Abstracts

Introduction

Analytical approaches to health-disease processes in the Brazilian Amazon should offer a framework that integrates the triad of health, environment, and economy. The agrarian economy ¹1. Brossier JB, Landais E. Systems studies in agriculture and rural development. Paris: NRA Editions; 1994., with its diverse agents and technical systems, is the main factor that mobilizes the elements that drive transformations in the social and natural landscapes of the Amazon. Such transformations create anthropogenic disturbances in the forest ecosystem and affect its structural and functional integrity ²2. Rodrigues ET. Paisagens florestais associadas à sistemas tecno-produtivos rurais na Amazônia Brasileira [Master's Thesis]. São José dos Campos: Instituto Nacional de Pesquisas Espaciais; 2024., changing land use and cover patterns, the role of socio-spatial units, and their arrangement in the contemporary Amazonian urban environment ³3. Cardoso ACD, dal'Asta AP, Monteiro AMV. O que é o urbano na Amazônia contemporânea? Implicações para a vigilância em saúde no bioma. Cad Saúde Pública 2023; 39:e00129723.. They also change the distribution of income, goods, and services associated with the chains involved in regional and local economies. In this perspective, these technical systems of production are directly related to the choice of a regional development path that can lead to the sustainability or unsustainability of the forest and social landscape, interconnected to the history of occupation of the Brazilian Amazon ⁴4. Costa FA. A brief economic history of the Amazon (1720-1970). Cambridge: Cambridge Scholars Publishing; 2019.^,55. Escada MIS, Amaral S, Fernandes DA. Dinâmicas de ocupação e as transformações das paisagens na Amazônia, Brasil. Cad Saúde Pública 2023; 39:e00021723.. In this context, the health dimension is scarcely addressed in these debates. In order to promote this analytical structure, a new territorial representation is necessary for the landscapes of the biome, organized as production landscape units (PLU), as we will argue throughout this article. The PLU are units associated with the modes of production operated by the rural techno-productive systems, present in the agrarian economy.

The PLU is conceived as a representation for the state of forest landscapes modified by technical systems associated with rural production. In the Amazon context, the PLU encompass a variety of ecosystems and and environments that are interspersed and overlap as a result of the interactions between production agents, such as farmers, extractivists, cattle ranchers and local communities, and the components of the landscape, such as rivers, forests, roads and cities. These elements are organized in varied and non-uniform spatial arrangements, reflecting the diversity and complexity of social interactions, economic and environmental interactions ⁶6. Dal'Asta AP, Lana RM, Amaral S, Codeço CT, Monteiro AMV. The urban gradient in malaria-endemic municipalities in Acre: revisiting the role of locality. Int J Environ Res Public Health 2018; 15:1254.^,77. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45.^,88. Souza AR. Os sistemas do açaí sob a perspectiva das trajetórias tecno-produtivas rurais e sua expressão na paisagem amazônica: elementos para análise da economia do açaí na região do Baixo Tocantins - Pará [Doctoral Dissertation]. São José dos Campos: Instituto Nacional de Pesquisas Espaciais; 2024.. Once identified and mapped, the PLU can be used as a proxy, that is, as a surrogate indicator variable, for the types of technical systems associated with rural production in the area under analysis. The rural techno-productive trajectories or technological trajectories (TTs) are analytical categories introduced by Costa ⁹9. Costa FA. Trajetórias tecnológicas como objeto de política de conhecimento para a Amazônia: uma metodologia de delineamento. Rev Bras Inov 2009; 8:35-86.^,1010. Costa FA. Structural diversity and change in rural Amazonia: a comparative assessment of the technological trajectories based on agricultural censuses (1995, 2006 and 2017). Nova Economia. 2021; 31:415-53. for the analysis of agrarian economics. They represent patterns of technical and productive solutions that are constituted as specific trajectories, adopted by various rural (either peasant or capitalist) production units.

The patterns observed in the modified landscapes are influenced by the ways of living and modes of production, associated with the choices of production models of peasant and capitalist agents. The analysis of specific spatial patterns, observed in modified landscapes, makes it possible to associate them with the corresponding PLU and TTs. As the PLU represent landscapes constantly shaped by the technical systems of production, these spatial patterns − typically present in the living spaces of these agents − can be identified and measured via satellite images.

Codeço et al. ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754. highlight the importance of integrated approaches considering economy, environment, and health-disease processes to (re)formulate health surveillance strategies. The article ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754. offers a framework to discuss the economic, environmental, and health dimensions at the municipal level.

The economic dimension is analyzed based on the analytical category defined by Costa ⁴4. Costa FA. A brief economic history of the Amazon (1720-1970). Cambridge: Cambridge Scholars Publishing; 2019.^,99. Costa FA. Trajetórias tecnológicas como objeto de política de conhecimento para a Amazônia: uma metodologia de delineamento. Rev Bras Inov 2009; 8:35-86. as a descriptor of the agrarian economy in the Amazon, the TTs. Costa ¹⁰10. Costa FA. Structural diversity and change in rural Amazonia: a comparative assessment of the technological trajectories based on agricultural censuses (1995, 2006 and 2017). Nova Economia. 2021; 31:415-53., using data from the agricultural censuses (1995, 2006, and 2017) in a methodology based on statistical regressions, principal components, and factor analysis, shows the presence of six types of TTs in the municipalities of the Brazilian Amazon. Codeço et al. ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754. associates the dominant TT in the municipal agrarian economy with environmental and epidemiological indicators from a database organized and harmonized at the municipal level ¹²12. Rorato AC, Dal'Asta AP, Lana RM, Santos RBN, Escada MIS, Vogt CM, et al. Trajetorias: a dataset of environmental, epidemiological, and economic indicators for the Brazilian Amazon. Sci Data 2023; 10:65..

