Introduction

The white-legged shrimp (Penaeus vannamei) is the main species produced in the world by aquaculture with a proportion of 52.9 % (≈4966.2 thousand tons) as compared to other crustacean species in 2018. Shrimp is an important source of foreign exchange income and has historically been among the most traded products; most of the production comes from Asia and Latin America, while one of the main markets is the United States of America (^{FAO, 2020}). Several studies have been conducted on shrimp culture, addressing different topics to determine the effect of environmental and management variables on the growth and survival of shrimp (^{McGraw et al., 2001}; ^{Moss & Moss, 2004}; ^{Islam et al., 2005}; ^{Milstein et al., 2005}; ^{Mena-Herrera et al., 2006}; ^{Araneda et al., 2008}; ^{Ponce-Palafox, et al., 2019}; ^{Jaffer et al., 2020}; ^{Rahmawati et.al., 2021}). Despite several scientific contributions to enhancing shrimp farming, only a few investigations have been conducted in commercial operations under different farming conditions: a) intensive affected by diseases such as white spot virus (^{Ruiz-Velazco et al., 2010a}; ^{González-Romero et al., 2014}) and disease-free (Ruiz-Velazco et al., 2010b; ^{Estrada-Pérez et al., 2016}); b) semi-intensive affected by the acute hepatopancreatic necrosis disease (AHPND) (^{Estrada-Pérez et al., 2020}) and disease-free (^{Estrada-Pérez et al., 2015}; ^{González-Romero et al., 2018}). In these studies, researchers found several relationships between environmental and management variables that may affect or benefit intensive and semi-intensive shrimp cultivation. The present study assessed the different relationships between environmental and management variables on growth, survival, and how they affected the production in some semi-intensive commercial Penaeus vannamei shrimp farms. This type of technology represents about 97 % of the currently operating farms in Mexico. The present study also addressed the relationships between the different variables involved and aimed to contribute to a better understanding of shrimp farming to optimize management practices which lead to better production.

Material and Methods

Data survey

Data were collected from a semi-intensive commercial shrimp farm located in the municipality of San Blas, Nayarit State, Mexico. The study units were eleven earth ponds ranging from 0.6 to 7 ha, that were in use during spring stocking. For each pond, the following dependent variables were analyzed: shrimp growth (final weight, g), final survival (%), and final biomass yield (kg ha^-1). The independent variables used were: water temperature (°C), salinity (ups), dissolved oxygen (mg L^-1), initial stocking density (post larvae per square meter, pL m^-2), pond size (ha), cultivation duration (days), and amount of feed used in kg ha^-1 (Table 1). According to the producers, the feeding strategy was with commercial feed on demand, using as an indicator the consumption baskets in the ponds. The protein percentages used were 40% during the first month and 35 % thereafter until harvest.

Oxygen and temperature were measured twice daily (06:00 and 18:00 h) using oximeters (Model 55, YSI, Yellow Springs, OH). Salinity was measured weekly using refractometers (Aquafauna Bio-Marine, Hawthorne, CA) at a precision of 1 ups. Shrimp weight was measured weekly using 0.01 g and 0.1 g precision balances (Ohaus Corp., Pine Brook, NJ). Survival was estimated from shrimp population samples caught with cast nets (1.5 m radius), depending on the shrimp size monofilament line cast nets had a mesh size of 3.2 or 25.4 mm.

Results

Simple linear regression models

There were presented the simple linear regression models that had significant differences (p < 0.05) in the correlation analysis between variables (environmental and management) and the production variables used in the simulation (Figure 1). The final weight was positively related to temperature (Figure 1a) and dissolved oxygen (Figure 1b) but negatively related to stocking density (Figure 1c), amount of feed used (Figure 1e), and duration of cultivation (Figure 1f). The final survival was negatively related to pond size (Figure 1d).

Figure 1 Simple linear regression models between environmental and management variables related to production parameters: final weight-temperature (a), dissolved oxygen (b), stocking density (c), amount of feed (e), duration of cultivation (f) and final survival-pond size (d). 

Discussion

According to the results of variance analysis, environmental variables such as temperature and dissolved oxygen showed the greatest variability between ponds and therefore, these variables had an effect on production. The exception among the environmental variables analyzed was salinity, as there were no statistically significant differences between the ponds, since it remained relatively constant, without an effect on the production parameters.

The literature has shown that growth in P. vannamei increases with increasing temperature up to 32 °C. (^{Ponce-Palafox et al., 2019}). ^{Ruiz-Velazco et al. (2010a}, b), ^{Estrada-Pérez et al. (2015}) and Estrada-Perez et al. (2020) found that at higher temperatures, the final weights of shrimp were better in intensive and semi-intensive farming so that the results of this investigation are consistent with the findings by these authors.

