Details
Zusammenfassung: <jats:p>A new diatom-based sea surface salinity (SSS) estimation has been applied to a collection of 27 taxa in 48 present-day sediment and surface water samples recovered in the Baltic Sea and Kattegat. The sediment core 303610-12 (2005) from the Eastern Gotland was chosen for study of Holocene sequence, ranging the last 8160 yrs BP. The Artificial Neuronal Network (ANN) method allows the estimation of spring SSS (March-April) values ranging between 7.04 ‰ and 8.25 ‰ at an averaged Root Mean Squared Error (RMSE) of 0.49 ‰. The rather low amplitude of salinity change might be caused by mixing of fresh water with upper surface layer of the Baltic Sea due to high precipitation and riverine input. The estimates of spring SSS from core 303610-12 were compared with independent geochemical proxies for salinity (K, Ti and S) derived from XRF Core Scanner record. Conspicuous correlation between salinity and sulphur records and reverse-correlation to K and Ti demonstrates that the ANN method combined with quantitative and qualitative analyses of diatoms provides a useful tool for palaeosalinity reconstructions in the Holocene sediments of the Baltic Sea</jats:p>
Umfang: 131-140
ISSN: 0067-3064
DOI: 10.5200/baltica.2014.27.22