LOW-COST TECHNOLOGY FOR FISH MONITORING APPLIED TO THE FISHING OF TWO SPECIES OF PACU IN AMAZONAS, BRAZIL*

The present study aimed to investigate the “pacu” species’ commercialized in the “Colônia dos Pescadores Z-31” in the municipality of Humaitá, Amazonas State, from May 2018 to April 2019. For this purpose, we developed a method for raising a diary database in the colony, obtaining variables such as production, effort, catch per unit effort (CPUE), fishing spots, amongst others. We have identified two species of “pacu” Mylossoma aureum and M. duriventre with productions of 10.5 and 9.5 tonnes respectively which generated an income of U$ 22,173.00. Captures occurred mostly randomly, due to the region’s fishing characteristics although the CPUE was higher when compared to bigger cities. We could also find evidence over their migratory cycles from a correlation between the catching sites and the period, where: i) during the falling water (May to July), the “pacus” start to leave the streams and concentrate in lakes; ii) during the dry season (August to October), the “pacus” perform a second migratory cycle, leaving streams and lakes and going to rivers; iii) finally, during the rising water, “pacus” once again move to small streams. The generated information can be used in fishing strategies in the region, reducing costs with inputs.


INTRODUCTION
Fishing is one of the most important human activities in Amazon, constituting a source of food, commerce, income and leisure for local populations, especially those living in the riversides of both large and small-sized rivers (Meggers, 1977). According to some authors (Cerdeira et al., 1997;Batista et al., 2004;Santos and Santos, 2005), fish consume index in the Amazon is one of the greatest in the world, varying from 369 g person -1 day -1 to 600 g person -1 day -1 , which characterizes it as one of the main protein sources for locals. Considering that there are other sources of protein available, these Igor Hister LOURENÇO 1 *  Marcelo Rodrigues dos ANJOS 1  João Pedro BARREIROS 2  numbers represent a high consume index (around one fish person -1 day -1 ).
Among the most harvested species, the pacu (Mylossoma spp.) is one of the most captured in Brazilian Central Amazon. From Porto Velho (Goulding, 1979;Doria and Lima, 2008;Doria et al., 2012) to Manaus (Petrere Junior, 1978); and from smaller municipalities like Iranduba and Benjamin Constant (Doria et al., 2018) to bigger ones like Santarém (Isaac et al., 1996), the pacu is always one of the most captured species, demonstrating its commercial acceptance and economical importance. However, lots of species from the sub-family Serrasalmidae (Myleus spp., Mylossoma aureum e M. duriventre, Metynnis spp.) are commonly known as "pacus", being the "pacu manteiga" (Mylossoma duriventre) one of the most representative in landings at Porto Velho, as pointed by Doria and Lima (2008).
This fact, when considering monitoring points such as the Municipal Market of Humaitá, can prove to be an obstacle in the exact determination of which species of "pacu" are explored in the region since the providers for data survey are not concerned about a taxonomic differentiation and thus compromising a precise estimate.
Despite recent studies performed in the region (Sales et al., 2011;Lima et al., 2016), little is known about the "pacu" and its capture dynamics. This study is then an opportunity to expand our knowledge on the "pacu", its bioecology and investigate the production from the artisanal fishing from the Middle Madeira in the adjacency of Humaitá municipality between May 2018 and April 2019 using low-cost social technology to achieve the data.

