Analysis of Migratory Catfish Production from Artisanal Fishing in the Middle Madeira Sub-Basin Using New Monitoring Methods, Southwestern Amazon

,


Introduction
Due to the richness of species found in the Amazon region, the fshing activity stands out compared to other neotropical regions, whether by the number of explored species or the dependence of riverine communities and urban populations on the activity [1].However, according to the Food and Agriculture Organization of the United Nations (FAO), much of the data from artisanal/small-scale fsheries is still unknown [2].
Amazonian fsh species adapt to seasonal changes in the environments they occupy.Terefore, knowing these changes is fundamental to understanding the abundance and composition of fsh resources and, consequently, defning assertive management policies [3][4][5][6].
Te ecological function of the Siluriform order, more specifcally from the Pimelodidae family, and their inter-or intraspecifc interactions are fundamental for environmental balance [7][8][9].Also known as "catfsh," they are responsible for dispersing seeds throughout the basin through zoochory, evidencing the important ecological role played by aquatic biota when using the fooded forests as habitat and source of food [10][11][12][13].
During the upstream journey of migratory catfsh (fooding season), while recruiting for reproduction, some species are a target of intense fshing, characterizing a period known as "piracema."During this period, which usually extends from July to November [14,15], the percentage of captures increases in some regions of the Madeira River basin.However, the beginning of the migration comprises the months of November, December, and January, with the spawning season usually occurring during the food period (which comprises the months of February, March, and April) and adapting to each region's hydrological cycles [16].
Te fshing activity developed in the middle Madeira River sub-basin presents small-scale/artisanal characteristics due to the use of simple fshing gear, short fshing trips of low catch, and multispecifc captures [17,18].From 2001 to 2017, an estimated average of 213 tons of fsh/year was captured in the municipality of Humaitá, but these numbers, especially in the lower portion of the Madeira River, have continually decreased over the years [19,20].Some authors attribute this decrease in catch to demographic expansion and higher demand for fsh [21], while others attribute it to the construction of hydropower dams [20].
Given the sizes that some Pimelodidae species can reach, they have excellent market acceptance.Also known as "large migratory catfsh," some species, such as the Pseudoplatystoma punticfer, P. tigrinum, Brachyplatystoma rousseauxii, and B. flamentosum, fgure as one of the most important commercial freshwater fsh species in South America [10,22,23].Tis good market acceptance, consequently, generates pressure over the exploration of some species of this family, explaining part of the decreases in catch observed in the last decades.
Parallelly, starting in 2008, the Jirau and Santo Antônio hydroelectric plants (UHE) were built upstream from Humaitá, Amazonas, in Porto Velho, Rondônia, and inaugurated in 2012, creating a physical barrier and fooding a considerable area.Evidence shows that the Pimelodidae family sufers strong infuence from the hydrological cycles [12,20,24], which coordinates the migration cycles, telling them when and where to go [10,25].Based on that, recent studies suggest that the construction of hydropower dams along the Madeira basin has occasioned a dramatic decrease in migrations and possible local extinctions in some parts of the basin [20,26,27].
Reliable catch data are essential to understand these processes better [28].However, the growing demand for fsh, associated with the lack of information from artisanal/smallscale fsheries, hampers the production of reliable datasets and, consequently, weakens the policies of the sector, occasioning fsh stock overexploitation and indirectly changing the structure and functioning of freshwater ecosystems in medium and long term [29,30].In addition, climatic efects, pollution, and habitat degradation can also cause changes in fsh stock abundance and capture dynamics [31,32].
Te efectiveness of fsheries' good economic performance and proftability relies on the exploited environments' productivity, the exploited species' value, and the fsheries' duration over time [33].Tus, understanding the dynamics of small-scale fsheries, the activity potential risks, and adopting prophylactic management decisions based on reliable data are the key steps to defne how to exploit this resource appropriately [34].According to Castello [35], however, artisanal/small-scale systems require a diferentiated management approach adapted to each region's social and ecological specifcity.
In this light, this study uses the data produced by a new and more refned fsheries monitoring method [18]-based on a social technology under a "citizen science" perspective, where the researchers daily monitor the fsh landings with the aid of the involved fshermen, photographing, taking biometric measures, and quantifying each of the landed species-in the municipality of Humaitá, Amazonas, Brazil, to characterize the migratory catfsh commercial production of the Pimelodidae family landed in the Z-31 Fishermen's Colony Dr. Renato Pereira Gonçalves, bringing new and more reliable information over the fsheries dynamics of pimelodids in inland artisanal/small-scale systems.

