Hydropower plant in Eastern Amazon and its impacts on the local population’s eating habits

This study assessed the direct impacts of Cachoeira Caldeirão Hydropower Plant (before and during its construction) on the eating habits (fish and wild meat consumption) of the residents in Porto Grande City, Amapá State, Brazil. The study was conducted in 2015, based on a questionnaire for data collection (Sample size = 53) and on comparative analysis for significance (Wilcoxon test, p < 0.05) and correlation (Pearson’s test) assessment. Results suggested significant changes in families’ eating habits. The Wilcoxon test detected three significant variables ( p < 0.05): family income, number of species (fish and wild meat consumption) and rate of wild meat consumption per family. Both the Wilcoxon and the Pearson’s correlation tests confirmed two variables ( p < 0.05 and r > 0.7): family income versus wild meat consumption before ( p value = 0.045; r = - 0.75) and family income versus fish consumption during ( p value = 0.0029; r = - 0.83) hydropower plant construction. It can be inferred that the decrease in family fish consumption may be related to changes in the families' eating habits. Changes in diet may have led families to a growing consumption of industrialized and/or ultra-processed foods, very common in geographically isolated Amazon riverside regions with limited access to electricity. Furthermore, families were forced to adapt to a new reality due to environmental changes in their territories.


INTRODUCTION
Economic growth planning in developing countries is usually based on major infrastructure projects, which highly worsens environmental impacts at the local scale. Economic growth and policies in Brazil have been based on large businessessuch as hydropower plantsthroughout the last fifty years (Guimarães, 2003;Moretto et al., 2012).
This infrastructure-project type leads to the following main impacts: downstream and upstream activity impairments, water quality decrease, water body silting, waterborne disease increase, natural resource loss around reservoir areas and, mainly, local population and economic activities' compulsory displacement from affected areas (Acselrad, 2004;Giusti, 2005;Bermann, 2007;Araújo and Belo, 2009;Santos et al., 2017).
Therefore, building hydropower plants is one of the most controversial activities affecting development rates in tropical countries, as well as the leading cause of environmental and social issues. For example, the 79 planned Amazon dams are likely to flood approximately 3% of the forest and cause irreversible damage to local biodiversity and to traditional local populations (Zhouri and Oliveira, 2007;Sieben and Cleps Junior, 2012;Fearnside, 2013). Amazonian river basins have been overexploited in order to meet energy demands, and this accelerates the socioenvironmental degradation process and disregards appropriate environmental management (Filizola et al., 2002;Holanda et al., 2011).
Although hydropower plants are acknowledged for their advanced environmental protection policy, they remain controversial due to their impacts, namely: physical effects, socio-environmental changes in territories where they are installed and insufficient corrective and/or mitigation measures (Richter et al., 2010;Cavalcante and Santos, 2012).
Accordingly, hydropower plants have been the leading cause of significant socioenvironmental impacts on the Araguari River. The local population has been the most affected segment during the building and installation processes (Sousa, 2000;Sá-Oliveira et al., 2016;Santos et al., 2017).
The local population is mainly affected by negative effects on aquatic and terrestrial fauna dynamicsmainly on the riversidesdue to hydropower plants, since they are socioeconomically, nutritionally and environmentally dependent on rivers (Begossi, 2004;Marques et al., 2018). Diagnoses on food sources and the social vulnerability of these populations can indicate the degree of dependence on, and the level of relationship with, local natural resources (Begossi, 1993;Zagallo and Ertzogue, 2018).
The relationship between populations and enterprises is asymmetrical, as popular access to the environment depends on other social strata. Thus, conflicts arise when natural resources from biodiverse systems are appropriated (Martinez-Alier, 1999;Acselrad, 2004).
Accordingly, the eating habits of local Amazonian residents are essential to assess natural resource sustainability and to measure the potential impacts of hydropower plants' construction, installation and operation (Lekwot et al., 2016;Zagallo and Ertzogue, 2018).
The present study assessed the factors affected by the construction of Cachoeira Caldeirão Hydropower Plant, as follows: family income in Porto Grande City (Amapá State, Brazil) residents, number of consumed fish and wildlife species, food consumption rate.

