Quantifying local-scale changes in Amazonian forest cover using phytoliths

Authors:
Open Access
Publisher website
Google Scholar
Abstract
The ecosystem services and immense biodiversity of Amazon rainforests are threatened by deforestation and forest degradation. A key goal of modern archaeology and paleoecology in Amazonia is to establish the extent and duration of past forest disturbance by humans. Fossil phytoliths are an established proxy to identify the duration of disturbance in lake sedimentary and soil archives. What is not known, is the spatial scale of such forest disturbances when identified by phytoliths. Here we use phytolith assemblages to detect local-scale forest openings, provide an estimate of extent, and consider long-term forest recovery. We use modern phytolith assemblages of 50 Amazonian lakes to i) assess how phytolith assemblages vary across forest cover at 5 spatial scales (100 m, 200 m, 500 m, 1 km, 2 km), ii) model which phytolith morphotypes can accurately predict forest cover at 5 spatial scales, and iii) compare phytoliths with pollen to quantify their relative ability to detect forest cover changes. DCA results show phytolith assemblages could be used to differentiate low, intermediate, and high forest cover values, but not to distinguish between biogeographical gradients across Amazonia. Beta regression models show Poaceae phytoliths can accurately predict forest cover within 200 m of Amazonian lakes. This modern calibration dataset can be used to make quantitative reconstructions of forest cover changes in Amazonia, to generate novel insights into long-term forest recovery. Combining phytoliths and pollen provides a unique opportunity to make qualitative and quantitative reconstructions of past vegetation changes, to better understand how human activities, environmental and climatic changes have shaped modern Amazonian forests.

This publication has 80 references indexed in Scilit: