Use of remote sensing for archaeological survey in the Lower Amazon Floodplain

Abstract

The extensive and dynamic floodplains of the Amazon Basin have seldom been explored from an archaeological perspective due to the dynamic processes of erosion and deposition that shape them. A frequent assumption is that most landforms of the whitewater floodplains have recently been reworked. To better understand where human activities may have left an archaeological record in these ecosystems, it is essential to enhance our comprehension of the floodplain's geomorphology. We seek to construct a method of mapping older fluvial levees, that are landforms where we suspect that archaeological contexts may be preserved. The format of the terrain's lateral accretion can be identified based on fluvial geomorphology. Due to the annual flooding of the landscape, it is hypothesized that the higher vegetation corresponds to elevated landscape positions of vertical accretion. This proposal was explored through data from remote sensing and ground truthing. We sought to identify this vegetation using different values of NDVI (Normalized Difference Vegetation Index) that may aid in locating levees and, consequently, pre-colonial archaeological sites. This study presents qualitative findings from ground-truthing conducted in the Alenquer floodplain region of the Lower Amazon. Even though our results are initial, they demonstrate successful identification of some levees and document a late Holocene archaeological site located on a levee. Our study suggests that despite the losses of portions of the floodplain to erosional processes, there is an archaeological record to be explored in the seasonally flooded forests that may improve our understanding of historical human relationships with these ecosystems.

1 Introduction

For more than half of the 20th Century, Amazonia was viewed as a place lacking human occupation, where many areas were left unexplored and uninhabited (Denevan, 1992). This was based on the assumption that living in the forests was extremely challenging and, as such, was linked to a constant struggle for survival. This perspective resulted in the opinion that it would be unlikely for human groups to establish large, thriving settlements (Meggers, 1971; Meggers et al., 1988). However, multidisciplinary research in recent decades reveals a varied history of Amazonia. Current scholarship highlights that Amazonian landscapes, which may appear untouched at first glance, were shaped by human actions over thousands of years. Examples of these transformations include anthropogenic forests, shellmounds, Amazonian Dark Earths, geoglyphs, and other types of earthworks (Erickson, 2010; Furquim et al., 2023; Hilbert et al., 2023; Lombardo et al., 2022; Moraes and Neves, 2019; Pugliese et al, 2024; Souza et al., 2019). These studies show that the Amazon rainforest has had the capacity of sustaining various populations for a long time and has a long-term indigenous history (Balée, 2018; Clement et al., 2015; Neves, 2022).

Through detailed studies, Amazonia—once perceived as vast expanses of green (forest) and blue (water)—is now recognized as a mosaic of diverse terrestrial and aquatic ecosystems (Junk et al., 2011; Shock and Moraes, 2019). According to Junk and colleagues’ (2011) classification system, there are 14 types of wetlands in the Amazon Basin, with floodplains belonging to the group whose water levels fluctuate. Sioli (1984) classified Amazonian rivers according to their course morphologies and the waters’ different physical and chemical properties: clear waters (e.g., Tapajós, Xingu, Iriri, and Tocantins Rivers), black waters (e.g., Negro, Coari, Jutaí, and Tefé Rivers) and white waters (e.g., Amazonas, Purus, Juruá, and Madeira Rivers). The white waters are known to have a high sediment load due to their origin in geologically more recent soils in the Andean Cordillera that are additionally recognized as being highly fertile (Cunha, 2011; Junk et al., 2012). This project focused on the floodplains of whitewater rivers (várzeas) that are known for their instability due to poorly consolidated and recently deposited sediments (Tandon and Sinha, 2007). Modern floodplain inhabitants speak of terras crescidas and terras caídas (‘grown’ and ‘fallen’ lands), revealing their understanding of sedimentary deposition and erosion that shape and remodel the floodplain, through the river's movements and its flood cycles (Sioli, 1984, 141–143).

The Amazon Basin (Figure 1), the largest river basin in the world, covers an area of 5,000,000 km² (Lakshmi et al., 2018), and the main course of the Amazon River is over 6,000 km long (Tandon and Sinha, 2007). Melack and Hess (2010) estimate that the Amazon River floodplain corresponds to 12% of the basin's wetlands and covers an area of 98,110 km². It is an ecosystem rich in biodiversity (Dagosta and De Pinna, 2019; Junk and Silva, 1997; Petermann, 1997; Wittmann et al., 2006; Wittmann et al., 2013). Despite the floodplain occupying a prominent position in some classical theories of pre-colonial human occupation in the Amazon (Carneiro, 1970; Denevan, 1996; Lathrap, 1970; Meggers, 1971), there is a lack of detailed information that can substantiate suppositions about the ancient history of human presence (or its “absence”) in these floodable terrains. Much of the archaeology around the floodplains has focused on their integration into a system with bluff habitation (e.g., Denevan, 1996). Evidence is needed to fill this gap in archaeological record.

Figure 1

The floodplain holds archaeological potential, even though river dynamics can affect the preservation of sites, particularly through erosion (destruction) and sedimentation (burial). Archaeological sites in the floodplain are registered for the Ucayali River (Upper Amazon River) (Lathrap 1968), the Solimões River (Middle Amazon River) (Araújo and Lima, 2023; Macedo et al., 2019) and the Lower¹ Amazon River floodplain (Costa, 2012; Hilbert, 1959; Lima, 2018; Nimuendajú, 2004; Roosevelt et al., 1991), but very few of these sites have been excavated.

Campos et al. (2024) developed a method based on principles of fluvial geomorphology and remote sensing to map fluvial levees in ancient parts of the floodplain, for the purpose of aiding archaeological survey in floodplain terrains. Fluvial levees are the most prominent topographic features of floodplains, formed along the river banks by the deposition of alluvial sediments during successive flooding events (Hudson, 2008). They are longitudinal as they parallel the river and wedge-shaped in cross section, mainly composed of coarse sediments [sand and silt], as the bedload is lost as soon as the flood waters leave the channel (Branß et al., 2022; Oloo and Xie, 2018). The sediment texture of depositional scroll bars has been shown to be, on average, 65% silt and 10%–20% sand with the rest clay (Mertes et al., 1996). This composition is expected to be overlain by silt and clay deposits derived from flood events after the channel responsible for the levee has been abandoned. According to Goldberg and MacPhail (2006), these landforms have the potential to preserve archaeological contexts.

Using satellite images from the Landsat collection, Campos et al. (2024) performed a temporal analysis to detect erosional and depositional contexts of the Amazonian floodplain in recent decades and the marks (scroll bars) left by the deposition of point bars. With an understanding of their local manifestation, the scroll bars and their interruptions were evaluated to propose hypothetical depositional sequences on the Amazonian islands (discerning between older and more recent deposition). Supervised classification was employed to map the non-flooded vegetation cover that might indicate the location of fluvial levees. After the development of this methodological design, the next step was to verify in the field whether vegetation mapping actually helped to locate fluvial levees and, simultaneously, archaeological sites. This article aims to qualitatively present the results of fieldwork, including ground-truthing and archaeological survey, conducted in a floodplain area of the Lower Amazon. It demonstrates the potential contributions of this type of research to Amazonian archaeology, as discussed further in section 2.1 (Regional archaeological context). We think that studying the floodplains will make it possible to discover more about multiple histories, including providing insights into issues of territoriality, mobility, and interactions.

2 Amazon floodplain

The study area initially analyzed (Campos et al., 2024) is located between channels of the Amazon River across four different municipalities: Óbidos, Curuá, Alenquer and Santarém (western Pará state, Lower Amazon), covering an area of approximately 860 km². To plan ground-truthing, a smaller area of floodplain was selected (Figure 2). This area is in Alenquer, specifically within the Atumã Agro-Extractivist Settlement Project (Projeto de Assentamento Agroextrativista—PAE). The PAE is part of Brazil's agrarian policy managed by the National Institute for Colonization and Agrarian Reform (Instituto Nacional de Colonização e Reforma Agrária—INCRA), with the aim of legalizing lands occupied by families from traditional communities in the floodplain and guiding sustainable use of the ecosystem (Thomas et al., 2014; Corrêa et al., 2019, p. 744).

Figure 2

The Amazon floodplains have their origins in the Quaternary period, with a division between the paleo-floodplains from Pleistocene sea level highstands and the modern floodplains that formed during the Holocene (Irion et al., 2010). Currently, the tide's daily influence can be felt 1,000 km upstream, as far as the Strait of Óbidos (Sioli, 1984, p. 133). The changes between glacial and interglacial climatic conditions caused variations in sea level, due to the retention, or not, of water in glaciers. Interglacial conditions, with high sea levels, had a direct influence on the formation of floodplains due to the low gradient of the Amazon River (Irion, 1984; Junk, 1984; Sioli, 1984). More details on formation processes can be found in Campos et al. (2024).

According to the Köppen's climate classification, the study area is Am, tropical monsoon (Alvares et al., 2013). The nearest temperature station in the floodplain (Óbidos) registered an average of 26.9°C (between 1973 and 2013) (Almeida et al., 2016). The closest weather station (Santarém) records 1677 ± 490 mm of precipitation in the winter from December to June, with peak rainfall between March and April, and 377 ± 139 mm in the summer from July to November, with drought conditions in September and October (McGrath et al., 1993, Moraes et al., 2005).

Morphologically speaking, the Middle and Lower Amazon River courses are classified as anabranching and anastomosing (Rozo et al., 2014). Fluvial landform diversity is evident, influenced by drainage patterns and varying intensities of fluvial dynamics (Ashworth and Lewin, 2012; Dunne and Aalto, 2013; Lewin and Ashworth, 2014). The Amazon River diversity encompasses small to large channels (e.g., furos, igarapés, paranãs), lakes (fluviolacustrine system), flooded fields, fluvial levees, and point bars (Dantas and Teixeira, 2013, p. 31). Pedologically hydromorphic soils predominate: haplic gleysols, fluvic neosols and haplic plinthosols (Dantas and Teixeira, 2013, p. 32).

Floodplain vegetation, like everything in the ecosystem, is linked to the water dynamics that occur between the dry and flood seasons, as well as their intermediate seasons, that establish a cycle of plant succession (Junk, 1984). When the water level is high (aquatic phase) macrophytes and phytoplanktons flourish in the depressions and lakes, meanwhile with low water levels (the terrestrial phase) herbaceous plants take root followed by (semi-)aquatic grasses (Junk, 1984, p. 221). Where topographically elevated areas develop (from sediment deposition), the grasses and herbaceous plants are replaced by trees and palms, making up the high floodplain forest (Salo et al., 1986) (Table 1).

The greatest freshwater ichthyological diversity in the world can be found in the Amazon Basin waters (Oberdorff et al., 2019). Species range in size from the small, that include the charuto [Pyrrhulina australis], with an adult size of 5 cm; the branquinha [Psectrogaster amazonica], with an average size of 19 cm and 175 g in weight; and the tamoatá [Hoplosternum littorale], with lengths up to 30 cm and weigh up to 450 g, to the large, such as the pirarucu (Arapaima gigas), and the piraíba (Brachyplathystoma filamentosum), which can be over 2 m long and weigh over 200 kg. There is an accompanying diversity of aquatic mammals, that include the manatee (Trichechus inunguis) and some species of large rodents as well as birds and reptiles, such as caimans and chelonians (Lathrap, 1970; Meggers, 1971).

