| Summary: | This thesis presents a multi-proxy reconstruction of Holocene sea-level change at the Steart Peninsula in Bridgwater Bay, Somerset. Single- and multi-proxy transfer functions were developed from diatom and foraminifera training data from two sites on the Severn Estuary coast. Constrained and partial constrained ordination estimated that 5.5% of the intertidal diatom variation and 11.4% of the intertidal foraminifera variation in the training sets was explained independently by elevation within the tidal frame. Foraminifera provided more favourable transfer function prediction statistics but were low in abundance in parts of the intertidal zone, resulting in a smaller than optimal training set. The preferred transfer function combined diatom and foraminifera data to optimise prediction statistics and intertidal coverage. An 11.22m fossil core was retrieved from the Steart Peninsula for stratigraphic and micropalaeontological analyses. Analogue matching revealed that combining the two proxies increased the number of fossil samples with good or fair modern analogues in relation to the single proxies separately. The preferred multi-proxy transfer function was applied to the fossil micropalaeontological data. The sediment stratigraphy, biostratigraphy and transfer function-based estimates indicated an overall rise in relative sea-level of about 13m between 7,582-7,345 and 1,804-1,690 cal. years before present (cal. yr BP), beginning with an initial rapid phase of sea-level rise and silty clay deposition in a minerogenic saltmarsh environment. Between 6,188-6,007 and 3,942-3,759 cal. years BP three main fluctuations in marine influence occurred, allowing organic upper saltmarsh conditions to develop periodically, but the data scatter and large vertical error bars prevent a detailed interpretation of absolute sea-level change. Diatoms and foraminifera were sparse or absent between 4,053-3,869 and 1,927-1,806 cal. years BP. Finally, foraminifera assemblages indicated a possible tidal flat environment and increase in marine influence between 1,927-1,806 and 1,682-1,619 cal. years BP. This research concludes that multi-proxy methods have the potential to improve the accuracy and precision of relative sea-level reconstruction in an extremely macrotidal setting such as that of the Severn Estuary, but that a number of issues such as sediment autocompaction and possible Holocene tidal range changes need to be quantified in order for the broad sea-level history presented to be constrained further.
|
|---|
