Astronomical solutions form the backbone of accurate dating for geology and paleoclimate studies. Beyond ∼50 Ma, however, the chaos inherent in the solar system makes it impossible to calculate a single unique astronomical solution. Geological data have been used to constrain this chaos in order to arrive at a geologic time scale up to the end-Cretaceous. Here, we adopt and extend this approach into the latest Cretaceous, by re-analyzing the Zumaia and Sopelana composite proxy records from the Maastrichtian. We find that the filtered sum total light reflectance (L*) record is most compatible with the astronomical solution ZB20a. However, these results are sensitive to parameter choices in our algorithm, which we describe in detail. Nevertheless, we present evidence in favor of solution ZB20a for cyclostratigraphy during the latest Cretaceous. Periods with very long eccentricity nodes (VLNs) (low amplitude in the short eccentricity cycle) in the astronomical solutions that coincide with large amplitudes in the short eccentricity-related peaks in the filtered sum proxy record rule out alternatives.

We use global coupled atmosphere-ocean-biogeochemistry models from the Coupled Model Intercomparison Project (CMIP5), under the RCP8.5 scenario, to show that the global interannual variability of the sea surface pCO (calculated as 1σ) could increase by 62 ± 22 % by 2090. This amplification is a consequence of a larger background pCO and a lower buffering capacity that enhance the response of pCO to surface temperature (T) and dissolved inorganic carbon (DIC) changes. The amplification is counteracted by a decrease in the sea-surface DIC interannual variability, which will likely cause a strong reduction on the pCO’s variability in the equatorial Pacific. The potential changes in seawater carbonate chemistry are simulated with higher consistency than those in the DIC and T anomalies driven by ocean circulation and biology. The changes in sea-surface pCO interannual variability are reflected in the ocean-atmosphere flux of CO and need to be accounted for future carbon projections.