Aggregative capacity of experimental anchored Fish Aggregating Devices (aFADs) in Northeastern Brazil revealed through electronic tagging data

Catches of pelagic fish associated to anchored Fish Aggregating Devices have been responsible for increases in income, fish consumption, and even cultural identity of artisanal fishing communities in many developing countries worldwide. Nonetheless, in Brazil, aFAD fishing is still poorly developed and studied. In this experiment, FADs were anchored offshore the city of Recife (Northeastern Brazil) to investigate the potential of moored buoys in the aggregation of commercially important pelagic species near the coast, as an alternative fishing site for artisanal fishers. The behavior of acoustically tagged fish was investigated to assess whether they were attracted to the FADs and how long they remained associated to them. The results indicated that, although economically important species were found near the FADs, they did not remain associated for long periods. From the four species tagged, Acanthocybium solandri, Coryphaena hippurus, Thunnus atlanticus , and Caranx crysos , only the two latter were detected at the FADs. Both species presented a preference for a specific FAD, with stronger site fidelity being recorded for C. crysos . This species presented Total Resident Times (TRTs) of more than a month and continuous residence times of more than 14 consecutive days. T. atlanticus , on the other hand, remained around the buoys for short time intervals, with a maximum TRT of only two days. Short diurnal excursions far from the FADs and few longer excursions during nighttime were recorded for C. crysos . These results do not support the possible use of moored FADs near the coast of Recife as an alternative fishing site for artisanal fisheries. It is possible that the geomorphological characteristics of the experimental area did not favor the aggregative behavior of large pelagic fish species, such as tunas, around FADs.


INtrODUctION
Fish Aggregating Devices (FADs) have been used by fishers since ancient times to increase their catches due to many pelagic fish species' natural behavior to aggregate around floating objects (Morales-Nin et al., 2000). At first, FADs consisted just of floating debris such as trunks and palm leaves, naturally found in the ocean (Jones, 1772). Besides using these natural FADs, fishers also started to construct them, primarily of bamboos and palm leaves (Morales-Nin et al., 2000). Since the 1960s, however, modern FADs, produced with man-made materials, like plastic buoys and metal rafts, have been deployed in oceanic and coastal regions (Taquet, 2013), reaching, nowadays, tens of thousands of FADs disseminated across all five oceans (Baske et al., 2012).
Coastal and oceanic anchored FADs (aFADs) are mainly used by small-scale and sport fishing, targeting tunas, and other pelagic species (Taquet, 2013). However, in some coastal countries like the Maldives, the pole and line tuna fisheries around aFADs have attained a semi-industrial level (Adam et al., 2015). Oceanic drifting FADs are primarily used by industrial purse seiners, having tunas as their target species (Taquet, 2013). Purse seine fishing around FADs is currently responsible for more than half of the tuna catches worldwide (Parker et al., 2014). Due to the great economic importance and environmental impacts of these activities, most of the studies dedicated to the behavior of FAD-associated species, and the relationship between the fish and the FAD, have focused on tunas (Dahlet et al., 2019;Moreno et al., 2019;Oshima et al., 2019).
Catches of pelagic fish associated to FADs, including non-tuna species, have also been discussed, particularly concerning their value to sport fishing and food security in coastal communities (Bell et al., 2015;Campbell et al., 2016;. Besides the increase in fishers income, fish consumption, and even cultural identity (Albert et al., 2014;Montes et al., 2019), shifts in fishing effort from demersal species with slow growth and high longevity to fast-growing pelagic fishes may benefit demersal fish (Mbaru et al., 2018). Despite the environmental, economic, and social importance of these pelagic species, however, limited research has focused on their associative behavior (Capello et al., 2012;Forget et al., 2015;Rodriguez-Tress et al., 2017;Soria et al., 2009;Taquet et al., 2007), leaving, still, a remarkable lack of information on their ecology, fishing potential and, consequently, on the status of their populations (Gaertner et al., 2008;Moreno et al., 2016).
In Brazil, the fishing for various tuna (Thunnus obesus, Thunnus albacares, Thunus alalunga, Thunnus atlanticus, and Katsuwonus pelamis) and non-tuna species (Coryphaena hippurus, Elagatis bipinnulata, and Acanthocybium solandri) associated with oil rigs, or even with anchored oceanographic buoys, have already demonstrated their use for artisanal fisheries (Carvalho et al., 2015;Silva et al., 2018Silva et al., , 2013, although no study has been so far conducted on the associative behavior of pelagic fish species around these anchored structures. The use of aFADs in Brazil was first registered in 1984, with the deployment of anchored devices in the Continental shelf break in the Southwest region, located far from the coast, aiming to reestablish and develop commercial skipjack fishing (Katsuwonus pelamis) (Scott, 1985). Even though the results were promising, the floating structures did not resist the harsh oceanic conditions (Silva et al., 2013). In 1998, due to an agreement signed between the Brazilian government and fishing companies, six aFADs were also deployed on the Southwest coast (Lima et al., 2000). Despite the increased tuna and non-tuna catches around them (around 700 tons), financial resources were discontinued (Lima et al., 2000). On the Northeast coast, increased catches of tuna and non-tuna species have been registered around an oceanographic buoy from the Pilot Moored Array in the Tropical Atlantic (PIRATA) (Silva et al., 2013). The buoys are moored in open waters, in depths exceeding 4,000m (Silveira, 2014). The high aggregation potential of deep anchored FADs has been well studied and established worldwide (Adam et al., 2015;Bell et al., 2015;Whitney et al., 2016), including the PIRATA buoy (Silva et al., 2013), but no information is available regarding the aggregative potential of shallow aFADs in Brazil.
In this study, Fish Aggregating Devices were anchored off the city of Recife, Pernambuco (Brazil), with the objectives to investigate if the buoys aggregate economically important pelagic species near the coast and evaluate how long they would retain the fish. Aiming these objectives, the associative behavior of acoustically tagged fish nearby the coastal aFAD array was investigated using passive acoustic telemetry, to evaluate the temporal persistence of commercially important fish around the FADs. This is the first study on the associative behavior of pelagic species associated with FADs in Brazil.

