Introduction

The Amazon is a global public good in constant transformation, mainly due to human action. It has great potential for aquaculture development, because it has abundant freshwater and edible fish¹. In addition, aquaculture contributes to food security, as it is a source of protein and economic income^2-4, prevents overfishing of rivers and lakes (some wild species are close to extinction) and promotes sustainable development in the region^5,6.

The Peruvian rainforest is made up of five departments: San Martin, Amazonas, Madre de Dios, Loreto and Ucayali, its average GDP (Gross Domestic Product) of Ucayali, Amazonas, Loreto, Madre de Dios from 2007 to 2016 was 4.05 %, San Martin 5.6 %⁷. Aquaculture in Peru has grown by 13 % in the last decade, based on three species (trout, fan shells and prawns) which account for 95 % of total production⁸. In 2019, Peru's aquaculture harvest was 161279.12 t⁹ The main actors in aquaculture in the Peruvian rainforest are communities and small farmers who entered into this activity as a result of State promotion with a welfare-based approach¹⁰. Both the development opportunities and their impact must be evaluated in their real magnitude, especially the limited resource aquaculture types (AREL)¹¹.

The Department of Ucayali is divided into 4 provinces and 17 districts¹². Aquaculture is practiced in all provinces, mainly in Padre Abad and Coronel Portillo, which together have 95 % of AREL type Aquaculture Centres (CA) and micro and small enterprise aquaculture (AMYPE), which, according to the legal norm, the Regional Production Directorate of the Regional Government of Ucayali promotes their development, but not adequately because it has an insufficient number of extensionists to attend them (Daniel Velarde, 2020, personal communication, director of the Aquaculture Area) and therefore it is not known how they carry out their activities¹³. Padre Abad is known as a coca-producing area, linked to illicit drug trafficking, and represents an area of interest for the Peruvian state¹⁴.

In 2014, the National Programme "A Comer Pescado" reported that in Loreto and Ucayali per capita consumption was 51.6 and 42.1 kg/year respectively, being the two regions with the highest fish consumption in the country¹⁵. Fishing in the Amazon basin is under pressure due to overfishing, which is showing signs of depletion and possible collapse of fish stocks¹⁶, fish farming represents the greatest opportunity for food security and increase in fishermen's income^17,18. The Amazon has the world's largest freshwater reserves and easily accessible land. Aquaculture is therefore a strategic choice to promote regional development⁵.

The objective of this study was to characterize the infrastructure, farming system, social and economic conditions of fish farmers in the department of Ucayali, comprising the districts of Padre Abad, Neshuya, Campoverde and Calleria.

Materials and methods

In Ucayali there are 637 CA, 93.7 % are in the provinces of Padre Abad and Coronel Portillo. From these, the districts with the highest number of CA were selected, Table 1, representing 54.6 % of the CA in the department. The sample consisted of 88 CA.

The sample was drawn considering access to the CA, the conditions of the state of health emergency due to the COVID-19 pandemic, as well as the protective measures issued by the Ministry of Health. Prior to the application of the survey, 3 aquaculture extensionists in the area were coordinated to contact the aquaculture farmers to be surveyed in order to ensure their acceptance. A questionnaire constructed ad-hoc for the research was applied to obtain data on economic and social conditions (educational level, housing, access to health system, family, technical training, employment, investment and aquaculture production) as well as facilities (infrastructure, water resources) and culture system (species, feeding, culture control, period and harvesting).

The surveys were cleaned by eliminating surveys that had complete or contradictory information, and a database was created in Excel.

The present work was carried out from July to December 2020. The field visit was carried out from 14 to 25 September 2020.

Results

In Ucayali, 73 % of the CA are AREL, 8 % have a production of 3.5 t, 42 % of the AMYPE category produce <3.5 t (Table 2).

