The bycatch of piramutaba, Brachyplatystoma vaillantii industrial fishing in a salinity and depth gradient in the Amazon estuary, Brazil

ABSTRACT The piramutaba, Brachyplatystoma vaillantii is a freshwater catfish that is the most abundant fishery resource in the Amazon estuary. Piramutaba trawling is done on industrial fishing scale and is characterized by the presence of many freshwater and marine bycatch species, with and without commercial value. Here we describe the bycatch of the industrial fishery of piramutaba in the Amazon estuary and evaluate the relationship of two important environmental factors, depth and salinity, with the accidental capture of freshwater and marine fishes in the Amazon estuary in the rainy and dry seasons. We identified 21 cartilaginous fish species (19.1% freshwater and 80.9% marine) and 125 bony fish species (25.6% freshwater and 74.4% marine). The bycatch included 64 species without commercial value (43% of all bycatch species), which are always discarded. Freshwater and estuarine fishes exhibited significantly higher abundances in shallower environments, while marine fishes were similarly abundant along the entire depth gradient. On the contrary, the abundance of freshwater fishes significantly decreased, and that of estuarine and marine fishes significantly increased with increasing salinity. Regarding the conservation status of the bycatch species, one is classified as vulnerable (VU), and seven as critically endangered (CR). The information on the bycatch of piramutaba fishery in the Amazon estuary is important to subsidize regional fisheries policies and the management of protected areas.


INTRODUCTION
The piramutaba, Brachyplatystoma vaillantii (Siluriformes: Pimelodidae) is a freshwater catfish that is the most abundant fishery resource of the Amazon estuary, formed by a panmitic population along the Solimões-Amazonas River (Barthem and Goulding 1997;Formiga et al. 2021).This catfish is explored by artisanal fisheries using gillnets and long lines and industrial fisheries using pair or triple trawling (Barthem, and Goulding 1997;Ruffino 2003).The industrial fishery of the piramutaba began in 1971 with the use of adapted shrimp boats (Silva et al. 2016), reached the largest landing (22,486 tonnes) in 1977, and occupied the ninth place on the list of exportation goods from Pará state, Brazil by 1980 (Dias-Neto et al. 1985;Barthem 1990;Dias-Neto 1991;Prestes et al. 2022).The piramutaba trawl fishery is limited to the Amazon estuary facing the Marajó Island and the coast of state of Amapá coast, Brazil (Prestes et al. 2022), a region characterized by the extensive input of sediments by the Amazon River (Geyer et al. 1996), with enormous oscillation in the discharge of the Amazon River, which, at its peak in May-June, more than doubles the minimum discharge, which occurs in November (Guimberteau et al. 2012).The region is located between the North Cape (01º42'S and 49º55'W) and the Maguari Cape (00º15'S and 48º25'W) (Figure 1), encompassing a 80-mile coastline stretch with a depth range between 3 and 20 m (Santos et al. 1984;Dias-Neto et al. 1985).
Trawling has low selectivity, therefore producing a high proportion of accidental captures (Perez and Pezzuto 1998;Diamond et al. 2000) of species with and without commercial value, the latter being discarded (Alverson et al. 1994;Clucas 1998).Accidental captures may have a significant impact on local biodiversity (Clucas 1997) by altering the community structure and food webs (Anderson et al. 2013), which makes trawling management challenging (Davies et al. 2009).The correct identification of the bycatch is fundamental for establishing the guidelines for fisheries zonation, which should be based on all the populations affected by the activity (sensu Manthey and Fridley 2009).Therefore, the incomplete knowledge of the fish diversity of the bycatch of the piramutaba industrial fishery hampers the development of effective protection measures for the local fauna and the management of the ecosystems affected by the activity (Thrush et al. 1998;Greenstreet and Rogers 2004;Juan and Demestres 2012).

AMAZONICA
In this study, we describe the bycatch of the industrial fishery of the piramutaba in the Amazon estuary by integrating secondary data from the literature with primary data from the Japan International Cooperation Agency (JICA), the Centro Nacional de Pesquisa e Conservação da Biodiversidade Marinha do Norte (CEPNOR-ICMBIO), and personal databases from collaborators.Considering that salinity and depth were the main forces that structured the distribution patterns of species in the Ariidae family in the Amazon estuary (Soares et al. 2021), we analyzed the influence of these two factors considering the abundance of all fish captured by trawls in the same area.For this, we grouped the fish into three groups, marine, estuarine and freshwater, and tested the abundance of these groups in response to salinity and depth in the periods of the year when the discharge of the Amazon River is maximum and minimum.

