Holocene relative sea-level histories of tropical islands

Abstract

Holocene relative sea-level (RSL) data offer important context for future sea-level changes. Predictions of the spatial patterns of RSL change to deeply uncertain, but potentially catastrophic ice-sheet melt rely on glacial isostatic adjustment (GIA) models that are calibrated to Holocene RSL data. In particular, RSL changes in tropical locations distal from the deglaciation centres of past ice sheets are sensitive to the input of meltwater into the oceans (otherwise known as ‘ice-equivalent sea-level changes’). These tropical Holocene RSL records provide key information to fine-tune the ice models assumed in GIA models. Holocene RSL data have also been used to unravel the relationships between past climates, ice sheets and sea level, offering insights into the processes and interactions driving RSL changes. However, our understanding of Holocene RSL changes is hindered by a lack of high-resolution data in tropical regions. The potential contribution of Antarctica to periods of rapid RSL rise (meltwater pulses) in the past are still debated due to the large uncertainty and temporal gaps in early- to mid-Holocene RSL data – despite Antarctica being the largest source of uncertainty in future sea-level projections.

In this thesis, I produced quality-controlled, high-resolution Holocene RSL data from tropical regions that have insofar been lacking. I produced a standardised, quality-controlled database of published Holocene RSL data from mid Pacific islands, which had been formerly used to validate GIA models and infer ice-equivalent sea-level changes. I then produced two new Holocene RSL reconstructions using two different sea-level proxies (coral microatolls and mangrove peat) in two different tropical settings. I demonstrate the utility of coral microatolls to serve as precise indicators of Late Holocene RSL in places like Singapore, where the lack of accommodation space due to RSL fall from GIA has formerly limited the use of sedimentary indicators like mangrove peat. I produced the longest, continuous Holocene RSL record stretching back 8500 yrs BP from mangrove peat in Belize. The success of this reconstruction attests to the use of sedimentary indicators in areas of continued post-glacial RSL rise, which creates the accommodation space necessary for the accretion of an undisturbed sequence of mangrove peat. The thesis demonstrates the challenges and successes in producing high-resolution, accurate tropical Holocene RSL reconstructions – and highlights the methods and potential avenues for producing more successful tropical Holocene RSL reconstructions moving forward.