NS22A-01 INVITED
Soils and Processes - Wave-based Characterization
The complex nature of soils and their intricate behavior in the context of engineering applications remain challenging after a century of extensive research and engineering advances. A particle-level review of these materials reveals the underlying mechanisms that are responsible for their macro-behavior in the context of engineering applications and permits identifying engineering properties that can be inferred from geophysical measurements. The presentation starts with a review of fundamental properties of near surface geomaterials including the different physical processes in fine and coarse grained soils, the relevance of the state of stress, unsaturation, cementation, and spatial variability. Then, the complementary nature of elastic waves and electromagnetic waves in near-surface characterization is analyzed. Applications include liquefaction monitoring, pressure diffusion, chemical diffusion, diagenetic cementation and and tomographic imaging the state of stress.
NS22A-02 INVITED
Combined Geoelectrical and Georadar Measurement for State Characterization of porous Rock
The state parameters volumetric pore water content and pore water composition influence geomechanical stability of near surface unconsolidated rock or soil. Changes of those characteristics can result in the instability of the rock and therefore in on set of a failure process. The geophysical monitoring of state parameters in space and time allows the assessment of compaction or soil water suction/pressure. The objective of the thereafter presented investigation was the quantitative determination of water saturation and water salinity using multimethod geophysical measurement. The application of only one geophysical method can give rise to gross uncertainties in the estimation of salinity or water saturation. The combination of a low frequency conductivity measurement (2 Hz) and a high frequency electromagnetic measurement (1 GHz) provides two petrophysical parameters: electrical conductivity and dielectric permittivity. Both parameters are strongly water saturation dependent and somewhat dependent on water salinity. A system of two nonlinear model-equations was used to determine salinity and saturation. An unique solution is possible in case of constant pore space structure. The experiments have been carried out on a rectangular designed sand box model and a hydraulically isolated sandstone block. Each of that has a size of 2m x 1m x 0.3 m. Three types of medium scale hydraulic experiments were monitored by multimethod geophysical measurements: - Water imbibition and drainage, - Displacement of water by density driven flow, - Displacement of water by forced convection 4-point electric conductivity measurement and radar transmission measurement have been carried out along several vertical profiles. The data were used to test the area of validity of the petrophysical model. The block model were divided into several descrete rock volumes. Water saturation and salinity were calculated for each single discrete rock volume. Independently water balance and hence the mean water saturation in the model was controlled by in- and outflow measurements. The mean salinity was estimated after the end of the experiment in the drained water outside the block model. The results show that a quantitative determination of saturation and salinity with sufficient accuracy is possible by using combined geophysical measurements. Parameter variation in space and time can be monitored with adequate resolution.
<a href='http://www.dgfz.de'>http://www.dgfz.de</a>
NS22A-03
Using Resistivity Cone Penetrometry (RCPT) to Constrain 2D and 2.5D Inversion of Electrical Resistivity Tomography (ERT) Field Data
Sand and gravel bodies in clay-rich till alter its engineering properties, and act as significant water flow routes. Electrical Resistivity Tomography (ERT) can detect these high-resistivity sand/gravel bodies; however, commonly-used cell-based ERT inversion routines reconstruct sharp geoelectrical boundaries poorly. In contrast, the invasive technique of Resistivity Cone Penetrometry (RCPT) can sample resistivity at centimetre vertical resolutions, although lateral resolution is low (i.e. distance between RCPT sampling points). We use synthetic and real data to investigate the potential for combining the two methods by using RCPT to constrain ERT inversion. A synthetic resistivity model of a high resistivity sand/gravel lens in a low resistivity clay till was constructed. Synthetic Wenner ERT field data were generated and contaminated with 100 sets of Gaussian noise with standard deviation of 2% and 5%. The ERT data were inverted with the University of British Columbia Geophysical Inversion Facility code DCINV2D. We inverted with reference models containing lenses and blocks of similar depth and thickness to the lens in the synthetic model but of variable width, as well as a homogenous reference model. Analysis showed that i) using a reference model that is close to the true model produces a final model that is close to the true model and ii) using a reference model that is significantly different to the true model results in a final model that is neither close to the true model nor duplicates the reference model. The final models were ranked by the sum of i) the least-squares data misfit, ii) the least squares misfit between the true model and the reference model and iii) the final model flatness. Ranking correctly identified the final models with the closest fit to the true model without a priori knowledge of the true model. ERT, RCPT and EM data were collected from a coastal site in Yorkshire, UK, where sand lenses in clay till were known to be present from cliff exposure. 2D Reference models were constructed with resistivity values and layer thicknesses based on RCPT data, and a range of resistivity anomaly widths based on EM data. The 2D inversion results were assessed qualitatively, by comparison to cliff exposures, and quantitatively, by using the methods identified by modelling. The sum of the least squares data misfit and the least squares model misfit allowed identification of the final model with the closest fit to the true geoelectrical structure. We extend this approach by carrying out 2.5D inversion of parallel lines of Wenner ERT data from the same field site using a range of 3D reference models derived from the RCPT. We discuss the implications for the design of combined ERT and RCPT investigations.
NS22A-04
Electrical Response measurements and models of Sand-clay mixtures under stress.
