*Article* **Pull the Emotional Trigger or the Rational String? A Multi-Group Analysis of Organic Food Consumption**

**Qiuqin Zheng 1, Haimei Zeng 2, Xintian Xiu <sup>3</sup> and Qiuhua Chen 1,\***


**Abstract:** The organic food industry in China has been developing fast with the increasing consumer demand for healthier, safer, and more nutritious foods since the epidemic outbreak. It is of great significance to understand the psychological preference of consumers for organic food and adjust the marketing strategy accordingly. In this study, we adopted the multi-group structural equation model (SEM) to analyze 571 questionnaire data and explored the effects of consumers' perception on the sensory appeal of organic food, perception on promotional stimulation, positive emotion, and perceived social value on the purchase intention of organic food. Based on the Stimulus–Organism–Response (S-O-R) model, this study divides the route affecting organic consumption behavior into the rational route and emotional route. It was proved that the emotional route (positive emotion) has a greater impact on the purchase intention of organic food than the rational route (perceived social value). In addition, there are different purchase intentions among different product types. Specifically, compared with organic tea, positive emotion has a greater effect on the purchase intention for organic rice. This study provides an important reference for the organic food-marketing strategy of enterprises.

**Keywords:** organic food consumption; positive emotion; sensory appeal; multi-group SEM

### **1. Introduction**

In the context of green and sustainable consumption, China and other emerging countries have begun to heavily promote organic consumption in recent years. In 2020, the global sales of organic food and beverage exceeded 120 billion Euros, and China accounted for 10.2 billion Euros (8.5%) [1]. China is the fourth largest organic food market in the world [1]. In 2022, China Central Document No. 1 stated that governments should continue to adjust and optimize the agricultural structure; strengthen the certification and management of green food, organic agricultural products and geographical indications of agricultural products; and increase the supply of high-quality green agricultural products. Compared with traditional food, organic food follows the production standards of organic agriculture, without using chemically synthesized fertilizers, pesticides, growth regulators, and other substances [2,3]. It contains no pesticide residues and does not use growth hormone and genetic engineering (GE) in the growing process, which is more healthy, nutritious, and natural [4]. Organic food plays an important role in promoting environmental protection and agricultural efficiency, enhancing the competitiveness of agricultural products and meeting the demand for safe and high-quality agricultural products. Consumers have linked organic food with health and nutrition since the outbreak of COVID-19. In the next few years, the organic food industry in China has great potential for further expansion [5,6].

Chinese governments have carried out a lot of publicity and education on organic food consumption. With the heavy promotion from governments, consumers are no longer

**Citation:** Zheng, Q.; Zeng, H.; Xiu, X.; Chen, Q. Pull the Emotional Trigger or the Rational String? A Multi-Group Analysis of Organic Food Consumption. *Foods* **2022**, *11*, 1375. https://doi.org/10.3390/ foods11101375

Academic Editors: Riccardo Testa, Giuseppina Migliore, Giorgio Schifani and József Tóth

Received: 5 April 2022 Accepted: 9 May 2022 Published: 10 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

unfamiliar with the concept of "organic." Yang et al. (2021) [7] found that the subjective cognition level of Chinese consumers on organic food is in a high position. However, the consumption capacity of organic food in China still lags behind other countries [1]. This indicates that only improving consumers' cognition is not enough. Ignoring the emotional appeal or emotional resonance of individuals cannot effectively promote the real organic consumption behavior of consumers [8]. Food consumption is never just to satisfy the appetite, which is closely related to culture and emotion [9–11]. Emotion is the attitude experience of individuals on whether objective things meet their needs. When objective things meet their needs, individuals show a positive attitude and this usually reflects in feelings of love, joy, happiness, etc. [12]. According to the broaden-and-build theory of positive emotions proposed by Fredrickson (1988) [13], various positive emotions such as happiness, interest, satisfaction, pride, and love, are more helpful to expand the scope of attention, cognition, and behavior of individuals; on that basis, people can more effectively obtain and analyze information, make more appropriate action choices, and have the effect of continuously enhancing personal resources for a long time. Moreover, Eastern culture is more emotional compared with the rationality of Western culture [8,14]. Therefore, it is of great significance to explore the role of emotion in organic consumption in China. In addition, the positive emotions of consumers can promote their cognitive flexibility and expand their scope of attention, making consumers aware of the differences between similar products [15,16]. This is also of great significance for brand competition.

Emotion is the most important endogenous factor in individual psychology and plays a very important role in decision-making [11,17]. More and more consumers tend to choose foods that have an emotional resonance with them. Holbrook and Hirschman (1982) [18] were the first to apply emotion to the research field of consumer behavior. They proposed that consumers were not always rational and emphasized the importance of emotion in purchase behavior. For example, Hsu and Tsou (2011) [19] employed the Stimulus–Organism–Response (S-O-R) model, starting from the impact of website quality on the repurchase intention of consumers, to measure consumers' emotional state through the Mood Scale, and selected it as the intermediary variable. They found that website quality can bring about individual feedback with positive emotions, while the positive emotion is helpful for consumers to repurchase and improve their repeat purchase intention. In marketing practice, especially in advertising design, marketers are increasingly using various emotional experiences to influence consumer behavior and decision-making. Emotion involves the whole process of consumer behavior, i.e., all the behaviors from searching and processing information, to product selection, and then to post-purchase, are closely related to the emotional state of consumers [20–22]. Therefore, it is crucial to incorporate emotional factors into the cognitive process of purchase behavior of organic consumers. The analysis of Dispoto (1977) [23] showed that the correlation coefficient between ecological emotion and ecological behavior was 0.15, and many people with little environmental knowledge still show strong emotional loyalty to the environment. This reveals that environmental knowledge and environmental emotion are two independent influencing variables. Kanchanapibul et al. (2014) [24] also considered that it is reasonable to take environmental emotion and environmental knowledge as two different variables to study their impacts on environmental consciousness and behavior.

However, only a few studies are focusing on the impact of emotion on ecological products, low-carbon consumption behavior, and organic products to date. Meneses et al. (2010) [25] analyzed emotional variables and cognitive variables and found that the recycling behavior of consumers was more related to emotional factors than cognitive factors. Based on the study of low-carbon purchase behaviors, Wang and Jing (2012) [26] found that the impact of low-carbon emotion is greater than that of low-carbon knowledge, suggesting that stimulating emotion was more effective than improving cognition in the influence of consumers' attitude towards low-carbon environmental protection. Lee and Yun (2015) [27] used the S-O-R model to study the two routes of organic consumption, i.e., the emotional route and rational route, and found that consumers are more inclined to

cognitive judgment than emotion in purchasing organic food. Jose and Kuriakose (2021) [28] took Indian consumers as the object and compared the effects of emotion, practice, and rationality on organic consumption intention. They found that rational factors such as environmental motivation have little influence, and emotion plays a leading role in organic consumption.

Therefore, to continuously and effectively expand the organic consumption market in China, it is necessary to understand the driving mechanism of the organic food purchase of consumers in China. Based on the S-O-R model and the broaden-and-build theory of positive emotions, this study adopts a multi-group structural equation model (SEM) to explore: (1) Is it the rational route or emotional route that can better enhance the purchase intention of Chinese consumers for organic food? (2) Are there significant differences in the impact routes of different product types (such as organic tea and organic rice)?

### **2. Theoretical Framework and Research Hypothesis**

### *2.1. Theoretical Framework*

A large number of studies on consumer purchase behavior took the S-O-R model as the theoretical basis since it was proposed by Russell (1974) [29]. The S-O-R model considers that the purchase behavior of consumers is mainly caused by external stimuli (products, situations, etc.), which change the psychological activity of individuals, thereby generating motivation, making purchase decisions, and implementing purchase behavior. Therefore, the S-O-R model can be regarded as a dynamic expression of the purchase behavior process of individuals. Many scholars considered that in this dynamic process, the internal change of the organism is due to the cognition and emotion of individuals to the external stimuli, which will be reflected in the subsequent behavioral response.

One of the main hypotheses of the theory of reasoned action (TRA) and the theory of planned behavior (TPB) is that people are rational in the decision-making process and action; thus, cognitive methods can be used to predict behavior [30]. However, the addition of affective variables has been recommended as a useful extension of the theory [12,28,31]. Following this suggestion and using the S-O-R model, we define the behavioral response as the purchase intention of organic food and divide the internal changes of organisms into two-dimensional dimensions, i.e., emotion and rational (positive emotion and perceived social value), which is helpful to check whether it is the emotional route or the rational route that plays a role in organic consumption behavior in China (Figure 1). Therefore, the S-O-R model is suitable for this study.