Riverside houses are often built on stilts in the highest parts of the floodplain² (Figure 3), differing from houses on non-flooded terra firme. This adaptation allows residents to remain above the water level during most floods, although extreme flood events may still inundate these structures. In the dry season, these stakes stand out in the landscape and the high-water marks from flooding are visible.

Figure 3

2.1 Regional archaeological context

The study of archaeological sites in the Lower Amazon region during the Late Holocene (4000–450 BP) includes the division of ceramic remains into named phases and traditions. In the research area those documented at sites on or near the floodplain include Zoned hachure, Pocó, Konduri, and Tapajó. The Zoned hachure ceramics were defined, in part, on the basis of surface collections from the Ponta do Jauari shellmound located in the floodplain between Curuá and Alenquer (Pará state), however they lack chronological placement (Hilbert, 2009; Meggers and Evans, 1961). For the Lower Amazon floodplains shellmounds are documented, including Taperinha (a shellmound located in the floodplains east of Santarém) where predominantly undecorated ceramics are dated to 7500 years ago (Roosevelt, 1995). This is different from the Ponta do Jauari shellmound with its Zoned hachure material. Few sites, however, are documented in the floodplains.

Pocó³ and Konduri ceramics have their origins in the Nhamundá-Trombetas rivers area (west of Óbidos/Pará), with the Pocó occupations being older (3000–1200 BP) than the Konduri occupations (1250–1650 AD) (Alves, 2020; Zuse, 2016). Across the Amazon, the sites with Tapajó material (900–1600 AD) center on the city of Santarém/Pará, in the Lower Tapajós region (Troufflard, 2016). It is important to note that the Tapajó, that gave name to the ceramics, and many other Indigenous ethnic groups, that may or may not have used the same styles of vessels, still reside in the Lower Tapajós region today (Peixoto et al., 2012, p. 280; Lima et al., 2020, p. 2).

Pocó ceramics have been found in the Nhamundá and Trombetas Rivers areas on the left bank, Juruti, and the region of Santarém on the right bank of the Amazon River (Hilbert and Hilbert, 1980; Guapindaia and Lopes, 2011; Gomes, 2011; Sá, 2016; Lima, 2018, p. 131). Meanwhile the more recent Konduri pottery's extent is from the middle Trombetas River region and the Óbidos municipality on the left bank to the municipalities of Parintins and Juruti on the right bank, not reaching Santarém (Alves, 2018).

The distribution of Tapajó pottery has been mapped over an estimated area of 2000 km² on the right bank of the Amazon River after it meets the mouth of the Tapajós River. A total of 156 archaeological sites related to the Tapajó occupation have already been found in the non-flooded lowland areas along rivers and on plateaus in the region of modern Santarém and Belterra, Pará state (Nimuendajú, 2004; Stenborg et al, 2012; Stenborg, 2016; Stenborg et al., 2018; Troufflard and Alves, 2019, p. 554). Sites are located in a diversity of modern environments including campinaranas (a particular type of savannas) on sandy soils, and forests on both sandy soils on lowland terraces near the rivers and clayey soils on the plateau (Lima, 2018). The sites in the urban area of Santarém are no longer interpreted as a Tapajó political center (according to a hierarchical view), but as part of a web of heterarchical relations between the larger and smaller villages in the region (Schaan, 2016; Troufflard and Alves, 2019). A map by Barreto and Alves (2023) extends the Tapajó ceramics distribution further to the east, approximately to the modern city of Prainha, Pará state.

In the Lower Amazon region, therefore, settlements with Konduri ceramics predominate on the Amazon River's left bank, while settlements with Tapajó material predominate on the right bank. Between these ceramic complexes there are extensive areas of floodplains. Ethnohistorical data and archaeological evidence indicate that there was a network of relationships through exchange and trade between various ethnolinguistic groups in the Lower Amazon region, including the Tapajó and Konduri (Troufflard, 2016). These interactions would have taken place via river and land routes, where people, material culture and ideas circulated (Schaan, 2016) establishing a network of relations (Lima, 2018) of short, medium, and long distance (going beyond the Lower Amazon).

Near Santarém archaeological evidence has been found in some floodplain communities, specifically Centro do Aritapera, Cabeça do Onça, and Carapanatuba. The material includes pottery sherds (some with decoration, such as basketry marks, painting, and incisions), a fragmented ceramic figurine, a ceramic pipe, polished axes, and a possibly anthropogenic ditch (Costa, 2012). Lima (2018, p. 126–127) associates these and other sites located in the Lower Amazon floodplain with a probable Tapajó occupation, although there is also a chance that some of the material can be historic.

The floodplain can be an important area for resource extraction. Zooarchaeological studies conducted in the Middle Amazon (Hatahara site, near Manaus, Amazonas state, Brazil) and southwestern Amazon (Monte Castelo shellmound in the Guaporé River region and Loma Salvatierra in the Llanos de Moxos, Bolivia) have attested to the importance of aquatic fauna in the diet of Amazonian societies in the pre-colonial period (Prestes-Carneiro et al., 2016; Prestes-Carneiro and Béarez, 2017; Prestes-Carneiro et al., 2019, Prestes-Carneiro et al., 2020). In the Lower Amazon, studies based in biology, anthropology, and economics have shown that the consumption of various fish and chelonians is a common practice in the daily life of communities located in the current floodplains, where fisherpeople have expert knowledge of the ecological, social, and reproductive behavior of aquatic fauna (Barboza, et al., 2013; Corrêa, et al., 2019; Laurido and Braga, 2018; Lima et al., 2019).

Archaeological sites such as the Ponta do Jauari shellmound, Taperinha (a shellmound located in the floodplains east of Santarém), and Caverna da Pedra Pintada (a shelter located in an area of rock formations in Monte Alegre/Pará), include aquatic fauna but more detailed analyses of these remains have yet to be published (Frikel, n.d;Roosevelt et al., 1991; Roosevelt et al., 1996). An archaeological rescue on a previously unexcavated area at the Porto site in Santarém revealed fish, aquatic mammals and chelonian remains (Lima, 2024; Silva and Pereira, 2021).

When considering aquatic resources it is important to note that they are generally more available (concentrated) when the waters recede. However, little is yet known about the seasonality of archaeological sites in the floodplains and whether resource acquisition was practiced exclusively by groups occupying the environment, or was also a factor for other short-term visits to the region (e.g., fishing stations). From modern observations we can state that aquatic resources are of great importance to various populations of the Amazon, both residing on and off of the floodplains. Thus, there are many questions about the relationships, such as those of exchange, that may have been established between sites located on the floodplains and those on the bluffs or, alternatively, if there were independent groups that had some other contacts. The partial, physical separation of sites with Tapajó ceramics, concentrated South of the Amazon River, and those with Konduri style, to the North, presents the possibility that the river corridor was a place for encounters, both productive and bellic. Further questions arise from observing the flood regime and modern houses, such as the types of architecture of archaeological houses if sites were permanent residences (e.g., stilt houses), and how permanent residence played into the relationships between groups that preferred different types of ceramics. The importance of aquatic environments and their occupations should be explored to better comprehend mobility and territoriality over time in the Amazon floodplain.

While we have detailed the pre-European invasion archaeological data for the floodplain, it is important to remember that the human presence has been unbroken. There are indigenous communities, traditional river-dwelling communities, and a further history within the African descendant communities (quilombos). During the historical period, the plantations associated with the 19th-century slave system used the floodplains, and there has been extensive exploitation of floodplains for jute production (Corchorus capsularis) as well as cattle and buffalo ranching (Py-Daniel et al., 2023). In sum, the floodplains have multiple histories that need to be discovered.

3 Materials and methods

3.1 Remote sensing

Our initial analysis of floodplain depositional episodes employed Landsat satellite imagery [from Landsat 5 Thematic Mapper (TM) sensor and Landsat 8, Operational Land Imager (OLI) sensor], characterized by a 30-m spatial resolution (Campos et al., 2024). In order to refine the treatment of vegetation, specifically to differentiate between vegetation types, a data source with higher spatial resolution was necessary. We chose CBERS-4A satellite data from Brazil's National Institute for Space Research (Instituto Nacional de Pesquisas Espaciais - INPE) catalog, which provides free and processed images (INPE, n.d.). The Multispectral and Panchromatic Wide-Scan Camera (WPM) was chosen from available sensors⁴. This imagery has a spatial resolution of 8 m for the multispectral bands and 2 m for the panchromatic band. Digital Elevation Model (DEM) topographic data were not used due to errors provoked by closed canopy forest and deviations in the readings themselves approaching 16 m (Gillespie et al., 2006; Hofton et al., 2006). Given the low variations in floodplain terrain these data would not be useful.

Given that the satellite was launched in December of 2019, the imagery searched for appropriate scenes dates from 2020 onwards. To ensure optimal visualization of dry season vegetation the imagery examined was narrowed to the period from September to November when the water levels are lowest, and plant stress due to drought is the greatest. Scene selection was based on visual observation of cloud cover, a crucial consideration for optical satellite imagery such as that provided by CBERS. The selected scene (orbit/path 219, point/row 116), captured on September 1st, 2023, exhibited minimal cloud cover within the study area. All geospatial processing was conducted utilizing the open-source QGIS LTR 3.34 Prizren software.

We chose to use a vegetation index, the Normalized Difference Vegetation Index (NDVI)⁵, to assess whether differences in vegetation spectral response would allow the separation of vegetation types by their response to soil and topographic conditions. The hypothesis was that different plants would be distributed according to differences in relief (more specifically, the elevation of the terrain that controls for the portion of the year that vegetation spends under flood conditions as well as the height of the water table in the soil profile) and that their photosynthesis processes would be indicators of vegetation groups. We chose NDVI (indicator of canopy structure and vegetation greenness, sensitive to factors such as vegetation density) because it is widely used in remote sensing, with well-established results in various areas of application (Carlson and Ripley, 1997; Lu et al., 2004; Palace et al., 2017; Wang et al., 2005; Xue and Su, 2017; Yang et al., 2012; Zhu and Liu, 2015). The NDVI formula is . To visualize the results, the rendering type was false-color single band with linear interpolation, Turbo color gradient, label accuracy 4, continuous mode, 9 classes (Figure 4).

Figure 4

Based on the hypothesis of separating vegetation types by NDVI value, we projected that the higher NDVI values would be associated with forested areas (denser vegetation and more closed canopies) and that average values could represent other types of floodplain vegetation, including fields and areas that spent more time submerged. We therefore sought to verify these differences in the field by separating plots of vegetation with ‘high’ NDVI values from those with ‘average’ NDVI values.

