Pre-Columbian land use and human impact in the Bolivian Amazon

There is a polarised debate amongst Neotropical archaeologists and ecologists over the extent of Pre-Columbian (pre-AD 1492) anthropogenic environmental impacts in Amazonia. While some maintain the old paradigm of pre-Columbian Amazonia as a “pristine wilderness”, which was sparsely populated by humans, others point to the discovery of an increasing number of archaeological sites across the Amazon basin as evidence for large, complex societies, supported by intensive agriculture and management of forest and aquatic resources. Under this model, pre- Columbian people had profound impacts on biogeochemical cycles through largescale deforestation and biomass burning, and left an indelible legacy on Amazonian ecosystems by altering edaphics, geomorphology, hydrology, and the distribution of plant taxa. However, so far this debate has continued amongst a paucity of palaeoecological data, meaning that we have little evidence to indicate the scale of impact, and often no palaeoenvironmental context in which to place these societies. Other aspects of pre-Columbian cultures, such as their chronology, land use practices and subsistence strategies, are also poorly understood. In this thesis, palaeoecological methods are applied to improve our understanding of the scale, nature, and legacy of land use associated with pre- Columbian geometric earthwork cultures in north-east Bolivia. The methodology employs analysis of fossil pollen and macroscopic charcoal from lake core sediments. In total 110 pollen and 628 charcoal samples were analysed from four lakes: Laguna Isireri (14°49’18”S, 65°40’57”W), Laguna Orícore (13°20’44”S, 63°31’31”W), Laguna Granja (13°15’44”S, 63°42’37”W) and Laguna La Luna (13°21’20”S, 63°35’2”W). Chronologies were built for these cores from 12 AMS 14C dates. Results are presented in four chapters (thesis chapters 2-5), in the form of academic papers. Chapter 2 describes a new laboratory technique co-developed with Whitney et al., which improves the recovery of cultigen pollen grains from sediments from large, Neotropical lake basins. This technique was applied in the laboratory preparations of sediments analysed for Chapters 3-5. Chapter 3 employs palaeoecological reconstructions from differently sized lakes (L. Granja and L. Orícore) in the same earthwork region, to obtain both a regional- and a local-scale history of environmental change/human impact around the archaeological site. Key findings of this paper are: 1) earthworks were built in an originally open savannah landscape, which existed under drier-than-present climatic conditions in the mid-to-late-Holocene; 2) forest expanded into this region from ~ 2000 (cal yrs) BP and was supressed locally around the settlement to maintain an open landscape; therefore, 3) earthwork construction across southern Amazonia, may not have required extensive deforestation, and pre-Columbian impacts on biogeochemical cycling may have been much less than some authors have suggested. Chapter 4 looks more closely at the local scale record provided by L. Granja. These data are integrated with phytolith data analysed by co-author J. Watling and existing archaeological data, to discuss the chronology of settlement on the site, the agricultural/land use strategies employed by its inhabitants, and the spatial scale and distribution of impacts locally around the site. Key findings are that: 1) first occupation of the site is much earlier than previously dated from archaeological contexts, beginning ~2500 BP; 2) maize was the staple crop grown on site; 3) land use involved more extensive and intensive burning of the landscape than compared to modern slash-and-burn agriculture; 4) site decline occurred ~ 500 BP, and may have been related to the Columbian Encounter of AD1492; 5) the close integration of local scale palaeoecological records with archaeology, is highly useful in discerning aspects of chronology and spatial variability of land use. Chapter 5 presents a 6000 year record of palaeoenvironmental change and land use on a pre-Columbian forest island site. Key findings are that: 1) As in Chapter 3, inhabitants exploited an originally open landscape and practiced forest suppression to maintain that open landscape; 2) the earliest recorded evidence for maize agriculture in the region is found at 2100 BP; 3) the economically useful species Theobroma cacao, which is abundant on the site today, is not detectable in the pollen record; 4) clear-cutting was not practiced on the site and previous population estimates, based on labour for deforestation, must be reconsidered. The work in this thesis reveals a new model of human-environment interactions, demonstrating that pre-Columbian earthwork cultures in southern Amazonia occupied and adapted to a region of dynamic, climatically controlled forest-savannah transition during the mid-to-late-Holocene. Obtaining a palaeoenvironmental context for archaeological landscapes, is shown to be a vital pre-requisite to inferring past environmental impacts. Furthermore, we demonstrate the valuable contribution that palaeoecology can make to a better understanding of the chronology and land use practices of pre-Columbian cultures.