Regarding dissolved oxygen, fluctuations throughout the day were evident, especially in ponds that had a high stocking density. During the early morning, due to biological oxygen demand critical oxygen levels may lead to mortalities (^{Hepher & Pruginin, 1981}; ^{Carranza, 2020}).

In the present study it should be noted that the concentrations of dissolved oxygen in some ponds were lower than those recommended by ^{McGraw et al. (2001}) for shrimp farming, yet were above the range considered (approximately 3 ups) as stress inducing (^{Brock & Main, 1994}). Despite this, in the present study no mass mortalities were observed in culture. However, the final weight showed that higher concentrations of dissolved oxygen lead to greater weight achieved at the end of the culture, coinciding with ^{Rahmawati et al. (2021}) in indoor raceway ponds. Therefore, it was observed that oxygen dissolved variations had a negative effect on shrimp growth which lead to a suboptimal production in low concentrations. ^{Ruiz-Velazco et al. (2010a}, b) and ^{Estrada-Pérez et al. (2020}) also found this relationship in intensive and semi-intensive commercial shrimp farming.

In this study, salinity concentrations higher than those recommended for shrimp culture were recorded (^{Bray et al., 1994}). However, salinities lower than seawater (15-30 ups) have been associated with better growth in P. vannamei culture by several studies (^{Hernandez-Llamas & Villareal-Colmenares, 1999}; ^{Ruiz-Velazco et al., 2010a}, ^b; ^{Estrada-Pérez, 2020}). High salinity concentrations in the present study made it impossible to determine whether they actually had an effect on the growth or survival of the shrimp. Although the literature is not conclusive with respect to optimum salinity for P. vannamei since the white shrimp is a euryhaline species and, therefore, is resistant to wide salt concentration fluctuations (0.5 to 50 ups) (^{Bray et al., 1994}; ^{Saoud et al., 2003}).

It has been reported a negative effect of the use of high stocking densities on the development of shrimps P. vannamei (^{Otoshi et al., 2007}; ^{Araneda et al., 2008}; ^{Schveitzer et al., 2013};). The results of the present study also showed a negative effect on final weight when high stocking densities were used, which led to poor production. Contrarily, ^{Casillas & Ibarra (1996}) found that the highest production in their study was obtained at high stocking densities. However, in this study, over 45% of the ponds used had stocking densities considered high for semi-intensive culture (stocking densities higher than normal for semi-intensive culture), but there were no aeration systems. This limitation probably affected shrimp performance, according to ^{Musa (2013}) and ^{Boyd (2017}), as dissolved oxygen played an important role in improving water quality. Also, ^{Mena-Herrera et al. (2006}) mention that the high oxygen consumption in high biomass shrimp ponds suggests that the pond's load capacity depends on water quality, in particular on dissolved oxygen fluctuations. In addition, ^{Martinez (1993}) suggested that the optimum stocking density for a semi-intensive production system of P. vannamei is 22 shrimp m^-2 for the spring-summer cycle. Therefore, one of the most important challenges to overcome when working with high densities is the limitation represented by oxygen availability (^{Krummenauer et al., 2011}), and the significant impacts that different aeration devices can have on shrimp growth in different culture systems (^{Lara et al., 2017}).

In intensive shrimp farming, density-dependent growth and survival is a typical response mainly due to a combination of factors, including reduced availability of natural food sources and space, increased cannibalism, decreased water quality, and accumulation of undesirable sediments (^{Arnold et al., 2006}). ^{Arambul-Muñoz et al. (2019}) mention that total ammonia nitrogen (TAN), nitrate, and total phosphorus water quality were significantly higher at a higher density, furthermore they observed that growth rates and survival decreased as density increased after 300 org m^-3 of an intensive photoheterotrophic white shrimp (Penaeus vannamei) system. ^{Krummenauer et al. (2011}) observed in a super-intensive culture of Penaeus vannamei in a Biofloc technology (BFT) system that, stocking density was pivotal to shrimp survival, verifying along with other studies that P. vannamei productivity is limited by reduced growth as stocking density increases (^{Moss & Moss, 2004}; ^{Coman et al., 2004}). An efficient aeration system let to increasing stocking density (^{Samocha et al., 2004}) agreeing with ^{Da Silveira et al. (2020}) who consider stocking densities of 500 shrimp m^-2 viable for BFT production systems; however, the author mentions that, this implies careful planning in terms of Dissolved Oxygen consumption, and that the application of new technologies may potentially lead to even higher densities in the future.

