MARINE AND FISHERY SCIENCES 36 (3): 245-265 (2023)

https://doi.org/10.47193/mafis.3632023010907

ABSTRACT. The Patagonian toothfish (Dissostichus eleginoides) is a commercially valuable

demersal fish present in the southern hemisphere. Therefore, understanding movements of the

species across time and spatial scales would enhance our understanding of its behavior within the

Argentine and Chilean Patagonian platform. The Instituto Nacional de Investigación y Desarrollo

Pesquero (INIDEP, Argentina) initiated a tag-recapture program for D. eleginoides in 2004. A total

of 5,907 specimens, mostly juveniles (<82 cm total length), have been tagged and released in waters

off the edge of the Argentine shelf and slope between 37° S and 47° S (northern Sector of the Argen-

tine fishing ground), and east of De los Estados Island and south of Tierra del Fuego (54° S-57° S

southern Sector of the Argentine fishing ground). A total of 121 specimens were recaptured: 25

(20.7%) were recovered in the northern Sector, 84 (69.4%) in the southern Sector and 12 (9.9%) in

waters of the Pacific Ocean in Chile. A total of 67.5% were recaptured within 20 nm (37 km) of the

release site and 15% traveled distances of less than 120 nm. A smaller fraction (5%) traveled dis-

tances between 120 and 400 nm and only 12.5% were recaptured at more than 400 nm. Using a Gen-

eralized Additive Model, it was determined that variables Days at liberty, Sector and Time of tag-

ging were influential in the distance traveled by specimens. From the present work, it seems evident

that this species possesses high site fidelity and lacks cyclic migratory movements involving a sub-

stantial component of the stock in the American southern cone.

Key words: Tagging, recapture, fishing ground, stock.

Patrones migratorios de la merluza negra (Dissostichus eleginoides) en el Océano Atlántico

Sudoccidental

RESUMEN. La merluza negra (Dissostichus eleginoides) es un pez demersal presente en el

hemisferio sur, muy valioso comercialmente. Por ello, comprender los movimientos en diferentes

escalas temporales y espaciales contribuiría a conocer más acerca del comportamiento que presenta

la especie en la plataforma patagónica argentina y chilena. Desde 2004, el Instituto Nacional de

Investigación y Desarrollo Pesquero (INIDEP, Argentina), inició un programa de marcado y recap-

tura de D. eleginoides. Un total de 5.907 ejemplares, en su mayoría juveniles (<82 cm de largo

total), fueron marcados y liberados en sectores ubicados en aguas del borde de la plataforma y talud

de la Argentina entre 37° S y 47° S (Sector norte del caladero argentino), y al este de la Isla de los

Estados y sur de Tierra del Fuego (54° S-57° S Sector sur del caladero argentino). Actualmente,

fueron recapturados 121 ejemplares, 25 (20,7%) se recuperaron en el Sector norte, 84 (69,4%) en

el Sector sur y 12 (9,9%) en aguas del Océano Pacífico en Chile. El 67,5% fue recapturado a menos

de 20 mn (37 km) del lugar de liberación y 15% recorrió distancias inferiores a las 120 mn. Una

fracción menor (5%) recorrió distancias entre 120 y 400 mn y solo 12,5% se recapturó a más de

400 mn. Mediante un Modelo Aditivo Generalizado se determinó que las variables Días en libertad,

Sector y Época de marcado influyeron en la distancia recorrida de los ejemplares. A partir del

245

*Correspondence:

gtroccoli@inidep.edu.ar

Received: 21 December 2022

Accepted: 2 June 2023

ISSN 2683-7595 (print)

ISSN 2683-7951 (online)

https://ojs.inidep.edu.ar

Journal of the Instituto Nacional de

Investigación y Desarrollo Pesquero

(INIDEP)

This work is licensed under a Creative

Commons Attribution-

NonCommercial-ShareAlike 4.0

International License

Marine and

Fishery Sciences

MAFIS

ORIGINAL RESEARCH

Migratory patterns of Patagonian toothfish (Dissostichus eleginoides) in the

southwestern Atlantic Ocean

GONZALO H. TROCCOLI*, PATRICIA A. MARTÍNEZ, EMILIANO J. DIMARCO, JUAN A. WAESSLE and OTTO C. WÖHLER

Instituto Nacional de Investigación y Desarrollo Pesquero (INIDEP), Paseo Victoria Ocampo Nº 1, Escollera Norte, B7602HSA -

Mar del Plata, Argentina. ORCID Gonzalo H. Troccoli https://orcid.org/0000-0002-0057-4325

INTRODUCTION

The Genus Dissostichus comprises two impor-

tant commercial species: D. mawsoni and D. elegi-

noides (Collins et al. 2010). The first one is dis-

tributed in Antarctic waters, while the second

reaches lower latitudes (Gon and Heemstra 1990).

