Location: Soil and Water Management ResearchTitle: A comparison of second order derivative based models for time domain reflectometry wave form analysis Author
|Wang, Zhuangji - Iowa State University|
|Kojima, Yuka - Gifu University|
|Chen, Yan - China Agricultural University|
|Horton, Robert - Iowa State University|
Submitted to: Vadose Zone Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/26/2017
Publication Date: 7/13/2017
Citation: Wang, Z., Schwartz, R.C., Kojima, Y., Chen, Y., Horton, R. 2017. A comparison of second order derivative based models for time domain reflectometry wave form analysis. Vadose Zone Journal. 16(7):1-10.
Interpretive Summary: Monitoring of soil water is important for the evaluation of water flows in the environment and management of irrigation. Time domain reflectometry is a standard method for monitoring soil water based on the travel time of a signal. A new corner preserving filter was developed to improve the accuracy of travel time estimation. The new algorithm yielded accurate travel time estimates across a wide range of electrical conductivities. Based on the observations of this study, use of the CPF may be superior to conventional filters to estimate water content in salt affected soils with TDR.
Technical Abstract: Adaptive waveform interpretation with Gaussian filtering (AWIGF) and second order bounded mean oscillation operator Z square 2(u,t,r) are TDR analysis methods based on second order differentiation. AWIGF was originally designed for relatively long probe (greater than 150 mm) TDR waveforms, while Z square 2(u,t,r) was originally designed for relatively short probe (<50 mm) TDR waveforms. The performances of AWIGF and Z square 2(u,t,r) on both long and short TDR probes have not been fully evaluated. Thus, the main objective of this study is to evaluate theoretically and experimentally the performances of AWIGF and Z square 2(u,t,r) on both long and short TDR probes. Theoretical evaluations are performed via mathematical analysis of AWIGF and Z square 2(u,t,r). Experimental analysis is performed via laboratory measurements of long probe and short probe waveforms obtained in calcium carbonate solutions with a range of electrical conductivity (EC) values, and testing the stability of Z square 2(u,t,r) and AWIGF after adding different magnitudes of Gaussian noise to the waveforms. In order to improve the stability of AWIGF on short TDR probe waveforms, an alternative corner preserving filter (CPF) is designed and embedded into AWIGF. Based on tests, the CPF filter preserves the second order derivatives of the TDR waveforms, and adaptively emphasizes the reflection positions compared to the original Gaussian filter in AWIFG. Both the theoretical and experimental tests illustrated the consistency among Z square 2(u,t,r), the original AWIGF and AWIGF with the CPF filter. The standard deviations of second reflection positions (t2) and relative permittivity (er) are less than 5% for all of the noise levels. In conclusion, Z square 2(u,t,r) and AWIGF can provide stable analysis for long and short probe TDR waveforms. Combined with the CPF filter, the AWIGF is capable of stably analyzing challenging short probe TDR waveforms.