Abstract/Summary

Dry forests are the most threatened tropical biome in South America, with only 10% of their original cover remaining. Chapter 1 reviews literature on tropical dry forests and palaeoecology to demonstrate why the impact of the drier climatic conditions predicted for much of South America in the coming century on these forests is uncertain. This thesis uses the mid-Holocene (6000 years BP) drier climatic period to provide insight into the long-term ecological responses of tropical dry forest to drier conditions. One of the reasons for uncertainty over the impacts of the mid-Holocene drier climate on these forests is the lack of large lakes from which to draw palaeoecological records. The large lakes that do exist within the dry forest biome in Bolivia are situated in complex catchments with hydrological connections to tropical wetlands. In this thesis we therefore first develop our understanding of a relatively new palaeoecological proxy –phytoliths –to assist in our interpretation of the mid-Holocene palaeo-vegetation record. Chapter 2 demonstrates the palaeoecological potential of phytoliths from lake sediments, including the first empirical evidence that phytoliths from lakes most strongly represent local vegetation. Chapter 3 builds on these findings to investigate the sensitivity of phytoliths to vegetation changes within key Bolivian ecosystems, and compares it to pollen, showing that phytoliths are most sensitive to changes within ecosystems with strong herbaceous components such as forest understories. Chapter 4 uses the results of the previous chapters to investigate the response of the Bolivian dry forest to mid-Holocene drier climatic conditions, demonstrating resilience of the dry forest as a biome, with a shift in composition towards more drought-tolerant taxa. However, savannah encroachment did occur at the ecotone during the mid-Holocene, with fire identified as potentially reinforcing the ecotone shift. The vulnerability of tropical dry forests to future climate change will therefore depend on complex interactions between climate, fire and human land-use.