SALT MARSHES IN WEST GREENLAND: NEW PALAEOENVIRONMENTAL ARCHIVES OF HIGH LATITUDE SEA LEVEL CHANGE

Temperate salt marshes are a proven environment for developing high precision (~ ± 0.10 to 0.20 m) relative sea level (RSL) records over the last millennium. However high latitude salt marshes in Greenland have not been explored as archives of RSL changes. Issues which face those developing and applying transfer functions in these environments include: (1) to use a modern training set of microfossil counts from marshes local to the fossil core, or incorporate a wider geographical area to better capture variability in microfossil distributions, (2) how the structure of the modern training set may affect the apparent performance of the transfer function. We develop diatom-based transfer functions using data from salt marshes from two sites, 250 km apart in West Greenland to explore these issues.

There are variations in diatom distributions between the two sites (with different tidal ranges) when standardised using the Standardised Water Level Index (SWLI). Therefore a local transfer function model must be superior to a combined model in this area, because of the different elevation ranges (in relation to tidal flooding) inhabited by the same diatom species at sites with different tidal ranges. Using a single site transfer function, RSL reconstructions are affected by species whose optima are poorly predicted by the regression model (Weighted Averaging Partial Least Squares), and have wide elevation tolerances and no clear optima in the modern elevation range. We therefore develop an alternative method using Visual Interpretation (VI) of modern and fossil diatom, %LOI and particle size data, using changes in multiple species over small height ranges in the upper marsh, taking into account both presence and absence and vertical trends in diatom abundances. The results from both approaches are in broad agreement, although transfer function reconstruction errors are generally larger than using the VI method.

This study shows that (1) high latitude salt marshes can be used for relative sea-level reconstructions using a transfer function, (2) local training sets are needed for accurate reconstructions and standardising between sites with different tidal ranges needs to be treated with care, and (3) the VI method provides an alternative, complementary method to reconstruct RSL changes which incorporates other available stratigraphic data.