Knowledge of the micro-structural changes in unconsolidated earth materials subjected to effective stress is of interest to geotechnical engineers and earth scientists. We analyze the effects of stress on the mechanical and electrical properties of sand and clay mixtures. Measurements of the spectral electrical responses (SER) and the physical as well as petrophysical properties of saturated clay-sand mixtures, and subjected to varying effective stress levels (5-25MPa)are performed in a laboratory environment. Relations between micro-structural changes in sand-clay mixtures due to stress changes and modifications in electrical parameters describing their spectral electrical responses are exploited and anlyzed. Spectral electrical response measurements of snad-clay mixtures are described with an equivalent circuit model which considers the mixtures a heterogeneous multi-component system. The validity and usefulness of the relations between the electrical parameters and the soil textural and mechanical properties are assessed using laboratory measurements of the spectral electrical response(0.01Hz to 10 kHz) of the mixtures under varying stresses. Other useful information from the SER and how they influenced by the micro-structural characteristics of the mixtures are also investigated.
NS22A-05
Decomposition of Spatially-Distributed Soil Moisture Into Multiple Component Variables With Applications to Interpolation.
The spatial variation of soil moisture is an important state variable in hydrology, ecology, agriculture, and climatology. Because collection of high resolution soil moisture data is difficult, it would be desirable to collect only sparse soil moisture observations and interpolate to higher resolutions. Interpolation of soil moisture is also difficult because soil moisture patterns result from a complex interaction of multiple hydrological processes and can exhibit distinct relationships to site properties depending on the season and the spatial scale. Numerous studies have noted the limitations of traditional regression techniques to estimate soil moisture because they treat soil moisture as a single random variable, which presumably displays consistent behavior in space. Any variation from the regression behavior is considered to be random error. However, it is also possible that deviations from the regression are not random errors but are deterministically associated with a different and independent physical process. In this study, it is shown that at a given time, soil moisture can be represented as multiple superimposed patterns or likewise as a multivariate spatial distribution. The component patterns are associated with different controlling processes. Ground-based soil moisture measurements from the Tarrawarra Catchment in Australia are used. Previously, it has been shown that soil moisture in this catchment is simultaneously controlled by lateral drainage and evapotranspiration (ET), with the relative importance of each process varying seasonally. Here, Empirical Orthogonal Function (EOF) analysis is used to decompose the soil moisture data into temporally constant component spatial patterns (EOFs) and time-varying weights (expansion coefficients). Two statistically significant EOF patterns are identified that account for 64% of the observed spatial variation over the sampling period, and these EOFs are found to be related to lateral drainage and ET. Soil moisture on any sampling date is shown in part to consist of a linear combination of these two patterns, yet the EOF patterns exhibit more consistent relationships with site properties (i.e. topographic properties) than soil moisture itself. Using sample variograms, it is also shown that the EOF patterns exhibit a more consistent spatial structure than soil moisture. The implication is that the EOFs can be more accurately interpolated from sparse soil moisture data and detailed site properties than soil moisture itself. The interpolated EOFs can then be used to reconstruct the soil moisture pattern. This interpolation method is demonstrated, and the results are compared to traditional techniques.
NS22A-06
Hydrogeophysical Characterization of a Texas Expansive Heavy Clay Soil (Vertisol) Using Electrical Resistivity Measurements
We applied field and laboratory electrical resistivity measurements to characterize seasonal wetting and drying of a Texas Vertisol. Vertisols are complex soils characterized by high clay content ($>$30%), high shrink/swell potential, microrelief expression of subsurface soil dynamics (gilgai), and low hydraulic conductivity. A total of thirty-three multi-electrode resistivity profiling lines were taken along the same profile between April 27, 2005 to February 26, 2006 using combined dipole-dipole and Schlumberger electrode configurations, at electrode spacings of 0.5 m. The field site is located at the USDA - Grassland Soil and Water Research Station, near Riesel, Texas. The profiles were 17.5 m long and intersected two sets of microhighs and microlows of the gilgai. Inversion of resistivity data was done by a rapid least-square technique. In order to constrain and evaluate field results, in-situ measurement of soil-moisture was made using impedance probes and auger sampling. We designed a laboratory experiment to measure variations of soil moisture with resistivity using soil samples. We also made in-situ measurement of crack depths, and the results were incorporated into the laboratory and field data. The laboratory results were used to estimate petrophysical parameters for the Vertisol using a modified Archie's relationship, and to calibrate field results. Results of field surveys show that the upper 1.4 m of the Vertisol can be divided into three distinct soil moisture regimes: the upper zone (0 to $\sim$0.5 m) which is the most dynamic with regards to wetting and drying in the Vertisol; the middle zone ($\sim$0.5 m to $\sim$1.1 m) which is relatively saturated and less dynamic, and the lower zone (below $\sim$1.1 m) which is relatively less saturated compared to the middle layer. Seasonal hydrodynamics of the Vertisol is a complex interrelationship of by-pass flow through cracks, microrelief variability and hydraulic conductivity of the soil. We attribute gilgai formation to periodic expansion of the "wetter" depth interval in a manner analogous to formation of mullions in contracting sedimentary sections.