Mejía, Luz María; Mendez-Vicente, Ana; Abrevaya, Lorena; Lawrence, Kira T; Ladlow, Caroline; Bolton, Clara T; Cacho, Isabel; Stoll, Heather M (2017): A diatom record of CO2 decline since the late Miocene [dataset publication series]. PANGAEA, https://doi.org/10.1594/PANGAEA.881946,

---

Supplement to: Mejía, LM et al. (2017): A diatom record of CO 2 decline since the late Miocene. Earth and Planetary Science Letters, 479, 18-33, https://doi.org/10.1016/j.epsl.2017.08.034

Extratropical sea surface temperature records from alkenones record a dramatic cooling of up to 17°C over the last ~14 Ma, but the relationship between this cooling and greenhouse gas forcing has been elusive due to sparse and contrasting reconstructions of atmospheric CO2 for the time period. Alkenone carbon isotopic fractionation during photosynthesis has previously been used to estimate changes in pCO2 over this interval, but is complicated by significant changes in cell size of the alkenone-producing coccolithophorids over this time period. In this study, we reconstruct carbon isotopic fractionation during photosynthesis (epsilon p) using organic compounds trapped within the frustules of pennate diatoms in sediments from the Eastern Equatorial Pacific Ocean at Ocean Drilling Program Site 846 over the last ~13 Ma. Physical separation of pennate diatoms prior to measuring carbon isotopic fractionation enables us to obtain a record with constant cell geometry, eliminating this factor of uncertainty in our pCO2 reconstruction. In the past ~11Ma, epsilon p declines from 15.5 to 10.3 per mil. Using the classic diffusive model and taking into account variations in opal content, alkenone concentration and coccolith Sr/Ca as indicators of past productivity and growth rate, and sea surface temperature records from the site, we estimate a decline in pCO2 from 454 (+/-41) to 250 (+/-15) ppmv between ~11 and 6 Ma. Models accounting for changing the significance of active carbon uptake for photosynthesis, which likely produce more accurate CO2 estimates, suggest a significant larger pCO2 decline of up to twice that shown by the classic diffusive model (in average from 794 (+/-233) ppmv at ~11 Ma to 288 (+/-25) ppmv at ~6 Ma, considering growth rates varying between 0.5 and 1.7 day-1). Large uncertainties in the pCO2 estimated between ~8 and 11 Ma using the active uptake model are related to the growth rate used for calculations. Together, these results suggest CO2 forcing for this period of steep decline in temperatures.