Ecoregionalisation of the Southern Ocean Using Radiolarians

The Southern Ocean is an important region for identifying ecoregions, both to assist in conservation efforts as well as for paleoecological studies. In this paper we use the Southern Ocean Radiolarian Dataset (SORAD), a compilation of surface sediment radiolarian census data, to determine radiolarian ecoregions throughout the Southern Ocean within the Atlantic, Indian and Southwest Pacific Sectors. The distribution of radiolarian species is explored using the unconstrained non-Metric Multidimensional Scaling to identify gradients in species assemblage variability, as well as distinct and sharp changes in assemblage composition. Ecoregions are assigned to groups of sites using Multivariate Regression Tree analysis, and environmental variables are assessed for their explanatory power. Radiolarian distribution was found to be heterogeneous throughout the Southern Ocean, with each Sector hosting a unique group of radiolarian ecoregions. A total of 6 ecoregions were identified for the Atlantic Sector, 8 for the Indian Sector and 6 for the Southwest Pacific Sector. Ecoregions were generally latitudinally distributed, but also appear to be influenced by significant oceanographic features, such as plateaux and islands, large eddy fields, hydrological fronts and sea ice. This study highlights the need for further exploration of radiolarian ecology, and for caution when using radiolarian assemblage variability in paleo-environmental studies.

Automated summary

The Southern Ocean is an important region for identifying ecoregions, and this study used the Southern Ocean Radiolarian Dataset to determine the radiolarian ecoregions in the Atlantic, Indian, and Southwest Pacific Sectors. The distribution of radiolarian species was found to be heterogeneous throughout the Southern Ocean, with each sector having its own unique group of ecoregions. This study highlights the importance of further exploring radiolarian ecology and being cautious when using radiolarian assemblage variability in paleo-environmental studies.