Research articleA new radiolarian transfer function for the Pacific Ocean and application to fossil records: Assessing potential and limitations for the last glacial-interglacial cycle
Introduction
The Pacific Ocean is the largest oceanographic basin and plays an important role in modulating global climate at multiple time scales, from seasonal to orbital (Cane, 1998; Gray et al., 2018; Newman et al., 2003; Winckler et al., 2016). Sediment history of this region provides strong constraints on the response of the climate system to temperature, atmospheric CO2 or ice-volume changes. Calibration of sedimentary proxies from marine records is important in order to provide adequate quantification of the natural climate change variability and sensitivity to external and internal forcing mechanisms. In particular, sea surface temperature (SST) is one of the most important indicators of change in the climate system. Various paleotemperature proxies have been calibrated in different parts of the Pacific Ocean, including alkenones (Prahl et al., 1995), Mg/Ca in planktonic foraminifera (Lea et al., 2000) or transfer functions based on faunal and floral census counts (Cortese and Prebble, 2015; Feldberg and Mix, 2002; Hernández-Almeida et al., 2017; Kucera et al., 2005b; Pisias and Mix, 1997; Saavedra-Pellitero et al., 2013). However, when some of these calibrations are applied on the same downcore marine records, some significant discrepancies have been observed (Bard, 2001; Kienast et al., 2012; Koutavas and Sachs, 2008; Leduc et al., 2010; Steinke et al., 2008). Although some problems are inherent to the type of proxies (e.g. organic matter preservation, transport, diagenesis, no-analog conditions) calcium carbonate proxies are particularly prone to dissolution in some regions of the Pacific Ocean (Lyle et al., 2002). Depth of the lysocline is relatively shallower today (3000 m) compared to the Atlantic Ocean, and it has been suggested that glacial lysocline was 600 m deeper (Farrell and Prell, 1989). This in fact has been shown to affect the integrity of calcium carbonate proxies in different regions of the Pacific Ocean (Le and Shackleton, 1992; Lea et al., 2006; Mekik et al., 2007). Organic proxies for SST also show significant overestimate of SSTs at many sites in the Pacific basin compared to faunal or floral based estimates (Bard, 2001). Kienast et al. (2012) observed a warm seasonal bias of alkenone-derived SST estimates in the near-coastal Peru upwelling zone that underestimated the amplitude of the glacial cooling, due to a highly dynamic environment or bias towards El Nino events.
One of the most promising proxies for reconstructing quantitatively SST is the composition of microfossil assemblages in marine sediments (Pisias et al., 1997). Compared to proxies based on carbonate microfossil remains, biogenic silica proxies are often abundant and well preserved in many parts of the Pacific Ocean, in particular in some high latitude regions. Although biogenic silica can be also affected by dissolution within the water column and at the water-sediment interface, biogenic silica is abundant in many regions of the Pacific Ocean with poor calcium carbonate preservation, such as the Eastern Equatorial Pacific (Haslett, 1992; Pisias and Mix, 1997), Arctic Ocean (Itaki et al., 2003), Bering Sea (Itaki et al., 2012; Okazaki et al., 2005), Sea of Okhotsk (Itaki et al., 2008; Okazaki et al., 2003) and Southern Ocean (Crosta et al., 2005; Gersonde et al., 2005). Therefore, investigation of biogenic silica microfossils in the sediments of the Pacific Ocean can potentially lead to the development of robust methods to quantitatively reconstruct past changes in oceanographic and climatic conditions. Thus, our ability to understand past and to predict future climatic change in this region is linked to our ability to make proper use of proxies with biogenic silica composition.
Radiolarians are siliceous micro-zooplankton that dwell mainly in pelagic and hemipelagic oceans, are highly diverse and have wide vertical distributions. Radiolarian assemblages reflect ocean conditions, making them potential carriers of information on the state of past ocean conditions (Pisias et al., 1997). Studies on distribution of radiolarians from surface sediments and in the water column have shown that many species follow large-scale patterns of productivity in the Pacific Ocean (Molina-Cruz, 1984). Studying the biogeographic distribution of modern radiolarian assemblages at basin-scale and their relationship to environmental variables is thus of fundamental importance to improve the accuracy of paleoceanographic reconstructions based on the relative abundance of their species fossil remains.
