GPR MEASUREMENTS OF WATER LEVEL IN SILTY SOILS. Sandeep Pyakurel

GPR MEASUREMENTS OF WATER LEVEL IN SILTY SOILS Sandeep Pyakurel Problem report submitted to the College of Engineering and Mineral Resources at West Virginia University in partial fulfillment of the requirements for the degree of Master of Science in Civil Engineering Udaya B. Halabe, Ph.D., P.E., Chair Hema J. Siriwardane, Ph.D., P.E., Co-Chair Hota V. S. GangaRao, Ph.D., P.E. Department of Civil and Environmental Engineering Morgantown, West Virginia 2006 Keywords: GPR, Saturation, Silty Soil, Water Level

ABSTRACT Detection of water level in silty soils can be complicated because of capillary action. In this study, the water level in silty soil samples was detected using Ground Penetrating Radar (GPR) technique in the laboratory. The soil samples had dimensions of 62 cm x 48 cm x 46 cm for no flow condition and of 30 cm x 25 cm x 120 for the transient seepage boundary condition. The soil samples were kept in a clear Plexiglas container which facilitated water level measurements. Two groundcoupled antennas with frequencies of 900 MHz and 1,500 MHz were used in this study. The soil sample was dry at the beginning of the experiment. The water level in the soil sample was raised to a pre-determined level and radar readings were taken at different times over a period of 24 hours. The moisture content in the soil sample above the water level increased with time due to capillary action. At the end of the experiment, the variation of moisture content with depth within the sample was experimentally determined. The GPR observations were compared with measured water saturation level in the soil sample. The study showed that radar detects water in the capillary rise zone as the soil-water interface. Also, the saturation level at the radar detected boundary typically varied from 20 to 40 percent. The corresponding volumetric water content was in the range of 9 to 20 percent.

ACKNOWLEDGEMENTS I would have never accomplished this research in a short time frame without the continuous encouragement and support from my advisor and Advisory and Examining Committee (AEC) Chair, Dr. Udaya B. Halabe. I m very much indebted for his guidance and advice in every step of difficult moments for the completion of this research. I m also thankful to Dr. Hema J. Siriwardane for serving as Co-chair in my committee and providing necessary geotechnical laboratory support and guidance for the completion of this research. I also want to thank both of them for providing an opportunity for me to participate and publish the papers based on this work in the QNDE 2006 Conference, Portland, Oregon. I would also like to express my appreciation to Dr. Hota V. S. GangaRao for serving as an AEC member. I m also grateful to the Department of Civil and Environmental Engineering, West Virginia University for the education and experience I ve received during my MSCE Program. Appreciation is extended to Federal Highway Administration whose support made this research possible. This work was supported through the Federal Highway Administration (USDOT - FHWA) Contract # DTFH61-05-C-00004. Last but not the least; I want to thank my wife, Archana, and my family for their continuous support and encouragement during the course of this research. iii

TABLE OF CONTENTS ABSTRACT... ii ACKNOWLEDGEMENTS... iii LIST OF FIGURES... vii LIST OF TABLES...x CHAPTER 1 - INTRODUCTION... 1 1.1 Problem Statement... 1 1.2 Objectives... 2 1.3 GPR as an NDT Method... 3 1.4 Previous Research Studies... 4 CHAPTER 2 BASIC THEORY... 6 2.1 GPR Principle... 6 2.2 GPR Instrumentation and Sensors... 9 2.2.1 Transmitter and Receiver Antenna... 9 2.2.2 Oscilloscope... 10 2.2.3 Timing Generator... 11 2.2.4 Computer System... 11 2.3 GPR Methods for Measuring Soil Water Content... 11 2.3.1 Using Reflected Waves... 12 2.3.1.1 Common (Single) Offset Method (Som)... 12 2.3.1.2 Multi-Offset Reflection Method (MOF)... 14 2.3.2 Using Ground Wave for Soil Water Content Measurement... 16 iv

2.3.3 Using Borehole GPR Measurements (GPR Tomography)... 19 2.3.4 Using Surface Reflection Method (Air Launched Antenna)... 20 2.4 GPR System Used in this Study... 21 2.5 GPR Resolution... 22 2.6 Sources of Errors and Environmental Effects... 23 2.7 Soil Properties... 24 2.7.1 Silty Soils... 24 2.7.2 Void Ratio... 25 2.7.3 Porosity... 25 2.7.4 Degree of Saturation... 26 2.7.5 Volumetric Water Content... 26 CHAPTER 3 NO FLOW OR NO SEEPAGE SETUP... 27 3.1 Experimental Setup... 27 3.2 GPR Data Acquisition... 29 3.3 GPR Data Processing... 30 3.4 Results and Discussions... 33 CHAPTER 4 TRANSIENT SETUP... 45 4.1 Experimental Setup... 45 4.2 GPR Data Acquisition... 47 4.3 Data Processing... 47 4.4 Results and Discussions... 51 v

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS...61 5.1 Conclusions... 61 5.2 Recommendations for Future Research... 63 REFERENCES... 64 vi

LIST OF FIGURES FIGURE 2.1. Schematic diagram showing the travel path for the GPR emitted EM waves.... 7 FIGURE 2.2. Illustration of a trace from the GPR signal traveling into the soil kept in a box with aluminum foil underneath... 7 FIGURE 2.3. Schematic of a typical GPR system... 10 FIGURE 2.4. Illustration of different types of waves observed in GPR records... 12 FIGURE 2.5.GPR traverse over an anomalous wetter zone above the ground water Table... 13 FIGURE 2.6. A typical survey layout during CMP (top) and WARR (bottom) method.... 15 FIGURE 2.7. WARR measurements with 225 MHz antenna... 16 FIGURE 2.8. Maps illustrating the increase in soil water content from irrigation.. 18 FIGURE 2.9. Comparison between the measured soil water content using Capacitance probe, ZOP and Mop method using borehole GPR... 18 FIGURE 2.10. Schematic of the GPR system used in the research... 22 FIGURE 3.1. Experimental setup showing GPR scanning setup and the Plexiglas box containing soil sample.... 28 FIGURE 3.2. Illustration of a GPR trace for position correction. (a) Trace before position correction, and (b) Trace after position correction.... 31 FIGURE 3.3. Illustration of applying linear gain to the position corrected GPR trace... 31 vii

FIGURE 3.4. (a) Raw radar data, and (b) Radar data after applying the position correction and range-gain..... 32 FIGURE 3.5. Processed radar data illustrating water level rise in the capillary zone... 34 FIGURE 3.6. GPR detected height of water level at different times 37 FIGURE 3.7. GPR detected height of water level at different times in soil type B... 38 FIGURE 3.8. GPR detected height of water level at different times in soil type C... 39 FIGURE 3.9. Saturation and Volumetric water content measurements in soil types A, B and C respectively... 40 FIGURE 3.10. GPR detected height of water level at different times in (a) soil type A and (b) soil type B... 41 FIGURE 3.11. GPR detected height of water level at different time for soil type C... 42 FIGURE 3.12. Saturation and Volumetric water content measurements for the soil type A... 42 FIGURE 3.13. Saturation and Volumetric water content measurements for soil types B and C respectively... 43 FIGURE 4.1. (a) Experimental setup showing GPR scanning setup, (b) View of sample box from different perspective.... 46 Figure 4.2. (a) Raw radar data, and (b) Radar data after applying the position correction and display gain... 49 viii

