Soil Properties at Near-Surface Conditions Posters
Presiding: M Prasad, Colorado School of Mines; J Ajo-Franklin, Stanford University; S A al Hagrey, University of Kiel
NS41B-01 0830h
Improved Characterization of the Shallow Subsurface by Nonlinear AVA Analysis of High-Resolution Seismic Data
The focus of high-resolution shallow seismic has so far been on obtaining the geometrical or the structural information of the subsurface. Estimation of soil properties from surface seismic data has remained restricted to some spatially average values. However, it is the local properties, and not the average properties, that are more meaningful to the geotechnical, civil and environmental engineers. Laboratory estimation of soil properties, especially shear modulus and density, usually suffers from sample disturbance, and in-situ estimates are hard to obtain in loose, soft soils. The recent developments in data acquisition and processing technology provide the opportunity to extract soil information using the amplitude of shallow seismic reflections. Our research to extract the detailed soil information combines a reliable macro velocity model derived from seismic reflection data with the local interfacial velocity information obtained from angle-versus-amplitude (AVA) inversion. The approach first decomposes the multi-component data into P- and S- waves, which are then processed separately. Local density fluctuations are also estimated. The final model contains fine details that cannot be revealed by conventional velocity analysis. Our research shows that for large incidence angles and sharp velocity and density contrasts, which typify transitions in the shallow subsoil, a full nonlinear AVA inversion is essential for reliable local property estimation. The new approach has been tested on both synthetic and real data. The detailed information obtained can be translated to soil properties relevant to the engineers.
NS41B-02 0830h
Determining the State Properties in the Near-Surface Sands From a Seismic-CPT Integrated Model
The relative density and the effective horizontal stress are two critical state properties that determine the deformation and failure behaviours of sandy soils. They are, therefore, important to the engineers. However, in-situ estimation of these two properties in shallow, unconsolidated sand deposits is at present virtually impossible. This leads to risks and accidents. To solve this problem, we propose a new idea of integrating two very different in-situ field measurements based on an understanding of the underlying physical models. We have addressed the models of seismic shear wave velocity and the cone tip resistance obtained by geotechnical cone penetration testing (CPT). Although these two indepenent field measurements correspond to very different levels of strain, for a given type of sand they depend primarily on the relative density (and hence porosity) and the effective horizontal stress. We have constrained the porosity-stress models of elastic shear modulus in unconsolidated soils using the field observed seismic shear-wave velocity. For modelling CPT, which is a direct push geotechnical measurement, we have used a cylindrical cavity expansion approach. The reliability of these models have been verified on field data. Finally, we have integrated these two models to obtain unique estimates of the relative density and the effective horizontal stress, which is not possible from either of the two masurements alone. When tested on actual field data, our approach appears promising.
NS41B-03 0830h
Shear Seismic Anisotropy Within a Relay Ramp Structure, Baton Rouge Fault System, Louisiana
Shear wave data were acquired to characterize the fracture pattern at depth within a relay ramp structure associated with a Pleistocene Growth Fault system in Louisiana. By using both the degree and maximum direction of shear seismic anisotropy, we estimate the extent and orientation preference of subsurface fractures. Multi-source shear seismic data were generated by striking an I-beam, cut to 18 inches in length, from either side. Data collected in two control surveys show an expected 10-15% seismic anisotropy between fast and slow polarization directions, with the maximum anisotropy produced with the survey coordinate system being rotated parallel to the fault scarp. Data collected within the relay ramp structure indicate the principle direction(s) of stress, with depth, as well as density of fracturing. The results of this experiment should aid in regional fluid flow modeling and in local infrastructure planning such as residential construction, groundwater usage evaluation, and waste disposal site selection. This experiment has defined a successful technique that should be used when conducting similar studies in the region.