Study site
Cachoeira Caldeirão Hydropower Plant (CCHPP) is located in Porto Grande City, Southern Amapá State, Brazil. Porto Grande City is 130 km away from Macapá (the capital city). It has a territorial area of 4.428.013 km 2 , estimated population of 21.484 inhabitants and demographic density of 3.82 inhabitants/km 2 (IBGE, 2018).
Porto Grande City is bathed by the Araguari River, whose extensive water network is approximately 617 km in length and records 0.955/km drainage rate (Cunha et al., 2011). According to the new Brazilian river basin classification by the National Water Agency (ANA), the Araguari River Basin is entirely located within the Amazon Basin region, most precisely within Amapá State's hinterlands ( Figure 1). Porto Grande City's urban zone was the site for the study, most precisely the residential area on the Araguari River's banks, since it was the most affected by the CCHPP installation.

Data collection and analysis procedures
The survey was conducted in 2015, based on questionnaires assigned to 53 families affected by Cachoeira Caldeirão HPP installation. The questionnaires held open-and closedended questions, as follows: a) What was your family income before and during the hydropower plant's installation? The currency rate (converted into dollars) was R$ 3.12, by June 2015.
b) Which were the wildlife species consumed by you and your family, and how often were they consumed per month, before and during the hydropower plant's installation? c) Which were the fish species consumed by you and your family, and how often were they consumed per month, before and during the hydropower plant's installation?
The questions referred to food consumption prior to the hydropower plant's installation (August 2013) and during its effective construction (October 2013 to June 2015).
Respondents reported the vernacular names of fish and wildlife species, so they could be gathered and used for data analysis. The families or genera that the fish species mentioned by respondents belonged to were assessed for taxonomic identification, based on the following studies: Ethnoecology of fishermen in the Amapá National Forest (FLONA), by Brandão and Silva (2008); Observation of the lifestyle of FLONA residents, by Costa (2013); Fishery classification in Coaracy Nunes (Amapá State) HPP Reservoir, by Sá- Oliveira et al. (2013).
These surveys were carried out near the study site, since they addressed local families affected by Cachoeira Caldeirão HPP. The International Union for Conservation of Nature's (IUCN, 2018) Red list of Threatened Species was used as reference to identify and audit the scientific names of fish and wildlife species.
Data surveys were carried out from April to June 2015. Residents were contacted prior to the survey in order to assess the number of households affected by the hydropower plant's construction. The total of 224 households was identified. Sampling at 90% confidence level and 10% sampling error indicated 53 affected families, as shown in the Equation 1 below.
Where in: n = sample; e = margin of error; z = z score; N = sample size.
Respondents were randomized through the snowball sampling method for questionnaire application, since the affected families were not residing in the affected areas. "Catalyst" participants were identified, so that they could successively indicate new participants.
Interviews were conducted in the respondents' respective households and interviewees were preferably the heads of the family. Family members who had information on the assessed subject were interviewed in their absence.
Data collected and recorded in the field survey were systematically arranged in Microsoft Excel 2010 and analyzed in BioEstat 5.3. software (Ayres et al., 2005).
The number of consumed fish and wild meat species, as well as consumption periodicity (before and during the hydropower plant construction) were assessed through Wilcoxon test for paired samples, since it measures and identifies differences between ordinal and numerical variables observed in a given timeframe, after an event. Correlations between variables were assessed through Pearson's correlation test.

Family income data
Although an expropriation program was conducted in directly affected areas, 20% (11) of the families were not expropriated, whereas 80% (42) of them were expropriated and received compensation for their properties.
Family income changed throughout the assessed periods: Although mean monthly family income was R$ 1.824,79 (586.56 USD) before the hydropower plant's construction, it decreased to R$ 1.087,54 (349.57 USD) during its constructionmean decrease by 40.40% (Figure 2). The Wilcoxon test for paired samples detected a significant difference (p value = 0.002) among family incomes before and during the hydropower plant's construction. Cities surrounding hydropower plants, or affected by them, experience changes in their socioeconomic dynamics over time (Tetteh et al., 2004). These socioeconomic changes can have different effects on each social stratum. In essence, dam projects strongly affect the economy of families living near them (Filizola et al., 2002;Holanda et al., 2011).
Other studies have shown similar impacts on the income of families affected by dams. The hydrological cycle of the São Francisco River was altered due to Xingó HPP construction, which compromised traditional fishing and agriculture activities (Holanda et al., 2005). Fishermen from Amapá State experienced negative change in their income due to Araguari River's damming during Ferreira Gomes Hydropower Plant's installation (Santos et al., 2017). Family income of São Paulo State's population affected by Tijuco Alto Dam was substantially altered because of the forced displacement from fertile areas (Jeronymo et al., 2012).
Changes in economy and health experienced by communities living near dams are some of the impacts of these projects; those living farther away from them do not experience such a reality (Tetteh et al., 2004). Accordingly, interviewed families showed substantial income deficits, and it was indicative of a negative relationship among dams' projects, local population's activities and income sources.
Although changes in family income have been evidenced by other surveys, it is worth noting that socioeconomic impacts depend on the following factors: dams' projects region and the interaction of different social groups with it; social production conditions; dams' projects conduction. Several case studies on the socioeconomic effects of hydropower plant installation have allowed some generalizations about the constancy and intensity of its effects on the affected population (Reis et al., 2003).