FIGURE 4.3. Soil sample in Transient boundary conditions showing the water level after few hours of the start of experiment... 50 FIGURE 4.4. Radargram illustrating the raised water level... 50 FIGURE 4.5. Radargram illustrating GPR detected height along the length of the samples at different times in soil A.(a) Before introducing water into the sample (b) After 1 hour, (c) After 3.5 hours and (d) After 22 hours... 54 FIGURE 4.6 GPR detected height along the length of the samples at different times for soil A.... 55 FIGURE 4.7. Radargram illustrating GPR detected height along the length of the samples at different times in soil B. (a) After 5 hours and, (b) After 24 hours.... 56 FIGURE 4.8. GPR detected height along the length of the samples at different times for the soil B... 56 FIGURE 4.9. GPR detected height along the length of the samples at different times in soil B. (a) After 1.5 hours and, (b) After 23.5 hours.... 57 FIGURE 4.10. GPR detected height along the length of the samples at different times in soil B.... 57 FIGURE 4.11. Radargram illustrating GPR detected height along the length of the samples at different times in soil C. (a) After 2.5 hours, (b) After 21.5 hours and (c) After 29 hours... 58 FIGURE 4.12. GPR detected height along the length of the samples at different times in soil C.... 59 ix

LIST OF TABLES Table 2.1. Approximate resolvable limit for the given wavelength... 23 Table 3.1. Degree of Saturation and volumetric water content at the radar detected height in the capillary rise zone for the soil types A, B and C.... 36 Table 4.1. Saturation and volumetric water content at the radar detected height in the capillary rise zone for the soil types A, B and C.... 60 x

CHAPTER 1 INTRODUCTION 1.1 PROBLEM STATEMENT Assessing soil moisture content is an important task in any geotechnical and structural work. Properties such as infiltration, fluid flow, contaminant transport, soil bearing capacity and frost susceptibility are largely dependent on the soil moisture content. Further, knowing the soil physical properties such as compressibility, bearing capacity and frost susceptibility helps for efficient highway and pavement designs especially in areas with expansive clays. These are also important for highways maintenance, monitoring and identification of subsurface drainage for culvert and pavement construction. Transportation engineers use water content of subgrade layer as an indicator to evaluate the pavement conditions and need for any subsurface drainage retrofits. Therefore, knowledge of soil water content has an important significance in design and maintenance of engineering infrastructures as well as in other fields such as agriculture, archaeological, hydrological, geotechnical and geological applications. Remote sensing has been widely used to measure the subsurface water content but the scope of the work is limited by the resolution. This method is only suitable for mapping water content in large and regional areas and is insensitive in smaller areas such as 1 m 2 or less. Similarly, Time Domain Reflectometry (TDR), Capacitance probe and Lysimeter measurements are used for measuring water content but these methods are tedious and provide only point measurements. 1

Therefore, there is a need for an intermediate method (between regional scale and point measurements) that rapidly provides dense and accurate information at an intermediate scale. Ground Penetrating Radar (GPR) method has the capability of measuring soil water content at an intermediate scale. Time domain Ground Penetrating Radar method has been widely used in areas of hydrology and geotechnical engineering in soil-water studies and for applications such as detection of contaminant zone in the groundwater and detecting the ground water table. The rapid data acquisition and non-invasive characteristics have made GPR popular for these applications. However, there has been very few published work relating water saturations to the GPR measurements and the GPR detected water level in silty soils. Therefore, there is a need to address some of the basic issues relating the GPR measurements of water level in silty soil. The detection of water level in fine-grained soils could be complicated due to capillary action. It is important to provide insight into the water level interface detected by GPR as soil-water boundary and the corresponding water saturation level. It was not clear if this interface is somewhere within the capillary rise zone or at the actual water level. The water saturation level at the GPR detected interface is not precisely known. 1.2 OBJECTIVES While past studies focused on GPR methods for measuring soil water content and mapping its variability, this research work is centered on identifying the phreatic surface (free surface of underground water) during seepage through silty 2

soils under transient and steady state (no flow) boundary conditions. A correlation between the degree of saturation/volumetric water content and the water level detected by GPR within the capillary zone was developed in this study. 1.3 GPR AS AN NDT METHOD The NDT techniques provide a means to assess the condition of infrastructure or materials without causing any physical damage. These methods are faster and safer to operate than conventional methods, and provide insight into the materials being tested. There are many NDT techniques being applied in civil engineering disciplines and each of these techniques has its own advantages and limitations. One type of technique may give excellent result for some specific application but may not be good for other applications. Hence, these techniques should be chosen in accordance with the objective of the NDT work. Some of these techniques commonly being applied in civil engineering applications are ultrasonic, infrared thermography, acoustics emission, and ground penetrating radar. GPR method of testing soils and infrastructure has emerged as a powerful, fast and reliable NDT technique over other techniques mentioned above. GPR has an ability to collect data at a rapid rate in contrast to methods such as acoustic emission and ultrasonic which are very slow in terms of data acquisition. Therefore, GPR can be used to survey larger areas in a short time frame. Further with the advancement in the antenna frequency, GPR can provide insight into materials as thin as 1 inch to approximately 200 ft thick which makes GPR a more powerful technique compared to ultrasonic and infrared thermography. Also, GPR has very little dependence on the 3

climatic conditions and hence an area can be surveyed at any time, such as during day or night or in cold as well as dry weather. In comparison, method such as infrared thermography is mostly dependent on the ambient conditions. A major problem associated with GPR is the difficultly in radar data interpretation. The GPR signals are often associated with the environmental noise (from power lines, etc.), multiple reflections and diffractions. This makes it complicated to isolate the reflection events corresponding to the target of interest from the GPR data with additional features (e.g., multiple reflections and diffractions). Also, GPR signal highly attenuates with the moisture content and depth, thus making it difficult for data acquisition and interpretation in those terrains. However, with improvements in the data quality, high signal to noise ratio and advancements in the software for data processing, GPR offers a promising future as an effective NDT technique. 1.4 PREVIOUS RESEARCH STUDIES Ground Penetrating Radar (GPR) has been extensively used in varieties of disciplines for estimating soil water content. Time domain GPR method has been widely used in areas of hydrology and geotechnical engineering in soil-water studies. Numerous articles on mapping and monitoring the soil water content have been published. Davis and Annan (1989) have described in detail about the theory and applications of the GPR methods. Huisman et al. (2005) illustrated various GPR methods for measuring volumetric soil water content. These GPR methods included use of ground wave, reflected wave and amplitude of the reflected waves to delineate 4

the soil water. These methods will be discussed in detail in the following section. Redman et al. (2002) used air-coupled GPR method for measuring subsurface water content using surface reflectivity measurements. Other researchers have used GPR systems to calculate bulk dielectric constant and soil water content (Serbin et. al, 2003 and Lunt et. al., 2005). Similarly, Gloaguen et al. (2001) estimated soil properties such as porosity and aquifer hydraulic conductivity using GPR and hydrostratigraphic data which were well correlated with the laboratory determined values from pumping and tracer tests. Miller et al. (2002) have demonstrated the effects of soil properties on the dielectric and GPR measurements. He also illustrated that the spatial variability in the soil water content will create large variations in the dielectric constant and the GPR response. He also showed that the attenuation in the radar signal increases with the higher soil water content. As indicated above, a large number of literatures can be cited where the GPR technology has been used for assessing soil water content and detecting the water table. However, there has been very few published work relating the saturation level in soils to GPR measurements involving the detection of water level in soils. Specifically, it is not clear if the interface detected by GPR is the actual water level or an interface somewhere within the capillary rise zone. Investigation of this issue is the focus of this research. In this study, the interface location obtained from GPR measurement has been correlated with the degree of saturation and soil water content at the same depth. 5

