Near-Surface Geophysics Applied to Geologic Hazards II: Land Hazards
Presiding: S Kruse, University of South Florida; M A Smith, U.S. Department of the Interior Minerals Management Service
NS14A-01 15:30h
Three-Dimensional Georadar Surveying of Active Faults
For reliable seismic hazard estimates and knowledge of the fundamental processes that govern surface rupturing earthquakes it is necessary to determine the locations and geometries of active faults, including those buried beneath recently deposited sediments. Although excavated trenches may supply important details on hidden faults, the subsurface volumes that can be examined in this manner are generally small. In an attempt to provide paleoseismologists with a non-invasive tool that will allow them to investigate regions between and beyond sparsely spaced trenches, we have developed cost-effective 3-D georadar data acquisition, processing and interpretation techniques. These techniques have been applied with varying degrees of success to a range of active strike-slip, normal, thrust and transpressional fault zones. Our 3-D georadar data across the northern San Andreas fault in California revealed, unexpectedly, the existence of two fault strands. An offset of a linear-trending feature suggested that 4.5 5.5 m of horizontal displacement had occurred on one of these strands, either during the 1906 San Francisco earthquake or earlier. Somewhat surprisingly, two fault stands were also detected in 3-D georadar data acquired across the Wellington strike-slip fault in New Zealand (NZ). The first-ever fault-plane reflections from an active strike-slip fault were observed in these data. Georadar data collected within the Taupo Volcanic Zone (NZ) contained two prominent parallel reflections that originated from the boundaries of Late Pleistocene lacustrine and tephra deposits. Distinct vertical offsets of these reflections allowed us to estimate displacements at individual normal fault strands across the entire inner graben of the Maleme Fault Zone. The total displacement represented by these offsets was roughly 10-20% greater than that inferred from geomorphological studies, demonstrating the limitations of surface observations for determining cumulative fault movements. In the Canterbury Plains west of Christchurch (NZ), the results of 3-D georadar surveying not only confirmed the location of a fault zone, the existence of which had been postulated on the basis of subtle geomorphological features, but also demonstrated that it comprised multiple strands. Finally, 3-D georadar data allowed us to map the detailed geometry of several thrust planes of the Ostler fault system (NZ).
NS14A-02 15:45h
Near-Surface Geophysical Investigations of the Green Valley and Hayward Faults
We conducted seismic refraction and ground-penetrating radar (GPR) surveys at two active right-lateral strike-slip fault zones in the San Francisco Bay Area in order to gain a better understanding of near-surface geology and to provide data needed to constrain fault slip rates. Seismic refraction data were recorded using a 24-channel seismograph with a sledgehammer source. GPR data were recorded using 50 MHz antennae and 0.5 m trace spacing. The site on the Green Valley fault was selected due to its geomorphic expression of classic linear valleys and offset stream channels. The primary geophysical targets are buried paleochannels that can serve as piercing points for the determination of fault offset. Seismic refraction data indicate three layers in the near-surface zone: an uppermost layer ~1 m thick, with seismic P-wave velocity (VP) ~100 m/s; a second layer 5-7 m thick, with VP ~600 m/s; and the top of the third layer at 8-10 m depth, with VP ~1700 m/s. We also recorded a grid of several GPR lines at this site. Some of the GPR lines appear to show paleochannel features at ~2-6 m depths. We will attempt to sample these features with an auger. Cone Penetrometer Test (CPT) holes were logged at several locations along two of the lines where seismic and GPR data were recorded. This site will be trenched within the next year to obtain paleoseismic data that can be used to better constrain the slip rate of the fault. The second site is Tyson's Lagoon, a sag pond between parallel strands of the southern Hayward fault. This site has been extensively trenched with the goal of determining the slip history of the fault. Detailed trench logs, borehole logs and CPT logs are available. Preliminary refraction data indicate a top layer typically 6-10 m thick with VP ~300 m/s, underlain by material with VP ~1900 m/s. The top layer reaches its maximum thickness at ~10 m depth in the central portion of Tyson's Lagoon, between the fault traces, and shallows to ~1 m on the west side of the western fault trace. Two GPR lines that we recorded at this site appear to image the same surface detected by seismic refraction. The 1968 borehole and 2002 CPT logs clearly indicate the presence of a surface of significant material contrast at similar depths, and this surface has been inferred to be the contact between Pleistocene gravels and overlying Holocene deposits. We interpret the surface detected at similar depths through seismic and GPR to be the same contact. None of the trenches in Tyson's Lagoon were deep enough to detect this surface in the central portion of the pond. In this study, we use GPR and seismic refraction as complementary techniques in support of paleoseismologic investigations: GPR often provides a more detailed subsurface image, while seismic refraction is less susceptible to cultural disturbances and provides better depth control. We continue to refine our processes for employing these two geophysical methods together in an attempt to 1) determine optimal locations for proposed trenches, 2) correlate geophysical data with ground-truth data (trench, CPT, borehole), and ultimately 3) develop "virtual trenching" methods that use geophysical techniques to extend real trenches beyond their actual length and depth dimensions.
