Soil water content has an important impact on many fundamental biophysical processes. The quantification of soil water content is necessary for different applications, ranging from large-scale calibration of global-scale climate models to field and catchment scale monitoring in hydrology and agriculture. Many techniques are available today for measuring soil water content, ranging from point scale soil water content sensors to global scale, active and passive, microwave satellites. Geophysical methods are important methods, used for several decades, to measure soil water content at different scales. Among these methods, ground penetrating radar has been shown to be one of the most reliable and promising ones. Soil water content measurement using ground penetrating radar requires the application of parametric equations that will convert the measured dielectric permittivity to water content. While several studies have been performed to test equations for soil water content sensors such as time domain reflectometry, a few studies have been performed to test different formulae for application to ground penetrating radar. In this study, we compare available formulae for converting dielectric permittivity obtained from detailed laboratory scale measurement of reflected waves using ground penetrating radar. Four soils covering a wide range of textures were used and the measured soil water contents were compared with values obtained from gravimetric measurements. Results showed that the dielectric mixing model of Roth et al. (1990) provided the best fit for individual soil textural classes, except for sandy soils. However, for all data combined the dielectric mixing model performed much better with significant difference in coefficient and determination and root mean square error. Empirical equations developed from calibration of time domain reflectometry performed poorly when applied to estimation of soil water content obtained from ground penetrating radar. Differences in sample volume, frequency of operation and data analysis between ground penetrating radar and time domain reflectometry, suggest to use more flexible and robust electromagnetic mixing formulae, allowing for incorporating the dielectric properties of constituents materials and geometrical features of the media. Sensitivity analysis was then performed to provide detailed information for the most accurate application of the selected dielectric model. Sensitivity analysis showed that the geometric parameter α and the dielectric permittivity of the solid phase ∊s are the two most sensitive parameters, determining important variations in the estimation of soil water content. Based on these results, these two parameters are suggested as fitting parameters, to be selected if the model is fitted to data. Otherwise, the model can be successfully used without calibration, as presented in this study, by using α = 0.5 (as also suggested by the authors) and ∊s = 4, which is an average value for soil minerals.
Anbazhagan P., Bittelli M., Pallepati R.R., Mahajan P. (2020). Comparison of soil water content estimation equations using ground penetrating radar. JOURNAL OF HYDROLOGY, 588, 1-9 [10.1016/j.jhydrol.2020.125039].
Comparison of soil water content estimation equations using ground penetrating radar
Bittelli M.
Investigation
;
2020
Abstract
Soil water content has an important impact on many fundamental biophysical processes. The quantification of soil water content is necessary for different applications, ranging from large-scale calibration of global-scale climate models to field and catchment scale monitoring in hydrology and agriculture. Many techniques are available today for measuring soil water content, ranging from point scale soil water content sensors to global scale, active and passive, microwave satellites. Geophysical methods are important methods, used for several decades, to measure soil water content at different scales. Among these methods, ground penetrating radar has been shown to be one of the most reliable and promising ones. Soil water content measurement using ground penetrating radar requires the application of parametric equations that will convert the measured dielectric permittivity to water content. While several studies have been performed to test equations for soil water content sensors such as time domain reflectometry, a few studies have been performed to test different formulae for application to ground penetrating radar. In this study, we compare available formulae for converting dielectric permittivity obtained from detailed laboratory scale measurement of reflected waves using ground penetrating radar. Four soils covering a wide range of textures were used and the measured soil water contents were compared with values obtained from gravimetric measurements. Results showed that the dielectric mixing model of Roth et al. (1990) provided the best fit for individual soil textural classes, except for sandy soils. However, for all data combined the dielectric mixing model performed much better with significant difference in coefficient and determination and root mean square error. Empirical equations developed from calibration of time domain reflectometry performed poorly when applied to estimation of soil water content obtained from ground penetrating radar. Differences in sample volume, frequency of operation and data analysis between ground penetrating radar and time domain reflectometry, suggest to use more flexible and robust electromagnetic mixing formulae, allowing for incorporating the dielectric properties of constituents materials and geometrical features of the media. Sensitivity analysis was then performed to provide detailed information for the most accurate application of the selected dielectric model. Sensitivity analysis showed that the geometric parameter α and the dielectric permittivity of the solid phase ∊s are the two most sensitive parameters, determining important variations in the estimation of soil water content. Based on these results, these two parameters are suggested as fitting parameters, to be selected if the model is fitted to data. Otherwise, the model can be successfully used without calibration, as presented in this study, by using α = 0.5 (as also suggested by the authors) and ∊s = 4, which is an average value for soil minerals.File | Dimensione | Formato | |
---|---|---|---|
Bittelli_ Comparison.pdf
Open Access dal 01/10/2022
Tipo:
Postprint
Licenza:
Licenza per Accesso Aperto. Creative Commons Attribuzione - Non commerciale - Non opere derivate (CCBYNCND)
Dimensione
1.02 MB
Formato
Adobe PDF
|
1.02 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.