[M-IS12] 古気候・古海洋変動

[MIS12-P25] Making sense of the tropical surface ocean “cold bias” in Mg/Ca temperature records of the Late Pliocene

*Elena Dominguez-Valdes¹、Cyrus Karas²、Jeroen Groeneveld¹、Dirk Nurnberg³、Heather Ford⁴、Erin McClymont⁵、Sze Ling Ho¹ (1.Institute of Oceanography, National Taiwan University, Taipei, Taiwan、2.Universidad de Santiago de Chile, Santiago, Chile、3.GEOMAR Helmholtz Centre for Ocean Research, Kiel, Germany、4.School of Geography, Queen Mary University of London, London, UK、5.Department of Geography, Durham University, Durham, UK)

キーワード：Pliocene, KM5c interglacial, Mg/Ca paleothermometer, Planktonic foraminifera, Sea Surface Temperatures, Carbonate dissolution

The future evoluton of tropical sea surface temperature (SST) patterns remains highly uncertain despite its importance to climate sensitivity (Rugenstein et al., 2023). The Late Pliocene KM5c interglacial (3.205 ± 0.01 Ma) offers an opportunity to investigate the equilibrated tropical response to higher-than-preindustrial CO₂ forcing in the context of continental and orbital configurations that resemble those of today. However, recent compilations of KM5c surface ocean temperature estimates by the PlioVAR PAGES working group have revealed the existence of conflicting signals between different proxies which crucially hamper data-model intercomparisons. In particular, available Mg/Ca data from a number of low latitude sites showed colder SSTs than pre-industrial (PI) (McClymont et al., 2020), in stark disagreement with alkenone-derived temperature records that indicate relalively uniform surface warming with respect to PI. This would imply a larger degree of tropical sea surface heterogeneity than predicted by model experiments which commonly consider the upper end of Pliocene CO₂ estimates (i.e. 400 ppm, de la Vega et al., 2020). Here we revisit the underlying assumptions behind estimations of PI-Pliocene SST anomalies from Mg/Ca using existing surface-dwelling planktonic foraminifera records, as well as new G. ruber data of the Indian and Pacific oceans spanning the KM5c. We evaluate downcore Mg/Ca dissolution corrections and calibration choices using a compilation of nearby coretops (Tierney et al., 2019) and inferred site-specific calcification depths. Together with presumed moderate changes in the Mg/Ca concentration of seawater since the Pliocene, as suggested by clumped isotope thermometry (Meinicke et al., 2021), our preliminary results point towards the existence of small positive PI-anomalies during the KM5c. We find that the culprits behind the so-called Mg/Ca “cold bias” were a combination of (1) homologous treatment of T. sacculifer and G. ruber Mg/Ca data despite differences in their habitat depth preferences for certain hydrographic settings, (2) the use of SST-PI data products to calculate PI-anomalies, and (3) the adoption of Mg/Ca dissolution corrections that in some cases would yield unrealistic calcification depths when applied to nearby coretops. Our re-assessment of Mg/Ca tropical PI-KM5c anomalies is compatible with alkenone-based SST records calibrated within the 28ºC proxy limit for sites where both temperature indicators have been measured. Moreover, it reduces the gap between data- and PlioMIP-derived zonal SST gradients of the Pacific. Nonetheless, interpretation of carbonate-based proxies would greatly benefit from a comprehensive understanding of dissolution effects which is currently lacking.