NS41B-04 0830h
Comparison of static and dynamic bulk moduli in unconsolidated sands
The dynamic and static bulk moduli of reconstituted samples of three dry, natural sands and one glass bead sample were measured over a number of pressure cycles from 0 to 20 MPa. The dynamic modulus, Kdyn, was calculated from ultrasonic compressional- and shear-wave velocity measurements, while the static modulus, Kstat, was measured from the corrected volumetric strains between pressure steps. For a given sample, the static bulk modulus demonstrates a great deal of variation based on the loading history of the sample, while the dynamic modulus is only slightly sensitive to the loading history. The Kdyn to Kstat ratio on the normally consolidated, initial loading path varies from between 2 and 10 for the various samples, and decreases slightly with increasing pressure for a given sample as the dynamic modulus rises faster than the bulk modulus. On the first unloading step of any pressure cycle the dynamic and bulk moduli are approximately equal, while with continued unloading the Kdyn to Kstat ratio rises from approximately 1 at the initial unloading step to near 3 at zero pressure for each of the samples. The significant variability in the Kdyn to Kstat ratio with pressure history indicates that a robust prediction of the static bulk modulus from dynamic measurements made in situ would require information on the loading history of the sample and on the current effective pressure. The Preisach-Mayergoyz space analysis was adapted to account for the effects of the plastic strains, in addition to those of the elastic hysteresis, on the relationship between the static and dynamic bulk moduli. Inaccuracies in the volumetric strain measurements lead to a quantitative mismatch between the dynamic modulus predicted from this analysis and that observed in the data. Nevertheless, this analysis does demonstrate the degree to which both the strain magnitude dependence of the static modulus and the occurrence of plastic strain contribute to the difference between the static and dynamic bulk moduli on the loading portions of the pressure path. On unloading, there is no plastic strain in the samples, so the strain magnitude dependence of the static modulus is the sole cause of the difference in the static and dynamic bulk moduli.
NS41B-05 0830h
Data-driven models for near-surface geophysical studies
Knowledge about porosity and permeability is essential to evaluate fluid content, contamination, and remediation success in soils. Information about subsurface formations is generally gathered at different scales, which vary in resolution, spatial coverage, and number of parameters measured. There is a need to up- or downscale to increase reliability of prediction. This paper will show applications of rock physics and experimental data to calibrate observations made in the near-surface geophysics surveys. We will present current data available and then show various possible applications of using this data and unconsolidated sediment models to understand soil properties. Experiments on sediments in controlled environment can allow us to develop empirical and theoretical trends and show the different petrophyical controls on near-surface geophysical signatures in sediments. We will show results of variations in, for example, porosity, cementation, pore-filling, and compaction, and how rock physics analyses can be implemented to diagnose unknown data sets, calibrate seismic data, for pressure prediction, and for modeling response of clay-sand mixtures.
NS41B-06 INVITED 0830h
Ground Truth for Geological and Geophysical Mapping of New York Harbor
As part of the project to deepen the harbors of the Ports of New York, Newark, and Elizabeth to 50 feet, we are mapping the geology of New York and New Jersey harbor, particularly the eastern edge of the Newark Basin, with geophysical measurements and core borings. The results are geological cross-sections, stratigraphic columns, and geological maps with resolutions approaching 1 foot. Sonar images map the surface of the bottom. Multichannel and single-channel, reflection seismic methods penetrate the sediment and rock to 100 feet below mean low water (MLW). The subsurface materials include black silt, gray silt, gray sands, Pleistocene till, Pleistocene varves, Jurassic diabase, Triassic sands and shales, Ordovician serpentinite, Cambro-Ordovician schist. We measure formation properties including lateral extent, thickness, strike and dip, contamination, fracture density, and diggability. The range of compressional velocity among these materials is a factor of six. Fundamental axioms of mapping, causality, equipresence, and frame indifference require sampling by core borings. Sampling is straightforward yet rarely required. We calibrate the water column for bathymetric studies and the subsurface for geotechnical studies. We measure Blow counts (or drill rate) Recovery efficiency Lithological description Digital photographs Rock quality designation (RQD) Fractures per foot Ultrasonic compressional wave velocity Unconfined compressive strength The measurements provide Ground truth Stratigraphic calibration Velocity modeling for time-to-depth conversion Material for correlation and experimentation The acoustic properties of the materials involved vary as a function of season. A sensitive issue is stability of the sediments in the channel slope. We place the hundreds of core borings into a single reference frame with the geophysical measurements. The geological maps and cross-sections follow from the interpretation of the processed geophysical results.