The methodology for determining the TTs rely on the data of the agricultural census, conducted every 10 years and at the municipal household level. However, the biome undergoes important changes, observed by official monitoring systems: the Real-Time Deforestation Detection System (DETER, acronym in Portuguese) and Amazon Deforestation Monitoring Program (PRODES, acronym in Portuguese), monthly and yearly, respectively ¹³13. Coordenação Geral de Observação da Terra, Instituto Nacional de Pesquisas Espaciais. Monitoramento do Desmatamento da Floresta Amazônica Brasileira por Satélite - PRODES. http://www.obt.inpe.br/OBT/assuntos/programas/amazonia/prodes (accessed on 04/Mar/2024).
http://www.obt.inpe.br/OBT/assuntos/prog... . Changes in land use and cover in deforested areas are observed by the TerraClass system (https://www.terraclass.gov.br/) ¹⁴14. Instituto Nacional de Pesquisas Espaciais. Projeto TerraClass 2012 - mapeamento do uso e cobertura da terra na Amazônia Legal. http://www.inpe.br/cra/projetos_pesquisas/dados_terraclass.php (accessed on Jan/2016).
http://www.inpe.br/cra/projetos_pesquisa... . These data derived from images have spatial resolutions of 30m in PRODES, 10m in TerraClass 2020, and 30m in the other years of TerraClass. These elements are considered to estimate the minimum spatial unit of mapping − 2.5 hectares (ha) in these two systems − which can be aggregated in municipal totalizations, for example ¹⁵15. Almeida CA, Coutinho AC, Esquerdo JCDM, Adami M, Venturieri A, Diniz CG, et al. High spatial resolution land use and land cover mapping of the Brazilian Legal Amazon in 2008 using Landsat-5 / TM and MODIS data. Acta Amaz 2016; 46:291-302..

On the other hand, data from health information systems, especially from surveillance and monitoring systems such as the Brazilian Information System for Notificable Diseases (SINAN, acronym in Portuguese) and the Malaria − Epidemiological Surveillance Information System (Sivep-Malária, acronym in Portuguese), among others, were structured to support the design and evaluation of local, municipal, state, and national policies and programs in public health, as well as scientific studies ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754.^,1616. Castro MC, Baeza A, Codeço CT, Cucunubá ZM, Dal'Asta AP, De Leo G, et al. Development, environmental degradation, and disease spread in the Brazilian Amazon. PLoS Biol 2019; 17:e3000526.^,1717. Braz RM, Barcellos C. Analysis of the process of malaria transmission elimination with a spatial approach to incidence variation in the brazilian Amazon, 2016. Epidemiol Serv Saúde 2018; 27:e2017253.. The minimum spatial unit of these systems is the “case”, centered on a specific address. To integrate health databases with environmental databases, the “case” can be aggregated to the municipal level, or to other levels, according to the analysis and specific periods. This integration is important because Amazonian municipalities present large territorial extensions and hold a diversity of socioenvironmental settings that have different possibilities for the analysis of the evolution of health-disease processes. To promote integrated health-environment-economy analyses in the Amazon, one must evaluate the state of TTs at least every year or every two years, in a more disaggregated spatial unit, appropriately located between the municipality and the “case”.

The technical systems associated with TTs transform the forest landscape and leave marks that can be captured on satellite images and be associated with other information from primary and secondary data ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45.. In this perspective, the representation of landscapes, associated with these technical systems and their TTs, constitute a synthesis territorial representation of the landscapes of production present in the regional agrarian economy. These mapped landscapes constitute a representation of the territories of production, which are territories of life and experiences of the agents involved in this economy.

The territories of production can be represented as a set of landscape units that capture the characteristics of technical systems associated with rural production and constitute part of the TTs. These landscape units can be identified and mapped with new approaches to classification of satellite-derived products, exploring machine learning with supervised classification and decision tree-based methods ¹⁸18. Quinlan JR. C5.0: an informal tutorial. https://www.rulequest.com/see5-unix.html#BOOSTING (accessed on 14/Aug/2021).
https://www.rulequest.com/see5-unix.html... . We call these classified units as PLU. The PLU are the integrating territorial element, the structure this article proposes for integrated studies of health-disease processes considering the triad of health, environment, and economy in the Amazon.

To show the feasibility of the PLU identification and mapping methodology, we conducted a case study in the municipalities of Mocajuba and Cametá, located in the Baixo Tocantins region, in the State of Pará, Brasil, for 2021. We describe in this article the proposed methods to identify and map the PLU located in the municipalities, in a spatial unit of intra-municipal reference, and how to associate them with the types of TTs present in the agrarian economy of the Amazon. Thus, we seek to promote a debate on the use of PLU in the structuring of integrated approaches in health, associated with agrarian territories of production on an intra-municipal basis.