61.4 % of CA depend on water provided by rain, limiting control of supply, AMYPE 12.5 % with low production. Other water sources (spring, stream, stream, well) with low flow Table 3. Water from rivers and lagoons, due to their high contamination, is not used²⁰. 22.7 % of the water is collected by pumping, 77.3 % is collected using the gravity of the rainwater that falls directly into the pond (no transfer system, 38.6 %) or because the water collection point (from well, stream, river and lagoon) is at a higher level than the ponds and is transferred mainly with PVC pipes of small diameter (54.6 %).

Water quality analysis of CA ponds producing <2 t in the AREL category is performed by 22.8 %, in the AMYPE category 2.3 % do analyse the water. Those producing 2-3.5 t among the ARELs and do not test the water are twice as many as those who do. Overall, 55.6 % of the CA do not test water, with the highest proportion (31.8 %) among those producing <2 t (Table 4).

Aquaculture production is atomised in small CA with a low level of production (Table 5). 54.6 % of the CA in Ucayali have between 1 and 2 culture ponds with a production level of less than 2 t per season. There are CA with 1 to 3 ponds producing between 3.5 to 10 t per season. Productions higher than 10 t are recorded in CA with 4 to 6 ponds. There is one case with more than 10 ponds and a production of 10 to 20 t.

Regarding the number of ENUs, 22.7 % of the CA have one inoperative pond, half of them have a total of two ponds. Those with two inoperative ponds account for 7.9 %, half of them have 4 ponds in total (Table 5). 6.8 % of the CA have 3 inoperative ponds, 9.0 % have 4 to 6 inoperative ponds. Table 6 shows the reasons for non-use of the ponds, the main reasons are: lack of fry (47.5 %) and lack of financing (27.5 %).

Those that produce <2 t have an average of 0.34 ha of water surface and an average investment of S/ 3382, and those that produce 2 to 3.5 t invest an average of S/ 12184, which are the extreme values (Table 7).

Regarding the investment/ha of the water mirror, <2 t of production invest S/10064.3/ha, in the range 2-3.5 t they invest S/13820.9/ha. In the 3.5-10 t range they only invest S/8602.4/ha, in the 10-20 t range they invest S/10365.5 per hectare, a value similar to those that produce <2 t. At the regional level, it has an average water mirror of 0.67 ha with an average investment of S/7131, while the investment per hectare of water mirror is S/10682.6 and the investment per fingerling planted is S/ 1.12.

Figure 1 Average water surface (ha), fry planted and investment (S) of the aquaculture centers, categorized by the number of ponds constructed 

The average value of fingerlings per water mirror (ha) grows with the number of ponds, from 3 thousand in CA with 1 pond to more than 20 thousand in those with 10 or more ponds. Figure 1 shows the correspondence of the water mirror with the number of ponds, it increases in the CA with 5 or more ponds. The data from the water mirror and fingerlings have a normal distribution (the Shapiro-Wilk test was applied). In the Pearson test, a high correlation (0.853) was recorded between the number of ponds with the investment and the hectares of the water mirror, including its linear regression with an R² of 0.728.

The use of lime varies from one species to another, in the crops with paco and paiche they use 2667 kg/ha. In the case of boquichico cultivation, they use only 71 kg of lime/ha (Table 8). 93 % of CA use lime. According to Table 9, the CA use an average of 422.7 kg of lime/ha. In the range of 10 to 20 t, they use 185.3 kg/ha of lime. The CA with the production range of 3.5 to 10 t use 545.2 kg cal/ha.

The average fertilizer used per hectare of water mirror in CA is 8588 kg, predominantly organic (43 %), especially those that produce <2 t. Those that produce 10 to 20 t use 257.8 kg/ha. Those that produce <2 t use an average of 7698.5 kg/ha. 36 % of CA do not use fertilizer.

Fingerlings buy it from IIAP because it sells it cheaper than private laboratories (LP), which are important producers of fingerlings (Table 10) that together supply 87.6 % of CA or 95.3 % of stocked fingerlings.

