What do paleoecological data tell us about the Amazon Anthropocene?

ABSTRACT

Recently, Roberts et al. (2023) defined three socioeconomic thresholds thought to have laid the foundations for the Anthropocene: the first, between 4000 and 1000 BCE, related to the beginnings of agriculture; the second, between 0 and 1500 CE, related to urbanism and the demographic consolidation of human populations; and the third, between 1500 and the present CE, associated with the advance of colonialism and capitalism. In this article, we discuss the results of paleoecological studies carried out in four Amazonian regions known to have high concentrations of archaeological sites (the geoglyphs of Acre, the raised fields of French Guiana, the dark earths of the lower Tapajós region, and the zanja sites of Iténez, Bolivia), and evaluate what they tell us about the nature of the Amazonian Anthropocene. We conclude that the largest and most destructive anthropogenic impacts occurred during the colonial period, especially in the last 50 years, associated with the arrival of capitalist economies. In contrast, Indigenous management practices, which begin to be visible from ca. 2500 BCE, and become highly transformative from ca. 0 CE, managed to maintain vital ecosystem services and increase the agrobiodiversity of the rainforest, building upon - rather than destroying - their relationships with other living beings.

KEYWORDS:
Paleoecology; Archaeology; Thresholds for the Anthropocene; Amazonia

RESUMO

Recentemente, Roberts et al. (2023) definiram três thresholds socioeconômicos que teriam sido a base do Antropoceno no mundo: o primeiro, entre 4.000 e 1.000 AEC, relacionado ao início da agricultura; o segundo, entre 0 e 1.500 EC, relacionado ao urbanismo e à consolidação demográfica das populações humanas; e o terceiro, entre 1.500 EC e o presente, associado ao avanço do colonialismo e do capitalismo. Neste artigo, discutimos os resultados de estudos paleoecológicos realizados em quatro regiões da Amazônia conhecidas por apresentarem altas concentrações de sítios arqueológicos (os geoglifos do Acre, os campos elevados da Guiana Francesa, as terras pretas do Baixo Tapajós, e os sítios de zanja em Iténez, Bolívia), e avaliamos o que eles nos dizem sobre a natureza desses thresholds e como eles podem nos informar sobre o Antropoceno amazônico. Concluímos que os maiores e mais destrutivos impactos antropogênicos têm acontecido no período colonial, especialmente nos últimos 50 anos, associados à incursão de economias capitalistas nestes locais. Em contraponto, as práticas de manejo indígenas, que começam a ser visíveis a partir de cerca de 2.500 AEC, e se tornam altamente transformativas a partir de cerca de 0 EC, conseguiram manter serviços ecossistêmicos vitais e aumentar a agrobiodiversidade da sua vegetação, construindo - ao invés de destruindo - relações com os outros seres vivos.

PALAVRAS-CHAVE:
Paleoecologia; Arqueologia; Thresholds do Antropoceno; Amazônia

Introduction

He concept of the Anthropocene denotes the profound impact of human societies on Earth’s geological history. Despite intense debate, recent efforts to formally define it as a geological epoch were ultimately rejected by the Subcommission on Quaternary Stratigraphy. However, the Anthropocene as an ongoing event remains widely accepted, supported by indisputable evidence of human-induced planetary transformation. This reflects a narrative of gradual but significant changes spanning centuries. Originally coined by Paul Crutzen in 2000, the term “Anthropocene” increasingly characterizes our era as one defined by human-driven planetary alterations, rooted in the industrialization of the late 18th century and accelerating significantly over the past five centuries with colonial expansion and the Industrial Revolution. In its non-chronostratigraphic sense, the Anthropocene is used informally to denote general processes of anthropogenic impact on the planet. Although these impacts have accelerated since the 19th century, archaeological and paleoecological evidence shows that humans have been altering the geosphere for thousands of years.

Tropical regions, such as the Amazon basin, are among the most important places to study the foundations of the Anthropocene, given the importance of their biodiversity and the role they play in regulating the world’s climate. Recently, Roberts et al. (2023) highlighted the need to research the deep history of interactions between humans and environments in tropical regions through multidisciplinary approaches. Roberts et al. identified three potential socioeconomic thresholds that lay the foundations of the Anthropocene: 1) changes in land cover related to the adoption of agriculture, suggested to have taken place by approximately 4000-1000 Before Common Era (BCE); 2) the emergence of urbanism and, with it, demographic consolidation, between 0-1500 Common Era (CE); and 3) European colonialism and the expansion of capitalist systems, between 1500 CE and the present.

Exploring critical thresholds in human-environment interactions can be effectively achieved through the multidisciplinary analysis of paleoecological and archaeological archives. In the Amazon basin, the past decade has witnessed a substantial increase in efforts to quantify the spatial and temporal scales of human impact on the environment (Mayle; Iriarte, 2014). It is now widely recognized that human groups managed fruit and nut species before 11,000 calibrated date (cal.) BCE, with these practices rapidly extending to the cultivation of annual crops in some regions (Iriarte et al., 2020a). Evidence also indicates that subsequent population growth and the establishment and spread of food production systems throughout the Middle and Late Holocene culminated in a peak of anthropogenic landscape formation between approximately 0 and 500 cal. CE (Arroyo-Kalin; Riris, 2020). This period was followed by the genocide of Indigenous populations and the colonial era, which heralded an “apogee of hectic regional biological, physical, cultural, and human devastation” (Roosevelt 2014, p. 84).

While evidence continues to accumulate on the nature of the Amazonian Anthropocene, many questions remains, including: 1) What was the role of deforestation in Indigenous management practices in the past?, 2) Was the clearing of large swiddens a pre-colonial phenomenon or, as some people propose (e.g. Denevan, 2006), post-colonial?, 3) To what extent were dark earths “made” for cultivation purposes?, 4) Can we generalize about an Amazonian “management mode”, or about the types and scales of ecological legacies left by peoples from different parts of the Amazon basin?, 5) To what extent did the Amazon rainforest “recover” after the demographic collapse caused by European invasion (e.g. Koch et al., 2019), and 6) How was the appearance of capitalist economies, that include different cycles of post-colonial extractivism, responsible for shaping the composition of their ecosystems (e.g. McMichael et al., 2017)?

In this contribution, we compile data from four regions in the Amazon where local reconstructions were possible through integrative and multi-proxy paleoecological approaches. These study regions include: 1) the geoglyphs of Acre; 2) the raised fields of French Guiana; 3) the dark earths of the lower Tapajós; and 4) the zanja sites in Iténez, Bolivia. After examining each case individually, we will discuss how the data from these different regions interact and whether they provide answers to the questions raised above. Finally, we will evaluate the extent to which it is possible to identify the different Anthropocene thresholds proposed by Roberts et al. (2023) for the global tropics.