Animal protein consumption
Family fish consumption decreased during the assessed period: 96.2% (51) of surveyed families consumed fish before the hydropower plant's construction, but this rate dropped to 73.6% (39) during its construction. The most consumed fish species were: Hoplias sp. (trairão), Tometes sp. (curupeté) and Myleus sp. (pacu-branco) ( Table 1).  Accordingly, family wild-meat consumption has sharply decreased: 96.2% (51) of surveyed families consumed this protein before the hydropower plant's consumption, but this rate dropped to 35.8% (19) during its construction. The most consumed wildlife species were Cuniculus paca, Priodontes maximus and Mazama americana ( Table 2).
The aforementioned fish species are compatible with those found in the Araguari River and with the ones consumed by riverside communities in the Araguari River Basin (Brandão and Silva, 2008;Sá-Oliveira et al., 2013;Oliveira et al., 2018). This finding proves that fishing is an essential part of the diet of the interviewed families. The consumed species in both assessed periods showed significant difference (p value = 0.007) in the Wilcoxon test for paired samples.
Results of species diversity and abundance reduction around dam reservoirs corroborate those in similar studies (Nascimento et al., 2011;Sá-Oliveira et al., 2013;Santana et al., 2014;Oliveira et al., 2018). Yet, the association between qualitative and quantitative decrease in fishery resources facing environmental changes, and uncontrolled fishing, is common around hydropower reservoirs (Agostinho et al., 1999;Gavarello et al., 2010). We Biological surveys of the ichtofauna, carried out before the construction of the Cachoeira Caldeirão HPP (IBAMA, 2007;Soares et al., 2012;Sá-Oliveira et al., 2013), showed the occurrence of fish species similar to those mentioned by the interviewees, and studies carried out after its construction (Morales, 2018;Oliveira et al., 2018;Costa et al., 2020) also found the presence of the mentioned species. Thus, it is possible to state that the diversity of fish species was not negatively impacted; however, the abundance may have been influenced by the construction of the Cachoeira Caldeirão HPP, as reported by the interviewees and statistically confirmed with the species consumption frequency data.
The Wilcoxon test for paired samples reported statistically significant difference (p value = 0.0037) in wild meat consumption between both periods (before and during the hydropower plant's construction). This outcome is indicative of the hydropower plants' impacts on the availability of animal species. Such impacts could be attributed to deforestation at the reservoir's surrounding areas, since it affects the local fauna.
Although wild meat consumption is common in the assessed region, none of the consumed species is on the IUCN Red List of Threatened Species (2018). All listed animals are widely spread and can be easily found throughout the Amazon Rainforest and Amapá State's territories.
The interviewed families described hunting and wild animals' sales as common daily activities for local populationsa fact evidenced in similar studies -, although these practices are illegal in Brazil (Dias Junior et al., 2014;Figueiredo and Barros, 2016).
The diet of Amazonian families culturally consists of fish and wild meat. Hunting is most common in the rainy season, since increased river levels hinder access to fish due to their dispersion in flooded areas. Thus, game meat is one of their main food alternatives and plays a vital role in their subsistence (Murrieta and Dufour, 2004;Murrieta et al., 2004;Adams et al., 2005;Silva, 2007;Figueiredo and Barros, 2016).
It is worth noting that hunting and wild meat consumption are related to food choice and consumption. These activities are influenced by ecological, economic and cultural features that shape the sociocultural context of populations that consume this protein type. However, few studies specify the different aspects of wild fauna management, mainly that in the Amazon Rainforest (Richards, 1939;Firth, 1961;De Garine, 1994;Descola, 1998;Viveiros De Castro, 9 Hydropower plant in Eastern Amazon and its … Rev. Ambient. Água vol. 16 n. 5, e2700 -Taubaté 2021 2002; Ziober and Zanirato, 2014;Figueiredo and Barros, 2016).

