The Ice Core Gas Age‐Ice Age Difference as a Proxy for Surface Temperature.

Ice cores provide detailed records of past climate change. Water stable isotopes are the most commonly used ice core climate proxy, but their quantitative interpretation remains challenging. Here, I argue that the gas age‐ice age difference (Δage) is a powerful proxy for past surface temperature. An analytical framework is derived that directly links past temperature to firn properties that can be reliably reconstructed (Δage, lock‐in depth). The framework is calibrated using both present‐day spatial patterns and last glacial maximum temperatures reconstructed via borehole thermometry. The usefulness of the method is demonstrated using three case studies from Greenland and Antarctic ice cores. The calibration suggests that several firn densification models, with the possible exception of the Herron‐Langway model, have insufficient sensitivity to accumulation rates. This low sensitivity, in combination with large amplitude temperature forcing, can explain historical difficulties of densification models in simulating ice age firn thickness in East Antarctica. Plain Language Summary: It is important for scientists to understand past natural climate change. Ice cores drilled in the polar regions contain ancient ice up to 800,000 years old and can be used to reconstruct past climate. Ice from the polar regions contains air bubbles that are trapped at the bottom of the thick (50–120 m) perennial snow pack called the firn. The air in these bubbles is younger than the ice that surrounds it, and this age difference is called Δage. This paper develops a new method to estimate past temperatures of the ice sheet surface using our knowledge of changes in ice core Δage. The method is very simple and fast and agrees very well with independent temperature reconstructions where available. Key Points: The ice core gas age‐ice age difference (Δage) is a powerful proxy for past surface temperatureA simple analytical framework allows for reliable past temperature estimation using empirical estimates of ΔageLow firn model sensitivity to accumulation rate contributes to published model‐data mismatch of ice age firn thickness in East Antarctica [ABSTRACT FROM AUTHOR]