Basin-wide datasets have the benefit of providing a standardized product that can be used by other researchers, and allows developing sea surface temperatures (SST) reconstructions in multiple locations using the same proxy. Existing distributional data collected for radiolarians from surface sediments from the Pacific Ocean by Pisias et al. (1997) (hereinafter P97) (n = 170) provided first insights into the biogeography of radiolarians in the Pacific Ocean. Although this study was a major achievement that allowed developing routinely quantitative reconstructions based on radiolarians, gaps in sample distribution limited its application in some situations. For example, temperature reconstructions using the P97 dataset in Western Pacific marginal seas showed little temperature variability compared to studies including local samples that also provided better modern analogs (Hernández-Almeida et al., 2017). Moreover, P97 analyses lack a statistically solid evaluation of other factors than temperature affecting species distribution regionally and calibrations are uniquely based on the Imbrie and Kipp (I&K) transfer function method (Imbrie and Kipp, 1971). Although this regression technique may produce realistic reconstructions, it is important to test alternative methods which have been demonstrated to produce more reliable results, such as the Modern Analogue Technique (MAT) (Hutson, 1977) or Weighted Average Partial Least Squares (WA-PLS) (ter Braak and Juggins, 1993).
Calibration of radiolarian assemblages to SST on a basin-wide scale requires accurate knowledge of radiolarian taxonomy and biogeography. Consequently, the principal aim of this study is to develop a new radiolarian dataset for the Pacific Ocean, based on published literature and taxonomically homogenized census datasets. This new calibration builds-up on P97, but also extends over regions in the Pacific Ocean under-represented in P97 (e.g. Sea of Okhotsk, East and South China Sea, Southwest Pacific), and thus covering multiple environmental ocean gradients. This new dataset offers a unique opportunity to fully investigate radiolarian biogeography in the Pacific Ocean and the environmental parameters behind it using a robust and thorough statistical approach. Finally, the derived radiolarian-based calibration model will be applied to radiolarian fossil assemblages from various locations in the Pacific Ocean in order to reconstruct quantitatively their environmental variability over last 165 ka.
Section snippets
Modern radiolarian datasets
The relative abundance of all species (including unidentified forms), recorded as a percentage of the total number of radiolarian shells counted, was compiled from 27 published sources, as well as unpublished datasets (Kruglikova, unpublished). These new radiolarian datasets were prepared using standard preparation techniques. Samples were dried and washed through a 45 μm mesh. The residue was pipeted and mounted on a glass slide using Canada balsam, and the radiolarians on the slide were
Multivariate statistics
Multivariate ordination analyses are shown in Fig. 3, Fig. 4. The two first PCA axes were significant according to the broken stick model and explained 69.7% of the variance in the environmental data. The PCA1 axis explained 48.2% of the variance, and it is negatively correlated to SST10, SST200, Sal10 and Sal200, and positively correlated with Phos200, Nit200, Sil10, AOU200 and Sil200 (Fig. 3). The PCA2 axis explains 19.1% of the variance, and it is positively correlated to Npp, and negatively
Transfer function and importance of SST for radiolarians
Ocean temperature in the surface ocean down to the thermocline is the most important variable not only for determining the optimal conditions at which organisms can dwell and reproduce, but it is also one of the two most important variables (together with salinity) controlling the vertical structure of the water column, which in turn plays an important role for primary productivity in the ocean (Sarmiento et al., 2004). Radiolarians are zooplankton organisms, and consequently they mainly depend
Conclusion
- •
In this study, we use a new 801 samples modern radiolarian dataset for the Pacific Ocean and apply it to 31 published radiolarian records in this basin, harmonizing core-top and downcore radiolarian datasets. Multivariate statistical analyses show that SST10 is the most important variable controlling the radiolarian distribution in surface sediments in the Pacific Ocean, followed by SST200.
- •
All regression methods showed similarly good statistical performance, and at some extent, they were
Authors contribution
IHA collected the data with the help of DB; SBK provided unpublished data. IHA analyzed the data and wrote the manuscript, with contributions of all co-authors.
Data availability
Data is available as supplementary material.
Declaration of Competing Interest
None.
