Sustainability Dimensions Assessment in Four Traditional Agricultural Systems in the Amazon

Although traditional agriculture carried out by ethnic groups is considered for its high biodiversity and important for food security and sovereignty, few studies have investigated the potential of these systems in the interest of promoting a sustainable agricultural development policy according to United Nations Sustainable Development Goals. Using the FAO's Sustainability Assessment of Food and Agriculture (SAFA) methodology, this study analyzed the sustainability of four traditional agricultural systems, three indigenous (Waorani, Shuar, and Kichwa) and one migrant settler populations in the Yasuní Biosphere Reserve (YBR) and identified synergies and trade-offs among the dimensions of sustainability. The results showed different dynamics in all dimensions of sustainability-specifically, trade-offs in the dimensions of good governance with environmental integrity and social well-being, economic resilience, and social well-being. It was identified that the differences in terms of sustainability are narrowing between the indigenous Shuar people's traditional agricultural systems and those of migrant settlers, which provides policymakers with specific information to design sustainable development policies and rescue traditional agricultural systems in the Amazon region.


INTRODUCTION
Traditional production systems have several attributes that contribute to sustainable development, such as high species diversity with strong environmental adaptability (Zhang et al., 2011) and resistance to pests and diseases (Flores-Delgadillo et al., 2011). They can act as instruments for resilience, mitigation, and adaptation to climate change (McCord et al., 2015;Torres et al., 2015Torres et al., , 2018Tesfaye and Tirivayi, 2020). Such systems can help manage risk and decrease rural poverty and food insecurity (Michler and Josephson, 2017;Waha et al., 2018;Lovo and Veronesi, 2019;Mustafa et al., 2019;Bellon et al., 2020) through providing income from local markets (Bellon et al., 2020) and nutrition and child health in impoverished rural households. Due to the great diversity of edible and commercial crop species (Perreault, 2005), traditional production systems have been supporting many indigenous populations who live in voluntary collective isolation. To protect the sustainability of these systems, instruments are needed to reduce the possibility to have these communities becoming or being forced to be in contact with the outside world (Cabodevilla, 2008;Beckerman et al., 2009;United Nations, 2009;Colleoni and Proaño, 2010;Heredia-R and Hernández, 2019), based on these antecedents, Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES; http://www.ipbes.net/) and the Convention on Biological Diversity (CBD) recognizes that the diverse social, cultural and environmental knowledge of indigenous peoples and local communities (IPLC) contributes extensively to sustainability across large parts of the globe, and thus has a major role to play in assessments and policy formulation for biodiversity and ecosystem services (IPBES, 2012;Tengö et al., 2017;Hill et al., 2020).
The Ecuadorian Amazon Region (EAR) is home to 11 Indigenous groups who have inhabited the land for thousands of years. Sixty years ago, several populations of migrant settlers arrived after the discovery of significant oil reserves in the northern EAR. Roads were opened by oil companies (Pichón, 1997;Bilsborrow et al., 2004), leading to a change in the demography with populations of heterogeneous life strategies . The northern EAR has become a multiethnic and multicultural system. Within this region, the Yasuní Biosphere Reserve is populated by Indigenous Waorani, Shuar, and Kichwa people as well as migrant settlers who arrived mainly from rural areas of the Ecuadorian highlands (Santos, 1996;Murphy et al., 1997;Barbieri et al., 2003Barbieri et al., , 2009Bilsborrow et al., 2004;Sellers et al., 2017).
These groups manage their own traditional agricultural systems, which are part of each cultural heritage and geographic origin, for their tangible and intangible assets (Rudel et al., 2002;Lu and Bilsborrow, 2011;Ima Omene, 2012;Torres et al., 2015;Zurita-Benavides, 2017). However, climate variability, deforestation, and poor land management practices have led local populations to depend on agrochemicals (Davis et al., 2012) or in the expansion of the agricultural frontier to increase incomes (Mena et al., 2006(Mena et al., , 2011Walsh et al., 2008). Land use changes and degradation have resulted in the decrease in agricultural sustainability as a means of livelihood. Agricultural intensification policies, migration processes, population growth, and low educational levels (Viteri-Salazar and Toledo, 2020) have led to socioeconomic and demographic changes that include alterations in household composition and other family life cycle factors . It is necessary to analyze these factors from a sustainability perspective to gradually define potential solution for the future under the 2030 Sustainable Development Goals.
The sustainability of traditional production systems in the Amazon has generally been assessed from an environmental perspective (Galford et al., 2013). Nevertheless, social, economic, and governance dimensions of sustainability are also important, but little studied in this region (SAFA, 2012). Sustainability assessment aims to better understand the conditions of a location (from a community or region to a country) to support future decisions making that support all three pillars, i.e., social, economic, and environmental, of sustainable development (Bond et al., 2012). Several approaches have been developed including (Belcher et al., 2004) the human well-being index (Kaivo-oja et al., 2014), the environmental performance index (Hsu and Zomer, 2014), and the sustainable development index (Barrera-Roldán and Saldívar-Valdés, 2002). A multicriteria approach is often encouraged to ensure that the social component is not forgotten (Maness and Farrell, 2004). This is especially important when examining sustainability in the context of resource management and therefore agricultural development (Asimeh et al., 2020). For traditional production systems, such assessment provides not only benchmarks but also help reveal the impacts of policy measures adopted at a national level on a region or individual farm (Van Calker et al., 2005).