MAterIAls AND MetHODs stUDy sIte AND FAD ArrAy INstrUMeNtAtION
FADs were anchored at 50 m and 200 m depth (Table 1), 20 miles off the Port of Recife, Brazil ( Figure 1). Both FADs consisted of a single float of equal size, a monitoring buoy, a stainless-steel chain, a positively buoyant rope, and four concrete block anchors (Figure 2). A third buoy from the "Programa Nacional de Boias" (G), anchored by the Brazilian Navy and The Global Ocean Observing System-Brasil (GOOS-Brasil) to collect oceanographic data, was located in the study area during the time of the experiment. Each FAD was equipped with a Vemco VR2W acoustic receiver (VEMCO, INNOVASEA, Canada). The receivers were attached at 15 m depth, from 3 November 2015 to 29 February 2016. Due to financial and logistical difficulties, a detection range test with the transmitters was not performed, but using the range calculator from Vemco´s website (www.vemco.com), it was possible to estimate the ranges for the V13 tags from 410 m to 550 m (for winds from 11 to 16 knots) and the V9 tags from 360 m to 500 m (for winds from 11 to 16 knots).

tAggINg prOceDUres
Two cruises were conducted to tag fish aboard the research boat Sinuelo (13 m length wood vessel), on 3 and 6 November 2015. The fish were captured using different techniques, including trolling, rod and reel, and handline, using circle hooks without barbs to minimize fish injuries. To capture large pelagic predator fish associated to the FADs, trolling was carried out with the boat navigating from one FAD to the other. To catch smaller fish, which are usually closely associated to the FADs, the boat was positioned right next to the buoy, and trolling was switched to   rod and reel and handline fishing. When hooked, fish were carefully transferred to a V-shaped table, where its eyes were covered with a wet cloth, and a hose was placed in its mouth to ensure the supply of oxygen. Only apparently healthy fish were measured (Fork Length-FL) and then tagged with coded Vemco V9 and V13-69kHz acoustic transmitters. The tags were surgically implanted in the fish's peritoneal cavity (Govinden et al., 2013;Meyer et al., 2000). After tagging, the fish was immediately released back to the water, and the GPS position was registered. The total duration of the tagging procedure did not exceed 2 minutes. All tagging procedures occurred during daytime hours (8:00 to 15:00). A total of 13 fish of four different species were tagged (Table 2): four Thunnus atlanticus, two Acanthocybium solandri, one Coryphaena hippurus, and six Caranx crysos. All tagged fish were captured and released closer to FAD2 than FAD1 (Figure 3).

DAtA ANAlysIs
To investigate the site fidelity and behavior of the tagged fish, the time spent by the fish around the FADs was characterized using (i) Total Residence Times (TRTs), defined as the total time spent by the fish in the FAD array, as detected by the acoustic receivers Robert et al., 2012), and (ii) Continuous Residence Times (CRTs), defined as the total detection time of a tagged fish by an acoustic receiver without absences of predetermined time intervals, known as Maximum Blanking Periods (MBP) . Two distinct CRTs were considered: (i) large-timescale (CRT), corresponding to continuous presences at the FADs without dayscale absences (MBP= 24 h), usually applied in studies on tunas and other large pelagic fish Girard et al., 2007;Govinden et al., 2013), and (ii) small-timescale (fCRT), used to analyze the fine-scale behavior of tagged fish (Capello et al., 2013a(Capello et al., , 2012;  Soria et al., 2009). fCRTs were obtained considering an MBP of 20 min, following the procedure described by Capello et al. (2015). Daytime was considered from 5:00 to 16:59 and nighttime from 17:00 to 4:59, based on sunrise and sunset times in Recife during summer-time. The departure times of C. crysos were also analyzed from fCRTs, to investigate day-night departure differences and determine if some fish could have left the FAD simultaneously, i.e., within a 20 min interval. This choice, which corresponded to the small-timescale MBP, minimized the effects of noise and collisions in the analysis . The total duration of excursions out of the hydrophone's detection range, also called Continuous Absence Time (CAT and fCAT) (Govinden et al., 2013), was obtained considering the time intervals between consecutive CRTs and consecutive fCRTs. Additionally, the maximum diel excursion distance (MED) traveled by a fish which departed and returned to the same FAD within less than 24 h was calculated considering a linear movement at a mean speed (one body length per second) (Capello et al., 2012). The maximum diel excursion distance (MED) was defined as follows: where v BL is the mean fish speed (units: m/min -1 ) and f CAT is the absence time of the fish out of the aFAD (units: min). Individual fork lengths reported in Table  2 were used to calculate the speed.
Current velocity and direction measurements were obtained from the G buoy, for the total experiment period, from 1 November 2015 to 10 December 2015. Current values were plotted against small-scale residence times (fCRTs) and absence times (fCATs) to check for current intensity and direction influences. Wind roses were plotted for the tagging experiment period (6 November 2015 and 7 December 2015) for each one of the current direction categories (North, Northeast, East, Southeast, South, Southwest, West, and Northwest). All analyses were conducted using the R statistical package (R Core Team, 2019).