Patagonian toothfish (D. eleginoides) has a wide

distribution in the southern hemisphere (Collins et

al. 2010; Nelson 2016), with its populations often

separated by vast expanses of deep ocean (Shaw et

al. 2004; Ashford et al. 2006; Rogers et al. 2006;

Ashford and Jones 2007). Patagonian toothfish is

a demersal species inhabiting shelves and slopes

of subantarctic islands of the Southern Ocean

(South Georgia Islands, South Sandwich Islands,

South Orkneys, Crozet, Kerguelen, Heard,

McDonald, Macquarie and Prince Edward, banks

such as Banzare, Ob and Lena, and regions of the

Ross Sea) and waters surrounding the southern

American cone, from Peru and Chile in the Pacific

Ocean to Argentina and Uruguay in the Atlantic

(Cousseau and Perrota 2004). The species inhabits

mainly the deep shelf, continental slope and sub-

marine canyons from 80 to 2,500 m depth (Prenski

and Almeyda 2000; Arkhipkin and Laptikhovsky

2010; Lee et al. 2021). Its ontogenetic cycle is

characterized by a pelagic behavior during early

development stages (eggs and larvae) (Evseenko

et al. 1995; North 2002) and a demersal behavior

from juvenile to adult. Several studies in the South

Atlantic (Agnew et al. 1999; Arkhipkin et al.

2003; Arkhipkin and Laptikhovsky 2010; Collins

et al. 2010; Péron et al. 2016; Lee et al. 2021) and

Indian Ocean (Duhamel 1981; Duhamel and

Pletikosic 1983; López Abellán 2005; Welsford et

al. 2011) have indicated a correlation between

larger fish and their distribution at greater depths.

This ontogenetic shift in habitat is common in

deep-sea fish (Agnew et al. 1999). However, spa-

tial segregation between individuals may also be

related to biological factors, such as trophic (Troc-

coli et al. 2020) or reproductive behavior (Brown

et al. 2013a; Boucher 2018) and environmental

variables, such as currents and their physical vari-

ables (Lee et al. 2021).

In the process of managing a fishing resource,

it is essential to be aware of the existence of uni-

tary stocks in order to determine whether a fish-

ery can be managed independently or integratedly

with other jurisdictions (Ying et al. 2011;

Hawkins et al. 2016; Kerr et al. 2017; Cadrin

2020). There are many factors contributing

knowledge about differentiation of stocks, such

as the understanding of fish movements in differ-

ent temporal and spatial scales (Brown et al.

2013a; Lee et al. 2022). The abundance of certain

populations may be subject to change in response

to any variation of relevant factors (Cianelli et al.

2013). The identity of the southwest Atlantic

toothfish stock (SA) has been the subject of vari-

ous studies. There is a well-documented genetic

differentiation between individuals from Patag-

onian shelf and those from South Georgia and

South Sandwich Islands (Shaw et al. 2004;

Rogers et al. 2006; Canales-Aguirre et al. 2018).

However, detailed population structure of the

species on the Argentine and Chilean Patagonian

shelf is still unknown. So far, genetic studies in

the area have not demonstrated any differentia-

tion justifying the presence of more than one pop-

ulation in the Chilean Pacific and Argentine

Atlantic (Shaw et al. 2004; Rogers et al. 2006;

Canales-Aguirre et al. 2018; Arkhipkin et al.

2022). Notwithstanding, other complementary

approaches, such as tag-recapture (Brown et al.

246 MARINE AND FISHERY SCIENCES 36 (3): 245-265 (2023)

presente trabajo, parece evidente que la especie en el cono sur americano carece de desplazamientos migratorios regulares que involu-

cren a una parte significativa del stock.

Palabras clave: Marcado, recaptura, caladero, stock.

2013a; Lee et al. 2022), otolith microchemistry

(Ashford et al. 2006, 2007; Ashford and Jones

2007), parasitology (Brickle et al. 2006; Brown et

al. 2013b), otolith shape (Lee et al. 2018), repro-

ductive characteristics (Laptikhovsky et al. 2006;

Pájaro et al. 2009; Boucher 2018) and oceano-

graphic modelling of dispersal patterns (Ashford

et al. 2012), indicate some heterogeneity along

the slope of the southern American cone, suggest-

ing the existence of different stocks in the area

(Wöhler et al. in press).