Materials and methods

TTs and the agrarian economy in the Amazon

The economist Francisco de Assis Costa introduced a new economic category for regional analysis: the TTs. According to the author ⁹9. Costa FA. Trajetórias tecnológicas como objeto de política de conhecimento para a Amazônia: uma metodologia de delineamento. Rev Bras Inov 2009; 8:35-86.^,1010. Costa FA. Structural diversity and change in rural Amazonia: a comparative assessment of the technological trajectories based on agricultural censuses (1995, 2006 and 2017). Nova Economia. 2021; 31:415-53., the technical solutions adopted at the level of production units are empirical materializations of TTs. These, in turn, are aggregates or convergences of these practices guided by a technological paradigm: “a ‘model’ or ‘standard’ of solutions of technological problems, selected based on principles derived from the natural sciences and on selected material technologies...” ¹⁹19. Dosi G. Technological paradigms and technological trajectories. A suggested interpretation of the determinants and directions of technical change. Res Policy 1982; 11:147-62. (p. 152). Costa ¹⁰10. Costa FA. Structural diversity and change in rural Amazonia: a comparative assessment of the technological trajectories based on agricultural censuses (1995, 2006 and 2017). Nova Economia. 2021; 31:415-53. highlights that the rural reality of the Amazon is characterized by a historical-geographical-cultural structural diversity. In this context, the evolutionary patterns related to the workforce, technologies, inputs, knowledge, and institutions that make up the agrarian economy of the Brazilian Amazon brings to light a set of TTs ⁴4. Costa FA. A brief economic history of the Amazon (1720-1970). Cambridge: Cambridge Scholars Publishing; 2019..

The TTs, guided by technological paradigms associated with production, combine agents and their microeconomic rationalities under two large groups: agroextractivist and mechanical-chemical-genetic ²⁰20. Folhes RT, Fernandes DA. A dominância do paradigma tecnológico mecânico-químico-genético nas políticas para o desenvolvimento da bioeconomia na Amazônia. Paper do NAEA 2022; 31:1-25.. These paradigms guide different arrangements for the organization of their workforce, land structure, capital, and the package of technological solutions in their relations with the forest. Associated with these paradigms are two main economic agents: peasant and capitalist. While the agro-extractive paradigm is associated with the idea of a living nature, in which the forest is a means of production, the mechanical-chemical-genetic paradigm is associated with the idea of a dead nature, in which the forest is input, a resource to be consumed.

Costa ⁹9. Costa FA. Trajetórias tecnológicas como objeto de política de conhecimento para a Amazônia: uma metodologia de delineamento. Rev Bras Inov 2009; 8:35-86.^,1010. Costa FA. Structural diversity and change in rural Amazonia: a comparative assessment of the technological trajectories based on agricultural censuses (1995, 2006 and 2017). Nova Economia. 2021; 31:415-53. identified six TTs present in the regional agrarian economy, three of which are associated with peasant agents and three with capitalist. Box 1 shows a synthesis of these TTs and describes the patterns and arrangements that can be observed for the elements present in the landscape, shaped by these agents and their productive technical systems.

Identifying the production landscape units

The first methodological step consists in choosing the resolution of the cell. Each cell represents a landscape unit, and their set forms a grid (cell space) ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45.. Land use and cover information retrieved from satellite imagery is generally provided systematically, with 10m, 20m, and 30m of average spatial resolution. Thus, to evaluate the spatial arrangement and composition of land use and cover classes that make up production landscapes, we worked with a cell grid, with a [2×2]km spatial resolution. The definition of cell size is done empirically ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45. and involves the use of medium-resolution satellite imagery, land cover use data, and field records, integrated and analyzed in a Geographic Information System (GIS). The definition of the cells size depends on the characteristics of the region, objects, and categories of interest ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45.. Therefore, we considered the size of the areas of agricultural use that vary in the study area, from 0.5 to 2ha for peasants and above this value for medium and large agricultural capitalist systems, reaching 200ha or more. Based on the defined cell size, 1,291 cells were generated for the study area: 290 for the municipality of Mocajuba and 1,001 for Cametá.

The construction of the PLU is influenced by the cells size that will delimit the patterns of the landscapes of interest. The PLU is described based on the observation, in the databases, of the spatial arrangements and the composition of the landscape elements (polygons and classes of land use and cover) found in the cell, in addition to the context information, such as density and size of rural establishments, proximity to rivers, location on island or floodplain areas. For this construction, we start from the principle that the technical systems of rural production present landscapes with specific spatial patterns, visually identifiable in the data derived from satellite images, showing correspondences with the TTs, but not necessarily in a direct relationship.