Sustainability assessments require a combination of methods and models to provide useful information on the impacts of past or proposed changes within systems (Thornton and Herrero, 2001). No single approach can satisfy the multiple purposes of sustainability assessment (Schader et al., 2014). Generally, the results of sustainability assessment are a starting point for discussion, reflection, and learning (De Olde et al., 2016). Within this context, our study seeks to: (a) examine the main agricultural and socio-demographic characteristics at the household level in four social groups (Waorani, Shuar, Kichwa, and migrant settlers) living in the YBR; (b) evaluate the sustainability dimensions of their traditional production systems; and (c) identify the existing interactions (synergies and trade-offs) between the sustainability dimensions of these traditional production systems.
The Traditional Agricultural Context of the Waorani, Shuar, Kichwa, and Migrant Settlers in the Yasuni Biosphere Reserve The food production system of the Waorani indigenous populations is based on reciprocity or individual production to share with their nuclear families (Rival, 2002), which is complemented by food gathering from the surrounding forests (Tirira et al., 2020). In their language (Wao-terëro), the generic name for the Waorani's traditional production system is the kinkore. According to Zurita-Benavides (2017), this ethnic group considers that the name of the traditional system depends on the main crop, which is usually accompanied with other crop species. For example, if cassava (Manihot esculenta Cranz) is the main crop, it is called kewenkore (a traditional system of cassava production), while if it is banana (Musa paradisiaca), the system is called penenkore (a traditional system of banana production). Likewise, the traditional system based on peanut (Arachis hypogaea) takes the name koromonkore, the one of corn (Zea mays), kaginkore, and the system with barbasco (Lonchocarpus utilis) (a product used for fishing) is called kompankore. These crop systems are generally oriented to subsistence production, with few produces destined for sale. These systems stand out for having a low impact on the landscape and do not use chemical fertilizers or agrochemicals for pest control . The size of the kinkore depends on the composition of the family group. Edible and medicinal plants are sown, as well as other species that allow for the extraction of natural dyes, seeds, and fiber for handicrafts (Zurita-Benavides, 2017). The kinkore are the places where knowledge is transferred to younger generations, commonly managed by Waorani women (Zurita-Benavides, 2017;Paniagua Blanc, 2019).
The Shuar people originate from the southern EAR and migrated to the northern part (Rubenstein, 2001), due to high population growth rate, land scarcity, and low agricultural production (Vasco et al., 2018). Their food supply is obtained through fishing, hunting, gathering forest products (Caballero-Serrano et al., 2017), and growing crops in a traditional system called aja. For the Shuar, the aja is a sacred space and is managed by women (González de Lema and Herrera, 2017). To establish an aja, first the man does the initial clearing of the forest (Karsten et al., 2000), then the woman has the main role of establishing (Pérez-Robalino et al., 2019) and maintaining the aja for life (Pohle and Gerique, 2008), since the aja for this culture constitutes the woman's identity (Carvajal and Shacay, 2004;González de Lema and Herrera, 2017). The transmission of knowledge about the ecology of planting and harvesting is only taught to the daughters (Cañadas and Sandoval, 2019). As the Shuar people have moved into the market economy, they have created an agricultural landscape that is more biodiverse than the mestizo (mixed-race) migrant settler system (Lu and Bilsborrow, 2011), although biodiversity is not as high as the primary forest (Vasco et al., 2017). Therefore, the agricultural practices carried out in an aja are considered compatible with the conservation of biodiversity (Rudel et al., 2002).
The traditional production systems of the migrant settlers are more market-oriented and, according to several authors, are less sustainable compared to traditional indigenous systems (Heredia-R et al., 2020a,b,c,d,e;Viteri-Salazar and Toledo, 2020). They include pastures, annual crops (rice, corn, and cassava), and perennial crops (coffee and cacao) to generate income (Mena, 2007;Monteros, 2011), resulting in intense deforestation (Pan and Bilsborrow, 2005;Sellers et al., 2017;Mullan et al., 2018). The dynamics of migrant settlers have forced neighboring indigenous communities to join the market economy and change some of their ancestral traditions of coexistence with the forest (Orta-Martínez and Finer, 2010). Northeastern Ecuador is one of the places where the cattle frontier in Ecuador is rapidly expanding (Sierra, 2000;Suárez et al., 2009), and is one of the main promoters of land-use change and deforestation in the EAR (Sierra, 2013). In this region, as of 2014, it was estimated that around one million hectares of pastureland, which represented 38% of the total agricultural area in the EAR (MAGAP, 2014), were dedicated to extensive cattle ranching (Lerner et al., 2014). Cattle production is for meat and milk to satisfy both the local demand and national markets (Knoke et al., 2014).

