Omni-Directional Extension of the Refraction Microtremor Method

We propose and test an extension of the refraction microtremor (ReMi) method allowing us to extend our analysis from linear to two-dimensional arrays. Our “omni-directional” approach allows resolution of the azimuth of arriving ambient microtremor, as well as of lower frequencies and thus deeper velocities in areas where deploying a long linear array is not practical. We have found from our study of various array configurations that the specific array geometry (circle, triangles, rectangle, U, etc.) is not overly important. For this study, we analyze data collected with an approximately U shaped array, and with a triangular array. The U shaped array consisted of 48 geophones at 50-foot spacing and was recorded in 1997 near the Reno/Tahoe Airport. The data consisted of Rayleigh waves from background sources, mostly morning traffic from the airport and the nearby I-80/US-395 interchange. Originally, the data were used to measure the Rayleigh-wave dispersion values along separated east-west and north-south arrays. However, the U shape of the complete array is amenable to measuring wave speed in any desired direction. Using the omni-directional ReMi analysis, we can now analyze the speed of Rayleigh waves traveling in any direction. We have created a map of Vs30 (modeled average velocity from the surface to 30 m depth) values radiating out from the airport site to compare with the initial measurements. We also added the sum of p-f images for each analyzed directional vector to give an average Vs30 estimate for the area under the array. In comparison with an F-K dispersion curve published for a larger triangle array at the site in 2001, the omni-directional analysis of the U-shaped array finds correct phase velocities. As well, we find phase velocities to lower frequencies than could the standard linear-array ReMi analysis. Next, arrays of nested triangles provided data that we compare against P-S log data from a borehole drilled at the same location. The omni-directional ReMi results were also compared to a standard linear ReMi array recorded at the site. Two different triangular configurations of geophone arrays were acquired. We found that a larger triangle of fewer geophones provided better resolution of deeper velocities than a smaller triangle having additional closely spaced geophones. For both triangle arrays, summing the p-f plots from different directions provided very robust analyses of Vs30 and wave azimuth. The velocities compared very well to the estimates from the linear ReMi array. Omni-directional analysis of the two-dimensional (triangular and U-shaped) arrays allowed us to get deeper with a smaller spatial acquisition footprint than the linear setup. We are also exploring the aspect of better detection of lateral velocity variations using omni-directional analysis.