For the chosen scene (9/1/2023), the NDVI calculation produced values between −0.3675889 and 0.7604396 (Figure 4). We used QGIS to select the pixels with values above 0.65 to highlight the distribution of the highest 5% of NDVI values, where the tree formations might be present. To visualize the distribution of these values, we created a binary mask in which all NDVI values > 0.65 (in the extension of the plots studied) were represented by the number 1 (positive) and all other values were represented by 0. Visually, it was possible to see that these distributions have linear shapes, suggesting that they could accompany some part of a geomorphic structure such as a fluvial levee. This raster layer was vectorized, thus drawing high NDVI vegetation areas from which it would be possible to select random points (Figure 5). We used QGIS random point selection tool to generate 10 random ground-truthing points at a minimum inter-point distance of 500 m to ensure spatial independence within each vegetation category. These points, representing potential levee forest vegetation, were labeled D01 to D10 (levee in Portuguese is dique, hence the capital letter D). Subsequently, another set of ten random points was generated within the ‘average’ NDVI range of 0.35–0.55 (representing the 70th to 90th percentile⁶), that was hypothesized to represent non-levee vegetation, and labeled ND21 to ND30 (Figure 6). These selected points were then transferred to a Garmin eTrex® 10 GPS device for subsequent ground-truthing.

Figure 5

Figure 6

3.2 Fieldwork

Fieldwork was conducted March 11–19, 2024. This period of the flood season was selected to ensure adequate accessibility by boat while maintaining sufficient land exposure for archaeological investigation. While differences in elevation would be more noticeable in the dry season, over-land access would be even more difficult and slower than access by boat. The primary author established a dialog with the Urucurituba community in the beginning of 2024 and upon arriving to conduct the fieldwork, held a meeting to share the research questions and aims. Two of the co-authors (José Reinaldo Campos Garcia and Antônio Gilmar Ribeiro Nogueira Filho) are community members who provided invaluable knowledge and assistance by guiding ground-truthing activities and directing the team to an archaeological site. Although it was not always possible to determine whether we were on a levee by visual criteria, our co-authors (Garcia and Nogueira Filho) who are local fishermen have a more intimate knowledge of the topography. They explained when we were on a rise with expressions such as “This is the last bit that gets submerged”.

The initial research design envisioned a quantitative, statistical approach with a target sample size of 20 data points, 10 in each of the zones by NDVI values. However, logistical constraints encountered during fieldwork led to a modification of the methodology. Significant time was consumed by travel, exacerbated by inclement weather conditions (rain) and challenging terrain (muddy conditions). Given the spatial extent of the study area and the associated travel demands (time and fuel consumption), it became impractical to collect data from all predetermined locations. We prioritized the points with the highest NDVI values (D01 to D10), which supposedly represent high forest vegetation on fluvial levees. This choice was also due to the archaeological objectives of the investigation.

The procedures for ground-truthing and archaeological survey consisted of: (1) environmental characterization (morphology of the terrain, estimated vegetation height, and distance to nearest water body) with photographic documentation (photos facing to the north/west/south/east taken from a single position at each location to show the surroundings); (2) unsystematic surface survey and collection (with provenience numbers) around the ground-truthing location and (3) an auger sample.

The auger sampling was performed to search for buried archaeological material and in order to collect sediment. Based on the GPS point, the decision was to dig in the highest portion of the area. Excavation proceeded in 20 cm depth intervals and, from each level, a sample of soil for basic characterization was collected (500 mL collected in a 1 kg plastic bag). For each level soil color (Munsell), texture, and bioturbations were noted on the recording form along with sample identification numbers. Texture was estimated by touch in laboratory, using the ball-filament-ring technique (Historic England, 2015). All soil was examined for possible archaeological remains. Due to equipment limitations, auger sampling terminated at a depth of 120 cm. Given the presence of archaeological materials on the surface, the sampling strategy of a 70 × 100 cm testpit excavated in 10 cm levels was established to acquire a larger sample of cultural remains. Subsequent to cultural levels, past two sterile levels, the auger was used to identify underlying strata.

Although we were prioritizing the levee (Dique) points, after a team dialogue, we decided that given time constraints it would not be worth checking out point D08 (a long way from our research base in Urucurituba). Instead, we decided to invest time in following up on the indication that the Bom Jesus locality (called BJ) was an archaeological site and to observe its topographic placement. Point D03 could not be properly checked due to access difficulties (more details in the Supplementary Material).

The production of a figure with soil texture data, Figure 7, was based on generative AI technology (ChatGPT, GPT-5.5 model; OpenAI, San Francisco, CA, USA) as detailed in the Supplementary Material (1. Ground-truthing data: Prompt used to generateFigure 7).

Figure 7

4 Results

4.1 Ground-truthing

The results of ground-truthing (Figure 6) have been condensed in Table 2 to summarize the locations that were visited and to facilitate understanding and analysis. The BJ point was included in the table because it was identified as part of the archaeological survey. During data collection, it was possible to record the locations’ names and the topographical knowledge of our community collaborators. A selection of images that illustrate the visible topography of the locations are presented in Figure 8. Photos that assist in the visualization of all locations and predominant vegetation are in the Supplementary Material along with the details of auger testing that are summarized in Figure 7 (Supplementary Table 1).

Figure 8

Buried Fluvial Levee (D06): The sample contains fine-grained sediments, including sandy loam and silty clay, overlying loose sand between 60 and 120 cm (Figure 9). This inverted stratigraphy deviates from typical levee aggradation and is more consistent with burial by overbank fines resulting from channel migration or a change in flood regime (Johnston et al., 2019). Mottling observed in the upper 40 cm suggests impeded drainage due to low-permeability, whereas the underlying sandy substrate remains aerobic and well-drained (Filgueira-Rivera et al., 2007; Vepraskas, 2015). The textural inversion at 40–60 cm likely represents a former levee surface. However, borehole data alone are insufficient to distinguish between burial, bioturbation, or in situ pedogenesis.

Figure 9

Fluvial Levees (D02 and D04): Both points have a substrate dominated by silty clay and are slightly moist throughout. Redoximorphic features are confined to the surface horizon (0–20 cm). At D02, brown sediment (10YR 4/3) is mottled with very dark grayish brown (10YR 3/2); at D04, brown (7.5YR 4/2) is mottled with red (2.5YR 4/6). These features indicate episodic, near-surface reducing conditions associated with seasonal inundation rather than prolonged saturation (Ferguson et al., 1999; Vepraskas, 2015). Below 20 cm, uniform colors without mottling indicate predominantly aerobic conditions (Bhattacharyya et al., 1998). During ground-truthing, both D02 and D04 were located on elevated terrain adjacent to a river channel, consistent with fluvial levees and currently covered by floodplain forest. D04 is texturally uniform, whereas D02 exhibits a fining-upward sequence, from clayey sand at depth to clay and silty clay toward the surface. This vertical trend is consistent with overbank deposition and levee aggradation (Branß et al., 2022; Johnston et al., 2019). Compaction decreases with depth at both points. The combination of limited surface mottling and aerobic subsurface conditions indicates moderately well-drained to well-drained profiles. At D02, the stratigraphic trend and field position converge in supporting a levee interpretation. At D04, the field position is consistent with a levee setting, but the absence of a clear vertical trend makes this interpretation less congruent.

Geomorphic variability within areas associated with high floodplain forest (D01, D05, D07, D09, D10): During ground-truthing, five points were located in areas identified by our local informants as fluvial levees and are currently covered by high floodplain forest. Elevation differences among these points are subtle or imperceptible in the field, and the soil profiles show marked variability. Across these samples, soil properties indicate substantial hydromorphic variability. Only D09 and, to a lesser extent, D07 show profiles compatible with relatively well-drained, elevated positions. D05, D10, and D01 exhibit stronger or more persistent saturation, suggesting lower or transitional settings within the floodplain. The correspondence between vegetation, traditional knowledge, and geomorphic position is therefore partial rather than systematic.

D09 and D07 show comparatively well-drained conditions. At D09, mottling is restricted to 20–40 cm, indicating brief and localized redox fluctuations; the remainder of the profile is uniform, suggesting predominantly aerobic conditions (Vepraskas, 2015). At D07, red mottles appear below 40 cm and persist downward, consistent with seasonal water table oscillation; their onset coincides with a sandy loam horizon (20–60 cm), which may facilitate lateral water movement (Ferguson et al., 1999; Vepraskas, 2015). Drainage in both cases is interpreted as moderately well-drained (Filgueira-Rivera et al., 2007). These profiles are compatible with relatively elevated positions within the floodplain. Their occurrence in high floodplain forest and their identification as levees in traditional knowledge are consistent with this interpretation, though not independently diagnostic.

D05 and D10 exhibit more extensive redoximorphic development. At D05, mottling spans 20–80 cm, with progressive matrix lightening and iron concentration, indicating increasing saturation depth and intensity (Bhattacharyya et al., 1998; Vepraskas, 2015). At D10, mottling extends from the surface to 60 cm, indicating active redox cycling in the upper profile (Bhattacharyya et al., 1998; Vepraskas, 2015). In both cases, colors become more uniform below the mottled interval, suggesting relatively improved aeration at depth, although prolonged anaerobic conditions cannot be excluded at D05. Drainage is interpreted as imperfect (Filgueira-Rivera et al., 2007). Despite their association with high floodplain forest and local classification as levees, these profiles indicate conditions consistent with lower or transitional floodplain positions subject to sustained seasonal saturation.

D01 presents a stratified redox pattern. Strong gleying between 20 and 60 cm indicates prolonged saturation and reducing conditions (Bhattacharyya et al., 1998; Vepraskas, 2015). Between 60 and 90 cm, a brown matrix mottled with gray indicates a fluctuating redox boundary (Ferguson et al., 1999). Below 90 cm, uniform brown sediments indicate improved aeration. The surface horizon (0–20 cm) shows a dark, organic-rich sequence consistent with forest cover (Kraus and Aslan, 2009; Martins-Oliveira et al., 2020). Drainage varies vertically, from imperfect to poor in the middle horizons to moderately well-drained at depth (Filgueira-Rivera et al., 2007).

The only ground-truthing points among the 10 in the area of ‘average’ NDVI (0.35 < NDVI < 0.55) visited were points ND26 (could not be auger tested because it was flooded) and ND27. Both had shrubby vegetation, grasses, and palm trees. This open vegetation is different from that observed at the ‘high’ NDVI (NDVI > 0.65) points. Further details on these points can be found in the Supplementary Material.

4.2 Additional ground-truthing of an archaeological site

At the site of Bom Jesus, preliminary investigations were undertaken on March 12 and 19, 2024. Our local co-authors reported archaeological material at Bom Jesus. Upon our team's arrival at the location, the local resident showed us material he had collected from the site: fragmented and partially intact ceramic figurines, incised rim sherds, appliqués, a perforated disk, and polished stone axe blades (Figure 10). Unsystematic surface survey revealed ceramic material in varying concentrations. Archaeological material was observed on the surface of an area of approximately 200 × 50 m. In a heavily forested area, where the soil surface was covered with roots and fallen leaves, we found the most remains, including axe blades. Observation of the area led to our distinguishing between two differences in the terrain: in elevation and in the vegetation.

Figure 10

In Figure 11 it is possible to observe that one portion of the site is higher and has dense forest vegetation, while the other area is lower and has shrubby vegetation. While both areas contain archaeological remains there is a difference in ceramic density, elevated on the higher ground (Figure 12), and fragmentation, that is greater in the lower area (Figure 13) but included a zoomorph (Figure 14) that was collected (Supplementary Material Figure S26).