Study Area. Te study area includes the middle Madeira
River sub-basin region, between the Aripuanã and the Machado rivers.Almost entirely distributed in the southern portion of the Amazonas state and a small part located in the state of Mato Grosso [22], the Aripuanã, Manicoré, Matauará, and Marmelos rivers stand out as its main tributaries.It covers rivers, streams, and lakes located near the municipality of Humaitá, Amazonas, Brazil, compressed between the coordinates (7 °53′17.06″Sand 62 °52′51.5 2″W and 6 °46′12.6″Sand 62 °27′34 0.0″W) (Figure 1).
Te middle Madeira region already has some generic ongoing fshing management policies in place, which consists of a closed season that aims to protect the reproduction of some fsh species, including some pimelodids, as established by the Normative Instruction MMA No. 18, from October 14th, 2004 (Hypophthalmus spp.);IBAMA Ordinance No. 48 from November 5th, 2007 (Hypophthalmus spp.);CEMAAM Resolution No. 18, from September 16th, 2014, and amended by CEMAAM No. 21, from October 27th, 2015 (Pseudoplatystoma tigrinum and Pseudoplatystoma punticfer), which prohibits the fshing of certain species of commercial interest from November 15th until March 15th.Although prohibited, the fshing of migratory catfsh did not stop in this period.

Data Collection.
We gathered the data used in this study by monitoring the artisanal fshing production of the Fishermen's Colony Z-31 Dr. Renato Pereira Gonçalveswhich counts 1,500 members and is located in the middle Madeira basin, in the Humaitá municipality, Southeast Amazonas, Brazil-landed from May 2018 to April 2019, using the low-cost social technology applied to the monitoring of the artisanal fshing proposed by Lourenço et al. [18]-a protocol based on the daily monitoring of the fsheries landings with structured questionnaires, adapted to the multispecies captures present in the region [19,30].Te method consists of four steps, described as follows: Step 1.When the embarkation arrives, the technician must head to the landing point, present the work, and ask for authorization to take the biometric measures of 2 Journal of Applied Ichthyology each species.Te technician then positions himself close to the transfer point and keeps track of how many and which species are being landed.In this part of the process, having a good relationship with the fshermen is essential once they can separate the species in different boxes during the landings, making the work easier.
Step 2. Technician is responsible for taking note of every new species landed and taking its biometric measures in consonance with the information provided instep 1 (e.g., species 1, species 2, and species n).
Step 3. Technician is responsible for fling the identifcation card according to the information provided insteps 1 and 2 and photographing one individual of each species for taxonomic identifcation.
Step 4. Technician is responsible for interviewing the fsherman and collecting information like when, where, and how the fsh were captured, along with other socioeconomic information, such as investments and proft.
By the end of each landing, all technicians must reunite and cross information to fll out the landing form.Tese pieces of information allow us to identify the following in each landing: the fshing grounds, communities, efort variables (fsherman.daysfshing −1 ), type of environment, species, individual and total mass, length, type of embarkation, and fshing gear used.Finally, using the available literature, we use the photographs to identify the landed species at the lowest possible taxonomic level [36].
During the monitoring, approved by the Brazilian Ethics Committee, we collected the following variables: total number of captured individuals and their respective species; weight; length; captured volume (kg); commercialization price of each species; days spent fshing; the number of fshermen per boat; type of vessel; where they captured it; and type of environment (e.g., stream, river, and lake).Te coordinates of each fshing ground and community were taken with in loco visits using a GPS (Global Position System).

