A new model for predicting soil thermal conductivity from matric potential

Abstract

Effective soil thermal conductivity (λeff) is required by a variety of science and engineering studies and associated numerical simulations. Because water content is the dominant factor determining the magnitude of λeff, modeling of λeff is usually based on the measured water content (θ) or degree of saturation (Sr). However, these approaches are largely soil dependent and no model was found for universal application. Prediction of λeff from energy states of soil water (i.e., ψ, the matric potential or matric suction) provides a unique way that takes pore size and geometry of soil into account and the resultant λeff(ψ) relationship is less sensitive to soil types compared to the λeff(θ) and λeff (Sr) relationships. The λeff(ψ) relationship developed by McCumber and Pielke (1981) (MP1981) has been incorporated into over 10 land surface, climate or hydrological models for wide applications. But the MP1981 model was reported to give unreasonable estimates especially at low ψ range. Lu et al. 2019 (LS2019) proposed an improved λeff(ψ) model but limited to applications in high ψ range. The objective of this study was to develop a new λeff(ψ) model for applications across the full range of matric potential. Experimental measurements of λeff(ψ) taken on 26 soils from four studies were used for the model development. A comparison of the new model with the MP1981, LS2019 and three other literature models showed the new model gave satisfactory estimates (RMSE < 0.2 W m−1 °C−1, AD ≈ 0 W m−1 °C−1 and NSE > 0.9) that outperformed all the other models. However, there is a lack of experimental measurements of λeff(ψ) to further validate and calibrate this model. Future efforts on the measurement of λeff(ψ) should be conducted and studies that investigate the effects of air entry and hysteresis on the behavior of λeff(ψ) relationship is recommended.