CHAPTER 2 BASIC THEORY 2.1 GPR PRINCIPLE GPR is a non-invasive technique that has been extensively used for locating and identifying the subsurface features. GPR works in a manner similar to the seismic method, since both methods rely on the reflection coefficient of the interface separating the two mediums and record the reflected waves from this interface. GPR is an electromagnetic wave based technique and the GPR system consists of source and receiver antennas. The electromagnetic waves are emitted by the source antenna and the receiver antenna records the reflected wave from an interface. Whenever the electromagnetic waves are emitted, portion of the waves is reflected back from the surface where it encounters sufficient changes in the dielectric properties of the materials (Figure 2.1). The amount of the reflected energy from an interface, where change in dielectric constant is encountered, can be expressed in terms of the reflection coefficient R 12 as (Halabe et al., 1993), ε ' 1 2 R 12 = (2.1) ε ' 1 + ε ' ε ' 2 where ε is the dielectric constant; subscripts 1 and 2 denote the dielectric constant of the medium above and below the interface. Therefore, as the contrast between the two media above and below the interface increases, more energy will be 6

Plexiglas Sample Box Air 1.5 GHz Monostatic Antenna Tx : Transmitter, Rx: Receiver Tx Rx Emitted Electromagnetic Wave Illumination Zone Soil Water Level FIGURE 2.1. Schematic diagram showing the travel path for the GPR emitted EM waves. Air A D Soil B Aluminum foil placed at the bottom of the soil sample A B FIGURE 2.2. Illustration of a trace from the GPR signal traveling into the soil kept in a box with aluminum foil underneath. 7

reflected, thus resulting in high amplitude in the signal. Figure 2.2 illustrates a typical GPR trace passing through the air soil aluminum foil medium. The aluminum foil was placed underneath the soil sample to obtain a sharp reflection for use in computing the average dielectric constant through the soil. As the wave strikes at the air-soil boundary, changes in the dielectric of these two mediums will create a high reflection coefficient. Since the amplitude is proportional to the reflection coefficient of the electromagnetic waves, high amplitude in the recorded signal can be interpreted as the air-soil boundary. Similarly, as the wave passes through the soil and encounters the aluminum foil, changes in the dielectric will lead to higher amplitude in the recorded signal trace. The air-soil and soil-aluminum boundaries are represented by A-A and B-B, respectively, in Figure 2.2. The travel time associated with the reflected waves in the soil is measured and used to compute the average dielectric constant of the medium through which the wave travels. These reflected waves are recorded, amplified and processed to gain insight into the material it traveled. The depth (D) to the interface is given by (Halabe, 1995), D = c t 2 ε '........(2.2) where ε' is the dielectric constant of the medium; D is the depth of the target of interest (detected water level, or soil sample s bottom in case of dry soil); c is the velocity of EM wave in vacuum = 3 x 10 8 m/s; and t is the roundtrip time taken by the EM pulse to reach the receiver antenna. 8

2.2 GPR INSTRUMENTATION AND SENSORS In general, the GPR equipment consists of a transmitter and receiver antenna with oscilloscope, tape recorder and computer hardware. The oscilloscope and tape recorder are normally associated with old GPR system whereas the GPR system after 1990 used the computer as both display and storage unit. A typical radar measurement system and setup is shown in Figure 2.3. The largest three manufactures for the GPR systems are the Geophysical Survey Systems, Mala Geosciences and Sensors and Software (Huisman et al., 2005). The available GPR systems typically use antennas with frequencies ranging from 10 MHz to 2.0 GHz. 2.2.1 Transmitter and Receiver Antenna In market, both ground coupled and air coupled antennas are available. The air coupled antenna can be suspended few feet above the ground and easily mounted on a vehicle which makes data acquisition easy for a large survey area. On the other hand, ground coupled antenna has to be in physical contact with the ground surface. In terms of sensitivity, ground coupled antenna offers more sensitivity and high signal penetration depth when compared with the air launched antenna because in air launched antenna a significant portion of the emitted electromagnetic energy is reflected off of the air-ground interface before it actually penetrates the ground. These antennas operate at frequency ranging from 10 MHz to 2 GHz. For most of the geotechnical works, lower frequency antennas are preferred since the signal penetrates to greater depth where as high frequency antennas are preferred for structural works such as for concrete and timber bridge decks because high frequency leads to better resolution of defects. It is also preferable to use shielded 9

antennas in the areas with high environmental noise (coming from power lines, etc.) that might interfere with the GPR received signal. The main purpose of the receiver antenna is to record the signal emitted from the transmitter. The receiver records the reflected, refracted, transmitted and direct wave arrivals as they come at different times to produce a time-domain waveform. In some cases the receiver and transmitter antennas are separable and in other cases they are housed inside a single unit. Battery Charger Central Panel Unit Central display Unit Transmitter/ Receiver Unit These units are housed into single unit on the modern GPR system Antenna system Oscilloscope Computer System FIGURE 2.3. Schematic of a typical GPR system (Halabe et al. 1995). 2.2.2 Oscilloscope The oscilloscope is used to visualize the waveform generated from reflection from the various interfaces in the medium. These waveforms can be digitized using a digitizing board. A GPR system, usually after 1990s, uses the computer for display purposes. 10

2.2.3 Timing Generator The main function of the timing generator is to produce short pulses of electromagnetic energy at short intervals, usually with duration of about 1.5 to 2 nanoseconds and spacing between adjacent pulses being of the order of 1 microsecond. 2.2.4 Computer System The recorded data can be processed and analyzed using a computer system with appropriate software. The software is used for filtering and processing the acquired radar data set. Often digital media with high storage capacity are used for storing the data. 2.3 GPR METHODS FOR MEASURING SOIL WATER CONTENT All the GPR methods for soil moisture mapping are based on the dielectric constant of the medium which is estimated using the travel time and/or amplitude of the emitted electromagnetic waves. Figure 2.4 represents a typical recorded GPR time-domain scan obtained using two antennas (a transmitter and a receiver). In the figure, a variety of wave events such as reflected waves, direct waves, and ground waves are observable. All the GPR methods are based on interpreting arrival time of some or all of these waves. 11

FIGURE 2.4.. Illustration of different types of waves observed in GPR records (Huisman et al., 2005). 2.3.1 Using Reflected Waves This method is based on the arrival time of the electromagnetic waves reflected from an interface with dielectric impedance contrast. In general, unsaturated and saturated soils have different dielectric values and the EM waves are reflected whenever a contrast between two dissimilar media is encountered. There are two methods for data acquisition using reflected waves. 2.3.1.1 Common (Single) Offset Method (SOM) In this method, the source and the receiver antennas are moved along the profile with a constant offset. The EM waves passing through the soil are partly reflected whenever changes in the dielectric properties are encountered. Figure 2.5(a) represents an ideal method of GPR data acquisition using SOM and Figure 2.5(b) presents a typical recorded signal. In Figure 2.5(b) it is clear that the reflected waves trace out a hyperbola with apex over the anomaly. This is the region with high dielectric constant. The shape of the hyperbola is due to the fact that the reflected wave takes shorter time to reach the receiver when the antenna set up is just above 12

the anomaly. The average velocity for the soil can be determined using the following equation for the hyperbola, 2 2 2 x + d v =..(2.3) t Where, x = horizontal distance of the antenna set up from the apex of the hyperbola, d = depth of the anomaly, and t = arrival time of the reflected waves at distance x. This time has to be zero corrected. Huisman et al. (2005) have given the velocity equation whenever there is significant antenna separation as, 2 2 2 2 ( x 0.5a) + d + ( x + 0.5a) + d v = (2.4) t The velocity is then used to back calculate soil permittivity and interpret the water content in the soil. (a) (b) FIGURE 2.5(a).GPR traverse over an anomalous wetter zone above the ground water Table using common antenna offset. FIGURE 2.5(b).Ideal recorded signal from the traverse. A, B and C represents reflections from airwave, anomalous wet zone and ground water table (Husiman et al., 2005). Although this method offers rapid method of data acquisition it has certain limitations in terms of measurement of soil water content. This method only gives the average wave velocity and hence only the average soil water content up to the 13