Animal consumption periodicity
Fish consumption did not significantly change between months. Although there was drastic decrease in fish consumption from 30 (-84.62%) to 12 times a month (-69.23%), the Wilcoxon test for paired samples did not show significant difference (p value = 0.5136) between the periods before and during the hydropower plant's construction. Such a result indicates that family consumption may be associated with fish purchase rather than with subsistence fishing in the area directly affected by the hydropower plant (Table 3). The results suggest that the decreased fish consumption may be related to changes in the eating habits of families that may have partially replaced it with other industrialized and/or ultra-processed foods, such as frozen chicken, canned food, sausages and bolognawhich are very common foods in geographically isolated Amazonian riverside areas lacking access to electricity. However, the hydropower plant prevented access to some fishing areas, mainly the surrounding ones. According to respondents, fish became scarce in traditional fishing areas due to increased flow of ships and progressive riverside deforestation.
River seasonality is yet another variable (not herein assessed) that can influence fish consumption. Fish availability can be high or low depending on the water level in both wet and dry periods (Pezzuti and Chaves, 2009).
It should be noticed that fish is one of the main cultural protein sources of the Brazilian people. Mean consumption in urban areas is 3.3 kg/year, whereas it ranges from 4 to 7.6 kg/year in rural areas and Amazonian towns (IBGE, 2010). Drastic and sudden decrease in protein consumption can aggravate local populations' vulnerability, mainly in low-income families that depend on fishing for socioeconomic and food subsistence.
With respect to wild meat consumption periodically, there was a drastic decrease in the regular consumption: 1, 4 and 8 times per month. The Wilcoxon test for paired samples showed significant difference (p value = 0.00322) in consumption between periods before and during the hydropower plant's construction (Table 3).
Hunting by other communities near the study site must also be taken into consideration when it comes to assessing decreased wild meat consumption, since this activity may have influenced the abundance of consumed species in the assessed period. However, data gathered so far is insufficient to test this hypothesis.
Therefore, animal-consumption changes depend on the environmental context of a given population, since access to alternative protein sources depends on resource availability. Thus, ecological and geographical conditions are key in determining protein source (Torres et al., 2018). In addition to these variables, forest coverage and distance from downtown can influence wildlife distribution (Foerster et al., 2012;Mgawe et al., 2012).
Hunting and wild meat consumption are basic conditions to ensure food provision and family income supplementation (Cymerys et al., 1995). Thus, identifying the eating habits of a given population is key to understanding the relationship between man and biological changes in the environment (Dufour and Teufel, 1995).
The correlation matrix showed significant difference and negative correlation for "family income" association with variables "wild meat consumption before the hydropower plant's installation" (p value = 0.045 and r = -0.75) and "fish consumption during the hydropower plant's installation" (p value = 0.029 and r = -0.83) (Figure 3). Results can indicate that families have prioritized other protein sourcessuch as fishover wild meat in the period before the hydropower plant's installation, since their income was higher in this period than during the installation process. On the other hand, families have seemingly prioritized the consumption of industrialized and/or ultra-processed food throughout the installation process, as the wild meat supply decreased. It is noteworthy that most fish and wild meat can be purchased in local markets, a very common practice in Amapá's cities, rather than obtained through hunting and subsistence fishing.

CONCLUSIONS
The impacts of the hydropower plant on families' fish and wild meat consumption is attributed to changes in the local environment, which affected species' distribution and abundance. These facts were identified in studies on environmental impacts caused by the hydropower plant and confirmed by interviewed families. Moreover, these impacts have seemingly affected local businesses and the consumption of some species by the local population.
Therefore, family eating habits have significantly changed. In addition to family income, the frequency of some fish and wild meat consumption and monthly wild meat consumption by families were also affected. Correlation significance between family income and protein consumption before and during the hydropower plant's construction indicates that families have changed their eating habits due the environmental transformations caused by the hydropower plant's construction in their territories, which imposed diet adaptation on them.
Although the hydropower plant has affected the diversity and abundance of local aquatic and terrestrial fauna, fish and wild meat consumption may have been partially compensated by products sold in neighboring cities. Therefore, the assessed variables could be completed by assessments of the intensity and extent of socio-environmental impacts on the local population.