To build the typology of the PLU for the region, it is necessary to consider its economic and agrarian context. The municipalities of Cametá and Mocajuba are located in the Baixo Tocantins region in Pará State, the second largest açaí producing region in Pará ²¹21. Instituto Brasileiro de Geografia e Estatística. Produção agrícola municipal. https://sidra.ibge.gov.br/tabela/1613#resultado (accessed on 28/Feb/2024).
https://sidra.ibge.gov.br/tabela/1613#re... . This region has significant agro-extractive and agricultural production, based on peasant technical systems. Açaí is produced in different systems, in areas of islands, floodplains, and, more recently, in areas of upland ²²22. Souza AR, Adorno BV, Gonçalves GC, Bragion GR, Oliveira KD, Escada MIS, et al. Paisagens e uso da terra em núcleos populacionais e estabelecimentos rurais da região do Baixo Tocantins - Pará. http://urlib.net/ibi/8JMKD3MGP3W34T/44STMLE (accessed on 09/Jul/2021).
http://urlib.net/ibi/8JMKD3MGP3W34T/44ST... . Small-scale agriculture, also presenting with an expressive economy, is located in areas of upland, associated with the cultivation of black pepper, cocoa, and cassava, the latter produced in a fallow system ²²22. Souza AR, Adorno BV, Gonçalves GC, Bragion GR, Oliveira KD, Escada MIS, et al. Paisagens e uso da terra em núcleos populacionais e estabelecimentos rurais da região do Baixo Tocantins - Pará. http://urlib.net/ibi/8JMKD3MGP3W34T/44STMLE (accessed on 09/Jul/2021).
http://urlib.net/ibi/8JMKD3MGP3W34T/44ST... ^,2323. Nogueira KNS, Costa FA, Adami M. Território e trabalho: análise geoeconômica com base em trajetórias camponesas. Novos Cadernos NAEA 2018; 21:117-37.. Besides the peasants and their technical systems, there are capitalists and their modes of production, especially associated with the production of açaí in upland. In addition to the açaí produced in irrigated and mechanized systems, there is the production of palm oil, black pepper, and cocoa, the latter two often consorted with açaí²²22. Souza AR, Adorno BV, Gonçalves GC, Bragion GR, Oliveira KD, Escada MIS, et al. Paisagens e uso da terra em núcleos populacionais e estabelecimentos rurais da região do Baixo Tocantins - Pará. http://urlib.net/ibi/8JMKD3MGP3W34T/44STMLE (accessed on 09/Jul/2021).
http://urlib.net/ibi/8JMKD3MGP3W34T/44ST... . Fish farming is also observed, developed along with agricultural production, in both capitalist and peasant systems. Notably, this area, unlike most regions of the Amazon, has little livestock activity ²²22. Souza AR, Adorno BV, Gonçalves GC, Bragion GR, Oliveira KD, Escada MIS, et al. Paisagens e uso da terra em núcleos populacionais e estabelecimentos rurais da região do Baixo Tocantins - Pará. http://urlib.net/ibi/8JMKD3MGP3W34T/44STMLE (accessed on 09/Jul/2021).
http://urlib.net/ibi/8JMKD3MGP3W34T/44ST... ^,2424. Cunha MA. Sensoriamento remoto para sistemas agrários associados à produção de açaí na região do baixo Tocantins, PA [Master Thesis] São José dos Campos: Instituto Nacional de Pesquisas Espaciais; 2023..

To build a typology representing the PLU categories present in the region for 2021, an important information was incorporated into the land use and cover data: the indicator of potential areas of occurrence of native açaí palms. Given the diverse nature of the açaí productive systems, the available systematic production data of land use and cover, such as those of TerraClass ¹⁴14. Instituto Nacional de Pesquisas Espaciais. Projeto TerraClass 2012 - mapeamento do uso e cobertura da terra na Amazônia Legal. http://www.inpe.br/cra/projetos_pesquisas/dados_terraclass.php (accessed on Jan/2016).
http://www.inpe.br/cra/projetos_pesquisa... and MapBiomas (https://brasil.mapbiomas.org/) ²⁵25. Souza Jr. CM, Shimbo JZ, Rosa MR, Parente LL, Alencar AA, Rudorff BFT, et al. Reconstructing three decades of land use and land cover changes in Brazilian biomes with landsat archive and earth engine. Remote Sens 2020; 12:2735., were not used, since they do not identify the small-scale agriculture nor do they differentiate stages of secondary vegetation, largely associated with fallow shift cultivation. The need for better identification of crops led to the classification of land using Sentinel-2A (https://www.engesat.com.br/sentinel-2/) satellite images with 10m spatial resolution, to obtain classes of use and coverage more detailed and adhering to the characteristics of the study area.

Image data cubes from Sentinel-2A ²⁶26. Simões R, Camara G, Queiroz G, Souza F, Andrade PR, Santos L, et al. Satellite image time series analysis for big earth observation data. Remote Sens 2021; 13:2428. were used for classification, as well as supervised classification using Random Forest ²⁷27. Breiman L. Bagging predictors. Mach Learn 1996; 24:123-40., implemented in the SITS package (Satellite Image Time Series Analysis on Earth Observation Data Cubes) ²⁶26. Simões R, Camara G, Queiroz G, Souza F, Andrade PR, Santos L, et al. Satellite image time series analysis for big earth observation data. Remote Sens 2021; 13:2428.. The Cube was produced with monthly images from January, 1st 2021 to Dicember, 31 2021. Ten spectral bands were used, except the water vapor band. A map of land use and cover was generated, presenting 94% overall accuracy and a 0.92 kappa coefficient and containing the following classes: (a) small-scale agriculture; (b) large-scale agriculture; (c) herbaceous pasture; (d) shrubs pasture; (e) initial secondary vegetation; (f) advanced secondary vegetation; and (g) hydrography. The final map includes the forest and non-forest classes from PRODES ¹³13. Coordenação Geral de Observação da Terra, Instituto Nacional de Pesquisas Espaciais. Monitoramento do Desmatamento da Floresta Amazônica Brasileira por Satélite - PRODES. http://www.obt.inpe.br/OBT/assuntos/programas/amazonia/prodes (accessed on 04/Mar/2024).
http://www.obt.inpe.br/OBT/assuntos/prog... .

Indicator of potential areas of occurrence of native açaí palms is prepared from the combination of forest cover data with data from hydro-topographic environments ²⁸28. Rennó CD, Nobre AD, Cuartas LA, Soares JV, Hodnett MG, Tomasella J, et al. HAND, a new terrain descriptor using SRTM-DEM: mapping terra-firme rainforest environments in Amazonia. Remote Sens Environ 2008; 112:3469-81.. A hydro-topographic map is generated to indicate the lowland areas, transition zones (between floodplains and slopes), slope areas, and plateau areas (upland). The data generated by hydro-topographic mapping were combined with forest cover data. Thus, the areas with high potential for occurrence of native açaí (Euterpe oleracea) were defined as having primary or secondary forest cover, in floodplain (lowland) areas or in transition zones. Moderate potential areas were identified in zones with forest cover in slope areas; low potential areas were identified by presenting secondary vegetation and/or forest on upland ²⁴24. Cunha MA. Sensoriamento remoto para sistemas agrários associados à produção de açaí na região do baixo Tocantins, PA [Master Thesis] São José dos Campos: Instituto Nacional de Pesquisas Espaciais; 2023.. The areas that did not present primary forest or secondary vegetation were defined as areas of absence of açaí palms. The class of high potential of occurrence of açaí palms was integrated with land use and cover data, forming a new unified database. This integration is important in determining the types of PLU present in the constructed typology.