68.3 % of CA AREL buy fingerlings from LP and/or IIAP, the same as 100 % of those that produce 3.5-10 t. 100 % of those who produce 10-20 t buy from IVITA or LP. Fingerlings from the natural environ

ment are of lower quality and contain other species, representing 0.7 % of the total sown. The Regional Government supplies those that produce <2 t. In Ucayali, 63.5 % cultivated paco, 9 % gamitana, and in combined sowings (paco - gamitana) 19.6 %, these two species represent 92 % of aquaculture production (Table 11).

In addition, in Ucayali, the prevention of fish diseases is not common. Only 20 % of the CA use some prevention method, being the most common method the treatment with salt, and more practiced among those who grow paco. The species they take care of the most is the paiche.

Table 12 shows the months of cultivation up to its commercial size, which depends on the type of food (the value in parentheses). The paco cultivated with commercial feed (balanced pellet feed) alone or combined with natural feed is 5.7 and 5.6 months old, respectively. When only natural food is given, the time is 12 months.

Table 13 shows the amount of water mirrors used in each cultivated species. Between paco and gamitana, 85.6 % of the total water mirror is dedicated to it, as well as 92 % of the fingerlings planted in this region, that is, they are planted with greater density compared to other species. It should be noted that when two species are cultivated in the same pond, the density increases by 50 % (15.9 thousand/ha), which makes it reasonable for their densities to be added.

Discussion

In Ucayali, 73 % of CA are of the AREL type, 27 % AMYPE, information similar to the cadastre PRODUCE²¹. It should be noted that 8 % of the ARELs have a production greater than that which corresponds to them (3.5 t), showing improvements in their production (Table 2).

The CA are dependent on the rains (61.4 %), this being their main water source, limiting the control of their supply and making them vulnerable to climate change²². They pointed out that over the years the rain cycle in the Amazon has presented alterations due to climate change causing droughts making them vulnerable for their development.

Likewise, the CA, AMYPE, with low production (representing 12.5 %) also depends on the rain. This form of aquaculture is far from the typical forms that exist^23,24. These conditions limit its expansion to increase aquaculture production in Ucayali. The other supply options (water hole, stream, well) are characterized by the low flow. The water of rivers and lagoons, due to its high contamination, are not suitable according to the aquaculturists. Similarly, in 2014 in the Ucayali river samples were recorded with the presence of coliforms, suspended solids, and lead, being a non-optimal water source for the development of aquaculture activity as indicated in the Basic Manual of Fish Health^20,25.

77.3 % of CA collect water using gravity since the water falls directly into the pond due to rain or because the point of water intake from the well, stream, river, and lagoon is at a level above the ponds. There are cases in which they are supplied by rain, but these have a pond to store water and need to pump it and transfer it through pipes to the culture ponds, for this reason, they are transferred through plastic pipes, which makes filling or water replacement slow. of the pond. Another characteristic is that all the ponds are earthen, they lack the typical structures for this type of activity²⁶. In times of intense rain, the water rises, allowing it to be replaced.

Pond water quality testing is a rare practice²⁷. The CA that produce <2 t in the AREL category (30.7 %) do not perform any water analysis, in the AMYPE 66 % do analyze the water. Those that produce 2-3.5 t among the AREL do not analyze the water is twice as many as do. Overall, 55.6 % of CA do not test their water. Little practice is common in those that produce <2 t (31.8 %). The importance of water quality guarantees us an optimal production of fish in earthen ponds, for this reason, one of the important parameters is the obtaining of dissolved oxygen, which is recommended for the growth and survival of this species²⁸, however, they are not suitable for the conditions mentioned for the native species of the Peruvian Amazon, which can develop optimal growth with 5.8±0.7 mg L^-1 of dissolved oxygen²⁹. Similarly, there is a study mentioned by the IIAP indicating that paco can resist low concentrations of dissolved oxygen 1-3 mg L^-1 without fish mortality³⁰.

