HUMANS, VEGETATION, AND DROUGHT IN LATE HOLOCENE AMAZONIA

Abstract

The resilience and adaptation of past societies to climate change remain unclear, especially in the tropical cloud-forests of the Andes and the aseasonal forests of lowland western Amazonia. Long thought to have been too humid for significant human settlement, a new image of cloud-forests is emerging in which they supported human settlements for millennia. Anthropogenic disturbance, such as forest clearance, fire activity, and land-use has been suggested to have intensified during times of drier climates. Even so, the mid-elevational Andes and western Amazonia probably supported smaller pre-Columbian populations than settings with a distinct dry season. Uncertainties persist, however, regarding the spatial and temporal extent of pre-Columbian occupation in western Amazonia. Prior paleoecological studies have reported human disturbance in this region as temporally discontinuous, with localized disturbance strongly influenced by distance from rivers, lakes, or seasonal flood-plain/savanna settings. The trajectory of Amazonian population growth is often portrayed to be exponential in the late Holocene, leading to the expectation that peak occupation and disturbance were interrupted by European colonization. Some recent studies from Amazonia point to peaks of land clearance occurring in the period 1150–950 cal. yr BP rather than around 450 cal. yr BP, perhaps complicating the simple narrative of steady population growth prior to collapse after European contact. Here, I derive paleoecological data from a mid-elevation lake, Lake Condores, in the Peruvian cloud-forest and from two lakes in lowland Amazonia, Lakes Ayauchi and Kumpaka . I use high-resolution pollen, phytolith, charcoal, sediment chemistry, and diatom data to provide detailed reconstructions of past land-use and connections to climate. By investigating the anthropogenic impacts of these settings over the past 2100–2500 years, I ask the key questions: Were times of drier climate characterized by more severe human disturbance? Did human manipulation of the lake basins peak at the time of European arrival? Lake Condores, Peru, was characterized by forest clearance, burning, and maize cultivation from the onset of the record ca. 2085 cal. yr BP until ca. 850 cal. yr BP, where increased times of land-use were associated with profound droughts. Forest regrowth and reduced maize cultivation reshaped the valley after 850 cal. yr BP, shortly preceding the start of burial activity in adjacent cliff tombs about 700 cal. yr BP amid a long-term shift to wetter conditions. In Ecuador, Lake Ayauchi provided a 2460-year long history of continuous maize cultivation until 550 cal. yr BP where fire was actively used as a tool for land clearance and forest suppression. Lake Kumpaka , Ecuador, was the least impacted of the three sites and provided a 2460-year history of discontinuous maize cultivation. Sparse evidence of fire usage suggested a preference for slash-and-mulch rather than slash-and-burn to clear forest. At Kumpaka , maize cultivation ceased ca. 750 cal. yr BP, long before European contact. Increased land-use at Ayauchi and Kumpaka coincided with periods of drier conditions whereas wetter periods were characterized by abandonment or reduced land-use. Neither of the three sites showed indications of a steady population growth interrupted by European arrival; rather, land-use appeared to peak before 650 cal. yr BP. Old-growth forest elements remained abundant (>40%), as intensified periods of land-use and forest clearance alternated with times of less intense landuse or abandonment. Overall, I find evidence of people exploiting these wet western Amazonian settings, opportunistically expanding into previously forested areas during dry events, but abandoning marginal areas or changing land-use as conditions became too wet. These data support both the importance of the environment in shaping human actions, and also the capacity for humans to prioritize their actions and adapt to new conditions.