depth of the reflector (Huisman et al., 2005). It is also important to know the depth of the reflecting horizon for accurate velocity estimation. Vellidis et al. (1990) used this method by calculating the average velocity from the reflected waves from buried pipes above the wetting front. He assumed that the soil water content was homogenous above the wetting front and was successful in monitoring the wetting front movement. Similarly, Grote et al. (2002) used shallow reflection times to estimate soil water content values within a constructed sandy test pit and obtained good correlation of 0.83 between lysimeter measurements and GPR measurement using antenna frequency of 1200 MHz. 2.3.1.2 Multi-Offset Reflection Method (MOF) In contrast to the single offset measurements, Multi-offset reflection method provides a means to estimate soil water content without requiring the knowledge of the depth for the reflecting horizon. This method can be utilized either using CMP (Common Mid Point) or WARR (Wide Angle Reflection and Refraction). In CMP method, the source and the receive distance are increased at steps maintaining a common mid point between them. In WARR system, the transmitter antenna is fixed at one location and the receiver distance is increased stepwise (Huisman et al., 2005). Figure 2.6 shows the schematic chart showing the layout of these two methods. Figure 2.7 shows a typical WARR radargram. The velocity of the reflected wave in the MOF method is calculated as, 2 2 2 d + (0.5a) v =.(2.5) t 14

The above equation gives the average velocity up to the depth of the reflector. This velocity is converted into interval velocity (v int ) of each soil layer using the Dix equation (from Huisman et al., 2005), v int = t. v t. v t t 2 2 n n n 1 n 1 n n 1.. (2.6) Where, v n = average velocity from the surface to the bottom of the n th layer, and v n-1 = average velocity from the surface to bottom of the (n-1) th layer, and t n and t n-1 represent the corresponding two-way time from the surface to the bottom of the layers. FIGURE 2.6. A typical survey layout during CMP (top) and WARR (bottom) method. S represents source antenna and R represents the receiver antenna (Huisman et al., 2005). 15

FIGURE 2.7. WARR measurements with 225 MHz antenna, v is the velocity of ground wave and c is the velocity of the air wave (Huisman et al., 2005). 2.3.2 Using Ground Wave for Soil Water Content Measurement The ground wave is a part of the emitted EM signal which travels directly from the source to the receiver through the top of the soil. In this method, there is no reflection event and the direct arrival time of the ground wave from the source to the receiver is recorded. Figure 2.7 represents ground wave and the evident slope of the ground wave can be directly used to calculate the ground wave velocity which is then used to estimate the soil water content. Both SOM and MOF are used for the ground wave arrival time analysis. MOF has advantage over SOM because there is no need to know the absolute direct ground wave arrival time and the velocity can be calculated easily using the slope of the time versus offset. A major drawback associated with MOF is its low spatial resolution and the method is tedious and time consuming. In contrast, SOM offers higher spatial resolution using shorter antenna offset and larger areas can be surveyed within a short time. However, the main problem associated with SOM 16

method is the problem and accuracy of identifying the leading edge of the ground wave. Huisman et al. (2005) clearly showed the accuracy of the ground wave method to map soil water content and compared the results with TDR measurements. As observed in Figure 2.8, there is a good relationship between the TDR and GPR data. Galagedara et al. (2003) used SOM and WARR with ground wave to measure soil water content using 100 MHz and 450 MHz antenna and compared the results with the TDR measurements. They also demonstrated that the low frequency antenna may underestimate the water content due to the influence from the deeper zones associated with the longer wave length. Therefore the use of ground wave is restricted to the uppermost layer in the soil. The influence depth of the ground wave has been shown to be nearly equal to 0.145 λ by Sperl 1999 (from Huisman et al. 2005). Thus, the depth ranges from 0.15m (ε = 4) to 0.1 m (ε = 20) for 224 MHz antenna. Sperl 1999 (from Huisman et al. 2005) also proposed a relationship to compute dielectric permittivity (ε ) from ground wave as, ( t ) 2 gw ) x c t aw + ε =.. (2.6) x Where, x is the antenna separation distance, t gw is the ground wave arrival time, t aw is the air wave arrival time, and c is the velocity of electromagnetic wave in air. Similarly, Grote et al. (2005) used 450 MHz and 900 MHz frequency with common offset GPR ground wave measurements and had highest correlation with water content values averaged between 0-0.2m range and least with 0.1-0.2m interval. 17

FIGURE 2.8. Maps illustrating the increase in soil water content (m3 m -3 ) from irrigation. Left map represents results obtained using GPR and right map presents the results from TDR (Huisman et al., 2005). FIGURE 2.9. Comparison between the measured soil water content using Capacitance probe, ZOP and MOP method using borehole GPR (Huisman et al., 2005). 18

2.3.3 Using Borehole GPR Measurements (GPR Tomography) In this method, the transmitter and receiver antennas are lowered into two boreholes and the arrival times of the direct wave is used to calculate the soil permittivity and the velocity. When the transmitter and receiver antenna are lowered together in boreholes maintaining same depth from their mid point the method is known as Zero offset Profile (ZOP). In Multi-Offset Profile measurement (MOP), the source is placed at one depth and receiver antenna depth is varied. Again the depth of the transmitter antenna is increased and the arrival times of the direct waves from varying receiver antenna depth is measured. The first arrival times of the direct waves in MOP are used to construct Tomographic Image of soil water content between the boreholes. The 2D-Tomogram is constructed by converting area between the boreholes into finite grids with cells of constant velocity. The velocity for each cell is calculated by minimizing the differences between the measured arrival times and the arrival times calculated for ray paths passing through the cells (Huisman et al. 2005). Alumbaugh et al. (2002) compared the volumetric water content measured from the GPR Tomogram and from the neutron log derived values and obtained a good match with the results obtained from the neutron log measurements. The comparison between the water contents derived from MOP and the capacitance probe measurements is shown in Figure 2.9 and a good correlation is seen between those two measurements. 19

2.3.4 Using Surface Reflection Method (Air Launched Antenna) Unlike the previous methods where both receiver and transmitter antennas are ground coupled, this method uses air launched receiver and transmitter antennas. The antennas are kept at some distance above the ground which enables rapid data acquisition which is very useful in surveying large areas. The method is based on the principle that the amplitude of the reflected wave is proportional to the portion of energy that is reflected. Therefore, high amplitude indicates large contrast in dielectric constant between the subsequent layers. For example, wet layer underlying dry layer will have high reflection amplitude in comparison to both the layers being either dry or wet. The soil property measured using air launched reflection method is the reflection coefficient which is determined from the measured amplitude of the reflected wave traces. The reflection coefficient (R.C) is given as Ulriksen (1982), RC. = K K 1 + K K 2 1 2 (2.7) Where, K 1 = dielectric constant for the medium 1, K 2 = dielectric constant for the medium 2. For the air-soil interface, the dielectric constant of the air is 1. Therefore, the Equation (2.7) reduces to, RC. 1 K 2 = 1 + K2.. (2.8) The reflection coefficient is determined by dividing the measured reflection amplitude, A r by the reflection amplitude from a large metal plate (a perfect 20