Acknowledgement
The authors acknowledge all scientists who made available published radiolarian datasets in public data repositories or sent them to the authors, in particular to F. Chang, K. Matsuzaki, R. Wang, A. Matul, L. Zhang, K. Matsuzaki, T. Itaki, L. Liu, Y. Okazaki, K. Ikehara and N.G. Pisias. We also thank H. Stoll for discussion and advice regarding this study. GC was supported by the New Zealand Government through the GNS Science Global Change through Time Program. We thank two reviewers for their
References (132)
- et al.
Reproducibility of species recognition in modern planktonic foraminifera and its implications for analyses of community structure
J. Micropalaeontol.
(2018) Comparison of alkenone estimates with other paleotemperature proxies
Geochem. Geophys. Geosyst.
(2001)- et al.
Photosynthetic rates derived from satellite-based chlorophyll concentration
Limnol. Oceanogr.
(1997) Quantitative palaeoenvironmental reconstructions
Numerical tools in palaeolimnology-progress, potentialities, and problems
J. Paleolimnol.
(1998)- et al.
Diatoms and pH Reconstruction. Philosophical transactions of the Royal Society of London. B
Biol. Sci.
(1990) - et al.
Strengths and weaknesses of quantitative climate reconstructions based on Late-quaternary
Open Ecol. J.
(2010) - et al.
Polycystine radiolarians in surface sediments in the Arctic Ocean basins and marginal seas
Mar. Micropaleontol.
(2003) - et al.
Amphimelissa setosa (Cleve)(Polycystina, Nassellaria)–a stratigraphic and paleoecological marker of migrating polar environments in the northern hemisphere during the quaternary
Stratigraphy
(2015)
Vertical distribution patterns of Radiolaria Polycystina (Protista) in the World Ocean: living ranges, isothermal submersion and settling shells
J. Plankton Res.
Biogeography of Radiolaria Polycystina (Protista) in the World Ocean
Prog. Oceanogr.
World atlas of distribution of recent Polycystina (Radiolaria)
Palaeontol. Electron.
Last Glacial Maximum to Holocene sea surface conditions at Umnak Plateau, Bering Sea, as inferred from diatom, alkenone, and stable isotope records
Paleoceanography
A role for the tropical Pacific
Radiolarian fauna in surface sediments of the northeastern East China Sea
Mar. Micropaleontol.
A Holocene paleotemperature record based on radiolaria from the northern Okinawa Trough (East China Sea)
Quat. Int.
R: A Language and Environment for Statistical Computing
A radiolarian-based modern analogue dataset for palaeoenvironmental reconstructions in the Southwest Pacific
Mar. Micropaleontol.
Late Pleistocene–Holocene radiolarian paleotemperatures in the Norwegian Sea based on artificial neural networks
Palaeogeogr. Palaeoclimatol. Palaeoecol.
The biogeography of major diatom taxa in Southern Ocean sediments: 2. Open ocean related species
Palaeogeogr. Palaeoclimatol. Palaeoecol.
Climatic change and CaCO3 preservation: an 800,000 year bathymetric reconstruction from the central equatorial Pacific Ocean
Paleoceanography
Sea-surface temperature estimates in the Southeast Pacific based on planktonic foraminiferal species; modern calibration and last Glacial Maximum
Mar. Micropaleontol.
World Ocean Atlas 2013
Sea-surface temperature and sea ice distribution of the Southern Ocean at the EPILOG last Glacial Maximum—a circum-Antarctic view based on siliceous microfossil records
Quat. Sci. Rev.
Deglacial upwelling, productivity and CO2 outgassing in the North Pacific Ocean
Nat. Geosci.
QSR Correspondence “is spatial autocorrelation introducing biases in the apparent accuracy of palaeoclimatic reconstructions?” Reply to Telford and Birks
Quat. Sci. Rev.
PAST: Paleontological statistics software package for education and data analysis
Characteristics of alkenones synthesized by a bloom of Emiliania huxleyi in the Bering Sea
Geochim. Cosmochim. Acta
Rapid fluctuation of alkenone temperature in the southwestern Okhotsk Sea during the past 120 ky
Glob. Planet. Chang.
Freshwater impacts recorded in tetraunsaturated alkenones and alkenone sea surface temperatures from the Okhotsk Sea across millennial-scale cycles
Paleoceanography
Early Pleistocene glacial-interglacial radiolarian assemblages from the eastern equatorial Pacific
J. Plankton Res.
The Sea of Okhotsk: a window on the ice age ocean
Deep-Sea Res. I Oceanogr. Res. Pap.