AcOUstIc tAggINg
Only seven fish were detected from the 13 tagged fish: two T. atlanticus (TATL1, TATL2) and five C. crysos (CCRY1, CCRY2, CCRY3, CCRY4, CCRY5). Except for one of the C. crysos (CCRY3), which visited both FADs, the other six fish were only detected at the FAD closer to their release position, FAD 2 (Table 2). One individual of C. crysos (CCRY6) was detected only twice and, thus, excluded from the analysis.

tOtAl resIDeNce tIMes (trt)
Clear differences in the Total Residence Times were observed between the two detected species (Welch t-test; t=3.07; p < 0.05) ( Table 3). The T. atlanticus were only detected during the first two days after the tagging. The C. crysos presented, in general, longer TRTs, close to or higher than 15 days, with a mean of 16.83 d (±11.70 S.D.). The interval between the release and first detection time was also different between the two species, mainly because the two T. atlanticus were released out of the receiver's detection range, while all C. crysos were released inside the detection range. The intervals for T. atlanticus were 1.05 and 17.83 h, while for C. crysos, they did not exceed 0.2 h.

lONg-scAle resIDeNce AND AbseNce tIMes (crts AND cAts)
Ten CRTs were obtained, with a maximum of two CRTs per fish (Table 4 and Figure 4). All seven detected specimens showed CRTs at FAD2 (the FAD closer to the tagging locations, see Figure 3), whereas only one individual was detected at the two FADs (CCRY3) (Figure 4). The T. atlanticus presented residence times of less than one day. TATL1 presented two CRTs, both at FAD2, one on the same day of the tagging, and the second one on the next day. Both CRTs were of short duration with a maximum of 16.32 min. TATL2 presented one CRT, also at FAD2, with a duration of 7.93 h. C. crysos, in general, were detected for much longer periods than T. atlanticus (Welch t-test; t=3.06; p < 0.05). Apart from CCRY1, detected only in the first day after tagging, with a residence time of 7.8 h, the remainder four detected C. crysos individuals remained associated to the FAD of tagging for around 15 consecutive days. CCRY5 returned to FAD2 almost seven days after departure and remained associated to the FAD for 7.09 h. CCRY3 was the only fish to make an excursion between both FADs, being detected at FAD1 15 days after its last detection at FAD2, with a residence time of 10.17 h.

sHOrt-scAle resIDeNce AND AbseNce tIMes (Fcrts AND FcAts)
The pattern of short-scale residence times was close to the long-term residence times. However, fCATs allowed characterizing more refined scale movements of the fish. Excursions shorter than 24

cUrreNt MeAsUreMeNts
There was a clear increase in current velocity throughout the experiment, with the lowest mean current strength being registered on the first day (8 cm s -1 ± 5 SD) and the highest mean value on the last day (48 cm s -1 ± 8 SD) (Figure 4). Three wind directions were most frequent and presented the highest values, North (N), Northeast (NE), and Southwest (SW) ( Figure 5). Three current intensity peaks were registered, but, in general, the mean current strength increased continuously (Figure 4). During this period, no predominant current direction was observed when current speeds were below 20 cm s -1 . However, at current intensities higher than 20 cm s -1 , they were exclusively from North and Northeast direction until the 30th day of the experiment (3 December) and from Southeast direction from day 31 until the end   (Figures 4 and 5). Also, all C. crysos left FAD2 when the mean current intensity started to increase strongly, with some fish being detected later, but for a couple of minutes only (Figure 4). CCRY3 showed up at FAD1 on the 34 th day of the experiment (7 December), when the Southwest current, present since day 30, was predominant and had the highest strength recorded for the whole tagging period ( Figure  4). However, possibly due to the small dataset, no clear correlation could be found between fish departure and current intensity.

