A new thermal conductivity model for sandy and peat soils

Soil thermal conductivity (lambda) is frequently required when studying coupled heat and water transfer in soils. While lambda models are widely available for mineral soil, they are not as readily available for organic soil, which plays an increasingly significant role in climate change and carbon sequestration. This study used a heat-pulse method to measure the thermal properties of six soil peat:mineral mixes spanning wide ranges of water contents and bulk densities. Based on this, a new generalized lambda model was proposed, which was tested by fitting the model to heat-pulse measurements, and then comparing with several other widely-used empirical or physically-based models of lambda. The new model accounts for the effects of porosity, degree of saturation, organic matter content and soil texture on the thermal conductivity. It is capable of describing the measured data more effectively than other models due to its simplicity and applicability, as indicated by the lowest values of root mean square error (0.028 W m(-1) K-1) and the highest value of Nash-Sutcliffe efficiency (0.995). The new model has shown potential for use in studying numerical algorithms for describing coupled heat and mass transfer processes in a variety of multiphase soils. It can also be incorporated into global land surface models, particularly in terms of identifying the significance of peat soils on climate change and carbon sequestration.