The unified land use and cover database was added to the [2×2]km cells, making it possible to define five types of PLU found in the study area. Three PLU were associated with peasant systems and two with agricultural capitalist systems, namely: (1) extractive peasant unit - PLU1; (2) agro-extractive peasant unit with temporary agriculture - PLU2; (3) peasant unit with açaí production, permanent and/or temporary agriculture, and pastures - PLU3; (4) capitalist unit with irrigated planted açaí - PLU4; and (5) capitalist unit with cattle farms - PLU5. Figure 1 illustrates the spatial patterns relative to the new land use and cover map, which characterize the types of PLU identified in the study area.

The PLU characterized in this study were developed considering that the technical systems associated with the production of açaí are predominant in the region. However, in upland environments, one can find other types of crops such as black pepper and palm oil, which may or may not include açaí, in addition to livestock areas that do not include açaí. As the PLU are inserted in different landscapes, such as island, floodplain, and upland environments, the crops and technological practices adopted in the management, harvesting, and storage of production are different. Therefore, the relations of these PLU with risk/protective factors in the context of local health are also different. Box 2 shows the final typology, covering the five types of PLU.

Mapping the production landscape units

The classification of the PLU was based on the typology developed (Box 2). Spatial attributes, composition, and context are extracted for each cell of [2×2]km for the classification of PLU, in which we use a supervised classifier based on decision trees with boosting method ¹⁸18. Quinlan JR. C5.0: an informal tutorial. https://www.rulequest.com/see5-unix.html#BOOSTING (accessed on 14/Aug/2021).
https://www.rulequest.com/see5-unix.html... . The attributes, or landscape descriptors, are used to train the algorithms that will perform PLU classification in a supervised manner. The sample collection process is carried out iteratively, until a consistent set of training samples is obtained to adequately represent the elaborated typology. The boosting method is used to improve classification accuracy by creating multiple decision trees based on a predefined number of iterations, in which at each iteration, the adjustment of training errors that occurred in the previous iteration is performed ¹⁸18. Quinlan JR. C5.0: an informal tutorial. https://www.rulequest.com/see5-unix.html#BOOSTING (accessed on 14/Aug/2021).
https://www.rulequest.com/see5-unix.html... . Combining several trees, with different metrics of composition, structure, and context of the landscape, the final decision of the model results from a voting system, which assigns to each cell the most frequent class imputed by the decision trees ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45.^,1818. Quinlan JR. C5.0: an informal tutorial. https://www.rulequest.com/see5-unix.html#BOOSTING (accessed on 14/Aug/2021).
https://www.rulequest.com/see5-unix.html... .

The classification of the PLU consisted of four steps and was carried out within the integrated environment Geographic Data Mining Analyst-GeoDMA ²⁹29. Korting TS, Fonseca LMG, Escada MIS, Da Silva FC, Dos Santos Silva MP. GeoDMA A novel system for spatial data mining. In: 2008 IEEE International Conference on Data Mining Workshops. https://ieeexplore.ieee.org/document/4734031?denied= (accessed on Mar/2020).
https://ieeexplore.ieee.org/document/473... :

i) Attribute extraction. Using land use and cover data, 246 metrics of landscape composition, structure, and context were extracted, such as distance to rivers, distance to roads, presence/absence of rivers, number and size of rural establishments, for each cell;

ii) Selection of training and test samples. A total of 212 samples were collected: 48 samples for PLU1, PLU2, PLU3, and PLU5 and 20 samples for PLU4. The lower number of samples for PLU4 is due to the lower representativeness of cells with the corresponding spatial pattern. Among the 212 samples collected, 70% were intended for classifier training and 30% were used in the validation phase as test samples;

iii) Classification of cells by boosted decision tree algorithm, with 99 iterations and independent classifications;

iv) Classification assessment. The assessment was performed using a confusion matrix with the set of test samples.

Results

Figure 2a presents the PLU mapping result, including an enlarged area. The peasant PLU (PLU1, PLU2, PLU3) cover 95% of the cells, evincing the dominance of the economic activities of these agents in the studied region (Figure 2b). PLU2 represents more than 45% (496 cells) of the total mapped cells, followed by PLU1, corresponding to almost 31% (335 cells), and PLU3 with 18% (202 cells). PLU4 corresponds to 4.1% (44 cells) while PLU5 does not reach 1% (8 cells). In alignment with the TTs, PLU1 and PLU2 are associated with the TT2, representing together almost 77% of the total mapped in the region, corroborating Nogueira et al. ²³23. Nogueira KNS, Costa FA, Adami M. Território e trabalho: análise geoeconômica com base em trajetórias camponesas. Novos Cadernos NAEA 2018; 21:117-37., and field records ²²22. Souza AR, Adorno BV, Gonçalves GC, Bragion GR, Oliveira KD, Escada MIS, et al. Paisagens e uso da terra em núcleos populacionais e estabelecimentos rurais da região do Baixo Tocantins - Pará. http://urlib.net/ibi/8JMKD3MGP3W34T/44STMLE (accessed on 09/Jul/2021).
http://urlib.net/ibi/8JMKD3MGP3W34T/44ST... , that indicate the predominance of activities of peasants in the region. Considering the last Agricultural Census (2017), the peasant trajectory TT2 is dominant in both municipalities. Regarding the capitalist trajectory, TT5&6 prevails in Cametá and TT7 prevails in Mocajuba ¹²12. Rorato AC, Dal'Asta AP, Lana RM, Santos RBN, Escada MIS, Vogt CM, et al. Trajetorias: a dataset of environmental, epidemiological, and economic indicators for the Brazilian Amazon. Sci Data 2023; 10:65..

