Abstract

The variability of diatom species composition in lake surface sediments was studied along transects in four lakes in northeastern Germany. Three dimictic lakes (Dudinghausener See, Tiefer See, and Cambser See) and one shallow lake (Gro beta Peetscher See) were sampled. Large differences in diatom composition were found between adjoined samples from different depths within one lake. These differences were mainly displayed by planktonic species. For example, the relative frequency of Stephanodiscus alpinus varied between 4% and 43% within the surface sediment samples of the open-water region of Dudinghausener See. Using transfer functions for total phosphorus (TP) based on the European Diatom data-base (EDDI) combined TP data-set and a local data-set, the inferred TP values differed strongly within one lake when using Weighted Averaging-Partial Least Squares (WA-PLS) regression. In Tiefer See (average of measured TP: 30 mu g l(-1)), the inferred TP values range from 45 to 110 mu g l(-1) using the transfer function based on WA-PLS regression and the EDDI data-set; and from 16 to 100 mu g l(-1) using WA-PLS and a local data-set. Performing Maximum Likelihood (ML) regression reduced the difference between measured and inferred values. For Tiefer See, the inferred TP values range between 16 and 45 mu g l(-1) using ML regression and the local data-set. Therefore, it seems that ML regression can deal better with the natural variability in species composition than WA-PLS regression. In general, it was shown that by using ML regression and the local data-set, the error of the inferred values was significant lower for all lakes than using WA-PLS regression and the EDDI data-set. The Root Mean Square Error of Prediction (RMSEP) was not useful in selecting the most stable transfer function.

Keywords

diatoms; total phosphorus; transfer function; WA-PLS regression; ML regression; transect analysis; EDDI data-set; RMSEP

Published in

Journal of Paleolimnology
2007, Volume: 37, number: 4, pages: 573-590
Publisher: SPRINGER

UKÄ Subject classification

Ecology


Publication identifier

DOI: https://doi.org/10.1007/s10933-006-9054-1

Permanent link to this page (URI)

https://res.slu.se/id/publ/51567