FIsH AssOcIAtIve beHAvIOr
From the 13 tagged fish, two blackfin tunas (T. atlanticus), two wahoos (A. solandri), and one dolphinfish (C. hippurus) were never detected. Up to date, there is few available information on the behavior of these species around FADs (Addis et al., 2006;Dempster, 2004;Dempster and Taquet, 2004;Girard et al., 2007;Sepulveda et al., 2011;  2019). Studies with ultrasonic transmitters have shown that fish return to the FAD vicinity after swimming longer distances than the receiver's detection range (Holland et al., 1990;Matsumoto et al., 2013;Weng et al., 2013). A detection range test was not performed in the present study, so it was not possible to determine the maximum detection distance of the fish from the FAD (Kessel et al., 2014); nonetheless, the VEMCO calculator presented a conservative detection interval of almost 150 m (410-550 m for V13 and 360-500 m for V9). Therefore, the likelihood of distant associations for these species remains low, suggesting they probably left the area right after the release, due to tagging stress , natural displacement and/or migrating behavior (Maguire et al., 2006) or because they had a higher attraction to the shelf break (Dubroca et al., 2013;Holland and Grubbs, 2007).
A preference for the deeper FAD (FAD2, 200 m depth) was registered for all species. This could be explained by the proximity to the tagging location (since the fish were released closer to FAD2 than FAD1). However, the tagging locations already suggested that fish were more abundant near the buoy deployed in deeper areas. Another possible explanation of the higher number of fish present at one of the two FADs relies on the competition among the two FADs, which were close to each other (distance less than three km). Previous studies demonstrated that tuna could detect a FAD up to distances of ten km (Girard et al., 2004;Holland et al., 1990;Marsac and Cayré, 1998). In the presence of social interactions (Capello et al., 2011;Sempo et al., 2013), even if the two FADs can be considered equivalent, one of the two FADs could present higher fish biomass than the other. This scenario would imply alternating associated biomass between the two FADs over time.
Both T. atlanticus individuals that were tagged and detected stayed around the FADs for short time intervals (maximum of 48 min), and a maximum Total Resident Time of only two days. Studies on the associative behavior of this species are not yet available, but residence times of tunas around FADs have been extensively studied, varying significantly among study sites, tuna species, and size classes Govinden et al., 2013;Rodriguez-Tress et al., 2017). Robert et al. (2013) categorized tuna behavior around FADs in three groups, briefly passing near a FAD, short association, or long association. The former is generally an association of a couple of minutes, as observed in the present study, suggesting they did not show associative behavior to the experimental FADs. Even with decades of scientific efforts to understand how tuna species behave around floating objects, the reasons why the fish are attracted or aggregate around these devices are still not well understood (Girard et al., 2004;Rodriguez-Tress et al., 2017). Several hypotheses are proposed to explain association periods (Fréon and Dagorn, 2000), but based on the short-term visits recorded in the present study, four main reasons were considered as most plausible. By the indicator-log hypothesis, fish would explore floating objects to assess information on the species present in the area, such as possible preys or predators. In the comfortability stipulation hypothesis, if local conditions are favorable at the time, fish could use the FADs as quick resting areas. Moreover, a floating device could equally be used as a meeting point to form larger fish schools. It should also be considered that, since both tunas presented the same behavioral pattern, they possibly did not associate to the buoys due to a combination of non-favorable local abiotic and biotic factors, such as water temperature , prey availability (Graham et al., 2007), and the presence of congeners associated with the FAD (Capello et al., 2011). The current velocities registered when the tunas visited the FAD were low, ruling out their influence in the short visits observed. However, since only two individuals were detected, further studies should be conducted to confirm the short association times found here.