Finding indigenous management practices in the paleoecological record

The ways in which Indigenous management practices transform Amazonian ecosystems only began to be studied by ecologists and, eventually, archaeologists, from the 1980s onwards (Posey; Balée, 1989). The emergence of Historical Ecology transformed the way of studying the interactions between people and the environment in the present and in the archaeological past: populations previously considered limited by the low productivity of tropical soils came to be seen as agents in the creation of agrobiodiversity, and thus their own living conditions. These processes, encompassed by the terms “management practices”, “landscape domestication”, and “cultural niche construction”, are both conscious and unconscious, but result from relationships built carefully and intentionally between humans and other beings that cohabit the world (Clement et al., 2021).

While the co-construction of Amazonian biodiversity by Indigenous peoples extends back to their arrival in the region before ca. 11,000 cal. BCE (Morcote-Rios et al., 2021; Neves et al., 2021), the empirical data available to assess how, where and when these processes took place are relatively few and limited to specific regions or archaeological sites. Although ecological studies conducted at both local and pan-regional scales (e.g. Junqueira et al., 2010; Levis et al., 2018) demonstrate a clear correlation between the agrobiodiversity present in a vegetation community and its proximity to an archaeological site; unravelling the temporal depth of these legacies is a methodological challenge that paleoecological research programs are able to address.

Paleoecology, in its broadest sense, is the discipline that aims to reconstruct the life histories of organisms and their relationships with each other over geological time. However, while it is about the ecology of the past, it is not done by ecologists, but by specialists, usually from the areas of geology or the environmental sciences, who study the fossil record of plants and animals (Lins de Paiva; Gallo, 2021). In the Amazon region, paleoecology has become almost synonymous with the study of pollen grains and other fossil remains preserved in lake and swamp sediments to understand changes in vegetation composition over time. Beginning in the 1970s, the major focus of these studies was on the effects of glacial cycles on the evolution of tropical vegetation; in particular the need to test the Forest Refuge Hypothesis (Colinvaux et al., 2000). It was not long, however, before scientists began to find evidence of human presence in palynological records, in the form of maize (Zea mays) pollen grains or signs of vegetation opening by fire (Bush; Colinvaux 1988). Such discoveries stimulated an interest in the influence humans had on Amazonian vegetation over time, a theme that has gained a lot of strength in the last decade because of its current implications with socio-biodiversity (Neves et al., 2021).

Methodological advances in recent years have allowed paleoecological studies to achieve greater analytical power. Improvements in the accuracy of radiocarbon dating, the reduction in sample sizes needed to date sediments, isotopic analyses, and the accessibility of regional reference collections (Flantua et al., 2015) have improved the detail and chronological resolution of palynological studies. Recently, phytoliths have also begun to be used to reconstruct vegetation communities from off-site soil profiles away from archaeological sites (Watling et al., 2017). Although they may not be as sensitive to differences between certain forest types (Hill et al., 2023), phytoliths can illuminate differences in past deforestation intensity and forest enrichment in useful species typical of human presence, such as palms. Moreover, since their deposition is highly local (<5 m), these data can complement palynological data - whose interpretative scales are mostly regional (>10 km) - or serve as an alternative source of information to pollen in places without the presence of bodies of water. Additionally, advances in charcoal analysis from sediment cores such as charcoal morphology (Leys et al. 2015), reflectance (Belcher et al., 2018), and chemistry (Gosling et al., 2019), and the creation of programmes and statistical packages, such as CharAnalysis, that enable advanced data manipulation (Higuera et al., 2011), have greatly improved the understanding of the spatial-temporal variability of past fires.

The last ten years have also seen greater integration between the fields of paleoecology and archeology, and a concern to carry out integrative and trans/multidisciplinary projects (Mayle; Iriarte, 2014; Iriarte et al., 2020a, Figure 1). Such an approach becomes important given the need, during the interpretation of the paleoecological record, to distinguish between “natural” phenomena versus anthropic ones. In practice, such a distinction can only be made by mobilizing expectations based on archaeological data (about where, how, and how many people were living in a certain period), and paleoclimatic/geotectonic data (about how rainfall and temperature regimes, or landmasses themselves evolved over time). With regard to archaeology, recent use has been made of site probability distribution curves (SPD), which use radiocarbon date frequencies as a proxy for past human demography (Arroyo-Kalin; Riris, 2020). Isotopic analyses performed on speleothems can provide paleoprecipitation records with sub-decadal accuracy (Novello et al., 2017).

Figure 1
Schematic diagram of an integrative and multi-proxy paleoecological approach.

Last but not least, archaeobotanical analyses carried out at archaeological sites have played a key role within integrative approaches by providing evidence on the nature and temporality of management practices and food production in the past. While several domesticated species, such as maize, beans (Phaseolus sp.), squash (Cucurbita sp.), cassava (Manihot esculenta), sweet potatoes (Ipomoea batatas), among others, are more easily identified through their microbota- nical remains (e.g. phytoliths and starch grains), macrobotanical remains allow better identification of perennial, not necessarily domesticated species, such as fruit trees or woody species (Caromano et al., 2013; Watling et al., 2018a). The marriage of on-site archaeobotanical data with off-site paleoecological data is important to understand the impacts that management practices have had on regional landscapes and can also be complemented with data from current floristic inventories to better understand the legacies of these practices on current vegetation (Maezumi et al., 2018a).

Study regions

Figure 2
Sites and regions discussed in the text are on the map: A) geoglyphs of the east of the state of Acre, Brazil, B) the zanja sites in Iténez, Bolívia, c) raised fields of the coast of French Guiana, and D) archaeological sites with dark earths near the lower Tapajós river, region of Santarém, Brazil.

The geoglyphs of Acre

The geoglyphs of Acre (Figure 2) are geometric sites composed of ditches measuring up to 11 m wide and 4 m deep, enclosing areas up to 300 m in diameter. There are more than 500 geoglyphs recorded for the eastern region of the state of Acre, where they have been documented since the 1990s when regional deforestation made these structures visible during flyovers. As of the 2000s, archaeological studies of geoglyphs have demonstrated that the vast majority were built and occupied throughout the first millennium of the CE, but that some date back to ca. 1000 cal. BCE (Parsinnen et al., 2009; Schaan et al., 2012). Data from excavations show that the interiors of the structures are characterized by a very low density of archaeological material while inside the ditches themselves, close to the entrances, features composed of whole and decorated vessels are found. These data, together with their architectural forms (circles, squares, and their combinations), point to a public/ceremonial, rather than domestic, use for these sites, where different social groups would gather around special occasions (Saunaluoma; Virtanen, 2015).

