Willcocks, S.Hasterok, D.Halpin, J.A.Walsh, J.Jennings, S.2025-10-232025-10-232025Tectonophysics, 2025; 911:230802-1-230802-100040-19511879-3266https://hdl.handle.net/2440/147951Modelling heat flux through the lithosphere requires accurate estimates of thermal conductivity, yet few regions have sufficient estimates to constrain thermal models. Global geochemical databases in contrast, have numerous samples, but lack physical property estimates. In this study, we combine thermal conductivity measurements on 1053 globally-distributed samples with known chemical composition—including 48 new analyses from Antarctica— to develop empirical relationships between conductivity and major element composition, modal mineralogy and normative mineralogy. Despite a skew in the residuals, all compositional models result in similar misfit (∼0.4 W m⁻¹ K⁻¹) with 95 % of samples within ±20 % of measured conductivities. We then apply this conductivity-composition relationship to the PetroChron Antarctica database to predict the thermal conductivity of 6995 igneous protoliths. We predict 95 % of thermal conductivity estimates for Antarctic geochemical samples range from 1.78 to 3.19 W m⁻¹ K⁻, with an average of 2.49 ± 0.31 W m⁻¹ K⁻¹. These empirical relationships provide a way to produce reasonable estimates of rock conductivity that can be used to improve heat flux estimates beneath glaciers and ice sheets where the composition of the rocks is knownen© 2025 Published by Elsevier B.V.Thermal conductivity; Rock properties; Geothermics; Antarctica; Geothermal heat flux; Heat flowJournal article10.1016/j.tecto.2025.230802744730Hasterok, D. [0000-0002-8257-7975]Walsh, J. [0000-0003-1183-9015]Jennings, S. [0000-0003-3762-7336]