The residence times obtained for C. crysos, with TRTs of more than a month and fCRTs of more than 14 consecutive days, suggested a strong site fidelity to the FADs. C. crysos probably explored the FADs in search for food supply and protection (Gooding and Magnuson, 1967;Hunter and Mitchell, 1968;Rountree, 1989). This species has a preference for coastal waters, but is also naturally found inhabiting areas near open-ocean features, including floating or fixed objects, due to the increased food availability around them (Brown et al., 2010;Castro et al., 2002;Holland and Grubbs, 2007). Sinopoli et al. (2019) have demonstrated a strong correlation between the extensive use of FADs and the expansion of C. crysos geographical distribution in the Mediterranean Sea, where the FAD network is probably facilitating the species' retention in coastal waters, providing food availability and protection against predators.
Faster excursions far from the FADs during the day and smaller number of longer excursions during the night were recorded for C. crysos. The daily excursions may be used to explore and feed, with the FAD acting as a reference point (Gooding and Magnuson, 1967). The maximum diurnal excursion distances found for C. crysos (267.2 m ± 68.3 S.D.) suggest that they did not move far from the FAD area and could not have visited the other FADs during these excursions. Nocturnal excursions were less common, with fish staying closer to the FADs. When the nocturnal excursion occurred, nonetheless, the maximum calculated excursion distances were higher (10.3 km ± 5.5 S.D.), and fish could have swum to much farther areas, not returning to the FAD.
The observed general pattern of independent C. crysos departures, plus the observation of certain synchronicity in departure events of some individuals, suggest the existence of small fish schools rather than a large one, with small groups leaving the buoy within short time intervals and others remaining associated to the FAD . However, the only simultaneous departure event does not guarantee that individuals have physically exited the FAD together. Similar synchronous patterns of small fish schools around FADs have also been observed for other small pelagics (Soria et al., 2009).
Although the departure of all C. crysos coincided with an increase in current intensity, the available data do not show a clear correlation between C. crysos departures and current speed. Capello et al. (2013) studied aggregations of a small Carangidae species, Selar crumenophthalmus, using acoustic tagging and found aggregations' position to be shifted upstream, with increasing distances from a moored FAD with increasing current intensity. As observed by Capello et al. (2013), C. crysos, which is also a Carangidae species, could have changed its distribution around the FAD in stronger currents, moving further from the FAD or upcurrent. The data indicate that fish left the FAD mostly when mean current values were between 18 and 20 cm s -1 ( Figure  5). However, the available data do not allow us to assess the tagged fish's position at a fine spatial scale. Remarkably, CCRY3 was observed to depart from FAD2 during a strong southwest current, occurring at FAD1 (located southwest of FAD2) 15 days later. The current direction could have influenced this fish after leaving FAD2 because the shallower FAD1 was located in the current path. However, other factors, such as the arrival of predators or changes in food availability, may explain the departure of C. crysos.

