (Thunnus thynnus, L.) reveals habitat North Atlantic, Bay of Biscay, and known spawning areas in the use and behaviors in the Mediterranean Sea. PLoS One, 10: western and central MEDI, but only tentatively to western stock e0116638. Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020 management units and not at all to Nordic regions. The reappear- Cosgrove, R., Stokesbury, M. J. W., Browne, D., Boustany, A., Block, ance of large ABT into the northeast Atlantic represents an op- B. A., and Farrell, M. 2008. Bluefin tuna tagging in Irish waters. Fisheries Resource Series, 6: 1–16. portunity to study the full cycle of foraging and reproductive Daniault, N., Mercier, H., Lherminier, P., Sarafanov, A., Falina, A., behaviours of this important cohort of the eastern Atlantic Zunino, P., Pe´rez, F. F. et al 2016. The northern North Atlantic spawning stock in a period of increasing fishing pressure in the Ocean mean circulation in the early 21st century. Progress in eastern Atlantic. Oceanography, 146: 142–158. Darbyson, E., Swain, D. P., Chabot, D., and Castonguay, M. 2003. Supplementary data Diel variation in feeding rate and prey composition of herring Supplementary material is available at the ICESJMS online ver- and mackerel in the southern Gulf of St Lawrence. Journal of Fish sion of the manuscript. Biology, 63: 1235–1257. Dias, M. P., Granadeiro, J. P., and Catry, P. 2012. Do seabirds differ Data availability from other migrants in their travel arrangements? On route strategies of Cory’s shearwater during its trans-equatorial journey. The data underlying this article are available in the article and in PLoS One, 7: e49376. its online supplementary material. Druon, J.-N., Fromentin, J.-M., Hanke, A. R., Arrizabalaga, H., Damalas, D., Ticina, V., Quı´lez-Badia, G. et al 2016. Habitat suit- Acknowledgements ability of the Atlantic bluefin tuna by size class: an ecological niche We thank skippers Adrian Molloy of the FV Evie Rose and approach. Progress in Oceanography, 142: 30–46. Michael Callaghan of the FV Leah C, without whom this work Faillettaz, R., Beaugrand, G., Goberville, E., and Kirby, R. R. 2019. would not have been possible. Atlantic Multidecadal Oscillations drive the basin-scale distribution of Atlantic bluefin tuna. Science Advances, 5: eaar6993. Funding Faillettaz, R., Beaugrand, G., Goberville, E., and Kirby, R. R. 2020. T.W.H. was supported by a University of Exeter Ph.D scholar- Atlantic Multidecadal Oscillations drive the basin-scale distribu- ship, M.J.W., L.A.H. and T.W.H. were supported by the tion of Atlantic bluefin tuna. Science Advances, 5: eaar6993. Ferter, K., Tracey, S., Hinriksson, J., Bjelland, O., Onandia, I., and European Maritime and Fisheries Fund (ENG2395) and Defra Nøttestad, L. 2019. Tagging of Atlantic bluefin tuna (Thunnus grant C7531. Fieldwork operations were supported by a grant thynnus) with pop-up satellite archival tags (PSAT) in western from the Irish Marine Institute. Norway during 2018. https://www.iccat.int/GBYP/Docs/Tagging_ Phase_8_IMR.pdf (last accessed 20 May 2020). References Fromentin, J.-M. 2009. Lessons from the past: investigating historical Abascal, F. J., Medina, A., De La Serna, J. M., Godoy, D., and Aranda, data from bluefin tuna fisheries. Fish and Fisheries, 10: 197–216. G. 2016. Tracking bluefin tuna reproductive migration into the Fromentin, J.