Located in high areas of the interfluve of the Purus River basin, far from river floodplains, the flora of this region was considered untouched by humans until their discovery. In 2011, a project began with the aim of studying the landscape history of the geoglyphs (Watling et al., 2017, 2018b). The questions raised by the research were: 1) What was the predominant vegetation in the region when the geoglyphs were built?, 2) What was the spatial scale of environmental impact associated with the construction and use of the sites?, 3) What other aspects of the environment were transformed by Indigenous management practices?, and 4) What effect did the abandonment of the geoglyphs have on the regional vegetation?

The study of pollen grains to reconstruct the region´s vegetation was unfeasible for two reasons: the location of the geoglyphs far from water bodies; and the nature of the lakes present in the region, which are mostly recent paleochannels. Therefore, the collection of paleoenvironmental data was done through soil profiles that were sampled for the analysis of phytoliths (as a vegetation proxy), charcoal (as a proxy for vegetation burning) and carbon isotopes (as a proxy for changes in the density of the forest canopy). Six profiles were analysed, and chronologies established through charcoal and soil humin dating. The first profile was opened in the centre of the Fazenda Colorada site (FC), a site composed of three geoglyphs, dated to between 250 and 1378 cal. CE. The other five were opened in the vicinity of the Jaco Sá site (JS), a site composed of two structures built between 750 and 966 cal. CE. The first profile was excavated in the centre of the site (0 km), and the others in locales 500m, 1.5 km, 3.5 km and 7.5 km away from it, to verify the spatial scale of the impacts associated with its construction and use.

The raised fields of French Guiana

In several regions of the Americas, annually flooded savanna environments were transformed into productive landscapes for agriculture through the construction of raised fields (Denevan, 2001; Whitmore; Turner, 2001). These structures range from small circular mounds to rectangular platforms measuring tens of meters in length, and would have provided several benefits for cultivation, such as soil drainage, better water retention during dry periods, increased fertility and crop yields, and the creation of channels and pools suitable for fish farming (Erickson; Walker, 2009). In the coastal strip of the Guianas, complexes of elevated fields have been recorded across an area of ​​600 km² (Rostain, 2010). In French Guiana, two raised field complexes were dated to 670-700 cal. CE (Bois Diable site) and 920-950 cal. CE (the K-VIII site) and associated with occupations by producers of Arauquinoid Tradition ceramics. These populations migrated to the Guyanese coast from a point of origin on the Middle Orinoco River ca. 500 CE (Rostain; Versteeg, 2004).

Between 2006 and 2012, a multidisciplinary team of researchers was formed to investigate the formation of the raised fields, and the landscapes in which they occur. The methodological approach consisted of: 1) the excavation of the habitation sites and the analysis of the microbotanical remains contained in ceramic artefacts, 2) the excavation and sampling of soils in different types of fields, for analysis of phytoliths, charcoal and geochemistry, and 3) pollen, charcoal and phytolith analysis of a sediment core extracted within a permanently flooded area of the K-VIII site. The results of these analyses (Iriarte et al., 2010, 2012), confirmed the anthropic nature of the raised fields (Rostain, 2010) and identified their use for maize cultivation. In this paper, we will focus on data from the K-VIII site where it was possible to observe broader landscape changes associated with raised field construction and abandonment.

The dark earths of the lower Tapajós

Amazonian Dark Earths (ADEs), anthropogenic soils that formed in habitation areas created through the decomposition of organic waste (Glaser; Birk, 2012; Woods et al., 2009) are one of the most compelling lines of evidence that humans transformed Amazonian landscapes (Arroyo-Kalin, 2016; Petersen et al., 2001). They are black to brown in colour and vary in their thickness and composition as well as in the densities of artefacts they contain (e.g. Erickson, 2003; Hecht, 2003). ADEs have high concentrations of organic matter and charcoal and elevated values of pH, P, Ca, and Mg which enable them to maintain nutrient levels over hundreds of years, making them some of the most fertile soils in the world (Teixeira et al., 2010).

ADEs are broadly distributed in Amazonia, but they are generally located on terra firme soils on topographically strategic locations (Denevan, 1996). With the exception of regions in the NW and SW Amazon, where ADEs date to 4000 cal. BCE (Mora et al., 1991; Mongeló, 2020), the majority began forming ca. 500 cal. BCE. While most sites are less than 2 ha in area (Kern et al., 2003), some sites, such as the Porto site in Santarem, cover over 90 ha (Roosevelt, 1987).

While our understanding of how these ADEs were formed, and by whom, has increased significantly in the last two decades, little is known about the Indigenous land use management systems they supported. Between 2014 and 2018 the “Pre-Columbian Amazon-Scale Transformations (PAST)” project aimed to address this issue and, more specifically: 1) What annual and perennial crops were consumed and planted in ADEs?, 2) What was the extent of forest clearing associated with ADEs?, 3) Were ADE residents enriching the forest with edible plants?, and 4) What was the role of fire in these management systems?

The project focused on the region of Santarém, at the confluence of the Tapajós and Amazon rivers, where there are extensive and well dated ADEs as well as a lake (Caranã) that is large and old enough to capture the pollen rain and charcoal particles that the ADEs’ land use systems produced. The results of these analyses (Maezumi et al., 2018a, 2018b) demonstrated polyculture agroforestry practices that began by 2500 cal. BCE. Here, we focus on data from Lake Caranã and the Maguari archaeological site where we observe pre- and post-colonial land use changes.

Zanja sites in the Llanos de Mojos, Bolívia

The Llanos de Moxos, located in the Bolivian department of the Beni, has some of the most diverse and extensive examples of earthworks in the Amazon Basin (Walker, 2018). The province of Iténez, in the NE of the region, is characterized by extensive ring-ditch earthworks known locally as zanjas (Prümers and Betancourt, 2014) that occur alongside causeways, ditched agricultural fields, and fish weirs. Remote sensing and ground-based archaeological surveys show that these sites cover an area of ca. 12,000 km² (Erickson, 2010).

The zanjas are located on small plateaus overlooking streams and can exhibit diverse shapes, (Erickson, 2010). With berms located outside of the ditches they recall the geoglyphs of Acre but, different from these, the interior of the zanjas contain abundant archaeological materials including ceramics, ADEs and human burials. Historical accounts from the 18^th century describe settlements of Baures language speaking groups with a defensive ring ditch enhanced by a palisade (Eder, 1985 [1772]) suggesting that these sites served as fortified villages.