Geographical Location
This study was conducted in the Diversity and Life Zone (DLZ) in the Yasuní Biosphere Reserve (YBR). The YBR was established in 1989 by the Ecuadorian government and has a surface area of 1.68 million hectares. The core conservation area is the Yasuni National Park (YNP), which consists of 9.82 thousand hectares (Taco, 2001;Finer et al., 2009). The DLZ was created in 2007 with the objective of managing in a special way the production systems and livelihoods of indigenous households and migrant settlers established there . Through two presidential decrees, the national government declared (1999) and mapped out (2007) a "Forbidden Zone, " also called the Intangible Zone or Tagaeri-Taromenane Intangible Zone (ZITT for its Spanish acronym) (Presidencia de la República de Ecuador, 1999Pappalardo et al., 2013). The YBR is in the northern part of the Ecuadorian Amazon Region (RAE) and extends between the provinces of Orellana (50.51%), Pastaza (39.40%), Napo (8.64%), Sucumbíos (1.45%) and it borders two areas of natural and cultural importance: the Sumaco Biosphere Reserve and the Cuyabeno Wildlife Production Reserve (Figure 1).
The YBR is part of an important mosaic of ecosystems. The climate is within the "Amazonian core" (Killeen and Solórzano, 2008), which is characterized by warm temperatures (annual average of 24 −27 • C), high precipitation (annual average 3,200 mm), and high relative humidity (annual average 80-94%) (Pitman, 2000). The soils are geologically young, fluvial sediments coming from the erosion of the Andes (Pitman et al., 2002;Valencia et al., 2004). The area is part of the hotspot called the "Western Amazon Highlands" (Myers, 1988;Myers et al., 2000). However, like many other areas of high biodiversity, it is threatened by habitat destruction and cultural alterations (Mittermeier et al., 1998;Beristain et al., 2009). The YBR faces high rates of deforestation and land-use change, with one of the main drivers being the opening of roads because of oil concessions (Jorgenson and Coppolillo, 2001;Mena et al., 2006;De la Torre and Cornejo, 2008;Suárez et al., 2009;Bonilla-Bedoya et al., 2018). New roads have facilitated colonization and subsequent deforestation by small-scale migrant farmers engaged in agriculture and cattle ranching (Wunder, 2003;Sierra, 2004). Indigenous peoples also migrated into the Diversity and Life Zone (DLZ) for cash-crop agriculture by indigenous peoples (Sierra, 2004;Lu et al., 2012;Davis et al., 2016).

Sample Design and Data Collection
The data come from a household survey conducted between August and November 2019 in the DLZ. A template of the Poverty and Environment Network (PEN) questionnaire (Cavendish, 2003) was adopted to obtain information on the demographic characteristics of the households, the population structure by social group, and their distributions by sex and age. The following indices were calculated: (1) proportions of young population (<14 years, P youngpop ); (2) adult population (between 15 and 64 years, P adults ); (3) ratio of children/women, defined as: the number of children under 5 years of age for every woman of reproductive age (R); (4) masculinity rate, consisting of the ratio of males for every 100 females in a given population, which is the first indicator to analyze the population's distribution by sex (MR); (5) the youth dependency index-the ratio between those (<15 years) who are potentially dependent and those of a potentially active age (between 15 and 64 years, I t dy ); (6) the working age population structure index-the ratio between those aged 40-64 years and those aged 15-39 years (I r ); and (7) the working age population replacement index-the ratio between those aged 60-64 years and those aged 15-19 years (I t r ); (Hinde, 2001;Rowland, 2004;Holdsworth et al., 2013;Wilson, 2016). Moreover, we determined the time needed to travel to: (a) the road and river, (b) traditional production systems for indigenous people, and (c) production systems for migrant settlers.
(1) x 100 Frontiers in Sustainable Food Systems | www.frontiersin.org The cluster sampling technique was used with a two-stage approach. During the first stage, two to five communities were selected per social groups, based on criteria of ethnicity and ownership of traditional production systems. To this end, 14 communities were identified ( Table 1). The variation in conditions between communities ensured a fair representation of the communities' diversity, which influences the robustness of the study (Cavendish, 2003). During the second stage, with the support of a key informant per social group, 10 households per social group were randomly selected for a total of 40 households (10 households per social group) for the interviews. The survey was translated from English to Spanish, Kichwa, Shuar, and Wao-terëro. The 40 households (10 households per social group) housed between 3 and 13 members per household ( Table 1). The survey was administered to the husband and wife of each household. Although the sample is not strictly called a probability sampling of communities and households, the 14 communities and 40 households provided a good representation of diverse ethnic or social groupings, geographies, and livelihood diversity, considering the complexity of conducting interviews in indigenous communities (Vasco et al., 2018). The process of obtaining free and informed consent was carried out with the support of indigenous leaders and settlers, through whom all approaches to households were made. Surveys were conducted according to the principles of ethical research (Vanclay et al., 2013) where the objectives, methodology, and timetable of the study were a priori explained to members of each community.

