Assessing the impact of quartz content on the prediction of soil thermal conductivity

Abstract

Accurate predictions of soil thermal conductivity k are strongly influenced by the volumetric fraction of quartz, Θ q , data for which are very scarce. This paper reveals a new approach to estimate Θ q from measured k records. First, an equation that relates the normalised k of soil (K e ) to the degree of saturation S r is fitted to experimental k data, and the k at full saturation is assessed; then Θ q is calculated from a geometric mean model. This modelling approach was applied to the k data of 10 Chinese soils obtained by Lu et al., also containing k measurements at full dryness, and to soils investigated by Kersten with measured quartz content data. The fitted Θ q data for Chinese soils are noticeably different from the sand mass fraction, commonly assumed in the past as an equivalent of quartz content, consequently leading to irrational k estimates. Acceptably good agreement was obtained between fitted and measured quartz content for Kersten's soils. Five K e (S r ) functions were tested against the experimental data for ten Chinese soils, supplemented with calculated k at full saturation. Overall, the normalised function by Lu et al. was the most suitable for the soils tested. The assumption that K e (S r ) = 0, applied to Johansen's model extended to full dryness, worked well for fine soils, and was acceptable for coarse soils.