NS41B-07 0830h
The GeoModel at Kiel University: A Full Scale Model for Hydro- and Biogeophysical Studies in the Vadose Zone
Our GeoModel, a full scale tank analog (3x5x2 m3), is a new approach to develop novel integrated hydro-biogeophysical 4D techniques of high spatiotemporal resolution (in the range of cm and seconds). Fully controlled experiments are conducted at vadose soils simulating various structural, textural, lithological and hydrological properties in their natural scale. The GeoModel is a bridge between scaled laboratory models (typical size of 1 m3) and field surveys (several km3) and takes advantage of both. The GeoModel is equipped with a multitude of hydrogeophysical instruments including dc resistivity, GPR, TDR, tensiometer and tracers. A special device simulates varying irrigation scenarios of precipitation, ponding and drip of different rates, intensities and contaminations. A bottom filter pebble layer is divided into different segments for monitoring lateral distribution of flow. A self-developed vacuum suction device is installed to compensate for the hydraulic capillary barrier at the interface between the pebble filter layer and the overlying fine sand. The hardware and software allow complex 3D data acquisition, processing and inversion using fine electrode grids from all sides in numerous arrays, radar antenna configurations from surface and boreholes, and a distribution of hydrological probes for water content and tension. Performance tests showed the very high quality and resolution of the data and the capability of the GeoModel for new experiments to determine petro-hydrogeophysical parameters. We present examples of 3D mapping of fine structures and monitoring (4D) of water and dye tracer flow in the GeoModel. The inversion of reference data sets resolves cells with a length smaller than 1 cm. In many infiltration experiments the monitored radargrams and resistivity models image clearly the proceeding of a heterogeneous preferential flow front with time. Radargrams show the continuously increasing travel times from a reference reflector caused by increasing water content. Time dependent measurements enable us to monitor water flux. We established empirical calibration relationships similar to that of Archie and Topp for the substrate of the GeoModel and transformed the resistivity and velocity models into water content distributions.
NS41B-08 0830h
Geophysical and Geotechnical Investigations for Proposed Dominica Airport
The results of geophysical and geotechnical investigations carried out at the site of the proposed International Airport at Melville Hall, Commonwealth of Dominica, West Indies, are presented. The geotechnical investigation confirms the findings of the previous geophysical investigation, which concludes that bedrock is not likely to be encountered within the proposed depths of excavation. The stratigraphic models of both geophysical and geotechnical investigations are compatible and suggest that the soil profile is one of deeply weathered pyroclastic tuff and ash deposits transitioning to a boulder conglomerate ash horizon. The main geophysical method used was seismic refraction, additionally ground penetrating radar, resistivity sounding and resistivity tomographic imaging were also performed at some of the sites. Analysis of the seismic data shows a gradual increase in velocity with depth for which a model has been determined. Ancillary models or predictions of porosity, density, and (natural) compaction with depth are given, based on the basic seismic model. The main geotechnical investigative methods comprised of boreholes to 30 m depth with Standard Penetrating Testing (SPT) and undisturbed Shelby tube, and disturbed Split Spoon soil sampling. Water content, plasiticity, and grainsize distribution characteristics are obtained from laboratory testing leading to a classification of elastic silts and elastic silts with sand using the Unified Classification System. Geophysical and geotechnical data correlations are presented. Seismic velocity and SPT-N blow counts appear to be well correlated by a linear model. A model relationship between SPT and seismic dynamic elastic modulus is developed derived from seismic velocity. SPT-N is better correlated with the dynamic elastic modulus than with seismic velocity. The results show that seismically derived dynamic elastic modulus can accurately predict soil strength as measured by SPT blow counts.