Assessment of Agricultural Sustainability
The program SAFA version 3.0 (SAFA, 2012), developed in 2012 by the FAO, was used to obtain a set of useful indicators that highlighted problems and identified solutions regarding the sustainability evaluation of the production systems, used in this case at the household level. The surveys consisted of questions based on the SAFA indicators (Food Agriculture Organization, 2014) related to agricultural systems and livelihoods. The SAFA has hierarchical levels: dimensions, themes, sub-themes, and indicators (Food Agriculture Organization, 2013). The dimensions include good governance (GG), environmental integrity (EI), economic resilience (ER), and social well-being (WB). One dimension is the pillar of sustainability, which is the highest and most general level in the SAFA structure (Food Agriculture Organization, 2014).
The SAFA usually consists of 21 themes, 58 sub-themes and 116 indicators (Food Agriculture Organization, 2013. In this research, however, the theme product quality and information (with its 3 sub-themes and 7 indicators) was not considered since none of the products produced and sold was labeled or had any traceability standards. Therefore, to assess the four dimensions of sustainability, the 20 SAFA sustainability themes, 55 sub-themes and 109 indicators ( Table 2) were targeted. They were measured on a scale of 1-5 (Food Agriculture Organization, 2013) with the following sustainability levels: insufficient (red, <1.0), moderate (orange, 1.1-2), fair (yellow, 2.1-3), superior (light green, 3.1-4) and best (dark green, >4.1) sustainability practices (Food Agriculture Organization, 2014).
The development of the SAFA included four stages: (1) mapping: description of the system to be evaluated; (2) contextualization: review of sub-themes and dynamic contexts; (3) indicators: selection of indicators and individual assessment; and (4) reporting of the information obtained at each stage, and presented in a sustainability polygon, using the SAFA version 3.0 and Excel version 2110 programs (Food Agriculture Organization, 2014). The SAFA program lists the metric tools and standards for data collection, which determine the level of data quality by attributing a variable score from 1 to 3, where 1 corresponds to low-quality data; 2 to moderate quality data; and 3 to high-quality data (Food Agriculture Organization, 2013). The survey had a duration of 60-85 min per household.
For the definition of the evaluated topics (Annex 1 in the Supplementary Material), three field visits were made to the study area between the months of May and June 2019. The direct observation technique was used because it helped data collection and information, allowed for a more detailed analysis of the facts and realities that made up each case within the study (Campos and Martínez, 2012) and was the best method for an in-situ evaluation (Sutton and Austin, 2015).

Statistical Method for Identifying Interactions Within the Dimensions of Sustainability
An analysis of variance (ANOVA) was used to compare the traditional production systems among the four social groups. A heteroscedastic model was fitted in the case of heterogeneous variances between social groupings. When normality was not met, we utilized the non-parametric Kruskal-Wallis test (Kruskal and Wallis, 1952). For significant differences at 5%, a Fisher's multiple comparison test was performed. For nominal variables (0, 1), Chi-square distribution tests were used.
A linear discriminant analysis was performed with the variables (sustainability issues) and social groupings with the aim of adjusting discriminant axes-a linear combination of the original variables-to maximize the differences between populations. The results were visualized in a biplot (variables and social groupings) in which the projection vector of the variables on the axes allows one to identify the relative contribution of the variables to the separation of the populations. Finally, synergies and trade-offs were evaluated using the non-parametric Spearman's correlation test (Spearman, 1904). Positive correlation coefficients (0 <r s ≤ +1) represent synergies, while negative correlation coefficients (0> r s ≥ −1) represent trade-offs. Analyses were performed with IBM SPSS Statistics version 21 (SPSS, 2012) and the statistical program Infostat (Di Rienzo et al., 2011).

RESULTS
This section details the sociodemographic characteristics and sustainability scores for traditional production systems; certain  significant differences are evidenced in demographic terms and a narrowing between the sustainability dynamics evaluated.