Study area
The study area was undertaken in the inner estuary of the Amazon River mouth, in the fishing area of the Brazilian bottom pair trawler fleet (Figure 1).The area is characterized by the shallow muddy bottom and the dynamism of the mixture of the freshwater of the Amazon River and the marine water (Curtin and Legeckis 1986;Eisma and Marel 1971).The annual discharge cycle of the Amazon River ranges from 230×103 m 3 s −1 in the first half of the year to 103×103 m³ s −1 in the second half (Guimberteau et al. 2012) and causes the displacement of wedge salt along the shelf of the Amazon mouth, being the most critical force acting on the salinity structure of the plume (Geyer et al. 1996).The dry season is marked by the arrival of brackish waters to the coast in May, and by the predominance of fresh waters from December (Barthem and Schwassmann 1994).

Database
We compiled a species list of bycatch from available data of published studies on the bycatch of the industrial outrigger trawling fleet of the North coast do Brazil (Barthem 1985;Jimenez et al. 2013;Silva et al. 2016;Klautau et al. 2016), primary data from the Japan International Cooperation Agency (JICA) from 1996 to 1997, and Research and Conservation National Center of Northern Marine Biodiversity (CEPNOR, Brazil) from 2016 to 2020.
The collections by JICA, Museu Paraense Emilio Goeldi (MPEG), and Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA) were performed between 1996 and 1997, sampling in an area divided into 1330 blocks of 3′ latitude by 3′ longitude each and nearly nine square nautical miles in total.Samplings were carried out during three trawling expeditions in the dry period of 1996 (August to September), the rainy period of 1997 (March to April), and the dry period of 1997 (August to September).Each seasonal survey comprised 120 trawling stations distributed randomly and proportionally to the area of three strata of isobaths: 5-10 m, 10-20 m, and 20-50 m.In each station, fishes were sampled by bottom trawling conducted by a pair of vessels operating for 30 min along the current direction at a constant 2-3 knots speed.Standardized trawls were carried out at daytime by trawlers equipped with bottom trawl nets of 58.2 m length and 14 cm between opposite knots, with all collected specimens identified and counted.
The collections by CEPNOR were carried out under SISBIO license # 44915-7 and all species collected by CEPNOR are represented by voucher specimens in the ichthyological collection of Museu Paraense Emílio Goeldi (MPEG) in Belém (Pará, Brazil), the zoological collection of Universidade Santa Cecília at Santos (AZUSC) in Santos (São Paulo, Brazil), and the Laboratory of Fish Biology and Genetics (LBP) at Universidade Estadual Paulista Júlio de Mesquita in Botucatu (São Paulo, Brazil).Images and voucher numbers of all marine species captured by CEPNOR are available in Marceniuk et al. (2020;2021a).Species were identified using the descriptions and collection keys provided by Van Der Sleen and Albert (2017) and Marceniuk et al. (2021a).Species were further classified as primary marine, estuarine, or freshwater following Nelson (2016) and as pelagic, demersal, or benthonic following Marceniuk et al. (2021b).The market value for the piramutaba and all other bycatch was obtained at Ver-o-Peso Market (Belém, Pará state, Brazil) by WSRS, APM and JPRJ.The conservation status of the species followed the Red Book of the Brazilian Fauna Threatened with Extinction (ICMBio, 2018), and its updates (MMA 2022).
Depth and salinity information were obtained during the sea-borne survey in the Amazon River Mouth by JICA, concomitantly with the collection of the primary data on fish composition.Salinity was measured from the surface to the bottom with an STD (Alec Denshi AST200PK STD, salinity measuring capability 0-40 ‰, depth range 0-200 m, measurement distance 0.2 m).