Figure 11

Figure 12

Figure 13

Figure 14

After discussing archaeology with the local resident, they granted us permission to verify the presence of buried archaeological deposits. The potential for these was given by the material on the surface. We aimed to collect a small sample of the ceramic material and, if possible, material for dating (charcoal or an OSL tube). The test pit was opened where an axe fragment was on the surface (Supplementary Material Figure S28) and oriented based on the space between the trees. Archaeological remains were limited to the upper 20 cm of the test pit (Figure 15).

Figure 15

BJ shows a sandy silt loam texture at 0–40 cm, transitioning to sandy loam between 40 and 140 cm. Surface horizons (0–40 cm) are uniformly dark (10YR 3/3–10YR 4/2) and lack mottling, with slightly moist conditions at 0–20 cm grading to dry below. In contrast, redoximorphic features are present from 40 to 140 cm, with recurring mottling between gray to brown matrices (10YR 4/1–10YR 5/3) and yellowish brown concentrations (10YR 5/6–10YR 5/8). These features indicate a fluctuating water table affecting the lower profile, while the unmottled upper horizons reflect well-drained conditions near the surface (Filgueira-Rivera et al. 2007; Vepraskas, 2015). Compaction is strongest in the upper 80 cm, decreasing with depth. This pattern may reflect the anthropogenic modification associated with occupation, but post-depositional loading and natural consolidation remain plausible alternatives in the absence of micromorphological data (Corrêa et al. 2003; Guccione et al. 1988). Periodic water table rise into the lower profile indicates that this stability was seasonal rather than permanent (Guccione et al., 1988; Mott, 2022). This pattern may reflect pre-occupation conditions, although this cannot be demonstrated without independent chronological control. The combination of sandy texture, well-drained surface horizons, and elevation above seasonal saturation is consistent with a levee setting.

The site of Bom Jesus is located on a fluvial levee, which demonstrates the potential of this topographic position to contain archaeological remains. The site was indicated to us by locals’ knowledge, but it is important to observe that it could have been among the randomly selected points in the ‘high’ NDVI area as it falls within this portion of the map.

5 Discussion

The vast and dynamic nature of the Amazon floodplain poses significant challenges for conducting archaeological surveys. However, this should not prevent us from carrying out research that seeks to enhance our knowledge of the relationships that human groups established with this ecosystem for millennia. The archaeological record is a precious source of information in this regard and needs to be approached with methodologies that are capable of working with the geomorphological particularities of the floodplains. This method of mapping fluvial levees using vegetation remote sensing, based on NDVI, represents an initial step toward developing these techniques.

The initial purpose of our research was to ground-truth this methodology's ability to map levees. The ground-truthing results indicate that NDVI > 0.65 captures sectors of the floodplain that include levee-like settings, but also encompass lower and transitional environments. Its utility as a predictor of geomorphic position or human occupation potential is therefore limited and requires further validation. Although we were unable to verify all 20 points, these preliminary results are promising. Furthermore, it is crucial to recognize that the scarcity of archaeological materials could also be linked to the limited number of random points and the depth of augering. Our experience in attempting to cover a large study area within a short time frame provided valuable insights into the investment needed for exploring floodplains. In areas with dense vegetation, more time is required for systematic studies. Therefore, additional ground-truthing should be conducted.

Although there are limitations, the archaeological material indicated to us by the local population in Bom Jesus is situated on a fluvial levee. This suggests that it is indeed possible to discover evidence of pre-colonial human occupations associated with this type of fluvial geomorphological formation. Thus, while still requiring broader testing, the hypothesis linking levees with archaeological contexts remains plausible. The floodplain is an ecological zone marked by biodiversity (Wittmann et al., 2006). In the future it may be possible to consider individual plant species that might be associated with human presence or archaeological sites (Balée, 1989) due to their economic and cultural relevance, such as the sapucaia nut (food), cuieira (artifact/craft), marajá and catauari (fishing), apuí (medicinal), tachi (construction), etc. Further exploration of vegetation may show that specific plants, alongside terrain, can better guide archaeological surveys.

The auger tests demonstrated that the locations visited (eight ‘high’ and one ‘average’ NDVI points, and the Bom Jesus locality) were subject to recent sedimentation in the Amazon floodplain. All auger tests began in clay-rich sediments consistent with floodwater deposition. Only at D06 did the 1 m profile reach underlying sandy material, suggesting a higher-energy depositional context possibly associated with a former levee. This contrast indicates that earlier floodplain surfaces may be buried beneath fine overbank deposits in parts of the study area. However, the limited depth of the auger tests constrains the detection of such features, and their depth and continuity remain unresolved. Although additional geomorphological research is required to fully understand paleolevee transformation under contemporary fluvial dynamics, levee morphology and, in certain instances, internal stratigraphy remain discernible. At D02, both stratigraphic and field-position evidence support a levee interpretation. In contrast, identification at D04 depends mainly on relative elevation, resulting in a less constrained interpretation. Furthermore, microrelief is partially expressed in vegetation structure, as areas occupying higher floodplain positions generally support plant communities exposed to less frequent or shorter periods of inundation. This pattern is evident in certain high-NDVI locations, including the Bom Jesus locality.

In other Brazilian fluvial contexts, it has been observed that the fluvial landforms are more clearly visible than those of the Amazon floodplains (Fernandes et al., 2023). The Amazon River is one of the world's largest sediment carrying rivers, and is the largest in South America in this category (Fagundes et al., 2021). This means that newly transported sediments are transforming the topography of ancient landforms (Ashworth and Lewin, 2012; Dunne and Aalto, 2013; Lewin and Ashworth, 2014). Remembering that the Amazon, with the highest flow rate in the world (200,000 m³ s⁻¹) according to Tandon and Sinha (2007), covers extensive areas during floods we expect that over time the morphologies of older levees will have been smoothed by the subsequent deposition of finer sediments from lower velocity floodwaters.

A study by Fernandes et al. (2023) along the Paraná River demonstrates that vegetation structure is linked to geomorphology and hydrology, and is a factor that influences sediment accumulation during the formation of fluvial levees and during floods. From this we understand that paleolevees will have variable heights in part due to the past existence of vegetation on them, and subsequent to their abandonment as levees, when they become part of the broader floodplain, this vegetation can lead to greater accumulations of sediments from floodwaters. The factor of vegetation may thus assist in preserving the variable relief of levees, but this is not always the case, and over successive floods, the levees may be hidden or topographically smoothed. All of the points with ‘high’ NDVI values were located in floodplain forest and the auger tests revealed layers of clay sediments from floods. Since in only one of the locations the sediments included sands, there is some confusion as to how much of the modern relief is a product of the forest vegetation that promotes a greater accumulation of sediments (aggravating or exaggerating previous landforms) and how much is the result of a paleolevee. Our present research indicates that, in the Lower Amazon, we still need to explore the conditions under which floodplain sediments either smooth or exaggerate prior landforms. An analysis of sedimentary deposition in the study area was previously conducted, but to test its validity, it is necessary to date the sediments (Campos et al., 2024). In this fieldwork, it was not possible to complete this step, but we can affirm that progress has been made in that we were able to conduct an initial verification of the location of levees and the existence of archaeological sites in the study area (initiating a more empirical phase of the research), which also opened up new avenues of investigation. We still need to correlate the history of landscape formation with the sedimentary record.

Two factors influenced the archaeological outcome of ground-truthing: the very small number of auger tests and their limited depths. The archaeological site that we were able to include in this research was based on the community members’ knowledge. Visiting this site, Bom Jesus, we observed archaeological material on the surface of an area of approximately 1 ha. This corresponds to a fluvial levee and its margins, but that the sediments containing archaeological materials are clayey, indicating that human occupation does not correspond to the period when the levee was formed. However, given the location of the site on a part of the floodplain that is thought to be older (Campos et al., 2024), there is a possibility for many periods of occupation. In fluvial environments there is always some doubt about the local or introduced origin of archaeological material due to possible fluvial transportation. Lathrap (1968, p. 71) suggests that the degree of pottery fragmentation and the rounding of the breaks can help to investigate this question of provenience. However, the texture of the sediments—clay—provides us with information about the energy of the fluvial system at the time of ceramic deposition.

The Bom Jesus site may be smaller than the total extent of artifact distribution since it was possible to observe in the area of the levee that the pottery sherds were larger/more intact, while on the lower slopes of the terrain there were many small sherds (measuring 2 cm or less), an indication of local erosion from the elevated area to lower areas or of some type of differential discard. Meanwhile, the fine-grained sediments indicate that the energy in the fluvial system, where the waters had already lost their load of suspended sands and silts, was low, making it unlikely that ceramics and lithics atop the levee were moved by a flood to the site. The extent of the deposit and the size of artefacts are indicative of an occupation. Geoarchaeological analysis will undoubtedly be necessary to determine if the cultural materials are contained within the original boundaries of this site or have expanded to a greater area by subsequent transportation/erosion. Areas with higher densities of ceramic material on the surface (Figure 12) are suggestive of little sediment deposition after occupation or possibly moments of local erosion as might have been provoked by recent human processes such as ranching. The possibility of recent erosion of the site is supported by the greatest densities of materials in areas with roots that assist in securing both the soil and material contained within it. Bom Jesus is the first site in the Alenquer floodplain known to the academic community. As residents of the Urucurituba community pointed to the existence of other areas with archaeological material, a future field season is envisioned to better understand the chronology of this and other sites and their relationships with the landscape. Across the Amazonian floodplains the known archaeological sites include shellmounds, dark earths and artificial islands, each with their own set of formation processes that need to be understood in concert with fluvial geomorphology.

The utilization of the floodplains during the last centuries may have had effects that further complicate the reading of this area's geomorphological history, and thus whether models can assist in archaeological survey. Further elaboration of methods that may direct survey should also reflect upon the dynamics from these more recent activities. These may have provoked erosion or other landform transformations. In the 18th and 19th centuries the fertile lands of the floodplains were used to plant cocoa (Theobroma cacao) and rubber, although these activities were more concentrated on terra firme. Meanwhile, the 20th century brought intensive exploitation with the cultivation of jute (Harris, 2011; Winklerprins, 2006). Additionally, buffalo and cattle ranching expanded, taking advantage of the floodplain fields as pasture (Sheikh et al., 2006). Agriculture and cattle ranching are practices that have greatly contributed to deforestation in Amazonia as a whole (Ayala et al., 2016; Hecht, 1993). The work by Renó and colleagues (Renó et al. 2011, Renó and Escada 2016) based on remote sensing showed that there was a significant decrease in the vegetation cover of the Lower Amazon floodplains between 1975 and 2008, attributable, probably, to ranching and farming activities. In most riverine communities fishing is the actual source of income for families, replacing agriculture, which had previously been commercial in nature; crops are now planted for personal consumption, for example watermelon (Citrullus lanatus), corn (Zea mays), beans (Phaseolus vulgaris), manioc (Manihot esculenta), etc. (McGrath et al., 1993; Winklerprins, 2006, p. 832). Subsequent economic activities that have negatively impacted the floodplain ecosystem are overfishing and mining (Castro and McGrath, 2003; McGrath et al., 1993; Meneses et al., 2022). In addition, alongside all of the changes in how the land is utilized, the impacts of climate change are becoming more visible in the Amazon Basin (Arias et al., 2020; Best; 2019; Bodmer et al., 2018; Correa et al., 2022; Espinoza et al., 2024; Santos de Lima et al., 2024).