Sociodemographic and Agricultural Characteristics
The population for the entire study area was composed of men 55.7% and women 44.3% (262 total), and the average age of both men and women was 27 years old. Household composition did not significantly differ among the social groups. Among the Waorani, the number of members per household varied between 4 and 13 people, while the Kichwa household had the lowest number of individuals i.e., 3 people. In the Shuar, Kichwa, and migrant settler households, the maximum number of individuals was 9 ( Table 3).
The age of male heads of household differed significantly between the Shuar groups at 29 years and the other three social groups (average of 49). In terms of education, there are no significant differences among the Shuar, Kichwa, and migrant settler heads of household. However, four of Waorani heads of household no formal education, bring the level of education significantly lower at 60% than the three other groups.
The results showed no significant to marginal differences among the four groups in terms of aid for sustainable development from national or international organizations ( Table 3). The Shuar and Kichwa tended to have received more help than the Waorani and migrant settlers. Access to financial credit appeared to be limited to all the four groups with only 10% of migrant settlers having some access. Access to mobile phones did not vary among the four groups but, migrant settlers tended to have highest number of mobile phones (80%), while the Kichwa the lowest. The distance from the household (in minutes) to the main road or to the river port significantly varied, with the longest time to reach the river port (72 min) was for the Kichwa group (specifically the Mandaripanga community). The Waorani households had the lowest times, with some only having a 2 min distance to reach the main road. For the Shuar and migrant settlers, the average time varied between 12 and 19 min.
Demographic indices, such as percentages of young and adult in the populations, were not significantly different among the four social groups (Table 4). Nonetheless, the Shuar appeared to have the lowest young population, comprised of 18.9% men and 8.1% women, and the highest adult population, made up of 37.8% men and 35.2% women. As expected, this led to having the lowest rate of youth dependency and the highest percent of working-age people, meaning that their replacement rate for future work was the lowest. The other three groups tended to have similar trends with more youth and less working-age population.
The average agricultural land size per household varied from 1.75 ha for the Shuar group to 3.9 ha for the migrant stellers, and these results were significant ( Table 5). The Shuar had the largest forested area followed by Waorani and Kichwa and finally the migrant settlers, which was significantly the lowest in surface area. While the number of staple crops remained similar for the four social groups, the Kichwa and migrant settlers tended to cultivate at least one more cash crop than the Waorani and Shuar (Table 5). While the Kichwa tended to be closer to their  agricultural system, the results were not significant due to high variation in distances for both migrant settlers and Waorani. The distance to the forest, however, significant differed among groups with the Shuar being closer to the firested areas than the other groups.

Sustainability Dimensions Assessment
The four dimensions of sustainability were analyzed using the 55 from the 58 SAFA sub-themes for the four traditional production systems. We found that 20 of the 55 (36.4%) of the evaluated subthemes had significant differences among the four social groups ( Table 6) and the sustainability polygons displayed in Figure 2.

Good Governance Dimension
The mean performance scores for the social groups were as follows: Waorani (1.92), followed by the Shuar group (2.59), the Migrant settler group (2.65), and the highest mean for the Kichwa (2.71). Of the 14 sub-themes evaluated, six showed significant differences among the four social groups. Mission statement, and legitimacy resource appropriation subthemes were always significantly lower for the Waorani group than the others. It was similar for the subthemes, stakeholder dialogue, and sustainability management plan. However, the Migrant settler group was not significantly different from the Waorani. In terms of overall scores, full cost accounting was the lowest followed by sustainability management plan, grievance procedures, and conflict resolutions.

Environmental Integrity Dimension
The average performance scores for the social groups were as follows: Kichwa (2.69), migrant settlers (2.78), Waorani (2.95), and the highest mean for the Shuar (3.33). Of the 14 sub-themes evaluated, four showed significant differences among the four social groups. Greenhouse gases were similar for the Waorani and Shuar but different for the Kichwa and migrant settlers, while ecosystem diversity, animal health and freedom from stress were significantly lower for the Waorani group than the others. In terms of overall scores, freedom from stress was the lowest, followed by animal health, air quality, and material use.

Economic Resilience Dimension
The mean performance scores for the social groups were as follows: the lowest mean for the Waorani (1.40), Kichwa (1.48), migrant settlers (2.48), and Shuar (2.50). Of the 11 sub-themes evaluated, six showed significant differences among the four social groups. The internal investment, community investment, long-ranging investment, and profitability were significantly lower for the Waorani and Kichwa groups while, the stability of market for the migrant settlers was slightly higher for the others and local procurement for the Kichwa was slightly lower than the other groups. In terms of overall scores, the stability of market was the lowest, followed by the stability of supply, liquidity, and risk management.

Social Welfare Dimension
The average performance scores for the social groups were as follows: the lowest average for the Shuar (2.41), Kichwa (2.44), Waorani (2.54), and the highest average for the Migrant settlers (2.47). Of the 16 sub-themes evaluated, six showed significant differences among the four social groups: the quality of life for the Waorani group; the indigenous knowledge and food sovereignty for the Shuar; the gender equality, support for vulnerable people and public health for the Kichwa were significantly lower than the others. The public health was significantly higher in Kichwa than the others; the indigenous knowledge is the main difference between the Shuar and Kichwa groups; the food sovereignty in the Kichwa group is significantly different from the others. In terms of overall scores, responsible buyers were the lowest, followed by the rights of suppliers, child labor, forced labor, freedom of association, and rights to bargaining.

Interactions Within Sustainability Dimensions and Themes
Before evaluating the sustainability interactions within the four dimensions and 20 themes according to SAFA, it was necessary to quantify their results ( Table 7). The results showed that the sustainability indices had a performance level of "moderate." In general, the traditional aja system (2.84) had the best score, followed by the migrant settlers' agroforestry system (2.65), considering that the themes with lower performance scores are: fair trading practices, labor rights, animal welfare and holistic management. They were followed by the Kichwa's chakra system, influenced by its lowest scoring themes: fair trading practices, investment, vulnerability, and labor rights. In last place was the Waorani's kewenkore whose lowest scores were in the themes: fair trading practices, investment, animal welfare, and holistic management. From the scores of the four dimensions and the 20 sustainability themes (Table 7), interactions within the sustainability themes were analyzed using a Spearman's correlation test. In general terms, trade-offs and synergies were found (50%) ( Table 8). There were trade-offs between good governance and environment integrity (15%), followed by social issues (11%), and synergies (25%) with respect to economic resilience. In addition, significant synergies were observed within environmental integrity with economic and social issues, and trade-offs between economic resilience and social well-being.

