ALBERT

Description: Composite core images for Ocean Drilling Program (ODP) Hole 1088B were generated by cutting individual core sections from core table photos (downloaded from JANUS: http://www-odp.tamu.edu/database/). Each section image was then compiled and scaled to the shipboard meters composite depth (mcd) using the Includes_Core_Table_Photos functions within Code for Ocean Drilling Data (CODD v1; www.codd-home.net; Wilkens et al. (2017)). To account for the single point lighting source used in core table photos, a lighting correction was also applied. As shipboard color reflectance data was not collected at Site 1088, core color information (hue, saturation, lightness, red, green, blue) was extracted every mm from the composite core images using the CODD “CreateHoleCoreImageProfiles” function. The extracted lightness (L) showed the clearest variability and was chosen to approximate Lstar data. The top 4 cm of each section was excluded from the L profile to remove a lighting artefact at section ends. To account for core heterogeneity, the L data was then despiked using the CODD “Median_DeSpike” function (despiked data over 2 s.d. from the median within a 400 point/0.4 m window) and then smoothed using an 11-point binomial smoothing. All extracted data is provided in supplementary table 1. The composite core images from Hole 1088B show clear evidence for core disturbance at the top of most cores (supplementary table 2). These intervals were excluded from the extracted lightness and shipboard magnetic susceptibility records. The shipboard magnetic susceptibility (MS) data was also despiked using the “Median_DeSpike” function (despiked data over 2 s.d. from the median within a 40 point/0.4 m window; supplementary tables 3). An astrochronology for Hole 1088B was generated between 4.2 and 7.9Ma (38-90 mcd), using 24 minimal tie points between L maxima and ~110 kyr eccentricity maxima (Ecc; Laskar et al. (2004). These additional age points are provided in supplementary table 4.

Description: This table T6 contains pCO2 and algal growth rate simulations from Tanner et al. 2020 with their reported 1 simga confidence interval based on a full Monte Carlo Simulation (n=10'000). The calculations of simulation 1-7 are based on a multilinear regression model (Stoll et al. 2019) that used the isotopic fractionation (εp) during photosynthetic fixation of carbon, algal growth rate (μ), algal cell radius and the photosynthetic available radiation (PAR) to estimate past CO2. Proxy data are from the Ocean Drilling Program (ODP) core 1088, spanning the time interval of the late Miocene. Simulation 8 follows the pCO2 calculations based on alkenones using the traditional, diffusive “b”-term approach (e.g. Zhang et al. 2013). Simulation 9 is a slight variation of this diffusive approach with coccolith size as a proxy for growth rate, according to Zhang et al. 2020. The dataset provides a high-resolution trend over the late Miocene Cooling ranging from ~8 to 4.3 Ma. The different CO2-trends in this dataset are the main takeaway from the publication (Tanner et al. 2020) and help the reader understand the different thoughts and approaches that were taken in computing the published trend (based on simulation 1 and 6).