reflector), A m. Thus Equation (2.8) leads to the following equation for obtaining the soil dielectric constant. 2 1+ R K soil = 1 R.. (2.9) Where, K soil = dielectric constant of the soil and R = A r / A m According to Redman et al. (2002), GPR systems with broadband source gives a good reflective measurement and high frequency antennas are preferred. This study states that errors in air launched antenna method are mostly associated with amplitude accuracy. Further, reflection coefficient is also lowered by surface roughness and varying soil water content profiles with depth. 2.4 GPR SYSTEM USED IN THIS STUDY The GPR system used in this study consisted of a main frame unit connected to a portable laptop computer (Figure 2.10). The laptop controls the GPR data acquisition and processing. GSSI SIR-20 unit was used as the main frame data acquisition system and RADAN 5.0 software was used for data processing and analysis. The mainframe system was connected to the ground coupled monostatic antenna for transmitting and receiving the EM signals. Two types of ground coupled antennas were used in this study, one operating at 900 MHz central frequency and the other at 1500 MHz (i.e., 1.5 GHz). A signal transmit rate of 100 KHz was set in the GPR mainframe setup. GPR scans were taken in a free run mode and survey wheel mode for the steady state and transient boundary conditions, respectively. In a free run mode the scans were taken at 75 21

scans/sec where as 30 scans/inch was used in the survey wheel mode. A record length of 512 points was used to represent each 15 ns signal. Figure 2 shows the schematic of the entire GPR system. Input Power Source (12V DC) Display & Data Processing (Portable Laptop) Mainframe (SIR-20) Transmitter / Receiver Antennas (1.5GHz & 900 MHz) FIGURE 2.10. Schematic of the GPR system used in the research. 2.5 GPR RESOLUTION An evaluation of GPR data resolution limits addresses basic questions concerning the ability of the radar dataset to resolve the top and bottom of a layer as distinctly separate reflections. Resolution limits are specified in terms of the minimum vertical and horizontal dimensions of an object that can be resolved. The vertical resolution limit defines the minimum thickness of a layer for which the reflection from the top is distinctly separate from the reflection from the base. The vertical resolution limit depends upon the wavelength of the electromagnetic wave and dielectric impedance contrast within the bounding layers (Sheriff, 1977). For this work, vertical resolution limit was approximated as a quarter of a wavelength obtained using the computed average dielectric constant of the soil (~ 2.5) and is presented in Table 2.1. 22

Table 2.1. Approximate resolvable limit for the given wavelength Frequency Wavelength in air Resolution in soil 1.5 GHz 20.0 cm = 7.87 inch 7.87/(4*sqrt(2.5)) = 1.244 inch 900 MHz 33.3 cm = 13.11 inch 13.11/(4*sqrt(2.5)) = 2.07 inch The resolution of the radar data is dependent on the frequency of the antenna used. Higher frequency antennas offer higher resolution but are often limited by lower depth penetrating capacity. Since the radar data attenuates with depth, there is a trade-off between the higher frequency antenna and the penetration depth. In this research, 1.5 GHz and 900 MHz frequency antennas were used for comparative study. The 1.5 GHz frequency is the highest commercially available frequency in the ground coupled antenna series. 2.6 SOURCES OF ERRORS AND ENVIRONMENTAL EFFECTS Like other geophysical NDT systems, GPR system is also not error free. The errors may occur during the data acquisition stage or later during the data processing and interpretation stage. During the GPR survey, environmental noise is the major contributing factor introducing errors in the recorded data. The sources of environmental noise may be from nearby power lines, interference from any electromagnetic devices or from other associated conductive features. Such noise affects air launched antennas much more than the ground coupled antennas. Depth and Resolution is another important aspect during the GPR measurements. The user should have tentative idea of the target depth and type of information that is needed, and accordingly choose the appropriate antenna 23

frequency. Using low frequency antenna for shallower target depth may not resolve near surface features and using high frequency antenna signals may not provide the required penetration depth to resolve deeper targets. There may also be errors during the interpretation stage. For example, the volumetric water content calculation are formulated based on particular type of soil, however in nature we may not get uniform soil composition in all the places. Secondly, certain soils contain higher proportion of magnetic minerals which can adversely affect the electromagnetic signal and penetration depth. 2.7 SOIL PROPERTIES The dielectric properties of soil that affect radar measurements are water content, texture, salinity and temperature (Serbin et. al., 2003). Coarse and low surface area soils such as coarse sand contains little bound water and has low electrical conductivity in contrast to fine grained soils like silt that has high electrical conductivity. Reflection coefficient at the interface between two media is proportional to the contrast in dielectric constant and electrical conductivity: however high electrical conductivity also increases the attenuation of the radar signal and the signals penetrate only to smaller depths (Serbin et. al, 2003). Important soil properties that are used for computing the soil water content and saturation are described in the following section. 2.7.1 Silty Soils These are fine grained soils with higher silt content. In general silty soils are fine grained soils with grain size ranging from 0.05mm to 0.002 mm (AASHTO 24

classification). Three different silty soils, Types A, B and C were engineered in the lab to have constant porosity throughout the volume. Soil A, B and C had clay content of 5%, 15% and 25%, respectively (Siriwardane et al. 2006). 2.7.2 VOID RATIO Void ratio is the ratio of the volume of the voids to the total volume of the solids in a given soil sample. In general, void ratio is given as, V e = V v s.. (2.10) Where, V v = volume of voids, V s = total volume of soil solids. 2.7.3 POROSITY Porosity is defined as the volume of voids to the total volume of the soil sample. It can be written as, Vv n =..(2.11) V Where, V = total volume of the soil sample Porosity (n) can be related to the void ratio (e) as, Vv n = = V Vv V + V s v = e 1 + e...(2.12) 25

2.7.4 Degree of Saturation Degree of saturation is defined as the ratio of volume of water in the voids spaces of the soil to the total volume of the voids. It is usually expressed in terms of percentage as, Where, V w = volume of water in the soil. Vw S (%) = 100....(2.13) V v 2.7.5 Volumetric Water Content It is the product of the porosity and saturation of the soil and usually expressed in terms of percentage. It is given as, V. W (%) = n S(%)...(2.14) 26

CHAPTER 3 NO FLOW OR NO SEEPAGE SETUP In this case, the experiments were designed to test the soil specimen in a fixed no flow or no seepage boundary condition. In case of no flow or no seepage boundary condition, the height of the water introduced in the soil was fixed on both sides of the soil sample box and the vertical capillary rise of water was determined. 3.1 EXPERIMENTAL SETUP Radar scans were taken over soil sample housed in a Plexiglas box of dimension 62 cm x 48 cm x 46 cm (24 x 19 x 18 ) as shown in Figure 3.1. The Plexiglas box was built as a part of a previous study (Siriwardane et al. 2006) and was also used in this study. Use of Plexiglas avoided interference in the radar signal and allowed easy observation of the water level in the soil sample. The soil used in the experiment was silty sand (a mixture of sand and clay), engineered in the laboratory to have a constant porosity throughout the sample (soil samples were provided by Siriwardane et al. 2006). Steady state seepage boundary condition was maintained by supplying water from both ends of the box into the sample. The water level was maintained at a constant height of 3 inches on both sides of the sample box (Figure 3.1). Aluminum foil was placed at the bottom of the soil chamber to get a clear radar reflection signal. GPR readings were taken over the sample at different times over 40 hours. At the end of the experiment, small soil samples were collected at different locations and heights to experimentally determine the soil moisture 27

content and degree of saturation. Three different silty soil types, A (5% clay content), B (15% clay content) and C (25% clay content) were tested. Each of the experiments was repeated for statistical accuracy. (a) FIGURE 3.1. Experimental setup showing GPR scanning setup and the Plexiglas box containing soil sample. (b) 28

