STABLE ISOTOPIC DERIVED VARIATIONS IN BOTTOM WATER TEMPERATURES OFF VESTFIRDIR, NW ICELAND: AN INDICATOR OF CHANGES IN ATLANTIC WATER OVER THE LAST 6000 CAL YRS

A mixing line for Iceland waters has been developed through an analysis of bottom water d18O SMOW and the water values for d18O SMOW on land. This relationship is used to estimate temperatures from the d18OVPDB determinations on benthic foraminifera from cores collected around Vestfirdir, NW Iceland. Bottom grabs and CTD casts provided data on foraminiferal distributions and their association with environmental gradients. Samples from recent sediments was also used to derive d18OVPDB values for the polar planktonic foraminifera, N. pachyderma s. and two species of benthic foraminifera, namely Cibicides lobatulus (epifaunal species) and Melonis barleeanus (infaunal species). Empirical relationships between water temperature, salinity, and the d18O values indicated that in these waters, most of the variance in the d18O is associated with temperature. Application of the normal equation for the calculation of d18OVPDB for equilibrium calcite to the Siglunes 50-yr hydrographic series, which showed an abrupt change from Atlantic to Arctic waters in 1969 associated with the Great Salinity Anomaly, confirmed the overwhelming importance of temperature in explaining variations in d18OVPDB (r=0.98 for d18O and T°C versus r=0.66 for d18O and S‰). The strong (r=0.78) positive, and potentially offsetting, relationship between temperature and salinity over this interval of time is not sufficient to produce a “flat” d18O series. Application to down core d18O data was made taking into account the “vital effect” and with a selection of salinity values. Bottom water temperature estimates from B997-328,-329, -330, and -341 indicate that the coldest interval in the last 5 cal ka BP was during the 18-16thC when bottom waters are estimated to have decreased to between 3 and 4 °C compared to between 5.5 and 7°C during the maximum of the last 5 cal ka B.P. which occurred between 1.5 and 1 cal ka B.P. An attempt is made to ascribe error limits to the temperature estimates.