In supervised classification ¹⁸18. Quinlan JR. C5.0: an informal tutorial. https://www.rulequest.com/see5-unix.html#BOOSTING (accessed on 14/Aug/2021).
https://www.rulequest.com/see5-unix.html... , six decision trees were generated to classify the PLU. The algorithm parameterization predicted 99 iterations, however, it was stabilized in the sixth training cycle. The adjustment error of the model varied from 0.6% to 4.4% until it stabilized in the last iteration at 0%. The choice of attributes is performed by an entropy measure that estimates the information gain of the variables during classification, discarding those with lower gains. Figure 3a shows the selected attributes. Of the 15 metrics used in the decision trees, four were used at the root, generating the initial branches: (1) area of advanced secondary vegetation (CA_VSA); (2) form index of the herbaceous pasture class (LSI_PL); (3) spot density of large-scale agriculture areas (PD_AGLE) and; (4) fractal dimension index of large-scale agriculture areas (AWMPFD_AGLE). Most metrics were used to differentiate PLU4 or PLU5 from the others, with the exception of CA_VSA, which separated PLU1 in the first cut. In the intermediate and final branches of the trees, other metrics stood out, such as medium size of small-scale agriculture class (MPS_AGPE), in addition to three context metrics related to minimum distance to rivers (HID_min), presence of rivers (P_HID), and hydrography interspersion and juxtaposition index (IJI_HID), used to separate PLU2 from PLU3. PLU2 is located near the rivers, in which activities such as açaí management and collection predominate, with smaller areas dedicated to agriculture, while PLU3, located farther from the rivers, has agriculture as the main economic activity, and may or may not include açaí crops.

In one of the trees, the TABO_alto metric was used, which measures the Area of the largest object in the class high potential for the occurrence of açaí palms. This metric separated PLU3 from PLU2. This result shows the importance of differentiating the classes of agriculture in the input data for the distinction of capitalist and peasant PLU. No metrics of landscape composition, density, and size of agricultural establishments were used. Improvements and updates to the input databases could improve, for example, the separation between peasant units and capitalist units. Figure 3b shows one of the six trees generated with the landscape structure and context metrics.

The confusion matrix of Figure 3c enables an evaluation of the classification with 30% of the test samples. There were few confusions, and the most significant occurred between PLU3, peasant, and PLU5, capitalist associated with livestock. This occurred because both PLU3 and PLU5, located on upland, can present classes such as shrubs pasture, initial secondary vegetation, and forest. Imbalance in training samples is a limitation of this method. PLU1 was confused with PLU2 in 7% of the cells. There were no confusions between PLU2 and PLU4, although PLU4 presents a limited number of samples.

Discussion

The classification is consistent with the expectations for this region, aligning with field records ²²22. Souza AR, Adorno BV, Gonçalves GC, Bragion GR, Oliveira KD, Escada MIS, et al. Paisagens e uso da terra em núcleos populacionais e estabelecimentos rurais da região do Baixo Tocantins - Pará. http://urlib.net/ibi/8JMKD3MGP3W34T/44STMLE (accessed on 09/Jul/2021).
http://urlib.net/ibi/8JMKD3MGP3W34T/44ST... and works incorporating the TTs ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754.^,2323. Nogueira KNS, Costa FA, Adami M. Território e trabalho: análise geoeconômica com base em trajetórias camponesas. Novos Cadernos NAEA 2018; 21:117-37.. The mapping of PLU at the intramunicipal scale showed the persistence in 2021 of peasant TTs associated with PLU1, PLU2, and PLU3. PLU1 and PLU2 predominate in the islands and floodplains of Baixo Tocantins and its tributaries. The presence of PLU3 is identified as it moves away from floodplain areas. PLU4 and PLU5, associated with capitalist TTs, although less expressive, are more distant from Baixo Tocantins, located close to roads and highways, in upland areas, in which activities such as livestock and açaí monoculture are developed, configuring homogeneous landscape patterns. The PLU associated with the TTs coexist and interact in different ways in the economic dynamics of rural Amazon. The changes in the structure and composition of forest landscapes of the ways of living and modes of production, associated with technical systems, are captured by the PLU. Thus, the PLU constitute a structure of representation that introduces new possibilities to (re)think the elements of exposure and contact of populations in these areas to old and new pathogens ³⁰30. Mancini MCS, Barreto JR, Carvalho RL, Muylaert RL, Arrais RC, Prist PR. Landscape ecology meets disease ecology in the tropical america: patterns, trends, and future directions. Curr Landscape Ecol Rep 2024; 9:31-62., along with the integrated health-environment-economy approaches.