**Figure 1.** Conceptual model.

Lee and Yun (2015) [27] proposed that it was more suitable to focus on the stimulation of food itself to explore the purchase intention of organic food than psychosocial stimulation, because they believed that the main determinant of the purchase intention of organic food is the product attributes related to health, environmental protection, and animal welfare. However, consumers in China still have little purchasing power for organic food, and they rely more on the publicity of the government and the outside world. Therefore, it is necessary to include both food stimuli (organic food characteristics) and external stimuli (promotional stimulation) in this study.

### *2.2. Research Hypothesis*

### 2.2.1. Sensory Appeal

According to the cue utilization theory [32], consumers usually evaluate products and make purchase decisions based on various internal and external information clues [33,34]. Internal information is inherent in the product itself, including taste and texture, while external information is other information related to the product, such as label, packaging, etc. [35] Acebron and Dopico (2000) [36] pointed out that both internal and external information about products can advance consumers to generate positive emotions and form judgments on product quality, which will further affect their purchase decision of consumers. These can be called sensory attributes [37]. In addition, products with different sensory attributes cause different emotional reactions in consumers [9]. Among these sensory attributes, vision is usually the first sense and overwhelms the perception of other information in attracting consumers' attention [38]. Visual cues are not only limited to the internal characteristics of the product itself but also involve external characteristics such as product packaging [39]. As the main physical characteristics of food, visual cues not only can indicate the quality of food but also link consumers to other emotional experiences [40,41]. Therefore, they are the attributes with more influence. Lee and Yun (2015) [27] found that consumers' perception of the sensory appeal of organic food can promote their purchase intention of consumers because organic food brings them positive emotions. Some studies also showed that the sensory attributes of organic food are usually related to pleasure, hedonism, and happiness [42]. Therefore, the following hypotheses are put forward:

**Hypothesis 1 (H1):** *Consumers' perception of the sensory appeal of organic food has a significantly positive impact on the positive emotion of consumers*.

**Hypothesis 2 (H2):** *Consumers' perception of the sensory appeal of organic food has a significantly positive impact on the perceived social value of consumers*.

### 2.2.2. Promotional Stimulation

Promotional stimulation is a form of communication used to raise consumers' awareness of the product and can distinguish them from their detractors. It can be used as a source of information to evaluate products and stores [43]. The premise of TRA is that people can completely control their behavioral intention. However, consumers are bound to be interfered with by external factors when making purchase decisions in real life [44]. When the real information about products is scarce or asymmetric, consumers will seek external help to obtain relevant purchase experience or suggestions, and then form the corresponding purchase intention [45]. Marketing promotion can be used as a source of information for evaluating products and stores [43]. The calls of environmental protection associations and governments and advertising factors can stimulate the organic purchase intention of consumers [46]. Chen and Antonelli (2020) [47] considered that external context factors can significantly improve the perceived value and purchase intention of consumers on products. Zhu et al. (2013) [48] found that situational factors of laws and policies can adjust the degree of influence of consumers' purchase intention on green purchase behavior. Miller et al. (2021) [49] found that situational effects have an impact on the perception and

cognitive links of consumers, which can interact with consumers' psychological perception and then affect their purchase intention. With strong calls from governments, environmental associations, etc., consumers can demonstrate their concern and responsibility for the environment by purchasing and using organic foods, gaining more social recognition and approval [50]. Therefore, the following hypotheses are put forward:

**Hypothesis 3 (H3):** *Consumers' perception of the promotional stimulation of organic food has a significantly positive impact on the positive emotion of consumers*.

**Hypothesis 4 (H4):** *Consumers' perception of the promotional stimulation of organic food has a significantly positive impact on the perceived social value of consumers*.

### 2.2.3. Positive Emotion

Positive emotion refers to the emotion with a positive valence. It is associated with the satisfaction of certain needs, accompanied by pleasant subjective experience, and can improve the enthusiasm and activity ability of individuals [51]. Gutjar et al. (2015) [10] indicated that the choice of food is mainly related to positive emotions. Meneses (2010) [25] also showed that the recycling behavior of consumers is more based on positive emotion than negative emotion. Based on his research, positive emotions include joy, contentment, interest, pride, gratitude, and love. According to the expansion theory of positive emotion proposed by Fredrickson (2001) [13], positive emotions can advance individuals to break through certain restrictions and produce more thoughts under general conditions, expand the scale of attention, enhance cognitive flexibility, and update and expand the cognitive map of individuals. Isen (2001) [52] found that positive emotions can provide more information for cognitive processing. Plenty of positive experiences with a product may lead to intuitive decisions on future purchases [53]. Existing studies have demonstrated that positive emotions can predict the ecological consumption intention of individuals [54] and also have a significantly positive impact on green purchase behavior and environmental protection behavior [25]. Therefore, the following hypothesis is put forward:

**Hypothesis 5 (H5):** *Consumers' positive emotion has a significantly positive impact on the purchase intention of consumers for organic food*.

#### 2.2.4. Perceived Social Value

Perceived value is the consumer's preference and evaluation of the product attributes and their utility that helps or hinders the achievement of goals in a given usage context [55]. In the era of consumerism, goods are purchased not only to satisfy individual functional needs but also to achieve the purpose of self-identity construction [56]. Sweeney and Soutar (2001) designed a scale for measuring the perceived value of durable goods, and they viewed perceived social value as the utility of a product to reinforce self-concept. They argued that when customers bought a product, they would consider the impression that the purchase would have on others [55]. Relevant studies from a social perspective have revealed that green consumption behavior stems from personal reputation and status, e.g., people are more willing to pay for environmental protection in public to gain extra points for their image [57,58]. Chinese consumers with collectivist cultural values face more than Western consumers with an individualistic culture [59]. That is, social motives influence Chinese consumers' green consumption behavior more profoundly than environmental and economic motives. Further, through organic food consumption, consumers can effectively present themselves to others [60]. This is because, compared to convention food, organic food is more expensive and pro-social, which can reflect the non-generic nature of consumers, thus helping them to gain more praise, social recognition, and good impressions [61]. Noppers et al. (2014) [62] stated that consumers brought green products because such consumption helped to project a positive image of themselves. Kohlova and Urban (2020) [63] pointed out that green consumption enhanced consumers' social status

because it helped them to demonstrate wealth-related competencies. People who need to confirm their social status or self-identity will prefer organic food [57]. Therefore, out of rational thinking, consumers will choose organic foods with greater utility in order to demonstrate their social status and value preferences. Therefore, the following hypothesis is put forward:

**Hypothesis 6 (H6):** *Consumers' perceived social value has a significantly positive impact on the purchase intention of consumers for organic food*.

### 2.2.5. Product Type

Products can be classified as hedonic products and practical products according to different classification standards [64]. Hedonic products can bring emotional joy, while practical products reflect rational cognition [65,66]. In general, practical products are mainly used by consumers to meet some specific tasks and obtain more efficiency, while hedonic products are mainly used to obtain emotional demands, and the consumption is to meet the subjective feeling of consumers [67]. Although these two kinds of products reflect different consumer psychology, they are not opposed to each other. Studies have shown that hedonic attributes and practical attributes have a positive correlation [68]. Some products have both the characteristics of hedonic products and practical products; to be specific, if a product is defined as a hedonic product, it means that the hedonic attributes of this product are greater than the practical attributes, rather than only having the characteristics of hedonic products without any characteristics of practical products [68]. In the situation of different product types, the influence route of purchase intentions of consumers is different [69]. Based on the functional consistency theory, compared with positive emotion, perceived social value emphasizes the practical value of organic food and can help consumers obtain more efficiency, which is matched with the attribute characteristics of practical products and is easy for consumers to generate a positive purchase intention [70]. Based on the selfconsistency theory, compared with perceived social value, positive emotion emphasizes the hedonic value, which is matched with the attribute characteristics of hedonic products and is easy for consumers to generate a positive purchase intention [71]. To easily compare the differences in purchase intention between different kinds of organic food, this study defines the tea as the hedonic product and rice as the practical product. Therefore, the following hypotheses are put forward:

**Hypothesis 7 (H7):** *In terms of tea, positive emotions have a greater impact on the purchase intention of organic food than perceived social value*.

**Hypothesis 8 (H8):** *In terms of rice, perceived social value has a greater impact on the purchase intention of organic food than positive emotion*.