NS41B-09 0830h
Complex Impedance of Swelling and Non-swelling Clay-Brine Suspensions
We carried out electrical impedance measurements over the frequency range of 1 Hz to 1 MHz on mixtures of swelling and non-swelling clay samples in NaCl solutions ranging from zero to 0.35 molar. The purpose of the measurements was to try to identify possible differences in the impedance response of the two clays under controlled conditions that could be exploited in the field. Measurements were carried out using a two-electrode liquid test fixture; the electrode gaps and applied voltages were varied to assure that these parameters did not affect the results. For the mixtures of 1g/L dispersed clay in the solution, no flocculation or settling occurred during the measurements. For both clays, the d.c. conductivity follows Archie's Law down to very low brine concentrations [0.001 molar - 0.01 S/m for Na-montmorillonite (swelling) clay; 0.0001 molar - 0.001 S/m for kaolinite (non-swelling) clay]. At lower NaCl concentrations the conductivity of the mixture is constant at a level consistent with the relative cation exchange capacity of each clay. For a dispersed clay mixture in brine, once the brine concentration exceeds the threshold conductivity, both the d.c. and complex conductivity is the same as that of a pure brine solution and the clay mixtures are indistinguishable. This change in conductivity behavior occurs at brine concentrations below those typically encountered in nature. This work was performed under the auspices of the U.S. Department of Energy by the University of California Lawrence Livermore National Laboratory under contract W-7405-ENG-48 and supported specifically by the Office of Basic Energy Science.
NS41B-10 0830h
Measurement and Modeling of the TDR Signal Propagation Through Layered Dielectric Media and of the Effective Permittivity of Sandy Soils
Layered dielectric materials are often encountered in the natural environment due to differences in water content caused either by a wetting or drying front. This is especially true for coarse-grained materials such as sandy soils, sediments and some rocks that have very distinctive layers of water content. This poster examines the issue of how the permittivity along a TDR probe is averaged as a function of layer thickness and probe orientation. Measurements of apparent permittivity, Ka, using TDR are presented for two, three and multi layer materials. TDR waveforms are modeled for multiple layers of varying thickness and show a change in the averaging of the apparent permittivity from refractive index to arithmetic when more thin layers are present. Analysis of the modeled results shows that the averaging regime is frequency-dependent. However, broadband techniques applied to materials with a few layers will generally produce refractive averaging. A transition to arithmetic averaging is found for systems having many (>4 layers). Narrow-band methods may be very sensitive to layering and may perform in a highly non-refractive way when layering with a strong permittivity contrast is present. Many empirical formulas relating TDR-measured permittivity, Ka, to volumetric water content have been proposed owing to the lack of a robust and accurate physically-based model describing this relationship across a range of soils. Soil-specific calibrations are often infeasible due to the time-consuming gravimetric sampling required for adequate calibration where limited resources often prevail. In this poster we propose a physically-based sample scale model for the permittivity - water content relationship in coarse-grained media using modeled or calibrated two-point anchoring. Materials tested include mono-size glass spheres and quartz sand grains in addition to two sandy soils. The physically-based model accounts for particle shape and bulk density using a two-phase, particle-scale mixing model and refers to a wetting or draining profile with a sharp wetting or drying front. Our measurements indicate the absence of dielectric hysteresis for the materials studied. An alternate calibration approach only requires the measured soil effective permittivities for dry and saturated conditions (i.e., two-phase mixtures) and knowledge of the bulk density. For the effective permittivity of dry sandy soils we recommend to use a common value of 3 for the solid phase permittivity. The results provide insight into the appropriate "refractive index" modeling of layered (wetting/drying) soil profiles with the grain-scale modeled two-phase permittivity providing bounds for the sample-scale three-phase porous medium.
NS41B-11 0830h
Saturation Profiles from Lab-scale Permittivity Measurements and 2-Phase Flow Models.
