Normalized Concept for Modelling Effective Soil Thermal Conductivity from Dryness to Saturation

Hailong He,Kosuke Noborio,Oistein Johansen,Miles F. Dyck,Jialong Lv

DOI: https://doi.org/10.1111/ejss.12820

IF: 4.178

2019-01-01

European Journal of Soil Science

Abstract:Effective soil thermal conductivity (λ eff) is a critical parameter for environmental and earth science as well as engineering applications. Models to predict λ eff are required in diverse global and community land surface schemes as well as climate models to investigate coupled water and heat transport in soils and heat exchange at the earth surface. Among the many soil thermal conductivity models, models based on the normalized concept are most often developed and utilized for estimating λ eff. However, at present no systematic study has been performed to investigate the origin and evolution of the normalized thermal conductivity models, nor to evaluate their performance with large datasets. The objectives of this study were to: (a) review the development and evolution of the normalized thermal conductivity models, and (b) assess their performance with datasets consisting of soils with a full range of water saturation and a wide range of soil textures and bulk densities. A total of 38 normalized thermal conductivity models were critically reviewed and their relationships were clearly outlined. Their performance was evaluated by five categories according to model characteristics with a compiled dataset consisting of 71 soils and 669 tests collected from nine studies. Our analysis demonstrated key roles of the quartz content, solid thermal conductivity and choice of the Kersten functions in the model applicability and accuracy of estimating λ eff. The results showed that the Y2018, CK2005, CK2006, J1975, L2007 and T2009 models have the best performance among the models without fitting parameters, but further improvements are required to apply them universally. Although the models of H2017, LD2015, M2006 and K2007 are the best performing models with fitting parameters, approaches to calculate these parameters are required so they can be easily applied. Future studies on parametrization of currently well‐performing models for wider and more accurate application, development of a soil thermal conductivity database for model evaluation and calibration purposes, and connecting soil thermal conductivity models to hydraulic properties are recommended.