This thesis investigates sea-level changes in order to test the hypothesis that the main contribution to early 20th century sea-level rise was Northern Hemisphere land-based ice melt. Multiproxy sea-level reconstructions were established for the Falkland Islands, a location where models suggest sea-level rise from Northern Hemisphere ice melt produces the largest signal. The Falklands reconstruction indicated sea levels in the early 20th century accelerated compared to the long-term rate, synchronous with accelerations observed globally. The magnitude of the acceleration in the Falklands reconstruction was greater than Northern Hemisphere rates, consistent with the spatial pattern from a Northern Hemisphere melt source, but likely less than in New Zealand and Australia. It is therefore not possible rule out other contributions to the observed sea-level acceleration. The Falklands reconstruction indicated a rapid sea-level jump around 8.4 ka BP, synchronous with a jump observed in the Northern Hemisphere, which has been attributed to the sudden drainage of Laurentide proglacial lake Agassiz-Ojibway associated with the 8.2 ka BP climatic downturn. A maximum estimate of 0.89 ± 0.22 m for this jump in the Falklands is considerably less than estimates from Northern Hemisphere records. This difference could indicate additional contributions from the Southern Hemisphere are being recorded in the Northern Hemisphere signal. This thesis also focused on developing testate amoebae as sea-level indicators. In the Falklands, testate amoebae transfer functions were able to reconstruct sea level with precision (±0.08 m) comparable to diatoms (±0.07 m). However, preservation issues were indicated in the fossil testate amoebae assemblages which limits their use as tools for sea-level reconstruction. In addition, contemporary distributions of salt-marsh testate amoebae were investigated over one annual cycle. Seasonal variations in the live assemblages were observed to be asynchronous between taxa. Variations in the death assemblages were also observed which were correlated with variations in the live assemblages. This observation suggests the commonly applied assumption in palaeoenvironmental studies that analysing the death population negates temporal bias is invalid. Further research is required to investigate the impact these observed variations have on reconstructive performance.

Document Type


Publication Date




Creative Commons License

Creative Commons Attribution-NonCommercial 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License