Environmental determinants of radiolarian assemblages in the western Pacific since the last deglaciation

Authors:
Publisher website
Google Scholar
Abstract
Radiolarians are a very diverse microzooplanktonic group, often distributed in regionally restricted assemblages and responding to specific environmental factors. These properties of radiolarian assemblages make the group more conducive for the development and application of basin‐wide ecological models. Here we use a new surface sediment data set from the western Pacific to demonstrate that ecological patterns derived from basin‐wide open‐ocean data sets cannot be transferred on semirestricted marginal seas. The data set consists of 160 surface sediment samples from three tropical‐subtropical regions (East China Sea, South China Sea, and western Pacific), combining 54 new assemblage counts with taxonomically harmonized data from previous studies. Multivariate statistical analyses indicate that winter sea surface temperature at 10 m depth (SSTw) was the most significant environmental variable affecting the composition of radiolarian assemblages, allowing the development of an optimal calibration model (Locally Weighted‐Weighted Averaging regression inverse deshrinking,R2cv = 0.88, root‐mean‐square error of prediction = 1.6°C). The dominant effect of SSTw on radiolarian assemblage composition in the western Pacific is attributed to the East Asian Winter Monsoon (EAWM), which is particularly strong in the marginal seas. To test the applicability of the calibration model on fossil radiolarian assemblages from the marginal seas, the calibration model was applied to two downcore records from the Okinawa Trough, covering the last 18 ka. We observe that these assemblages find most appropriate analogs among modern samples from the marginal basins (East China Sea and South China Sea). Downcore temperature reconstructions at both sites show similarities to known regional SST reconstructions, providing proof of concept for the new radiolarian‐based SSTw calibration model.
Funding Information
  • Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (P300P2_164634)

This publication has 80 references indexed in Scilit: