THE APPLICATION OF LOCAL AND REGIONAL TRANSFER FUNCTIONS TO RECONSTRUCT FORMER SEA LEVELS, NORTH NORFOLK, ENGLAND

Foraminiferal assemblages from Thornham and Brancaster marshes (Norfolk, UK) illustrate that elevation with respect to mean sea level is the most statistically significant controlling variable. We developed local (Thornham and Brancaster marshes) and regional (data from Thornham and Brancaster marshes combined with those from eleven sampling areas around the UK) predictive foraminifera-based transfer functions, and applied them to a sediment sequence from Holkham North Norfolk, UK. The chronological framework was provided by four radiocarbon dates and five infrared stimulated luminescence ages.

Although there is a general similarity between the local and regional transfer functions, there are significant differences associated with the precision of the reconstructions. It is no surprise that the error ranges of local function are more precise than their regional counterparts as the data used to develop the latter transfer function are derived from thirteen sites with a great variety of tidal, sedimentological and floral characteristics. However, the magnitude of this difference is small when compared to the overall magnitude of the error terms associated with each transfer function and so does not represent a significant improvement in performance. The elevation of the relative sea levels also differed somewhat between the two transfer functions, primarily due to the lack of modern analogues in the local transfer function. We suggest that interpretations based on transfer functions derived from local training sets should be treated with caution unless the constancy of environmental conditions at the sample site can be. Intuitively, these conditions will be most satisfactorily met when dealing with ‘young’ high marsh sediments, but will be subject to increasing conjecture as the age of the sample and its depth relative to the tidal frame increases. The resulting relative sea-level curve from the regional transfer function showed a general trend of rising RSL throughout the Holocene, which is consistent with observations and model predictions from areas close to the limit, and beyond, of the British ice sheet at the Last Glacial Maximum.