A review and evaluation of thermal conductivity models of saturated soils

Thermal conductivity of saturated soil (λsat) is a critical parameter for environmental and engineering applications. A wide variety of models used to predict λsat have been presented in the literature, but there is a lack of systematic evaluation of their performance. The objective of this study was to evaluate the performance of available λsat models with a compiled dataset containing λsat measurements from various soil types. A total of 52 models, including 14 theoretical models, 10 semi-empirical models and 28 empirical models, were assessed using a large dataset consisting of more than 295 measurements on 168 soils from 23 studies. The results showed that none of the empirical models gave satisfactory estimates of λsat. One semi-empirical model (i.e. LK1926II or the geometric mean model) and seven theoretical models (i.e. MA1951, DV1963, KT1984, KT2007-TV18, HG1986, CS2010, and TV2015) satisfactorily modeled λsat of almost all soil types with Nash–Sutcliffe Efficiency > 0.8 and RMSE < 0.35 W m−1 °C. The geometric mean model (i.e. LK1926II) was the most parsimonious model with a high degree of predictive power.