Copy the page URI to the clipboard
Julier, Adele Charlotte Mary
(2018).
DOI: https://doi.org/10.21954/ou.ro.0000d450
Abstract
Understanding the pollen assemblages produced by modern tropical vegetation, and improving the taxonomic resolution of pollen identification are vital in generating high quality interpretations of fossil pollen assemblages. Lake Bosumtwi, Ghana, is a 1 million year old meteorite impact crater-lake, in which pollen has been deposited tracking vegetation change, including the expansion and contraction of grass dominated landscapes evidenced by very high proportions of Poaceae pollen. This thesis explores the modern pollen-vegetation relationships of a forest-savannah transitional mosaic, a moist semi-deciduous forest and a wet evergreen rainforest in Ghana, using artificial pollen traps. It also tests the ability of Fourier Transform Infra-red Spectroscopy (FTIR) to identify Poaceae pollen to below family level.
Characteristic taxa from the forest-savannah mosaic landscape were Poaceae and Melastomataceae/Combretaceae, with these reaching 61% and 73% respectively. The moist semi-deciduous site was characterised by Celtis (up to 89%) and Triplochiton (up to 20%), and the wet evergreen rainforest was characterised by Cynometra, Drypetes, Vitex and Homalium (around 10% of pollen sum each). It was found that, using FTIR spectroscopy, it is possible to achieve an 80% classification success rate of pollen to sub-family level within the Poaceae. These results suggest that the threshold of 55% Poaceae pollen previously used to mark the transition between grass dominated and forested landscapes in the Lake Bosumtwi record may be too high, and should be closer to 40%. It is also concluded that the assemblages recovered from interglacial periods of Lake Bosumtwi may represent vegetation that was less similar to wet rainforest and closer to forest-savannah mosaic.
This thesis should inform future studies of fossil pollen assemblages recovered from West Africa, and further work on the Lake Bosumtwi record, to enable higher resolution interpretations of fossil pollen assemblages and, therefore, a better understanding of Earth’s past.