Aquaculture production is atomized in small CA with low production levels (Table 5), likewise, the Piscicultura technical guide, it mentions that one of the factors to obtain good results in production depends on the size of the pond³¹, however, the 10 % of the CA surveyed have an average water mirror of 600 m² for each pond. 54.6 % of the CA of Ucayali that have between 1 and 2 culture ponds have a production level of less than 2 t/campaign, and these have 1000 m² to 1750 m² respectively, a stocking density of 1.5 fingerlings/m² and 0.7 fingerlings/m², this would be one of the factors for the correct management of production since the recommended area in its ponds for this activity is 1000 to 5000 m² and the depth of 1.20 m in the supply and 1.50 m in the drainage point because at greater depth it makes it difficult harvesting and depth less than 70 cm allows overheating of the water with the risk of fish death³².

In addition, aquaculture production will also depend on the number of inoperative ponds found in 22.7 % of the CA surveyed due to lack of fingerlings (45 %) and lack of financing (22.5 %) as main causes. These factors limit the growth of aquaculture activity in Ucayali in the short term.

According to PRODUCE³³, the water mirror varies between 0.0072 to 0.4396 ha with a total of 76.45 ha. Likewise, in the survey carried out, different values were found, having a total sum of 58.74 ha. The CA with the largest water mirror have 3.6 ha, a correlation was observed that the greater the production, the larger the water mirror, except at the level of 10 to 20 t since it has an average of 1.16 ha, which could be explained because the ponds They are smaller than the general average. According to DIREPRO³³, the average water mirror was 0.12 ha, while in CA they have an average of 0.67 ha.

There is a decreasing amount of purchase of fingerlings no greater than 3.5 thousand for 9 ponds presenting a low investment and this would give low productivity according to some authors. Thus, for species such as paco and gamitana³⁴ they use a density of 1 fish/m^2/35 they sow 1.5 to 2.5 fingerlings/m² of water mirror³⁶ they cultured at 2.6 fish/m², and Deza et al.³⁷ cultivated paco at 0.5 to 1.5 fingerlings/m² indicating that the higher the density, the more profitable the crop. According to these reports, the average density value can be taken at 2 fish/m² or 20 thousand fingerlings/ha. But we report that the CA of Ucayali sows 9.5 thousand/ha, evidencing the low productivity in Ucayali. If 100 % of all existing ponds were used, aquaculture production would increase by 30 %. With proper techniques³⁸ production could easily be doubled. If all the ponds were used and with adequate techniques, the production would be multiplied by 3.5 times the current production.

The data from the water mirror and fingerlings have a normal distribution (the Shapiro-Wilk test was applied). In the Pearson test, a high correlation (0.853) was observed between the number of ponds with the investment and the hectares of the water mirror (Figure 1), its linear regression is included with an R² of 0.7281. These results show that not all the ponds have the same dimension and that the greater number of ponds increases the water mirror in greater proportion because they are larger ponds. There is a moderate correlation (0.666) between the hectares of water mirror and the number of fingerlings stocked, which shows that there is no single criterion to establish stocking density in culture ponds, which is not based on technical criteria.

The "water mirror - investment" correlation was 0.504, showing little criteria to invest, especially when there is no technical management to establish the density of cultivation. The correlation (Spearman's Rho) “investment - stocked fingerlings” was 0.442, a low value that justifies the diversity in the productivity of the CA. The correlation "number of ponds - fingerlings stocked" was 0.624, this low value is due to the fact that CA with more than 8 pounds have low crop loads, demonstrating limited technical knowledge. The correlation "investment - number of ponds" was 0.782 (with Spearman's Rho), a relatively high value, showing that investments in relation to available ponds are managed with better criteria in CA. The use of lime for ponds is a common practice in Amazonian aquaculture because the soil has a low pH and eliminates contaminating microorganisms³⁹. The use of lime by cultivated species is different. The CA that grow paco and paiche use the quantity recommended by the authors (2667 kg/ha). In the case of the boquichico crop, since it is a species of lower value, almost no lime is used, only 71 kg/ha) (Table 8). 93 % of the AC surveyed in Ucayali use lime, however, the amount used is not recommended, so in the production range of 3.5 to 10 t, they are the ones that use the most lime per hectare (545.2 kg) when expected, according to Boyd⁴⁰ it was 3000 at 5000 kg/ha of hydrated lime, although Deza et al.³⁷ recommend 1500 kg/ha. That is to say, the practice of using lime is not very effective since they only use it in Ucayali from 185.3 to 545.2 kg/ha.