NS14A-03 16:00h
Downhole Measurement of Shear-Wave Splitting in Holocene Sediments: Application to Landslide Hazards in the Ottawa Valley, Canada
Stress-aligned inclusions (pore spaces and microfractures) cause seismic shear waves to exhibit directional polarizations in response to propagation through azimuthally anisotropic sediments. This phenomenon, called shear-wave splitting, is observable by measuring differences in shear-wave velocity between waves traveling parallel (S1 - fast shear wave) and perpendicular (S2 - slow shear wave) to the trend of the inclusions. This investigation provides evidence of shear-wave splitting in Holocene sediments of the Ottawa Valley (southern Ontario), Canada. Preliminary analysis of a multicomponent downhole data set, recorded using hammer and mass energy sources, indicates shear-wave splitting related to in-situ stresses associated with steep, landslide-prone, slopes bounding the South Nation River. Orientation of the split shear waves, determined from particle motion plots of the three-component downhole records, shows alignment of the polarization directions with the local topography and analysis of travel-time delays between S1 and S2 suggests anisotropy values of approximately 7% for the upper 30 m of the subsurface. Temporal monitoring of changes in shear-wave splitting may provide a method to evaluate factors related to landsliding, and assist in identifying potentially unstable slopes.
NS14A-04 16:15h
Ground Penetrating Radar Imaging of Tephra Fallout and Surge Deposits
GPR profiles on Cerro Negro volcano, Nicaragua, and Poás, Irazú, and Arenal volcanoes, Costa Rica, show this method has utility for mapping tephra blanket and surge deposit thicknesses, as well as ballistics distributions. These data are useful for estimating eruption volumes, particularly close to vents where deposits may be thicker than trenching depths. In the dry, highly resistive tephra of the Cerro Negro basaltic cinder cone, distinct deposits are clearly imaged between 2 and 20 m depth. The lowermost coherent reflection is presumed to be the contact with underlying pre-Cerro Negro lavas and weathered tephra deposits. Within the 2-20 m package, individual reflecting horizons are clearly resolved, and reflection attributes, particularly phase, may contain useful information on the nature of contacts, such as abrupt changes in granulometry. Because of the very high velocities at Cerro Negro (0.14 m/ns), even with 200 MHz antennas strata shallower than 2 m are difficult to resolve. In contrast, wetter ash, pumice, paleosol, and surge deposits on Irazú and Poás volcanoes show velocities as low as 0.045 m/ns. The corresponding shorter wavelengths permit strata as shallow as 40-70 cm to be imaged with 200 MHz antennas, with depth penetration typically 5 to 8 m. Comparison of trench observations and radar profiles indicates that strong radar reflections are produced by iron-rich zones at the water table and soil-ash contacts. Other features visible in the profiles are small (tens of cm) sub-vertical offsets of nearly horizontal units, and diffractions or disruptions in horizontal units presumed to reflect >30 cm blocks.
NS14A-05 16:30h
P-wave and S-wave Tomography for Characterizing an Impoundment Dam
A data set consisting of vertical and shear component seismic data was recorded on the crest of an impoundment dam. The data were recorded for the purpose of characterizing the integrity of the dam before transferal of the property to the public domain. Cone penetrometer results were also available for comparison. Four sets of 24 channels were used to record data from 3-component spiked geophones and vertical component data from gimbaled geophones in a parallel land streamer configuration. A sledge hammer with flat plate and spiked shear source was used for vertical, transverse, and inline source orientations. Left-right and front-back hits were used to generate the shear data. At each source location, five orientations of the source were used: vertical, transverse left-right, inline front-back. Thus the final data set consists of nine-component data. A spread consisted of 24 one meter spaced receiver locations. The source was moved through the stationary spread starting and ending two meters off the spread ends for 28 shots per spread. A total of three spreads were shot with spread move-up of one-half spread length. Data processing was done using a diving wave tomography approach and Rayfract software. The tomography approach uses first arrivals. First arrivals for the shear data were picked from differenced transverse shot gathers to remove compressional wave contamination. The transverse shear gathers were equalized before differencing. The resulting P-wave and S-wave velocity profiles were subsequently used to calculate a Poisson's ratio profile for the dam. The P-wave velocity profile shows zones of high velocity, approximately that of water, correlating with suspected zones of seepage. These zones occur at a depth of nine meters. Near the top of the P-wave profile, two low velocity anomalies and one higher velocity anomaly are apparent. No independent data exist to confirm the nature of these anomalies. The S-wave velocity profile does not exhibit the anomalous zones near the suspected seepage zones seen in the P-wave profile. One of the cone penetrometer locations occurs at the center of the S-wave profile. The S-wave profile exhibits somewhat higher velocities at this location that generally corresponds with higher cone penetrometer S-wave values. A Poisson's ratio profile was calculated from the P-wave and S-wave profiles. Although, the Poisson's ratio profile contains a large number of dubious values, two zones show values that correlate well with identifiable velocity zones in the P-wave and S-wave profiles.