Researching the floodplains from an archaeological point of view means trying to understand issues related to the mobility and territoriality of human groups, the insertion of occupations and settlements in the landscape and whether there were preferences for occupying/exploiting certain types of locations in the floodplains over time. The plausible importance of aquatic resources to human societies should lead to reflections on the possibilities of management that could have occurred in the floodplains, how management practices could have altered the environment and, consequently, the possibility that it is a modified landscape (Prestes-Carneiro et al., 2021). Hence, there is a substantial amount of research that remains to be done on the Amazonian floodplains to address the questions posed for regional archaeology (mentioned in section 2.1), for example, and Amazonian archaeology as a whole. Future research should be based on both the knowledge of the floodplain dwellers and surveys that can be directed by modeling.

The Late Holocene is marked by being a period of population densification, with many occupations in different areas of the Amazon at the same time, as exemplified by the Tapajó occupation in western Pará. The hypothesis that, during this period, human groups were also occupying and living in these large floodplains, exploiting and managing the vast amount of resources available along the world's largest freshwater river, the Amazon, is perfectly possible.

6 Concluding remarks

To assess whether mapping fluvial levees is useful for archaeological survey, we used the vegetation index NDVI to distinguish between two categories that were hypothesized to be linked to topographic positions: high floodplain forest in elevated areas (probable levees) and other types of vegetation situated in lower areas (such as flooded fields). Adverse conditions in the field did not allow us to test the model according to scientific standards of statistical rigor. However, our results have qualitative relevance, as an exploration of an area hitherto unknown archaeologically that allowed us to record an archaeological site known to floodplain dwellers and to strengthen our hypothesis for the locations of fluvial levees.

The archaeological record of the Amazon floodplains can contribute to Amazonian archaeology once survey has encountered the existing sites. Our model can be improved, in the future, by the incorporation of other vegetation indices and more extensive testing. As the modern uses of the floodplain vary between people who are residents of permanent communities or those who use resources seasonally, we expect that the archaeological record will expose diverse occupations. The value of applying modern tools and methodologies to uncover new perspectives on ancestral histories has been shown by studies that are taking place across Amazonia (e.g., Peripato et al., 2023; Prumers et al., 2022; Rostain et al., 2024; Souza et al., 2018).

To gain a clearer understanding of the formation of the floodplain landscape and how fluvial geomorphological processes impact archaeological sites, it is necessary to expand knowledge of the sedimentary deposition history of the floodplain, as well as to understand the formation processes of fluvial levees. In addition to the need for deeper excavations (using appropriate equipment) and increasing the number of auger tests across all plots, we suggest the use of more advanced technologies to obtain a more precise topographic survey of the localities, such as LiDAR. This knowledge can help us further comprehend the archaeological sites already known in the floodplain and also support future surveys.

Archaeology provides an opportunity to understand the ancestral management practices that helped to shape the Amazon's rich biodiversity if we can locate sites. Thus, there is increasing worry that this history may be lost to a system dominated by the greed of a few humans, who seem incapable of noticing the harm they are doing to the living beings.

References

• 1

  AlmeidaC. T.Oliveira-JúniorJ. F.DelgadoR. C.CuboP.RamosM. C.(2016). Spatiotemporal rainfall and temperature trends throughout the Brazilian Legal Amazon, 1973–2013. Int. J. Climatol. 37 (4), 2013–2026.

  • Google Scholar

• 2

  AlvaresC. A.StapeJ. L.SentelhasP. C.GonçalvesJ. D. M.SparovekG. (2013). Köppen’s climate classification map for Brazil. Meteorologische Zeitschrift22 (6), 711–728. 10.1127/0941-2948/2013/0507

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AlvesM. L. (2018). Para além de santarém: os vasos de gargalo na bacia do rio trombetas. Boletim do museu Paraense emílio goeldi. Ciências Humanas13, 11–36. 10.1590/1981.81222018000100002

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  AlvesM. L. (2020). Revisitando os alter egos: figuras sobrepostas na iconografia konduri e sua relação com o xamanismo. Boletim do museu Paraense emílio goeldi. Ciências Humanas15 (3), e20190105. 10.1590/2178-2547-BGOELDI-2019-0105

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  AraújoG. L. H.LimaA. M. A. (2023). Narrativas sobre o modo de vida dos povos amazônicos do passado e do presente em comunidades do médio solimões. Revista Arqueologia Pública18, e023004–e023004. 10.20396/rap.v18i00.8670853

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  AriasM. E.FarinosiF.LeeE.LivinoA.BriscoeJ.MoorcroftP. R. (2020). Impacts of climate change and deforestation on hydropower planning in the Brazilian Amazon. Nat. Sustain.3 (6), 430–436. 10.1038/s41893-020-0492-y

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  AshworthP. J.LewinJ. (2012). How do big rivers come to be different?Earth Sci. Rev.114 (1-2), 84–107. 10.1016/j.earscirev.2012.05.003

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  AyalaL. M.van EupenM.ZhangG.Pérez-SobaM.MartoranoL. G.LisboaL. S.et al (2016). Impact of agricultural expansion on water footprint in the Amazon under climate change scenarios. Sci. Total Environ.569, 1159–1173. 10.1016/j.scitotenv.2016.06.191

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  BaléeW. (1989). The culture of amazonian forests. Adv. Econ. Bot.7, 1–21.

  • Google Scholar
  • View reference in article

• 10

  BaléeW. (2018). Brief review of historical ecology. Les Nouvelles de L'archéologie152, 7–10. 10.4000/nda.4150

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  BarbozaR. S. L.BarbozaM. S. L.PezzutiJ. C. B. (2013). Estava pescando de malhadeira, vi na Praia uns cascos brilhando, era luar, abeirei a terra e fui pegar": práticas de pesca de quelônios na várzea amazônica (santarém-PA). Amazônica-Revista de Antropologia5 (3), 622–653. 10.18542/amazonica.v5i3.1583

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  BarretoC.AlvesM. (2023). “Birds of prey in ancient Amazonia: predation and perspective in ceramic iconography,” in The Art and Archaeology of Human Engagements with Birds of Prey: From Prehistory to the Present, ed. WallisR.J. (London: Bloomsbury Academic), 152–165. 10.5040/9781350268029.ch-009

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  BestJ. (2019). Anthropogenic stresses on the world’s big rivers. Nat. Geosci.12 (1), 7–21. 10.1038/s41561-018-0262-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  BhattacharyyaT.PalD. K.MondalC. (1998). Soil formation as influenced by geomorphic processes in the Brahmaputra flood plains of Assam. J. Indian Soc. Soil Sci.46 (3), 351–357.

  • Google Scholar
  • View reference in article

• 15

  BodmerR.MayorP.AntunezM.ChotaK.FangT.PuertasP.et al (2018). Major shifts in Amazon wildlife populations from recent intensification of floods and drought. Conserv. Biol.32 (2), 333–344. 10.1111/cobi.12993

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  BragaP. I. S. (1979). Subdivisão fitogeográfica, tipos de vegetação, conservação e inventário florístico da floresta amazônica. Acta Amazonica9, 53–80. 10.1590/1809-43921979094s053

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  BranßT.Núñez-GonzálezF.AberleJ. (2022). Fluvial levees in compound channels: a review on formation processes and the impact of bedforms and vegetation. Environ. Fluid Mechan.22 (2), 559–585. 10.1007/s10652-022-09850-9

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  CamposV. S.ShockM. P.Prestes-CarneiroG. (2024). Terras crescidas e terras antigas: novas aplicações do sensoriamento remoto à prospecção de sítios arqueológicos em várzeas amazônicas. Boletim do museu Paraense emílio goeldi. Ciências Humanas19, e20220059. 10.1590/2178-2547-BGOELDI-2022-0059

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  CarlsonT. N.RipleyD. A. (1997). On the relation between NDVI, fractional vegetation cover, and leaf area index. Remote Sens. Environ.62 (3), 241–252. 10.1016/S0034-4257(97)00104-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  CarneiroR. L. (1970). A theory of the origin of the state: traditional theories of state origins are considered and rejected in favor of a new ecological hypothesis. Science169 (3947), 733–738. 10.1126/science.169.3947.733

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  CastroF. D.McGrathD. G. (2003). Moving toward sustainability in the local management of floodplain lake fisheries in the Brazilian Amazon. Hum. Organ.62 (2), 123–133. 10.17730/humo.62.2.9bkh58xeekj6bg0m

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  ClementC. R.DenevanW. M.HeckenbergerM. J.JunqueiraA. B.NevesE. G.TeixeiraW. G.et al (2015). The domestication of Amazonia before European conquest. Proceed. Royal Soc. B: Biol. Sci.282 (1812), 10.1098/rspb.2015.0813

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  CorreaS. B.Van Der SleenP.SiddiquiS. F.Bogotá-GregoryJ. D.ArantesC. C.BarnettA. A.et al (2022). Biotic indicators for ecological state change in Amazonian floodplains. Bioscience72 (8), 753–768. 10.1093/biosci/biac038

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  CorrêaJ. C. S. L.BragaT. M. P.LauridoS. F. (2019). Usos de recursos faunísticos pelos moradores da comunidade boca do arapiri, assentamento agroextrativista atumã, em alenquer, pará, brasil. Amazônica-Revista de Antropologia11 (2), 741–769. 10.18542/amazonica.v11i2.7954

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  CorrêaM. M.KerJ. C.FontesL. E. F.MendonçaR. F.AzevedoA. C. (2003). Atributos físicos, químicos e mineralógicos de solos da região das várzeas de sousa (PB). Revista Brasileira de Ciência do Solo27 (2), 311–324. 10.1590/S0100-06832003000200011

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  CostaA. G. M. (2012). “Arqueologia da região do aritapera,” in Arqueologia, Patrimônio e Multiculturalismo na Beira da Estrada: Pesquisando ao Longo das Rodovias Transamazônica e Santarém-Cuiabá, ed. SchaanD. P. (Belém: GK Noronha), 93–97.

  • Google Scholar
  • View reference in article

• 27

  CunhaS. B. (2011). “Bacias hidrográficas,” in Geomorfologia do Brasil, 8th ed. eds. da CunhaS. B.GuerraA. J. T. (Rio de Janeiro: Bertrand Brasil), 229–271.

  • Google Scholar
  • View reference in article

• 28

  DagostaF. C.De PinnaM. (2019). The fishes of the Amazon: distribution and biogeographical patterns, with a comprehensive list of species. Bull. Am. Mus. Nat. Hist.2019 (431), 1–163. 10.1206/0003-0090.431.1.1

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  DantasM. E.TeixeiraS. G. (2013). “Origem das paisagens,” in Geodiversidade do Estado do Pará, eds. X. S. J. João, S. G. Teixeira, and D. D. F. Fonseca (Belém: CPRM), 23–52.