Diseases such as visceral leishmaniasis (VL), cutaneous leishmaniasis (CL), malaria, dengue, Chagas disease, for example, prevail in the Amazon to the present day ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754.^,3131. Sousa Júnior AS, Gonçalves NV, Miranda CSC, Santos BO, Oliveira RAC, Costa RJF, et al. Cutaneous leishmaniasis spatial distribution and epidemiological and environmental risk factors in Cametá, state of Pará, Brazil. Braz J Infect Dis 2020; 24:330-6.. The Trajetorias¹²12. Rorato AC, Dal'Asta AP, Lana RM, Santos RBN, Escada MIS, Vogt CM, et al. Trajetorias: a dataset of environmental, epidemiological, and economic indicators for the Brazilian Amazon. Sci Data 2023; 10:65. database, constructed for all the municipalities of the Brazilian Legal Amazon, presents the incidence rates of the main diseases found in the studied area, stratified by urban and rural areas. In Cametá, the Trajetorias¹²12. Rorato AC, Dal'Asta AP, Lana RM, Santos RBN, Escada MIS, Vogt CM, et al. Trajetorias: a dataset of environmental, epidemiological, and economic indicators for the Brazilian Amazon. Sci Data 2023; 10:65. database shows, from 2015 to 2019, an incidence of malaria (Plasmodium vivax) of 4,620.19 in the rural environment, considering the rate per 100,000 inhabitants. Chagas disease had a rate of 22.61, while VL had 27.60 and CL, 6.65. In Mocajuba, no cases of Chagas disease were registered from 2015 to 2019 in the rural environment. The incidence rate of malaria (Plasmodium vivax) was 2,466.58, VL was 31.54, and the CL, 40.15.

PLU1 and PLU2 present socio-environmental conditions that favor the occurrence of malaria (Plasmodium vivax), Chagas disease, and CL in the study area. In communities with açaí production in Baixo Tocantins, cases of Chagas disease with oral transmission were recorded ³²32. Carneiro ER, Sousa RL. Doença de chagas: fatores de risco associados ao consumo da polpa de açaí em uma comunidade rural, Abaetetuba, Pará. Biodiversidade 2019; 3:148-51.. Generally, it is in island and floodplain areas that peasants grow açaí, represented by PLU1 and PLU2. The processes of fruit harvesting, sanitation, and storage introduce risks for oral transmission, since the fruits can be contaminated with feces and urine of the vector, containing Trypanosoma cruzi³³33. Nascimento LPGR, Nogueira LMV, Rodrigues ILA, André SR, Graça VV, Monteiro NJ. Prevalência da doença de chagas associada ao modo de infecção. Cogit Enferm 2021; 26:e73951., at any of these stages. Although farmers receive guidance by agents of local institutions and by the industries that buy the açaí fruit, in practice they are still insufficient for effective control of the disease, resulting in the incidence of Chagas in the region. In the case of CL, the study by Sousa Júnior ³¹31. Sousa Júnior AS, Gonçalves NV, Miranda CSC, Santos BO, Oliveira RAC, Costa RJF, et al. Cutaneous leishmaniasis spatial distribution and epidemiological and environmental risk factors in Cametá, state of Pará, Brazil. Braz J Infect Dis 2020; 24:330-6. shows that from 2007 to 2016, islands and floodplains areas in Cametá had the highest numbers of CL cases.

Leishmaniasis is predominantly transmitted by the bite of sandflies (birigüi, mosquito-palha) of the genus Lutzomyia. These mosquitoes use moist places, with shade and organic matter, such as leaves and fruits, as well as animal feces, to reproduce. Animal burrows, the base or trunk of trees are potential breeding grounds. They feed on the blood of various species of wild animals, synanthropic animals, and, in specific transmission cycles, domestic animals ³¹31. Sousa Júnior AS, Gonçalves NV, Miranda CSC, Santos BO, Oliveira RAC, Costa RJF, et al. Cutaneous leishmaniasis spatial distribution and epidemiological and environmental risk factors in Cametá, state of Pará, Brazil. Braz J Infect Dis 2020; 24:330-6.. These can be natural hosts or (potential) wild reservoirs. PLU1 and PLU2 present situations of houses very close to the forest and forest environments close to the home environment and production areas, and there is also the presence of domestic and wild animals, in particular rodents. These are landscapes that increase the risks of exposure and contact of sandflies with peasant producers. Guidelines related to the management of production territories, which in the case of peasants involves their home and peridomicile, such as cleaning backyards and their cultivation and extraction areas, can be a strategy designed by health surveillance, integrated with other municipal policies.

PLU3, although peasant, presents a differentiated technical system. It is located in more diverse landscapes, with açaí production, permanent and/or temporary agriculture, and pastures. It presents socio-environmental conditions that favor the risk of CL and VL transmission. It occurs on the mainland, encompassing the urban fringes, distant from rivers but close to vicinal roads. It is located in regions where forest ecosystem disturbances are more pronounced than in PLU1 and PLU2. Deforestation, proximity to vicinal roads, and proximity to dense urban areas, characteristics of these PLU, increase the risks for leishmaniasis (cutaneous and visceral), as observed in Baixo Tocantins ³¹31. Sousa Júnior AS, Gonçalves NV, Miranda CSC, Santos BO, Oliveira RAC, Costa RJF, et al. Cutaneous leishmaniasis spatial distribution and epidemiological and environmental risk factors in Cametá, state of Pará, Brazil. Braz J Infect Dis 2020; 24:330-6.^,3434. Monteiro NP, Costa RTM, Fonseca PCC, Santos GRC, Pontes CDN, Maneschy RB, et al. Aspectos clínicos e epidemiológicos da leishmaniose visceral na região do baixo Tocantins/PA. Revista Ibero-Americana de Ciências Ambientais 2020; 11:224-33.. In the case of this PLU, the issues related to the processes of the açaí production chain, collection, sanitation, and storage, associated with Chagas disease, are also present ³³33. Nascimento LPGR, Nogueira LMV, Rodrigues ILA, André SR, Graça VV, Monteiro NJ. Prevalência da doença de chagas associada ao modo de infecção. Cogit Enferm 2021; 26:e73951..

