Comparison of three methods to obtain the apparent dielectric constant from time domain reflectometry wave traces

Authors: Timlin, Dennis; PACHEPSKY YAKOV - DUKE UNIVERSITY, NC

Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/16/1995
Publication Date: N/A
Citation: Timlin, D.J., Pachepsky Yakov 1996. Comparison of three methods to obtain the apparent dielectric constant from time domain reflectometry wave traces. Soil Science Society of America Journal. 60:970-977.

Interpretive Summary: Measurement of the water content in soil is an extremely important activity in the fields of agriculture and engineering. Soil water content is used to forecast irrigation requirements and can be used to track the movement of chemicals in soil. Time domain reflectrometry (TDR) is an accurate and flexible method to measure water content in soil and can be easily automated. TDR measures water content by sending a voltage pulse along metal rods that have been inserted into the soil and measuring the time it takes for the voltage pulse to return to the instrument The travel time is proportional to water content. We compared three methods to obtain this travel time, a manual method, use of a computer algorithm and a method that simulates the TDR signal using the theory of voltage propagation in an electrical circuit. An optimization procedure adjusts the travel time until the simulated signal matches the measured one. The simulation method reproduces the measured signal accurately and gives travel times that are not significantly different from the other two methods. The simulation method is an improvement over the other two methods because it is not subject to operator error, is robust, and can be easily automated. The simulation method can also be a useful tool in theoretical studies of signals from TDR's in soil-water systems.

Technical Abstract: Calibration of time domain reflectometry (TDR) probes to measure soil water in a particular soil is desirable since no universal relationship between water content and apparent dielectric constant (Ka) has been found. We compared three calibration procedures based on three documented methods of computing Ka from TDR traces: a manual method that uses a plot of the wavetrace, a derivative-based computer algorithm, and fitting a simulated TDR trace to a measured one using a simple multiple reflection model. The latter method was modified to include the rise time of the TDR instrument. TDR traces were measured using a Tektronix 1502B TDR in-situ on a Beltsville Silt Loam, a Fine-loamy mixed mesic Typic Fragiudult. Fifteen-cm probes having three rods were used. The water contents which were measured gravimetrically using soil cores, ranged from 0.132 to 0.332 m3/m3. A linear regression of the square root of Ka and soil water was sufficient to describe the dependence of Ka on soil water content. Calibration using apparent dielectric constants obtained from the derivative-based algorithm gave the smallest standard error of predictions, and the manual method was better than the wave simulation method. The intercepts and slopes of the calibration equations for the three methods, however, were not significantly different from each other. The wave simulation method gives physically meaningful values for Ka and characteristic impedances and can be a useful tool for studies of theoretically based calibration of TDR.