  • Google Scholar
  • View reference in article

• 30

  De SouzaJ. G.RobinsonM.MaezumiS. Y.CaprilesJ.HoggarthJ. A.LombardoU.et al (2019). Climate change and cultural resilience in late pre-Columbian Amazonia. Nat. Ecol. Evol.3 (7), 1007–1017. 10.1038/s41559-019-0924-0

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  De SouzaJ. G.SchaanD. P.RobinsonM.BarbosaA. D.AragãoL. E.MarimonB. H.Jret al (2018). pre-Columbian earth-builders settled along the entire southern rim of the Amazon. Nat. Commun.9 (1), 1125. 10.1038/s41467-018-03510-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  DenevanW. M. (1992). The pristine myth: the landscape of the Americas in 1492. Ann. Assoc. Am. Geograph.82 (3), 369–385. 10.1111/j.1467-8306.1992.tb01965.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  DenevanW. M. (1996). A bluff model of riverine settlement in prehistoric Amazonia. Ann. Assoc. Am. Geograph.86 (4), 654–681. 10.1111/j.1467-8306.1996.tb01771.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  DunneT.AaltoR. E. (2013). “Large river floodplains,” in Treatise on Geomorphology Fluvial Geomorphology, 1st edn. eds. ShroderJ. FWohlE (San Diego: Academic Press; Elsevier), 645–678. 10.1016/B978-0-12-374739-6.00258-X

  • Google Scholar
  • View reference in article

• 35

  EricksonC. L. (2010). The transformation of environment into landscape: the historical ecology of monumental earthwork construction in the Bolivian Amazon. Diversity. (Basel)2 (4), 618–652. 10.3390/d2040619

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  EspinozaJ. C.JimenezJ. C.MarengoJ. A.SchongartJ.RonchailJ.Lavado-CasimiroW.et al (2024). The new record of drought and warmth in the Amazon in 2023 related to regional and global climatic features. Sci. Rep.14 (1), 8107. 10.1038/s41598-024-58782-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  FagundesH. D. O.FanF. M.PaivaR. D.SiqueiraV. A.BuarqueD. C.KornowskiL. W.et al (2021). Sediment flows in South America supported by daily hydrologic-hydrodynamic modeling. Water Resour. Res.57 (2), e2020WR027884. 10.1029/2020WR027884

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  FergusonR. J.BrierleyP. A.BrierleyG. J. (1999). Levee morphology and sedimentology along the lower tuross river, south-eastern Australia. Sedimentology46 (4), 627–648. 10.1046/j.1365-3091.1999.00235.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  FernandesL. A. F.LeliI. T.StevauxJ. C.KawakitaK.RomagnoloM. B. (2023). Estrutura da vegetação e relação com os processos geomorfológicos e regime fluvial no arquipélago três ilhas, alto rio paraná. Revista Brasileira de Geomorfologia24 (1), 1–22. 10.20502/rbg.v24i1.2166

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  Filgueira-RiveraM.VendittiJ. J.HassanM. A.Leite GuerreiroI. (2007). Controls on natural levée development in the Columbia river, British Columbia, Canada. Sedimentology54 (4), 905–919. 10.1111/j.1365-3091.2007.00865.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  FrikelP. (n.d.). O Jauarí do Lago Grande do Curuá. Notas e observações acêrca de um sernambí e de uma estação indígena arqueológica. Unpublished reports on research from 1939 to 1941.

  • Google Scholar
  • View reference in article

• 42

  FurquimL. P.NevesE. G.ShockM. P.WatlingJ. (2023). “The constructed biodiversity, forest management and use of fire in ancient Amazon: an archaeological testimony on the last 14,000 years of Indigenous history,” in Global Ecology in Historical Perspective, eds. IkeyaK.BaléeW. (Singapore: Springer Nature), 259–281. 10.1007/978-981-19-6557-9_15

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  GillespieT. W.ZuttaB. R.EarlyM. K.SaatchiS. (2006). Predicting and quantifying the structure of tropical dry forests in south Florida and the neotropics using spaceborne imagery. Glob. Ecol. Biogeogr.15 (3), 225–236. 10.1111/j.1466-8238.2005.00203.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  GoldbergP.MacphailR. I. (2006). Practical and Theoretical Geoarchaeology. Malden, MA: Blackwell Publishing.

  • Google Scholar
  • View reference in article

• 45

  GomesD. M. C. (2011). Cronologia e conexões culturais na amazônia: as sociedades formativas da região de santarém - PA. Revista de Antropologia54 (1), 269–314. Available online at:http://www.jstor.org/stable/41825785 (Accessed October 2, 2024).

  • Google Scholar
  • View reference in article

• 46

  GuapindaiaV.LopesD. (2011). Estudos arqueológicos na região de porto trombetas, PA. Revista de Arqueologia24 (2), 50–73. 10.24885/sab.v24i2.327

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  GuccioneM. J.SaucierJ. L.ReynoldsM. D. (1988). Environmental constraints of human settlement in an evolving holocene alluvial system, the lower Mississippi valley. Geoarchaeol.: Int. J.3 (1), 65–84. 10.1002/gea.3340030105

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  HarrisM. (2011). “The floodplain of the Lower Amazon as a historical place,” in The Amazon Várzea: The Decade Past and the Decade Ahead, eds. Pinedo-VasquezM.RuffinoM.PadochC.BrondízioE. (Dordrecht: Springer), 37–51. 10.1007/978-94-007-0146-5_3

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  HechtS. B. (1993). The logic of livestock and deforestation in Amazonia. BioScience43 (10), 687–695. 10.2307/1312340

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  HilbertK. (2009). Uma biografia de peter paul hilbert: a história de quem partiu para ver a amazônia. Boletim do museu Paraense emílio goeldi. Ciências Humanas4, 134–154. 10.1590/S1981-81222009000100012

  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  HilbertL.AlvesD. T.NevesE. G.IriarteJ. (2023). A glimpse into shell mound builders’ diet during mid-to-late holocene on Marajó island. Veg. Hist. Archaeobot., 34 (2), 1–10. 10.1007/s00334-023-00930-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  HilbertP. P. (1959). Achados arqueológicos num sambaqui do baixo amazonas. Instituto de Antropologia e Etnologia do Pará26 (4), 578–578. 10.2307/278769

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  HilbertP. P.HilbertK. (1980). Resultados preliminares da pesquisa arqueológica nos rios nhamundá e trombetas, baixo amazonas. Boletim do Museu Paraense Emilio Goeldi. Antropologia n°75, 1–15. Available online at:https://repositorio.museu-goeldi.br/handle/mgoeldi/633 (Accessed October 2, 2024).

  • Google Scholar
  • View reference in article

• 54

  Historic England (2015). Geoarchaeology: Using Earth Sciences to Understand the Archaeological Record. Swindon, UK: Historic England. Available online at:https://historicengland.org.uk/images-books/publications/geoarchaeology-earth-sciences-to-understand-archaeological-record/ (Accessed March 18, 2026).

  • Google Scholar
  • View reference in article

• 55

  HoftonM.DubayahR.BlairJ. B.RabineD. (2006). Validation of SRTM elevations over vegetated and non-vegetated terrain using medium footprint LiDar. Photogramm. Eng. Remote. Sens.72 (3), 279–285. 10.14358/PERS.72.3.279

  • CrossRef
  • Google Scholar
  • View reference in article

• 56

  HudsonP. F. (2008). “Natural levees,” in Encyclopedia of Water Science, 1st edn. ed. TrimbleS. W. (Boca Raton, FL: CRC Press/Taylor & Francis), 763–767. 10.1201/NOE0849396274

  • CrossRef
  • Google Scholar
  • View reference in article

• 57

  Instituto Nacional de Pesquisas Espaciais (INPE). (n.d.). Image Catalog. Available online at:https://www.dgi.inpe.br/catalogo/explore(Accessed February 22, 2024).

  • Google Scholar
  • View reference in article

• 58

  IrionG. (1984). “Sedimentation and sediments of amazonian rivers and evolution of the amazonian landscape since pliocene times,” in The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin, ed. H. Sioli (Dordrecht: Springer Netherlands), 201–214. 10.1007/978-94-009-6542-3_7

  • CrossRef
  • Google Scholar
  • View reference in article

• 59

  IrionG.MelloJ. A. S. N.MoraisJ.PiedadeM. T. F.JunkW. J.GarmingL. (2010). “Development of the Amazon valley during the middle to late quaternary: sedimentological and climatological observations,” in Amazonian Floodplain Forests, eds. JunkW.PiedadeM.WittmannF.SchöngartJ.ParolinP. (Dordrecht: Springer), 27–42. 10.1007/978-90-481-8725-6_2

  • CrossRef
  • Google Scholar
  • View reference in article

• 60

  JohnsonB. (2014). Effects of pansharpening on vegetation indices. ISPRS. Int. J. Geoinf.3 (2), 507–522. 10.3390/ijgi3020507

  • CrossRef
  • Google Scholar
  • View reference in article

• 61

  JohnstonG. H.DavidM. A.SmithN. D. (2019). Connecting fluvial levee deposition to flood-basin hydrology. J. Geophys. Res.: Earth Surf.124 (2), 561–585. 10.1029/2019JF005014

  • CrossRef
  • Google Scholar
  • View reference in article

• 62

  JunkW. J. (1984). “Ecology of the várzea, floodplain of amazonian whitewater rivers,” in The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin, ed. H. Sioli (Dordrecht: Springer Netherlands), 215–243. 10.1007/978-94-009-6542-3_8

  • CrossRef
  • Google Scholar
  • View reference in article

• 63

  JunkW. J.da SilvaV. M. F. (1997). “Mammals, reptiles and amphibians,” in The Central Amazon Floodplain: Ecology of a Pulsing System, ed. W. J. Junk (Berlin, Heidelberg: Springer Berlin Heidelberg), 409–417. 10.1007/978-3-662-03416-3_21

  • CrossRef
  • Google Scholar
  • View reference in article

• 64

  JunkW. J.PiedadeM. T. F.SchöngartJ.Cohn-HaftM.AdeneyJ. M.WittmannF. (2011). A classification of major naturally-occurring amazonian lowland wetlands. Wetlands31, 623–640. 10.1007/s13157-011-0190-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 65

  JunkW. J.PiedadeM. T. F.SchöngartJ.WittmannF. (2012). A classification of major natural habitats of Amazonian white-water river floodplains (várzeas). Wetl. Ecol. Manag.20, 461–475. 10.1007/s11273-012-9268-0

  • CrossRef
  • Google Scholar
  • View reference in article

• 66

  KrausM. J.AslanT. M. (2009). “Palaeosol sequences in floodplain environments: a hierarchical approach,” in Fluvial Sedimentology VII, eds. SmithN. D.RogersJ. (Oxford: Blackwell), 303–321. 10.1002/9781444304190.ch12

  • CrossRef
  • Google Scholar
  • View reference in article

• 67

  LakshmiV.FayneJ.BoltenJ. (2018). A comparative study of available water in the major river basins of the world. J. Hydrol. (Amst)567, 510–532. 10.1016/j.jhydrol.2018.10.038

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 68

  LathrapD. W. (1968). Aboriginal Occupation and changes in river channel on the central Ucayali, Peru. Am. Antiq.33 (1), 62–79. 10.2307/277773

  • CrossRef
  • Google Scholar
  • View reference in article

• 69

  LathrapD. W. (1970). The Upper Amazon. London: Thames and Hudson.