To explore the history of environmental changes caused by human land use we implemented a multi-proxy approach to compare local-scale vegetation and fire histories (through archaeological excavations and analysis of soil profiles) with those on a regional-scale (through palynological analysis of lakes). The fieldwork occurred between 2016 and 2017 and results were published a few years later (Iriarte et al., 2020a, 2020b; Maezumi et al., 2022; Robinson et al., 2020). Here we summarize the data from the Lake Versalles sediment core and soil profiles from the Triunfo archaeological site, located on the margin of this lake. Triunfo has ADE deposits up to 1 m in depth and contains occupations assigned to three distinct ceramic phases. Construction of the zanja, which is a double oval ditch, is dated to 1300-1400 cal. CE, with a second constructive phase (including post holes indicative of a palisade) between 1630 and 1800 cal. CE, that is associated with people who used late Versalles phase ceramics (Robinson et al., 2020). This chronology is consistent with dates obtained from other zanjas in the region (Prümers; Betancourt, 2014; Walker, 2018).

The geoglyphs of Acre

Charcoal inversions (the appearance of younger charcoal beneath older charcoal) occur in four of the six profiles (Figure 3) and demonstrate the inherent difficulty in analysing soils which, unlike lake sediments, often suffer from mixing through bioturbation. Nevertheless, the broad chronological integrity of the profiles was ascertained based on the overall consistency of the data both within and across different profiles, and concordance between the dating of events in the profiles, and dates obtained from the archaeological sites.

The basal dates obtained for the profiles fall between 4500 and 2500 cal. BCE, during the Middle Holocene. Phytolith assemblages from this period, but also throughout the profile, are consistent with those from bamboo forests. Today, this type of vegetation is found across an area of ​​ca. 161,500 km² and is dominated by the genus Guadua sp., a woody bamboo whose aggressiveness towards larger trees results in forests that are structurally more open.

From 2500 cal. BCE, increases in charcoal in some of the profiles are interpreted as signs of vegetation burning by humans. This hypothesis is based on paleoecological records that demonstrate an increase in rainfall during this period corresponding to the beginning of the Late Holocene, a fact that would have reduced the natural flammability of the local vegetation. The period also coincides with the oldest archaeological dates in the region (Saunaluoma; Schaan 2012).

Concurrent with this increase in charcoal across the records, the profiles closest to the geoglyphs (FC1, JS1 and JS2) begin to record a parallel, albeit gradual, increase in palm phytoliths (Arecaceae). This increase was not expected due to the increased precipitation in this period (Baker et al., 2001) that would have favoured a denser forest canopy that would be less favourable for the establishment of these species. This suggests that the forests in these locations were being progressively enriched with useful plants before the geoglyphs were built. It is interesting to note that the same phenomenon happens in JS4 (albeit less accentuated), but is not repeated in profiles JS3 and JS5, although the frequency of charcoal in these profiles also increases over time.

Figure 3
Phytolith, charcoal and stable carbon isotope data from the analysis of six soil profiles in, and nearby, the Fazenda Colorada e Jaco Sá geoglyphs, in the east of the state of Acre.

Charcoal peaks at levels 45-50 cm (from FC1) and 30-35 cm (from JS1), dated respectively to 296 cal. BCE and 440 cal. CE, represent episodes of vegetation burning that are contemporary with the beginning of geoglyph construction as registered in archaeological excavations. At both sites, these events accompany decreases in arboreal phytoliths and the continued rise of palm trees. Meanwhile at JS3-5 the construction of the Jaco Sá site apparently did not impact the local vegetation, revealing a localized deforestation pattern restricted to the site itself.

The maintenance of low percentages (<10%) of grasses, as well as δ¹³C values (in JS1), as present in levels below 35cm, suggest that the vegetation growing upon the geoglyphs was not kept completely open during their use. In fact, the highest values ​​of grasses (20%), charcoal (> 400 particles/cm²), and δ¹³C (-19.7%) found in the study come from surface samples (0-5 cm) and reflect a landscape deforested ~40 years ago for cattle farming. The uniqueness of this modern signal is strong evidence that pre-Columbian populations never practiced deforestation on a similar scale in the past.

Finally, where palms had increased over time, there is an abrupt reduction in these phytoliths near the top of the profiles, dated to 1280-1350 cal. CE in JS2 and JS4. This period corresponds to the end of the use of the geoglyph sites and the proliferation of circular mound villages in the region (Saunaluoma et al., 2018). It is possible that the abandonment of the geoglyphs meant that palm populations - previously maintained abundant by the management practices of those populations - were outcompeted by slow-growing tree species.

Raised fields of French Guiana

Figure 4
Summarized data of the pollen, phytoliths and charcoal extracted from the sediment core from the K-VIII site.

The K-VIII site sediment core reached a depth of 31.5 cm and gave a baseline date of 350-290 cal. BCE. The main results are summarized in Figure 4 and separated into three zones, delimited from the trends observed in the data.

In Zone 1, the high percentages of Cyperaceae in the pollen and phytolith records, as well as the presence of Marantaceae phytoliths, indicate that the landscape before the construction of the raised fields was a flooded savanna, as these two families consist of aquatic herbs typical of these environments. Charcoal influx into this zone, measured in particles per cm³ per year, remains low.

From ca. 1200 cal. CE, there is an abrupt drop in wetland taxa and a sudden increase in grass (Poaceae) pollen and phytoliths that marks the beginning of Zone 2 and the period of construction and use of the raised fields. The increase in grasses in this period was probably the result of the creation of new areas of terra firme - that is, the dry surfaces of the raised fields - while a concomitant increase in the subfamily Oryzoideae suggests an expansion of wetland margins. Maize pollen grains also appear for the first time in this zone, and the presence of this crop, as well as the colonization of the savanna by grasses, is repeated in phytolith assemblages extracted from the raised fields themselves (Iriarte et al., 2010). This zone also sees the influx of charcoal nearly triple.

An abrupt increase in charcoal ca. 1500 cal. CE marks the beginning of Zone 3 and coincides with the arrival of Europeans on the Guyanese coast. After a hundred-year hiatus, maize reappears ca. 1600 cal. CE, but is replaced by manioc and sweet potato pollen when the first permanent European settlements are established in the mid-17th century. It is uncertain whether these crops were planted in the raised fields or not, since their absence in phytolith assemblages from the fields is not conclusive: sweet potato does not produce diagnostic phytoliths, while manioc produces them in very low amounts. However, the presence of new structures, such as drainage channels and cultivation platforms built upon forested terraces near the site, suggests that agricultural practices during this period shifted to upland areas, and involved more frequent burning of regional vegetation.

