On the Polar Bias in Ice Core ¹⁰Be Data

Cosmogenic radionuclide records from polar ice cores provide unique insights into past cosmic ray flux variations. They allow reconstructions of past solar activity, space weather, and geomagnetic field changes, and provide insights into past carbon cycle changes. However, all these applications rely on the proportionality of the ice core radionuclide records to the global mean production rate changes. This premise has been long debated from a model and data-perspective. Here, we address this issue through atmospheric mixing model experiments and comparison to independent data. We find that all mixing scenarios, which do not assume complete tropospheric mixing, result in a polar bias. This bias is more prominent for geomagnetic field changes than solar modulation changes. The most likely scenario, supported by independent geomagnetic field records and marine Be-10 during the Laschamps geomagnetic field minimum, results in a dampening of geomagnetic field induced changes by 23%-37% and an enhancement of solar-induced changes by 7%-8%. During the Holocene, we do not find conclusive evidence for a polar bias. We propose a correction function that allows deconvolving the glacial ice core record in order to restore proportionality to the global mean signal.