### **3. Methods**

### *3.1. Questionnaire Design*

The questionnaire is divided into two parts. The first part is the main body of the questionnaire, including the scales of each variable, and the second part is the personal information of the respondents. Assuming that the measurement items of each variable in the model are from the maturity scale that has been widely used in the relevant literature, and have been appropriately modified based on expert opinions and the specific consumption situation of organic agricultural products. The Likert 7-point scale was used as the form of all scales.

As can be see in Table 1, the measurement of sensory appeal (SA) and positive emotion (PE) uses the scale of Lee and Yun (2015) [27]. The measurement of promotional stimulation (PS) uses a modified scale based on the design of Wang et al. (2018) [72], including the publicity of governments and academic institutions, and is suitable for the current mode of promoting organic agricultural products in China. The measurement of perceived social value (PSV) uses the scale of Wang et al. (2017) [8] and Sweeney and Soutar (2001) [55]. The measurement of the purchase intention of organic agricultural products (PI) uses a modified scale based on the design of Kim and Lee (2019) [73].

**Table 1.** The measures.


#### *3.2. Data Collection and the Sample*

The questionnaire method was used in this study, and the data were collected online based on the professional questionnaire platform (Credamo). In this survey, consumers who purchased organic tea (rice) were taken as the survey object. This is because only those consumers who purchased these products can perceive the relevant organic attributes. Meanwhile, trap questions were set in the questionnaire, i.e., "100 + 100 = ?". Those who answered wrong were regarded as not seriously filling in the questionnaire.

Before the survey, we conducted a small-scale pilot survey, and a total of 30 pilot survey questionnaires were distributed. Using SPSS24.0, we removed the measurement items with a Cronbach's α value of less than 0.6. According to the information and suggestions fed back by the pilot survey, the items with unclear semantics and confusion in the questionnaire were adjusted and revised, which can ensure the effectiveness of the questionnaire. In this study, two sets of questionnaires focusing on organic tea and organic rice, respectively, were designed. A total of 571 valid questionnaires were collected, including 290 questionnaires on organic rice and 281 questionnaires on organic tea.

### *3.3. Research Methods*

The data analysis in this study was divided into three steps. First, SPSS24.0 and AMOS24.0 were used to test the reliability and validity of variables to ensure the goodness of fit of the structural model. Second, AMOS24.0 was used to conduct a hypothesis test on the structural model to verify the relationship between sensory appeal, promotional stimulation, positive emotion, perceived social value, and purchase intention. Thirdly, the multi-group SEM was used to analyze the regulation of different types of organic foods. The multi-group SEM analysis can explore whether the route model suitable for one sample is also suitable for other samples. Existing studies have mainly focused on single organic food [74], and the studies for comparing and analyzing the differences between the purchase intentions of different types of organic foods are rare.

### **4. Results**

#### *4.1. Descriptive Statistics Analysis*

As shown in Table 2, the number of female samples (55%) is greater than male samples (45%), which is consistent with previous research results in which women are the main buyers of families in China [7,75]. The respondents aged 25~34 account for the largest proportion (62.5%), followed by those aged 35~44, accounting for 22.8%, which means that young consumers show more purchase intention for organic foods. This result is consistent with Chekima et al. (2017) [76] and Yadav and Pathak (2016) [77]. In addition, most respondents had a bachelor's degree. In terms of family population, families consisting of approximately three to four accounted for the largest proportion, followed by those consisting of approximately five to six people. The monthly income was at the level of 6501 RMB and above. Overall, the survey samples of this study are more in line with the actual situation of organic consumption in China and can be used for further analysis.


**Table 2.** Descriptive statistics of consumer social demographic characteristics.

Note: Chinese currency symbols, abbreviated as RenMiBi (RMB).

### *4.2. Reliability and Validity Test of Samples*

The composite reliability (CR) value was used to test the reliability of the questionnaire. From the measurement results of the model in Table 3, the CR values are greater than 0.7, suggesting that the indexes of each dimension have sufficient reliability and internal consistency [78]. The measurement of validity is tested by convergent validity and discriminant validity, in which the convergent validity is mainly reflected by normalized factor loading, Z-value, and average variance extracted (AVE). The results show that the normalized factor loadings are greater than 0.6 and significant, and the AVEs are greater than or close to 0.5, indicating that the scale has high convergence validity [79]. Meanwhile, the correlation coefficient between any two variables is less than the square root of AVE

of each variable, as shown in Table 4. Therefore, the scale has good discriminant validity, which lays a foundation for the analysis of the structural model.


**Table 3.** Results of measurement model analysis.

Note: \*\*\* *p* < 0.001. Composite reliability (CR), average variance extracted (AVE), sensory appeal (SA), promotional stimulation (PS), promotional emotion (PE), perceived social value (PSV), purchase intention (PI).

**Table 4.** Results of discriminant validity test.


Note: The items on the diagonal represent the square roots of the AVE; off-diagonal elements are the correlation estimates. Sensory appeal (SA), promotional stimulation (PS), promotional emotion (PE), perceived social value (PSV), purchase intention (PI).

#### *4.3. Test of the Measurement Model*

The measurement model is evaluated using the maximum likelihood method based on AMOS24.0. From Table 5, it can be seen that the overall test results of goodness of fit of the model are χ2/df = 2.505, GFI = 0.953, AGFI = 0.932, CFI = 0.966, and RMSEA = 0.051. These indexes of the model meet the standard, indicating that the model fits well.



Note: Goodness of fit index (GFI), root mean square error of approximation (RMSEA), standardized root mean square residual (SRMR), comparative fit index (CFI), Tucker–Lewis Index (TLI).

#### *4.4. Test of Structural Equation Model*

As shown in Table 6, all hypotheses passed the significance test. As expected, consumers' perceptions of sensory appeal and promotional stimulation had a significant

positive impact on the positive emotion of consumers (β = 0.647, *p* < 0.001 and β = 0.329, *p* < 0.001, respectively). The stronger the consumers' perception of the sensory appeal of organic food or the stronger the consumers' perception of external publicity, the higher the positive emotion of consumers. Then, H1 and H2 were supported. Meanwhile, consumers' perceptions of sensory appeal and promotional stimulation had a significantly positive impact on the perceived social value of consumers (β = 0.505, *p* < 0.001 and β = 0.329, *p* < 0.001, respectively). Then, H3 and H4 were supported. Further, consumers' positive emotions and perceived social value had a significantly positive impact on the organic purchase intention of consumers (β = 0.579, *p* < 0.001 and β = 0.242, *p* < 0.001, respectively). Then, H5 and H6 were supported.

**Table 6.** Results of the hypothesis test.


Note: \*\*\* *p* < 0.001. Sensory appeal (SA), promotional stimulation (PS), promotional emotion (PE), perceived social value (PSV), purchase intention (PI).

The Bootstrapping method is used to test the mediating effect. Hayes et al. (2009) suggested that the Bootstrapping method should be repeated at least 5000 times during the mediating effect test. In SPSS 24.0, we adopted the plug-in unit process 4.0 to set the sampling times to 5000 times and the confidence was 95% [80]. The results are shown in Table 7.


**Table 7.** Results of mediating effect test.

Note: Sensory appeal (SA), promotional stimulation (PS), promotional emotion (PE), perceived social value (PSV), purchase intention (PI).

The confidence interval (CI) of the indirect effect was used to judge whether the mediating effect exists. If the CI did not include 0, we rejected the original hypothesis, which means that the indirect effect was not 0 and the mediating effect existed [81]. As shown in Table 6, the indirect effect existed and was significant, indicating the existence of the mediating effect; the direct effect was less than the total effect and significant, indicating that there are partial mediating effects.

### *4.5. Multi-Group Analysis*

To check whether the route model suitable for the whole sample was also suitable for the specific sample group, and further to check whether different product types have the same influence route [82], this study selected different product types (organic tea and organic rice) as adjustment variables to conduct the multi-group SEM test on the S-O-R model of organic foods. The operation in this part follows the measurement invariance procedure of the composite model proposed by Byrne (2004) [83]. The results are shown in Table 8.


**Table 8.** Fit indices for multi-group invariance tests.

Note: Goodness of fit index (GFI), adjusted goodness of fit index (AGFI), comparative fit index (CFI), root mean square error of approximation (RMSEA).

From Table 8, the *p*-value of all competition models is less than 0.05, and the ΔCFI between any two models is less than 0.01, indicating that the multi-group measurement invariance is valid [83]. The research conclusion is applicable to all consumers. Then, the difference in route coefficient between organic tea and organic rice was in-depth investigated, and the results are shown in Table 9.