The dark earths of the lower Tapajós

In the lower Tapajós region, from the beginning of the record (Figure 5), the elevated quantities of tree pollen, the presence of maize, and the low amounts of herbs and charcoal are suggestive that polyculture agroforestry was practiced by populations from ca. 2500 cal. BCE.

After ca. 550 cal. BCE there is an increase in the number of edible plants (from ~45% to > 70% of terrestrial pollen taxa), a decrease in the number of plants not documented as edible¹ (from ~50% to 30%), followed by the arrival of manioc ca. 300 cal. BCE. This pattern is not associated with significant change in the regional precipitation data, suggesting that forest enrichment by Indigenous populations at that time. The increase in regional fire activity between 300 and 1450 cal. CE, evidenced by increased charcoal influx, is associated with demographic growth over time.

The archaeobotanical data from Serra do Maguari 1, a 15-ha mounded village surrounded by a mosaic of ADE, show that construction occurred between 1420 and 1500 CE by users of Tapajônico (Santarém) period ceramics. Squash phytoliths are present in layers before the formation of the ADE, consistent with the presence of squash in the lake ca. 1350 cal. CE, however, maize only appears when the ADE forms.

Figure 5
Pollen and charcoal data from the sediment core of lake Lago Caranã, in the lower Tapajós region, presented alongside SPD values (“Regional archaeology), precipitation data (δ18O) from the Paraíso cave (Wang et al., 2017), and charcoal and phytolith data from a soil profile in the ADE of the Serra de Maguari 1 site.

Charcoal also increases with the formation of ADE soils from ca. 1420 cal. EC, suggesting that in-field burning was likely implemented in the process of creating these soils. However, sediment charcoal from Lake Caranã indicates an overall decrease in fire activity in the watershed at that time, synchronous with the driest regional climate conditions in the past 5500 years. Today, prolonged dry periods in the eastern Amazon are associated with great increases in fire activity in the region; thus, this suggests that pre-Colonial populations implemented management practices that suppressed wildfires. This interpretation is further supported by the continued presence of rainforest vegetation at Lake Carana (~30-45%) and the Maguari site (~56-82%), indicating that large-scale deforestation did not occur on the plateau over the past 4500 years.

From 1700 cal. CE onwards, there is a reduction in pollen from palms and edible tree species (from 70 to 45%), while maize, manioc, and squash remain present. Following a decline in charcoal influx, which coincides with a fall in SPD values, between 1450 and 1850 cal. CE, charcoal starts to return gradually during the last century, and rapidly from the 1970s onwards. The highest fire magnitude values are attributed to the last decade, accompanied by a 30% increase in herb pollen (primarily Ambrosia sp.)

The zanja sites of Iténez, Bolivia

Rainforest vegetation is present throughout the duration of the record, indicated by over 40% Moraceae/Urticaceae pollen (Figure 6). Maize pollen is present after 3750 cal. BCE along with the onset of low-level of fire activity. SPD values are consistent with human presence in the region. Climate data from the Pumacocha speleothem indicate climate conditions drier than the present around 4830 BCE that become progressively wetter after 3050 cal. BCE. Therefore, the increase in charcoal influx after 2550 cal. BCE is attributed to anthropogenic burning, which reached record levels ca. 850 cal. BCE.

Between 1000 and 700 BCE charcoal influx decreases, coupled with a decline in trees and shrubs (8%), an increase in palm pollen (3%), an increase in Mauritia/Mauritiella pollen concentration (20%), and the presence of maize. Data from the Triunfo site indicate the presence of manioc and leren (Calathea sp.) and episodes of local burning prior to, and during, the formation of ADEs. ADE soil formation begins ca. 400 BCE during the Chocolatal ceramic phase. Between 250 and 850 CE there is an increase in regional burning, a decrease in the concentration of Mauritia/Mauritiella and of trees and shrubs; meanwhile palms in general increase (10%) and there is the continued presence of maize and sweet potato. This period corresponds with slightly increased regional human activity indicated by the SPD values around 400 BCE. Stable, wetter climatic conditions are indicated by the δ¹⁸O values from Pumacocha (Bird et al., 2011).

After 850 cal. CE, the lake record indicates a decline in burning and maize pollen diminishes as it was only recorded in one sample dated to 1030 cal. CE. While palynology indicates that the regional vegetation was relatively stable, at the Triunfo site there is an increase in herb phytoliths (20%) at the expense of trees and shrubs, an increase in charcoal, and the continued presence of manioc, maize and leren. These changes occur in a part of the profile with darker ADE and ceramics from the Early Versalles phase. This archaeological phase (850-1150 cal. CE) corresponds to an abrupt increase in regional human activity indicated by the SPD and drier conditions associated with the Medieval Climate Anomaly.

After 1150 cal. CE the sediment core records the continuation of burning and cultivation activities associates with an abrupt increase in herbs at the expense of trees in location vegetation. This period corresponds with the construction of ditches (between 1300 and 1600 cal. CE) and is associated with increased SPD values as well as an increase in regional precipitation around 1250 to 1750 cal. CE associated with the Little Ice Age (Della Libera et al., 2022; Novello et al., 2017).

The modern signals show less palm pollen and an absence of crop pollen in association with the lowest charcoal values in the last 3500 years. While it is currently used by the local Itonama Indigenous community of Versalles, there are no settlements on the lake.

Figure 6
Pollen and charcoal data from the sediment core at lake Versalles, Iténez, Bolívia, alongside precipitation data (δ18O) from Cuica Cave (Della Libera et al., 2022), SPD values for the region (“Regional archaeology”), and the charcoal and phytoliths from a soil profile located at the Triunfo zanja site.

Discussion

In Table 1, we summarize the paleoecological data according to the socioeconomic thresholds suggested by Roberts et al. (2023). We do not intend to discuss the applicability of these categories in and of themselves, we simply assess the existence of changes in land use during the three main periods: origins of agriculture (4000-1000 BCE), demographic consolidation (0-1500 CE) and capitalism (1500 CE-present)..

Table 1
Summaries of data regarding land use in the three threshold periods for the Anthropocene reached by Roberts et al. (2023)

	4000-1000 BCE	0 - 1500 CE	1500 CE - present
Acre, Brasil	Increase ca. 2500 BCE	Increase from 0 CE	Increase 1970 CE
	fire activity	geoglyph construction	fire activity
	forest enrichment	fire activity	permanent deforestation
		forest enrichment
Coast of Guiana Francesa	No record	Increase ca. 1200 CE	Increase ca. 1500 CE and in the 20th century
		raised field construction	construction of canals
		fire activity	fire activity
		savanna vegetation transformation	terra firme deforestation
Lower Tapajós, Brasil	Increase ca. 2500 BCE	Increase from 0 EC	Increase 1970 CE
	fire activity	fire activity	fire activity
	polyculture agroforestry	forest enrichment
Iténez, Bolívia	Increase ca. 2500 BCE	Increase from 500 BCE	Decline
	fire activity	construction of ADEs and zanjas
		low-level fire activity
		palm enrichment

• Source: Created for this publication.