-M., Bonhommeau, S., Arrizabalaga, H., and Kell, L. T. Mediterranean Sea with electronic pop-up satellite archival tags 2014a. The spectre of uncertainty in management of exploited fish using two tagging procedures. Fisheries Oceanography, 25: 54–66. stocks: the illustrative case of Atlantic bluefin tuna. Marine Policy, Amante, C., and Eakins, B. W. 2009. ETOPO1 1 Arc-Minute Global 47: 8–14. Relief Model: Procedures, Data Sources and Analysis. NOAA Fromentin, J.-M., Reygondeau, G., Bonhommeau, S., and Beaugrand, Technical Memorandum NESDIS NGDC-24. National G. 2014b. Oceanographic changes and exploitation drive the Geophysical Data Center, NOAA. spatio-temporal dynamics of Atlantic bluefin tuna (Thunnus thyn- Aranda, G., Abascal, F. J., Varela, J. L., and Medina, A. 2013. nus). Fisheries Oceanography, 23: 147–156. Spawning behaviour and post-spawning migration patterns of Galuardi, B., Golet, W., Lutcavage, M., Logan, J., Royer, F., and Atlantic bluefin Tuna (Thunnus thynnus) ascertained from satel- Neilson, J. 2010. Complex migration routes of Atlantic bluefin lite archival tags. PLoS One, 8: e76445. tuna (Thunnus thynnus) question current population structure Arregui, I., Galuardi, B., Goni,~ N., Lam, C. H., Fraile, I., Santiago, J., paradigm. Canadian Journal of Fisheries and Aquatic Sciences, Lutcavage, M. et al 2018. Movements and geographic distribution 67: 966–976. of juvenile bluefin tuna in the Northeast Atlantic, described Galuardi, B., and Lutcavage, M. 2012. Dispersal routes and habitat through internal and satellite archival tags. ICES Journal of utilization of juvenile atlantic bluefin tuna, Thunnus thynnus, Marine Science, 75: 1560–1572. tracked with mini PSAT and archival tags. PLoS One, 7: e37829. Bates, D., Ma¨chler, M., Bolker, B. and Walker, S. 2014. Fitting linear Gilly, W., Markaida, U., Baxter, C., Block, B., Boustany, A., Zeidberg, mixed-effects models using lme4. arXiv preprint arXiv:1406.5823. L., Reisenbichler, K. et al 2006. Vertical and horizontal migrations Block, B. A., Teo, S. L. H., Walli, A., Boustany, A., Stokesbury, M. J. by the jumbo squid Dosidicus gigas revealed by electronic tagging. W., Farwell, C. J., Weng, K. C. et al 2005. Electronic tagging and Marine Ecology Progress Series, 324: 1–17. population structure of Atlantic bluefin tuna. Nature, 434: Hanke, A., Macdonnell, A., Dalton, A., Busawon, D., Rooker, J. R., 1121–1127. and Secor, D. H. 2017. Stock mixing rates of bluefin tuna from Brunnschweiler, J. M., Baensch, H., Pierce, S. J., and Sims, D. W. Canadian landings: 1975-2015. Collective Volume of Scientific 2009. Deep-diving behaviour of a whale shark Rhincodon typus Papers. ICCAT, 74: 2622–2634. 12 T. W. Horton et al.

Hazen, E. L., Carlisle, A. B., Wilson, S. G., Ganong, J. E., Castleton, Rodrı´guez-Marıń, E., Arrizabalaga, H., Ortiz, M., Rodrı´guez-Cabello, M. R., Schallert, R. J., Stokesbury, M. J. et al 2016. Quantifying C., Moreno, G., and Kell, L. T. 2003. Standardization of bluefin overlap between the Deepwater Horizon oil spill and predicted tuna, Thunnus thynnus, catch per unit effort in the baitboat fish- bluefin tuna spawning habitat in the Gulf of Mexico. Scientific ery of the Bay of Biscay (Eastern Atlantic). ICES Journal of Reports, 6: 33824. Marine Science, 60: 1216–1231. ICCAT. 2017. Report of the 2017 ICCAT bluefin stock assessment Rodrı´guez-Ezpeleta, N., Dıáz-Arce, N., Walter, J. F., Richardson, D. E., meeting. Collective Volume of Scientific Papers. ICCAT, 74: Rooker, J. R., Nøttestad, L., Hanke, A. R. et al 2019. Determining 2372–2535. natal origin for improved management of Atlantic bluefin tuna. ICCAT. 2018. BFT 18-02: recommendation by ICCAT establishing a Frontiers in Ecology and the Environment, 17: 439–436.