Tag-recapture programs established in differ-

ent fisheries around the world (Williams and

Lamb 2002; Williams et al. 2002; Marlow et al.

2003; Tuck et al. 2003; Agnew et al. 2006b; Dunn

et al. 2007; Roberts and Agnew 2008; Brown et

al. 2013a; Rubilar et al. 2013; Lee et al. 2022)

have been a very useful tool to understand the

geographic and bathymetric movements of fish,

contributing to understand fish growth, behavior

and population structure (Collins et al. 2010). The

so-called ‘conventional tags’ enable the determi-

nation of the distance traveled from the release

point to the recapture point, which can be corre-

lated with a specific timeframe. Specimens of the

Genus Dissostichus are suitable for tagging and

recovery studies as they are relatively robust and,

given the lack of a swim bladder, do not sustain

serious decompression injuries when captured

and brought to the surface from the depths in

which they inhabit (Agnew et al. 2006a). As a

result, numerous studies have investigated their

movement patterns through this approach

(Williams et al. 2002; Marlow et al. 2003; Tuck et

al. 2003; Agnew et al. 2006b; Roberts and Agnew

2008; Petrov and Tatarnikov 2010; Stacy et al.

2021; Lee et al. 2022; Grilly et al. 2022).

The Patagonian toothfish tag and recapture

program has been in place in Argentine waters

since 2004 to learn more about possible migrato-

ry movements of the species in the southwestern

Atlantic. This article analyzed results derived

from the aforementioned program as a contribu-

tion to our understanding of migratory move-

ments of the juvenile toothfish fraction and the

population structure present in the Argentine

Patagonian shelf.

MATERIALS AND METHODS

In 2004, researchers from the Instituto

Nacional de Investigación y Desarrollo Pesquero

(INIDEP) launched the Tag and Recapture Pro-

gram based on an experience conducted onboard

of a commercial longliner. Individuals were

tagged with an Australian-made standard anchor

T-bar tag (Hallprint, Adelaide), similar to the ones

used in various toothfish tagging programs at the

Commission for the Conservation of Antarctic

Marine Living Resources (CCAMLR). The Pro-

gram became a mandatory activity for fishing

vessels with an individual transferable quota fish-

ing for the species at the beginning of 2007. The

program focuses on specimens that have been

tagged in two sectors of the fishing ground of the

Argentine commercial fleet: one corresponding to

the northern Sector (NS), bordering the slope

between 37° S and 47° S, where the longline fleet

operates; and the other to the southern Sector

(SS), located to the east of De los Estados Island

(54° 30'S-55° 00'S and 64° 00'W-58° 00'W)

and south of Tierra del Fuego (55° 00'S-57° 30'

S and 67° 00'W-64° 00'W), which is visited by

both longliner and trawler fleets.

With the progressive withdrawal of longliners

as from 2017, specimens were only tagged in

research surveys conducted onboard of short-set

trawlers at shallower depths, contemplating a

conditioning time for the tagged fish in water

tanks before discarding those that did not recover

properly and therefore increase the survival rate

of those released. Fish were selected, prioritizing

those with the best general condition, measured

(cm) and weighed (g). Tags were inserted with a

gun on the left side of the dorsal part of the fish,

in the area of muscle between the second and

247

TROCCOLI ET AL.: MIGRATORY PATTERNS OF THE PATAGONIAN TOOTHFISH IN THE SOUTHWESTERN ATLANTIC OCEAN

third spine of the first dorsal fin (Figure 1). Data

on geographical position (longitude and latitude),

date and depth of the capture and subsequently

release sites were recorded. The release position

remained the same or very close to the one from

which the capture was obtained.

All tagged specimens were recaptured by com-

mercial fleets. When a tagged toothfish was

caught, observers onboard recorded its tag num-

ber, size, weight, and sex. Vessel operators

recorded data if no observer was present. Recap-

tures in Chilean waters were recorded by both

freezer and artisanal fleets. Observer records

included tag number, size, weight, sex, date, time,

depth, latitude, and longitude. In some cases, tags

were recovered after processing, leaving only the

date and position of recapture available.

Estimated traveled distances by specimens and

potential trajectories were visualized from tag and

recapture position data by using Google Earth v.

7.3.3 Pro. They were divided into four categories:

Short Distances (<20 nm), Short Intermediate Dis-

tances (20-120 nm), Long Intermediate Distances

(120-400 nm) and Long Distances (>400 nm).