In almost all the records analyzed, stepped increases in land use - that is, in the human impact on the environment - ​​can be observed hat roughly coincide with the thresholds established by Roberts et al. (2023).

Starting around 2500 cal. BCE, we see an increase in burning activity associated with maize cultivation (lower Tapajós and Iténez) and the management of vegetation to favor useful plants (Acre and lower Tapajós). How these trends relate to the climate of the early Late Holocene is not clear: while the climate started to stabilize in Acre and the lower Tapajós, in Iténez, less stable climatic conditions persisted for longer.

The period 0-1500 cal. CE is characterized by regional increases in burning that correlate with population growth and the construction of geoglyphs, raised fields, zanjas and ADEs. Dark earths and raised fields have been interpreted as evidence for the intensification of Indigenous management systems (Iriarte et al., 2020a), transforming large areas of infertile soils into areas productive for cultivation. At the same time, paleoecological and archaeobotanical data from sites located in forest environments (Acre, lower Tapajós and Iténez) show that, these same populations: 1) enriched the vegetation with useful plants, increasing agrobiodiversity, and 2) maintained regionally stable forest cover (Iriarte et al., 2020b; Maezumi et al., 2018a; 2022; Watling et al., 2017). This forest management scenario (polyculture agroforestry) contrasts with the situation in the Old World, where Neolithic farmers destroyed large regions of forest that never recovered.

Circumstantial evidence from three of the regions (all except Acre) suggests that this was possible through differentiated burning practices, which helped to control or even suppress severe fires. In French Guiana, the increase in burning that occurred shortly after the raised fields were abandoned suggests that fire was previously suppressed in the fields during the dry season (Iriarte et al., 2012); in the lower Tapajós, the forest cover above the ADEs was maintained through low-severity burning (Maezumi et al., 2018b); while in Iténez, although there was a reduction in the quantity of charcoal reaching the lake in this period, the increase in Mauritia/Mauritella pollen concentration suggests that, regionally, low severity burns were increasing (Maezumi et al., 2022).

These management systems left enduring legacies on today´s forest composition. Botanical inventories conducted in Acre and the lower Tapajós together with this research show increases in the frequency and richness of useful plants close to the geoglyphs and on dark earths soils, when compared to adjacent areas.

After the European invasion ca. 1500 CE, none of the paleoecological records demonstrate a “bounce back” of forest cover. In fact, in Acre, the lower Tapajós and Iténez, cultural changes which happened previously are partly reflected in our records, coinciding with the Medieval Climate Anomaly (De Souza et al., 2019). Such a scenario calls into question the hypothesis that the “Great Dying” of Indigenous populations, and the subsequent reforestation of the Amazon, would have been large enough to cause the Little Ice Age (Koch et al., 2019). This may be attribute to the fact that Indigenous land use practices, such as agroforestry (Maezumi et al., 2018a) did not practice large-scale forest clearance and rather maintained closed-rainforest canopies around the sites.

In three of the studied regions (excluding Iténez), activities related to the burning and removal of native vegetation have increased dramatically during the last 500 years. On the French Guiana coast, early European colonization of the region accompanied an abrupt change in fire regimes ca. 1500 cal. CE, which intensified further during the 20^th century. In Acre and the lower Tapajós, however, these processes only seem to have gained strength in the second half of the 20th century.

At Lake Caranã, the Rubber Boom - whose first phase lasted from the 19th century until 1912, was characterized by the lowest burning levels in the entire record. The Tapajós river was one of the places most affected by rubber extraction in the Amazon. Given the quantity of people that migrated to the area, and the fact that fire was considered one of the biggest threats to rubber plantations (Schroth et al., 2003), the lack of fires in this period is interpreted as the use of fire suppression practices by these populations (Maezumi et al., 2018b). It is only from the 1970s, when the FLONA reserve was established and public policies tried to suppress the use of fire in the region, that the frequency and severity of regional fires reached a much higher level than at any previous time (Maezumi et al., 2018b). This trend is synchronous with the incursion of deforestation practices for cattle ranching and soybean plantations, which aggravate forest drying already in course by anthropogenic climate change (Maezumi et al., 2018b). In Acre, the sharp increase in fire and vegetation removal in the top of the profiles is associated with the colonization of the region from the 1970s onwards and the establishment of cattle ranching (Watling et al., 2017).

Data from Laguna Versalles contradict this pattern and show a decline in anthropogenic impact during the colonial period. In general, this region has suffered less from the incursion of agriculture and other extractivist cycles, and until today the majority of the population in Versalles is Indigenous and practices traditional landscape management. Palynological studies of Lagunas Oricore and La Granja, in the neighbouring region of Bella Vista, also in the province of Iténez, (Carson et al., 2014), show the same pattern of reduced regional fires during the last 500 years.

Conclusion

While more multidisciplinary and integrative work needs to be done in other Amazonian regions, the four case studies presented converge on some important points that have implications for how we think about the conditions that precipitated the Amazonian Anthropocene.

First, it seems that the period beginning ca. 2500 cal. BCE marked one of large-scale environmental transformations, and that there was a more abrupt increase in these transformations from ca. 0 CE. Whether this last phase can be considered as the beginning of an Amazonian Anthropocene is a topic open for debate. On the one hand, there is no doubt that the composition of vegetation and soils was highly modified in this period, and that the construction of large earthworks resulted in new physical and symbolic landscapes. However, paleoecological data suggest that such transformations, rather than cause negative ruptures within the existing ecosystems, were able to maintain vital ecosystem services, mainly through the maintenance of forest cover. Instead of being destroyed, people built new relationships with other forest beings in this period. In the Amazonian case, this version of the Anthropocene did not imply a split with the “natural” world - a fact that perhaps was the result of unique Amerindian ontologies and ways-of-inhabiting-the-world (e.g. Fausto; Neves, 2018).

Finally, in several regions of the Amazon, the most intensive and destructive anthropic impacts happened after the European invasion, especially during the 20^th century. Large-scale deforestation and the occurrence of wildfires should concern us, as the lack of analogues for this type of land use in the past implies that the recovery of these ecosystems in the future is not a given. Furthermore, it is likely no coincidence that Iténez, which was the only region to record a decline in human impact during the colonial period, is the least-colonized region of the four examined in this study.