Primary production is a complementary food for farmed fish and contributes to the supply of dissolved oxygen, so it is common to add fertilizer to promote photosynthetic flowering⁴¹. 58 % of the CA use fertilizer, 43 % use the organic type (43 %), especially those that produce <2 t, use an average of 8588 kg/ha of water mirror. Those that produce <2 t use an average of 7698.5 kg/ha, this is 5 times what is recommended by Deza et al.³⁷, 1500 kg/ha. The CA that produce 3.5-10 t use a quantity of 1398.2 kg/ha of water mirror, a value close to what is recommended.

Regarding the purchase of fingerlings, the CA buy mainly from the IIAP because it is cheaper than the LPs, which are important producers of fingerlings (Table 10), as a whole they supply 87.6 % of the CA or 95.3 % of stocked fingerlings, however, this does not satisfy the demand of the aquaculturists, because it is one of the causes of inoperability of the ponds (45 %). Fingerlings from the natural environment are of lower quality and contain other species, which is why they represent 0.7 % of the total sown. The Regional Government supplies those that produce <2 t. In Ucayali, paco and gamitana farming represents 92 % of aquaculture production (Table 11). It should be noted that in 2019, 55.8 % paco-gamitana and 44.1 % paiche were produced. In 2020, due to the pandemic, the production of paiche fell, highlighting paco and gamitana. In Ucayali, paco³⁸ is traditionally grown.

In Ucayali it is not common to take measures to prevent diseases in fish, since they are considered resistant³⁸. Only 20 % of the CA use some prevention method, the most common is salt treatment and it is more practiced among those who grow paco. The species that cares the most about its health is the paiche because of its higher commercial value.

The most used type of food is balanced food. There are cases in which natural food consisting of vegetables and fruits from the area is used for several months of fattening, which are well accepted³⁸. Therefore, in Table 12 are the months of cultivation up to its commercial size, which depends on the type of food (the value in parentheses). The paco cultivated with commercial feed (balanced pellet feed) alone or combined with natural feed is 5.7 and 5.6 months old, respectively. When they only give natural food, the time is 12 months. Small producers are the ones that use more natural food, but their cultivation periods are longer. In the case of paiche, they use more balanced feed because they have a higher commercial value.

Table 13 the amount of water mirror used in each cultivated species. Between paco and gamitana, 85.6 % of the total water mirror is dedicated to it, as well as 92 % of the fingerlings planted in this region, that is, they are planted with greater density compared to other species. It should be noted that when two species are cultivated in the same pond, the density increases by 50 % (15.9 thousand/ha), which makes it reasonable for their densities to be added.

In conclusion, the main source of water is rain for 67 % of the ARELs and 46 % of the AMYPEs, aquaculture production is fragmented due to the low capacity of the CA, 1 out of 4 do not using at least one pond due to the lack of fingerlings (45 %) and financing (22.5 %), 93 % apply lime in their ponds, although in very low quantities (185.3 to 545.2 kg/ha), questioning its effectiveness. In the paiche and paco crops, they use adequate amounts of lime, the most cultivated species are paco and gamitana (92.04 %) due to COVID-19 that limited the cultivation of paiche, the main producers of fingerlings are the IIAP and the LP (71.6 %), combine balanced feed with natural feed to lower production costs.