INsIgHts FOr tHe DevelOpMeNt OF FAD FIsHerIes IN cOAstAl wAters IN NOrtHeAst brAzIl
This study assessed for the first time the potential use of shallow aFADs off the Northeast coast of Brazil. Even though the overall number of tagged individuals was low (N=13), half of the dataset was constituted by large pelagic fish species of commercial interest. The large tagged fish were either never detected or showed short residence times at the experimental FADs. Even though passive acoustic studies on pelagic fish of commercial interest around FADs usually consider a higher number of tagged individuals (Forget et al., 2015;Govinden et al., 2013;Stehfest et al., 2013;Taquet et al., 2007), the proportion of detected fish was considerably higher (>70%) than in the present study (<30%). Such discrepancy corroborates the hypothesis that these species may not have had aggregative behavior around the studied FADs. Few studies have also obtained information regarding pelagic fish associative behavior, such as tunas and the dolphinfish, with few individuals tagged (N=5 and 13) (Muir et al., 2012;Whitney et al., 2016). Telemetry studies with pelagic fish other than elasmobranchs in Brazil have only been published for Thunnus albacares with three tagged individuals (Travassos et al., 2009), and for Istiophorus platypterus with four fish (Mourato et al., 2014), highlighting the importance of the present results. The acoustic telemetry articles published with sharks and rays (Bezerra et al., 2019;Branco-Nunes et al., 2016;Ferreira et al., 2013;Niella et al., 2017) also dealt with similar or lower number of tagged fish compared to this study.
From the four tagged species (T. atlanticus, A. solandri, C. hippurus and, C. crysos), only C. crysos had a strong site fidelity to the buoys; nonetheless, all individuals left the study area within a month. Despite this species being frequently captured by recreational and artisanal fishing in Brazil (mostly line and trap fishing), and one of the main Carangidae species captured in the Northeast region, its commercial value is much lower than the other tagged species (Haimovici et al., 2014;Lessa and Nóbrega, 2000;Lima et al., 2018).
Dolphinfish and tunas are commonly found associated to other coastal aFADs around the world and have been responsible for increases in income, food security and livelihoods of local artisanal fishers in different regions (Albert et al., 2014;Bell et al., 2015;Montes et al., 2019). Such aggregating devices were deployed in oceanic islands, where insular shelves are relatively narrower than continental ones (Quartau et al., 2014). In these regions, despite aFADs being deployed only a few miles from the islands, they were located in open waters, where the surface buoy was probably the only reference point used by the fish in search of food supply, protection against predators (Hunter and Mitchell, 1968;Rountree, 1989), cleaning stations (Gooding and Magnuson, 1967), or meeting points (Dagorn et al., 1997;Fréon and Dagorn, 2000). Therefore, the geomorphological characteristics of our study area might not have favored the aggregative behavior of large pelagic fish around the aFADs.
Small-scale artisanal fishing boats, mainly from Northeastern Brazil, are currently targeting yellowfin, T. albacares, and bigeye tunas, Thunnus obesus (Lowe, 1839), close to the open-ocean buoys from the Pilot Moored Array in the Tropical Atlantic (PIRATA) (Silva et al., 2013). The buoys are moored in open waters more than 300 nm from the nearest port (Areia Branca-RN) (Silveira, 2014). Thus, an anchored FAD's deployment close to the shore was considered an alternative to aggregate tunas in a more accessible site to local fishers. However, the present results may indicate that the location the FADs were moored did not favor fish aggregation because of its proximity to the coast. Eighani et al. (2019) conducted experiments with shallow coastal FADs in the Persian Gulf and reported similar results due to the short distance to shore. Nearshore aFADs, thus, may not be an effective alternative strategy to artisanal fishers from Brazilian Northeast coast. Notwithstanding, further studies, based on larger numbers of individuals, should be conducted to confirm these findings.