References

• ARROYO-KALIN, M. Landscaping, landscape legacies, and landesque capital in pre-Columbian Amazonia. In: The Oxford Handbook of Historical Ecology and Applied Archaeology. Oxford: Oxford Univ. Press, 2016.
• ARROYO-KALIN, M.; RIRIS, P. Did pre-Columbian populations of the Amazonian biome reach carrying capacity. PTRS B: Biological Sciences, v.376, n.1816, p.20190715, 2020.
• BAKER, P. A. et al. The history of South American tropical precipitation for the past 25,000 years. Science, v.29, n.5504, p.640-3, 2001.
• BELCHER, C. M. et al. What Can Charcoal Reflectance Tell Us About Energy Release in Wildfires and the Properties of Pyrogenic Carbon? FES, v.6, p.169, 2018.
• BIRD, B. W. et al. A 2,300-year-long annually resolved record of the South American summer monsoon from the Peruvian Andes. PNAS, v.108, n.21, p.8583-8, 2011.
• BUSH, M. B.; COLINVAUX, P. A. A 7000-year pollen record from the Amazon lowlands, Ecuador. Vegetation, v.76, p.141-54, 1988.
• CAROMANO, C. F. et al. Revealing fires and rich diets: macro-and micro-archaeobotanical analysis at the Hatahara Site, Central Amazonia. Tipití, v.11, n.2, p.40-51, 2013.
• CARSON, J. F. et al. Environmental impact of geometric earthwork construction in pre-Columbian Amazonia. Proceedings of the National Academy of Sciences, v.111, n.29, p.10497-502, 2014.
• CLEMENT, C. R. et al. Disentangling Domestication from Food Production Systems in the Neotropics. Quaternary, v.4, n.1, p.4, 2021.
• COLINVAUX, P. A.; DE OLIVEIRA, P. E.; BUSH, M. B. Amazonian and neotropical plant communities on glacial time-scales: The failure of the aridity and refuge hypotheses. QSR, v.19, n.1-5, p.141-69, 2000.
• CRUTZEN, P. J.; STOERMER, E. F. The ‘Anthropocene’ (2000). In: The Future of Nature: Documents of Global Change. New Haven: Yale Univ. Press, 2013. p.479-90.
• DE SOUZA, J. G. et al. Climate change and cultural resilience in late pre-Columbian Amazonia. NEE, v.3, n.7, p.1007-17, 2019.
• DELLA LIBERA, M. E. et al. Paleoclimatic and paleoenvironmental changes in Amazonian lowlands over the last three millennia. QSR, v.279, p.107383, 2022.
• DENEVAN, W. M. A bluff model of riverine settlement in prehistoric Amazonia. Annals of the AAG, v.86, n.4, p.654-81, 1996.
• _______. Cultivated landscapes of Native America and the Andes. Oxford: Oxford Univ. Press, 2001.
• _______. Pre-European forest cultivation in Amazonia. In: Time and Complexity in Historical Ecology. New York: Columbia Univ. Press, 2006. p.153-64.
• EDER, F. J. Breve Descrición de las Reducciones de Mojos. Historia Boliviana. 1985.
• ERICKSON, C. L. Historical ecology and future explorations. In: Amazonian Dark Earths: Origins, Properties, Management. New York: Springer, 2003. p.455-500.
• _______. The transformation of environment into landscape: the historical ecology of monumental earthwork construction in the Bolivian Amazon. Diversity, v.2, n.4, p.618-52, 2010.
• ERICKSON, C. L.; WALKER, J. H. Pre-Columbian causeways and canals as landesque capital. In: Landscapes of Movement: Trails, paths and roads in anthropological perspective. Philadelphia: Penn Museum Press and Univ. of Pennsylvania Press, 2009. p.232-52.
• FAUSTO, C.; NEVES, E. G. Timeless gardens: deep indigenous history and the making of biodiversity in the Amazon. In: Tropical Forest Conservation: Integrating natural and cultural diversity for sustainability, a global perspective. Mexico City: Unesco Mexico, 2018.
• FLANTUA, S. G. et al. Updated site compilation of the Latin America pollen database. RPP, v.223, p.104-15, 2015.
• GLASER, B.; BIRK, J. J. State of the scientific knowledge on properties and genesis of Anthropogenic Dark Earths in Central Amazonia. Geochimica et Cosmochimica Acta, v.82, p.39-51, 2012.
• GOSLING, W. D.; CORNELISSEN, H. L.; MCMICHAEL, C. N. H. Reconstructing past fire temperatures from ancient charcoal material. PPP, v.520, p.128-37, 2019.
• HECHT, S. B. Indigenous soil management and the creation of Amazonian Dark Earths: implications of Kayapó practice. In: Amazonian Dark Earths: Origins, Properties and Management. New York: Springer, 2003. p.335-72.
• HIGUERA, P. E. et al. Peak detection in sediment-charcoal records: Impacts of alternative data analysis methods on fire-history interpretations. IJWF, v.19, p.996-1014, 2011.
• HILL, J. et al. Differing local-scale responses of Bolivian Amazon forest ecotones to middle Holocene drought based upon multiproxy soil data. JQS, p.1-21, 2023.
• IRIARTE, J. et al. Late Holocene Neotropical agricultural landscapes: phytolith and stable carbon isotope analysis of raised fields from French Guianan coastal savannahs. JAS, v.37, n.12, p.2984-94, 2010.
• IRIARTE, J. et al. Fire-free land use in pre-1492 Amazonian savannas. PNAS, v.109, n.17, p.6473-8, 2012.
• IRIARTE, J. et al. The origins of Amazonian landscapes: Plant cultivation, domestication and the spread of food production in tropical South America. QSR, v.248, p.106582, 2020a.
• IRIARTE, J. et al. Geometry by design: Contribution of lidar to the understanding of settlement patterns of the mound villages in SW Amazonia. JCAA, v.3, n.1, p.151-69, 2020b.
• JUNQUEIRA, A. B.; SHEPARD, G. H.; CLEMENT, C. R. Secondary forests on anthropogenic soils in Brazilian Amazonia conserve agrobiodiversity. Biodiversity and Conservation, v.19, n.7, p1933-61, 2010.
• KERN, D. C. et al. Distribution of Amazonian Dark Earths in the Brazilian Amazon. In: Amazonian Dark Earths: Origins, Properties, Management. New York: Springer, 2003. p.51-75.
• KOCH, A. et al. Earth system impacts of the European arrival and Great Dying in the Americas after 1492. QSR, v.207, p.13-36, 2019.
