4.2 Mineral-specific acid fractionation factor
Residuals of aragonite clumped isotope data around the low-temperature (<30°C) York regression (0.042‰; 1σ; see section 3. 1 and Fig. 2 ) is predominantly explained by analytical uncertainty on Δ47 measurements (0.026‰ and 0.046‰; 1σ; see section 2.3 ). Uncertainty on formation temperatures in the low-temperature dataset (±0.8°C; 1σ; see S4 ) would add an additional uncertainty of 0.0024‰ (1σ) if applied to the weighted average formation temperature of all low-temperature (<30°C) data points (22.0°C; see S4 ). Variability around the linear ∆47-\(\frac{1}{T^{2}}\) regression is almost completely explained by uncertainty in the calibration dataset, leaving little uncertainty on the temperature relationship in the low-temperature domain (<30°C; see section 3.2 ; Fig. 2 ). If clumped isotope fractionation during acid digestion is indeed different between aragonite and calcite (as suggested in Müller et al., 2017; Petersen et al., 2019), the difference in AFF is highly unlikely (95% CL) to exceed +0.007‰ or -0.001‰ (ΔΔ47 = 0.003 ± 0.004‰ relative to Anderson et al., 2021; 95% CL; S7 ). The calcite AFF in Petersen et al. (2019), which are included in the I-CDES reference scale (Bernasconi et al., 2021) can therefore be used for aragonite samples.