Reconciling ice core CO₂ and land-use change following New World-Old World contact

Precise measurements of atmospheric CO2 throughout the last two millennia, derived from ice cores, are crucial for understanding the global carbon cycle and providing a pre-industrial baseline for comparison with the current, anthropogenic CO2 rise. High-resolution records are limited to the Law Dome and WAIS Divide ice cores, each with strengths and limitations. The Law Dome core offers high temporal resolution due to its high accumulation rate, but its data have undergone multiple revisions. The WAIS Divide core provides higher sample resolution and precision but suffers from greater smoothing due to its lower accumulation rate. Discrepancies exist between these records, particularly around a distinctive CO2 minimum in the Law Dome record around 1610 CE, known as the 'Orbis Spike', which has been linked to land-use changes in the Americas following the Columbian Exchange. However, this minimum is not clearly present in the WAIS Divide record. This study aims to resolve these discrepancies by presenting high-resolution CO2 data from the Skytrain ice core, providing an independent dataset to test the validity of the existing records and their interpretations.

Existing high-resolution CO₂ records for the Common Era primarily come from the Law Dome and WAIS Divide ice cores. The Law Dome record, while possessing high temporal resolution due to its high accumulation rate, has undergone several revisions and contains features, like the 1610 CE minimum, that are difficult to reconcile with our current understanding of the carbon cycle. The WAIS Divide record, with a lower accumulation rate, shows less high-frequency variability and lacks the prominent 1610 CE minimum. Several studies have attempted to explain these discrepancies, proposing links between rapid CO₂ changes and historical events such as plagues, wars, and climate-CO₂ feedback. The 1610 CE minimum, in particular, has been attributed to large-scale land abandonment and subsequent regrowth in the Americas following the arrival of Europeans, a hypothesis that is central to ongoing debates about the onset of the Anthropocene. However, the lack of a clear minimum in the WAIS Divide record introduces significant uncertainty.

This study analyzed CO₂ at 31 depth intervals from the Skytrain ice core, spanning 1454-1688 CE. Samples were measured using the Oregon State University (OSU) crusher system with 2-3 replicates per sample. Smoothing splines were generated for the Skytrain, Law Dome, and WAIS Divide records using a bootstrapped Monte Carlo simulated cubic smoothing spline procedure. The influence of firn smoothing (low-pass filtering) on the gas records was investigated using firn filters generated by the OSU firn model, which accounts for gas diffusion, advection, mixing, and bubble closure. The Law Dome record was convolved with the WAIS Divide and Skytrain firn filters to assess whether the 1610 CE minimum could be solely attributed to firn-based smoothing. To validate the firn filters, the same approach was applied to CH₄ records from the three cores. Finally, single deconvolution experiments were performed to reconstruct net carbon fluxes (land-atmosphere and ocean-atmosphere) using the OSU Carbon Cycle Box Model, with atmospheric CO₂ histories derived from the spline fits to the ice core data. The results were then compared to bottom-up scenarios of anthropogenic land-use change using different land-use datasets (HYDE and KK10).

The Skytrain CO₂ record reveals a gradual decrease in CO₂ into the 17th century (8.0 ppm over 157 years, or 0.5 ppm per decade), consistent with the WAIS Divide record but contrasting with the more rapid decrease in the Law Dome record (9.9 ppm over 84 years, or 1.2 ppm per decade). Analysis of firn smoothing using the OSU firn model demonstrated that the 1610 CE minimum observed in the Law Dome record could not be solely attributed to firn-based processes, as this minimum is absent in the WAIS Divide and Skytrain records. Even artificially increasing the firn smoothing parameters to eliminate the 1610 CE minimum in the Law Dome CO2 convolution results in overly smoothed methane records, showing this amount of smoothing is not physically plausible. Deconvolution experiments revealed that the land carbon sink during the period was much shallower (0.26 ± 0.05 PgC per year) based on WAIS Divide, Skytrain, and combined reconstructions, compared to that based on the Law Dome record (0.47 ± 0.19 PgC per year). The shallower sink is consistent with high-end scenarios of land-use change in the Americas, while the more drastic sink derived from the Law Dome data is incompatible with even the most extreme land-use scenarios. These findings strongly suggest that the 1610 CE CO₂ minimum in the Law Dome record is not a true representation of atmospheric CO₂ but rather a potential artifact.

This study's findings challenge the interpretation of the 1610 CE CO₂ minimum in the Law Dome record as a primary indicator of the onset of the Anthropocene driven by rapid land-use change in the Americas following New World-Old World contact. The more gradual CO₂ decrease observed in the Skytrain and WAIS Divide records aligns better with the currently accepted understanding of climate-carbon feedbacks and high-end scenarios of land-use change. While large-scale land abandonment likely played a role in the CO₂ decrease, it was likely a more gradual process than previously suggested by the interpretation of the Law Dome data. Although natural climate variations could have influenced the CO₂ decrease, the lack of similarly large CO₂ decreases during other periods of comparable cooling in the Northern Hemisphere makes this hypothesis less likely. The study highlights the need for caution when interpreting high-frequency variability in ice core records, especially those with lower accumulation rates, and suggests re-examination of existing Law Dome data or acquiring new high-resolution records from other Antarctic regions.

The analysis of the Skytrain ice core data reveals a gradual decrease in atmospheric CO₂ around 1600 CE, challenging the interpretation of a sharp drop at 1610 CE in the Law Dome record as a true atmospheric signal. The findings corroborate models of land-use change in the Americas following New World-Old World contact but suggest this change was more gradual. Future research should focus on clarifying the origin of the discrepancy between the Law Dome and other ice core records, possibly through re-measurement of existing samples or the acquisition of new high-resolution cores. Advanced measurement techniques and improved understanding of firn processes will further refine our understanding of rapid carbon cycle changes during this critical period.

The study's interpretation relies on indirect evidence regarding the validity of the 1610 CE CO₂ minimum in the Law Dome record. While the convolution analysis using different firn filters and the comparison with CH₄ data strongly suggest an artifact, direct evidence for the source of the discrepancy remains elusive. Furthermore, the resolution of existing land-use change models is relatively coarse, limiting the precision of comparisons between model outputs and the high-resolution ice core data. Finally, this analysis focuses solely on CO2 and CH4; inclusion of other relevant isotopic and atmospheric trace gas data would enhance our understanding of the processes driving changes in this period.