A modified normalized model for predicting effective soil thermal conductivity

Effective soil thermal conductivity (lambda (eff)) describes the ability of a multiphase soil to transmit heat by conduction under unit temperature gradient. It is a critical parameter for environmental science, earth and planetary science, and engineering applications. Numerous models are available in the literature, but their applicability is generally restricted to certain soil types or water contents (theta). The objective of this study was to develop a new model in the similar form of the Johansen 1975 model to simulate the lambda (eff)(theta) relationship of soils of various soil textures and water contents. An exponential type model with two parameters is developed and a new function for calculating dry soil thermal conductivity is presented. Performance of the new model and six other normalized models were evaluated with published datasets. The results show that the new model is able to well mimic lambda (eff)(theta) relationship of soils from sand to silt loam and from oven dry to full saturation. In addition, it has the best performance among the seven models under test (with root-mean-square error of 0.059 W m(-1) degrees C-1, average deviations of 0.0009 W m(-1) degrees C-1, and Nash-Sutcliffe efficiency of 0.994). The new model has potential to improve the reliability of soil thermal conductivity estimation and be incorporated into numerical modeling for environmental, earth and engineering studies.