Multivariate Discriminant Analysis by Social Groups
Similarities and antagonisms in terms of sustainability among the social groups were identified with a multivariate discriminant analysis, using the averages of the 116 sustainability indicators of the four dimensions of the SAFA methodology (Figure 3). The results showed that in the horizontal component (60% of the variation between groups), the social groups were ordered by placing the traditional systems of the Shuar population on the extreme right together with those of migrant settlers because they showed similarities. On the far left were the Waorani and Kichwa cases, which differed in certain aspects to the two aforementioned groups. These differences between the social groups were mainly explained by a contrast or negative correlation between the investment, atmosphere, and materials themes vs. the participation. In this sense, the Shuar and migrant settler populations were characterized by high levels of investment, atmosphere, and materials (although also in biodiversity, animal

Assessment of Sustainability Dimensions
When comparing the four dimensions of sustainability between the traditional production systems (Waorani, Shuar, Kichwa, and migrant settlers), we found that the sustainability profiles by sub-themes are more statistically different in the dimensions of good governance and social well-being in contrast to the dimensions of environmental integrity and economic resilience ( Table 7). This is consistent with the dynamics of management in traditional production systems among the Waorani (Zurita-Benavides, 2017), Shuar (Karsten et al., 2000), Kichwa (Coq-Huelva et al., 2017;Luzuriaga-Quichimbo et al., 2019), and migrant settlers (Sellers et al., 2017;Heredia-R et al., 2020a,b,c;Viteri-Salazar and Toledo, 2020). It is related to different cultural and migratory dynamics (Barbieri et al., 2009), such as, mining and land redistribution in favor of settlers. Some Shuar populations had to migrate to the province of Orellana and other provinces of the Ecuadorian Amazon (Holt et al., 2004), when the migrant settlers arrived mainly from the highlands and coast (Barbieri and Carr, 2005;Wasserstrom, 2010;Viteri-Salazar and Toledo, 2020). None of the traditional production systems works uniformly for all themes (Table 3) and sub-themes (Figure 2) of sustainability. This may be related to social or environmental conflicts (Scheidel et al., 2020) and market access, as has been demonstrated in other indigenous populations, for instance the Baka of Cameroon (Africa), the Punan Tubu of Indonesia (Asia), and the Tsimane' of Bolivia (South America), where three relatively isolated societies, largely dependent on traditional subsistence systems (Joiris, 2003;Levang et al., 2007;Leclerc, 2012;Reyes-García et al., 2014;Díaz-Reviriego et al., 2016), saw their diets alter when living closer to markets. This shifted their consumption habits from nutritional food (i.e., fruit, vegetables, and food of animal origin) to an increased consumption of fats and sweet products due to market integration, thus decreasing work in traditional production systems (Reyes-García et al., 2019).
The low scores of the Waorani (kinkore) traditional production systems for good governance ( Table 7) are influenced by the holistic management theme and correlated with the sub-themes of sustainability management plan and full cost accounting (Figure 2 and Table 6). This contrasts with the common belief that indigenous populations are associated with a holistic management of their traditional systems (long fallows, according to Boserup, 1965), with collective responsibility, participation and, therefore, compatibility with resource conservation (Dufour, 1990;Schwartzman et al., 2000). However, several studies (Henrich, 1997;Zimmerman et al., 2001;Rudel et al., 2002;Godoy et al., 2005;Gray et al., 2008;Porro et al., 2014) indicate that indigenous peoples sometimes also engage in unsustainable practices (Food Agriculture Organization, 2005), driven by the presence of oil activity (Rivera-Parra et al., 2020), which has negative implications (Diantini et al., 2020). The Waorani stand in contrast to the Shuar's aja, which had the highest scores, thus complying with Boserup's assertion (Boserup, 1965) regarding systems with more sustainable levels of governance (Rudel et al., 2002).
In the environmental integrity dimension, the low scores obtained by the migrant settlers are related to the themes of animal welfare (driven by the sub-themes of animal health and freedom from stress) and atmosphere (driven by the sub-themes of greenhouse gases and air quality). An explanation for this finding is that pastoral systems in the EAR are associated with extensive operations (Lerner et al., 2014;MAGAP, 2014). Therefore, they require larger areas FIGURE 2 | Comparison of the dimensions of good governance, environmental integrity, economic resilience, and social well-being in the four traditional production systems in the Yasuní Biosphere Reserve, Amazon region of Ecuador. of land. In tropical regions, livestock development is being promoted with an economically, socially, and environmentally sustainable approach (Ness et al., 2007). Cattle ranching offers an alternative source of capital that the rural poor can accumulate as a "savings account" to protect themselves against income fluctuations (Broom et al., 2013;Lemaire et al., 2014). Therefore, raising cattle in rural households is considered an alternative form of insurance, allowing them to earn income from the sale of animals in times of crisis (Hoddinott, 2006;Mogues, 2011). The welfare of ethnic minority households with livestock production tends to be higher than those without livestock production (Truong et al., 2020), and sustainability depends on the availability of resources for transformation to agrosilvopastoral systems (Lopera-Marín et al., 2020).