cONclUsIONs
From the four pelagic species (T. atlanticus, C. hippurus, A. solandri and C. crysos) tagged around the experimental aFADs, only C. crysos demonstrated site fidelity behavior, possibly using the FADs for food supply and protection. The possible non-aggregative behavior observed for the other species may be due to a combination of non-favorable local conditions, such as depth, prey availability, water temperature, and the presence of congeners associated with the aFADs.
This work offers a first assessment of the potential use of nearshore anchored FADs as a possible fishing alternative, in a region where the artisanal fisheries sector is in a difficult situation due to overfishing, consequently facing declining productivity of most exploited stocks, such as lobsters and demersal fish species (Silva et al., 2018). However, the results suggested that the studied aFADs may not be effective in aggregating pelagic fish of commercial importance, such as tunas and the dolphinfish.
For future work on this field, we suggest telemetry experiments with an increased number of tagged species and specimens. Long-term monitoring periods, as well as using complementary techniques, such as active telemetry and acoustics, would also contribute to a better understanding of fish behavior, fish spatial distribution and movement patterns around FADs. Proper monitoring of the fishing activities and research on fish composition, reproduction, behavior, and stock status will also be essential to ensure the aggregative potential and sustainability of such a new fishing alternative.

AcKNOwleDgeMeNts
We are deeply grateful to the staff responsible for the tagging logistics, Rafael Muniz, Bruno César, Gleidson Tavares, and Carlinhos. We would like to thank Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for granting the Master's degree fellowship to LQV. We thank the two anonymous reviewers for the important suggestions made, which contributed considerably to improving the manuscript's quality. This study was supported by "Projeto de Implantação de Atratores de Tunídeos e Afins em Meia Água na Plataforma Externa do Litoral de Pernambuco-ATUNA", funded by FINEP and CAPES-CIMAR II.