Bias in foraminiferal multispecies reconstructions of paleohydrographic conditions caused by foraminiferal abundance variations and bioturbational mixing: A model approach

A simple box model was applied to study the combined effect of bioturbation and foraminiferal abundance variations using synthetic isotope records constructed from the Greenland oxygen isotope record (GISP2). Apart from the well known general smoothing of the signal, the attenuation of peak heights, and the up- or downwards shift of the proxy records at intervals of changing abundance, the modeling results draw attention to two effects of special importance to climatic reconstructions based on the comparison of proxy signals from two or more planktonic foraminifer species. First, the bioturbational mixing of tests out of abundance maxima and into over- and underlying sediment can lead to the introduction of stratigraphic shifts between different species when species with opposing abundance patterns are used. This shifting of the signals can easily result in apparent increases or decreases in the isotopic gradients between different species, especially at intervals of rapidly changing climate conditions. Second, in intervals with rapidly changing environmental conditions, such as Dansgaard–Oeschger cycles, where species with different preferences often show opposing abundance patterns, the mixing of foraminifer tests out of their respective abundance maxima can result in a general, but artificial, offset between the two species. This offset then may disappear during intervals of slower variability. These artifacts have the potential of seriously biasing any paleoceanographic and paleoclimatic reconstructions based on multispecies analysis where the studied species show opposing abundance patterns. It is therefore highly recommended that species with as low abundance variations as possible, or at least with similar abundance variations, are used whenever multispecies reconstructions are attempted.