In relation to economic resilience, the low scores of the Waorani (kinkore) traditional production systems are related to the theme of investment and the sub-themes internal investment, community investment, long-term investment, and profitability. First, the Waorani have recently become integrated into the markets (Sierra et al., 1999;Franzen and James, 2007;Lu, 2007), driven by certain factors: accessibility, monetary income, trade of agricultural products, and wild meat, and a cultural erosion of exchange (Donders and Barriocanal, 2020). Oil exploitation in the Amazon basin is an additional factor, whereby some Waorani are economically dependent on salaried jobs in the oil industry (Rivera-Parra et al., 2020), which has caused time constraints for interaction with traditional production systems (Diantini et al., 2020). The following factors have also been registered: water contamination (Balseiro-Romero et al., 2018;Maurice et al., 2019), the presence of heavy metals in the food chain (Adzigbli and Yuewen, 2018;Maletić et al., 2019), including crops (Veil et al., 2004;Barraza et al., 2018) health risks for indigenous people and migrant settlers (Barraza et al., 2017;Merchán-Rivera and Chiogna, 2017), a loss of biodiversity (Durango-Cordero et al., 2018) and acculturation processes (Swing et al., 2012). The social and environmental history of the oil industries and stateowned companies in the Amazon basin  have not been detrimental for the indigenous and migrant settler populations (Valdivia, 2007;Swing et al., 2012;Bebbington and Bury, 2013) due to the resulting economic dependence (Rival, 1992). This is similar to the scenario in Central Africa, where the Chad-Cameroon pipeline project did not improve the living conditions of the local indigenous community of Bagyeli, poverty increased and the quality of life decreased (Horta, 2007;Husar and Finka, 2016).
Although the production systems evaluated had high scores in most of the social issues, the lowest scores were in the Shuar's aja, mostly determined by fair trading practices and the sub-themes of responsible buyers, rights of suppliers, employee relations, forced labor, and child labor. This is not related to indigenous cultural dynamics, because the management of production systems, from a social perspective, is the result of a set of values, deeply rooted in the indigenous worldview (Goyes et al., 2021;Santafe-Troncoso and Loring, 2021;Varese, 2021), and ancestral knowledge is transmitted from generation to generation through harmonious relationships, based on solidarity and reciprocity (Sierra, 2000). These dynamics can also be identified in traditional production systems along the Napo riverbank, located to the north of the YBR (Heredia-R et al., 2020d,e). It is important to highlight that the contributions of indigenous and local knowledge in research are increasingly being considered in the science of sustainability (Mistry and Berardi, 2016;Tengö et al., 2017).
The discriminant analysis showed attributes of similarity (sustainability issues) between migrant settlers and the Shuar, corroborating the results of other researchers, who suggest that the Shuar are increasingly integrating into the market economy, participating in extensive cattle raising and commercial agriculture, and have similarly to migrant settlers (Rubenstein, 2001;Zimmerman et al., 2001;Rudel et al., 2002;Godoy et al., 2005;Gray et al., 2008). Meanwhile, greater divergence is evident among the Waorani and Kichwa in terms of sustainability, probably influenced by the following factors: (1) Indigenous people adopt technologies for labor in traditional production systems (Sellers and Bilsborrow, 2020); (2) Oil activity (Rivera-Parra et al., 2020) has intensified, and this potentially represents a major transformation of their social, economic and environmental context (Codato et al., 2019), and to compound these changes, oil companies may offer access to employment, cash payments or health, and transportation services to indigenous communities in order to facilitate their work and/or comply with legal or internal "corporate social responsibility" mandates (O'Faircheallaigh, 2013;Billo, 2015); (3) The agricultural and extractive (hunting/fishing) frontier Heredia-R and Hernández, 2019;Houssou et al., 2019) has advanced due to unsustainable agricultural activities (Heredia-R et al., 2020a,b,c), caused by a dependence on salaried jobs in the oil sector, according to the literature on the Dutch disease (Bozigar et al., 2016;Arsel et al., 2019;Erdogan et al., 2020), correlated with the majority of the EAR population being impoverished (Lang, 2017;Torres et al., 2017a), due to its externalities (Vasco et al., 2015). In addition, inequality in terms of sustainability in the groups evaluated is evidenced by the existence of new forms of land use and management, a greater use of land for planting cocoa, African palm species, and pastureland for cattle. These activities are completely linked to the global market, and also imply an increased use of chemicals and fertilizers, given the low quality of Amazonian soils and the quantity of pests. Therefore, modifications are occurring and even more intensely due to the presence of other actors, such as extractive companies and the State (Vasco et al., 2015;Salinas Castro et al., 2020).

