Extended Data Fig. 2: Insolation imprint in δAr/N2 ratios of the air trapped in ice containing (a) bubbles and (b) clathrates only.
From: Early Pleistocene East Antarctic temperature in phase with local insolation

The concept here is similar to Fig. 1 in the main text, with Dome Fuji (red)28, Dome C (blue)27, and Allan Hills S27 (gray)30 datasets. Clathrate-based δAr/N2 is negatively correlated with local Dec 21st insolation intensity in both Dome Fuji (sample N = 219, r2 = 0.24, two-tailed p < 0.01) and Dome C ice (N = 40, r2 = 0.21, p < 0.01). There is also a significant (two-tailed p < 0.05) correlation between bubble-based δAr/N2 and insolation in Dome Fuji (N = 238, r2 = 0.24), Dome C (N = 29, r2 = 0.20), and S27 (N = 24, r2 = 0.23) ice. Dashed lines represent 95% confidence interval of the slope. Note that the number of available δAr/N2 data points is smaller than the number of δO2/N2 data, because not all δO2/N2-reporting studies include δAr/N2.