Data Analyses.
To analyze the pimelodids captures' relation with the regional hydrological cycles, we collected the average quotas data from the fuviometric station "15630000 Humaitá" and defned the minimum and maximum hydrological values for each period.Tese values allowed the visualization of how the average monthly quotes of the Madeira River levels changed during the analyzed period and to delimitate the following periods: fooding (from August to October), food (from November to January), ebb (from February to April), and drought (from May to July).We analyzed the landings' catch (kg) per species using descriptive statistics to obtain the following parameters: mean, standard deviation, minimum, maximum, coefcient of variation, and kurtosis.To verify the distribution of the captured volumes (kg), we used Shapiro-Wilk's normality test [37].For the homogeneity of variances analysis, we used Levene's test [38,39].To verify the existence of signifcant diferences in the landed volume between the Pimelodidae species, we applied the nonparametric Kruskal-Wallis [40] and Dunn's post hoc test of multiple comparisons, which allowed us to visualize the diferences expressed by p values.
We adopted ordering and classifcation analyses to assess the spatial-temporal fshing dynamics, reducing the amount of interpreted information to obtain an interpretive grouping.Te two multivariate techniques applied in this study intended to search for emerging interpretable patterns, namely: (i) nonmetric multidimensional scaling (NMDS), which produces orderings from distance matrices, but instead of using the real distances, it only considers the distance "ranks" [41][42][43], and (ii) cluster analysis for indicating the aggregation of the most similar data and, through this, allowing comparisons of the data [44].
We also used nonmetric multidimensional scaling (NMDS) to defne the fshing seasonality as well as the spatial-temporal dynamics of the analyzed Pimelodidae species, uncovering the solution for a certain number of dimensions by ordering it to minimize the STRESS function (standardized residual sum of squares).To do that, we elaborated a quantitative matrix of the captured volumes (kg) for each species, using the Euclidean measure to determine the geometric distance in the multidimensional space.Ordering patterns difer according to the number of ordering axes; for the validation of the evidenced structure, we used a signifcance test based on the "null model" of the Monte Carlo permutation method [42,45], present in the PAST 4.05 statistical package [46].
To determine the similarity of the captured volumes between the capture sites, we used a classifcation analysis, where we created a quantitative matrix of the captured volumes (kg) using the Bray-Curtis index [47], along with an agglomerative method based on the group means (UPGMA) indicated for decreasing the distortion of the original matrix during the construction of the dendrogram [48].We also used the cophenetic correlation coefcient to validate the dendrogram with the original matrix [49].
To assess the efect of the monotonic relationships between the descriptor variables, we applied Spearman's correlation coefcient (r s ) for the following variables: hydrological level, type of environment, days spent fshing, number of support vessels, and distance between the capture sites and the landing site [50].Finally, we used the canonical correspondence analysis (CCA) to verify the infuence of each variable on the distribution of captured volumes (kg) of the thirteen analyzed fsh species [51,52].

Results
Te total production of catfsh species of the Pimelodidae family addressed in this study was 6013.93 kg for the entire period, totaling around 1,689 fsh from 10 genera and 13 species.We evaluated a total of 186 landings from 122 fshermen, from which 98 were motorized canoes and 24 fshing boats, from 24 diferent fshing grounds; the total captured volume (kg) was distributed between rivers (71%), lakes (18.8%), and streams (10.2%).Te identifed species are available in Table 1.
Te average production of pimelodids in this sample was 32.33 ± 89.84 kg (0.27 min-980.00max) per fshing trip, which lasted an average of 3 days.Te average sale value in each landing was R$5.57/kg (2.00 min-12.00max) paid by the "magarafes," who resale the products after processing for up to R$20.00/kg.
Te highest volumes of migratory catfsh landings occurred between July and September, corresponding to the region's transition period from the ebb to the drought period.Te fshing gears most commonly used were stationary gillnets (96.72%), followed by throwing nets (1.64%), and "espinhéis" (stationary longlines with several hooks) (1.64%).Te gillnets mesh sizes varied from 45 mm to 190 mm, and the 70 mm mesh was the most used.
Descriptive analyses showed variation in the volume caught (Table 2) and revenue (Table 3) per fshing trip for the thirteen species of migratory catfsh analyzed.For the period studied, the highest values observed for each fshing trip, both in volume caught and in revenue (R$), were for the "surubim" fsh, Pseudoplatystoma punticfer.
Te values obtained from the Kruskal-Wallis nonparametric analysis of variance, used to test the diferences in volumes captured for each trip, were represented by H � 73.48 and p � 1.83E −28 , demonstrating that the medians for the analyzed samples presented signifcant differences, and represented by the multiple comparisons posterior Mann-Whitney test (Table 4).
Trough nonparametric multidimensional scaling (NMDS), we ordered the captured volumes (kg) according to the location, type of capture environment, and seasonality.Te groupings were arranged according to the amount of biomass registered for the diferent months of the year (Figure 2).Trough NMDS, it was possible to verify that the highest production for the analyzed locations occurred between the ebb and drought periods, with low production during the food period, a fact possibly linked to the beginning of the closed-season period.It was possible to observe that the volumes varied spatiotemporally due to the availability of diferent environments during diferent periods of the year.Most of the monitored fshing grounds presented low captures, evidencing the artisanal/small characteristics of the studied region.Linear adjustment of NMDS for the captured volumes in kilograms (kg) presented a determination coefcient (R 2 ) of 0.8262 for dimension 1 and 0.492 for dimension 2, with a residual sum of squares equal to STRESS of 0.1269.
Trough the similarity dendrogram (Figure 3), we identifed fve groups with adjustments higher than 55%    Te monotonic relationships tested for Spearman's correlation coefcient (r s ) between the descriptor variables presented signifcant values for nonlinear associations (Table 5; Figure 4).Spearman's correlation coefcient (r s ) allowed us to detect the monotonic relationships between the descriptor variables pair by pair, highlighting the uncorrelated variables.