MATERIAL AND METHODS
The study area includes rivers, streams, and lakes around the municipality of Humaitá, Amazonas State, Brazil (Figure 1) located under coordinates 7º30'22"S and 63º01'15"W.
Considering the nature and territorial extension of the Middle Madeira Basin (Gibbs, 1967;Goulding et al., 2003;McClain and Naiman, 2008;Queiroz et al., 2013), and other similar regional studies (Sales et al., 2011;Lima et al., 2016), the authors opted for sampling in ports due to: convenience in data collection; logistics, since all catches should pass through the port (Fisherman's Colony); and the already ongoing partnership between the LIOP (Madeira River Valley Laboratory of Ichthyology and Fishery Ordering) and The Colony.
The Colony in question (Z-31), accounts for nearly one thousand and five hundred associated fishermen and is located at the municipality of Humaitá, Amazonas. We -unlike Sales et al. (2011) and Lima et al. (2016), who obtained the data from the local market -obtained our data from daily monitoring the arriving boats in the rush hours (6 am to 9 am) with the aid of teams. For that purpose, we developed an adequate method for the local conditions, which is described below.
Each team responsible for daily data collection had four members with different duties during landings: a) the first one counted and took notes of the catches' common names; b) the second one collected at least one representative individual from each species to take its biometric information (size and weight), paying attention to identify each individual according to the common name proposed by the first member; c) the third one took pictures of each representative individual and identified it according to its common name; and d) the fourth member interviewed the owner of the fishing boat for the collection of socioeconomic and other useful information (days spent fishing, number of persons involved in the activity, when it happened, type of embarkation, description of the capture location, among others). Finally, all the data had been systematically written down with the aid of the landing form. Subsequently, in the lab, another team was responsible for analyzing the photographic records and to identify the collected specimens to the highest possible taxonomic level.
From that methodology, we were able to obtain: total number of captured individuals; weight; length; fishery production; the commercial value of each species; days spent fishing; the number of fishermen per boat; type of boat; and local of capture. Catch Per Unit of Effort (CPUE) was then estimated. All these variables were submitted to descriptive statistics with the aid of the R Software, defining mean, standard deviation (SD), maximum and minimal values, and the Coefficient of Variation (CV).
We obtained the Effort (Equation 1), the CPUE (Equation 2) and the CV (Equation 3) according to the following equations: where: NDF is the number of days spent on fishing; and the NF is the number of fishermen involved where: CV is the coefficient of variation; SD is the standard deviation; and X is the mean.
The authors opted for represent the variable "Effort" with the number of fishermen involved times the amount of days spent fishing, as recently used by Silva Junior et al. (2017), due to the difficult to extract information about how the fishing gears were used (number of repetitions and how long the remained in place), considering the used method to collect information. In addition to that, it is a classic method proposed by Petrere Jr. (1978), to measure "Effort", especially when considering the Amazonas State particularities.