PLU4 is associated with açaí monoculture. This PLU has a technological solution package with the use of mechanics, chemistry, genetics, with specialized technical assistance. However, it produces a little diverse environment, with little forest cover. The circulation of wildlife is restricted to nearby forest fragments and the new “forest” of planted palms, in general, intercropped with another culture. This environment favors the transmission of diseases that have wild animals in their cycle, which are hosts and reservoirs of a wide range of parasitic species. Açaí in consortium with cocoa presents socio-environmental conditions favorable to contamination by VL ³⁴34. Monteiro NP, Costa RTM, Fonseca PCC, Santos GRC, Pontes CDN, Maneschy RB, et al. Aspectos clínicos e epidemiológicos da leishmaniose visceral na região do baixo Tocantins/PA. Revista Ibero-Americana de Ciências Ambientais 2020; 11:224-33.. On the other hand, the population density in these areas is lower, since it employs little labor, reducing the chances of contact and risks of transmission on site or in neighboring areas.

PLU5 is characterized by extensive livestock farming. The technical systems of extensive livestock farming significantly consume the forest cover, which is replaced by pasture. Thus, they draw transition zones between pasture areas and forest areas, which become conducive habitats for wild rodents ³⁵35. Guerra JAO, Guerra MGVB, Vasconcelos ZS, Freitas NS, Fonseca FR, Silva Junior RCA, et al. Socioenvironmental aspects of the Purus Region - Brazilian Amazon: why relate them to the occurrence of American tegumentary Leishmaniasis? PLoS One 2019; 14:e0211785.. The landscapes of livestock production in Cametá and near the banks of the Tocantins River recorded occurrences of CL ³¹31. Sousa Júnior AS, Gonçalves NV, Miranda CSC, Santos BO, Oliveira RAC, Costa RJF, et al. Cutaneous leishmaniasis spatial distribution and epidemiological and environmental risk factors in Cametá, state of Pará, Brazil. Braz J Infect Dis 2020; 24:330-6.. For these territories, of types PLU4 and PLU5, health surveillance can design specific actions and establish, together with other municipal public sectors and with capitalist, procedures to monitor and control the presence and circulation of wild animals, in the interest of local public health.

In this analysis, still initial and descriptive, we observed how to identify, characterize, and map the PLU, helping to understand the differentiated risks of human exposure to pathogens. An analytical framework that integrates socio-environmental elements in the context of agricultural production is a promising structure to be investigated. Although these PLU have been mapped for only one year, in a specific region, they can be produced for several years and adapted to other regions, allowing one to observe the dynamics of technical systems and TTs ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45., their evolution, and in which directions their dynamics change the landscapes of the biome, at the intra-municipal scale. Spatial data, derived from satellite images, make it possible to generate refined analyzes at different spatial and temporal scales ⁷7. Souza AR, Escada MIS, Silva LGF, Fernandes DA, Monteiro AMV. A economia dos sistemas tecno-produtivos rurais sob a perspectiva das trajetórias tecno-produtivas e sua expressão na paisagem Amazônica. Revista da UFMG 2021; 28:109-45.^,2424. Cunha MA. Sensoriamento remoto para sistemas agrários associados à produção de açaí na região do baixo Tocantins, PA [Master Thesis] São José dos Campos: Instituto Nacional de Pesquisas Espaciais; 2023..

In this article, we did not explore PLU as a variable in statistical or mixed models to study the relationships between its dynamics and the emergence, re-emergence, or intensification of health-disease processes. Our focus was to point out this new integrative structure as a territorial representation of the health-environment-economy triad and its potentials in integrated health studies. Understanding the interaction between the social agents that promote economies in the biome, the landscapes they transform, and their impacts on health-disease processes is essential to develop (new) surveillance and control strategies or rethink existing ones. In this process, one should consider the constant transformation of the forest ecosystem, driven by the decisions of its economic agents, and their impacts on public health ¹¹11. Codeço CT, Dal'Asta AP, Rorato AC, Lana RM, Neves TC, Adreazzi CS, et al. Epidemiology, biodiversity, and technological trajectories in the Brazilian Amazon: from malaria to COVID-19. Front Public Health 2021; 9:647754.^,1616. Castro MC, Baeza A, Codeço CT, Cucunubá ZM, Dal'Asta AP, De Leo G, et al. Development, environmental degradation, and disease spread in the Brazilian Amazon. PLoS Biol 2019; 17:e3000526., affecting the communities and all Amazonian populations.

The structure presented can help to instrumentalize the debates in the health dimension regarding policies and proposals for more effective strategies, associated with choices of paths for agrarian economic development in the Brazilian Amazon. At the moment, a dispute over Bioeconomy ²⁰20. Folhes RT, Fernandes DA. A dominância do paradigma tecnológico mecânico-químico-genético nas políticas para o desenvolvimento da bioeconomia na Amazônia. Paper do NAEA 2022; 31:1-25. projects is underway as a strategy for regional development. Understanding what Bioeconomy we are talking about and, in particular, how our choices for one project or another impact the health of the forest ecosystem and the health of populations in the Amazon, is necessary and urgent.

Acknowledgments

This work was funded by Brazilian National Research Council (CNPq, Brazilian Center for Synthesis on Biodiversity and Ecosytem Services − SinBiose, process: 442357/2019-2). It was carried out with the support of the Brazilian Coordination for the Improvent of Higher Education Personnel (CAPES, funding code 001).

References

Publication Dates

History