Journal of Applied Ichthyology
Te canonical correspondence analysis (Figure 5) presents the infuence of environmental descriptors (months, hydrological level, type of environment, number of days fshed, number of support vessels, and distance between     the capture sites and the landing site) on the volume captured in kilograms of Pimelodidae catfsh for the analyzed area.Leiarius marmoratus, B. vaillantii, P. pirinampus, H. marginatus, and Phractocephalus hemioliopterus positively responded to changes in the hydrological level and the number of days fshed.While the production of P. punticfer, S. lima, P. tigrinum, and S. planiceps was positively infuenced by the environment and the number of support boats, the distance between the capture sites and the landing site negatively infuenced the observed catches, reinforcing the observed higher production values found in "Igarapé do Buiçú," "Tambaquizinho," and "Três Casas lake."Te frst two axes of the canonical correspondence analysis explained 67.87% of the variation in the catch distribution of Pimelodidae species.Te descriptors considered to be signifcant in the distribution of abundances in the catch of the species were the month, hydrological level, type of environment, number of days fshed, number of support vessels, and the distance between the capture sites and the landing site (Table 6).

Discussion
Te characteristics present in the gathered data, such as multispecifc exploration, diversity of equipment used, limited types of environments, and variations in food pulses, in addition to the ecological conditions of the populations, infuence fshing strategies, abundances, and composition [53][54][55][56][57], characterizing the fshing of migratory catfsh in the middle Madeira River sub-basin region as a multispecifc and subsistence commercial fshery.
Pseudoplatystoma punticfer, popularly known as "surubim," "pintado," or "cachara," is one of the Pimelodidae species with the highest commercial value and the one with the higher proportion in landings for the analyzed region.Tis species was responsible for 69.7% of the total production among Pimelodidae species, with approximately 4 tons and 8.3% of the 46.7 tons landed in the region, ranking ffth in total production, corroborating other studies [58,59].According to Goulding [22], the genus Pseudoplatystoma Bleeker, 1862, was the most important among those exploited in fsheries on the Guaporé and Mamoré rivers in the 1970s.According to Payne [60], 95% of fsh landed in Trinidad and Bolivia were composed of four main species, two of which are within the Pimelodidae family: P. punticfer and P. tigrinum.Santos [61], describing the composition and situation of fshing in the state of Rondônia, mentions that in 1984, the production of P. tigrinum was 41.2% of the total 104.5 t landed in Vila Pimenteira on the banks of the Guaporé river.Petrere Jr.
[62] estimated P. tigrinum as 48% of the captures in the Mamoré River.More recently, Lopes et al. [30] observed that Pseudoplatystoma spp accounted for more than 10% of the fsh landed in Boca do Acre, Amazonas, in 2012.
Te main diferences observed in production and travel revenue for each Pimelodidae species were analyzed using exploratory techniques and multivariate statistics for interpretation purposes [63,64].Tus, the descriptive analyses showed signifcant diferences when analyzed separately for each species, where P. punticfer had the highest captured volume and revenue values.Te Kruskal-Wallis nonparametric analysis of variance detected the diferences observed for the production of the evaluated species, and their values were represented by the multiple comparisons post hoc Dunn's test, pointing once again to P. punticfer as having the most signifcant production and revenue, unlike the other Pimelodidae species analyzed.Tis fact can be explained by the great fexibility of the species P. punticfer in occupying diversifed environments throughout their entire ontogenetic development [12,65], which in turn justifes that their production is seven times greater than that of P. tigrinum, which is the second most frequently captured large catfsh species in the study area.
Regarding the parties involved, we identifed three production agents working in the landings of commercial fshing in the middle Madeira River region: fshing boats (who focus their catches on the mainstream of the rivers), motorized canoes (which explore lakes and streams), and the "magarefes" [19,22].
A peculiar feature is the fshing in stream environments during the drought season, an environment frequently explored by fshermen in motorized canoes [19].Tis modality presents great expressiveness in production, reaching the highest values and surpassing, in some cases, the production of periods expected to be more productive for migratory catfsh.On the other hand, during this period, shoals are more frequent in the main channel of large rivers, and the fshing efort of the feet is mainly concentrated in these types of environments [15,22] with trips lasting on average less than one day [19].However, for the studied region, these trips last around three days.
Another factor that explains the diferences observed in production between fshing trips is due to the numerous reproductive strategies that exist among Pimelodidae species, such as other migratory species from tropical environments, which have developed mechanisms that enable them to optimize the environment [66][67][68][69].Spawning occurs at the head of rivers during the rainy season when the rivers are full of turbulent and oxygenated running water.Young and adult individuals of species such as P. punticfer are found in adjacent fooded areas during the food period [70].During the ebb and drought seasons, the distribution is diferent.Young fsh are found in streams and marginal lakes, while adults are found only in the main riverbeds  [71,72].Tis diference explains the high capture rates of P. punticfer throughout the period but with the highest relative frequencies during the ebb and drought periods, thus corroborating the data observed in the present study.Nonparametric multidimensional scaling analysis showed that capture volume for diferent Pimelodidae species varied spatiotemporally as a function of the types of environment and months.Tis tendency for variation in catch volumes is caused by the preference of some species for certain types of habitats during diferent times of the year, which consequently leads fshermen to undertake a higher capture efort in environments such as lakes and small adjacent streams [11,15].Te period considered most productive was during the ebb and drought season, comprising the months from May to October, with the highest values of capture per trip occurring in May and September, respectively.