A reduced estimate of the strength of the ocean's biological carbon pump
Geophys. Res. Lett.
Tropical ocean temperatures over the past 3.5 million years
Science
Environmental determinants of radiolarian assemblages in the western Pacific since the last deglaciation
Paleoceanography
Detrended correspondence analysis: an improved ordination technique
Vegetatio
Transfer functions under no-analog conditions: experiments with Indian Ocean planktonic foraminifera
Quat. Res.
Paleoenvironmental changes in the northern area of the East China Sea during the past 42,000 years
Palaeogeogr. Palaeoclimatol. Palaeoecol.
A new micropaleontological method for quantitative paleoclimatology: application to a late Pleistocene Caribbean core
Alkenone-sea surface temperatures in the Japan Sea over the past 36 kyr: warm temperatures at the last glacial maximum
Org. Geochem.
Depth-related radiolarian assemblage in the water-column and surface sediments of the Japan Sea
Mar. Micropaleontol.
Depth distribution of radiolarians from the Chukchi and Beaufort Seas, western Arctic
Deep-Sea Res. I Oceanogr. Res. Pap.
Last glacial–Holocene water structure in the southwestern Okhotsk Sea inferred from radiolarian assemblages
Mar. Micropaleontol.
Millennial-scale variations of late Pleistocene radiolarian assemblages in the Bering Sea related to environments in shallow and deep waters
Deep-Sea Res. II
Stopping rules in principal components analysis: a comparison of heuristical and statistical approaches
Ecology
Rioja: Analysis of Quaternary Science Data. R Package Version 0.5–6
Quantitative reconstructions in palaeolimnology: new paradigm or sick science?
Quat. Sci. Rev.
Quantitative environmental reconstructions from biological data, Tracking environmental change using lake sediments
Changes in long chain alkenone distributions and Isochrysidales groups along the Baltic Sea salinity gradient
Org. Geochem.
Cited by (12)
New Southern Ocean transfer function for subsurface temperature prediction using radiolarian assemblages
2023, Marine MicropaleontologyCitation Excerpt :While IKM can extrapolate no analogue situations and therefore account for them, MAT interpolates predictions from the information contained in the modern analogs exclusively present in the reference dataset, requiring a training dataset including as many stations as possible. The MAT variation, WMAT, was added here based on the results of Hernández-Almeida et al. (2020). The authors showed that WMAT performed best with their Pacific dataset when compared to the Weighted-Averaged Partial Least Squares method (WAPLS) from Ter Braak and Juggins (1993).
Southern Ocean sea surface temperature synthesis: Part 1. Evaluation of temperature proxies at glacial-interglacial time scales
2021, Quaternary Science ReviewsAntarctic Polar Front migrations in the Kerguelen Plateau region, Southern Ocean, over the past 360 kyrs
2021, Global and Planetary ChangeRadiolarian biogeographic contrast between spring of 2017 and winter of 2017–2018 in the South China sea and Malacca Strait
2020, Continental Shelf ResearchLate quaternary sea-ice and sedimentary redox conditions in the eastern Bering Sea – Implications for ventilation of the mid-depth North Pacific and an Atlantic-Pacific seesaw mechanism
2020, Quaternary Science ReviewsCitation Excerpt :From ∼15 ka, however, sea-ice biomarker records demonstrate a consistent decrease in the Bering Sea and the subarctic North Pacific sea-ice cover (Fig. 5) (Méheust et al., 2018, 2016); while the record from the central Sea of Okhotsk indicates a transition towards marginal sea-ice conditions (Lo et al., 2018). This is in line with substantial hemisphere-wide atmospheric (Shakun et al., 2012) and region-wide SST warming at the onset of the BA (Caissie et al., 2010; Hernández-Almeida et al., 2020; Max et al., 2012; Méheust et al., 2018; Meyer et al., 2016; Praetorius et al., 2020; Riethdorf et al., 2013) (Supplementary Figure 8). Consistent with warming during the BA, sea-ice biomarker records at Site U1343 suggest predominantly ice-free conditions, with a renewed increase in MIZ sedimentation during the late BA, as indicated by increasing HBI III concentrations (Fig. 2).