Statistical analysis
We evaluated the importance of seasonality and two environmental variables (depth and salinity) to the abundance of freshwater, estuarine, and marine fish using the data sampled by JICA between 1996 and 1997.As the data distribution was heteroscedastic (according to Levene's test), we used the Kruskal-Wallis H test to evaluate differences in the abundance of freshwater, estuarine, and marine fish between the rainy and dry seasons.We then evaluated the importance of depth (average depth at hauling) and salinity (average salinity at hauling) to the average abundance of freshwater, estuarine, and marine fishes by fitting generalized additive models (GAMs) integrating a smoothing parameter ACTA AMAZONICA to consider the nonlinear relationship of the variables.GAMs were performed using the default gam function from the mgcv package (Wood 2011) in the R environment (R Core Team 2021).All analyses considered a 5% level of significance.
The bycatch included 64 species (43% of all bycatch species) with no commercial value (Figure 2c), which are always discarded, and 81 species (57%) with some commercial value (Table 1).Among the latter, nine were more valuable than piramutaba, 11 had similar value to piramutaba, and 61 had lower value than piramutaba and were only occasionally traded (Table 1).Regarding the conservation status of the bycatch species (Figure 2d), 109 (74.7%) were classified as safe or least concern (LC), six as near threatened (NT), one as vulnerable (VU), and seven as critically endangered (CR) (Table 1).Among the critically endangered species, six were cartilaginous fishes, including the endemic Isogomphodon oxyrhynchus, the sawfish Pristis pristis and three species of hammerhead sharks (Sphyrna), and only one was a bony fish, the Atlantic goliath grouper, Epinephelus itajara.We highlight that 23 of the identified species were not evaluated (NE) or data deficient (DD), eight of them endemic to the area (Table 1).

Effect of depth and salinity
The standardized samples obtained by JICA occurred in sites with depth ranging from 6 to 50 m, and salinity near the substrate ranging from 0.05 to 36 ppm.The occurrence and abundance of freshwater, estuarine, and marine fishes between the seasons and along the depth and salinity gradients was not random.While estuarine fishes did not differ in abundance Comm value = commercial value: target species (A), commercial value higher than target species (>), commercial value lower than target species (<), equal commercial value equal to target species (=), no commercial value (*).Habitat: freshwater (FW), brackish water (BW), marine and estuarine (M).Habit: pelagic (I), demersal (J), benthic (K).End = endemic to the influence area of the Amazonian-Orinoco plume and its drainages.Status = conservation status in the Brazilian List of Endangered Fauna (ICMBio 2018): CR = critically endangered (CR), vulnerable (V), near threatened (NT), least concern (LC), data deficient (DD), not evaluated (X).ACTA AMAZONICA ACTA AMAZONICA between the dry and rainy season (H (1) = 0.851; p = 0.356), the abundance of both freshwater (H (1) = 20.712;p < 0.001) and marine fishes (H (1) = 6.462; p = 0.011) was significantly higher during the rainy season.Depth and salinity affected the abundance of freshwater, estuarine, and marine fishes differently (Figure 4).Freshwater (GAM; R² = 0.132; p < 0.001) and estuarine fishes (R² = 0.024; p = 0.002) exhibited significantly higher abundance in shallower environments, while marine fishes had similar abundance along the entire depth gradient (R² = 0.001; p = 0.221).On the contrary, the abundance of freshwater fishes significantly decreased (R² = 0.444; p < 0.001), and that of estuarine (R² = 0.067; p < 0.001) and marine fishes (R² = 0.211; p < 0.001) significantly increased with increasing salinity.