**Table 9.** MGA test results.


Note: \*\*\* *p* < 0.001; \* *p* < 0.1. Sensory appeal (SA), promotional stimulation (PS), promotional emotion (PE), perceived social value (PSV), purchase Intention (PI).

As can be seen in Table 9, in the four routes of H1–H4, the groups that purchase organic tea and organic rice are both significant at the level of 0.001, and there is no significant difference in route coefficient value and direction. In the route of H5, positive emotion has a significantly positive impact on purchase intention, i.e., β<sup>T</sup> = 0.382, *p* < 0.001 and β<sup>R</sup> = 0.714, *p* < 0.001, respectively. In the route of H6, perceived social value has a significantly positive impact on purchase intention, i.e., β<sup>T</sup> = 0.373, *p* < 0.001 and β<sup>R</sup> = 0.168, *p* < 0.05, respectively. This shows that the positive emotion of organic rice consumers has a greater impact than that of organic tea consumers on purchase intention; the positive emotion of organic rice consumers has a greater impact than perceived social value. In terms of organic tea, both the positive emotion and perceived social value have a positive impact on purchase intention, and there is no significant difference.

#### **5. Discussion**

Previous studies have shown that ignoring individual emotions cannot effectively promote the real organic consumption behavior of consumers [12]. The current research mainly focuses on the influencing factors of organic consumption behavior based on the theory of reasoned action (TRA), theory of planned behavior (TPB), motivation-abilityopportunity (MAO) theory, and value-belief-norm (VBN) theory [84]. It is necessary to add emotional variables to enhance the explanatory power of the existing research [12,28,31]. In this study, we adopt the S-O-R model and incorporate rational cognitive factors and emotional cognitive factors for research. By referring to the study of Fredrickson (1998) [13] and Arvola et al. (2008) [85], this study adopts positive emotion rather than negative emotion as the emotional factor. The reason is that food purchase is more based on positive emotions [10]. Our results also confirm this point, that is, in the situation of organic food consumption, positive emotion is a useful influencing factor.

Firstly, compared with the rational route (perceived social value), the emotional route (positive emotion) has a greater effect on the purchase intention of consumers for organic food. Moreover, the positive emotion plays a partial mediating role between the perception of the sensory appeal of organic food and purchase intention, and between the perception of promotional stimulation of organic food and purchase intention. Different from the research of Lee and Yun (2015) [27], this study supports the conclusion of Wang (2015) [12] and Jose (2021) [28]. Specifically, cognition usually features with transience, shallowness, situationally, and low involvement, while emotion is profound and highly involved [12]; if only cognitive education is conducted for consumers without arousing their emotional resonance, it is difficult to turn their cognition into practical behavior. This study believes that external stimuli (sensory appeal and promotional stimulation) can advance consumers to generate a sense of pleasure and satisfaction with their appropriate behavior. They will consciously purchase organic food to maintain and increase this happy emotional experience. The research of Rana and Paul (2017) [86] also showed that the demand of consumers for organic food in developed countries is mainly due to the requirement for meeting their high-level emotional needs, such as respect and self-realization.

Secondly, this study proved that consumers' perceptions of the sensory appeal of organic food can positively affect the positive emotion and perceived social value of consumers. The increase in the sensory appeal of organic food by one unit increases the positive emotion and perceived social value of consumers by 64.7% and 50.5% respectively. Meanwhile, consumers' perceptions of promotional stimulation of organic food can positively affect the positive emotion and perceived social value of consumers. The addition to the promotional stimulation of organic food by one unit increases the positive emotion and perceived social value of consumers by 32.9% and 32.9%, respectively. These are consistent with the previous research conclusions [27,28,87], that is, consumers' choice of organic food is based on complex judgment from perceived external information (such as packaging, price, publicity, etc.) [35]. Based on various internal and external clue information, consumers can form the value judgment or sensory expectation for organic food [36]. In addition, the promotional stimulation of organic food is useful since it can provide consumers with more organic knowledge to help consumers distinguish the positive attributes of organic food from traditional food. Compared with promotional stimulation, consumers' perception of the sensory appeal of organic food has a greater impact on the positive emotion and perceived social value of consumers [88]. According to the study of Lee and Yun (2015) [27], sensory appeal is usually linked to hedonic attitude and good experience, and the stronger the sensory appeal of organic food, the more pleasant experience it can bring to consumers.

Thirdly, unlike previous studies that focused more on product function and economic value to analyze consumer purchase behavior, this study also reveals the underlying mechanisms that influence organic food consumption from the perspective of perceived social value, further confirming the existence of social motives in organic consumption. In a collectivist culture, where people care more about the connection with people around them, Chinese consumers driven by a sense of face perceive organic food as having higher perceived value. This is consistent with the findings of [14,59]. Therefore, in terms of corporate marketing, the benefits of organic food consumption for others can be promoted so that consumers feel how others around them want them to behave, rather than simply emphasizing the functional or environmental value that organic products bring to consumers.

Finally, based on the multi-group analysis results, there are differences in the relationship between the organic product type and positive emotion and purchase intention. The existing studies are mainly the multi-group analysis of the impact of demographic characteristics or regional differences on consumer behavior [85,89], and studies on the regulation of product types are rare. In this study, we used the multi-group SEM to conduct the test. Although the overall model difference remains unchanged, from the specific impact path, the positive emotion has a greater impact than the perceived social value on purchase intention in terms of organic rice, which is a surprising discovery. This is due

to the fact that according to the previous research conclusions, compared with practical products, hedonic products have emotional and symbolic attributes, which match with positive emotions and can form a positive consumer response. Compared with rice, tea has more hedonic properties. One possible explanation for this is that for Chinese consumers, organic tea and organic rice are both hedonic products due to the high price of organic foods [5,6]. Moreover, the price of organic tea is much higher than the price of organic rice. From the perspective of availability, it is easier for consumers to convert their emotions on organic rice to purchase intention. This suggests that organic retailers should reduce the cost of organic food through various channels to reduce the price on one hand and improve the awareness of organic food of consumers to break the price barrier on the other hand.

This study still has the following deficiencies: First, this study only measures the purchase intention of consumers, and there is a big gap between intention and behavior. Future research can perform the measurement of actual purchase behavior. Second, although we choose the positive emotion as the intermediary according to the expansion theory of positive emotion and the characteristics of organic food, there have been some studies to explore the relationship between negative emotion and the purchase intention of consumers [90]. Therefore, negative emotion variables can be added for more systematic comparison in future studies. Third, based on the consumption and price of rice and tea in China, we treat tea as hedonic and rice as utilitarian. This is an empirical judgement that should be confirmed by adding pre-testing in future studies.

### **6. Conclusions**

The coronavirus pandemic has intensified consumer demand for healthier, safer, and more nutritious food. Organic food is considered healthier and safer than traditional food, with a self-owned "health halo" [91]. The health needs of consumers will further advance the development of the organic food industry. To successfully satisfy the growing market demand for organic food, marketers and decision-makers should understand the psychological preference of consumers for organic food and adjust marketing strategies accordingly to change their consumption decisions for organic food. This study classifies the routes affecting organic consumption behavior as a rational route and emotional route, and proves the influence of the emotional route (positive emotion) on organic food consumption behavior. Two different kinds of products, i.e., organic tea and organic rice, are taken to conduct the multi-group SEM analysis, and it is found that the product type has a certain impact. In addition, this study also figures out the antecedents of organic food consumption behavior, namely sensory appeal, promotional stimulation, positive emotion, and perceived social value. On that basis, the following suggestions are put forward:


should strengthen technological innovation, increase production, and reduce production costs, so that consumers can afford healthy and safe organic food.

**Author Contributions:** Conceptualization, Q.Z.; methodology, Q.Z.; software, Q.Z.; validation, Q.Z., H.Z., X.X. and Q.C.; formal analysis, Q.Z.; investigation, H.Z. and X.X.; data curation, X.X.; writing original draft preparation, Q.Z.; writing—review and editing, X.X. and Q.C.; funding acquisition, X.X. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Social Science Foundation of China, grant number 19BJY048.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We would like to thank the anonymous referees for commenting on this paper.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Gerg ˝o Gyalog 1,\*, Julieth Paola Cubillos Tovar 2,\* and Emese Békefi <sup>1</sup>**


**Abstract:** This paper provides a comparative overview of decadal changes in aquaculture production in the European Union (EU-27) and Latin America and the Caribbean (LAC). Contrary to other regions of the world, freshwater fish farming in these two territories is a marginal sub-segment of the aquaculture sector. Using an indicator-based approach, we track development tendencies in freshwater aquaculture, focusing on the main established and emerging species, diversification, and shifts in the mean trophic level of farmed animals. Geographical patterns in production trends are revealed in both regions. The study attempts to explain between-region and between-country differences in aquaculture growth by analyzing freshwater resource endowments at region-level and country-level, using total renewable water resources (TRWR) as an indicator of water-abundancy. Thermal optimum of main produced species is matched against climate conditions prevailing in main producer countries to provide further understanding of spatial heterogeneity in growth rates of aquaculture sector.

