Silicate melt viscosity database for gpvisc

Overall description This document is the silicate melt viscosity database used to calibrate the machine learning model gpvisc (Le Losq et al., 2025, doi.org/10.1016/j.epsl.2025.119287). It is a compilation of most of the existing experimental data about silicate melt viscosities. All compositional data is expressed in mol% and normalized to 100, P in GPa, T in K, and viscosity is log10 viscosity expressed in Pa.s.

Sheet description This xlsx file is composed of three sheets: - lr represents the low pressure (1 atm) dataset - hr represents the high pressure (>1 atm) dataset - References is the list of all publications used in both lr and hr datasets All sheets are also available independently in csv format This work was supported by the Data Intelligence Institute of Paris (diiP) and IdEx Université Paris Cité (ANR-18-IdEx-0001), and by the LabEx UnivEarthS (ANR-10-LABX-0023 and ANR-18-IDEX-0001).