DISCUSSION
The bycatch of piramutaba fishery is composed of species with and without commercial value, being that 20 to 30% of the total capture is discarded (Dias-Neto et al.Ruffino 2003;Jimenez et al. 2013).We showed that almost 45% of all captured species in the bycatch have no commercial value, while the remaining species range from lower to higher market value than that of the piramutaba.This information is crucial for optimizing management criteria for the use of the bycatch of piramutaba.Some bycatch species are already overexploited fishery resources, such as Brachyplatystoma rousseauxii (Garcia-Vasquez et al. 2009;Agudelo-Córdoba et al. 2013).
The bycatch of the industrial fishery that targets piramutaba, a primary freshwater species, is mainly composed of marine fishes, while about 25% of the species are freshwater fishes (see Table 1).This association between freshwater and marine species in the Amazonian estuary is unique compared to other oceanic basins worldwide, due to the enormous oscillation in the discharge of the Amazon River (Guimberteau et al. 2012).
The abundance of estuarine fishes did not vary significantly between the dry and the rainy seasons (minimum and maximum discharge of the Amazon River), while freshwater species were more abundant during the rainy season and marine species were more abundant during the dry seasons.Seasonal variation in rainfall is a major environmental factor determining the taxonomic and functional composition of fish communities in estuaries (Barletta et al. 2005;Castillo-Rivera 2013;Molina et al. 2020).The composition turnover is usually associated to: (i) life-history traits by species using different habitats for reproduction (Dantas et al. 2010;Fontoura et al. 2019) and (ii) physiological constraints due to changes in salinity level (Lisboa et al. 2015;Smyth and Elliott 2016).
The seasonal variation in salinity is more important for primary marine or primary freshwater species than for estuarine species, because the latter display higher osmoregulatory efficiency and can better cope with this environmental variation (Becker et al. 2011;Marshall 2012;Park et al. 2020).Consequently, the abundance of estuarine fishes tends to be stable between seasons, while abundance of marine and freshwater fishes is higher in the season when salinity levels are lower (rainy season) or higher (dry season), respectively, as corroborated by our GAM results.
The association of fish species with depth is more complex than with salinity, depending on habitat use and osmoregulation.Freshwater fishes have lower osmoregulatory efficiency and might occupy the shallower waters along the coastline both because of lower salinity levels and of adequate habitat structure.Estuarine fishes can better cope with salinity variation, but are usually inhabit or migrate seasonally to shallower areas and more structured environments (Akin et al. 2003;Barletta et al. 2005;Dantas et al. 2010).Marine species might use shallower areas with lower salinity levels in specific periods of their life cycle (Smyth and Elliott 2016).Therefore, the spatial pattern of variation in salinity levels in the Amazon estuary, where lower salinity is frequently ACTA AMAZONICA associated with shallower areas, is important in shaping the distribution of fishes and the turnover of species with different affinity levels (Soares et al. 2021).
The majority (90%) of threatened species in the bycatch are elasmobranchs, six of them critically endangered and three vulnerable.Other threatened elasmobranch species that occur in the area, but were not reported in the piramutaba bycatch, include the stingrays Fontitrygon colarensis (VU), Paratrygon aiereba (CR) and Pseudobatos percellens (VU) (ICMBio 2018; MMA 2022).The potential negative impact of the fishery on the populations of these endangered species has not yet been assessed.

CONCLUSIONS
The present study poses an important advance on the knowledge of the fauna captured during piramutaba trawling in the mouth of the Amazon River and the influence of the hydrological cycle of the river on the structure of this community.This information will support the establishment of criteria for management for the use of the bycatch of the piramutaba fishery in order to improve the definition of areas and periods allowed for the operation of this fishing fleet.Given the observed gaps in the knowledge of the ichthyofauna in the region, we highlight the importance to increase the knowledge on the biology and fishery stocks of the endemic fauna, especially sharks and batoids.

Figure 1 .
Figure 1.Location of the Amazon estuary (small map, in red) and detailment of the study area between North Cape (one asterisk) and Maguari Cape (two asterisks) by the mouth of the Amazon River.This figure is in color in the electronic version.

Figure 2 .
Figure 2. Composition of the bycatch of piramutaba fisheries in the Amazon estuary, northern coast of Brazil.A -Taxonomic and marine-estuarine/freshwater composition (%) of Elasmobranchii; B -Taxonomic and marine-estuarine/freshwater composition (%) of Teleostei; C -Economic value of the by-catch species: higher than Brachyplatystoma vaillantii (red), equal to B. vaillantii (orange), lower than B. vaillantii (yellow), no commercial value (grey); D -Conservation status: critically endangered (CR), vulnerable (VU), near threatened (NT), data deficient (DD), not evaluated (NE), least concern (LC).This figure is in color in the electronic version.

Figure 4 .
Figure 4. Generalized additive models (GAM) with smoothing parameters describing the relationship of depth (upper) and salinity (bottom) with the average abundance of freshwater (left), estuarine (middle), and marine fishes (right) of the bycatch of the piramutaba fisheries in the Amazon estuary.This figure is in color in the electronic version.