multi-annual management plan for bluefin tuna in the eastern Downloaded from https://academic.oup.com/icesjms/article-abstract/doi/10.1093/icesjms/fsaa090/5878157 by guest on 30 July 2020 Rooker, J. R., Fraile, I., Liu, H., Abid, N., Dance, M. A., Itoh, T., Atlantic and the Mediterranean Sea. 18-02. Madrid, Spain. 1–46 Kimoto, A. et al 2019. Wide-ranging temporal variation in trans- pp. https://www.iccat.int/Documents/Recs/compendiopdf-e/ 2018-02-e.pdf (last accessed 20 May 2020). oceanic movement and population mixing of bluefin tuna in the North Atlantic Ocean. Frontiers in Marine Science, 6: 1–13. Jansen, T., Campbell, A., Kelly, C., Ha´tu´ n, H., and Payne, M. R. 2012. Migration and fisheries of north east Atlantic Mackerel (Scomber Rooker, J. R., Secor, D. H., De Metrio, G., Schloesser, R., Block, B. A., scombrus) in autumn and winter. PLoS One, 7: e51541. and Neilson, J. D. 2008. Natal homing and connectivity in Jansen, T., Post, S., Olafsdottir, A. H., Reynisson, P., O´ skarsson, G. J., Atlantic bluefin tuna populations. Science, 322: 742–744. and Arendt, K. E. 2019. Diel vertical feeding behaviour of Atlantic Silva, M. A., Prieto, R., Jonsen, I., Baumgartner, M. F., and Santos, R. mackerel (Scomber scombrus) in the Irminger current. Fisheries S. 2013. North Atlantic blue and fin whales suspend their spring Research, 214: 25–34. migration to forage in middle latitudes: building up energy Kimoto, A., and Itoh, T. 2017. The standardized bluefin CPUE of reserves for the journey? PLoS One, 8: e76507. Japanese longline fishery in the Atlantic up to 2017 fishing year. Stokesbury, M. J. W., Cosgrove, R., Boustany, A., Browne, D., Teo, S. ICCAT SCRS, 25: 1–23. L. H., O’Dor, R. K., and Block, B. A. 2007. Results of satellite tag- Lawson, G. L., Castleton, M. R., and Block, B. A. 2010. Movements ging of Atlantic bluefin tuna, Thunnus thynnus, off the coast of and diving behavior of Atlantic bluefin tuna Thunnus thynnus in Ireland. Hydrobiologia, 582: 91–97. relation to water column structure in the Northwestern Atlantic. Taylor, N. G., Mcallister, M. K., Lawson, G. L., Carruthers, T., and Marine Ecology Progress Series, 400: 245–265. Block, B. A. 2011. Atlantic bluefin tuna: a novel multistock spatial Longhurst, A. R., 2010. Ecological geography of the sea, 2nd edn. model for assessing population biomass. PLoS One, 6: e27693. Elsevier. 560 pp. Teo, S. L. H., Boustany, A., Dewar, H., Stokesbury, M. J. W., Weng, MacKenzie, B. R., Aarestrup, K., Birnie-Gauvin, K., Cardinale, M., K. C., Beemer, S., Seitz, A. C. et al 2007. Annual migrations, div- Casini, M., Harkes, I., Onandia, I. et al 2020. Electronic tagging of ing behavior, and thermal biology of Atlantic bluefin tuna, adult bluefin tunas by sport fishery in the Skagerrak. Collective Thunnus thynnus, on their Gulf of Mexico breeding grounds. Volume of Scientific Papers. ICCAT, 76: 650–664. Marine Biology, 151: 1–18. O´ Maoileídigh, N., Connolly, P., Drumm, A., Neill, O., Maxwell, H., Co, J., Bunn, R. et al 2018. Final report on tagging activities in the Thorrold, S. R., Afonso, P., Fontes, J., Braun, C. D., Santos, R. S., Celtic Seas Area 2018. (ICCAT GBYP 07/2018-PHASE 8). 