Multiple Proxy Estimates of Atmospheric CO₂ From an Early Paleocene Rainforest

Abstract

Proxy estimates of atmospheric CO2 are necessary to reconstruct Earth's climate history. Confidence in paleo‐CO2 estimates can be increased by comparing results from multiple proxies at a single site, but so far this strategy has been implemented only for marine‐based techniques. Here we present CO2 estimates for the well‐studied early Paleocene Castle Rock site in Colorado using four paleobotanical proxies. Median estimates range from 470 to 813 ppm, demonstrating fair correspondence. The synthesis yields a median of 616 ppm (352–1110 ppm at 95% confidence), considerably higher than previous early Paleocene CO2 estimates (~300 ppm). Ash bed geochronology by the high‐precision U‐Pb method places the Castle Rock assemblage at 63.844 ± 0.097 Ma (fully propagated 2σ error). When these results are placed into the broader context of other Cenozoic CO2 estimates from plant‐gas‐exchange approaches and coeval estimates of global mean surface temperature, a pattern emerges of an Earth system sensitivity around 3 °C per CO2 doubling during the Paleocene and Eocene, a time with little land ice, then steepening to >7 °C after the Eocene once land ice was present on Antarctica.

No Evidence for a Large Atmospheric CO2 Spike Across the Cretaceous‐Paleogene Boundary

“…Stenochlaena, values come from S. palustris reported here; for G. adiantoides, values come from field-grown G. biloba at~400 ppm CO 2 (Kowalczyk et al, 2018). For aff.…”

“…For aff. To solve for atmospheric CO 2 , we use the Kowalczyk et al (2018) code written in R (v.3.4.4; R Core Team, 2018). For other inputs not directly measured, we used the recommended values from Franks et al (2014) or appropriate values from the literature (see Data Set S1).…”

“…For aff. Stenochlaena, values come from S. palustris reported here; for G. adiantoides, values come from field-grown G. biloba at~400 ppm CO 2 (Kowalczyk et al, 2018). For other inputs not directly measured, we used the recommended values from Franks et al (2014) or appropriate values from the literature (see Data Set S1).…”

“…For other inputs not directly measured, we used the recommended values from Franks et al (2014) or appropriate values from the literature (see Data Set S1). To solve for atmospheric CO 2 , we use the Kowalczyk et al (2018) code written in R (v.3.4.4; R Core Team, 2018).…”

“…The model has been used to reconstruct CO 2 during the Phanerozoic (Franks et al, 2014), including the late Paleozoic , middle Cretaceous (Richey et al, 2018), Late Cretaceous (Martínez et al, 2018), early Paleocene (Kowalczyk et al, 2018), middle Eocene (Maxbauer et al, 2014;Wolfe et al, 2017), Oligocene-Miocene boundary (Reichgelt et al, 2016;Tesfamichael et al, 2017) and early Miocene .…”