Carbon cycle instability and orbital forcing during the Middle Eocene Climatic Optimum

The Middle Eocene Climatic Optimum (MECO) is a global warming event that occurred at about 40 Ma. In comparison to the most known global warming events of the Paleogene, the MECO has some peculiar features that make its interpretation controversial. The main peculiarities of the MECO are a duration of ~500 kyr and a carbon isotope signature that varies from site to site. Here we present new carbon and oxygen stable isotopes records (δ13C and δ18O) from three foraminiferal genera dwelling at different depths throughout the water column and the sea bottom during the middle Eocene, from eastern Turkey. We document that the MECO is related to major oceanographic and climatic changes in the Neo-Tethys and also in other oceanic basins. The carbon isotope signature of the MECO is difficult to interpret because it is highly variable from site to site. We hypothesize that such δ13C signature indicates highly unstable oceanographic and carbon cycle conditions, which may have been forced by the coincidence between a 400 kyr and a 2.4 Myr orbital eccentricity minimum. Such forcing has been also suggested for the Cretaceous Oceanic Anoxic Events, which resemble the MECO event more than the Cenozoic hyperthermals.

occurring nanno-, micro-and macrofossils . However, a possible contribution of early burial diagenesis cannot be excluded. In order to exclude variations due to the difference in lithofacies, bulk stable isotopes measurements of the Baskil section were only performed on the hemipelagic marl.
The bulk carbonate δ 13 C curve has highly fluctuating values, with 0.5‰-1.0‰ variations between different lithological beds (Fig. S1). According to the previous considerations, these rapid variations mainly reflect different carbonate sources or possible early burial diagenesis. However, there are more gradual variations that occur consistently throughout different lithological beds. Thus, despite the noise of the rapid fluctuations, it can be recognized an average trend that can be attributed to secular changes in the marine reservoir of dissolved inorganic carbon (DIC). The same interpretation can be applied to the δ 18 O, probably with a more significant contribution of diagenesis in the rapid fluctuations, as oxygen is more sensitive than carbon to early diagenetic processes (e.g. Swart, 2015). After these considerations, and given the paleoceanographic and paleoclimatic target of this work, we chose to rely on the clearer δ 13 C and δ 18 O signals obtained from individual genera of foraminifera.

2) Significance of δ 13 C and δ 18 O data from individual genera of foraminifera
Specimens of foraminifera have been selected from three genera (Acarinina, Subbotina, and Cibicidoides) for stable isotopes analyses. The number of specimens analysed depended on the size of the tests (see methods). These three genera have been selected as representative of three different environments across the water column (surface water, thermocline, and bottom, respectively). Therefore, the isotopic composition of each individual genus depends on the environmental conditions of its specific living environment and, differently from the bulk, is not mixed with carbonate from other sources. However, diagenesis may bias the δ 13 C and δ 18 O compositions in ancient carbonate microfossils (e.g. Edgar et al., 2015). SEM imaging of selected specimens representative of the three studied genera of foraminifera reveal a general very good preservation (Fig. S2). However, all the observed specimens display evidence of recrystallization of the test wall, particularly those of Acarinina that have smaller and more fragile shells (Fig. S2, e-f). Moreover, it seems that significant infill is present. These features indicate that diagenesis may have altered the primary stable isotope signals. Considering the stable isotopes values of the foraminifera from the Baskil section, we observe that the average δ 13 C of Cibicidoides is 0.83‰ lower than Subbotina, which is 0.76‰ lower than Acarinina Acarinina (e, f), from samples T321, T69.5, and T73, of the Baskil section, respectively. probably not always representative of the primary parameters, particularly for δ 18 O of Acarinina, however, the trends can be confidently interpreted.
Another possible bias related to stable isotopes in foraminifera is vital effect. This occurs when, for some reasons, the carbonate of a specific fossil shell does not precipitate in isotopic equilibrium with the environment (i.e. Sharp, 2007). We tried to minimize the vital effect by measuring multiple specimens per sample (3 to 5 for Cibicidoides and Subbotina, 10 to 15 for Acarinina, which is significantly smaller) and, as mentioned before, by basing our interpretation only on curves smoothed by 3 points moving average and on variations of 0.5‰ amplitude or larger.

3) Paleoceanographic and paleoecological changes during the MECO
The stable isotopes shifts that identify the MECO in the Baskil section are accompanied by important changes in calcareous nannofossils assemblages (Fig. S3).
Various groups of calcareous nannofossils increase during or soon after the MECO, whereas others decrease and some disappear (Fig. S3). Despite the paleoecological interpretation of several groups of nannofossils is still controversial, we can infer that Helicosphaera during the MECO can be related to a shift towards warmer and more eutrophic conditions. This agrees with mineralogical and geochemical data from Rego et al. (2018) and shown in the figure 3 of the main article, which display calcite decrease, detrital mineral increase, and relative increase in smectite and decrease in illite accompanying the increase in ∆ 18 O. As discussed in the main text and in , this suggests the onset of warm and humid conditions with prevailing chemical weathering at the MECO, which increased the runoff from the land, freshened the sea surface, and then enhanced the density contrast between the surface and the thermocline.