Considering that toothfish is a deep-sea species, a

trajectory on the edge of the slope between the

release position and the recapture position was

assumed, and the minimum distances that the fish

would have traveled in nautical miles (1 nm =

1,852 km) were measured. Subsequently, displace-

ment of specimens was analyzed with a General-

ized Additive Model (GAM), using the ‘mgcv’

package of the R program version 4.3.0. These

models are an extension of general linear models

replacing the linear predictors with additive pre-

dictors through a smoothing function or spline

(Hastie and Tibshirani 1986; Wood 2006). Four

different transformations were tested on the

response variable (log, square root, cube root and

box cox) to reduce the spread of data. The log

transformation provided the best-validated model

fit. Continuous explanatory variables considered

influential in the traveled distance were the length

of tagged specimens (TL), the period of time

between the release and the capture (Days at liber-

ty DL), and the tagging Depth (D). Factors con-

sidered were the sector in which specimens were

tagged (northern and southern) and the season

(spring, summer, autumn and winter). The general

model was the following:

248 MARINE AND FISHERY SCIENCES 36 (3): 245-265 (2023)

Figure 1. Position of the body where the T-bar style tag was implanted. In some specimens two tags were implanted.

Log(Distance) =m+s(TL) +s(DL) +s(D) +Sector

+Season +Error

Family =Gaussian, Link =‘Identity’

where e~N (0, s2e). First, the existence of

collinearity between variables was verified to

eliminate any variable correlated with another

(Zuur et al. 2009). Once definitive variables were

obtained, the best model was chosen based on the

stepwise selection process according to the

Akaike Information Criteria and Generalized

Cross Validation (Akaike 1973; Burnham and

Anderson 2004; Wood 2012). In both cases, the

lower the value the better the model obtained will

be. In addition, the adjusted multiple determina-

tion coefficient was calculated for each model,

and the explanatory capacity of the model built

through the explained deviance was calculated.

Subsequently, influential points were eliminated,

obtaining the final model. Finally, its graphic val-

idation was carried out using the gam.chek func-

tion (‘mgcv’) and those variables significant to

the final model were plotted (Fernández Casal et

al. 2022).

RESULTS

A total of 5,907 toothfish have been tagged in

the northern Sector (NS) and southern Sector

(SS) since the beginning of the Program in 2004.

A total of 5,528 specimens were tagged onboard

the longline fleet and 379 on the trawler, from

which, 26.04% were tagged in the NS, while the

remaining 73.96% were tagged in the SS (Figure

2; Table 1). To date, a total of 121 specimens have

been recaptured, of which 25 (20.7%) were

recaptured in the NS, 84 (69.4%) in the SS, and

12 (9.9%) in Pacific waters off the Chilean coast

(Table 1; Figure 3). At the same time, 3 speci-

mens from Chilean waters and 2 tagged in the

249

TROCCOLI ET AL.: MIGRATORY PATTERNS OF THE PATAGONIAN TOOTHFISH IN THE SOUTHWESTERN ATLANTIC OCEAN

Figure 2. Release position of tagged toothfish specimens in the Argentine Exclusive Economic Zone (EEZ).

area corresponding to the Argentine Patagonian

shelf, around Malvinas Islands, were also recap-

tured in the SS. The majority of tagged specimens

in each year were juveniles in size (Figure 4).

Juvenile fish consisted of 75.18% of fish tagged

across the sample period, ranging from 53.75%

(2012) to 100% (2019).

From their release site, 67.5% were recaptured

at a distance of less than 20 nm, while 15% trav-

eled short intermediate distances less than

120 nm. The smaller fraction (5%) covered dis-

tances between 120 and 400 nm (longer interme-

diate distances) and 12.5% were recaptured at

greater distances (more than 400 nm) from their

tagging locations. Only 9.9% of the fish left the

Argentine fishing ground. Most of specimens

(87.5%) covered distances of under 400 nm,

remaining inside the same sector, whose length

was estimated to be around 370 nm for both sec-

tors. In the NS, 83% (25) of tagged specimens

moved within 400 nm. Seventy-seven percent of

them (23) were even recaptured at less than 20

nm. Only 17% (5) of the total was recaptured at

distances greater than 400 nm, exceeding the lim-

its of that sector of the Argentine fishing ground.

In the SS, 89% (81) of the fish moved less than

400 nm from their release site, and 64% of the

total (58) moved less than 20 nm. The remaining

250 MARINE AND FISHERY SCIENCES 36 (3): 245-265 (2023)

Table 1. Number and percentage of toothfish specimens tagged and recaptured in the northern, southern, and Pacific Ocean sec-

tors, indicating recapture rate per year (%).