According to the analysis of correlation and simple linear regression, shrimp growth was affected by high stocking densities and low dissolved oxygen concentrations (Table 3), which suggested that the load capacity was exceeded. It is encouraged the use of aerators for this type of culture as reported in previous reports (^{McGraw et al., 2001}; ^{Da Silveira et al., 2020}).

Reported attempts to increase production have concentrated to increase harvest size, for which two options are available according to ^{Casillas & Villarreal (1995}): 1) An increased culture period and 2) a reduction of the stocking density. In the present study, the weight gain of the shrimp was close to 1 g per week, similar to what is considered appropriate for commercial farms. However, shrimp held for a longer period of time in ponds stocked at high densities were affected in final weight and therefore their growth was suboptimal (^{Otoshi et al., 2007}; ^{Araneda et al., 2008}; ^{Schveitzer et al., 2013}). Another factor in consideration was the apparent negative effect on final weight related to the amount of feed used, this could have been a consequence of inadequate feed management leading to suboptimal production and also problems related to water quality ( ^{Tacón et al., 2013}). However, this could be also linked to high stocking densities in small-sized ponds according to the correlation analysis.

On the other hand, there are several indications related to the negative correlation between dissolved oxygen and the duration of culture. The most convincing was the high dissolved oxygen demand by the shrimp during growth. ^{Rahmawati et al. (2021}) found a decrease in dissolved oxygen levels as days of culture increased, this was due to increased biomass of P. vannamei, indicating an increase in oxygen consumption rates (^{Kureshy & Davis, 2000}). Another indication was the accumulation of organic matter (uneaten food and feces) as the culture progressed. Uneaten feed absorbs water and eventually settles at the pond bottom, decomposing later, adding excess nutrients to the water, and causing oxygen depletion (^{Iber & Kasan, 2021}). Therefore, the biological oxygen demand tended to increase, making necessary supplemental oxygen through aeration.

In the present study, it was found a negative correlation between the size of the pond and survival. This may be explained by the use of a small pond size, which made its management more convenient for survival. The water flow produced by water exchanges resulted in more efficient and faster control if any complications aroused. In small ponds, it is easier to monitor water quality, shrimp populations, and feeding than in large ponds. The action of wind in small ponds is more effective in mixing, or solids from the bottom of the water (^{Brune & Drapcho, 1991}).

In the same sense, it has been emphasized the use of production units with small pond sizes for best management practices (^{Hernandez-Llamas & Villareal-Colmenares, 1999}; ^{Milstein et al., 2005}; ^{Magallón, 2006}). Several studies using multiple linear regression models support that in small-sized units lower mortality rates are obtained (^{Ruiz-Velazco et al., 2010b}; ^{Estrada-Pérez et al., 2020}). Similarly, a negative correlation between large-sized ponds and survival was observed. The smaller ponds stocked with the highest stocking density presented better survival (Table 2) which indicated the benefit of small ponds monitoring. It is expected, that until a certain limit in which a culture lasts, the final weight of the shrimp increases, this situation turned out as expected, the longer the culture duration, the greater the final weight is obtained (^{Ruiz-Velazco et al., 2010b}; ^{Ruiz-Velazco et al., 2013}; ^{Estrada-Pérez et al., 2015}; ^{Estrada-Pérez et al., 2020}). This is apparently contrary to what was found in this research; however, it is more likely that this is due to an effect of stocking density and dissolved oxygen in the water, since the ponds that lasted longer were the ponds stocked at higher densities, and these same ponds that lasted longer had lower dissolved oxygen values in the water (Table 2). In other words, the final shrimp weights reached better sizes in shorter cultures, because they were stocked at lower stocking densities and had higher dissolved oxygen values, which resulted in faster harvesting.

Simulation analysis with simple linear regression models showed that the variation in dissolved oxygen was the most influential variable as shown by the production results (biomass), followed by temperature, stocking density, and finally pond size (Table 4).

The present study confirmed that the fluctuations of dissolved oxygen affected the production. Additionally, as mentioned above, stocking densities higher than those recommended for semi-intensive shrimp culture could have an effect on the dissolved oxygen concentration. The stocking density combined with other variables such as temperature, could affect the growth of organisms and hence yield production.

Further studies are needed to determine if an increase in the concentration of dissolved oxygen, particularly for this kind of farming (small pond size and appropriate stocking densities) could enhance the growth and survival of cultured shrimp and therefore improve farmed shrimp.

Conclusions

Given that dissolved oxygen was the most sensitive variable, supplemental aeration is essential at these stocking densities. A combination of high oxygen values (controlled with supplemental aeration), lower stocking densities, and short culture durations in small ponds, could increase production in this type of system.