**Keywords:** aquaculture; renewable water resources; climate; trophic level

### **1. Introduction**

Since the mid-1990s, nearly all growth in seafood supply has originated from aquaculture. At the global level, the contribution of freshwater fish production to total aquaculture output increased from 55.6% to 61.2% between 1995 and 2019 [1], indicating that the growth rate of freshwater aquaculture outpaces that of mariculture. In the European Union (EU-27) and Latin America and the Caribbean (LAC), the profile of the aquaculture industry is different from the other regions, since coastal (marine or brackish water) aquaculture dominates the sector in both regions. In 2019, freshwater aquaculture only contributed 25.0% and 27.4% to total fish production in LAC and EU-27, respectively [1], and the rate of its growth was lower than that of marine aquaculture in both regions.

Nevertheless, freshwater aquaculture production experienced considerable growth in the latest decades in LAC [2]. In contrast, freshwater production in the EU has stagnated for decades, however, there is large heterogeneity between growth rates of member states. Opportunities for aquaculture growth are not the same in the two regions, as they differ from each other in terms of markets, regulation environment, and resource availability. Per capita fish consumption in the EU-27 is relatively high with a value of 24 kg/year, corresponding to a yearly consumed quantity of 12.3 million tons. With only a 41% selfsufficiency rate, the EU is the most important seafood importer in the world [3]. Moreover, in the category of freshwater fish, the self-sufficiency rate of the EU-27 is only 37% [4]. Conversely, LAC has the lowest per capita seafood consumption in the world with only 10.5 kg/year which is equivalent to a demand of 6.7 million tons, largely met by marine

**Citation:** Gyalog, G.; Cubillos Tovar, J.P.; Békefi, E. Freshwater Aquaculture Development in EU and Latin-America: Insight on Production Trends and Resource Endowments. *Sustainability* **2022**, *14*, 6443. https:// doi.org/10.3390/su14116443

Academic Editors: Giuseppina Migliore, Riccardo Testa, József Tóth and Giorgio Schifani

Received: 1 April 2022 Accepted: 23 May 2022 Published: 25 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

fisheries [5]. Latin American aquaculture is a net aquatic food exporter, and even though the majority of exports originated from the marine environment, tilapia, farmed in freshwater, is also marketed in large quantities to the USA [6]. However, domestic demand for aquatic food is increasing, as among all regions of the world the highest growth rate (+18% between 2016 and 2030) in per capita seafood consumption is projected for Latin America [5].

All in all, freshwater aquaculture has a marginal role in total fish production and aquatic food supply both in the EU and LAC, but domestic markets exist and are being developed for freshwater aquatic products, and for the latter, region export markets would also offer growth potential if competitiveness was further improved. This paper attempts to review the trends of freshwater aquaculture production under these circumstances. Although there are a variety of socio-economic and regulatory conditions in which the two regions differ from each other, it was not the intention of this study to explore these. Rather, using aggregate statistics we tracked the internal tendencies of the sector. As such, the paper presents, both at a regional and country-level, how the production volume changed over the last decade and investigates which species contributed to the growth. By using an index for diversification, we conclude whether freshwater aquaculture tends towards diversification or concentration. Between-country differences in production tendencies are revealed in both regions, and we attempt to explain these with differences in freshwater resource endowments and climatic conditions.

#### **2. Data Sources for the Analysis**

Data on aquaculture production (both quantity and value) was obtained from FAO FishstatJ [1]. The unit value of production was calculated by dividing production value by production quantity. Population information, which was used for calculating production growth per capita, was derived from the World Bank database [7]. Trophic levels (TL) in aquaculture were considered in our study. TL for each species was extracted from FishBase [8]. The TL of interspecific hybrids was assigned based on the TL of parental lines. Renewable freshwater estimates were obtained from the FAO Aquastat program website [9]. At the country-level, we used the indicator 'Total annual renewable water resources (TRWR) per inhabitant' to represent the water endowment of major aquaculture producer countries. In order to calculate the region-level (EU-27 and LAC) values for the availability of renewable water, first, we summed country-level data on 'Total internal annual renewable water resources (TIRWR)' (i.e., not counting external water resources) in order to avoid the problem of multiple accounting of resources shared by more than one country [10]. Second, the sum of the country-level TIRWR values was divided by the population of the region [7] to calculate the per-capita availability of renewable water resources in the EU-27 and LAC. We presented climate information for the analysis, which was extracted from the Climate Change Portal of the World Bank Group [11]. For our study, we utilized monthly mean temperature data recorded in the 1990–2020 reference period. The thermal optimum of cultured species was copied from the META (Maritime and Environmental Thresholds for Aquaculture) database [12].

To measure the diversification degree of the aquaculture sector, we implemented an index from Hofherr et al. (2012) [13]. This index was calculated both at the country-level and region-level. This diversification index (DIV) considers the Herfindahl-Hirschman Index (HHI), which is a calculation of variety that takes into account both richness (i.e., the number of farmed items) and evenness (i.e., how evenly the quantity produced is distributed among these items). The range of DIV is set from 0 to 1, where a score close to 1 indicates a highly diversified industry in terms of the families produced, and a score close to 0 indicates a sector that is highly concentrated on one family [14]. The calculation formula of DIV is as follows:

$$\text{DIV} = 1 - \sum\_{i=1}^{N} s\_i^2$$

where *Si*: share of production of species belonging to a family in total aquaculture production, and *N* is the number of fish families farmed in the aquaculture sector.

### **3. Aquaculture Production Trends in the Two Regions**

Although at the global level, freshwater aquaculture is expanding rapidly, there is spatial heterogeneity in development patterns both between regions and within each region. Figure 1 provides an overview of those countries in the two regions considered in this study, where freshwater aquaculture output fell over the last ten years.

**Figure 1.** Geographical scope of the study. Blue- and red-colored countries represent increasing and falling freshwater aquaculture production between 2007–2009 and 2017–2019, respectively.

### *3.1. Production in LAC*

Figure 2 presents the decadal changes in Latin American freshwater aquaculture production. During this period the output grew by 95% (from 476 to 927 kT), which is considerably higher than the growth rate of the global freshwater aquaculture level (60%) [1,5]. Brazil is by far the largest producer of LAC; it is the only non-Asian country in the top 10 of the global list of freshwater aquaculture producers (ranking 7th in 2019), and the 2.1-fold growth in Brazilian production over a decade is considerably higher than in other large global producers. However, in other major producers of LAC (Peru, Mexico, and Colombia) the sector grew at a rate even higher than in Brazil. Altogether the top-4 producers (Brazil, Colombia, Mexico, and Peru) account for 85% of total freshwater aquaculture output in the region, and contributed to 98% of the increment in production volume over a decade. Annual production in these four countries increased from 338 to 783 kT. On the contrary, there was a drop in output in some countries, including Ecuador and Chile in South America, and many of the Central American and Caribbean states (Cuba, Costa Rica, Jamaica, Panama).

Regional aquaculture development was centered around the growth of tilapia (mainly Nile tilapia) production, a non-native tropical fish with standardized rearing protocols which has robust domestic and export (USA and European) markets [15,16]. With a yearly output of 543 kT, tilapia contributes to 59% of regional production. Farming of characins, a family of tropical species native to LAC (mainly cachama, pirapatinga, pacu, and their interspecific hybrids), is produced entirely for domestic markets, and cold-water salmonids (almost exclusively represented by non-native rainbow trout) is also a rapidly growing segment in the region. Carp farming, a traditional and formerly important sub-sector in LAC aquaculture, has gradually lost its weight over the last decade (Figure 1).

**Figure 2.** Freshwater aquaculture production in LAC. Data source: [1].

### *3.2. Production in EU*

Contrary to significant development in Latin American and global freshwater aquaculture, output in the EU has not grown for decades. Production has slightly decreased from 284 to 280 kT over the last decade (Figure 3). Similar to Latin America, big differences exist between the development patterns of individual countries. There are marked west-east and south-north gradients in industry growth rates: aquaculture output in most of the Western and Mediterranean countries fell, on the contrary, Eastern and Northern EU states increased their fish production (Figure 1).