3.2 GPR DATA ACQUISITION In this setup, the data was collected in free run mode. In free run mode, antenna is kept stationary and GPR data is acquired based on the scans per second rate set in the header file, which then controls the number of scans per second emitted by the antenna. The antenna was placed over the dry soil surface and the scans were collected at different times to monitor the rise in water level due to capillary action. The main frame unit was connected to a portable laptop computer that controls the GPR data acquisition and processing. The ground coupled monostatic antenna used for transmitting and receiving the EM signals was connected to the mainframe system. The same antenna served as the transmitter and receiver. Two types of ground coupled antennas were used for the no flow boundary condition. First one was operating at 900 MHz central frequency and the other was at 1500 MHz (i.e., 1.5 GHz). The 900 MHz antenna offers higher depth penetration but lower resolution while the 1.5 GHz antenna provides higher resolution but lower depth penetration. As mentioned in the theory section, there is always a trade off between the penetration depth and signal quality. A signal transmit rate of 100 KHz was set in the GPR mainframe setup. GPR scans were taken in a free run mode at 75 scans/sec. A record length of 512 points was used to represent each 15 ns signal. 29

3.3 GPR DATA PROCESSING The radar data were processed and analyzed using the RADAN 5.0 software. The data processing was fairly simple and straight forward since the objective was only to determine the height of the water level over time in the capillary rise zone. The data were corrected for the position correction and range-gain was applied for visual enhancements. Position Correction is a procedure for setting the zero time position relative to the soil top surface. It shifts and aligns each trace of the radar signal to the first contact point of the antenna with the soil. Figure 3.2(a) and 3.2(b) shows the process of position correction. The top of the soil surface is represented by high amplitude echo at the start of the GPR trace in Figure 3.2(a). This high amplitude echo is moved to zero position to align it with the actual soil surface as illustrated in the Figure 3.2(b). Range Gain is a procedure to enhance the signal quality by compensating for the attenuation of radar signal with depth. Based on the user specified value, the data are gained (amplified) which enhances the signal quality. Figure 3.3 illustrates this process. The red lines below the GPR trace show the relative values of the applied gain at various points in the radar data. The gain is piecewise linear between the two consecutive points. Figures 3.4(a) and 3.4(b) illustrates actual recorded GPR radargram data, before and after the position correction. In the radargram shown in Figure 3.4(a), top of the soil surface is represented by white-black-white bands. The maximum 30

(a) (b) FIGURE 3.2. Illustration of a GPR trace for position correction. (a) Trace before position correction, and (b) Trace after position correction. FIGURE 3.3. Illustration of applying linear gain to the position corrected GPR trace. The linear line below the trace shows the magnitude of the applied gain. 31

FIGURE 3.4. (a) Raw radar data, and (b) Radar data after applying the position correction and range-gain. The top of the soil surface, represented by white-blackwhite band has been shifted to zero position in the position corrected data. 32

amplitude of the first white band around 4.5 inches corresponds to the top of the soil surface. Therefore the radar data was shifted upward by 4.5 inches in order to align the maximum positive amplitude of the first white band to the zero position as illustrated in the Figure 3.4(b). This corresponds to the first contact point of antenna with the soil. The white-black-white bands in the processed data in Figure 3.4(b) are much more distinct than the unprocessed data in the Figure 3.4(a). This shows the importance of gaining the radar data for better visualization. 3.4 RESULTS AND DISCUSSIONS The radar data acquired over successive periods of time showed a gradual rise in water level with time. An upward shift in the white-black-white bands in the middle portion of the radargram (Figures 3.5 (a), (b) and (c)) is observed over time, which illustrated the upward rise in the water level as detected by GPR. This phenomenon can be explained in terms of the dielectric impedance contrast discussed earlier under GPR theory. The dielectric constant for air and water is 1 and 80 respectively (Davis et. al. (1989). The dielectric constant of dry soil was around 2.5 which is much lower than that of the moist soil in the capillary zone. This creates an interface between dry and moist soil, which is detected by GPR. As a result, a distinct white-black-white band appears in the radargram somewhere between the top and bottom of the soil sample indicating the presence of water in the capillary rise zone. The horizontal interface around 10 inches depth from the top in Figure 3.5(b) marks the water level in the capillary rise zone. When the water starts to rise in the soil due to capillary action, this interface also rises (Figure 3.5(c)). 33

T B (a) T W B (b) T W (c) B FIGURE 3.5. Processed radar data illustrating water level rise in the capillary zone. T: top of the soil surface, B: Bottom of the soil surface and W: Detected water level. (a) Before the introduction of water into the sample, (b) 14.5 hours after the introduction of water and (c) 25 hours after the introduction water into the sample. 34

The bottom of the soil layer appears to be at a higher depth than the actual depth in the radargrams. As observed in the radargram in Figure 3.5(b), the apparent depth of the soil bottom is at approximately 20 inches whereas the actual depth of the soil was at 15 inches. The downward shift in depth of the bottom of the soil layer can be explained in terms of the changes in the dielectric constant of the soil layer with the gradual rise in the water level through capillary action. As water starts to rise due to capillary action and causes an increase in the degree of saturation of the silty soil, its dielectric constant increases which slows down the traveling electromagnetic waves in this moist region. As a result, the depth of the soil bottom appears to increase in the radargram. The actual height of the water level can be obtained by subtracting the depth of the GPR predicted water level (that is, the interface between dry and moist soil) from the total soil height. This actual height detected by radar over time for the different soil types is illustrated in the various graphs presented in Figures 3.6, 3.7, and 3.8, respectively. The height of the input water level was fixed at 3 inches throughout the experiments but over a period of time the water level within the soil detected by GPR reached a height of about 11 inches. This indicated that GPR actually detected the water level in the capillary rise zone. Clearly, GPR was detecting the interface with saturation level far below 100 percent. At the end of the GPR experiments, small samples of soil at various locations and heights were extracted for laboratory measurement of porosity and water saturation as per ASTM standard procedure (ASTM 1999). The water content data 35

for the respective soil samples were collected and provided by Siriwardane, Ingram and Kiriakidis (2006). A plot of saturation and volumetric water content is presented in Figure 3.9. The plots shown in Figures 3.10 to 3.13 are results from the repeated experiments, which showed a fair degree of consistency in the results. The degree of saturation at the height of water level detected by radar was computed by interpolation of the actual water saturation data obtained at different locations and heights within the sample. This saturation value, obtained at the radar detected height in the capillary rise zone, is tabulated in Table 3.1. The saturation level at the radar detected boundary typically varied between 20 and 40 percent. The results showed that radar is detecting water level in soil, with degree of saturation as low as 19 percent and up to a maximum of 50 percent, as a distinct water level boundary within the soil. The corresponding volumetric water content detected by the radar typically fell in the range of 9 to 20 percent. Table 3.1. Degree of Saturation and volumetric water content at the radar detected height in the capillary rise zone for the soil types A, B and C. Soil Clay type content Height of water level detected by GPR (inch) Using 1.5 GHz 1 Degree of Saturation (%) 1 Volumetric water content (%) Total time (hrs) Height of water level detected by GPR (inch) Using 900 MHz 1 Degree of Saturation (%) 1 Volumetric water content (%) Total time (hrs) 5% 8.2 23.5 11.29 23.63 A 5% 6.2 50 24.01 5.83 6 53 25.45 5.83 15% 11.66 21 10.00 96.83 9.85 29 13.81 96.83 B 15% 7.4 41 19.69 22.35 25% 11.8 18.8 9.72 105 C 25% 9.2 19 9.12 52.23 8 32 15.37 52.23 1. From Siriwardane, Ingram and Kiriakidis (2006) 36

Water level rise detected by GPR using 1.5 GHz antenna Soil type A (5% clay) Height (inch) 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 CZ SZ 0 1 2 3 4 Time interval (hours) 5 6 7 (a) Water level rise detected by GPR using 900 MHz antenna Soil type A (5% clay) 15 14 13 12 11 10 Height (inch) 9 8 7 6 5 4 3 2 1 0 0 1 2 CZ SZ 3 4 5 6 7 Time interval (hours) (b) FIGURE 3.6. GPR detected height of water level at different times in soil A. The solid dashed line denotes the water level fixed at 3 inches height. The subscript CZ denotes capillary rise zone and SZ denotes water saturated zone. (a) Using 900 Mhz antenna (b) Using 1.5 GHz antenna. 37