Tagged Recaptured

Year Northern Southern Northern Northern Pacific Percentage

2004 247 247 0 0 0 0.00

2005 0 0 0 0 1 0.00

2006 0 0 0 0 0 0.00

2007 14 763 0 0 0 0.00

2008 300 755 4 3 0 0.66

2009 368 685 4 5 0 0.85

2010 223 705 2 23 2 2.91

2011 37 465 1 23 1 4.98

2012 99 154 9 13 1 9.09

2013 92 244 1 4 4 2.68

2014 0 0 0 2 0 0.00

2015 0 0 2 2 0 0.00

2016 0 0 2 2 0 0.00

2017 28 32 0 1 2 5.00

2018 130 223 0 3 0 0.85

2019 0 40 0 2 1 7.50

2020 0 56 0 1 0 1.79

Total 1,538 4,369 25 84 12 2.05

Percentage 26.04 73.96 20.66 69.42 9.92

11% (10 fish) were recaptured at a distance

greater than 400 nm from the release site (Figure

5). The predominance of short distances traveled

after being tagged and released demonstrates the

affinity of the species for its residence sites in the

NS and SS of the Argentine fishing ground.

There was no clearly defined direction pattern

observed in the fish tagged in the NS (Figure 6).

Most of the specimens exhibited a distinct pattern

of movement towards the southeast and east.

However, one specimen that traveled a moderate

distance went in the opposite direction towards

the northeast. Furthermore, the specimens that

traveled the greatest distances all moved towards

the south. Only five of the recaptured fish crossed

the NS of the Argentine fishing ground, migrating

over long distances (up to approximately 1,500

nm), three of which were recaptured near the

Namuncurá/Burdwood Bank, and the rest in

southern Chile (Figure 7). A unified directional

movement was also not observed for fish tagged

and recaptured in the SS of the Argentine fishing

ground (Figure 8). Both westward and eastward

directions were observed for fish undertaking

movements less than 100 nm. Those that moved

greater distances did essentially to the west-

northwest (Figure 9). Regarding specimens

tagged in the SS, 11 of them exceeded the limits

of the Argentine fishing ground, eight of which

traveled distances in the direction of central-north

Chile, between 477 nm (885 km) and 2,100 nm

(3,889 km). The vast majority of the tagged fish

(64%) remained close to their release sites. A sin-

gle specimen was reported to the east, tagged at

54° 53' S-59° 08'W and caught at 55° 00' S-57°

37' W by the longliner ‘CFL Hunter’ in waters

near the Malvinas Islands (Martínez, unpublished

data).

The variables Days at liberty, Sector, and Sea-

son of tagging influenced on the traveled distance

in GAM results (Table 2). The final model was

defined by these three variables, which explained

34.4% of the variability of the model. The length

of fish at the time of tagging and the depth of

release were not significant in the movements of

toothfish. This hypothesis and the homogeneity

of variances seems to be adequate from validation

graphs for normality (Figure 10).

251

TROCCOLI ET AL.: MIGRATORY PATTERNS OF THE PATAGONIAN TOOTHFISH IN THE SOUTHWESTERN ATLANTIC OCEAN

Figure 3. Reported recapture positions of tagged toothfish specimens in the Argentine Exclusive Economic Zone (EEZ).

252 MARINE AND FISHERY SCIENCES 36 (3): 245-265 (2023)

Figure 4. Distribution of relative frequencies of grouped lengths of tagged toothfish specimens per year. The black dotted line

represents the length of first maturity estimated at 82 cm (Pájaro et al. 2009).

2018 2019 2020

2012 2013 2017

2009 2010 2011

2004 2007 2008

25 50 75 100 125 25 50 75 100 125 25 50 75 100 125

0.00

0.02

0.04

0.06

0.00

0.02

0.04

0.06

0.00

0.02

0.04

0.06

0.00

0.02

0.04

0.06

Total length (cm)

Relative frecuency

Figure 5. Ratio of recaptured toothfish specimens based on the range of distance traveled in the northern and southern Sector.

0-20 20-120 120-400 >400

Northern Southern

Traveled distance range (nm)

Ratio

0.0

0.2

0.4

0.6

0.8

1.0

253

TROCCOLI ET AL.: MIGRATORY PATTERNS OF THE PATAGONIAN TOOTHFISH IN THE SOUTHWESTERN ATLANTIC OCEAN

Figure 7. Possible migratory route of specimens tagged in the northern Sector of the Argentine fishing ground and recaptured

more than 400 nm away from it. Orange circles indicate the tagging position and yellow ones the recapture position. The

five specimens that traveled long distances did so in a southerly direction.

Figure 6. Migratory movements corresponding to distances of less than 400 nm in the northern Sector of the Argentine fishing

ground.