**Figure 3.** Freshwater aquaculture production in the EU-27 and in the top 8 producer countries (bar charts). Pie charts represent share of major groups in total production of EU-27. Data source: [1].

EU aquaculture is heavily concentrated on two species, which altogether account for 83% of production. Rainbow trout, a predatory species predominant in the aquaculture of Northern, Western, and Mediterranean countries, are farmed in cold-water systems. The production of this species fell in the period investigated, from 167 to 152 kT. The second most important farmed organism is the common carp (70 kT in 2007–2009 and 73 kT in 2017–2019), which is cultured at lower trophic levels in warm-water aquaculture, mainly in Eastern European EU states. Production statistics suggest that geographical patterns in aquaculture development are more important than general differences in growth pathways of different species, since carp production shrank significantly in France and Germany, despite the general growth of the carp industry in Eastern Europe.

In addition to rainbow trout and common carp, several other *Salmonidae* and *Cyprinidae* species are farmed as well, but in lower volumes. Next to salmonids and cyprinids, higher TL value species (*catfishes*, *sturgeons*, *perciform* sp., eel, pike) are cultured in the EU, which have a higher market value than cyprinids.

### *3.3. Trophic Level and Unit Value of Species Produced*

At a global level, capture fisheries supply markets with carnivorous species, whereas aquaculture focuses on species that are lower in the food chain, and carnivorous species make up less than 10% of farmed output [17]. In line with global trends, the majority of species farmed in freshwater aquaculture in LAC are omnivorous and herbivorous fish, and carnivorous species (TL > 3.5) account for less than 12% of total production. Unlike global and Latin American aquaculture, EU-27 fish farming is focused on carnivorous fish, which contribute to 66% of production, while herbivorous and omnivorous species at TL < 3.5 account for only 34% of the production.

At a global level, carnivorous species are traded with higher value and have larger production costs due to protein-rich feeds applied in farming [17]. For the LAC and EU, Figures 4 and 5 match the trophic level (TL) against the unit value of cultured species (including interspecific hybrids). Unlike general patterns in global markets of cultured species, in Latin America there is no (statistically) significant correlation between trophic level and market value; most of the carnivorous species are traded with values (<3 USD/kg) similar to those at lower trophic levels, with the exception of rainbow trout and arapaima that command a higher price on the markets. On the other hand, blue tilapia has a relatively high market value in spite of its herbivorous nature.

**Figure 4.** Bubble plot of the trophic level versus the unit value for the top-25 species in LAC aquaculture (calculated for 2019). The size of the bubbles relates to the production volume of a particular species. Cichlidae and Characidae species are marked in red and green, respectively. Items in italics are not species but higher-level aggregates. Data sources: [1,8].

**Figure 5.** Bubble plot of the trophic level versus the unit value for the top-25 species in EU-27 aquaculture (calculated for 2019). The size of the bubble relates to the production volume of a particular species. Salmonidae and Cyprinidae species are marked in red and green, respectively. Items in italics are not species but higher-level aggregates. Data sources: [1,8].

Unit values of fish in the EU are found in a wider range, from 1.6 (bighead carp) to 10.9 (eel) USD/kg, with a positive gradient along the trophic chain. There is a statistically significant correlation (r = 0.48, *p* = 0.02) between TL and the unit value of species, implying that European consumers have a willingness to pay higher prices for carnivorous species.

Diverting culture practices toward low trophic level species is identified as a strategy for sustainable aquaculture, to reduce nutrient loading and the demand for high-protein terrestrial or marine feed sources [18]. Each level up the trophic chain decreases the efficiency of utilizing energy produced by photosynthetic organisms. For this reason, metrics calculated with the trophic level are often used as indicators for sustainability [19,20]. Though the original meaning of TL has been blurred recently with the increasing share of vegetablebased ingredients in diets of farmed carnivorous species [21], the protein content (either it is sourced from vegetable or animal ingredients) and cost of aquafeed recipes are still higher for carnivorous species than for herbivores. Therefore, we continue to consider TL as a proxy indicator of the level of requirement for costly nutrients during the culture of fish species. Figure A1 illustrates the change in mean trophic level of freshwater aquaculture production (both at the region-level and country-level) between 2007–2009 and 2017–2019. In Latin America, there was only a slight increase in the mean trophic level of the regional aquaculture, from 2.30 to 2.32, which indicates the unchanged dominance of herbivore and omnivore species. On the contrary, the mean trophic level of EU aquaculture is relatively high (3.64 calculated for 2017–2019), but slightly decreasing with a rising share of carp in total production.

### *3.4. Diversification and Emerging Species*

Species diversification increases the resilience of industry by reducing its vulnerability to market shocks and species-specific disease outbreaks [22–24]. To analyze the diversity of the aquaculture sector, we used metrics reflecting the degree to which fish production is evenly distributed among more species. Figure A2 presents the calculated diversification index (DIV) and its change between 2007–2009 and 2017–2019, for the two regions considered. Higher values indicate higher diversity. The DIV calculated for the Latin American freshwater aquaculture was reduced from 0.68 to 0.59 in the last decade, which suggests that concentration of the industry has taken place, and the sector became less diversified at the regional level. The reduction in the DIV is mainly attributed to the increasing dominance of Nile tilapia in Latin American aquaculture (Table A1). In Brazil, Mexico, and Peru the diversification of fish production was reduced significantly, corresponding to a development pattern where an already dominant species becomes even more dominant in production (tilapia for Mexico and Brazil, trout for Peru). This reflects that the aquaculture industry sees the opportunity in concentrating efforts, investments, and infrastructure on the production of these species. However, rainbow trout and tilapia are non-native species, and most recent aquaculture plans (Peru, Colombia, Brazil) identify the culture of native species as a priority and promote this as a path to sustainability [I]. Figure 6 provides an overview of the aquaculture development of native emerging species.

**Figure 6.** Emerging species in freshwater aquaculture of EU-27: production quantities and unit values in 2007–2009 and 2017–2019.

In contrast with the Latin American freshwater aquaculture, the species diversity slightly increased in the EU-27 in the last years from 0.54 to 0.56. This is mainly attributed to the shrinking contribution of trout to total production, but the increasing output of emerging species (Figure 6) also contributes to increased diversity in European aquaculture. Most of these novel species are carnivorous species with high but falling market value. African catfish (and its hybrid, the Hetero-clarias catfish) is exceptional in that it is marketed at low prices. Being an air-breathing organism and its wide tolerance for water quality, the African catfish is cultured in high densities [25] with low per-unit fixed costs, allowing farmers to position it as a low-value species. Thanks to its low price, it has a stable domestic market, and its calculated market value increased over the last decade. However, being a non-native invasive species, there are ecological concerns over escapees from culture units [26]. EU countries are important contributors to global sturgeon meat and caviar output, originating from aquaculture, but in recent years demand for these products was lower than the offer [27]. This is reflected in the decreasing unit prices (Figure 7), which has a negative impact on the growth prospects of this industry. Production of perciform species (pikeperch, perch, and hybrid striped bass) increased double-fold over the period investigated. Pikeperch is the most important native percid fish in Europe, with a very

solid market price. Yet, various technological problems hamper the growth of pikeperch farming, such as unpredictability in reproductive performance and juvenile production [28]. Char farming is also an important emerging segment in EU aquaculture, especially in the Northern states, with the potential to diversify salmonid production [29]. Land-based Atlantic salmon farming is in its infancy, production is being upscaled in large RAS systems. Total RAS production of Atlantic salmon is larger than what is indicated in Figure 7, since there are land-based systems that produce salmon in salt water, and their production is reported under marine aquaculture production [29].

**Figure 7.** Emerging species in freshwater aquaculture of EU-27: production quantities and unit values in 2007–2009 and 2017–2019.

European aquaculture producers face import competition mainly from mid-value salmon and low-value pangasius originating from countries (Norway and Vietnam, respectively) where climatic and geographic conditions are ideal for these species, and the EU market penetration of these species is supported by a well-developed value chain. Conventional species (trout, carp) farmed in the EU do not have the perspective to increase domestic market share, therefore, European fish farmers try to find breakthrough points by diversifying production with species that are destined for supplying niche markets where international competition is lower.