  • Google Scholar
  • View reference in article

• 70

  LauridoS. F.BragaT. M. P. (2018). Caracterização da pesca na boca do arapirí, uma comunidade no assentamento agroextrativista atumã em alenquer, pará. DESAFIOS-Revista Interdisciplinar da Universidade Federal do Tocantins5 (4), 15–27. 10.20873/uft.23593652201854p15

  • CrossRef
  • Google Scholar
  • View reference in article

• 71

  LewinJ.AshworthP. J. (2014). The negative relief of large river floodplains. Earth Sci. Rev.129, 1–23. 10.1016/j.earscirev.2013.10.014

  • CrossRef
  • Google Scholar
  • View reference in article

• 72

  LimaA. M. A. (2018). A ecologia de assentamentos, interações sociais ameríndias e o contexto geográfico dos muiraquitãs no baixo amazonas. Cadernos do LEPAARQ (UFPEL)15 (30), 121–141. 10.15210/lepaarq.v15i30.13816

  • CrossRef
  • Google Scholar
  • View reference in article

• 73

  LimaA. M. A. (2024) Os tapajó enterravam seus mortos?: reflexões sobre tratamentos funerários entre os tapajó. Master thesis. Museu Paraense Emílio Goeldi.

  • Google Scholar
  • View reference in article

• 74

  LimaA. M. A.MoraesC. P.SáM. S. R. (2020). Os discos perfurados do período tapajônico: análise tecnológica e questões contextuais. Boletim do museu Paraense emílio goeldi. Ciências Humanas15, e20190104. 10.1590/2178-2547-BGOELDI-2019-0104

  • CrossRef
  • Google Scholar
  • View reference in article

• 75

  LimaE. M.SantosP. B.BragaT. P.McGrathD. G. (2019). A pesca de acari (pterygoplichthys pardalis) na várzea do baixo amazonas, pará, brasil: aspectos estruturais e socioeconômicos. Gaia Scientia13 (4), 70–85. 10.22478/ufpb.1981-1268.2019v13n4.48781

  • CrossRef
  • Google Scholar
  • View reference in article

• 76

  LimaM. d. N. (2022) Entrelaçando histórias: antigas formas de habitar os lagos do médio solimões. Doctoral dissertation. University of São Paulo, Museum of Archaeology and Ethnology.

  • Google Scholar
  • View reference in article

• 77

  LombardoU.Arroyo-KalinM.SchmidtM.HuismanH.LimaH. P.MoraesC. P.et al (2022). Evidence confirms an anthropic origin of Amazonian dark earths. Nat. Commun.13 (1), 3444. 10.1038/s41467-022-31064-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 78

  LuD.MauselP.Brondı´zioE.MoranE. (2004). Relationships between forest stand parameters and landsat TM spectral responses in the Brazilian Amazon basin. For. Ecol. Manag.198 (1–3), 149–167. 10.1016/j.foreco.2004.03.048

  • CrossRef
  • Google Scholar
  • View reference in article

• 79

  MacedoR. S.TeixeiraW. G.LimaH. N.SouzaA. C. G. D.SilvaF. W. R.EncinasO. C.et al (2019). Amazonian dark earths in the fertile floodplains of the Amazon river, Brazil: an example of non-intentional formation of anthropic soils in the central Amazon region. Boletim do Museu Paraense Emílio Goeldi. Ciências Humanas14 (1), 207–227. 10.1590/1981-81222019000100013

  • CrossRef
  • Google Scholar
  • View reference in article

• 80

  Martins-OliveiraA. C.BritoS. R.SantosM. L. (2020). Morphological, physical and chemical attributes of the soil profile to the right margin of the Paraguay river, baia da campina, cáceres (MT). Scientif. Electron. Arch.13 (10), 76–84. 10.36560/131020201076

  • CrossRef
  • Google Scholar
  • View reference in article

• 81

  McGrathD. G.De CastroF.FutemmaC.de AmaralB. D.CalabriaJ. (1993). Fisheries and the evolution of resource management on the Lower Amazon floodplain. Hum. Ecol.21, 167–195. 10.1007/BF00889358

  • CrossRef
  • Google Scholar
  • View reference in article

• 82

  MeggersB. J. (1971). Amazonia; man and Culture in a Counterfeit Paradise. Chicago: Aldine.

  • Google Scholar
  • View reference in article

• 83

  MeggersB. J.DiasO. F.MillerE. T.PerotaC. (1988). “Implications of archaeological distributions in Amazonia,” in Proceedings of a Workshop on Neotropical Distribution Patterns, eds. P. E. Vanzolini and W. R. Heyer (Rio de Janeiro: Academia Brasileira de Ciências), 275–294.

  • Google Scholar
  • View reference in article

• 84

  MeggersB. J.EvansC. (1961). “An experimental formulation of horizon styles in the tropical forest area of South America,” in Essays in pre-columbian art and Archaeology, ed. S. K. Lothrop (Cambridge, MA: Harvard University Press), 372–388. 10.4159/harvard.9780674864917.c27

  • CrossRef
  • Google Scholar
  • View reference in article

• 85

  MelackJ. M.HessL. L. (2010). “Remote sensing of the distribution and extent of wetlands in the Amazon basin,” in Amazonian Floodplain Forests: Ecophysiology, Biodiversity and Sustainable Management, eds. JunkW.PiedadeM.WittmannF.SchöngartJ.ParolinP. (Dordrecht: Springer), 43–59. 10.1007/978-90-481-8725-6_3

  • CrossRef
  • Google Scholar
  • View reference in article

• 86

  MenesesH. D. N. D. M.Oliveira-da-CostaM.BastaP. C.MoraisC. G.PereiraR. J. B.de SouzaS. M. S.et al (2022). Mercury contamination: a growing threat to riverine and urban communities in the Brazilian Amazon. Int. J. Environ. Res. Public. Health.19 (5), 2816. 10.3390/ijerph19052816

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 87

  MertesL. A.DunneT.MartinelliL. A. (1996). Channel-floodplain geomorphology along the Solimões-Amazon river, Brazil. Geol. Soc. Am. Bull.108 (9), 1089–1107. 10.1130/0016-7606(1996)108%3C1089:CFGATS%3E2.3.CO;2

  • Google Scholar
  • View reference in article

• 88

  MoraesB. C. D.CostaJ. M. N. D.CostaA. C. L. D.CostaM. H. (2005). Variação espacial e temporal da precipitação no estado do pará. Acta Amazonica35, 207–214. 10.1590/S0044-59672005000200010

  • CrossRef
  • Google Scholar
  • View reference in article

• 89

  MoraesC. P.NevesE. G. (2019). “Earthworks of the Amazon,” in Encyclopedia of Global Archaeology, ed. C. Smith (Switzerland AG: Springer Nature), p. 1–13. 10.1007/978-3-319-51726-1_3026-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 90

  MottW. A. (2022) Gênese de solos aluviais e implicações ambientais no cariri paraibano. M.S. Thesis. Universidade Federal de Viçosa.

  • Google Scholar
  • View reference in article

• 91

  NevesE. G. (2022). Sob os Tempos do Equinócio: Oito mil Anos de História da Amazônia Central. São Paulo: Ubu Editora.

  • Google Scholar
  • View reference in article

• 92

  NevesE. G.GuapindaiaV. L.LimaH. P.CostaB. L.GomesJ. (2014). “A tradição pocó-açutuba e os primeiros sinais visíveis de modificações de paisagens na calha do amazonas,” in Amazonía: Memorias de las Conferencias Magistrales del 3er Encuentro Internacional de Arqueología Amazónia, ed. RostainS. (Quito: Ministerio Coordinador de Conocimiento y Talento Humano e IKIAM), 137–158.

  • Google Scholar
  • View reference in article

• 93

  NimuendajúC. (2004). [1922-1927]. in Pursuit of a Past Amazon: Archaeological Researches in the Brazilian Guyana and in the Amazon Region. Gothenburg: Elanders Infologistik.

  • Google Scholar
  • View reference in article

• 94

  OberdorffT.DiasM. S.JézéquelC.AlbertJ. S.ArantesC. C.BigorneR.et al (2019). Unexpected fish diversity gradients in the Amazon basin. Sci. Adv.5 (9), eaav8681. 10.1126/sciadv.aav8681

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 95

  OlooM. A.XieC. (2018). A method to integrate geological knowledge in variogram modeling of facies: a case study of a fluvial deltaic reservoir. Int. J. Geosci.9 (06), 337. 10.4236/ijg.2018.96021

  • CrossRef
  • Google Scholar
  • View reference in article

• 96

  PalaceM. W.McMichaelC. N.BraswellB. H.HagenS. C.BushM. B.NevesE.et al (2017). Ancient amazonian populations left lasting impacts on forest structure. Ecosphere.8 (12), e02035. 10.1002/ecs2.2035

  • CrossRef
  • Google Scholar
  • View reference in article

• 97

  PeixotoR. C.ArenzK.FigueiredoK. (2012). O movimento indígena no baixo tapajós: etnogênese, território, estado e conflito. Novos Cadernos NAEA15 (2), 279–313. 10.5801/ncn.v15i2.719

  • CrossRef
  • Google Scholar
  • View reference in article

• 98

  PeripatoV.LevisC.MoreiraG. A.GamermanD.SteegeH. T.PitmanN. C. A.et al (2023). More than 10,000 pre-Columbian earthworks are still hidden throughout Amazonia. Science382 (6666), 103–109. 10.1126/science.ade2541

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 99

  PetermannP. (1997). “The birds,” in The Central Amazon Floodplain: Ecology of a Pulsing System, ed. W. J. Junk (Berlin, Heidelberg: Springer Berlin Heidelberg), 419–452. 10.1007/978-3-662-03416-3_22

  • CrossRef
  • Google Scholar
  • View reference in article

• 100

  Prestes-CarneiroG.BéarezP. (2017). Swamp-eel (synbranchus spp.) fishing in Amazonia from pre-Columbian to present times. J. Ethnobiol.37 (3), 380–397. 10.2993/0278-0771-37.3.380

  • CrossRef
  • Google Scholar
  • View reference in article

• 101

  Prestes-CarneiroG.BéarezP.BailonS.Py-DanielA. R.NevesE. G. (2016). Subsistence fishery at hatahara (750–1230 CE), a pre-Columbian central Amazonian village. J. Archaeol. Sci. Rep.8, 454–462. 10.1016/j.jasrep.2015.10.033

  • CrossRef
  • Google Scholar
  • View reference in article

• 102

  Prestes-CarneiroG.BéarezP.PuglieseF.ShockM. P.ZimpelC. A.PouillyM.et al (2020). Archaeological history of middle holocene environmental change from fish proxies at the monte castelo archaeological shell mound, southwestern Amazonia. Holocene30 (11), 1606–1621. 10.1177/0959683620941108

  • CrossRef
  • Google Scholar
  • View reference in article

• 103

  Prestes-CarneiroG.BéarezP.ShockM. P.PrümersH.Jaimes BetancourtC. (2019). Pre-Hispanic fishing practices in interfluvial Amazonia: zooarchaeological evidence from managed landscapes on the llanos de mojos savanna. PLoS One14 (5), e0214638. 10.1371/journal.pone.0214638