2018. Skomal, G. B., and Berumen, M. L. 2014. Extreme diving behav- Olson, R. J., Young, J. W., Meńard, F., Potier, M., Allain, V., Goni,~ N., iour in devil rays links surface waters and the deep ocean. Nature Logan, J. M. et al 2016. Bioenergetics, trophic ecology, and niche Communications, 5: 1–7. separation of tunas. Advances in Marine Biology, 35: 199–344. Van Aken, H. M. 2001. The hydrography of the mid-latitude Pleizier, N. K., Campana, S. E., Schallert, R. J., Wilson, S. G., and Northeast Atlantic Ocean—Part III: the subducted thermocline Block, B. A. 2012. Atlantic bluefin tuna (Thunnus thynnus) diet in water mass. Deep-Sea Research Part I: Oceanographic Research the Gulf of St. Lawrence and on the eastern Scotian Shelf. Journal Papers, 48: 237–267. of Northwest Atlantic Fishery Science, 44: 67–76. Van De Poll, W. H., Kulk, G., Timmermans, K. R., Brussaard, C. P. Queiroz, N., Humphries, N. E., Mucientes, G., Hammerschlag, N., D., Van Der Woerd, H. J., Kehoe, M. J., Mojica, K. D. A. et al Lima, F. P., Scales, K. L., Miller, P. I. et al 2016. Ocean-wide track- 2013. Phytoplankton chlorophyll a biomass, composition, and ing of pelagic sharks reveals extent of overlap with longline fishing productivity along a temperature and stratification gradient in the hotspots. Proceedings of the National Academy of Sciences of the northeast Atlantic Ocean. Biogeosciences, 10: 4227–4240. United States of America, 113: 1582–1587. Walli, A., Teo, S. L., Boustany, A., Farwell, C. J., Williams, T., Dewar, Ravier, C., and Fromentin, J. M. 2004. Are the long-term fluctuations H., Prince, E. et al 2009. Seasonal movements, aggregations and in Atlantic bluefin tuna (Thunnus thynnus) population related to diving behavior of Atlantic bluefin tuna (Thunnus thynnus) environmental changes? Fisheries Oceanography, 13: 145–160. revealed with archival tags. PLoS One, 4: e6151. Reglero, P., Ortega, A., Balbıń, R., Abascal, F. J., Medina, A., Blanco, Whitlock, R. E., Hazen, E. L., Walli, A., Farwell, C., Bograd, S. J., E., de la Gańdara, F. et al 2018. Atlantic bluefin tuna spawn at Foley, D. G., Castleton, M. et al 2015. Direct quantification of en- suboptimal temperatures for their offspring. Proceedings of the ergy intake in an apex marine predator suggests physiology is a Royal Society B: Biological Sciences, 285: 20171405. key driver of migrations. Science Advances, 1: e1400270–10. Richardson, D. E., Marancik, K. E., Guyon, J. R., Lutcavage, M. E., Galuardi, B., Lam, C. H., Walsh, H. J. et al 2016. Discovery of a Wilson, S. G., and Block, B. A. 2009. Habitat use in Atlantic bluefin spawning ground reveals diverse migration strategies in Atlantic tuna Thunnus thynnus inferred from diving behavior. Endangered bluefin tuna (Thunnus thynnus). Proceedings of the National Species Research, 10: 355–367. Academy of Sciences of the United States of America, 113: Wilson, S. G., Jonsen, I. D., Schallert, R. J., Ganong, J. E., Castleton, 3299–3304. M. R., Spares, A. D., Boustany, A. M. et al 2015. Tracking the fi- Rideout, R. M., and Tomkiewicz, J. 2011. Skipped spawning in fishes: delity of Atlantic bluefin tuna released in Canadian waters to the more common than you might think. Marine and Coastal Gulf of Mexico spawning grounds. Canadian Journal of Fisheries Fisheries, 3: 176–189. and Aquatic Sciences, 72: 1700–1717.

Handling editor: David Secor