### **4. Water Use and Resources in LAC and EU Aquaculture**

### *4.1. Water Resource Intensity of LAC and EU Aquaculture Production*

The intensity of resource use varies widely between culture systems. Therefore, first, we review major types of rearing systems used in aquaculture in LAC and EU before discussing the relationship between growth and freshwater resources. Although statistical reports do not break down production data by different farming systems for the Latin American region, based on literature sources it is obvious that earthen pond culture is dominant, especially in tilapia and characid sectors [2,16,30,31]. Pond farming technologies vary from low to high intensity, which differs in stocking densities, nutrient input, water quality, flow management, etc. A smaller part of the production takes place in staticwater ponds under extensive (non-fed) conditions, where supplementary water is only withdrawn to replace what is lost through evaporation [16]. The largest portion of the output is farmed under semi-intensive conditions in fed and fertilized earthen ponds, where water management is either similar to that of extensive systems [31] or a moderate flow rate is provided [30]. Intensive technologies in earthen ponds are operated with high flow rates (proportional to biomass density) to provide constant water refreshment [30]. In

addition to pond farming, the use of reservoirs for aquaculture is also common in LAC, either as a place for extensive management or intensive culture in floating cages [32–34]. Although recirculating aquaculture systems (RAS) are more and more widely used in culture of marine species using saltwater, they do not represent a significant share in Latin American freshwater aquaculture [16].

The main factors affecting specific (per kg) direct water use in production systems are yields (kg produced per m<sup>3</sup> or ha) and flow regime (frequency of water intake, intensity of water exchange). Feed-associated (indirect) water use is also significant in fed-systems, in the range of 1–2.5 m3/kg production [35,36] System-associated water use takes place on the production site, but, by contrast, feed-associated water consumption is often incorporated into imported crop ingredients (e.g., soybean), with implications on water resources found in regions/countries far away from the fish production site. Therefore, with the consequent aim of investigating how the development of the aquaculture sector in LAC and the EU depends on the spatial availability of water resources, we focus on the system-associated water requirements of different systems below.

At a global level, RAS and cage systems are considered to use blue water resources most efficiently, with a minimal (<0.5 m3/kg) water footprint [37–39]. However, accounting freshwater use to cage culture when multiple uses occur in water bodies is not consistent [14]. On the other end, flow-through systems are considered to be the least efficient systems in terms of using blue waters, usually with a footprint >50 m3/kg [38–40]. Pond systems, which are the dominant environment for freshwater fish production both globally and in LAC, are in between RAS/cage and flow-through systems in terms of water use, with footprint values between 3 and 40 m3/kg, depending on yields, evaporation and seepage conditions at the production site, and water refreshment regime applied [37]. Generally, it is considered that specific water use has an asymptotic relationship with aquaculture production intensity, since more intensive production systems were found to use water resources more efficiently (per kg of fish produced) than extensive production systems [35,37,39,41,42]. Results of studies assessing water use in LAC and EU are summarized in Table 1.


**Table 1.** Per kg water use in typical fish production systems in LAC and EU as per literature sources.


**Table 1.** *Cont.*

<sup>1</sup> Definitions of direct (system-associated) water use vary across studies. Most studies calculate the total amount of water withdrawn for production, which is larger than consumptive water use. <sup>2</sup> Non-fed, fertilized system with 1.5 kg/m<sup>3</sup> max. biomass density. <sup>3</sup> Fed and fertilized system with a daily 30% water exchange. Max biomass density is 25 kg/m3. <sup>4</sup> Fed system with a daily 100–400% water exchange. Max biomass density is 40 kg/m3. <sup>5</sup> Fed, fertilized and aerated system, with supplementary water intake (offset evaporation loss) 9–14 t/ha. <sup>6</sup> Fed system in static water body. Max density is 37–43 kg/m3. <sup>7</sup> Water exchange is 0.1 m3/kg feed. Culture density is >300 kg/m3. <sup>8</sup> Fertilized and fed system with supplementary water intake (offset evaporation loss). Yield is 710 kg/ha. <sup>9</sup> Constant water flow diverted from a river; oxygen supply is provided.

Calculated per kg water demands of species farmed in Latin American systems (Table 1) fall in line with finding for other regions of the world discussed above. Although results are not supposed to be directly compared since different studies use different methodologies with different system boundaries, it is important to note that a recent study found that intensive tilapia culture was associated with a higher blue water footprint than extensive farming due to high flow rates of refreshing water in the former technology [30]. This is contradictory to common findings for other regions, as discussed above, and it may challenge the view that intensification in LAC comes with water resource efficiency.

For EU-27, the statistical office of the European Union reports aquaculture production data by production method (farming system) [45]. Based on data available it is estimated that 48% of freshwater production originates from flow-through pond/tank/raceway systems, 38% is produced in static-water earthen ponds, while RAS systems and cage/pen aquaculture account for 10% and 4% of production, respectively. Under flow-through conditions mainly trout [46], and to a lesser extent, African catfish, are cultured. Coldwater trout are often reared in surface water diverted from smaller water courses, while warm-water catfish are farmed in subterranean geothermal water. In the pond farming segment, typically a semi-intensive carp-dominant polyculture is practiced with low (<1 t/ha) yields [47–49]. Contrary to Latin America, European RAS systems are constructed primarily to farm freshwater species, mainly trout, catfishes, and sturgeons [50]. There are farms also that rear Atlantic salmon and eel in a freshwater RAS environment [29]. Unlike many regions of the world, where cage farming is an important segment of both freshwater and marine aquaculture, in the EU cage systems are not typical in freshwater environments [51], only some facilities exist to farm carp and sturgeon in reservoirs and on cooling water of thermal power plants. To minimize the discharge of trout farms and comply with strict environmental regulations, partial recirculation of water was a tendency in Denmark, one of the largest producers in the EU. The main advantage of these systems is reduced nutrient emission, but there are some disadvantages that limit the development of RAS culture, such as high capital cost and worse energy efficiency due to automation [29,50,52,53].

The water demand of the flow-through trout farming segment is high (50–100 m3/kg), and this can be reduced by up to two orders of magnitude (to 0.1–2 m3/kg) if systems are converted to RAS [36,38,54]. Carp produced in semi-intensive pond production in an Eastern European climate have a water demand of around 20 m3/kg ([55] and calculations in Table 1).

### *4.2. Role of Water Resources in Aquaculture Development*

In the previous section, it was highlighted that aquaculture production growth requires some 5–50 m3 of water per kg of additional capacity, depending on the species and production system. Tilapia and carp aquacultures in most typical semi-intensive systems demand 10–30 m3/kg, while trout produced in conventional flow-through require

more than 50 m3/kg. Here, we match between-region and between-country differences in growth rates with differences in freshwater resource endowments. In our study, we examined two regions: LAC, which are abundant in water resources with a TRWR value of 21,476 m3/capita/year, and the EU-27, which have a TRWR less by an order of magnitude (3041 m3/capita/year). Growth in annual freshwater aquaculture production over the last ten years was −0.01 kg/cap (the EU) and 0.70 kg/cap (LAC).

Figure 8 plots the per-capita availability of annually renewed freshwater resources against per-capita growth in the aquaculture sector in the last decade for the top 12 producing countries in each region. Per-capita growth if aquaculture was calculated as the difference between per-capita production in 2017–2019 and in 2007–2009. Therefore, countries with increasing populations and slightly increasing production may have negative values for per-capita change in fish production (e.g., Denmark). The calculated Pearson-r correlation between the two variables is 0.53 (*p* = 0.08) for Latin American countries, while for European countries it is 0.75 (*p* < 0.01) if outlier data for Bulgaria was excluded. These values suggest a positive relationship between per capita freshwater aquaculture development and per capita freshwater availability. In Latin America, Peru, Colombia, and Brazil are the most water-abundant countries, and these countries are ranked 2nd, 1st, and 4th in terms of per capita aquaculture growth, respectively. On the other hand, Cuba is characterized by the lowest water resource availability in LAC, and this corresponds to the biggest reduction in aquaculture production.

**Figure 8.** Bubble plot of the Total Renewable Water Resources (TRWR, 2018–2022) versus per capita growth of annual freshwater aquaculture production over a 10-year period (from 2007–2009 to 2017–2019) for top-12 freshwater aquaculture producers in LAC (**upper**) and EU (**lower**) graph. The size of the bubble relates to the freshwater aquaculture production (t/year) of the corresponding country (avg. for 2017–2019). Note that the x-axis is log-scaled, and scales differ between the two graphs. Data sources: [1,9].

