The Relationship Between the Global Mean Deep‐Sea and Surface Temperature During the Early Eocene

Abstract

Link to publication General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the "Taverne" license above, please follow below link for the End User Agreement: https://www.ou.nl/taverne-agreement

The Temperature of the Deep Ocean Is a Robust Proxy for Global Mean Surface Temperature During the Cenozoic

“…However, proxy data compilations of the Pliocene and early Palaeogene fall within uncertainty of a 1:1 T d -GMST relationship, suggesting that (some) other process(es) act to balance polar amplification. Using both the DeepMIP set of simulations (Lunt et al, 2021) with varying CO 2 and fixed paleogeography and a set of Cenozoic HadCM3L simulations with covarying paleogeography and CO 2 (Valdes et al, 2021), we show that these processes are: (a) that SST in the regions and season of deep water formation is less sensitive to climate forcing than high latitude mean annual SST (evident in the DeepMIP ensemble but not clearly the case in the HadCM3L V21 simulations), and (b) the fact that the land surface warming is more sensitive than the ocean surface (see also Goudsmit-Harzevoort et al (2023)) to CO 2 and paleogeographic-driven climate change over the Cenozoic. While this provides a mechanistic basis for a 1:1 T d -GMST relationship prior to the (mid)Pliocene, we note that some HadCM3L V21 simulations do not adhere to this, with GMST underestimated by up to 3°C during the Miocene (Figure 4c).…”

“…As shown by Goudsmit-Harzevoort et al (2023), the DeepMIP simulations are characterized by a T d -GMST slope close to 1 based on the entire ensemble (excluding the 9 × CO 2 CESM simulation and PI controls), with m = 1.026 (Figure 4a). Anchoring a 1:1 line to the mean of the lowest CO 2 simulation for each model with Eocene paleogeography (in order to broadly follow the assumption of Hansen et al, 2013) demonstrates that all simulations fall within ±2°C, with the majority falling within ±1°C (mean average error = 0.82°C).…”

“…The results of the first two of these analyses is shown in Figures 6 and 7 for the DeepMIP and HadCM3L V21 simulations respectively, in all cases anchored to (a) the simulation with the lowest pCO 2 but Eocene paleogeography in the case of the DeepMIP simulations, as the closest possible representation of Pliocene-like conditions, and (b) to the Pliocene in the case of HadCM3 V21 , again, because the principal aim of this study is to test the pre-Pliocene T d -GMST assumption of Hansen et al (2013). Note that while a similar interrogation of the relationship between deep ocean temperature and GMST was performed by Goudsmit-Harzevoort et al (2023), the analysis presented here differs in that we interrogate all simulations by determining the relative change from the Pliocene or that with the lowest CO 2 and paleo boundary conditions, although we replot the data in absolute terms for comparison (e.g., Figures 6c and 6f). This is an important distinction, given that it is the relative change in T d and GMST prior to the base of the Pliocene that we are principally interested in, and doing so avoids potential bias derived from model-specific skill in capturing past climate states.…”

“…10.1029/2023PA004788 temperature estimate of Adkins et al (2002). An estimate of the relationship between T d and GMST based on clumped isotope deep ocean temperatures is not given here as a result of the sparsity of data in certain key intervals, but note that a comparison between Δ 47 -derived T d and GMST is given by Goudsmit-Harzevoort et al (2023).…”

“…In the following sections, we: identify the potential issues with this simple transformation of deep ocean temperature (inferred via benthic foraminifera δ 18 O) into GMST in Section 2, critically evaluate the quality of the Cenozoic deep ocean temperature data sets (Section 3.1), and then test whether the methodology can usefully approximate GMST by comparing to a combination of curated proxy data compilations and two sets of climate model simulations, including the DeepMIP model database (Sections 3.2 and 3.3). In comparison to previous work, especially that of Goudsmit-Harzevoort et al (2023) and Valdes et al (2021), our analysis differs in several key respects in that (a) we do not limit our analysis to climate model output, (b) we explicitly set out to test what relationship between GMST and deep ocean temperature is expected, for the entirety of the Cenozoic, and (c) present several advances in the way in which climate model data are evaluated for these purposes, detailed below (Section. 3.3).…”