Synergies and Trade-Offs Within Sustainability Dimensions
To be sustainable, a system must be robust, sufficiently productive, and use resources efficiently to achieve its equilibrium. However, there are often synergies or tradeoffs between the various goals of sustainability and their related objectives toward which sustainability is directed (NAS, 2010). Greater synergies are observed between good governance dimension and economic resilience, as well as between environmental integrity and social well-being, which emphasizes the importance of good governance in influencing the level of performance in the other dimensions of sustainability (Schader et al., 2016) demonstrating the importance of the dimensions of sustainability in rural contexts and their influence on the strategic decision-making of indigenous populations and migrant settlers for the management of traditional production systems (Coteur et al., 2020). This implies that a specific focus on good governance can improve sustainability performance in traditional production systems through synergistic interactions with other sustainability dimensions. However, several studies show that trade-offs and synergistic relationships are dynamic and can change over time (Haase et al., 2012;Li et al., 2019), and vary across a wide range of conditions (Walker and Salt, 2012). Therefore, interactions that enhance sustainability should be prioritized according to the context and state of traditional production systems, while policies can facilitate the transition to more sustainable production systems.

Policy Implications to Safeguard Traditional Agricultural Systems
It is evident that the differences in terms of agricultural sustainability (Figure 2) between indigenous populations and migrant settlers (Figure 3) are narrowing (Rubenstein, 2001;Sellers and Bilsborrow, 2020). The existence of identified synergies and trade-offs (Table 8) provide policymakers with specific information to design sustainable development policies and rescue traditional agricultural systems in the DLZ.
Considering the global importance of the YBR and the Amazon basin in terms of conserving its high diversity, the findings of this research contribute to the management model developed for the communities that make up the DLZ in 2015. They were carried out by several national and local public organizations (MAGAP, 2016), who aimed to (a) strengthen governance in the DLZ; (b) achieve a balance between environmental conservation and living conditions; and (c) salvage cultural dynamics. The management model is implemented through a special regime that guarantees sustainability and minimizes risk factors for indigenous communities and migrant settlers in the DLZ (MAGAP, 2016).
At the national level, the Ecuadorian government has created different territorial management instruments that seek to restore degraded and deforested areas. Traditional agroforestry systems, such as the chakra, aja, kinkore, and agroforestry system, can contribute to the implementation of the Agenda for Productive Transformation in the Amazon (APTA) concerning its objective of bringing about a move toward more sustainable production systems (MAGAP, 2014). In this respect, it should also encourage the rescue of traditional systems that have been in use for millennia by the indigenous groups.
In terms of future research, based on our findings, we recommend studying the relationship between rural household livelihoods (livelihood strategies)  and forest ecosystems (forest governance) (Torres et al., 2013), as well as investigating biotrade alternatives in the DLZ-for livelihood strategies (Torres et al., 2017b;Boeri et al., 2020), climate change mitigation and the conservation of flora and fauna diversity-as a tool for biological corridors.

CONCLUSIONS
The differences in terms of sustainability between the traditional production systems evaluated are influenced by the sociodemographic conditions and characteristics, access routes, and extractive activities in the Amazon basin. This confirms the theory that differences between indigenous people and mixedrace settlers are narrowing, since there is greater similarity between the social, environmental, economic, and governing dynamics between the Shuar and migrant settler populations, while the sustainability weightings are farther apart between the Waorani and Kichwa populations.
Synergies exist between the dimensions of good governance and economic resilience among all groups assessed, following the finding of Schader et al. (2016), who emphasized the importance of good governance in influencing performance levels for other dimensions of sustainability. This implies that a specific focus on the good governance dimension can improve the sustainability performance of traditional production systems through its synergistic interactions with the other dimensions of sustainability. Thus, decision-makers at the local level and future research projects can intervene in the two trade-offs found between good governance and environment integrity, as well as good governance and social well-being, to contribute toward United Nations Sustainable Development Goals. The aim would be to optimize resources and time in the execution of plans, programs, and projects with the populations of the Amazon basin and at the same time verify if the same theory by Schader et al. (2016) works with the trade-off found between economic resilience and social well-being.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Ethics Committee, Universidad Estatal Amazònica. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Ethics clearance was also not required at Brock University due to use of secondary data (April 2021).

AUTHOR CONTRIBUTIONS
MH-R and DB performed the fieldwork. MH-R, BT, LP, and LV performed the statistical analyses. MH-R, BT, LV, and CD-A wrote the first draft of the manuscript. All authors contributed to the conception and design of the study, revision of the manuscript, and reading and approval of the submitted version of the manuscript.

FUNDING
The fieldwork of this research was carried out with small funds from the project: Analysis of sustainability and climate change in indigenous farmers and settlers of the Yasuní Biosphere Reserve, Ecuadorian Amazon of the Universidad Estatal Amazonica (Ecuador).