Among the major freshwater fish producer countries in the EU, Sweden, Hungary, and Romania have the largest volume of water resources, corresponding to positive growth rates of aquaculture on a per-capita basis in these countries. Sweden has the highest water abundance among the major producers in the EU and this enables the high growth rate of trout production in flow-through systems, which have the highest water demand among European systems. Recirculation aquaculture is relatively undeveloped in Sweden [HH]. Among countries that have a TRWR value of less than 4000 m3/capita/year, it can be seen that countries where carp-based pond aquaculture is dominant (Czechia, Poland, Bulgaria) increased their production, while per-capita aquaculture output fell in countries, where aquaculture sector is based on flow-through through systems (France, Germany, Italy, and Spain). Considering that flow-through systems (with a footprint of >50 m3/kg) are more sensitive to water stress than carp-based pond farming (~20 m3/kg), aquaculture development patterns can be partly explained by the difference in the degree of vulnerability of different systems to temporal water shortages, which are more frequent with climate change [25,56–59].

In water-poor regions, one strategy to maximize production value per m3 of water used is to farm high-value species in recirculation aquaculture systems (RAS), which minimize water footprint. RAS aquaculture (farming sturgeons, eel, catfish, trout) has developed rapidly, especially in the European countries where per capita water renewable resources are below 4000 m3. Denmark, France, Germany, Poland, and Spain altogether account for 75% of RAS production in the EU [50].

In addition to freshwater resource availability, the potential growth of freshwater aquaculture is also determined by climatic conditions since water temperature in most culture systems is under the control of the climate. The species for culture must be selected so that the range in temperature preference and tolerance of the species chosen is in harmony with the local climate [60]. Figure 9 provides an overview of the climatic conditions of the top 10 producers in both regions matched with the thermal preferences of the main target species. Although the graph represents data for air temperatures, it is often assumed that there is a linear relationship between air and water temperatures [57,61].

**Figure 9.** Climatic conditions (range of mean monthly air temperatures) versus thermal optimum of major cultured species in the top 10 fish producing countries of LAC (**left**) and EU (**right**). The size of the bubbles relate to the production volume of species. The whiskers encompass the optimal water temperature range for each species. Source of data: [11,12,62].

Most of the Latin American countries have a tropical climate with little variation in monthly temperatures, which favors aquaculture production by enabling them to plan production cycles without seasonality. Even in sub-tropical countries (Mexico), the range is narrower than in European countries, and warm-water species can be fattened in the colder season. Cold-water trout farmers at higher altitudes can also benefit from near-toconstant temperatures, as is shown in rainbow trout aquaculture development in Peru, the country with the coldest annual mean temperature among the major producer countries of the region.

Most of the EU territories are under a temperate climate, with large variations in monthly temperatures, therefore there is a strong seasonality in fish production cycles in open systems. The graphical tool helps the understanding of the difference in growth between trout and carp production over the last decade. Carp is a robust species with wide temperature tolerance and low biological sensitivity to environmental changes [63]. Temperature increase, which is an ongoing tendency, is forecasted to favor the metabolic activity and growth rate of carp under Eastern European conditions, since prevalent temperatures are far away from its upper limit of thermal preference [57,61]. On the one hand, trout is a species with a relatively low upper thermal limit, and consequently, warming may significantly enhance trout mortality and affect productivity, especially in Mediterranean countries [58]. In light of this, climate change contributes to the explanation of the difference in aquaculture production changes between Mediterranean and Northern European trout farming countries.

#### **5. Emission of Aquaculture Production**

Aquaculture generates emissions either to the air or to the aquatic space. The most pronounced environmental concerns are over (i) the release of nitrogenous or phosphorus, which may stimulate eutrophication processes in the receiving water body, and (ii) greenhouse gas (GHG) emission [64]. Unlike water footprint, which is mainly generated during on-farm activities, the majority of aquaculture-related GHGs are emitted during feed production, thus carbon footprint is largely determined by the feed conversion rates and the ingredients used in aquafeeds [65,66]. This implies that the nutritional habit of the cultured species and the regional availability of ingredients matching these nutritional requirements have a major influence on climate change mitigation. A recent study using relatively narrow system boundaries and standardized methodology across different systems and species found that tilapia farming in LAC has a significantly lower carbon footprint (2 kgCO2eq/kg fish, in live weight) than the global average tilapia production (3.7 kgCO2eq/kg fish), and this emission efficiency is mainly attributed to regional feeds with lower footprints and lower use of fossil energy during on-farm processes [66]. In the same study, the GHG emission of European carp production is calculated to be lower (1.6 kgCO2eq/kg fish) than the global average carp carbon footprint (3.2 kgCO2eq/kg fish). However, if the system boundaries of the analysis are expanded, the carbon footprint of carp production is found to be significantly higher (6 kgCO2eq/kg fish), as infrastructure maintenance (pond dredging) and post-harvest operations (packaging and transport) are responsible for a large amount of greenhouse gas emission [67]. While most systematic review studies conclude that per-unit GHG emissions of tilapia and carp production are in a similar range, there is disagreement on whether the carbon footprint of salmonids (including trout) is higher or lower than that of carp and tilapia [65,66,68]. Similarly, there is a lack of consensus in answering the question of whether RAS produced trout have a higher carbon footprint than one from a flow-through system, as the GHG emission during RAS production is largely dependent on the source (renewable or fossil) of the electricity used for operating the system [69]. Nevertheless, on-farm energy use in RAS technology is higher than in other systems, but in the post-harvest stage the fuel demand is often lower with shorter transportation routes because RAS facilities are built in the proximity of markets [70].

Nutrient emissions of aquaculture segments are determined by the utilization (retention) efficiency of input nutrients in farmed organisms, and treatment/recovery of the non-utilized part of nutrients. While the former factor is more species-specific, the latter one is system-specific. In flow-through and cage systems the non-retained part of nutrients is generally discharged with water exchange [53,71]. In RAS systems effluents are treated and solid wastes are collected [72], while in static-water pond culture part of the non-retained nutrient input is recycled through the food web and recovered in the plankton biomass [57]. For this reason, feed nitrogen and phosphorous conversion efficiency are relatively high (>40%) in European and Latin American pond cultures [47,73,74]. If the total (feed and fertilizer) nutrient inputs are considered, then pond farming has low conversion efficiency (<20%) in comparison to other systems, because nutrients present in fertilizers are not directly utilized by target fish species and transition losses arise with nutrients transferred through three levels of the trophic web (fertilizer-phytoplankton-zooplanktontilapia/carp) [57,75]. However, we argue that the nutrient efficiency of fertilized systems cannot be directly compared with fed systems, as for the latter one fertilizer input during the production of crops used as aquafeed ingredients should also be accounted for.

### **6. Conclusions and Perspectives**

There are several factors that play a significant role in aquaculture development, including market demand, environmental concerns, licensing regulations, and institutional capacity [76]. This study was not written with the objective to discuss socio-economic influences that may limit the exploitation of resources, rather it concentrated on production trends and underlying factors endowments as available from aggregate statistics. We investigated the climate and availability of freshwater resources, which are crucial factors in aquaculture development [77], and shed light on their influence on the growth prospects of the aquaculture sector. The LAC, accounting for one-third of the world's total runoff [78], is well-endowed with currently underutilized renewable water resources [10] and still has a huge scope for expansion. In the European context, it is often cited that bureaucracy and restricting environmental regulations are barriers to growth [73,79], but it needs to be further understood whether regions poor in natural resources tend to have more strict environmental rules to ensure the conservation of biodiversity and ecosystem functioning and whether socio-economic and institutional influences are themselves consequences of resource scarcity. In fact, many of the European producers see the future potential of the industry rely on subsidies, rather than expansion of physical output [80,81].

**Author Contributions:** G.G.: Conceptualization; methodology, validation; formal analysis, investigation; writing—review and editing, visualization, supervision; J.P.C.T.: Conceptualization, Data curation, formal analysis, investigation, writing—original draft preparation; E.B.: Conceptualization; writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding author upon request.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

### **Appendix A**

**Table A1.** Aquaculture production of top 25 species in LAC and EU-27. Items in red are not species, but larger aggregates as per given by Aquatic Sciences and Fisheries Information System (ASFIS). Source: [1].


**Figure A1.** Average trophic level of freshwater aquaculture production at the region-level and country-level (calculated for 8–8 largest producers) in 2007–2009 and 2017–2019. Data sources: [1,8].

**Figure A2.** Calculated diversification index (DIV, between 0 and 1) in freshwater aquaculture production at the region-level and country-level in 2007–2009 and 2017–2019. Note that DIV is 0 for Honduras because one family accounts for 100% of production. Data source: [1].

### **References**

