HIGH RESOLUTION SEA-LEVEL RECONSTRUCTION AT SAND HILL POINT, NORTH CAROLINA: A COMPARISON OF TWO APPROACHES IN A MICROTIDAL SETTING

Coastal North Carolina (USA) is particularly vulnerable to future sea-level rise. Salt marshes are at risk due to the low-lying coastal plain and low gradient. Sea-level reconstructions can provide a clearer context for changes in sea-level trends and insight into the likely effects of future changes. The expansive salt marshes of North Carolina provide an excellent environment from which to produce relative sea-level reconstructions based on salt marsh foraminifera, the distribution of which is controlled by tidal elevation. We assess the advantages and disadvantages of using a transfer function approach in this microtidal setting. The microtidal regime of northeastern North Carolina is unusual, the elevation between mean lower low water (MLLW) and mean higher high water (MHHW) is often less than 0.3m. The recently developed regression model, locally weighted-weighted average (LWWA), was used to reconstruct paleomarsh elevation and to assess the effectiveness of this transfer function method (RMSEP = 74.7; average error = ± 0.11m) as compared to a non-quantitative approach that recognizes, a priori, a division into low marsh and high marsh. Foraminiferal assemblages of surface samples (n = 23) from two transects at Sand Hill Point (Cedar Island, North Carolina) were added to an existing modern training set (n = 205) of paired observations of tidal elevation and foraminiferal assemblages and provided new local analogues. The foraminiferal assemblages preserved in a radiocarbon-dated core of salt marsh peat was utilized to produce a continuous reconstruction of relative sea-level during the late Holocene using the two separate and distinct methodologies. This comparison allows us to assess the advantages and disadvantages of using transfer functions versus traditional methods in restricted environments with microtidal ranges.