The soil water content is an important variable in soil physics and agricultural production. At small scales, TDR (time domain reflectometry) probes can measure the water content and especially changes of water content with very high accuracy (Jacobsen and Schj0nning, 1993; Nissen et al., 1999; Roth et al., 1990). But there is still a lack of methods suitable for larger areas and measurements of heterogeneity of soil water content. Using TDR, a high number of probes have to be installed, which leads to considerable cost and work. To determine small-scale heterogeneities, the TDR also has the disadvantage of disturbing the area by installing probes. For these cases, ground-penetrating radar (GPR) is an alternative measuring device. With respect to water content variability analysis, two different procedures based on GPR measurements can be distinguished.
Various authors use tomographic or multi-offset radar methods to determine the velocity of an electromagnetic wave through the subsurface. The dielectric constant can be calculated approximately from the velocity. Water content is then usually derived from the dielectric constant using the popular empirical Topp equation (Topp et al., 1980) or mixture formulae (i.e., Roth et al., 1990) which describe the relationships between dielectric and hydraulic parameters. Using this procedure, Hubbard et al. (1997), Parkin et al. (2000), Greaves et al. (1996), and Sénéchal et al. (2000a) investigated the use of radar data for estimating subsurface water content.
Another method of estimating water content distributions from GPR data is to look for a variable that statistically or geostatistically describes this distribution. In this case, the analysis of radar data is used for characterizing the heterogeneity of the subsurface. To calculate any kind of distribution from radar data, the recorded radar traces are analyzed using different attributes. This method is analogous to seismic trace analyses. Chen and Sidney (1997) gave an overview of seismic attributes that are specific measurements of geometric, kinematic, dynamic, or statistical features derived from the recorded data. Using this procedure for electromagnetic applications, Sénéchal et al. (2000b) gave a complex interpretation of a three-dimensional GPR data set using attributes calculated from amplitude analysis of reflected radar waves. They got an understanding of the lateral continuities and discontinuities of the reflectors, the geometry of structures, and their dynamic characteristics. Knight et al. (1997) used the amplitude values recorded in the radar traces for a geostatistical analysis of the GPR data. The spatial variation in dielectric properties in the subsurface was closely related to the spatial variation in grain size. The geostatistical analysis captured information about the spatial distribution of the dominant sedimentological features. Rea and Knight (1998) found agreement using geostatistical analysis of a digitized photograph and the radar data (amplitudes). The GPR data imaged the spatial distribution of lithologies and could be used to quantify the correlation structure of the sedimentary unit. They also observed agreement between the spatial variation in dielectric properties in the subsurface and the spatial variation of the grain size, both indicating the heterogeneity of the subsurface. The authors hypothesized that information extracted from the GPR data can be used to describe spatial variability of hydraulic properties. Also Tercier et al. (2000) found that geostatistical analysis of GPR data gave an effective way of quantifying the correlation structure of the two-dimensional GPR image. Charlton (2000) used GPR techniques for spatially distributed measurements of volumetric soil moisture. He found significant relationships between maximum amplitude and moisture content, indicating the potential of GPR for a quantitative assessment of soil moisture at different depths.
This study summarizes and completes—by showing new data—two earlier experimental works on GPR application in well-defined systems (Schmalz et al., 2002, 2003; Stoffregen et al., 2002). At first, we focused on the general suitability of the GPR technique to depict temporal soil water content changes by analyzing the electromagnetic wave propagation. A more detailed analysis of the GPR signal was performed in a second step in order to get more insight into the spatial heterogeneity of the soil moisture distribution. We selected certain attributes of the individual radar traces, analyzed them statistically, and compared the computations to soil water content distributions as derived from hydraulic simulation studies. It is important to note that the GPR technique can be used to obtain different kinds of information. At first, the velocity analysis of the magnetic wave can be applied to obtain absolute soil water content information as an integral over the soil depth which is penetrated by the waves. At second, certain attributes of the GPR signal can be analyzed in order to obtain depth-resolved information of the underground.
Was this article helpful?