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 104

  Prestes-CarneiroG.Sá Leitão BarbozaR.Sá Leitão BarbozaM.MoraesC. D. P.BéarezP. (2021). Waterscapes domestication: an alternative approach for interactions among humans, animals, and aquatic environments in Amazonia across time. Animal Frontiers11 (3), 92–103. 10.1093/af/vfab019

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 105

  PrümersH.BetancourtC. J.IriarteJ.RobinsonM.SchaichM. (2022). Lidar reveals pre-hispanic low-density urbanism in the Bolivian Amazon. Nature606 (7913), 325–328. 10.1038/s41586-022-04780-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 106

  PuglieseF.CayónL.NevesE. (2024). “Building meanings as places: the archaeology of the shellmounds and the Indigenous infrastructure of the Amazon,” in Infrastructure in Archaeological Discourse, eds. M. G. Ellis, C. M. DeSanto and M. C. L. Howey (Abingdon: Routledge), 219–240. 10.4324/9781032678450-16

  • CrossRef
  • Google Scholar
  • View reference in article

• 107

  Py-DanielA. R.O'DwyerE. C.MoraesC. P.MoreiraC. S.PintoR. S.PintoE. S. (2023). Dinâmicas de ocupação da região do rio ituqui/paranã do maicá, Santarém/PA: 8000 anos até o presente. Revista de Arqueología Pública: Revista Eletrônica do Laboratório de Arqueologia Pública de Unicamp18 (1), 5. 10.20396/rap.v18i00.8673267

  • CrossRef
  • Google Scholar
  • View reference in article

• 108

  RenóV. F.NovoE. M. L. M.Almeida-FilhoR.SuemitsuC. (2011). Mapeamento da antiga cobertura vegetal de várzea do baixo Amazonas a partir de imagens históricas (1975–1981) do sensor MSS-landsat. Acta Amazonica41, 47–56. 10.1590/S0044-59672011000100006

  • CrossRef
  • Google Scholar
  • View reference in article

• 109

  RenóV.NovoE.EscadaM. (2016). Forest fragmentation in the Lower Amazon floodplain: implications for biodiversity and ecosystem service provision to riverine populations. Remote. Sens. (Basel)8 (11), 886. 10.3390/rs8110886

  • CrossRef
  • Google Scholar
  • View reference in article

• 110

  RooseveltA. C. (1995). “Early pottery in the Amazon: twenty years of scholarly obscurity,” in The Emergence of Pottery: Technology and Innovation in Ancient Societies, eds. BarnettW.HoopesJ. (Washington, DC: Smithsonian Institution), 115–131.

  • Google Scholar
  • View reference in article

• 111

  RooseveltA. C.HousleyR. A.SilveiraM. I.MarancaS.JohnsonR. (1991). Eighth millennium pottery from a prehistoric shell midden in the Brazilian Amazon. Science254 (5038), 1621–1624. 10.1126/science.254.5038.1621

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 112

  RooseveltA. C.Lima da CostaM.MachadoC. L.MichabM.MercierN.ValladasH.et al (1996). Paleoindian cave dwellers in the Amazon: the peopling of the Americas. Science272 (5260), 373–384. 10.1126/science.272.5260.373

  • CrossRef
  • Google Scholar
  • View reference in article

• 113

  RostainS.DorisonA.De SaulieuG.PrümersH.Le PennecJ. L.Mejía MejíaF.et al (2024). Two thousand years of garden urbanism in the upper Amazon. Science383 (6679), 183–189. 10.1126/science.adi6317

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 114

  RozoM. G.NogueiraA. C. R.CastroC. S. (2014). Remote sensing-based analysis of the planform changes in the upper Amazon river over the period 1986–2006. J. South. Am. Earth. Sci.51, 28–44. 10.1016/j.jsames.2013.12.004

  • CrossRef
  • Google Scholar
  • View reference in article

• 115

  SaloJ.KalliolaR.HäkkinenI.MäkinenY.NiemeläP.PuhakkaM.et al (1986). River dynamics and the diversity of Amazon lowland forest. Nature322 (6076), 254–258. 10.1038/322254a0

  • CrossRef
  • Google Scholar
  • View reference in article

• 116

  Santos de LimaL.SilvaF. E. O. E.Dorio AnastácioP. R.KolanskiM. M. D. P.Pires PereiraA. C.MenezesM. S. R.et al (2024). Severe droughts reduce river navigability and isolate communities in the Brazilian Amazon. Commun. Earth Environ.5 (1), 370. 10.1038/s43247-024-01530-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 117

  SáL. P. (2016). “Cerâmicas pocó e konduri no baixo amazonas,” in Cerâmicas Arqueológicas da Amazônia: Rumo a uma nova Síntese, eds. BarretoC.LimaH. P.BetancourtC. J. (Belém: IPHAN), 293–301.

  • Google Scholar
  • View reference in article

• 118

  SchaanD. P. (2016). “Discussing centre-periphery relations within the Tapajó domain, Lower Amazon,” in Beyond Waters: Archaeology and Environmental History of the Amazonian Inland, ed. StenborgP. (Gothenburg: University of Gothenburg), 23–36.

  • Google Scholar
  • View reference in article

• 119

  SheikhP. A.MerryF. D.McGrathD. G. (2006). Water Buffalo and cattle ranching in the Lower Amazon basin: comparisons and conflicts. Agric. Sys.87 (3), 313–330. 10.1016/j.agsy.2005.02.003

  • CrossRef
  • Google Scholar
  • View reference in article

• 120

  ShockM. P.MoraesC. P. (2019). A floresta é o domus: a importância das evidências arqueobotânicas e arqueológicas das ocupações humanas amazônicas na transição pleistoceno/holoceno. Boletim do Museu Paraense Emílio Goeldi. Ciências Humanas14, 263–289. 10.1590/1981.81222019000200003

  • CrossRef
  • Google Scholar

• 121

  SilvaW.PereiraE. (2021). “Programa de gestão do patrimônio arqueológico na Área do empreendimento petróleo sabbá S.A - base de santarém, município de santarém – pará,” in Partial Technical and Scientific Report on the Archaeological Rescue of the Porto PA-ST-42 Site, Belém: Inside Consultoria Científica.

  • Google Scholar

• 122

  SioliH. (1984). “The Amazon and its main affluents: hydrography, morphology of the river courses, and river types,” in The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin, ed. SioliH. (Dordrecht: Springer Netherlands), 127–165. 10.1007/978-94-009-6542-3_5

  • CrossRef
  • Google Scholar

• 123

  StenborgP. (2016). Beyond Waters: Archaeology and Environmental History of the Amazonian Inland (Gotarc Series A Gothenburg Archaeological Studies 6). Gothenburg: University of Gothenburg-Department of Historical Studies. 10.26530/OAPEN_632432

  • CrossRef
  • Google Scholar

• 124

  StenborgP.SchaanD. P.FigueiredoC. G. (2018). Contours of the past: LiDAR data expands the limits of late pre-Columbian human settlement in the Santarém region, Lower Amazon. J. Field Archaeol.43 (1), 44–57. 10.1080/00934690.2017.1417198

  • CrossRef
  • Google Scholar

• 125

  StenborgP.SchaanD. P.LimaM. A. (2012). Precolumbian land use and settlement pattern in the Santarém region, Lower Amazon. Amazônica-Revista de Antropologia4 (1), 222–250. 10.18542/amazonica.v4i1.886

  • CrossRef
  • Google Scholar

• 126

  TandonS. K.SinhaR. (2007). Geology of Large River Systems. Large Rivers: Geomorphology and Management. 7–28. 10.1002/9780470723722.ch2

  • CrossRef
  • Google Scholar

• 127

  ThomasS.AlmeidaO.PinheiroE.RiveroS.MoriJ. (2014). Projeto assentamento agroextrativista e a construção do desenvolvimento sustentável das várzeas amazônicas (paper 346). Papers do NAEA1 (1), 3–19. 10.18542/papersnaea.v23i1.11274

  • CrossRef
  • Google Scholar

• 128

  TroufflardJ. (2016). “Cerâmicas da cultura santarém, baixo tapajós,” in Cerâmicas Arqueológicas da Amazônia: Rumo a uma nova Síntese, eds. BarretoC.LimaH. P.Jaimes BetancourtC. (Belém: IPHAN), 237–252.

  • Google Scholar

• 129

  TroufflardJ.AlvesD. T. (2019). Uma abordagem interdisciplinar do sítio arqueológico cedro, baixo Amazonas. Boletim do museu Paraense emílio goeldi. Ciências Humanas14, 553–580. 10.1590/1981.81222019000200015

  • CrossRef
  • Google Scholar

• 130

  VepraskasM. J. (2015). Redoximorphic Features for Identifying Aquic Conditions. Technical Bulletin 301. Raleigh, NC: North Carolina State University.

  • Google Scholar

• 131

  WangQ.AdikuS.TenhunenJ.GranierA. (2005). On the relationship of NDVI with leaf area index in a deciduous forest site. Remote Sens. Environ.94 (2), 244–255. 10.1016/j.rse.2004.10.006

  • CrossRef
  • Google Scholar

• 132

  WinklerPrinsA. M. (2006). Jute cultivation in the Lower Amazon, 1940–1990: an ethnographic account from Santarém, Pará, Brazil. J. Hist. Geogr.32 (4), 818–838. 10.1016/j.jhg.2005.09.028

  • CrossRef
  • Google Scholar

• 133

  WittmannF.HouseholderE.PiedadeM. T.de AssisR. L.SchöngartJ.ParolinP.et al (2013). Habitat specifity, endemism and the neotropical distribution of Amazonian white-water floodplain trees. Ecography36 (6), 690–707. 10.1111/j.1600-0587.2012.07723.x

  • CrossRef
  • Google Scholar

• 134

  WittmannF.SchöngartJ.MonteroJ. C.MotzerT.JunkW. J.PiedadeM. T.et al (2006). Tree species composition and diversity gradients in white-water forests across the Amazon basin. J. Biogeogr.33 (8), 1334–1347. 10.1111/j.1365-2699.2006.01495.x

  • CrossRef
  • Google Scholar

• 135

  XueJ.SuB. (2017). Significant remote sensing vegetation indices: a review of developments and applications. J. Sens.2017 (1), 1353691. 10.1155/2017/1353691

  • CrossRef
  • Google Scholar

• 136

  YangF.MatsushitaB.FukushimaT.YangW. (2012). Temporal mixture analysis for estimating impervious surface area from multi-temporal MODIS NDVI data in Japan. ISPRS. J. Photogramm. Remote. Sens.72, 90–98. 10.1016/j.isprsjprs.2012.05.016

  • CrossRef
  • Google Scholar

• 137

  ZhuX.LiuD. (2015). Improving forest aboveground biomass estimation using seasonal landsat NDVI time-series. ISPRS. J. Photogramm. Remote. Sens.102, 222–231. 10.1016/j.isprsjprs.2014.08.014

  • CrossRef
  • Google Scholar

• 138

  ZuseS. (2016). “Variabilidade cerâmica e diversidade cultural no alto rio Madeira,” in Cerâmicas Arqueológicas da Amazônia: Rumo a uma nova Síntese, eds. BarretoC.LimaH. P.BetancourtC. Jaimes (Belém: IPHAN), 385–401.

  • Google Scholar