CZ SZ (a) Water level rise detected by GPR using 1.5 GHz antenna Soil type B (15% clay) 14 12 10 Height (inch) 8 6 4 CZ 2 SZ 0 0 5 10 15 20 25 30 35 40 45 50 55 60 Time interval (hours) 65 70 75 80 85 90 95 100 (b) FIGURE 3.7. GPR detected height of water level at different times in soil B. The solid dashed line denotes the water level fixed at 3 inches height. The subscript CZ denotes capillary rise zone and SZ denotes water saturated zone. (a) Using 900 MHz antenna and (b) Using 1.5 GHz antenna. 38

14 Water level rise detected by GPR using 900 MHz antenna Soil type C (25% clay) 12 10 Height (inch) 8 6 4 CZ 2 SZ 0 0 5 10 15 20 25 30 35 40 45 50 55 Time interval (hours) (a) 14 Water level rise detected by GPR using 1.5GHz antenna Soil type C (25% clay) 12 10 Height (inch) 8 6 4 2 CZ SZ 0 0 5 10 15 20 25 30 35 Time interval (hours) 40 45 50 55 (b) FIGURE 3.8. GPR detected height of water level at different times in soil C. The solid dashed line denotes the water level fixed at 3 inches height. The subscript CZ denotes capillary rise zone and SZ denotes water saturated zone. (a) Using 900 MHz antenna and (b) Using 1.5 GHz antenna. 39

Soil type A 14 Avg. Saturation Avg. Volumetric water content 12 10 Height (inch) 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Saturation(%) 14 Soil type B Avg. Saturation Avg. Volumetric water content 12 10 Height (inch) 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Saturation (%) Soil type C 14 Avg. Saturation Avg. Volumetric water content 12 10 Height (inch) 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Saturation (%) FIGURE 3.9. Saturation and Volumetric water content measurements for soil types A, B and C, respectively. 40

14 Water level Rise detected by GPR using 1.5 GHz antenna Soil type A ( 5% clay) 12 10 Height (inch) 8 6 4 2 0 0 5 10 15 Time interval (hours) 20 25 (a) 14 Water level Rise detected by GPR using 1.5 GHz antenna Soil type B (15% clay) 12 Height (inch) 10 8 6 4 2 0 0 5 10 15 Time interval (hours) 20 25 (b) FIGURE 3.10. GPR detected height of water level at different times using 1.5 GHz antenna. The solid dashed line denotes the water level fixed at 3 inches height. The subscript CZ denotes capillary rise zone and SZ denotes water saturated zone. (a) Soil Type A, and (b) Soil type B. 41

14 Water level rise detected by GPR Soil Type C (25% clay) 12 10 Height (inch) 8 6 4 2 0 0 10 20 30 40 50 60 Time interval (hours) FIGURE 3.11. GPR detected height of water level at different time for soil type C using 1.5 GHz antenna. The solid dashed line denotes the water level fixed at 3 inches height. The subscript CZ denotes capillary rise zone and SZ denotes water saturated zone. 70 80 90 100 110 Soil type A 14 12 Avg. Saturation Avg. Volumetric Saturation Height (inch) 10 8 6 4 2 0 0 10 20 30 Saturation (%) FIGURE 3.12. Saturation and Volumetric water content measurements for the soil type A. 40 50 60 70 42

Soil type B 14 Avg. Saturation Avg. Volumetric water content 12 10 Height (inch) 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Saturation(%) 14 12 Soil type C Avg. Saturation Avg. Volumetric water content 10 Height (inch) 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 Saturation (%) FIGURE 3.13. Saturation and Volumetric water content measurements for soil types B and C, respectively. 43

A comparison between the heights of water level detected by 1.5 GHz and 900 MHz antennas is also shown in Table 3.1. As expected, the higher frequency 1.5 GHz antenna was able to provide higher resolution of detection. The height of water level (above the soil bottom) detected by the 1.5 GHz antenna was about one inch higher than the height detected by 900 MHz antenna. This means the depth of the detected water level, as measured from the top surface of the soil, is lower for the 1.5 GHz antenna. The corresponding saturation level and volumetric water content for 1.5 GHz antenna is also lower than for the 900 MHz antenna, which indicates that the 1.5 GHz antenna has detected a lower saturation zone due to its higher sensitivity and resolution capability. Also, both GPR antennas were not able to detect capillary rise when it reached near the surface of the soil due to superposition of the two reflection echoes from the soil surface and the capillary interface. 44

CHAPTER 4 TRANSIENT SETUP In this case, the experiments were designed to test the soil specimen under transient boundary conditions. In the transient boundary condition, water in the soil sample is allowed to freely flow from one end to the other end. Water is introduced from one end and allowed to exit from the other end of the sample while keeping the height of water at the inlet side fixed. 4.1 EXPERIMENTAL SETUP Radar scans were taken over soil housed in a Plexiglas box of dimension 30 cm x 25 cm x 120 cm (12 x 10 x 48 ) as shown in Figure 4.1. The Plexiglas box was built as a part of a previous study (Kiriakidis 2006) and was also used in this study. The Plexiglas was used to avoid interference in the radar signal and observe the water level in the soil sample. The soil used in the experiment was silty sand, engineered in the lab to have a constant porosity throughout the sample (soil samples were provided by Siriwardane et al. 2006). Transient seepage boundary conditions were generated by allowing the water to enter the sample from one end and exit from the other end of the sample. The water level at the entry point was maintained at a constant height of 4 inches. Aluminum foil was placed at the bottom of the soil sample to get a clear reflection of radar signals. GPR readings were taken along the 45

(a) (b) FIGURE 4.1. (a) Experimental setup showing GPR scanning setup, (b) View of sample box from different perspective. Water level maintained at height of 4 at the inlet side. 46

length of the sample at different times over 36 hours. At the end of the experiment, small samples were taken at different locations of the soil bed to compute the degree of saturation along the length and height of the sample. Three different soil types, A (5% clay content), B (15% clay content) and C (25% clay content) were tested and some of the experiments were repeated for statistical accuracy. 4.2 GPR DATA ACQUISITION In this setup, the GPR data were acquired using the survey wheel. The forward movement of the survey wheel controlled the data acquisition and allowed scanning at precise distance intervals. GSSI SIR-20 unit was used as the main frame data acquisition system. Scans were taken along the length of the sample as the wheel moved forward. Monostatic ground coupled antenna of frequency 1.5 GHz was used for transmitting and receiving the EM signals. The signal transmit rate was set at 100 khz. The antenna was placed over the dry soil surface and the scans were collected over a period of time to capture the gradual upward rise of water in the soil. For the transient setup using survey wheel, a 512 point signal was used to represent 12 ns range and the survey wheel was calibrated to collect the data at 30 scans/inch. 4.3 DATA PROCESSING The radar data was processed and analyzed using the RADAN 5.0 software. The data processing for the transient boundary case was similar to the no flow case. Since the main objective was to determine the height of the water level rise in the capillary zone, the GPR data needed to be processed only for the position correction 47

and for range gain. Position Correction shifted and aligned each trace of the radar signal to the first contact point of the antenna with the soil. Range gain filter enhanced the data by compensating for the signal attenuation with increasing depth and it is also important for the visual enhancement of the radargrams. 10 points linear gain was applied to the data set to enhance the data. In addition, display gain was applied for further visual enhancement of the radargrams. Figures 4.2(a) and 4.2(b) illustrates the data before and after the position correction. In Figure 4.2(a), top of the soil surface is represented by white-blackwhite bands in the radargram. The maximum amplitude of the first white band around 2.5 inches corresponds to the soil surface. Therefore the radar data is shifted upward by 2.5 inches which aligns the maximum positive amplitude of the first white band at the zero position. This corresponds to the first contact point of the antenna with the soil as illustrated in the Figure 4.2(b). The distinct white-blackwhite bands in the processed data shows the importance of gaining the display for better visualization. 48

