Relationship between thermal and electrical conductivity curves of soils with a unimodal pore size distribution: Part 2. Estimating bulk electrical conductivity from thermal conductivity

Soil thermal conductivity (lambda) and electrical conductivity (a) influence heat conduction and electrical conduction through soils. In Part 1 of this two-part series, we demonstrated that for soils with a unimodal pore size dis-tribution, the lambda and a curves were interrelated and could be described with a unified series-parallel resistor model. Based on the conceptual model presented in Part 1, the "mirror image" phenomenon in the lambda-a rela-tionship was further evaluated. Starting with the Lu et al. (2007) lambda model, the "mirror image" phenomenon was used to derive a new normalized a model. The new a model was dependent on degree of water saturation (S), and shared the same shape parameters as those in the lambda model. Here, the new a model is examined using new datasets consisting of simultaneous thermo-TDR sensor measurements of soil water content (B), lambda and a. New model a values interpolated between known dry and saturated a values agreed well with measured a values, with RMSE values within 0.102 dS m-1 and bias values between-0.083 and 0.014 dS m-1 for a variety of soil samples. Using repeated in situ lambda measurements made in soils during an evaporative drying period, the new a model estimated normalized a values with RMSEs within 0.015 dS m-1. The new a model offers an effective way to estimate a of unsaturated soils.