Investigating the relationship between Total Air Content (TAC) variations in polar ice cores and surface climate conditions

Ice cores constitute a valuable archive for reconstructing climate and atmospheric composition from glacial-interglacial to annual timescales. The Total Air Content (TAC), corresponding to the total volume of air trapped in ice, reflects changes in atmospheric pressure, temperature, and pore volume at the bubble close-off at the bottom of the firn. Building on these properties, TAC has been employed as a paleoelevation proxy and more recently as an orbital dating tool. Pore volume at bubble close-off depends not only on atmospheric pressure but also on local surface conditions driving firn densification and air entrapment efficiency through processes like compaction and snow grain metamorphism.

Investigating the relative impact of different surface climate parameters on the TAC signal, requires evaluating variables such as local insolation, accumulation rate, and seasonal temperature variations. Previous studies have mainly focused on site-specific analyses, limiting broader insights into regional and global patterns. To address this gap, we compiled TAC data from 30 ice cores across Antarctica and Greenland, combining published datasets with new measurements from the EDC, EDML and TALDICE ice cores. This data compilation includes sites with highly contrasting local climatic conditions, in terms of accumulation rates (1150 to 22 mm w.e. yr^-1) and surface temperatures (-14 to -58°C). In addition to surface parameters (e.g. reconstructed annual surface temperatures and accumulation rates, Half Year local Summer Insolation index and atmospheric pressures), simulated summer temperatures from an Earth system model of intermediate complexity were used to be compared to past TAC changes. Then, we apply a series of statistical analyses on the compiled dataset, including linear and multiple regression analyses as well as residual analyses, to evaluate the relationships between TAC and the different environmental parameters at orbital and millennial scales. We also compare the measured TAC datasets with TAC outputs from the IGE firn densification model.

Our results highlight regional contrasts in the relationship between TAC variations and the different surface climate parameters. For Greenlandic ice cores we observed strong correlations observed between TAC and climatic parameters. For instance, at NGRIP and GRIP sites, coefficients of determination (R²) between TAC and Half Year Summer Insolation are higher than 0.6. Antarctic sites, including those on the East plateau, exhibited more variable and site-specific responses. For example, at EDC and Dome Fuji sites, the R² between TAC and Half Year Summer Insolation is respectively 0.3 and 0.6. These findings underline the critical importance of addressing site-dependent dynamics to use TAC as a robust environmental proxy and orbital dating tool.