Figure 4.2. (a) Raw radar data, and (b) Radar data after applying the position correction and display gain. 49

FIGURE 4.3. Soil sample in Transient boundary conditions showing the water level after few hours of the start of experiment. The raised water level can be easily observed from outside of the box, as illustrated by arrows in the figure. FIGURE 4.4. Radargram illustrating the raised water level. The blue line indicates the sloping water surface while the red line denotes the apparent depth of the bottom of the soil. 50

4.4 RESULTS AND DISCUSSION The radar data acquired over the successive periods of time showed a gradual rise in the water level with time. The water level could be observed through the Plexiglas box as shown in Figure 4.3. As explained in Section 3.4, the differences in the dielectric impedance contrast of the two medium (dry soil and moist soil) creates an interface which the GPR detected as soil-water boundary in this case. This interface appears as a distinct white-black-white band in the radargram somewhere between the top and the bottom of soil sample (shown by blue dashed line in Figure 4.4) indicating the presence of water in the capillary rise zone. The upward shift of this white-black-white band over time indicates the upward rise of the water over time due to capillary action in the soil. The water interface denoted by blue dashed line in the Figure 4.4 represents a sloping water interface. Since the height of water was maintained at 4 inch at the left side of the figure, the water level is nearer to the surface at this end in comparison to the right end where the water is draining out of the box. This sloping interface was observed to rise with time as a result of water rise in the capillary zone. The sloping interface in the radargram can also be easily recognized when compared with actual water level observed outside of the box in the Figure 4.3. The interface illustrated by a red line in the Figure 4.4 marks the bottom of the soil sample. The soil bottom appeared as a rising interface because it is the apparent depth observed in the radargram. As the water starts to rise and saturate the soil at the left end, the dielectric constant of the saturated portion increases which 51

slows down the traveling electromagnetic waves in this region. As a result the depth appears to increase in the radargram on the left side. Thus with the water interface sloping downwards from left to right, the apparent depth of the soil bottom appears to be sloping upwards. Figures 4.5 (a), (b), (c), (d) display a series of radargram showing the gradual rise in water level over time in soil A. The bottom of the soil sample, before adding water into the sample, is at a depth of 10 inch from the top and represented by two red arrows at the either ends in the Figure 4.5(a). As soon as water is added into the sample from the left end, water slowly seeped through the soil. After one hour, water moved though a horizontal distance of about 23 inches represented by point B in the Figure 4.5(b). The sloping interface represented by white-black-white bands between points A and B represent the dipping water surface while the segment between points B and C represents the bottom of the soil sample (the soil is still dry in this region). After 3.5 hours, the sloping white-black-white bands completely reach the right end at C (Figure 4.5(c)) indicating that water has reached the right end. Also the changes in the height of this interface shows the water was slowly rising (due to capillary action) which was detected by the GPR. After 22 hours of the start of the experiment, the white-black-white bands (showing the water surface) at the left end observed in the Figures 4.5(b) and 4.5 (c) disappeared and merged with another white-black-white bands at around 3 inch depth as shown in Figure 4.5(d). However, the white-black-white bands corresponding to the water surface still showed up at distance 15 inch from the left end at 6 inch depth as indicated by point D in Figure 4.5 (d). The disappearance of 52

white-black-white bands at the left side in Figure 4.5(d) can be explained in terms of the sizable wavelength of the electromagnetic waves which prevented the two reflection echoes from the top of the soil and water surface to exist as separate events as the water surface approached near the top soil surface. As illustrated in the Table 2.1, the wavelength for 1.5 GHz antenna was about 5 inches which means any feature within the five inch limit can not be easily resolved by the GPR data due to overlapping radar echoes. Clearly in our case the water level was at 4 inches depth on the left side which was within this wavelength. However, at point D, the waves were able to exist as separate reflection events and the soil-water boundary was visible after this point all the way to the right of Figure 4.5(d) until point C. Similar rising trends were observed in the radargram illustrated in the following Figures 4.9 and 4.11 for the soil types B and C, respectively. The GPR interpreted height of the water level showing rise in water level at different times is plotted in the graph, distance vs. height and shown in Figures 4.6, 4.8, 4.10 and 4.12. Each of the profiles in the figures represents the rising height of water level over different time periods. The height of the water level at the inlet side of the sample was kept constant at 4 inches throughout the experiment. However, over some time period, the GPR detected water level was found to be approximately 6 inches near the inlet side. This indicates that GPR is actually detecting water level in the capillary rise zone. Clearly, GPR was detecting the water level with saturation significantly below 100%. The uppermost profile shown by red line in each of the graphs indicates the height of the water surface when the sample was tested for determining the moisture content variation with depth. 53

(a) (b) (c) (d) FIGURE 4.5. Radargram illustrating GPR detected height along the length of the samples at different times in soil A.(a) Before introducing water into the sample (b) After 1 hour, (c) After 3.5 hours and (d) After 22 hours. 54

Water level Rise detected by GPR Soil type A (5% clay) 0.5 hr 1hr 1.5 hr 10 Height (inch) 2 hr 9 2.5 hr 8 3.25 hr 7 4 hr 6 4.75 hr 5 5.75 hr 4 6.5 hr 3 7.5 hr 2 8.28 hr 1 9 hr 0 0 5 10 15 20 25 30 35 40 45 9.75 50 hr Distance (inch) 10.5 hr FIGURE 4.6 GPR detected height along the length of the samples at different times for the soil A. The uppermost line denotes the height corresponding to the end of the experiment when the actual soil saturation level was measured. 55

(a) FIGURE 4.7. Radargram illustrating GPR detected height along the length of the samples at different times in soil B. (a) After 5 hours and, (b) After 24 hours. (b) 10 9 8 7 Water level Rise detected by GPR Soil type B (15% clay) 0.3 hrs 1 hr 1.66 hr 3.25 hr 4.5 hr 6 hr Height (inch) 6 5 4 3 2 1 7.5 hr 9.75 hr 11.5 hr 22 hr 24 hr 26 hr 27.5 hr 0 0 5 10 15 20 25 30 35 40 45 50 Distance (inch) FIGURE 4.8. GPR detected height along the length of the samples at different times for the soil B. The uppermost line denotes the height corresponding to the end of the experiment when the actual soil saturation level was measured. 56

(a) (b) FIGURE 4.9. GPR detected height along the length of the samples at different times in soil B. The uppermost line denotes the height corresponding to the end of the experiment when the actual soil saturation level was measured (a) After 1.5 hours and, (b) After 23.5 hours. 10 9 8 7 6 Water level Rise detected by GPR Soil type B (15% clay) 0.5 hr 1.5 hr 2.5 hr 3.5 hr 5 hr 6.5 hr Height (in) 5 4 3 2 1 8 hr 9.5 hr 11.5 hr 22 hr 23.5 hr 25 hr 27 hr 0 29 hr 0 5 10 15 20 25 30 35 40 45 50 Distance (in) FIGURE 4.10. GPR detected height along the length of the samples at different times for the soil B. The uppermost line denotes the height corresponding to the end of the experiment when the actual soil saturation level was measured. 57

(a) (b) (c) FIGURE 4.11. Radargram illustrating GPR detected height along the length of the samples at different times in soil C. The uppermost line denotes the height corresponding to the end of the experiment when the actual soil saturation level was measured (a) After 2.5 hours, (b) After 21.5 hours and (c) After 29 hours. 58