(Received 12 October 1999; accepted 2 November 2000)
Published Online: 01 June 2001
| ||Format||Pages||Price|| |
|PDF (236K)||13||$25||  ADD TO CART|
Cite this document
Time domain reflectometry (TDR) systems are being used widely for in situ measurement of soil water content. Water content is calculated based on the measured dielectric properties of the soil system. TDR systems for water content measurement were originally developed for use in soils with low bulk densities similar to those found in agriculture. Four highway test sites in Tennessee were instrumented with five-segment TDR probes in the soil subgrade and in the unbound aggregate sub-base (Rainwater et al. 1999). These materials are more dense than agricultural soils. After several months of data collection and field verification, the TDR predicted water contents and gravimetric water contents did not coincide. For this reason, a study was performed with the TDR equipment using each of the four test site subgrade soils and an unbound aggregate sub-base sample. Ten previously published TDR water content relationships were evaluated to determine which relationship most accurately predicted water content for the subgrade soils and for the unbound aggregate sub-base using the five-segment probe. During the course of the study, it was necessary to evaluate the manufacturer's relationship between the measured propagation time and the corrected propagation time, which accounts for the epoxy fill between the probe waveguides. The relationship was revised to reduce significant variation in corrected propagation times between segments.
The relationship between inverse signal velocity and soil water content proposed by Herkelrath et al. (1991) most accurately predicted water content for all subgrade soils; however, this relationship required the derivation of a soil-specific slope and intercept. The equation proposed by Baran (1994) most accurately predicted water contents for the unbound aggregate sub-base. Both of these expressions appear to provide a better prediction than the widely used Topp's relationship.
Research associate, The University of Tennessee, Knoxville, TN
Professor and head, The University of Tennessee, Knoxville, TN
Research engineer, The University of Tennessee, Knoxville, TN
Professor, The University of Tennessee, Knoxville, TN
Stock #: GTJ11337J