Cluster analysis showed the formation of groups defned by the similarity of volumes captured for the considered fshing grounds, demonstrating the efect of the types of environments for each location (Figure 3).It was possible to observe that the capture volumes varied among the diferent locations, infuenced by fuctuations in hydrological level, which interfered with the capture dynamics for the analyzed area.
Studies on Pimelodidae have shown that the highest revenues revolve around the species B. rousseauxii and B. vaillantii due to the greater demand for these species, as well as a higher fshing efort [34,73], mainly in the river mouth region [74].However, artisanal fshing, especially in the middle Madeira River subbasin region, does not have a feet of boats with the proper characteristics to capture these species.Generally, specifc equipment and vessels with a large storage capacity (tons) are necessary to capture larger catfsh species [75].
Based on the local reality, although Barthem and Goulding [23] describe the capture of species of the Pimelodidae family as occurring from July onward, in the middle Madeira region, captures of P. punticfer and P. tigrinum were observed occurring from April onwards, which for the study area corresponds to the fnal period of reproduction for these species when individuals fnd themselves with a low accumulation of reserves and depleted gonads [76,77].
It is essential to consider that fsh assemblages can be directly afected by changes in environmental conditions, such as the increase and/or reduction of fooded areas, food availability, and harbor areas [78].Te results showed that for the analyzed period, the Buiuçú stream was the most productive, with a total capture of 1,333 kg, followed by the Tambaquizinho with 618 kg and, in third place, Três Casas Lake with a production of 573 kg.Tese numbers highlight the importance of adjacent lakes and streams for fshing activity in the region, especially during drought periods when these environments present higher captures [79].
Some authors, when dealing with the production of catfsh, observed that the greatest catchability occurs during the food period due to the beginning of recruitment to carry out the reproductive migration.However, shoals form during the ebb, with the decrease in the supply of environments, increasing their catch [34].Tis study corroborates with our data, where the highest production recorded occurred in August, September, and October, respectively, during the lowest water level recorded [80].
We can observe that P. punticfer and P. tigrinum are responsible for the highest abundances in captured volume when considering production per trip and total biomass, refecting the observed revenue.However, it is noteworthy that the region's fsheries are nontargeted hybrids.Fishermen catch whatever gets trapped in the mesh, which explains the low number of species and individuals captured when analyzed separately [18,81].
Another critical factor is that during fshing trips, vessels do not have adequate conditions for fsh storage and transport, which, in turn, also limits the captures and consequently generates low values in revenue due to the efect of underexploration.Most boats used in catfsh fshing in the studied area are motorized canoes with expanded polystyrene (Styrofoam) boxes [82].Studies report that motorized canoes, popularly called "rabeta," represent the majority of vessels and are responsible for more than 90% of the storage and transport capacity of fsheries in the region [83].
Vessels like motorized canoes limit fshermen to shorter and more frequent trips, usually to places close to the landing areas, which demands a low cost for fsheries, leading to a higher number of fshing trips, but with fewer days fshed, increasing the number of records [55,84,85].In this way, fshermen adapt fshing techniques to exploit abundance cycles and local market demands [34].Te information generated with the data obtained can explain the values observed for the revenues in relation to the evaluated locations and species.In addition, the precariousness and lack of infrastructure during the capture's landing compromise the fshing activity's development, increasing costs for its execution, which may refect in the price of commercial fsh.
Tus, catfsh have a higher commercial value when compared to the other species that come from this study area [18], whether because of their commercial appeal as well as the difculties imposed by the specifc catches observed for the group, with emphasis on the minimum lengths and weights allowed established by ordinance No. 48/2007 and some state legislations [1].Other difculties include the lack of access to ice for fsh conservation, the lack of guarantees of access to closed-season insurance, and the lack of necessary inputs to manufacture equipment, preventing the development of fshing activity in the region [86].
Due to the requirements imposed by local commercial consumption, as well as the limitations in storage space for large volumes of fsh, some species of catfsh have the head and viscera removed before being brought to the landing site; these body parts represent between 20 and 30% of the individual fsh's volume, adding to the total weight and thus infuencing the fnal total weight of fsh caught.Tis practice occurs in several regions, involving the fshing of Pimelodidae catfsh [87].Undoubtedly, this factor may have contributed to the low production observed for some species when considering the weight obtained.