RESULTS AND DISCUSSION
During the study, we inventoried a total of 162 fishing boats with the presence of pacu. The total production for the pacu between May 2018 and April 2019 was approximately 20 tonnes, divided into two species: Mylossoma aureum and M. duriventre, with 10.5 and 9.5 tons respectively.
The mean production per boat along the year (Table 1) was 113.36 kg, with a CV of 173.12%. This low productivity associated to a high CV can be explained by two factors: a) the composition of the fishing fleet, which is composed predominantly by motorized canoes and a small parcel of fishing boats ( Figure 2); and b) the non-selective fishing methods practiced by a considerable part of the fishers, which lands many species of low commercial value, commonly known as "salada". The same explanation can be applied to the variable effort.
The information acquired corroborates other studies developed in the region by Sales et al. (2011) and Lima et al. (2016). The existence of two types of boats (small and big) (Figure 2), directly influence the transport capacity, which explains the maximum value of production ( Table 1). The non-selective fishing explains the minimum production value (Table 1) since the captured pacu in these cases are the product of random captures ("salada") . In other words, the local fishermen don't go out to capture a specific specie, they capture whatever get caught in the mesh. This generates a great problem for the fishing monitoring in the region, once the disembarked fishes are composed by several species, commonly known as "salada". In this light, the present method can be considered a success for managing to identify, at a taxonomic level, which species are being captured in the region. Despite only the "pacu" appear in this study, the data used is only a small cut of a bigger ongoing research.
The same explanation applies to the differences observed between the observed mean and CV for the variables Effort and CPUE (Table 1), in other words, the capacity of a canoe and a fishing boat to carry fishermen and fish varies significantly. Although, the mean value found for CPUE (24.25 kg fisherman -1 day -1 ) is still above the indexes found in bigger cities, such as Porto Velho, RO (upstream of Humaitá, AM), which is around 15±21 kg fisherman -1 day -1 (Doria and Lima, 2008), evidencing a high fishery production in the studied area, possibly for being a less explored region, when compared to the upriver capital. However, when compared to other regions of the Amazon, the high values of CPUE do not reflect an expressive production (Petrere Junior, 1978;Goulding, 1979;Isaac et al., 1996;Doria and Lima, 2008;Doria et al., 2012Doria et al., , 2018 justified once again by the predominance of motorized canoes (Figure 2).
The pacu price did not vary significantly along the year, as stated by the low CV found for this variable (Table 1), but there are some variations related to the supply and demand relations, where the highest price (USD 1.55) occurs at the end of the closed season (around March), while the lowest price (USD 0.66) occurs when the supply overcomes the demand, at the beginning of the flood season (around May and April). Considering the mean price (USD 1.14), by multiplying it with the total production (≈ 20 t), we have an approximated value of USD 22,173.00, once again, evidencing the socioeconomic importance of the pacu in the region.
Regarding the mean length, the observed values are inside the limits established by the decree IBAMA GEREX AM Nº 01/2001, which defines a minimum capture size of 15 cm for the pacu. Despite the minimum value of 12 cm (1.23% of the landings), most captures respected the size established by decree, which is proved by the SD.
We noticed that the captured pacu concentrates in certain times of the year (Figure 3) associated with two hydrological periods: the beginning of the lowering water (May/July) and the final of the drought (August/October), corroborating with Doria and Lima (2008).
The predominant captures in specific periods can be associated with the reproductive ecology observed in some species of Mylossoma, which are short/medium migrators, typic from "várzea" systems (Cella-Ribeiro et al., 2016). This suggests that the capture peaks in the drought period ( Figure 3) can be associated with the beginning of the migration period, which begins at the rising water. Despite classic literature consider October as a drought season, as pointed by Barthem and Goulding (1997), recent studies (Cella-Ribeiro et al., 2016) suggests that October characterizes the de beginning of the rising waters.
Consequently, the flood season presents the lowest production rates (Figure 3), once the fish shoals are in small streams and other flood environments for the spawning season. Finally, during the end of the drought and beginning of the lowering water, when the fish shoals return to the rivers after spawning, it's possible to notice the return of captures, characterizing a second migratory cycle.
The evidence above is reinforced by the environments of catch along the year (streams, lakes, rivers, and igapó/várzea), which varied with the hydrological season ( Figure 4).
The variation of capture environments ( Figure 4) and the information provided by Cella-Ribeiro et al. (2016) about the ecology and biology of fishes from Madeira River, when used  together, allow us to trace the pacu migratory routes along the year, where: i) during the lowering water, the pacus that moved to small streams and flooded environments during the flood season begin to return to the lakes; ii) during the drought, the pacus perform a second cycle, moving from small streams and lakes to rivers; iii) at last, during the flood season (not present in this study due to the lack of information during the closed season), the fishes migrate once more to small streams and flooded environments.
The generated information about the capture environments according to the hydrological season, in addition to corroborating with research about fish ecology in the region, may also be used to orientate strategies to lower the input expenses of the activity. In other words, by knowing when and where a certain fish species will be in a certain period of the year, it is possible to lower the number of days spent on fishing, and consequently, expenses related to supplies (food, fuel, ice, etc.).
Finally, it is important to mention that the collected data used a social technology of low costs based on an experimental method adapted to local fishing reality. The present method integrated interdisciplinary participation of people from different social fields, such as the researchers, who elaborated the monitoring strategies; the fishermen, who provided the information; and students, who were responsible for the data collection and systematization.

CONCLUSIONS
During the study period, that occurred from May 2018 to April 2019, we identified two pacu species (Mylossoma spp.), which, combined, sum up a total production of 20 tons, where 10.5 tonnes are composed by M. aureum and 9.5 tonnes by M. duriventre.
The regional fishing is almost completely artisanal, with the predominance of motorized canoes with a high diversity of low commercial species commonly called "salada". This, by its turn, drastically influenced the variables Effort, CPUE, and Production. However, despite this high fishing variability, the CPUE kept above the levels found in studies performed in close urban centers, evidencing the conservation levels of the inventoried stocks.
We could also find a relation from the capture environments with the hydrological season, evidencing the pacu (Mylossoma spp.) cycles. This information can be used to orientate regional fishing strategies, enabling the application of the economics principle in the expanses with transportation, feeding, and storage of the product.