Coccolith clumped isotopes suggest a more modest Miocene North Atlantic polar amplification

Abstract:

Current climate models often struggle to reproduce the high latitude warmth that widely-used proxies suggest during warm climates. This model-data mismatch, in which proxies imply an increased polar amplification and extreme fattening of the latitudinal thermal gradient, remains one of our biggest challenges as a paleoclimate community.

Here we produce the first clumped isotope (∆₄₇) temperature record of the North Atlantic over the last 16 Ma from >91% pure fractions of well-preserved coccoliths (ODP Site 982). We prove the reliability of this proxy to reproduce production temperatures by comparing ∆₄₇ temperatures from a monospecific sediment trap sample, and well-constrained satellite temperatures of this Coccolithus pelagicus bloom.

An analysis of production dynamics in the area shows that most production occurs during winter-spring, and that there is a larger than acknowledged deeper, light-limited production. Therefore, for our area, coccolith ∆₄₇ likely represents mixed-layer winter-spring temperatures.

Although coccolith ∆₄₇ and alkenone (UK'37) cooling trends agree well, coccolith ∆₄₇ temperatures are substantially colder. Here we suggest that alkenone production may be affected by non-thermal mechanisms, such as light limitation, wherewith the application of available calibrations to UK'37 could lead to SST overestimations in the North Atlantic. Moreover, our UK'37 SSTs suggest a mid-Miocene to mid-Holocene cooling 1.7 times larger than that shown by coccolith ∆₄₇ temperatures. The lower coccolith ∆₄₇-derived cooling agrees better with that suggested by a recent Earth System model of mid-Miocene climate (Crichton et al.., preprint, Climate of the past) for our study site. The extreme polar amplification and negligible thermal gradient during the mid-Miocene suggested by alkenones contrasts with a more modest polar amplification and less extreme latitudinal thermal fattening (>10 °C latitudinal thermal differences) implied both by ODP Site 982 coccolith ∆₄₇ and by the model.

A continuous pre mid-Miocene coccolith ∆₄₇ record could help to better understand polar amplification during warm climates. Colder than previously thought pre mid-Miocene high latitude temperatures would imply a more optimistic perspective of high latitude response to anthropogenic CO₂.