Conclusions
Te Pimelodidae catfsh production in the middle Madeira River subbasin region is a small-scale fshery and plays a vital role by providing food and a source of income for local populations.When analyzing the catfsh production expressed as volume and the type of capture environments, we found that, despite being partially organized, fshing in the region needs investments concerning the acquisition of vessels with greater transport capacity and storage, improvements in the foating platform where the fsh are landed, improvements in working conditions, better access to ice, and necessary supplies for the development of this type of fshing, in addition to updating public policies aimed at this sector, based on local realities.
Te presence of large hydroelectric projects in the region may represent one of the causes of low productivity found in this study, especially when considering that these types of projects consist of artifcial barriers that can interfere with the reproductive dynamics of the analyzed group, and the registered mean productions presented here are lower when compared to other similar studies that cover periods before the installation of the hydroelectric dams.
Finally, despite the short history dataset available in this study, the refned data produced by the monitoring method allowed us to characterize the pimelodids' production dynamics in the region.Tis characterization is the frst step to circumvent the lack of systematic and continuous information available about artisanal fshery production in Amazon's inland waters and highlights the importance of the expansion of more refned monitoring methods adapted to each region's specifc realities.

Figure 1 :
Figure 1: Capture areas and characterization of diferent types of environments recorded for fshing landings made in the municipality of Humaitá, Amazonas, of migratory catfsh (Siluriformes: Pimelodidae) registered in artisanal fshing in the middle Madeira sub-basin, between 2018 and 2019.

Figure 2 :
Figure 2: Nonparametric multidimensional scaling for fsh catch volumes (kg) for each fshing trip in diferent types of environments and catch months for Pimelodidae species recorded in artisanal fsheries in the middle Madeira sub-basin, between 2018 and 2019, in relation to localities (dots correspond to locations and capture environments; black dots represent rivers; red dots represent lakes; blue dots represent "igarapés" or stream).

Figure 3 :
Figure 3: Similarity dendrogram generated through cluster analysis of captured volumes (kg) of Pimelodidae species, for each location, using the Bray-Curtis index as a function of months and capture locations (cophenetic correlation coefcient � 0.8849).

Table 1 :
Recorded Pimelodidae catfsh captured during the artisanal fsheries monitoring in the middle Madeira sub-basin, Amazonas, Brazil, between 2018 and 2019.Te frequency of species occurrence (presence/absence) is expressed as a percentage.

Table 2 :
Descriptive statistics of biomass production (kg) of migratory catfsh (Siluriformes: Pimelodidae) recorded in artisanal fshing in the Madeira River sub-basin, between 2018 and 2019, according to location and hydrological period.

Table 5 :
Scores of statistical values and two-tailed probabilities for nonexistent correlations among those described considered for the analysis.

Table 6 :
Canonical correspondence analysis (CCA) score for the infuence of descriptor variables on the volume of fsh caught (kg) of Pimelodidae species for the middle Madeira River region between 2018 and 2019.