• LEVIS, C. et al. How people domesticated Amazonian forests. FEE, v.5, p.171, 2018.
• LEYS, B. et al. Fire history reconstruction in grassland ecosystems: Amount of charcoal reflects local area burned. ERL, v.10, n.11, p.114009, 2015.
• LINS DE PAIVA, H. C.; GALLO, V. Paleoecologia x Ecologia: Uma interação necessária? SUSTINERE, v.9, n.1, p.254-96, 2021.
• MAEZUMI, S. Y. et al. The legacy of 4,500 years of polyculture agroforestry in the eastern Amazon. Nature Plants, v.4, n.8, p.540-47, 2018a.
• MAEZUMI, S. Y. et al. New insights from pre-Columbian land use and fire management in Amazonian Dark Earth forests. FEE, v.6, n.111, p.1-23, 2018b.
• MAEZUMI, S. Y. et al. Legacies of Indigenous land use and cultural burning in the Bolivian Amazon rainforest ecotone. PTRS B, v.377, n.1849, p.20200499, 2022.
• MAYLE, F. E.; IRIARTE, J. Integrated palaeoecology and archaeology - a powerful approach for understanding pre-Columbian Amazonia. JAS, v.51, p.54-64, 2014.
• MCMICHAEL, C. H. et al. Comment on “Persistent effects of pre-Columbian plant domestication on Amazonian forest composition”. Science, v.355, n.6328, p.925-31, 2017.
• MONGELÓ, G. Z. Ocupações humanas do Holoceno inicial e médio no sudoeste amazônico. Bol. MPEG, v.15, n.2, p.e20190079, 2020.
• MORA, S. et al. Cultivars, Anthropic Soils, and Stability. Pittsburgh: Univ. of Pittsburgh, LAA, 1991.
• MORCOTE-RÍOS, G. et al. Colonisation and early peopling of the Colombian Amazon during the Late Pleistocene and the Early Holocene: New evidence from La Serranía La Lindosa. Quaternary International, v.578, p.5-19, 2021.
• NEVES, E. G. et al. Peoples of the Amazon before European colonization. In: Science Panel for the Amazon: Amazon Assessment Report 2021. New York: UNSDSN, 2021. p.1-40.
• NOVELLO, V. F. et al. A high-resolution history of the South American Monsoon from last glacial maximum to the Holocene. SR, v.7, p.1-8, 2017.
• PETERSEN, J. B.; NEVES, E. G.; HECKENBERGER, M. J. Gift from the past: terra preta and prehistoric Amerindian occupation in Amazonia. In: Unknown Amazon: Culture and nature in ancient Brazil. London: The British Museum Press, 2001. p.86-106.
• POSEY, D. A.; BALÉE, W. (Ed.) Resource Management in Amazonia: Indigenous and folk strategies, AEB, v.7. New York: New York Botanical Garden, 1989.
• PRÜMERS, H.; JAIMES BETANCOURT, C. 100 anos de investigación arqueológica en los Llanos de Mojos. Arqueoantropológicas, v.4, p.11-54, 2014.
• ROBERTS, P. et al. Mapping our reliance on the tropics can reveal the roots of the Anthropocene. NEE, v.7, p.632-36, 2023.
• ROBINSON, M. et al. Anthropogenic soil and settlement organisation in the Bolivian Amazon. Geoarchaeology, v.36, n.3, p.388-403, 2020.
• ROOSEVELT, A. C. Chiefdoms in the Amazon and Orinoco. In: Chiefdoms in the Americas. Lanham, Maryland: Uni. Press of America, 1987. p.153-84.
• ROOSEVELT, A. C. The Amazon and the Anthropocene. Anthropocene, v.4, p.69-87, 2014.
• ROSTAIN, S. Pre-Columbian earthworks in coastal Amazonia. Diversity, v.2, n.3, p331-52, 2010.
• ROSTAIN, S.; VERSTEEG, A. J. The Arauquinoid Tradition in the Guianas. In: Late Ceramic Age Societies in the Eastern Caribbean. Oxford: BAR International Series 1273, 2004. p.223-50.
• SAUNALUOMA, S.; VIRTANEN, P. K. Variable models for organization of earthworking communities in Upper Purus, southwestern Amazonia: Archaeological and ethnographic perspectives. Tipití, v.13, n.1, p.23-43, 2015.
• SAUNALUOMA, S.; PÄRSSINEN, M.; SCHAAN, D. Diversity of pre-colonial earthworks in the Brazilian state of Acre, southwestern Amazonia. JFA, v.43, n.5, p.362-79, 2018.
• SAUNALUOMA, S.; SCHAAN, D. Monumentality in Western Amazonian formative societies: geometric ditched enclosures in the Brazilian state of Acre. Antiqua, v.2, n.1 p.1-11, 2012.
• SCHAAN, D. et al. New radiometric dates for precolumbian (2000-700 BP) earthworks in western Amazonia, Brazil. JFA, v.37, n.2, p.132-42, 2012.
• SCHROTH, G. et al. Rubber agroforests at the Tapajós river, Brazilian Amazon. AEE, v.97, p.151-65, 2003.
• TEIXEIRA, W. G. et al. As Terras Pretas de Índio da Amazônia. Manaus: Embrapa, 2010.
• WALKER, J. H. Island, River and Field: Landscape archaeology in the Llanos de Mojos. Albuquerque (NM): UNM Press, 2018.
• WANG, X. et al. Hydroclimate changes across the Amazon lowlands over the past 45,000 years. Nature, v.541, n.7636, p.204-7, 2017.
• WATLING, J. et al. Impact of pre-Columbian “geoglyph” builders on Amazonian forests. PNAS, v.114, n.8, p.1868-73, 2017.
• WATLING, J. et al. Direct archaeological evidence for Southwestern Amazonia as an early plant domestication and food production centre. PLoS One, v.13, n.7, p.e0199868, 2018a.
• WATLING, J.; MAYLE, F. E.; SCHAAN, D. Historical ecology, human niche construction and landscape in pre-Columbian Amazonia: a case study of the geoglyph builders of Acre, Brazil. JAA, v.50, p.128-39, 2018b.
• WHITMORE, T. M.; TURNER II, B. L., Cultivated Landscapes of Middle America on the Eve of Conquest. Oxford: Oxford Univ.Press, 2001.
• WOODS, W. I. et al. (Ed.) Amazonian Dark Earths: Wim Sombroek’s vision. [S.l.]: Springer, 2009.

Note

• 1
  This division of plants, between edible and without documentation as edible focuses on only the recorded uses by humans and is limited by extant ethnographic documentation. The research sought to document transformations provoked by management and considered edible all genera where at least one species was so documented in ethnographic literature.

Publication Dates

• Publication in this collection
  29 Nov 2024
• Date of issue
  Sep-Dec 2024

History

• Received
  06 June 2023
• Accepted
  03 June 2024