Laboratory fluid-flow experiments through soils are important to study the different fluid-flow processes that occur in the subsurface of the Earth. Most of them, measure the flow conditions at the entry and exit points of the sample, while, the saturation profile in the sample, and its change with time, is unknown. Until now, only very expensive and time-consuming techniques, such as CT or MRI scans, could provide information about the flow in the sample. Now, for the first time, we have developed a technique to measure the saturation profile of soil samples that is cheap and easy. With this technique, we obtain the saturation profiles from the inversion of permittivity measurements. The experimental set-up consists of a coaxial transmission line with a large sample holder (3 cm of diameter and 10 cm long) that allows for fluid-flow through the sample. The complex electric permittivity is reconstructed, per frequency, from the electromagnetic reflection and transmission responses of the line. Relative changes in the permittivity in the order of 1% can be detected over a wide frequency band up to 3 GHz, while the lowest usable frequency depends on the permittivity of the material filling the sample holder. The accurate measurements of permittivity performed with the described tool assume that the measured quantity is an effective property. The assumption then is that the sample is homogeneous. Under non-flowing conditions this can be a reasonable assumption, but during a flow experiment it cannot. This is clear from the fact that the reconstruction of the permittivity of these samples is not successful. The model of the sample must be modified to include the resulting heterogeneity. This is done by assuming a multilayered sample, each layer considered to be homogeneous with a constant permittivity and a known frequency dependence. Minimizing a cost function involving the measured reflection and transmission coefficients and the modelled response of the multilayered sample, we obtain the permittivity profile. The permittivity is mainly determined by the saturation, then, the permittivity profile within the sample is equivalent to the saturation profile. One can be obtained from the other using the Complex Refractive Index Model (CRIM). The saturation profile can also be obtained with traditional 2-phase flow modelling. Preliminary results show that the saturation profile obtained from the permittivity is more close to reality than the one inferred from 2-phase flow modelling.
NS41B-12 0830h
Factors Affecting Interfacial Polarization and Dielectric Behavior of Wet Soils
The presence of free charges and numerous discontinuities separating liquid, gas, and solid phases in partially-saturated soils give rise to a relatively understudied phenomenon of interfacial polarization that could impact bulk dielectric permittivity measurements using TDR and similar soil-embedded sensors. Evidence suggests a complex interplay between the response of complex dielectric permittivity spectrum to changes in ambient temperature and amount of ionic charges present in the wet soil. Model calculations based on Hanai's theory are supported by direct measurements using Network Analyzer showing that at low frequency (<10 MHz), the bulk dielectric permittivity increases with increasing temperature with a magnitude determined by the bulk electrical conductivity (amount of free ionic charges). Beyond a certain critical frequency, the decrease in free water permittivity dominates and the bulk dielectric permittivity decreases with increasing temperature. The dielectric permittivity inferred from TDR waveform travel time analyses is not significantly influenced by changes in the low frequency range (<10 MHz). In contrast, soil dielectric permittivity sensors operating at frequencies in the range of 10 MHz are likely to exhibit significant sensitivity to ambient conditions (temperature and electrical conductivity) affecting interfacial polarization, hence require special care in measurement interpretation.
NS41B-13 0830h
Capillary Sealing and Overpressure Regime in the Anadarko Basin
The Anadarko Basin in southwestern Oklahoma is known to contain today areas of extensive overpressures (pressures higher than hydrostatic pressure). Explaining the origin and maintenance of overpressured pore-fluids in the basin over long periods of time cannot be achieved by invoking classical, common causes, such as compaction disequilibrium or gas generation. We propose a capillary sealing mechanism that is responsible for both generating and maintaining almost all overpressure observed today in the Anadarko Basin. Capillary sealing occurs in a sedimentary basin when capillary forces act at gas-water interfaces between coarse- and fine-grained clastic rocks. Detecting capillary seals and estimating the magnitude of their pressure sealing implies two main aspects: (1) measuring the pore throat radius of coarse- and fined-grained clastic rocks, and (2) detecting the presence of gas-bearing layers using geophysical logs and other data. Measurements by injecting mercury into rock pores allow estimation of the pore throat radii controlling the capillary sealing. 21 fine-grained rock samples from the Anadarko Basin were thus measured and the average pore throat radius was found to be 2.5 x 10-8 m. The proposed model also requires the presence of gas-bearing layers interbedded into shale layers. Using a suite of geophysical logs from more than 100 wells, we were able to identify such gas-saturated layers in more than 50 wells. Further calculation indicates that a capillary sealing mechanism in the overpressured area of the Anadarko Basin may produce ~40 MPa of pressure, or ~80% of the maximum observed overpressure in the basin.