Journal of Earth Science, Vol. 21, No. 6, p. 961–967, December 2010 Printed in China DOI: 10.1007/s12583-010-0149-2
ISSN 1674-487X
Complex Effective Relative Permittivity of Soil Samples from the Taunus Region (Germany) Katja Lauer*, Christian Albrecht Institute of Soil Science and Soil Conservation, Research Centre for BioSystems, Land Use and Nutrition (IFZ), Justus-Liebig-University Giessen, Germany Christina Salat Department of Underground Space for Storage and Economic Use, Federal Institute for Geosciences and Natural Resources, Hannover, Germany Peter Felix-Henningsen Institute of Soil Science and Soil Conservation, Research Centre for BioSystems, Land Use and Nutrition (IFZ), Justus-Liebig-University Giessen, Germany ABSTRACT: The most important parameter affecting ground-penetrating radar (GPR) measurements is the complex effective relative permittivity GPR pulses. Knowing
ε
* r,eff
* ε r,eff
because it controls the propagation velocity and the reflection of
of soils passed through by electromagnetic waves increases accuracy in soil
thickness and interface identification. Complex effective relative permittivity
* ' '' ε r,eff = ε r,eff − jε r,eff
of 25
soil samples with textures ranging from loamy sand to silty clay was measured using the two-electrode parallelplate method. The measurements were conducted at defined water contents for frequencies from 1 MHz to 3 GHz. The results confirm the frequency dependence of
* ε r,eff
and show that the dielectric behavior of
soil-water mixtures is a function of water content. Applying the experimental data of this study with predictions based on the empirical model by Topp et al. (1980), we find that Topp et al.’s curve tends to underestimate the real part of
* ε r,eff
measured. Along with frequency and water content, soil texture and organic matter affect
soil permittivity. Moreover, the real part of calibration model enables us to predict
* ε r,eff
* ε r,eff
increases at higher dry bulk densities. Output from our
for the soil samples which were tested under the actual in situ
soil water content. This results in high accuracy of soil thickness prediction. KEY WORDS: ground-penetrating radar (GPR), complex effective relative permittivity, soil sample.
INTRODUCTION In the mountainous regions of the Rhenish massif, Germany, high stone contents limit the effectiveness This study was supported by the German Research Foundation (DFG) (No. SFB 299). *Corresponding author:
[email protected] © China University of Geosciences and Springer-Verlag Berlin Heidelberg 2010 Manuscript received June 7, 2010. Manuscript accepted August 10, 2010.
of traditional soil surveying tools like drilling or soil profile analysis (Gerber, 2009). We used groundpenetrating radar (GPR) as a non-destructive technique to examine these soils and map their thickness over bedrock. To receive the highest possible accuracy in soil thickness prediction, we established a laboratory-based calibration model. GPR uses electromagnetic energy to image the shallow subsurface by transmitting radio-wave pulses into the soil. While propagating through the ground, they are scattered, attenuated and reflected at interfaces (stratigraphic layers, soil horizons, etc.) that
962
Katja Lauer, Christian Albrecht, Christina Salat and Peter Felix-Henningsen
have contrasting dielectric properties. When pulses have returned to the surface, they are detected as a function of the two-way travel time t (ns) by the receiving antenna (Inman et al., 2001; Knoll, 1996). To convert t into depth information, the relative propagation velocity vr of the electromagnetic waves has to be known. Assuming homogenous and isotropic soils, vr can be calculated from the following expression * vr = c ε r,eff (1) * where c is the speed of light in vacuum; and ε r,eff is the complex effective relative permittivity (Daniels, 2004). The depth d to a reflector is derived from (2) d=vr(t/2) Equation (1) shows that wave propagation ve* locities in the ground mainly depend on ε r,eff . In ad* dition to the velocity, ε r,eff also controls the reflection of GPR pulses at interfaces. By controlling the * velocity and the reflection of GPR signals, ε r,eff is the second most important parameter affecting GPR applications after complex effective electrical conduc* * tivity σ eff . In general, ε r,eff consists of a real part,
' ε r,eff , representing the polarisability of soil materials, " and an imaginary part, ε r,eff , representing the energy
losses due to polarisation and ionic conductivity (e.g., Shang et al., 2000). The objective of this study was to establish a re* gional calibration model for ε r,eff as a function of * the volumetric water content θ. To this end, ε r,eff of 25 soil samples with various textures was measured at defined water contents under lab conditions. Since the soil texture and the in situ soil water content at the time of GPR survey can be determined in the field, the calibration model serves to provide accurate predictions of the wave propagation velocities. This information is the key factor for accurate soil thickness indications. MATERIALS AND METHODS Origin of Soil Samples Tested Soil samples examined in this study originate from soil profiles in the Taunus area, in the south-eastern part of the Rhenish massif (Fig. 1). The
Figure 1. Map of study area (Taunus, south-eastern Rhenish Massif) in Germany. a.s.l. Above sea level.
Complex Effective Relative Permittivity of Soil Samples from the Taunus Region (Germany)
geology of the Taunus is set up of slate, sandstone and quartzite of Devonian formation. With the exception of floodplains, the study area is almost completely covered by periglacial slope deposits formed in the Late Pleistocene by solifluction processes and loessdeposition (Gerber, 2009). Typical soils of the Taunus region are Regosols, Cambisols, stagnic Luvisols and Planosols as well as Gleysols in river valleys. Measurement Technique * Complex effective relative permittivity ε r,eff was measured within the frequency range from 1 MHz to 3 GHz at room temperature and atmospheric pressure with an Agilent impedance analyser (E4991A RF) and a Novocontrol high-frequency measuring cell (Fig. 2). The cell consists of two parallel-arranged plates of 12-mm diameter forming a capacitor. The soil samples were sandwiched between electrodes and enclosed by a Teflon ring. After stimulating the sample with an alternating-current (AC) source, the real and imagi* nary parts of ε r,eff were calculated from the capacitance Cp and the dissipation factor D. For further details about the measuring technique, see Salat and Junge (2010).
Figure 2. Two-parallel-plate measuring cell used in this study. Sample Preparation and Measurement Procedure Samples were taken from horizons of nine pedons representing typical soils of the Taunus region. Each sample was dried at 40 ℃ and passed through a 2-mm sieve. Soil texture was determined by the combined sieve and pipette method (Schlichting et al., 1995) using Na4P2O7 as the dispersant. Samples rich in soil organic matter (SOM) were pre-treated with H2O2. SOM content was measured with Vario EL III elementar analyser at 1 200 ℃ for A horizons. The
963
properties of the soil samples tested are summarized in Table 1. Soil textures exhibit a range from loamy sand to silty clay with medium to high contents of organic matter in the topsoil. Table 1
Soil texture and soil organic matter (SOM) contents of the samples tested Soil
Sand
Silt
Sample
0.063 to
0.002 to
No.
