In Situ Determination Of Soil Bulk Density Using Thermo-TDR Technique:A Thermal Conductivity-Based Approach

Soil bulk density (pb) is a critical parameter in soil science, ecology, and agronomy as well as engineering constructions. Due to the spatial and temporal nature of soil structure, it has been a challenge to monitor pb in situ. In recent years, the thermo-time domain reflectometry (Thermo-TDR) technique has been applied to determine ?b continuously under field conditions. The technique is based on the measurement of soil volumetric heat capacity (C) with heat pulse (HP) method and volumetric water content (0) with TDR, and ?b is estimated indirectly using the de Vries (1963) C model. Several studies have shown that the C-based method is affected strongly by errors in HP C data. The objectives of this study are to improve the HP method for estimating soil thermal properties, and to develop a new thermal conductivity model, both well improve the Thermo-TDR method for measuring pb. The dissertation includes the following four sections.In the first part, we improve the traditional pulsed-infinite-line-source (PILS) theory for estimating soil thermal properties by using the late-time data of the temperature-by-time curve, with the purpose to eliminate the errors resulted from finite probe properties and soil-probe thermal contact resistance. Based on the heat conduction equation, the PILS model was fitted to the late-time data (refered to PILS-Late method), and determined the optimized C, thermal diffusivity (?) and thermal conductivity (?). The PILS-Late method was validated using three independent experiments. In the first experiment, the specific heat of soil solids (cs) on various soils were determined with the improved HP method and were compared to those measured using the technique of differential scanning calorimetry (DSC). The results showed that the PILS-Late method provided more reliable cs and reduced the relative errors to within 3.2%. In the second experiment, we tested the accuracy of HP measured water contents. We demonstrated that the PILS-Late method reduced the overestimation error in HP ? effectively. In the third experiment, a one-dimensional heat transfer column was established to evaluate the performance of PILS-Late approach for measuring ?. When the PILS-Late approach was used, HP ? values agreed well with the values obtained with that from the heat flux plate method. In addition, compared with the estimates from the identical-cylindrical-perfect-conductors (ICPC) theory, the PILS-Late method provided more accurate thermal property data by reducing the errors from finite probe property as well as from soil-probe thermal contact resistance.In the second part, we proposed a new model for estimating ? from ?, ?b, and soil texture. The new model has two shape factors, which are related to the fractions of sand and clay, and ?b. The shape factors were calibrated on X curves measured on seven soils. The new model was evaluated using ? data obtained with repacked soil samples as well as from published datasets. The root mean square error (RMSE) values and bias of the new model were within 0.15 W m-1 K-1 and 0.10 W m-1 K-1, respectively. Due to the simple form and fewer parameters, the new model has the potential to be used in quantifying the interactions among ?, ?b, and ?.In the third part, we developed a ?-based approach for determining in situ ?b with the Thermo-TDR method. Soil ? and ? are determined with a Thermo-TDR sensor in situ, and ?b were estimated inversely with the new ? model with known texture information. The ?-based approach was tested on repacked soil cores as well as in situ field measurements under four tillage treatments. The results demonstrated that the X-based approach gave reliable ?b values with RMSE less than 0.17 g cm-3. Particularly, the X-based approach gave reliable Pb on fine-textured soils, but exhibited relatively large errors on coarse-textured soils. Under field conditions, the ?-based approach provided ?b values with relative error within 10%.In the last part, we monitored ?b and thermal property dynamics in a tilled layer during post-tillage period on a loam soil. Using a marked wood stick installed vertically in soil profile, we were able to monitor the changes of soil surface level under the influences of rainfall and wetting/drying cycles. During this process, ?b measurements with core method and Thermo-TDR method responded dynamically to rainfall events:both increased gradually with the continuing of rainfall and wetting/drying cycles.The results from this study have important implications in understanding of coupled processes of water, heat, and solute in steady and unsteady soils.